The Oxford Handbook of African Languages 9780199609895, 0199609896
This book provides a comprehensive overview of current research in African languages, drawing on insights from anthropol
148 76 13MB
English Pages 1104 Year 2020
Polecaj historie
Table of contents :
Cover
The Oxford Handbook of AFRICAN LANGUAGES
Copyright
Contents
Acknowledgments
List of Maps, Figures, and Tables
List of Abbreviations
List of Contributors
Part I: INTRODUCTION
Chapter 1: Introduction
1.1 Part I: Introduction
1.2 Part II: Domains of grammar
1.3 Part III: Language comparison
1.4 Part IV: Language phyla and families
1.5 Part V: Language structure: case studies
1.6 Part VI: Languages, cognition, and culture
1.7 Part VII: Language and society
1.8 Part VIII: Language and history
1.9 Part IX: Language and orature
Part II: DOMAINS OF GRAMMAR
Chapter 2: Phonology and phonetics
2.1 Introduction
2.2 Autosegmental phonology
2.3 Segmental and prosodic inventories
2.4 Vowel harmony
2.5 V-V and C-C clusters
2.6 syllabification
2.7 Reduplication
2.8 Phonological phrasing
2.9 Conclusion
References
Chapter 3: Tone
3.1 Introduction
3.2 Levels and contours: the primitives of tone
3.2.1 Levels
3.2.2 Contours
3.3 Tone features
3.4 Downstep and upstep
3.5 What tones do that is special
3.5.1 Local versus long-distance assimilation and shift
3.5.2 Melodies
3.5.3 Segmental influences
3.6 Summary
References
Chapter 4: Morphology
4.1 Introduction
4.2 Morphological typology of african languages
4.2.1 Isolating languages
4.2.2 Agglutinating languages
4.2.3 Fusional languages
4.3 Salient concepts in african morphology
4.3.1 Derivation
4.3.1.1 Inter-word class derivation
4.3.2.2 Intra-word class derivation
4.3.2 Compounding
4.3.3 Inflection
4.3.4 Reduplication
4.3.5 Morphological classes
4.4 Historical morphology
References
Chapter 5: Syntax
5.1 Introduction
5.2 The sentence: basic constituent order
5.3 The noun phrase
5.4 Pronouns
5.5 Adpositions
5.6 Adjectives and adverbs
5.7 Relative clauses
5.8 Multiclausal constructions, serial verbs and auxiliaries
5.9 Topic and focus
5.10 Question formation
5.11 Passive and raising
5.12 Conclusion
References
Part III: LANGUAGE COMPARISON
Chapter 6: African language types
6.1 Introduction
6.2 Previous studies in african comparative typology: a selective foray
6.3 Salient typological features
6.3.1 Phonological characteristics
6.3.2 Morphosyntactic characteristics
6.3.2.1 Noun class systems
6.3.2.2 Case marking
6.3.2.3 Verbal extensions
6.3.2.4 Serial verb constructions
6.4 Concluding remarks
References
Chapter 7: Dialectology and linguistic geography
7.1 Introduction
7.2 Dialectology
7.2.1 Introduction to methods in dialectometry
7.2.2 Dialectometry analysis and African languages: the case of Berber lexis
7.2.2.1 The linguistic atlas of the Rif as a data source
7.2.2.2 Data matrix of Rif-Berber lexis
7.2.2.3 Distance matrix of Rif-Berber lexis
7.2.2.3.1 Binary distance (Bin)
7.2.2.3.2 "Gewichteter identitatswert" distance (GIW)
7.2.2.3.3 Levenshtein distance (Lv)
7.2.2.4 Numerical dialectometry analyses of Rif-Berber lexis
7.2.2.5 Visualization of dialectometry analyses of Rif-Berber lexis
7.2.3 Contrastive results
7.3 Linguistic geography
7.4 Conclusion
References
Chapter 8: Reflections on the history of african language classification
8.1 Introduction
8.2 Types of classification
8.3 Language data, large scale collections, and first attempts at classification
8.4 The age of enlightenment romanticism, and the rise of comparative linguistics
8.5 The nineteenth century
8.6 The twentieth century
References
Part IV: LANGUAGE PHYLA AND FAMILIES
Chapter 9: Niger-congo, with a special focus on benue-congo
9.1 What tdo we mean by niger-congo?
9.2 Language relationships within niger-congo languages
9.2.1 The history of Niger-Congo classification
9.2.2 Subgrouping within Niger-Congo
9.2.3 What proves relationships in Niger-Congo?
9.3 Benue-congo: niger-congo's largest subgroup
9.3.1 Delineating Benue-Congo and Bantu
9.3.2 Benue-Congo subgroups
9.4 Niger-congo language typology
9.4.1 The areality of major Niger-Congo features
9.4.2 The typology of Benue-Congo languages
9.5 Advancing the comparative picture
References
Chapter 10: Atlantic
10.1 Overview
10.2 Atlantic as a geographically and typologically motivated grouping
10.3 Chances for the comparative method?
10.4 A different basis for atlantic?
10.5 Current research and a way foreward
Quoted web pages
References
Chapter 11: Mande
11.1 Introduction
11.2 Historical sources and early scholarship
11.3 External and internal classification
11.4 Linguistic reconstruction
11.5 Language contact and typology
References
Chapter 12: KWA
References
Chapter 13: Gur
13.1 Introduction
13.2 Linguistic approaches
13.3 Classification
13.4 Some limguistic characteristics
13.4.1 Phonemes, tones, syllables
13.4.2 Word order
13.4.3 Noun phrase
13.4.4 Verb phrase
Acknowledgements
References
Chapter 14: Bantu and bantoid
14.1 Introduction
14.2 Early studies of bantu languages, the recondtruction of proto-bantu, and twentieth-century developments
14.3 Internal classifications
14.4 Bantu dispersals
14.5 Bantu and bantoid
14.6 Non-genetic classifications
References
Chapter 15: Adamawa
15.1 Adamawa as alinguistic unit
15.2 Adamawa languages and groups
15.3 Classification
15.4 Linguistic characteristics
15.4.1 Verb
15.4.2 Word order
15.5 Conclusion
Acknowledgment
References
Chapter 16: Ubangi
16.1 Introduction
16.2 Language documentation and classification
16.3 Grammatical features
References
Chapter 17: Kpodofanian
17.1 Introduction
17.2 Kodofanian languages: A BRIEF OVERVIEW
17.3 Kodofanian studies
17.3.1 The challenge of studying Kordofanian languages
17.3.2 The main types of approaches
17.3.2.1 First testimonies: before World War I
17.3.2.2 The missionaries’ linguistic work
17.3.2.3 The surveys
17.3.2.4 The study of individual languages
17.4 Some salient features of kordofanian languages
17.4.1 Phonology
17.4.2 Morphology
17.4.2.1 Noun classes
17.4.2.2 Verb extensions
17.4.3 Lexicon
17.4.3.1 Heibanian and Katloid
17.4.3.2 Heibanian and Rashadian
17.4.3.3 Heibanian and Talodian
17.4.3.4 Heibanian and Lafofa
17.4.3.5 Other comparisons between Kordofanian families:Katloid Rashadian and Talodian
17.4.3.6 Proto-Kordofanian roots
17.5 Conclusion and prospects
17.6 Appendices: worldlists of cognates between various kordofanian families
17.6.1 Appendix 1: Plausible lexical cognates between Heibanian and Katloid
17.6.2 Appendix 2: Plausible lexical cognates between Rashadian and Heibanian
17.6.3 Appendix 3: Plausible lexical cognates between Talodian and Heibanian
17.6.4 Appendix 4: Plausible lexical cognates between Lafofa and Heibanian
17.6.5 Appendix 5: Plausible lexical cognates between Katloid, Rashadian and Talodian
Acknowledgments
List of abbreviations
Abbreviated references
References
Chapter 18: Afro-asiatic overview
References
Chapter 19: Egyptian
19.1 Introduction
19.2 Language history
19.3 Phonology
19.4 Morphosyntax
19.5 Lexicon
References
Chapter 20: Berber
Acknoledgment
References
Chapter 21: East Cushitic
21.1 Background
21.2 EC as cushitic par excellence?
21.3 The internal organization of EC
21.4 HEC
21.5 Disentangling LEC
21.6 Southern LEC: omo-tana and oromoid vs. dully and yaaku
References
Chapter 22: North cushitic
22.1 Introduction
22.2 The history of the classicication of beja
22.3 Lexicostatistic studies
22.4 Etymology and the comparative method
22.5 The issues of language contact
22.5.1 Contact with Arabic
22.5.2 Contact with NS
22.6 Dialectology and the internal classification of Beja
References
Chapter 23: Central cushitic
23.1 Introduction
23.2 Salient phonological features
23.2.1 Presence of /ŋ/
23.2.2 Presence of /z/
23.2.3 Presence of /ɣ/
23.2.4 Presence of central vowels
23.2.5 Lack of ejectives
23.2.6 Lack of implosives
23.2.7 Lack of pharyngeals
23.2.8 Lack of /ɲ/
23.2.9 Lack of vowel length
23.2.10 Lack of consonant gemination
23.3 Salient Morphological features
23.3.1 Personal pronouns
23.3.2 The unmarked nominative/marked accusative case
23.3.3 Agreeing genitives
23.4 Constituent order
23.5 Lexical similarity
23.6 Concluding remarks
References
Chapter 24: South Cushitic
24.1 Geographical didtribution and speakers
24.2 Sources
24.3 External and internal classification
24.4 Internal Classification
24.4.1 Typological profile of PSC
24.4.2 Internal development of the modern WRlanguages from PSC
24.5 The history of SC as a series of language contacts
References
Chapter 25 Omotic
25.1 Introduction
25.2 History of classification
25.2.1 Pre-“Omotic” endeavors
25.2.2 The establishment of “Omotic”
25.2.3 Post-“Omotic” debates
25.3 Outlook
References
Chapter 26: Chadic
26.1 Introduction
26.2 A survey of chadic languages
26.2.1 List and classification of languages
26.2.2 Geographic situation
26.2.3 Number of speakers
26.3 The typical chadic language
26.3.1 Phonology
26.3.2 Morphology
26.3.3 Sentence structure and function marking
26.4 Chadic and afro-asiatic
26.5 The chadic family and linguistic areas
26.5.1 Africa as a linguistic area
26.5.2 The Macro-Sudan belt
26.2 Conclusion
References
Chapter 27: Ethio-semitic
References
Chapter 28: Nilo-saharan and its limits
28.1 Introduction
28.2 Some diagnostic features
28.2.1 The n/k alternation
28.2.2 The causative
28.2.3 The reflexive (middle voice) marker
28.2.4 Pronominal reference
28.2.5 A stable preposition
28.2.6 Postpositions and case marking in Northeastern NS
28.2.7 Some further grammatical cognates
28.3 Some Final Observations
References
Chapter 29: Saharan
29.1 Introduction
29.2 South overview of research history
29.3 Distribution of the saharan languages
29.4 Common saharan genetic features
29.4.1 Common Saharan person elements
29.5 Common saharan verb structure
29.5.1 Plurality
29.5.2 Common features shared by the Saharan languages
29.6 Summary
References
Chapter 30: Eastern sudanic
30.1 Introduction
30.2 Phonological variation
30.3 Morphosyntactic variation
30.3.1 Number marking on nouns
30.3.2 Constituent order within the noun phrase
30.3.3 Pronominal reference
30.3.4 Constituent order and head marking versus dependent marking at the clausal level
References
Chapter 31: Central sudanic
31.1 Introduction
31.2 Inventory
31.3 Lexical evidence
31.4 Sound correspondences
31.5 Morphology
31.6 World order
31.7 Conclusion
References
Chapter 32: Khoisan
32.1 Introduction
32.2 Surviving languages
32.3 Field studies: dictionaries and grammars
32.4 Phonology and instrumental studies
32.5 Historical studies and classification
32.6 Digitization
32.7 Sociolinguistics
32.8 The current situation
References
Chapter 33: Linguistic isolates
References
Part V: LANGUAGE STRUCTURES: CASE STUDIES
Chapter 34: Bom-Kim
34.1 Introduction
34.2 Phonology
34.3 Word categories and morphology
34.3.1 Nouns, nominal morphology, and pronouns
34.3.2 Verbs and inflectional morphology
34.3.3 Negation
34.3.4 Verb extensions
34.3.5 Adpositions
34.3.6 Adjectives and other minor categories
34.4 Syntax
34.4.1 Word order, the verb phrase: SVO ~ S–AUX–OP–V
34.4.2 Relativization and focus
34.4.3 Questions
34.4.4 Subordinate clauses
Acknowledgments
References
Chapter 35: Dan
35.1 Introduction
35.2 Phonology
35.3 Morphology
35.3.1 Nouns
35.3.1.1 Locative nouns
35.3.1.2 Relational and free nouns
35.3.1.3 Plural marking
35.3.2 Adjectives
35.3.3 Numerals
35.3.4 Pronominal predicative markers (PPM)
35.3.5 Noun phrase structure
35.3.5.1 Genitive-like noun phrases
35.3.5.2 Attributive noun phrases
35.3.5.3 Coordinative noun phrases
35.3.6 Pronouns
35.3.6.1 Personal pronouns
35.3.6.2 Interrogative pronouns
35.3.6.3 Demonstrative pronouns and adverbs
35.3.7 Verbs
35.3.7.1 Lability
35.3.7.2 TAM and polarity
35.4 Syntax
35.4.1 Simple sentences
35.4.2 Complex sentences
References
Chapter 36: Biali
36.1 Introduction
36.2 Phonology
36.2.1 Vowels
36.2.2 Consonants
36.3 The noun phrase
36.3.1 Noun classes
36.3.2 Nominal derivation
36.3.2.1 Nouns from verbs
36.3.2.2 Nouns from nouns
36.3.2.3 Derivation by noun reduplication
36.3.3 Nominal compounding
36.3.4 Noun phrases
36.3.5 Nominal qualifiers
36.3.6 Pronouns
36.3.6.1 Allocutive pronouns
36.3.6.2 Substitutive pronouns
36.3.6.3 Interrogative pronouns
36.3.7 Numerals
36.4 The verb
36.4.1 Verbal derivation
36.4.2 Aspect
36.4.2.1 Formal markers
36.4.2.2 Tone markers
36.4.3 Tense
36.4.4 Mood
36.4.5 Negation
36.4.6 Focalization
36.4.7 Adverbs
References
Chapter 37: Yukuben
37.1 Demographic and socio-geolinguistic background
37.2 Phonology
37.2.1 Syllable structure
37.2.2 Vowels
37.2.3 Vowel sequences—the distribution and phonotactics of vowels
37.2.4 Consonants
37.2.5 Phonological alternations and processes
37.3 Morphology
37.3.1 Noun
37.3.1.1 Inflection
37.3.1.2 Derivation and compounding
37.3.2 Adjectivals and associatives
37.3.3 Numerals
37.3.4 Pronouns
37.3.4.1 Personal
37.3.4.2 Possessive
37.3.4.3 Relating concord
37.3.4.4 Interrogative
37.3.5 Demonstrative
37.3.6 Verb
37.3.6.1 Verb phrase
37.3.6.2 Tense, aspect, mood
37.3.7 Adverbials
37.4 Syntax
References
Chapter 38: Bende
38.1 General information
38.2 Phonology
38.2.1 Sounds and phonemes
38.2.2 Tones
38.2.3 Glide, vowel coalescence, and nasalization
38.3 Morphology
38.3.1 Nominal morphology
38.3.1.1 Nouns
38.3.1.2 Adjectives
38.3.1.3 Numerals
38.3.2 Verbal morphology
38.4 Syntax
38.4.1 Word order
38.4.2 Major sentence types
38.4.2.1 Simple sentences
38.4.2.2 Complex sentences
References
Chapter 39: Waja
39.1 Introduction
39.2 Phonology
39.2.1 Consonants
39.2.2 Vowels
39.2.3 Tone
39.3 Morphosyntax
39.3.1 Noun and noun phrase
39.3.2 Concord
39.3.3 Adjectives
39.3.4 Personal pronouns
39.3.5 Verb and verb phrase
39.3.5.1 The verb forms
39.3.5.2 Motion and posture verbs with copy pronouns
39.3.5.3 Verb extensions
39.3.5.4 The tense–aspect–mood system (TAM)
39.3.5.5 Use of the definite verb form
39.3.5.6 Use of the indefinite verb form
39.3.5.7 Use of the verbal noun
39.3.5.8 Past marker na
39.3.5.9 Non-verbal predication
39.3.5.10 Negation
Acknowledgments
References
Chapter 40: Zande
40.1 Introduction
40.2 Phonology
40.3 Word formation
40.4 The noun phrase
40.5 The verb phrase
40.6 Number and numerals
40.7 Gender
40.7 Syntax
40.8.1 Floating determiners
40.8.2 Secondary predicates
40.8.3 Complementizers
40.8.4 Negation
40.8.5 Grammar of space
References
Chapter 41: Zenaga
41.1 Introduction
41.2 Phonology
41.2.1 Vowels
41.2.2 Consonants
41.2.2.1 Voicing opposition
41.2.2.2 Emphasis opposition
41.2.2.3 Length opposition
41.2.3 Syllable structure
41.3 Morphology
41.3.1 Nouns
41.3.1.1 Inflection
41.3.1.2 Derivation
41.3.1.2.1 Deverbal nouns
41.3.1.2.2 Diminutives
41.3.2 Nominal determiners
41.3.3 Adjectives
41.3.4 Numerals
41.3.5 Pronouns
41.3.5.1 Personal
41.3.5.1.1 Independent
41.3.5.1.2 Suffixal
41.3.5.2 Possessive
41.3.5.3 Demonstrative
41.3.6 Relative
41.3.7 Interrogative
41.3.8 Verbs
41.3.8.1 Inflection
41.3.8.2 Derivation
41.3.8.3 Tense, aspect, mood
41.3.9 Adverbials
41.4 Syntax
41.4.1 Word (constituent) order
41.4.2 Major sentence types
41.4.2.1 Simple sentences
41.4.2.2 Complex sentences
41.4.2.2.1 Coordination
41.4.2.2.2 Subordination
41.4.2.3 Negation
41.4.2.4 Questions
41.4.2.5 Other important types
41.4.3 Noun phrase structure
41.4.4 Clause structure
References
Chapter 42: Sidaama
42.1 Introduction
42.2 Phonology
42.2.1 Segmentals
42.2.2 Morphophonemic rules
42.2.3 Suprasegmentals
42.2.3.1 Syllables
42.2.3.2 Pitch accent
42.2.3.3 Intonation
42.3 Parts of speech
42.3.1 Open classes
42.3.1.1 Nouns
42.3.1.2 Verbs
42.3.1.3 Adjectives
42.3.1.4 Adverbs
42.3.2 Closed classes
42.4 Morphology
42.4.1 Affixes and reduplication
42.4.2 Morphological case-marking system
42.4.3 Suffix orders
42.5 Syntax
42.5.1 Grammatical relations
42.5.2 Word order
42.5.3 External possessor constructions
References
Chapter 43: Kolisi
43.1 bIntroduction
43.2 Phonology
43.2.1 Vowels
43.2.2 Consonants
43.2.3 Syllable structure
43.2.4 Tone
43.3 Morphology
43.3.1 Pronouns
43.3.1.1 Personal pronouns
43.3.1.2 Demonstrative pronouns
43.3.1.3 Reflexive pronouns
43.3.2 Nouns
43.3.3 Verbs
43.3.3.1 Negation
43.3.3.2 Interrogation
43.3.3.3 Passive
43.3.3.4 Causative
43.3.3.5 Imperative
43.3.3.6 Converb
43.3.3.7 Jussive
43.3.4 Adjectives
43.3.5 Adverbs
43.3.6 Derived nominals
43.3.7 Conjunctions
43.4 Syntax
43.4.1 Word order
43.4.2 Verb phrase
43.4.3 Noun phrase
43.4.4 Subordination
43.5 Finals remarks
Acknowledgments
References
Chapter 44: Iraqw
44.1 The language and its speakers
44.2 Phonology
44.2.1 Vowels
44.2.2 Consonants
44.2.3 Tone
44.2.4 Shape of syllables, roots, and words
44.3 Nominal morphology
44.3.1 Gender of nouns
44.3.2 Number of nouns
44.3.3 Number and gender agreement in adjectives
44.3.4 Numerals
44.3.5 Pronouns
44.3.5.1 Indefinite pronouns and suffixes
44.3.5.2 Interrogative pronouns and other question words
44.3.6 Noun phrase structure, antigenitive (construct) case
44.4 Verbal morphology
44.4.1 Verbal inflection
44.4.2 The verbs ‘to be’ or so-called “selectors”
44.4.3 Verbal derivation
44.4.4 Adverbials
44.4.4.1 Verbal adverbs
44.4.4.2 Case clitics
44.4.4.3 Sentential adverbs and adverbials
44.4.4.4 Interjections
44.4.4.5 Ideophones
44.5 Syntax
44.5.1 Constituent order and basic syntactic structures of simple sentences
44.5.2 Negative clauses and questions
44.5.3 Complex sentences
Acknowledgments
References
Chapter 45: Wandala
45.1 Introduction
45.2 Phonology
45.3 Lexical Categories
45.4 Morphology
45.5 the Structure of the noun phrase
45.8 Semantic relations within the clause
45.9 Number
45.10 Locative predication
45.11 Locative ectensions
45.12 Adjuncts
45.13 Verbless predication
45.14 Aspect
45.15 Tense
45.6 Mood
45.17 Negation
45.18 Questions
45.19 Comment clause
45.20 Topicalization
45.21 Focus
45.22 Reference system
45.23 Paratactic constructions
45.24 Complementation
45.25 Conditional and Temporal clauses
45.26 Adjunct clauses
45.27 The relative clause
References
Chapter 46: Kumam
46.1 Introduction
46.2 Phonology
46.2.1 Consonants
46.2.2 Vowels
46.2.3 Syllable canon
46.3 Tone
46.3.1 Inventory of tones
46.3.2 Tone sandhi
46.3.3 Downstep and double downstep
46.3.4 Syntactic functions of tone
46.4 Major categories
46.4.1 Verbs
46.4.2 Nouns
46.4.3 Adjectives
46.4.4 Prepositions
46.4.5 Pronouns
46.4.6 Demonstratives
46.4.7 Interrogative pronouns
46.4.8 Numerals
46.5 Syntax
46.5.1 Basic order in sentences
46.5.2 Intransitive, transitive, and middle sentences
46.5.3 Predicate nominal and predicate adjectival constructions
46.5.4 Existence, location, and possession
46.5.5 Topicalization
46.5.6 Complex sentences
46.6 Negation
46.7 Questions
References
Chapter 47: Baale
47.1 Introduction
47.2 Phonology and morphological alternations
47.3 Morphology and morphosyntax
Acknowledgments
References
Chapter 48: Sonoghay languages
48.1 Introduction
48.2 Phonology
48.3 Derivation
48.4 Nominals
48.5 Arguments
48.6 Predicates
48.7 Information structure
48.8 Lexicon
References
Chapter 49: Cara
49.1 Introduction
49.2 Phonology
49.2.1 Vowels
49.2.2 Consonants
49.2.3 Tones
49.3 Morphology
49.3.1 Nominals
49.3.1.1 Gender and number
49.3.1.2 Subject–object relation
49.3.1.3 Associative particle dì
49.3.2 Adjectives
49.3.3 Numerals
49.3.4 Pronominal forms
49.3.4.1 Personal pronouns
49.3.4.2 Demonstratives
49.3.4.3 Possessives
49.3.4.4 Interrogatives
49.3.5 Verbals
49.3.5.1 Structure of finite verbs
49.3.5.2 Verbal extensions
49.3.5.3 Juncture
49.3.5.4 Passive
49.3.5.5 Tense and aspect
49.3.5.6 Mood/modality
49.3.5.7 Negation
49.4 Syntax
49.4.1 Word order
49.4.2 Simple sentences
49.4.2.1 Declaratives
49.4.2.2 Questions
49.4.3 Complex sentences
49.4.3.1 Coordination and subordination
49.4.3.2 Relative clauses
References
Chapter 50: ǁX’egwi
50.1 Introduction
50.2 Phonology
50.2.1 Vowels
50.2.2 Consonants
50.2.3 Tone
50.3 Morphology
50.3.1 Noun
50.3.2 Adjectives
50.3.3 Numerals
50.3.4 Pronouns
50.3.4.1 Personal pronouns
50.3.4.2 Demonstrative pronouns
50.3.4.3 Interrogative pronouns
50.3.5 Verb
50.3.6 Adverbials
50.4 Syntax
References
Chapter 51: Sandawe
50.1 Introduction
51.2 Phonology
51.2.1 VowelsVowels
51.2.2 Consonants
51.2.3 Tone
51.3 Morphology
51.3.1 Noun
51.3.1.1 Inflection
51.3.1.2 Derivation
51.3.2 Adjectives and numerals
51.3.3 Pronouns
51.3.4 Verb
51.3.4.1 Inflection
51.3.4.2 Derivation
51.3.4.3 Tense, aspect, mood
51.3.5 Adverbials
51.4 Syntax
51.4.1 Constituent order
51.4.2 Major sentence types
51.4.2.1 Complex sentences
51.4.2.2 Questions
52.4.2.3 Copular constructions
References
Part VI: LANGUAGE, COGNITION, AND CULTURE
Chapter 52: Ideophones
52.1 Introduction
52.2 Methodological approaches: from purely formal to more discourse-prafmatic definitions
52.3 Typological definition
References
Chapter 53: Color term systems: Genetic vs. areal distribution in sub-Saharan Africa
53.1 Introduction
53.2 Color term cocepts and systems
53.3 Data and methodology of the seventh african languages study
53.4 Genetic and areal distribution of systems
References
Chapter 54: Experiencer constructions
54.1 Introduction
54.2 Experiencer constructions, definition, elements
54.2.1 Wolof (Niger-Congo, Atlantic)
54.2.2 South Cushitic: Iraqw, Alagwa, Burunge
54.2.3 Amharic (Afro-Asiatic, South Semitic)
54.2.4 Likpe (Niger-Congo, Kwa)
54.2.5 Wolaitta (Afro-Asiatic, Omotic)
54.2.6 Hausa (Afro-Asiatic, Chadic)
54.3 Summary
Reeferences
Chaptrer 55: Language and ethnobotany
55.1 Introduction
55.2 Folk taxonomy in selected bantu languages
55.2.1 Unique beginner
55.2.2 Life forms
55.2.3 Common names
55.2.3.1 Generics/generic taxa
55.2.3.2 Specifics/specific (subgeneric) taxa
55.3 Folk taxonomy and bantu noun classes
55.3.1 The unique beginner and life forms
55.3.2 Generics and specifics/subgeneric taxa
55.4 Conclusion
Acknowledgments
References
Chapter 56: Distinctive languages
56.1 Introduction
56.2 Functions of distinctive languages
56.3 Categorization
56.3.1 General notes
56.3.2 Subcategories of distinctive languages
56.3.2.1 Guild languages
56.3.2.2 Ritual languages
56.3.2.3 Reverence languages
56.3.2.4 Anti-languages
56.3.2.5 Urban youth languages
56.3.3 Overview
56.4 Linguistic strategies
56.4.1 General notes
56.4.2 Phonological strategies
56.4.2.1 Affixation
56.4.2.2 Permutation
56.4.2.3 Substitution of phonemes
56.4.2.4 Truncation
56.4.2.5 Word-forming pattern
56.4.3 Lexical-semantic strategies
56.4.3.1 Paraphrase
56.4.3.2 Derivation and composition
56.4.3.3 Synonymy
56.4.3.4 Semantic shift
56.4.4 Overview
56.5 Conclusion
References
Chapter 57: Conversation Analysis
57.1 Introduction
57.2 Methodological preliminaries
57.3 Organization of turm-taking
57.4 Sequence organization
57.5 Future perspectives
References
Chapter 58: Cognition and language
58.1 Introduction
58.2 Review: the development ofcognitive linguistics in relation to african languages
58.2.1 On the relation between language and cognition
58.2.1.1 Nativism vs. cognitivist assumptions of embodimentand linguistic categories
58.2.1.2 Lexical categorization
58.2.1.3 Conceptual metaphor
58.2.2 Insight contributed through the study of Africanlanguages: grammaticalization and spatial expressions
58.2.2.1 Grammaticalization
58.2.2.2 Spatial orientation
58.3 Current developments: culture as a relevant variable
58.3.1 Philology and historical semantics: the benefitfrom cognitive linguistic insight
58.3.2 Temporality
58.4 Areas of convergent conceptual patterns and the prospect of cognitive research on african languages
References
Part VII: LANGUAGE AND SOCIETY
Chapter 59: Indigenous african scripts
59.1 Introduction
59.2 Tifinagh
59.3 West african syllabaries
59.3.1 The Vai syllabary *
59.3.2 The Mende syllabary
59.3.3 The Loma syllabary
59.3.4 The Kpelle syllabary
59.3.5 The Bambara syllabary
59.4 West african alphabets
59.4.1 Bassa Vah alphabet
59.4.2 Nko script
59.4.3 Wolof Garay script
59.4.4 Hausa scripts
59.4.5 Adlam alphabet
59.5 Script of cameroon
59.5.1 Bamum script *
59.5.2 Bagam script
59.6 Script of somalia and ethiopia
59.6.1 Somali alphabets
59.6.2 Oromo script
59.7 Script of northeastern Africa
59.7.1 Beria script *
59.7.2 Nubian scripts
59.8 Script of religious communities
59.8.1 Medefaidrin (Obεri Ɔkaimε) script
59.8.2 Mandombe script
59.8.3 Yoruba “holy” writing
59.9 Individual writing systems
59.9.1 Bété script *
59.9.2 Nwagu Aneke Igbo syllabary
59.9.3 Aka Umuagbara script
59.9.4 Mwangwego script *
59.9.5Other alphabetic scripts
59.10 Scripts of the diaspora
References
Chapter 60: Language policy and politics
60.1 Introduction
60.2 Public recognition
60.2.1 Official multilingualism
60.2.2 Language rationalization
60.3 Conclusion
References
Chapter 61: Language and education
61.1 Introduction
61.2 Literature
61.3 African or european LOIs?
61.3.1 Terminology
61.3.2 African experiences
61.4 Which african LOIs?
61.5 Conclusion
References
Chapter 62: Language endangerment, documentation, and revitalization
62.1 Introduction
62.2 Types of Endangerment in africa
62.2.1 How languages become endangered
62.3 Documentation
62.4 Reviatlization
62.5 Conclusion
References
Chapter 63: Language birth: Youth/town language
63.1 Introduction
63.2 Examples of youth/town languages from african cities
63.2.1 Nouchi (Ivory Coast)
63.2.2 Town Bemba (Zambia)
63.2.3 Tsotsitaal (South Africa)
63.2.4 Camfranglais (Cameroon)
63.2.5 Sheng (Kenya)
63.3 Discussion
63.4 Conclusion
References
Chapter 64: Language contact
64.1 Introduction
64.2 Early work on language contact in Africa
64.3 Africa as a contact zone: recent findings
64.4 Theoretical framework of language contact
64.5 Language contact in souroudougou
64.6 Conclusion
References
Chapter 65: Mixed languages: The case of Ma’á/Mbugu
65.1 Background
65.2 Normal mbugu and inner mbugu or ma'A
65.3 Patterns of inner mbugu usage
65.4 Historical scenario
65.5 Language shift or not
65.6 Conclusion
References
Chapter 66: African languages in the diaspora
66.1 Introduction
66.2 The African diaspora
66.3 Linguistic practices and african languages in the diaspora
66.4 Conclusion
Acknowledgments
References
Chapter 67: Pidgin and creole languages
67.1 Introduction
67.2 Definitions and basic concepts
67.3 Distribution, classification, and major characteristics
67.3.1 Exoglossic language-based languages and/or varieties
67.3.1.1 Portuguese-based PC
67.3.1.2 French-based PC
67.3.1.3 English-based PC
67.3.1.4 Dutch- or Afrikaans-based PC
67.3.2 Arabic-based PC
67.3.3 Endoglossic language-based languages and/or varieties
67.3.3.1 Bantu-based PC
67.3.3.2 Other African language-based PC
References
For further reading
Chapter 68: Sign languages
68.1 Introduction
68.2 Emergence and creation
68.3 Sign languages outside the context of deaf education
68.4 Language change and contact
68.5 Multimodal communication
68.6 Language policies and applied benefits
68.7 Conclusion
References
Chapter 69: Arabic in africa
69.1 Native arabic: the dialects
69.1.1 Population distribution and basic history
69.1.2 The dialects
69.1.2.1 The Maghreb
69.1.2.2 Sudanic Arabic
69.1.2.3 Egyptian Arabic
69.1.2.4 Eastern Libya
69.1.3 Contact
69.1.3.1 Coptic, Berber
69.1.3.2 Other African languages
69.1.3.3 Code-switching
69.1.3.4 Pidgin and Creole Arabic
69.1.4 Sociolinguistics
69.2 Alassical and standard arabic
69.2.1 The domain of Classical Arabic
69.2.2 Writing
69.3 Conclusion
References
Chapter 70: Orthography standardization
70.1 Introduction
70.2 Language vitality in africa
70.3 Design challenges specific to Africa
70.4 Motivations for orthography standardization
70.4.1 Endangered languages
70.4.2 Linguistic rights and cultural heritage
70.4.3 Universal quality primary education
70.4.4 Bible translation
70.5 Models for orthography standardization
70.5.1 Laissez-faire model
70.5.2 Top-down model
70.5.3 Community-based model
70.6 Orthography standardization practices
70.6.1 Historical background
70.6.2 Intra-language standardization
70.6.3 Language family standardization
70.3.4 National harmonization
70.6.5 International harmonization
70.7 Case studies
70.7.1 Ik (Uganda)
70.7.2 Aja (Benin and Togo)
70.8 Conlcusion: Best practices
References
Chapter 71: Pragmatics and communication
71.1 The scope of pragmatics
71.2 A methodology for the study of african cultures and societies
71.3 Development pragmatics
71.4 Pragmatic operators
71.4.1 Pragmatic aspects of nominal classification
71.4.2 Focus and implicature
71.5 Historical and political dimensions of linguistic choice
71.5.1 The wider scope of pragmatics: dual staging
71.5.2 The interplay of language use and dual staging
71.5.3 Dual staging in literary expression and publicadvertisement
71.6 Face and politeness
71.6.1 Strategies for mitigating social power
71.6.2 Conventional vs. conversational implicatures of dualdirective repertories
71.6.3 Linguistic politeness as a factor of social cohesion
71.7 Address and greeting routines
71.7.1 Enhancing the social value of positive face
71.7.2 Strategic omission of politeness markers
71.7.3 Pragmatics as a module of language description
References
Chapter 72: African languages in information and communications techonology
72.1 Introduction
72.2 Bridging or increading the digital divide?
72.3 Technological dimensions and interface language
72.4 Internet content
72.5 Texting
72.6 Conclusion
References
Part VIII: LANGUAGE AND HISTORY
Chapter 73: Words, things, and meaning: Linguistics as a tool for historical reconstruction
73.1 Introduction
73.2 Words, things, concepts: tighter and looser bundles
73.3 New directions
73.4 Conclusion
Acknowledgments
References
Chapter 74: Language and archaeology
74.1 Introduction
74.2 Stratigraphy
74.3 Seriation
74.4 Geograpghic distribution
74.5 Absolute dating and time-depth
74.6 Towards archaeo-linguistics?
Acknowledgments
References
Part IX: LANGUAGE AND ORATURE
Chapter 75: Narratives
75.1 Western perception of african narrative art
75.2 General features of oral literature
75.3 Specoific types of oral literature: genres
75.3.1 Documentary literature
75.3.2 Fictional narrations
75.3.2.1 Fables
75.3.2.2 Animal tales
75.3.2.3 Fairy tales
75.3.3 Borderline narrations between truth and fiction
75.3.3.1 Anecdotes
75.3.3.2 Legends
75.3.3.3 Epic narrations
75.4 Modern scholarly view on oral literature in africa
References
Chapter 76: Proverbs
76.1 Introduction
76.2 Historical heritage
76.3 Culture heritage
76.4 Oral tradition
76.5 Language
76.6 Literature
76.7 Orature
76.8 Proverbs
76.8.1 Origin, creation, collection, and spread of proverbs
76.8.2 Types of proverbs
76.8.2.1 Form of proverbs
76.8.2.1.1 Linguistic features
76.8.2.1.1.1 Phonetics and phonology
76.8.2.1.1.2 Onomatopoeia and ideophones
76.8.2.1.1.3 Homophones, homographs, and homonyms
76.8.2.1.1.4 Elision and contraction
76.8.2.1.1.5 Assonance
76.8.2.1.1.6 Morphology and systax/sentence structure
76.8.2.1.2 Literary style of proverbs
76.8.2.1.2.1 Paradixical proverbs
76.8.2.1.2.2 Metaphors
76.8.2.1.2.3 Rhythm and meter
76.8.2.1.2.4 Puns/parnomasia
76.8.2.1.2.5 Aphorisms
76.8.2.1.2.6 Parabolic proverbs
76.8.2.1.2.7 Antithetical/contradictory proverbs
76.8.2.1.2.8 Repetition
76.8.2.1.2.9 Rhyme
76.8.2.1.2.10 Alliteration
76.8.2.1.2.11 Republication
76.8.2.1.2.11 Parody
76.8.2.1.2.13 Simile in proverbs: "like" or "as" proverbs
76.8.2.1.2.14 Climax
76.8.2.1.3 Other (general) features
76.8.2.1.3.1 Numerical proverbs
76.8.2.1.3.2 Integral proverbs
76.8.2.1.3.3 Synthetic proverbs
76.8.2.1.3.4 Onomastic proverbs or personifcation
76.8.2.1.3.5 Perverbs
76.8.2.2 Content
76.8.2.2.1 Oracy vs. literacy
76.8.2.2.2 Historical proverbs
76.8.2.2.3 Culture heritage
76.8.2.2.4 Chatterboxes
76.8.2.2.5 Invitability of the communality of work
76.8.2.2.6 Marriage and familky
76.8.2.2.7 Economics/banking
76.8.2.2.8 Justice/jurisdiction/lawsuits
76.8.2.2.9 Education
76.8.2.2.10 Politics
76.8.2.2.11 Old age and respect for elderly people
76.8.2.2.12 Spiritual and religious proverbs
76.8.2.2.13 Popular proverbs
76.8.2.2.14 Immediate action/opportunity
76.8.2.2.15 Reciprocity
76.8.2.2.16 Prioritization
76.8.2.2.17 Roaming/ineptitide
76.8.2.2.18 Caution
76.8.2.2.19 Truth/honesty
76.9 Impact of proverbs
76.10 Conclusion
References
Further Reading
Chapter 77: Poetry
77.1 Introduction
77.2 Oral poetry
77.2.1 Praise poetry
77.2.2 Epic-heroic poetry
77.2.3 War and hunting poetry
77.2.4 Elegiac poetry
77.2.5 Religious poetry
77.2.6 Lyric poetry and songs
77.3 Written poetry and manuscript cultures
77.4 Poetry in latin script and new media
77.5 Poetry in the postcolony
References
Author Index
Language Index
Subject Index
Citation preview
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
AFRICAN LANGUAGES
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
O X F O R D H A N D B O O K S IN LI N G U I S T I C S Recently published
THE OXFORD HANDBOOK OF ERGATIVITY Edited by Jessica Coon, Diane Massam, and Lisa deMena Travis
THE OXFORD HANDBOOK OF WORLD ENGLISHES Edited by Markku Filppula, Juhani Klemola, and Devyani Sharma
THE OXFORD HANDBOOK OF POLYSYNTHESIS Edited by Michael Fortescue, Marianne Mithun, and Nicholas Evans
THE OXFORD HANDBOOK OF EVIDENTIALITY Edited by Alexandra Y. Aikhenvald
THE OXFORD HANDBOOK OF LANGUAGE POLICY AND PLANNING Edited by James W. Tollefson and Miguel Pérez-Milans
THE OXFORD HANDBOOK OF PERSIAN LINGUISTICS Edited by Anousha Sedighi and Pouneh Shabani-Jadidi
THE OXFORD HANDBOOK OF ENDANGERED LANGUAGES Edited by Kenneth L. Rehg and Lyle Campbell
THE OXFORD HANDBOOK OF ELLIPSIS Edited by Jeroen van Craenenbroeck and Tanja Temmerman
THE OXFORD HANDBOOK OF LYING Edited by Jörg Meibauer
THE OXFORD HANDBOOK OF TABOO WORDS AND LANGUAGE Edited by Keith Allan
THE OXFORD HANDBOOK OF MORPHOLOGICAL THEORY Edited by Jenny Audring and Francesca Masini
THE OXFORD HANDBOOK OF REFERENCE Edited by Jeanette Gundel and Barbara Abbott
THE OXFORD HANDBOOK OF EXPERIMENTAL SEMANTICS AND PRAGMATICS Edited by Chris Cummins and Napoleon Katsos
THE OXFORD HANDBOOK OF EVENT STRUCTURE Edited by Robert Truswell
THE OXFORD HANDBOOK OF LANGUAGE ATTRITION Edited by Monika S. Schmid and Barbara Köpke
THE OXFORD HANDBOOK OF LANGUAGE CONTACT Edited by Anthony P. Grant
THE OXFORD HANDBOOK OF ENGLISH GRAMMAR Edited by Bas Aarts, Jill Bowie, and Gergana Popova
THE OXFORD HANDBOOK OF AFRICAN LANGUAGES Edited by Rainer Vossen and Gerrit J. Dimmendaal For a complete list of Oxford Handbooks in Linguistics, please see pp. –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.........................................................................................................................................
AFRICAN LANGUAGES ......................................................................................................................................... Edited by
RAINER VOSSEN and
GERRIT J. DIMMENDAAL
1
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
3
Great Clarendon Street, Oxford, , United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © editorial matter and organization Rainer Vossen and Gerrit J. Dimmendaal © the chapters their several authors The moral rights of the authors have been asserted First Edition published in Impression: All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press Madison Avenue, New York, NY , United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: ISBN –––– Printed and bound by CPI Group (UK) Ltd, Croydon, Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
C
...................................
Acknowledgments List of Maps, Figures, and Tables List of Abbreviations List of Contributors
xii xiii xxi xxxii
PART I INTRODUCTION . Introduction G J. D R V
PART II DOMAINS OF GRAMMAR . Phonology and phonetics M J. K
. Tone D O
. Morphology K B
. Syntax J Z
PART III LANGUAGE COMPARISON . African language types R V
. Dialectology and linguistic geography M B. L
. Reflections on the history of African language classification L G
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
vi
PART IV LANGUAGE PHYLA AND FAMILIES . Niger-Congo, with a special focus on Benue-Congo J G
. Atlantic F L̈
. Mande H S
. Kwa M E K D
. Gur G M
. Bantu and Bantoid L M
. Adamawa U K̈
. Ubangi H P
. Kordofanian N Q
. Afro-Asiatic overview V P
. Egyptian Bˊ J. I-N
. Berber M K
. East Cushitic M T
. North Cushitic M V
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
vii
. Central Cushitic Z L
. South Cushitic R Kß
. Omotic B K̈
. Chadic B C
. Ethio-Semitic V P
. Nilo-Saharan and its limits G J. D
. Saharan N C
. Eastern Sudanic G J. D A J
. Central Sudanic P B
. Khoisan H H
. Linguistic isolates G J. D
PART V LANGUAGE STRUCTURES: CASE STUDIES . Bom-Kim G. T C
. Dan V V
. Biali C Sˊ
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
viii
. Yukuben R-J A
. Bende Y A
. Waja U K̈
. Zande H P
. Zenaga C T-C
. Sidaama K K
. Kolisi Z L
. Iraqw M M
. Wandala Z F
. Kumam O H
. Baale G J. D
. Songhay languages L S
. Cara R V
. ǁX’egwi H H
. Sandawe H E
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
ix
PART VI LANGUAGE, COGNITION, AND CULTURE . Ideophones C K-H . Color term systems: Genetic vs. areal distribution in sub-Saharan Africa D L. P
. Experiencer constructions U Z
. Language and ethnobotany K Lˋ
. Distinctive languages P F
. Conversation analysis M R̈
. Cognition and language A F
PART VII LANGUAGE AND SOCIETY . Indigenous African scripts A R S B
. Language policy and politics K S
. Language and education I S
. Language endangerment, documentation, and revitalization J E
. Language birth: Youth/town language E H
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
x
. Language contact K B
. Mixed languages: The case of Ma’á/Mbugu M M
. African languages in the Diaspora A H
. Pidgin and creole languages G S
. Sign languages V N
. Arabic in Africa J O
. Orthography standardization E K D R
. Pragmatics and communication T B
. African languages in information and communications technology K V L
PART VIII LANGUAGE AND HISTORY . Words, things, and meaning: Linguistics as a tool for historical reconstruction D L. S . Language and archaeology K B
PART IX LANGUAGE AND ORATURE . Narratives W J. G. M̈
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xi
. Proverbs S K. B
. Poetry C V
Author Index Language Index Subject Index
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
A .....................................................................
Special thanks are due to the cartographer of the Institute for African Studies and Egyptology, University of Cologne; Monika Feinen for the range of maps she produced and for her professional attitude; as well as to Jan Knipping for his help with the indexes. The editors would also like to express their deeply felt gratitude to Lakshmanan Sethuraman and Joy Mellor for their dedication and patience during the proofreading stage. Gerrit Dimmendaal would like to thank the Rector of the University of Cologne, the wonderful Prof. Axel Freimuth, for bestowing the Leo Spitzer Prize for excellent research upon him, which made it possible to engage the following two people in order to complete this project. John Vanderelst kindly offered his help in formatting several chapters, at the same time pointing toward several editorial issues for which we are very grateful. His support as well as the help of Mary Chambers in shaping up the texts of authors whose first language is not English is also gratefully acknowledged here. This support would not have been possible without this financial support from the University of Cologne.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
L M, F, T
.............................................................................................................................
Maps
.
Geographical position of languages described briefly in Part V
.
African language families and isolates following recent proposals
.
Niger-Congo subgroups following recent proposals
.
The geographical distribution of Atlantic and Mande languages
.
The geographical distribution of Mande languages
.
The geographical distribution of Kwa languages
.
The geographical distribution of Gur languages
.
The geographical distribution of Bantu languages
.
Bantu language groupings (based on Guthrie )
.
The geographical distribution of Adamawa languages
.
The geographical distribution of Ubangi languages
.
The geographical distribution of Kordofanian languages
.
The geographical distribution of Afro-Asiatic languages
.
The geographical distribution of Berber languages
.
The geographical distribution of East Cushitic languages
.
Distribution of Central Cushitic languages
.
The geographical distribution of South Cushitic languages
.
The geographical distribution of Omotic languages
.
The geographical distribution of Chadic languages
.
Hausa extension
.
Chadic (minus Hausa) and phonological zones in Africa (based on Clements and Rialland )
.
The geographical distribution of Ethio-Semitic languages
.
The geographical distribution of Nilo-Saharan languages
.
The geographical distribution of Saharan languages
.
The position of the verb in Eastern Sudanic
.
The geographical distribution of Central Sudanic languages
.
The geographical distribution of Khoisan languages
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xiv
, ,
.
African linguistic isolates
.
The distribution of pidgins and creoles in Africa
Figures
.
Map of the georeferenced survey points of the Rif (Lafkioui )
.
Dendrogram vs. CA map—Lv—all lexis
.
CALv maps—seven clusters vs. three clusters vs. two clusters—all lexis
.
CALv vs. MDSLv vs. CACLv maps
.
CAC—Bin vs. GIW vs. Lv maps—all lexis
.
The six phonological zones of Africa (based on Clements and Rialland )
.
New grouping superseding Atlantic (Pozdniakov and Segerer )
.
“SplitsTree” diagram with reticulations for Atlantic languages (Pozdniakov and Segerer )
.
“SplitsTree” diagram without reticulations (Pozdniakov and Segerer )
.
New classification of Atlantic (Pozdniakov and Segerer forthcoming)
.
Bantu within Niger-Congo (after Williamson and Blench )
.
Main subgroups of Grollemund et al.’s () classification
.
A lexicostatistical classification of Bantoid (after Piron : )
.
Main groupings of the phylogenetic classification of Bantoid of Grollemund et al. (forthcoming)
.
Blažek’s () classification of Berber founded on Starostin’s calibrated glottochronology based on minimal values (language names adapted)
.
The Cushitic languages (from Tosco a)
.
The classification of Cushitic according to Voigt ()
.
A revised classificatory tree of East Cushitic
.
Synopsis of the historical development of PSC (PSC ¼ PWR) Kießling and Mous (: )
.
The subclassification of Nilo-Saharan
.
The subclassification of Eastern Sudanic
.
Classification of Saharan languages
.
The subclassification of Eastern Sudanic
.
South Atlantic classification (Voeltz ; Childs )
.
Vowel trapezium in Yukuben
.
Hypothesized basic color term systems (based on Kay and Maffi , )
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
xv
.
General distinction of special languages
.
Categorization of distinctive languages
.
Word category continuum for nouns across the Bantu languages
.
Comparison of Nouchi and standard French grammar and vocabulary (Vakunta )
.
Example of Town Bemba compared to standard Bemba (from Kabinga : )
.
Examples of Tsotsitaal with a Xhosa and Zulu base
.
Camfranglais via CamP syntax (Schröder )
.
Camfranglais via French syntax (Schröder )
.
Example of Sheng compared to Standard Swahili (Githiora : )
.
መስቀል፡ስክዌር፡፡ (Mäsk’äl Square)
Tables
.
Examples of click-containing words in Khoisan
.
Inventory of noun classes in Yeyi (Bantu; see Seidel : –)
.
General procedure of computational dialectometry analysis
.
Data excerpt in encoded format of ALR
.
Excerpt from the binary distance matrix of the lexeme ‘heel’
.
Excerpt from the GIW distance matrix of the lexeme ‘heel’
.
Excerpt from the Lv distance matrix of the lexeme ‘heel’
.
Cost of operations allowing modification of awrez into inerz (heel)
.
Example of calculation of Lv distance for modifying awrez into inerz (heel)
.
Possible cognate noun class markers in Niger-Congo following Schadeberg (: )
.
Possible cognate verbal extensions in Niger-Congo following Hyman (: )
.
Proto-Bantu noun classes
.
Possible Konyagi correspondences for Sereer b- (Pozdniakov : )
.
A classification of Mande languages, major branches, and principal representatives
.
Comparison of reconstructions of initial consonants: Central-Southwest (Kastenholz ); Southern (Vydrin ); Niger-Volta (Schreiber )
Main groups of Bantu languages (based on Guthrie –; Maho )
.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xvi
, ,
.
Adamawa groups and isolated languages
.
Some similarities between noun class markers of various semantic classes across Kordofanian families
.
Tima (Katloid) and Koalib (Heibanian, noun class language) language names nominal derivation
.
Some verb extensions attested in Koalib (Heibanian)
.
Some plausible Proto-Kordofanian roots
.
Core vocabulary items between CC languages and the rest
.
Selected cognates in CC languages and Ethio-Semitic (Amharic)
.
The Chadic language family: classification (Newman a: –)
.
Chadic consonants (Newman : )
.
Proto-Chadic and Afro-Asiatic pronouns
.
Internal -a- imperfect/habitative in Chadic languages (Jungraithmayr )
.
Distribution of linguistic features
.
Nilo-Saharan and its limits
.
The accusative case marker in Nilo-Saharan
.
The question word ‘who’ in Nilo-Saharan
.
Rates of cognates in Saharan languages (verbs and non-verbs)
.
Underlying pronominal morphemes in Saharan languages
.
Verb class I in Saharan languages
.
Verb class II in Saharan languages
.
Verb class III formative n ‘to say’
.
Verb class III in Saharan languages
.
Eastern Sudanic subgroups
.
The consonant inventory of Kenzi
.
The consonants of Ngaalam
.
Pronominal possessives in Me’en
.
Subject marking on verbs in Anywa
.
Previous groupings of Central Sudanic languages
.a
Lexicostatistical distances between twenty-nine CS languages (based on a hundred-word Swadesh list; Branch Average)
.b
Identification of subgroups and languages
.a
Distribution patterns of likely cognates in the CS subgroups (excluding indeterminate languages) and number of occurrences
.b
Absolute occurrences and percentages (out of ) of shared likely cognates in pairs of subgroups
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
.
Central Sudanic correspondence formulas for labiovelar consonants
.
Central Sudanic correspondence formulas for glottalized consonants
.
Evidence of the verbal prefixes through the subgroups
.
A selection of WALS typological features for CS languages (Dryer and WALS )
.
Languages possibly constituting linguistic isolates
.
Bom-Kim consonantal inventory
.
Bom-Kim vowel inventory
.
Personal pronouns in Bom-Kim
.
The Bom-Kim verbal complex
.
Dan vowels
.
Dan consonants
.
Derivation of plural and intensive adjectival forms in Dan
.
Dan numerals
.
Pronominal predicative markers (PPM) in Dan
.
Personal pronouns in Dan
.
Affirmative TAM constructions in Dan
.
Negative TAM constructions in Dan
.
The vowel system of Biali
.
Vowel reduction in Biali
.
The consonant system of Biali
.
Allophonic variation in Biali
.
Qualifying adjectives in Biali
.
Demonstrative adjectives in Biali
.
Interrogative adjectives in Biali
.
Substitutive pronouns in Biali
.
Verbal derivation in Biali
.
Imperative/hortative in Biali
.
Adverbs in Biali
.
Non-complex (or ‘simple’) consonants in Yukuben
.
Yukuben noun classes
.
‘Relating concords’ in Yukuben
.
The emphatic personal pronouns in Yukuben
.
Pronouns in object position in Yukuben
.
The possessive pronouns in Yukuben
xvii
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xviii
, ,
.
Speaker–hearer’s orientation and demonstratives/deictic markers in Yukuben
.
Yukuben TAM marking strategies
.
Phonemes and phones in Bende
.
Suprasegmental diacritics in Bende
.
Tonal types according to the number of syllables (σ) in Bende
.
Glide formation and vowel coalescence (VV) in Bende
.
Regular noun class pairings in Bende
.
Numerals – (only – take a numeral prefix) in Bende
.
Verbal slots and affixes in Bende
.
Concordance with persons and classes in Bende
.
Verb conjugation (simple forms) in Bende
.
Consonants in Waja
.
Vowels in Waja
.
Concord markers in Waja
.
Class pronouns and agreeing nominal suffixes in Waja
.
Noun class system in Waja
.
Genders and single classes in Waja
.
Personal pronouns in Waja
.
Verb forms of simple verbs in Waja
.
Motion verb ‘mount, climb’ in Waja
.
Posture verb ‘sit (down); stay’ in Waja
.
Productive verb extensions in Waja
.
Subject pronouns in Waja
.
Zande personal pronouns
.
Zenaga consonant inventory
.
Nominal determiners in Zenaga
.
Independent personal pronouns in Zenaga
.
Suffixal personal pronouns in Zenaga ()
.
Suffixal personal pronouns in Zenaga ()
.
Possessive pronouns in Zenaga
.
Demonstrative pronouns in Zenaga
.
Verbal inflection in Zenaga
.
Tense, aspect, mood in Zenaga
.
Sidaama consonant phonemes
.
Aspectual differences between dynamic verbs in Sidaama
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
.
Nominative, genitive, and dative-locative case suffixes in Sidaama
.
Kolisi vowels
.
Kolisi consonants
.
Personal pronouns in Kolisi
.
Object pronouns in Kolisi
.
Possessive pronouns in Kolisi
.
Reflexive pronouns in Kolisi
.
Iraqw consonant phonemes in orthographical symbols
.
Gender in demonstratives in Iraqw
.
Iraqw personal pronouns
.
Iraqw possessives pronouns and suffixes
.
Iraqw demonstrative pronouns and suffixes
.
Iraqw indefinite suffixes and pronouns
.
Iraqw question words
.
Construct case suffixes and pronouns in Iraqw
.
Indicative present in Iraqw
.
Imperative forms in Iraqw
.
Affixes to selectors in Iraqw
.
The first imperfective in Iraqw
.
Underlying consonants and glides in Wandala
.
Underlying vowels in Wandala
.
Phonetic vowels in Wandala
.
Consonant phonemes in Kumam
.
Vowel phonemes in Kumam
.
Baale consonants
.
Pronominal possessives in Baale
.
Perfective and imperfective verb forms in Baale
.
Consonant system of Proto-Songhay
.
Independent pronouns in Proto-Songhay
.
Mood/aspect distinction in the Songhay family
.
Cara vowel phonemes
.
Egressive consonants in Cara
.
Cara click inventory
.
Cara PGN enclitics
.
The personal (subject) pronouns of Cara
.
ǁX’egwi lexemes with “uvular” consonants
xix
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xx
, ,
.
The vocalic system of ǁX’egwi
.
ǁX’egwi pulmonic consonants
.
ǁX’egwi lingual consonants
.
Singular/plural classes in ǁX’egwi
.
ǁX’egwi pronominal system per Lanham and Hallowes ()
.
Demonstratives and locative adverbs in ǁX’egwi
.
Active verb tense forms in ǁX’egwi
.
Sandawe vowel phonemes
.
Egressive consonant phonemes in Sandawe
.
Ingressive consonant phonemes in Sandawe
.
PGN suffixes in Sandawe
.
Sandawe demonstratives
.
Sandawe object markers
.
Verb to verb derivation in Sandawe
.
Realis, irrealis, and imperative/subjunctive in Sandawe
.
Narrative, repetitive, and subjunctive conjunctions in Sandawe
.
Verbal direct speech vs. ideophonic direct speech
.
Gestures and morphemes
.
Distribution of total African languages compared to study sample
.
The seventy-language sample
.
Generalizations of apparent African BCT systems
.
Properties of prototypical agents, prototypical patients, and experiencers (cf. Kießling : )
.
Noun class allocation of life forms in selected Bantu languages
.
Noun class distribution of plant names in selected Bantu languages
.
Interrelations of functions and categories of distinctive languages
.
Interrelations of strategies and categories of distinctive languages compared to Kießling and Mous (, )
.
Glossary transcript symbols used in CA
.
Vai syllabary
.
Mende syllabary (an extract)
.
Nko alphabet
.
Bamum script
.
Somali scripts
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
L A ..................................................................................
AA AdjP AIDS ALR Alt (), () Ar. ASB ASL () ATR BB BBC BC BCT
Afro-Asiatic anti-agreement ablative absolutive accusative Anno Domini adjective adjective phrase adverb affirmative agentive agreement Acquired Immune Deficiency Syndrome Atlas linguistique des variétés berbères du Rif alternative morpheme anaphoric (pronoun) animate (gender) anterior aorist adverbial phrase; antipassive applicative Arabic Ait Segrouchen Berber American Sign Language aspect associative advanced tongue root attitude marker attributive marker auxiliary Bantu Botatwe British Broadcasting Corporation before Christ Benue-Congo before current era basic color term
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xxii
BD Bin Bk BLR B-M BR Bu C CA CAC CAM CamP CAR CASAS () CB CC CISOLA , . CNRS () () CP
Bɔ̄dʊ̄ŋázʊ dialect benefactive background binary distance black Bantu Lexical Reconstructions Biu-Mandara before present Base Reduplicant blue consonant cluster analysis; Conversation Analysis Composite Analysis Cluster Cameroon Cameroon Pidgin English Central African Republic Centre for Advanced Studies of African Society causative Common Bantu Central Cushitic current era circumstant Comité International du Suivi de l’Orthographe de la Langue Aja conjoint (noun) class class marker; conceptual metaphor; concord marker common case completive Centre National de la Recherche Scientifique comitative complementizer completive construct case concord; construct case concessive conditional conjunction connective consecutive continuous copula Convergent Pedagogy
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
CREA CS DC () DKB DNA DRC DW EC E-C .. EGIDS El () ES
Centre de Recherches et d’Études Anthropologiques construct state; Central Semitic; Central Sudanic causative consecutive dative double Djenné Chiini (not defined in source) declarative definite(ness) (marker/article) demonstrative dependent clause derivational suffix desiderative determiner detransitive diminutive directional distal disjoint Documenting the Kim and Bom languages of Sierra Leone deoxyribonucleic acid direct object determiner phrase Democratic Republic of Congo derivational suffix different subject converb dual derivative verb extension Deutsche Welle (Voice of Germany) East Cushitic East Chadic exponent of a function Expanded Graded Intergenerational Disruption Scale extra-low tone emotion emphasis; emphatic enclitic epenthetic; expletive pronoun external possessor construction ergative Eastern Sudanic
xxiii
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xxiv
E-S ESA , excl. () F FPP G GD GIW GTM GUSO H HA HEC HS H-S HVHS IADI IAV ICT
Ethio-Semitic Epigraphic South Arabian exclusive existential series of personal predicative markers existential experitum experiencer expletive extension falling tone feminine fundamental frequency (indicating the lowest frequency counting from zero) factative focus marker final particle first pair parts future tense green gender Ga-Dangme genitive (marker/linker) gerund “Gewichteter Identitätswert” goal general preposition Ghana-Togo-Mountain (languages) Gbe Uniform Standard Orthography high tone Haya habitual Highland East Cushitic hither Humburi Senni Hamito-Semitic hearsay evidential human (gender) high vowel harmony set hypothetical Ik Agenda Development Initiative “immediately after the verb” position Information and Communications Technology
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
IDP IFAN () INALCO () , incl. ()() () (). IP IPA () ISO ITB K KC KG KN KS KSL KY L L
xxv
indirect object internally displaced person Institut Fondamental d’Afrique Noire imperative (series of personal predicative markers) imperfect imperfective impersonal verbal extension ‘in’ Institut National des Langues et Civilisations Orientales inanimate (gender) inclusive indicative indefinite independent (pronoun) infinitive instrumental intensive interrogative interjection intransitive inversion indirect object marker intonation phrase International Phonetic Alphabet internal possessor construction imperfective irrealis International Organization for Standardization Imdlawn Tashlhiyt dialect of Moroccan Berber conjoint series of pronominal predicative markers; conjoint verbal form juncture jussive Köhler Koyra Chiini Kwangali Kwanyama Koyraboro Senni Kenyan Sign Language Kikuyu low tone first language
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xxvi
L LACITO L&H LD LEC lit. LND LOI LOT LSP Lv LZ M MAS MDS MEF MIT MMD MP MSA MTE N, n. NA NACALCO NC . . ND
second language Laboratoire de Langues et Civilisations à Tradition Orale Lanham and Hallowes Bálíssà (or Lissa) dialect Lowland East Cushitic literally linker Lendu locative (case) logophoric pronoun language of instruction Landelijke Onderzoekschool Taalwetenschap language for specific purposes Levenshtein Lozi mid-tone masculine manner-marking function of the preposition ni Mangbetu(-Asua) multidimensional scaling mediopassive Mangbutu-Efe middle (extension/voice) Massachusetts Institute of Technology Moru-Madi modified, modification Mpiemo modal particle Modern South Arabian mother tongue education noun; (syllabic) nasal neuter non-attested; action noun National Association of Cameroon Language Committees North Cushitic; Niger-Congo noun class; narrative conjunction nominal case inflection noun class marker narrative connective noun class pronoun, hu class noun class system Ndonga
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
NGO NI NKNS NL NN NP NS ()() OCP OSU OT P P-AA () () PB PBT p.c. PC P-C PCC PCCl PEB PEC
negation, negative neutral aspect nongovernmental organization non-high Nigeria Niger-Kordofanian–Nilo-Saharan national language nominalizer North Nyanza nominative; nominalizer noun phrase non-perfective number Nilo-Saharan non-subject series of personal pronouns number; numeral object oblique (case) Obligatory Contour Principle other-initiated repair object marker oblique object object pronoun optative series of pronominal predicative markers Ohio State University Optimality Theory predicate Proto-Afro-Asiatic particle participle passive past tense Proto-Bantu Proto-Bantu-Potou-Tano personal communication Proto-Cushitic; pidgin and creole; pronominal clitic Proto-Chadic Proto-Central Cushitic Preverbal Clitic Cluster Proto-East Bantu Proto-East Cushitic perfect
xxvii
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xxviii
() PHEC PIRQ PLEC PNC png PNN PNWR () PP PPM ()
perfective pertensive marker perfective predicate focus perfective perfective person–gender–number Proto-Highland East Cushitic Proto-Iraqwoid plural Proto-Lowland East Cushitic plural verb stem predicative marker Proto-Niger-Congo person–number–gender Proto-North Nyanza Proto-Northern West Rift postposition place of articulation polite possessive point-of-view marker phonological phrase; prepositional phrase preprefix pronominal predicative marker predicative marker prefix preposition present tense; presentative preterite perfective; perfect series of pronominal predicative markers pronominal predicative marker pronoun pronominal proclitic progressive prohibitive series of pronominal predicative markers prohibitive prospective series of pronominal predicative markers proximal presumptive series of pronominal predicative markers particle
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
P-S PSam PSC PSWR PT PWR quant/ quot R () () () RFI RJ RN SB SBB SC SELAF
Proto-Semitic Proto-Sam Proto-South Cushitic postposition Proto-Southern West Rift Potou-Tano pre-vowel Proto-West Rift question marker; reconstructed form proposed by Quint quotative quantifier quotation marker red; rising tone root repetitive conjunction; relating concord realis recent reciprocal reduplication referential (pronominal element) reflexive (pronoun) relative marker Radio France Internationale Stiftelsen Riksbankens Jubileumsfond realis root number reactive token subject Common Southern Bantu Sara-Bongo-Bagirmi subject subjunctive subjunctive South Cushitic single class; subjunctive conjunction subject cross-reference marker Sociétéd’ÉtudesLinguistiquesetAnthropologiquesdeFrance semelfactive sequential action subject focus suffix singular
xxix
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xxx
SH SHARP SIL SL SPLM-N SPP SOAS SS () . SVC SW TA TAM TBU TCU TETU , . TSK TUFS Tþ T T- UCLA UDSM UN
superhigh tone School Health and Reading Program subject indefinite Summer Institute of Linguistics sign language; superlow tone selective focalization marker subject marker singular verb stem specifier specificity Sudan People’s Liberation Movement-North second pair parts School of Oriental and African Studies South Semitic same subject converb stative stimulus subject subordinator, subordination; subjective subordinate clause subjunctive subsecutive marker serial verb construction Swahili tense and aspect tense–aspect–mood/modality tone-bearing unit turn-constructional unit “the emergence of the unmarked” terminal marker tense topic marker transitive transitivity Tondi Songway Kiini Tokyo University of Foreign Studies repair solution turn of repair initiation trouble-creating turn University of California at Los Angeles University of Dar es Salaam unrelated
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
UNAIDS UNESCO UNICEF US(A) USAID V v vcd vcls VD VLE VP W WALS WC W-C WCS WHO WNil WR WT X, x Y ya Z Ø ¼ . # ↑
þ –
xxxi
United Nations Programme on HIV/AIDS United Nations Educational, Scientific and Cultural Organization United Nations International Children’s Emergency Fund United States (of America) United States Agency for International Development vowel verb voiced voiceless Vidunda vowel harmony Vernacular Liberian English verbal noun voice volitional verb phrase Westphal; white World Atlas of Language Structures West Cushitic West Chadic World Color Survey World Health Organization Western Nilotic West Rift Words and Things (method of) constituent of variable category yellow years ago Ziervogel zero clitic boundary morpheme boundary syllable boundary downstepped tone upstepped/raised tone present absent first person second person third person
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
L C ...............................................................................
Yuko Abe, Tokyo University of Foreign Studies, Japan Rose-Juliet Anyanwu, Nasarawa State University, Keffi, Nigeria, and Goethe University, Frankfurt am Main, Germany Thomas Bearth, University of Zurich, Switzerland Sebastian K. Bemile, Ghana Institute of Languages, Accra, Ghana Klaus Beyer, Goethe University, Frankfurt am Main, Germany Koen Bostoen, Ghent University, Belgium Pascal Boyeldieu, LLACAN, Villejuif, France Solomija Buk, Ivan Franko National University of Lviv, Ukraine Bernard Caron, LLACAN, Villejuif, France G. Tucker Childs, Portland State University, OR, USA Norbert Cyffer, University of Vienna, Austria Gerrit J. Dimmendaal, University of Cologne, Germany Helen Eaton, SIL International, Uganda-Tanzania Branch James Essegbey, University of Florida, Gainesville, FL, USA Axel Fleisch, Goethe University, Frankfurt am Main, Germany Zygmunt Frajzyngier, University of Colorado, Boulder, CO, USA Patricia Friedrich, Munich, Germany Ludwig Gerhardt, University of Hamburg, Germany Jeff Good, State University of New York, Buffalo, NY, USA Osamu Hieda, Tokyo University of Foreign Studies, Japan Andrea Hollington, University of Cologne, Germany Henry Honken{, Sarona, WI, USA Ellen Hurst, University of Cape Town, South Africa Balázs J. Irsay-Nagy, Eötvös Loránd University, Budapest, Hungary
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xxxiii
Angelika Jakobi, University of Cologne, Germany Elke Karan, SIL International, Dallas, TX, USA Kazuhiro Kawachi, National Defense Academy of Japan Michael J. Kenstowicz, MIT, Cambridge, MA, USA Roland Kießling, University of Hamburg, Germany Christa Kilian-Hatz, Goethe University, Frankfurt am Main, Germany Ulrich Kleinewillinghöfer, Johannes Gutenberg University, Mainz, Germany Bernhard Köhler, Goethe University, Frankfurt am Main, Germany Maarten Kossmann, Leiden University, the Netherlands Mary Esther Kropp Dakubu{, University of Ghana, Legon, Accra, Ghana Mena B. Lafkioui, École des Hautes Études en Sciences Sociales, Paris, Université Sorbonne, Paris Cité, and LLACAN, Villejuif, France Karsten Legère, University of Gothenburg, Sweden Kristin Vold Lexander, University of Oslo, Norway Zelealem Leyew, Addis Ababa University, Ethiopia Friederike Lüpke, University of Helsinki, Finland Lutz Marten, School of Oriental and African Studies, University of London, UK Gudrun Miehe, University of Bayreuth, Germany Wilhelm J. G. Möhlig, University of Cologne, Germany Maarten Mous, Leiden University, the Netherlands Victoria Nyst, Leiden University, the Netherlands David Odden, Ohio State University, Columbus, OH, USA Jonathan Owens, University of Bayreuth, Germany Helma Pasch, University of Cologne, Germany Doris L. Payne, University of Oregon, Eugene, OR, USA Victor Porkhomovsky, Russian Academy of Sciences and Lomonossov State University, Moscow, Russia Nicolas Quint, LLACAN, Villejuif, France David Roberts, Independent Researcher, Togo Andrij Rovenchak, Ivan Franko National University of Lviv, Ukraine Maren Rüsch, University of Cologne, Germany
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
xxxiv
Coffi Sambiéni, University of Abomey-Calavi, Benin David L. Schoenbrun, Northwestern University, Evanston, IL, USA Henning Schreiber, University of Hamburg, Germany Ingse Skattum, University of Oslo, Norway Gabriele Sommer, University of Bayreuth, Germany Lameen Souag, LACITO, Paris, France Kembo Sure, Moi University, Eldoret, Kenya Catherine Taine-Cheikh, LACITO, Paris, France Mauro Tosco, University of Turin, Italy Martine Vanhove, LLACAN, Villejuif, France Clarissa Vierke, University of Bayreuth, Germany Rainer Vossen, Goethe University, Frankfurt am Main, Germany Valentin Vydrin, LLACAN, Villejuif, France Jochen Zeller, University of KwaZulu-Natal, Durban, South Africa Ulrike Zoch, Goethe University, Frankfurt am Main, Germany
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
..............................................................................................................
INTRODUCTION ..............................................................................................................
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
.
A , languages are said to exist on earth, nearly one-third of them are located on the African continent. According to a widely accepted hypothesis, African languages can be classified into four larger genetic units (so-called phyla); these have been labeled Afro-Asiatic, Niger-Congo, Nilo-Saharan, and Khoisan. Niger-Congo represents the largest, Khoisan the smallest unit. All phyla subdivide into a number of language families. However, a good number of African languages present themselves as hardly classifiable and are, therefore, considered unaffiliated for the time being. A few—more or less successful—attempts have been made to reconstruct the histories of African language phyla. There have also been ventures to establish genetic links between language phyla, such as Niger-Congo and Nilo-Saharan, but the results were not overwhelmingly convincing. For sure, the most promising historical reconstructions were attained at the language family and/or subgroup level. This volume is a reference book, providing a set of descriptive, typological, historicocomparative, sociolinguistic, and other analytical statements about African languages. It is meant to meet the interest of different populations of readers, especially students, Africanist as well as general linguists, but also non-linguistic scholars specializing in Africa. A large amount of the data presented in the book is drawn from the authors’ own field experience and collections assembled over the last thirty years approximately. The book consists of nine parts divided into seventy-seven chapters, as described below.
. P I: I
.................................................................................................................................. This chapter explains the goals and describes the overall structure of the book. It gives reasons for the range of eight major subject areas (Parts II through IX) into which the work is divided, and that are considered to be of particular relevance in African linguistics today.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
. P II: D
.................................................................................................................................. This part deals with the major domains of grammar. The four chapters by Michael Kenstowicz on phonetics and phonology (chapter ), David Odden on tonology (chapter ), Klaus Beyer on morphology (chapter ), and by Jochen Zeller on syntax (chapter ) each summarize the current research situation and contain a wealth of vivid examples illustrating occurrences and, at times, patterns of distribution of subject-specific phenomena in Africa. These presentations are of fundamental importance for the comprehension of the book as a whole.
. P III: L
.................................................................................................................................. Three chapters are devoted to aspects of language comparison that partly implies but at the same time goes beyond genetic relationship. While the overall discussion of structural language types by Rainer Vossen (chapter ) appears to be a compelling task of a handbook such as the present one, dialectology and language geography are hardly ever taken into account in manuals on African linguistics. Mena Lafkioui (chapter ) is trying to come up to current expectations against the background of a remarkable upswing in the field of linguistic variation on the African continent. Chapter , by Ludwig Gerhardt, touches upon the fascinating research history of African language classification. It is regarded a prerequisite of understanding current attempts at linguistic taxonomy within and across the four language phyla.
. P IV: L
.................................................................................................................................. As shown by Ludwig Gerhardt in chapter on the history of African language classification, the situation is far from settled. Part IV, therefore, takes the “Greenbergian” classification into four phyla for the African continent (Afro-Asiatic, Niger-Congo, Nilo-Saharan, and Khoisan) as a starting point, but focuses on well-established families which themselves may or may not be part of one of these larger phyla, as for Mande or Ubangi (and their disputed Niger-Congo affiliation), or Khoisan, which is primarily an areal grouping according to specialists. As shown by the chapters on these smaller families in Part IV, some authors focus more on typological variation within these smaller families, whereas others are more concerned with subclassification issues. The state of the art in our understanding of Africa’s largest family, Niger-Congo, is discussed by Jeff Good (in the introductory chapter ) with a special focus on BenueCongo. The tremendous genetic and typological diversity within this family becomes clear from the discussion of Atlantic by Friederike Lüpke (chapter ), Mande by Henning
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Schreiber (chapter ), Kwa by the late Mary Esther Kropp Dakubu (chapter ), Gur by Gudrun Miehe (chapter ), and Bantu and Bantoid by Lutz Marten (chapter ). These are followed by chapters on what the late Joseph H. Greenberg claimed to be two coordinate branches of a Niger-Congo branch, Adamawa, by Ulrich Kleinewillinghöfer (chapter ) and Ubangi (called “Eastern” by Greenberg), which is surveyed by Helma Pasch (chapter ). The survey of Niger-Congo is concluded by a discussion of the current state of the art regarding Kordofanian, by Nicolas Quint (chapter ). As these various chapters should show, considerable progress has been made in recent times with respect to the investigation of a range of Niger-Congo languages, also from a historical perspective. The best-established genetic grouping on the African continent is formed by AfroAsiatic, as discussed in the overview by Victor Porkhomovsky (chapter ). The current state of the art for Egyptian is treated by Balázs J. Irsay-Nagy (chapter ), whereas Maarten Kossmann presents the reader with an update on Berber (chapter ). The subfamily with perhaps the largest internal diversity is formed by Cushitic, whose four subgroups are discussed in chapter by Mauro Tosco (East Cushitic), chapter by Martine Vanhove (North Cushitic), chapter by Zelealem Leyew (Central Cushitic), and chapter by Roland Kießling (South Cushitic). The genetic diversity within Omotic (referred to as West Cushitic by Greenberg ) is also considerable, as shown by Bernhard Köhler (chapter ), but its status as a branch of Afro-Asiatic is no longer disputed. The largest subgroup in terms of number of languages is formed by Chadic, for which Bernard Caron (chapter ) gives the reader an update. Given the rather dramatic changes within the Ethio-Semitic subgroup of Semitic, these are discussed in chapter by Victor Porkhomovsky. In a brief overview in chapter , Gerrit J. Dimmendaal discusses current research with respect to Nilo-Saharan, the most controversial of Greenberg’s African big-four groupings, sometimes even characterized as a “wastebasket phylum” involving “leftover languages”. Three of the major subgroups of this family are discussed in subsequent chapters: Saharan by Norbert Cyffer (chapter ), Eastern Sudanic by Gerrit J. Dimmendaal and Angelika Jakobi (chapter ), and Central Sudanic by Pascal Boyeldieu (chapter ). The current view among Khoisan specialists is that the name reflects an areal rather than a genetic grouping, as shown by the late Henry Honken (chapter ). Greenberg () also included the Hadza language as a member of the Khoisan family, but this language and a range of languages either included by Greenberg in one of his three major phyla or identified as distinct languages in subsequent research, probably constitute linguistic isolates, as discussed in the final chapter of Part IV (chapter ) by Gerrit J. Dimmendaal.
. P V: L : C
.................................................................................................................................. Africa is sometimes presented as a “hotbed” for noun classes and serial verbs. But, if we leave aside gender-based noun classes, these features are mainly confined to (part of) the Niger-Congo phylum. The tremendous genetic and typological diversity in some areas on the African continent should become clear from the description of language structures in Part V (Map . shows the geographical distribution of these languages). The sketch of the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
. Geographical position of languages described briefly in Part V
Atlantic language Bom-Kim by G. Tucker Childs (chapter ) shows a language which has inherited the prototypical Niger-Congo noun class system as well as the system of verbal derivational markers. As shown by Valentin Vydrin in his structural survey of the Mande language Dan (chapter ), this language is typologically and grammatically distinct from Niger-Congo languages like Bom-Kim, or from the Gur language Biali, whose features are discussed by Coffi Sambiéni (chapter ). The Jukunoid (i.e. Benue-Congo) language Yukuben, as discussed by Rose-Juliet Anyanwu (chapter ), has retained the noun class system that is typical of the family it belongs to, but like other languages in the area it developed verb serialization. In this respect, Bende, as described by Yuko Abe (chapter ), is more conservative and characteristic of the Narrow Bantu group within Benue-Congo to which it belongs.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Whereas in Greenberg () Adamawa and Ubangi (called “Eastern” by him) are treated together as a sub-branch of Niger-Congo, more recent studies could not confirm this closer genetic affinity. Structural properties of a typical Adamawa language, Waja, are discussed by Ulrich Kleinewillinghöfer (chapter ), whereas Helma Pasch presents an outline of the typologically rather distinct Ubangi language Zande (in chapter ). Chapter is devoted to the study of Zenaga, a less well-known Berber language spoken in Mauretania and described by Catherine Taine-Cheikh. Three of the four primary branches of Cushitic (with only North Cushitic missing) are represented by the East Cushitic language Sidaama, as described in chapter by Kazuhiro Kawachi; by the Central Cushitic language Kolisi in chapter , authored by Zelealem Leyew; and by the South Cushitic language Iraqw in chapter , contributed by Maarten Mous. Finally, in chapter , an outline grammar of Wandala, a Chadic language, is provided by Zygmunt Frajzyngier. The Nilo-Saharan family is represented by the Nilotic language Kumam, whose main structural features are described by Osamu Hieda (chapter ), whereas major structural properties of the Surmic language Baale are outlined by Gerrit J. Dimmendaal (chapter ). The genetic position of the Songhay cluster in West Africa as a Nilo-Saharan subbranch or as an independent language family is controversial. As shown by Lameen Souag (chapter ), Songhay forms a dialect continuum with considerable grammatical variation as a result of contact with other language families. Structural features of one of the few surviving Central Khoisan languages, Cara, are discussed by Rainer Vossen (chapter ), whereas kX’egwi, a moribund, if not extinct member of the Southern Khoisan branch, is treated by the late Henry Honken (chapter ). The genetic affiliation of Hadza in Tanzania is controversial, but the affiliation of another language with clicks in the same country, Sandawe, with Central Khoisan is not or no longer contested. Structural properties of this language are treated by Helen Eaton (chapter ).
. P VI: L, ,
.................................................................................................................................. Whereas the first five parts address issues of core linguistics, both synchronically and diachronically, the following four major sections address interdisciplinary research involving language and its interaction with cognition and culture. The important role played by ideophones in day-to-day interaction was already noted by early researchers like Diedrich Westermann. They have since received much attention, as shown by Christa Kilian-Hatz in her historical survey (chapter ). The division of the color spectrum and the constraints set by our cognitive system, or the role played by cultural selection is a classic topic in the ongoing debate between “universalists” and “relativists” in anthropological linguistics, as discussed by Doris Payne (chapter ). Experiencer constructions have been subject to a debate in cognitive linguistics. Their semantic properties are central to chapter , by Ulrike Zoch, who shows that languages vary with regard to the available means for codifying emotional and physical state situations.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Ethnobotanical studies, as discussed in Karsten Legère’s contribution (chapter ), have also been the subject of debate between universalists and relativists, in particular the question whether indigenous groups follow a taxonomic model parallel to a Linnaean or Darwinian taxonomy of the natural environment in biology, with a set of taxa specifiable by relations of affinity and contrast, or a more functional (natural core) model, with gradable membership and categorial contiguity. This latter model accounts better for the fact that phenomena in the natural environment are not necessarily categorized (or lexicalized) and may be left unclassified instead. A further topic of interest with respect to language studies involves its social functions, in particular concealment and marking of group membership, as shown by Patricia Friedrich in what she calls “distinctive languages” in chapter , where additional concepts for such (technical) languages are discussed, such as guild languages, ritual languages, reverence languages, and urban youth languages. When it comes to the organization of talk-in-interaction, a range of features are shared by speech communities across the world, as demonstrated by Maren Rüsch in her survey of conversation analyses (chapter ). But as shown in the same contribution, at the same time there are also interesting differences, for example concerning pauses or the extent to which overlap is allowed in turn-taking. These, as well as differences in “body language”, provide interesting new insights into language and culture-dependent strategies. The distinctive role played by African languages in the development of grammaticalization models is one of the themes discussed by Axel Fleisch, who further takes up cognitive research on African languages as a central theme in chapter .
. P VII: L
.................................................................................................................................. Writing traditions on the African continent go back at least , years. But apart from the well-known scripts developed during Pharaonic times, or the more recent Arabic and Latin scripts, there are different indigenous scripts, as shown by Andrij Rovenchak and Solomija Buk (chapter ). Orthographies play an important role in language policies, in particular for educational and administrative purposes. These and related issues with respect to current language politics are discussed by Kembo Sure (chapter ). In the next chapter (), Ingse Skattum addresses the actual role played by African languages in educational systems, in particular the development of written material as part of corpus planning. Whereas several languages are thriving as spoken or written regional or national contact languages, others are endangered. In chapter , James Essegbey discusses modern initiatives to document (or sometimes revitalize) such endangered languages on the African continent. Language loss or obsolescence is just one outcome of language contact. These include the emergence of new varieties of African or European languages primarily in urban areas, as shown by Ellen Hurst in chapter . As Klaus Beyer shows in his general survey of language contact phenomena in chapter , a number of other outcomes are possible. As shown by Maarten Mous in chapter , one of these is the emergence of mixed languages, which are also sometimes referred to as syncretic or intertwined languages, as the term “Mischsprache” (mixed language) is associated with the colonial era in the field of
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
African linguistics, more specifically the end of the nineteenth and the beginning of the twentieth century, when language and race (as well as culture) were assumed to correlate with each other, resulting in a distinction between pure languages (and races) and mixed languages (and races). The early days of Western expansion and colonialism also resulted in a diaspora of African languages mainly due to the massive transportation of slaves into the Americas. Andrea Hollington presents a survey of these languages in chapter , which have also been maintained in other parts of the world. Another result of this diaspora was the creation of pidginized and creolized varieties of European languages, as discussed in the chapter by Gabriele Sommer (chapter ). As with conversation analysis, the investigation of African sign languages constitutes a relatively new, important area of research in the field of African linguistics; Victoria Nyst introduces the reader into this emerging field in chapter . Next to major lingua francas emerging from contact situations, such as Hausa and Swahili, Arabic is among the most important in a range of countries in the Sahel region. Its expansion also gives rise to various new varieties of this language on the African continent, as shown in the survey by Jonathan Owens (chapter ). Variation in language, due to dialect differences or the use of different registers, also raises the question of standardization, a topic which is central to the contribution by Elke Karan and David Roberts (chapter ). Apart from written communication, there is the oral dimension, which is still of key importance in an African context, as discussed by Thomas Bearth in chapter on pragmatics. But here, too, there are interesting new developments concerning communications technology, as Kristin Vold Lexander (chapter ) shows in the final chapter on language and society.
. P VIII: L
.................................................................................................................................. The study of languages also plays an important part in the overall reconstruction of the African past. Given the relative scarcity of written testimonies, especially for distant time periods and eras, historical linguistics in cooperation with other diachronically oriented disciplines may serve as one appropriate tool for uncovering historical events. For this reason, two chapters are dedicated to the interplay between language and history: David L. Schoenbrun’s essay on words, things, and meaning (chapter ) and Koen Bostoen’s strongly methodology-oriented analysis of linguistic–archaeological relations (chapter ).
. P IX: L
.................................................................................................................................. In general, Africa and its indigenous populations are viewed as a stronghold of orality. It is, therefore, imperative to devote a major section of this book to language and its role in African orature (or “oral literature”). While chapter by Wilhelm J. G. Möhlig on narratives gives an overview of types of oral literature and story-telling genres,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Sebastian K. Bemile (chapter ) gets right to the bottom of the broad field of proverbs that play an extraordinarily important role in African oral discourse. The chapter shows, inter alia, the intercultural relevance of many proverbs, even far beyond the borders of the African continent. Part IX concludes with a chapter () by Clarissa Vierke on African poetry, touching upon various types of oral poetry (e.g. praise, epic-heroic, war and hunting, religious poetry, and songs) as well as written poetry and manuscript cultures.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
..............................................................................................................
DOMAINS OF GRAMMAR ..............................................................................................................
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
.
. I
.................................................................................................................................. W the rise of generative grammar and other approaches that view the analysis of individual languages through the lens of a general theory, the search for universals, both formal and substantive, and limitations on the scope of linguistic variation have become primary research objectives. The languages of Africa have played a very prominent role in this endeavor. In this chapter we survey a few of the highlights from segmental and prosodic structure. First we review some of the evidence that supports extending autosegmental representations from tone to segmental and prosodic phonology. We then survey the contributions of African languages to the typologies of vowel harmony, the treatment of vowel and consonant clusters, syllabification, reduplication, and phonological phrasing.
. A
.................................................................................................................................. One of the fundamental breakthroughs in the development of phonology was Goldsmith’s () proposal that tone be represented on a separate level or tier from the segments that bear it. This notation allows one–many and many–one relations between tones and segments to be formally expressed. Many of the data that originally motivated such representations derive from African languages including Igbo, Mende, and Tiv (Goldsmith ), as well as Hausa (Leben ) and Margi (Williams ). To take a simple example, autosegmental representation permits an eminently simple formalization of the widespread assimilation rule of High-Tone-Spread, as in the two-tone Togolese Gur language Lama (Kenstowicz et al. ) where the combination of a High (H) tone followed by a Low (L) is realized as a High plus Fall: yó ‘child’, rì ‘mother’, yó rî ‘child’s mother’. This assimilation can be expressed as an extension of the H [+hi] of yó to the vowel of rì. The resulting
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
representation () includes a H–L sequence on a single syllable to represent the falling tone of rî as well as a H that extends over two successive syllables. ()
segments
yó
rí
yó
rî
tones
H
L → H
L
This representation finds independent motivation in processes of downstep that typically treat the spread H as a single unit. Likewise tonal polarity of affixes supports the analysis of rising and falling (contour) tones as bitonal sequences LH and HL, respectively, associated with a single syllable: cf. Margi à sá gù ‘you go astray’, á dlà gú ‘you fall’, á və̌l gù ‘you fly’ (Hoffmann ; Pulleyblank ). Once introduced, autosegmental representations were quickly extended from tone to segmental phonology. Here as well African languages played a pivotal role. The phenomena elucidated by such representations include vowel harmony (see section . below) as well as the interface between sound segments and prosodic structure. We survey a few examples. McCarthy () showed that the consonants composing the lexical roots of Semitic paradigms such as Arabic katab-a ‘he wrote’, ya-ktub-u ‘he writes’, ma-ktab ‘office’, ma-kaatib ‘offices’, can be thought of as comprising a radical /ktb/ that is associated with CVCVC, CVCCVC, CVCVVCVC templates parallel to the way that tones map to syllables. This notation predicts that a single segment can have multiple exponents. One of the most stunning confirmations of this prediction is found in Ethiopian Semitic Chaha (McCarthy ). First, just as a tone may show up on an adjacent syllable when its segmental associate has been deleted (as in Lomongo /là bɔ́nà/ > [lɔ̌nà] ‘with baby’, where the L of là has combined with the H of bɔ́nà; Odden ), vocalic features exhibit a similar behavior. The SG.M object marker is a suffix -u in Levantine Arabic; but in Chaha it is a [+round] autosegment that searches in the preceding root to find the closest non-coronal consonant that can host it (). ()
nəkəb ‘he found’ kəfət ‘he opened’ qəTər ‘he killed’
nəkəbw ‘he found it.’ kəf wət ‘he opened it.’ qwəTər ‘he killed it.’
(cf. Levantine ʔaṭal, ʔaṭal-u)
Bi-radical roots such as /sk/ ‘plant’ and /gm/ ‘chip’ () have a single consonantal segment that is associated with two consonantal slots in the CVCVC template (a “long-distance” geminate). This representation suggests that when such a segment hosts the labial autosegment, then the labialization will be expressed on two consonants simultaneously. Forms such as the impersonal (which also takes the [+round] suffix) are a striking confirmation of this prediction: səkwəkw. ()
base səkək gəməm root tier template
impersonal səkwəkw ‘plant’ w w gəm əm ‘chip’ s k
CəCəCə
s kw
CəCəC
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Kenstowicz () finds evidence for similar representations at the segmental level in Tigrinya. In this Semitic language of Eritrea a velar stop is spirantized after a vowel: cf. /rkb/ in mi-rkab ‘find’ but rəxəb-ə ‘he found’. Postvocalic geminate stops do not spirantize: yi-rəkkib ‘he finds’. But successive velar stops created by morpheme concatenation do: /mirak-ka/ > mirax-ka ‘your calf ’. This difference is explained by representing a true geminate as a single, multiply linked segment, while “fake” geminates arising from concatenation are singly linked (). A uniformity requirement that each link in the geminate must satisfy the rule blocks spirantization in true geminates because the second link is not postvocalic. ()
k
k k
C C true geminate
C C fake geminate
The distribution of these two alternative representations for a phonetically long segment became a topic of considerable discussion in the phonological literature under the heading Obligatory Contour Principle (OCP). One early proposal by McCarthy () was that the multiply linked structure is normally restricted to tautomorphemic adjacent consonants, while those arising from morpheme concatenation are singly linked. Tigrinya supports this proposal. In addition, Tigrinya has a rule that completely assimilates the stop of the passive prefix tə- to a following consonant: /yi-tə-rəkəb/ > /yi-t-rəkəb/ > [yirrəxəb] ‘find (passive jussive)’. This is naturally represented in autosegmental notation by delinking the stop and associating its skeletal C slot to the following consonant, creating a heteromorphemic but multiply linked geminate (). This representation predicts that a velar stop will not spirantize, a prediction that is confirmed by forms such as ‘open’ /yi-tə-kəfət/ > [yikkəfət], *[yixkəfət]. ()
segment
y i -t - r ə k ə b
template
CVCCVCVC
Another reflex of the multiply linked representation for geminates is their unexpected failure to be split apart by rules of epenthesis (geminate integrity). Guerssel’s (, ) studies of Ait Segrouchen Berber (ASB) was one of the first analyses of this phenomenon in autosegmental terms. In ASB a schwa is inserted between the first and second members of a triconsonantal cluster, as in paradigms such as s wudi ‘with butter’ but sə tmaziGt ‘in Berber’ and sə zzit ‘with oil’. When the first two consonants of the cluster form a geminate, then epenthesis is blocked: while /ams-tən/ ‘rub them’ is realized as [aməs-tən], /ass-tən/ ‘tie them’ is [asstən]. If the geminate is represented as in (a), then the epenthetic vowel cannot be spelled out on the segmental tier without crossing an association line and thereby yielding an ill-formed autosegmental representation. Guerssel shows that in ASB heteromorphemic fake geminates can be split, while those arising from complete assimilation cannot. In the construct state a prefixal vowel is deleted and any CCC cluster is broken even when the first two consonants are identical: /t-a-∫i∫aw-t/ ‘chick’ > [t∫i∫awt], /t-a-tbirt/ ‘pigeon’ > /ttbirt/ > [tətbirt] (b). But the genitive prefix n- completely assimilates
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
to a following sonorant consonant: n-taddart ‘house’ but l-litub ‘book’, w-wadu R‘wind’. This process creates a multiply linked geminate that blocks epenthesis: cf. nə-tratt a ‘net’ but l-lwiski (*ləlwiski) ‘whisky’ and w-wtəm (*wəwtəm) ‘male’ (c). ()
a.
s
ə s ə
→
C C
CVC
b. t t b i r t → CCCVCC
tə tbi r t CVCCVCC
c. n l w i s k i
n l w i s k i →
C C C V C C V
C C C V C C V
To take one more example, Clements () proposes that the Bantu phenomenon of compensatory vowel lengthening that often accompanies the processes of high-vowel devocalization and prenasalization can be insightfully formalized as the reassociation of the vocoid or nasal to the adjacent consonant to create a complex single consonant; its erstwhile slot in the skeletal timing tier is filled by the adjacent vowel. The examples in () from Luganda illustrate. The three different variants of the class prefix mu- seen in (a) arise from the processes depicted in (b). ()
a. mu-limi ‘farmer’ mw-aana ‘child’ muu-ntu ‘person’
b. segmental tier
mu l i m i
m u a n a
m u n t u
skeletal tier
XXXXXX
X X X X X
X X X X X
. S
.................................................................................................................................. Vowel systems in the African languages range from the minimal /i u a/ plus predictable schwa in Magrebi Arabic and Berber to more elaborated systems with three or four degrees of height. The five-vowel [i u e o a] is probably the most common. But even in languages with the canonical five vowels the distribution of mid vowels may be restricted to particular positions in roots, such as the initial syllable in Shona (Beckman ). Many African languages elaborate this basic stock with contrasts in length and nasality. Less often does one find contrasts in rounding or voice quality, such as those prevalent in Turkic and East Asian languages. Consonant systems include the generic stops, fricatives, nasals, as well as prenasalized stops; pharyngeals are largely restricted to the Afro-Asiatic family. Doubly
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
articulated labial-velars kp and gb are widely distributed in the Niger-Congo languages. The Khoisan and some neighboring Bantu languages are justly famous for their clicks. Besides the ubiquitous opposition in voice, laryngeal distinctions include glottalized ejectives and implosives. On the other hand, three-way contrasts of voiceless, voiced, aspirate are not so common. One of the most elaborated inventories is found in !Xóõ with consonants ( different clicks and non-clicks) and vowels with contrasts in nasality, breathy voice, and pharyngealization overlaid on the canonical five-vowel inventory (Traill ). On the prosodic level most African languages eschew syllables with complex onsets or codas. And perhaps because tonal distinctions are so prevalent, stress plays a less central role in the segmental or prosodic phonology compared to languages like English.
. V
.................................................................................................................................. One of the most common supplementary vowel height distinctions is provided by the feature advanced tongue root [ATR]. It is employed in languages of West Africa as well as Bantu and Nilotic. Using cine-X-ray instrumentation, Ladefoged () identified its phonetic correlate as pharyngeal cavity expansion by tongue root advancement, a discovery that has been reconfirmed with ultrasound by Gick et al. (). The major acoustic reflex is a lowering of the first formant. In a survey of African languages, Casali () finds that over half () freely distribute the [ATR] feature over the canonical five-vowel system to yield a ten-vowel inventory: [+ATR] [i u e o ə] and [–ATR] [ɪ ʊ ε ɔ a] (a prime example of Clements’s Feature Economy). The rest restrict the contrast to either the mid vowels [e o] vs. [ε ɔ] or the high vowels [i u] vs. [ɪ ʊ], with the former outnumbering the latter to . In no case are the higher vowels passed over in favor of a low-vowel [ə] vs. [a] distinction. While lexical contrasts in [ATR] are often restricted to particular positions such as the root, the feature is frequently extended to the edges of the word by vowel harmony. Akan exhibits such “root control” (Clements ) with both prefixes and suffixes assimilating to the root: o-fiti-i ‘he pierced it’ vs. ɔ-cɪrε-ɪ ‘he showed it’. The low vowel in Akan is not the site of an [ATR] contrast. It freely combines with both [+ATR] and [–ATR] vowels in the root. Moreover, it triggers [–ATR] on prefixal vowels when it is root-initial and on suffixal vowels when root-final (a). Such behavior was crucial to the rejection of the Root Marker Theory of vowel harmony (Lightner ) in favor of an autosegment that can have scope over several successive syllables but where the root is factored into separate harmonic spans, as in (b). ()
a. kari ‘weigh’ yarɪ ‘be sick’ ɔ-kari-i ‘He weighed.’
b.
[–A] [+A] ɔ -kar i-i
bisa ‘ask’ pɪra ‘sweep’ o-bisa-ɪ ‘He asked.’
[+A] [–A] o-bisa-i
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
In his survey of [ATR] harmony, Casali () finds that while [+ATR] is the dominant feature for high vowels, [–ATR] is dominant for mid vowels. For example, Kinande has an [i ɪ u ʊ ε ɔ a] inventory. Mutaka () reports that [+ATR] may spread backwards across a word boundary to a high vowel ( /εmɪtɪ mikuhi/ > [εmɪti mikuhi] ‘short trees’) but [–ATR] cannot (ekikəli kɪrɪ ‘tall woman’). On the other hand, for languages with an [ATR] contrast in the mid vowels, [–ATR] predominates. The Adangbe dialect of Ewe (Clements ) furnishes an example; the topicalization particle -é assimilates to a preceding [ε ɔ]: əwo-é ‘it’s you’, but ə̀ vlε-ε´ ‘it’s a weaver bird’, ə̀ sɔ́-ε´ ‘it’s a horse’. In the typology of harmony, vowels lacking a harmonic counterpart may block the spread of the harmonic feature (opaque vowels) or they may allow the change to pass through without themselves showing an effect of the harmonic feature (at the phonetic surface). Kinande exhibits the former behavior (Valinande ; Mutaka ; Gick et al. ; Kenstowicz ). It has seven vowels with an [ATR] contrast in the high vowels: [i ɪ u ʊ ε ɔ a]. A suffixal high vowel like the applicative alternates between [+ATR] and [–ATR] as a function of the root (a). But when combined with disyllabic CVCaC roots with an [ATR] contrast in the first syllable (b), the applicative shows a consistent [–ATR]. This has been explained with the proposal that [a] has a [–ATR] autosegment that blocks spreading by the ban on line crossing (c). ()
a. base eri-lim-a εrɪ-lɪm-a εrɪ-sab-a
applicative eri-lim-ir-a εrɪ-lɪm-ɪr-a εrɪ-sab-ɪr-a
‘exterminate’ ‘cultivate’ ‘ask’
b. eri-liban-a eri-liban-ɪr-a ‘disappear’ εrɪ-hɪtan-a εrɪ-hɪtan-ɪr-a ‘get angry’
c. l i b a n i r a +A –A
*liban ira +A –A
Yoruba provides an intriguing example of transparent vowels as well as a directionality effect (Orie ). It contrasts mid vowels for [ATR]. Adjacent mid vowels must agree for this feature: oko ‘farm’, ɔkɔ ‘husband’, ὲdɔ̀ ‘liver’, ògèdè ‘incantations’. The low and high vowels are compatible with both values of [ATR] to their right: abɔ ‘female’, ahoro ‘ruins’, igbó ‘bush’, itɔ́ ‘saliva’. But they impose restrictions on vowels to their left. The low vowel requires a [–ATR] mid vowel to precede: ɔba ‘king’, *oba. High vowels show a dialect difference. In Standard Yoruba mid vowels that precede a high vowel must be [–ATR] (ὲbi ‘guilt’, ὲrù ‘fear’), while in the Ife dialect mid vowels surface as [+ATR] in this context (èbi ‘guilt’, èrù ‘fear’). Thus, both dialects impose harmony from right to left with the low vowel triggering the expected [–ATR]. For the high vowels Ife imposes harmony based on the surface [+ATR] value of [i] and [u] (e.g. èbi) while Standard Yoruba does not and instead allows the default [–ATR] to appear on the preceding mid vowels (e.g. ὲbi). Trisyllabic words with a medial high vowel flanking mid vowels allow the transparency phenomenon to be tested. In the Ife dialect an initial mid vowel assimilates [–ATR] across the high vowel: oríwo ‘boil, tumor’ vs. ɔ̀títɔ́ ‘truth’. It thus appears that the [ATR] contrast in the final vowel has been copied into the initial mid vowel, ignoring the intervening high vowel (a). Such “gapped” structures contradict the basic premise of autosegmental theory
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
that a harmony span constitutes a temporal continuum and violate the line-crossing ban that enforces it (b). () a. ɔ t i t ɔ
–A +A –A
b. ɔ t i t ɔ
+A –A
Curiously, Standard Yoruba shows [+ATR] in same context: oríwo vs. òtítɔ́. Nevins () interprets this dialectal difference as a parameter on the features an underspecified vowel archiphoneme searches for in order to obtain a value for [ATR]. In Ife only the vowels that lexically contrast for [ATR] may contribute a value, thereby forcing harmony from the more distant mid vowel in /Otitɔ/ > ɔ̀títɔ́ . In Standard Yoruba any vowel may do so, and consequently the adjacent high vowel is preferred, since it is closer: /Otitɔ/ > òtítɔ́. An alternative analysis postulates that Ife harmonizes from a higher sonority mid or low vowel (cf. ɔ̀kuta ‘stone’) if one is present, but defaults to a high vowel trigger if no lower vowel is available (èrù ‘fear’). Under either account Standard Yoruba òtítɔ́, òkuta but ὲbi, ὲrù remains puzzling: medial high vowels trigger harmony but word-final ones do not. The African vowel harmony systems exhibit directional asymmetries as well. In Kinande a [+ATR] high vowel spreads [+ATR] to all preceding vowels, creating [e o ə] variants that do not occur independently: compare εrɪ-hεk-a ‘carry’ but mo-twə-hek-ir-ε ‘recent past’. As we have seen, the low vowel blocks left-to-right spread of the harmonic feature (recall eri-liban-ɪr-a). Hyman () sees this directional asymmetry as the phonologization of a common phonetic coarticulation asymmetry: vowel height transitions are typically anticipated in the preceding syllable rather than carried over into the following syllable. This behavior contrasts with tone: tonal height distinctions are more typically carried over to the following syllable rather than anticipated in phonetic coarticulation. The phonological counterpart is that H and L tones more commonly spread to the right rather than to the left. Finally, the complex [ATR] harmony system in the Nilotic language Lango (Bavin et al. ) has attracted considerable theoretical attention. The assimilation of [+ATR] from an adjacent syllable is influenced by direction as well as the height and backness of the target and trigger and whether or not a consonant cluster intervenes. For each factor there is a phonetically grounded (dis)preference (Archangeli and Pulleyblank ), which, however, can be overridden by a combination of preferences pulling in the opposite direction—a socalled “gang effect”. For example, a back high vowel raises a preceding mid vowel /lε-wu/ > [lewu] ‘axe’ (PL), but not across a consonant cluster /dεk-wu/ > [dεkwu] ‘stew’ (PL). On the other hand, a front high vowel does raise the root vowel in this context: /dεk-ni/ > [dekki] ‘stew’ (SG). In Smolensky’s () Optimality Theoretic analysis, a general harmony constraint requires two adjacent syllables to agree for [ATR]. Harmony can be overridden by a combination of markedness dispreferences such as spreading across a cluster and from a back-vowel trigger. Neither markedness constraint is sufficient to block harmony by itself (Faithfulness » Markedness) but in tandem they do do so (Markedness1 & Markedness2 » Faithfulness). Potts et al. () offer an alternative analysis where constraints are assigned weights so that the summation of two or more markedness dispreferences may override the weight of the harmony constraint.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
. V–V C–C
.................................................................................................................................. Morpheme and word concatenation can lead to underlying sound sequences that are at variance with the CV syllable shape favored by so many of the languages of Africa. The different ways in which the CV optimum is enforced offers a rich typological resource. Here we review two of the most prevalent cases—the resolution of hiatus and NC clusters. The sound changes that resolve hiatus include elision, devocalization, coalescence, epenthesis, and diphthongization. The first three are particularly favored in the African context and have been studied in detail by Casali (). He reports two major findings. First, for elision Casali finds a strong preference to delete V1 in hiatus. This preference may be overridden when V1 belongs to a lexical category: Preserve-Vlex » Preserve-V2 » Preserve V1 in Optimality Theoretic terms. Emai (Benue-Congo, Nigeria; Schaefer ) illustrates this profile well. In /ɔli ebe/ ‘the book’ > [olebe] and /ebe ɔna/ ‘this leaf ’ > [ebena], the vowels of the root /ebe/ are preserved and affixal vowels elide. The same preference appears in /ɔli oa ɔna/ ‘the house this’ > [ɔloana]. But in /oa isi ɔi/ ‘house of his’ > [oasɔi], hiatus at the juncture of the two grammatical morphemes is resolved by deletion of V1. The same V1 deletion preference appears at word junctures as well as between two roots in a compound: /kɔ ema/ ‘plant yam’ > [kema] and /u-kpe-akɔn/ ‘prefix-wash-teeth’ > [ukpakɔn] ‘chew-stick’. A preference for preserving lower vowels in hiatus resolution is found in a number of languages such as ‘Afar (Cushitic, Ethiopia; Bliese ): cale irgice ‘I cut a mountain’ > [calergice], anu okme ‘I ate’ > [anokme], daro akme ‘I eat grain’ > [darakme], adeena uble ‘toothbrush I saw’ > [adeenable]. In many languages lower vowels are longer than higher ones (Lehiste ). The ‘Afar strategy could reflect a preference to retain the more salient (longer) vowel. Casali finds that the preference for lower vowels also governs vowel coalescence. In the following examples from Anufɔ (Kwa, Ghana/Togo/Benin; Adjekum et al. ), the first vowel is devocalized if round and otherwise deleted; but its openness shows up in a lowering of the following high vowel (Preserve-[low] » Preserve[high]): /fa-i/ > [fε:] ‘take it’, /fa-u/ > [fɔ:] ‘take you’, /n-de-u/ > [ndo:] ‘I will take it from you’, /bo-i/ > [bwe:] ‘beat it’, /sɔ-i/ > [swε] ‘carry it’. In many of the Bantu languages, certain noun class prefixes (e.g. class ) terminate in a consonantal nasal archiphoneme that assimilates the place features of the following consonant. Place assimilation is often accompanied by a change in the manner (nasal hardening) as well as laryngeal features of the following consonant. The Zulu data from Halpert () in () illustrate: ()
coalescence /iN-ɓali/ imbali ‘truth’ /iN-thando/ int’ando ‘free will’ /iN-sindiso/ intsindiso ‘salvation’
no coalescence /um-ɓali/ umɓali ‘color’ /um-thando/ umthando ‘love’ /um-∫ikisho/ um∫ikisho ‘friction’
In Zulu the vowel of class mu- has been lost creating mC sequences. Here the nasal retains its place of articulation and the following consonant remains intact as well, which shows that simplification of C in NC depends upon place assimilation. Halpert () analyzes the behavior of the class prefix /iN/ as fusion of the nasal and the following consonant into a
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
single oral gesture drawing on the resources of the Articulatory Phonology model of Browman and Goldstein (). This hypothesis is supported by the finding that assimilated clusters have phonetic durations that approximate a single segment, while the unassimilated m-C is considerably longer. The loss of implosion, aspiration, and continuancy are markedness reductions defined over the complex NC gesture.
. S
.................................................................................................................................. The syllabification process in the Imdlawn Tashlhiyt dialect of Moroccan Berber (ITB) described and analyzed by Dell and Elmedlaoui () is remarkable in its own right and has had a significant impact on phonological theory. Like other Berber dialects, ITB has an /i u a/ vowel inventory. Underlying forms may contain long strings of consonants, which are realized without epenthetic vowels (unlike in Moroccan Arabic). Nevertheless, Dell and Elmedlaoui use evidence from native speaker intuition, grammar-internal rules, as well as metrical verse to show that such strings are parsed into OnsetNucleusCoda sequences just as sequences of consonants and vowels are. As shown by the data in (a), any consonant including a stop is a possible nucleus: ()
a. ikr̩zawn tx̩zn̩t tf̩tk̩t
‘He plowed for you.’ ‘You stored.’ ‘You sprained.’
b. sm̩d ‘add’ vs. z̩dm̩ ‘gather firewood’ xn̩g ‘choke’ n̩gd̩ ‘drown’ Given this fact, for any string there are many potential parses depending on which consonant anchors the nucleus. Unlike in Moroccan Arabic, where syllabification applies directionally, Dell and Elmedlaoui demonstrate that ITB syllabification seeks out local sonority peaks to place the nucleus (b). Their algorithm scans the segmental string searching for segments according to the sonority hierarchy (low vowel > high vocoid > liquid > nasal > fricative > stop) to parse out Onset-Nucleus pairs; any leftover segments are then added as codas at the end of this routine. The syllabification of tizr̩walin ‘those from Tizrwalin’ and disyllabic ratlult ‘you will be born’ vs. trisyllabic ratr̩gl ̩t ‘you will lock’ proceeds as in (). ()
/t-IzrUal-In/ tIzr(wa)lIn (ti)zr(wa)(li)n (ti)(zr̩)(wa)(li)n (ti)(zr̩)(wa)(lin)
/ra-t-lUl-t/ (ra)tlUlt (ra)t(lu)lt ----------(rat)(lult)
/ra-t-rgl-t/ (ra)trglt -----------(ra)(tr̩)(gl ̩)t (ra)(tr̩)(gl ̩t)
low vowel high vowel liquid coda
An important rider to the algorithm is that the parse blocks if it gives rise to hiatus (two successive nuclei). This explains why the final stop in /smd/ (b) is not parsed into a nucleus as well as why for /IħaUltn/ > iħawl ̩tn̩ ‘he made them plentiful’ the third syllable’s nucleus has lower sonority than the onset.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
() a. /smd/ (sm̩ )d -----(sm̩ d)
/zdm/ z(dm̩ ) nasal (z̩)(dm̩ ) fricative -------- coda
b. /IħaUltn/ I(ħa)Ultn *(i)(ħa)(u)ltn (i)(ħa)(wl̩)tn (i)(ħa)(wl̩)(tn̩)
low vowel high vowel liquid nasal
(blocked by hiatus ban)
ITB syllabification presents a significant challenge to the traditional bottom-up, step-bystep construction of prosodic structure, since a small difference at one point in the segmental string can have consequences for more remote positions. This problem provided the original impetus for Optimality Theory (OT; Prince and Smolensky []) in which the grammar chooses among fully formed structures according to a preference hierarchy (here Onset » Harmonic Nucleus) rather than constructing the prosody incrementally by small local changes that apply in isolation from other changes. See, however, Frampton () for an analysis of ITB in a more traditional step-by-step format, as well as Pater () for an analysis in Harmonic Grammar, a framework that employs ranked OT constraints to build the output gradually by a sequence of minimal modifications.
. R
.................................................................................................................................. In the typology of reduplication it is customary to distinguish the reduplicant (copy) in terms of its size and position with respect to the adjacent base. In complete or unbounded reduplication the entire base is copied; bounded or partial reduplication is most commonly restricted to two syllables or two moras. This pattern is evident in the Kinande paradigm in (). ()
eri-huma eri-humwa eri-soloma eri-swa eri-ta ery-ómbola ery-óta
eri-huma-huma eri-humwa-humwa eri-solo-soloma eri-swaswa-swa eri-tata-ta ery-ómbola-mbóla ery-ótot-óta
‘hit’ ‘be hit’ ‘harvest’ ‘grind’ ‘bury’ ‘take discreetly’ ‘bask’
When the base is two syllables then both copy, and when longer, then just the first two syllables can be copied.1 When the base is just one syllable, then it is copied twice to ensure that the reduplicant is disyllabic. Another complication is that when the stem begins with a vowel, then the copying process skips to next syllable in order to avoid creating a V–V
1
When the base is polysyllabic, then reduplication more often blocks, a morpheme integrity condition (Mutaka and Hyman ) that creates a paradigm gap.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
sequence: *eryómbo-ombóla. Such misalignment is found in a number of other Bantu languages such as Hehe (Odden and Odden ). The domain of reduplication is argued by Downing () and others to be defined by the P-stem, a phonological category derived from the morphological stem by shifting the boundary to align with a more unmarked CV syllable. Jones () shows that several other processes of Kinande phonology are sensitive to this P-stem but that reference to the morphological stem is also needed. ()
M
M
hu ma
o mbo la
P
P
Reduplicants are often the site of simplifications so that the copy is not exact, a phenomenon termed “the emergence of the unmarked” (TETU) by McCarthy and Prince () and formalized in Optimality Theory by markedness constraints dominating the BaseReduplicant Faithfulness that demands an exact match. The complete verbal reduplications in Shona vs. Chewa illustrate this point nicely (). In Shona the reduplicant fails to copy the marked H tone of the stem (*H » Ident-BR) while Chewa opts for faithfulness (Ident-BR » *H). ()
Shona (Odden ) handáká-bikísa handáká-bikísa-bikisa handáka-tóréserá handáka-tóréserá-toresera
‘I didn’t make cook frequently.’ ‘I didn’t make take for frequently’
Chewa (Myers and Carleton ) nda-namizá da-namizá-namizá ‘I have deceived repeatedly’ ndíma-sangalátsa ndíma-sangalátsa-sangalátsa ‘I please repeatedly’ The TETU phenomenon is also found in the Fong-Gbe Ewe dialects described by Capo (), as illustrated in the paradigms in (a). Wací copies the following base faithfully except for H tone, while Gen additionally simplifies CR onsets and Pecí blocks nasality. The differences between the dialects are formalized by the ranking of reduplicative faithfulness relative to various markedness constraints that disprefer high tones, complex syllable onsets, and nasal vowels (b). ()
a. Wací ɖuɖu za̰zá̰ kla̰klá̰
Gen ɖuɖu za̰zá̰ ka̰klá̰
Pecí ɖuɖu ‘eat’ zazá̰ ‘use’ kaklá̰ ‘separate’
b. Wací: *H » Ident-BR » *Complex Onset, *V̰ Gen: *H, *Complex Onset » Ident-BR » *V̰ Pecí: *H, *Complex Onset, *V̰ » Ident-BR When reduplication is formalized as a constraint on candidate output forms rather than as a copying rule, then the possibility of back-copy arises in which the base is altered to bring
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
it closer to the reduplicant. This situation can occur when the reduplicant undergoes a sound change triggered by an element outside the [base-reduplicant] or [reduplicant-base] structure or by the base itself. Kenstowicz and Petros Banksira () argue for such a case in the realization of verbal radicals in Chaha. As illustrated by the data in (), [k] is derived from /x/ when followed by a continuant /f s z ʕ/. () /frx/ /xrm/ /xfr/ /rxs/ /fxr/
yə-frəx yə-xrəm yə-kfir yə-ŋkis yə-fxər
‘tolerate’ ‘spend year’ ‘separate’ ‘bite’ ‘multiply’
Chaha has several different reduplications defined over its verbal radicals shown in (). In /C1C2/ > C1C2C1C2 both radicals are reduplicated. For /xs/ and /xyf/ roots, each /x/ is followed by a fricative that triggers dissimilation. But in /xt/ and /xm/ the sequence of fricatives arises from reduplication itself. Instead of just /xt/ > /ktxt/, the second /x/ is also replaced by /k/, arguably to satisfy base-reduplicant identity. A similar phenomenon is found in /C1C2/ > C1C2C2 and /C1C2C3/ > C1C2C3C3 structures that reduplicate the final radical. () /xs/ /xyf/ /xt/ /xm/ /sx/ /mrx/
kəskis kyəfkyif > kyəkyif kətkit kəmkim > kəkim sikik tə-mərkək
‘smash’ ‘sprinkle’ ‘crush’ ‘trim’ ‘drive a peg’ ‘kneel down’
The tableau in () shows the derivation of sikik under the ranking Ident-BR[contin], *x . . . [+contin] » Ident-IO[contin]. ()
/sx-red/ Ident-BR[contin] *x . . . [+contin] Ident-IO[contin] >sikik * sixix *! *! sikix
. P
.................................................................................................................................. In most languages a word will take on different phonological shapes depending on its location in the utterance. These include such phonetic effects as F0 declination as well as lengthening before pause. In many African languages these effects have been phonologized, giving rise to quite intricate alternations. In addition, phonological processes changing one sound in the context of another may operate between words that stand in particular
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
grammatical/prosodic relationships. For many languages it is useful to assume that the utterance is organized into hierarchical domains that mirror aspects of its surface syntactic structure. Exactly how many levels appear in the hierarchy and how they are derived from the syntax is an ongoing area of research. Here as well the languages of Africa have significantly advanced our understanding of the issues. We have space to illustrate this point with two case studies from Bantu. Hyman () shows that Kinande has two levels of sentential prosodic structure: the phonological phrase (PP) and a more inclusive intonation phrase (IP). Each level is associated with a boundary tone that appears at the right edge of the domain. But their exact realization depends on the tonal and prosodic/morphological structure of the word they attach to. The paradigm in () shows the varying realizations of the word /e-kɪ-ryatʊ/ ‘shoe’ depending on its location in the utterance. The subject is phrased separately from the verb, which is typically grouped with its complements. () citation: declarative: phrase-medial: phrase-final: interrogative: list:
è-kì-ryátʊ̀ mwátùmà è-kì-ryátʊ̀ è-kì-ryàtʊ̀ kì-rítò è-kì-ryàtʊ́ kì-kâ-wâ tùmà è-kì-ryàtʊ́ wásè mwátùmà è-kì-ryátʊ́ mótwàtùmè è-kì-ryátʊ́, nà èkíkòbá, . . .
‘shoe’ ‘He bought a shoe.’ ‘heavy shoe’ ‘Shoe is falling.’ ‘Send a shoe and come!’ ‘Did he buy a shoe?’ ‘We bought a shoe, and a rope, ...’
In phrase-medial position the underlying structure /e-kɪ-ryatʊ/ appears with default low tones [è-kì-ryàtʊ̀]; but in phrase-final position a H boundary tone is inserted on the final syllable: [è-kì-ryàtʊ́]. This H shifts to the penult when it appears at the end of a declarative sentence or in citation: [è-kì-ryátʊ̀]. The H retraction is plausibly treated by Hyman as a response to the insertion of a L marking the right edge of the IP. Since Kinande generally avoids complex tones, the phrasal H moves off the final syllable to make room for the IP L. As in many other languages the interrogative and list contexts are marked by a final H, which also induces retraction of the phrasal H: [è-kì-ryátʊ́]. A further complication is that an underlying L blocks the insertion of the phrasal H (but not the IP H): /e-kɪ-tsʊngʊ̀/ ‘potato’ is realized as [è-kì-tsʊ̀ngʊ̀] in citation and subject position but as [è-kì-tsʊ̀ngʊ́] in interrogatives and lists. Thus, while a final pitch rise in interrogatives and a final fall in declaratives are cross-linguistically very common, they interact with the tonal phonology of Kinande in subtle ways rather than constituting a simple F0 overlay. This interaction motivates treating such intonational effects in the same autosegmental terms as the lexical tones as parts of an integrated system. Three other Bantu languages have figured prominently in the phrasing literature for their alternative parses of a verb plus two complement NPs. The realization of vowel length as well as tonal alternations helps to diagnose these differences. In Mwiini (Kisseberth and Abasheikh ) an underlying long vowel is realized phonetically so long as it appears in a final three-syllable window (a). The utterances in (b) show that the window is defined at the level of the phonological phrase. Underlying /maaya/ shortens in (mayi malada) but retains its length in (maayi) (ni malada) since the subject forms a separate phrase. The data in (c) suggest that the verb phrases with its first complement (NP) (V NP) while the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
second complement forms a separate phrase (NP) (V NP) (NP). This parse allows the long vowel of Nuuru to surface. () a. x-sooma x-soomesha x-somehana
‘to read’ ‘to teach’ ‘to teach each other’
b. maayi ni malada ‘Water is fresh.’ mayi malada ‘fresh water’ c. mwaana somele chibuuku ‘The child read a book.’ mwaana somele Nuuru chibuuku ‘The child read a book to Nuuru.’ In Chewa (Kanerva ) there is no vowel length contrast; but the penultimate vowel is lengthened at the end of a phrase. In addition, a word-final H tone is retracted to the lengthened penult (a). The fact that ‘house’ and ‘child’ surface in their phrase-medial form in (b) argues that in Chewa the verb groups both complement into a single phrase phonological: (V NP NP). () a.
isolation /nyumbá/ nyuúmba /mwaná/ mwaána /njiingá/ njiínga
phrase-medial nyumbá ‘house’ mwaná ‘child’ njingá ‘bicycle’
b. (anaményá nyumbá ndí mwáála) ‘He hit the house with a rock.’ (tinapátsá mwaná njíínga) ‘We gave the child a bicycle.’ Finally, in Matuumbi (Odden ) underlying long stem vowels shorten when not located in the final word of a phrase: /mpuunga/ ‘rice’ > (mpunga wa baándu) ‘rice of the people’. Also, a H is inserted on the final syllable of utterance-medial, phrase-final words: V ! V́ / ___ ) (as in (mpuungá) (waabói) ‘the rice spoiled’. In the double object sentence in () the underlying long vowel of kikóloombe ‘shell’ surfaces phonetically. This fact suggests that kikóloombe occupies a separate phrase from Mambóondo. But it lacks the final H found on phrase-final forms. Both properties can be accounted for if a recursive structure is assigned in which the second NP is adjoined to the phrase formed by the verb and the first complement: (V NP) NP). In this parse the second complement lacks a left edge and hence fails to satisfy the requirement for H insertion. ()
naampéi kikóloombe mambóondo
‘I gave the shell to Mamboondo.’
Truckenbrodt () typologizes the phrasing differences for the V NP NP structures in these three Bantu languages with the help of a Wrap constraint that requires the head verb to join with both complements into a single phonological phrase. In the double-object constructions Wrap-XP conflicts with Align-XP-Right, which requires a phrase break at the right edge of an XP. Finally, a dispreference for recursive structures completes the typology. The different phrasings for the V–NP–NP structure follow from the rankings in (a). The comparative tableau in (b) shows the violation profile for the competing
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
candidates in the typology. Given the rankings in (a), the alternative parses will have been eliminated by higher ranking constraints. () a. Chewa: Wrap-XP, *Recursion » Align-XP-Right Mwiini: Align-XP-Right, *Recursion » Wrap-XP Matuumbi: Align-XP-Right, Wrap-XP » *Recursion b.
Align-XP-R Wrap-XP *Recursion Chewa (V NP NP) * Mwiini (V NP) (NP) * Matuumbi (V NP) NP) *
. C
.................................................................................................................................. The languages of Africa have had a significant impact on phonological theory and are a precious resource for understanding the range of typological variation. Their continued documentation and analysis should be prime objectives for students of linguistics and African studies.
R Adjekum, G., Holman, M. E., and Holman, T. W. (). Phonological Processes in Anufɔ. Legon: Institute of African Studies, University of Ghana. Archangeli, D., and Pulleyblank, D. (). Grounded Phonology. Cambridge, MA: MIT Press. Bavin Woock, E., and Noonan, M. (). ‘Vowel harmony in Lango’, Proceedings of the Chicago Linguistics Society : –. Beckman, J. (). ‘Positional faithfulness, positional neutralization, and Shona vowel harmony’, Phonology : –. Bliese, L. F. (). A Generative Grammar of Afar. Dallas, TX: Summer Institute of Linguistics and University of Texas at Arlington. Browman, C., and Goldstein, L. (). ‘Articulatory phonology: an overview’, Phonetica (–): –. Capo, H. (). A Comparative Phonology of Gbe. New York, NY: Foris. Casali, R. (). Resolving Hiatus. Ph.D. thesis. Los Angeles, CA: University of California. Casali, R. (). ‘[ATR] value asymmetries and underlying vowel inventory in Niger-Congo and Nilo-Saharan’, Linguistic Typology : –. Clements, G. N. (). ‘Vowel harmony in Ewe’, Studies in African Linguistics : –. Clements, G. N. (). Vowel Harmony in Nonlinear Generative Phonology. Bloomington, IN: Indiana University Linguistics Club. Clements, G. N. (). ‘Compensatory lengthening and consonant gemination in LuGanda’, in L. Wetzels and E. Sezer (eds.), Studies in Compensatory Lengthening. Dordrecht: Foris, –. Clements, G. N. (). ‘Feature economy in sound systems’, Phonology (): –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Dell, F., and Elmedlaoui, M. (). ‘Syllabic consonants and syllabification in Imdlawn Tashlhiyt Berber’, Journal of African Languages and Linguistics : –. Downing, L. (). ‘Morphological and prosodic constraints on Kinande verbal reduplication’, Phonology : –. Frampton, J. (). ‘GDE syllabification—a generalization of Dell and Elmedlaoui’s syllabification algorithm’, The Linguistic Review : –. Gick, B., Pulleyblank, D., Campbell, F., and Mutaka, N. (). ‘Low vowels and transparency in Kinande vowel harmony’, Phonology : –. Goldsmith, J. (). Autosegmental Phonology. Ph.D. thesis. Cambridge, MA: Massachusetts Institute of Technology. Guerssel, M. (). ‘Constraints on phonological rules’, Linguistic Analysis : –. Guerssel, M. (). ‘A condition on assimilation rules’, Linguistic Analysis : –. Halpert, C. (). ‘Overlap-driven consequences of nasal place assimilation’, in P. Hoole, L. Bombien, M. Pouplier, C. Mooshammer, and B. Kühnert (eds.), Consonant Clusters and Structural Complexity. Berlin: Mouton de Gruyter, –. Hoffmann, C. (). A Grammar of the Margi Language. London: Oxford University Press. Hyman, L. (). ‘Boundary phonology and the prosodic hierarchy’, in S. Inkelas and D. Zec (eds.), –. Hyman, L. (). Is there a right-to-left bias in vowel harmony? Paper presented at the Ninth International Phonology Meeting, Vienna. Inkelas, S., and Zec, D. (eds.) (). The Phonology–Syntax Connection. Chicago, IL: University of Chicago Press Jones, P. (). ‘New evidence for a phonological stem domain in Kinande’. Proceedings of the West Coast Conference on Formal Linguistics . Somerville, MA: Cascadilla Proceedings Project, –. Kanerva, J. (). ‘Focus on phonological phrases in Chichewa’, in S. Inkelas and D. Zec (eds.), –. Kenstowicz, M. (). ‘Gemination and spirantization in Tigrinya’, Studies in the Linguistic Sciences (): –. Kenstowicz, M. (). ‘Two notes on Kinande vowel harmony’, Language Sciences : –. Kenstowicz, M., and Petros Banksira, D. (). ‘Reduplicative identity in Chaha’, Linguistic Inquiry : –. Kenstowicz, M., Nikiema, E., and Ourso, M. (). ‘Tonal polarity in two Gur languages’, Studies in the Linguistic Sciences (): –. Kisseberth, C., and Abasheikh, M. I. (). ‘Vowel length in Chi-Mwi:ni: a case study of the role of grammar in phonology’, in A. Bruck, R. Allen Fox, and M. W La Galy (eds.), Parasession on Natural Phonology. Chicago, IL: Chicago Linguistics Society, –. Ladefoged, P. (). A Phonetic Study of West African Languages: An auditory-instrumental study. Cambridge: Cambridge University Press. Leben, W. (). Suprasegmental Phonology. Ph.D. thesis. Cambridge, MA: Massachussetts Institute of Technology. Lehiste, I. (). Suprasegmentals. Cambridge, MA: MIT Press. Lightner, T. (). ‘On the description of vowel and consonant harmony’, Word : –. McCarthy, J. (). Formal Problems in Semitic Phonology and Morphology. Ph.D. thesis. Cambridge, MA: Massachusetts Institute of Technology. McCarthy, J. (). ‘Consonantal morphology in the Chaha verb’, in M. Cobler, S. MacKaye, and M. T. Wescoat (eds.), Proceedings of the West Coast Conference on Formal Linguistics . Stanford, CA: Stanford Linguistics Association, –. McCarthy, J. (). ‘OCP effects: gemination and antigemination’, Linguistic Inquiry : –. McCarthy, J., and Prince, A. (). ‘The emergence of the unmarked: optimality in prosodic morphology’, Proceedings of the Northeastern Linguistics Society : –. Mutaka, N. (). ‘Vowel harmony in Kinande’, Journal of West African Languages (): –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Mutaka, N., and Hyman, L. (). ‘Syllables and morpheme integrity in Kinande reduplication’, Phonology : –. Myers, S., and Carleton, T. (). ‘Tonal transfer in Chichewa’, Phonology : –. Nevins, A. (). Locality in Vowel Harmony. Cambridge, MA: MIT Press. Odden, D. (). ‘Stem tone assignment in Shona’, in G. N. Clements and J. Goldsmith (eds.), Autosegmental Studies in Bantu Tone. Dordrecht: Foris, –. Odden, D. (). The Phonology and Morphology of Kimatuumbi. Oxford: Clarendon Press. Odden, D. (). Introducing Phonology. Cambridge: Cambridge University Press. Odden, D., and Odden, M. (). ‘Ordered reduplication in Kihehe’, Linguistic Inquiry : –. Orie, O. O. (). ‘An alignment-based account of vowel harmony in Ife Yoruba’, Journal of African Languages and Linguistics : –. Pater, J. (). ‘Serial Harmonic Grammar and Berber syllabification’, in T. Borowsky, S. Kawahara, T. Shinya, and M. Sugahara (eds.), Prosody Matters: Essays in honor of Elisabeth O. Selkirk. London: Equinox Press, –. Potts, C., Pater, J., Jesney, K., Bhatt, R., and Becker, M. (). ‘Harmonic grammar with linear programming: from linear systems to linguistic typology’, Phonology : –. Prince, A., and Smolensky, P. ( []). Optimality Theory: Constraint interaction in generative grammar. Oxford: Basil Blackwell. Pulleyblank, D. (). ‘Extratonality and polarity’, in M. Cobler, S. MacKaye, and M. T. Wescoat (eds.), Proceedings of the West Coast Conference on Formal Linguistics . Stanford, CA: Stanford Linguistics Association, –. Schaefer, R. (). An Initial Orthography and Lexicon for Emai. Bloomington, IN: Indiana University Linguistics Club. Smolensky, P. (). ‘Optimality in phonology II: harmonic completeness, local constraint conjunction, and feature domain markedness’, in P. Smolensky and G. Legendre (eds.), The Harmonic Mind: From neural computation to Optimality-Theoretic grammar. Vol. . Cambridge, MA: MIT Press, –. Traill, A. (). Phonetic and Phonological Studies of !Xóõ Bushman. Hamburg: Helmut Buske. Truckenbrodt, H. (). ‘On the relation between syntactic phrases and phonological phrases’, Linguistic Inquiry : –. Valinande, N. K. (). The Structure of Kinande. Ph.D. thesis. Washington, DC: Georgetown University. Williams, E. (). ‘Underlying tone in Margi and Igbo’, Linguistic Inquiry : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. A tone systems have played a major role in phonological theory thanks to research on autosegmental phonology by Africanists such as Clements, Kisseberth, Hyman, Goldsmith, and Leben. Spread and shift of tone, floating tones and tone melodies, and downstep are welldocumented properties of African tone, manifesting the fact that tones owe little allegiance to the segments realizing them. See Leben (), Goldsmith (), Clements and Ford (), and Odden () for general overviews of African tone and autosemental phonology. Controversies arise over whether certain languages are tonal, for example Somali (Hyman ), and some tonal systems are referred to a typological midpoint in a continuum between tone (e.g. Ewe) and stress (e.g. Swahili), termed pitch-accent. The primary objection to the idea that languages like Somali are tonal is that they differ from Chinese, which has become the tonal archetype in the literature. In so far as the identifying property of so-called pitch-accent languages is that tonal structure is simple and contrast is limited, this overview considers languages like Somali, Ganda, and Kinga to be proper tone languages, albeit with more restrictions on tone than found in Kru languages. See Pulleyblank (), Odden (), Hyman (), Downing () for a discussion of “accent”.
. L :
.................................................................................................................................. The most basic question about tone is: what is tone? By assumption, it is a featural object, like [i]. Analogous to the question “what are the possible vowels”, one can ask about the inventory of tones.
.. Levels Tone is primarily the contrastive use of pitch in grammar and lexicon, including movement from level to level. The first question is: what are the fundamental pitch levels? The simplest
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
systems have a two-way contrast between higher and lower pitch, H and L,1 and the majority of African languages fall into this category. Gur, Atlantic, Mande, Dogon, Nilo-Saharan, Chadic, and Cushitic languages usually have two levels; Bantu languages are almost universally two-level. Three-level languages are not rare, but they are a distinct minority. Examples of three-level languages are Ngas, Peki Ewe, Ebira, Ga’anda, Kasem, Kotoko, Kpelle, Logo, Mbay, and Yoruba. Languages with four tone levels are much less common, and include Bariba, Anlo Ewe, Grebo, Igede, Kamba, and Wobe. Five levels are quite rare, occurring in Benchnon and the Santa dialect of Dan; and only Chori is reported to have six. Terminology and notations for tone vary. In two-tone systems, tones are usually termed H and L, transcribed v́, v̀, though also v́, v or rarely v, v̀. Three-level systems are usually analyzed via the terms H, Mid (M), and L notated variously as v́ v v̀, v́ v̄ v̀ or v́ v̍ v̀. Terminology and notation for four-level systems is quite variable, ranging over terms such as High, Mid, Low, Superhigh (SH), Superlow (SL), Top, Raised, Upper Mid, Mid-H, Mid-L. Terminology and transcription are normalized as follows: ()
-level -level -level /-level
H = v́ L = v̀ H = v́ M = v̄ L = v̀ SH = v̋ H = v́ L = v̀ SL = v̏ integers
A further complication in the notion of tone level is the phenomenon of downstep and upstep, discussed in section .. Integer notation is used when the point is specific pitch, given downstep.
.. Contours In some languages (Shona, Pare, Mbololo Taita, Miya), syllables are either H or L, without phonological rising or falling tones, which involve F movement from level to level. Kotoko has the three-tone system H M L, but no contour tones. Many languages have phonological contour tones. Some allow contour tones only on long syllables, for example Hausa and many Bantu languages (Tachoni, Dembwa Taita) have just falling tone (F) and only on long syllables. Kaficho has one lexical contour (Addo ; Theil ), rising (R) restricted to long syllables,2 and Didinga, Khoekhoegowab, and Matumbi3 have rising and falling tone, again only on long vowels. Nara (Dawd and Hayward ), which has only H and L on monomoraic syllables, has rising and falling contours on long syllables and tritonal R–F on trimoraic (XVVC) syllables.4 In these languages, short syllables can only have level tone. Languages like Tachoni, which limit tones to one per mora, are “contourless” in the
1 There is a further distinction within two-level languages between equipollent—H vs. L—and privative—H vs. Ø or rarely L vs. Ø—analyses. Mande languages are typically analyzed as opposing H and L, whereas most Bantu languages are treated as contrasting H and toneless. 2 Falling tones are also allowed on heteromorphemic vowel sequences. 3 Unusually, long level-H vowels are not allowed. 4 Interestingly, there is no locational contrast within the syllable: tones map one-to-one left-to-right, allowing trimoraic LHH, HLL but not LLH, HHL.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
moraic sense, but from the syllabic perspective they allow contours. In contrast, Miya has long vowels but no R or F, and Tachoni still has only one of the expected contours on long vowels. Many languages have contours on short vowels; thus Gen and Temne have H, L, R, and F tones; Ngas has three tone levels and the four rising and falling contours which do not end with M tone; Benchnon (Wedekind ) has five levels but only one contour, a – rising tone. These languages lack long syllables. The Santa dialect of Dan (Bearth and Zemp ; Flik ), which has five levels and contrastive length, allows one short contour (– fall) but five long contours (rises –, – and falls –, –, –), far fewer than the twenty possible contours. On short vowels, Guéré (Paradis ), which has four tone levels only, allows L–SL and H–SL falling tones and SL–L, SL–SH, and L–H rising tones, but has no significant restrictions on tone on long vowels and diphthongs.5 Closely related Wobe (Bearth and Link ), which has long syllables and four level tones, has at least three rising tones and four falling tones on short vowels, but no obvious restrictions on sequences of tones on long syllables. A few languages have tritonal complexes—R–F (Nyang, Mende, Kono, Wobe, Mongo, Nara) and F–R (Mongo). Kenyang allows rising and falling tones on long and short vowels, but has R–F only on long vowels, and as noted above in section .., Nara allows R–F only on trimoraic syllables. Inventory restrictions on contours may thus separately limit the sequence of tones in a syllable, as well as the number of tones per mora. It has long been recognized that contour tones analytically decompose into sequences of levels, where F is a sequence HL, R is a sequence LH, and so on. This many-to-one mapping is widely discussed in the autosegmental literature, which draws support from the ease with which it formally identifies this unity. ()
H
L
bá b à
=
HL bâ
L
H
bà bá
=
L H b aˇ
Two facts supporting this treatment of contours are melodic parallelism and tone preservation. Melodic parallelism is where words with different numbers of syllables have the same abstract tone pattern, realized differently due to how tones compress onto single vowels in shorter words. In Bobo (Morse ), plural and singular nouns have parallel tone patterns, so contours are found in the singular, which has one tone-bearing unit (TBU), but in the plural each vowel has a distinct level tone. ()
5
lu᷅ zɔ᷅ yɔ᷆ sɔ̃̌ pε᷅
lìā zɔ̀ε ̄ yɔ̄rè sàmá̃ pàgā
‘cold’ ‘fishnet’ ‘soul’ ‘person’ ‘forked stick’
Paradis treats all long vowels and long diphthongs as disyllabic.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Tone preservation refers to the situation where vowels merge into one which preserves all of the tones, creating contour tones, illustrated in () with the well-known case of Lomongo. ()
H+L ! F L+H ! R H+R ! FR L+F ! RF
mpùlú ìnέ là bɔ́nà bàlóngá bǎkáé fàkàlà ɔ̂tswà
! ! ! !
mpùjwînέ lɔ̌nà bàlónga᷉káé fàkàlɔ᷈tswà
‘these birds’ ‘with the baby’ ‘his blood’ ‘Fakala comes in’
. T
.................................................................................................................................. The evidence for decomposing contour tones into sequences of level specifications on a single unit is so strong that the decomposition is typically taken for granted, though see Bearth and Link (), Newman (). The fundamental question of how to analyze tones is: what are the appropriate features for describing tone levels? A widely adopted theory is that proposed by Yip (), terminologically modified by Pulleyblank (), which divides tone space into upper and lower registers, subdividing registers by distinguishing raised vs. non-raised tones. Accordingly, tones are assigned feature values as follows. ()
SH H L SL upper + + raised + +
The only conclusive evidence for feature assignment is whether it describes tonal groupings: any theory can identify “all tones” or “a specific tone”. A test of a theory requires rules identifying a proper subset of tones, so tone features can only be tested in a language with at least three levels. The presumption is that if tones {A,B} are targeted to the exclusion of {C,D} (e.g. {A,B} trigger some change which {C,D} do not), then {A,B} have in common some feature which {C,D} lack. Convincing evidence most likely will come from a language with four levels. Weaker evidence is partial-assimilation evidence, where A assimilatorily changes to B in the context of C: A takes on some properties of C, but not all properties. An assimilatory argument for () comes from Anlo Ewe (Clements ), which raises L to SH when surrounded by H. Other processes spread SH to H syllables, and lower phrase-final SH. ()
àkplɔ̀ mègbé ! àkplɔ̀ mègbé ‘behind a spear’ èkpé mègbé ! èkpe̋ me̋gbé ‘behind a stone’
Strikingly, L goes “in the direction” of the surrounding H, but overshoots that target. The process is described formally by (), which assigns the upper-register feature borne by H to L (which is the [+raised] tone of the lower register), making it the raised tone Ø in the upper register, i.e. SH.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
H
L x −upper
+upper −raised
H +upper +raised
−raised
Further support for the model is that it allows a coherent characterization of a morphosyntactically conditioned tone-raising in Igede (Bergman ), where suffixed verbs raise SL to L and H to SH. The underlying tones are [raised], contrasting in the register feature [upper], and become [+raised], effecting a one-step shift from SL to L and H to SH. Three-level languages can easily be described: the problem is the ambiguity of possible specifications for three tones. There are four specification patterns, where two of the three tones act together, excluding the third.
upper
a. H M + +
L –
raised
+
+
()
–
b. H M L + + – +
– –
c. H M L + – – + + –
d. H M L + – – –
+ –
Support for (d) exists in Sara Mbay, where H becomes L when preceded by H and followed by M (Keegan and p.c.), i.e. [raised] H takes on the [upper] register of M, becoming L. ()
síndá-gə̄ ! síndàgə̄ m-ɔ́də́-ī ! mɔ́dìī
‘horses’ ‘I touch you’
súú-gə̄ ! súù-gə̄
‘spiders’
(single H in m̄ -ndá-ī ‘I hit you’ unaffected)
It is not self-evident that the process is assimilation, so under analysis (a) this can be treated as dissimilation of [+upper] in H to [upper] L before [+upper, raised] M. Under analysis (c) the process could be dissimilation of [+raised], which H and M have in common—this creates otherwise non-existent [+upper, raised] tone, which can be corrected with a structure-preserving readjustment to [upper]. Finally, it is difficult to rule out on empirically justified theoretical grounds a statement of this process as a formally arbitrary feature change, e.g. under analysis (b) [+upper, +raised] ! [upper, raised] / [+upper, +raised] __ [+upper, raised]. Since the elements involved in the process are single tones, the question of what unites tones does not arise. An example involving two triggering segments is assimilatory raising in Mbui (Hyman and Schuh ) whereby L becomes M before M and H.6 () nì-bʉ́ʉ́ ! nī-bʉ́ʉ́ ‘breast’ nì-sɔ̄ŋ ! nī-sɔ̄ŋ ‘tooth’ nì-kùò ‘four’ This can be expressed as spread of [+upper] in analysis (b) and spread of [+raised] in analysis (c), but recall that analysis (b) entailed an arbitrary feature change for Mbay. Mupun (Frajzyngier ) has a conceptually similar assimilatory lowering rule whereby H becomes M after L and M. 6
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Analysis (d)—the one which enabled an assimilatory account of Mbay—is the one analysis where H and M have no features in common. The only analysis possible under analysis (a) is an arbitrary feature change. In comparing these languages, we see that Mbay tones are best analyzed under feature assignment (d), but Mbui is best analyzed under assignments (b) and (c)—meaning that there is no universal mapping between inventory and analysis. Rather, the featural analysis must be discovered by seeing how tonemes group under phonological rules. More problematic is the tonology of the Ali associative construction (Bradshaw ), whereby N raises its tone by one level, so that L becomes M and M becomes H, triggered presumably by a floating tone morpheme. ()
zù + + yérè ! zū yérè gbà̃là̃ + + yà ! gbā̃lā̃ yà nū + + kpánà ! nú kpánà sālā + + tὲ ! sálá tὲ
‘buffalo head’ ‘his bone’ ‘jar mouth’ ‘body hair’
No analysis using () expresses the notion “raises one step”. Some four-level languages superficially have a worst-case problem in that three out of four tones act as a class. In Bariba, L becomes SL after SL when followed by SH, H, or L—no feature can identify the set SH, H, L. But the reason why SL does not participate in the triggering class is that SL independently becomes SH when surrounded by SL tones, thus SL–L # SL simply does not exist. In Boko, verbal SH becomes H when surrounded by lower tones (H, L, SL) (Jones ). ()
ȁlε̋ n̩̋ pa̋pȁ ȁlε̋ wá pápȁ f ɔ̀ kpa̋ ma̋! ʔa̋ kpa̋ wȁ! f ɔ̀ kpá wȁ! zù kpá wȁ! má zɔ̀lε̋ mí zɔ̀lέ ò ma̋ bȉı ̋ ma̋ bȉí zã̏
‘He is stoning them’ ‘He is stoning monkey’ ‘Greet them!’ ‘Give him a yam!’ ‘Greet him!’ ‘Give him a cow!’ ‘I sat down’ ‘I didn’t sit down’ ‘I will shine it’ ‘I will shine it far’
Similarly, Fe’Fe’ Bamileke SL becomes L before any higher tone (SH, H, L)—SL–SL does not change. In these cases, the excluded tone is identical to the target tone, and the generalization can be restated as applying unless before or adjacent to an identical tone. Identification of an identical tone can be reduced to a representational distinction in autosegmental theory, by presuming that adjacent identical tones fuse into one multiply associated tone (examples from Boko).7 7
An alternative approach is the constrained algebraic theory of Reiss () where rules may explicitly state featural identity relations or their negation, thus in Boko lowering SH if surrounded by non-identical tones, and in Bamileke SL raises before a non-identical tone. Evidence for “higher” or “lower” tones distinct from reference to “non-identical” tone would in principle be a case where a rule
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
SH
()
SL
ʔ a kp a w a
L
SH
f ɔ kp a ma
L fɔ
SH → H SL kp a
wa
Given such a representation, SH-lowering can be stated simply as lowering a singly linked SH in phrase-medial position, a condition not met in the case of SH before SH. Thus, such data do not necessarily pose a problem for the concept of featurally defined natural classes. Kamba provides evidence for grouping SH and SL together, which is impossible under the standard theory, and supports a feature [extreme] characterizing highest and lowest tones. ()
SH H L SL H + + – – extreme + – – +
Evidence is a rule raising H to SH before SL (see Roberts-Kohno for additional details of Kamba tone). Infinitives have a final SL, which spreads to the second mora of a long penult. Verbs are also lexically differentiated as to whether their first root mora has an underlying H versus L. ()
L verbs
kò-kòn-ȁ kò-kὲε̏l-ȁ H verbs kò-kélàkèly-ȁ kò-kétèk-ȁ
‘hit’ ‘strain’ ‘tickle’ ‘occur’
kò-kɔ̀lɔ̀k-ȁ kò-sìtàȁk-ȁ kò-kóòlòk-ȁ kò-tálàȁŋg-ȁ
‘stir’ ‘accuse’ ‘advance’ ‘count randomly’
Whenever H comes before SL, that H becomes SH, which spreads to a final SL vowel. () kò-ta̋l-a̋ ‘count’ kò-kőȍly-ȁ ‘ask’ kò-tẅ-a̋ ‘pluck’ /kòtálȁ/ /kòkóȍlyȁ/ /kò-tẅá ̏/ Raising H to SH before SL is easily comprehensible as () if the language employs a feature [extreme] grouping SH and SL together. ()
H V
V [+extreme]
The previously mentioned rule of Bariba whereby SL becomes SH when surrounded by SL is now easily understood as dissimilation of L to H, retaining underlying [+extreme]. refers to SL followed by L, H, SH but also L followed by H, SH, that is, a probative rule would have to involve more than one input tone. A rule referring to L followed by H, SH is insufficient, since the class H, SH is easily referred to as the [+upper] class.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
These examples illustrate problems that linguistic theory has faced in reaching firm conclusions regarding featural analysis of tones: evidence can only support a model given well-justified primary theoretical principles, and to the extent that such principles are controversial, the potency of such data for adjudicating the correct set of tonal features is lessened. The presumption that there even exists a single universally applicable set of features is a venerable, yet controversial assumption (Morén ; Blaho ; Odden ). An obvious problem for featural analysis is that two binary features only allow a fourway contrast in levels; yet five-way contrasts do exist. Benchnon presents a six-way minimal pair involving five levels and one (rising) contour. ()
kar5 ‘clear’ kar4 ‘broad leaf ’ kar3 ‘to be round’ kar2 ‘wasp’ kar1 ‘pudenda’ kar23 ‘game played with stones’
This language along with Santa Dan lacks phonological processes elucidating the analysis of contrasts, only showing the need to add some feature. The most challenging language is Chori (Dihoff ), with six tone levels. According to Dihoff, only tones , , and are required in underlying contrasts of major word classes (nouns and verbs). Nevertheless, all six levels play a role in the grammar. Levels , , and generally derive by rule, or are limited to certain grammatical constructions. Levels and can derive from compression of two tones onto a single syllable in connected speech (+ ! , + ! ). ()
kel2 gha2 di1 [ŋa1 a6] ~ kel2 gha2 di1 [ŋa3] ‘His wife gave it to him’ ŋa1 gos1 sub2 [mba2 a6] ~ŋa1 gos1 sub1 [mba5]
‘He bought the mushroom (non-)’
Level may derive from a partial tone assimilation rule raising to before . () mi1 bzar1 [nap1] ke6 a2 ‘I touched this branch’ mi1 bzar1 [fɛb4]e6 a2 ‘I touched this cockroach’ mi1 bzar1 [nyagh6]e2 a1 ‘I touched this cow’
[nap1] ke1 ‘This is a branch’ [fɛb2] e1 ‘This is a cockroach’ [nyagh6] e1 ‘This is a cow’
Levels , , and also play a role in the system of grammatical tone. The underlying /, , / contrast in verbs is subject to inflectionally governed modifications. In the present tense, // becomes [] after a SG subject, and /, / become [] after a SG subject. ()
subject na6 a2 fal6 na6 a2 tsep4 na6 a2 dur6
subject mi6 i2 fsal1 mi6 i2 tsep3 mi6 i2 jur3
gloss ‘cooking’ ‘writing’ ‘catching’
Similarly, in the past completive object focus form // becomes [] and /, / become [], regardless of subject. The tone of the subject pronoun varies according to tense, so the SG
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
pronoun /mi/ in the past habitual has tone . Analysis of Chori in light of contemporary theory would be enlightening, and important for the theory of tone.
. D
.................................................................................................................................. Downdrift and downstep complicate the study of tonal inventories. Downdrift refers to recursive automatic lowering of the pitch of H separated from the previous H by L. In Karanga Shona [hàvázàkázòndíbìkírà] ‘they haven’t ever cooked for me’, after the first H, each H is pronounced at a lower level than the previous H, rendered numerically as [hà3vá1zà4ká2zò5ndí3bì6kí4rà7]—the level of L tones also decreases, perhaps at a lesser rate. This global process of tone-register lowering is phonemically transparent, so lowered H appears if and only if it is preceded by a L-toned vowel. In principle, one might be led analytically astray by the fact that the language presents very many systematically used pitch levels and wrongly conclude that the language has a dozen or more tone levels, but ordinary phonemicization methods quickly yield a simpler system with just two distinctive levels and a recursive phonetic interpretation of those two levels. It is an interesting question how a theory of phonetics accounts for this process (see Schachter and Fromkin ; Fromkin ; Clifton for early proposals; Clements for an autosegmentalmetrical theory), but since there is a bidirectional mapping between phonological representation with just H and L and their continuous physical realization, downdrift need not concern phonological theory. Some languages have this effect without overt L, that is, global lowering may be contrastive. Often these facts are analyzed—incorrectly, it turns out—in terms of an additional mid level, for example in Efik and Igbo (Green ) or Twi (Christaller ). Efik has been analyzed as having three tone levels based on the minimal triple [ɔ́bɔŋ́ ] ‘mosquito’, [ɔ́bɔŋ̄ ] ‘chief ’, [ɔ́bɔŋ̀ ] ‘piece of cane’. Such data give prima facie evidence for a three-level system. As observed by Winston (), the properties of supposed M are very peculiar. It only exists distinctively after H and M: it does not exist unless there is at least one preceding H, and is an allotone of H after L. It conditions phonetic variants of subsequent H in a complex fashion. M of the syllable [nyoŋ] in (b) causes the following Hs in [édí ú] to have a special lowered variant—the pitch of a string of H syllables after M becomes identical to the pitch of M-toned [nyōŋ]. Likewise, M of [dī] in (c) conditions the lower-variant pitch on following [ú]. (d) shows that M itself lowers when preceded by M, and lowering is cumulative, so M of [nyoŋ] in (d) causes all subsequent tones to lower by one step, and that of [dī] causes further lowering. () a. b. c. d.
èkéŋ1 é1dí1 ú1fɔ̀k èkpé1nyōŋ2 é2dí2 ú2fɔ̀k èkéŋ1 é1dī2 ú2fɔ̀k èkpé1nyōŋ2 é2dī3 ú3fɔ̀k
‘It was Ekeng came to the house’ ‘It was Ekpenyong came to the house’ ‘Ekeng came to the house’ ‘Ekpenyong came to the house’
Explaining Efik by reference to mid tones necessitates proliferating tone variants—M, M, M, H, H, H, etc. The distribution of these tones is restricted so H never appears immediately after any M; H only appears after M or H; H only appears after M or H.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A simple generalization is being missed: M and H define a pitch ceiling which subsequent tones cannot exceed, and M is distinct from H only in lowering that pitch ceiling. The key is recognizing the identity of this ceiling-defining and lowering property of “M”, and the lowering effect of downdrift: dowstep is simply contrastive downdrift. Winston combines the phonemic tone levels H and M, introduces a register-lowering operator “downstep”, notated !, and significantly simplifies the description of Efik tone. The realization of H and L tones is computed relative to a local register value which decreases whenever downstep occurs. Accordingly, (d) is phonologically [èkpé!nyóŋ é!dí úfɔ̀k]. It follows that H cannot be higher than M because M is H, simply one that occurs right after downstep. Downstep is usually restricted to two-level languages, but a few three-level languages have downstep. Yala Ikom (Armstrong ) has three levels plus downstep affecting H and M; thus, H!H and M!M are found. In Ga’anda (Ma Newman ), both floating M and L cause downdrift of H, though M tone itself is not downstepped (M!M appears to not exist). Languages with more than two tone levels are important for understanding the nature of downstep, since they raise the question of possible phonetic equivalence of H!H and HM. In Ga’anda and Yala Ikom, it is reported that the second tone in H!H is higher than in HM; thus, !H is phonetically distinct from M. The degree of pitch drop characterized by downstep is subject to language-specific variation, and a small number of languages have total downstep, where the sequence TiTj is identical to Ti!Ti (Ti being the tone which is one step higher than Tj). Kikuyu (Clements and Ford ) and Päri (Andersen ) are two-level languages with total downstep: H!H is phonetically identical to HL, the difference between the sequences being that when placed before L, pitch drops again after H!H, and cannot rise above the level of !H. In contrast, pitch does not drop when HL is placed before L, and pitch does rise to the level of H when H is placed after HL. Babanki, which has three surface levels, is reported to have total downstep, so that the drop in H!H is said to be phonetically identical to that in HM— however, M in Babanki has a very restricted distribution and is only followed by H (Hyman ), so the phonetic equivalence of H!H and HM cannot be unambiguously tested. The four-level language Bwamu (Riccitelli ) has total downstep which can be both contrastive8 and also automatic—SH is subject to automatic downstepping after H and L (not SL), but the level of all tones decreases after downstep. ()
ȍ4 mòò3 ! da̋2si 2̋ wa̋2 ! ba̋3ga̋3wa̋3 nà5 ! ma̋a4̋ sὲὲ6 ! ma̋a̋5 yí6 ! bi̋n6 tȁȁ?9 ‘He saw the soldiers’ shirts which cannot be washed there, didn’t he?’
Typically, downstep only exists between H tones, so LH will not contrast with L!H, nor would L!L contrast with LL. Utterance-initial H does not usually contrast with !H. Unusual languages are attested, however. Dschang Bamileke contrasts LL vs. L!L, LH vs. L!H, as well as utterance-initial H vs. !H; finally, it contrasts single- and double-downstep (Hyman and Tadadjeu ).
8 Riccitelli briefly mentions a number of contexts where non-transparent downstep occurs, for example H downsteps after long level H; under certain circumstances, H assimilates SH from the plural suffix -wa̋ which causes downstepping before another assimilated-H noun.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() tɔ́ŋ ! kɔ́ŋ à kè tɔ́ŋɔ́ sə́ŋ à kè tɔŋ́ ɔ́ !mɔ́ à kè tɔŋ́ ɔ́ !səŋ́ à kè tɔŋ́ ɔ́ !!mɔ́
‘call yesterday conditional’ ‘like yesterday conditional’ ‘if he called a bird’ ‘if he called a child’ ‘He called a bird’ ‘He called a child’
Another language reported to have double-downstep is Kumam (Hieda ). The opposite process, where tone register is raised, upstep (notated ¡), is reported in a few languages. Recursive phonemic upstep is only reported in one Mesoamerican language, Acatlan Mixtec. As noted by Hyman (), upstep in Engenni, Mankon, and Cahi Rimi have the peculiar property that upstep cannot follow upstep without an intervening resetting of pitch register. Engenni has automatic upstep of H before L, resulting in the alternation [mì ̣ mọ́ní wọ́] ‘I saw you’, [mì ̣ mọ́ní ¡wọ́ bhẹ̀è] ‘I did see you’, and it also has phonemic upstep when a L-toned vowel elides between H syllables, as in /ó dírè ẹ́dà/ ! [ó ¡dír ẹ́dà] ‘she will cook beans’. Pre-L [ẹ́] is not automatically raised—it already is raised because of the preceding raised syllable. A crucial property of downstep distinguishing it from an additional tone level is recursivity, i.e. successive unbounded pitch lowering. African languages with assumed phonemic upstep all appear to lack that crucial diagnostic of global register shift, and it is possible to analyze upstep in terms of a derived SH plus, in languages like Engenni, progressive assimilation of H to SH after SH. It is controversial how downstep should be formally represented. The traditional view connects downdrift and downstep, positing that downstep is the surface result of floating L between Hs; thus, phonologically speaking, downstep is a floating L.9 This is well motivated in many cases, e.g. Ngizim, Twi, Bambara, Ga, Tiriki. A minor problem with the floating L analysis is that in some cases, downstep exists without downdrift (Kikuyu, Dschang, Bwamu), where the phonetic process must identify only floating tones as triggering register lowering. The second problem is that in some languages, downstep automatically emerges when Hs are concatenated. Shambaa illustrates this: any time a word ending with H is placed before a word beginning with H, downstep always appears. () ŋgó!tó ‘sheep’ dú ‘only’ ízafá ‘They died’ ŋgó!tó !dú !ízafá ‘Only sheep died’ This leads to a competing analysis of downstep, where H!H is represented as two separate H tones but HH is represented as a single multiply linked H.
9
Unusually, downstep in Päri is triggered by floating H.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
H ŋwá ná ‘child’
H
H
ngó tó ‘sheep’
See Odden (), Bickmore () for such analyses, and Paster and Kim () for defense of the floating L analysis. The main weakness of the H-concatenation theory is that such analyses often require ad hoc rules of tone fusion combining separate H autosegments—the situation in Shambaa and Logoori where concatenation of H robustly yields downstep turns out to be unusual, and more languages are like Namwanga, which requires multiple fusion rules. The H-concatenation theory cannot be a universal representation of register lowering, given languages like Dschang where downstep is contrastive in contexts such as L(!)H and where double-downstep contrasts with singledownstep. It is possible that both models are correct, so some languages have automatic register lowering when identical tonal autosegments are concatenated, and others have a floating tone entity. While many languages present evidence that downstep is an ordered and movable entity, no language seems to present direct evidence that downstep and linked L act together in triggering phonological processes.
. W
.................................................................................................................................. As a featural object, tones do what other featural objects do—they change by assimilation, dissimilation, and contextually neutralize. Tone is also special in certain ways. Tone spread commonly comes in two varieties, iterative vs. local, whereas assimilatory segmental harmony rules apply as long as their structural descriptions are satisfied. Tones can also move from place to place, and can define multi-element grammatical “melodies” imposed on a string.
.. Local versus long-distance assimilation and shift A common tonal process is unbounded tone spread, illustrated in () with Shambaa data, where H spreads rightwards through a word to (excluding) the final vowel.10 () kùghòshòà ní kúghóshóà kùghòshòàghòshòà kùchíghóshóághóshóà
‘to do’ ‘it is to do’ ‘to do repeatedly’ ‘to do it repeatedly’
This process is formally parallel to well-known vowel harmony. Unlike vowel harmony, tone spreading is often restricted to spreading just once, a process termed “tone doubling”. The Korovere dialect of Makua (Cheng and Kisseberth ) illustrates this. In the negative 10
The source of the H tone is underlined in the examples below.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
perfect, H is assigned to the second stem vowel (underlined), and spreads to the following non-final vowel. () àkì-thùkálè àkì-lòkóthálè àkì-lòkótánìnhè àkì-màálíhàlè
‘I have not tied’ ‘I have not picked up’ ‘I have not picked up (plural)’ ‘I have not silenced’
A related process which is unusual from the perspective of other features but not unusual in tone is tone shift, a process whereby a tone moves from one underlying or phonologically derived position to another. An example of “local” shift is found in Jita (Downing ), where H shifts to the following non-prepausal syllable. The underlying H of /bón/ remains in situ when its syllable is the penult. When (L-toned) suffixes are added, H moves one syllable rightward. () òkùbónà òkùbònírà òkùbònánà òkùbònírànà
‘to hit’ ‘to hit for’ ‘to hit each other’ ‘to hit for each other’
An unbounded shift to the penult is found in Zigua (Kenstowicz and Kisseberth ), illustrated with the H-toned verb root /ón/ ‘see’, whose tone shifts to the penult of the following noun. () màlàpùlà:pù ‘scrubbing instrument’ nàònà màlàpùlá:pù ‘I see a scrubbing instrument’ Segmental features such as rounding, vowel height, and nasality do not shift, indicating that tone is special.
.. Melodies The concept of tone melody which can be abstracted away from the segments realizing the tones was an essential argument for autosegmental theory. The classical melody is that of Mende, which is said to restrict tone patterns to just five types, whose realization (especially the distribution of contours) depends on stem length.11 ()
11
H L HL LH LHL
háwámá kpàkàlì félàmà ndàvúlá nìkílì
‘waistline’ ‘tripod chair’ ‘junction’ ‘sling’ ‘groundnut’
pέlέ bὲlὲ ngílà nàvó nyàhâ
‘house’ ‘trousers’ ‘dog’ ‘money’ ‘woman’
kɔ ́ kpà mbû mbǎ mbǎˋ
‘war’ ‘debt’ ‘owl’ ‘rice’ ‘companion’
Mende words do not actually exhibit only these five patterns (see Dwyer ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The unity of the LHL pattern is graphically evident in (). ()
L H L nik il i
L H L ny a h a
LH L mb a
Melodies are especially prominent in and well motivated by systems of verbal inflection, where tone changes frequently are a part of tense–aspect inflection. () illustrates three tense-related melodies in Tiv. Roots lexically have H or L on the first syllable, and remaining tones are supplied by a grammatical melody.12 ()
general past (L) vá ‘come’ úngwà ‘hear’ yévèsè ‘flee’ past habitual (H) vá ‘come’ úngwá ‘hear’ yévésé ‘flee’ recent past (HL) vá ‘come’ úngwá ‘hear’ yévésè ‘flee’
dzà ‘go’ vèndè ‘refuse’ ngòhòrò ‘accept’ dzá ‘go’ vèndé ‘refuse’ ngòhóró ‘accept’ dzá ‘go’ vèndé ‘refuse’ ngòhórò ‘accept’
The general rule for distributing the tone melody is that the first tone of the melody goes to the second stem vowel—the first vowel lacking an underlying tone specification—and the second tone, if any, goes to the third vowel. Bantu languages are particularly rich with melodic patterns marking verb tense inflections (see Odden and Bickmore and the associated volume). One example is (Karanga) Shona. In lieu of a melodic tone, L stems only have L, and H stems have H on the first three syllables. () kù-bìkà kù-bìkìsà kù-bìkìsìrà kù-bìkìsìrànà
‘to cook’ ‘to make cook’ ‘to make cook for’ ‘to make cook for each other’
kù-tórá kù-tórésá kù-tórésérà kù-tórésérànà
‘to take’ ‘to make take’ ‘to make take for’ ‘to make take for each other’
A melodic H (double-underlined) appears in all negative and subordinate clause tenses, appearing on the second stem syllable in L roots (spreading once to the right), but in lexically H verbs the tone is on the final syllable, with an obligatory L between the lexical tone and the melodic H.
12
Tiv has no contour tones, so L is eliminated when combined with H tones; thus, predicted past habitual dzǎ becomes [dzá] and general past vâ becomes [vá].
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
L
H
hàvá-bìká hàvá-bìkísà hàvá-bìkísírà hàvá-bìkísírànà hàvá-tórà hàvá-tórèsá hàvá-tórésèrá hàvá-tóréséràná
‘They didn’t cook’ ‘They didn’t make cook’ ‘They didn’t make cook for’ ‘They didn’t make cook for each other’ ‘They didn’t take’ ‘They didn’t make take’ ‘They didn’t make take for’ ‘They didn’t make take for each other’
Bantu verbal tone melody patterns can be extremely complex; see, for example, the Luyia languages (Marlo , ). These complications involve not just particular tones defining the grammatical affix, but also complex mappings which differ from tense to tense, and which can be sensitive to lexical tone of the root, presence of object prefixes, and phonological properties of the stem such as syllable and mora count. In the following indefinite future data from Marachi (Marlo and p.c.), there are Hs towards the left edge of the stem, but exactly where those Hs begin and how far they extend to the right, or even whether they are present at all, is the result of a complex interaction between the presence of an object prefix, the lexical tone of the root, and the shape of verb stem. () L
H
‘he will V’ àlì-syá àlì-lóbá àlì-réébá àlì-bákálà àlì-lómálòmà àlì-lékhúúlà àlì-lyà àlì-bèkà àlì-rèèrá àlì-bùkúlá àlì-bùkúláángà àlì-bòdóókhànà
‘grind’ ‘refuse’ ‘ask’ ‘spread’ ‘talk’ ‘release’ ‘eat’ ‘shave’ ‘bring’ ‘take’ ‘be taking’ ‘go around’
‘he will V him’ àlì-mù-syá àlì-mù-lóbà àlì-mù-réébà àlì-mù-lé!khérá àlì-mù-ló!málómèrà àlì-mù-lé!khúúlà àlì-mù-ryá àlì-mù-békà àlì-mù-réèrá àlì-mù-bú!kúlá àlì-mù-bú!kúlírà àlì-mù-bó!dóókhànà
‘grind’ ‘refuse’ ‘ask’ ‘forgive’ ‘talk for’ ‘release’ ‘fear’ ‘shave’ ‘bring’ ‘take’ ‘take for’ ‘go around’
Melodic patterns are also common in Adamawa-Ubangian languages. The Ngbaka dialect of Henrix et al. () has three melodic tones added to roots, and the dialect of Englund () has six melodies: SH–L*, SL–SH, SH*, final L, initial L, and H. Verbs in Mbodɔmɔ (Boyd ) have no lexical tone, and receive one of three melodies determined by tense– aspect inflection, as is also true of Tupuri. Similar tonal melodies associated with tense– aspect are found in Gur languages, Grebo, and also occur in Chadic languages (Kotoko, Muyang, Dangla), as well as in Nilo-Saharan (Mbay).
.. Segmental influences It is particularly common to find phonological tone changes conditioned by the distinction between voiced and voiceless consonants, as discussed in Bradshaw (). An example is
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
the modification in Suma of the imperfective melodic H, which spreads to the left but not to a vowel after a voiced obstruent. () ɓúk ‘applaud’ fɔ́ɗí ‘stir briskly’ nĩ ḱ ĩ ŕ ĩ ́ ‘exaggerate’ bǒm ‘be blind’ bùsí ‘be bland’ zìkíɗí ‘delay’ Voiced obstruents are the most likely to trigger L-tone behavior. Voiced sonorants behave variably; for example, one rule in Ewe pertaining to imperative tone includes sonorants in the set of tone-lowering segments, whereas the depressor effect in Suma is triggered only by voiced obstruents. Almost universally, implosives behave as non-depressors, except in Zina Kotoko, where they are non-depressors for lowering of H to M, but depressors for the lowering of M to L. A synchronically very unusual segmental influence on tone is found in Tupuri (Odden ), where imperative verbs alternate when the first vowel is non-high between H and SL, determined by the initial consonant (SL after depressors, H otherwise) as in [dȁ] ‘want’ versus [há] ‘give’, but the tone is uniformly SH when the first vowel is high [hɪ ̋k] ‘dry’, [dűk] ‘vomit’).
. S
.................................................................................................................................. Despite substantial progress made in describing tone in African languages over the past forty years, there remain many unanswered questions about tone, primarily because of the lack of thorough descriptions of tone systems. Simple questions such as the maximum number of tone levels in languages or how tone levels pattern in terms of class behavior cannot be confidently answered, since languages relevant to testing theories are most often vastly under-studied. Such a lack of knowledge does not signal serious impediments to research into African tone systems: on the contrary, it indicates that there are innumerable opportunities, because something so simple as the discovery of a four-level language implies the possibility of finding substantive evidence bearing on how tones group into natural classes.
R Addo, T. (). The Tonology of Kafi Noonoo. M.A. thesis. Trondheim: Norwegian University of Science and Technology. Andersen, T. (). ‘Downstep in Päri: the tone system of a Western Nilotic language’, Studies in African Linguistics : –. Armstrong, R. (). ‘Yala (Ikom): a terraced-level language with three tones’, Journal of West African Languages : –. Bearth, T., and Link, C. (). ‘The tone puzzle of Wobe’, Studies in African Linguistics : –. Bearth, T., and Zemp, H. (). ‘The phonology of Dan (Santa)’, Journal of African Languages : –. Bergman, R. (). ‘Vowel sandhi and word division in Igede’, Journal of West African Languages : –. Bickmore, L. (). ‘Downstep and fusion in Namwanga’, Phonology : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Blaho, S. (). The Syntax of Phonology: A Radically Substance-Free Approach. Ph.D. thesis. University of Tromsø. Boyd, V. (). A Phonology and Grammar of Mbódɔ̀mɔ̀. M.A. thesis. Arlington, TX: University of Texas. Bradshaw, M. (). ‘One-step tone raising in Ali’, OSU Working Papers in Linguistics :–. Bradshaw, M. (). A Crosslinguistic Study of Consonant-Tone Interaction. Ph.D. thesis. Columbus, OH: Ohio State University. Cheng, Chin-Chuan, and Kisseberth, C. W. (). ‘Ikorovere Makua tonology (Part )’, Studies in the Linguistic Sciences : –. Christaller, J. G. (). A Grammar of the Asante and Fante Language called Tshi Chwee, Twi based on the Akuapem dialect with reference to the other (Akan and Fante) dialects. Basle: Evangelical Missionary Society. Clements, G. N. (). ‘Tone and syntax in Ewe’, in D. J. Napoli (ed.), Elements of Tone, Stress, and Intonation. Washington, DC: Georgetown University Press, –. Clements, G. N. (). ‘The hierarchical representation of tone features’, in I. Dihoff (ed.), Current Approaches to African Linguistics. Dordrecht: Foris, –. Clements, G. N., and Ford, K. C. (). ‘Kikuyu tone shift and its synchronic consequences’, Linguistic Inquiry : –. Clements, G. N., and Ford, K. C. (). ‘On the phonological status of downstep in Kikuyu’, in D. L. Goyvaerts (ed.), Phonology in the ’s. Ghent: Story-Scientia, –. Clifton, J. M. (). ‘Downdrift and rule ordering’, Studies in African Linguistics : –. Dawd, A., and Hayward, R. J. (). ‘Nara’, Journal of the International Phonetic Association : –. Dihoff, I. (). Aspects of the Tonal Structure of Chori. Ph.D. thesis. Madison, WI: University of Wisconsin. Downing, L. (). The Tonal Phonology of Jita. Munich: Lincom Europa. Downing, L. (). ‘What African languages tell us about accent typology’, ZAS Papers in Linguistics : –. Dwyer, D. (). ‘What sort of tone language is Mende?’, Studies in African Linguistics : –. Englund, P. (). Ngbaka: Phonology and Verb Morphology. Ph.D. thesis. Evanston, IL: Northwestern University. Flik, E. (). ‘Tone glides and registers in five Dan dialects’, Linguistics : –. Frajzyngier, Z. (). A Grammar of Mupun. Berlin: Dietrich Reimer. Fromkin, V. (). ‘Tone features and tone rules’, Studies in African Linguistics : –. Goldsmith, J. (). Autosegmental Phonology. Ph.D. thesis. Cambridge, MA: MIT. Green, M. (). ‘The classification of West African tone languages: Igbo and Efik’, Africa : –. Henrix, M., van den Eynde, K., and Meeuwis, M. (). Description grammaticale de la langue ngbaka: Phonologie, tonologie et morphosyntaxe. Munich: Lincom Europa. Hieda, O. (). ‘Tonal system in Kumam, a double downstep language’, Journal of Asian and African Studies : –. Hyman, L. M. (). ‘Tonology of the Babanki noun’, Studies in African Linguistics : –. Hyman, L. M. (). ‚Tonal accent in Somali’, Studies in African Linguistics : –. Hyman, L. M. (). ‘Register tones and tonal geometry’, in H. van der Hulst and K. Snider (eds.), The Phonology of Tone: the representation of tonal register. Berlin and New York: Mouton de Gruyter, –. Hyman, L. M. (). ‘Privative tone in Bantu’, in S. Kaji (ed.), Cross-Linguistic Studies of Tonal Phenomena. Tokyo: Institute for the Study of Languages and Cultures (TUFS), –. Hyman, L. M., and Schuh, R. (). ‘Universals of tone rules: evidence from West Africa’, Linguistic Inquiry : –. Hyman, L. M., and Tadadjeu, M. (). ‘Floating tones in Mbam-Nkam’, in L. M. Hyman (ed.), Studies in Bantu Tonology. Los Angeles, CA: University of Southern California, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Jones, R. (). The Boko/Busa Language Cluster. Munich: Lincom Europa. Keegan, J. (). A Reference Grammar of Mbay. Munich: Lincom Europa. Kenstowicz, M. J., and Kisseberth, C. W. (). ‘Chizigula tonology: the word and beyond’, in S. Inkelas and D. Zec (eds.), The Phonology-Syntax Connection. Chicago, IL: University of Chicago Press, –. Leben, W. (). Suprasegmental Phonology. Ph.D. thesis. Cambridge, MA: MIT. Ma Newman, R. (). ‘Downstep in Ga’anda’, Journal of African Languages : –. Marlo, M. R. (). The Verbal Tonology of Lunyala and Lumarachi: Two Dialects of Luluyia (Bantu, J., Kenya). Ph.D thesis. Ann Arbor, MI: University of Michigan. Marlo, M. R. (). ‘Tura verbal tonology’, Studies in African Linguistics : –. Morén, B. (). ‘The parallel structures model of feature geometry’, Working Papers of the Cornell Phonetics Laboratory : –. Morse, M. (). A Sketch of the Phonology and Morphology of Bobo (Upper Volta). Ph.D. thesis. New York, NY: Columbia University. Newman, P. (). ‘Contour tones as phonemic primes in Grebo’, in K. Bogers, H. van der Hulst, and M. Mous (eds.), The Phonological Representation of Suprasegmentals. Dordrecht: Foris, –. Odden, D. (). ‘Tonal phenomena in Shambaa’, Studies in African Linguistics : –. Odden, D. (). ‘Tone: African languages’, in J. Goldsmith (ed.), The Handbook of Phonological Theory. Oxford: Blackwell, –. Odden, D. (). ‘Typological issues in tone and stress in Bantu’, in S. Kaji (ed.), Cross-Linguistic Studies of Tonal Phenomena: Tonogenesis, typology, and related topics. Tokyo: Institute for the Study of Languages and Cultures (TUFS), –. Odden, D. (). ‘Features impinging on tone’, in J. Goldsmith, E. Hume, and L. Wetzels (eds.), Tones and Features. Berlin: Mouton de Gruyter, –. Odden, D., and Bickmore, L. (). ‘Melodic tone in Bantu’, Africana Linguistica : –. Paradis, C. (). Description phonologique du guéré. Abidjian: Institut de linguistique appliquée. Paster, M., and Kim, Y. (). ‘Downstep in Tiriki’, Linguistic Discovery : –. Pulleyblank, D. (). Tone in Lexical Phonology. Dordrecht: Riedel. Reiss, C. (). ‘Quantification in structural descriptions: attested and unattested patterns’, The Linguistic Review : –. Riccitelli, J. (). ‘Tone analysis: a practical approach’, The Bible Translator : –. Roberts-Kohno, R. R. (). Kikamba Phonology and Morphology. Ph.D. thesis. Columbus, OH: Ohio State University. Schachter, P., and Fromkin, V. (). A Phonology of Akan: Akuapem, Asante, Fante. UCLA Working Papers in Phonetics . Los Angeles, CA: University of California. Theil, R. (). ‘Kafa phonology’, Journal of African Languages and Linguistics : –. Wedekind, K. (). ‘A six-tone language in Ethiopia’, Journal of Ethiopian Studies : –. Welmers, W. (). ‘Notes on the structure of Bariba’, Language : –. Winston, F. D. D. (). ‘The “Mid tone” in Efik’, African Language Studies : –. Yip, M. (). The Tonal Phonology of Chinese. Ph.D. thesis. Cambridge, MA: MIT.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. M is often conceptualized as the center of linguistics. Being the study of word structure, it is considered the connecting part between the lower levels of linguistic description, i.e. phonetics and phonology, and the higher descriptive tiers of syntax and semantics (Spencer and Zwicky : ). Its connecting function is also reflected by the compositional terms “morphophonology” and “morphosyntax” that refer to the sometimes fuzzy transition zones between these conceptual layers. These transition zones are, however, only pertinent when the syntagmatic dimension of morphology is addressed, i.e. the sequential expression of meaningful elements. The paradigmatic dimension, on the other hand, leads to the concept of word (or morpheme) classes in which all possible elements for a given syntagmatic position are listed. The descriptive unit of morphology is the morpheme, conventionally signaled by curly brackets. A general and recurrent definition states that it is “the smallest meaningful unit” in a given language. Possible meanings of morphemes range from pure lexical references that usually form open classes to highly complex grammatical functions, usually in finite or closed sets. The syntagmatic status of a morpheme may be anything between a completely self-dependent element (i.e. a free, lexical morpheme) and a unit neatly tied to another (bound or unbound) element expressing a pure grammatical relation (i.e. a bound, grammatical morpheme). The morphological systems found in languages south of the Sahara display a wide variety and reflect nearly the whole range of morphological types found worldwide. There are languages of the “isolating” type that display very little bound morphology and are therefore nearly void of inflection and derivation. Those languages are often, though not exclusively, located in West Africa. Many other languages, e.g. those from the Bantu family, adhere to the “agglutinating” type. They display elaborated paradigms of bound morphology with a one-to-one relation of grammatical function and morpheme. A third recurrent system, though slightly less important in numbers, is found, for instance, among AfroAsiatic languages like Hausa. They display bound morphology and a one-to-many relation between form and function. In such a system a single morpheme may cover, for example, the grammatical categories gender, number, and aspect without allowing a clear separation of these functions on the surface level. Such an element is also called a “portemanteau” morpheme as it carries several functions in one morphological unit.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Morphemes that express grammatical relations display a great number of structural variations in terms of position and shape. Languages may, for instance, prefer the morphological marking of syntactic heads over dependent elements or vice versa, or even mark both parts. They may use prefixes at the beginning of a word, suffixes at its end, circumfixes on both sides, or even infixes that split the base to which they are attached. Some languages also may use tonal exponents for grammatical functions, while others again work with stem-internal modifications of vowels and/or consonants. From a functional viewpoint, language is a means of communication between people and the above-mentioned morphological types merely reflect different stages of “complexification” (Trudgill ) of this means. Therefore, one can assume that there are also social factors that account for morphological complexity. The social factors that seem most influential are low amounts of adult language contact, high social stability, small group size, dense social networks, and large amounts of communally shared information (Trudgill : ). These factors are not the only ones responsible for the development or reduction of complexity in morphological systems, but they seem to play important roles. They are closely related to even more general factors like “time” and “usage” that also play their parts in the development of those systems. Generally speaking, every morphological system changes over time. Languages tend to lose bits and pieces of their morphological inventory through processes of fusion and erosion triggered through use by their speakers in regular communication. Obviously, languages are also able to (re)introduce formerly lost functions and even create new ones by exploiting internal linguistic means or eventually even using external (borrowed) linguistic material. Such diachronic processes of “morphologization” or “morphemization” are described by the theory of grammaticalization and embrace a cyclic understanding of the development of morphological systems. Most of the above-mentioned phenomena will be illustrated in the following sections of this chapter with the most salient and recurrent features put center stage, but also with occasional hints to some morphological quirks found in African languages.
. M A
.................................................................................................................................. A classification of languages according to ideal morphological types was already developed in the late eighteenth and early nineteenth century by Wilhelm von Humboldt and August Wilhelm Schlegel. They coined the notions of “analytic” and “synthetic” morphology according to the form of grammatical elements. The synthetic type employs at least some affixes for the expression of grammatical function, while the analytic type makes extensive use of free grammatical morphemes (no segmental affixes) and strict word order. A more frequent morphological typology uses the term “isolating” (for analytic) to describe languages that show a one-to-one correspondence between words and morphemes. In other words, most morphemes are lexical or free, while bound morphology plays no role in the grammar. In Africa, languages from the Kwa and Mande families are often quoted as typical representatives of isolating languages. Following this terminology, languages are further divided according to the ratio of grammatical function per
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
morpheme. Thus, an “agglutinating” language ideally has a one-to-one ratio between morpheme and grammatical function, which is a typical feature of Bantu languages. A so-called “fusional” language covers several functions with just one morpheme (e.g. person, number, tense, mood) and morpheme boundaries are less clear-cut than in agglutinating languages. Afro-Asiatic languages like Hausa are frequently quoted as African examples. A third synthetic type is constituted by “polysynthetic” languages. They tend to combine several grammatical and lexical elements in just one word (Comrie : –). However, no reference cases from Africa are known for this grammatically highly complex type. Another approach defines morphological language types along a two-dimensional grid where one axis provides the grade of synthesis and the other the grade of fusion. The first parameter counts morphemes per word and oscillates between isolating and polysynthetic type. The second parameter looks at the segmentability of morphemes within the word; its two extremes being agglutination where segmentation is straightforward, and fusion where segmentation is not possible (Comrie : ). Although this approach has also difficulties with the definition of basic elements like words and morphemes and the question whether or not concatenation of elements is obligatory, it accounts for the independence of synthesis and fusion. More recently, the possibility of whole-language morphological typologies was rejected altogether. The morphological scale running from isolation via agglutination to fusion conflates three different parameters that do not necessarily co-vary, as was assumed for long (Bickel and Nichols ). These parameters are fusion, exponence, and flexion, and they combine with one another in individual morphological processes within any given language (Velupillai : ). Although most researchers subscribe to this view and admit that pure morphological types rarely (if at all) exist in the world’s languages, such typologies are still used as a heuristic for the morphological characterization of languages (Childs : , ; Katamba : ; Childs : , ). For instance, the Malian language Supyire (Niger-Congo, peripheral Gur) is isolating with regard to tense and case formatives and expresses only one TAM category per morpheme. It also shows some fusional tendencies as it marks up to three inflectional categories on the verb (Bickel and Nichols ). Furthermore, Carlson (: –) reports tendencies of object incorporation, which is a typical feature of polysynthetic languages. Consider example (), where the isolating markers {be} and {} coexist with an agglutinative definite suffix and the fusional imperfective marking on the verb. Moreover, the object noun ‘mouth’ is reanalyzed as part of a compound verb. This can be seen from its appearance in the bare indefinite singular form which would otherwise be impossible because, as part of an ordinary possessive construction, it would obligatorily be marked for definiteness and plural, thus cross-referencing the possessor: ()
Guvernamí-ŋi ɲyε na pi ɲwɔ càà. government- be their mouth seek. ‘The government feeds him.’
(Carlson : )
This example illustrates what is generally true for language typologies: clear-cut cases of ideal morphological types can hardly be found! It is, therefore, also difficult to quantitatively estimate the various morphological types that exist on the African continent.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Hence, the following characterizations of some exemplary languages according to morphological types can only be understood as a broad approximation. Definitely the individual languages need to be described in detail to account for their often very complex and interrelated morphological structures.
.. Isolating languages As already stated, many languages from the Mande family are often cited as textbook cases for isolating languages. Here are some phrases from Bambara (Manding): ()
a. Hawa tέ jέgε sàn. Hawa : fish kill ‘Hawa doesn’t buy fish.’ b. Hawa yé jέgε sàn. Hawa fish kill ‘Hawa did buy fish.’
(Dumestre : )
All elements appear in their bare lexical form that would not change under most syntactic circumstances apart from vowel syncope in fluent speech. The copula tέ is one of the rare instances in the language where a one-to-many relationship, i.e. imperfective + negation, is explicit. There is, however, a complexification that was often overseen by researchers looking at typological characteristics. For instance, Bambara, a close relative to Manding, is a tone language with both, lexical and grammatical tone. The definite marker in the following example is expressed only by the falling tone on the final vowel of the noun and therefore adds some grammatical complexity, hence challenging the definition of the morpheme (see also () and () in section .. below): ()
dònsó` ká fárin. hunter. brave ‘The hunter is brave.’
(Dumestre : )
The one-to-one relation of morphemes and lexical units is pretty obvious with the exception of the definite marker. To compensate for such an analytic morphological structure where not much syntactic cross-referencing or semantic role assignment on any single element is possible, languages of this type usually employ a very strict syntactic order. Many languages of the predominantly isolating type from the Macro-Sudan belt (Güldemann ) of West Africa therefore display a strict S–A–O–V–(other) order of meaningful elements (cf. Vossen, chapter of this volume and Lafkioui, chapter of this volume). Another case of a fairly isolating-analytic language is found much farther south among South African Khoisan languages. The Northern Khoisan language !Xuun is a case in point. Its basic word order is S–V–O and it displays lexical and grammatical tone: ()
Mī kú shē ’āà. see ‘I see you.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
[ . . . ] ta n!ēí !hm̄ m̄ ke nǀāò pòhó [ . . . ] and then leopard treat.well jackal ‘[ . . . ] and then the leopard treated the jackal well [ . . . ]’ (Heine and König : )
An interesting quirk in the morphology of the !Xuun dialects is a set of suppletive verbs where two different lexical forms refer to singular and plural participants respectively. Thus, in transitive constructions verb stem selection reflects the number of the clausal object: ()
a. n!hȁì !hún gúmí. lion kill: cattle ‘The lion killed the cow.’ b. n!hȁì !’óán gúmí. lion kill: cattle ‘The lion killed the cattle.’
(Heine and König : )
Contrary to this, the languages from the Khoe part of the family display a rich inflectional and derivational morphology, as will be shown in section .. below.
.. Agglutinating languages The textbook case of an agglutinating structure in Africa is found in Bantu languages. Swahili, for instance, displays noun classes with agreement not only within the noun phrase but also cross-referencing the noun classes on the verb for subject and, in transitive constructions, for object position. Agreement is overtly marked so that in case of phonetic identity between noun class and agreement marker an alliterating structure evolves: ()
a. ki-kapu ki-kubwa ki-moja -basket -large -one ‘One large basket fell.’
ki-li-anguk-a. .--fall-
b. vi-kapu vi-kubwa vi-tatu vi-li-anguk-a. -basket -large -three .--fall- ‘Three large baskets fell.’
(Welmers : )
Swahili also displays a rich derivational and inflectional morphology on the verb. If one compares Swahili with English, it becomes clear that the former uses bound morphology where the latter employs lexical and free grammatical morphemes. Put differently, an English five-word sentence comes out as a single word in Swahili: ()
ni-na-mw-andik-i-a. :-:-:-write-- ‘I am writing to him.’
Rich noun class morphology and a wide range of derivational means are prevalent in Bantu languages but are also known from many other languages of the Niger-Congo phylum.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
For instance, the mainly agglutinating language Fulfulde from the Atlantic family1 is equally rich in this respect. Depending on the dialect, there are up to twenty-five noun classes and a multitude of derivational morphemes that can also combine to create verb stems of impressive length. The following examples of verbal derivation are from the westernmost Fulfulde dialect, Pulaar, spoken in Guinea and adjacent areas: ()
a. winnd-ugol winnd-id winnd-an winnd-indir winnd-in winnd-inkin winnd-it winnd-or winnd-ir winnd-oy
‘to write’ ‘write together’ ‘write for someone’ ‘write each other/write in’ ‘let write’ ‘pretend to write’ ‘rewrite’ ‘write occasionally’ ‘write with sth.’ ‘write at another place/time’
/
+ + + add-it-an-oy-gol ‘to bring again for someone + + from far away’ + add-an-oy-gol ‘to bring from far for someone’ + + (Diallo : ) [English translation of glosses is mine; KB]
b. add-ugol add-it-ugol add-it-an-gol
‘to bring’ ‘to bring again’ ‘to bring again for someone’
The concatenation of verbal derivatives follows a specific order. As becomes clear from the examples above, for instance, the repetitive suffix {-it} always precedes the causative and the distantive elements. First in order always comes {-ɗ } which derives verbs from nouns or adjectives; {-or} and {-an} are interchangeable in their position (see also section .. below): () -ɗ, -in, -it, -inkin, -id, -(i)ndir, -ir, -or/-an, -oy
(Diallo : )
As already stated, rich noun class morphology and elaborated systems of derivational morphological means are some of the main features of the Niger-Congo phylum of African languages. More often than not, the morphological exponents of those features express just one function per morpheme, as in the Bantu examples above, so that one may even consider agglutinating structure to be the unmarked morphological type of Niger-Congo languages.
Whether Atlantic is an integral part of Niger-Congo is debatable (see Lüpke, chapter of this volume). 1
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Fusional languages Fusional languages employ mainly morphemes which comprise more than one grammatical meaning under a single morphological umbrella. In Africa south of the Sahara, such one-tomany morphological relationships are largely known from Afro-Asiatic languages. For instance, Qafar (East Cushitic, Cushitic), which is mainly spoken in the “Danakil Depression” of Ethiopia, displays an interesting mix of pre- and suffixed portmanteau morphemes some of which even combine with ablaut (Hayward : –). There are mainly two verb classes that behave differently in terms of pre- and suffixing morphology, but both sets carry information on gender, number, and aspect: (a)
Class I .: y-eeɖegeh ‘he knew’ .: t-eeɖegeh ‘she knew’ : n-eeɖegeh ‘we knew’
Class II fak-eh ‘he opened’ fak-t-eh ‘she opened’ fak-n-eh ‘we opened’
A second paradigm (related to transitivity) also shows only suffixes for class I, but now in combination with ablaut for aspect and certain mood distinctions, whereas class II verbs confine themselves to the corresponding suffixes: (b)
Class I .: yeeɖeg-e .: yaeɖig-e .: yaeɖág-ay .: iɖig
Class II ‘he knew’ fak-e ‘he knows’ fak-a ‘let him know!’ fak-ay ‘know!’ fak
‘he opened’ ‘he opens’ ‘let him open!’ ‘open!’ (Hayward : )
The Central Khoisan language Khoekhoe, also known as Nama/Damara and mainly spoken in Namibia, exhibits some characteristics of a fusional language. Even though Haacke (: ) characterizes the language as of a “predominantly isolating type”, the portmanteau characteristics of its PGN (person–gender–number) markers for nouns show clear fusional tendencies. These PGNs come in two different flavors that relate to the syntactic position of the nouns they specify. Syntactically, they often function as pronouns and are then considered as free (isolated) portmanteau morphemes. But they also serve as noun suffixes specifying semantic and syntactic function. Consider the following paradigm: ()
Person
Masculine
Feminine
Neuter
I
II
III
I
II
III
I
III
III
ta
ts
b/mi/ni
ta
s
s
~i
Dual
khom
kho
kha
m
ro
ra
m
ro
ra
Plural
ge
go
gu
se
so
di
da
du
n
Singular
(Haacke : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A surface noun comprises a lexical stem and a clitical PGN, thus conveying syntactical meaning too (). For inanimate objects, the PGN also differentiate semantic meaning, where masculine nouns tend to signify larger or elongated objects, while feminine nouns denote relatively small and roundish objects, as in (): ()
tara-s ‘woman’ (.) tara-ta ‘I, a/the woman’ ()
() !ā-b hai-b dao-b om-mi
Masculine ‘river’ ‘stick; tall, narrow tree’ ‘path, way, track’ ‘large building’
!ā-s hai-s dao-s om-s
Feminine ‘village, town’ ‘plant; bush; tree (in general)’ ‘doorway’ ‘hut, house’ (Haacke : )
The language also disposes of a morphosyntactic feature that is commonly attributed to highly synthetic, i.e. polysynthetic languages. In the same vein as the Malian Supyire mentioned earlier in this section, Khoekhoe can incorporate object nouns, postpositions, and simple postpositional phrases into the verb (Haacke : ): ()
ao-b ge !ari-sa ra !au. ! ao-b ge ra !ari!au. man steenbuck hunt man steenbuck.hunt ‘The man is hunting steenbuck.’ ‘The man is steenbuck hunting.’
This incorporating device is thus also corroborating the claim that Khoekhoe is not a “purely” isolating language at all. It serves likewise as an example of the appearance of polysynthetic features in African languages, while textbook cases of this language type have not (yet?) been found in Africa.
. S A
.................................................................................................................................. Given the impossibility of describing all morphological devices that can be found in African languages, we look at some of the more salient means which may not be specific for the continent alone but do play an important role in the majority of its languages. To this end, we firstly consider the continuum of word formation strategies ranging from inter- and intra-word class derivation via reduplication to compounding. Secondly, we look at inflectional systems, including noun class morphology and morphological verb classes. In this section, we also consider grammatical tone and other suprasegmental devices that convey grammatical meaning.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Derivation There are innumerable derivational means that cover the whole range of possibilities, including switches between word classes and intra-word class operations leading, for instance, to valence change of verbs and/or more or less far-reaching changes in meaning.
... Inter-word class derivation Like most languages of the world, African languages have devices to derive nouns from verbs. The resulting nouns denote the underlying verbal state of affairs as a fact, an act, or a quality. There are also cases where verbs are derived from nouns, though to a lesser degree. As a matter of fact, African languages often exhibit no or just a small and closed class of genuine adjectives. Accordingly, word class-changing derivations predominantly commute between verbs and nouns. In Dagara (Gur, Niger-Congo), for example, an element {-d-} is used to derive a verbal adjective (or participle) from a verb. The resulting qualifying phrase only adapts in number to the head noun (as Dagara’s noun class system is highly eroded): () bʊ́ dɪ-d-á bʊ́ dɪ-d-í thing eat-- thing eat-- ‘edible thing(s)’ ‘food’
(Delplanque : )
A related nominalizing {*R} is found in adjacent languages where it is used for the same purpose. In Mampruli (Gur, Niger-Congo) deverbal nouns may also occur in nominal compounds (see also section .. below). As in Dagara, the well-known strategy for the formation of general terms involves a dummy noun meaning ‘thing’: ()
dayεlli bunni bunni
‘stick’ ‘palaver’ ‘thing’ ‘thing’
+ + + +
ba sɔˈai bɔ’ai yigi
‘weigh down’ ‘hide’ ‘crawl’ ‘fly’
> > > >
dabaarigu yεlisɔaˈra bumbɔ’arigu bunyigrigu
‘roof batten’ ‘secret matters’ ‘creepy-crawly, reptile’ ‘flying creature’ (Naden : )
In Gur languages we find time and again a formative {*R} that is involved in deverbal derivation. Be it the aforementioned verbal adjectives or the formation of agent and instrument nouns, it is most likely an imperfective verb stem expanded with {*R} and endowed with the appropriate noun class marker (where noun classes are still active) that leads to respective nominal forms as, for example, in Pana: ()
cébí ‘cut.’ + mègí ‘beat.’ +
-*R- + -*R- +
/ > / >
cébírì, cébírò ‘cutter(s)’ mègírì, mègírò ‘fighter(s)’ (Beyer : )
That this kind of deverbal derivation in combination with nominal compounding is highly productive is documented by neologisms in Mampruli:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() bɔli ‘ball, football’ + bisa ‘breast’ +
ŋmε ‘hit’ > ɡoaii ‘restrain’ >
bɔliŋmeera ‘footballer’ bisiɡoara ‘bra’ (Naden : )
The presentation of {*R} as a derivational formative with the common function {} in Gur languages blurs the fact that its underlying consonants [r d l] appear in many more derivative and inflectional functions. Sometimes they are reduced to verbal extensions with unclear or only broadly related meanings. This holds true not only for the Gur family (Manessy ; Naden ), as evidenced by individual languages such as Mooré (Kaboré ; Nikiema ), but is also the case on the Niger-Congo level as a whole (Voeltz ; for an overview, see Beyer ). While derivation of agentive, instrumental, and locative nouns is common practice in many African languages, there are many more category-changing means with sometimes more quirky outcomes. For instance, though less frequent and more restricted, verbs derived from nouns also occur. In Hausa (Chadic, Afro-Asiatic) a restricted set from the closed class of Abstract Nouns of Sensory Quality is able to derive verbs. These take a suffix {-aCa(a)} and build the socalled Hausa verb grades through application of the respective tone patterns. Grade verbs convey transitive and grade verbs intransitive-inchoative meaning: () Noun Grade verb káushíi ‘roughness’ káusàsáa ‘roughen’ táuríi ‘hardness’ táuràráa ‘harden’ tsárkíi ‘cleanliness’ tsárkàkáa ‘purify’
Grade verb kàusásà ‘become rough’ tàurárà ‘become hard’ tsàrkákà ‘become pure, clean’ (Newman : )
... Intra-word class derivation Another means of changing the semantics (but not the word class) of a nominal base is very frequent among languages displaying noun class morphology. The transposition of a lexeme from its default gender to another gender often signifies additional semantic connotations. In Diola-Fogny (Atlantic, Niger-Congo) diminutive and augmentative readings are rendered by noun class transpositions. Consider the nominal bases for -ko ‘head’ and -ɲil ‘child’: ()
Singular Normal fu-ko/a-ɲɪl Augmentative bu-ko/ fʊ-ɲɪl Diminutive ji-ko/ jɪ-ɲɪl
Plural Collective ku-ko/kʊ-ɲɪl ɲi-ko/ ɲɪ-ɲɪl mu-ko/ mʊ-ɲɪl bə-ko/ ba-ɲɪl
(Sapir : )
Sometimes, the derivational base is not clear. In Kulango (peripheral Gur, Niger-Congo) a class of ambivalent lexical bases exists that does not indicate the direction of the derivation: () bi-yò, bu-wù ‘child(ren)’ < > bi-gè, bu-hù ‘fruit(s), nut(s), seedling(s)’ kɪlɪ-ɡɔ̀ ‘stranger (to a place)’ < > kɪlɪ-yɔ,̀ kɪlɪ-bɔ̀ ‘unknown person(s)’ (Elders : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
In the verbal realm derivation is a common mechanism in African languages for changing meaning, valence, and aspectual notions of verbs. Particularly well-developed systems of suffix derivation are known from Bantu languages. Schadeberg (: ) lists eleven common suffixes called “verb extensions” that can be reconstructed for Proto-Bantu. In most languages they can combine and thus form derivational chains as long as it makes sense on semantic grounds. Generally speaking, the more productive extensions follow less productive ones, and any additional extension takes the already derived verb as its phonological and semantic base (Schadeberg : ). For instance, in Southern Sotho a derivational chain may look like this: () -sheba ‘watch X’ + > -sheb-is-a ‘cause Y to watch X’ + > -sheb-is-ets-a ‘cause Y to watch X on behalf of Z’ + > -sheb-is-ets-an-a ‘cause Y to watch X on behalf of each other’ (Doke and Mofokeng : –)
.. Compounding In African languages, compounding is also a widely used means from the morphological toolbox. The most conventional types are determinative compounds with a modifier–head structure; they are usually further divided into endo- and exocentric constructions. Generally speaking, the head of an endocentric determinative compound assigns its morphosyntactic properties and the semantic basis to the construction as a whole. Thus, in Supyire the second noun in a noun–noun compound is usually the syntactic head of the construction, as can be seen from its noun class assignment, and it also provides its semantic basis, as becomes clear from the translation: () kampe-e ‘finger-’ + fègè-wè > kampe-fègè-wè ɲyi-ii ‘eye-’ + lwɔ-hɔ > ɲyi-lwɔ-hɔ ŋ̀kù-ù ‘chicken-’ + cere > ŋ̀kù-cèrè
‘ring-’ ‘fingerring-’ ‘water-’ ‘eye-water-/tears’ ‘egg-’ ‘chicken egg-’ (Carlson : –)
An endocentric compound may again serve as input for an even longer compound, with a third noun acting as head in such a triple noun structure: () ɲuŋɔ ‘head.’ + wyi-i ‘hole-’ ‘head-hole-/outer ear canal’ + fu-ro ‘excrement-’ > niŋ-gyí-í > niŋ-gyí-fú-ró ‘head-hole-excrement-/ear wax’ (Carlson : –) In Supyire, as in many other African languages, nouns like ‘wife’, ‘child’, and ‘husband’ gain more derivative-like properties, as can be seen (at least partially) from the noun class assignments in such compounds, where one would expect the gender (G) which is the default for human (animate) beings:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() bòm-pὲὲ-gὲ sika-pèrè pwunɲ-cwɔ̀ ŋ̀kùli-cwɔ́-ɡɔ́ nù-pyà kacii-pyá
‘baboon-male- / male baboon’ ‘goat-male. / billy-goat’ ‘dog-female. / bitch’ ‘cockroach-female- / female cockroach’ ‘cow-child. / calf ’ ‘ax-child. / ax blade’ (Carlson : –)
Exocentric compounds where the meaning of the construction is not transparent from its individual components are also quite frequent. The metaphoric connection may or may not still be visible. In the following data from Pana the last two compounds take a deverbal agentive noun and a nominalized verb meaning ‘the state of ’ as their heads: () ʝùsó ‘hunt’ + kà̰á̰ ‘wife’ > ʝùsòká̰ ‘second wife’ bṵ̀ntòló ‘billy-goat’ + lùré ‘testicle’ > bṵ̀ntòlòlúrè ‘aubergine’ yὲrέ ‘body’ + ɲòará ‘fact of decaying’ > yὲrέɲóarà ‘agony, suffering’ nὲ̰έ ̰ ‘hand’ + mó ‘hold’ + -rì ‘:’ > nὲ̰móorì ‘witchdoctor’ (Beyer : )
.. Inflection Inflectional systems cover the whole range of verbal and nominal categories and are generally agglutinating or fusional in nature. There are also languages that are usually categorized as “isolating” but display inflectional categories through tonal changes, which calls this categorization into question. In Twi (Kwa, Niger-Congo), for instance, a regular verbal predication involves a subject pronoun, a TAM marker, and the bare verb. In some cases the segmental TAM marker is missing so that inflectional categories are rendered solely by tonal means. Consider the difference between habitual and stative verb forms: () ɔ́-kɔ̀ ‘he walks’ (habitual) ɔ̀-sòmá ‘he sends’ (habitual) ɔ̀-wɔ̀ há ‘he/she is here’ (stative) ɔ-̀ ɡyìnà hɔ́ ‘he stands there’ (stative) (Reineke : ) Habitual (or generic) reading is expressed without any segmental support through specific tone patterns expanding over the pronoun and the verb: HL for monosyllabic and LLH or HLL for disyllabic verbs. A stative reading is expressed by a general low tone pattern expanding over the whole construction (Reineke : –). The Manding languages, often cited as prototypical isolating, also have a purely tonal inflection. In Bambara (Mande, Niger-Congo), for instance, a so-called “floating” low tone marks definiteness. Its presence is not tied to any segmental representation but is reflected in tonal changes of surrounding elements. In the following negative statement the lowered high tone on the predication marker {PFV:NEG} is due to the floating low tone marking definiteness on the preceding noun: () fúlákέ Fulbe.man fúlákέ Fulbe.man:
má : mā :
nà. come nà. come
‘No Fulbe man has come.’ ‘The Fulbe man has not come.’ (Creissels : , cited from Tröbs : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Cases like () and () are strong arguments against the classification as “isolating”, because, with the suprasegmental marking of grammatical relationships, the one-to-one relation between word and morpheme is no longer existent. Tonal inflection is also widely known in agglutinating and fusional languages. Often it is a combination of tone and segmental change that carries the grammatical load. A rather complicated system of interrelated tone and segmental patterning is reported from Kanakuru (Chadic, Afro-Asiatic) where grammatically determined tone patterns interact—for instance, in the marking of tenses—with underlying lexical tones of the verbal stem, with the syntactic position of the verb, and with different stem types. Without going into detail, we just present a small comparative chart where (tense) pronouns and verbs carry different tones according to the different subjunctive sets they are used in: () bə̀là nà nái gwá bə̀là náa nài gwá bə̀là wù nái gwá bə̀là wún nái gwá
‘that I drink water’ ‘that I drink water’ ‘that we drink water’ ‘that they drink water’
(Newman : )
The only semantic difference between the two subjunctive patterns refers to logophoricity: the use of the second form indicates that a third person subject of the main clause is not the same as the subject of the embedded clause (Newman : –).
.. Reduplication Cross-linguistically, reduplication serves a wide variety of functions. It may cover both derivation and inflection in that it creates either new lexical meaning or just changes the word form (e.g. from singular to plural). The duplicated parts may range from complete repetition of the base to partial duplication of single syllables. The grammatical and syntactical functions range from mere plural marking to expression of emphasis and causative derivation. Ewe (Kwa, Niger-Congo) has a rich reduplicating morphology that serves several purposes. In () reduplication derives attributive adjectives from stative verbs: ()
kɔ̌ ‘be tall’ > kɔ́kɔ́ ‘high, tall’ sḛ̌ ‘be strong’ > sésẽ ́ ‘strong, hard, difficult’ fǎ ‘be cold’ > fáfá/fáfέ ‘cold, cool’
(Dzablu-Kumah : )
In () a reduplicated monosyllabic verb enlarged by a bilabial nasal {-ḿ} yields the present progressive in a construction with the auxiliary le ‘to be’. If an object is present in the VP, it is expressed in preverbal position and reduplication no longer occurs: () yi ‘go’ > mele wɔ ‘do’ > mele > mele dɔ́
yiyiḿ. ‘I am going.’ wɔwɔḿ. ‘I am doing.’ wɔḿ. ‘I am doing work.’
(Westermann : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Reduplication—by its mere appearance—has some iconic force to it. So it is not surprising that it often expresses emphasis, intensity, plurality, and the like. All this can be seen in Hausa, where (whole and partial) reduplication covers a wide range of functional processes. Here is a small list: ()
Pluralization
jóojì bâm
‘judge’ ‘bomb’
> >
jóojì-jóojì ‘judges’ bàmàabàmái ‘bombs’
Intensification
sáafé
‘early in the morning’ ‘quickly’
>
sássáafé
>
‘call’ ‘get lodged in’
> >
mármázá ‘very quickly’ (mázá mázá) kírkíráa ‘call many/often’ máƙálƙàlée ‘get all lodged in’ (Newman : )
mázá Pluractional verb kíráa máƙàlée
‘very early . . . ’
A quite peculiar case of double reduplication is reported from another Chadic (AfroAsiatic) language: Mandara. In this language, both perfective aspect and reference to a plural object may be expressed by reduplication of the verb. The reduplicated form indicating the plural object serves as input for the formation of the perfective, which in turn infixes a subject pronoun (which is a Ø-morpheme for ) between its two parts: () wá-wà shoot-: wá-rú-wà shoot-- wáwà-wáwà shoot:-: wàwá-r-wàwá shoot:--
gígálè. rat ɡíɡálè. rat gìgálà-xà. rat- gìgálà-xà. rat-
‘He/she shot a rat.’ ‘They shot a rat.’ ‘He/she shot rats.’ ‘They shot rats.’ (Frajzyngier : –)
.. Morphological classes More generally speaking, morphological classes are divided into open and closed sets. While nouns and verbs are usually elements of open classes of which no single speaker knows all members, other elements, e.g. demonstratives, usually form closed sets where all members are known to the speakers. A further division of morphological classes relates to their syntactic status in a given language. Where free forms may occur on their own, bound forms need some support for them to be realized correctly. Of course, all these generalizations only work from a strictly synchronic perspective. From a diachronic point of view, even the most hermetically closed class may change over time (see section . below). Yet another, final remark on one of sub-Saharan Africa’s most famous morphological structures seems in order here. The ubiquitous noun class systems of Niger-Congo
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
languages comprise in their most elaborated forms both derivational and inflectional characteristics. While noun class assignment can be seen as derivational means on the noun form, all agreement patterns in syntactic contexts fall into the realm of inflection. The differentiation between those two sets of relations is important in order to understand and analyze the features and functions correctly. Although this has already been noted in Corbett (), it still happens quite frequently that agreement classes are treated together with morphological classes of nouns under the same cover term “noun class system”. In many (not only) African languages, agreement class and morphological class of the noun may behave differently so that they need to be kept apart neatly in order to analyze synchronic structure appropriately (Van der Velde and Idiatov ).
. H
.................................................................................................................................. Diachronic viewpoints on morphology comprise the theory of grammaticalization, and— earlier in the development of African linguistics—the reconstruction of morphological systems. Given the fact that historical linguistics was the main occupation of Africanist linguists from the late nineteenth far into the twentieth century, it is not surprising that comparison and reconstruction of morphological systems play a decisive role in the establishment of language families. While the establishing of phonological correspondences between related languages in addition to lexical comparison dominated the early historical work on African languages (Dimmendaal : –) the establishment of morphological soundmeaning correspondences for the purpose of reconstruction of proto-languages became more important from the late s. This is merely due to a sort of “qualitative shift” leading away from mass comparison and lexicostatistics to more fine-grained analysis of morphological systems. Often, the argument was that morphology is less prone to contactinduced change and therefore more reliable for reconstruction (Newman : ). This seems to be especially true of low-level groupings of language families where reconstructions of morphological systems add substantially to the recognition of genetic relationship. A case in point is the work of Manessy, who, in his reconstructions of the Gurunsi () and Oti-Volta () subgroups and the resulting Proto-Central Gur family (), relied heavily on comparisons of noun class suffixes. As most agreement systems in many languages of this family are defective, he mainly used noun class suffixes and the still mostly intact series of independent class pronouns for the establishment of the groups. He always kept an eye on the meanings of the noun forms belonging to a given class, so that his reconstructions also gave some hints as to the semantic content of the noun classes (Manessy : –). In this way he established a system of sixteen noun class suffixes that were organized in seven SG/PL gender pairs plus five transnumeral classes, all with some hints as to the inherited semantic content of the noun forms involved (Manessy : –). Manessy’s reconstructions have remained valid to this day, which is also true for the derived internal classification of the Gur/Voltaic family. However, the application of Manessy’s morphological and lexical criteria led Kleinewillinghöfer () to add some languages of the Adamawa family (spoken some , km farther east in eastern Nigeria and northern Cameroon) to this Central Gur family, thus forming what is nowadays called (Central) Adamawa-Gur (see also Miehe et al. : –).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
In other parts of the grammar and/or for higher level groupings, morphological reconstruction may be hampered by several constraints. Hence, an attempt to reconstruct verb extensions for the whole Niger-Congo family came up with some extremely vague proto-forms concerning both form and meaning (Voeltz ). The problems of such enterprises are obvious. Firstly, it is extremely difficult to establish corresponding series of elements that are predominantly suffixed to lexical roots. Such elements easily undergo all sorts of reduction and reorganization processes that may alter their shape completely. Secondly, on a synchronic level the derivational meanings are often already quite hard to pinpoint, and precise functions hard to determine accordingly. So, comparing unclear phonetic substances with diffuse functional loads across supposedly very remotely—if at all—related languages becomes extremely hazardous (Beyer : ). The above-mentioned fact that grammatical elements undergo phonetic reduction processes and also functional splits is the key assumption in the theory of grammaticalization. In their seminal work on Grammaticalization and Reanalysis in African Languages (), Heine and Reh demonstrate how processes of linguistic erosion and reanalysis of other elements shape languages over time. It soon becomes clear that such processes are ubiquitous in languages and that some of the processes observed follow universal tendencies; for example, verbs of directed motion like ‘come to’ and ‘go to’ develop into future markers (Heine and Kuteva : , ). Besides the huge contribution to general theories of language change based on African empirical findings, grammaticalization theory also threw new light on the old question of how noun classes and concomitant agreement systems in Niger-Congo may have developed. The most convincing answer to this question evokes grammaticalization paths that start from lexical class terms (like English ‘X-berry’) and/or measure terms (‘pound of X’) to pass to lexico-grammatical classifier systems (e.g. body part nouns). From there, they may further develop fully grammatical systems with agreement on syntactic targets, NP-internally as well as NP-externally, to finally reach the “classic” overtly marked noun class cum agreement systems (Kießling : ). Such grammaticalization paths can be found in various stages all over the Niger-Congo-speaking area and reflect both the cognitive models of noun ordering and the underlying need for syntactic relation marking, often building on agreement structures known from genitival constructions (Kießling : –).
R Beyer, K. (). ‘Verbderivation und Wurzelstruktur im Pana: Versuch einer integrativen Beschreibung’, in T. Schumann, M. Reh, R. Kießling, and L. Gerhardt (eds.), Aktuelle Forschungen zu afrikanischen Sprachen. Sprachwissenschaftliche Beiträge zum . Afrikanistentag, Hamburg, .–. Oktober . Cologne: Rüdiger Köppe, –. Beyer, K. (). La langue pana: Description linguistique, lexique, textes. Cologne: Rüdiger Köppe. Bickel, B., and Nichols, J. (). ‘Fusion of selected inflectional formatives’, in M. S. Dryer and M. Haspelmath (eds.), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology, http://wals.info/chapter/ accessed Apr. . Carlson, R. (). A Grammar of Supyire. Berlin and New York: Mouton de Gruyter. Childs, G. T. (). An Introduction to African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Comrie, B. (). Language Universals and Linguistic Typology: Syntax and morphology. Chicago, IL: University of Chicago Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Corbett, G. G. (). Gender. Cambridge: Cambridge University Press. Creissels, D. (). Perçu sur les structures phonologiques des langues négro-africaines. Grenoble: Éditions littéraires et linguistiques de l’Université de Grenoble. Delplanque, A. (). ‘Dérivation et production des adjectives en dagara’, Gur Papers/Cahiers Voltaïques : –. Diallo, A. (). Lehrbuch des Pular. Cologne: Rüdiger Köppe. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Doke, M. C., and Mofokeng, M. (). Textbook of Southern Sotho Grammar. London: Longmans, Green & Co. Dumestre, G. (). Le bambara du Mali: Essais de description linguistique. Lille: Atelier national de reproduction des thèses. Dumestre, G. (). Grammaire fondamentale du bambara. Paris: Karthala. Dzablu-Kumah, S. W. (). Basic Ewe: For foreign students. Cologne: Institut für Afrikanistik. Elders, S. (). Grammaire kulango (parler de Bouna, Côte d’Ivoire). Cologne: Rüdiger Köppe. Frajzyngier, Z. (). A Grammar of Hdi. Berlin and New York: Mouton de Gruyter. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, –. Haacke, W. H. G. (). ‘Namibian Khoekhoe’, in R. Vossen (ed.), –. Hayward, R. J. (). ‘Qafar (West Cushitic)’, in A. Spencer and A. M. Zwicky (eds.), –. Heine, B., and König, C. (). ‘Northern Khoesan: !Xun’, in R. Vossen (ed.), –. Heine, B., and Kuteva, T. (). World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Heine, B., and Reh, M. (). Grammaticalization and Reanalysis in African Languages. Hamburg: Helmut Buske. Kaboré, R. (). Essai d’analyse de la langue mὺʋré (parler de Wàoǵdgò). Thèse d’ètat, Université de Paris . Katamba, F. (). Morphology. Basingstoke: Palgrave Macmillan. Kießling, R. (). ‘On the origin of Niger-Congo nominal classification’, in R. Kikusawa and L. A. Reid (eds.), Historical Linguistics : Selected papers from the th International Conference on Historical Linguistics, Osaka, – July , –. Kleinewillinghöfer, U. (). ‘Relationship between Adamawa and Gur languages: the case of Waja and Tula’, Gur Papers/Cahiers Voltaïques : –. Manessy, G. (). Les langues gurunsi: Essai d’application de la méthode comparative à un groupe de langues voltaïques. vols. Paris: SELAF. Manessy, G. (). Les langues Oti-Volta: Classification généalogique d’un groupe de langues voltaïques. Paris: SELAF. Manessy, G. (). Contribution à la classification généalogique des langues voltaïques. Paris: SELAF. Miehe, G., Reineke, B., and Winkelmann, K. (). Noun Class Systems in Gur Languages. Volume II: North Central Languages. Cologne: Rüdiger Köppe. Naden, A. J. (). ‘Gur’, in J. T. Bendor-Samuel (ed.), The Niger-Congo Languages: A classification and description of Africa’s largest language family. Lanham, MD, New York, and London: University Press of America, –. Naden, A. J. (). ‘Verb to nominal derivation in Mampruli and friends’, Gur Papers/Cahiers Voltaïques : –. Newman, P. (). The Kanakuru Language. Leeds: Institute of Modern English Language Studies, University of Leeds. Newman, P. (). The Hausa Language: An encyclopedic reference grammar. New Haven, CT: Yale University Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Nikiema, R. (). ‘Différences de comportement et rapports entre consonne finale de radical CVC et consonne initiale de suffixe en moore’, Studies in African Linguistics : –. Reineke, B. (). ‘Twi (Kwa)’, in G. E. Booij, C. Lehmann, J. Mugdan, and S. Skopeteas (eds.), Morphologie: Ein internationales Handbuch zur Flexion und Wortbildung. Berlin and New York: Mouton de Gruyter, –. Sapir, J. D. (). A Grammar of Diola-Fogny. Ibadan: Cambridge University Press. Schadeberg, T. C. (). ‘Historical linguistics’, in D. Nurse and G. Philippson (eds.), The Bantu Languages. London and New York: Routledge, –. Spencer, A., and Zwicky, A. M. (eds.) (). Handbook of Morphology. Oxford: Blackwell. Tröbs, H. (). Sprachtypologie, TAM-Systeme und historische Syntax im Manding (West-Mande). Cologne: Rüdiger Köppe. Trudgill, P. (). Sociolinguistic Typology: Social determinants of linguistic complexity. Oxford: Oxford Linguistics. Van der Velde, M., and Idiatov, D. (). ‘Morphological classes and gender in əna-Yungur’, in S. Kaji (ed.), Proceedings of the th World Congress of African Linguistics (WOCAL). Tokyo: Research Institute for Languages and Cultures of Asia and Africa, Tokyo University of Foreign Studies, –. Velupillai, V. (). An Introduction to Linguistic Typology. Amsterdam and Philadelphia, PA: John Benjamins. Voeltz, E. (). Proto-Niger-Congo Verb Extensions. Ph.D. thesis. Los Angeles: University of California. Vossen, R. (ed.) (). The Khoesan Languages. London and New York: Routledge. Welmers, W. E. (). African Language Structures. Berkeley and Los Angeles, CA: University of California Press. Westermann, D. (). A Study of the Ewe Language. London: Oxford University Press.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. I this chapter I present an overview of the basic—and what I consider the most intriguing—syntactic properties of the languages spoken on the African continent. I discuss the major syntactic word categories and general aspects of word order typology, but I also address specific topics that have attracted considerable attention in the fields of African linguistics and theoretical syntax, including topic and focus constructions, wh-questions, serial verbs, and the passive. In this review, I highlight those syntactic phenomena that are mainly or exclusively found in African languages, such as logophoricity, or the so-called “verb medial” (S–Aux–O–V–X) constituent order. Occasionally, I also mention prominent generative analyses of the constructions I review.1
. T :
.................................................................................................................................. The “basic” order of constituents in a language is typically defined by the position of subject (S), verb (V), and object (O) in declarative, affirmative, active main clauses which are morphologically and pragmatically unmarked.2 According to Heine (, ), the 1 In this review, I mainly provide examples from the literature. I have made no attempt to unify the phonetic notations, and in general have adopted each language example from the source in its original form. By and large, this also applies to the glosses, although here I have tried to some extent to harmonize the abbreviations. I have also on occasion added or modified glosses when I was sufficiently familiar with the language or when I was able to find the necessary information in another source. In a few cases, I have taken the liberty of simplifying the glosses when a relevant function or distinction was not crucial for the particular phenomenon I wanted to illustrate by an example. In some of the Bantu examples, the basic noun class of the nouns and noun class agreement on grammatical elements are marked through numbers in brackets, according to Meinhof ’s () numbering system of Proto-Bantu. 2 Notice that the notion of “basic” constituent order must not be confused with that of “underlying” constituent order in generative syntactic theories, which is determined by a (possibly universal) initial syntactic configuration from which surface word orders are derived by movement.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
proportion of languages with S–V–O constituent order is much higher in Africa than globally; it is the basic order of approximately % of African languages (Heine : ). S–V–O languages are common in all four phyla. The majority of the Niger-Congo languages are S–V–O; in fact, this constituent order is almost without exception in the Atlantic and Bantu branches (Heine ; Watters ). S–V–O languages in the NiloSaharan phylum include some Central Sudanic and Western Nilotic languages (Heine ; Vossen ; Creider ; Bender ). Most Chadic languages of Afro-Asiatic (Newman ; Schuh ) and the Northern Khoisan languages (Güldemann and Vossen ; Heine and König ) are S–V–O: () Takko wii Demmba mi yiyii ɓe. Takko told Demmba I saw them ‘Takko told Demmba that I saw them.’
[Fula; Atlantic; Niger-Congo]
() zân sayá wà mátātā rígā. : buy wife: dress ‘I’ll buy a dress for my wife.’
[Hausa; Chadic; Afro-Asiatic]
(Potsdam : )
(Green : )
The relative order of multiple objects in S–V–O languages is mostly determined by semantic properties such as animacy or definiteness; indirect objects typically precede direct objects. Note that in African languages with valency-changing morphology, the number of NP-objects3 is not limited to two. For example, in the Bantu language Chaga, up to five object NPs are licensed when the verb combines with an applicative marker (Moshi ). As noted by Heine (), S–O–V constituent order is less common among African languages than worldwide. There are only a few S–O–V languages in Niger-Congo, e.g. the Ijoid and Dogon languages, and the Kordofanian language Tegem (Williamson ; Heine ; Williamson and Blench ). Kanuri, Maba, Kunama, and the Nubian languages are prominent examples of S–O–V languages in the Nilo-Saharan phylum (Bender ). The Ethio-Semitic, Omotic, and Cushitic languages of Afro-Asiatic are verb-final (Heine ; Comrie ; Watters ; Tosco ), and among the Khoisan languages, S–O–V is the basic constituent order in Khoe (Central Khoisan), in Kwadi, and in Sandawe (Rust , Dalgish ; Güldemann and Vossen ; den Besten ; Vossen : chapter ):4 () avá-nzə-yè shí-rò kúŋənà cîn. father-- he- money: give: ‘His father gives him money.’
[Kanuri; Saharan; Nilo-Saharan] (Lukas : ; glosses added)
3 I use the abbreviation “NP” as a cover term for all noun phrases, including those analyzed as DPs (determiner phrases) in generative theories. 4 Afrikaans, a West Germanic verb-second language spoken in South Africa and Namibia, also has a basic S-O-V constituent order (Robbers ; den Besten ). Note incidentally that there are also African languages that have been classified as verb-second languages. I briefly discuss this fact in section ..
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() Johanneb ge ti ǁnaoba goro gurin ei-ǃâ ge mû. John . uncle five years ago see [Khoekhoegowab; Central Khoisan; Khoisan] ‘John saw my uncle five years ago.’ (den Besten : ) “Rigid” verb-final languages, in which all types of objects precede the verb, are rare in Africa (Creissels ). For example, of the approximately African languages surveyed by Heine (), only % have the word order S–AP–V, compared to % with S–O–V constituent order (Heine : ). In some West African languages, verbs systematically follow their objects (and auxiliaries), but precede oblique arguments and adjuncts, giving rise to S–Aux–O–V–X constituent order (Gensler , ; Creissels ; Nikitina , ). In Niger-Congo, this constituent order is the rule in Mande and in the Senufo languages of the Gur family (Carlson ; Creissels ), and it is also attested in Songhay, a Nilo-Saharan dialect cluster closely related to Mande (Nicolaï ): () sékù jé mădù délíí wáríí #lá kúnún. [Bambara; Mande; Niger-Congo] Sékou Madou ask money yesterday ‘Sekou asked Madou for money yesterday.’ (Creissels : ex. (b)) A related phenomenon is the S–V–O ~ S–O–V constituent order alternation that is found in languages from all phyla. In Niger-Congo, the Kwa, Kru, and some Gur languages (Koopman ; Marchese ; Fabb ; Manfredi ; Aboh ), as well as the Atlantic language Kisi (Childs ), are best known for showing this variable word order, but the V–O ~ O–V alternation is also found in the Southern Cushitic branch of AfroAsiatic (Heine ; Güldemann ), in some Central Sudanic and Western Nilotic languages of the Nilo-Saharan phylum (Heine ; Creider ), and in Northern Khoisan (Güldemann ; König a). The choice between the V–O- and the O–Vvariant is conditioned by polarity, finiteness, or temporal-aspectual properties of the sentence. Therefore, the O–V-variant is often (but not always) attested when an auxiliary is present, in which case S–V–O alternates with S–Aux–O–V: () a. ɔ pī sāyὲ. cook: meat ‘He cooked meat.’
(S–V–O)
b. ɔ́ nà sāyὲ pī. (S–Aux–O–V) meat cook ‘He has cooked meat.’
[Dewoin; Kru; Niger-Congo]
(Marchese : )
In Givón (), the O–V-variant of the word order alternation illustrated by () is interpreted as a reflex of an earlier stage of Proto-Niger-Congo, for which a basic O–Vsyntax is postulated. Claudi () argues for a historical development in the opposite direction, suggesting that the O–V-pattern is the result of a grammaticalization process in which the first transitive verb of a serial verb construction with S–V–O–V order was reanalyzed as an auxiliary. Generative theories account for the alternation in terms of movement transformations. While Koopman () argues that the V–O-order is derived
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
from the O–V-order by verb movement to a higher head position to the left of the object, others take the V–O-order as basic and assume that the O–V-order is derived via leftward movement of the direct object (see, e.g., Fabb , Manfredi , Kandybowicz and Baker , and Aboh for different versions of this idea). Güldemann () analyzes the alternation in terms of information structure, proposing that the preverbal position of the object is correlated with a decrease in pragmatic salience. In some languages with the V–O ~ O–V alternation, oblique arguments may still follow the verb, even when the direct object appears preverbally. This gives rise to the order S–Aux–O–V–X, which was noted as being the only possible order in Mande, Senufo, and Songhay. Since most of the languages with this word order are spoken in the same geographic region (in the so-called “Macro-Sudan belt”; Güldemann ), they are sometimes grouped together and treated on a par. (For example, they are classified as “verb-medial” languages by Gensler and , or as “Type B” languages in Heine .) Although it is controversial whether the S–Aux–O–V–X constituent order is actually a uniform phenomenon (see Creissels and Good for discussion), it is generally accepted that it is a characteristic property of a subset of African languages which is very rarely (if ever) found outside Africa. The proportion of African languages with basic V–S–O constituent order seems to correspond roughly to the relative number of V–S–O languages worldwide. The least controversial examples of African V–S–O languages are found in Nilo-Saharan, namely in the Kuliak, Surmic, and the Eastern and Southern Nilotic branches of Eastern Sudanic (Dimmendaal ; Vossen ; Creider ; Bender ; König b). Verb-initial languages in Afro-Asiatic are Arabic, Ancient Egyptian, and (more controversially) Berber, as well as a few Chadic languages (Comrie ; Creissels ; Schuh ). Hadza, a Tanzanian language sometimes grouped with the Khoisan languages, has been classified as V–S–O (Heine ; Watters ), but there seem to be no Niger-Congo languages with basic V–S–O constituent order: () é-dúŋ εn-kεráɩ́ en-kíné. -cut .-child: .-goat: ‘The child will cut the goat.’
[Maasai; Eastern Nilotic; Nilo-Saharan] (Payne : )
Finally, the universally rare basic O–V–S constituent order is found in transitive sentences in the Western Nilotic ergative language Päri (Andersen ), and Malagasy, an Austronesian language spoken on the island of Madagascar, has V–O–S constituent order (Keenan ; Pearson ).
. T
.................................................................................................................................. In African languages, NPs whose nominal head precedes demonstrative, numeral, and adjectival modifiers are more frequent than noun-final NPs, and the noun-initial order is more common in Africa than elsewhere in the world (Heine , ; Creissels ):
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() (Kɔ́kú xɔ̀) távò ɖàxó xóxó àtɔ̀n éhè lɔ́ lέ. (Koku buy:) table big old [þdef] ‘(Koku bought) these specific three big old tables.’ [Gungbe; Kwa; Niger-Congo] (Aboh : ) () hikwa-’ée’ ku-dá kooʾan lówa hhoo’ [Iraqw; Cushitic; Afro-Asiatic] cattle-‘. ./- five very nice ‘those five very nice cows of mine’ (Mous : ) Globally, the constituent order N–Adj–Num–Dem illustrated by the Gungbe example in () is the most frequent order in noun-initial NPs (Greenberg ; Hawkins ; Cinque ). The Iraqw example in (), which shows the order N–Dem–Num–Adj(P), is interesting because Iraqw, despite being noun-initial, is a verb-final language. According to Creissels (), the tendency of S–O–V languages to exhibit head-final order inside the NP is actually quite weak in African languages, and one finds many languages with a rigid noun-initial constituent order among the verb-final languages. Most East Cushitic languages are like Iraqw in this respect (e.g. Somali, Gawwada, and Boni; see Tosco , ), as well as Kanuri, Tubu, Nubian, and Fur, which belong to the Nilo-Saharan phylum (Heine ). A noun-initial NP order which is considered to be quite rare universally is the order N–Num–Adj–Dem (Rijkhoff calls this constituent order “non-iconic”). Interestingly, however, according to Heine (), this order is attested in various African languages, namely in Gabra (Cushitic; Afro-Asiatic), in Logoli (Bantu; Niger-Congo), and in Luo (Western Nilotic; Nilo-Saharan). According to Cinque (), the NP order N–Adj–Dem– Num is also found in only a few languages worldwide, but it is attested in the Kru language Godié (Marchese : ). NPs with noun-final constituent order are found in the Ethio-Semitic (Hayward ; Kramer ) and some Cushitic languages of Afro-Asiatic (Schneider-Blum ), in the Central Khoisan languages (Heine ; Hagman ; Güldemann and Vossen ; Vossen : chapter ), and in the Ijoid languages of Niger-Congo:5 () ɨnnäñña-n sost tɨllɨk’ bet-otʃtʃ (ayyä-hu). [Amharic; Ethio-Semitic; Afro-Asiatic] those- three big house- see- ‘(I saw) those three big houses.’ (Ruth Kramer, p.c.) ()
nùmà mààmà gògòrì those two red ‘those two red axes’
ŋgì axe
[Defaka; Ijoid; Niger-Congo] (William Bennett, p.c.)
Notice that the order of the constituents in () and () is Dem–Num–Adj–N, which has been argued to be the only possible order in noun-final NPs (cf. Greenberg’s Universal ; see Hawkins ; Cinque ).
5
I am indebted to Ruth Kramer and Will Bennett for providing me with the Amharic and Defaka data.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
NPs may also be “noun-medial”, with N selecting some modifiers to its right and some to its left. Some of the N-medial orders that are uncommon universally do again occur in African languages. For example, according to Cinque (: –), the orders Dem–N–Num–Adj and Dem–Adj–N–Num are found in very few languages worldwide. However, the former word order is attested in Maasai (Koopman ), and the latter in Zande (Rijkhoff ): ()
kù-n-dâ --that
mέsa-i table-.:
àré ..two:
‘those two nice tables’ ()
sìdân nice:: [Maasai; Eastern Nilotic; Nilo-Saharan] (Koopman : )
gi rarai a-mangu biata-re [Zande, Adamawa-Ubangi, Niger-Congo] heavy -box three-/here ‘these three heavy boxes’ (Rijkhoff : ; attributed to Christopher Leone Daffalla, p.c.)
Finally, Creissels (: ) notes that in some African languages, adjectives are the only modifiers that precede the noun, contrary to the generalization that, cross-linguistically, adjectives tend to follow the noun. The order Adj–N–Num–Dem, for example, has been noted to exist in Gude (Afro-Asiatic) and in the Ngbandi-based creole Sango (Niger-Congo) (see Cinque : , n. ). Possessive modifiers commonly appear in post-nominal position in African languages (cf. Heine ), although there are some West African languages in which possessors are the only modifiers that precede the noun (see Marchese ; Claudi ; Creissel ). A well-studied type of complex NP that is used to express possession in many African languages is the associative construction (Welmers ), in which the possessor and the possessee are linked by means of a grammatical marker. In the Bantu languages, the associative marker -a- shows agreement with the noun class of the preceding head noun, the possessee (see Güldemann ; Carstens ): () a. kikombe cha Mariamu cup: : Maria ‘Maria’s cup’ b. vikombe vya Mariamu cup: : Maria ‘Maria’s cups’
[Swahili; Bantu; Niger-Congo]
(Rainer Vossen, p.c.)
Associative constructions are also found in the Central Khoisan languages. When the possessor precedes its governing noun, as in (a), a form of the associative marker di appears between the possessor and the possessee, which optionally agrees with the following possessee. However, when the possessee is placed before the head noun, as in (b), agreement is obligatory (Güldemann and Vossen ):
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
a. hàúgù-m dì(-s) ǀ’óán-sà dog-. (-.) bone-. ‘dog’s bone’ b. ǀ’óán-sà hàúgù-m dì-sà bone-. dog-. -. ‘dog’s bone’
[Naro; Central Khoisan; Khoisan]
(Güldemann and Vossen : )
The associative construction is typically not restricted to possessor relations and can express a much wider range of semantic associations, such as time, function, and quantity (Welmers ; Güldemann ). Welmers () notes that, in this respect, the associative in Niger-Congo is similar to the so-called “construct state” (CS) construction which is used to express similar semantic relations (including possession) in the Semitic languages of Afro-Asiatic (Borer ; Benmamoun ; Ouhalla ): () (*l-)kitaab-u ṭ-ṭaalib-i (the-)book- the-student- ‘the student’s book’
[Standard Arabic; Semitic; Afro-Asiatic] (Benmamoun : )
A CS consists of the nominal head and a following possessor NP. The two members of a CS construction must be adjacent and form a prosodic unit. As () shows, only the rightmost noun can be marked for definiteness and combine with a determiner. According to Creissels (: ), definite articles are fairly common in African languages. As the examples in () and () show, they are often realized as clitics or affixes and attach to the first or last word of the NP with which they combine.
. P
.................................................................................................................................. Pronouns in African languages may occur as independent word forms (so-called strong pronouns), which usually have the same distribution as full NPs (DPs). Weak pronouns, however, which are typically realized as clitics or affixes, are more common. In many languages, the weak pronouns can co-occur with coreferential full NPs. They are, therefore, often analyzed as subject or object agreement markers, and may indeed be functionally ambiguous in some African languages (see Bresnan and Mchombo for Bantu). A peculiar system of pronominal reference that exists in many African languages is logophoricity (see, e.g., Hagège ; Clements ; Hyman and Comrie ; Culy ; Curnow ; and many others). Logophoric pronouns are used to refer to the person whose speech or thoughts are reported. They, therefore, typically express obligatory coreference between the subject or object argument of an embedded clause and the subject of the matrix clause. In (a), for example, the use of the logophoric pronoun in the embedded subject position indicates that the person whose statement is reported is also the agent of the reported event. The use of a regular, unmarked pronoun in the same context indicates non-coreferentiality, (b):
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
a. wu sat nɘ́ ɗi nas an. . say that .. beat ‘He1 said that he1 beat me.’ b. wu sat nɘ́ wu nas an. . say that . beat ‘He1 said that he2 beat me.’
[Mupun; Chadic; Afro-Asiatic]
(Frajzyngier : )
Logophoric pronouns are found in Nilo-Saharan, Niger-Congo, and Afro-Asiatic languages spoken in the Macro-Sudan belt (Güldemann , ). It has been noted that logophoricity is an exclusively African phenomenon and that genuine logophoric pronouns do not exist outside Africa (Heine ).
. A
.................................................................................................................................. According to Watters (: ), African languages tend to have fewer adpositions than European languages, because semantic relations that are typically expressed by prepositions in the latter are often expressed by other grammatical means in the former. For example, while locatives are PPs in a language such as English, they are realized as NPs with locative morphology in most Bantu languages. Interestingly, however, in some Bantu languages, locative morphology seems to have undergone a process of degrammaticalization, and locative prefixes have become reanalyzed synchronically as prepositions (Marten ). In many other African languages, adpositions are derived from nouns referring to body parts or from verbs with locative or existential meanings (Heine ; Ameka ; Nikitina ). Therefore, the synchronic classification of an element as belonging to the word category “adposition” is not always straightforward. In the Northern Khoisan language Juǀ’hoan, a postpositional locative phrase that follows the theme NP in a ditransitive construction is obligatorily preceded by the element kò or kē (depending on the dialect), giving rise to a construction that resembles a circumpositional phrase, (a). Interestingly, however, the order of locative and theme can also be inverted in Juǀ’hoan, (b): ()
a. Uto dchuun-a ǀKaece ko n!ama n!ang. car hit- ǀKaece road in ‘A car hit ǀKaece in the road.’ [Juǀ’hoan (!Xun); Northern Khoisan; Khoisan] b. Uto dchuun-a n!ama n!ang ko ǀKaece. car hit- road in ǀKaece ‘A car hit ǀKaece in the road.’ (Collins : )
The marker kò/kē does not only combine with locatives and themes, but also with instrumentals and benefactives. Therefore, König (a) analyzes kò/kē as a semantically
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
empty preposition. Güldemann and Vossen () label this element a “multipurpose oblique marker”, Dickens () refers to it as a “transitive particle”, Heine and König () propose the term “transitive preposition”, and Baker and Collins () simply call it a “linker”. Baker and Collins show that the linker element is also found in the Benue-Congo languages Yoruba and Kinande (Bantu). In the latter language, the linker agrees in noun class with the preceding NP: ()
a. Mo-n-a-hir-ire ---put-
okugulu leg:
k’ :
omo-kihuna. :-hole: [Kinande; Bantu; Niger-Congo]
‘I put the leg in the hole.’ b. Mo-n-a-hir-ire omo-kihuna ---put- :-hole: ‘I put the leg in the hole.’
m’ okugulu. : leg: (Baker and Collins : )
Because of its agreement properties, Baker and Collins () and Collins () do not analyze the linker as a preposition, but as the head of a functional projection within the extended VP whose specifier can host different arguments of the verb.
. A
.................................................................................................................................. Adjectives are generally rare in African languages, particularly in the Niger-Congo phylum, where many languages only have a very small number of genuine, non-derived adjectives (Creissels : ). This is partly because many predicates which are expressed as adjectives in the European languages are stative verbs in African languages (Watters : ). Attributively used adjectives that modify nouns are often realized as subject relative clauses. The overwhelming majority (%) of Heine’s sample of c. African languages exhibits the AdjP-internal order Adj–Adv (Heine : ). Examples of languages in which adverbs precede adjectives are found in the Afro-Asiatic phylum, e.g. in the Ethio-Semitic language Amharic (Kramer ) and in the Cushitic languages Iraqw and Qafar (Hayward ; Mous ): () nabám xeera very tall/long ‘very tall/long’
[Qafar; Cushitic; Afro-Asiatic] (Hayward : )
In (), degree modification is expressed by an element belonging to the syntactic category Adv, but similar adverbial functions may also be encoded by auxiliary verbs in some African languages (Anderson ).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. R
.................................................................................................................................. In most African languages, relative clauses follow their head nouns: ()
ŋʷon ga ma m-ak-iny na boy -see-o. ‘the boy who I saw’
[Bagirmi; Central Sudanic; Nilo-Saharan] (Dryer : )
The object relative clause in () is introduced by a segmental relative marker and includes a resumptive pronominal clitic which is coreferential with the head noun. One also finds African languages in which relative clauses are marked by verbal inflection or tone (e.g. Bantu languages such as Bemba; Kula and Cheng ), or in which relative clauses are marked through constituent order (which according to Vossen is the case in the Eastern Nilotic language Oxoryok). Prenominal relative clauses are found in Amharic (Hudson ; Kramer ; see Kayne and Ouhalla for generative analyses), in Khoe (den Besten ), and in Ijo (Givón ; Jenewari ): () Bomá wá fẹ inji mẹ sọ́árị. Boma we buy: fish the cook: ‘Boma is cooking the fish that we bought.’
[Ijo; Ijoid, Niger-Congo] (Jenewari : )
Internally-headed relative clauses have been noted to exist in the Mande language Bambara (see Bird ; Comrie ; Watters ) and in the Gur languages Mooré (Tellier ) and Buli (Hiraiwa ): () n̩
dε̌ ate
Àtìm Atim
lī
dà bought
mánɡò-kūːy mango-
‘I ate a mango that Atim bought yesterday.’
dìe͂ lá. yesterday [Buli; Gur; Niger-Congo] (Hiraiwa : )
Note that the constituent order in Mooré and Buli is S–V–O. Therefore, the existence of internally headed relative clauses in these Gur languages contradicts the typological generalization that, universally, internally headed relatives are only found in O–V languages (see Hiraiwa for discussion).
. M , ,
.................................................................................................................................. Although coordination and subordination are often expressed simply via juxtaposition in African languages (Creissels ; Watters ), the use of complementizers and
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
conjunctions is not uncommon. In many African languages, complementizers are grammaticalized forms of a verb of “saying”: nakàn. (cf. gàn ‘say’) meat [Nupe; Benue-Congo; Niger-Congo] ‘Musa knows that the chief cooked the meat.’ (Kandybowicz : )
() Musa Musa
kpe know
gànán C
etsu chief
du cook
Many African languages coordinate sentences by means of the so-called consecutive construction, which is used to express a succession of events that chronologically follow each other. The first verb determines the tense, while the following verb (or series of verbs) is formally marked as “sequential”, “subsecutive”, or “narrative” (Hyman ; Carlson ; Creissels ; Watters ): () ὲ-à-ɪ̀mʊ̀j-ɪ̀ --eat-
εkàsukowùt old.man:
k-iyar-a-kin-ì .̥ -belch-- [Turkana; Eastern Nilotic; Nilo-Saharan] ‘The old man ate and then belched.’ (Dimmendaal : ; Carlson : )
Consecutive constructions are sometimes compared to the converbial constructions that are characteristic of the Ethio-Semitic and Cushitic languages of Afro-Asiatic, and that also exist in some Nilo-Saharan languages (Amha and Dimmendaal ). Converbs are similar to subsecutive verbs in that they may be specified for PNG-agreement and aspect, but not for tense, which is only expressed on the main verb. Another type of complex clause with multiple verbs which is particularly common in West African Niger-Congo languages is the serial verb construction (SVC). A SVC is a succession of two or more verbs (plus their complements, if they are selected) within the same clause (there is no overt conjunction or complementizer). All verbs share a grammatical subject and have the same tense/aspect specification, and the different actions expressed by the verbs in a SVC are conceptualized as a single event (see, e.g., Welmers ; Bamgbos: e ; Ekundayo and Akinnaso ; Baker ; Collins ; Carstens ; Aboh , among many others): () àhì hû ólo chu. we take load put.on.head ‘We carried the load.’
[Igede; Benue-Congo; Niger-Congo] (Bamgbos: e : )
Another characteristic feature of SVCs illustrated by the example in () is argument sharing: when the first verb in the sequence is transitive, its internal argument is typically also an argument of the second verb. Baker () argues that the “shared” argument in SVCs is theta-marked by both verbs. Collins () proposes instead that the second verb projects its own VP and selects an unpronounced pronominal argument which is controlled by the internal argument of the first verb. In contrast, Aboh () analyzes the first
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
verb in SVCs as a realization of a functional “light” verb and argues that the object NP is in fact not a shared argument of both verbs, but only theta-marked by the second. SVCs are also found in Nilo-Saharan (Mekoulnodji et al. ), in the Chadic languages of Afro-Asiatic (Frajzyngier ), and in the Khoisan languages (Sebba ; Collins ; Dickens ; Kilian-Hatz ; König a): () Mi m a nǀoa ’m !ha. [Juǀ’hoan (!Xun); Northern Khoisan; Khoisan] cook eat meat ‘I will cook and eat meat (repeatedly).’ (Collins : ) As () shows, the two verbs in SVCs in Northern Khoisan must be adjacent; a shared object will always follow the verb complex, which is never interrupted by any grammatical material. Therefore, Collins (, ) terms the verb complexes in Northern Khoisan SVCs “verbal compounds”. He suggests that constructions such as () have the same underlying syntax as SVCs in Niger-Congo, but that verbal compounds in Northern Khoisan are derived by movement of the lower verb past the object NP to a position adjacent to the higher verb. According to Anderson (), the SVC is one of the three major source constructions from which monoclausal constructions with auxiliary verbs have developed in African languages (the other two being embedded/nominalized and clause-chained structures). This may explain why the “doubled inflection” pattern, in which inflectional features such as subject agreement are morphologically encoded on both the auxiliary and the lexical verb, is relatively common in auxiliary-verb constructions in African languages, particularly in Bantu (see Anderson ).
. T
.................................................................................................................................. African languages use a variety of syntactic means to express aspects of information structure (for recent overviews and individual analyses, see, e.g., Bearth ; Aboh et al. ; Ermisch ; Fiedler and Schwarz ; Güldemann et al. ). The standard way of marking an element as a topic is by means of left or right dislocation. Typically, the fronted or extraposed topic is picked up by a resumptive pronoun or pronominal clitic in the comment clause (see, e.g., Bresnan and Mchombo ). As for focus, the most common strategy of marking something as new information, which probably exists in all African languages (Watters : ), is the cleft construction. In clefts, the focused constituent is introduced by a copula and modified by a relative clause: () nɘhna ina nɘ-ʔabrɘhat zɘ-rɘʔena. : we -Abrahat -saw: ‘It is we who saw Abrahat.’
[Tigrinya; Semitic; Afro-Asiatic] (Gragg : )
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Focus can also be marked by displacement of the focused constituent (focus movement) in many African languages: () mààlóŋ ó có cùùcúúwó ní. rice he sow ‘It is rice that he is sowing.’
[Kisi; Southern Atlantic; Niger-Congo] (Childs : )
In addition, many African languages mark information structure by grammatical elements, such as topic or focus particles or special verbal affixes. For example, Kisi focus constructions such as () include the invariant clause-final focus marker ní (Childs ). In Somali (East Cushitic, Afro-Asiatic), focus markers are obligatory in declarative main clauses and must immediately follow the preverbal focused constituent (Saeed ; Svolacchia et al. ; Lecarme ). In !Xun (Northern Khoisan), a suffixal focus marker is attached to the focused constituent in the left periphery, while the rest of the clause is separated from the focus by means of a topic marker (König a, c). In Hausa (Chadic, Afro-Asiatic), focus constructions are marked by an optional focus-marking copula as well as special morphology on the verb (Tuller ; Wolff ; Newman ; Jaggar ; Green ). In Efik (Benue-Congo), the choice between different tense allomorphs depends on whether or not the verb is included in the focus (Hyman and Watters ). The Central Khoisan SOV-language Khoekhoegowab (Nama) has a sentence-initial position in which both topicalized and focused constituents (including wh-phrases) can occur (Rust ; Westphal ; Hagman ; den Besten ; Haacke ): () Tara-s-a b woman-.- he
ge
ao-b-a ra mû. man-.- see [Khoekhoegowab; Central Khoisan; Khoisan] ‘The woman, he—the man—is seeing.’ (Haacke : )
The focused object in () is followed by a subject clitic (b), which always attaches to the fronted constitutent, while the lexical subject NP follows the declarative particle ge (Haacke ). According to den Besten (), the declarative particle signals the second position of the clause, and he suggests that constructions such as () are to some extent comparable to verb-second constructions in the Germanic languages. A similar pattern is attested in the Western Nilotic language Dinka, where the verb or auxiliary appears in the second position and is typically preceded by a topic or a wh-phrase. Consequently, Dinka has been classified as a verb-second language (Andersen ; Dimmendaal ; Richards and van Urk ). In some African languages, focused constituents appear in clause-final position (Watters : ). This is the case, for example, in the Bantu language Rundi (Ndayiragije ): ()
Yohani a-á-oógeje néezá imiduga. [Rundi; Bantu; Niger-Congo] John: :--wash: well cars: ‘John washed cars well (not trucks).’ (Ndayiragije : )
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
In many other Bantu languages, focused constituents instead appear in the so-called IAV (“immediately after the verb”) position (Watters ; Buell ; Van der Wal ). For example, in the Grassfield Bantu language Aghem, an SVO-language, contrastive focus on the subject is expressed by V–S–O constituent order: () à bέ-#kɔ́. mɔ̀ zı ́ á-f ıń eat friends fufu ‘The friends ate the fufu.’
[Aghem; Bantu; Niger-Congo] (Watters : )
The IAV-position is also involved in focus phenomena in some verb-initial languages. According to Koopman (: ), focused objects can appear between the verb and the subject in the Eastern Nilotic VSO-language Maasai, producing V–O–S constituent order, and Dimmendaal (: ) notes that the same position is available for certain contrastively focused adverbs in Turkana (another verb-initial Eastern Nilotic language). Sands (: –) observes that V–O–S order is possible with focused objects in Hadza (Khoisan), and Tuller () shows that the Chadic VSO-language Podoko (Afro-Asiatic) also licenses focused material in the IAV position. Finally, some African languages use verb copying to mark certain types of focus: ()
lē à lē sẚka. eat we eat rice ‘We are really eating rice.’
() Musa è gí bise gí. Musa eat hen eat ‘Musa is in fact eating a hen.’
[Vata; Kru; Niger-Congo] (Koopman : ) [Nupe; Benue-Congo; Niger-Congo] (Kandybowicz : )
In (), verb focus is expressed by means of a predicate cleft construction in which a focused verb in the left periphery of the clause is doubled by a resumptive verbal copy (Dimmendaal ; Koopman ; Childs ). () is a so-called “emphatic declarative”, which expresses polarity focus by means of verbal repetition, i.e. the occurrence of two non-distinct verbal copies within the same clause (Smith ; Kandybowicz ).
. Q
.................................................................................................................................. In most African languages, the same strategies that are used in focus constructions are also used to form content (constituent, wh-) questions. Wh-clefts are common (see, e.g., Adesola ; Zerbian ; Rose et al. ), and wh-movement constructions and yes/no (polar) questions are often accompanied by focus markers (see, e.g., Aboh ; Torrence and Kandybowicz ). Indirect questions in African languages are typically formed by means of a generic noun plus relative clause (“I ask the thing you want”), although some
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
African languages also express indirect questions by means of wh-pronouns (Watters ). The Berber wh-construction in () illustrates the so-called “anti-agreement effect”, which is associated with subject questions in various African languages: ()
Man tamghart ay yzrin (*t-zra) Mohand? which woman:. saw: (.-see) Mohand ‘Which woman saw Mohand?’
[Berber; Afro-Asiatic] (Ouhalla : )
Ouhalla () observes that in Berber, an extracted third-person wh-subject cannot trigger third-person agreement on the verb. Instead of the verb form tzra, an invariant, nonagreeing form of the verb is used in (). A similar deviation from the default third-person agreement form is observed in subject questions and relative clauses in some Niger-Congo languages. In Kinande and Bemba (Bantu), for example, the regular subject agreement marker of noun class (a-) is systematically replaced by u- in such constructions (Schneider-Zioga ; Henderson ). Although the anti-agreement effect is usually taken to be a reflex of (local) operator movement, it also shows up in wh-in situ constructions in some African languages, for example in the Benue-Congo language Ibibio (Baker ). Another interesting phenomenon that is associated with subject questions in NigerCongo languages such as Vata (Kru) or Yoruba (Benue-Congo) is illustrated by (): () ta ni *(ó) ń ta isu? who sell yams ‘Who is selling yams in the market?’
[Yoruba; Benue-Congo; Niger-Congo] (Carstens : )
As () shows, subject questions in Yoruba require a resumptive pronoun in the subject position when the subject has undergone wh-movement. Object wh-phrases, in contrast, do not license resumption (see Koopman ; Carstens ; Sonaiya ).
. P
.................................................................................................................................. Passive constructions exist in many African languages. The following examples are from Kinyarwanda: () a. Igitabo cy-a-haa-w-e umugóre n’-úmugabo. book: :--give-- woman: by-man: [Kinyarwanda; Bantu; Niger-Congo] ‘The book was given to the woman by the man.’ b. Umugóre y-a-haa-w-e igitabo n’-úmugabo. woman: :--give-- book: by-man: ‘The woman was given the book by the man.’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() Ibíro bíne bi-pim-w-a kilos: four: :-weigh-- Lit.: ‘Four kilos are weighed by this book.’
n’-íki by.this:
gitabo. book: (Kimenyi : –)
In the examples in () and (), the logical object is realized as the subject of the sentence, a transformation that is analyzed in terms of NP-movement in generative theories. The thematic subject argument no longer functions as the grammatical subject, but instead is optionally realized as the complement of a preposition (the so-called “by-phrase”). As () shows, Kinyarwanda is a “symmetrical” language, because either object of a ditransitive verb can be passivized. In contrast, “asymmetrical” Bantu languages such as Chewa allow only one internal argument (usually the indirect object) of a ditransitive verb to be promoted to subject position (see Bresnan and Moshi ). () shows that Kinyarwanda also allows for the passivization of “adverbial” objects, with a non-agentive subject in the by-phrase (see Kimenyi : –). Not all African languages have the type of passive construction illustrated by (). For example, Heine (: ) notes that “genuine” passive constructions do not exist in many West African languages. Instead, some of these languages use the so-called “they”-passive, i.e. an active sentence with a generically used pronominal third-person plural subject. Khwe (Central Khoisan) has a regular passive in which the theme/patient is realized as the subject and the agent as a by-phrase, (a). However, in an alternative passive construction, no agent subject can occur, and the theme/patient remains grammatically marked as the object, (b) (Kilian-Hatz ): () a. ǂ’ṹ à ǂ’ṹ-can-atà ápa-a kà. food eat-- dog- by ‘The food was eaten by the dog.’
[Khwe; Central Khoisan; Khoisan]
b. hèútù-hὲ ὲ tc’ãã́ -́ i-tà. car-. steal-- ‘The car was stolen.’ / ‘One has stolen the car.’
(Kilian-Hatz : , )
Passive constructions similar to (b) are also found elsewhere in Africa. For example, according to Creider (), passives are marked morphologically in the Eastern Nilotic languages, but the thematic object does not receive nominative case and is not promoted to subject position. The opposite seems to be the case in Supyire (Gur; Niger-Congo). Carlson () notes that passives are not marked morphologically in Supyire, but that a passive construction is formed by realizing the thematic object of a transitive verb as the subject. Another type of construction that is analyzed in terms of NP-movement in generative syntactic theories is raising. In standard raising constructions, the subject of an embedded infinitive appears as the subject or object of the main clause (e.g. Johni seems ti to love Mary). However, many African languages allow for so-called “Hyperraising” constructions in which the logical subject of an embedded finite clause is realized as the matrix subject or object. In (), for example, the thematic subject argument of the embedded verb has become the subject of the main clause, where it triggers subject agreement with the verb:
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() I-nzovu z-aa-mεnyeekan-ye kó z-iish-e -elephant: :--be.known- that :-kill- báa-ba-ntu. [Rundi; Bantu; Niger-Congo] :--people ‘Elephants are renowned for having killed those people.’ (Harford (Perez) : ) Hyperraising constructions are found in many Bantu languages (see Harford (Perez) ; Zeller ; Carstens and Diercks ; Halpert ) and also elsewhere in Niger-Congo (see, e.g., Ura ), in Nilo-Saharan (Jake and Odden ; Creider ), and in AfroAsiatic (Sadiqi ).
. C
.................................................................................................................................. In this chapter I have discussed various aspects of the syntax of African languages. I have shown which basic constituent orders are attested and how particular grammatical constructions are realized in different languages. I have drawn attention to construction types which are attested in typologically unrelated languages from different families or phyla, and I have highlighted attributes that are characteristic of African languages but that are rarely, or not at all, found outside Africa. The phenomena discussed in this review illustrate the enormous wealth of interesting data from African languages, many of which raise challenging questions, and sometimes pose non-trivial problems, for existing syntactic theories.
R Aboh, E. O. (). The Morphosyntax of Complement-Head Sequences: Clause structure and word order patterns in Kwa. Oxford: Oxford University Press. Aboh, E. O. (). ‘Focused versus non-focused wh-phrases’, in E. O. Aboh, K. Hartmann, and M. Zimmermann (eds.), –. Aboh, E. O. (). ‘Clause structure and verb series’, Linguistic Inquiry : –. Aboh, E. O., Hartmann, K., and Zimmermann, M. (eds.) (). Focus Strategies: Evidence from African languages. Berlin and New York: Mouton de Gruyter. Adesola, O. (). ‘On the absence of superiority and weak crossover effects in Yoruba’, Linguistic Inquiry : –. Ameka, F. K. (). ‘The linguistic construction of space in Ewe’, Cognitive Linguistics : –. Amha, A., and Dimmendaal, G. J. (). ‘Converbs in an African perspective’, in F. K. Ameka, A. Dench, and N. Evans (eds.), Catching Language: The standing challenge of grammar writing. Berlin and New York: Mouton de Gruyter, –. Andersen, T. (). ‘Ergativity in Pari, a Nilotic OVS language’, Lingua : –. Andersen, T. (). ‘Subject and topic in Dinka’, Studies in Language : –. Anderson, G. D. S. (). ‘Auxiliary verb constructions in the languages of Africa’, Studies in African Linguistics : –. Baker, M. (). ‘Object sharing and projection in serial verb constructions’, Linguistic Inquiry : –. Baker, M. (). ‘On the nature of the antiagreement effect: evidence from Wh-in-situ in Ibibio’, Linguistic Inquiry : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Baker, M., and Collins, C. (). ‘Linkers and the internal structure of vP’, Natural Language and Linguistic Theory : –. Bamgbos: e, A. (). ‘On serial verbs and verbal status’, Journal of West African Languages : –. Bearth, T. (). ‘The contribution of African linguistics towards a general theory of focus: update and critical review’, Journal of African Languages and Linguistics : –. Bender, M. L. (ed.) (). Nilo-Saharan Language Studies. East Lansing, MI: African Studies Center. Bender, M. L. (). ‘Nilo-Saharan’, in B. Heine and D. Nurse (eds.), –. Bender, M. L., Takács, G., and Appleyard, D. L. (eds.) (). Selected Comparative-Historical Afrasian Linguistic Studies, in Memory of Igor M. Diakonoff. Munich: Lincom Europa. Benmamoun, E. (). The Feature Structure of Functional Categories. Oxford: Oxford University Press. Bird, C. (). ‘Relative clauses in Bambara’, Journal of West African Languages : –. Borer, H. (). ‘Deconstructing the construct’, in K. Johnson and I. Roberts (eds.), Beyond Principles and Parameters. Dordrecht, Boston, MA, and London: Kluwer, –. Bresnan, J., and Mchombo, S. A. (). ‘Topic, pronoun, and agreement in Chicheŵa’, Language : –. Bresnan, J., and Moshi, L. (). ‘Object asymmetries in comparative Bantu syntax’, Linguistic Inquiry : –. Buell, L. (). ‘Evaluating the immediate postverbal position as a focus position in Zulu’, in M. Matondo et al. (eds.), Selected Proceedings of the th Annual Conference on African Linguistics. Somerville, MA: Cascadilla Proceedings Project, –. Carlson, R. (). ‘Narrative, subjunctive, and finiteness’, Journal of African Languages and Linguistics : –. Carlson, R. (). A Grammar of Supyire. Berlin: Mouton de Gruyter. Carstens, V. (). ‘Wh-movement in Yoruba’, Studies in African Linguistics, Supplement : –. Carstens, V. (). ‘Concord in minimalist theory’, Linguistic Inquiry : –. Carstens, V. (). ‘Antisymmetry and word order in serial constructions’, Language : –. Carstens, V., and Diercks, M. (). ‘Parameterizing case and activity: hyper-raising in Bantu’, in S. Kan, C. Moore-Cantwell, and R. Staubs (eds.), Proceedings of the North East Linguistic Society . Amherst, MA: Graduate Linguistic Student Association, –. Childs, G. T. (). A Grammar of Kisi: A Southern Atlantic language. Berlin and New York: Mouton de Gruyter. Childs, G. T. (). An Introduction to African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Childs, G. T. (). ‘Focus in Mani and Kisi’, in S. Ermisch (ed.), –. Cinque, G. (). ‘Deriving Greenberg’s Universal and its exceptions’, Linguistic Inquiry : –. Claudi, U. (). Die Stellung von Verb und Objekt in Niger-Kongo-Sprachen: Ein Beitrag zur Rekonstruktion historischer Syntax. Cologne: Institut für Afrikanistik, Universität zu Köln. Clements, G. (). ‘The logophoric pronoun in Ewe: its role in discourse’, Journal of West African Languages : –. Collins, C. (). ‘Argument sharing in serial verb constructions’, Linguistic Inquiry : –. Collins, C. (). ‘Multiple verb movement in ǂHoan’, Linguistic Inquiry : –. Collins, C. (). ‘The internal structure of vP in Juǀ’hoansi and ǂHoan’, Studia Linguistica : –. Comrie, B. (). Language Universals and Linguistic Typology, nd edn . Oxford: Basil Blackwell. Creider, C. A. (). The Syntax of the Nilotic Languages. Berlin: Dietrich Reimer. Creissels, D. (). ‘Typology’, in B. Heine and D. Nurse (eds.), –. Creissels, D. (). ‘S-O-V-X constituent order and constituent order alternations in West African languages’, Berkeley Linguistics Society : –. Culy, C. (). ‘Aspects of logophoric marking’, Linguistics : –. Curnow, T. J. (). ‘Three types of verbal logophoricity in African languages’, Studies in African Linguistics : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Dalgish, G. M. (). ‘Subject identification strategies and free word order: the case of Sandawe’, Studies in African Linguistics : –. den Besten, H. (). ‘Khoekhoe syntax and its implications for L acquisition of Dutch and Afrikaans’, Journal of Germanic Linguistics : –. Dickens, P. J. (). A Concise Grammar of Juǀ’hoan. Cologne: Rüdiger Köppe. Dimmendaal, G. J. (). The Turkana Language. Dordrecht: Foris. Dimmendaal, G. J. (). ‘Head marking, dependent marking and constituent order in the Nilotic area’, in F. K. E. Voeltz (ed.), Studies in African Language Typology. Amsterdam and Philadelphia, PA: John Benjamins, –. Dimmendaal, G. J. (ed.) (). Coding Participant Marking: Construction types in twelve African languages. Amsterdam and Philadelphia, PA: John Benjamins. Dryer, M. S. (). ‘Noun phrase structure’, in T. Shopen (ed), Language Typology and Syntactic Description. Cambridge: Cambridge University Press, –. Ekundayo, S. A., and Akinnaso, F. N. (). ‘Yoruba serial verb string commutability constraints’, Lingua : –. Ermisch, S. (ed.) (). Focus and Topic in African Languages. Cologne: Rüdiger Köppe. Fabb, N. (). ‘Reduplication and object movement in Ewe and Fon’, Journal of African Languages and Linguistics : –. Fiedler, I., and Schwarz, A. (eds.) (). The Expression of Information Structure: A documentation of its diversity across Africa. Amsterdam and Philadelphia, PA: John Benjamins. Frajzyngier, Z. (). A Grammar of Mupun. Berlin: Dietrich Reimer. Gensler, O. D. (). ‘On reconstructing the syntagm S-Aux-O-V-Other to Proto-Niger-Congo’, in Proceedings of the th Annual Meeting of the Berkeley Linguistics Society (Special Session on Historical Issues in African Linguistics). Berkeley, CA: Berkeley Linguistic Society, –. Gensler, O. D. (). ‘Grammaticalization, typology, and Niger-Congo word order: progress on a still-unsolved problem. Review of Claudi ()’, Journal of African Languages and Linguistics : –. Givón, T. (). ‘Serial verbs and syntactic change: Niger-Congo’, in C. N. Li (ed.), Word Order and Word Order Change. Austin, TX, and London: University of Texas Press, –. Good, J. (). ‘When arguments become adjuncts: negation and object preposing in Leggbó’, in J. E. Cihlar, A. L. Franklin, D. W. Kaiser, and I. Kimbara (eds.), Proceedings of Chicago Linguistic Society : Main session. Chicago: Chicago Linguistic Society, –, http://www.acsu.buffalo.edu/ ~jcgood/jcgood-Leggbo_CLS.pdf accessed Apr. , . Gragg, G. (). ‘Cleft sentences in Tigrinya’, Journal of African Languages : –. Green, M. (). Focus in Hausa. Oxford and Boston, MA: Basil Blackwell. Greenberg, J. H. (). ‘Some universals of grammar with particular reference to the order of meaningful elements, in J. H. Greenberg (ed.), Universals of Language. Cambridge, MA: MIT Press, –. Güldemann, T. (). ‘Toward a grammaticalization and typological account of the ka-possessive in southern Nguni’, Journal of African Languages and Linguistics : –. Güldemann, T. (). ‘Logophoricity in Africa: an attempt to explain and evaluate the significance of its modern distribution’, Sprachtypologie und Universalienforschung : –. Güldemann, T. (). ‘Preverbal objects and information structure in Benue-Congo’, Journal of African Languages and Linguistics : –. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, –. Güldemann, T., and Vossen, R. (). ‘Khoisan’, in B. Heine and D. Nurse (eds.), –. Haacke, W. H. G. (). ‘Syntactic focus marking in Khoekhoe (“Nama/Damara”)’, in I. Fiedler and A. Schwarz (eds.), Papers on Information Structure in African Languages. Berlin: Zentrum für Allgemeine Sprachwissenschaft, –. Hagège, C. (). ‘Les pronoms logophoriques’, Bulletin de la Société de Linguistique de Paris : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Hagman, R. S. (). Nama Hottentot Grammar. Bloomington, IN: Indiana University Publications. Halpert, C. (). Argument Licensing and Agreement. New York: Oxford University Press. Harford(Perez), C. (). Aspects of Complementation in Three Bantu Languages. Ph.D. thesis. Madison, WI: University of Wisconsin. Hawkins, J. (). Word Order Universals. New York: Academic Press. Hayward, R. J. (). ‘Qafar (East Cushitic)’, in A. Spencer and A. M. Zwicky (eds.), The Handbook of Morphology. Oxford: Basil Blackwell, –. Heine, B. (). A Typology of African Languages, based on the order of meaningful elements. Berlin: Dietrich Reimer. Heine, B. (). ‘Determination in some East African languages’, in G. Brettschneider and C. Lehmann (eds.), Wege zur Universalienforschung: Sprachwissenschaftliche Beiträge zum . Geburtstag von Hansjakob Seiler. Tübingen: Gunter Narr, –. Heine, B. (). ‘Adpositions in African languages’, Linguistique africaine : –. Heine, B. (). ‘Africa as a linguistic area’, in K. Brown and S. Ogilvie (eds.), Concise Encyclopedia of Languages of the World. Amsterdam: Elsevier, –. Heine, B., and Nurse, D. (eds.) (). African Languages: An introduction. Cambridge: Cambridge University Press. Henderson, B. (). ‘Agreement and person in anti-agreement’, Natural Language & Linguistic Theory (): –. Hiraiwa, K. (). ‘Relativization in Buli’, in G. Akanlig-Pare and M. Kenstowicz (eds.), Studies in Buli Grammar. Cambridge, MA: MIT Working Papers on Endangered and Less Familiar Languages : –. Hudson, G. (). ‘Amharic and Argobba’, in R. Hetzron (ed.), The Semitic Languages. London and New York: Routledge, –. Hyman, L. M. (). ‘Consecutivization in Fe’fe’’, Journal of African Languages : –. Hyman, L. M., and Comrie, B. (). ‘Logophoric reference in Gokana’, Journal of African Languages and Linguistics : –. Hyman, L. M., and Watters, J. R. (). ‘Auxiliary focus’, Studies in African Linguistics : –. Jaggar, P. (). Hausa. Amsterdam and Philadelphia, PA: John Benjamins. Jake, J., and Odden, D. (). ‘Raising in Kipsigis’, Studies in the Linguistic Sciences : –. Jenewari, C. E. W. (). ‘Defaka, Ijo’s closest linguistic relative’, in I. R. Dihoff (ed.), Current Approaches to African Linguistics. Vol. . Dordrecht: Foris, –. Kandybowicz, J. (). The Grammar of Repetition: Nupe grammar at the syntax-phonology interface. Amsterdam and Philadelphia, PA: John Benjamins. Kandybowicz, J., and Baker, M. C. (). ‘On directionality and the structure of the verb phrase: evidence from Nupe’, Syntax : –. Kayne, R. (). The Antisymmetry of Syntax. Cambridge, MA: MIT Press. Keenan, E. (). ‘Predicate-argument structure in Malagasy’, in C. S. Burgess, K. Dziwirek, and D. B. Gerdts (eds.), Grammatical Relations: Theoretical approaches to empirical issues. Stanford, CA: Center for the Study of Language and Information, –. Kilian-Hatz, C. (). ‘Serial verb constructions in Khwe (Central-Khoisan)’, in A. Y. Aikhenvald and R. M. W. Dixon (eds.), Serial Verb Constructions: A cross-linguistic typology. Oxford: Oxford University Press, –. Kilian-Hatz, C. (). ‘Khwe’, in G. J. Dimmendaal (ed.), –. Kimenyi, A. (). A Relational Grammar of Kinyarwanda. Ph.D. thesis. Los Angeles, CA: University of California. König, C. (a). ‘!Xun’, in G. J. Dimmendaal (ed.), –. König, C. (b). ‘Ik’, in G. J. Dimmendaal (ed.), –. König, C. (c). ‘Focus in !Xun’, in S. Ermisch (ed.), –. Koopman, H. (). ‘Control from COMP and comparative syntax’, The Linguistic Review : –. Koopman, H. (). The Syntax of Verbs: From verb movement rules in the Kru languages to Universal Grammar. Dordrecht: Foris.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Koopman, H. (). ‘On the parallelism of DPs and clauses’, in A. Carnie, H. Harley, and S. A. Dooley (eds.), Verb First. On the syntax of verb-initial languages. Amsterdam and Philadelphia, PA: John Benjamins, –. Kramer, R. (). Definite Markers, phi-Features, and Agreement: A Morphosyntactic Investigation of the Amharic DP. Ph.D. thesis. Santa Cruz, CA: University of California. Kula, N., and Cheng, L. (). ‘Phonological and syntactic phrasing in Bemba relatives’, Journal of African Languages and Linguistics : –. Lecarme, J. (). ‘Focus in Somali’, in G. Rebuschi and L. A. Tuller (eds.), The Grammar of Focus. Amsterdam and Philadelphia, PA: John Benjamins, –. Lukas, J. (). A Study of the Kanuri Language. London: Dawsons of Pall Mall. Manfredi, V. (). ‘Aspectual licensing and object shift’, in R.-M. Déchaine and V. Manfredi (eds.), Object Positions in Benue-Kwa. The Hague: Holland Academic Graphics, –. Marchese, L. (). Tense/Aspect and the Development of Auxiliaries in Kru Languages. Dallas, TX: Summer Institute of Linguistics. Marten, L. (). ‘The great SiSwati locative shift’, in A. Breitbarth, C. Lucas, S. Watts, and D. Willis (eds.), Continuity and Change in Grammar. Amsterdam and Philadelphia, PA: John Benjamins, –. Meinhof, C. (). Grundzüge einer vergleichenden Grammatik der Bantusprachen. Berlin: Dietrich Reimer. Mekoulnodji, N., Melick, C., and Moeller, S. (). A Brief Grammatical Sketch of Ngambay. GIALens: Electronic Notes Series ., http://www.gial.edu/GIALens/vol-/MelickMoellerMekoulnodj-Ngambay.pdf accessed May , . Moshi, L. (). ‘Word order in multiple object constructions in KiVunjo-Chaga’, Journal of African Languages and Linguistics : –. Mous, M. (). A Grammar of Iraqw. Hamburg: Helmut Buske. Ndayiragije, J. (). ‘Checking economy’, Linguistic Inquiry : –. Newman, P. (). The Classification of Chadic within Afroasiatic. Leiden: Universitaire Pers. Newman, P. (). The Hausa Language: An encyclopedia reference grammar. New Haven, CT: Yale University Press. Nicolaï, R. (). ‘Position, structure and classification of Songay’, in M. L. Bender (ed.), –. Nikitina, T. (). ‘The syntax of postpositional phrases in Wan, an “SOVX” language’, Studies in Language : –. Nikitina, T. (). ‘Categorial reanalysis and the origin of the S-O-V-X word order in Mande’, Journal of African Languages and Linguistics : –. Ouhalla, J. (). ‘Subject extraction, negation and the anti-agreement effect’, Natural Language and Linguistic Theory : –. Ouhalla, J. (). ‘Semitic relatives’, Linguistic Inquiry : –. Payne, D. L. (). ‘Argument structure and locus of affect in the Maasai external possession construction’, in A. C. Bailey, K. E. Moore, and J. L. Moxley (eds.), Proceedings of the TwentyThird Annual Meeting of the Berkeley Linguistics Society: Special session on syntax and semantics in Africa. Berkeley, CA: Berkeley Linguistics Society, –. Pearson, M. (). The Clause Structure of Malagasy: A Minimalist Approach. Ph.D. thesis. Los Angeles, CA: University of California. Potsdam, E. (). ‘The long-distance anaphor in Fula’, in A. Akinlabi (ed.), Theoretical Approaches to African Linguistics. Trenton, NJ: Africa World Press, –. Richards, N., and van Urk, C. (). ‘Two components of long-distance extraction: successivecyclicity in Dinka’, Linguistic Inquiry : –. Rijkhoff, J. (). ‘Descriptive and discourse-referential modifiers in a layered model of the noun phrase’, Linguistics : –. Robbers, K. (). Non-Finite Verbal Complements in Afrikaans: A comparative approach. The Hague: Holland Academic Graphics. Rose, Sh., Ackerman, F., Gibbard, G., Jenks, P., Kertz, L., and Rohde, H. (). ‘In-situ and ex-situ wh-question constructions in Moro’, Journal of African Languages and Linguistics : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Rust, F. (). Praktische Namagrammatik. Cape Town and Amsterdam: A. A. Balkema. Sadiqi, F. (). ‘Raising in Berber’, Studies in African Linguistics : –. Saeed, J. I. (). The Syntax of Focus and Topic in Somali. Hamburg: Helmut Buske. Sands, B. (). ‘Hadza’, in R. Vossen (ed.), –. Schneider-Blum, G. (). ‘Alaaba’, in G. J. Dimmendaal (ed.), –. Schneider-Zioga, P. (). ‘Anti-agreement, anti-locality and minimality: the syntax of dislocated subjects’, Natural Language and Linguistic Theory : –. Schuh, R. (). ‘Chadic overview’, in M. L. Bender, G. Takács, and D. L. Appleyard (eds.), Selected Comparative-Historical Afrasian Linguistic Studies, in Memory of Igor M. Diakonoff. Munich: Lincom Europa, –. Sebba, M. (). ‘Some remarks on Juǀ’hoan serial verbs’, in A. Traill, R. Vossen, and M. Biesele (eds.), The Complete Linguist: Papers in memory of Patrick J. Dickens. Cologne: Rüdiger Köppe, –. Smith, N. V. (). ‘Repetition of the verb in Nupe’, African Language Studies : –. Sonaiya, R. (). ‘Wh-movement and proper government in Yoruba’, in P. Newman and R. D. Botne (eds.), Current Approaches to African Linguistics. Vol. . Dordrecht: Foris, –. Svolacchia, M., Mereu, L., and Puglielli, A. (). ‘Aspects of discourse configurationality in Somali’, in K. É. Kiss (ed.), Discourse Configurational Languages. Oxford University Press: New York and Oxford, –. Tellier, C. (). ‘Head-internal relatives and parasitic gaps in Mooré’, in I. Haïk and L. Tuller (eds.), Current Approaches to African Linguistics. Vol. . Dordrecht: Foris, –. Torrence, H., and Kandybowicz, J. (). ‘Wh-question formation in Krachi’, Journal of African Languages and Linguistics : –. Tosco, M. (). ‘The historical syntax of East Cushitic: a first sketch’, in T. Bearth, W. J. G. Möhlig, B. Sottas, and E. Suter (eds.), Perspektiven afrikanistischer Forschung: Beiträge zur Linguistik, Ethnologie, Geschichte, Philosophie und Literatur / X. Afrikanistentag (Zürich, – September ). Cologne: Rüdiger Köppe, –. Tosco, M. (). ‘Cushitic and Omotic overview’, in M. L. Bender, G. Takács, and D. Appleyard (eds.), –. Tuller, L. A. (). Bijective Relations in Universal Grammar and the Syntax of Hausa. Ph.D. thesis. Los Angeles, CA: University of California. Tuller, L. A. (). ‘The syntax of postverbal focus constructions in Chadic’, Natural Language and Linguistic Theory : –. Ura, H. (). ‘Checking, economy, and copy raising in Igbo’, Linguistic Analysis : –. Van der Wal, J. (). Word Order and Information Structure in Makhuwa-Enahara. Utrecht: Landelijke Onderzoekschool Taalwetenschap. Vossen, R. (). ‘Comparative Eastern Nilotic’, in M. L. Bender (ed.), –. Vossen, R. (ed.) (). The Khoesan Languages. London and New York: Routledge. Watters, J. R. (). ‘Focus in Aghem: a study of its formal correlates and typology’, in L. M. Hyman (ed.), Aghem Grammatical Structure. Los Angeles, CA: Department of Linguistics, University of Southern California, –. Watters, J. R. (). ‘Syntax’, in B. Heine and D. Nurse (eds.), –. Welmers, W. E. (). African Language Structures. Berkeley, Los Angeles, CA, and London: University of California Press. Westphal, E. O. J. (). ‘The click languages of southern and eastern Africa’, in T. A. Sebeok (ed.), Linguistics in Sub-Saharan Africa. The Hague and Paris: Mouton, –. Williamson, K. (). A Grammar of the Kolokuma Dialect of Ịjọ. Cambridge: Cambridge University Press. Williamson, K., and Blench, R. (). ‘Niger-Congo’, in B. Heine and D. Nurse (eds.), –. Wolff, H. E. (). Referenzgrammatik des Hausa. Münster: LIT. Zeller, J. (). ‘Raising out of finite CP in Nguni: the case of fanele’, Southern African Linguistics and Applied Language Studies : –. Zerbian, S. (). Expression of Information Structure in the Bantu Language Northern Sotho. Ph.D. thesis. Berlin: Zentrum für Allgemeine Sprachwissenschaft.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
..............................................................................................................
LANGUAGE COMPARISON ..............................................................................................................
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. I principle, language typology is a globally oriented rather than geographically restricted linguistic discipline. Within the scope of the present handbook, however, geographical confinement seems inevitable and even well justified, not only because of the fact that there have already been a number of more or less shining examples in the history of African linguistics, some of which will briefly be called to mind below (section .). The goal of this chapter is to highlight major salient features in the structures of African languages and their approximate distribution. A typological classification is not aimed at. Traditionally, linguistic typological studies used to be of a grammatical nature. Morphological and word order (i.e. syntactical) typology, the polarity of “grammatical” versus “lexical” languages, subject–object relations, and phonological typology turned out to be the chief fields of investigation. Morphological typology examines the expressive power of non-lexical meanings such as number, case, person, definiteness, tense, aspect, and mood (in contrast to lexical categories such as substantives, verbs, adjectives, or, perhaps, adverbs) and the ways in which these are represented: by inflectional, morphologically bound elements (affixes) or by functional items or particles (e.g. articles, adpositions, pronouns, auxiliaries) (see also Beyer, chapter of this volume). Within word order typology, two major findings deserve mentioning here: the type-specific distinction between free and fixed (or rigid) word order, on the one hand, and the tendency of modifiers to consistently precede or follow the governing word, on the other hand. The distinction between “grammatical” and “lexical” languages is based essentially on the observation that the former owe their expressive power in particular to the syntactical combination of lexical items, i.e. words, whereas the latter depend, above all, on words and their morphological combination. Research into subject–object relations was considered elementary and was triggered by the observation that in certain, mainly non-European languages it is not the subject (agent) of a transitive verb but its object (patient) that is encoded in agreement with the subject of intransitive verbs; in other words, a transitive sentence in those languages
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
resembles a passive sentence in European standard languages. This phenomenon, known today as ergativity, marked the beginning of generally questioning traditional categories such as subject and object. Finally, phonological typology lays emphasis on sound patterns such as syllable structure, consonantal versus vocalic dominance, regressive versus progressive assimilation processes, and tonality. Against the background of broadened knowledge of individual languages generally, improved description of grammatical categories and patterns, and an advanced methodology with respect to descriptive generalizations, attempts have been made in recent times to harmonize the above-mentioned typological fields in order to understand the ways in which these fields are interrelated. In addition, new fields were opened up or deepened with the introduction of lexical, discourse, and areal typology. For a very long time, lexical typology was a notoriously neglected, if not non-existent, field. Without going into detail here for reasons of economy, suffice it to say that this was due to misconceptions in theoretical linguistics of the lexicon as a fundamental body of language. For the lexicon is not just an inventory of lexical items of a language but needs to be apprehended as a complex structure that is built upon categories and relations. It is exactly this systematic aspect of the lexicon that is relevant to grammar. Lexical entries fall into a number of lexical classes of which the most significant are word classes. More importantly, lexical classes and grammatical categories are mutually conditional. Lexical classes prescribe the grammatical behavior of lexical entries and, as a consequence, their function in grammar (cf. Lehmann ). Central phenomena of lexical typology are productive word formation processes: strategies of lexical extension, lexical rules, cognitive linguistics (strategies for semantic extension), and semantic relations within the lexicon (e.g. polysemy, word fields). In contrast to syntactic typology, which focuses on grammatical sentence structure, discourse typology goes beyond the sentence level by analyzing and establishing principles of textual formation and discourse strategies. Discourse analysis seeks to determine the elementary components of a given text body such as discourse segments (e.g. intonation patterns, pause structure) and grammatical markers (e.g. discourse particles), but also semantic devices like plot structure. This last aspect is closely connected with the problem of textual genres, as it seems clear that, for instance, narrative texts (e.g. stories, fairy tales, riddles) generate plot structures that are different from, for example, topical texts. Discourse typology requires extensive and wide-ranging text documentation which in Africa, however, is still in the making. For an instructive example of discourse analysis, see Serzisko (). Areal typology examines linguistic features shared by languages which are spoken in geographically contiguous areas but need not be genetically interrelated. When these features occur on various—if not all—linguistic levels and can be shown to go back to mutual language contact the term Sprachbund (linguistic area) is commonly used to refer to such a situation.1 More often than not, however, the verification of a Sprachbund situation is a difficult task because linguistic resemblances may have reasons other than language
1 A good number of recent studies on areal typology in Africa are contained in Heine and Nurse () and Hieda et al. (); of particular interest here is Güldemann’s () contribution on the so-called “Macro-Sudan belt”.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
contact (i.e. genetic relationship, chance, or linguistic universals) (cf., e.g., Comrie ). Contact-induced phenomena tend to occur initially on the lexical level; however, long-lasting and intensive contact between languages may lead to far-reaching structural concurrences that frequently make them look more “similar” to one another than genetically related languages (cf., e.g., Thomason and Kaufman ).2
. P A :
.................................................................................................................................. In the African context, typological studies comprise a broad spectrum of objectives and methodological operations. Basically, three principles of typological classification appear to have been applied in all these studies: the “ideal type” principle, the “prototype” principle, and the “discrete type” principle. Ideal type classifications are characterized by the fact that the criteria employed for the distinction of types are so distant from linguistic reality that they can hardly, if at all, be fulfilled. In other words, the ideal type is a theoretical construct that languages may approximate to but will never correspond to entirely. In contrast to this, prototype classifications are geared with actually existing languages, i.e., they are based exclusively on features which are linguistically attested. Finally, discrete typologies are typified by the fact that a language can only be assigned to a given typological class if it meets all the criteria of definition set up for that class. Although all three principles of typological classification represent legitimate concerns of linguistic research, it appears that discrete typologies are of greatest importance with respect to both grammatical theory and research in linguistic universals. Another facet of great importance in the history of language typology is the dichotomy of full-system versus partial-system typology. While the former lays claim to making statements about the whole structure of languages, the latter is confined to statements about specific parts of linguistic structure (e.g. phonology, word order). In the African context, classic examples of full-system typology are Lepsius () and Meinhof (). Lepsius’s classification, known as Drei-Zonen-Theorie (three-zone theory), was based on phonological (distinctive tones, syllable structure), morphological (e.g. prefixes vs. suffixes, gender vs. non-gender, grammatical agreement), and syntactical (e.g. S–V–O vs. others, prepositions vs. postpositions, genitive word order) criteria and was meant chiefly to work out the dividing line between Bantu and the then so-called “Hamitic” languages. Although Lepsius’s statements were of a synchronic typological nature, they pursued a historical objective. The focus of Meinhof ’s study () was on explaining the emergence of inflectional languages, thereby leaning on Schleicher’s () and contemporaries’ general conception of a tripartite distinction between isolating (“Sudanic”), agglutinating (Bantu), and inflectional (“Hamitic”, Semitic) languages. (For details, see Beyer, chapter of this volume.) 2
For more on language contact in Africa, see Beyer (chapter of this volume).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A more recent and more advanced attempt at providing a full-system typology is Houis (), who arrived at a bipartite division of African languages by using criteria from all linguistic levels. Part of his methodology was to declare features as existent (+) or nonexistent ( ), i.e. make predicative statements for assigning typological features in order to outline contrasts. He also characterized languages by making comparative statements such as that languages can be morphologically richer (i.e. have a complex internal word structure) or less rich (i.e. operate largely with bare lexical stems). Leaving aside here the details of his typological classification, an implicit result may be seen in the insight that predicative statements are of limited validity, whereas comparative statements appear to be more appropriate. Numerical statements may be considered most accurate because of the measurability of typological proximity or distance. As previously mentioned, partial-system typologies rely on selective criteria. An early example is Schmidt’s () globally oriented study. He examined the distribution of number categories such as dual and trial, the occurrence of inclusive vs. exclusive personal pronouns, and noun classification patterns, as well as the order of words and morphemes. The last-mentioned criterion in particular turned out to be of significance for later research on typology because of some implications that Schmidt was able to demonstrate. Hence, if languages make use of prepositions, then the genitive and the adjective follow the head noun and the object follows the verb. In contrast to this, genitive–noun, object–verb, and adjective–noun orders are observed when languages have postpositions. Perhaps one of the most influential partial-system typologies was established by Heine () in the realm of word order. Taking Greenberg’s (a) universally oriented syntactic typology as a point of departure, Heine employed parameters in his investigation of more than African languages (across genetic families) and came up with basically four major syntactic types: () type D, whose dominant word order is S–O–V with a typical preverbal positioning of the adverbial phrase; () type C, which is characterized by sentence-initial predication/verb positioning (V–S–O); () type B, which, like type D, has dominant S–O–V and genitive-noun order and uses postpositions instead of prepositions but places the adverbial phrase after the verb; () type A, which is S–V–O in principle and represents the unmarked type from which B/C/D can be derived. Each type is characterized by additional syntactic features, and subtypes have been established too. There are very few discrete typologies in African linguistics. A comprehensive attempt was made by Heine and Vossen (), who used occurrence/non-occurrence patterns of three morphological (noun class system (NCS), nominal case inflection (NCI), and derivative verb extensions (DVE)), plus one syntactical feature (order of meaningful elements, based on Heine ), as criteria for defining linguistic types. Approximately languages from all genetic phyla and families (cf. Greenberg b) were investigated and, as a result, seven (out of eight logically possible) types were established. The noun class criterion was chosen because a large majority of African languages from all genetic phyla and all parts of the continent possess such a system. Languages were considered to have a noun class system when the nouns can be allocated to classes on the basis of agreement markers. The nominal case criterion was regarded as met if minimally nominative and accusative are distinguished by way of nominal inflection. Verbal extensions can also be considered a striking typological feature of African languages. More than ten such extensions can
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
frequently be found in languages from all phyla and all parts of the continent. Finally, Heine’s word order typology was taken as a welcome and sound syntactical amplification. Subject to all conceivable combinations of features of the classificatory criteria discussed above, a maximum of eight major types (called through ) could logically be distinguished. Each type could be further subdivided into four subtypes (A through D) when consulting the results of Heine’s () word order typology. The quantitative evaluation yielded a number of interesting figures as well as insights. Thus, a remarkable majority of languages (.%) was found to belong to type , which is defined by the feature combination [+NCS / NCI / +DVE]. The lowest representation was ascertained for type (.%), characterized by [ NCS / +NCI / DVE], whereas one type (: [+NCS / +NCI / DVE]) had to be described as virtually non-existent. For the remaining types figures of between .% and .% had been calculated. Probably more important than percentages were some typological implications that ensued from the feature combinations, for example: () Languages that have both a noun class and a nominal case-marking system always also possess the possibility of changing the verbal function and, hence, valence by means of derivative affixes. () Languages which place the genitive before the head noun and the adverbial phrase after the verb, i.e. Heine’s Type B languages, are not equipped with nominal case inflection. () Languages that place the verb/predicate at the beginning of the sentence (Heine’s Type C) always have derivative extensions. This discrete typology generally confirms the fundamental different nature of genetic and typological language classifications which is conditioned by the diversity of criteria of comparison. When contrasting Greenberg’s putative language phyla with the seven language types resulting from Heine and Vossen’s study, one finds that (a) Niger-Congo/ Kordofanian languages show up in all seven types, (b) Nilo-Saharan and Afro-Asiatic languages are distributed over six types each, and even the numerically smallest of all phyla, (c) the Khoisan languages, are still to be found in three of the seven types. However, when one looks at individual families within the phyla, a restrictive correlation between genetic and typological relationship can be discovered. For example, Central Sudanic languages occur exclusively as type ([ NCS / NCI / +DVE]) or type ([ NCS / NCI / DVE]) representatives, whereas Semitic and East Cushitic distribute over type ([+NCS / +NCI / +DVE]) and type . All Atlantic languages except Gola belong to type . A similar statement can be made for Mande, which, with the exception of Khasonke and Bisa, appears to fall within type .
. S
.................................................................................................................................. For the distribution of linguistic features worldwide, readers’ attention is directed to Dryer and Haspelmath’s () famous atlas project.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Phonological characteristics African languages show a wide range of phonological types, with some features highly marked in a worldwide sample. Since phonological systems of African languages are treated systematically in Kenstowicz (chapter of this volume), we shall here confine ourselves to some selected prominent examples. Among consonants a highly promising candidate for finding (nearly) exclusive or geographically restricted segments is stops because, generally speaking, they occur in a multitude of shapes in African languages. While simple stops (explosives) such as p, t, k show up as a rule, implosives stops (e.g. ɗ, ɠ ) are relatively rare and, in most cases, do not cooccur with their explosive counterparts in one and the same language. Common complex stops are affricates (e.g. px, tx, kx) and aspirated plosives (e.g. ph, th, kh); glottalization (ejectives such as p’, t’, k’ ) is known to exist widely in Afro-Asiatic and Khoisan language families. Phonemic stop clusters (e.g. px’, tx’, kx’ ) also occur, but in a limited way. Several language families or groups are well known for making extensive use of marked sounds. West African Kwa and Gur phonological systems are to be mentioned here as the most prominent examples for the occurrence of labial-velars kp and gb, which do not appear in word-final position, though. Bantu languages are first-class representatives of prenasalized stops, such as mb, nd, ŋg, and also fricatives (e.g. nz ). These segments are phonologically distinct and must not be confused with (non-phonemic) nasal+stop combinations, which also exist in Bantu. Thus, /mb/ in the Swahili adjective -embamba ‘narrow’ is a phoneme, whereas [mb] in the noun mbwa ‘dog’ is not because of the morpheme boundary that exists between the nasal class / prefix /m/ and the stem-initial stop /b/ (hence, /m-bwa/). The most extended consonant inventories on the African continent are to be found in Khoisan languages of southern Africa, in particular in North (!Xun) and South Khoisan (Tuu). Apart from complex egressive consonant systems, these languages are peculiarly characterized by additional subsets of ingressive, so-called click consonants.3 As a result of long-enduring and extensive language contact, a number of (presently and/or formerly) neighboring Bantu languages (e.g. Nguni group, Kavango languages) plus East Cushitic Dahalo in Kenya4 also possess click consonants (cf., e.g., Vossen a). Clicks consist of two components, the basic click type (primary articulation) and the click accompaniment (secondary articulation). There are five basic click types: lateral (symbolized by ǁ), dental (ǀ), alveolar (ǃ), palatal (ǂ), and labial (ʘ). Alveolar and palatal clicks are stops; the other clicks are fricatives. While the labial click is restricted to Tuu languages and ǂHoan, the rest are widely distributed over the Khoisan families. Click accompaniments can be as many as twenty but, just like basic types, they are not evenly distributed in Khoisan. Accompaniments mark the transition from basic click to the following vowel in a strictly CVCV/CVV/CVN phonological word structure. They are
3
The distinction between ingressive and egressive is largely based on the underlying (pulmonic and lingual) air stream mechanisms that form part of the reproduction process. 4 East Africa is seen by many scholars (see, e.g., Ehret ) as a region that was formerly inhabited by Khoisan-speaking populations. Today, the only East African languages counted as Khoisan are Hadza and Sandawe; both spoken in Tanzania and thus far away from Dahalo.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Examples of click-containing words in Khoisan Language
Click type
Accompaniment
Lexeme
Gloss
Juǀh’oan
dental
voiced
gǀúbú
‘crop of a bird’ (Dickens 1994: 205)
!Xun
alveolar
prenasalized
n!ùàn
‘throw’ (König and Heine 2008: 54)
Khwe
lateral
glottalized
ǁ’áé
‘camp, home’ (Kilian-Hatz 2003: 206)
Naro
palatal
aspirated
ǂhùní
‘elbow’ (Vossen 1997: 430)
!Xóõ
labial
aspirated-uvular
ʘqhú’u
‘be jammed in between’ (Traill 2018: 42)
derived from egressive consonantal properties, i.e. click accompaniments as a rule form part of the egressive inventory too. Over time, clicks tend to get lost and replaced by clickless—that is, egressive—consonants. For an overview of such processes and mechanisms, the reader is referred to Traill and Vossen (). Click phonemes are exclusively found in word-, i.e. root-initial position. They do, however, occur word-medially in compounds; e.g. Cara cxóánǁnàà ‘elephant tusk’ < cxóá+nǁnáà ‘elephant+horn’ (Vossen b: ). Table . gives some lexical examples.5 Probably all African languages possess oral vowels, and many have nasal vowels in addition. However, other distinctive colorings are also known to exist in a number of language families or groups. In Northern (!Xun) and Southern Khoisan languages, for instance, we find glottalized, pharyngealized, and breathy vowels next to oral and nasal ones, combinations of which are also possible, thus enlarging vocalic phoneme inventories to a considerable extent. Some Western Nilotic (Eastern Sudanic; Nilo-Saharan) languages also have breathy vowels, and there are up to three significant degrees of vowel length (short, long, and medium) in languages like Dinka and Nuer (Tucker and Bryan : –). [ATR] feature-based vowel harmony, mostly of the cross-height type with distinctive sets of [+ATR] (“closed” or “tense”) and [ ATR] (“open” or “lax”) vowels, is another important phenomenon that occurs in different parts of Africa, e.g. in Kwa, some Chadic, and basically all Nilotic languages.
.. Morphosyntactic characteristics ... Noun class systems As stated earlier (see section ..), a widely found and prominent phenomenon in African languages is noun class systems. Here, a distinction may be made between gender and class
5
A sound and condensed description of clicks is contained in Ladefoged and Maddieson (: –).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
systems. However, “gender” and “class” will here be replaced by sex-oriented and natureoriented noun class systems respectively. Sex-oriented systems are characterized minimally by a grammatical distinction of masculine and feminine. Most typically such systems have only a limited number of nominal classes. To the best of my knowledge, no African language is known to have more than three classes (“genders”). If a third class does exist, it is, according to its semantic content, mostly designated as neuter, common, or diminutive (sometimes locative). Sex-oriented languages occur overwhelmingly in Afro-Asiatic, but can also be found in some Nilo-Saharan (Eastern and Southern Nilotic) and Central Khoisan languages. Noun class allocation in languages with a nature-oriented class system is based above all on the opposition human/non-human, on the one hand, and animate/non-animate, on the other. But animals, plants, tools, uncountable quantities, collective and abstract concepts, small and big objects, etc. may also function as classificatory devices. The number of such noun classes normally ranges from two to thirty and is on an average much larger than that of sex-oriented systems. Nature-oriented systems are predominant in a majority of NigerCongo (especially Bantu; cf. Table . for Yeyi) languages, but they also occur in Northern and Southern Khoisan. A few African languages (e.g. Zande, Ubangi; cf. Pasch, chapter of this volume) show characteristics of both types of noun class systems. A further distinction can be made between overt and covert noun class systems. The former are characterized by obligatory class markers on the noun, while the latter do not
Table . Inventory of noun classes in Yeyi (Bantu; see Seidel : –) Noun class
Nominal prefix
Semantic scope
1 1a 2 2x 3 4 5
muØ-, ubabamumili-
6 7 8 9 9a 10 10a 11 12 13 14 15 16 17 18
mashizii(N)izi(N)zirukatuwukupakumu-
human animals, kitchen tools, etc. plural to 1 and 1a honorifics plants, natural phenomena, manner/style, etc. plural to 3 paired things, natural phenomena, augmentatives, singulatives, etc. liquids, uncountables, collectives, plural to1a, 5, 9, 9a, 14 inanimates, pejoratives, etc. plural to 7 miscellaneous miscellaneous, loans plural to 9 plural to 9a elongated things, groups, singulatives, etc. household things, diminutives, etc. diminutives, plural to 12 abstracts, groups Infinitives, etc. explicit location general location interior location
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
have such markers. Overt class marking is achieved through prefixes, suffixes, or both (circumfixes or bilateral affixes), or else by suprasegmental means (tones). A close relationship between class and number is evident in the noun class systems of African languages.
... Case marking Apart from the primary distinction between nominative and accusative, which by itself is relatively rare (estimated %) in African languages, nominal case is morphologically marked in a number of languages for other cases, e.g. locative and genitive. In recent years, ergativity has increasingly come into play as a subject of investigation. Sometimes, it is not easy to determine whether case is expressed through inflection or independent words, for example postpositions. In a majority of languages, however, suffixes are used to mark case distinctions. Some employ tone. Case inflection is often bound to the noun. In many languages it occurs at the end of the noun phrase, though. African case languages can finally be typified according to their usage as the absolute nominal (i.e. citation) form of either nominative or accusative. Most languages fall within the former category. In a number of African languages verbal case marking is more important than, or compensates for, nominal case. For an illustrative and altogether path-breaking typology of case manifestations on the African continent, the reader is referred to König’s () eye-opening monograph, which also discusses the role of grammaticalization with case and the relationships between, for instance, definiteness, topic, focus, and case.
... Verbal extensions Derivative verb extensions are another widespread and morphosyntactically very important feature in African languages and are almost always marked by affixes, predominantly suffixes. These extensions are by themselves frequently derived from verbs which often have survived as such and, hence, exist side by side with the grammaticalized verbal formative, e.g. ǂHaba (Central Khoisan) -mà ‘dative ~ benefactive’ : mâ ‘to give’ (Vossen : , –). Verbal extensions fulfill an extraordinary variety of different functions that can be grouped into three larger functional domains: () Verbal extensions with prevalently syntactic functional load: these are used primarily to alter the sentence structure (verb valence) by removing or adding a syntactic role represented by a nominal or adverbial phrase. For example, the passive formative -(I)w- in Swahili (and other Bantu languages) is applied, above all, to delete the sentence or phrase agent. Derivations with prevalently syntactic function are causative, intransitive, passive, applicative (“prepositional”), reciprocal, reflexive, instrumental, ablative, comitative, associative, and contactive. They form the predominant group of verbal extensions. Generally, more than half of all extensions in African languages belong to this group, and in a certain number of languages syntactic extensions are even exclusive. () Verbal extensions with prevalently semantic function: these serve, above all, for altering or modifying the meaning of the action contained in the verbal root. The ventive extension, for instance, denotes a verbally expressed motion towards the speaker, whereas the opposite action, a motion away from the speaker, is conveyed
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
through the itive extension. In African languages, intensive, ventive, itive, reversive, and frequentative have a wider distribution. () Other verbal extensions: these represent the smallest of the three groups. In most cases they exercise functions that are normally filled by inflectional morphemes such as number marking on subject and object, verb negation, or verbal aspect marking.
... Serial verb constructions Most typically, serial verb constructions (SVC) are to be found in languages with little morphology, i.e. in so-called isolating languages. SVCs consist of two or more verbs whose individual meanings, taken together, often lay the foundation for the emergence of a new, semantically related meaning, e.g., ‘take’ + ‘go’ + ‘give’ > ‘take something out (to somebody)’. According to Collins (: ), “A serial verb construction is a succession of verbs and their complements (if any) with one subject and one tense value that are not separated by any overt marker of coordination or subordination.” In Africa, Kwa (Niger-Congo) languages are among the most promising candidates for SVCs. In Ewe, for instance, all verbs participating in a given SVC are marked for tense or aspect, as in () and (), where they occur in zero-marked aorist forms: ()
()
àdèlá sìà f ɔ́ yì àdègbé. hunter stand.up go hunt ‘This hunter stood up and went hunting.’
(Pasch : )
Kofi tsɔ́ kpé gbà̰ àtùkpá lá. Kofi take stone break.to.pieces bottle ‘Kofi has smashed the bottle with a stone.’
(Pasch : )
In (), the verbs in the series appear as future-marked stems, prefixed by á-: ()
Kofi á-mlì kpé-á á-tsó tó-á Kofi -roll stone- -come.from mountain- ‘Kofi will roll the stone from the mountain top.’
dzí. (Pasch : )
Also, Khoisan languages make use of SVCs. For Juǀ’hoan (Northern Khoisan), Dickens (: –) distinguishes two strategies: transitive and intransitive. In the former construction a verb is followed by a transitive verb plus object, as in (): ()
dȁ’àma khù gǀáíá kàtȍngá. child jump go.out.of box ‘The child jumped out of the box.’
The intransitive strategy shows verb plus intransitive verb plus subject word order, as shown in (): ()
ȕtò nǂàq’ú !àò jù. car knock fall.over person ‘The car knocked the person over.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Although (at least some) Central Khoisan languages may count as morphologically rich, they do employ SVCs to a certain extent. In Khwe, for instance, SVCs are a means of expressing adjectival concepts, as Kilian-Hatz (: ) reports: ()
ǁ’ó-ó-xu-a tì ̰ĩ nǁgóέ die- -- stand moon ’last stage of waning moon’
(lit. ‘dead standing moon’)
In Khwe narratives verb serialization also occurs when introducing a new episode (KilianHatz : ): ()
a. djiri yà kx’áà-à-tè. monkey come drink- I- b. djiri cì kx’áà-à-tè. monkey arrive drink- I- ‘When coming/arriving, the monkey drinks.’
The semantic and syntactic complexity of serial verbs has also been observed in Eastern ! Xóõ (Southern Khoisan) (Traill, quoted in Güldemann a: –): ()
í qúba sâa ǁgâã ’ǁnā̰h’-na sâ’ã go spend.day chop.out- seed. ‘and are said to have chopped out (tsamma) seeds all day long’
()
sí sâa ǂhābi tshûu ǀîi go get.on.top sit ‘to have gone to get on top’
. C
.................................................................................................................................. In recent times, the focus of African typological research has more and more been directed away from continent-wide investigations towards studies on various linguistic phenomena in individual languages or language areas. For the former target group, volumes of collected papers on negation (cf. Cyffer et al. ) or phasal polarity (Kramer and Kießling forthcoming) may be mentioned here as examples, while Güldemann’s (b) typological survey of the Khoisan-speaking area is singled out as an excellent representative of the latter. A broad spectrum of typological topics has been dealt with in Voeltz’s () proceedings of an international symposium that was held at the beginning of this century. There can be no doubt that all these and future follow-up contributions will continue to lead to a better comprehension of African linguistic structures and types.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Collins, C. (). ‘Multiple verb movement in ǂHoan’, Linguistic Inquiry : –. Comrie, B. (). Language Universals and Linguistic Typology. Cambridge: Cambridge University Press. Cyffer, N., Ebermann, E., and Ziegelmeyer, G. (eds.) (). Negation Patterns in West African Languages and Beyond. Amsterdam and Philadelphia, PA: John Benjamins. Dickens, P. J. (). (.) English Juǀ’hoan Ju-ǀ’hoan English Dictionary, nd edn. Cologne: Rüdiger Köppe. Dickens, P. J. (). A Concise Grammar of Juǀ’hoan, with a Juǀ’hoan–English glossary and a subject index. Edited by R. Vossen and M. Biesele. Cologne: Rüdiger Köppe. Dryer, M. S., and Haspelmath, M. (). The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology, http://wals.info accessed July , . Ehret, C. (). ‘The extinct Khoesan languages of eastern Africa’, in R. Vossen (ed.), –. Greenberg, J. H. (a). ‘Some universals of grammar with particular reference to the order of meaningful elements’, in J. H. Greenberg (ed.), Universals of Language. Cambridge, MA: MIT Press, –. Greenberg, J. H. (b). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), –. Güldemann, T. (a). ‘Syntax: Southern Khoesan (Tuu)’, in R. Vossen (ed.), –. Güldemann, T. (b). ‘Typology’, in R. Vossen (ed.), –. Heine, B. (). A Typology of African Languages, based on the order of meaningful elements. Berlin: Dietrich Reimer. Heine, B., and Nurse, D. (eds.) (). A Linguistic Geography of Africa. Cambridge: Cambridge University Press. Heine, B., and Vossen, R. (). ‘Sprachtypologie’, in B. Heine, T. C. Schadeberg, and E. Wolff (eds.), Die Sprachen Afrikas. Hamburg: Helmut Buske, –. Hieda, O., König, C., and Nakagawa, H. (eds.) (). Geographical Typology and Linguistic Areas. Amsterdam and Philadelphia, PA: John Benjamins. Houis, M. (). ‘Réflexion sur une double corrélation typologique’, Journal of West African Languages : –. Kilian-Hatz, C. (). Khwe Dictionary. Cologne: Rüdiger Köppe. Kilian-Hatz, C. (). ‘Syntax: Kxoe subgroup—Khwe’, in R. Vossen (ed.), –. König, C. (). Case in Africa. Oxford: Oxford University Press. König, C., and Heine, B. (). A Concise Dictionary of Northwestern !Xun. Cologne: Rüdiger Köppe. Kramer, R. L., and Kießling, R. (eds.) (forthcoming). The Expression of Phasal Polarity in African Languages. Berlin and New York: De Gruyter Mouton. Ladefoged, P., and Maddieson, I. (). The Sounds of the World’s Languages. Oxford: Blackwell. Lehmann, C. (). ‘Towards lexical typology’, in W. Croft, K. Denning, and S. Kemmer (eds.), Studies in Typology and Diachrony: Papers presented to Joseph H. Greenberg on his th birthday. Amsterdam and Philadelphia, PA: John Benjamins, –. Lepsius, R. (). Nubische Grammatik, mit einer Einleitung über die Völker und Sprachen Afrika’s. Berlin: Wilhelm Hertz. Meinhof, C. (). Die Entstehung flektierender Sprachen. Berlin: Dietrich Reimer. Pasch, H. (). Kurzgrammatik des Ewe. Cologne: Rüdiger Köppe. Schleicher, A. (). Die Sprachen Europas. Bonn: H. B. König. Schmidt, W. (). Die Sprachfamilien und Sprachenkreise der Erde. Heidelberg: Carl Winter. Seidel, F. (). A Grammar of Yeyi: A Bantu language of southern Africa. Cologne: Rüdiger Köppe. Serzisko, F. (). Sprechhandlungen und Pausen: Diskursorientierte Sprachbeschreibung am Beispiel des Ik. Tübingen: Max Niemeyer.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Thomason, S. G., and Kaufman, T. (). Language Contact, Creolization, and Genetic Linguistics. Berkeley, Los Angeles, CA, and London: University of California Press. Traill, A. (). A Trilingual !Xóõ Dictionary: !Xóõ–English–Setswana. Edited by H. Nakagawa and A. Chebanne. Cologne: Rüdiger Köppe. Traill, A., and Vossen, R. (). ‘Sound change in the Khoisan languages: new data on click loss and click replacement’, Journal of African Languages and Linguistics : –. Tucker, A. N., and Bryan, M. A. (). Linguistic Analyses: The non-Bantu languages of north-eastern Africa. London, New York, and Cape Town: Oxford University Press. Voeltz, F. K. E. (ed.) (). Studies in African Linguistic Typology. Amsterdam and Philadelphia, PA: John Benjamins. Vossen, R. (a). ‘What click sounds got to do in Bantu: reconstructing the history of language contacts in southern Africa’, in B. Smieja and M. Tasch (eds.), Human Contact through Language and Linguistics (Festschrift René Dirven). Frankfurt am Main: Peter Lang, –. Vossen, R. (b). Die Khoe-Sprachen: Ein Beitrag zur Erforschung der Sprachgeschichte Afrikas. Cologne: Rüdiger Köppe. Vossen, R. (ed.) (). The Khoesan Languages. London and New York: Routledge.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
.
. I
.................................................................................................................................. B dialectology and linguistic geography study language variation, change, and diffusion. The former examines geolinguistic data synchronically and diachronically on an intragenetic level (e.g. Bantu, Berber) or an intergenetic phylum level (e.g. Niger-Congo, Afro-Asiatic), whereas the latter focuses on stable, contact-induced linguistic changes on an intergenetic phylum level in adjacent areas.1 In the present chapter, I address these linguistic disciplines from an African perspective, explaining their fundamental principles by means of specific African-related cases. Furthermore, I compare the different approaches to African areal linguistics, pointing out their respective potentials and limitations. This introduction is followed by a preliminary explanation of the dialectological framework, problematizing some of its theoretical and methodological concepts, such as “dialect consciousness”, “inter-intelligibility”, “language continuum”, and the “language versus dialect” dichotomy (section .). In the same section, a detailed dialectometry classification and its various methods are presented. In section ., a critical overview of the existing studies in African areal classification is given. The chapter ends with a number of conclusions about language variation and its connection with language diffusion and evolution with respect to the African area (section .).
. D
.................................................................................................................................. Dialectology considers variation in language based most commonly on geographical distribution and its correlated phonological, grammatical, and lexical features. Synchronic dialectology studies the synchronic linguistic proximity of geographically contiguous 1
The term “genetic” is used here in the sense of genealogical inheritance.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
language varieties. Diachronic or historical dialectology describes and explains by means of socio-historical reconstructions the language evolution and diffusion processes which led to the current variation in form, meaning, and structure. The two approaches are interrelated and therefore no clear-cut distinction can be made between them. Indeed, in variational linguistic research (i.e. dialectology), any geographical diversity necessarily refers to a temporal diversity (Saussure []; Lafkioui ). In addition to the horizontal (geographical) and vertical (historical) dimensions, linguistic variation also involves social diversity and individual diversity. The former includes parameters such as habitat (rural/ urban), country of origin/immigration, gender, ethnicity, age, social class, educational level, and educational type; the latter overlaps with social variation but also incorporates ideological and psycho-historical aspects of the individual speakers (Lafkioui , a, ). In the case of Africa, it is historical dialectology which, of the two sub-disciplines, is the one in greater need of additional and advanced study. A notable exception to this lack is the research conducted in creolistics, which benefits not only dialectological but also linguisticgeographical studies (e.g. Gilman ; Mufwene ).2 Although the terms “dialectology” and “dialect” do not involve any value judgments in the eyes of dialectologists, they may be subject to controversy, in particular when the latter is used in contrast with the concept of “language” in sociolinguistic situations where linguistic varieties (lects or geolects) are in sociopolitical conflict, especially as the language–dialect distinction cannot be established on scientific criteria. Concepts like “dialect consciousness” and “inter-intelligibility”, even though often used as defining parameters (e.g. Daan ), do not suffice to distinguish languages from dialects or dialects from each other. Nor do they form adequate geographical demarcation devices for the varieties in question (Lafkioui , a: –), mainly because they are based on the subjective experience of speakers whose values—which are difficult to quantify—depend heavily on their personal disposition (ideological, social, and historical). As regards dialect consciousness, even if the speakers regard themselves as belonging to the same linguistic variety or group of varieties—calling themselves by the same anthroponym for example— their language attitudes often seem to be motivated by the social status accorded to the variety in question and the degree of in-group solidarity (Lafkioui ). Moreover, the validity of the inter-intelligibility criterion is primarily determined by the conscious or unconscious willingness of speakers, which does not always permit the researcher to decide whether mutual comprehension actually stems from no more than their linguistic competencies. It is not uncommon to find speakers who claim not to understand because they refuse to understand (Chambers and Trudgill : ). Therefore, these two extralinguistic methodological tools should be employed with caution when delimitating linguistic varieties and selecting (field) survey points for this purpose; their use should be confined to subsidiary verifications of the data and analyses drawn from linguistic inventories. The prime objective of research in dialectology is to analyze and classify the language variation of a specific geographical zone (diatopic variation) and so to identify the causal linguistic boundaries (isoglosses) of the different encountered varieties and variety groups, even when these cut across administrative and thus political borders. Essential to this approach is the construction of a linguistic atlas of the area being studied, that is, a
For African historical dialectology, see, e.g., Vossen (, ); Sommer and Vossen (); and Lafkioui (, c, a, b, , b). 2
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
structured set of geolinguistic maps displaying the precise distribution in space of particular linguistic phenomena. As discussed in Lafkioui (), a linguistic atlas—especially when it deals with non- or little-documented and studied varieties—should comprise, and give equal consideration to, all linguistic domains, namely phonetics, phonology, morphology, syntax, and lexicon. Such an approach facilitates further linguistic studies such as dialectometry classification, a quantitative methodology for measuring dialectal proximity or distance, which is one of the most appropriate developments of this research perspective and is explained in more detail in the next section. A major problem for classification (both quantitative and qualitative types), and particularly for statistical classifications such as dialectometrics, is the existence of so-called “dialect continua”. These impede a clear-cut demarcation of linguistic varieties (dialects and languages) because the progressive change which they exhibit prevents convergence to precise linguistic borderlines (isoglosses and bundles of isoglosses). Well-known examples of African dialect continua are the Bantu continua (e.g. Guthrie –; Möhlig ; Heine and Möhlig ), one of which is that found in Kenya, where the status of “language” has been assigned to varieties such as Gikuyu, Meru, and Kamba mainly on the basis of non-linguistic parameters (Heine and Möhlig : f.).3 Another case of dialect continua is that of Berber (Basset , ; Lafkioui ), of which the Rif-Berber continuum (North Morocco) has been the subject of a thorough dialectometry classification that is demonstrated in the following section (see also Lafkioui , a, b, a, b).
.. Introduction to methods in dialectometry Dialectometry is a quantitative methodology for calculating linguistic distances between linguistic varieties. The most frequently used dialectometry methods can be divided into the categories of traditional and computational methods.4 The most well-known traditional approaches are those based on the concept of isogloss, which is a line that bisects a geographic map into separate zones according to the detected linguistic features. The classification of the varieties is deducted from the arrangement of isoglosses, clusters of isoglosses (Goossens ), or clusters of demarcative isoglosses (Stankiewicz ; Garde ) on the geolinguistic map.5 Although this method allows for Nevertheless, Möhlig (, ) and Heine (, ) were the first to apply statistical methods to African corpora, the former in order to get a dialectometry classification of Bantu varieties, whereas the latter is interested in lexicostatistics and its importance to African historical classification. See also Bastin () for a lexicostatistic study of Bantu. 4 There also exist different perceptual approaches that permit drawing sociolinguistic borders based on the speaker’s dialectal conscience (e.g. Weijnen , ; Rensink ; Daan and Blok ; Gooskens ). 5 The qualifying term “demarcative”, added to the common dialectology criterion of “isogloss clusters” (Goossens : ), refers to the structural value of the isoglosses relating to the material aspect of the phenomena as well as to their relative distribution (direction and density). Thus, not only the quantitative dimension (number) of isoglosses is relevant to the typology of classification, but also the qualitative aspect, i.e. their degree of importance. However, non-demarcative isoglosses may also be of great significance for the classification, especially when they allow for an evaluation of the results. On the relationship between structuralism and dialectology, see, e.g., Weinreich (); Fourquet (); Grosse (); and Martinet (). 3
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
verification of the visualized facts, it has several disadvantages, including the difficulty of finding clusters of isoglosses that precisely divide the geolinguistic area examined (Kessler ).6 Another traditional technique is the geolinguistic structuring method, which divides a geographic area depending on the linguistic structure of its varieties (e.g. Moulton ; Goossens ). For instance, varieties with the same phonemic system are part of the same geolinguistic group. However, classifications based on this method are mainly phonological and therefore lack an interpretation basis that is connected with other linguistic dimensions.7 The computational dialectometry methods are numerous and are currently considered most adequate for reasons I will explain later on in this section. The foundations of digital automated dialectometry were established by Séguy () with his analytical method for calculating the linguistic differences between varieties of Gascon. The comparison is based on an algorithm which classifies data as identical or non-identical. The sum of the measured distances between two varieties matches their linguistic distance. The visualization of the classification analysis is conducted through lines of various types (bold/non-bold, dotted/non-dotted, etc.), which divide the region according to the linguistic differences of the varieties. As a counterpart, Goebl (, ) has calculated the similarities between varieties from Italy and Southern Switzerland. Even though the results of the calculation of Séguy and Goebl have the merit of being objective, they lack refinement, because their technique excludes distance graduation. The main computational methods based on the frequency of linguistic variants are the “Corpus Frequency Method” (Hoppenbrouwers and Hoppenbrouwers , ) and the “Frequency per Word Method” (Nerbonne and Heeringa , ). The basic principle of the first approach is that the degree of difference/similarity between two varieties is derived from comparing the frequency of the marked linguistic features of their variants. The problem with this approach is that the entity “word” is not considered as a linguistic unit. However, this obstacle is removed by the second approach, which assigns to words the status of “units” functioning as such. Nevertheless, the two classification tools do not take into account the order of the phonic units in the sequence. The “Levenshtein distance”, on the other hand, allows incorporating the parameter of sequential ordering of phonic units in the classification, which makes it more appropriate than other digital/numerical methods. This tool has been introduced in dialectometry by Kessler (), who has applied it to a corpus of Irish Gaelic. The Levenshtein distance measure corresponds to the numerical value of the lowest cost of operations (insertions, deletions, and substitutions) needed to convert a string of characters into another (Kruskal ). One of the most commonly employed techniques of comparison is the “phone string comparison”, in which all operations have the same cost, regardless of the degree of affinity between the phonic units: the pair [t, d] has the same cost as the pair [u, t] and [u, u:]. Yet, with the technique of “feature string comparison”, phonetic features of phonic units can be compared: the cost of the pairs [u, t] and [u, u:] is not equal because the phonetic affinity between the phonic units of [u, u:] is greater than that of [u, t]. 6 A significant critique of this method is that it cannot completely exclude some subjectivity, because isoglosses might be chosen, a priori, according to the linguistic borders they yield (Goossens ). 7 Although the frequency of the variants compared is taken into account (Kocks ), this approach does not seem to be the most appropriate (Heeringa : –).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
.. Dialectometry analysis and African languages: the case of Berber lexis Of the different existing dialectometry approaches, I prefer the computational methods because they allow handling large data corpora with a certain ease, while ensuring the accuracy and consistency of the analyses. These aims can be achieved because: – – – –
distances and frequencies are measured automatically; data are classified digitally; mapping can be assisted by computer; statistical analyses can be made and displayed automatically.
The dialectometry analyses that I present in this section were performed with the free software of Kleiweg (RuG/L).8 In order to complete a displayed dialectometry analysis, all the procedural steps summarized in Table . below are indispensable:
Table . General procedure of computational dialectometry analysis Step 1
linguistic atlas = georeferenced data source
Step 2
data matrix
Step 3
distance matrix
Step 4
analysis
Step 5
visualization
... The linguistic atlas of the Rif as a data source The Rif is the region of North Morocco stretching from the Strait of Gibraltar in the west to the Algerian frontier in the east. The Rif-Berber varieties (Tarifit) belong to the northern Berber languages and thus are part of the large Afro-Asiatic language phylum. The Berberspeaking area of the Rif is delimited as follows: – in the west, by the varieties of the Ktama tribe (the so-called Senhaja varieties); – in the south, by the koine of Gersif, which is the ultimate geographic point where Rif-Berber is spoken before reaching the corridor of Taza; – in the east, by the varieties of Iznasen, which have spread to the regions of Arabicspeaking varieties on the Morocco-Algerian border. The lexical data compared and classified in this study are collected from the Atlas linguistique des variétés berbères du Rif (ALR; Lafkioui , https://atlasrif.wordpress.com/). The digital data corpus consists of lexemes regarding the human body (maps –), kinship (maps –), animals (maps –), colors (maps –), numbers (maps –), as well as a subset of various nouns and verbs (maps –). Of these lexemes, eleven have only one variant per variety; the remaining lexemes display the co-occurrence of multiple variants for each lexeme. 8
See http://www.let.rug.nl/~kleiweg/L04
A
C
B
D
F
E
124 135
50
G
38
H 3934
61 43
45
44
29 40
30 15 16
21
N
13 1 2 3
12 10 8
4 5
27
22 7
48
69 65 67
119
116 117 134
103 78
104 100
110
106 138
62
P
Q
101 105
87
L 131
122 123
28 25
121
97
95
88
130
71
70 89
O 92
23 9
26 24
11
68
84 81
132 120
K
R 107
108
113
98
91
94
6
93 90
99
96 109
S
T
U
102
W
V
X 139
141
140
. Map of the georeferenced survey points of the Rif (Lafkioui )
112 114 111 115
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
20
75 79
118
82
83
73
74 76 77
46
47
80 85
51
6463
41
86
127 125 128
129
72
66 19
14
M
52
58 59 54 57 53
49
42
18 17
60
J
56
55
36
31 35 37 33
32
126
I
133
137
136
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
Due to the completion of the automated ALR, the data obtained from it are already in digital format, which has rendered the onerous task of digitizing unnecessary. However, an adaptive conversion to the software RuG/L (Kleiweg) was necessary. The ALR also offers a precise digital map of the Rif region (see Figure .), which is essential for the visualization of the dialectometry analyses, except for the dendrogram. In total, georeferenced points—belonging to Rif tribes—were selected from a group of localities in the Rif according to their degree of linguistic variation (Lafkioui ).9
... Data matrix of Rif-Berber lexis The data matrix is composed of digital lexical excerpts from the ALR (Lafkioui ) converted following the format of the software RuG/L (Kleiweg). Table . below gives a small sample in digital format of the ALR (Mapinfo Professional format):
Table . Data excerpt in encoded format of ALR
9
The survey points were selected on the basis of the principle of equidistance dividing the inquiry field into several grids to which were assigned points that could match with localities on the field. The
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
... Distance matrix of Rif-Berber lexis This section contrasts the three most commonly employed digital comparison techniques: the binary distance (Hamming algorithm), the “Gewichteter Identitätswert” distance (weighted identity value), and the Levenshtein distance. I will apply these techniques on the Rif-Berber lexical corpus to test their validity and to select the most appropriate to Berber. Each distance measurement allows acquiring precise numerical values derived from the linguistic comparison between the varieties of the Rif area. These values make up the distance matrices (symmetric matrices N N, N = sum of varieties), whose configuration differs depending on the adopted algorithm. .... B (B) The binary distance is used to classify lexical units as being identical or non-identical: comparison of type -; = resemblance and = difference. Table . presents an excerpt from the binary distance matrix of the lexeme ‘heel’ (ALR, map ):
Table . Excerpt from the binary distance matrix of the lexeme ‘heel’
.... “G I” (GIW) The “Gewichteter Identitätswert” distance deviates from the binary distance in that the frequency of the lexical variants is considered in the comparison: low-frequency variants weigh heavier than high-frequency variants. The distance obtained by this technique varies between and ; { d }. Table . presents an extract from the distance matrix of the lexeme ‘heel’.
greater the variation was, the more the grids were reduced. The 452 locations selected for this research were for the most part chosen so that they could, a priori, indicate linguistic borders. This selection mainly stemmed from the scientific and empirical knowledge of the investigator about the different varieties spoken in the Rif area.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
Table . Excerpt from the GIW distance matrix of the lexeme ‘heel’
.... L (L) The distance values resulting from a Levenshtein-based comparison—an algorithm taking into account the sequential order of phonic units composing lexemes—fluctuate between and ; { d }, as shown in Table ..
Table . Excerpt from the Lv distance matrix of the lexeme ‘heel’
These values result from the selection of the least costly calculation to transform a lexical unit—as a string of phonic units—into another. Table . depicts the lowest cost of operations which allow modifying the string awrez into inerz (‘heel’). The lowest cost of operations amending awrez into inerz is , which implies that the distance between these two lexemes is / ( being the total number of features);
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Cost of operations allowing modification of awrez into inerz (heel)
i n e r z
0 0.5 1 1.5 2 2.5
a
w
r
e
z
0.5 1 1.5 2 2.5 3
1 1.5 2 2.5 3 3.5
1.5 2 2.5 3 2.5 3
2 2.5 3 2.5 3 3.5
2.5 3 3.5 3 3.5 3
consequently, the Levenshtein distance is %. These calculations are based on operations that cost . for an insertion or deletion and for a substitution. Table . illustrates this calculation technique.
Table . Example of calculation of Lv distance for modifying awrez into inerz (heel) Tamadda Tizi Lv distance
a i 1
w n 1
r 0,5
e e 0
r 0,5
z z 0
3/5 * 100 = 60 %
... Numerical dialectometry analyses of Rif-Berber lexis From the distance matrices, numerical comparative analyses of Berber lexis can be accomplished through two techniques: cluster analysis and multidimensional scaling. Cluster analysis (CA) consists of regrouping data by reducing the distance matrix by means of various algorithms. Using Kleiweg (RuG/L), I have implemented the Ward algorithm (minimum variance), which is generally regarded as one of the most appropriate algorithms for this type of analysis. On the other hand, multidimensional scaling (MDS) is: a technique that, using a table of differences, tries to position a set of elements into some space, such that the relative distances in that space between all elements corresponds as close as possible to those in the table of differences. (Kleiweg, RuG/L)
... Visualization of dialectometry analyses of Rif-Berber lexis Classification by clustering necessarily uses a dendrogram for its display. A dendrogram is a complex ranking structure, usually in color, whose branches represent the linguistic varieties. It can be matched with a digital map, resulting in a geolinguistic map that shows the distribution of linguistic varieties depending on the linguistic differences and
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
the selected classification criteria. In contrast, analyses by MDS directly offer maps on which the relative linguistic variation is gradually represented by different colors. Figure . below shows the lexis classification obtained through Lv distance corresponding to an asymmetrical configuration of seven clusters which are structured into two major clusters distanced from one another by .. The matching dendrogram shares the same linguistic main delimitation (between groups and –) with the other dendrograms based on the Bin and GIW algorithm (for their visualization, see Lafkioui , b). This observation is corroborated by the CALv maps displayed in Figure ., of which the two-cluster map clearly indicates the most distinctive linguistic boundary. It is important to note that the CALv map (Figure .) shows a distribution of the varieties similar to the CABin distribution (Lafkioui , b), even though the composition of their respective dendrograms is divergent.
.. Contrastive results Because of its accuracy, the MDS method is most appropriate for dialectometry analysis of Berber lexis. Accordingly, it provides a yardstick against which other dialectometry methods can be contrasted (Figure .). Of the cluster analysis classifications, CABin and CALv best join the distribution maps displayed by MDS (seven groups). Moreover, the CALv classification shows a further refinement because it takes into account the phonic variation of the units as much as their arrangement in the lexemes (Lafkioui , b). However, any analysis based on Lv distance (CA as well as MDS) ignores the existence of the hierarchy between the phonic units (phonetic units = phonemic units), unless various weights are granted to them through a specific parametering. The CA classification has the benefit of precisely indicating significant linguistic boundaries. The CAC maps (Composite Analysis Cluster map; Figure .) designate these boundaries by dark lines. Compared to the distinctive boundaries drawn by the dendrogram and CA maps of Figures . and ., the principal linguistic delimitation of the CAC map of Figure . is drawn farther to the west. It is important, nevertheless, to note that the CAC maps do not seem suitable for displaying the classification of Rif-Berber lexis because of the difficulty of interpreting the data, due to their rather chaotic representation (Figure .).10 Dialectometry analysis, especially when based on the Levenshtein algorithm, is most valuable for the geographical classification of linguistic varieties and when the phone strings examined are tokenised in pairwise alignments. This procedure is feasible on the level of languages (e.g. Rif-Berber) as well as language families (e.g. Berber) and phyla (e.g. Afro-Asiatic) but is only productive—because it requires a great deal of time and energy—when applied to a considerable corpus. An interesting development would be to build a common phonological parameter structure for each phylum, drawn from the micro-analysis of its various languages, which might lead to a general linguistic
10
Kleiweg (RuG/L04) offers some alternatives to the Ward algorithm, which seems to be causing visual disorder by which CAGIW maps are most affected.
1
Bni Wkil Afsu Saka Hasi Berkan Ayt Hidra Driwc Ayt Dawed Ayt Muhend U ’Abdellah
2
Tafughalt Tawrirt Bni Buzeggu Zayyu Qaryat Arkman-Mellah
3
Ayt Hmu ’Mar Suyyah Bured Ayt Hazem
Amejjaw Tazaghin Iyar n Tzaxt Sidi Hsayn La’zib n Sidi C’ayb u Meftah Ayt Bu Ya’kub Ayt Mayyit Ayt Tayar-Sidi Dris Bu’diyya Budinar Ayt Yarur Buhfura Ben Teyyeb Tariwin Ayt ’Azza Iyarmawas-Friha Talamghacht Iyarmawas-Ijarayen Midar Alto-Icennuden
4
Aghir Umedgha Imezzujen-Ayt Nadur-Ice’aren Imezzujen-Ifarxanen-I’emranen Ayt Sidar-Rabe’ n Trat Burtwal Xadeb Had n Ayt Sicar Icemraren Imehhuten Tibuda
5
6
7
Aduz Tawssart Tara Yusef Izemmuren Tafnessa Ayt Bu’iyyach Mulay ’Abd Qader Ayt Hicem-Idij Imzuren Ayt Hdifa-Tazrut Tizi Aghir Hmed Wad Mahkim Sidi Bucetta Ufis Tufist-Imuruten Igzennayen A’raben Tamadda Mazuz Ssaqya Aghennuy Tizirt Lmexzen
’Arwi Tizdudin Iqedduren Gersif ’Ayn Zura-Ayt Heqqun Wlad Melluk Mezgitam-Ayt Hmed Hbircat Ifettuchen Ayt Waklan Thgasrut Zegzel Berkan Cabo de Agua-Imrabden/Ayt Yusef Acnur- Tighezdratin I’etmanen-Rbardun Ajdir-Tara Tazeghwaxt Ayt Mhend Tizi Wesri - Ayt Ziyyan Dar Kebdani Isarhiwen Meqdada Amzzawru Raba’ n Trugut Ayt Marghnin-I’ewwaden Ayt Ta’ban Ayt Ya’qub Ifasiyen Ayt Buhidus Yarzuqen Hammuda Mhajar Tawarda Igarduhen Tlata Uzlaf Tawrirt n wuccen La’zib n Midar Raq Azirar
4
7
At Buyefrur-Zghenghen Selwan Ayt Nsar Bumiyya Tifasur Cabo de Tres Forcas-Qabddenya Ibuyqeddiden Ihninaten Imezzujen-Ifarxanen-Ijuhraten I’zanen-Sidi Lehsen
5
4
6 3
1
2
Bades Taghza Bughembew Maya Ayt ’Abdellah-Icibanen Tmasint-Zawiyet n Sidi ’Isa Timerzga-Tafsast Ayt Qamra-Rwaz Mnud-Ayt Hicem Azru n tili Bni Budjay Asammar Ayt Hmed Tazrut L’ ars Bu’ di Iharunen Azila Luta Lqel’a Tarya Wersan Asammer Ssahel
0
2
4
6
8
. Dendrogram vs. CA map—Lv—all lexis Note: For a full-colour version of this figure, please see the plates at the end of the chapter.
.
. CALv maps—seven clusters vs. three clusters vs. two clusters—all lexis
CALv
MDSLv
CACLv
. CALv vs. MDSLv vs. CACLv maps
Bin
GIW
Lv
. CAC—Bin vs. GIW vs. Lv maps—all lexis Note: For full-colour versions of the figures on this page, please see the plates at the end of the chapter.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
classification grid of Africa that allows comparing a substantial amount of features in a structured way, combining phonology, grammar, and lexicon.
. L
.................................................................................................................................. The term “linguistic geography” refers to the analysis of contact-induced language change and diffusion. It aims at delimiting linguistic zones geographically, on a micro- and macrolevel, by reconstructing the long-standing affiliation patterns ensuing from non-genetic contact situations. Accordingly, a linguistic area is identified when the properties of a linguistic system specific to the region in question cannot be explained by genetic inheritance, universal principles, drift, or chance. Well-known examples of established linguistic areas, i.e. linguistic convergence zones, are, for instance, Standard Average European (Haspelmath ) and the Mesoamerican linguistic area (Campbell et al. ). From an African perspective, the key question that arises from this approach is the following: is Africa shaped areally? Not only is the answer to this question enlightening for Africa’s linguistic classification— areally, (inter)genetically, and typologically speaking—but it also allows for a better understanding of its cultural, historical, and ecological developments. Given the intricate and fused nature of these classification types, answering this question is a particularly tricky task. However, some seminal research on the areal classification of the languages of Africa has been carried out, but there is still a lot of work ahead, both on a macro-level, considering Africa as one defined area, and on a micro-level, which is confined to a specific African subarea and its subsystems. The delayed interest in this research field is mainly associated with the lack of detailed comparative analyses of genetically related languages, which are essential to areal as well as intergenetic classifications. One of the first scholars to take an interest in the study of African linguistic geography from a macro point of view was Greenberg (, ). He distinguished Africa’s linguistic zones by means of the areal patterns of certain properties (mainly lexical polysemies, e.g. the employment of the same lexeme to express the notion of ‘eat’/‘defeat’/‘have sexual intercourse’ or to express the notion of ‘meat’/‘wild animal’), thus providing important accounts that have served as a basis for further studies (e.g. Heine and Nurse ). Although Greenberg’s classification work has been criticized as unclear and inaccurate with regard to the distinction between areal and genetic parameters (Güldemann ), it has the merit of drawing attention to some of Africa’s linguistic traits, the genetic origin of which he also reconstructed. Larochette’s areal-typological contribution () is perhaps less well known than Greenberg’s but is as important to the study of African linguistic geography. Most of his proposed features concern the languages of Congolese Bantu (Kikongo, Luba, Mongo), Mangbetu (Central Sudanic), and Zande (Ubangi), and are still employed to establish and expand other classifications (Meeussen ; Heine ; Heine and Leyew ). Even if there is reason to believe that Africa constitutes (to a certain extent) a linguistic area, scholars have failed to find a significant cluster of features that are attested solely on
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
African territory. However, a considerable effort to classify Africa, and more precisely subSaharan Africa, has been made by Heine () and lately by Heine and Leyew () on the basis of eleven typological features that have been the subject of a survey of languages belonging to the four African phyla, namely Afro-Asiatic, Nilo-Saharan, Niger-Congo, and Khoisan. Because of the poor results obtained from the isopleth mapping method, which consists of delimiting areas of languages with the same number of features regardless of their nature (van der Auwera : ), Heine and Leyew considered typological resemblance on the basis of not only the presence of a feature but also its absence. Their findings corroborate the tendency to regard the Sudanic belt (West-Central Africa) as the area that contains most of Africa’s linguistic characteristics, cutting across the boundaries of three genetic language phyla: Afro-Asiatic, Nilo-Saharan, and Niger-Congo (Westermann ; Greenberg , ). Fundamental properties distinguishing this major convergence area, such as the ‘meat’/‘animal’ polysemy and labial-velar stops, indicate Niger-Congo as its nucleus and, hence, confirm Greenberg’s () reconstructions. However, Güldemann’s () study of the Macro-Sudan belt, a northern sub-Saharan area extending from West Africa to Ethiopia in the east, disputes the validity of this macroclassification of Africa, pointing to the danger of carrying out areal research from a Sudanic-biased or even sub-Saharan-biased perspective. Indeed, it is true that North and North-East Africa are typically excluded from this predisposed classification trend because of their dissimilar typological profile related to their common prevailing Afro-Asiatic genetic background. And yet these zones provide accounts of one of the indisputably typological linguistic features of Africa. This is the marked nominative, a feature which is barely attested outside Africa but is present in most case languages within it, including those of East Africa (König , ). In these systems, it is the nominative that is functionally marked with reference to the accusative, even if the languages in question may differ as regards morphosyntactic marking procedures, as shown in the following examples: ()
ὲ-sak-ì a-pà a-k-ìmuj. -want- father: food: ‘Father wants food.’
[Turkana (Eastern Nilotic, Nilo-Saharan)] (Dimmendaal : )
The nominative is indicated by a floating low tone, whereas the accusative is the same as the basic form, also called “absolutive” or “absolute”. All other cases are marked by fixed tonal schemes. ()
ɪ-swæ
ʊ-rgæz æmæn .-drank: man- water- ‘The man drank water.’
[Tamazight (Berber, Afro-Asiatic)]
The nominative is encoded by the morphological marker ʊ- (dependent state marker) and a continuous intonation unifying the noun with the preceding verbal syntagm, while the accusative is encoded by the absence of these morphemes and so matches the unmarked form, i.e. æmæn (accusative, independent state) vs. w-æmæn (nominative, dependent state). An exception to the above-mentioned tendency to bias in the study of African linguistic geography is perhaps the contribution of Clements and Rialland (), who present a
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
NORTH
SUDANIC EAST
CENTER
RIFT
SOUTH
. The six phonological zones of Africa (based on Clements and Rialland )
convincing areal classification of Africa as a whole that might be connected with contactinduced developments going back to prehistory. They propose to divide Africa into six major zones—North, East, Sudanic, Center, South, and Rift Valley—according to the presence or absence of the following specific phonological features: () absence of P-sounds, () emphatic consonants, () non-tonal prosody, () labial flaps, () labial-velar stops, () implosives, () nasal vowels, () two series of high vowels, () three+ tone levels, () “lax” question markers, () ejective and aspirated stops, () clicks, and () slack voiced stops. Although no single phonological characteristic common to the entire African continent is found, a certain areal regrouping is put forward, even if the dividing lines are considered permeable and unstable (Figure .). From the six clusters, the first three display a clear areal demarcation cutting across genetic boundaries: the North zone is mainly defined by the properties (), (), and (); the Sudanic belt is characterized by the properties (), (), (), (), (), (), and (); the South is clearly distinguished by the properties (), (), and (). The other three areal zones are less well delimited: apart from a shared Afro-Asiatic heritage, the East zone shows a prevalent presence of implosives, ejectives, and tone systems; besides the Common Bantu properties, the Center zone shares the vowel harmony system; as to the complex Rift Valley zone, no conclusive phonological evidence diffused over the entire area and so intersecting the genetic phyla is attested, which definitely calls for further study.11 On a mid- and micro-level, a considerable amount of research on African areal classification has been accomplished, although not always backed up with conclusive data. The Ethio-Eritrean area—also called the Ethiopian area—is considered the most defined—if not 11
For more about the Rift Valley area, see Kießling et al. ().
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
the only—convergence zone of Africa (e.g. Greenberg ; Ferguson , ; Heine ; Appleyard ; Zaborski , ; Hayward ; Crass ; Bender ). This area is demarcated and expanded differently according to the traits used for analysis. Heine’s () constituent order classification, for example, enlarges the area by including Lake Chad, while Tosco’s () interest in the study of the verb paradigms limits it to the Gurage–Sidamo area. With the exception of Tosco (), the ongoing debate generally supports the hypothesis of the Ethio-Eritrean convergence area, which has recently been extended to the entire Horn of Africa and certain parts of Sudan, Tanzania, and Uganda (Zaborski ).
. C
.................................................................................................................................. Language is continuously modulated through innovations that may surface in structurally layered and causal formations dictated by properties that are system-internal and/or system-external. It goes without saying that social factors such as prestige and community solidarity are important for the diffusion of variants, which may follow different adoption and acquisition scenarios. However, the Berber data concerning the phonological and morphological phenomena produced by vocalization of the liquids /r/ and /r: / in Central Rif-Berber (North Morocco) and examined in Lafkioui (c, b, b) reveal the importance of certain system-internal factors in the diffusion of new linguistic items. For example, economy and code conformity are important functional parameters for the dispersal of the vocalized liquids in the Central Rif area. Moreover, structure adequacy and systematization/generalization of specific morphological patterns play a significant role in the success of their diffusion across the Rif-Berber varieties. Contrary to the language evolutionary claim that only social factors are responsible for variant selection (Milroy and Milroy ; Milroy : –; Croft : –, ), Lafkioui demonstrates how functional and social (socioeconomic) factors can interact in the selection and hence diffusion of language forms, and how in some cases, as in the restructuring of the verbal paradigm, system-internal properties may prevail. Another interesting finding related to linguistic variational research is that language change is gradual not just on an extralinguistic level (geographical and social variation) but also on a linguistic one. Therefore, it is important to consider the continuous selection process of variants not exclusively from a social perspective (“propagation” of Croft : , ) but also in terms of how the variants are formally and functionally incorporated into ever-changing linguistic structures (Lafkioui c, b, b). It should also be noted that there exists an important correlation between linguistic variation and ecology (besides physical geography and technology). The ecological risk hypothesis states that in non-industrial societies, such as certain West African ones (Nettle ), social networks serve as a kind of protection against the risks of agriculture (Scott ; Braun and Plog ).12 In this approach, geolinguistic diversity is employed as historical evidence to explain the particular social and economic networks formed in 12
The bigger the provision problem, the larger the social network.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
order to adapt to the ecological situation at hand and so reduce the risk of undernourishment. It would be of great interest to consider things the other way around, that is, to study the effects of ecology on language, considering linguistic diversity also as an outcome of people’s conduct rather than just a consequence of physical geography. Unfortunately, this field of study is by and large lacking in African geolinguistics.13
R Appleyard, D. (). ‘The relative verb in focus construction: an Ethiopian areal feature’, Journal of Semitic Studies : –. Basset, A. (). Études de géographie linguistique en Kabylie. Paris: Leroux. Basset, A. (). Articles de dialectologie berbère. Paris: Klincksieck. Bastin, Y. (). ‘Classification lexicostatistique des langues bantoues ( relevés)’, Le Bulletin des Séances de l’Académie Royale des Sciences d’Outre-Mer : –. Bender, M. L. (). ‘Northeast Africa: a case study in genetic and areal linguistics’, Annual Publication in African Linguistics : –. Braun, D. P., and Plog, S. (). ‘Evolution of tribal social networks: theory and prehistoric North American evidence’, American Antiquity : –. Brunk, K., Ibriszimow, D., and Jungraithmayr, H. (). ‘L’atlas linguistique d’Afrique sahélosoudanienne (Alass): À la recherche des isoglosses intergénétiques dans le domaine zoonymique’, in C. Baroin and J. Boutrais (eds.), L’homme et l’animal dans le bassin du lac Tchad. Paris: IRD, –. Campbell, L., Kaufman, T., and Smith Stark, T. C. (). ‘Meso-America as a linguistic area’, Language : –. Chambers, J. K., and Trudgill, P. (). Dialectology. Cambridge: Cambridge University Press. Clements, N., and Rialland, A. (). ‘Africa as a phonological area’, in B. Heine and D. Nurse (eds.), –. Crass, J. (). ‘Ejectives and pharyngeal fricatives: two features of the Ethiopian language area’, in B. Yimam, R. Pankhurst, D. Chapple, Y. Admasu, A. Pankhurst, and B. Teferra (eds.), Ethiopian Studies at the End of the Second Millennium: Proceedings of the th International Conference of Ethiopian Studies, – Nov. . Addis Ababa: Institute of Ethiopian Studies, –. Croft, W. (). Explaining Language Change: An evolutionary approach. Harlow: Longman. Daan, J. C. (). ‘Dialects’, in D. R. Preston (ed.), Handbook of Perceptual Dialectology. Vol. . Amsterdam and Philadelphia, PA: John Benjamins, –. Daan, J. C., and Blok, D. P. (). Van Randstad tot Landrand; toelichting bij de kaart: Dialecten en Naamkunde. Amsterdam: Noord Hollandsche Uitgevers Maatschappij. Dimmendaal, G. J. (). The Turkana Language. Dordrecht: Foris. Ferguson, C. A. (). ‘The Ethiopian language area’, Journal of Ethiopian Studies : –. Ferguson, C. A. (). ‘The Ethiopian language area’, in M. L. Bender, J. D. Bowen, R. L. Cooper, and C. A. Ferguson (eds.), Language in Ethiopia. Oxford: Oxford University Press, –. Fourquet, J. (). ‘Linguistique structurale et dialectologie’, in E. Karg-Gasterstädt and J. Erben (eds.), Fragen und Forschungen im Bereich und Umkreis der germanischen Philologie: Festgabe für Theodor Frings zum . Geburtstag. Berlin: Akademie-Verlag, –. Garde, P. (). ‘Réflexions sur les différences phonétiques entre les langues slaves’, Word : –. Gilman, C. (). ‘African areal characteristics: sprachbund, not substrate?’, Journal of Pidgin and Creole Languages : –. Goebl, H. (). Dialektometrie: Prinzipien und Methoden des Einsatzes der numerischen Taxonomie im Bereich der Dialektgeographie. Vienna: Verlag der österreichischen Akademie der Wissenschaften. 13
Except perhaps for Brunk et al. ().
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
Goebl, H. (). ‘Probleme und Methoden der Dialektometrie: Geolinguistik in globaler Perspektive’, in W. Viereck (ed.), Proceedings of the International Congress of Dialectologists. Stuttgart: Franz Steiner, –. Gooskens, C. (). On the Role of Prosodic and Verbal Information in the Perception of Dutch and English Varieties. Ph.D. thesis. Nijmegen: Katholieke Universiteit. Goossens, J. (). Die niederländische Strukturgeographie und die ‘Reeks Nederlandse Dialectatlassen’. Amsterdam: Noord-Hollandsche Uitgevers Maatschappij. Goossens, J. (). Strukturelle Sprachgeographie: Eine Einführung in Methodik und Ergebnisse. Heidelberg: Winter. Goossens, J. (). Inleiding tot de Nederlandse Dialectologie. Groningen: Wolters-Noordhof. Greenberg, J. H. (). ‘Africa as a linguistic area’, in W. R. Bascom and M. J. Herskovitz (eds.), Continuity and Change in African Cultures. Chicago, IL: Chicago University Press, –. Greenberg, J. H. (). ‘Some areal characteristics of African languages’, in I. R. Dihoff (ed.), Current Approaches to African Linguistics . Dordrecht and Cinnaminson, NJ: Foris, –. Grosse, R. (). ‘Strukturalismus und Dialektgeographie’, Biuletyn Fonograficzny : –. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), –. Guthrie, M. (–). Comparative Bantu: An introduction to the comparative linguistics and prehistory of the Bantu languages. vols. Farnborough: Gregg International. Haspelmath, M. (). ‘The European linguistic area: Standard Average European’, in M. Haspelmath, E. Konig, W. Oesterreicher, and W. Raible (eds.), Language Typology and Language Universals: An international handbook. Vol. . Berlin: Mouton de Gruyter, –. Hayward, R. J. (). ‘Is there a metric for convergence?’, in C. Renfrew, A. McMahon, and L. Trask (eds.), Time Depth in Historical Linguistics. Vol. . Cambridge: McDonald Institute for Archeological Research, –. Heeringa, W. (). Measuring Dialect Pronunciation Differences Using Levenshtein Distance. Ph.D. thesis. Groningen: Rijksuniversiteit. Heine, B. (). ‘Kalenjin glottochronology: preliminary hypotheses’, Mila : –. Heine, B. (). ‘Historical linguistics and lexicostatistics in Africa’, Journal of African Languages : –. Heine, B. (). ‘Language typology and convergence areas in Africa’, Linguistics : –. Heine, B., and Leyew, Z. (). ‘Is Africa a linguistic area?’, in B. Heine and D. Nurse (eds.), –. Heine, B., and Möhlig, W. J. G. (). Language and Dialect Atlas of Kenya. Vol. I: Geographical and Historical Introduction. Berlin: Dietrich Reimer. Heine, B., and Nurse, D. (eds.) (). A Linguistic Geography of Africa. Cambridge: Cambridge University Press. Hoppenbrouwers, C., and Hoppenbrouwers, G. (). ‘De feature frequentie-methode en de classificatie van Nederlandse dialecten’, Bulletin voor taalwetenschap : –. Hoppenbrouwers, C., and Hoppenbrouwers, G. (). De indeling van de Nederlandse streektalen: Dialecten van steden en dorpen geklasseerd volgens de FFM (feature frequentie methode). Assen: Koninklijke Van Gorcum. Kessler, B. (). ‘Computational dialectology in Irish Gaelic’, in V. Ginsburgh and S. Weber (eds.), Proceedings of the th Conference of the European Chapter of the Association for Computational Linguistics. San Francisco, CA: Morgan Kaufmann, –. Kießling R., Mous, M., and Nurse, D. (). ‘The Tanzanian Rift Valley area’, in B. Heine and D. Nurse (eds.), –. Kocks, G. H. (). Die Dialekte von Südostdrente und anliegenden Gebieten: Eine strukturgeographische Untersuchung. Ph.D. thesis. Groningen: Rijksuniversiteit. König, C. (). ‘Marked nominative in Africa’, Studies in Language : –. König, C. (). ‘The marked-nominative languages of eastern Africa’, in B. Heine and D. Nurse (eds.), –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Kruskal, J. (). ‘An overview of sequence comparison’, in D. Sankoff and J. Kruskal (eds.), Time Warps, String Edits and Macromolecules: The theory and practice of sequence comparison. Stanford, CA: CSLI Press, –. Lafkioui, M. B. (). Atlas linguistique des variétés berbères du Rif. Cologne: Rüdiger Köppe. Lafkioui, M. B. (). ‘Dialectometry analyses of Berber lexis’, Folia Orientalia : –. Lafkioui, M. B. (a). ‘Études de géographie linguistique berbère: variation géolinguistique et classification dialectométrique’, Le Bulletin des Séances de l’Académie des Sciences d’Outre-Mer : –. Lafkioui, M. B. (b). ‘Analyses dialectométriques du lexique berbère du Rif ’, in D. Ibriszimow, R. Vossen, and H. Stroomer (eds.), Études berbères IV. Essais lexicologiques et lexicographiques et autres articles. Cologne: Rüdiger Köppe, –. Lafkioui, M. B. (c). ‘Synchronic and diachronic linguistic variation as an indicator of language change and diffusion in Tarifit’, Studi Maġribini (Nuova Serie) : –. Lafkioui, M. B. (a). Variation et structuration linguistiques et sociolinguistiques en berbère du Rif. Cologne: Rüdiger Köppe. Lafkioui, M. B. (b). ‘How system-internal linguistic factors indicate language change and diffusion: a geolinguistic analysis of Berber data’, Dialectologia et Geolinguistica : –. Lafkioui, M. B. (). ‘Méthodologie de recherche en géolinguistique’, CORPUS : –. Lafkioui, M. B. (a). ‘The Rif Berber language continuum. An algorithmic geolinguistic study’, in T. Colleman, J. De Caluwe, V. De Tier, A.-S. Ghyselen, L. Triest, R. Vandenberghe, and U. Vogl (eds.), Woorden om te bewaren. Hulde-Album voor Jacques Van Keymeulen. Gent: Skribis, –. Lafkioui, M. B. (b). ‘Geolinguistic complexity in Berber. Structural and algorithmic perspectives’, Dialectologia et Geolinguistica : –. Larochette, J. (). ‘Overeenkomst tussen Mangbetu, Zande en bantu-talen’, in Handelingen van het XXIIIe Vlaams Filologencongres: Brussel – April , –. Martinet, A. (). ‘Structural dialectology’, Pakha Sanjam (special volume), article n . Meeussen, A. E. (). Possible Linguistic Africanisms. Fifth Hans Wolff Memorial Lecture. Bloomington, IN: Indiana University. Milroy, J. (). Linguistic Variation and Change. Oxford: Blackwell. Milroy, J., and Milroy, L. (). ‘Linguistic change, social network, and speaker innovation’, Journal of Linguistics : –. Möhlig, W. J. G. (). Die Stellung der Bergdialekte im Osten des Mt. Kenya: Ein Beitrag zur Sprachgliederung im Bantu. Berlin: Dietrich Reimer. Möhlig, W. J. G. (). ‘Dialektgrenzen und Dialektcontinua im Bantu-Sprachgebiet von Kenia: Zum Problem der Grenzfindung und Grenzgewichtung’, in J. Göschel, P. Ivić, and K. Kehr (eds.), Dialekt und Dialektologie. Ergebnisse des Internationalen Symposiums „Zur Theorie des Dialekts“, Marburg/Lahn, .–. September . Wiesbaden: Franz Steiner, –. Möhlig, W. J. G. (). ‘L’atlas linguistique du Kenya: méthodes et résultats’, in L. Bouquiaux, G. Guarisma, and G. Manessy (eds.), Problèmes de comparatisme et de dialectologie dans des langues africaines. Paris: SELAF, –. Moulton, W. G. (). ‘The short vowel systems of northern Switzerland’, Word : –. Mufwene, S. (). ‘From genetic creolistics to historical dialectology: ecological and population genetics perspectives’, American Speech : –. Nerbonne, J., and Heeringa, W. (). ‘Computationale vergelijking en classificatie van dialecten’, Taal en Tongval : –. Nerbonne, J., and Heeringa, W. (). ‘Computational comparison and classification of dialects’, Dialectologia et Geolinguistica : –. Nettle, D. (). ‘Language diversity in West Africa: an ecological approach’, Journal of Anthropological Archaeology : –. Rensink, W. G. (). ‘Dialectindeling naar opgaven van medewerkers’, Mededelingen der Centrale Commissie voor Onderzoek van het Nederlandse Volkseigen : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
Saussure, F. de. ( []). Cours de linguistique générale. Edited by T. de Mauro. Paris: Payothèque. Scott, E. P. (). Indigenous Systems of Exchange and Decision-Making among Small-Holders in Rural Hausaland. Ann Arbor, MI: Department of Geography Publications. Séguy, J. (). ‘La dialectométrie dans l’Atlas linguistique de la Gascogne’, Revue de linguistique romane : –. Sommer, G., and Vossen, R. (). ‘Dialects, sectiolects, or simply lects? The Maa language in time perspective’, in T. Spear and R. Waller (eds.), Being Maasai: Ethnicity and identity in East Africa. London: James Currey; Dar es Salaam: Mkuki na Nyota; Nairobi: EAEP; Athens, OH: Ohio University Press, –. Stankiewicz, E. (). ‘On discreteness and continuity in structural dialectology’, Word : –. Tosco, M. (). ‘The Northern Highland East Cushitic verb in an areal perspective’, in C. Griefenow-Mewis and R. M. Voigt (eds.), Cushitic and Omotic Languages: Proceedings of the Third International Symposium (Berlin, March –, ). Cologne: Rüdiger Köppe, –. Tosco, M. (). ‘Is there an “Ethiopian language area”?’, Anthropological Linguistics : –. van der Auwera, J. (). ‘Revisiting the Balkan and Meso-American linguistic areas’, Language Sciences : –. Vossen, R. (). Towards a Comparative Study of the Maa Dialects of Kenya and Tanzania. Hamburg: Helmut Buske. Vossen, R. (). ‘How to comprehend the obvious? Linguistic variation and dialect geography: the African experience’, Sprache und Geschichte in Afrika : –. Weijnen, A. (). ‘De grenzen tussen de oost-noord-Brabantse dialecten onderling’, Bijdragen en mededelingen der Dialectencommissie van de Koninklijke Nederlandse Akademie van Wetenschappen te Amsterdam : –. Weijnen, A. (). Nederlandse dialectkunde. Assen: Van Gorcum. Weinreich, U. (). ‘Is a structural dialectology possible?’, Word : –. Westermann, D. H. (). Die Sudansprachen. Hamburg: L. Friederichsen. Zaborski, A. (). ‘Ethiopian language subareas’, in S. Piłaszewicz and E. Rzewuski (eds.), Unwritten Testimonies of the African Past. Warsaw: Wydawnictwa Uniwesytetu Warszawskiego, –. Zaborski, A. (). ‘Ethiopian language macroarea’, Sprawozdania z Posiedze ń Komisji Naukowych Oddziału PAN w Krakowie : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. A classification is intended to reduce the complexity of reality and “to evolve order out of chaos” (Doke : ). As a consequence classifications exist in nearly every discipline, e.g. biology, economics, librarianship, and linguistics. In the field of African languages it is, obviously, much easier to handle the four language phyla established by the most widely accepted classification, i.e. Greenberg (), than to tackle African speech forms on a case by case basis—the number of languages that have been identified so far by students of African languages being on the order of ,. In contradistinction to other disciplines, however, linguists trying to classify languages find it difficult to specify what in fact they intend to classify with any degree of precision. This is due to the fact that definitions of the concept of “language” that have been formulated in the history of linguistics are at great variance. This is not the place to enter into the age-old discussion about the difference between language and dialect, but three arbitrarily chosen examples should suffice to illustrate the difficulties linguists face. Firstly, if we had to accept the definition given by Max Weinreich, that the difference between a language and a dialect is that the former has an army and a navy behind it,1 the number of African languages to be classified would be drastically reduced, and not simply because not all African states have access to the ocean. Furthermore, in every new edition of Ethnologue (Gordon ) the number of African languages increases, which is as remarkable as it is irritating—irritating because at the same time language death in Africa, deplored by linguists and social anthropologists, seems to be proceeding at an alarming 1
“A shprakh iz a diyalekt mit an armey un a flot” (Weinreich : ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
rate. A closer inspection of Ethnologue, however, reveals that the increase in the number of languages is mainly due to the fact that speech forms hitherto labeled as “dialect” have been raised to the status of “language”. Finally, the first volume of the Handbook of African Languages by André Basset is entitled La langue berbère, which suggests that there is, in the vast Berberophone area of North Africa, just one language. Most Berber scholars would nowadays recognize a considerable number of different languages, not just “variations dialectales” as Basset (: v) suggests. For purposes of the present chapter “language” will be assumed to be a speech form distinct from others; this speech form is used by a speech community as its medium of communication. It is fair to say that , and possibly more than , such speech forms presently exist in Africa.
. T
.................................................................................................................................. Classifications assess factors of similarity and distinctness. Languages displaying similarities can be classified into groups. A higher in relation to a lesser degree of similarity allows for further subclassifications; on the other hand, a lesser degree allows the integration of groups into larger classes of a higher order. In the course of history several types of classification have evolved. According to the methods applied, There are three methods of language classification which are of major significance: the genetic, the typological, and the areal. Of these the genetic is the only one which is at once non-arbitrary, exhaustive, and unique. . . . This is because genetic classification reflects historical events which must have occurred or not occurred. (Greenberg : )
What “genetic relationship” means has been defined by many scholars, e.g. John Lyons (: ): “the term ‘relationship’ is being used, as is customary in linguistics, to refer to a historical, or ‘genetic’, relationship. To say that two languages are related is equivalent to saying that they have developed from some earlier single language.” Leonard Bloomfield (: ) has the same thing in mind when he says: “we mean that these [genetically related] languages are later forms of a single earlier language.” Typological classifications, on the other hand, are arbitrary because they depend on the criteria selected for classification. Seen from the point of view of our own times, typological classification during the nineteenth and the first half of the twentieth century was a more or less convenient method of indexing certain linguistic facts. Typological classifications of that time: were concerned with grammatical features such as the presence or absence of grammatical gender, without particular regard to the particular morpheme by which this was indicated . . . One weakness of such an approach was that, while it was taken for granted that the classification was a genealogical one, the implications of this were not fully worked out. (Winston : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
However, since genetically related languages frequently display similar grammatical features, it follows that in many cases typological classifications are not completely independent of genetic classifications. In the early days, areal classifications in the field of African linguistics were often the result of first attempts to give an overview of the languages of certain areas. There are several examples of researchers interpreting the original language-geographic character in a more genetic way. A good example is Sigismund W. Koelle’s classification (), which was originally conceived as an areal one. Later scholars such as Diedrich Westermann (, ) and Joseph H. Greenberg () gave it a clearly genetic orientation (Dalby ). In a similar way Malcolm Guthrie’s classification of Bantu languages () became more and more “genetic” after initially being characterized by its author as “practical” or “geographic”.
. L , ,
.................................................................................................................................. Not surprisingly classifications were only possible after a minimum of reliable data on African languages was available. They can only be as accurate and as complete as these data allow. This process began in the course of the seventeenth century: The humanistic interests of the Renaissance and the broadening knowledge of the world as a whole resulting from the voyages of exploration had produced in Europe by the th century a general attitude of curiosity about man in all his varieties that had, as one of its facets, an interest in the languages of non-Western peoples. (Greenberg (: )
It is from these times that we have the first reliable grammatical descriptions2 of subSaharan African languages as well as first ideas of how to classify them. Initially, and because of good documentation, the languages of the Near East as well as those of Ethiopia could be used as a base for classificatory endeavor. A good example of such activity can be found in one of the German editions of Olfert Dapper’s monumental book on Africa. It shows that ideas about relationships between languages and the possibility of classifying them were formulated almost as soon as Europe came into contact with Africa and her languages. In this passage Dapper tries to make clear that Ancient Egyptian is quite different from other Semitic languages, offering the words for “father” and “son” in Hebrew, Chaldean, Arabic, Syrian, and Abyssinian as an example. He contrasts them with their Armenian and Old Coptic counterparts and asserts: “From this you can see that oriental [i.e. Semitic] languages, although they are closely related to each other, have no relationship with Egyptian.” Moreover, “another author has claimed relationship between 2
Here the grammar of Kikongo by Giacinto Brusciotto () and the descriptions of Ge’ez and Amharic by Hiob Ludolf (, ) should get first mention.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
all these languages based only on one or two words. It is widely known that there is no language which does not have some words which correspond to words in other languages” (Dapper : ; translation LG). Here we see that at the end of the seventeenth century it was possible for scholars to formulate conditions for the assumption of a linguistic relationship, to differentiate between the reasons why languages were similar, and to classify them accordingly. For Dapper and his contemporaries some similarities were more important for language classification than others: those that shared a common origin had more relevance for language classification than those which were judged to be accidental, and could thus be neglected. The rise of the transatlantic slave trade led to a general lack of interest in the languages of Africa—you need not care about the languages of people you treat as chattels. Consequently, further progress in descriptive as well as in comparative studies occurred only with the abolition of the slave trade at the end of the eighteenth and the beginning of the nineteenth century. That period also saw the rise of intensive missionary activities all over Africa which proved to be of extreme importance for African linguistics as a whole. Protestant missionaries proved to be remarkably successful in providing primary data on hitherto unknown languages and the linguistic situation in their particular fields. At that time, too, the results of the linguistic work of men such as Bruce, Park, Salt, Lichtenstein, and Barth were finding due recognition in the scientific world. At the same time collecting data on exotic languages became fashionable. Empress Catherine the Great of Russia, among others, supported the publication of multivolume editions of data from languages all over the world and collections such as Mithridates (Adelung and Vater –) were published. They contain data on quite a number of African languages, and the classification they present displays features which can still be found in modern classifications.
. T A E, ,
.................................................................................................................................. Around the turn of the eighteenth to the nineteenth century comparative Indo-European studies gained momentum, and the application of watertight methods in comparative linguistics replaced the sometimes fanciful, mostly theologically motivated search for “the original language of mankind”.3 Consequently, a broad interest in the prehistory of known European languages, scientific proof of their “genetic” relations and, more generally, an interest in forms of linguistic change, replaced speculative theories about the original language of humankind.
3
To give one example: Richard Bentley (–), a famous scholar of his time, wrote in his “Dissertation on the Epistles of Phalaris”: “we are sure, from the names of persons and places mentioned in Scripture before the deluge, not to insist upon other arguments, that Hebrew was the primitive language of mankind.”
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Comparative linguistics around the turn of the nineteenth century is a daughter of the “Enlightenment” and of the romanticism that was part of it. The most influential representative of “enlightened” philosophy in Germany, Immanuel Kant, had a profound influence on his contemporaries and subsequent generations; in his extremely popular anthropological lectures, he propagated a kind of cultural-historical classification of “races” in which Africans, who, according to him, used language mainly for magical purposes, had a very low status, despite a wealth of evidence known to him.4 It is mainly through him that the concept of “race” was introduced into the scientific discourse. Another important representative of this romantic zeitgeist, Johann Gottfried von Herder, maintained that there is a strong correlation between the language of a people and its culture. Together with Alexander von Humboldt he paved the way for a discipline, Völkerpsychologie (ethnopsychology), that pursued this correlation in its research. Three generations later it was still a strong influence on African linguists. Carl Meinhof was one of these. Thus, three cultural-historical streams stood at the beginning of African linguistic studies: missionary activities, the rise of comparative linguistics in the Indo-European and Oriental fields, and general anthropology. Thoughts about African languages in the course of the nineteenth century were strongly influenced by the ideas of Charles Darwin, especially after these ideas were transferred to the study of language by August Schleicher in an open letter to Ernst Haeckel, a strong supporter of Darwin’s ideas on the continent (Schleicher ). Haeckel, professor of biology and a colleague of Schleicher, wrote a preface to Wilhelm H. I. Bleek’s treatise on the “Origin of language” (Bleek ) in which he stated: As is generally known the peoples of South Africa, the Hottentots, Bushmen, Kaffers and others that are normally regarded as Negro tribes of the race that is designed as longheaded (dolichocephal) have remained on the lowest level of human development, and have distanced themselves the least from the apes. As for all their physical and moral properties this holds true also for their languages. (Haeckel : iv)
. T
.................................................................................................................................. This was the zeitgeist when the first continent-wide classifications of African languages were worked out. One of them deserves special mention, because it was taken—more or less tacitly—as the basis of many subsequent classifications: Koelle’s Polyglotta Africana, published in London in by the Church Missionary Society. It contains wordlists of more than items from some languages all over Africa, written down in a uniform transcription. Koelle’s language informants were freed slaves living in Freetown on the western margin of the continent. The criteria for grouping languages together were 4
Kant’s remarks about Africa appear in quite a number of his works; bibliographic details are given in Smidt (, ). Over the last few years there has been lively discussion of Kant’s role in forming the negative image of Africa. Smidt’s studies contain a summary of this discussion together with a wealth of relevant bibliographical sources.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
apparent similarity between speech forms and geographic proximity. Together with language data, Koelle provided geographical information about the area where the languages were spoken. The London Church Missionary Society, which had commissioned this volume, hoped to get as much information and as precise an idea as possible of the linguistic situation in the inner parts of Africa, with the aim of gaining insights into the individual languages in order to further missionary activities, i.e. to work as effectively as possible in the African field. The German missionary Ludwig Krapf, who published an “East African Polyglotta” () containing vocabularies of Bantu and some Cushitic languages,5 had similar objectives. Krapf understood the main objective of his work to be “apostolic, not scientific”, notwithstanding his—and for that matter also Koelle’s—enormous contribution to our knowledge of African languages. The missionary objectives behind classificatory work on African languages can be observed over and over again, e.g.: There is no need to call the attention of anyone to the importance of Bantu. Independently of its scientific interest, it is a key for opening one half of an immense continent to Christian Civilization. (Torrend : vii)
As a marginal note, it might be added that Meinhof (: ) criticized Torrend’s inclination towards speculation, because the latter had viewed the Bantu noun class prefixes as a mirror of the process of creation. This incursion into African linguistic studies on the part of missionaries was relevant for classification in so far as the practical needs of the mission (e.g. the usefulness of a particular language for Bible translation) played an important role in the work of later research workers such as Lepsius, Meinhof,6 Westermann,7 and Guthrie—even if some of them were not missionaries—and motivated such publications as Ethnologue. The second major factor in the development of African language studies and classificatory efforts was the rise of systematic comparative linguistics. However, it took several decades before the first comparative study of African languages saw the light of day, i.e. a study that was on a par with the methodological standards of the seminal handbooks of Indo-European studies that had appeared half a century earlier. In the meantime much comparative work, independent of work in the field of comparative Indo-European linguistics, resulted in a great number of classifications published in the second half/last quarter of the nineteenth century. Some of them deserve mention here, partly because they were influential beyond the time of their publication, and partly because the underlying methodology deviates from the methods used in our time. A very widespread prejudice about African languages prevalent among European intellectuals almost certainly hindered a breakthrough in historical African linguistics: the
5
In addition, Krapf had accurate insights into the interrelatedness of the Bantu languages (: III). Meinhof ’s enormous scientific output (about 600 publications) contains more than 140 titles that treat mission-related topics (cf. Meyer /). 7 Westermann, after working as an employee in the postal service, decided to become a missionary. In the course of his education, his extraordinary language skills became known, skills which resulted in him becoming professor of African linguistics in Berlin. 6
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
conviction that African languages lacked a history worthy of research, for the simple reason that most of them did not have a written tradition. The most important of these intellectuals was Wilhelm H. I. Bleek (–). His A Comparative Grammar of South African Languages appeared between and and corresponds to a contemporary view of a methodically adequate approach. Bleek had earned a Ph.D. in Bonn in and later became the librarian of Sir George Grey in Cape Town where he had access to nearly all known data about African languages. Accordingly he was able to enlarge the classification presented by Koelle (Bleek ). As a basis for his comparative studies he chose one particular Bantu language, Tswana, which for phonological reasons he regarded as the most archaic. In fact, this prevented him from correctly analyzing and interpreting certain sound shifts which had merged in Tswana but were still differentiated in other Bantu languages. In the years –, Friedrich Müller accompanied the Fregatte Novarra on a journey around the world. During that voyage he collected a tremendous wealth of primary linguistic data. His classification (Müller –), however, brimmed with racial overtones in so far as he classified languages according to the somatic features of the speakers. Because he did not believe in linguistic monogenesis, he thought it necessary to go back to a “higher principle” beyond language, i.e. race. Thus, he distinguished the languages of the woolyhaired, the sleek-haired, and the wavy-haired races as major linguistic units. This illconceived theory became influential by the fact that the natural scientist and philosopher, Ernst Haeckel (see section . above), propagated it in his important work Natürliche Schöpfungsgeschichte (A Natural History of Creation, ). In addition, Müller developed far-reaching theories about migrations that heavily influenced language classifications in the decades that followed. One hypothesis that was to exert an extremely negative effect was that the Hamitic immigrants into Africa were culturally superior to the autochthonous population (see Köhler : –). The next very influential classification of the languages of the continent was established by Karl Richard Lepsius (–), an Egyptologist who traveled to Egypt and Nubia in –. The result was a remarkable grammar of Nubian which contained a long introduction about the people and languages of Africa (Lepsius ). Commissioned by the Church Missionary Society, Lepsius also developed a system for a uniform transcription of African languages which came to be known as Lepsius’s “standard alphabet”. This was later expanded by Meinhof and demonstrated its usefulness in being used for the transcription of c. African languages. Lepsius’s classification is meant to be genetic but has many typological features. He regarded the Bantu languages as the main type of African languages, for he thought they were the most archaic. Another type he identified was the genderdenoting Hamitic languages.8 In between these two families were the languages of the Sudan, languages, in his opinion, of mixed speech forms (Lepsius : xvii), i.e. they were “formless” in so far as they had lost grammatical elements. He regarded the monosyllabic character of these languages as secondary. The hypothesis of mixed speech forms was taken up in the decades that followed by many scholars. Meinhof ’s “Hamitic theory” is a direct “descendant”
8 That Lepsius’s classification was not free from racist thinking follows from the fact that he excluded Hausa from his Hamitic family because the physical type of its speakers differs from the Hamitic type. Hausa is classified among the “central mixed Negro languages”.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
of Lepsius’s ideas in which he, Lepsius, combined cultural-historical and typological criteria to prove the genetic relationships of some African languages. To sum up: much of the classificatory work done during the nineteenth century was inspired or actually done by missionaries active in Africa. From the middle of the century on, systematic linguistics began to play a definite role. Racist ideas that had their origins in the philosophy of the Enlightenment and German romanticism rarely assumed major importance.
. T
.................................................................................................................................. The Bantu languages have always played a special role in African linguistics. In spite of the fact that they are spoken over a vast area, they share many similarities. African language groups with a comparable degree of similarity, such as subgroups of Benue-Congo or subgroups within Nilo-Saharan and Afro-Asiatic, are spoken in much more restricted areas. In the history of African linguistics (and of African language classification) Bantu languages were also important because they served as a test case for the application of the comparative method in Africa. Until the publication of Meinhof ’s modest volume, the first edition of his comparative phonology, only languages with a long(er) written tradition had been the subject of systematic comparative work. Meinhof ’s comparative work on Bantu (Meinhof [st edn] and ) caused a sensation in the scientific world and was vital for his getting the world’s first chair of African linguistics in Hamburg.9 Shortly after he started work there, he published a series of important lectures, “Hamburger Vorträge” (–), one of which contained a classification of the languages of Africa that was to set the state of the art in this field for the next forty years. African languages were subdivided into Hamitic languages in the north, Sudan languages in a belt that stretched above Central Africa, Bantu languages in the vast southern triangle and the click languages in South-West Africa. This classification was translated by Alice Werner from the School of Oriental and African Studies and had enormous influence in the English-speaking world. In fact, this classification preserved many of the ideas that had been proposed by nineteenth-century scholars. One particular publication of Meinhof should be mentioned because it was detrimental to Meinhof ’s reputation as a linguist: Die Sprachen der Hamiten (). This book was an unexpected deviation from his previous work in so far as it was not based upon sound comparative method. His intention was to classify languages spoken by ethnic groups with a similar lifestyle (cattle nomads, whose languages belong in all four major African language phyla as we see them today) as a single family. Because the speakers of these languages shared certain somatic features and because Meinhof held the view that these languages were imported into Africa through migration and that the immigrants possessed a superior culture compared with the autochthonous population, this work has strong racist undertones. However, it had a tremendous influence on anthropological, ethnological, and historical thinking, at least in Europe. It was only through the work on Fulfulde of 9
An evaluation of Meinhof ’s work can be found in Gerhardt ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
another linguist, August Klingenheben (/), that this hypothesis lost its linguistic foundation. Classifications based on other principles continued to appear; for instance that of Alfred Drexel (–), who chose certain cultural features (e.g. the form of houses—triangular vs. rectangular) as his basis for language classification. But these studies did not influence the mainstream of classificatory work. During the twentieth century there was a marked tendency to concentrate classificatory endeavors on smaller groups of languages. This coincided with the growing specializations of linguists and an ever-increasing output of publications on African languages. In this way language families were established on a lower level, but with a higher degree of precision; this precision was frequently arrived at by application of the comparative method and by working out regular sound correspondences. In the middle of the century two projects were started which culminated in the publication of two very different classifications of the languages of the continent. One was the four volumes of the Handbook of African Languages published by the International African Institute between and ; the second was Greenberg’s classification, which was published in two separate versions: Greenberg (/, ) and Greenberg (). In a certain sense they were typical of the different schools that produced them. The Handbook tried to bring together all available information on African languages and to record the accepted state of the art in language classification. However, this classification was extremely conservative. While the bulk of West African languages in the Handbook are classified into genetic subgroups, some case languages are lumped together in accordance with typological criteria. Languages of northern Nigeria and adjacent parts of Cameroon, for example, are classified as “Class languages” vs. “Non class languages”. The languages subsumed under these headings not only belong to different subgroups of Niger-Congo (Adamawa-Eastern, Kwa, and Benue-Congo) but also to different language phyla (AfroAsiatic, Niger-Congo). Nevertheless, the wealth of detailed information contained in these volumes means that they are still a useful source some sixty years after publication. With Greenberg’s classification North America made her first internationally visible appearance on the stage of African linguistics—and this with definite aplomb. Greenberg’s intention was to classify African languages according to linguistic criteria alone. This was in contrast with nearly all previous classifications. His critique of the existing classifications, therefore, was rather harsh—and provoked hostile reactions. Greenberg himself has stated in an article outlining the basis of his methodology that the methods he used could establish “valid hypotheses concerning genetic relationship among languages” (Greenberg : ) The method he propagated compared resemblances in both the sound and meaning of morphemes—lexical and grammatical—in as many languages as possible. This method has come to be known as “mass comparison”. While Greenberg’s classification initially met with a lot of hostility from European scholars, most of the basic assumptions underlying his classification have now been accepted. One of the first (positive) reviews compared Greenberg’s classification to an eighteenth-century map of Africa where the outline of the continent had finally been mapped correctly— despite a lot of white spots showing unknown areas. Generally, however, Greenberg’s classification has been shown to be correct—many of the imaginary white spots having been filled in. In the meantime, a great number of languages that were unknown in have been described and included in the picture. Furthermore, some major adjustments
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
have been made. In most cases the “bush-like” subgroupings have been restructured into a more tree-shaped form. It could, however, still be maintained with Thomason and Kaufman (: ) that the: Comparative method is a means by which a hypothesis of genetic relationship is demonstrated through the following kinds of evidence: () the establishment of phonetic correspondences () the reconstruction of a phonetic system () the establishment of grammatical correspondences, and () the reconstruction of grammatical systems to whatever extent is possible.
They continue: “Where one or more points ()–() cannot be attained, some doubts, ranging from minor to serious, can be cast on a hypothesis of genetic relationship.” Thus, some doubts concerning the reliability of the prevailing classifications seem to be justified. After Greenberg nobody has had the courage to present a new continent-wide genetic classification. Oswin Köhler () presents a single-handed attempt to evaluate the results of the competing classifications—most of which have not been mentioned here for lack of space. In general, interest in genetic classification has diminished in the last few decades. Instead, together with the development of new methods in general linguistics, a renewed interest in typological classifications has emerged, focusing on syntactic features such as word order. At the same time, the study of language contact has led to a completely new revaluation of the concept of language mixture and to new models of language relations. Considerations of the social conditions under which language mixture can occur are a major point of interest and will certainly lead to further coherent insights into language relations in Africa.
R Adelung, J. C., and Vater, J. S. (–). Mithridates oder allgemeine Sprachenkunde: Mit dem Vater Unser als Sprachprobe in bey nahe fünfhundert Sprachen und Mundarten. Mit Benützung einiger Papiere desselben fortgesetzt, und aus zum Theil ganz neuen oder wenig bekannten Hülfsmitteln bearbeitet von Dr. Johann Severin Vater. Berlin: Vossische Buchhandlung. Basset, A. (). La langue berbère. London: International African Institute. Bleek, W. H. I. (). The Library of His Excellency Sir George Grey, K.C.B. Philology. Vol. I, Part I: South Africa (within the limits of British influence). Part II: Africa (north of the Tropic of Capricorn). London and Leipzig: Trübner. Bleek, W. H. I. (–). A Comparative Grammar of South African Languages. Part : Phonology (). Part : The Concord, Section , The Noun (). London and Cape Town: Trübner. Bleek, W. H. I. (). Über den Ursprung der Sprache. Weimar: Hermann Böhlau. Bloomfield, L. (). Language. London: George Allen & Unwin. Brusciotto, G. (). Regulae quaedam pro difficillimi congensium idiomatis faciliori captu ad grammaticae normam redactae. [Some rules for the easier understanding of the most difficult language of the Congo people, brought into the form of a grammar.] Rome: Sacra Congregatio de Propaganda Fide.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Dalby, D. (). ‘Reflections on the classification of African languages, with special reference to the work of Sigismund Wilhelm Koelle and Malcolm Guthrie’, African Language Studies (in honor of Malcolm Guthrie): –. Dapper, O. (). Umbständliche und Eigentliche Beschreibung von Africa. Amsterdam: Meurs. Doke, C. M. (). Bantu: Modern grammatical, phonetical, and lexicographical studies since . London: Dawsons of Pall Mall. Drexel, A. (–). ‘Die Gliederung der afrikanischen Sprachen’, Anthropos /: –; /: –; : –, –. Gerhardt, L. (). ‘The place of Carl Meinhof in African linguistics’, Afrika und Übersee : –. Gordon, R. G. (). Ethnologue, th edn. Dallas, TX: Summer Institute of Linguistics. Greenberg, J. H. (/, ). ‘Studies in African linguistic classification’, Southwestern Journal of Anthropology : –, –, –; : –. Greenberg, J. H. (). Essays in Linguistics. Chicago, IL: University of Chicago Press. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Greenberg, J. H. (). ‘The history and present status of African linguistic studies’, in A. S. D. Dil (ed.), Language, Culture and Communication. Stanford, CA: Stanford University Press, –. Guthrie, M. (). The Classification of the Bantu Languages. London: Oxford University Press. Haeckel, E. (). ‘Vorwort’, in W. H. I. Bleek, Über den Ursprung der Sprache. Weimar: Hermann Böhlau, i–viii. Haeckel, E. (). Natürliche Schöpfungsgeschichte. Berlin: Dietrich Reimer. Klingenheben, A. (/). ‘Die Präfix-Klassen des Ful’, Zeitschrift für Eingeborenen-Sprachen : –, –. Koelle, S. W. (). Polyglotta Africana: Comparative vocabularies of more than distinct African languages. London: Church Missionary Society. Köhler, O. (). ‘Geschichte und Probleme der Gliederung der Sprachen Afrikas: Von den Anfängen bis zur Gegenwart’, in H. Baumann (ed.), Die Völker Afrikas und ihre traditionellen Kulturen. Teil : Allgemeiner Teil und südliches Afrika. Wiesbaden: Franz Steiner, –. Krapf, L. (). Vocabulary of Six East-African Languages. Tübingen: F. Fues. Lepsius, K. R. (). Nubische Grammatik: Mit einer Einleitung über die Völker und Sprachen Afrika’s. Berlin: Wilhelm Hertz. Ludolf, H. (). Grammatica Aethiopica. London. Ludolf, H. (). Grammatica Linguae Amharicae. Frankfurt am Main. Lyons, J. (). Introduction to Theoretical Linguistics. London: Cambridge University Press. Meinhof, C. (). Grundriß einer vergleichenden Lautlehre der Bantusprachen (st edn , Berlin) Berlin: Dietrich Reimer. Meinhof, C. (). Die Sprachen der Hamiten. Hamburg: L. Friederichsen. Meinhof, C. (). Grundzüge einer vergleichenden Grammatik der Bantusprachen. nd edn. Hamburg: Eckardt & Messtorff. Meyer, E. (/). ‘Das Werk von Carl Meinhof—dargestellt in seinen Schriften’, Zeitschrift für Eingeborenen-Sprachen : –. Müller, F. (–). Grundriß der Sprachwissenschaft. vols. Vienna: Hölder. Schleicher, A. (). Die Darwinsche Theorie und die Sprachwissenschaft—Offenes Sendschreiben an Herrn Dr. Ernst Häckel . . . Weimar: Hermann Böhlau. Smidt, W. (). Afrika im Schatten der Aufklärung: Das Afrikabild von Immanuel Kant und Johann Gottfried Herder. Berlin: Holos. Smidt, W. (). ‘Afrika im Werk Kants, oder: Wie die „läppischen“Afrikaner dass Ziel der Geschichte verfehlten’, in Th. Schumann, M. Reh, R. Kießling, and L. Gerhardt (eds.), Aktuelle Forschungen zu afrikanischen Sprachen. Cologne: Rüdiger Köppe, –. Thomason, S. G., and Kaufman, T. (). Language Contact, Creolization and Genetic Linguistics. Berkeley, Los Angeles, and London: University of California Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Torrend, J. A. (). A Comparative Grammar of the South-African Bantu Languages. London: Kegan Paul, Trench, Trübner & Co. Weinreich, M. (). ‘Der yivo un di problemen fun undzer tsayt’, Yivo-bleter : –. Westermann, D. (). Die westlichen Sudansprachen und ihre Beziehungen zum Bantu. Berlin: de Gruyter. Westermann, D. (). ‘Nominalklassen in westafrikanischen Klassensprachen und in Bantusprachen’, Mitteilungen des Seminars für Orientalische Sprachen : –. Winston, F. D. D. (). ‘Greenberg’s classification of African languages’, African Language Studies : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
..............................................................................................................
LANGUAGE PHYLA AND FAMILIES ..............................................................................................................
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
- , - ......................................................................................................................
. W N-C?
.................................................................................................................................. F proposed by Greenberg (), Niger-Congo (NC) has for decades been treated as one of the four major phyla of African languages.1 The term, as presently used, however, is not without its difficulties. On the one hand, it is employed as a referential label for a group of over , languages, putting it among the largest commonly cited language groups in the world. On the other hand, the term is also intended to embody a hypothesis of genealogical relationship between the referential NC languages that has not been proven. Reference sources through recent decades, such as Williamson (b) and Williamson and Blench (), have tended to equate NC in its referential and genealogical senses. More current presentations, such as Dimmendaal (: –), clearly differentiate between the two, separating out a “core” NC comprising a set of language groups where evidence of genealogical relationship is comparatively strong from a second set where it is much weaker. Below, the intended sense of “Niger-Congo” will be made explicit where necessary. The distinction between referential and genealogical NC is well illustrated through the comparison of Maps . and .. Map . depicts a relatively uncontroversial genealogical core of NC, alongside other African language families and isolates. Map . breaks down core NC into commonly cited subgroups. In addition to the subgroups on Map ., referential NC would also include the following groups found on Map ., going from east to west: Mande, Dogon, Ijo/Defaka (more typically known under the label of Ijoid), and Ubangian, as well as some of the languages indicated as isolates, such as Bayot or Ega.
The label “Niger-Kordofanian” is also used for this group, following Greenberg (: ) and reflecting a genealogical claim—not currently generally accepted (Williamson b: )—that Kordofanian languages are a coordinate branch with the rest of NC. Today, NC dominates as the general term for the group, inclusive of Kordofanian languages believed to be part of the family. 1
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. African language families and isolates following recent proposals
There are a number of existing works providing overviews of NC. The most thorough of these is the edited volume Bendor-Samuel (). Others include various chapters in Sebeok (), as well as Williamson and Blench (), Pulleyblank (), and Dimmendaal and Storch (). More general references on comparative African linguistics, such as Sands () or Dimmendaal (: –), also contain helpful overviews. While inevitably repeating some of the content found in these works, this chapter will also try to complement their coverage. A targeted and critical overview of work on language relationships will be given in section .. In section ., there will then be more detailed discussion of Benue-Congo (BC), the subgroup of NC of special focus here. A summary of major typological and areal features of NC languages will be given in section .. The chapter will conclude with a brief consideration of what a productive research agenda for comparative NC linguistics in the twenty-first century might look like in section ..
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
. Niger-Congo subgroups following recent proposals
. L N-C
..................................................................................................................................
.. The history of Niger-Congo classification The widespread acceptance of a language family associated with the name Niger-Congo can be first attributed to Greenberg (), with Greenberg () being more typically cited as a general reference. However, Greenberg’s proposals are best understood as a refinement and extension of ideas developed by earlier scholars, in particular Westermann (, ; see also Wallis ). Greenberg’s most important original contribution to NC classification
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
is almost certainly his explicit “demotion” of the Bantu languages to the status of a subgroup of BC (see section .). Previously, Bantu languages had been treated as a separate language family due to their internal cohesion and the fact that the comparatively large size of many of them encouraged the availability of descriptive materials from an early stage (see Cole for relevant discussion). Other sources adequately cover the history of NC language classification (see, for example, Williamson , a, b; Greenberg ; Dimmendaal and Storch ).2 Olson () should be singled out for its attention to the details of the evidence provided for many of the subgroups, clarifying, in particular, why a number must be considered unproven. Many of the early classifications are primarily of historiographic interest, though they occasionally contain insights which are of significance for contemporary scholarship. For instance, the work of Johnston () made use of the term “Semi-Bantu” to characterize languages with clear “Bantu” affinities but not showing the full range of characteristics that he accorded to Bantu languages (see, for example, Johnston : –). Among other possibilities, he discusses whether Semi-Bantu languages could have developed from a “fusion” of Bantu languages with other languages (Johnston : ), a view more concretely and explicitly adopted by the prominent Bantuist Malcolm Guthrie (: ). After the work of Greenberg, hypotheses such as these were often summarily rejected. Welmers (: –), for instance, completely dismisses the ideas of Johnston () on language mixing. While it is certainly the case that many of the specific ideas found in older works need to be immediately dismissed (with their more racist elements standing out in particular), one sometimes finds parallels between their claims and more recent scholarly trends. Johnston’s idea of language creation via “fusion”, for instance, clearly anticipates the notion of “mixed” languages, and, given that the development of a Bantu Cushitic mixture as found in Ma’a/Mbugu is now amply documented (Mous ), we could just as well expect many (much harder to detect) mixed languages to have developed among grammatically and lexically similar NC languages over the millennia, making a label like “Semi-Bantu” perhaps not as inappropriate as it might first appear.
.. Subgrouping within Niger-Congo Previous overview presentations of NC, such as Williamson (b), are strongly oriented towards tree-based subgrouping over other kinds of historical relationships, and this limits their value to the non-specialist, who is likely to be interested in issues of language contact and linguistic prehistory as well (see also Dalby and Heine : –). Moreover, the evidentiary basis of the presented subgroupings is never entirely clear, and their discovery often relies on relatively coarse methods, e.g., lexicostatistics, as in Bennett and Sterk () (see Schadeberg for further discussion of their methods). These issues are present in the whole-family tree-based classification given in Williamson and Blench (: ), for instance. Unfortunately, due to a lack of more definitive classifications, the comprehensive nature of such trees means they are republished in other reference sources (see, e.g., 2 See also Doneux () for a comprehensive overview of African language classification through and Lüpke and Storch (: –) for a concise overview of classifications from the early nineteenth century through the early twentieth century.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
Childs : and Schadeberg b: ), and this process of “scholarly inertia” (Childs : ) makes such classifications appear more valid than the evidence warrants. Accordingly, no specific tree-based classification of the NC languages is included here. One can get a more accurate impression of the present state of our knowledge regarding the subgrouping of the family simply by inspecting Map ., as long as a number of qualifications are made (see also Dimmendaal : – for more detailed discussion). First, the implication of Map . that Bantoid and Bantu are separate major subgroups of NC is simply false. Bantu is universally viewed as a subgroup of Bantoid, and Bantoid is viewed as a subgroup of BC. This “splitting” representation is due to the importance of Bantu in the context of comparative NC studies. Second, while languages in the various subgroups seen in Map . are generally considered to belong to NC, the genealogical unity of the subgroups themselves is not necessarily established. For example, on the one hand, one finds Kru, whose status as a subgroup is not controversial (Marchese ). On the other hand, there is Atlantic, whose status as a subgroup is far from clear (Childs : –).3 Moreover, as with classificatory trees, one should be wary of assuming that a subgroup is coherent merely because it has been repeatedly listed in the reference literature. Adamawa provides a good example here. It has been part of standard presentations for decades, even though Kleinewillinghöfer (: ) states that its coherence had not yet been convincingly demonstrated, and, to the best of my knowledge, no work has established it since, either. Other groups raise different problems. Kwa, for instance, is widely viewed as part of a larger complex with BC, and it is not clear either where the border between Kwa and BC lies or if Kwa itself is a coherent subgroup within this larger complex (see, e.g., Stewart : –; Williamson and Blench : –; and Kropp Dakubu ). There are similar problems with respect to Gur (Bendor-Samuel : ; Naden : –), and links between Gur and Adamawa languages have also been proposed (Boyd : –; Kleinewillinghöfer ; Elders : –).4 In the case of Kordofanian, while reasonable evidence for its inclusion within NC has been presented, its membership has changed from when it was first proposed, and it is, again, not established if the Kordofanian languages are a coherent subgroup or represent distinct NC lineages that independently entered the Nuba mountain refugium (see Schadeberg a, ; Dimmendaal : –; Blench ). It otherwise seems premature to present any articulation of the NC tree above the level of these subgroups. The documentary coverage of the family is also quite uneven, which has an impact on our understanding of the comparative linguistics of the family. BC and Kwa, for instance, are relatively well documented, whereas most other groups need more attention. The groups just discussed are all, at present, uncontroversial members of genealogical or “core” NC (see section .). Something should also briefly be said here about additional subgroups, shown in Map ., which have been treated as part of referential NC. The most well-studied of these is Mande (see Dwyer ; Kastenholz /). Mande’s status as a coherent group has been considered clear since at least Koelle () 3
See Pozdniakov () for discussion of the difficulties involved in conducting comparative work on Atlantic, and Pozdniakov () for a more updated discussion of Atlantic classification. 4 The extensive comparative work on the noun class systems of Gur languages reported in in Miehe and Winkelmann () and Miehe et al. () will hopefully add more clarity to the Gur picture in the coming years.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
(Dwyer : ). Dogon languages were poorly known until quite recently, and, as documentation has become more available, good evidence for their inclusion in NC has not materialized. Moran and Prokić () focus on the internal classification of Dogon but also discuss the history of its external classification. The group labeled “Ijo/Defaka” in Map . consists of a small language cluster (Ijoid) along with the Defaka language (Jenewari ). A close relationship between Ijoid and Defaka is not yet entirely clear, and Connell et al. () contains an up-to-date overview of the relevant issues. Mande and Ijoid languages will be briefly discussed again in section .. to clarify why their inclusion within genealogical NC should be considered tenuous at present. The status of Ubangian (or “Eastern” in older publications) as a valid subgroup is not completely clear. Samarin (: ) suggests that it is “obvious”, Boyd (: –) is more equivocal, and Moñino () indicates that the unity of the group is no longer obvious (see also Moñino ). Greenberg () proposed grouping Ubangian languages together with Adamawa languages, but good evidence for this has not materialized. Unlike Adamawa languages, which have been found to have good connections to other NC languages, good evidence for Ubangian’s inclusion with the rest of NC has not been presented, which is why recent overview treatments, such as Dimmendaal (: ), treat it as separate from NC.
.. What proves relationships in Niger-Congo? The immediately preceding discussion naturally leads to the question of what comparative evidence exists for NC in the first place. Greenberg (: –) famously attributed his classification to the use of so-called mass comparison, where many languages are looked at in parallel. While this technique may be appropriate for hypothesis raising, there is general agreement that it cannot constitute proof of relationship (see, e.g., Newman : for a perspective that is sympathetic to Greenberg’s work and Campbell and Poser : – for one that is not). The strongest evidence for a genealogical unit under the heading of NC undoubtedly involves comparison of noun class systems among members of the family. Such a comparison, adapted from Schadeberg (: ), as part of a justification for inclusion of Kordofanian in NC, is given in Table ..5 The table lists class markers that are candidates for being cognates, from six reconstructed NC groups. As indicated, Oti-Volta is a subgroup of Gur, and Ghana-Togo Mountain Languages (formerly referred to as Togo Remnant languages; Stewart : ; Blench : ) are used to exemplify Kwa langages. The class numbering follows Bantuist conventions (see Katamba , with the addition of class a following Welmers : ). Classes and form a singular/plural pairing, as do classes and . Each class is associated with a characterization of its semantics or specific
5 The sources for the reconstructions are as follows: Kordofanian (Schadeberg : ; see also Schadeberg b: ), Atlantic (Doneux : ), Togo Remnant (Heine : ), Oti-Volta (Manessy : ), Benue-Congo (De Wolf : –), Bantu (Meeussen : ). Schadeberg () also includes data from the Ubangian language Mba (Carrington : ), which is removed here due to Ubangian’s questionable status within NC (see section ..). Some adjustments have been made based on consultation of original sources and secondary sources such as Dimmendaal (: ).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
lexical elements reconstructed as belonging to that class (given in single quotes). The ProtoBantu (PB) forms give two distinct series of noun class markers, one for the prefixes appearing on nouns themselves and another for a series of concord markers associated with third person pronouns, following Meeussen (: ).
Table . Possible cognate noun class markers in Niger-Congo following Schadeberg (: ) class 1
3
4 ‘tree’
5
6
li-
ŋu-
6a
kordofanian
gu- humans gu- j-
atlantic
gu- humans gʊ- Ci- ‘trees’
de- ga-
‘head, name’ ma- liquids
oti–volta (gur)
-ʊ
humans -bʊ -Ci ‘tree’
-ɖɪ -a
‘egg, head’
-ma liquids
ghana–togo (kwa)
o-
humans o-
i-
a-
‘egg, head, name’
N-
benue-congo
u-
humans u-
(t)i- ‘tree’
a-
‘egg, head, name’
ma- liquids
‘egg, name’
ma- liquids ga-
‘firewood’ li-
bantu noun mu- humans mu- mi- ‘tree’ pronoun jugu- gi-
li-
i ̧madi- ga-
‘egg’
ŋ- liquids
liquids
The presentation in Table . is strongly suggestive of a genealogical relationship between the relevant groups (see also Williamson b: – and Kießling : for comparable and more extensive presentations). The form–function matching across classes, reconstruction of pairings of specific singular/plural classes, as well as lexically idiosyncratic groupings of words in certain classes (e.g. ‘egg’, ‘head’, and ‘name’ in classes /) are not mere typological resemblances.6 No general reconstruction of the noun class system of Proto-Niger-Congo (PNC) has been developed (Grinevald and Seifart : ; Kießling : ), which is surprising given its importance in establishing relatedness among members of the family. As indicated in Table ., in one group, Oti-Volta, noun class suffixes are reconstructed, when, elsewhere, noun class prefixes are reconstructed. Variation between prefixing and suffixing class markers is, in fact, relatively widespread within NC and has received a fair amount of attention. It is generally attributed to patterns of morphological loss (e.g. of prefixes on the noun) with subsequent renewal via concordant elements following the noun that morphologize as suffixes (see, e.g., Hoffmann : –; De Wolf : –; Welmers b: ; Greenberg , ; Childs ; Williamson b: –; and Dimmendaal : –). One point that must be emphasized in the use of noun classes to establish family membership is the fact that there are languages that are uncontroversially considered to be NC despite lacking synchronic noun class systems. This is especially the case for many 6 Hammarström () argues that the data presented in Table . is not, in fact, sufficient to establish that Kordofanian is connected to the rest of NC, even if it is suggestive of a relationship. This may be true of the other groups as well, but I am not aware of targeted studies.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Kwa languages. Lexical evidence has been proposed strongly linking Kwa languages without noun classes to nearby languages with noun classes (see, e.g., Bennett and Sterk ), and there is also sometimes evidence of remnant noun class markers on lexical items in these languages making it reasonable to see these as NC languages that once had noun class systems but lost them (see section ..). In a group such as Mande, however, one does not find noun class systems, and the points of lexical similarity to the rest of NC are quite weak (Welmers a: –; Bennett and Sterk : ; Williamson b: ; Williamson and Blench : –), which is why its inclusion within NC is seriously questioned. While Ijoid languages do have noun classes, they are very different in structure from the canonical NC system, based on distinctions in animacy and sex (Jenewari : –), and, again, lexical linkages are not strong (Bennett and Sterk : ). Thus, it is not merely the lack of correspondences like those in Table . that has resulted in controversy regarding the placement of groups like these within NC, but also the lack of other strong evidence for their inclusion. Next to noun class systems, verb extensions are the other major grammatical feature of NC languages that has been proposed for family-level reconstruction.7 These are suffixes which create derived verb stems, often altering verbal argument structure in some way. The most extensive comparative study of these verb extensions is found in Voeltz (), and Hyman () provides an up-to-date overview (see also Creissels : –). A listing of some possibly cognate extensions in Bantu and Atlantic, alongside the NC reconstructions of Voeltz (: –; adapted from Hyman : ), is given in Table .. The PB reconstructions (two of which are multimorphemic) are drawn from Schadeberg (a: ) and the Proto-Atlantic ones from Doneux (: –).
Table . Possible cognate verbal extensions in Niger-Congo following Hyman (: )
applicative causative passive reciprocal reversive
proto -nigercongo
proto bantu
proto atlantic
-de -ci, -ti -o -na -to
-ɪd -ic-i -ɪb-ʊ -an -ʊd
-ed -an -V[+back] -ad -ɪt
Since Bantu and Atlantic languages are at opposite geographic ends of NC and do not otherwise show a close relationship, possible cognates in these two branches support a NC level reconstruction. However, while the parallels between the verbal extensions in PB and Proto-Atlantic are suggestive of some genealogical connection, there are both formal and functional complications, as discussed in Hyman (). Moreover, they do not form a 7
See also Babaev () for a reconstruction of NC person marking based on an extensive survey of the family.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
paradigm in the way that noun class markers do. This makes their value for proving a genealogical relationship comparatively weak. Another possible source of evidence which could firmly establish NC relationships would be, of course, detailed lexical reconstructions. At the NC level, the only serious efforts in this regard appear to be due to Stewart (e.g. , ). This work focuses on comparison between PB and the Potou-Tano subgroup of Kwa, suggesting that this can serve as a foundation for PNC. However, as Mous (: ) points out in an obituary, Stewart “regretted the fact that so few people joined him in the strict application of the comparative method to the reconstruction of West African languages.” Indeed, the level of “proof” available, in general, for the genealogical unity of NC falls well short of the highest standards. Nevertheless, available comparisons and the intuitions of specialists, at least regarding so-called “core” NC, should not be dismissed lightly, given their exposure to the relevant data. In this regard, the highly negative assessment of NC found in Campbell and Poser (: –) seems overly pessimistic.
. B-C: N-C’
..................................................................................................................................
.. Delineating Benue-Congo and Bantu BC is the largest commonly cited subgroup of NC, and its members include the Bantu languages, which dominate the southern part of sub-Saharan Africa. Non-Bantu BC languages are found in a geographically contiguous region of southern Nigeria and adjacent parts of Cameroon, with Bantu languages spread out over a much larger region to the south and east of the rest of the group. Current reference sources place around two-thirds of all NC languages within BC. In terms of nomenclature and classification, BC suffers from the same range of problems as NC: its status as a genealogical unit is not clear, convincing criteria for establishing which groups belong to it have not been presented, and its subgroups are not firmly established beyond comparatively low-level ones. Williamson (a), though now somewhat dated and limited in scope, remains the most extensive published general reference for the group. The most prominent effort at reconstruction is represented by De Wolf ’s () work on noun class systems (see also Good ). Babaev (, ) is the most recent systematic attempt at BC reconstruction, focusing on person marking. By virtue of their emphasis on family-wide reconstruction, these latter works are also useful for their collection of references on the family that are more up-to-date than some of the more standardly cited sources. The problems in delineating a clear-cut set of BC languages can be illustrated by consideration of two prominent classificatory concerns. The first is the division between BC and Kwa. A number of language groups of Nigeria at the eastern border of Kwa and western border of BC have been variously classified in one or the other subgroup. These include Yoruboid (Capo ), Igboid (Manfredi ), Edoid (Elugbe ), Nupoid (Blench ), and Idomoid (Armstrong ), all relatively shallow in genealogical terms.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
While standard presentations have placed these groups in BC for decades on the basis of proposals in Bennett and Sterk (), no particularly convincing corroborating evidence has subsequently emerged, leading instead to proposals to “return” them to Kwa (see, e.g., Blench a: §.). Significant here is that these groups pattern typologically with the isolating languages that are uncontroversially treated as Kwa, rather than the agglutinative languages found in much of BC (especially Bantu) (see, e.g., Williamson , Hyman , Good for relevant discussion). This has led scholars to use the label “Kwa” to cover both genealogical and typological notions (see, e.g., Aboh and Essegbey : xi for a recent non-genealogical use of “Kwa”), a source of potential confusion in the literature. One encounters comparable issues with respect to the delineation of Bantu languages from the rest of BC. Despite relatively intense study, it has not yet been possible to devise criteria which distinguish a “Bantu” group from its nearest Bantoid relatives (Nurse and Philippson b: –). The only proposal for a Bantu-specific innovation that gained serious consideration centered around the presence of nasal consonants in certain Bantu noun classes (Crabb : ; see also Hyman : and Table .). Voorhoeve () discusses difficulties with this proposal arising from consideration of shared vocabulary alongside noun class innovations for certain Bantoid languages of the Cameroon Grassfields. More strikingly, Miehe () suggests a much broader provenance for nasal noun classes, well outside of Bantu. Thus, somewhat ironically, even the most well-studied subgroup within NC cannot be considered to be firmly established as a true genealogical entity.
.. Benue-Congo subgroups As just discussed, the precise borders of BC can fluctuate depending on the classification. In this section, the less controversial subgroups of BC will be briefly introduced. The more difficult cases of subgroups whose classification has shifted between Kwa and BC were listed in section ... Bantu and its closest relatives comprise Bantoid, BC’s largest uncontroversial subgroup, which probably contains two-thirds of all BC languages. Of these, around – are nonBantu Bantoid languages. These occupy a compact region on either side of the Nigeria– Cameroon border, though it should be said that, as with Bantu, precise boundaries for Bantoid have not been established. The other units of “core” BC include Cross River, a group of sixty or so languages spoken in southeastern Nigeria (Faraclas ), as well as the groups Kainji, Plateau, and Jukunoid (Gerhardt ). These are spoken in scattered areas of central Nigeria (with some Jukunoid languages also found in Cameroon). Kainji is associated with around sixty languages, Plateau around seventy, and Jukunoid around twenty. These groups have been classified together (see Gerhardt : –), though referential classifications now treat them as separate from each other. The languages of the Kainji group are especially distinctive, and, as data on them becomes increasingly available, it seems likely that they will play an important role in coming to a better understanding of Proto-BC (see McGill and Blench ; Blench and McGill ; Blench b). In addition to these major groups, there are a number of languages that are placed within BC but of otherwise uncertain affiliation, for instance, Ukaan (Salffner and Sands ). If we exclude the Bantu languages, the most striking overall pattern of core BC is the great language
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
density of the region where they are spoken (see, e.g., Stallcup : ) and the degree of grammatical diversity one finds even within languages belonging to a single subgroup. The surprising extent of their grammatical diversity can be seen in, for instance, the study of Cross River noun classes found in Faraclas (), where he argues that the diversity of noun class systems within this single subgroup can stand in as a model for all of BC. Gerhardt () makes similar arguments for the Plateau group and Good et al. () describe a remarkable degree of noun class system diversity in the small cluster of languages of the referential Yemne-Kimbi group. This is not to say that BC languages do not share many typological similarities (see section .). Rather, the extent to which the range of variation found within the whole of BC can also be found within its subgroups is striking. Finally, it is worth noting that, due to the prominence of Bantu, the label BC is variously used either in the “regular” way for a subgroup of NC which includes the Bantu languages or to refer only to non-Bantu BC languages. In particular, if a given source labels a language as belonging to “Benue-Congo”, this will almost certainly mean “non-Bantu BenueCongo”. Similarly, “Bantoid” is often used to mean “non-Bantu Bantoid”.
. N-C
..................................................................................................................................
.. The areality of major Niger-Congo features Due to NC’s dominance of the African continent, there is considerable overlap between the typology of NC languages and those of Africa in general. On the one hand, this means that continent-wide overviews such as Clements and Rialland () and Creissels et al. () can serve as a useful reference for NC typology, and each covers the relevant issues in much more detail than is possible here. On the other hand, it also means that there are relatively few typological features that specifically distinguish NC from other languages of the continent. In particular, since it spans a number of linguistic areas, diverse influences have hindered the maintenance of a unique typological profile for the family (see Dimmendaal ; Good ). The distribution of labial-velar consonants (e.g. kp and gb) illustrates these points relatively clearly. These consonants are exceptionally rare from a worldwide perspective, except in a region of Africa that has been termed the Macro-Sudan belt, running roughly east–west from the Atlantic Ocean to the Ethiopian highlands and bounded to the north by the Sahara and the south by the Central African rainforest (see Güldemann ). NC languages dominate this belt, and it is languages of this family where such consonants are most likely to be found (Maddieson ). At the same time, labial-velars are found in languages outside of NC as well, such as in the Chadic (Afro-Asiatic) language Kotoko (Bouny : ) but are almost non-existent in Bantu languages except for those northern Bantu languages bordering the Macro-Sudan region (Güldemann : –). Thus, they are an “African” feature that is very much present in NC, but which does not characterize the whole family. The distribution of [ATR], or cross-height, harmony in NC shows a similar pattern. Languages in the Macro-Sudan region often exhibit vowel harmony where one set of high
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
and mid vowels forms a harmony class against another set of high and mid vowels, specifically, /i u e o/ vs. /ɪ ʊ ε ɔ/ (Casali ). Bantu languages, by contrast, tend to show different, typologically unusual vowel harmony patterns (Clements and Rialland : –; see also Hyman ). There is a comparable distributional pattern with respect to the presence of alternative O–V word orders alongside the more typically African V–O word order. S–V–O word order dominates NC, either as the only frequent word order or as one of two common word orders, the other being S–(Aux)–O–V (Güldemann : –).8 For instance, in the Cross River (BC) language Leggbó, S–V–O word order is most typical, but, in negative sentences, the word order shifts to S–O–V, as in () (Good : –). () a. wàdum sέ e-dzi lídzil. man -eat food ‘The man ate food.’ b. wàdum sέ lídzil eè-dzi. man food :-eat ‘The man didn’t eat food.’ As recognized in the influential work of Heine (: –; see also Heine ), this type of word order split between languages is correlated with other features, with languages with relatively rigid V–O word order typically showing noun–genitive order and those exhibiting possible O–V order showing genitive–noun order. Languages of this latter type are areally restricted to the western half of the Macro-Sudan belt, encompassing a number of NC subgroups, in particular Atlantic, Kru, Kwa, and (following the pattern seen above) non-Bantu BC, but not Bantu. This word order pattern is also characteristic of Mande languages, and it seems likely that its presence in core NC is due to Mande influence of some kind, again presenting us with an “African” feature in which NC participates, but is only partly implicated. Even if many NC characteristics can be classified as “areal”, it is not the case that we can divide the family neatly into a handful of discrete areas. This is seen, for instance, in the results of the survey of African tone patterns in Wedekind (: ). The survey reveals concentric patterns of tonal complexity where pockets of languages with five distinctive tone levels in Côte d’Ivoire, the southern Nigeria-Cameroon border, and southwestern Ethiopia are surrounded by regions with four distinctive levels, which are, in turn, all part of a large region, corresponding to much of the Macro-Sudan area, with three distinctive tone levels. This region is itself embedded in a two-tone area comprising the bulk of subSaharan Africa. Thus, there are a number of “sub-areas” within the larger areas. Noun class systems are the one NC typological feature that most clearly escapes this areal patterning. While their genealogical implications receive more attention (see section ..), they also have distinctive typological properties. For instance, NC noun class systems are cross-linguistically exceptional for not being based on a distinction between sexes 8 A more recently documented phenomenon with similar distribution is the presence of portmanteau morphemes encoding subject, person, and number along with information on tense, mood, or aspect (Anderson ).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
(see Corbett ). While non-sex-based systems are found elsewhere in the world, they are a minority type and, except for NC, rare in Africa. Kießling (: –) gives an overview of other major typological features of NC noun class systems (see also Welmers : – and Grinevald and Seifart : –). Another noteworthy typological feature of these systems is that, while classes may have specific associations with a given number (e.g. singular, plural, or mass), number does not typically exist as a significant morphological category. Rather, each class affix is “autonomous and mono-morphemic” (Welmers : ).9 This can be seen by consideration of the PB noun class system, as presented in Table . and adapted from Maho (: ) and Katamba (: –). The labels for semantic categories associated with each class should be taken as only approximate (see Denny and Creider , ; Moxley ; Maho : –; and Dingemanse for further discussion).
Table . Proto-Bantu noun classes singular
plural
semantics
1 3 5 7 9 11 12 14 15 16 17 18 19
2 4 6 8 10
humans trees, plants mixed/cl. 6 liquids mixed animals, mixed mixed augmentative, diminutive, etc. abstract infinitive location on location at location in diminutive
mùmùlìkìnìlùkàβùkùpàkùmùpì̧-
βàmìmàβì ̧lì̧-nì-
13 tù-
Various generalizations emerge from Table ., supporting the claim that the noun classes are morphologically autonomous and number is not morphologically encoded. First, there is no consistent formal relationship between paired singular and plural class markers. Second, several of the classes have semantics where number simply does not appear to be a relevant notion (e.g. the locative classes). Third, while simplified presentations of Bantu noun classes imply that singular/plural pairings are relatively consistent, it is not uncommon for there to be lexically conditioned variability in how a given singular noun forms its plural and vice versa (see, e.g., Maho : for discussion specific to Bantu and Good : – and Lovegren : – for relevant examples from a Bantoid language). Finally, we also see in the Bantu case how a single class, namely class , 9
There are, however, cases where number exists as an active morphological category (see, e.g., Schadeberg on Swahili and Creissels : – on certain Atlantic languages).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
simultaneously serves plural and mass encoding functions, again suggesting that a singular/ plural opposition is not basic to the system. The NC noun class system, therefore, constitutes both a genealogical indicator of the family and a typological feature of many NC languages. Reduction of the noun class system in many NC languages means it is not a grammatical feature common across the family. However, extreme patterns of loss are associated with broader patterns of reduction centered around the “Kwa” area (see section ..), meaning that, unlike other “NigerCongo” features, this type of noun class system is not restricted to a specific region but, rather, is found generally except within a specific area. Most striking in this regard is that the most well-developed noun class systems within NC are found in the “peripheral” groups of Bantu (see Table .) and Atlantic (see, e.g., the descriptions of Fula noun class systems as found in Arnott : – and Breedveld : –), where mutual contact cannot have been a factor in their maintenance.
.. The typology of Benue-Congo languages Given the geographic extent of the BC group and the fact that it straddles major areal boundaries, the general comments about the typology of NC languages discussed in section . are largely applicable to BC as well. Beyond this, the most striking feature of BC language typology is almost certainly the diversity of the morphological systems found within the family, ranging from the isolating “Kwa” type, which encompasses many languages of the referential Kwa group as well as BC languages spoken nearby to Kwa languages (see section ..), to the agglutinating Bantu languages, with languages spoken in between these groups showing profiles intermediate between the two (see Hyman ; Good ). This morphological diversity can be illustrated by comparison of an example like that in (), from the “Kwa”-like BC language Yoruba, with the example in (), from the Bantu language Chewa. The Yoruba example employs a serial verb structure to encode a benefactive construction and shows no morphological complexity in its words. In the Chewa example, there are instances of affixation on both the nouns (e.g. the class chi- marker on the word chitsîru ‘fool’) and the verb, which contains a subject agreement prefix, a tense marker, an applicative extension, and the so-called “final vowel”, an inflectional marker found throughout Bantu not associated with a clear-cut functional category (see Nurse : –). The Yoruba example can be further compared with the example in (), from the Kwa language Fon Gbe, which shows a morphosyntactically parallel construction (with an apparent vestigial prefix in the first vowel of the word àsɔn ́ ‘crab’; Good : ). The example from the Bantoid language Naki in () illustrates the gradient nature of this typological divergence within BC. Here, one finds words like àcōm and kâm, both of which exhibit prefixal morphology coding class (with the shapes a- and k- respectively). However, while verbs in this language show tonal morphology, their segmental morphology is quite limited. Subject agreement is not found for instance, though there are some prefixal elements that participate in tense–mood–aspect marking, one of which, with shape a, is seen on the verb ágé in () and marks a kind of consecutivization. The noun nya᷆m ‘.animal’ in this example has no segmental affix, but its tone pattern changes to code number. In its comparative lack of segmental morphology, Naki is more “Kwa”-like, but with respect to overt morphological coding of grammatical categories, it is more “Bantu”-like.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, - () mo mú ìwé wá fún ẹ. take book come give you ‘I brought you a book.’
[Yoruba (BC)] (Stahlke : )
() chitsîru chi-na-gúl-ír-á atsíkána mphátso. .fool --buy-- .girl .gift ‘The fool bought a gift for the girls.’ () Kɔ̀kú sɔ́ àsɔ́n ɔ́ ná Àsíbá. Koku take crab give Asiba ‘Koku gave the crab to Asiba.’
[Chewa (Bantu)] (Alsina and Mchombo : ) [Fon Gbe (Kwa)]
(Lefebvre and Brousseau : )
() àcōm kâm dzε᷅ ágé-kū nya᷆m fyə́, .story .my stand. go.-catch .animal .two bú Fìkō. Fìmə́kwə̄mə́ .Chameleon and .Hare ‘My story is about two animals, Chameleon and Hare.’
[Naki (Bantoid)]
(Good : )
As seen, this morphological divergence within BC affects both nominal and verbal morphology. It is also associated with a general reduction in word size in Kwa-type languages. A matter of open debate is whether the highly agglutinating morphological structures found in Bantu represent the conservative situation for BC, with a language like Naki, as exemplified in (), representing a partial reduction en route to the Kwa type, or whether Bantu became more morphologically complex at some stage with Proto-BC being of a more intermediate type. Güldemann (: –) suggests, for example, that the Bantu verbal prefixes may result from a relatively historically shallow grammaticalization process involving an S–Aux–O–V structure of the sort discussed in section .. (see also Güldemann and Hyman ). Given that BC diversity is reflective of NC diversity more broadly, this issue has clear implications for the family as a whole. Overall, the most straightforward way of understanding BC typology involves recognizing that its northwestern area is found within the Macro-Sudan region, and, thereby, participates in the various areal processes found in the Macro-Sudan and associated subregions, while the Bantu spread resulted in BC languages being spoken well outside this area, with a resulting change in their historical trajectory. The Bantoid area, as well as northwestern Bantu, partly participated in the historical processes of the Macro-Sudan area, though not as extensively as, say, Yoruba (see also Nurse and Philippson b: ).
. A
.................................................................................................................................. The intertwining of early linguistic work on African languages with the larger imperialist enterprise led to a number of unfortunate distortions of the continent’s linguistic picture (Irvine ). Given this, it is hardly surprising that much work in the latter half of the
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
twentieth century emphasized a treatment of African language history based on the results of Indo-European scholarship (see, e.g., Welmers : –) as a kind of corrective to earlier racist ideologies. At this point, however, the limitations of reductionist, tree-based approaches to the historical modeling of NC linguistic development have become clear, and the most promising way forward almost certainly involves the integration of social facts into our models of the family’s history (see also Lüpke and Storch : –). In particular, comparative work has not paid enough attention to the pervasive multilingualism of sub-Saharan African societies (see, e.g., Fardon and Furniss : ). Multilingualism almost certainly has a crucial role to play in understanding well-known comparative problems, such as the difficulty of finding a clear boundary between Bantu and Bantoid or BC and Kwa (see section ..) or even how, despite covering a vast stretch of territory, the Bantu languages behave more like a dialect continuum than a series of discrete languages (Schadeberg b: ; see also Möhlig , ). While multilingualism is most typically understood as a force for convergence, we must also consider its potential role in linguistic divergence. Within NC societies, there is evidence that divergence may involve more socially active processes than implied either by the standard interpretations of the tree-based and wave-based models of language change, where divergence is understood to be largely the result of “drift” due to geographic expansion (Heggarty et al. : ). African linguistic cultures are well documented as employing different kinds of manipulated language (see Storch ), which are a likely source of many localized patterns of change. The fact that speakers in highly multilingual societies know many languages facilitates not only convergence but also deliberate divergence. Both processes require knowledge of the language being used as a model—for purposes of either emulation or distancing (see Schadeberg c: and Di Carlo and Good : – for possible NC examples of these latter processes). If there is a general point to be taken from this survey, then, it is almost certainly that a bias towards presenting classificatory trees, as opposed to the nature of the evidence underlying each position in the tree, has hindered progress in comparative NC linguistics. What has been lacking, in particular, is studies that develop a historical model of the evolution of NC at various levels via the integration of data from lexicon and grammar with what is also known about the socio-historical context in which these languages have been spoken.
R Aboh, E. O., and Essegbey, J. (). ‘Introduction’, in E. O. Aboh and J. Essegbey (eds.), Topics in Kwa Syntax. Dordrecht: Springer, xi–xv. Alsina, A., and Mchombo, S. A. (). ‘Object asymmetries and the Chichewa applicative construction’, in S. A. Mchombo (ed.), Theoretical Aspects of Bantu Grammar. Stanford, CA: Center for the Study of Language and Information, –. Anderson, G. D. S. (). ‘STAMP morphs in the Macro-Sudan belt’, in D. L. Payne, S. Pacchiaro, and M. Bosire (eds.), Diversity in African Languages: Selected papers from the th Annual Conference on African Linguistics. Berlin: Language Science Press, –. Armstrong, R. G. (). ‘Idomoid’, in J. Bendor-Samuel (ed.), –. Arnott, D. W. (). The Nominal and Verbal Systems of Fula. Oxford: Oxford University Press. Babaev, K. V. (). ‘Reconstructing Benue-Congo person marking I: Proto-Bantoid’, Journal of West African Languages : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
Babaev, K. V. (). ‘Reconstructing Benue-Congo person marking II’, Journal of Language Relationship/Voprosy iazykovogo rodstva : –. Babaev, K. V. (). Нигеро-конголезский праязык: Личные местоимения [Proto-Niger-Congo: Personal pronouns]. Moscow: LRC Publishing House. Bendor-Samuel, J. (). ‘Niger-Congo, Gur’, in T. A. Sebeok (ed.), –. Bendor-Samuel, J. (ed.) (). The Niger-Congo Languages: A classification and description of Africa’s largest language family. Lanham, MD: University Press of America. Bennett, P. R., and Sterk, J. P. (). ‘South Central Niger-Congo: a reclassification’, Studies in African Linguistics : –. Blench, R. M. (). ‘Nupoid’, in J. Bendor-Samuel (ed.), –. Blench, R. M. (). ‘Do the Ghana-Togo Mountain languages constitute a genetic group?’, Journal of West African Languages : –. Blench, R. M. (). ‘Splitting up Kordofanian’, in T. C. Schadeberg and R. M. Blench (eds.), –. Blench, R. M. (a). ‘The Bantoid languages’, in Oxford Handbooks Online: Linguistics. Oxford: Oxford University Press, https://www.oxfordhandbooks.com/view/./oxfordhb/../ oxfordhb--e- accessed Jul. , . Blench, R. M. (b). Nominal affixing in the Kainji languages of northwestern and central Nigeria. MS, https://www.academia.edu//Nominal_affixing_in_the_Kainji_languages_of_ northwestern_and_central_Nigeria accessed Jul. , . Blench, R. M., and McGill, S. (). The Kainji languages of Northwestern and Central Nigeria. MS, https://www.academia.edu//The_Kainji_Languages_of_Northwestern_and_Central_ Nigeria accessed Jul. , . Bouny, P. (). ‘Inventaire phonetique d’un parler kotoko: le mandagué de Mara’, in J.-P. Caprile (ed.), Études phonologiques tchadiennes. Paris: SELAF, –. Boyd, R. (). ‘Adamawa-Ubangi’, in J. Bendor-Samuel (ed.), –. Breedveld, J. O. (). Form and Meaning in Fulfulde: A morphophonological study of Maasinankoore. Leiden: Research School Center of Non-Western Studies. Campbell, L., and Poser, W. J. (). Language Classification: History and method. Cambridge: Cambridge University Press. Capo, H. B. C. (). ‘Defoid’, in J. Bendor-Samuel (ed.), –. Carrington, J. F. (). ‘Esquisse de la langue mba (kimanga)’, Kongo-Overzee : –. Casali, R. F. (). ‘ATR harmony in African languages’, Language and Linguistics Compass : –. Childs, G. T. (). ‘Noun class affix renewal in Southern West Atlantic’, in J. D. Kaye, H. J. Koopman, D. Sportiche, and A. Dugas (eds.), Current Approaches to African Linguistics . Dordrecht: Foris, –. Childs, G. T. (). An Introduction to African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Clements, G. N., and Rialland, A. (). ‘Africa as a phonological area’, in B. Heine and D. Nurse (eds.), –. Cole, D. T. (). ‘The history of African linguistics to ’, in T. A. Sebeok (ed.), –. Connell, B., Bennett, W., Essien, I., Obikudo, E., Akinlabi, A., and Ndimele, O.-M. (). ‘Defaka and Ijọ: a reassessment of the Ijoid relationship’, in M. Brenzinger and A.-M. Fehn (eds.), Proceedings of the th World Congress of African Linguistics (WOCAL ). Cologne: Rüdiger Köppe, –. Corbett, G. G. (). ‘Sex-based and non-sex-based gender systems’, in M. S. Dryer and M. Haspelmath (eds.), The World Atlas of Language Structures online. Leipzig: Max Planck Institute for Evolutionary Anthropology, http://wals.info/chapter/ accessed Jul. , . Crabb, D. W. (). Ekoid Bantu Languages of Ogoja. Cambridge: Cambridge University Press. Creissels, D. (). ‘Niger-Congo’, in R. Lieber and P. Štekauer (eds.), The Oxford Handbook of Derivational Morphology. Oxford: Oxford University Press, –. Creissels, D. (). ‘Typologie des classes nominales dans les langues atlantique’, in D. Creissels and K. Pozdniakov (eds.), –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Creissels, D., Dimmendaal, G. J., Frajzyngier, Z., and König, C. (). ‘Africa as a morphosyntactic area’, in B. Heine and D. Nurse (eds.), –. Creissels, D., and Pozdniakov, K. (eds.) (). Les classes nominales dans les langues atlantiques. Cologne: Rüdiger Köppe. Dalby, D. (). ‘A referential approach to the classification of African languages’, in C.-W. Kim and H. Stahlke (eds.), Papers in African Linguistics. Carbondale, IL, and Edmonton, AB: Linguistic Research, –. Denny, J. P., and Creider, C. A. (). ‘The semantics of noun classes in Proto-Bantu’, Studies in African Linguistics : –. Denny, J. P., and Creider, C. A. (). ‘The semantics of noun classes in Proto-Bantu’, in C. G. Craig (ed.), Noun Classes and Categorization. Amsterdam and Philadelphia, PA: John Benjamins, –. De Wolf, P. P. (). The Noun Class System of Proto-Benue-Congo. The Hague: Mouton. Di Carlo, P., and Good, J. (). ‘What are we trying to preserve? Diversity, change, and ideology at the edge of the Cameroonian Grassfields’, in P. K. Austin and J. Sallabank (eds.), Endangered Languages: Beliefs and ideologies in language documentation and revitalization. Oxford: Oxford University Press, –. Dimmendaal, G. J. (). ‘Areal diffusion versus genetic inheritance: an African perspective’, in A. Y. Aikhenvald and R. M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance: Problems in comparative linguistics. Oxford: Oxford University Press, –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Dimmendaal, G. J., and Storch, A. (). ‘Niger-Congo: a brief state of the art’’, in Oxford Handbooks Online: Linguistics. Oxford: Oxford University Press, https://www.oxfordhandbooks.com/view/ ./oxfordhb/../oxfordhb--e- accessed Jul. , . Dingemanse, M. (). The Semantics of Bantu Noun Classification: A Review and Comparison of Three Approaches. MA thesis. Leiden University, http://hdl.handle.net//-M--CA- accessed Jul. , . Doneux, J. L. (). ‘Hypothèses pour la comparative des langues atlantiques’, Africana Linguistica : –. Doneux, J. L. (). Histoire de la linguistique africaine: Des précurseurs aux années . Aix-enProvence: Publications de l’Université de Provence. Dwyer, D. J. (). ‘Mande’, in J. Bendor-Samuel (ed.), –. Dwyer, D. J. (). ‘The place of Mande in Niger-Congo’, in I. Maddieson and T. J. Hinnebusch (eds.), –. Elders, S. (). ‘Issues in comparative Kebi-Benue (Adamawa)’, Africana Linguistica : –. Elugbe, B. O. (). ‘Edoid’, in J. Bendor-Samuel (ed.), –. Faraclas, N. G. (). ‘Cross River as a model for the evolution of Benue-Congo nominal class/ concord systems’, Studies in African Linguistics : –. Faraclas, N. G. (). ‘Cross River’, in J. Bendor-Samuel (ed.), –. Fardon, R., and Furniss, G. (). ‘Introduction: frontiers and boundaries—African languages as political environment’, in R. Fardon and G. Furniss (eds.), African Languages, Development and the State. London: Routledge, –. Gerhardt, L. (). ‘Kainji and Platoid’, in J. Bendor-Samuel (ed.), –. Gerhardt, L. (). ‘Western Plateau as a model for the development of Benue-Congo noun-class systems’, Afrika und Übersee : –. Good, J. (). ‘When arguments become adjuncts: split configurationality in Leggbó’, in D. W. Kaiser, J. E. Cihlar, A. L. Franklin, and I. Kimbara (eds.), Proceedings of the Chicago Linguistic Society –: Main session. Chicago, IL: Chicago Linguistic Society, –. Good, J. (). ‘Topic and focus fields in Naki’, in I. Fiedler and A. Schwarz (eds.), The Expression of Information Structure: A documentation of its diversity across Africa. Amsterdam and Philadelphia, PA: John Benjamins, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
Good, J. (). ‘How to become a “Kwa” noun’, Morphology : –. Good, J. (). ‘Niger-Congo languages’, in R. Hickey (ed.), The Cambridge Handbook of Areal Linguistics. Cambridge: Cambridge University Press, –. Good, J. (). ‘Eastern Benue-Congo noun classes, with a focus on morphological behavior’, in J. R. Watters (ed.), Proto-East-Benue-Congo. Berlin: Language Science Press –. Good, J., Lovegren, J., Mve, J. P., Tchiemouo, N. C., Voll, R., and Di Carlo, P. (). ‘The languages of the Lower Fungom region of Cameroon: grammatical overview’, Africana Linguistica : –. Greenberg, J. H. (). ‘Studies in African linguistic classification I: the Niger-Congo family’, Southwestern Journal of Anthropology : –. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Greenberg, J. H. (). ‘Niger-Congo noun class markers: prefixes, suffixes, both or neither’, Studies in African Linguistics, Supplement : –. Greenberg, J. H. (). ‘How does a language acquire gender markers?’, in J. H. Greenberg (ed.), Universals of Human Language. Vol. : Word Structure. Stanford, CA: Stanford University Press, –. Greenberg, J. H. (). ‘African linguistic classification’, in J. Ki-Zerbo (ed.), General History of Africa. Vol. I: Methodology and African Prehistory. Paris: UNESCO, –. Grinevald, C., and Seifart, F. (). ‘Noun classes in African and Amazonian languages: towards a comparison’, Linguistic Typology : –. Güldemann, T. (). ‘Grammaticalization’, in D. Nurse and G. Philippson (eds.), –. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), –. Güldemann, T. (). ‘Proto-Bantu and Proto-Niger-Congo: macro-areal typology and linguistic reconstruction’, in O. Hieda, C. König, and H. Nakagawa (eds.), Geographical Typology and Linguistic Areas, with special reference to Africa. Amsterdam and Philadelphia, PA: John Benjamins, –. Guthrie, M. (). ‘Bantu origins: a tentative new hypothesis’, Journal of African Languages : –. Hammarström, H. (). ‘Noun class parallels in Kordofanian and Niger-Congo: evidence of genealogical inheritance?’, in T. C. Schadeberg and R. M. Blench (eds.), –. Heggarty, P., Maguire, W., and McMahon, A. (). ‘Splits or waves? Trees or webs? How divergence measures and network analysis can unravel language histories’, Philosophical Transactions of the Royal Society. B: Biological Sciences : –. Heine, B. (). Die Verbreitung und Gliederung der Togorestsprachen. Berlin: Dietrich Reimer. Heine, B. (). ‘Language typology and convergence areas in Africa’, Linguistics : –. Heine, B. (). A Typology of African Languages, based on the order of meaningful elements. Berlin: Dietrich Reimer. Heine, B. (). ‘Methods in comparative Bantu linguistics (the problem of Bantu linguistic classification)’, in L. Bouquiaux (ed.), L’expansion bantoue: Actes du colloque international du CNRS, Viviers (France) – avril . Volume II. Paris: SELAF, –. Heine, B., and Nurse, D. (eds.) (). A Linguistic Geography of Africa. Cambridge: Cambridge University Press. Hoffmann, C. (). ‘An outline of the Dakarkari noun class system and the relation between prefix and suffix noun class systems’, in M. G. Manessy (ed.), La classification nominale dans les langues négro-africaines, Aix-en-Provence, – juillet . Paris: CNRS, –. Hyman, L. M. (). ‘Reflections on the nasal classes in Bantu’, in L. M. Hyman (ed.), Noun Classes in the Grassfields Bantu Borderland. Los Angeles, CA: University of Southern California, Department of Linguistics, –. Hyman, L. M. (). ‘The historical interpretation of vowel harmony in Bantu’, in J.-M. Hombert and L. M. Hyman (eds.), Bantu Historical Linguistics: Theoretical and empirical perspectives. Stanford, CA: Center for the Study of Language and Information, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Hyman, L. M. (). ‘How to become a “Kwa” verb’, Journal of West African Languages : –. Hyman, L. M. (). ‘Niger-Congo verb extensions: overview and discussion’, in D. L. Payne and J. Peña (eds.), Selected Proceedings of the th Annual Conference on African Linguistics. Somerville, MA: Cascadilla, –. Hyman, L. M. (). ‘The Macro-Sudan Belt and Niger-Congo reconstruction’, Language Dynamics and Change : –. Hyman, L. M., and Voorhoeve, J. (eds.) (). L’expansion bantoue: Actes du colloque international du CNRS, Viviers (France) – avril . Vol. I: Les classes nominaux dans le bantou des Grassfields. Paris: SELAF. Irvine, J. T. (). ‘Subjected words: African linguistics and the colonial encounter’, Language & Communication : –. Jenewari, C. E. W. (). ‘Ijoid’, in J. Bendor-Samuel (ed.), –. Johnston, Sir H. H. (). A Comparative Study of the Bantu and Semi-Bantu Languages. Vol. . Oxford: Clarendon. Kastenholz, R. (/). ‘Comparative Mande studies: state of the art’, Sprache und Geschichte in Afrika /: –. Katamba, F. X. (). ‘Bantu nominal morphology’, in D. Nurse and G. Philippson (eds.), –. Kießling, R. (). ‘On the origin of Niger-Congo nominal classification’, in R. Kikusawa and L. A. Reid (eds.), Historical Linguistics : Selected papers from the Twentieth International Conference on Historical Linguistics, Osaka, – July . Amsterdam and Philadelphia, PA: John Benjamins, –. Kleinewillinghöfer, U. (). ‘Relationship between Adamawa and Gur languages: the case of Waja and Tula’, Cahiers Voltaïques/Gur Papers : –. Koelle, S. W. (). Polyglotta Africana, or a comparative vocabulary of nearly three hundred words and phrases, in more than one hundred distinct African languages. London: Church Missionary House (repr. : Akademische Druck- u. Verlagsanstalt, Graz, Austria). Kropp Dakubu, M. E. (). Towards a phonology of Proto-Kwa: Onwards from Stewart’s “Potou-Akanic-Bantu”. Paper presented at the International Congress towards Proto-NigerCongo: Comparison and Reconstruction, September –, Paris, France, http://llacan.vjf.cnrs.fr/ fichiers/nigercongo/fichiers/Kropp_Dakubu-Proto-Kwa.pdf accessed Jul. , . Lefebvre, C., and Brousseau, A.-M. (). A Grammar of Fongbe. Berlin and New York, NY: Mouton de Gruyter. Lovegren, J. (). Mungbam Grammar. Ph.D. dissertation. Buffalo, NY: University at Buffalo. Lüpke, F., and Storch, A. (). Repertoires and Choices in African Languages. Berlin: Mouton de Gruyter. McGill, S., and Blench, R. M. (). ‘Documentation, development, and ideology in the northwestern Kainji languages’, in P. K. Austin and S. McGill (eds.), Language Documentation and Description. Vol. . London: SOAS, –. Maddieson, I. (). ‘Presence of uncommon consonants’, in M. S. Dryer and M. Haspelmath (eds.), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology, https://wals.info/chapter/ Maddieson, I., and Hinnebusch, T. J. (eds.) (). Language History and Linguistic Description in Africa. Trenton, NJ: Africa World Press. Maho, J. (). A Comparative Study of Bantu Noun Classes. Gothenburg: Acta Universitatis Gothoburgensis. Manessy, M. G. (). Les langues Oti-Volta: Classification généalogique d’un groupe de langues voltaïques. Paris: SELAF. Manfredi, V. (). ‘Igboid’, in J. Bendor-Samuel (ed.), –. Marchese, L. (). ‘Kru’, in J. Bendor-Samuel (ed.), –. Meeussen, A. E. (). ‘Bantu grammatical reconstructions’, Africana Linguistica : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
-, -
Miehe, G. (). Die Präfixnasale im Benue-Congo und Kwa: Versuch einer Widerlegung der Hypothese von der Nasalinnovation des Bantu. Berlin: Dietrich Reimer. Miehe, G., Reineke, B., and Winkelmann, K. (eds.) (). Noun Class Systems in Gur Languages. Vol. : North Central Gur Languages. Cologne: Rüdiger Köppe. Miehe, G., and Winkelmann, K. (eds.) (). Noun Class Systems in Gur Languages. Vol. : Southwestern Gur Languages (without Gurunsi). Cologne: Rüdiger Köppe. Möhlig, W. J. G. (). ‘The Bantu nucleus: its conditional nature and its prehistorical significance’, Sprache und Geschichte in Afrika : –. Möhlig, W. J. G. (). ‘Stratification in the history of the Bantu languages’, Sprache und Geschichte in Afrika : –. Moñino, Y. (). ‘Introduction: cousines ou voisines?’, in Y. Moñino (ed.), Lexique comparatif des langues oubanguiennes. Paris: Paul Geuthner, –. Moñino, Y. (). The position of Gbaya-Manza-Ngbaka group among the Niger-Congo languages. Paper presented at the workshop “Genealogical classification in Africa beyond Greenberg”, Berlin, Humboldt University, February –, , https://www.iaaw.hu-berlin.de/de/afrika/linguistik-undsprachen/veranstaltungen/greenberg-workshop/the-position-of-gbaya-manza-ngbaka-group-amongthe-niger-congo-languages accessed Jul. , . Moran, S., and Prokić, J. (). ‘Investigating the relatedness of the endangered Dogon languages’, Literary and Linguistic Computing : –. Mous, M. (). The Making of a Mixed Language: The case of Ma’a/Mbugu. Amsterdam and Philadelphia, PA: John Benjamins. Mous, M. (). ‘John M. Stewart, –’, Journal of African Languages and Linguistics : –. Moxley, J. L. (). ‘Semantic structure of Swahili noun classes’, in I. Maddieson and T. J. Hinnebusch (eds.), –. Naden, T. (). ‘Gur’, in J. Bendor-Samuel (ed.), –. Newman, P. (). On Being Right: Greenberg’s African linguistic classification and the methodological principles which underlie it. Bloomington, IN: Institute for the Study of Nigerian Languages and Cultures, African Studies Program, Indiana University. Nurse, D. (). Tense and Aspect in Bantu. Oxford: Oxford University Press. Nurse, D., and Philippson, G. (eds.) (a). The Bantu Languages. London and New York: Routledge. Nurse, D., and Philippson, G. (b). ‘Introduction’, in D. Nurse and G. Philippson (eds.), –. Olson, K. S. (). ‘On Niger-Congo classification’, in H. I. Aronson, D. L. Dyer, V. A. Friedman, D. S. Hristova, and J. M. Sadock (eds.), The Bill Question: Contributions to the study of linguistics and languages in honor of Bill J. Darden on the occasion of his sixty-sixth birthday. Bloomington, IN: Slavica, –. Pozdniakov, K. (). ‘Niveaux linguistiques et problèmes de reconstruction dans les langues atlantiques’, Sprache und Geschichte in Afrika : –. Pozdniakov, K. (). ‘Diachronie des classes nominales atlantiques: morphophonologie, mophologies, sémantique’, in D. Creissels and K. Pozdniakov (eds.), –. Pulleyblank, D. (). ‘Niger-Kordofanian languages’, in B. Comrie (ed.), The World’s Major Languages, nd edn. Abingdon and New York: Routledge, –. Salffner, S., and Sands, B. (). Why Ukaan is hard to classify. Paper presented at the International Congress towards Proto-Niger-Congo: Comparison and Reconstruction, September –, Paris, http://llacan.vjf.cnrs.fr/fichiers/nigercongo/abstracts/salffner_sands.pdf, accessed Jul. , . Samarin, W. J. (). ‘Adamawa-Eastern’, in T. A. Sebeok (ed.), –. Sands, B. (). ‘Africa’s linguistic diversity’, Language and Linguistics Compass : –. Schadeberg, T. C. (a). ‘The classification of the Kadugli language group’, in T. C. Schadeberg and M. L. Bender (eds.), Nilo-Saharan: Proceedings of the First Nilo-Saharan Linguistics Colloquium, Leiden, September –, . Dordrecht: Foris, –. Schadeberg, T. C. (b). ‘Das Kordofanische’, in B. Heine, T. C. Schadeberg, and E. Wolff (eds.), Die Sprachen Afrikas. Hamburg: Helmut Buske, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Schadeberg, T. C. (c). ‘Die Geschichte der Nominalklassen des Laru (Kordofanisch)’, in H. Jungraithmayr (ed.), Berliner afrikanistische Vorträge (xxi. Deutscher Orientalistentag, Berlin .–..). Berlin: Dietrich Reimer, –. Schadeberg, T. C. (). ‘The lexicostatistic base of Bennett & Sterk’s reclassification of Niger-Congo with particular reference to the cohesion of Bantu’, Studies in African Linguistics : –. Schadeberg, T. C. (). ‘Kordofanian’, in J. Bendor-Samuel (ed.), –. Schadeberg, T. C. (). ‘Number in Swahili grammar’, Afrikanistische Arbeitspapiere : –. Schadeberg, T. C. (a). ‘Derivation’, in D. Nurse and G. Philippson (eds.), –. Schadeberg, T. C. (b). ‘Historical linguistics’, in D. Nurse and G. Philippson (eds.), –. Schadeberg, T. C., and Blench, R. M. (eds.) (). Nuba Mountain Language Studies. Cologne: Rüdiger Köppe. Sebeok, T. A. (ed.) (). Current Trends in Linguistics. Vol. : Linguistics in Sub-Saharan Africa. The Hague: Mouton. Stahlke, H. F. W. (). ‘Serial verbs’, Studies in African Linguistics : –. Stallcup, K. (). ‘La Géographie linguistique des Grassfields’, in L. M. Hyman and J. Voorhoeve (eds.), –. Stewart, J. M. (). ‘Kwa’, in J. Bendor-Samuel (ed.), –. Stewart, J. M. (). ‘The potential of Proto-Potou-Akanic-Bantu as a pilot Proto-Niger-Congo, and the reconstructions updated’, Journal of African Languages and Linguistics : –. Stewart, J. M. (). ‘Consonant mutation in Proto-Potou-Akanic-Bantu and the Fula-type languages of Senegal and Guinea’, in W. A. A. Wilson and A. Storch (eds.), Guinea Languages of the Atlantic Group: Description and internal classification. Frankfurt am Main: Peter Lang, –. Storch, A. (). Secret Manipulations: Language and context in Africa. Oxford: Oxford University Press. Voeltz, F. K. E. (). Proto Niger-Congo Verb Extensions. Ph.D. dissertation. Los Angeles, CA: UCLA. Voorhoeve, J. (). ‘Bantu et bane’, in L. M. Hyman and J. Voorhoeve (eds.), –. Wallis, B. M. (). Diedrich Westermann’s Die westlichen Sudansprachen and the Classification of the Languages of West Africa. Ph.D. dissertation. Evanston, IL: Northwestern University. Wedekind, K. (). ‘Thoughts when drawing a map of tone languages’, Afrikanistische Arbeitspapiere : –. Welmers, W. E. (a). ‘Niger-Congo, Mande’, in T. A. Sebeok (ed.), –. Welmers, W. E. (b). ‘The typology of the Proto-Niger-Kordofanian noun class system’, in ChinWu Kim and H. Stahlke (eds.), Papers in African Linguistics. Carbondale, IL and Edmonton, AB: Linguistic Research Inc., –. Welmers, W. E. (). African Language Structures. Berkeley, CA: University of California Press. Westermann, D. (). Die Sudansprachen: Eine sprachvergleichende Studie. Hamburg: L. Friederichsen. Westermann, D. (). Die westlichen Sudansprachen und ihre Beziehungen zum Bantu. Berlin: Walter de Gruyter. Williamson, K. (). ‘The Benue-Congo languages and Ịjọ’, in T. A. Sebeok (ed.), –. Williamson, K. (). ‘How to become a Kwa language’, in A. Makkai and A. K. Melby (eds.), Linguistics and Philosophy: Essays in honor of Rulon S. Wells. Amsterdam: John Benjamins, –. Williamson, K. (a). ‘Benue-Congo overview’, in J. Bendor-Samuel (ed.), –. Williamson, K. (b). ‘Niger-Congo overview’, in J. Bendor-Samuel (ed.), –. Williamson, K., and Blench, R. M. (). ‘Niger-Congo’, in B. Heine and D. Nurse (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
̈
. O
.................................................................................................................................. Hitherto, the comparative study of African languages has tended to be oversimplistic, based on assumptions that the complex language-map of today has evolved from a much simpler language-map in the past. The consequent belief that the many hundreds of modern languages in Africa may be traced back to a handful of ‘proto-languages’ has led comparative linguists to concentrate on the minority of common material uniting large numbers of African languages (i.e., the overlap between their apparent cores of direct retention) to the neglect of the majority of their non-common material; it has also led them to concentrate on the process of divergence among languages with an assumed common origin, to the exclusion of the process of convergence among unrelated languages or of re-convergence among related languages. (Dalby : )
O hundred and sixty-six years after the first mention of a group of “North West Atlantic languages” in Koelle’s Polyglotta Africana (), our understanding of the relatedness of the languages subsumed under this and similar labels has come almost full circle. What was a geographically and typologically motivated grouping in the beginning, followed by attempts to integrate the languages into the Niger-Congo family tree as a genetic group of languages, is now again seen as a conglomerate of very diverse languages, with a shrinking and disputed genetically interpreted core. What motivates grouping them together as Atlantic languages thus most convincingly remains their location in a geographical area of West Africa, along the Atlantic coast, the longstanding social interactions of its inhabitants, and the major typological difference of Atlantic languages from the neighboring Mande languages. In addition, it remains remarkable that many of the languages share the features of noun classes, initial consonant mutation, [ATR] vowel harmony, labiovelar consonants, and verb extensions or a subset of them, whose status as inherited, typological, or acquired through language contact is uncertain. With the exception of Fula (distributed through the entire Sahel), Atlantic languages can be found in an area extending from Liberia in the south to Mauretania in the north,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
¨
reaching from the Atlantic coast inland and interspersed with other languages or families. Because of the prominence of Mande for the conceptualization of Atlantic, for historical interaction and entangled populations, Map . shows the approximative distribution of both Atlantic and Mande languages. Apart from Wolof, Temne, and the different Fula varieties, most Atlantic languages are spoken by small multilingual speech communities, many of which are multilingual in or in contact with speakers of other Atlantic and/ or Mande languages, Portuguese-based Creole, English-based Krio, and the ex-colonial languages French, English, and Portuguese. In the past, a strong tendency to underdifferentiate linguistic diversity and regard ethnic identity as equivalent with linguistic identity resulted in subsuming different languages under one language label, as is the case of Joola or Baïnounk. Recent descriptive research points to the existence of a larger number of languages than reported so far, and of the misalignment between languages and identity. Many Atlantic languages can be classified as endangered, which means that basic research on them is of great urgency. Wilson () presents a good overview of the inventory of languages, the history and ongoing debate surrounding their classification, and their typological characteristics. The reader is referred to his article and to Childs () for recent accounts of Atlantic. The facts presented in these works will not be reiterated here, but new hypotheses, challenges, findings, and avenues for future research will be discussed instead.
. The geographical distribution of Atlantic and Mande languages
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. A
.................................................................................................................................. The two features that make Atlantic a meaningful entity are typology and geographical distribution. It is striking that typologically the Mande languages are, in important points, the very reverse of Atlantic and are genetically totally unrelatable to them. Despite the impossibility of proving genetic relationship among Atlantic as a whole—though it is clear within several small sets of the languages—membership of Atlantic has in general been unchallenged. (Wilson : –)
There is no basis to claims that Atlantic languages as a whole constitute a genetic group. Recent lexicostatistical calculations (Pozdniakov and Segerer ) based on new lexical data not available at the time of Sapir’s () classification confirm the low percentages of resemblances presented in Sapir’s table. While Sapir is able to identify subgroupings, he admits that: The major bone of contention must remain the over-all unity of West Atlantic. The low percentages (less than percent) leave us pretty much in the same place as Westermann did in his study. (Sapir : )
Sapir cautiously suggests that the languages might be genetically related, but concedes that the resemblances must remain hypothetical: matches were accepted as putative cognates if each phoneme in a CVC sequence was either identical in the compared forms or varied by no more than a single phonetic feature. Thus b=p and p=f were accepted, but not b=f. N=N was always accepted, but N=NC was accepted only if homo-organic, i.e. m=mp but not m=ŋg etc. When an identical match was found between contiguous but not closely related languages (e.g. Noon: Wolof) the cognate was rejected as a loan. . . . The aim of this count was to permit reasonable hypotheses as to the degree of relationship. Absolute proof is something else and must come, of course, via the comparative method. (Sapir : )
Sapir’s heuristic to identify putative cognates through differences no greater than one phonetic feature instead of reconstructing sound changes is not appropriate. Additionally, it is rendered problematic by the complex morphophonology of many Atlantic languages, which exhibit initial consonant mutation (see section ..). Since Sapir, new methods applied essentially to the same data have confirmed Sapir’s percentages and shed new light on their interpretation. The application of new methods of computing phylogenetic networks such as the programme “SplitsTree” (Huson and Bryant ) by Pozdniakov and Segerer () presents evidence that many of the resemblances
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
¨ “Final” result - 3 large groupings - North 1 - North 2 - South - Differences compared to Sapir - wolof-nyun-buy - bijogo with Bak - South with Mel
balante manjaku, mankanya, pepel joola central karon, mlomp kwaatay bayot bijogo nyun, buy wolof cangin tenda biafada, jaad peul sereer nalu, baga mboteni baga temne kisi sherbro, krim sua gola limba
. New grouping superseding Atlantic (Pozdniakov and Segerer )
between Atlantic languages are not based on descent from a common ancestor language but on horizontal transfer between languages. Local subgroupings of languages, such as the Tenda languages (Bapen, Tanda, Basari, Bedik, Konyagi), the Cangin languages (Lehar, Noon, Safen, Ndut, Palor), the Joola cluster (including Kerak, Kasa, Bliss, Gusilay, Banjal, Fogny, Ejamat, Karon, Mlomp, and Kwaatay) and the Manjaku-Bok, Mankanya, and Pepel continuum, are confirmed as genetically related. Tree-like structures emerge for three groupings given in Figure ., based on the “SplitsTree” network diagrams in Figures . and .. However, given the extreme distances between some of the languages in the new groupings, also mirrored in the low percentage of resemblances given by Sapir (), between, e.g. Wolof and Nyun-Nyamone (or Baïnounk Guñaamolo) of a mere %, these groupings appear more doubtful than the ones uniting languages spoken in adjacent locations, like the Joola languages. Pozdniakov and Segerer (forthcoming) have abandoned working on phylogenetic networks and now pursue genetic classification based on shared innovations. The result is a radically shrunk Atlantic family with two branches (Figures .–), and the elimination of their () south branch, although this classification is not shared by Childs (forthcoming), who posits that genetic unity can only be proven at the family level (Bak, Mel, etc.) and not beyond. It is unfortunate that Pozdniakov and Segerer have abandoned modeling affinities through SplitsTree or Neighbor-net (better suited to visualize networks rather than trees, see McMahon and McMahon ), since it would be of great relevance for a better understanding of the relatedness of Atlantic languages to not only include tree-like structures, indicating vertical relations, but also consider the network-like structures, visualized through horizontal reticulations, and not contained in Figure .. These reticulations point to resemblances not inherited from a proto-language but presenting hybridizations caused by language contact. Even in a diagram such as Figure . that does not take
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Result (1) Joola-Gusilay Joola-Banjal Joola-Ejamat Joola-Bliss Kwaatay Joola-Kasa Joola-Kerak Joola-Fogny Mankanya Joola-Mlomp Manjaku-Bok Karon Pepel Bayot Bijogo-Kagbaaga Balante-Fca Jaad
Nalu
Biafada
Baga-Mboteni Konyagi
ROOT
Tanda Basari Bapen Bedik
Gola Kisi Krim Sherbro
Pular Sereer
Sua Ndut Lehar PalorNoon Safen
Temne Nyun-Nyamone Limba-Biriwa Wolof Buy-Kobiana
. “SplitsTree” diagram with reticulations for Atlantic languages (Pozdniakov and Segerer )
Result (2) Manjaku-Bok Balante-Fca Nalu Pepel Sua Mankanya Baga-Mboteni Sherbro Bayot Kwaatay Krim Joola-Mlomp Kisi Karon Temne Joola-Ejamat Joola-Fogny Gola Joola-Banjal Limba-Biriwa Joola-Gusilay Joola-Bliss Joola-Kasa Joola-Kerak Pular Sereer Bijogo-Kagbaaga Biafada Jaad Buy-Kobiana Nyun-Nyamone Wolof
Ndut Palor Lehar Noon Safen
Tanda Basari Bedik Bapen Konyagi
. “SplitsTree” diagram without reticulations (Pozdniakov and Segerer )
ROOT
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
¨ WOLOF Wolof, Lebu Nyun (Gunyaamolo, Gujaher, Gubëeher...); Buy (Kasanga, Kobiana) TENDA Basari, Tanda, Bedik, Bapen; Konyagi
NYUN-BUY NORTH
JAAD Biafada; Badiaranke FULA-SEREER Fula (Pular, Pulaar, Fulfulde...); Sereer
CANGIN Palor, Ndut; Noon, Laala, Saafi ATLANTIC
NALU Nalu; Bage Fore; Baga Mboteni BALANT Ganja, Kentohe, Fraase JOOLA Fogny, Banjal, Kasa, Kwaatay, BAK
Karon, Ejamat, Keeraak...; Bayot ?
MANJAKU Bok, Cur, Bassarel; Pepel; Mankanya BIJOGO Kamona, Kagbaaga, Kajoko
. New classification of Atlantic (Pozdniakov and Segerer forthcoming)
the largest part of the lexical inventory of these languages into account by including only items from the Swadesh list judged to be most resistant to borrowing, reticulations point to the transfer of lexical material between languages through contact. If more lexical data were to be included, a very different, but much more complete picture of relatedness would emerge.
. C ?
.................................................................................................................................. Pourquoi ne dispose-t-on pour aucun groupe de langues, excepté le bantu, d’un système fiable de correspondances phonétiques, et par conséquent, des reconstructions convaincantes? À qui la faute? Aux linguistes, à la méthode, ou aux langues elles-mêmes? (Pozdniakov : )
If one wants to retain a genealogical perspective, ultimately, the genetic status of the suggested new groupings superseding older Atlantic classifications as genetic, and their relation to Niger-Congo can only be determined through the application of the comparative method—a method that traces the history of a group of languages suspected to be genealogically related based on a range of linguistic features by using word lists in order to identify resemblances and, by assuming regular sound changes, reconstructing the vocabulary of proto-languages and identifying cognates, that is, words inherited from them. So far, the comparative method has not been employed for Atlantic languages, and there is little hope that it will be in the near future. The main reason for this lies in the challenges posed by the complex morphophonology of these languages, and in particular in the presence of initial consonant mutation, as convincingly argued by Pozdniakov (). Most Atlantic languages exhibit alternations of the initial consonant of the radical that serve to express number, noun class membership, and other functions. The alternations
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
consist in a change of mode of articulation, in maximally three series; cf. Sereer saax ‘village’, a-caax ‘villages’, and o-njaax ‘small village’ (McLaughlin –: ). The presence of consonant mutation renders the application of the comparative method all but impossible, since a given consonant can belong either to one series of alternations or to several series, and can exhibit different behavior in different languages, thus extending the range of possible sound correspondences to a set too wide to allow the identification of regular sound changes based on sound–meaning correspondences. The proliferation of possible correspondences has been exemplified by Pozdniakov (: ) for Sereer and Konyagi in Table . below.
Table . Possible Konyagi correspondences for Sereer b- (Pozdniakov : ) Sereer
b
b
b
b
b
b
b
b
b
b
b
b
Konyagi
mp
b
w
p
p
ϕ
nkw
gw
w
kw
kw
xw
As research progresses, we must take note of the sustained inapplicability of the comparative method and the continually decreasing evidence for genetic relatedness of Atlantic languages. More crucially, epistemological concerns about the validity of the tree model, the regularity of change, and the separability of monoparental diversification from contact-induced influence as instigators of language change (e.g. McMahon and McMahon ; Mufwene ;) put its explanatory power into question more than ever.
. A A?
.................................................................................................................................. contact might be just as important if not more important than genetic inheritance for understanding the structures and relatedness of African languages. (Childs : )
If Atlantic as a genetically motivated unit has become obsolete or very limited in scope, is there any rationale for maintaining the label Atlantic other than as a term of convenience to refer to a group of languages spoken at the fringe of the Mande area that are typologically maximally distinct from Mande languages? It appears that integrated frameworks, which do not attempt to distinguish between inherited and contact-induced language features, are of greater promise for comprehending the entanglements of inhabitants of the area through history and the traces they have left in language. While language contact is an important factor in creating affinities between languages, alongside descent from a common ancestor language, it can only be set apart from shared ancestry through a knowledge of history. Yet language contact has never been treated on a par with genealogical relatedness by the dominating frameworks of historical linguistics, and, as a consequence, our understanding of language history is skewed towards looking at common ancestry only, usually modeled in a tree-like structure.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
¨
Recent work, however, has emphasized the importance of language contact for Africa as a whole and a number of areas within it (Heine and Nurse ), and for the zone of Mande–Atlantic contact (Childs ). With respect to the features defining his “MacroSudan belt”—logophoricity, labiovelars, labial flap consonants, [ATR] vowel harmony, S–AUX–O–V–X word order, and V–O–NEG ordering—Güldemann (: ) concedes that: Some lineages, which in geographical terms are all peripheral but adjacent to the core, display an ambiguous behavior regarding linguistic commonalities with this area. These lineages are Atlantic, Dogon, Songhay, Chadic, Ijoid, Narrow Bantu, and Nilotic.
Presenting extensive evidence from Kisi, and drawing on the history of the area, Childs () also emphasizes the importance of language contact for Africa and the Atlantic space, stating: Africa features a number of long-standing contact situations between groups speaking unrelated languages. In a broad band across the sub-Saharan region from east to west many such situations can be identified, including the Atlantic Mande contact region of western West Africa. The interaction between speakers of Atlantic languages and speakers of Mande languages has pointed predominantly in only one direction as to (linguistic) influence, namely, from Mande to Atlantic. Why this is so can be explained with reference to historical and socio-cultural factors. Although there are exceptions to this directionality, the exceptions actually reinforce these explanations. (Childs : )
While contact between Atlantic autochthones and Mande-speaking incomers certainly has shaped the languages in question, contact of the relatively small Atlantic communities with speakers of other Atlantic languages alongside Mande languages through a number of past and present social practices, by far not all of them involving conflict, also remains part and parcel of everyday life for most language communities. Alongside contact caused by asymmetrical power relations between settlers and intruders (which has occurred between populations speaking Atlantic languages as well, for instance between speakers of Joola languages and speakers of Baïnounk languages; cf. Brooks ; Bühnen ), contact and multilingualism are also conditioned by more peaceful exchanges deeply embedded into the fabric of Atlantic societies. To summarize, crucial for the coexistence of several languages in individuals and communities are the following strategies for exchange, mobility, and cohesion (see Cobbinah et al. ; Lüpke and Storch ; Lüpke , ), in addition to those identified by Childs (): • patrimonial language ideologies territorializing language through associating the language of its male founders to a place, and giving these strangers land rights and the authority to settle strangers; • marriage patterns with incoming wives speaking other languages, often through longstanding marriage networks; • language acquisition in peer groups and age classes cross-cutting language boundaries in multilingual communities, with federal initiation ceremonies uniting generations of initiates;
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
• widespread fostering across language boundaries; • joking relationships, secret societies, religious systems, and patronymic equivalences beyond ethnolinguistic boundaries; • mobility and migration of both men and women for trade, ritual, religious, economic, language-learning, and educational purposes. Despite its great role in shaping the lexicon and structures of languages throughout West Africa, language contact has not been a topic of major concern in African linguistics until very recently (see also Childs ), not just because of a preoccupation with genealogical relatedness, as also noted by Pozdniakov (), but also because of a lack of data on languages in contact, if this contact does not involve a European and a major African language. However, only a systematic investigation of societal and individual patterns and motivations of language contact and multilingualism and their linguistic consequences will allow a conclusive answer to the question whether Atlantic can be re-conceptualized as a linguistic area. There are some recent contributions raising awareness of contact phenomena and calling for more and concerted research (Cobbinah ; Lüpke a, b; Lüpke and Chambers ) in the Mande–Atlantic sphere and beyond. Yet language contact studies quickly run into similar epistemological problems as genealogical studies do: the simple fact that it is an artifact of nineteenth-century language ideologies to insist on distinguishing between two kinds of linguistic change (see McMahon and McMahon ; Mufwene ; Watson ; Lüpke forthcoming a; and Lüpke and Watson forthcoming), in particular in areas where contact is common between genetically close languages, whose features are therefore constantly reconfigured.
. C
.................................................................................................................................. Nihil est in lingua quod non prius fuerit in oratione. (Benveniste : )
The descriptive situation for Atlantic languages is dramatically improving, due to fieldbased description, documentation, and sociolinguistic studies taking place after long periods of inaccessibility of many language areas due to civil wars and unrest. An important number of Atlantic languages have been documented and described in the past decade, either individually (see the grammatical descriptions united in Lüpke forthcoming b) or within the “Sénélangues” project, a major research project that regrouped twenty-four researchers from Europe and Senegal (http://senelangues.huma-num.fr/ and Creissels and Pozdniakov ). These new descriptions facilitate work on the historical-comparative challenge that Atlantic languages pose, since new and more reliable lexical material on more languages becomes available. Lexical material on Atlantic languages and beyond is being compiled by Guillaume Segerer in the online database, “RefLex” (http://reflex.cnrs.fr/database/), a reference lexicon for African languages.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
¨
Crucially, detailed descriptions and documentations of language use for more Atlantic languages, and in particular those in contact with each other will enable future research to go beyond lexical comparison and the presence of typological features such as vowel harmony, consonant mutation, or noun classes and start investigating structural outcomes of language contact such as semantic and syntactic calques or metatypies, areas of convergence and divergence in different domains, overlaps and differences in prosody and gesture, language acquisition in multilingual contexts, and the investigation of variation and change in speech situations involving multilingual networks. A growing body of anthropological linguistic and sociolinguistic research on the dynamics of multilingualism and code interaction in the area, on the linguistic and social representation of languages through history, and on the complex interplay of factors determining linguistic and ethnic identity (which do not coincide in many of the area’s settings) is available (e.g. Haust ; Juillard ; Dreyfus and Juillard ; Irvine ; McLaughlin , ; Ndecky ; Juffermans ; Nunez ; Nunez and Léglise ; Lüpke , , forthcoming a; Cobbinah et al. ; Goodchild ; van den Avenne ), and the website of the Crossroads project (https://soascrossroads.org/). Most of this research leaves aside old dichotomies between genetic and contact-induced features and between inheritance, borrowing, and code-switching in spoken and written language use. Rather, these studies focus on the social patterns motivating and constraining multilingual speech, language ideologies, and their impact on language use and the categorization of multilingual speech, as well as on historical continuity and change in social and linguistic organization. In order to advance our understanding of Atlantic, it seems most promising to combine these strands of research—historical-comparative research, synchronic description and documentation, and detailed sociolinguistic and anthropological-linguistic studies—to reveal all types of relatedness of languages in the geographical area of the Atlantic space, and to unveil the social, cultural, and political factors that led and are leading speakers to the kind of linguistic behavior that is reflected in the divergent and convergent features in their languages today. As Campbell () and Mufwene () remind us, only verifiable histories of speakers and their intermingling allow interpreting linguistic convergences and divergences. To give only one example of the need to combine historical and interdisciplinary linguistic perspectives: Baïnounk languages are the only cluster in the Nyun-Buy group of Atlantic that do not exhibit stem-initial consonant mutation. Pozdniakov and Segerer (forthcoming) ascribe the absence of this feature to longstanding contact with Joola languages, none of which exhibits consonant mutation. However, Baïnounk Gujaher, whose speakers have been in close contact and have intermarried with speakers of Kasanga (also in the Nyun-Buy group and featuring consonant mutation), since at least the sixteenth century (Bühnen ), does not exhibit consonant mutation. Mere contact and/or multilingualism is thus not sufficient in order to predict language change. A knowledge of micro-historical settlements and exchanges combined with an exploration of social indexical functions of languages (Silverstein ) is required to allow interpreting convergences and divergences. In the case of the astonishing distance between Kasanga and Baïnounk Gujaher (see Lüpke forthcoming a for a detailed discussion), it is existence of a political and matrimonial alliance based on difference and indexed linguistically that has kept them apart through centuries.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Recent works by descriptive linguists stress the importance of holistic approaches to the study of African languages combining structural and sociolinguistic viewpoints (Lüpke a; Childs et al. ), and sociolinguistic scholars have emphasized the central importance of multilingualism and language contact for variation and change on this continent (Barasa ; Anderson and Ansah ; Mesthrie ). More of such research in an interdisciplinary perspective is therefore desirable, as it constitutes the most promising avenue to disclosing why Atlantic languages or a subset of them resemble each other, and in which features; why other languages and features do not, despite longstanding proximity; and why these opposing forces cannot be captured (or only to a very limited extent) in genetic terms.
Q Crossroads Project, https://soascrossroads.org/, accessed //. RefLex. http://reflex.cnrs.fr/database/, accessed //. Sénélangues. http://senelangues.huma-num.fr/, accessed //.
R Anderson, J. A., and Ansah, G. N. (). ‘A sociolinguistic mosaic of West Africa: challenges and prospects’, in D. Smakman and P. Heinrich (eds.), –. Auzeanneau, M., Bento, M., and Leclère, M. (eds.) (). Pratiques plurilingues, mobilités et éducation: Éclairages d’Afrique ou d’ailleurs. Paris: Édition des Archives Contemporaines. Barasa, S. N. (). ‘Ala! Kumbe? “Oh my! Is it so?”: multilingualism controversies in East Africa’, in D. Smakman and P. Heinrich (eds.), –. Benveniste, E. (). Problèmes de linguistique générale. Paris: Gallimard (I). Brooks, G. E. (). Landlords and Strangers: Ecology, society and trade in western Africa, –. Boulder, CO: Westview Press. Bühnen, S. (). Geschichte der Bainunk und Kasanga. Ph.D. thesis. Gießen: Justus-Liebig-Universität. Campbell, L. (). ‘Areal linguistics: a closer scrutiny’, in Y. Matras et al. (eds.), –. Childs, G. T. (). An Introduction to African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Childs, G. T. (). ‘The Mande and Atlantic groups of Niger-Congo: prolonged contact with assymetrical consequences’, in F. Lüpke and M. Chambers (eds.), –. Childs, G. T. (forthcoming). ‘Genetic clusters in Atlantic’, in F. Lüpke (ed.). Childs, G. T., Good, J., and Mitchell, A. (). ‘Beyond the ancestral code: towards a model for sociolinguistic language documentation’, Language Documentation and Conservation : –. Cobbinah, A. (). ‘The Casamance as an area of intense language contact: the case of Baïnounk Gubaher’, in F. Lüpke and M. Chambers (eds.), –. Cobbinah, A., Hantgan, A., Lüpke, F., and Watson, R. (). ‘Carrefour des langues, carrefour des paradigmes’, in M. Auzeanneau et al. (eds.), –. Creissels, D., and Podzniakov, K. I. (eds.) (). Les classes nominales dans les langues atlantiques. Cologne: Rüdiger Köppe. Dalby, D. (). ‘A referential approach to the classification of African languages’, in Chin-Wu Kim and H. Stahlke (eds.), Papers in African Linguistics. Carbondale, IL, and Edmonton, AB: Linguistic Research Inc., –. Dreyfus, M., and Juillard, C. (). Le plurilinguisme au Sénégal: Langues et identités en devenir. Paris: Karthala.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
¨
Goodchild, S. (). ‘ “Which language(s) are you for?” “I am for all the languages.” Reflections on breaking through the ancestral code: trials of sociolinguistic documentation’, SOAS Working Papers in Linguistics : –. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), –. Haust, D. (). Codeswitching in Gambia: Eine soziolinguistische Untersuchung von Mandinka, Wolof und Englisch in Kontakt. Cologne: Rüdiger Köppe. Heine, B., and Nurse, D. (). A Linguistic Geography of Africa. Cambridge and New York: Cambridge University Press. Huson, D. H., and Bryant, D. (). ‘Application of phylogenetic networks in evolutionary studies’, Molecular Biology and Evolution : –. Irvine, J. T. (). ‘Subjected words: African linguistics and the colonial encounter’, Language & Communication : –. Juffermans, K. (). Local Languaging, Literacy and Multilingualism in a West African Society. Bristol: Multilingual Matters. Juillard, C. (). Sociolinguistique urbaine: La vie des langues à Ziguinchor (Sénégal). Paris: Presses du CNRS. Koelle, S. W. (). Polyglotta Africana (repr. of London edn). Graz: Akademische Druck- und Verlagsanstalt. Lüpke, F. (a). ‘Language and identity in flux: in search of Baïnouk’, in F. Lüpke and M. Chambers (eds.), –. Lüpke, F. (b). ‘Multilingualism and language contact in West Africa: towards a holistic perspective’, in F. Lüpke and M. Chambers (eds.), –. Lüpke, F. (). ‘Pure fiction—the interplay of indexical and essentialist ideologies and heterogeneous practices: A view from Agnack’, Language Documentation and Conservation—Special Publication : –. Lüpke, F. (). ‘Multiple choice: language use and cultural practice in rural Casamance between convergence and divergence’, in J. Knörr and W. T. Filho (eds.), Creolization and Pidginization in Contexts of Postcolonial Diversity: Language, culture, identity. Leiden: Brill, –. Lüpke, F. (forthcoming a). ‘Language contact in West Africa’, in M. Aronoff (ed.), Oxford Research Encyclopedia in Linguistics. Oxford: Oxford University Press. Lüpke, F. (ed.) (forthcoming b). The Oxford Guide to the Atlantic Languages of West Africa. Oxford: Oxford University Press. Lüpke, F., and Chambers, M. (eds.) (). ‘Multilingualism and language contact in West Africa: towards a holistic perspective‘, Special Issue of Journal of Language Contact. Thema Series (), https://brill.com/view/journals/jlc///jlc..issue-.xml Lüpke, F., and Storch, A. (). Repertoires and Choices in African Languages. Berlin and New York: Mouton de Gruyter. Lüpke, F., and Watson, R. (forthcoming). ‘West Africa’, in E. Adamou and Y. Matras (eds.), The Routledge Handbook of Language Contact. London: Routledge. Matras, Y., McMahon, A., and Vincent, N. (eds.) (). Linguistic Areas: Convergences in historical and typological perspective. Houndmills: Palgrave Macmillan. McLaughlin, F. (–). ‘Consonant mutation in Seereer-Siin’, Studies in African Linguistics : –. McLaughlin, F. (). ‘On the origins of Urban Wolof: evidence from Louis Descemet’s phrase book’, Language in Society : –. McLaughlin, F. (). ‘The social meanings of Wolof and French: contact dialects, language ideology, and competing modernities in Senegal’, in R. Bassiouney (ed.), Identity and Dialect Performance: A study of communities and dialects. London: Routledge, –. McMahon, A., and McMahon, R. (). ‘Keeping contact in the family: approaches to language classification and contact-induced change’, in Y. Matras et al. (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Mesthrie, R. (). ‘Towards a distributed sociolinguistics of postcolonial multilingual societies: the case of Southern Africa’, in D. Smakman and P. Heinrich (eds.), –. Mufwene, S. (). ‘Population movements and contacts in language evolution’, Journal of Language Contact THEMA : –. Ndecky, A. (). Pratiques et répresentations des parlers mancagnes de Goudomp (Sénégal). Ph.D. thesis. Amiens: Université de Picardie Jules Vernes, Laboratoire d’Études Sociolinguistiques sur les Contacts des Langues et la Politique Linguistiques (LESClAP). Nunez, J.-F. (). L’alternance entre créole afro-portugais de Casamance, français et wolof au Sénégal: Une contribution trilingue à l’étude du contact de langues. Ph.D. thesis. Paris: INALCO. Nunez, J.-F., and Léglise, I. (). ‘Ce que les pratiques langagières plurilingues au Sénégal disent à la linguistique du contact’, in M. Auzeanneau et al., –. Pozdniakov, K. (). ‘Études comparatives atlantiques—questions de métholodogie’, in Société de Linguistique de Paris (ed.), Tradition et rupture dans les grammaires comparées de différentes familles. Leuven: Peeters, –. Pozdniakov, K., and Segerer, G. (). Classification des langues atlantiques. Ndayane: Sénélangues. Pozdniakov, K., and Segerer, G. (forthcoming). ‘A genealogical classification of Atlantic languages’, in F. Lüpke (ed.). Sapir, J. D. (). ‘West Atlantic’, in T. A. Sebeok (ed.), Current Trends in Linguistics. Vol. : Linguistics in Sub-Saharan Africa. The Hague: Mouton, –. Silverstein, M. (). ‘Indexical order and the dialectics of sociolinguistic life’, Language & Communication : –. Smakman, D., and Heinrich, P. (eds.) (). Globalising Sociolinguistics: Challenging and expanding theory. London and New York: Routledge. van den Avenne, C. (). De la bouche même des indigènes: Échanges linguistiques en Afrique coloniale. Paris: Éditions Vendémiaire. Watson, R. (). ‘Patterns of lexical correlation and divergence in Casamance’, Language & Communication, –. DOI: ./j.langcom.... Wilson, W. A. A. (). ‘Atlantic’, in J. Bendor-Samuel (ed.), The Niger-Congo Languages. New York and London: Lanham, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. M speech communities are found all over West sub-Saharan Africa, from the Atlantic coast to the banks of the river Niger in Nigeria (see Map .). The most recent estimates of languages in terms of linguistically and sociopolitically distinct units range from to (: Kastenholz ; : Dwyer ; : Vydrin ; : Lewis ), according to varying criteria of language–dialect distinction. The geolinguistic distribution pattern of this genealogical unit reflects the multilayered history of its speech communities and, noteworthy, of the states that Manding groups had established during the medieval golden age of transSaharan trade in slaves, gold, salt, and ivory. The linguistic patterning of Mande nevertheless indicates also other kinds of destinies, as isolated and threatened linguistic minorities. The functional status of Mande languages in today’s African societies and states is likewise diverse. None of them has become an official language, and only languages with larger speech communities have been declared national languages. The Manding vernaculars subsumed under the term “Jula” (Dioula, Dyula) have been widely used for centuries as a means of interethnic communication. It is Manding languages such as Bambara and Maninka with millions of speakers that defend their former linguistic dominance in quotidian domains against the official language, French, in most of the Mandephone countries, while the status of the written press even in the dominant and widespread Manding languages is rather insignificant because of the overall low level of reading practice and, thus, remains at an initial stage (Dumestre and Canut ). However, Mande languages are becoming more and more dominant on the radio and in the media, and in popular culture in almost all countries of former French West Africa. Recently, state-owned and private radio programs in the Mandephone countries have been broadcasting to a much higher degree in national languages and producing online streaming audio content that is also accessible to the diaspora.
. H
.................................................................................................................................. Written documents in Mande languages prior to the colonial period are scarce. In , however, Ajami glosses and colophons on Arabic manuscripts from the eighteenth and
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Mande languages
nineteenth century of an unidentified variety of Soninke have been recovered by Nikolay Dobronravin in the collection of Trinity College, Dublin (TCD MSS , , , , ). No early medieval manuscripts and translations from Arabic are known at present; yet to this day only a few manuscript collections have been established, and the cultural impact of Islam on the Mali Empire at that time leads us to assume at least some activity in literacy. An edition of more recent ta:rikhas in Arabic and Ajami Mandinka has been published by Giesing and Vydrin (). Mande kingdoms such as Vai and Kpelle have later developed their own innovative Mande writing systems (see Rovenchak et al. and Rovenchak and Buk, chapter of this volume). While these did not spread widely and are no longer used in current publications or in the press, the more recent N’ko writing system as developed by Soulaymane Kanté () competes with the Latin script of the official orthographies and derived systems. N’Ko was initially only used for Mandinka but seems now to be on the advance in the Mande sphere. Linguistic interest in Mande languages commenced with the exploration of the West African coast and colonial and missionary activities in the later French colonies. The earliest piece of work, a wordlist containing lexical entries from Mandinka and Soninke (Sarakole), was compiled by the Compagnie Royale du Sénégal in or around and was later published by Avesac in (Schreiber ). This earliest collection differs from a later Manding vocabulary which had erroneously been attributed to Barbot by Fodor (; see Hair : ). While single words and compiled wordlists are the only specimens until the end of the eighteenth century (Kilham ; Norris ), full grammars and textbooks were produced in the context of the British endeavors in geographic discovery and missionary work in West Africa by the first half of the nineteenth century. In Dard published his Dictionnaire français-wolof et français-bambara, and the first grammatical description of a Mande language was presented by MacBrair in . At that time also the first comparative studies were undertaken. Manding is used in Lantham’s report dating from as an explicit linguistic designation for a linguistic unit, and Steinthal () is
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
generally reported to have been the first author to emphasize the distinction between the Manding language cluster and a genealogical unit Mande. Delafosse () was the first scholar who established Mande in terms of a language family as such and likewise refined his internal classification of Mande on the basis of the word for ‘ten’, the once well-known but today obsolete tan/fu dichotomy.
. E
.................................................................................................................................. Mande languages generally share a common stock of phonetically similar cultural items denoting objects like cattle, bow, and arrow (Dwyer ). Nevertheless, most of these lexical items have a wider areal distribution and proto-forms for the items mentioned are neither unambiguously reconstructible nor distinguishable from reconstructions in other neighboring linguistic groups. Westermann and Bryan () had considered Mande an isolated family with some lexical relations to other families, a view that has its followers even today. In his publication, Greenberg represented this family as the first spin-off in his genealogical representation of Niger-Kordofanian. This hypothesis has since been reproduced in overview articles and introductory textbooks but is more controversial than the unanimous presentation suggests. The most important criterion of Greenberg on this level of time-depth seemingly was not of lexical but morphological origin: the nonexistence of a productive noun class system in Mande. The assignment of Mande to NigerCongo on typological grounds thus was questioned by some authors such as Campbell (: ). Furthermore, the loss hypothesis is problematic for placing Mande within the Niger-Congo tree model because the type of the disappeared noun class system of Mande is unidentified—whether it was based on prefixing or suffixing, head/dependent marking, cross-referencing. Hence, the innovative loss neither indicates the relative time-depth of the split nor the genealogical distance to Atlantic, Benue-Congo, or Gur. Gregersen () and later Blench () resolved this issue by including Mande as a family in an even larger genealogical unit, Kongo/Nilo-Saharan. Mukarovsky () opposed the integration of Mande into Niger-Congo and argued for genealogical ties between Mande and Chadic. However, the evidence presented by him merely consists of unsystematic lexical similarities that may likewise be interpreted in terms of ancient language contact. Strictly speaking, the position of Mande in the the Niger-Congo macro-family has not been defnitively solved but is currently undebated, however. By contrast, the view of Mande as forming a genealogical unit has never been challenged since Delafosse (). With regard to the internal classification (see Table .), the tan/fu dichotomy of Delafosse () stood uncontradicted until its rejection by both Prost () and Welmers (). While Prost did not explicitly propose an alternative classification but argued on the basis of his first-time collection of data on Eastern Mande languages, Welmers was the first scholar to present a classification based on lexicostatistics and to introduce the division into Eastern and Western Mande that has remained valid to this day. Several subsequent lexicostatistical classifications (Galtier ; Dwyer ; Grégoire and de Halleux ) generally confirmed this distinction and classified smaller languages that
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . A classification of Mande languages, major branches, and principal representatives Eastern Mande
Western Mande
Niger-Volta (Eastern Eastern) Bisa (Lebri; Barka) Samo (San; Sane) Boko-Busa (Busa; Boko; Bokobaru) Kyanga-Shanga (Kyanga; Shanga) Monzanze (extinct)
Northwestern Bobo (B.-Madaré; Konabéré) Azer (extinct) Soninke-Bozo (Soninke; Bozo (Hainyaxo; Tiemacèwè; Tiéyaxo; Jenaama)) Samogo (Jowulu; Duun-Seenku (Seenku; Duungooma ; Dzùùngoo; Bankagooma))
Mani-Bandama (Southern Eastern) Guro-Yaure (Guro; Yaure) Tura-Dan (Tura; Dan (Dan; Kla-Dan)); Mano-Mwan (Mano; Mwan) Gban Wan Beng Gbin (extinct) N.: Northern S.: Southern E.: Eastern W.: Western italics: subgroup/language cluster numerical index; see Map 11.1; distinct numbering for Eastern and Western branch
Central-Southwestern Susu-Yalunka (Susu; Yalunka) Central Jogo-Jeri (Jeri; Jogo; Jalkunan) Vai-Kono (Vai; Kono) Mokole (Mogofin; Kakabe; Kuranko; Lele) Manding (Bolon; Jahanka; Eastern Manding (Bamana; Jula (L1/Kong); Marka; Western Manding (Kagoro; W.Maninka; E.Maninka; Xasonka); Mandinka)) Southwestern Mende-Loko (Mende; Loko; Bandi; Zialo) Loma-Kpelle (Loma; Kpelle)
had not been included in previous studies. However, lexicostatistical classification in Mande faces widely known and, in addition, some more specific problems. Consonant mutation systems and phonotactic restrictions of medial consonants increase the phonological distance ratio of even closely related languages. The calculation of cognate rates is furthermore disturbed by a low rate of semantic simplicity in basic vocabulary, and a high amount of nominal and verbal compounds leads to semi-cognate forms (Grégoire and de Halleux : ). As a consequence, the rate of shared cognates appears to be low, even between languages that are closely related according to the comparative method. Therefore, Kastenholz () applied the alternative classification method of shared lexical innovations as a spot-check procedure for Western Mande in the course of his reconstruction of Central/South-Western Mande. The subsequently identified phonological innovations and morphophonological traits confirm the results of the precursory analysis of the Central and South-Western group. The discrepancies between the proposals of Kastenholz () and Vydrin (b) can be interpreted as being due to lexicostatistical vs. reconstructive classification. On a higher level, the morphological correspondences between the Northwestern Mande groups of Bobo, Samogo, and Soninke do not appear to correspond to a high degree of similarity in standard basic vocabulary and are thus questioned by
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Vydrin (b: ) in his lexicostatistical approach. Vydrin also groups Susu-Yalunka with South-Western and not with Central Mande—despite the fact that Susu-Yalunka exhibits an overall low percentage of cognates, has no consonant mutation system, and lacks certain indicative lexical innovations.
. L
.................................................................................................................................. More extensive approaches to phonological and lexical reconstruction have been proposed by Podzniakov (), Bimson (), Claire Grégoire (labial consonants; ), HenriClaude Grégoire (“proto-mande-sud”; ), Kastenholz (Central/South-Western; ), Vydrin (Mani-Bandama [Eastern Southern]; ), and Schreiber (Niger-Volta [EasternEastern]; ). However, Mande has not been reconstructed up to the proto-level, and there are only a few smaller articles dealing with some aspects of proto-morphology (Dwyer ), and in spite of the general interest in the issue of noun classes in Mande, only Vydrin () has dealt with this topic. Lexical roots in Mande languages are mostly reconstructed as disyllabic *CVCV, but Eastern Mande languages tended to reduce the stems to a CCV or CV structure. As can be seen from Table ., a controversial issue is the allophones of the proto-phonemes and the reconstruction of consonant mutation that is still productive in South-Western Mande and Soninke. While authors like Bimson (), Henri-Claude Grégoire (), and Vydrin () do not set up such a system for Northern and South-Eastern Mande, Kastenholz () and Schreiber () do reconstruct this kind of paradigmatic allophonic variation. Other problematic issues are the reconstruction of medial consonants, the phonemic status of velarized and palatalized plosives, labiovelars, and nasality. Vowel systems are considered unstable due to a high degree of internal (non-transparent) variation even in closely related languages or dialects (Kastenholz : –). They have mostly been ignored or explicitly excluded from phonological reconstruction.
Table . Comparison of reconstructions of initial consonants: Central-Southwest (Kastenholz ); Southern (Vydrin ); Niger-Volta (Schreiber ) *p/*p/*p~f *t/*t/*t *ty/*c(?)/[* ] *k/*k/*k *kp/*kp/[* ] *b~ɓ/*b/*b~w *d~ɗ/*d/*d~l *j~y/*j/[* ] *g~w~y/*g/*g~w~y [* ]/[* ]/[* ] [* ]/[* ]/*kw [* ]/*gw(??)/*gw *ɓ/*ɓ~m/[* ] *ɗ/*ɗ ~n/[* ] [* ]/*ƙ/[* ] [* ]/*ƙw/[* ] *f/*f/[* ] *s/*s/*s *s y/[* ]/[* ] [* ]/*x(?)/[* ] [* ]/[* ]/*h(?) [* ]/*z/*z *m/[* ]/*m *n/[* ]/[* ] [* ]/*y~ɲ/*y [* ]/*w~w̃/[* ]
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. L
.................................................................................................................................. Loan relations and language contact are attested for Songhay and Chadic languages (Nicolaï ). The consecutive cultural dominance and linguistic hegemony of Manding must have had a strong linguistic impact on neighboring languages, as is the case for Senufo languages (Gur; Dombrowsky-Hahn ). The structural-internal convergence of basic syntactic patterns such as word order, negation, and non-verbal predication in Mande has been proven for “Greater Manding”, a contact zone comprising some North-Western and Manding languages (Kastenholz ). Vydrin () indicates intense contact between Mani-Bandama (South Mande) and Kwa/Kru languages. Other authors recently noticed a potential “Type B-Sprachbund” (Heine and Nurse : ). Although historical sources of Arab writers do not indicate specific historical speech community events, the geopolitical position of Mande at the fringe of the Sahel, on the one hand, and historical trade with speakers of Berber or Chadic languages and Arabic, on the other, may have played a role regarding internal divisions of Mande, although such a migratory event is likely to have occurred prior to the establishment of trade contacts. From a typological perspective, the phonological systems of Mande languages generally point more to a membership of linguistic areas of West Africa than to genealogical affiliation (see Clements and Rialland ). The distribution of implosives and labiovelars/labialized plosives in Mande follows the areal pattern as identified in Greenberg (). Furthermore, like the adjacent Kru/Kwa languages, Southern Mande languages pertain to the linguistic area of languages lacking phonemic nasal consonants (see le Saout ) and showing articulatory underspecified nasal consonants (Kastenholz : –). The prototypical vowel system of a Western Mande language has seven vowels; there is no [ATR] harmony and there are no phonemic nasal vowels. Eastern Mande languages generally show a symmetric set of / oral and five phonemic nasal vowels, but also no [ATR] harmony. Exceptions are found among the South Mande languages (Vydrin ). Syllable structure is canonical CV, CCV, or CVCV, but in Eastern Mande there are strong restrictions on root-medial consonants (either nasals or liquids; le Saout ). Western Mande languages such as Manding and most of the others have a restricted tone system with monotonal or bitonal (H, H-L) patterns. Occasionally, floating tones and L-H patterns are also attested in South-Western and Northern Mande. The domain of tonal patterns is in most languages the phonological word, which may consist of several syntactic units, as in Manding, or lexical units, as in Soninke (Vydrin ). Most Eastern Mande languages have a three-tone system, extra-high and floating tones. More complex tone systems are mostly found in Southern Mande languages due to language contact. The morphological type is much different from prototypical Niger-Congo representatives such as Bantu, not only because of the often cited lack of noun classes and derivative verb morphology. In most cases grammatical words in Mande correspond with morphemes. Nevertheless, phonological words can have high word–morpheme ratio due to
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
productive word formation by affixation, fusion of subject pronoun and auxiliaries, and clitization of pronominal object markers. Mande languages are strictly speaking not polysynthetic but show a high degree of compounding and different kinds of object incorporation (cf. Creissels ). They do not have a noun class or nominal classifier system. Nominal class membership is syntactically realized for the alienable/inalienable and free/relational distinction (Welmers : –). Languages like Koranko (Kastenholz ) overtly show a tripartite number system of singular, plural, and general number, and specificity marking is found in languages of different sub-branches (Welmers : ). Most Mande languages are neither head- nor dependent-marking, as there is neither an overt nor covert marking of classifying singular/plural morphemes on nominal heads, nouns specifiers, or cross referencing. Verbal extensions, which are frequently attested for Benue-Congo languages, do not appear in Mande. Nevertheless, the argument structure of complex predicates is altered by the use of inherent postpositions. Property and quality assignment is expressed by stative verbs, light verbs, and postpositional verbs; object incorporation is found in both major branches. With regard to tense, aspect, and mood, transitive verbs are generally distinguished from intransitive verbs, and negation is expressed by either negative auxiliaries or negation particles. The pattern of perfect marking on the verb stem and imperfective marking by auxiliaries is frequent, for example in Manding (see Tröbs ). The word order patterns in Mande languages are disharmonic in terms of the status of possessive pronouns preceding the noun and the auxiliary preceding the verb but postpositions with a strict S–O–V order. The basic word order pattern is S–AUX–O–V–X without any exception, as identified by Heine (; Type B, MANDING subtype). This atypical behavior has attracted much discussion in Mande linguistics and grammaticalization theory (cf., e.g., Claudi ; Bearth ). Non-verbal predication is syntactically distinguished from verbal predication. In Manding languages like Bambara the relative clause does not follow or precede the head noun but follows the phrase, while in Eastern Mande languages like Bisa the relative phrase follows the noun and precedes the phrasefinal determiner. Serial verb constructions are not found in Mande, but Manding languages show clause chaining. Information structure is mostly expressed by focus markers; frontshifting of heavy NPs is only found in pseudo-cleft constructions.
R Avesac, A. (). Vocabulaires guiolof, mandingue, foule, saracole, séraire, bagnon et floupe, recueillis à la côte d’Afrique pour le service de l’ancienne Compagnie Royale du Sénégal, et publiés pour la première fois. Paris: Institut National des Langues et Civilisations Orientales. Bearth, T. (). ‘Nominal periphrases and the origin of the predicative marker in Mande languages—an alternative view’, Afrikanistische Arbeitspapiere : –. Bimson, K. D. (). Comparative Reconstruction of Proto-Northern-Western Mande. Ph.D. thesis. Los Angeles: University of California. Blench, R. M. (). ‘Is Niger-Congo simply a branch of Nilo-Saharan?’, in R. Nicolaï and F. Rottland (eds.), Proceedings of the Fifth Nilo-Saharan Linguistics Colloquium, Nice University, August th–th, . Cologne: Rüdiger Köppe, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Campbell, L. (). Historical Linguistics: An introduction, nd edn. Edinburgh: Edinburgh University Press. Claudi, U. (). ‘Word order change as category change: the Mande case’, in W. Pagliuca (ed.), Perspectives on Grammaticalization. Amsterdam and Philadelphia, PA: John Benjamins, –. Clements, G. N., and Rialland, A. (). ‘Africa as a phonological area’, in B. Heine and D. Nurse (eds.) (), –. Creissels, D. (). ‘L’Incorporation en mandinka’, in D. Amiot (ed.), La composition dans une perspective typologique. Arras: Artois Presses Université, –. Dard, J. (). Dictionnaire français-wolof et français-bambara, suivi du dictionnaire wolof-français. Paris and Dakar: Imprimerie Royale. Delafosse, M. (). Manuel pratique de la langue mandé ou mandingue. Paris: Ernest Leroux. Delafosse, M. (). La langue mandingue et ses dialectes (malinké, bambara, dioula). Vol. : Introduction, grammaire, lexique français-mandingue. Paris: Geuthner. Dombrowsky-Hahn, K. (). Phénomènes de contact entre les langues minyanka et bambara (sud du Mali). Cologne: Rüdiger Köppe. Dumestre, G., and Canut, C. (). Stratégies communicatives au Mali. Aix-en-Provence and Paris: Institut d’études créoles et francophones, URA du CNRS, Université de Provence. Dwyer, D. (). ‘Towards a proto Mande morphology’, Mandenkan /: –. Dwyer, D. (). ‘Mande’, in J. Bendor-Samuel with R. L. Hartell (eds.), The Niger-Congo Languages: A classification and description of Africa’s largest language family. Lanham, MD, New York, and London: University Press of America, –. Dwyer, D. (). ‘Mande languages’, in K. Brown (ed.), Encyclopedia of Language and Linguistics, nd edn. Amsterdam: Elsevier, –. Fodor, I. (). Pallas und andere afrikanische Vokabularien vor dem . Jahrhundert. Hamburg: Helmut Buske. Galtier, G. (). ‘Classification révisée des langues du groupe mandé’, in G. Dumestre (ed.), Journées d’Études, Langue et Linguistique Manding, Orcement, – juillet . Communications. Paris: CNRS and INALCO, –. Giesing, C., and Vydrin, V. (). Ta:rikh Mandinka de Bijini (Guinée-Bissau): La mémoire des Mandinka et des Sòoninkee du Kaabu. Leiden: Brill. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Greenberg, J. H. (). ‘Some areal characteristics of African languages’, in I. R. Dihoff (ed.), Current Approaches to African Linguistics. Vol. . Dordrecht and Cinnaminson, NJ: Foris, –. Gregersen, E. A. (). ‘Kongo-Saharan’, Journal of African Linguistics : –. Grégoire, C. (). ‘An attempt to reconstruct labial consonants in Mande’, in M. Dominicy and J. Dor (eds.), Phonological Reconstruction: Problems and methods. (Belgian Journal of Linguistics, .) Brussels: Éditions de l’Université de Bruxelles, –. Grégoire, C., and de Halleux, B. (). ‘Étude lexicostatistique de quarante-trois langues et dialectes mandé’, Africana Linguistica : –. Grégoire, H.-C. (). ‘Indices pour une reconstruction du proto-mandé-sud (sept langues de Côted’Ivoire)’, Verbum : –. Hair, P. E. H. (). Barbot’s West African Vocabularies of c.. Liverpool: Centre of African Studies, University of Liverpool. Heine, B. (). A Typology of African Languages, based on the order of meaningful elements. Berlin: Dietrich Reimer. Heine, B., and Nurse, D. (eds.). (). A Linguistic Geography of Africa. Cambridge and New York: Cambridge University Press. Kanté, S. (). Méthode pratique d’écriture N’ko. Kankan: Self-published. Kastenholz, R. (). Das Koranko. Ein Beitrag zur Erforschung der Nord-Mande-Sprachen. Bonn: Mundus.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Kastenholz, R. (). Sprachgeschichte im West-Mande: Methoden und Rekonstruktionen. Cologne: Rüdiger Köppe. Kastenholz, R. (). ‘On distributed predicative syntax in Mande’, in B. Caron and P. Zima (eds.), Sprachbund in the West African Sahel. Paris and Leuven: Peeters, –. Kilham, H. (). Specimens of African Languages, spoken in the colony of Sierra Leone. London: Society of Friends for Promoting African Instruction. Lantham, R. G. (). ‘On the present state and recent progress of ethnographical philology. Part : Africa’, in Report of the Meeting of the British Association for the Advancement of Science, held at Oxford in June , –. le Saout, J. (). ‘Langues sans consonnes nasales’, Annales de l’Université d’Abidjan, Série H, Linguistique : –. le Saout, J. (). ‘Remarques phonético-phonologiques sur les séquences de type CVLV’, Annales de l’Université d’Abidjan, Série H, Linguistique : –. Lewis, M. P. (). Ethnologue: Languages of the world, th edn. Dallas, TX: SIL International. MacBrair, R. M. (). A Grammar of the Mandingo Language, with vocabularies. London: Wesleyan-Methodist Missionary Society. Mukarovsky, H. G. (). Mande-Chadic Common Stock: A study of phonological and lexical evidence. Vienna: Afro-Pub. Nicolaï, R. (). Parentés linguistiques (à propos du songhay). Paris: Éditions du CNRS. Norris, E. (). Outline of a Vocabulary of a Few of the Principal Languages of Western and Central Africa: Compiled for the use of the Niger expedition. London: John W. Parker. Perrot, J. (). Les langues dans le monde ancien et moderne. Paris: Éditions du CNRS. Podzniakov, K. I. (). Mande Languages: A Historical Comparative Analysis. Ph.D. thesis. Moscow University. Prost, A. (). Les langues mandé-sud du groupe mana-busa. Dakar: IFAN. Rovenchak, A. J. M., and Riley, C. (). ‘Distribution of complexities in the Vai script’, http://arxiv. org/abs/. accessed July , . Schreiber, H. (). Eine historische Phonologie der Niger-Volta-Sprachen: Ein Beitrag zur Erforschung der Sprachgeschichte der östlichen Ost-Mandesprachen. Cologne: Rüdiger Köppe. Schreiber, H. (). ‘The Compagnie Royale Dictionnaire Manuscript’, Afrika und Übersee : –. Steinthal, H. (). Die Mande-Negersprachen psychologisch und phonetisch betrachtet. Berlin: Ferdinand Dümmler. Tröbs, H. (). Sprachtypologie, TAM-Systeme und historische Syntax im Manding (West-Mande). Cologne: Rüdiger Köppe. Vydrin, V. F. (). ‘Traces of nominal classification in the Mande languages: the Soninke evidence’, St. Petersburg Journal of African Studies : –. Vydrin, V. F. (). ‘Some hasty notes on the ways of the evolution of Mande tonal systems’, in R. Nicolaï and P. Zima (eds.), Lexical and Structural Diffusion. Nice: Publications de la Faculté des Lettres, Arts et Sciences Humaines de Nice et de la Faculté des Études Humaines, Université Charles de Prague, –. Vydrin, V. F. (). ‘South(ern) Mande reconstruction: initial consonants’, Orientalia et classica : –. Vydrin, V. F. (a). ‘Areal features in South Mande and Kru languages’, in N. Cyffer and G. Ziegelmeyer (eds.), When Languages Meet: Language contact and change in West Africa. Cologne: Rüdiger Köppe, –. Vydrin, V. F. (b). ‘On the problem of the Proto-Mande homeland’, Journal of Language Relationship : –. Vydrin, V. F., Bergman, T. G., and Benjamin, M. (). ‘Mandé language family of West Africa: Location and genetic classification’, https://web.archive.org/web//http://www.sil. org/silesr//-/silesr-.htm
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Welmers, W. E. (). ‘The Mande languages’, in W. Austin (ed.), Report of the th Annual Round Table Meeting on Linguistics and Language Studies. Washington, DC: Georgetown University Press, –. Welmers, W. E. (). African Language Structures. Berkeley, Los Angeles, CA, and London: University of California Press. Westermann, D., and Bryan, M. A. (). Languages of West Africa. London, New York, and Toronto: Oxford University Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
T Kwa languages as enumerated in Stewart () and in Williamson and Blench () include the following, moving from southwestern Ivory Coast through southern Ghana and Togo to the southern border between Nigeria and the Republic of Benin (see Map .): . Six languages spoken by small populations in the lagoon area of southeastern Ivory Coast: Avikam and Alladian, the languages Adjukru (Adioukrou), Abidji, Abbey spoken in the area of the Agneby river, and Akye (Attié). . The Potou-Tano (PT) family consisting of approximately 22 languages: Ega, the Potou group consisting of Ebrié (or Kyama/Cama) and Mbatto, and the Tano group. Tano has four branches: Krobu, spoken in the Agneby area; Western Tano or Ono, consisting of Aburé (Aboure) and Eotilé (Greenberg’s Metyibo); Central Tano; and Guang. Central Tano in turn has two subgroups: Bia, which includes Anyi, Baule, Anufo (Chakosi), Aowin, Sehwi, Nzema and Ahanta, and Akan, by far the most widely spoken Kwa language, with a number of mutually intelligible dialects, including Asante Twi, Akuapem Twi, Fante, and Bono (Brong). The Guang languages are spoken in a broad arc approximately following the Volta river, and are usually divided into north Guang, consisting of Gonja, Nchumuru, Nawuri, Gikyode and Ginyanga, Kaachi (Krakye, Krachi), Yeji, and a number of other very small and little-known varieties in the north; and south: Efutu and Awutu at the coast, Okere (Kyerepong), Lεtε (Larteh), and Anum farther inland. Classification of Nkonya as north or south is uncertain (Snider ). . The two-language group, Ga and Dangme. . The Ghana-Togo-Mountain (GTM) languages, formerly known as the Togorestsprachen or Togo Remnant (Westermann and Bryan ) or Central Togo languages (Dakubu ). Heine () classified them into two groups with names based on the lexical root for ‘meat’: NA, consisting of Buem (Lelemi and Lefana), Siwu (Akpafu) and Siwui (Lolobi), Sekpelε (Likpe), Sεlεε (Santrokofi), Ikpana (Logba), Sεderε (Adele), and Basila; and KA, consisting of the cluster Siya (Avatime),Tεgbo (Tafi), Tutrugbu (Nyagbo); Kposo, Ahlo, Tuwuli (Bowiri), Kebu, and Animere. The last-mentioned has almost disappeared.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Kwa languages
. The Gbe group is subclassified by Capo () into five major clusters: Vhe (Ewe), Gen, Aja, Fon (including Gun, Maxi), and Phla-Pherá (including Alada). They constitute a classic continuum, with sound changes shared between adjacent clusters, but not between those at the eastern and western extremes, Ewe and Phla-Pherá (Capo : ). These groupings are mainly uncontested, but there is dispute concerning the internal coherence of GTM. Following Heine (), both Stewart () and Williamson and Blench () divide them into the NA and KA groups, but neither treats these groups as forming a genetic unit. We shall return to this problem. On the basis of the lexicostatistical reclassification of Bennett and Sterk (), Stewart but not Williamson and Blench divided all of Kwa into two groups, “Nyo” and “Left Bank”. “Left Bank” included Gbe and the GTM-KA languages. “Nyo” included the rest, on the basis of the claim that a root -ɲɔ ‘two’ is a lexical innovation common to those languages. Normally, it should not be necessary to discuss the validity of a genetic grouping. As Newman (: ) argued, “the Comparative Method is a technique for reconstructing aspects of a proto-language by the systematic comparison of languages already understood to be related” (emphasis original). However, the question arises because the two scholars who have seriously examined the Kwa group since Westermann (), Jack Berry (),
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and especially John M. Stewart, called it into question.1 Stewart chose to concentrate on the relationship between what I will continue to call Proto-Potou-Tano (and not, as he later preferred, Potou-Akanic), i.e. the group consisting of the Tano languages (Stewart : ) plus the two Potou languages, which is indisputably an established group, and Proto-Bantu, finding more systematic correspondences between those two families than between Proto-Potou-Tano and the other Kwa languages and groups.2 On the basis of comparison with Guthrie’s reconstruction of Common Bantu (Guthrie –), he reconstructed a series of implosive stops3 in Proto-Bantu-Potou-Tano that accounted for correspondences with the implosive series he reconstructed for Proto-Potou-Tano.4 The Tano languages do not have these sounds, but one of the Potou languages (Cama/Ebrié) does, in items that regularly correspond both with Tano and with Bantu items. Recognition that this was an old feature inherited from a distant ancestral proto-form thus enabled reconstruction of at least part of the phonetic history of the Potou-Tano languages and of their relationship to the Bantu languages and therefore, in theory, to the rest of Benue-Congo. Stewart () finally claimed that this work was a pilot reconstruction of Proto-Niger-Congo. The alternative to Kwa, favored at different times by both Stewart and Williamson, is a flat “Benue-Kwa” tree, uniting current Kwa and Benue-Congo. The supposed similarities between Kwa languages on the one hand and Benue-Congo languages on the other are then the result of area features spreading through contact (Stewart a). Stewart suggested more than once that apparent similarities between the languages were due to all other languages in the area having been influenced by Tano languages over a long period. However, it is very likely that the current distribution on the map, giving the impression that the Tano languages, Ga-Dangme (GD), and Gbe all converge on the lower Volta valley, is relatively recent. Most of the language groups apparently arrived in their current positions within the past , years or less. It is accepted by both historians and linguists that Akan, the largest Tano language, spread eastward to come into contact with Gbe and Ga-Dangme within the past few hundred years, while Ewe has spread westward. Further, although some Tano-Guang languages probably had early contact with western Gbe (Ewe), Ga-Dangme, and the GTM languages before the eastward spread of Akan, there is no reason to think they were ever particularly “dominant”. In an unpublished paper, Stewart (a) classified the Ivory Coast “Lagoon” languages Avikam and Alladian as belonging to Tano, and Ega to Potou-Tano. However, he excluded Adjukru, Abidji, Abbey, and Akye from this group, although he did not deny genetic relationship at some higher level. He also proposed that on the basis of Bennett and Sterk’s () lexicostatistics, Ga (and by implication Dangme) was excluded not just from Tano and Potou-Tano but from Kwa, and thus from his expanded version of Benue-Congo that includes Kwa (and in effect is equivalent to Volta-Congo minus Gur and Adamawa; see Stewart ).
1 Mukarovsky (: ) accepted the “Western Kwa” of earlier authors, including Kru but excluding GTM, but explicitly did not consider its validity. 2 In his late work Stewart referred to Tano as Akanic, hence Potou-Akanic, e.g. Stewart (). 3 He originally reconstructed them as lenis, but later decided they were implosive. 4 This work was presented in a long series of papers, not all of them published, including Stewart (, , a, b, ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The division into Nyo and Left Bank, rejection of the genetic unity of GTM, and (tentative) rejection of any common ancestor for Ga-Dangme and Benue-Congo or Kwa later than Proto-Niger-Congo were all based on lexicostatistics, specifically Bennett and Sterk’s () lexicostatistics. The position taken here is that lexicostatistics is useful as a preliminary assessment of relations, even of the approximate time-depth of relations, but cannot establish degrees of genetic relationship and in particular cannot compete with the comparative method. None of the three proposals is based on application of the comparative method, and they are therefore discounted here. One means by which scholars have tried to justify (or dismiss) claims for the genetic unity of a group is by finding lexical items unique to the group, which they propose as constituting “lexical innovations”. The concept has not been precisely defined, but, in any case, lexical innovation is certainly a weaker criterion than conditioned sound changes, which, unlike lexical innovations (Williamson ), must occur across a whole group and can therefore be used as criteria for exclusion. Bennett and Sterk () proposed four items, ‘three’, ‘breast’, ‘firewood’, and ‘neck’, as distinguishing Kwa from Benue-Congo, which therefore constituted common lexical innovations in Kwa and evidence for its genetic validity. They were discussed by Williamson () in her Benue-Congo overview for The Niger-Congo Languages, and for good reasons she discarded the first two. She tentatively maintained ‘firewood’, with very different roots in the two groups, although she doubted whether either the Kwa (*je/*nye) or the Benue-Congo (*koni) was in fact an innovation, and proposed the form *toŋ ‘neck’ as an innovation in Benue-Congo (: ).5 She then proposed nine lexical items, including the two just mentioned, as possible innovations in Benue-Congo. The others are ‘lick’, ‘one’, ‘count’, ‘man/male/husband/marry’, ‘back’, ‘dance’, and ‘roast/cook/burn’. She also considered words that Bennett and Sterk proposed as defining subdivisions of Benue-Congo, ‘ten’ and ‘woman’. If none of the proposed innovations exists in any branch of Kwa, they perhaps are indeed exclusive to Benue-Congo, but if one of them exists in one or more branches of Kwa, then obviously it does not support the separation of these two sets of language groups.6 In fact, however, there are very likely Kwa cognates for all but three of the proposed Benue-Congo innovations: ‘firewood’, ‘neck’, ‘dance’, and possibly also ‘count’. The item ‘ten’ (wo) was probably borrowed by Gbe languages from Yoruba. Moreover, of the roots considered by Williamson to be innovations into branches of Benue-Congo following the break-up of that proto-language, four have possible cognates in Kwa, namely Nupoid *-gberε ‘throat’: compare Gbe gbe, Ga-Dangme *gbɩdῖ ‘voice’; Proto-Bantu *-mìdò ‘throat’: compare Ga-Dangme *mĩ ĩ́ ‘throat’; Proto-Yoruba *ɔ-kɔ ‘husband’ (and similar words in other branches): compare Tano-Akan -kunu, Ga-Dangme *hu, GTM-Logba ń-gù, and probably Ewe ŋu-tsu ‘man’ (first syllable). Proto-Upper Cross *-túB, túmà ‘dance’ 5
For detailed examination of these items, see Dakubu (b). In ensuing discussions I use Capo’s Proto-Gbe, but where Capo does not reconstruct an item, Westermann () is often used—Westermann, in fact, used “Ewe” to cover all of Gbe. Proto-GaDangme is my own reconstruction; see, e.g., Dakubu (). Reconstructions for GTM are Heine’s (), but if no item is reconstructed, likely reflexes are proposed from several sources, including Kropp (), Dorvlo () (for Logba), and data provided by participants at the “Workshop on GTM Languages” held in , for which I am grateful. 6
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
resembles forms in at least two GTM-NA languages, Santrokofi tũ, Lelemi tɔ̃. The list of supposedly exclusively Benue-Congo items is thus considerably reduced, and does not advance the problem of the unity of Kwa or its distinction from Benue-Congo one way or the other. An alternative and potentially more fruitful strategy is to build on Stewart’s reconstructions, especially his rules for deriving Proto-Potou-Tano from Proto-Bantu-Potou-Tano (PBT), as stated in Stewart (). Unfortunately, several of the rules are difficult, if not impossible to apply, for lack of comparable items.7 However, at least one of these sound changes appears to be common to all branches of Kwa, although the evidence from GTM is admittedly weak. Stewart’s (: ) Rule : [V –hi –lo +nas] > [V –hi –lo –nas] . The rule states that in a first syllable, a nasalized consonant before a nasal vowel followed by a nasalized lateral and also the nasal vowel both lost nasalization before a second syllable beginning in n. However, as stated, this subsumes and conflates more than one rule: one in which a prenasalized stop becomes a sonorant after a nasal vowel, another in which the preceding vowel then becomes non-nasal, and a third by which an initial nasalized sonorant also loses nasality. Stewart does not reconstruct several of the items cited as attesting Rule with an old initial nasal consonant, only a vowel in the first syllable which lost nasality before the (earlier) nasalized second syllable. The reflexes in other branches of Kwa have generally lost the second syllables, so that they mainly attest the loss of nasality in the initial syllable.8 This common innovation allows us to posit a Proto-Kwa language. Gloss
PBT >
PT
GTM
GD
Gbe (Ewe)
‘neck’ ‘get wet’ ‘look after a child’ ‘moon’ ‘white’11
*-ƙɔ̃ntɩ *-ɓɔ̃mbɩ *-ῖε̃ῖῖ *-ɰ̃ ε̃ῖῖ *-ɰ̃ ε̃ῖῖ
*-ƙɔnῖ *-bɔʋɩ *-lεnῖ *-ɰεnῖ *-ɰεnῖ
Likpe ɔ́-kɔ *-fólu – –
*kɔn-9 *pɔ̀ *lε– *hiέ- (v) *hiέn- (adj)
kɔ̀ fo (tsıˇ) ɣletí10 *wé (ɣé)
Another of Stewart’s rules states that two Proto-Bantu-Potou-Tano consonants merged in Proto-Potou-Tano. They merged not only in Potou-Tano but also in GTM and Gbe. Stewart’s (: ) Rule : c > t/C.
7
Dakubu (a) is a more exhaustive discussion of the application of these rules to the rest of Kwa. I omit some additional items in Stewart () that seem to have no reflexes in Gbe, GD, or GTM. 9 ‘shoulder’. 10 Capo does not reconstruct this word for Proto-Gbe, but reconstructs ɣ as *w (see ‘white’). 11 Stewart () does not indicate that this item is an example of the rule, but since it seems to be entirely homophonous with ‘moon’, and given the number of mistakes in the publication, I include it here. 8
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
That is, the Proto-Bantu-Potou-Tano voiceless palatal affricate *c became Proto-PotouTano non-affricate *t in root-initial position. Since a voiceless alveolar stop already existed, these two consonants merged in Potou-Tano, and it turns out also in Proto-GTM and Proto-Gbe.12 However, they did not merge in Proto-Ga-Dangme. We conclude, therefore, that Ga-Dangme was the first split from Proto-Kwa. Gloss
PBT *c > PT *t GTM *t GD *s
‘underneath’ *-cɩ ‘ground’ *-cɩ ‘father’ *-cɩ
*-tɩ *-tɩ *-tɩ
*ka-tí – *-tɔ
Gbe *t
*sĩ ̄té *sĩ-, *zĩ to-me *tsε *-tɔ
Proto-Bantu-Potou-Tano *t on the other hand remained unchanged throughout. Gloss
PBT *t > PT *t GTM *t GD *t Gbe *t
‘ear, edge’ *-tʊ̃ ‘three’ *-tãntɩ
*-tʊ̃ *-tã
*ku-túe *-itá
*tó*étε̃
*-tó etɔ̃
There is as yet insufficient evidence to propose any further subgrouping (but see Dakubu ). There are, however, a considerable number of correspondence series between pairs of Kwa languages, although Stewart provided no Potou-Tano or even Tano reconstructions for many of the items. We conclude with a few examples. PBT *ɓ > PT *ɓ > Tano *b
Gloss
‘be cooked’ *ɓῖlῖ̃ ‘ask’ ‘calculate’
*-ɓɩnῖ *ɓVCV
GTM *b GD *b
*bɩ Akan bisa *bíé Akan bu *bu
Gbe *b
*bè *bí *bya ?*biya *bí *bù *bu
Heine () does not reconstruct ‘be cooked’, but on the basis of available evidence it appears to be present and reconstructible at least for the NA branch. The same is true of ‘calculate’. A [ ATR] vowel for ‘be cooked’ is proposed here. *ɓ is proposed as underlying both Potou-Tano and Proto-Gbe ‘ask’ in Stewart (n.d.). Gloss
PBT *m > Potou-Tano *m > Tano *m GTM *m
‘swallow’ *mῖl ῖ̃ ‘inside’ ‘fresh’
12
*-mῖlῖ
GD *m
Gbe *m
Logba mε *mĩ ̀*bĩ [*mĩ] ́ Akan mῦ *-miá *mĩ ̄ĩ *-me *mῦdɩ-ε *-muɔdῖ *mṹ (adj)
Potou-Tano *t later shifted to *s in Tano, so that Akan has s in all these forms.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Bendor-Samuel, J. (ed.) (). The Niger-Congo Languages: A classification and description of Africa’s largest language family. Lanham, MD, New York, and London: University Press of America. Bennett, P. R., and Sterk, J. (). ‘South-central Niger-Congo: a reclassification’, Studies in African Linguistics : –. Berry, J. (). Structural Affinities of the Volta River Languages and Their Significance for Linguistic Classification. Ph.D. thesis. University of London. Capo, H. (). A Comparative Phonology of Gbe. Berlin and New York: Foris. Dakubu, M. E. Kropp (ed.) (). The Languages of Ghana. London: Kegan Paul. Dakubu, M. E. Kropp. (). ‘Earliest Ga-Dangme culture, from a linguistic point of view’, Research Review (Supplement No. ): Accra before Colonial Times. Proceedings of a Colloquium on Early Accra, –. Dakubu, M. E. Kropp. (). Sub-classifying the languages of the lower Volta valley: a way forward? Paper presented to the International Workshop on GTM Languages, Ho, August –. Dakubu, M. E. Kropp. (a). The relationship of GTM, Gbe and Ga-Dangme to Potou-Tano: the Bantu-Potou-Tano correspondences. Paper presented to the International Workshop on GTM Languages, Ho, August –. Dakubu, M. E. Kropp. (b). Kwa and Benue-Congo: lexical relations. Paper distributed at the International Workshop on GTM Languages, Ho, August –. Dorvlo, K. (). A Grammar of Logba (Ikpana). Utrecht: LOT. Guthrie, M. (–). Comparative Bantu: An introduction to the comparative linguistics and prehistory of the Bantu languages. vols. Farnborough: Gregg. Heine, B. (). Die Verbreitung und Gliederung der Togorestsprachen. Berlin: Dietrich Reimer. Kropp, M. E. (). Lefana, Akpafu and Avatime with English Gloss. (Comparative African Wordlists no. .) Legon: Institute of African Studies. Mukarovsky, H. (). A Study of Western Nigritic. vols. Vienna: Institut für Ägyptologie und Afrikanistik der Universität Wien. Newman, P. (). On Being Right: Greenberg’s African linguistic classification and the methodological principles which underlie it. Bloomington; IN: West African Languages Institute, Indiana University. Snider, K. (). ‘The noun class system of Proto-Guang and its implications for internal classification’, Journal of African Languages and Linguistics : –. Stewart, J. M. (). ‘The lenis stops of the Potou Lagoon languages and their significance for preBantu reconstruction’, Papers in Ghanaian Linguistics (Supplement no. ): –. Stewart, J. M. (). Towards Volta-Congo Reconstruction. Leiden: Leiden University Press. Stewart, J. M. (). ‘Kwa’, in J. Bendor-Samuel (ed.), –. Stewart, J. M. (). ‘The second Tano consonant shift and its likeness to Grimm’s Law’, The Journal of West African Languages : –. Stewart, J. M. (a). South Volta-Congo (Benue-Kwa) subclassification: the position of Tano (Akanoid). Paper presented at the nd Annual Conference on African Linguistics, Berkeley, CA. Stewart, J. M. (b). The stem-initial consonant system of proto-Potou-Tano-Bantu: an update. Paper presented at the nd Annual Conference on African Linguistics, Benue-Congo Workshop, Berkeley, CA. Stewart, J. M. (). ‘The potential of Proto-Potou-Akanic-Bantu as a pilot Proto-Niger-Congo, and the reconstructions updated’, Journal of African Languages and Linguistics : –. Stewart, J. M. (n.d.) Implosives, homorganic nasals and nasalized vowels in Volta-Congo. Cyclostyled paper. Westermann, D. (). Die westlichen Sudansprachen und ihre Beziehungen zum Bantu. Berlin: Walter de Gruyter. Westermann, D. (). Wörterbuch der Ewe-Sprache. Berlin: Akademie-Verlag. Westermann, D., and Bryan, M. A. (). Handbook of African Languages. Part II: Languages of West Africa. London: Oxford University Press. Williamson, K. (). ‘Niger-Congo overview’, in J. Bendor-Samuel (ed.), –. Williamson, K., and Blench, R. (). ‘Niger-Congo’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T Gur or Voltaic languages are spoken in central western Africa south of the Sahara, mainly in Burkina Faso and the adjacent regions of Mali, Ivory Coast, Ghana, Togo, and Benin, as well as in northwestern Nigeria (see Map .). The name “Gur”, coined by G. A. Krause in a letter to J. G. Christaller (reported in /: ), because of the many language names containing the syllable -gur- is still in use by Anglo- and Germanophones, while the term “Voltaic”, preferred today by Francophones, was introduced by M. Delafosse (/: ). Bodomo (a) proposed the term “Mabia” as a new cover term for the whole family, which he (a) had originally introduced only in order to replace the name “Western Oti Volta” (one of the subgroups of the Northern branch of “Central Gur”).1
. L
.................................................................................................................................. To this day, Gur languages are among the less well-documented West African languages. Early vocabularies appeared in the late nineteenth century. And when Koelle interviewed slaves who had returned from America in Freetown and recorded their languages, these included speakers of Gur languages (see Köhler ). Apart from grammatical sketches and vocabularies provided by early scholars like Christaller, Delafosse, and Westermann, it was only after the World War II that more information on Gur languages became available. As is known from the historiography of other African languages, the main impetus came 1 During the s, following the general trend of that time, G. Manessy coined the cover terms within Gur classification by changing the names of the higher units from language-oriented names to geographical river-oriented names. Among others, he replaced “Moore-Dagbane” by “Oti-Volta Occidental”. Bodomo (a: , note ) explains the etymology of the term “Mabia” as a compound formed from ma ‘mother’ + bia ‘child’ in order to denote “a sibling relationship” of the respective languages. Unfortunately, languages outside of “Western Oti Volta” do not have the same word for ‘mother’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Gur languages
from missionaries.2 Among them, one of the most effective researchers on Gur was A. Prost, who analyzed and provided data he had collected himself or which were delivered to him by other missionaries. He developed a special questionnaire, the output of which was organized on the same principle as all his other numerous publications. Another important field researcher, who paved the way for further linguistic approaches in Gur, was W. Welmers. To this day, his meticulous studies of Senufo (, ) and Bariba (Baatɔnum; ) are indispensable prerequisites for anyone starting research on these languages. On the academic side in Europe, two scholars in particular should be mentioned: O. Köhler and G. Manessy. The former provided a comprehensive overview of early research on Gur languages (, /) that also contains lexical reconstructions as well as a classification of the languages based on lexical and morphological comparison. Manessy, likewise applying the philological method, started in the s with the study of Bwamu and its dialects and went step by step through all important subfamilies, ending with the Senufo area to which he devoted his last publications (b, a, b, c, d). 2 During colonial times and thereafter, mostly initiated by the French Catholic missions, later by Protestant missions, in particular by the Wycliffe Bible Translators (known under the name “SIL” (Société Internationale de Linguistique/Summer Institute of Linguistics).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Apart from his overviews of whole language groups or families, he provided comprehensive comparative studies on special morphological subjects (noun classes, adjectives, verb typologies, relatives, negation, etc.). It is through his meticulous comparisons and reconstructions that the foundation was laid for all further linguistic research on Gur.3 Independently of each other, both scholars, by connecting their linguistic results with oral traditions and the knowledge of historians, designed scenarios for the migration streams that created today’s fragmented linguistic situation.4 Parallel and subsequent to Manessy’s philological approaches, a new impetus and new methodology came in with the linguistic departments of the newly founded African universities, as well as with the linguistic activities of the Bible translators, which is well documented in the overviews written by Bendor-Samuel () and Naden ().5 Further overviews and comparative studies are Bodomo (a) on serial constructions; Reineke (), Hiraiwa (), and Dombrowsky-Hahn () on relative clauses; Miehe () on personal pronouns; Miehe (d and ) on numerals; Delplanque () on typological features of languages spoken in Burkina Faso; Fiedler, Reineke, and Schwarz (), Fiedler and Schwarz (, ), and Schwarz () on focus;6 Nicole (), Miehe and Winkelmann (), and Miehe, Reineke, and Winkelmann ()7 on noun class systems; Winkelmann and Miehe () on negation, and Winkelmann () on tense and aspect. As far as historical reconstructions are concerned, we note a parallel with Bantu and other Benue-Congo languages in so far as—apart from the lexicon—only reconstructions of noun class markers and some derivative morphemes are available. This evidence points to the creativity and dynamics of human speech that evidently innovates the linguistic means used to express predication (in particular, the range of markers designating mode, aspect, or tense) more often than lexemes and their inflectional morphemes. Reconstructions in Gur were offered by Köhler (, /), Manessy (in almost all his overviews), Swadesh et al. (), Naden,8 and recently by Sambiéni (). Concerning other descriptions of Gur languages, the state of the art is unbalanced in so far as phonological and tonological studies prevail. Extensive grammatical descriptions are relatively rare but have increased in number during the past decade.9 Dialect surveys stem
3
For a summary of his work, see Miehe (b). See Köhler () and Manessy (, , ); a similar attempt has been made by Naden for the Ghanian Gur languages (a) or by Bodomo (a) for the “Mabia linguistic group”. 5 See also the studies published in volume of the Gur Papers () by Carlson on Senufo, Kleinewillinghöfer on Gurunsi, and Winkelmann on the Southwestern Gur languages. 6 As a result of research within the Collaborative Research Program (SFB ) of the Universities of Potsdam and Berlin (Humboldt). 7 As a result of joint research projects carried out by the Gur working groups in Bayreuth and Berlin: volume () on Southwestern Gur languages, and volume () on North Central Gur languages. A third volume on “Gurunsi” (ed. by U. Kleinewillinghöfer) is in preparation. 8 In his numerous unpublished “Lexinotes”. 9 See, for instance, Kropp-Dakubu () on Dagaare; Kra () on Kulango; Beyer () on Pana (Gurunsi); Schwarz () on Buli; Cahill () on Konni; Niggli () on Ninkare (Western OtiVolta); Micheli () on Kulango; Elders () on Kulango; Kropp-Dakubu () on Farifare (Gurene); Reinhard () on Moba; Pali () on Miyobe (Togo/Benin); Delplanque () on Dagara-Lobr; Niggli () on Kusaal (Burkina Faso); Bakpa () on Ngem (Gurma); DombrowskyHahn () on Syer (Senufo); Traoré () on Senar (Senufo); and Palm () on Ja̰a̰ne. 4
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
from Nicole on Nawdm () and Malgoubri on Moore (), recently on Nuni (). For sociolinguistic studies, see Showalter on Dogose-Ka̰a̰sa () and part II in Miehe, Owens, and Roncador (). Contact phenomena between Gur languages and Gur languages in contact with neighboring languages of other language families have been considered, for instance, by Soma (), Dombrowsky-Hahn (b), Kleinewillinghöfer (, , to appear b), and Beyer ().
. C
.................................................................................................................................. The latest external classification (Williamson and Blench ) sees Gur together with Adamawa as a member of North Volta-Congo.10 The history of internal classification displays as many attempts as different methods to capture the diversity of the languages or dialects in question. Early classifications included Dogon, which today is considered a linguistic unit on its own, as well as Seme, which seemingly belongs to Kru. Koromfe, classified by Westermann and Köhler as belonging to Gurunsi, was placed by Manessy () among Northern Central Gur languages, as a single sub-branch because of its “in-between” features, which show affinities with Gurunsi on the one hand and Moore on the other. A similar case is Mɔyɔbε which has tentatively been associated with Gurma by Manessy (: ). However, it shows in fact resemblances in its noun class prefixes to Oti-Volta noun classes (see Rongier ) but is quite distant on the lexical level. Naden (: , note), therefore, had doubts about its relationship to Gur. Today, we note general consensus concerning the internal subgrouping of Central Gur with its major branches “North” and “South”. In addition, there are a few other languages spoken to the (south)west that show many more peculiarities so that they are placed outside Central Gur. Nevertheless, there is no doubt about their “Gurness”. This is particularly true of Senufo languages and Kulango, whose affiliation was doubted by Naden (: ). Even if the former differ in word order, there exist strong morphological traits that link the Senufo languages more closely to Gur than to Guang or to Ghana-TogoMountain languages (as Naden argues).11 Likewise, Kulango and in particular its close relatives to the north (Teen and Loma) show common traits with the Kaansa Dogose group, especially within the verbal system (see the appendix in Elders ). Apart from these two language groups, which can be classified as coordinates to the Central Gur block, we note four further languages/language group(s) (Cεfɔ, Viemo, Samu group, and Tusian) which are designated as “isolated” units but yet are closer to Gur than to any other language in the region. For the following, see the version of Miehe et al. (: –):12
10
Apart from the different labels in the table and the text that follows, the somewhat crude arrangement of several Adamawa groups as coordinates of Central Gur languages (: ) lacks linguistic comment. Moreover, this classification fails to mention Kleinewillinghöfer (), who presented the first serious linguistic arguments for regrouping some (but not all) of the Adamawa languages together with Central Gur. 11 For details, see Miehe (a). 12 Which in turn is based on Manessy (, ) and Naden () but modified and amended by the author’s own research results. The names of the individual languages (shown in bold) are written so as to reflect their phonological form as closely as possible. The names of the language groups are written in a somewhat standardized form.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
C G
North Central Gur Oti-Volta:
Bwamu: Koromfe South Central Gur Gurunsi:14
Buli-Konni: Buli, Kɔnni Oti-Volta East:13 Byali, Ditammari, Nateni, Waama, Mbεlimε Oti-Volta West: Northwest: Dagara cluster, Gurenε, Moore, Nootre, Safalaba, Kantoonsi Southwest: Dagbani, Hanga, Kusaal, Mampruli, Nabit, Talni Gurma: Akasele, Ngbem (Gangam), Dye (Gangam), Gulmancema, Konkomba, Moba, Ncam (Bassari) Yom-Nawdm: Nawdm, Yom Bwamu, Boomu Lama Kabiye: Kabɪyε, Kufalo, Masεε (Yaka, Sarakawa), Ləkpa Tem South-East: Bago, Kusuntu, Cala (Bogoŋ), Dulo (Dilo, Delo) Gurunsi North: Kasim: Kasɪm, Furi (Western KasIm) Nυnɪ (North) Nυnɪ (South) Lyele: Lyele, Nebule (Batoondo) Pana Samoma (Kalemse) Gurunsi West: Winyε Phwie (Puien) Sisaala cluster: Buuni (Northern Sisaala) Passale (Southern Sisaala) Tumulung (Sisaala) Busillu (Western Sisaala) Tampulma Cakali Dεg Vagla Siti (Kyitu) Cerma, Cuuramma Lobiri, Ja̰a̰ne Ka̰a̰sa, Gbatɔgsɔ, Dɔgɔsε, Khisa Sυ-Khe, Dɔgɔsυ
Gurunsi East:
Cerma-Curama: Lobiri-Ja̰a̰ne: Kaansa-Dogose: Khe-Dogosu: Baatɔnum C G (“ C G”) Kulango group: North-Kulango, South-Kulango, Loron (Teen, Loma) Senufo: Northern Senufo: Minyanka, Nanεrghε, Supyire (Supyire, Sucite, Shεnpire), Senar 13 14
Also known as Atakora languages (named after the mountains where the languages are spoken). Classification according to Kleinewillinghöfer (p.c.).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Karaboro: Kar, Syer (Tenyer) Central Senufo: Shenara, Papara, Tyebari, Cebaara, Nafara, Dugubεsyε̰εri Southern Senufo: Palaka (Pilara), Nyarafolo, Fodonon, Tagbana, Tafire, Jimini, Nafaanra
Cεfɔ Viemo Samu group: Samu (Wara), Samwe (Natioro), Palεn Tusian: Win (South Tusian), Tir (North Tusian) Moyɔbε
. S
..................................................................................................................................
.. Phonemes, tones, syllables All phoneme inventories of Gur languages contain the consonant phonemes of the common five points of articulation, including labio-velars in many but not all languages (not in Moore, for instance). In addition, there are some regional particularities: retroflex articulation is allophonic to alveolar stops (e.g. in Gurunsi East), fricative /h/ appears as an allophone of /f/, /s/, or /k/ (e.g. in Oti-Volta). Moreover, /h/ may trigger nasalization of the following vowel (e.g. in Oti-Volta East). Glottal stops and pre-glottalized obstruents require special attention: though Manessy (: ) reconstructed Proto-Central Gur glottalic obstruents / ɓ ɗ ɗy ˀy ˀ/ in addition to the voiced and unvoiced obstruents series,15 the occurrence of glottalized stops in modern Gur appears to be a secondary development. While glottal stops have either replaced velars or appear as their allophones (Cεfɔ, Tusian, Supyire (marginal), Kaansa-Dogose),16 the full series is attested only in two related languages: Lobiri (ˀb~ɓ ˀl ˀy ˀw) and Ja̰an ̰ e (ˀb~ɓ ˀl ˀn (ˀLV̰)), a special case being Dagbane, where glottal stops are restricted to some pre-pausal occurrences. Hyman () proposes treating the glottal stop in this language as a prosodic feature. Dombrowsky-Hahn () reports that lateral fricatives resulting from consonant clusters are observed in extreme southwestern Burkina Faso in two neighboring but nonrelated languages, in Syer (Senufo) and in traces in Cuuramma (South-Central branch). Voiced/voiceless alternation of initial plosives is found everywhere in the Gur area. This alternation concerns mainly bilabial but also alveolar and velar obstruents and occurs language-internally as well as externally, the most prominent examples being the anaphoric pronoun of class *ba that is realized as pV in phrase-initial position in Senufo, Tusian, Kaansa-Dogose, and in KabIyε.17 Even if most cases concern an alternation from voiced to 15 Likewise, he reconstructed only one voiced obstruent series /ɓ ɗ ˀy/ for Proto-Senufo (Manessy c: ). 16 In Ka̰a̰sa, it even appears as an allophone of /r/ and /s/. 17 Köhler (: –) discusses the phenomenon in detail, calling it “spontaneous sound change”, and argues that it might go back to “Proto-Sudanic”. He assumes that the modern reflexes might be “traces of old variable consonants” (: ). Like Westermann, he points to a similar phenomenon in Ghana-Togo-Mountain languages.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
voiceless plosives, examples of changes the other way round are also known.18 The phenomenon needs further comparative analysis. With regard to the vowel systems, there is a remarkable contiguous occurrence of tenor nine-vowel systems19 in the southwestern region (with Senufo forming the western border), which fall to mainly seven-vowel systems in the central and eastern regions. Accordingly, vowel harmony rules (based on the [ATR] traits) are more frequent in the western region. Nasal vowels occur in all languages (Miehe ). As far as tone systems are concerned, the only characteristic underlying all Gur languages is the terraced tone principle, with more or less developed contour tones. Otherwise, we note very divergent systems, ranging from four-tone levels in Senufo via two-tone and downstep systems in the central area (Oti-Volta West, some Gurunsi languages), to three-tone levels, for instance, in Lobi or Oti-Volta East. In between, tendencies towards more pitch- or stress-oriented languages (e.g. Ka̰a̰sa) or to more grammatically oriented systems20 are nevertheless documented. There are two languages (Koromfe and Kulango) about which some authors have doubts concerning the phonemic status of the tones because of the non-existence of lexical contrasts. Tone depressor consonants (b, v, d, j, g, gb) are reported from Ja̰a̰ne (Palm ). As documented by Elders () for Kulango, grammatical paradigms as well as clause units are often characterized by typical tone sequences. Syllable structures are commonly CV or CVC(V). Often triggered by reduction processes, liquids or nasals may form the coda of former CVCV sequences.
.. Word order Like almost all West African Niger-Congo languages, Gur has the modifier modified order in associative constructions, while the evidence is not so homogeneous on the syntactic level. With the exception of Senufo, Viemo, Win, and the Samu group, all spoken in the Western region, the other Gur languages generally have S P O order, which can, however, be altered to S O P when used in certain verb paradigms (like negation or aspect; see Delplanque ).
.. Noun phrase Within noun phrases, qualifiers and determiners follow the head nouns, which in turn are generally deprived of their class markers. Thus, noun phrases generally follow the principle of “last elements only bear the burden of morphological marking”, which is widespread in western Niger-Congo languages. Nouns are defined by class markers that are suffixed to the stem. However, bilateral marking occurs regionally, mainly in the eastern area (Gurma, Oti-Volta East, Gurunsi East), and also in the southeast (Sʋ-Khe). In all known cases, the prefixes (which are more or less grammaticalized as inflectional morphemes, i.e. they occur in citation form) are derived from the respective pronominal class markers. Whereas in some cases these prefixes function as definite markers, in some other languages special
Westermann (: ) even assumed that the change took place from voiceless to voiced consonants. In spite of the fact that Manessy (: ) has reconstructed systems of only seven proto-vowels. 20 For instance Ja̰a̰ne, where verbs are not marked for tone in the lexicon; likewise, grammatical tone plays an important role in Cerma. 18 19
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
definite markers have been developed either equally from the class pronouns or from special particles (former focus markers?). Deverbal derivation processes considerably enlarge the nominal inventories of both substantives and qualifiers. There are languages like Nateni (Oti-Volta East) which recruit almost all attributive qualifiers from verbal stems that are inserted between noun stem and class suffix without any derivative morpheme (see Nédellec ). Apart from this extreme formation process, almost all other Gur languages derive deverbal adjectives or participles by morphological means. “Primary” (i.e. non-derived) adjectives are generally few in number. While in many languages, qualifiers take on the suffixes of the head nouns, there are others (in particular in Oti-Volta West) where the adjectives bear fixed class markers (recruited from the existing class system) which agree with the head noun in number only. In contrast, Cerma (besides adjectives in concordance with head nouns) has qualifiers the suffixes of which do not stem from the common class markers (for comparative details, see Manessy ). Generally, cardinal numbers deviate from other constituents within the noun phrase in that their (pronominal) concord markers precede the numerals, i.e. the agreement marker is not suffixed, as is the case with qualifiers (Miehe d). As for the marker position of determiners, suffixing prevails; however, prefixing may also occur. Further on, it depends on the function of the determiners whether they receive the pronominal or the nominal class markers. There is no language that does not show any traces of a former (supposedly) welldeveloped noun class system. The Gur class systems, which in general show a close relationship to the noun class systems of other Niger-Congo languages, contain different singular–plural pairings as well as single class genders with semantically more or less welldefined class content. As emphasized several times by Manessy in his comparative studies, number-based class-pairing is not the decisive characteristic of Gur, but a more semantically complex system of classes, the contents of which, according to Manessy (: ; : ), may express “non numéralité, unicité, pluralité, collectivité, totalité ou globalité d’un ensemble, particulièrement grand et particulièrement petit”. The most extensive use of nominal class suffixes to express different stages of qualification has been described by Nédellec () for Nateni (Eastern Oti-Volta). There is no other known Gur language that has developed such a detailed semantic specification of its nominal class markers. In addition, we observe triple genders, mostly differentiating between SG–PL–non-countable PL, in several Gur languages belonging to different subgroups. Moreover, in one or the other language, almost all classes may function as single class genders. The degree of syntactical agreement varies from elaborated systems with full concord systems to less differentiated ones where the use of former anaphoric pronouns is reduced to the distinction between animates and non-animates.21 Interestingly, the different reduction processes are observed as crossing any classification boundary. The chart below is copied from the overview “Tentative reconstructions of noun classes in Gur”, as discussed in Miehe, Kleinewillinghöfer, Roncador, and Winkelmann (: –):22
21 Moore, for instance, distinguishes only singular and plural subject pronouns; see Naden (c) for an overview of Oti-Volta West languages. 22 In general, the numbering follows Benue-Congo evidence and some numbers are added. There are also some “provisional reconstructions” that have not yet been integrated into the numbering system. It goes without saying that ongoing research on Adamawa languages will contribute to a more consistent reconstruction of the Gur nominal class systems (see Kleinewillinghöfer , a, and b).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Class Proto-Gur 1 *ʊ, *a
Pronoun Suffix ʊ~ɪ, a -ʊ, -a-a
Function Forms gender with SG 2, 4, 6, 10, 13, 13~21
1a
* ʊ, *a (-Ø) ʊ~ɪ, a
-Ø
SG
2a
2
*ba
-ba~-a
PL
1, 5, 12, 15
2a
*ba (-X-*ba) ba
PL
1, 1a
SG
4, 6, 13, *ci
3
*ŋʊ
4 5
ba
-X-ba 23
ŋʊ ~ʊ
-ŋʊ~-ʊ-ŋi
*i
i
-i
PL
1, 3, 5, 12, 14, 15, 19, 20, 23
*ɖi
ɖi
-ɖi
SG
2, 4, 6, 10, 13, 21
6
*ŋa
ŋa~ka~a -ka~-a
PL
1, 3, 5, 12, 14, 15, 19, 22~23
7
no evidence in Gur (BC *ki)
8
no convincing evidence in Gur (BC*bi)
9
no convincing evidence in Gur (BC*ì, PB *N)
10
*ni
ni na, da -ni -na24
U PL
1, 5, 12, 14, 15, 19, 20, 21
11
no evidence in Gur (BC *lu)
12
*ka
ka
-ka
SG
2, 4, 6, 10, 13, 21, 22
13
*sɪ
sɪ
-sɪ
PL
1, 3, 5, 12, 14, 15, 19, 21
14
*bʊ
bʊ~bɪ
-bʊ~-bɪ
SG U
4, 6, 10, 13, 21, *ci
15
*kʊ
kʊ~kɪ
-kʊ~-kɪ
SG
2, 4, 6, 10, 13, 21
16
no convincing evidence in Gur (PB*pa)
17 18
*kʊ kʊ -kʊ25 no evidence in Gur (PB *mu)
(LOC)
19
*fʊ
SG
4, 6, 10, 13, 21, *ci 4, 10, 21
fʊ~fɪ
-fʊ~-fɪ
20
*da
da, dɪ
-la
SG
21
*tʊ
tʊ~tɪ
-tʊ~-tɪ
SG U PL 1, 5, 10, 12, 14, 15, 19, 20
22
*mʊ
mʊ
-mʊ
U PL
12, 14
23
*ma
ma
-ma
U
4, 6
24
A class *-na- was proposed in Volume I. After gaining deeper insights, we now refrain from separating it from class 2a.
25
*wa *na *ni
– – ni28
-wa, -ba, -ya, -a26 SG -na,27 -an SG -ni -yI -E SG
10, 13, 22
*ci
ci
-ci, -k, -d -fi29
3, 3~15, 14, 19
*JE
This was tentatively established in Volume I, but now see plural class*ci-
*ŋwe
–
23 24 25 26 27 28 29 30
-ŋwV30
PL U
Gurunsi. Eastern Oti-Volta. Nateni, Kulango, and Ja̰a̰ne. Samu group, Kaansa-Dogose, Viemo, Kulango. -na in the Samu group; -an in Win. ni in Cεfɔ; -ni in Samwe and Win; -yɪ in Kulango; -E in Ja̰a̰ne. Ncam and Koromfe. Viemo and Dogɔsε.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Verb phrase The verb phrase has the canonical order S–(tense)–V–(aspect/tense)–(enclitics). Generally, the anaphoric subject pronoun is not repeated when it is preceded by the respective nominal subject. However, due to modifications within the information structure, subject pronouns may be repeated, as is the general rule in Bantu languages. An extraordinary order with verb-initial phrases is reported from Ka̰a̰sa and Ja̰a̰ne, two neighboring languages that belong to different sub-branches of Southern Central Gur. In both languages, it is the perfect paradigm which is marked by this extraordinary order (alongside another paradigm: past perfect in Ka̰a̰sa, and habitual in Ja̰a̰ne).31 In predicates, tense and modality are usually expressed by preverbal markers, while aspect marking is found with suffixes (more or less combined with stem vowel or tone variation); post-verbal clitics (among others) denote assertion. Aspect marking is widespread in Gur languages, but so multifaceted in form and function that Manessy (: ) refrained from reconstructing any morpheme. Apart from dichotomous systems with either imperfective or perfective as “marked” forms, there are triple systems with a “neutral” aspect from which the other two are obviously derived, even though often in unpredictable forms (see Winkelmann for a comprehensive overview of tense and aspect systems in Gur). “Narratives” and “aorists” on the one hand and imperatives and subjunctives on the other, as the most unmarked forms within the verb inflectional systems, are often identified as this third, i.e. neutral aspect form (see Carlson ). However, the formation of imperatives may differ considerably, as shown by Miehe () for Cerma, Ka̰a̰sa, and Ja̰a̰ne. As far as negation is concerned, we usually note preverbal markers, often combined with post-phrasal particles. Common to all languages is a separate prohibitive morpheme (see Winkelmann and Miehe for comparative evidence). To sum up, the interaction of the rich paradigmatic resources for building predicates varies considerably within the Gur family. In conclusion, two further characteristic features that are very common in Gur languages should here be mentioned. These are serial constructions, which are found in all languages, but more or less in complementary occurrence to verbal derivations (Reineke ), and the great amount of transitive “flexible” or “labile” verbs that allow diathesis only by moving patients to the subject position (and at the same time suppressing agents) without any further morphological means within an otherwise strict frame.32
A I am grateful to Kerstin Winkelmann and Ulrich Kleinewillinghöfer for comments on earlier versions, and to Ruth Schubert for polishing up my English.
31
For details, see Miehe (). Details are given by Lebikaza () for Kabiye; Carlson () for Supyire; Nicole () for Nawdm; and Reineke and Miehe () for Ditammari, Byali, and Ka̰a̰sa. 32
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Before citing the references given in the text, it should be pointed out that there is a special monograph series exclusively reserved for Gur languages: Gur Monographs/ Monographies voltaïques, edited by G. Miehe, B. Reineke, and M. von Roncador (Cologne: Rüdiger Köppe). In addition, there are two journals, Gur Papers/Cahiers Voltaïques and Studies in the Languages of the Volta Basin, devoted exclusively, or mainly, to topics concerning Gur languages. However, both journals ceased publication some years ago. The following charts give an idea of the contents: Gur Papers/Cahiers Voltaïques (edited by M. von Roncador, K. Winkelmann, B. Reineke, A. Schwarz, K. Beyer, and U. Kleinewillinghöfer)33
Phonology and tonology Nouns Adjectives Verbs and verbal systems Multi-verb constructions Derivation Syntax/information structure Lexicon
1 2 3 4 5 6 7 (1996) (1997) (1998) (1999) (2000) (2003) (2006) x x x x x x x x x x x x x x x x x x x x x x x x x x x x
Studies in the Languages of the Volta Basin (edited by M. E. Kropp Dakubu, E. K. Osam, Nana Aba Amfo, K. K. Saah, and G. Akanlig-Pare)34
1 2 3 4 5 6/1 6/2 (2003) (2004) (2005) (2007) (2008) (2010) (2010) Phonology and tonology x x x Noun classes x x x x Adjectives x x x x Verbs and verbal systems x x x x x Multi-verb constructions x x x x Derivation Syntax/information structure x x x Lexicon x x
33
Initiated by the Gur Working Group at the University of Bayreuth and the Humboldt University of Berlin. 34 The journal contains the results of the Legon-Trondheim Linguistic Project, which started in and ran for five years.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The following list contains only titles that are not included in Roncador and Miehe () or in Nicole (): Bakpa, M. (). Étude du ngbem, parler gangam de Koumongou. Münster: LIT. Beyer, K. (). La langue pana. Cologne: Rüdiger Köppe. Beyer, K. (). ‘Multilingual actors: examples from a West-African contact zone’, in C. de Féral, M. G. Kossmann, and M. Tosco (eds.), In and Out of Africa: Languages in question. In honour of Robert Nicolaï. Vol. : Language Contact and Language Change in Africa. Paris and Leuven: Peeters, –. Cahill, M. (). Aspects of the Morphology and Phonology of Konni. Dallas, TX: SIL. Carlson, R. J. (). ‘Event-views and transitivity in the Supyire verbal system’, Gur Papers/Cahiers Voltaïques : –. Dakubu, M. E. Kropp. (). Collected Language Notes on Dagaare Grammar. (Collected Language Notes No..) Legon: Institute of African Studies. Dakubu, M. E. Kropp. (). Parlons farefari (gurenè). Paris: L’Harmattan. Delplanque, A. (). ‘Identité des langues gur du Burkina Faso’, https://hal.archives-ouvertes.fr/ hal-v/document Delplanque, A. (). Le dagara-lobr. Leuven and Paris: Peeters. Dombrowsky-Hahn, K. (). A Grammar of Syer: Phonology, morphology and argument realization (Western Karaboro, Senufo). Cologne: Rüdiger Köppe. Dombrowsky-Hahn, K. (). ‘Relative clauses in Syer (Western Karaboro, Senufo, Gur)’, in R. Kramer and R. Kießling (eds.), –. Elders, S. (). Grammaire kulango (parler de Bouna, Côte d’Ivoire). Cologne: Rüdiger Köppe. Fiedler, I., Reineke, B., and Schwarz, A. (). ‘Let’s focus it: Fokus in Gur- und Kwasprachen’, in G. Heusing (ed.), Sprach- und literaturwissenschaftliche Beiträge zum : Afrikanistentag. Hamburg: LIT, –. Fiedler, I., and Schwarz, A. (). ‘Out-of-focus encoding in Gur and Kwa’, in I. Shinichiro, M. Schmitz, and A. Schwarz, A. (eds.), Interdisciplinary Studies on Information Structure . (Working Papers of the SFB .) Potsdam: University of Potsdam, –. Hiraiwa, K. (). ‘The head-internal relativization parameter in Gur: D and its typological implications’, The Proceedings of the North East Linguistic Society : –. Kleinewillinghöfer, U. (). ‘Kontaktphänomene im Südost-Gurunsi’, in T. Schumann, M. Reh, R. Kießling, and L. Gerhardt (eds.), Aktuelle Forschungen zu afrikanischen Sprachen. Cologne: Rüdiger Köppe, –. Kleinewillinghöfer, U. (). Correlations of the noun class systems of Central Adamawa and Proto Central Gur. Paper presented at the th International Conference on Gur Languages. Gur Internal and External Relationships, Bayreuth, July –, . Kleinewillinghöfer, U. (a). ‘The northern fringe of the Jos Plateau, a prehistorical contact zone of Benue-Plateau and Adamawa-Gur languages: the evidence of the cultural vocabulary’, in R. Kramer and R. Kießling (eds.), –. Kleinewillinghöfer, U. (b). Die ‘Ding-Klasse’ ne im Lɔŋto und seinen Verwandten. Paper presented at the Kolloquium: ‘Prof. Dr. Raimund Kastenholz’. Johannes Gutenberg-Universität Mainz, March , , https://www.blogs.uni-mainz.de/fb-adamawa/files///Ding-Klasse-ne-im-Lo %C%Bto.pdf accessed July , . Kra, K. A. E. (). Parlons kulango. Paris: L’Harmattan. Kramer, R., and Kießling, R. (eds.) (). Current Approaches to Adamawa and Gur Languages. Cologne: Rüdiger Köppe. Malgoubri, P. (). Recherches dialectologiques et dialectométriques nuni (une langue gurunsi du Burkina Faso). Doctoral dissertation. University of Leiden. Manessy, G. (). ‘Langues et histoire des peuples voltaïques: signification et limites de la comparaison historique’, Gur Papers/Cahiers Voltaïques : –. Micheli, I. (). Profilo grammaticale e vocabulario della lingua Kulango (Côte d’Ivoire). Doctoral dissertation. Università di Napoli “L’Orientale”.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Miehe, G. (). ‘Les pronoms personnels dans les langues gur’, in D. Ibriszimow and G. Ségerer (eds.), Systèmes de marques personnelles en Afrique. Paris: Peeters, –. Miehe, G. (). ‘Nasal(ised) vowels in Gur (Voltaic) languages: an assessment report’, in I. KraskaSzlenk and B. Wójtowisz (eds.), Current Research in African Studies: Papers in honour of Mwalimu Dr. Eugeniusz Rzewuski. Warsaw: Elipsa, –. Miehe, G. (). ‘Imperatives in Gur languages: evidence from Cerma, Kãasa and Jãane’, in R. Kramer and R. Kießling, R. (eds), –. Miehe, G., Kleinewillinghöfer, U., von Roncador, M., and Winkelmann, K. (). ‘Overview of noun classes in Gur II (revised and enlarged version)’, in G. Miehe, B. Reineke, and K. Winkelmann (eds.), –. Miehe, G., Owens, J., and von Roncador, M. (eds.) (). Language in African Urban Context. Part II: Burkina Faso. Berlin: LIT. Miehe, G., Reineke, B. and Winkelmann, K. (eds.) (). Noun Class Systems in Gur Languages. Vol. : North Central Gur Languages. Cologne: Rüdiger Köppe. Miehe, G., and Winkelmann, K. (eds.) (). Noun Class Systems in Gur Languages. Vol. : Southwestern Gur Languages (without Gurunsi). Cologne: Rüdiger Köppe. Nédellec, B. (). L’expression de la qualification en nateni. Cologne: Rüdiger Köppe. Nédellec, B. (). ‘Nateni’, in H. Tröbs, E. Rothmaler, and K. Winkelmann (eds.), La qualification dans les langues africaines: Qualification in African Languages. Cologne: Rüdiger Köppe, –. Nicole, J. (). Les classes nominales dans les langues voltaïques. Collection recherches techniques. Togo, Benin: SIL. Nicole, J. (). ‘ “La chèvre ne mange pas bien”: syntaxe et discours dans la phrase simple en nawdm’, Gur Papers/Cahiers Voltaïques : –. Nicole, J., and Nicole, M.-C. (). Les peuples voltaïques (gur): Bibliographie commentée des peuples parlant les langues voltaïques (ou gur). Cologne: Rüdiger Köppe. Niggli, U., and Niggli, I. (). Esquisse grammaticale du ninkare au Burkina Faso. (Cahiers de Recherches Linguistiques de la SIL et l’ANTBA Burkina Faso, No..) Ouagadougou: SIL. Niggli, U., and Niggli, I. (). The Structure of Burkina Kusaal. Ouagadougou: SIL. Pali, T. (). Description systématique de la langue miyobe (Togo/Benin). Doctorat nouveau régime, thèse de doctorat en cotutelle internationale: Université de Bordeaux/Université de Lomé (Togo). Palm, A. (). Description systématique du já̰á̰né, parler de Diebougou: Phonologie, morphologie et syntaxe. Thèse de doctorat unique en linguistique, Université de Ouagadougou (Burkina Faso). Reineke, B., and Miehe, G. (). ‘Diathesis alternation in some Gur languages’, in F. K. E. Voeltz (ed.), Studies in African Linguistic Typology. Amsterdam: John Benjamins, –. Reinhard, P. (). Parlons moba. Paris: L’Harmattan. Sambiéni, C. (). Le Proto-Oti-Volta-Oriental. Cologne: Rüdiger Köppe. Schwarz, A. (). Aspekte der Morphosyntax und Tonologie im Buli: Mit Schwerpunkt auf dem Buli von Wiaga, edoc online, urn:nbn:de:kobv: -, Doctoral dissertation, HumboldtUniversität Berlin. Schwarz, A. (). ‘Verb- and predication focus marker in Gur’, in I. Fiedler and A. Schwarz (eds.), The Expression of Information Structure: A documentation of its diversity across Africa. Amsterdam and Philadelphia, PA: John Benjamins, –. Showalter, S. D. (). The Same but Different: Language use and attitudes in four communities of Burkina Faso. SIL International. Traoré, D. (). Le senar (langue senufo du Burkina Faso). Göttingen: Cuvillier. von Roncador, M., and Miehe, G. (). Les langues gur (voltaïques): Bibliographie commentée et inventaire des appellations des langues. Cologne: Rüdiger Köppe. Williamson, K., and Blench, R. (). ‘Niger-Congo’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Winkelmann, K. (). ‘The basic structure of tense and aspect systems in Gur languages’, in R. Kramer and R. Kießling (eds), –. Winkelmann, K., and Miehe, G. (). ‘Negation in Gur: genetic, areal and unique features’, in N. Cyffer, E. Ebermann, and G. Ziegelmeyer (eds.), Negation Patterns in West African Languages and Beyond. Amsterdam: John Benjamins, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T Bantu family is the largest African language family in terms of geographic and demographic size. There are an estimated – Bantu languages, spoken by about million speakers in countries, from the Nigerian-Cameroonian borderland in the northwest of the Bantu-speaking area to Kenya in the northeast and to South Africa in the south (see Map .). The family includes languages with millions of speakers, such as Zulu, Shona, Nyanja, Kinyarwanda, and Swahili, but also smaller languages, many of which are poorly or not at all documented. However, overall Bantu languages are the most well-described language group in Africa, and the first to attract comparative work, resulting in a reconstructed proto-language, Proto-Bantu. In terms of classification, Bantu is deeply embedded within the Niger-Congo phylum and belongs to the Benue-Congo family within it (Figure .). The place of Bantu within Benue-Congo and, indeed, the exact membership and internal classification of Bantu are subject to some debate, and despite the early establishment of a reconstruction for the family, a comprehensive subgrouping of Bantu is still being addressed. Furthermore, in the linguistically complex Nigerian-Cameroonian borderland, about – Bantoid languages are spoken which are closely related to Bantu, and whose exact position with respect to “narrow” Bantu remains to be established. This, of course, has implications for what we mean by Bantu and for the properties of Proto-Bantu. Despite the antiquity of comparative-historical Bantu studies within African linguistics, there remains considerable scope for scholarly discovery.
. E B , P-B,
.................................................................................................................................. The earliest surviving records of Bantu languages consist of a few words noted down by early travelers in Africa, such as the Arab historian and traveler Al-Mas’udi, who visited the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Bantu languages
east African coast in the tenth century and recorded the Swahili word wafalme with the meaning ‘son of the Great Lord’ (Freeman-Grenville : ). More comprehensive descriptions of Bantu languages were the result of the onset of missionary activity in Angola and the Congo in the second half of the seventeenth century: in , the Italian missionary Giacinto Brusciotto published a grammar of Kikongo, the first known grammar of an African language, and in the following decades missionaries, travelers, and explorers continued to publish sketches of different Bantu languages. In the latter half of the eighteenth century, Abbé Proyart () noted the similarity between different Bantu languages spoken on the coast around the Congo estuary, as did the German missionary Oldendorp (), who recorded the languages of west African slaves in the Danish West Indies in the s. In southern Africa, Lichtenstein () compared Nguni and Tswana
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Niger-Congo Kordofanian Atlantic
Mande Ijoid Dogon East Volta-Congo Benue-Congo East Benue-Congo
West Volta-Congo Kwa
West Benue-Congo
Bantoid Bantu
. Bantu within Niger-Congo (after Williamson and Blench )
words and phrases he had learned when traveling in the Cape with Tsonga examples collected by White () in Maputo, and noted their relatedness. Meanwhile in London, the orientalist William Marsden, on the basis of these accounts and of data supplied by his Mozambican servant, proposed that central, southern, and eastern Bantu languages are genetically related (Tuckey : –). With increasing colonial expansion in the nineteenth century, interest in African languages and their comparative study increased too—a correlation which, as Said () notes, is not entirely accidental: the classification and grouping of the languages of Asia and Africa by European missionaries, travelers, and scholars was often exploited by colonial policies of dividing and ruling Asian and African territory and societies. Grammatical descriptions of Bantu languages increased, and, by the middle of the century, the material was rich enough for comparative studies such as Bleek (/), who introduced the term “Bantu” for the family, based on the common root -ntu ‘person’ and the animate plural prefix ba-, so meaning ‘people’, as well as the numbering convention for Bantu noun classes. Meinhof (, , ) expanded this work and at the end of the nineteenth century, on the basis of a small number of geographically balanced test languages (Duala, Herero, Kongo, Nyakyusa, Sango, Northern Sotho, Swahili, and Zulu), proposed the reconstructed Ur-Bantu or Proto-Bantu, which, in broad outline, has remained more or less accepted until today, and provided a main foundation and reference point for subsequent work in Bantu linguistics. The understanding of the position of Bantu languages within Niger-Congo resulted from comparative work on west African languages: Westermann’s () results implied that Bantu was part of the Niger-Congo phylum, and Greenberg (, ) developed a detailed classification in which Bantu became deeply embedded within Niger-Congo and one coordinate branch within the Benue-Congo group. While this does not give cause for surprise today, it was remarkable at the time, given the presumed importance of the Bantu languages in Africa (cf. Greenberg : ). Work on Bantu itself was subsequently significantly influenced by Guthrie’s (, –) comparative Bantu studies, further discussed in section . below, and the work of the “Lolemi” research group at Tervuren, inaugurated by Achille Meeussen, which resulted in numerous studies on lexical and grammatical aspects of Bantu, including
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Meeussen’s () Bantu grammatical reconstruction. Over the last forty years or so, Bantu languages have also become central to questions of theoretical linguistics, for example in the areas of phonology and especially tone (Hyman and Kisseberth ), syntax (Mchombo ), grammaticalization, tense aspect, and information structure (see Nurse and Philippson a).
. I
.................................................................................................................................. Bantu as a language family was firmly established by the beginning of the twentieth century. However, the internal subclassification of Bantu languages remained outstanding, and, despite several studies, this remains true today. An early, comprehensive study addressing the question is Johnston (/). However, it was not until almost fifty years later that the most influential study of the internal classification of Bantu appeared, the geographicalreferential classification by Malcolm Guthrie (–). On the basis of comparative material of over languages, Guthrie divided Bantu languages into language groups (Table .), which are assigned to geographical zones. Each zone is assigned an alphabetical letter from A in the northwest to S in the south (Map .), and within each zone, languages or language groups are assigned a letter and a number, so that, for example, the Swahili group is G, and within this Swahili itself G (see Maho ). In addition, Guthrie proposed a set of reconstructions, to add to the forms proposed by Meinhof and his followers, as well as a long list of what he called “starred forms”. These were not meant to be historical hypotheses, as reconstructed forms are, but merely summary statements of the variation found in his data (“Common Bantu” as opposed to Proto-Bantu). Yet the distinction is subtle, and in practice the forms often serve as reference points for historical work, and they also formed one of the bases for the large online database of Bantu Lexical Reconstructions BLR (BLR). Guthrie’s zonal classification has sometimes been criticized for not corresponding well to genetically more plausible groupings. For example, a genetically justified group of interlacustrine (i.e. between the east African Great Lakes) languages (including, for example, Rwanda, Rundi, Haya, and Luganda) straddles Guthrie’s zones D and E, and sometimes a new zone, zone J, is used to refer to this group. But in a sense, the criticism is beside the point. The only measurable quality of a referential classification such as Guthrie’s is whether it is used widely (which it is), while a genetic classification may result in principle in entirely different groupings (cf. Maho ). The problems with a truly genetic grouping of Bantu have been noted frequently (e.g. Meinhof : ; ; Nurse /; Vansina : ; Schadeberg ): the large amount of data needed—from all or so Bantu languages—for a comprehensive comparative study, the effect of language contact and borrowing, which may obscure historical divergence, and the question whether a traditional genetic tree-based model, assuming binary splits of diversion, is adequate for a situation which might better be characterized as a convergence area, or “a huge single pool of dialects” (Vansina : ). The situation has prompted mainly two kinds of response: the detailed study of subgroups, so as to develop an overall classification from bottom up, and large-scale comparison of only a limited set of data, often with innovative methods (in addition, non-genetic classifications have been proposed as well, discussed in section .). Among the former are
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Main groups of Bantu languages (based on Guthrie –; Maho ) A Lundu-Balong A Duala A Bube-Benga A Basa A Bafia A Sanaga A Yaunde-Fang A Maka-Njem A Kaka
D Ruanda-Rundi
B Myene B Kele B Tsogo B Shira-Punu B Njabi B Mbete B Teke B Tende-Yanzi
F Tongwe F Sukuma-Nyamwezi F Ilamba Irangi
C Ngundi C Mboshi C Bangi-Ntumba C Ngombe C Soko-Kele C Mongo-Nkundu C Tetela C Kuba D Mbole-Ena D Lega-Kalanga D Bira-Huku D Konjo D Bembe-Kabwari
E Nyoro-Ganda E Haya-Jita E Masaba-Luhya E Ragoli-Kuria E Kikuyu-Kamba E Chaga E Nyika-Taita
L Luba L Kaonde L Lunda L Nkoya M Fipa-Mambwe M Nyika-Safwa M Konde M Bemba M Bisa-Lamba M Lenje-Tonga N Manda N Tumbuka N Nyanja N Senga-Sena
G Gogo G Shambala G Zigula-Zaramo G Swahili G Pogolo G Bena-Kinga
P Matumbi P Yao P Makua
H Kikongo H Kimbundu H Kiyaka H Kimbala
R Umbundu R Ndonga R Herero R Yeye
K Chokwe-Luchazi K Lozi K Luyana K Subiya
S Shona S Venda S Sotho-Tswana S Nguni S Tswa-Ronga S Inhambane
L Pende L Songe
Nurse and Hinnebusch’s () establishment of a Northeast-Coast Bantu group, interlacustrine Bantu languages (Schoenbrun ), or southern Bantu (Doke ; Janson /; Herbert and Bailey ). From this and similar work, a more comprehensive classification of Bantu overall can gradually be built up, although this requires work on many more subgroups (cf. also Möhlig ). The second possible response is to work with a large number of languages, but with only a restricted data set. This has been done in particular with lexical data sets employing lexicostatistic, and later phylogenetic methods, in a number of studies (e.g. Heine et al. , Bastin et al. , Holden , Holden and Gray , Grollemund et al. ; cf. Marten ). Heine et al.’s () results show a
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. Bantu language groupings (based on Guthrie )
distinction between three main Bantu subgroups: a northwestern group, a western-central group, and, as the largest group, a southeastern group. Bastin et al.’s () extensive lexicostatistical study is based on a -word list of basic vocabulary from Bantu languages. The results of the study are, however, ambiguous. In their overall analysis, Bastin et al. () conclude that, first, hierarchical, binary tree relations are not well suited for expressing their results; second, Bantu languages share histories of divergence as well as contact and convergence, and these histories are sometimes impossible to tell apart; and, third, by varying the computational parameters used to analyze the data, a series of trees results which expresses the ambiguous relations between the Bantu languages of the sample
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Bantu Northwest Central-west Westwest Southwest
East
. Main subgroups of Grollemund et al.’s () classification
better than a single tree. The computational parameters are connectivity (the degree of coherence within a given group of languages) and exclusivity (the degree of similarity with members from outside the group), as well as different thresholds of similarity between different nodes in the tree. The resulting family of trees provides a “different perspective on linguistic relationships which are all part of the total picture” (: ), and so there is no single correct tree to model the relationship between different Bantu languages. Vansina (), commenting on prepublication results of the study, takes the argument further, and proposes that the relationship between Bantu languages cannot be represented by the treemodel, but fits rather well a wave-model of language change which allows overlapping changes and multiple inheritance. Another approach is taken by studies using phylogenetic methods adopted from evolutionary biology, such as Holden (), Rexová et al. (), and Grollemund et al. (), which allow computation over the set of possible trees compatible with the data to establish the most consistent or “consensus” tree. Grollemund et al.’s () study is based on -word lists from languages and does result in such a consensus tree with hierarchical subgroupings, in which five main branches are identified: northwest, central-west, westwest, southwest, and east Bantu (Figure .). A few studies have also been conducted comparing non-lexical data (Bastin ; Ehret ; Nurse and Philippson b). Nurse and Philippson (b), for example, use a variety of phonological and morphological features. They argue that lexical data are not well suited for subclassification and the historical study of language development, since large parts of a language’s vocabulary might be borrowed, and if, as in the case of Bantu, borrowing is often from closely related languages, borrowed forms are hard or even impossible to detect. On the basis of their non-lexical features, Nurse and Philippson (b) propose a number of language groups, which can be combined into four large groups: west, forest, west-central, and northeastern savannah. No southeastern group is postulated, in contrast to many previous studies, and Nurse and Philippson (b) note that no such group is supported by the features adopted. In sum, most proposed subclassifications of Bantu languages identify a northwestern group, including more or less the Bantu languages of Cameroon, Gabon, Congo, and the northwestern part of the DRC, and a large eastern or southeastern group of Bantu languages from Kenya to South Africa. In addition, two further groups are often distinguished: a southwestern group to the south of the northwestern group, including languages of Angola and Namibia, and a central group of languages between the eastern and the western groups. Even though no fully worked-out subclassification of Bantu has been proposed so far—despite considerable efforts—the results which have been achieved had a great influence on the historical study of the areas where Bantu languages are spoken today, in particular with respect to the “Bantu expansion”, or “Bantu dispersals”.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. B
.................................................................................................................................. A particular concern of comparative Bantu has for a long time been the relation between classification and social history, and hypotheses about the spread of Bantu languages across central, eastern, and southern Africa have had considerable influence on models of African history. Because Bantu languages are quite similar, it has often been assumed that the fact that they are spoken over a wide geographic area is a comparatively recent phenomenon. This then led to the question of how Bantu languages, or their speakers, became so widely dispersed. The three interrelated questions—in which direction did the dispersal of Bantu languages take place, what were the factors involved in the process, and when did it happen—have received different answers over the last century. In an early hypothesis, Johnston (/) proposes that Bantu languages originated from Bantoid languages of Nigeria, which he called “Semi-Bantu” languages, and then were dispersed eastwards and southwards. As it turned out, this was very much along the lines of what was later shown to be the best explanation. However, in the s and s, the question was subject to a lively discussion between Guthrie (), Greenberg (), and the historian Oliver (). Greenberg argued that Bantu languages were originally spoken in the Cameroon-Nigerian borderland, the area with the highest complexity and greatest diversity of Bantu languages, and then subsequently came to be spoken throughout the rest of today’s Bantu-speaking area. Guthrie, in contrast, argued that the origin of Bantu languages, the “Bantu nucleus”, lies in the Congo basin. His main argument against Greenberg’s proposal was that, in order for Bantu languages to have spread from Cameroon, they (or their speakers) would have had to have crossed the tropical rainforest of Central Africa, which he considered impossible. Yet Guthrie had to accept, reluctantly, that languages of Cameroon and Nigeria were related to Bantu, and so that speakers (or languages) had to be able to cross the rainforest, in his case from the Congo to Cameroon. But that, of course, meant that Guthrie had to concede that movement from Cameroon to the Congo was possible after all, since, as Greenberg pointed out, “if there is a road from New York to Chicago, there has to be a road from Chicago to New York” (: ). Greenberg’s proposal was further substantiated by Heine et al.’s () lexicostatistical, as well as by Grollemund et al.’s () phylogenetic results, and is widely accepted today. The most likely historical interpretation of the linguistic facts (cf. Vansina , ; Ehret ) is that Bantu languages spread through a succession of small migrations, involving backwards and forwards movements, rather than through one large “expansion”, so that the present distribution of Bantu languages is the outcome of “many complex historical dynamics involving successive dispersals of individual languages over a time span of millennia and involving reversals as well as successes” (Vansina : ). The migrations were accompanied by language shift of speakers of other languages to Bantu languages, so that the process involved both the physical movement of speakers and the dispersal of languages without speaker movements. The reasons why Bantu languages appeared attractive may be related to technical, economic, or social differences. Early Bantu communities may have been larger than the more widely dispersed communities they met with, and their languages may have been associated with more perceived advantages because of this. Given the frequent postulation of a main split between northwestern, western, and southeastern Bantu languages
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
(see section .), a possible model of the historical dispersal is an initial spread from the original Bantu-speaking area eastwards and southwards through the rainforest, with groups of Bantu speakers emerging to the south of the rainforest, reflected in the western Bantu group, and to the east of the forest, eventually leading to the further spread of Bantu languages through the drier savannah and shrub lands of eastern and southern Africa, although all these processes may have been accompanied by and overlapped with parallel and subsequent movements. The most recent hypothesis, developed by Grollemund et al. (), is that early Bantu speakers took advantage of the emergence of a rainforest “corridor” resulting from periodic climate change around , whereby for some time the rainforest receded, allowing for much easier passage through it. The time of the Bantu dispersal can further be estimated by correlation of Proto-Bantu vocabulary with archaeological evidence (Vansina ): Proto-Bantu vocabulary for pottery and farming indicates that Bantu languages had not dispersed before the earliest archaeological attestation of these activities in western Cameroon (around ). On the other hand, the absence of iron-smelting vocabulary, and its distribution across different Bantu languages show that by the time the technology is first attested (around the Great Lakes c. ) the Bantu dispersals were well under way.
. B B
.................................................................................................................................. For the subclassification of Bantu languages, as well as for the understanding of Bantu languages within the wider Niger-Congo phylum, the relation between Bantu and its closest neighbors within the Benue-Congo group of Niger-Congo is of central importance. Yet many questions remain at present unanswered about this relation. Geographically, the question concerns the northwest of the Bantu area: Bantu languages border on Khoisan (and Germanic) languages in the south, and on Nilo-Saharan, Cushitic, and Ubangian languages in the north and northeast, and the difference between these languages and Bantu are on the whole quite clear. However, in the northwest of the Bantu area Bantu languages are spoken in the neighborhood of closely related Bantoid languages, which are felt by many observers, including such early studies as Koelle () and Johnston (/) to be similar to Bantu in terms of vocabulary or structure, but yet not similar enough to be obviously part of “narrow Bantu”. There are some Bantoid (excluding narrow Bantu) languages spoken in western Cameroon and eastern Nigeria (Blench ). Some disagreement exists as to the linguistic grouping of the languages, and their classification is still in progress. Lewis () includes the following thirteen groups (with the number of languages in the group provided in brackets): Beboid (), Dakoid (), Ekoid (), Fam (), Jarawan (), Mambiloid (), Mamfe (), Mbam (), Mbe (), Ndemli (), Tikar (), Tivoid (), and Wide Grassfields (), in addition to six unclassified languages. However, the precise membership and internal structure of these groups is often still not definitely established (e.g. Piron ; Williamson and Blench ). Grassfields languages are sometimes called “Grassfields Bantu” (e.g. Watters ), and the term “Ekoid Bantu” has been used as well (e.g. Crabb ), showing the perceived closeness to narrow Bantu, while the Mbam group is also called “Mbam Bantu” and includes languages of the Bantu A and A groups. Not only geographically, but also in terms of structural
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
characteristics, Bantoid languages lie between other Benue-Congo and Bantu languages. Often several lexical items can be related to Proto-Bantu forms, and many Bantoid languages have (traces of) noun classes similar to Bantu noun classes. However, it is often difficult to distinguish between genuine shared Bantu properties and those which Bantu and Bantoid languages share because they are wider Niger-Congo features. The problems related to establishing a classification of Bantoid languages, and their relation to northwestern Bantu languages, are similar to those facing attempts to subclassify Bantu languages (section .): conflicting isoglosses and potentially high levels of borrowing which obscures linguistic inheritance. In addition, for a number of languages, often small and undescribed, there are not enough data to judge their relationship with other languages in a complex linguistic area with a high degree of linguistic variation. Finally, presumably due to widespread processes of phonological attrition, many complex lexical roots in Bantu correspond to short, monosyllabic roots in Bantoid, which makes comparison more difficult. Responses to these problems are, again, similar to Bantu subclassification: the construction of small groups so as to develop larger relationships bottom-up, the use of both lexical and non-lexical data, and the use of lexicostatistical methods, and of alternative models of language relation (see section .). Watters () provides an overview and synthesis of earlier classifications and distinguishes a northern and a southern Bantoid group, the latter including narrow Bantu. A comprehensive lexicostatistical study of Bantoid languages is provided in Piron (). As with other lexicostatistical studies (e.g. Bastin et al. , discussed in section . above), the results are not easy to interpret, and Piron’s study results in a number of different trees, depending on the particular analytical parameters employed. One tree, based on branch average distance, is given in Figure .. The classification identifies three different Bantu groups, but separates southeast Bantu from other south Bantoid languages (including central and northwest Bantu), and central Bantu from northwest and non-Bantu south Bantoid. Grollemund et al. (forthcoming) is a more recent phylogenetic study of Bantoid based on -word lists of languages. It distinguishes eleven main groups (Bantu, Jarawan, Mabm-Bubi, Grassfields, Beboid, Bantoid Dakoid Mambiloid Tikar
South Bantoid South and East Bamtu Central Bantu
Non-Bantu South Bantoid
Northwest Bantu
Jarawan Tivoid Beboid Ekoid
B10
Mbam Wide Grassfields
A50
B30
A + B20
. A lexicostatistical classification of Bantoid (after Piron : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Bantoid
Ekoid Bendi Dakoid Jukunoid Beboid Kenyang Tivoid Mambiloid
Tikar Grassfield Mbam-Bubi Jarawan Bantu
. Main groupings of the phylogenetic classification of Bantoid of Grollemund et al. (forthcoming)
Mambiloid, Tivoid, Jukunoid, Dakoid, Bendi, and Ekoid) and two isolates (Tikar and Kenyang). As in previous work, a group of Bantu A languages (A, A, A, and possibly A), the Mbam-Bubi languages, are outside of narrow Bantu and are grouped closer to the Jarawan group. More controversially, Jukunoid is classified as part of Bantoid, rather than as a higher grouping in Benue-Congo. A simplified representation of the classification is given in Figure .. Even though more data are likely to change the picture of the classification of Bantoid languages, it is clear that the situation is complex and has considerable potential impact on the classification and reconstruction of narrow Bantu: one implication of current work on Bantoid is that what is currently termed “Bantu” might not be a coherent genetic unit, and that current Bantu reconstructions do not take sufficient account of northwestern Bantu languages. On the other hand, as with Bantu, the complexity of the relation between Bantoid languages has also led to exploration of non-genetic approaches to classification.
. N-
.................................................................................................................................. The challenges encountered in establishing robust subgroups of Bantu and Bantoid languages has led to the application of different and more complex methods for the study of linguistic relationships, including lexicostatistical and phylogenetic methods, as well as the use of non-lexical data, such as morphological innovations. These approaches share the assumption that ultimately Bantu languages can be classified in some tree-like arrangement, reflecting shared linguistic innovations. However, an alternative response has been to regard the interrelation of Bantu/Bantoid languages not as an empirical challenge, but as a conceptual one, and that a potential genetic classification of Bantu needs to be supplemented, or even be replaced, by an alternative classification. The most widely used of these alternatives are typological classifications. In fact, typological criteria were often mixed with genetic criteria in the history of Bantu classification without sharply distinguishing between the two. Guthrie’s (: –) criteria for the identification of Bantu languages, as a famous example, are mainly typological, including presence of a noun class system encoded
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
by prefixes, CVC lexical roots, and a symmetric vowel system, but they also include the genetic criterion of reconstructible vocabulary (Gerhardt ). A range of comparative studies have been undertaken by members of the “Lolemi” school of Bantu linguistics in Tervuren. Although not specifically typological in theoretical outlook, cross-Bantu studies of, for example, relative clauses (Nsuka-Nkutsi ) or locative classes (Grégoire ) provide comprehensive case studies of similarities and variation across Bantu. The striking overall typological similarity of Bantu languages and the distribution of different typological features across different subgroups of languages are reminiscent of historical diffusion rather than historical diversification, reflecting a long equilibrium rather than punctuation in the terms of Dixon (). Alternatively, in areal terms, the Bantu area can be thought of as a spread zone (Nichols ), in which linguistic features diffuse, resulting in overall similarity, and a crisscrossing of similarities and differences across many different varieties, without clearly distinguishing groups of varieties. This idea is developed, for example, by Möhlig (), who proposes several historical Bantu convergence zones (“stratificational nuclei”), and more recently by Güldemann () who proposes a Bantu spread zone (including mainly central and southeastern Bantu languages) and a wider Macro-Sudan belt, which also includes northwestern Bantu languages at its periphery. An areal approach provides a new perspective on Bantoid languages as well: located between the Bantu spread zone, on the one hand, and the typologically quite distinct Kwa languages, Bantoid languages show similarities with both, constituting a typological “buffer zone” between the two (Good ). A central aspect of all these explanations is language contact. A different contact-based approach to Bantu classification is to compare languages with different roles in intra-Bantu contact situations, where it appears that convergence is particularly driven by languages with high numbers of second-language speakers (cf. Marten et al. ). The adoption of new models of explanation, as well as continuing work within more established paradigms, shows that, despite the relative antiquity of the family, Bantu and Bantoid classification remains a challenging and rewarding field of study.
R Bastin, Y. (). ‘Essai de classification de quatre-vingts langues bantoues par la statistique grammaticale’, Africana Linguistica : –. Bastin, Y., Coupez, A., and Mann, M. (). Continuity and Divergence in the Bantu Languages: Perspectives from a lexicostatistic study. Tervuren: Musée Royal d’Afrique Centrale. Bleek, W. H. I. (/). A Comparative Grammar of South African Languages. vols. London: Trübner. Blench, R. (). Archaeology, Language, and the African Past. Oxford: Alta Mira Press. BLR, Bantu Lexical Reconstructions . Online database, http://www.metafro.be/blr accessed October , . Brusciotto, G. (). Regulae quaedam pro difficillimi Congensium idiomatis faciliori captu ad grammaticae normam redactae. Rome: Typis Sacra Congregatio de Propaganda Fide. Crabb, D. W. (). Ekoid Bantu Languages of Ogoja. Part . Cambridge: Cambridge University Press. Dixon, R. M. W. (). The Rise and Fall of Languages. Cambridge: Cambridge University Press. Doke, C. M. (). The Southern Bantu Languages. London: Oxford University Press. Ehret, C. (). An African Classical Age: Eastern and southern Africa in world history, B.C. to A.D. . Charlottesville, VA: University Press of Virginia.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Ehret, C. (). ‘Subclassifying Bantu: the evidence of stem morpheme innovations’, in J.-M. Hombert and L. M. Hyman (eds.), Bantu Historical Linguistics: Theoretical and empirical perspectives. Stanford, CA: Center for the Study of Language and Information, –. Freeman-Grenville, G. S. P. (). The East African Coast: Select documents from the first to the earlier nineteenth century. Oxford: Clarendon. Gerhardt, L. (). ‘Sprachvergleich und Rekonstruktion’, in B. Heine, T. C. Schadeberg, and E. Wolff (eds.), Die Sprachen Afrikas. Hamburg: Helmut Buske, –. Good, J. (). ‘A (micro-)accretion zone in a remnant zone? Lower Fungom in areal-historical perspective’, in B. Bickel, L. A. Grenoble, D. A. Peterson, and A. Timberlake (eds.), Language Typology and Historical Contingency. Amsterdam and Philadelphia, PA: John Benjamins, –. Greenberg, J. H. (). ‘Studies in African linguistic classification III: the position of Bantu’, Southwestern Journal of Anthropology : –. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Greenberg, J. H. (). ‘Linguistic evidence regarding Bantu origins’, Journal of African History : – [repr. in K. Denning and S. Kemmer (eds.), On Language: Selected writings of Joseph H. Greenberg. Stanford, CA: Stanford University Press, –]. Grégoire, C. (). Les locatifs en Bantou. Tervuren: Musée Royal de l’Afrique Centrale. Grollemund, R., Branford, S., Bostoen, K., Meade, A., Venditti, C., and Pagel, M. (). ‘Bantu expansion shows habitat alters the route and pace of human dispersals’, Proceedings of the National Academy of Sciences () –. doi.org/./pnas. Grollemund, R., Branford, S., Hombert, J.-M., and Watters, J. R. (forthcoming). ‘Towards a new phylogenetic classification of Bantoid languages’, in J. R. Watters (ed.), Towards Proto-NigerCongo: Comparison and reconstruction. Berlin: Language Science Press. Güldemann, T. (). ‘Proto-Bantu and Proto-Niger-Congo: macro-areal typology and linguistic reconstruction’, in O. Hieda, C. König, and H. Nakagawa (eds), Geographical Typology and Linguistic Areas, with special reference to Africa: International Symposium, Tokyo University of Foreign Studies, May –, . Amsterdam and Philadelphia, PA: John Benjamins, –. Guthrie, M. (). The Classification of the Bantu Languages. London: Oxford University Press for the International African Institute. Guthrie, M. (). ‘Bantu origins: a tentative new hypothesis’, Journal of African Languages : –. Guthrie, M. (–). Comparative Bantu: An introduction to the comparative linguistics and prehistory of Bantu languages. vols. Farnborough: Gregg. Heine, B., Hoff, H., and Vossen, R. (). ‘Neuere Ergebnisse zur Territorialgeschichte der Bantu’, in W. J. G. Möhlig, F. Rottland, and B. Heine (eds.), Zur Sprachgeschichte und Ethnohistorie in Afrika: Neue Beiträge afrikanistischer Forschungen. Berlin: Dietrich Reimer, –. Herbert, R. K., and Bailey, R. (). ‘The Bantu languages: sociohistorical perspectives’, in R. Mesthrie (ed.), Language in South Africa. Cambridge: Cambridge University Press, –. Holden, C. (). ‘Bantu language trees reflect the spread of farming across sub-Saharan Africa: a maximum parsimony analysis’, Proceedings of the Royal Society of London Series B : –. Holden, C., and Gray, R. (). ‘Rapid radiation, borrowing and dialect continua in the Bantu languages’, in P. Foster and C. Renfrew (eds.), Phylogenetic Methods and the Prehistory of Languages. Cambridge: McDonald Institute for Archaeological Research, –. Hyman, L. M., and Kisseberth, C. (eds.) (). Theoretical Aspects of Bantu Tone. Stanford, CA: Center for the Study of Language and Information. Janson, T. (/). ‘Southern Bantu and Makua’, Sprache und Geschichte in Afrika /: –. Johnston, H. H. (/). A Comparative Study of the Bantu and Semi-Bantu Languages. vols. Oxford: Clarendon. Koelle, S. W. (). Polyglotta Africana. London: Church Missionary House. Lewis, M. P. (ed.) (). Ethnologue: Languages of the World, th edn. Dallas, TX: SIL International, http://www.ethnologue.com accessed July , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Lichtenstein, M. H. K. (). ‘Bemerkungen über die Sprachen der südafricanischen wilden Völkerstämme, nebst einem kleinen Wörterverzeichnisse aus den gebräuchlichsten Dialecten der Hottentotten und Kaffern’, Allgemeines Archiv für Ethnographie und Linguistik : –. Maho, J. (). ‘A classification of the Bantu languages: an update of Guthrie’s referential system’, in D. Nurse and G. Philippson (eds.), –. Marten, L. (). ‘Bantu classification, Bantu trees and phylogenetic methods’, in P. Foster and C. Renfrew (eds.), Phylogenetic Methods and the Prehistory of Languages. Cambridge: McDonald Institute for Archaeological Research, –. Marten, L., Kula, N. C., and Thwala, N. (). ‘Parameters of morphosyntactic variation in Bantu’, Transactions of the Philological Society : –. Mchombo, S. A. (ed.) (). Theoretical Aspects of Bantu Grammar. Stanford, CA: Center for the Study of Language and Information. Meeussen, A. E. (). ‘Bantu grammatical reconstructions’, Africana Linguistica : –. Meinhof, C. (). Grundriss einer Lautlehre der Bantusprachen. Leipzig: Brockhaus. Meinhof, C. (). Grundriss einer Lautlehre der Bantusprachen, nd edn. Berlin: Dietrich Reimer. Meinhof, C. (). Introduction to the Phonology of the Bantu Languages. Translated by N. van Warmelo. Berlin: Dietrich Reimer. Meinhof, C. (). ‘Die Gliederung der Bantusprachen’, Wiener Zeitschrift für die Kunde des Morgenlandes : –. Möhlig, W. J. G. (). ‘Stratification in the history of the Bantu languages’, Sprache und Geschichte in Afrika : –. Nichols, J. (). Linguistic Diversity in Time and Space. Chicago, IL: Chicago University Press. Nsuka-Nkutsi, F. (). Les structures fondamentales du relatif dans les langages bantoues. Tervuren: Musée Royal de l’Afrique Centrale. Nurse, D. (/). ‘Historical classifications of the Bantu languages’, Azania : –. Nurse, D., and Hinnebusch, T. J. (). Swahili and Sabaki: A linguistic history. Berkeley, CA: University of California Press. Nurse, D., and Philippson, G. (eds.) (a). The Bantu Languages. London and New York: Routledge. Nurse, D., and Philippson, G. (b). ‘Towards a historical classification of the Bantu languages’, in D. Nurse and G. Philippson (eds.), –. Oldendorp, C. G. A. (). Historie der caribischen Inseln Sanct Thomas, Sanct Crux und Sanct Jan, insbesondere der dasigen Neger und der Mission der evangelischen Brüder unter denselben. Erster Teil. Kommentierte Ausgabe des vollständigen Manuskriptes aus dem Archiv der Evangelischen Brüder-Unität Herrnhut. Edited by G. Meier, S. Palmié, P. Stein, and H. Ulbricht. Berlin: Verlag für Wissenschaft und Bildung. Oliver, R. (). ‘The problem of the Bantu expansion’, Journal of African History : –. Piron, P. (). Classification interne du groupe bantoïde. vols. Munich: Lincom Europa. Proyart, L. B. (). Histoire de Loango, Kakongo et autres royaumes d’Afrique. Paris [repr. . Farnborough: Gregg]. Rexová, K., Bastin, Y., and Frynta, D. (). ‘Cladistic analysis of Bantu languages: a new tree based on combined lexical and grammatical data’, Naturwissenschaften : –. Said, E. (). Orientalism. New York, NY: Vintage Books. Schadeberg, T. C. (). ‘Historical linguistics’, in D. Nurse and G. Philippson (eds.), –. Schoenbrun, D. L. (). Early History in Eastern Africa’s Great Lakes Region: Linguistic, Ecological and Archaeological Approaches. Ph.D. thesis. Los Angeles, CA: University of California. Tuckey, J. K. (). Narrative of an Expedition to Explore the River Zaire, Usually Called the Congo, in South Africa in . London: John Murray. Vansina, J. (). Paths in the Rainforests: Toward a history of political tradition in Equatorial Africa. London: James Currey. Vansina, J. (). ‘New linguistic evidence and “the Bantu expansion” ’, Journal of African History : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Watters, J. R. (). ‘Bantoid overview’, in J. Bendor-Samuel (ed.), The Niger-Congo languages: A classification and description of Africa’s largest language family. Lanham, MD: University Press of America, –. Watters, J. R. (). ‘Grassfields Bantu’, in D. Nurse and G. Philippson (eds.), –. Westermann, D. H. (). Die westlichen Sudansprachen und ihre Beziehungen zum Bantu. Berlin: de Gruyter. White, W. M. (). Journal of a voyage performed in the Lion Extra Indiaman, from Madras to Columbo and Da Lagoa Bay, on the eastern coast of Africa (where the ship was condemned), in the year : with some account of the manners and customs of the inhabitants of Da Lagoa Bay and a vocabulary of the language. London: Stockdale. Williamson, K., and Blench, R. (). ‘Niger-Congo’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
̈
. A
.................................................................................................................................. “A” is the classificatory label for about ninety languages, spoken in the central part of the sub-Saharan savannah belt (Sudan zone), in an area which stretches from the mountains bordering the basins of the Middle Benue and the Lower Gongola in northeast Nigeria across the north of Cameroon to the east into Chad and the Central African Republic (see Map .). Adamawa languages are among the least documented in Africa. Many of them are confined to just a few settlements; several are on the brink of extinction. By far the largest language is Mumuye with more than , speakers, followed by the Kebi-Benue languages Mundang and Tupuri with about , and , speakers respectively (Lewis et al. ). The term “Adamawa” derives from Lamido Modibo Adama (–), the founder of the Emirate of Fombina (centered at Yola), and depicts the territory he brought under his control.1 Greenberg () introduced Adamawa as the name of one of originally fourteen branches of the Niger-Congo language phylum. In , he joined “Adamawa” with the neighboring “Eastern Branch” to form “Adamawa-Eastern”, one of six families constituting Niger-Congo. Adamawa-Eastern was renamed “Adamawa-Ubangi” by Samarin (), the term still being broadly used, although no study has ever reasonably demonstrated its genetic validity. On the contrary: the coherence of Adamawa-Ubangi as a whole and the unity of either branch have seriously been contested. There are still various conflicting ideas regarding what “Adamawa” stands for. In addition to their insufficient documentation, the marked linguistic diversity of Adamawa languages greatly adds to its contested status. Undoubtedly, the intense contact the languages have been exposed to in the past has contributed significantly to their remarkable divergence. Being located in the center of the sub-Saharan savannah belt, south of the Lake Chad basin, where three of Africa’s language phyla meet, the area must have been a hotspot of language contact for a long time (see also Güldemann ). Severe climatic changes (desertification, periodical dry spells) (cf. Maley and Vernet ), the 1
Cf. “Adamawa State” in Nigeria and “Région d’Adamaoua” and “Plateau d’Adamaoua” in Cameroon.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Adamawa languages
emergence of (feudalistic) states, and their subsequent raids on the minority groups of the wider region are probably the most crucial events which caused extensive relocations of peoples resulting in numerous varying contact scenarios in the past millennia.2 The strongest evidence contesting the validity of Adamawa as a genetic unit and its status as a branch of an Adamawa-Ubangi family consists in the obvious relations of (certain) Adamawa groups with Central Gur. This alone led Köhler (: ) to already regard Adamawa as distinct from Ubangi. Bennett and Sterk () and in particular Bennett () opine on the other hand that Adamawa-Ubangi languages rather form a continuum with Gur. They named the continuum “North Central Niger-Congo”, which was altered to “North Volta-Congo” in Williamson ().3 Central to their point of view are the results of lexicostatistical studies which indicate that “Tula-Longuda . . . is by the figures of this test as close to—or as far from—Gurunsi as it is to Boa-Kula and Chamba-Namshi”4 2 The Kwararafa State, for example, a confederacy of peoples centered in the fifteenth/sixteenth century at the northwestern periphery of the Adamawa area (Webster ), apparently brought forth a linguistic area comprising Adamawa, Chadic, and Benue-Congo languages (Kleinewillinghöfer ). The rise of the Fulɓe Emirates (Gombe, Muri, Adamawa, Rey-Bouba) in the nineteenth century led to a lasting intense impact of Fulfulɗe and Hausa on many Adamawa minorities (see also Kleinewillinghöfer ). 3 Whether or not Kru is part of North Volta-Congo is irrelevant to this presentation. 4 Tula-Longuda combines the northwestern groups Tula-Waja and Longuda. Boa-Kula signifies the Bua group, and Chamba-Namshi comprises Samba-Duru, Mumuye, and Yandang.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
(Bennett : –). In conclusion, Bennett subdivides the Adamawa-Eastern part of the continuum into four branches: “Trans-Benue” (including “Tula-Longuda”), forming the link to the Gur branches, “Boa-Kula”, “Cameroun-Ubangi”, and “Gbaya”.5 CamerounUbangi combines the groups Chamba-Namshi, Mbum-Mangbai, and Ubangi, which Bennett views as forming a cluster (: ). The particular status of Tula—and Waja—with respect to their potentially linking Adamawa and Gur is examined in Kleinewillinghöfer (b). It is shown there, above all, that the correspondences of the noun class systems with those reconstructed for Central Gur are indeed substantial, corroborating the lexical data. The evidence clearly indicates common genetic descent. Recent research on Samba-Duru,6 however, revealed noun class systems which impressively correspond in terms of characteristic features—some of which are not found in Tula-Waja—with those of Central Gur groups. Supported by lexical correlates and typological analogies (see section . below), Samba-Duru evidently links with Central Gur as well. A thorough review of the “Adamawa-Gur” complex is required to determine whether or not the two Adamawa groups are more closely related to each other than either of them is to Central Gur (one of its subgroups or in general). Kleinewillinghöfer () shows that part of the cultural vocabulary of “proto-central” (Gur), which relates to a “civilisation néolithique agricole . . . dans un milieu savane boisée” (Manessy : ), i.e. names of trees, terms referring to landscape and subsistence such as ‘millet/guinea-corn farmers’, is shared by Tula-Waja, Samba-Duru, and other groups. Their common ancestors apparently inhabited a contiguous area in the savannah belt, probably situated in modern Nigeria (cf. Manessy : ) before the continuum became separated, possibly reinforced by the southward expansion of Chadic and the emergence of Savannah states. Although no study ever backed Greenberg’s Adamawa-Ubangi, the notion of its validity is repeatedly sustained in subsequent publications. Yet authors generally consent to a North Volta-Congo unit. The classification of Niger-Congo by Williamson and Blench displays a “Gur-Adamawa” continuum (comprising an Ubangi branch) in the tree diagram (: ),7 but also an Adamawa-Ubangi language family consisting of an Adamawa and an Ubangi branch (: ). Dimmendaal (: ), on the other hand, accepts that Gur and Adamawa form a subgroup of Niger-Congo but contests its relation to Ubangi, which he considers constitutes “an independent language family” unrelated to Niger-Congo (see also Miehe, chapter of this volume, and Pasch, chapter of this volume).
. A
.................................................................................................................................. Comprehensive accounts of the languages and groups are presented by Samarin () and Boyd (). Blench () offers an extensive language and dialect list; the Ethnologue
5 6 7
See Moñino () on the classification of Gbaya. Samba-Duru combines the Samba-Leko and Duru groups. See Blench () for an update.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Adamawa groups and isolated languages Group / isolated Alternative name(s) Greenberg language (1963)
(Approximate number of members), additional sources
Location1
Tula-Waja əna-Mboi Kam Bikwin-Jen Longuda Mumuye Yandang
Waja Yungur Nyiŋɔm, Nyingwom Burak-Jen2 Nʊngʊra
NI NI NI NI NI NI NI
Kwa* Samba-Duru
Baa (Chamba) Leko, Duru
(8) (6) (1) (7) (dialect cluster) (?) cf. Shimizu (1979) (?) cf. Kato et al. (n.d.) (1) ( 17) cf. Lode (1992) cluster oflects 3 (1) ( 16) cf. Boyd (1974); Elders (2006) (10) (3) (1) (1?) (1)
CAM CAM CAM, CH, CAR
Fali* Nimbari ?4 Kebi-Benue Bua Kim Day* Gueve-Duli* Chamba-Daka*
Yendang
Bari, Niam-niam Mbum, Lakka
A1 A7 A8 A9 A10 A5 A5 A2 & A4 A11 A12 A6 A13 A14
Gewe Daka, Dakoid
A3
NI NI, CAM
CH CH CH CAM NI
* Comments: Baa (Kwa) (Hansford et al. ) and Day (Nougayrol ) are isolated languages (re)identified as Adamawa (cf. Boyd ). Fali is regarded as misclassified in Adamawa (Boyd ). Blench () excludes it from the GurAdamawa(-Ubangi) unit. Despite the apparent poor lexical evidence, there are nevertheless morphological cognates which link Fali with Adamawa-Gur. Gewe (Gueve) and Duli, as reported in Baudelaire () and Strümpell (/), are names of two settlements where apparently variants of a single language were spoken. Both variants are probably extinct. Chamba-Daka is excluded from Adamawa and affiliated with Benue-Congo in Boyd () and Williamson and Blench (). Boyd () reclassifies it as a “peripheral group of Adamawa”. Laɓi, as listed in the Ethnologue, is a language of initiation (see Elders : –). 1 NI = Nigeria, CAM = Cameroon, CH = Chad, CAR = Central African Republic. 2 The Burak group (Hansford et al. ) is joined to Jen (Bennett ) and renamed “Bikwin” (Kleinewillinghöfer a). 3 See Kramer (: –) for the current state of the art. 4 See Kastenholz and Kleinewillinghöfer () and Ayotte and Ayotte (: –).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
(Lewis et al. ) provides current statistical information on (almost) all Adamawa languages, as well as language maps. Cameroonian languages are listed and mapped in Dieu and Renaud (), and Nigerian languages in Hansford et al. () and Crozier and Blench (). Greenberg’s numbering of Adamawa groups still serves as a widely used reference; it is, therefore, included in Table ..8
. C
.................................................................................................................................. Leaving aside the still undefined composition of an Adamawa-Gur unit, the question of how the Adamawa groups are interrelated among themselves is still largely unresolved. Up to now higher-level groupings beyond Greenberg’s classification have not been demonstrated by meticulous historical-comparative studies and appear to be largely based on counts of putative cognate words, randomly compounding items of horizontal and vertical (genetic) transfer. Geographical proximity appears to be decisively correlated to presumed states of relationship. The most comprehensive approaches are the lexicostatistical study by Bennett () as well as the classification of Boyd ().9 The latter seemingly also forms the basis for the subdivisions of Adamawa in Williamson and Blench (). A modified subgrouping of Adamawa is presented by Blench (). Boyd’s statement that a “core consists of the languages of group . . . which seem to have the greatest lexical affinities with the small isolated northern groups (Bua), (Kim) and Day” (: ) appears to be the essential published evidence supporting one of the major Adamawa divisions (see also Elders : ). The validity of a northwestern division, “Trans-Benue” in Bennett (), comprising Tula-Waja, Longuda, Bikwin-Jen, and Yungur, is certainly disputable (see also Boyd : ). While Tula-Waja morphologically clearly links to Central Gur, the relations to the other three groups, as well as their interrelations, are still unresolved (Kleinewillinghöfer a, b). As regards the noun class morphology, Longuda outwardly links with TulaWaja as well, but the evidence still lacks conclusiveness. Low rates of lexical cognates of Yungur (Ɓəna-Mboi) with the other three groups, its discrete set of personal pronouns (cf. Babaev ), and diverging noun class morphology suggest an affiliation of Ɓəna-Mboi to a separate branch. A central(/southwestern) division, including Samba-Duru, Mumuye, and Yendang, is, according to Boyd (: ), primarily supported by lexical correspondences.10 SambaDuru, in turn, strongly links up with Central Gur, most evidently through the impressive number of morphological cognates, reinforced by lexical evidence (Kleinewillinghöfer a, b). The gender systems of Bua languages share remarkable morphological 8
See also Kleinewillinghöfer (–) for additional information. Boyd () is apparently the basis of higher-level groupings of Adamawa languages in current editions of the Ethnologue. Updates on or revisions of Adamawa in recent editions are only implemented for individual languages/groups where their associates reported. Thus, out-dated and/or misleading information may be repeated, creating an air of correctness. The somewhat arbitrary designations used for still unproven higher-level groupings are, unfortunately, often quoted as established facts. 10 Cf. “Chamba-Namshi” in Bennett (). The extinct “Gueve-Duli” and “Nimbari” correspond most clearly with Samba-Duru and thus belong to this grouping as well. 9
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
features with those in Samba-Duru. Supportive lexical and typological correspondences indicate that Bua may link up as well with this unit.11 The lack of comparable data from most Adamawa groups (see section . below), however, largely hampers the diagnostic significance of class morphology in the attempt to firmly establish the structure of the wider Adamawa-Gur complex. What is particularly uncertain is also how the eastern groups (e.g. Kebi-Benue) relate to Ubangi. The vestigial noun class systems of the Mba languages (Pasch ), nevertheless, show some structural and morphological similarities with genders in Adamawa/Gur.
. L
.................................................................................................................................. The comprehensive studies by Samarin () and Boyd () address early research as well as the major phonological, morphological, and typological characteristics of Adamawa languages. As for their accounts of phonological characteristics, we may add that [ATR] vowel harmony systems also occur in Adamawa, namely in Tula-Waja, Longuda, Bikwin, and in Bua (Elders ; Kastenholz n.d.). Noun class morphology—in addition to lexical resemblances—is the very characteristic that Greenberg used to confirm Adamawa as a branch of Niger-Congo. It also provides the decisive evidence for establishing the internal structure of Gur (Manessy ). In Adamawa, however, class languages—where class membership is still marked by characteristic suffixes and where sets of pronouns and bound morphemes are used to mark grammatical agreement (“concord”)—amount to merely a quarter. They occur in only five groups: Tula-Waja (Kleinewillinghöfer b), Longuda (Jungraithmayr /), Yungur (Kleinewillinghöfer ), Samba-Duru (Kleinewillinghöfer n.d.), and Bua (Boyeldieu et al. ; Kastenholz n.d.). Notwithstanding the considerable diversity that exists between the noun class systems in the various groups and languages, in their fundamental features they all relate to VoltaCongo class morphology. Typologically, they correspond to Gur, i.e. class markers are generally suffixed. Partial prefixing is only known from Tula and Yungur. In both cases it occurs alongside the standard suffixing of class markers (Kleinewillinghöfer , b). Exclusive to Adamawa, however, are the morphological correspondences between SambaDuru class languages and Central Gur. Most remarkable are the noun classes and/or morphemes which agree in terms of form, semantics, and function with Proto-Central Gur *ni, *tʊ, *sɩ, *mʊ, *bʊ, and *da (cf. Manessy ; Miehe et al. ; Kleinewillinghöfer n.d.). Similarly remarkable morphological cognation with Central Gur was previously reported only from Tula-Waja (Kleinewillinghöfer b). What remains enigmatic is the origin of the locative class P in Waja. It is reminiscent of Bantu *pa but unknown in Adamawa-Gur (Kleinewillinghöfer b: ). Bua languages show varying degrees of erosion of a former more complex gender system with regular singular/plural pairings of nouns; partial/restricted concord marking is only 11 This hypothetical grouping coincides with a typological analogy. It combines Adamawa groups where possessors precede the possessed in associative constructions. This word order is the standard in Gur but not in other Adamawa groups.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
known from Kulaal (Boyeldieu ). The concord morphemes reported appear to be vestiges of the former class system. Most of them correspond to Samba-Duru noun classes. The majority of Adamawa languages, however, only show various stages of transition from class languages to a stage where (most/all) traces have vanished. Plural marking of nouns—the most obvious area where former noun classes manifest—has largely disappeared. In Kebi-Benue nominal plural marking occurs in one subgroup only (Elders ). From Kuasap (Kim), Nyiŋɔm (Kam), Day, and Fali of Mongo just a single morpheme is known which marks nominal plurals. Exceptional for Adamawa is plural marking with a prefix, as in Kwa and Bikwin-Jen (Kleinewillinghöfer a). These languages are spoken in the Benue valley, where contact with prefixing Jukunoid and Jarawan Bantu languages is/ was highly probable. Vestiges of class morphology may also manifest in derivations and regular syllable patterns correlating with semantic concepts (Shimizu ; Elders ; Anonby ; Kramer ); or, as demonstrated by Boyd () for “proto-lakka”, they may be the base of irregular correspondences of final consonants. Unrelated to former noun classes are “nominal pre-formatives”, bound morphemes preposed to noun stems. They are common in Kebi-Benue languages (Elders ); solitary cases are also known from Fali (Kramer ), Waja, and Tula.
.. Verb The lack of adequate data from about three-quarters of the languages does not permit a summarizing overview of the evidently very heterogeneous verbal systems occurring in Adamawa. Considerable deviations are reported even between closely related languages. Aspect marking seems to be general (see Nurse et al. n.d.), though highly multifaceted. Common to a number of languages is a contrast of two verb forms, distinguished by tone and/or segmental morphology, around which—by the use of suffixes, clitics, auxiliaries, reduplication of verb stems, and modifications of the tone patterns—“inflectional systems of varying complexity are constructed” (Boyd : ). Serial verb constructions are reported from several languages. Verb extensions, “verb suffixes which may increase valence”, “decrease valence”, “(re-)orient action”, “or mark aspect” (Hyman : ) are Niger-Congo heritage. Verb extensions varying in number and function are also found in Adamawa, though often they are non-productive.12 Boyd (: ) reconstructs as “widespread” in Adamawa “intensive”, “iterative”, “benefactive”, and “causative”.
.. Word order Within noun phrases the head noun generally precedes modifiers such as adjectives/adjectivetype words, demonstratives, numerals, and relative clauses. In most Adamawa languages the
12 Waja has the productive extensions “pluractional”, “passive-intansitive”, “instrumental”, “destinative”, and “ventive” (Kleinewillinghöfer b: ). Wiering and Wiering (: ) list fourteen distinct extensions for Doyayo (Samba-Duru).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
order of constituents in associative/possessive (or genitive) constructions follows the same pattern, i.e. the possessed (head noun) precedes the possessor (determining noun/pronoun). In one division of Adamawa (Samba-Duru, Mumuye, Yandang, Bua, and Kim), however, the possessor as a rule precedes (or /may precede) the possessed (= head). Several languages, particularly in Samba-Duru, distinguish nominal possessors from possessive pronouns. While the former precede the possessed noun, possessive pronouns generally follow. ()
Samba: Àbdú nà ‘la vache d’Abdou’ (Abdu’s cow) nà mə́ ‘ma vache’ (cow my)
(Fabre : )
A subset of these languages (Mumuye, Yandang, part of Samba-Duru) evidently varies in the order of constituents, i.e. the possessor may precede or follow the possessed noun. While no reason for the variance in Mumuye is reported (Shimizu ), in other languages it is correlated to the formal distinction of associative constructions based on their respective semantics. Variance in the word orders (co-occurring with morphological marking) distinguishes (i) alienable versus inalienable possession (in Yandang, Lɔŋto)13 and (ii) constructions which express possession in the narrow sense from constructions where the possessor rather modifies the possessed noun, referring to its origin, quality, etc. (in Beiya, Gəmme, Vəmnəm) (Kleinewillinghöfer ). ()
a. Beiya: gbilə̀ river b. máám water
máám ‘river’s water; water in the river’ water gbilə-m ‘river water; water from a river’ river-
In clauses, Adamawa languages generally display a basic SVO order of constituents. Additional S+AUX+O+V+X and/or S+P+O+V+X syntax is reported for languages belonging to the Samba-Duru and Bua groups (Elders , ; Kleinewillinghöfer ). A verb object negative word order is widespread14 where negation markers often appear in clause-final position. In several languages, final negation markers are merely part of a more complex negation marking (Kleinewillinghöfer and Littig ).
. C
.................................................................................................................................. Most (or all?) languages that Greenberg () classified as Adamawa are rather part of a larger Adamawa-Gur complex. The issue of its composition and how it relates to Ubangi and Gbaya is still unresolved. In the light of present evidence “Adamawa” may be maintained when understood as denoting “Adamawa-Gur” languages spoken in the Adamawa area.
13 This contrast also occurs elsewhere in Adamawa (Boyd : ). It is generally marked morphologically. 14 It is to be found across a large area in the central part of Africa (Güldemann ; Dryer ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
A I am grateful to Jörg Adelberger, Bruce Connell, Larry Hyman, and Raimund Kastenholz for comments on earlier versions, and to Lucinda Bowles for corrections of my usage of English.
R Anonby, E. J. (). Phonology and Morphology of Mambay (Niger-Congo, Adamawa). Ph.D. thesis. Leiden University. Ayotte, M., and Ayotte, C. (). Sociolinguistic Language Survey of Dama, Mono, Pam, Ndai, and Oblo. Dallas, TX: SIL International. Babaev, K. (). Niger-Congo person marking: testing kinship with morphology. Paper presented at the International Congress “Towards Proto Niger-Congo: Comparison and Reconstruction”, Paris, – September . Baudelaire, H. (). ‘La numération de à dans les dialectes Habé de Garoua, Guider, Poli et Rey Bouba’, Bulletin de la Société des Études Camerounaises : –. Bennett, P. R. (). ‘Adamawa-Eastern: problems and prospects’, in I. R. Dihoff (ed.), Current Approaches to African Linguistics . Dordrecht: Foris, –. Bennett, P. R., and Sterk, J. P. (). ‘South Central Niger-Congo: a reclassification’, Studies in African Linguistics : –. Blench, R. (). The Adamawa Languages, http://www.rogerblench.info/Language/Niger-Congo/ AU/Adamawa%language%list.pdf Blench, R. (). Niger-Congo: An alternative view, http://www.rogerblench.info/Language/NigerCongo/General/Niger-Congo%an%alternative%view.pdf Boyd, R. (). Étude comparative dans le groupe Adamawa. Paris: SELAF. Boyd, R. (). ‘Adamawa-Ubangi’, in J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham, New York, and London: SIL, –. Boyd, R. (). ‘The syntax and semantics of the Chamba-Daka verbal noun’, Afrika und Übersee : –. Boyeldieu, P., Kastenholz, R., Kleinewillinghöfer, U., and Lionnet, F. (). ‘The Bua Group languages (Chad, Adamawa ): A comparative perspective’, in R. Kramer and R. Kießling (eds.), Current Approaches to Adamawa and Gur Languages. Cologne: Rüdiger Köppe, –. Boyeldieu, P. (). Les langues du groupe boua (Adamawa de J.-H. Greenberg). Paper presented at the International Congress “Towards Proto Niger-Congo: Comparison and Reconstruction”, Paris, – September . Crozier, D. H., and Blench, R. M. (eds.) (). An Index of Nigerian Languages, nd edn. Dallas, TX: SIL. Dieu, M., and Renaud, P. (eds.) (). Atlas linguistique de l’Afrique Centrale: Atlas linguistique du Cameroun (ALCAM). Situation linguistique en Afrique Centrale. Inventaire préliminaire: le Cameroun. Paris: Agence de Coopération Culturelle et Technique; Yaoundé: Centre International de Recherche et de Documentation sur les Traditions et les Langues Africaines/Délégation Générale à la Recherche Scientifique et Technologique. Dimmendaal, G. J. (). ‘Language ecology and linguistic diversity on the African continent’, Language and Linguistics Compass /: –. Dryer, M. S. (). ‘Verb-object-negative order in Central Africa’, in N. Cyffer, E. Ebermann, and G. Ziegelmeyer (eds.), Negation Patterns in West African Languages and Beyond. Amsterdam and Philadelphia, PA: John Benjamins, –. Elders, S. (). A survey of distributed predicative syntax in Adamawa-Ubangi. Paper presented at the Workshop on “Distributed Predicative Syntax”, th World Congress of African Linguistics (WOCAL ), Rutgers University, New Brunswick, NJ, June –, .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Elders, S. (). ‘Distributed predicative syntax in Doyayo: constituent order alternations and cliticization’, in A. Akinlabi and O. Adesola (eds.), Proceedings of the th World Congress of African Linguistics. Rutgers University, New Brunswick, NJ. Köln: Rüdiger Köppe, –. Elders, S. (). ‘Issues in comparative Kebi-Benue’, Africana Linguistica : –. Fabre, A. G. (). Étude du samba leko, parler d’Allani (Cameroun du nord, famille Adamawa. Doctoral thesis. Université de Paris III. Greenberg, J. H. (). ‘Studies in African linguistic classification. I. The Niger-Congo family’, Southwestern Journal of Anthropology : –. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, –. Hansford, K., Bendor-Samuel, J., and Stanford, R. (). ‘An index of Nigerian languages’, Studies in Nigerian Languages . Hyman, L. M. (). ‘Niger-Congo verb extensions: overview and discussion’, in D. L. Payne and J. Peña (eds.), Selected Proceedings of the th Annual Conference on African Linguistics. Somerville, MA: Cascadilla Proceedings Project, –. Jungraithmayr, H. (/). ‘Class languages of Tangale-Waja District’, Afrika und Übersee : –. Kastenholz, R. (n.d.). Comparative list of Bua Group languages, Day, and Laal. MS. Kastenholz, R., and Kleinewillinghöfer, U. (). Nimbari as a Language Name, http://www.blogs. uni-mainz.de/fb-adamawa/files///Nimbari.pdf accessed July , . Kato, B., Yoder, Z., and Blench, R. (n.d.) The Maya [Yendang] Languages, http://www.rogerblench. info/Language/Niger-Congo/AU/Yandang%group/Comparatie%Yandang.pdf Kleinewillinghöfer, U. (). ‘Nominalklassen im Mboi und Ɓəna Lala der Yungur- bzw. ƁənaMboi-Gruppe’, Afrika und Übersee : –. Kleinewillinghöfer, U. (). ‘Don’t use the name of my dead father: A reason for lexical change in some Northwestern Adamawa languages (northeastern Nigeria)’, Afrika und Übersee : –. Kleinewillinghöfer, U. (a). ‘Die nordwestlichen Adamawa-Sprachen—eine Übersicht’, in U. Seibert (ed.), Afrikanische Sprachen zwischen Gestern und Morgen. (Frankfurter Afrikanistische Blätter : –.) Köln: Rüdiger Köppe. Kleinewillinghöfer, U. (b). ‘Relationship between Adamawa and Gur languages: the case of Waja and Tula’, Cahiers Voltaïques / Gur Papers : –. Kleinewillinghöfer, U. (). ‘Assoziative Konstruktionen im Vere-Gimme und Lɔŋto, Central Adamawa’, in R. Kramer, H. Tröbs, and R. Kastenholz (eds.), Afrikanische Sprachen im Fokus: Linguistische Beiträge zum . Afrikanistentag, Mainz, .–. April . Köln: Rüdiger Köppe, –. Kleinewillinghöfer, U. (–). Adamawa Language Groups, http://www.blogs.uni-mainz.de/fbadamawa/adamawa-languages/. Kleinewillinghöfer, U. (a). Correlations of the noun class systems of Central Adamawa and Proto-Central Gur. Paper presented at the “th International Conference on Gur Languages. Gur Internal and External Relationships”, Bayreuth, July –, . Kleinewillinghöfer, U. (b). Samba-Duru, a core group of Central Adamawa. Paper presented at the International Congress “Towards Proto Niger-Congo: Comparison and Reconstruction”, Paris, – September . Kleinewillinghöfer, U. (). ‘The northern fringe of the Jos Plateau, a prehistorical contact zone of Benue-Plateau and Adamawa-Gur languages: the evidence of the cultural vocabulary’, in R. Kramer and R. Kießling (eds.), Current Approaches to Adamawa and Gur Languages. Cologne: Rüdiger Köppe, –. Kleinewillinghöfer, U. (n.d.). ‘Noun class systems in Central Adamawa’. MS. Kleinewillinghöfer, U., and Littig, S. (). Negation patterns in Sama-Duru languages. Paper presented at the “th World Congress of African Linguistics (WOCAL )”, University of Buea, August –, .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
Köhler, O. (). ‘Geschichte und Probleme der Gliederung der Sprachen Afrikas: Von den Anfängen bis zur Gegenwart’, in H. Baumann (ed.), Die Völker Afrikas und ihre traditionellen Kulturen. Teil I: Allgemeiner Teil und südliches Afrika. Wiesbaden: Franz Steiner, –. Kramer, R. (). Die Sprache der Fali in Nordkamerun: Eine funktionale Beschreibung. Köln: Rüdiger Köppe. Lewis, M. P., Simons, G. F., and Fennig, C. D. (eds.) (). Ethnologue: Languages of the World, th edn. Dallas, TX: SIL International, http://www.ethnologue.com Lode, L. (). Aperçu des langues parlées dans le département de Faro. Poli, Cameroun. MS. Maley, J., and Vernet, R. (). ‘Peuples et évolution climatique en Afrique nord-tropicale, de la fin du Néolithique à l’aube de l’époque moderne’, Afriques, débats, méthodes et terrains d’histoire. Vol. . CEMAf-UMR , http://afriques.revues.org/ Manessy, G. (). Contribution à la classification généalogique des langues voltaïques. Paris: SELAF. Manessy, G. (). ‘Langues et histoire des peuples voltaïques: signification et limites de la comparaison historique’, Gur Papers / Cahiers Voltaïques : –. Miehe, G., Kleinewillinghöfer, U., von Roncador, M., and Winkelmann, K. (). ‘Overview of noun classes in Gur II (revised and enlarged version)’, in G. Miehe, B. Reineke, K. Winkelmann (eds.), Noun Class Systems in Gur Languages. Vol. II: North Central Gur Languages. Köln: Rüdiger Köppe, –. Moñino, Y. (). The position of Gbaya-Manza-Ngbaka group among the Niger-Congo languages. Paper presented at the Workshop on “Genealogical Classification in Africa beyond Greenberg”, Berlin, Humboldt University, February –, . Nougayrol, P. (). Le day de Bouna (Tchad). I: Phonologie, syntagmatique nominale, synthématique. Paris: SELAF. Nurse, D., Rose, S., and Hewson, J. (in collaboration with C. Beaudoin-Lietz) (n.d.) Verbal Categories in Niger-Congo Languages, https://www.mun.ca/linguistics/more/e-books/NigerCongo_Book_Introduction.pdf Pasch, H. (). Die Mba-Sprachen. Hamburg: Helmut Buske. Samarin, W. J. (). ‘Adamawa-Eastern’, in T. A. Sebeok (ed.), Linguistics in Sub-Saharan Africa. (Current Trends in Linguistics, .) The Hague: Mouton, –. Shimizu, K. (). A Comparative Study of the Mumuye Dialects (Nigeria). Berlin: Dietrich Reimer. Shimizu, K. (). The Zing Dialect of Mumuye: A descriptive grammar. Hamburg: Helmut Buske. Strümpell, F. (/). ‘Wörterverzeichnis der Heidensprachen des Mandaragebirges’, Zeitschrift für Eingeborenensprachen : –, –. Webster, J. B. (). Kwararafa: the traditional face of the coin. Paper prepared for Boston University Center for African Studies Conference, April –, . Wiering, E., and Wiering, M. (). The Doyayo Language: Selected studies. Dallas, TX: SIL. Williamson, K. (). ‘Niger-Congo overview’, in J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham, MD: University Press of America, –. Williamson, K., and Blench, R. (). ‘Niger-Congo’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T Ubangi languages are spoken primarily in the Central African Republic (CAR), but also in the southeast of South Sudan, the north of the Democratic Republic of Congo (DRC) and the Republic of Congo, in East Cameroon, and in two pockets in Chad (see Map .). Delafosse and Caquot (: ) were the first to name this language family “groupe oubanguien” (after the river on whose banks most of these languages are spoken), a designation manifested by Samarin in . According to Ethnologue, languages belong to this family (Lewis ). The languages with the largest numbers of speakers are Banda, Ngbaka, and Zande, each of which has approximately one million speakers or more, Ngbandi, with some , speakers, and Gbaya, with more than ,. Mention should also be made of Sango, the national official medium of CAR, which by developed out of a stabilized pidgin (Samarin : ) and later on creolized. Its lexifier language is Ngbandi. Sango is spoken by more than three million people.
. L
.................................................................................................................................. The first descriptions and wordlists of Ubangi languages were produced by early travelers or, rather, the persons who edited their recordings for publication, e.g. Amãlo by Junker (); Müller (); and Czekanowski (: –); Ndungale by Thonner (); Mbane by Stuhlmann (); and Zande by Schweinfurth () and Czekanowski (: –). More comprehensive descriptions and dictionaries as well as classifications were presented by Christian missionaries, e.g. Zande by Lagae (); Lagae and Vanden Plas (, ); and Gore (); Ngbandi by Lekens (, , ); Banda by Tisserant (, ); Zande and Banda by Santandrea (); Mbane by Carrington (); and Gbaya by Hilberth (, ). Professional linguistic research commenced with the Bantu border
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Ubangi languages
line expedition in the Belgian Congo (–) for the International African Institute in London, supervised by A. N. Tucker. Delafosse and Caquot’s “groupe oubanguien”, with members attached to their large “négro-africaine” family on the basis of typological and morphological criteria, corresponds less to Greenberg’s () classification of Ubangian languages than has occasionally been claimed (see, e.g., Moñino a: ). It also includes Kresh and Mittu, which Greenberg (: , ) clearly allocated to Central Sudanic (II.E..). Delafosse and Caquot’s Ubangian languages are Zande plus (the dialects) Nzakara and Barambo, Mayogo–Mundu, Sere plus Ndakko and Golo, Ngbandi plus Dendi and Gbwaga, Banda, Gbanziri, Monjombo, Manja, Gbaya, Ngbaka–Ma’bo plus Goundi and Bomassa, and Gbaya. The “groupe oubanguien” is rather close to Müller’s “äquatoriale Sprachfamilie” (, ) which includes Central Sudanic Kresh and Mangbetu besides the Ubangi languages Zande, Barambo, Amãlo (language of the “A-Madi”), Mayogo–Mundu (“Maigo–Mungo” in Müller’s spelling), and Golo, a variant of Sere. Because of typological resemblances with the Kwa-Kru languages, Westermann (: ) saw a relationship to the “Nigritic” section of the Sudan languages, and Köhler (: ) characterized the Ubangi languages as “Kwa of the East”. He also saw a resemblance between the plural affixes o- in Ngbaka-Gbaya and -wó in Ewe. Westermann’s (: –) “Nigritic” group is larger than that of Delafosse and Caquot and comprises besides the Ubangi languages not only languages of the Adamawa and Central Sudanic families but also Kunama, Barea, Fur, and Nubian. Next to these, Westermann listed the following Ubangi language groups, again without hierarchical order: Ndogo–Sere– Belanda and other languages, Babuckur–Mundu, Banda, Mbwaka–Limba–Monjombo, Manza–Gbaya, Bangba–Zande (with the dialects Nzakara and Barambu) and other languages, Banda (with the dialects Mbanza, Togbo, Ndi, Manza–Gbea, Linda, Gobu),
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Ngbandi–Sango, and Ngbaka–Ma’bo–Monzombo. He was uncertain whether Gbanziri should be grouped with the last or with the penultimate group. Early attempts to classify the Ubangi languages were based on information about a few languages only, which makes classification difficult. This is well reflected in the classification of the Mba languages (Pasch : –). The close relationship between Mbane and Ndungale became apparent when the two were “discovered” one after the other by travelers coming from the west and were classified by van Bulck () as “old Nigritic Sudanic languages”. On the other hand Amãlo, located further northeast, was “discovered” by travelers coming from the east and was allocated by van Bulck to a different group: the “new Sudanic languages”. The lack of comprehensive data should be responsible for the fact that until today Ubangi languages have usually been given just a few lines in many monographs dealing with the classification of African languages (see, e.g., Heine and Nurse , ). Only Delafosse and Caquot (: –) and Köhler (: –) devoted several pages to this group. Systematic comparative research on Ubangi languages began as late as . In this year, Luc Bouquiaux and Jacqueline Thomas started a project with a group of scholars working on different Ubangi languages, in order to carry out comparative work in cooperation (Moñino a: ). Discontented with classifications that were based exclusively on lexical evidence, they also wanted to take into consideration sociolinguistic factors such as intermarriage patterns, which would allow them to find mixed languages. Greenberg’s (: ) classification, which was the first to rely solely on resemblances in sound and meaning, is as follows: . . . . . . . .
Gbaya, Manza, Ngbaka Banda Ngbandi, Sango, Yakoma Zande, Nzakara, Barambo, Pambia Bwaka (Ngbaka–Ma’bo), Monzombo, Gbanziri, Mundu, Mayogo, Banga Ndogo, Bai, Bvori, Golo, Sere, Tagbo, Feroge, Indri, Mangaya, Togoyo Amãlo Ndungale, Mbane
Since then the subgrouping has considerably been modified. To begin with, Bouquiaux and Thomas proposed reclassifying Greenberg’s groups “” and “” as the third subdivided section of “” (Samarin : ), hence: . Gbaya–Ngbandi–Monzombo–Ndogo (a) Gbaya, Manza, Ngbaka (b) Ngbandi, Sango, Yakoma (c) (i) Ngbaka–Ma’bo, Monzombo, Gbanziri, Mundu, Mayogo, Bangba (ii) Ndogo, Bai, Bviri, Golo, Sere, Tagbo, Feroge, Indri, Mangaya, Togoyo Because of their geographical distance Samarin considered the inclusion of Greenberg’s group and “c.ii” next to “c.i” in group “” as “unlikely” since there is an enormous distance
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
between the areas of distribution of groups c.i and c.ii. Furthermore, he prefered the geographical names for the groups: . Western, . Central, . Southern, . Southeastern, . Southcentral. On the basis of Moñino’s () lexical comparisons Boyd (: ) proposed the following subgroups: . Gbaya . (A) Banda (B) Ngbandi (C) (I) Sere (II) a. Ngbaka b. Mba . Zande The subgroups have also been subject to regroupings. Pasch () classified the Mba languages as follows: .a.i Mbane, .a.ii Ndungale, .b. ‘Dongoko, . Amãlo. Boyd excluded Kpatiri from the Zande group (cf. Tucker and Bryan ) and allocated it to the Ngbandi group; moreover, with Geme he discovered a new dialect of Zande (Boyd a, b). Baka, which for a long time was considered a variant of Ngbaka–Ma’bo, is now deemed an Abstand language (Kilian-Hatz : ). It is the only Ubangian Pygmy language.
. G
.................................................................................................................................. Five- or seven-vowel systems predominate. There are restrictions on the sequences of vowels in a word; several languages require identical vowels in disyllabic words or common features such as back/front or open/closed. Three of the Mba languages mark plural by umlaut (Pasch ). Apart from Gbaya, which allows word-final consonants, Ubangi languages have CV(CV) word structure. The inventory of initial and non-initial consonants in most Ubangi languages is similar. Exceptions are Sere, which has reduced its intervocalic inventory, and Gbaya, which has apparently adopted the Adamawa pattern. Word order is usually SVO, adjectives and other attributive words precede the noun, and in possessive constructions, the possessed typically precedes the possessor; prepositions rather than postpositions are used. With regard to other features a considerable divergence is observed. While the plural prefix a- of Banda, Ngbandi, and Zande is widespread in the area, even beyond Ubangi languages, it is nda- in Ndogo-Sere, ka- in Bviri, o- in Ngbaka-Gbaya, and wu- in Mundu. Zande and three of the Mba languages have pronominal genders, which gives them a special position within Ubangi. The Mba languages Mba, Ndunga, ‘Dongo, and Ma have noun classes, the first three with and the last without agreement, which makes their position even more noteworthy, and so does the system of possessive classifiers that ‘Dongoko has developed (Pasch ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Köhler () considered the coexistence of these two class systems as unique in Africa. He considered the gender system more recent than the noun class system; however, the interrelatedness of the two systems makes it difficult to decide which one is older. The Mba noun class system has been used as an argument in support of the allocation of the Ubangi family to the Niger-Congo phylum (Greenberg : ), although the relationship to other nouns class systems has never been proved. Furthermore, there is little evidence of a prior noun class system in the other languages of the family. While the genealogical classification of the Ubangi languages is difficult, there are no Ubangian features, but a number of lexical, morphological, and phonological features are also found in languages of other families spoken in the same or adjacent areas. On the basis of these features Westermann () developed the hypothesis of a Sudanic family, and Gregersen () and Blench (, ) even postulated that Niger-Kordofanian and Nilo-Saharan be united in a “Kongo-Saharan” or “Niger-Saharan” respectively, a macrophylum which Güldemann () called NKNS. Güldemann’s NKNS confirmed the findings of Bender (: ) and Dimmendaal (: –), who attempted to reconstruct logophoric markers for the proto-language of this group. Boyd () identified a substantial number of resemblances between the two units, to the extent that ultimately he found it difficult to determine whether a given Adamawa-Ubangian language belonged to Niger-Congo or to Nilo-Saharan. Cloarec-Heiss () has shown, however, that the many properties which Banda shares with Central Sudanic languages may well be the result of convergence between two languages. Using the model of Thomason and Kaufman (), she claimed that in a situation of intensive contact speakers of a Central Sudanic language gave up their language, of which they retained a number of phonetic-phonological and morphosyntactic features when they adopted—not quite perfectly—an early version of present-day Banda. On the basis of word order features Heine (: ) grouped Gbaya, Sango, Banda, Zande, Ndogo, Bai Bviri, Tagbu, and Sere, together with some Adamawa languages as well as Efik, Bamileke, and Hausa, as the “ ‘Banda subtype’ of A languages”. He observed that most languages of this subtype belong to the Adamawa-Ubangi group and are located in a geographically definable area north of the Congo River. While he was able to demonstrate the diffusion of certain areal features from a nucleus into the periphery in other regions of Africa, he did not do so with regard to the area under consideration here. Güldemann (: ) considered the mere presence of the specific structural linguistic features insufficient evidence for a common origin as long as borrowing cannot be ruled out. He argued for “a basically areal hypothesis” and mapped the geographic distribution of six typological features: () labiovelar plosives, () labiodental flaps, () [ATR] vowel harmony, () logophoricity, () word order V–O–NEG, and () word order S–(AUX–) O–V–X. The areas of distribution of features () through () cover the area where Ubangi languages are spoken, an “areal hotbed” (Güldemann : ). They are, however, shared by four adjacent language groups: Benue-Congo, Adamawa, Bongo Bagirmi, and Moru Mangbetu (Güldemann : –). The distribution of the features as indicated on Map . must, however, be interpreted with care. While labiovelar plosives are found in all Ubangi languages, the labiodental flap is practically absent in the Zande group. Olson and Hajek (: ) know of only one example from Nzakara, documented by Santandrea (: ). [ATR] vowel
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
harmony, which is well attested in neighboring languages, is not attested in the Ubangi languages, though, with the exception of Zande (Boyd : ; Landi to appear). Logophoricity apparently occurs throughout Ubangi, and so does clause-final negation, but not all languages have discontinuous negation marking. The origin of the clause-final negation marker in Sango might derive from Bantu (Pasch ).
R Bender, M. L. (). ‘Nilo-Saharan’, in B. Heine and D. Nurse (eds.), –. Blench, R. (). ‘Is Niger-Congo simply a branch of Nilo-Saharan?’, in R. Nicolaï and F. Rottland (eds.), Proceedings of the Fifth Nilo-Saharan Linguistics Colloquium, Nice, – August, . Cologne: Rüdiger Köppe, –. Blench, R. (). The Niger-Saharan Macrophylum, http://www.rogerblench.info/Language/NiloSaharan/General/Niger-Saharan%book.pdf accessed July , . Boyd, R. (). ‘Les langues adamawa’, in D. Barreteau (ed.), Inventaire des études linguistiques sur les pays d’Afrique noire d’expression française et sur Madagascar. Paris: Conseil international de la langue française, –. Boyd, R. (a). ‘Le kpatiri ou gbāyī, une nouvelle langue du groupe ngbandi’, in Y. Moñino (ed.), –. Boyd, R. (b). ‘Le geme ou jὲmέ, une nouvelle langue du groupe zande’, in Y. Moñino (ed.), –. Boyd, R. (). ‘Adamawa-Ubangi’, in J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham, MD: University of America, –. van Bulck, V. (). Existe-t-il un groupe de langues soudanaises équatoriales? Paper presented at the “Troisième congrès international des sciences éthnologiques et anthropologiques”, Brussels, August, . Carrington, J. F. (). ‘Esquisse de la langue mba (kimanga)’, Kongo-Overzee : –. Cloarec-Heiss, F. (). ‘Emprunts ou substrat? Analyse des convergences entre le groupe banda et les langues du Soudan central’, in R. Nicolaï and F. Rottland (eds.), Proceedings of the Fifth NiloSaharan Linguistics Colloquium, Nice, – August, . Cologne: Rüdiger Köppe, –. Czekanowski, J. (). ‘Ethnographie-Anthropologie’, in J. Czekanowski (ed.), Wissenschaftliche Ergebnisse der Deutschen Zentralafrika Expedition – unter Führung Adolph Friedrichs, Herzog zu Mecklenburg. Vol. , part . Leipzig: Klinkhard und Biermann, –. Delafosse, M., and Caquot, A. ( []). ‘Les langues du Soudan et de la Guinée’, in A. Meillet and M. Cohen (eds.), Les langues du monde. Paris: Librairie Ancienne Honoré Champion, –. [: –.] Dimmendaal, G. J. (). ‘Logophoric marking and represented speech in African languages as evidential hedging strategies’, Australian Journal of Linguistics : –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Gore, C. E.C. (). A Grammar of Zande. London: Sheldon Press. Greenberg, J. H. ( []). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Gregersen, E. (). ‘Kongo-Saharan’, Journal of African Languages : –. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), –. Heine, B. (). A Typology of African Languages, based on the order of meaningful elements. Berlin: Dietrich Reimer. Heine, B., and Nurse, D. (eds.) (). African Languages: An introduction. Cambridge: Cambridge University Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Heine, B., and Nurse, D. (eds.) (). A Linguistic Geography of Africa. Cambridge: Cambridge University Press. Hilberth, J. (). Vocabulaire français-gbaya: Vocabulaire gbaya-français. Örebro and Berberati: Mission Baptiste Suédoise. Hilberth, J. (). Notes sur la grammaire gbaya. Örebro and Berberati: Mission Baptiste Suédoise. Junker, W. (). ‘Verzeichnis von Wörtern Zentralafrikanischer Sprachen’, Zeitschrift für Afrikanische Sprachen : –. Kilian-Hatz, C. (). Das Baka: Gründzüge einer Grammatik aus der Grammatikalisierungsperspektive. Cologne: Institut für Afrikanistik, Universität zu Köln. Köhler, O. (). ‘Geschichte und Probleme der Gliederung der Sprachen Afrikas: Von den Anfängen bis zur Gegenwart’, in H. Baumann (ed.), Die Völker Afrikas und ihre traditionellen Kulturen. Teil : Allgemeiner Teil und südliches Afrika. Wiesbaden: Franz Steiner, –. Lagae, C. R. (). La langue des Azande. Vol. : Grammaire, exercices, légendes. Éditions Dominicaines VERITAS. Lagae, C. R., and Vanden Plas, V. H. (). La langue des Azande. Vol. : Dictionnaire français-zande. Ghent: Éditions Dominicaines VERITAS. Lagae, C. R., and Vanden Plas, V. H. (). La langue des Azande. Vol. : Dictionnaire zande-français. Ghent: Éditions Dominicaines VERITAS. Landi, G. (to appear). Phonologie du zande. Ph.D. dissertation. Universität zu Köln. Lekens, B. (). Spraakkunst der Ngbanditaal (Grammaire ngbandi). Bruges: K. Beyaert & C. Houdmont-Cortoriendt. Lekens, B. (). Ngbandi-idioticon. I: Nederlands-frans en ngbandi. Tervuren: Commissie voor Afrikaanse Taalkunde. Lekens, B. (). Ngbandi-idioticon. II: Ngbandi en frans-nederlands. Tervuren: Commissie voor Afrikaanse Taalkunde. Lewis, M. P. (ed.) (). Ethnologue: Languages of the world, th edn. Dallas, TX: SIL International, http://www.ethnologue.com/ accessed August , . Moñino, Y. (a). ‘Introduction: cousines ou voisines?’, in Y. Moñino (ed.), –. Moñino, Y. (ed.) (b). Lexique comparatif des langues oubanguiennes. Paris: Paul Geuthner. Müller, F. (). ‘Die äquatoriale Sprachfamilie in Central-Afrika’, in Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften : –. Vienna: F. Tempsky. Müller, F. (). ‘Nachträge zur Abhandlung: Die äquatoriale Sprachfamilie in Central-Afrika’, in Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften : –. Vienna: Carl Gerold’s Sohn. Olson, K. S., and Hajek, J. (). ‘A crosslinguistic lexicon of the labial flap’, Linguistic Discovery : –, http://linguistic-discovery.dartmouth.edu/cgi-bin/WebObjects/Journals.woa//xmlpage// article/ accessed November , . Pasch, H. (). ‘Possession and possessive classifiers in “Dongo-ko”, Afrika und Übersee : –. Pasch, H. (). Die Mba-Sprachen. Hamburg: Helmut Buske. Pasch, H. (). ‘Sango’, in S. G. Thomason (ed.), Contact Languages: A wider perspective. Amsterdam and Philadelphia, PA: John Benjamins, –. Samarin, W. (). ‘Adamawa-Eastern languages’, in T. A. Sebeok (ed.), Current Trends in Linguistics. Vol. : Linguistics in Sub-Saharan Africa. The Hague: Mouton, –. Santandrea, S. (). Languages of the Banda and Zande Groups: A contribution to a comparative study. Naples: Istituto Universitario Orientale. Schweinfurth, G. (). Im Herzen von Afrika: Reisen und Entdeckungen im centralen ÄquatorialAfrika während der Jahre bis . vols. Leipzig: F. A. Brockhaus. Stuhlmann, F. (). Die Tagebücher von Dr. Emin Pascha. Vol. . Hamburg: G. Westermann. Thomason, S. G., and Kaufman, T. (). Language Contact, Creolization, and Genetic Linguistics. Berkeley, CA: University of California Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Thonner, F. (). Vom Kongo zum Ubangi: Meine zweite Reise in Mittelafrika. Berlin: Dietrich Reimer. Tisserant, C. (). Essai sur la grammaire banda. Paris: Institut d’Ethnologie. Tisserant, C. (). Dictionnaire banda-français. Paris: Institut d’Ethnologie. Tucker, A. N., and Bryan, M. A. (). The Non-Bantu Languages of North-Eastern Africa. London: Oxford University Press for the International African Institute. Westermann, D. (). Die Sudansprachen: Eine sprachvergleichende Studie. Hamburg: L. Friederichsen. Westermann, D. (). ‘Sprache und Erziehung’, in H. Baumann, R. Thurnwald, and D. Westermann (eds.), Völkerkunde von Afrika. Essen: Essener Verlagsanstalt, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. U the secession of South Sudan in , Kordofan was the most central state of the Sudan. This vast area located in the very middle of the Sahelian belt has given its name to a peculiar language group, Kordofanian, whose members are all endogenous to Kordofan, more specifically to South Kordofan, a rocky area also called the Nuba Mountains. The Nuba are the indigenous inhabitants of South Kordofan: they belong to many cultural groups and speech communities, and their languages fit into two unrelated language families (Quint a), namely Nilo-Saharan and Kordofanian, itself a branch of NigerCongo. Basically, Nilo-Saharan languages occupy the Western Nuba Mountains (or Jebels), while Kordofanian languages are mainly spread across the Eastern Jebels. In this chapter, I shall endeavor to introduce the reader to the Kordofanian languages. First, I will give a brief overview of those tongues. Secondly, I will deal with Kordofanian studies, insisting in particular on the difficulties of the fieldwork and the successive stages of the development of research in that area. Thirdly, I will present some salient features of Kordofanian languages and discuss the available hypotheses about the phylogenetic relationships between the different Kordofanian sub-branches and between Kordofanian and other Niger-Congo languages. Finally, I will conclude with the challenges that await the scholarly community regarding Kordofanian languages.
. K :
.................................................................................................................................. There are roughly two dozen Kordofanian languages, split into five main families (see Map .): (a) Heibanian (ten languages): Heiban, Ko, Koalib, Laro, Lukha (= Logol), Moro, Otoro, Shwai (= Shirumba), Tira, and Werni (= Warnang). Those languages are spoken in the northern and central parts of the Eastern Jebels and get their name from the city of Heiban, which is situated among these mountains. The area occupied by the
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. The geographical distribution of Kordofanian languages
Heibanian languages is mostly continuous1 and, on the ground, the linguistic boundaries between the main recognized linguistic communities are far from being precise. For instance, at the boundary between Koalib and Heiban speech areas, some villages have a mixed population and the local varieties of each Koalib and Heiban language display more features in common with the neighboring language.2 Heibanian languages have many more speakers than any other Kordofanian family: at least two Heibanian languages (Koalib and Moro) have more than , speakers, while two others (Otoro and Tira) have more than ,. (b) Talodian (nine languages): Acheron, Dagik (Masakin, Dengebu), Lumun, Nding (Eliri), Ngile (Daloka), Tasomi (Jomang, Talodi), Tocho, Tolona, and Tuwal (Norton and Kuku Alaki ). All Talodian languages are spoken in the vicinity of the city of Talodi (hence their name), in the southern and southeastern parts of the Nuba Mountains. Their total number of speakers amounts to approximately ,. (c) Lafofa (three languages or variants): El-Amira, Lafofa, and Tegem. The speech communities (numbering between , and , speakers) which make up this 1 With the exception of the easternmost members of the family, namely Ko and Werni, which are clearly outliers. 2 As a matter of fact, in the present state of knowledge about Kordofanian languages, their exact number cannot be stated. In large linguistic areas (like Koalib or Moro) with dialectal continuums, it is not always easy to decide on linguistic grounds only whether we are faced with closely related languages or varieties of one and the same language.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
small group live close to Talodian-speaking communities, with whom they seem to share some linguistic features. (d) Rashadian (two dialect clusters): Tagoi and Tegali, which are spoken in two mountainous ranges located in the northeastern parts of the Nuba Mountains. Tegali used to be the dominant language of the kingdom of Tegali (Ewald ) until the time of the Mahdiya,3 when the kindgom was devastated by the Mahdists and most of its population taken away to the Nile Valley in slavery (Elles ; Ewald ). At the end of the Mahdiya, when the Tegalians came back to their mountains, many of them had switched to Arabic as their vernacular. Be that as it may, there are still native speakers of Tegali and Tagoi, although they probably represent only a minority of today’s local population. Note that Tagoi varieties have a noun class system, while closely related Tegali varieties lack this system altogether. (e) Katla-Tima or Katloid (two languages): these two communities (Katla and Tima) live traditionally in the northwestern part of the Nuba Mountains. They are surrounded by Nilo-Saharan-speaking people and geographically separated from the remaining Kordofanian languages. Internally, each of these families shows a high degree of similarity which can be compared to the relationships existing between the members of the Romance family or between those of the Germanic family. However, the exact nature of the relationship between these families is a much more controversial issue, always actively discussed by the few specialists of Kordofanian (Schadeberg ; Hammarström ; Blench , , and forthcoming; Dimmendaal , ).
. K
..................................................................................................................................
.. The challenge of studying Kordofanian languages Kordofanian languages remain poorly known and described mainly because of the difficulty in accessing the field. As a matter of fact, since the nineteenth century (i.e. the beginning of Kordofanian studies: see section ... below), the Nuba Mountains have faced various periods of war and civil unrest (Turkish-Sudanese wars, Mahdiya, several local uprisings against British rule, second Sudanese civil war, ongoing conflict between the government and the Sudan People’s Liberation Movement-North [SPLM-N] since ) which have
3 The Mahdiya (–) was a crucial period in the recent history of Sudan. After several decades of Turkish-Egyptian rule over Sudanese territory, the Sudanese rose against their foreign occupiers under the leadership of Muhammad Ahmad, a Dongolawi from northern Sudan, who proclaimed himself the Mahdi, i.e. the ultimate Prophet of Islam. The upheaval gained momentum and soon the Mahdist rebels were able to overcome large Egyptian forces until they finally took Khartoum in January . Afterwards, a Mahdist state was established which would wage almost permanent war against foreign powers and local disobedient communities (such as the Tegalians and many other Nuba tribes). The Mahdist state was finally destroyed by a British expedition in .
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
considerably hindered scholarly research into the languages spoken there (including Kordofanian languages).
.. The main types of approaches ... First testimonies: before World War I There are no data available for Kordofanian languages before the nineteenth century, when some travelers (e.g. Rüppell ) began to compile wordlists of the various languages spoken by the people they came across during their journeys. Following these first testimonies come the works of various scholars (e.g. Tutshek , ; Meinhof / and –; Seligmann /) who endeavored to understand the structure of several languages spoken in the Nuba Mountains. Meinhof (/) himself laid stress on the study of a Kordofanian language, Heiban (Heibanian family).
... The missionaries’ linguistic work At the beginning of the twentieth century, after the Mahdiya, some English-speaking Christian missionaries settled in several localities in the Nuba Mountains and began to produce religious material in the vernacular languages in order to propagate their faith. This resulted in the publication of some portions of the Bible (generally the New Testament) in several Kordofanian languages, namely Koalib, Otoro, Heiban, and Moro. In several cases, those missionaries also produced some teaching material, including textbooks (e.g. Anonymous- and - for Koalib) and grammars (Black and Black on Moro). Among those missionaries, Stevenson occupies a special place: he devoted a considerable amount of time to the scientific study of the Nuba languages, wrote grammars of at least two Kordofanian languages (Otoro () and Tira (), both re-edited by Schadeberg in ), and produced and published a monumental Ph.D. thesis about the Nuba Mountain languages (/) in which almost all of these languages are mentioned and partially documented (through wordlists and paradigms). Until today, the missionaries’ works and publications constitute indeed the main available source for many a language spoken in South Kordofan, including various Kordofanian tongues. However, the missionaries’ approach to Nuba languages suffered from a solid bias: in general this approach was quite logically guided by religious considerations, which led the missionaries to study in priority the languages of those people who had been less touched by the spread of Islam and who were liable to be more easily converted to Christianity. This is why several Heibanian languages were studied and written down by the missionaries, as most Heibanian-speaking people at that time were mainly animists and as the relatively large number of speakers of several communities (see section . above) justified the effort in translating the Bible into their languages. Smaller communities (such as the Talodian people) or Muslim groups (such as the Rashadians), whose conversion to Christianity was probably deemed harder to achieve, were therefore generally neglected, and their languages left undocumented by the missionaries.
... The surveys Some scholars endeavored to classify the dozens of tongues spoken in the Nuba Mountains, including the Kordofanian groupings. Several surveys were launched in order to check
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
exactly (or approximately) how many languages were spoken in the region and how many groups they could fit into. In /, the MacDiarmids (a missionary couple) conducted such a survey (published in ) and, collecting the translation of the same wordlist in diverse local languages, were able to identify several language families, including the main families of Kordofanian (“Kawalib”, i.e. Heibanian; “Katla”; “Lafofa”; “Talodi-Masakin”, i.e. Talodian; and “Tegali”, i.e. Rashadian): basically, the genetic groupings proposed by the MacDiarmids still remain valid today. The second significant survey was carried out by Schadeberg, a trained linguist, in the s. Schadeberg had his informants translating Swadesh’s -word list plus the numbers from to and he focused on Kordofanian and Kadu4 languages. Schadeberg’s lists and analyses have been published for Heibanian (a), and Talodian (b), and they remain to date the most authoritative comparative work on those languages.5 Schadeberg published at a later date () his Rashadian raw lexical data, together with some short grammatical observations. In an unpublished work (c.), he had also elaborated some reconstructed forms and sounds for Proto-Rashadian: these reconstructions are mentioned in the present chapter.
... The study of individual languages More recently, some trained linguists have undertaken the description of several individual languages belonging to the Kordofanian grouping. This scholarly enterprise began in and is now fully developing due to two different trends: (i) the growing interest on the part of the scientific community and civil society in documenting and describing understudied languages (in order to preserve linguistic diversity), which helps researchers secure funding for such studies; (ii) a growing commitment of Sudanese scholars and speakers of Kordofanian languages to the study of those languages. This new dynamics has notably resulted in (i) the recent publication of a sizable number of volumes and papers dedicated to several Kordofanian languages from diverse families: (a) Heibanian: Heiban (Schadeberg and Kossmann ), Koalib (e.g. Quint b, , a, b, , forthcoming), Moro (e.g. Jenks and Rose ; Rose et al. ); (b) Katloid: Katla (e.g. Hellwig ), and Tima (e.g. Alamin a, b; Schneider-Blum ; Dimmendaal ; Bashir ); (c) Talodian: Dagik (e.g. Vanderelst ), Lumun (e.g. Smits ); (d) Rashadian: Tagoi (e.g. Bashir ); (ii) the organization of a Nuba Mountain Languages Conference every three years since in which several dozen scholars (among whom many specialists in Kordofanian languages) participate on a regular basis.
4
The Kadu (or Kaduglian) languages seem to belong to the Nilo-Saharan phylum. However, they have a noun class system, which might have been borrowed through contact from the neighboring Kordofanian communities (see below and also Quint : –; b: ; Schadeberg c, ). 5 Regarding Talodian, mention should also be made of Norton and Kuku Alaki’s recent study ().
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Thus, Kordofanian studies, after two decades of near interruption (between Schadeberg’s comparative publications in the early s and the beginning of the twenty-first century), are now developing at an ever accelerating pace, and it is probable that the coming years will see many more publications devoted to Kordofanian languages than have been hitherto published.
. S K
..................................................................................................................................
.. Phonology Three features at least are worth mentioning about Kordofanian languages: (i) probably all Kordofanian languages are tone languages (with generally two tone levels, high (H) and low (L)). In general, tone seems to have a higher functional load in morphology () than in contrasting lexical pairs (). ()
Morphological tone contrasts Koalib kwìcì LL vs. kwícì HL human.being: human.being: Koalib kwèny-ὲεcέ. LLH vs. kwény-ὲεcέ. HLH ‘I will see him/her.’ ‘We will see him/her.’ vs. kwény-έεcέ HHH ‘S/he will see me.’
()
Lexical tone contrasts Tima kʊ̀dá LH vs. kʊ̀dà LL vs. kʊ́dà ‘tree sp.’ ‘shoe’ ‘python’ Koalib ŋwɔ́ny H vs. ŋwɔ̀ny L ‘saliva’ ‘eggs’
HL (Schneider-Blum :)
(ii) most (if not all) Kordofanian languages also display phenomena of vowel harmony, i.e., the vowels are distributed into two sets, and all the vowels of a given word may belong to only one of these sets. The parameter commanding the sets is typically [ATR] (e.g. Dagik; see Vanderelst : –), while in some cases, it seems to be height (e.g. Koalib opposes a high set /i, ɐ, u/ to a low set /e, ε, a, ɔ, o/ (see Quint b: –, : –);6
Dimmendaal (: ) proposes that [ATR] and non-[ATR] systems occupy different geographical zones in the Nuba Mountains, but the delimitations of such zones are at least partially based on Stevenson’s material (/), which is not fully reliable as regards vowel harmony in Kordofanian languages. 6
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
(iii) a majority of Kordofanian languages have five places of articulation for their consonants, i.e. basically: (A) labial, (A) dental/interdental, (A) alveolar/ retroflex, (A) palatal, and (A) velar. The contrast dental vs. alveolar (already mentioned by Williamson and Blench (: ) as a Kordofanian trait) is quite widespread among Kordofanian languages and it has a high functional load: ()
Koalib (Heibanian) òtté [òt t̪ é̪ ] vs. òţţé [òʈʈé] ‘make rough (a grindstone)’ ‘gather (sorghum)’
()
Dagik (Talodi) (m)ən̪ ɔ vs. (k)ənːɔ ‘swim’ (v.) ‘taste’ (v.)
()
Tima (Katloid) t̪úh vs. ʈúh ‘uproot’ (v.) ‘hang’ (v.)
(Vanderelst : ) (Bashir : )
Only Rashadian languages do not seem to have the contrast (A) dental/interdental vs. (A) alveolar/retroflex (Stevenson / []: ; Schadeberg c.: , , ;) However, recent research by Bashir (, p.c.) seems to show that this contrast also exists at least in some Tagoi varieties. Be that as it may, many non-Kordofanian languages spoken in the Nuba Mountains also have tones, vowel harmony, and five places of articulation. Although they are shared by most Kordofanian languages, these three phonological features may therefore be areal and not inherited.
.. Morphology ... Noun classes They are considered the most typical characteristic of Kordofanian languages and the main criterion which justifies the inclusion by Greenberg ( []: –) of these languages in the Niger-Congo phylum. As a matter of fact, in most Kordofanian languages, each noun belongs to a noun class characterized by a prefix generally borne by the noun and which commands agreement on several modifiers of the noun and the verb. ()
Koalib (Heibanian; Quint : ) a. kw-ór kw-ínyí kw-òppá. kw-man kw-: kw-be.big: ‘My husband is big.’ (kw- agreement of the possessive and verbal prefixes with the noun kwór.) b. l-ə̀pə́ntì l-ínyí l-òppá. -teacher -: -be.big: ‘My teacher is big.’ (l- agreement of the possessive and verbal prefixes with the noun lə̀pə́ntì.)
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
Dagik (Talodi; Vanderelst : ) p-at ː̪ ɪ i-b-ige b-asɔ. p-man -p-sick p-come: ‘The sick man came.’
()
Tagoi (Rashadian; Stevenson / []: ) a. f-aran f-irmo f-cow f-black ‘black cow’ b. y-εrŋan y-irmo y-stone y-black ‘black stone’
In addition to the presence of noun classes in many Kordofanian languages, several of these noun classes are quite similar across families (see Table .).
Table . Some similarities between noun class markers of various semantic classes across Kordofanian families Semantic core of the class Noun class marker1 Heibanian
Lafofa
Rashadian
Talodian
Proto-form
(a) |.:|
(a) /kw/
(a) /b~p/
(a) /w/
(a) /p/
*/kw/
(b) ||
(b) /ŋ/
(b) /ŋ~ɲ/
(b) /ŋ/
(b) /ŋ/
*/ŋ/
(c) |.:|
(c) /t /̪
(c) /t /̪ ?
(c) /t~t /̪ ?
(c) /t /̪
*/t /̪
(d) ||
(d) /ŋ~t /̪
(d) ?
(d) /ŋ/
(d) /t ~ ̪ ŋ/
*/ŋ~t /̪
(e) | |
(e) /ŋ~t /̪
(e) ?
(e) /ŋ/
(e) /k~t /̪ 2
(f) |:|
(f) /ɲ/
(f) ?
(f) /ɲ/
(f) /ɲ~ŋ/
3
*/ŋ~t /̪ */ɲ/
1 References: Heibanian: Schadeberg (a); Lafofa: Schadeberg (b); Rashadian: Schadeberg (, c.); Talodian: Norton (: ); Vanderelst (: ); Schadeberg (b); protoforms: Schadeberg (). Some proto-forms in Table . differ slightly from the ones proposed by Schadeberg; I am responsible for these differences. 2 Norton and Kuku Alaki ( ) give only /k/ as a reconstructed form for noun class markers in Talodian. However, at least three Talodian languages exhibit with a noun class marker akin to Proto-Talodian /t /̪ : Dagik, Daloka, and Tuwal, which makes quite reasonable to postulate the existence of a /t /̪ variant in Proto-Talodian, all the more so as /t /̪ is frequently used as an noun class marker in Talodian and the categories of and are semantically interrelated in many Kordofanian languages (see the discussion of (e) in section ...). 3 At least two Talodian languages (Norton and Alaki : ), namely Dagik and Tuwal, have a noun class marker /ŋ/ for the |:|, e.g. Dagik ŋura ‘little/tiny calf ’ (VE:). However, within a comparative perspective, we have only retained the Talodian /ɲ/ marker, whose form obviously coincides with its homologues in Heibanian and Tegalian.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
a. (a) Koalib kwór ‘man’, (a) bid̪εgέm ‘Tegem [=Lafofa] person’ (TSd 7:), bəlibwere ‘blacksmith’ (TSd:), p-uma b-owi ‘woman’ (SV[]:), (a) wàrórək ‘thief ’ (MS:), Orig wóoríg ‘person from J[ebel] Turjuk’ (MS:), (a) Dagik p-ʊɽa ‘thief ’ (VE:)
Phonetically, the correspondence between the four families seems quite plausible: (i) conservation of Proto-Kordofanian */kw/ in Heibanian; (ii) passage /kw/ > /p/ in Talodian, an evolution path widely attested in other languages such as Indo-European, e.g. Latin (Italic) quinque /kwinkwe/ ‘five’ vs. Breton (Celtic) pemp or Latin quis /kwis/ ‘who..:’ vs. Oscan (Italic) pis; (iii) loss of the plosive element /k/ of Proto-Kordofanian in Rashadian; (iv) Lafofa /b~p/ variation is finally in line with Talodian /p/, as there is fluctuation within Talodian itself as to the voiced/unvoiced realization of Proto-Talodian /p/ (see results of Schadeberg’s survey (b) for the words ‘husband’ (p. ), ‘man’ (p. ) and ‘person’ (p. )). b. (b) Koalib ŋâo ‘water’, ŋèelà ‘oil, fat’, (b) Lafofa ŋ-εi ‘urine’ (TSd:), ŋo ‘milk’ (TSd:), ɲie~ɲee ‘blood’ (TSd:), ɲ-íí ‘water’ (TSd:), (b) Orig ŋàì ‘water’ (MS:), ŋoc ‘urine’ (MS:), (b) Dagik ŋa ‘oil’ (VE:), ŋeru ‘blood’ (VE:), Lumun ŋaák ‘fat’ (n.) (HS), ŋʊccʊk ‘blood’ (HS) Here the correspondence between the four families is nearly perfect. Regarding Lafofa, we can note that all known items having both an initial /ɲ/ and a meaning have a palatal vowel as the first element of their root (ɲie~ɲee, ɲ-íí ). This palatal vowel could account for the shift to /ɲ/ (palatal nasal) of an original class marker */ŋ/. c. (c) Koalib tée [t ̪ée] ‘arm, upper limb’, tél [t ̪él] ‘horn’, téɽàny [t ̪éɽàɲ] ‘boundary’, (c) t ̪-ɔ́wáa-áy ‘arm’, t ̪-û-í ‘horn’, (c) Tagoi t ̪-əwan ‘rope’ (ST [H]:), túúr (iŋ) ‘horn’ (TS:), téŋlàk ‘tongue’ (TS:), (c) Dengebu t ̪-ʊ́lʊŋ́ ε̂ ‘tongue’ (TSd:), Tocho t ̪-úúbε ‘horn’ (TSd:) d. (d) Koalib ŋémà ‘strength’, ŋèpèetáŋ ‘whiteness’ (< pèeté ‘be white’), ŋètὲny ‘fear’, táakà [t ̪áagà] ‘marriage’ (< àaké ‘marry’), Tira ŋáòð ‘work’, ð̩́bɽá ‘strength’, (d) Orig ŋírís ‘fear’ (MS:), ŋúmán ‘shame’ (MS:), (d) Dagik ŋərε ‘work’ (VE:), ŋəma ‘strength’ (VE:) e. (e) Koalib ŋɐ̀tkù ‘Arabic’, Werni t ̪èrðà ‘Arabic’, Tira ðìccùl ‘Arabic’, (e) Orig ŋɔɔ́ ́ríg ‘Turjuk [= Orig] language’ (MS:), ŋȩ́gdìráá ‘Arabic’ (MS:), (e) Dagik ð-əlːamε ‘Arabic’ (VE:) What is striking is the general hesitation in the three Kordofanian groups for which this semantic class is documented between /ŋ/ and /t̪/ markers, which are also used to mark abstract nouns. The reason for the parallelism between and is probably the fact that a can be conceived of as the abstraction par excellence of a given cultural community (e.g. “Arabic” is the emblematic associated with Arab people, etc.).
For abbreviated references, see key preceding the “References” section. Unreferenced data were collected by the author. 7
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
f. (f) Koalib nyór ‘children, nyàkró [ɲàgró] ‘chicks’ (< yàkró ‘hen’), Werni ɲákóró ‘chicks’ (< kákóró ‘hen’), Tira ɲìmná ‘kids’, (fR) Orig ɲímnàn ‘kids’ (MS:), ɲìrŋán ‘small stones’ (< yìrŋán ‘stone’, MS:), Tagoi ɲíŋ ‘children’ (TSb:), (fT) Joomang ɲáanûŋ ‘children’ (TSd:), Lumun ɲʊkʊl ‘children’ (Smits )
However, despite their widespread use in Kordofanian and the striking similarities they display through different families of this grouping, noun classes are not a consensual criterion to justify that Kordofanian regroups languages belonging to the same phylogenetic stock and therefore makes up a particular branch of Niger-Congo. (i) Some Kordofanian languages have no noun classes. This scenario is represented by Tegali (Eastern Rashadian) languages8 as opposed to Tagoi (Western Rashadian) languages which have noun classes. Regarding Rashadian, the core of the debate is to choose between two hypotheses: () either all Rashadian languages used to have noun classes and these were lost at a given period by Tegali languages only, whereas Tagoi languages kept them to this day (see Schadeberg , c.), or () all Rashadian languages used to have non noun classes and Tagoi languages acquired those classes due to the influence of the neighboring Kordofanian noun class languages, in particular Heibanian (see Stevenson / []: ; Blench : –). Although I personally support () (due to the many resemblances shared by Rashadian and other Kordofanian languages), I think that no decisive evidence has been adduced in favor of either of these hypotheses. (ii) Other Kordofanian languages only have a reduced system thereof. This scenario corresponds to Katloid languages which have very few distinct noun classes; for example, in Tima (Alamin b: –), there are two morphologically marked general singular noun classes (characterized by prefixes /kV/- and /t/- respectively), and two locative classes plus an abstract noun class (prefix /dV/-), the latter being used in particular to derive language names (see Table .). However, reduced as it may be, some features of Tima class morphology are clearly reminiscent of typical Kordofanian class languages. As shown in Table ., the class derivation > is realized both in Tima and Koalib through prefixal alternation.
8
In fact, Schadeberg (c.: ) signals the existence in Tegali of at least two productive prefixes
ŋV- and t-, used to derive language- and place-names respectively, and Tucker and Bryan (: ) also mention (citing Stevenson but without the exact reference) the use of ŋV- to derive abstract nouns.
Although these prefixes do not command class agreement on nominal dependents, one cannot but notice their evident formal proximity to noun class markers attested in other Kordofanian languages: (a) ŋ- is widely used with nouns belonging to the and - semantic classes in many (see Table .) and (b) a t(V)-/ð(V)- marker used to produce locative nouns exists both in Tagoi (e.g. toorig ‘the village of Jebel Turjuk’ (where Orig is spoken, MS:) and at least in some Talodian varieties (e.g. Dagik tɔ̀-sɔ́ ‘Lumun/Tocho/Asheron area’, VE:).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Tima (Katloid) and Koalib (Heibanian, noun class language) language names nominal derivation Language
Form
Meaning
Form
Meaning
Tima (Katloid)
kɘ̀mààdáŋ
‘Katla person’
dɘ̀mààdáŋ
‘Katla language’
Koalib (Heibanian)
kwèjὲŋkέ
‘Dinka person’
ŋèjὲŋkέ
‘Dinka language
(iii) Some non-Kordofanian languages spoken in the Nuba Mountains also have noun classes. This is in particular the case of Kadu languages, traditionally included in Nilo-Saharan (Schadeberg c, ) and spoken in the southwestern part of the Nuba Mountains, in the vicinity of Talodian and Heibanian languages. The existence of noun classes in Kadu is unquestionable: ()
a. miḑε y-adagbo man -big ‘this big man’
y-ɔ -
b. ka m-adagbo m-ɔ woman -big - ‘this big woman’ However, the central semantic distinction in the Kadu class system is between / and / (see (a) and (b), taken from Stevenson / []: –), whereas in most Kordofanian (and more generally Niger-Congo) languages, the basic contrast is usually between and - (Creissels : –; : ). This strongly suggests that Kadu languages might have borrowed (or copied) the morphological structure (prefixal concord markers) of Kordofanian noun classes while maintaining the essentials of their own semantic organization, which is clearly at variance with Kordofanian as a whole.
... Verb extensions All Kordofanian families have numerous verbal extensions and suffixes conveying various meanings, such as those exemplified by Koalib (Quint a) (see Table .). Similar verb extensions are found in all other Kordofanian languages for which sufficient data are available: • In Tima (Katloid), we find (Alamin b: –; Alamin et al. ) an antipassive, two applicatives (benefactive/recipient and instrumental), a causative, a middle voice (with both reflexive and reciprocal values), a separative (“movement out of some original position”; see Alamin b: ), a reversive, and a ventive. • In Dagik (Talodian), we find two applicatives (benefactive and locative), a causative, an associative, an inchoative-stative, a middle voice, and an iterative (Vanderelst : –). • In Tumale (Rashadian), we find at least a benefactive and possibly a transitive (Stevenson / []: ).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Some verb extensions attested in Koalib (Heibanian) Verb extension
Form of the extension
Example 1
Segmental
Tonal
Applicative (benefactive)
-ccE2
H(L)nH
àé ‘die’ > ɐ́ìccí ‘die for s.o.’
Applicative (malefactive)
-(a)tA
H(L)n
nyìimí ‘steal sth.’ > nyíimɐ̀tɐ̀ ‘steal sth. from s.o.’
Associative
-VtE
(L)nH
àppé ‘carry’ > àppàté ‘carry together’
Causative
-i + HVHS
ájlè ‘be weak’ > ɐ̀jlí ‘weaken’
Excessive
-AttE
ɔ́blὲ ‘be short’ > òblàtté ‘be too short’
Immediate
-(t)AnnE
ìiɗí ‘fall’ > ìiɗɐ̀nní ‘fall suddenly’, ţùú ‘go out’ > ţùutɐ̀nní ‘go out at once’
Locative/transitive
-AcE
ɽùuní ‘bring s.o. up’ > ɽùunɐ̀cí ‘bring s.o. up somewhere’
Passive
-(t)VnnE
ìppí ‘beat’ > ìppìnní ‘be beaten’, yὲέ ‘eat (tr)’ > yὲεtènnέ ‘be eaten’
Reciprocal
-AtEcE
ὲɽnyέ ‘kill’ > èɽnyàtècé ‘kill each other’
Reflexive/middle
-VnnE/ -VtnE
ìppí ‘beat’ > ìppɐ̀nní ‘beat oneself ’, yὲέ ‘eat (tr)’ > ὲtnέ ‘eat (intr = for oneself)’
3
H = high tone; L = low tone; n = number of syllables with n . Except for causative (see next note), the vowels of the verb extensions are indicated in uppercase as their actual realization depends on the Vowel Harmonic Set of the lexical root they attach to: /i, ɐ, u/ belong to the high set, and /e, ε, a, ɔ, o/ belong to the low set. A Koalib phonological word can only contain vowels belonging to one and the same set (i.e. both sets cannot co-occur in a given Koalib lexical item). For more details about vowel harmony rules in Koalib, see Quint (:-, b:–). 3 HVHS = High Vowel Harmonic Set. The causative extension generally triggers the use of vowels belonging to the high set in the lexical root itself. 1 2
Note that, except for the Rashadian transitive, all these Kordofanian extensions behave according to the same morphologic pattern, as they are produced through suffixes which attach to the verb root. In spite of the fact that semantically related verbal extensions are found in many African languages (Hyman forthcoming), for a linguist who has worked in-depth on a Kordofanian language (i.e. Koalib for the author), it is obvious that: (i) the verb extensions described by other linguists for other Kordofanian languages do share many common points between themselves.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
(ii) these same verb extensions also present striking similarities with other Niger-Congo languages, such as those belonging to the Atlantic stock (as I have been able to discover myself while working on a member of this grouping, namely the Nyun variety of the village of Djifanghor, Senegal; see Quint forthcoming) or the Bantu family (e.g. the surprising resemblance of some specific uses of applicative derivations in both Zulu (Bantu) and Koalib (Kordofanian) described in Quint a: –).
.. Lexicon Kordofanian languages also present similarities regarding lexicon. However, systematic lexical comparisons remain difficult, as, for a majority of Kordofanian languages, we still only have at our disposal meager wordlists (generally adapted from Swadesh) comprising at best or items, and this scant material does not allow much more than impressionistic judgments and intuitions. At least two important lexical collections are now available: a Tima dictionary (Schneider-Blum ; c., entries) and a Koalib dictionary (Ali Karmal Kokko and Quint forthcoming; c., entries). One can also mention the nearly , entries of the lexicon recently produced by Vanderelst (: –) for Dagik and several other recent publications (e.g. Bashir , ; Alamin b) which mention many lexical items in their analyses. I will show how the existence of these new resources allows us to find many more cognates than had been noticed until now between the various Kordofanian families.
... Heibanian and Katloid A rapid examination of the available documentation for Tima yields around twenty plausible cognates with Proto-Heibanian (as reconstructed by Schadeberg a) or Koalib (see Appendix ).
... Heibanian and Rashadian Here too, the available material allows one to recognize dozens of convincing cognates. Actually the number of specific lexical coincidences with Koalib (see Appendix ) is striking (e.g. ‘lower leg’ or ‘six’). One may attribute some of these lexical similarities to the fact that the Koalib linguistic area is in direct contact with the Rashadian (Tagoi) languages. Note, however, that (i) the Koalib data presented here come from Rere (Quint ), which is spoken in the very center of the Koalib country (and therefore not in direct contact with Rashadian), and (ii) the lexical resources available for Rashadian languages are quite limited, which renders all the more significant this relatively high number of lexical similarities that I have been able to identify between the two families.
... Heibanian and Talodian Some clear cognates can be identified (see Appendix ). Note that, despite the comparative work produced by both Schadeberg (a) and Norton and Kuku Alaki (), it is harder to find cognates between Heibanian and Talodian than between Heibanian and
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Rashadian. This result is contrary to several recent publications, such as Blench ()9 and Dimmendaal (: ), which tend to advocate the existence of a stronger link between Heibanian and Talodian in contrast to the remaining Kordofanian families.
... Heibanian and Lafofa There are indeed some convincing cognates between those two families (see Appendix ). However, their number is quite limited (fewer than twenty for the moment). Two remarks are in order about this scarcity of cognates: (i) the result of these comparisons illustrates once more (Schadeberg b; Hammarström ; Norton and Kuku Alaki ) the specific status of Lafofa inside Kordofanian. Indeed, it seems more reasonable to consider Lafofa as an independent Kordofanian family and not as a branch of Talodian; (ii) be that as it may, the relatively reduced number of cognates available for Lafofa is also due to the lack of available data for this Kordofanian family, which is probably now the least studied of the whole Kordofanian branch.
... Other comparisons between Kordofanian families: KatloidRashadian and Talodian Many more pairs (or triplets . . . ) of the five Kordofanian families could be compared in the same way as I have done in sections ...–... above for Heibanian and each of the other four Kordofanian families. A comparison involving Katloid, Rashadian, and Talodian has in particular been undertaken, suggested, or discussed by several scholars. Stevenson (/ []: ) already noticed some lexical similarities primarily between Katloid and Rashadian languages and secondarily between these and Talodian. These shared items (which do not seem to be chance correspondences) are one of the main arguments used by some contemporary scholars to postulate a common origin for Katloid and Rashadian (Blench : ; Dimmendaal : ), which would make up together one of the primary sub-branches of Kordofanian. In Appendix , I present a list of these potential cognates (excluding cases for which a plausible cognate can also be found in Heibanian). Note that (i) although these items clearly belong to the core lexicon of the languages at stake and display striking semantic and segmental similarities, their actual number is quite low; (ii) Stevenson based his comparative list (which represents the majority of known potential cognates) on individual languages and not reconstructed forms of proto-families (which remain to be proposed for Katloid). However, in most cases, the now available reconstructed forms do not seem to invalidate Stevenson’s observations. Within the perspective advocated in this chapter (namely that the five Kordofanian families probably are genetically related), I will consider that, when Katloid, Rashadian, and Talodian forms display a cognate unattested in Heibanian or Lafofa, this cognate can plausibly be traced back to Proto-Kordofanian. Indeed, as Heibanian is geographically
9 However, in a more recent paper (Blench ), the same author also questions the existence of a specific HeibanianTalodi grouping, in view of the relative paucity of cognates and other similar characteristics between the two families.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
situated between Talodian, Rashadian, and Katloid, it is more likely to suppose that items shared by the three later families represent retentions from a common ancestor.10
... Proto-Kordofanian roots The comparisons produced in this chapter allow one to propose Proto-Kordofanian roots for items which are attested in similar forms in at least three of the five Kordofanian families (see Table .).
Table . Some plausible Proto-Kordofanian roots Family Talodian
Proto-Kordofanian
Meaning Heibanian
Katloid (Tima)
Lafofa
Rashadian
bark (n.)
Koalib kìmùukùl [kìmùugùl]
UN
UN
Orig kìmbàkɔ́l *t-ə-mək
*-mVk-
bark (n.)1
UN
kúúr
c-iíri
*g-ware
UN
*k(V)VrE~c (V)VrE
beat/hit
*-bid̪-
UN
. . . biŋ
Orig bí(r)/ pù(ró)
*-gob-/ *kə-bɔ *-bV-
belly
*g-+-aare
kúɽúún
t̪-úur-i
UN
*j-+-arag / *ca-rək
*-VrVk~ *-VɽVk
bite/eat
*-iy-
UN
. . . jiε
*yεk
*-gVjog/ *kə-ɟɔ
*-CVk
blood
UN
Katla ija~iya
UN
Rashad wiya
*ŋ-+ -ittsug / *ŋ-ɪccʊk
*(C)iCpV
breast
UN
Tima kᵼ̀míndì
UN
*d-miɲ
*j-+ -intsig / *c-ə-mmik
*CVmiC
clothes
*g-+-ered̪
UN
UN
Tagoi kεr(έ)w
*k-εrεt̪
*kErEC
dry
*-unDo
UN
UN
*-uddi
*-an d̪o
*-OndV (continued )
10
The other possibility would be that Katloid, Rashadian, and Talodian speakers came in contact with each other during their migration to the Nuba Mountains or at a time when the speakers of the Proto-Kordofanian families (i.e. Proto-Heibanian, Proto-Talodian, etc.) occupied locations different from the ones they occupy now in the Nuba Mountains. As there is no reliable source documenting the linguistic geography of the region before the nineteenth century, a lot of different scenarios can theoretically be envisaged. However, (i) the important linguistic differentiation observed between the various languages of each Kordofanian family and (ii) the dialect continuums observed in several places (e.g. between Koalib and Heiban; see also note ) seem to support the fact that the various Kordofanian communities have been residing in their present-day areas (or in nearby places) for a considerable period of time (Quint b: –, : –).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Continued Family Meaning Heibanian
Proto-Kordofanian
Katloid (Tima)
Lafofa
Rashadian
Talodian
kɔ́.nɔ̀
UN
Tagoi finin, Tegali (a)nuu (n)
*g-+ -eenu / *k-εnu
*kVnV
elephant *d-+-oŋor
UN
yu:ŋi
*(fV)ŋVn
NA
*-VŋV(C)
eye
UN
Katla gᵼgöt̗
UN
*y-ngid
*j-+-igg / *c-it *?+-git
foot/leg
Koalib káakà [káagà]
UN
l-ia-ga
*d-εgεn
*ts-+ -agag
*-AkA(C)
goat
UN
Tima címìð
εεmi
*mbᵼt
*w-+ -emig / *u-mit
*Em(b)iT
green/ wet
*-iigla
UN
b-ʊ́ɒji-lli
Tagoi -ijilú, Tegali -rígε̂l
UN
*-iklV~ ijlV
hair
UN
Tima káàm
UN
*g-aam
*d-+ -ʊgaŋ / *-ŋən̪
*kaam~ *gaam
head
UN
Katla gas UN
*g-aj
*j-+-ats, *c-ac
*gaCDP~ CPaCDP
left (side)
*-awur
UN
kúlɪ
*-awwir
*-gule / *-gulε *-CVul-~ *-CVur-
mud
Koalib kèlòo
Tima kʌ́ʌ́lu
UN
Orig ŋí ̧lɔ́
UN
near
Koalib kέttɔ̀k [kέt ̪t ̪ɔ̀k]
Tima mὲt ̪έn
UN
Tagoi gattɔŋ / *-iddu / *-t ̪t ̪o- *-Et ̪t O ̪ C tɔgɔt t̪
one
*-aDDe
Tima àtíín
UN
Orig wàttá
UN
*attV ~ *addV
rain
*g-+-aw
UN
k-állɔ́-y
*(y)au
*k-abɪk
*kaw ~ *kal
red
*-UUre
Tima -rdí UN (SB:)
*-araw
*-oode (TSd) / *-OrdE *-ḍε (NK:)
sheep
Koalib káaŋàl (k)áŋàl
ßa:ŋi
Orig kàgóy
*t̪ʊ-ŋgat̪
smoke (n.)
*g-+-ulu
kʊ̀ʊ̀ɽʊ́n
c-oor-í
Tagoi k(ə)rək, UN Tegali tulέ
sun
ø-+-*aŋin
kínèè
UN
*-aane
*j-+-iŋgi, *c-ə- *-VCNV ŋgi
tongue
d̪-+*-ŋela
kìlíŋíì
l-íáŋ-i
*d-aŋil(-ag)
*d ̪-+ -(V)lVŋe *-d̪Vŋl(V) ~ / *t ̪ʊ-ləŋε -d̪VlVŋ(V)
vomit
*-wey-
-húwʌ̀
lwâ*VdVk d̪aŋ, . . . lwa
ear
*g-/n-+ -aani
-VddV / *ukdε
*-ElO
*kACVAC *-uCLVBV
*-UdA ~ *-UwA
For the existence of two Proto-Kordofanian roots for ‘bark’, see Appendix and the footnote referring to this item therein. 1
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Some of the above items, relating in particular to body parts (e.g. ‘belly’, ‘ear’, ‘foot/leg’, ‘tongue’), qualifiers (e.g. ‘dry’, ‘green/wet’, ‘red’), common verbs (e.g. ‘beat/hit’, ‘vomit’) or other basic notions (e.g. ‘left’, ‘rain’) are clearly part of the core vocabulary of any language and the similarities attested between the different Kordofanian families may plausibly be linked to a common origin (i.e. a Proto-Kordofanian stage). For some other items, more caution is in order. This is the case for words such as (i) ‘clothes’: until the twentieth century, a significant proportion of Kordofanianspeaking communities did not wear any clothes and therefore the notion of clothing was probably not as lexically central as body parts or the ‘rain’; (ii) ‘sheep’: although sheep-breeding seems to have been practiced for several millennia in East Africa (Blench and MacDonald ), the degree of similarity between Koalib (Heibanian) and Tima (Katloid) forms is too high not to be suspicious if one thinks that these groupings, admitting that they both derive from one and the same “Ur-Kordofanian” ancestor, must have been separated for at least , to , years, judging from the many divergences they display. The similarity between Katla and Koalib might as well be due to borrowing either (a) at a time when Proto-Katloid and Proto-Heibanian speakers lived side by side or (b), if we take into account the fact that all Kordofanian branches exhibit convincing cognates for ‘sheep’, at a relatively recent time, when a specific variety of sheep expanded through the Nuba Mountains and the name of this new variety accompanied this expansion. Scenario (b) is supported by the fact that other widespread lexical roots in the Nuba Mountains, i.e. /mVrtV/ ‘horse’ or /ʃVrtV/ ‘iron’, have convincingly been shown to be the result of lexical diffusion through borrowing, the ultimate source of these wandering words being generally the Nile Valley (see Quint : – and references therein). As summary as it may be, the results of the small lexical comparison undertaken in this chapter clearly show that all Kordofanian branches do share some common lexical roots in their core vocabulary. This result strengthens the hypothesis of the existence of a ProtoKordofanian stage from which Heibanian, Katloid, Lafofa, Rashadian, and Talodian would all be derived.
. C
.................................................................................................................................. Much more research should be done and much more data should be gathered in order to develop sound comparisons between the different branches of Kordofanian and between Kordofanian and the other families which make up the Niger-Congo (or NigerKordofanian) phylum. At any rate, the grammatical and lexical commonalities shared by many Kordofanian languages seem to support the hypothesis of a common origin for these languages. It is to be hoped that the increasing knowledge assembled by the new generations working in this field will enable us to solve one of the most interesting puzzles for Niger-Congo linguists by providing more detailed answers to the three following questions: (i) What really are the precise linguistic characteristics of Kordofanian languages? (ii) What is the exact relationship between Kordofanian languages and how can we account for the divergences observed in the attested varieties? (iii) How does Kordofanian fit into Niger-Congo and what historical processes can be adduced to explain the existence of this
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
isolated pocket separated (both by a large distance and different communities) from the rest of its Niger-Congo sister-languages?
. : K
..................................................................................................................................
.. Appendix : Plausible lexical cognates between Heibanian and Katloid Katloid1
Heibanian
Meaning
Tima
Proto-Heibanian
Koalib
belly
kúɽúún (SB:)
*g-+-aare
káaré
buy
əəl-/éèl (AL:, BSb:)
NA
èelá
cough (n.)
kwààr.lɘ́l (BSa:)
NA
kwàrlé
ear
kɔ́.nɔ̀ (AL:, BSa:, SB:)
*g-/n-+-aani
kɐ́ɐnì
fish
(ku)-mòŋ (SB:)
*gu-+-Um
kwôm
food
t ̪ʊ̀rì (SB:)
NA
tûl [t ̪ûl] (: tùlì) ‘sorghum porridge’
hammer
d̪ɔr (BSa:–)
NA
ţɔ̀r [ʈɔ̀r]
lungs
pàʈpàʈ (BSb:)
NA
kèpɔ́tpɔ̀t [kèvɔ́ðvɔ̀t ̪]
male
ø-kɔ́ŋkɔ́r (BSa:, SB:)
*-oŋor
kwóoŋòr
mud
k-ʌ́ʌ́lu (AL:)
NA
kèlòo ‘salty mud’
navel
kᵼ̀-lʌ́lù (AL:)
NA
kwúllɐ̀
near
mὲt ̪έn (SB:)
NA
kέttɔ̀k [kέt̪t ɔ̪ ̀k]
one(RN)
àtíín (SB:)
*-aDDe
kwέttέ
red
-rdí (SB:)
*-UUre
óorè
scar
yuulʌ (SB:)
NA
lóolè ‘scarification’
kVɽ-/kVr- (majority of Heibanian)
òkwɽé
sew
#
kúrúú (BSa:)
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
sheep
(k)áŋàl (SB:)
NA
káaŋàl
smoke (n.)
kʊ̀ʊ̀ɽʊ́n (BSa:)
*g-+-ulu
kwùulú
snake
kᵼ ̀-mᵼ ́nʌ̀ (BSa:)
UN
kímɐ̀u
sun
kínèè (BSa:)
*-+-aŋin
àaŋwòn
thirst
yeedi (SB:)
NA
ɐ́ɐtɐ̀
tongue
kìlíŋíì (AL:, BSa:) *d̪-/d-+-ŋela
téŋlà
uncle
mʌ́múŋ (BSb:)
NA
mɐ́m ‘mother’s brother’
vomit
-húwʌ̀ (SB:)
*-wey-
ùɐ́
waterbag
tùlkù (SB: )
NA
kwúlkùŋ [kwúlgùŋ] ‘goatskin’
white(ness)
bı ̀tʌ̀ k (BSa:)
*-biid̪o
pèeté [f/vèeðé]
After each Tima item, I give the initials of the authors consulted (see “Abbreviated references” above) followed by the page number where the item has been found. For Proto-Heibanian, all data come from Schadeberg (a). For Koalib, the data are mine. The segments in bold letters are those that present most similarities between Katloid and Heibanian items. Whenever an English meaning is followed by (RN), it means that Russell Norton suggested to me the comparison for the item in question. 1
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Appendix : Plausible lexical cognates between Rashadian and Heibanian Rashadian1
Heibanian
Meaning
Orig/Tagoi (Tagoi)
Gom (Tegali)
ProtoProtoKoalib (Heibanian) Rashadian Heibanian
Achilles’ tendon
kiric (MS:)
NA
NA
NA
kìrìi ‘tendon’
back of head
yìndɔ́ŋ (MS:)
NA
NA
NA
lèɗɔ̀ŋ
baobab
wɔr (MS:)
NA
NA
NA
kwòkwɔ̀r
bark
kìmbàkɔ́l (MS:)
UN
UN
UN
kìmùukùl
beard
kábú (MS:)
NA
NA
NA
kɐ́ɐbú
beat/hit
bí(r)/pù(ró) (MS:)
NA
UN
*-bid̪-
ìppí
brain
yáràɲ (MS:)
NA
NA
NA
ŋàɽény
canine tooth
tèrèm (MS:, )
NA
NA
NA
téeɽám
cave
kùbáŋ (MS:)
NA
NA
NA
kèbàŋ
cheek
yìnèènè (MS:)
NA
NA
NA
lènέεnὲ
cloth, dress
kóbàŋ (MS:)
UN
NA
NA
kwópàn [kwóvàn]
clothing
kεrέ(ε)w (TSb:)
UN
NA
*g-+-ered̪ kèrέt
doum (tree)
wàndɔ́k, fruit: yàmdɔ́k NA (MS:)
NA
NA
àmtó
dry
-uttə́/-ottó (TSb:)
*-uddi (TSc:)
*-unDo
ònté [ònd̪é] (: òntò)
dust
tú̧láŋ (MS:), kólólóŋ tólóŋ (TSb:) NA (TSb:)
NA
kúulɐ́ (: kùulɐ́ŋí)
ear
fɩnɩn (MS:)
nu/nũũ/ núun/ NA anuun (TSb:)
*g- + aani
kɐ́ɐnì
eat
yεk (TSb:)
yεk (TSb:)
*-iy-
yὲέ
eight
tùppá (MS:)
tuppá (TSb:) *tuppá (TSc:)
*-uuba(ŋ) ɗòpɔ̀kkwóppà [ɗòvɔ̀kkwóppà]
elbow
tulu (MS:, )
NA
NA
elephant
fàŋán (MS:), fu-ŋin/ ŋin(ε)/ŋεni fə-ŋᵼn (SV-:) (SV-:)
-údî (TSb:)
*yεk (Q)
NA
*(fV)ŋVn *d-+(Q) oŋor
tóolóŋ tóoŋwór [t̪óoŋwór]
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
fat
ŋaarəf (TSb:)
iri ‘oil’ (TSb:)
*-aarif (TSc:)
*ŋ-+eeLa
ŋèelà
flower
kàbláɲ (MS:), kabalíɲ (TSb:)
oofənjεlak (TSb:)
NA
*-biraŋ?
ópɽény [óvɽéɲ]
foot/leg
tὲgàn (MS:), tέgεn (TSb:)
εgέn (TSb:)
*d-εgεn (TSc:)
UN
káakà [káagà]
go
ndé (MS:), έέ-dέ/έndε (TSb:)
ε-ndεk (TSb:)
*ε-ndε(k) *-nD(Q)
green
-ijilú (TSb:)
-rígε̂l (TSb:) NA
*-iigla
-íklì [íglì]
guest
w-εrᵼn (SV-:)
NA
NA
NA
kwìirìn
hare
cín (MS:)
NA
NA
NA
ŋìinì
kid
címnàŋ (MS:)
NA
NA
NA
ţémnà [ʈémnà] ‘young goat’
áwə́r (TSb:) *-awwir (TSc:)
*-awur
ŋɐ̀ɐkùr [ŋɐ̀ɐgùr] ténţény [t̪énɖéɲ] ‘front of the lower leg’
left (side) ŋáwwúr (TSb:)
ὲεlέ (: ntu~nti [ndu~ndi])
lower leg
tέndὲny (MS:, )
NA
NA
NA
many
-ə́tta (TSb:)
UN
NA
*-uDDun -éţţàcér
milk
ŋá̧n (MS:), ŋan (SV- NA :)
NA
NA
mud
ŋí ̧lɔ́ (MS:)
near
gattɔŋ (TSb:), tɔgɔt (TSb:)
UN?
NA
UN
kέttɔ̀k
night
kìrìm (MS:, ), kərəm (TSb:)
úgri (TSb:)
NA
UN
kìrìm ‘darkness’
old (6¼ new)
-ɒrɒ́w (TSb:)
arɔ́ (TSb:)
*-VrV (Q)
*-ooɽan
-óoɽón
one
wàttá (MS:, ), -ε ndá (TSb:)
w-uttá (TSb:)
NA
*-aDDe
kwέţţέ
NA
NA
NA
kwôn
porcupine wor (MS:)
ŋán kèlòo ‘salty mud’
python?
wàm (MS:)
NA
NA
NA
kwâm
rain
yàʊ̀ (MS:), yau/ yaaw (TSb:)
áú (TSb:)
*(y)au (Q)
*g-+-aw
kâo ‘rain’, yâo ‘much water’
red
-àráʊ́ (MS:), -aaráw -aryáu (TSb:) (TSb:)
*-araw (TSc:)
*-UUre
-óorè
salt
mɔ́do/mə́ndu (TSb:)
mude (TSb:) *mVdV (Q)
*-+muda
kwúmtɐ̀ [kwúmðɐ̀]
scorpion
w-udeny (SV-:)
NA
NA
lὲbɗὲny
NA
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
sheep
kàgóy (MS:)
NA
NA
NA
káaŋàl
shoulder
killi (MS:, )
NA
NA
NA
kéllè
six
ɲérér (MS:), ɲέέrέy/ ɲέέrέr ɲεεɽεɽ (TSb:) (TSb:)
*ɲεεrεC (Q)
*ɲ-d-εɽil
nyérlél
smoke (n.)
kɪrëk (MS:), k(ə)rək tulέ (TSb:) (TSb:)
NA
*g-+-ulu
kwùulú
soul, shade
tàglám (MS:)
NA
NA
NA
tíkɽìm [tígɽìm]
sun
yàn (MS:), yáán (TSb:)
áánέ (TSb:)
*-aane (TSc:)
*-+-aŋin àaŋwòn
tail
teyək/tiik (TSb:)
(e)yək (TSb:)
NA
*d̪-+-ia
têa
taste (v.)
-daŋal (MS:)
NA
NA
NA
nyὲŋlέ
tongue
téŋlàk (MS:, ), táŋə́lak/taŋəlk (TSb:)
aŋa (TSb:)
*d-aŋil (-ag) (TSc:)
*d̪- + ŋela
téŋlà
vomit
(w)odək (TSb:)
udak (TSb:) *VdVk (Q)
*-wey-
ùɐ́ (: uɐtu [ùɐðù])
water
ŋa(a)y (TSb:)
UN
NA
*ŋ-+-aw
ŋâo
yam
kìrár (MS:)
NA
NA
NA
kíɽɐ̀r ‘edible tuber sp.’
1 See preceding note to Appendix for the reference system and for Proto-Heibanian and Koalib data. In some cases, when all Rashadian forms present a high degree of similarity, I have myself proposed a Proto-Rashadian form (indicated by Q). The remaining Proto-Rashadian forms all come from Schadeberg (c.).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Appendix : Plausible lexical cognates between Talodian and Heibanian Talodian1
Heibanian
Meaning
Languages
Proto-Talodian
bad
Jomang -ʊ́kki (TSd:), Tocho pɔ-ɡi-d̪ak (NK:)
*-Vggi (TSd), *-ɡi *-gge (NK:)
-kε̂ε (: -kèe)
bark (n.)
Lumun təmmə̂k (HS), Tuwal x-əmək (NK:)
*t-ə-mək (NK:) UN
kìmùukùl [kìmùugùl]
beat/hit
-ɔppí/kɔbɔ (TSd:), Tocho ɔppɔ (NK:)
*-gob- (TSd), *kə- *-bid̪bɔ (NK:)
-ìppí
belly
carə́k (HS), Tasomi j-áárə̯́k (TSd:)
*j-+-arag (TSd), *ca-rək (NK:)
bite/eat
Tasomi d̪á-gíjɔk (TSd:) *-gVjog (TSd), *kə-ɟɔ (NK:)
*-iy-
yὲέ
black
-on̪i (VE: ), -ɔɲî (HS) *-uɲi (TSd)
*-uŋuna
-úŋnì
butter
pəði (VE:)
NA
púutì [fúuðì]
claw/nail
Tasomi g-áárə̯́li (TSd:) *g-+-aareli (TSd), *k-adəḷi (NK:)
UN
kèɽèllè
cloth(es)
Lumun k-εrεt (NK:), k-ərέt (HS)
*k-εrεt ̪ (NK:)
*g-+-ered̪ kèrέt [kèrέt ̪]
cold
-it ̪ːi (VE:, )
*-id̪d̪iŋ (TSd), *-it ̪t ̪iŋ (NK:)
UN
-tε̂ε [t̪ε̂ε] (: -tèe)
drum
Dagik p-amba (NK:), *p-ambaŋ Acheron b-ambaŋ (NK: (NK:) )
NA
ámpà [ámbà]
dry
-an̪d̪ɔma (VE:), -ɔnt ̪ɔ́mat (HS)
*-unDo
-ònté [òn̪d̪é] (: -òntò)
dust
Torona k-ʊɽʊ-baŋ (NK: *t-ə-d̟ʊ(-baŋ) ), Tocho t-uru-bâŋ (NK:)
NA
kèpɽá [kèvɽá] (: kèpɽáŋé )
ear
kenu (VE:)
*g-+-eenu (TSd), *k-εnu (NK:)
*g-+-aani kɐ́ɐnì
fire
Lumun t̪-ík (NK:), Acheron d̪-ɪk (NK:)
*d̪-+-ig (TSd), t̪-ɪk *ø-+-iiga (NK:)
ìikɐ́ [ìigɐ́]
food
Lumun t̪-ʊ̀ɽít (NK:)
*t̪ʊ-l̟ɪ (NK:)
NA
tûl [t̪ûl] (: tùlì) ‘sorghum porridge’
foot/leg
səgək (VE:), Tasodi s-ágâk (TSd:)
*ts-+-agag
UN
káakà [káagà] ‘foot’
NA
*-and̪o (TSd)
ProtoKoalib or other Heibanian
*g-+-aare káaré
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
forehead
suðu (VE:), Torona k- c-ə-d̪uC (NK:) úðu-t/n-uðu-l (NK:)
NA
kìtùl [kìðùl]
fowl
ðakːərʊ (VE:)
NA
NA
kàkró [kàgró] ‘hen’
hard
-ɔn̪d̪ɔ (VE:), -ɔnt ̪ɔ́mat (HS)
NA
NA
-ònté [òn̪d̪é] (: -òntò )
hole
ðu (VE:), c-ɪbʊ́ (NK:)
*d̪-?+-ubu (TSd), *-bʊ (NK:)
*li-+bugul
túu [t̪úu] ‘hole’, lìbùŋ ‘hole (dug into ground)’
human/ person
pʊr (VE:)
*b-+-ur (TSd), *p- *gu-+-ur ʊḷʊ (NK:) ‘man’
left (side) Tuwal ŋɔ-kʊrε (NK:), *-gule (TSd), *gulε *-awur Torona t ̪ɔ-ɣʊrε (NK:) (NK:) *p-εrεn-p-εrεn (NK:)
kwór ‘man’ ŋɐ̀ɐkùr ‘left side’ (tɐ́ɐkúr [t ̪ɐ́ɐgúr] ‘left-handed person’) kwúmpɐ̀rɐ̀ [kwúmbɐ̀rɐ̀]
lyre
Tuwal p-εrε-mb-εrε (NK:)
narrow
-ómmo (TSd: ), Torona *-ʊmmʊn̪ pəmməŋ (NK:) (NK:)
*-muma? ùmmɐ̀ttí
near
gat ̪t ̪î/ɔ́nɔ́gʊ̯tuk̥ (TSd:), *-iddu (TSd), kə-t ̪t ̪u (NK:) *-t ̪t ̪ʊ-t ̪ (NK:)
NA
kέttɔ̀k [kέt ̪t ̪ɔ̀k]
pig
Nding t-úduru (NK:), *t-ə-t ̪t ̪əruk (RN:) Lumun t-út ̪t ̪rúk (NK:)
NA
Ko kùʈr̩̀, Tira úuʈʈúr, Werni (ù)uʈúr
NA
lùmù ‘hedgehog’
*g-+-aw
kâo
porcupine amumu (VE:)
NA
NA
rain
Torona kabɪk (NK:), *k-abɪk (NK:) Acheron ɡabɪk (NK:)
red
-ɔɽε (VE:, )
*-oode (TSd), *-ḍε *-UUre (NK:)
-óorè
rope
ðɔr (VE:)
*d̪-+-oorag (TSd), *d̪-+-ar *t ̪-ɔḷək (RN:)
lár
sheep
ðuŋga (VE: ), Tocho t ̪ʊ-ŋɡak (NK:)
*t ̪ʊ-ŋɡat ̪ (NK:) NA
káaŋàl
strength
ŋəma (VE:)
NA
ŋémà
sun
j-íŋgi/c-íŋgi (TSd:)
*ø-+-aŋin àaŋwòn *j-+-iŋgi (TSd), *c-ə-ŋɡi (NK:)
thief
pʊɽa (VE:, )
NA
NA
kwòoɽàm
thorn
túgi (Ngile, TSd:)
*d̪-+-a(g)i
*d-+-ugi
túukí [t̪úugí]
throat
ðəguru (VE:, )
NA
NA
tɔ́kwɽɔ̀ [ʈɔ́gwɽɔ̀]
tongue
ðuləŋε (VE:)
*d̪-+-(V)lVŋe (TSd), *t ̪ʊ-ləŋε (NK:)
*d̪-+-ŋela téŋlà
NA
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
*gu-+aaɽe
kwáaɽé
tree
b-ííɽâ/p-ííɽâ (TSd:)
*b-/g-+-ɪɪda (TSd), *p-ɪda (NK:)
two
Dengebu -έεɽá, Torona m-εɽa (NK:)
*-eedag (TSd), *C- NA εḍac (NK:)
-iɽɐn
vomit
uttε (HS)
*-VddV (TSd), *uk-dε (NK:)
ùɐ́ (: uɐtu [ùɐðù])
*-wey(TSa)
See note to Appendix for the reference system and for Proto-Heibanian and Koalib data. When a Heibanian language other than Koalib is mentioned, its name is explicitly mentioned and the data are mine. For Proto-Talodian, data come from Schadeberg (b) and Norton and Kuku Alaki (). 1
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Appendix : Plausible lexical cognates between Lafofa and Heibanian Lafofa1 Meaning
Heibanian ProtoHeibanian
Koalib
belly
t ̪-úur-i (TSd:)
*g-+-aare
káaré
bite
. . . jiε (TSd:)
*-iy-
yὲέ
bone
t-ʊ́ám-i (TSd:)
*li-+uya
lèá
elephant yuːŋi (TSd:)
*d-+-oŋor
tóoŋwór [t ̪óoŋwór]
fly
bai [β~b] (TSd:)
NA
kwáeó [kwáajó]
foot
l-ia-ga (TSd:)
UN
káakà [káagà] ‘foot’
hit
. . . biŋ (TSd:)
*-bid̪-
ìppí
*-awur
ŋɐ̀ɐkùr ‘left side’ (tɐ́ɐkúr [t ̪ɐ́ɐgúr] ‘left-handed person’)
left (side) kúlɪ (TSd:) rain
k-állɔ́-y (TSd:)
*g-+-aw
kâo
seed
t-ʊ́ʊl-ɪ (TSd:)
*li-+-UUla
lòolà
sheep
ßa:ŋi (TSd:)
NA
káaŋàl (Koalib)
sky
l-ʊ́ʊr-í (TSd:)
NA
lέεrέ
smoke (n.)
c-ʊʊr-í (TSd:)
*g-+-ulu
kwùulú
star
t-ɽɔ́ɔ́ (TSd:)
*li-+-UDam
t ̪òoròm
tongue
l-íáŋ-i (TSd:)
*d̪-+-ŋela
t̪éŋlà
tooth
t ̪-έεŋ-í (TSd:)
*li-+-ŋad̪
lèŋàt̪
vomit
lwâ-d̪aŋ, . . . lwa (TSd:)
*-wey-
ùɐ́
wet
b-ʊ́ɒji-lli (TSd:)
*-iigla
íklì [íglì]
woods
t-ɔ́ɔ́ɽɔ́ (TSd:)
NA
lúuɽí (: ţúuɽí) ‘piece of wood’
worm
kʊ-dʊŋ-í (TSd:)
*gu-+-ɔnd̪ɔ
kwɔ̀nţɔ́ [kwɔ̀nḑɔ́] (: kwɔ̀nţɔ́ŋé )
1
See note to Appendix for the reference system and for Proto-Heibanian and Koalib data.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Appendix : Plausible lexical cognates between Katloid, Rashadian and Talodian Katloid1
Rashadian
Talodian
Meaning
Languages
ProtoLanguages Rashadian
Proto-Talodian
bark(RN)2 Katla kúúr
Tagoi kówár, Tagom *g-ware warε (TSb:) (TSc:)
UN
UN
blood
Katla ija~iya (SV-:)
Rashad wiya (SV*ŋ-oya :), oyá (TSb:) (TSc:)
Tocho ŋ-íccɔk (TSd:)
*ŋ-+-ittsug (TSd:), *ŋɪccʊk (NK:)
bone
Tima kuh (SV:)
Tagoi kuh (TSb:) *g-usru (TSc:)
UN
UN
breast
Katla munji (SV- Orig tə́miɲ :), Tima (TSb:), Rashad kᵼ̀míndì (SB:) min (SV-:)
*d-miɲ (TSc:)
Talodi j-injᵼk *j-+-intsig (SV-:) (TSd:),3 *c-əmmik (NK:)
eye
Katla gᵼgöt ̗ (SV:)
Tagoi yígət (TSb:), Rashad (ŋ)gət (SV-:)
*y-ngid (TSc:)
*j-+-iig (TSd:), Lumun cí̧t (HS), Ngile *c-it (NK:) s-ígi (TSd:)
four(RN)
Tima ìhálə̀m (SB:)̀
Tagoi -arəm, Tagom *(h)aram (Q) áarám (TSb:), Tegali harom (TSc:)
goat4
Tima címìð (SB:)
Orig m̀bə́t, Tagoi/ Rashad mbᵼt (SV:)
good(RN) Tima -àmmɔ̀l, Tagoi -dɔ́mbɔ́r, -màl (SB:, ) Turjok -ambalɔ́, Tagom kamɔr
UN
UN
*mbᵼt (Q) Lumun imit (SV-:)
*w-+-emig (TSd:), *u-mit (NK:)
*-am(b)ɔr UN (Q)
UN
*g-aam (TSc:)
Tocho k-ʊ́ʊ́ɡaŋ (TSd:)
*d-+-ʊgaŋ (TSd:), *ŋən̪ (NK:)
hair
Katla gagam (: Tagoi/Orig ka(a)m agam) (SV-:), (TSb:), Rashad am (SV-:) Tima káàm (SB:)
head
Katla gas (SV:)
Tagoi/Orig káy, *g-aj Rashad as (SV-:) (TSc:)
Talodi j-as (SV-:)
*j-+-ats (TSd:), *c-ac (NK:)
tortoise
Tima k-ɔ̀rʊ̀l (AL:)
NA
Dagik kə́ɽɔ́ (VE:)
NA
NA
See note to Appendix for the reference system and for the sources of the data (including protolanguages). 2 This root for ‘bark’ may well be shared with Lafofa c-iíri (TSd:). However, it is distinct from another root attested in Heibanian, Talodian, and also Orig, a Rashadian language (see Table .). The existence of two roots for ‘bark’ might be the reflexion of a lexical doublet already present in Proto-Kordofanian. It could also be the case that these two roots refer to two different kinds of ‘bark’ which we are not able to distinguish, due to the lack of fine-grained lexicographic studies devoted to Kordofanian languages. 3 TSd () has j-+-instig but, according to the collected forms for ‘breast’ (TSd: ) and the comment on the form (“NG: mm < *nts ?”, p. ), instig is most probably a typographical error for j-+-intsig. 4 Note that for this item, Lafofa also displays a related form, εεmi ‘goat’ (TSd:). 1
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
A This work is partially supported by a public grant overseen by the French National Research Agency (ANR) as part of the progam “Investissements d’Avenir” (reference: ANR--LABX-), Labex EFL (Axe , Research Group RT). I also wish to acknowledge herein my dear colleagues Abeer Bashir, Roger Blench, Gerrit Dimmendaal, Harald Hammarström, Russell Norton, Thilo Schadeberg, Gertrud Schneider-Blum, Helen Smits, and John Vanderelst, who were kind enough to share with me such a wealth of advice, publications, and data. A heartfelt thank you to Claude Rilly, who read and commented on the entire first draft of this chapter. All remaining mistakes and imperfections are mine.
L A = central vowel; = attributive marker; CDP = dental or palatal consonant; CLVB = liquid or vibrant consonant belonging to the set /l, r, ɽ/; CN = nasal consonant; CP = palatal consonant; CV = velar consonant; DEM = demonstrative; DIM = diminutive; E = front vowel; F = feminine; IMP = imperative; M = masculine; N = noun class marker; O = back vowel / object; PFV = perfective; POSS = possessive; Q = reconstructed form proposed by Quint; (RN) = suggested by Russell Norton; T = dental/interdental or alveolar obstruent
A AL = Alamin (b); BSa = Bashir (); BSb = Bashir (); HS = Smits (); MS = Schadeberg and Elias (); NK = Norton and Kuku Alaki (); N = suggested by Russell Norton; SB = Schneider-Blum (); SV = Stevenson (/); TSa = Schadeberg (a); TSb = Schadeberg (); TSc = Schadeberg (ca. ); TSd = Schadeberg (b); VE = Vanderelst ()
R Alamin, S. (a). ‘Negation strategies in Tima’, Occasional Papers on Sudanese Languages : –. Alamin, S. (b). The Nominal and Verbal Morphology of Tima, a Niger-Congo language spoken in the Nuba Mountains. Cologne: Rüdiger Köppe. Alamin, S., Schneider-Blum, G., and Dimmendaal, G. (). ‘Finding your way in Tima’, in A. Mietzner and U. Claudi (eds.), Directionality in Grammar and Discourse: Case studies from Africa. Cologne: Rüdiger Köppe, –. Ali Karmal Kokko, S., and Quint, N. (forthcoming). Dictionnaire koalibfrançais. Paris: L’Harmattan. Anonymous- (c.?). Lessons in ŋirere. Typescript manual of Rere (ŋireɽe) comprising lessons. Anonymous- (c.?). Kawaleep Language, Yitam –. Books –. Abri(?). Bashir, Abeer M. A. (). A Phonetic and Phonological Study of the Tima Language. Ph.D. dissertation. Khartoum: University of Khartoum. Bashir, Abeer M. A. (). ‘A description of Tima sounds’, Adab : –. Bashir, Abeer M. A. (). ‘Noun class genders in Tagoi’, in G. Schneider-Blum, B. Hellwig, and G. J. Dimmendaal (eds.), –. Black, Mr. and -Mrs. K. (). The Moro Language; Grammar and dictionary. Khartoum: University of Khartoum.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Blench, R. (). ‘Splitting up Kordofanian’, in R. Blench and T. C. Schadeberg (eds.), –. Blench, R. (forthcoming a). ‘Do Heiban and Talodi form a genetic group and how are they related to Niger-Congo?’, in G. Schneider-Blum, B. Hellwig, and G. J. Dimmendaal (eds.), –. Blench, R. (forthcoming). ‘Kordofanian and Niger-Congo: new and revised lexical evidence’, http://www.rogerblench.info/Language/Niger-Congo/Kordofanian/Kordofanian%and%NigerCongo.pdf accessed May , . Blench, R., and MacDonald, K. (). The Origins and Development of African Livestock: Archaeology, genetics, linguistics, and ethnography. London and New York: Routledge. Blench, R., and Schadeberg, T. C. (eds.) (). Nuba Mountain Language Studies. Cologne: Rüdiger Köppe. Creissels, D. (). Description des langues négro-africaines et théorie syntaxique. Grenoble: Ellug. Creissels, D. (). ‘Typology’, in B. Heine and D. Nurse (eds.), –. Dimmendaal, G. J. (). ‘Where have all the noun-classes gone in Tima?’, in C. de Féral, M. Kossmann, and M. Tosco (eds.), In and Out of Africa: Languages in question. In honour of Robert Nicolaï. Volume : Language contact and language change in Africa. Louvain: Peeters, –. Dimmendaal, G. J. (). ‘Accretion zones and the absence of language union’, in G. J. Dimmendaal, The Leopard’s Spots: Essays on language, cognition and culture. Leiden: Brill, –. Elles, R. J. (). ‘The Kingdom of Tegali’, Sudan Notes and Records : –. Ewald, J. J. (). Soldiers, Traders and Slaves: State formation and economic transformation in the Greater Nile Valley, –. Madison, WI: University of Wisconsin Press. Greenberg, J. H. ( []). The Languages of Africa. Bloomington, IN: Indiana; The Hague: Mouton. Hammarström, H. (). ‘Noun class parallels in Kordofanian and Niger-Congo: evidence of genealogic inheritance?’, in R. Blench and T. C. Schadeberg (eds.), –. Heine, B., and Nurse, D. (eds.) (). African Languages: An introduction. Cambridge: Cambridge University Press. Hellwig, B. (). ‘Verbal morphology in Katla’, in R. Blench and T. C. Schadeberg (eds.), –. Hyman, L. (forthcoming). ‘Nuba Mountain verb extensions in African perspective’, in A. Jakobi et al. (eds.). Jakobi, A., Manfredi, S., Rose, Sh., and Quint, N. (eds.) (forthcoming). Proceedings of the nd Nuba Mountain Language Conference, Paris, August –, . Cologne: Rüdiger Köppe. Jenks, P., and Rose, Sh. (). ‘High tone in Moro: effects of prosodic categories and morphological domains’, Natural Language and Linguistic Theory : –. MacDiarmid, P. A., and MacDiarmid, D. N. (). ‘The languages of the Nuba Mountains’, Sudan Notes and Records : –. Meinhof, C. (/ []). ‘Sudansprachen und Hamitensprachen’, Zeitschrift für Kolonialsprachen : –. Meinhof, C. (– []). ‘Sprachstudien im egyptischen Sudan’, Zeitschrift für Kolonialsprachen (/) –, –; (/) –, –, –, –; (/) –, –; (/) –, –, – [Vaduz: Kraus Reprint]. Meinhof, C. (/ []). ‘Das Heiban in Kordofan’, Zeitschrift für Eingeborenen-Sprachen : –. Norton, R., and Kuku Alaki, T. (). ‘The Talodi languages: a comparative-historical analysis’, Occasional Papers in the Study of Sudanese Languages : –. Quint, N. (a). ‘Do you speak Kordofanian?’, in Proceedings of the th International Sudan Studies Conference, April th–th, . Bergen: Universitetet i Bergen [CD-ROM]. Quint, N. (b). Phonologie de la langue koalibe. Paris: L’Harmattan. Quint, N. (). The Phonology of Koalib, a Kordofanian language from the Nuba Mountains (Sudan). Cologne: Rüdiger Köppe. Quint, N. (a). ‘Benefactive and malefactive verb extensions in the Koalib verb system’, in S. Kittilä and F. Zúñiga (eds.), Benefactives and Malefactives: Typological perspectives and case studies. Amsterdam and Philadelphia, PA: John Benjamins, –. Quint, N. (b). ‘La littérature orale en koalib (Sud-Kordofan)’, Études littéraires africaines : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Quint, N. (). ‘Integration of borrowed nouns in Koalib, a noun class language’, in R. Blench and T. C. Schadeberg (eds.), –. Quint, N. (). ‘An assessment of the Arabic lexical contribution to contemporary spoken Koalib’, in S. Manfredi and M. Tosco (eds.), Arabic in Contact. (coll. Arabic Linguistics) Amsterdam and Philadelphia, PA: John Benjamins, –. Quint, N. (forthcoming). ‘Djifanghor Nyun (Bainouck)’, in F. Lüpke (ed.), Oxford Guide to the World’s Languages: Atlantic. Oxford: Oxford University Press. Rose, Sh., Ackerman, F., Gibbard, G., Jenks, P., Kertz, L., and Rohde, H. (). ‘Wh-question constructions in Moro’, Journal of African Languages and Linguistics : –. Rüppell, E. (). Reisen in Nubien, Kordofan und dem peträischen Arabien vorzüglich in geographisch-statisticher Hinsicht. Frankfurt am Main: Friedrich Wilmans. Schadeberg, T. C. (c.). The Rashad Group [Outline of a volume on the Rashadian languages]. Unpublished typescript. Schadeberg, T. C. (a). A Survey of Kordofanian. Vol. : The Heiban Group. Hamburg: Helmut Buske. Schadeberg, T. C. (b). A Survey of Kordofanian. Vol. : The Talodi Group. Hamburg: Helmut Buske. Schadeberg, T. C. (c). ‘The classification of the Kadugli language group’, in T. C. Schadeberg and L. M. Bender (eds.), Nilo-Saharan: Proceedings of the First Nilo-Saharan Linguistics Colloquium, Leiden, September –, . Dordrecht and Cinnaminson, NJ: Foris, –. Schadeberg, T. C. (). ‘Kordofanian’, in J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham, MD, New York, and London: University Press of America, –. Schadeberg, T. C. (). ‘Comparative Kadu wordlists’, Afrikanistische Arbeitspapiere : –. Schadeberg, T. C. (). ‘Rashad survey data’, in R. Blench and T. C. Schadeberg (eds.), –. Schadeberg, T. C., and Elias, P. (). A Description of the Orig Language (Southern Kordofan), based on the notes of Fr. Carlo Muratori. Tervuren: Musée Royal de l’Afrique Centrale. Schadeberg, T. C., and Kossmann, M. (). ‘Participant reference in the Ebang verbal complex (Heiban, Kordofanian)’, Journal of African Languages and Linguistics : –. Schneider-Blum, G. (). A TimaEnglish Dictionary. Cologne: Rüdiger Köppe. Schneider-Blum, G., Hellwig, B., and Dimmendaal, G. J. (eds.) (). Nuba Mountain Language Studies. New Insights. Cologne: Rüdiger Köppe. Seligman, B. Z. (/ []). ‘Note on the language of the Nubas of southern Kordofan’, Zeitschrift für Kolonialsprachen : –. Smits, H. (). ‘The locative-applicative suffix in Lumun’, in R. Blench and T. C. Schadeberg (eds.), –. Smits, H. (). Lumun List. Unpublished material. Stevenson, R. C. (). The Tira Language. Typescript. Stevenson, R. C. (). The Otoro Language. Typescript. Stevenson, R. C. (/). ‘A survey of the phonetics and grammatical structure of the Nuba Mountain languages’, Afrika und Übersee : –; : –; : –. Stevenson, R. C. (). Tira and Otoro: Two Kordofanian grammars. Edited by T. C. Schadeberg. Cologne: Rüdiger Köppe. Tucker, A. N., and Bryan, M. A. (). ‘The Koalib-Tagoi (class) languages’, in A. N. Tucker and M. A. Bryan (eds.), Linguistic Analyses: The non-Bantu languages of north-eastern Africa. London, New York, and Cape Town: Oxford University Press, –. Tutshek [Tutschek], L. (). ‘On the Tumali language’, Proceedings of the Philological Society for – and – (): –. Tutshek [Tutschek], L. (). ‘On the Tumali alphabet’, Proceedings of the Philological Society for – and – (): –. Vanderelst, J. (). A Grammar of Dagik, a Kordofanian language of Sudan. Cologne: Rüdiger Köppe. Williamson, K., and Blench, R. (). ‘Niger-Congo’, in B. Heine and D. Nurse (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
- ......................................................................................................................
A-A (or Afroasiatic; also known as “Hamito-Semitic”) is an entity of genetically related languages which is often labeled a macro-family or language phylum due to the number and typological diversity of its member languages and the chronological depth of this entity. In any case the choice of the label is arbitrary and depends on taxonomic preferences of particular schools and individual authors within the framework of genealogical classification, since in this sphere there is no generally accepted and scientifically based taxonomy. The name of this language phylum has a long history. The term “Semitic” was proposed by Leibniz in (French translation: Leibniz ). He was followed by German orientalist Schlözer1 (: ) who wrote: From the Mediterranean Sea to the Euphrates and from Mesopotamia down to Arabia, as is known, only one language reigned. The Syrians, Babylonians, Hebrews and Arabs were one people. Even the Phoenicians who were Hamites spoke this language, which I might call Semitic.
The term “Semitic” was derived from the name of the elder son of Noah, Shem (Sem), who was the eponymous ancestor of certain peoples speaking the languages in question according to the biblical genealogy of peoples (Gen. ). Following Schlözer, Renan () introduced the term “Hamitic” (chamitique in French) for languages that exhibit common features with Semitic languages but are too different from them to be included in the Semitic entity, namely Ancient Egyptian, Coptic, and Berber. Renan also used the term “Cushitic” for the non-Semitic languages spoken in Ethiopia and the Horn of Africa. At first, these languages were labeled Ethiopian, but later this name was applied only to Semitic languages of Ethiopia. According to the biblical genealogy of peoples, Ham (Kam, Kham) is the second son of Noah, and Cush and Mizraim (Egypt in biblical Hebrew) are the sons of Ham. Cush (Kush) in Ancient Near East roughly corresponded to Nubia.
1
Incidentally, Schlözer spent several years in St. Petersburg and became one of the founding fathers of the academic ancient history of Russia.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Thus, Ancient Egyptian and Cushitic fit well into the biblical genealogy as members of the Hamitic entity. It must be noted that the biblical genealogy of peoples may in no way be considered linguistically based; rather, it reflects the geopolitical realities of the epoch, i.e. Canaan, according to this genealogy, is the son of Ham, while Elam is the son of Sem.2 Quite logically, from the s the term “Hamito-Semitic” (henceforth H-S; also “SemitoHamitic”) became widely used; among the pioneers in the field, R. Lepsius, L. Reinisch, T. Benfey, and F. Müller must be mentioned. Further development of H-S studies mainly consisted in the rapid growth of relevant linguistic data and the number of languages considered H-S. All languages showing certain parallels with but not sufficient parallels to classify them as Semitic, were treated as “Hamitic” within the H-S entity. In this context the problem of the affiliation of Hausa and Chadic languages in general became a most important issue. Already in the mid-nineteenth century the idea of linking Chadic with H-S was expressed by Newman (), and this hypothesis was taken up by other pioneers of H-S studies. It is necessary to emphasize here that Chadic languages form a unique group as to its composition, including more than languages and dialect clusters with rather limited numbers of speakers, plus Hausa—the second biggest language of Africa south of Sahara after Swahili—and possessing a very high functional status in comparison with all other Chadic languages. Hausa shows various evident morphological parallels with classical Semitic languages and Ancient Egyptian, first of all in personal pronouns and preverbal markers. It may be surmised that if the sociolinguistic role of Hausa had been played by one of the Central Chadic languages, where isoglosses of this kind are mostly absent, the study of Chadic in the H-S context would probably have been delayed for decades. The increase in linguistic data made it clear that the languages traditionally classified as “Hamitic” do not form a coherent linguistic unit within H-S, the only common feature being their not belonging to the Semitic family. Following scholars supporting this idea (Cohen and others), Greenberg (, ) in his studies of African linguistic classification argued that, since no “Hamitic” entity exists, all language groups included in it should be considered as independent branches within H-S on a par with Semitic languages. As a next step he proposed eliminating the term “Hamitic” from genealogical linguistic nomenclature and replacing H-S with “Afro-Asiatic” (AA), as this is the only important language unit embracing both Asia and Africa. However, this term is not very convenient because it may mean all the languages of Asia and Africa. Therefore, Diakonoff () proposed changing it to “Afrasian”, which is widely used in Russian studies and also by some Anglophone authors. Yet other variants proposed by certain scholars—“Erythraic”, “Lisramic”, “Noahitic”—were not accepted by the linguistic community. At present, both variants are in wide use: AA and H-S or “Semito-Hamitic” (following European academic traditions), often on a par with AA. It may be noted that today the reason for eliminating the term “Hamitic” is not as valid as it was in the s, when it implied a dichotomy into Semitic and Hamitic branches. The label H-S is just a conventional one, analogous to “Indo-European” or “Indogermanisch”; nobody would consider the latter a binary unit consisting of Indian and European (or even Indian and Germanic) language groups.3 2
Canaanite and Ancient Hebrew are closely related languages; hence both are Semitic, while Elamite is not included in the H-S unit at all. 3 For more details and bibliographical references on the history of H-S studies, see Hodge () and Cohen ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
The AA language phylum is one of the most intriguing and complex issues within the framework of comparative linguistics. It includes languages with ancient written traditions (Ancient Egyptian, Akkadian, Eblaite, Amorite, Phoenician, Ugaritic, etc.) as well as dozens of unwritten languages in sub-Saharan Africa. The chronological gap between classical dead and modern unwritten languages is uncontested in historical linguistics. Moreover, and in contrast to Indo-European, the ancient written monuments are attested only in two groups of AA: Semitic and Egyptian. The paradigm of comparative linguistics (in the sense of Kuhn ) was drawn up in the s within Indo-European studies by a group of scholars known as neogrammarians. However, non-paradigmatic conceptions in the field of linguistic reconstruction and genetic classification played an important part in AA studies, especially the attempts to explain lexical, phonetic, and grammatical parallels within AA by language mixing as a result of areal contacts and ancient processes of convergence. One must not forget that, in the mentality of the nineteenth and the first decades of the twentieth century it was very hard to accept the idea of a common origin of the languages of the great ancient civilizations and world religions on the one hand and “minor” unwritten languages, as spoken in western and central Sudan, on the other. This approach reached its apogee in the work of the distinguished German Africanist Meinhof (). He built his Hamitic hypothesis on the Hamitic theory as well known from the field of physical and social anthropology and history, according to which pastoralists of Hamitic origin brought to Black Africa a highly developed civilization and then mixed with local people. Meinhof surmised that “Hamitic” languages in West Africa emerged as a result of the mixture of the languages of incoming pastoral Hamites with local languages. Following Lepsius (), he emphasized the role of grammatical gender as a key linguistic criterion for classifying a language as “Hamitic”. According to the neogrammarian paradigm, however, typological criteria are not appropriate for postulating linguistic affinity, even though in Bantu comparative linguistics nominal classification is traditionally considered obligatory for classifying a language as Bantu (cf. Guthrie –). Most probably Meinhof, as a pioneer of Bantu comparative studies (Meinhof ), had chosen grammatical gender as a primary criterion for his classification of “Hamitic” because of its typological parallelism to the nominal classification of Bantu. Hence, he classified as “Hamitic” some Chadic languages possessing grammatical gender, and labeled this group as “Chado-Hamitic”, while other languages obviously related to them were classified as Chadic and non-H-S. On the basis of the nominal classification as the only linguistic criterion, he also included in his “Hamitic” family Fula (Peul), part of the Nilotic languages (referred to as “Nilo-Hamitic”), and some other languages of East and southern Africa. Meinhof ’s nomenclature partially survived into the mid-twentieth century, when Greenberg published his classification of H-S and later changed its name to AA. Greenberg also excluded from this unit all languages classified by Meinhof as “Hamitic” on the mere basis of the typological criterion of grammatical gender or nominal classification. Other scholars, albeit non-supporters of Meinhof ’s approach, were reluctant to consider Chadic languages to be H-S. Such an attitude was somehow preserved until the s, when Chadic languages were included in a volume on sub-Saharan languages (see Manessy ) but not in the H-S volume of the same series (see Cohen ). The evolution of the status of Chadic within AA may be described as centripetal; the position of these languages as a self-standing unit within AA is no longer contested.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
In contrast to this the evolution of the genealogical status of Cushitic languages was centrifugal. The idea of excluding Bedauye (Beja) from Cushitic, while keeping it in AA, got no support. But the hypothesis of considering West Cushitic as an independent branch (i.e. Omotic) of AA (Fleming ; Bender , ) is now supported by the majority of experts. Omotic is even considered by some scholars to be the most divergent branch of AA, while some others still support the idea of its Cushitic affiliation. This controversy reflects a general theoretical problem of genetic classification within the neogrammarian paradigm. The genealogical tree remains a sound and well-established model for presenting diachronic splitting within a genetic entity of languages. Postulating such an entity (absolute classification) as well as most closely interrelated language groups on the highest levels of the genealogical tree does not normally provoke much doubt. However, splitting on the medium level often remains obscure, since it may be very difficult or even impossible to unambiguously distinguish between genetically relevant isoglosses and areal ones induced by contacts between related languages. Thus, it may be concluded that in the present state of research, the AA language phylum consists of five or six branches: Semitic, Egyptian, Berber, Cushitic, Omotic (or a single Cush-Omotic), and Chadic (see Map .). As to the external links of AA, the idea of relating Semitic to Indo-European was first proposed at the beginning of the twentieth century and was later developed for the whole AA phylum within the framework of Nostratic theory, bringing together several language families and phyla: Indo-European, AA, Uralic, Altaic, Kartvelian, and Dravidian. The further development of Nostratic studies may help to resolve the problem of the AA
. The geographical distribution of Afro-Asiatic languages
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Urheimat (original homeland). At present the AA-internal arguments indicate eastern Sahara and adjacent territories as the most probable AA Urheimat, while Nostratic theory makes the Fertile Crescent hypothesis more probable. According to different hypotheses, the possible time span for the differentiation of Proto-AA varies from , to , ; a more realistic estimate was provided by Diakonoff (: ): “the speakers of Egyptian were the first to break away from the basic Proto-Afrasian nucleus not later than the th millennium B.C.” The present state of the art in comparative AA linguistics may be characterized as follows: regular sound correspondences and important and diverse morphological isoglosses, as well as a rich common lexicon, create a solid base for future research. Despite its long history, however, AA comparative linguistics is still in its infancy. The few existing comparative vocabularies of AA are far from standard norms. A comprehensive monographic overview of AA has been published by Diakonoff (). The comparison of this study with its earlier version (Diakonoff ) makes it possible to appreciate the changes in methodological approaches and main trends in AA comparative studies. Further progress is to be expected from comparative studies and linguistic reconstructions on lower levels of all AA groups and subgroups. In view of the enormous chronological gap between various branches and the time-depth of Proto-AA, only stepby-step reconstructions on intermediate levels may lead to conclusive results. A large number of languages are not documented and analyzed at all; for many others only short wordlists and very brief grammatical notes are available; thus, field research and documentation of non-described languages have absolute priority in AA linguistics.
R Bender, M. L. (). Omotic: A new Afroasiatic language family. Carbondale, IL: Southern Illinois University. Bender, M. L. (). The Non-Semitic Languages of Ethiopia. East Lansing, MI: African Studies Center, Michigan State University. Cohen, D. (a). ‘Introduction’, in D. Cohen (ed.) (b), –. Cohen, D. (b). ‘Le Chamito-sémitique’, in D. Cohen (ed.), Les langues dans le monde ancien et moderne. Part III. Paris: Éditions du CNRS, –. Cohen, M. (). ‘Langues chamito-sémitiques’, in A. Meillet and M. Cohen (eds.), Les langues du monde. Paris: Editions du CNRS, –. Diakonoff, I. M. (). Semito-Hamitic Languages: An essay in classification. Moscow: Nauka. Diakonoff, I. M. (). Afrasian Languages. Moscow: Nauka. Fleming, H. C. (). ‘Classification of West Cushitic within Hamito-Semitic’, in D. F. McCall, N. R. Bennett, and J. Butler (eds.), Eastern African History. New York, Washington, DC, and London: Frederick A. Praeger, –. Greenberg, J. H. (). Studies in African Linguistic Classification. New Haven, CT: Compass Publishing. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Guthrie, M. (–). Comparative Bantu: An introduction to the comparative linguistics and prehistory of the Bantu languages. vols. Farnborough: Gregg. Hodge, C. T. (). ‘Afroasiatic: an overview’, in T. A. Sebeok (ed.), Linguistics in South West Asia and North Africa. (Current Trends in Linguistics, .) The Hague and Paris: Mouton, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Kuhn, T. S. (). The Structure of Scientific Revolutions (nd enlarged edn, ). Chicago: University of Chicago Press. Leibniz, G. W. (). ‘Brevis designatio meditationum de Originibus Gentium, ductis potissimum ex indicio linguarum’, in G. W. Leibniz, Miscellanea Berolinensia ad incrementum scientiarum ex scriptis Societatis Regiae Scientiarum exhibitis edita : –. Leibniz, G. W. (). ‘Bref essai sur l’origine des peuples déduite principalement des indications fournies par les langues’, in G. W. Leibniz (ed.), L’Harmonie des langues. Paris: Points, –. Lepsius, R. (). Standard Alphabet for Reducing Unwritten Languages and Foreign Graphic Systems to a Uniform Orthography in European Letters, nd edn. London: Williams and Norgate. Manessy, G. (). ‘Les langues de l’Afrique noire’, in J. Perrot (ed.), Les langues dans le monde ancien et moderne. Part I. Les langues de l’Afrique subsaharienne. Paris: Éditions du CNRS, –. Meinhof, C. (). Grundzüge einer vergleichenden Grammatik der Bantusprachen. Berlin: Dietrich Reimer. Meinhof, C. (). Die Sprachen der Hamiten. Hamburg: L. Friederichsen. Newman, T. N. (). ‘Remarks on the Hausa language’, in J. C. Prichard (ed.), Researches into the Physical History of Mankind. Vol. , rd edn. London, –. Renan, E. (). Histoire générale et système comparé des langues sémitiques. Paris: Imprimerie impériale. Schlözer, A. L. von. (). ‘Von den Chaldäern’, Repertorium für Biblische und Morgenländische Litteratur : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
ˊ . -
. I
.................................................................................................................................. E is a branch of Afro-Asiatic that consists of only one language. Ancient Egyptian was spoken in the Lower (Northern) Nile Valley and in the Nile Delta; it was the written language of the Ancient Egyptian civilization from the appearance of writing in the late fourth millennium . Its latest form, Coptic, was the language of Egyptian Christians until the tenth century , when it was replaced by the Semitic language Arabic; however, to this day Coptic remains the liturgical language of the Coptic Church in Egypt. (Copts are a minority of approximately % of the population of modern Egypt.) Egyptian is unique among the world’s languages in being documented over a time period of more than four millennia, the longest period of such documentation known to research. Egyptian shares many basic archaic features with other Afro-Asiatic languages, like the system of verbal roots (lexical meaning is carried by a consonant skeleton, whilst different patterns of vowels are the means of expressing mood and voice, and also of forming verbal nouns), a dual gender system (masculine vs. feminine), and common, etymologically related sets of pronouns. Egyptian roots consist in most cases of two to four consonants, although roots with fewer or more also occur. Stems (derivation of verbs by root extension and different vowel patterns) are common in Afro-Asiatic, but only two are found in Ancient Egyptian: in addition to the basic stem, the causative form with prefixed ś- (which became unproductive early). Regarding typology, Egyptian was originally a VSO language; however, SVO structures developed at a relatively early stage as well. Egyptian completely lacks the prefix conjugation, which is an archaic feature in Semitic, Cushitic, and Berber languages; on the other hand, it practices suffix conjugation like Semitic and Berber to express the stative (Peust ). There are a few traces of what could be remnants of noun declension, though it is hardly a functioning system during the written phases of the language.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˊ . -
. L
.................................................................................................................................. The earliest written evidence appears around in Abydos, in a royal tomb marked as U-j (for the history of writing in Egypt, see Woods et al. : –). Following a short hiatus, writing reappears in the so-called Dynasty (about ), and spreads throughout Egypt during the First and Second Dynasties (late fourth and early third millennium ). The purpose of these earliest inscriptions is data storage (recording names of persons, objects, years, etc.); only at the end of the Second Dynasty does writing become a medium of communication, with the appearance of the first short texts (Kahl ). Written evidence until the Third Dynasty has only been analyzed lexicographically (Kahl et al. –). Longer inscriptions are recorded first during the Fourth Dynasty (mid-third millennium ). The Pyramid Texts, the earliest known religious texts, constitute a special case, as, although they appear at the end of the Fifth Dynasty, their grammar is much more archaic than that of contemporary inscriptions. On the basis of grammar, Ancient Egyptian can be divided into different phases (cf. Allen ): A. Earlier Egyptian . Old Egyptian (Fourth to Tenth Dynasty, c.– ) . Middle or Classical Egyptian (Eleventh to Eighteenth Dynasty, c.– ) B. Later Egyptian . Late Egyptian (Nineteenth to Twenty-Fifth Dynasty, c.– ) . Demotic (c. –mid-fifth century ) . Coptic (from the second century ) In Earlier Egyptian there is a predominance of synthetic structures, and it can be characterized as a flexional language. Though it has an accusative profile (objects are the marked elements in transitive clauses, and the only known valence reduction is the passive), pronouns in intransitive verbal clauses can be analyzed in terms of split ergativity (split-S or active type; see Dixon ; cf. also Satzinger ). A considerable grammatical change is evident in Late Egyptian, with the appearance of many analytical features (auxiliary verbs; definite, indefinite, and possessive articles). In Coptic, these originally independent elements have merged with the verbs and the nouns to form polysynthetic structures as the language has become fusional (again). There is a considerable degree of overlapping between the different phases in the written sources. Middle (or Classical) Egyptian was used until the end of the fourth century (Jansen-Winkeln ; Engsheden ; Kurth ). According to one theory, parallel usage of Classical and Late Egyptian is not the consequence of active language competence of various registers but rather the result of reliance on archived text templates during text composition (von Lieven : –). Demotic and Coptic are also notable for being contemporaneous for centuries, the only difference in the second century being the choice of writing, indigenous versus Greek-based (which, however, from the third century on also means pagan versus Christian).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Synchronic geographical variation can be observed only in the case of Coptic—Sahidic, Akhmimic, Lycopolitan, and Bohairic being the most important dialects. (Mini-grammars for all Coptic dialects, written by Rodolphe Kasser, Peter Nagel, Helmut Satzinger, Ariel Shisha-Halevy, and Hans-Martin Schenke, can be found in volume of Atiya .)
. P
.................................................................................................................................. Regarding the phonological interpretation of Egyptian graphemes, two theories have been proposed. The main difference lies in the interpretation of the plosive opposition; in traditional phonology it is understood as voiced/voiceless (Vergote ). In an influential article published in (Rössler : –), Otto Rössler argued for an emphatic (i.e. ejective or post-glottalized) vs. non-emphatic opposition, and additionally also in Old Egyptian for a different value of the consonant traditionally transliterated as pharyngeal fricative (the cayin). A traditional phonology is Takács (); on the other hand, Peust () is Rösslerian. The debate is far from being resolved: a recent critical review of the c ayin theory is Gensler (); a compromise between the two systems is proposed in Satzinger ().
. M
.................................................................................................................................. In the earliest days of Egyptology, traditional European and Semitic grammar served as a model for the interpretation of the morphosyntax of Earlier Egyptian, particularly Arabic, whose features were considered the most archaic within Semitic. Opposing this approach, based on structuralism, Hans-Jakob Polotsky (in a series of articles beginning in ) developed an alternative model that was later labeled “Standard Theory” (Polotsky : –, –, –). According to this, most verbal roots are realized in actual speech only as nouns, adjectives, or adverbs, in processes called transpositions, and therefore only nominal, adjectival, or adverbial clauses exist in Egyptian (see Junge ). Subsequent research led to the realization that Standard Theory has to be either revised or abandoned (Loprieno ). Reference grammars of Old Egyptian are Schweitzer () (Fourth Dynasty, revised Standard Theory), Edel (–) (Fifth and Sixth Dynasties, traditional), and Allen () (verbal system of the Pyramid Texts, Standard Theory). Although conceived as an introduction, the monumental Gardiner () remains the most important traditional grammar of Classical Egyptian for its wealth of detail. Recent Middle Egyptian reference grammars and textbooks offer various degrees of modification of the Standard Theory (Hoch ; Malaise and Winand ; Schenkel ; Ockinga ; Allen ), or its rejection (Ritter ; Graefe and Kahl ; Borghouts ). In the same articles, in which the Standard Theory was proposed, Polotsky also offered a descriptive model of the Coptic verbal system that has been employed as the basis for all phases of Later Egyptian, i.e. Late Egyptian (Winand ; Černý and Groll ;
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˊ . -
Neveu ; Junge ), Demotic (Johnson , ; Simpson ), and Coptic (for Sahidic: Brankaer ; Layton ; for Bohairic: Shisha-Halevy ). For Sahidic Coptic, Polotsky (/) is a further refinement of this theory upon which in turn some of the grammars are based (Plisch ; Eberle ). On the other hand, Reintges (a) abandons the Polotskyan framework and terminology (cf. also Reintges b).
. L
.................................................................................................................................. Egyptian lexicon (Erman and Grapow –; Wilson ; Hannig , , ) displays, unsurprisingly, a good number of words with cognates in the Afro-Asiatic languages (Takács –). During the course of its history many phonological and semantic changes occurred (Late Egyptian: Lesko –; Demotic: Johnson ; Coptic: Crum ; Coptic etymological dictionaries: Westendorf –; Černý ; Vycichl ). From the second half of the second millennium hundreds of Semitic words were borrowed from Canaanite (and later also Aramaic) languages (Hoch ; for Semitic words in Coptic, see Quack ). Approximately one hundred Greek loanwords entered Demotic; in the Coptic lexicon the number rises to %, though in written Coptic texts the average percentage is only (cf. Kasser in Atiya , vol. : –; Förster ; for language contacts, see Schneider et al. ; Hasznos ).
R Allen, J. P. (). The Inflection of the Verb in the Pyramid Texts. Malibu, CA: Undena Publications. Allen, J. P. (). The Ancient Egyptian Languages: A historical study. Cambridge: Cambridge University Press. Allen, J. P. (). Middle Egyptian: An introduction to the language and culture of hieroglyphs, rd revised and reorganized edn. Cambridge: Cambridge University Press. Atiya, A. S. (ed.) (). The Coptic Encyclopedia. Vols –. New York and Toronto: Macmillan and Collier Macmillan. Borghouts, J. F. (). Egyptian: An introduction to the writing and language of the Middle Kingdom. vols. Leiden and Leuven: Nederlands Instituut voor het Nabije Oosten–Peeters. Brankaer, J. (). Coptic: A learning grammar. Wiesbaden: Otto Harrassowitz. Černý, J. (). Coptic Etymological Dictionary. Cambridge: Cambridge University Press. Černý, J., and Groll, S. I. (). A Late Egyptian Grammar, th edn with additions. Rome: Biblical Institute Press. Crum, W. E. (). A Coptic Dictionary. Oxford: Clarendon Press. Dixon, R. M. W. (). Ergativity. Cambridge: Cambridge University Press. Eberle, A., with Schultz, R. (). Koptisch: Ein Leitfaden durch das Saidische. Munich: LINCOM. Edel, E. (–). Altägyptische Grammatik. vols. Rome: Pontificium Institutum Biblicum. Engsheden, Å. (). La reconstitution du verbe en égyptien de tradition – avant J.-C. Uppsala: Uppsala University. Erman, A., and Grapow, H. (–). Wörterbuch der ägyptischen Sprache. vols. Berlin: AkademieVerlag. Förster, H. (). Wörterbuch der griechischen Wörter in den koptischen dokumentarischen Texten. Berlin and New York: Walter de Gruyter.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Gardiner, A. (). Egyptian Grammar: Being an introduction to the study of hieroglyphs, rd rev. edn. Oxford: Griffith Institute. Gensler, O. D. (). ‘A typological look at Egyptian *d > ʕ’, in E. Grossman, M. Haspelmath, and T. S. Richter (eds.), Egyptian-Coptic Linguistics in Typological Perspective. Berlin, Munich, and Boston, MA: Walter de Gruyter, –. Graefe, E., and Kahl, J. (). Mittelägyptische Grammatik für Anfänger, th edn. Wiesbaden: Otto Harrassowitz. Hannig, R. (). Die Sprache der Pharaonen: Großes Handwörterbuch Deutsch-Ägyptisch (– v. Chr.). Mainz: Philipp von Zabern. Hannig, R. (). Ägyptisches Wörterbuch. I. Altes Reich und Erste Zwischenzeit. Mainz: Philipp von Zabern. Hannig, R. (). Ägyptisches Wörterbuch. II. Mittleres Reich und Zweite Zwischenzeit. Part and . Mainz: Philipp von Zabern. Hasznos, A. (). Graeco-Coptica: Greek and Coptic clause patterns. Wiesbaden: Otto Harrassowitz. Hoch, J. E. (). Semitic Words in Egyptian Texts of the New Kingdom and Third Intermediate Period. Princeton, NJ: Princeton University Press. Hoch, J. E. (). Middle Egyptian Grammar. Mississauga: Benben Publications. Jansen-Winkeln, K. (). Spätmittelägyptische Grammatik der Texte der . Zwischenzeit. Wiesbaden: Otto Harrassowitz. Johnson, J. H. (). Thus Wrote cOnchsheshonqy: An introductory grammar of Demotic, nd rev. edn. Chicago, IL: Oriental Institute. Johnson, J. H. (). The Demotic Dictionary of the Oriental Institute of the University of Chicago, http://oi.uchicago.edu/research/pubs/catalog/cdd/ accessed September , . Johnson, J. H. (). The Demotic Verbal System, nd printing with corrections. Chicago, IL: Oriental Institute. Junge, F. (). Syntax der mittelägyptischen Literatursprache: Grundlagen einer Strukturtheorie. Mainz: Philipp von Zabern. Junge, F. (). Einführung in die Grammatik des Neuägyptischen, rd edn. Wiesbaden: Otto Harrassowitz. Kahl, J. (). Das System der ägyptischen Hieroglyphenschrift in der .–. Dynastie. Wiesbaden: Otto Harrassowitz. Kahl, J., Bretschneider, M., and Kneissler, B. (). Frühägyptisches Wörterbuch. Wiesbaden: Otto Harrassowitz. Kurth, D. (). Einführung in das Ptolemäische: Eine Grammatik mit Zeichenliste und Übungsstücken. Part and . Hützel: Backe. Layton, B. (). A Coptic Grammar with Chrestomathy and Glossary: Sahidic dialect, rd edn. Wiesbaden: Otto Harrassowitz. Lesko, L. H. (–). A Dictionary of Late Egyptian. Vols –. Providence, RI : B. C. Scribe Publications. Loprieno, A. (). Ancient Egyptian: A linguistic introduction. Cambridge: Cambridge University Press. Malaise, M., and Winand, J. (). Grammaire raisonnée de l’égyptien classique. Liège: Centre Informatique de Philosophie et Lettres. Neveu, F. (). La langue des Ramsès: grammaire du néo-égyptien. Paris: Khéops. Ockinga, B. G. (). A Concise Grammar of Middle Egyptian: An outline of Middle Egyptian grammar, rd rev. edn. Mainz: Philipp von Zabern. Peust, C. (). Egyptian Phonology: An introduction to the phonology of a dead language. Göttingen: Peust & Gutschmidt. Peust, C. (). ‘On the subgrouping of Afroasiatic or: how to use an unrooted phylogenetic tree in historical linguistics’, Lingua Aegyptia : –. Plisch, U.-K. (). Einführung in die koptische Sprache. Wiesbaden: Ludwig Reichert. Polotsky, H. J. (). Collected Papers. Jerusalem: Magness.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˊ . -
Polotsky, H. J. (/). Grundlagen des koptischen Satzbaus. vols. Decatur and Atlanta, GA: American Society of Papyrologists. Quack, J. F. (). ‘Zu den vorarabischen semitischen Lehnwörtern im Koptischen’, in B. Burtea, J. Tropper, and H. Younansardaroud (eds.), Studia Semitica et Semitohamitica: Festschrift für Rainer Voigt anläßlich seines . Geburtstages am . Januar . Münster: Ugarit-Verlag, –. Reintges, C. H. (a). Coptic Grammar (Sahidic Dialect): A learner’s grammar. Cologne: Rüdiger Köppe. Reintges, C. H. (b). ‘Second tenses don’t exist!’, in M. Immerzeel and J. van der Vliet (eds.), Coptic Studies on the Threshold of a New Millennium: Proceedings of the Seventh International Congress of Coptic Studies, Leiden, August– September . Volume . Leuven, Paris, and Dudley, MA: Peeters and Departement Oosterse Studies, –. Ritter, T. (). Das Verbalsystem der königlichen und privaten Inschriften: XVIII. Dynastie bis einschließlich Amenophis III. Wiesbaden: Otto Harrassowitz. Rössler, O. (). Gesammelte Schriften zur Semitohamitistik. Edited by T. Schneider. Münster: Ugarit-Verlag. Satzinger, H. (). ‘On ergativity in Egyptian’, in A. Zaborski (ed.), New Data and New Methods in Afroasiatic Linguistics: Robert Hetzron in memoriam. Wiesbaden: O. Harrassowitz, –. Satzinger, H. (). ‘What happened to the voiced consonants of Egyptian?’, in P. Kousoulis and N. Lazaridis (eds.), Proceedings of the Tenth International Congress of Egyptologists, University of the Aegean, Rhodes – May . Vol. . Leuven, Paris, and Bristol, CT: Peeters, –. Schenkel, W. (). Tübinger Einführung in die klassisch-ägyptische Sprache und Schrift, th rev. edn. Tübingen: Wolfgang Schenkel. Schneider, T., Breyer, F., Kaelin, O., and Knigge, C. (eds.) (). Das Ägyptische und die Sprachen Vorderasiens, Nordafrikas und der Ägäis: Akten des Basler Kolloquiums zum ägyptisch– nichtsemitischen Sprachkontakt, Basel .–. Juli . Münster: Ugarit-Verlag. Schweitzer, S. D. (). Schrift und Sprache der . Dynastie. Wiesbaden: Otto Harrassowitz. Shisha-Halevy, A. (). Topics in Coptic Syntax: Structural studies in the Bohairic dialect. Leuven, Paris, and Dudley, MA: Peeters. Simpson, R. S. (). Demotic Grammar in the Ptolemaic Sacerdotal Decrees. Oxford: Griffith Institute. Takács, G. (). Etymological Dictionary of Egyptian. Vol. : A Phonological Introduction. Leiden: Brill. Takács, G. (–). Etymological Dictionary of Egyptian. Vol. . Leiden: Brill. Vergote, J. (). Phonétique historique de l’égyptien: Les consonnes. Leuven: Muséon. von Lieven, A. (). Grundriss des Laufes der Sterne: Das sogenannte Nutbuch. Vol. : Text. Copenhagen: Museum Tusculanum Press. Vycichl, W. (). Dictionnaire étymologique de la langue copte. Leuven: Peeters. Westendorf, W. (–). Koptisches Handwörterbuch. Heidelberg: Carl Winter. Wilson, P. (). A Ptolemaic Lexicon: A lexicographical study of the texts in the temple of Edfu. Leuven: Peeters. Winand, J. (). Études de néo-égyptien. . La morphologie verbale. Liège: Centre Informatique de Philosophie et Lettres. Woods, C. (ed.) (). Visible Language: Inventions of writing in the ancient Middle East and beyond. Chicago, IL: Oriental Institute.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
B languages are a close-knit language group, whose internal differences remind one of those in Germanic or Romance. In Berber scientific and political discourse, there is a tendency to play down the differences, and often Berber is represented as one single language with only some superficial regional variation (e.g. Chaker : ). On the other hand, some sources outside the world of Berberology use a very liberal definition of “language” where Berber is concerned; thus the Ethnologue (Lewis et al. ) has no fewer than different Berber languages (see Map .). Instead of providing a tree-model of the different varieties, I shall proceed by defining a number of “blocks”, i.e. bundles of varieties that exhibit a fair degree of consistency with one another. Two such blocks are easily defined. First, Zenaga in Mauritania stands on its own. It is in many points very different from other Berber languages, e.g. by showing different developments in phonology, and several highly original morphological traits (for details, see Zenaga, chapter of this volume). Zenaga has one sister language, Tetserret, which is spoken by parts of the Ayttawari Seslem, a small subgroup of the Iwellemmeden Tuareg ethnicity in Niger, thousands of kilometers from Zenaga territory. As shown conclusively by Lux (), this language has many innovations in common with Zenaga and should be considered a sister variety of it. The other clearly defined group is Tuareg, which has several regional variants in Mali, Niger, and Algeria and, to a lesser extent, in Burkina Faso (due to immigration from the eighteenth century onwards) and Libya. Again, this group is defined by a large number of common innovations. Internal diversification is important, and speakers from different varieties may have difficulties in understanding each other. Zenaga, Tetserrét, and Tuareg are the only Berber languages that fall outside the northern Berber dialect continuum that stretches from the Atlantic coast in Morocco to the Siwa oasis in Egypt. Although sometimes large Arabic-speaking territories lie between several patches of Berber speakers, adjacent (or the like) dialects are normally mutually understandable, and communication using Berber is possible. The effect of this continuum situation is that the spread of linguistic innovations is not blocked by linguistic obstacles; they can freely spread over the continuum, leading to an intricate network of isoglosses which only rarely define clear linguistic boundaries.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. The geographical distribution of Berber languages
Berber dialectology is predominantly synchronic, trying to define different synchronic blocks by using different approaches (e.g. Willms ; Ameur ; Lafkioui ; cf. also Lafkioui, chapter of this volume). While some minor breaks appear, most of the results confirm a basically geographic cline, not so different from the situation invoked by André Basset, the foremost Berber dialectologist of the twentieth century: ce n’est même pas, comme on le croit trop généralement, une langue divisée en quelques dialectes. . . . Il en résulte que cette langue s’éparpille directement ou à peu près en une poussière de parlers, de à mille peut-être pour quelques cinq millions d’individus. (Basset : ) [It does not even constitute—as is too often believed—a language divided into a small number of dialects. . . . As a result, this language fragments, immediately or almost so, into a dust cloud of varieties, four to five thousand for a population of about five million people.]
Different results are obtained when taking a historical point of view (e.g. Kossmann , : –; Naït-Zerrad ; Souag : –). This follows two itineraries. First, certain innovations that are believed to be quite early in the development of Berber, and which are very commonly found, can be used in order to single out varieties that did not undergo this development (i.e. have archaic features where the others shared an innovation). Secondly, when a consistent bundle of isoglosses defines a territory, one may venture the idea that these varieties once formed a unity. The many much less consistent isoglosses cutting through these territories are then interpreted as later innovations.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Using these methods, a number of historically defined entities can be distinguished, which will be described from west to east. In the western part of Morocco, there is a large continuous territory which covers the Anti-Atlas, the Sous plains, the High Atlas, and most of the Middle Atlas. The medieval material contained in a number of texts (van den Boogert , ) clearly belongs to this group. While the varieties spoken at the extremities of the continuum are very different from each other, there are no clear internal boundaries inside it—not even the abrupt relief of the High Atlas chain seems to constitute a major break. For practical reasons, a distinction is often made between two main varieties, Tashelhiyt, spoken in southwestern Morocco in the Anti-Atlas, the Sous plains, and the western High Atlas, and Tamazight (also known as Middle Atlas Berber),1 spoken in the southeastern Moroccan oases, in the eastern High Atlas, and in the western and central Middle Atlas. While differentiating between these entities may be useful for practical purposes and reflects the amount of linguistic variation, it is misleading in suggesting a clear boundary between the two. In northwestern Morocco, two more varieties are spoken that may belong to the same western Morocco block: Ghomara Berber and Senhadja de Sraïr.2 It seems, however, that they have been without contact with the other western Moroccan varieties for a while, which has led to many atypical retentions and innovations; a more detailed investigation might reveal that one of them (Ghomara) or both are best considered entities on their own. Eastern Morocco, most of Algeria, Tunisia, and parts of Libya belong to one single historical block, which has been called Zenatic since the late nineteenth century (for an early overview, cf. Destaing ).3 This block is characterized by a large number of innovations in phonology and morphology (Kossmann ; Naït-Zerrad ; Souag ). Taking these innovations as a basis, Zenatic has clear boundaries in Morocco and in Algeria. In Morocco, this boundary separates Ghomara and Senhadja de Sraïr (nonZenatic) from Riffian (Zenatic); in the Middle Atlas, it separates Ayt Seghrouchen (Zenatic) from its non-Zenatic western neighbors. In northern Algeria, Zenatic comprises all varieties except Kabyle. Zenatic is also the language of the major oases in the northern Sahara, Figuig, Gourara, Mzab, and Ouargla. Farther to the east, the boundaries of Zenatic become blurred. The Tunisian dialects (as far as we know about them) and Zuara (Libya) are still classical Zenatic varieties. However, the oases of Sokna, Elfoqaha, and Siwa share some features with Zenatic, but lack other features (cf. Souag ). Thus, while Zenatic has clear boundaries to the west and with Kabyle, it is more in a relation of continuum with eastern varieties.4 In view of the large and—because of the expansion of Arabic—rather scattered geographical distribution, Zenatic has been split up in some accounts into many different 1 Tamazight is the name of the Berber language in a large number of Berber varieties (among others Riffian and, with altered phonology, Tuareg); therefore, its use for one specific variety is unfortunate. The more neutral term “Middle Atlas Berber”, which is often used instead, is unfortunate too, as it also comprises parts of the High Atlas mountains. “Central Moroccan Berber” seems to be the least problematic term, but has not been used yet by many authors. 2 In spite of claims to the contrary, both varieties are thriving (Lafkioui ; Mourigh ). 3 The name derives from the important historical Zanāta tribe. Whether there is a link between the historical and the linguistic entity is an open question. Note that our delimitation is different from that found in, for instance, the Ethnologue (Lewis et al. ). 4 This suggests that the propagation of the Zenatic dialects—whether by demic or merely by linguistic expansion—went from east to west.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
“languages” (e.g. Lewis et al. ). This is rather arbitrary; in fact, there is sometimes remarkable mutual understanding over long distances. Thus, speakers of Figuig Berber (eastern Morocco) traveling to Libya were astonished that they could understand Zuaran without a problem, while they would not understand Moroccan Tashelhiyt. Kabyle seems to stand alone. It has significant dialectal fragmentation, and especially the most eastern varieties (eastern part of “Petite Kabylie”) are very different from what is found elsewhere. In a number of cases, Kabyle has undergone similar innovations to the western Moroccan block. It is difficult to decide, at this point, whether this points to an earlier extension of this block towards Algeria (separated by the incursion of Zenatic), or whether they represent parallel developments. The situation in Libya and Egypt is the most complicated in Berber dialectology; unfortunately, we lack good documentation for some of the key Libyan dialects (especially Sokna and Elfoqaha, both of them probably extinct now). As mentioned above, the fishing port of Zuara has a classical Zenatic dialect. The oasis dialects of Sokna and Elfoqaha seem to represent varieties close to Zenatic, but not quite part of it. Siwa (Egypt) is relatively similar to Sokna and Elfoqaha, but has undergone major innovations, especially in verbal morphology. These innovations are so profound that one suspects that it has undergone a kind of reshuffling due to the presence of a large community of non-native speakers in the oasis—something well-known from the history of the place (Souag : –). Two Libyan oasis dialects stand apart: Ghadames in western and Awjila in eastern Libya. While they share a number of important archaisms, there is hardly any sign of common innovations, and they are best considered different entities. Both are very different from the other Libyan and Egyptian dialects, although Awjila has some recent innovations in common with nearby Siwa, especially in syntax. The dialect of Djebel Nefusa in western Libya has a special position. On the one hand, it reveals a number of very archaic features, which place it outside Zenatic and its eastern continuation, e.g. the retention of a continuant pronunciation of *β before a consonant (Kossmann : ). Moreover, it has a number of developments shared with Ghadames (esp. forms such as Nefusa ufəs, Ghadames ofəs ‘hand’ instead of general Berber (a)fus). On the other hand, some of the defining Zenatic developments are also found in Nefusan, thus establishing a link with this block. One notes that Nefusan texts are quite easy to process with a knowledge of a Zenatic Berber variety, while this is much more difficult with, for instance, Ghadames texts. Maybe Nefusan is best viewed as a basically non-Zenatic dialect which, at a certain moment, underwent a very strong influence from neighboring Zenatic varieties. On the other hand, Souag (: ) suggests the opposite scenario, Nefusan being a basically Zenatic dialect that underwent significant influence from neighboring Ghadames. At this point the importance of the Ibadite network should not be underestimated (Brugnatelli ). The Ibadites constitute an early branching of the Islamic creed; in northern Africa, they persist in the Djebel Nefusa and in the Zenatic-speaking communities of Mzab, Djerba, and Zuara. The continuous contacts between these brothers-in-creed may well have brought about a certain “zenatification” of Nefusan. In a block-like classification of Berber languages, one has, therefore, the following historically defined entities: . Zenaga block (Zenaga of Mauritania, Tetserrét in Niger) . Tuareg block
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
a. Western Moroccan block (southwestern Morocco, central Morocco, i.e. Tashelhiyt and most of Tamazight) b. possibly including northwest Moroccan Berber (Ghomara, Senhadja de Sraïr) . Zenatic block (eastern Morocco, western Algeria, Saharan oases, Tunisia, Zuara) extending towards the east with Sokna, Elfoqaha, Siwa . Kabyle (northern Algeria), possibly linked to the western Moroccan block . Ghadames (Libya), probably linked to Djebel Nefusa (Libya) . Awjila (Libya) It should be noted that some of the most salient dialectal phonetic developments in Berber cut across these groups and seem to represent later innovations. This is the case with the lenition of stops (called spirantization in Berber studies), which is found all over northern Morocco, northern Algeria, and Tunisia, as well as in Zenaga, and which cuts both the western Moroccan block and the Zenatic block in two. It is also the case with the loss of the accentual system, which happened in all non-Tuareg varieties of Algeria and Morocco, and with the reduction in the system of short vowels, which is found all over the northern continuum, except for Ghadames. Moreover, one notes the existence of some salient features with a highly erratic distribution, cutting across many of the above blocks. One example is the distribution of the pronunciation ṭ instead of ḍ (e.g. aṭar ‘foot’ vs. aḍar). While ḍ is found throughout the Berber territory, including Zenaga and Tuareg, ṭ is found in a scattering of dialects in the east (Siwa, Awjila, Nefusa), in Algeria (eastern Kabylia), in the Zenatic block (Ayt Warayn in the eastern Middle Atlas), and in the western Moroccan block (Dades region and Ghomara). Similarly, the SG subject marking on the verb has three variants with a discontinuous distribution. Most generally attested is t- . . . -əd. However, a number of varieties have a pharyngealized consonant in the suffix (t- . . . -əḍ or t- . . . -əṭ), e.g. Siwa and Kabyle, while others have a non-pharyngealized voiceless suffix (t- . . . -(ə)t), e.g. Awjila, Ghadames, Ghomara, and Tashelhiyt. This suggests that in earlier times, before the emergence of the currently definable blocks, other entities existed and that the groups defined above include members of different earlier groups. As a consequence, one may doubt whether the tree model is suitable for the description of the Berber language family. Its continuous history of convergence and differentiation along new lines makes any definition of branches arbitrary. Moreover, mutual intelligibility and mutual influence render notions such as “split” or “branching” rather difficult to apply except, maybe, in the case of Zenaga and Tuareg. Still, a number of elaborate attempts at subclassification have been made, using lexicostatistical methods.5 In Figure ., I present the results of one of these, Blažek’s () tree based on twenty-two varieties using Starostin’s “calibrated” glottochronological method. These results are not too different from the seven-block compartmentalization presented above; the main difference is the place of Djebel Nefusa. Still, even if one accepts the basic 5
An unconvincing classification has been provided by Aikhenvald and Militarev (), cited and criticized by Blažek (). At many points this classification seems to be arbitrary. Moreover, at points it classifies dialects which are fully undocumented (e.g. Tmessa in Libya) or which are not Berber at all (e.g. Tadaksahak, which is Northern Songhay, and the Kufra oases, which are Teda-speaking). Unfortunately, some of the main lines of this classification have been taken over by the Ethnologue (Lewis et al. ).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
–600
–400
–200
0
+200
+400
+600
+800
+1000
+1200
West
Zenaga +800/88.3%
–680 58.8% North
Zenati
–370 64.7%
+950/90.9% +150 75.5%
-600 60.6%
+310/78.8%
+590/84.3% +1220/95.1% +820/88.7% +730/87.0%
+410/80.8%
+770/87.9% +940/90.8%
+50 73.5%
Tashelhiyt Tamazight Figuig Rif Beni Snous Metmata Chawia Mzab Ouargla Sened Zuara Nefusa Kabyle Ghadames
East
+510/82.8% –20/72.0%
+930/94.5%
El Foqaha Sokna Siwa Awjila
South
+940/90.8% +1180/94.5%
Ahaggar Tadghaq E. Iwellemmeden
. Blažek’s () classification of Berber founded on Starostin’s calibrated glottochronology based on minimal values (language names adapted)
tenets of the method, its application to Berber is difficult. The method discards loanwords from the set of items to be measured. This is relatively easy in the case of loanwords from Arabic. However, when it comes to borrowing between Berber varieties, it is hardly possible to distinguish loans from common heritage, both practically (how to see the difference) and theoretically (whether a spreading lexical innovation is to be considered common heritage or borrowing). The dating of Proto-Berber is a difficult affair. As mentioned above, linguistic differentiation seems to be similar to that in Germanic or Romance, which would put it somewhere in the first millennium (cf. Louali and Philippson ). A similar date ( ) is provided by Blažek. Other researchers have provided much earlier dates. Blench (: ), for example, has proposed a date for Proto-Berber around , explaining the high degree of uniformity by “highly mobile populations already speaking closely related languages, constantly encountering one another in open terrain”. Whatever the merits of the model as such may be, it is hardly applicable to northern Africa, which is for a large part mountain area.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
In view of the continuous movement of convergence, one may ask whether the establishment of a Proto-Berber entity is possible at all (cf. Galand : ). Múrcia Sànchez (, vol. II: ) has suggested that at a certain period Berber would have constituted a multinuclear koine rather than a unity, i.e., large-scale convergence would have blurred the distinctions between originally much more different varieties without obliterating them entirely. Maybe a relatively late date could be proposed for this koine. One of the remarkable facts about Berber in antique sources (Múrcia Sànchez ) is the general use of gu/gw where one would expect (non-geminated) w, and c rather than ɣ, as found in almost all modern Berber varieties.6 Does this mean that in Antiquity (or later?) these phonemes were still pronounced as stops and that their lenition was able to spread all over the Berberspeaking territory? Were these changes accompanied by other innovations, less visible in the sources? To what extent do our reconstructions in fact represent this late Antique koineization rather than Proto-Berber? If this is the case, it is doubtful that one could even get a glimpse of the proto-language through the veil of the grand convergence movement(s) which followed it. Two entities fall outside the normal definition of Berber in its synchronic sense, but clearly have some relationship with it. First, sometime in the first millennium writings appear in northern Africa which use an alphabetic script called “Libyan”, which seems to have been developed independently to a large degree. Kerr () points to important structural parallels with the Punic script and posits its invention in the second century . Most specialists prefer an earlier dating (e.g. Pichler ), and it is conceivable that Kerr’s arguments concern an orthographic reform rather than the earliest design of the script. Although there are thousands of inscriptions in this script, their language is not easy to determine, as they consist almost exclusively of personal names. The few inscriptions that have somewhat more text show a language which has clear parallels with Berber; however, it is difficult to define this relationship with any precision (cf. Galand : –). The other entity is the ancient language of the Canary Islands,7 commonly known as Guanche. This language died out sometime in the seventeenth century, and we only have limited resources on it, which more often than not are difficult to interpret (cf. Wölfel ). There can be no doubt that the language contains Berber elements; there are many words that have clear and unproblematic cognates in Berber, such as ilfe ‘pig’ (cf. the generally attested northern Berber form iləf ‘pig’). On the other hand, one is struck by the presence of a large stock of vocabulary that does not have any resemblance to Berber whatsoever. Moreover, as remarked by Galand more than once (e.g. : –), the few short texts resist any interpretation from a Berber point of view, and no Berber inflectional elements seem to appear in them. As to the lexicon, one notes that most of the Berber material concerns agricultural terms pertaining to crops, livestock, and related concepts. Most (but not all) terms referring to basic concepts do not have a clear correlate in Berber. One could explain this by positing a double layering in the language: it would be basically non-Berber, but due to the assimilation of a later influx of Berber speakers, who may have introduced new agricultural practices and livestock, large numbers of Berber words entered the lexicon. Zenaga has glottal stop as the cognate of ɣ elsewhere; in Ghadames, the phoneme has irregularly split into two: ɣ and ʕ. Finally, there are some unexpected cognates with ɣ all over Berber which have ẓ in Tuareg. 7 I should like to thank Marijn van Putten for his help in this matter. 6
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Such an explanation is tempting but must remain speculation because of the scarcity of documentation; thus, we are much better informed about the cultural lexicon than about the lexicon relating to body parts or basic verbs.
A I wish to thank Carles Múrcia Sànchez and Thilo Schadeberg for their comments on this chapter.
R Aikhenvald, A. Y., and Militarev, A. Y. (). ‘Ливийско-гуанчские языки’ [‘Lybia-Guanche languages’], in I. M. Diakonoff and G. S. Sharbatov (eds.), Языки Азии и Африки [Languages of Asia and Africa]. Vol. .: Афразийские языки [Afro-Asiatic Languages]. Moscow: Nauka, – [not consulted]. Ameur, M. (). ‘À propos de la classification des dialectes berbères’, Études et Documents Berbères : –. Basset, A. (). La langue berbère. London, New York, and Toronto: Oxford University Press. Blažek, V. (). On classification of Berber. Paper presented at the th Colloquium of African Languages and Linguistics, Leiden, August –, . Blench, R. (). ‘Types of language spread and their archaeological correlates: the example of Berber’, Origini : –. Brugnatelli, V. (). ‘D’une langue de contact entre berbères ibadites’, in M. Lafkioui and V. Brugnatelli (eds.), Berber in Contact: Linguistic and sociolinguistic perspectives. Cologne: Rüdiger Köppe, –. Chaker, S. (). Linguistique berbère: Études de syntaxe et de diachronie. Paris and Leuven: Peeters. Destaing, E. (). ‘Note sur la conjugaison des verbes de formes C1eC2’, Mémoires de la Société de Linguistique de Paris : –. Galand, L. (). Regards sur le berbère. Milan: Centro Studi Camito-Semitici. Kerr, R. (). ‘Some thoughts on the origins of the Libyco-Berber alphabet’, in H. Stroomer, M. Kossmann, D. Ibriszimow, and R. Vossen (eds.), Études berbères V: Essais sur des variations dialectales et autres articles. Cologne: Rüdiger Köppe, –. Kossmann, M. (). Essai sur la phonologie du proto-berbère. Cologne: Rüdiger Köppe. Kossmann, M. (). The Arabic Influence on Northern Berber. Leiden and Boston, MA: E. J. Brill. Lafkioui, M. (). Atlas linguistique des variétés berbères du Rif. Cologne: Rüdiger Köppe. Lafkioui, M. (). ‘Analyses dialectométriques du lexique berbère du Rif ’, in R. Vossen, D. Ibriszimow, and H. Stroomer (eds.), Études berbères IV: Essais lexi cologiques et lexicographiques et autres articles. Cologne: Rüdiger Köppe, –. Lewis, M. P., Simons, G. F., and Fennig, C. D. (eds.) (). Ethnologue: Languages of the World, th edn. Dallas, TX: SIL International. Louali, N., and Philippson, G. (). ‘Berber expansion into and within north-west Africa: a linguistic contribution’, Afrika und Übersee : –. Lux, C. (). Le tetserret, langue berbère du Niger: Description phonétique, phonologique et morphologique, dans une perspective comparative. Cologne: Rüdiger Köppe. Mourigh, K. (). A Grammar of Ghomara Berber (North-West Morocco). Cologne: Rüdiger Köppe. Múrcia Sànchez, C. (). La llengua amaziga a l’antiguitat a partir de les fonts gregues i llatines. Barcelona: Promocions i Publicacions Universitàries. Naït-Zerrad, K. (). ‘Esquisse d’une classification linguistique des parlers berbères’, Al-AndalusMaghreb –: –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Pichler, W. (). Origin and Development of the Libyco-Berber Script. Cologne: Rüdiger Köppe. Souag, L. (). Berber and Arabic in Siwa (Egypt): A study of linguistic contact. Cologne: Rüdiger Köppe. van den Boogert, N. (). The Berber Literary Tradition of the Sous. Leiden: Nederlands Instituut voor het Nabije Oosten. van den Boogert, N. (). ‘Medieval Berber orthography’, in S. Chaker and A. Zaborski (eds.), Études berbères et chamito-sémitiques: Mélanges offerts à Karl-G. Prasse. Paris and Leuven: Peeters, –. Willms, A. (). Die dialektalische Differenzierung des Berberischen. Berlin: Dietrich Reimer. Wölfel, D. J. (). Monumenta Linguae Canariae. Die kanarischen Sprachdenkmäler: Eine Studie zur Vor- und Frühgeschichte Weißafrikas. Graz: Akademische Druck- und Verlagsanstalt.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. B
.................................................................................................................................. T sheer majority of Cushitic languages belong to one single subgroup: East Cushitic (EC). Its speech area extends from Eritrea with Saho to Mount Kenya with the extinct Yaaku language, and from the Indian Ocean with ‘Afar and Somali to the Sudanese-Ethiopian border with Dhaasanac (see Map .). Thirty-three of the forty-five Cushitic languages—a few of them actually members of macro-languages—counted by Ethnologue are classified under EC. Ongota, a highly endangered language of southwest Ethiopia of uncertain stock, may be yet another EC language (Savà and Tosco ). In demographic terms, a few languages take the lion’s share: at least five languages (Oromo, Somali, Sidamo, ‘Afar, and Hadiya) have more than one million speakers each, (and the first two—Oromo and Somali—much more than ten million each). At the other end of the spectrum, a couple of EC languages are extinct: Elmolo and Yaaku, both formerly spoken in Kenya, belong here. Many more have a few thousand or hundred speakers at most. In this chapter, only the genetic classification of EC will be discussed. The genetic classification of languages follows biological taxonomy in striving for perfect nestedness, i.e. to assign to each and every language a unique position in a branched tree, thus representing its genealogical position. It will, however, be assumed that genetic classification does not show perfect “nestedness” because certain types of natural languages (pidgins and creoles, true mixed languages, and planned languages) are beyond its scope. The establishment of isoglosses is central in order to justify subgrouping, and genetic classification follows biological taxonomy in being cladistic rather than phenetic: not just any isogloss or shared similarity between languages counts as decisive; only shared innovations (synapomorphies) do. Genetic classification does not depict the history of a language but merely shows its position within a classificatory tree. Of course, because languages notoriously adopt so many features through contact, classification per se does not show much of what a language looks like. In the case of Cushitic, and EC in particular, this involves the difficult task of disentangling the role of the much debated Ethiopian language area(s) (see Tosco b for a critical view).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of East Cushitic languages
. EC C ?
.................................................................................................................................. The usual classificatory tree of Cushitic involves four parallel branches. The richness of the EC branch stands out clearly (Figure .) The other three branches of Cushitic (North, Central, and South) are each much smaller and less diverse: to one extreme, North Cushitic (NC) corresponds to a single language, Beja (see Vanhove, chapter of this volume). The divergence between Beja and the rest of Cushitic has been remarked upon many times, and even its classificatory position outside Cushitic has been proposed. Zaborski (: ) has aptly defined the whole question as “a classical rumour” devoid of any scientific basis. Still, the differences between Beja and the rest of Cushitic are huge at all levels: Ehret (: ) noted a percentage of
Cushitic
Northern (= Beja)
Bilin
Khamtanga
Central (= Agaw)
Southern
Kemant Awngi
(source: Appleyard (1988: 590)) Highland (HEC)
Hadiyya
Eastern
Kambata Sidamo Gede’ o Burji
(based upon Hudson 1981:120)
Yaaku-Dullay
*Yaaku
Ma’ a
Lowland
Dullay (cluster)
Omo-Tana
Western
Central
Dhaasanac Arbore *Elmolo Baiso
Oromoid
Dahalo
Saho-‘Afar
Eastern Oromo Konsoid (cluster)
Rendille Boni
Somali
Rift
West
Iraqwoid
East
Kwadza Asa
Iraqw Gorwaa Alagwa Burunge (Southern Cushitic from Ehret 1980; West Rift: Kiessling 2002)
. The Cushitic languages (from Tosco a)
common similarity between Beja and South Cushitic (SC) of only to (going up to to as compared with Central Cushitic (CC)). With regard to the morphological level, the complex problem of the verbal system of Beja has been discussed many times, most notably by Voigt (, ) and Zaborski (, , , a, b, ). CC (see Zelealem, chapter of this volume) is made up of a group of relatively closely related languages traditionally called Agaw, whose internal classification has been elaborated by Appleyard (). Hetzron () proposed linking Agaw to a subgroup of EC, as will be discussed in some detail in section . below. The doubts about SC (see Kießling, chapter of this volume), notwithstanding Ehret’s () phonological reconstruction, have never been dissipated. As noted by Hetzron (: ), “the fact that these languages do form a genetic unit does not necessarily mean that Southern Cushitic must be a branch coordinated with the other major branches of Cushitic, as has been assumed. It may still be a sub-branch of one of the major branches.” Consequently, he tried to “place Iraqw, and with it the whole of the Southern Cushitic group, within Southern Lowland Cushitic”, substantially on morphological grounds (traces of EC adpositions and subject marking in the Iraqw cluster). With regard to the other SC languages, nothing new can be said about the (quite certainly extinct) Aasá(x) and Kw’adza languages due to lack of data (cf. Petrollino and Mous for a review of the status of Aasá(x)). The position of Dahalo has been discussed in Tosco (a), where its reclassification within EC was proposed. Ma’a (or Mbugu) presents special problems due to its status as essentially a secret and partially planned language (Mous ). The most radical proposal about the inclusion of much of Cushitic within EC—or rather the watering-down of EC within Cushitic as a whole—has been put forward by Voigt
Cushitic
North (= Beja)
Saho-‘Afar
South
Omo-Tana
–Central (= Agaw) –Highland –…
. The classification of Cushitic according to Voigt ()
(). On the basis of some peculiarities of its prefix conjugation, he put Beja on equal status with all the rest of Cushitic, as schematically shown in Figure .. In Voigt’s proposal EC is denied any value as a classificatory unit and its subgroups are placed on the same footing with Agaw (CC) and SC. However, no final proof for or against the inclusion of SC within EC will be offered here; attention will rather be paid to the “traditional” EC languages and their classification.
. T EC
.................................................................................................................................. It will come as no surprise that the internal classification of EC is a complex and not fully resolved problem. It must also be said that, after the wealth of comparative studies and proposals of the s and s (one of the best examples being Sasse’s historical phonology of EC), recent years have rather witnessed a much welcomed rise in descriptive work, while comparison and reconstruction have been lagging behind. Certain subgroups stand out clearly, but a few languages defy any easy classification. A major problem consists in the status of “Lowland EC” (LEC) against the well-established and sounder “Highland EC” (HEC). Although the distinction goes back to Leo Reinisch’s work on different Cushitic languages in the nineteenth century, and Black () attempted a reconstruction of LEC (from which Dullay, called “Werizoid”, and Yaaku—still unknown at that time—were excluded), the exact limits and the genetic (rather than typological) status of this group are still unclear. In the standard view, EC is made up of two or maybe three branches: Highland, Lowland, and, as the case may be, a much smaller separate group consisting of Dullay and Yaaku. Lowland is further divided into three branches: Omo-Tana, Oromoid, and Saho-‘Afar. Although the last is shown as a single macro-language in Figure . above, two separate languages have to be recognized. Still, it remains a very small and coherent subgroup. The classification and historical phonology of Omo-Tana have been worked out by Sasse () and, for its Eastern branch, by
Heine (), who labeled it “Sam”. Oromoid, as its name implies, is dominated by a single macro-language (Oromo) and a few other as yet very imperfectly known languages of southwest Ethiopia (Konso, Bussa, Diraasha—formerly known rather as Gidole). Dullay and Yaaku (especially the latter, an extinct language) are still poorly known. As shown above, even their membership in a single subgroup is debatable. In contrast to this, the unity of the HEC group seems evident.
. HEC
.................................................................................................................................. This subgroup and its membership are quite well established. Only recently a number of high-quality grammatical descriptions of new varieties have been made available (Crass for K’abeena; Schneider-Blum for Alaaba; Treis for Kambaata). A major factor in shaping the HEC languages has certainly been the long and strong contact with neighboring Omotic languages of the Ometo group. Syntactically, the HEC languages are characterized by S O V order at sentence level and modifier head order at phrasal level (against head modifier of LEC). They have quite complex systems of nominal cases and postpositions. Phonologically, they have ejectives but no implosives, distinctive vowel length but no [ATR] feature, and no tones. Per se, all these features offer no proof of being more than a collection of areal features, with no or little decisive evidence of genetic unity. But the HEC languages also have a highly idiosyncratic lexical stock, as evidenced in Sasse’s () comparative dictionary and Hudson ( and —the latter on the internal organization of the subgroup). While the cohesion of the HEC languages is clear, their external relations have been the matter of some contention, in particular with regard to CC (Agaw). In what remains probably one of his most audacious proposals, Hetzron () reclassified CC with HEC. It is important to note that most of Hetzron’s arguments apply to Awngi (the southernmost Agaw language) and Burji (the southernmost and most isolated member of HEC, but also the least affected by Omotic influence). He added that his hypothesis “is not proposed here ‘aggressively’, with a full commitment by the author to it” (Hetzron : ). As is common with Hetzron’s work, his isoglosses were morphological and therefore particularly strong, but Hetzron himself provided a stringent criticism of them, pointing to their areal or borrowed character, or to their status as archaisms rather than common innovations. In the end, the strongest piece of evidence, according to Hetzron, is the genitive agreement, whereby in Agaw and HEC the genitive markers attached to the possessor agree in gender (and also in number in Agaw) with the possessed (Hetzron : ; agreement with the possessor is obviously well attested elsewhere). Although they are syntactic in nature (the actual markers are the pan-Cushitic -k for masculine and -t for feminine), the highly idiosyncratic nature of the construction (with gender markers on the possessor agreeing with the possessed, in line with a consistent S–O–V syntax) may be accepted as a good Agaw–HEC isogloss. In the verbal system, what sets HEC apart from the other EC languages is the development of “extended” paradigms, which have come to dominate the verbal system at least in main declarative clauses. Even if possibly triggered by areal contact (Tosco ), this would still be a synapomorphy and, therefore, decisive in classification.
What is left of EC after HEC underwent its specific developments in the verbal system makes up LEC. It has preserved the old Cushitic suffix conjugation (and also the still older prefix conjugation) to a larger extent than HEC (and of CC). In this sense, LEC encompasses not only the “classical” Lowland languages but Yaaku and Dullay too.
. D LEC
.................................................................................................................................. Within LEC, a distinctive feature of Saho-‘Afar is the conservation (and development) of the prefix conjugation. Obviously, this conservative feature can hardly qualify in itself for genetic classification. Other well-known but isolated features in the basic lexicon (such as the replacement of the common EC word for ‘blood’) will also be disregarded here as possibly due to contact. Syntactically, all LEC languages with the exception of Saho-‘Afar are characterized by a head–modifier phrasal order; being typologically unusual, this order is all the more interesting (Tosco considers it original in EC). Still, it is just a typological, syntactic feature, and scarcely a proof of genetic subgrouping. At the morphosyntactic level, the development of preverbal subject clitics and “case markers” is common to all the LEC languages with the exception again of Saho-‘Afar. Hetzron () discusses this feature at length as a truly innovative feature separating Saho‘Afar from the rest. On the other hand, Hetzron himself (: , n. ) also points to its broadly areal character. What Hetzron calls case markers is actually adpositions, and, following Biber’s () analysis for Somali, possibly former postpositions which got separated from their NP and moved before the verb. As to preverbal subject marking, the process whereby preverbal clitics came to replace or accompany the suffixal marking on the verb is not dissimilar from what was found elsewhere, as, for instance, in many Romance languages (cf. Tosco for a comparative typological analysis in terms of feature geometry), and is in any case doubtful as a base for genetic classification. Again, these syntactic developments do not seem to warrant a genetic split between Saho-‘Afar and the rest of LEC. The case for the subject clitics to be a bona fide innovation of LEC is, nevertheless, reinforced by the actual morphemes, which bear a good resemblance across languages. Of course, the subject clitics go back to other pronominal series, and similarity is expected. The main shared features are an element Vn for the first person found, to quote just a few examples, in Somali and Elmolo (Omo-Tana), in the whole Oromoid branch, in Dullay, and in Yaaku. As discussed in Tosco (), the LEC languages seem to start their subject clitic series from the first person singular (against Romance languages, where the second person singular is the first and often the only subject clitic). In LEC, for the second person -*t is preserved in Dirayta (or D’iraasha; Oromoid), or undergoes the usual phonological processes of the language (e.g. voicing, yielding, e.g. aad in Somali). In Dullay it is assimilated to a following consonant (yielding, e.g. ʔaC= in Gawwada). All things considered, the development of preverbal subject and preverbal adpositions and of subject clitics, coupled with, for the latter, their actual exponents, seems robust enough to be taken as a proof of genetic unity of all the LEC languages with the exception of
Saho-‘Afar. Following Hetzron (), Tosco (a) proposed calling this branch “Southern Lowland East Cushitic”.
. S LEC: O-T O . D Y
.................................................................................................................................. As for the single branches of LEC, the existence of Omo-Tana and Oromoid is relatively uncontroversial. The validity of the former has been questioned by Lamberti (in Haberland and Lamberti ): according to him, its Western branch (Dhaasanac, Arbore, and the now extinct Elmolo) would be a separate group within EC (which he called “Galaboid”, from galab, the Amharic denomination of the Dhaasanac), and coordinate to the Eastern branch (the “Sam” languages of Heine , i.e. Somali, Rendille, and Boni) together with Baiso (whose special links to southern Somali dialects have been noted; cf., e.g., Ehret and Nuuh Ali ). Lamberti’s proposal is not convincing (cf. Tosco a for a few critical remarks), and Omo-Tana remains a robust subgrouping. One of its strongest isoglosses is the imperative singular ending of the middle (autobenefactive) verbs in -o (with loss of final *-t), as in Somali furo, Dhaasanac fúru ‘open for yourself !’. A good starting point when looking for synapomorphies is the verbal system, which is the most complex part of Cushitic morphology. In particular, if we discount the grammaticalization of auxiliary constructions and analogical leveling, the positive paradigms are fairly well preserved in most Lowland languages. Not so the negative paradigms, which have been independently restructured in the different branches and languages. Banti (/) offers a very detailed and exhaustive analysis of most negative verbal forms in their Afro-Asiatic context. In both Omo-Tana and Oromoid one finds an invariable negative past opposed to a person-marked negative present. The negative past is marked by a nasal affix -Vn or -nV, with the former being more typical of Omo-Tana and the latter found in Oromoid (but not exclusively). The presence of the invariable past negative is tentatively proposed as a synapomorphy for the setting up of a Nuclear Southern Lowland group (cf. Tosco a), made up of Oromoid and Omo-Tana against the other Southern Lowland languages (Yaaku and Dullay). A genetic link between Yaaku and Dullay was suggested by Ehret (, ) and substantiated by Hayward () with lexical and phonological isoglosses. Although Hayward () withdrew from his list of shared phonological innovations the devoicing of obstruents (which is not common to the whole of Dullay and is not unknown elsewhere in EC), the hypothesis that Dullay and Yaaku make up a true genetic subgroup still has some weight. To turn to the verbal system, Yaaku and Dullay also show an interesting parallelism in their negative paradigms. No invariable negative form exists in Dullay; the negative perfective has a single form for all persons of the singular (marked by the suffix -ú in the Gawwada variety), while the plural persons have separate affixes. Interestingly, the same, typologically quite odd pattern is found in Yaaku. Yaaku (Heine /) distinguishes definite and indefinite paradigms. The indefinite negative has a single form for all persons in the singular of certain verbs (while other verbs have distinct
forms); as in Dullay, no reduction occurs in the plural. It must be noted, on the other hand, that the Yaaku indefinite does not cover the Dullay perfective, and that it is, rather, present in meaning. Instead, Dullay and Yaaku are very different in the other negative paradigms. The Dullay negative imperfective has basically the same syncretisms of the positive paradigms (with a single form shared by the first person singular and the third person masculine singular, and another shared by the second person singular and the third person feminine singular). The Yaaku negative definite has a singular form -n enlarged to -nèn in the plural. Banti (/) has noted the “nouny” character of this Yaaku paradigm and the parallel with other concordless negative past forms, proposing their origin from nominal forms (Yaaku -n is obviously reminiscent of the Omo-Tana and Oromoid invariable negative past mentioned above, and even its enlargement to -nèn in the plural finds parallels, for example, in Karre of southern Somalia; cf. Tosco ). Accepting Hayward’s () evidence as well as this common development in the negative verbal forms, one can tentatively propose a Peripheral Southern Lowland group (called “Transversal” in Tosco a) made up of Dullay and Yaaku. Alternatively, Dullay and Yaaku would not be a proper genetic group with its own synapomorphies but simply what was left of Southern Lowland after the rise of the Nuclear subgroup (Omo-Tana and Oromoid).
East Cushitic
Highland
Lowland
Saho-‘Afar
Southern
Nuclear
Omo-Tana
Peripheral (?)
Oromoid
Dullay
Yaaku
. A revised classificatory tree of East Cushitic
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
To summarize: EC is divided into a Highland and a Lowland branch (see Figure .). LEC is made up of what was left from common EC after HEC underwent its specific developments in the verbal system. Among the Lowland languages, Saho-‘Afar is the residual language; in its somewhat isolated location, Saho-‘Afar did not participate in the development of preverbal subject clitics (as per Hetzron ), which gave rise to Southern LEC. This was further divided by the rise of a Nuclear subgroup (Omo-Tana and Oromoid) characterized by specific developments in the negative paradigms. What was left, Dullay and Yaaku, may either form a Peripheral branch, or simply be what was left over after the rise of the Nuclear subgroup. The peripheral location of both languages seems to lend support to this view.
R Appleyard, D. L. (). ‘The Agaw languages: a comparative morphological perspective’, in T. Beyene (ed.), Proceedings of the Eighth International Conference of Ethiopian Studies (Addis Ababa, ). Vol. . Huntingdon: Elm Press, –. Banti, G. (/). ‘Some concord-less verbal paradigms in Omo-Tana’, Sprache und Geschichte in Afrika /: –. Biber, D. (). ‘The diachronic development of preverbal case markers in Somali’, Journal of African Languages and Linguistics : –. Black, P. (). Lowland East Cushitic: Subgrouping and Reconstruction. Ph.D. thesis. Yale University, Crass, J. (). Das K’abeena. Cologne: Rüdiger Köppe. Ehret, C. (). Ethiopians and East Africans: The problem of contacts. Nairobi: East African Publishing House. Ehret, C. (). ‘Cushitic prehistory’, in M. L. Bender (ed.), The Non-Cushitic Languages of Ethiopia. East Lansing, MI: African Studies Center, Michigan State University, –. Ehret, C. (). The Historical Reconstruction of Southern Cushitic Phonology and Vocabulary. Berlin: Dietrich Reimer. Ehret, C., and Nuuh Ali, M. (). ‘Soomaali classification’, in T. Labahn (ed.), Proceedings of the Second International Congress of Somali Studies (Hamburg, August –, ). Vol. : Linguistics and Literature. Hamburg: Helmut Buske, –. Griefenow-Mewis, C., and Voigt, R. M. (eds.) (). Cushitic and Omotic Languages: Proceedings of the Third International Symposium (Berlin, March –, ). Cologne: Rüdiger Köppe. Haberland, E., and Lamberti, M. (). Ibaaddo ka-Ba’iso: Culture and language of the Ba’iso. Heidelberg: Carl Winter. Hayward, R. J. (). ‘The Qawko dialects and Yaaku’, Abbay : –. Hayward, R. J. (). ‘Comparative notes on the language of the S’aamakko’, Journal of Afroasiatic Languages : –. Heine, B. (/). ‘Notes on the Yaaku language (Kenya)’, Afrika und Übersee : –, –. Heine, B. (). ‘The Sam languages: a history of Rendille, Boni, and Somali’, Afroasiatic Linguistics .: –. Hetzron, R. (). ‘The limits of Cushitic’, Sprache und Geschichte in Afrika : –. Hudson, G. (). ‘The Highland East Cushitic family vine’, Sprache und Geschichte in Afrika : –. Hudson, G. (). Highland East Cushitic Dictionary. Hamburg: Helmut Buske. Mous, M. (). The Making of a Mixed Language: The case of Ma’a/Mbugu. Amsterdam and Philadelphia, PA: John Benjamins. Petrollino, S., and Mous, M. (). ‘Recollecting words and expressions in Aasá, a dead language in Tanzania’, Anthropological Linguistics : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Sasse, H.-J. (). The extension of Macro-Somali. Paper presented at the “Colloque international sur les langues couchitiques et les peuples qui les parlent, Paris, – septembre ”. Sasse, H.-J. (). ‘The consonant phonemes of Proto-East-Cushitic (PEC): a first approximation’, Afroasiatic Linguistics .: –. Sasse, H.-J. (). An Etymological Dictionary of Burji. Hamburg: Helmut Buske. Savà, G., and Tosco, M. (). ‘A sketch of Ongota, a dying language of southwest Ethiopia’, Studies in African Linguistics : –. Schneider-Blum, G. (). A Grammar of Alaaba. Cologne: Rüdiger Köppe. Tosco, M. (). Schizzo grammaticale del dialetto Karre di Qoryooley. Rome: Ministero degli Affari Esteri. Tosco, M. (). ‘The historical syntax of East Cushitic: a first sketch’, in T. Bearth, W. J. G. Möhlig, B. Sottas, and E. Suter (eds.), Perspektiven afrikanistischer Forschung: Beiträge zur Linguistik, Ethnologie, Geschichte, Philosophie und Literatur. X. Afrikanistentag (Zürich, – September ). Cologne: Rüdiger Köppe, –. Tosco, M. (). ‘The Northern Highland East Cushitic verb in an areal perspective’, in C. Griefenow-Mewis and R. M. Voigt (eds.) (), –. Tosco, M. (a). ‘Cushitic overview’, Journal of Ethiopian Studies : –. Tosco, M. (b). ‘Is there an “Ethiopian Language Area”?’, Anthropological Linguistics : –. Tosco, M. (). ‘Feature-geometry and diachrony: the development of the subject clitics in Cushitic and Romance’, Diachronica : –. Treis, Y. (). A Grammar of Kambaata. Cologne: Rüdiger Köppe. Voigt, R. M. (). ‘Zur Bildung des Präsens im Bedauye’, in M. Bechhaus-Gerst and F. Serzisko (eds.), Cushitic–Omotic. Papers from the International Symposium on Cushitic and Omotic Languages, Cologne, January –, . Hamburg: Helmut Buske, –. Voigt, R. M. (). ‘Zur Gliederung des Kuschitischen: Die Präfixkonjugationen’, in C. GriefenowMewis and R. M. Voigt (eds.) (), –. Zaborski, A. (). The Verb in Cushitic. Cracow: Zeszyty Naukowe Uniwersytetu Jagielloskiego, Prace Jezykoznawcze. Zaborski, A. (). ‘Remarks on the genetic classification and the relative chronology of the Cushitic languages’, in J. Bynon (ed.), Current Progress in Afro-Asiatic Linguistics. Amsterdam and Philadelphia, PA: John Benjamins, –. Zaborski, A. (). ‘Remarks on recent developments in Cushitic’, in G. Bernini and V. Brugnatelli (eds.), Atti della a Giornata di Studi Camito-semitici e Indoeuropei. Milan: Unicopli, –. Zaborski, A. (a). ‘The position of Cushitic and Berber within Afroasiatic dialects’, in A. Bausi and M. Tosco (eds.), Afroasiatica Neapolitana. Naples: Dipartimento di Studi e Ricerche su Africa e Paesi Arabi, Istituto Universitario Orientale, –. Zaborski, A. (b). ‘Problems of the Beja present seven years later’, Lingua Posnaniensis : –. Zaborski, A. (). ‘La linguistique chamito-sémitique cinquante années après l’Essai Comparatif de Marcel Cohen’, in M. El Medlaoui, S. Gafaiti, and F. Saa (eds.), Actes du er congrès Chamito– Sémitique de Fès (– mars ). Fès: Université Sidi Mohamed Ben Abdellah, Faculté des Lettres et des Sciences Humaines Saïs-Fès, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T North Cushitic (NC) branch of the Afro-Asiatic (AA) phylum (formerly known as Hamito-Semitic) consists of only one language, Beja, named beɖawije:t by the Beja people. Recent archeological discoveries show growing evidence that Beja is related to the extinct languages of the Medjay (from which the ethnonym ‘Beja’ is derived: Rilly, : ) and Blemmye tribes, first attested on Egyptian inscriptions of the twelfth dynasty for the former and on a Napatan stela of the late seventh century for the latter; for recent discussions, see Browne (); El-Sayed (); Zibelius-Chen (); Rilly (, ). Beja is mainly spoken in eastern Sudan by some ,, speakers ( census, probably at least double that amount now) and in northern Eritrea (approximately , speakers). It used to be spoken in southern Egypt, but it seems that all or almost all speakers have now shifted to Arabic. Twenty years ago, Morin (: ) mentioned speakers at Aswan and Daraw, and Wedekind () still does so. It is necessary to recall the question of Meroitic, the language of the so-called “black pharaohs” attested in the inscriptions of the monuments along the Sudanese part of the Nile river. Since the beginning of the twentieth century, scholars have been discussing three possible classifications: Nilo-Saharan (NS; Nubian and Nara), AA (Beja or Omotic), or an isolated status (for an overview, see Rilly ). Even though the NS hypothesis has been favored since the second half the twentieth century, it was only recently that Rilly () could prove by means of the comparative method that Meroitic indeed belongs to NS— more precisely to an Eastern Sudanic branch which includes also Nubian, Nara, Taman, and Nyima—and not to NC. For a long time it had been believed that the Blemmyes, already mentioned by the Egyptians and the Greeks, were the ancestors of the Beja. Only recently Browne () was able to prove that the language of the Blemmyes was the ancestor of the Beja language. In his opuscule written in Latin he showed that the beginning of the text of a Blemmye ostracon from Saqqara (seventh century ) contains words which are cognate with Beja words, as well as the beginning of Psalm in which Beja morphological devices are recognizable in addition to the lexicon. He also showed that numerous Blemmye anthroponyms could be linked to Beja nouns, in particular those ending in tek or tak, ‘man’, in
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Beja. According to Rilly (: ), the first possible attestation of Beja may go as far back as , to an Egyptian papyrus written by a scribe named Djehouty-Mose which contains a magic formula starting with a word meaning ‘go away!’ in Beja (sigi), an opening typical of Egyptian magic formulas. If this hypothesis is correct, it would mean that the Beja language has been present in the Nile valley for thousands of years.
. T B
.................................................................................................................................. In the second half of the nineteenth century, Joseph Halévy (), a French orientalist, had already noticed the close links that exist between Agaw, Beja, Danakil (i.e. ‘Afar), Oromo, and Somali. Two decades later, Reinisch (/) proposed a bipartite classification of Cushitic languages and grouped Beja together with Saho-‘Afar, Somali, and Oromo in a “Lowland Cushitic” unit. As Lamberti (: –) pointed out, Reinisch’s classification made sense towards the end of the nineteenth century, considering the poor knowledge at the time, but is no longer tenable today. Lamberti (: ) deduced a more refined classification from Cerulli’s various publications on Cushitic languages, which appeared during the second quarter of the twentieth century. This classification sets Beja apart as the only member of the NC branch and divides the other languages into three groups: Central, Lowland, and Sidama. During the same time period Moreno () too argued for a twobranch division, somewhat different from Reinisch’s. He did not see Beja as an independent branch from the rest of Cushitic but grouped it with Central Cushitic (CC) in addition to Lowland Cushitic. Among the specialists in Cushitic languages, Cerulli’s classification of Beja prevailed against Moreno’s, but for various reasons it remained disputed until the beginning of the twenty-first century. An important public debate took place at the “Conférence internationale sur les langues couchitiques et les peuples qui les parlent”, which was held in Paris in , and whose transcripts and audio files are now available in Enguehard et al. (). During this conference, the late Joseph Tubiana initiated a debate about what defined and characterized Cushitic languages. Typological and genetic arguments were at the core of the numerous discussions that took place, and very prominently for Beja. Zaborski was one of the participants who strongly advocated, on genetic comparative grounds, an NC branch comprising only Beja.1 Several participants noted that the study of Beja was marginalized at that time and were struck by the very low number of cognates it shares with Oromo and Somali, and by its proximity to Semitic languages, in particular concerning the root structure and the role of apophony. Some five years later, Hetzron (: in particular –), who did not attend the conference, discussed the position of Beja within AA. He considered it too different from other Cushitic languages “in too many respects” and consequently proposed setting it apart from Cushitic as an independent branch of AA. Hetzron’s proposal was criticized by other linguists (e.g. Zaborski ; Tosco ), and Appleyard () showed that Hetzron’s features were
1
His position was further explained in detail in his ‘Cushitic overview’ in Bender ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
in fact limited to five Cushitic morphological innovations that Beja does not share with the other Cushitic groups, and to a different chronology of a putative word order change, an argument of typological nature. Our increasing knowledge of Cushitic languages has not changed Cerulli’s classification of Beja so that it still prevails to this day. In the s, another proposal was made from a very different perspective. Tucker () proposed applying a typological classification to Cushitic languages. In this work, he considered Beja along with Agaw and East Cushitic (EC), a grouping reminiscent of Reinisch’s even if based on different arguments, as an “orthodox” Cushitic language, as opposed to the other language groups described as “fringe” Cushitic. A few decades later, the late Morin () again took up the question of Beja’s position within Cushitic and made a cautious attempt to bridge the gap between Beja and another branch of Cushitic, Lowland EC, and more precisely ‘Afar and Saho. He grounded his hypothesis on the geographical contiguity that existed between the three languages in historical times and exemplified his comparative linguistic arguments with a few parallel tendencies found in phonetic and lexical features. The phonetic ones concern a tendency shared by Beja and cAsaurta, a dialect of Saho-‘Afar, to weaken the second or third (often final) consonant of a root, and to create diphthongs in fast speech due to glide deletion between vowels. The lexical argument concerns fewer than a dozen cultural (specialized, local, or “rare”) cognate or putatively cognate terms shared by all three languages. So far his attempt has not seen further advances.
. L
.................................................................................................................................. Lexicostatistics and glottochronology, which calculate degrees of retention in the basic vocabulary, have also been applied to classify Beja and Cushitic languages in general. Cohen (: ) showed that Beja has only approximately % of the basic vocabulary (on the basis of a revised list of items) in common with its closest neighbors, ‘Afar and Somali (EC), and with Agaw (CC). Moreover, it was hardly half of that ratio for the more distant EC languages, Sidamo and Oromo, for example. Nevertheless, Cohen noted that these figures were somewhat similar to what had been calculated for the different subgroups of Cushitic (e.g. % of common basic vocabulary between Oromo and ‘Afar but only % to % between Sidamo, Kambata, and Hadiyya), thus providing no reinforcement for Moreno’s classification. In a more recent glottochronological study, Blažek () added more data and languages and used the recalibrated method proposed by Starostin (). From his work we learn that the percentage of cognates between CC, a rather homogenous branch with respect to basic vocabulary, ranging from % to %, and Beja equals the calculation made by Cohen () (i.e. % of cognates in Blažek’s list of items). EC languages are much more heterogeneous and percentages of cognates rarely exceed %, and can at times be as low as % between the most distant language pairs. Blažek calculated a percentage of cognates of approximately % between Beja and Proto-EC, which he described as a “remarkable” closeness. But this relative proximity is due to the heterogeneity of the EC data and its important time-depth, which goes back roughly eight millenniums, as compared to a modern language, Beja. It thus does not contradict Cohen’s previous findings about the proportion of cognates between Beja and ‘Afar.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. E
.................................................................................................................................. Blažek (a, b, , ; Blažek and Dufková ) has been compiling a comparative and etymological dictionary of Beja, organized in semantic fields, and so far the results for fauna, kinship and social terminology, natural phenomena, time, and geographical terminology have been published. He did not attempt on this basis to revise the classification of Beja, but these works, based on regular phonetic correspondences and the identification of borrowings (not always specified as such though), pave the way for indepth comparative studies and provide ample data for further studies on the reconstruction and genetic classification of the language. A team of researchers headed by Guillaume Segerer has compiled a huge database of lexicons and dictionaries of African languages freely available online (http://www.reflex. cnrs.fr/database/). Some Cushitic languages, among them Roper’s () vocabulary of Beja, and AA languages are included. The various sets of tools for processing and analyzing the ,, lexical entries covering languages from , sources are meant to ease comparative research but remain to be exploited for Cushitic languages.
. T
.................................................................................................................................. When one is considering the classification and reconstruction of Beja, one big issue concerns the extent to which language contact with Semitic languages, Arabic in Sudan and Tigre in Eritrea, other Cushitic languages and NS, in particular Nile Nubian, has had, and still has, an influence on Beja vocabulary, morphology, and syntax. A lot of research remains to be done in this domain, but there are already a few studies of language contact between Arabic and Beja and between NS and Beja that can be mentioned.
.. Contact with Arabic It is uncertain when exactly the Beja people started to convert to Islam. What we know for sure is that after the Egyptian conquest of Sudan in the fourteenth century , the Beja people gradually converted to Islam and it is generally believed that all of them had become Muslims as far back as the sixteenth century (some think eighteenth or nineteenth century). The Arabic language arrived with Islam, and the Beja people, especially males, are nowadays very often bilingual (with various degrees of proficiency in Sudanese and “classical” Arabic). In Sudan, we are thus dealing with at least five centuries of language contact in recent history, which probably increased dramatically during the last fifty years, very different from the contacts Beja and other Cushitic languages have undergone in Ethiopia, Eritrea, and northern Kenya. Such a situation probably partially explains why Beja is so unlike other Cushitic languages in many respects. It is often mentioned that the Beja lexicon is full of borrowings from Arabic and EthioSemitic languages (e.g. Blažek a: ), which seems to be the case, but we are still lacking
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
explicit figures about the proportion of the vocabulary, and types of semantic fields concerned. Cohen (: ) noted that Beja is the Cushitic language with by far the largest proportion of triconsonantal roots (.%), and states that they are in majority of Arabic origin. Vanhove (forthcoming), on the basis of Reinisch’s dictionary () which mentions cognates with Semitic roots, calculated that % of verbal roots have a Semitic cognate. To what extent these are Semitic loans still remains to be clarified. At the morphological level, Beja is the language that has better than most retained the old prefix conjugation for all the indicative paradigms: % of the verbs are concerned, according to Cohen’s counting (: ), as opposed to % in ‘Afar and only five verbs in Somali and South Agaw. To what extent this is due to contact with Arabic remains to be clarified, since the two systems are organized very differently: while the prefix and suffix conjugations are lexically assigned in Beja, in Arabic the morphological distinction is aspect-based. Beja is also the Cushitic language where the non-concatenative morphology organized in stems and patterns, including stems with vocalic alternation, is the most pervasive, and typologically very similar to the particularly robust Arabic one. Nonconcatenative morphology is also found, but to a much lesser extent and without patterns with ablaut in the stem, in two neighboring Cushitic languages, ‘Afar and Saho, but not in any other Cushitic language. Vanhove () showed that despite the apparent similarity, Beja and Arabic have developed their own pattern system in terms of both semantics and forms, as flectional and derivational devices for the noun, verb, and adjective categories. Beja does not show any clear case of borrowing, copying, or replication from the patterns of dominant and prestigious Arabic. Instead, sociolinguistic and linguistic data favor an interpretation in terms of a convergence phenomenon in a large part of the Beja morphology, for which Arabic acted as a strong factor in the preservation of a cross-linguistically uncommon system. Such a situation is particularly tricky to disentangle, since we are dealing with related languages. It is even trickier because contact does not result in a change but is a matter of preservation of a proto-system (see also Owens, chapter of this volume).
.. Contact with NS Blažek () studied the lexical borrowings between Beja and the three NS languages that were in contact with it: Nile Nubian (Nobiin, Kenzi, and Dongola varieties), Kunama, and Nara. The author discusses the direction of thirty-six borrowings, which consist mostly of the comparisons provided by Reinisch. It turns out that Beja borrowings from NS amount to only nine lexical items, plus one dubious one, seven from Nubian, two from Kunama, and one from Nara (the item for ‘white’ was counted twice, as a borrowing from Nubian and from Nara). Two more loans, which came ultimately from Egyptian via Nubian, could be added to these figures; for ten items the direction of borrowing is not identifiable (and the twelve remaining items were borrowed from Beja into NS). All in all, it seems thus that contact with NS had very little impact on Beja, at least at the lexical level.
. D B
.................................................................................................................................. It has become a tradition to classify the different varieties of Beja by the names of the corresponding Beja tribes, with only few typological features mentioned. Morin (: –)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
is the first to have proposed a geographically based classification that he deduced from the observations made by Almkvist (–) and Roper (). One of the two varieties represents a northern and the other a southern dialect, to which Morin adds a transition zone. Apart from lexical specificities, the main distinctions he mentions concern the phonology of certain morphemes characterized by i in the south and u in the north, and a phonetic tendency to lengthen vowels (stressed or unstressed) in the south. These main divisions are further subdivided into sociolects: the Halenga tribe variety with its characteristic ablative morpheme s often features prominently in the literature about Beja. Each variety is also subdivided into local dialects, such as the variety of the area of Erkowit which is situated in the mountains above Port Sudan, close to Sinkat, and is considered as the “purest” Beja variety by Beja speakers (Vanhove )—for reasons that are still unclear to me. Morin refines his classification by distinguishing two southern varieties in the migration zone of the Hadendowa tribe, that of Sinkat, a contact zone with the Ammar’ar tribe, who speak a northern variety R (miːmhit beɖawije, also spoken by the Bishariyyin tribe), and that of the Gash area (gaː it beɖawije). Wedekind () proposes a somewhat different geographical division with three dialectal areas: North, Central, and South, to which he assigns the names of tribes or clans (p. ): Bishari, Atman, and Hadendowa. In this order they cover, to the East of the Nile river, (i) southern Egypt and northern Sudan as far as Haylab; (ii) the area north of Port Sudan (including some districts of Port Sudan) and Suakin; (iii) the region from Port Sudan to Kassala in Sudan and Teseney in Eritrea. He also mentions sociolectal and local variations. To the features mentioned by Morin he adds differences in the use of pitch accent in plural formation, the vowels of the definite article, and sociolinguistically based phonetic variants in borrowings from Arabic between urban and rural varieties. He also asserts the proximity of the three varieties using the Swadesh -wordlist, with differences ranging from % to %. This is similar to the retention rate calculated by Blažek () by means of recalibrated glottochronology.
R Almkvist, H. (–). Die Bischari-Sprache Tuu-Beḍaawie in Nordost-Afrika. Uppsala: Kaiserliche Akademie der Wissenschaften. Appleyard, D. (). ‘Beja as a Cushitic language’, in G. Takács (ed.), Egyptian and Semito-Hamitic (Afro-Asiatic) Studies in memoriam Werner Vycichl. Leiden and Boston, MA: Brill, –. Bender, M. L. (ed.) (). The Non-Semitic Languages of Ethiopia. East Lansing, MI: Michigan State University, African Studies Center, –. Blažek, V. (). ‘Cushitic lexicostatistics: the second attempt’, in A. Bausi and M. Tosco (eds.), Afroasiatica Neapolitana: Contributi presentati all’o Incontro di Linguistica Afroasiatica (CamitoSemitica), Napoli . Naples: Istituto Universitario Orientale, –. Blažek, V. (a). ‘Beja kinship and social terminology’, in M. Hasitzka-Johannes, D. G. Dembski, and J. Diethart (eds.), Das alte Ägypten und seine Nachbarn: Festschrift zum . Geburtstag von Helmut Satzinger. Krems: Österreichisches Literaturforum, –. Blažek, V. (b). ‘Fauna in Beja lexicon: A fragment of a comparative-etymological dictionary of Beja’, in L. Kogan and A. J. Militarev (eds.), Studia Semitica: Festschrift for Alexander Militarev. Moscow: Russian State University for Humanities, –. Blažek, V. (). ‘Natural phenomena, time and geographical terminology in Beja lexicon. (Fragment of a comparative and etymological dictionary of Beja.)’, in N. Koslova, S. Loesov, S. Tishchenko, and L. E. Kogan (eds.), Babel und Bibel : In memoriam Igor M. Diakonoff. Winona Lake, IN: Eisenbrauns, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Blažek, V. (). ‘Natural phenomena, time and geographical terminology in Beja lexicon. (Fragment of a comparative and etymological dictionary of Beja.)’, in L. E. Kogan, N. Koslova, S. Loesov, and S. Tishchenko (eds.), Babel und Bibel . Winona Lake, IN: Eisenbrauns, –. Blažek, V. (). ‘Beja lexicostatistics’, Lingua Posnaniensis : –. Blažek, V. (). ‘Beja versus Nilo-Saharan: on the lexical witness of mutual contacts’, Folia Orientalia : –. Blažek, V., and Dufková, K. (). Flora in Beja lexicon. Paper presented at the “Colloquium on African Languages and Linguistics”, August –, , Leiden. Browne, G. M. (). Textus Blemmycus: Aetatis Christianae. Champaign, IL: Stipes. Cohen, D. (). ‘Bédja’, in D. Cohen (ed.), Les langues dans le monde ancien et moderne: Langues chamito-sémitiques. Paris: Éditions du CNRS, –. El-Sayed, R. (). Afrikanischstämmiger Lehnwortschatz im älteren Ägyptisch: Untersuchungen zur ägyptisch-afrikanischen lexikalischen Interferenz im dritten und zweiten Jahrtausend v. Chr. Orientalia Lovanensia Analecta . Leeuven, Paris, and Walpole, MA: Peeters. Enguehard, F., Simeone-Senelle, M.-C., Vanhove, M., and Treis, Y. (eds.) (). Proceedings of the First International Colloquium on the Cushitic Languages and the Peoples Who Speak Them. Cologne: Rüdiger Köppe. Halévy, J. (). Essai sur la langue agaou: Le dialecte des Falachas (Juifs d’Abyssinie). Paris: Maisonneuve. Hetzron, R. (). ‘The limits of Cushitic’, Sprache und Geschichte in Afrika : –. Lamberti, M. (). ‘Cushitic and its classifications’, Anthropos : –. Moreno, M. (). Manuale di sidamo. Milan: Armondo Mondadori. Morin, D. (). ‘Des paroles douces comme la soie’: Introduction aux contes dans l’aire couchitique (bedja, afar, saho, somali). Paris and Louvain: Peeters. Morin, D. (). ‘Bridging the gap between Northern and Eastern Cushitic’, in A. Zaborski (ed.), New Data and New Methods in Afroasiatic Linguistics: Robert Hetzron in memoriam. Wiesbaden: Otto Harrassowitz, –. Reinisch, L. (/). Die Beḍauye-Sprache in Nordost-Afrika. vols. Vienna: Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften. Reinisch, L. (). Wörterbuch der Beḍauye-Sprache. Vienna: Alfred Hölder. Rilly, C. (). Le méroïtique et sa famille linguistique. Paris and Louvain: Peeters. Rilly, C. (). ‘Language and ethnicity in ancient Sudan’, in J. Anderson and D. Welsby (eds.), The Fourth Cataract and Beyond: Proceedings of the th International Conference for Nubian Studies. Leuven, Paris, and Walpole, MA: Peeters, –. Rilly, C. (). ‘Languages of Ancient Nubia’, in Dietrich Raue (ed.), Handbook of Ancient Nubia. Berlin and New York: De Gruyter, –. Roper, E. M. (). Tu Beḍawie: Grammar, texts and vocabulary. Hertford: Stephen Austin and Sons. Segerer, G., and Flavier, S. (–). RefLex: Reference Lexicon of the Languages of Africa, Version .. Paris and Lyons, http://reflex.cnrs.fr/database/ accessed July , . Starostin, S. (). ‘Sravniteno-istoričeskoe jazykoznanie i leksikostatistika’, Lingvističeskaja rekonstrukcija i drevnejšaja istorija Vostoka, –. Tosco, M. (). ‘Cushitic overview’, Journal of Ethiopian Studies : –. Tucker, A. N. (). ‘Fringe Cushitic: an experiment in typological comparison’, Bulletin of the School of Oriental and African Studies : –. Vanhove, M. (). ‘Roots and patterns in Beja (Cushitic): the issue of language contact with Arabic’, in M. Vanhove, T. Stolz, H. Otsuka, and A. Urdze (eds.), Morphologies in Contact. Berlin: Akademie-Verlag, –. Vanhove, M. (). Le bedja. Leuven and Paris: Peeters. Vanhove, M. (forthcoming). ‘Beja-Arabic contact’, in C. Lucas and S. Manfredi (eds.), Arabic and Contact-induced Change. Berlin: Language Science Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Wedekind, K. (). ‘Sociolinguistic developments affecting Beja dialects’, in M. Brenzinger and A.-M. Fehn (eds.), Proceedings of the th World Congress of African Linguistics, Cologne, – August . Cologne: Rüdiger Köppe, –. Zaborski, A. (). ‘Cushitic overview’, in M. L. Bender (ed.) (), –. Zaborski, A. (). ‘The position of Cushitic and Berber within Hamito-Semitic dialects’, in A. Bausi and M. Tosco (eds.), Afroasiatica Neapolitana: Contributi presentati all’ o Incontro di Linguistica Afroasiatica (Camito-Semitica), Napoli, – Gennaio . Naples: Istituto Universitario Orientale, –. Zibelius-Chen, K. (). ‘Sprachen Nubiens in pharaonischer Zeit’, Lingua Aegyptia: Journal of Egyptian Language Studies : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. C C (henceforth CC) is one of the four major branches within the Cushitic language family, widely known by non-linguists as Agäw. Linguists coined CC on account of the geographical location that speakers of the Agäw languages inhabit in contrast to the other Cushitic subfamilies, namely North Cushitic (NC), South Cushitic (SC), East Cushitic (EC), and West Cushitic (WC; later classified as “Omotic”). Despite ongoing doubts about the internal classification of Cushitic as to the status of Beja, and Southern Cushitic (Dahalo, for instance), Cushitic as a branch of Afro-Asiatic (AA) is well established (see Greenberg ; Zaborski ; Hetzron ; Ehret ; Hayward ; Sasse ; Mous ). The inclusion of CC within the Cushitic macro-family has been accepted ever since the first serious classificatory attempts by European scholars about a century ago. Reinisch () classified present-day CC as “Hoch-Kuschitisch” for the first time, together with the then Sidaama languages. In the s, Cerulli () introduced the label “CC” in his classification of Cushitic into NC, CC, Low Cushitic, and Sidaama. Moreno () subclassified Cushitic languages into two branches based on the ani-ati and ta-ne distinction in personal pronouns. The ani-ati group refers to CC, NC, and EC, and the ta-ne group to WC (i.e. Omotic). In , Hetzron reinforced Moreno’s classification and divided Cushitic into NC, CC, EC, and WC. Hetzron () proposed that Beja should be non-Cushitic, subsumed SC under EC, and introduced the term “Rift Valley Cushitic” by bringing together Highland East Cushitic (HEC) and CC languages on the basis of shared morphological features. Excluding Beja, Voigt () subdivided Cushitic into Saho-‘Afar, Omo-Tana, Agäw, and HEC, whereas Tosco () considered CC a distinct branch of Cushitic together with Beja and a large Eastern family. Hayward () proposed a sub-grouping into six branches: Northern, Central (Agäw), HEC, Lowland East Cushitic (LEC), Dullay, and SC. Zaborski (: ) envisioned bringing Agäw and LEC together by saying that “Agäw classified as ‘Central Cushitic’ so far is most probably just a branch of East Cushitic.” The position of Agäw within Cushitic has further been strengthened in the work of Ehret () and Sasse (). Appleyard (, , ) studied Agäw languages in some detail and underscored the fact that CC is a distinct branch with no special genetic proximity to any other branch of Cushitic. A similar view was held by Hetzron Palmer () and (, ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The most reliable internal classification of CC, containing Awŋi, Bilin, Kemantney, Xamt’aŋa, and a number of smaller and endangered varieties, goes back to Appleyard () and is shown below:
Proto-Agäw (CC)
Northern Agäw
Bilin
Eastern Agäw
Western Agäw
Southern Agäw
Xamir Xamt’a
Kemant, Quara, etc. Awŋi, Kunfäl
Apart from Bilin, which is now spoken in Eritrea, the rest are located in Ethiopia (see Map .). Awŋi is spoken by around , people, including monolinguals, and is a medium of instruction up to grade . Bilin has some , speakers, most of them bilinguals (Fallon ). Xamt’aŋa is spoken by roughly , people. Contrary to
. Distribution of Central Cushitic languages.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Grimes (), who declared the language extinct, it is now being used as a medium of instruction and, in addition, has been a school subject since . Kemantney is the most seriously endangered Agäw language spoken by fewer than , people, who are more proficient in Amharic than in Kemantney (Zelealem ). At present, the once widely spoken Agäw languages are endangered at different levels as a result of strong pressure from Amharic and Tigrinya. The CC languages exhibit peculiar grammatical features, especially in phonology and morphology, that clearly set them apart from other members of Cushitic. These salient features are outlined in the following sections, on the basis of data from primary and secondary sources.
. S
..................................................................................................................................
.. Presence of /ŋ/ The velar nasal appears word-medially and -finally in all Agäw languages. Word-initially it appears in Awŋi and Xamt’aŋa (see also Appleyard ). () Kemantney (Zelealem ) kiŋa ‘death’ säyiŋ ‘cloth’
Awŋi (Hetzron ) ŋaʧ ‘bone’ giŋ‘run’
() Xamt’aŋa (Appleyard ) käzɨŋ ‘cliff ’ ŋɨn ‘house’
Bilin (Appleyard ) gɨdɨŋ ‘dog’ kɨrɨŋa ‘stone’
Whereas the nasal series in CC is /m n ŋ/, it is /m n ɲ/ in other Cushitic languages.
.. Presence of /z/ The voiceless alveolar fricative s is one of the most common sounds in all Cushitic languages and was part of Proto-Cushitic. In contrast to this, the voiced alveolar fricative z is restricted to just a few languages. It is absent in Beja (NC) (Roper ; Wedekind ), SC languages (Elderkin ; Kießling , ), and LEC languages. HEC (Hudson ) and all CC languages, however, do have /z/ as a phoneme. () Awŋi (Hetzron ) Kemantney (Zelealem ) Xamt’aŋa (Appleyard ) zɨr ‘intestine’ zɨgra ‘guinea fowl’ zirwa ‘wheat’ sedza ‘four’ gɨzɨŋ ‘dog’ taz ‘maize’ As noted in Appleyard (), /z/ is one of the Proto-Agäw phonemes.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Presence of /ɣ/ The voiced velar fricative ɣ is one of the most frequently occurring phonemes in Agäw languages, where it shows up in all positions. It is not included in the consonant inventories of Proto-West Rift (PWR) of SC and EC. () Awŋi (Hetzron ) ɣəmbi ‘nose’ əɣoɣ‘laugh’ liɣ ‘hundred’
Kemantney (Zelealem ) liɣ ‘hundred’ aɣoy ‘head’ laɣla ‘bee’
According to Hetzron (), the frequent occurrence of the voiceless and voiced velar fricatives makes the occurrence of the glottal h rare in Agäw languages. In Awŋi, ɣ replaces h in Amharic loanwords, while, in Xamt’aŋa, h is used in free variation with x (Appleyard ). Appleyard () proved the existence of x and ɣ but not of h in his reconstructed Proto-Agäw consonant inventory.
.. Presence of central vowels Sasse () writes that almost all Cushitic languages have a five-vowel system /i e a o u/, which is a feature of Proto-Cushitic. Compare Zaborski () and Ehret () for Cushitic; Kießling (, ) for SC; Hudson () for HEC; and Bliese (), Owens (), and Saeed () for LEC languages. On the other hand, CC has a seven-vowel system including the two central vowels ɨ and ä. According to Appleyard (: ), “Long contact with Ethiopian Semitic, principally the prestige languages of ‘Abyssinian’ culture, first Ge’ez and then Amharic and Tigrinya, has had a considerable effect on all the Agäw languages individually and collectively.” Cf. the examples in ()–(), which include some minimal pairs. () Kemantney (Zelealem ) ɨrfa ‘maize’ käb- ‘cut’ arfa ‘month, moon’ kab- ‘help’ () Awŋi (Hetzron ) tsɨlli ‘little’ ɨʃʃi ‘meat’ dɨmmi ‘red’ ʤɨgɨr ‘wall’ () Xamt’aŋa (Appleyard ) nän ‘hand’ ɨqat- ‘wash’ bär- ‘leave’ ɨʧɨwa ‘mouse’ () Bilin (Appleyard ) amära ‘year’ sɨna ‘butter’ fär‘go’ kɨr- ‘die’ Appleyard (, ) included the two central vowels in the Proto-Agäw vowel chart. Hetzron () and Joswig () doubted the phonemic status of these vowels, as they
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
mostly appear as epenthetic vowels. The rare occurrence of especially ɨ in word-initial position, as in ɨŋi ‘ant’ and ɨri ‘rain’ in Awŋi, ɨʧ ɨwa ‘mouse’ in Xamt’aŋa, and ɨxwär ‘dried dung’ in Bilin, might prove that it is a marginal phoneme. Besides, speakers of Agäw languages do not replace ä and ɨ by non-central vowels as is the case among speakers of other Cushitic languages.
.. Lack of ejectives Ejective sounds are absent in CC but widespread in the rest of Cushitic and AA at large. In CC languages, ejectives are recorded only in loanwords from Semitic languages (Hetzron ). Appleyard (, ) pointed out that ejectives were not part of the Proto-Agäw consonant inventory. In Kemantney, four ejectives were recoded: t’, ʧ ’, k’, and s’. However, they all appear in loanwords from Amharic, as in t’ena ‘health’, biʧ’a ‘yellow’, s’af- ‘write’, k’äst ‘arrow’, etc. (Zelealem ). In Xamt’aŋa, Appleyard () identified a few words with ejectives: k’äk’ma ‘antelope’, ʧ’ik’a ‘bad’, ŋɨʧ ’ir ‘black’, gäs’ ‘face’, s’amra ‘friend’, and a few others. Any close examination of these few words proves that most of them are not genuine Agäw. The words for ‘bad’, ‘face’, ‘black’, and ‘friend’ are Semitic, with phonological and semantic modifications. Hetzron () mentioned Bilin k’ and k’w as the only ejectives found in genuine Agäw words. But according to Palmer (), these ejectives are a result of influence of Semitic, especially Tigrinya. They are replaced by the uvular q or the velar fricative x, as in Bilin k’wal-, Xamt’aŋa qal-, Kemantney xal-, and Awŋi kant- ‘see’. Fallon () identified three ejectives, t’, ʧ ’, and k’, in Bilin; cf. laŋk’ ‘tongue’, k’wɨlɨħ-a ‘child’, and lak’ɨl ‘bee’. However, there are very few words in the language with ejectives (the most frequently occurring is perhaps k’). Moreover, the first two words are of Semitic origin (cf. Ge’ez laŋk’a ‘palate’ and Tigrinya k’wälɨħ-a ‘child’). Although Fallon () argued that ejectives (notably the velar ejective) can be attributed to Proto-Agäw, and, hence, not all ejectives come from Ethiopian Semitic, this requires further investigation. Besides, the cognates of these words contain non-ejective sounds as in läk’il (Bilin) (Appleyard recorded laxla), läla (Xamt’aŋa), laɣla (Kemantney), and tsɨɣari (Awŋi) ‘bee’; laŋk’(i) (Bilin), laq (Xamt’aŋa), lanxɨ (Kemantney), and tsaŋ (Awŋi) ‘tongue’. Bilingual Awŋi native speakers substitute ejectives with their non-ejective counterparts in their Amharic speech, as shown in () below. () Amharic k’et’o bak’ela aʧ’ʧ’edmärrät’ʧ’äw
‘green pepper’ ‘bean’ ‘harvest’ ‘choose’ ‘salt’
Awŋi keto bakela atsedmeretʧ iwi
This substitution may indicate that ejectives are non-indigenous in Agäw. As the consonant inventory of Beja too contains no ejective phonemes (see Roper ; Wedekind ), the lack of ejectives is a shared feature of CC and NC.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Lack of implosives The absence of implosive sounds is another phonological feature that sets CC apart from the Cushitic languages spoken in Ethiopia. Implosives are widespread in EC languages and rare in SC, whereas the Beja consonant charts in Roper () and Wedekind ) contain no such implosives.1 Thus, the lack of ejectives too moves the CC languages close to NC.
.. Lack of pharyngeals The lack of pharyngeal sounds (ʕ and ħ) is typical of modern Agäw languages (Appleyard ). Mous (n.d.) pointed out that HEC languages have ejectives but no pharyngeals; they do not appear in Beja either. SC languages have retained pharyngeals which, according to (Kießling , ), are even reconstructible for Proto-SC and Proto-West Rift. They also exist in some EC languages beyond HEC. An exception within Agäw is Bilin, which has ʕəl ‘eye’ and ʕawad ‘cotton’ (Palmer ). Compared to əl and yəl in other Agäw languages where the latter word in particular turns out to be of Arabic origin, their genuine Agäw descent is debatable. Appleyard () also claims that the Bilin words däħna ‘well’, ħagay ‘dry season’, ʕatär ‘chickpea’, ʕastär ‘sky’, and yet others are all of Semitic origin.
.. Lack of /ɲ/ The palatal nasal occurs in the consonant inventories of all Cushitic branches except CC. It is part of PWR, as reconstructed by Kießling and Mous (), of HEC, as shown in Hudson (), and of Beja (Wedekind ). ɲ is recorded in CC Kolisi, but only in the SG personal pronoun ɲi , where it is merely due to the effect of the high front vowel i. The same pronoun is ŋi in Awŋi, ni in Bilin, and ŋäŋ in Kemantney and Xamt’aŋa. In Agäw, ɲ is attested only in loanwords from neighboring Semitic languages, mainly Amharic.
.. Lack of vowel length One of the major features of CC languages is the absence of long vowels. The oppositions of short and long vowels in the rest of Cushitic are: i : ii, e : ee, u : uu, o : oo, and a : aa. Vowel length is significant in Beja (NC) (Wedekind ), EC (e.g. see Gragg for Oromo) and SC as well (for examples, see Kießling for Burunge and Mous for Iraqw).
.. Lack of consonant gemination Gemination of consonants is common in Cushitic (Mous ). In Oromo (EC), for instance, gemination is phonemic, and all consonants geminate with the exception of h. In Konso (EC), Wedekind () recorded one example, ɗehay ‘people’, which might be due to borrowing or hypercorrection on the part of the informant or perhaps transcription inaccuracy. 1
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
gemination takes place with all consonants (Ongaye ). In K’abeena (EC) all except the glottal consonants geminate (Crass ). According to Tosco (), “Most Dahalo [SC] consonants can be geminated, and gemination plays an important role in verbal morphology.” Only one example is recorded from Beja (NC): asa ‘arise!’ vs. assaa ‘do!’ (Wedekind ); hence it seems difficult to consider gemination a phonemic feature in NC. In CC languages, consonant gemination is not phonemic. It occurs at morpheme boundaries as a result of assimilation, as in ɨy-wa > ɨy-ya ‘me’. Hetzron () provides some data from Awŋi, where gemination plays a phonemic role. These are anetsɨŋ ‘churn’ vs. annetsɨŋ ‘carpenter’ and dada ‘path’ vs. dadda ‘on the road’. The lack of distinctive gemination in Bilin, Kemantney, and Xamt’aŋa and its occurrence only in Amharic loanwords prove that this feature is lacking in Agäw in general. Besides, Appleyard’s () reconstruction of Proto-Agäw does not show consonant gemination as a distinctive feature.
. S
..................................................................................................................................
.. Personal pronouns NC makes gender distinction in both SG and SG. SC languages show gender distinction in SG only. In EC it is confined to SG. In CC, however, neither SG nor SG are marked for gender. The only exception among Agäw languages is Xamt’aŋa, where gender distinction is observed in SG. This unique feature appears to be the result of influence by Ethio-Semitic languages, presumably Tigrinya (Appleyard ). The Awŋi (CC) personal pronouns are given in () below. () polite polite
SG ani ɨnti ɨntu ɲi nay
PL nu ɨntu ŋa
The main difference between Agäw and other Cushitic languages is in the third person pronoun, where a gender distinction is not attested. CC deviates from EC and SC, which have retained the AA s~ʃ in SG. Beja (NC) to some extent shares this feature in its clitic forms -hoos (SG) and -hoosna (PL) for object, and -s (SG) and -sna (PL) for the possessive. As shown in () below, Proto-Agäw plural pronouns (see Appleyard ) exhibit different forms from the rest of Cushitic. ()
Singular Subject Oblique *an *yɨ *ɨnt *kwɨ *ŋi *ŋɨ
Plural Subject Oblique *ɨnn *ɨn(a)*ɨntɨn ~ än *ɨnt(a)*ŋa *ŋa-
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Another difference between CC and the rest could be the occurrence of independent honorific forms for the second and third persons in CC, similar to Ethiopian Semitic languages.
.. The unmarked nominative/marked accusative case The case system in Proto-Cushitic is reconstructible as *-i (also *-u) for subject and *a for absolute case (Sasse ). The traces of -i masculine subject case are still attested in Agäw; however, the situation is complicated where the old subject case got lost with the development of a marked object case (Appleyard ). König (: ) writes that “Within Cushitic, marked nominative is found in EC and NC only. CC has accusative and SC no case system at all.” According to Kießling (), though, SC languages do not mark case on nominal arguments themselves; they follow the strategy of identifying syntactic subjects and objects through clitics. EC languages mark nominative case in the majority of cases by -i/-n. Beja (NC) pronouns distinguish nominative from accusative (Vanhove ). Hence, in both singular and plural, the nominative has accusative counterparts. In declarative sentences, the nominative subject and the accusative object are identified simply by word order (Roper ). In CC languages, nominative is unmarked but accusative is marked by the suffix -s for masculine and by -t for feminine. Awŋi has -o/-wa (Hetzron ) and Kolisi -wa (Zelealem, chapter of this volume). Object pronouns in Bilin, Kemantney, and Xamt’aŋa also show the ending -t (Appleyard ). ()
()
yir zaf-ɨs käb-ø-ɨɣw. man tree- cut-.- ‘The man cut the tree.’ lämma fiʧirä-yän-ti kiw-u. Lemma goat-- kill:-. ‘Lemma killed the goat.’
[Kemantney] (Zelealem ) [Xamt’aŋa] (Appleyard )
The marked nominative hardly survives in Agäw due to the influence of Semitic. As shown in examples () and () above, the accusative is very prominent instead.
.. Agreeing genitives In genitive constructions, determiners, and adjectives agree with the head noun in gender and number and sometimes case, as in the following Awŋi PP constructions: tsɨnkut-des ŋɨn-des () a. an-des that- nice- house- ‘from that nice house’ b. ann-i-des tsɨnkut-ka-des ŋɨn-ka-des that-- nice-- house-- ‘from those nice houses’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
In the Awŋi NPs in (), the nouns agree in gender: ()
a. muri-w aqi village-(:) man(:) ‘the man of the village’ b. muri-t ɣuna village-(:) woman(:) ‘the woman of the village’
(Hetzron )
Agreeing relative verbs were recorded from Kemantney (Zelealem ). Palmer () recorded genitival agreement involving number, gender, and case in Bilin. In Xamt’aŋa, the case marker is attached to the possessor noun in a possessor–possessee construction, preceding the agreement suffix. Outside of CC, Hetzron () stated that genitive agreement in number and gender is infrequently found only in HEC, Burji, and Gedeo (see also Hudson ).
. C
.................................................................................................................................. Syntactically, CC languages display the Cushitic basic SOV word order. Following Heine’s () classification, Agäw languages are members of type D. In CC, the finite verb is clause-final. Postpositions dominate over prepositions. Tosco () presents a typological overview of the word order patterns of NP in selected Cushitic languages where the CC languages have not been represented. According to him, the following are the widespread NP patterns in CC: () Awŋi Bilin Kemantney Xamt’aŋa
: + + + +
: + + + +
: + + + +
: + + + +
: + + + +
: + + + +
Given this, CC languages tend to be very similar to the HEC languages and to Saho-‘Afar, a LEC language. This proves that CC languages are predominantly head-final, which sets them apart from LEC, SC, and NC languages, which all exhibit head-initial word order (see Tosco ; Mous ). In CC, head-final order also occurs in verbal, prepositional, and adjectival phrases. Appleyard () notes that CC preserves the Cushitic system of vocalic suffix inflection chiefly in subordinate verb forms. Hence, one of the major features that distinguish CC from the rest of the Cushitic family is exhibited in the suffix conjugation of main and subordinate verbs (Appleyard ): main verbs follow the pattern STEM+AGR+TAM; subordinate verbs show STEM+AGR+SUBORDINATOR order.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. L
.................................................................................................................................. Bender () suggested –% lexical similarity between CC and NC and –% similarity between CC and EC (see also Zaborski ). There also exist a number of cognates between Proto-Agaw and Proto-EC, e.g. *ʔay(y) and *ʔaw ‘who’, *ʃamb and *samb ‘lung’, *gilb and *gɨrb ‘knee’, etc. (Appleyard ). Compare the following cardinal numerals in the four branches of Cushitic: ()
Beja (NC) ngàal mhaloo mhày faDig (ʔ)áy (ʔ)asagwir (ʔ)asaramaa (ʔ)asamhày (ʔ)as’s‘aDìg tamin
PEC *mitto/*tokko2 *lamma *sazziħ *ʔafar/ʃoole3 *omute *laħ *lamala *sadzee *honso *taman
PSC *wak *ɬaama *tami *haaka *koo’ani *lahhoo’u *faanqu *dakati *gwaleeli *mebi
PCC *la *laŋa *säƔwa *säʒʒa *ʔankwa *wal *laŋätta *säɣwätta *säʧta *ʧika
Gloss ‘one’ ‘two’ ‘three’ ‘four’ ‘five’ ‘six’ ‘seven’ ‘eight’ ‘nine’ ‘ten’
Whereas there are at least two similarities between EC and SC (cf. the numerals ‘’ and ‘’), Beja and CC are quite different. Of the two, Beja shares cognates with EC and SC in the numeral ‘’ and distantly in ‘’. CC languages share cognates with EC and SC in the numerals ‘’ and ‘’. These digits (–) have the following shapes with striking similarities in the four CC languages in which CC numerals show inherent gender and, hence, the digits and the base ‘’ are all feminine (see () below).4 The same property has been identified in SC Iraqw (Mous ). ()
Awŋi láɣú láŋa ʃúɣa sedza ánkwa wólta láŋéta sóƔéta sésta ʧikka (Hetzron )
Bilin laxw läŋa säxwa säʤa ʔankwa wälta läŋäta säxwäta sässa ʃɨka (Hudson )
Kemantney laɣwa (l)niŋa seɣwa säʤʤa ankwa wälta näŋäta saƔwäta sässa ʃika (Zelealem )
Xamt’aŋa läw liŋa ʃaqwa siza akwa walta laŋta säwta s’ayʧ’a s’ɨk’a (Appleyard )
Gloss ‘one’ ‘two’ ‘three’ ‘four’ ‘five’ ‘six’ ‘seven’ ‘eight’ ‘nine’ ‘ten’
2 No single proto-form can be reconstructed for ‘one’ in PEC: *mitto (PHEC), *tokko (PLEC), *kow (PSam), etc. 3 *ʔafar (PHEC), *ʃoole (PLEC). It also includes other forms like *salaħ (Proto-Dullay). 4 For details on individual Agäw languages and their proto-forms, see Appleyard ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Core vocabulary items between CC languages and the rest Beja (NC)
Iraqw (SC)
Oromo (EC)
Awŋi (CC)
Kemantney (CC)
Gloss
ragad/lagad kwire hami/d.ifi gin'a gau/’anda hindi ʃ a/dof yam d.efa/yaf siyam/ʃ uʃ hadal/hadod adal/adar darab reba fena/gwiʃ ’a ‘amis/’adi n’e mek bissa/kaffa yas/mani ʃ ’a/yiwe hada galaba/karai kwa san de baba s’a/t.ata.m diw/nari tam/’am gw’a/ʃ ifi dir hadid/kwinh ‘iyai/bilil dah/l’a dis/dabali win/ragaga
yaaee sihhinoo se’eeengw muuná do’ slupi fu’naay ma’ay piindo gitsoo boo daa/aat loohi tlooma *laabala (PSC) *hhada ’asla daqwaay nyauw seeaay slee diraangw *bahaa (PSC) hat’ay nana aayi baaba iwiit guu‘ aag wah gaas ’oo’ *’iiraw (PSC) *du/iya (PSC) *niinaw (PSC) *dir (PSC)
miila/luka ilkee dabbasaa onnee mana mukha foon biʃ an balbala ʧ ’itaa gurraʧ ʧ a diimaa karaa/godaana tuullu waraana ulee/dullaa ibidda haare adure seere sa’a lenʧ ’a waraabo obboleeytii obboleessa haaɗa aabba taa’uu rafuu ɲaaʧ ʧ u ɗugaaiti aʤʤeesuu dubbattu hap’ii furdaa t’innoo guddaa/dagaaga
lɨkw ɨrkwí ʧ iʧ ifí ɨʃ ew ŋɨn kani ɨʃ ʃ i aɣu lɨmʧ i/sank sigwi ʧ árkí dɨmmí dad kán werém gɨmb leg dɨɣwarí anguʧ ʧ a gɨséŋ ɨllwa wuʤi ɨɣwí séná sén ʧ wá tablí ɨnʤikwɣuryɣwzɨqkwdibsɨnʧ u morí ʧ ɨlí dɨngulí
lɨkw ɨrkw ʃ ibka lɨbäka nɨŋ kana sɨya axw bäla ʃ anka ʃ ämäna säraɣ gorwa dɨba ʃ ämärgina kɨnbɨ wɨzɨŋ dɨɣora damiya gɨzɨŋ käma gämäna wäya ʃ än zän gäna aba täkosɨmgänʤxwʤaxkwgämärk’ät’änwäfärʃ igwey fɨraq
‘foot’ ‘tooth’ ‘hair’ ‘heart’ ‘house’ ‘wood’ ‘meat’ ‘water’ ‘door’ ‘grass’ ‘black’ ‘red’ ‘road’ ‘mountain’ ‘spear’ ‘stick’ () ‘fire’ ‘donkey’ ‘cat’ ‘dog’ ‘cow’ ‘lion’ ‘hyena’ ‘sister’ ‘brother’ ‘mother’ ‘father’ ‘sit’ ‘sleep’ ‘eat’ ‘drink’ ‘kill’ ‘speak’ ‘thin’ ‘fat’ ‘small’ ‘big’
(Roper )
(Kießling ) (Mahdi )
(Appleyard ) (Zelealem )
The core vocabulary items from five selected languages in Table . also illustrate the distant relation between the CC languages and the rest. Through time the CC lexicon has shown striking similarities with Ethio-Semitic languages. The sample of cognates in Table . is taken from Appleyard’s () dictionary.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Selected cognates in CC languages and Ethio-Semitic (Amharic) Amharic
Awŋi
Bilin
Kemantney
Xamt’aŋa
Gloss
kässäsmäkkärɨnsɨsa awre bära/boha/ mälat’a tɨhwan ammänmärrärakal k’ulk‘wal doro märrätgärräzbunna mido t‘ɨt’ agär azzo k’ura ɨwway al-
keseskusɨŋɨnsɨsi bɨʤi baɣwi tɨɣwana amnmeretakalat kulkùlí duri wogɨngerezbun ʃukat tɨti ager azo qura ɨway y-
käsäsmäxwäxwʔɨnsus wanin bärrah tɨxwana ʔämnmärärgärob k’wɨlänk‘wal dirwa gwälälsämätärbun kɨlal tɨt bɨra ħargäʧ ’ kwaxwɨra waʔ y-
käsäsmäkärɨnsɨsa awre bära täxwana amänmärarakal kwäkwäla dirwa märät‘gäräzbuna mido tɨya agär azo xwäräy ɨw y-
kässmäxɨrɨnsɨsi aräwi mɨlat’ dɨhwana amnmärrakäl
‘accuse’ ‘advise’ ‘animal’ ‘animal’ ‘bald’ ‘bedbug’ ‘believe’ ‘bitter’ ‘body’ ‘cactus’ ‘chicken’ ‘choose’ ‘circumcise’ ‘coffee’ ‘comb’ ‘cotton’ ‘country’ ‘crocodile’ ‘crow’ ‘cry’
ʤirwa mars‘garzbun mida t’it’ ħagɨr aqsa qurā’ qayy-
Bender () suggested a –% lexical similarity between CC and Semitic. Appleyard (: ) pointed out that “All the Agäw languages contain a sizable body of lexical material which has cognates in Ethiopian Semitic.” Bilin is a bit divergent within Agäw. The words wanin ‘wild animal’, gärob ‘body’, and ħargäʧ’ ‘crocodile’ are all loanwords from Tigre, another member of northern Ethio-Semitic (Appleyard ). The word sämätär- ‘circumcise’ is connected to Ge’ez matata ‘cut off ’ and (dialectal) Amharic mättär- ‘chop meat into pieces’. In the same way, the Awŋi word wogɨn- is related to Amharic wäggän- ‘side’. Borrowing is not unidirectional. As there are several loanwords from Semitic in Agäw languages, it is believed that Agäw has provided old loanwords for Semitic including Ge’ez (see Appleyard ). Amharic wuha ‘water’ is a loanword from Agäw axw. On historical mutual influence between Agäw and especially northern Ethio-Semitic Appleyard (: ) writes: Because of their location, the Agäw have been over the centuries in large part absorbed into the Semitic language and culture first of the Aksumite kingdom and then of Christian Ethiopia. Agäw languages thus provide the deepest-level and major linguistic substratum of Tigrinya and Amharic. There are, indeed, already some traces of Agäw linguistic influence in Ge’ez, mostly lexical but also from the area of morphosyntax.
The lexical correspondence between CC and Ethio-Semitic can be explained in terms of Proto-AA origin as well as language contact. The similarities between CC and other Cushitic phyla are due to Proto-Cushitic origin, as explained in Ehret (, ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. C
.................................................................................................................................. CC is a well-defined language subfamily whose speakers are isolated from the rest of Cushitic but surrounded by Semitic speakers. The Agäw people played a pivotal role in sociocultural and sociolinguistic interactions, especially in medieval Ethiopia. Historical linguistics shows that Cushitic, in particular Agäw, influenced mainly the syntax of Ethio-Semitic languages. Agäw languages have also been influenced by Ethio-Semitic languages. Currently, the Agäw languages are under strong pressure from Semitic languages. The influence exerted between CC and Ethio-Semitic languages is a result of centuries of language contact. Compared to the other subfamilies of Cushitic, CC is perhaps the least studied. There are a number of varieties within this unit that are not well known as yet. Thus, the varieties of Xamt’aŋa still need to be studied. The relationship between Kemantney, Quara, Dambia, Kailigna, and Falashan awaits further clarification. Safe, endangered, and extinct languages ought to be identified. Dambia and Falashan are no longer spoken varieties. Quara was reported to be on the verge of extinction (Appleyard ). Kailigna is a derogatory name used by the Kemant people to refer to the variety spoken by the Bete-Israel people (Zelealem ). Kolisi (Kunfäl) ought to be described in depth in order to determine its status either as a dialect of Awŋi or as an independent language. Finally, it also appears that the synchronic lexical comparison between CC and Ethio-Semitic could perhaps yield more striking similarities. CC is subjected to scholarly scrutiny with its own exotic features. In addition to the aforementioned linguistic features that are believed to be found widely in CC (but are absent or rarely found in the rest of Cushitic), it is worth suspecting that CC might also exhibit the following possible salient features: • • • •
rare occurrence of ʧ phonemic status of labialized consonants /kw, gw, ŋw, xw, ɣw/, AGR and TAM inflections on verbs, and heterogeneous plural marking.
The -ak (-Vk) auxiliary in the main verb conjugation is purely an Agäw innovation (Appleyard ). Future descriptions should verify whether or not CC languages have tones, be they grammatical, lexical, or both—like in the rest of Cushitic. Whether or not the lack of clear tonal distinction in at least present-day Bilin, Kemantney, and Xamt’aŋa could be the result of the influence of Ethio-Semitic languages needs further investigation.
R Appleyard, D. (). ‘The internal classification of the Agäw languages: a comparative and historical phonology’, in J. Bynon (ed.), Papers of the Third International Hamito-Semitic Congress. Amsterdam and Philadelphia, PA: John Benjamins, –. Appleyard, D. (). ‘Agäw, Cushitic and Afroasiatic: the personal pronoun revisited’, Journal of Semitic Studies : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Appleyard, D. (). ‘A grammatical sketch of Khamtanga’, Bulletin of the School of Oriental and African Studies : –, –. Appleyard, D. (). ‘The vowel systems of Agäw: reconstruction and historical inferences’, in H. G. Mukarovsky (ed.), Proceedings of the Fifth International Hamito-Semitic Congress. Vol. . Vienna: Afro-Pub, –. Appleyard, D. (). ‘The position of Agäw within Cushitic’, in P. Zemanek (ed.), Studies in Near Eastern Languages and Literatures: Memorial volume for Karel Petráček. Prague: Oriental Institute, –. Appleyard, D. (). ‘Language death: the case of Qwarenya (Ethiopia)’, in M. Brenzinger (ed.), Endangered Languages in Africa. Cologne: Rüdiger Köppe, –. Appleyard, D. (). A Comparative Dictionary of the Agäw Languages. Cologne: Rüdiger Köppe. Bender, M. L. (). ‘The languages of Ethiopia: a new lexicostatistic classification and some problems of diffusion’, Anthropological Linguistics : –. Bender, M. L. (ed.) (). The Non-Semitic Languages of Ethiopia. East Lansing, MI: African Studies Center, Michigan State University. Bliese, L. F. (). A Generative Grammar of Afar. Dallas, TX: SIL. Cerulli, E. (). ‘Note su alcune popolazione Sidama dell’Abyssinia meridionale. II: I sidama dell’Omo’, Rassegna di Studi Orientali : –. Crass, J. (). Das K’abeena: Deskriptive Grammatik einer hochlandostkuschitschen Sprache. Cologne: Rüdiger Köppe. Ehret, C. (). ‘Proto-Cushitic reconstruction’, Sprache und Geschichte in Afrika : –. Ehret, C. (). Reconstructing Proto-Afroasiatic (Proto-Afrasian): Vowels, tone, consonants and vocabulary. Berkeley, CA: University of California Press. Elderkin, E. D. (). ‘Southern Cushitic’, in M. L. Bender (ed.), –. Fallon, P. (). ‘The velar ejective in Proto-Agäw’, in O. Akinloye and L. Moshi (eds.), Selected Proceedings of the th Annual Conference on African Linguistics. Somerville, MA: Cascadilla, –. Gragg, G. B. (). ‘Oromo of Wollega’, in M. L. Bender (ed.), –. Greenberg, J. H. (). Studies in African Linguistic Classification. New Haven, CT: Compass Publishing. Grimes, B. F. (ed.) (). Ethnologue: Languages of the World. Dallas, TX: SIL International. Hayward, R. (). ‘Afroasiatic’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –. Heine, B. (). A Typology of African Languages, based on the order of meaningful elements. Berlin: Dietrich Reimer. Hetzron, R. (). ‘Agaw numerals and incongruence in Semitic’, Journal of Semitic Studies : –. Hetzron, R. (). The Verbal System of Southern Agäw. Berkeley and Los Angeles, CA: University of California Press. Hetzron, R. (). ‘The Agaw languages’, Afroasiatic Linguistics : –. Hetzron, R. (). ‘The nominal system of Awngi (Southern Agäw)’, Bulletin of the School of Oriental and African Studies : –. Hetzron, R. (). ‘The limits of Cushitic’, Sprache und Geschichte in Afrika : –. Hetzron, R. (). ‘Genitival agreement in Awngi: Variation on an Afroasiatic theme’, in F. Plank (ed.), Double Case Agreement. Oxford: Oxford University Press, –. Hudson, G. (). ‘Highland East Cushitic’, in M. L. Bender (ed.), –. Joswig, A. (). ‘The status of the high central vowel in Awngi’, in S. Uhlig (ed.), Proceedings of the XVth International Conference of Ethiopian Studies. Wiesbaden: Harrassowitz, –. Kießling, R. (). Eine Grammatik des Burunge. Hamburg: Research and Progress. Kießling, R. (). ‘Some salient features of Southern Cushitic (Common West Rift)’, Lingua Posnaniensis : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Kießling, R. (). ‘South Cushitic links to East Cushitic’, in A. Zaborski (ed.) (a), –. Kießling, R., and Mous, M. (). The Lexical Reconstruction of West-Rift Southern Cushitic. Cologne: Rüdiger Köppe. König, C. (). Case in Africa. Oxford: Oxford University Press. König, C. (). ‘Case marking and linguistic geography’, in O. Hieda, C. König, and H. Nakagawa (eds.), A Geographical Typology of African Languages. Amsterdam and Philadelphia, PA: John Benjamins, –. Mahdi Hamid Muudee (). Oromo Dictionary I. Atlanta, GA: Sagalee Oromoo Publishing. Moreno, M. M. (). Manuale di sidamo. Milan: Mondadori. Mous, M. (). A Grammar of Iraqw. Hamburg: Helmut Buske. Mous, M. (). ‘Cushitic’, in Z. Frajzyngier and E. Shay (eds.), The Afroasiatic Languages. Cambridge: Cambridge University Press, –. Ongaye, O. (). A Grammar of Konso. Utrecht: Landelijke Onderzoekschool Taalwetenschap. Owens, J. (). A Grammar of Harar Oromo including a text and a glossary. Hamburg: Helmut Buske. Palmer, F. R. (). ‘The noun in Bilin’, Bulletin of the School of Oriental and African Studies : –. Palmer, F. R. (). ‘An outline of Bilin phonology’, in Atti del Convegno Internazionale di Studi Etiopici. Rome: Accademia Nazionale dei Lincei, –. Palmer, F. R. (). ‘Cushitic’, in C. T. Hodge (ed.), Afroasiatic: A survey. The Hague: Mouton, –. Reinisch, L. (). Die Chamirsprache in Abbessenien. Vienna: Sitzungsberichte der kaiserlichen Akademie der Wissenschaftern, Phil.-hist. Classe , . Roper, E. M. (). Tu Bed.awiε: Grammar, texts and vocabulary. Hertford: Stephen Austin and Sons. Saeed, J. I. (). Somali Reference Grammar. Wheaton, MD: Dunwoody Press. Sasse, H.-J. (). ‘Die kuschitischen Sprachen’, in B. Heine, T. C. Schadeberg, and E. Wolff (eds.), Die Sprachen Afrikas. Hamburg: Helmut Buske, –. Sasse, H.-J. (). ‘Cushitic languages’, in K. Brown, A. H. Anderson, L. Bauer, M. Berns, J. Miller, and G. Hirst (eds.), International Encyclopedia of Linguistics : –. Oxford: Oxford University Press. Tosco, M. (). A Grammatical Sketch of Dahalo. Hamburg: Helmut Buske. Tosco, M. (). ‘A historical syntax of East Cushitic: a first sketch’, in T. Bearth, W. J. G. Möhlig, B. Sottas, E. Suter, and N. Cyffer (eds.), Perspektiven afrikanistischer Forschung. Beiträge zur Linguistik, Ethnologie, Geschichte, Philosophie und Literatur. Cologne: Rüdiger Köppe, –. Tosco, M. (). ‘Cushitic overview’, Journal of Ethiopian Studies : –. Vanhove, M. (). ‘The independent personal pronouns in Beja: synchronic functions and diachronic perspective’, in A. Amha and M. Mous (eds.), Omotic and Cushitic Language Studies. Cologne: Rüdiger Köppe, –. Voigt, R. (). ‘Zur Gliederung des Kuschitischen: die Präfixkonjugationen’, in C. Griefenow-Mewis and R. Voigt (eds.), Cushitic and Omotic Languages: Proceedings of the Third International Symposium. Cologne: Rüdiger Köppe, –. Wedekind, K. (). Beja Pedagogical Grammar, http://www.afrikanistik-online.de/archiv// , accessed July , . Zaborski, A. (). ‘Cushitic overview’, in M. L. Bender (ed.), –. Zaborski, A. (ed.) (a). New Data and New Methods in Afroasiatic Linguistics: Robert Hetzron in memoriam. Wiesbaden: Harrassowitz. Zaborski, A. (b). ‘A note on the classification of Agäw as a branch of East Cushitic’, in A. Zaborski (ed.) (a), –. Zelealem Leyew (). The Kemantney Language: A sociolinguistic and grammatical study of language replacement. Cologne: Rüdiger Köppe.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
ß
. G
.................................................................................................................................. T South Cushitic (SC) languages have been established as a subgroup of the Cushitic family since Greenberg (: ). According to the most inclusive view (Ehret ), SC consists of eight languages, spoken in non-contiguous areas of central Tanzania with outliers in the Pare Mountains (northeast Tanzania) and in Kenya (see Map .). According to the Ethnologue (Eberhard et al. ), these are: Iraqw (Mbulu; , ()),1 Gorwaa (Fiome; , ()), Alagwa (Chasi; , ()), Burunge (Mbulungwe; , ()), Qw’adza (Ngomvia; extinct), Aasáx (Aramanik; ()), Dahalo (Sanye; ()), and Mbugu (Ma’a; , ()). A less inclusive view (Kießling ; Kießling and Mous ) restricts SC membership to the core of the West Rift (WR) languages Iraqw, Gorwaa, Alagwa, and Burunge, excluding the others for various methodological reasons.
. S
.................................................................................................................................. Whiteley’s pioneering studies (, ), the descriptive monographs of Nordbustad () and Mous (), the text collection of Berger and Kießling (), and the dictionary of Mous et al. () make Iraqw the best described SC language. While Burunge (Kießling ), Dahalo (Elderkin ; Tosco ), Ma’a/Mbugu (Mous ) Alagwa (Mous ), and Gorwaa (Harvey ) have received monographic treatments, Qw’adza and Aasáx are only represented in Claus () and Fleming (), beside the lexical data assembled by Ehret ().
1
Information given in brackets: alternative name(s), number of speakers.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ß
. The geographical distribution of South Cushitic languages
The lexical and phonological reconstruction of SC by Ehret () has been updated, expanded, and methodologically improved by Kießling and Mous (). Detailed grammatical and socio-historical reconstructions are provided by Kießling ().
. E
.................................................................................................................................. Drawing on preconceptions such as the “Iraqw group” (Meinhof ; Struck ), Greenberg’s establishment of SC was based on only nine SC attestations of lexical resemblances from a total of seventy-eight Afro-Asiatic comparative series. Due to this weak evidence, Tucker () still had Iraqw excluded as “Fringe Cushitic” from Cushitic proper. While Greenberg affiliates SC as a sister branch to North (NC), Central (CC), and East Cushitic (EC), Fleming () regards SC (“Old East African Cushitic”) as the most ancient split from Proto-Cushitic (PC), coordinate to the entire rest of northeast African Cushitic proper. Hetzron (: –) and Ehret (: ) propose a closer affiliation of SC with EC which is also reflected in the most recent classification (Tosco ) that groups SC as part of “Transversal Southern Lowland Eastern Cushitic”.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The difficulty with determining the exact genetic position of SC boils down to a cluster of several problems: (a) the lack of diagnostic phonological innovations which are independent of morphological innovation; (b) the impossibility of sorting out shared innovations from shared retentions (resulting from the absence of a comparable corpus of morphological reconstructions at the level of the other Cushitic branches, especially EC); and (c) the rather morphosyntactic and typological nature of many features which link SC to EC, e.g. the preverbal case marking system including benefactive {sa} and instrumental/comitative {ra}. The lexical gap between SC and the rest of Cushitic is fairly wide, as reflected in the scarcity of potential lexical cognates (fewer than of more than , PWR lexical reconstructions in Kießling and Mous ). While this might point to an ancient split of SC from the rest of Cushitic, as suggested by Fleming (), the absence of firmly established SC:EC sound correspondences, the formal identity of some basic body part terminology (e.g. *ʔafa ‘mouth’, *ʔila ‘eye’) and the over-representation of the extremely marked phoneme *tɬ’ in SC might well reflect the result of a period of strong relexification, probably from Khoisan sources (*tɬ’ possibly representing the imperfect rendition of various clicks by Cushitic L speakers), under conditions of extended and intensive language contact cum shift—which would have required much less time-depth. The most probable candidates for shared SC and EC grammatical innovations which would argue for a genetic link to the exclusion of CC and NC are imperative suffixes *-i (SG) vs. *-a (PL), preverbal clitic subject marking including a contrast of *a (first/second person) vs. *i (third person), a system of nominal plural suffixes including *-oo, *-ee, *-u, and *-aCza, the adjectival plural suffix *-an, the singulative suffix *-inoo, and the postpositions *i (allative) vs. *koo (background) (Kießling , ). However, as long as no reliable reconstruction of EC and PC grammar is available, the issue remains unresolved, since there is no means of separating shared innovations within Cushitic from shared retentions from PC times.
. I
.................................................................................................................................. The internal genetic classification of SC provided by Ehret (: ) sees a major split into three branches: Rift, Mbugu, and Dahalo. While Mbugu and Dahalo continue as single languages, i.e. Ma’a/Mbugu and Dahalo, respectively, the Rift branch consists of a Western and an Eastern section. WR splits into Iraqw and Gorwaa on the one side and Alagwa and Burunge on the other. East Rift includes Qw’adza and Aasáx. However, due to methodological dilemmas and lack of adequate documentation, SC has to be restricted to the WR subgroup. For both Qw’adza and Aasáx, central parts of the grammar, e.g. verbal inflection, remain completely unknown, and documentation is far too poor to allow for reliable incorporation into an in-depth phonological and morphological reconstruction. Moreover, the limited data available have been collected from “rememberers” in a situation of imminent language death, which makes it methodologically impossible to disentangle terminal speakerinduced distortions from genetically diagnostic innovations representative of the prior speech community at large. Ma’a/Mbugu, since Meinhof () notorious as an example of a mixed language and an extreme case of intertwining of Bantu and Cushitic (see also Hollington, chapter of this volume), is the result of at least two language shifts (Mous ), the first from a Cushitic
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ß
L to Pare and the most recent representing a deliberate effort at creating an ethnolinguistic register distinct from Pare by drawing on many available sources, including South Cushitic, Oromo, Dahalo, several Eastern Bantu languages, and Maasai. Dahalo straddles the borderline “between South and East Cushitic”, genetically distant from both (Rowe ), lexical and morphological affinities to Yaaku and Dullay rather pointing to an EC affiliation of Dahalo (Tosco ; Blažek and Tosco ). Since the WR group, therefore, remains the only undisputable SC core, the phonological, morphosyntactic, and lexical reconstruction of WR presented in Kießling () and Kießling and Mous ()—as the only methodologically sound representation of SC—should be used as a basis for comparison on a general Cushitic level. Once the position of WR-based SC within Cushitic has been resolved, it can be used as a yardstick for determining the position of Ma’a/Mbugu, Dahalo, and those vestiges of “East-Rift” that happen to be documented.
.. Typological profile of PSC PSC reconstructs with five vowels, plus length opposition and thirty-two consonants, including laryngealized occlusives (ts’, tɬ’, q’), pharyngeals (ħ, ʕ), labialized dorsals, a lateral fricative ɬ, while completely lacking voiced fricatives. The position of the pitch accent, only marginally distinctive, was largely predictable by phonotactic criteria. In nominal morphology, a covert gender system contrasts feminine (*t) with masculine (*kʷ) plus a marginal neuter class, broadly associated with plural. The complex number system reconstructs with eighteen plural and seven singulative suffixes, constituting a three-way number opposition of generic vs. singulative vs. multiple reference form, reflecting intricate interaction with the gender system: the number suffix assigns gender, while the gender of the singular preselects the plural marker. In verbal derivation, PSC employs suffixes for causative (*-is), mediopassive (*-it), progressive (*-aCz, *-VVCz), intensive (*-aaCz), durative (*-im), inchoative (*-aw), frequentative (*CV-), distributive (*CVCV-), applicative (*hiin-), and comitative (*ʔila-). Inflectional categories are distributed as verbal suffixes (person, number, and gender of subject: SG zero, SG/SG.F *-it, SG.M *-i, PL *-an, PL/PL = SG/SG.M + *-’i/*-ri; aspect (perfective *-i vs. imperfective *-a); mood (subjunctive *-ee); negation (*-baɬi); interrogation (*-V) and as proclitics in a Preverbal Clitic Cluster (PCCl) for person of subject (*ha / vs. *hi ), person, number, and gender of object (SG *ni, SG.F *ki, SG.M *ku, SG.F *ga, SG. M *gu, PL/N *gi, PL *haanti, PL *kunu/*huunku), case relations (comitative/ instrumental *ri, benefactive *sa), and tense (past *aa, perfect *in, sequential). This renders the basic structure of transitive clauses in PSC as S + O + PCCl + V, or S + PCCl + O + V. Syntactically, PSC reconstructs with moderate SOVcharacteristics: the finite verb is clausefinal, postpositions outnumber prepositions, but in the NP head modifier order prevails.
.. Internal development of the modern WR languages from PSC The linguistic history of SC (Figure .) is shaped by genetic lines of internal divergence (tree model) and two types of contact-induced change: internal convergence (dotted boxes) and convergence towards external targets (arrows pointing towards relevant splits).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Pre-West-Rift
Proto-Southern-Nilotic/Proto-Omotik-Datooga: lexicon, suffixes of nominal derivation
Proto-West-Rift (PWR) Pre-Datooga: lexicon
Pre-Datooga: lexicon, morphophonology, syntax, semantics Datooga: lexicon, morphophonology Swahili: lexicon
Pre-Alagwa
Proto-East African Khoisan: lexicon Proto-East African Bantu: lexicon, suffixes of nominal derivation, morphemes of verbal inflection Proto-East African Bantu: lexicon, syntax, morphosemantics
Proto-North-West-Rift (PNWR)
heterogeneous groups of Pre-Datooga and Bantu: phonology, morphological reduction and fusion
Proto-South-West-Rift (PSWR)
Proto-Iraqwoid (PIRQ) Pre-Burunge Iraqw
Gorwaa
Alagwa
convergence: lexicon
XatsooAlagwa?
Burunge
Proto-East African Bantu: lexicon, nominal and verbal inflectional morphemes Sandawe: lexicon Langi: lexicon Swahili: lexicon
convergence: lexicon, phonology, verbal inflectional morphology, syntax, semantics
. Synopsis of the historical development of PSC (PSC = PWR) Kießling and Mous (: )
PSC split into two different dialects: Southern (PSWR), with Burunge as its sole continuation, and Northern (PNWR), which includes Alagwa and Proto-Iraqwoid (PIRQ), the predecessor of Gorwaa and Iraqw. Internal convergence of Alagwa towards PSWR produced bundles of Burunge/Alagwa isoglosses in lexicon and derivational morphology, which could easily be wrongly taken to reflect genetic inheritance if it were not for contrary evidence from inflectional morphology which clearly links Alagwa to the Iraqwoid group within PNWR genetically. Transition from PWR to PSWR is marked by a split of *ŋʷ into *ŋ and *w, grammaticalization of *ʔafa ‘mouth’ as malefactive verbal prefix, grammaticalization of adverb *loʔi ‘really’ as optative, and Bantu-inspired changes in syntax and semantics, e.g. reorganization of the basic word order *SOV towards a pragmatically conditioned variation of *SOV ~ *SVO, continued identification of grammatical gender with sex, and convergence of the neuter gender with the semantic category of plural. Major PNWR innovations include intervocalic dental lenition (*d > *r), grammaticalization of locative postposition *ee for predicative function, substitution of demonstrative anaphoric bases by possessive ones, innovation of an addressee-proximal demonstrative suffix *-(ʔ)isi, incipient grammaticalization of numeral *wak ‘one’ as indefinite, innovation of verbal extensions *hara- (simultaneous action) and *ʔala- (posterior), and innovation of a prohibitive by grammaticalizing a periphrastic construction with *maw ‘abandon’. The transition from PNWR to Alagwa is marked by (a) apocope of vocalic suffixes restricting aspectual vowels *-i (perfective) vs. *-a (imperfective) to negative and interrogative, (b) infix genesis resulting from haplological contraction of stem and reduplicatory suffixes, e.g. kwari ‘year’ vs. kwaaraa ‘years’ which derives from *kwar-araa > *kwaØaraa (Kießling ), and (c) various contact-induced transfers from PSWR (e.g. optative, basic word order reorganization, and semantic reanalysis of gender). The PSWR intrusion into Alagwa is also visible in lexical double reflexes, e.g. yaʕaam (PNWR) ~ ʕiimb (PSWR) ‘agree’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ß
The most dramatic period of change in SC history is the transition from PNWR to PIRQ. Apart from morphological innovations (e.g. substitution of negative suffix *-baɬi with *-káŋ < *kaħ ‘lack’, and grammaticalization and fusion of *ma ‘what?’, *tam ‘even’, *bara ‘when’ in the PCCl), it is characterized by radical reductions which produce morphophonological alternations in nearly all grammatical domains, e.g. lenition of voiced stops (), vowel quantity alternations (, b), and tonal oppositions/tonogenesis (a). () Lenition a. Verbal inflection *fáar ‘count (SG)’ < *faad-a *fád ‘count (SG)’ < *faad-it-a *fàar ‘count (SG.M)’ < *faad-i-a
()
> *fáarà > *fáarø > *fáadtà > *fádtà > *fàaríyà > *fàarø
b. Verbal derivation *fadit ‘be counting’
< *faad-ad-it > *faaddit > *faddit
c. Nominal derivation *dawa ‘arm’ *dabee ‘arms’
< *daba < *dab-abee
> *fádø
> *fadøit
> *dabbee > *dabøee
Quantity alternations a. Nominal derivation *tuuʔaa ‘corpse’ *tuʔu ‘corpses’
*tuuʔʔu > *tuʔʔu > *tuʔøu
b. Verbal derivation *fool ‘dig’ *folit ‘be digging’
*foollit
> *follit
> *foløit
The emergence of Burunge is characterized by innovations that postdateR the period of PSWR influence on Pre-Alagwa: palato-alveolar realization of *ts’ as t ’, intervocalic voicing of *t and *k in grammatical morphemes, expansion of the PCCl by a subsystem of new tenses, and grammaticalization of the preverb ħaanti as a prohibitive marker.
. T SC
.................................................................................................................................. As center of the Tanzanian Rift Valley contact zone (Kießling et al. ), SC has been shaped by external and internal convergence to such an extent that all genetic splits could also be attributed to substantial transfers from contact partners such as Southern Nilotic (Datooga), East African Bantu (Langi, Gogo), and Sandawe. Roughly speaking, PSWR and Burunge have been formed by Bantu impact, while Alagwa results from “Burungization” caused by integration of Burunge wives into an offshoot of PNWR society. Processes of partial morphological reduction and contraction paving the way for the emergence of PIRQ were caused by imperfect PNWR acquisition by shifting groups of heterogeneous ethnic
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and linguistic backgrounds during a period of emigration, probably from the original home in Maʕaangwatáy. Increasing dominance of Datooga L-speakers among these language shifters left a Datooga imprint on PIRQ grammar, manifest in various types of morphophonological fusions, and the expression of emotional concepts and spatial relationships (e.g. innovation of prepositions ‘in’ and ‘near’ from body part nouns *guraʔ ‘belly’ and *ʔafa ‘mouth’, exploitation of the bovimorphic model for the innovation of prepositions, e.g. *gurúu ‘in; under’ < *guraʔ ‘belly’ and *daandúu ‘on; above’ < *daanda ‘back’). Every period of external influence on SC is reflected in specific loanword sets. On the other hand, loanword strata in East Bantu languages from central Kenya to southern Tanzania testify to the historical presence of former SC-speaking communities which have become extinct (Ehret ; Ehret and Nurse ; Nurse ).
R Berger, P., and Kießling, R. (). Iraqw Texts. Cologne: Rüdiger Köppe. Blažek, V., and Tosco, M. (). Between South and East Cushitic: reconsidering the position of Dahalo. Paper read at the “. Afrikanistentag”, September , , Cologne. Claus, H. (). ‘Die Wangómwia’, Zeitschrift für Ethnologie : –. Eberhard, David M., Simons, Gary F., and Fennig, Charles D. (eds.). (). Ethnologue: Languages of the World, nd edn. Dallas, TX: SIL International. Online version: http://www.ethnologue.com Ehret, C. (). Ethiopians and East Africans: The problem of contacts. Nairobi: East African Publishing House. Ehret, C. (). The Historical Reconstruction of Southern Cushitic Phonology and Vocabulary. Berlin: Dietrich Reimer. Ehret, C. (). Reconstructing Proto-Afroasiatic (Proto-Afrasian). Berkeley, CA: University of California Press. Ehret, C., and Nurse, D. (). ‘The Taita Cushites’, Sprache und Geschichte in Afrika : –. Elderkin, E. D. (). The Phonology of the Syllable and the Morphology of the Inflected Word in Dahalo. M.A. thesis. University of Nairobi. Fleming, H. C. (). ‘Asa and Aramanik: Cushitic hunters in Masai-land’, Ethnology : –. Fleming, H. C. (). ‘Chadic external relations’, in E. Wolff and H. Meyer-Bahlburg (eds.), Studies in Chadic and Afroasiatic Linguistics. Hamburg: Helmut Buske, –. Greenberg, J. H. (). Studies in African Linguistic Classification. New Haven, CT: Compass Publishing. Harvey, Andrew (). The Gorwaa Noun: Toward a Description of the Gorwaa Language. Ph.D thesis, SOAS, University of London. DOI: ./zenodo. https://doi.org/./zenodo . Heepe, M. (). ‘Hamitica I: Fiome-Texte’, Mitteilungen des Seminars für Orientalische Sprachen : –. Hetzron, R. (). ‘The limits of Cushitic’, Sprache und Geschichte in Afrika : –. Kießling, R. (). Eine Grammatik des Burunge. Cologne: Rüdiger Köppe. Kießling, R. (). ‘South Cushitic links to East Cushitic’, in A. Zaborski (ed.), –. Kießling, R. (). Die Rekonstruktion der südkuschitischen Sprachen (West-Rift): Von den systemlinguistischen Manifestationen zum gesellschaftlichen Rahmen des Sprachwandels. Cologne: Rüdiger Köppe. Kießling, R. (). ‘Infix genesis in South Cushitic’, in M. L. Bender, G. Takács, and D. L. Appleyard (eds.), Selected Comparative-Historical Afrasian Linguistic Studies in memory of Igor M. Diakonoff. Munich: Lincom Europa, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ß
Kießling, R. (). ‘Alagwa functional sentence perspective and “incorporation” ’, in A. Amha, M. Mous, and G. Savà (eds.), Omotic and Cushitic Language Studies: Papers from the Fourth Cushitic Omotic Conference, Leiden, – April . Cologne: Rüdiger Köppe, –. Kießling, R., and Mous, M. (). The Lexical Reconstruction of West Rift (Southern Cushitic). Cologne: Rüdiger Köppe. Kießling, R., Mous, M., and Nurse, D. (). ‘The Tanzanian Rift Valley area’, in B. Heine and D. Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, –. Meinhof, C. (). ‘Linguistische Studien in Ostafrika, X: Mbugu’, Mitteilungen des Seminars für Orientalische Sprachen : –. Mous, M. (). A Grammar of Iraqw. Cologne: Rüdiger Köppe. Mous, M. (). ‘Basic Alagwa syntax’, in A. Zaborski (ed.), –. Mous, M. (). The Making of a Mixed Language: The case of Ma’a/Mbugu. Amsterdam and Philadelphia, PA: John Benjamins. Mous, Maarten (). Alagwa—A South Cushitic Language of Tanzania: Grammar, texts and lexicon. Cologne: Rüdiger Köppe. Mous, M., Qorro, M., and Kießling, R. (). Iraqw–English Dictionary, with an English and a thesaurus index. Cologne: Rüdiger Köppe. Nordbustad, F. (). Iraqw Grammar: An analytical study of the Iraqw language. Berlin: Dietrich Reimer. Nurse, D. (). ‘Extinct Southern Cushitic communities in East Africa’, in M. Bechhaust-Gerst and F. Serzisko (eds.), Cushitic–Omotic. Hamburg: Helmut Buske, –. Rowe, J. (). Finding the Place of Dahalo in Cushitic: An investigation into the genetic classification of an East African language. Senior Honor’s thesis. University of Michigan. Struck, B. (). ‘Über die Sprachen der Tatoga und Irakuleute’, Mitteilungen aus den Deutschen Schutzgebieten, Ergänzungsheft : –. Tosco, M. (). A Grammatical Sketch of Dahalo. Hamburg: Helmut Buske. Tosco, M. (). ‘Cushitic overview’, Journal of Ethiopian Studies : –. Tucker, A. N. (). ‘Fringe Cushitic: an experiment in typological comparison’, Bulletin of the School of Oriental and African Studies : –. Whiteley, W. H. (). A Short Description of Item Categories in Iraqw. Kampala: East African Institute of Social Research. Whiteley, W. H. (). ‘The verbal radical in Iraqw’, African Language Studies : –. Zaborski, A. (ed.) (). New Data and New Methods in Afroasiatic Linguistics. Wiesbaden: Otto Harrassowitz.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
̈
. I
.................................................................................................................................. T Omotic languages are situated exclusively in southwest Ethiopia (see Map .). The total number of speakers is about four million, with two million belonging to the Wolaitta cluster (Hayward : –). It seems that the label “Omotic” was first used in print by Bender (), although Fleming () defined this entity in opposition to the Cushitic language family. The name derives from the river Omo, near which the languages are spoken; deviant labels are mentioned in section .. Following Bender (: ), the roughly thirty Omotic languages are classified thus: Northwest Ometo (Wolaitta cluster, Malo, Oyda, Basketo, Maale, etc.); Southeast Ometo (Koorete, Zayse, etc.); C’ara; Gimira (Benc’-She); Yem; Kefoid (Bworo, Anfillo, KafaMocha); Dizoid (Dizi, Sheko, Nayi); Mao (Hozo-Sezo, Bambeshi-Diddesa, Ganza); Aroid (Ari, Hamar-Bana-Kara, Dime). The degrees of resemblances between these nine groups vary considerably; for instance, the units from Northwest Ometo to Kefoid are also called “ta/ne languages” because of shared personal pronouns (see section ..). To this day, the Omotic family as outlined by Fleming () has been subject to vigorous debate. On various grounds, some authors called into question the idea of a homogeneous and relatively independent linguistic unit “Omotic” and, consequently, rejected this name. The discussion about the descent of Omotic languages is particularly complicated because of the Ethiopian linguistic area (cf. Ferguson ), whose extent is far from clear. Hence, genetic and areal considerations intermingle. The present chapter intends to shed light on the history of Omotic classification (section .) and to argue for an Omotic family besides Semitic, Egyptian, Berber, Chadic, and Cushitic (section .).
. H
.................................................................................................................................. The main focus of this account is on external relations, i.e. how Omotic languages are classified with regard to other language groups. Internal relations are, on the one hand,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
. The geographical distribution of Omotic languages
naturally affected by such external views. On the other hand, internal classification has remained quite stable over time—changes were rarely proposed. The history of comparative research until around has also been outlined by Fleming (a: –). This history is partly characterized by polemic debates.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Pre-“Omotic” endeavors Two travelers may be considered the first “Omoticists”: Abbadie and Beke. In his first classificatory attempt Abbadie (: –) listed twenty-eight Ethiopian languages, e.g. Djandjörou, Warata, and Gomara or Kafa. However, these three languages were grouped among those whose affinities are still unknown, whereas Semitic and Agaw were already considered families. Two years later Abbadie () provided more information about Ethiopian varieties and presented a label for the formerly unclassified ones, namely “Chamitic”. As such are counted—together with Cushitic Galla, Bĭdja, ‘Afar, and even Semitic Amhara—the Omotic representatives Sĭdama (that is Kafa in his view), Dawrooa, Yămma or Yangara, Gonga and Shay, and Nao. Among the “Negro tongues” are mentioned Dokko, Gamo, and Arouro, but it is uncertain whether “Chamitic” and “Negro” form a dichotomy. Abbadie based his considerations on lexical data consisting of up to , words for individual languages. Beke’s (: –) proposal was based exclusively on lexical comparisons. His wordlist of thirteen languages comprises Gonga, Kaffa (Abbadie’s “Sidama”), Worátta (Abbadie’s “Dawrooa”), Woláitsa, and Yángaro; other languages derive mainly from Agau (Central Cushitic [CC]) and Semitic. Beke (: –) described these Omotic languages plus Derbábbo (?), Mócha, and Afíllo as a group, even though he considered Yángaro quite deviant. Furthermore, he distinguished between “the Gonga and Agau languages”, i.e. between parts of Omotic and (Central) Cushitic. Next, Abbadie (: –) presented a classification of fifty-eight Ethiopian languages divided into five groups: Semitic, Kamitic, Sub-Semitic, Sub-Kamitic, and unclassified; it was based on a comparison of up to , words. Gonga, Kaface, Dawroa, Gazamba, Xe, Yamma, Xaka, Na’a, and Dollo, together with Cushitic Agaw, Bija, etc., are presented as Kamitic, whereas Otollo, two Xinaxa varieties, Zayse-Garduri and Balta remain unclassified. Furthermore, the “langues nègres” contain Zargulla and Koyra—and all these are only the names which match those of modern Omotic varieties. Reinisch (: –) divided the Ethiopian “Chamiten” into Low Cushitic (Bedscha, Saho, ‘Afar, Somali, Oromo) and High Cushitic (Agau, Sidama, Barea-Kunama).1 Here, the term “Sidama” comprises modern Omotic and modern Highland East Cushitic. He mentioned geographical and linguistic criteria for his classification. Praetorius (: –) put forward a more modern view: given Reinisch’s High Cushitic, he accepted Agau as Cushitic but doubted the Cushitic or even “Hamitic” affiliation of the other members. He thereby anticipated the fact that Omotic is a separate family and Barea-Kunama is part of the NiloSaharan (NS) phylum. Notwithstanding these doubts, Reinisch later () retained his dichotomy of Low Cushitic and High Cushitic in a broader morphological context. The Italian period commenced with Conti Rossini (: –) who looked at Reinisch’s Sidama only and divided it into four parts: Western (Cafa, Gonga), Eastern (today’s Highland East Cushitic), Southern (Dawaro, Ualamo or Ualaitsa, Haruro or Gatzamba, Badditu, etc.), and Upper Ghibié (Zengerò or Yangaro or Yemmā). Highland East Cushitic is mixed with three divisions of Omotic; the evidence is lexical, with a focus 1
The niderkuschitischen and the hochkuschitischen languages; the English versions are from Fleming (a: –).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
on numerals. Conti Rossini thus anticipated the attempt by Cerulli (: ). The latter assumed the same divisions within Sidāmā, but two names differ: the groups are Western, Omō or Omêti (Wålaittā, etc.), Eastern, and Northern (Yammā). Cerulli employed the river name “Omo” for a linguistic group, even if it only designates modern Ometo and not Omotic as a whole. Significant progress was made by Moreno (: –). He compared parts of grammar and lexicon to distinguish Cushitic and Omotic and developed a classification with concrete linguistic forms. The main heading is “Lingue Cuscitiche”, but then two divisions follow: A. “ani/ati” languages: North Cushitic (NC) (Bégia), CC (Agàu), EC (Low Cushitic, various languages; Burgi-Sidamo); B. “tā/nē” languages: West Cushitic (WC) (Yamna, Ometo, Ghimira, Gonga). Moreno’s crucial observation concerns the personal pronouns with Cushitic first person singular ani, second person singular ati vs. Omotic first person singular tā, second person singular nē. This implies the first clear borderline between Cushitic and Omotic in research history. Conti Rossini accepted Moreno’s view: he stated that the “Sidama” concept might be wrong (Conti Rossini : ) and adopted the new classification including WC—even if he assigned the recently discovered Aroid (i.e. South Omotic) languages Bacco, Amar, and Banna to Nilotic (Conti Rossini : –, –). Bryan () published an account of several languages of the region including Ganzo, Kerre, Dime, Amarr, and Mali, and placed them in very tentative geographic groups. After that she offered another version of “Sidama” (Bryan : –, especially –). She listed NC (Beja), CC (Agau, Sidama), and East Cushitic (EC) (Saho-‘Afar, Somali, Galla, Burji-Geleba), Sidama comprising the “languages” Janjero, Anfillo, and Shinasha, and the “dialect clusters” Sidamo (Highland East Cushitic without Burji), Ometo, Gimira, and Kaffa. Amarr and Bako are considered “non-Cushitic” (Bryan : ). Cerulli (: –) retained his Sidama concept with new evidence from Caffino. Tucker and Bryan (: ) began their Cushitic chapter with Moreno’s “sharp cleavage” between first person singular an, second person singular at in Bedawiye, Agau, Saho-‘Afar, Sidamo, Galla, and Somali, and first person singular ta, second person singular ne in Janjero, Ometo, Gimira, and Kaffa; Konso-Geleba is said to be “insufficiently known”. Some Mao varieties are classified as Koman; Bako and others constitute a group of their own (Tucker and Bryan : –, ). Deeper linguistic insights are not given. Parts of the lexicon and a few morphemes characterize the Afro-Asiatic (AA) classification by Greenberg (: –). His Cushitic subclassification is very modern, with Northern (Beja), Central (later Agaw), Eastern, Western, and Southern branches. As WC he counted thirty varieties twenty-nine of which, including Anfillo, Bako, and Amar, deserve this affiliation; only Tsamai is incorrect. In their grammar-based approach, Tucker and Bryan (: ) emphasized their threefold distinction within Cushitic by contrasting Cushitic languages (Bedauye, Agau, Saho-‘Afar, Sidamo, Galla, and Somali), showing “the main features”, with “partially” Cushitic languages (Janjero, Ometo, Gimira, and Kaffa), differing “in many important respects”, and Geleba, a “language with little or no claim to be” Cushitic. Not much later, Tucker (: ) created “Orthodox” Cushitic (Greenberg’s Northern, Central, and Eastern, minus Geleba-Mogogodo, plus Sanye) and “Fringe” Cushitic (Greenberg’s Western, Southern, minus Sanye, plus Geleba-Mogogodo, Ika-Tepes, and Hadza). Having compared personal pronouns and some verbal morphemes, Tucker (: ) concluded that Greenberg’s WC lacks some Cushitic patterns.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. The establishment of “Omotic” The paper by Fleming () is a revelation in comparative Omotic studies. He conjectured that “West Cushitic” probably is a sixth family within “Hamito-Semitic” based on “protoAri-Kafa” (p. ). Attention is paid to the Ari-Banna subgroup (pp. –) with several Nilotic morphemes and NS pronouns and lexemes, but a special phonology; Fleming considered Ari-Banna the earliest offshoot of WC. The general description (pp. –) focuses on lexical evidence to support Moreno’s assumption that WC clearly differs from other Cushitic groups. According to Fleming (p. ), other scholars had overlooked Moreno’s dichotomy. In fact, the name “Omotic” was introduced by Bender (: –), although he acknowledges Fleming for coining it. Apart from this new nomenclature, Bender repeated Greenberg’s classification for Ethiopian languages and included Ari, Banna, etc., in Eastern, and Maji, Southern Mao, etc., besides Ometo, in Western Omotic; Tsamai now belongs to Lowland East Cushitic. The lexicostatistical analysis shows that Ometo languages share more than % of basic lexicon, whereas the figures within other Omotic are between % and % (Bender : ); grammar is not considered. Nevertheless, Bender provided a treasure of lexical data. An overview of comparative research into Cushitic and Omotic languages was provided by Black (). Zaborski (: ) accepted the new view, although he saw a close Cushitic–Omotic relationship. Details on further languages were given by Fleming () to support his approach, whereas Bender (b) offered grammatical and especially lexical arguments for assigning the Mao languages to Omotic. His monograph (Bender a) touches upon Afro-Asiatic grammar and lexicon quite generally and comprises even species like Proto-Central Khoisan, Proto-Bantu, and Proto-Indo-European. Here, Bender (a: ) states that Omotic is AA, but is “the most divergent” member. In three reference papers, Fleming (, a, b) refined divisions and features of the Omotic family. He listed three South Omotic and fifteen North Omotic groups (Fleming : ); interestingly, the southern part is also “Eastern” and the northern part “Western” (Fleming b: ), but the north–south distinction remains canonical. Fleming (a: –) used pronominal elements, some morphemes, and about sixty lexemes to show that North and South Omotic belong together. To characterize Omotic, Fleming (b: –) mentioned features such as lack of pharyngeals, reduced grammatical gender, special pronominal patterns, and distinct words for concepts like “eye” (af-, ap- vs. Cushitic il-). These observations have become meaningful in later discussions.
.. Post-“Omotic” debates Like Bender, Ehret (: –) investigated some AA lexemes and regarded Omotic as the first to split off; the remainder is Erythraic (Chadic, Cushitic, North AA). Similar classifications of – were summarized by Ruhlen (: ), e.g. Newman’s exclusion of Omotic from AA and Bender’s South AA with Cushitic and Omotic. Bender (: –) reconfirmed the Mao languages as Omotic and included Ganza due to lexical evidence. Later he repeated this idea and clarified that Koman and Gumuz were rather NS (Bender ). Zaborski, expressing doubts about the genetic affiliation of Omotic (Zaborski : –) but also stating that it is “very different from Cushitic” (Zaborski : ), posed
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
a provocative question: he wondered whether Omotic could be revived as WC (Zaborski ). With reservations due to interferences and the scarce morphological data, Zaborski expressed the idea that Omotic might be “the most innovating group within Cushitic” (p. ). In the years that followed, internal relations of Omotic were investigated more than external ones. Bender (, ) reconstructed Proto-Omotic phonemes and lexemes and developed his ninefold subclassification from “O” North Ometo to “O” Aroid. Objections came from Hetzron (), who considered pronouns and verbal suffixes to prove that North Omotic was AA, but South Omotic remained doubtful. Nevertheless, Bender (, a) continued his studies on comparative Omotic grammar and defined “Cushomotic” with two closely related but independent families (Bender a: ). Hayward (b) edited the first collective volume completely devoted to Omotic languages, with his statement on the methodological benefits of separate Cushitic and Omotic families (Hayward a: viii), Bender’s (b: –) view of NS Komuz vs. Omotic Mao, and Zaborski’s () remarks on case suffixes as another possible Cushitic– Omotic link. Preserving the Italian tradition, Lamberti (: –) divided Cushitic into six families: Central (Agaw), East, Southern, Northern (Beja), West, and Ari-Banna. He later (Lamberti : –) gave reasons like shared nominal gender, number, and case suffixes and initiated a harsh conflict with Fleming (), who detected several shortcomings in Lamberti’s argumentation. As an outcome, Lamberti (: ) accused Fleming of unfounded “insults” and the latter (Fleming ) corroborated some points. It seems true that “the duel is not now serving any useful scientific purpose” (Fleming : ). Bender (: ) again compared lexical data from rather isolated languages and concluded that Hozo probably and Ari possibly belonged to Omotic. In yet another lexicostatistical contribution Bender () accepted “Cushomotic” as a valid grouping and strengthened his Omotic subclassification. Soon thereafter, however, Bender (: ) returned to his and Ehret’s idea that Omotic was the first to split off within Afrasian, where Macro-Cushitic comprises Berber, Semitic, and Cushitic. The evidence is grammatical, syntactical, and lexical. A new provocation was formulated by Zaborski (: ), who stated: “It is sad to see how easily the hypothesis about an allegedly independent status of Omotic has been uncritically repeated usually by people who know nothing about these languages.” He claimed that morphology and lexicon point towards WC rather than separate Omotic, although he also remained skeptical because of lacking or contradictory data and possible interferences. As a matter of fact, morphology and lexicon are the focus of Bender’s fundamental books on comparative Omotic studies. While the focus is on relations between the nine parts “O” to “O” as outlined already in the s, neither of the two monographs (Bender , ) places Omotic in a wider context. However, both works include lots of Proto-Omotic reconstructions which are based on a huge reservoir of data and show that the groups from Ometo to Mao and Aroid can be considered one genetic unit. Tosco (: –) also noticed the deviance of Aroid and especially Mao without doubting their Omotic membership. He stated that Omotic languages tend to have fewer phonemes than Cushitic ones. Hayward (: –) supported the hypothesis that Omotic with features like sibilant harmony was untypical within AA. Hudson (: –) presented a classification of seventy-five Ethiopian languages occurring in the national census, twenty-three of which are Omotic.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Another paper by Zaborski () again raised the question whether WC must be reinvented. According to him, many Omotic grammatical morphemes resemble Cushitic, AA, or NS ones or are inconclusive, but some decisive verbal paradigms point to Cushitic. Furthermore, Aroid and Mao are declared to be NS, mainly because of their personal pronouns. Theil () extended the debate by doubting at all the AA ancestry of Omotic. After a vast evaluation of grammatical and lexical data by Fleming, Bender, and others, he considered Omotic an isolated language family (p. ). The classification by Militarev (: –), entirely based on lexical comparisons and glottochronology, contains new evidence in favor of “Cushomotic” or South Afrasian (which emerged in ), where Ongota is South Omotic, vs. North Afrasian (which emerged in ). The pioneer Fleming (: –) promoted Ongotan as a primary branch of AA while pleading for Omotic as the latter’s earliest separate. The major arguments still lie in the lexicon. In his final comparative paper, Bender (: –, ) elaborated on the ta/ne pronouns and expressed the view that Omotic has some clear Afrasian or Cushitic features and is otherwise somehow divergent. Blažek () undertook another lexicostatistical study of Omotic within its AA context. He came up with quite high cognate rates between Omotic and EC (but also Chadic) and argued for an independent Omotic branch because the EC resemblances can be explained at least partially by borrowing (p. ). The position paper by Bechhaus-Gerst (: ) ends with methodological reflections on the sense of ancient proto-families as opposed to small language groups, but not with statements about the Cushitic–Omotic relationship.
. O
.................................................................................................................................. In spite of Bechhaus-Gerst’s pessimism regarding definite hints as to the position of Omotic, a few things can be said about linguistic structures. Phonology and lexicon are less informative for this purpose than morphology; this has been Zaborski’s main concern for decades. And here, for example, Omotic is characterized by a remarkable wealth of verbal portmanteau suffixes (Bender : ), where Cushitic tends to use particles, socalled “selectors” (Mous ). With regard to interrogative sentences, Omotic languages generally have question markers in their verbal morphologies (Köhler ); some languages, mainly from Ometo, even mark yes/no-questions by deleting a morpheme which is present in declarative sentences (Köhler ). Such morphological traits of Omotic should not be underestimated. There have been roughly four different approaches to the classification of Omotic languages: they form part of Cushitic (Cerulli, Lamberti, Zaborski); a separate branch of AA with affinities to Cushitic (Moreno, Fleming, Bender); the earliest offshoot of AA (Bender, Ehret, Fleming); or a separate family outside AA (Newman, Theil). Several authors have changed their mind over time; others have remained rather skeptical about external relations of Omotic languages. In view of this diversified picture, it would be a methodological compromise to say that Omotic has an AA heritage, but not a distinctly Cushitic one. This would also fit Hayward’s opinion about the advantages of a unified Omotic language group.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
The question of inherited vs. borrowed features is crucial, especially regarding Ethiopian languages. Zaborski’s earlier skepticism, however, has fortunately become somehow outdated: the last decades have seen several ample descriptions of single varieties. Interestingly enough, Ethiopian scholars have not yet developed wider classifications; instead, they continue to provide vital linguistic data from the field. The final clue to the development of Omotic has not been found yet, so further thorough investigations must follow. Nevertheless, it surely makes much sense to consider Omotic a genetic unit with an AA background.
R Abbadie, A. T. de. (). ‘Lettres de M. d’Abbadie à M. Jules Mohl. I. Sur les langues éthiopiennes’, Journal Asiatique (): –. Abbadie, A. T. de. (). ‘To the Rev. G. C. Renouard, Foreign Secretary of the Royal Geographical Society’, The Athenæum (April , ): –. Abbadie, A. T. de (). ‘Lettre de M. d’Abbadie à M. Mohl’, Journal Asiatique (): –. Bechhaus-Gerst, M. (). ‘The classification of Cushitic and Omotic: a critique’, in G. Takács (ed.), Semito-Hamitic. Festschrift for A.B. Dolgopolsky and H. Jungraithmayr. Berlin: Dietrich Reimer, –. Bechhaus-Gerst, M., and Serzisko, F. (eds.) (). Cushitic–Omotic: Papers from the International Symposium on Cushitic and Omotic Languages, Cologne, January –, . Hamburg: Helmut Buske. Beke, C. T. (; repr. ). ‘On the languages and dialects of Abyssinia and the countries to the south’, in Proceedings of the Philological Society. Vol. . Amsterdam: Swets & Zeitlinger, –. Bender, M. L. (). ‘The languages of Ethiopia: A new lexicostatistic classification and some problems of diffusion’, Anthropological Linguistics : –. Bender, M. L. (a). Omotic: A new Afroasiatic language family. Carbondale, IL: University Museum, Southern Illinois University. Bender, M. L. (b). ‘The beginnings of ethnohistory in Western Wellegga: the Mao problem’, in R. K. Herbert (ed.), Patterns in Language, Culture, and Society: Sub-Saharan Africa. Proceedings of the Symposium on African Language, Culture, and Society, held at the Ohio State University, Columbus, April , . Columbus, OH: Department of Linguistics, Ohio State University, –. Bender, M. L. (). ‘Remnant languages of Ethiopia and Sudan’, in M. L. Bender (ed.), Nilo-Saharan Language Studies. East Lansing, MI: African Studies Center, Michigan State University, –. Bender, M. L. (). ‘Gumuz, Koman, Mao, and Omotic’, in R. G. Schuh (ed.), Précis from the th Conference on African Linguistics: UCLA, March –, (= Studies in African Linguistics, Supplement ). Los Angeles, CA: Department of Linguistics and African Studies Center, University of California, –. Bender, M. L. (). ‘First steps toward Proto-Omotic’, in D. A. Odden (ed.), Current Approaches to African Linguistics . Dordrecht and Providence, RI: Foris, –. Bender, M. L. (). ‘Proto-Omotic phonology and lexicon’, in M. Bechhaus-Gerst and F. Serzisko (eds.), –. Bender, M. L. (). ‘Gender in Omotic’, Journal of Afroasiatic Languages : –. Bender, M. L. (a). ‘A survey of Omotic grammemes’, in P. Baldi (ed.), Linguistic Change and Reconstruction Methodology. Berlin and New York: Mouton de Gruyter, –. Bender, M. L. (b). ‘The limits of Omotic’, in R. J. Hayward (ed.) (b), –. Bender, M. L. (). ‘The mystery languages of Ethiopia’, in H. G. Marcus (ed.), New Trends in Ethiopian Studies: Papers of the th International Conference of Ethiopian Studies, Michigan State University, – September . Vol. I: Humanities and Human Resources. Lawrenceville, GA: Red Sea Press, –. Bender, M. L. (). ‘The limits of Omotic revisited’, in C. Griefenow-Mewis and R. M. Voigt (eds.), Cushitic and Omotic Languages: Proceedings of the Third International Symposium, Berlin, March –, . Cologne: Rüdiger Köppe, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Bender, M. L. (). ‘Upside-down Afrasian’, Afrikanistische Arbeitspapiere : –. Bender, M. L. (). Comparative Morphology of the Omotic Languages. Munich: Lincom Europa. Bender, M. L. (). Omotic Lexicon and Phonology. Carbondale, IL: Southern Illinois University. Bender, M. L. (). ‘Topics in Omotic morphology’, in A. S. Kaye (ed.), Morphologies of Asia and Africa. Vol. . Winona Lake, IN: Eisenbrauns, –. Black, P. (). ‘Cushitic and Omotic classification’, Language Sciences : –. Blažek, V. (). ‘A lexicostatistical comparison of Omotic languages’, in J. D. Bengtson (ed.), In Hot Pursuit of Language in Prehistory: Essays in the four fields of Anthropology. In honor of Harold Crane Fleming. Amsterdam and Philadelphia, PA: John Benjamins, –. Bryan, M. A. (; repr. ). ‘A linguistic no-man’s land’, Africa : –. Bryan, M. A. (). The Distribution of the Semitic and Cushitic Languages of Africa: An outline of available information. London, New York, and Toronto: Oxford University Press. Cerulli, E. (). ‘Note su alcune popolazioni sidama dell’Abissinia meridionale’, Rivista degli Studi Orientali : –. Cerulli, E. (). Studi Etiopici. IV. La lingua caffina. Rome: Istituto per l’Oriente. Conti Rossini, C. (). ‘Studi su popolazioni dell’Etiopia (continuazione e fine, v. Vol. III, –; IV, –). V. Appunti di lingua gonga’, Rivista degli Studi Orientali : –. Conti Rossini, C. (). ‘Prefazione’, in V. L. Grottanelli (ed.), Missione etnografica nel Uollega occidentale. Vol. : I Mao. Rome: Reale Accademia d’Italia, –. Conti Rossini, C. (). ‘Pubblicazioni Etiopistiche dal al ’, Rassegna di Studi Etiopici : –. Ehret, C. (). ‘Omotic and the subgrouping of the Afroasiatic language family’, in R. L. Hess (ed.), Proceedings of the Fifth International Conference on Ethiopian Studies. Session B. April –, . Chicago, USA. Chicago, IL: Office of Publications Services, University of Illinois at Chicago Circle, –. Ferguson, C. A. (). ‘The Ethiopian language area’, Journal of Ethiopian Studies (): –. Fleming, H. C. (). ‘The classification of West Cushitic within Hamito-Semitic’, in D. F. McCall, N. R. Bennett, and J. Butler (eds.), Eastern African History. New York, Washington, DC, and London: Frederick A. Praeger, –. Fleming, H. C. (). ‘Omotic as an Afroasiatic family’, in W. R. Leben (ed.), Papers from the Fifth Annual Conference on African Linguistics (= Studies in African Linguistics, Supplement ). Los Angeles, CA: Department of Linguistics and African Studies Center, University of California, –. Fleming, H. C. (). ‘Recent research in Omotic-speaking areas’, in H. G. Marcus (ed.), Proceedings of the First United States Conference on Ethiopian Studies, Michigan State University, – May . East Lansing, MI: African Studies Center, Michigan State University, –. Fleming, H. C. (a). ‘Omotic overview’, in M. L. Bender (ed.), The Non-Semitic Languages of Ethiopia. East Lansing, MI, and Carbondale, IL: African Studies Center, Michigan State University, and Southern Illinois University, –. Fleming, H. C. (b). ‘Non-Semitic languages: Cushitic and Omotic’, in M. L. Bender, J. D. Bowen, R. L. Cooper, and C. A. Ferguson (eds.), Language in Ethiopia. London: Oxford University Press, –. Fleming, H. C. (). ‘Omotic and Cushitic: a reply to Lamberti’, Anthropos : –. Fleming, H. C. (). ‘Second reply to Lamberti’, Anthropos : –. Fleming, H. C. (). Ongota: A decisive language in African prehistory. Wiesbaden: Otto Harrassowitz. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Hayward, R. J. (a). ‘Introduction’, in R. J. Hayward (ed.) (b), vii–xix. Hayward, R. J. (ed.) (b). Omotic Language Studies. London: SOAS. Hayward, R. J. (). ‘Afroasiatic’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –. Hayward, R. J. (). ‘Omotic: the “Empty Quarter” of Afroasiatic linguistics’, in J. Lecarme (ed.), Research in Afroasiatic Grammar II: Selected papers from the Fifth Conference on Afroasiatic Languages, Paris, . Amsterdam and Philadelphia, PA: John Benjamins, –. Hetzron, R. (). ‘The position of Omotic’, in M. Bechhaus-Gerst and F. Serzisko (eds.), –. Hudson, G. M. (). ‘Languages of Ethiopia and languages of the Ethiopian Census’, Aethiopica: International Journal of Ethiopian and Eritrean Studies : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
̈
Köhler, B. (). Towards the typology of interrogative verbal morphologies in Omotic. Paper read at the conference “Omotic Utterance Type, Mood and Attitude Markers and Linguistic Typology”, Leiden, Netherlands, October , . Köhler, B. (). ‘Interrogative zero-marking in some Ometo languages’, in M.-C. Simeone-Senelle and M. Vanhove (eds.), Proceedings of the th International Conference on Cushitic and Omotic Languages, Paris, – April . Cologne: Rüdiger Köppe, –. Lamberti, M. (). Kuliak and Cushitic: A comparative study. Heidelberg: Carl Winter. Lamberti, M. (). ‘Cushitic and its classifications’, Anthropos : –. Lamberti, M. (). ‘Omotic and Cushitic: a reply to Fleming’, Anthropos : –. Militarev, A. Y. (). ‘Once more about glottochronology and the comparative method: the OmoticAfrasian case’, in A. V. Dybo, V. A. Dybo, O. A. Mudrak, and G. S. Starostin (eds.), Aspects of Comparative Linguistics . Moscow: Russian State University for the Humanities, –. Moreno, M. M. (). Manuale di Sidamo. Grammatica—esercizi: Testi—glossario. Milan: A. Mondadori. Mous, M. (). ‘Selectors in Cushitic’, in F. K. E. Voeltz (ed.), Studies in African Linguistic Typology. Amsterdam and Philadelphia, PA: John Benjamins, –. Praetorius, F. (; repr. ). ‘Über die hamitischen Sprachen Ostafrika’s’, Beiträge zur Assyriologie und vergleichenden semitischen Sprachwissenschaft : –. Reinisch, S. L. (). ‘Das Zalwort vier und neun in den chamitisch-semitischen Sprachen’, in Sitzungsberichte der philosophisch-historischen Classe der Kaiserlichen Akademie der Wissenschaften. Vol. . Vienna: F. Tempsky, –. Reinisch, S. L. (). Das persönliche Fürwort und die Verbalflexion in den chamito-semitischen Sprachen. Vienna: Alfred Hölder. Ruhlen, M. (). A Guide to the World’s Languages. Vol. : Classification. Stanford, CA: Stanford University Press. Theil, R. (). Is Omotic Afroasiatic? A critical discussion. MS, http://www.uio.no/studier/emner/ hf/iln/LING/v/THEIL%Is%Omotic%Afroasiatic.pdf accessed July , . Tosco, M. (). ‘Cushitic and Omotic overview’, in M. L. Bender, G. Takács, and D. L. Appleyard (eds.), Selected Comparative-Historical Afrasian Linguistic Studies: In memory of Igor M. Diakonoff. Munich: Lincom Europa, –. Tucker, A. N. (). ‘Fringe Cushitic: An experiment in typological comparison’, Bulletin of the School of Oriental and African Studies : –. Tucker, A. N., and Bryan, M. A. (). The Non-Bantu Languages of North-Eastern Africa. London, New York, and Cape Town: Oxford University Press. Tucker, A. N., and Bryan, M. A. (). Linguistic Analyses: The Non-Bantu languages of NorthEastern Africa. London, New York, and Cape Town: Oxford University Press. Zaborski, A. (). Studies in Hamito-Semitic I: The verb in Cushitic. Cracow: Nakładem Uniwersytetu Jagiellońskiego. Zaborski, A. (). ‘Basic numerals in the Omotic languages’, in S. Segert and A. J. E. Bodrogligeti (eds.), Ethiopian Studies, dedicated to Wolf Leslau on the occasion of his seventy-fifth birthday. Wiesbaden: Otto Harrassowitz, –. Zaborski, A. (). ‘Remarks on the genetic classification and the relative chronology of the Cushitic languages’, in J. Bynon (ed.), Current Progress in Afro-Asiatic Linguistics: Papers of the Third International Hamito-Semitic Congress. Amsterdam and Philadelphia, PA: John Benjamins, –. Zaborski, A. (). ‘Can Omotic be reclassified as West Cushitic?’, in G. Goldenberg (ed.), Ethiopian Studies: Proceedings of the Sixth International Conference, Tel-Aviv, – April . Rotterdam and Boston, MA: A. A. Balkema, –. Zaborski, A. (). ‘Preliminary remarks on case morphemes in Omotic’, in R. J. Hayward (ed.) (b), –. Zaborski, A. (). ‘The position of Cushitic and Berber within Hamitosemitic dialects’, in A. Bausi and M. Tosco (eds.), Afroasiatica Neapolitana: Papers from the th Italian Meeting of Afroasiatic (HamitoSemitic) Linguistics, Naples, January –, . Naples: Istituto Universitario Orientale, –. Zaborski, A. (). ‘West Cushitic—a genetic reality’, Lingua Posnaniensis : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T or so languages that make up the Chadic family are spoken over an area fanning south of Lake Chad, in the Niger Republic, Nigeria, Chad, and Cameroon. Their sociolinguistic situation varies greatly. At one end, Hausa, one of the national languages of Nigeria and the Niger Republic and spoken by twenty-five million people as a first and twenty million more as a second language, is dominating and threatening many of its Chadic cousins and many Niger-Congo minority languages spoken farther south. After absorbing Chadic languages that were spoken in its present territory, Hausa has become a lingua franca over a large part of West Africa, where it is only matched by Nigerian Pidgin and its seventy million speakers. The other Chadic languages count between , and a couple of thousand speakers. Some have recently become extinct, while many are spoken fluently only by the older generation, the younger speakers being more fluent in Hausa, Fulfulde, or, for the modern elite, the colonial languages and their creolized version. This follows the pattern prevailing in West Africa whereby multilingualism is the rule and monolingualism an exception. The aim in this chapter is to characterize the identity of the Chadic family. Given their common genetic origin, a fact that is now taken for granted by the specialists of language classification, what does it mean to be a Chadic language, not only in terms of retentions and innovations from their common Afro-Asiatic origin, but also from a typological point of view?
. A C
..................................................................................................................................
.. List and classification of languages This chapter uses Newman’s (a) internal classification of Chadic and its system of reference (see Table .).1 In this classification, Chadic languages are divided into four 1 This new version of Newman’s classification includes recent developments in Chadic linguistics, e.g. the new “Barawa” cluster .C. sub-branch (Jaggar ; Caron , , , , ; Haruna ) and a new internal classification of East Chadic B (Lovestrand ). Newman is uncertain concerning the status of the II.B unit: “Whether this II.B unit (especially the Kotoko group) should be treated as a sub-branch within B-M rather than as a separate, independent branch of Chadic remains to be determined” (Newman a: ).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . The Chadic language family: classification (Newman a: –) Branches
SubGroups branches A.
Subgroups Languages
. Hausa
Hausa, Gwandara
. Bole-Tangale a
b
Bole, B’ele, Daza, Deno, Galambu, Gera, Geruma, Karekare, Kirfi, Maka, Ngamo, Bure† Tangale, Kanakuru, Kupto, Kushi, Kwami, Nyam, Pero, Piya, Widala Angas, Chip, Jorto, Kofyar, Mushere, Sura/ Mupun Goemai, Koenoem, Montol, Pyapun, Tal
c
Gerka
a
Ron, Kulere, Mundat, Sha
b
Fyer, Tambas
b
I. West Chadic (W-C)
. Angas
. Ron
B.
a
Bade, Duwai, Ngizim, Auyo†, Teshena†
. Bade . Warji
a b
C.
A.
Bubbure, Guruntum, Jimi, Ju
. Tera
a
Tera, Jara
b
Ga’anda, Hona
II. Biu-Mandara (B-M) (=Central Chadic)
Boghom, Kir, Mangas
. Bura
Bura, Chibak, Kilba, Margi
. Higi
Higi, Bana, Hya, Kapsiki
. Matakam
C.
Dass, Geji, Polchi, Saya, Zari, Zeem, Luri†
. South-Bauchi a (= “Barawa b cluster”) c
. Mandara
B.
Warji, Diri, Jimbin, Kariya, Mburku, Miya, Siri, Tsagu Pa’a
a b
Mandara/Wandala, Dghwede, Glavda, Guduf, Gvoko, Malgwa, Podoko Cineni, Hdi, Lamang, Vemgo
a
Mafa/Matakam, Chuvok
b c
Dugwor, Giziga, Mada, Mefele, Merey, Mofu, Moloko, Muyang, Ouldémé, Zulgo Muktele
d
Mbuko, Vame
. Sukur
Sukur
. Daba
Daba, Buwal, Mazagway, Mbedam, Mina
. Bata a
Bata, Bachama, Gude, Gudu, Jimi, Ngwaba, Nzanyi, Sharwa, Tsuvan, Zizilivakan Kotoko, Jilbe, Logone, Majera, Zina
b
Buduma
. Kotoko . Musgu
Musgu, Mbara, Muskum†
. Gidar
Gidar
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Branches
SubGroups branches A.
Subgroups Languages
. Somrai
Somrai/Sibine, Buso, Gadang, Miltu, Mire, Ndam, Sarwa, Tumak Lele, Gabri, Kabalai, Kimre, Nancere, Tobanga Kera, Kwang
. Kera . Dangla-Mubi a b
Dangla/Dangaléat, Bidiya, Birgit, Bourmataguil, Migama, Mogum, Toram Mubi, Kajakse, Masmaje, Zirenkel
c
Kujarge
. Mukulu/ Mokilko
Mukulu/Mokilko
. Sokoro
Sokoro, Mawa, Saba, Tamki, Ubi
. Barain
Barain
. Masa
Masa, Gizey, Marba, Mesme, Musey, Zime, Zumaya†
IV. Masa (M-S)
III. East Chadic (E-C)
. Lele
B.
†
= extinct
coordinate branches indicated by Roman numerals (I to IV). Sub-branches are indicated by capital letters (A, B, C). These are divided into groups indicated by Arabic numerals (, , etc.) When needed, subgroups are identified by lower case letters (a, b, c, etc.). Kapsiki, for example, is classified as II.A., that is, it belongs to the A sub-branch of Biu-Mandara (II), and within this sub-branch, to the Higi group ().
.. Geographic situation In their present location (see Map .), Chadic languages are in contact with or overlap languages from other phylogenetic families, such as Berber (Tuareg) in the northwest (Niger Republic); Semitic (Shuwa Arabic) in the south of the Chad Republic; Nilo-Saharan languages, ranging from Songhay, Kanuri, Kanembu (from the northwest to the northeast) to Central Sudanic languages such as Daju, Bagirmi, Sara (southwest Chad); and finally Niger-Congo, including Fulfulde (northern Nigeria, Adamawa, and northern Cameroon) and numerous Benue-Congo languages in the southwest (Nigeria, Cameroon).
CHAD Lake Chad II.B.1
Gashua I.B.1
HAUSA
Kano
FUL
B-C
I.B.1 Potiskum
FUL
Maiduguri
Damaturu
II.B.1
I.C
II.A.3
II.A.1 II.A.1
I.A.1
II.A.8 Numan Yola
III.A.3
III.A.1
Yelwa
A-U
Soro III.A.2
IV.A.1
N-S
FUL
FUL
Ch ar i
III.A.3
FUL
Garoua A-U
nu e
Pankshin I.A.3
I.A.1
II.C.1
N-S
IV.A.1 FUL
Be
I.A.4
Mubi II.A.8
B-C
II.A.5 Maroua II.A.7 II.A.5
AR
Bara
one
II.A.2
Tafawa Balewa ABUJA
FUL
III.B.1
B3
II.A.6
Biu Gombe
Bauchi B-C
III.B.1
Logone Birni
Log
I.A.2
Kaduna
KWA
III.B.1
AR Gwoza II.A.4 II.A.4
I.B.2 B-C
AR
III
Zaria
I.A.1
III.B.1 AR
Misau
HAUSA
Karay
N-S
N-S N’DJAMENA
NIGERIA
AR
FUL
A-U
Sarh
C.A.R.
CAMEROON B-C
KWA
I
CHADIC LANGUAGES: NON-CHADIC LANGUAGES:
A-U
West Chadic Adamawa-Ubangi
II
B-C
Biu-Mandara
Benue-Congo
. The geographical distribution of Chadic languages
III
FUL
Fulfude
East-Chadic N-S
Nilo-Saharan
Masa
IV
AR
Arabic
D. Bonardelle - Llacan CNRS - 2015
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
N-S
NIGER
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Number of speakers The number of speakers of Chadic languages is difficult to assess. One reason is the lack of clear-cut distinction between languages and dialects. A second reason is due to a tendency of SIL, which is our main source of reference, to inflate the numbers of both languages2 and speakers. Finally, in Nigeria, where locally dominant ethnic entities are granted political and economic advantages by the “federal character” of the constitution, census figures tend to be tampered with and are not to be trusted. Bearing these facts in mind, all the figures quoted below from Ethnologue (Lewis et al. ) can only be tentative and correspond to a rough estimate. If we compare these figures with those given by Barreteau and Newman (: ) almost forty years ago, we see that the numbers have been roughly multiplied by two, and the hierarchy of the languages has been preserved: • Hausa (see Map .), with over forty-five million speakers (twenty-five million L and twenty million L speakers), weighs more than all the other languages put together. • Twelve languages count , speakers or more: Bura (,), Ngas (,), Zaar (,), Higi (,), Mwaghavul (,), Musey (,), Bade (,), Masa (,), Mafa-Mefele (,), Tangale (,), Gera (,), and Goemai (,). • Nineteen languages count between , and , speakers: Cibak (,), Ron-Bokkos (,), Margi (,), Margi-South (,), Musgu (,), Bata (,), Karekare (,), Tera (,), Bole (,), Mofu-Gudur (,), Nancere (,), Giziga (,), Ngizim (,), Warji (,), Gidar (,), Dangla (,), Guduf (,), Buduma (,), and Boghom (,). • Sixteen languages count between , and , speakers: Wandala (,), Lamang (,), Mubi (,), Gabri (,), Hona (,), Glavda (,), Kilba (,), Dghwede (,), Mofu-Duvang (,), Gwandara (,), Galambu (,), Pero (,), Daba (,), Migama (,), Gvoko (,), and Kanakuru (,). • Many languages count fewer than , speakers, while some are now extinct (e.g. Bure, Anyo, Teshena, Luri, Zeem, Lushi, Tulai, Muskum, and Zumaya).
As a summary, while many languages are in serious danger of disappearing under the pressure of Hausa, the majority are faring quite well and benefiting from the rapid demographic growth in the north of West Africa. Unfortunately, due to the poverty prevailing in the majority of the population this growth is not accompanied by language development.
The number of Chadic languages is given as in Lewis et al. (), in Newman (a), and in Schuh (). 2
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
NIGER
MALI CHAD er Nig
Lake Chad BURKINA FASO
NIGERIA Niger
BENIN
GHANA
TOGO
i ar Ch
CÔTE D’IVOIRE
CAMEROON
Gulf of Guinea 0 0
200 200
400 mi
400 km
First language
Second language
Second language within the Muslim community
. Hausa extension
. T C
..................................................................................................................................
.. Phonology Chadic languages have a complex consonantal system (see Table .) balanced by far fewer vowels. In addition to the consonants listed in this table, many languages have prenasalized obstruents (mb, nd, ŋg) and extra series of palatalized (k y, g y, x y) and
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Chadic consonants (Newman : ) Labial Stops
Affricates
Lateral continuants
p
t
č
k
b
d
j
g
glottalized
ɓ
ɗ
ɗy
ɠ/ƙ
š
x
voiceless
ts dz
voiceless
f
voiced
v
voiceless
s z
ɣ
ɬ ɮ m
Liquids Glides
Velar
voiceless
voiced Nasals
Palatal
voiced
voiced Continuants
Alveolar
n
(ɲ)
ŋ
r w
y
labialized (kw, gw, xw) velars.3 The lateral fricatives (ɬ, ɮ) are present in B-M and a few W-C languages, but absent from E-C.4 Neither the glottal stop /ʔ/ nor /h/ have been reconstructed for Proto-Chadic (P-C) (Newman ). They are either a non-distinctive feature of word onset, or a secondary development from /’y/, /’w/ or /x/, /ɣ/ respectively. Finally, the presence of labial-velars and labial flap is a contact-induced phenomenon. The number of vowels postulated for any Chadic language largely depends on the analytical tools used for the description. A language can be analyzed as having as few as two vowels (a and ə, with y and w prosodies), while others count as many as nine vowels, plus vowel length. [ATR] vowel harmony, a typical Sudanic feature, is only present in East Dangla and Tangale, where it has been proved to be the result of historical contacts with Waja, an Adamawa language (Kleinewillinghöfer ). Schuh () postulates for Chadic a common system of three vowels doubled by length + central ə = [ɨ], (i (ī), u (ū), a (ā), ə), the length of ī and ū resulting from *V + homorganic glide. Noting how close the system is to Semitic and Berber, he comments on the fact that it has resisted the influence of Niger-Congo languages. Tone is present in all Chadic languages, and operates different systems: a high (H) and low (L) two-level tone system, with downdrift, as in Hausa; a three-tone level, with a Mid
3 4
They were reconstructed for P-C in Newman (). In his table of “Common Chadic Consonants”, Newman (: ) writes l (= [ɮ] ?) and comments: Although /ɮ/ is phonetically the voiced counterpart of /ɬ/, and in some languages functions as its pair, its historical connection is probably with the liquid /l/. That is, from a historical perspective, /ɮ/ should be thought of as a liquid, albeit with fricative articulation, as contrasted with /ɬ/, which should be thought of as a fricative albeit with lateral articulation.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
(M) tone added, as in Zaar; a “terrace level” system, with downstepped H (!H), as in Miya: átsə́tsalìy [ ̅ – – _ ] H!HHL ‘cowry’. Although Chadic languages vary with respect to their phonotactic rules, they do not allow consonant clusters in the syllabic coda, and in most languages the set of consonants allowed in the coda is smaller than the set allowed in the onset. Syllable weight (where light syllables have a short vowel nucleus, and heavy syllables have a long vowel nucleus or are closed by a consonant) is an essential factor in the morphophonology of Chadic languages (Newman ). Frajzyngier (a) explores rules of vowel deletion to mark phraseinternal position, vowel retention, and the addition of segments to mark phrase-final position.
.. Morphology Chadic languages possess the following lexical categories: verbs, nouns, adjectives, adverbs, adpositions, and ideophones. Nominal system. Gender and plurality are well represented in Chadic. Gender opposes masculine to feminine in the singular only. When these categories are represented in a language, adjectives agree for number and gender with the noun they qualify. Chadic languages that lack gender (e.g. Pero, Zaar, and Mafa) have lost it over their evolution. Some languages (e.g. Hausa) have a very rich system of means to form noun and adjective plurals, including numerous suffixes, internal vowel ablaut, germination, and reduplication,5 while in others plurality is not marked on the noun but can be marked on the noun phrase. For example, in Geji (I.C..a) the noun is invariable but plurality is expressed in the suffixed deictics (lelpi-gḛ̀ː/lelpi-ʤḛ̀ː ‘this antelope/these antelopes’), or in the genitive link used to modify a noun (e.g. puka gə/̀ ʤì Áwdù; lit. ‘knife of (/) Audu’). All three branches have an associative plural, e.g. Zaar (I.C..a) gjáː, in gjáː Lokʃíl ‘Lokshil and his friends’. The existence of independent pronouns, different from the other pronominal paradigms, is a property that Chadic languages have kept from Proto-Afro-Asiatic (P-AA). These pronouns appear in the function of topic, focus, or preposition complement. Other paradigms are specialized for different syntactic functions, e.g. subject, and direct and indirect object. In some languages, contrary to Hausa, which has a distinct indirect object paradigm, the direct and indirect paradigms are identical, with the indirect object function expressed elsewhere, e.g. in Zaar by the sentence-final benefactive particle mən ́ taking the second object as an argument. The reflexive and reciprocal pronoun paradigms are often derived from the name of body parts, e.g. Hausa kâina ‘myself ’ (lit. ‘head of me’). A typical Chadic paradigm contains eight pronouns indicating person, number, and gender, with gender opposition in the second and third person singular, and an inclusive/exclusive opposition in the first person plural. Logophoric pronouns are present in E-C, but marginal elsewhere. Verbs. Chadic verbs are organized into classes whose morphosyntactic behavior is dependent on their phonological structure, involving tone, the number of consonants, and the status and number of vowels. Chadic root structure, and particularly that of verbs, is typically biconsonantal (Hodge ; Schuh ). In words with third consonants, these In Newman’s (a) encyclopedic grammar of Hausa, chapter , entirely devoted to nominal plurals, is thirty-five pages long! 5
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
have been shown to be frozen former grammatical formatives (Jungraithmayr , ). Newman () reconstructed two lexical classes of verbs with two consonants and final ə/a vowels: CVCə/a, doubled by a H/L tone opposition. Schuh (: ) links the tones and final vowels of the reconstructed verb classes to their syllable weight. Verb conjugation involves three coding means: a preverbal complex called “subject pronoun”, expressing person, tense, aspect, and mood; verb stem alternation, including change of vowel and tone pattern; suffixation, especially an -n suffix marking plural agreement (e.g. Karekare nà ŋgàtáa kò ‘I fell down’, mù ŋgàt-án kò ‘we fell down’ (Newman and Schuh )).6 There is much variation in the categories involved in the TAM systems of Chadic languages, from a minimal three-term system (e.g. Dott (Caron ) whose basic TAMs are the aorist, the continuous, and the future) to very complex systems (e.g. Zaar for which Caron () describes more than thirty TAM paradigms, with six aspects, four tenses, three moods, and their combinations). Two elements are pervasive in the Chadic verbal system: (i) the formation of a continuous based on a locative construction (‘be at’ + verbal noun); (ii) dependent verb forms in focus, post-nominal relatives, and sequential constructions, as, e.g., in Hausa, Kanakuru, and Tera (Newman and Schuh : ). Verbal derivation by affixation includes the formation of verbal nouns, and the suffixation of verbal extensions to mark spatial relationships (e.g. ventive and centrifugal (Frajzyngier )), and semantic roles on the verb (e.g. causative, middle, goal, and point of view (Frajzyngier ; Frajzyngier and Munkaila )). Finally, verb plurality is expressed all over Chadic languages in what has been called “pluractionals” since (Newman : ), denoting either plurality of participants (subject of intransitive verb, object of transitive verb) or plurality of action. Newman (: –) mentions its presence in thirty languages, scattered over eighteen out of the twenty-six sub-branches of the family, from Hausa in the west to Migama in the east. Pluractionals with -a- ablaut have been described, e.g., in Zaar (.C..a): ʤom/ʤwáːm ‘pile up, gather’ (with a plural object); nʤol/nʤwáːl ‘leave, go’ (with a plural subject); ɮop/ɮwáːp ‘sting’; vjer/vjáːr ‘insult’ (Caron : ). The category has consistently been described as not a grammatically obligatory agreement feature (Frajzyngier ).
.. Sentence structure and function marking Although nowadays the vast majority of Chadic languages have SVO as basic main clause order, Frajzyngier () and Schuh () independently argue for an initial VSO word order in P-C. Most languages have abandoned post-verbal placement of subjects in any function, leaving just a few W-C languages that postpose subjects for questioning and focus and a few B-M languages that retain basic VSO sentence structure. A large majority of SVO and some VSO languages have in situ ordering for question words. Some SVO languages that use in situ ordering for non-subjects have the unusual feature of placing questioned subjects in post-verbal position. Non-verbal predication is used in Chadic to express locative stative, equational, and existential predications. The constituent order in nonverbal predication agrees with that of verbal predication, e.g. SVO and subject-predicate in Hausa, VSO and predicate-subject in Hdi (Frajzyngier and Shay ). 6
Cf. Newman’s () analysis concerning the diachronic importance of this plural agreement pattern.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Syntactic functions are not expressed by case marking in Chadic, but they involve word order, prepositions, or inflectional coding on the verb. Nominal indirect objects are expressed through prepositional phrases, but the pronominal indirect objects are realized as verbal clitics preceding direct objects, whether nominal or pronominal. Serial verb constructions, which are present in some W-C languages, are the result of language contact, as they are similar in their forms and functions to analogical structures in Niger-Congo languages.
. C A-A
.................................................................................................................................. Johannes Lukas, in the s, initiated the study of the Chadic family as such when he presented its first classification (Lukas ) and published a good many descriptions of its members. The modern classification as a member of the Afro-Asiatic phylum was firmly established by Greenberg (, ) on the basis of mass lexical comparison and shared morphology. Following Greenberg, Newman and Ma () published the first authoritative comparative study and classification of Chadic languages, which was gradually amended in subsequent revisions (Newman , , , a). This section lists the historical arguments establishing the status of Chadic as a family inside the Afro-Asiatic phylum, and examines how they have fared in the light of new developments in Chadic linguistics. Linguists generally agree on a few strong lexical correspondences pointing at the genetic unity of Chadic and its affiliation with the rest of Afro-Asiatic, e.g. Schuh (: ): *m-t‘die’ (Bole [W] motu, Musey [M] mít—cf. Egyptian mt, Arabic māt); *s-n- ‘know’ (Miya [W] sən, Mokilko [E] sū̀ ne—cf. Egyptian swn, Tuareg əssən); *b-n- ‘build; house’ (Bole [W] ̂ ‘grass hut’—cf. Arabic banā); and *fəɗu ‘four’ (cf. Egyptian bònò ‘house’, Mokilko [E] bī no fdw) and *ɬəw- ‘meat’ (cf. Tuareg isan). Apart from these, the available reconstructions of P-C lexicon (Newman ; Jungraithmayr and Ibriszimow ) suffer from the small number of reconstructed items and some methodological flaws. Chadic needs a more extensive reconstruction of the proto-language, from the bottom up, starting with a detailed reconstruction of the four branches of the family (Newman : ). Out of the seven morphological arguments listed by Hayward () for the relatedness of Afro-Asiatic languages, only two are verified in Chadic: (a) the paradigm of personal pronouns, and (b) a feminine gender marking element -(a)t. Newman () had a more convincing list: (a) a feminine formative t with the triple function “female/diminutive/ singulative”; (b) an n/t/n gender-number-marking pattern in the deictic system (Greenberg ); (c) an m- prefix forming nouns of place, instrument, and agent; (d) the formation of noun plurals, among other ways, by the suffixation of -n and by the insertion of an internal vowel -a- (Greenberg ); (e) the formation of intensive (“pluractional”) verbs by internal consonant gemination, also often accompanied by the use of an internal -a-; and (f) an asymmetrical conjugational system involving suffixed feminine and plural markers in addition to pronominal prefixes. If the correspondence between P-C common pronouns as reconstructed by Newman (: ) and P-AA non-subject pronouns (i.e. either object or possessive pronouns) proposed by Hayward (: ) is still as impressive (see Table .), the recent developments in the knowledge and analysis of Chadic languages have modified our general vision of the relationship between Chadic and Afro-Asiatic.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Proto-Chadic and Afro-Asiatic pronouns Pronouns
P-C
P-AA
1
ní
i ~ yi
2m
ka
ku, ka
2f
ki(m)
kim
3m
nì
si, isi
3f
ta
1pl
mun (incl.), na (excl.)
2pl
kun
kuuna
3pl
sun
su ~ usu
(*-na ~ *-nu ~ *-ni)1 ?
Hayward () does not give the form for P-AA ‘us, our’ but gives the following reconstructions for the six branches: P-E *-ina; P-S *-na ~ *-nu ~ *-ni; Berber -nɣ; P-C *-na ~ *-nu ~ *-ni, *ʔina ~ etc.; Cushitic na; Omotic in. 1
Newman () hesitates to consider various formatives of nominal plurals found in Chadic as retentions both for the internal ablaut to -a- (“plurals that look like internal formations . . . probably derive historically from external suffixes” (: )) and for the medial consonant gemination (“In most cases, gemination is not the sole marker of noun plurality, but rather accompanies other overt markers, i.e. it contributes to the ‘strengthening’ of a stem without being the marker of plurality per se” (: ).). Newman still reconstructs several suffixes as plural formatives for P-C, e.g. *-n, *-ai, *-i, *-aki, and possibly *-ɗi (: ). Among those, Newman suggests that only *-n and *-ai are Afro-Asiatic retentions (: ). Contrary to what is still widely repeated, Chadic languages have no verbal affixes expressing reflexivity, reciprocity, or the passive, and no -s causative. The Hausa verbal extensions, traditionally called respectively “causative” (also known as grade ) and “passive” (also known as grade ) have been reanalyzed as marking “affectedness” for the -u suffix of grade (Frajzyngier a: ) and “efferentiality” for the -as suffix of grade (Newman b: ). The Hausa “true” causative is expressed analytically through a sequence of full clauses. It has long been admitted that the preverbal pronominal elements used in the conjugation of Chadic languages (“subject pronouns”) are different from the prefix-conjugation, and represent an innovation. More damaging is the reanalysis of the examples of -a- ablaut and consonant gemination mentioned by Hayward (), quoting Jungraithmayr () as reflexes of the P-AA “present stem” associated to prefixed conjugation (cf. Mubi, Ron, and Migama imperfect/habitative formations in Table .). These have since been reanalyzed as “innovative redesignations of either plural verb stems, or of the proto-VN stem”. If the P-AA “present stem” has survived in Chadic, it is outside the aspect system, in pluractional verbs (cf. section .. above). According to Wolff (), the Chadic verbal system is not organized around a perfect/imperfect opposition, but syntactically on the basis of the feature [+/ nominal] and semantically on the basis of the feature [+/ plural].
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Internal -a- imperfect/habitative in Chadic languages (Jungraithmayr ) Mubi ‘bite’
níʔūwát (‘imperfect’)
níʔēwít (perfect)
Ron-Daffo ‘die’
ʔí mwáat (‘habitative’)
ʔí mot (‘Grundaspekt’)
Migama ‘wash’
náʔàpàllá (‘imperfect’)
náʔápìlé (‘perfect’)
Newman (: ) describes a pattern of number agreement, hence different from pluractionals, as an element of the P-AA prefixed conjugation preserved in Chadic verb morphology: “Verbs in Proto-AA were conjugated for person and number by means of pronominal-like prefixes. In the nd and rd person plural, the verbs required a suffix as well. This suffix had the form *-(Vn).” Good examples are given in the West and BiuMandara branches, for Kirfi (I.A..a): mù làɓɓinshù ‘we stepped on them’ vs. nà làɓɓishu ‘I stepped on them’), and Giziga (II.A..b) ‘killed’: ʔi kaɗ/kəɗ-am ‘/’, kə kaɗ/kəɗ-am ‘/’, ʔa kaɗ/kəɗ-am ‘/’ (Newman : ). Pelu, a dialect of Geji (I.C..a), exhibits a similar situation where, in the subjunctive, the functional load of marking the plurality of subject is shifted from the second person pronoun (which does not mark number) to the verb, which displays an -ən suffix in the plural: sə̀ laː tə̀k poːl/póːl-ən hunká (: want . sweep/sweep- room), ‘they want you to sweep (/) the room (Caron : ). Although Diakonoff () postulates that P-AA must have been a tonal language, on the basis of the fact that distinctive tone is present in two distant branches of Afro-Asiatic, i.e. Chadic and Cushitic, it is probably an innovation: Wolff () convincingly argues that tonogenesis in Chadic has been triggered by two phonological processes linked to the interaction between consonant and tone: the first is phrase-internal vowel deletion, the second is neutralization of syllable-final stops and continuants to corresponding sonorants. Deleted voiceless consonants would have created high tones, and voiced consonants low tones. Schuh (: ) thinks tonogenesis is due to early contact with non-Afro-Asiatic tone languages. A conjunction of both origins is equally feasible.
. T C
.................................................................................................................................. This section explores the possibility that Chadic languages share a certain number of typological characteristics, and whether this comes from their Afro-Asiatic heritage or from contact and interaction with neighboring languages. These sets of typological features, when they do not coincide with genetic boundaries, may define a geographical area that straddles those boundaries (Wolff ), and correspond to the African continent or to smaller units.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Africa as a linguistic area The idea of the existence of Africa as linguistic area has been on the agenda of African linguists for many years since Greenberg (, ) first launched it. Heine and Leyew (), working on Greenberg’s initial intuitions, give a first list of properties “essentially restricted to Africa”, e.g. clicks, labial flaps, and case inflections expressed exclusively by tone, and a second list of properties that are “distinctly more common in Africa than elsewhere”, e.g. labial-velar stops, implosives (= non-obstruent stops), noun class systems, and SOVX word order. Chadic languages fare quite well in this second set, as some of them possess six out of nine of those properties which would identify them as “typical African languages”. However, Heine and Leyew () state that this second set of properties is either genetically or areally restricted in their occurrence, and more precisely, to an area Güldemann () named the Macro-Sudan belt. As a conclusion, it would be circular to list the properties of Chadic languages pointing at their typical African character if Chadic and other languages of the area are at the origin of the very properties used as a standard to characterize African languages in general. The exploration of smaller areal units has proved more relevant.
.. The Macro-Sudan belt Clements and Rialland () have established the existence of a “Sudanic phonological area” on a sound methodology based on the systematic comparison of phonological features of languages spoken on the African continent, regardless of their genetic affiliation, to non-African languages. They concluded that “there is no characteristically African phonological property that is common to the continent as a whole” (p. ) and defined six smaller, more specific regions (see Map .). Among Afro-Asiatic languages, Chadic languages belong to the “Sudanic” zone, while Arabic and Berber languages define the north zone, and the Ethiopian Semitic, the Cushitic, and the Omotic belong to the east zone. Together with Chadic, the Sudan zone contains all non-Bantu (and some Bantu) languages of the Niger-Congo phylum, southern varieties of Arabic, and most NiloSaharan languages. Table . summarizes the features shared by Chadic languages and these three zones (north, east, Sudan). What defines the phonological characteristics of the Sudanic zone is “labial flaps, labial-velar stops, implosive, nasal vowels, vowel systems [ATR vowel harmony], multiple tone levels and ‘lax’ question prosodies” (Clements and Rialland : ). The study of the Sudanic area is completed by Güldemann (), who includes morphological properties and interprets the data in terms of their phylogenetic consequences for the finer genetic families (e.g. Berber, Chadic, Gur, and Nilotic) rather than Greenberg’s superordinate units (Afro-Asiatic, Niger-Congo, etc.). His analysis leads him to define the Macro-Sudan belt as consisting of three concentric circles, where the sharing of features is particularly prominent in the innermost one, the “hotbed” [Benue-Congo, Adamawa-Ubangi, Moru-Mangbetu, BongoBagirmi], is still strong in the intermediate one, the “core” [the previous four + Atlantic, Mande, Kru, Gur, Kwa], but peters out in the outermost one, the “macro-zone” as a whole. (Güldemann : )
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
NORTH
SUDANIC EAST
T RIF
CENTER
SOUTH
0
500
1000 km
. Chadic (minus Hausa) and phonological zones in Africa (based on Clements and Rialland )
Certain characteristic Sudanic features, absent in the north zone, are represented in Chadic without being ubiquitous. These are labial flaps, [ATR] vowel harmony, logophoricity, and final negation, which are present in languages spoken in east Nigeria, north Cameroon, and southeast Chad. These properties cannot be reconstructed for the protolanguage, and they are better accounted for by contact-induced interference with the two Niger-Congo families belonging to the “hotbed” of the Macro-Sudan belt, i.e. Benue-Congo and Adamawa-Ubangi. Chadic, like Nilotic and Narrow Bantu, thus belongs to the periphery of the Macro-Sudan belt, having retained features inherited from the AfroAsiatic stock, but acquired many through contact with the Benue-Congo and AdamawaUbangi families. As Güldemann (: ) puts it:
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Distribution of linguistic features Sudanic
North
East
Chadic
absence of P-sounds
x
xxx
xx
–
emphatic consonants
–
xx
–
–
non-tonal prosody
–
xxx
x
–
labial flaps
x
–
–
x
labial-velar stops
xx
–
-
x
implosives
xx
–
(xx)
xxx
xx
–
-
–
xxx
–
(x)
–
3+ tone levels
xx
–
(x)
xx
ejective stops
–
–
xx
–
nasal vowels two series ofhigh vowels
“lax” question markers
xx
–
–
xx
logophoricity
xx
–
–
x
V–O–Neg
xx
–
–
x
(xx = very common or ubiquitous; xx = common; x = infrequent; – = very rare or absent; (x) and (xx) = Omotic and/or Cushitic)
[Chadic does] not really belong to the area, as the features are mostly untypical for [it]; but they occur recurrently in some member languages which border on the area and which could thus be viewed as partaking in it. . . . Chadic and Nilotic repeatedly show a pattern whereby a feature in individual languages or subgroups is fairly clearly due to contact with languages belonging to the core area.
. C
.................................................................................................................................. As can be seen in the latest extensive bibliography of Hausa and Chadic linguistics published by Newman (b), a wealth of information concerning Chadic languages is now available, providing ample material for linguists interested in typology or language comparison. The family can boast of very high-quality grammars, among them three large reference grammars devoted to Hausa alone (Wolff ; Newman a; Jaggar ), and some remarkable descriptions of less well-known Chadic languages, e.g. Miya (Schuh ). Thanks to the democratization and miniaturization of technology, new areas of research are now available for hitherto little-documented languages, such as information structure and intonation (Leben ; Green ; Green and Jaggar ; Zimmermann ; Hartmann and Zimmermann , , ; Schuh et al. ; Caron ; Caron et al. ). Finally, Frajzyngier should be singled out for his grammatical description of
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
various languages (Frajzyngier , , , , b), but also for his large-scale and typologically innovative review of Chadic languages (Frajzyngier a) which has made this work both easier and tricky. Paradoxically, what is most lacking now is an extensive and systematic Proto-Chadic lexical reconstruction.
R Barreteau, D., and Newman, P. (). ‘Les langues tchadiques’, in D. Barreteau (ed.), Inventaire des études linguistiques sur les pays d’Afrique Noire d’expression française et sur Madagascar. Paris: Conseil International de la Langue Française, –. Caron, B. (). ‘Guus, aka Sigidi (Chadic, West-B, South-Bauchi): grammatical notes and vocabulary’, Afrika und Übersee : –. Caron, B. (). ‘Dott, aka Zoɗi (Chadic, West-B, South-Bauchi): grammatical notes, vocabulary and text’, Afrika und Übersee : –. Caron, B. (). ‘Le Luri: quelques notes sur une langue tchadique du Nigeria’, in P. Boyeldieu and P. Nougayrol (eds.), Langues et cultures: Terrains d’Afrique. Hommages à France Cloarec-Heiss. Leuven and Paris: Peeters, –, http://llacan.vjf.cnrs.fr/PDF/Publications/Caron/Caron_HF_EDLuri.pdf accessed July , . Caron, B. (). ‘South-Bauchi West pronominal and TAM systems’, in B. Caron and P. Zima (eds.), Sprachbund in the West African Sahel. Leuven and Paris: Peeters, –. Caron, B. (). ‘Zaar grammatical sketch’, in A. Mettouchi, M. Vanhove, and D. Caubet (eds.), The CorpAfroAs Corpus. ANR CorpAfroAs: A corpus for Afro-Asiatic languages, http://halshs.archivesouvertes.fr/halshs- accessed July , . Caron, B. (). ‘Number in South-Bauchi West languages (Chadic, Nigeria)’, in A. Storch and G. J. Dimmendaal (eds.), Number-Constructions and Semantics: Case studies from Africa, Amazonia, India and Oceania. Amsterdam and Philadelphia, PA: John Benjamins, –. Caron, B. (). ‘Tone and intonation’, in A. Mettouchi, M. Vanhove, and D. Caubet (eds.), –. Caron, B., Lux, C., Manfredi, S., and Pereira, C. (). ‘The intonation of topic and focus: Zaar (Nigeria), Tamasheq (Niger), Juba Arabic (South Sudan) and Tripoli (Libya)’, in A. Mettouchi, M. Vanhove, and D. Caubet (eds.), –. Clements, G. N., and Rialland, A. (). ‘Africa as a phonological area’, in B. Heine and D. Nurse (eds.), –. Diakonoff, I. M. (). Afrasian Languages. Moscow: Nauka. Frajzyngier, Z. (). ‘Marking syntactic relations in Proto-Chadic’, in E. Wolff and H. MeyerBahlburg (eds.), Studies in Chadic and Afroasiatic Linguistics: Papers from the International Colloquium on the Chadic Language Family and the Symposium on Chadic within Afroasiatic, University of Hamburg, September –, . Hamburg: Helmut Buske, –. Frajzyngier, Z. (). ‘Ventive and centrifugal’, Afrika und Übersee : –. Frajzyngier, Z. (). A Grammar of Pero. Berlin: Dietrich Reimer. Frajzyngier, Z. (). A Grammar of Mupun. Berlin: Dietrich Reimer. Frajzyngier, Z. (). A Grammar of Lele. Stanford, CA: Center for the Study of Language and Information, http://www.loc.gov/catdir/toc/uchi/.html http://www.loc.gov/catdir/ description/cam/.html accessed July , . Frajzyngier, Z. (). ‘L’augment télique (“goal”) dans les langues tchadiques’, in A. Lonnet and A. Mettouchi (eds.), Les langues chamito-sémitiques. Paris and Gap: Ophrys, –. Frajzyngier, Z. (). A Grammar of Gidar. Frankfurt am Main: Peter Lang. Frajzyngier, Z. (a). ‘Chadic’, in Z. Frajzyngier and E. Shay (eds.), The Afroasiatic Languages. Cambridge and New York: Cambridge University Press, –. Frajzyngier, Z. (b). A Grammar of Wandala. Berlin and New York: Mouton de Gruyter.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Frajzyngier, Z., and Munkaila, M. M. (). Grammatical and Semantic Relations in Hausa: “Point of view”, “goal” and “affected object”. Cologne: Rüdiger Köppe, http://www.loc.gov/catdir/toc/ fy/.html accessed July , . Frajzyngier, Z., and Shay, E. (). A Grammar of Hdi. Berlin and New York: Mouton de Gruyter. Greenberg, J. H. (). ‘Studies in African linguistic classification. IV. Hamito-Semitic’, Southwest Journal of Anthropology : –. Greenberg, J. H. (). ‘Internal a-plurals in Afroasiatic (Hamito-Semitic)’, in J. Lukas (ed.), Afrikanistische Studien. Berlin, –. Greenberg, J. H. (). ‘Africa as a linguistic area’, in W. R. Bascom and M. J. Herskovits (eds.), Continuity and Change in African Cultures. Chicago, IL: University of Chicago Press, –. Greenberg, J. H. (). ‘An Afro-Asiatic pattern of gender and number agreement’, Journal of American Oriental Society : –. Greenberg, J. H. (). The Languages of Africa. The Hague: Mouton. Greenberg, J. H. (). ‘Some areal characteristics of African languages’, in I. R. Dihoff (ed.), Current Approaches to African Linguistics. Dordrecht: Foris, –. Green, M. (). Focus and Related Phenomena and the Syntax of Hausa. Ph.D. thesis. University of London. Green, M., and Jaggar, P. J. (). Ex-situ and In-situ Focus in Hausa. (Cognitive Science Research Papers, .) University of Sussex: School of Cognitive and Computing Sciences. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), –. Hartmann, K., and Zimmermann, M. (). ‘In place-out of place? Focus in Hausa’, in K. Schwabe and S. Winkler (eds.), On Information Structure, Meaning and Form: Generalizing across languages. Amsterdam and Philadelphia, PA: John Benjamins, –. Hartmann, K., and Zimmermann, M. (). ‘Morphological focus marking in Gùrùntùm (West Chadic)’, Lingua : –. Hartmann, K., and Zimmermann, M. (). ‘Focus marking in Bura: semantic uniformity matches syntactic heterogeneity’, Natural Language & Linguistic Theory : –, http://www.jstor.org/ stable/ accessed April , . Haruna, A. (). A Grammatical Outline of Gùrdùŋ/Gùrùntùm. Cologne: Rüdiger Köppe. Hayward, R. J. (). ‘Afroasiatic’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –. Heine, B., and Leyew, Z. (). ‘Is Africa a linguistic area?’, in B. Heine and D. Nurse (eds.), –. Heine, B., and Nurse, D. (eds.) (). A Linguistic Geography of Africa. Cambridge: Cambridge University Press. Hodge, C. T. (). ‘Afroasiatic: an overview’, in T. A. Sebeok (ed.), Southwest Asia and North Africa. (Current Trends in Linguistics, .) The Hague: Mouton, –. Jaggar, P. J. (). ‘Guruntum (gùrdùŋ) (West Chadic-B): linguistic notes and wordlist’, African Languages and Cultures : –. Jaggar, P. J. (). Hausa. Amsterdam and Philadelphia, PA: John Benjamins. Jungraithmayr, H. (). ‘On root augmentation in Hausa’, Journal of African Languages : –. Jungraithmayr, H. (). ‘Reflections on the root structure in Chadohamitic (Chadic)’, in Actes du e Congrès International de Linguistique Africaine, Abidjan – mars . Vol. . Abidjan: Université d’Abidjan, –. Jungraithmayr, H. (). ‘A tentative four stage model for the development of Chadic languages’, in P. Fronzarolli (ed.), Atti del secondo congresso internazionale di linguistica camito-semitica (Firenze, – aprile ). Florence: Istituto di Linguistica e di Lingue Orientali, Università di Firenze, –. Jungraithmayr, H., and Ibriszimow, D. (). Chadic Lexical Roots. vols. Berlin: Dietrich Reimer. Kleinewillinghöfer, U. (). ‘Aspects of vowel harmony in Waja and Tangale common vocabulary’, Frankfurter Afrikanistische Blätter : –. Leben, W. R. (). ‘Intonation in Chadic languages’, Studies in African Linguistics, Supplement : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Lewis, M. P., Simons, G. F., and Fennig, C. D. (eds.) (). Ethnologue: Languages of the World, th edn. Dallas, TX: SIL International. http://www.ethnologue.com accessed July , . Lovestrand, J. (). Classification and Description of the Chadic Languages of the Guéra (East Chadic B). (SIL Electronic Working Papers -). Dallas, TX: SIL International. Lukas, J. (). ‘The linguistic situation in the Lake Chad area in Central Africa’, Africa : –. Mettouchi, A., Vanhove, M., and Caubet, D. (eds.) (). Corpus-Based Studies of Lesser-Described Languages: The CorpAfroAs corpus of spoken AfroAsiatic languages. Amsterdam and Philadelphia, PA: John Benjamins. Newman, P. (). ‘Syllable weight as a phonological variable’, Studies in African Linguistics : –. Newman, P. (). ‘Proto-Chadic verb classes’, Folia Orientalia : –. Newman, P. (). ‘The origin of Hausa /h/’, Studies in African Linguistics, Supplement : –. Newman, P. (). ‘Chadic classification and reconstructions’, Afroasiatic Linguistics : –. Newman, P. (). The Classification of Chadic within Afroasiatic. Leiden: Universitaire Pers Leiden. Newman, P. (). Nominal and Verbal Plurality in Chadic. Berlin: Walter de Gruyter. Newman, P. (a). The Hausa Language: An encyclopedic reference grammar. New Haven, CT: Yale University Press. Newman, P. (b). ‘Causatives (and related formations)’, in P. Newman (a), –. Newman, P. (). ‘Comparative Chadic revisited’, in P. Newman and L. M. Hyman (eds.), West African Linguistics: Papers in honor of Russell G. Schuh. (Studies in African Linguistics, Supplement .) Columbus, OH: Department of Linguistics and the Center for African Studies, Ohio State University, –. Newman, P. (a). The Chadic Language Family: Classification and name index. Mega-Chad Research Network / Réseau Méga-Tchad, http://www.megatchad.net/publications/Newman-Chadic-Classification-and-Index.pdf accessed September , . Newman, P. (b). Online Bibliography of Chadic and Hausa Linguistics, https://scholarworks.iu. edu/dspace/handle// accessed February , . Newman, P., and Ma, R. (). ‘Comparative Chadic: phonology and lexicon’, Journal of African Languages : –. Newman, P., and Schuh, R. G. . ‘The Hausa aspect system’, Afroasiatic Linguistics : –. Schuh, R. G. (). ‘The Chadic verbal system and its Afroasiatic nature’, Afroasiatic Linguistics : –. Schuh, R. G. (). ‘West Chadic verb classes’, in P. Newman and R. Ma Newman (eds.), Papers in Chadic Linguistics. Leiden: Afrika-Studiecentrum, –. Schuh, R. G. (). A Grammar of Miya. Berkeley and Los Angeles, CA: University of California Press. Schuh, R. G. (). ‘Chadic overview’, in M. L. Bender, G. Takács, and D. L. Appleyard (eds.), Selected Comparative-Historical Afrasian Linguistic in Memory of Igor M. Diakonoff. Munich: LINCOM Europa, –. Schuh, R. G., Gimba, A. M., and Ritchart, A. (). ‘Bole intonation’, UCLA Working Papers in Phonetics : –. Wolff, H. (). ‘Subsystem typologies and area linguistics’, Anthropological Linguistics : –. Wolff, H. E. (). ‘Grammatical categories of verb stems and the marking of mood, aktionsart, and aspect in Chadic’, Afroasiatic Linguistics : –. Wolff, H. E. (). ‘Consonant-tone interference in Chadic and its implications for a theory of tonogenesis in Afroasiatic’, in D. Barreteau (ed.), Langues et cultures dans le bassin du Lac Tchad. Paris: ORSTOM, –. Wolff, H. E. (). Referenzgrammatik des Hausa: Zur Begleitung des Fremdsprachenunterrichts und zur Einführung in das Selbststudium. Münster: LIT. Zimmermann, M. (). ‘Focus in western Chadic: a unified OT account’, in C. Davis, A.-R. Deal, and Y. Zabbal (eds.), NELS : Proceedings of the Thirty-Sixth Annual Meeting of the North East Linguistic Society. Amhearst, MA: University of Massachusetts, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
- ......................................................................................................................
E-S (Ethiosemitic, Ethiopian Semitic; henceforth E-S) languages form a group within the Semitic family of the Afro-Asiatic (Hamito-Semitic) language phylum. These languages are spoken in the Horn of Africa and adjacent territories of East Africa, mostly in Ethiopia and Eritrea, as well as in some border areas of Sudan (see Map .). When Eritrea became independent, there were attempts to change the name of the group into a politically more correct one (Afro-Semitic, Ethio-Eritrean Semitic), but these attempts were not successful. E-S is a long-established language entity and its composition was never questioned in so far as there were no attempts to include in this group languages from outside the defined geographical area or to exclude from it languages already classified as E-S. Thus, in all the surveys of languages and language families of the world the E-S group is always present and is by default treated as a genetic unit within Semitic. However, in the course of comparative studies the genealogical unity of E-S was put in doubt. For example, one recent publication describes only three isoglosses which may be considered as shared innovations, making it possible to consider E-S a separate genetic entity (Weninger : ): () agentive nouns of the pattern *CaaCaCi(y), which replaced the Proto-Semitic participle *CaaCiC-; () the existential verb *hlw; () infinitive ending *-ot. It is quite evident that the above arguments are too weak to postulate E-S as a genetic entity in its own right. Still, it seems quite justified to preserve E-S as a taxonomic unit within Semitic. Its traditional status is supported by a long history of population mixing, language contacts, and interference within the framework of specific sociocultural, geographical, and ecological paradigms. All these factors contributed to forming a particular linguistic milieu, different from all other Semitic languages. It must be added that genetic classification on intermediate levels within well-established language families is always one of the most contested issues in comparative linguistics, since there are no infallible and generally accepted methods for segregating genetically valid isoglosses from areal ones, induced by intensive and protracted contacts between cognate languages (for more details, see Porkhomovsky, chapter of this volume). Thus, it is generally accepted that many specific common E-S features on all language levels are due to a Cushitic (mostly Agaw) substratum.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Ethio-Semitic languages
This controversy about the genetic unity of E-S makes it possible to surmise two possible scenarios for the migration of Semitic-speaking population from South Arabia to eastern Africa, namely, that these migrants either spoke dialects belonging to a single dialect cluster or were of heterogeneous descent within Semitic. At the same time it is necessary to emphasize major changes as far as the position of E-S within common Semitic is concerned. All these changes resulted from radical transformation
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
of the panorama of comparative Semitic studies in recent decades. Suffice it to focus on the two most important aspects: the general shift of emphasis from Arabic, Canaanite (especially Ancient Hebrew), and Aramaic towards Akkadian in Proto-Semitic reconstructions, and full-scale inclusion of modern South Arabian languages in comparative Semitic linguistics. In the early stages of Semitic comparative studies E-S languages were included in South Semitic (SS) together with Arabic, Epigraphic South Arabian (ESA), and Modern South Arabian (MSA). One of the major isoglosses for postulating SS, i.e. broken plurals in nominal morphology, has lost its classification value for various reasons. Then Arabic was excluded from SS, mostly on the basis of its prefix-conjugated imperfective verbal form with reduced vocalism (see below); later ESA was also excluded from this group for the same reason. As to the classification of MSA and E-S as constituting SS, only the most important isoglosses used for postulating such an entity will be discussed below due to the limited size of the present chapter. The presence of -k- based suffixed personal markers in perfective in both MSA and E-S as opposed to -t- based markers in Arabic and Northwest Semitic (Canaanite, Aramaic) is now considered an areal rather than genetic isogloss, since it is also attested in some Arabic dialects of South Arabia and in ESA. Glottalized (ejective) representation of common Semitic emphatics has long been noted as a specific feature of E-S, as opposed to pharyngealization in traditional Arabic pronunciation, which was then interpreted as a reflex of the common Semitic prototype. At this early stage glottalized E-S pronunciation was explained as an areal African feature, most probably induced by contact with Cushitic. When glottalized representation of common Semitic emphatics in MSA languages (Mehri, Socotri, etc.) was acknowledged as valid for Proto-Semitic comparative studies, and glottalization was also reconstructed for Akkadian emphatics, it became clear that glottalization of emphatics in both E-S and MSA is a shared Semitic archaism and cannot be used as an argument for genetic classification within the alleged SS group. A somewhat similar evolution is attested for the interpretation of the most important morphological isogloss: prefix conjugation. There are two different patterns of prefix conjugation in Semitic: a fully vocalized form [pattern of the basic stem – yvCvC(C)vC] and a form with reduced vocalism [yvCCvC]. The fully vocalized form functions as imperfective in Akkadian, E-S, and MSA; it is absent in Central Semitic (CS) languages— Northwest Semitic, Arabic, and ESA. The form with reduced vocalism is attested as perfective and as jussive/subjunctive (with some changes) in Akkadian, as jussive/subjunctive in E-S and MSA, and as imperfective and as jussive/subjunctive (with different changes in particular languages) in CS. A suffix-conjugated form is also attested in Semitic: as stative in Akkadian (where it is not a finite verbal form) and as perfective in all other Semitic languages. All three verbal patterns exhibit obvious formal parallels in all Semitic languages; however, the reconstruction of Proto-Semitic archetypes has always been the most difficult task. For example, Meillet emphasized: Les langues sémitiques sont plus semblables entre elles que ne le sont les langues indoeuropéennes; à les observer, on a souvent l’impression de formes diverses d’une même langue plutôt que de langues vraiment différenciées, comme le sont les langues indo-européennes; et malgré cela, on n’arrive pas à poser un « sémitique commun », un ursemitisch, comme on pose un «indo-européen commun», un urindogermanisch. (Meillet : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Despite all the differences in the comparative analysis of the Semitic verbal system (cf. the model of Proto-Semitic verbal reconstruction as proposed by the present author; Porkhomovsky ) it is obvious that the fully vocalized prefix-conjugated pattern attested as imperfective in Akkadian, E-S and MSA belongs to common Semitic. This isogloss cannot be used for classifying E-S and MSA together as an integral SS subgroup within Semitic, since it is not a shared innovation (Porkhomovsky : ). Hence, the place of E-S in the classification of Semitic languages may be described as follows: E-S forms a unit on a par with MSA and CS, the latter including North-Central Semitic (Canaanite, Aramaic, Ugaritic), Arabic, and ESA. These three groups (CS, MSA, E-S) form West Semitic, which is opposed to East Semitic, including Akkadian and possibly Eblaite (cf. Huehnergard and Rubin : ). This bifurcation into West and East Semitic is the first split of Proto-Semitic. The internal genetic classification of E-S is a problematic issue: different hypotheses and classification models have been proposed in the course of Ethiopian studies. The problem of the status of E-S as a genetic or as an areal unit, as described above, only aggravates the situation. E-S includes Ge‘ez (Old Ethiopic, Classical Ethiopic) with written records going back to the first centuries ,1 plus different unwritten languages or languages with comparatively short written traditions (for more on this, see below). In the early stages of Ethiopian studies Ge‘ez was considered the ancestral language of all E-S languages, just as Sanskrit was for Indo-European languages at the beginning of Indo-European studies. Again, as in the case of Sanskrit, the hypothesis that Ge‘ez should be regarded as a proto-language for all E-S languages was abandoned, and they were divided into Northern and Southern groups (Cohen ). Ge‘ez, Tigre, and Tigrinya (Tigray) were included in the Northern group and all other E-S languages formed the Southern group. No known E-S language may be considered a direct descendant of Ge‘ez. According to Hetzron’s classification (, ), Southern E-S languages are divided into Outer and Transversal groups. Hetzron demonstrated that the traditional Gurage cluster is not a genetic entity. East Gurage (Selti, Wolane, Ulbarag, Inneqor, Zway) together with Amharic, the national language of Ethiopia, Argobba, and Harari (Adari) form Transversal South Ethiopian, while North Gurage (Soddo, Kistane, Aymellel) and West Gurage (Chaha, Ezha, Gumer, Ennemor (Inor), Gyeto, Endegeñ, Masqan, Muher, Goggot) belong to Outer South Ethiopian together with extinct Gafat. The internal classification of E-S provokes many controversies on all levels including the language ~ dialect level. Thus, the bifurcation of E-S into a Northern and a Southern group was challenged by Voigt (: ), who examined all linguistic data in support of this binary classification and concluded “that there is no longer any argument for a strict genealogical separation into two language groups in Ethio-Semitic, a Northern and a Southern group.” Ethiopia occupies an exclusive place in African studies thanks to the long written tradition of Ge’ez. Writing was brought into Ethiopia from Southern Arabia. The ESA (Sabaean) consonant alphabet adopted for Ge’ez was later reformed, most probably under Greek and Indian (Kharoshti) influence, and became a syllabic one. This change from a consonantal to a syllabic writing system with a specific way of presenting vowels by adding
1
It ceased to be used as a spoken language around the turn of the millennium but is still being employed in Ethiopian Christian Church liturgy.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
markers to the respective consonant characters is attested only in Indian and Ethiopian writing systems (Friedrich ). The direction of writing was also changed from right to left to left to right. The Ethiopian writing system is used for some other E-S languages, namely Amharic (since the sixteenth century), Tigrinya (since the nineteenth century), and Tigre (since the end of the nineteenth century). The Ethiopian syllabic system is applied to some unwritten E-S languages (as well as to some Cushitic languages of Ethiopia), or is replacing Arabic writing in Islamic areas.2 Ethiopian linguistics forms one of the well-established branches of Oriental and African studies, with a long-standing tradition and rich literature. The scope of this chapter does not permit presenting the history of Ethiopian linguistics (for which, see, for example, Leslau ; Cohen ; Weninger ), but one name deserves special mention here: Wolf Leslau (–), the patriarch of Ethiopian studies. In the course of many decades of field research in Ethiopia Leslau collected abundant data on numerous languages that had not been documented previously, and thus made an unprecedented contribution to Ethiopian linguistics and philology. He published grammars and dictionaries of many E-S languages, including etymological and comparative dictionaries, as well as a large collection of synchronic and comparative historical studies.
R Cohen, D. (). ‘Le sémitique en Éthiopie’, in J. Perrot (ed.), Les langues dans le monde ancien et moderne. Part III: Les langues chamito-sémitiques. Paris: Éditions du CNRS, –. Cohen, M. (). Études d’éthiopien méridional. Paris: Geuthner. Friedrich, J. (). Geschichte der Schrift. Heidelberg: Carl Winter. Hetzron, R. (). Ethiopian Semitic: Studies in classification. Manchester: Manchester University Press. Hetzron, R. (). ‘Genetic classification and Ethiopian Semitic’, in J. and T. Bynon (eds.), HamitoSemitica: Proceedings of a colloquium held at the School of Oriental and African Studies, University of London, on the th, th and th of March . The Hague and Paris: Mouton. Huehnergard, J., and Rubin, A. D. (). ‘Phyla and waves: Models of classification of the Semitic languages’, in S. Weninger et al. (eds.), –. Leslau, W. (). ‘Ethiopic and South Arabian’, in T. S. Sebeok (ed.), Linguistics in South West Asia and North Africa. (Current Trends in Linguistics, .) The Hague and Paris: Mouton, –. Meillet, A. (). ‘Sur le degré de précision qu’admet la définition de la parenté linguistique’, in F. Boas, O. Dempwolff, G. Panconcelli-Calzia, A. Werner, and D. Westermann, (eds.). Festschrift Meinhof: Sprachwissenschaftliche und andere Studien. Hamburg: L. Friederichsen, –. Porkhomovsky, V. (). ‘Modern South Arabian Languages from a Semitic and Hamito-Semitic perspective’, Proceedings of the Seminar for Arabian Studies. : –. Porkhomovsky, V. (). ‘Hamito-Semitic aspect system: the case of Semitic and Berber’, Sprache und Geschichte in Afrika : –. Voigt, R. (). ‘North vs. South Ethiopian Semitic’, in S. Uhlig (ed.), Proceedings of the th International Conference of Ethiopian Studies. Vol. IV. Trondheim: NTNU-tryck, –. Weninger, S. (). ‘Ethio-Semitic in general’, in S. Weninger et al. (eds.), –. Weninger, S., Khan, G., Streck, M. P., and Watson, J. C. E. (eds.) (). The Semitic Languages: An international Handbook. Berlin and Boston, MA: de Gruyter Mouton.
2
For example, Harari had an Arabic-based writing system frome the eighteenth century; now it is replaced by Ethiopian writing except for Islamic religious purposes.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
- ......................................................................................................................
.
. I
.................................................................................................................................. I an attempt to eliminate from the genetic classification of African languages racial or cultural biases, the anthropologist and linguist Joseph H. Greenberg published a series of articles in the Southwestern Journal of Anthropology between and . Among the innovative ideas initiated in these studies, which appeared as a monograph shortly after (Greenberg ), was the establishment of a Macro-Sudanic family. This genetic grouping, which was renamed Chari-Nile following a suggestion by the linguist William Welmers, consisted of two major branches, Central Sudanic (CS) and Eastern Sudanic (ES), plus the Berta and Kunama language cluster. Other families identified as genetic groupings mainly due to the pioneering descriptive work of linguists like Margaret Bryan, Roland Stevenson, and Archibald Tucker were subsequently pulled together in a new phylum of which ChariNile formed the core, termed Nilo-Saharan (NS) by Greenberg () (see Map .). Table . summarizes Greenberg’s ideas on the genetic classification of NS, based on socalled mass comparison as well as the position of the present author on the subject. Further details as well as a survey of comparative work on lower-level units within NS are given in Dimmendaal (: –, –) and Dimmendaal et al. (). The methodological background to Greenberg’s mass comparison has been discussed extensively in the literature (e.g. by Newman ; Croft : xiii–xxxvi; and Dimmendaal : –, –). Without his application of multilateral comparisons, in particular with respect to grammatical morphemes, the reality of NS as a phylum probably would have gone unnoticed until today. At the same time, it is clear that the results emerging from Greenberg’s pioneering work are in need of refinement. The problem may be illustrated by way of the following example. Whereas Greenberg () treats Kuliak as part of the NS phylum, Sands () prefers to treat this group of languages in northeastern Uganda as an independent stock. Dimmendaal () adheres to the Greenbergian position, because, amongst other reasons, Kuliak (Rub) manifests reflexes of widespread grammatical morphemes in NS, such as the accusative case marker -ka (< *-ka), or derivational prefixes on
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
. The geographical distribution of Nilo-Saharan languages
the verb, such as the causative marker ɪ-/i-, which is one of the most widespread and stable diagnostic properties of NS. One of the languages not mentioned in Greenberg (), but added as a possible candidate for a genetic affiliation with NS in Greenberg (), is the extinct language of the Meroitic kingdom in northern Sudan. Trigger () was the first author to suggest that this language is a member of the ES branch within NS. Due to the detailed investigation by Rilly (), evidence emerged in favor of a closer relationship with ES subgroups such as Nubian, Taman, Nara, and Nyimang, plus Afitti. As the brief following survey should make clear, to date neither the internal classification nor the limits of NS have been settled in a range of other cases. In his comparative study of NS, Bender () has argued that the Kadu (or Tumtum) languages along the southern edge of the Nuba Mountains (Sudan), which were classified as Kordofanian (Niger-Congo) by Greenberg (), also belong to NS. Ehret (: ) calls Kadu (Kadugli-Krongo) “a group of doubtful connection to the Nilo-Saharan family”, a position shared by the present author. Given the (presumably) tremendous time depth and corresponding typological diversification involved, it is important to identify as many likely grammatical cognates as possible, as these tend to be more stable and less subject to borrowing than lexical morphemes. Such a list with “diagnostic properties” is used in the present chapter to include or exclude specific languages or groups, for example the Kadu languages. This group shows some similarities in its pronominal system to the pattern described in
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Table . Nilo-Saharan and its limits Greenberg 1955
Greenberg 1963
Classification by the present author
Songhai
isolate
Songhai
isolate
Central Saharan
isolate
Saharan
Nilo-Saharan
Maban
isolate
Maban
Nilo-Saharan
Mimi
isolate
Mimi
isolate
Fur
isolate
Fur
Nilo-Saharan
Nyangiyan
isolate
Nyangiya, Teuso
Nilo-Saharan
Temainian
isolate
Temein, Keiga Jirru
Nilo-Saharan
Nubian
Nubian
Beir-Didinga
Surmic
Barea
Nara
Tabi
(West) Jebel
Nyimang Merarit
Eastern Sudanic
Nyimang, Dinik MacroSudanic
Dagu
Taman
Nilo-Saharan
Daju
Nilotic
Nilotic
Great Lakes
Nilotic
Central Sudanic
Central Sudanic
Nilo-Saharan
Berta
Berta
Nilo-Saharan
Kunama
Kunama
Nilo-Saharan
Koman, B’aga
isolate?
Koman, Gumuz
isolate
section .. below, but lacks other prominent NS features. The Shabo language in southwestern Ethiopia is another case in point. It has an accusative case marker -ka(k), reminiscent of the Northeastern NS marker with the same function, but there is no evidence (as far as present knowledge goes) for additional cognate grammatical markers. Both consequently are treated as isolates by the present author (see Dimmendaal ). The three languages called Mimi (in Chad) have also caused problems in terms of genetic classification. One language called Mimi has turned out to be closely related to the NS language Fur, and is also referred to as Amdang. The genetic position of Mimi of Decorse (after the person who collected a wordlist around , published by GaudefroyDemombynes : –) and Mimi of Nachtigal (after the explorer who reported its
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Songhay Maban Fur and Amdang Northeastern
Kunama Saharan
Nilo-Saharan Eastern Sudanic Kuliak Central Sudanic
Koman, B‘aga
. The subclassification of Nilo-Saharan
existence) is unclear, because the few lexical similarities with the Maban group within NS may well be due to borrowing. Either or both of these may, therefore, constitute linguistic isolates; see Starostin () for a balanced account. Nicolaï () has shown that most of the evidence adduced in Greenberg () for the affiliation of the Songhai cluster with NS is not convincing. The situation is less clear with respect to the Koman cluster, which may indeed be NS, as argued in section .. below. Similarly, Gumuz and two further languages, Daats’iin and Kadallo (which were identified more recently), forming the B’aga language cluster, may or may not be part of NS. It is well known from the comparative study of Indo-European as well as other language families that it is extremely difficult to arrive at more refined subgroupings for deeper historical levels. Consequently, NS specialists like Bender () and Ehret () disagree on the correct subclassification (Figure .). Although the present author’s position is more in line with Ehret (), a relatively “flat” structure is assumed in the present contribution involving a primary split between CS and Northeastern NS (as it is called here), leaving the subgrouping of most lower-level units open for the time being.1 Within the NS phylum (as it is understood by the present author), there is a rather sharp typological split between CS and the remaining branches. As this divergence can be related to common innovations either in these remaining groups or in CS, we take this division to correspond to a genetic split. This split involves, inter alia, changes in constituent order, the emergence of case, extensive number marking, and converbs in Northeastern NS. CS languages on the other hand manifest a historical drift towards segmental reduction and extensive use of compounding. These diverging historical developments probably are the result of areal contact with typologically different languages, as argued in section . below. 1 The family tree presented here (see Figure .) roughly corresponds to the tree presented by Ehret (), except for the fact that he adds Rub (Kuliak) to ES (or Eastern Sahelian in Ehret’s terminology) rather than treating it as a more isolated unit, as in the present study.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
ES constitutes the largest subgroup within NS, and consists of the following branches (Figure .): Taman Meroitic Northern
Nubian Nara
Nyimang plus Afitti
Eastern Sudanic
Berta Jebel Southern Daju Temeinian Surmic Nilotic
. The subclassification of Eastern Sudanic
The “Northern” subgrouping is based on Rilly (). The grouping of the Daju cluster, Temeinian, Nilotic, and Surmic into one sub-branch, called Southern ES is supported, for example, by the innovation of an irregular (suppletive) alternation for ‘cow’ (Dimmendaal : ): () Proto-Nilotic Majang (Surmic) Temein Daju of Lagowa Gaahmg (Jebel)
singular d̪εŋ taŋ ńtέŋ teɲe tɔɔ
plural/collective d̪ʊk togi kítúk tukke tɔg
The evidence from Gaahmg (probably the only Jebel language still spoken today) is inconclusive in this respect because the alternation for ‘cow(s)’ is not irregular. However, due to the detailed monograph by Stirtz () on this language, it is now possible to arrive at a more definitive subgrouping for this language and thereby for the Jebel group within ES. Like several Nilotic and Surmic languages, Gaahmg uses a case marker -ε for post-verbal subjects, a feature absent from Northern members of ES (which do not allow for post-verbal subjects). Consequently, Gaahmg (Jebel) probably is more closely related to the Southern than the Northern group within ES. The grammatical morphemes of the Kuliak (Rub) languages do not show any obvious affinity with Northern or Southern ES; they consequently appear to constitute an earlier split from Northeastern NS rather than being part of ES, as claimed by Greenberg ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
. S
..................................................................................................................................
.. The n/k alternation Among the most stable properties of the NS phylum is the alternation, expressed primarily on the verb, between singular n and plural k, a common feature already identified by Greenberg (: –, ). Greenberg (: ) refers to demonstratives in the CS language Moru, where this alternation is attested, e.g. ana (SG)/kai (PL) ‘that/those’. Moser (: ) mentions a referential demonstrative -n in the CS language Kabba which “follows a given, known or presupposed piece or information”. The logophoric (third person singular or plural) subject pronoun ǹ- occurring on verbs in the same language presumably is another reflex of this original deictic marker. Tucker (: ) observes that the plural suffixes may also “be attached to verbs or verbal auxiliaries in sentences where the subject is plural”. In CS languages like Ma’di, the plural marker -kɪ is not (or no longer used) with nouns, but it is still used with this function as a verbal (pluractional) marker: () odra-ki-ra. die-- ‘They died.’ This kind of system attested in CS languages synchronically provides a natural explanation for their use as singulative (-n-) versus pluractional (-k-) markers on verbs in Northeastern NS language groups like Maban (data on Maba from Weiss ): () à-wá:n-á. à-wà:k-ír-ì. ‘I poured a bit.’ ‘I poured a lot.’ Reflexes of the n/k alternation may also be found in the ES branch of NS, for example in demonstratives in Surmic languages like Didinga (Lokonobei and de Jong : ) with corresponding voicing of the intervocalic stop:
proximate distant cε-nɪ ca-ni cε-gɪ ca-gi
But in parallel to the Maban group within Northeastern NS, one also finds reflexes of the n/k alternation in verb forms or nominal modifiers based on verbs in ES groups. Compare the following alternations in the Nilotic language Turkana: () singular ɪ-mɔ-n-a. -ka-mɔn-anɪ ̥ ε-ka-mɔn-anɪ ̥
plural ɪ-mɔ-k-a. ‘You (/) are mean.’ -ka-mɔn-ak ‘mean’ (nominal modifier) ŋɪ-ka-mɔn-ak ‘miser’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
.. The causative A second archaic property reconstructable for the earliest stages of NS is a causative marker *i- (Dimmendaal ); this prefix might have alternated with *ɪ-, if the proto-language had [ATR]-harmony. The causative marker *i- has been retained in different subgroups of Northeastern NS, for example in Nilotic and Surmic, as shown by the following examples from the Nilotic language Kipsikiis: () -nέr ‘be fat’
-ì-nέε̂r ‘fatten’
Unseth () presents a survey of this archaic feature in Surmic languages, for example in Me’en (in ()) and Majang (in ()): () dibis ‘be full’
-idibis ‘fill’
() bod- ‘be safe, well’ ɪ-bod- ‘rescue, save’ In CS languages, this prefix is part of a set of (often highly lexicalized) valency-changing prefixes, as first described in Tucker ( []: –). Blackings and Fabb (: ) give the following examples for Madi: () zɔ̄ ‘to grow’ ī ̄z̵ ɔ́ ‘to bring up’ tū ‘to climb up’ ītú ‘to make climb up; promote’ Different NS languages show indirect reflexes of the original prefix *i-, e.g. ES Gaahmg, rāgg ‘stop (itr.)’/rə́ə̄-d ‘stop (tr.)’ (Stirtz : ), or the Daju language Shatt, sia- ‘eat’, zzi-a ‘feed’ (Tucker and Bryan : ). Tucker and Bryan (: ) also point towards the presence of a class of mainly transitive verbs with an initial (petrified) vowel in Kunama and their similarity to Nilotic, e.g. -isasa ‘tell’, or -ite ‘find’. Other derivational prefixes attested in CS, such as the ventive or iterative, are expressed by way of suffixes or case markers on noun phrases in Northeastern NS, although sometimes these former vocalic prefixes appear to have been retained as lexicalized forms. One such NS branch with common V-CVC verb stems is Kuliak. Compare the following reconstructions for Proto-Kuliak from Heine (): () *ekakw *ekat *iykesk *ẹbit *eset *ikkεt
‘herd animals’ ‘increase’ ‘know’ ‘plant’ ‘return something’ ‘strangle’
This pattern coexists with (the cross-linguistically common) -CVC root pattern in Kuliak. Similar -V-CVC patterns are attested in Maban languages, though not so frequently as in Kuliak. Synchronically, these vocalic elements (other than the causative) do not have a separate morphological status in Northeastern NS. Historically, they may be cognate with
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
the prefixes still found in CS languages, but this requires a systematic comparison of lexical roots. A borrowing scenario from CS in, for example, Kuliak, of course, cannot be excluded, but is extremely unlikely for the following reasons: first, verbs in Kuliak usually end in a consonant, i.e. have the canonical shape -V-CVC, whereas the typical CS pattern is -V-CV. Second, -V-CVC verb stems are extremely common in Kuliak. Third, Kuliak languages are not geographically contiguous—at least today—to CS languages. There is a second causative marker in Northeastern NS, attested as -ɪta- (plus allomorphs) in Nilotic, da- in Nara, yitə (plus allomorphs) in Saharan, and -ndV- in the Maban group, probably going back to a combination of causative *i- plus an auxiliary verb *ta. () base Nara dengi ‘wait’ Kanuri lúwùkìn ‘I leave’ Maba yàŋàn ‘having drunk’
causative da-dengi ‘cause to wait’ tù-lúwùkìn ‘I take out’ nd-àŋ-á ‘make drink!’
Spontaneous nasalization of an earlier vocalic causative prefix (i-n-C < *i-iC) as in Maban is also found synchronically in the Nilotic language Maasai whenever a preceding prefix ends in a high front vowel and the first stem consonant is a plosive or sh (example adapted from Tucker and Mpaayei : ): ()
kɪ-ɪba ) kɪmba ‘we hate’ -hate
The causative prefix is sometimes combined with a derivational suffix -(ɪ)sɪ in Nilotic or Surmic languages, for example. Pace Ehret (: ), this suffix is not a causative marker itself (nor is it linked to the widespread Afro-Asiatic (AA) causative -is). Instead, this suffix usually performs a different role, namely detransitivization or indirect causation (‘cause to do . . . ’), as in Southern Nilotic Kipsikiis (data from Toweett ): ()
kɪɪ-laaŋ-sɪ laakwεεt keetit .-climb- child: tree: ‘to make the child climb the tree’
.. The reflexive (middle voice) marker Among the brilliant intuitions of the late Joseph Greenberg, which led to the postulation of a NS phylum, is a third archaic property, a so-called “verb reflexive in r” (Greenberg : ), based on the word for ‘self ’ or ‘body’, still found as such in CS languages, e.g. Lugbara ro or Kabba rɔ. In the Saharan language Kanuri, one finds ro as a reflex; this form is also used to express comparison in this language (parallel to the use of the word for ‘body’, lic, to express a comparison ‘like’ in Old English).2 The form with a corresponding front vowel in this formative probably is a reflex of an archaic third person logophoric pronoun *ε. In CS
2
Ehret (: ) proposes an original NS root form *ru or *ruh ‘self/body’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
languages like Kabba, for example, this noun is often followed by the third person pronoun (rɔ-ε > rεέ): ()
ń-tɔl rεέ. -kill body: ‘He killed himself.’
Reflexes may also be found in ES, where this marker usually expresses middle voice. Across Nilotic, for example, one also finds reflexes of this former reflexive marker, commonly functioning as a middle voice (or “unergative”) marker for Eastern Nilotic Turkana -rε. Note also the following infinitive complements of main verbs from Turkana, with a reflex of the archaic Nilo-Saharan n/k alternation: a-tɔ-n-εrε (SG)/a-tɔ-k-εrε (PL) ‘die’.
.. Pronominal reference Pronominal alignment systems across NS vary in a number of respects from a typological point of view. First, languages may use independent (syntactic) pronouns as reference markers, without any subject or object marking or agreement on the verb, as is the case in ES languages like Bari or Nyimang. This pattern, however, is rare. The more common system is to use bound pronominal elements on the verb in combination with free pronouns. These bound markers may involve prefixes, suffixes, proclitics, or enclitics, sometimes in combination with each other. Given these differences and the apparent historical dynamics (or instability) involved in this grammatical subsystem, it is rather remarkable that several pronominals appear to be cognate, as already pointed out by Greenberg (). Greenberg (: –, –, –) identified a “first person singular pronoun in a” (p. ), and a “second person singular pronoun in i” (p. ). Most likely, however, the two pronominal elements also contained an initial consonant. In CS languages, an initial nasal m- is found with first and second person independent pronouns, as shown in the survey by Tucker (: –).3 Examples from Lugbara (Crazzolara (: ): ma (alternative áma) mi (alternative émi)
àma èmï
However, given the fact that the same nasal is found with all pronouns, the former probably constituted a separate morpheme. This is supported by the fact that shorter forms without the nasal occur in dependent verb forms in Lugbara (examples from Tucker and Bryan : ): () á-tsɔ -beat í-tsɔ -beat 3
mvá ‘that I may beat the child’ child mvá! ‘Beat the child!’ child
In Madi, the second person has a palatal nasal (nyi/a-nyi), whereas in Lendu it is ni (Tucker : ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
It is these forms for first and second person singular in CS which should be compared with corresponding forms in Northeastern NS. Whereas in various languages belonging to that sub-branch, the first and second person pronoun indeed consist of a vowel, there are also Northeastern NS languages which have a distinct initial consonant for the two pronouns. Reflexes suggest that the original first person pronoun was either *qa or *ka, and the second person pronoun *ji. In ES languages like Tama (Taman group), the first person singular pronoun is wa, whereas in Temein the corresponding prefix is ŋa. Western Nilotic Dinka has ɣa, whereas in the Surmic language Baale the reflex is ka-. The presence of an original uvular or velar stop is also suggested by allomorphs of the first person marker on the verb in Eastern Nilotic (Maasai ka-; Teso-Turkana ka-). Similarly, the second person singular reflexes in different Northeastern NS branches suggest that the original form may have been *ji. Whereas a reflex i is widespread across the family (including the CS branch), forms such as ji in Fur suggest the presence of an original initial palatal consonant. Forms with a vowel -u- for the second person singular (also in possessive constructions, as in Kunama) probably resulted from the honorific use of a second person plural form (which is more common cross-linguistically). Greenberg (: ) in fact lists a second person plural in w-. As pointed out by Greenberg (: –), a- (first person singular) and i- (second person singular) are also attested in Koman languages. Of course, one reason why Greenberg () included the Koman languages in his NS phylum was the attestation of these forms for first and second person singular in these languages, which match up with those above. The actual forms for the third person singular pronoun vary considerably across NS but a third person subject marker kV- (whereby the vowel tends to be ɔ) is quite common (Greenberg : –). To this, one more potential cognate may be added, a logophoric third person singular pronoun ε, as in example () from Kabba above. (Greenberg : suggests an original form e.) In Eastern Nilotic languages, this has become the regular third person subject pronoun on verbs. The discussion of cognate personal pronoun markers here is far from exhaustive, and is simply restricted to the least controversial cases. For a more extensive discussion, the reader is referred to the extensive discussion in Ehret (: –).
.. A stable preposition Whereas postpositions and postnominal modifiers (or postpositional nouns) are common across NS, there is one prepositional element which is attested both in CS and in Northeastern NS, thereby manifesting a remarkable historical stability. Thus, in a typical Northeastern language like Fur, where there are several postpositions specifying the search domain for some object as well as enclitical case markers (Waag ), there is one preposition commonly used to express semantic roles such as manner or instrument, kɪ. ()
kí ʒərri ‘quickly’ running
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
This situation has a parallel in CS languages. For example, in Kabba the only non-derived preposition is kə (Moser : ). In the Bongo-Bagirmi group it is also used as an infinitive marker, as in Sara Mbai kɪ-sa ‘to eat’, k-au ‘to go’.4 The same marker probably constituted the basis for what has come to be known as the “movable-k” element in NS, occurring “basically on nominal constructions in an apparently capricious manner” (Greenberg : ). Contrary to Greenberg, however, the present author assumes that this element was not derived from a demonstrative (i.e. a so-called Stage III article), but rather from the preposition *kɪ. This preposition is used in many modern Nilo-Saharan languages as an element introducing attributive phrases with nouns. Such noun phrases expressing particular features involving appearance or behavioral patterns are often headless (i.e. no head noun occurs). This strategy is a common device for the derivation of names for natural phenomena or human beings, for example in Nilotic languages like Nandi, as shown by the numerous examples in the dictionary by Creider and Creider (). The prefix -p- in such constructions is an archaic genitive marker for intimate or inalienable possession, which is also attested in Western Nilotic (Storch : , and passim): () keny ‘long ago’ kaliang ‘flies’
ki-p-keny ‘old-timer’ kipkaliang’it ‘fly whisk’
This strategy is also very common with personal names in Nandi, which often reflect the situation under which the person was born, e.g . Ki-p-lagat (after lagat ‘evening’) for a boy born in the evening. A further reflex of this original preposition probably occurs as a verbal extension expressing different but related semantic roles, such as the dative marker -ke in Ik (Heine ), whereby the raised vowel indicates a vowel which is devoiced before pause. Rottland (: ) reconstructs a presumably cognate suffix *-cɪ(:) for Southern Nilotic. The applied suffix on verbs -kə in Saharan languages like Kanuri (Hutchison : –) probably is a further reflex of this Proto-NS preposition. A presumably cognate preposition ki/gi is found in different Koman languages as well, e.g. ki in Uduk or gi in Kwama. According to Griscom (), there are three possible explanations for this property shared, for example, with neighboring Nilotic languages: chance, inheritance, or contact-induced change. Several Koman languages are currently being investigated in more detail by a number of scholars. Detailed grammatical descriptions should help to clarify the genetic status of these languages spoken in a residual zone in the border area between Ethiopia and Sudan, enclosed by expansion zones such as Nilotic, Surmic, and Eastern Cushitic. A second widespread preposition (usually realized as a proclitic or prefix), ka, which covers a locative meaning, is well attested in Northwestern NS, e.g. Kanuri ka-. Like the preposition kɪ, it is also used to introduce (headless) attributive noun phrases, as in the Turkana examples above. This Nilotic language still uses this preposition to introduce adverbs of place, as in ka neeni ‘from here’. See also the description by Storch (: –, and passim) of this element as well as other prefixal elements in Western Nilotic, e.g. in Luwo kà-r-níínɔ̀ ‘bed, 4
See also the discussion in Ehret (: –).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
sleeping place’. Greenberg (: –) mentions a k- nominal derivative in this respect, introducing infinitives as well as agent and place of action, and gives examples from CS languages like Sara oji ‘to smoke’, kweji ‘tobacco’ or Mangbetu abuo ‘to advise’, ukubo ‘advice’. If the forms in CS are indeed cognate with those in Northeastern NS, the preposition *ka is also reconstructable for the earliest stages of NS.
.. Postpositions and case marking in Northeastern NS Postpositions or postnominal modifiers derived from nouns (in particular body part terminology) are common in CS and Northeastern NS, even in verb-initial Surmic languages like Tennet, as illustrated by the following examples: ()
tswέ tree kéét-á tree-
dri ‘on the top of the tree’ (Moru) head táddên-a ‘from up in the tree’ (Tennet) up-
In addition, however, CS languages have postpositions not derived from other word classes whose semantic role is also more abstract or “grammatical” in nature. A number of these appear to correspond to case markers in Northeastern NS. These are not analyzed as case markers in CS because core syntactic functions like subject and object do not co-occur with such markers. It is not clear what the etymological source of the accusative case marker (*ka/*ga) in Northeastern NS is (see Table .). In CS languages like Keliko, Lugbara, or Madi, there is a postposition ga expressing semantic roles such as locative (Tucker : ). But it is not clear whether this is the actual etymological source.5
Table . The accusative case marker in Nilo-Saharan Saharan
KanuriTubu
ga, aga
Nubian
Ghulfan Kadaru Nobiin
ga ka
Kuliak
5
Ik
-gi
a
-k
Fur
(-gɪ)
Kunama
(-k-)
There is a common tendency cross-linguistically to reinterpret more peripheral case markers as markers of core case especially when referring to human (or animate) entities. Hayward and Tsuge () present such a case for Omotic, where the South Omotic dative/benefactive case marker *-n in North Omotic corresponds to the South Omotic accusative case marker with the same form, the latter function being an innovation.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
It is not clear either where the accusative markers, such as Nyimang -ʊŋ, Tama -ɪŋ, or Maba -gu are direct reflexes or whether their forms resulted from the fusion with definiteness markers. In Kunama and Fur, we find reflexes of the original accusative marker only in a restricted environment. The common accusative marker for syntactic objects in these two Northeastern NS branches is -si, which is attested as a peripheral case marker elsewhere in Northeastern NS. A possibly cognate postposition si (expressing several more peripheral semantic roles such as instrument or source) is attested in CS languages again belonging to the Moru-Madi cluster (see Tucker : ). If additional shared innovations for Kunama and Fur can be found, this would point towards a closer genetic relationship between them than between other Northeastern NS groups. A further element attested in CS with cognates functioning as case markers in Northeastern NS is a “genitive in n” in Greenberg (: ). Ehret (: ) reconstructs an associative marker *-(ε:)ni for an early stage in NS called “Sudanic” in his subclassification. It may be related to or have the same origin as a nominal genitive suffix reconstructed as *-na by Ehret (: ), with possible reflexes in an alveolar nasal in Maban, Nubian, and Surmic. In CS as well as many Northeastern NS languages the order of genitive constructions tends to be nomen rectum plus nomen regens, i.e., the genitive marker is attached to the possessor, which precedes the possessed: ()
tí ni cow te-n cow-
zè ‘cow dung’ (Madi; CS) ɔl ‘cow’s teats’ (Kadaru; Nubian) teats
Other grammatical markers such as the locative (or directional) case marker reconstructed as *-tha/*-th̪ a by Ehret (: ), with reflexes such as Kunama -tta ‘to(wards), until’, Tama -ta ‘at, to’, or Berta -Ɂa ‘at, to, in’ do not appear to be related to or derived from any obvious cognate postposition in CS. For a more detailed discussion of case markers in NS, the reader is referred to Ehret (: –) and Dimmendaal et al. (). This typological distinction between CS and Northeastern NS corresponds to a difference in constituent order. Whereas CS languages tend to manifest an alternation between S–Aux–O–V and S–V–O depending on tense–aspect marking in a clause, or S–V–O as a basic constituent order, Northeastern NS languages tend to be verb-final; within the latter sub-branch Southern members of the ES branch deviate from this pattern, in that they have a verb-initial or verb-second structure coinciding with a reduced case-marking system. If case marking in Northeastern NS emerged through areal contact with AA languages (or an AA type of language), as suggested in Dimmendaal (), one may also expect some neighboring CS languages to have been affected by this areal diffusion. At least one such language, Sinyar (spoken in Chad and Sudan) has been identified by Boyeldieu (), who shows that this CS language has a marked nominative case-marking system, i.e., it marks subjects rather than objects for case. This system is also attested in two Southern members of the ES branch in NS, Nilotic and Surmic. However, in these branches this system only applies to post-verbal (as against preverbal) subjects, whereas in Sinyar subjects precede the verb. As further pointed out by Boyeldieu (: ), “the social and
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
historical genesis of Sinyar still represents an open question in the present state of research.” Whereas case marking appears to have been lost altogether in Temeinian as well as the Daju cluster, it is still prominent in Gaahmg (Jebel), Surmic, and to a lesser extent in Nilotic. The actual subject case markers in members of these branches appear to go back to an original instrumental case marker *-ε and a genitive case marker *-i (Dimmendaal a). The innovation of these latter case markers coincides with a typological shift in the Southern ES languages towards a head-marking strategy by way of verbal extensions, and a corresponding reduction in case marking. Surmic and some Nilotic languages still have a number of peripheral case markers, but not as extensive as their Northern relatives within ES. The extensive case marking of Northeastern NS branches such as Maban, Fur, Saharan, or Northern ES corresponds to a verb-final structure, the use of converb plus light verb constructions and converbs in narrative discourse as well as extensive number-marking systems in all these branches. Again, these typological properties are also attested in verbfinal AA languages in Ethiopia, more specifically in Cushitic and Omotic (Dimmendaal ). The converb with the widest distribution across Northeastern NS is a verb ‘say’, probably reconstructable as *ni, with reflexes in Saharan (-n), Taman (-n), Berta (-n), and Nilotic ni, where this marker frequently introduces ideophonic adverbs.
.. Some further grammatical cognates CS languages have a few plural-marking number suffixes, often used with human or animate nouns only. This contrasts dramatically with number suffixes in Northeastern NS, where it is common to have a three-way distinction between singulatives, plurals, and replacement (Dimmendaal ). This latter system constitutes another typological property shared with AA languages in Ethiopia. In spite of this typological divergence, there appear to be a number of cognate plural suffixes going back to the earliest stages of NS, namely -k and -i, as again pointed out by Greenberg (: ); see also Ehret (: –) for a discussion of number inflection. () singular lέdr bʊ̀r-ti wel
plural lέdr-ì ‘person’ (Moru) bʊ́r-í ‘road’ (Masalit) wel-i ‘dog’ (Dongolese Nubian)
Other plural markers, such as -n or -t plurals and singulative suffixes in -t (as in Masalit, a member of the Maban group, above), probably do not go back to the earliest stages of NS, but instead appear to have been innovated in Northeastern NS. Among the additional set of grammatical properties listed by Greenberg (: , ) indicative of a common genetic origin for a range of languages “sandwiched” between two other major phyla, AA and Niger-Congo, is a negation marker in k; see also Ehret (: ff.) for a discussion. Given its mobile syntactic position in CS and Surmic and its behavior as an auxiliary verb, this element probably was verbal in origin (meaning ‘lack’). An example from Moru (adapted from Tucker and Bryan : ):
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
() má-nyá ŋgá kʊ . -eat thing ‘I did not eat anything.’ má kʊ̀ ŋgá ɔ̀ -nya. thing -eat ‘I am not eating anything.’ The marker contains a high back vowel in distantly related members like CS (Moru kʊ), or Surmic, where kʊ/ku is still inflected as a verb in languages like Majang (example adapted from Unseth : ). The initial position for this verb reflects a more general typological property of Majang and other Surmic (or Nilotic) languages, i.e. verb-initial constituent order. ()
ku-u-ko bokaye ɗεpey ɓokot-it. -- Bokaye lion kill- ‘Bokaye did not kill a lion.’
Reflexes with alternative vowels (such as Nara ka-) presumably result from the fusion with pronominal suffixes or enclitics. In addition, there is a widespread negation marker ɓa/ba, which appears to have been used in order to express existential negation (‘there is/are not’). Reflexes are found in CS (e.g. Mangbetu ɓa), Saharan (Kanuri -ba), Fur (where it is part of a double negation strategy a . . . ba), as well as in ES (e.g. ba in the Daju language Nyalgulgule). For further observations on negation markers in NS, the reader is referred to Ehret (: –). One of the best attested cognates in the domain of question words is the lexeme for ‘who’, with cognates across the family, as shown in Table .:
Table . The question word ‘who’ in Nilo-Saharan Maban
Maba
Kunama Eastern Sudanic
Central Sudanic
nya na
Taman
na
Nara
na
Nubian, Nobiin
na
Berta
na(n), nanne
Kaba
ná
Additional cognates can be found with the relative clause pronoun in other NS branches, for example in the ES (Jebel) language Gaahmg, na, or in Eastern Nilotic, where the feminine relative pronoun is na (the majority of nouns having feminine gender in this Nilotic branch). The invariable forms thus most likely go back to a Proto-Nilo-Saharan form *na.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
. S
.................................................................................................................................. The historical-comparative study of NS is still in its infancy. The picture emerging from our current understanding of this phylum, characterized by tremendous typological and genetic diversification, is a phylum whose proto-language was closer in terms of its morphological and syntactic structure to CS languages. Case marking, extensive number marking for nouns, or the use of converbs as a strategy in narrative discourse, for example, cannot be reconstructed for its earliest stages. Instead, they appear to have been acquired in Northeastern NS through areal contact with AA (or “Afro-Asiatic type” languages). CS languages have been innovating phonologically in that members of this primary branch have undergone segmental reduction lexically and grammatically (*-C(C)VC > -C(C)V or -C), possibly an adaptation to their Ubangian neighbors; consonant sequences appear to have included prenasalized stops, which are still found as such in CS and Maban, for example. As shown by Boyeldieu et al. () in their comparative study of the Bongo-Bagirmi cluster within CS, verbal -C(C)V-roots are frequently preceded by CV- or V- derivational prefixes, as in the Gula Mere reflex of the widespread verbal root for ‘eat’, ɔ-ɲɔ, below. Nominal roots again ending in open syllables in CS also correspond to forms frequently ending in a consonant in several (presumably) cognate forms in Northeastern NS, although individual languages in this branch sometimes have last final obstruents too (as in Fur below). For reasons of space, it is not possible to discuss these correspondences in more detail here. The following two examples thus serve as an illustration of the nature of these correspondences. ()
‘eat’ Moru -nya Lendu -nyo Fur aam Kalenjin am Turkana -ɲam
‘mouth’ Moru ti Lendu tso Fur udo Temein it ̪uk Turkana -kɪ-tʊk
In spite of the tremendous time-depth presumably involved for this phylum, there are several widespread lexical cognates next to the inflectional morphemes discussed above. However, the number of convincing lexical cognates is not enough to be able to reconstruct a complete phonological system. Given the fact that the phonological distinction between voiced implosive, voiced plosive, and voiceless implosives is widespread across NS, this contrast probably goes back to its earliest stages. The ejectives found in Surmic languages and the lack of voicing distinctions for stops in Southern Nilotic probably are due to contact with neighboring languages belonging to AA and Bantu, respectively. Whereas [ATR]-harmony (involving a set of nine or ten vowels) is widespread across the phylum, it should be kept in mind that such prosodic features easily spread across areas (regardless of whether such languages are genetically related or not). The picture emerging from grammatical forms where the vowels have remained virtually unaltered is an original system with the following vowels: *i, *ɪ, *e, *ε, *a, *ɔ, *o, *ʊ, *u.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
A brief survey such as the present, focusing on inflectional properties involving presumably cognate morphemes, does not do justice to the rich and fascinating properties of the more than a hundred languages belonging to the NS phylum. Widespread derivational strategies are discussed in Dimmendaal (b). Areal contact clearly played a role in the historical development of different NS branches, and is discussed in more detail in Dimmendaal (). Since the “center of gravity” for NS, i.e. the area with the highest degree of genetic diversity, is found in the border area between Chad, Sudan, and the Central African Republic, this may well have been the original homeland of this phylum. The presumed spreading of Northeastern NS from this area possibly was related to climatological changes in the area, more specifically the emergence of a major riverine system (the Yellow Nile or Wadi Howar) some , years ago, as well as cultural innovations, namely the introduction of pastoralism around . Evidence for this scenario, which is supported by the spreading of specific archaeological traditions and osteological investigations of human remains, as well as research on DNA structures among different populations in the area, is discussed in more detail in Dimmendaal et al. (). The disappearance of the Yellow Nile around appears to have resulted in the diffusion of different groups speaking NS languages (as argued in Dimmendaal et al. ). Whereas some NS groups moved into the Nuba Mountains in Sudan, which became a residual zone as a result of the migration of speakers of Niger-Congo and Kadu languages, other groups (Nilotic and Surmic speakers) moved southward into South Sudan and neighboring countries, thereby creating new expansion zones, which only came to a halt during colonial times in eastern Africa.
R Bender, M. L. (). The Nilo-Saharan Languages. Munich: LINCOM Europa. Blackings, M., and Fabb, N. (). A Grammar of Ma’di. Berlin and New York: Mouton de Gruyter. Boyeldieu, P. (). Case alignment(s) in Sinyar, in A. Mietzner and A. Storch (eds.), Nilo-SaharanModels and Descriptions. Cologne: Rüdiger Köppe, –. Boyeldieu, P., Nougayrol, P., and Palayer, P. (). Lexique comparative historique des langues sarabongo-baguirmiennes. Paris: CNRS-LLACAN. Crazzolara, J. P. (). A Study of the Logbara (Ma’di) Language. London: Oxford University Press. Creider, J. T., and Creider, C. (). A Dictionary of the Nandi Language. Cologne: Rüdiger Köppe. Croft, W. (). ‘Editor’s introduction’, in W. Croft (ed.), Joseph H. Greenberg—Genetic Linguistics: Essays on theory and method. Oxford: Oxford University Press, xiii–xxxvi. Dimmendaal, G. J. (). ‘The two morphological word classes in Nilotic’, in R. Vossen and M. Bechhaus-Gerst (eds.), Nilotic Studies: Proceedings of the International Symposium on Languages and History of Nilotic Peoples, Cologne, January –, . Vol. . Berlin: Dietrich Reimer, –. Dimmendaal, G. J. (). ‘Number marking and noun categorization in Nilo-Saharan languages’, Anthropological Linguistics : –. Dimmendaal, G. J. (). ‘Eastern Sudanic and the Wadi Howar and Wadi Milk diaspora’, Sprache und Geschichte in Afrika : –. Dimmendaal, G. J. (). ‘Africa’s verb-final languages’, in B. Heine and D. Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Dimmendaal, G. J. (a). ‘Marked nominative systems in Eastern Sudanic and their historical origin’, Afrikanistik Aegyptologie -Online , https://www.afrikanistik-aegyptologie-online.de/ archiv// accessed July , . Dimmendaal, G. J. (b). ‘Derivation in Nilo-Saharan’, in R. Lieber and P. Štekauer (eds.), Oxford Handbook of Derivation. Oxford: Oxford University Press, –. Dimmendaal, G. J. (). ‘Nilo-Saharan’, in R. Hackley (ed.), Oxford Handbook of Areal Linguistics. Oxford: Oxford University Press, –. Dimmendaal, G. J., with Ahland, C., Jakobi, A., and Kutsch Lojenga, C. (). ‘Linguistic features and typologies in languages commonly referred to as “Nilo-Saharan” ’, in H. E. Wolff (ed.), The Cambridge Handbook of African Linguistics. Cambridge: Cambridge University Press, –. Ehret, C. (). A Historical-Comparative Reconstruction of Nilo-Saharan. Cologne: Rüdiger Köppe. Gaudefroy-Demombynes, M. (). ‘Documents sur les langues de l’Oubangui-Chari’, in Actes du XIV congrès international des orientalistes. Paris: E. Leroux. Greenberg, J. H. (). Studies in African Linguistic Classification. New Haven, CT: Compass Publishing. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Greenberg, J. H. (). ‘Nilo-Saharan and Meroitic’, in T. A. Sebeok (ed.), Current Trends in Linguistics. Vol. : Linguistics in Sub-Saharan Africa. The Hague: Mouton, –. Greenberg, J. H. (). ‘Nilo-Saharan movable-k as a stage III article (with a Penutian typological parallel)’, Journal of African Languages and Linguistics : –. Griscom, R. (). KI constructions in the Luo and Koman families: contact or inheritance? Paper presented at the th Nilo-Saharan Linguistics Colloquium, University of Cologne. Hayward, R. J., and Tsuge, Y. (). ‘Concerning case in Omotic’, Afrika und Übersee : –. Heine, B. (). The Kuliak Languages of Eastern Uganda. Nairobi: East African Publishing House. Heine, B. (). ‘The dative in Ik and Kanuri’, in W. Croft and S. Kemmer (eds.), Studies in Typology and Diachrony: Papers presented to Joseph H. Greenberg on his th birthday. Amsterdam and Philadelphia, PA: John Benjamins, –. Hutchison, J. P. (). The Kanuri Language: A reference grammar. Madison, WI: African Studies Program, University of Wisconsin. Lokonobei, L. L., and de Jong, N. (). ‘On the position of Boya in relation to Murle and Didinga’, Occasional Papers in the Study of Sudanese Languages : –. Moser, R. (). Kabba: A Nilo-Saharan language of the Central African Republic. Munich: LINCOM Europa. Newman, P. (). On Being Right: Greenberg’s African Linguistic Classification and the methodological principles which underlie it. Bloomington, IN: Institute for the Study of Nigerian Languages and Cultures, African Studies Program, Indiana University. Nicolaï, R. (). La force des choses ou l’épreuve ‘nilo-saharienne’: Questions sur les reconstructions archéologiques et l’évolution des langues. Cologne: Rüdiger Köppe. Rilly, C. (). Le méroïtique et sa famille linguistique. Leuven and Paris: Peeters. Rottland, F. (). Die südnilotischen Sprachen: Beschreibung, Vergleichung und Rekonstruktion. Berlin: Dietrich Reimer. Sands, B. (). ‘Africa’s linguistic diversity’, Language and Linguistics Compass : –. Starostin, G. (). ‘On Mimi’, Journal of Language Relationship : –. Stirtz, T. M. (). A Grammar of Gaahmg, a Nilo-Saharan language of Sudan. Utrecht: LOT. Storch, A. (). The Noun Morphology of Western Nilotic. Cologne: Rüdiger Köppe. Toweett, T. (). A Study of Kalenjin Linguistics. Nairobi: Kenya Literature Bureau. Trigger, B. G. (). ‘Meroitic and Eastern Sudanic: a linguistic relationship?’, Journal of African History : –. Tucker, A. N. ( []). The Eastern Sudanic Languages. Vol. . London: Dawsons of Pall Mall for the International African Institute.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Tucker, A. N., and Bryan, M. A. (). Linguistic Analyses: The Non-Bantu languages of northeastern Africa. London, New York, and Cape Town: Oxford University Press. Tucker, A. N., and ole Mpaayei, J. T. (). A Maasai Grammar, with vocabulary. London, New York, and Toronto: Longmans, Green and Co. Unseth, P. (). ‘Word order shift in negative sentences of Surma languages’, Afrikanistische Arbeitspapiere : –. Unseth, P. (). ‘Two old causative affixes in Surmic languages’, in G. J. Dimmendaal and M. Last (eds.), Surmic Languages and Cultures. Cologne: Rüdiger Köppe, –. Waag, C. (). The Fur Verb and its Context. Cologne: Rüdiger Köppe. Weiss, D. (). Phonologie et morphosyntaxe du maba. Doctoral dissertation. Université Lumière Lyon .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T Saharan language family consists of a small number of languages located in the central region of the Sahara, the area around Lake Chad and parts of Darfur in western Sudan and the adjacent area in Chad (see Map .). They are Teda-Daza (Tubu), Kanuri (including Kanembu), and Beria (Zaghawa). The Berti language, a close relative of Beria, became extinct at the beginning of the twentieth century.
. S
.................................................................................................................................. The earliest indication about a relationship between the languages, which are known today as Saharan, was given by Heinrich Barth, a German traveler to Central Africa (Barth –). He recognized the close relation between Teda-Daza in the central Sahara and Kanuri-Kanembu around Lake Chad. Another traveler, Gustav Nachtigal, made the family complete and included the Beria language, also known by the Arabic exonym Zaghawa. He referred to this family as the Tubu-Kanuri-Baele group (Nachtigal : ). Johannes Lukas supported the former findings and called the family East Saharan (Lukas /). In Tucker and Bryan () it was renamed Central Saharan. Joseph Greenberg () reduced the name to Saharan. In principle this classification was never challenged in the past years. Later attempts geared at the search for higher genetic units, or phyla, which would integrate the Saharan family. Lukas (/) did not make any proposals about higher genealogical levels. Tucker and Bryan () treated the family as an “Isolated Unit”. After several decades of attempts to reshape the genealogical landscape of African languages, Greenberg proposed the Nilo-Saharan phylum, of which Saharan languages formed a branch. Attempts in recent decades to relocate the Saharan languages to other phyla did not find much support. Most controversies were fought on the validity of a Nilo-Saharan phylum and on the internal structure of classification. Among the Saharan languages Kanuri has the longest tradition of documentation. The first comprehensive grammatical description by Sigismund Wilhelm Koelle appeared in (Koelle a). At the same time an extensive Kanuri text collection appeared
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
L I B YA Kanuri Teda
Teda Teda Kanuri
NIGER
SUDAN
Daza Daza
Zaghawa (Beria)
Kanembu Kanuri
Lake Chad CHAD
. The geographical distribution of Saharan languages
(Koelle b). The grammar shows a high level of linguistic competence, according to the state of knowledge at that time. Lukas published his Kanuri grammar in . It was the most important reference work for the decades that followed. From the s the number of publications on Kanuri increased considerably. For a long time, knowledge about the grammar of Teda-Daza (“Tubu”) still depended on sketchy grammatical descriptions by Lukas () and Charles and Marguerite Le Cœur (). In the meantime, however, this research gap could be closed by the grammar of the Daza language by Josiah K. Walters (). The first comprehensive grammar of Beria (Zaghawa) was published by Jakobi and Crass (). Earlier works by Karel Petráček supported membership in the Saharan family (Petráček , , ).
. D S
.................................................................................................................................. Map . above roughly illustrates the distribution of Saharan languages in the area. According to Nachtigal (), the origin of the family can be assumed to be in the central Sahara in the Tibesti mountain region. From here the separation of the Saharan languages took place. The Beria became established in their present habitat in the southeast. Another group moved to the south and settled in the Kanem Province in Chad. From there the area of Borno was settled and the influential dynasty of Kanem-Borno evolved. The political center was shifted to Borno, to the west of Lake Chad. In the ensuing period Kanuri became the first or second language in the oases of Bilma and Fachi in Niger as well as in some places in Fezzan (Libya). More recently, the Kanuri language, like other languages, became established in parts of Sudan as a result of the pilgrim route to Mecca (see Figure .).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Proto-Saharan
Kanuri/Kanembu
Teda-Daza (Tubu)
Beria (Zaghawa)/Berti
. Classification of Saharan languages
. C S
.................................................................................................................................. A common Saharan vocabulary is apparent in the basic vocabulary, though the number of cognates between the individual languages varies considerably.1 ‘mouth’ ‘tongue’ ‘ear’ ‘liver’ ‘knee’ ‘person’ ‘leaf ’ ‘big’ ‘that’ ‘to die’ ‘to come’ ‘to see’ ‘to drink’ ‘to say’
Kanuri cî tə́lam sə́mo kəmáttən ngurumngurum âm () kálú kúra túdu nú ís rú yá, sá n
Teda-Daza kai tirmẽś u šímo maasen amo kólú kɔra te̥ye nus ri ru ya n
Beria āā tàmsī màī kórú ɔɔ̄ ̄ ɔ́gύr ύgύrī tɔ̄ nύí tíí ír̥ì yá n
Proto-Saharan *kai *tiram *simo *masin *kurum *am *kur *kut *tu *nu *it *tu *sa *n
Most Proto-Saharan reconstructions have been proposed by Ari Awagana (). As the phonological and tonal analysis requires further investigation in Teda-Daza, the analyses should be considered tentative. The lexical coherence in the Saharan family varies. As a general observation it can be stated that the largest coherence exists between Teda-Daza and Kanuri, and the lowest between Kanuri and Beria, as shown in Table ..
1 In Kanuri the following tone marking is applied: á high tone, a low tone, â falling tone. For Beria the convention used by Jakobi and Crass () is applied: á high, à low, ā mid. Tone in Teda-Daza is unmarked because of unreliable available data.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Rates of cognates in Saharan languages (verbs and non-verbs) Teda-Daza
Kanuri
Teda-Daza
Beria
60%
Kanuri
50%
Beria
40%
20% 20%
Verbs
30% Non-verbs
These approximate figures of lexical agreement also confirm the hypothesis already mentioned by Nachtigal () about coherence.
.. Common Saharan person elements All Saharan languages share identical underlying pronominal morphemes (Cyffer a). In Table . they are illustrated as reconstructed forms.
Table . Underlying pronominal morphemes in Saharan languages Pronoun
Subject
Object
Possessive Bound form
Teda-Daza
Kanuri
1sg
T
T
2sg
N
N?
B K
3sg
T
T
1pl
T
T
2pl
N
N?
B
3pl
T
T
K
T
K
K
1sg 2sg
N
3sg
T
T
1pl
T
T B
2pl 3pl Beria
N
T
1sg
T
2sg
N
3sg T
2pl
N
3pl
T N T
N
N?
T
T
T
N
N
N
T
T
T
T
N
B T
K N
T
B K
B K
N
N
K
K
T
T
T
N
N
K
K
T
K
N
T K
B
T
T B
B
T T
T
K
1pl
T N
Free form
N K
K
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
In a condensed presentation these pronominal forms can be summarized as follows: Person Series A Series B Other 1 T1 K1 2 N B 3 T2 K2 Ø noun, e.g. ‘person’
. C S
.................................................................................................................................. One of the most striking attestations for the validity of the Saharan family is the occurrence of verb classes which are based on similar formatives and morphological arrangement. All Saharan languages have or had three verb classes. Verb classes I and II differ in their morphological setup. Class III applies an additional formative. This also means that the common verbal system was in existence in a Proto-Saharan period. Class I verbs occur today in Teda-Daza and Beria. They are more or less extinct in Kanuri. However, traces of this class are still present. Its basic characteristic features are prefixation of subject markers and zero marking of the third person subject (see Table .).
Table . Verb class I in Saharan languages Class I
Teda-Daza ər ‘come’
Basic structure
Beria má ‘ripen’
sj
vr
sj
vr
sj
vr
1sg
+
+
t
ər
i
á
má
rí
2sg
+
+
n
ər
i
ná
má
rí
3sg
Ø
+
Ø
ər
i
Ø
má
rí
1pl
+
+
t
ər
du
tá
má
rí
2pl
+
+
n
ər
du
ná
má
ri
3pl
Ø
+
Ø
ər
du
Ø
mâ
ri
In Kanuri all verbs which can be traced as original class I verbs are intransitive. However, in Beria and Teda-Daza there are also some transitive verbs in that class. Another possible argument that classes I and II were originally distinguished by the transitive and intransitive distinction is the fact that class II verbs outnumber class I verbs by far.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Verb class II in Saharan languages Class II
Basic structure sj
Teda-Daza ruk ‘go out’
sj
sj
sj pl vr sj tam
+
+
ruk
r
lad sk
in
óyá: g
rí
2sg
+
+
ruk
m
lad m
in
óyá: n
rí
lad
in
lad ye
in
+
2pl 3pl
+
+
k
+
+
+
+
+
pl
sj vr sj tam
vr
3sg
vr
Beria óyá: ‘speak’
1sg
1pl
sj
Kanuri lad ‘sell’
ruk
k
s
ruk
t
r
ruk
t
m
ruk
t
lad w s
a lad
in in
k óyá: óyá: d óyá: b k óya:
rí rí rí rí
Class II verbs differ structurally from those of class I. Whether the original semantic difference is due to the transitive intransitive distinction is yet to be established (see Table .). Those verbs in Kanuri which originally belonged in Kanuri class I were either integrated into class II, still showing some class I features, or were replaced by new class III verbs. In all Saharan languages most verbs belong to class III (see Table .). On the one hand, this class is old enough to belong to each language; on the other hand, it is highly productive. For example, in Kanuri more than % of the verbs belong to class III. With the help of an auxiliary verb n ‘say, think’, almost all nouns, adjectives, and ideophones can be verbalized (see Table .). This auxiliary verb itself can also occur as an independent verb of class II. Therefore, the possible number of verbs is unlimited.
Table . Verb class III formative n ‘to say’ Kanuri sj
pl
1sg 2sg 3sg
s
vr
sj
tam
‘say’
n
sk
in
ngin
‘I say’
n
m
in
nəmin
‘you say’
in
shin
‘he/she says’
Ø
1pl
n
ye
in
nyen
‘we say’
2pl
n
w
in
nuwin
‘you say’
in
sai(n)
‘they say’
3pl
s
a
Ø
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Verb class III in Saharan languages Class III
Basic structure
Teda-Daza wa-n ‘refuse’
mc sj aux sj mc sj aux pl sj +
+ wa
n
sk
in
nɔ :́
Ø
g rí
wá
n
m
in
nɔ :́
Ø
n rí
wá s
Ø
in
nɔ :́ gé
n
rí
in
nɔ :́
Ø
d rí
in
nɔ́:
Ø
n rí
nɔ :́ gê
n
rí
+
2sg
+
3sg
+
1pl
+
+
+ wa
n
t
r
wá
n
ye
2pl
+
+
+ wa
n
t m
wá
n
w
3pl
+
n
t
wá s
in
+
+ wa
mc sj pl aux sj tam mc sj aux sj tam
1sg
+
+ +
wa ki
+
wa ki
Beria nɔ ́:-n ‘see’
Kanuri wá-n ‘refuse’
n
r
wá
n
m
Ø
a
Ø
The derived word takes the function of the meaning carrier. Examples from Kanuri: Basic form
Verbalized form
kúra Adj ‘big’ kuranyéna ‘we have become important’ ngáwo N ‘back’ ngawozána ‘they have moved back’ badak Ideo ‘describes beating [karə́gənzə]́ badakcîn ‘[his heart] is beating’ of heart’
.. Plurality Saharan languages share features in the formation of plurality, e.g. in nouns, pronouns, and verbs (Cyffer b). The common formatives for plural marking are T and A, e.g.: Plural marker A Teda-Daza
Kanuri
Beria
‘tree’ ake aka ‘house’ fáto fatowá ‘monkey’ dεgil dεgila ‘boy, son’ táda tadawá ‘house’ jεgε jεga ‘bicycle’ báskur baskurwá
(pluralization by tonal contrast)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Pluractional A Teda-Daza Base Pluractional ‘to dye’ dil ‘to dig’ lu ‘to hang up’ lus
dal la las
Plural marker T Teda-Daza: Possessive pronouns ps 3ps 1 2 3
s
sg form 1/2ps n r n ər n m n əm n s ən
3ps
s
form n n n
pl pl t t t(V)
r m
1/2ps ntər ntəm sənto
Plural marker TA: Berti (Petráček ) ps 1 2 3
sg su nage kore
pl su-ta nage-ta kore-ta
Kanuri : lewóno ‘he/she went’ : leyá.da ‘they went’ : bákkam ‘you beat me’ : bákka.da.m ‘you beat us’
.. Common features shared by the Saharan languages Kanuri Teda-Daza Beria Lexical coherence Three verb classes Person marking Identical morphemes in and Different morphemes in and Marking plurality Intransitive/“passive” derived from verb Applied derived from verb Causative derived from verb Locative-imperfective correlation
+ +
+ +
+ +
(+) + + + + + +
(+) + + + + + +
+ + + + + +
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. S
.................................................................................................................................. The internal coherence of the Saharan languages is undoubted. Several common features of lexicon and grammar are attributed to the common genetic property of the family. It is likewise evident that the Saharan languages in their current form also share features with other languages of the contact area. This especially applies to the syntax of complex sentences. It can be assumed that contact features were adopted when the Saharan languages had separated into independent languages and formed individual categories through direct borrowing or calque processes. In addition, the elaboration of the TAM system also appears to be a result of contact.
R Awagana, A. (). ‘Racines lexicales sahariennes: préludes à la reconstruction du vocabulaire de base’, in D. Löhr, E. Rothmaler, and G. Ziegelmeyer (eds.), Kanuri, Borno and Beyond: Current studies on the Lake Chad region. Cologne: Rüdiger Köppe, –. Barth, H. (–). Collection of Vocabularies of Central African Languages. Gotha: J. Perthes [repr. , London: Frank Cass]. Cyffer, N. (a). ‘The person elements in Saharan languages: a step towards the creation of protoSaharan’, in T. C. Schadeberg and M. L. Bender (eds.), Nilo-Saharan: Proceedings of the First Nilo– Saharan Linguistics Colloquium, Leiden September –, . Dordrecht and Cinnaminson, NJ: Foris, –. Cyffer, N. (b). ‘Pluralization in Saharan languages’, Afrika und Übersee : –. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Jakobi, A., and Crass, J. (). Grammaire du beria (langue saharienne). Cologne: Rüdiger Köppe. Koelle, S. W. (a). Grammar of the Bornu or Kanuri Language. London: Church Missionary House. Koelle, S. W. (b). African Native Literature—or Proverbs, Tales, Fables, & Historical Fragments in the Kanuri or Bornu Language. London: Church Missionary Society [repr. , Graz: Akademische Druck- und Verlagsanstalt]. Le Cœur, C., and Le Cœur, M. (). Grammaire et textes teda-daza. Dakar: Mémoires de l’IFAN. Lukas, J. (). A Study of the Kanuri Language: Grammar and vocabulary. London: Oxford University Press [repr. , London: Dawsons of Pall Mall]. Lukas, J. (/). ‘Umrisse einer ostsaharanischen Sprachgruppe’, Afrika und Übersee : –. Lukas, J. (). Die Sprache der Tubu in der Zentralen Sahara. Berlin: Akademie-Verlag. Nachtigal, G. (). Sahara und Sudan. Vol. II. Berlin: Weidmannsche Buchhandlung und Verlagsbuchhandlung Paul Pare. [repr. , Graz: Akademische Druck- und Verlagsanstalt.] Petráček, K. (). ‘Phonologische Systeme der zentralsaharanischen Sprachen (Konsonantische Phoneme)’, Archiv Orientalni : –. Petráček, K. (). ‘Phonologische Systeme der zentralsaharanischen Sprachen (Vokalische Phoneme)’, in D. Cohen (ed.), Mélanges M. Cohen. The Hague: Mouton, –. Petráček, K. (). ‘Die Zahlwörtersysteme der zentralsaharischen Sprachen’, in V. Six, N. Cyffer, E. Wolff, L. Gerhardt, and H. Meyer-Bahlburg (eds.), Afrikanische Sprachen und Kulturen—ein Querschnitt. Hamburg: Deutsches Institut für Afrika-Forschung, –. Tucker, A. N., and Bryan, M. A. (). The Non-Bantu Languages of North-Eastern Africa. London: Oxford University Press. Walters, J. K. (). A Grammar of Dazaga (Leiden: Brill).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
.
. I
.................................................................................................................................. I his continent-wide classification of African languages, Greenberg () also established a genetic unit called “Eastern Sudanic”, which constituted one of the sixteen African language families identified by him at the time. Eastern Sudanic (ES) consisted of eight subgroups still recognized as valid branches of this family in more recent studies. In his classification, Greenberg added Nyangiyan and Temainian (classified as isolates in Greenberg ) as additional subgroups. Moreover, the subgroups referred to as Nilotic and Great Lakes were combined in a single branch, called Nilotic (following the pioneering work of Köhler ). Greenberg (: –) already argued that ES is part of a larger genetic grouping which includes Kunama, Central Sudanic (CS), and Berta, for which he proposed the name “Macrosudanic”. This family was renamed “Chari-Nile” in Greenberg (), where it was established as one of the six primary branches of a new phylum established by him, Nilo-Saharan (NS; see Dimmendaal, chapter of this volume). Table . presents the eleven subgroups constituting ES according to Greenberg (), and also gives their current names. As shown in Table ., Greenberg () also included the Kuliak languages (referred to as “Rub” by Ehret ) in his ES branch of NS, but he appears to have been misled by the numerous borrowings from Nilotic languages. Schrock () has shown that % of the lexicon in the Kuliak language Ik, for example, consists of borrowings from the Nilotic Teso-Turkana languages. As further pointed out in sections .–., several diagnostic features suggesting that the remaining ten subgroups form a valid subgrouping within NS called ES are absent from the Kuliak languages, which most likely form an independent, early offshoot within NS (see Dimmendaal, chapter of this volume). On the basis of the main consonantal exponents of the first person singular pronoun, Bender (b: ) argued for a division of Eastern (or East) Sudanic into two subfamilies:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Eastern Sudanic subgroups Greenberg ()
Greenberg ()
Current name
Nubian
Nubian
Nubian
Beir-Didinga
Murle, Longarim, Didinga, Suri, Surmic Mekan, Murzu, Surma, Masongo
Barea
Barea
Nara
Tabi
Ingassana
Gaahmg, a member of (Eastern) Jebel (the other three members are probably extinct)
Nyimang
Nyima, Afitti
Nyimang and Afitti
Merarit
Merarit, Tama, Sungor
Taman
Dagu
Daju Dagu of Darfur, Baygo, Sila, Dagu of Dar Dagu, Dagu of Western Kordofan, Njalgulgule, Shatt, Liguri
Nilotic
Western Nilotic
Nilotic
Great Lakes
Eastern Nilotic and Southern Nilotic Temainian (linguistic isolate) Temein, Teis-um-Danab
Temeinian
Nyangiyan (linguistic isolate) Nyangiya, Teuso
Kuliak (or Rub)
“Ek”, consisting of Nubian, Nera (Nara), Nyima (Nyimang), and Tama, where this pronoun is assumed to contain a velar stop, and “En”, consisting of Surma (Surmic), Jebel, Daju, Temainian, and Nilotic, where the first person singular pronoun is assumed to contain a nasal. Although not all language groups support this proposed diagnostic feature (the Taman first person pronoun, for example, is wa), Bender’s intuition on the subgrouping of ES appears to have been right. More recently, Rilly () also arrived at the conclusion that Nubian, Nara, Nyimang, and Afitti, as well as Tama(n), form a subgroup, called “Northern East Sudanic” by him. On the basis of an extensive comparison of the lexicon and morphological structure of these languages, he also concluded that the extinct Meroitic language belongs to the same subgroup within ES. The classification of the remaining ES subgroups, the Daju cluster, Temeinian, Nilotic, and Surmic, into a second sub-branch called “Southern Eastern Sudanic” is supported, for example, by the innovation of an irregular (suppletive) alternation for ‘cow’ (Dimmendaal : ): () Proto-Nilotic Majang (Surmic) Temein Daju of Lagowa
Singular *d̪εŋ taŋ ńtέŋ teɲe
Plural/Collective *d̪ʊk togi kítúk tukke
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
The evidence from Gaahmg (probably the only Jebel language still spoken today) is inconclusive in this respect because the alternation for ‘cow(s)’ is not suppletive: tɔɔ / tɔg. The plural (collective) form, tɔg, is cognate with the plural form in the other four subgroups (as shown in () above). Moreover, Gaahmg (Jebel) shares the innovation of two ergative case markers, *-ı/-i and *-ε/-e for post-verbal A-roles (subjects), a feature absent from Northern members of ES, which do not allow for post-verbal subjects. Exactly the same two ergative case markers show up in Tennet (Surmic) and Päri (Nilotic), as further discussed in section ... Consequently, Gaahmg is probably more closely related to the Southern than the Northern branch of ES. Recent research has shown that Berta is not an isolated member of Chari-Nile (as argued in Greenberg ), but instead forms a subgroup within ES with Gaahmg. The word for ‘cow’ in Berta (taaŋ) does not alternate, and hence does not provide evidence for the shared innovation discussed above. Post-verbal subjects in Berta are marked for (nominative) case, but as this inflection is expressed tonally, it cannot be shown to be a reflex of the widespread i-/-ε case marker found in the Southern branch of ES. But Bremer () shows on the basis of an extensive lexical as well as a grammatical comparison with Gaahmg, the only Jebel language for which extensive data are available, that the latter language forms a genetic unit with Berta. The subgrouping of the ES branch adhered to in the present study is shown in Figure ..1 Taman Meroitic Northern
Nubian Nara Nyimang plus Afitti
Eastern Sudanic
Berta Jebel Southern Daju Temeinian
Surmic Nilotic
. The subclassification of Eastern Sudanic
1 According to Ehret (: –), Berta and the (West) Jebel group form a subgroup, which in turn forms a genetic unit with “Kir” called “Kir-Abbaian”. The Kir group comprises Temeinian, Daju, Surmic, and Nilotic, but also Nyimang and Afitti (Dinik).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. P
.................................................................................................................................. From a typological perspective, the consonant inventories of the Northern branch within ES languages are either moderately small or average-sized. The prototypical inventory, as presented in Table ., is found in the Nile Nubian language Kenzi (adopted from Abdel-Hafiz ).
Table . The consonant inventory of Kenzi
voiced plosives
labial
alveolar
palatal
b
d
ɟ
voiceless plosives voiceless fricatives nasals
t
approximants
glottal
g k
f
s
ʃ
m
n
(ɲ)
liquids
velar
h ŋ
l, r w
j
With the exception of Midob (Nubian), where the voiceless labial [p] stop has phonemic status, there is a characteristic absence of [p] as a phoneme in Nubian, Taman, Nyima, and Nara. Due to borrowing from Arabic, which is the lingua franca in the area, the consonant inventory may also comprise the velar fricatives [x] and [ɣ], the pharyngeal [ħ] as well as the alveolar [z], the last being attested in Nile Nubian and Nara. Kenzi does not distinguish between dental and (post-)alveolar stops, but this contrast is found in Kordofan Nubian languages and Nyimang, as well as in the Southern branch of ES. Southern members of ES in general tend to have slightly richer consonant inventories, which include a set of glottalized (implosive) consonants, as in the Surmic language Ngaalam (Yigezu and Dimmendaal to appear) (see Table .).
Table . The consonants of Ngaalam bilabial
dental t̪
voiceless stops
p
voiced stops
b
implosives
ɓ
fricatives
f
alveolar
palatal
t
k
d
g
ɗ
ʄ
voiceless affricates
tʃ dʒ
nasals
m
n
ɲ
l
trill glides
r w
ɠ
ð
voiceless affricates laterals
velar
j
ŋ
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Implosive consonants occur also in other Surmic languages as well as in Daju (see Palayer and Yaya : for Daju of Eref), and are also reconstructed for Proto-Berta by Bremer (: ) as well as for Proto-Nilotic by Dimmendaal (). Within the Northern branch, they appear to be restricted to Tama (Dimmendaal ). A second type of glottalized consonants, ejectives, is restricted to Berta, and the Southeastern branch of Surmic (which forms a coordinate branch with Southwestern Surmic called “Southern Surmic”; this last in turn forms a coordinate branch with Northern Surmic, consisting of one language, Majang). Yigezu (: ) reconstructs ejectives (*t’, *tʃ’, *k’) for Proto-Southeastern Surmic and concludes that they “are archaic and old in the system, and [that they] are not the result of contact or borrowing from the neighbouring [Omotic] languages.” The following data are drawn from Yigezu (: ). () Chai tílā tʃɔrε kεrε
Mursi tila tʃɔ́rε kerre
Me’en t’ilá (< *t’ila) ‘food’ tʃ’ɔrε (< *tʃ ’ɔrε) ‘hair’ k’erdʒe (< *k’erdʒe) ‘horn’
Given the geographical vicinity of Berta and Surmic to Koman and B’aga (which may or may not be part of NS), the ejectives may be the result of unadapted borrowing, though this has not been investigated in detail yet. This phenomenon can also be observed in Western Nilotic (WNil) Alur, which has gɓ and kp as a result of extensive (unadapted) borrowing from neighboring CS languages (Dimmendaal ). ES languages frequently distinguish between l and two types of rhotics, as in the Kordofan Nubian languages, or Nyimang, which has l, r, and ɽ, as in bʊlaŋ ‘broad hoe’, bʊraŋ ‘adulteress’, and bʊɽaŋ ‘jackal’ (Stevenson : ). There are usually restrictions on the occurrence of liquids, for example in Nubian languages, where liquids are not admitted in word-initial position, or Nilotic languages like Maasai, where the contrast between r and ɽ is neutralized word-finally. Prenasalized consonants are rare in ES. The Higir dialect of Nara has a series of prenasalized plosives, mb, nd, ŋg, and ŋgw which are, however, only attested in initial position (Hayward ). In Afitti, the plosive mb is occasionally found in initial position, where it corresponds to Nyimang b or m, e.g. mbɪrr vs. bar ‘cow’, mbərr vs. mir ‘fire’ (Tucker and Bryan : ). Whereas the areal source for these clusters in Northern members of ES is unknown, Ubangian and CS, as well as Bantu languages, played an important role in the phonological restructuring of WNil languages (Dimmendaal ; Storch ). In parallel to several Northern members of ES (and other NS languages), Southern members tend to have nine or ten vowels with [ATR] harmony. This system is found in Kordofan Nubian, Taman, Temeinian, Nilotic, and Surmic. The Nubian language Tagle, spoken in the Nuba Mountains in Sudan, for instance, has a vowel inventory comprising a set of five [+ATR] vowels and a set of five [ ATR] vowels.2 Within phonological words vowels either have [+ATR] or [ ATR] features, the vowel of syllabic suffixes taking the
2 In their study of the Tagle phoneme system, Ibrahim and Huttenga () analyzed the vowel system as comprising nine vowels, four [ ATR] vowels and five [+ATR] vowels. However, more recently Gumma Ibrahim (p.c.) found that there are two sets, five [+ATR] vowels and five [ ATR] vowels.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
[ATR] feature of the root vowel. The Tagle plural suffix, for instance, has two allomorphs, -ɪ after a [ ATR] root vowel and -i after a [+ATR] vowel (Ibrahim and Huttenga ). () kʊ́l- ‘wells’ kól-ì ‘houses’ Apart from [ATR] harmony and height assimilation, there is also fronting harmony, i.e. vowel assimilation involving the position of the tongue blade, for example in the Nubian language Karko. The inventory comprises eight short vowels, i, e, ε, a, o, ɔ, u, ə, which are associated with a system of progressive vowel assimilation, i.e., syllabic suffixes have an underspecified vowel which “copies” the vowel quality of the root vowel, as illustrated by the plural suffix -Vnd, whose vowel quality depends on the preceding root vowel, e.g. ēbēnd ‘tail-’, ām-ānd ‘ram-’, and ōr-ōnd ‘head-’ (Jakobi and Hamdan ). In Gaahmg, which has six vowels, i, ε, a, ɔ, ə, u, suffix vowels regularly adapt in terms of height or rounding. Examples with the completive suffix on verbs from Stirtz (: ): () bὲὲ-sà kɔ́m-sɔ̄ cīg-sə̄ rùm-sū
‘tell, say’ ‘cut, chop’ ‘wear’ ‘clear path’
The Daju language of Eref (in Chad), for example, has five short and five long vowels (Palayer and Yaya : ). Five-vowel systems, i, e, a, o, u, are attested in the Nile Nubian languages and Nara; six-vowel systems (i, e, a, o, ə, u) occur in Afitti and the Darfur Nubian language Midob, whereas Southwestern Surmic languages and Nyimang have seven vowels (Rilly : ). In general, the geographically more peripheral members tend to have more reduced vowel systems, suggesting influence from other languages, although the areal source is not always obvious. The most dramatic vowel changes occurred in one group of WNil languages consisting of Dinka, Nuer, and Atuot. Andersen () shows that in Dinka the original WNil system with five [ ATR] and five [+ATR] vowels, developed into fourteen breathy and creaky voice vowels ([ ATR] *ɪ, *ε, *a, *ɔ, *ʊ > (creaky voice) i ̰, e ̰, ε̰, a ̰, ɔ̰, o̰, ṵ ; [+ATR] *i, *e, *ʌ, *o, *u > (breathy voice) i̤, e̤, ε̤, a̤, ɔ̤, o̤, ṳ. The areal source for this dramatic restructuring (also at the morphological level) is unknown. Most ES languages are tonal, with either two or three registers, which may be combined to form complex (contour) tones at the structural level. As shown by Tucker (: –) for the Nilotic language Luo, languages with only two tonemes, low and high, may have up to nine distinct tonal realizations phonetically (including downstep and upstep). A number of Daju languages as well as Nara appear to have pitch-accent systems, allowing for the assignment of only one high tone per word (see Hayward for a discussion of this principle in languages of Ethiopia). Bender (: –) provides an informative survey of phonological inventories for a range of ES languages, and also includes an extensive set of potential lexical cognates for this NS branch (Bender : –). These, together with the lexical comparison of Northern members of ES in Rilly () and reconstructions for ES sub-branches like Daju (Thelwall ), Nilotic (Dimmendaal ), Nubian (Jakobi to appear), Taman (Edgar ), or Surmic (Yigezu ) will form a solid basis for the establishment of regular sound correspondences in the near future, a project already initiated by Bender (: –).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
. M
..................................................................................................................................
.. Number marking on nouns The classical tripartite number marking for Northeastern NS languages comprises singulative suffixes on unmarked plural stems, plural (plurative) suffixes on unmarked singular stems, and pairs of singular and plural suffixes, i.e. a replacement pattern (Dimmendaal ). Within the Northern branch, this system is attested in Taman and Nubian. The following examples, drawn from the Nubian language Birgid (Thelwall ), illustrate this tripartite division. () Singular kur-ti ɑŋɑl wɑ:ɲ-di
Plural kur ‘man’ ɑŋɑl-si ‘elephant’ wɑ:ɲ-e ‘star’
As a variant, the replacive pattern may be realized solely by tonally contrasting singular/ plural stems, as in Birgid úr /ùr ‘head’ (Thelwall ). In the Kordofan Nubian languages, the tonal replacement pattern is restricted to kinship terms, for example, Karko ʈîj /ʈıˇj ‘maternal uncle’ (Jakobi and Hamdan ). Where singulative marking was lost, as in Nara or Tama, the replacement pattern was generalized (next to the inherited system of plural marking). Examples from Tama (Dimmendaal : ):3 () tààmʊ́-t tààmʊ́-k ‘Tama person’ kòy-ít kòy-é ‘belly’ kʊ́m-á kʊ̀m-ʊ́k ‘chicken’ The archaic tripartite system of number marking has been retained as a lexical strategy in the Southern branch of ES, more specifically in Nilotic, Surmic, and Temeinian, but it was lost in some of the geographically peripheral zones of ES, for example in Nile Nubian, Nyimang, and Afitti, or the Southern Lwoo group within WNil (Storch ). In one Lwoo language, Luo (spoken in Kenya, Tanzania, and Uganda) the former suffixation system (restricted to two plural markers) is gradually being replaced by a system of noun class prefixes as a result of intensive borrowing from neighboring Bantu languages and the reinterpretation of former lexical roots as prefixes (Dimmendaal : ; data from Tucker ). () mɪ-sʊmba wa-sumb-nî ‘bachelor’ ja-lúô jo-lúô ‘Luo person’ (< *jal/jol ‘guest, stranger’)
3 Tama uses t and k in various pairs of replacive suffixes. Bryan () was the first to observe this widespread pattern in ES, but attributed this to a substratum (from an unknown source). The present authors, however, take this alternation as a feature inherited from its common ancestor.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Constituent order within the noun phrase Adjectives, quantifiers, numerals, and relative clauses commonly follow the head noun in ES, as they do in general in NS (regardless of constituent order at the clausal level). The position of the possessor as well as of demonstratives, on the other hand, varies within as well as between languages. In Kordofan Nubian languages, nouns in genitive function and demonstratives may either precede or follow the head noun, as in Tagle (Jakobi and Ibrahim ). The question to what extent these alternative constructions are associated with information packaging in this language has not been investigated in detail yet. () ìʃí íŋì ~ íŋì ìʃ ì ‘these hands’ ʃántù-nà kéé ~ kéé ʃántù-nà ‘Shantu’s horse’ Double marking (as in French) whereby the possessive marker also shows whether the head noun is singular or plural is common in the Southern branch of ES, as pointed out by Unseth (: ).4 This pattern is illustrated for the Surmic language Me’en in Table . (based on Unseth : ).
Table . Pronominal possessives in Me’en Possessor
Singular
Plural possessed
1sg
di-a-n
g-a-n-u
2sg
de-n-u
g-e-n-u
3sg
de-n-e
g-e-n-e
1pl:incl
de-g-ay
g-e-g-ay
1pl:excl
de-g-ey
g-e-g-ey
2pl
de-g-u
g-e-g-u
3pl
de-g-e
g-e-g-e
The -n- /-g- alternation in the genitive linkers above is a reflex (presumably) of the common n/k singular/plural alternation in deictic markers found across NS (Greenberg : ), which constitutes one of the most stable diagnostic features of this phylum.
.. Pronominal reference The dramatic variation between Northeastern NS languages in the way pronominal subject marking is expressed on the verb suggests that, like constituent order, this is a rather unstable grammatical phenomenon historically. The following variation occurs:5 This pattern is also attested elsewhere in NS, for example in Fur (Jakobi : –). Nyimang and the Nilotic language Bari (and its closest relatives) are rather unusual in that no subject reference markers occur on verbs. Similar variation occurs with respect to pronominal object marking, but this is not further discussed here for reasons of space. 4 5
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Table . Subject marking on verbs in Anywa Prefixes
Suffixes
1sg
ā-
-ā
2sg
ī-
-Yì
3sg
ɛ̄-
-ɛ̄ (or tonal inflection)
1pl:incl
ɔ ́-
-ɔ ́
1pl:excl
wā-
-wā, -wà
2pl
ū-
-Wū, -Wù
3pl
gī-
-gī,- gì
(a) prefixes; (b) prefixes combined with enclitics; (c) prefixes or suffixes depending on the thematic structure of a clause; (d) suffixes or enclitics. Occasionally, such variation occurs in one and the same language, as shown in Table . for WNil Anywa (Reh : –). Pronominal prefixes in Anywa occur when the verb is not marked for tense–aspect–mood (or if the last marker is non-segmental). Suffixation in Anywa occurs, for example, with NP-initial clauses (Reh : ). More generally, the occurrence of preverbal or postverbal pronominal subject marking across ES and the generalization of one of the two strategies (as with the suffixation pattern in WNil Dinka, Nubian, Berta, or Nara) appear to be linked to the organization of thematic structure, or more generally, information packaging in a clause. Greenberg (: –) already pointed out that the first and second person singular markers are cognate across a wide range of ES and, in fact, NS languages. The first person singular in Anywa, for example, may be compared with Dinka ɣa (Nilotic branch), Temein ŋa, (Temeinian branch), or Tama wa (Taman branch). Because of pronominal reference marking on the verb for subjects, independent pronouns perform a pragmatic role, i.e. as a (contrastive) topic marker (‘as for . . . ’) or in order to express focus. In Anywa and, in fact, in a wide range of Nilotic languages, the copula or predicative marker ni is added with independent pronouns as in ‘āa-nī ‘SG’ (Reh : ). In the closely related Surmic languages, this former specifier has been incorporated (i.e. increment occurred) and a new specifier has been added, as in Murle aneta ‘SG’ (< *a-nɪ-ɪta).
.. Constituent order and head marking versus dependent marking at the clausal level Northern members of ES share a range of typological features with more distantly related Northeastern NS subgroups, such as, Fur, Kunama, Maban, or Saharan. These include extensive number-marking systems (with singulatives), a verb-final constituent order, coverbs as well as converbs, or extensive case-marking systems, features also found in Afro-Asiatic languages in Ethiopia (Dimmendaal ; see also Map . for the geographical spreading of these features). Northern members of ES tend to have rather rich inventories of case markers. The Nubian language Dongolawi, for example, marks the accusative, locative, instrumental,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The position of the verb in Eastern Sudanic
comitative, allative, ablative, adessive, similative, and genitive (Jakobi and El-Guzuuli ). They have a nominative accusative case-marking system, whereby the nominative is morphologically unmarked—with the exception of Nara, which combines a system of
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
differential subject marking (with the case marker -nu /-n) with differential object marking (by means of an accusative case marker -go and allomorphs; Kievit : ).6 The term “differential” refers to the nature of such case inflections, which is governed by semantic principles such as definiteness, specificness, or animacy as well as thematic role in a sentence (Dimmendaal ). In the Kordofan Nubian language Uncunwee, for example, non-emphatic object pronouns are not marked for case and procliticize onto the following verb. However, when the object carries focus, it receives accusative case marking and constitutes an independent word (Jakobi ). () yě à=ɖʊ́k-kὲrὲ. =beat-: ‘I will beat you.’ () yě à-gì ɖʊ́k-kὲrὲ. - beat-: ‘I will beat you (picked out of a group of people).’ Accusative case marking disappeared as a shared innovation of the Southern branch of ES, where objects are never marked for case. Instead (depending on the language), the subjects of both a transitive verb and an intransitive verb are marked for case, i.e. a marked nominative system occurs; or the subject of a transitive verb is marked for case, i.e. an ergative system occurs. Case marking is obligatory for post-verbal subjects in all Southern members of ES that have case systems. The common ergative case markers -ɪ /-i and -ε /-e are attested in all three branches of Southern ES: Gaahmg (Jebel), Päri (Nilotic), and Tennet (Surmic). An example from Gaahmg (Stirtz : ): ()
mīī ma᷇n nāmán=ε᷇ έ ūlg=ì mâŋ wá. goat certain beat.= thirst.= well ‘There was once a very thirsty goat (lit.: a certain goat, thirst beat badly).’
Since exactly the same case markers (-ɪ /-i and -ε /-e) are attested in the Saharan branch of NS, they probably go back to their latest common ancestor, Proto-Saharan-ES. As argued in Dimmendaal (), marked nominative case systems, which are found in Berta as well as different Surmic and Nilotic languages, probably resulted from a generalization of case marking for transitive post-verbal subjects towards post-verbal subjects of intransitive predications. The reduction in case marking in Southern members of ES not only corresponds to a shift in constituent order from a verb-final to a verb-second or verb-initial structure, but also to an increase in verbal valency-marking strategies (i.e. towards head marking at the clausal level). This drift towards head marking at the clausal level is most prominent in Nilotic languages. In Southern Nilotic Kalenjin, for example, the only dependentmarking strategy still found is that between (marked) nominative and absolutive case. The verb on the other hand takes pronominal subject and object markers and verbal markers expressing additional semantic roles; see Rottland (: –), who reconstructs 6
Within Northeastern NS, this latter type of optional case marking on syntactic subjects as well as objects is also found in the Saharan language Dazaga (Walters ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
a range of Proto-Southern Nilotic suffixes expressing instrument, applicative, essive, two types of andative (itive) and ventive (with and without associated motion), as well as inceptive. Pluractionality marking is also a common feature of ES and other NS languages. In the Kordofan Nubian languages, for instance, pluractionality (also known by the term “verbal number”) is realized by singular and plural verb stems. Apart from expressing event number, i.e. aspectual functions such as habitual, iterative, or distributed actions, these stems are also sensitive to transitivity and participant number. Participant number, i.e. the number of the transitive object (patient) and the intransitive subject, triggers the selection of singular or plural stems, as illustrated by the transitive clauses () to () and the intransitive clauses () and () from Uncu (examples drawn from Comfort ). ()
ktʊ́ kūj-ōóŋ. door. open-: ‘You () opened a door.’
()
ktʊ́ kūj-ūúŋ. door. open-: ‘You () opened a door.’
() kn kūj-ēr-ōóŋ. door. open--: ‘You () opened doors.’ ()
kn kūj-ēr-ūúŋ. door. open--: ‘You () opened doors.’
The following intransitive clauses illustrate that the number of the subject triggers the selection of singular and plural verb stems. () ŋāj-ōóŋ. walk-: ‘You () walked.’ ()
ŋāj-ēr-ūúŋ.
walk--: ‘You () walked.’ The interaction between pluractionality and the encoding of grammatical relations is attested in ditransitive clauses, too. In these clauses it is the number of the direct object (theme) that is sensitive to the selection of singular and plural stems, but the number of the indirect object (the recipient-like argument) is insensitive to this selection. That is, the theme morphosyntactically behaves like a patient of a transitive clause, but the recipient-like argument behaves differently. This relation, which is known as indirective alignment (Malchukov et al. ), is illustrated by the Nubian language Karko with examples drawn from Jakobi ().7 7
For further details, including the interaction between verbal number and degree of transitivity, the interested reader is referred to Jakobi ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
. kə̄k-ə̄nd=ə́g ɔ̀g ε̄g-ɲὲn! stone-= : roll.-: ‘Roll the stone for me!’
() kə̄k-ə̄r=ə́g ɔ̀g ε̄gε̄j-nε̂n! stone-= : roll.-: ‘Roll the stones for me!’ The Nile Nubian languages exhibit secundative alignment, i.e. here the selection of singular and plural verb stems is sensitive to the number of the indirect object (the recipient-like argument) but insensitive to the number of the direct object (theme). This alignment pattern is illustrated by the following Nobiin examples (Khalil ). () ay torbar=ka aŋŋaree=g kaay-a-tis. farmer= bed= make--:: ‘I made the farmer a bed.’ ()
ay torbar-ii=ga aŋŋaree=g kaay-a-tic-c-is. sg farmer-= bed= make----: ‘I made the farmers a bed.’
() ay torbar=ka aŋŋaree-ɲci=ga kaay-a-tis. farmer= bed-= make--:: ‘I made the farmer beds.’ Apart from variation between Northern and Southern ES languages in the degree of head marking on the verb, there is another difference in terms of lexical strategies. Languages belonging to the Northern branch of ES and elsewhere in Northeastern NS also have coverb plus light verb constructions. While coverbs provide the lexical meaning, light verbs (usually derived from the verb ‘do’ or ‘say’) serve as the base for inflectional and derivational markers. In the Saharan languages, for instance, the common light verb is ni (or -n). A cognate form is found in ES subgroups like Taman, for example Tama nV-, which combines with nouns, adjectives, or adverbs to form a complex predicate. Coverbs as such have disappeared from the Southern branch of ES, but the common habitual or inceptive-marking suffix -n in this branch presumably is a reflex of this former strategy, as in Berta, where the cognate light verb has been reinterpreted as a suffix-marking inception. The following is an example from Tucker and Bryan (: ), where the noun kere ‘dawn’ may be used predicatively by adding the suffix (or enclitic) -n: () kere-n ‘dawn (lit.: light do/say)’ Converbs, or dependent verb forms in complex sentences, are also commonly used in the Northern branch of ES as a narrative-discourse strategy enhancing the storyline. Nubian languages like Uncunwee also make a formal distinction between events involving the same subject and those involving different subjects, i.e. switch reference is expressed on the dependent verb form (examples from Comfort ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
kàtʊ̀ bārʃ- kàrέ kár-ʃ ε̄r-ε̂. () yὲ állʊ́ŋ tomorrow field harvest- sorghum.flour grind-- ‘I will harvest the field and grind the sorghum-flour tomorrow.’ () yε̌ kʊ̀kʊ̂r=g úr-ùr-èné ʃántʊ̀ kɔ̄l-lʊ́ŋ. chicken= kill--: Shantu eat-: ‘I killed the chicken and Shantu ate it.’ Such converbs frequently develop close semantic ties with the following main verb. Gulfan (: –) gives examples from a Kordofan Nubian language, Taglennaa. () ol-i ŋa-i! come.out- walk- ‘Die!’ In the Southern branch of ES, a radically different type of narrative discourse is used in order to enhance the (primary) storyline, involving the use of a consecutive or subsecutive verb form following (rather than preceding) the main verb. In his survey of constituent order and its link with the organization of the storyline, Longacre (: –) gives examples of this strategy from Nilotic languages like Anywa, Luwo, Sabaot, and Toposa. Reh (: –) remains one of the most detailed analyses of narrative discourse strategies used in a Southern member of ES. The following example illustrates the syntactically dependent (or consecutive) verb form ‘embrace’, which follows the main verb in Anywa. () màr ā-òo máal òo gʌ̀ʌm ócòk aunt:his -come up and embrace: nephew:her òo tèen gɔ́. and cook::: him ‘His aunt got up, embraced her nephew and cooked for him.’ Considerable progress has been made with the synchronic analyses of ES languages over the past decades, although genetically isolated languages like Nara or the Temeinian cluster remain poorly understood. Additional in-depth comparative studies of the various subgroups in the near future most likely will result in a better understanding of the historical origin of ES languages, their genetic divergence, and astonishing typological disparity.
R Abdel-Hafiz, A. S. (). A Reference Grammar of Kunuz Nubian. Ph.D. thesis. Buffalo, NY: State University of New York. Andersen, T. (). ‘Vowel length in Western Nilotic languages’, Acta Linguistica Hafniensia : –. Bender, M. L. (ed.) (a). Proceedings of the Fourth Nilo-Saharan Conference, Bayreuth, August –September , . Hamburg: Helmut Buske. Bender, M. L. (b). ‘Sub-classification of Nilo-Saharan’, in M. L. Bender (ed.), –. Bender, M. L. (). The East Sudanic Languages: Lexicon and phonology. Carbondale, IL: Southern Illinois University Printing.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Bremer, N. D. (). ‘Berta and the East Jebel subfamily: reinvestigating a Nilo-Saharan isolate’, in A. Mietzner and A. Storch (eds.), Nilo-Saharan Models and Descriptions. Cologne: Rüdiger Köppe, –. Bryan, M.A. (). ‘The T/K languages: a new substratum’, Africa : –. Comfort, J. (). ‘Converbs in Uncunwee (Kordofan Nubian)’, in T. C. Schadeberg and R. Blench (eds.), –. Comfort, J. (). ‘Verbal number in the Uncu language (Kordofan Nubian)’, Dotawo : –. Dimmendaal, G. J. (). ‘The lexical reconstruction of Proto-Nilotic: a first reconnaissance’, Afrikanistische Arbeitspapiere : –. Dimmendaal, G. J. (). ‘The role of bilingualism in Nilotic sound change’, in M. Dominicy and D. Demolin (eds.), Sound Change (Belgian Journal of Linguistics, ). Amsterdam: John Benjamins, –. Dimmendaal, G. J. (). ‘Number marking and noun categorization in Nilo-Saharan languages’, Anthropological Linguistics : –. Dimmendaal G. J. (). ‘Language shift and morphological convergence in the Nilotic area’, Sprache und Geschichte in Afrika /: –. Dimmendaal, G. J. (). ‘Eastern Sudanic and the Wadi Howar and Wadi El Milk diaspora’, Sprache und Geschichte in Afrika : –. Dimmendaal, G. J. (). ‘Tama’, in G. J. Dimmendaal (ed.), Coding Participant Marking: Construction types in twelve African languages. Amsterdam and Philadelphia, PA: John Benjamins, –. Dimmendaal, G. J. (). ‘Differential object marking in Nilo-Saharan’, Journal of African Languages and Linguistics : –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Dimmendaal, G. J. (). ‘Nilo-Saharan’, in R. Lieber and P. Štekauer (eds.), The Oxford Handbook of Derivational Morphology. Oxford: Oxford University Press, –. Edgar, J. (). ‘First steps toward Proto-Tama’, in M. L. Bender (ed.), –. Ehret, C. (). A Historical-Comparative Reconstruction of Nilo-Saharan. Cologne: Rüdiger Köppe. Greenberg, J. H. (). Studies in African Linguistic Classification. New Haven, CT: Compass. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Gulfan, G. I. (). ‘Converbs in Taglennaa (Kordofan Nubian)’, in T. C. Schadeberg and R. Blench (eds.), –. Hayward, R. J. (). ‘Observations on tone in the Higir dialect of Nara’, in R. Vossen, A. Mietzner, and A. Meißner (eds.), Mehr als nur Worte . . . Afrikanistische Beiträge zum : Geburtstag von Franz Rottland. Cologne: Rüdiger Köppe, –. Ibrahim, G., and Huttenga, P. (). ‘The phoneme system of Tagle, a Kordofanian Nubian language’, in D. Payne and M. Reh (eds.), Advances in Nilo-Saharan Linguistics: Proceedings of the th Nilo-Saharan Linguistics Colloquium, Hamburg, August –, . Cologne: Rüdiger Köppe, –. Jakobi, A. (). A Fur Grammar. Hamburg: Helmut Buske. Jakobi, A. (). Differential object marking in Uncunwee (Kordofan Nubian). Paper presented at the th World Congress of African Linguistics, University of Cologne. Jakobi, A. (). ‘Verbal number and transitivity in Karko (Kordofan Nubian)’, Sprachtypologie und Universalienforschung : –. Jakobi, A. (to appear). Kordofan Nubian: A Synchronic and Diachronic Study. Jakobi, A., and El-Guzuuli, El-Shafie. (). ‘Heterosemy of case markers and clause-linkers in Andaandi (Nile Nubian)’, Studies in African Linguistics : –. Jakobi, A., and Hamdan, A. (). ‘Number marking on Karko nouns’, Dotawo : –. Jakobi, A., and Ibrahim, A. (). Number marking and noun classification in Tagle (Kordofan Nubian). Paper presented to the Nilo-Saharan Linguistics Colloquium at Nairobi, September –, .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Khalil, M. K. (). ‘The verbal plural marker in Nobiin (Nile Nubian)’, Dotawo : –. Kievit, D. (). ‘Nara’, in S. Uhlig (ed.), Encyclopedia Aethiopica. Vol. . Wiesbaden: Otto Harrassowitz, –. Köhler, O. (). Die nilotischen Sprachen: Darstellung ihres Lautsystems, nebst einer Einleitung über die Geschichte ihrer Erforschung, ihrer Verbreitung und Gliederung. Ph.D. thesis. Berlin: Humboldt-Universität. Longacre, R. E. (). Storyline Concerns and Word Order Typology in East and West Africa. Los Angeles: James S. Coleman African Studies Center and the Department of Linguistics, University of California at Los Angeles. Malchukov, A., Haspelmath, M., and Comrie, B. (). ‘Ditransitive constructions: a typological overview’, in A. Malchukov, M. Haspelmath, and B. Comrie (eds.), Studies in Ditransitive Constructions. Berlin und NewYork: de Gruyter, –. Palayer, P., with de Yaya, C. (). Grammaire du dadjo d’Eref (Tchad). Leuven and Paris: Peeters. Reh, M. (). Anywa Language: Description and internal reconstructions. Cologne: Rüdiger Köppe. Rilly, C. (). Le méroïtique et sa famille linguistique. Leuven and Paris: Peeters. Rottland, F. (). Die südnilotischen Sprachen: Beschreibung, Vergleichung und Rekonstruktion. Berlin: Dietrich Reimer. Schadeberg, T. C., and Blench, R. (eds.) (). Nuba Mountain Language Studies. Cologne: Rüdiger Köppe. Schrock, T. (). A Grammar of Ik. Utrecht: LOT. Stevenson, R. C. (). ‘A survey of the phonetics and grammatical structure of the Nuba mountain languages, with particular reference to Otoro, Katcha and Nyimang’, Afrika und Übersee : –, –, –. Stirtz, T. (). A Grammar of Gaahmg, a Nilo-Saharan language of Sudan, https://openaccess. leidenuniv.nl/handle// accessed July , . Stirtz, T. (). ‘Ergative, antipassive and other verb derivational morphemes in Gaahmg’, Journal of African Languages and Linguistics : –. Storch, A. (). ‘Dynamics of interacting populations: language contact in Bahr el-Ghazal’, Studies in African Linguistics : –. Storch, A. (). A Grammar of Luwo: An anthropological approach. Amsterdam and Philadelphia, PA: John Benjamins. Thelwall, R. (). ‘A Birgid vocabulary list and its links with Daju’, in H. Ganslmayr and H. Jungraithmayr (eds.), Gedenkschrift Gustav Nachtigal –. Bremen: Übersee-Museum, –. Thelwall, R. (). The Daju Language Group. Ph.D. thesis. New University of Ulster. Tucker, A. N. (). A Grammar of Kenya Luo (Dholuo). Edited by C. A. Creider. Cologne: Rüdiger Köppe. Tucker, A. N., and Bryan, M. A. (). Linguistic Analyses: The non-Bantu languages of north-eastern Africa. London: Oxford University Press. Unseth, P. (). ‘Possessive markers in Surmic languages’, in M. L. Bender (ed.), –. Walters, J. K. (). A Grammar of Dazaga. Leiden and Boston, MA: Brill. Yigezu, M. (). A Comparative Study of the Phonetics and Phonology of Surmic Languages. Ph.D. thesis, Université Libre de Bruxelles. Yigezu, M., and Dimmendaal, G. J. (to appear). A Grammatical Sketch of Ngaalam.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T label “Central Sudanic” (CS) traditionally applies to a group of some sixty languages that are spoken in the center of the African continent and cover parts of Chad, Sudan, Central African Republic, Democratic Republic of Congo (DRC), and Uganda (see Map .). With the exception of the Sara languages of Chad and Moru of Sudan/DRC (more than one million each)1 most languages have fewer than , speakers, and many of them just a few thousand. Several languages are declining, namely under the influence of Chadian or Sudanese Arabic. Typological features include short word shapes (V, CV, VCV, and CVCV), frequent [ATR] vowel feature (in the southeastern part of the area), widespread glottalized consonants (see section .), and tone systems that may include up to four distinctive levels. Word order is diverse (see section .). Tucker and Bryan (, ) divided these languages in two “larger units” (. MoruMangbetu, . Bongo-Bagirmi) despite some similarities in lexicon and morphology (: –). Greenberg () coined the term “Central Sudanic” and brought all the languages together—first as a part of Chari-Nile, then directly (Greenberg )—within his NiloSaharan (NS) phylum. Finally, Bender () divided CS (renamed “Family F”) in two subgroups (“Peripheral” and “Central”) that roughly correspond to Tucker and Bryan’s “larger units”. These classifications are summarized in Table .. The aim of this chapter is to evaluate the situation of these languages in the light of several criteria. I will argue that they probably constitute a genetic, historical unit made up of five more or less equidistant subgroups, with the exception of some indeterminate languages the situation of which remains uncertain for the time being. I will not address the relation of CS with other NS languages.
1
Estimates from Bender (: ).
The Central Sudanic Languages kuka bulala kuka medogo
beraku
Sinyar
kenga
NIGERIA ‘barma
gele
CHAD
fer kulfa na tiye gula sime ‘dem gulay sar ngambay ndoka luto mvang yulu ‘bedjond ngama mbay wad vale nduga kaba
Moru-Madi
Mangbutu-Efe
Lendu
SBB
Mangbetu
Others
yulu binga Dongo yulu binga Kresh ‘bu‘bu Aja
SUDAN ETHIOPIA bongo
Woro
CENTRAL AFRICAN REPUBLIC Birri bagiro
CAMEROON
modo ‘beli etc. morokodo nyamusa etc. baka
furu
DEM. REP. CONGO CONGO
UGANDA
0
100
500 km
. The geographical distribution of Central Sudanic languages
Table . Previous groupings of Central Sudanic languages Tucker & Bryan () Greenberg ()
. M-M Moru-Ma’di (LG) Mangbutu-Efe (LG) Mangbetu (LG) Lendu (SU)
. B-B Bongo (LG) Kresh (LG) Sinyar (SU) Kara (SU) Sara (LG) Bagirmi (LG)
C S ( C-N NS) . Moru, Avukaya, Logo, Keliko, Lugbara, Madi . Mangbetu, Lombi, Popoi, Makere, Meje, Asua . Mangbutu, Mamvu, Lese, Mvuba, Efe . Lendu
Greenberg ()
Bender ()
C S ( NS) . Moru-Madi . Mangbetu . Mangbutu-Efe . Lendu
F F P F: Moru-Madi F: Mangbutu F: Mangbetu F: Kresh F: Baadha (= Lendu)
. Bongo-Bagirmi . Bongo, Baka, A. Bongo Morokodo, [ . . . ], B. Sara, Kara, Sara dialects ([ . . . ]), Bagirmi [ . . . ], Bagirmi, Kuka, C. Yulu Kenga, [ . . . ] . Kreish . Kreish . Binga, Yulu, Kara
(LG = language group; SU = single unit)
C Fa: Bagirmi-Sar Fb: Yulu-Binga Fc: Fongoro F: Shemya (= Sinyar) F: Bongoid
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. I
.................................................................................................................................. CS includes the following subgroups that basically correspond to those identified by earlier scholars (some representative languages are indicated in brackets): . . . . .
[MMD] Moru-Madi (Moru, Avokaya, Logo, Lugbara, Ma’di) [LND] Lendu (Lendu, Ngiti) [MAS] Mangbetu(-Asua) (Mangbetu, Meje, Lombi, Asua) [MEF] Mangbutu-Efe (Mangbutu, Ndo, Mamvu, Lese, Efe) [SBB] Sara-Bongo-Bagirmi (Modo, Baka, Bongo, Yulu, Gula dialects, Ndoka, Bagiro, Na, Kenga, Bagirmi, Sara languages)
Every subgroup is relatively well documented for at least two languages. Given its linguistic coherence, each one can be assumed to constitute a historical unit. SBB, which represents the largest and most diversified subset, is also the best-established genetic unit (Boyeldieu b; Boyeldieu et al. ). The tricky question that is addressed in this chapter, then, concerns the historical unity of CS as a whole. Lastly, and although they are usually considered as CS, I provisionally regard the following languages as indeterminate outsiders. Despite obvious similarities, Kresh, Dongo, Woro, Aja, and Birri (Santandrea /, ) do not show lexical and sound correspondences that are significant enough with the above-mentioned subgroups. On the basis of limited data (Doornbos and Bender ), Sinyar has been counted among the SBB subgroup. However, recent fieldwork of my own shows that Sinyar also differs from SBB in many important features, notably the existence of a case-marking system. The specific development of Sinyar, therefore, remains an open question.
. L
.................................................................................................................................. Two types of estimates have been made regarding the lexicon. Tables .a and .b provide the results of a lexicostatistical study for a sample of languages (based on a Swadesh -wordlist; Branch Average).2 MAS shows an internal similarity rate of %, MMD, LND, and MEF a rate of %, while the internal rate of SBB is markedly lower (%). The similarity rates between subgroups vary from % (MMD with LND) to % (all subgroups). While the general figure clearly indicates five coherent subgroups, the higher connections between the latter rely on rates that are much less distinctive. As for the indeterminate languages, Sinyar is close to SBB (%). Kresh, Dongo, Aja, and Birri also stand closer to SBB, although with a low rate of %.
2
I am indebted to Thilo Schadeberg for his LEXISTAT software.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table .a Lexicostatistical distances between twenty-nine CS languages (based on a hundred-word Swadesh list; Branch Average) [BA] % 100 90 80 70 60 50 40 30 20 10 0
MORU-MADI LENDU M. –ASUA M. –EFE SBB 1 2 3 4 5 6 13 14 10 11 12 7 8 9 15 16 28 17 18 19 20 21 22 23 24 25 27 26 29 mz av lo ka lr ma ni lu as mt li mv ls ef kr dn bi aj md bk bn yu gk gz bo na sr ke si
Table .b Identification of subgroups and languages MORU-MADI LENDU
MANGBETUASUA
1 mz Moru Miza 13 ni Ngiti 10 as Asua 2 av Avukaya 3 lo Logo
MANGBUTUEFE
7 mv Mamvu 18 md Modo
14 lu Lendu 11 mt Mangbetu 8 ls 12 li
Lombi
SBB
9 ef
Indeterminate
15 kr Kresh
Lese
19 bk Baka
16 dn Dongo
Efe
20 bn Bongo
17 aj Aja 28 bi Birri
4 ka Kaliko
21 yu Yulu
5 lr Lugbara
22 gk Gula Koto 29 si Sinyar
6 ma Ma’di
23 gz Gula Zura 24 bo Bagiro 25 na Na 26 ke Kenga 27 sr Sar
Table .a displays the distribution patterns of likely cognates between the five subgroups.3 Two remarks should be made here. First, the number of likely cognates shared by all the five subgroups () or even by four of them () is noticeably low. Second, the number of cognates shared by two subgroups (Table .b) varies from =.% (MEF with LND, and SBB with LND) to =.% (MAS with LND), thus revealing no clear-cut predominance in their distribution. Here, again, the overall figure does not bring out clearly any marked proximity between any of the subgroups, which rather appear as more or less equidistant. 3
The counting is based on a personal database. An occurrence in one language is considered enough to represent a subgroup. Indeterminate languages are not taken into account.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table .a Distribution patterns of likely cognates in the CS subgroups (excluding indeterminate languages) and number of occurrences Distr. 5x
4x
3x
2x
MMD
LND
MAS
MEF
SBB
Total
Total
+
+
+
+
+
20
20
+
+
+
+
–
10
+
+
+
–
+
14
+
+
–
+
+
7
+
–
+
+
+
11
–
+
+
+
+
9
+
+
+
–
–
14
+
+
–
+
–
11
+
–
+
+
–
12
–
+
+
+
–
7
+
+
–
–
+
6
+
–
+
–
+
12
–
+
+
–
+
3
+
–
–
+
+
9
–
+
–
+
+
4
–
–
+
+
+
13
+
+
–
–
–
23
+
–
+
–
–
16
–
+
+
–
–
7
+
–
–
+
–
8
–
+
–
+
–
3
–
–
+
+
–
8
+
–
–
–
+
8
–
+
–
–
+
8
–
–
+
–
+
17
–
–
–
+
+
7 267
51
91
105
267
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table .b Absolute occurrences and percentages (out of ) of shared likely cognates in pairs of subgroups MMD
LND
MAS
MEF
LND
105 39.3%
MAS
109 40.8%
84 31.5%
MEF
88 32.9%
71 26.6%
90 33.7%
SBB
87 32.6%
71 26.6%
99 37.1%
80 30.0%
Table . Central Sudanic correspondence formulas for labiovelar consonants CS
*VCV
*kp
*gb
*ngb
*ŋm
MMD
*(V)CV
*kw-a/E/i
*gw-ɛ/i, (*g-u ?)
*ngw
*ŋw-a/u
LND
*VCV
*kp / *ts
*gb / *dz
*ngb / *ndz
*m ?
MAS
*VCV
*kw-a/ɛ
*g ?
*ngw-a/ɛ
*ŋw-a/ɛ ?
MEF
*(C)VCV
*kp
*? / *g-u ?
*ngb ?
*?
SBB
*(C)VCV
*kp
*gb / *g-u ?
(*ngb ?)
*ŋm
. S
.................................................................................................................................. Establishing regular sound correspondences proves to be a difficult task given the restricted number of general cognates, the variety of likely reflexes (especially for vowels and tones), and the shortness of the word shapes to be reconstructed (mostly *VCV). Bender () made a serious attempt at it, although his results are not quite convincing in my opinion.4 In a very limited way Boyeldieu () establishes regular correspondences for labiovelar consonants (intermediate reflexes are summarized in Table .). Glottalized consonants also show persuasive series of regular reflexes. LND languages display an uncommon contrast of voiceless/voiced glottalized consonants (Dimmendaal ; Kutsch Lojenga ) that must be assigned to the likely CS sound system. This contrast was lost and reflexes have merged in all other subgroups (intermediate reflexes are summarized in Table .).
4 For instance, out of his four series illustrating a *kp formula, only one is valid for me (Boyeldieu ). As for the glottalized consonants, the author disregards the important voiceless/voiced contrast that characterizes the LND languages.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Central consonants
Sudanic
correspondence *ɓ
*ɗ
*ƈ
*ɗ (/ *ɗw)
*ʄ
*ƭ
*ƈ
for
*VCV
*ƥ
MMD
*(V)CV
*ɓ
LND
*VCV
*ƥ
MAS
*VCV
*ɓ / *ʙ
*ɗ
*ʄ ?
*ʔ
MEF
*(C)VCV
*ɓ
*l
*j
*VØV
SBB
*(C)VCV
*ɓ
*ɗ
*ʄ
*ʔ
*ɗ
*ʄ
glottalized
CS
*ɓ
*ƭ
formulas
*ʔ *ʔ
*ʄ
*ʔ
. M
.................................................................................................................................. Tucker and Bryan (, ) pertinently pointed up morphological similarities that can be observed in most CS subgroups. This regards the following three verbal prefixes: T(V)-: verbal intensive (Tucker and Bryan : ) [MMD], e.g. Logo tV- (Vallaeys : ) ’di ‘percer’ ti’di ‘percer’ (répétitif/intensif) [LND] no evidence [MAS], e.g. Mangbetu: possible remnants of -Vt- ? (after Larochette : passim) -εtápú ‘s’évanouir, anéantir’ cf. -ɔpú ‘tuer, détruire, anéantir’ [MEF], e.g. Mamvu ṭì-/ṭì- (Vorbichler : –, and passim) ɔ̀ma ‘schlagen’ ṭìma ‘schlagen (viele, PL)’ [SBB], e.g. Gula t- (Nougayrol : –) èh e ‘nommer’ téhē (Gula Koto) ‘nommer à plusieurs reprises’ K(V)-: Deverbative (Tucker and Bryan : ) [MMD], e.g. Moru Miza k-: formative of adjectives, e.g. in contrast with Avukaya Ojila (Tucker : ) kɔ̀zɪ (cf. Avukaya Ojila ɔnzí) ‘bad’ [LND] no evidence [MAS], e.g. Mangbetu k(u)-: verbal nouns (Larochette : –) -εgu ‘voler’ nέkεgú ‘vol, voleur’ [MEF], e.g. Mamvu q-: deverbative function (Vorbichler : –) òmvu ‘faulen’ qomvù ‘feucht, naß’ [SBB], e.g. Bagiro k-̀: infinitives and nouns/adjectives (Boyeldieu a: –) gyɔ̀ ‘il enfante’ k-ɔ̀gyɔ̀ ‘enfanter’ kɔ̀gyɔ̀ ‘parent’ O-/E- or A-/E-: a. itive/ventive; b. neuter/causative (Tucker and Bryan : –) [MMD], e.g. Moru (Tucker : –, –) a. ɔ-gɔ ‘to go back’ εgɔ ‘to come back’ b. o-ŋga ‘to get up’ eŋga ‘to raise up’ [LND], e.g. Ngiti (Kutsch Lojenga : –, –) a. otseta ‘to run’ (itive) itsétá ‘to run’ (ventive) b. avhítá ‘to get lost’ ivhítá ‘to lose something’ [MAS], e.g. Mangbetu (Larochette : –, –)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Evidence of the verbal prefixes through the subgroups MMD LND MAS MEF SBB T(V)-: verbal intensive
+
–
?
+
+
K(V)-: deverbative
+
–
+
+
+
O-/E- or A-/E-: a. itive/ventive; b. neuter/causative +
+
+
+
–
a. -ɔku ‘repartir’ -εkú ‘revenir’ b. -owú ‘égoutter’ -ewú ‘laisser égoutter’ [MEF], e.g. Mamvu (Vorbichler : , ) a. ɔ̀si ‘gehen’ εsí ‘kommen’ b. òfe ‘zu Ende gehen’ ìfe ‘beenden’ [SBB] no evidence Although the distribution of these similarities between the subgroups reveal some gaps (especially in LND; see Table .), there is no clear indication that a specific subgroup— especially not SBB—should be left out at this point.
. W
.................................................................................................................................. Tucker and Bryan (: –) further justified their distinction of two larger units, BongoBagirmi (SBB) and Moru-Mangbetu (other subgroups) by a contrast in word order. However, things are not so straightforward. Table . displays a selection of typological features as characterized by Dryer and WALS () for CS languages. Notwithstanding clear contrasts of MMD and LND with SBB (features B and A), Mangbetu (MAS) and Mamvu (MEF) show the same S–V–O order as SBB (A), while Moru (MMD) shows the same noun–genitive order as SBB (A). Most probably SBB, which extends into the northwestern part of the area, reinforced innovations in this respect but, here again, there is no evidence for a clear-cut contrast between SBB and the other subgroups altogether.
. C
.................................................................................................................................. The various observations made in the preceding sections lead to the following concluding remarks: (a) There is no clear evidence that SBB should be contrasted with the other subgroups as a whole; CS, rather, comprises five more or less equidistant subsets. (b) There is some evidence that the five subgroups may well result from the historical splitting of a single language (or dialect group); however, considerable work is still necessary to confirm this assumption and to explain the diversity of the present-day languages, a diversity that seems somewhat contradictory with their geographical proximity. (c) Lastly, the affiliation of some indeterminate languages remains uncertain; their similarities with CS could turn out to be the result of different types of contact.
81A
81B
Position of pronominal possessive affixes
Order of subject, object and verb
Languages with Order of adposition and noun phrase two dominant orders of subject, object, and verb
Order of genitive and noun
–
no dominant order
SOV or SVO
postpositions
no dominant order
Logoti
–
no dominant order
SOV or SVO
postpositions
genitive noun
Lugbara
–
no dominant order
SOV or SVO
postpositions
genitive noun
Ma’di
–
no dominant order
SOV or SVO
postpositions
no dominant order
Moru
possessive suffixes
no dominant order
SOV or SVO
postpositions
noun genitive
Lendu
possessive prefixes
no dominant order
SOV or SVO
postpositions
genitive noun
Ngiti
possessive suffixes
no dominant order
SOV or SVO
postpositions
genitive noun
MAS
Mangbetu
possessive suffixes
SVO
–
–
no dominant order
MEF
Lese
–
no dominant order
–
–
no dominant order
Mamvu
possessive prefixes
SVO
–
–
no dominant order
Bagirmi
possessive suffixes
SVO
–
prepositions
noun genitive
Bagiro
possessive suffixes
SVO
–
no dominant order
noun genitive
Baka (Sud.) –
SVO
–
prepositions
noun genitive
Binga
possessive suffixes
–
–
–
–
Bongo
possessive suffixes
SVO
–
prepositions
noun genitive
MMD Avokaya
LND
SBB
85A
86A
57A
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . A selection of WALS typological features for CS languages (Dryer and WALS )
SVO
–
no dominant order
noun genitive
Jur Mödö
SVO
–
prepositions
noun genitive
Kara (CAR) possessive suffixes
SVO
–
prepositions
noun genitive
Kenga
possessive suffixes
SVO
–
no dominant order
noun genitive
Mango
–
–
–
–
noun genitive
Mbay
possessive suffixes
SVO
–
no dominant order
noun genitive
Ngambay
possessive suffixes
SVO
–
prepositions
noun genitive
Yulu
possessive suffixes
SVO
–
prepositions
noun genitive
–
SVO
–
prepositions
noun genitive
Kresh
possessive suffixes
SVO
–
prepositions
noun genitive
Birri
–
–
–
–
–
Others Aja
–
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Gula (CAR) –
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Bender, M. L. (). ‘Central Sudanic segmental and lexical reconstructions’, Afrikanistische Arbeitspapiere : –. Bender, M. L. (). ‘Nilo-Saharan’, in B. Heine and D. Nurse (eds.), African Languages; An introduction. Cambridge: Cambridge University Press, –. Boyeldieu, P. (a). La langue bagiro (République Centrafricaine): Systématique, textes et lexique. Frankfurt am Main: Peter Lang. Boyeldieu, P. (b). Identité tonale et filiation des langues sara-bongo-baguirmiennes (Afrique centrale). Cologne: Rüdiger Köppe. Boyeldieu, P. (). ‘Reflexes of a labiovelar series in Central Sudanic’, in A. Abu-Manga, L. Gilley, and A. Storch (eds.), Insights into Nilo-Saharan Language, History and Culture. Cologne: Rüdiger Köppe, –. Boyeldieu, P., Nougayrol, P., and Palayer, P. (). Lexique comparatif historique des langues sara-bongobaguirmiennes. Paris: CNRS-LLACAN, http://sumale.vjf.cnrs.fr/SBB/ accessed July , . Dimmendaal, G. J. (). ‘Language typology, comparative linguistics and injective consonants in Lendu’, Afrika und Übersee : –. Doornbos, P., and Bender, M. L. (). ‘Languages in Wadai-Darfur’, in M. L. Bender (ed.), NiloSaharan Language Studies. East Lansing, MI: Michigan State University, –. Dryer, M. S. and the WALS author team (). ‘Family Central Sudanic’, in M. S. Dryer and M. Haspelmath (eds.), The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, https://wals.info/languoid/family/centralsudanic#/./. accessed October , . Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Greenberg, J. H. (). ‘Nilo-Saharan and Meroitic’, in T. Sebeok (ed.), Current Trends in Linguistics. Vol. : Linguistics in Sub-Saharan Africa. The Hague: Mouton, –. Kutsch Lojenga, C. (). ‘Lendu: a new perspective on implosives and glottalized consonants’, Afrika und Übersee : –. Kutsch Lojenga, C. (). Ngiti: A Central-Sudanic language of Zaire. Cologne: Rüdiger Köppe. Larochette, J. (). Grammaire des dialectes mangbetu et medje. Tervuren: Annales du Musée Royal du Congo Belge. Nougayrol, P. (). Les parlers gula (Centrafrique, Soudan, Tchad). Grammaire et lexique. Paris: Editions du CNRS. Santandrea, S. (/). ‘The Birri language: brief elementary notes’, Afrika und Übersee : –, –. Santandrea, S. (). The Kresh Group: Aja and Baka languages (Sudan). Naples: Istituto Universitario Orientale. Tucker, A. N. (). The Eastern Sudanic Languages. Vol. . London, New York, and Toronto: Oxford University Press. Tucker, A. N., and Bryan, M. A. (). The Non-Bantu Languages of North-Eastern Africa. London, New York, and Cape Town: Oxford University Press. Tucker, A. N., and Bryan, M. A. (). Linguistic Analyses: The non-Bantu languages of north-eastern Africa. London, New York, and Cape Town: Oxford University Press. Vallaeys, A. (). Dictionnaire logo-français, suivi d’un index français-logo. Tervuren: Musée Royal de l’Afrique Centrale. Vorbichler, A. (). Die Sprache der Mamvu. Hamburg: J. J. Augustin.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. K studies essentially began in the nineteenth century with the work of Wilhelm Bleek and Lucy Lloyd. Bleek was able to work with speakers of ǀXam who had been paroled from the Cape Town Gaol and House of Correction to live in his house. His sister-in-law Lucy Lloyd, also resident in Bleek’s house, began to assist with the transcriptions and eventually took over most of the research.1 Bleek’s daughter Dorothea Bleek continued their research, carrying out studies of Naro, ǂKx’auǁ’ẽ, ǀAuni, Hadza, and others (e.g. Bleek , –, , a, b). Subsequent research was sporadic, short studies of individual languages, such as Doke’s (, ) field studies of !Xũ and ǂKhomani. In , Ernst Westphal carried out a field trip covering all the Khoisan groups in the southern African area known at that time (see Map .), recording a -word vocabulary and a grammatical sketch for each. In the same time period, Oswin Köhler began his studies of Kxoe (“Khwe” in official orthography) (Khoe family, West Kalahari branch) and Juǀ’hoan (Ju family), and Westphal’s disciple Jan Snyman (, ) produced the first grammar and dictionary of a Khoisan language outside Khoekhoe. Although Khoisan researchers still remain a small minority compared to linguists specializing in other African languages, Khoisan research has expanded considerably since the s. A world conference on Khoisan was held every three years in Riezlern, Kleinwalsertal (Austria) between and , and a follow-up meeting was organized in South Africa in .
. S
.................................................................................................................................. Most of the Khoisan languages still spoken are severely endangered. Current population figures (taken from Ethnologue) are , for Juǀ’hoan, , for Khoekhoe, and , for Sandawe, but the number of fluent speakers for the majority of the surviving languages range from a handful of elderly speakers (as is the case for Nǀuu) to a few hundred (e.g. Hadza with some speakers). 1
For a major joint publication as a result of this cooperation, see Bleek and Lloyd ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. The geographical distribution of Khoisan languages
Even those languages with relatively large numbers of speakers are in danger, since Khoisan languages are not regarded, often even by their own speakers, as prestigious languages, and there is constant pressure on speakers to shift to the national languages or the dominant local language.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Some communities have begun to fight to maintain their language. A prominent example is the Nyae Nyae Conservancy, which promotes the development of educational and reading materials in Juǀ’hoan. Similar projects have been initiated for Naro and Khwe.
. F :
.................................................................................................................................. Many of the surviving Khoisan languages are still little known. Although Westphal, Köhler, Traill, and Vossen all carried out field surveys of the Khoe languages, the only researchers studying East Kalahari varieties at present are Chebanne (esp. Cirecire), Fehn (e.g. ), and Collins (e.g. Collins and Chebanne forthcoming). A valuable in-depth study of Tsua tonology has been provided by Mathes (). In general, however, the majority of the East Kalahari languages are little known, if not unknown, in the literature. Two Ju languages, Juǀ’hoan and Ekoka !Xũ (!Xun), have been studied in detail, and Snyman carried out a dialect survey of the Ju languages in . Considerable progress was made recently through the publication of Heine and König’s () comparative grammar of !Xun. Nevertheless, the present situation has improved in basic resources. There are now dictionaries and grammars available for a number of Khoisan languages and others in progress. In the Ju family (formerly Northern Khoisan), two dictionaries are available for Juǀ’hoan: Snyman’s original dictionary and Dickens (). Although the latter seems to be preferred by current researchers, the two dictionaries only partially overlap and are actually complementary. The same is true of the two grammars, Snyman () and Dickens (). The Dickens grammar is a short sketch which adds new information to Snyman’s description of Juǀ’hoan, but at the same time does not cover many topics in Snyman. Köhler () is a detailed sketch of the variety of Juǀ’hoan spoken at Dikundu (Namibia). König and Heine treat this as a separate language, but it is close enough to Juǀ’hoan spoken in the Tsumkwe district to be considered a dialect. The Northwestern branch of Ju is represented by König and Heine’s () dictionary of Ekoka !Xun. A detailed grammar of Ekoka !Xun is underway, but an early version of the grammar was published in the Working Papers series of the Khoisan Forum (König and Heine ). Ekoka !Xun is very close to the !Xun spoken in Ovamboland, for which there is a substantial wordlist, as well as a detailed grammatical sketch (Heikkinen ). Khoe (formerly Central Khoisan) is the largest of the remaining Khoisan families, although most of the languages belonging to this family which are still spoken are on the verge of extinction. Among the Khoe languages, the Nama and Damara dialects of Khoekhoe have the longest continuous tradition of study. The only modern grammar of Khoekhoe is Hagman (), based on his doctoral thesis. However, Krönlein’s () original Nama dictionary and Rust’s () updated version have been replaced by Haacke and Eiseb’s dictionary (), which greatly improves on previous dictionaries of this language, with its , headword and compound entries, fully tone-marked and supplemented by numerous examples. In addition to this, Haacke () has written a study of Khoekhoe tonology and is carrying out research on Khoekhoe dialects. Brugman’s () thesis, a study of Khoekhoe phonology, offers an alternative analysis of Khoekhoe tonology and is available online.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Khwe, a member of the Northern subfamily of the West Kalahari branch of Khoe, is well represented. There is a dictionary (Killian-Hatz ) and a grammar (Kilian-Hatz ). Köhler’s monumental study of Khwe texts covers both Khwe culture and language and is now up to four volumes (Köhler , , , ). Naro, a member of the Naro-Gǁana branch of West Kalahari, has greatly benefited from the Naro language project, headed by Hessel Visser. The Naro dictionary (Visser ) has around , entries. The project has also published literacy materials, primers, and Bible translations. Gǀui, another Naro-Gǁana language, has been subject of series of field studies carried out by Hiroshi Nakagawa, Hitomi Ono, Akira Takada, and other Japanese researchers. Nakagawa’s () doctoral thesis dealt with Gǀui phonology, and a Gǀui dictionary is in progress. Nakagawa’s studies have expanded to include ǂHaba and Gǁana, two other Naro-Gǁana languages. In Southern Khoisan, now generally referred to as Tuu, one of the most dramatic events in recent Khoisan studies was Nigel Crawhall’s () discovery of surviving speakers of Nǀuu, which had been thought to be extinct. The few surviving full speakers have taken part in a series of field studies of their remarkable language. Collins and Namaseb () have written a short grammar of Nǀuu, published in the well-established QKF2 monograph series, and a dictionary is being prepared by Bonny Sands online. Traill’s () dictionary of Lone Tree !Xóõ (!Xoon) set a standard for Khoisan dictionaries. It contains over , entries, detailed lists of personal names, birds, and plants (identified by their scientific nomenclature), and other noteworthy features. Each entry contains basic grammatical information and numerous example sentences. The DoBeS project3 has been extending this research to Western !Xóõ. Geoffrey Gruber’s ǂHõã (ǂHoan) wordlist () has been digitized and Collins and Gruber (, ) combined the results of their field studies in a ǂHõã grammar. There have been other field studies of ǂHõã, the most recent being Sands, Chebanne, and Shah (results not yet published). Most recently, an enormously detailed study of the phonology of N!aqriaxe, a variety of ǂ’Amkoe, which is closely related to ǂHõã, was published by Gerlach (). Grammars by Eaton () and Steeman () have largely enriched our knowledge of Sandawe, and a first comprehensive dictionary based on ten Raa’s field collections was provided by Ehret and Ehret (). In addition, Elderkin has published on many aspects of Sandawe structure and phonology (see, for example, his contributions in Vossen ). No full-length study of Hadza has yet been published, but Sands’ work, which followed Sands et al.’s () study of Hadza phonology, is being continued by Kirk Miller. For publications on Hadza grammar, the reader is referred to Sands’s contributions in Vossen ().
. P
.................................................................................................................................. Following Traill’s () study of !Xóõ phonology, the use of the sound spectrograph, X-ray photography, and other tools of modern phonology has increased considerably. 2
Quellen zur Khoisan-Forschung (Research in Khoisan Studies), edited by Rainer Vossen. Dokumentation bedrohter Sprachen (Documentation of Endangered Languages), funded by the Volkswagen Foundation. 3
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Indeed, instrumental techniques are almost obligatory for the study of sounds as complex as clicks. Two studies in the UCLA Fieldwork Studies of Targeted Languages project dealt with Khoisan languages, Sandawe and Hadza, using instrumental techniques to good effect in resolving inconsistencies which had arisen in earlier studies. Collins and Bell () is a study of ǂHoan click types. One interesting result of this study is confirmation of a three-way contrast between plain aspirated clicks, clicks with aspirated uvular release (i.e. aspirated linguo-pulmonic clicks), and clicks with delayed aspiration. This pattern is not found in Ju languages or Khoe languages like Khoekhoe, Naro, and Khwe. But it is also found in Gǀui (Khoe), Nǀuu (!Ui branch of Tuu), and Western !Xoon (Taa branch of Tuu), suggesting the spread of phonetic features in an areal fashion, since these languages are roughly contiguous. Amanda Miller has contributed numerous studies on the phonetic properties of Khoisan languages, including studies of Juǀ’hoan tone, reduplication patterns in Juǀ’hoan, and unusual click types in the Ju languages. The last of these has expanded the traditional set of five clicks: labial, dental, palatal, alveolar, and lateral to include the forward released palatal which replaces the palatal click in Ekoka !Xun and the retroflex click first described by Doke (cf. Sands and Miller-Ockhuizen ). Particularly notable is Miller et al. (), a description of Nǀuu phonology which proposes a new interpretation of the articulatory mechanisms underlying the production of clicks. It has long been known that clicks require two closures, one anterior and one posterior. In the traditional interpretation, the posterior closure is velar. In the reinterpretation in Sands et al., the posterior closure varies according to the click type but is usually uvular or pharyngeal. Clicks which have traditionally been called uvular clicks are in the new interpretation regarded as contour consonants with a separate audible release of the back closure. There has been much discussion recently of whether the clicks should always be regarded as unitary consonants, no matter how complex, or in some cases be analyzed as clusters. For a summation of the arguments, see Güldemann ().
. H
.................................................................................................................................. There is still no consensus on the precise interrelationships of the Khoisan languages, although most researchers believe they fall into a number of unrelated families. Few accept Greenberg’s () original classification of all Khoisan languages, including Sandawe and Hadza, into a single Khoisan phylum. Starostin () has posted online an attempt at reconstruction. Ehret is preparing an etymological dictionary comparing Khoekhoe, Juǀ’hoan and Lone Tree !Xóõ. For current views on the classification of Khoisan, the reader is referred to Güldemann and Vossen (), Honken (), and Güldemann and Fehn (). Vossen () is the first complete reconstruction of any Khoisan family. His study is based on his own data for fourteen Khoe languages, few of which had been studied before, combined with the published data on Khoekhoe, Khwe, etc., and reconstructs several hundred stems for Proto-Khoe, as well as basic grammatical structure. Heine and Honken
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() provide evidence linking Ju with the ǂHoan language, formerly considered an isolate. Güldemann and Elderkin () propose that Kwadi and Sandawe are related to Khoe. Most historical studies in the Khoisan area deal with particular linguistic systems. Elderkin () and Honken (), for instance, reconstruct the Proto-Khoe tonal system. Sands and Miller-Ockhuizen () are concerned with evidence for reconstructing a “retroflex” click in Proto-Ju.
. D
.................................................................................................................................. The continuing growth of the Internet and the use of computers have opened up new avenues of research. New papers in many fields are now regularly posted on the Internet. The most significant event of this sort for Khoisan researchers is the digitization of the original notebooks of Wilhelm Bleek, Lucy Lloyd, and Dorothea Bleek (which have been posted at: http://lloydbleekcollection.cs.uct.ac.za). Also notable is Traill (), which has made available a number of sound recordings of extinct Khoisan languages on a CD, including a recording of Jopi Mabinda, the last speaker of the Tuu language ǁX’egwi (see Honken, chapter of this volume). A great deal of such archival material appears to exist, including recordings of speech and song, short films, and unpublished notes and papers. Traill (), for example, makes available R. Story’s unpublished paper on K’uǀhaːsi, almost the only material recorded on this language.
. S
.................................................................................................................................. A word must be said about sociolinguistic studies. Though the number of researchers who concentrate on this aspect of Khoisan is relatively small, their results may be of great significance to other branches of Khoisan studies. Obviously, languages do not exist in a vacuum. Boden and Barnard () as well as Hitomi Ono (e.g. ) have taken an interest in kinship systems. This was also a central concern of Barnard (). Akira Takada and Kazuyoshi Sugawara have published several studies of speech interaction and semantic concepts among the Gǀui. Recently, Sugawara and Jiro Tanaka () have compiled an encyclopedia of Gǀui-Gǁana culture, which combines a huge amount of cultural and sociological information with the results of Nakagawa’s linguistic research. Kemmonye Monaka and Laurentius Davids have also been primarily concerned with language use in a sociological context.
. T
.................................................................................................................................. The term “Khoisan” is generally maintained as a convenient way to refer to those languages of southern and eastern Africa which employ clicks as basic speech sounds and appear to be indigenous. That is to say, it does not include languages like Xhosa, which are thought to have borrowed the clicks.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Though some researchers still support Greenberg’s original classification of all these languages into one click family, the majority of researchers in the field agree that they comprise a number of unrelated families. Many of the surviving languages are spoken by diminished populations and the outlook for their continued survival is not hopeful. In some cases, Khoisan populations have adopted a more vigorous political strategy in an attempt to maintain their languages. The population of Khoisan researchers themselves remains small, but several projects have been initiated to document the more poorly known languages. The number of published grammars and dictionaries has been growing steadily. Considerably more is known about the Khoisan languages than was known fifty years ago, and the increased activity in Khoisan research has produced several surprises. One was the discovery by Traill of ǂHoan, a language which at the time could not be assigned to any of the traditional three families. Another was the discovery that Nǀuu, a !Ui language thought to be extinct, had still a small number of fluent speakers. The use of instrumental techniques in analyzing the clicks and the complex vowel sounds found in Ju and Tuu languages has greatly illuminated the phonologies of these languages and will in the long run be of inestimable value to historical linguistic research as well. A great loss was suffered by the field with the untimely deaths of Jan Snyman, Anthony Traill, Hans den Besten, and others. They will be missed. Fortunately, the Khoisan field has attracted a considerable number of young researchers and the present outlook for Khoisan studies remains bright.
R Barnard, A. (). Hunters and Herders of Southern Africa: A comparative ethnography of the Khoisan peoples. Cambridge: Cambridge University Press. Barnard, A., and Boden, G. (eds.) (). Southern African Khoisan Kinship Systems. Cologne: Rüdiger Köppe. Bleek, D. F. (). The Naron: A Bushman tribe of the Central Kalahari. Cambridge: Cambridge University Press. Bleek, D. F. (–). ‘Bushman grammar: a grammatical sketch of the language of the ǀxam-ka-!k’e’, Zeitschrift für Eingeborenen-Sprachen : –; : –. Bleek, D. F. (). Comparative Vocabularies of Bushman Languages. Cambridge: Cambridge University Press. Bleek, D. F. (a). ‘The Hadzapi or Watindega of Tanganyika territory’, Africa : –. Bleek, D. F. (b). ‘Traces of former Bushman occupation in Tanganyika territory’, South African Journal of Science : –. Bleek, W. H. I., and Lloyd, L. C. (). Specimens of Bushman Folklore. London: George Allen. Brenzinger, M., and König, C. (eds.) (). Khoisan Languages and Linguistics: Proceedings of the st International Symposium, January –, , Riezlern/Kleinwalsertal. Cologne: Rüdiger Köppe. Brugman, J. C. (). Segments, Tones and Distribution in Khoekhoe Prosody. Ph.D. thesis. Cornell University. Collins, C., and Bell, A. (). ‘ǂHoan and the typology of click accompaniments in Khoesan’, Cornell Working Papers in Linguistics : –. Collins, C., and Chebanne, A. (forthcoming). A Grammatical Sketch of Kuasi. Cologne: Rüdiger Köppe. Collins, C., and Gruber, J. S. (). A Grammar of ǂHȍã. Cologne: Rüdiger Köppe. Collins, C., and Gruber, J. S. (). A Grammar of ǂHȍã: Vocabulary and texts. Cologne: Rüdiger Köppe. Collins, C., and Namaseb, L. (). A Grammatical Sketch of Nǀ uuki with Stories. Cologne: Rüdiger Köppe.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Crawhall, N. (). Too Good to Leave Behind: The Nǀuu language and ǂKhomani people of Gordonia District. Ph.D. thesis. University of Cape Town. Dickens, P. (). English Juǀ’hoan, Juǀ’hoan English Dictionary. Cologne: Rüdiger Köppe. Dickens, P. (). A Concise Grammar of Juǀ’hoan, with a Juǀ’hoan English glossary and a subject index. Edited by R. Vossen and M. Biesele. Cologne: Rüdiger Köppe. Doke, C. M. (). ‘An outline of the phonetics of the language of the Chu: Bushmen of the northwest Kalahari’, Bantu Studies : –. Doke, C. M. (). ‘An outline of Khomani Bushman phonetics’, Bantu Studies : –. Eaton, H. (). The Grammar of Focus in Sandawe. Ph.D. thesis. University of Reading. Published online in as SIL e-book . Ehret, C., and Ehret, P. (eds.) (). A Dictionary of Sandawe, from the field collections of E. ten Raa with the advice of E. D. Elderkin. Cologne: Rüdiger Köppe. Elderkin, E. D. (). ‘Proto-Khoe tones in Western Kalahari’, in S. Ermisch (ed.), –. Ermisch, S. (ed.) (). Khoisan Languages and Linguistics: Proceedings of the nd International Symposium, January –, , Riezlern/Kleinwalsertal. Cologne: Rüdiger Köppe. Fehn, A.-M. (). A Grammar of Ts’ixa (Kalahari Khoe). Ph.D. thesis. Universität zu Köln. Gerlach, L. (). N!aqriaxe: The Phonology of an Endangered Language of Botswana. Wiesbaden: Otto Harrassowitz. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Gruber, J. (). ǂHoan vocabulary. Unpublished field notes. Güldemann, T. (). ‘Phonological regularities of consonant systems in genetic lineages of Khoisan’, in R. Vossen and W. H. G. Haacke (eds.), –. Güldemann, T., and Elderkin, E. D. (). ‘On external genealogical relationships of the Khoe family’, in M. Brenzinger and C. König (eds.), –. Güldemann, T., and Fehn, A.-M. (). Beyond ‘Khoisan’: Historical relations in the Kalahari Basin. Amsterdam and Philadelphia, PA: John Benjamins. Güldemann, T., and Vossen, R. (). ‘Khoisan’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –. Haacke, W. H. G. (). The Tonology of Khoekhoe (Nama/Damara). Cologne: Rüdiger Köppe. Haacke, W. H. G., and Eiseb, E. (). A Khoekhoegowap Dictionary, with an English Khoekhoegowab index. Windhoek: Gamsberg Macmillan. Hagman, R. S. (). Nama Hottentot Grammar. Bloomington, IN: Indiana University. Heikkinen, T. (). An Outline of the !Xũ Language, spoken in Ovamboland and West Kavango. Pretoria: African Languages Association of Southern Africa. Heine, B., and Honken, H. (). ‘The Kx’a family: A new Khoisan genealogy’, Journal of Asian and African Studies : –. Heine, B., and König, C. (). The Ekoka !Xun: A demographic and linguistic report. (Khoisan Forum, Working Paper .) Cologne: Institut für Afrikanistik, Universität zu Köln. Heine, B., and König, C. (). The !Xun Language: A dialect grammar of Northern Khoisan. Cologne: Rüdiger Köppe. Honken, H. (). ‘The split tones in Central Khoisan’, in S. Ermisch (ed.), –. Honken, H. (). ‘Genetic relationships: an overview of the evidence’, in R. Vossen (ed.), –. Kilian-Hatz, C. (). Khwe Dictionary. Cologne: Rüdiger Köppe. Kilian-Hatz, C. (). A Grammar of Modern Khwe (Central Khoisan). Cologne: Rüdiger Köppe. Köhler, O. (). ‘Les langues khoïsan’, in J. Perrot (ed.), Les langues dans le monde ancien et moderne. Part : Les langues de l’Afrique subsaharienne, edited by G. Manessy. Paris: Éditions du CNRS, –. Köhler, O. (). Die Welt der Kxoé-Buschleute im südlichen Afrika: Eine Selbstdarstellung in ihrer eigenen Sprache. Vol. : Die Kxoé-Buschleute und ihre ethnische Umgebung. Berlin: Dietrich Reimer. Köhler, O. (). Die Welt der Kxoé-Buschleute im südlichen Afrika: Eine Selbstdarstellung in ihrer eigenen Sprache. Vol. : Grundlagen des Lebens: Wasser, Sammeln und Jagd, Bodenbau und Tierhaltung. Berlin: Dietrich Reimer.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Köhler, O. (). Die Welt der Kxoé-Buschleute im südlichen Afrika: Eine Selbstdarstellung in ihrer eigenen Sprache. Vol. : Materielle Ausrüstung: Werden und Wandel. Wohnplatz und Buschlager. Berlin: Dietrich Reimer. Köhler, O. (). The World of the Khwe Bushmen in Southern Africa: A self-portrait in their own language. Vol. .: Family and Society. Customary Law: Norms and Practices. Text historically analysed and edited by G. Boden, with interlinear glosses and English translations by A.-M. Fehn, with assistance from T. Chedau. Berlin: Dietrich Reimer. König, C., and Heine, B. (). A Concise Dictionary of Northwestern !Xun. Cologne: Rüdiger Köppe. Krönlein, J. G. (). Wortschatz der Khoi-khoin (Namaqua-Hottentotten). Berlin: Deutsche Kolonialgesellschaft. Mathes, T. K. (). Consonant-Tone Interaction in the Khoisan Language Tsua. Ph.D. thesis. New York University. Miller, A. L., Brugman, J., Sands, B., Namaseb, L., Exter, M., and Collins, C. (). ‘Differences in airstream and posterior place of articulation among Nǀuu clicks’, Journal of the International Phonetic Association : –. Nakagawa, H. (). Aspects of the Phonetic and Phonological Structure of the Gǀui Language. Ph.D. thesis. Johannesburg: University of the Witwatersrand. Ono, H. (). ‘A comparison of kinship terminologies of West Kalahari Khoe: ǂHaba, Tshila, Gǀui, Gǁana, and Naro’, in R. Vossen and W. H. G. Haacke (eds.), –. Rust, F. (). Nama Wörterbuch (Krönlein redivivus), J. G. Krönlein’s Wortschatz der Khoikhoin, überarbeitet und ergänzt. Pietermaritzburg: University of Natal Press. Sands, B. (). ‘Ju subgroups based on phonological patterns’, in M. Brenzinger and C. König (eds.), –. Sands, B. E., Maddieson, I., and Ladefoged, P. (). ‘The phonetic structures of Hadza’, Studies in African Linguistics : –. Sands, B. E., and Miller-Ockhuizen, A. (). Comparative evidence for new click types in Northern Khoisan. Paper presented at the th Annual Meeting of the Linguistics Society of America, Chicago, Illinois. Snyman, J. W. (). An Introduction to the !Xũ (!Kung) Language. Cape Town: A. A. Balkema. Snyman, J. W. (). Žuǀ’hõasi Fonologie & Woordeboek. Cape Town and Rotterdam: A. A. Balkema. Snyman, J. W. (). ‘A preliminary classification of the !Xũũ and Žu’hõasi dialects’, in W. H. G. Haacke and E. D. Elderkin (eds.), Namibian Languages: Reports and papers. Cologne: Rüdiger Köppe, –. Starostin, G. (). ‘A lexicostatistical approach towards reconstructing Proto-Khoisan’, Mother Tongue : –. Steeman, S. (). A Grammar of Sandawe: A Khoisan language of Tanzania. Utrecht: LOT. Sugawara, K., and Tanaka, J. (eds.) (). An Encyclopedia of ǀGui and ǁGana Culture and Society. Kyoto University: Laboratory of Cultural Anthropology. Traill, A. (). Phonetic and Phonological Studies of !Xóõ Bushman. Hamburg: Helmut Buske. Traill, A. (). A !Xóõ Dictionary. Cologne: Rüdiger Köppe. Traill, A. (). Extinct: South African Khoisan Languages. (Compact disk and booklet.) Johannesburg: Department of Linguistics, University of the Witwatersrand. Traill, A. (ed.) (). R. Story, K’uǀ ha:si Manuscript. (Khoisan Forum, Working Paper .) Cologne: Institut für Afrikanistik, Universität zu Köln. Visser, H. (). Naro Dictionary: Naro–English, English–Naro, th edn. Gantsi, Botswana: Naro Language Project. Vossen, R. (). Die Khoe-Sprachen: Ein Beitrag zur Erforschung der Sprachgeschichte Afrikas. Cologne: Rüdiger Köppe. Vossen, R. (ed.) (). The Khoesan Languages. London and New York: Routledge. Vossen, R., and Haacke, W. H. G. (eds.) (). Lone Tree—Scholarship in the Service of the Koon: Essays in memory of Anthony T. Traill. Cologne: Rüdiger Köppe.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
.
M languages spoken on the African continent today belong to one of the major phyla, Afro-Asiatic, Nilo-Saharan, or Niger-Congo. Within each of these stocks, there are usually clearly defined branches themselves covering major areas. Within Afro-Asiatic, for example, the Cushitic and Chadic branches are spoken across large areas; Niger-Congo subgroups such as Bantu or Kwa, or Nilo-Saharan branches such as Central Sudanic or Nilotic are also spread across wider zones. Next to these three major phyla, there are language families, sometimes also covering major areas, which appear to constitute independent stocks. Mande and Ubangian, for example, are classified as members of the Niger-Congo phylum by Greenberg (), but are probably better treated as independent families (see Dimmendaal ). The same applies to Ijo plus Defaka (in Nigeria). Also, Dogon is probably better classified as an independent stock rather than as part of Kwa or another Niger-Congo branch (e.g. Gur, as in Greenberg ). In all cases, the actual grammatical evidence for a genetic affiliation with the Niger-Congo phylum is rather slim. This also applies to a group of languages grouped under the name Tumtum and classified as Kordofanian (i.e. Niger-Congo) by Greenberg (: ). According to Schadeberg (), this group, referred to by the author as Kadugli (after the biggest town in the area), is not part of Kordofanian, or Niger-Congo for that matter. The Songhay cluster (spoken mainly along the Niger River in West Africa) and Koman (in the border area between Ethiopia and Sudan), which were classified as members of the Nilo-Saharan phylum by Greenberg (), again are better treated as independent families (see Dimmendaal, chapter of this volume). Whereas Greenberg () also grouped Northern Khoisan, Central Khoisan, and Southern Khoisan, plus Sandawe and Hadza into one phylum, specialists for these languages have been deconstructing his classification since, and prefer to treat the first three groups as independent language families (cf. Vossen ; Güldemann and Fehn ). Central Khoisan and the virtually extinct Kwadi language in Angola probably constitute a genetic unit with Sandawe in Tanzania (Güldemann and Elderkin ). But Hadza (also known as Kindiga or Tindiga) is one of several languages on the African continent with no obvious relatives, probably constituting an isolate (Sands , ). As is true for other isolates on the continent, it is surrounded by language families themselves constituting expansion zones, more specifically (Southern) Cushitic, (Southern) Nilotic, and (Eastern) Bantu.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Farther towards the north, in the linguistically complex southwestern corner of Ethiopia another isolate is found, Ongota (also known as Biraile), a language first described by Teferra () and Mikesh and Seelig (). There are only a few speakers left for this language, as most ethnic Ongota have shifted towards the neighboring Cushitic language Tsamakko as their first language. As is the case for other isolates, Ongota shares features with neighboring Omotic and Cushitic (i.e. Afro-Asiatic) languages as well as with NiloSaharan. Correspondingly, investigators of this language have argued that it belongs to Nilo-Saharan (e.g. Blažek ) or Afro-Asiatic (Fleming ). The most detailed accounts of Ongota are to be found in Savà and Tosco (, , ), who also prefer to treat this language as a linguistic isolate. The same authors are in the process of documenting Ongota with what appears to be the last full speakers of this language. The name Shabo refers to a language spoken by another ethnic minority in southwestern Ethiopia who call themselves Sabu. The alternative name by which the Sabu or their language is known, Mikeyir, derives from the language of the Majang with whom they live in a symbiotic relationship and who speak a Surmic (i.e. Nilo-Saharan) language. The Shabo also interact with the Sheko, whose language belongs to Omotic (Afro-Asiatic). Bender (: ) points out that initial data on this language were collected by Harvey Hoekstra. More recently, Teferra and Unseth () presented data on the Shabo language. As shown by Schnoebelen (), no convincing “phylogenetic” signal emerges when trying to classify this language. The area where languages like Shabo or Ongota are spoken constitutes a residual or accretion zone, i.e. an area characterized by a high degree of genetic and structural diversity less affected by expansion zones such as Cushitic, Nilotic, or Surmic. Apart from the Nilo-Saharan language Bertha and the Koman languages, Gumuz is spoken in this area. This dialect continuum spoken by around , people in western Ethiopia and southeastern Sudan, and classified as a member of the Koman group within Nilo-Saharan by Greenberg (), is treated as an isolate by the present author, since it lacks the core grammatical features that define Nilo-Saharan and its limits. Across the Sahel zone towards West Africa, a number of other isolates are found, again spoken along the edge of expansion zones. One example of such an isolate surrounded by language families covering large areas is Laal, a language spoken by around people in three villages in southern Chad. The language name derives from láà:l, the characteristic speech of Gori (known as lá in Laal), one of the villages where this language is spoken. Boyeldieu (, a, b, ) gives a detailed account of Laal, and points out that its speakers are in permanent contact with people surrounding the Laal area and speaking languages belonging to Chadic (Afro-Asiatic) and Adamawa (Niger-Congo); numerous lexical borrowings from the Bua languages attest to influence from this Adamawa cluster, for example. More recently, the Central Sudanic language Bagirmi, a lingua franca of the area (and language of the former Bagirmi kingdom) has become a dominant language. Typologically, Laal shares properties with different language families surrounding it. Internal morphology (e.g. with pluractionality marking) with verb stems is also found in neighboring Adamawa and Chadic languages. Next to a plural marking number suffix for nouns, root-internal changes occur, as in neighboring Adamawa as well as Chadic languages. The singulative number suffix with nouns again is reminiscent of Northeastern Nilo-Saharan (see Dimmendaal, chapter of this volume), although the actual form -aL does not appear to be attested there. Nouns and pronouns are furthermore characterized by
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
a distinction between masculine, feminine, and neuter gender, a typological property also attested in Ubangian languages farther towards the east, but also in the Kadu languages along the southern range of the Nuba Mountains, Sudan, as well as the linguistic isolate Shabo in southwestern Ethiopia (and, of course, Eastern Nilotic and Afro-Asiatic languages). However, in all cases the similarities are typological in nature, i.e. there does not appear to be any evidence that the actual forms are cognate. Interesting from a typological point of view is the fact that a formal distinction occurs between masculine and feminine pronouns for the first person but not for the second person singular in Laal. Languages may remain unclassified not only due to a lack of reliable data but also due to the confounding influence of language contact. This probably applies to another language spoken in Chad as well as Sudan, Kujarge, which may belong to Chadic, although its pronominal system does not support such a claim. In the linguistic fragmentation belt of northeastern Nigeria, another isolate on the brink of extinction, Jalaa, has been identified by Kleinewillinghöfer (), who observes that the major part of the lexicon seems to differ entirely from all the surrounding languages, which themselves belong to different language families. These surrounding languages are part of Adamawa, Benue-Congo, and Chadic, as well as Atlantic (namely Fulfulde, a language today spoken in a zone ranging from Senegal in the west to Ethiopia in the east). Several borrowings in neighboring Adamawa languages like Tso attest to the former presence of “another language related to Jalaa, which has already vanished” (Kleinewillinghöfer : ). One reason why these neighboring languages borrowed from Jalaa (and possibly other members of a language family of which Jalaa is now the last representative) was the need to replace lexemes “which had also been used as a name of someone who had died” (Kleinewillinghöfer : ). Mainly due to the pioneering work of the British anthropologist and linguist Roger Blench, a number of other linguistic isolates have been identified, all situated in West Africa. Mpra (Mpre) is the name of a language without full or semi-speakers left today which once was spoken in the village of Butei, north-central Ghana. It was probably first referred to in the literature by Cardinall (). As is common with vanishing languages, these “remembers” are only capable of producing individual words and parts of some fixed expressions, e.g. the first clauses of proverbs (Blench a). Baŋgi Me is a linguistic isolate spoken in seven villages in northern Mali by people who call themselves Baŋgi Na. It was probably first reported by Bertho (), who referred to the language as Yeni, after one of the villages where this language is spoken. As is the case with other isolated languages discussed here, Baŋgi Me shares typological similarities with neighboring language groups, more specifically with Dogon (assumed to constitute an independent stock by the present author) and to a lesser extent with the Mande language Bozo (Blench ). The analytical problem linguists interested in genetic classification are confronted with in all cases listed above is the fact that these languages do not have any obvious close relatives, which would otherwise help to identify a first set of inherited lexicon and grammar. It is difficult to arrive at a uniform classification because the lexical and grammatical similarities with other languages or language families quite often do not allow us to distinguish between borrowed and inherited material. Dimmendaal () discusses the hypothetical case of English as the only remaining Indo-European language, with all traces of other members of this language family having vanished. Around % of
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
the English lexicon apparently is non-Germanic in origin. In the case of this language, one might still be able to identify similarities with Semitic languages like Arabic, e.g. algebra, from Arabic al-jabr ‘the science of reuniting’, or tomato and potato (or their reflexes), from Amerindian families like Uto-Aztecan and Carib, respectively. But, as pointed out in Dimmendaal (: ): [l]anguages like Laal or Ong’ota . . . manifest themselves to us in an essentially a-historical Gestalt. The various similarities to Nilo-Saharan and Afroasiatic languages in the area could be due to borrowing or shared inheritance from one of the two phyla. But we probably will not be able to show which hypotheses are more plausible, unless a range of morphological elements can be shown to be cognate with widespread morphological elements in either NiloSaharan or Afroasiatic or some other language phylum.
When plotting linguistic isolates on the African continent on a map (Map .), it becomes clear that many of them are spoken along the edge of areas occupied by language families with a relatively low degree of internal genetic diversity covering large geographical areas, i.e. at the edge of spread zones (in the sense of Nichols ). This in turn suggests that within these expansion zones linguistic isolates probably disappeared as a result of language shift. Indirect evidence for this scenario derives from lexical substrates in languages belonging to expansion zones, as the following example may help to illustrate. The expansion of the Nilotic and Surmic branch within Eastern Sudanic (Nilo-Saharan) is the result probably of the dominant role played by pastoralism, which had a dramatic effect on the mobility of speech communities. One submerged group referred to in oral traditions of speakers of the Nilotic language cluster Teso-Turkana is Oropom, purportedly once spoken in the border area between northeastern Uganda and northwestern Kenya where TesoTurkana languages are spoken today. Wilson () presents around words, but unfortunately no data are available on the grammar of Oropom. When studying this wordlist, one can observe similarities between Oropom and Turkana related to, for example, color, such as kopurat ‘red’ (Turkana -kɪpʊrat ‘violet, pink’), pusia ‘blue’ (Turkana -pʊsɪ ‘green, blue’), or meri ‘spotted’ (Turkana -mεrɪ ‘spotted’), but also with isolates such as Hadza or Shabo; compare the root for ‘ear’ in Oropom, ki-ito, and Shabo kiti. But this still leaves us with a core of vocabulary (e.g. for ‘cow’, ‘dog’, ‘enemy’, ‘meat’, ‘nose’, ‘tree’, ‘water’) which does not show any lexical similarity to other languages or language families in the area, including the neighboring Kuliak (Rub) languages, which form an isolated branch within Nilo-Saharan (as argued in Dimmendaal, chapter of this volume). Whereas the subsistence economies of several African communities speaking isolated languages have been described in terms of hunting and gathering, these do not necessarily represent continuations of early stages in the development of humankind. Kuliak languages in northeastern Uganda, for example, are spoken by people for whom hunting and gathering is part of their daily activities. But since a lexicon related to the “bovine idiom” can be reconstructed for their common ancestor (Heine ), it is most likely that the ancestral community practiced pastoralism, in particular since the terms involved do not appear to be borrowed from neighboring language families like Nilotic or Surmic. Another cause for the loss of linguistic diversity (apart from the introduction of pastoralism) probably is to be sought in the expansion of agriculturalists into territories formerly inhabited by hunter-gatherers. The expansion of languages spoken by communities with new
. African linguistic isolates
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
means of subsistence or technological innovations often results in language shift among such communities. The expansion of Bantu languages into Central, Eastern, and Southern Africa constitutes a particularly telling example. The only non-Bantu language spoken in, for example, Gabon today probably is Baka, an Ubangian language. In addition, there is Irimba, a language whose wider affiliation still needs to be clarified (Hombert and Philippson ). Technological and social innovations in combination with climate change probably had an important effect on the linguistic map of Africa (Dimmendaal ). The spreading of the Nubian branch within Nilo-Saharan is a case in point. Today, Nubian languages are spoken in the Darfur region of Sudan and the Nuba Mountains in Central Sudan, as well as along the Nile in northern Sudan and southern Egypt. Their diffusion most likely is related to dramatic climatological changes, more specifically the disappearance of a major riverine system (the Yellow Nile or Wadi Howar) in the area, which forced people to migrate to less arid zones (Rilly ; Dimmendaal : –). The list of linguistic isolates presented above is not necessarily exhaustive, as specific areas are still poorly studied. Other potential candidates include Gomba, Weyto, and Rer Bare, all in Ethiopia. Although these are listed as extinct in Ethnologue, a recent survey by Ethiopian linguists has shown that there may still be speakers around. Apart from these and other names listed above, Hombert and Philippson () mention a number of other languages possibly constituting isolates. One of these, Meroitic, clearly belongs to the Northern branch of Eastern Sudanic (Nilo-Saharan), as shown by Rilly (). The name Kara in their list refers to different ethnic groups and their languages in Chad as well as across the border in Sudan and the Central African Republic (Nougayrol ). Kara of Birao (a sub-prefecture in Chad) is listed as an isolate in Ethnologue, but probably belongs to the Central Sudanic branch within Nilo-Saharan. The three languages called Mimi in Chad have also caused problems for classification. One has turned out to be closely related to Fur, and is also known as Amdang. The genetic position of Mimi of Decorse (after the person who collected a wordlist around published by Gaudefroy-Demombynes : –) and Mimi of Nachtigal (after the explorer who reported its existence) is unclear, because the few lexical similarities with the Maban group within Nilo-Saharan may well be due to borrowing. Either or both of these may, therefore, constitute a linguistic isolate; see Starostin () for a balanced account. There are, in fact, a range of such languages constituting either linguistic isolates or unclassified members within an established family. Blench () mentions Ega, a language spoken in the western part of Ivory Coast. Dompo is another language of unknown provenance in Ghana showing lexical similarities to neighboring Kwa languages, but also to Senufo, another relatively isolated group of Niger-Congo languages (Blench b). According to Dave Roberts (p.c.), Meyobe (also known as Solla) is an endangered language spoken in Togo and Benin, showing lexical similarity to Oti-Volta (Gur) languages as well as Togo Mountain (Kwa) languages. Bayot in Senegal is another example of a language whose wider genetic affiliations are unclear. In the absence of (extensive) data on the grammatical structure of such languages, it is not possible to solve the issue. A further analytical complication for genetic classification is due to extensive borrowing also of basic vocabulary as a result of taboo on the pronunciation of certain words, as the Jalaa borrowings in Adamawa languages discussed above make clear. Storch () discusses manipulated languages in an African context, and shows that deliberate language change in fact is extremely common. It is also common in language
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
shift situations to transfer specialized lexicon from the vanishing source language into the target language. Such multiple layers, of course, complicate genetic classifications, especially in the absence of grammatical data. At times, only indirect traces tell us of languages which have vanished otherwise, as with the numerous cases discussed in the literature of substrate languages in this respect, for example in specialized lexicon of unknown genetic affinity in Bantu, Central Sudanic and Ubangian languages spoken by Pygmy groups, or the Austronesian language Malagasy. Connell (, ) mentions different languages whose genetic affiliation remains unclear due to a paucity of data, including Luo (Cameroon), which is only known through song texts. These and other cases mentioned by him and other authors in the literature are listed in Table ..
Table . Languages possibly constituting linguistic isolates Language name
Classification according to Greenberg ()
Bangi Me (Mali)
not listed
Bayot (Senegal)
not listed
Dompo (Ghana)
not listed
Ega (Ivory Coast)
not listed (possibly Kwa)
Gomba (Ethiopia)
not listed
Gumuz (Ethiopia, Sudan)
Nilo-Saharan
Hadza (Tanzania)
Khoisan
Irimba (Gabon)
not listed
Jalaa (Nigeria)
not listed
Kujarge (Chad)
not listed
Laal (Chad)
not listed
Lufu (Nigeria)
not listed (possibly Jukunoid, Benue-Congo)
Luo (Nigeria)
not listed
Mawa (Nigeria)
not listed
Meyobe (Benin, Togo)
not listed
Mimi of Decorse and Mimi of Nachtigal
Nilo-Saharan
Mpra (Ghana)
not listed
Oblo
not listed (possibly Adamawa)
Ongota (Ethiopia)
not listed
Oropom (Kenya, Uganda)
not listed
Rer Bare (Ethiopia)
not listed
Shabo (Ethiopia)
not listed
Weyto (Ethiopia)
not listed
Wutana (Nigeria)
not listed
Yeni (Cameroon)
not listed (possibly Mambiloid, Benue-Congo)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The resulting linguistic map is somewhat different from what one of the pioneers of the genetic classification of African languages, the late Joseph H. Greenberg, proposed. Apart from three major phyla, there are at least nine smaller families probably constituting independent language families, and an unknown number of isolates as the last representatives of additional language families which have become extinct otherwise. In this respect, the linguistic map of Africa is not that different from Eurasia (with families covering large areas next to families with a more restricted geographical spreading, and isolates such as Basque, Burushaski, Etruscan, Kusunda, or Sumerian) or other parts of the world.
R Bender, M. L. (). ‘The Surma language group: a preliminary report’, Studies in African Linguistics Supplement : –. Bertho, J. (). ‘La place des dialectes dogon de la falaise de Bandiagara parmi les autres groupes linguistiques de la zone soudanaise’, Bulletin de l’IFAN : –. Blažek, V. (). ‘Nilo-Saharan stratum of Ongota’, in D. Payne and M. Reh (eds.), Advances in NiloSaharan Linguistics. Cologne: Rüdiger Köppe, –. Blench, R. (). The Ega language of Côte d’Ivoire: etymologies and implications for classification, http://www.rogerblench.info/Language/Niger-Congo/Kwa/Ega%data.pdf accessed July , . Blench, R. (). Baŋgi Me, a language of unknown affiliation in northern Mali, http://www. rogerblench.info/Language/Isolates/Bangime%wordlist%paper.pdf accessed July , . Blench, R. (a). Recovering data on Mpra [=Mpre], a possible language isolate in north-central Ghana, http://citeseerx.ist.psu.edu/viewdoc/download?doi=....&rep=rep&type=pdf accessed July , . Blench, R. (b). The Dompo language of central Ghana and its affinities, http://www.ddl.cnrs.fr/ projets/clhass/PageWeb/ressources/Isolats/Dompo%Blench%.pdf accessed July , . Boyeldieu, P. (). ‘Éléments pour une phonologie du laal de Gori (Moyen-Chari)’, in J.-P. Caprile (ed.), Études phonologiques tchadiennes. Paris: SELAF, –. Boyeldieu, P. (a). Deux études laal (Moyen-Chari, Tchad). Berlin: Dietrich Reimer. Boyeldieu, P. (b). ‘Quelques questions portant sur la classification du laal (Tchad)’, in H. Jungraithmayr (ed.), The Chad Languages in the Hamitosemitic-Nigritic Border Area (Papers of the Marburg Symposium, ). Berlin: Dietrich Reimer, –. Boyeldieu, P. (). ‘Détermination directe/indirecte en laal’, in P. Boyeldieu (ed.), La maison du chef et la tête du cabri: Des degrés de la détermination nominale dans les langues d’Afrique centrale. Paris: Geuthner, –. Cardinall, A. W. (). ‘A survival’, Gold Coast Review (): –. Connell, B. (). ‘Moribund languages of the Nigeria Cameroon borderland’, in M. Brenzinger (ed.), Endangered Languages in Africa. Cologne: Rüdiger Köppe, –. Connell, B. (). ‘Endangered languages in Central Africa’, in M. Brenzinger (ed.), Language Diversity Endangered. Berlin: Mouton de Gruyter, –. Dimmendaal, G. J. (). ‘Do some languages have a multi-genetic or non-genetic origin? An exercise in taxonomy’, in R. Nicolaï and F. Rottland (eds.), Proceedings of the Fifth Nilo-Saharan Conference, Nice . Cologne: Rüdiger Köppe, –. Dimmendaal, G. J. (). ‘Language ecology and linguistic diversity on the African continent’, Language and Linguistics Compass (): –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Fleming, H. (). Ongota: A Decisive Language in African Prehistory. Wiesbaden: Harrassowitz.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Gaudefroy-Demombynes, M. (). ‘Documents sur les langues de l’Oubangui Chari’, in Actes du XIVe Congrès des Orientalistes (Alger ). Part II. Paris: Ernest Leroux, –. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University; The Hague: Mouton. Güldemann, T., and Elderkin, E. D. (). ‘On external genealogical relationships of the Khoe family’, in M. Brenzinger and C. König (eds.), Khoisan Languages and Linguistics: Proceedings of the st International Symposium, January –, , Riezlern/Kleinwalsertal. Cologne: Rüdiger Köppe, –. Güldemann, T., and Fehn, A.-M. (eds.) (). Beyond ‘Khoisan’: Historical relations in the Kalahari basin. Amsterdam and Philadelphia, PA: John Benjamins. Heine, B. (). The Kuliak Languages of Eastern Uganda. Nairobi: East African Publishing House. Hombert, J.-M., and Philippson, G. (). The Linguistic Importance of Language Isolates: The African Case. Unpublished manuscript. Kleinewillinghöfer, U. (). ‘Jalaa—an almost forgotten language of northeastern Nigeria: a language Isolate?’, in D. Nurse (ed.), Historical Language Contact in Africa (Special issue of Sprache und Geschichte in Afrika. Vol. /). Cologne: Rüdiger Köppe, –. Mikesh, P., and Seelig, J. M. (). ‘Ongota or Birale: a moribund language of Gemu-Gofa (Ethiopia)’, Journal of Afroasiatic Languages (): –. Nichols, J. (). Linguistic Diversity in Space and Time. Chicago, IL: University of Chicago Press. Nougayrol, P. (). Les parlers gula, Centrafrique, Soudan, Tchad: Grammaire et lexique. Paris: Editions du CNRS. Rilly, C. (). Le méroitique et sa famille linguistique. Leuven and Paris: Peeters. Sands, B. E. (). ‘The linguistic relationship between Hadza and Khoisan’, in M. Schladt (ed.), Language, Identity, and Conceptualization among the Khoisan. Cologne: Rüdiger Köppe, –. Sands, B. E. (). ‘Africa’s linguistic diversity’, Language and Linguistics Compass (): –. Savà, G., and Tosco, M. (). ‘A sketch of Ongota, a dying language of southwest Ethiopia’, Studies in African Linguistics (): –. Savà, G., and Tosco, M. (). ‘The classification of Ongota’, in M. L. Bender, G. Takács, and D. L. Appleyard (eds.), Selected Comparative-Historical Afrasian Linguistic Studies in Memory of Igor M. Diakonoff. Munich: Lincom Europa, –. Savà, G., and Tosco, M. (). ‘Review of Fleming ()’, Aethiopica : –. Schadeberg, T. C. (). ‘The classification of the Kadugli language group’, in T. C. Schadeberg and M. L. Bender (eds.), Nilo-Saharan: Proceedings of the First Nilo-Saharan Linguistics Colloquium, Leiden, September –, . Dordrecht and Cinnaminson, NJ: Foris, –. Schnoebelen, T. (). ‘(Un)classifying Shabo: phylogenetic methods and results’, in P. Austin, O. Bond, M. Charette, D. Nathan, and P. Sells (eds.), Proceedings of Conference on Language Documentation and Linguistic Theory . London: SOAS, –. Starostin, G. (). ‘On Mimi’, Journal of Language Relationship : –. Storch, A. (). Secret Manipulations: Language and context in Africa. Oxford: Oxford University Press. Teferra, A. (). ‘A sketch of Shabo grammar’, in M. L. Bender (ed.), Proceedings of the Fourth NiloSaharan Conference, Bayreuth, Aug. –September , . Hamburg: Helmut Buske, –. Teferra, A., and Unseth, P. (). ‘Toward the classification of Shabo (Mikeyir)’, in M. L. Bender (ed.), Topics in Nilo-Saharan Linguistics. Hamburg: Helmut Buske, –. Vossen, R. (ed.) (). The Khoesan Languages. London and New York: Routledge. Wilson, J. G. (). ‘Preliminary observations on the Oropom people of Karamoja, their ethnic status, culture, and postulated relation to the peoples of the Late Stone Age’, The Uganda Journal (): –.
.............................................................................................................
LANGUAGE STRUCTURES CASE STUDIES .............................................................................................................
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
- ......................................................................................................................
.
. I
.................................................................................................................................. T brief sketch introduces the Bom-Kim language, a complex of at least two closely related and highly endangered varieties spoken on the southern coast of Sierra Leone. From a linguistic perspective Kim and Bom represent two (mutually intelligible) dialects of a single language, despite having two separate ISO codes: krm and bmf. It is possible they form a dialect continuum with Sherbro and other local varieties spoken in the same general area and thus represent a typical variant of a larger set of genetically related languages. The two varieties are spoken today by a limited number of isolated and venerable residents in the area. The two varieties are more distinct socio-historically (there are land wars in the remembered past) and politically (separate chiefdoms) than they are linguistically. Speakers say that the languages are different but also admit that they can understand each other. Linguistic analysis showed some limited lexical and grammatical differences between the two. In the phonology, for example, both varieties have been affected by a diachronic rule, l > n __#, which marks them as distinct from other closely related languages (Sherbro, Mani, and Kisi), but the rule is not so well established in Bom, since there are still some words ending in [l], making it more similar to Sherbro, to which it is geographically closer. Another example is found in the morphosyntax: the locative postposition ko in Kim is used also as a post-verbal particle realized always as [go] in Kim because of its near clitic-like status, a distribution not found in Bom. Both languages are moribund, Kim the more so of the two: there were twelve speakers in at the end of a documentation project (Documenting the Kim and Bom languages of Sierra Leone; or DKB). Most of them were found in the twin towns of Nyandehun and Tamuke. The only previous documentation of Bom-Kim is a brief grammatical sketch of Kim (“Krim” in Pichl ). Bom speakers were more numerous, likely numbering some several hundred, but were scattered throughout the Bom chiefdom, with small concentrations in the towns of Sampor and Sogbaleh. All Bom-Kim speakers are over and no children are learning the language. All speakers are bilingual in Mende, an unrelated Mande language, one of the major languages of Sierra Leone. A few men know Krio, the resident pidgin, and virtually all speakers are illiterate.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. South Atlantic A. Mel languages
1. Temne (tem), Pukur (Baga Binari; bcg), Baga Manduri (bmd), Baga Tchitem (Sitemu; bsp), Lɔŋiʧ (Baga Fore, Mbulungish; mbv), Landoma (Kogoli; ldm) 2. Bolom languages: Kisi (Northern Kisi (kqs), Southern Kisi (kss)), Mani (Bullom So; buy), Sherbro (bun), Bom-Kim (bmf and krm) 3. Gola (gol)
B. Limba (East Limba (lma) and West-Central Limba (lia)) C. Mansoanka (Sua; msw)
. South Atlantic classification (Voeltz ; Childs )
Bom-Kim is classified as in Figure .. South Atlantic is one of the early branches off the Niger-Congo stock and is now considered separate from North Atlantic and the isolate Bijogo (Blench ), all of which were once combined in (West) Atlantic (Wilson ). The display of South Atlantic provides other well-known names and ISO codes after each language in parentheses; extinct languages are not shown. As can be seen by the ISO codes, a number of South Atlantic varieties have been overly differentiated, not only Bom-Kim but also Kisi and perhaps Limba. It is also the case that the Ethnologue figures (Lewis ) overstate the number of Bom-Kim speakers (, Bom speakers and Kim speakers).
. P
.................................................................................................................................. The phonemic inventory of Bom-Kim features an areally common set of consonants and vowels, and a relatively modal phonemic inventory (Maddieson ), as illustrated in Table .. There is a voiced labialvelar stop in addition to four pairs of stops contrasting in voicing at different places of articulation, several voiceless fricatives, a liquid, and two glides. The feature [nasal] is important for both a series of nasals (at four places of articulation) and a series of prenasalized stops (at several places of articulation), as well as for several phonological processes. The sounds in parentheses represent peripheral phonemes. Orthographic conventions follow general Africanist practice: “ny” represents the palatal nasal [ɲ]; “c” and “j” are the affricates [ʧ] and [ʤ]. Phonemes in parentheses are of limited distribution. The vowel inventory (Table .) features a symmetrical seven-vowel system, three front and back pairs, with only the low vowel /ɑ/ lacking a counterpart. Length is contrastive for the monophthongs but not for the four diphthongs, which pattern phonologically with long vowels. The one non-IPA orthographic convention in the representation of vowels is the symbol “a” for IPA [ɑ]. Following general Africanist practice once again and avoiding diacritics, short vowels are represented with a single symbol, thus “e”, and long vowels with a double symbol “ee”.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Table . Bom-Kim consonantal inventory Labial
Dental
Palatal
Velar
Nasals
m
n
ny
ŋ
Stops
pb
td
cj
kg
nt nd
(nyj)
ŋk ŋg
Pre-nasal (mp) mb Fricatives
f (v)
Labiovelar Glottal gb
s
Liquids
h
l
Glides
j
w
Table . Bom-Kim vowel inventory i
u e
o ε
ɔ a
Diphthongs: ɔi, ai, ui, uε, ua
Unconditioned alternations between front and back vowels occur, especially with the high vowels, e.g., bami/bamu ‘my father’, gbeŋ/gboŋ ‘finish’. Such alternations are found in other closely related languages within the Bolom subgroup. Other alternations in vowel quality occur, e.g., bεsi- ‘broom’ and basi ‘sweep’, ye/ya ‘dance’, and caŋ/cεŋgi ‘pass’, yet none with any systematicity. In contrast, such alternations in Kisi have grammatical significance, e.g., the vowel [ɑ] changes to [e] when verbs with the vowel are negated. Syllable structure is CV(V)(C) where the only the liquid /l/ and nasals are allowed in syllable codas. The distribution differs slightly at the ends of the words, where only nasals are allowed, excluding the palatal nasal. Only nasals can be syllabic at the beginning of words, always before a homorganic stop. Onset “strengthening” occurs when a word ends in a nasal and precedes a word beginning with [l]. In this environment the [l] changes to [d], as also occurs in Mani and Kisi. In some cases no [l] need be present for the [d] to appear: jὲndàgbὲn ~ jὲndàgbὲndí ‘story’. Other notable segmental changes in Bom-Kim involve intervocalic “weakening”, particularly with regard to the voiced velar stop: [g] alternates with [ɣ] and [w] and even with zero in some cases. The liquid [l] also often disappears in this position. Occasionally an (optional) epenthetic [i] will occur at the end of words with a final closed syllable, e.g., gbahun(i) ‘door’ (and ‘story’ above). With regard to prosody, lexical tone can be reconstructed for South Atlantic; yet BomKim retains little evidence of such marking. Only a few inflectional morphemes, e.g., the negative marker, and some particles, e.g., the definite article, show the trace of what was probably once a high tone. Tone has not been marked on the following examples except where relevant to the discussion. As mentioned above when discussing the consonantal inventory, nasality functions prominently in the phonology. In addition to using [nasal] as a contrastive feature,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Bom-Kim has a notable process of anticipatory vowel nasalization when a coda is filled with a nasal, heaviest when the nasal is the velar nasal [ŋ] and the vowels are low. In addition, vowels are heavily nasalized after [h] (“rhinoglottophilia” in Matisoff ), particularly the low vowel [a] and the diphthong [ai].
. W
.................................................................................................................................. Nouns and verbs are the primary lexical categories; (underived) adjectives comprise a small class with fewer than twenty members. There is also a small set of adverbs and a few ideophones. Bom-Kim has both prepositions and postpositions along with a small set of connectives. The number system is five- and twenty-based (quinary-vigesimal).
.. Nouns, nominal morphology, and pronouns Nouns are the easiest word category to identify because of the pervasiveness of a noun class system, which features a salient and closed set of affixes to mark nouns and dependent elements. All nouns belong to one of nine classes, membership in which is typically signalled by a prefix on nouns but always determines agreement markers on dependent elements such as adjectives and low numbers. Not all nouns carry such marks, but all control agreement. Relevant semantic features are number, shape, or configuration, and animacy, with animacy often overriding class membership when the two conflict (“animate concord”). As with most noun class languages, there is overlap between the personal pronouns and the noun class pronouns. The third person personal pronouns, singular and plural, are identical to the singular and plural animate classes of the noun class system. Another morphosyntactic criterion identifying nouns is the almost ubiquitous presence of a following definite article. Nouns rarely appear without the cliticized article: /-ε/ in Kim vs. /-lε/ in Bom, a difference probably arising from l-weakening in intervocalic position. Inalienable nouns always appear with a possessive pronoun. This subcategory includes family relations, e.g., wante mi/wante mu ‘my/your sister’, and body parts, e.g., cεmi/cεmu ‘my/your chest’. As might be surmised, Bom-Kim is a head-initial language, with dependent elements following the head noun. The order of elements is: -()-()-, as in the example below. ()
1
The Bom-Kim noun phrase pεntimulε, pεntimu toondε1 pεnti-mu-lε pεnti-mu toon-lε brother-- brother- small- ‘your brother, your small brother’
(B.)
The [d] in the citation form shows the strengthening of [l] after nasals described above. Generally speaking, the first line in the examples represents something closer to a phonetic transcription than the morpheme-by-morpheme analysis in the second line.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Table . Personal pronouns in Bom-Kim Subject
Object
Possessive
Emphatic
ya, a
mi, -(i)m
mi, -(i)m
yaŋ, yan
n, mu, hu(m)
mu
mu
mun
wɔ, ɔ
wɔ
wɔ
wɔn, wɔŋ
yi, (h)i
(h)i
(h)i
hin
ha
nu, hum
nu
han
(h)a
ha
(h)a
han
Personal pronouns but not nouns show case distinctions (nominative vs. objectivepossessive). Case is marked, however, on only the first and second person singular pronouns. Other formal differences in the pronouns are phonological, conditioned by syllable structure (see the discussion of example () below), or morphophonological, as with the addition of the final emphatic nasal. All pronouns have an emphatic form with a final nasal, e.g., non-emphatic ya ‘’ vs. yaŋ/yan ‘.’. Table . shows some of these alternations. The phonological alternations are predictable on the basis of syllable structure. For example, the object pronoun is realized variably below as [mi] and [m]. In example (a), the verb kεtiŋ ‘cut’ (an extended form of the verb in (b)) ends in a closed syllable and therefore is followed by the full form mi; in example (b) the final syllable of the verb kεti ends in an open syllable and is followed by the reduced form -m. ()
a. bogalε hu kεtimmi. boga-lε hu kε-tiŋ-mi cutlass- . cut-- ‘The cutlass cut me.’
(B.)
b. bεlalε hu kεtim. bεla-lε hu kεti-m axe- . cut- ‘The axe cut me.’
(B.)
In addition to these two pronominal paradigms, there is a set of interrogative pronouns, some of them likely compound forms, as well as a number of indefinite pronouns. Interrogative: hina ‘who’ yε or yεna ‘what’ lɔ ‘what, where, when, how’ halɔ ‘what’ Indefinite: pε ‘someone, they, people’ la ‘something, it’
la ‘where’ lɔndɔ ‘where’ la or hu ‘how’ layεnda ‘why (for what)’ lɔ ‘somewhere, there’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Table . The Bom-Kim verbal complex
(Tense) ka/na Past
(Aspect) Verb (Aspect) (Aspect) ge/go ~ gon 0 -(y)έ ge ~ go ~ gon (Bom) Progressive Habitual Completive Perfect
.. Verbs and inflectional morphology Verbal distinctions, primarily aspectual, are generally signalled by pre- and post-verbal particles. The perfective (here called “completive” to contrast more sharply with “perfect”) however, is marked by a high-toned suffix -(y)έ. The habitual features the simple verb with no inflection, although there is a change in word order when pronominal objects are present. Comparison with closely related Mani suggests there was once a tonal aspect marker on the pronoun (see discussion of word order in section .. below). Table . shows the structure of the verbal complex. The two varieties have slight differences between them and some variation. The forward slash “/” separates a Bom form from a Kim form where the two differ. The tilde “~” separates alternates within a given variety. ()
batinuwε wɔ na ge bεmpa tεεngε. batnu-ε wɔ na ge bεmpa tεεngε farmer- make platform ‘The farmer was making a platform.’
(K.)
Imperatives optionally feature a second person pronoun and the verb a high tone or accent on the first syllable of the verb (here shown only on the verb). ()
Imperatives a. hún! ‘Come!’ b. (n) gbási yege logi! (n) gbási yege logi cut cassava this ‘Cut this cassava!’
.. Negation Negation is generally signalled by a floating high tone and an alveolar nasal suffixed to the tensed element in a verb phrase. Note how the object fe ‘money’ has no article in (). This is one of the few environments where nouns can appear without an article. ()
Negation: -ń suffixed to verb or auxiliary a. ya bilέn fe. ya bil-έ-n fe have-- money ‘I have no money.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
b. ya cén kɔ. ya ce-́n kɔ - go ‘I am not going.’ The imperative, however, is negated with a preverbal particle ma, which can also be used to negate nominal elements. Several speakers insisted on a final particle wε in the negative imperative, as featured in () below. ()
The imperative negated ma (n) gbási yege lε wε! ma (n) gbási yege lε wε () cut cassava ‘Don’t cut the cassava!’
()
Nominal elements negated with the negator ma bεpɔgi logi, ha ga kakpa ma wɔ go. bεpɔgi logi ha ka kakpa ma wɔ go chiefdom this vote ‘This chiefdom have voted not him (to remove him).’
(B.)
.. Verb extensions Bom-Kim has a limited number of verb extensions, only one of which, the instrumental, is currently productive, surviving perhaps, somewhat ironically, because it has been reanalyzed as a preposition. Other verb extensions are the causative and middle. The expected benefactive is commonly expressed analytically by means of a preposition la ‘for, on behalf of ’. With extensions following roots, this part of the verbal complex follows the Proto-NigerCongo pattern of the “inflectional verb” as outlined in Nurse (, ), an amazingly stable pattern throughout the phylum. ()
()
()
Instrumental -ka [ga] ha bεmpaga blɔklε isundε. ha bεmpa-ka blɔk-lε i-sun-lε make- block- -sand- ‘They make blocks with sand.’
(B.)
Causative -i ha ga ce go hini ntɔpɔlε? ha ka ce go hin-i n-tɔpɔ-lε lie- -rat.trap- ‘Were you laying rat traps?’
(B.)
Middle -n hi gɔ tɔn puε go. hi kɔ tɔ-n puε go go wash- waterside at ‘We go bathe at the waterside.’
(B.)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
With regard to the instrumental extension -ka, for some speakers it has been reanalyzed as a preposition, probably due to the influence of Mende, the language to which speakers are shifting, a highly analytic language. The example in () shows ka as a separate word after the direct object (cf. example ()). ()
a kɔn kεti tɔgilε ka gbεlalε. a kɔn kεti tɔgi-lε ka gbεla-lε go cut tree- with axe- ‘I cut the tree with the axe.’
(B.)
.. Adpositions The two varieties differ as to where the adposition ko/go is found: before the object in Kim and after the object in Bom. When it is before the object, as in Kim, the object can be followed by a more specific postposition further identifying the relation of the object to the verb. ()
()
a. Kim ya kwεn kɔn ko teiɔn. ya kwεn kɔ n ko tei-ɔn tomorrow go to Tei-in ‘Tomorrow I will go to Tei.’
(K.)
b. Bom hi hin lεε go. hi hin lεε go lie.down ground on ‘We slept on the ground.’
(B.)
pε ga gbem wɔ | bεngε go. pε ka gbem wɔ bεngε go they bore Bεngεh in ‘My father was born in Bεngεh.’
(B.)
Most adpositions precede their object, e.g., la ‘for’.
.. Adjectives and other minor categories Bom-Kim has very few adjectives, which all represent basic concepts. Adjectives can be derived from verbs by a productive process but the total number of underived adjectives in the language is below twenty. I list some examples below. ()
a. wɔ go dintε. ‘It is white.’ b. wɔ go pitε. ‘It is black.’
(B.)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Other adjectives, all representing basic concepts, are sana ‘new’, sa ‘red, brown, ripe’, wεi ‘bad’, kεnε ‘good’, bom ‘big’, teen ‘small’, ven ‘tall’. Verbs and nouns convey such meanings as ‘be old’ or ‘have cleverness’. Similarly, there are a number of intensifiers and some adverbs, locatives and temporal words. Colorful words known as ideophones register sensory perceptions. ()
si kɔ mu bεti gbɔlɔhɔn, gbip! si kɔ mu bεti gbɔlɔhɔn gbip then cut throat ‘Then (she) goes to cut (your throat) gbip!’ (as a goat’s is cut)
(B.)
Ideophones number some fifty or so (total lexicon ,+ words) but were likely more numerous when the language was more vital (see Childs ).
. S
.................................................................................................................................. Bom-Kim shows the effect of prolonged and extensive contact with the highly analytic Mande languages, as do all South Atlantic languages, but the mixed word order described in section .. below with regard to pronouns (S–AUX–OP–V) likely goes back to ProtoNiger-Congo; see, for example, Nurse ).
.. Word order, the verb phrase: SVO ~ S–AUX–OP–V What is striking about Bom-Kim is the dramatic differences in pronoun location with simple vs. complex predicates. In the former, with no auxiliary present, all objects appear after the verb, as shown in (). ()
SVO word order with simple predicates a. Baton wɔ sɔm sɔmma wε Baton eat food that ‘Baton ate that food.’
(K.)
b. wɔ kee mi. see ‘She saw me.’ c. ha sɔm ngefeyεε eat rice.gruel ‘They ate rice gruel.’
(K.)
When predicates consist of more than one element (technically, when tense is not marked on the verb but rather on some preverbal element such as an auxiliary or a pronoun), pronominal objects appear between the verb and AUX.2 Lexical noun phrases still appear 2
Even more specifically, the pronouns appear immediately after tense before the negative marker (see Childs for a detailed treatment).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
after the (lexical) verb. In the case of the hortative/future, TAM is marked on the pronoun, and pronoun objects appear immediately after the subject pronoun before the verb (.––). ()
wɔ hun yipi. wɔ hun yipi . bring ‘She should/will bring it.’
(K..)
Indirect objects when pronominalized will also appear between tense and the verb. ()
wɔm yoyo kuluŋ. wɔ-m yoyo kuluŋ - send goat ‘He will send the goat to me.’
(B.)
This order can be compared to the completive where the pronoun is found after the verb. ()
a. wɔ yipε sɔgi mɔ. wɔ yipε sɔk mɔ bring chicken one ‘She brought one chicken.’ b. wɔ yipε hun. wɔ yipε hun bring . ‘She brought it.’
(K.)
.. Relativization and focus Relative clauses begin with a pronoun agreeing with the noun being relativized, followed by a relative clause with a gap where the relativized noun should appear, with the relative clause terminating in the particle -yὲ. ()
Relative constructions a. a ka wɔm pεlεlε hu wɔ pin yε. a ka wɔn pεlε-lε hu wɔ pin yε give . rice- . buy ‘I gave him the rice she bought.’
(B.)
b. yi gbɔm laŋgbaŋgwε wɔ wugε. yi gbɔm laŋgbaŋg-ε wɔ wug-yε meet man- be.missing- ‘We met the man who was missing.’
(K.)
Focus constructions are similar except that there is no final “binding” particle. Focus, then, involves a fronted element, which is followed by its pronoun and a clause with a gap, except when the relativized noun is the subject, as in (b).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
- ()
Focus constructions a. sɔmma ma Baton wɔ sɔmyε. sɔmma ma baton wɔ sɔm-yε food Baton eat- ‘It’s the food that Baton has eaten.’
(K.)
b. Baton hu wɔ sɔmyε sɔmma. baton hu wɔ sɔm-yε sɔmma Baton eat- food ‘It’s Baton who has eaten the food.’
(K.)
.. Questions -questions require a fronted question word with a final particle -a. ()
a. lɔndɔ m pin pεlaa? lɔndɔ m pin pεla-a where buy rice- ‘Where did you buy the rice?’ b. hina wɔ ka ce Bεŋgesana gowa? hina wɔ ka ce bεŋgesana go-a who Bengesana - ‘Who was in Bengesana?’
(B.)
(B.)
Yes-no questions are marked only by rising intonation (examples not shown).
.. Subordinate clauses Subordinate adverbial clauses are introduced by one of several subordinating conjunctions such as the clause initial le and si in (). The first, le, is a general complementizer, and si is also used in the sense of ‘and then’ or ‘next’. There are no formal differences in the “subordinate” clause. ()
Subordinate clauses le a sɔm sɔmma, si ya han tisohin, minminanyε wɔ to mi. le a sɔm sɔmma si ya han tisohin when eat food if do sneeze minminanyε wɔ to mi ghost. beg ‘When I am eating, if I sneeze, [it’s a sign that] the spirit is petitioning me.’
(K.)
In the following sentence both types of connectives appear, here both with the optional binding particle yε found at the end of relative clauses.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
. Subordinate and coordinate clauses le wɔ woo kogan yε, nε wɔ dug yε, gbεŋwε negin. le wɔ woo kogan yε nε wɔ dug yε when mount up and fall gbεŋwε negin foot.. hurt ‘When he climbed up (the tree), and he fell, he broke his foot.’
(K.)
A Many thanks to the Bom and Kim peoples of the Bum and Kwamebai Chiefdoms in Sierra Leone. This material is based upon work supported by the National Science Foundation under Grant No. and by the Hans Rausing Endangered Languages Programme (MDP ).
R Blench, R. M. (). Archaeology, Language, and the African Past. Lanham, MD: AltaMira Press. Childs, G. T. (). ʻWhere have all the ideophones gone? The death of a word category in Zuluʼ, Toronto Working Papers in Linguistics : –. Childs, G. T. (). ʻThe Atlantic and Mande groups of Niger-Congo: a study in contrasts, a study in interactionʼ, Journal of West African Languages : –. Childs, G. T. (). ʻThe S-AUX-O-V syntagm in the Atlantic languagesʼ, Studies in African Linguistics : –. Lewis, M. P. (ed.) (). Ethnologue: Languages of the World, th edn. Dallas, TX: SIL International. Maddieson, I. (). Patterns of Sounds. Cambridge: Cambridge University Press. Matisoff, J. A. (). ʻRhinoglottophilia: the mysterious connection between nasality and glottalityʼ, in C. A. Ferguson, L. G. Hyman, and J. J. Ohala (eds.), Nasálfest. Stanford, CA: Stanford Universals Project, –. Nurse, D. (). ʻDid the Proto-Bantu verb have a synthetic or analytical structure?ʼ, in N. C. Kula and L. Marten (eds.), Bantu in Bloomsbury: Special issue on Bantu linguistics. London: Department of Linguistics, School of Oriental and African Studies, University of London, –. Nurse, D. (). Tense and Aspect in Bantu. Oxford: Oxford University Press. Pichl, W. J. (). The Krim Language in Sierra Leone. Pittsburgh, PA: Duquesne University Press Voeltz, F. K. E. (). Les langues de la Guinée. (Cahiers dʼÉtude des Langues Guinéennes, .) Conakry: Université de Conakry, Faculté des Lettres et Sciences Humaines. Wilson, W. A. A. (). ʻAtlanticʼ, in J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham, MD, and London: University Press of America, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. D is spoken by more than . million people in Côte d’Ivoire and Liberia; there are several Dan villages in Guinea as well. It is usually referred to as Yakuba in Côte d’Ivoire, and as Gio in Liberia. There is considerable variation among the Dan dialects; the variety described here is the dialect of the Gwεεtaa area (which is taken for the basis of the language norm for the Northeastern Dan area). Additional information on Dan can be found in Vydrin (a, b, , , ) and Vydrin and Kességbeu ().
. P
.................................................................................................................................. Table . summarizes the system of oral and nasal vowels in Dan. Phonetically long vowels are interpreted as combinations of two identical vowels (rather than long phonemes). /ɒ:/ is always long. ææ, ɒ:, æ̰ æ̰ , ɒɒ̰ ̰ very often vary with εε, ɔɔ, ε̰ε̰, ɔɔ̰ ̰ respectively. Semi-closed oral vowels /e, ɤ, o/ are realized as more closed, [ɩ, ұ, ʋ] under extra-high tone or in the position after an extra-high tone: ɓőɓő [ɓʋ̋ɓʋ̋] ‘deaf-mute’, dɤŋ̋ ̋dɤŋ̏ ̏ [dұ̋ŋ̋dұ̏ŋ̏] ‘miserly’, főőfȍȍ [fʋ̋ʋ̋fʋ̏ʋ̏] ‘useless (person)’. There is a trend to phonologization of [ұ] triggered by morphologically conditioned alternation of tones, which is not further discussed here for reasons of space. /ŋ/ can be considered as a vowel with a limited distribution. Table . summarizes the Dan consonant system. In a nasal feet, any consonant is nasalized; phonemes /ɓ, ɗ, y, w/ are represented by their allophones [m, n, ɲ, w̃ ]. Foot-internal -l- is realized as [-r-] when preceded by dental or palatal consonants, and as [-l-] after labial and velar consonants. Combinations sl-, zl- are realized respectively as [ɬ], [ɮ]: slʌʌ̄ ̄ [ɬʌʌ̄ ̄] ‘turn’, zláȁ [ɮáȁ] ‘younger sibling’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Dan vowels Oral vowels
Nasal vowels
Front
Back non-labial
Back labial
i
ɯ (ұ) ɤ ʌ a
u
e ε æ
ŋ
o ɔ ɒ:
Front
Back non-labial
Back labial
ḭ
ɯ̰
ṵ
ε̰ æ̰
ʌ̰ a̰
ɔ̰ ɒ̰
Table . Dan consonants Labial
Dental
Voiceless stops
p
Voiced stops Voiceless fricatives
Velar
Velar labialized
Labiovelar
t
k
kw
kp
b
d
g
gw
gb (gm)
f
s
Voiced fricatives
v
z
Implosives/sonorants
ɓ (m)
ɗ (n)
Lateral sonorants
Palatal
y (ɲ)
w (w̰)
l
There are five level tones: () () () () ()
extra-high: ka̋a̋ ‘scabies’, high: káá ‘you ()’, negative imperfective series of pronominal predicative markers, mid: kāā ‘you ()’, prospective series of pronominal predicative markers, low: kàà ‘to scratch’ (a conjoint construction form), extra-low: kȁȁ ‘reed’.
There are three modulated tones (all are available on single vowels), they are less frequent than level ones: () high-falling: gbε̰̂ ‘dog’, () mid-falling: da̰᷆ ‘hunger’, () extra-high-falling: zīīzı̋ˋɗ e ‘extremely ancient’. There are lexical tones and grammatical tones (tonal morphemes); the latter can substitute the lexical tone of a word (pɔ̰́ ‘to dig’ ! pɔ̰̏ ‘to dig’, neutral aspect) or be suffixed (pɔ̰̂ ‘to dig’, infinitive). Contextual modifications of tones are very few. Syllabic structures allowed in Dan are: V, CV, CСV. The only consonant admissible in the C position is /l/, it cannot be preceded by gw-, l-, v-, y-.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The main rhythmical unit is the metric foot. The types of feet attested in Dan are: V, ŋ, CV, CVV, CVŋ, CVVV, CVVŋ, ClV, ClVV, ClVŋ. A foot is characterized by: ) nasal harmony, ) restrictions on vowel combination, and ) restrictions on tonal combinations. Only identical vowels occur in the ClVV feet. In CVV and CVVŋ feet, if both vowels (-ŋ not included) are not identical, the first vowel can be only i or u, and the second one is almost exclusively back unrounded non-high, i.e. ɤ, ʌ, a. In the CVVV feet, three identical vowels are impossible; V can be either i or u, V and V are always identical, non-high, and most often contrasting with V by localization: sı̋ʌʌ̏ ̏ ‘pretentiousness’, sű e e ‘fearful’. All tones in a sequence within the word can be identical; otherwise, if the final tone is extra-high, it can be preceded by any other tone; if the first tone is mid, it can be followed by any tone, except for low. Other tonal combinations are extremely rare; modulated tones do not appear in heavy feet (consisting of more than one syllable). See also Bearth and Zemp () for further details.
. M
..................................................................................................................................
.. Nouns Nouns are subdivided into the classes of nouns proper and “locative nouns”. The latter have an emerging inflectional category of case; the former have none.
... Locative nouns Locative nouns result from a fusion of nouns with postpositions. Six cases can be singled out, with a great deal of irregularity in their formation: () common case is most often marked with the suffix -ɗ ε̏, the stem is often modified (if compared with the corresponding noun proper): the vowel may be reduplicated, the tone may be modified, e.g. gε̰̏ ‘leg, foot’ (noun) ! gε̄ŋ̰ ɗ̏ ε̏ (locative noun, common case). The common case form of certain other locative nouns can be identical with the corresponding noun: yṵ̄ ‘nose’, sε̋ ‘earth’; () locative case is derived from the common case form through substitution of the suffix -ɗε̏ by the suffix -ɗɤ̄ (identical with the adverbial suffix); sometimes the tone of the stem is modified (which can be explained by the fact that at least some of the locative case forms have incorporated the inessive postposition gɯ́ as well): ɗ ε̄ŋɗ̏ ε̏ ‘farm camp’ (common case), ɗ ε̄ŋɗ̄ ɤ̄ (locative case). Certain lexemes derive their locative case without any suffix; () superessive case results from fusion with the postposition ɓȁ ‘on (flat surface)’, e.g. kɔɔ́ ɗ́ ε̏ ‘house’ (common case), kɔɔ́ ̏ ‘on the wall of the house’ (subessive case);
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() inessive case results from fusion with the postposition gɯ́ ‘in’, e.g. ya̰̋ ‘eye, eyes’ (common case), yʌ̰̋ŋ̋ ‘in(to) the eyes’ (inessive case); () adessive, a rare case, results from fusion with the postposition tȁ ‘on (flat surface)’, e.g. kɔ̏ ‘hand’ (common case), kɔɔ̏ ̏ ‘on the hand(s)’ (adessive case); () comitative/instrumental, a rare case, results from fusion with the postposition ká ‘with, by’, e.g. kɔɔ̄ ̄ ‘by the hand(s)’ (comitative/instrumental case). The declension of the locative nouns is defective; there is no lexeme which would have forms of all the six cases. In fact, only common and locative cases are relatively regular, the other cases being available for % (the superessive case) to % (the adessive and comitative cases) of the locative nouns.
... Relational and free nouns The distinction between relational and free nouns is valid for both the nouns proper and the locative nouns; it manifests itself in a possessive noun phrase where the possessor (the dependent word) is a person. If the head noun (the possessum) is a free noun, it is connected to the possessor noun by a connective word ɓȁ ~ ȁ: Gbȁtȍ ɓȁ ɗʌ̰́-ɗ ṵ̏. ‘Gbato’s children’. ɓȁ is optionally replaced by gɔ̏ if the NP appears in the circumstant function: ()
n̄ gɔ̏ kɔ́ɔ́ɗɤ̄ . . house. ‘at my home’
If the possessed noun is relational, there is no connective word: Gbȁtȍ dʌ̄ ‘Gbato’s father’. Most body part names and kinship terms belong to the class of relational nouns; however, there are some important exceptions: ɗʌ̰́ ‘child’, bέε̏ ‘nephew (sister’s son)’, wṵ̄ ‘hair (head of hair)’, fʌ́yi̋ ‘sweat’ are free nouns, while ɓāŋ̄ ‘child’ (a synonym of ɗʌ̰́ ), wēe̋ ‘urine’, kāȁ ‘body hair’ are relational.
... Plural marking The standard plural marker is -ɗ ṵ̏, which follows the noun. If the noun is modified by an adjective (and/or a modifier), -ɗ ṵ̏ most often follows the adjective; however, it can follow the noun itself, or appear in both positions (cf. section ...). A limited group of nouns for younger male kin and members of professional groups form their plural by a derivative suffix -zʌ̏, followed by -ɗṵ̏; if the singular form has a suffix -ɓḭ̄ (designating a human being, par excellence a male), it is omitted: yε̄ε̄-ɓḭ̄ ‘griot’ (member of an inferior caste) ! yε̄ε̄-zʌ̏-ɗṵ̏; bέε̏ ‘nephew’ ! bέε̏zʌ̏-ɗ ṵ̏. Otherwise, there is only one case of irregular plural formation: ɗēbʌ̏ ‘woman, wife’ ! ɗōō-ɗ ṵ̏, ɗ ōŋ̄-ɗ ṵ̏ (a regular form, ɗ ēbʌ̏-ɗ ṵ̏, is also possible). The majority of locative nouns have no plural forms in the oblique cases, and when they do, it is usually derived as follows: the common case suffix is added after the locative case suffix, and the standard plural marker is added at the end, e.g. ɗēŋ̏-ɗɤ̄ ‘field, farm’ ! ɗēŋ̏-ɗɤ̄-ɗ ε̏-ɗ ṵ̏.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Adjectives Syntactic functions characteristic of adjectives are: () noun modifier in an attributive NP: ()
yʌ̰́ŋ̏ gbe̋e y ɤ̏ fɤ̄yi̋ gā lȍȍ ɓε̰̄ ɓȁ. sun. difficult . sweat bone arrive- human on ‘Hot sun makes one sweat.’
() predicate in a postposition-less construction: n̄ kɔ ̏ yɤ̏ kpʌ̋ʌ̋kpʌ̏ʌ̏. . hand. . numb ‘My hand is numb.’
()
() predicate in a postpositional construction: ()
ȁ gε̰̏ yɤ̏ kpa̰̋a̰̋ ká. . foot . bare with ‘He is barefoot.’
For the adjectives, there are semi-regular derivative models for plural and intensive meaning based on reduplication, modification of tones, suffixation, and transfixation; see Table ..
Table . Derivation of plural and intensive adjectival forms in Dan Basicsingular
Basic plural
Singular intensive
Plural intensive
Singular superintensive
gbḭ́ŋ̏ ‘heavy’
–
gbḭ́ŋ́gbḭ̏ŋ̏
gbḭ́ŋ̏gbḭ́ŋ̏
gbı̰̋kı̰̋gbḭ̏kḭ̏
glɔ̋ɔ̋glɔ̏ɔ̏ ‘blunt’
–
glɔ̄ɔ̋glɔ̄ɔ̋-sɯ̏
–
glɔ̋kɔ̋glɔ̏kɔ̏
sε̰̄ε̰̋nʌ́ ‘small’
sέ̰έ̰nʌ́
sε̰̄ε̰̋ε̰̏ε̰̏nʌ́
sέ̰sέ̰nʌ́
–
All adjectives have a selective form derived by the suffix -sɯ̏ (homonymous to the gerund suffix and the adjectivizing suffix). There are two non-productive suffixes deriving adjectives from nouns, -ɗē (pȉɤ̏gā ‘iron’ ! gbɔ̄ pȉɤ̏gā-ɗē ‘iron pot’) and -sɯ̏ (pēŋ̄ ‘fire’ ! pēŋ̄-sɯ̏ ‘hot’).
.. Numerals Numerals (see Table .) from six to nine are etymologically compounds (+, , , ; probably, a connective element gā ‘bone; unit’ is incorporated). kɒ̏ɒ̏ŋ̏ ‘ten’ has free variants kɔŋ̏ ̏ and kɔ̏; this numeral is derived from the noun kɔ̏. Numbers of units are connected with the numbers of dozens with the expression ɤ̄ gā ‘its unit’. kʌŋ̄ ̄ ‘hundred’ is probably
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Dan numerals
dō plὲ~p e edʌ̄ yȁȁgā yȉȉsīɤ̄ sɔ̋ɔ̋ɗű sɒ̋ɒ̏dō sɒ̋ɒ̏plὲ sa̋ȁgā sű esīɤ̄~sű esīɤ̄
kɒ̏ɒ̏ŋ̏dō kɒ̏ɒ̏ŋ̏dōɤ̄gādō kɒ̏ɒ̏ŋ̏dōɤ̄gāplὲ kɒ̏ɒ̏ŋ̏dōɤ̄gāyȁȁgā kɒ̏ɒ̏ŋ̏plὲ kɒ̏ɒ̏ŋ̏plὲɤ̄gādō kɒ̏ɒ̏ŋ̏plὲɤ̄gāplὲ~p e edʌ̄ kɒ̏ɒ̏ŋ̏yȁȁgā kɒ̏ɒ̏ŋ̏yȉȉsīɤ̄
, , , , ,,
kɒ̏ɒ̏ŋ̏sű esīɤ̄~sű esīɤ̄ kʌ̄ŋ̄dō kʌ̄ŋ̄plὲ gblɯ̋dō gblɯ̋plὲ gblɯ̋kɒ̏ɒ̏ŋ̏dō gblɯ̋kʌ̄ŋ̄dō gblɯ̋ɓε̰̄ε̰̏dō
borrowed from Manding kὲmέ. Gblɯ̋ ‘thousand’ is by origin the name of a basket storing , cola nuts (in the Dan country, cola nuts were traditionally the major export item). Nouns modified by numerals appear in their single form: sű plὲ ‘two months’. Ordinal numbers are derived by the suffix -ɗ ȁ̰ȁ̰ (sɔɔ̋ ɗ̋ ű-ɗ ȁ̰ȁ̰ ‘the fifth’); only ‘first’ has a suppletive form, blε̏ε̏sɯ̏.
.. Pronominal predicative markers (PPM) Pronominal predicative markers (see Table .) stem from the fusion of personal pronouns in the subjective function with post-subject auxiliary elements (predicative markers). A PPM appears as a head word of a verbal phrase; it is necessarily present in any verbal sentence (while the subject may be null); it is also obligatory in most types of non-verbal utterances. A PPM expresses (alone or in combination with grammatical tones on the verb, verbal suffixes, and auxiliary verbs) different TAM meanings and polarity.
Table . Pronominal predicative markers (PPM) in Dan Singular Person
Existential
ā
ī/ɯ̄
Conjoint
á
í/ɯ́
Perfect
ɓá̰
ɓá
yɤ̏/ yɤ̀ Ø/ɤ́/ yɤ́ yà/yȁ
Prospective
ɓā̰ā̰ ɓīī
Imperative
–
Optative
á
Presumptive
ɓā̰ȁ̰ ɓāȁ
Ø/ ɓɤ̏ í/ɯ́
Plural
Dual
Plural
Logophoric ɤ̄
excl.
incl.
incl.
yī
kō
kwā
kā
wȍ
Logophoric wō
ɤ́
yí
kó
kwá
ká
wó
wó
yá
yá
kó
kwá
ká
yɤ̄ɤ̄
–
yīī
kōō
wà/ wȁ kwāā kāā wōō
–
–
–
kȍ
kwȁ
kȁ
–
–
Ø/yɤ̏
ɤ́
yí
kó
kwá
ká
wȍ
wó
yāȁ
–
yāȁ
kōȍ
kwāȁ kāȁ wāȁ
wá –
–
Negative imperfective
ɓá̰á̰ ɓáá
yáá
–
yáá
kóó
kwáá
káá
wáá
–
Negative perfective
ɓḭ́ḭ́
ɓíí
yíí
–
yíí
kóó
kwíí
kíí
wíí
–
Prohibitive
ɓá̰
ɓá
yá
–
yá
kó
kwá
ká
wá
–
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A logophoric PPM appears in a subordinate sentence or a non-first sentence in a narrative chain coreferent to the subject of the main (preceding) sentence. However, the controller of the logophoric PPM can be a “semantic subject” different from the syntactic subject, as in (). ()
yɤ̏ kʌ̏ Zâ̰i gɯ́ ɗ ὲ . become\ Zan in that kʌ̄ klɤ̋ɤ̋klɤ̏. yȁj/yái ./.. become healthy ‘Zan thinks that he has recovered.’ (Zâ̰ is an oblique object.)
.. Noun phrase structure ... Genitive-like noun phrases In a genitive-like noun phrase, the head noun follows the dependent one. Genitive NPs are of two types: () unmarked: bāā ɗɯ̋ (lit. ‘cassava tree’) ‘cassava stalk’, and () head-marked, where the syntactic connection is marked on the head noun by an extra-low tone: bāā ɗɯ̏ ‘cassava plant’. To the class of genitive-like NPs also belongs the postpositional NP. In such NPs, the connection between the constituents is marked by a postposition: ɗūɤ̄ɤ̄ ɓȁ gbīŋ̄gā (lit. ‘raffia. palm on caterpillar’) ‘raffia palm caterpillar’. Possessive NP (ɗùùtí ȉ ɓȁ ɗ ȁȁ ‘village chief ’s knife’) can be also regarded as a variety of the postpositional NP.
... Attributive noun phrases In an attributive NP, the head word precedes the dependent one: gbε̰̂ tīī ‘black dog’. In the attributive NP, determinatives (plural marker, definite article, etc.) normally follow the dependent word: gbε̰̂ tīī-ɗ ṵ̏ ‘black dogs’, gbε̰̂ tīī ɓā ‘the black dog’. Very atypically of Mande languages, the plural marker can optionally follow the head word (gbε̰̂-ɗ ṵ̏ tīī ‘black dogs’) or appear in both positions (gbε̰̂-ɗṵ̏ tīī-ɗ ṵ̏ ‘black dogs’); both strategies are subject to numerous lexical restrictions.
... Coordinative noun phrases The coordinative NP is constructed with coordinative pronouns (cf. section ...).
.. Pronouns ... Personal pronouns Non-subject pronouns (see Table .) appear in any argument position, with the exception of subject; they can also appear as the dependent member of a genitive-like NP. A nonsubject pronoun can be modified by an adjective, as in ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Personal pronouns in Dan Singular
Plural
Dual
Plural
Person
Refl excl
incl
incl
Refl
Non-subject
N̄
ī/ɯ̄
ȁ
ɤ̄
kō
kwā
kā
ȁnȕ
wō
Possessive
mā
ɯ̄ ɓā ȁ ɓȁ
ɤ̄ ɓā yīȉ
kōȍ
kwāȁ
kāȁ
ȁnȁȁ
wōȍ
ɓī
yɤ̄
–
yī
kō
kwā
kā
wō
–
yɤ̄ɤ̄
–
yīī
kōō
kwāā
kāā
wōō
–
Autonomous mā
yī
Contrastive
māŋ̄ ɓīī
Simple coordinative
yāā/ kāā/ wāā/ – yāȁ kāȁ wāȁ
yāā/ – yāȁ . . . nȕ
kwāā/kwāȁ ( . . . nȕ)
kāā/ wāā/ – kāȁ . . . nȕ wāȁ . . . nȕ
Fused coordinative
yāȁ
–
yāȁ . . . nȕ –
kwāȁ ( . . . nȕ)
kāȁ . . . nȕ wāȁ . . . nȕ –
kāȁ
wāȁ
Coordinative yε̄ŋ̄/ kε̄ŋ̄/ wε̄ŋ̄/ – portemanteau yε̄ŋ̏ kε̄ŋ̏ wε̄ŋ̏ & & &
()
yε̄ŋ̏ nȕ (&)+
kε̄ŋ̄/ kε̄ŋ̏ wε̄ŋ̄/ wε̄ŋ̏ – kwε̄ŋ̄/ kwε̄ŋ̄/ kwε̄ŋ̏ kwε̄ŋ̏ . . . nȕ nȕ (&) nȕ (&) (&)+ & + +
kɤ̄ kíí ɓā wʌ́ʌ̏gā ɗὲŋ̀, míí that that..>. money lose .. ɗó ȁ wε̋ε̋ nū’. go . other give- ‘If you lose your money, I won’t give other.’
The non-subject pronoun is on the way of grammaticalization as a prepositional definite article, (). ()
Yà ɗṵ̄ ȁ ɓε̰̄ főkőfȍkȍ-sɯ̏ ká. . come . human useless.- with ‘He has brought the most useless person among the people in question.’
A pronoun of the simple coordinative series appears as the first member of a coordinative phrase, while the second member is expressed by a non-pronominal NP. The fused coordinative pronouns are used when the second conjunct includes, at the underlying level, a non-subject pronoun ȁ, which is incorporated into the coordinative pronoun. The coordinative portmanteau pronouns appear when both conjuncts are pronominalized.
... Interrogative pronouns Interrogative pronouns are used for special questions: dē ‘who?’, ɓʌ̰̏ ‘what?’, ɗὲ ‘how many?, how much?’, ɓε̰̏ε̰̏ ‘which?’, ɓέ̰ ‘where?’. As a rule, they occupy the same position as the substituted noun; however, an interrogative pronoun (or an NP containing such a pronoun) can be moved to the leftmost position, in which case a resumptive pronoun occupies its syntactic position, as in ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
ɓʌ̰̏i
ɤ́ ɯ́ ȁ yɤ̄ kpḭ̏ŋ̏ tȁ? what . . see. road on ‘What have you seen on the road?’
A general question is expressed by the particle e e at the end of the sentence.
... Demonstrative pronouns and adverbs There are five demonstrative adverbs differentiated according to different parameters: yɤ̄ and yā indicate a place near the speaker or at an equal distance from the speaker and the addressee, within the limits of visibility. ɓā indicates a place near the addressee, within or outside the visibility of the speaker. ɓɯ̰̄ stands for a place outside the visibility of both participants in the communication, which can be described, but not indicated, or designates a movement from the deictic center, with an indefinite final point. tīȉɗ ɤ̏ (the intensive form is tīītı̋ɗ ɤ̏) designates a remote object at the limit of visibility or just beyond; it is necessarily accompanied by an indication by hand or by finger. yā has undergone grammaticalization into a demonstrative pronoun, and a definite article has evolved from the adverb ɓā. In both cases, a relativization construction is involved, as in (). ()
kɔ ́ ɤ́ ɤ́ yā /ɓā ! house . here there ‘house which is here/there’ (near the addressee) !
kɔ ́ yā /ɓā house this ‘this house/the house’
.. Verbs ... Lability There is no category of voice in Dan. On the other hand, lability is widely spread, especially involving decausative interpretations, as in (a, b). ()
a. yà fá̰ȁ̰ yɔ̰̀ɔ̰̀ ɤ̄ ya̰̋ tȁ. . headgear move . eyes. on ‘He has pulled the cap over his eyes.’ b. yɤ̄ɤ̄ yɔ̰̀ɔ̰̀ kɔ ́ ɗ ı̋ ɓȁ . . . . move house mouth on ‘She was going to approach the door of the house . . . ’
Passive lability (a, b) is also available, although it is not as widely spread as in Manding. ()
a. ɗēbʌ̏ɗʌ̰́ ɓā yà kɔ̏ɔ̏ wɯ́. woman . gourd break ‘The woman has broken a gourd.’ b. kɔ̏ɔ̏ ɓā yà wɯ́ ɗēbʌ̏ɗʌ̰́ ɓā ȁ kɔ̄ɔ̄/gɔ̏. gourd . break woman . by ‘The gourd has been broken by the woman.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
... TAM and polarity TAM and polarity are expressed by combinations of the following means: () () () ()
PPM (cf. section ..); verbal suffixes; grammatical tones on the verb; auxiliary verbs.
This is a typical linear system. The affirmative and the negative subsystems are asymmetrical in their nuclear parts and should be represented separately. The list of TAM/polarity constructions is shown in Tables . and ..
Table . Affirmative TAM constructions in Dan Neutral aspect EXI – V\El (the lexical tone of the verb is substituted by an extra-low grammatical tone). Semantics: habitual, irreal condition, progressive, imperfect, resultative/ perfect, completive. Conjoint construction
JNT – V\JNT, tone of the verb is lowered (according to an intricate pattern) or remains intact. Appears as an analogue to the neutral aspect construction in syntactically bound contexts (in dependent clauses, sequential strings, clauses containing focalization constructions, etc.).
Retrospective
EXI kʌ̄ V\El, kʌ̄ comes back to the verb kʌ̄ ‘to do; to become’.
Imperfect
EXI gṵ̄ \El – V-sīʌ̄, the auxiliary verb gṵ̄ ‘to be in the past’ is in the neutral aspect form; the content verb is in the durative form. Semantics: continuous action in the past.
Perfect
PRF – V.
Durative
EXI – V-sīʌ̄. Semantics: progressive, stative (with the verbs of knowledge, etc.).
Dependent progressive
JNT – V-ɗ ȁ̰ȁ̰/-ŋ̏. Appears in relative clauses, of limited use.
Resultative
EXI – V-sɯ̏ ká. -sɯ̏ is the gerund suffix, ká is a comitative postposition.
Prospective
PROS – V.
Future
EXI ɗó \El – V\INF; ɗó is an auxiliary verb ‘to go’ in the neutral aspect form; infinitive is marked by a non-segmental extra-low tone suffix on the verb.
Continuative
EXI tṵ̏ \El – V\INF, the auxiliary verb tṵ̏ ‘to continue; to remain’ is in the neutral aspect form.
Posteriority taxis
is expressed by two synonymous constructions: (a) OPT gṵ̄ – V\INF-ɗ ȁ̰ȁ̰ /-ŋ̏ ; (b) OPT gṵ̄-ɗ ȁ̰ȁ̰ – V\INF. Semantics: the action is to occur after another action mentioned in the first clause of a sentence.
Optative
OPT – V.
Imperative
IMP – V.
Presumptive
PRSM ɗṵ̏ – V\INF. Semantics: evidential (presumptive and/or inferentive) and probabilitive.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Negative TAM constructions in Dan Negative imperfective
NEG.IPFV – V.
Negative perfective
NEG.PFV – V.
Negative durative
NEG.IPFV – V-sīʌ̄.
Negative prospective
NEG.IPFV – PROS – V (the only construction requiring two PPMs; presumably, there was an auxiliary verb after the negative imperfective PPM).
Negative retrospective
NEG.IPFV kʌ̄ – V or NEG.PFV kʌ̄ – V.
Negative future
NEG.PFV ɗó – (O) V\INF.
Prohibitive
PROH ɗó – (O) – V\INF.
. S
..................................................................................................................................
.. Simple sentences The basic word order in a simple verbal sentence is (S) PPM (DO) V (OO/CIRC). Dan is a null-subject language (explicit presence of a subject NP is unnecessary, for the subject is expressed in the PPM). Presence of a DO makes a verb transitive; if the DO position is void, the verb is intransitive. Non-verbal sentences are formed with copulae (identification, negative identification, presentative) and with PPM (negative identification, locative, qualitative, equative). Identification, S ɓɯ̰̏ ‘It is S’ (if the object referred to by S is in the range of visibility), S ɓā ‘It is S’ (S is outside the range of visibility). One of the equative sentence type, S ɓɯ̰̏ O ká ‘S is O’, can be regarded as a variety of the identification sentence. Negative identification, (S) NEG.IPFV ɓɯ̰̏. ‘It is not S.’. Presentative sentence, S ɗὲ̰. ‘Here is S.’. In the (affirmative) identification and the presentative sentences, a subject is obligatory and a PPM is missing; all other sentence types in Dan contain a PPM. Locative sentence, (S) EXI OO ‘S is located in OO.’, the negative counterpart: (S) NEG. IPFV OO. Apart from the locative, many other meanings are expressed by this sentence type as well: possessive, equative, physical and psychological states, etc. Qualitative sentence, (S) EXI ADJ, negative counterpart: (S) NEG.IPFV ADJ. In non-autonomous clauses, EXI is replaced with JNT.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Complex sentences The two most current conjunctions are yɤ́ (rather coordinative) and kɤ̄ (rather subordinative); in both cases the existential series of personal predicative markers tend to be substituted by conjoint series of pronominal predicative markers, and the neutral aspect verbal form (extra-low-toned) by the conjoint form (tonal lowering according to a rather complex pattern). An adversative clause can be introduced by the conjunction kε̋ε̋, and an object clause by the conjunction ɗὲ. There are two relativization strategies. () Correlative strategy: the full-fledged relativized NP, followed by the relative conjunction ɤ́ (sometimes kɤ́), appears in the leftmost position in the subordinated clause. The relativized NP is resumed in its own syntactical position by a PPM or a non-subject pronoun; it is resumed again, by a pronominal predicative marker or a non-subject pronoun, in the main clause. The relative clause can be put either to the left of the main clause or embedded into the main clause. () Left-side relativization: the relativized noun is preceded by a nominalized verb (the nominalization is marked by an extra-low tone on the verb if it is separated from the relativized noun by an oblique object or an adverbial; otherwise, the nominalization remains morphologically unmarked). The relative clause appears in a reduced form; pronominalized subject and direct object are omitted.
R Bearth, T., and Zemp, H. (). ‘The phonology of Dan (Santa)’, Journal of African Languages : –. Vydrin, V. (a). ‘Les adjectifs en dan-gwèètaa’, Mandenkan : –. Vydrin, V. (b). ‘South Mande reconstruction: initial consonants’, in A.N. Meshcheriakov (ed.), Aspekty Komparativistiki II. Moscow: RGGU, –. Vydrin, V. (). ‘Co-ordinative pronouns in Southern and South-Western Mande: A second compound pronouns area in Africa?’, in K. Pozdniakov, V. Vydrin, and A. Zheltov (eds.), Personal Pronouns in Niger-Congo Languages: International Workshop. St. Petersburg, September –, . St. Petersburg: St. Petersburg University Press, –. Vydrin, V. (). Déclinaison nominale en dan-gwèètaa (groupe mandé-sud, Côte-d’Ivoire). Faits de langues : –. Vydrin, V. (). ‘Tonal inflection in Mande languages: the cases of Bamana and Dan-Gwεεtaa’, in E. L. Palancar, J. L. Léonard (eds.), Tone and Inflection: New facts and new perspectives. Berlin: De Gruyter Mouton, –. Vydrin, V., and Kességbeu, M. A. (). Dictionnaire dan-français (dan de l’est) avec une esquisse de grammaire du dan de l’est et un index français–dan. St. Petersburg: Nestor-Istoria.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. B is spoken in the north of the Republic of Benin, in the region of Atakora, as well as in Togo and Burkina Faso, by people who call themselves bīālībā.1 Biali consists of four dialectal variants in Benin: kāpāī spoken in Gouandé, lāsuàlī spoken in Dassari, piāngī spoken in Piangou, and mátèī spoken in Matéri (see also Sambiéni ). According to Manessy’s () and Naden’s () classifications, Biali belongs to the eastern subgroup of Gur-Oti-Volta languages, which belong to the Niger-Congo phylum; see Sambiéni () for additional information.
. P
..................................................................................................................................
.. Vowels According to Sambiéni (: ), Biali has five vocalic phonemes, which are shown in Table ..
Table . The vowel system of Biali Front
Back
Close
i
u
Mid
e
o
Open
1
Central
a
bīālī bā is the plural form of biālā-ū ‘Biali speaker’.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
All five vowels of Biali are nasalized when preceding nasal consonants.2 ()
hanfa pinhu huntu puom cendili
[hãfə] [pĩhũ] [hũdu] [puɔ̃m] [cε̃dələ]
‘guinea fowl’ ‘arrow’ ‘sauce’ ‘to hit’ ‘trade’
In some cases, particularly in medial and final positions, all five vowels are likely to be reduced to schwa (Sambiéni : –). This vowel reduction is shown by the borrowings from French represented in Table ..
Table . Vowel reduction in Biali Vowels
French transcription
Borrowings in Biali
Gloss
a i u o e
kaʀabin kapitεn bʀuεt ʃosεt s(ə)kʀetεʀ
kàràbīnī [kàrə̀bīnī] kàpìténī [kàpə̀ténī] pùrùwētī [pùrə̀wētī] sòsíétū [sə̀syéhū] sìkèteī [sìgə̀tēī]
‘rifle’ ‘captain’ ‘wheelbarrow’ ‘sock’ ‘secretary’
.. Consonants According to Sambiéni (), Biali has fifteen consonants, illustrated in Table ..
Table . The consonant system of Biali Labial
Alveolar
Palatal
Velar
Stop
p, b
t, d
c
k
Nasal
m
n
ɲ
Fricative
f
s
Lateral Approximant
Glottal
h
l w
y
Some consonantal phonemes have predictable variants in specific contexts. These are summarized in Table . and illustrated in example ().
2
/e/ and /o/ are realized as [ε̃] and [ɔ̃] instead of [ẽ] and [õ] when nasalized.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Allophonic variation in Biali Phonemes
Phonetic realization
Contexts
k
-ŋ-g-, -ɣ-r-d-
after n in intervocalic position (depending on the dialect of the speaker) in intervocalic position after n
d t
()
a. sanka sansi neka nesi naaki saki saka b. puudi piidi c. huntu mintu
! ! ! ! ! ! ! ! ! ! !
[sangə] ! [saŋə] [sansi] [negə] [nesi] [naaɣi] [saɣi] [saɣa] [puurə] [piirə] [hundu] [mĩndu]
‘horse’ ‘horses’ ‘bird’ ‘birds’ ‘to want’ ‘hut’ ‘huts’ ‘to blow’ ‘to take’ ‘sauce’ ‘excrement’
According to N’ouéni (: ), Biali contrasts three tonal heights: high tone (H), mid tone (M), and low tone (L). ()
búúkā ‘dog’ būūkā ‘goat’ bùùkā ‘kind of fetish’
Biali has three types of syllable structures: • Monosyllabic items with V and CV structures appear especially with pronouns and particles: ()
ā ù ī pá, pà dā
‘you ()’ ‘he, she, it’ ‘you ()’ ‘particle marking negation’ ‘particle marking coordination’
• Disyllabic items with VCV, CVV, and CVCV structures: ()
àlā àlī lōú lēí bārú tēbū
‘here’ ‘this’ ‘wood’ ‘woods’ ‘river’ ‘baobab’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
• Trisyllabic items with CVCVCV and CVCVV structures: ()
báríkā kàlábū márāū márībā
‘box’ ‘loincloth’ ‘smith’ ‘smiths’
. T
..................................................................................................................................
.. Noun classes The nominal lexical stock of Biali is divided into eight genders and fifteen noun classes. Each gender is represented with underlying class markers, as described in Sambiéni (). ()
Gender I (u/ba) lā-ū pūāhā-ū mámárā-ū cāārā-ū
()
()
[lā-ū] [pwāhā-ū] [mámárā-ū] [cāārā-ū]
Gender II (hú/hi) lō-hú [lō-ú] cō-hú [cō-ú] hwá-b-hú [hwá-b-ú] hàn-k-hú [hàn-g-ú] pā-hú [pā-hú] bà-hú [bà-hú] Gender III (fa/hi) nā-fā [nā-fə̄] ‘ox’ wō-fā [wō-fə̄] ‘snake’ cìì-fā [cìì-fə̄] ‘bee’ sí-fā [sí-fə̄] ‘navel’
() Gender IV (li/ha) kúndí-lī [kūndí-ī] tō-lī [tō-ī] tūn-nī [tūn-ī] cīīn-ní [cīīn-ī]
‘man’ ‘woman’ ‘smith’ ‘hunter’
‘wood’ ‘path’ ‘bone’ ‘stem’ ‘swamp’ ‘branch’
nā-hí wō-hí cì-hí sí-hí
lā-bā pūō-bā mámárā-bā cāārā-bā
lē-hí cē-hí hwá-b-hí hàn-k-hí pā-hí bà-hí [nā-hí] [wō-hí] [cì-í] [sí-í]
[lā-bə̄] [pūō-bə̄] [mámárə̄-bə̄] [cāārə̄-bə̄]
[lē-í] [cē-í] [hwá-b-í] [hàn-g-í] [pā-hí] [bà-hí]
‘woods’ ‘paths’ ‘bones’ ‘stems’ ‘swamps’ ‘branches’
‘oxen’ ‘snakes’ ‘bees’ ‘navels’
‘meat’ kúndí-há [kúnd-á] ‘ear’ tō-hā [tō-ā] ‘bean’ tūn-hā [tūn-ā] ‘knot’ cīīn-hā [cīīn-ā]
‘men’ ‘women’ ‘smiths’ ‘hunters’
‘meat’ ‘ears’ ‘beans’ ‘knots’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
()
()
Gender V (bu/na) cēē-bū [cēē-bū] tīē-bū [tyē-bū] nēē-bū [nēē-bū] tīī-bū [tīī-bū]
‘soap’ ‘bow’ ‘net’ ‘medicine’
Gender VI (hu/tu) sān-hū [sān-hū] kúnkún-hū [kúnkún-hū] múó-hū [mwó-hū] kù-hū [kù-hū] sē-hū [sē-hū] fēē-hū [fēē-hū] Gender VII (ka/si) bíí-kā [bíí-gə̄] nē-kā [nē-gə̄] búú-kā [búú-ɣə̄] kū-kā [kū-ɣə̄] lāpí-kā [lāpə́-kə̄] sēbí-kā [sēbə́-kə̄]
() Gender VIII (m) níī-m [níī-m] túó-m [twó-m] làā-m [làā-m] síí-m [síí-m]
‘car’ ‘cotton’ ‘grass’ ‘field’ ‘mat’ ‘leaf ’
‘child’ ‘bird’ ‘dog’ ‘chicken’ ‘boy’ ‘girl’
[cēē-nā] [tyē-nā] [nēē-nā] [tīī-nā]
cēē-nā tīē-na nēē-nā tīī-nā
bíí-sí nē-sí búú-sí kū-sí lāpí-sí sēbí-sí
sān-tū kúnkún-tū múó-tū kù-tū sā-tū fìā-tū [bíí-sí] [nē-sí] [búú-sí] [kū-sí] [lāpə́-sí] [sēbə́-sí]
‘soaps’ ‘bows’ ‘nets’ ‘medicine’ [sān-dū] [kúnkún-dū] [mwō-rū] [kù-rū] [sā-tū] [fyà-tū]
‘cars’ ‘cotton’ ‘grasses’ ‘fields’ ‘mats’ ‘leaves’
‘children’ ‘birds’ ‘dogs’ ‘chickens’ ‘boys’ ‘girls’
‘water’ ‘sand’ ‘drink’ ‘cold’
.. Nominal derivation With the nominal class markers -u, -hú, -li, -ha, -bu, -hu, -tu, -ka, -si, -m, two types of derivation are possible: nouns from verbs and nouns from others nouns.
... Nouns from verbs Biali offers considerable possibilities for deriving nouns from verbs by adding class markers to the root of the verb. ()
a. bá cāārī bá cēndī b. bá múórí bá pārí
[bə́ cāārə̄] [bə́ cēndə̄] [bə́ mwórə́] [bə́ pārə́]
‘to hunt’ ‘to trade’ ‘to swell up’ ‘to plait’
cāārā-ū cēndā-ū múór-hú [mwór-ú] pár-hú [pār-ú]
‘hunter’ ‘trader’ ‘swelling’ ‘plaited rope’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
c. bá cīná bá tōūká d. bá nàkìsì bá dùòndì e. bá pá bá yīsá f. bá kūndì bá sòú g. bá bárí bá hùḿ h. bá císīm bá hósī i. bá súbīkī bá yíndī j. bá fùsì bá lāāsí
[bə́ cīná] [bə́ tōūɣá] [bə́ nàɣə̀sə̀] [bə́ dwòndì] [bə́ pá] [bə́ yīsá] [bə́ kūndə̀] [bə́ sòú] [bə́ bárə́] [bə́ hùḿ] [bə́ císə̄m] [bə́ hósə̄] [bə́ súbə̄kē] [bə́ yíndə̄] [bə́ fùsə̀] [bə́ lāāsə́]
‘to tie’ ‘to load’ ‘to compete with’ ‘to praise’ ‘to give’ ‘to ban’ ‘to baptize’ ‘to pray to’ ‘to dress’ ‘to quarrel’ ‘to sneeze’ ‘to bark’ ‘to become heavy’ ‘to become stupid’ ‘to breathe’ ‘to laugh’
cīn-nī tōūkī-lī nàkìs-hā dùònd-hā pā-bū yī-bū kūndì-hū sò-hū bá-tū hùn-tū císīm-kā hósī-kā súbí-sí yíndī-sí fùsì-ḿ lāāsí-m
[cīn-ī] [tōūɣ-ī] [nàɣə̀s-ā] [dwònd-ā] [pā-bū] [yī-bū] [kūndə̀-hū] [sò-hū] [bá-tū] [hùn-du] [císə̄m-kə̄] [hósə̄-kə̄] [súbə́-sí] [yíndə̄-sí] [fùsə́-m] [lāāsə́-m]
‘knot’ ‘load’ ‘rivalries’ ‘praises’ ‘donation’ ‘banning’ ‘baptism’ ‘prayer’ ‘clothes’ ‘quarrel’ ‘sneeze’ ‘barking’ ‘heaviness’ ‘stupidity’ ‘breath’ ‘laugh’
... Nouns from nouns Biali also has various strategies for deriving nouns from other nouns by replacing the respective class markers. () a. yàbì-lī tà-lī b. tín-nī nùō-lī sān-kā c. kàlá-bū mīātì-lī d. sābī-kā hānr̄-kā lāpí-kā
[yàbì-ī] [tà-lə̄] [tín-ī] [nwō-ī] [sān-gə̄] [kàlá-bū] [myātì-ī] [sābə̄-kə̄] [hānrə̄-kə̄] [lāpə́-kə̄]
‘penis’ ‘leg’ ‘groundnut’ ‘yam’ ‘horse’ ‘loincloth’ ‘teacher’ ‘Lannea’3 ‘Vitex’4 ‘boy’
yàb-hú tà-l-hú tín-hū nùō-hū sān-hū kàlá-kā mīātì-kā sābī-ḿ hānrī-ḿ lāpí-m
[yàb-ú] [tà-l-ú] [tín-hū] [nwō-hū] [sān-hū] [kàlá-ɣə̄] [myātə̀-kə̄] [sābə́-m] [hānrə́-m] [lāpə́-m]
‘very small penis’ ‘very small leg’ ‘plant of groundnut’ ‘plant of yam’ ‘vehicle (big horse)’ ‘small loincloth’ ‘small teacher’ ‘fruit of Lannea’ ‘fruit of Vitex’ ‘small boy’
... Derivation by noun reduplication In addition to affixation, reduplication is used as a word formation strategy in Biali. ()
3 4
sān-nī ‘time’ sān-nī sān-nī ‘a long time’ bàn-nī ‘distance’ bàn-nī bàn-nī ‘a long way’ bīēn-nī ‘year’ bīēn-nī bīēn-nī ‘a long time (many years)’
Scientific name of a tree. Scientific name of a tree.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Nominal compounding Nominal compounding in Biali is characterized by the use of two or more lexemes expanded with class markers. The compound noun retains only the class marker of the last noun. There are two types of compound nouns: completive nouns and qualifying nouns: Completive nouns: ()
! + ! pùē-ū ‘father’ + pūāhā-ū ‘woman’ + ! tè-kā nā-fā
‘foot’ ‘ox’
+ bíí-lī + lā-ū
‘child’ ‘male’
tè-bíí-lī ‘toe’ tēē-kā ‘non-castrated’ nà-lā-tē-kā ‘bull’ tà-ū ‘sister’ + bíí-kā ‘child’ pùē-pū-tà-bí-kā ‘cousin’
Qualifying nouns: () lā-ū ‘man’ + bà-tū ‘cheeks’ ! lā-bà-tū ‘man with big cheeks’ yūō-lī ‘head’ + sìīb-hú ‘blackness’ ! yūō-sìīb-hú ‘bad luck’
.. Noun phrases There are two ways of constructing noun phrases with morphologically related markers: conjoining noun phrases using da, and noun phrases expressing an alternative (‘or’) using ní da. da bíí-kā bà nūām hīdā-cēdā. () a. pùē-ū father- and child- be person-one ‘The father and the child are one person.’ b. ā pāá bō-lī ní dá tā-bū āá? go party- war- ‘Are you going to party or to war?’
.. Nominal qualifiers Adjectives do not exist in Biali as an independent grammatical category. In the case of nominal qualification, adjectives are derived from stative verbs by taking the class marker of the noun they are qualifying (see Table .); this is only possible when they have an attributive function. If the verb has a predicative function, there is no gender agreement. However, the demonstrative and interrogative adjectives have a lexical base to which the class markers are assigned (see Tables . and .).
.. Pronouns ... Allocutive pronouns The allocutive pronouns represent persons taking part in the speech act.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Qualifying adjectives in Biali Stative verbs
Gloss
Adjectives
Gloss
bá ḿ púóà
‘to be white’
bá ḿ míēsī
‘to be slender’
bá ń sūōs
‘to be red’
bá ń sìībì
‘to be black’
bá ń tōū
‘to be big’
bá ń lāākī
‘to be wide’
bá ń sūlī
‘to be beautiful’
bá ń léésī
‘to be clean’
hīdī-púóà-ū hīdī-púó-bā lō-míēs-hú lō-míēs-hí nà-sūōsī-fā nà-sūōs-hí nìn-sīībì-lī nìn-sīīb-hā tēētōtō-bū tūōntōtō-hā dùān-lāākī-hū dùān-lāākī-tū bíí-sūlī-kā bíí-sūlī-sí níí-léésī-m
‘white person’ ‘white persons’ ‘slender timber’ ‘slender timbers’ ‘red ox’ ‘red oxen’ ‘black tooth’ ‘black teeth’ ‘big baobab’ ‘big baobabs’ ‘wide room’ ‘wide rooms’ ‘beautiful child’ ‘beautiful children’ ‘clean water’
Table . Demonstrative adjectives in Biali Gender Demonstrative adjectives Gloss sg
pl
I
à-kwē
à-bā
II
à-hú
à-hí
III
à-fā
à-hí
IV
à-lī
à-há
V
à-bū
à-há
VI
à-hū
à-tū
VII
à-kā
à-sí
VIII
‘this, that (sg)’ ‘those (pl)’
à-mum
‘this, that, these those’
Table . Interrogative adjectives in Biali Gender
Interrogative adjectives sg
pl
I
wētī
mám-bā-tī
II
mán-hú-tī
mán-hí-tī
III
mán-fā-tī
mán-hí-tī
IV
mán-nī-tī
mán-há-tī
V
mán-bū-tī
mán-há-tī
VI
mán-hū-tī
mán-tū-tī
VII
mán-kā-tī
mán-sí-tī
VIII
mám-mum-tī
Gloss
‘which’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
n̄ ‘I’ tī ‘we’ ā ‘you ()’ ī ‘you ()’
... Substitutive pronouns The substitutive pronouns act as substitutes for nouns; there are fifteen possible substitutive pronouns, corresponding to the fifteen noun classes of Biali (see Table .).
Table . Substitutive pronouns in Biali Gender Substitutive pronouns Gloss sg
pl
I
ù
bà
II
hù
hì
III
fà
hì
IV
lì
hà
V
bù
nà
VI
hù
tù
VII
kà
sì
VIII
m̀
‘he, she / they’
‘he, she, it / they’
‘he, she, it’
... Interrogative pronouns Formally the interrogative pronouns are the same as the interrogative adjectives presented in Table ..
.. Numerals In Biali, all the numerals from one to ten are simple constituents. This is also the case for , ,, ,,, and ,,,. The numerals , , ,, ,,, and ,,, take the class markers in the same way as nouns. () càdām díà tāādī nààsì nùm kō-kā tùsī-cān-nī
‘one’ ‘two’ ‘three’ ‘four’ ‘five’ ‘hundred’ ‘million’
hàdùàm pìléí nēī wáí púī-kā tùsī-lī tùsī-bàm
‘six’ ‘seven’ ‘eight’ ‘nine’ ‘ten’ ‘thousand’ ‘billion’
Coordinated numeral phrases are characterized by the use of the coordination marker dā followed by n̄:
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() púī-kā dā n̄ càdām ‘eleven’ púī-kā dā n̄ tāādī ‘thirteen’ púī-kā dā n̄ wáí ‘nineteen’ Multiplication is used in order to express decimals: () púí-sí púí-sí kō-sí kō-sí tùs-hā
díà tāādī díà pìléí púī-kā
‘twenty’ ‘thirty’ ‘two hundred’ ‘seven hundred’ ‘ten thousand’
The tens and hundreds apart, the other numerals are expressed by the combination of a completive numeral phrase coordinated with another numeral. () púí-sí díà dā n̄ wáí ‘twenty-nine’ púí-sí wáí dā n̄ nùm ‘ninety-five’ kō-sí wáí dā n̄ wáí ‘nine hundred nine’
. T
..................................................................................................................................
.. Verbal derivation Inchoative and factative verbs can be derived from stative verbs, as illustrated in Table .. Inchoative verbs are marked with -k-; the factative combines the inchoative form with the factative marker -a.
Table . Verbal derivation in Biali Stative verbs
Inchoative verbs
Factative verbs
Gloss1
bá ń tìì
bá tìì-k-ì
bá tìì-k-á
‘to be hot’
bá ń súbí-ì
bá súbī-k-ī
bá súbī-k-á
‘to be heavy’
bá ń tōū
bá tōū-k-ī
bá tōū-k-á
‘to be fat’
bá ń bíárí
bá bíárī-k-ī
bá bíárī-k-á
‘to be short’
1
The meanings here are those of the stative verbs from which the inchoative and factative forms are derived.
.. Aspect The aspectual system of Biali expresses a binary opposition: perfective vs. imperfective. The perfective is distinguished from the imperfective through formal markers or through tonal
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
markers. In the first case, the perfective is marked either with -a or with -i. The two perfective markers are substituted for -u in imperfective forms. In the second case, the opposition between the perfective vs. imperfective is expressed through tonal differences.
... Formal markers () Perfective a. ù hūb-á nā-fā. tie- ox- ‘He/she has tied the ox.’ b. ù sānd-í kúé-sí. give.birth.to- twin- ‘She has given birth to twins.’ () Imperfective a. ù hūb-ū nā-fā. tie- ox- ‘He/she ties the ox.’ b. ù sānn-ú kúé-sí. give.birth.to- twin- ‘She gives birth to twins.’
... Tone markers () Perfective a. wō-fā níím kū-bíí-m. snake- swallow: chicken-child- ‘The snake has swallowed the chick.’ b. ù yìím kúnd-há. fry: meat- ‘He/she has fried meat.’ () Imperfective a. wō-fā níīm kū-bíí-m. snake- swallow: chicken-child- ‘The snake is swallowing the chick.’ b. ù yìīm kúnd-há. fry: meat- ‘He/she is frying meat.’
.. Tense In terms of grammatical tense, only the past and the future are marked through middle and high tones.5 On the segmental level the forms are identical. 5
Metrical tense markers also exist but they are not considered here.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() a. ù tīím yùākīsī-bā. help poor- ‘He/she helps the poor.’ b. ù yí tīīm yùākīsī-bā. help poor- ‘He/she will help the poor.’ c. ù yī tīím yùākīsī-bā. help poor- ‘He/she has helped the poor.’
.. Mood In Biali, only the infinitive and the imperative or hortative are formally marked as modal categories. The infinitive is marked with bá and the imperative (Table .) with -ni. In the case of the hortative, the use of the is necessary. ()
bá sāndí ‘to give birth’ bá písá ‘to bed (sb.)’ bá tūtí ‘to share’ bá hūbá ‘to tie (sb., sth.)’ bá pùàtì ‘to dance’ bá cīnná ‘to tie up’
Table . Imperative/hortative in Biali Imperative
Hortative
Gloss
2pl
1pl
Imperative
Hortative
máání-nī tēí-nī yísī-nī
tí máání-nī tí tēí-nī tí yísī-nī
‘Comeǃ’ ‘Goǃ’ ‘Get upǃ’
‘Let’s comeǃ’ ‘Let’s go!’ ‘Let’get upǃ’
.. Negation The negative particle pa bears two different tones: low tone in the declarative negative sentence and high tone in the imperative negative sentence. () a. ā pà písí yákā. go.to.bed today ‘You haven’t gone to bed today.’ b. ā pá písí yáka! go.to.bed today ‘Don’t go to bed today!’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Focalization The focus marker lè serves as both a nominal and a verbal focus marker. Verbs marked for focus receive a progressive reading. ()
a. bíí-kā lè. child- ‘That is the child.’ m bíí-kā lè. child- ‘That is my child.’ b. m pāá lè. go ‘I’m going.’ m sūūm lè. swim ‘I’m swimming.’
.. Adverbs Adverbs in Biali are invariable words expressing semantic notions like place, time, manner, and quantity (Table .).
Table . Adverbs in Biali Adverbs
Gloss
Adverbs
Gloss
Place ɲānnī àlā àhām túmū dīshú bàmbànnī Time sānnī sīnnī yāndālī
‘there’ Manner tùòmtùòm ‘quickly’ ‘here’ ‘over there’ pímpím ‘knowingly’ ‘outside’ ‘in front of ’ yùākīyùākī ‘smoothly, gently, etc.’ ‘far’ ‘when’ Quantity kùànkùàkā ‘not much, very’ ‘yesterday’ púómpúóm ‘a lot, much, many’ ‘before yesterday’ tūātí ‘how much’
R Manessy, G. (). Les langues oti–volta: Classification généalogique d’un groupe de langues voltaïques. Paris: SELAF. Naden, T. (). ‘Gur’, in J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham, MD, New York, and London: University Press of America, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
N’ouéni, R. W. (). Contribution à l’étude phonologique du parler biali de Matéri avec application à l’établissement d’une orthographe pratique: mémoire de maîtrise. Université Nationale du Bénin. Sambiéni, C. (). Une approche du système nominal du biali: mémoire de maîtrise. Université Nationale du Bénin. Sambiéni, C. (). Le proto-oti-volta-oriental: Essai d’application de la méthode historique comparative. Cologne: Rüdiger Köppe. Sambiéni, C. (). ‘La reconstruction interne du biali, gur oriental, Bénin’, Nordic Journal of African Studies : –. Sambiéni, C. (). ‘Quelques éléments indicatifs pour une orthographe standard du biali, gur oriental, Bénin’, Gbe Studies : –. Sambiéni, C. (). ‘Du statut phonologique du schwa en biali, langue gur orientale, Bénin’, in M. Brenzinger and A.-M. Fehn (eds.), Proceedings of the th World Congress of African Linguistics, Cologne . Cologne: Rüdiger Köppe, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
-
. D
.................................................................................................................................. Y speakers are estimated to be about ,. There are several dialects; the most prominent of which are Bɔ̄dʊ̄ŋázʊ̀ and Bálíssà (henceforth BD and LD, respectively). Yukuben settlements are dispersed along the feet of the mountains separating Cameroon and Nigeria, east and south of Takum, Taraba State, Nigeria. Further landmarks include both the Gamana and Kashimbila rivers. The Yukuben people are traditionally farmers and hunters.
. P
..................................................................................................................................
.. Syllable structure Yukuben has the following canonic structures: V, VV, CV, CVC, CVV.
.. Vowels There are nine oral vowels (Figure .), which tend to show distinction in duration: i, ɪ, u, ʊ, o, ɔ, e, ε, a. Some dialects have, in addition to these, seven word-final nasal vowels that seem to have evolved from the elision of velar nasals in word-final position. Apart from these vowels, the language also has a schwa (ə), which tends to be nonphonemic. Hence, it is in the vowel trapezium. Minimal pairs are not easily found. Here are, however, some examples: () Vowel length—minimal and near/pseudo-minimal pairs ɔɔ́ ́-tsí/ɔ́ɔ́-tʃɪ ‘tree’ ɔ̄ɔ̄-tsíí ‘witch’ j j bà-ᵑg íí ‘oil’ báá-ᵑg í ‘people’ sú ‘sing’ súú ‘agree, admit’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
close
front i
back u
high
(ә) ı half close
ʊ
e
o mid
half open
ε
±ː (exception: ә)
ɔ
low open
a
. Vowel trapezium in Yukuben
.. Vowel sequences—the distribution and phonotactics of vowels The language admits word-final vowels, thereby creating an open-syllable structure. ()
Vowel in stem-final position kā-rā ‘tree name’ kì-váá ‘calabash’ kóó-zō ‘farm’ bùù-zóó ‘pieces’
Likewise, vowels occur as part of the prefix complex in initial position. ()
Vowel in word-initial position: CV- and CVV-structures bè-k͡péé ‘gods’ béé-lə̄m ‘male children’ í-ɟú ‘melon’ íí-lāŋ ‘sweet potato’
The vowel harmony (VH) is based on the ATR principle. ()
VH [+ATR] Word kī-tūū ìì-jōó éé-cíí (< /*éé-kjíí/)
Gloss ‘neck’ ‘friend’ ‘heads’
[–ATR] Word kī-lʊ̄k˺ kīī-hɔ́k˺ έέ-bīr
Gloss ‘husk’ ‘lizard’ ‘nests’
Nasal vowels occur more often in the final position than in any other position. ()
The distribution of nasal vowels a. à-pã́ã́ ‘two’ b. kí-rẽ ̄ ‘a smooth basket used in filtering’ c. kí-cɔ͂̀ ‘stone’ d. ã́-hã̀ ‘stories’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Consonants Yukuben has only egressive consonants, which can be classified as simple and complex, i.e. coarticulated, consonants. The simple consonants are those that have one and only one POA and no delayed release. They are phonemic, and twenty-four in number (Table .).
Table . Non-complex (or ‘simple’) consonants in Yukuben
Plosive
labial
dental/alveolar
palatal
velar
-voice
p
t
c
k
+ voice
b
d
ɟ
g
m
n
ɲ
ŋ
Nasal Liquid Fricative Approx.
trill
r
lateral
l
-voice
f
s
ʃ
+ voice
v
z
ʒ
non-lateral
j
lateral
ʎ
glottal
h ɣ
[t̺] and [ɾ] are dialectal variants of (presumably) the same phoneme, say /*T/, as in the word for the numeral ‘three’: [āːt̺āː] (BD) vs. [āːɾāː] (LD). Yukuben has different types of complex consonants: ()
Complex consonants a. kūū-k͡pā ‘skin’ b. ùù-g͡bá ‘rock’ c. būū-ŋ͡mā ‘salt’ d. ī-wáá ‘quarrel’ e. kú-tʃáá ‘thigh’ f. ā-dʒānóó ‘ghost’ g. ɔɔ̄ ̄-tˢíí ‘witch’ h. ūū-dzīī ‘thief ’ i. ᵐbé ‘only/just’ j. ú-ⁿdú ‘mouth’ k. bàà-ᶮɟʊ̄rʊ̄ ‘honey’
All consonant types can occur in stem-initial position. Only /k/ and /b/ are found in prefixinitial position, whereby the former occurs with singular forms and the latter with plural and uncount (e.g. mass and abstract) nouns. Only the following consonants can occur in word-final position: /p, k, m, n, ŋ, r/.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
.. Phonological alternations and processes Among the phonological processes attested are the following: assimilation, (cluster) simplification, elision, lenition, and fortition. ()
Complete regressive (or “anticipatory”) assimilation /kí + ó-kwú/ ! [kóːkwú]1 ‘it’ ‘dead’ ‘It’s dead.’
()
Partial assimilation /í + rí + [] ‘What is she eating?’
áʎí + ‘eat’
è-ré?/ ! [íréáʎíꜛèré] ‘what’
Long vowels are shortened at juncture (i.e. at word boundary). In the following example, for instance, one of the à’s of N gets deleted when juxtaposed to the preceding N which ends in a vowel. ()
Neutralization of vowel distinction—in N prefix /óó-ɲɟí/ ‘person’ + /àà-tsí/ ‘medicine’ ! [óːɲɟéꜛàꜜtsí] ‘medicine man’
Internal simplification can occur when, for instance, any of the velar plosives with palatal glide release (/*kj, *gj/) loses its [i]-coloring per bleaching, thereby resulting in palatalization [c] or [ɟ], as the case may be. ()
Internal simplification /*ū-gjīī/ ! [ūɟīː] ‘year’
Aphaeresis entails the loss of segments in word-initial position. ()
Elision of vowels in word-initial position—aphaeresis /īī-zə̄p˺/ + /ʊ́-lá/ ! [īːzə̄bꜜlá] / [īːzə̄βꜜlá] -dreg. -fire ‘charcoal’
Syncope involves the elision of segments in word-medial position. ()
Elision of segments in intervocalic position—syncope /bū + ìyíyé/ ! [būꜛìjé] ‘do’ ‘again’ ‘do again’
As in many other Central Nigerian Benue-Congo languages (including some Central Jukunoid languages), many lexical items seem to have lost their vowels in word-final position due to apocope.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
Example of lexical items that seem to have undergone apocope a. ū.kə̄.p˺ ‘bone’ b. kíí-lɔ̀k˺ ‘pool’ c. kὲὲ-lə́m ‘tusk’ d. έέ-bīr ‘nests’ e. ē-cŕ̩ ‘star’
Especially in fast speech, consonants undergo optional lenition in initial and medial positions. ()
Lenition in fast speech a. /kú-kēmbrə̄/ b. /bé-məŋ́ / c. /ē-bī/ d. /īī-zə̄p˺/ + /ʊ́-lá/ ‘dregs’ ‘fire’ ‘charcoal’
! ! ! !
[gúɣēmbrə̄] or [xúɣēmbrə̄] [b̥émə́ŋ], [βémə́ŋ] [ēβī] or [ēβ̞ī]2 [īːzə̄bꜜlá] / [īːzə̄βꜜlá]
‘pregnant’ ‘water’ ‘rain’
Whereby the g in [gúɣēmbrə̄] and the b in bé-mə́ŋ are practically devoiced in initial position: [g]̥ and [b̥], respectively. The only two plosives /k, p/ that occur in final position are characterized by no audible release (k˺, p˺). /r/ is sometimes syllabic, which implies the transfer of syllabicity onto r from an elided adjacent vowel. ()
Syllabic r̩ a. bū-bŕ̩ ‘tattoo’ b. ē-cŕ̩ ‘star’
. M
.................................................................................................................................. This section is divided into the following subsections: nouns, adjectives, numerals, pronouns, verbs, and adverbials.
.. Noun This section is divided into inflection and derivation.
... Inflection With some reordering based on features specific to Yukuben, the numbering of the noun classes (see Table .) follows partly the largely accepted system used for comparative purposes based on Bleek () and Meinhof (), the reconstructed systems of ProtoBenue-Congo (de Wolf ), as well as Proto-Jukunoid (Shimizu ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Table . Yukuben noun classes Noun class Number Class marker Semantic characteristics a
U-
humans; ( of exclusively a)
a
BA-
( of exclusively a)
b
A-
family and kinship, occupational terms; ( of b)
b
BE
loanwords, friends, job descriptions and titles, language names; ( of b, , )
U-
trees, natural and meteorological phenomena, some human organ names; ( of , )
I-
( of )
KI-
pairing objects, tree names, (shell-)fish/animals, colors; ( of a, , )
a
A-
( of , , , )
b
BE-
liquids
KI-
things, tools and instruments; ( of b, a, , )
BU-
of , ,
I-
animals, vigor designations; ( of b, )
/
I-
animals, uncount nouns; ( of , , )
BU-
uncount nouns, augmentatives, natural realities, extent and degree; (, of , )
KU-
dangling and movable objects; ( of a, , )
To mark plurality, the plural noun prefix simply substitutes the singular one. ()
Number marking via noun prefix substitutions Class pair Noun Gloss Noun Gloss a/a úú-ɲàlə́m ‘man’ báá-ɲàlə́m ‘men’ b/b ā-dʒ͡ ānóó ‘ghost’ bā-d͡ʒānóó ‘ghosts’ / ū-lə̄k˺ ‘rope’ ī-lə̄k˺ ‘ropes’ /a kíí-dín ‘tooth’ áá-dín ‘teeth’ / kāā-mā ‘knife’ būū-māŋ ‘knives’
Central to the nominal agreement of a noun class system are agreement markers. They eventually provide markings on, for instance, adjectives and numerals that match the class of the reference/head noun. ()
Illustrating agreement on elements in a sentence í-ꜜhṹ ꜜá-ꜛndà ū-də̄ŋ ŋ́ ꜜbú-ꜜdúɳ ꜛbà-mə́n bó-tòŋ -pay -give -chief the .-child -goat -five nəŋ̄ ꜛbà-dzá. kíí-ꜜcí3 -head this -adultery ‘He had to pay the chief five kids (i.e. young goats) for adultery.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . ‘Relating concords’ in Yukuben No.
RC
Pairing
Content
à-mí
of
animate (mainly humans)
à-bí
of ; goes also with ‘water’
animate (mainly humans)
à-wú
of
inanimate
à-kí
of
inanimate
à-yí
of and
inanimate
Relevant for the marking of pronouns for the noun class of the reference noun are the “relating concords” (RCs). There are two ways in which an adjective can be marked. It can occur directly following the noun or it can be introduced by any of the five RCs in Table ., especially in emphatic situations, or when fulfilling a contrastive deictic function. In the following example, the referential marker ú- is in agreement with the reference noun āá-kə̄b (< /*āá-kə̄p̚/) or its substitute ī- in /ī-hĩ /̄ , thereby agreeing also in shape with the noun’s class marker U-. ()
Adjectival marking without RC ī-hĩ ̄ (āá-kə̄b) ʊ́-cɔ̃̀. - (.-woman) -big ‘She (the woman) is fat.’
()
Adjectival marking with a RC ē-hĩ ꜛà-kí kē-tə̄n. - - -clean ‘It is the clean one (i.e. the hut).’
... Derivation and compounding A deverbal noun can participate in the derivation of a noun. () Derivation of noun with deverbal noun /ē-tú + è-k͡pé/ ! [ētúꜛèꜜk͡pé] -body nominalization-finish ‘fatigue, tiredness’ One way of forming a gerund is by encliticizing a H-toned prefix found between the verb and its L-toned prefix. ()
Gerund ù-ú-ꜜŋ͡mg͡bá ‘drinking’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Deadjectival nouns can be derived by the prefixation of a formative, such as /wV-/. ()
Deadjectival nouns wù-ú-ꜜbáx ꜜkí-ꜜbí ɲī Nominalization--hot--animal-Definite ‘hotness of the meat’
Compounds are formed by combining two different nouns. ()
Compound noun /kì-kú + bé-mə́ŋ/ ! -hole -water ‘well’
[kìkúꜜbéꜜmə́ŋ]
.. Adjectivals and associatives Adjectives have prefixes of three basic morphological forms: () kV-, () bV-, and () V̀ V́ -. The V of the kV- structure is either i or e with any of the following tone patterns: H, L, M. The forms with ki- are more predominant than those with ke-. Adjectives may or may not be accompanied by a RC of the type described above. ()
kV- adjectives (e.g. kí-cɔ̃̀ ‘big’) a. Inanimate () ɔɔ́ ́-tsí/ɔɔ́ ́-tʃɪ ʊ́-cɔ̀̃ ( ɲī) .-tree ..-big (the) ‘(the) big tree’ b. Inanimate () έέ-tsí /έέ-tʃí í-cɔ̀̃ ( ɲī) .-tree ..-big (the) ‘(the) big trees’
bV- adjectives (e.g. adjectives of evaluation) with a value adjective bɔ-̄ rəp̄ ̚ ‘be beautiful; be fine’ ()
āá-kə̄b bɔ̄-rə̄p˺ ( ɲī) -woman .-beautiful (the) ‘(the) beautiful woman’ ([āákə̄b] < /āá-kə̄p˺/ ‘woman’)
() V̀ V́ -adjectives (e.g. adjectives of physical propensity)—with ùú-nū ‘bitter’ gwórò (ꜛà-kí) ùú-nū cola (nut) (-) .-be.bitter ‘bitter cola (nut)’ (lit. ‘cola nut (which) has become bitter in a bitter state’)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Associative construction (without overt linker) ()
/ī-hέ + ɔ́ɔ́-tsí/ɔ́ɔ́-tʃɪ ̾ ! [īhέɔ́tsí] /[īhέɔ́tʃ ī] .-root .-tree ‘root of tree / tree root’
Associative construction (with linker) ()
bè-cə́ŋ nə̄ŋ ū-hwũ̄ .-liquid of -soup ‘liquidity of soup’
.. Numerals The smallest unit of counting is the number ‘’, while the highest is ‘’. Beyond this, numbering becomes complex, wearisome, and more complicated. With the exception of ‘’, all numerals follow the noun. All numerals from six to eight are formed based on the numeral ‘’. This is placed in N position, forming an associative construction with the numeral in N position. The numeral ‘’, for instance, is derived by juxtaposing the numeral ‘’ óó-tòŋ in N position with the numeral ‘’ à-pã́ã́ in N position. ()
The numeral ‘’ /óó-tòŋ + à-pã́ã́/ ! ‘five’ ‘two’
(a) [óꜛtòŋꜛàbã́] or (b) [óꜛtòŋꜛàâã́] (in fast speech)
.. Pronouns The pronominal forms presented here are: personal, possessive, relating concord, and interrogative.
... Personal The personal pronoun distinguishes emphatic and non-emphatic forms (Table .).
Table . The emphatic personal pronouns in Yukuben Emphatic (long/full) pronoun
Non-emphatic (short) pronoun
Person
Gloss
à-mà
mà/má/mā; mè/mì
I
à-mú/ò-mú
ú-/ù-/ū- (clitic)
you
è-mí
í-/ì-/ī- (clitic)
he/she/it
è-zí/è-dí/à-dí
dí, dì
we
è-ɲí/à-ɲí
ɲé/ɲè/ɲē
you
à-bí
bí/bì/bī
they
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
()
The in an emphasized (non-cliticized) context è-ꜜmí ꜜrí sí í-sí. - [] walk -walk ‘S/he or it is walking.’ (lit. ‘s/he or it is walking walk’)
()
The as clitic—in a non-emphasized context ī-ꜜrí sí í-sí. -[] walk -walk ‘S/he or it is walking.’ (lit. ‘s/he or it is walking walk’)
The in object position (see Table .).
Table . Pronouns in object position in Yukuben
()
Object position forms
Person
Gloss
mì
me
mū
you
mí/mī
her/him/it
dí/dī
us
ɲé/ɲè/ɲē
you
bì/bī
them
īn í-sà à-bá má ꜜsáŋ mī. if - -come see ‘If he came, I would see him.’ (where sà < /*sì/)
... Possessive Possessive pronouns (Table .) are usually preceded by the nouns they modify.
Table . The possessive pronouns in Yukuben Person
Number
Possessive pronoun
Gloss
é-nə́m
my
á-ndú
your
ɲ
á- ɟí
his/her
é-zí
our
é-ɲí
your
j
their
é-b í
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
Possessive é-nə́m ‘’ ù-mbá (ꜜé-)ꜜnə́m .-brother (-). ‘my brother’
... Relating concord Actually, there is no relative pronoun. There is instead a relational concord, the function of which is assumed by the same set of RCs described in section .... ()
RCs à-bí and à-kí bá-ɲɟí ꜛà-bí ꜛà-fá ꜜbí ŋ́ ꜛò-lúx (< /ò-lúk̚/) .-people -. -hate -burn. ꜜá-ꜛndà ꜜá-bā ꜜɲɟí ŋ́. ꜜkέέ-ꜜtím ꜛà-kí ꜜí-ꜜmbí .-house -.sg -build -give -father . ‘The house that he built for his father has been burnt down by the enemies.’ (lit. ‘people who hate them burnt down the house that he built for his father’) (Here, ‘build give’ is a serial verb construction doing the function of ‘build for’.)
... Interrogative Interrogative pronouns include the following: ē-yī ‘who’, è-lé ‘what’. ()
Sentence with ē-yī ‘who’ è-yí rí bá ꜛà-mə́ŋ? -who [] come -here ‘Who is coming here?’
()
Sentences with èlé/èré ‘what’ béé-lə̄m ɲī ꜜrá ljī ꜛè-lé? .-youth the [] eat -what ‘What do the children eat?’
.. Demonstrative Demonstratives (see Table .) are three-dimensional in character. ()
Demonstrative nə̄ŋ ‘this/these’ ù-kəŋ́ / è-kəŋ́ nə̄ŋ .-road / .-road demonstrative[] ‘this/these road(s)’
.. Verb Generally, with the exception of few locative and positional verbs, non-finite verbs do not have prefixes.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Table . Speaker–hearer’s orientation and demonstratives/deictic markers in Yukuben Speaker–hearer
Locatives and spatial deixis
Gloss
near and visible to both
nə́ŋ
‘this/these’
à-mə́ŋ
‘here’ (emphatic marker)
lə́ŋ () / láŋ ()
phrase-final intensifier
ó-hó
‘that/those’
far but visible to both far and not visible to either
m
ó-ꜜ bó (ꜜó-ꜜhó)
‘that/those over there’
Non-finite verbs can have any one of the following structures: CV, CVC, and CVV(C), e.g. bú ‘use’, hɔ̀ŋ ‘pluck’, lūū ‘keep’, hēēŋ ‘feed’. The verb stem is either simple or complex. ()
The verb stem a. gú ì-gú b. bā c. bū d. kí-lē e. kō-vē-rì f. kí-té-bé-rì
‘fight’ (lit. ‘fight fight’) ‘come’ + cīn ‘enter’ > ‘arrive’ ‘do/make’ + ì-yíyé / ì-yé ‘again’ > ‘re-do’ ‘exist, dwell, settle, stay’ ‘bow down’ ‘kneel down’
Syntactically, the verb as predicate head follows the subject and precedes the complement or object (e.g. nouns and pronouns). ()
A finite verb with (cognate) object bè rè gú ꜛì-gú. [] fight -fight[ ] ‘They are fighting.’ (lit. ‘they are fighting fight’)
()
A simple sentence—action í-gú m̀ . -beat. ‘S(h)e hit me.’
... Verb phrase The verb in its finite form can be preceded by a prefix (or rather a proclitic), the function of which, depending on sentence type and structure, can be referential (or in some cases anaphoric), often agreeing in number with the reference noun. For ease of application, it is glossed here as ‘-’ (i.e. referential pronominal element).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
Sentence with verb phrase containing a verbal prefix (-) a. Singular subject è-mə́n ‘goat’ ε̄ε̄-mə́n ɲī ꜛì-mbá ꜜí-ꜜdúŋ məŋ́ . .-goat the .-deliver. .-child already. ‘The goat has dropped its young (kid).’ b. Plural subject bà-mə́n ‘goats’ bà-mə́n ɲī ꜛà-mbá ꜜbú-ꜜdúŋ mə́ŋ. .-goat the .-deliver. .-child already. ‘The goats have dropped their kids.’
A new verb can be derived via, for instance, a serial verb construction (SVC). ()
SVC yielding a new verb ē-hĩ ́ ꜛà-má tú bé-mə́ŋ ɲí ꜜkrú. - - push. -water the pour ‘The water was spilled by me.’ (lit. ‘it is I/me who pushed the water pour’)
A verb may be converted to a verbal noun by the affixation of a complex prefix unit, e.g. V̀ V́ -. ()
SVC yielding a gerund a. ùú-ꜜŋ͡mg͡bá ‘drinking’ b. àá-nà ꜜbú-ꜜlá ‘going to bed’ (where nà bú-lá = ‘sleeping; fall asleep’) c. ìí-mbà ‘giving birth’
... Tense, aspect, mood As in many other African languages, the three notions tense, aspect, and mood (TAM) frequently go hand in hand, whereby especially aspect and mood are more central to the grammar of the verb and the clause than tense (cf. also Ameka and Kropp Dakubu ; Anyanwu ). At the center of the TAM systems are their markers (Table .). ()
Expressing simple past/preterite with time adverbial mà ʎí ꜜá-ꜜllá í-ʎī. eat. -food -yesterday ‘I ate yesterday.’ (lit. ‘I eat food yesterday’)
Table . Yukuben TAM marking strategies [+]
[–]
Perfective
Imperfective
Tonally
[H]; [ꜜH]
[L]; [ꜜH]; [H]
[H]; [ꜜH]
Morphologically
mə́ŋ; ꜜmə́ŋ
lè/rì; lé/rí; ꜜlé/ꜜrí (subject to VH as well as tonal rules)
rápú; fé; ʃɔ́n
Syntactically
participles
gerunds, verbal nouns
non-finite verbs (e.g. infinitives)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
()
Perfective with participial form and mə́ŋ ‘already’ kó ꜛòó-kWú (ꜜmə́ŋ). [] .-be.dead (already.) ‘It (the dog) is dead.’ (lit. ‘it has died already it is (now) in a state of death’) (where [kóꜛòó-kWú] < /kí + òó-kwú/)
()
Perfect with mə́ŋ ‘already’ mà ʎí ꜜá-ꜜllá mə́ŋ eat. -food already. ‘I have eaten already.’ (lit. ‘I eat food already’)
()
Imperfective (habitual narrative) ì-bú ꜛkì-zú kɔ᷇ŋ ꜛè-kpé ꜜí-ꜜré ꜜá-ꜜʎí, -do -(some)thing -but -[] -eating ꜜí-ꜜrí ꜜŋ͡mgbá, ꜜí-ꜜrí ꜜná ꜜbú-ꜜlá. -[] drinking -[] sleep -sleep ‘He does nothing but eat, drink, and sleep.’
()
Imperfective (present progressive) mè rí ꜜlʊ́k̚ á-lʊ́ ɲī. [] crush -grain the ‘I am crushing the grain.’
()
Future simple má ꜜʎí (ꜜá-ꜜllá) ú-rú-ꜜkpé ŋ́. (LD) eat. (-food) -sun-finish ‘I will eat this evening.’ (lit. ‘I’ll eat at sunset—for sure’)
()
Injunctive/prohibitive with fé fó ꜜsú kɔ᷇ŋ; rʊ́ŋ tə́p̚ ; sing remain still/quiet rā nā ꜜbú-ꜜlá. bá-ɲɟí .-person [] sleep -sleep ‘Don’t sing; be quiet; people are sleeping.’
()
Wish mè rápú má ꜜʎí ꜛkì-zú. eat. -(some)thing ‘I want to eat (something).’ (lit. ‘I need to eat something’)
.. Adverbials The following are examples of true adverbs: à-rəṕ ̚ ‘well’.
bé ‘only/just’, à-məŋ́ ‘here’, ì-ndǎŋ ‘now’,
m
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
An adverb in context í-ꜜbá ì-ndàŋ mbé. -come. -now only/just ‘He will come shortly.’
. S
.................................................................................................................................. Yukuben has the following basic word order: S‒V‒O. The major sentence types in Yukuben include simple and complex sentences. ()
Simple sentence with a dynamic verb mè bɔ́ŋ ꜛè-ə́rk˺. kill. .-rabbit ‘I killed a rabbit.’
()
Coordination ì-cɔŋ́ ꜛè-bẽ ̄ ꜜmbé gìɟàŋ ìí-ꜜkú, -be.ill. -much .nearly/almost ..-be.dead ŋ́ ꜛì-fá bò-kù mə̄ŋ. ꜛè-kpé ì-ndà -but -now -recover -health .already ‘He was very ill and he almost died, but now he is well.’
()
Subordinate modifier clause í-rí ꜛŋú bā-tə̄r ꜜɲɟí í-ndá -[] remove .-cloth . -because í-ꜜmgbé ꜛè-tú. -wash. -body ‘He is getting undressed so that he can wash (himself).’ (lit. ‘he is removing his clothes because he intends to wash body’)
The most predominant type of negation in Yukuben is predicate negation. In this case, the sentence as a whole is negated. This is accomplished by inserting the negative form kɔ᷇ŋ (BD) or ké-nəŋ̄ (LD) in final position. ()
Negation with a dynamic verb í-fá ꜜí-ꜜyáx ꜛà-yə́ ꜛndà é-ꜜʎí kɔ᷇ŋ. -find. .-place - give -eat N ‘He did not find anywhere to eat.’ (lit. ‘he found place where for his eating not’)
Yukuben has: () yes-no questions, and () question-word questions (or here, the è-yí/ è-ré questions).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
The yes-no question is marked by áá-ꜜá, which is also accompanied by a rise in intonation. ()
Yes-no question n ú-tə́m ꜛkì-zú ꜛndà ꜜí-ꜜbá dú ááꜜá? -prepare. -(some)thing give -husband . Q ‘Did you cook something for your husband?’ (Where the verbal complex -tə́m . . . ꜜ n dá ‘prepare give’ is a SVC for the notion ‘cook’.)
The majority of the question-word questions are generally marked by the morpheme complex è-yí/ è-ré (and also ‘à-yí/è-lé’) ()
Questions with è-yí ‘who’ è-yí sáŋ mú í-ndà yə́ ꜛù-də́ŋ -who see. -moment while -go.[] ké-hə́n è-mbè? -village -there ‘Who saw you going to the village?’
()
Questions with èré/èlé ‘what’ Ú-rí bú ꜛè-ré? -[] do -what ‘What are you doing?’
The NP constitutes the noun or its substitute (which is usually a pronoun) and a specifier. The last may be a modifier/deictic. ()
NP with determiner é-tím ɲī -work the ‘the work’
()
NP with indefinite marker bá-kə̄b ꜛà-bí .-woman - ‘some women’
()
NP with restrictive clause m è-hĩ ́ ꜛì-bú ꜜnə́m è-ɲìm kɔ᷇ŋ; é-hĩ́ ꜛà-mə́ bé. - dog .oss -bark - - just/only ‘It was not my dog that barked; it was another one.’
The following are some of the clause types attested: ()
SVA mè ꜜlé ī Jālīngò. / Jalingo ‘I live in Jalingo.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
SVCompl à-dé ꜛè-hĩ ́ Úúhūm-Kīɟī. - - Úúhūm-Kīɟī ‘We are Úúhūm-Kīɟī (alias Yukuben).’
()
SVO mè ré tú ꜜkí-ꜜbí. [] cook -meat ‘I am cooking meat.’
()
SVOA ì-bú ká ꜜí-ꜜdúŋ ɲī vr̩̄ ꜛè-tú ꜜŋgí .-dog cut. .-child the around -body . í-ʎī. léꜜláŋ all -yesterday ‘The dog bit the child all over his body yesterday.’
()
SVOO mè ndá ꜜí-ꜜdúŋ ɲī ꜜŋ͡mgbá bé-mə́ŋ. give. .-child[IO] the drink -water[DO] ‘I made the boy (= child) drink some water.’
()
SV mé ꜜdə́ŋ. go ‘I will go.’
Agreement occurs at the level of the noun and its modifiers in the predicate. This agreement (or concord) is accomplished by means of the noun classes. The marking of the subject by means of an agreement marker in the verbal phrase is compulsory. ()
Agreement a. ē-hĩ ́ ..- ‘He is a man.’ b. bé-hĩ ́ ..- ‘They are men.’
ꜜɔ́-ꜜɲɟí. .-man ꜜbá-ꜜɲɟí. .-man
c. bá-ɲɟí ꜜlé-ꜜláŋ ꜛà-pū bá-ꜛà-ʎī. .-person -all -need .--eat ‘Every man needs to eat.’ (lit. ‘persons all they need they eat’)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
R Ameka, F. K., and Kroop Dakubu, M. E. (eds.) (). Aspect and Modality in Kwa Languages. Amsterdam and Philadelphia, PA: John Benjamins. Anyanwu, R.-J. (). Tense, Aspect, and Mood in Benue-Congo Languages. Cologne: Rüdiger Köppe. Bleek, W. H. I. (). De Nominum Generibus Linguarum Africae Australis, Copticae, Semiticarum Aliarum Sexualium. Dissertation. Bonn. de Wolf, P. P. (). The Noun Class System of Proto-Benue-Congo. The Hague and Paris: Mouton. Meinhof, C. (). Grundzüge einer vergleichenden Grammatik der Bantusprachen. Berlin: Dietrich Reimer. Shimizu, K. (). Comparative Jukunoid. Vienna: Institut für Afrikanistik und Ägyptologie der Universität Wien.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. G
.................................................................................................................................. B is a Bantu language (Niger-Congo family) spoken in the Mpanda district of Katavi region in western Tanzania. The speakers of Bende are estimated at , by the Ethnologue (Lewis ) or , by the Languages of Tanzania Project (Chuo kikuu cha Dar es Salaam ), although neither of those estimates can easily be verified at the moment. The Bende language is usually considered as being in a brotherly relation to Tongwe or Bende-Tongwe (Eberhard et al. ). Tongwe and Bende are listed as F and F respectively in Guthrie’s referential classification (Guthrie ). The two languages are very similar to each other, except for a few items of vocabulary and minor morphological features (e.g. future negative form). However, the two ethnic groups were divided by history, and each of them has undergone different linguistic changes over time. The traditional classification has identified Bende and Tongwe, among the languages belonging to zones F, M, and JD, as having unique features that are different from those of other F languages and neighboring languages in significant ways (Nurse ).
. P
..................................................................................................................................
.. Sounds and phonemes Most Bende syllables are open, consisting of (N)(C)(G)V. There are ten vowel phonemes (/a, aa, e, ee, i, ii, o, oo, u, uu/), and nineteen consonant phonemes (obstruents /p, b, t, d, c, ɟ, k, g, f, s, z, h/, glides /j, w/, nasals /m, n, ɲ, ŋ/, homorganic nasal /N/) (see Table .). Note that since /c, z/ appear only in loanwords, they are assumed to be relatively recent phonemes in Bende. The homorganic nasal /N/ assimilates regressively to the following consonantal point of articulation, i.e. [mp], [mb], [nt], [nd], [ŋk], [ŋg], [ɱf], [ns], and [nV]. Vowels may be accompanied by suprasegmental features, i.e. tone and nasalization (Table .).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Phonemes and phones in Bende a [ɑ], aa [ɑː], e [ɛ], ee [ɛː], i [i], ii[iː], o [ɔ], oo [ɔː], u [u], uu [uː]
Vowels: Consonants:
Plosive
p [p], b [b~β], t [t], d [d~l], c [c], ɟ [ɟ], k [k], g [g~ɣ]
Fricative
f [f], s [s], z [z], h [h~p]
Nasal
m [m], n [n], ɲ [ɲ], ŋ [ŋ], N [m~ɱ~n~ ŋ]
Approximant
w [w], j [j], l [l]
Table . Suprasegmental diacritics in Bende Tonal:
á (High), a (Low), ǎ (Rising), â (Falling)
Nasalized:
ã
.. Tones Bende has a distinctive tonal system, as do many other Bantu languages. Tone is not assigned to each syllable, but a tonal type is assigned to each root or to a root with suffix(es) (hereafter “base”). Tonal types are three: H(igh), L(ow), and ø (unspecified). The H type is realized with all high tone syllables, while the L type has a low tone on the first syllable of its root. The ø type has only one high tone consistently on the penultimate syllable of its word form. The ø type includes many loanwords from Swahili. Table . shows tonal types by number of syllables. Four tones—H(igh), L(ow), R(ising), and F(alling)—are observed in the surface form. Tones over a whole word are realized with the combination of affix tones, which are high or unspecified. Tones in a sentence are realized differently, depending on the context of each utterance.
Table . Tonal types according to the number of syllables (σ) in Bende 1σ
2σ
3σ
4σ
-2σ verb
Pr-2σ noun
H type
H
HH
HHH
HHHH
-βá.mbá ‘stretch (leather)’
má-ɟá.βó ‘fruits’ (cl6)
L type
R
LH
LHH
LHHH
-βa.ndá ‘prong’
má-fi.ɣá ‘cooking-stones’ (cl6)
ø type
R
RL
LHL
LLHL
-soó.ka ‘descend’
ma-βeé.le ‘breast’ (cl6)
HL
LHL
LLHL
ø’ type
si-tímu ‘light’ (cl7)
.. Glide, vowel coalescence, and nasalization Glide formation, vowel coalescence, and nasalization result from morphopholonogical processes (Table .). Glide formation and vowel coalescence occur in the vowel sequence
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Glide formation and vowel coalescence (VV) in Bende V2
i
e
a
o
u
i
ii
je
ja
jo
ju
e
ii
ee
ja
jo
jo
V1
a
ee
ee
aa
oo
oo
o
we
we
wa
oo
oo
u
wi
we
wa
oo
uu
(VV). When V and V are the same, the sound remains the same (e.g. ii > ii), whereas when V and V are not the same, V may form a glide (e.g. *iu > ju), or V or V may assimilate into the other vowel, i.e. vowel coalescence (e.g. *ei > ii). ()
Vowel coalescence and glide formation mooŋgá (< *mu-oŋgá) mjoŋgá (< *mi-oŋgá) -river -river ‘river’ ‘rivers’
In Bende, nasalization occurs in the presence of a nasal consonant and [h] triggers nasalization of the following vowel. There are two contexts for nasalization: type hṼ < *NhV, and type NVhṼ < *NVhV. () Nasalization type hṼ < *NhV hũliíké. (< *N-hulik-ílé) uhuliíké. (< *ú-hulik-ílé) -hear- -hear- ‘I hear.’ ‘You hear.’ (no nasalization) ()
Nasalization type NVhṼ < *NVhV ́ (< *mu-hiiɣ-i) âahiíɣi (< *ba-hiiɣ-i) muhĩĩ ɣi -hunt- -hunt- ‘hunter’ ‘hunters’ (no nasalization)
. M
.................................................................................................................................. Eight parts of speech are identified in Bende: nouns, adjectives, demonstratives, numerals, adverbs, prepositions, conjunctions, and verbs. From the morphological point of view, they are grouped into three: nominal (nouns, adjectives, demonstratives, and numerals), verbs, and particles (adverbs, prepositions, and conjunctions). A nominal consists of “prefix–noun”, and a verb has the form “prefix–root–suffix”.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Nominal morphology Nominals (nouns, adjectives, demonstratives, and numerals), which basically consist of “prefix–noun”, form a noun class system based on the prefixes and their grammatical concords. Bende has a fairly typical Bantu class system with eighteen noun classes, including three locative classes. The prefixes are further divided into three types with minor differences: noun prefixes, pronominal prefixes, and numeral prefixes. () Prefixes for nominals a. Noun prefix b. Pronominal prefix c. Numeral prefix mú-ntú mú-ntú ɣw-ápí mú-ntú ɣú-mwı̌ -person -person -black -person -one ‘person’ ‘black person’ ‘one person’
... Nouns Most nouns have inherent singular and plural classes, i.e. class (singular) has its paired class (plural). Table . shows typical singular/plural pairs in Bende. Uncountable nouns are found in class (ma-ansí ‘water’), class (lú-heehí ‘dry cold season’), and class (kú-tendá ‘to say’, a gerund of -tendá ‘say’). Classes , , and are locative classes. These three are distinguished as follows: class refers to a definite place which is usually visible, class refers to an indefinite, usually invisible place, and class refers to the inside of the following noun. Class – prefixes precede a noun with a noun prefix, i.e. the locative classes usually take double prefixes. ()
Locative class with lú-ɣǒ ‘house’ (class ) Class (definite) Class (indefinite) Class (inside) há-lu-ɣó kú-lu-ɣó mú-lu-ɣó ‘at, to, from the house’ ‘at, to, from the house’ ‘in the house’
Pairings of classes and may express a part–whole relationship. () - class pairs (part–whole) âw-ásí lw-ásí -grass -grass ‘a rootstock of grass’ ‘a string of grass’
... Adjectives Some adjectives require a noun prefix or a pronominal prefix, depending on the stem, although those with a noun prefix are more common. The adjectives with a pronominal prefix are limited to basic colors (white, black, and red) and quantifiers. () Noun prefix-adjective ma-ɟaâo mú-kulú -fruit -big ‘big fruits’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Regular noun class pairings in Bende SG
PL
Noun prefix
Example
1
2
mu- ~ ø-
βa(a)-
múntú, βántú
‘person’
3
4
mu-
mi-
mwéɣó, mjéɣó
‘heart’
5
6
ii- ~ li(i)-
ma(a)-
lítándá, mátándá
‘spider’
7
8
si-
fi-
sîntú, fîntú
‘thing’
9
10
N- ~ ø-
N- ~ ø-
nsimbá, -
‘lion’
lúɲélé, ɲélé
‘hair’
káɲóɲí, túɲóɲí
‘bird’
11 12
lu13
ka(a)-
tu(u)-
14
βu-
βulwěle, malwěle
‘disease’
15
ku-
kúβokó, máβokó
‘arm’
16
ha-
háluɣó
‘at the house’
17
ku-
kúluɣó
‘at the house’
18
mu-
múluɣó
‘in the house’
() Pronominal prefix-adjective ma-ɟaβo ɣá-áβé -fruit -red ‘ripe fruits’
... Numerals Cardinal numbers, which have a numeral prefix, are limited to numbers one through five, and are concordant with the modified noun by class. The other cardinal numbers (six and above) do not take agreement with the modified noun (Table .).
Table . Numerals – (only – take a numeral prefix) in Bende 1
-mwı̌
6
múkaaɣá
2
-βilí
7
ndúi
3
-tatú
8
múnaané
4
-neé
9
kéndá
5
-taánu
10
íikumí
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() Numerals (with and without agreement) mú-ntú ɣú-mwı̌ βá-ntú ndúi -person -one -person seven ‘one person’ ‘seven persons’
Table . Verbal slots and affixes in Bende PROC= Subject- TAMn- Objectn- RootSMkuká-
te= ke= se= a= βo= ti= tu= rel=
objli-
sialóónááliláákúnaákóóká-
Extensionn-
Final-
=ENC
-aɣ -am -al -uk -ul -ik -an -il -i~-isi -u~-iβu
-a -í -á -é -âŋga -íle
=ho =ko =mo =tu =elo =itje =hě =ehi =nu
Functions of the affixes PROC
na-
semelfactive
Final
te=
negative
á-
consecutive
-a
indicative
ke=
‘stop’
kóó-
habitual
-í
negative
se=
‘come’
ká-
distal, itive
-á
imperative
a=
associative
Object
-é
subjunctive
βo=
focus
li-
ti=
quotative
Extension
-íle
tu=
‘only’
-aɣ
iterative
ENC
-am
positional
=ho locative,16 =ko
Subject
reflexive
-âŋga intensive anterior
ku-
gerund
-al
extensive
ká-
itive
-uk
separative-intransitive =mo locative,18
-ul
separative-transitive
=tu
TAM
locative,17 ‘only’
si-
negative
-ik
impositive, stative
=elo affirmative
a-
past
-an
reciprocal
=itje emphatic
lóó-
future
-il
applicative
=hě ‘where’
náá-
irrealis
-i~-isi
causative
=ehi negative
li-
historical past
-u~-iβu passive
láákú- immediate future
=nu ‘now’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Verbal morphology Basic Bende verb structure consists of a root, which then takes prefixes and suffixes on both sides. Thus, the structure has the following slots: subject—TAMn–objectn–root– extensionn–final. The TAM, object prefixes and the extension suffix slots may take multiple affixes, shown as [n]. In addition to basic structure, clitics (proclitics/enclitics) may be attached to both ends of the structure, as shown in Table .. Both inflectional and derivational affixes appear in each slot, as listed in Table .. Derivational suffix(es) in extension slots may derive the meaning of the verb root. The subject and TAM slots mainly contain inflections. A minimal construction consists of
Table . Concordance with persons and classes in Bende Personal pronoun Subject marker (sm) Object marker (obj) sg 1 úuné 2 úuɣwě 3 ɲené
N- ~ iNu- ~ ɣua ~ e- ~ ɣa-
Nkumu-
pl 1 úufwě 2 úumwě 3 βeené
tumuβa-
tumi- ~ muβa-
Class Dem. (near)
Subject marker (sm) Object marker (obj)
1
ɣoóɣu
a- ~ e- ~ ɣa-
mu-
2
βaáβa
βa-
βa-
3
ɣoóɣu
ɣu-
mu-
4
ɟeéɟi
ɟi-
mi-
5
leéli ~ ljěli
li-
li-
6
ɣaáɣa
ɣa-
ɣa-
7
seési ~ sjěsi
si-
si-
8
feéfi ~ fjěfi
fi-
fi-
9
ɟeéɟi
ɟi-
ɟi-
10
seési ~ sjěsi
si-
si-
11
loólu
lu-
lu-
12
kaáka
ka-
ka-
13
toótu
tu-
tu-
14
βoóβu
βu-
βu-
15
koóku
ku-
ku-
16
haáha
ha-
ha-
17
koóku
ku-
ku-
18
moómu
mu-
mu-
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Verb conjugation (simple forms) in Bende a. Past affirmative tw-a-tend-á tw-a-ká-tend-á tw-a-tend-âŋga tw-a-ká-tend-âŋga tw-a-na-tend-á tw-a-na-ká-tend-á tú-tens-ílé tú-ká-tens-ílé tw-a-li-tend-á tw-a-li-tend-âŋga tú-náá-tend-e tú-náá-ka-tend-e tú-náá-na-tend-e tú-náá-na-ka-tend-e
‘we said just now’ (today) ‘we said’ (prehodiernal) ‘we have already said’ (today) ‘we have already said’ (prehodiernal) ‘we have finished saying’ (today) ‘we have finished saying’ (prehodiernal) ‘we were saying’ (now, today) ‘we were saying’ (prehodiernal) ‘it is true that we said’ ‘it is absolutely true that we said’ ‘if we said (today) . . . ’ ‘if we said (prehodiernal) . . . ’ ‘even if we said . . . ’ (today) ‘even if we said . . . ’ (prehodiernal)
b. Present affirmative tu-kóó-tend-á tu-si-tend-á tu-sj-a-tend-á tu-na-tend-a
‘we usually say’ ‘we are still saying’ ‘we are still saying’ ‘we are able to say’
c. Future affirmative tu-lóó-tend-á tu-loo-ká-tend-á tu-láákú-tend-á tú-tend-e tú-ká-tend-e tú-tend-êŋge tú-ká-tend-êŋge tw-á-tend-á
‘we will say’ (today) ‘we will say’ (posthodiernal) ‘we will say immediately’ ‘we shall say’ (today) ‘we shall say’ (posthodiernal) ‘we shall certainly say’ (today) ‘we shall certainly say’ (posthodiernal) ‘and then, we will say/we said’
d. Past negative te=tw-a-tend-á te=tw-a-ká-tend-á te=tw-a-tend-âŋga te=tw-a-ká-tend-âŋga te=tw-a-na-tend-á te=tw-a-na-ká-tend-á te=tú-tens-ílé te=tu-ká-tens-ílé te=tú-náá-tend-e te=tú-náá-ka-tend-e te=tú-náá-na-tend-e te=tú-náá-na-ka-tend-e
‘we didn’t say’ (today) ‘we didn’t say’ (prehodiernal) ‘we didn’t say any longer’ (today) ‘we didn’t say any longer’ (prehodiernal) ‘we have said, haven’t we?’ (today) ‘we have said, haven’t we?’ (prehodiernal) ‘we were not saying’ (today) ‘we were not saying’ (prehodiernal) ‘if we didn’t say . . . ’ (today) ‘if we didn’t say . . . ’ (prehodiernal) ‘even if we didn’t say . . . ’ (today) ‘even if we didn’t say . . . ’ (prehodiernal)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
e. Present negative te=tu-kóó-tend-á te=tu-na-tend-a
‘we don’t usually say’ ‘we are not able to say’
f. Future negative te=tw-a-tend-e te=tw-a-ká-tend-e tú-si-tend-í tú-si-ká-tend-í te=tw-â-tend-e
‘we will not say’ (today) ‘we will not say’ (posthodiernal) ‘we shall not say’ (today) ‘we shall not say’ (posthodiernal) ‘and then, we will not say/we didn’t say’
“root–final”, which forms an imperative (lól-á ‘look!’ ()). The TAM slot takes maximally four prefixes, the object slot maximally two, while the derivational slot does not have a precise limit. ()
Basic verb structure with clitics te=tw-a-ká-si-mu-ɣús-j-âŋge=éhi? =-----buy--= ‘We didn’t sell it to him/her, did we?’ (prehodiernal past)
A subject marker or an object marker may appear in the subject or object prefix slot respectively, with the agreement of subject or object by person or class (see Table .). Bende has a rich conjugation system with twenty-six tense and aspect patterns for simple affirmatives, and nineteen tense and aspect patterns for simple negatives. Table . shows the list of simple forms, i.e. main clause forms in a single verb structure. It gives sample sentences of the first person plural (tu- ~ tw-) and a verb root (-tend ‘say’) with free translations, using several of the affixes listed in Table ..
. S
..................................................................................................................................
.. Word order The canonical word order of Bende is S–V–O and NA (noun followed by adjective), like the majority of Bantu languages. Bende tolerates some variation in word order, depending on the focus of the sentence, although the agreement of a logical subject is obliged to occur in the subject slot of the verb structure except in the case of locative inversion. For ditransitive verbs (-ha ‘give’), a dative (indirect/person) object is primarily chosen as a single in a verb structure (a). However, two s are allowed if the first refers to a non-human and the second to a human (only for , , and markers that
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
include a nasal sound), as in (b). The word order of the object nouns in both (a) and (b) is obliged to occur in the order indirect object followed by direct object (‘the children’ and ‘book’). ()
Ditransitive verb requiring two objects a. One in object slot úune na-a-βa-h-ǎ βa-aná si-taáβu. . ---give- -child -book ‘I gave the children a book.’ b. Two s in object slot úune na-a-si-mu-h-ǎ̃ βa-aná si-taáβu. . ----give- -child -book ‘I gave the children a book.’
In the case of passivization, only the dative object can be the subject. ()
Passivized version of () βa-aná βa-a-h-w-ǎ si-taáβu no=óné. -child --give-- -book by=. ‘The children were given a present by me.’
Locative inversion is common in Bende. A locative phrase (formal subject) appears in a preverbal position, and the verb agrees with the locative in class, while a logical subject following the verb is not marked for agreement on the verb. ()
Locative inversion (formal subject: hákiɟiíɟi, logical subject: múntú ) há-ki-ɟiíɟi ha-ká-fw-ı̌lé mú-ntú. --village --die- -person ‘A person died in the village.’
Except for locative inversions, Bende does not tolerate subject inversion. Post-verbal subjects may appear, as in (), although here the verb agrees with the logical subject. ()
Post-verbal subject in a song βa-ká-fw-ílé holdámu ná ka-aná ka-aɟe βe-é-fúmbéét-é. --die- Holdam and -child -: --grasp- ‘Holdam and his child died while they were holding each other.’
.. Major sentence types ... Simple sentences Bende has four simple sentence types: declarative, interrogative, imperative, and exclamatory. The declarative type of sentence is the most common, producing statements with a declining intonation. Two types of interrogative sentences, a yes/no question and a content question, have an outstanding high pitch on the H tone(s) of the final word of the sentence.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The imperative has two specific moods: the simple imperative and the subjunctive. The latter implies politeness, and it often takes a proclitic ke=, which was historically derived from the imperative léké ‘Stop!’. ()
imperative sentence types a. Simple imperative mood ɟís-á! ɟís-é! come-. come-. ‘Come!’ () ‘Come!’ () b. Subjunctive mood u-tende! kj=ó-tend-e! -say- =-say- ‘Please say!’ () ‘Wait, please say!’ ()
Exclamatory sentences appear with the interjections ee βaáβo! ‘Hey, please!’ or βúti ‘How!’, or with emphatic clitics like tí= or =itje. The tag question enclitics =élo, and =éhi are also used in exclamatory sentences.
... Complex sentences Coordination consists of two or more simple sentences, while the following four clause types appear to create subordination: subjunctive, consecutive, temporal, and relative clauses. A subjunctive clause takes a verb conjugation with -e in the final slot. A consecutive clause takes a consecutive prefix á- in the TAM slot. A temporal clause takes the class relative clitic ho= preceding a ‘when’ clause. () Subjunctive clause (irrealis TAM marker náá- and subjunctive final -e) ú-náá-tend-e mpáke é-ɣúlú li-tándúk-e n-si-na-fum-a. --say- till -sky -split- ---go.out- ‘I will not go out, even if you keep saying (so) until the sky splits up.’ ()
Consecutive clause mw-â-mú-lól-á ɲánsálá mw-â-mw-íháɣ-íl-a koóko=é. ---see- Nyansala ---kill-- .= ‘If you see Nyansala, you should kill her there.’
()
Temporal clause hó=βa-a-hulik-a βú-lwěle bhw-âke βa-a-jís-a .=--hear- -disease -his --come- Mú-kú-mú-lól-á. ---see- ‘When they heard about his sickness, they came to see him.’
Relative clauses are marked only by a H tone on a SM of the verb, when the subject is relativized. A relative clitic appears only when the object is relativized.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() Relative clauses a. Subject is relativized (H tone on the SM) mú-ntú á-koo-ɱ-fúnááníl-á ɣa-a-fw-a. -person ---help- --die- ‘A person who always helped me has died.’ b. Object is relativized (relative clitic) mw-enda ɣô=n-teh-ílé ɣw-a-hw-ílé=ehě? -cloth .=-like- --finish-=where ‘Where was the cloth which I liked carried away to?’
R Chuo kikuu cha Dar es Salaam. (). Atlasi ya lugha za Tanzania. Dar es Salaam: LoT Project, University of Dar es Salaam. Eberhard, D. M., Simons, G., and Fennig, D. (eds.) (). Ethnologue: Languages of the world, nd edn. Dallas, TX: SIL International, https://www.ethnologue.com accessed July , . Guthrie, M. (). The Classification of Bantu Languages. London, New York, and Toronto: Oxford University Press. Lewis, P., Simons, G., and Fennig, D. (eds.) (). Ethnologue: Languages of the world, th edn. Dallas, TX: SIL International, https://www.ethnologue.com accessed July , . Nurse, D. (). ‘Towards a historical classification of East African Bantu languages’, in J.-M. Hombert and L. Hyman (eds.), Bantu Historical Linguistics: Theoretical and empirical perspectives. Stanford, CA: Center for the Study of Language and Information Publications, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. W (nyan wɩyáὺ) is the largest language of Tula-Waja, a subgroup of Adamawa-Gur.1 Waja is mainly spoken to the west of the Lower Gongola, in Gombe State, Nigeria. The main dialect area, Wɩɩ, comprises the foothills and plains north of the Longuda Plateau; Deri (nyan derúwò) is the dialect spoken on the plateau. The core groups of the Waja people are autochthones whose numbers were significantly increased by migrants from the neighboring groups (Kleinewillinghöfer ).2 This sketch is based on the Wɩɩ lect spoken in Dong/Talasse. Waja is a noun class language in which concord marking is observed. The basic word order in clauses is S–V–O; the head noun generally precedes modifiers.
. P
..................................................................................................................................
.. Consonants All stops, /m/, /r/, and /y/ have labialized variants. Prenasalized stops, common in Wɩɩ lects, often correspond to simple nasals in Deri. /h/ is rare in Wɩɩ lexemes, but common in Deri (see Table .).
Cf. Kleinewillinghöfer, chapter of this volume on Adamawa. Waja and its Chadic neighbors share a substantial common vocabulary, as well as several grammatical features. 1 2
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Consonants in Waja p pw
b mb bw
t tw s sw
m mw
d nd dw
c cw
n r rw l
j nj jw
k
g ŋg
kw
gw
ny
ŋ
y yw
w
(h)
Table . Vowels in Waja [+ATR] i
[–ATR] u
ɩ
[ә] e
ʋ [ɜ]
o
ε
ɔ a
.. Vowels The nine vowels (Table .) belong to two harmony sets differentiated by the feature advanced [+ATR] versus non-advanced tongue root [–ATR]. Harmony rules govern the distribution of these sets: (i) word stems contain either [+ATR] or [–ATR] vowels; (ii) vowels of affixes and most clitics adapt to the [ATR] feature of the respective stems, with the exception of /a/, which in most bound morphemes has no [+ATR] variant. The central vowels [ә] and [ɜ] only occur in inter-consonantal position, where they are generally allophones of /i/, /u/, /ɩ/ or /ʋ/. Where their underlying phonemic status is uncertain, the surface form has to be maintained in the transcription. ()
dәne ‘name’ dәnè ‘knee’3 bәgù ‘giving birth’ pɜgὺ ‘planting’
Vowel length, phonologically a sequence of single vowels, is distinctive. Length also results from the suffixation of grammatical elements.
3
Compare Tula: dәn/dimbi ‘name’; Tso: diǹ ‘name’; and Tula: dә̀-dun ‘knee’; Cham: dun ‘knee’.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
puri ‘grave’ ma mɔ laa. =build. house ‘You built a house.’
puuri ‘sheep’ ma mɔ-ɔ laa. =build.- house ‘You built a house over there.’
.. Tone Waja has two tonemes: high (H) and low (L). As result of ‘pausal lowering’,4 an additional non-high (NH) surfaces lexically. Tone-bearing units (TBU) are vowels and syllabic consonants. Falling tones (H-L) are H + L linked to one TBU. ()
HH LH HL H-L
tέnyέ laa nεrὲ yô
‘how?’ ‘house’ ‘person, man’ ‘we’
The major function of tone-marking in Waja is grammatical. Only a few nouns are lexically differentiated by tone (cf. ()); tonal contrast between verb stems likewise results predominantly from inflection/derivation. ()
lɩya! ‘Delay someone!’ lίyà! ‘Be lost!’ yárà! ‘Fall!’ yará! ‘Land!’
Tonal contrasts mark aspect and mood (cf. Table .) and valence on verbs (); tone distinguishes noun class pronouns and suffixes (cf. Table .), and direct and indirect object pronouns. ()
a pίrá-yír-ò. a pίrá-yir-o. =ask-.- =ask-.- ‘S/he asked us.’ ‘S/he asked for us.’
Several grammatical morphemes are associated with a floating tone which links within the phrase to a following morpheme, as in (), with the floating H of the imperfective -rε´. ()
mɔ.rε´. mɔ.r láa. ()build. ()build. house (cf. laa ‘house’ in ()) ‘S/he builds.’ ‘S/he builds a house.’
Downstep |!|, the lowering of the level of subsequent tones within the phrase, results, e.g., from the segmental deletion of L-toned morphemes, while the tones remain. The pluractional -ŋà in (a) is deleted when an object pronoun follows. The tonetic realizations contrast the phrases in (b) and ().
4 Phrase-initial H and phrase-final H are generally lowered to NH. Phrase-initial L is realized as NH, and phrase-final L surfaces as L. Sequences of two H in either phrase-initial or phrase-final position remain unchanged.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
a. a pίrá-ŋà. ‘They ~ s/he asked (repeatedly).’ b. a pίrá-!yírò. ‘They ~ s/he asked us (repeatedly).’ a pίrá-!yiro. ‘They ~ s/he asked (repeatedly) for us.’
. M
..................................................................................................................................
.. Noun and noun phrase All Waja nouns are assigned to classes; the class markers (.) are suffixed. Concord (grammatical/syntactical agreement) is marked by a set of distinct concord morphemes: class pronouns (.) and suffixes (.). The set of . varies significantly in number and form from the set of concord markers (cf. Table .). Taking into consideration that discrete semantic properties are associated with specific ., and/or specific class pairings (genders), the set of concords is regarded as a reduction of a formerly more elaborate system. The (proto-)system given in Table . is, therefore, presented as a diachronically biased extrapolation. The single classes at the
Table . Concord markers in Waja1 cl.pn wU wU´ kU dә pU´
cl.suf -E -Ú -Ù -Ì -Ú
cl.pn bә sә tә´
cl.suf -wa -Ì -Í
1 Vowels written with capitals and the central vowel [ə] follow harmony rules; N [m ~ n ~ ŋ] adapts to the phonological environment.
Table . Class pronouns and agreeing nominal suffixes in Waja cl.pn wu
wu´
nmlz.suf -E, -NE ~ -N -o ~ -a -Ú -ŋ́ ~ -ḿ
cl.pn bә
tә
ku
-gÙ ~ -Ù
dә
-I, -E -rE, -nE
tә bә
-Í -ma
pu´
-Ú
sә
nmlz.suf -wa ~ -a -mba -ndi -I -dI ~ -[ɾI] -rI -ŋI -ndI
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
bottom mainly contain masses, liquids, and abstract nouns where number distinction is not salient; is a locative class.
Table . Noun class system in Waja 1
we
ba
2
1a
-ø
-mba 2a
25
o
i ~ si 4, 13
3
Ú
-di
*ci
ni
10
5
di
ŋi
6
15
ku
ti
21
?
-I´
23
ma
21
ti1
16
PÚ
1 The large number of mass and abstract nouns and derivations in cl. warrants its establishment as single class.
Identification with cognate/corresponding genders and classes reconstructed for (Central) Gur is shown by the numbers and markers to the left and right. They follow the system outlined in Miehe et al. (: –).5 Table . exemplifies the class system in Table .. It also shows the major means of forming plurals, by substitution of the singular ..
Table . Genders and single classes in Waja /
nεr-ὲ / nʋ-wa nʋr-ε / nʋrɜ-mba
‘person, man’ ‘woman, wife’
/
sɜŋ-ε / sɜŋ-ndɩ jiŋ-ne / jiŋ-ndi
‘python’ ‘fish’
/-dɩ
jum-e / jum-di
‘horn’
/
puur-e / puur-i
‘sheep’
/
gun-ò / gun-ì
‘goat’
/ . . . /-dɪ
deré-ú / deré-ndi kεl-ʋ / kεl-rɩ
‘scar’ ‘fallow’ Continued
5
Plural suffix -ndɩ in Wɩɩ corresponds to -nɩ in Deri and Tula; suffix -dɩ is assumed to be cognate to class *ci in Gur (Miehe et al. :–).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Continued /ŋɪ
bak-ὶ / bakә-ŋɩ cʋr-ὲ / cʋrɜ-ŋɩ mwa-nὲ / mwa-ŋɩ
‘forest’ ‘heart’ ‘food’
/
gund-ù / gundu-ri
‘water’
/
swaat-ʋ / swaat-ɩ
‘bamboo’
-í -í/(ŋɪ)
rʋn-ί yɩrɩ (/yɩrɩ-ŋɩ)
‘provocation’ ‘thing’
/()
tɩɩ-mà / (tɩɩma-ndɩ)
‘blood’ / (‘types of ~’)
pɩ ~ pʋ
‘place’
.. Concord Concord is marked on attributive modifiers, pronouns, associative/possessive constructions, and relative/locative/temporal clauses. Bilateral marking is the basic pattern: dependent constituents/morphemes/clauses are “bracketed” between the two markers, . and .. Examples are adjectives (), the numeral ‘one’ (), absolute pronouns (), associative constructions (), and relative (a) and locative/temporal clauses (b). ()
nεr wʋ duw-e ‘big man / person’
()
sәmá bi-ino-wà ‘one land / generation’
() k-ɔ-mɔ́-ὺ ‘your own’ (cl. ku) ()
nendi sә́ la-ɩ ‘domestic animals’ (lit. ‘animals for the house’) dәn dә́ nεr-ɩ̂ ‘(a) name for a person’ (cf. ())
()
a. nɔɔ́ bύ sɔm-bʋ nύrύ-wà ‘the people who don’t know the woman’6 b. pʋ nέr jε mwan-ὺ, . . . ‘when / where the person ate the meal, . . . ’7 kʋ baa-wʋ, yi tusó wύraὶ. ‘When he came, we started the work.’
Concord on demonstrative/locative pronouns () and the copula -ɔ́ɔ́ () is marked with the . only, whereas in possessive constructions () the . is the only marker. ()
guno wʋ-nà ‘this goat here’ pә-na ‘here’ pʋ-wɔ ‘there (near you)’ pw-ɩɩ ‘over there’
() gundu k-ɔ́ɔ́. ‘Water is there/available.’ yidi s-ɔ́ɔ́ yemengû. ‘Birds are flying.’ 6 Relative clauses may also often be formed with a demonstrative pronoun in the position of the .: bwala kʋ ma tusó kύwɔ, ‘The fire that you started there’. 7 . marks locative and temporal clauses, . only temporal.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
laa-mɔ́-ύ gundu-mó-ù dәn nεr-ɩ̂ nyɩɩmá bʋya wύna-wà
‘your house’ ‘your water’ ‘(the) name of the person’ (cf. ()) ‘milk of this child’
.. Adjectives Adjectives form their own word class. Some have singular and plural forms like nouns. A large subclass terminates with -k.8 A small number of adjectives () are not attested with an independent citation form; when cited, they generally relate to a head and bear the agreeing .. ‘pleasant’ banyɩna / nyɩtɩtɩ ‘small, little’ () tɩŋ dúrúmbolòm ‘round’ lεpὲ-k / lεpε-ndɩ ‘wide’ ()
kwεrὲk ‘bitter’ cɔkɔrɔk ‘long, high’ bʋbὺk ‘soft’ pɔpʋlɔk ‘white’
()
duu- ‘big’ bwɩɩ- ‘bad, destructive’ puc- ‘new’ pεε- ‘red’
In attribution, adjectives are typically bracketed between the markers (). Final -k is often but not always substituted by the .. ()
kεn dɩ pɔpʋlɔ́-ὶ ~ kεn dɩ pɔpʋlɔk ‘white metal; zinc’
.. Personal pronouns Waja differentiates six persons.9 and differ from the . of gender / (Table .).
Table . Personal pronouns in Waja independent
8 9
possessive
object direct ()
indirect ()
mi mɔ̂ yɔ̂
-mí-mÓ-yÓ-
-ŋ́g-mə́n-`
-ŋən-mən-
yô mbâ pḿmà
-yír-mbÍr-pḿmɔ-
-yír-` -mbÍr-`
-yir-mbIr-
Adjectives ending in -k are also common in neighboring Chadic languages. Unlike most Tula-Waja languages, Waja does not distinguish inclusive/exclusive in .
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Object pronouns are suffixed to the verb stem, preceding TAM markers. and have no specific object pronouns; independent pronouns are used instead, following the TAMmarked verb like a nominal object. () a pɔ-yίr-ò. =catch-.- ‘S/he caught us.’ a pwέ-ὲ pḿmà. =catch- they ‘S/he caught them.’ ()
cá-ŋзn-ύ=wὲὶ. ()cook-.= ‘S/he will cook for me.’ cá-mán-ύ=wὲὶ yɔ̂. ()cook--= s/he ‘S/he will cook for her/him.’
.. Verb and verb phrase ... The verb forms Waja verbs have three basic forms: two contrasting aspectual verb forms, definite () and indefinite (), and the verbal noun (). The principal markers are -o, -a, and -yo, -ya with CV-verbs (cf. Table .(a), (b)). They mark of all simple and extended verbs, except a verbal subclass which uses the suffix -Ì instead (cf. Table .(c)), and verbs extended with the altrilocal–ventive extension (cf. Table .(d)). is marked by suffix variants with two distinct tone patterns: -(y)I´, -rE´ ~ -dE´, ~ -dI´, ~ -nE´, and -dÎ ~ dÊ ~ -ŕÈ, -ÍÌ. Their distribution relates to the phonological structure of the stems; -ÍÌ is restricted to a subset of intransitives (cf. Table .(d)).
Table . Verb forms of simple verbs in Waja
gloss
(a) náà ( /ṭ/, /ẓ/ > /s: /, and /ɣ/ > /x/).
... Emphasis opposition Zenaga, like all other Berber languages, has two main emphatic phonemes, /ḍ/ and /ẓ/. The other emphatic consonants often appear to arise as contextually determined emphatic 1 This emphasis, which is characterized by pharyngealization (or velarization), is transcribed here by a subscript dot (i.e. [ṭ] for [tˤ] and [tɣ]). The labiovelarization of the bilabial consonants, whose phonological status is unclear, is also written as a subscript dot.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
variants. Their status as phonemes (either rare or with low frequency) results from the fact that some of their occurrences are not contextually conditioned. Note that the labiovelarization of labial and bilabial consonants correlates with vowel backing ([u] for the close vowel).
... Length opposition Apart from some rare phonemes (mainly the laryngeal consonants and the trills), length is usually distinctive for the consonants. This contrast, however, tends to be neutralized when it is only marked by a difference in length and/or tenseness. (Tenseness alone characterizes the coda position, i.e. for the main part the final position.) In fact, in most cases, the length contrast between short and geminated consonants is more complex. It includes the following pairs: • • • •
a continuant and a non-continuant ([đ] vs. [dd], [đ̣] vs. [ḍḍ], [ɣ] vs. [qq]), a voiceless interdental consonant and a voiced sibilant ([θ] vs. [zz], [θ̣] vs. [ẓẓ]), a hushing sibilant and a hissing sibilant ([š] vs. [ss]), a semi-vowel and a liquid ([y] vs. [ll]).
The common characteristic of these contrasts is that the short consonants tend to be more lax than the geminated consonants (apart from a few particular contexts). The phonetic difference might be a real distinction in aperture or a small difference in pressure (as with the voiced hushing consonant, [ž̯] vs. [žž]).2 Zenaga has, along with several non-southern Berber languages, a tendency to weaken short consonants. However, most of those languages only spirantize the dental stops which are not adjacent to a nasal (Louali-Raynal ). Besides the phonetic particularities of the relationship between short and geminated consonants, Zenaga is characterized by some features which are common in southern varieties of Berber (or at least in some of those varieties): a diachronic development of hissing sibilants to hushing sibilants (/s/ > /š/, /z/ > /ž), the presence of palatalized consonants, and a tendency to preserve the laryngeal consonants. However, in Zenaga, this tendency affects the glottal stop more than the fricative h (preserved only after a nasal). Even if, historically, the fricative /ɣ/ has evolved into /ʔ/ (apart from in exceptional cases), the origin of many occurrences of /ʔ/ is a glottal stop. The other Berber languages have only kept partial or indirect traces of this laryngeal.
.. Syllable structure Open syllables are almost as frequent as closed syllables. Both extra-long syllables (with a long vowel followed by a simple coda or a short vowel followed by a double coda) and vocalic onsets at the beginning of words are attested. Syllable structure influences the realization of consonants, but the realization of vowels to a lesser extent. 2
For the sake of simplicity, phonetic variations of short consonants which do not change the phonological status of the segment are transcribed using subscript or superscript symbols.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Lexemes in general consist of two or three syllables; sometimes they are monosyllabic and often they consist of more than three syllables (in particular inflected verbs). Stress is rarely marked and is not fixed. It is important as phrasal accent and moves towards the end of the word if one or more clitics are present.
. M
.................................................................................................................................. Morphological variations consist of affixation and/or vocalic alternations for both nouns and verbs. The various forms of a lexeme and the various etymologically related lexemes usually have only the ordered radical consonants in common.3 However, phonetic evolution and alteration often impede the synchronic identification of this root, particularly because of the tendency of the laryngeal consonants to disappear. In Zenaga, the fricative h becomes silent (except after a nasal), and the glottal stop is dropped in final position (but reappears before a suffix).
.. Nouns ... Inflection There are two genders, masculine () and feminine (), and two number categories, singular () and plural (). The distinction between the collective and the unity nouns is sometimes expressed by a difference in gender, e.g. əššiymi ‘fish ()’ vs. tšiymiˀđ ‘(a) fish’. Apart from rare exceptions (often pan-Berber words) like fäđ ‘thirst’, all masculine nouns begin with a vowel (usually /a/ ~ /ā/): aɣayär ‘kid (young goat)’. In the case of feminine nouns, this vowel is preceded by the prefix t-: taɣayärt ‘female kid’. In the singular, most feminine nouns are also characterized by a suffix -t which often assimilates to the last radical consonant: täygaḌ ‘young female kid’ vs. äygäđ̣ ‘young kid’. In the plural, this suffix -t usually disappears before the plural marker, which contains a nasal. The plural marker is (in most cases) -ən for the feminine (tiygađ̣ən ‘young female kids’) and -än for the masculine (iygađ̣an ‘young kids’). The alternation of the pre-radical vowel ä~a / ə~i is frequent, but other alternations are possible in both genders: īđ̣ ‘night’ āđ̣an, ižiˀmär ‘lamb’ əžaˀmärän. This shows that there is not just the one plural nominal pattern (Taine-Cheikh ).
... Derivation Nouns can be derived both from verbs and from other nouns. .... • The nominal pattern with the prefix m- (sometimes n-) gives rise to forms expressing habits, trades or properties. For example, ämäddäwkty ‘(boy-) friend, husband’ əmdukkäyän, tämäddäwkəL ‘(girl-)friend, wife’ təmdukkäyən. 3
In Zenaga, such forms/lexemes can also share a long vowel resulting from the loss of a laryngeal fricative (*h).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
• The action nouns () are often masculine nouns with a closed vowel like əddən ‘process of covering’ or feminine nouns with an open vowel like täyḍḍar: t ‘process of becoming fat’. However, action nouns derived from complex verb forms usually have a prefix aˀ- as in aˀšbiˀži ‘process of making s.b. speak’. .... In contrast with the other Berber languages, where the diminutive is often expressed through a gender modification, Zenaga derives diminutive forms through affixes (Vycichl ; Taine-Cheikh ).4 • : prefix aɣ- and suffix -t (from the ): aɣdiymämt diminutive from ađiymäm ‘baby’, • : prefix äyđ-: äyđīDyəđän diminutive from īDyəđän ‘goatskins’, • and : prefix tyā-: tyāwäžuḌ diminutive from täwäžuḌ ‘hen’, tyāđätən diminutive from tātən ‘ewe’.
.. Nominal determiners Zenaga has three sets of demonstrative clitics. Only the first set (proximal deictic expressions) varies in terms of number (Table .).
Table . Nominal determiners in Zenaga Proximal
Intermediate (+ anaphoric reference)
Distal
-äđ -iđ
-iˀđ
-ān
Other determiners exist, for example to express the indefinite: äkki ‘each’ (invariable) and yuđ̣an, uđ̣anən ‘another’ (following the head noun but not cliticizing to it), as in äyiˀm yuđ̣an ‘another camel’.
.. Adjectives The few existing adjectives form a specific syntactic category (Taine-Cheikh b, ). They have no prefix and their suffixal gender and number markers differ from those of the nouns: . -äđ, -əđ (if the last radical consonant is a glottal stop, the forms are . -đäđ, -đəđ); for example, ‘blue, green’ . bäyđig, . bäyđigäđ, bäyđigəđ.
4
Tetserret alone presents a similar derivation pattern (Attayoub : ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Adjectives can be inflected like verbs, with the difference that they can only be modified by suffixes. For example, maẓẓūg-ag ‘I am short’.
.. Numerals Zenaga has kept the original Berber numerals (Blažek ; Taine-Cheikh b). When followed by a noun denoting the object which is being counted, the numerals one to ten agree in gender (suffix -äđ~ät for the feminine). For the following tens, the unit is preceded by əđ ‘with’, e.g. märäg əđ akkuẓ̄ ‘fourteen’. The following numerals are formed with tmärīn ‘tens’, tmađ̣aˀn ‘hundreds’, and ävđ̣an ‘thousands’, e.g. kar: đ̣at-tmärīn ‘thirty’ (lit. ‘three tens’).
.. Pronouns ... Personal .... The independent pronoun system of Zenaga is shown in Table ..
Table . Independent personal pronouns in Zenaga
Singular
Plural
niˀK
nəkni
nəknaˀNyäđ kəK
nətni
kəmm
nətnaˀgəmnyäđ/nətnamnyäđ
nəttä/əntä
nəhni
nəttäˀhäđ/ntäˀhäđ
nəhnaˀNyäđ
.... The clitic pronouns (Tables . and .), usually affixed to the verb, can precede the verb under specific conditions. The pronouns encoding the direct object have morphological
Table . Suffixal personal pronouns in Zenaga () Direct object Unmarked context
-i(h), -iˀh, -iˀn, -iˀđ1
-ki -käm
-ti -täđ / -tiyäđ
1
Indirect object
After preposition (except after ən(n))
-iˀ(h)
-iˀh
After a glottal stop
-āg -ki -ām -käm -iˀh -iyäđ
-āš
-š
The forms -iˀn and -iˀđ result from a fusion with the initial consonant (n or d) of the directional particles näh and däh.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Table . Suffixal personal pronouns in Zenaga () Direct Object Unmarked context After a glottal stop
-aˀnäg
-kūn -kəmny(äđ)
-āgūn -āgəmny(äđ)
-tän -təNy(äđ)
-(ə)nän -(ə)Ny(äđ)
Indirect object After preposition (except after ən(n)) -näg -kūn -kəmny(äđ) -āšän -āšəNy(äđ)
-šän -šəNy(äđ)
idiosyncrasies if the last radical consonant of the verb is a glottal stop. The pronouns used after prepositions (except after ən(n)) sometimes resemble the direct object affixes (in the second person), and sometimes the indirect object affixes (in the third person).
... Possessive Two sets of affixal pronouns (Table .) are used after the preposition ən(n) ‘of ’. One of the two, like the preceding sets, expresses possession: täyimt=ən=k ‘your she-camel’. The other, which is rare, expresses pronominal possession: ənn=əđän ‘theirs’.
Table . Possessive pronouns in Zenaga Affixal pronouns
Independent pronouns
-(i)ˀn
-(ə)nn=äˀn
ənn=uˀh~uˀn
ənn=äkkän
-(ə)n=k -(ə)nn=ūn ənn=ak~äkkəkägi ənn=ättän -(ə)m -(ə)nn=äđkəmny(äđ) ənn=äkkäm ənn=äđkəmny(äđ)
-(ə)n=š -(ə)n=šän -(ə)n=šəNy(äđ)
ənn=ih
ənn=əđän
... Demonstrative Demonstrative pronouns inflect for gender and number (Table .).
Table . Demonstrative pronouns in Zenaga Singular
Proximal Intermediate (or “ . . . at issue”) Distal
Plural
neuter
äyđ äˀđ
äđ iˀđ ān
täđ tiˀđ tān
əđniđ əđniˀđ əđnān
təđniđ təđniˀđ təđnān
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
There is, in Zenaga, an invariable predicative auxiliary äyđ=äđ ‘it is’ and three sets of presentatives, each of which is made of a demonstrative pronoun preceded by the invariable deictic word äđ-: äđ=äyđ ‘this is’, äđ=äˀd ‘that is’, äđ=ān ‘that (over there) is’.
.. Relative Only the following three invariable forms are real relative words: ār, āf, and āš (used for relativization of an indirect object or an adverbial word). These exist only in Zenaga and can also be used as oath particles (Taine-Cheikh , a).
.. Interrogative Interrogative expressions are invariable: m=äđ~taˀk=äđ ‘who?’, taˀk=äyđ ‘what?’, maˀhäg/ maˀK ‘where?’, maˀK=oˀgđ/taˀK=oˀgđ ‘when?’, taˀK=āf/iy=wäh ‘why?, for what?’.
.. Verbs ... Inflection Verbal inflection in Zenaga is illustrated in Table .. The personal indexes are always the same with only two exceptions. • the imperative, which is only used in the second person, has only suffixal plural markers: oˀgum ‘run!’, oˀgum-äm ‘run ( )!’, oˀgum-əmny(äđ) ‘run ( )!’. • the form called “participle” (Galand ), only used in the third person, is characterized by the addition of -n: . y-äskär-än, . t-äskär-än, (+) äskär-nin.
Table . Verbal inflection in Zenaga Personal indexes
‘to make’
. . . -äg
äskär-äg
t- . . . -äđ
t-äskär-äđ
y- . . . t- . . .
y-äskär t-äskär
n- . . .
n-äskär
t- . . . -äm t- . . . -əmnäđ
t-äskər-äm t-äskär-əmnäđ
. . . -än . . . -əNäđ
äskər-än äskär-əNäđ
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
... Derivation Basic verbs can have various meanings (state verbs, active verbs, middle verbs, reflexive verbs, or a combination of these meanings). The derived forms are semantically specialized. • The prefix s- (whose phonetic realization depends on the nature of the radical consonants) derives a causative or a factative form. • The prefix m(m) ~ n(n)-, which is rarely used, derives a middle or a reflexive form. • The prefix T y- added to an agentive verb (basic or prefixed by s-) usually derives a passive form.
... Tense, aspect, mood The system distinguishes three main forms: two aspectual forms, which are in opposition to each other (“perfective” vs. “imperfective”), and a basically modal form called “aorist”. The perfective and the imperfective forms often vary according to the context (positive vs. negative). Aorist forms vs. perfective forms, and positive forms vs. negative forms, are usually distinguished by vocalic alternations (Cohen and Taine-Cheikh ). The imperfective forms are usually characterized by a consonantal modification (either the prefix t(t)- or gemination of a radical consonant), at least for the basic verbs. Table . contains two disyllabic verbs: the first illustrates the most frequent case, whereas the second illustrates the vocalic alternations which are characteristic of the “internal” verbs (Cohen ). Apart from the regular disyllabic verbs like ‘take’, verbs usually have at least two stems which fuse with each other. The basic system is completed by some grammaticalized verb forms, for example the verb yänhäyä (lit. ‘be busy’) is used to express the future.
Table . Tense, aspect, mood in Zenaga Aorist ‘take’
Negative Perfective
yärməš yərməš
Positive Perfective
Negative Imperfective
Positive Imperfective
yərmäš
yirəmməš
yirämmäš
‘fear’
yəgif
yägif̣
‘shiver’
yərgəgi
yärgägä
‘get caught’
y
yäT ärmäš
yəttəgəf yəttərgəgi y
yəT ärmäš
yəTyərməš
.. Adverbials Apart from two directional (or orientational) particles (-d(äh) and -n(äh)) which function as clitics, like the affixal pronouns, the adverbs form a small class of invariable words. Deictic elements are often part of location and time adverbs.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. S
..................................................................................................................................
.. Word (constituent) order The basic order is head+determiner and head+argument/adjunct. The direct object tends to precede the indirect object, both preceding prepositional objects. However, if both direct object and indirect object are encoded as affixal pronouns, the indirect object can possibly precede the direct object, if the latter is in the third person. Furthermore, the verbal “satellites” (i.e. pronominal and adverbial clitics) can be moved and affixed onto other hosts (e.g. onto negation).
.. Major sentence types ... Simple sentences In a predication whose core is a verb, the personal index alone can encode the subject. A verbal form like yäˀṃuššä ‘he committed perjury’ is thus a predication as complete as toˀfukt ̣ t-əddäm ‘the sun has set’. The same holds where there are other verbal arguments: compare (nəttä) y-əzzəg īrki əḏ^täšši ‘(it is he who) he has tied the calf with the cow’ and y-uwäḏ^täḏ^ḏäh aɣma=ˀn ‘[then] my brother married her’ (lit. ‘brought her here brother of mine’). The lexeme encoding the personal index is placed before or after the verb according to the way the information is structured. Neither this lexeme nor the direct object is marked for case. Nouns, adverbs, and adjectives can also be the core of the predication. In this case, the subject is obligatorily encoded as a noun or a pronoun (except maybe for adjectives inflected in the first and in the second person). The order is: S+predicate. The invariable copula äḏ is necessary to predicate non-verbal elements other than adjectives. Compare (niˀK) käygiš-ag ‘(me) I am blind’ and täyimt äḏ^tməgḏīẒ ‘the she-camel is a dairy one’.
... Complex sentences .... Sentences are usually coordinated by juxtaposition. Sometimes, however, imperative forms and imperfective forms which are not the first verbal form in a sequence of actions are replaced by aorist forms. naˀ ‘or’ expresses the alternative between two clauses. .... There exist several types of subordination. Among sentences which modify a verb, obligatory ones always immediately follow the verb. The presence and the choice (especially for äḏ vs. äyš) of the subordination marker, and the choice of the verbal form (aorist vs. (im)perfective) depend on the main verb and on the type of subordinate clause involved. One of the peculiarities of Zenaga is that the particle äḏ, which has a deictic origin, is used in subordinate clauses introduced by the verb ‘say’ (a construction used for reported speech), in consecutive clauses, and in the protasis of conditional clauses (Taine-Cheikh b).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
In subordinate clauses, verbal satellites are usually moved and affixed to the subordination marker. This movement also happens in relative clauses in the absence of a relative marker (āš, ār, or āf ) and preposition. In this case, a host with a deictic origin must be added. The relativization of the subject is marked on the verb (affixes -än or -(n)in of the “participle”), whereas the relativization of the direct object is not overtly marked (“gap strategy”) (see Galand : –, ; Taine-Cheikh ).
... Negation The invariable particle wär is used with all the verbal predicates. Its presence triggers the negative (im)perfective form of the verb. The verbal satellites (including the participle marker) are suffixed onto the negator: compare yəssän=ti ‘he knows it’ with wär=ti yəssən ‘he does not know it’. In the case of a non-verbal predicate (except for the adjectival predicate), a form which has a verbal origin (< yigä ‘become’) replaces the copula äḏ and tends to coalesce with the negator wär: ( . . . ) wäygi kəK ‘( . . . ) it is not you’. The adverbs äNyin and đäštyän used after the particle wär mean ‘never’ and ‘not anymore’, respectively (see also the auxiliaries wär^yiššiy ‘not . . . anymore’ and wär^yumrä ‘not . . . yet’). See Taine-Cheikh () for further details.
... Questions Interrogative markers are always sentence-first. A peculiarity of Berber languages is that interrogative markers like m=äḏ ‘who?’ (lit. ‘what-this?’), if they are not used alone, are always followed by a participial relative clause: m=äḏ^ti yuẓẓuˀr-an ‘who saw him?’.
... Other important types Commands are expressed by the imperative in the second person. To express prohibitions, the negative particle wär can be followed by an imperative form based on the positive imperfective. It is, nevertheless, more common to use an injunctive sentence. In such a sentence, the verb is in the aorist form and preceded by the particle äḏ. Oaths are characterized by the use of the positive perfective (without wär, even for a negative oath) combined with specific particles. For further discussion, see Taine-Cheikh (a).
.. Noun phrase structure If the modifier is a noun, it is usually introduced by ən: aˀgäy ən^tiˀräyn ‘(a/the) country of stones’. This relational marker alternates with ḏäg ‘among’ after numerals: tmađ̣ih ən~ḏäg^tiˀymən ‘twenty she-camels’. It is, however, absent with numerals less than eleven: ittäm^äyiˀm ‘eight camels’. In such cases, the head noun remains in the singular and the numeral agrees with it in gender (the units of the numerals from eleven to nineteen also agree in gender with the head noun): ittäm-äḏ^täyimt ‘eight she-camels’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
If the modifier is an adjective, it follows its head noun directly and is marked as a participle by the suffixes -än and -nin (like the verb of the relative clause whose antecedent is a subject): täyimt äḏäy-än ‘(a/the) black she-camel’.
.. Clause structure On the basis of the utterance context and the information structure of the sentence, the clause may contain one or more phrases. The predicative phrase can be more or less complex and can vary in terms of its argument structure with one and the same verb (for example if it is a labile verb): (ižiˀgär) y-ižäm ‘([the] rope) is plaited’, (nəttä) y-ižäm ižiˀgär ‘(it is he who) he has plaited a/the rope’. Topic and focus are marked both intonationally and morphosyntactically. The topicalized noun phrase is moved to the beginning of the sentence (as a general rule), while a pronominal clitic remains in the base position: ižiˀgär y-ižäm=ti ‘[the] rope, he has plaited it’. However, if the subject noun phrase (which is non-obligatory if there is a personal marker on the verb) is topicalized, an independent pronoun is found in the base position. A noun phrase can usually be focalized using a copula or a presentative construction: äyḏ=äḏ (wäygi) niˀK äḏ^y-ittur=än ‘it is (not) I who work’.
R Aikhenvald, A. (). ‘A structural and typological classification of Berber languages’, in S. Brauner and E. Wolff (eds.), Progressive Tradition in African and Oriental Studies. Berlin: Akademie Verlag, –. Attayoub, A. K. (). La tətsərret des Ayttawari Seslem: Identification socio-linguistique d’un parler berbère non-documenté chez les touaregs de l’Azawagh (Niger). Mémoire de maîtrise. Institut national des langues et civilisations orientales. Blažek, V. (). ‘Berber numerals’, Archiv Orientalni : –. Chaker, S. (). Linguistique berbère: Études de syntaxe et de diachronie. Leuven and Paris: Peeters. Cohen, D., and Taine-Cheikh, C. (). ‘À propos du zénaga: vocalisme et morphologie verbale en berbère’, Bulletin de la Société de Linguistique de Paris : –. Cohen, M. (). ‘Verbes déponents internes (ou verbes adhérents) en sémitique’, Mémoires de la Société de Linguistique de Paris : –. Galand, L. (). ‘Le Berbère’, in J. Perrot (ed.), Les langues dans le monde ancien et moderne. Vol. III: Les langues chamito-sémitiques. Paris: CNRS, –. Galand, L. (). ‘Propositions relatives, rhématisation et thématisation’, Études de linguistique berbère. Leuven and Paris: Peeters, –. Galand, L. (). ‘Le “participe” berbère’, Faits de Langues : –. Galand, L. (). ‘Typology of relative clauses: the case of Berber’, Language Typology and Universals : –. Kossmann, M. (). The Arabic Influence on the Northern Berber Languages. Leiden: Brill. Lamberti, M., and Tonelli, L. (eds.) (). Afroasiatica Tergestina. Padua: Unipress. Louali-Raynal, N. (). ‘La spirantisation en berbère’, in M. Lamberti and L. Tonelli (eds.), –. Lux, C. (). La tetserret, langue berbère du Niger: Description phonétique, phonologique et morphologique, dans une perspective comparative. Cologne: Rüdiger Köppe. Taine-Cheikh, C. (). ‘Le zénaga de Mauritanie à la lumière du berbère commun’, in M. Lamberti and L. Tonelli (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Taine-Cheikh, C. (). ‘Morphologie et morphogenèse du diminutif en zénaga (berbère de Mauritanie)’, in K. Naït-Zerrad (ed.), Articles de linguistique berbère: Mémorial Werner Vycichl. Paris: L’Harmattan, –. Taine-Cheikh, C. (a). ‘La corrélation de gémination consonantique en zénaga (berbère de Mauritanie)’, Comptes rendus du Groupe Linguistique d’Études Chamito-Sémitiques (–): –. Taine-Cheikh, C. (b). ‘L’adjectif et la conjugaison suffixale en berbère zénaga’, in J. Lentin and A. Lonnet (eds.), Mélanges David Cohen. Paris: Maisonneuve et Larose, –. Taine-Cheikh, C. (a). ‘Du rôle de la quantité vocalique en morphogénie: réflexions à partir de l’arabe et du berbère de Mauritanie’, Faits de Langues : –. Taine-Cheikh, C. (b). ‘Les numéraux en zénaga: contribution à la syntaxe des noms de nombre en berbère’, Studi Maghrebini (N.S.) : –. Taine-Cheikh, C. (). ‘Alternances vocaliques et affixations dans la morphologie nominale du berbère: le pluriel en zénaga’, in R. Vossen, H. Stroomer, and D. Ibriszimow (eds.), Études berbères III: Le nom, le pronom et autres articles. Cologne: Rüdiger Köppe, –. Taine-Cheikh, C. (). ‘Les propositions relatives en zénaga et le problème des relateurs en berbère’, in M. Moriggi (ed.), Atti del XII Incontro Italiano di Linguistici Camito-semitica (Afroasiatica). Soveria Mannelli: Rubbettino, –. Taine-Cheikh, C. (a). ‘Ordre, injonction, souhait et serment en zénaga (étude comparative)’, in H. Stroomer, M. Kossmann, D. Ibriszimow, and R. Vossen (eds.), Études berbères V: Essais sur des variations dialectales et autres articles. Cologne: Rüdiger Köppe, –. Taine-Cheikh, C. (b). ‘The role of the Berber deictic and TAM markers in dependent clauses in Zenaga’, in I. Bril (ed.), Clause Linking and Clause Hierarchy: Syntax and pragmatics. Amsterdam and Philadelphia, PA: John Benjamins, –. Taine-Cheikh, C. (). ‘L’énoncé négatif en berbère zénaga’, in A. Mettouchi (ed.), Parcours berbères: Mélanges offerts à Paulette Galand-Pernet et Lionel Galand pour leur ᵉ anniversaire. Cologne: Rüdiger Köppe, –. Taine-Cheikh, C. (). ‘Qualification and comparison in Berber: the verb-noun distinction and its fluctuations’, Language Typology and Universals : –. Vycichl, W. (). ‘Diminutiv und Augmentativ im Berberischen’, Zeitschrift der Deutschen Morgenländischen Gesellschaft (): –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. S belongs to the Highland East branch of the Cushitic language family, which is subsumed under the Afro-Asiatic language phylum.1 It is spoken in the Sidaama zone of south-central Ethiopia, whose capital, Hawassa, is located km south of Addis Ababa. According to the Ethiopian Census, the population of the Sidaama people comprised ,, people. They are the fifth-largest ethnic group in Ethiopia, and Sidaama is the language spoken by the fifth-largest number of speakers in Ethiopia. The language of instruction at school in the Sidaama zone was Amharic until , when the new Ethiopian government permitted the use of local languages for primary education; since then, the language of instruction at public primary schools has been Sidaama, and more and more Sidaama speakers have come to use the Latin alphabet. Sidaama has two major dialect groups, Highland and Lowland, which differ lexically and phonologically. The Highland dialects are considered more authentic and conservative than the Lowland dialects, which have been influenced by other languages, especially Amharic. There are not many previous in-depth studies on this language, although there is a grammatical sketch of Highland East Cushitic languages by Hudson (), grammars of Sidaama by Moreno (), Teferra (), and Kawachi (), a Sidaama–English dictionary by Gasparini (), a collection of words translated between Highland East Cushitic languages and English by Hudson (), and papers by Kawachi (e.g. , , a).
. P
..................................................................................................................................
.. Segmentals Sidaama has five vowel phonemes: /i/, /e/, /a/, /o/, /u/. Sidaama is also often mistakenly called “Sidamo”, which was used around to refer to the province that included the Sidaama zone and other neighboring zones. 1
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Sidaama consonant phonemes Place of articulation
Manner of articulation Stop
Bilabial Labio-dental Alveolar Post-alve-olar Pala-tal Velar Glottal
Plosive Ejective
b p’
td
kɡ
t’
k’
Implosive
’
ɗ
Affricate
čǰ Ejective
č’
Fricative Nasal
f
s (z)
m
n
Flap
r
Lateral Approximant
l
Approximant
w
š
h ñ
y
The consonant phonemes of Sidaama are shown in Table .. The maximum number of consonants that can occur successively in Sidaama is two. Consonant clusters only occur across syllable boundaries. The large majority of consonant clusters are sequences of (i) a sonorant (/m/, /n/, /r/, /l/) plus an obstruent, or (ii) a glottal stop plus a sonorant (/m/, /n/, /r/, /l/, /y/). All the Sidaama consonants have geminate forms, which occur across syllables, and some of them (/p’/, /b/, /ɗ/) have to be geminated word-medially. Gemination can also occur as a grammatical process (e.g. ɡimboola ‘bamboo basket (number unmarked)’ ! ɡimboolla (PL)).
.. Morphophonemic rules Sidaama has a variety of morphophonemic rules, including those on epenthesis, metathesis, assimilation, glottalization, and gemination. These are discussed in Abebe Gebre-Tsadik () and Kawachi (). They are, for the most part, motivated by the phonotactics of this language.
.. Suprasegmentals ... Syllables The syllable structures attested in Sidaama are V (e.g. á.ma ‘mother’), CV (e.g. ro.dó.o ‘sibling’), VC (e.g. án.na ‘father’), and CVC (e.g. dán.ča ‘good’). V and CV can occupy any position of a word, but VC and CVC can occur only word-initially or word-medially, not word-finally, because all Sidaama words end in a vowel. Most Sidaama words consist of two or more syllables, and there are only a small number of monosyllabic words (e.g. t’a ‘now’, te ‘this (F)’, ko ‘this (M)’). Any of the consonants can appear in either the onset or coda of a syllable, except /’/, which word-medially can occur only in the coda, never in the onset.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
In this language, two different consecutive vowels (e.g. [ai]) are interpreted as a sequence of two different vowel phonemes (/a/+/i/) constituting two syllables, rather than a diphthong. Similarly, a long vowel (e.g. [a:]) is also interpreted as a sequence of two identical vowel phonemes (/a/+/a/) constituting two syllables.
... Pitch accent Sidaama is a pitch accent language. Prominence is indicated by high pitch, which is not associated with the duration of the vowel. High pitch in the citation form of a noun or adjective is normally located on the penultimate vowel (i.e. the penultimate syllable), e.g. be.ét.to ‘child’, se.é.da ‘long’, ha.í.t’e ‘barley’.
... Intonation Intonation differentiates sentence types in Sidaama. In declarative sentences ending with a verb predicate, high pitch occurs on the penultimate vowel segment of the main verb in the imperfect, distant perfect, or imperfect, if the subject is in the second person plural or the third person (e.g. áni it-oó-mm-a (. eat-..--) ‘I (F) ate.’); it occurs on the ultimate vowel segment of the main verb in any of these aspectual categories whose subject is in any other person-number combination (e.g. ísi it-ø-inó (.. eat-...) ‘He ate.’). In sentences ending in nominal or adjectival predicates, pitch is normally high on the penultimate vowel segment of the noun or adjective, and is lowered toward the end of the sentence (e.g. íse dánča=te (.. good=..) ‘She is good.’). In any type of negative predicate, for which the negative proclitic di= is used, the vowel of this clitic has an extra-high pitch. In (both positive and negative) polar questions involving a verb predicate, a pitch higher than in the declarative counterpart occurs in the vowel segment of the verb where a high pitch would occur if the sentence were declarative (e.g. extra-high pitch on the final o in ísi it-ø-inó? (.. eat-.-..) ‘Did he eat?’). In nominal or adjectival predicates, the location of an extra-high pitch is the ultimate vowel segment of the noun or adjective (e.g. extra-high pitch on the second á of danča in íse dančá=te (.. good=..) ‘Is she good?’). (When the interrogative marker =ni attaches to a verbal or non-verbal predicate, the vowel of this marker carries an extrahigh pitch.)
. P
..................................................................................................................................
.. Open classes ... Nouns In Sidaama, nouns usually come at the end of noun phrases, and may be marked with affixes for case, number, and gender. Certain abstract nouns are usually accompanied by a modifier to form an adjunct: specifically, about twenty locational nouns (e.g. ɡíddo ‘inside’,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
aána ‘top’, álba ‘front’, mereéro ‘middle’, múle ‘nearness’), the comitative noun lédo, and ɡeéšša ‘degree, extent’.
... Verbs Finite verbs contain either (i) one of the aspectual suffixes in addition to the subject person–number(–gender) suffix(es) or (ii) one of the mood suffixes. Both the aspectual and mood suffixes have different forms depending on the person–number(–gender) of the subject. Non-finite verbs may contain either the connective converb suffix or the infinitive suffix, as well as the subject person–number(–gender) suffix (e.g. it-t-e (eat-.-)). Sidaama verb forms showing a transitivity contrast usually follow one of three patterns. First, intransitive verbs may be turned into transitive verbs by adding the causative suffix (e.g. e’- ‘enter’, ee-s- ‘move something in’). Second, transitive verbs with the passive suffix may act as intransitive verbs (e.g. hiikk’-am- ‘get broken’, hiikk’- ‘break something’). Third, pairs of idiomatic expressions where y- ‘say’ and ass- ‘do’ follow an ideophonic or onomatopoeic element show a transitivity contrast (e.g. dolli y- ‘lie down’, dolli ass- ‘lay down someone/something’). Sidaama has only a limited number of intransitive–transitive pairs of morphologically unrelated lexical verbs that seem to show a semantic contrast (e.g. re- ‘die’, š- ‘kill someone’), and has only one labile verb, to my knowledge (hanaf- ‘begin (intransitive/transitive)’). A small number of transitive verbs (mostly of feeling) can take an impersonal third person singular masculine subject, which usually emerges only as the subject suffix on the verb (e.g. ɗiw-ø-ino-’e (cause.sickness-.-..-) ‘I became/am sick.’; lit. ‘. caused sickness to me.’). Sidaama verbs, which are all dynamic rather than stative, can be classified into action or state-change verbs depending on whether they can be used in the continuous aspect, and the meaning of each of their three aspectual forms, as shown in Table ..
Table . Aspectual differences between dynamic verbs in Sidaama Aspect
Verb type
Action verbs (e.g. dod- ‘run’)
State-change verbs (e.g. uurr- ‘stand up’)
Present progressive ongoing action (e.g.dod-d-á-nni no) gradual state-change (e.g. uurr-i-t-ánni no) (stand-ep-3sg.f-inf-abl:ins(run-3sg.f-inf-abl:ins come.to. come.to.exist.dist:prf.3) ‘She is in the exist.dist:prf.3) ‘She is running.’ process of standing up.’ Continuous
*
continuous state up to the present (e.g. uurr-i-t-e no) (stand-ep-3sg.f-cvb come.to.exist.dist:prf.3) ‘She has been standing.’
Recent or distant perfect
completed action (e.g. dod-d-inó) (run-3sg.f-dist:prf.3) ‘She ran.’
completed state-change/current state as the result of a state-change (e.g. uurr-i-t-inó) (stand-ep-3sg.f-dist: prf.3) ‘She stood up/is standing.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
... Adjectives Sidaama has only a limited number of non-derived, true adjectives, though it has many adjectives derived from verbs or nouns. Sidaama adjectives share some morphosyntactic properties with nouns and some with verbs. Like nouns, many adjectives can form constituents that have grammatical relations in sentences (i.e. argument noun phrases), and can be accompanied by modifiers (Kawachi b). Also like nouns, adjectives are marked for case, and agree in case, number, and gender with the noun that they modify. Many adjectives can be modified by degree adverbials (e.g. lowó ɡeešša (large degree) ‘very much’) and can be used in comparative constructions. These properties are also found in a limited number of verbs.
... Adverbs The class of lexical adverbs in Sidaama, many of which are time adverbs, is small in size compared to the other open classes. Most of them are invariant in form in any syntactic environment, and do not take any inflectional suffixes.
.. Closed classes As closed-class categories, Sidaama has pronouns (personal, reflexive, and reciprocal pronouns), demonstratives, and clitics. Most of the demonstratives, which cross-cut different parts of speech and take different forms depending on their grammatical function and gender (preserving the proto-Afro-Asiatic demonstrative t (F) vs. k (M) contrast), make a four-way distinction: ‘close to the speaker’, ‘close to the listener, or known to the speaker and listener from a previous discourse’, ‘far from both the speaker and listener’, and ‘very far from both the speaker and listener’. Sidaama has one proclitic and a number of enclitics. The negative proclitic di= usually attaches to the beginning of a clause-final constituent, or sometimes to the beginning of a constituent in the preverbal focus position. Out of a number of enclitics, two are used to form noun phrases. One of them, =ta (F) / =ha (M), which apparently originates from the proto-Afro-Asiatic t (F) / k (M), attaches (i) to a genitive noun phrase or a relative clause to form an argument noun phrase, (ii) to an adjective, a noun phrase, a genitive noun phrase, or a relative clause to form a predicate, or (iii) to a clause, to form a clausal complement (Kawachi b). (The plural form =re, which can be used for animate entities, is limited to (i).) When this enclitic attaches to a predicate noun, its form is determined by whether the noun is common or proper, and if it is common, whether it is accompanied by any modifier, the possessive pronominal suffix, or both (“Modified”, henceforth) (=ti), or neither (“Unmodified”, henceforth) (=te (F) / =ho (M)) (Kawachi and Tekleselassie ). The other noun phrase enclitic, =wa, attaches to a genitive noun phrase, an adjective, or a relative clause to form a noun phrase that refers to a location, goal, or source (e.g. hakkó kinč-í=wa (that.. rock-..=place) ‘at/to that rock’). When =wa is bound to a genitive noun phrase that refers to an object, the object is treated as a location.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Sidaama also has another complementizer enclitic, and conjunctive enclitics for causal, concessive, and temporal clauses, as well as a few other enclitics and enclitic complexes.
. M
..................................................................................................................................
.. Affixes and reduplication Sidaama is an agglutinative language in which almost all derivational morphology and all inflectional morphology involves affixation. The reduplication of a verb root to express repeated or intensified actions is not uncommon, but not productive.
.. Morphological case-marking system Sidaama is a nominative–accusative language. It uses a suprafix (high pitch on the final vowel segment) for the accusative–oblique case, suffixes for the nominative, dative–locative, allative, and ablative–instrumental cases, and both a suprafix and a suffix for the genitive case. The suprafixes and the suffixes concern the ordinary, syntactic notion of “modification” and the less usual notion of “Modification” (section ..), respectively. The Modification distinction, together with gender, determines the choice of the allomorphs of some of the case suffixes. For example, the nominative suffix on Unmodified masculine common nouns is -u, as in (), whereas that on Modified ones is -i, as in () and (). ()
()
beétt-u da-ø-inó. child-: come-.-:. ‘A/the boy came.’ a. kú’u that.: b. búš-u bad-: c. isé .: d. íse .:
saɡalé food.:
u-i-t-ino give--.-:.
beett-i da-ø-inó. child-:: come-.-:. (a) ‘That boy / (b) ‘A/the bad boy / (c) ‘Her son / (d) ‘A/the boy who she gave food to came.’2 2 Nevertheless, the use of the allomorph -u of the nominative suffix for Unmodified masculine nouns is possible even when the masculine noun is modified by an adnominal demonstrative (e.g. in (2a), kú’u beett-u). However, this applies to neither the genitive nor the dative–locative case suffix.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
beétt-i-se da-ø-inó. child-::-.: come-.-:. ‘Her son came.’
()
Table . Nominative, genitive, and dative-locative case suffixes in Sidaama Case and gender Modification Unmodified
Nominative
Genitive
F
M
F
M
F
M
-u
-te
-u
-te
-ho
-i
-ø
-i
-ø
Modified
Dative locative
-ra
The allomorphs of the nominative, genitive, and dative–locative case suffixes are listed in Table .. Note that -u and -i replace the final vowel of a stem noun, whereas -te, -ho, and -ra are added to a stem noun; -te and -ho are each added to an accusative–oblique noun stem (e.g. beettó-ho (child..-..) ‘to the boy’), while -ra, one of the allomorphs of the dative–locative suffix as well as the allative suffix, and -nni, the ablative–instrumental suffix, which is not listed in the table, are each added to a genitive noun stem (e.g. beett-í-se-ra (child-..-..-..) ‘to her son’). The syntactic modification distinction plays a role in suprafixation. Nominative forms of unmodified nouns have high pitch on their penultimate vowel segments, whereas their accusative–oblique and genitive forms have high pitch on their final vowel segments. Nouns modified by a modifier usually carry no high pitch. When modified by an adnominal demonstrative or an adjective, the modifier follows the pitch accent pattern of the case of the noun phrase (e.g. (a), (b)).
.. Suffix orders Any Sidaama word follows the order: root–derivational suffix(es)–inflectional suffix(es), when it contains both types of suffixes (Greenberg : ). The ordering relationships of the noun suffixes and those of the adjective suffixes are shown in () and (), respectively. ()
Noun suffixes ROOT
Nominalizing Case (Nominative/ Abstracting Genitive) Number Gender
()
Adjective suffixes ROOT
Adjectivizing Number Gender
Case
Possessive Case (Dative–locative/ Allative/ Ablative–instrumental)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
One of the derivational processes (nominalization) and one or both of the inflectional categories (number and gender) listed for the first noun suffix slot in () are often expressed by a portmanteau morpheme, but more than one suffix cannot be used simultaneously for more than one of these categories. As shown in (), after the adjective root, the adjectivizing suffix precedes the number suffix, which in turn precedes the case suffix. The ordering relation between the gender suffix on the one hand and the adjectivizing and number suffixes on the other is indeterminate. Some of the adjectivizing suffixes also show a gender contrast. Possible combinations of derivational verb suffixes are shown in (). ()
a. Passive–Double-causative (e.g. k’alt’-an-siis- (choke--.-) ‘cause someone to choke/hang himself/herself ’) b. Reciprocal–Double-causative (e.g. t’on-an-šiiš- (insult--.-) ‘cause some people to insult each other’) c. Causative–Passive (e.g. huf-i-s-am- (boil----) ‘become boiled (by someone)’) d. Causative–Middle (e.g. ra’-i-s-i-ɗ- (become.cooked-----) ‘cook something for oneself ’) e. Double-causative–Middle (e.g. ra’-i-siis-i-ɗ- (become.cooked--.--) ‘cause someone to cook for himself/herself ’) f. Middle–Double-causative (e.g. buu-ɗɗ-i-siis- (smear.butter---.-) ‘cause someone to smear butter on himself/herself ’)
While the object person–number(–gender) suffix comes after any other inflectional verb suffix, the order of the subject person–number(–gender) suffix(es) and the aspectual or infinitive suffix relative to the verb stem varies depending on the person–number(–gender) combination.
. S
..................................................................................................................................
.. Grammatical relations In Sidaama, grammatical relations are coded by case marking and verb suffixes. Subjects, direct objects, and indirect objects are in the nominative, accusative, and dative-locative cases, respectively. The pronominal subject suffix on the verb provides information on the person–number(–gender) of the subject. The pronominal object suffix on the verb usually but not necessarily indicates the primary object; strictly speaking, it refers to the human event participant whom the speaker perceives as the most prominently affected by the event, either positively or negatively. Sidaama has an imperative construction and subject-control constructions, which characteristically involve the notion of subject. Two pairs of multi-verb/clause constructions, the connective converb construction and the temporal overlap construction, take different forms depending on whether the subjects of the main verb and the non-main verb(s) are the same or
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
different. Constructions to which direct objects are relevant in Sidaama are a passive construction and a direct object-control construction.
.. Word order The word order is predominantly S–O–V. The focus position is immediately before the verb. O–S–V and other orders are possible only when the preverbal constituent is in focus. V–S and V–S–O word orders emphasize the action or state-change that the verb expresses, i.e. they express verb focus. In Sidaama, an adnominal constituent usually precedes the noun that it modifies. When more than one type of modifier occurs in a noun phrase to modify the noun, the preferred order is adnominal demonstrative–numeral–adjective–genitive noun phrase– noun. Any modifier dislocated from its normal position has to form a noun phrase in apposition with the other noun phrase(s) having the same referent. A dislocated genitive noun phrase has to be accompanied by the noun phrase clitic =ta (F) or =ha (M) (see section ..). A relative clause also usually precedes a noun that it modifies. It modifies a noun with no other modifier without using the noun phrase clitic. However, when it modifies a noun with a modifier or modifiers, it often forms an appositive noun phrase with the noun phrase clitic at its end. In Sidaama, the subordinate clause usually precedes the main clause, and the converb normally precedes the main verb. Sidaama also shows the ordering relationships in (): ()
a. b. c. d.
Standard of comparison–comparative marker–adjective/verb Manner adverb–verb Complement clause–complementizer Interrogative word–verb
.. External possessor constructions Sidaama has two types of external possessor constructions (EPC), accusative– oblique possessum EPCs and dative–locative possessor EPCs (Kawachi , ), where the possessor and the possessum noun phrases do not form a constituent. In the accusative– oblique EPCs, the possessum noun phrase, which could be translated literally as ‘with respect to the possessum’, is in the accusative–oblique case, while in the dative–locative EPCs, the possessor noun phrase is in the dative–locative case. Possessive relations in either type of construction can also be expressed by internal possessor constructions (IPCs), in other words, within noun phrases. The predicate of either type of EPC may be an adjective, an intransitive verb, or a transitive verb. In the accusative–oblique EPCs, the possessor noun phrase assumes the syntactic role that the possessum noun phrase in the IPC counterpart would fill, whereas in the dative–locative EPCs, the possessum noun phrase has the same syntactic status as in the IPC counterpart. The possessor in an accusative– oblique EPC sentence and the possessum in a dative–locative EPC sentence each serve as
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
subject when the predicate is an adjective or intransitive verb, and as object when the predicate is a transitive verb. Transitive examples of the accusative–oblique and dative–locative EPCs and their IPC counterpart are shown in (a), (b), and (c), respectively. ()
a. t’arap’eesá table.: íse .:
lekká leg.:
b. t’arap’eesá-ho lekká table.:- leg.: :: c. t’arap’ees-ú table-.
hiikk’-i-t-inó. break--.-:.
lekka leg(:)
‘She broke the leg of the table.’ (lit. ‘she broke (a) the table with respect to the leg / (b) the leg to the table / (c) the table’s leg’) The types of possessor–possessum combinations possible in the dative–locative EPCs are (i) animate entity–body part, (ii) animate entity–kinsman, (iii) animate entity–artifact/ belongings, and (iv) inanimate entity–part. Those possible in the accusative–oblique EPCs are only (i) and (iv), although accusative–oblique EPCs also allow the possessum to be a property or abstract entity that the possessor has.
R Abebe Gebre-Tsadik, M. (). ‘An overview of the morphological structure of Sidamo verbs’, in M. Abebe Gebre-Tsadik, R. J. Sim, C. Wedekind, and K. Wedekind, The Verb Morphophonemics of Five Highland East Cushitic Languages, including Burji: Afrikanistische Arbeitspapiere . Cologne: Institut für Afrikanistik, University of Cologne, –. Gasparini, A. (). Sidamo–English Dictionary. Bologna: Editrice Missionaria Italiana. Greenberg, J. H. (). ‘Some universals of grammar with particular reference to the order of meaningful elements’, in J. H. Greenberg (ed.), Universals of Language. Cambridge, MA: MIT Press, –. Hudson, G. (). ‘Highland East Cushitic’, in L. M. Bender (ed.), The Non-Semitic Languages of Ethiopia. East Lansing, MI: African Studies Center, Michigan State University, –. Hudson, G. (). Highland East Cushitic Dictionary. Hamburg: Helmut Buske. Kawachi, K. (a). A Grammar of Sidaama (Sidamo), a Cushitic language of Ethiopia. Ph.D. dissertation. Buffalo: State University of New York. Kawachi, K. (b). ‘Feelings in Sidaama’, LACUS Forum : –. Kawachi, K. (a). ‘Can Ethiopian languages be considered languages in the African linguistic area? The case of Highland East Cushitic, particularly Sidaama and Kambaata’, in O. Hieda, C. König, and H. Nakagawa (eds.), Geographical Typology and Linguistic Areas—with special reference to Africa. Amsterdam and Philadelphia, PA: John Benjamins, –. Kawachi, K. (b). ‘Noun phrases without nouns in Sidaama (Sidamo)’, LACUS Forum : –. Kawachi, K. (). ‘External possessor constructions in Sidaama’, CLS : –. Kawachi, K. (). ‘Dative external possessor constructions in Sidaama’, Berkeley Linguistic Society : –. Kawachi, K., and Tekleselassie, A. A. (). ‘Modification within a noun phrase in Sidaama (Sidamo)’ Berkeley Linguistic Society : –. Moreno, M. M. (). Manuale di Sidamo. Milan: Mondadori. Teferra, A. (). A Grammar of Sidaama. Ph.D. dissertation. Hebrew University of Jerusalem.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. K is the name of a little-known, small ethnolinguistic group in northwestern Ethiopia. Their neighboring ethnolinguistic groups call them Kunfäl, a name considered abusive by the Kolisi, who basically call themselves and their language kol-isi Agäw ‘lowlander Agäw’ (kol ‘hot/lowland’ + isi (adjectivizer)) and the Awi (speakers of Awŋi) däg-isi Agäw ‘highlander Agäw’ (däg ‘cold/highland’ + isi (adjectivizer)). The Awi also consider the Kolisi bikalka ‘hybrid’ because of their complexion, which brings together Awi and Gumuz (Nilo-Saharan) physical features. Cowley et al. () call them K’oll-äɲɲ-oʧʧ ‘lowlanders’ and write that their skin color is darker than the Awi but lighter than the Gumuz. The Kolisi inhabit the areas of Jawi and Quara districts of the Amhara regional state and Dangur of the Benishangul-Gumuz regional state in a few scattered villages. The Kolisi are all Orthodox Christians and build their houses from wood and grass as well as sheets of corrugated iron. For subsistence, they rely mainly on crop and honey production. Kolisi is classified under the Central Cushitic language family, traditionally known as Agäw (Hetzron ). The only works on the Kolisi people and their language are the sixpage article by Cowley et al. () and a Ph.D. dissertation by Allene (). It contains brief sociolinguistic information and a wordlist containing a hundred words and indicates that the language is very similar to Awŋi. Kolisi is spoken by almost all members of the small ethnic group, including monolinguals. Their language is used mainly in the home domain. As a result of the expansion of towns, mechanized farms, and resettlement areas, the swamping effect of Amharic seems to be inevitable. Kolisi children are sent to Awŋimedium schools and this is challenging the ethnolinguistic identity of the Kolisi people.
. P
..................................................................................................................................
.. Vowels Kolisi has a seven-vowel system (Table .). The existence of the central vowels /ɨ/ and /ä/ is one of the features distinguishing Agäw languages from the rest of Cushitic. It is believed
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Kolisi vowels1 i e
(ɨ) (ä) a
u (o)
1
The three vowels in parentheses need further investigation to determine their status.
Table . Kolisi consonants p b f
m w
t d s z
ʃ ʒ ts ʤ
n l, r
k g x ɣ ʧ
q h
ŋ
y
that these vowels emerged from contact with Ethio-Semitic languages. Appleyard () included the two central vowels in the Proto-Agäw vowel chart, although their status as reconstructed phonemes is doubtful. /ɨ/ occurs in all but predictable environments. Examples are: ɨŋi ‘ant’, fɨr ‘well’, and ɨnt(ɨ) ‘you ()’. /ä/ is recorded in few instances: läb- ‘fall’, worämi ‘spear’, bägi ‘sheep’, and ʃeräng ‘flute’. It occurs more frequently in loanwords from Amharic (see discussions on these marginal vowels in Hetzron ; Appleyard , ; Joswig ). The status of /o/, which occurs rarely and in loanwords from Amharic, is also problematic. As in other Agäw languages, vowel length is not phonemic in Kolisi either. In the following examples in (), the lengthening of the penultimate vowel shows emphasis and plays the role of the degree adverbial ‘very’ in English. ()
sɨnkuti ‘good’ sɨnkuuti ‘very good’ liggismi ‘long’ liggiismi ‘very long’ ɨzzani ‘wide’ ɨzzaani ‘very wide’
.. Consonants There are twenty-five consonant segments in Kolisi (Table .). As a typical Agäw language, /ŋ/ is a frequently occurring phoneme in Kolisi. /p/, /q/, and /ts/ are part of the consonant inventory but rare in other Cushitic languages. The velar fricatives are also among the archaic phonemes in Agäw languages. Ejectives are recorded only from Amharic loanwords and are replaced by the non-ejective counterparts by monolingual speakers as in k’et’o > keto ‘green pepper’, ʧ ’amma > ʧami ‘shoe’, bak’ela > bakela ‘bean’, aʧ ’ʧ ’ed- > atsed- ‘harvest’, märrät’- >meret- ‘choose’, and ʧ ’äw > ʧiwi ‘salt’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The nasal series lacks /ɲ/, which is not among the Proto-Agaw consonant phoneme, occurs only in one instance (see section ...). Words with the palatal nasal are hence predictably loans from neighboring languages, particularly Amharic. Labialized velars kw, gw, xw, ɣw, and ŋw are archiphonemes which perhaps need to be specified in further studies.
.. Syllable structure Kolisi has both open and closed syllables. The noun kwa ‘cow’, nu ‘we’, ag ‘uncle’, and the verbs y- ‘give’, xw- ‘eat’, and s- ‘weep’ are among the shortest verb stems. Other monosyllabic words include saŋ ‘tongue’ and ʤer ‘child’. Since Agäw languages do not have phonemic vowel length, all the syllables show non-heavy syllables with minimal words that consist of either a VC or a CV syllable. ()
()
Disyllabic words VCV asi VCCV ɨnni VCCVC ɨnkɨrCVCCVC berber
‘fish’ CVCV biri ‘ox’ ‘elephant’ VCCV amli ‘cabbage’ ‘laugh’ CVCCV bɨqli ‘mule’ ‘pepper’
Trisyllabic words CVCVCV sɨlali ‘root’ VCCVCV antsɨni ‘father-in-law’ CVCVCCV tsaɣinsi ‘flee’ VCVCVC ɨɣaɣar- ‘curse’
It appears that words are bi- or trisyllabic in the majority of cases. There are, however, few instances of quadrisyllabic words as in musisini ‘pillar’ (CVCVCVCV) and ɨɣaɣari ‘curse’ (VCVCVCV).
.. Tone The status of tone in Agäw languages has not been well studied. Hetzron () writes that Awŋi and Bilin are tonal languages. In Kolisi, there are few instances where the contrast between high and low tone corresponds to meaning distinctions in the lexicon. ()
yìwi yíwi mùri múri
‘wrist’ ‘mine’ ‘village’ ‘snake’
kùwi kúwi áqi àqi
‘who killed’ ‘your ()’ ‘person’ ‘persons’
The homophones such as ɨnni ‘this’ vs. ɨnni ‘elephant’, ani ‘I’ vs. ani ‘that ()’, and awi ‘sun’ vs. awi ‘Agäw person’ should have once been distinguished through tonal differences. Whether tone is grammatical or lexical or both and whether or not it is distinct from stress or pitch accent still need further investigation. It seems that Agäw languages might have dropped tone or tonal accent due to frequent contact with Semitic languages.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. M
..................................................................................................................................
.. Pronouns ... Personal pronouns There are three singular and two plural subject personal pronouns as well as two polite forms for the second and third person singular (Table .). The singular and plural forms share the same bases: n for the first person, nt for the second person, and ŋ for the third person.The corresponding plural marker in the first and second persons is -u and -a in the third person.The occurrence of ɲ only in the third person in Kolisi is merely the effect of the vowel i (cf. ŋi in Awŋi, ni in Bilin and Kemantney, and ŋäŋ in Xamt’aŋa for the ). No gender distinction is shown in both the second and third persons. The second person polite form and the share the same form and the third person polite form is identical to the second and third person polite forms of Kemantney (Zelealem ). Object pronouns (Table .) are not derived from their subject counterparts. They are bound pronouns similar to the possessive pronouns and identified by the accusative marker -wa attached to them invariably in all persons. Note that, in the , w assimilates to the last segment of the base, resulting in the gemination of the latter.
Table . Personal pronouns in Kolisi Singular
Plural
ani
nu
ɨnti
ɨntu
(honorific)
ɨntu
ɲi nay
(honorific)
ŋa
Table . Object pronouns in Kolisi Singular
Plural
ɨy-ya
nu-wa
ku-wa
ɨntu-wa
(honorific)
ɨntu-wa
ɲi-wa
(honorific)
nay-wa
ŋa-wa
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Possessive pronouns in Kolisi Singular
Plural
yi-wi
nu-wi
ku-wi
ɨntu-wi
(honorific)
ɨntu-wi
ɲi-wi
(honorific)
nay-wi
ŋa-wi
The bases of the bound possessive pronouns are derived from the object pronouns. The difference between the two lies in the suffixes: -wa (accusative suffix) and -wi (genitive suffix). The possessive pronouns in Table . hence have the literal meaning ‘of me/mine’, ‘of you/your’, ‘of him/his’, and so on. Before possessed nouns, possessive pronouns appear independently by dropping i of the genitive suffix -wi, as shown in () below. Literally, the following genitive NPs mean ‘house of me’, ‘house of you’, ‘house of him’, etc. ()
yi-w gɨn ‘my house’ nu-w gɨn ‘our house’ ku-w gɨn ‘your () house’ ɨntu-w gɨn ‘your () house’ ɲi-w gɨn ‘his/her house’ ŋa-w gɨn ‘their house’
In unalienable possessive constructions like yi-abba > yibba ‘my father’, yi-gana ‘my mother’, yi-zan ‘my brother’, the base forms function as possessive prefixes without genitive endings.
... Demonstrative pronouns The proximal demonstrative pronouns are ɨnn-i and ɨnn-a for the singular and ɨnni-ŋa for the plural. They precede the noun as in ɨnni-gɨseŋ-i ‘this dog’, ɨnna-gɨseŋ-a ‘this bitch’, and ɨnniŋa-gɨseŋ-ka ‘these dogs/bitches.’ The distal demonstrative pronouns are an-i and an-a for the singular and ani-ŋa for the plural, as in an-i aqi ‘that man’, an-a xuna ‘that woman’, and aniŋa-aqi ‘those persons’ for the plural. The plural demonstratives are derived from the masculine singular form by attaching the plural suffix -ŋa (the same form for the subject pronoun). Kolisi shows the i/a masculine and feminine gender distinction in both proximal and distal singular demonstratives. The proximal deictic is ɨn-da ‘here’ and the distal an-da ‘there’ combine the proximal deictic ɨn- and the distal deictic an- with the locative marker -da.
... Reflexive pronouns In reflexive pronouns (Table .), the pronoun appears preceding the reflexive, which contains the noun bal ‘possessor/owner/husband’ and the locative marker -da.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Reflexive pronouns in Kolisi an bal-di ɨntɨ bal-da ɨntɨ bal-di ɲi bal-di ɲi bal-da nu bal-di ɨntu bal-di ŋa bal-di
‘I myself ’ ‘you () yourself ’ ‘you () yourself ’ ‘he himself ’ ‘she herself ’ ‘we ourselves’ ‘you () yourselves’ ‘they themselves’
The reflexive pronouns show gender distinction in the second and third singular: a for feminine and i for masculine. The occurrence of bal ‘possessor/owner/husband’ makes Kolisi different from the other Agäw languages, which use the word for ‘head’ in reflexive pronouns (ŋara in Awŋi and aɣoy in Kemantney).
.. Nouns Definiteness is not marked in Kolisi and hence no formal difference is exhibited between nouns in the indefinite and definite. ()
aqi kaz-ɨɣwa. man go-:. ‘A/the man went.’
()
gɨsaŋ koy-ɨɣwa. dog bark-:. ‘A/the dog barked.’
Kolisi nouns show singular and plural distinctions. Singular is unmarked and is identical with the citation form. Plural is marked by way of different number suffixes which sometimes makes the prediction of nouns in the plural difficult. The most productive plural formation is by suffixing -ka. ()
ʤeri gɨni sankɨti gebeli gɨsaŋ
> > > > >
ʤer-ka gɨn-ka sankɨt-ka gebel-ka gɨsaŋ-ka
‘child/children’ ‘house/houses’ ‘piece of clothing/clothes’ ‘market/markets’ ‘dog/dogs’
The gender marker i is elided before -ka. Other plural formations include internal vocalic change, as in biri ‘ox’ vs. bera ‘oxen’. Plural formation through complete reduplication is recorded in two instances: xuna > xuna-xuna ‘woman/women’ and siɣa > siɣa-siɣa
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
‘girl/girls’. With proper nouns, the morpheme -ʤi appears as plural marker as in kasa-ʤi ‘Kasa and others’, alämitu-ʤi ‘Alemitu and others’. The morpheme -ʤi appears as plural marker in Awŋi subject pronouns. Kolisi has a two-way gender system in nouns: masculine and feminine. It is only in the singular that gender distinction is marked. As mentioned in sections ... and ... above, feminine is marked by a and masculine by i as in bäg-a sheep () vs. bäg-i ‘sheep’ (), bald-a ‘owner’ () vs. bald-i ‘owner’ (), fɨlay-a ‘goat’ () vs. fɨlay-i ‘goat’ (), antsɨn-a ‘mother-in-law’ vs. antsɨn-i ‘father-in-law’, etc. Most nouns end in i and are thus masculine. Some naturally genderless nouns such as dabali ‘skin’, bɨni ‘river’, ɨri ‘rain’, and gumiti ‘stem’ have a grammatically masculine representation. Nouns that end in consonants, such as ʤer ‘child’, ɨlaf ‘face’, gɨsaŋ ‘dog’, tend to be masculine with augmentative connotation. Nouns that end in a, such as qalqala ‘uvula’, arfa ‘moon’, and the digits (one to nine) have a grammatically feminine and diminutive denotation. As shown in section ..., demonstratives make a gender distinction. Gender distinction is also shown in verbs with and subjects by pronominal suffixes through a zero morpheme and -u for masculine and -t and -i/-y for feminine. In Kolisi, nominative is unmarked. Subject nouns are identical to their citation forms. The marked morphological cases are accusative and genitive. Accusative is marked in definite nouns by the suffix -wa after vowels and -u after consonants. ()
ani ɲe-wa kant-u-ɣwa. I .- see-- ‘I saw him.’
() ɲi iy-ya kant-u-ɣwa. he - see-.- ‘He saw me.’ ()
ʤer-i mäsaf-u kats-ɨɣwa. child- book- take-:. ‘The child took the book.’
The other morphological case is the genitive. In the genitive NPs in () where the source of the head noun is expressed, the suffix -i is the genitive marker. ()
tɨt-i sankɨta antsin-i warama ʧik-i gɨna sɨk-i sɨlaɣa gomm-i ʧampi ŋarg-i mɨʃa
‘cloth from cotton’ ‘spear from iron’ ‘house from mud’ ‘beer from sorghum’ ‘shoe from rubber’ ‘mead from honey’
The genitive NPs in () show the purpose of the head noun. In this case, the genitive suffix is -wu attached to the complement noun.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
sankɨti-wu tɨti waramu-wu antsini gɨn-wu ʧiki sɨlaɣa-wu sɨki ʧampu-wu gomma mɨʃi-wu ŋargi
‘cotton for cloth’ ‘iron for spear’ ‘mud for house’ ‘sorghum for beer’ ‘rubber for shoe’ ‘honey for mead’
Compound nouns such as kɨnti gɨna ‘education house/school’, buni gɨna ‘coffee house/bar’ and bɨri gana ‘blood root’ behave like genitive NPs. Oblique cases are assigned to the respective nouns by postpositions preceding them. The dative or indirect object is marked by the postposition -s/-z as in: () ani kasa-s mas’af ʤep-u-ɣwa. I Kasa- book buy-- ‘I bought a book for Kasa.’ ()
ani ɲi-zi mas’af ʤep-u-ɣwa. I .- book buy-- ‘I bought a book for him.’
The ablative case ‘from’ is marked by the postposition -das. Note that the verb -nt is one of the relics of verbs with prefix conjugation. () ani pawi-das a-nt-u-ɣwa. I Pawi- -come-- ‘I came from Pawi.’ The allative ‘to’ is assigned by the postpositions -a and -u/-w (-u after vowels and -w after consonants). ()
Kas-i kɨntigɨn-a kaz-ɨɣwa. Kasa- school- go-:. ‘Kasa went to school.’
()
ani pawi-w kazt-u-ɣwa. I Pawi- go-- ‘I went to Pawi.’
Directionality is marked by the postposition -ʃi as in ɲi bɨʃtan-ʃi kaz-ɣwa [he churchtowards go-] ‘He went to church.’ The postposition -di marks commutative ‘with’ as shown in () and () below: () ani kasa-di kaz-u-ɣwa. I Kasa- go-- ‘I went with Kasa.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() ani yi-ʧu-di xw-ɨɣwa. I :-mother- eat-: ‘I ate with my mother.’ The locative case is marked by continuous and discontinuous postpositions. The postposition -aɣada (-aɣa ‘in’ and -da ‘by/side’) marks locative ‘in/inside/at’, as in gɨn-aɣada ‘in the house’. Location is also marked by the morpheme -(ampa)da (-ampa ‘top of ’ and -da ‘by/ on’) as in kani-(ampa)da ‘on (top of ) a tree’. Another locative is marked by aɣɨs (-aɣ ‘beneath/under’ and -(ɨ)s ‘by’) as in kaga-aɣɨs ‘under the bed’. As in bɨʃtan-ɨngɨra-s ‘behind the church’ and bɨʃtan-das fɨnfɨna ‘in front of the church’, the compound postpositions ɨngɨra-s (ɨngɨra ‘back’ and -s ‘by’) and -das fɨnfɨna (-das ‘for’ and fɨnfɨna ‘face by face’) mark location. The suffix -ss marks instrument as in waram-(ɨ)ss ‘with spear’ and direction as in maskot-ɨss ‘through the window’. The adpositional suffixes -des-fɨnda (a combination of des ‘by’ and fɨnda ‘face’), as in yi-des-fɨnda ‘before me’, and -des-ɨngɨra (a combination of des ‘by’ and ɨngɨra ‘back’), as in ‘yi-des-ɨngɨra ‘after me’ express time and location. The benefactive is marked by the dative postpositional suffix -s, as in ani kasa-s mesekerɣwa ‘I witnessed for the benefit Kasa.’ The malefactive is marked by the locative postposition -da, as in ani kasa-da mesekerɣwa ‘I witnessed to the detriment of Kasa.’
.. Verbs Kolisi exhibits a complex inflectional morphology. The verb is inflected for tense, aspect, and mood. The subject of any sentence is cross-referenced in verbs through portmanteau morphemes of person, number, and gender: three persons (first, second, and third), two numbers ( and ), and two genders ( and ). The subject is, therefore, an optional constituent in a sentence. Except for very few verbs with prefix conjugation, all verbal inflections appear following the verb stem and hence Kolisi exhibits suffix conjugation. In the ‘be’-verb conjugation, the present and past ‘be’-verbs show suppletive verb stems. The present ‘be’-verb stem is -ax for all persons which makes the overt realization of the subject obligatory.1 ()
ani ax ɨntɨ ax ɲi ax nu ax ɨntu ax ŋa ax
‘I am.’ ‘You () are.’ ‘(S)he is.’ ‘We are.’ ‘You (.) are.’ ‘They are.’
The past ‘be’-verb exhibits a complex internal morphology. Its stem has two forms ɨʃ- and ʤ-. ɨʃ- means ‘spend the day’ and ʤ- means ‘live (locative existential verb)’. Unlike the present ‘be’-verb, markers appear in all persons. In the first and second singular, -t; in Hetzron () calls such verbs which deviate from the full conjugation without inflections “impoverished” conjugation. 1
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
the ., Ø; and in the ., -t show (note also that ɨʃ- changes to ʤ- in the .). In the , -in, in the , -t, and in the , -i show (here too the verb stem changes: ʤ-, and , and ɨʃ-, ). The perfect tense is shown by -u for the first and second singular, -ɣwa for the . and , -i for the ., and -kwa for the and . () ani ɨʃ-t-u ɨntɨ ɨʃ-t-u ɲi ʤɨ-ɨɣwa ɲi ɨʃ-t-i nu ʤ-in-ɨɣwa ɨntu ɨʃ-t-ɨkwa ŋa ʤ-i-kwa
‘I was’ ‘you () were’ ‘he was’ ‘she was’ ‘we were’ ‘you () were’ ‘they were’
In Kolisi, there are two aspects: imperfect and perfect. The imperfect aspect expresses present and future and the perfect aspect expresses the past tense. Hetzron () calls the future an imperfect tense and definite aspect and the conditional future an imperfect tense and indefinite aspect in Awŋi. Compare the examples in ()–() below in which the imperfect is marked by the suffixes -aɣa with the and -a with the .: () ani hulu-ʧibi xw-aɣa. I every-time eat-: ‘I eat every time.’ () ani ʧa xw-aɣa. I tomorrow eat-: ‘I will eat tomorrow.’ () kas-i hulu-ʧibi ɨnkɨr-a. Kasa- every-time laugh-:. ‘Kasa laughs every time.’ () Kas-i ʧa ɨnkɨr-a. Kasa- tomorrow laugh-:. ‘Kasa will laugh tomorrow.’ In the perfect, actions performed in the past are expressed by the suffix -ɣwa. Hetzron () divides this tense into “perfect definite”, which means simple past, and “perfect indefinite”, which means present perfect. () ani ayiŋa xw-ɨɣwa. I yesterday eat-: ‘I ate yesterday.’ () Kas-i ayiŋa ɨnkɨr-ɨɣwa. Kasa- yesterday laugh-:. ‘Kasa laughed yesterday.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A verb in the present progressive tense shows the auxiliary prefixes /z-/ or /zɨkw-/ to the verb stem followed by and the imperfect tense markers. The past progressive has the pattern verb stem plus the auxiliary suffixes /-(ɨ)zɨʃt/ or /-az/ plus the marker and finally the perfect tense marker. The auxiliaries are derived from verbs which mean ‘live’ or ‘be present’ or ‘exist’. () ani ŋɨʃi z-ɨxw-aɣa. I now -eat- ‘I am eating now.’ () ani ayiŋa xw-ɨzɨʃt-ɨɣwa. I yesterday eat-- ‘I was eating yesterday.’ ()
kas-i ŋɨʃi zɨkw-ankɨr-a. Kasa- now -laugh-:. ‘Kasa is laughing now.’
() kas-i ayiŋa ɨnkɨr-az-ɨɣwa. Kasa- yesterday laugh-present/exist-:. ‘Kasa was laughing yesterday.’ Transitive verbs agree in person, number, and gender with their respective subjects. In the following conjugation, the and . are marked by Ø, the and . by i, the by (ɨ)n, the by i plus a change of the past marker initial segment from ɣ to k, and finally the only by the change of the past marker initial segment from ɣ to k. As shown in () below, transitive verbs are inflected for subject markers only. ()
ani kasa-wa kant-ɨɣwa. I Kasa- see-: ɨntɨ kasa-wa kant-i-ɣwa. you Kasa- see-- ɲi kasa-wa kant-ɨɣwa. he Kasa- see-:. ɲi kasa-wa kant-i-ɣwa. she Kasa- see-.- nu kasa-wa kant-ɨn-ɣwa. we Kasa- see-- ɨntu kasa-wa kant-i-kwa. you Kasa- see-- ŋa kasa-wa kant-ɨkwa. they Kasa- see-:
‘I saw Kasa.’ ‘You () saw Kasa.’ ‘He saw Kasa.’ ‘She saw Kasa.’ ‘We saw Kasa.’ ‘You () saw Kasa.’ ‘They saw Kasa.’
In the intransitive verb in (), the conjugation is very similar except for the slight change in the markers.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() ani xur-u-ɣwa. I sleep-- ɨntɨ xur-i-ɣwa. you sleep-- ɲi xur-u-ɣwa. he sleep-.- ɲi xur-i-ɣwa. she sleep-.- nu xur-n-ɨɣwa. we sleep-- ɨntu xur-i-kwa. you sleep-- ŋa xur-kwa. they sleep-:
‘I slept.’ ‘You () slept.’ ‘He slept.’ ‘She slept.’ ‘We slept.’ ‘You () slept.’ ‘They slept.’
Whereas each person is distinctly marked, the verbs in the first and third person masculine singular exhibit identical forms. In (), where the verb stem has a CVC syllable pattern, the , ., and marker morpheme i sounds like a diphthong, iy. In the conjugation in (), the and . are marked by Ø, and , ., and are marked by the suffix t(ɨ). ()
ani ɨnkɨr-ɣwa. I laugh-: ɨntɨ ɨnkɨr-t-ɨɣwa. you laugh-- ɲi ɨnkɨr-ɣwa. he laugh-:. ɲi ɨnkɨr-t-ɨɣwa. she laugh-.- nu ɨnkɨr-n-ɨɣwa. we laugh-- ɨntu ɨnkɨr-t-ɨkwa. you laugh-- ŋa ɨnkɨr-kwa. they laugh-:
‘I laughed.’ ‘You () laughed.’ ‘He laughed.’ ‘She laughed.’ ‘We laughed.’ ‘You () laughed.’ ‘They laughed.’
It is known that Agäw languages exhibit several different conjugational patterns in verbs. Kolisi too shares this property and hence the conjugations shown above represent only a few of the verb classes.
... Negation Negation is one of the verbal inflections. In Kolisi, the negative marker is suffixed to verb stems following markers. In () below, the negative suffix is -ya in the perfect. The verb stem is xur- ‘sleep’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() xur-a-ya. sleep-:- xur-i-ya. sleep-- xur-a-ya. sleep-.- xur-i-ya. sleep-.- xur-na-ya. sleep-- xur-ika-ya. sleep-- xur-ika-ya. sleep--
‘I have not slept.’ ‘You () have not slept.’ ‘He has not slept.’ ‘She has not slept.’ ‘We have not slept.’ ‘You () have not slept.’ ‘They have not slept.’
As shown in () below, in the imperfect, the negative marker is the most common suffix -la occurring right after the agreement markers. The negation marker is slightly modified in the second and third plural as -la/-li. () xur-a-la. sleep-- xur-i-la. sleep-- xur-a-la. sleep-.- xur-i-la. sleep-.- xur-ina-la. sleep-- xur-i-li. sleep-- xur-a-li. sleep--
‘I will not/do not sleep.’ ‘You (.) will not/do not sleep.’ ‘He will not/does not sleep.’ ‘She will not/does not sleep.’ ‘We will not/do not sleep.’ ‘You (.) will not/do not sleep.’ ‘They will not/do not sleep.’
The negative instrumental suffix is -ag (-g is commonly found in subordinate verbs). The suffix -ata, in this instance, seems to be an instrumental postpositional suffix. It might be a negation marker, as can be attested from the second person negative imperative -ta in Quara and Xamir, or it might be similar to the Awŋi atti- ‘lack/not have’. In the second case, there exists double negation. () waram-ag-ata spear-without gɨb-ag-ata stick-without adom-ag-ata hoe-without akaf-ag-ata spade-without
‘without spear’ ‘without stick’ ‘without hoe’ ‘without spade’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
As shown in ... below, the negation marker in the imperative forms is -ya.
... Interrogation In Agäw languages, -ma is the most common interrogative marker for ‘yes’ or ‘no’ polar interrogative questions. This is obvious from greetings, as in dikit-i-ma ‘Are you (.) fine?’, dekat-a-ma ‘Are you (.) fine?’, and dekat-ma ‘Are you () fine?’. Compare the following: () a-nt-u-ɣwa-ma? -come--- ti-nt-u-ɣwa-ma? -come--- ti-nt-u-ɣwa-ma? -come--- yi-nt-ɨnu-ɣwa-ma? -come--- ti-nt-ɨkwa-ma? -come-- yi-nt-kwa-ma? -come--
‘Did I come?’ ‘Did you () come?’ ‘Did (s)he come?’ ‘Did we come?’ ‘Did you () come?’ ‘Did they come?’
In other instances, question words can be used in interrogative constructions. As shown in () below, question words appear preceding a predicate or in sentence-initial position. () ay kazu? who went wan gɨnda? which house dar tarika? what story ɲi darda kazil? he for what went ɲi wani kazu? he when went ɲi waza/wada kazu? he where went ayiw biri dedaɣɨstu? of who ox stolen
‘Who went?’ ‘Which house?’ ‘What story?’ ‘Why did he go?’ ‘When did he go?’ ‘Where did he go?’ ‘Whose ox is stolen?’
The interrogative pronoun ayiw consists of two morphemes: ay ‘who’ and the possessive pronominal suffix -iw. dar-da ‘why’ combines dar ‘what’ and the locative -da. In the same way, waza and wada are combinations of the base of the interrogative pronoun -wa and the locative postpositions.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
... Passive Passive is one of the verbal extensions. An active verb changes to a passive by suffixing the morpheme -(ɨ)st as illustrated in () below: ()
xwzɨqkasguzɨnkwarädkw -
‘eat’ ‘drink’ ‘take’ ‘lead’ ‘call’ ‘slaughter’ ‘kill’
xw-ɨstzɨqw-ɨstkat-stguz-ɨstɨnkw-ɨstarädɨstkw-ɨst-
‘be eaten’ ‘be drank’ ‘be taken’ ‘be led’ ‘be called’ ‘be slaughtered’ ‘be killed’
... Causative The causative is morphologically marked by the suffix -ɨts following the verb stem: () agxwfaykɨntankɨr-
‘bring’ ‘eat’ ‘search’ ‘learn’ ‘laugh’
ag-ɨts-ɣwa xw-ɨts-ɣwa fay-ɨts-ɣwa kɨnt-ɨts-ɨɣwa ankɨr-ts-ɨɣwa
‘He caused to bring.’ ‘He caused to eat.’ ‘He caused to search.’ ‘He caused to learn.’ ‘He caused to laugh.’
... Imperative The imperative for the is the first syllable of the verb stem having dropped the markers. In the , the verb stem appears with the plural marker suffix. () kazxwmiyxur-
‘go’ ‘eat’ ‘hold’ ‘sleep’
ka! xw! mit! xur-i!
‘go! ()’ ‘eat! ()’ ‘hold! ()’ ‘sleep! ()’
kaz-an! xw-an! mit-an! xur-an!
‘go! ()’ ‘eat! ()’ ‘hold! ()’ ‘sleep! ()’
The negative imperative is marked by the morpheme -ya identical to the negative marker in the imperfect. Whereas this negative formative is very common in Kolisi, in other Agäw languages, it is rare. () xwzɨqkɨntgɨʃxur-
‘eat’ ‘drink’ ‘learn’ ‘dig’ ‘sleep’
xw! zɨq! kɨnt! gɨʃ! xuri!
‘eat!’ ‘drink!’ ‘learn!’ ‘dig!’ ‘sleep!’
xw-ɨy-ya! zɨq-i-ya! kɨnt-i-ya! gɨʃ-i-ya! xur-i-ya!
‘Do not eat!’ ‘Do not drink!’ ‘Do not learn!’ ‘Do not dig!’ ‘Do not sleep!’
As in many other Ethiopian languages (see Ferguson ), the imperative for the verb -nt‘come’ in Kolisi is aw! ‘Come!’ for second singular and awan! for second plural. Its negative form is ant-i-ya ‘Do not come!’. Another verb is ag- ‘bring’, which has a suppletive imperative form kataw ‘Bring!’ and the negative imperative ak-i-ya ‘Do not bring!’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
... Converb Converbs (also called gerundives) have the following pattern: verb stem + markers + gerundive suffix. Note that the gerundive marker has three forms: -ata (, , .), -ama (., , ), and -ana (). () xur-ata sleep-: xur-iy-ata sleep-- xur-ama sleep-:. xur-iy-ata sleep-.- xur-n-ana sleep-- xur-ik-ama sleep-- xur-k-ama sleep--
‘I, having slept’ ‘you (.), having slept’ ‘he, having slept’ ‘she, having slept’ ‘we, having slept’ ‘you (.), having slept’ ‘they, having slept’
Converbs do not stand on their own and hence take main verbs to convey a complete message, as shown in () and () below: xur-u-ɣwa. () ani zɨq-ata I drink-: sleep-- ‘I, having drunk, slept.’ () ɲi zɨq-ama xur-e-ɣwa. he drink-:. sleep-.- ‘He, having drunk, slept.’ Like any other complex clauses, gerundive clauses precede main clauses (see further in section ..).
... Jussive The jussive is marked by the morpheme (ɨ)s. The internal structure of a jussive verb is verb stem + + jussive morpheme. Compare the jussive conjugations for the verb -nt- ‘come’ and kaz- ‘go’ in (). Note that the first exhibits a prefix conjugation and the second a suffix conjugation. () a-nt-ɨs. -come- yɨ-nt-ɨs. .-come-
‘Let me come.’ ‘Let him come.’
kaz-ɨs. go- kaz-ɨs. go-
‘Let me go.’ ‘Let him go.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
tɨ-nt-ɨs. ‘Let her come.’ .-come-JUSS yɨ-nt-ɨn-ɨs. ‘Let them come.’ -come--
kat-ɨs. ‘Let her go.’ go- kaz-ɨn-ɨs. ‘Let them go.’ go--
In the first conjugation, for the , the prefix is a-. For the ., it is y- and for the . t-. In the jussive, is marked by the discontinuous morphemes y- . . . -n. In the second paradigm, where the usual suffix conjugation takes place, the and . have a zero marker. In the ., the underlying representation kaz-t- changes to kat-, deleting the final segment of the verb stem z and replacing it with the feminine t. In the , is marked by the most frequently occurring morpheme -n.
.. Adjectives In Kolisi, simple adjectives are rare. Rather, most of the modifiers have a relative verb structure. Like nouns, adjectives show gender distinction and hence those ending in -i are masculine and those ending in -a are feminine, as in ɨskawi/ɨskawa ‘new’, wiliʤi/wiliʤa ‘old’, fuʧʧi/fuʧʧa ‘white’, and tsɨlli/tsɨlla ‘small’. Compare the following sentential examples: () a. ani ɨmpɨl-i buz-i tsark-i gɨsaŋ-i kant-ɨɣwa. I one- fat- black- dog- see-: ‘I saw one fat black dog.’ b. ani ɨmpɨl-a buz-a tsark-a I one- fat- black- ‘I saw one fat black bitch.’
gɨsaŋ-a kant-ɨɣwa. dog- see-:
In NPs, adjectives appear preceding their noun head as in dɨmmi dunisi ‘red potato’, waɣa arfa ‘full moon’, minika aqi ‘many persons’, and yɨzan-ka kan-ka ‘strong woods’. As shown in (), adjectives can appear successively before their noun heads. However, there is no strict rule on the sequence of adjectives so long as they all precede the noun. In sentences like aqi dɨxi aɣ ‘The man is poor.’ and bɨri dɨmmi aɣ ‘Blood is red.’, adjectives are part of the predicate following the noun modified. Numerals, as quantifiers, can be cardinals and ordinals, which are the only expanded modifiers in Kolisi. The digits from one to ten are: laɣw ‘one’, laŋa ‘two’, ʃuɣwa ‘three’, sezza ‘four’, ankwa ‘five’, walta ‘six’, laŋatta ‘seven’, suɣatta ‘eight’, sesta ‘nine’, and sɨkka ‘ten’. The digits all end in -a, which means that they have a feminine interpretation. Other nouns, however, end in the vowel -i, which indicates the masculine gender. The cardinal numeral ‘one’ has two variants: laɣw and ɨmpɨl. Whereas the former appears in isolation and when counting, the latter appears as a modifier, as in ɨmpɨl aqi ‘one person’. The numerals from twenty up to ninety are decimal-based compound nouns consisting of the digits and the tens as in sizzi-sɨkka ‘forty’, ankw-sɨkka ‘fifty’. The word for hundred is lix. The ordinal numerals have the structure: cardinal + agentive nominal marker-anti, as in ɨmpɨl-anti ‘first’, laŋ-anti ‘second’, and ʃuɣw-anti ‘third’. The final vowel of the cardinal numeral base is deleted before the suffix to avoid vowel sequencing not allowed in the language.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Adverbs Time adverbials include ayŋa ‘yesterday’, ʧa ‘tomorrow’, ŋɨʃi ‘now’, naka ‘today’, ɨmbɨt ‘quickly’, lɨbbuz ‘slowly’, maʧi ‘Monday’, and säbiti ‘Tuesday’. In sentences, adverbs always appear before the verbs they modify. Compare the examples below: () kas-i mäs’afo lɨbbuz anäbäb-ɣwa. Kasa- book slowly read-:. ‘Kasa read the book slowly.’ ()
ʤeri ayŋa ɨmbɨta ɨmbɨtama kɨntigɨn-a kaz-ɨɣwa. the child yesterday quick.quick school-to go-:. ‘The child went quickly to school yesterday.’
() ani ʧa xw-aɣa. tomorrow eat-: ‘I will eat tomorrow.’
.. Derived nominals Nominal derivation in Kolisi is deverbal through suffixation. In that way, agent nominals are derived from their verbal counterparts by suffixing -ant. Note that the final vowel is a masculine marker. ()
axxwbambzɨqkɨnt-
‘know’ ‘eat’ ‘swim’ ‘drink’ ‘learn’
ax-ant-i xw-ant-i bamb-ant-i zɨq-ant-i kɨnt-ant-i
‘knowledgeable’ ‘eater’ ‘swimmer’ ‘drinker’ ‘student’
Infinitival nominals are derived from verbs by suffixing -(i)ŋ in: () xwgɨbɨrtmiyɨnkuy-
‘eat’ ‘work’ ‘hold’ ‘call’
xw-ɨŋ gɨbɨrt-ɨŋ miy-ɨŋ ɨnkuy-ɨŋ
‘to eat/eating’ ‘to work/working’ ‘to hold/holding’ ‘to call/calling’
Abstract nominals are derived by suffixing the morpheme -ti: ()
xuna nɨʧtar dɨɣi habtam
‘female’ ‘male’ ‘poor’ ‘rich’
xun-ti nɨʧtar-ti dɨɣ-ti habtam-ti
‘femaleness’ ‘maleness’ ‘poorness’ ‘richness’
Instrumental nominals are derived by suffixing the morpheme ts as in gɨŋ-ts-i ‘a thing to run with’, zɨq-ts-i ‘a thing to drink with’, and kew-ts-i ‘a thing to cut with’. Derived nominals in the majority of cases end in -i and hence are masculine.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Conjunctions The morpheme -(ɨ)sta (also a subordinator ‘as/like’) is the conjunction which conjoins two nouns. It appears attached to the preceding noun. () ani-sta ɨntɨ I-and you ‘you and I’ () kasa-sta Aster Kasa-and Aster ‘Kasa and Aster’ () fɨlay-ɨsta bagi goat-and sheep ‘goat and sheep’ Disjunction between nouns is shown by the morpheme woyki ‘or’ (a modified borrowing from the Amharic woyim). () ani woyki ɨntɨ I or you ‘I or you’ () kas-i woyki aster Kasa- or Aster ‘Kasa or Aster’ () fɨlay woyki bagi goat or sheep ‘goat or sheep’ The connector -aŋzi ‘because’ conjoins compound sentences, as in ɲi yi-nt-u-aŋzi an kazala ‘I will not go because he will come.’
. S
..................................................................................................................................
.. Word order Kolisi, like the rest of the Cushitic and Semitic languages, has an S–O–V word order as the dominant word order. () an anki xw-ɨɣwa. I Injera eat-: ‘I ate Injera.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() ɲi beru kas-u-ɣwa. he gun take-.- ‘He took the gun.’ Since the verb is morphologically complex, Kolisi is a pro-drop language. Thus, especially the subject is an optional constituent in a sentence and as a result the sentences in () and () above can appear without the subject. When a sentence contains a ditransitive verb, the direct object precedes the indirect object. () kas-i ʤer-ɨs mas’afu ag-u-ɣwa. Kasa- boy-for book give-.- ‘Kasa gave the book to the boy.’ () ŋa y-ɨs sankɨti ʤew-ɨkwa. they me-for cloth buy-: ‘They bought cloth for me.’ The unmarked pattern is S + DO + IO as shown in () and () above. However, the alternation of the position of DO and IO is also possible. Hence, the above sentences can also be constructed as kasi mas’afu ʤerɨs aguɣwa and ŋa sankɨti yɨs ʤewɨkwa. In complex sentences like ɲi kani-da fɨŋu kal-a [he tree-on climb can-IPFV:M.SG] ‘He can climb a tree.’, the subject is followed by a prepositional phrase, then an infinitival verb, and finally an auxiliary verb.
.. Verb phrase The smallest verb phrase in a predicate is the verb itself. As a head-final language, verbs appear following their complements in all Kolisi phrasal categories. In () and (), where the verb is an intransitive verb, the VP = V: () kas-i xur-ɨɣwa. Kasa- sleep-:. ‘Kasa slept.’ () mare ɨnkɨrt-ɨɣwa. Mare laugh-:. ‘Mare laughed.’ Bigger VPs, as in () and (), contain a PP and a verb. When a sentence contains postpositional phrases, these phrases appear between the subject of the sentence and the verb. () ani ɲi-di kɨntigɨn-a ant-ɨɣwa. I she-with school-to come-: ‘I came to school with her.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() ani ɲi-da ɨnkɨr-ɨɣwa. I he-at laugh-: ‘I laughed at him.’
.. Noun phrase Kolisi is a head-final language. Hence, NPs have a modifier + noun structure. In NP constructions like buzi aqi ‘fat ape’, laŋa gɨni ‘two houses’, sɨnkut aɣw ‘clean water’, and minʧika gɨnka ‘many houses’, quantifiers appear preceding their qualified heads. When an NP has more than one modifier, the pattern looks like quantifier + attributive adjective + noun. () laŋa kuntali tabi ‘two sacks of teff ’ ɨmpɨl liggizmi aqi ‘one tall person’ When the head noun is in the plural, its modifiers agree in number and gender, as shown in () below. ()
laŋa leggazɨm-ka wofram-ka kan-ka two tall- thick- tree- ‘two tall thick trees’
() ɨmpɨl-i buz-i tsark-i gɨsaŋ-i one- fat- black- dog- ‘one fat black dog’ () ɨmpɨl-a buz-a tsark-a gɨsaŋ-a one- fat- black- dog- ‘one fat black bitch’ In the following NPs, the head noun aqi ‘man’ () and ʤeri ‘child’ () are modified by relative modifiers preceding them. () ani ayiŋa kant-u aqi I yesterday see-: man ‘the man whom I saw yesterday’ () fatana fuw-i ʤeri exam pass-:. child ‘the boy who passed the exam’ In relative clauses, the verb agrees with its subject.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Subordination In Kolisi, the subordinate clause precedes the main clause. () ɲi yi-nt-a-ni an kaz-a. he .-come-- I go-: ‘When he comes, I will go.’ () nu xur-n-u-ta ani ax-a z-ɨkwa. we sleep--- I know-: exist/present- ‘I know that we slept.’ So long as the subordinate verb precedes the main verb, the constructions above are acceptable, as in ani ɲi yintani kaza and ani nu xurnuta axa zɨkwa. When the main and subordinate clauses share the same subject, as in () and (), the subject appears once at the beginning of the sentence. () ani bunu zɨq-a-ta saxan kats-a-la. I coffee drink-- well take--: Lit. ‘If I drink coffee, sleep doesn’t take me well.’ () ɲi xur-a-ni xɨmpɨp-ta. she sleep-- snore-:. ‘When she sleeps, she snores.’ Kolisi subordinators appear following the VSTEM and elements. In temporal subordinate verbs, the subordinator is -ni, as in an axw-a-ni ‘when I eat’, ɨntɨ xw-ɨye-ni ‘when you () eat’, etc. The same suffix appears in conditional subordinate verbs, as in an a-nt-u-ni ‘if I come’, ɨntɨ t-yint-u-ni ‘if you () come’, etc. The -ta suffix also marks temporal adverbial subordinate clauses, as in ɲi ʃiwa yint-u-ta xur-e-ɣwa ‘He slept immediately after he came.’ The suffix -das also appears in conditional clauses replacing -ni, as in ɲi zɨq-u-das sekar-a ‘If he drinks, he becomes drunk.’, kasi xw-u-das xur-ɨs ‘If Kasa eats, let him sleep!’ The thatclause is indicated by the subordinator -ta, as in ani ɨntɨ xur-ɨw-ta ax-askwa ‘I know that you slept.’ In reason adverbial subordinate clauses, the subordinator is -ŋɨzi, as in kasi wɨfär-ŋɨzi ɨnzɨŋ guʤ-u-ɣwa ‘Because Kasa became fat, he could not walk.’ Locational subordination takes place with the subordinator -da (also locative), as in kasi an zɨku sɨrbi-da yi-nt-a ‘Kasa will come where I live.’ In relative clauses, a relativizer is attached to a relative verb, as in ŋa ayŋa kant-ɨn-u aqi kɨr-u-ɣwa ‘The man whom they saw yesterday died.’ The suffixes -ama/-ata mark subordination in gerundive subordinate clause, as in kasi yim-ama xur-u-ɣwa ‘Kasa, having prayed, slept.’ The subordinator in concessive clauses is -kɨla, as in aster mitsini-kɨla ʤeru kaman-ɨŋɨs fay-te-la ‘Even though Aster gets married, she does not like to give birth.’ Compare also the following adverbial subordinate clauses connected to time, manner, location, etc.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() aster ani ʤuŋ-das fɨnu gɨbri-da ka-ti. Aster I wake up- front work-to go-:. ‘Aster went to work before I woke up.’ ()
aster tabli kɨru-das kɨta mits-e-ɣwa. Aster her father die- behind marry-.- ‘Aster got married after her father died.’
() kas-i giŋ-ani ɨwɨts-a. K- run- cough-:. ‘Kasa coughs while he runs.’ () kas-i yi-nt-ɨnba ani fur-a. Kasa- come- I rest-:. ‘I will take rest until Kasa comes.’ () ani kaz-u-das zur-u-ɣwa. I went-- returned-: ‘I returned from where I went.’ () ani aras-ŋɨs biri fayt-a. I plow- ox need-:. ‘I need ox to plow.’ Some of the subordinators such as -des, -da, -s, -ba, and -ta are postpositions. The subordinator -ŋɨs combines two morphemes (the infinitival nominal marker -ŋ and the postposition -ɨs). Other subordinators such as the relative markers -i, -u, and gerundive markers -ta/-ma are complementizers that make independent verbs dependent.
. F
.................................................................................................................................. This grammatical sketch of Kolisi can only give some insight into the language. The vowel system and the verb morphology need particular attention. An in-depth description is extremely vital for mapping the status of the language, whether or not it is a dialect of Awŋi. Beyond shared grammatical features, extralinguistic issues complicate determining the relationship between the Kolisi and Awi ethnoliguistic groups and their respective languages. The Awi, who feel superior, say, “The Kolisi speak a funny Agäw, not a refined one like ours.” They also hesitate to accept their common ancestry. The Kolisi admit slight variations between their codes but recognize their common descent. The striking similarities of Kolisi with Kemantney and Quara cannot be ignored. Regrettably, both Quara and Kemantney are seriously endangered languages. Therefore, a deeprooted language survey is required in order to identify the languages and dialects in this part of the country.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A The fieldwork for this research was sponsored by the German Society for Endangered Languages (DGBS), to which I am grateful. I am deeply indebted to my informants Ato Negatu Wase and Ato Agedew Ayal. The data were elicited in January from native speakers of the language in Jawi, about km to the northwest of Addis Ababa.
R Allene, Y. (). ‘A Grammar of Kulazngi’. Ph.D. thesis, Addis Ababa University. Appleyard, D. (). ‘The internal classification of the Agäw languages: a comparative and historical phonology’, in J. Bynon (ed.), Current Progress in Afro-Asiatic Linguistics: Papers of the Third International Hamito-Semitic Congress. Amsterdam: John Benjamins, –. Appleyard, D. (). A Comparative Dictionary of the Agaw Languages. Cologne: Rüdiger Köppe. Cowley, R. W., Birru, T., and Adal, Z. (). ‘The Kunfäl people and their language’, Journal of Ethiopian Studies : –. Ferguson, C. A. (). ‘The Ethiopian language area’, in M. L. Bender, J. D. Bowen, R.L. Cooper, and C. A. Ferguson (eds.), Language in Ethiopia. Oxford: Oxford University Press, –. Hetzron, R. (). The Verbal System of Southern Agäw. Berkeley and Los Angeles, CA: University of California Press. Hetzron, R. (). ‘The Agäw languages’, Afroasiatic Linguistics : –. Joswig, A. (). ‘The status of the high central vowel in Awngi’, in S. Uhlig (ed.), Proceedings of the XVth International Conference of Ethiopian Studies.Wiesbaden: Otto Harrassowitz, –. Zelealem, L. (). The Kemantney Language: A sociolinguistic and grammatical study of language replacement. Cologne: Rüdiger Köppe.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. T
.................................................................................................................................. I is spoken in northern Tanzania on the high plateau between Lake Manyara and Lake Eyasi. It is the largest Southern Cushitic language, with roughly half a million speakers. Its closest relatives are Gorwaa, Alagwa, and Burunge. These four languages form the southernmost group of Cushitic languages (see Kießling, chapter of this volume). The dialect variation within Iraqw is negligible. Major linguistic sources for Iraqw are Whiteley (), Nordbustad (), Maghway (), Mous (), Berger and Kießling (), and Mous et al. ().
. P
..................................................................................................................................
.. Vowels There are five vowels i, e, a, o, u, which have long counterparts ii, ee, aa, oo, uu; in addition there are two rising diphthongs ay, and aw. In vowel coalescence the first vowel is dropped, except for ou and ayu, which both result in o. A rounded vowel preceded by a velar or uvular consonant and followed by a nonrounded vowel becomes a glide w. Epenthetic vowels assimilate to the following vowel if the intervening consonant is “back” (glottal, pharyngeal, uvular, or velar), and to the preceding vowel if the intervening consonant is a glottal stop.
.. Consonants The Iraqw consonant phonemes are given in Table .. Few lexical items contain palatal consonants. Iraqw sounds very different from the surrounding languages, which is primarily due to the pharyngeal fricatives and the voiceless uvular stop, which can be realized as an affricate.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
f
s
sh
nasal
m
n
ny
trill glide
kw
x
xw
ng
ngw
q
/
’
hh
h
qw
tl
fricative resonant
k
glottal
ts
pharyngeal
ch
p
eject. affricate
gw
rounded uvular
t
voiceless stop
g
uvular
j
labialised velar
palatal
d
velar
alveolar
b
voiced stop
lateral
labial
Table . Iraqw consonant phonemes in orthographical symbols
hl
l r y
w
The following non-IPA symbols and digraphs are used in this chapter: ng is a voiced velar nasal stop, and between vowels it is followed by an oral velar stop; ts is a voiceless ejective alveolar affricate; tl is a voiceless ejective palatal stop with lateral release; hl is a voiceless lateral fricative; hh is a pharyngeal voiceless fricative; / is a pharyngeal voiced fricative followed by creaky voice. Some of the most common phonological processes are the following: ) If an alveolar consonant is added in morphology to another alveolar oral consonant, the first one is dropped, e.g. dasi-r-dá’ (girl--) ! dasidá ‘that girl’. ) The glottal consonants ’ and h are dropped before an oral consonant, e.g. duunga’-ká ! duungaká ‘this nose’. ) A glide or labialisation is dropped before a rounded vowel, e.g. sidw-o ! sido ‘pots’ and before a labialized consonant, e.g. masasakw-mo ! masasakmo ‘ant’.
.. Tone Iraqw is a restricted tone language with a high tone (´) and a low tone (left unmarked in this chapter), which have a grammatical function and virtually no lexical function. Some suffixes have a high tone; notably demonstrative and possessive suffixes are all hightoned. Tone plays a role in verbal inflection; for example, in the past tense all final vowels are high-toned. Tone is distinctive on final syllables only; by adding high-toned suffixes a series of high tones can occur, e.g. ino’ín-á-dá’ (they--) ‘those ones’, but there is maximally one tonal movement on a word: once the tone goes up, it remains high till the last syllable in the word, as in this example. The tone patterns are L* and L*H*. Most words have no (high) tone (L*), which is why I do not consider Iraqw a (pitch) accent language.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Shape of syllables, roots, and words All syllables have an onset (note that the glottal stop as word onset is not written); the peak consists of a long or short vowel, and the coda is empty or consists of a single consonant or a homorganic nasal-oral stop combination. The minimum word consists of a heavy syllable; a CV-structure cannot form an independent word if the vowel is short. The preferred pattern in noun roots is Ci/a/uC(C)ee/aa/ooC- or CViCViC-; the first pattern has a short high or low vowel in the first syllable and a long mid or low vowel in the second; the second pattern consists of identical vowels (any of the five). These preferred patterns are violated when morphemes are added or have been added historically, and are not valid for loans. Three subsequent syllables with short vowels are avoided by deleting the vowel of the middle syllable, e.g. gawid-en (difficult-) ! gawden.
. N
..................................................................................................................................
.. Gender of nouns There is a three-way gender distinction, which includes an interesting feature “plural” () besides masculine () and feminine (). The gender of a noun is defined by agreement. Subject agreement on the verb is purely with gender, not with number. The examples in () below show the three gender values defined by subject agreement. These examples are chosen to highlight some of the remarkable features of Iraqw. The word daaqaay for ‘boys’ is plural in number, but it triggers . agreement on the verb and hence is masculine in gender. The word hhayse for ‘tails’ is plural in number, but it triggers . agreement on the verb and hence is feminine in gender. The singular of this word, hhayso, triggers agreement on the verb and hence is “plural” in gender even though it is singular in number. Note that singular and plural forms of the same lexical item do not need to have the same gender. These more general properties of Cushitic gender and number are discussed in some detail in Mous (). ()
a. daaqay i giilín. boys fight:. ‘The boys are fighting.’
i giilín. fight:. ‘He is fighting.’
b. hhayse i harweeriiríin. i harweeriiríin. tails make.circles:. make.circles: . ‘The tails make circles.’ ‘She is making circles.’ c. hhayso i harweeriiriná’. tail make.circles: ‘The tail is making circles.’
i harweeriiriná’. make.circles: ‘They are making circles.’
Nominal modifiers agree in gender with the head noun. Demonstrative suffixes are preceded by a linker agreeing with the gender of the noun, u for , r for , zero for , as shown in Table .. Demonstrative and possessive pronouns agree in gender with the noun they refer to, but the agreement markers of and have merged to ku.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Gender in demonstratives in Iraqw Word
Meaning
Noun + demonstrative suffix
Demonstrative pronoun Gender
tsamas dasi hhayso
‘giraffe’ ‘girl’ ‘tail’
tsamas-u-qá’ dasi-r-qá’ hhayso-qá’
ku-qá’ ta-qá’ ku-qá’
The gender of a noun is not predictable, either by its shape or by its meaning. The singular and the plural of a noun are in general different in terms of gender. Words for male beings or entities can be feminine such as hayda () ‘bull’, barisee () ‘elders, older men’, and na/ani () ‘penis’, and the other way around, such as /atlaráy () ‘old cows’, dasu () ‘girls’, and isaangw () ‘nipple, breast’. From the shape of the noun one can guess its gender, but not predict it. There are two kinds of gender indications: number suffixes have their inherent gender; thus, the number suffix tells you which gender the noun has, but not all nouns contain a number suffix. A second kind of gender indication is that noun roots ending in e or i tend to be feminine, and those in u masculine. Loanwords from Swahili follow this pattern. Nouns ending in a or o can have any gender. Swahili loans in a tend to be feminine. Most nouns in fact end in a vowel. Nouns ending in r or l are masculine; those in t are feminine.
.. Number of nouns Both singular and plural forms may be derived by suffixation; compare (a, b) with (e). There are many different suffixes forming a plural, and it is unpredictable which plural suffix a particular noun will take; compare (a) with (b). Moreover, not every noun has a singular and a plural form. Some nouns have no plural form, although this is not always evident from their meaning. Compare (c) with (d); the single number form is singular in (c) but plural in (e) and unspecified for number in (f). Some nouns have several plural forms, sometimes with a difference in meaning, as in (h), sometimes without a difference in meaning, as in (g). The gender of the plural noun is often different from the gender of the singular noun. The fact that a singulative exists has no influence on the number interpretation of the base form, as (i) illustrates. Most plural formations involve suffixation but some involve infixation of -ee- before the final root consonant as well, (h). ()
guhhul-áy / guhhul-i’i ‘stick, club’ kukum-áy / kukum-aawee ‘sling’ sori ‘a reddish bird’ poqi / poq-a ‘red or yellow bishop’s bird’ daaqay ‘boys’ tluway ‘rain’ / mashoot / two derived plurals: masoo-du or mashoot-ma’ ‘leather bag’ hhamhhamo ‘one eyebrow’, hhamhham-’i ‘eyebrows, -lashes, -lids of one person’, hhamhhee>m-o ‘of many people’, hhamhham-i ’one hair’ i. tsuunkáy ‘dung’ , singulative: tsuunk-i ‘one piece of dung’ a. b. c. d. e. f. g. h.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Number and gender agreement in adjectives There is a two-way number distinction in terms of agreement, singular and plural. The adjective follows the noun and is the only word class that agrees in number with the noun. Nouns that are plural in number agree with the plural form of the adjective in terms of gender, even if the noun itself is () or (), and not (), as in (). The plural is formed by the suffix -en, or the suffix -t, which shortens the preceding syllable. ()
duunga’ () ‘nose’ /ameena () ‘women’ garmo () ‘boy’ hhaysó tleer ‘a long tail’
duungáa ur ‘big nose’ /ameenár uren ‘big women’ garmóo niiná ‘a small boy’ hhaysér tlet ‘long tails’
Adjectival number agreement is semantically based with nouns that either refer to a collection (singular adjective) or a distribution of entities (plural adjective). For example, the loan notay (, /) ‘bank note(s), paper money’ modified by a singular adjective ur ‘big’ refers to ‘a lot of notes, a lot of money’, notóo úr, but with a plural adjective the interpretation of distributed notes each of which is big or of high denomination is enforced, notóo urén. Gender agreement of adjectives is purely tonal in nature. Adjectives following a () noun have a high tone on the final syllable. Adjectives following () or () nouns have a low final syllable, as in (). Gender is also marked in the gender linker that is suffixed to the modified noun, which occurs in an antigenitive (construct) case form. The fact that adjectives have gender and number agreement independently shows that it is not possible to reanalyze the gender as a value of number. When used predicatively, adjectives follow the adjectival copula, which also agrees in gender with the head noun, as in (): ()
na/ii ki ququmat. hhayso ki ququmaar. children. . short:: tail. . short: ‘The children are short.’ ‘The tail is short.’
The gender is in three ways related to plurality as an exponent of number marking: ) The subject agreement on the verb for nouns is that of the third person plural; see (1c). ) All number suffixes which impose the gender value have multiple reference but not all multiple reference suffixes are , and singular nouns do exist. ) Coordinated nouns take agreement on the verb, as in (5). ()
kitangw nee mesa i gwarangwarimit-iyá’ asmá kunseeli. chair and table shake- because earthquake ‘The chair and the table shake because of the earthquake.’
.. Numerals All numerals under ten are feminine and have a high tone: wák () ‘one’, tsár () ‘two’, tám () ‘three’, tsiyáhh () ‘four’, koo’án () ‘five’, lahhóo’ () ‘six’, fáanqw () ‘seven’, dakáat ()
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
‘eight’, and gwaléel () ‘nine’. The numerals for , , and , are masculine and have low tone: mibangw () ‘ten’, : mibeeri (), tsiiru () ‘hundred’, : tsiiree (); and kuma () ‘thousand’, : kumee (). They form the basis for higher numbers and therefore have plural forms. The base of the numeral system is ‘ten’. Higher numbers are made by addition of one of the lower numbers and multiples of ‘ten’, ‘hundred’, and ‘thousand’. Multiplication is by juxtaposition, addition by nee ‘and’, which is used only once in a numeral expression. Multiplication by ‘one’ is not expressed for ‘ten’, but it is for ‘hundred’ and ‘thousand’. ()
mibangw nee wák /+/ ‘eleven’ kume tsár tsirú wák mibeeri tsár nee faanqw /. . .+/ ‘’
.. Pronouns The personal pronouns are presented in Table .. A cross-linguistically rare phenomenon of Iraqw is the presence of a sex distinction in the second person (singular) but not in the third person; the possessive suffixes show no sex distinction at all. In the object pronouns there is a gender/sex distinction in both third and second person (singular). The personal pronouns occur in the same syntactic environments as other nouns. The full forms can receive nominal suffixes, but the short forms cannot. The personal pronouns are used for contrast or emphasis only.
Table . Iraqw personal pronouns Meaning
Full form
Short form
aníng
án
.
kúung
kú
.
kíing
kí
inós
ís
atén
át
kuungá’
–
ino’ín
inín
The possessive pronouns (Table .) consist of the possessive suffix preceded by ko, if referring to a noun of the genders or , and by to if referring to . The possessive suffixes are preceded by a linker which agrees in gender with the head noun. The possessive pronouns are used in sentences like (). ()
hiimuwí a kwe’ée’. ‘This rope is mine.’ dasirós a taqá’. ‘His girl is that one.’
hiimuwí kwe’ée’ ‘this rope of mine’ dasirós taqá’ ‘that girl of his’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Iraqw possessives pronouns and suffixes ,
possessive suffix
tsamas ‘giraffe’
dasi ‘girl’
hhayso ‘tail’
kwe’ée’
te’ée’
’ée’
tsamas-u-’ée’
dasi-r-’ée’
hhayso-’ée’
kók
tók
ók
tsamas-uw-ók
dasi-r-ók
hhays-ók
kós
tós
ós
tsamas-uw-ós
dasi-r-ós
hhays-ós
torén
rén
tsamas-u-rén
dasi-rén
hhayso-rén
kohúng tohúng húng
korén
tsamas-u-húng
dasi-r-húng
hhayso-húng
ko’ín
tsamas-u-’ín
dasi-r-’ín
hhayso-’ín
’ín
to’ín
There are four degrees in deixis in the demonstratives (Table .): . near me, . near you, . near him, but visible, . far away. Demonstrative suffixes and can be reduplicated. Demonstrative is used for referential meaning. The proximal demonstrative for gender nouns is ká and not í.
Table . Iraqw demonstrative pronouns and suffixes
hhasam () ‘dilema’
gi’i () ‘ghost’
hhasamarí
gi’iká
kwisíng tisíng kusíng síng hiimusíng
hhasamasíng
gi’isíng
kuqá’
tiqá’
kuqá’
qá’
hiimuqá’
hhasamarqá’
gi’iqá’
kudá’
tidá’
kudá’
dá’
hiimudá’
hhasamadá’
gi’idá’
kwí
1
sfx
tí
kuká
í, ká hiimuwí
1
hiima () ‘rope’
Also taqa’.
The demonstrative pronouns are used to replace the head noun, but also if the head noun is already modified by another (possessive) suffix.
... Indefinite pronouns and suffixes The indefinite suffixes (Table .) are used only if the speaker considers it important to mark that the noun is not specific. Nouns can be indefinite without the indefinite suffix. The indefinite suffixes have a low tone, while the inherently definite demonstrative and possessive suffixes all have a high tone.
Table . Iraqw indefinite suffixes and pronouns Indef sfx
Gender
Example
Meaning
Pronoun
ko ka kaariya’
tsamas-u-ko dasi-r-ka hhayso-kaariya’
‘a giraffe’ ‘a girl’ ‘a tail’
koo-ko too-ka koo-kaariya’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
There are no relative pronouns in the language. Instead, the relative clause follows the head noun immediately and the head noun is marked with the antigenitive (construct) case (section ...).
... Interrogative pronouns and other question words Most question words (but not the last four in Table .) are based on nouns and formed by the addition of a high-toned suffix -má or -lá.
Table . Iraqw question words Question word
Meaning
Base
Meaning
diimá aamá heemá adoomá idoomá daqmá gaalá xaylá milá magá’ gaalay
‘where’ ‘where’ ‘who’ ‘how’ ‘how’ ‘at what time’ ‘which’ ‘when’ ‘what’ ‘how many’ ‘where, why not’
di () aamo () hee () adoo () idoo () daqa () ga ()
‘place’ ‘place’ ‘man’ ‘manner’ ‘manner’ ‘moment’ ‘thing’
.. Noun phrase structure, antigenitive (construct) case In a typological sense, the structure of the noun phrase is not in line with the verb-final basic word order, as modifiers follow the head noun in Iraqw. The head noun that is modified by an adjective, another noun, a numeral, or a relative clause is marked as being modified by an antigenitive case marker termed construct case () (see Table .). This marking consists of a gender linker and a final high tone on the noun. If the head noun is understood, it may be replaced by a construct pronoun (c). Relative clauses are final within the noun phrase but possessives precede demonstratives or indefinite modifiers.
Table . Construct case suffixes and pronouns in Iraqw Construct sfx
Gender
Example
Meaning
Construct pronoun
-u-H -r-H -a-H
tsamas-ú dasír hhaysá
‘a giraffe’ ‘a girl’ ‘a tail’
oo ar awa
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
a. tiqtí-r hlee-r-ós disease:- cow--. ‘the disease of his cow’ b. mará tam awa ur-en houses:: three . big- ‘three big houses’ c. tsir/i tam ar wák ar [aa dakúus] bird even . . : miss:: ‘a bird, even one, that he missed’
. V
.................................................................................................................................. There are two types of verbs, the verbs ‘to be ‘and other verbs. Every sentence has to contain a verb ‘to be’ (also those containing other verbs), whether the complement is nominal or verbal, as in () and (). The verbs ‘to be’ are discussed in section .. below. ()
tlee/usee a Manda. potters Bantu ‘Potters are Bantu.’
() aníng a goo’-íim. write-: ‘I am writing.’
.. Verbal inflection There are three conjugation classes: verbs ending in a nasal, verbs ending in a short vowel plus w, and other types. The first class contains verbs with the durative suffix -iim; the second contains verbs with the inchoative suffix -uw. The third class is the most numerous. Verbs are inflected for person and gender features of the subject. The third person singular feminine form is always identical to the second person singular form. There are two third person plural forms for verbs of the third class, -ir and -iyá’, which do not differ in meaning. Verbs are also inflected for tense (present or past) and for mood (indicative or subjunctive). The paradigms for the three verb classes in the indicative present are given in Table .. Note that all forms end in a high tone, except the . form of verbs like doohl, i.e. those of the third class. The indicative past is characterized by a final high tone, distinctive only for the . form of verbs like doohl, and by -é’ in and instead of -á’. The past tense is also marked on the verb ‘to be’. The subjunctive present is characterized by a final low tone and a suffix -i for the . form of verbs like doohl. The past of the subjunctive is identical to the indicative past.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Indicative present in Iraqw Person
‘ask’
‘leave’
‘hoe’
a firíim
a tláw
a dóohl
a firíin
a tléer
a dóhl
.
i firín
i tláy
i doohl
i tléer
i dóhl
.
i firíin
a firiimáan a tlawáan a doohláan
a firiindá’
a tleerá’
a dohlá’
i firiná’
a tlayá’
i doohliyá’
i doohlír
The second person singular shows a number of different forms due to some historical processes involving the original alveolar stop as second person marker. The second and third person plural forms are based on their corresponding singular forms, e.g. firin-a’ (.-). The first person plural forms of some verbs have a short vowel in the stem-final syllable. This is an indication that the first person singular is marked by vowel lengthening. The simple imperative is just the verb stem with a high tone; no ‘to be’ verb is used for the positive imperatives. Other imperative forms are the plural (of addressee), the transitive forms (with an object), and the hither forms (towards me or to my benefit). When the imperative is negated, the only remaining distinction is singular versus plural addressee. Sentences with a negative imperative (or prohibitive) verb form also require a prohibitive mood morpheme on the ‘to be’ verb (see Table .).
Table . Imperative forms in Iraqw Form
Meaning
Suffix
Gloss
dóohl doohl-é’ doohl-eek doohl-aak doohl-áng doohl-aré’ huw-ang huw-are’ ma doohl-aar ma doohl-ara’
‘Dig!’ ‘Dig!’ (to many) ‘Dig it!’ ‘Dig it!’ (to many) ‘Dig for me!’ ‘Dig for me!’ (to many) ‘Bring it to me!’ ‘Bring it to me!’ (to many) ‘Do not dig!’ ‘Do not dig!’ (to many)
-é’ -eek -aak -áng -aré’ low tone low tone-aré’ -aar -ara’
. .. .. .. .. -.. .. ..
.. The verbs ‘to be’ or so-called “selectors” The complexity of the Iraqw language lies in the rich inflectional element that forms a separate word independent of the verb. Bradfield () presents his analysis as a game to
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
account for endless numbers of different surface forms. These so-called selectors are present in every sentence except for imperatives and are used with nominal and verbal complements and for that reason I see a verb ‘to be’ as the root for these inflectional forms. With nominal complements four different types can be distinguished. The first type is a copulative. The sentence is an equation. The form of ‘to be’ is a. ()
iraqw a doohlitee. Iraqw farmers ‘Iraqw are farmers.’
The second type is a locative. The complement is a locative expression. The form is a for first and second person subjects and i for third person subjects. ()
inós i bará qaymo. s/he in field ‘S/he is in the field.’
The third type is a temporal construction. The complement expresses the state into which the subject enters or has been. The form is ta. The suffix -wa in () is an aspect marker indicating simultaneous actions, here glossed as “background”. ()
án ta-wa na/ay. I - child ‘When I was a child.’
() ta-y dí-r af-kú tlawi. - place:- mouth-: lake ‘They get at the edge of a lake.’ The fourth type is adjectival in nature. The complement is an adjective. The verb ‘to be’ agrees with the subject in person and gender. ()
inós ku hhóo’. inós ka hhoo’. . nice . nice ‘He is nice.’ ‘She is nice.’
These verbs ‘to be’ are also used with a verbal complement. The main verb inflects for subject and tense, which makes the complement verbal and not nominal(ized). The copulative ‘to be’ is used in the perfect with a perfect aspect marker -(g)a suffixed to ‘to be’, (). The locative ‘to be’ is used in the present tense (), in the expectative “tense” with an aspect marker -n, and in the imperfective past with an aspect marker -na suffixed to ‘to be’. The temporal ‘to be’ is used for subjunctive mood (), simultaneous aspect -wa, and consecutive aspect -ri. The adjectival ‘to be’ is used in “passive” sentences, ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() inós a-a /áay. - eat..: ‘He has eaten’ ()
inós i /ét gawá xa’ano-wa alé. . fall:. on tree- ‘She falls from the tree.’
()
ta dohl. . dig:: ‘You should dig.’
() qaymo ka doohl. field . dig ‘The field is cultivated.’ To these four different verbs ‘to be’ a number of affixes and clitics may be added to form the inflection complex. These affixes are presented in Table .. It is impossible to cover the range of different forms of the selectors. I present just a few cases to illustrate the complexity and the richness of the system. In sentence () the first selector is una, in which the object pronoun u refers to the masculine object that precedes it. If the object follows the selector, no object pronoun is used. The shape also reveals that the subject is first or second person, because a third person subject would have required a third person subject prefix. The object pronoun replaces the verb ‘to be’ or merges with it (au ! u) depending on one’s analysis. The form is suffixed with a past tense marker -na. The second selector in the sentence is nguri with the consecutive marker ri which is typical for such a sequence of events; the initial velar nasal shows that the selector is from the dependent series which is required by the consecutive marker; u refers again to the stone that is used as instrument; the fact that the object is an instrument is indicated by the
Table . Affixes to selectors in Iraqw -
-
-
-
-ar -as -i
-a
bartammm-
-n -na -(g)a -ri -wa
bar conditional tam concessive m questioning m prohibitive
n expectational r instrumental a past s reason na past i directional (g)a perfect ri consecutive wa background
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
case clitic r, which is positioned immediately before the verb tsaxaar and which cliticizes to the left onto dayshimo ‘snake’ (and which happens not to be the complement of the instrumental clitic). The base of the third selector, iri, is i, the most common selector indicating third person subject. This i is the same form for main and dependent clause selectors but this would be different for other subjects. () tla/ano u-na tatáahh; ng-u-ri dayshimo-r stone .- take: .-.- snake- tsaxáar; i-ri gwáa’. hit - die ‘I took a stone and hit the snake with it; it died.’ Sentence () has niwa as first selector. The suffix wa is used to mark the clause as background to the next clause. This suffix requires the dependent base, which is ni for first person (ta for second person). This ni is homophonous with the hither selector ni, and because the verb hardah ‘arrive’ tends to take the hither form, I have decided to gloss it as ‘hither’. The second selector is i for third person subject. There is an object in this clause but it follows the selector for backgrounding: the activity is essential, not the object, the letters, which are an automatic result of the activity. ()
aníng ni-wa hardáh /isá, inós i barwadá go’-ín. - arrive yesterday letters:: write-: ‘When I arrived yesterday, he was writing two letters.’
In sentence (), the feminine object indicated by a is the following clause and the prefix gindicates that there is a third person subject. The dependent clause has the dependent selector, ni for first person and is marked for background by -wa. () baabá g-a hlaa’ ni-wa doohl. father -. want .- cultivate ‘Father wants me to cultivate.’ The dependent selector can be used in main clauses, as in () for optative or prohibitive clauses. This clause is of the latter type, as is evident from the prohibitive mood prefix m-; the necessary vowel that follows this prefix is a copy of the following vowel from the first person plural object pronoun referring to atén, which could have been left out without changing the meaning. Grammatically, the subject of () could be the first person singular, as this would render the same verb form and the same selector form but semantically this interpretation would be odd without a proper context. () atén mi-ti-wa múux. -.- beat ‘He is not to beat us.’ The other mood prefixes are illustrated in the next examples. The conditional bar- is shown in (), where the vowel is assimilated to a following vowel i for third person subject but this
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
vowel is subsequently deleted by the reduction rule affecting three subsequent CV syllables. The concessive tam- is illustrated in (), and the questioning prefix m- in (); the selector with the questioning prefix m- asking for ‘what?’ does not contain an object pronoun. () ino’ín bir-na hamtlin’i, a tlaw-aan. they :- bath:: / leave- ‘If they have bathed, we will go.’ () atén tam-ta-wa gilaaliim-áan, we -- fight:- ta baló wakaawak-aan-ii-ká. ever hate:--:- ‘We may be fighting, we never hated each other.’ () laarí m-a /ay-áan? today -/ eat- ‘What are we eating today?’
.. Verbal derivation The verbal derivational morphemes include a durative suffix -m, reduplication of the penultimate stem consonant for imperfective, reduplication of the initial root consonant for frequentative, middle suffix -t, causative suffix -s, and inchoative suffix -w. The durative and middle morphemes exclude each other. The inchoative suffix can combine with reduplication but not with a durative suffix. Only one of the two types of reduplication is possible at a time. The causative suffix has to occur at the end of the verb; this also holds in the case of a frozen causative. The vowel preceding the final derivational suffix, whether -m, -t, or -s, is ii. This vowel ii assimilates to a or u in the preceding syllable if the intervening consonant is back, i.e. glottal, pharyngeal, uvular, or velar (van der Hulst and Mous ); this same vowel becomes uu after a rounded consonant (which becomes unrounded), e.g. tsaakuum /tsaakw-iim /scoop-/.Verbalizing suffixes include the causative, durative, stative, and inchoative derivation but preceded by uu or ee rather than ii. () wa/aari i aníng tiq-m-iis-iyá’. vomiting. be:ill--- ‘Vomiting is making me ill.’ () ur boo/ gawid niina
‘big’ ‘black’ ‘difficult’ ‘small’
uraw boo/aw gawduw niinawuw
The suffix -m expresses duration (). Verbs with the middle suffix -t are often used for verbs which have a subject that does not control the action, and for verbs that cannot have a
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . The first imperfective in Iraqw Stem
st imperfect
Meaning
Process
al/aay /akuut aahh ya/aam goo’iim waa/ huu’ dah
al/agiim /akwmiit aahhiit ya/a/aam goo’ii’iim wa/aamiim huuhu’uum darah
‘deceive’ ‘jump’ ‘hate’ ‘agree’ ‘write’ ‘vomit’ ‘fall’ ‘enter’
durative -m infixed middle -t reduplication of penultimate reduplication and ii -aamiim durative -m and reduplication infixation of
person as subject, but the body itself, e.g. yuqumiit ‘to be cold’ (subject is the body, not the person); see Mous and Qorro (). The meaning of the suffix can also be durative, habitual, or stative, e.g. siiqiit ‘to be cutting’; the suffix has an overlapping meaning with the durative suffix -m. The expression of imperfective aspect through derivational markers is interesting. The interpretation depends on which derivations are available for a given verb lexeme. I distinguish a first and second imperfective: The first derivation is durative in interpretation; the second is habitual. Thus, it is not the actual form of the morpheme that determines durative or habitual, but whether the form constitutes a first or a second derivation. A second derivation is always built on an existing first derivation and consists of reduplication of the penultimate stem consonant. If there is no second derivation, the first derivation covers the entire domain of imperfective aspect. The various ways to arrive at a first imperfective derivation are presented in Table .. Nouns can be incorporated into the verb, as in () in section .., and from this construction noun-verb compounds arise. There is a small set of semi-productive preverb compound elements. The most common is al ‘together’: () atén a al-doohl-áan. / together-cultivate- ‘We cultivate together.’
.. Adverbials ... Verbal adverbs Iraqw has a closed set of verbal adverbs that occur in the preverbal position and modify the verb. These are adverbs like ada ‘quickly’, lak ‘almost’, mal? ‘again’, the common emphasis marker qoo (see Elders and Mous ), and the evidential adverb hhoo (Mous and Qorro ). Comparison is expressed using the verbal adverb ak, (). A verbal adverb like lówa ‘very’ is used to modify an adjective, in which case it requires the construct case just like the head noun, ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() ayto’o-r-í ka ak hhó’ ta ak maize-- .::. more nice: than more dír to-qa’-ee. place:- .-- ‘This maize is better than that maize there.’ ()
i-na amó-r lo’wá-r saaw káy. - place:- very:- far go:. ‘He went very far.’
... Case clitics The position of case clitics is immediately before the verb: -i directional, (), -wa ablative, -ar instrumental, and -as reason. The clitics attach to the host preceding them; see also the discussion of example () above. If the host is a noun, there is no gender linker preceding the case clitic, except for feminine nouns. Moving the case-marked constituent outside the verbal complex requires the addition of a resumptive verbal element alé, (). () tsir/i i gawá daandú awi iwiiwíit. bird top:of back:of bull: sit::. ‘The bird is on top of the bull.’ ()
i-na basi-r-ar daqay-ká tluwo-sa alé. - bus-- leave:.- rain:- ‘He didn’t go by bus because of the rain.’
... Sentential adverbs and adverbials The sentential adverbs do not occur in the position of verbal adverbs and they do not modify the verb but the clause instead. Examples are áangw ‘long ago, already, before’, (), wané ‘maybe (beyond control of the subject)’, and /eesi ‘always’. Often the distinction between noun and adverb is difficult to make. The adverbial hám, hamí, or hamtá ‘now’ can be analyzed as a noun ham () ‘moment’ followed by a demonstrative suffix. Some adverbs have their origin in fixed expressions: anga/áw, ‘maybe (in control of subject)’ from: an- ga/áw (/- see:) lit. ‘I see’ or ‘I suppose’ (). () atén áangw a-ga di-r-qá’ hoot-áan. before /- place-- live- ‘Before, we lived there.’ ()
atén anga/aw a may.be / ‘Maybe we will leave.’
tlaw-áan. leave- (We will see, we haven’t decided yet.)
Not all expressions of time are adverbs. Many are nouns: nouns with a background suffix, a demonstrative suffix, or a directive suffix, e.g. laa-r-í (day--) ‘today’. Most of the following day-relators are complex: /isá () ‘yesterday’, lat-’áangw ‘day before yesterday’ (lit. day–before),
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
bal-’áangw ‘second day before yesterday’ (lit. second day–before), matlo () ‘tomorrow’, baloqá (day-) ‘day after tomorrow’, baldane ‘second day after tomorrow’, tudane ‘third day after tomorrow’, taqane ‘fourth day after tomorrow’, and tudadane ‘fifth day after tomorrow’. Adverbs can be used to link a clause to a preceding clause: see asma ‘because’ in (). () kúung-a gurham-t-a asma ta-wa ..-/ regret:.:- because /- place: dí-r do’-óg-i ki/ place:- house-.- return:: ‘Are you unhappy because you are going home?’
... Interjections Interjections are used in direct speech in addressing people only. Some interjections are xáygan, which is used to call the attention of somebody, ar’ee, expressing pity and sympathy, and are, when offering somebody something.
... Ideophones Ideophones are often used as complement of verbs ‘to say’. Verbs can be formed on the basis of ideophones by adding one of the verbal derivational suffixes: () ta-na óo’: xáhl; ta-na xahlmíit. - say:. xahl - keep:silent::. ‘They said: xahl (sound of silence); they kept silent.’
. S
..................................................................................................................................
.. Constituent order and basic syntactic structures of simple sentences Every sentence needs an inflectional element that is a word category on its own, the verb ‘to be’, termed “selector” by Whiteley (). This phenomenon is not unusual among Cushitic languages. The lexical base of these -words can be characterized as verbs ‘to be’ since the various bases occur with nominal complements, as well as in verbal sentences: see section .. above. The core of an Iraqw sentence consists of a verbal complex which is a brace with ‘to be’ as the left edge and the verb as the right edge. Within this brace the following phrases occur in any combination and in any order: the object (O), an adverbial case clitic with its complement (oblique object or OO), or a verbal adverb (Adv), as is characterized in () with an example in (). () verbal complex: Infl O/Adv/OO Verb
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() inós i hhartá hhawati malé hanmiis. s/he : stick:: man: again give:::he O OO Adv Verb ‘He is giving a stick to the man again.’ Subject and tense are obligatorily marked on the verb. The subject is also marked on ‘to be’ but here the distinction is only between speech participants (first and second person) versus third person. Tense is marked both on the verb and on ‘to be’. Tense marking on the verb is marginal but obligatory. Objects are often moved leftwards. The semantic motivation of the leftward movement of the object is foregrounding. This movement is very common in past or perfect tenses (), and rare in present tense forms, in line with cross-linguistic focus patterns. Present tense is not marked and is actually non-past; future or intention is the unmarked interpretation for non-past sentences with a foregrounded object, as () shows. Note that the object inside the verbal complex requires construct case. Sentences that focus on simultaneous action do not show object foregrounding, (). Objects of negative sentences tend to be foregrounded, (). Inherent objects, (), and objects in general statements are not foregrounded, (). The object is preferably external and sentence-initial in sentences with impersonal subjects, since the object is naturally the theme if the subject is not to be mentioned, (). In () it is more appropriate for the object to be external because otherwise the meaning would be that work in general is finished once and for all, which is not intended. () baynu g-i-na /aymís. pigs .-.- feed: ‘She fed the pigs.’ () aníng kasíis a huuríim. potatoes . cook: ‘I’ll cook potatoes.’ aníng a kasíir huuríim. / potatoes:: cook: ‘I cook potatoes.’ () kar ís i hikwá dáaf, dasi i-n /akmamít. well cattle: return girl - run: ‘While he returns the cattle, the girl is running.’ () dóo úr g-w-a tleehh-ii-ká. house: big: .-.- build-:- ‘He didn’t build a big house.’ () a fa/ár huuriind-a-ká. / porridge:: cook:-- ‘You are not going to cook porridge.’ fa/a a huuriind-a-ká. porridge . cook:.-- ‘You don’t cook porridge.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() a kahawú wáh. kahawa u wáh. / coffee: drink:. coffee . drink: ‘I drink coffee.’ ‘I drink coffee.’ () yaamu k-i-na piimúus. land: .:-.- measure: ‘The land was measured.’ () gadyéet ga faak. work .:. finish:: ‘He finishes the work.’ Complements of adverbial case clitics are fronted for foregrounding purposes and just as with objects such a movement requires an object pronoun in the selector. The case clitic itself is not fronted and stays behind in the verbal complex, cliticizing to its leftmost element even if this is the object of the verb rather than to the complement of the case clitic; see the discussion of () above. Adverbs and oblique objects (i.e. noun phrases with an adverbial case clitic) can be moved to the right of the verb, in which case the moved constituent must be followed by the resumptive pronoun alé. The adverbial clitic moves along with its complement in this rightward displacement; see tluwosa alé ‘because of rain’ in () above. Bare noun objects (no construct case) can incorporate into the verb to form a noun-verb compound, (). () anága hee-gaas. :/- man-kill ‘I committed manslaughter.’
.. Negative clauses and questions A clause is negated by adding a nominalizer (glossed initive) and a negative marker ka to the verb, which is mostly the final word in the sentence. The infinitive marker is ii for third person masculine and for past tense, (), otherwise aa, (). Polar questions are formed by sentence-final question intonation and the same infinitive marker on the verb. The question intonation entails an extra-high tone on the penultimate syllable and a subsequent fall. Yes/no questions are often used to avoid a negative statement: yes/no questions are often leading questions to ‘no’, as () illustrates. () a axâs-a? / listen::- ‘Do you listen?’
(Expected answer: ‘no’)
Content questions can be formed by adding the questioning prefix to the selector (example () above), or by using a question word. No question intonation is used. A common strategy for question formation is according to the model “the man who is talking to you is who?” (); that is to say a copula clause consisting of a general noun (hee ‘man’) and a
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
question word built on that general noun and a question suffix (hee-má with a relative clause added to the initial general word. One can leave out the second part, copula plus question word, and the incomplete sentence consisting of a general head noun plus relative clause is still interpreted as a question. Questions can also be formed by a nominal clause consisting of a noun plus copula plus question word as in (). But one can also stop at the end of the relative clause, in which case one ends up with the questioning element sentenceinitially and an incomplete sentence marking it as a question, (). () do’ók a gaalá? house-your which ‘Which is your house?’ ()
hée kúung u axwées (a heemá)? man: you. . say. who ‘Who is talking to you?’
.. Complex sentences Complement clauses are arguments and can trigger the use of an object pronoun, as in () above. Complement clauses can also consist of a head noun like adoo ‘manner’ plus a relative clause translation like “how+clause”. The Datooga loan gidabá ‘that’ is used as a complementizer, as () illustrates. () inós g-a alk/íit gidabá i tláy. -. tell:: that leave: ‘He said that he is leaving.’ A common way to link clauses as consecutive action is by use of the consecutive suffix to the selector of the second clause, as in () above, or by marking the first clause as backgrounded with the background suffix added to the selector, (). Consecutive clauses are often linked by the general preposition/conjunction nee ‘and, with’. The conceptual link between clauses can be indicated by adverbs such as asma ‘because’ (see () above) and bar ‘if ’.
A I would like to thank Khalid Mourigh for extensive comments and suggestions for more clarity.
R Berger, P., and Kießling, R. (eds.) (). Iraqw Texts. Cologne: Rüdiger Köppe. Bradfield, M. (). ‘ “It” in Iraqw: an analysis of the object-selector series of the Iraqw verb’, African Languages/Langues africaines : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Elders, S., and Mous, M. (). ‘The adverbial qo in Iraqw’, in U. Claudi and D. Mendel (eds.), Ägypten im afro-orientalischen Kontext: Aufsätze zur Archäologie, Geschichte und Sprache eines unbegrenzten Raumes. Cologne: Universität zu Köln, –. Maghway, J. B. (). ‘Iraqw vocabulary’, Afrikanistische Arbeitspapiere : –. Mous, M. (). A Grammar of Iraqw. Hamburg: Helmut Buske. Mous, M. (). ‘Number as exponent of gender in Cushitic’, in Z. Frajzyngier and E. Shay (eds.), Interaction of Morphology and Syntax: Case studies in Afroasiatic. Amsterdam and Philadelphia, PA: John Benjamins, –. Mous, M., and Qorro, M. (). ‘Middle voice in Iraqw’, in K. Kahigi, Y. Kihore, and M. Mous (eds.), Lugha za Tanzania / Languages of Tanzania: Studies dedicated to the memory of Prof. Clement Maganga. Leiden: Centre of Non-Western Studies, –. Mous, M., and Qorro, M. (). ‘The evidential adverb hhóo in Iraqw’, in G. Takács (ed.), SemitoHamitic Festschrift for A. B. Dolgopolsky and H. Jungraithmayr. Berlin: Dietrich Reimer, –. Mous, M., Qorro, M., and Kießling, R. (). An Iraqw–English Dictionary. Cologne: Rüdiger Kölpe. Nordbustad, F. (). Iraqw Grammar: An analytical study of the Iraqw language. Berlin: Dietrich Reimer. van der Hulst, H., and Mous, M. (). ‘Transparent consonants’, in R. Bok-Bennema and R. van Hout (eds.), Linguistics in the Netherlands. Amsterdam: John Benjamins, –. Whiteley, W. H. (). A Short Description of Item Categories in Iraqw (with material on Gorowa, Alagwa and Burunge). Kampala: East African Institute of Social Research.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. W is a Central Chadic language spoken by more than , speakers in the far North Province of Cameroon and by a similar number in northeastern Nigeria. The main interest of this language is the existence of system-wide two inflectional markers, vowel reduction and the addition of the suffix a, which are not word-formation means but rather sentence-formation means. While neither of the processes indicates any specific semantic relation, they both lead to inferences about a large number of semantic relations within the clause and within the sentence. The vowel e at the end of the morphemes codes the clause- or sentence-final position, or a pause in all other positions (Frajzyngier ). The coding of grammatical relations subject and object is distributed over a wide range of morphemes. With respect to these relations, the language is neither head-marking nor dependent-marking. For further details, see Frajzyngier ().
. P
.................................................................................................................................. The language has a rich consonantal inventory (Table .) with three underlying vowels, as shown in Table .. The phonetic vowels (broad transcription) include those shown in Table .. There are significant differences in the phonotactics of consonants and vowels in the underlying structures and in phonetic realizations. Although a number of morphemes underlyingly end in a consonant or consist of a consonant (and tone) only, no word other than ideophones and a few very recent borrowings in the surface structure can end in a consonant or a glide in isolation. That suggests that some of the final vowels in clause- or sentence-final position are epenthetic. Word- and morpheme-final vowels constitute a major coding means for the syntactic system of Wandala. Only two vowels may occur in the word-final pre-pausal position: a, for
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Underlying consonants and glides in Wandala Stops
Voiceless Voiced Prenasalized Glottalized
Lab
Alv
Pal
Pal-Vel
Vel
Lab-Vel
p b mb ɓ
t d nd ɗ
py dy
ky gy
kp
y’
ɠy
k g ŋg
ts dz
c j
s z n r l
ʃ
Affricates Continuants
Voiceless Voiced
Nasals Liquids
f v m
ɠw
h ž ɲ ry
ŋ
ɬ ɮ
Lateral continuants Voiceless Voiced Glides
Glottal
w
y
Table . Underlying vowels in Wandala i
u a
Table . Phonetic vowels in Wandala i e
ə
u o a
the majority of lexical items, and e. The vowel e in the clause-internal position signals a pause. The vowel e is epenthetic in this position and is suffixed to morphemes that underlyingly end in a consonant. In the clause-internal non-pre-pausal position, most lexical items and grammatical morphemes can end in either a consonant or the vowel a. Although the underlying forms can begin with complex consonant clusters, including geminated consonants, only some clusters are allowed in the word-initial position in the phonetic realization. The presence of the word-final vowel codes the phrase-final position. The absence of the word-final vowel codes the clause-internal position. Vowels that are grammatical morphemes by themselves rather than part of a grammatical morpheme are not deleted. Wandala has labial and palatal harmonies. The trigger of the labial harmony is a round vowel, and its target is a velar consonant. The labial harmony operates across other segments:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
nóŋwán yénjátófká pàtrònárwà. nó ŋán yá njà á ù wáfk-á patron-á-rwà. sit before face- boss-- ‘Here I am sitting in front of my boss.’
The low vowel a is raised to [e] when it precedes or follows a palatal glide or a high-front vowel in the preceding syllable: ()
à ʃɓ-ì-vé zènápà. hide-- Zenapa ‘Zenapa hid me.’
()
yà ʃɓə̀-và zènápà. hide- Zenapa ‘I hid Zenapa.’
Wandala has two tones. The functions of the tones differ in various classes of morphemes. In nouns, both tones can be part of the underlying structure. Tones are not part of the underlying structure of verbs. Tones have an important grammatical function, coding tense, mood, aspects, relationships between nouns, and semantic relationships within a proposition. Less often tones distinguish between lexical items. The rhythmic structure of the utterance does not coincide with the lexical, morphological, and syntactic divisions of the utterance. Vowel fusion, vowel insertion, and vowel deletion are accompanied by complex rules of tonal behavior. When two vowels fuse, the low tone is replaced by the high tone.
. L
.................................................................................................................................. In addition to nouns and verbs, Wandala has a relatively large number of adverbs, morphologically marked adjectives, a small class of prepositions, a small class of spatial specifiers, and a category “locative predicator”. The locative predicator is deployed in locative predication when the main verb of the clause is not inherently locative. The locative predicator does not have its own tense, aspect, and mood characteristic. Verbs differ from other lexical categories in that they can never begin with a vowel. There is a class of nouns characterized by the reduplicated second syllable; the class comprises a large number of body parts, names of animals, and insects.
. M
.................................................................................................................................. Affixation in Wandala includes suffixes, very few prefixes, and one infix. Prefixes have been observed only in the nominal system, and in most cases are not productive.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Infixation is limited to just one morpheme a, potentially an Afro-Asiatic retention (Greenberg ) coding the plurality of the verb. Suffixes occur in all lexical categories. They are added to nouns or verbs before these lexical items are inserted into a higher formal structures, e.g. a phrase or a clause. Added to numerals and adjectives, suffixes derive nouns. There are different types and different functions of reduplication across lexical categories and even with respect to the same lexical category. Reduplication may also have phrases in its scope. The reduplication of the interrogative phrase jìbà rà ‘thing Q’ yields a nominal category jìbà-r-jìbà-rà ‘all kinds of things, what not’. Verbs have two types of complete reduplication, each type coding a different aspect. Reduplication is also a means to derive adverbs from other lexical categories. Gemination of consonants is a coding means for verbs only. In the imperfective aspect, it marks transitive verbs when there is no overt nominal or pronominal object in the clause. It also codes the imperative mood. Nominal inflectional morphology is limited to the plural suffix added to some nouns only and the genitive marker coding one type of relationship between two nouns. Verbal inflectional morphology, i.e. reduplication, infixation, suffixation, and tone code aspect (reduplication, tone); mood (tone); number (infixation and reduplication); subject; direct and indirect object; goal; ventive; point of view; spatial configuration of the event; and co-participation. The language has no passive, i.e. a form indicating that the subject of a transitive verb is not the expected controller but rather the patient. In addition to pronominal subjects that precede the verb in the affirmative and follow the verb in the negative mood, verbs can have subject suffixes that mark aspectual and modal functions. Inflectional changes on subject pronouns that precede the verb code tense. Auxiliary verbs and prepositions code aspectual systems, modality, altrilocality, and sequential clauses. All lexical and morphological categories can undergo inflectional changes indicating the role of the following constituent. This is illustrated by the behavior of question words m ‘where’, w ‘what’, and wàr ‘who’. In the phrase-final form they have the epenthetic vowel e, i.e. mè, wè, and wàrè. The consonantal ending indicates that the following form is the object. The form with the vowel a, i.e. mà, wà, and wàrà, indicates that the ensuing form is the subject.
. T
.................................................................................................................................. Nominal and adjectival modifiers, determiners, and quantifiers follow the head. Some adjectives, however, can precede the noun and be linked to the following noun by the pertensive marker. In the domain of modification by nouns or pronouns, Wandala makes a distinction between two subdomains: the modification of inherently relational nouns and the modification of all other categories including other types of nouns. The modification of non-relational categories has the form X-á[PERT]-PRO/N. The modification of relational nouns has the form N[ROOT]-PRO for pronominal modifiers, and N-á-PRO gə̀ N for nominal modifiers.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. T
.................................................................................................................................. Clauses in which no noun phrase is topicalized, focused, or represents switch reference for the subject have most often one noun phrase, which occurs after the verb. Subject pronouns precede the verb (except in the backgrounding aspect) even if there is a nominal subject in the clause. The following is a schematic summary of the pragmatically neutral clause: PRO (nàbà) V (Q/NEG) (NP) (NP) (PP) or V-PRO (Q/NEG) (NP) (PP). The gloss PRO stands for the obligatory pronominal subject. The terms Q and NEG refer to a morpheme from a closed set that includes the negative marker and one of the question words. There is no case marking. All prepositional phrases begin a new phonological phrase, as evidenced by the presence of the vowel e on the preceding constituent.
. G
.................................................................................................................................. The consonantal ending on the constituent X, where X could be any category other than a verb, indicates that the following noun phrase is the object. The vowel à ending on the constituent X indicates that the following noun phrase is the subject. The coding of the grammatical roles of the noun phrases that occurs after the verb depends on the type of predication and on the aspect. Either nominal objects or nominal subjects occur after the verb, but they do not have to follow the verb. The fronting of a constituent codes topicalization, focus, or switch reference. The formal means of coding of the grammatical roles of noun phrases is achieved through the alternation between the root, root + a, and root + á forms that precede nominal arguments. The root + a form of these morphemes indicates that the following noun phrase is the subject. The root form, i.e. the form without the final vowel, indicates that the following noun phrase is the object. When the noun directly follows the verb, three types of predication are marked. First, the object-directional predication indicates that the subject is directing the action at another participant, the object. Among verbs that are inherently object-directional are ɓə́ɠyà ‘find’, ɓàlà ‘announce’, jà ‘hit’, cìɓà ‘kill ()’, ícà ‘cut’, xɗà ‘bury’, wà ‘bite’, fà ‘put’, pwà ‘put’, tàtàyà ‘search’, tàpà ‘taste’, and ndrà ‘build’. The object role of the NP following the verb is marked by the vowel deletion on the preceding verb: ()
tà ɓə́ɠyì dàdà á njá-r láŋnè á-m : find father sit:- like_that - xùɗ-á vgə̀ ŋánnà. belly- grave ‘They found the father sitting in the grave, just like that.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The subject role of the NP following the verb is marked by the suffix -á added to the verb: ()
à ɓə̀ɠy-á dàdà. : find- father ‘The father found him.’
The non-affected subject non-affected object predication does not indicate any effects of the event on either the subject or the object. The predicates inherently coding this predication are intransitive verbs of movement, verbs of perception, and the following verbs: pàk ‘wait’, ɬyà ‘happen’, ksà ‘grab’, màgà ‘make’, šà ‘reach’, ɮè ‘finish’, vyà ‘forget’, tàtàyà ‘search for’, tə̀gà ‘share, hold’, and kàtà ‘want’. The verb ends in the vowel à, with low tone when it is followed by a subject of intransitive or by an object of a transitive verb. ()
Intransitive verb à nàbà ɮàlà páll-á-tàrè. : then depart:.. -- ‘And then, one of them left.’
()
Object of the transitive verb tà šà hàyè. : reach river ‘They reached the river.’
To indicate that the noun phrase following the verb is the subject, and that there is an object in the proposition, the verb must have an object pronoun and the goal marker: ()
à tàttàyà-n-á hàlímà. look_for-- Halima ‘Halima looked for him.’
Affected subject predication indicates that the subject of the clause has undergone a change of form or posture (but not place). Verbs that inherently mark this predication include tsè ‘rise, stop’, kyà ‘break (a round object), break into several pieces, disperse’, the corresponding plural verb kyàɓà ‘break into many pieces’, mbè ‘heal’, and tsàkwà ‘settle (about water)’. The vowel à at the end of the verb indicates that the noun phrase that follows is the affected subject: ()
à tsà nábbà. : stop Nabba ‘Nabba stopped.’
To indicate that the noun phrase that follows the verb is a controlling (rather than affected) subject of an event that includes an object, the third-person singular object pronoun n with the goal marker á must be added to the verb:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() à tsà-n-á nábbà. : stop-:- Nabba ‘Nabba stopped him.’ To indicate that the noun phrase that follows the verb is an object under control the goal marker á must be added to the verb: ()
à ts-á nábbà. : stop- Nabba ‘He stopped Nabba.’
. S
.................................................................................................................................. Semantic relations coded in the clause include the indirect object, the affected subject, the co-participant in the event, the associative argument, the source, and the target for predicates involving movement. The coding of indirect objects depends on the inherent properties of verbs. Verbs that take an indirect object as their inherent characteristic, such as the verb ‘give’ and some verbs of saying, add an object pronoun. This pronoun indicates the indirect object. Verbs that do not inherently take indirect objects must have an object pronoun followed by the third person singular object pronoun. The third person singular object pronoun codes the indirect object function of the preceding object pronoun. The affectedness of the subject or another argument is coded by the point of view of subject extension v. An event performed for the benefit of the subject or another argument is coded by the ventive extension -w´, realized as ú in the inter-consonantal position. The associative relation is coded by the extension án, identical with the associative preposition. The spatial relations for locative predications and the semantic broadening of erstwhile spatial relations are coded by the source extension s, the target extension t, the inner-space extension m, and the ‘on’ extension -ar.
. N
.................................................................................................................................. The verbal system codes number in two ways. The plural number of addressees in the imperative is coded through the suffix -wa. The infix a codes plurality of the event, which in the case of transitive verbs may imply plurality of the object. There is one plural marker in the nominal system, a or ahà. This marker is used mainly with human nouns. It does not have to be used if the noun is modified by a numeral larger than one.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. L
.................................................................................................................................. The domain of locative predication comprises the following functions: directional predication, stative predication, point of view, spatial orientation with respect to the locative complement, manner, and altrilocality. The forms of locative predications in all functions depend on the inherent properties of the predicates and the complements. The juxtaposition of a [+locative] verb and a toponym referring to a Wandala town or a noun from a set limited to ‘home’, ‘compound’, ‘town’, is the coding means for locative predication. Locative predication with other nouns, including other toponyms, requires the use of prepositions. If the predicate is not inherently locative, the locative predicator á must precede the complement. If the predicate is locative but separated from the complement by other material, such as the subject or the object, the predicator á must also precede the complement. Spatial specifiers that situate the subject or object with respect to the locative complement follow the preposition. Directionality of the event is coded by auxiliary verbs dà ‘depart’ and sà ‘come from’, which follow the main verb. The markers of directionality may occur even if the main verbs are dà ‘depart’ and sà ‘come from’. Verbs coding the manner of movement precede verbs coding directionality. Altrilocality indicates that the event happens at a place different from the place at which subject is or will be. Altrilocality can be coded either from the point of view of the preceding place or from the point of view of the subsequent place. Altrilocality coding differs from the coding of directionality in that the markers of altrilocality, derived from the verbs dà ‘depart’ and sà ‘come from’, are first in the construction, and the other verbs are considered complements of the verbs coding altrilocality.
. L
.................................................................................................................................. Wandala has grammaticalized verbal extensions coding the movement of the subject or object. The ventive codes movement toward the speaker, the place of speech, and the point of view of the subject. The inner-space extension m indicates movement of the subject or object into an enclosed space. The target extension t indicates movement toward a target. The source extension s indicates movement from the source. The extensions add directionality to the verbs of movement and the features movement and directionality to the verbs that do not code movement. All extensions have acquired other, non-locative functions.
. A
.................................................................................................................................. Adjuncts can be coded by: () inherent lexical adverbs, () adverbs derived through the reduplication of any lexical item,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() the destinative preposition gə̀, or () complexes consisting of the predicator á followed by preposition t or m. Adverbs can occur in the clause-initial or clause-final position. Equal comparison is coded by the preposition skè, which precedes the target of comparison. Unequal comparison is coded through the verb jà ‘surpass’. The target of predication can be marked by the preposition ár ‘on’, by the destinative preposition gə̀ ‘to’, or by the associative preposition án.
. V
.................................................................................................................................. Verbless predications serve a variety of functions whose only common characteristic is the absence of the verb. Locative verbless predication must have the locative predicator á in the structure S á (PREP) NP. Equational predication has the form NP NP, with the first NP being the subject. Attributive predication has property concepts preceding the subject. Property concepts have the root + a form preceding the subject even if their citation forms end in the vowel e. Affirmative existential predication has the predicate ánkwà preceding or following the element whose existence is asserted. Negative existential predication is coded by the form ɓákà following the element whose existence is denied. Possessive predication has the form NP1 án NP2, where NP1 represents the possessor and NP2 the possessum.
. A
.................................................................................................................................. There are different aspectual systems in affirmative and negative clauses. In the affirmative clause, the aspectual system distinguishes the perfective; an imperfective formed with the verbal noun; an imperfective formed through the subject pronoun followed by the reduplicated verbal noun; the backgrounding aspect, which has the form R1PROR2 (R first reduplicant, R the second reduplicant); the punctual aspect, formed with the suffix -hè; the inceptive aspect, formed with the auxiliary tsè; and the stative aspect, formed with the associative preposition án. Aspects carry different discourse functions. The backgrounding aspect provides the background for the interpretation of the following clauses within the same sentence or the comment on the preceding clauses within the same sentence. The punctual aspect provides the discourse background for the following narrative.
. T
.................................................................................................................................. The tense system of Wandala includes the unmarked form coded by the subject pronouns with the low tone. This form can have general past or present time reference. This form is in
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
contrast with the specific present, coded by the high tone on the subject pronoun. The future tense is coded by the auxiliary d, derived from the verb ‘go to’, or by the verb kàtà ‘want’. The specific past time reference is coded by the high tone on the final syllable of bisyllabic and on monosyllabic verbs. The subject pronoun has the low tone in the specific past. Trisyllabic verbs code the specific past by the high tone on the first syllable.
. M
.................................................................................................................................. The unmarked clause in Wandala indicates what the speaker would like the listener to believe. All other moods must be overtly marked. The hypothetical mood and hedging on the truth can be coded by a variety of means, including the marker má in the clause-initial position and the expression bà kà ‘you say’. The deontic modality has the subdomain of orders to the addressee(s) (imperative), coded through the gemination of the initial consonant, and subjunctive, i.e. wishes with respect to all persons, including the second person. The subjunctive is coded through the high tone subject pronoun followed by the imperative form of the verb.
. N
.................................................................................................................................. The negative marker k (kà in the phrase-final position) follows the verb and precedes the nominal subject or object. The distinction between the two arguments is marked on the form of the negative marker. Negation can also be coded by the morpheme tà placed between the subject pronoun and the verb. This morpheme must co-occur with the negative marker k. The system of tenses and aspects has been reduced in negative clauses to the past and non-past tense. The non-past tense is coded by the form PRO V[ROOT] NEG. The past tense is coded by the form V-a-PRO NEG. Prohibitive predication has the root form of the verb with the high tone followed by the negative marker kà.
. Q
.................................................................................................................................. Polar questions can be coded by intonation or by the particle hè, which occurs after the predicate but before the arguments or adjuncts. Polar questions can also be coded by the disjoint conjunction mtú ‘or’. Polar questions can have the full range of tenses and aspects. Content questions can be marked in two ways. One is through question words that code the semantic classes human, non-human, and place. Other classes, such as reason, are derived through the use of prepositions with a content question word. Content question words, like the polar question marker hè, occur after the verb but before the nominal arguments and adjuncts. The grammatical role of the content question words is coded through the same means through which the grammatical role of the nominal arguments is coded (as described in the section .).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The other means of coding content questions is through the use of various nouns together with the content interrogative marker rà. The tense and aspectual system in content questions is reduced in that no reduplicated form of the verb can be used in content questions.
. C
.................................................................................................................................. Wandala has grammaticalized a type of dependent clause called here comment clause marked by the form wá preceding the clause. The form wá marks the following clause as being a comment on what precedes it.
. T
.................................................................................................................................. The main means of coding topicalization is the fronting of the topicalized element with a subsequent separation of the topicalized noun phrase from the comment clause. This separation may have the form of a pause, a root + a form, or of determiners that follow the topicalized element. The determiners indicate that the listener should identify the topicalized element as being previously mentioned in the discourse. The grammatical relations of the topicalized noun phrase are coded through subject and object pronouns. The grammatical relation object is coded through the third-person object pronoun n suffixed to the verb. Adjuncts are coded through the markers of adjuncts preceding the noun phrase. The oblique argument role of the topicalized noun phrase is coded through the marker of the oblique that remains in situ. The absence of the object pronoun on the verb and the absence of any markers of adjuncts indicate that the topicalized noun phrase is the subject.
. F
.................................................................................................................................. There are two focus constructions in Wandala. One uses the particle bà preceding the element in focus. The focused element, which can be any component of the utterance, including the verb remains in situ. The other focus construction, contrastive focus, uses the form kínì, which follows the element in focus. Contrastive focus on nominal subjects often involves placing the subject in the clause-initial position. Such subjects are followed by the particle kínì.
. R
.................................................................................................................................. The major subdomains within the domain of reference in Wandala are deixis, deixis with previous mention, previous mention (without deixis), deduced reference, indefiniteness,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and switch reference. The coding means involved in the coding of these subdomains are deictic markers and anaphoric markers, overt use of nouns, pronouns, and determiners, word order, and the use of an existential verb to code indefiniteness.
. P
.................................................................................................................................. Like most Chadic languages (Frajzyngier ), Wandala does not have a coordinated clausal conjunction. Instead, clauses are marked for a specific semantic relation either with a preceding or with an ensuing clause, and some markers indicate a connection with both the preceding and the following material. The specific semantic relations between clauses include temporal simultaneity, coded by asyndetic parataxis; counter-expectation, marked mainly by the borrowed marker à(m)má; exclusive clauses, coded by the marker mtú ‘or’; temporal apodosis clauses, coded by the form nábà; and temporal protasis clauses coded by the form də̀. The scope of all overt markers could be a sentence or a discourse, in the sense that all of these markers can occur in sentences consisting of a single clause.
. C
.................................................................................................................................. The coding means within the domain of complementation include complementizers, the nominalized form of the complement clause, and to a lesser extent the clausal order. The most frequent clausal order is matrix clause-complement clause. Contentinterrogative complements of the verbs of saying may precede the matrix clause. Direct speech may be preceded by complementizer gə̀ní or the complementizer də̀gìyá ‘behold’, or it may be marked by a pause preceding it. The comment-clause marker wá precedes the complement clause of a variety of verbs when some other material, such as a nominal subject, occurs between the verb and the complement clause. The complement clause preceded by the comment marker wá may be direct or indirect speech, and may have the indicative or deontic modality. Matrix coding of the subject of the complement clause marks direct perception. The absence of matrix coding codes marks indirect perception.
. C
.................................................................................................................................. Conditional and temporal clauses share the property of having both protasis and apodosis clauses. While apodosis clauses of the two types of sentences share many similarities, the protasis clauses are quite different. The conditional protasis is coded by the hypothetical marker má. The temporal protasis clauses can code the time simultaneous with the time of
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
the apodosis clause, the time before the time of the apodosis clause, and the time after the time of the apodosis clause. These different time relations are coded by the use of prepositions. The time before that of the protasis clause may also be marked through the backgrounding aspect, i.e. the form R1PROR2.
. A
.................................................................................................................................. Adjunct clauses follow the matrix clause. They are marked by the subordinators wàrà and mbàtə̀, by the predicator á, and by the destinative preposition gə̀. The use of the altrilocality constructions with the auxiliary verbs s ‘come’ and d ‘go’ implies that the clause that follows the auxiliaries is a purpose clause.
. T
.................................................................................................................................. The head always precedes the relative clause. There is a fundamental difference in Wandala between the relativization of the subject and the relativization of other grammatical relations. The relativization of the subject involves placing the nominal subject in the clause-initial position, and additional coding of the subject through the subject pronoun. The relativization of the object involves placing the nominal object in the clause-initial position and the pronominal subject before the verb and a nominal or pronominal subject after the verb. The remaining grammatical and semantic relations of the head are coded by prepositions.
R Frajzyngier, Z. (). Grammaticalization of the Complex Sentence: A case study in Chadic. Amsterdam and Philadelphia, PA: John Benjamins. Frajzyngier, Z. (). A Grammar of Wandala. Berlin and New York: Mouton de Gruyter. Frajzyngier, Z. (). ‘Inflectional markers of sentential parsing’, Lingua : –. Greenberg, J. H. (). ‘Internal a-plurals in Afroasiatic (Hamito-Semitic)’, in J. Lukas (ed.), Afrikanistische Studien. Berlin: Akademie Verlag, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. K is a member of the Southern Lwoo group within Western Nilotic, a branch of the Nilotic family within the Nilo-Saharan phylum. Kumam is spoken in central Uganda. The number of speakers is given as , in Ethnologue (Lewis et al. ). Kumam borders on Lango, a Western Nilotic language, and on Teso, an Eastern Nilotic language. With regard to grammar and lexicon, Kumam is closely related to Lango (see Hieda ). However, Kumam has a lot of lexical entries Eastern Nilotic languages. The lexical forms common in Eastern Nilotic languages are likely to have been recently borrowed from Teso, due to close contact that still continues today.
. P
..................................................................................................................................
.. Consonants The posited inventory of consonants is given in Table .. Fricatives in Kumam are observed only in borrowed words. Geminated consonants are observed in forms derived via morphological processes. For example, when nouns ending in a consonant are followed by the possessive suffix -ná ‘my’, the alveolar nasal of the suffix is assimilated to the preceding consonant and forms a geminated consonant (e.g. del ‘skin’ + -ná ! dellá ‘my skin’). The semi-vowels /w/ and /y/ may occur in syllable onsets as glides with the following vowels (e.g. tyεn ‘leg’). The semi-vowels /w/ and /y/ are included in the inventory of consonantal phonemes, because they may fill the syllable onset position (e.g. waŋ ‘eye’).
.. Vowels Kumam has ten vowel phonemes (Table .).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Consonant phonemes in Kumam Bilabial Alveolar Palatal Velar Voiceless stop p
t
c
k
Voiced stop
b
d
j
g
Fricative
(f )
(s)
ɲ
ŋ
Lateral
l
Trill
r
Nasal
m
Semi-vowel
w
n
y
Table . Vowel phonemes in Kumam [ATR]
[+ATR]
front
back
front
back
high
ɪ
ʊ
i
u
mid
ε
ɔ
e
o
low
a
ɑ
The vowels are divided into two sets with regard to vowel harmony. The [+ATR] vowels control vowel harmony. Vowel harmony is applied to affixes (e.g. bundú!kú ‘gun’ + -ná ! bundú!kú-ná ‘my gun’), but not to clitics (e.g. a= ‘I’ + tédó ‘cook’ ! a=tédó ‘I cook.’). Vowel harmony is limited to only the syllable immediately preceding a syllable containing a [+ATR] vowel, in regressive assimilation (e.g. ɔ=jwajwatɔ + í ! ɔ=jwajwɑtí ‘He hit you repeatedly’). Vowels have no distinctive opposition of length. However, distinctive long vowels occur as a result of morphophonological alternations in forms derived through morphosyntactic processes. When the transitive infinitive suffix -nɔ is attached to a verb stem, the stem vowel is lengthened in compensation for the loss of a consonant (e.g. ted- ‘cook’ + -ne ! *teddo ! teedo ‘to cook’). Phonetic vowel coalescence occurs when vowels are unstressed. When a word ending in a vowel is followed by one beginning with a vowel, the final vowel of the preceding word is basically deleted. The preserved vowel assumes the [ATR] value of the deleted vowel. The toneme associated primarily with the deleted vowel is preserved. ()
dákɔ́ ɑ-gér ! [dáká!gér] woman -fierce ‘fierce woman’
Vowel sandhi also take place at the morphophonological level. When a morpheme ending in a vowel is followed by one beginning with a vowel, the preceding vowel is deleted. ()
téd-ɔ-έrέ ! cook--
té!d-éré ‘be cooked’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Syllable canon The syllable canon can be described by the following formula: (C) (G) V (C). Glides may follow any consonant except for fricatives. Syllables constitute a tone-bearing unit (TBU).
. T
..................................................................................................................................
.. Inventory of tones Kumam is a tone language, exhibiting a low, a high, a falling, a rising, a downstep high, and a double downstep high tone which contrast at the phonetic level (Hieda a). ()
low high falling rising downstep high double downstep high
[à] [á] [â] [ǎ] [!á] [!!á]
A high and a low toneme are posited in underlying representations. The low tone is not written in the phonological representation of words in the present contribution. The binary opposition results in six tones in surface representations in Kumam, as shown in section ...
.. Tone sandhi Kumam has the tone sandhi rules “High Spread” and “Floating High Assignment”, which have the effect of altering underlying tonal representations. When a word bearing a low toneme in the leftmost position follows a word bearing a high toneme in the rightmost position, the high toneme spreads over the boundary to the following word. ()
High Spread ɑbúké waŋ ! eyelash eye
ɑbúké wâŋ
‘eyelash’
When a word bearing a floating high toneme in the rightmost position is followed by a word beginning with a low toneme, the floating high toneme is not assigned to the rightmost TBU of the preceding word but to the leftmost TBU of the following word. ()
Floating High Assignment cogó rac ! cogo râc. bone bad ‘The bone is bad.’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Downstep and double downstep When a TBU associated with a high toneme follows a floating low toneme, the TBU is pronounced with a downstep high tone. ()
Downstep LHLHL
H
a=ted-o
cam
1sg=perf:cook-tr
food
→
L H (L)H
(L)H
a=té!dó
!cám.
‘I cooked food.’
A double downstep high tone is observed in the verbal morphology. The subject clitics bear a lexical tonal pattern LH. Perfect aspect is marked with a low toneme. Every simple verb stem bears a high toneme. The transitive formative suffix -ɔ bears a low toneme. The object suffixes bear a high toneme. The rightmost TBU is pronounced with a double downstep high tone after a sequence of a floating low, a floating high and a floating low toneme.1 ()
Double downstep LHLHLH
ε=nεk-ɔ-a 3sg=perf:kill-tr-1sg
L H (L)(H) (L)H
→
ε=nέ!!k-á.2 ‘(S)he killed me.’
.. Syntactic functions of tone Tone has a syntactic function to distinguish direct from indirect speech constructions. ()
Indirect speech dákɔ́ ɔ=wac-ɔ nέ í!cʊ́ɔ bé á=!téd-ó !cám. woman /=:say- to man that =:cook- food ‘The woman said to the man that I (speaker) cooked food (habitually).’
()
Direct speech dákɔ́ ɔ=wac-ɔ nέ í!cʊ́ɔ bé a=tédó !cám. woman /=:say- to man that =:cook food ‘The woman said to the man: “I cook food.” ’
1 2
This is not the case for the switch-reference third person singular in the perfect. ɔ + a ! a.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
High spread is blocked by a suprasegmental boundary, for example between the complementizer bé ‘that’ and the following subordinate clauses in direct speech. There is also a suprasegmental boundary after topics, which blocks the application of tone sandhi.
. M
..................................................................................................................................
.. Verbs Transitive stems are constructed by attaching the transitive formative suffix -ɔ to verb stems (e.g. yεŋ-ɔ ‘satisfy’, yεŋ ‘be satisfied’). Some transitive verbs have middle forms which are derived by attaching the suffix -έrέ(L) to transitive stems (e.g. nέn-ɔ + έrέ(L) ! nέ!nέrέ(L) ‘be seen’). When a verb stem ends in a vowel, the initial vowel of the middle suffix is assimilated to the preceding vowel after the transitive formative suffix is deleted (e.g. dí-ɔ + έrέ(L) ! dí!írέ(L) ‘be pressed’). Frequentative transitive forms are derived by reduplicating an initial consonant and a core vowel of the stem (e.g. put- ‘uproot’ ! puput-o ‘uproot frequently’). Some frequentative transitive forms have an intransitive counterpart derived by adding the suffix -un to the frequentative stem (e.g. puput- ! puput-un ‘be plucked frequently’). A few transitive verbs have a neuter form derived by lengthening a stem vowel and attaching the suffix -έ (e.g. rɪd- ‘squeeze’ ! rɪɪd-έ ‘be squeezed’).3 Neuter denotes a continuous state of events and actions, but does not presuppose the existence of agents of actions or events. Peripheral intransitive forms are derived by attaching the suffixes -Vkɪn or -ar (e.g. rʊkʊm‘incubate’ ! rʊkʊm-ʊkɪn ‘be incubated’, jurus- ‘take off ’ ! jurus-ɑr ‘be taken off ’). A few intransitive forms ending in -Vkɪn pair with transitive counterparts ending in -Vkɪnɔ. Transitive infinitive forms are derived by attaching the suffix -nɔ to verb stems. The alveolar nasal of the suffix is assimilated to the preceding consonant. Stem vowels are lengthened in compensation for the loss of a consonant (e.g. ted-nɔ ! *ted-do ! teedo ‘to cook’). When stems end in a vowel, consonantal assimilation does not occur. However, stem vowels are lengthened because of rhythm adjustment (e.g. mɪ-nɔ ! mɪɪnɔ ‘to give’). The negative particle lɪka may be followed by an infinitive. However, constructions consisting of the noun derived from the verb dagɪ ‘refuse’ and an infinitive are often used as negative infinitives. tɪɪmɔ tic a-rac. () lɪka/dágí /refusal do: job -bad ‘Not doing the job is bad.’ Verbal complexes consist of subject clitics and verb stems. Tense is not marked in verbal complexes. Aspect is obligatorily marked with a suprasegmental morpheme. Perfect aspect is marked with a low toneme. Imperfect aspect is morphologically unmarked. The plural suffix -únú is used in order to distinguish plural from singular only in the second person. 3
Some neuter forms have a free variant form derived by reduplicating a consonant in stem-final position.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
Imperfect a=téd-ó !cám i=téd-ó !cám ε=téd-ó !cám
Perfect a=té!d-ó !cám i=té!d-ó !cám ε=té!d-ó !cám (non switch-reference) ɔ=ted-o cám (switch-reference) ɔ=téd-ó !cám ɔ=té!d-ó !cám i=téd-ú!nú cám i=té!d-ú!nú cám gɪ=téd-ó !cám gɪ=té!d-ó !cám4 ‘I cook food’, etc. ‘I cooked food’, etc.
Whenever there is a nominal subject, the verb is not inflected for person (or number). ()
Imperfect ɪcʊɔ́ /cʊɔ téd-ó man/men /::cook- ‘The man/men cook(s) food.’ Perfect ɪcʊɔ́ /cʊɔ ɔ=ted-o man/men /=:cook- ‘The man/men cooked food.’
!cám. food
!cám. food
The third person singular subject clitic ɔ= is used to indicate that the subject in the subordinate clause is not coreferential with the subject of the matrix (i.e. for switch reference). The subject clitic ε= indicates that the subject in the subordinate clause may or may not be the same as the subject of the matrix. The object suffixes are as follows: ()
Imperfect ε=dí!p-á ε=dí!p-í ε=dí!p-έ ε=díp-ó-!wá ε=díp-ó-!wu ε=díp-ó-!gí ‘He hits me’, etc.
Perfect ε=dí!!p-á ε=dí!!p-í ε=dí!!p-έ ε=dí!p-ó-!wá ε=dí!p-ó-!wú ε=dí!p-ó-!gí ‘He hit me’, etc.
Object suffixes are attached to verb stems before the plurality suffix -únú. () Imperfect i=nέ!k-á-!nú. =:kill-- ‘You () kill me.’
4
When a verb is preceded by the third person plural independent pronoun gɪn, the third person plural subject clitic ɪ= is attached to the verb (e.g. gín í=!tédó in the imperfect, gín í=!té!dó in the perfect).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Perfect i=nέ!!k-á-!nú. =:kill-- ‘You () killed me.’ Past tense is expressed by the adverbial particle ɔʊdɔ, which may be accompanied by a verbal complex in the imperfect or the perfect. The adverbial particle is located quite freely in sentences. ()
Past imperfect ɔʊdɔ a=téd-ó =:cook- ‘I cooked food.’ Past perfect ɔʊdɔ a=té!d-ó =:cook- ‘I had cooked food.’
!cám. food
!cám. food
Future tense is expressed by the verb yaarɔ ‘decide’ and an infinitive. The verb yaarɔ ‘decide’ is always marked with perfect aspect in the future tense. () Future a=yá!r-ɔ́ teedo cám. =:decide- cook: food ‘I will cook food.’ The irrealis is expressed by the combination of the adverbial particle ɔʊdɔ and the future tense. The irrealis denotes events or actions that may or may not be performed in the future. Every simple verb stem bears a tonal pattern LH in the imperative and the subjunctive mood. The plurality suffix -únú is attached in plural imperatives. When a verb is not followed by an object, the subjunctive suffix -ɪ is attached to the verb stem. ()
cam rí!ŋó eat: meat ‘Eat meat!’
()
cam-í eat: ‘Eat!’
5
ɪ + u ! u.
cɑm-únú ríŋ!ó eat:- meat ‘Eat () meat!’
cɑm-únú5 eat:- ‘Eat ()!’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
The first person dual is distinguished from the plural only in the subjunctive. The plurality suffix -únú is attached to stems marked for the first person plural in the subjunctive. When a verb is not followed by an object, the subjunctive suffix –ɪ is attached to the verb stem (e.g. myero a=cá!m-í ‘I should eat.’). ()
myero a=cám !rí!ŋó myero i=cám !rí!ŋó myero ε=cám !rí!ŋó myero ɔ=cám !rí!ŋó myero ɔ=cá!m-únú rí!ŋó myero i=cá!m-ú!nú rí!ŋó myero gɪ=cám !rí!ŋó ‘I should eat meat’, etc.
Kumam has no morphological device for deriving benefactive stems. When a benefactive nominal phrase follows a verb in the subjunctive, the preposition nέ ‘for’ may be attached to the preceding verb. The alveolar nasal of the preposition is assimilated to the preceding consonant (e.g. wɪl + nε > wɪllε > wɪɪlε ‘buy for’). !bé dákɔ́ wɪ̌l nέ/wɪɪlέ á!tín ɪtabʊ́. () a=mítɔ́ =:want that woman buy: for/buy:: child book ‘I want the woman to buy the book for the child.’ Gerunds are constructed from verb stems, being prefixed with a- and suffixed with -á. ()
kal-mέ ɑ-ryég-á millet- -grind- ‘millet for grinding’
A gerund may follow a verbal complex in order to express a repetition of the verbal event.
.. Nouns Although most nouns have lost number marking as an inflectional property, a few nouns still preserve distinctive singular and plural forms, which are relics of archaic morphological operations. Many borrowed words (probably from Teso) also preserve distinctive singular and plural forms. () ye dyaŋ (singulative) apʊ́s(L)
yédé(L) ‘boat’ dóké ‘cow’ ɑpusin
‘cat’
Kumam makes a morphological distinction between alienable and inalienable possession, as shown in () below.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Although many nouns have lost the original forms for inalienable possession, Kumam has developed a new contrast in order to distinguish inalienable from alienable possession. To express inalienable possession, speakers sometimes use these forms, derived by applying consonantal assimilation and compensatory lengthening (e.g. lέέbέ (inalienable), lέb-mέrέ (alienable) ‘his tongue’). The alveolar nasal of the inalienable possessive suffix -ná ‘my’ may be assimilated to the preceding consonant. Moreover, the preceding vowel may be lengthened in compensation for consonantal loss. Consequently, three variants of possessive forms coexist (e.g. ít + -ná ! ít-ná ~ ittá ~ íítá ‘my ear’).6 A few nouns preserve archaic forms for inalienable possession. ()
Inalienable wi-á wi-í wi-é wi-wá wi-wú wi-gí ‘my head’, etc.
Alienable wic-ná wic-ní wic-mέrέ wic-wá wic-wú wic-gí ‘my head’, etc.
When possessors are animate, possessed nouns and possessors in genitive constructions are linked with a low toneme in inalienable possession (e.g. dɔǵ !dákɔ́ ‘mouth of a woman’), and with the possessive particle á (L) ‘of ’ in alienable possession (e.g. ɑgúlú-á !dákɔ́ ‘pot of a woman’). When possessors are inanimate, the possessive particle mέ (L) ‘of ’ is used in genitive construction. Both of the possessive particles á (L) and mέ (L) can be used for animate possessors in alienable possession. The possessive particle á (L) is exclusively used when possessors are kinship terms or the noun ‘God’. () jámé-mé egóe things- cloth ‘things of cloth’ *jámé-á egóe things- cloth ‘things of cloth’ () *jámé-mé papa things- father ‘things of father’ jámé-á papa things- father ‘things of father’ The first component of a compound cannot be modified by a qualifier. The elements of a compound are simply juxtaposed without any morpheme linking them.
6
When a nominal stem ends in a consonant, a nasal or a glide interrupts vowel harmony.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() atín sukú!lú-ná child school-my ‘my pupil’ The suffixes -mɔrɔ () and -mɔgɔ () are used to express indefiniteness. Nouns that are not affixed by the indefinite suffixes are interpreted as definite. () yat-mɔ́rɔ́ yén-!mógó ‘a tree’ ‘some trees’ Nouns are followed by modifiers, which usually have the following order: nouns, possessives, adjectives, numerals, and demonstratives. Numerals or demonstratives can be replaced with adjectives. () del dɔ́g-á á-!tí acyέl skin. lip- -big one ‘my one big lip’ These are illustrated in section .. below.
.. Adjectives Some adjectives have distinctive forms for singular and plural. Adjectives agree with modified nouns with respect to number. In associative constructions, modified nouns and adjectives are linked with the attributive particle a(H). () atín a-rác ɪdʊέ́ ɑ-réco ‘bad child, children’ Adjectives behave like verbs in many respects. Adjectives constitute complexes with subject clitics only in imperfect aspect. Past tense is expressed by the adverbial particle ɔʊdɔ (e.g. ɔʊdɔ a=rác ‘I was bad.’). ()
a=rác i=rác ε=rác ɔ=réco i=ré!cúnú gɪ=réco ‘I am bad’, etc.
Nouns are derived from adjectives by adding a suffix -ɔ with a specific tonal pattern (e.g. lácɔ́(L) ‘width’, lac ‘wide’).7 Adjectives beginning with ɪ/i have no morphologically derived noun. The infinitive beedo ‘to stay’, when used with an adjective, expresses the nominal equivalent (e.g. beedo ɪsεdεɲεɲ ‘shallowness’). 7
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Some adjectives also occur as reduplicated forms. The reduplicated forms express reductive meanings, or are used in semantically limited domains (e.g. bεbεr ‘somehow good’, bεr ‘good’, cɔcɔl ‘black (only for color)’, cɔl ‘black, dark’). The reduplication of a full form expresses an expanded meaning (e.g. rac rac ‘really bad’, rac ‘bad’). The repetition of adverbs denotes a restriction in terms of scope. Repetitive forms are constructed from stems prefixed with ɑ- and sometimes suffixed with –ɑ (e.g. ɲoró ɑɲorɑ ‘just yesterday’, ɲoró ‘yesterday’).
.. Prepositions Prepositions take pronominal possessive suffixes in order to express inalienable possession. When a possessive suffixe is attached, some prepositions take on a particular form which is followed by a modifier. ()
pí ‘because of ’ pɪr-á pir-í pɪr-έ pɪr-wá pɪr-wú pɪr-gí ‘because of me’, etc.
The prepositions constitute genitive constructions with the following nouns in inalienable possession. Some prepositions are subject to vowel harmony. () í-ɔt ‘in the house’ pɪ dákɔ́ ‘because of the woman’ The preposition í (L) ‘in, on, at’ combines with other prepositions or nouns (e.g. í-wi ‘on’, í-!dúd ‘under’).
.. Pronouns Kumam has the following independent personal pronouns: ()
áŋɔ́ (L) ín έn wán wún gín
Independent associative pronouns are based on the possessive particle mέ (L) ‘of ’, conjugated for persons.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
possessed mέ!rá mé!rí mέ!rέ mέ!rwá mέ!rwú mέr!gí ‘mine’, etc.
possessed mέέ!gá méé!gí mέέ!gέ mέέ!gwá mέέ!gwú mέέ!ggí
Independent reflexive pronouns are based on kέn ‘alone’, conjugated for persons. ()
kέná kéní kέnέ kέnwá kέnwú kέngí ‘myself, me alone,’ etc.
.. Demonstratives The deictic affixes express three relations of spatial deixis. The demonstratives for near and far spatial deixis form contracted forms with the preceding noun in the singular. () ɔt-nɪ ɔɔt-ɪ ‘this house’ ɔt-no ɔɔt-o8 ‘that house’ ɔt-ca ‘that house far away’
ude-gi ‘these houses’ ude-go ‘those houses’ ude-kɑ ‘those houses far away’
Independent demonstrative pronouns take the following form: ()
‘this’ mán (L) mágí (L) ‘that’ mánó (L) mágó (L) ‘that, far’ mácá (L) máká (L)
.. Interrogative pronouns Interrogatives are inflected for singular as against plural. ()
8
‘who’ ŋáí ‘what’ ɲó ‘which’ mέnέ
ɑlú ɲoígo mέgέ
[-high] vowels do not cause vowel harmony in regressive assimilation.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Numerals The basic cardinal numbers are as follows: ()
acέl(L) ɑré ɑdéek ɔŋwɔ́n kaɲ
kaɲɑpé kaɲɑré kaɲɑúní(L) kaɲaŋɔ́n tɔmɔ́n
The numbers ‘’ to ‘’ are formed by combining the numerals ‘’ to ‘’ with tɔmɔ́n ‘’ followed by íwi-é ‘on top of it’. () tɔmɔn íwi-é á!cέl(L) tɔmɔn íwi-é á!ré The numbers ‘’ to ‘’ are formed by modifying the noun ɔt with the numerals ‘’ to ‘’. The numbers ‘’ to ‘’ are formed by modifying the noun tɔĺ with the numerals ‘’ to ‘’. () ɔt ɑré tɔĺ á!cέl(L)
. S
..................................................................................................................................
.. Basic order in sentences Nouns or noun phrases are not marked for case in Kumam, in contrast with a range of other Western Nilotic languages. Word order is the primary device for indicating grammatical relations in simple sentences. The unmarked order of elements is: () (topic) (S) verb (benefactive/dative) (DO) (prepositional phrases) (adverbials) Subjects, predicates, benefactive/datives, and direct objects constitute the core constituents of sentences. Other constituents are peripheral elements, which cannot intervene between core elements.
.. Intransitive, transitive, and middle sentences Transitive verbs obligatorily require two arguments in Kumam.9 Kumam does not allow transitive verbs to behave as intransitives without morphological marking.
9
Kumam has only a few ditransitive verbs.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() Intransitive a=yέŋ kede cám. =:be satisfied.with food ‘I am satisfied with food.’ () Transitive cám ɔ́=yεŋ-ɔ atín. food /=:satisfy- child ‘Food satisfied a child.’ () *a=cá!m-ɔ́ =:eat- ‘I ate.’ Middle verbs allow only one argument. Middle forms are divided into two subclasses. When the arguments that would be direct objects of transitive verbs are subjects of the corresponding middle forms, the middle forms are called Middle . When the arguments that would be subjects of transitive verbs are subjects of the corresponding middle forms, the middle forms are called Middle . The arguments that the transitive counterparts of Middle take in sentences are sometimes psychologically connected to each other. Actions that are expressed by Middle sometimes inherently include reflexive meanings. () Transitive a=ʊ́d-ɔ́ atín. =:find- child ‘I see the child.’ () Middle atín ɔ=ʊd-έrέ. child /=:find- ‘The child has been found.’ () Middle welo ɔ=mot-ere. visitor(s) /=:greet- ‘The visitors greeted each other.’
.. Predicate nominal and predicate adjectival constructions Nominal or adjectival predicate constructions consist of a subject followed by a predicative nominal or an adjective without a copula. () áŋɔ́ emúrón. doctor ‘I am a doctor.’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() atín rac. child bad ‘The child is bad.’
.. Existence, location, and possession Existential and locative constructions are formed with the verb tye ‘be’. Subjects in existential constructions are usually accompanied by the indefinite suffixes. ()
Existential ogwók-mɔ́rɔ́ tyé ɪ ɔt. dog- /::be in house ‘There is a dog in the house.’
() Locative eŋú tyé ɪ amɔ́ní. lion /::be in forest ‘The lion is in the forest.’ Possessive constructions are formed with the existential predicate tye ‘be’ and the preposition kede ‘with’. ()
a=tyé kede ɪtabʊ́. =:be with book ‘I have the book.’
.. Topicalization The topic slot is located in sentence-initial position. When a subject or object is topicalized, the subject or object slot of the corresponding sentence without topicalization is empty.10 When a noun is topicalized from a prepositional phrase, the noun slot is filled by a pronoun interpreted as coreferential with the topic. Topicalization can be applied to any noun phrase in a sentence, including noun phrases from a subordinate clause. () gwen, okélo ɔ=kwal-ɔ. chickens Okelo /=:steal- ‘The chickens, Okelo stole.’ ()
atín, dákɔ́ ɔ=tedo cám nέ. child1 woman /=:cook food for:1 ‘The child, the woman cooked for.’
() okélo, a=wí!ɲó !bé ɔ́=kwal-ɔ gwen. Okelo1 =:hear /=:steal- chickens ‘Okelo, I heard that he stole the chickens.’ 10
The slot is filled with a φ-anaphor.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Complex sentences Kumam has no relative pronouns. Relativization involves the same syntactic operation as topicalization except that the relative clause is preceded by the relative marker ɑmé (L) or a (H).11 The same syntactic phenomena are observed in relative constructions as in topicalization. When a subject or an object is relativized, the subject or object slot is empty in the relative clause. When a noun is relativized from a prepositional phrase, the noun slot is filled with a pronoun interpreted as coreferential with the antecedent. Relativization can also be applied to a noun phrase in a subordinate clause. () gwεnɔ έn á!mé ɪcʊ́ɔ ɔ=kwal-ɔ. chicken man /=:steal- ‘It is the chicken which the man stole.’ () pala έn á!mé !dákɔ́ ɔ=ŋɔl-ɔ riŋo ké!dé. knife1 woman /=:cut- meat with:1 ‘It is the knife with which the woman cut meat.’ When a time adverb is relativized, the adverbial slot is filled by the preposition kede ‘with’ accompanied by a pronoun coreferential with the adverb. () ɲoro έn á!mé a=dɔ̂k kedé. yesterday1 =:go.back with:1 ‘It was yesterday when I went back.’ Kumam has five types of complementation (Hieda b): hypotactic indicative, “paratactic” indicative, hypotactic subjunctive, “paratactic” subjunctive, and infinitive.12 Hypotactic constructions are formed with subordinate clauses preceded by the complementizer bé ‘that’. “Paratactic”constructions are not accompanied by a complementizer. The complement types are chosen depending on syntactico-semantic characteristics of main verbs. () Hypotactic indicative a=tá!m-ɔ́ !bé í!cʊ́ɔ ɔ=kwal-ɔ gwen. =:think- man /=:steal- chickens ‘I thought that the man stole the chickens.’ () “Paratactic” indicative a=nέ!n-ɔ́ ɪcʊ́ɔ ɔ=kwal-ɔ gwen. =:see- man /::steal- chickens ‘I saw the man stealing the chickens.’
11 12
The relative marker a(H) has the same form as the attributive particle. These are V-complements. Kumam has also N-complements.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() Hypotactic subjunctive a=sʊ́!pɔ́ cuɲ ɪcʊ́ɔ bé έ=kwal gwén. SG=PRF:encourage heart of:man =steal: chickens ‘I persuaded the man to steal the chickens.’ () “Paratactic” subjunctive a=yé!í ɪcʊ́ɔ kwal gwén. =:accept man /:steal: chickens ‘I agreed that the man should steal chickens.’ () Infinitive a=mí!ɔ́ ɪcʊ́ɔ kwalo gwen. =:give man steal: chickens ‘I made the man steal the chickens.’ The subordinate clause is logically dependent on the main clause in hypotactic constructions, while the assertions in a subordinate clause are independent of the main clause in “paratactic” constructions. These characteristics allow us to establish strategies for identifying the construction types (see Hieda for further details). Conditional expressions are found with subordinate clauses. () ká!mé abáka ɔ=bino, ɔ=yá!rɔ́ teedo cám. if king /=:come =:decide cook: food ‘If the king comes, we will cook food.’ Time expressions are found with relative constructions preceded by the word ká!mé ‘when’, which consists of the noun ka (H) ‘place’ and the relative marker ɑmé (L). () ká!mé a=dɔ̂k kedé, a=ʊ́!d-ɔ́ dákɔ́-!ná pacɔ. when1 =:go.back with:1 =:find wife- home ‘When I went back, I found my wife at home.’ Nouns are conjoined with the preposition kede ‘with’. () apʊ́ɔ kede eŋú ɔ́ʊdɔ tyé ɑwótín. hare with lion /::be friends ‘The hare and the lion were friends.’ The verb =kɔ́ ‘do’ is used in conjoining sentences, and is always conjugated for perfect aspect, taking an infinitive complement. A coordinating conjunction is not necessary between such sentences. () dákɔ́ ɔ=ted-o cám (dí) woman /=:cook- food (and) í!cʊɔ́ ɔ=kɔ́ caam-έ. man /=:and eat:- ‘The woman cooked food and the man ate it.’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
In other coordinating constructions, sentences are connected by conjunctions. () dákɔ́ ɔ=ted-o cám dɔ́ á!tín líká ɔ=cam-έ. woman /=- food but child /=:eat- ‘The woman cooked food but the child did not eat it.’ Topicalization and relativization cannot be applied to noun phrases in the second sentence of coordinated sentences.
. N
.................................................................................................................................. The negative particles líká and kʊ́r are used in indicatives and in subjunctives, respectively. Negative subjunctives in the second person are used for negative imperatives. A negative particle is always followed by a predicate; therefore, negative nominal phrases are expressed by relative constructions. () Indicative dákɔ́ !líká ɔ=tedo cám. woman /=:cook food ‘The woman did not cook food.’ ()
Subjunctive a=mí!ɔ́ ɪcʊ́ɔ lworo bé kʊ́r ε=kwal gwέnɔ. =:give man threat =steal: chicken ‘I warned the man not to steal the chicken.’
() Imperative kʊ́r í=!cám =eat: ‘Do not eat meat!’ Imperative kʊ́r í=!cá!m-ú!nú =eat:- ‘Do not eat () meat!’
!rí!ŋó. meat
riŋó. meat
. Q
.................................................................................................................................. Yes-no questions are distinguished from their corresponding declaratives by intonation only. Interrogative sentences have a rising intonation. Although interrogative pronouns are usually located in the same slots as the relevant noun phrases in the corresponding declaratives (i.e. in situ), they are quite freely located within sentences. When an interrogative pronoun has the grammatical role of subject, a relative construction is used for the interrogative sentence.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() okélo ɔ=kwal-ɔ ɲó ɲoro? Okelo /=:steal- what yesterday ‘What did Okelo steal yesterday?’ () ŋáí á!mé ɔ=kwal-ɔ gwen? who /=:steal chickens ‘Who stole the chickens?’
R Hieda, O. (a). ‘Tonal system in Kumam, a double downstep language’, Journal of Asian and African Studies : –. Hieda, O. (b). ‘Complementation in Kumam’, in O. Hieda (ed.), Descriptive Studies of Nilotic Morphosyntax. Tokyo: Research Institute for Languages and Cultures of Asia and Africa, –. Hieda, O. (). Kumam Vocabulary with Grammatical Notes. Tokyo: Research Institute for Languages and Cultures of Asia and Africa. Hieda, O. (). A Grammar of Kumam: The interaction between syntax and pragmatics. Tokyo: Research Institute for Languages and Cultures of Asia and Africa. Lewis, M. P., Simons, G. F., and C. D. Fennig, C. D. (). Ethnologue: Languages of the World, th edn. Dallas, TX: SIL.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
.
. I
.................................................................................................................................. B (also known as Kachipo or Zilmamu) is spoken by approximately , people in the border area between southern Ethiopia and South Sudan. Together with the DidingaMurle languages in South Sudan and Ngaalam in Ethiopia (Yigesu to appear), it forms the Southwestern branch of the Surmic family within Nilo-Saharan (Dimmendaal a: ). The speech community using Baale as a first language forms an ethnic unit known as Suri (or Surma) with the neighboring Tirma and Chai, who speak a Southeastern Surmic language and who live mainly on the Ethiopian side of the border. Southwestern and Southeastern Surmic together form the Southern branch of Surmic, one of the two primary branches of this family, with Majang constituting the Northern branch. Surmic in turn forms a genetic subgroup with Nilotic, Temeinian, the Daju group, and Jebel within the Eastern Sudanic branch of Nilo-Saharan (Dimmendaal, chapter of this volume). Whereas Baale people also speak Tirma and/or Chai (which form a dialect continuum with Murle), only few Chai or Tirma speakers appear to know Baale. The close interaction between speakers of Baale with their ethnic kin speaking Tirma or Chai apparently has resulted in a restructuring of Baale at the phonological, morphosyntactic, and pragmatic level, whereby Tirma-Chai as the intercommunity language cluster provided the model for grammatical calquing, a phenomenon which has come to be known as metatypy. Dimmendaal (: –) describes the phonological and grammatical effects of this process for Baale, and also shows how this affected the lexical structure, for example in that compounds do not have a head– modifier structure, as in the closely related Didinga-Murle languages, but rather a modifier–head order, as in the distantly related but geographically proximate Tirma-Chai language.
. P
.................................................................................................................................. Characteristically for Surmic and many other language families belonging to the Eastern Sudanic branch of Nilo-Saharan, vowels in Baale are divided into two harmony sets:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
[–ATR] ɪ, ε, a, ɔ and ʊ, and [+ATR] i, e, o, and u; these vowels can be either short or long. Vowel harmony rules in Baale follow a pattern which is more common in Surmic and neighboring Nilotic languages, known as feature control. Hence, whenever a [+ATR] vowel occurs either in a root or a suffix, [–ATR] vowels in adjacent morphemes shift towards their [+ATR] counterpart. Thus, in the following example the [+ATR] vowel in the singulative suffix causes preceding root vowels to shift to [+ATR]: ()
εmmέ ‘bones’ emmee-ní ‘bone’
Whereas the closely related Didinga-Murle languages have a classic system with ten vowels, the [+ATR] vowel *ä (*ʌ) has shifted to and merged with its [–ATR] counterpart *a in Baale. As a result, the vowel a is neutral, i.e. compatible with [+ATR] as well as [–ATR] vowels within the same word (Möller ); moreover, the vowel a does not alternate, as the following examples with roots containing [+ATR] as against [–ATR] vowels helps to show: ()
kaalέ ‘birds’ kaalé-jì ‘bird’
The Baale consonant system (Table .) is also characteristic for the Surmic family as a whole, in that it shows a contrast between voiced and voiceless stops as well as (voiced) implosives.1 This contrast has been reconstructed for the earliest stages of Southern Surmic by Yigezu (). The palatal approximant y in the present contribution represents the IPA symbol j, whereas j corresponds to the IPA symbol ɟ; vcls stands for “voiceless”, and vcd for “voiced”. Geminate consonants contrast with single consonants intervocalically in Baale. Morphophonological alternations suggest that single voiceless stops automatically become
Table . Baale consonants bilabial dental alveolar palatal velar glottal stop
t
c
k
vcd b
vcls
d
j
g
ɓ
ɗ
ʄ
ɠ
fricative vcls
s
ʃ
vcd
z
glott
nasal
m
n
lateral
l
trill
r
glide
1
t̪
h
ɲ
ŋ
y
w
Southeastern Surmic languages like Bodi (but not Tirma and Chai) also have ejectives, probably due to areal contact with neighboring Omotic (Afro-Asiatic) languages.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
voiced fricatives intervocalically, as in the following singular/plural alternation for the imperative of ‘kill’, where the voiceless velar stop in initial position automatically becomes voiced (and slightly fricativized) intervocalically: ()
kat̪ɔǃ [kāðɔ̄] ‘Kill!’ (antipassive singular) a-gat̪-t̪ɔ́ǃ [āɣāt̪t ɔ̪ ́] ‘Kill!’ (antipassive plural)
As voiced stops are also realized as voiced fricatives in this position, the voicing distinction as such is neutralized intervocalically. The voiceless dental stop in initial position alternates with [ð] intervocalically; there is, however, no independent voiced dental stop [d̪] wordinitially. Consequently, [ð], [ɽ], [ʑ], and [ɣ] are to be treated as the intervocalic allophones of the voiceless stops /t̪/, /t/, /c/, and /k/, respectively; voiced stops (/d/, /j/, and /g/) are also fricativized intervocally (i.e. they are realized as [ɽ], [ʑ], and [ɣ]), thereby resulting in a neutralization of the phonetic distinction with corresponding voiceless stops in this position. Where there is no morphophonemic alternation (as with ‘kill’), such intervocalic consonants are represented as voiced consonants in this chapter. Corresponding geminate consonants are always realized as (voiceless or voiced) stops, as in examples () and (). ()
agá [āɣá] ‘tongue’ aggá [āggá] ‘we’ (absolutive case)
There are also restrictions on the distribution of word-final consonants in Baale. There is language-internal as well as historical-comparative evidence that stops were lost historically in this position in Baale (Dimmendaal a, b). Liaison phenomena in the synchronic grammar of Baale, for example, suggest that the voiceless stops *t̪, *t, *c, and *k were lost word-finally, but retained or protected from loss before a following suffix or clitic, thereby giving rise to an intricate system of morphophonemic alternations. These liaison consonants are represented as < C> in this chapter. Compare the following example, where such a “floating consonant” is followed by an enclitic genitive linker a plus the pronominal possessive naandí: ()
kówú< t̪> kóút̪-a-naandí mεεlέ< k> mεεlέk-a-naandɪ
[kówú] [kóúðānāāndí] [mε̄ε̄lέ] [mε̄ε̄lέɣānāāndí]
‘chicken’ ‘my chicken’ ‘axe’ ‘my axe’
The liaison consonant is also realized when a suffix follows, e.g. the number suffix -Ca (whereby C represents a consonant underspecified for point of articulation): ()
kout̪-t̪á ‘chicken’ mεεlεk-ká ‘axes’
Such word-final obstruents only show up before bound morphemes (suffixes and clitics), not before following words, even if these have an initial vowel (which consequently would allow the creation of a possible syllable structure in the language). Thus, the word for ‘earth,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ground’ lo< t̪> has a final liaison consonant which shows up before the oblique case marker, but not before the predicative marker a- in the following word: ()
aí lot̪-á. [āí lōðá] :be ground- ‘It is on the ground.’
()
lo< t̪> á-iɗo [lō áīɗō] earth -cloud ‘The country is cloudy.’
The absence of word-final stops at the phonetic level parallels the system found in the neighboring Tirma and Chai languages, and thus is best explained as an instance of areal contact. The Baale tonal system is still poorly understood. According to Yigezu and Dimmendaal (: –), it has three register tones: high (´), mid (which is left unmarked in the present contribution), and low (`), whereby the low tone appears to have a limited distribution; there does not appear to be any downdrift in Baale. As shown in Yigezu (), Baale has borrowed heavily from Tirma and Chai, including basic vocabulary terms for body parts and kinship terminology. These phonological and lexical properties as well as the strong morphosyntactic convergence towards Tirma and Chai probably are a consequence of the frequent shift towards these two languages in dayto-day interaction. Nevertheless, the Baale people as a minority group did not give up their first language, presumably because of their “double identity”, as Suri people (like the Tirma and Chai) and as Baale, a community whose main means of subsistence is agriculture, unlike their Tirma and Chai neighbors, who see themselves as pastoralists.
. M
.................................................................................................................................. One of the stable features of Northeastern Nilo-Saharan languages (as defined in Dimmendaal, chapter of this volume) is the tripartite system of number marking for nouns (Dimmendaal a), which has been retained as such in Baale. It involves singulative marking (for nouns which are inherently plural or collective in their morphologically basic form), plural marking, and replacement (involving number suffixation in the singular as well as the plural). One productive pattern, already introduced in () above, involves a suffix -Ca which copies its consonantal features from the root-final consonant: ()
Singular Plural kaala< t> kaalat̪-t̪á ‘shin’ horó< c> horoc-cá ‘cheek’
The geminates also show the inherent voiceless character of the final stop.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
A further productive type of plural formation involves a [+ATR] high front vowel -i, which triggers harmony shift on preceding [–ATR] root vowels, as in the word for ‘body’ below: () Singular Plural εεlέ< t̪> eelet̪-í ‘body’ uŋŋé< t̪> uŋŋet̪-í ‘nose’ Apart from roots, suffixes may also end in a floating (i.e. hidden) stop. The plural forms of the following words end in a high front vowel when pronounced in isolation, but when they are followed by another suffix or clitic, a floating < t̪> appears (realized as [ð] intervocalically). ()
Singular Plural rúm rumm-í< t̪> ‘cloth’ rumm-í< t̪>-a-gaandí [rūmmíðāgāāndí] ‘my clothes’
The following suffixation types are also semi-productive: ()
Singular ɗílá cól
Plural ɗɪla-ɲá col-ɲá
‘spear’ ‘little spear’
síró hallá
sireen hallεεn
‘antelope’ ‘finger’
ʊt̪ʊ́< k>
ut̪ug-et̪í ‘mouth’ túttú< k> tuttug-et̪í ‘doorway’
Singulative marking occurs for nouns referring to entities occurring in groups, as collectives, or in pairs. The most productive marker is the suffix -Vni, which causes preceding [–ATR] vowels (others than a) to shift harmony class: ()
Singular ŋɪgɪtaaní ɓallógóóní máríccóg-óní segerééní emmeeni
Plural ŋígítá ɓallɔ́gá maricco sεgέrá εmmέ
‘tooth’ ‘leaf ’ ‘pole’ ‘cowry shell’ ‘bone’
Alternatively, there is replacement as a number-marking strategy: () Singular Plural táɗʊ́gɔ́-n taɗʊgɔ-í< t̪> ‘mucus’ ɓallóg-í ɓallɔ́g-á ‘leaf ’ Such final liaison consonants may in fact constitute separate morphemes themselves, e.g. the singulative marker -c, which consequently is only realized before another suffix or clitic:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
só-< c> [só] ‘leg’ sóc-ā-nāāndí [sóʑānāāndí] ‘my leg’
Kinship terms in Baale take a plural prefix ko-: () Singular Plural bábbá kó-bábbá ‘father’ yáyá kó-yáyá ‘mother’ Nominal modifiers other than relative clauses follow the head noun in Baale. ()
tɪɪná rámmá ‘two cows’ tɪɪná ɲjaagá ‘those cows’ tɪɪná rámmá koolyé ‘two dark (black) cows’
Like other Surmic languages, Baale has a quinary numeral system. ()
óóɗe rámmá íyyó wé túr
‘one’ ‘two’ ‘three’ ‘four’ ‘five’
tɔrkɔ́nɔ tʊrgέrε túrgé tɔrgɔ́gɔ ɔmɔtɔ́
‘six’ ‘seven’ ‘eight’ ‘nine’ ‘ten’
As is more common in the area, the word for ‘man’ or ‘person’ is used metaphorically to refer to ‘twenty’: eeccí óóɗe ‘one person, twenty’. Like numerals and adjectives, demonstratives follow the head noun in Baale. ()
Proximate Distant ɲiiní ɲaaná (ɲ)jeegí (ɲ)jaagá
Whenever a noun is modified by a demonstrative or pronominal possessive, the former takes a specifier or definiteness marker, as it is called by Lapranova (), who also points out that the order within the noun phrase is noun–specifier–relative clause–adjective– demonstrative. A similar pattern is found in the Southern Nilotic Kalenjin group and in Daju (Dimmendaal to appear). The demonstratives above and the pronominal possessives below (Table .) manifest reflexes of the common Nilo-Saharan deictic elements *n (), *k (). A typologically interesting property of Baale and other Nilo-Saharan languages with a tripartite numbermarking system is the morphological distinction on the pronominal possessive depending on whether the nomen rectum is a singular or plural noun.2
It should be noted that ‘name’ in Baale functions as a plural noun: sárā-ā-gāāndí ‘my (collection of) name(s)’. As pointed out by Greenberg (: ), this property is widespread in Nilo-Saharan. 2
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Table . Pronominal possessives in Baale Possessor
Nomen regens sg
Nomen regens pl
Possessor
Nomen regens sg
Nomen regens pl
1sg
naandá
gaandá
1pl:excl
náa
gáa
1pl:incl
naá
gaá
2sg
nuundé
guundé
2pl
nuú
guú
3sg
niindέ
gɪɪndέ
3pl
nií
gií
Genitive constructions are rendered by way of a connective marker between the possessum and the possessor. If the possessor is a noun, the latter takes a genitive case marker -n/-o () / -u (): () eɗíŋ a έέss-ɔ meat goat-: ‘goat’s meat’ eɗíŋ a έέyá-ú meat goats-: ‘(the) meat of goats’ Compared to Northern members of Eastern Sudanic (e.g. Taman or Nubian languages), Baale and other Surmic languages have a reduced case system. Apart from genitive case, Baale has an oblique case marker for adjuncts (covering different semantic roles), and nominative case marking for postverbal subjects of transitive and intransitive verbs. Preverbal subjects or objects (regardless of their position) are not inflected for case, and take absolutive case. This is the common pattern in Surmic (and various Nilotic languages as well as the Jebel language Gaahmg). Like other Surmic and Nilotic languages, and unlike Taman or Nubian languages, Baale also has an extensive system of head marking on the verb (involving the expression of semantic roles like direction, instrument, and beneficiary). These latter properties are discussed below after a brief survey of the case system. Whereas Eastern Sudanic languages belonging to the Northern branch (e.g. Taman, Nubian, or Nyimang plus Afitti) are verb-final, Southern members (Jebel, Surmic, Nilotic, Daju, and Temeinian) are either verb-initial or verb-second. This typological split corresponds to a number of other differences. Whereas Northern members have extensive casemarking systems, Southern members of the Eastern Sudanic branch within Nilo-Saharan show a much reduced system.3 Moreover, instead of marking objects for case, Southern members of Eastern Sudanic mark the subject (when the latter occurs in post-verbal position). Baale and other Southern Surmic languages have an oblique case marker -a covering different semantic roles, such as instrument or location. In combination with motion verbs, they may also refer to the direction or source of some verbal event. As is 3
As remnants of a more peripheral case-marking strategy are still found in Surmic and Nilotic (Dimmendaal to appear), the situation in Northern Eastern Sudanic probably constitutes a retention, also because such systems are found in more distantly related Nilo-Saharan groups such as Saharan, Fur, Kunama, or Maban.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
common in all Southern Eastern Sudanic languages with case (i.e. in Surmic and Nilotic), subjects in Baale take a case marker when they follow the verb, but not when they precede the latter, whereas objects are never marked for case. ()
ka-úkki wórgí gaáttíc-á. :-search gold: pen- ‘I am searching for gold with a gold pen.’
Apart from the oblique case marker, there are post-nominal modifiers or postpositions derived from nouns further specifying the search domain for objects, which also take an oblique case marker: () kíssa lot̪-á house. earth- ‘underneath the house’ kíssa ɓágúj-á house. earth- ‘behind the house’ In addition, there are a few prepositions, e.g. ká (location) and kó (accompaniment). These also appear in combination with some nouns, presumably reflecting former (headless) phrases, e.g. with the location-marking preposition ka, as in kaɓɔ́gɔ́ ‘hole’ (compare ɓɔgɔ ‘dig’). This pattern coexists with a pattern of nominal derivation involving suffixation in Baale, e.g. mʊ́lʊ́g-ɔ́n ‘sweat’ (compare -mʊlʊ< k> ‘sweat (verb)’, ɗiir-í< c> ‘broom’, from ɗir-‘sweep’). Because the formal marking of post-verbal subjects applies to subjects of transitive as well as intransitive predications (i.e. A and S roles), Baale may be said to have a marked nominative system. Whereas the closely related Didinga-Murle languages are verb-initial, constituent order in Baale is more flexible, patterning along with the system found in the neighboring Tirma and Chai languages. Together with the phonological features discussed in section . above, and the extensive structural and lexical borrowing (including basic vocabulary, as shown in Yigezu ()), these are manifestations of metatypy in Baale. In Dimmendaal (b: ), it is argued “that the common syntactic patterning in Surmic languages that are not strictly verb-initial is a verb-second structure, with alternatives (e.g. SVO versus OVS, or even SOV) being conditioned by pragmatics or discourse.” Preverbal subjects in verbal and non-verbal predications as well as objects (regardless of their syntactic positions) or nouns in isolation always occur in the morphologically unmarked case, the absolutive form, in Baale and other Surmic languages. Nominative case in Baale is expressed by way of a case suffix -i/-e, or - /-i in the singular, and -na in the plural, or by way of tonal inflection (depending on the noun). () á-gʊr teerúj-è. :-drink woman- ‘The woman drinks it.’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
The verb appears to form one phonological word with the following subject noun, as suggested by morphophonological alternations. For example, the root-final consonant of the verbal root kin- ‘drink’ in the example above is changed into r. The intervocalic voicing of the root-initial stop between vowels is regular; moreover, implosive ɗ does not occur in syllable-final position within a word. The corresponding implosive is found, for example, in the imperative singular form kʊɗ-ɔ, where the suffix -ɔ expresses an antipassive and forms a syllable with the preceding implosive (kʊ.ɗɔ). The same lenition rule (ɗ changing to r) occurs when other focused constituents, for example, object nouns, follow the verb: () á-gʊr ɗááccí. :-drink honey.beer: ‘(S)he drinks honey beer.’ Within Eastern Sudanic, Surmic languages like Baale tend to have more extensive verb morphologies than their distant relatives, such as the Taman or Nubian group. Whereas it is common to mark derivational roles affecting core constituents, such as voice markers, on the verb in all these languages, Surmic languages developed a more extensive head-marking strategy on the verb affecting peripheral roles like instrument or direction (which are expressed by way of case in Northern Eastern Sudanic languages). The verb constitutes the most complex and most intricate part of Baale grammar, and is characterized by a high degree of morphological complexity with the following morphological slots: SU–ASP– root–DER–DER–SU–OB. The complexity of the system derives from the morphophonological alternations accompanying these derivational and inflectional processes involving [ATR] harmony and liaison consonant as well as consonant alternation between adjacent morphemes. Baale makes a formal distinction between perfective and imperfective verbs. The imperfective is expressed by a prefix -a-, whereas the perfective is expressed by an underspecified vowel prefix -V- which takes its features from the first vowel of the following verb stem, as illustrated in Table .. Note also that the [+ATR] vowels of the second person singular and plural suffix (-u and -tu, respectively) are dominant suffixes causing preceding [–ATR] vowels to shift to their corresponding [+ATR] counterparts. The plural suffix -ta is widespread in Southeastern Eastern Sudanic languages; it is also found in the Jebel language Gaahmg or in Nilotic languages. De Jong () shows that the actual forms for the perfective versus imperfective in the closely related language Didinga further depend on whether realis as against irrealis mood is involved. Imperfective realis is used, for example, in order to give background information, whereas imperfective irrealis is used for a warning in Didinga. Realis/irrealis distinctions also appear to play a role in Baale, where the perfective is used after the auxiliary element nɔɔ́, in order to express a hortative/permissive, for example: () nɔɔ́ kʊ-gʊ́ɗ-ɔ! :-drink- ‘Let me drink!’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The imperfective marker a- is also used with non-verbal predications, where it expresses not only identity relations but also semantic relations such as ‘having the properties or qualities of ’. () rúm ammaa. cloth .water ‘The cloth is wet.’ The following examples illustrate some of the morphophonological alternations occurring between verbal roots and (inflectional and derivational) suffixes. The examples involve first person singular as against plural verb forms in the perfective. () kirít̪ee kesséjee kajájee kubúɗee kuttúloo
‘I have coughed.’ ‘I have hidden.’ ‘I have exchanged.’ ‘I have looked back.’ ‘I have plaited hair.’
kirít̪t ̪aa kesséccaa kajáccaa kubúʄaa kuttulyaa
‘We have coughed.’ ‘We have hidden.’ ‘We have exchanged.’ ‘We have looked back.’ ‘We have plaited hair.’
Whereas subjects are expressed by way of prefixes plus suffixes in Baale, corresponding pronominal objects are marked by way of suffixes. Verbal roots or stems occur in the conjoint form when followed by these cross-reference markers. () -na -nee -ni -nuuŋ -Ø -Ø An important structural property of Baale, also expressed in Table . above, is the distinction between, what is called here conjoint and disjoint verb forms.4 Conjoint verb forms in Baale occur when a syntactic element immediately following the verb carries focus. This may be an object or an incorporated peripheral constituent (which requires derivational morphology on the verb, as pointed out above), but it could also be a subject. Whether this structural property is more common in Surmic or unique to Baale is not known. The conjoint form also occurs with “derived objects”, i.e. when adjunctival forms are “incorporated” into the verbal complex by way of a derivational extension on the verb (see below). As shown in example () above, the conjoint verb form plus this syntactic phrase carrying focus form one phonological word. In other words, such phrases are characterized by the same sequence structure conditions as lexical words (consisting of morphologically simple or complex forms). The conjoint form is also used with predicative constructions:
4 This terminological distinction derives from the study of Bantu languages, where similar phenomena occur. The Baale system shows how such a system in Bantu may have come about historically, i.e. through antipassive marking.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Table . Perfective and imperfective verb forms in Baale Perfective
Imperfective
Conjoint
Disjoint
Conjoint
Disjoint
1sg
kV- . . . -a
kV- . . . -VV
ka- . . . -i
ka- . . . -VV
2sg
V- . . . -u
V- . . . -VV
a- . . . -i
a- . . . -VV
3sg
V- . . . -Ø
V- . . . -ɔɔ-
a- . . . -Ø
a- . . . -VV
1pl:excl
kV- . . . -ta
kV- . . . -taa
ka- . . . -na
ka- . . . -naa
1pl:incl
kV- . . . -(C)V?
kV- . . . -(C)VV
ka- . . . -(C)V?
ka- . . . -(C)VV?
2pl
V- . . . -tu
V- . . . -tuo
a- . . . -nu
a- . . . -nuo
3pl
V- . . . -ɪ
V- . . . -ɪta
a- . . . -nɛ
a- . . . -nɛɛ
() andá kééni baalejíní. :be Baale ‘As for me, I am a Baale.’ () aggá kegínna baalέ. :be Baale ‘We are Baale.’ Disjoint verb forms do not form a phonological unit with the following word.5 In the case of transitive disjoint verb forms, the verb also contains an antipassive marker -a/-o (which may fuse with preceding inflectional or derivational suffixes on the same verb), as shown in the following example: ()
kát̪-ɔ́! kill- ‘Kill!’ kát̪ taŋŋá! kill cow ‘Kill the cow!’
With first and second person singular and plural verb forms, for example, the antipassive suffix fuses with preceding person markers, resulting in morphophonemic alternations: ()
5
Conjoint form Disjoint form ka-ɗág-i ka-ɗág-ee < *ka-ɗagi-o
As shown by Joswig (), similar contrasts between conjoint and disjoint occur in Surmic languages like Majang, Me’en, and Tirma(ga).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The disjoint verb form is common with negative verb forms in Baale. The latter are expressed by way of a proclitic ŋa=, a marker probably going back to a verb ŋay-o attested across the family both as a free form and as a clitic or affix (Dimmendaal b: ). Rootinitial voiceless stops in Baale automatically become voiced when a person-marking prefix is added, as shown in example () above. As the same root-initial voiceless stops remain voiceless when the negation marker is added, the latter presumably is a clitic rather than a prefix. ()
ŋa=-ka-ɠéy-oó. [ŋākāɠéyōó]
=-eat- ‘I don’t eat (it).’
Apart from the antipassive as a valency-changing suffix, Baale has a passive, formed by way of a suffix -(C)w (as well as additional forms, sometimes combined with a prefix k-, depending on the tense–aspect form); the (optional) phrase expressing the agent takes the oblique case marker: () á-jεɓεε polís-à. :-tie. police- ‘(S)he is being tied up/handcuffed by the police.’ Lapranova (: ) mentions a reciprocal suffix -(ɔ)yɔ / -(o)yo. Because of cross-reference marking for subject and object on the verb, independent pronouns play a role only at the pragmatic level (with topicalization or focus marking). In the following example, an independent pronoun introduces a topic (or topicalized subject) followed by the object plus verb. The latter consists of a verb root plus instrumental marking, and is followed by a question word occurring in the post-verbal focus position. Similar topic OV (or SOV) order as well as verb-second structures with a post-verbal constituent carrying focus is attested in Tirma and Chai (Last and Lucassen ). ()
nɔ́gɔ eɗíŋ á-ɗag iɲá? meat -eat why ‘(As for them) why are they eating meat?’
() eɗíŋ kà-ɗág-ì nāāná. meat: :-eat-: : ‘I am eating meat.’ The focus position immediately after the verb is also the preferred position for question words, except when referring to the subject or agent; in that case, the question word precedes the verb. () áŋŋεεnέ á-ɗag eɗíŋ? who -eat meat ‘Who is eating meat?’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Corresponding negative constructions tend to occur with the antipassive, i.e. in the disjunctive form, although it is possible to combine these with conjoint verb forms, as in the following example: () ŋa-ɔggɔr tíína a búméc-óǃ -steal: cows Nyangatom- ‘Don’t steal the cows of the Nyangatom (people)!’ The morphophonology of the Baale verb is complex due to the interaction of the various suffixes with each other and with the root. [ATR] harmony, fusion of vowels and consonants, and consonant alternation together result in a sheer endless set of verb forms; also, verb roots may be expanded with different derivational suffixes, such as the passive and antipassive above, as well as the ventive, itive, dative, and instrument. Because of these complex alternations, it is not always easy to identify the formal structure of these formatives. There are two directional forms, ventive -n- and itive -a-. The dative is expressed by way of a suffix -(V)k-, and expresses semantic roles such as experiencer or benefactive. () á-uk-k-aa laʄo :-affect-- cold ‘I feel cold.’ (lit. ‘The cold affects me.’) Instrumental marking on the verb by way of a suffix -(C)ay reflects the presence of an instrumental role involved in the verbal event. The instrument marking constituent may either be expressed or be understood from the discourse context (i.e. its cognitive status may be inferentially or situationally available). The combination of verbal roots with derivational as well as inflectional suffixes results in a sheer infinite number of verb forms, a feature which Baale shares with the closely related Didinga-Murle languages. The interested reader is referred to Stirtz () for a description of this phenomenon in Didinga. Complement clauses follow the main clause in Baale: () wá kʊssígaá kanní kakkʊ́naa. :.tell .say :.come ‘I said that I would come.’ Adverbial clauses are introduced by conjunctions and may either precede or follow the main clause, depending on whether they constitute background information or important new information, respectively. In some constructions, such initial conjunctive markers are combined with clause-final enclitics, as with conditional clauses, which take an enclitic marker -go: () wássi ka-ɠá-gò . . . -know- ‘If I know . . . ’
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A Data on Baale were collected in southwestern Ethiopia by the present author in the s. I would like to express my sincere thanks to the Institute of Ethiopian Studies, Addis Ababa University, for making this research possible. The present author’s data derived mainly from the speech of Alemu Olecagi, a brilliant young man and wonderful language consultant who was killed in an unfortunate accident in . This contribution is dedicated to his memory. Additional data on Baale were collected by Moges Yigezu (see Yigezu , ).
R de Jong, N. (). ‘The verbal matrix in Didinga’, in A. Abu-Manga, L. Gilley, and A. Storch (eds.), Insights into Nilo-Saharan Language, History, and Culture. Cologne: Rüdiger Köppe, –. Dimmendaal, G. J. (a). ‘A syntactic typology of Surmic from an areal and historical-comparative point of view’, in G. J. Dimmendaal and M. Last (eds.), –. Dimmendaal, G. J. (b). ‘Surmic languages and cultures: an introduction’, in G. J. Dimmendaal and M. Last (eds.), –. Dimmendaal, G. J. (a). ‘Number marking and noun categorization in Nilo-Saharan languages’, Anthropological Linguistics : –. Dimmendaal, G. J. (b). ‘Noun classification in Baale’, in R. Vossen, A. Mietzner, and A. Meissner (eds.), “Mehr als nur Worte . . . ”: Afrikanistische Beitäge zum . Geburtstag von Franz Rottland. Cologne: Rüdiger Köppe, –. Dimmendaal, G. J. (). ‘Language diffusion and genetic relationship: an African perspective’, in R. M. W. Dixon and A. Y. Aikhenvald (eds.), Language Diffusion and Genetic Linguistics. Oxford: Oxford University Press, –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Dimmendaal, G. J. (to appear). ‘Nilo-Saharan: an overview’, in Bedilu Wakjira, R. Meyer, Y. Treis, and Zelealem Leyew (eds.), The Oxford Handbook of Ethiopian Languages. London: Oxford University Press. Dimmendaal, G. J., and M. Last (eds.) (). Surmic Languages and Cultures. Cologne: Rüdiger Köppe. Greenberg, J. (). The Languages of Africa. Bloomington, IN: Indiana University Press. Joswig, A. (). ‘Syntactic sensitivity and preferred clause structure in Majang’, in A. Storch and A. Mietzner (eds.), Nilo-Saharan—Models and Description. Cologne: Rüdiger Köppe, –. Lapranova, L. (). A Brief Grammar of the Kacipo-Balesi Language. Addis Ababa: SIL Ethiopia. Last, M., and Lucassen, D. (). ‘A grammatical sketch of Chai, a Southeastern Surmic language’, in G. J. Dimmendaal and M. Last (eds.), –. Möller, M. (). Vowel Harmony in Bale: a study of ATR harmony in a Surmic language. M.A. thesis. Department of Linguistics, University of Stockholm. Stirtz, T. M. (). ‘Laarim tone’, SIL Electronic Working Papers -. Yigezu, M. (). A Comparative Study of the Phonetics and Phonology of Surmic Languages. D.Phil. thesis. Université Libre de Bruxelles. Yigezu, M. (). ‘Convergence of Baale, a Southwest Surmic language, to the Southeast Surmic group: lexical evidence’, Annual Publication in African Linguistics : –. Yigezu, M. (to appear). ‘Ngaalam: an endangered Nilo-Saharan language of southwest Ethiopia—a sociolinguistic survey on language vitality and endangerment’, Nairobi Journal of Language and Linguistics. Yigezu, M., and Dimmendaal, G. J. (). ‘Notes on Baale’, in G. J. Dimmendaal and M. Last (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. S is a language family of the West African Sahel region with more than four million speakers, distributed mainly in western Niger and northeastern Mali along the Niger river, but including outlying pockets scattered as far afield as Algeria, Ghana, and Sudan. In spite of this small population, Songhay’s role in regional history has been substantial: the Songhay capital, Gao, was one of the earliest recorded termini of the trans-Saharan trade, while the Songhay Empire later dominated most of the western Sahel, and Songhay remained an important trade language into the nineteenth century. Strong contact effects along the periphery—notably involving Berber in the north, Bariba and Hausa in the southeast, and Mande languages at an earlier stage—have resulted in a remarkable situation where different varieties may have nearly identical basic vocabularies but conspicuously different typologies, varying in, for example, whether direct objects are preverbal or post-verbal; whether noun plurality is marked morphologically or syntactically; whether the genitive may or may not follow the head; whether verb serialization is possible or not; whether tone or pharyngealization is phonemic or not, etc. Cladistic and lexical evidence shows that much of the observed variation reflects contact influences quite different from those obtaining today, including influences from non-Tuareg Berber languages that are now nearly extinct in the region. After a brief introduction detailing the distribution, history, and phonology of Songhay and listing the principal sources for its grammar and lexicon, this chapter examines morphology and syntax across the family. Particular attention is given to innovations distinguishing the three principal subgroups (Eastern, Western, and Northern), and to the motivations for these changes. Finally, the structure and history of the lexicon is briefly discussed, with an emphasis on borrowing as a means of expansion. Examples are drawn partly from the author’s fieldwork on Kwarandzyey/Korandjé (Algeria). Songhay is a close-knit language family spoken by more than four million people, including some % of Malians and % of Nigeriens. The best-known hypothesis on its wider genetic affiliation is Greenberg’s (), linking it to Nilo-Saharan, but the comparisons are questionable (Lacroix ). Nicolaï’s (, ) Afro-Asiatic lexical comparisons are equally speculative (Dimmendaal ), although his proposal that Songhay
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
originated as a contact language with a Mande substratum is promising. Over the past millennium, Songhay has played an important role in Sahelian history, embracing centers of trans-Saharan trade and scholarship such as Gao and Timbuktu; in the fifteenth and sixteenth centuries, the Songhay Empire controlled most of the western Sahel. Perhaps owing to the historic use of Songhay as a lingua franca, several Songhay languages show strikingly high levels of contact influence, leading to a remarkable typological diversity given the homogeneity of the basic vocabulary. Songhay may be subgrouped as follows: • Eastern: • Benin Dendi (restructured)—Djougou, Kandi (Benin) (Zima ) • Niger Dendi—Karimama (southwestern Niger) (Tersis-Surugue ) • Zarma—western Niger (Tersis-Surugue ; Bernard and White-Kaba ; Sibomana ), pockets in Nigeria • Humburi Senni (HS)/Central Songhay—Hombori (Mali), pockets in Burkina Faso (Heath ) • Kaado—northwestern Niger (Ducroz and Charles ; Hanafiou ) • Koyraboro Senni (KS)—northeastern Mali around Gao (Prost ; Heath a), pockets in Sudan (Abu-Manga ) • Tondi Songway Kiini (TSK)—Kikara (Mali) (Heath ) • Northwestern: • Western: (Heath b) • Koyra Chiini (KC)—around Timbuktu (Mali) • Djenné Chiini (DC)—Djenné(Mali) • Northern: • Kwarandzyey / Korandjé—Tabelbala (Algeria) (Souag ) • Nomadic: • Tadaksahak—northeastern Mali (Christiansen-Bolli ) • Tagdal—northern Niger • Tasawaq—Ingal (Niger) (Alidou ; Kossmann ) • Emghedesie (extinct)—Agades (Niger) (Barth ; Lacroix ) The family is centered on western Niger, with a continuous distribution along the Niger River from Diré (Mali) to the Nigerian border, and a discontinuous scattering of more divergent varieties away from the river. Despite its historical interest, the more divergent Northwestern subgroup accounts for only about % of speakers, many of them culturally Tuareg.
. P
.................................................................................................................................. Proto-Songhay had the consonant system given in Table ., following Nicolaï () except in splitting *r (never initial) from *z (rarely medial). Except in HS and TSK, *ỹ, *w̃ usually > /ñ/, /ŋ/; in Northern they disappear from the phonology. Labiovelarized stops are sometimes lost. /k/ and /g/ are usually palatalized
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Consonant system of Proto-Songhay b m f
td n sz l r
ỹ(ñ)
kg
kʷ gʷ w̃ (ŋʷ)
y
w
h
before front vowels, often creating (alongside loanwords) phonemic /c/, /j/. Loanwords have introduced marginal /p/ to most Eastern varieties, along with other phonemes, such as KS /š/, /ž/. Under Mande influence, Western has merged *z > /j/; under Gur influence, Benin Dendi has turned labiovelarized stops into labiovelars /kp/, /gb/, /ŋm/. Berber’s effect on Northern has been substantial; pharyngealization has typically become phonemic for coronals, and new dorsals /q/, /ɣ/, /x/ have been added, even in inherited words. Some varieties have also borrowed Arabic pharyngeals /ħ/, /ʕ/. The proto- (and pan-)Songhay vowel system is /a e i o u/, with vowel length contrastive. Nasal vowels are attested following h. In the southernmost varieties /e/ and /o/ have split under areal influence, yielding new phonemes /ε/ and /ɔ/ (e.g. Djenné čee ‘time’ vs. čεε ‘foot’). Northern has gained a new central vowel /ə/ from Tuareg. Kwarandzyey’s vowel system has drifted furthest; vowel length is non-phonemic, mid vowels have merged to high, and *a has split, e.g. ha ‘ask’ vs. hɑ ‘play’. The canonical proto-Songhay syllable was CVV/CVC, with CVVC possible word-finally; most words are one to three syllables long. Initial V and syllabic n are rare but reconstructible. In coda position, voicing was non-contrastive, and fricatives absent; except in Northern, even coda stops have been lenited (Souag ). Proto-Songhay had two level tones; contour tones, though widely attested, usually result from syllable loss (e.g. Zarma tǔ:rì < *tùgúdì ‘wood’) or borrowing. In compounding/ affixation, the first stem’s tones are often lowered. Under Berber/Arabic influence, tone has been lost in KS, Western, and most Northern; *hàmnì ‘flour’ and *hámní ‘fly’ become homonyms. In nomadic varieties, this was balanced by developing contrastive stress: for inherited words, stress is penultimate from original HL (e.g. tu'gudu ‘tree’), final otherwise (Nicolaï : ).
. D
.................................................................................................................................. A few nominal derivational bound suffixes are found, including diminutive *-íyòw, e.g. Gao aru ‘man’ > aryu ‘little man’; *-tàrây (cf. *táráy ‘outside’) for natures or homelands, e.g. Zarma zármá ‘Zarma’ > zàrmà-tàrây ‘Dosso region’; and (pan-Sahelian) gentilic *-ànké, e.g. Zarma gùrmá ‘Gourma, south bank’ > gùrm-àncé ‘Gourma man’. Only *-tàrây survives in Northern. (All derivational processes mentioned are absent from Northern Songhay unless otherwise indicated.)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Affixation may once have been more extensive. Prost (: ) suggests that final vowels mark noun class; most nouns ending in -i refer to liquids/collectives/abstracts, and many flora take -ey. The prefix n-/(h)àn-, mainly for animal names (cf. Mande), varies across varieties (e.g. Timbuktu ŋkondo ‘ant’ vs. Kaado kòndò). Equally striking is the suffix *-mè in *àrmè ‘brother’, *wàymè ‘sister’ (cf. *àrù ‘man’, *wày ‘woman’). Many word pairs, such as *hàndú ‘moon’ vs. *hàndàrí ‘star’, or *kòbì ‘finger’ vs. *kòbsi ‘hoof ’, suggest fossilized suffixes, but have not been systematically analyzed. Compounding is a commoner noun formation method. Several nouns are particularly productive as finals: *ízè ‘child’ forms names of parts, fruits, juveniles, etc., e.g. Timbuktu malfa ‘rifle’ > malfajje ‘bullet’; *ỹâ: ‘mother’ refers to plants, e.g. Gao dugu ‘incense’ > duguñaa ‘incense plant’; *àrù ‘man’ and *wày ‘woman’ specify sex; *kóỳ ‘master’ / *-kòyní forms characteristic nouns. In Northern, compounding is less productive, but ‘woman’ and ‘master’ have become suffixes, e.g. Kwarandzyey izi ‘boy’ > izwəy ‘girl’, kʷɑrɑ ‘village’ > kʷɑrɑkʷəy ‘villager’. More spectacular compounds are attested sporadically, including nominalizations of whole sentences, along with reduplicated forms. Predicative “adjectives” are stative verbs; attributive adjectives, however, form a syntactically distinct class. Some attributives are derived by tonal change, zero-derivation, or reduplication; most, by suffixing *-òw (also Northern), e.g. Kaado móorú ‘be sour’ > móoró ‘sour’. They may be formed from other verbs by suffixing participial *-ànté (cf. Soninke -nte), e.g. Zarma tón ‘burn, grill’ > tònànté ‘grilled’; this suffix also forms ordinals from numerals (cf. Soninke -ndi). Northern has almost entirely lost *-ànté (participial -an is from Berber), but shares other strategies discussed by Kossmann (), e.g. Kwarandzyey tsirəy ‘be red = red’, ku ‘be long’ > kuku ‘long’, gəb ‘be tough’ > gabuw ‘tough’. Reduplication marks distributivity, especially for numerals (including in Northern). Many verbs are labile, with zero-derived passives (cf. Mande, Berber). Otherwise, diathesis changes are handled by suffixation or tonal change. A single suffix *-ándì forms causatives from intransitive verbs (and a few transitive verbs) and potential passives from transitive verbs, e.g. Gao too ‘arrive’ > toondi ‘deliver’, gar ‘find’ > garandi ‘be found’ (findable). (Compare Mandinka, Soninke causative -ndi, Bozo -ni.) In Northern Songhay, the only inherited causative suffix is *-ǹdá, probably reshaped from *-ándì under the influence of *ǹdá ‘with’, calquing Berber, in which causative s- resembles instrumental s. KS and Hombori also have *-à, forming resultative passives or indicating an unspecified object, e.g. Hombori čín ‘build’ > čìnà ‘be built’ (cf. Soninke, Bozo -e). Pairs like *yê: ‘return (intr.)’ vs. *yè:tí ‘return (tr.)’ suggest other fossilized diathesis morphemes. Comitative *ǹdá can be suffixed to verbs in several varieties, mainly forming verbs of possession or transport, e.g. Gao kaa ‘come’ > kaanda ‘bring’. Pluractionals are formed through reduplication (also in Northern), e.g. Zarma fûn ‘pierce, be pierced’ > fúnfùn ‘pierce little holes’. Nouns of instrument are formed using *-(dí)gì, e.g. Zarma fûn ‘pierce’ > fùníjì ‘drill’ (traces in Northern). Characteristic nouns are formed from verbs by suffixing *-kówò or rarer *-kòḿ, e.g. Zarma wí ‘kill’ > wìikô ‘killer’; Northwestern substitutes nominal *-kóỳ, e.g. Timbuktu baa ‘want’ > baakoy ‘close friend’. The formation of verbal nouns is somewhat irregular. The most productive suffixes are identical to plural markers: (indefinite) plural *yaw̃ and (definite plural) *-ây. In Eastern this is mere homophony, since they take definite suffixes, e.g. Gao dur-yaŋ-oo (pound-) ‘pounding’; however, in Kwarandzyey this is polyfunctionality, since the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
marker is separable, e.g. gʷɑ=f ʷ hənn-u=yu (sit=one good-=) ‘some good sitting-down’. *-í (traces in Western) and *-dì > -rì are both productive in Eastern, e.g. Zarma dòonù ‘sing’ > dòoní ‘song’, Kaado dúmbú ‘cut’ > dùmbàrì ‘piece’. In Northern, zero-derivation is also well-attested (its status elsewhere is arguable). Other lexically restricted minor suffixes include *-déy > -rey (traces in Northern), *-mí, and *-âw. Northern Songhay compensates for the loss of most inherited morphology by borrowing derivational processes, mostly limited to Berber stems but occasionally extended to inherited ones. In nomadic varieties, diathesis morphology only applies Berber affixes to Berber stems, so that most inherited verbs have suppletive diathesis derivations, e.g. Tadaksahak ŋá ‘eat’ > š-íkša ‘make eat’. Similarly, in Kwarandzyey, Arabic stative loanverbs take Arabic factatives, e.g. yəmḍɑ ‘be sharp’ > mʌḍḍɑ ‘sharpen’. Berber nominal affixes . a-, . ta–t are used to mark sex or size, even on inherited words, e.g. Tadaksahak karfú ‘rope’ > diminutive takarfútt, Kwarandzyey fumbu ‘stinky’ > tsafumbuyts ‘stinky woman’.
. N
.................................................................................................................................. In most of Eastern, one suffix marks plurality and definiteness: *-oˇ :/ǎ:, : *-ěy, : *-yáw̃. This is syntactically independent, being attached to the last element of the core NP, e.g. Gao woy beeri hiŋk-ant-oo (woman big two--:) ‘the second big woman’. In HS the : suffix has lost definite force, becoming obligatory even for indefinites except before modifiers. Demonstratives are separately inflected for number, e.g. Gao bor-ey w-ey (person-: -:) ‘these people’. Northwestern and Benin Dendi show a simpler system with no reflexes of the definite suffixes, in which *-yáw̃ only marks plurality. Usually, *-yáw̃ retains syntactic independence; thus, Kwarandzyey bɑ bya=ɣ=yu (person big==) ‘these big people’. In nomadic varieties, however, the corresponding suffix -en (cf. Berber -en) has become an agreement marker, appearing on heads and modifiers alike, as in Berber, e.g. Tadaksahak bi-yén giŋ-gimán-ən (shade- -good-:) ‘good shades’. Throughout Northern, Berber/Arabic words have been borrowed together with their plurals, independently leading to double marking, e.g. Kwarandzyey lħwayəž fts=yu (things bad=) ‘bad things’ vs. lħašt ‘thing’. In general, Songhay languages show only one adnominal spatial demonstrative, *wô: (cf. Mande, Berber), bound in TSK, alongside *dì(ń) for discourse-old referents. Northern usually has (V)ɣo < *Vgo for the former, but Tadaksahak contrasts aɣo (topicality/salience marker) with -o ‘this’, suggesting that the two are etymologically independent. Further spatial distinctions are usually made using demonstrative adverbs; of these, *ně: ‘here’ and *nôw̃ ‘there’ can be reconstructed for proto-Songhay, while others are probably derived from nouns meaning ‘place/area’ plus demonstratives. Northern varieties (like Berber) show more adnominal spatial distinctions, e.g. Kwarandzyey ɣuna ‘that (distal)’ or Tadaksahak áyda ‘this (close to the addressee)’. For a modifier to act as a head noun, an affix or a dummy head is often required. Adjectives are nominalized by prefixing i- (TSK ŋ́-; relative marker Vɣo in Northern),
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
which also nominalizes lower numerals in most Eastern varieties. *à- nominalizes lower numerals in Northwestern Songhay and (for indefinites only) TSK; elsewhere, it only nominalizes ‘one’ (*à-fó). Most varieties allow the spatial demonstrative to stand alone, but require a head *wô: for *dì; a few peripheral ones require a prefix for the spatial one too, e.g. Kwarandzyey u-ɣu (independent) vs. -ɣu, TSK hɔ̀r-ɔ̂ (independent) vs. -ɔ̂. Absolute genitives are formed with *wánè, e.g. Zarma ày wánè ‘mine’. The structure of the core noun phrase is possessor–head–adjective–numeral–demonstrative (cf. Mande), e.g. Timbuktu bor bibi hiŋka woo di (person black two ) ‘these two black men’, Djougou bààbá bàrí (father horse) ‘the father’s horse’. Some variation is attested, mainly due to contact. In Western the possessor may, and in Northern it must, be followed by a genitive marker, wane or n, respectively (the latter influenced by Berber n ‘of ’). Northern also allows adjective-like postnominal genitives using the full form of *wánè, notably to identify material/content/purpose. Lower numerals always follow the head, as in most sub-Saharan languages. However, in Northern, numerals over ‘ten’ precede the noun, as in Berber and Arabic, no doubt facilitated by extensive lexical borrowing of numerals, e.g. Kwarandzyey xəmsin ɑr: u ‘fifty men’ (= local Arabic xəmsin :r ajəl).
. A
.................................................................................................................................. Personal pronouns distinguish three persons and two numbers; Western and TSK also have distinct logophoric/reflexive third person pronouns. Some varieties have regularized plural pronouns with a plural marker, e.g. Djougou níyò ‘you ()’. Proto-Songhay independent forms were roughly those shown in Table .. Alongside independent pronouns, all varieties have bound forms restricted to particular syntactic environments, notably subjects, possessors, preverbal objects, and objects of postpositions. In most Eastern varieties, these can be replaced by independent forms to mark emphasis or logophoricity. Proto-Songhay already had third person bound forms (e.g. as subject and possessor markers), not obviously derivable from the independent pronouns, and strikingly similar to Mande *à, *ì. Many varieties have extended the distinction to other persons, e.g. Zarma nîn ‘ (independent)’ vs. ní ‘ (subject/ possessor)’. In Northern Songhay, bound forms are often used pleonastically even when the referent is in situ. Sometimes more distinctions have emerged: thus, for ,
Table . Independent pronouns in Proto-Songhay Singular
Plural
*àgáy
*yér(ì)
*ní(ǹ)
*ńgà
logophoric/reflexive
*ǹgu
*wár(àn) Northern: *índì Eastern: *ńgǎy Northwestern: *ńgì *ǹgu-yów̃
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Kwarandzyey distinguishes aɣəy (independent), ʕ- (subject, possessor), and ɣəy (object of verb/postposition). Some varieties have preserved reflexes of *ya, *ma in suppletive datives and subjunctives. Hombori has a system of pronominal suffixes marking possessors or objects of postpositions: -ey, -eyndì, -an, -andòŋ; a similar usage of *-oˇ : to mark the possessor of inalienable nouns is a little more widespread. Core arguments are marked by position: the fairly inflexible basic order in Eastern is subject–mood/aspect/negation–object–verb–other (SAOVX), like Mande, while Northwestern instead uses SAVOX. Even in Eastern some verbs take post-verbal objects, and clausal complements are always post-verbal, as is the second object in double object constructions, e.g. Djougou áà Abdù nó hánsì (he+ Abdou give dog) ‘He gave Abdou a dog’. Otherwise, postpositional *sê marks datives. The basic locative postpositions are *rá ‘in, at’ (cf. Mande; not Northern), *gá ‘on, against’ (= ‘body’). More specific spatiotemporal relationships are handled by locational nouns often derived from body parts, e.g. *bòw̃ò ‘head’ > ‘on’ in Zarma. As commonly in the Sahel, locative adpositions are indifferent to movement: *rá marks source, goal, or location alike, depending on the predicate’s semantics. However, when the path is included, prepositions are used: *zǎ: ‘starting from, since’, *hál ‘until, up to’ (< Berber). *ǹdá (cf. Chadic, Berber) is both an instrumental/comitative/comparative preposition and a conjunction ‘and’. Relative clauses follow the core NP, using a marker *kâ(ŋ) (Northern -ɣo). For most functions, a gap strategy is used, with postpositions pied-piping to follow the relative marker; resumptive strategies are occasionally attested.
. P
.................................................................................................................................. The existential/locative predicator is *goˇ :, replaced by *bárà in certain syntactic contexts, and generally in Northern; its negative is *sí:. *nôw̃ ‘there’ is widely used for identification, e.g. Timbuktu ay nono ‘It’s me.’ The equational copula is perfective *tí / *kí, lost in Tadaksahak and Kwarandzyey. Throughout the family, a four-way mood/aspect distinction is primary (except Kwarandzyey negatives), as shown in Table .. Usually, these are separate words, placed between the subject and the preverbal object or verb; however, they often fuse with adjacent pronouns, and in Northern they are simply preverbal prefixes. Many varieties have developed further distinctions, e.g. future markers,
Table . Mood/aspect distinction in the Songhay family Positive
Negative
Imperative
*Ø, *wà
= subjunctive
Perfective indicative
*Ø (*nà if subject would otherwise be adjacent to object)
*ná (often *màná in Eastern; cf. Mande)
Imperfective ga/go/o/etc. < existential *goˇ indicative Northern b < *bárà
< negative existential *sí
Subjunctive *mà
*mà sí
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
emphatic indicatives. Except in Northern, verb serialization is common, using infinitival *ká (cf. Mande); adverbial, aspectual, and modal functions are often handled using verb serialization, e.g. TSK à w̃ǎ: ká bén ( eat finish) ‘He is done eating.’ Some (postverbal) adverbs are found, e.g. *koˇ yné ‘again’, *tàmbá ‘early, quickly’. Northwestern and KS have developed a centripetal suffix *-kàté ‘hither’ (= ‘bring’), e.g. Gao yee ‘return’ > yee-kate ‘come back’. Under Berber influence, Northern has added centrifugal *-nàn ‘away’ (cf. *náŋ ‘let’, Berber centrifugal -n). Except in Northern, many verbs may be followed by lexeme-specific intensifying ideophone-like interjections, often phonologically anomalous. While these are rather variable, often reflecting areal contact, some may be proto-Songhay, e.g. *tík to intensify ‘black’.
. I
.................................................................................................................................. The main interrogative pronouns are built on *má ‘what?’ (cf. Berber): *máy ‘who?’, *mán(ì) ‘where?’ (cf. Berber), *márgè ‘how many?’, *mátè ‘how?’. *má-kíné ‘what manner?’ loses *má and becomes ‘what?’ in KS, Tadaksahak, Kwarandzyey. Distinct from these is *-fò ‘which?’. Polar questions are formed intonationally or with *wàlá ‘or’ (< Arabic). Elements in contrastive focus are fronted and followed by a particle, notably presentational *nôw̃. In situ focus particles vary across varieties, but postposed *mò ‘also’ and *dà ‘exactly’ are widespread. Contrastive topics are usually initial and followed by a particle *bìndé (not Northern); weak topicalization is often handled in situ by adding a particle, e.g. Timbuktu ta.
. L
.................................................................................................................................. Most of the core vocabulary shows no systematic similarities to other languages. While most Nilo-Saharan etymologies are dubious, good comparisons are available for some words, such as *hárí ‘water’, and especially with Saharan. Plausible grammatical comparisons with Mande can be made; proposed lexical comparisons are less promising, but include basic forms like *kámbé ‘hand’ (Creissels ). More recent loans are widespread outside Northern, e.g. Zarma jàsàráy ‘griot’ < Soninke, Gao maatiga ‘peanut’ < Bambara. Fulani loans are also widely attested outside Northern, e.g. Gao daamal ‘spleen’; in KS they include entire herding-related lexical fields. There may be Gur loans in proto-Songhay, e.g. *hìllí ‘horn’, but Songhay influence on Gur seems more extensive, including horse terminology and social ranks (Prost ). Some Afro-Asiatic loans are almost pan-Songhay, e.g. Berber *tásà ‘liver’, *ǹzórfù ‘silver’, Arabic *kúlù ‘all’, *àlkámà ‘wheat’; most are less widespread, often cultural. Hausa loans are commonest in Niger but have reached Mali, e.g. Gao kowya ‘boondocks’. Songhay has also influenced Afro-Asiatic neighbors, contributing terms to Tuareg (Heath ), Hassaniya (Heath ), and Hausa, possibly including the ethnonym ‘Hausa’ itself (Skinner ). French loans are frequent for modern items, e.g. Gao lekkol ‘school’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Northern Songhay is a special case: its inherited vocabulary has been reduced to a few hundred words, with minimal synonymy, only the top elements of hierarchies like color or number, and few terms for parts/subclasses. Most of the vocabulary is usually Berber, including basic terms like Kwarandzyey tsaɣmməs ‘teeth’, zəgzəg ‘green/blue’. Not all such loans derive from Tuareg (Souag a, b); for instance, Tadaksahak táašinda ‘twenty’ is attested only in Zenaga. Arabic also plays a significant role: in non-nomadic varieties most numerals are Arabic loans, and in Kwarandzyey Arabic loans may outnumber Berber ones. Benin Dendi is a less extreme parallel; while the vocabulary remains predominantly Songhay, regional loans include fairly basic vocabulary like càrgù ‘frog’, góósì ‘forehead’.
R Abu-Manga, A. (). ‘The Songhai speech communities in the Sudan with special reference to the Songhai speakers of the Blue Nile’, in R. Nicolaï and F. Rottland (eds.), Proceedings of the Fifth Nilo-Saharan Linguistics Colloquium, Nice, – August . Cologne: Rüdiger Köppe, –. Alidou, O. (). Tasawaq d’In-Gall: Esquisse linguistique d’une langue dite ‘mixte’. MA thesis. University of Niamey. Barth, H. (). ‘Progress of the African mission, consisting of Messrs. Richardson, Barth, and Overweg, to Central Africa’, Journal of the Royal Geographical Society of London : –. Bernard, Y., and White-Kaba, M. (). Dictionnaire zarma–français (République du Niger). Paris: Agence de coopération culturelle et technique. Christiansen-Bolli, R. (). A Grammar of Tadaksahak: A berberised Songhay language (Mali). Cologne: Rüdiger Köppe. Creissels, D. (). ‘De la possibilité de rapprochements entre le songhay et les langues niger-congo (en particulier Mandé)’, in T. C. Schadeberg and M. L. Bender (eds.), Nilo-Saharan: Proceedings of the First Nilo-Saharan Linguistics Colloquium, Leiden, September –, . Dordrecht: Foris, –. Dimmendaal, G. J. (). ‘Review of Nicolaï ’, Bulletin of the School of Oriental and African Studies : –. Ducroz, J.-M., and Charles, M.-C. (). Lexique soŋey (Songay)–français: Parler kaado du Gorouol. Paris: L’Harmattan. Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University Press. Hanafiou, H. S. (). Eléments de description du kaado d’Ayorou-Goungokore (parler songhay du Niger). Ph.D. thesis. Grenoble: Université Stendhal. Heath, J. (a). A Grammar of Koyraboro (Koroboro) Senni: The Songhay of Gao, Mali. Cologne: Rüdiger Köppe. Heath, J. (b). A Grammar of Koyra Chiini: The Songhay of Timbuktu. Berlin and New York: Mouton de Gruyter. Heath, J. (). Hassaniya Arabic (Mali)-English–French Dictionary. Wiesbaden: Otto Harrassowitz. Heath, J. (). Tondi Songway Kiini (Songhay, Mali): Reference grammar and TSK-English–French dictionary. Stanford, CA: Center for the Study of Language and Information Publications. Heath, J. (). Dictionnaire touareg du Mali: Tamachek–anglais–français. Paris: Karthala. Heath, J. (). Humburi Senni Grammar, https://deepblue.lib.umich.edu/bitstream/handle/ .//Grammar%of%Humburi%Senni%downsized.pdf?sequence=&isAllowed=y accessed September , . Kossmann, M. (). ‘Grammatical borrowing in Tasawaq’, in Y. Matras and J. Sakel (eds.), Grammatical Borrowing in Cross-Linguistic Perspective. Berlin and New York: Mouton de Gruyter, –. Kossmann, M. (). ‘Adjectives in Northern Songhay’. Afrika und Übersee : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Lacroix, P. F. (). ‘L’ensemble songhay-jerma: Problèmes et thèmes de travail’, in M. Houis (ed.), Actes du e congrès international de linguistique africaine, Abidjan, – mars . Abidjan: Université d’Abidjan, –. Lacroix, P. F. (). ‘Emghedesie “Songhay Language of Agades” à travers les documents de Barth’, in E.R.A. du CNRS ‘Langage et Culture en Afrique de l’Ouest’ (ed.), Itinérances—en pays peul et ailleurs: mélanges. Paris: Société des Africanistes. Nicolaï, R. (). Les dialectes du songhay: Contribution à l’étude des changements linguistiques. Paris: SELAF—Peeters. Nicolaï, R. (). Parentés linguistiques (à propos du songhay). Paris: Éditions du CNRS. Nicolaï, R. (). La force des choses, ou, l’épreuve ‘nilo-saharienne’: Questions sur les reconstructions archéologiques et l’évolution des langues. Cologne: Rüdiger Köppe. Prost, A. (). La langue soṅay et ses dialectes. Dakar: IFAN. Sibomana, L. (). Le zarma parlé: Esquisse grammaticale, lexique, textes. Berlin and Münster: LIT. Skinner, N. (). Hausa Comparative Dictionary. Cologne: Rüdiger Köppe. Souag, L. (). Grammatical Contact in the Sahara: Arabic, Berber, and Songhay in Tabelbala and Siwa. Ph.D. thesis. School of Oriental and African Studies, University of London. Souag, L. (). ‘The subclassification of Songhay and its historical implications’, Journal of African Languages and Linguistics : –. Souag, L. (a). ‘Explaining Korandjé: Language contact, plantations, and the trans-Saharan trade’, Journal of Pidgin and Creole Linguistics : –. Souag, L. (b). ‘Non-Tuareg Berber and the genesis of Nomadic Northern Songhay’, Journal of African Languages and Linguistics : –. Tersis-Surugue, N. (). Le dendi (Niger): Phonologie, lexique dendi–français, emprunts (arabe, hausa, français, anglais). Paris: SELAF. Tersis-Surugue, N. (). Économie d’un système: Unités et relations syntaxiques en zarma (Niger). Paris: SELAF—Peeters. Zima, P. (). Lexique dendi (songhay): Djougou, Bénin: avec un index français–dendi. Cologne: Rüdiger Köppe.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. C (Tçara, Cárádàm) is an endangered, little documented Central Khoisan (Khoe) language of the East Kalahari sub-branch (Tshwa subgroup), spoken by an unknown, yet certainly small number of people in and around Gweta, Nata, and Mopipi (Central District, Botswana).
. P
..................................................................................................................................
.. Vowels Cara has five oral and three nasal vowels, as shown in Table .. Within the given disyllabic word structures CV–CV, CV–V, and CV–N, oral vowels are not constrained for positioning, whereas nasal vowels can occur in CVV roots only. There are also nine oral (/ai ae ao au oe oa ui ue ua/) and four nasal vowel sequences (/ãĩ ãũ õã ũĩ/), which are restricted to CVV roots. The back mid-vowel < õ > appears only in sequences.
.. Consonants The consonantal inventory consists of thirty-one egressive (non-click) and nineteen ingressive (click) phonemes. Egressive consonants distribute over seven points and four manners of articulation, as shown in Table .. Only /p t b g s h m y/ can occur in C and C position; other consonants are restricted to C. In intervocalic position, /d/ and /b/ are represented by their allophones [r] and [β] respectively. Free variation exists between /dz/ and [z] and between /c/ and [ky] as well. There are also six foreign phonemes (not entered in Table .): /l mb ŋk ŋg/ were found to appear word-medially, and /nd w/ word-initially.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Cara vowel phonemes oral
nasal ĩ
u
i e
ũ
o a
ã
Table . Egressive consonants in Cara Bilabial
Alveolar
Post-alv.
Palatal
Velar
Uvular
Glottal
p
t
ts
c
k
q
ʔ
th
tsh
ch
kh
d
dz
j
g
Voiceless Aspirated Voiced
b
ʔy
Implosive Vcl. uvul. fricative
tx
tsx
cx
Ejective
t’
ts’
c’
k’
ɲ
ŋ
Voiceless
s m
x
n
y
The click phonemes result from combinations of basic click types (primary articulation) and click accompaniments (secondary articulation). There are four basic types and eight accompaniments (Table .).
Table . Cara click inventory
Dental
Alveolar
Palatal
Lateral
ǀ
ǃ
ǂ
ǁ
Voiceless
ǀ
ǃ
Voiced
ǀg
Nasalized-voiced
nǀg
Prenasalized
nǀn
Voiceless uvular
ǀq
Voiceless uvular fricative
ǀx
Aspirated
ǀh
Glottalized
ǀ’
ǁ ǁg nǁg
nǃn
nǁn ǁx
ǃh
ǁh ǂ’
ǁ’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
The inventory is largely asymmetric as the fricative dental and lateral basic types exceed the alveolar and palatal stop series. This unbalance is due to click loss and replacement processes that have affected East Kalahari Khoe languages to an extreme extent (cf. Traill and Vossen ). In my database, none of the four click stops listed in Table . occurs in more than two lexical items.
.. Tones The tonal system has been examined to some degree, with a focus on the verb. L–L, H–H, H–L, and L–H sequences were observed to exist on the surface. An in-depth study of more than Cara verbs—simple, extended, as well as compound forms—resulted in establishing eight (largely evenly balanced) tone classes, whereby each of the above-mentioned sequences represents two classes. The classes differ from one another in their conjugational tone patterns (see Vossen : – for some details).
. M
.................................................................................................................................. Cara is still a morphologically rich language although it has—like all East Kalahari sister languages—commenced suffering from attrition. Both inflection and derivation as major word formation processes are observed for verbs but not for nouns; the latter are optionally inflected, though.
.. Nominals ... Gender and number Nouns are inflected for gender and number. There are three genders (masculine, feminine, and common) and three numbers (singular, dual, and plural). These are expressed on the nominal stem by means of combined enclitics, so-called person–gender–number (PGN) markers (see Table .). The use of such markers is not obligatory; they can be, and in fact mostly are, omitted if the context so permits. Word categories which are syntactically dependent on the governing noun formally agree with it. Such categories are adjectives, numerals, demonstratives, possessives, and associatives. However, PGN marking on dependent categories is largely optional. The use of PGN enclitics affects both animate and inanimate nouns. On animate nouns it denotes natural sex, e.g. khóé-nà ‘people’ (common plural); on inanimate nouns any PGN
Table . Cara PGN enclitics Singular
Dual
Plural
Masculine
-mà
-dzàrà
-kùà
Feminine
-sà
-sàrà
-dzì
Common
–
-khòrà
-nà
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
marker may principally be applicable even though gender is fixed. However, a systematic analysis of PGN semantics is still to be carried out.
... Subject–object relation Cara lacks case as a grammatical category but is provided with a couple of structural elements to express inter-nominal relationship such as subject–object relations. Hence, an object may be (but is mostly not) marked by the suffix -’à or -à, while the subject remains unmarked. This phenomenon has been referred to as differential object marking by Dimmendaal (p.c.). Another (non-morphological) means of indicating object function is word order (see section ..).
... Associative particle dì In associative (“genitive”) constructions a particle dì is sometimes employed between possessed and possessor noun, as in (a). Its use seems to trigger obligatory suffixation of an element -kà, whose meaning is anything but clear, on the possessed noun. As a rule, associative constructions make use of juxtaposition, though, as in (b). ()
a. k’áró dì
ǀxá-’ò-kà
b. k’áró ǀxá-’ò boy body-(-?) ‘a/the boy’s body’
.. Adjectives Adjectival concepts are often conveyed by verbal adjectives of the “to be X”-type. Proper adjectives are few (for examples, see Vossen : ). Predicative adjectives follow (a) and attributive ones precede (b) the head noun. PGN marking is very rare. ()
a. ìí ǁháó ǀ’áǹ ’è. : hoe bad ‘This hoe is bad.’ b. tá’à t’úĩ k’òhú khàà! : nice meat give ‘Give me nice meat!’
.. Numerals As a rule, Cara speakers count up to four; see (a). Numerals above ‘four’ tend to be descriptive and are hardly ever used; for an example, see (b). ()
a. ‘one’ ‘two’ ‘three’ ‘four’
ǀúí ǀám̀ ǁ’óbé hǎtsá
b. ‘five’
ǀúí-tshàù (‘one-hand’)
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Ordinal numbers are partly based on cardinals, as shown in the “lion” paradigm in (); they precede the head noun: ()
k’áí-ǀxà xam tá’à forehead-body lion : ǀám-ǁ’àì tsá’à two-? .: ǁ’óbe-ǁ’àì ʔémà’à three-? .: hǎtsa-ǁ’àì four-?
‘The first lion saw me.’
múũ-á-ha. see--
‘The second lion saw you.’ ‘The third lion saw him.’ ǁgàì-à-tá. ‘The fourth lion ran away.’ run--:
.. Pronominal forms Cara has five types of pronominal forms: personal subject and object pronouns, demonstratives, possessives, and interrogatives. There are no relative pronouns.
... Personal pronouns Both subject and object forms occur exclusively as independent pronouns. They are formally interrelated with the nominal PGN markers (third person forms; see section ... above) and are basically identical except that object pronouns take on the general object marker -’à (section ...). Their syntactic positions are entirely in line with the general S–O–V order of chief constituents. Hence, object pronouns are placed between the subject and the verb/predicate (cf., e.g., the examples in () above). The subject paradigm is given in Table ..
Table . The personal (subject) pronouns of Cara Singular Masculine
Feminine
tá
tá
tsá
sá
ʔé.mà
ʔé.sà
Common
ʔí
Dual
dzáḿ
sáḿ
kháḿ
dzáró
sáró
kháró
ʔé.dzàrà
ʔé.sàrà
ʔé.khòrà
Plural
jé
sé
dzé
gáó
sáó
dó
ʔé.kùà
ʔé.dzì
ʔé.nà
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
In all genders and numbers the third person base forms are prefixed by an element ’é-, whose origin is unknown. The . pronoun is indefinite and may best be translated as ‘it’.
... Demonstratives Cara has two demonstratives: proximal ìí ‘this’ and distal ùú ‘that’. As nominal qualifiers they can be employed with or without PGN markers, as in (a). When used independently, they must be nominalized by a PGN marker throughout (b). ()
a. ìí(-dzì) :(-.) ùú(-kùà) :(-.)
khóé-dzì ‘these women’ person-. khóé-kùà ‘those men’ person-.
b. ìí-khòrà :-.
‘these ones’
... Possessives Strictly speaking, there is no morphological category of possessive pronouns in its own right. In order to express pronominal possession two strategies are possible: (i) juxtaposition of the personal pronoun and the noun referent (a), which is the preferred variant; or (ii) employment of the associative marker di (cf. section ...) in combination with the personal pronoun. (In this context, however, the first and third person pronouns undergo formal changes: tá ! tí, . ’é.mà ! ’é.m̀ , . ’é.sà ! ’é.sì.) Personal pronoun and associative marker can alternatively be placed before or after the noun referent without any perceptible semantic change (b). For reasons of agreement, the associative marker may be PGN-marked (c), in which case the possessive phrase must follow the noun referent. ()
a. tí kai
‘my house’
b. kàì tí dì tí di kàí c. kàí(-mà) tí dì-mà house(-.) -.
... Interrogatives Basically, there are two interrogatives: má (alternatively ná) ‘who, which?’ and ndú ‘what?’. má and ná may also combine and they can optionally be PGN-marked, as in (a), whereas ndú is never marked for agreement (b). In possessive contexts (‘whose?’), má(.ná) follows the same strategies as possessives (see (i) and (ii) in section ... above), as in (c). Interrogatives can be placed sentence-initially (a), sentence-medially (d), or sentence-finally (b). ()
a. má ǁga-khoe ùú ’è? who female-person : ‘Who is that woman?’ má.na-’è tá’à ke k’ãĩ? who- : laugh ‘Who is laughing at me?’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
b.
ìí xùù ndú ’è? : thing what ‘What is this (thing)?’
c.
má.n jua-’è tsá kè káá? who stick- . want ‘Whose stick do you (.) want?’ ìí jíbe má.n dì ’ì? : axe who ‘Whose axe is this?’
d.
tsá ke ndú kàà? . what want ‘What do you want?’
.. Verbals ... Structure of finite verbs There appear to be two structural sequences of the finite verb in Cara whose application depends on tense. In present and future, the following order (structure type A) holds: {tense | negation} – verb + derivative suffix(es) + passive + negation {} tense. “+” indicates affixation, “–” the positioning of free morphemes in a given verbal phrase. Of the six structural slots maximally five can simultaneously be filled (–, –, or – plus ). Positions (tense only) and alternate. However, if is filled by the negation marker and by negative present, future meaning is implied, as in (). ()
’é.nà tátánà ǀíí nǁgáì-tà. . song sing-: ‘They will not sing a song.’
As becomes obvious from this example, the sequence of verbal elements can be interrupted by other parts of speech, in this case a nominal object. If negation fills slot , future is again the logical consequence, as negative present is a tense in its own right. Here, the future marker may occur either in or . ()
tá kùà híĩ-m̀. do- ‘I shall not do it.’
In (), examples are given for the employment of affirmative present (a) and future (b). In the latter sentence, the object ‘porridge’ is topicalized.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() a. tsá ke ’yánù ’yˇuũ. . porridge eat ‘You (are) eat(ing) porridge.’ b. ’yánù ’yˇuũ kùà tsá. porridge eat . ‘You will (be) eat(ing) porridge.’ Structure type B is used when the predicate occurs in negative present or one of the available past tenses. The sequential order is: verb + derivative suffix(es) + negation + {passive | juncture} + tense/aspect. All slots can be filled simultaneously. The juncture in represents active constructions. In (a), all five positions are filled; in (b) slots , , and . ()
a. khóé ǀxòà tá ǁ’àé-kú-tòm-à-ha. person with meet---- ‘I have met nobody.’ b. tsá ǁgaba ts’áã-e-há. . bow steal-- ‘Your bow was stolen.’
... Verbal extensions Verbs can be derived from verbs by means of suffixes which follow the verbal base immediately. As a rule, derivative suffixes modify the semantic content of the verbal base and may alter the valence and, hence, the argument structure of the derived verb. In other words, they impact on both the semantic and the syntactical level. Nine functions have been found so far to exist in Cara: two causatives, iterative, dative, or applied form, reflexive, reciprocal form, directional–locative, terminative–itive, and intentional form; they are all more or less productive. Examples are listed in () below. One of the two causatives is formed by reduplication, rather than a specific suffix, but reduplication also stands for the iterative function. In fact, for some verbs the semantics of these two functions are fairly closely related so that they could be part of the same derivational pattern. Sometimes, causative extensions make intransitive verbs become transitive, as the word pairs ‘recover | cure’ (causative I) and ‘move away | move sth./sb.’ (causative II) below illustrate. The reverse is the case when the reflexive or reciprocal extension is applied. The intentional suffix cannot occur in finite verb forms. ()
I: -kahu sáã ‘recover’ > sáã-káhu ’yˇuũ ‘eat’ > ’yùũ-káhú II: reduplication tòé ‘move away’ > tòé-tóé ǀ’áé ‘fall’ > ǀ’áé-ǀ’àè
‘cure’ ‘feed’ ‘move sth./sb.’ ‘let fall, drop’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
: reduplication nǁgáì ‘sing’ > nǁgáì-nǁgàì ǁ’áḿ ‘beat’ > ǁáḿ-ǁáḿ / : -ma séè ‘take, seize’ > séè-má nǀgóá ‘cook’ > nǀgóá-ma : -sin ǀ’ˇuũ ‘kill’ > ǀ’ˇuũ-sin kóḿ ‘hear’ > kóḿ-sin : -ku k’ùí ‘speak’ > k’ùí-kú !hùú ‘push’ > !hùú-kú –: -’o táó ‘pound’ > táó-’o k’áà ‘drink’ > k’áò-’o –: -hu khǒã ‘break’ > khǒã-hu ’úé ‘crack, smash’ > ’úé-hu : -a c’èé ‘cry’ > c’èé-a gàḿ ‘throw’ > gàḿ-a
‘sing continuously’ ‘beat repeatedly’ ‘take for, seize for’ ‘cook for’ ‘commit suicide’ ‘hear oneself ’ ‘quarrel’ ‘push each other’ ‘pound in sth.’ ‘drink from sth.’ ‘break off ’ ‘crack open’ ‘in order to cry’ ‘in order to throw’
Verbal extensions can also be combined in various ways. In (a), the dative extension is followed by the reflexive form, whereas in (b) we have a sequence of causative I plus the reciprocal form. (c) offers an example of three combined extensions, i.e. directionallocative plus causative I plus reflexive. The order of extensions in such complex derivative constructions appears to be generally fixed. ()
a. nǀgóá ‘cook’ > nǀgóá-má-sin ‘cook for oneself ’ b. kóḿ ‘hear’ > kóḿ-kàhu-ku ‘listen to one another’ c. k’áà ‘drink’ > k'áà-’o-kàhú-sin ‘make oneself drink from sth.’
... Juncture When constructing finite verbs, Cara employs a certain element that links the verbal base to the following tense/aspect marker. This element was first discovered by Köhler (), who called it “juncture”. Recently, a more systematical study (for the whole Central Khoisan family) was published by Vossen (), who preferred to use the term “linker” for this element. In Cara, the juncture is used only with verbs that are marked for past tenses. Its morphological base is /-a-/ with two allomorphs, [-ra-] and [-na-], that are largely conditioned environmentally. The use of the base form is confined to verb stems with a final nasal consonant or any vowel except /ã/. If the verb stem ends in /ã/, this vowel becomes oral (i.e. [a]) and the nasality is shifted to the juncture in the shape of a nasal consonant; hence, . . . ã+a-! . . . a-na-. The allomorph [ra] is employed on verbs that end in /a/, but
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
there are also several counterexamples. The case is not entirely clear yet. For the time being, one may regard its use as verb-specific. CVCV and polysyllabic verb stems never accept the juncture. From a paradigmatic perspective, however, one may speak of a zero juncture morpheme. The appearance of the juncture can be discerned in a number of examples throughout this chapter.
... Passive The passive morpheme is /-e-/. It appears between the verb stem and the tense/aspect marker. For an example, see (b).
... Tense and aspect On the basis of the available data, a sharp distinction can hardly be made between the two categories, neither morphosyntactically nor semantically. As a rule, TA markers cannot cooccur. Certainly, there are ways to distinguish anteriority from present and posteriority. While present (ke) and future (kùà) are expressed by free morphemes, the two past tense markers (recent -tá and “indefinite” -ha) are suffixed to the verb. Negated present (-tà), which is also suffixed, is a tense form in its own right. Examples: () a. tá ke b. tá kua c. tá d. tá e. tá
xúnú. snore xúnú. xúnú-tá. xúnú-ha. xúnú-tà.
‘I (am) snore/(snoring).’
(present)
‘I shall snore.’ ‘I snored (e.g. yesterday).’ ‘I (have) snored.’ ‘I do/(am) not snore/(snoring).’
(future) (recent past) (“indefinite” past) (present negative)
... Mood/modality Imperatives are expressed by the bare verb stem in both singular and plural, e.g. nǁgáì ! ‘sing!’ (/). In the negative, the particle tí follows the verb stem: nǁgái tí ! ‘Don’t sing!’ There is no category subjunctive or conjunctive. Modality is indicated as shown in (): ()
a. jé kua ndú nǁgáì? . what sing ‘What shall we (.) sing?’ b. a tá kúá ‘May I sing?’
nǁgáì rè? sing
c. tá tsá ǀxoà hàá káá-tà. . with come want-: ‘I cannot come with you (.).’ d. tá ke ’áé kà nyúú. stay at home ‘I must stay at home.’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Evidently, tense markers (a–b/d) and modal verbs like ‘want’ (c) are employed to describe modal notions such as ‘shall’, ‘may’, ‘can’, or ‘must’.
... Negation Verbal negation is basically dependent on tense/aspect marking. In the present, a specific suffix -tà is used, as can be seen in (e). For the remaining TAs, a considerable degree of variability obtains. Both past forms can be negated by the compound element -ma.na- which is placed between the verb stem and the TA suffix: () a. tá k’ò-hú k’òó-mà.nà-tá. eat.meat-thing eat.meat--: ‘(Today) I have not (yet) eaten meat.’ b.
k’òó-má.ná-ha. eat.meat-- ‘I ate/have eaten no meat.’
However, there also exist separate means of expressing negation: -tòm̀.tám̀ for recent and -tom- for “indefinite” past. The former presents itself like a tense form in its own right, comparable to the present negative. Whether or not this is true cannot be decided on the basis of the available data. The marker -tom- can, but need not, be followed by the juncture. c. tá k’ò-hú k’òó-tòm̀ .tám̀ . eat.meat-thing eat.meat--: ‘(Today) I have not (yet) eaten meat.’ d. k’òó-tòm̀(-á)-ha. eat.meat-- ‘I ate/have eaten no meat.’ There are two strategies for negating future tense. On the one hand, the trisyllabic element tátánà, possibly a compound form, can be employed in combination with the negative present suffix -tà, as in (a). If, however, the future marker kùà is used, the negation marker -m̀ must be suffixed to the verb stem (b). Which one of the two strategies is more common remains an open question. ()
a. ’é.nà tátánà ǀíí nǁgáì-tà. . song sing-: ‘They will not sing a song.’ b. tá kùà híĩ-m̀ . do- ‘I will not do (it).’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. S
..................................................................................................................................
.. Word order The basic order of chief constituents appears to be S–O–V. The majority of examples in my database represent this sequence structure. In (a) the object is placed between the tense/ aspect marker and the verb remains morphologically unmarked. Occasionally, the subject occurs sentence-initially and the object follows the verb, hence S–V–O. Although (b) is an interrogative sentence, S–V–O order does not seem to be restricted to questions. ()
a. khóé-nà ké kàí ǀgùì. person-. house thatch ‘The people are thatching the/a house.’ b. a tsá ké múũ-re ìí k’áḿ-ǀge? . see- : animal ‘Do you (.) see this animal?’
As to the positioning of adverbials no restrictions have been observed. () gives two examples of locative adverbs that occur in pre-final position, immediately before the verb and after the subject. In (a) the order of constituents is O–S–Adv–V, while (b) displays S–Aux–Adv–V. ()
a. má-n ǀxòà tsá ’áé kà ǁ’àé-kù-à-tá? who-. with . home at meet---: ‘Whom have you (.) (just) met at the village?’ b. ǁgàà-khóé ke gùnu ’ò ’á táo. female-person mortar in pound ‘The woman is grinding in the mortar.’
A somewhat different picture obtains for temporal adverbials. In (a) the adverbial is placed sentence-initially, followed by the chief constituents in S–O–V order. In the other two examples (b/c) the adverbial follows the subject and precedes the object and the verb; hence: S–(Aux–)Adv–O–V. () a. ’ókuri tá ’aká súrì thíyá gòm̀-à-tá. last.year tobacco much smoke--: ‘Last year I smoked a lot of tobacco.’ b. a tsá ’úíka súrì xúń-á-hà rè? . yesterday tobacco grind-- ‘Did you (.) grind the tobacco yesterday?’ c. tá kùà ’úwá ’áíyá-khòè sĩĩ-a-ma. tomorrow chief work-- ‘I shall work for the chief tomorrow.’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Simple sentences ... Declaratives In (), examples of copulative constructions are presented. They show that the affirmative copula is ’è, which has been assimilated to the final vowel of the preceding adjective in (d). (c) shows the use of the negative copula m̀. ()
a. ìí tí ǀ’on ’è. : : name ‘This is my name.’ b. tá Cárákhòè ’è. Cara.person ‘I am Carakhoe.’ c. tá Shúákhòè m̀. Shua.person : ‘I am not Shuakhoe.’ d. jé ǁara gùrí ’ì. . field large ‘Our (.) field is large.’
In the following sentence a paraphrastic ‘have’-construction is contained. Note that hàã ‘exist’ is a defective verb: () ǀúm-khoe thíyá bee ǀxoa hàã. chief many cattle with exist ‘The chief has many cattle.’ In negative constructions (), hàã is replaced by the regular verb ’áí ‘to not have’ followed by the negative present marker -tà: () ǀàó ǀ’óã ’ai-tà. snake bone not.exist-: ‘A snake has no legs.’
... Questions Since wh-questions have already been discussed in the section on interrogatives (section ...), the present section will exclusively be devoted to yes/no-questions. These tend to commence with the question marker à and end with the interrogative particle rè, as in (). () a. à ìí màà tí di ri? : head : ‘Is this head mine?’
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
b. à tsá ke dànì ǀóã ’úũ-re-a-má rè? . honey child collect--- ‘Are you (.) collecting honey for (your) child(ren)?’ Note that in (a) the interrogative particle would seem to imply copulative function. Its shape rì has undergone progressive assimilation to the final vowel of the preceding associative marker di. In the following examples the interrogative particle occurs sentence-medially. The two versions in () are said to be interchangeable. () a. à tsá ke kóḿ . hear
rè tí ǀ’on? : name
b. à kóḿ ke re tsa tí ǀ’on? hear . : name ‘Do you (.) hear my name?’ Double occurrence of the interrogative particle is observed in (). The fact that it shows up after both predicative adjectives may be further evidence for the assumption of the particle’s inherent copulative power. () à tsá ǁáó-á-ha ǀáò gùrí rè re cúĩ rè? . shoot-- buffalo big small ‘Was the buffalo that you (.) shot big or small?’
.. Complex sentences ... Coordination and subordination Given the scantiness of data, it seems difficult to decide whether there exists a clear-cut dividing line between coordination and subordination. Coordinate clauses are mostly not linked by conjunctions but, rather, occur in juxtaposition, as in (). () a. k’áá-khoe ké ’é.ku ǀôã ǀxòà gùnì ’ó kûũ, male-person .: child with hunt go ǁgaa-khóé kúk’ari ké ’é.dzi ǀôã ǀxòà ’úũ-re ’ò kûũ. female-person all .: child with gather- go ‘The men go hunting with their sons (and) the women go gathering with their daughters.’ b. tí ǁàbò k’òá ’è, tí kũĩ-k’áì ǁàbò ǁnáó-ha. : sandal new : sister sandal be.old- ‘My sandals are new, (but) my sister’s sandals are old.’ The sentence in () has a subordinate structure in English but is coordinate in Cara:
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
() ’é.mà kúũ-á-ha tóã-m ǀxòá k’òò-hú . go-- friend-. with meat ’úíka, ’è.m ǀ’ùũ-á-ha.e ’áé kúũ-á-ha. yesterday . kill-- home go-- ‘Yesterday he went out with a friend (to) carry the hunted meat home.’ Example () represents a subordinate participle construction. The verb in the dependent clause is always followed by the participle clitic si: () à tsá k’áró múũ-á-ha . boy see-- k’áró ke k’òò-hú ts’âã boy meat steal ‘Did you see the boy stealing meat?’
re sì.
Note that the object of the main clause (k’áró ) is resumed preverbally in the dependent clause (as agent of the participle phrase). Finally, () illustrates an intriguing instance of clause-embedding: () tá ke tí cahu-m dànì ká páá-é-hà k’àĩ. : brother-. bee bite-- laugh ‘I am laughing (because) my brother has been stung by a bee.’ Literally, the sentence reads: ‘I am | my brother by a bee has been bitten | laughing.’ That is, the embedded clause (representing the subordinate clause) breaks up the verb phrase of the main clause.
... Relative clauses Cara has no relative pronouns. There is, however, a post-verbal particle e (with varying tone) which is used in relative clause-like constructions, as in (). ()
a. màí tsá tá’à khàà-à-tá é kùrí-ha. egg . : give-- finish- ‘The eggs which you gave me are finished.’ b. tá’à tséé-mà k’áà-khòè xóḿ ’àã e! : send- male-person land know ‘Send me a man who knows the area!’
R Köhler, O. (). ‘Les langues khoïsan’, in J. Perrot (ed.), Les langues dans le monde ancien et moderne. Part I: Les langues de l’Afrique subsaharienne, edited by G. Manessy. Paris: Éditions du CNRS, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Traill, A., and Vossen, R. (). ‘Sound change in the Khoisan languages: new data on click loss and click replacement’, Journal of African Languages and Linguistics : –. Vossen, R. (). Die Khoe-Sprachen: Ein Beitrag zur Erforschung der Sprachgeschichte Afrikas. Cologne: Rüdiger Köppe. Vossen, R. (). ‘The verbal “linker” in Central Khoisan (Khoe) in the context of verbal derivation’, Journal of Asian and African Studies : –. Vossen, R. (). The Khoesan Languages. London and New York: Routledge.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
ǁ X ’ ......................................................................................................................
. I
.................................................................................................................................. ǁX’, the furthest east of the !Ui languages, was spoken in the Lake Chrissie area of the Eastern Transvaal at the time it was studied. Güldemann () classifies it as the outlier branch of the Tuu family. The ǁX’egwi language has been studied in the field five times, but the only description of ǁX’egwi culture is a short monograph by Potgieter (). The earliest data are from Dorothea Bleek’s Comparative Vocabularies () where ǁX’egwi is cited as S. This is a short vocabulary of fewer than items. Two subsequent studies, Ziervogel () and Lanham and Hallowes (), have been published. The material collected by Westphal (n.d.) and Köhler (n.d.) is available only in field notes and has never been published. In Traill (), there is a recording of main ǁX’egwi informant Jopi Mabinda telling a folktale. There are many disagreements between the five researchers, but if all the accounts are taken together, a consistent picture emerges. The present chapter presents the most likely picture of ǁX’egwi based on a syncretism of all five studies. The language is variously referred to in the literature as “Batwa”, “Eastern Bushman”, “ǁXegwi Kl’egwi”, “ǁX’egwi”. I have adopted the last, the transcription used by Westphal and Köhler in the two most recent studies.
. P
.................................................................................................................................. The major disagreement on the phonological system of ǁX’egwi concerns the status of the pulmonic consonants which Lanham and Hallowes identify as uvulars. Westphal () records these as affricated velars and Ziervogel as palatals. It is clear at least that there were two contrasting series, since all the researchers transcribe the “uvulars” differently (Table .).
.. Vowels ǁX’egwi had five oral vowels (a, o, u, e, i ) and three nasal vowels, a typical Khoisan system. As in other Khoisan languages, the low back vowel is centralized before a high back vowel (Lanham and Hallowes : }), either following or in the next syllable.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ǁX ’
Table . ǁX’egwi lexemes with “uvular” consonants Gloss
Lanham and Hallowes
Ziervogel
Westphal
Köhler
‘heart’
keleŋ
kele
keleŋ
kèlèŋ
‘hole’
gumi
guːmi
gɦumi
gˇuːmì
‘throat’
qhobo
kxobo
kxhoobo
khòːßò
‘woman’
q’iŋ
gyëiŋ [’gɤyəiŋ]
gɤəĩ
gèn͡y, kx’èŋ
Table . The vocalic system of ǁX’egwi Oral vowels u o
i e a
Nasal vowels ũ
ĩ ã
Pharyngealized vowels uˀu, ũˀũ iˀi, ĩˀ ĩ oˀo eˀe aˀa, ãˀã
Lanham and Hallowes distinguish five nasal vowels, but they give only one example each of ẽ (ʔunẽẽ ‘food’)1 and õ (ʘõõ ‘son’). Westphal has these as ’unẽŋ ~ ’uni and ʘoŋ and, indeed, the pervasive variation between a final nasal vowel and a sequence -Vŋ makes the status of nasal vowels generally unclear. All of the South African Khoisan languages outside Khoe have a contrast between modal and pharyngealized vowels.2 In the case of ǁX’egwi, none of the researchers records pharyngealization as such, but all record sequences of vowels interrupted by a glottal stop. Traill () transcribes these as strongly pharyngealized, and some of them correspond to Nǀuu words with pharyngealized vowels: ǁX’egwi t’a’àne ‘walk’, Nǀuu ɟaʕn, ǁX’egwi gɦa’a ‘sky’, Nǀuu ǃaˁa, suggesting that ǁX’egwi also had a modal/non-modal contrast of some kind in the vowels (Table .). The following vowel sequences have been observed: ai /ei, eu, oe, ui, ẽũ, ũĩ. ǁX’egwi has a strong tendency to level the vowels in a stem; about % of disyllables have like vowels. The same tendency applies to vowel sequences, e.g. ǁX’eu-gwi > ǁX’egwi.
.. Consonants ǁX’egwi has a large phonemic inventory, with pulmonic consonants (non-clicks) (Table .) and lingual consonants (clicks) (Table .), a total of . The central alveolar affricate ʦ is marginal; l (as a contrasting phoneme) and f are restricted to loans. Tables . and . are based on a comparison of the data recorded by Ziervogel, Lanham and Hallowes, Westphal, and Köhler and may be incomplete. Lanham and Hallowes have additional contrasts, including a series of palatals recorded from some of their informants, labialized clicks and non-clicks, and four uvulars. 1 Probably borrowed from Khoe Shua or Tshwa subgroups: cf. Kua ʔyũˇ ‘eat’ + common plural suffix -ni (Vossen : ). 2 !Xóõ and Juǀ’hoan have additional contrasts of breathiness, laryngealization, etc.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . ǁX’egwi pulmonic consonants Bilabial
Alveolar Central
Stop
p, pʰ, b
Palato-alveolar
Velar
Uvular
Glottal
k, kʰ, g
q ɢ
ʔ
Lateral
t, tʰ, d
Affricate
(ʦ) ʣ, ʦ͡x, t͡x
k͡ɬʰ d͡l
ʧ, ʧʰ ʤ
Glottalic
ʦˀ
k͡ɬˀ
ʧˀ
k͡xˀ
ɲ
ŋ
ʃ, ʒ
x
Nasal
m
n
Fricative
(f ), v
s, z
ɬ
ɾ
(l)
Liquid
h
Table . ǁX’egwi lingual consonants Labial
Dental
Alveolar Central
Lateral
Stop
ʘ
ǀ, ǀʰ, ᵍǀ
ǃ, ǃʰ, ᵍǃ
ǁ, ǁʰ, ᵍǁ
Nasal
ᵑ̥ʘʰ, ᵑʘ
ᵑ̥ǀʰ, ǀˀ, ᵑǀ
ᵑ̥ǃʰ, !ˀ, ᵑ!
ᵑ̥ǁʰ, ǁˀ, ᵑǁ
ǀ͡x
ǃ͡x
ǁ͡x
ǀ͡xˀ
!͡xˀ
ǁ͡xˀ
Affricate Linguo-glottalic affricate
ʘ͡xˀ
The labialization on Cʷ, Qʷ can be reinterpreted as a rounded vowel following the consonant, CoV, QoV, to conform with the general phonological profile of South African Khoisan, and there is clear evidence for only two uvulars. Since all Khoisan languages with uvular non-clicks also have linguo-pulmonic clicks (i.e. clicks with an audible uvular release), ǁX’egwi probably did also, but there is no direct evidence for this.
.. Tone Little can be said about ǁX’egwi tone. Only Köhler marks tone consistently. On the basis of Köhler’s data, ǁX’egwi probably had a four-tone system like !Xóõ.
. M
.................................................................................................................................. ǁX’egwi structure is highly analytic with little inflection, mostly of the agglutinative type. There is evidence in ǁX’egwi for nine word classes: noun, pronoun, verb, adjective, adverb, qualifier, preposition, conjunction, and particle.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ǁX ’
.. Noun ǁX’egwi nouns are inflected only for number. The nominal number system is extremely complex, with at least fourteen classes. Examples of ǁX’egwi singular/plural classes are given in Table ., with data from Köhler and Westphal.
Table . Singular/plural classes in ǁX’egwi Gloss
Singular
Plural
Class ‘person’
K
Class ‘fat’
W
Class ‘father’
K
Class ‘locust’
W
Class ‘face’
W
-Ø xù
-ŋ xù-ŋ
Class ‘star’ ‘tree’
K K
-zi ǀ ̫ùnì-zì ʘ̃ mò-zì, ʘờ-zì
-Ø, -ŋ ǀ ̫ùnì, ǀúnĩ ̀ ʘ̃ mò-ŋ
Class ‘knee’
K
Ø, -ŋ ǀnòmàŋ
-le gè ǀnòmà-lè
Class ‘knife’
W
-be dhè-bè
-ŋ kí-dẽ̥-ŋ
Class ‘dog’
K
-ī klwèŋ
-mi k͡lúmĩ ̀
Class ‘nose’
K
-Ø ̃ ǀnù:
-nù ̃ ke ǀnúː-nù
Class ‘child’
K
-ri ʘà-rì, ʘà-rí
-ni ʘèː-nì
Class ‘foot’ ‘tooth’
K K
Ø ǀhĩ’ĩ́ ̀ ǁèŋ
gí . . . (ŋ) gè ǀhĩ’ĩ́ ̀ gè ǁèŋ
Class ‘crab’
K
kàːrà
ri-, lirìkàrà
Class ‘lion’
W
li(li)bʰuɓezi-zi
mamábhúɓezi
Suppletive tle
kwí
mass nouns, plural only swiĩ -Ø ’aː
-gũ ’àː-gù-ŋ singular = plural ǁxoo
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Like all non-Khoe Khoisan languages, ǁX’egwi has a number of nouns with suppletive plurals: ‘person’ kwii, tuŋ; ‘man’ tl’oo, tl’e; ‘woman’ gɤəĩ (Z gyëiŋ, L&H q’iŋ), gɤaŋ (Z gyaŋ, L&H qaŋ); ‘child’ ǁ’ele’e, tlho-ŋ; ‘girl’ chigaa, chiǀo; ‘hand’ kxhĩ̥ŋ (L&H qhii ), ǀx’’aŋ. None of the researchers discusses the mass noun/count noun distinction, but a number of nouns such as ‘fat’, ‘sand’, ‘flour’, etc., are recorded only in plural contexts. The most common plural marker is the suffix -ŋ, which often co-occurs with the plural prefix ge- discussed in the next section. Singular nouns are normally unmarked, but some take the suffix -zi, which may have been a singulative like Nǀuu -si (Lanham and Hallowes ). Like the Nǀuu suffix, it is frequently suffixed to loans: ‘table’ li-tafula-zi (< Zulu itafula); ‘Friday’ li-hlanu-zi (< Zulu -hlanu ‘five’). Remarkably, ǁX’egwi has borrowed a set of singular/plural prefixes from Bantu, which are sometimes used with loans from nonBantu languages: ‘joint’ lì-lùŋàzì (< Swazi ilunga); ‘plate’ (li)pleiti, mapleiti. ǁX’egwi has two possessive constructions; juxtaposition and use of a possessive marker ki, which both published sources treat as stylistic variants. However, a careful analysis of the examples in Lanham and Hallowes and Ziervogel, combined with the material in Westphal’s and Köhler’s unpublished notes, suggests that there is a semantic distinction between the two constructions. Part of the confusion arises from the fact that ǁX’egwi has three lexemes with an initial velar followed by a front vowel: a verb gi ‘to have’, a plural marker ge-, and a possessive marker ki. The forms given are from Köhler, but there is a considerable amount of variation, both among and within the sources. Almost % of the nouns used with the juxtaposition construction are body part nouns; the rest are typical alienable nouns such as ‘knife’, ‘house’, and ‘ancestor’. The order is possessor–possessed and the possessed noun normally takes the ge- plural. ()
ɲa-ŋ-gi-ǁkhi-ŋ.
:---tooth- ‘They are my teeth.’ ()
ɲa-ŋ-ge-!ũʔu.
:---bowel ‘They are my intestines.’ ()
()
(L&H)
(L&H)
hà ŋ́ ǁèŋ. : tooth ‘It is my tooth.’
(K)
nyá í gè ǁèŋ. : tooth ‘They are our teeth.’
(K)
In the second construction, the word order is reversed and the nouns are linked by the possessive marker ki: possessed + ki + possessor. ()
kha ki ’aye water you ‘your water’
(Z)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ǁX ’ ()
()
klhumiŋ gi ǀkxoŋ dog: European ‘dogs of the European’ ń tùːbì k͡ lwa ki-ǁx’éu k͡ le. hear affair people ‘I hear the people’s affair.’
(L&H)
(K)
.. Adjectives Adjectives form a distinct subclass of verbs in ǁX’egwi, characterized by the use of the aspect suffix -a in the present and the use of ge as a past marker. In an adjectival phrase, the adjective follows the noun and must be relativized. There seem to be two classes of adjectival verbs, depending on whether they take the aspect suffix in the present. Examples of the second class, without the suffix are: tšwe ‘short’ and kl’ini ‘small’. ()
()
ɦá ‘ŋ́ !’ele íŋ ǁ’ow-a. throat dry- ‘My throat is dry.’ gè ǁòːw-à. dry- ‘I was thirsty.’
ŋ́
() ń thùnìy-à. tired- ‘I’m tired.’ ()
()
()
(W)
ń gè thùnìy-à. tired- ‘I was tired.’
(K)
(K)
(K)
haa-kwii ʔela kl’ini. -:-person small ‘It’s a small person.’
(L&H)
ha-ge-kl’ini. --small ‘He used to be small.’
(L&H)
() haa giǁŋa ʔe !e-ya. -: thing red- ‘It’s something red.’
(L&H)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Numerals ǁX’egwi, like most Khoisan languages, probably lacked a full number system and had terms for only through : ǃwaa ‘one’, c’uu ~ k’uu ‘two’, gwana ‘three’. Like other nominal modifiers, the numerals follow the noun and are linked to it by a relative suffix. At the time Köhler recorded his data, ǁX’egwi had filled out the system with borrowings from Bantu, but Köhler is the only researcher to record all the terms from one to ten and it is not clear exactly how the larger numbers were used in practice. Among these higher numbers, ‘four’ is interesting being constructed of ‘two’ plus a plural suffix c’uu-ŋ. haa-!kxwa ʔena !waa. :-cow one ‘It’s one cow.’
()
(L&H)
() ɲa-kl’e ʔena gwana. :-people three ‘It’s three people.’ ()
(L&H)
ŋ́
ǀnì ǀxá-ŋ é lá gwàːnà. see: hand- three ‘I see three hands.’
(K)
.. Pronouns ... Personal pronouns ǁX’egwi had several different kinds of pronominal elements (see Table .); roughly, these divide into simple clitic-like elements closely associated with a nominal or verbal head and complex pronouns composed of a base and a recurring suffix. Lanham and Hallowes refer to the latter as “selfstanding pronouns”. All of the ǁX’egwi selfstanding pronouns have a suffix -ʔe which is dropped in the subjectival and possessive clitics and absorbed into the base in the objectival pronominal clitics. The subject markers, also used as possessives, precede the verb; the object pronouns
Table . ǁX’egwi pronominal system per Lanham and Hallowes () Selfstanding
Subject
Possessive
Object
Singular
ʔiŋ-ʔe
iŋ-
ŋ-
-ne
Singular
ʔa-ʔe
aŋ-, a-
ˀa-
-ae
Singular
ha-ʔe
haŋ-, ha-
ˀa-
-yaˀe, -waˀe
Plural
ʔi-ʔe
i-
i-
-ye
Plural
ʔu-ʔe
uŋ-, u-
u-
-we
Plural
hiŋ-ʔe
hi-, ŋ-
?
-ŋʔe, -ʔŋ
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ǁX ’
follow. The exact status of these markers is, of course, no longer determinable. Lanham and Hallowes write them joined to their head; the other sources all write them separate. Lanham and Hallowes use the term “clitic”. The examples given in the two published papers suggest that these series interacted in a complex way not easy to explain from the evidence presented. Both Ziervogel and Lanham and Hallowes remark on the copious use of pronouns. Ziervogel suggests that this type of construction serves to disambiguate situations where the verbal pronouns are too similar (Ziervogel ). In the third person, the forms with initial h- are the most common but there are additional third person forms with an initial affricate, though it is not clear how these contrast with the h- forms. Lanham and Hallowes (: ) note that they function to reduce confusion between first person ’iŋ-e and third person hiŋ-’e.3 ()
ha-a-ŋʔe. --I ‘It is me.’
() dzi-ya-ŋʔe. --they ‘It is them.’
(L&H)
(L&H)
() ’u-waː zi, ’n ’a !imi-ye zi. take- want- ‘Take it, I don’t want it.’
(Z)
Köhler shows a contrast between third person singular dzi and third person plural le. ()
dzí à ǁèŋ. it :- tooth ‘It is his tooth.’
() lè áŋgè ǁèŋ. they :- tooth: ‘They are their teeth.’
... Demonstrative pronouns ǁX’egwi had proximate, distal, and remote demonstratives, the first two of which could be suffixed to the third person pronouns for emphasis (see Table .). They could be pluralized by -ŋ. While Lanham and Hallowes (: ) claim ʔela, ʔena, ʔeta are all synonymous, Ziervogel and Westphal record a difference of meaning. Both Ziervogel and Westphal say ʔe-la, and ʔe-na are alternatives for ‘this’, with the first the more common. Both agree that ʔe-ta contrasts as ‘that’. The ʔe- stem of the 3
In the example sentences, I will give the selfstanding pronouns full glosses and use symbols 1, 2, etc., for the clitics.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Demonstratives and locative adverbs in ǁX’egwi Gloss
Westphal Köhler
‘this’ tsélá ‘that’
Ziervogel
Lanham and Hallowes
’e la, 'e na
ʔe-na, ʔe-la, ʔe-ta
tsela
ʔétá
éːtè, éːtà ’e ta
‘here’
ɟee̋’e
gé’è
‘there’
tseèyà
‘yonder’
zéːkwésò
‘that yonder’
’e ta ǀe kweso ke’e, ge’e
c’e’e ~ ɟe’e
ze kweso
demonstrative could function as a relative. The -na suffix was also found in other words with a meaning ‘this’: ‘today’ ’nǁ’umu naː (Z); ŋ. ǁumuna (W) (< ŋ + ǁumi ‘sun’ + na ); ‘this year’ ŋǃxé’ũnaa (W) (ˀŋ + ǃxe’ũ ‘year’ + naa ). Demonstratives could be suffixed to the third person pronominal/demonstrative form dzi-, e.g. há tsélá (K). () n-’énà yí-à yí gè ǁàlàŋ. - - shoulder.blade ‘These are our shoulder blades.’
(K)
() ŋ́ ǀnâː éːtà há ǁèŋ. saw : house ‘I saw that house.’
(K)
The third person pronoun bases, alone or in combination with demonstratives, precede most nouns in a sentence and seem to function to some extent as definite articles. ki ya ha ’e ta ha kwi () ha kxelo :-blanket :-person ‘the blanket of that person’
(Z)
... Interrogative pronouns Westphal gives the following interrogative pronouns: too-na ‘who’ (), too-gũ ‘who’ () (Lanham and Hallowes to-wa ~ twa-ŋ), ‘what’ gwana (Ziervogel ku na). In addition, Lanham and Hallowes give tee ‘where’, thiĩ ‘what’ (Ziervogel tiŋ), kentwa ‘when’, and ntu ‘how’.
.. Verb ǁX’egwi verbs can be divided into four classes, according to the behavior of associated tense–aspect markers (TAM):
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ǁX ’ () () () ()
the copula, existential verbs, stative-adjectival verbs, and active verbs.
The copula is -ya-, which fuses with the pronominal proclitics. In negative and past tense sentences, it is replaced by the negative and past tense markers. () iŋʔe iɲaʃaa. I --handsome ‘I am handsome.’
(L&H)
() ha-ʔa-ŋʔe. --I ‘It is not me.’
(L&H)
() ha-ge-ʔa-kwii. ---person ‘He was not a person.’
(L&H)
Existential verbs are verbs of being and having and are characterized by not being able to take the tense suffixes -e and -a. () ŋ-kl’e ŋ-ʔa. :-people -be-.present ‘The people are present.’
(L&H)
() u-gi chaŋ. -have milk ‘You () have milk.’
(L&H)
()
ha ge ki ǁ’ele’e. have child ‘He had a child.’
(Z)
Adjectival verbs take the -a aspect suffix in present tense. Active verbs take the full range of tense–aspect markers. Although Lanham and Hallowes (: ) say that the verbal suffixes have a preference for certain verbs and are otherwise meaningless, Ziervogel, Westphal, and Köhler all agree that they mark tense. ǁX’egwi had three tenses: present, past, and future. Present tense was marked by the suffix -e, past by the suffix -a, and future by the particle za preceding the verb. Verbs appear to have fallen into three classes, based on the way in which the TAM suffixes were chosen and attached to the stem. In class , the TAM suffixes fused with the stem, in the second class the future was based on the present stem, and in the third on the past stem. The examples in Table . are taken from Köhler’s notes.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Active verb tense forms in ǁX’egwi Verb
Present
Past
Future
Infinitive
zà ǁ’è zà sé
sè̱:zĩ ̏
zà ǃhàlà
ǃhàlàzĩ ̀
zà ǀxèyà
ǀx”èzĩ ̀
Class : irregular, stem fuses with suffix ‘go’ ‘come’
ǁ’à sà
ǁ’éyá se̍ya̍
Class : future built on present stem ‘plow’
ǃhàlàyè
ǃhàlâː
Class : future built on perfect ‘speak’
ǀxè
ǀxèyâː
Other TAM such as ge (past continuous) and so (progressive) precede the verb.
.. Adverbials Only a small number of adverbials have been recorded for ǁX’egwi. The most notable feature was a prefix ŋ, which was ubiquitous with temporal adverbs: n-dlobe ‘tomorrow’, ŋ-!kxaa ‘yesterday’, ŋ-ǁumǀiso ‘afternoon’ (perhaps based on ǁumi ‘sun’), and ŋ-ʔuye ‘now’.
. S
.................................................................................................................................. The basic word order was S–V–O in both statements and questions. Nominal modifying elements—adjectives, demonstratives, numbers, and other quantifiers—followed the noun, usually linked to it by a relative pronoun. Adverbial elements were usually placed at the end of the sentence. The verb, depending on its subclass, was followed by a tense–aspect suffix and might be preceded by one or more particles. The negative particle directly preceded the verb. In ditransitive sentences, the basic order was S–V–IO–DO. () in-sa-ae ʔunẽẽ. -give- food ‘I give you food.’ ()
in-sa-ae gi-ǁŋa. -give- -thing ‘I give you a thing.’
(L&H)
(L&H)
Verbs of wanting, knowing, etc., took a complement formed with the nominalizing suffix -ziŋ. This is the only complement structure recorded in the data. Examples () and () illustrate a transformation from simple VP to verb with complement. The present tense
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ǁX ’
suffix of tà’àne ‘travel’ is replaced by the infinitive suffix and the new clause then embedded in a sentence with ‘want’. () ń tà’àne ǁná àyè. travel- you ‘I travel with you.’ ()
ń ǃimíyà tá’ázì ǁná áːyè. want- travel- you ‘I wanted to travel with you.’
(L&H)
(L&H)
Aside from ǁŋa- ‘with/and’, which can be used to join nouns, all other conjunctions recorded for ǁX’egwi are borrowed from Bantu.
R Bleek, D. F. (). Comparative Vocabularies of Bushman Languages. Cambridge: School of African Life and Languages at the University of Cape Town. Güldemann, T. (). ‘Tuu as a language family’, in T. Güldemann (ed.), Studies in Tuu (Southern Khoisan). Leipzig: Institut für Afrikanistik, Universität Leipzig, –. Köhler, O. (n.d.). Unpublished notes housed in the Oswin-Köhler-Archiv, Institut für Afrikanistik, Goethe-Universität Frankfurt am Main. Lanham, L. W., and Hallowes, D. P. (). ‘An outline of the structure of Eastern Bushman’, African Studies : –. Potgieter, E. F. (). The Disappearing Bushmen of Lake Chrissie: A preliminary survey. Pretoria: J. L. Van Schaik. Traill, A. (). ‘Phonetic diversity in the Khoisan languages’, in J. W. Snyman (ed.), Bushman and Hottentot Linguistic Studies . Pretoria: University of South Africa, –. Vossen, R. (). Die Khoe-Sprachen: Ein Beitrag zur Erforschung der Sprachgeschichte Afrikas. Cologne: Rüdiger Köppe. Westphal, E. O. J. (). ‘The click languages of southern and eastern Africa’, in T. A. Sebeok (ed.), Current Trends in Linguistics. Vol. : Linguistics in Sub-Saharan Africa. The Hague: Mouton, –. Westphal, E. O. J. (n.d.). ǁx’’e (Bushmen of Amsterdam, district Ermelo). (BC (C), E O J Westphal Papers). Unpublished notes housed at the Rare Documents and Manuscripts Department, University of Cape Town. pp. Note: Xegwi (in pencil). Ziervogel, D. (). ‘Notes on the language of the Eastern Transvaal Bushmen’, in E. F. Potgieter (ed.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. S is spoken by approximately , people in central Tanzania. Published research on the Sandawe language includes a grammar by Dempwolff (). More recent works have focused on phonetics and phonology (Elderkin ; Wright et al. ; Eaton ; Hunziker et al. ) or grammar and discourse (Elderkin , ; Eaton a, b; Steeman ). A dictionary has also been produced (Ehret and Ehret ).
. P
..................................................................................................................................
.. Vowels Sandawe has fifteen vowel phonemes (Table .).
Table . Sandawe vowel phonemes Front unrounded Central unrounded Back rounded High
i, iː, ĩː
u, uː, ũː
Mid
e, eː, ẽː
o, oː, õː
Low
a, aː, ãː
Oral vowels have a phonemic length contrast, whereas nasal vowels do not. Low-toned /i/ and /u/ are realized as voiceless vowels in word-final position, except where they follow consonant clusters. Sequences of adjacent vowels are uncommon and only involve oral vowels.
.. Consonants Sandawe has twenty-nine egressive (Table .) and fifteen ingressive (click) (Table .) consonant phonemes.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Egressive consonant phonemes in Sandawe
Plosive
Bilabial Labiodental
Alveolar (apical)
p, pʰ, b
t, tʰ, d
Alveolar lateral Post(laminal) alveolar
tɬ, dɮ ts’
Nasal
m
Labial- Glottal velar ʔ
k, kʰ, g
Affricate Ejective
Velar
tʃ, tʃʰ, dʒ
tɬ’
k’
ɬ
x
n ɾ
Tap Fricative
f
s
Approximant
l
h
j
w
Table . Ingressive consonant phonemes in Sandawe Dental Post-alveolar (apical) Lateral (laminal) Voiceless unaspirated
ǀ
ǃ
ǁ
Voiceless aspirated
ǀʰ
ǃʰ
ǁʰ
Voiced
ᶢǀ
ᶢǃ
ᶢǁ
Voiced nasalized
ⁿǀ
ⁿǃ
ⁿǁ
(Post-)glottalized
ǀ’
ǃ’
ǁ’
Syllables are open, with the exception of those ending in a glottal stop and those created by vowel elision or through borrowing words with prenasalized plosives or nasal plus plosive sequences. Clicks occur both word-initially and word-medially. Voiced clicks are rare.
.. Tone Sandawe has two underlying tone levels, high (ˊ) and low (ˋ). A surface mid tone is analyzed as a downstepped high. Falling tones (ˆ, on short and long vowels) and rising tones (ˇ, on long vowels only) are analyzed as sequences of level tones. Rightward tone spread occurs in non-word-final morae (Hunziker et al. : –).
. M
..................................................................................................................................
.. Noun ... Inflection Most Sandawe nouns are not inflected for gender, but some nouns referring to people contain person–gender–number (PGN) suffixes (Elderkin : ), e.g. máxà-éː ‘man’ and dìʔsěː-sú ‘old woman’. These suffixes also have other functions in the language (Table .).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . PGN suffixes in Sandawe Low-toned
High-toned
1
-sì ̥
-sé
2
-pò
-pó
3.
-Ø / -è / -ù / -mù̥
-é / -éː
3.
-sù̥
1
-sú / -éːsú -sũː́
2
-sũ̀ː / -sà -sĩː̀
-sĩː́
3.
-sò
-só
3.
-ʔ`̥wà
-ʔwáː / -ʔ`̥wáː
Nouns are often unmarked for number, but plural marking by means of the suffix -xéː together with the specificity suffix -`̴ː is possible, e.g. bô-xéː-`̴ː ‘the (specific) words’. When the specificity suffix is attached to a noun without this plural suffix, it must be followed by a low-toned PGN suffix, e.g. ǀʰιˇ ːà-`̴ː-sù̥ ‘the (specific, female) dik-dik’.
... Derivation Nouns can be derived from verbs and adjectives by the suffix -ʔ`õː, e.g. ǁáːsì ̥-ʔ`õː ‘flight’ (from ǁáːsì ̥ ‘flee’), méː-ʔ`õː ‘largeness, size’ (from méː ‘big’). The suffix -óː has the same function, but is restricted to certain verbs. It replaces the final vowel in the verb to which it attaches and the resulting form has all high tones, e.g. ǁántóː ‘satisfaction’ (from ǁàntá ‘satisfy’). A third action nominalization suffix is -sà, which must be followed by a postpositional suffix, e.g. ʔítʃʰà-sà-nà ‘for looking for’ (from ʔítʃʰà ‘look for’). Agentive nominalization is by means of the suffix -ì, e.g. fàɾé-ì ‘liar’ (from fàɾé ‘lie’), ǀʰímé-ì ‘singer’ (from ǀʰímé ‘sing’). Nouns with an instrumental meaning can be derived from verbs with the instrumental suffix -ʔĩ̀ː, e.g. kê-ʔĩ̀ː ‘something to climb with’ (from kê ‘climb’).
.. Adjectives and numerals Sandawe has a limited number of underived adjectives, including bútɬ’ì ̥ ‘red’, k’ùnk’ú ‘blunt’, and méː ‘big’. Adjectives can also be derived from verbs by suffixation. The possessive suffix -sí followed by a low-toned PGN suffix is commonly used for this, e.g. gàndà-sí-è ‘thin..’ (from gàndà ‘be thin’). The suffix -`̥tʰéː is a productive means of deriving adjectives and quantifiers from intransitive verbs, e.g. bàʔé-`̥tʰéː ‘big’ (from bàʔé ‘be big’), děː-`̥tʰéː ‘many’ (from děː ‘be many’). The suffix -ʔ`̥tó has the same function, but is limited to a few examples, such as ǀ’ʷěː-ʔ`̥tó ‘narrow’ (from ǀ’ʷěː ‘be narrow’). The suffix -tò productively derives an adjective from a verb containing object marking, such as ⁿǀěː-tò ‘cut (for modifying a singular noun)’ (from ⁿǀěː ‘cut...’). The Sandawe numeral system is quintenary and uses the word dáːndà ‘at the side’ for compounds, e.g. kʷàʔḁ̀ná dáːndà ts’éxè ‘six’ (lit. ‘five at the side one’).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Pronouns There are seven personal pronouns in Sandawe: tʃí ‘I’, hàpú ‘you.’, hèwé ‘he’, hèsú ‘she’, sṹː ‘we’, sĩː́ ‘you.’, and hèsó ‘they’. These pronouns also function as possessive determiners when they immediately precede a noun with a lowered tone pattern, e.g. tʃí mìːndʒò ‘my journey’ (from mιˇ ːndʒó ‘journey’). Pronominal possessive constructions can be formed with the pronominal suffix -ì, e.g. ⁿǀʷǎː-ì ‘elephant’s one’.
Table . Sandawe demonstratives Masculine
Feminine
Plural (animate)
Plural (inanimate)
Proximal
hěːù
hěːsù̥
hěːsò
hěːxʷéː / hěːxʷè
Referential
hèwé
hèsú
hèsó
hèwéxéː / hèwéxè
Distal
hǎːù (hĩ ̌ːgò)
hǎːsù̥ (hĩ ̌ːsù̥)
hǎːsò (hĩ ̌ːsò)
hǎːxʷéː / hǎːxʷè (hĩ ̌ːxʷè)
Sandawe demonstratives can be used as pronouns or as modifiers (Table .). The plural inanimate forms may also be used with animate nouns, if the nouns do not have PGN suffixes. The forms in brackets are less commonly occurring variants associated with the Eastern dialect of the language. In Sandawe, the specificity suffix -`̴ː also functions as a relativizer. In subject relatives, the specificity suffix is preceded by the possessive suffix -sí and a low-toned PGN suffix, e.g. màkǎː ǁ’àntá-sí-kù̥-sí-è-`̴ː (thing be.satisfied----.-) ‘thing which does not satisfy’. In object relatives, the applicative suffix -ts’ê precedes the specificity suffix if there is no object marking in the verb. If the verb contains object marking, the specificity suffix is preceded by either the pronominal suffix -ì (for . objects), e.g. hábúsà-`̴ː tʃí=áː pó-é-ì-`̴ː (condition- I= .-..--) ‘the condition which I gave you’ or the appropriate low-toned PGN suffix (for other objects). Interrogatives include the following: hô ‘who’, hótʃò ‘what’, hótʃò-kìm ̥ éː (what-because) ‘why’, hôsì ̥ ‘why’, há-kù̥ ‘where-at’, há-nà ‘where-to’, háʔ`̥sè ‘when’, and híkí ‘how’.
.. Verb ... Inflection In realis clauses, the verb may be followed by a pronominal clitic (PC) indicating the person, gender, and number of the subject, e.g. ǀí=sì ̥ ‘I come/came’. Some intransitive verbs have suppletive stems which reflect whether the subject is singular or plural: ǀí ‘come..’, ⁿǀàtí ‘come..’; hík’ì ̥ ‘go..’, níʔ ‘go..’. The pluractional suffix -wà attaches to a verb to indicate multiple occurrences of the action of the verb, e.g. jàʔbé-wà ‘work (repeatedly)’. The less common iterative suffix -mé has a similar function, e.g. ǁ’àkí-mé ‘get down (repeatedly)’. Both suffixes can also be indicators of a plural subject in intransitive verbs, depending on the context, and -wà can mark habitual aspect. The object markers in Table . may be suffixed to the verb.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Sandawe object markers Direct
Benefactive
Applicative
1
-sé
-xì ̥-sé
-ts’ê-sé [-ʔtʃʰe], [-ʔse]
-pó
-xì ̥-pó
-ts’ê-pó [-ʔpo]
.
-é, -éː
-kù-è [-kʷe], -kù-éː [-kʷeː]
-ts’ê-è [ts’e], -ts’è-éː [-ts’eː]
.
-ésú, -éːsú
-kù-èsú [-kʷesu], -kù-éːsú [-kʷeːsu]
-ts’ê-ésú [-ts’eːsu]
-sṹː
-xì ̥-sṹː
-ts’ê-sṹː [-ʔtʃʰũː], [-ʔsũː]
-sĩː́
-xì ̥-sĩː́
-ts’ê-sĩː́ [-ʔtʃʰĩː], [-ʔsĩː]
.
-ʔĩː́
-xì ̥-ʔĩː́
-ts’ê-ʔĩː́ [-ts’ĩː]
.
-wáː
-kù-wáː [-kʷaː]
-wà-ts’ê-é [-wats’eː]
A noun phrase which is the object of a transitive verb which does not contain object marking is usually followed by the postposition ts’ì ̥ when the object is specific, for example ts’â-`̴ː=ts’ì ̥=sà ⁿǃʷáné (water-=at=... ask.for) ‘she was asking for the water’. Some verbs show an alternation between the vowels /e/ and /a/ in their stem and objectmarked forms respectively, e.g. sʷéː ‘peel’, sʷáː ‘peel...’; ts’éː ‘drink’, ts’áː ‘drink.. .’. A few transitive verbs mark an object by means of suppletive stems, e.g. pěː ‘put.. .’, kǎː ‘put..’; sí-é ‘take-..’, tɬ’ǎː ‘take..’.
... Derivation Table . gives some examples of verbs derived from other verbs by suffixation.
Table . Verb to verb derivation in Sandawe
dàɾà ‘wait’ mántʃʰà ‘eat’ k’áⁿǃá ‘be lost’
dàɾà-sí-kù̥ ‘make wait’ mántʃʰà-kù̥ ‘feed’ k’àⁿǃà-sé ‘lose’
/
màɬé ⁿǁókʰò
‘choose’ ‘wash’
màɬé-ts’ì̥ ‘be chosen, choose oneself ’ ⁿǁókʰò-ts’í ‘be washed, wash oneself ’
bìkʰé
‘leave’
bìkʰé-kí
‘leave each other’
/
ǃʰóː kʰǎː
‘fall’ ‘hit’
ǃʰóː-ká kʰǎː-ká
‘fall.. with’ ‘hit with’
sàjò
‘speak’
sàjò-m`̥sé
‘want to speak’
The choice between the different causative suffixes and the different middle/reflexive suffixes is lexically dependent. The suffix combination -sí-kù̥ (-, realized as [-súkù̥]) is the most productive of the causatives. The choice between the middle/reflexive and comitative/instrumental interpretations of these respective suffixes depends on context. Some derived verbs have non-transparent meanings, e.g. měːnà-ts’í ‘be happy’ (from měːnà ‘like’).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The suffix -ts’í productively derives verbs from adjectives and adverbs, e.g. k’ánk’áɾà-ts’í ‘be/become black’ (from k’ánk’áɾà ‘black’) and dʒàkʰá-ts’í ‘be/become an outsider’ (from dʒàkʰá ‘outside’). Verbs can be derived from nouns with the suffix -sì ̥, e.g. tʰũ̌ː-sì ̥ ‘be dark’ (from tʰũ̌ ː ‘darkness’) or the combination of possessive and causative suffixes -sí-kù̥, e.g. máɾóː-sí-kù̥ [máɾóːsúkù̥] ‘make someone clever’ (from máɾóː ‘cleverness’). A few verbs are derived from non-verbal grammatical categories by means of the suffixes - ̴́ːkí, -kí, or -kì ̥, e.g. tʃʰěː-kí ‘be finished’ (from the adjective tʃʰěː ‘absent’).
... Tense, aspect, mood Sandawe distinguishes between realis and irrealis (indicative) and imperative/subjunctive by means of the following morphemes (Table .).
Table . Realis, irrealis, and imperative/subjunctive in Sandawe Realis Realis negative
Irrealis
Irrealis negative
/
sì ̥
-tʃʰì-sé [-tʃʰeː]
-ì-sì ̥ [-si̥]
-ì-sì ̥-ts’é [-si̥ts’e]
èʔ
ì
-tʃʰì-pó [-poː]
-ì-pò [-po]
-ì-pò-ts’é [-pots’e]
kò / `̥kʷáː
. à
-tʃʰì-éː [-tʃʰeː]
-ì-Ø [-i]
-ì-Ø-ts’é [-its’e]
kʷà / xì ̥à
.
sà
ò
è
-tʃʰì-sú [-tʃʰuː] -ì-sù̥ [-su̥] -ì-sù̥-ts’é [-su̥ts’e] xì ̥sà -tʃʰì-sũː́ [-tʃʰũː] -ì-sũ̀ː [-sũː] -ì-sũ̀ː-ts’é [-sũːts’e] òʔ -tʃʰì-sĩː́ [-tʃʰĩː] -ì-sĩː̀ [-sĩː] -ì-sĩ̀ː-ts’é [-sĩːts’e] kʷè
àʔ
-tʃʰì-só [-tʃʰoː]
-ì-sò [-so]
-ì-sò-ts’é [-sots’e]
kʷàʔà
The realis affirmative and imperative/subjunctive morphemes are pronominal clitics which may be attached to non-subject clause constituents, such as objects and adverbs, as well as, or instead of, to the verb. The subject in realis clauses may be attached with the subject focus (SF) clitic áː if it is focused. The distribution of these morphemes in a clause is restricted by the following conditions (see also Elderkin : ): ) A verb without a PC must not precede the first PC or SF marker of a clause. ) A verb with a PC must not be preceded by another PC or a SF marker in the same clause. Focused constituents in realis clauses carry a PC or SF morpheme as appropriate (see Eaton a for further details). The other three morpheme sets shown above only attach to the verb. Affirmative realis clauses have a present or past time reference. The difference in time reference is usually made clear by context, the presence of time adverbs, or aspectual marking. Negative realis verbs have a past time reference. Irrealis verbs commonly refer to a future time, but may instead describe habitual past situations which no longer hold, e.g. tʰímé-ì-sò ‘they will cook’ or ‘they used to cook’. Negative irrealis verbs may have a present time reference. Negative imperatives and subjunctives are formed using the particle měː, e.g. měː=kò bìkʰé-sé (=.. leave-.) ‘don’t leave me!’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Sandawe marks perfective and imperfective aspect by means of verb coordination, e.g. mànà-é=àʔ- ː̴́ ǃ’ǒːkʰà (know-..=..- stop; lit. ‘they knew it and stopped’) ‘they already knew it’ and ǃ’ʷǎː=tà=àʔ něː- ː̴́ pùndùsé (puddle=in=.. stay..- swim; lit. ‘they stay and swim in a puddle’) ‘they are swimming in a puddle’. In addition, in transitive verbs, object marking expresses perfective aspect and the absence of object marking expresses imperfective aspect, e.g. mìndà=sì ̥ ɬǒːme (field=. . cultivate...) ‘I cultivated a field (and finished)’ and mìndà=sì ̥ ɬòmé (field=.. cultivate) ‘I cultivated/was cultivating a field’.
.. Adverbials Adverbs in Sandawe may not be recognized by any formal morphological properties such as suffixes or tonal characteristics, but rather by their syntactic properties. Some adverbs do not also belong to other word classes (ɬáː ‘well’), whereas others are also nouns (pʰútɬ’úmà ‘peacefully, safely; peace, safety’) or verbs (xâ ‘badly; be bad’). Postpositions in Sandawe include the following: ts’ì ̥ ‘at’, tà ‘in’, nà ‘to’, tànà ‘into’, tʃè ‘from’, tàtʃè ‘out of ’, ǁ’à(ʔ) ‘belong’, ʔĩ̀ː ‘with’, xéʔ`̥ ‘like’, and kì̥méː ‘because’, e.g. kʰǒː=tà ‘in the house’.
. S
..................................................................................................................................
.. Constituent order The dominant constituent order in Sandawe is SOV, but preposing and postposing constituents for discourse considerations is possible (Eaton a). The following example of the constituent order OSV results from object topicalization: ()
hèwéxéː ⁿǁǒːkò ⁿǀínì ̥-ì-sò .. children eat.meat-- ‘these (meats), the children would eat’
Temporal adjuncts usually occur sentence-initially, while other adjuncts usually occur either before the object or immediately before or after the verb. The unmarked order of constituents in a NP is: demonstrative–possessive–noun–adverb– adjective–numeral/quantifier, e.g. hěːù hàpú hùmbù-`̴ː (this you cow-) ‘this cow of yours’, sómbá ʔúɾĩː́ méː (fish very big) ‘a very big fish’. The order noun–demonstrative is used when the noun has previously been established in the discourse. In a NP with the structure possessive–noun, the modifier may be a pronoun or another noun. The following head is realized with a lowered tone pattern to indicate the genitive relationship, e.g. ɬǎː ts’ʷàː (goat tail; cf. ts’ʷǎː ‘tail’) ‘goat’s tail’.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Major sentence types ... Complex sentences Independent clauses can be joined by a conjunction which is marked in agreement with the subject of the clause (Table .).
Table . Narrative, repetitive, and subjunctive conjunctions in Sandawe
. .
Narrative
Repetitive
Subjunctive
síː píː kʷiː páː kʷáː sáː póː kóː péː ʔáː
sîxì ̥ pîxì ̥ pâxì ̥ kʷâxì ̥ sâxì ̥ pôxì ̥
ʔéː kóː kʷíː kʷáː páː sáː ʔóː
pêxì ̥ ʔâxì ̥
kʷéː kʷáʔáː
The example below illustrates the narrative conjunction (NC): ()
sáː málák’ʷà=à ǃ’àwé páː ǁ’ ĩ ː̂ -`̴ː=áː tʃʰóù .. slip=. fall .. fire-= go.out ‘then she slipped and fell, and the fire went out’
Repetitive conjunctions (RCs) show that the situation described by the following clause is occurring habitually or for a second time: ()
pâxì̥ ǁ’ĩ ̂ː-`̴ː tʃʰóù . fire- go.out ‘then the fire went out again’
Subjunctive conjunctions (SCs) introduce subjunctive clauses: ()
ǁ’àkí=kò ʔóː níʔǃ get.down=.. . go.. ‘get down and let’s go’
A verb following a NC retains its lexical tone pattern when the clause does not contain any realis pronominal clitics or the SF marker, whereas the tone pattern of a verb following a SC is lowered.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
A subordinate clause is constructed using the conjunction hí ‘when’ marked with a realis PC in agreement with the subject of the clause, and a clause-final ìʔ: ()
hí=sà ʔíxĩː̂ =sà ké’é=ìʔ when=... thus=... hear=. sáː ʔúrì̥=sà kʷátì̥. .. very=... be.afraid ‘when she heard thus, she was very afraid’
A NC can link two clauses in a causal relationship by occurring at the end of the second clause and agreeing with the subject of the first clause: ()
ⁿǃê day
ts’éxè one
kéùtò pig
ǃ’ìné-óː=nà=ò níʔ hunt-=to=.. go..
póː. wàré té-è mìndà=àʔ mántʃʰà-é [friend other-. field]=.. eat-.. . ‘one day we went hunting pigs, because they had eaten another friend’s field’ In a similar way, the SC may occur at the end of a clause, linking it to the previous clause and expressing the meaning ‘otherwise’: ()
ǀ’ìnké kʰéʔé-ì-sò kóː! ʔàʔá dóːlò=kò no slowly=.. chew hear-- . ‘no, chew slowly, otherwise they will hear’
The narrative connective clitic à can conjoin VPs in clauses containing NCs (see example ()). The connective suffix - ː̴́ also conjoins VPs and is not restricted to clauses containing NCs.
... Questions Interrogative pronouns usually occur sentence-initially. Questions which expect the answer ‘yes’ or ‘no’ include the interrogative clitic nè, which is attached before a pronominal clitic or SF marker in realis clauses and to the predicate in a copula or the verb in irrealis clauses, e.g. hík’ì ̥-ì-pò=nè (go..--=) ‘will you go?’.
... Copular constructions There is no overt copula in Sandawe. The predicate in a copular clause is usually suffixed with a PGN suffix, e.g. tʃí ⁿǁǒː-sì ̥ ‘I am a child’. Negative copular clauses include the clitic ts’é and irrealis copular clauses are constructed by suffixing the verbalizer -ts’í to the predicate and treating the resulting form as a regular verb.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Dempwolff, O. (). Die Sandawe: Linguistisches und ethnographisches Material aus DeutschOstafrika. Hamburg: L. Friedrichsen. Eaton, H. (). ‘Sandawe’, Journal of the International Phonetic Association : –. Eaton, H. (a). ‘Information structure marking in Sandawe texts’, in I. Fiedler, and A. Schwarz (eds.), The Expression of Information Structure: A documentation of its diversity across Africa. Amsterdam and Philadelphia, PA: John Benjamins, –. Eaton, H. (b). A Sandawe Grammar. Dallas, TX: SIL International, http://citeseerx.ist.psu.edu/ viewdoc/download?doi=....&rep=rep&type=pdf accessed July , . Ehret, C., and Ehret, P. (eds.) (). A Dictionary of Sandawe: The lexicon and culture of a Khoisan people of Tanzania, compiled from the field collections of Eric ten Raa. Cologne: Rüdiger Köppe. Elderkin, E. D. (). ‘Diachronic inferences from basic sentence and noun structure in Central Khoisan and Sandawe’, Sprache und Geschichte in Afrika (): –. Elderkin, E. D. (). The Significance and Origin of the Use of Pitch in Sandawe. Ph.D. thesis. University of York. Elderkin, E. D. (). ‘Sandawe’, in R. Vossen (ed.), The Khoesan Languages. London and New York, NY: Routledge, –, –, –, –. Hunziker, D. A., Hunziker, E., and Eaton, H. (). A Description of the Phonology of the Sandawe Language. Dallas, TX: SIL International, http://citeseerx.ist.psu.edu/viewdoc/download? doi=....&rep=rep&type=pdf accessed July , . Steeman, S. (). A Grammar of Sandawe: A Khoisan language of Tanzania. Utrecht: LOT Publications. Wright, R., Maddieson, I., Ladefoged, P., and Sands, B. (). ‘A phonetic study of Sandawe clicks’, UCLA Working Papers in Phonetics : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.............................................................................................................
LANGUAGE, COGNITION, AND CULTURE .............................................................................................................
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
-
. I
.................................................................................................................................. T major word classes nouns, verbs, adjectives, adverbs, even interjections or exclamations as well as function words like adpositions and particles are natural subjects of detailed descriptions in grammars. By contrast, words subsumed under the term “ideophone” are widely ignored in most grammars or hard to find as a separate word class. This is irritating because some linguists confirm that ideophones may form a huge class of words with even hundreds of members in many languages, as e.g. Gbeya (Samarin : ), Zulu (Staden : ), Ewe (Jungraithmayr : –), Bahnar (Diffloth : ), Baka (KilianHatz : ), Korean (Martin : ), or Xhosa (Weakley : ). The disproportion with respect to the well-established word classes is due to the fact that the study of ideophones is still young, dating back only to the end of the nineteenth century. The most important pioneers were Diedrich Westermann (, ) and Clement Doke (). Westermann created the term Lautbilder (“picture-words”) for a striking onomatopoetic class of words in west African languages “which describe an object or an event as a whole” (Westermann : ). Doke (: ) reported on a similar class of words in many Bantu languages and called them “ideophones” according to their function as “a vivid representation of an idea in sound. A word, often onomatopoeic, which describes a predicate, qualificative or adverb in respect to manner, colour, sound, smell, action, state or intensity”. Under the influence of Westermann’s and Doke’s descriptions, ideophones were almost exclusively described for African languages until the s. Meanwhile, the use of ideophone-like words is observed in all language families of the world so that their existence seems to be a universal feature of human language. Examples of ideophone-like words are given in () und () from the Caribbean language Macushi and from English, respectively. ()
Macushi taw taw kíyîn kíyîn sîpa sîpa sísiu
‘sound of axe chopping wood’ ‘caterpillar or snake moving’ ‘action of waves of rough water’ ‘lightening’
(Abbott : f.)
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
- English boing, zack, rummms, piff-paff-poff, buuuuum
More or less intuitively these words seem to differ phonetically, morphologically, and/or syntactically somehow from other word classes: they allow sounds or sound combinations in addition to the inventory of phonemes, they seem to be invariable simplicia, and they commonly change their position in an utterance. However, there are no clear formal criteria that define a word class of ideophones cross-linguistically. Therefore, descriptions of ideophones are rather language-specific and there still exists confusion about how to classify ideophone-like words in a given language. Are they subclasses of major word classes, or do they form a separate class? Are they even words in a strict sense? Samarin () claims the need for good linguistic criteria and lists some in a comparative study of ideophones in Bantu languages. Unfortunately, many of the formal criteria are valid for ideophones of the homogeneous group of Bantu but not applicable to other African languages. Moreover, no single formal feature applies to all ideophones so that Childs (, ) prefers a prototype approach to describe African ideophones. A complementary perspective on ideophones is provided by the discourse-pragmatic approach, while Kilian-Hatz () presented the first typological study of ideophones worldwide by offering a cross-linguistic definition which combines universal, formal, and discourse-pragmatic criteria.
. M : -
.................................................................................................................................. Like other lexical word classes, ideophones are primarily classified as a syntactic word class. A leading scholar of this approach is Newman (). In some languages, ideophones seem to occupy a fixed slot in a clause, e.g. that of a noun, a verb, or an adverb. In these cases they are commonly categorized as a subclass of the prevailing major word class with some phonological and morphological anomalies. Thus, they are called, for instance, “ideophonic nouns, verbs, or adverbs”. We find such special nouns in African languages such as Yoruba (Rowlands ) or Somali (Dhoorre and Tosco ), special verbs in the Bantu language Tsonga (Marivate : ), or special adverbs in the Caribbean language Yagua (Payne and Payne : ). But ideophones frequently occupy two or more slots; consequently, they are at the same time subclasses of two or more major word classes. This is the case in most languages, so that ideophones are often extremely mobile, as, for example, in the Ubangian language Gbaya-Bogoto (Peli /) or the Amerindian language Pastaza Quechua (Nuckolls : , –) where they may occupy the position of almost any kind of major word class. The syntactical mobility of ideophones led Kunene () to reject a purely syntactical definition as misleading. Ideophones are independent of the position which they occupy in
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
a clause; they are “syntactically aloof ”, because they are grammatically and pragmatically different from the major word classes; ideophones are generally characterized as “dramaturgic predicates” comparable to vocal gestures. Crucial in Kunene’s study on Southern Sotho (Bantu) ideophones is the shift of participant roles whereby “the speaker becomes the actor of the ideophonic event and the hearer becomes a witness” (Kunene : –). In that he followed Fortune (: ), who characterized ideophones in the Bantu language Shona as part of the event level in being “a vivid re-presentation or re-creation of an event in sound.” Most authors now focus on the special expressive function of ideophones as part of an expressive register. They call them “expressive adverbs”, as Dumestre () did for the Mande language Bambara; “expressive onomatopes”, as Kakumasu () did for the Amerindian Guarani language Urubu-Kaapor; or simply “expressives”, as Diffloth () did for the Mon-Khmer language Semai. According to Kock (: ), “the communicative purpose of ideophones . . . [is] to actualize that which it [sic] describes.” Similarly, Nuckolls (: ) states for ideophones in Pastaza Quachua that “the distinction between a speech event and a narrated event is blurred [because] the speech event becomes the narrated event”. And for Kita () Japanese ideophones (“mimetics”) are part of the affectoimagistic dimension, which is complementary to the analytic dimension containing descriptive information. Thus, ideophones are not just expressive variants of other words but belong to a complementary mode or dimension of speech. They consist of a combination of motivated sounds that imitate an event rather than describe it.
. T
.................................................................................................................................. In combining formal and discourse-pragmatic criteria, one can define an ideophone crosslinguistically as having the following features (cf. Kilian-Hatz ): ideophones are part of a mode of utterance which is quite different from that of interjections and descriptive words (verbs, nouns, adjectives, and adverbs). Unlike descriptive words, which describe an event, the use of ideophones forms a new kind of perception. It allows speaker and hearer to share a direct experience of an extra-linguistic event by imitation of it through the vocal tract. This causes a social proximity between speaker and hearer (cf. Nuckolls : ). This is a main feature of informal speech and explains in turn why ideophones are part of it exclusively. Therefore, narrations and similar kinds of texts where ideophones are found are limited to “phonically and graphically spoken language” in the sense of Koch and Oesterreicher (). From this perspective, the huge number of ideophones found in texts from many African languages may merely be due to their oral tradition. Ideophones are content words whose semantics is learned and identifiable by the speech community. Ideophones provide a direct sensitive perception (audible, visible, tactile, osmic, gustatory, or psychic) of the event. Ideophones in the languages of the world differ typologically according to the distribution of possible sensory experiences: there are languages whose ideophones denote uniquely audible events because they imitate only acoustic sensations. And then there exist languages that allow ideophones to denote audible plus other kinds of sensations, as mentioned above.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
Descriptive words form only one constituent of a more complex utterance. Therefore, they denote only one single aspect of the whole event described by a clause. By contrast, ideophones have a more complex meaning. An ideophone contains information about the actor as well as its performed action or its state. Therefore, ideophones are equivalents to complete, clausal utterances. Their clause-like character is finally reflected by the fact that they commonly form an intonation unit of their own, irrespective of the context in which they may be embedded. The simulation of an event by an ideophone implies that the quoted event really takes place in (the fiction of ) the narration. Therefore, ideophonic utterances are inherently affirmative and declarative in nature. That is why ideophones are limited to incorporation in affirmative declarative sentences. And, in turn, they represent neither instances of negation nor interrogations. Grammatical relations between an ideophonic clause and other constituents of the descriptive context are not expressed morphologically. Therefore, ideophones are syntactically aloof and morphologically invariable. The interaction of ideophones and descriptive context takes place on the narration level, where the descriptive text is interrupted by ideophonic utterances. The descriptive context can be understood as the frame wherein the ideophonic event is realized. The clause-like nature of ideophones allows two ways of integration into the context: (a) Ideophones may replace a whole descriptive clause. They are then completely isolated and form a parataxis to other sentences or clauses, as demonstrated in (3) below, taken from the Ubangian language Baka with a long sequence of ideophones; or they form one clause in a series of clauses, as in the English example “Our eyes met. Zing. Cupid!” (Oswalt : ). (b) Ideophones accompany the whole event or a nuance of it (denoted in a verb, noun, adjective) like a non-verbal gesture, either by anticipating the event, as, for example, in (4a) from the Ubangian language Gbaya and in (5a) from German, or by paraphrasing it in an apposition, as, for example, in (4b) from Gbaya and (5b–c) from German; such ideophones as in (5b) would probably be categorized as a subclass of adverbs in the formal approach. ()
()
Baka Wòàwòàwòàwòà, ‘The hunters are discussing, kung, spear strikes the chimpanzee, pao, the chimpanzee breaks a branch,
(Kilian-Hatz : )
pɔ̀ ɔ̀ , the chimpanzee interrupts eating, wóoò, the chimpanzee falls down, tung. hard falling the chimpanzee arrives on the ground.’
Gbaya a. Rut, a yeé kɔ. ‘Flash, it entered a hole.’ b. A yeé kɔ rut. ‘It entered a hole rut (like a flash).’ or ‘It entered a hole—flash.’
(Noss : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
German a. Peng! Da knallt die Tür. bang there bang.: the door ‘Bang! There the door bangs.’ b. Da knallt – peng – die Tür. there bang.: bang the door ‘There the door bangs: Bang!’ c. Da knallt die Tür: Peng! there bang.: the door bang ‘There the door bangs: bang!’
Whereas descriptive words form single constituents that are combined analytically in a clause to denote an event, an ideophonic clause-like utterance consists of one word only. Consequently, there are no further grammatical relations and morphological marking otherwise known from constituents in a descriptive sentence. Ideophonic one-word utterances are therefore simplicia in which all the information is bundled up synthetically. The nature of ideophones is inherently sound-symbolic. Complex information is encoded in language-specific, sound-symbolic, sub-morphemic units that follow each other iconically. The units may consist of a syllable, a sound, or a tone. Each unit represents a phase of an event, a sequence of units its chronological course (cf. Weakley ). The onomatopoeic nature finally causes their special performative function of verbalized dramatization. Ideophones are often characterized as phonetically aberrant. However, this is a rather peripheral phenomenon cross-linguistically and is also found in interjections. Most ideophones of a language do not make use of sounds or sound combinations that are not part of its phoneme inventory. Herein ideophones behave in a rather regular or “tame” way in the sense of Rhodes (). A merely “wild” (Rhodes ), i.e. extreme phonetic realization of the given sounds from the sound inventory is, however, possible. This happens if two or more aspects of an event are encoded simultaneously in one sub-morphemic unit, which leads to so-called Überblendungen (“phono-semantic overlappings”; cf. Wissemann ), resulting in variation and/or formation of phono-semantic clusters that are otherwise not permitted. It seems worth mentioning at this point that only so-called “tame” ideophones are prone to changing the word class and, thus, may develop into nouns or verbs, or vice versa. Ideophonic utterances share many formal and functional properties with utterances of direct speech. The parallels are even so striking that one has to define ideophones as a complementary subtype of it (cf. Voeltz and Kunene for Bantu languages and more generally Roncador : ). Like direct speech, ideophones may be introduced either obligatorily or optionally to the context by a verbum dicendi, as demonstrated in (a) from Southern Sotho, or a quotation marker/complementizer, as shown in () from Tamil and in (a) from Somali. Ideophones are also commonly introduced by dummy verbs (cf. Childs : –) such as ‘do’, ‘give’, ‘have’, ‘be’, as in () and (b), or by other language-specific morphemes of quotes. Other elements that commonly introduce them are perception verbs, as in () and (). Like embedded clauses of direct speech, the clause-like ideophonic “dialogue” forms its own intonation unit. However, a morphologically marked shift from speech to event level that is
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
usually signaled by changing deictic elements (personal pronouns, demonstratives, tense, or local deictics) is not possible with ideophones. ()
Southern Sotho (Kunene : ) a. bá thóla bá-re tú. or: they kept.quiet they-say/do tu b. bá thóla tú. they kept.quiet tu ‘They kept quiet—(they kept) dead-quiet.’
()
Tamil (Asher : ) Ava kupu kupu-ɳɳu aʑutaa. she - weep::. ‘She wept with a sobbing sound.’
()
Baka (Kilian-Hatz : ) Mópipìi à mεὲ piípiípiípií. bird: do piípiípiípií ‘The mopipi-bird is calling [lit. does] piípiípiípií.’
()
Somali (Dhoorre and Tosco : ) a. Sabuulkii baa dhilig yiri. corn-cob.: dhilig said: ‘The corn-cob broke off.’ b. Gacantiisa ayaan dhilig siiyay. arm:. : dhilig gave: ‘I broke his arm.’
() Baka (Kilian-Hatz : ) Bùku à gɔ, à sià: bukuluuu. ball go see bukuluuu ‘The ball is rolling away, look: bukuluuu.’ ()
Southern Sotho (Kunene : ) Ka-utlwa swahlaswahla. ‘I heard a rustle-rustle.’
As in a live broadcast on the radio, the speaker, by using ideophones or direct speech, raises the illusion that the verbalized event or situation happens simultaneously at the moment of its production/pronunciation. Ideophones and direct speech differ from one another only in so far as direct speech quotes an utterance, whereas ideophones report an extralinguistic event like a sound, a smell, a taste, a visual impression, a movement, or a psychic emotion. The complementary nature of the two types of direct speech is summarized in Table ..
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Verbal direct speech vs. ideophonic direct speech Verbal direct speech Ideophonic direct speech Model of quote
speech act
Sensational experience audible Reference
extralinguistic state or event: acoustic sound, movement, taste, feeling . . . audible, visible, tactile, osmic, gustatory, psychic
+deictic
deictic
Languages differ typologically in (a) how ideophones are embedded in the context: with or without an introducing element, such as parataxis, paraphrase, or both; (b) their language-specific sound symbolism; and (c) the range of sensations they may quote. In order to distinguish ideophones from other word classes and gestures we use the features listed in Table .. Gestures are very similar to ideophones and interjections, but they are non-verbal utterances. Ideophones are iconic, grammatically aloof, highly auto-semantic, and clause-like. They differ from interjections only in that interjections are always appellative. Not so ideophones. In contrast to this, descriptive content words are symbolic, synsemantic, and separate clausal constituents. Finally, function words are highly symbolic and grammatically bound.1
Table . Gestures and morphemes Gestures
Ideophones “wild”
dependent on situation grammatically aloof auto-semantic clause-like icon speech and event level synthetic
1
Interjections “tame”
Descriptive word
Function words
independent of situation grammatically bound synsemantic clausal constituent symbol speech level analytic
In recent years, a good number of new studies on ideophones that could not be considered here for editorial reasons have been published. For examples, see the works of Mark Dingemanse (e.g. ) and John Haiman ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
-
R Abbott, M. (). ‘Macushi’, in D. C. Derbyshire and G. K. Pullum (eds.), Handbook of Amazonian Languages. Vol. . Berlin: Mouton de Gruyter, –. Asher, R. E. (). Tamil. Amsterdam: North-Holland Publishing Company. Childs, G. T. (). A prototype definition for ideophones. Paper presented at the “Eighteenth Annual Conference on African Linguistics”, April –, University of Quebec, Montreal. Childs, G. T. (). ‘African ideophones’, in L. Hinton, J. Nichols, and J.J. Ohala (eds.), –. Dhoorre, C. S., and Tosco, M. (). ‘Somali ideophones’, Journal of African Cultural Studies : –. Diffloth, G. (). ‘Expressives in Semai’, in P. N. Jenner, L. C. Thompson, and S. Starosta (eds.), Austroasiatic Studies , Oceanic Linguistics. Honolulu: University Press of Hawaii, –. Diffloth, G. (). ‘I: big, a: small’, in L. Hinton, J. Nichols, and J. J. Ohala (eds.), –. Dingemanse, M. (). ‘Redrawing the margins of language: lessons from research on ideophones’, Glossa: A Journal of General Linguistics : –. Doke, C. M. (). Bantu Linguistic Terminology. London: Longmans. Dumestre, G. (). ‘Idéophones et adverbes expressifs en bambara’, Afrique et Langage : –. Fortune, G. (). Ideophones in Shona: An inaugural lecture. London: Oxford University Press. Haiman, J. (). Ideophones and the Evolution of Language. Cambridge: Cambridge University Press. Hinton, L., Nichols, J., and Ohala, J. J. (eds.) (). Sound Symbolism. Cambridge: Cambridge University Press. Jungraithmayr, H. (). ‘Moderne Entwicklungstendenzen in afrikanischen Sprachen’, Internationales Afrikaforum : –. Kakumasu, J. (). ‘Urubu-Kaapor’, in D. C. Derbyshire and G. K. Pullum (eds.), Handbook of Amazonian Languages. Vol. . Berlin: Mouton de Gruyter, –. Kilian-Hatz, C. (). ‘Die Baka-Ideophone: Ihre Struktur und Funktion in narrativen Texten’, Afrikanistische Arbeitspapiere : –. Kilian-Hatz, C. (). Ideophone: Eine typologische Untersuchung unter besonderer Berücksichtigung afrikanischer Sprachen. Habilitationsschrift. Cologne: Universität zu Köln. Kita, S. (). ‘Two-dimensional semiotic analysis of Japanese mimetics’, Linguistics : –. Koch, P., and Oesterreicher, W. (). ‘Sprache der Nähe—Sprache der Distanz: Mündlichkeit und Schriftlichkeit im Spannungsfeld von Sprachtheorie und Sprachgeschichte’, Romanistisches Jahrbuch : –. Kock, I. J. (). ‘The speech act theory: a preliminary investigation’, South African Journal of African Languages : –. Kunene, D. P. (). ‘The ideophone in Southern Sotho’, Journal of African Languages : –. Kunene, D. P. (). The Ideophone in Southern Sotho. Berlin: Dietrich Reimer. Marivate, C. T. D. (). ‘The ideophones as a syntactic category in the southern Bantu languages’, Studies in African Linguistics, Supplement : –. Martin, S. E. (). ‘Phonetic symbolism in Korean’, in N. Poppe (ed.), American Studies in Altaic Linguistics. Bloomington, IN: Indiana University Publications, –. Newman, P. (). ‘Ideophones from a syntactic point of view’, Journal of West African Languages : –. Noss, P. A. (). ‘The ideophone in Gbaya syntax’, Current Approaches in African Linguistics : –. Nuckolls, J. B. (). Sounds like Life: Sound-symbolic grammar, performance, and cognition in Pastaza Quechua. New York: Oxford University Press. Oswalt, R. L. (). ‘Inanimate imitations in English’, in L. Hinton, J. Nichols, and J. J. Ohala (eds.), –. Payne, D. L., and Payne, T. E. (). ‘Yagua’, in D. C. Derbyshire and G. K. Pullum (eds.), Handbook of Amazonian Languages. Vol. . Berlin: Mouton de Gruyter, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Peli, G. (/). Les idéophones en gbaya-bogoto: mémoire de maîtrise. Université de Bangui, Centrafrique. Rhodes, R. A. (). ‘Aural images’, in L. Hinton, J. Nichols, and J. J. Ohala (eds.), –. Roncador, M. von. (). Zwischen direkter und indirekter Rede: Nichtwörtliche direkte Rede, logophorische Konstruktionen und Verwandtes. Tübingen: Max Niemeyer. Rowlands, E. C. (). ‘Ideophones in Yoruba’, African Language Studies : –. Samarin, W. J. (). ‘Measuring variation in the use of Gbeya ideophones’, in Actes du Huitième Congrès International de Linguistique Africaine, Abidjan, – mars . Abidjan: Université d’Abidjan, –. Samarin, W. J. (). ‘Survey of Bantu ideophones’, African Language Studies : –. Staden, P. M. S. von. (). ‘Some remarks on ideophones in Zulu’, African Studies : –. Voeltz, E. F. K. (). ‘Towards the syntax of the ideophone in Zulu’, in Chin-Wu Kim (ed.), Papers in African Linguistics. Edmonton, AB: Linguistic Research, –. Weakley, A. J. (). An Introduction to Xhosa Ideophone Derivation and Syntax. Grahamstown: Department of African Languages, Rhodes University. Westermann, D. (). ‘Laut, Ton und Sinn in westafrikanischen Sudansprachen’, in F. Boas (ed.), Festschrift Meinhof. Glückstadt and Hamburg: J. J. Augustin, –. Westermann, D. (). ‘Laut und Sinn in einigen westafrikanischen Sprachen’, Archiv für die gesamte Phonetik : –, –. Wissemann, H. (). Untersuchungen zur Onomatopoiie. Teil I: Die sprachpsychologischen Versuche. Heidelberg: Carl Winter.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
Genetic vs. areal distribution in sub-Saharan Africa ......................................................................................................................
.
. I
.................................................................................................................................. D a long tradition of work on color systems worldwide, there remain remarkably few detailed studies of color terms and color term systems for individual African languages. This chapter reports results of a survey of color terms in seventy sub-Saharan African languages. The survey includes languages from Niger-Congo, Nilo-Saharan, Afro-Asiatic, and Khoisan phyla or groupings, taken from West, Central, East and southern Africa. In each phylum, family, and geographic region there is variation in the complexity of “basic color term” (BCT) systems, but the data overall suggest that Africa is marked by an areal distribution of different BCT systems (Berlin and Kay ). There is, however, the strong imprint of a probable early Niger-Congo three-term system on the areal distribution, as well as an obvious impact of salient cultural activities, and certainly urbanization, in the evolution of color term systems. Color-naming systems containing three basic terms are strongly attested in languages of western Africa, and appear to extend across what has been called the macro-Sudan belt (Güldemann ), into the Narrow Bantu languages of East Africa. Most languages of this broad region fall into Greenberg’s () “Niger-Kordofanian” (Niger-Congo) phylum, but the languages within this Niger-Congo construct are still highly diverse, and there is debate about whether all can be shown to hold together as a linguistic-genetic grouping (cf. Good, chapter of this volume). The status of Ubangian languages, Mande, and some of the proposed Atlantic group as part of Niger-Congo has been particularly questioned (Dimmendaal ). Southern African languages, both those pertaining to southern Narrow Bantu and those of the Khoisan grouping, appear to have moderately elaborate color term systems with around five basic terms. “Khoisan” is used in this chapter as a cover term for what is now considered an areal and typologically based collection of at least three language families plus isolates (Dimmendaal ; cf. also Honken, chapter of this volume). The greatest elaboration—and actually variation from simple to complex color term systems—is found in Afro-Asiatic and Nilo-Saharan languages of northeastern and eastern Africa. These may have between (rarely) three or four to seven or eight basic color terms plus salient “color+design” vocabulary. The status of Afro-Asiatic as a genetic unit is
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
not questioned (cf. Porkhomovsky, chapter of this volume). Nilo-Saharan is hugely diverse typologically and much remains to be done in understanding its internal genetic relationships and contact influences (cf. Dimmendaal, chapter of this volume). This chapter first presents a basic introduction to color term concepts relevant for understanding the sample and methodologies of the current study. It then presents the findings of the study and hypotheses for further work.
. C
.................................................................................................................................. Initiated by the ground-breaking work of Berlin and Kay (), over the past half century an extensive body of psycholinguistic and ethnolinguistic literature has developed on color terminology and systems around the world (see, e.g., Kay , Kay and McDaniel , Kay et al. , , MacLaury , Hardin and Maffi , Corbett and Davies , Kay and Maffi , , for some of the main threads). This short section cannot do justice to this literature and particularly the controversies it raises, but presents some basic issues as background for the current study. Berlin and Kay () proposed several criteria for evaluating whether a word in any given language should be considered a BCT. Building on this, Davies and Corbett () and Corbett and Davies () suggest two major criteria types and methodologies for evaluating whether a given term in some language should be considered a BCT. First, a color term is basic if it is psychologically salient within a culture or language. A term is deemed salient if it is one of the first terms that comes to speakers’ minds when they are asked for color words, if most speakers of the language have the term in their idiolect (frequency), if there is general agreement about what the term refers to (consistency), and if reaction times to the term are faster than for other terms. They also suggest one can usefully consider frequency of terms in text corpora, word length, and number of derived forms in which a given term appears. Secondly, a basic color term must be general, meaning that it does not name a subtype of any other color term and it is not restricted in its application. For example, in Maa (Nilo-Saharan, Nilotic) the color adjective ómo ‘light brown, tan’ can only be used for sheep, and hence is not general. On the basis of a sample of about twenty languages, Berlin and Kay () hypothesized an evolutionary scheme for how BCT systems would develop in languages. They first proposed the possibility of eleven BCTs, claimed to be added to systems from left to right roughly as follows: & > > / > > > , , , . The proposal spurred a tremendous amount of subsequent research, including the World Color Survey (WCS), carried out in the s on unwritten languages. The WCS employed a standardized set of color chips and specific experimental protocols (see Kay et al. , ; Hardin and Maffi ; Kay ). A major conclusion from the WCS is that speakers of most languages tend to agree about where focal color points are in standardized color chip arrays. This is true even if languages and their speakers have significantly different inventories of color terms, disagree over the boundaries of categories named by given terms, or have different associations as to the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
most characteristic item named by a term. For example, Murle (Nilo-Saharan, Surmic) was one of the languages included in the World Color Survey. Arensen (: ) describes the WCS Murle research, and reports that Murle colai is clearly a “grue” term, covering all shades of both and . However, its focal point is in bright and the term is saliently associated with the “fresh new leaves that come out on the trees after the rains”; significantly, an additional sense of colai is ‘new’. The WCS and many additional studies, including some from Africa, have chipped away at the original Berlin and Kay () hypotheses. For example, it has been argued that some languages have two BCTs for a given color hue. This is probably the case for Maa (NiloSaharan, Nilotic), where two out of five stative-verb color roots correspond to (Payne ): dɔ ‘be (dark) red’ and nyokie ‘be red’. Both terms are highly salient, and for some speakers both can name more or less the same range of items and hues. However, dɔ usually has an association with deep dark red and is saliently associated with the color of blood (though dɔ is general, and is completely distinct from osárgé ‘blood’). Nyokie can name cattle pelts, certain human skin colors, and possibly any hues; for some speakers, it may be a hypernym for dɔ. Other studies have observed the psychological salience in some cultures of texture for choice of color terms, the salience of particular objects associated with colors, and degree of (de)saturation or fluorescence (e.g. Arensen : – for Murle). African languages have also contributed to the conclusion that BCTs for and can apparently be added at any stage in the elaboration of color-naming systems (Kay et al. ; Davies and Corbett ). A particular challenge in the domain of / terms is posed by strongly pastoralist cultures of East Africa. One of the implications of Berlin and Kay’s original proposal was that if a language had as a basic term, then it should also have separate basic terms for all of , , , , , and . However, if much of society, interpersonal relationships, wealth, and survival are built around livestock, there is high value for being able to distinguish one’s own cattle, goats, and sheep from others’. Livestock pelt colors and patterns are, of course, a prime means for doing so, and morphologically simple terms that name color+design combinations are highly salient and widely shared (e.g. ‘white with large brown splotches’, ‘black+white finely dispersed’, ‘mostly white but with red head’). In fact, Turton () went so far as to claim that Mursi (Nilo-Saharan, Surmic) had no color terms aside from those which name animal pelts. A major question for the Berlin and Kay type of investigation, of course, is whether any of the multiple brown-related and color+design terms would count as BCTs using the WCS methodologies. Where there is high cultural reason to have many terms in the tan-brownwhite-black-reddish domain and also to have color+design combinations, these may have the same or higher psychological salience than (elaboration of ) , , and terms. But the very nature of the WCS stimuli (not including chips varying in texture and combining color+design) a priori precludes exploration of some of these issues. Findings from the WCS plus other studies eventually led to much refined and more elaborated (statistical) proposals about the nature and development of color term systems. According to Kay and Maffi (, ), the current understanding of the evolution of BCT systems may be summarized as follows. Firstly, three types of terms primarily designate colors (i.e. such terms do not primarily denote an object associated with some color). Primary color terms name (Bk),
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
I 2 BCTs
W⋃R Bk
⋃
⋃
G
Y
⋃
Bu
II 3 BCTs W R Y Bk
W R⋃Y Bk
⋃
G
III 4 BCTs
⋃
⋃
G
⋃
Bu
IV 5 BCTs W R Y G Bk
⋃
Bu
or
Bu W R⋃Y
G ⋃ Bu Bk
W R Y G ⋃ Bu Bk
V 6 BCTs
W R Y G Bu Bk
VI >6 BCTs
W R Y G Bu Bk plus derived terms
. Hypothesized basic color term systems (based on Kay and Maffi , )
(W), (Y), (R), (G), and (Bu). These categories are in part physiologically grounded, based on the human inability to sense a combination of and , or of and . The achromatic contrast between and is also considered a primary opposition. Composite color terms name fuzzy unions of the primaries (Kay and McDaniel ). For example, the following Abidji (Niger-Congo, Kwa; Kay ) terms which name the following unions are all composite terms: lóbù ‘black/ green/blue’, lébè ‘red/yellow’, gàlè ‘green/blue’ (also known as ‘grue’). In Abidji, only lófù ‘white’ appears to be a primary term. Derived color terms name fuzzy intersections of primaries. For example, English pink names the intersection of and , and gray names the intersection of and . Secondly, Kay and Maffi (, ) interpret data from out of the WCS languages as supporting an evolutionary sequence roughly as presented in Figure .. The simplest color term system would consist of two composite BCT terms as in column I. As a system increases in complexity, composites are gradually broken down into primaries, as in columns II through V. In later stages, derived terms may be added (e.g. column VI). ‘Gray’ and ‘brown’ terms are “wild”, meaning they may be added at any stage.
. D A
.................................................................................................................................. In the World Color Survey, only sixteen (.%) of the languages were from Africa, even though African languages constitute approximately % of the world’s languages (Lewis ). Further, the maps in Kay and Maffi (), reflecting the WCS sample plus ten additional languages, show that the African WCS subsample distributes primarily across what Güldemann () has called the “Macro-Sudan” belt. In terms of language families, the African WCS subsample has the following distribution:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
• thirteen putative Niger-Congo ( Mande, Kwa, Kru, Gur, Bantoid from Nigeria/ Cameroon, Adamawa-Ubangian) • two Nilo-Saharan (both Surmic) • one Afro-Asiatic (Chadic) • no putative Khoisan Issues of linguistic-genetic diversity and possible contact, which may be reflected in geographic distribution, clearly need to be better addressed. The current study thus examines a broader sample of African languages from a wider geographic sub-Saharan distribution. The composition of the sample is based partly on convenience, but has also been chosen with an eye to internal complexity of proposed phyla. Thus, certain (putative) families are more strongly represented in the sample, compared to the actual percentage of languages from that group. In particular, the sample is biased against Greenberg’s NigerCongo and Narrow Bantu, and increases relative representation of Nilo-Saharan, AfroAsiatic, and “Khoisan”, as shown in Table .. The Afro-Asiatic sampling does not include Berber or Mediterranean-region Semitic. Nilo-Saharan subfamilies are not evenly represented and focus on eastern Africa. The “total attestation” figures in Table . are based on Lewis (). All sixteen WCS African languages are included in the current sample. Table . presents greater detail on the family relationships of languages used in the current study and sources of data. The source “WCS” indicates the World Color Survey (Kay ). Roman numerals in Table . indicate my evaluation of the likely degree of complexity of each language’s BCT system relative to the categories in Figure .. Any language which has some derived terms in its inventory prior to stage V (i.e. prior to the resolution of all composite terms) is classified according to the number and type of composite terms. For example, Arensen () suggests that in Murle the derived color terms rεgε ‘pink’ and gidaŋ ‘gray/brown’ qualify as basic, since they show up frequently in the WCS experiments; but in this study I classify Murle (Nilo-Saharan, Surmic) as stage IV, since it contains a composite ‘blue/green’ basic term. Maa (Nilo-Saharan, Nilotic) has derived terms and is classified as stage VI, since there are distinct, salient, and general terms for all the primary colors.
Table . Distribution of total African languages compared to study sample Phylum or grouping Niger-Congo (excluding Narrow Bantu and Adamawa Ubangi)
Total attestation (Lewis ) = % (, total Niger-Congo)
Current sample = %
Narrow Bantu
= %
= %
Adamawa Ubangi
=
=
%
%
Nilo-Saharan
= %
= %
Afro-Asiatic
= %
= %
Khoisan grouping Total
=
%
, = %
=
%
= %
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . The seventy-language sample Phylum or grouping
Family/region (no. of lgs.)
Languages BCT system (sources)
NigerCongo
Atlantic ()
Kisi II (G. Tucker Childs, p.c.), Wolof II-III (Munro and Gaye )
Mande ()
Jɔ II (Carlson ), Yacouba II (WCS)
Kwa ()
Abidji II-III, Chumburung IV-V (both WCS)
Kru ()
Wobé II, Bété II (both WCS)
Gur ()
Konkomba IV, Mampruli IV, Nafaanra II, Vagla II, Dyimini IV-V (all WCS)
Kainji ()
C’Lela III (Doro )
Bantoid ()
Gunu II, Ejagham II (both WCS)
Grassfields () (Cameroon)
Kom III (Jones et al. ), Ngomba II (Satre )
Cameroon ()
Basaa II (Njock ), Badwe’e I (Beavon ), Koonzime II (Beavon and Beavon ), Mpyemo II (Beavon and Beavon ), Njyem II (Beavon )
Narrow Bantu
Adamawa Ubangi
NiloSaharan
East Africa ()
Swahili II (various sources); Hehe II (Marko Mwipopo, p.c.)
Southern () Africa
Chewa IV, Ndebele IV, Xhosa IV, Tswana III-IV (all Davies and Corbett )
Cameroon ()
Mbodomo II (Boyd )
CAR ()
Ngbugu II (Olson and Mbomate )
DRC ()
Mono III (Olson et al. )
Sudan ()
Mundu II (WCS)
Surmic ()
Mursi IV (Siebert and Caudwell ; Turton ), Suri-chai IV (Lucassen and Last ), Murle IV (WCS; Arensen ), Didinga IV (WCS)
Nilotic ()
Acholi IV (Eidson ; Crazzolara ), Anywa III (Reh et al. ), Nandi III-IV (Creider and Creider ), Bari V-VI (Spagnolo ), Toposa V-VI (Schroeder ), Maa VI (Payne )
Other ()
Ik IV-V (Heine ), Gumuz III (Colleen Ahland, p.c.), Komo III (Manuel Otero, p.c.), Manga V (Jarrett ), Katcha III (Stevenson, n.d.), Temein III (Blench, n.d.)
Afro-Asiatic Chadic ()
“Khoisan”
Mbuko IV (Gravina et al. ), Zulgo IV-VI (Haller ), Ngizim V (Schuh ), Hausa V (Newman ), Lele III (WCS)
Omotic ()
Northern Mao III (Michael Ahland, p.c.); Maale (Amha ); Wolaytta IV, Gamo III, Dorze III (all Abebe ), Zergulla IV (Siebert and Hoeft )
Cushitic ()
Bayso II? (Siebert and Hoeft ); Arbore IV, Tsamai IV (both Wedekind )
Ethiopian () Semitic
Argobba IV (Leyew and Siebert )
Unclassified ()
Ongota IV (Wedekind )
()
!Xóõ IV, Ju/’hoan IV, Khoekhoe IV, Sandawe III; Hadza III (all http://www.sscnet.ucla.edu/history/ehret/Khoisanword.pdf accessed October , )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
It may be questioned whether identification of a color term as basic in the sense of Berlin and Kay () or Corbett and Davies () can be adequately determined from lexical lists, dictionaries, and grammar materials. However, from such materials it can first be determined that certain hue categories are at least positively named. If the available lexical materials are extensive, we can sometimes assess salience of a term from its occurrence in compounds, phrases, and as parts of other color expressions. If the materials are very extensive (e.g. over , items) silences in the data may tentatively be taken as indicating either infrequently used or non-existent terms. Further, cross-language investigator bias coupled with the now considerable literature on color term systems may be of assistance if carefully handled. For example, if in the lexical materials for a given language there is a term glossed as English ‘green’, but we do not find a term glossed as English ‘blue’ (clearly a BCT in English), we might hypothesize that the target language does not name a distinct category of ; but it may subsume hues under a composite term that also names or . Other times, there are multiple terms listed as meaning ‘red’, ‘green’, etc. Kay and Maffi () suggest that even if one is not certain about the BCT status of particular terms due to synonymy, we can still say with some confidence that the given color category is distinguished in the system. From such observations we can make hypotheses about what and how many categories are distinctly named in a language. In general the current study has not counted words as possible BCTs if they are obvious borrowings, or if they name objects with characteristic colors. For example, Munro and Gaye’s () Wolof dictionary includes three terms for ‘green’, one of which is a verb form transparently the same as the root for ‘parrot’, and another of which is a borrowing from French. There is another verb bulóo for ‘blue’, likely a French or English borrowing. From the lexical materials one cannot know for sure that, despite their fairly transparent sources, these could not have developed within Wolof to have the salience, frequency, and generality of BCTs (much as orange is now a BCT in English). But it seems clear that the origins of the Wolof system correspond to a stage II (or possibly III if there is a ‘green/grue’ BCT).
. G
.................................................................................................................................. The data of the seventy language sample (cf. Table .) suggest the following hypotheses about the distribution of BCT systems in Africa. The generalizations are tentative, and there are exceptions even in the data consulted. For example, Gur languages used in the WCS and this study, from geographically close Ghana and Côte d’Ivoire, range from stages II to VI or possibly V. The distributions in Table . suggest some areal influences. The stage II BCT system, involving at least two composite terms, extends from West Africa, along the Macro-Sudan belt,1 1
Kay and Maffi () comment that complexity of color term systems in the World Atlas of Language Structures (WALS; Haspelmath et al. ) data seems to vary with distance from the equator. However, the WALS maps suggest an African confounder, as many of the equatorial three-term BCT systems in the WCS sample are in the Macro-Sudan belt.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Generalizations of apparent African BCT systems Most Niger Congo (western Africa, extending across the Macro-Sudan belt
II–III (with some minority extensions to IV–V)
Northwestern (Cameroonian) Narrow Bantu
II
Adamawa-Ubangi (central-western Africa)
II–III
Southeastern Narrow Bantu
IV
Afro-Asiatic (western and eastern Africa)
III–V
Nilo-Saharan (central and eastern Africa)
III–VI, plus salient “color+design” vocabulary
Khoisan (southern and eastern Africa)
III–V
and into East Africa. Most West and Central African languages fall into Greenberg’s putative Niger-Congo phylum. We have noted that there is strong debate about whether these comprise a single genetic group, but the stage II system is found in both families indisputably Niger-Congo (Kwa, Kru, Bantu) and those whose membership has been questioned (Atlantic, Mande, Ubangi). Grassfields and Narrow Bantu languages of both West and East Africa appear to display this system. For example, the modern East African Narrow Bantu languages Hehe and Swahili have monomorphemic terms for ‘black’, ‘white’, and ‘red’. ‘Green’ is a complex expression related to the color of new leaves, and ‘blue’ is a borrowing from English. Complex expressions on the pattern of ‘the color of x’ may be conventionalized across speakers for certain other color concepts in both Swahili and Hehe, but the pattern of data clearly suggests stage II origins of the color term systems of these East African Bantu languages. A stage II system in Proto-Bantu is supported by the “Bantu lexical reconstructions ” (BLR) at the Royal Museum for Central Africa, which builds on the earlier work of Meeussen () and others. The BLR database contains the widespread reconstructed forms *pi ‘black’, *pe/pemb ‘white’, and *kʊ́ndʊ́ ‘red’, but there are no reconstructions for other colors like ‘green, blue’, or ‘yellow’. Though no stage I systems were found in the seventy-language sample, it is noteworthy that Bimson () cannot reconstruct color terms except for ‘black’ and ‘white’ in Mande. For Northern Mande, reconstruction is unsuccessful for essentially any color terms besides a */gb/ element for ‘white’ (p. ), though terms for ‘black’ and ‘white’ are reconstructed for lower-level nodes within Northern Mande. For Proto-Southwestern Mande Bimson (: ) reconstructs *kòlé ‘white’ and *tὲlí ‘black’. For Proto-Northwestern Mande he is unable to reconstruct any color terms (p. ). He is unable to reconstruct any Mande term for ‘red’. Though speculative, this hints at the possibility of the simplest, stage I BCT type of system in at least one early West African family. The Afro-Asiatic languages of Nigeria and Cameroon in the sample are exceptions to the dominant West African BCT stage II (though with some extension to III) pattern, as they evidence from stages III to IV or V. Thus, they are more like their eastern Afro-Asiatic counterparts, which appear to be at stages III and IV. Chadic, Cushitic, and Ethiopian Semitic tend to have one or more ‘brown’ terms; but ‘brown’ is largely absent in the Omotic sources.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
While composite color terms are common across the Macro-Sudan belt, Davies and Corbett () find weak evidence for their existence in the four Southern Bantu languages they examined. Both Khoisan and Southern Narrow Bantu languages appear to have moderately elaborated stage IV systems. Clearly there was contact between Southern Bantu and Khoisan languages, which may have influenced the Bantu color systems. In Southern Bantu languages, terms for ‘brown’ can be basic (i.e. is “wild”), unlike in the East African Bantu languages. Bonny Sands (p.c.) suggests the presence of terms for animal colors in Khoisan languages; this may influence the salience of terms. The greatest elaboration, as well as the greatest variation, in complexity of color term systems occurs in Afro-Asiatic and Nilo-Saharan languages centered in northern and eastern Africa. Some of the Nilo-Saharan languages have up to seven or eight BCTs plus salient color+design vocabulary. One of the more challenging issues in the seventy-language sample concerns the status of derived color terms (e.g. , , , ). For most languages, the available lexical materials do not contain terms for these concepts, though some are attested in at least some Nilo-Saharan Surmic and Eastern Nilotic languages (e.g. Maa bárláí ‘orange, pastel’; Didinga rega ‘pink, purple’ (Kay ); Murle gidaŋ ‘gray/brown’). It is difficult to know whether the silences in the data indicate total absence of such terms; but it is unlikely that derived terms other than possibly or - have status as BCTs. A concluding observation relates to parts of speech of color terms in African languages. Not all the lexical materials consulted give clear or convincing designation of part of speech, but where they do it is noteworthy that BCTs commonly include at least some terms lexicalized as verbs. This is the case in at least Atlantic, Mande, some Grassfields Bantu, Nilotic and other Nilo-Saharan such as Kuliak and Gumuz, and Omotic.
R Abebe, A. (). Ometo Dialect Plot Survey Report, https://www.sil.org/resources/publications/ entry/ accessed July , . Amha, A. (). The Maale Language. Leiden: University of Leiden. Arensen, J. (). Mice Are Men: Language and society among the Murle of Sudan. Dallas, TX: International Museum of Cultures. Beavon, K. (). Badwe’e–French Lexicon, https://www.sil.org/resources/archives/ accessed September , . Beavon, K. (). Njyem–French–English Lexicon, https://www.sil.org/resources/archives/ accessed September , . Beavon, K., and Beavon, M. (). Lexique koonzime–français, https://www.sil.org/resources/ archives/ accessed September , . Beavon, K., and Beavon, M. (). Mpyemo–French Lexicon (provisional), https://www.sil.org/ resources/archives/ accessed September , . Berlin, B., and Kay, P. (). Basic Color Terms: Their universality and evolution. Berkeley, CA: University of California Press. Bimson, K. (). Comparative Reconstruction of Proto-Northern-Western Mande. Ph.D. thesis. Los Angeles, CA: University of California. Blench, R. (n.d.). Comparative Temein Wordlists, http://www.rogerblench.info/Language/NiloSaharan/Eastern%Sudanic/Temein%cluster/Comparative%Temein%wordlist.pdf accessed July , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Boyd, G. (). Petit lexique gbaya mbódɔ̀mɔ̀-français: Dialecte de Gandima Tongo, https://www.sil. org/resources/archives/ accessed September , . Carlson, R. (). ‘A sketch of Jɔ: a Mande language with a feminine pronoun’, Mandenkan : –. Corbett, G. G., and Davies, I. R. L. (). ‘Establishing basic color terms: measures and techniques’, in C. L. Hardin and L. Maffi (eds.), –. Crazzolara, J. P. (). A Study of the Acooli Language—Grammar and vocabulary. London: Oxford University Press. Creider, J. T., and Creider, C. A. (). A Dictionary of the Nandi Language. Cologne: Rüdiger Köppe. Davies, I. R. L., and Corbett, G. G. (). ‘Colour categories in African languages: a test of the Berlin and Kay theory of colour universals’, in R. K. Herbert (ed.), African Linguistics at the Crossroads: Papers from Kwaluseni, Cologne: Rüdiger Köppe, –. Dimmendaal, G. (). ‘Language ecology and linguistic diversity on the African continent’, Language and Linguistics Compass : –. Doro Rikoto, B. (). K’Batksa C’Lela–C’Anasara–C’Gana. (Dictionary C’Lela– English–Hausa.) Nigeria: Lelna Language Development/Translation Association. Eidson, R. (). Color terms in Acholi: derivation, usage and classification. MS. Eugene, OR: University of Oregon. Gravina, R., Nelezek, R., and Tchalalao, R. (). Mbuko–French–English Lexicon, https://www. silcam.org/documents/lexicons/mbuko/index.html accessed July , . Greenberg, J. H. (). The Languages of Africa. Bloomington, IN: Indiana University [International Journal of American Linguistics ., Part II]. Güldemann, T. (). ‘The Macro-Sudan belt: towards identifying a linguistic area in northern subSaharan Africa’, in B. Heine and D. Nurse (eds.), A Linguistic Geography of Africa. Cambridge: Cambridge University Press, –. Haller, B. (). Zulgo Lexicon, https://www.sil.org/system/files/reapdata//// /Zulgolexicon.pdf accessed September , . Hardin, C. L., and Maffi, L. (eds.) (). Color Categories in Thought and Language. Cambridge: Cambridge University Press. Haspelmath, M., Dryer, M. S., Gil, D., and Comrie, B. (eds.). (). The World Atlas of Language Structures. Oxford: Oxford University Press. Heine, B. (). Ik Dictionary. Cologne: Rüdiger Köppe. Jarrett, K. (). A Dictionary of Manga, a Kanuri language of Eastern Niger and NE Nigeria. Edited by Roger Blench, http://www.rogerblench.info/Language/Nilo-Saharan/Saharan/Manga%dictionary%Unicode.pdf accessed July , . Jones, R., with Loh Pius and the Kom Language Development Committee (). Provisional Kom– English Lexicon, https://www.silcam.org/documents/lexicons/kom/index.html accessed July , . Kay, P. (). ‘Synchronic variability and diachronic change in basic color terms’, Language in Society : –. Kay, P. ( []). The World Color Survey. Stanford, CA: Center for the Study of Language and Information. Kay, P., Berlin, B., Maffi, L., and Merrifield, W. (). ‘Color naming across languages’, in C. L. Hardin and L. Maffi (eds.), –. Kay, P., Berlin, B., and Merrifield, W. (). ‘Biocultural implications of systems of color naming’, Journal of Linguistic Anthropology : –. Kay, P., and McDaniel, C. K. (). ‘The linguistic significance of the meanings of basic color terms’, Language : –. Kay, P., and Maffi, L. (). ‘Color appearance and the emergence and evolution of color lexicons’, American Anthropologist : –. Kay, P., and Maffi, L. (). ‘Number of basic colour categories’, in M. Dryer and M. Haspelmath (eds.), The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, chapter , https://wals.info/chapter/ accessed July , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Lewis, P. (ed.) (). Ethnologue: Languages of the World, th edn. Dallas, TX: SIL International, https://www.ethnologue.com accessed July , . Leyew, Z., and Siebert, R. (). Sociolinguistic survey report of the Argobba language of Ethiopia, https://www.sil.org/resources/publications/entry/ accessed July , . Lucassen, D., and Last, M. (). A Chai word list, https://www.sil.org/system/files/reapdata/// //SILESR_.pdf accessed July , [Data collected in ]. MacLaury, R. (). Color and Cognition in Mesoamerica: Constructing categories as vantages. Austin, TX: University of Texas Press. Meeussen, A. E. (). Bantu Lexical Reconstructions. Tervuren: Royal Museum for Central Africa (repr. with index). Munro, P., and Gaye, D. (). Ay Baati Wolof. (A Wolof Dictionary). Los Angeles, CA: University of California. Newman, R. Ma. (). An English–Hausa Dictionary. New Haven, CT: Yale University Press. Njock, P. E. (). Basaa–English–French–German Dictionary, https://www.silcam.org/documents/ lexicons/basaa/index.html accessed July , . Olson, K. S., and Mbomate, J. V. (). ‘Ngbugu digital wordlist: presentation form’, Linguistic Discovery : –. Olson, K. S., Schrag, B. E., Schrag, B., Mbanza, A. G., Tembenekuzu, K., Olson, R. E., Liedtke, M., Kuntz, J., and Huttar, L. (). Mono -item digital wordlist: presentation form. SIL Language and Culture Documentation and Description, . Dallas, TX: SIL International, https://www.sil.org/ resources/publications/entry/ accessed October , . Payne, D. L. (). ‘Maa color terms and their use as human descriptors’, Anthropological Linguistics : –. Reh, M., Akwey, S., and Uriat, C. (). Anywa–English and English–Anywa Dictionary. Cologne: Rüdiger Köppe. Royal Museum for Central Africa, Tervuren. Bantu Lexical Reconstructions , https://www.africamuseum.be/collections/browsecollections/humansciences/blr accessed December , . Satre, S., and Ngomba Language Committee (). Provisional Ngomba–French–English Lexicon, https://www.silcam.org/documents/lexicons/ngomba/index.html accessed July , . Schroeder, M. (). Dictionary Toposa–English, English–Toposa. Nairobi: SIL International. Schuh, R. (). A Dictionary of Ngizim. Berkeley, CA: University of California Press. Siebert, R., and Caudwell, S. (). Sociolinguistic survey report of the Melo (Malo) and Mursi languages of Ethiopia, https://www.sil.org/resources/publications/entry/ accessed July , . Siebert, R., and Hoeft, L. (). Sociolinguistic report of the languages of the Abbaya/Chamo area of Ethiopia, Part I, https://www.sil.org/resources/archives/ accessed July , . Spagnolo, L. M. (). Bari Grammar. Verona: Missioni Africane. Stevenson, R. (n.d.). Dictionary of Katcha, http://www.rogerblench.info/Language/Nilo-Saharan/ Kadu/Katcha%Dictionary.pdf accessed July , . Turton, D. (). ‘There’s no such beast: cattle and colour naming among the Mursi’, Man : –. Wedekind, K. (). Sociolinguistic survey report of the languages of the Gawwada (Dullay), Diraasha (Gidole), Muusiye (Bussa) areas, https://www.sil.org/resources/archives/ accessed September , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. I experiential situations one of the participants is a sentient being who experiences an emotion, a physical sensation, a wish, or a perception. Consider the following examples: ()
Molly loves Jim.
()
Aster azzənə-čč. Aster be.sad:-. ‘Aster is sad.’1
[Amharic] (Amberber : )
()
náa jí tsooɾóo. . feel fear ‘I am afraid.’
[Hausa]
()
tsooɾóo yáa káamaa ní. fear . catch . ‘Fear took hold of me.’
[Hausa]
()
suma xol dafa tàng. my heart : be.hot ‘I am angry.’ (Lit.: ‘My heart is hot.’)
[Wolof ]
()
1
cwíɲé yòm. liver:. soft:. ‘He’s happy.’ (Lit.: ‘His liver is soft.’)
(Becher : ) [Lango] (Noonan : )
Generally, interlinearizations are unchanged from the source. For purposes of homogeneity, however, some abbreviations have been adapted, and in some cases (literal) translations were added.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
I am cold.
()
Mich friert. . freeze ‘I am cold.’
()
Kofí nyá gε. [Ewe] Kofi know Accra ‘Kofi knows Accra.’
[German]
(Ameka : )
Typically, situations belonging to the semantic domains of , , , , and are regarded as experiential. Crosslinguistically, experiencer constructions tend to show certain peculiarities which have been studied quite extensively for South-Asian (“dative subject”; e.g. Verma and Mohanan ; Bhaskararao and Subbarao ) and some Indo-European languages (Icelandic, “quirky subjects”; e.g. Rögnavaldsson ; Sigurðsson ). Languages tend not to have a dedicated syntactic strategy to codify experiential situations. Instead, a number of constructions are employed which seem to be more suited to expressing material processes (as opposed to physical, sensory, or psychological experiences , cf. Halliday & Matthiesen : –). This may be the use of a transitive verb, as in () and (), or of an intransitive verb, as in (). In these sentences the meaning of the verb is congruent with the meaning of the sentence as a whole, i.e. the verb simply labels the experience that the sentences describe. Therefore, such verbs are called “labeling verbs” and the constructions “congruent” (cf. Reh a: ). The verbs in () and (), on the other hand, refer to a material process (‘catch’) and a physical state (‘be hot’), respectively, and not to and , as the examples as a whole do. Such constructions are metaphoric, and they are quite common in some semantic domains of experiencer constructions, e.g. . Sometimes the syntactic behavior of experiencer verbs differs from the standard transitive or intransitive pattern in a particular language, as in (), where it is unusual that the single argument of a verb is marked by the accusative case. () and () illustrate the use of terms for certain body parts to express experiential situations. Another possibility is identificational, locative, or possessive copula constructions, as in (). Simon and Reh (: ) identify three conceptual entities that constitute experiential situations: the experiencer (a sentient being), the experitum (the emotion/sensation/perception being experienced by the experiencer), and the stimulus (the participant that causes the emotion/sensation, etc., in the experiencer). Croft (: –) differentiates experiencer-subject verbs from experiencer-object verbs, i.e. verbs which assign the experiencer to the subject or object position, respectively. Regarding mental state verbs, he observes that “there is cross-linguistic variation in the assignment of experiencer to subject or object status.” When both possibilities are available, “then the subject-experiencer version is interpreted as implying more volition or direction of attention to the stimulus than the object-experiencer version. The reason for this is that the subject is conceptualized as having control, or at least more control, over the state of affairs denoted by the verb.” This observation also applies to other experiential situations. From a semantic point of view experiencers are ambiguous. As summarized in Table ., they share some features with prototypical agents (conscience, sentience) and
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Table . Properties of prototypical agents, prototypical patients, and experiencers (cf. Kießling : ) Prototypical agent
Prototypical patient
Experiencer
Animacy
+
–
+
Volition
+
–
+/–
Initiation
+
–
+/–
Causation
+
–
+/–
Sentience
+
–
+
Control
+
–
+/–
Affectedness
–
+
+
others with prototypical patients (affectedness). With respect to volition, causation, and control, the experiencer may be conceptualized as being involved to different degrees, i.e. sometimes more like an agent and sometimes more like a patient. This can be related to Comrie’s (: ) description of semantic roles as points on a continuum of control. The endpoints of the continuum—i.e. the prototypical agent and the prototypical patient—have been described by characteristic sets of features (e.g. Bugenhagen : ; Dowty : ). Concerning other, intermediate semantic roles, Bugenhagen (: –) states: Given these two clear endpoints along the “continuum of control”, other roles like Force and Instrument are intermediate, instancing only some of the characteristics of the endpoints. Experiencers constitute another such intermediate semantic role.
Due to the ambiguity of the experiencer role—semantically right between a prototypical agent and a patient—many languages make use of different kinds of constructions that encode experiential situations, highlighting or downplaying certain semantic aspects of the experiencer role, and thereby specifying the exertion of volition, initiation, causation, and control on the part of the experiencer, or of the stimulus. Generally, using a prototypical transitive construction with the experiencer encoded as the subject evokes a more agentive interpretation of the experiencer role, whereas an experiencer encoded in the object position conveys a less volitional view on the experiencer. The aim of this chapter is to give an overview of some lexico-grammatical means which African languages use to codify experiential situations, especially with regard to modulation of certain semantic aspects of the experiencer role. The rest of the chapter is structured as follows. The next section briefly summarizes a number of parameters that need to be taken into account when comparing experiencer constructions cross-linguistically from a typological point of view. The main part of the chapter is devoted to experiencer constructions in various African languages of different genetic affiliation (i.e. Wolof, three South Cushitic languages, Amharic, Likpe, Wolaitta, and Hausa), focusing on their outstanding features.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. E , ,
.................................................................................................................................. When describing experiential constructions in a variety of languages, it is useful to take a consistent set of parameters into account. The choice of parameters used in this overview is based on the framework for typological description of experiencer constructions as outlined in Reh (a, b). As mentioned in section . above, a broad distinction can be made between congruent and metaphoric constructions. Whereas the meaning of the predicate in congruent constructions is consistent with the meaning of the sentence as a whole, predicates of metaphoric constructions may belong to different semantic domains, as in () and (). Languages vary in their preferred ways to encode experiencer situations. When assessing the characteristics of the codification of experiencer constructions in a given language, it is important to take the pragmatic markedness into account, especially of metaphoric constructions. For instance, while in English a sentence like ‘my heart is hot’ is pragmatically highly marked, although perhaps comprehensible to a degree, the equivalent in Wolof (cf. () above) is a pragmatically rather neutral way to express anger (and therefore the appropriate translation in English is ‘I am angry’). Examples () and () show other pragmatically rather neutral metaphoric sentences in Lango and Ewe. () cwíɲé cwε̂r. liver:: :drip: ‘He’s dispirited.’ (Lit.: ‘His liver drips.’) ()
ŋɔ́
dzí-m. worm ooze- ‘I was frightened.’
[Lango] (Noonan : ) [Ewe] (Ameka : )
In other languages, metaphoric constructions are pragmatically marked and may indicate a higher degree of intensity and/or underline that the experiencer is not volitionally involved in the situation at all; see () from Amharic. ()
aster-ɨn fɨrhat yaz-at. Aster- fear catch::.-: ‘Aster is overcome by fear.’
[Amharic] (Amberber : )
However, as we will see in section .. on Wolof below, metaphoric constructions can also be used to indicate lower intensity. Furthermore, experiencer constructions may be differentiated according to the element coded as subject, i.e. the situational orientation (cf. Reh a: –) of the construction. Example () illustrates an experiencer-oriented construction. The experiencer is the subject of an intransitive verb which congruently expresses the emotion experienced by the subject:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
()
[kuuŋgáʔ] a [buh-áʔ / gurhaamis-áʔ]. . / be.angry-: / be.jealous-: ‘You () are angry/jealous.’
[Iraqw] (Kießling : )2
Sometimes a body part replaces the experiencer. We could then speak of a body partoriented construction, as in () above in Lango, with the literal translation ‘his liver is soft’, meaning ‘he is happy’. In the following stimulus-oriented construction the experiencer is grammatically encoded as the object, and the verb expresses the experitum: ()
[ʔidarak] i-ni [ʔana] [baʕatis]. words: - : make.tired ‘These words made me tired/bored me.’
[Alagwa] (Kießling : )
Examples () and () show another stimulus-oriented construction with an experiencerobject verb, i.e. a verb which requires the experiencer to be encoded as an object. ()
[daf] [ma] [bett]. : : surprise ‘It surprised me.’
() dəssɨta yɨ-ssəmma-ññ-al. happiness :-feel-:- ‘I feel happy.’ (Lit.: ‘Happiness (it) feels (to) me.’)
[Wolof ] (Becher : ) [Amharic] (Amberber : )
Regarding the construction types used to render experiential situations, we have already mentioned labeling verbs, i.e. underived verbs which simply name the experitum. On the other hand, verbs may be derived, e.g. by a causative morpheme, as seen in () (compare also () without the causative). ()
dafa tàngal suma xol. : be.hot: my heart ‘It makes me angry.’ (Lit.: ‘It heats my heart.’)
[Wolof ] (Becher : )
Anthropomorphic constructions, in which the experitum is somehow personified (cf. Reh a: ), are usually instantiated by a verb denoting a material process. They can have any situational orientation. See a subject-oriented construction in () and an experitum-oriented construction in (). ()
2
mer moo ma jàpp. anger : : seize ‘Anger seized me.’
[Wolof ] (Becher : )
In Kießling () glosses and translations are in German. In this and the following examples in this section, indexed square brackets have been added.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() tsoro ya cika zukata-n iyaye a Yobe. [Hausa] fear : fill hearts-. parents at Yobe ‘The parents in Yobe were frightened.’ (Lit.: ‘Fear filled the hearts of the parents at Yobe.’) (http://hausa.leadership.ng/node/ accessed March , ) For an in-depth description of a wider range of construction types, see Reh (a). In the following sections language-specific characteristics of the codification of experiential situations are briefly summarized.
.. Wolof (Niger-Congo, Atlantic) As we have seen in section . above, metaphoric renderings of experiential situations often indicate higher intensity. In Wolof, however, metaphoric emotional expressions indicate a relatively lower degree of intensity. For a small number of mental state and inner physical state verbs, two different causative suffixes are available to form stimulus-oriented experiential constructions which assign different degrees of volitionality to the causee, i.e. the stimulus. Becher () subdivides experiencer constructions in Wolof into three semantic subgroups that partly coincide with certain lexico-grammatical characteristics: emotion, non-emotional mental experiences, and inner physical experiences. Emotional (, , , , , , etc.) and inner physical experiences ( , /, , etc.) are encoded as mental states by intransitive predicates, the experiencer being the only participant. () dafa rus. : be.ashamed ‘She is ashamed.’ ()
dama liw. : feel.cold ‘I feel cold.’
[Wolof ] (Becher : ) [Wolof ] (Becher : )
The stimulus can be made explicit by way of using a derived causative verb which transfers the stimulus to subject position and the experiencer to object position. () daf ma rus-loo. : : be.ashamed-:: (?) ‘It made me feel ashamed.’
[Wolof ] (Becher : )
It is a characteristic of emotional expressions that each congruent experiencer-oriented rendering corresponds to a metaphoric body part-oriented construction: () a. dafa mer. : be.angry ‘She is angry.’
[Wolof ] (Becher : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
b. suma xol dafa tàng. my heart : be.hot ‘My heart is hot.’
[Wolof ] (Becher : )
While both constructions share roughly the same low degree of pragmatic markedness, the congruent construction (a) is preferred for emotions which are visibly expressed by a bodily reaction, while the metaphoric body part construction,3 as in (b), is used for emotions of probably minor intensity which do not (yet) trigger a physical reaction (cf. Becher : –). Some emotions and inner physical experiences may be rendered by experitum-oriented metaphoric constructions. In these cases the experiencer is perceived as lacking control over the situation and the personified emotion (a deverbally derived noun) is the subject: () mer moo ma jàpp. anger : : seize ‘Anger seized me.’
[Wolof ] (Becher : )
and are expressed by experiencer-object verbs in transitive, stimulusoriented constructions with the experiencer as object, e.g.: () daf ma bett. : : surprise ‘It surprised me.’
[Wolof ] (Becher : )
These verbs have to be derived in order to be used in experiencer-oriented constructions: () dama waar-u. : surprise- ‘I am surprised.’
[Wolof ] (Becher : )
In stimulus-oriented causative constructions, intentionality of the causer may be indicated by the use of different causative morphemes: () a. dafa tiit-al xale bi. [Wolof ] : fear- child ‘She frightened the child.’ (volitionally induced, causer typically a human being) b. daf ma tiit-loo. : : fear-:: ‘It made me feel afraid.’ (non-volitionally induced)
(Becher : )
The choice between the two causative extensions is possible only for a small number of verbs. Generally, the suffix -al is compatible with quality verbs and -loo with stative verbs 3 According to Becher (: ), certain body parts are associated with specific emotions: heart ! , , , ; mind ! , , ; body ! , ; skin ! .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
denoting mental or inner physical states or physical signs of mental states (Becher : –). In other metaphoric constructions the experitum can occur as a possession of the experiencer, as in (). () am naa naqar. have : bitterness ‘I am sad.’ (Lit.: ‘I have bitterness.’)
[Wolof ] (Becher : )
Non-emotional mental experiences, like , , , , and but also the emotions , and , are typically constructed as mental activities in congruent transitive predications with the experiencer as subject and the stimulus as object. () xam [naa] [loolu / kooku] know : that.thing / that.person ‘I know it/him/her.’
[Wolof ] (Becher : )
Metaphoric constructions with non-emotional mental experiences, which often use material verbs as predicates, may serve to either highlight or downplay volition and control on the part of the experiencer. () [mu] jublu démb. : turn yesterday ‘She remembered the past.’ ()
sama xel daj na ci moom. my mind hit : him/her ‘I remembered him/her.’ (Lit.: ‘My mind hit on him/her.’)
[Wolof ] (Becher : ) [Wolof] (Becher : )
Though both constructions are experiencer-oriented, in () the experiencer is understood as being volitionally involved in the act of remembering, while in (), where the experiencer is replaced by a body part, he does not control the act of remembering.
.. South Cushitic: Iraqw, Alagwa, Burunge In the South Cushitic languages Iraqw, Alagwa, and Burunge two derivational morphemes are used to evoke more or less volitional readings of the semantic role of the syntactic subject in experiencer constructions. While the causative introduces an additional argument in subject position—semantically endowed with control over the verbal action—the use of the mediopassive assigns affectedness and lack of control to the syntactic subject (cf. Kießling : –). In pragmatically unmarked congruent expressions with a nonderived predicate, the experiencer may either be encoded as an untypical agent with an experiencer-subject intransitive verb, as in (), or an untypical patient with an experiencer-object transitive verb; see ():
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() kuuŋgáʔ a buh-áʔ / gurhaamisáʔ. : / be_angry-: / be.jealous-: ‘You () are angry/jealous.’ ()
ʔidarak
i-ni ʔana baʕatis. words- - make.tired ‘These words made me tired/bored me.’
[Iraqw] (Kießling : ) [Alagwa] (Kießling : )
Metaphoric constructions with the personified stimulus as subject are pragmatically marked. They are used to express undesirability and lack of control on the part of the experiencer; see (). [Iraqw] () muunúu-ʔéeʔ gurhaamí-r yaariir gw-aa ʔéer. heart-: sadness- big: :.- enter:.: ‘Great sadness has come upon my heart.’ (Kießling : ) In general, the use of the mediopassive correlates with certain semantic traits of verbs. Verbs often used with the mediopassive are characterized by reflexivity (movement verbs), lack of an agent (processes), and lack of control (bodily reactions). With regard to experiential situations, Kießling (: –) concludes that the use of the mediopassive renders the subject position semantically more suitable for an experiencer. ()
ʔana
[Alagwa] o-no digiduut. :.- remember: () ‘I remember you.’ (without conscious effort and control; in German: ‘du fällst mir ein’) (Kießling : )
The causative, on the other hand, when it does not introduce an additional participant, ascribes more control and volitionality to the subject. Compare () and () with the mediopassive and causative suffix, respectively. [Alagwa] () [ʔana] [o-no] [digiduus] :.- remember: () ‘I recall you.’ (with conscious effort and control; in German: ‘ich rufe dich mir in Erinnerung’) (Kießling : )
.. Amharic (Afro-Asiatic, South Semitic) In Amharic, certain construction types coincide with experiential subfields. The group of verbs denoting (uncomfortable) physical states, which introduce the experiencer as the object, is particularly clear-cut and can be regarded as a specialized subclass of transitive verbs. Metaphoric constructions primarily denote high intensity. Emotions like , , and are expressed by congruent experienceroriented constructions with intransitive verbs.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() aster azzənə-čč. Aster be.sad:-: ‘Aster is sad.’
[Amharic] (Amberber : )
and as well as (, , and ) are instantiated by transitive constructions with the experiencer as subject and the stimulus as object. () aster ləmma-n wəddədə-čči-w. Aster Lemma- love:-:-: ‘Aster loved Lemma.’ () sɨlə-ɨhɨt-u assəb-ə about-sister-: think:-: ‘He thought about his sister.’
[Amharic] (Amberber : ) [Amharic] (Amberber : )
Uncomfortable physical states of the body like , , , and , which an experiencer would not bring about himself voluntarily, are lexicalized as experiencerobject verbs. () rabb-ə-ññ. hunger.-.-. ‘I am hungry.’ (Lit.: ‘It hungers me.’)
[Amharic] (Amberber : )
These verbs can be considered as a subclass of transitive verbs. In the literature they are sometimes referred to as “impersonal verbs” because they are used nearly exclusively with the third person masculine singular (‘it’) as subject, denoting the stimulus (cf. Leslau : ). The stimulus, though, may also be expressed by a nominal phrase, as in (). () hɨməm yɨ-ssəmma-ññ-al. pain :-feel-:- ‘I feel pain.’ (Lit.: ‘Pain feels (to) me.’)
[Amharic] (Amberber : )
When the passive/reflexive prefix t(ə)- is used with a transitive experiencer-subject verb, the experiencer is moved to object function and is regarded as less agentive, denoting “thinking which occurs without one’s full control” (Amberber : ). () t-assəb-ə-ññ. /-think:-:-: ‘It occurred to me.’ (Lit.: ‘It was thought to me.’)
[Amharic] (Amberber : )
In comparison, () above shows the same verb without derivate affixes, denoting deliberate thinking. Both the passive/reflexive and the causative derivation may reverse the allocation of semantic roles to subject and object function, respectively, with different effects depending on the type of verb they are combined with. When experiencer-object verbs are constructed with the passive/reflexive derivation, the experiencer is encoded as the subject, as in ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() aster tə-č’ənnək‘ə-čč. Aster /-worry:-: ‘Aster is worried.’ (Probably lit.: ‘A. worried.’)
[Amharic] (Amberber : )
Compare the sentence in () with the underived verb ‘worry’ in a construction with the experiencer as object. () aster-(ɨn) č’ənnək’-at. Aster-() worry:::-: ‘Aster is worried.’ (Lit.: ‘It worried Aster.’)
[Amharic] (Amberber : )
With experiencer-subject verbs the causative derivation as- introduces the stimulus as subject and the experiencer as the object. [Amharic] () ləmma aster-ɨn as-dəssət-at. Lemma Aster- -be.happy:::-: ‘Lemma pleased Aster.’ (Lit.: ‘caused her to be happy/joyful’) (Amberber : ) When used with basic experiencer-object verbs, the causative may be used to indicate more control on the part of the experiencer, which is then encoded as subject. Consider () without the causative. The experiencer is encoded as the object and perceived as exerting a low degree of control over the event of remembering. () mədhanit-u. t-awwəs-at. medicine- /-remember:::-: ‘She remembered the medicine.’ (‘It came to mind.’)
[Amharic] (Amberber : )
In (), on the other hand, Aster (the experiencer and subject of the sentence) remembers the medicine by way of a conscious effort. () aster mədhanit-u-n as-t-awwəs-əčč. Aster medicine-- --remember:-: ‘Aster remembered the medicine.’
[Amharic] (Amberber : )
Metaphoric renderings of experiential situations seem to primarily indicate higher intensity and are probably pragmatically marked. In the following sentence (cf. (), here repeated as ()) is personalized and encoded as subject; the experiencer is the object. () aster-ɨn fɨrhat yaz-at. Aster- fear catch:::-: ‘Aster is overcome by fear.’ (Lit.: ‘Fear caught Aster.’)
[Amharic] (Amberber : )
Certain body parts are associated with emotions, as () illustrates. In these cases, the body part usually replaces the experiencer.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() dəm-u fəlla. blood-: boil:: ‘He is incensed.’ (Lit.: ‘His blood boiled.’)
[Amharic] (Amberber : )
.. Likpe (Niger-Congo, Kwa) In Likpe, subject and object functions are marked by word order. With some verbs denoting situational orientation is underspecified, i.e., without modification they may either be used with the experiencer in subject or object position. Pragmatically neutral metaphoric constructions are used for a great variety of experiential situations. In two specialized metaphoric constructions expressing the experiencer is encoded as an indirect object. and are rendered by transitive, experiencer-oriented constructions with labeling verbs: () u-sió ə́-mə á-nɔ e-tiki ə́-mə. -woman - -hear -shout - ‘The woman heard the shout.’ ()
n-klomá fə. -remember ‘I remember you.’ (i.e. ‘I miss you.’)
[Likpe] (Ameka : ) [Likpe] (Ameka : )
As mentioned at the beginning of this section, some verbs belonging to the semantic field of may encode the experiencer either as subject or as object. a-taabí. () a. o-hiā́ a -need :-cowry ‘He needs money.’ b. a-taabí hiā́ wə. .-cowry need ‘He needs money.’ (Lit.: ‘Money needs him.’)
[Likpe]
(Ameka : )
A number of experiential situations are expressed metaphorically, not necessarily indicating pragmatic markedness. In such constructions, varying degrees of affectedness of the experiencer role may be indicated by using different material verbs. The experiencer can be construed as either the subject or the object, as in (a) and (b), indicating different degrees of affectedness. ()
a. li-kpefí nə́-mə́ lέ bo-fi. -child - hold -sickness ‘The child has fallen sick.’ b. bo-fi a-sɔ́ li-kpéfí nə́-mə.́ -sickness -hit -child - ‘Sickness has befallen the child.’
[Likpe]
(Ameka : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
, like some other emotive and cognitive sensations, is expressed by stating that the experiencer is showing the experiential state. () li-kpéfí nə́-mə́ lέ si-kpi bo-tέ. -child - hold -fear -show ‘The child is afraid.’ (Lit.: ‘The child is showing fear.)
[Likpe] (Ameka : )
can be conveyed by metaphoric constructions involving the stomach. The body part is realized as the subject argument, and the experiencer in subject position is understood to be its possessor. ()
ka-fo a-nyī ́ mε. -stomach -smell ‘I am angry.’ (Lit.: ‘Stomach smells to me.’)
[Likpe]
() i-fi ba-kpεlé ka-fo tíntín -pain :-Likpe -stomach very.much ‘It angered the Likpe very much.’ (Lit.: ‘It pained the stomach of the Likpe very much.’) (Ameka : ) is commonly expressed by a metaphoric construction with either the material verb lέ ‘hold’ or the locative verb kpé ‘be in’. In these constructions the experiencer appears as the indirect object and ɔ́kɔε ‘neck’ as the seat of the emotion. () fufu lέ mε ɔ́kɔε. fufu hold -neck ‘I have cravings for fufu.’ (Lit.: ‘Fufu holds for me neck.’) () ɔ-kuέ bú-su kpé mε ɔ́-kɔε. -farm -go be.in -neck ‘I long to go to the farm.’ (Lit.: ‘Farm going is in my neck.’)
[Likpe]
(Ameka : )
Note that this construction is restricted to expressing ; in order to encode actual location a locative marker would precede the location (Ameka : ).
.. Wolaitta (Afro-Asiatic, Omotic) In Wolaitta, most verbs can neatly be grouped into one-, two-, and three-place verbs, but experiencer verbs do not strictly adhere to this grouping. Some, like and , may have a single argument which is marked for accusative. In the passive construction the experiencer is marked as nominative. () a. táná sákk-eesi. : do.pain-::: ‘I am sick.’
[Wolaitta]
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
b. táaní sák-étt-aisi : do.pain--::: ‘I am sick.’
(Amha : )
Amha’s supplements to the otherwise identical translations of the examples, i.e. (a) ‘I feel pain as I speak’ and (b) ‘state, longer period of sickness’, might indicate that the semantic difference between the two constructions can hinge on the relative permanence of the situation. It is not clear whether this can be related in any way to different degrees of volition and/or control on the part of the experiencer. The following examples show how word order and case marking interact to assign the experiencer role more or less control over the situation described by the predication. () a. táná tukkeé ʔámoy-iisi. : coffee: crave.for-::: ‘I longed for coffee.’ (Lit.: ‘Coffee craved (to) him.’) b. táání tukk-íya ʔámott-aasi. : coffee-: crave.for-::: ‘I craved/longed for coffee.’
[Wolaitta]
(Amha : )
In (a), the experiencer is marked with the accusative, which evokes the reading that he does not have control over his cravings. In (b), however, the experiencer is conceptualized as exerting more control over his feelings. Additionally, (a) displays the unusual word order OSV. It seems to be the case that “when the experiencer is in control, the word order follows the basic SOV pattern. When the experiencer has no control over the situation, it is OSV” (Amha : ). Probably these two strategies of argument assignment interact to highlight the ambiguous attributes of the experiencer: the word order (the experiencer is in subject position, though marked by accusative) reflects his more agentand therefore subject-like properties; the accusative marking, on the other hand, highlights his affectedness.
.. Hausa (Afro-Asiatic, Chadic) Hausa (like some other West Chadic languages; cf. Zoch ) is characterized by a remarkable lexicalization pattern. Many semantic concepts which would typically be realized by intransitive S-verbs (intransitive verbs with an agentive subject) are expressed by action nouns, among them spontaneous bodily reactions like “coughing” and “sneezing”, but also “laughing”, “working”, “lying”, “trading”, and many others. Emotive and physical subdomains like , , , and are likewise instantiated as action nouns. This lexicalized personification of the experitum leads to a host of metaphoric constructions of all possible situational orientations and various degrees of pragmatic markedness. In a pragmatically neutral way these concepts are instantiated in specifying constructions with the verbs yí ‘do’ or jí ‘feel’, as in the following example:
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
() Rasha ta yi fushi da Turkiyya ka-n Syria. [Hausa] Russia : do anger with Turkey head-: Syria ‘Russia is angry with Turkey about/because of Syria.’ (BBC May , )4 Example () illustrates pragmatically less neutral constructions. In (a) the emotion is conceptualized as a container in which the experiencer subject of the clause is immersed. In (b) it is the experiencer object that is filled with the emotion. () a. Magoya baya-n Mursi na cikin fushi. supporter: back-: Mursi in anger ‘Mursi’s supporters are angry.’ (Lit.: ‘ . . . are in anger’.)
[Hausa] (DW: July , )
[Hausa] b. tsoro ya cika zukata-n iyaye a Yobe. fear : fill hearts-: parents at Yobe ‘( . . . ) the parents in Yobe were frightened.’ (Lit.: ‘fear filled the hearts of the parents at Yobe’) (http://hausa.leadership.ng/node/ accessed March , ) In the following experitum-oriented construction the experiencer appears in object position and is perceived as lacking control over the emotion: () tsooɾóo yáa káamaa ní. fear : catch : ‘Fear took hold of me.’
[Hausa]
In () the emotion is attributed to the experiencer subject by a comitative construction: () Waziri ya koma gida da fushi, ( . . . ) Waziri : return house with anger ‘Waziri returned home in an angry mood, ( . . . ).’
[Hausa] (Imam : )
(as well as ), on the other hand, is largely rendered by congruent experiencer-oriented constructions. () har abada ba zan manta da nasara-r ba. [Hausa] until forever : forget with victory- ‘I will never forget this victory.’ (http://www.bbc.co.uk/hausa/sport/// _bale_victory.shtml) Especially with regard to the great variety of metaphoric constructions, it is not easy in many cases to assess pragmatic markedness as well as the perceived level of control and initiation on the part of the participants. Likewise, the conceptual consequences of lexically grouping emotions, physical states, etc., with action nouns merit further study.
4
Most of the Hausa examples are taken from Hausa web pages of news agencies like Deutsche Welle (DW), the British Broadcasting Corporation (BBC), and Radio France Internationale (RFI).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. S
.................................................................................................................................. We have seen that certain broad generalizations may be made for experiencer constructions in probably all the languages taken into account; for example, while experiential constructions in the domains of and tend to be congruent and subject-oriented, languages vary greatly with regard to the available means of codifying emotional and physical state situations. Semantically, the choice of a particular construction often hinges on the perceived level of control on the part of the experiencer (or another participant in the situation). The actual lexico-grammatical means used for the instantiation of such situations depend on the structures available in the particular language.
R Amberber, M. (). ‘Testing emotional universals in Amharic’, in J. Harkins and A. Wierzbicka (eds.), Emotions in Crosslinguistic Perspective. Berlin and New York: Mouton de Gruyter, –. Amberber, M. (). ‘The grammatical encoding of “thinking” in Amharic’, Cognitive Linguistics : –. Amberber, M. (). ‘Remember, remind, and forget in Amharic’, in M. Amberber (ed.), The Language of Memory in a Crosslinguistic Perspective. Amsterdam and Philadelphia, PA: John Benjamins, –. Ameka, F. K. (). ‘The grammatical packaging of experiencers in Ewe: a study in the semantics of syntax’, Australian Journal of Linguistics : –. Ameka, F. K. (). ‘Likpe’, in G. J. Dimmendaal (ed.), –. Amha, A. (). ‘Wolaitta’, in G. J. Dimmendaal (ed.), –. Becher, J. (). ‘Experiencer constructions in Wolof ’, Hamburger Afrikanistische Arbeitspapiere : –. Bhaskararao, P., and Subbarao, K. V. (eds.) (). Non-Nominative Subjects. vols. Amsterdam and Philadelphia, PA: John Benjamins. Bugenhagen, R. D. (). ‘Experiential constructions in Mangap-Mbula’, Australian Journal of Linguistics : –. Comrie, B. (). Language Universals and Linguistic Typology: Syntax and morphology, nd edn. Oxford: Blackwell. Croft, W. (). ‘Case marking and the semantics of mental verbs’, in J. Pustejovsky (ed.), Semantics and the Lexicon. Dordrecht: Kluwer, –. Dimmendaal, G. J. (ed.) (). Coding Participant Marking: Construction types in twelve African languages. Amsterdam and Philadelphia, PA: John Benjamins. Dowty, D. R. (). ‘Thematic proto-roles and argument selection’, Language : –. Halliday, M. A. K., and Matthiesen, C. M. I. M. (). An Introduction to Functional Grammar, rd edn. London: Hodder Arnold. Imam, A. (). Magana Jari Ce. Littafi na Farko. Zariya: Gaskiya Corporation. Kießling, R. (). ‘Wille, Initiierung und Kontrolle: zur Morphosemantik von Experiensverben im Südkuschitischen’, in T. Schumann, M. Reh, R. Kießling, and L. Gerhardt (eds.), Aktuelle Forschungen zu afrikanischen Sprachen (Tagungsband des . Afrikanistentags). Cologne: Rüdiger Köppe, –. Leslau, W. (). Reference Grammar of Amharic. Wiesbaden: Otto Harrassowitz. Noonan, M. (). A Grammar of Lango. Berlin and New York: Mouton de Gruyter.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Reh, M. (a). ‘Experiens-Konstruktionen in afrikanischen Sprachen: Ansätze, Erkenntnisinteressen, Ergebnisse’, in M. Reh, C. Simon, and K. Koops (eds.), –. Reh, M. (b). ‘The role of metaphor in grammar: experiencer constructions’, in M. Reh, C. Simon, and K. Koops (eds.), – Reh, M., Simon, C. and Koops, K. (eds.) (). Experiens-Kodierung in afrikanischen Sprachen typologisch gesehen: Formen und ihre Motivierungen. Hamburg: Institut für Afrikanistik und Äthiopistik. Rögnavaldsson, E. (). ‘Quirky subjects in Old Icelandic’, Papers from the Twelfth Scandinavian Conference of Linguistics, –. Sigurðsson, H. Á. (). ‘The case of quirky subjects’, Working Papers in Scandinavian Syntax : –. Simon, C., and Reh, M. (). ‘Experienskonstruktionen in afrikanischen Sprachen’, Sprachtypologie und Universalienforschung : –. Verma, M., and Mohanan, K. P. (eds.) (). Experiencer Subjects in South Asian Languages. Menlo Park, CA: Center for the Study of Language and Information. Zoch, U. (). ‘Split intransitivity in some West Chadic and Biu-Mandara languages?’, in O. Hieda (ed.), Challenges in Nilotic Linguistics and More: Phonology, morphology and syntax. (Studies in Nilotic Linguistics, .) Tokyo: Research Institute for Languages and Cultures of Asia and Africa, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
ˋ
. I
.................................................................................................................................. E is defined as the study of people–plant relationships (Anderson et al. : ). It refers to a branch of ethnology which is focused on plant-related issues mainly in traditional communities/ethnic groups in, e.g., South America or Africa. The interaction of people with nature is most prominently observed in the countryside, while in urban areas substantial elements of nature are non-existent or blurred. Traditional rural communities have developed their own way of living in harmony with the world around them, making use of nature and natural products or striking a balance with it. The intensive contact of (wo)man and nature with regard to the plant kingdom is expressed in many ways. An important aspect that is an integral part of ethnobotany is the way in which plants are valued and how people bring order to the multitude of plants existing in nature, where in areas of rich biodiversity their number amounts to several thousand species.1 Rural communities, like ordinary people elsewhere, use their own system of coming to grips with plants of different sizes, shapes (plant morphology), habitats, or uses. They do not follow the Linnean way of structuring the plant kingdom, although there are many similarities. This chapter is, first of all, about people’s approach to systematizing plants that they are familiar with by establishing hierarchies and identifying similarities as well as variations. Hence, the chapter discusses the use of African languages in folk taxonomy, which is different from scientific taxonomy and conceptualization. The analysis and illustration of hierarchies, rankings, and other folk taxonomic aspects are impossible without appropriate linguistic terms. Another aspect to be addressed is the interdependence of folk taxonomy and language structures. In this respect, people’s views about plants, and their sizes or uses finds their reflection inter alia in grammatical means and categories that will be dealt with below. 1
Namibia is said to be home to approximately , species.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
This chapter is written from a linguistic perspective, touching upon non-linguistic issues as well, so far as they have to do with human language, and drawing its examples and arguments from African languages. In this respect, the discussion below refers mainly to the Bantu language area. This limitation of not covering the whole of Africa is largely due to the multifaceted linguistic profile of the continent. Accordingly, it was deemed more realistic and plausible for the sake of generalizations to make a selection focusing on a language family that is spread over large parts of the sub-Saharan region and displays a high degree of linguistic cohesion. Moreover, the author is most familiar with Bantu languages and highly experienced in ethnobotanical fieldwork in Tanzania and Namibia.
. F B
.................................................................................................................................. Folk taxonomy, i.e. “a set of categories or taxa arranged so that every taxon is included within one . . . higher order class” (Anderson et al. : ), has been studied all over the world with respect to communities and languages, including African ones. Research results and generalizations on cognitive processes and conceptualization in relation to folk taxonomy are contained in a good number of sources. In particular Berlin and his co-authors’ early findings in the s as well as Berlin () have stimulated ethnobotanic research and data interpretation about taxonomy and plant hierarchies in non-Western communities. Thus, various studies by Heine and co-authors2 as well as the Swahili plant research project, the results of which were presented in Heine and Legère (), benefited much from these ethnobotanic contributions. The categories applied in establishing folk taxonomic hierarchies in the plant kingdom are: • • • •
the unique beginner ( being the most inclusive unit), life forms (limited in number), such as , , generics/generic taxa, and specifics/specific taxa.
These categories will be focused on with regard to Bantu languages in the following sections.
.. Unique beginner The identification of a unique beginner looks quite simple, because the equivalent of the taxon in Bantu languages is widespread. Swahili (SW; G3) dictionaries record the term mmea4 ‘plant’, whose meaning is given in Kamusi (: ) as kitu chenye uhai
Published in a series of five books entitled Plant Concepts and Plant Use: An ethnobotanical survey of the semi-arid and arid lands of East Africa. 3 These codes for Bantu languages are based on the referential classification in Guthrie (–). 4 Derived from the verb stem -mea ‘grow, germinate’. 2
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˋ
chenye majani, shina na mizizi ‘something growing possessing leaves, a stem, and roots’. Similarly, in Ndonga (ND; R) of Northwest Namibia, the term oshimeno (< -mena ‘germinate, grow, strike roots’; Tirronen : ) is defined as ‘plant (like a tree, a tobacco, etc.)’ and confirmed by Ndonga speakers as a unique beginner. Moreover, in Kwangali (KG; K) spoken in western parts of Namibia’s Kavango Province, the lexical item simeno ‘plant’ (< -mena ‘germinate, sprout’; Bredell : ) represents a taxon which may function as an all-inclusive term. Also in Vidunda (VD; G, core areas are southern parts of Kilosa District, Central Tanzania),5 the noun limea ‘plant’ exists. In contrast to this, Mbukushu (K), as spoken in eastern Kavango and northern Botswana, distinguishes thimena ‘plant’ from other life forms such as muhonyi ‘grass(es)’ (singulare tantum), thus not being all-inclusive. Christina Thornell (p.c., January , ) reported that in Mpiemo (MP; Ac, Central African Republic and Cameroon), it was impossible for her to identify any unique beginner. This is, in fact, an observation made by other scholars—the taxon is not accepted as the unique beginner, but only as a life form (sometimes even restricted to cultigens) like a few others, but not higher in the hierarchy. The unique beginner issue does not seem to be straightforward and in harmony with Berlin’s approach in the Bantu languages presented here. Controversial, contradicting views about the role of the taxon and the assumed equivalent in the studied languages transpire also from, e.g., the Zanzibar-based SW dictionary, which describes mmea as jamii ya miti ‘group of trees’ (BAKIZA : ). See also the discussion in Heine and Legère (: –).
.. Life forms Next to the unique beginner is another hierarchical category that refers to plant life forms, also called growth forms, which are described as follows: Plants are classified taxonomically into families, genera, species, varieties, etc. This, however, is not the only way to classify plants. Species and individuals can be grouped into life form or growth form classes on the basis of their similarities in structure and function.6
In SW, the following life forms were identified: • mti ‘tree’ (PL miti), • mmea ‘small, non-woody (annual) plant’ (PL mimea); the basic meaning of this taxon was elicited ultimately as ‘non-tree’. The status of nyasi (plural of unyasi ‘Imperata cylindrical’), which—according to Kamusi (: )—is mmea pori ‘wild plant’ as life form and uyoga ‘mushroom, fungus’ as life form , is disputable.
5 For a Vidunda-related description of common plant names, see Legère et al. () and Legère (). 6 Source: http://landau.faculty.unlv.edu/lifeforms.htm accessed October , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The complexity of the life form issue in SW is apparent also from the following discrepancies. In Zanzibar a few plants are categorized as mboga ‘vegetable’, such as mchicha ‘spinach’ and the cultigens kabeji ‘cabbage’. The mainland approach is different: mchicha is mmea, but kabichi is mboga. The life forms traced in VD are the following: • • • •
idzabi(dzabi) ‘climber, vine’ ( species), igodi ‘tree’ ( species), isolo ‘herb, grass’ ( species), limea ‘plant’ ( species).
This list illustrates that the taxon igodi ‘tree’ plays an important role in folk taxonomy, followed by limea ‘plant’. The taxon idzabi(dzabi) ‘climber, vine’ is peculiar to the VD area and its montane vegetation. In addition, some herbs and grasses constitute the life form isolo ‘grass’. In ND the following life forms were elicited: • • • • • •
eno ‘thorn bush’, omuti ‘tree’, omwiidhi ‘grass, weed’, oshigwanga ‘herb’, oshihwa ‘bush, shrub’, oshimeno ‘plant’ (partly understood as cultigen or species which is distinct from other life forms such as oshigwanga, as well as partly being the unique beginner).
It is interesting to note here the life form oshihwa ‘bush, shrub’ that was identified neither in SW7 nor in VD. In addition, the taxon eno ‘thorn bush’ was found. It is related both to omuti and oshihwa, but resource persons insisted on using this taxon for plants like omungolwe ‘red spike-thorn’ (Maytenus senegalensis), or omunkono ‘blue thorn, yellow bark acacia’ (Acacia erubescens), and some others. Okanyata (not identified) and onziku ‘fragrant grass’ (Kyllinga alba) are those specimens that were quoted as omwiidhi ‘grass, weed’. Furthermore, a look at life forms in KG was taken. The following taxa were traced: • • • •
sihwa ‘bush’ ( specimens), simeno ‘plant’ ( specimens), sitji ‘tree’ ( specimens), wayi ‘herb, grass’ ( specimens).
The emphasis of the folk taxonomic approach in KG is obvious: simeno ‘plant’ is the most frequently used taxon, followed by sitji ‘tree’ and wayi ‘herb, grass’; sihwa ‘bush’ plays a subordinate role. The prominent use of simeno admits of two interpretations. On the one
7
The lexical item kichaka ‘bush’ refers to habitat; it is not a life form.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˋ
hand, this taxon is used by some people as the unique beginner. On the other hand, whenever an appropriate life form is not at hand or difficult to determine, reference is made to simeno.8 Life form data is also available from ethnobotanical studies by Thornell et al. (), Thornell (), and Muzale (), who published a Haya (HA; E) plant name dictionary. According to Thornell et al. (), life forms in MP are • mbon ‘vine, climber’, • le ‘tree’, • ka ‘herb’. These life forms are attested in transcripts of interviews about MP plants and their uses. Accordingly, the taxon mbon ‘vine, climber’ is due to the existence of dense forests close to the equator. The HA data is restricted to plant names in this language, while plant descriptions and uses are summarized in Swahili and English. The following taxa were extracted from this plant dictionary, and HA equivalents (in brackets) added by the author:9 • • • • •
‘bush, shrub’ (ekikamba, ekishaka),10 ‘grass’ (bunyaasi or bunyansi, SG orunyāsi), ‘weed’ (ekishambu; omwata, PL emyata), ‘herb’ (omubâzi, PL emibâzi), ‘tree’ (omúti, PL emíti).
Of these taxa only omúti ‘tree’ is currently accepted as a life form. Substantial folk taxonomic insights were provided by Koni Muluwa (a, b) for a group of B languages spoken in the DRC. Thus, e.g., the following (plural) taxa were elicited in Mpiin: • • • •
bítir ‘herbs’, miʃí ‘climbers, vines’, mitı ‘trees’, nswâs ‘shrubs’.
A similar system exists in Mbuun and Nsong. The neighboring languages Hungan and Ngong, however, lack the life form . Finally, for a summary of life forms in a non-Bantu language of East Africa (i.e. Eastern Nilotic Samburu) Heine et al. () was consulted. In this Kenyan language three life forms are definitely distinguished, namely:
8
A plausible answer to this open question is still pending. Author’s personal communication (Henry Muzale, Dar es Salaam, February , ). 10 As in Swahili, this lexical item refers to the area where particular plants grow, as clarified by Henry Muzale. 9
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
• l-keék ‘trees’, • nyɔɔrtɔ ‘weed’, • n-kújít ‘grass’.11 It is interesting to note that nyɔɔrtɔ ‘weed’ and n-kújít ‘grass’ are treated as distinct taxa. Heine et al. () argue that in Samburu, sub-life forms could be established within the category , i.e. two kinds of thorn trees, as well as leaf trees and latex trees. This taxonomic approach holds water also in languages under study but is not limited to trees only, as the data below illustrate. Thus in VD, sub-life forms include idzabidzabi lya (kadzabi ka) muhulo ‘(small) forest climber’, igodi lya wastani ‘medium-sized tree’ (additionally igodi ikulu (ng’hatu)/ligodi ‘(very) big tree’ and igodi idodo/kagodi ‘small tree’), igodi lya mulyo ‘fruit tree’, etc. In KG, simeno sokuharava ‘creeper’, wayi woyikumba ‘bedding grass’, and the compound noun sitjigona (also with diminutive ka- > katjigona ‘small/short tree’) are sub-life forms. In MP, le lɔn is a savannah tree, mbon mεgwɔ a thorn vine, le mεgwɔ a thorn tree—all proving these taxa to be sub-life forms.12 HA bunyaasi or bunyansi ‘grass’, ekishambu and omwata ‘weed’ and omubâzi ‘herb’ can also be treated as sub-life forms, together with ekiti /akiti ‘small/young tree’. The SW dictionaries offer a wealth of sub-life form taxa, such as mmea wa baharini ‘sea/ aquatic plant’ (BAKIZA : ), mmea wa nyuzinyuzi ‘fibrous plant’ (p. ), mmea unaotambaa ‘creeper’ (pp. , ), and mti wa mwituni ‘forest plant’ (p. ). Mainland dictionary entries list sub-life forms such as mti wenye miba ‘thorn tree’ (Kamusi : ) or mmea pori, quoted above.
.. Common names ... Generics/generic taxa In folk taxonomy the most basic rank is the generic taxon that is monotypic or polytypic, i.e. with or without sub-taxa/specific taxa. Foley (: ) argues that at the generic rank, “the largest number of taxa . . . is found . . . , but these rarely exceed items.”13 It is further emphasized that the generic taxa “are the most salient for the native: they are simple lexemes, most frequently used, learned early by children . . . and most easily elicited from informants” (: ). With regard to the higher order life form the relation is something like “a kind of . . . ”. Generics may further branch into specific taxa or subgenerics. As a rule, many generic taxa are single lexical items, whereas subgenerics/specifics display a more complex linguistic structure. Initially, two generic names are listed below (with the plural noun-class prefix shown in brackets in the first example only). They were traced in East Africa as well as in Namibia (and certainly elsewhere, too). Subsequently,
11 In addition, some Samburu people establish a fourth category which comprises food plants. n-dakí (all data quoted from Heine et al. ). 12 Data extracted from Thornell et al. (); subsequently, upon contact, Christina Thornell made further information about class allocation available. 13 “The least complex taxonomy would consist of an inventory of several hundred such generic taxa”, as Anderson et al. (: ) argued.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˋ
some generics in KG, ND, and VD are listed. Wherever available, the botanical and the English common names are added. euyu (ma-) KG, mubuyu (mi-) LZ (Lozi), omukwa (omi-) ND, divuyu (ma-) MB (Mbukushu), mbuyu (mi-) SW, ibuyu (mi-) VD (Adansonia digitata) ‘baobab’14 rumono KG, mono MB, olumono ND, mbono SW (Ricinus communis) ‘Castor bean/oil plant’ KG ecava (Fockea angustifolia) ‘python vine’ kandjata (Cenchrus biflorus) ‘Indian sandbur’ mbungururu (Mundulea sericea) ‘cork bush, silverbush’ muhengeva (Acacia sieberiana) ‘paperbark thorn’ namayara (Neptunia oleracea) ‘water mimosa’ nkuguru (Cyperus fulgens) simaka (Grewia falcistipula/monticola) ‘raisin bush (silver raisin)’ uhoro (Ficus burkei/thonningii) ‘strangler or (common) wild fig’ ND, KN (Kwanyama) evo, edo (Nymphaea capensis) ‘blue water lily’ kapwidhi (Euphorbia tirucalli) ‘rubber euphorbia’ nakaho (Brachystelma grossartii/arnotii) okaheneyidhi, okalyani (Eragrostis annulata/viscose) ‘ring windgrass’ oluzigo, oudiyo (Adenium boehmianum) ‘quiver tree’ ombindangolo, ombidangolo (Eragrostis lehmanniana) ‘lovegrass’ omugolo(golo), omwoolo (Terminalia sericea) ‘silver cluster-leaf ’ ondjoho, oshosholo (Tribulus zeyheri/terrestris) ‘devil’s thorn’ oshilagula, oshilunda (Tapinanthus oleifolius) ‘mistle toe’ VD chihakahaka (Phytolacca dodecandra) ‘African soapberry’ ihomi (Imperata cylindrical) ‘cogon grass’ mp’hudza (Cucurbita maxima) ‘squash’. What can be inferred from the above overview is that: () the single (monomial) lexemes denote monotypic taxa which correspond to binomial species in botanical nomenclature; () the initial sound, syllable, or segment represents a particular grammatical classification approach which, to a large extent, reflects cognitive criteria. This issue will be dealt with in more detail in section .. Berlin and others argue that, in folk taxonomy, most generics are identical with Linnean ranking and systematic botany species. However, a number of common plant names do not reflect those minor differences in plant morphology that should normally result in the 14
Whenever possible, common names in English are included.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
application of specific names. Thus, ND/KN ondjoho, oshosholo ‘devil’s thorn’ covers two slightly different species, i.e. Tribulus zeyheri and Tribulus terrestris. Or, KG simaka ‘raisin bush (silver raisin)’ means both Grewia falcistipula and Grewia monticola.15 In some cases a generic taxon is used for different species. Thus, VD has chimkwakwatu (Trichodesma zeylanicum) ‘camel/cattle bush’ for a herb, and chimkwakwatu (Spermacoce princeae) for a creeper. It further turns out that the botanist identifies plant samples as belonging to the same species, although different generic taxa/common names were elicited, such as VD ichenje and ilusina, both being Albizia gummifera ‘peacock flower’.16 Variation of generics is widespread in a language like SW spoken over a large area where the observation was made that the smaller a plant is, the more generics/common names are found, sometimes even diverging from one village to the next; hence, broad dialectal variation is not unusual in the case of lingua francas.17
... Specifics/specific (subgeneric) taxa Like sub-life forms, subgenerics/specifics are another way of ranking species. Subgenerics are either: (a) a generic head noun that is modified by an adnominal construction (an adjective or an associative+NP construction), thus being ditypical and at least binomial; or (b) a complex construction that is constituted by a quasi-generic head noun plus adnominal (as (a)) without the head noun functioning on its own.18 In the first case, various generics exist as names of species without modifiers. By adding a modifier, which occupies an adnominal position, a particular species is distinguished from another or several other species. In this case the generic taxon is no longer monotypic, i.e. just naming one species, but at least ditypic for the modifier(s). Below are examples, where an adjective (‘female’, ‘male’, ‘small’) modifies the generic taxon, thus conceptualizing a species of its own in ND. ontululwa (Solanum delagoense) (cf. KN onululu) ontululwa onkiintu (-kiintu ‘female ontululwa’, Solanum kwebense) ontululwa ondumentu (-lumentu ‘male ontululwa’, Solanum incanum) (thorn apple’ ondungu (Capsicum annuum) ‘chili pepper’ ondungu oshona (-shona ‘small ondungu’; Capsicum frutescens) ‘chili’. 15 In this respect, botanist Herta Kolberg (p.c.) of the National Herbarium in Windhoek admits that the various Grewia species are extremely difficult to identify because of the slight variation in plant morphology. Accordingly, folk taxonomy cannot be expected to solve this problematic issue when even professional botanists have not been able to do so. 16 VD plant experts argued against this identification of plant samples by UDSM herbarium botanist Frank Mbago, who used an identical taxon for plants that, in their opinion, are obviously different. 17 See Heine and Legère (: ) who give three reasons for a Swahili generic taxon to be identical all over the Swahili-speaking area (i.e. the plant being culturally or economically important as well as perceptually salient). 18 For structural aspects of complex plant names, see, e.g., Thornell (: –) for MP; Legère () for VD.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˋ
These plant names comprise closely related species, as demonstrated by the botanic names where all belong to the Solanum and Capsicum genera respectively.19 In the following example, omundjulu gwakola, the generic class head noun omundjulu is modified by an adnominal second NP of the associative type (class concord marker guplus associative -a and N, which is ekola ‘crow’, with preprefix e- being dropped). This kind of folk taxonomic conceptualization is frequent, as the examples (the generic taxon on top and subgeneric taxa below) below demonstrate: ND omundjulu (Sesuvium sesuvioides) ‘edible spinach’, also omundjulu gwo gwene ‘real spinach’ omundjulu gwakola (Gisekia pharnaceoides) ‘crow’s spinach’ o- mu- ndjulu gu- akola - - ndjulu - - .kola ekunde (Vigna unguiculata) ‘cow pea’ ekunde lyombambi (Rhynchosia venulosa) ‘Munkoyo roots’ (cl. head noun and modifier as before; ombambi ‘duiker buck’)20 ontanga (Citrullus lanatus) ‘wild water melon, tsamma’ ontanga yethithi (Citrullus lanatus) (ethithi ‘white monster, albino’) ontanga yo(ka)mbalakata ‘wild gourd’ ontanga yontsi (Lagenaria siceraria) ‘calabash, bottle gourd’ (ontsi ‘dust’); a kind of pumpkin (Tirronen : )21 ontanga kaathithi (Cucumis metuliferus) ‘horned cucumber, jelly melon’ (aathithi ‘ancestral spirits’) ontanga kamuthithi (Cucumis anguria) ‘maroon cucumber, West Indian gourd’ (omuthithi ‘ancestral spirit’) KG generics plus adjective or NP rufugu (Sansevieria scabrifolia) ‘bowstring hemp, mother-in-law’s tongue’ rufugu rosigara (Sansevieria pearsonii) ‘spiky mother-in-law’s tongue’ (sigara ‘male kind’) rufugu rosikadi (Sansevieria aethiopica) (sikadi ‘female kind’) sindjembere (Grewia bicolor) ‘white-leaved raisin’, also cultigen ‘grape’ sindjembere somowiza (Indigofera filipes) (mowiza ‘in the forest’, i.e. ‘wild sindjembere’). It is interesting to note that for small plants and even for what is perceived as trees, life form taxa are sometimes applied in the conceptualization process, probably for lack of adequate generics, e.g. ND:
For a list of modifiers, see Legère (: ) for VD and SW; Heine and Legère (: –) for SW; and Koni Muluwa (: –) for B languages of DRC. 20 In closely related KN it is inayembibo, which is a generic taxon. 21 Cf. KG rutanga ‘pumpkin plant’ and rutanga ronontende (Cucumis meeusei) ‘rough-leaved cucumis’, as well as rutanga rwakatjama (Citrus lannatus), all associative adnominals. 19
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
omuti gwaalodhi (Crocoxylon transvaalense) ‘witchdoctor’s tree’, sometimes expanded by gonomegumbo ‘being close to the homestead’ omuti gwelyalya (omukiintu ‘female’) (elyalya ‘cereal-like plant’, Zornia glochidiata) omuti gwelyalya (omulumentu ‘male’) (Polygala schinziana) omuti gwendjandja (Ehretia rigida) ‘puzzle bush’ (endjandja ‘stomach disease’, hence ‘stomach disease tree’); also called omuti gwongombe (ongombe ‘cattle’) omuti gweyoka (Sesbania pachycarpa) ‘snake tree’ (eyoka ‘snake’) omwiidhi gwomugologolo (Sporobolus fimbriatus) (omugologolo meaning not elicited; cf. uugolo ‘gout, rheumatism’, meaning ‘gout medicine grass’) omwiidhi gwontululu (Acroceras macrum) ‘Nile grass’ (ontululu meaning not elicited). Sometimes the folk-naming strategy builds on a complex construction that is constituted by a quasi-generic head noun plus adjective or NP without the head noun functioning on its own, e.g. KG mulilira gomure (Ozoroa longipes) ‘round-leaved resin-bush’ (-re ‘long’) mulilira gomusupi (Ozoroa schinzii) ‘short mulilira’ (-supi ‘short’) VD ichelejembe ikodzi (idala) (Piliostigma thonningii) ‘camel’s foot tree, monkey bread’ (-kodzi /-dala ‘female’) ichelejembe ilume (Terminalia fatraea) (-lume ‘male’) idaka (lya hasi) (Crinum kirkii) ‘Pajama lily’ (-a hasi ‘of low (valley) area’) idaka idung’hu lya kutambala (Brachystegia microphylla) (-dung’hu ‘red’, -tambala ‘creep’; i.e. ‘red creeping idaka’) ikavukavu lya muhulo (Polysphaeria parvifolia) (-a muhulo ‘of river’, i.e. ‘river ikavukavu’) ikavukavu lya uhani (Turraea nilotica) ‘bushveld honeysuckle-tree’ (-a uhani ‘of warm area’). In the above examples, the head noun only occurs in combination with another lexical element (adjective or ASS+NP). Nonetheless, it is also acceptable to delete the head noun modifier, which is parenthesized, like idaka (lya hasi) above, because the other subgenerics would still mark the difference. Another salient conceptualization issue is addressed in the following section, where folk taxonomy in relation to the Bantu noun class system will be analyzed.
. F B
.................................................................................................................................. The taxa, i.e. unique beginner, life form, generic and specific taxa presented so far, display the typical Bantu noun structure, i.e. noun class prefix + noun stem.22 22
The structure of the noun stem will not be further analyzed, since it is of minor interest here for taxonomic reasons.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˋ
Depending on the zone, the noun class prefix may further be divided into a preprefix (also called “initial vowel” or “augment” by some authors), as in ND or HA, and the noun class marker. As a rule, the noun class prefix indicates whether the noun represents singular or plural, but there are exceptions to this rule (e.g. abstract and mass nouns, liquids, etc.). The noun structure of plant names is illustrated by the following examples: ND omusati (Colophospermum mopane) ‘mopane tree’ (cf. KN omufyati) --sati (noun stem) VD chiyowelo (Plectranthus tenuiflorus) -yowelo (noun stem). Nouns in Bantu languages are generally distributed over fifteen noun classes. In addition, there are three locative classes and, in some zones, specific noun classes (augmentatives, diminutives) that are not part and parcel of the whole Bantu area (see also Kleinewillinghöfer, chapter of this volume). Class allocation is determined by the noun class prefix, which in its turn is selected in concordance with the semantic content of the noun stem. Both constituents are equally relevant for the conceptualization process that underlies the development of the Bantu lexical stock, generic and specific taxa included. In recent years, the semantics of the noun classes was summarized by Katamba (: ), who underlined the importance of class pair (also called “gender” by some authors) / for plant names.23 As everybody knows, Katamba and other Bantuists alike have profited fundamentally from the Bantu historical reconstructions of Meinhof (), who named class pair / the “tree class”, and Guthrie (e.g. –). While the results of their contributions deserve acknowledgement, in view of the many / plant names that have meanwhile been elicited, it is time to critically review the whole noun class issue with regard to a more diversified approach to plant taxa class allocation. Thus, it was already demonstrated by Heine () (quoted in Heine and Legère ) that non-tree generics and specifics were traced in classes other than /, such as , /, /, and . Similar conclusions emanate from Legère (), Thornell (), Koni Muluwa (a, b), or can be drawn from dictionaries and word lists such as Gachathi () for Kikuyu (KY), Legère and Munganda () for MB, and from other sources. In fact, the way plant taxa are treated in the grammatical system of Bantu languages is another salient aspect of Bantu folk taxonomy. This view also holds true for the taxa of the folk taxonomic hierarchies touched upon previously.
.. The unique beginner and life forms To judge from Proto-Bantu and individual language studies, class pair / turned out to be the favorite “gender” to which the life form was assigned, e.g. in SW, in B DRC languages, and in ND, as well as for the plural of VD taxa igodi ‘tree’, limea ‘plant, herb’, idzabi ‘climber, creeper, vine’, and isolo ‘grass’.
23
For a profound analysis of noun classes, see also Maho ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Class pair / assignment is relevant for ND thorn tree(s), and the singular class for all VD life forms. Furthermore, / is partly associated with smallness. This concept pertains to the ND life forms oshigwanga, oshihwa, and oshimeno, to MB thimena, as well as to KG sihwa and simeno. In KG, and also in MB, even the life form , i.e. sitji (MB thitondo), belongs to this class pair, a fact that cannot convincingly be commented on as long as no detailed analysis of class pair / in KG and MB has been made. It is assumed, however, that in KG a semantic split has taken place, whereby class noun mutji, which originally meant ‘tree’, at a certain point in history added the meaning ‘medicine’ to its semantic field (cf. ND omuti ‘tree, medicine’ and SW mitishamba ‘traditional medicine’). The taxon was subsequently dropped, since class taxon sitji (PL yitji) had obviously come into existence, thus making the original class noun meaning redundant.24 Class pair / is represented by the ND life form only. Moreover, various sample languages25 (excluding Standard SW) contain class pair / life forms, which are diminutives. Finally, class is relevant for KG wayi ‘grass’, which appears to be singulare tantum. It seems that the allocation of generic plant names and their varieties (see Table .) to a particular Bantu noun class (pair) is, at least partially, determined by the noun class a life form belongs to. Evidently, in the process of conceptualization the allocation of almost all tree taxa in SW to class / has to do with the fact that the life form mti26 belongs to this class. This is also true for ND and the B DRC languages, and with regard to class pair / in KG as well as for the (a)/ “gender” in VD, i.e. VD life forms and (by analogy) common plant names in the singular which belong to class (having drifted away from class ) have still retained the “tree” class plural. The situation is more complex in the case of other life forms, as Table . shows for the sample languages.
.. Generics and specifics/subgeneric taxa The allocation of plant names and taxa to noun classes is, like that of life forms, an outstanding component of folk taxonomy in Bantu languages. This is demonstrated below with reference to the broad and multifaceted approach that was observed in the field and presented in relevant publications. As already stated, plant names were mainly allocated to class pair /. However, folk taxonomy apparently modified the allocation strategy in some languages for reasons that one can merely speculate about: perhaps for a wider knowledge of nature going beyond the identification and naming of trees. Thus, when other useful plants that did not share the tree morphology were found and used, e.g. mushrooms or herbs, the conceptualization process took place in a way which was different from Proto-Bantu class pair /. Accordingly, although probably still maintaining the relevance of class pair /, at least for a while the rich reservoir of other semantically salient class prefixes was (and still is) tapped in the naming process, where, for instance, particular prefixes were associated with smallness
Note, however, the large number of generics in KG class . See Table . for details. 26 Provided that mmea ‘plant’ is accepted as unique beginner in SW, this term is another class pair / example which may be important for / class allocation. 24 25
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
ˋ
Table . Noun class allocation of life forms in selected Bantu languages Class Life form
Language
SG of miʃí, mitı omuti muhonyi omuti mmea, mti miʃí, mitı emiti omiti migodi, mimea, midzabi, misolo mimea, miti eno igodi, imea, idzabi, isolo ligodi, limea, lidzabi, lisolo
DRC HA MB ND SW DRC HA ND VD SW ND VD VD
omano
ND
SG of bítir sihwa, simeno, sitji oshigwanga, oshihwa, oshimeno thimena, thitondo Ø.le, Ø.ka, Ø.mbon bítir yihwa, yimeno, yitji yimena, yitondo bile, bika, bimbon oyigwanga, oyihwa, oyimeno omwiidhi
DRC KG ND MB MP DRC KG MB MP ND ND
oomwiidhi
ND
a
(vine)
DRC
(o)ka (diminutive)
KG, ND, VD
(o)tu (diminutive)
KG, ND, VD
(mushroom) wayi
DRC KG
(class pairs / and /), the concept of long/thin objects (cl. ), or that of specifying one of many (cl. ). With regard to the actual noun class allocation of generic and specific taxa, the following overview illustrates the situation. Noun class pair / was prominently used in the plant-naming process in SW, KY, in B DRC languages, HA, Namibian languages of zone K and R such as KN and ND, KG (only for the singular), LZ, and MB, as well as for the plural of plant names in VD, to give
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
but a few examples. This list could easily be expanded by considering more languages. This class pair is not relevant anymore in MP, and the singular class in VD is almost lost. In KG, there is no longer a class plural morpheme for class nouns (plant names included) because class marker is prefixed to the singular taxon. To judge from the noun class allocation of life forms and elicited plant names, another prominent class pair is /. It is widely used in MP, zone B DRC sample languages, KN, and ND (see Table .), and also in LZ and KG (sixty elicited plant names: trees, but mainly shrubs and smaller plants). In KG a situation is observed where the life form belongs
Table . Noun class distribution of plant names in selected Bantu languages Class Language
Class Language1
a
KG
a
KG
DRC HA KG MB ND SW
DRC % HA Ø MB ND SW VD ( (Heine et al. : ). As Dimmendaal (: ) points out, claims of diachronic unidirectionality need to be corroborated by sound historical linguistic reconstruction. Notwithstanding, the above sequence represents at least an overwhelming tendency. and are central notions in this sequence and have played an instrumental role in the study of grammaticalization in African languages.
... Spatial orientation Because of its great significance in human cognition, the domain of space has also been central in the development of cognitive linguistics since its beginnings in the s. Besides the human body, notions of space are thought to determine some of the most fundamental processes underlying human cognition, on the basis of the tenet that three-dimensional space and our interaction with it is virtually identical for all human beings. The human body has an intrinsic vertical axis, a front-back distinction based on sight and direction of movement, and external forces such as gravity apply to all human beings irrespective of their environment, language, or any other cultural attributes. These aspects are not totally invariable, though, and with the study of grammaticalization in particular with regard to African languages, certain differences between languages were pointed out early on. Already Heine () draws attention to the fact that in some African languages, the body part noun ‘back’ develops into a spatial expression denoting -region (‘on, above’), rather than the expected ‘behind, back, after’. Significantly these
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
are languages used by herder communities, and in proposing his “pastoralist model”, Heine suggests that culturally significant attributes may overrule universal tendencies. In this particular case, while the body schema is still at work, it is not the human body that underlies the construal of spatial notions, but that of the domestic animals kept by the respective language communities. It has thus been a particular strength of scholars of African languages to emphasize not only the universal aspects that cognitive linguists are most interested in, but also the language-specific variations. In addition to allowing for a more relativistic position, the most important aspect of this is the diachronic interest that adds importantly to the early cognitive linguistic insight. The initial interest of cognitive linguists was mostly with the synchronic construal of cognitive categories and their impact on language. African language studies had always paid more attention to diachronic aspects of language change and historical linguistics. The two research strands came into contact when Africanists began to draw increasingly on cognitive linguistic insight since the s and s, and at the same time many general linguists took a stronger interest in diachronic aspects of language change (e.g. Sweetser ; Hopper et al. ).
. C :
.................................................................................................................................. The aforementioned contributions are by no means the only examples of scholarly production at the intersection of African language description and cognitive linguistics (both formally and functionally oriented). The array is rather wide, ranging from Ameka’s work on Ewe grammatical constructions and illocutionary devices () relying on the framework of “Natural Semantic Metalanguage” (Wierzbicka ; Goddard and Wierzbicka , ) to work on experiencer constructions in African languages (Reh ; see also Zoch, chapter of this volume), the linguistic construal of emotion (Ameka ), the grammar of space (Ameka ), and motion/directionality (Treis and Mietzner ; Mietzner and Claudi ). In this diverse body of research a general tendency can be spotted. It is increasingly recognized that the relation between language and thought cannot comprehensively be investigated by linguists, psychologists, and neuroscientists, but rather needs to involve culture as a fundamental ingredient (see Dimmendaal : –). In addition to the longdebated relation between language and thought, culture as a third variable has found its way into this line of research. Where “pure” cognitive linguistics is at times more concerned with mechanisms of synchronic construal (notwithstanding exceptions, as mentioned earlier, e.g. Sweetser ; Jurafsky ), the interest in diachrony made this research tradition interesting to a perhaps less expected audience—cultural historians.
.. Philology and historical semantics: the benefit from cognitive linguistic insight Among general linguists working within the cognitive framework, interest in diachronic aspects of language (and cognition) has grown stronger since the early s. As shown
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
above, this allowed for a synergy not only between cognitivists and descriptivists/typologists working on African languages, but also with philologists and historians. From the perspective of philologists and historians, linguistic universals are less interesting in themselves. Instead, they are more interested in tracing the particularities of individual cases. Here, learning about regularity in semantic change has helped to improve their toolset. Lexical reconstructions are crucial for them, since it is here that cultural history leaves its residue. For long this has been a challenging arena. While the techniques for reconstructing linguistic form by way of the comparative method are quite widely accepted, semantic change was usually deemed accidental. Bréal refutes the idea that words could have inherent tendencies pushing them toward undergoing certain semantic changes as “l’erreur qu’on peut résumer sous le nom de tendances des mots. Rien, au fond, n’est plus chimérique. Comment les mots auraient-ils des tendances?” [the error that can be summarized under the notion of tendencies of words. In the end, nothing could be more delusive. How could words have tendencies?; emphasis Bréal’s] (Bréal : –). When it comes to semantic reconstruction, Anttila (: ) points out that it must ultimately rely on common sense and individual experience. Schadeberg () provides an insightful account of how, sometimes, even rather quirky semantic developments can be traced and explained systematically. His examples from Bantu languages include three unexpected semantic extensions: ‘bending’ ~ ‘nape’ ~ ‘returning home from work’; ‘old age’ ~ ‘thousand’ ~ ‘redemption’, and ‘scraping’ ~ ‘polygamy’. In addition to great familiarity and individual experience, Schadeberg relies on a growing corpus of rich data from Bantu languages, “Bantu Lexical Reconstructions ”. The quality and amount of available data are certainly crucial, but it should also be pointed out that better knowledge of what can be assumed as a plausible change in lexical meaning and regularities in semantic change (Traugott and Dasher ) has also contributed to richer historical semantic accounts. While it is not (and probably will never be) possible to reconstruct proto-meanings in the same way as we can rather reliably reconstruct proto-forms of lexical items, a lot more is known about the plausibility of certain changes in meaning (see Fleisch ). In order to explore this further, the semantic range of spatial expressions in some closely related languages of the Nguni group has been studied in more detail (Fleisch ). The idea was to compare the lexical meaning of these terms across languages in order to learn about the history of this language group. In working toward this aim, a question that is highly relevant to cognitive concerns needs to be addressed: to what extent are diverging meanings of lexical items synchronically construed ad hoc variations, and to what extent do variations in meaning attest to underlying conventionalized lexical polysemies? In other words, how specific is lexical meaning? By applying polysemy tests (Croft and Cruse ; Cruse ), looking into collocation patterns in electronic written text corpora, and by applying certain elicitation techniques, it has been possible to construct radial or network-like structures for these individual terms (for more on the techniques and methods, see contributions in Achard and Kemmer ; and Evans ). Interestingly, some of the attested, fully conventionalized lexical meaning of such highly polysemous Nguni terms as phezulu ‘on, above, over’ were found to be rather detached from what forms otherwise a fairly coherent radial structure à la Lakoff ()/ Brugman (). These are not unlikely to be semantic loans, as in the case of phezulu— usually implying a vertical arrangement in Nguni—being used in Zulu expressions referring to a figure located in relation to a two-dimensional landmark. This replicates the complex
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
English semantic history of ‘on’ and ‘upon’ (cf. houses on the street, Stratford-upon-Avon), with conceptual influence from French (cf. Arles-sur-Tech), but it is difficult to motivate the emergence of this lexical sense from within the semantic makeup of the Nguni term. This approach, which aims at linking diachronic semantics with historical linguistics, is not fully developed at this stage, but it has repercussions on widely employed historical methods such as “words-and-things” (Vansina , ; Schoenbrun ; Stephens , ) and cultural historical reconstructions concerning, e.g., the spread of certain technological innovations (Bostoen , ).
.. Temporality Another important area in which cognitive approaches to African languages have resulted in important insight that has fed back into general linguistic theory is the analysis of tense, aspect, and mood (TAM). African languages differ considerably among themselves with regard to how complex their grammaticalized TAM systems are. Particularly complex systems are found among Bantu languages, many of which distinguish dozens of TAM categories with distinct grammatical markers (for overviews see, e.g., Nurse ; Botne ). Standard views of tense distinguish past, future, and present in relation to the time of utterance, or possibly some other temporal reference point. Even adding notions like “current relevance” at some temporal reference point (allowing for categories like “anterior” or “perfect”) and remoteness distinctions (e.g. characterizing different past categories as “hodiernal”, “hesternal”, “recent”, or “remote”) has not led to a sufficiently stringent and rich model capable of accounting for the complexity in many Bantu languages. On the basis of proposals by Breu (, ) and Sasse (a, b), Drolc () has suggested that the interaction of lexical verb properties and grammatical aspect plays an important role in the TAM systems of Bantu languages (see also Breu ). Similar observations have been made elsewhere in Bantu (Besha ; Mreta ; Fleisch ) as well as for other African languages in many of which a systematic distinction between stative and dynamic verbs or activity vs. change-of-state verbs is made. While the model proposed by Breu and Sasse relies mainly on the boundaries of states of affairs, i.e. situations expressed by verbs, Botne (e.g. ) makes a different proposal concerning the construal of lexical aspect, distinguishing between an onset phase, a nucleus, and a coda. To what extent these components are lexically coded is part of the specific lexical information of a given verb. In addition to this different assumption concerning the construal of how a verbal situation unfolds, Botne () refines the understanding of remoteness distinctions and, perhaps most importantly, Botne and Kershner () argue that two different possibilities for viewing time unfold can be addressed separately by distinct sets of grammatical markers in Bantu languages (and beyond). This is in contrast to most standard views of tense which conceptualize time as an arrow, with the moment of speech marked on this timeline: ego’s here and now. For Botne and Kershner, a significant difference in the construal of time depends on whether ego moves along a more static timeline, as opposed to time moving (either passing by a static ego or “floating” with ego). Different grammatical constructions can target these distinctively, and morphological markers at sub-construction level can be explained very effectively (for more on this, see Botne and Kershner and other work by Botne).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Grammatical descriptions of Bantu languages that have incorporated this insight include Seidel () and Crane (n.d.). It is currently an open question whether this cognitive model of TAM in Bantu should be viewed as specific to these languages or as a universal model that would replace the prevailing current views on how time is cognitively construed. Botne himself proposes this enriched model of temporal construal as an alternative to the standard views that is, in principle, applicable to all languages, although most languages will only exploit part of the overall complexity/maximum number of T/A categories in this model. Bearing in mind the evidence for linguistic relativity (e.g. the work by Levinson and many co-workers at the Max Planck Institute in Nijmegen and elsewhere) and thus quite far-reaching cognitive differences between speakers of different languages, such construals are possibly culturespecific, and explanatory models for grammatical analysis could thus differ considerably between languages.
. A A
.................................................................................................................................. Early contributions within the cognitive paradigm argued that linguistic universals are better explained as similar reactions to comparable human conditions across speaker communities, rather than assuming that individual languages are simply (highly constrained) variations of the same underlying theme, i.e. the human language capacity. Rejecting the idea that ultimately languages show only shallow differences in the realizations of the same (innate) ingredients, this research strand has since increasingly acknowledged that languages are diverse in potentially much more fundamental ways than was recognized earlier. This fits in with a central assumption of “neo-relativistic” research, namely that linguistically entrenched patterns of thought have profound (language-specific) effects on human cognition, addressed briefly in previous sections. The perhaps most fundamental consequence of linguistic relativity would be that a reduced linguistic diversity also implies a reduced cognitive diversity across human societies and cultures. One could expect that the cognitivist stance and neo-relativistic research in combination would force us to postulate a multitude of different worldviews and conceptual universes among the numerous language-cultures on the African continent. Certainly, the particularly high diversity on the African continent, with about , languages, calls for an explanation: how did this fragmented picture come into being? At the same time, it must be questioned whether the actual degree of diversity is indeed as great as the large number of individual languages seems to imply. One possible objection is that, certainly, the diversity—even though possibly more fundamental than acknowledged by some—is still not totally chaotic. After all, there are many shared linguistic as well as cultural properties. An important aspect in this regard has become the study of how areal features in African languages spread. Horizontal
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
transmission, which historically must have been much more significant than some approaches (those focusing almost exclusively on unilinear descent) would have it, has become an important area of study in the field. It has long been known that certain linguistic features show an areal distribution which cuts across genetic family ties between languages. Sprachbund phenomena have been worked on in different areas in Africa (see, e.g., Ferguson , Güldemann and Elderkin ; cf. also Dimmendaal ). Lexical semantics and conceptual patterns have been less thoroughly studied within this approach (but see, e.g., Heine and Kuteva ; Heine ). Some emblematic features have been singled out, and it has even been proposed that some of these imply Africa as a whole could be fruitfully understood as a linguistic area (e.g. the widespread semantic extension of “drinking” for “smoking”, Boyeldieu : ; for a broader range of contributions along similar lines, see Hieda et al. ). Not for Africa as a whole, but possibly on a slightly smaller scale, looking at macro-regions, this approach is reminiscent of what Dixon () proposes for Australia: in a very specific sociolinguistic regime with languages mostly lacking prestige asymmetries, the high degree of fragmentation in terms of individual languages is counterbalanced by a relative homogeneity of conceptual patterns. While the lexicon and morphology of individual languages may differ considerably, the underlying patterns and mechanisms are very similar, enhancing translatability between languages. One of the great challenges that I see at this point in time concerns the functioning of semantics with regard to this diachronic domain. Any approach in which language is understood as simply representing reality will not have much to say on this. In such a limited view, semantics simply reflects what is out there. But if language gains a more constitutive character, we may begin to wonder how semantic properties and conceptual patterns fare in terms of diachronic stability or malleability. In technical terms, this will call for a kind of linguistic anthropological work that is sensitive to detail and of high resolution, but at the same time not totally particularistic. In order to provide richer empirical foundations for historical and anthropological linguistic understanding, it should instead be informed by larger trends in cognitive linguistics, focusing on those which elucidate regularity in semantic change, on the basis of historical pragmatic considerations like those made by Traugott and Dasher () and others who— contrary to Bréal’s statement quoted in section .. above—point out that semantic change is not fully unconstrained (e.g. Heine ). While the renewed acknowledgement of linguistic relativity (in cautious and more sophisticated understandings, e.g. in terms of Bourdieu’s “habitual thought”) seems to provide us with this possibility, one caveat should be mentioned. Any method from within such a mindset will necessarily have to rely to some extent on a notion of “lexical semantic inventories” and conceptual patterns in a given language as a relatively static “residue” of prior collective experience of that speaker community. Over recent years, many linguists— including, in particular, cognitivists—have been moving towards more dynamic views that assume any kind of linguistic construal as an “online process” which may turn out to be problematic when trying to match with a more conventional view of lexical inventories and stable patterns—both necessarily reifying what many cognitivists would assume to be “construed on the go”; cf., e.g., construction grammars (Croft , Evans ) or dynamic syntax proposals (Marten ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
This might create a counter-tendency to the Africanist philological endeavor which, to a considerable extent, needs to rely on a more static notion of languages as conventionalized systems, and the lexicon as a stable residue of prior historical experience of a cultural and linguistic community. How exactly this will play out in the future is a question that is paramount to academic disciplines with a strong philological backdrop. Among Africanist linguists (see Lüpke and Storch ), it is currently widely discussed, and innovative ideas bridging the gap between the “documentary” disciplinary tradition and novel approaches to the study of language, culture, and cognition are likely to emerge from this field.
R Achard, M., and Kemmer, S. (eds.) (). Language, Culture, and Mind. Chicago, IL: The University of Chicago Press. Ameka, F. K. (). Ewe: Its Grammatical Constructions and Illocutionary Devices. Ph.D. thesis. Canberra: Australian National University. Ameka, F. K. (). ‘The linguistic construction of space in Ewe’, Cognitive Linguistics : –. Ameka, F. K. (). ‘Cultural scripting of body parts for emotions: on “jealousy” and related emotions in Ewe’, Pragmatics and Cognition : –. Ameka, F. K. (). ‘Grammar and cultural practices: the grammaticalisation of triadic communication in West African languages’, Journal of West African Languages : –. Anttila, R. (). Historical and Comparative Linguistics. Amsterdam and Philadelphia, PA: John Benjamins. Besha, R. M. (). A Study of Tense and Aspect in Shambala. Berlin: Dietrich Reimer. Bostoen, K. (). ‘The vocabulary of pottery fashioning techniques in Great Lakes Bantu: a comparative onomasiological study’, in A. Akinlabi and O. Adesola (eds.), Proceedings of the th World Congress of African Linguistics, New Brunswick . Cologne: Rüdiger Köppe, –. Bostoen, K. (). ‘Semantic vagueness and cross-linguistic lexical fragmentation in Bantu— impeding factors for linguistic palaeontology’, in W. J. G. Möhlig, F. Seidel, and M. Seifert (eds.), –. Botne, R. (). ‘Cognitive schemas and motion verbs: COMING and GOING in Chindali (Eastern Bantu)’, Cognitive Linguistics : –. Botne, R. (). ‘Remoteness distinctions’, in R. I. Binnick (ed.), The Oxford Handbook of Tense and Aspect. Oxford: Oxford University Press, –. Botne, R., and Kershner, T. L. (). ‘Tense and cognitive space: on the organization of tense/aspect systems in Bantu languages and beyond’, Cognitive Linguistics : –. Boyeldieu, P. (). ‘From semantic change to polysemy: the cases of “meat/animal” and “drink” ’, in M. Vanhove (ed.), From Polysemy to Semantic Change: Towards a typology of lexical semantic associations. Amsterdam and Philadelphia, PA: John Benjamins, –. Bréal, M. (). Essai de sémantique (science des significations). Paris: Hachette. Breu, W. (). ‘Handlungsgrenzen als Grundlage der Verbalklassifikation’, in W. Lehfeldt (ed.), Slavistische Linguistik : Referate des X. Konstanzer Slavistischen Arbeitstreffens. Munich: Otto Sagner, –. Breu, W. (). ‘Interactions between lexical, temporal and aspectual meanings’, Studies in Language : –. Breu, W. (). ‘Der Verbalaspekt im Spannungsfeld zwischen Grammatik und Lexik’, Sprachwissenschaft : –. Brown, P., and Levinson, S. C. (). ‘ “Left” and “right” in Tenejapa: investigating a linguistic and conceptual gap’, Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Brugman, C. (). The Story of ‘over’: Polysemy, semantics, and the structure of the lexicon. New York: Garland Press. Chomsky, N. (). ‘Things no amount of learning can teach’, Noam Chomsky interviewed by J. Gliedman, Omni : November , , http://www.chomsky.info/interviews/–.htm accessed July , . Claudi, U. (). Die Stellung von Verb und Objekt in Niger-Kongo-Sprachen: Ein Beitrag zur Rekonstruktion historischer Syntax. Cologne: Institut für Afrikanistik, Universität zu Köln. Crane, T. M. (n.d.). Completion and dissociation in Totela tense and aspect. Paper based on a presentation at the conference “Bantu ”. Berlin: Zentrum für Allgemeine Sprachwissenschaft and Humboldt-Universität, April , . Croft, W. A. (). Radical Construction Grammar: Syntactic theory in typological perspective. Oxford: Oxford University Press. Croft, W. A., and Cruse, D. A. (). Cognitive Linguistics. Cambridge: Cambridge University Press. Cruse, D. A. (). Meaning in Language: An introduction to semantics and pragmatics. Oxford: Oxford University Press. de Schryver, G.-M., and Nabirye, M. (). ‘A quantitative analysis of the morphology, morphophonology and semantic import of the Lusoga noun’, Africana Linguistica : –. Dimmendaal, G. J. (). ‘Language ecology and linguistic diversity on the African continent’, Language and Linguistics Compass : –. Dimmendaal, G. J. (). ‘Esoterogeny and localist strategies in a Nuba Mountain community’, in W. J. G. Möhlig, F. Seidel, and M. Seifert (eds.), –. Dimmendaal, G. J. (). ‘Perception and the invisible hand among Teso-Turkana groups’, in A. Storch (ed.), Perception of the Invisible: Religion, historical semantics and the role of perceptive verbs (Special issue of Sprache und Geschichte in Afrika ). Cologne: Rüdiger Köppe, –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Dixon, R. M. W. (). The Dyirbal Language of North Queensland. Cambridge: Cambridge University Press. Dixon, R. M. W. (). The Rise and Fall of Languages. Cambridge: Cambridge University Press. Drolc, U. (). ‘Zur Typologie des Perfekts (am Beispiel des Swahili)’, in W. Breu (ed.), Probleme der Interaktion von Lexik und Aspekt (ILA). Tübingen: Max Niemeyer, –. Evans, V. (). ‘The meaning of time: polysemy, the lexicon and conceptual structure’, Journal of Linguistics : –. Evans, V. (). How Words Mean: Lexical concepts, cognitive models and meaning construction. Oxford: Oxford University Press. Ferguson, C. (). ‘The Ethiopian language area’, Journal of Ethiopian Studies : –. Fleisch, A. (). Lucazi Grammar: A morphosemantic analysis. Cologne: Rüdiger Köppe. Fleisch, A. (). ‘Sprachliche Ausdrücke aus dem Bereich “Emotion” im Zulu und Ndebele’, in A. Storch and R. Kastenholz (eds.), Sprache und Wissen in Afrika. Cologne: Rüdiger Köppe, –. Fleisch, A. (). ‘A cognitive semantic approach to and in Southern Ndebele’, in B. Zawada (ed.), Southern African Linguistics and Applied Language Studies (Special issue of Cognitive Linguistics in South Africa ), – . Fleisch, A. (). ‘The reconstruction of lexical semantics in Bantu’, in D. Ibriszimow (ed.), Problems of Linguistic-Historical Reconstruction in Africa (Special issue of Sprache und Geschichte in Afrika ). Cologne: Rüdiger Köppe, –. Goddard, C., and Wierzbicka, A. (eds.) (). Semantic and Lexical Universals: Theory and empirical findings. Amsterdam and Philadelphia, PA: John Benjamins. Goddard, C., and Wierzbicka, A. (eds.) (). Meaning and Universal Grammar: Theory and empirical findings. vols. Amsterdam and Philadelphia, PA: John Benjamins. Güldemann, T., and Elderkin, E.D. (). ‘On external genealogical relationships of the Khoe family’, in M. Brenzinger and C. König (eds.), Khoisan Languages and Linguistics: Proceedings of the st International Symposium, January –, , Riezlern/Kleinwalsertal. Cologne: Rüdiger Köppe, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Gumperz, J. J., and Levinson, S. C. (eds.) (). Rethinking Linguistic Relativity. Cambridge: Cambridge University Press. Heine, B. (). ‘Adpositions in African languages’, Linguistique africaine : –. Heine, B. (). ‘Areas of grammaticalization and geographical typology’, in O. Hieda, C. König, and H. Nakagawa (eds.), –. Heine, B., Claudi, U., and Hünnemeyer, F. (). Grammaticalization: A conceptual framework. Chicago, IL: University of Chicago Press. Heine, B., and Kuteva, T. (). Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heine, B., and Reh, M. (). Grammaticalization and Reanalysis in African Languages. Hamburg: Helmut Buske. Hendrikse, A. P. (). ‘Number as a categorizing parameter in Southern Bantu: an exploration in cognitive grammar’, South African Journal of African Languages : –. Hendrikse, A. P. (). ‘Systemic polysemy in the Southern Bantu noun class system’, in H. Cuyckens and B. E. Zawada (eds.), Polysemy in Cognitive Linguistics: Selected papers from the International Cognitive Linguistics Conference, Amsterdam . Amsterdam and Philadelphia, PA: John Benjamins, –. Hendrikse, A. P., and Poulos, G. (). ‘A continuum interpretation of the Bantu noun class system’, in D. F. Gowlett (ed.), African Linguistic Contributions, presented in honour of Ernst Westphal. Pretoria: Via Afrika, –. Hieda, O., König, C., and Nakagawa, H. (eds.) (). Geographical Typology and Linguistic Areas. Amsterdam and Philadelphia, PA: John Benjamins. Hopper, P. J., and Traugott, E. Closs (). Grammaticalization. Cambridge: Cambridge University Press. Jurafsky, D. (). ‘Universal tendencies in the semantics of the diminutive’, Language (): –, http://www.stanford.edu/~jurafsky/jurafsky.pdf accessed July , . Kashina, K. (). The Silozi Clause: A study of the structure and distribution of its constituents. Munich: Lincom. Kövecses, Z. (). The Language of Love: The semantics of passion in conversational English. Lewisberg, PA: Bucknell University Press; London and Toronto: Associated University Press. Kövecses, Z. (). Metaphor in Culture: Universality and variation. Cambridge and New York: Cambridge University Press. Lakoff, G. (). Women, Fire, and Dangerous Things: What categories reveal about the mind. Chicago, IL: University of Chicago Press. Lakoff, G., and Johnson, M. (). Metaphors We Live By. Chicago, IL: University of Chicago Press. Levinson, S. C. (). Space in Language and Cognition: Explorations in cognitive diversity. Cambridge: Cambridge University Press. Lüpke, F., and Storch, A. (). Repertoires and Choices in African Languages. Berlin and New York: Mouton de Gruyter. Marten, L. (). At the Syntax-Pragmatics Interface: Verbal underspecification and concept formation in dynamic syntax. Oxford: Oxford University Press. Mietzner, A., and Claudi, U. (eds.) (). Directionality in Grammar and Discourse: Case studies from Africa. Cologne: Rüdiger Köppe. Möhlig, W. J. G., Seidel, F., and Seifert, M. (eds.) (). Language Contact, Language Change and History based on Language Sources (Special issue of Sprache und Geschichte in Afrika ). Cologne: Rüdiger Köppe. Morolong, M., and Hyman, L. M. (). ‘Animacy, objects and clitics in Sesotho’, Studies in African Linguistics : –. Mreta, A. Y. (). An Analysis of Tense and Aspect in Chasu: Their form and meaning in the affirmative constructions. Hamburg: LIT. Nurse, D. (). Tense and Aspect in Bantu. Oxford: Oxford University Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Payne, D. L. (). ‘Verb initial languages and information order’, in P. A. Downing and M. Noonan (ed.), Word Order in Discourse. Amsterdam and Philadelphia, PA: John Benjamins, –. Pinker, S. (). The Language Instinct: How the mind creates language. New York: William Morrow. Reh, M. (ed.) (). Experiens-Kodierung in afrikanischen Sprachen typologisch gesehen: Formen und ihre Motivierungen. Hamburg: Institut für Afrikanistik und Äthiopistik, Universität Hamburg. Rosch, E. (). ‘Cognitive representations of semantic categories’, Journal of Experimental Psychology: General : –. Rosch, E., and Mervis, C. B. (). ‘Family resemblances: studies in the internal structure of categories’, Cognitive Psychology : –. Sasse, H.-J. (a). ‘Aspekttheorie’, in H.-J. Sasse (ed.), Aspektsysteme. Cologne: Institut für Sprachwissenschaft, Universität zu Köln, –. Sasse, H.-J. (b). ‘Aspect and aktionsart: a reconciliation’, in C. Vetters and W. Vandeweghe (eds.), Perspectives on Aspect and Aktionsart. Brussels: Éditions de l’Université de Bruxelles, –. Schadeberg, T. C. (). ‘Progress in Bantu lexical reconstruction’, Journal of African Languages and Linguistics : –. Schoenbrun, D. L. (). A Green Place, a Good Place: Agrarian change, gender, and social identity in the Great Lakes region to the th century. London: Heinemann. Seidel, F. (). A Grammar of Yeyi: A Bantu language of southern Africa. Cologne: Rüdiger Köppe. Simango, R. (). The Syntax of Bantu Double Object Constructions. Ph.D. thesis. Columbia, SC: University of South Carolina. Stephens, R. (). ‘Lineage and society in precolonial Uganda’, Journal of African History : –. Stephens, R. (). A History of African Motherhood: The case of Uganda, –. Cambridge: Cambridge University Press. Sweetser, E. E. (). From Etymology to Pragmatics: Metaphorical and cultural aspects of semantic structure. Cambridge: Cambridge University Press. Taylor, J. R., and Mbense, T. G. (). ‘Red dogs and rotten mealies: how Zulus talk about anger’, in A. Athanasiadou and E. Tabakowska (eds.), Speaking of Emotions: Conceptualisation and expression. Berlin and New York: Mouton de Gruyter, –. Traugott, E. Closs, and Dasher, R. B. (). Regularity in Semantic Change. Cambridge: Cambridge University Press. Treis, Y., and Mietzner, A. (). Encoding Motion: Case studies from Africa. Cologne: Rüdiger Köppe. Vansina, J. (). Paths in the Rainforests: Toward a history of political tradition in Equatorial Africa. London: James Currey. Vansina, J. (). How Societies Are Born: Governance in West Central Africa before . Charlottesville, VA: University of Virginia Press. Wierzbicka, A. (). Semantic Primitives. Frankfurt: Athenäum.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.............................................................................................................
LANGUAGE AND SOCIETY .............................................................................................................
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T tradition of writing in Africa has a long history dating back to the fourth millennium , the time of the appearance of the Egyptian hieroglyphs, which still compete for priority in script invention with the Sumerian cuneiform of Mesopotamia (cf. Mitchell ). This first African script later developed into the hieratic and demotic scripts. Recently, perhaps the oldest alphabet was discovered in Wadi el-Ḥôl (Western Desert in Egypt; see Darnell et al. ). Later, other scripts spread through different parts of northern Africa, namely Phoenician, Punic, Lybico-Berber, Meroitic, Coptic, and Nubian. Since the Middle Ages, numerous attempts to adapt Arabic script to write African languages have been known. Presently, Roman-based orthographies are absolutely predominant for official usage in sub-Saharan Africa. The only exception is the Ethiopic script for Amharic and several other languages of Ethiopia and Eritrea. In many parts of Africa indigenous graphic systems were developed. Typologically, they mostly belong to ideographic scripts and even “sentence writing”. Such systems include Bambara and Dogon symbols (Zahan ), sona of Chokwe and tusona of Luchazi (Kubik ), Ewe proverbs (Friedrich ), Gikandi, Adinkra, Nsibidi, and many others (Kubik ; Ogunleye ). In the following sections, we will briefly discuss several dozen modern indigenous African scripts, mostly from the sub-Saharan region. Many of these scripts are syllabaries, which means that separate signs stand for CV (where C can denote both consonant and consonant cluster and V is a vowel, both oral and nasalized) and the shapes of syllables with the same C but different V are unrelated. The most widely used modern syllabary in the world is Japanese kana. One should distinguish another type of syllabic writing system, namely alphasyllabary or abugida (Daniels ). In such scripts, special modifiers are used with some basic sign to denote different vowels following the same consonant. The bestknown examples include various Brahmi-derived scripts of India, namely Devanagari, Bengali, etc. In Africa, the Ethiopic script is an abugida, and in fact the term “abugida” was suggested from the traditional order of this very writing system.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
The issue of classifying indigenous African scripts cannot be addressed from a single point of view only. Often there is no direct evidence of links between the writing systems. For this chapter, we therefore suggest a mixed structure; i.e. a geographic grouping of scripts (Tifinagh family of scripts; West African syllabaries and alphabets; scripts of Cameroon; East African scripts) supplemented by a separate section on writing within religious communities, then analyzing several recent attempts of orthography development from the second half of the twentieth century and the beginning of the twenty-first century and, finally, giving brief accounts of scripts within African diasporas.
. T
.................................................................................................................................. A family of scripts known under a general term “Tifinagh” originated from the LybicoBerber script current in northern Africa in the first millennium . Several varieties of Tifinagh are known to be in use among present-day Amazighs (Berbers). They can be split into two groups: abjad orthographies (known also as consonant alphabets, i.e. scripts with a limited or no notation of vowels) being almost direct descendants of the original script, and Neo-Tifinagh alphabets using the shapes of the original letters (Elghamis ). In southern Algeria, Niger, Mali, and Burkina Faso some modifications of the Tifinagh abjad are used. A typical feature for these scripts is an extensive use of ligatures, which allows distinguishing between consonant clusters and consecutive consonants separated by a vowel; for the latter individual signs are employed. Tifinagh was used in personal communication and has recently started to occur in printed media too. In Neo-Tifinagh alphabets, different systems of vowel notation were introduced. These scripts are also used for personal records and in printed publications as well. In , one of the Neo-Tifinagh ( ) modifications was adopted as the official orthography for the Tamasheq language in Morocco.
. W A
..................................................................................................................................
.. The Vai syllabary * The beginnings of indigenous script invention in Africa in modern times can be traced back to the early nineteenth century. In the s, the Vai script originated in Liberia. The invention of this writing system is traditionally credited to Mɔmɔlu Duwalu Bukεlε, who was assisted by five friends (Dalby ). There is some evidence that the creation of the script could to some extent be the result of “stimulus diffusion” from the Cherokee syllabary, which might have become known in Vai country via the American mission in the s (Tuchscherer ). The script underwent several attempts at standardization. In , Momolu Massaquoi filled in the gaps in the original syllabary to reflect in particular recent phonetic loans. Another standardization was made by a committee at the University of Liberia with elders from several towns of Vai country in .
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Vai syllabary
The Vai script is a syllabary having signs for CV, V, and a syllabic nasal (see Table .). Originally, it also contained some logographic signs, two of which are currently in use as well: ꘓ /feŋ/ ‘thing’ and ꘘ /faa/ ‘die’ (on tombstones). The script is based on morae rather than syllables; for example, nasalization is marked by < ꘋ > /ŋ/, and for long vowels either the vowel is repeated or the syllables from h- or wseries are used (Singler ). Originally, there was no spacing in Vai texts but modern practice is to separate words by a blank (not always coinciding with that in the Roman orthography for Vai) as well as to use Western punctuation instead Vai. It is believed that the Vai syllabary became the stimulus for some other indigenous African scripts, namely Mende, Kpelle, Loma, etc.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Mende syllabary (an extract)
.. The Mende syllabary The creation of the Mende script is linked to two names, Kisimi Kamara (?–), who is credited as inventor in most sources, and Mohammed Turay, who probably began the work on the script in (Tuchscherer ). The first forty-two symbols of the syllabary are in fact an abugida (different vowels are marked by modifications of the basic sign; see Table .). The remaining part is not systematic with respect to shapes. Unlike the Vai script, the Mende syllabary is written right to left, in a fashion similar to Arabic script. There is an interesting system of numerical notation associated with Mende script, which stands almost a step from a true positional (decimal) one; for example, ,, was written as a J shape with six strokes above (Chrisomalis ).
.. The Loma syllabary The Loma script was created in the s by Widɔ Zoɓo and Moriba from Boneketa (Liberia). The syllabary consists of at least symbols for V, CV syllables, and CVV, i.e. consonant + diphthong, syllables. Small crosses (“+” shape) are used as word separators (Joffre ; Dalby ). Certain hints support the claim that the syllabary was used to keep lists of workers at the Firestone rubber plantation. To a limited extent, the script is still used for private records.
.. The Kpelle syllabary The Kpelle syllabary was created in the s by Gbili, a paramount chief of Sanoyea (Bong County, Liberia). Two variants conventionally called “A” and “B” are known. The attested character set of the first variant is more complete compared to the second (Dalby ). A quite recent list of the syllabic signs (Stone ) contains about ninety symbols, including ten numerals (–). It seems that some marginal tone marking was used in
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
the Kpelle syllabary: a small circle preceding a syllable denoted “prefixed low tone”; placed after a syllable it symbolized “syllabic low tone”. One symbol is used to denote syllables with phonologically similar initial consonants belonging to the so-called “mutational pairs” (Dalby ).
.. The Bambara syllabary The Bambara syllabary called Ma-sa-ba ( ) was created in by Woyo Couloubayi from the Kaarta region (western Mali). The symbols of the syllabary follow simple CV or V structure, with only two of them being used to write consonant clusters with a prenasalized /z/: nza and nzε (Galtier ). Initially, no tonal distinction was marked in the script, neither were the vowels marked long or nasal. Later, several diacritics were introduced into the system: a small vertical stroke preceding the syllable was used to mark vowel nasalization, long vowels with tone modulation were marked by a horizontal stroke over the syllabic sign, and short high-tone vowels were marked by an arc on the bottom-right of the respective syllable.
. W A
..................................................................................................................................
.. Bassa Vah alphabet An original alphabet for the Bassa language was created by Dr. Thomas Flo (Gbianvoodeh) Lewis in the early s. The script is known by an indigenous name, “Vah” ( ). The alphabet consists of twenty-three consonant signs, seven vowel signs, and five tonal marks. These marks are placed inside vowels to denote high, mid, low, mid-low, and highlow tones (Dalby ). The systematic use of tonal diacritics distinguishes the Bassa script from most of the African scripts, as even in modern Roman orthographies tone marking is rather rare (Bird ).
.. Nko script Nko ( ) is an alphabet invented in by the Guinean encyclopedist and enlightener Sòlomána Kántε (–). The script was intended as a writing system for the Manding languages of West Africa. It is used primarily for Guinean Maninka (Maninka-Mori). The script also gained some popularity in Liberia, Mali, Côte d’Ivoire, and within the Manding diaspora in Nigeria, Egypt, and elsewhere. Nko is a right-to-left running alphabet consisting of twenty-nine letters: seven vowels, syllabic nasal, and nineteen consonants, plus two characters to denote combinatory phonetic transformations (cf. Dalby ; Vydrin ; see Table .). Seven diacritics are used to mark tones together with vowel length; these are placed above the vowels. Note that the Nko way of tone marking differs from the “Western” linguistics approach, as syllables are not tone-bearing units in Maninka (Vydrin ). The nasalization mark is a dot placed underneath the vowel. Additional diacritics are used in Nko to modify the letters
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Nko alphabet
in order to represent foreign sounds. The Nko script uses indigenous punctuation and decimal digits which are written with the highest digit placed on the right, according to the direction of writing (cf. Arabic script, where the highest digit comes last). The characters within a word are written continuously, joined by a horizontal bar resting on the baseline. The changes between isolated, initial, medial, and final shapes are systematic for all the characters, unlike Arabic script. The Nko script was designed specifically for the Manding languages, and the orthography is quite “shallow” (cf. Coulmas ). The deviations from a one-to-one correspondence between phonemes and graphemes are mostly due to the so-called gbàralí (contraction) rule: if the same short (oral) vowel with the same tone appears in two consecutive syllables, only the second one is written: bólô ‘hand’ is written < > b(ó)ló. There are some exceptions to this rule, in particular for vowels following identical consonants, in polysyllabic words, etc.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.. Wolof Garay script A script for the Wolof language was invented in by Assane Faye. This alphabet named “Garay” ( ) by the inventor has twenty-five letters for consonants. Four vowel symbols can be combined to produce ten vowels. Four diacritical marks are used to denote vowel mutation, long vowels, geminate consonants, and syllable reduplication. Special shapes of letters with a decorative arc are used sentence-initially and also for the first letter of a name (Dalby ). The script is written right to left. Ten symbols for digits follow the Arabic convention, with higher digits standing on the left (as used in the Roman script but contrary to Nko).
.. Hausa scripts While Hausa has a long-established written tradition in both Arabic and Roman scripts, several attempts to introduce a new orthography are known for this language as well. These are mostly individual scripts (see section . below). One of the scripts was promoted in the s by a group of writers called “Raina Kama” from Kano (Nigeria). The proposed script is a right-to-left running alphabet. Some of the publications of this group appeared in the new script (Furniss ). The Hausa Gobiri script, which is connected with the name of Adaré Salifou, was created before and was used to the north of Maradi in Niger (Awagana ). It is a left-toright running alphabet. Some diacritics are used to mark nasalization and long vowels. Indigenous punctuation and decimal numerals are also known. In the s, Musa D. Abdullahi proposed a script for the major Nigerian languages Hausa, Igbo, and Yoruba. This script named “Tafi” by the inventor can be written in a simple linear style, which is predominant. In a second style, vowels are attached below the consonantal signs, which might be slightly modified. This second style resembles alphasyllabary as compared to the alphabet in linear writing. Tones can be marked in Tafi. The symbol set can be extended by symbols for foreign sounds, numerals, punctuation and mathematical symbols, and even some ideographs to denote directions, the five senses, etc.
.. Adlam alphabet An alphabet called Adlam ( ) for the Fula language was created in by Ibrahima and Abdoulaye Barry, brothers from Nzérékoré in southern Guinea (Davydov ). It is a bicameral script containing twenty-eight basic letters for consonants and vowels and some additional letters for foreign phonemes. Adlam is written right to left, with letters being joined within words, which causes slight modifications of their shapes. Ten digits are employed to write numerals, with the highest digit placed to the right, as in Nko. Specially devised Adlam symbols for question and exclamation marks are placed at the beginning of sentences while the Arabic question mark < ¿ > and ordinary exclamation mark < ! > are placed at the end. Several diacritical marks are used in Adlam to indicate vowel length, consonant gemination, and foreign phonemes.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. S C
..................................................................................................................................
.. Bamum script *
,
An original script for the Bamum language (Foumban, Grassfields region, Cameroon) was devised by the local king (mfɔn) Ibrahim Ndjoya with a contribution from Nji Mama Pekeko and Nji “Adjia” Mborong. The first version of the script was created in ; it contained over symbols of an ideographic nature. This version is known as lewa ‘book’. Later, the script underwent a rapid development over six more stages (Schmitt ). The penultimate version from known as a-ka-u-ku from the names of its first syllables, contained eighty syllabic signs, of which ten were also used as numerals (see Table .). In the final version, (mfemfe) a-ka-u-ku ( ), the number of symbols remained the same. Structurally, this writing system is a syllabary with some alphabetic elements. A special sign ndʒəmli is used before names. Two diacritical signs are used to indicate a change in the pronunciation of symbols: kɔʔndɔn (of the circumflex “^” shape, which mostly marks glottalization) and tukwentis (an overline, which mostly cancels the vowel of a syllable converting the sign to a consonant). Other syllables absent from the set of symbols can be written using combinations of two syllables. Indigenous punctuation is present in the script as well. The direction of writing is left to right. The script was used for the Bamum language called “Shüpamɔm” in vernacular and “Shümɔm” for the secret language invented by Njoya.
.. Bagam script The Bagam script was created around , most probably under the influence of the adjacent Bamum script. According to oral tradition, it was invented by a local king (fon), Pufong, assisted by a royal retainer, Nde Temfong (Tuchscherer , ). The script was used for record-keeping and farming calendars, and probably for private correspondence as well. However, the script never gained a wide currency among the Eghap people. So far no samples have been found in the Bagam script material; the only available data are contained in a manuscript deposited by Louis William Gordon Malcolm, a British military officer, in the Haddon Library of Cambridge University (Tuchscherer ). The number of recorded characters in the Bagam script is more than a hundred. Considering the structure of the script, the total number of characters may well amount to several hundred.
. S S E
..................................................................................................................................
.. Somali alphabets The Somali language has a relatively short written tradition. Despite long-established contacts with Arabs, the first attempts to write Somali in Arabic script were made only
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Bamum script
as late as the nineteenth century (Tosco ). Indigenous attempts to create a script for Somali are even younger. The first alphabet was created in by Cismaan Yuusuf Kenadiid. This script is known by the creator’s name as “Osmanya” or “Osmania” ( ). In its final stage of development, the alphabet had twenty-two letters for consonants, five for short, and three for the long vowels /a, e, o/. The remaining two long vowels, /i/ and /u/, were written with letters for /y/ and /w/, respectively, thus showing the influence of Arabic. Ten digits also have special signs in the Osmanya script.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Table . Somali scripts
In , an alphabet was invented by Cabduraxmaan Nuur from the Gadabuursi clan in the Borama District (northern Somalia). Accordingly, this writing system is known as “Gadabuursi” or “Borama” script (Lewis ). Another alphabet was proposed in by Xuseen Sheek Axmed Kaddare from the Abgaal clan. In comparison to the two above-mentioned scripts, this writing system known as “Kaddariya” has the most cursive style (see Table .).
.. Oromo script An alphasyllabary for the Oromo language was developed in by Shaykh Bakri Saṗalō (Hayward and Hassan ). Structurally, it was modeled on the Ethiopic script, with strong adaptations to Oromo phonology. For instance, there are seven vowel rows in the Ethiopic script as compared to thirteen in the Oromo script. Modifications of the basic consonant shape were used to denote geminate consonants.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. S A
..................................................................................................................................
.. Beria script * An indigenous script for the Beria language was proposed in by Adam Tajir (Ustaz Adam Abdalla). The shapes of symbols were based on Beria cattle brands that were mainly used to mark camels; therefore, the script is also known as “camel alphabet”. This original alphabet contained about forty letters showing a strong impact of Arabic phonology in the consonant set; it was supplemented by numerous vowel letters. In , Siddick Adam Issa improved the script by reducing the number of letters to twenty-four, thus making it better suited for the Beria phonetic system. In this version of the Beria alphabet capital and lower case forms are distinguished, and special diacritics are used to mark stress, rising and falling tones, and the [ATR] feature.
.. Nubian scripts The Nubian language has one of the longest written traditions not only in Africa but in the world as a whole. With due care and attention it can be traced back to the Meroitic script, the first proper sub-Saharan writing system. There is some evidence indicating that Meroitic and Old Nubian were related languages (Rilly ). Old Nubian was written in an alphabet derived from Coptic, with the addition of three letters for Nubian-specific sounds, presumably derived from Meroitic. In modern times, several writing systems are known for Nubian. These include modifications of Arabic and Roman scripts as well as orthography based on the Old Nubian alphabet. This does not exclude additional indigenous efforts at writing Nubian. An original script was proposed c. by Omar Hassan Al-Daboodi for the Ken(u)zi dialect of Nubian. The script is a right-to-left running alphabet (Jaeger ). The letter shapes show significant influence of Arabic script. Vowels are marked with special diacritics placed on the top-left corner of the preceding consonant letter, which again may be considered as the influence of harakāt (Arabic short vowel marks). Al-Daboodi attempted to popularize : his alphabet by distributing copies of a hand-written primer ‘ ﻫﻴﺎ ﻨﺘﻌﻠﻢ ﺎﻟﻠﻐﺔ ﺎﻟﻨﻮﺑﻴﺔLet’s learn Nubian!’) (c.). It is also worth mentioning an attempt by El Tayeed Saeed, a graphic designer from northern Sudan (presently residing in the Netherlands); in , he proposed another alphabet for Nubian whose letter shapes are based on those of the Meroitic script. Saeed’s script is a simple left-to-right running alphabet with twenty-six letters.
. S
..................................................................................................................................
.. Medefaidrin (Obεri Ɔkaimε) script The Medefaidfrin alphabet was created by Michael Ukpɔn and Akpan Udɔfia for an artificial language used within an indigenous Christian church, Obεri Ɔkaimε, in the s near Itu on
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
the western bank of Cross-River (Calabar Province, Nigeria). The alphabet was strongly influenced by English orthography: not only are capital and small letters distinguished by shapes, which is uncommon for most indigenous scripts, but a separate ideograph is used for “I” (atiu). Vigesimal (base ) numerals are associated with the script (Hau ).
.. Mandombe script The Mandombe ( ) script was invented by David Wabeladio Payi in . It is used by representatives of the Kimbanguist Church, mainly for Kikongo, Lingala, Tshiluba, and Swahili but can also be employed to write Wolof, Yoruba, Igbo, French, etc. While the main languages for which Mandombe was designed belong to five-vowel Bantu, Lingala has seven vowels. However, as noted by Bokamba (), in urban Lingala of Kinshasa there is a tendency to shift open [ε] and [ɔ] to their closed counterparts [e] and [o]. Not surprisingly, only five vowels are distinguished in Mandombe, plus the recent addition of a symbol for /y/ ( German /ü/). Structurally, Mandombe is an abugida: a syllable is formed by conjoining a consonantal sign with a vowel which appears in different orientations. There are special marks to denote nasalization, diphthongs, consonant clusters with /r/ and /l/, etc. Also, syllable-final consonants have separate signs. The direction of writing is left to right, with syllables within words separated by a blank and a dot serving as a word separator. Indigenous punctuation and digits are used in Mandombe (Rovenchak and Glavy ).
.. Yoruba “holy” writing Josiah Olunowo Oshitelu, a founder of the Church of the Lord (Aladura) in western Nigeria, created a script for writing an artificial language (Dalby ) presumably used for a while within the Aladura Church. The script is a right-to-left running alphabet that has remained undeciphered to this day. Except for letters provisionally identified as vowels and consonants, there are at least two “seal” signs: one used as the personal Oshitelu symbol and the other being the seal of the Church.
. I
.................................................................................................................................. Recent inventions of writing systems in Africa constitute an interesting phenomenon. Almost all script types are attested here, ranging from scripts with ideographic elements to simple alphabets. Such attempts are numerous, and in this section we will only describe a selection of scripts with some interesting features.
.. Bété script * The Bété script was invented in by an Ivorian artist, Frédéric Bruly Bouabré (–). The syllabary amounts to more than symbols of highly pictorial shapes. Reduplication of
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
symbols is a productive strategy for obtaining new characters, and a special mark is used to separate two consecutive identical characters if required (Riley ). Often several symbols correspond to one syllable, and this can mean that tonal distinctions are incorporated as a feature of some syllables.
.. Nwagu Aneke Igbo syllabary A syllabary for the Umueri (Umuleri) dialect of the Igbo language (Anambra State, Nigeria) was proposed around by Ogbuevi Nwagu Aneke. Before his death in a car accident in , he managed to write about a hundred notebooks with texts on mainly religious and anticolonial topics (Azuonye n.d.). The syllabary is rather imperfect. Many symbols have the same or very similar shapes, and there are no signs for individual vowels. Thus, for instance, the widespread name “Obiọma” is written with just two syllabic signs, bi-ma. Several dozen ideographic symbols are also used in the Nwagu Aneke script in addition to syllabic signs.
.. Aka Umuagbara script The script called “Aka Umuagbara” was developed by Dr. Ogonna Anaagudo-Agu from Calabar (Nigeria) in the s. It was designed for writing Igbo, and symbol shapes are based on traditional graphic designs within Igbo known as Uli symbols (Anaagudo-Agu ). The writing system is of a mixed type and is still being further developed (Rovenchak and Glavy ). Several dozen symbols are ideographs; apart from this, the basic unit of the script is the syllable. Vowels are attached to the basic consonantal sign as diacritics in different positions; in many cases the symbols without a diacritic can have the phonetic value of “consonant + /a/”. In several cases, consonant symbols show the influence of Roman script. Some syllables are written differently when used as grammatical markers (third person marker ya, past tense identifier -lu, continuous aspect formative na-).
.. Mwangwego script * This rather new writing system was invented in by Nolence Moses Mwangwego from Malawi. It was designed for writing languages spoken in Malawi and neighboring countries. The Mwangwego script is one of just a few African examples of abugida. An inherent vowel is /a/, and systematic shape changes produce syllables with the vowels /ε, i, ɔ, u/. Consonant clusters are formed by means of special diacritics called mituyo. An acute-like diacritic is used to mark the high tone when required. The Mwangwego script is unique in the sense that it has symbols for click consonants represented in Bantu Roman orthographies by and (Rovenchak and Glavy ). In , a book entitled (A Malawi tili pati?, ‘Malawians, where are we?) was published using this script in the Chewa language.
.. Other alphabetic scripts At least two alphabets for the Fula language have been known since the s–s. In , Adama Ba from Mali invented a script for his poetry and private correspondence
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
(Dalby ). Between and Oumar Dembele from Mali promoted the invented script called “Dita” (Dalby ). A special sign with the phonetic value /ã/ was used to mark nasalization. Both Fula alphabets have indigenous signs for decimal digits. Even today, further attempts to introduce indigenous orthographies for Fula are being made (cf. also Adlam, section ..). At least three new Fula scripts circulate in Guinea, presumably incited by the existence of Nko for the Maninka language (Vydrin , p.c.). The Yoruba FaYé alphabet was invented in by Remi-Niyi Alaran from Nigeria. Interestingly, in this script tone distinction is achieved by using mirrored shapes of letters for vowels (Rovenchak and Glavy ). Finally, we will mention some new scripts from the last decade. They are of a different nature but contain alphabetic elements: Luo or Lakeside script (development started in ) by Kefa Ombewa from Nairobi, Kenya, and Paul Sidandi from Francistown, Botswana; Akagu script for Igbo, suggested in as a part of Neo-Nsibidi, a mixed-type writing, partly ideographic and partly alphabetic, inspired by Nsibidi; a writing system called Ditema tsa Dinoko in Sotho or IsiBheqe SoHlamvu in Zulu was devised in / for South African languages.
. S D
.................................................................................................................................. Writing systems linked to Africa by their inspiration and origin and also scripts for African languages proper are known from within the African Diaspora. These scripts are typologically different, as are the scripts of Africa. For instance, Anaforuana, a graphic system of the Abakua secret society in Cuba, can be traced to Nsibidi or similar systems from Calabar (Mafundikwa ). The Afaka syllabary for the Ndjuka language (an English-based creole of Surinam) was created in . It contains fifty-six syllabic signs and an ideograph for Halley’s comet (Gonggryp ). An alphasyllabary for the Bantu and Bantoid languages was proposed in by Franklin Romuald Ngamga, of Cameroonian origin, and Hye Yeon Nam. In this writing system named “Kii” ( ) several basic consonantal shapes are modified by diacritics for vowels and diphthongs. In , the Ndebe ( ) Igbo script was invented by Lotanna Ọdunzẹ, a native of Anambra (Nigeria). This writing system is an alphasyllabary; vowel diacritics are placed above the consonants that they follow. Different marks are used for vowels with different tones. Words are separated by a middle-line dot (Rovenchak and Glavy ). In , an alphabet for the Wolof language was suggested by Saliou Mbaye, a native of Thiès (Senegal). The letter shapes of this left-to-right running alphabet are based on the acrophonic principle (Rovenchak and Glavy ).
R Anaagudo-Agu, O. C. (). ‘Aka Umuagbara script: Uli essence and symbols in the development of an indigenous Igbo script’, The Parnasus: University of Uyo Journal of Cultural Research : –. Awagana, A. (). ‘Die Hausa-Schrift’, in A. Kootz and H. Pasch (eds.), Jahre Schrift in Afrika. Köln: Universitäts- und Stadtbibliothek, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Azuonye, C. (n.d.). The Nwagu Aneke Igbo Script: its origins, features and potentials as a medium of alternative literacy in African languages. MS. Bird, S. (). ‘Strategies for representing tone in African writing systems’, Written Language and Literacy : –. Bokamba, E. G. (). ‘The spread of Lingala in the Congo Basin’, in F. McLaughlin (ed.), The Languages of Urban Africa. Continuum International Publishing, –. Chrisomalis, S. (). Numerical Notation: A comparative history. New York, NY: Cambridge University Press. Coulmas, F. (). The Blackwell Encyclopedia of Writing Systems. Malden, MA, Oxford, and Carlton, Victoria: Blackwell. Dalby, D. (). ‘A survey of the indigenous scripts of Liberia and Sierra Leone: Vai, Mende, Loma, Kpelle and Bassa’, African Language Studies : –. Dalby, D. (). ‘Further indigenous scripts of West Africa: Manding, Wolof and Fula alphabets and Yoruba “holy” writing’, African Language Studies : –. Daniels, P. T. (). ‘Fundamentals of grammatology’, Journal of the American Oriental Society : –. Daniels, P. T., and Bright, W. (eds.) (). The World’s Writing Systems. Oxford: Oxford University Press. Darnell, J. C., Dobbs-Allsopp, F. W., Lundberg, M. J., McCarter, P. K., Zuckerman, B., and Manassa, C. (). ‘Two early alphabetic inscriptions from the Wadi el-Ḥôl: new evidence for the origin of the alphabet from the Western Desert of Egypt’, The Annual of the American Schools of Oriental Research : , , –, –, –. Davydov A. V. (). ‘Miriden and Adlam—two new alphabetic scripts for the Maninka and Fula languages in Guinea’, in A. Zheltov (ed.), African Collection—. St. Petersburg: Museum of Anthropology and Ethnography, –. Elghamis, R. (). Le tifinagh au Niger contemporain: étude sur l’écriture indigène des Touaregs. Doctoral thesis. Leiden University. Friedrich, J. (). Geschichte der Schrift unter besonderer Berücksichtigung ihrer geistigen Entwicklung. Heidelberg: Carl Winter. Furniss, G. (). ‘Documenting Hausa “market” literature’, in T. A. Barringer (ed.), Africa Bibliography . Edinburgh: Edinburgh University Press, vii–xxxiii. Galtier, G. (). ‘Un exemple d’écriture traditionnelle mandingue: le “Masaba” des Bambara-Masasi du Mali’, Journal des Africanistes : –. Gonggryp, J. W. (). ‘The evolution of a Djuka-script in Surinam’, Nieuwe West-Indische Gids : –. Hau, K. (). ‘Obεri Ɔkaimε script, text, and counting system’, Bulletin de l’IFAN : –. Hayward, R. J., and Hassan, M. (). ‘The Oromo orthography of Shaykh Bakri Saṗalō’, Bulletin of the School of Oriental and African Studies : –. Jaeger, M. (). ‘Indigenous efforts to revitalize and digitize the Nubian languages’, Sudan Studies Association Newsletter : –. Joffre, J. (). ‘A new West African alphabet, used by Toma, French Guinea and Liberia’, Man : –. Kubik, G. (). ‘African graphic systems’, Muntu, revue scientifique et culturelle du Centre International des Civilisations Bantu (CICIBA) –: –. Kubik, G. (). Tusona-Luchazi Ideographs: A graphic tradition of West-Central Africa, nd edn. Berlin, Hamburg, and Münster: LIT. Lewis, I. M. (). ‘The Gadabuursi Somali script’, Bulletin of the School of Oriental and African Studies : –. Mafundikwa, S. (). African Alphabets: The story of writing in Africa. New York: Mark Butty. Mitchell, L. (). ‘Earliest Egyptian glyphs’, Archeology : –. Ogunleye, T. M. (). ‘Àrokò, Mmomomme, Twe, Nsibidi, Ogede, and Tusona: Africanisms in Florida’s self-emancipated Africans’ resistance to enslavement and war stratagems’, Journal of Black Studies : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Riley, C. (). Field notes on the Bété syllabary repertoire. MS. Rilly, C. (). La langue du Royaume de Méroé. Paris: Éditions Champion. Rovenchak, A., and Glavy, J. M. (). African Writing Systems of the Modern Age: The sub-Saharan region. New Haven, CT, Buena Park, CA, New Rochelle, NY, London, Lviv, and Abidjan: Athinkra. Schmitt, A. (). Die Bamum-Schrift. vols. Wiesbaden: Otto Harrassowitz. Singler, J. V. (). ‘Scripts of West Africa’, in P. T. Daniels and W. Bright (eds.), –. Stone, R. M. (). ‘Ingenious invention: the indigenous Kpelle script in the late twentieth century’, Liberian Studies Journal : –. Tosco, M. (). ‘Somali writings’, in A. Kootz and H. Pasch (eds.), Jahre Schrift in Afrika. Special issue of Afrikanistik Aegyptologie Online, https://www.afrikanistik-aegyptologie-online.de/ archiv// accessed July , . Tuchscherer, K. (). ‘The lost script of the Bagam’, African Affairs (): –. Tuchscherer, K. (). ‘History of writing in Africa’, in K. A. Appiah and H. L. Gates Jr. (eds.), Africana: The Encyclopedia of the African and African American Experience, nd edn. New York: Oxford University Press, –. Tuchscherer, K. (). ‘Recording, communicating and making visible: a history of writing and systems of graphic symbolism in Africa’, in C. Mullen Kreamer, M. Nooter Roberts, E. Harney, and A. Purpura (eds.), Inscribing Meaning: Writing and graphic systems in African art. Los Angeles: Smithsonian Institution, –. Vydrin, V. (). Manding English Dictionary (Maninka, Bamana). Vol. . St. Petersburg: Dimitry Bulanin. Zahan, D. (). ‘Pictographic writing in the Western Sudan’, Man : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. W on the ideology of language planning theory, Tollefson () argues that the study of language planning cannot be done legitimately outside the historical-political context of the community of speakers. Treating languages and their speakers as abstract constructs, evaluating success of language policies against the set objectives is using an irrelevant positivist paradigm which ignores the nuances of human relations issues and sociopolitical manipulations that underlie language policies. For his part Fairclough proposes a “critical language study” which would “correct a widespread underestimation of the significance of language in the production, maintenance, and change of social relations of power” (Fairclough : ; emphasis added). Language policy has social, legal, political, ethical, and educational consequences that can be best understood only when we consider the politics and history of a country. These consequences often engender disputes which touch on human rights, values, recognition, freedom, equity, fairness, etc. The resolution of the disputes lies in correcting the unequal and inequitable distribution of opportunities, goods and services, and reduced or prohibited public recognition of languages, all of which depend on existing power relationships. In other words, the state must decide which language(s) will be used to carry out certain functions in public institutions like hospitals, schools, the workplace, etc. The question is how this is done without disadvantaging any group(s). The present multilingual configurations in Africa are partly a function of the colonial past and also that of the recent wave of globalization processes being driven by ICT and its attendant language forms and discourse protocols. I will not dwell on colonial language policies, but on contemporary language policy issues in selected African countries. I will use a modified conceptual framework adopted from Patten (). According to him, “problems of language policy that are most contested are those of public recognition and individual autonomy. In his definition . . . a language enjoys public recognition when it
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
is possible to access services and/or conduct public business in that language” (Patten : ). On the other hand, a person is deemed to be enjoying linguistic autonomy “to the extent that he is free from state interference to select which language he will use in various non-public domains” (p. ). For the purposes of this chapter, I will consider only public recognition concerns.
. P
.................................................................................................................................. The immediate intuitive reaction to the debate on state control of language reminds us of the famous warning by the early descriptivists: Leave our language alone! If all liberal democracies have no official religions, why should they have national and official languages? The very act of legislating which language(s) to use in school, parliamentary debates, or for provision of health services is to bestow privilege on the chosen language(s) and, of course, their speakers. And naturally the speakers of the non-chosen ones will feel aggrieved in the same way that members of a non-official religion would feel discriminated against. However, individuals can practice their faiths privately or clandestinely in groups, but how possible is it for the state not to intervene in matters involving which language to teach children in or which one to use on tax forms and other official documents? All public services and business will inevitably require the use of one language or another. This is particularly true in the complex linguistic plurality in Africa and many parts of the developing world. To discuss this in a more orderly way, Patten breaks down public recognition into what he calls “models of thinking about language issues”. These are official multilingualism, language rationalization, and language maintenance models. For the purposes of this chapter I will discuss only the first two models.
.. Official multilingualism The official multilingual model maintains that each of the various languages spoken in the community should be accorded the same recognition (Patten : ). This is good news for Linguistic Human Rights activists who invoke moral imperatives in language planning projects, that is, all people have the right to learn and use their own language in all significant public and private situations. This assumes that all mother tongues are best for effective communication and provide the requisite social prestige and cultural identity in all spheres of a person’s life (May ; Skutnabb-Kangas ). But how practicable is this in multilingual settings with several competing mother tongues? In an approximation of ideal official multilingualism the South Africa constitution designates eleven languages “official” and all other indigenous languages “national”. The eleven languages are assumed to represent all the language interests in South Africa. Similarly, Ethiopia has declared English and Amharic official at the federal level and recognizes other Ethiopian languages as regional official languages. In both cases the state is officially multilingual and is trying to placate all language families in the case of South Africa and all language territories in the case of Ethiopia, if not all individual languages. The South African language policy is a function of its colonial political history, where English and Afrikaans were the only official languages;
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
African languages were restricted to education for black children and as subjects of academic research. In feudal Ethiopia, Amharic was the official language until the overthrow of the last emperor in . Subsequent governments have liberalized language planning to include all local languages in one way or another so that all communities should feel included. This is in conformity with the federal constitution in place. The ideological underpinning is that all languages or language groups receive official recognition and treatment in and by all public institutions. This is a case of the state supporting linguistic nationalisms and celebrating linguistic diversity. But do they all enjoy same recognition? Hardly! In practice, the result of the seemingly egalitarian policy has been inequality whereby people neglect or reject outright their mother tongue and demand to be given more of others’ language(s). There often emerge disparities as one or more languages acquire more public space and social prestige than others through official manipulation. For example, in South Africa, even a relatively strong official language, like Afrikaans, is being threatened as the country begins to go the “English-only” way. Some of the former Afrikaans-medium schools have converted to English-medium and parliamentary debates take place predominantly in English only (Ridge ; Brock-Utne et al. ). The official multilingual policy has created a fertile ground for English as a linguistic hegemon, leaving the ten other languages as ceremonial symbols. Another serious backlash of the policy is the growing unpopularity of the so-called Bantu languages. There is declining demand for places to study African languages at South African universities, which is threatening the survival of departments of African languages which used to be vibrant and attractive during the apartheid era. Rejection of the African mother tongues by black South Africans is a statement that they are now free from the apartheid-era language policy that forced them to learn in the Bantu languages as a way of keeping them down. The preference for English is not simply a policy issue but also a function of the colonial history of South Africa and other former British colonies, where the patterns of language use and the social values that were associated with them became part of the value systems. “This backdrop is significant in the location of the English language in the psyche of these societies on the one hand as the marker of a past of political subjugation and on the other as a key to power and better life” (Brock-Utne et al. : ). Of course, when it is a choice between the sentimental rejection of the past and real-time socioeconomic benefits, the latter always wins. In Ethiopia, some regional states have opted to make Amharic the official language, even though it is not indigenous in those regions; they teach it as a subject, and use it as medium of instruction and also in the regional legislative assemblies. This means that Amharic is the hegemon at the regional level, stifling the growth of the other mother tongues. English is the medium of instruction from secondary school to university despite the monumental communicative and educational problems that are associated with the policy (Heugh et al. ). English is preferred by officialdom and non-Amhara citizens alike; this is in total disregard of the language-of-wider-communication role played by Amharic. Amharic facilitates governance at both regional and federal levels, and general political discourse. English still features prominently in written official communication, whereas Amharic is the de facto official language in spoken communication in public offices and in business across communities. In spite of a fairly well-standardized Amharic, the politics of language
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
cannot allow the country to boldly adopt it as the only official language of Ethiopia without offending the sensitivities of the other language communities. This model is a good product of liberalism with a tinge of egalitarianism that superficially seems to celebrate all languages, but in the real world that we live in, it actually strengthens the strong languages but exposes the weaker languages to the vagaries of relentless linguistic competition that often has disastrous consequences. This is what Patten (: ) describes as equal “distribution of institutional space and capacities . . . which should have people as equal as possible in the degree to which the languages they speak are publicly recognized.” At best this is a wonderful utopian model, but it is not feasible in a complex multilingual Africa. A more democratic policy would be that which compels public institutions to promote the maintenance of vulnerable languages so as to stem language shift. If the model has worked well in Switzerland with four languages and in Belgium with three, it is because it is accompanied by rigid observance of politically defined boundaries and other socio-historical factors. Mother tongue speakers must remain sufficiently vigilant and motivated politically so that no overlaps occur, especially those encouraged by official policy.
.. Language rationalization The ideology of language rationalization involves a “program of promoting convergence on a privileged public language (or set of languages) by limiting or denying recognition of other languages in certain spheres of language use” (Patten : ). In most African countries this was the popular model, and it explains the privileged position of the European languages in and by public institutions. The myths surrounding this policy were: () () () () ()
They were neutral and so would facilitate national unity. They were suitable for technology and higher education. They would facilitate intercommunity communication. They would provide equal opportunity for social mobility. They would encourage political participation by all.
One of the most misguided policy decisions was that of adopting a European language as a uniting factor in an African country with several competing indigenous languages. The underlying ideology is that diversity begets disunity and, since it would be politically imprudent to pick one of the local languages, a neutral European language was the best bet. It is like arguing that the European evil would be shared equally by all instead of privileging one of our own. Rwanda, Burundi, Somalia, and, to a small extent, Botswana are almost unilingual. There is in each of these countries a dominant African language with its dialects and a few genetically related or non-related minority local languages. Before colonization, these speech communities coexisted with minimal communication problems. By introducing colonial administration and adopting a European language as a symbol of state power, intercommunity boundaries seem to have hardened and communities grew less tolerant of each other; even dialects emerged as autonomous languages. There were ethnic skirmishes before colonialism but nothing on as a massive scale as what we witnessed in Rwanda in or the persistent lawlessness in Somalia running into three
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
decades. The nineteenth-century European ideology of reifying language and making it a symbol of differentiation of communities on the basis of as little evidence as phonemic differences magnified the differences and created grave enmity between communities who otherwise spoke the same language. Besides, new chasms emerged to distinguish those who were fluent in a European language and those who were not. The rationalization of a public language brought in a new form of social categorization hitherto unknown in these parts of the world. In Tanzania language rationalization involved choice of an indigenous official language as a linguistic strand of the Ujamaa philosophy, a Tanzanian version of African socialism. Swahilization of Tanzania was aimed at forging a new nation with a national language to symbolize unity, strength, and independence. This project also failed because the elites who were pushing for a “Kiswahili-only” nation ignored the grassroots political and social dynamics (Blommaert ). For example, as the government was pushing for Swahilimedium primary education, Tanzanians were sending their children to Kenya to receive English-medium education. The ideology aspect of the language policy was in conflict with the practice aspect (Spolsky ; Shohamy ). After the crumbling of the Soviet Empire in the late eighties and especially after the retirement of President Julius Nyerere, the new Tanzanian leadership started de-emphasizing the “Kiswahili-only” policy by allowing the use of English medium in private primary schools; and it did this without even repealing the law. The unity promised by the one-language policy turned out to be an ideological mirage; Tanzanian parents suspected their government of trying to deny them progress, modernity, and power, all of which they thought were encoded only in the English language. It was an attempt to introduce a universalist culture where language is decontextualized and Tanzanian people asked to embrace the new language and new political culture, what others call “apocalyptic discourse” (Ridge ). But as Sacks (: ) warns, “Universalism turns the truth of a single culture into the measure of humanity. The results are often tragic and always an affront to human dignity.” Swahili may be indigenous but the culture of its speakers is unique to them. The political elite culture encoded in Swahili was imposed on the rest of theTanzanians and they simply subverted the project (Blommaert ). They saw it as a challenge to their own sense of cultural autonomy and also as a denial of access to the language of power and progress. However, the Swahilization project was immensely successful in spreading the government’s socialist agenda and instilling a rare sense of civic nationalism that is uniquely Tanzanian. Perhaps things would have worked even better if there was no radical reduction of public space for English whereby it remained only as medium of education in secondary school and university, with Swahili taking all other public functions and spaces. The myth that standard European languages were suitable for education and industrialization has also been exposed. It is what is sometimes called “nationism” as opposed to nationalism (Fishman ). The results of research in Africa show that foreign languages are not suitable for the economic, social, educational, and academic development of the continent (Bamgboṣe ; Prah ; Brock-Utne et al. ). A project on language of instruction in Tanzania and South Africa (LOITASA) has provided empirical evidence from classroom practices in the two countries that more effective learning occurs in mother-tongue-medium classrooms (Brock-Utne and Desai ). Similar results have
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
been reported from Ethiopia, where students taught in Amharic have been found to do better than those taught in English in most school subjects (Heugh et al. ). Research from different parts of the world show that teaching and learning in L becomes beneficial after at least six years of L-medium instruction. In the majority of cases in Africa, L instruction begins in fourth grade, but in overenthusiastic sections of the population it begins from year one. With the overwhelming evidence that early transit does not benefit African children, in whose interest is the early transit? The associated ideology is that the more hours children are exposed to a foreign language, the better their chances of success in school, no matter the quality and quantity of the linguistic input they receive. This certainly works for middle-class children for whom a second language is almost a mother tongue; but for the rest it inhibits creativity, kills enthusiasm, and delays development of intellectual skills. The desire of the political classes at independence was to promote intercommunity communication to ensure ethnic harmony and nurture the nascent new sense of civic nationalism. Most countries opted to retain a colonial language in the hope that they would avert violent ethnic conflicts. However, the tragic stories from Somalia, Rwanda, and Burundi have proved this ideology awfully wrong. Although Nigeria, with the highest number of indigenous languages, has had its share of interethnic conflicts, it has probably been faster to bring the different groups to agree to coexist than it has been to bring the almost unilingual Somalis to work together as a nation. Perhaps we need to look elsewhere for the true source of conflicts; language is certainly being used as a scapegoat for political and economic injustices that cause irritation and disaffection in some groups or regions. Some linguistic behaviors are a reflection of or reaction to these injustices. Language per se has not been proven to be the cause of strife, unless we are dealing with a case of metaphorical speech here. The fact is that people fight over resources and opportunities and may express their frustrations and unhappiness through different linguistic forms available to them. The medium must not be confused with the substance. Another challenge facing the political leadership is ensuring equal opportunity for all so that social mobility is not determined by accident of birth, color, tribe, class, or language. Providing equal access to only one language or a set of languages to be used in public institutions and for all important public functions is to enact an egalitarian public policy, or so the logic goes. Recently, Kenya was cited as one of the most unequal societies on earth, although for the last fifty-two years of independence we have had English as the official language and language of instruction for most of the school years. For example, in an annual survey of educational performance in Tanzania, Kenya, and Uganda, the results show that, in all three countries, there is glaring unequal access to quality education between regions and between social groups. In the three countries, education is provided in one official medium, which is not the mother tongue of all. According to the report, “The figure shows a large difference between the poor and non-poor . . . This data demonstrates that commitment to education for all aside, the reality is that opportunity to develop skills is highly unequal across East Africa” (Uwezo : ). This is true of the other African countries too, including those that opted for indigenous languages like Algeria, Sudan, and Egypt. The last claim I will discuss here is that imported European languages promote political participation or “democratic deliberation, in which free and equal citizens exchange reasons and are sometimes moved by them to change their opinions and preferences”
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
(Patten : ). The linguistic ideology underlying the claim is that diversity works against harmonious political discourse and that the tendency is for groups to engage in political activity through a home language to exclude the out-group members. But this is an oversimplification of the linguistic behavior of multilingual communities in Africa. For example, in Kenya the presidential speech on national days is always bilingual, not by law but by tradition. The formal written version is in English but after reading it, the “proper” heart-to-heart address occurs in Swahili, an indication that Kenyans expect their president to be bilingual in English and Swahili. Besides, citizens of multilingual societies normally regard code-switching as their mother tongue. For example, after the presidential address, Kenyans would engage in animated discussion of the speech in Swahili, English, and the numerous mother tongues spoken in the country. Language rationalization cannot be defended simply on the fear that people would fail to develop vibrant political conversation across linguistic boundaries. In any case, even in “monolingual” countries there are discourse communities defined by registers, genres, and dialects. The kind of homogeneity envisaged by this ideology is untenable in the present age of varied and complex linguistic competencies. Lastly, related to the myth of political participation is one political identity in the mold of the nineteenth-century “monoglot ideology” of one-country-one-language. Tanzania came closest to making Swahili the symbol of the people and the nation. But as examples from eastern and central Europe attest, homogenization efforts run counter to the natural tendencies of humans to celebrate their cultural peculiarities expressed in their languages. In fact, the more aggressively the state attempts to construct a single national identity, the more militant minorities become in the quest for public recognition. For example, in Kenya, little-known linguistic groups like the Ogiek, Abasuba, Terik, etc. have emerged recently to demand recognition of their languages in the same way that Croatian, Serbian, and Albanian have emerged as national languages in their respective countries. In Tanzania, there is a massive project to document the more than a hundred local languages, after nearly five decades of a “Kiswahili-only” national push. Minorities are continually seeking a voice, and any attempt at homogenization is vigorously contested.
. C
.................................................................................................................................. Distribution of public space is obviously a potent source of conflict among language groups in a country, and a public policy is required to deal with it. However, as long as monolingualism is assumed to be the only natural order of things and language is treated as a piece of technology for naming the world and thereby drawing communal boundaries, we shall continue to have unresolved language disputes, which might have tragic endings. Perhaps the wise counsel from Fishman (: ) may help change things: the entire world needs a diversity of ethno-linguistic entities for its own salvation, for its greater creativity, for the more certain solution of human problems, for the constant re-humanization of humanity in the face of materialism, for fostering greater aesthetic, intellectual and emotional capacities for humanity as a whole; indeed for arriving at a higher stage of human functioning.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Bamgboṣe, A. (). ‘Lessons from the Yoruba experience’, in B. Brock-Utne and R. K. Hopson (eds.), Language of Instruction for African Emancipation. Cape Town: CASAS, –. Blommaert, J. (). ‘Language policy and national identity’, in T. Ricento (ed.), –. Blommaert, J. (). State Ideology and Language in Tanzania. Tilburg: Tilburg University. Brock-Utne, B. (). ‘Research and policy on the language of instruction in Africa’, Internal Journal of Educational Development : –. Brock-Utne, B., and Desai, Z. (). ‘Expressing oneself through writing: a comparative study of learners’ writing skills in Tanzania and South Africa’, in Brock-Utne et al. (eds.), –. Brock-Utne, B., Desai, Z., Qorro, M., and Pitman, A. (eds.) (). Language of Instruction in Tanzania and South Africa. Highlights from a project. Rotterdam: Sense Publishers. Fairclough, N. (). Language and Power. London: Longman. Fishman, J.A. (). ‘Nationality-nationalism and nation-nationism’, in J.A. Fishman, C. A. Ferguson, and J. Das Gupta (eds.), Language Problems of Developing Nations. New York: Wiley, –. Fishman, J.A. (). ‘Whorfianism of third kind: ethno-linguistic diversity as a worldwide societal asset’, in J. A. Fishman, M. H. Gertner, E. G. Lowy, and W. G. Milán (eds.), The Rise and Fall of Ethnic Revival: Perspectives on language and ethnicity. Berlin and New York: Mouton de Gruyter, –. Heugh, K., Benson, C., Bogale, B., and Johaness, M. A. G. (eds.) (). Final Report: Study on Medium of Instruction in Primary School in Ethiopia. Addis Ababa: Ministry of Education. May, S. (). ‘Rearticulating the case for minority language rights’, Current Issues in Language Planning : –. Patten, A. (). ‘Political theory and language policy’, Political Theory : –. Prah, K. K. (). Mother-tongue education in Africa for emancipation and development: Towards the intellectualisation of African languages. In B. Brock-Utne and I. Skattum (eds.), Languages and Education in Africa: A comparative and transdisciplinary Analysis Cambridge: Cambridge University Press, –. Ricento, T. (ed.) (). Introduction to Language Policy: Research and method. Oxford: Blackwell. Ridge, S. G. (). ‘Language policy and democratic practice’, in S. B. Makoni and N. Kamwangamalu (eds.), Language and Institutions in Africa. Cape Town: CASAS, –. Sacks, J. (). The Dignity of Difference: How to avoid the clash of civilizations. London: Continuum. Shohamy, E. (). Language Policy: Hidden agenda and new approaches. London and New York: Routledge. Skutnabb-Kangas, T. (). ‘Language policy and human rights’, in T. Ricento (ed.), –. Spolsky, B. (). Language Policy. Cambridge: Cambridge University Press. Tollefson, J. W. (). Planning Language, Planning Inequality. London: Longman. Uwezo. (). Literacy and Numeracy across East Africa. Nairobi: Uwezo.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T of knowledge takes place mainly through language. The language question is thus of the utmost importance in education. Education is, in turn, one of the most important instruments for strengthening the status and use of a language, be it as means or matter of instruction. In Africa, where modern education was introduced by European colonial powers, the language of instruction (LOI) was that of the colonizers. At independence (which in most cases took place in the early s), the former colonies kept the European languages (English, French, Portuguese, and Spanish) as official language and LOI, alone or together with African languages. This situation still prevails in sub-Saharan Africa and the Indian Ocean.1 It would seem obvious that a child should learn to read and write in a language she/he understands and not have to grapple with initial literacy and a new language at the same time. Yet this is what happens in many African countries, especially in states with official Romance languages (French, Portuguese, and Spanish). Maintaining a European, exogenous (non-indigenous) language as official language and main LOI has had far-reaching consequences. A majority of Africans do not master their country’s official language because it is not spoken around them but acquired mainly at school, which many have never or only briefly attended. This excludes people from democratic processes and power positions, as well as from advances in health, science, Among the fifty-four states considered as African (fifty on the continent plus four in the Indian Ocean), one finds the following distribution of official languages, alone or together with other languages: English (twenty-two states), French (twenty-two), Arabic (twelve), African languages (twenty-eight), Portuguese (six), contact languages (Creoles) (two), Spanish (one), and Afrikaans (one). Only Ethiopia (which was never colonized) has an African language as sole official language. The six Arab-dominated countries north of the Sahara fall outside the scope of this chapter, which focuses on sub-Saharan Africa and the Indian Ocean (forty-eight countries). The term “African” here refers to these countries. 1
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and technology. Language-in-education policy is thus central to development issues, at the individual as well as the societal level. After a brief survey of literature on language and education (section .), we shall see that different sub-Saharan countries deal with these challenges in various ways. Empirical evidence from some of the forty-eight African countries will be examined in the light of two major questions: African or European LOIs? (section .); and Which African LOIs? (section .) before I conclude in section ..
. L
.................................................................................................................................. Focus will here be on formal education, though informal education as well as sign language (see Nyst, chapter of this volume) are related and relevant issues. The geographical and academic scope is, however, such that even this limitation forbids any claim to exhaustiveness. Language and education are a transdisciplinary field of research. Of particular interest in educational sciences are didactics, teacher training, classroom observation, competence tests, schoolbook analysis, and examinations results. Within linguistics, relevant subdisciplines are sociolinguistics (language planning and policy, bilingualism, language attitudes) and applied linguistics (language acquisition, language learning and teaching, cognitive development). Development studies also take up educational questions but, curiously enough, do not always consider the LOI. Bilingualism (often used for bi- and multilingualism) and bilingual education were first studied in Western countries. Important theoretical insights are presented in general works like Romaine (), Baker and Prys Jones (), Baker (), and García (). These theories are not always applicable to Africa, mainly because of the unequal status of European and African languages (Skattum , Brock-Utne ; see, however, Heugh for the opposite view). Cummins’s theories on the relationship between L and L instruction (presented in Baker : –) are, however, frequently referred to in works on African education. Cummins maintains that a satisfactory level of cognitive competence in L should be attained before moving to L instruction, and also that concepts and vocabulary acquired in L facilitate their acquisition in L. This leads to a strong correlation between L and L in reading and writing skills as well as comprehension (Cummins : ). Though some criticism and contradictory evidence is also put forward, e.g., by Bamgboṣe () and Romaine (), his “linguistic interdependence” hypothesis () is widely supported: Experiences in Africa and many other parts of the world show that cognitive development is achieved faster if the mother tongue, rather than an LWC (Language of Wider Communication), is used as the language of instruction in primary education. (Hovens : )
Bilingual education may refer not only to programs where two languages are used as media of instruction, but also to monolingual instruction in a language different from the children’s L (Baker : ). I will use it here in the first sense. Typologies of such education are more or less detailed. Hornberger’s () typology is the easiest to adapt to
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
an African context. She distinguishes between education models and program types. The former are more abstract and defined according to their goals, which may be linguistic, cultural, and/or social, while the programs have contextual (school-external) and structural (school-internal) characteristics. Models are of three types: () The transitional type aims to shift the child from its first language (L) to the second language (L) and to assimilate it to the majority culture. L instruction compensates for—temporarily—the lack of L competency. () The maintenance type aims at cultural pluralism through the protection of L and the strengthening of ethnic identity and minority groups’ civil rights. () The enrichment type not only aims at maintaining cultural pluralism but works towards the development and expansion of minority languages and the ethnic groups’ autonomy and integration within the nation.2
. A E LOI?
..................................................................................................................................
.. Terminology The terms “L” or “L instruction” are defined not according to the order in which the languages are learnt or the speaker’s degree of competence (as in sociolinguistics), but according to their mode of acquisition: L is a language acquired naturally in the child’s environment, while L is learnt at school. L in this sense does not refer to a foreign language taught as a subject, like English in Senegal, but to a LOI which, without being the students’ native tongue, has a special status in the polity, as official language and/or LOI, like French in Senegal.3 “Home language” distinguishes the child’s language from that used at school, while taking into account that many children are bi- or multilingual, either because they grow up in multilingual families or because they are minority language speakers and acquire the regional lingua franca. In English, “mother tongue” education (MTE) is frequently used in a broad sense “to refer to the use of an indigenous language in education for any purpose and at any level” (Bamgboṣe : ). In Francophone4 Africa, one often speaks of “national language instruction”. Both terms gloss over the fact that African LOIs hardly exist except in bilingual programs. “National language” (NL) carries various meanings in different countries. It may refer to selected languages with certain public functions, particularly in education and the media
The dichotomy “minority/majority” languages refers to immigrants’ languages as opposed to European languages in Western countries (e.g. Urdu/English in Britain). In Africa, “dominated/dominating” is a more appropriate opposition, since the European languages are imported, minority languages while African languages can be majority or minority languages demographically speaking and also more or less dominating according to other criteria (historical, social, economic, etc.). 3 Cf. the French distinction between FLE (Français Langue Étrangère) and FLS (Français Langue Seconde). 4 The terms “Anglo”-, “Franco”-, “Lusophone” (Portuguese-speaking), and “Romance-speaking” are inaccurate, as only a minority of the populations speak these official languages, but they reflect the differences in administrative and educational systems left by the former colonial powers. 2
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
(four in DRC, thirteen in Mali, etc.), or to a mainly symbolic status, as in Senegal, where all codified languages—twenty-one in —are NLs (in , six of them were admitted in Parliament, with interpreters financed by the European Union). It may also apply to all vernaculars without distinction (e.g. in Burkina Faso, Ethiopia5), or be used in daily speech without any legal implication.
.. African experiences The African experiences under discussion are each special in its own way and yet representative of different tendencies. Only those which give both L and L instruction will be considered here. Tanzania is one of nine states that combine a European official language with an African language. President Nyerere at independence succeeded in promoting Swahili as the nation’s unifying symbol, introducing it as LOI in primary school already in . Today Swahili has “a much wider use than African languages usually do south of the Sahara” (Brock-Utne and Skattum : ). It is not only the country’s official language alongside English, but also “the language of Parliament, of the lower courts, and of the government offices, besides being the language most frequently used in the written press, the television and the radio” (p. ), and it is taught in Teacher Colleges (which is not necessarily the case in other countries using African LOIs). Swahili is spoken by more than % of the population as L or as L between speakers of the approximately local languages, while only % of the citizens speak English (Mwinsheikhe : ).6 Despite this wide acceptance of both spoken and written Swahili, it was not until that the authorities decided to extend its use to secondary and tertiary education—a decision which has not yet () been implemented. Mwinsheikhe studied the consequences of the switch from Swahili to English after primary school (grades –) through an experiment carried out in science classes in secondary schools. Some teachers were asked to use Swahili as LOI and others to teach in English only, which meant diverting them from their usual code-switching practice (explaining in Swahili to ensure the pupils understood what was said in English). Observation and test results showed that Swahili instruction tended to ameliorate performances, for English it was the opposite, and code-switching gave in-between results. Mwinsheikhe concludes that English hampers learning, while the use of a familiar language enhances achievement. She underlines that teachers should have an acceptable level of English competence whether they teach it as a subject (English Language Teaching) or use it as LOI; otherwise, students will suffer as regards both the subject matter and the correct use of English. Ethiopia is the only country with an African language, Amharic, as sole official language. MTE was here introduced even earlier (/), and applied throughout lower primary school (grades –). To begin with, Amharic was the only LOI, but unlike in Tanzania, this dominance was contested. In all “nationality languages” became candidates for MTE, In Ethiopia the term is “nationality language”. As usual in Africa, language statistics are approximate: according to Rubagumya (: ), % of the population speak Swahili (% as L), while % have some knowledge of English. 5 6
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and twenty-two of the approximately eighty-five vernaculars were actually introduced (Lanza and Woldemariam : , ). Today, thirty vernaculars serve as LOIs—a record in sub-Saharan Africa.7 Despite the prominent role given to NLs, there is a shift from MTE to English in secondary education. The reform, however, recommended extending MTE to upper primary education (grades –). Of the then eleven (Mekonnen ) regional states, four answered the call by adopting a NL as LOI, while seven regions continued to teach in English. To see how the different LOIs affected students’ achievement, Mekonnen undertook a statistic analysis of grade national assessments in mathematics, science, and English for the years and . The sample included ten of the eleven regional states and all the LOIs. Considering and as a whole, Mekonnen found that MTE facilitated students’ learning capability in all the subjects studied, including English. The common belief that using English as LOI will help students speak, write, and read English was thus shown to be erroneous. Mekonnen concludes that in an African setting, where local languages are used in daily communication, English should be taught as the foreign language it is. Mekonnen’s quantitative study confirms the results of Mwinsheikhe’s qualitative experiment, both exceptionally conducted after the (lower) primary level. Though a “late exit” from MTE has proven more effective than an “early exit”8 (Baker and Prys Jones : ), and despite calls for cultural autonomy, the tendency after independence has paradoxically been to diminish African language instruction, either by going from a late to an early exit or by cutting out MTE altogether (Bamgboṣe : ). South Africa illustrates this trend, though the country holds a continental record with eleven official languages (English, Afrikaans, and nine of the approximately twenty-five vernaculars) (Baldauf and Kaplan b: ). The diminution of MTE is linked to the apartheid past, when Black people were confined to Bantustans conceived along ethnolinguistic lines. These homelands had their own official and instructional languages in addition to English and Afrikaans. The Black, however, claimed equal access to English, a gateway to social success. So when the Bantu Education Act extended existing African language education (four to six years) to all eight years of primary school, it aroused resentment and eventually led to the Soweto rebellion of students in . Amendments were made in , limiting MTE to the first four years of school (Mahlalela-Thusi and Heugh : ). The elections put an end to apartheid legislation, and in “all languages that were officially recognized in the old South Africa (in one way or another) were also recognized by the new Constitution” (Du Plessis : ). Du Plessis, however, characterizes South Africa as a “non-active endoglossic nation” (p. ), one that pretends to but does not actually develop local languages. This concurs with Mahlalela-Thusi and Heugh’s () analysis of mother tongue and bilingual education today, which “have been both systematically implemented and then removed from the education system of the country” (p. ). Schoolbooks in African languages for eight years of primary school had helped A new record is reported: fifty-one NLs are used in MTE (Sajita Bashir, World Bank manager, Education, Eastern Africa, NORAD conference at Oslo, May rd, ). 8 Baker (: ) defines “early exit” as maximum two years of L instruction, but the African practice of – years L instruction is also early, considering the weak L input outside class compared to that of immigrant children in the West. 7
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
develop terminology, but books and terminology have now gone out of use, and Xhosa English biliteracy programs, for instance, lack materials in Xhosa. The reimplementation of MTE is commonly countered with the arguments that costs would be too high, that there is insufficient terminology in African languages, and that for these reasons, publishers will not / cannot produce textbooks in African languages. Such arguments have been refuted by many proponents of MTE, such as Heugh () and an international team of researchers commissioned by the Association for the Development of Education in Africa and other international agencies (Alidou : –). They found that L instruction is more economic in the long run because of high repeat and dropout rates in L instruction. They also consider many African languages well instrumented today and observe that using the languages will encourage further instrumentation. The francophone countries are divided into three groups by Babault (: –): those which give L instruction only, those which give or have given L instruction in experimental bilingual schools, and those which have introduced MTE without experimentation. In the second group, experiments are often supported by foreign donors and therefore unstable. In Burkina Faso, a short period of bilingual teaching in French/NL in the s was abruptly stopped. Burkinabe authorities later allowed a Swiss NGO to reintroduce bilingual schools. In , forty schools gave instruction in seven of the approximately sixty NLs. Though this alternative education lasts five instead of the normal six years, from – it gave pupils a mean score of % success on final exams against % for the monolingual French schools (Ilboudo : ). Burkina Faso has also introduced “satellite schools” in remote villages when the “mother school” is too far away. Here three years of bilingual education is given before pupils join French schools. In , there were satellite schools (Burkina Faso ). Results are, however, unsatisfactory (Alidou : ). For though both types have final exams in French, which triggers more teaching time in this language, satellite schools switch abruptly to French in the fourth grade, while bilingual schools practice a gradual transfer throughout primary school (–%, with a /% distribution in the pivotal third grade; Ilboudo : ). Mali is a case apart in this group, having given bilingual education without interruption (though with ups and downs) since as part of public policy (Skattum ) and having extended it to , primary schools (.% of public schools) in eleven NLs in / (Skattum ). The “first generation” gave two to four years of instruction before switching to French. Though lack of follow-up led to dropping results, it gave better results than monolingual French instruction, as shown, e.g., by a language test in asking fifth graders who had their first year of French-only instruction to write a story in French from a picture. The same task was given to fifth graders in a neighboring “classical” school who had received French-only instruction from the first grade. The latter performed less well in every aspect: orthography, syntax, and vocabulary, as well as text length and textual cohesion (Skattum ). The “second generation”, called Convergent Pedagogy (CP), adopted a gradual transition throughout the first cycle of primary school (grades –). CP was meant to be implemented throughout the country, but here also results dropped with less follow-up. By , CP had lost ground, with , schools representing .% of public schools. Besides, Maurer () criticizes CP for lack of distinction between L and L didactics, while Skattum () points to the mismatch between the government’s declared goals,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
which imply a maintenance model (cultural autonomy, equal linguistic competence in NL and French, individual and social development) and the means used to reach these goals. These means indicate a transitional model (more teaching hours and more textbooks in French, teacher training focused nearly exclusively on French). To rectify the weaknesses, in the Malian authorities stopped further expansion. Today (), they have abandoned it completely. Only NGOs give some Bambara instruction in grades –. In the third group, two countries, Madagascar and Guinea, have changed abruptly from French to NL and back to French, without experimentation, for ideological reasons. In Madagascar, growing discontent with continued French cultural dominance after independence led to riots in . President Ratsiraka, who took power in , introduced Malagasy as sole LOI in primary and secondary schools. Declining school results, however, led to negative attitudes towards this “malgachization”—attitudes which still prevail—so French was reintroduced in secondary schools in , and in primary schools from grade in (Babault : –). Today, Malagasy is officially used in grades and but in reality longer, tests in revealing that “only .% of primary education teachers were proficient enough to teach in French” (Rabenoro : ). The Lusophone countries all teach nearly exclusively in Portuguese. In Mozambique, “[t]he psychological and pedagogical arguments advanced for an hypothetical scenario of mother-tongue (i.e. Bantu) instruction . . . are not clearly understood by parents, even by middle-class adults” (Lopes : ). A five year bilingual pilot project (–) for two primary schools with two different Bantu languages, in a gradual transition model, was favorably assessed but nevertheless stopped. Today Mozambique is changing this policy. In , the Ministry of Education announced a plan to use sixteen Mozambican languages alongside Portuguese in primary schools across the nation (UNICEF ). This plan was implemented in (Mozambique, Ministry of Education ). Guinea Bissau conducted a ten-year experiment (–) with Kiriol (Creole) as LOI for the first two years. Kiriol is spoken by % of the population and functions as a lingua franca among the approximately twenty NLs. This experiment, which included some thirty experimental classes with approximately , pupils, was also ended, though results were better than for monolingual Portuguese instruction. Despite the early exit and abrupt transition to L, Hovens (: ) found that pupils transferred their acquired reading skills from L to L, an aptitude he attributes to Cummins’s interdependence theory.
. W A LOI?
.................................................................................................................................. Once a country decides to give mother tongue or bilingual education, the question arises which African language(s) to choose. The risk of ethnic rivalry is recognized even by Africans who believe in strengthening their role: Deciding which indigenous language to promote as the national language within [national] boundaries carries the danger of ethnic rivalry. . . . Any move to make Hausa the national language of Nigeria could precipitate a national crisis in Yorubaland and Igboland. (Mazrui and Mazrui : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Some multilingual states nevertheless promote one NL to the detriment of others, with varying results. In Tanzania, Swahili is widely accepted as a non-tribal lingua franca. But even where the domination of one language is accepted, the difference in exposure is bound to have negative effects for many pupils: For example, in Tanzania, where Swahili is used as a medium of instruction in primary schools, there are four distinct categories of children for whom the practice has different implications: a minority of native speakers, a large majority of second language speakers, speakers of Bantu languages like Swahili, and speakers of non-Bantu languages. (Bamgboṣe : )
In Malawi, on the other hand, the imposition of Chewa as the sole “national language” (Kayambazinthu : ) since , with Chewa as LOI in grades – (p. ), has met with resistance. This is due to the choice of the first president’s own Chewa dialect as a standard, which has created a hostile climate hampering Chewa’s development into a neutral lingua franca (p. ). Multiparty elections in put promotion of less-used languages on the agenda. But, as Kayambazinthu observes, a pluralistic alternative also brings pluralistic dilemmas, “politically, economically and educationally. Where should the line be drawn? If equity is the criterion, then all languages should be treated equally, an undertaking that the government cannot afford” (: ). Faced with these problems, Malawi’s language-in-education policy is unclear. In Botswana, policy in favor of one language, Tswana, is clearer: Tswana instruction is given for the first four years (Vossen : ), while the minority languages which were taught before independence were “banned from use in school at independence” (NyatiRamahobo : ). They are now threatened with extinction because of a general intolerance (pp. –). Both Malawi and Botswana pupils, of course, face the same unequal exposure to the dominating endogenous language as Tanzanian children. In Mali, on the other hand, multilingualism is embedded in the constitution () and people have taken pride in their multicultural tradition dating from the great medieval empires (Skattum ). The jihadist trouble which started in northern Mali in , and recently spread to central Mali, has threatened this peaceful ethnic cohabitation. Linguistically, however, these positive attitudes still prevail despite the fact that Mali has a vigorous lingua franca, Bambara, spoken by % of the population as L or L (RGPH : ), which could have taken on the role of a unifying symbol. De facto effects of Bambara domination are however not unlike those imposed de jure in other countries, entailing language shift from some minority languages and unequal exposure to the national lingua franca for pupils of other NLs. Multicultural policy does not necessarily stem from tolerance of other ethnic groups but may be the result of ideological battles. Ethiopia is a striking example. Amharic has been the country’s official language since , and is also “arguably the most advanced African language in Africa” (Lanza and Woldemariam : ), written for centuries in a script unique to the country (called Ethiopic or Fidel), and possessing a rich literary tradition. Amharic, however, comes second in demographic terms (.% of L speakers ( census)), just after Oromo (.%) but before Somali (.%) and Tigrinya (.%), the language of the – leadership. The Tigrayan People’s Liberation Front installed a new political
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
structure of ethnic federalism which was open to MTE in any vernacular. However, as the “Ethioling” project () points out, “except for Amharic, the rest have mainly been spoken languages, not written. Therefore, there are problems of implementation related to attitudes, standardization, curriculum, teaching material, and manpower.” Divisiveness is not restricted to multilingual contexts but may also arise in linguistically homogeneous polities. Madagascar is one of a handful (near) monolingual African countries, and yet Malagasy people commonly consider their country triglossic: “[T]he official variety of Malagasy dominates the other varieties, while Standard French . . . dominates Malagasy” (Rabenoro : ). Though Malagasy dialects are mutually comprehensible, the official variety, based on the Merina dialect and associated with the Merina kingdom, is contested, confirming Babault’s (: ) remark that “language representations accurately reflect the social or political divisions that characterize the country” (my translation). Rabenoro proposes introducing the main non-official varieties as media of teaching and training teachers accordingly, but does not specify which or how many varieties nor discuss how teachers could be posted according to their dialects. As Fishman (: ) observes: [D]ifferences do not need to be divisive. Divisiveness is an ideologized position and it can magnify minor differences . . . . Similarly, unification is also an ideologized position and it can minimize seemingly major differences or ignore them entirely [author’s emphasis].
. C
.................................................................................................................................. Two main issues have been raised in this chapter: should education be given in European or African languages, and in the latter case, which endogenous language(s) should be chosen? For answers, I have looked to African experiences from Anglophone, Francophone, and Lusophone countries, plus the only “Afrophone” country, Ethiopia (which, however, is Anglophone as concerns secondary and higher education). Anglophone countries are the most favorable, Francophone countries more skeptical, and Lusophone countries the most hostile towards African language instruction in primary school. Practically none uses African languages as a means of instruction in secondary or higher education. These attitudes are reflected in the choice of educational models. Of forty-eight African countries, only six practice a late exit from MTE, which corresponds to the maintenance model, shown by international research to be the most effective for children’s cognitive development in a multilingual context. The trend after independence has rather been to move from late to early exit or to stop MTE altogether, so that fifteen countries today adopt a transitional model aimed at easing the acquisition of European languages and culture, with an early exit after two to four years. The enrichment model is hard to apply in Africa because of the asymmetrical relationship between oral, endogenous African languages and written, exogenous European languages. French and Portuguese assimilation policy has left the Franco- and Lusophone countries with little L instruction, given mostly in experimental bilingual schools supported by (foreign) donors and stopped when they withdraw. Exceptions are, on the one hand, Mali,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
with fairly extensive government supported bilingual education, and, on the other hand, Madagascar and Guinea, which have conducted abrupt reforms changing the LOI from L to L and back to L, with negative consequences particularly for teachers’ competencies. The answer to the first question (Should LOIs be African or European?) is that even a transitional model usually gives better results than L-only instruction, supporting Cummins’s interdependence theory. Classroom observation shows more dynamic interaction between teachers and pupils, and a better level of content learning. Statistical analysis of examination results and language tests has also given better results for bilingual than for monolingual L schools. However, L instruction in itself is not enough: contextual and structural conditions need to be satisfied. These are, inter alia, a gradual transition from L to L, teacher training in L didactics and in the language itself, L textbooks, linguistic research to ensure lexical development and an academically sound standard orthography, and information to parents and teachers to promote more positive attitudes towards L instruction. The second question (Which African LOIs should be chosen?) contains a “pluralistic dilemma”. Giving L instruction in all endogenous languages (approximately , on the continent; Heine and Nurse : ) is beyond the means of any country and impossible for practical reasons. For a written code to be established, a language needs to have a certain demographic weight, with a reasonable readership for literature, the written press, a written landscape, etc., and also it needs to develop dictionaries and grammars. Some countries select one major African language or dialect as LOI. This is economically efficient but culturally inequitable. Other countries give MTE in several African languages, chosen for demographic, political, economic, social, or historical reasons. Claims of authenticity are better satisfied, but unequal conditions for children of other NLs remain, if not in the same proportion. This still seems the best solution. Neither efficiency nor authenticity considerations should, however, hinder African language instruction. Not only do pedagogical and even economic reasons plead in favor of L instruction, but it is also a democratic issue of great importance to the development of African societies. As one of the Nestors in the field of language-in-education observes, official imported European languages immediately create two classes of citizens: the included and the excluded. The responsibility of the former in maintaining the status quo is heavy: The included are a major stumbling block in the use of African languages in a wider range of domains. Apart from lack of political will by those in authority, perhaps the most important factor impeding the increased use of African languages is lack of interest by the elite. (Bamgboṣe : )
R Alidou, H. (). ‘Promoting multilingual and multicultural education in Francophone Africa’, in B. Brock-Utne and I. Skattum (eds.), –. Babault, S. (). Langues, école et société à Madagascar: Normes scolaires, pratiques langagières, enjeux sociaux. Paris: L’Harmattan. Baker, C. (). Foundations of Bilingual Education and Bilingualism, th edn. Bristol, Buffalo, NY, and Toronto: Multilingual Matters.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Baker, C., and Prys Jones, S. (eds.) (). Encyclopedia of Bilingualism and Bilingual Education. Clevedon: Multilingual Matters. Baldauf, R. B. Jr., and Kaplan, R. B. (eds.) (a). Language Planning and Policy in Africa. Vol. : Botswana, Malawi, Mozambique and South Africa. Clevedon, Buffalo, NY, and Toronto: Multilingual Matters. Baldauf, R. B. Jr., and Kaplan, R. B. (b). ‘Language policy and planning in Botswana, Malawi, Mozambique and South Africa: some common issues’, in R. B. Baldauf Jr. and R. B. Kaplan (eds.), –. Bamgboṣe, A. (). Language and the Nation: The language question in sub-Saharan Africa. Edinburgh: Edinburgh University Press. Bamgboṣe, A. (). Language and Exclusion: The consequences of language policies in Africa. Hamburg: LIT. Brock-Utne, B. (). ‘The adoption of the Western paradigm of bilingual teaching: why does it not fit the African situation?’, in K. K. Prah and B. Brock-Utne (eds.), Multilingualism, an African Advantage: A paradigm shift in African languages of instruction policies. Cape Town: Centre for Advanced Studies in African Society, –. Brock-Utne, B., and Skattum, I. (eds.) (). Languages and Education in Africa: A comparative and transdisciplinary analysis. Oxford: Symposium Books. Burkina Faso (). Communication de l’équipe du Burkina FASO sur les écoles communautaires. Niamey, Forum International sur les Écoles Communautaires, – juillet , www.repta.net/ repta/telechargements/communication_niamey.pdf accessed November , . Cummins, J. (). ‘Linguistic interdependence and the educational development of bilingual children’, Review of Educational Research : –. Du Plessis, T. (). ‘South Africa: from two to eleven official languages’, in K. Deprez and T. Du Plessis (eds.), Multilingualism and Government: Belgium, Luxembourg, Switzerland, former Yugoslavia, South Africa. Pretoria: Van Schaik, –. ETHIOLING. Vernaculars in Ethiopian Schools, http://www.hf.uio.no/iln/english/research/projects/ ethioling/ accessed August , . Fishman, J. A. (). ‘Nationality-Nationalism and Nation-Nationism’, in J. A. Fishman, C. A. Ferguson, and J. Das Gupta (eds.), Language Problems of Developing Nations. New York: John Wiley, –. García, O. (). Bilingual Education in the st Century: A global perspective. With contributions by Hugo Baetens Beardsmore. Chichester: Wiley-Blackwell. Heine, B., and Nurse, D. (). Les langues africaines. Paris: Karthala. (Translated from the English edition: African Languages. An Introduction, Cambridge: Cambridge University Press, .) Heugh, K. (). The Case against Bilingual and Multilingual Education in South Africa. Cape Town: Project for the Study of Alternative Education in South Africa. Hornberger, N. H. (). ‘Extending enrichment bilingual education: revisiting typologies and redirecting policy’, in O. García (ed.), Bilingual Education: Focus-schrift in honor of Joshua A. Fishman. Amsterdam and Philadelphia, PA: John Benjamins, –. Hovens, M. (). ‘Bilingual education in West Africa: does it work?, International Journal of Bilingual Education and Bilingualism (): –. Ilboudo, P. T. (). L’éducation bilingue au Burkina Faso: Une formule alternative pour une éducation de base de qualité. Tunis: Association for the Development of Education in Africa. Kayambazinthu, E. (). ‘The language planning situation in Malawi’, in R.B. Baldauf Jr. and R. B. Kaplan (eds.), –. Lanza, E., and Woldemariam, H. (). ‘Language ideology and linguistic landscape: language policy and globalization in a regional capital of Ethiopia’, in E. Shohamy and D. Gorter (eds.), Linguistic Landscape: Exploring the scenery. New York: Routledge, –. Lopes, A. J. (). ‘The language situation in Mozambique’, in R. B. Baldauf Jr. and R. B. Kaplan (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Mahlalela-Thusi, B., and Heugh, K. (). ‘Terminology and school books in southern African languages’, in B. Brock-Utne, Z. Desai, M. A. S. Quorro, and A. Pitman (eds.), Language of Instruction in Tanzania and South Africa—Highlights from a project. Rotterdam, Boston, MA, and Tapei: Sense Publishers, –. Maurer, B. (). De la pédagogie convergente à la didactique intégrée: Langues africaines—langue française. Paris: L’Harmattan. Mazrui, A. A., and Mazrui, A. M. (). The Power of Babel: Language and governance in the African experience. Oxford: James Currey. Mekonnen, A. G. Y. (). ‘Implications of the use of mother tongues versus English as languages of instruction for academic achievement in Ethiopia’, in B. Brock-Utne and I. Skattum (eds.), –. Mozambique, Ministry of Education (). Situação actual do ensino bilingue em Moçambique. Maputo: Ministry of Education. Mwinsheikhe, H. M. (). ‘Spare no means: battling with the English/Kiswahili dilemma in Tanzania secondary school classrooms’, in B. Brock-Utne and I. Skattum (eds.), –. Nyati-Ramahobo, L. (). ‘The language situation in Botswana’, in R. B. Baldauf Jr. and R. B. Kaplan (eds.), –. Rabenoro, I. (). ‘National language teaching as a tool for Malagasy learners’ integration into globalisation’, in B. Brock-Utne and I. Skattum (eds.), –. RGPH (ème recensement général de la population et de l’habitat du Mali) (). ‘Analyse des résultats définitifs. Thème : État et structure de la population’, www.instat-mali.org/contenu/ rgph/rastr_rgph.pdf accessed Aug. , . Romaine, S. (). Bilingualism, nd edn. Cambridge, MA: Blackwell. Rubagumya, C. M. (). ‘Language in Tanzania’, in C. M. Rubagumya (ed.), Language in Education in Africa: A Tanzanian perspective. Clevedon: Multilingual Matters, –. Skattum, I. (). ‘L’éducation bilingue dans un contexte d’oralité et d’exoglossie: théories et réalités du terrain au Mali’, Nordic Journal of African Studies (): –, https://www.duo.uio.no/handle/ / accessed July , . Skattum, I. (ed.) (). ‘L’école et les langues nationales au Mali’, Nordic Journal of African Studies () (special issue), www.njas.helsinki.fi/ accessed July , . Skattum, I. (). ‘L’apprentissage du français dans un pays “francophone”: le cas du Mali’, in J. Jansen (ed.), Mandé—Manding: Background reading for ethnographic research in the region south of Bamako. Leiden: University of Leiden, Department of Cultural Anthropology and Development Sociology, –, http://www.hf.uio.no/ikos/personer/vit/ingse/L_apprentissage_du_francais_au_ Mali.pdf accessed July , . Skattum, I. (). ‘Mali: in defence of cultural and linguistic pluralism’, in A. Simpson (ed.), Language and National Identity in Africa. Oxford: Oxford University Press, –. Skattum, I. (). ‘L’introduction des langues nationales dans le système éducatif au Mali: objectifs et conséquences’, Journal of Language Contact : –, https://brill.com/view/journals/jlc/// article-p_.xml?lang=en accessed July , . Vossen, R. (). ‘Sprache und Schule in Afrika—das Beispiel Botswana’, in D. Ibriszimow, R. Leger, and U. Seibert (eds.), Von Ägypten zum Tschadsee: Festschrift für Herrmann Jungraithmayr zum . Geburtstag. Würzburg: Ergon, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
, , ......................................................................................................................
In the last years about half the known languages of the world have disappeared. (Sasse : ).
. I
.................................................................................................................................. T world’s languages are currently numbered between , (Wurm ) and , (Romaine ). Sasse’s quotation above, which has been stated in different forms by different authors (see, e.g., Romaine ), shows that languages have been disappearing for a long time now. In the last two and a half decades, the threat to languages has aroused the interest not only of linguists but also of journalists, non-governmental organizations, and state agencies. Himmelmann () and, following him, Ameka () point out that the current interest was triggered by a talk given by Bechert at the International Congress of Linguists in Berlin in . In that talk Bechert, following Mühlhäusler (), wondered why linguists did not appear to be concerned that a major part of their empirical basis was about to disappear (cf. Bechert ). Bechert’s talk led to the passing of a motion by the Comité International Permanent des Linguistes to bring the issue of language endangerment to the general public. Among the activities that came out of the motion was a symposium on endangered languages organized at the Linguistic Society of America Conference in by Ken Hale, where Kraus gave a presentation which contains the following quotation: Obviously we must do some serious rethinking of our priorities, lest linguistics go down in history as the only science that presided obliviously over the disappearance of % of the very field to which it is dedicated. (Krauss : )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
This quotation is often cited as the clarion call that got linguists into action. It also gave birth to a sub-discipline of linguistics, i.e. documentary linguistics. In Africa, just like the other parts of the world, many languages are said to be endangered. However, while Africa is said to have about a third of the world’s languages, there is the perception that it has not been given its due place in the accounts of endangerment. Lüpke (: ) puts it most succinctly thus: “it seems fair to state that Africa is underrepresented in the description and documentation of endangered languages and in language endangerment discourse.” In what follows, I discuss issues relating to language endangerment (section .), documentation (section .), revitalization (section .), with a subsequent conclusion (section .).
. T A
.................................................................................................................................. Batibo () observes: “Most of the more than , minority languages in Africa are endangered, to varying degrees.” UNESCO distinguishes between safe, vulnerable, definitely endangered, severely endangered, and critically endangered. Depending on the perspective they adopt, Africanists come up with different classifications. For instance, Brenzinger et al. () use the following in their discussion of language death in Africa: (i) extinct languages, (ii) languages which are either extinct or in the process of extinction, (iii) languages which are in the process of extinction or threatened with extinction, and (iv) languages threatened with extinction. In contrast, Blench (: ) distinguishes the following, which he refers to as the “provisional status” of the languages in West Africa: (i) fully vital, (ii) definitely threatened, (iii) moribund, (iv) probably extinct, and (v) declining. A common thread that runs through the endangerment discourse in Africa, however, is the lack of detailed verifiable data to properly determine the status of current languages as well as those which are already extinct. Brenzinger et al. () note that due to the fact that writing started much later in Africa, the time span for investigating extinct languages does not extend beyond the middle of the nineteenth century. They, however, mention Ge’ez, an Ethio-Semitic language which used to be spoken in Ethiopia, and Coptic, which used to be spoken in Egypt. For Central Africa, Connell () reports that the number of endangered languages appear to constitute only a small portion of the total number of languages in the area, but adds that “this may be no more than a reflection of our ignorance of the situation.” For West Africa, Blench () writes that there is no information on languages, and adds that the majority of them are small languages and, therefore, very likely to be threatened. For eastern Africa, Dimmendaal (: ) writes in a footnote: “[we] know very little about the actual processes of structural decay in the dying languages apart from a few case studies.”
.. How languages become endangered Wurm () discusses two ways in which languages die or become endangered:1 the first occurs when speakers of the language die as a result of war, genocide, or disease. In such 1 There are numerous definitions of language death or endangerment, which space does not allow me to cover.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
, , .
cases the language is left with no speakers and, therefore, disappears. For the African continent, Brenzinger and Dimmendaal () note that the only (rare) instances of languages disappearing through genocide occur in the Khoisan language family. The second type of language death, which is the most common, occurs when speakers of a language give up the language for another. Wurm refers to this as ecology and distinguishes two subtypes: one occurs when the: cultural and social settings in which a given language had been functioning, usually for a long time, have been replaced by new and quite different ones as a result of irresistible culture contact and clash. (Wurm : )
In Australia and the Americas, the clash has been between indigenous cultures and the colonial ones, and has resulted in the abandonment of the indigenous languages for colonial languages like English, Spanish, French, and Portuguese. The situation in Africa has been argued to be different. Several Africanists have noted that dominant regional African languages are the cause of the endangerment of African languages (cf. Mous ). Legère (: ), for instance, notes: It is well known that the spread of Swahili as a supra-ethnic means of communication in East Africa in general, and Tanzania in particular, took place on the basis of a considerable ethnic and linguistic heterogeneity.
In other parts of East Africa, hunter-gatherers shifted to the language of pastoralists with whom they came into contact (cf. Dimmendaal and references cited there). While African languages play a leading role in the disappearance of indigenous languages, the colonial languages are not particularly blameless, and this has become even more so in recent years (cf. Connell ). It used to be the case that these languages were only spoken by an elite few in Africa and were, therefore, not considered a threat (see Mufwene ; Lüpke ). However, the number of people who use them exclusively has been expanding beyond the elite class, especially in recent years. Djite (: ), for instance, reports that: [many] middle-class families are now bringing up children in European languages, and many from the masses also seek to become, or have their children become, proficient in the same languages, having all come to the conclusion that the uplifting of their socio-economic status lies in these European languages [emphasis mine].
A study by Schaefer and Egbokhare () corroborates this assertion and shows that English plays an important role in the endangerment of indigenous languages in the rural Bendel region of southern Nigeria (cf. Connell ; Ahadzi et al. ). Threat from pidgins and creoles, which have emerged as a result of contact with the colonial languages, are even more serious. Williamson (), for instance, argues that virtually all the indigenous Nigerian languages are under threat from Nigerian Pidgin English. The situation is no different in Cameroon where Good (: –) reports: “in the Lower Fungom context, the greatest threat to linguistic diversity does not appear to be larger local languages, but rather, the spreading lingua franca of Cameroonian Pidgin.”
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The second subtype of ecology which Wurm discusses involves the speech community, not the language itself. He cites instances where most of the young male members of an indigenous community move into urban areas to seek economic betterment. African cities have been growing steadily since the latter part of colonization mostly as a result of the movement of people from rural (often minority-language-speaking) areas into the urban centers. For instance, Esther Dogbe, who wants to document Dompo, an isolate spoken in Ghana, recently reported that most of the native speakers have left Dompofie, where Dompo is spoken, “in search of greener pastures”, leaving only twenty people, the youngest of whom is while the eldest is over years (p.c., July , ). Most of the people who move go to urban areas. Lüpke () cautions that the claim that urbanization leads to language endangerment is debatable. She notes that most people who move to the city maintain ties with the rural area and thereby maintain their languages. She draws, for example, on Vigouroux and Mufwene’s () representation of some portions of the city as smaller extensions of the villages they come from. This is because most new arrivals in the city usually stay with relatives or friends from the village for a long time and, thereby, create pockets of communities in the cities where they are able to use the indigenous languages. Lüpke’s caution is warranted because of an important African trait which differs from places like North America or Australia. Vigouroux and Mufwene () report that Africans have a long tradition of societal multilingualism, where the acquisition of the language of wider communication does not put pressure on people to give up their heritage language. People simply add new languages they find useful to their linguistic repertoire as and when they encounter them. Evans (), referencing Moore, shows that some people in parts of Cameroon are proficient in as many as eight languages. However, multilingualism appears to be on the decrease in urban areas. McLaughlin (: ) reports a telling situation in Dakar, the capital of Senegal, which I quote in extenso: Consider the example of the Dakar family cited in McLaughlin (b). The parents, now in their late sixties are both native speakers of Seereer who came to Dakar in their twenties but have always spoken fluent Wolof, the dominant urban language, as well. They speak Seereer to each other. Their children, in their thirties and forties, have a passive knowledge of Seereer but are not very fluent. When the parents speak to their children the conversations are often bilingual . . . or completely in Wolof. The next generation, that of the grandchildren, does not understand Seereer. They speak only Wolof to their parents and grandparents and consequently Seereer is not used as much in the household as it used to be. In this situation, which is a common one in urban Senegal and is mirrored in cities throughout the continent, the urban language contributes to generational language shift and the youngest generation is no longer multilingual in African languages [emphasis mine].
I emphasize the final part because of my experience when I was documenting Nyagbo, a Ghana-Togo-Mountain language (Kwa, Niger-Congo). I visited a young Nyagbo family in Accra, the capital of Ghana. The man had gone back to his hometown to marry and bring his wife to live with him in Accra. When I visited them, they had a baby who was a few months old. The man spoke Akan to the wife at home. When I asked why, he said he started teaching his wife Akan when she arrived in Accra so that she would be able to get
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
, , .
by, and it ended up becoming their language. I have no doubt that their child will grow up speaking Akan first and, maybe, Nyagbo down the line if they later decide to teach him to speak it. Tellingly, McLaughlin () distinguishes between societal and individual multilingualism, and notes that while the former is rife in the cities, the latter is declining there. I would like to add another dimension to Wurm’s second subtype of ecology. This involves populations that have been chased from their original places as a result of civil strife, drought, or famine. Henderson () reports that over twelve million internally displaced persons (IDPs) constituting more than a third of current IDPs are in sub-Saharan Africa. Lüpke argues that factors such as wars, climate change, and famine play an important role in wiping out languages in Africa and should, therefore, be formalized in language endangerment criteria by such institutions as UNESCO. Note that even though it includes wars, this situation differs from the first type of endangerment described by Wurm because it involves displacement, not the extinction, of speakers of the language.
. D
.................................................................................................................................. The heightened awareness of language endangerment and efforts to address it have led to the emergence of documentary linguistics as a sub-discipline of linguistics. According to Himmelmann (: ), it is primarily concerned with the compilation and preservation of primary linguistic data and interfaces between primary data and various types of analyses based on these data (see also Woodbury ). He proposes that a documentation should “cover all registers and varieties, social or local; . . . contain evidence for language as a social practice as well as a cognitive faculty; [and] include specimens of spoken and written language” so as to serve as a multipurpose and lasting product. For documentary linguistics, then, the goal is not, like the traditional method, to produce a grammar, dictionary, and a few representative texts. Woodbury (: ) stresses that “direct representation of naturally occurring discourse is the primary project while description and analysis are contingent, emergent byproducts which grew alongside primary documentation but are always changeable and parasitic on it.” The key features for documentary linguistics then, according to Himmelmann (: ), are the following: • • • • •
focus on primary data explicit concern for accountability concern for long-term storage and preservation of primary data work in interdisciplinary teams close cooperation with and direct involvement of the community
Documentary linguistics was facilitated by technological changes at the time. Woodbury (: ) describes the situation thus: Suddenly with powerful laptops, digital audio, video, and the world wide web, it at least seems that we should be able to capture and store enormous amounts of information; we should be able to search through this information with unprecedented speed and precision; we should be able to link transcriptions with audio- and videotapes, and entries and dictionaries or
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
statements in grammars with large databases of illustrative examples; we should be able to disseminate around the globe the material now collecting dust in attics or rotting in basements; and we should be able to keep huge amounts of information safe in perpetuity.
A number of prominent Africanists have been critical of some of the tenets of documentary linguistics. Newman (), for example, argues that linguists, as scientists, need to provide scientific justification for working on a language. Ladefoged () also wonders why linguists should be concerned that people are shifting to languages which they find more useful, and cautions that making a case for a language based on “political reasons” is paternalistic. Mufwene (: ) also decries linguists’ focus on simply documenting endangered languages without making the effort to understand “the social dynamics and mechanisms that have produced language loss.” Ameka (: ) writes: [d]espite the skepticism, many Africanists, professionals as well as think tanks, especially on the continent, have taken on the challenge to document, stabilize and promote the maintenance and preservation of threatened and minority languages in Africa.
For instance, Batibo (: ) notes that most of the papers presented in three parallel sessions at the Fifth World Congress of African Linguistics in Addis Ababa in August (i.e. language mapping in Africa, endangered languages, and approaches to language documentation) were case studies involving description and documentation of particular languages. Ameka () also notes that universities are establishing documentary linguistics programs on the continent, among which are the University of Uyo, the Université de Cocody, and Addis Ababa University.
. R
.................................................................................................................................. I mentioned that documentary linguistics requires the involvement of community members in the planning and execution of a documentation project and that the community should also benefit from the output of a documentation project. Himmelmann (), for instance, recommends that documentation be useful for language planning decisions and preparing educational materials. Ultimately, it is supposed to facilitate language revitalization and language maintenance.2 However, McGill and Austin () note that calls for linguists to be involved in activities that aid language revitalization are not usually supported by Africanists. Newman, for instance, calls such engagement “linguistic social work”, which should not be encouraged, and Dimmendaal () points out that revitalization efforts in Africa, just like that in the former Soviet Union by Grenoble and Whaley (), are doomed to fail. Basing his reasons on his work on Tima, a Kordofanian I share Mufwene’s () misgivings about the tendency to use the term “language preservation” interchangeably with “language maintenance”. He notes that maintenance suggests that the entity is in good standing and efforts are directed towards keeping this condition up, while preservation, by contrast, brings up the image of a lifeless entity that is maintained in a lifelike form in a jar. 2
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
, , .
language spoken in Sudan, Dimmendaal notes that such languages do not have adequate financial backing, which guarantees that revitalization efforts are doomed to fail. There is evidence that language documentation which involved the community has enabled the revitalization of some languages in Africa. The most remarkable instance is that of Naro, a Central Khoisan language spoken by about , people in western Botswana and eastern Namibia, as illustrated by the extensive quote below from Batibo (: ): The Naro Language Project has, as of , translated % of the Bible (http://en.wikipedia. org/wiki/Naro_language_project, ). A popular booklet, known as Daily Manna, containing Bible passages is published regularly. With an average of copies per issue, the booklet has helped in enhancing literacy skills and sensitizing the Naro people to make spiritual reflections. As a result, the Naro language has become vibrant and therefore has attracted second language speakers, particularly the ǀGwi and ǁGana speakers, who find it as a key to better living socially and economically, particularly in having access to the many Naro publications. One of the effects of the documentation-based empowerment of Naro is that the language is now classified as one of the regional dominant languages in Botswana because of its vitality, use in some public domains and ability to attract speakers of other languages.
One might object that a language with , speakers could not be said to have been endangered. The Khoisan languages are among the most endangered languages on the African continent, however, and, as we saw in section ., number is not a proper indicator of endangerment in Africa. I do think the term “revitalization” is appropriate here to the extent that the language has become attractive not only to native speakers but also to neighboring Khoisan language speakers. It is true that there are several languages with many fewer speakers and, as Lüpke () argues, the heterogeneity of African communities makes it unfeasible to invest in the development of literacy materials for all the minority communities. She argues that even if materials may be produced for some communities, it is practicable to use the writing system of a dominant regional language so as not to expend the limited resources on developing a completely new orthography for all languages. In other cases, it may make sense not to develop any materials in the language itself. This latter proposal would mean that there is no way to learn to read and write in the language. Ameka () takes a different tack regarding the development of writing systems for minority languages. He argues that writing systems often lead to unintended consequences which include the elimination of variation. This is because specific dialects, lexemes, and structures are promoted over others in an effort to standardize orthography. In other cases (such as Ewe), completely new structures are created from those of different dialects, giving rise to a version of the language that is not spoken by native speakers of any dialect. Ameka (: ), therefore, encourages Africanists to “pay more attention to multimodal modes of representation and in particular take advantage of the oral nature of the socio-cultural communities of practice and develop methods that promote diversity rather than kill it.” I agree with Ameka that standardization, which is introduced to ensure the use of orthography for literacy purposes, leads to the homogenization of language varieties.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
In Essegbey () I argue that the way to resolve the problem is to promote vernacular orthography in the sense of Elbow () for minority languages in Africa. This system involves the establishment of a minimum sound symbol correspondence which at the same time allows speakers to “speak as they write”. Variations that are inherent in the language and which are perfectly understandable for speakers are accepted in this system. This has the advantage that speakers do not have to undergo tedious literacy training to be able to read and write, and materials for language development can be organically produced by members of the community themselves.
. C
.................................................................................................................................. In this chapter, I have discussed language endangerment, documentation, and revitalization in Africa. I have noted that African languages lack detailed verifiable documentation to properly categorize the status of languages. While the high degree of multilingualism in Africa makes the situation there different from that of places like North America and Australia, things are changing. Increasing urbanization is leading to decreasing individual multilingualism. Also, although minority African language speakers used to shift to dominant regional African languages, in more recent times, colonial languages and the pidgins and creoles created from contact with colonial languages have been posing an increasing threat to the vitality of the minority languages. I proposed that rather than abandon the development of writing systems for minority language communities because of lack of resources, we should develop vernacular orthographies. These are easier to learn and help maintain the linguistic variation.
R Ahadzie, S., Ameka, F., and Essegbey, J. (). ‘Language use at home and performance in English composition in multilingual Ghana’, Afrikanistik-Aegyptologie-Online , urn:nbn:de:- https://www.afrikanistik-aegyptologie-online.de/archiv/// accessed July , . Ameka, F. (). ‘Unintended consequences of methodological and practical responses to language endangerment in Africa’, in J. Essegbey, B. Henderson, and F. McLaughlin (eds.), –. Batibo, H. (). ‘Language documentation as a strategy for the empowerment of the minority languages of Africa’, in M. Matondo, F. McLaughlin, and E. Potsdam (eds.), Selected Proceedings of the th Annual Conference on African Linguistics. Somerville, MA: Cascadilla Proceedings Project, –. Bechert, J. (). ‘Universalienforschung und Ethnozentrismus’, in W. Bahner, J. Schildt, and D. Viehweger (eds.), Proceedings of the th International Congress of Linguists , Berlin/GDR. vols. Berlin: Akademie-Verlag, –. Blench, R. M. (). ‘Endangered languages in West Africa’, in M. Brenzinger (ed.), –. Brenzinger, M. (ed.) (). Language Death: Factual and theoretical explorations with special reference to East Africa. Berlin: Mouton de Gruyter. Brenzinger, M. (ed.) (). Language Diversity Endangered. Berlin: Mouton de Gruyter. Brenzinger, M., and Dimmendaal, G. J. (). ‘Social contexts of language death’, in M. Brenzinger (ed.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
, , .
Brenzinger, M., Heine, B., and Sommer, G. (). ‘Language death in Africa’, in R. H. Robins and E. M. Uhlenbeck (eds.), –. Connell, B. (). ‘Endangered languages in Central Africa’, in M. Brenzinger (ed.), –. Connell, B. (). ‘The role of colonial languages in language endangerment in Africa’, in J. Essegbey, B. Henderson, and F. McLaughlin (eds.), –. Dimmendaal, G. J. (). ‘On language death in eastern Africa’, in N. Dorian (ed.), Investigating Obsolescence. Cambridge: Cambridge University Press, –. Dimmendaal, G. J. (). ‘Different cultures, different attitudes: but how different is the “African situation” really?’, in J. Essegbey, B. Henderson, and F. McLaughlin (eds.), –. Djité, P. G. (). The Sociolinguistics of Development in Africa. Clevedon: Multilingual Matters. Elbow, P. (). Vernacular Eloquence: What speech can bring to writing. Oxford: Oxford University Press. Essegbey, J. (). ‘ “Is this my language?” Developing a writing system for an endangered-language community’, in J. Essegbey, B. Henderson, and F. McLaughlin (eds.), –. Essegbey, J., Henderson, B., and McLaughlin, F. (eds.) (). Language Documentation and Endangerment in Africa. Amsterdam and Philadelphia, PA: John Benjamins. Evans, N. (). Dying Words: Endangered languages and what they have to tell us. Malden and Oxford: Wiley-Blackwell. Good, J. (). ‘ “Community” collaboration in Africa: experiences from northwest Cameroon’, Language Documentation and Description : –. Grenoble, L. A., and Whaley, L. J. (). Saving Languages. Cambridge: Cambridge University Press. Henderson, B. (). ‘Out of context: documenting languages in immigrant and refugee communities’, in J. Essegbey, B. Henderson, and F. McLaughlin (eds.), –. Himmelmann, N. P. (). ‘Language documentation: what is it and what is it good for?’, in J. Gippert, N. Himmelmann, and U. Mosel (eds.), Essentials of Language Documentation. Berlin and New York: Mouton de Gruyter, –. Krauss, M. (). ‘The world’s languages in crisis’, Language : –. Ladefoged, P. (). ‘Another view of endangered languages’, Language : –. Legère, K. (). ‘Language shift in Tanzania’, in M. Brenzinger (ed.), –. Lüpke, F. (). ‘Ideologies and typologies of language endangerment in Africa’, in J. Essegbey, B. Henderson, and F. McLaughlin (eds.), –. McGill, S., and Austin, P. K. (eds.) (). Language Documentation and Description. London: SOAS. McLaughlin, F. (). ‘Introduction to the languages of urban Africa’, in F. McLaughlin (ed.), Languages of Urban Africa. London: Continuum, –. Mous, M. (). ‘Loss of linguistic diversity in Africa’, in M. Janse and S. Tol (eds.), Language Death and Language Maintenance: Theoretical, practical and descriptive approaches. Amsterdam and Philadelphia, PA: John Benjamins, –. Mufwene, S. S. (). ‘Language endangerment: an embarrassment for linguistics’, Chicago Linguistics Society : –. Mufwene, S. S. (). Language Evolution: Contact, competition and change. London: Continuum. Mühlhäusler, P. (). ‘Review of Young People’s Dyirbal by Annette Schmidt’, Linguistics : –. Newman, P. (). ‘We have seen the enemy and it is us: the endangered languages issue as a hopeless cause’, Studies in the Linguistic Sciences : –. Newman, P. (). The law of unintended consequences: how the endangered languages movement undermines field linguistics as a scientific enterprise. Talk given at the Linguistics Department, School of Oriental and African Studies, October , . Robins, R. H., and Uhlenbeck, E. M. (eds.) (), Endangered Languages. Oxford and New York: Berg. Romaine, S. (). ‘The impact of language policy of endangered languages’, in M. Koenig and P. Guchteneire (eds.), Democracy and Human Rights in Multicultural Societies. Burlington, VT: Ashgate Publishing.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Sasse, H.-J. (). ‘Theory of language death’, in M. Brenzinger (ed.), –. Schaefer, R. P., and Egbokhare, F. O. (). ‘English and the pace of endangerment in Nigeria’, World Englishes : –. Vigouroux, C., and Mufwene, S. S. (). ‘Colonization, globalization and language vitality in Africa: an introduction’, in C. Vigouroux and S. S. Mufwene (eds.), Globalization and Language Vitality: Perspectives from Africa. London: Continuum, –. Williamson, K. (). Language endangerment in southern Nigeria. Paper presented at the Symposium on Language Endangerment in Africa, Second World Congress of African Linguistics, Leipzig, July– August. Woodbury, A. C. (). ‘Defining documentary linguistics’, in P. K. Austin (ed.), Language Documentation and Description. Vol. : Documentary Linguistics. London: SOAS, –. Wurm, S. (). ‘Language death and disappearance: causes and circumstances’, in R. H. Robins and E. M. Uhlenbeck (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
Youth/town language ......................................................................................................................
. I
.................................................................................................................................. “Y” or “town” languages are developing in contemporary African urban centers amongst conditions of multilingualism, globalization, and superdiversity. A number of these languages have had particular names applied to them, and have been described by researchers, notably Nouchi (Ivory Coast), Town Bemba (Zambia), Tsotsitaal (South Africa), Camfranglais (Cameroon), and Sheng (Kenya). This chapter seeks to outline the current understanding of scholars working on the topic and to use examples from the varieties above to consider the following questions: are there common linguistic strategies or characteristics that can be described for these varieties? What differences between the examples arise as a result of national contextual factors? Can these examples of language variation be explained by the concept of language “birth”, or are the varieties merely “slang” registers? The chapter makes the argument that overly simplistic categorization may obscure the diverse and creative strategies and styles that are being employed in these complex contexts. Blommaert and Rampton (: ) in their position paper “Language and Superdiversity” argue for the strong placement of sociolinguistics and linguistic anthropology to engage with the “contemporary social changes associated with superdiversity”. African cities are fertile sites for social and linguistic change and interaction. To take South Africa as an example, since its inauguration as a democratic country in at the end of apartheid it has attempted to negotiate eleven official languages (although English has increasingly taken center stage as the global language which enables access to knowledge; cf. Crystal ), and has seen a dramatic increase in immigrants internally to urban centers from historically segregated rural “homelands”, continental immigration from other African countries, and a steady stream of international immigrants entering the country since democratization, resulting in the rapid growth of cities (Kok : ). Additionally, the influx of modern information and communications technologies (ICT) has led to the intrusion of the global economy and the knowledge economy with resultant effects on
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
business, investment, advertising and consumer brands, media and technology, music, and film, leading to rapid modernization and transformation particularly in the highly networked urban centers (Muller et al. ). This complexity is equaled in other urban centers around the continent. Kube-Barth () describes the situation in Abidjan, the largest city in Ivory Coast, where twenty years of “massive migration from rural areas and from neighbouring countries” has led to a situation wherein % of the current population are non-Ivorian, and the Ivorians belong to all sixty ethnic groups. These urban contexts contain multiple languages: in Kenya as many as sixty-nine languages are spoken, in Cameroon , Zambia seventy, and Ivory Coast sixty, and the multitude of local languages compete in these environments with colonial languages such as French and English (Lewis ). Within these diverse urban locales, a number of “informal” or “non-standard” languages are being reported. The following sections describe some of the examples in more detail, in order to inform the subsequent discussion regarding their status.
. E / A ..................................................................................................................................
.. Nouchi (Ivory Coast) Nouchi is described by Newell (: ) as a register which is “widely associated with urban youth, criminality, and impoverished education” which is symbolic of a new Ivorian national identity. Kube-Barth (: ) calls it an “urban hybrid language”, whilst Newell argues that it is not simply a slang or an anti-language, because of its autonomy (being derived from the streetwise status of Nouchi), and that today it is being claimed as an Ivorian language which “should potentially be considered as its own standard” (Newell : ). It is spoken by youth between and years old (Kouadio ) and is gendered; Newell suggests it has “something to do with masculinity” (Newell : ), although this is not directly explored in her article. According to Ivorians, street gangs developed Nouchi prior to the s as a secret language to prevent the police from understanding them, and the term “nouchi” originally applied to this social group of “juvenile delinquents” (Ploog : ). Today it is used “stylistically” (Newell : ) by youth around the country across all classes and is also used in the media (Newell ; Kube-Barth : ). Nouchi appears to have entered into popular discourse around and was accompanied by: a new urban Ivoirian style involving handkerchiefs tied around the head and the knees of ripped jeans, as well as a variety of new dance forms through which Ivoirian youth mise-envaleur [placed value on] [sic] their clothes and bodies. (Newell : )
Today Nouchi youth wear brand names imported from the US. One respondent in Newell states: “You can tell who is Nouchi by the way they are dressed, but especially by their
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: /
(1) J’ai un gba avec un mogo. ‘J’ai un rendez-vous avec une fille.’ [I have a date with a girl.] (2) Ça c’ est ma go. ‘Voici ma petite amie.’ [Here is my girlfriend.] (3) Le gboo a behou. ‘Le groupe a fui.’ [The gang has vanished.] (4) Le professeur a mouillé la go. ‘Le professeur a couché avec la fille.’ [The teacher had sex with the student.]
. Comparison of Nouchi and standard French grammar and vocabulary (Vakunta )
mouth, by the way they stand and move.” It also involves a particular style of walking with a limp (Newell : ). Newell claims it is now also spoken in the home “and thus the first language of many Ivoirians” (Newell : ). Ploog (: ) states: “The term now designates the local linguistic identity, including non-marked daily use of spoken, non-standard French. Native urbans grow up with Nouchi-French.” Kube-Barth () describes three main roles that her respondents attribute to Nouchi: a medium of interethnic communication, a future national language, and an indicator of the failure of the national language policy in education. Ploog (: ) and Kube-Barth (: ) appear to agree that it is based on français populaire, the local variety of French (see Figure .), with lexical incorporations from other languages (Newell : ). Words in Nouchi come from, inter alia, English, Spanish, German, Dioula, Baoulé, and Bété (Newell ; Vakunta ). Linguistically, Nouchi is described as a “moving target” (Newell : ), which shifts to remain impenetrable. It involves “generativity and performance” and the “purposeful introduction of neologisms”. According to Vakunta (), Nouchi has an SVO word order. However, there are some transgressions from standard French syntax, for example, omission of the impersonal pronoun il [it], ellipses, and absence of the passive voice. Here though, we must caution that these features may be attributable to français populaire, and not unique to what is termed Nouchi; indeed Ploog () in her analysis of the “features of the microsystem of LA” treats Abidjanese French and Nouchi as the same object of analysis syntactically. This leads to the question, what constitutes Nouchi? There is little consensus on this partly due to its inherently variable nature. Newell (: ) states: Nouchi exists in that all Ivoirians believe it exists; but it has no objective, identifiable existence, in the sense that there is no agreement on precisely what Nouchi is, but rather a continual struggle over the right to define it.
She associates the language with a postcolonial identity, which she argues involves the “enregisterment of modernity in language”. She goes on the say that “such indexicalities represent processual struggles over the semiotic associations of speech which shape local understandings of indigeneity, ‘civilization’, and citizenship” (Newell : ). The current status of Nouchi relates to the place of French in Ivory Coast; she argues that
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Town Bemba:
Awe, layifi yaku Kitwe yaba iyashupa. no life in Kitwe be.pst difficult ‘No, life in Kitwe is difficult.’
Standard Bemba:
Ubwikalo bwaku Kitwe bwaliba ubwayafya. life in Kitwe sm is.pst difficult ‘Life in Kitwe is difficult.’
. Example of Town Bemba compared to standard Bemba (from Kabinga : )
French has been an instrument of state by which its project of Francocentric modernity has been imposed. Nouchi in contrast offers an alternative grassroots national identity. However, she outlines how Nouchi itself has become exclusionary by focusing on a “performance of urban cosmopolitanism” (Newell : ) and alienating those considered “rural” and “northern”. One of her respondents highlighted the belief that Nouchi is progressing beyond slang status: “All countries have some kind of Nouchi, but not as strong as ours. Ours is a real language. It is part of what makes Ivoirians different from other Africans” (Newell : , my emphasis).
.. Town Bemba (Zambia) Spitulnik (: –) describes Town Bemba as a variety of Bemba, or a hybrid language based on Bemba, and proposes that it should be understood as “a cover term for a set of Bemba-based multilingual practices that exemplify urbanity but need not be tied exclusively to urban locales.” She describes Town Bemba as a “moving target”, and adds that while it is not clear what constitutes it precisely, the Bemba base confers systematicity on the variety. Features of the variety include “leveling of the Bemba tense system, a pattern of regularizing irregular verbs, and some reduction of tone and vowel length distinctions”, although Spitulnik stresses that it is not a simplified form of Bemba, and retains its linguistic complexity. She identifies approximately –% of forms in Town Bemba as corresponding to standard Bemba (Figure . compares Town Bemba and standard Bemba). Town Bemba features loan words, coinage, and code-switching or mixing (Spitulnik : ). Borrowing tends to be accompanied by semantic transformation (Kabinga : ), and most of the borrowing has historically come from English, Fanagalo, Afrikaans, and Nyanja, although Kabinga (: ) suggests that the Afrikaans and Fanagalo borrowings are in decline, replaced primarily by English borrowings. Town Bemba is very widespread: it originated as an urban variety of Bemba used as a lingua franca for first language speakers of Lovale, Nyanja, and Lozi as well as Bemba in the Copperbelt mining region of Zambia during the s (Epstein : ; Spitulnik : ). Today it is also a first language and a language of children (Spitulnik : ; Kabinga : ). It is used in everyday life, in the workplace, at social meetings, and in political discussions (Kashoki : ) by both males and females (Kabinga : ). However, Kabinga suggests that still today, “young men, criminals or prisoners, and bakaponya ‘male street vendors’ coin most of the new [Town Bemba] vocabulary” (Kabinga : ). Despite its proliferation, and perhaps due to these links to criminal subcultures, Town Bemba suffers from a negative reputation and has been described as “slang” and “broken
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: /
Bemba” (Richardson : ). Kabinga (: ) compares it to South African Tsotsitaal in that “male speakers of [Town Bemba] and Tsotsitaal speak and dress in a ‘stylish’ way.” Town Bemba has connotations of modern, non-traditional, and urban life; it is positioned as “urban” in opposition to “rural” standard Bemba and carries “contradictory social values of being both a prestigious, cosmopolitan code and a corrupted, even devious code” (Spitulnik : ). Speakers of standard Bemba describe Town Bemba as corrupting of “pure” or “deep” Bemba (Spitulnik : ) or “a disrespectful language” (Ferguson : ), while speakers of Town Bemba describe it as “switched on” and speakers of rural standard Bemba as “backward” (Spitulnik : ) or as bakamunshi, which means ‘villager’ (Kabinga : ). On the other hand, Spitulnik (: ) explains that it is often also the unmarked choice: Thus, Town Bemba encompasses everything from the tough street language of mishanga boys (street vendors who sell individual cigarettes) to the playful language of upper-class university students, from the sophisticated language of business people to the everyday language of many urban families.
She suggests there may be specific marked varieties of Town Bemba spoken by urban youth and the urban criminal culture; she subsumes all these varieties under the term Town Bemba but states that documentation is needed of the degrees of overlap (Spitulnik : ). Thus, Town Bemba requires us to clarify the distinctions between urban languages and youth languages, and whether this distinction is useful or practical where a continuum of registers within a variety exhibits no clear divisions along these lines. Ultimately Spitulnik argues that despite its longevity Town Bemba remains ill-defined because it is in flux, and likens this flux to the city and modernity: “[t]he heteroglossic nature of Town Bemba thus reflects the heterogeneous nature of urban life and the way that notions of modernity have taken root in Zambia” (Spitulnik : ).
.. Tsotsitaal (South Africa) Tsotsitaal in South Africa has received somewhat more attention in the literature perhaps than some of the other varieties discussed here (Ndlovu ; Msimang ; Dube ; Ntshangase ; Calteaux ; Khumalo ; Makhudu ; Ngwenya ; Childs ; Mulaudzi and Poulos ; Molamu ; Brookes ; Sekere ; Rudwick ; Bembe ; Bembe and Beukes ; Aycard ; Mesthrie ; Nkosi ; Hurst , ); yet some of this has been misleading. Slabbert and Myers-Scotton’s () well-cited article discussing the status of Tsotsitaal and Iscamtho is one such, mainly due to problems around naming of the varieties. Building on the work of Ntshangase (), the authors made the claim that Tsotsitaal is an Afrikaans-based variety, while Iscamtho is an African language-based variety. This claim was most likely a result of the geographical sources of the data they analyzed, and has led to subsequent authors treating these varieties as quite distinct. However, Tsotsitaal is a term used in South Africa today to describe a number of different informal varieties that utilize a range of different base languages (Mesthrie and Hurst ). Iscamtho, on the other hand, is a term used to refer to some Sowetan tsotsitaals (Aycard ). In this chapter,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
I will use the name Tsotsitaal to refer to all these varieties, and the noun tsotsitaals to refer to the underlying phenomenon. The history of these varieties can be traced to an Afrikaans-based slang in Sophiatown, a racially mixed suburb of Johannesburg in the s, which was originally referred to as “Flytaal”. The term tsotsi emerged to describe firstly a style of trousers, and later petty criminals who used the style (Glaser ). The term is still commonly used today to refer to street hooligans or members of street corner gangs in South Africa’s townships. Taal is the Afrikaans word for ‘language’. Thus tsotsitaal has become the coverall term for the informal varieties that have blossomed in townships over the past seventy years. Lexical items and style relate these varieties—they are still today often accompanied by modern urban dress styles (with a focus on imported American brand names) and musical preferences (particularly forms of hip-hop), as well as particular body language that add a communicative layer (Mesthrie ; Hurst ; Hurst and Mesthrie ). In an attempt to delineate the tsotsitaal phenomenon I coined the term “stylect” to refer to varieties such as this (Hurst ). The meaning of stylect is that of a “lexicon (lect), inseparable from a discursive practice (style)” (Hurst and Mesthrie : ). Tsotsitaal shares many of Nouchi’s features. It is a moving target, subject to a lack of consensus on who speaks it and what constitutes it. It is reportedly spoken by “tsotsis”, who, similarly to “nouchis” are urban criminals (Hurst ; Newell : ). One etymology of nouchi is that it derives from American movies: from nou-chi ‘hair of nose’ in Susu, a Mande language from Guinea (Ploog : ), it is said to refer to the moustaches of cowboys in westerns (Newell : ), while tsotsi is said to have derived from zoot suits, the term used for a style of clothing worn especially by gangsters in American movies (Nixon : ). Newell (: ) also mentions that speakers of Nouchi are associated with a particular way of walking with a limp, which has also been identified as a body language feature of Tsotsitaal speakers (Hurst ). Tsotsitaal has been observed utilizing a Zulu, Sotho, Xhosa, and Afrikaans base, and therefore appears to attach itself to whichever language is the first language of its users (Hurst and Buthelezi : ). In Johannesburg townships such as Soweto this is accompanied by a large amount of code-switching (a phenomenon that arises from the highly multilingual contexts of the township) (Aycard ). In Cape Town on the other hand, where I did my Ph.D. fieldwork, speakers are primarily proficient in Xhosa, and this is the base language observed there (Hurst ). Figure . provides examples of Xhosa and Zulu base. (1) Xhosa example with derivations (author’s data, 2006) Xa u-blomme n-amajita, amajita a-khipha ngestyle. X X-A X-T T X-E -X X- E ‘When you hang out with the guys, they keep you in style.’ [X = Xhosa; A = Afrikaans; T = Tsotsitaal; E = English] (2) Zulu example: isiTsotsi and standard Zulu comparison (Rudwick 2005) Vusi: Serious majitha, OK asambe siyoyibona … ikhona la ngichibuka khona. [Iqiniso bafowethu, kulungile asihambe siyoyibona … yikhona-ke la ngihlala khona.] ‘Serious guys, OK, let’s go and see my room … alright, this is where I am staying.’
. Examples of Tsotsitaal with a Xhosa and Zulu base
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: /
This serves to underline that Tsotsitaal is not a code-switching phenomenon; speakers who are not bilingual or multilingual do not exhibit code-switching patterns, but merely utilize borrowings from other languages (alongside lexical innovation or neologisms). These borrowed lexical items are often semantically transformed to the point that their meaning is only relevant within the variety. An analysis of the literature (Hurst and Mesthrie : ) revealed a set of core terms that are common to all instances of Tsotsitaal, notwithstanding the grammatical frame. Many of the terms are longstanding and some even occurred in the original Afrikaans version in the s (Molamu ). In common with the other varieties described here, Tsotsitaal is (and was historically) positioned as urban in opposition to rural “pure” versions of the standard languages that it takes as its base (Hurst ). A figure which has been part of the code historically is the “clever”; a streetwise urban youth who speaks Tsotsitaal fluently (Glaser ). Today it is no longer restricted to speakers involved in criminal activity (although there remain “deep” versions of Tsotsitaal spoken by criminals) and is spoken broadly across urban African language townships by young males and some young females (although there is still stigma attached to female speakers) (Aycard ; Hurst ). There are some claims that in the townships surrounding Johannesburg it is being used in the home as a first language; however, the social status of speakers and its use in formal domains such as the classroom by children who speak it in the home remain to be examined.
.. Camfranglais (Cameroon) The next example, Camfranglais from Cameroon, is said to have originated as a language of criminals in the s but has become widespread amongst youth (Schröder : ; Féral ). It incorporates aspects of many of the languages spoken in Cameroon and has several names (Langage de bandits de Douala, Pidgin French, Franglais, Camspeak, Francanglais). It can be used via Cameroon Pidgin English (CamP) or via standard French (Schröder : ). Schröder’s analysis leads her to conclude that Camfranglais via CamP syntax (Figure .) can be considered as a relexicalized CamP (Schröder : ). Similarly, she asserts that via standard French Camfranglais (Figure .) keeps the morphosyntax, “whereas most of the lexicon has been replaced by new coinages” (Schröder Camfranglais
Kikman dem flop fo Douala.
CamP
Tifman dem plenty fo Douala. ‘There are plenty of thieves in Douala.’
. Camfranglais via CamP syntax (Schröder )
Camfranglais
Les klangdou sont à Sawa flop.
Standard French
Les voitures clandestines sont nombreuses à Douala. ‘There are many clandestine cars in Douala.’
. Camfranglais via French syntax (Schröder )
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: ). She poses the question: is Camfranglais, therefore, “relexified French, relexified CamP, or a language in its own right?” (Schröder : ). Schröder (: ) argues that while it is tempting to claim it is a relexified version of existing languages: a knowledge of English, French, or any other of the languages involved does not guarantee the hearer comprehension, as lexical items frequently acquire a new and often unexpected meaning, which only an initiated insider would be able to decode. Camfranglais thus seems to qualify as a so-called ‘mixed language’: i.e. a language that “combine[s] the grammatical system of one language with the lexicon from another language.”
Furthermore she asserts that Camfranglais qualifies as what Smith defines as a “symbiotic mixed language”. Smith’s () explanation of a symbiotic mixed language has some parallels with Halliday’s () definition of an anti-language (a language spoken by an anti-society, which is a society set up in opposition to the norm): he says the symbiotic mixed language exists in relationship to the dominant unmixed languages whose grammar it utilizes, is never the only language of its speakers, and is often used as a secret language (Smith : ). Schröder makes the argument, in agreement with Smith, that varieties like this should not be seen as “mere varieties of non-mixed languages” (Schröder : ) due to two qualities that distinguish them from a simple variety: a different “cultural identity”, and a different lexis, which Smith says makes the code “incomprehensible to the host community” (Smith : ). Schröder describes how Camfranglais is a marker of youth identity, is used as a “secret in-group language”, and is “not the result of unconscious language mixing: i.e. codeswitching or code-mixing” (Schröder : ). Like Tsotsitaal it is not necessarily spoken by bilinguals and is primarily used by “urban juvenile francophones in the francophone part of the country” (Schröder : ). She concludes that: it is linked to similar linguistic phenomena described for other African countries, such as the rise of Sheng and Engsh in Kenya and Tsotsitaal in South Africa. With these it shares more than just some parallels in use and lexical preoccupations. What all these codes represent is the urge of young Africans to impregnate dominant Standard and colonial languages with local flavor and to add local color to a global language. (Schröder : )
In the case of African language-based tsotsitaals and Town Bemba, speakers are in fact doing the reverse: adding a global flavor to local languages. This begins to highlight some of the contextual factors that impact on these different varieties—the varying statuses of local and colonial languages within African nations.
.. Sheng (Kenya) The final example to be discussed here is that of Sheng from Kenya. Sheng is an acronym for “Swahili-English slang” (Githinji : ) and it is believed to have originated in poor residential areas of Nairobi sometime between the s and s, possibly originating in
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: / Sheng
Sina eni lakini nataka kuvuruta.
Standard Swahili
Sina kitu (pesa) lakini nataka kunywa pombe.
‘I haven’t anything (money) but I want to drink.’
. Example of Sheng compared to Standard Swahili (Githiora : )
the underworld amongst pickpockets (Mazrui : ; Ogechi : ). It has defied classification as a pidgin, creole, slang, or jargon. Its development has been attributed to, inter alia, contact, secrecy, and lack of fluency in standard languages (Githinji : ). Ogechi (: ) states: “Regardless of this lack of consensus on origins, it is accepted that there is a variety that sounds like Kiswahili . . . but with a distinct and an unstable vocabulary.” Today it is a phenomenon used by youth in “multiethnic urban areas” in Kenya (Ogechi : ; Githinji : ). Githinji’s () article focuses on lexical variation and identity in Sheng. Its lexicon is derived from mainly Swahili and English (see Figure .) but also from Kenyan languages such as Dholuo, Kamba, and Gikuyu as well as Hindi and American westerns (Ogechi : ; Githinji : ) and its base is Swahili (Githinji : ). Githiora (: ) proposes that Sheng “is an age-marked, urban dialect of Kenyan Swahili” indicated by English lexical insertions. An English-based variety is also possible, which is sometimes referred to as Engsh (an inversion of Sheng) (Githinji : ). As with varieties discussed above, Githinji disputes the validity of “code-switching” to describe the variety, and refers to the new meanings ascribed to borrowings when they enter Sheng. He states that “Sheng is a lexical issue” (Githinji : ) and that its lexical variations are markers of speakers’ identities. Strategies of lexification include borrowing, arbitrary coinage, syllable inversion, compounding, and clipping (Githinji : –). It is described by Ogechi (: ) as a “peer group language [which] succeeds in isolating the non-initiated speaker because of the language’s innovativeness in creating . . . content words (nouns and verbs)”, suggesting an in-group/out-group dynamic. Sheng is spoken more by males than females (Githinji : ). Githiora (: ) describes how females, even those from Sheng-speaking residential areas, reported less use of Sheng than males, and preferred the broader term “slang”. He links this to research on gender by researchers such as Labov () and Milroy (), who have identified similar conservative linguistic behavior by women in the USA and UK. He suggests females are aware of the negative social significance of Sheng; or, perhaps, they are aware of its covert masculine prestige “associated with toughness, masculinity and local solidarity” (Githiora : ). This does recall examples from the literature on these varieties which suggest that lesbians may be an exception to female avoidance of the varieties (e.g. Rudwick et al. ; Maribe and Brookes ). The code does not restructure colonial languages, as was suggested for Camfranglais and Nouchi above, but rather makes use of a dominant African language. Spyropoulos () suggests its success may be a side effect of Kenya’s language policy by providing speakers with a code which makes good lexical use of all the major languages spoken in the city (Githiora : ), which is a suggestion worth bearing in mind when looking at all these examples of informal varieties.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. D
.................................................................................................................................. All of the examples described above have a number of features in common, both linguistic and otherwise. They involve innovation in lexicon, including neologisms and borrowing accompanied by semantic transformation. They are often referred to as “moving targets”, as there is little to distinguish them linguistically from the base languages they rely on, other than a shifting lexicon. Most of them are said to have developed from criminal argots (except Town Bemba, which, nevertheless, is said to take lexical items from criminal speech) and are used primarily by male youth, although Town Bemba appears to have been adopted as a lingua franca, and similar claims are being made for some of the other varieties. The gender aspect, however, appears to have been under-researched at this juncture. There is a clear urban vs. rural dimension, where these varieties are markers of modernity and urbanity. There are also clear links to extralinguistic style in many of the varieties, not that there is a single style associated with them, but that they involve wider communicative strategies which serve to communicate modern, streetwise identities. And there is some indication of a subversive relation to colonial languages. But there are significant differences between the varieties as well. Nouchi and Camfranglais utilize colonial languages as their base, while the other examples utilize African languages—in the case of Tsotsitaal there has been a progression from colonial to African base. Town Bemba may be significantly different from the other examples, the main differentiating features being that it uses only one base language: standard Bemba, that its use is not domain-restricted, that it is spoken by all generations and genders, and that it emerged from contact rather than criminal slang. These particulars may distinguish it enough from the other varieties considered here for Town Bemba to be regarded as a vernacular which is now spoken by children as their L. However, Kabinga (: ) still describes it as an “informal urban variety” and highlights commonalities with Tsotsitaal. Due to some of these similarities, we may be seeing in that example the future progress of some of the other varieties that are currently in the process of developing an L community. Often the reporting of these varieties is accompanied by rhetoric that informal languages are destroying speakers’ abilities in the standard varieties of languages in the country. Language death is a concern in many parts of Africa, as dominant languages such as English and Swahili take over the domains of home languages. But on the reverse side, new patterns of code-switching and informal language use seem to be emerging as prevalent practices, which in some cases have been claimed as “new languages”. So can these emerging informal varieties be described as the birth of languages? Kießling and Mous () describe the varieties as “Urban Youth Languages” and show that they are linked to national project identities that respond to colonialism and that they are attempts at formulating a coherent national identity by youth growing up amongst the complexities of an Africa intersected by global phenomena. Indeed the languages appear to be accompanied by other aspects of identity performance—subcultural styles, which include clothing brands, body language, music, and ideologies, which provide evidence that African youth cultures are traversed by global media and diasporic performances. As mentioned earlier, this author has made the argument that the connection to other aspects
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: /
of identity performance such as clothing, music, and language, warrants a new term for these varieties: stylect. Coupland () describes style in language and the practices of speakers in their linguistic performances as discursive identity constructions, featuring creativity and agency. In the case of stylects, these performances can be playful and innovative, drawn from a dazzling array of resources, both local and global, and involving multimodal enactments of identity. In these varieties, the use of the lexical items is accompanied by code-switching and other linguistic performances which reflect a complicated process of styling within specific contexts, for specific audiences. If we move away from ideas of “code-switching” between standard languages and consider the practice itself, then we begin to uncover dynamic language practices in context which can tell us something more about whether these languages can be considered as “language birth” or not. Authors such as Blommaert () and Rampton () have promoted the idea of repertoires, moving away from “multilingualism” or switching between multiple, discrete languages towards “a sociolinguistics of speech and of resources, of the real bits and chunks of language that make up a repertoire, and of real ways of using this repertoire in communication” (Blommaert : ) or what Harris and Rampton (: ) describe as “situated interaction which, we argue, provides a much better framework for understanding the dynamics of syncretic practice.” The varieties discussed here are prime examples of communicative practices that draw on a range of resources and repertoires in their urban multilingual contexts. Taking Town Bemba as an example, the trajectory of these varieties may follow the path: youth language ! wider language of communication ! language of the home. In this sense, these varieties may “fossilise” (Slabbert and Myers-Scotton ), or may retain their dynamism, yet still become a mother tongue, as Aycard () claims is happening to Iscamtho, a “Tsotsitaal” in Soweto. The potential here is that they may ultimately become new mixed forms of African languages that will need to be acknowledged and officially adopted and taught in schools. However, some factors may resist these trajectories. There is evidence to suggest that these youth slang styles are yet gendered and exclusionary, and most relevant as part of interactional performances relating to “streetwise” status within disempowered neighborhoods (Hurst ). Furthermore, there is still little consensus on the linguistic features of the varieties discussed here, whether they are simply overlexicalized versions of existing languages, or whether they are syntactically distinct from the languages they utilize as their base. Here the discussion of the distinction between “named” youth varieties and urban varieties of African languages needs to be briefly highlighted. McLaughlin (: ) distinguishes “youth and other specialized languages” from urban vernaculars and states that “by their very nature, youth languages are short lived and rapidly changing because they are premised on the assumption that others cannot understand them.” This highlights one of the tensions—that their anti-language nature inherently resists standardization. However, McLaughlin (: ) suggests that “after they become established as youth languages and their speakers grow older they may be adopted by the general urban population and can subsequently become urban vernaculars themselves.” She claims that this process is taking place with Sheng; yet it is unclear whether the other varieties discussed here are on this path of development. The relationship between urban languages
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and youth languages is at any rate a close one: Mesthrie (: ) posits that tsotsitaals have to utilize the syntax of “the most informal/urban/non-standard variety of the base language” rather than rural standard varieties. The status of these languages as modern and urban hinges on this.
. C
.................................................................................................................................. Questions remain regarding the varieties described here: whether this new array of varieties can be considered language birth; whether they are likely to become mother tongues which will need to be acknowledged and perhaps officially adopted (perhaps more likely for contact varieties such as Town Bemba than for youth varieties); whether it is possible, if it is even necessary, to codify these varieties, create grammars, and develop them for formal domains; and whether they threaten existing cultural and linguistic forms. Alternatively, these phenomena may be viewed as youthful repertoires which coexist with vernaculars, standard and formal languages. The answers to these questions are likely to be different depending on the contextual factors in each case. Luckily, the rapid changes that are still taking place in urban centers on the African continent give us the ideal ‘melting pot’ contexts to explore these questions and begin to answer them in the relatively short term. Equally, we need to consider new ways of understanding language in postcolonial contexts. We can reject the temptation to classify and categorize and instead map these languages as fluid practices without clear boundaries between one language and the next, involving innovative and creative communicative cooperations. Analyses of this sort will demand new methodological and theoretical approaches, as well as careful linguistic documentation of language in action. The African contexts described here provide a window on similar practices which are taking place throughout the world in our superdiverse urban contexts; they are worthy of our attention.
R Aycard, P. (). Speak as You Want to Speak: Just be Free! A Linguistic-anthropological Monograph of First-language Iscamtho-speaking Youth in White City, Soweto. MA thesis. University of Leiden. Bembe, P. (). The Use of Slang among Black Youth in Gauteng. MA thesis. University of the Witwatersrand, Johannesburg. Bembe, P., and Beukes, A.-M. (). ‘The use of slang by black youth in Gauteng’, Southern African Linguistics and Applied Language Studies : –. Blommaert, J. (). The Sociolinguistics of Globalization. Cambridge: Cambridge University Press. Blommaert, J., and Rampton, B. (). ‘Language and superdiversity: a position paper’, Working Papers in Urban Language & Literacies, Paper , https://www.kcl.ac.uk/ecs/research/researchcentres/ldc/publications/workingpapers/abstracts/wp-language-and-superdiversity-a-positionpaper- accessed July , . Brookes, H. (). ‘A repertoire of South African quotable gestures’, Journal of Linguistic Anthropology : –. Calteaux, K. (). A Sociolinguistic Analysis of a Multilingual Community. Ph.D. thesis. Rand Afrikaans University, Johannesburg.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: /
Childs, G. T. (). ‘The status of Isicamtho, an Nguni-based urban variety of Soweto’, in A. Spears and D. Winford (eds.), The Structure and Status of Pidgins and Creoles. Amsterdam and Philadelphia, PA: John Benjamins, –. Coupland, N. (). Style: Language Variation and Identity. Cambridge and New York: Cambridge University Press. Crystal, D. (). English as a Global Language. Cambridge and New York: Cambridge University Press. Dube, M. (). Language Attitudes in Soweto—The Place of the Indigenous Languages. MA thesis. Vista University, Pretoria. Epstein, A. L. (). ‘Linguistic innovation and culture on the Copperbelt’, Southwestern Journal of Anthropology : –. Féral, C. de. (). ‘Ce que parler camfranglais n’est pas: de quelques problèmes posés par la description d’un “parler jeune” ’, in M. Auzanneau (ed.), Langues, cultures et interaction. Paris: L’Harmattan, –. Ferguson, J. (). Expectations of Modernity: Myths and meaning of urban life on the Zambian Copperbelt. Berkeley, CA: University of California Press. Githinji, P. (). ‘Bazes and their shibboleths: lexical variation and Sheng speakers’ identity in Nairobi’, Nordic Journal of African Studies : –. Githiora, C. (). ‘Sheng: peer language, Swahili dialect or emerging Creole?’, Journal of African Cultural Studies : –. Glaser, C. (). Bo-tsotsi: The youth gangs of Soweto, –. Oxford and Cape Town: Heinemann. Halliday, M. A. K. (). ‘Anti-languages’, American Anthropologist : –. Harris, R., and Rampton, B. (). ‘Creole metaphors in cultural analysis: on the limits and possibilities of (socio)linguistics’, Critique of Anthropology : –. Hurst, E. (). Style Structure and Function in Cape Town Tsotsitaal. Ph.D. thesis. University of Cape Town. Hurst, E. (). ‘Tsotsitaal, global culture and local style: identity and recontextualisation in twentyfirst century South African townships’, Social Dynamics : –. Hurst, E., and Buthelezi, M. (). ‘A visual and linguistic comparison of features of Durban and Cape Town tsotsitaal’, Southern African Linguistics and Applied Language Studies : –. Hurst, E., and Mesthrie, R. (). ‘ “When you hang out with the guys they keep you in style”: the case for considering style in descriptions of South African tsotsitaals’, Language Matters : –. Kabinga, M. (). A Comparative Study of the Morphosyntax and Phonetics of Town Bemba and Standard Bemba of the Copperbelt, Zambia. MA thesis. University of Cape Town. Kashoki, E. M. (). ‘Town Bemba: a sketch of its main characteristics’, African Social Research : –. Khumalo, N. H. E. (). The Language Contact Situation in Daveyton. MA thesis. Vista University, Pretoria. Kießling, R., and Mous, M. (). ‘Urban youth languages in Africa’, Anthropological Linguistics : –. Kok, P., Gelderblom, D., Oucho, J. O., and Van Zyl, J. (eds.) (). Migration in South and Southern Africa: Dynamics and Determinants. Pretoria: Human Sciences Research Council. Kouadio, J. N. (). ‘Le nouchi et les rapports dioula-français’, Le français en Afrique (ROFCAN): –. Kube-Barth, S. (). ‘The multiple facets of the urban language form, Nouchi’, in F. McLaughlin (ed.), –. Labov, W. (). Language in the Inner City: Studies in the Black English vernacular. Philadelphia, PA: University of Pennsylvania Press. Lewis, M. P. (ed.) (). Ethnologue: Languages of the world, th edn. Dallas, TX: SIL International, http://www.ethnologue.com/ accessed September , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Makhudu, K. D. P. (). ‘An introduction to Flaaitaal’, in R. Mesthrie (ed.), Language and Social History: Studies in South African sociolinguistics. Cape Town: David Philip, –. Maribe, T., and Brookes, H. (). ‘Male youth talk in the construction of black lesbian identities’, Southern African Linguistics and Applied Language Studies : –. Mazrui, A. (). ‘Slang and code-switching: the case of Sheng in Kenya’, Afrikanistische Arbeitspapiere : –. McLaughlin, F. (ed.) (). The Languages of Urban Africa. London: Continuum. Mesthrie, R., and Hurst, E. (). ‘Slang registers, code-switching and restructured urban varieties in South Africa: an analytic overview of tsotsitaals with special reference to the Cape Town variety’, Journal of Pidgin & Creole Languages : –. Milroy, L. (). Language and Social Networks. Oxford: Blackwell. Molamu, L. (). Tsotsi-taal: A dictionary of the language of Sophiatown. Pretoria: University of South Africa Press. Msimang, C. T. (). ‘Impact of Zulu on Tsotsitaal’, Suid-Afrikaanse Tydskrif vir Afrikatale : –. Mulaudzi, P. A., and Poulos, G. (). ‘The “Tsotsi” language variety of Venda’, South African Journal of African Languages : –. Muller, J., Cloete, N., and Badat, S. (eds.) (). Challenges of Globalisation: South African debates with Manuel Castells. Cape Town: Maskew Miller Longman. Ndlovu, R. S. (). A Short Study of Slang in Zulu and the Role it Plays in the Development of the Language. MA thesis. University of Natal, Durban. Newell, S. (). ‘Enregistering modernity, bluffing criminality: how Nouchi speech reinvented (and fractured) the nation’, Journal of Linguistic Anthropology : –. Ngwenya, A. V. (). The Static and Dynamic Elements of Tsotsitaal with Special Reference to Zulu —A Sociolinguistic Research. MA thesis. University of South Africa, Pretoria. Nixon, R. (). Homelands, Harlem and Hollywood: South African culture and the world beyond. London: Routledge. Nkosi, D. Mm. (). Language Variation and Change in a Soshanguve High School. MA thesis. University of South Africa, Pretoria. Ntshangase, D. K. (). The Social History of Iscamtho. MA thesis. University of the Witwatersrand, Johannesburg. Ogechi, N. (). ‘On lexicalization in Sheng’, Nordic Journal of African Studies : –. Ploog, K. (). ‘Subversion of language structure in heterogeneous speech communities: the work of discourse and the part of contact’, Journal of Language Contact : –. Rampton, B. (). ‘Style contrast, migration and social class’, Journal of Pragmatics : –. Richardson, I. (). ‘Some observations on the status of town Bemba in northern Rhodesia’, Journal of African Language Studies : –. Rudwick, S. (). ‘Township language dynamics: isiZulu and isiTsotsi in Umlazi’, Southern African Linguistics and Applied Language Studies : –. Rudwick, S., Nkomo, Kh., and Shange, M. (). ‘Ulimi Iwenkululeko: township “women’s language of empowerment” and homosexual linguistic identities’, Agenda : –. Schröder, A. (). ‘Camfranglais—a language with several (sur)faces and important sociolinguistic functions’, in A. Bartels and D. Wiemann (eds.), Global Fragments: (Dis)Orientation in the new world order. Amsterdam: Rodopi, –. Sekere, N. B. (). Sociolinguistic Variation in Spoken and Written Sesotho: A Case Study of Speech Varieties in Qwaqwa. MA thesis. University of South Africa, Pretoria. Slabbert, S., and Myers-Scotton, C. (). ‘The structure of Tsotsitaal and Iscamtho: codeswitching and in-group identity in South African townships’, Linguistics : –. Smith, N. (). ‘Symbiotic mixed languages: a question of terminology’, Bilingualism: Language and cognition : –. Spitulnik, D. (). ‘The language of the city: Town Bemba as urban hybridity’, Journal of Linguistic Anthropology : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: /
Spyropoulos, M. (). ‘Sheng: some preliminary investigations into a recently emerged Nairobi street language’, Journal of the Anthropological Society of Oxford : –. Vakunta, P. W. (). Ivoirian Nouchi: Cousin to Cameroonian Camfranglais, https://www. postnewsline.com///ivorian-nouchi-cousin-to-cameroonian-camfranglais.html accessed September , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T issue of language contact came to the full awareness of Africanist linguists no earlier than a couple of decades ago. The reasons for such long neglect are manifold but are mostly related to mainstream directions in general linguistics and the special circumstances of the African colonial past. Throughout the late nineteenth and the first half of the twentieth century, research in African linguistics was predominantly occupied with the description and classification of the huge amount of African languages. Divergence as the governing principle of the comparative method led to the setup of language trees with neatly separated branches. Any complication motivated by convergence was avoided in order to cope with this enormous task. In the second half of the twentieth century “global” linguistics was dominated by the generative paradigm: research on the different levels of grammar presupposed coherent language systems residing in the language faculty of a given competent speaker. Any contact-related distortion of the ideally monolingual linguistic structure was set aside as representing corrupted speech and, as such, being out of the scope of the core linguistic discipline. Complications such as code-switching, heavy borrowing, or completely mixed languages were regarded as irregularities triggered by individual behavior or special social contexts, but not within a coherent theory of language. For various reasons, among which the development of the “global village” and huge streams of migrants figure prominently, the dominant linguistic paradigm changed, and researchers became more and more aware of contact-induced language developments. It is now common wisdom that most languages of the world are shaped to a very large extent by contact-induced features that defy essentialist views of closed language systems. Accordingly, language history can no longer be modeled exclusively on the genetic metaphor of diverging dendrograms. Nowadays, an enormous quantity of literature that puts language contact issues center stage is produced, and data from all over the world make it obvious that contact-driven language change is the norm rather than the exception. The prevailing major challenge consists in the formulation of an integrated theory of language contact that accounts for all levels of the subject, ranging from multilingual interpersonal encounters through social and
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
linguistic characteristics of multilingualism on the speech group level to large-scale promotions of areal features. In the sections . and ., I will briefly look at early works on language contact in Africa before I sum up more recent findings and developments in this emerging research field. Section . is devoted to theoretical and methodological issues. After a short account of the main aspects of an integrated approach to language contact (bilingual and second language acquisition, societal multilingualism, code-switching, and borrowing) different layers of contact-induced language change will be highlighted by means of a brief case study from West Africa (section .).
. E A
.................................................................................................................................. In the nineteenth and early twentieth century the contact issue (without naming it as such) was only discussed in terms of the so-called “Hamitic hypothesis”. Following a population movement scenario first designed by Speke (), researchers on African languages believed that language development in sub-Saharan Africa reflected the idea of “superior” people who had immigrated into Black Africa from the Near East. Lepsius ()1 and later Meinhof () put forward an evolutionary linguistic scenario of “superior” Hamitic languages affecting “less developed” black African languages, thereby yielding the mixed2 Bantu languages. It was Greenberg () whose work led to a widespread rejection of the Hamitic hypothesis. Post-World War II African linguists, then, were mainly occupied with classification and reconstruction tasks, with only a few exceptions. Heine’s overview of African lingua francas () was the first serious step in considering linguistic contact relations in Africa. In this book, he described the history and sociolinguistic setting of thirty-eight African languages of wider communication, including African-based pidgins and/or creoles. The latter languages inspired him to reflect on the origins and possible classification of this kind of so-called mixed language. Consequently, in , Heine presented data on seven pidgin languages from the Bantu area and explored whether their genetic relationship can successfully be analyzed within the comparative framework. Based on lexical comparison using the Swadesh -wordlist, regular sound correspondences, and a comparison of nonlexical forms (e.g. noun class prefixes, verbal derivation), he concluded that the analyzed pidgins are not mixed languages and that any pidginization model based on language mixing must therefore be rejected (: ). However, contact phenomena play a role in so far as the sociolinguistic histories of the analyzed pidgins all show a sudden change in the relations between different speech communities that had not been in contact before and are suddenly forced to communicate in specialized domains (: –).
But see Schuchhardt for a critic as early as . Meinhof (/: ) regarded Bantu languages as an offspring “von hamitischem Vater und nigritischer Mutter [from Hamitic father and Nigritic mother]”. 1 2
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
From to the French “Laboratory of Languages and Civilizations of Oral Tradition” (LACITO) initiated a series of five publications named “Contact of Languages and Contact of Cultures” (Contact de langues et contact de cultures).3 These booklets unite contributions mainly from a contact zone in the geographical center of Africa, namely southern Chad, Cameroon, and the Central African Republic. Different authors cover a wide array of subjects, ranging from linguistic demography and the language repartition among multilingual speakers through the analysis of lexical borrowings and neologisms based on metropolitan French to detailed studies of the respective sociolinguistic backgrounds for the presented contact phenomena. Even though the editor of the series, JeanPierre Caprile, only intended a reflection on practical problems resulting from language contact and multilingualism, e.g. for education and nation-building, the overview is also a first attempt at unraveling the multiple levels and different aspects of language contact in Africa.
. A :
.................................................................................................................................. The highly influential book Language Contact, Creolization and Genetic Linguistics by Thomason and Kaufman () was the starting point for a growing body of research on contact-induced language change. In Africa, this development gained momentum from the s onwards when many researchers started to question the “traditional” language change and development models and became interested in language convergence phenomena. Research along this line of interest included work on linguistic and social aspects of code-switching, contact-induced language change, and the dynamics of language maintenance and shift in multilingual communities. While all this must be considered as part of the general study of languages in contact, “we are still far from integrating them into a coherent and comprehensive theoretical/methodological framework” (Winford : ). That being said, a short outline of some major achievements in this broad field from an Africanist’s point of view will be in order. Following up on Dixon’s () provocative essay The Rise and Fall of Languages, the question of Areal Diffusion and Genetic Inheritance (Aikhenvald and Dixon ) became top of comparative linguistics’ agenda. The African language phyla in particular came under scrutiny, and the Greenbergian paradigm was reconsidered. Some of the features formerly regarded as prime elements of an eventual reconstruction of Proto-Niger-Congo (PNC) were reanalyzed, taking convergence as an alternative or at least a supplementary scenario into account. Dimmendaal (: –) demonstrated that the attempt to reconstruct a ten- (or nine-)vowel system for PNC may be erroneous. Based on the assumption that [ATR] vowel harmony of the cross-height type generally stabilizes a given vowel system (Stewart : ), many Niger-Congo (NC) specialists believed that the reconstructed ten-vowel systems The first four volumes were edited by Caprile (a, b, , ). The last volume, edited by Herrmann Jungraithmayr (), was finished in but only published five years later. 3
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
within Benue-Congo (Williamson /) and Kwa (Stewart , ) also represented the PNC forerunner. Consequently, languages with fewer than ten (or nine) vowels must have lost them over time (Bendor-Samuel : ). However, there are also good arguments for an areal diffusion of vowel harmony and subsequent expansion of vowel systems like the internal variability within NC families (Dimmendaal : ) and the fact that harmony systems apparently spread into neighboring Afro-Asiatic and Nilo-Saharan languages (Dimmendaal : ). Dimmendaal further examined other phonological (nasalized vowels), morphological (noun class systems), and syntactical (serial verbs) features that were commonly held to be in favor of the genealogical unit NC. While he also found evidence for areal diffusion of, for example, noun class systems, he ultimately stuck to divergence as the main principle for language change in Africa, concluding that “there is little evidence for extensive morphological borrowing in African languages” (Dimmendaal : ). As Africa is comparable to other regions in the world where multilingualism and a general lack of normative forces are also the rule but where many more contact-induced features can be found (Aikhenvald ; Dixon ), an explanation is hard to find. Dimmendaal’s solution is hardly convincing: This would leave a deeply rooted difference in the role played by language in contact situation in these African communities as against Australia and the Amazon as the only plausible, alternative explanation for the observed distinct outcomes. (Dimmendaal : )
In the same volume, Heine and Kuteva () presented the case of Nile Nubian languages where far-reaching language convergence phenomena have been detected4 that also affect integral parts of derivational and inflectional morphology. While they state that cases like Nile Nubian are not very common in Africa, they also admit that this may be due to a general lack of interest in such contact situations and their underlying sociolinguistic settings (Heine and Kuteva : ). Their proposition is to look more closely and also search for cases beyond the usual lexical, phonological, and morphological borrowing phenomena. One such possible field of research is a contact-induced process named “metatypy” (Ross , ). The idea behind it is that speakers using two languages continuously side by side tend to minimize their cognitive and linguistic effort of translating between two languages. The outcome is the metatypic process that makes literal intertranslatability easier between the languages involved, that reorganizes the range of meaning of words in each language so that they become closer, and that reshapes the way events are structured into predicates and arguments so that they become also closer to each other (Ross : ). Heine and Kuteva searched for such processes in African languages and extended the notion to a “grammaticalization process that is due to the influence of one language on another” (Heine and Kuteva : ). And they did find such grammaticalizing metatypy in Africa when looking at grammatical encoding of comparative constructions and reflexive markers. Based on a balanced selection of language data (Stassen ), they found that comparative constructions in African languages are time
The original account of this contact scenario was first presented by Bechhaus-Gerst (, ) who based her findings on combined linguistic, archaeological, and Egyptological evidence. 4
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and again based on an event schema—X (comparee) is Y (predicate) surpass Z (standard of comparison)—involving an action verb meaning something like ‘surpass, defeat, exceed’. This schema is found in thirteen out of twenty-three African languages and it cuts across genetic and regional boundaries (Heine and Kuteva : –). The same observation holds true for reflexive markers with over % of African languages from the sample grammaticalizing either ‘body’ or ‘head’ to become a reflexive pronoun. They summarized their findings by stating that there are actually patterns indicative of an areal rather than a genetically defined distribution that involve meaning and the way it is encoded. They also expressed a desideratum for research on contact-induced language relationship being still in its infancy: What we now need most urgently is, first, sociolinguistic micro-studies of speech communities in contact that would allow us to describe in more detail why and how exactly people adopt propositional schemas of the kind discussed in this section from other speech communities. (Heine and Kuteva : )
Despite these insights, contact linguistic research in Africa at the beginning of the new millennium is still at the “hunting and gathering stage” (Haspelmath : ). But there is growing awareness of areal distribution of linguistic elements and research on convergence phenomena in all language families of the continent. The state of the art of areal typology in Africa is documented by Heine and Nurse (). They bring together data on language convergence from all linguistic levels, and conceive Africa as a major linguistic area with specific linguistic subregions defying classic genealogical views. For instance, Güldemann () presents the “Macro-Sudan belt” as a vast convergence area sandwiched between the Sahel in the north and the Congo basin in the south, spanning the continent from the Atlantic Ocean in the west to the Ethiopian plateau in the east. Based on the analysis of six cross-linguistically highly marked features (e.g. logophoricity, labiovelar consonants, S–(Aux)–O–V–X word order) frequently present in the area’s languages but absent or rare outside, Güldemann has demonstrated a geographical cohesion cutting across major language families and subfamilies. Within this MacroSudan belt he identified an areal “hotbed” roughly covering the central parts of Africa where Benue-Congo (minus Narrow Bantu), Adamawa-Ubangi, Bongo-Bagirmi, and Moro-Mangbetu languages are spoken. In this region most of the features cluster together in a regular fashion. The “hotbed” itself is surrounded by a “core” area where at least three of the six features are present (: –). Güldemann refrained from developing a historical scenario for the linguistic findings but assumed from the specific clustering within the Macro-Sudan that the features did not spread simultaneously from one source only. The only non-linguistic arguments for the existence of the convergence area are the geographical boundaries in the north (Sahara) and south (the Congo basin rainforest) that constantly reinforced population movements along the west–east axis. So, according to Güldemann (: –), the area did not emerge due to accidental historical events but rather as a result of its long-lasting geographical integrity. As for Africa, the emerging field of contact linguistics propelled mostly large-scale areal overviews and critical reviews of the genetically based language families. What is still
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
missing is detailed accounts of the socio-historical processes leading to contact-induced language change: Any survey of the literature on language contact in Africa reveals the dearth of such information [on socio-historical conditions] despite the abundance of linguistic analysis. At times linguists seem preoccupied with the unusual or exotic results of language contact much to the detriment of understanding how such results were achieved. (Childs : )
One remarkable exception to this observation is the insightful book on Repertoires and Choices in African Languages by Lüpke and Storch (). In this book the authors highlight historical and contemporary practices of language use in Africa that all show how multilingual repertoires of speakers finally defy any essentialist view of traditional languages. In their view, the term: ‘language’ is only meaningful in geopolitical terms, lending a discrete identity, status, and power to otherwise fluctuating, hybrid, and changing linguistic practices and creating the illusion of an undifferentiated and homogenous associated ‘community’. (Lüpke and Storch : )
What is still largely missing, however, is a theory that aims at integrating all the different forces and various layers at play when speakers with highly heterogeneous linguistic backgrounds and repertoires at hand, nonetheless, start talking to each other. In the following section, I will briefly outline some key elements for such an integrated theoretical framework of language contact before I sketch an actual approach to such a situation by means of data from a micro-study in a West African contact zone.
. T
.................................................................................................................................. The most recent attempt to integrate synchronic and diachronic aspects of language contact in a unified theoretical framework comes from Matras (). He presents empirically informed data that range from language acquisition and production of bilinguals (or multilinguals) and second language learners through functional aspects of speakers’ multilingual repertoires in appropriate social settings to the large-scale appraisal of societal multilingualism and subsequent language change. His focus is on “the language processing apparatus of the individual multilingual speaker and the employment of this apparatus in communicative interaction” (Matras : ). His central assumption is that languages are not organized as discrete systems in the multilingual mind but rather as a complex interrelated repertoire of linguistic structures and routines that become associated with a range of social activities. The multilingual speaker is thus constantly challenged to maintain control over her/his complex repertoire of forms and structures and to select those forms that are context-appropriate (Matras : ). The outcome in a multilingual setting may then be a creative, innovative hybrid construction that is both task-effective and context-appropriate from the speaker’s point of view. This creative process—“pivot-matching” in Matras’ terms—leads to the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
enrichment of the constructions in the speaker’s repertoire that may become appropriate in a given set of (multilingual) interaction contexts (Matras : ). Once such contact-induced interference is produced by a single speaker, the question of its propagation and establishment in the recipient language is of crucial importance. This process is mainly socially conditioned, as the innovation needs to be accepted by the speaker’s interlocutor and—in order to become an established language change—has to be repeated by her-/himself and others. Surely, there are also linguistic reasons for the ease or difficulty of transfer, but at this point in the contact-induced language change process it is the social setting that governs the propagation throughout a speech community. Matras claims that linguistic areas arise when “multilingual communities” show “lax normative attitudes” and “flexible identity boundaries” (Matras : ). This seems an appropriate characterization for the social background of at least some African linguistic areas, too. On the linguistic level it seems that some elements are more “vulnerable” to substitution in a contact situation than others. Different attempts have been made to cover the susceptibility of linguistic elements in contact-induced transfer with so-called “borrowing scales”. While Thomason and Kaufman () related the elements of their scale to the supposed intensity of contact between speech groups, others tried to capture the borrowability of categories through frequency counts in a given language corpus (Haugen ; Muysken ), or to relate the relative ease of borrowing to the structural factors of the categories themselves in a kind of universalist approach (Moravcsik ). What is never addressed in all these approaches is the question of the prime motivation for borrowing. It is generally assumed that factors like the prestige of the donor language or a generalization of the repertoire for the ease of translatability are the main forces behind linguistic transfers. Again, a different approach is promoted by Matras (, , ), who demonstrates that borrowability of linguistic categories is also motivated by the roles and functions these categories play in discourse. His speaker–hearer-oriented cognitivefunctional approach explains multilingual speech production as the result of two opposing factors: on the one hand, a multilingual speaker in a multilingual setting produces speech acts that adhere to prevailing social norms by selecting appropriate forms from her/his multilingual repertoire that conform to the setting and goal of a given conversation. The mixing of two or more languages is then just a functional rationale (e.g. transparency of concept, speaker–hearer identity considerations, hearer-sided comfort) for the purpose of goal-oriented communication. On the other hand, there are also less functional and goaloriented slips of elements from contact languages in a multilingual setting. Such uncontrolled interferences happen when a speaker is occupied with the mental task of directing the hearer side processing of the propositional content of her/his speech. Thus, in a multilingual encounter, elements of the “monitoring-and-directing apparatus” that stem from a “pragmatically dominant”5 language, appear quite frequently at discourse-sensitive positions in the utterances of multilinguals (Matras : , ). 5 The notion of the “monitoring and-directing apparatus” refers to a set of linguistic operators that is involved in the hearer’s processing of propositional content (Matras ). It contains all sorts of “utterance modifiers” like discourse markers, focus and modal particles, interjections, tags, etc. The term “pragmatically dominant language” is coined for “the language that has been the target of the speaker’s accommodation efforts until shortly before the latest change of interaction setting” (Matras : ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. L S
.................................................................................................................................. The main purpose of the conjoint research project6 on Mande–Gurunsi language contact was to contribute to the general understanding of language change and the outcomes of language contact in a complex multilingual setting. The goal of the micro-study in the Sourou region of northwest Burkina Faso and adjacent areas in Mali was to document diachronic and synchronic outcomes of contact-induced language change and to correlate them with both linguistic and extralinguistic factors. The contact zone, named after the Sourou river, a tributary of the Mouhoun (Black Volta), has always been an area of retreat for smaller ethnolinguistic groups withdrawing from the dominant West African power centers.7 The northern fringe is dominated by Dogon languages, the east is occupied by Northern Samo-speaking groups, in the center, the endangered Gurunsi language Pana is spoken, and the southern and western parts are dominated by Marka-Dafing. Linguistically, these groups are only distantly (if at all) related. Additionally, a few speakers of Fulfulde, the long-standing vehicular language Jula, and French are part of the linguistic landscape of the Souroudougou. Although (apart from French) all these languages are traditionally part of the NC phylum, they are typologically quite diverse. All ethnolinguistic groups have been in place for at least years, but autochthony is disputed. They are organized in village alliances that cut across linguistic and blood boundaries (Barbieri et al. ) and were originally set up for defensive purposes (Beyer ; Beyer and Schreiber ). The long-standing history of active exchange and regular communication between the different groups resulted in a complex situation of multilingualism that left its effects on all languages and on all linguistic levels. In the following, I concentrate on the endangered Northern Gurunsi language Pana (Beyer a) and sketch some of the main contact-induced changes in this language. A cursory overview of lexical borrowings8 in Pana revealed different layers of interference with the languages of the region. The two main sources for lexical borrowing in Pana are Manding (Marka and Jula) and Northern Samo. The former source primarily lends items from the semantic field “natural environment”, while the latter primarily brings in “crop plants”. Interestingly, Pana lends lexical items to neighboring Dogon languages, but not vice versa (Beyer b). 6
Many thanks go to the Deutsche Forschungsgemeinschaft (German Research Foundation) for having funded this project from to . Research was conducted by Henning Schreiber, then Goethe University, Frankfurt, for the Mande part and Klaus Beyer, Humboldt University, Berlin, for the Gurunsi side. 7 The main forces in the wider region have been the Moose kingdoms, the Fulani Empire of Massina, and the Bambara Kingdom of Segou. While none of these ever held power in Souroudougou they used the region as “slave-reservoir” (Hubbell ). 8 On the basis of items from the cultural vocabulary and basic words, forty-one loans (.%) were identified. A closer look with more data and a better understanding of the languages will certainly increase this number.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Contact-induced language change is also visible on the morphosyntactic layer. The original twelve Proto-Central Gur genders have been reduced to only three in Pana. This attrition process continues with the current petrifaction or loss of singular morphemes and the general replacement of plural suffixes with a uniform plural morpheme {-ya}. Pana is apparently witnessing a complete shift to the plural marking structure of Mande languages, where a bare singular stem takes only one universal plural suffix. The negation pattern also displays signs of contact-induced change. Contrary to its closest genealogical kin, the Northern Gurunsi languages Lyele and Samoma, where negation changes word order from S–V–O to S–Aux–O–V, the imperfective aspect in Pana maintains the S–V–O pattern regardless of polarity. A negative morpheme {-ré} is suffixed to the verb in its imperfective form, which is completely alien to other Central Gur languages. In this instance, evidence hints to adjacent Dogon varieties where a general negative meaning of a suffix {-rV} has been documented.9 There are also manifestations of contact-driven language change on the phonological level. An example is the synchronic variation between labialized and non-labialized consonants. Labiovelars have been reconstructed for Pana’s ancestors Proto-Gurunsi and Proto-Central Gur (Manessy , ). Reflexes of the reconstructed *kp/*gb are the labialized variants kw/gw. But Pana goes even further, as labialization is now a variant in the context of all word-initial obstruents, e.g. kwìrí ~ kùrí ‘louse’, bwèné ~ bòné ‘level’, pwèré ~ pòlé ‘yeast’, swèní ~ sòní ‘twenty’. While Jula, Dogon, and French do not display this feature, both Northern Samo and Marka-Dafing show labialization of velars as a variation in their synchronic phonological systems (Diallo ; Schreiber ). Combined qualitative and quantitative sociolinguistic analyses revealed a correlation between speakers’ integration into local community networks and the synchronic variation of this phonetic feature: central actors are frequent “labializers”, while non-central actors tend towards a non-labializing pronunciation. In terms of the Social Network Approach,10 this reflects ongoing language change, as central actors stick to the established vernacular speech norm, while non-central actors easily pick up new, often prestigious ways of speaking and therefore are likely to lead ongoing language change.11 A similar tendency can be observed in the code-switching behavior of local actors. A survey of code-switching in free speech brought to light two interesting points: firstly, although robust statistical data are not yet available, a first analysis hints at a pattern where central actors are less inclined to code-switch than non-central actors. Secondly, of all of Pana’s possible contact languages switches only occur to French and Jula, with the latter serving mainly as source for specific reference and the former mainly providing discourse markers (Beyer ). Thus, currently ongoing contact-induced language change in Pana is mainly oriented towards Jula and, to a minor degree, French. Given that Jula has been the chief lingua franca of the region for at least three centuries, this is hardly surprising. Jula is not only part of every Pana speaker’s linguistic repertoire; it also is a prestigious language in the region. 9
See Beyer and Schreiber () for further details on this and other morphosyntactic borrowings. A comprehensive overview of the Social Network Approach and its application to African contact settings can be found in Beyer and Schreiber (). 11 A detailed account of ongoing de-labialization in Pana and of the application of the network approach in this research can be found in Schreiber () and Beyer (). 10
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Even though French has not got the status of a daily used lingua franca in Souroudougou, it currently is the prime source for all sorts of discourse-related particles used with growing frequencies in local vernacular Pana. Matras’s explanation (Matras : ) of utterance modifiers in multilingual discourse being a kind of repercussion of a “pragmatically dominant language” does not seem to hold here. The recurrent French utterance modifiers in Pana seem rather to be situated on the continuum between code-switching for the purpose of competence and identity flagging, on the one hand, and already integrated borrowings that even start to spread to the most conservative speakers, on the other. The presented sketch of data allows for a differentiated picture of language change in Pana: various contact languages, possibly at different times, are responsible for established contact-induced change on the lexical and morphosyntactic level. Ongoing change is led by non-central actors who introduce items from Jula and French on the phonological and lexical level into the local vernacular norm. The discourse data hint at the continuum from occasional switches to integrated borrowing. Extrapolating the synchronic sociolinguistic account to a diachronic scenario leads to the following: socially marginalized speakers use elements from prestigious or otherwise dominant contact languages (Samo at one point in time, possibly later Marka, and currently Jula) and thus reshape an established vernacular speech norm. Given the longstanding multilingual setup with very flexible identity boundaries in the Souroudougou and the fact that normative control is only executed by local community networks, the high proportion of non-Gur elements in Pana becomes comprehensible. However, there are also contradictions in the data. Although there is not much lexical material stemming from Dogon, Pana’s imperfect negation pattern seems to be built on a template stemming from those languages. As the synchronic data do not display that kind of syntactic restructuring, we need to ask whether currently ongoing language change in Pana does not touch upon syntactic relations or—perhaps more likely—we have not found them yet. A reasonable hypothesis would be that syntactic change like the attrition of noun classes or the restructuring of negation patterns is a much slower process that is not connected to prestige and social networks but more to long-term exchange like, for instance, recurring patterns of women exchange between the different language communities.
. C
.................................................................................................................................. Language contact is by now a well-researched theme in African linguistics. While the majority of research work still concentrates on classification and reconstruction tasks, we are still lacking detailed studies focusing on the underlying social and cognitive processes that trigger contact-induced language change. Given widespread linguistic areas like the Macro-Sudan belt and a prevailing, in some parts (like urban centers) even rising, multilingual competence of the majority of speakers of African languages, African linguistics provides a vast and promising field for that kind of research. This advantage may at the same time be an obstacle, as this kind of research requires far-reaching language competences in two or even more languages as well as the will and the open-mindedness to delve deep into multilingual communities and study contact-induced language change as it happens.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Aikhenvald, A. Y. (). ‘Areal diffusion, genetic inheritance, and problems of subgrouping: a North Arawak case study’, in A. Y. Aikhenvald and R. M. W. Dixon (eds.), –. Aikhenvald, A. Y., and Dixon, R. M. W. (eds.) (). Areal Diffusion and Genetic Inheritance. Oxford: Oxford University Press. Barbieri, C., Whitten, M., Beyer, K., Schreiber, H., Li, M., and Pakendorf, B. (). ‘Contrasting Maternal and Paternal Histories in the Linguistic Context of Burkina Faso’, Molecular Biology and Evolution (): –, https://doi.org/./molbev/msr Bechhaus-Gerst, M. (). ‘Sprachliche und historische Rekonstruktionen im Bereich des Nubischen unter besonderer Berücksichtigung des Nilnubischen’, Sprache und Geschichte in Afrika : –. Bechhaus-Gerst, M. (). Sprachwandel durch Sprachkontakt am Beispiel des Nubischen im Niltal: Möglichkeiten und Grenzen der diachronen Soziolinguistik. Cologne: Rüdiger Köppe. Bendor-Samuel, J. (). ‘Niger-Congo’, in W. Bright (ed.), The International Encyclopaedia of Linguistics. Oxford: Oxford University Press, –. Beyer, K. (a). La langue pana: Description linguistique, lexique, textes. Cologne: Rüdiger Köppe. Beyer, K. (b). ‘Das Pana im Netzwerk arealer Beziehungen: das Lexikon’, in D. Ibriszimow and K. Winkelmann (eds.), Zwischen Bantu und Burkina: Festschrift für Gudrun Miehe zum . Geburtstag. Cologne: Rüdiger Köppe, –. Beyer, K. (). ‘Language contact and change: social factors in an African rural environment’, Journal of Language Contact (Thema III) : –. Beyer, K. (). ‘Multilingual actors: examples from a West-African contact zone’, in C. de Féral, M. Kossmann, and M. Tosco (eds.), In and Out of Africa: Languages in question. In honour of Robert Nicolaï. Vol. : Language Contact and Language Change in Africa. Paris and Leuven: Peeters, –. Beyer, K., and Schreiber, H. (). ‘Intermingling speech groups: morpho-syntactic outcomes of language contact in a linguistic area in Burkina Faso (West Africa)’, in I. Léglise and C. Chamoreau (eds.), The Interplay of Variation and Change in Contact Settings—Morphosyntactic Studies. Amsterdam and Philadelphia, PA: John Benjamins, –. Beyer, K., and Schreiber, H. (). ‘Social Network Approach in African sociolinguistics’, Oxford Research Encyclopedia of Linguistics, https://oxfordindex.oup.com/view/./acrefore/ .. accessed July , . Caprile, J.-P. (ed.) (a). Contact de langues et contacts de cultures. Vol. : Démographie linguistique: Approche quantitative. Paris: SELAF. Caprile, J.-P. (ed.) (b). Contact de langues et contacts de cultures. Vol. : La situation du Tchad: Approche globale au niveau national. Paris: SELAF. Caprile, J.-P. (ed.) (). Contact de langues et contacts de cultures. Vol. : La création lexicale spontanée en Afrique centrale par emprunt au français. Paris: SELAF. Caprile, J.-P. (ed.) (). Contact de langues et contacts de cultures. Vol. : L’expansion des langues africaines: Peul, sango, kikongo, ciluba, swahili. Paris: SELAF. Childs, G. T. (). ‘Language contact in Africa: a selected review’, in R. Hickey (ed.), –. Diallo, M. (). Eléments de systématique et de dialectologie du marka-kan (Burkina Faso). Ph.D. thesis. Université Stendhal III, Grenoble. Dimmendaal, J. G. (). ‘Areal diffusion versus genetic inheritance: an African perspective’, in A. Y. Aikhenvald and R. M. W. Dixon (eds.), –. Dixon, R. M. W. (). The Rise and Fall of Languages. Cambridge: Cambridge University Press. Dixon, R. M. W. (). ‘The Australian linguistic area’, in A. Y. Aikhenvald and R. M. W. Dixon (eds.), –. Greenberg, J. H. (). ‘Studies in African linguistic classification’, Southwestern Journal of Anthropology : –. Güldemann, T. (). ‘The Macro-Sudan belt’, in B. Heine and D. Nurse (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Haspelmath, M. (). ‘How hopeless is genealogical linguistics, and how advanced is areal linguistics?’, Studies in Language : –. Haugen, E. (). ‘The analysis of linguistic borrowing’, Language : –. Heine, B. (). Afrikanische Verkehrssprachen. Cologne: Infratest. Heine, B. (). Pidgin-Sprachen im Bantu-Bereich. Berlin: Dietrich Reimer. Heine, B., and Kuteva, T. (). ‘Convergence and divergence in the development of African languages’, in A. Y. Aikhenvald and R. M. W. Dixon (eds.), –. Heine, B., and Kuteva, T. (). ‘On contact-induced grammaticalization’, Studies in Language : –. Heine, B., and Nurse, D. (eds.) (). A Linguistic Geography of Africa. Cambridge: Cambridge University Press. Hickey, R. (ed.). The Handbook of Language Contact. Oxford: Blackwell. Hubbell, A. (). ‘A view of the slave trade from the margin: Souroudougou in the late nineteenthcentury slave trade of the Niger bend’, Journal of African History : –. Jungraithmayr, H. (ed.) (). Contact de langues et contacts de cultures. Vol. : Langues tchadiques et langues non-tchadiques en contact en Afrique centrale. Paris: SELAF. Lepsius, C. R. (). Nubische Grammatik, mit einer Einleitung über die Völker und Sprachen Afrika’s. Berlin: Wilhelm Hertz. Lüpke, F., and Storch, A. (). Repertoires and Choices in African Languages. Boston, MA, and Berlin: Mouton de Gruyter. Manessy, G. (). Les langues gurunsi: Essai d’application de la méthode comparative à une groupe de langues voltaïques. Paris: SELAF. Manessy, G. (). Contribution à la classification généalogique des langues voltaïques. Paris: SELAF. Matras, Y. (). ‘Utterance modifiers and universals of grammatical borrowing’, Linguistics : –. Matras, Y. (). ‘Fusion and the bilingual basis for discourse markers’, International Journal of Bilingualism : –. Matras, Y. (). Language Contact. Cambridge: Cambridge University Press. Matras, Y. (). ‘Contact, convergence, and typology’, in R. Hickey (ed.), –. Meinhof, C. (/). ‘Sudansprachen und Hamitensprachen’. Zeitschrift für Kolonialsprachen : –. Meinhof, C. (). Die Sprachen der Hamiten. Hamburg: L. Friederichsen. Moravcsik, E. (). ‘Universals of language contact’, in J. H. Greenberg (ed.), Universals of Human Language. Stanford, CA: Stanford University Press, –. Muysken, P. (). ‘Halfway between Quechua and Spanish: the case for relexification’, in A. Highfield and A. Valdmann (eds.), Historicity and Variation in Creole Studies. Ann Arbor, MI: Karoma, –. Ross, M. D. (). ‘Contact-induced change and the comparative method: cases from Papua New Guinea’, in M. Durie and M. Ross (eds.), The Comparative Method Reviewed: Regularity and irregularity in language change. New York: Oxford University Press, –. Ross, M. D. (). ‘Social networks and kinds of speech-community events’, in R. Blench and M. Spriggs (eds.), Archaeology and Language. Vol. : Theoretical and Methodological Orientations. London and New York: Routledge. Schreiber, H. (). Eine historische Phonologie der Niger-Volta-Sprachen. Cologne: Rüdiger Köppe. Schreiber, H. (). ‘Social networks, linguistic variation and micro-change in an African context: a case study in the borderland of Mali and Burkina Faso’, in W. J. G. Möhlig, F. Seidel, and M. Seifert (eds.), Language Contact, Language Change and History Based on Language Sources in Africa. (Special issue of Sprache und Geschichte in Afrika.) Cologne: Rüdiger Köppe, –. Schuchardt, H. (). ‘Zur afrikanischen Sprachmischung’, Das Ausland: Wochenschrift für Länder und Völkerkunde : –. Speke, J. H. (). Journal of the Discovery of the Source of the Nile. London: Blackwoods. Stassen, L. (). Comparison and Universal Grammar: An essay in Universal Grammar. Oxford: Blackwell.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Stewart, J. M. (). ‘Niger-Congo, Kwa’, in T. E. Sebeok (ed.), Current Trends in Linguistics. Vol. : Linguistics in Sub-Saharan Africa. The Hague: Mouton: –. Stewart, J. M. (). Towards Volta-Congo Reconstruction. Leiden: Leiden University Press. Stewart, J. M. (). ‘The comparative phonology of Gbe and its significance for that of Kwa and Volta-Congo’, Journal of African Languages and Linguistics : –. Thomason, S. G., and Kaufman, T. (). Language Contact, Creolization and Genetic Linguistics. Berkeley, CA: University of California Press. Williamson, K. (/). ‘Vowel merger in harmony languages’, Journal of the Linguistic Association of Nigeria : –. Winford, D. (). ‘Some issues in the study of language contact’, Journal of Language Contact, Thema : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
The case of Ma’á/Mbugu ......................................................................................................................
. B
.................................................................................................................................. T term “mixed language” is highly controversial. Mixed languages do not exist according to Greenberg () and Kortlandt (). The term Mischsprache played an important role in the early scholarship on African historical linguistics. The term and the methodology were critically reviewed by, among others, Heine (). But Mbugu remained the worrying exception that mixed languages, however rare, do exist. Bakker and Mous (: –) propose that a language can be called mixed if % or more of the lexical roots in an average text are from a different origin than the grammar. This definition is purely descriptive and does not claim anything about its history. Historical linguistics is concerned about the processes, and one can encounter ordinary processes with extraordinary results (Thomason ). There is not necessarily a meaningful category “mixed languages”, and for every challenging case we should reconstruct the history and draw conclusions from a comparison of histories, not from a comparison of the extraordinary end results. One intriguing example is that of Ma’á/Mbugu and I will discuss only this case in this chapter and briefly mention some other cases at the very end. The challenge of a truly mixed language and why it should not exist is the fundamental insight of historical linguistics that languages develop from one ancestral language and not from two or more. The fundamental structures and vocabulary come from its ancestral language, no matter how strongly additional lexicon and grammar may have been taken from other languages (see Dimmendaal ). Against this background the “strange case of Mbugu” (Goodman ) has puzzled linguists from the first reports of this language with “a Bantu grammar but a Cushitic vocabulary”. When Meinhof decided to work on Mbugu in it was already known through earlier wordlists that Mbugu had Cushitic lexicon. Meinhof ’s publication showed for the first time the mixed nature of the language. Thomason analyzed the previously collected material in an influential article in (). Mbugu took a crucial place in the even more influential publication by Thomason and Kaufman () as evidence of the fact that grammar can be borrowed, and to a spectacular
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
degree. I hope to have demystified the case of Ma’á/Mbugu in my book (Mous a). I will repeat the main conclusions here briefly and then address the issue whether Ma’á/Mbugu underwent a shift or borrowed grammatical material to a large extent.
. N M M M’ˊ
.................................................................................................................................. When I started my fieldwork on Mbugu, I could rely on an excellent overview by Brenzinger () of all previous publications on and about Mbugu. The prevalent idea was that there were two languages, a (relatively) unmixed Bantu language and a mixed one. The former was called Mbugu and the latter Ma’á in the linguistic literature at the time. But to the speakers these two terms refer to the same thing, and actually the two languages proved to be the same in a way: they share the same grammar but there is a double and parallel lexicon, and the language is called Mbugu, normal Mbugu, and inner Mbugu or Ma’á, normal Ma’á, and inner Ma’á depending on whether one uses the normal vocabulary or the “deep” vocabulary. I have shown that the grammar is in detail identical and that the vocabulary is truly parallel for the most part. The origin of the “deep” vocabulary is, however, not from one source but from several sources: () a South Cushitic source that is close to Iraqw, () Maasai, () an East Cushitic source, and () some other less important sources, including Bantu languages. Interestingly, a number of lexical items in the inner Mbugu lexicon appear to be manipulated forms of the ordinary words. This has led me to the conclusion that language manipulation played a role in the development or creation of the inner Mbugu lexicon, the so-called mixed language.
. P M
.................................................................................................................................. Why do the people have this extra vocabulary, this extra language? All Mbugu speakers know the normal vocabulary, and many but not all know the deep vocabulary. Most speakers who do not know the deep vocabulary are members of the Gonja and Nhkandu clans, and these were the last to come from the Pare Mountains into the Usambara Mountains in Tanzania near the border with Kenya and the Indian Ocean. The vast majority of the people in the Usambara Mountains speak the Bantu language Shambaa (Riedel ), and also most Mbugu speak this language, too. The non-mixed normal Mbugu is quite different from Shambaa but very close to Pare (also known as Chasu). Both Mbugu and some Pare moved from the Pare Mountains to the Usambara Mountains. The Mbugu usually live higher in the mountains, concentrate more on cattle-keeping, and leave the cultivation of bananas to the Shambaa and to the sizeable population of Pare speakers who are present in the area. Mbugu live in roughly three areas: Magamba, near Bumbuli, and near Rangwi, and there are even small linguistic differences between the three areas. But in all three they live in close contact with the Shambaa and the Pare. In my material I find surprisingly little mixing of the codes. Most conversations are in one of the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: ’/
two variants. There is no specific reason to use one or the other variety, except that one uses the normal Mbugu with those who do not speak the inner Mbugu version. I have a tape of an interview in normal Mbugu in which a clear-cut shift to inner Mbugu is made at the moment that somebody passes by and greets in inner Mbugu. Thus one cannot speak of a register because there are no functions related to one or the other variant. The important function of inner Mbugu is its existence. Through its existence (and using it) the Mbugu people have a language of their own that is definitely deviant from the Bantu languages around them. It is a strong marker of being culturally different. All Mbugu consider themselves to be member of the one ethnic identity of Mbugu/Ma’á regardless of whether they can speak inner Mbugu or not. The existence of non-speakers of inner Mbugu within the same ethnic group may explain why they still need the nonmixed normal Mbugu. The acquisition of Mbugu in both its forms is from early childhood. I noticed that small children know both varieties.
. H
.................................................................................................................................. The big question is: how did this situation arise? In the absence of old enough sources, all we can do is speculate on the most likely scenario. A scenario will inevitably link the linguistic history with that of the people. To this end, I collected a number of oral histories in a number of different areas. Although oral histories typically do not give an accurate account of the actual course of history, they do provide valuable input for likely scenarios. Many speak of Vudee as a place in the Pare Mountains from where they were dispersed and where they still went back to until recently for initiation. Some of the Mbugu lived among the Maasai for a long time as a kind of servant group, but the story goes that one day they stole their cattle back from the Maasai and then had to flee into the Usambara Mountains. The histories of the various Mbugu clans differ with respect to the routes through which they entered the Usambara Mountains. Several came through Mombo, where the main road now also leads into the mountains. Others went from South Pare to the nearby North Usambara near Rangwi and those latter clans are the ones who tend not to speak inner Mbugu. The high proportion of Maasai words in the inner vocabulary confirms the oral tradition of living among the Maasai. In fact, to the earliest European travelers the Mbugu looked like Maasai in dress and outfit. In general, nouns are much more commonly borrowed than verbs, but there is a high proportion of verbs among the words from Maasai, and this fact suggest to me that the transfer of Maasai words was already at a stage where the Mbugu were expanding a parallel lexicon to build it into a different “language”. It is very difficult to know when they created this extra lexicon. It is not unlikely that it already started in the Pare Mountains. The Mbugu must have lived for a long time together with the Pare in the Pare Mountains because their grammar at least is basically Pare. Assuming that they were different in culture, and in complexion, and that they spoke a different language, they shifted to Pare in that period. In the last stages of language shift the former language only existed in the new one in the form of a few lexical items that were put on a par with the corresponding ones in the new language; hence a (limited) parallel lexicon. I assume the Mbugu once spoke a Cushitic language. The oldest stratum of the inner Mbugu vocabulary comes from this original East Cushitic language.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The tradition says that the Mbugu came from Lukipia or Likipia in Kenya. Likipia is not far from where the Cushitic language Yaaku was still spoken until recently, but it is also the place from where many Maasai dispersed (Galaty ). The most likely scenario is that the Mbugu left (with or without the Maasai) central Kenya, where they spoke a Cushitic language and had a cattle culture, for the Pare Mountains, where they settled and shifted to Pare. Once the shift was about complete, they expanded their emergent parallel lexicon to one which could serve as a language in an effort to have their own language as expression of their own different culture. They had enough access to non-Bantu vocabulary among the Maasai to enlarge this lexicon, and they had access to a language close to Iraqw. Some of the traditions speak of a period in Mbugwe, an area close to Gorwaa, Iraqw’s closest relative. The names Mbugwe and Mbugu are essentially the same. This may or may not suggest a historic link. Arriving in the Usambara Mountains, various Mbugu groups met and formed an ethnic unit for which the inner Mbugu lexicon served as a clear marker of their deviant ethnic and cultural identity. Possibly the long initiation in Vudee helped in the strengthening of the inner Mbugu lexicon. Nothing is known about what happened in the initiation in Vudee, but such long initiation periods commonly involve learning new (surrogate) languages. The creation of a separate and parallel lexicon for purposes of respect or identity is not uncommon, in general, and in Africa, which shows many examples of registers of respect such as Hlonipha and argots of smiths, hippo-hunters, etc. (Mous , b; see also Friedrich, chapter of this volume).
. L
.................................................................................................................................. Having demystified the strange case of Mbugu, it is no longer a challenge to the one-parent axiom of historical linguistics. A crucial question that still remains is whether the Mbugu shifted, as I suggested in section . above, or whether they incorporated the Bantu grammar into their own language, as Thomason (, ) maintains. Crucial to this debate is the presence of grammar features of the original language in the present-day language. There are a limited number of grammatical features that are not clearly Bantu. I have discussed these before (Mous a: –) and shown that most that are put forward in Thomason () are in fact not good candidates for being grammatical features of the original language for a variety of reasons. The strongest indications for non-Bantu grammar are the causative in -ti and details of the noun class agreement patterns. Quite a sizeable number of verbs in the inner Mbugu lexicon contain an ending -ti that can be identified as causative marker. That causative is not productive (Mous a: –): new causative verbs are formed by using the common and Bantu derivation -ij. This causative -ti appears mainly on verbs form the oldest East Cushitic lexical layer. I see the closest etymological link in a predecessor of Dahalo, which has a causative suffix -Vd which is in the conjugation often followed by a vowel i, and one of the allomorphs contains t rather than d. This causative suffix is an important argument for Thomason, because she claims that it shows that Cushitic grammar remained active for a long time in the Mbugu language. She argues that the suffix is used both on verbs that have undergone the historical truncation rule and on those that have not. To me the truncation rule is part of the form
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: ’/
manipulation that was used in the expansion of the parallel lexicon and, therefore, not a historical sound rule that applied once and without exceptions but one that applied to a number of lexemes, as is common with manipulation processes (Mous a, /). The most interesting issue is the question whether the noun class system is borrowed, or how the noun class system and its agreement developed. Mbugu has a full and standard Bantu noun class system. In some details inner Mbugu and normal Mbugu deviate. This is in the agreement patterns of possessives and demonstratives. Inner Mbugu demonstratives are invariant; they never show agreement. Interestingly, possessives do show agreement when used predicatively but not when used attributively: ()
ki-gí yá ní chá-shaba. -thing this is -of.copper ‘This thing is of copper.’
()
luhambá yá ní lu-ké. -machete this is -your ‘This machete is yours.’
()
ma-ginera yá ma-ké vs. ma-ginera ké -root this -your: -roots your: ‘these roots of yours’ ‘your roots’
The inner Mbugu modifiers that immediately follow the noun have non-Bantu forms and show no agreement. The bilingual speakers gave up the Cushitic gender agreement and took the most common forms of the Cushitic independent demonstrative pronouns as invariant demonstrative words showing a three-way distinction, as in their dominant language, Pare. The same applies to the possessives. Non-agreeing nominal modifiers are not an anomaly in a Bantu noun class system and occur in many Bantu languages for certain modifiers such as numbers; it is also what happens in Swahili/English codeswitching. This initial stage of using invariant modifiers in the tightest connection to the noun survived, but at a later stage they used Pare agreement prefixes for the sentence-level agreement and within the noun phrase for the associative construction (completely Pare) and for the possessives when not immediately following the noun. At this moment they employed the core of Pare grammar also when speaking inner Mbugu. I think that qualifies as language shift. The latest stage in the development of the noun class system in inner Mbugu was the addition of noun class prefixes. Meinhof (: ) reported instability in the noun class system. However, what he referred to is extremely stable and still exists: in the Bumbuli area (where Meinhof did his fieldwork) there are a number of nouns that lack the CV noun class prefix that is present in the same words in the rest of the Mbugu-speaking areas. Meinhof reported a word like aɬa ‘fire’ (from South Cushitic aɬa ‘fire’), which is class and has all the agreement properties of a class word but lacks the prefix mw which it does have in the other Mbugu areas. This is still the case, and there are a handful of words like that, the same ones that Meinhof mentioned in his publication. This is not instability but an indication that the noun class prefix is the last element of the noun class system to develop. I discuss this process of noun class development in more detail in my book (Mous a).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. C
.................................................................................................................................. In conclusion, Ma’á/Mbugu is a fascinating double language for which the inner Mbugu part is truly mixed. It originated in an attempt to recover a language after shift (Sasse ) or, rather, to build a language that is sufficiently different from the language that the people shifted to by expanding a parallel lexicon. The functional need for the language was and is to have a strong marker of ethnic identity and to mark the difference from the surrounding, dominant and different Bantu languages and culture. Ma’á/Mbugu was always the famous challenging mixed language, but there are more instances of extreme language mixture in Africa. To mention just a few, in several big cities youth slangs have arisen as a consequence of the urbanization of the last century (Kießling and Mous ). In some cases these slangs have developed so dramatically that they have become the dominant language of the city, young and old, and of modernity; Nairobi’s Sheng and Tsotsitaal in Gauteng, South Africa for example (see Hurst, chapter of this volume). The Northern Songhay languages are notorious for their mixed nature (see Souag, chapter of this volume). Wolff and Alidou () present the mixed nature of Tasawaq and offer hypotheses of its origin. Kossmann () shows how these (and Berber) languages developed high levels of mixture not only of lexicon but also of grammar by using a double morphological system depending on the etymological source of nouns and verbs. Closer to the Ma’á/Mbugu case is the Bantu language Ilwana that became rather mixed by heavy borrowing from Orma (Cushitic), including transfer of phonological processes and noun suffixes (Nurse ). All these “mixed” languages are the outcome of different linguistic processes in different sociolinguistic settings that led to different but fascinating and challenging results.
R Bakker, P., and Mous, M. (). ‘Introduction’, in P. Bakker and M. Mous (eds.), Mixed Languages: case studies in language. Amsterdam: Institute for Functional Research into Language and Language Use (IFOTT), –. Brenzinger, M. (). Die sprachliche und kulturelle Stellung der Mbugu (Ma’a). MA thesis. University of Cologne. Dimmendaal, G. J. (). ‘Review of Robert Nicolaï (): Parentés linguistique à propos du songhay’, Bulletin of SOAS : –. Galaty, J. G. (). ‘Maasai expansion and the new East African pastoralism’, in T. Spear and R. Waller (eds.), Being Maasai: Ethnicity and identity in East Africa. London: James Currey; Dar es Salaam: Mkuki na Nyota; Nairobi: EAEP; Athens,OH: Ohio University Press, –. Goodman, M. (). ‘The strange case of Mbugu’, in D. Hymes (ed.), Pidginization and Creolization of Languages. Cambridge: Cambridge University Press, –. Greenberg, J. H. (). ‘Are there mixed languages?’, in L. Fleishman, M. Gasparov, T. Nikolaeva, A. Ospovat, V. Toporov, A. Vigasin, R. Vroon, and A. Zalizniak (eds.), Essays in Poetics, Literary History and Linguistics presented to Viacheslav Vsevolodovich Ivanov on the occasion of his seventieth birthday. Moscow: Obyedinyenoye Gumanitarnoye Izdatelstvo (OGI), –. Heine, B. (). ‘Zur Frage der Sprachmischung in Afrika’, in W. Voigt (ed.), XVII. Deutscher Orientalistentag vom . bis . Juli in Würzburg. Wiesbaden: Franz Steiner, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
: ’/
Kießling, R., and Mous, M. (). ‘Urban youth languages in Africa’, Anthropological Linguistics (): –. Kortlandt, F. (). ‘On Russenorsk’, in D. Boutkan and A. Quack (eds.), Language Contact, Substratum, Superstratum, Adstratum in Germanic Languages. Amsterdam: Rodopi, –. Kossmann, M. (). ‘Parallel system borrowing: parallel morphological systems due to the borrowing of paradigms’, Diachronica : –. Meinhof, C. (). ‘Linguistische Studien in Ostafrika X: Mbugu’, Mitteilungen des Seminars für orientalische Sprachen : –. Mous, M. (). ‘Paralexification in language intertwining’, in N. Smith and T. Veenstra (eds.), Creolization and Contact. Amsterdam and Philadelphia, PA: John Benjamins, –. Mous, M. (a). The Making of a Mixed Language: The case of Ma’a/Mbugu. Amsterdam and Philadelphia, PA: John Benjamins. Mous, M. (b). ‘The linguistic properties of lexical manipulation and its relevance for Ma’a and for mixed languages in general’, in P. Bakker and Y. Matras (eds.), The Mixed Language Debate. Berlin: Mouton de Gruyter, –. Mous, M. (/). ‘Non-Bantu grammatical elements in the mixed language Ma’a’, Folia Orientalia /: –. Nurse, D. (). Inheritance, Contact, and Change in Two East African Languages. Cologne: Rüdiger Köppe. Riedel, K. (). The Syntax of Object Marking in Sambaa: A comparative Bantu perspective. Utrecht: Landelijke Onderzoekschool Taalwetenschap (LOT). Sasse, H.-J. (). ‘Theory of language death’, in M. Brenzinger (ed.), Language Death: Factual and theoretical explorations with special reference to East Africa. Berlin: Mouton de Gruyter, –. Thomason, S. G. (). ‘Genetic relationship and the case of Ma’a (Mbugu)’, Studies in African Linguistics : –. Thomason, S. G. (). ‘Language mixture: ordinary processes, extraordinary results’, in C. SilvaCorvalán (ed.), Spanish in Four Continents: Studies in language contact and bilingualism. Washington, DC: Georgetown University Press, –. Thomason, S. G. (). ‘Ma’a (Mbugu)’, in S. G. Thomason (ed.), Contact Languages: A wider perspective. Amsterdam and Philadelphia, PA: John Benjamins, –. Thomason, S. G., and Kaufman, T. (). Language Contact, Creolization, and Genetic Linguistics. Berkeley, CA: University of California Press. Wolff, H. E., and Alidou, O. (). ‘On the non-linear ancestry of Tasawaq (Niger), or: how “mixed” can a language be?’, in D. Nurse (ed.), Historical Language Contact in Africa (Special issue of Sprache und Geschichte in Afrika. Vol. /). Cologne: Rüdiger Köppe, –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
......................................................................................................................
Somos una gente Rooted in the memory of Mandinga, Mende and Kush We push forward, upward like branches, Demanding recognition, remembrance We are Nubia’s sunkissed descendants Speaking languages built on ancient civilizations Whose lyric, scripted in Olmec monuments, teach us Never to forget the mother tongue that birthed us (Sherehe Yamaisha Roze )1
. I
.................................................................................................................................. T African continent, like other parts of the world, has been characterized by migration and mobility for millennia. This involves not only intra-African mobility but also the migration of Africans to other parts of the world, which constitutes a major research theme from human prehistory down to the present day. Africans have migrated (and continue to migrate) voluntarily and involuntarily for different reasons and with various (linguistic and other) backgrounds to different places and with various objectives, as individuals, families, workers, bonded laborers, students, refugees, etc. In other words, the term “African Diaspora” refers to very complex and multifaceted phenomena relating to migration and displacement. A prototypical and quite well-studied African Diaspora emerged from the forced migration of enslaved Africans “through the middle passage” to the Americas (see Gilroy or Lovejoy and Trotman ). Many studies and popular articles concerned with the African Diaspora address the Black societies in North and South America and the This epigraph is an excerpt from Shereha Yamaisha Roze’s () poem “Somos una gente”, http:// www.weboaal.com/africandiaspora.htm accessed October , . 1
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Caribbean. Similarly, the slave trade also brought many enslaved Africans to the Indian Ocean (De Silva Jayasuriya and Pankhurst ). However, there are more dimensions to the African Diaspora (beyond the scope of the slave trade) and many diaspora communities in various parts of the world, from Lusophone Africans in Macau (see Bodomo and Teixeira-E-Silva ) to Zimbabweans in the United Kingdom. Africans and people of African descent are present in virtually all parts of the world, and African languages are spoken in various African communities in the Diaspora, such as Swahili spoken by Kenyans in Cologne (Germany) or West African languages in certain districts of Brussels (Belgium) or Paris (France). Furthermore, phenomena described under the umbrella of language contact play an important role with regard to Africans and their linguistic practices abroad. African languages and linguistic heritage linger on in contact languages, lexical Africanisms, calques, grammatical structures, gestures, cultural conceptualizations, ways of speaking, etc. This has, most prominently, been investigated with regard to the phenomena of “pidginization” and “creolization” in societies characterized by unequal power relations and bonded labor (prototypically plantation societies) which emerged through imperialism, colonial expansion, and the slave trade, especially in the contexts of the Atlantic and Indian Oceans (see, for instance, Parkvall ; De Silva Jayasuriya and Pankhurst ; Michaelis ; Bartens and Baker ). However, DeGraff () and other scholars have deconstructed the concept of “creole exceptionalism” and have shown that certain linguistic phenomena (including contact-induced change) also occur in other linguistic settings (see also Ansaldo and Matthews ). In other words, there are linguistic contact phenomena involving African languages in the Diaspora beyond what has classically been described as “creolization”, for example with regard to African youth languages (see Stein-Kanjora ) and in urban multilingual communication (see Blommaert ). Within the scope of this short chapter, it is not possible to do justice to the complex and numerous linguistic practices and processes that affect and characterize African languages in the large and multifaceted Diaspora. Rather, this chapter will give an overview of various phenomena involving languages and communicative practices in the Diaspora and exemplify some of the processes, features, and contact scenarios by using examples from different African diasporic communities. Looking at African languages in the Diaspora should also involve taking into consideration language attitudes and ideologies, metalinguistic knowledge, and the role of language as an emblematic marker of identity and cultural heritage, because these aspects may significantly influence the way speakers in diasporic communities make decisions and choices concerning their linguistic repertoire (including their use of African languages). However, within the scope of this chapter we can only address these issues very briefly. Moreover, the aspects of new media and online forms of communication in our globalized world play an important role with regard to diasporic communities. Mobile devices and the digital space not only enable faster, “smarter”, and boundless communication, but they also endow people with the possibilities of forming digital communities of practice and of representing themselves and their identity in new ways (e.g. through social media). Various scholars have stressed the important role of digital communication with regard to diaspora communities and their engagement in joint communicative (including linguistic) practices (see, e.g., Hinrichs ). Before turning to African languages in the Diaspora, the next section will briefly discuss the concept of the African Diaspora.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
. T A D
.................................................................................................................................. As implied in the introduction (section .), the concept of the African Diaspora is very complex and involves many different individuals and communities existing in different spaces and times. Large parts of the African Diaspora are the result of displacement through slave trade, especially in the Atlantic and Indian Oceans. The involuntary migration of Africans to the Americas brought about by the Transatlantic Slave Trade can be regarded as the first modern mass migration and accounted for more than ten million enslaved Africans leaving their homelands between the sixteenth and the nineteenth centuries (see, e.g., Lovejoy and Trotman ). The African Diaspora in the Indian Ocean comprises descendants of enslaved Africans, as well as voluntary migrants, and exhibits an enormous time-depth, with various waves of enslaved Africans being taken to the Indian Ocean region from the fourth century onwards (De Silva Jayasuriya and Pankhurst : ). Furthermore, the African Diaspora is also constituted by communities in other parts of the world which are not (or not exclusively) rooted in the slave trade but which also experience a form of displacement due to migration. This includes, for instance African diasporic communities in North America, Europe, China, and Australia. The histories underlying these various communities are very different in terms of their time-depth, their political, cultural, and social contexts and conditions (with regard to “home” as well as host society), and their size. The different ways of speaking and communicating in the African Diaspora reveal the role of African languages in manifold ways: the use of African languages in the Diaspora and their adaptation to the various sociolinguistic contexts, the emergence of new contact languages which show influences of African languages, contact-induced change on the levels of phonology, morphology, syntax, semantics, and pragmatics. Before looking at these phenomena in more detail, this section discusses perspectives on the African Diaspora which try to connect the diverse histories and developments of African diasporic societies. The term “diaspora” is of Greek origin and relates to experiences of scattering and dispersion. It has been employed mainly for displaced communities—in most cases “religious” or “ethnic”—living in a host society. In this sense, it has mainly been used for larger dispersed communities that migrated involuntarily, such as the Jewish Diaspora or the African Diaspora in the New World which emerged through the Transatlantic Slave Trade. However, the term is also applied to various immigrant communities which seek to maintain a relationship with their home society (see, for instance, Anderson ; Powers ). In efforts to account for the versatile phenomena associated with the term “Diaspora”, James Clifford (: ) “discusses problems of defining a traveling term, in changing global conditions” and concludes that “it is not possible to define ‘diaspora’ sharply, either by recourse to essential features or to privative oppositions. But it is possible to perceive a loosely coherent, adaptive constellation of responses to dwelling-in-displacement” (p. ). An important aspect in this regard is constituted by the efforts of displaced people(s) to maintain, revive, or invent connections with their homeland, or, as Clifford puts it: “The empowering paradox of diaspora is that dwelling here assumes a solidarity and connection
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
there. But there is not necessarily a single place or an exclusivist nation” (p. ). This last aspect plays an important role with regard to the African Diaspora and displacement experiences, which do not always (as in the case of the slave trade) make a direct tracing of the “original homeland” possible. As Edward Chamberlin () illustrates, in order to remember and connect to a (real or imagined) homeland, people often use powerful stories or narratives which retell the history of displacement, remember the experiences of a shared home, and authenticate the connection. Language thus plays an important role with regard to the narratives of a shared home and the expression of cultural identity in diasporas, as it serves highly representative functions. Representation and identity are also investigated by Stuart Hall (), who views these concepts as being closely connected to the formation of cultural identity, and argues that identity construction always happens within representation. Moreover, Hall reflects on the complex processes of identity formation in diaspora communities, where people seek to maintain ties with their “homeland” while being aware of belonging to different places, and where the “interlocking histories and cultures” inevitably lead to processes of “translation”; hybridity, in this sense, is not merely a mix of different cultural and linguistic practices, but the emergence of something new (Hall : ; see also Bhabha ). This also becomes evident when looking at African languages in the Diaspora and the manifold ways in which they are translated and transformed, which reveal various instances and processes of language contact.
. L A D
.................................................................................................................................. In the various and complex scenarios of the African Diaspora, African languages live on in multiple ways. The spread of African languages beyond the African continent yields many forms of transformation and language contact. The emergence of contact languages, in particular, constitutes a much-studied phenomenon. Most prominently, many contact languages that emerged in the African Diaspora in the Americas and the Indian Ocean world have been described as creole languages. The study of creole languages (which also exist in other parts of the world like the Pacific Ocean region and the African continent; with regard to the latter, see Sommer, chapter of this volume) culminated in the formation of creole studies as a branch of linguistics and an academic discipline.2 Prototypical creole languages emerged in plantation societies during European imperialism and slavery3 and are usually based on a European “lexifier language” which often also constitutes the “superstrate language” (i.e. the language of
2
However, it is important to note that the concept of a creole language is not unproblematic, and the idea of “creole exceptionalism” that emerged from studies on creole languages has received much criticism (see, e.g., DeGraff ; Ansaldo and Matthews ; Dimmendaal ; Hollington ). 3 However, there are many so-called creole languages which have emerged in other sociocultural contexts.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
prestige and power), while African languages are referred to as “substrate languages” which influence the emerging contact variety.4 Scholars have applied different approaches in order to investigate the extent of substrate influence. One influential theory in this regard, which made a strong claim for the influence of substrate languages, is the relexification hypothesis, according to which (the phonological forms of) lexical items of the “substrate language” are substituted with those of the target language (the “lexifier” or “superstrate”; cf. Muysken ). An important proponent of this theory and its further developments is Claire Lefebvre, who researched Fongbe influences in Haitian (e.g. Lefebvre , ; see also the critique by DeGraff ). There is a body of literature on African influences in contact languages of the African Diaspora; and Africanisms5 have been described at various levels of language and communication: phonology and suprasegmentals, morphology, syntax, semantics, and pragmatics. To review this literature in detail goes beyond the scope of this chapter, but we will look at some examples which illustrate the influences of African languages in a range of contact varieties. Since much of the literature on African influences in contact languages deals with the Atlantic World, the following examples concentrate on that area. Mikael Parkvall’s () book provides a compact overview of the topic and considers many contact languages of the Americas. On the level of phonology, African influences have been suggested to account for a range of features in the these contact varieties, among them, for instance, the presence of high nasal vowels, coarticulated stops, prenasalized stops, and fricatives (Parkvall ). With regard to phonotactics, scholars have discussed the preference for a CV syllable structure observable in many contact languages of the Atlantic world and also common in NigerCongo languages (Parkvall ). Furthermore, some researchers have argued for (traces of ) vowel harmony as an African linguistic feature in the New World (Berry ; see also Devonish , ). The grammatical structure constitutes a rich field with regard to the study of linguistic Africanisms. Many features of the grammar of contact languages have been discussed with regard to African influences, for instance reflexive constructions involving a word for ‘body’ or a human body part, the use of the third person plural pronoun to mark nominal plural, and certain cases of reduplication (Parkvall ; for reduplication, see also Kouwenberg and La Charité , ). Furthermore, syntactic strategies such as serial verb constructions (SVCs) and predicate clefting (fronting) as a focus construction in the contact languages reveal strong parallels to African languages. The following examples illustrate SVCs in Caribbean and West African languages. SVCs are based on a grammatical strategy where two or more verbs are used to form a single predicate. SVCs are very common in West African Niger-Congo languages but do not exist in the European languages involved in the contact scenarios of the African Diaspora in the New World.
4 For more discussion of these concepts and for various perspectives on (African) substrate influence, see Muysken and Smith () and Michaelis (). 5 Especially with regard to the Atlantic world; see, for instance, Mufwene ().
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
()
Li pran crab ale nan mache. take crab go in market ‘He brought the crab to the market.’
[Haitian] (adapted from Lefebvre : )
()
Di man tek Marcia go a farin. [Jamaican] man take Marcia go foreign ‘The man took Marcia abroad./The man went abroad with Marcia.’ (Hollington : )
()
É sɔ ́ àsɔ́n yì àxì mὲ. take crab go market in ‘He brought the crab to the market.’
()
[Fongbe] (adapted from Lefebvre : )
Wó-tsɔnε yi Keta. one-take.him go Keta ‘One carries him to Keta.’
[Ewe] (Westermann : )
The lexicon of contact languages of the African Diaspora has attracted the attention of scholars for a long time and many studies dedicated to the examination of lexical Africanisms discuss issues of etymology (see, e.g., the various contributions in Bartens and Baker ). It has been stated that the lexicon is the level of languages which is most easily affected by language contact and outside influences (Thomason and Kaufman ; Bartens ). Philip Baker investigated lexical Africanisms in French-based varieties in the Atlantic and Indian Oceans (e.g. Baker , and his contributions in Bartens and Baker ); Farquharson () studied African contributions to the Jamaican lexicon; Lipski () and Bartens () discuss possible Africanisms in the lexicon of Latin American Spanish and Portuguese, to name just a few examples. With regard to Mauritian, Baker suggests eastern Bantu languages, Mande, Wolof, Gbe, Yoruba, and Malagasy as sources of African lexemes (Baker ). ()
buduf ‘a fat, flabby person’
()
kalipa ‘wrestler; a well-built male’
Vw), and a rule changing final /n/ into /m/ (see Sharif ). As Greenberg pointed out (, ), Arabic loanwords in turn turned into “travel” words, spread not by Arabs but rather via the agents of the culture of Islam. For instance, since the Kanuri have been a key ethnic group in Islamic scholarship, much of the Arabic-derived vocabulary in the central Saharan and sub-Saharan region spread outwards from Kanem-Bornu. 2
As well as Chad, though here which variety is official is ambiguous (Pommerol ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
At the same time, if the non-lexical structural influence from spoken Arabic on subSaharan African languages has been slight, it is almost non-existent as far as Classical Arabic goes. This, it should be noted, is hardly surprising. Even in Arabic countries structural change is driven by contact between spoken varieties, not by influence from the standard language (Al-Wer ). Still, underlying the significant lexical influence from Classical Arabic is another important domain of influence, namely that until the advent of European languages in the nineteenth century, Arabic served as the main language of written communication in Islamic areas of Africa. The vast majority of texts which have come down to us relate to religious matters (e.g. O’Fahey ; Hunwick ) and closely follow models imported from the heartland of the Arabic-Islamic world in the Middle East, and more importantly as far as West Africa goes, from North Africa. Still, there are significant documents which describe local affairs. Certainly, the most original one is that of Ibn Furtu, a Kanuri, who wrote two detailed accounts of the sixteenth-century Bornu Mai Idris, one describing his attacks on Kanem to the east of Lake Chad, the other his attacks on peoples to the west of the lake (Lange ). Beyond these there are important lists of kings (e.g. Kano Chronicle), as well as various other documents, for instance the writings of the Kanembu-Arab scholargeneral Al-Kanemi in the mid-nineteenth century and in general the writings of the Nigerian Jihadists (the Bello dynasty). The transmission of Arabic in African societies has traditionally taken place via a long process of rote learning of the Qur’an, a process which can take up to fifteen or twenty years. Upon memorizing the Qur’an, students can graduate to the traditional Islamic sciences. This system of Qur’anic learning continues to command a high degree of respect in African Islamic societies (Owens a), even if in recent years there has been a movement to integrate the learning of Arabic into schools which are based on national educational curricular standard (e.g. the so-called Islamiyya schools in Nigeria; Pfennig ). Arabic has also reached Africa via migrations from Asia. Beyond the secret languages mentioned in section ... above, Muslim immigrants from India and Pakistan have brought Arabic with them, and in some instances Arabic has become an important emblem of communal identity (e.g. the Gujurati community in South Africa; Jeppie ).
.. Writing Another significant influence of the classical language on African languages has been in the use of the Arabic script to write African languages, the so-called ajami (foreign) script (Mumin and Versteegh ). Perhaps the earliest extant documents in which the Arabic script is used for writing an African language come from Berber (thirteenth/fourteenth century).3 In the seventeenth century a significant tradition of Qur’anic interpretation (tafsiyr) developed among the Kanembu in which the Arabic script was used to write Kanembu commentary (Bondarev ), and from the eighteenth century the Arabic script was used extensively for Swahili (Büttner ; Miehe ). The Arabic script has been used to a significant extent to write at least the following African languages (Daniels :
3
That is, the original manuscript.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
): ‘Afar, Afrikaans, Amharic, Argobba, Bambara, Beja, Berber (different varieties), Chimwiini, Dagbani, Djoula, Fulfulde, Gbanyito, Gonja, Harari, Hausa, Kanuri, Kotokoli, Maba, Makua, Malagasy, Mamprule, Mande, Nubian, Nupe, Oromo, Serer, Silt’e, Somali, Songhay, Soninke, Swahili, Tigrinya, Wolof, Yoruba, Zerma. Today the use of ajami scripts has receded drastically. Even during their era of ascendancy they tended to be marked by a considerable degree of internal variation (see, for instance, many remarks in Allen ). One significant indirect exception is the development of the Nko script for Mandinka. While this is not an ajami script, as it does not use Arabic letters, the inspiration for its developer, Suleyman Kante, a scholar learned in the Arabic sciences, was clearly the Arabic script. Nko ingeniously refunctionalizes constitutive elements of Arabic writing, like using short vowel diacritics to represent different tones (Oyler ).
. C
.................................................................................................................................. Although Arabic is the largest language on the African continent, and the most variegated given its different dialectal variants, historical offshoots in Creole Arabic, and the special religious-cultural status of Classical/Standard Arabic, it is nonetheless one which has been unfortunately neglected by specialists in African languages themselves. Recent “comprehensive” studies of para- or secret languages (Mous and Kießling ), or language contact in West Africa (Lüpke ), for instance, neglect Arabic in their critical frame. Generously, this may be because Arabic alone is too large a language, and one with its own “orientalist” tradition, to deal with among African languages. However, one suspects that equally important is the fact that in the West, Arabic has traditionally belonged to the Middle East and to the orientalists, making it off limits to Africanist traditions. The significant exceptions to this scholarly bureaucratization, such as is in evidence in the work of Greenberg, should encourage a closer integration of Arabic into Africanist linguistic traditions.
R Allen, J. (). ‘Arabic script for students of Swahili’, Tanganyika Notes and Records, Supplement to Volume . Dar-es-Salaam. Al-Wer, E. (). Sociolinguistics’, in J. Owens (ed.), –. Bassiouney, R. (). Function of Code Switching in Egypt. Leiden: Brill. Bassiouney, R. (). ‘Identity and code-choice in the speech of educated women and men in Egypt: evidence from talk shows’, in R. Bassiouney (ed.), Arabic and the Media. Leiden: Brill, –. Beaujard, P. (). Le parler secret arabico-malgache du sud-est de Madagascar. Paris: L’Harmattan. Behnstedt, P., and Woidich, M. (–). Die ägyptisch-arabischen Dialekte. vols. Wiesbaden: Dr. Ludwig Reichert. Bentahila, A. (). Language Attitudes among Arabic–French Bilinguals in Morocco. Clevedon: Multilingual Matters. Bentahila, A., and Davies, E. E. (). ‘The syntax of Arabic–French code-switching’, Lingua : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Bentahila, A., and Davies, E. E. (). ‘Code-switching and language dominance’, in R. J. Harris (ed.), Cognitive Processing in Bilinguals. Amsterdam: Elsevier, –. Bentahila, A., and Davies, E. E. (). ‘Patterns of code-switching and patterns of language contact’, Lingua : –. Bentahila, A., and Davies, E. E. (). ‘Code-switching: an unequal partnership?’, in R. Jacobson (ed.), Code-switching Worldwide. Berlin and New York: Mouton de Gruyter, –. Bentahila, A., Davies, E. E., and Owens, J. (). ‘Code-switching’, in J. Owens (ed.), –. Berjaoui, N. (). ‘Parlers secrets d’El-Jadida: notes préliminaires’, Estudios de Dialectología Norteafricana y Andalusí : –. Bondarev, D. (). ‘An archaic form of Kanuri/Kanembu: a translation tool for Qur’anic studies’, Journal of Qur’anic Studies : –. Büttner, C. (). Suaheli-Schriftstücke. Berlin: Seminar für orientalische Sprachen. Chebchoub, F. (). A Sociolinguistic Study of the Use of Arabic and French in Algiers. Ph.D. thesis. University of Edinburgh. Chtatou, M. (). ‘The influence of the Berber language on Moroccan Arabic’, Journal of the Sociology of Language : –. Daniels, P. (). ‘The Arabic writing system’, in J. Owens (ed.), –. El-Dash, L., and Tucker, R. (). ‘Subjective reactions to various speech styles in Egypt’, International Journal of the Sociology of Language : –. Ferguson, C. (). ‘The Arabic koine’, Language : –. Fischer, W., and Jastrow, O. (eds.). (). Handbuch der arabischen Dialekte. Wiesbaden: Otto Harrassowitz. Greenberg, J. H. (). ‘Arabic loanwords in Hausa’, Word : –. Greenberg, J. H. (). ‘Linguistic evidence for the influence of the Kanuri on the Hausa’, Journal of African History : –. Hachimi, A. (). ‘Becoming Casablancan: Fessis in Casablanca as a case study’, in Miller et al. (eds.), –. Heath, J. (). From Code-switching to Borrowing: A case study of Moroccan Arabic. London and New York: Kegan Paul. Heath, J. (). Jewish and Muslim Dialects of Moroccan Arabic. London: Routledge Curzon. Hunwick, J. (). Arabic Literature of Africa. Vol. : The Writings of Western Sudanic Africa. Leiden: Brill. Jakobi, A. (). A Fur Grammar. Hamburg: Helmut Buske. Jeppie, S. (). Language, Identity, Modernity: The Arabic study circle of Durban. Cape Town: Human Sciences Research Council Press. Khamis, C., and Owens, J. (). ‘Nubi’, in J. Holm and P. Patrick (eds.), Creole Syntax: Parallel outlines of seventeen Creole grammars. London: Battlebridge Press, –. Kossmann, M. (a). ‘Borrowing’, in J. Owens (ed.), –. Kossmann, M. (b). The Arabic Influence on Northern Berber. Leiden: Brill. Kossmann, M. (). ‘On substratum: the history of the focus marker d in Jijel Arabic (Algeria)’, in C. de Féral, M. Kossmann, and M. Tosco (eds.), In and Out of Africa: Languages in Question. In Honour of Robert Nicolaï. Vol. : Language Contact and Language Change in Africa. Paris and Leuven: Peeters, –. Lange, D. (). A Sudanic Chronicle: The Borno expeditions of Idrīs Alauma. Wiesbaden: Franz Steiner. Leslau, W. (). Arabic Loanwords in Ethiopian Semitic. Wiesbaden: Otto Harrassowitz. Luffin, X. (). Un créole arabe: Le kinubi de Mombasa, Kenya. Munich: Lincom. Luffin, X. (). ‘Bongor Arabic’, in K. Versteegh (ed.), Encyclopedia of Arabic Languages and Linguistics. Vol. . Leiden: Brill, –. Lüpke, F. (). ‘Multilingualism and language contact in West Africa: towards a holistic perspective’, Journal of Language Contact : –. Mahmud, U. (). Variation in the Aspectual System of Juba Arabic. Ph.D. thesis. Washington, DC: Georgetown University.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Mahmud, U. (). Arabic in the Southern Sudan: History and spread of a Pidgin-Creole. Khartoum: Khartoum University Press. Manfredi, S. (). Arabi juba: Un pidgin-créole du soudan du sud. Louvain: Peeters. Marçais, P. (). Parlers arabes du Fezzân. Geneva: Droz. Mejdell, G. (). Mixed Styles in Spoken Arabic in Egypt. Leiden: Brill. Miehe, G. (). Liyongo Songs. Cologne: Rüdiger Köppe. Miller, C. (). ‘Aperçu du système verbal en Juba-Arabic’, Comptes rendus du GLECS –: –. Miller, C. (). ‘Restructuration morpho-syntaxique en Juba-arabic et ki-nubi: à propos du débat universaux/substrat et superstrat dans les études créoles’, in Matériaux Arabes et Sudarabiques . Paris: Paul Geuthner, –. Miller, C. (). ‘Between accommodation and resistance: Upper Egyptian migrants in Cairo’, Linguistics : –. Miller, C. (). ‘Introduction’, in C. Miller et al. (eds.), –. Miller, C., Caubet, D., Watson, J., and Al-Wer, E. (eds.) (). Arabic in the City. London and New York: Routledge. Mous, M., and Kießling, R. (). ‘Urban youth language in Africa’, Anthropological Linguistics : –. Mumin, M., and Versteegh, K. (). The Arabic Script in Africa. Leiden: Brill. Myers-Scotton, C. (). Contact Linguistics: Bilingual encounters and grammatical outcomes. Oxford: Oxford University Press. Nöldeke, T. (). Die semitischen Sprachen: Eine Skizze. Leipzig: Tauchnitz. O’Fahey, R. (). Arabic Literature of Africa: The writings of eastern Sudanic Africa to c.. Vol. . Leiden: Brill. Owens, J. (). A Grammar of Eastern Libyan Arabic. Wiesbaden: Otto Harrassowitz. Owens, J. (). ‘The origins of East African Nubi’, Anthropological Linguistics : –. Owens, J. (a). ‘Language in the graphics mode: Arabic among the Kanuri of Nigeria’, Language Sciences : –. Owens, J. (b). ‘Minority languages and urban norms: a case study’, Linguistics : –. Owens, J. (). ‘Idiomatic structure and the theory of genetic relationship’, Diachronica : –. Owens, J. (). ‘Arabic-based pidgins and creoles’, in S. Thomason (ed.), Contact Languages: A wider perspective. Amsterdam and Philadelphia, PA: John Benjamins, –. Owens, J. (). Neighborhood and Ancestry: Variation in the spoken Arabic of Maiduguri, Nigeria. Amsterdam and Philadelphia, PA: John Benjamins. Owens, J. (). ‘Uniformity and discontinuity: toward a characterization of speech communities’, Linguistics : –. Owens, J. (ed.) (). Arabic as a Minority Language. Berlin and New York: Mouton de Gruyter. Owens, J. (). ‘Creole Arabic: the orphan of all orphans’, Anthropological Linguistics : –. Owens, J. (). ‘Processing the world piece by piece: iconicity, lexical insertion and possessives in Nigerian Arabic code-switching’, Language Variation and Change : –. Owens, J. (a). ‘Hierarchicalized matrices: code-switching among urban Nigerian Arabs’, Linguistics : –. Owens, J. (b). ‘Bare forms and lexical insertions in code-switching: a processing-based account’, Bilingualism: Language and Cognition : –. Owens, J. (). ‘Close encounters of a different kind: two types of insertion in Nigerian Arabic code-switching’, in C. Miller et al. (eds.), –. Owens, J. (). A Linguistic History of Arabic, nd edn. Oxford: Oxford University Press. Owens, J. (ed.) (). The Oxford Handbook of Arabic Linguistics. Oxford: Oxford University Press. Owens, J. (a). ‘The origins of Nubi morphology’, Journal of Pidgin and Creole Linguistics : –. Owens, J. (b). ‘Many heads are better than one: the spread of motivated opacity via contact’, Linguistics : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Owens, J. (). ‘Equilibrium, punctuation, dia-planar diffusion: Towards understanding early Aramaic - Arabic contact’, Al-Qanṭara : –. Owens, J., and Brahimi, F. (). ‘Language legitimization: Arabic in multiethnic contexts’, in J. Owens (ed.), –. Owens, J., and Hassan, J. (). ‘Making a fish of a friend. Waris: the secret language of Arabic Koran school students in Borno’, in J. Owens (ed.), –. Oyler, D. W. (). The History of the N’ko Alphabet and its Role in Mandé Transnational Identity. Cherry Hill, NJ: Africana Homestead Legacy. Pasch, H., and Thelwall, R. (). ‘Losses and innovations in Nubi’, in P. Maurer and T. Stolz (eds.), Varia Creolica. Bochum: Brockmeyer, –. Pfennig, C. (). Arabisch an modernen islamischen Schulen: Eine soziolinguistische Untersuchung in Maiduguri. Ph.D. thesis. Bayreuth University. Pommerol, J. de. (). l’arabe tchadien: émergence d’une langue véhiculaire. Paris: L’Harmattan. Reichmuth, S. (). Der arabische Dialekt der Šukriyya im Ostsudan. Hildesheim: Georg Olms. Retsö, J. (). ‘Kaskasa, t-passives and the dialect geography of ancient Arabia’, Oriente Moderno (new series) : –. Sharif, A. (). Loanwords in Kanuri: A Descriptive Analysis. Ph.D. thesis. University of Maiduguri. Taine-Cheikh, C. (). Dictionnaire zénaga–français. Cologne: Rüdiger Köppe. Tosco, M. (). ‘A Pidgin verbal system: the case of Juba Arabic’, Anthropological Linguistics : –. Tosco, M., and Manfredi, S. (). ‘Pidgins and creoles’, in J. Owens (ed.), –. Tosco, M., and Owens, J. (). ‘Turku: a descriptive and comparative study’, Sprache und Geschichte in Afrika : –. Wolfer, C. (). ‘Arabic secret languages’, Folia Orientalia : –. Ziamari, K. (). ‘Moroccan Arabic–French code-switching and information structure’, in J. Owens and A. Elgibali (eds.), Information Structure in Spoken Arabic. London and New York: Routledge, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. T past century has seen huge strides in the development of orthographies for previously unwritten languages worldwide, and Africa has been in the forefront, with hundreds of new orthographies emerging in what have traditionally been overwhelmingly oral cultures.1 In general, the process began first among the large, more dominant language groups of several million speakers (Swahili, Yoruba, Xhosa, etc.) and proceeded to mid-sized language groups. By now, no language group is deemed too small to have its own orthography. This chapter uses the term “orthography” for standardized as well as non-standardized forms, and treats standardization as a social process. Interestingly, definitions of “language standardization” which make reference to the process by which one variety becomes accepted as the norm sometimes neglect to mention specific decision-makers, change agents, or stakeholders (e.g. Ferguson : ; Wardhaugh : ). In contrast, this chapter will place people at the heart of orthography development, acceptance, and implementation. First, we address the question of whether African languages are sufficiently vital to invest in development of orthographies for them at all (section .); then we consider orthography design challenges that are specific to African languages (section .). We investigate four motivations that have driven standardization efforts over the last twenty-five years (section .) and sketch three social models of orthography development on the broader canvas of language planning (section .). We review orthography standardization practices across Africa (section .) and illustrate these with two case studies (section .). We conclude with a discussion of best practices in orthography standardization (section .).
1 According to one source, out of , living, oral African languages, , (%) are “known to be written” (though not necessarily standardized), and (%) are “known to be unwritten”. Information is lacking for the remaining (%) (https://scriptsource.org/cms/scripts/page.php?item_id=entry_detail&uid=wekytddkkc accessed Dec. , ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. L A
.................................................................................................................................. A fundamental question that needs answering first is whether orthography development is really worth the effort in Africa. Are her minority languages really capable of resisting the worldwide tendency to be overwhelmed by more powerful languages? If not, developing orthographies for them could be a lost cause. In fact, language vitality statistics reveal that Africa is faring relatively well: while more than % of languages in the Americas and more than % of those in the rest of the world are reported as dying or in trouble,2 less than % of Africa’s languages fall into these categories. Apparently, motivation to abandon vernaculars is not as rampant in Africa as elsewhere. However, encouraging statistics cannot be interpreted as indicators of positive language attitudes. The Sango orthography was standardized by government decree in and six years later a constitutional change accorded Sango the status of official language alongside French in the Central African Republic (Koyt ). Yet the attitude that “Sango is for illiterates” persisted (Karan : ). Research in Ghana and Burkina Faso indicates that although citizens may want the status of their local language enhanced, resistance to its use as medium of instruction in school is common (Beyogle ). Since relatively few of Africa’s languages currently have a standardized orthography,3 language vitality clearly does not depend on orthography standardization. Some even argue that standardization can actually work against revitalization goals (Bourcier ). However, there are many cases where the use of the mother tongue in the written domain has helped prevent both language loss and negative attitudes.
. D A
.................................................................................................................................. Largely for historical and political reasons, the great majority of emerging orthographies in Africa use Roman script. But adapting a script that was originally created for European
2 Eberhard et al. () classify a language as “dying” if the childbearing generation is no longer able to transmit the language to the next generation, since the only fluent users (if any remain) are above that age. This category covers those labeled “moribund”, “nearly extinct”, and “dormant” on the Expanded Graded Intergenerational Disruption Scale (EGIDS; Lewis and Simons ). Languages classified as “in trouble” are those labeled “threatened” (b) or “shifting” () on the EGIDS. In these cases, intergenerational transmission is possible, since the childbearing generation can still use the language but may be choosing not to. 3 Precise figures are hard to pin down, but Eberhard et al. () list African languages (.%) with the status “Institutional” (implying that their orthographies are probably standardized), and (.%) with the status “Developing” (implying that their orthographies may or may not yet be standardized). This suggests that as many as , (.%) probably do not have a standardized orthography.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
languages to the diverse sounds and structures of Africa’s ,+ languages presents considerable challenges. In general, the representation of consonants is the least problematic. For phonemes beyond the usual alphabet inventory, five strategies are common:4 () employ an otherwise unused character from the Latin alphabet (e.g. to represent the voiceless alveolo-palatal affricate /tɕ͡ / in Bambara, Mali; to represent the apical click /ǃ/ in Sotho, South Africa; to represent the voiceless uvular fricative /χ/ in Susu, Guinea); () recruit a new symbol (e.g. to represent the retroflex stop /ʈ/ in Kabiye, Togo; to represent the voiceless dental fricative /Ɵ/ in Mankanya, Guinea-Bissau); () use digraphs (e.g. to represent the voiced dental fricative /ð/ in Wambo, Angola) or trigraphs (e.g. to represent the voiced aveolo-palatal affricate /dʒ͡ / in Venda, South Africa); () use double letters—a subset of digraphs—(e.g. to represent the voiced glottal fricative /ɦ/ in Northern Ndebele, Zimbabwe; to represent the alveolar and velar ejectives /t’ k’/, respectively, in Komo and Gwama, Ethiopia); () use diacritics to differentiate between similar distinctive sounds (e.g. to set apart the voiced alveolar stop /d/ from its dental counterpart /d̪/ in Herero, Namibia; to distinguish between the voiceless alveolar sibilant /s/ and its retroflex counterpart /ʂ/ in Ronga, Mozambique). In addition, the dieresis is used in some Francophone settings, following its use in the French orthography, as an anticoagulant to distinguish digraphs from sequences (e.g. to set apart the voiced labio-velar stop /ɡb͡ / from the velar and labial stop sequence /gb/ in Nawdm, Togo; to differentiate the palatal nasal /ɲ/ from the alveolar nasal and palatal approximant sequence /ny/ in Tem, Togo). Special characters are often employed for vowels as well (e.g. to represent the high central vowels /ɨ, ʉ/, respectively, in Nzakara, Central African Republic; to represent the mid central unrounded vowel /ə/ in Ewondo, Cameroon). The use of digraphs such as to represent the open vowels /æ ɒ/, respectively, in Eastern Dan, Côte d’Ivoire is less common, and is usually avoided in languages with diphthongs.5 Vocalic elision is often marked with an apostrophe. Contrastive vowel length is ubiquitously represented by doubling the letter (e.g. to represent /a: e: i: o: u:/, respectively, in Bemba, Zambia, and many others). Vowel harmony is invariably written transparently, with each allomorph spelled as pronounced. A worrying number of African orthographies underrepresent vowels. A typical example is Tswana (Botswana), where the grapheme represents both /e/ and /ɪ/, and the grapheme represents both /o/ and /ʊ/. When such underrepresentation occurs, it may be due to the phonological contrasts not having been correctly identified—vowel Many of the examples in this section find their source in Omniglot, the online encyclopedia of writing systems and languages (Ager ). We cite only the principal country in which the language is spoken. 5 Since writing this chapter, the Eastern Dan orthography has undergone extensive reform. The two phonemes in question are now written as < æ ɶ> (Roberts et al. ; Vydrin et al. ). 4
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
contrasts are often less easily perceived than consonants—or to the pervasive influence of a prestigious orthography that has only five vowel symbols (e.g. English and Swahili in East Africa). But underrepresentation may also have been adopted deliberately on the grounds that an orthography devoid of special characters is more maniable for computer and smartphone users. Most African languages are tonal, and it is perhaps here that the greatest challenges emerge.6 The most widespread practice is to mark tone with diacritics (e.g. H, M, L level tones; HL, LH contours). Floating tones are sometimes marked by apostrophes, as in a proposition for the Bamanankan definite article (Konta and Vydrin : ). But there are divergent local traditions. In the Central African Republic, indicates H tone, M tone, and —absence of an accent—L tone. Certain Mande, Kru, and Kwa languages of Côte d’Ivoire use word-initial and -final punctuation to indicate tone (ʺ◌ extra-H; ‘◌ H; ◌ M; -vL; =◌ xL; ‑◌ʺL—H contour; ʺ◌‑xH-L contour, or variations of this basic inventory to meet particular needs; ILA ). Ghana is by no means the only country in which one legacy of English education is that orthography developers resist marking tone at all (Cahill ). Deciding how to represent tone is particularly challenging when diacritics are already being used on vowels for other purposes, whether it is the tilde (e.g. to represent the nasal vowels /ã ẽ õ ũ/, respectively, in Umbundu, Angola), or the dieresis (e.g. to represent breathy vowels /a̤ e̤ o̤/ respectively, in Dinka, South Sudan). Some orthographies handle this by placing in subscript position the non-tonal diacritics, such as pointing ( to mark [–ATR] vowels /ε ɔ/, respectively, in Yoruba, Nigeria) or the tilde (e.g. to mark nasal vowels /ĩ ε̃ ã ɔ̃ ũ/, respectively, in Nateni and Mbelime, Benin). Other authors have documented tone orthography strategies in more detail than is possible here (e.g. Kutsch Lojenga ; Bird ), but there are still few uncontested guidelines and many outstanding questions. One thing that has become increasingly clear is that the lexical and grammatical functional load of tone varies enormously between languages, so it is unwise to generalize about whether to mark tone and if so how much. Word break issues tend to be underreported and underrated. Kutsch Lojenga () is one of the few contributions to provide practical guidelines. The author identifies three criteria—grammatical, phonological and semantic—for assessing where to place word breaks. Any survey of the challenges inherent in African orthography development must also pay tribute to the numerous indigenous non-Roman script orthographies which are the subject of Rovenchak and Buk, chapter of this volume. As orthography developers grapple with the numerous challenges enumerated above, they would do well to pay attention to the increasing body of research on these scripts to see what lessons can be learned from their effectiveness (or lack thereof ). Finally, if orthography design could simply be equated with the kind of technical considerations summarized above, it would be a “task that could be achieved easily by trained linguists in much less time and for much lower cost than generally assumed by the
6
In this paragraph, the following abbreviations are used to refer to tones: xH = extra-high; H = high; M = mid; L = low; xL = extra-low.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
uninformed public” (Wolff : ). But, as Wolff goes on to explain, technical decisions are only a small part of orthography development. The rest of this chapter is dedicated to exploring the lion’s share of the process.
. M
.................................................................................................................................. What motivates individuals and communities to strive for orthography standardization? In general, there is perhaps a natural desire, at least for speakers of prestigious languages, to conform to what is normative, i.e. for orthographies to “have prescriptive power—to be standardized and authoritative” (Jaffe : ). Furthermore, governments are often motivated to standardize orthographies out of a concern for promoting national unity and for effective management of limited resources (Bergman : v). But four other additional factors have emerged over the last twenty-five years that have contributed to a sense of urgency for standardization, and it is to these we turn now: • • • •
the movement to save endangered languages (section ..); campaigns for linguistic rights and cultural heritage (section ..); efforts to provide universal quality primary education (section ..); the increased pace of Bible translation (section ..).
.. Endangered languages The realization that a large percentage of the world’s languages may die by the end of the twenty-first century if nothing is done to save them7 has led to the growth of language documentation as an entirely new subdomain of linguistics (see Essegbey, chapter of this volume). Even though language death is less of an immediate concern in Africa than elsewhere, as noted in section ., this new discipline is nevertheless making important contributions to orthography development on the continent. For documentationalists, detailed phonetic transcriptions of oral recordings are paramount, and they usually use the International Phonetic Alphabet (IPA) for this purpose. Even if they employ a more phonemic transcription for dictionaries and grammars, such notations serve only as tools in their race against time; acceptance by the speakers and government approval are not of primary concern. However, many documentationalists consider orthography development to be part of their job descriptions (Grenoble and Whaley ; Seifart ; Lüpke ). This takes skill and sensitivity, since criteria other than linguistic exactitude—pedagogical, psycholinguistic, and sociolinguistic—come into play. The orthography needs to be teacher- and user-friendly, and needs community ownership to enhance motivation to learn and use it.
Recent estimates report that % of the world’s languages are likely to die out (Wiecha ). Earlier estimates were much higher. 7
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Mosel (: ) suggests that the speech community, under such conditions, “decides on their own norms”. However, experience indicates that this is not a given; the speech community is not always consulted on orthography design or reform issues.
.. Linguistic rights and cultural heritage The past generation has seen a growing international awareness that freedom of expression, as a basic human right, is indissociable from the language of the person claiming that right. If a society does not value the first language of one of its members, that person’s entire cultural identity is eroded, especially if another, more powerful language is imposed in its place. International campaigns have emphasized first languages, reminding us that they are the primary vessel for the preservation of an individual’s cultural heritage, indigenous knowledge, artisan skills, and local technology. Such campaigns have also stressed the need to preserve linguistic diversity in the face of globalization. All these factors create a favorable environment for literacy in minority languages, and thence orthography standardization (Skutnabb-Kangas et al. ; UNESCO ).
.. Universal quality primary education Despite the multilingual context, most African schools use only an international language or a dominant national language for instruction. Children who do not know this language when they begin school are at a great disadvantage. Advocates are campaigning for the use of the primary language for the first few years of schooling with scheduled, gradual introduction of the other language(s), in order to enhance achievement and reduce dropout rates (Bender et al. : ), not to mention making education relevant. The effectiveness of the mother tongue in formal education is clearly supported by scientific research (Skutnabb-Kangas and Heugh ), and some ministries of education are listening. The emphasis on local languages for primary education creates a favorable climate for orthography development. This raises the question as to how far along in the standardization process a language needs to be before it can be used for formal education. A written school standard is definitely required, but spelling variation outside of school is not necessarily a barrier.
.. Bible translation Across sub-Saharan Africa, churches are a potent force for mother-tongue literacy. For well over a century, Bible translation agencies have been in the vanguard of orthography development. Bible translation in Africa has accelerated markedly in the last twenty-five years, as the emerging indigenous church increasingly takes responsibility for the work. And, since translation cannot begin without a written form, one outcome of the change in pace is a corresponding increase in orthography development. Religious materials are powerful influencers, inviting loyalty to a particular written code. This can lead to acceptance and standardization, particularly if various denominations agree on a single written form.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. M
.................................................................................................................................. Now let us turn to the social processes through which orthographies may be standardized. Broadly, there are three possible models: laissez-faire (section ..), top-down (section ..), and community-based (section ..).
.. Laissez-faire model Orthography standardization can come about through gradual, natural development (Crystal : ), or unforeseen circumstances, which Edwards (: ) refers to as “historical accident”. In Europe, it took centuries for written standards to evolve, with each scribe and publishing house using their own conventions, even up to years after printing had been invented (Charpentier : ). Sebba () challenges the view that a single, standardized norm is even necessary and makes no reference to standardization, norms, or correctness in his definition of orthography (Sebba : ). Stebbins (), writing about an endangered language of Canada, uses the term “orthography” in the context of multiple or competing writing systems. Essegbey (), discussing the documentation of Nyagbo, an endangered Ghana-Togo Mountain language of Ghana, advocates “orthographic transcription” in the place of standardization, allowing people to write as they speak, including colloquialisms and dialect variants. Prescriptivism is not necessarily viewed in a positive light. A spelling standard may be punitive when it cannot easily be mastered and errors result in ridicule, thus discouraging writing.8 Even the French orthography, which has a dominant place in African education and has traditionally been regarded as prestigious and inviolable, has recently been challenged by two “unofficial” adaptations beyond the African continent, Alfonic (Martinet ) and l’ortograf alternativ (Baril ). Baril even avoids using the terms “official” or “standard”, opting instead for l’orthographe traditionelle. This label diminishes the power of the standard and the ideology that there is only one way of writing a language. In any case, speakers may not appreciate uniformity or the imposition of a system by an authority.9 Research has shown that literacy skills are highly transferable from one language to another, even when symbols and conventions differ (Gardner-Chloros : ; Bernard : ; Page ). This suggests that transfer between orthographies of a single language ought not to be problematic for those whose literacy skills are well established.
8
Thonhauser () reports on Lebanese young adults’ attitudes toward writing in English, French, and Standard Arabic which they had been taught in school. They felt they had not mastered spelling in these languages; thus, their motivation to write in them was low. 9 Cassidy (: , cited in Baker : ) reports that use of the Jamaican Creole standard was limited to academic circles, since having a consistent way of writing the language was not a popular concern. Brody () indicates that Yukatec Maya speakers accept and expect graphic and word level variation in their language. Linguists’ efforts to provide a standard were resented.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Sango, the national and official language of the Central African Republic is a case in point. Although its orthography was fixed by decree in , it was not enforced: most citizens remained ignorant of its existence, and published literature continued to use a wide variety of spelling practices. Fasold (: ) critiques this laissez-faire approach, but Karan (: ) notes that, in the face of extremely limited resources, and initial negative reactions to the standard by various stakeholders, it may have been the most pragmatic way of preventing Sango literature production from coming to a complete halt. Alternating between various spelling practices certainly did not pose serious problems for readers. Although the standard orthography initially met with resistance, over time publications increasingly aligned themselves with it, with the exception of what some considered to be excessive tone notation.
.. Top-down model Orthography standardization is more often the result of a deliberate intervention by certain agents encouraging one language variety as a standard, and is presented as such in the literature (Bright : ; Mesthrie et al. : , ). This model considers that the design, implementation, standardization, and reform of an orthography are all elements in the broader domain of language planning.10 This process often involves top-down, authoritative action, usually by a government agency, with all its consequent political ramifications (Skutnabb-Kangas et al. ; Coulmas ). But governments are not the only highly placed players in the hierarchy. Linguists, as technicians with specialist knowledge and professional status, may also communicate (often unawares) unquestionable authority, especially if they act unilaterally. Whoever is responsible for such interventions, imposing a standard when it is unwanted can have the opposite effect to the one intended. Milroy and Milroy (: ) point out negative repercussions for the unchosen speech varieties, especially stigmatization of the non-standard variants. Fishman (: ix) refers to a “world full of unanticipated and often undesirable consequences”.
.. Community-based model Models of language management (Spolsky , ; Nekvapil ), in contrast, account for the larger social context, not equating potential agents with politicians or linguists only, but also individual speakers, parents, religious and educational leaders, entrepreneurs, media and communication specialists, and others who may have an influence on language choices or practices.
The term “language planning” was initially rendered as planification linguistique in French linguistic resources, but over time was replaced by politique linguistique. As sociolinguistic factors increasingly figured in discussions and the literature, the term aménagement linguistique gained popularity in the Francophone world. The work of Canadian linguist Jean-Claude Corbeil played a role in this change (Corbeil , cited in Loubier ). 10
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
This widens the circle of responsibility and opportunity: anyone with an interest in or affiliation with a speech community can potentially contribute to or impede a language’s development whether there is an explicit plan or not. Real language decisions happen in the home, the marketplace, places of worship, classrooms—wherever life happens. The community-based model does not preclude top-down decision making, but it considers community response—even if it is in the form of passive resistance or open rebellion—as a valid part of the process. There are numerous ways of involving the community in the orthography development process, including stakeholder consultations, community testing of different orthographic options (Karan , a), or even a participatory approach to primary linguistic research (Kutsch Lojenga ; Fitzgerald et al. ). When the orthography standardization process is more inclusive, there is greater likelihood that the speech community will be in favor of the decisions made, because they were involved in the process and their opinions are perceived to have carried weight. Who should make orthography decisions is not merely a practical question but also an ethical one (Eira : ), since the language belongs to the community as their birthright. Fishman (: ix) considers that the debate is not about whether to intervene, but rather how. Standardization efforts require sensitivity and wisdom, which may, in part, be acquired from past experiences—positive and negative. Whichever model is chosen, paying attention to social factors is paramount (Ferguson : ).
. O
..................................................................................................................................
.. Historical background Before the IPA was first devised and published in , Karl Richard Lepsius’s Standard Alphabet, developed in , was used for transcribing Egyptian hieroglyphics, and for a short time, to write various African languages. An adaptation by Wilhelm Schmidt was published in , which was supported and used by Carl Meinhof and Karl Endemann (Jensen : ). The Africa Alphabet was developed in under the leadership of Diedrich Westermann (IIALC ). Interest in developing a pan-African alphabet continued, and momentum increased with the independence movements. UNESCO-sponsored meetings, attended by African linguists interested in harmonization across African languages, were held in Bamako in , in Cotonou in , and in Niamey in . The African Reference Alphabet was proposed at the Niamey meeting (UNESCO ). A revised version, which eliminated uppercase letters, did not meet with acceptance (Baker : ). Participants at a follow-up meeting in Niamey in discussed principles for establishing African orthographies. However, since needs, practices, and preferences for symbolization diverge so widely, more recent harmonization efforts have focused on intra-language (section ..), language family (section ..), national (section ..), and international (section ..) levels rather than on the continental level.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Intra-language standardization Generally, it is not in a community’s best interest for a single language to employ parallel orthographies or to use different spelling conventions for various dialects; a single orthography and one set of literature would be ideal (for a counter example, see the case study in section ..). But standardization may not always be possible. Relationships between dialect communities may be less than cordial, resulting in a desire for separate identities rather than unification. Conversely, dialect communities may enjoy close social affiliation, but there may be too many phonological and lexical differences to share written materials. Citing two Ghanaian examples, Ring () illustrates that standardization efforts may not bear fruit when groups value distinctiveness or if linguistic differences are too great. In the s, speakers of the various varieties of the Dagaare-Waale-Birifor dialect continuum,11 despite a high degree of intercomprehension, resisted choosing a standard dialect, for various reasons relating to social identity. Consequently, five different primers were developed for literacy instruction. The hope remained that, once literacy skills were established, transfer to a single government-promoted standard and sharing of a single set of materials would be feasible. Recent discussions with Ring reveal that by now, three of the five groups—the central, northern, and northwestern dialects—do indeed share one orthography, which is used in literature produced by the government and the churches. Speakers of the closely related Waale and Birifor varieties, however, continue to employ distinct orthographies. In contrast, the Konkomba, despite dialect differences, had such a strong desire from the beginning to express solidarity that they opted for a single orthography and sharing of materials based on the Saboba dialect. The speakers of the Komba dialect, further removed geographically and linguistically, after initially being in favor of this decision found that the orthography did not serve them well. They later opted to make their own orthography decisions and develop their own materials, thus emphasizing their separate identity (Cahill and Karan : ). Still in Ghana, Akan unification efforts illustrate that standardization takes time. A group that was to study how three Akan dialects could be served by a unified orthography was appointed as early as , but there was little progress. The work of the official Committee of Unified Akan, set up in the mid s, came to a standstill since members had emotional attachments to their own dialects and orthographies and were not interested in compromise (Krampah and Gyekye-Aboagye ; Brock-Utne ). A new set of members made progress in the s and s, despite severe budget cuts; Akan orthographic wordlists were finally published in and spelling rules in . Intra-language standardization hardly ever comes without cost. The history of the Igbo standard orthography in Nigeria, including a thirty-two-year-long controversy, demonstrates the complications that can arise in standardization efforts, especially if there are historical, competing orthographies in addition to a complex dialect situation (Obiamalu ).
11
Ring () uses the term “Mole” for the whole continuum.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. Language family standardization The Centre for Advanced Studies of African Society (CASAS) has heavily invested in proposals for the harmonization of orthographies of languages within the same family and has published numerous monographs on the subject (e.g. for Bantu, Banda ; for Gur, Nikiéma et al. ; see also sections .. and ..). The difficulty is not generally with linguists agreeing on technical solutions; the real challenge is implementation and popular acceptance so that the work of experts does not remain a mere academic exercise (Chanda : –). Ideally, teachers and writers would be trained in workshops, literature and teaching materials would be developed, and the orthographies introduced into the formal education system (Brock-Utne : ). But this is a tall order, especially if the communities concerned were not fully involved in the decision-making process. Motivation to read and write in local languages is usually low, and introducing them into schools may require changes in language attitudes and national language policies. Orthography reforms are difficult to implement since habits are not easily changed. Indeed, some may question why family-level standardization should be a priority at all. Efforts such as the harmonized orthography for Gur languages remain largely unknown outside of academic circles, partly because the populations concerned have no motivation to transfer their mother-tongue literacy skills to another minority language within the same family. Most end users are more interested in ease of transfer to national lingua francas and international languages because they desire economic advancement and access to the wider world.
.. National harmonization Governments commonly delegate language planning responsibilities, including orthography standardization, to an official agency. Whether this functions under a government ministry or a university, a group of nationals has decision-making power on language issues. National and foreign linguists usually work under or in collaboration with such agencies and abide by their guidelines, such as using a national alphabet or harmonizing symbol choices with those of a dominant regional language. How active such an agency is and how much control it exerts may depend on how well resourced it is. Which agency is in charge may change over time. For instance, in Cameroon, CREA (Centre de Recherches et d’Études Anthropologiques) was a powerhouse in the s, strictly enforcing adherence to the General Alphabet of Cameroon Languages (Tadadjeu and Sadembouo ) for all new orthographies, although the standard had never been made official (Bird : ). When CREA ceased to exist in the early s, NACALCO (National Association of Cameroon Language Committees) spearheaded language development work but did not exert such tight controls. The standard alphabet came to be regarded as a set of flexible guidelines rather than law. A similar pragmatic strategy is also operational in Benin, where the national alphabet (Igué and Capo ) gets periodically reviewed through a process of inclusive consultation and modified on the basis of new linguistic findings (see CENALA for the most recent version).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.. International harmonization Another major challenge for harmonization is that, since national boundaries in Africa were not drawn along linguistic lines, standardization efforts at family and national level are more often than not faced with cross-border issues. Standardization efforts for Swahili began in , when an Inter-territorial Language Committee, with representatives from Kenya, Tanganyika (now Tanzania), Uganda, and Zanzibar, was given the dual responsibility of devising a standard orthography and unifying the use of new and existing words (Temu : ). Writing Swahili with Arabic script also continued, but certain limitations kept the system from gaining wider usage (Kaye : ). Two competing Roman script Hausa orthographies, one used in Nigeria and the other in Niger, were successfully harmonized at a meeting in Niamey in . The revised system now serves citizens in both nations (Wolff : ), notwithstanding Warren Rothlin’s (: ) concern that there is still room for a greater degree of standardization. The progress is particularly noteworthy since Nigeria’s official international language of education is English, while that of Niger is French. Efforts to harmonize Yoruba in Nigeria and Benin have not been as successful. Yoruba is one of the oldest Roman script orthographies in Africa (Crowther ), so writing traditions are strongly engrained. In particular, there is an attachment to the use of subscript pointing to represent [–ATR] vowels (see section .), and this is reinforced by similar conventions in other Nigerian languages. So when, much more recently, the Beninese Yoruba adopted special characters to represent [–ATR] vowels, this convention was rejected by the Nigerian Yoruba (Bada ). Since only a small fraction of Yoruba speakers reside in Benin, Oyetade () appeals for them to compromise on two vowel symbols, and for the Nigerian contingent to change one, but there is as yet no evidence that this proposal has won acceptance. After harmonization efforts of Akan languages had been effective in Ghana (see section ..), meetings were held in Côte d’Ivoire in and to extend the unified orthography to Akan communities there. Three action points are noteworthy: First, attendees encouraged the Ivorian contingent to do further research and test whether three of the vowels represented in their orthography were really phonemic or whether they could be eliminated. Second, the group proposed that palatalized sounds should be written with digraphs rather than with the Ivorian convention of single letters respectively. Third, on the matter of underdifferentiation of nasal vowels, which was the practice in Ghana, it was left up to the Ivorian group to decide whether to continue marking nasalization with , as they had been doing, or adopt the Ghanaian Akan convention of non-representation (Kouadio et al. ). Finally, we should note that whatever the scope of the standardization envisaged—be it at intra-language, family, national, or international level—there is a major difference between, on the one hand, merely agreeing to share a common character inventory and, on the other hand, the actual standardization of an orthography (i.e. shared characters, shared spelling rules, and shared literature). The former is far easier to achieve than the latter.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. C
.................................................................................................................................. We now illustrate standardization processes with two more detailed case studies: Ik, a Ugandan language (section ..), and Aja, a language spoken in Togo and Benin (section ..).
.. Ik (Uganda) Ik is a Nilo-Saharan language spoken in Uganda by about , people, otherwise known as Icetod. It illustrates that first attempts at developing an orthography might not bring results that ought to be promoted as a standard, that parallel orthographies can serve two different purposes, and that the process of standardization should not be hurried (Karan b). Over several decades, a series of linguists did fieldwork on Ik, culminating in the publication of an Ik-English dictionary (Heine ), but no orthography was ever popularized. A local language advocate designed an alphabet in , which was presented and considered at a orthography workshop attended by eight Ik community members who had a personal interest in writing their language (Schrock ). The workshop was facilitated by a linguist who was sensitive to the predominance of sociolinguistic issues. Language representatives, feeling that the Ik had been habitually marginalized, wanted their orthography to reflect their unique ethnic identity. They thus preferred unique symbols over those of neighboring languages. One of the participants, a local writer, used the resulting trial orthography for several years, but it did not gain widespread usage. There were aesthetic issues: four of the graphemes—some of them high-frequency—consisted of letter–apostrophe combinations, making written texts look cluttered. Consequently, some writers left out the apostrophes, resulting in serious underrepresentation. A new trial alphabet, an adaptation of the orthography used in the dictionary, was tested in . Special characters were prominent, and four symbols, representing generational pronunciation differences, were replaced by a single letter. But only a few writers ever used this alphabet. [–ATR] vowels were not represented in the or the orthographies, and neither became popular. Then some externally funded interventions aroused an interest in additional Ik orthography discussions. In , UNESCO awarded a grant to the “Ik Agenda Development Initiative” (IADI). The aim was to preserve the Ik’s intangible cultural heritage by documenting cultural knowledge and practices. This required transcribing audio and video recordings, creating the need to write Ik consistently. Additionally, USAID had been funding a “School Health and Reading Program” (SHARP), which mostly benefitted speakers of Uganda’s more dominant languages. A special workshop was held in to evaluate the orthographies used in SHARP projects. Motivated by possible inclusion in future development interventions, and the desire to be better prepared for the UNESCO project tasks, IADI and SIL-Uganda held a one-month Ik orthography workshop with dual objectives: to raise linguistic awareness among the participants, and to agree on a standardized form of written Ik.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
The outcome of the workshop was surprising: there was agreement among participants to promote not one standard but two, to cater for differing audiences, objectives, and needs. Because previous orthographies had posed difficulties for those who had attempted to read and write Ik, it was felt that for popular literacy to take root, writing needed to be as simple as possible. Thus, a popular orthography was designed which marks neither tone (because the functional load of tone was not considered to be sufficiently heavy) nor the difference between [+ATR] and [–ATR] vowels (because vowel harmony and assimilation rules obscure the difference and the average native speaker could not distinguish them). But the participants also agreed that there should be a way to record Ik more accurately in various resources for linguistic purposes, language learning, and language documentation. Thus, a linguistic orthography was designed which does mark tone, symbolizes all vowels, and uses a deeper representation when underlying forms and surface forms differ. Both systems have flexibility built into them. Schrock (: ) states that the Ik orthography is still developing, and believes that forcing a standardized system on people early on would be counterproductive and discourage writing. It is expected that through practice, over time, a single orthography may emerge. For now, dual orthographies serve different users, purposes, and audiences.
.. Aja (Benin and Togo) The second case study focuses on Aja,12 a Gbe language that straddles the border between Benin and Togo, yet another legacy of the linguistic naivety of the Berlin conference delegates. Such international divides, extremely common in Africa, often result in parallel orthographies, because one group of stakeholders is ignorant of the other’s existence or because of the sheer logistical difficulties of arranging cross-border meetings or as a valid response to sociolinguistic issues that are present on one side of the border but absent on the other. The Aja experience illustrates how these challenges can be overcome through sensitive and timely intervention. Preceding standardization, there were numerous individual and largely uncoordinated initiatives, such as a two-volume Aja–French dictionary (Harguindéguy ) and a descriptive grammar (Tchitchi ). In Benin, the government produced reading instruction books for use in government-run literacy programs (e.g. DAPR n.d.), while in Togo, the Southern Baptist Convention published their own (Abalo et al. ). A sociolinguistic survey conducted in (Kluge and Tompkins ) took account of language use and attitudes in both Togo and Benin. It identified five Aja varieties: Sipki, and Dogbo in Benin; Tala and Tado in Togo, and Hwe in both countries. In addition, Capo () in collaboration with CASAS (Capo et al. ), was promoting an entirely separate initiative. The Gbe Uniform Standard Orthography (GUSO), he hoped, would serve the needs of all Gbe languages, including Aja. By its nature, GUSO aims to cater for both Togo and Benin, but it is not well known among the Aja and has not gained widespread acceptance. Capo (: )—whose approach is based primarily on linguistic analysis rather than sociolinguistic and historical realities—admits that his proposals 12
We would like to thank Joshua Ham and Terry Sullivan for furnishing most of the information in this case study.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
remain “laboratory products”, and that there is resistance to reform due to attachment to existing systems. At more or less the same time, there was a growing feeling in the Aja community that the time was ripe for a coordinated and inclusive approach. In , the Comité International du Suivi de l’Orthographe de la Langue Aja (CISOLA) was formed. The word “International” deliberately stresses the cross-border profile of the language community, a gesture toward the Togolese, who might easily feel marginalized, since Azove, the major market and crossroad, is in Benin, and most CISOLA meetings were held in Benin. Also, while representatives of the Beninese government were present at CISOLA meetings, their Togolese counterparts, constrained by a less inclusive language policy (Afeli ), were absent. In addition to government representatives, CISOLA was composed of delegates from Catholic and Protestant denominations, and from all the NGOs involved in literacy. The committee held an initial orthography forum (CISOLA ), and met regularly thereafter to fine-tune the orthography. The results of oral reading and writing tests illuminated their discussions. CISOLA showed no interest in embracing Capo’s plea for language family standardization. Little by little the orthography crystallized, to the point where a guide for those wishing to transition from French to Aja (Beavon-Ham and Ega ), and a new bilingual dictionary (Atchi et al. ) could be published. The whole process spanned a decade and culminated in a ‘ratification forum’ where decisions, some of which already reflected existing practice, were formally adopted and made public (CISOLA a, b). The process involved compromises. The Togolese Aja wanted to spell nasal vowels with the tilde (e.g. [akã] ‘charcoal’), following the practice in Ewe, a related Gbe language which functions as a vehicular language for the south of Togo, and also has a much longer written tradition. The Beninese Aja, on the other hand, freer from the influence of Ewe, almost all wanted to represent nasal vowels with a vowel–nasal sequence (e.g. [akã]), to facilitate transition between Aja and French, the official language of both countries. This issue was settled by majority vote, with the latter convention winning. The border divide was also evident in the question of how to write palatalized and labialized consonants. Should the verb ‘to ask’ be written or ? Should the noun ‘syringe’ be written or ? The Togolese tended to write vowels, while the Beninese preferred consonants. However, the committee members, even those who were Beninese, opted almost unanimously for using vowel letters. So two issues that could have proved contentious were amicably resolved. This was facilitated by the two forums and the committee meetings which gave everyone the opportunity to air their views and be heard. There was also a sense of seizing the moment: stakeholders had come to a point where they were willing to put old allegiances and prejudices aside and come to firm decisions that would benefit everyone. The Aja standard orthography is a success at one level because CISOLA achieved bilateral consensus across an international border. But the real measure of the orthography’s success will be its use. There is a distinctive local tradition among taxi motorbike drivers of displaying slogans in Aja on the back of their vehicles. This is a visible barometer of literacy practice, revealing how writers choose to spell their language spontaneously. So far there is no indication that the taxi motorbike drivers are even aware of the
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
standard. If ever they do choose to conform to it, it will be a small indication that CISOLA’s standardization work has truly taken root.
. C:
.................................................................................................................................. This chapter has demonstrated the diversity of orthography development in Africa, in terms of design challenges, motivations, models, and practices. By way of conclusion, we summarize each section and highlight what we consider to be key points and best practices. We began by observing that Africa’s languages are considerably more resistant to the forces of globalization than those on other continents, and that because the vast majority are not immediately endangered, orthography development is a worthwhile investment. We discussed design challenges specific to Africa, highlighting five widespread strategies for representing consonants, and commenting that underrepresentation of vowels is common. We also noted that many questions remain unresolved with regard to the representation of tone, and that word break issues are often underestimated and underreported. We concluded this overview, however, by arguing that linguistic decisions should be considered only a small part of the task. We proceeded to identify four motivations for orthography standardization that have emerged in the last twenty-five years: the endangered language movement, campaigns for linguistic rights, the drive for universal quality primary education, and the increased pace of Bible translation. Contemporary orthography development in Africa would do well to harness the potential of these motivations for the benefit of the many languages that do not yet have a written form. We explored three models of orthography standardization, noting that inclusive, community-based decision-making is almost always more successful than a laissez-faire or a top-down approach, but that the laissez-faire model may be appropriate in some circumstances. The desired outcome is that the community would use the orthography, and the best way of achieving this is for stakeholders to be involved in the decision-making process, facilitated by supportive infrastructures. The community may participate in stakeholder consultations, orthography testing, and even in primary linguistic research. Communitybased decision-making is also a more ethical approach, since the language belongs to them by birthright. In a survey of orthography standardization practices across Africa, we identified several levels at which initiatives may take place—intra-language, language family, national, and international levels. At each level, there may be valid social reasons why communities seek cohesion or separate development, and these almost always override linguistic considerations. We also noted that, whatever the level, orthography standardization is much harder to achieve than the mere sharing of a common repertoire of symbols. Finally, we illustrated the recent process of orthography standardization in two languages. The Ik case study illustrates that initial attempts at developing an orthography might not bring results that ought to be promoted as a standard, and that parallel orthographies can serve two different purposes. The Aja case study demonstrates how cross-border issues can be successfully resolved with timely, sensitive intervention. Both case studies illustrate that
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
designing, standardizing, and implementing an orthography are best done in stages, and that there is no hurry: the process may well span years, if not decades. Once an orthography exists, it takes time and effort to implement it, and to create the conditions in which it is used, so that literacy impacts the community. An ongoing literacy program, literature production, and an increase in environmental print all contribute to conditions necessary for a society to value reading and writing. It is not the provision of a code that brings life to the written form of a language; motivation to learn it and use it will.
R Abalo, A. K. Y., Adjakossi, E. Y., Gligo, H. A., Honoungbe, S., Kedalan, G. K., Mawussi, A. F., and Sodahlon, K. M. (). Nan hlὲn àjà gbe (Syllabaire aja). Lomé: Convention Baptiste du Togo. Afeli, A. (). Politique et aménagement linguistiques au Togo: bilan et perspectives. Thèse d’État. Lomé: Université de Lomé. Ager, S. (). Omniglot: the online encyclopedia of writing systems and languages, http://www. omniglot.com/writing/index.htm accessed Dec. , . Atchi, E. A. F., Philippe K., Sodji, P. K., and Hounongbe, J. de Dieu K. (). Dictionnaire adja àjàgbe–nyɔkwiwémégã version français–adja et adja–français. Lomé: Imprimerie de la Cité. Bada, M. D. (). Règlement de la question de l’orthographe du yoruba au Bénin compte tenu de la pratique nigériane. Paper presented at the Atelier sur la révision des outils d’écritures des langues béninoises, Lokossa, Benin, May –, . Baker, P. (). ‘Developing ways of writing vernaculars: problems and solutions in a historical perspective’, in A. Tabouret-Keller et al. (eds.), –. Banda, F. (ed.) (). Language across Borders: The harmonisation and standardisation of orthographic conventions of Bantu languages within and across the borders of Malawi and Zambia. Cape Town: CASAS. Baril, D. (). ‘L’orthograf alternativ: un konpromi optimal’, iForum : , http://fracademic.com/ dic.nsf/frwiki/#La_notation_alfonic accessed Dec. , . Beavon-Ham, V., and Ega, É. (). Lire et écrire l’ajagbé: Guide pratique pour ceux qui savent lire le français. Cotonou: SIL Bénin. Bender, P., Dutcher, N., Klaus, D., Shore, J., and Tesar, C. (). In their own language: education for all. Washington, DC: World Bank, http://hdl.handle.net// accessed Dec. , . Bergman, T. G. (ed.) (). Proceedings of the Round Table on Assuring the Feasibility of Standardization within Dialect Chains, Noordwijkerhout, the Netherlands, September . Nairobi: SIL. Bernard, H. Russell (). ‘Languages and scripts in contact: historical perspectives’, in D. Wagner, R. L. Venezky, and B. V. Street (eds.), Literacy: An international handbook. Boulder, CO: Westview Press, –. Beyogle, R. (). ‘Language contact and language attitudes in two Dagara-speaking border communities in Burkina Faso and Ghana’, in E. C. Zsiga, O. T. Boyer, and R. Kramer (eds.), Languages in Africa: Multilingualism, language policy, and education. Washington, DC: Georgetown University Press, –. Bird, S. (). ‘Strategies for representing tone in African writing systems’, Written Language and Literacy : –. Bird, S. (). ‘Orthography and identity in Cameroon’, Written Language and Literacy : –. Bourcier, A. (). ‘Standardization in language revitalization’, in C. Dyck, T. Granadillo, K. Rice, and J. E. R. Labrada (eds.), Dialogue on Dialect Standardization. Newcastle upon Tyne: Cambridge Scholars Publishing. Bright, W. (). ‘Social factors in language change’, in F. Coulmas (ed.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Brock-Utne, B. (). ‘Language-in-education policies and practices in Africa with a special focus on Tanzania and South Africa: insights from research in progress’, in A. M. Y. Lin and P. W. Martin (eds.), Decolonisation, Globalisation: Language-in-education policy and practice. Clevedon: Multilingual Matters, –. Brody, M. (). ‘To the letter: a microanalysis of currently contested graphemes in the Maya of Yucatan’, in R. M. Brown (ed.), Endangered Languages and their Literatures: Proceedings of the th FEL Conference, Antigua, Guatemala, – August, . Bath: Foundation for Endangered Languages. Cahill, M. (). ‘Avoiding tone marks: a remnant of English education?’, Notes on Literacy : –. Cahill, M., and Karan, E. (). ‘Factors in designing effective orthographies for unwritten languages’, SIL Electronic Working Papers, https://www.sil.org/resources/archives/ accessed April , . Cahill, M., and Rice, K. (eds.) (). Developing Orthographies for Unwritten Languages. Dallas, TX: SIL International. Capo, H. B. C. (). ‘The pan-dialectal approach to orthographic conventions: the case of the Gbe languages of West Africa’, in K. K. Prah (ed.), Writing African: The harmonisation of orthographic conventions in African languages. Cape Town: CASAS, –. Capo, H. B. C., Bedou-Jondoh, E., Bolouvi, L.-P., Gagnon, D., Gbeto, F., Gnamiato, V., Kinhou, S.-M., Semadegbe, J., and Tohoun, B. (). L’orthographe harmonisée des langues gbé du Ghana, du Togo, du Bénin et du Nigeria. Cape Town: CASAS. Cassidy, F. G. (). ‘Short note on Creole orthography’, Journal of Pidgin and Creole Languages : –. CENALA (). Alphabet des langues nationales béninoises, th edn. Cotonou: Centre national de linguistique appliquée avec le concours de l’initiative francophone pour la formation à distance des maîtres (IFADEM). Chanda, V. M. (). ‘Orthography planning across languages and countries: some thoughts and proposals’, in F. Banda (ed.), –. Charpentier, J.-M. (). ‘Literacy in a pidgin vernacular’, in A. Tabouret-Keller et al. (eds.), –. CISOLA. (). Actes du forum sur l’orthographe de la langue aja, du au mars . Azové, Benin: Comité International du Suivi de l’Orthographe de la Langue Aja. CISOLA. (a). Guide de l’orthographe de l’ajagbe. Azové, Benin: Comité International du Suivi de l’Orthographe de la Langue Aja. CISOLA. (b). Orthographe pratique standard de l’ajagbe, st edn. Azové, Benin: Comité International du Suivi de l’Orthographe de la Langue Aja. Corbeil, J.-C. (). L’aménagement linguistique du Québec: Langue et société. Montreal: Guérin. Coulmas, F. (ed.) (). The Handbook of Sociolinguistics. Oxford: Blackwell. Coulmas, F. (). ‘The nationalization of writing’, Studies in the Linguistic Sciences : –. Crowther, S. A. (). Yoruba Grammar. London: Church Missionary Society. Crystal, D. (). A Dictionary of Linguistics and Phonetics, th edn. Oxford: Blackwell. DAPR (n.d.). Ajagbe wema ŋukɔtɔ (Livret Aja tome ). Cotonou: Direction de l’Alphabétisation du Ministère Béninois de la Culture et des Communications. Dyck, C., Granadillo, T., Rice, K., and Labrada, J. E. R. (). Dialogue on Dialect Standardization. Newcastle upon Tyne: Cambridge Scholars Publishing. Eberhard, D. M., Simons, G. F., and Fennig, C. D. (eds.) (). Ethnologue: Languages of the World, nd edn. Dallas, TX: SIL International, http://www.ethnologue.com accessed Dec. , . Edwards, J. R. (). Language and Disadvantage. London: Arnold. Egbokhare, F. O., and Oyetade, O. S. (eds.) (). Harmonisation and Standardisation of Nigerian Languages. Cape Town: CASAS. Eira, C. (). ‘Authority and discourse: towards a model of orthography selection’, Written Language and Literacy : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Essegbey, J. (). ‘Is this my language? Developing a writing system for an endangered-language community’, in J. Essegbey, B. Henderson, and F. McLaughlin (eds.), Language Documentation and Endangerment in Africa. Amsterdam: John Benjamins, –. Fasold, R. W. (). ‘Motivations and attitudes influencing vernacular literacy: four African assessments’, in A. Tabouret-Keller et al. (eds.), –. Ferguson, C. A. (). ‘Language development’, in J. A. Fishman et al. (eds.), – [Repr. in T. Huebner (ed.), –]. Fishman, J. A. (). Do Not Leave Your Language Alone: The hidden status agendas within corpus planning in language policy. Mahwah, NJ: Lawrence Erlbaum. Fishman, J. A., Ferguson, C. A., and Das Gupta, J. (eds.) (). Language Problems of Developing Nations. New York: Wiley and Sons. Fitzgerald, C., Hyslop, G., Rice, K., and Stenzel, K. (). Orthography development: the ‘midwife’ approach. Paper presented at the Institute on Collaborative Language Research, University of Kansas, June– July . Gardner-Chloros, P. (). ‘Vernacular literacy in new minority settings in Europe’, in A. TabouretKeller et al. (eds.), –. Gippert, J., Himmelmann, N. P., and Mosel, U. (eds.) (). Essentials of Language Documentation: Trends in linguistics. Berlin and New York: Mouton de Gruyter. Grenoble, L. A., and Whaley, L. J. (). ‘Orthography’, in L. A. Grenoble and L. J. Whaley, Saving Languages: An introduction to language revitalization. Cambridge: Cambridge University Press, –. Harguindéguy, R. (). Premiers éléments pour un dictionnaire adja–français. Volumes I et II. Lyon: Afrique et Langage. Heine, B. (). Ik Dictionary. Cologne: Rüdiger Köppe. Huebner, T. (ed.) (). Sociolinguistic Perspectives: Papers on language in society –. New York: Oxford University Press. Iguée, A., and Capo, H. (). ‘Benin’, in R. L. Hartell (ed.), Alphabets of Africa. Dakar: UNESCODakar Regional Office and SIL, –. IIALC (). Practical Orthography of African Languages: Memorandum . Oxford: International Institute of African Languages and Cultures and Oxford University Press. ILA (). Une orthographe pratique des langues ivoiriennes. Abidjan: Institut de linguistique appliquée, Université d’Abidjan. Jaffe, A. (). ‘Introduction: non-standard orthography and non-standard speech’, Journal of Sociolinguistics : –. Jensen, H. (). Sign, Symbol and Script: An account of man’s efforts to write, rd rev. and enlarged edn, trans. from the German by G. Unwin [Original title: Die Schrift in Vergangenheit und Gegenwart, ]. London: George Allen and Unwin. Karan, E. (). Writing System Development and Reform—A Process. MA thesis. Grand Forks, ND: University of North Dakota. Karan, E. (). The ABD of Orthography Testing: Integrating formal or informal testing into the orthography development process (Work Papers of the Summer Institute of Linguistics, University of North Dakota Session ), https://commons.und.edu/sil-work-papers/vol/iss// accessed Dec. , . Karan, E. (a). The ABD of Orthography Testing: Practical guidelines (Work Papers of the Summer Institute of Linguistics, University of North Dakota Session ), https://commons.und.edu/silwork-papers/vol/iss// accessed Dec. . Karan, E. (b). ‘Standardization: what’s the hurry?’, in M. Cahill and K. Rice (eds.), –. Kaye, A. S. (). ‘Adaptations of Arabic script’, in P. Daniels and W. Bright (eds.), The World’s Writing Systems. New York: Oxford University Press, –. Kluge, A., and Tompkins, B. (). ‘Sociolinguistic survey of the Aja language area’, SIL Electronic Survey Reports -:, https://www.sil.org/resources/archives/ accessed Dec. , .
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Konta, M., and Vydrin, V. (). ‘Propositions pour l’orthographe du bamanankan’, Mandenkan : –. Kouadio, N’Guessan J., Agyekum, K., Saah, K. K., Gyekye-Aboagye, J., Tera, K., and Kouame, D. T. (). A Unified Orthography for the Akan Languages of Ghana and Ivory Coast: General unified spelling rules. Cape Town: CASAS. Koyt, M. M. (). ‘La situation du sango en République Centrafricaine’, in H. Pasch (ed.), Sango: The national official language of the Central African Republic: Proceedings of the colloquium “The status and uses of Sango in the Central African Republic”, Cologne, September –, . Cologne: Rüdiger Köppe, –. Krampah, D. E. K., and Gyekye-Aboagye, J. (). ‘The Akan experience at harmonisation and standardisation’, in K. K. Prah (ed.), –. Kutsch Lojenga, C. (). ‘The writing and reading of tone in Bantu languages’, Notes on Literacy : –. Kutsch Lojenga, C. (). ‘Participatory research in linguistics’, Notes on Linguistics : –. Kutsch Lojenga, C. (). ‘Basic principles for establishing word boundaries’, in M. Cahill and K. Rice (eds.), –. Lewis, M. P., and Simons, G. F. (). ‘Assessing endangerment: expanding Fishman’s GIDS’, Revue Roumaine de Linguistique : –. Loubier, C. (). L’aménagement linguistique: Fondements de l’aménagement linguistique. Office de la langue française, http://linglang.uqac.ca/IMG/pdf/loubier.pdf accessed Dec. , . Lüpke, F. (). ‘Orthography development’, in P. Austin and J. Sallabank (eds.), Handbook of Endangered Languages. Cambridge: Cambridge University Press, –. Martinet, J. (). Je parle, donc j’écris, https://www.jeparledoncjecris.inforef.be/projet.html accessed Dec. , . Mesthrie, R., Swann, J., Deumert, A., and Leap, W. L. (). Introducing Sociolinguistics. Philadelphia, PA: John Benjamins. Milroy, J., and Milroy, L. (). ‘Varieties and variation’, in F. Coulmas (ed.), –. Mosel, U. (). ‘Fieldwork and community language work’, in J. Gippert et al. (eds.), –. Nekvapil, J. (). ‘From language planning to language management’, Sociolinguistica : –. Nikiéma, N., Silué, J. S., Takassi, I., Tchagbalé, Z., Kassan, B., Mawani, A. M., Atintono, A. S., and Kedrebéogo, G. (). Orthographe standard et harmonisée des langues gur. Cape Town: CASAS. Obiamalu, G. O. C. (). ‘The development of Igbo standard orthography: a historical survey’, in F. O. Egbokhare and S. O. Oyetade (eds.), –. Oyetade, S. O. (). ‘Towards the unification of orthographic practices in Defoid’, in F. O. Egbokhare and S. O. Oyetade (eds.), –. Page, C. (). ‘Biliteracy across scripts: implications for language development in Southeast Asia’, Journal of the Southeast Asian Linguistics Society : –. Prah, K. K. (ed.) (). Between Distinction and Extinction: The harmonisation and standardisation of African languages. Johannesburg: Witwatersrand University Press. Ring, J. A. (). ‘Three case studies involving dialect standardization strategies in northern Ghana’, in G. E. Kindell (ed.), Proceedings of the Summer Institute of Linguistics International Language Assessment Conference, Horsleys Green, – May . Dallas, TX: SIL International, –. Roberts, D., Basnight-Brown, D., and Vydrin V. (). ‘Marking tone with punctuation: an orthography experiment in Eastern Dan’, in Y. Haralambous (ed.), Proceedings of the Conference “Grafematik: Graphemics in the st century from graphemes to knowledge”. Brest: Fluxus, –. Schrock, T. B. (). A Guide to the Developing Orthography of Icetod. Kampala: SIL Uganda and the Ik Agenda Development Initiative, https://www.sil.org/resources/archives/ accessed Dec. , . Sebba, M. (). Spelling and Society: The culture and politics of orthography. Cambridge: Cambridge University Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Sebba, M. (). ‘Sociolinguistic approaches to writing systems research’, Writing Systems Research : –. Seifart, F. (). ‘Orthography development’, in J. Gippert et al. (eds.), –. Skutnabb-Kangas, T., and Heugh, K. (eds.) (). Multilingual Education and Sustainable Diversity Work: From periphery to center. London and New York: Routledge. Skutnabb-Kangas, T., Phillipson, R., and Rannut, M. (eds.) (). Linguistic Human Rights: Overcoming linguistic discrimination. Berlin and New York: Mouton de Gruyter. Spolsky, B. (). Language Policy. Cambridge: Cambridge University Press. Spolsky, B. (). Language Management. New York: Cambridge University Press. Stebbins, T. (). ‘Emergent spelling patterns in Sm’algyax (Tsimshian, British Columbia)’, Written Language and Literacy : –. Tabouret-Keller, A., Le Page, R., Gardner-Chloros, B. P., and Varro, G. (eds.) (). Vernacular Literacy: A re-evaluation. Oxford and New York: Clarendon and Oxford University Press. Tadadjeu, M., and Sadembouo, E. (). General Alphabet of Cameroon Languages. Bilingual version. Propelca . Yaoundé: University of Yaoundé. Tchitchi, T. Y. (). Systématique de l’ajagbé. Thèse de doctorat. Université de la Sorbonne Nouvelle Paris III. Temu, A. (). ‘The development of national language: a survey of Kiswahili in Tanzania’, in K. K. Prah (ed.), –. Thonhauser, I. (). ‘ “Written language but easily to use!” Perceptions of continuity and discontinuity between written/oral modes in the Lebanese context of biliteracy and diglossia’, Written Language and Literacy : –. UNESCO (). African languages: Proceedings of the meeting of experts on the transcription and harmonization of African languages, Niamey (Niger), July –, . Niamey: UNESCO, http:// www.bisharat.net/Documents/Niameyen.htm accessed Dec. , . UNESCO. (). Safeguarding the intangible cultural heritage, https://www.unesdoc.unesco.org/ ark://pf accessed Dec. , . Vydrin, V., Zeh, E., and Gué, N. (). Guide de nouvelle orthographe dan de l’est pour les apprenants familiers avec l’ancienne orthographe [Guide to the new Eastern Dan orthography for learners already familiar with the old orthography]. Man: EDILIS. Wardhaugh, R. (). An Introduction to Sociolinguistics, th edn. Malden, MA: Wiley-Blackwell. Warren-Rothlin, A. (). “West African scripts and Arabic-script orthographies in socio-political context”, in M. Mumin and K. Versteegh (eds.), The Arabic Script in Africa: Studies in the use of a writing system, Vol. . London: Brill, –. Leiden: Brill. Wiecha, K. (). New estimates on the rate of global language loss, http://rosettaproject.org/blog/ /mar//new-estimates-on-rate-of-language-loss accessed Dec. , . Wolff, E. (). ‘Language and society’, in B. Heine and D. Nurse (eds.), African Languages: An introduction. Cambridge: Cambridge University Press, –. Wolff, E. (). ‘Background and history: language politics and planning in Africa’, in A. Ouane and C. Glanz (eds.), Optimising Learning, Education and Publishing in Africa: The language factor. A review and analysis of theory and practice in mother-tongue and bilingual education in subSaharan Africa. Hamburg: UNESCO Institute for Lifelong Learning and Tunis Belvédère: ADEA, –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. T
.................................................................................................................................. P integrates the user perspective into the analysis of language, from the viewpoint of its design (langue) and from that of its operational mode (parole, discourse). Sharing its object of inquiry with time-honored disciplines of core linguistics such as lexicology, syntax, and semantics, it has acquired disciplinary status within the broader field of study of language since the early s, when it developed its own analytical and terminological apparatus: a methodological framework which for its heuristics, description, and explanation draws on categories and methods rooted in various disciplines ranging from the cognitive sciences (e.g. inferences) to sociology (interaction, power, politeness), ethno-methodology, anthropology, discourse analysis, and semiotics. As “users” of natural language we find ourselves in alternating roles of (co-)producers and interpreters of discourse. As a consequence of broadening the notion of “use” from speech production to processing of speech in complex interactional settings, How to Do Things with Words (Austin ) has come to mean many more things than what speech act theory based on this seminal concept originally implied: “Conversation, as opposed to monologue, offers the analyst an invaluable analytical resource: as each turn is responded to by a second, we find displayed in that second an analysis of the first by the [current] speaker” (Levinson : ; see section .. below). Suspected mismatches between producer intention and receiver’s understanding account for linguistically coded nonlinear and parallel procedures in the construction of discourse cohesion (Bearth ). The query “What did you mean?” points not so much to a “speech accident” as to metacommunication, validation, and formal closure as ingredients in human communication and in a viable model of communication. As a regulative force operating at the interface between society’s norms and language users, the pragmatic component further refers to the knowledge that native speakers interiorize to various degrees concerning the way in which semiotic resources (linguistic, gestural, mimicry) must or must not be used in a given culture or in specific recurrent
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
situations, and what their use implies in a given setting. Language taboos illustrate the negative, greetings the positive side of normative pragmatics.
. A A
.................................................................................................................................. Pragmatic knowledge as enacted through African languages is rooted in African culture, reflecting its diversity, but also its commonalities, often associated with commonplaces about, e.g., deference towards the elder (not just old people!). Systematic studies such as that on communicative behavior in West Africa by Ameka and Breedveld () or on speech acts in African languages (Sommer and Vierke ) are exceptional. Claims for African pragmatic universals, while plausible in some cases, remain, therefore, elusive. The challenge remains, on the other hand, to explore and describe this vast and diverse empirical field as a token of universal pragmatics. For heuristic purposes, two approaches can, nevertheless, be readily illustrated from existing work: observing speech in action, contextualizing it in terms of social and situational variables such as age, gender, status, and purpose of communication, and narrowing it down to specific types of acts, e.g. request (Obeng ), rebuke (Agyekum ), or apology (Obeng ), and thus do justice to Dell Hymes’ early plea for including research on speech acts in the description of the world’s languages (Hymes ). Alternatively, a pragmatic approach to language may be pursued from specific cross-sectional viewpoints (politeness: Yahya-Othman ; silence: Agyekum ), or from the perspective of oralliterary genres (proverbs: Yankah ; comics: Beck ). On a different scale, the scope of domains of inquiry may range from individual speech forms (e.g. proper names) to full-fledged conversational case studies (public meetings, health campaigns, casual exchanges). Methods of inquiry vary accordingly. Participant observation is a requirement for the study of speech in action, particularly in traditional settings such as palaver and ceremonies where highly formalized interactional routines are performed. These provide a largely untapped wealth of insights into communicative knowhow which still regulate “dos” and “don’t’s” in daily intercourse in much of Africa’s emancipated society of the twenty-first century. African orality, which includes naturalized varieties of English, French, and Portuguese, opens, beyond the anecdotic, large avenues for scientific exploration and practical applications.
. D
.................................................................................................................................. In the context of development communication, pragmatics is needed for addressing behavioral aspects of the language/knowledge interface, which in turn are indissolubly linked to progress in any field of human development. “Comprehensive knowledge”, according to a definition of UNAIDS, is interactive; it includes the ability to deal with counter-discourse and, in respect to the HIV pandemia, “involves not only accurate
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
understanding of the primary modes of HIV transmission . . . but also awareness of the inaccuracy of common misconceptions about HIV” (UNAIDS : ). AIDS communication, in facing taboo and social stigma, mobilizes a panoply of communicative resources, including theater (Chinyowa ), poetry (Mutembei ), cartoons (Beck ), songwriting (Tourneux et al. ), and popular metaphor (Sow ), thus providing an opening for applied pragmatics which, however, has to come under thorough review in order to ensure “ownership of knowledge by the entire population in its linguistic and cultural diversity” (Bearth et al. ; N’Guessan and Cissé ). The language-dependency of acquisition of the kind of articulate knowledge that is needed in increasingly complex natural and social environments makes it appear a matter of commonsense to look for communicative resources designed to address relevant issues in a meaningful and socioculturally compatible way (Nwosu ). In traditional African societies, sustainable and transmissible knowledge is not primarily owned by the individual, but is the result of a drive within communities to reach consensus. Understanding culturally embedded communicative knowledge and preferences is, therefore, essential to meet challenges relevant to development at a lower transaction cost and with a higher yield in crucial areas such as health, agricultural technology, and environmental responsibilization, as illustrated for the last domain by the Kono procedure of the Tura (Bearth and Baya ).
. P
.................................................................................................................................. Grammaticalization of pragmatic categories occurs where these interfere in a regular way with semantic contents expressed by terms or propositions, generating morphologically structured fields devoted to the expression of specific types of pragmatic meaning. Information structure, dialogue markers, social deixis, and sentence connectives exemplify domains in the scope of pragmatic operators interacting with independently meaningful lexico-grammatical structures as operands.
.. Pragmatic aspects of nominal classification In Swahili, as in many Bantu languages, nominal concord coded by class pair / singular/ plural prefixes is governed by the semantic feature of animacy inherent in the head noun, independently of its morphological class membership. By contrast, the morphological class of the head noun is the sole criterion for assignment of concord to all other noun classes (Reynolds and Eastman ). In addition, however, switching morphological class membership is exploited for producing special semantic effects. For example, classes / (li-/ma-), used derivatively, and less commonly classes / (m-/mi-), stand for augmentative, classes / (ki-/vi-) for diminutive subcategorization. Applied to human beings, dimensional subcategorization generally implies deviation from a norm. Augmentative connotation tends to be pejoratively interpreted as oversize and, by extension, as moral depravity (man ! monster). Adolescent pregnancies provoke strong censure, notably from fathers having paid fees for now “aborted” school attendance of their female offspring. This is the case of Tenge in
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Mwachofi’s play Mama ee (: ). When her brother arrives at the scene of the ordeal, unaware of her fault, he refers to his sister with class subject marker a- and asks: A-li-fanya nini? ‘She=PAST-do what?’ (‘What has she done?’). The father cuts him short, using class subject prefix li- (combined with pejorative lexicon) to denounce his misfit daughter, and the sentential negative marker si to lament his son’s failure to close the inferential gap at once (Bearth : ): Si li=mepachikwa mimba na George? ‘Can’tyou-see-that she=has-allowed-herself-to-get pregnant by George?’ Having caught up with the situation, the brother joins the father in manhandling Tenge. He reminds her that he had warned her against “that George of yours”, switching to “no-longer-human” class concords on pre- and postposed modifiers as a verbal means of throwing discredit on the lover, absent from the scene: Hi-l-o (this-NC-back_ref) George l-a-ko (NC-of-yours). Tenge’s case may be contrasted with that of a young girl moving from the countryside to the city for the first time (Mohamed : ). Upon arrival, she finds herself lodged in a brothel. Her first encounter is with a client whose physical attributes are described in augmentative terms (madevu (NC) ndevu ‘beard’, manywele (NC) nywele ‘hair’; mfua (NC) fua ‘breast’), while the man himself is ji-tu hi-li ‘giant-person this-NC; mtu huyu ‘this man’)—as much a description of the man’s imposing stature as of the overwhelming impression it leaves on the newcomer. Whereas in this case dimensionally marked class assignment is motivated by physical attributes of the character, the same is not true for Tenge or George. In their case, the derogatory effect rests on the symbolic value of the replacement of concord of class by that of class , a distinct and deliberate act of disjunction between reference to human agents— which is maintained, for they are held responsible for their deed—and the linguistic right to be identified as humans—which is denied to them. As shown above, the derogatory effect of hilo George lako is derivable via a continuum of semantic extensions. One might, therefore, ask, under Occam’s razor: What justifies its analysis as a distinctly pragmatic fact? The answer is that class agreement applied to normal-size humans is not a freely available option, but is subject to a particular type of situational constraint, the nature of which is most accurately captured in an early definition by Crystal (: ): “Pragmatics studies the factors that govern our choice of language in social interaction and the effects of our choice on others.”
.. Focus and implicature Information structure, highly relevant to this chapter, is too multifaceted to be covered extensively. For a recent survey, the reader is referred to Fiedler and Schwarz (). Inference and implicature, too, are at the heart of pragmatics, but implicatures as operands of focus have generally escaped attention (Bearth ). A short case study from Tura (Mande, Ivory Coast) will show how (i) focus marking (FOC = postclitic -le) triggers conventional implicature, (ii) assertion interacts with pragmatic presupposition, (iii) countervalue resulting from (i) and (ii) turns propositional truth into argument. Here goes the story. Tura female partners decided to use at their discretion the financial benefit they had reaped from participating in a development project sponsored by a foreign NGO. Their decision was approved by the NGO, but not by their husbands. Under agnatic law, matrimonial rights include male control over goods acquired by the female side.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Fully aware that arguing their case in these terms, notably evoking the bride price as a justification, would be counterproductive, the men (disambiguation of discourse reference) made a declaration which was uncontroversial in terms of propositional content—we married the women—but which, by virtue of its inferential pregnancy, settled the issue for ears tuned to Tura pragmatics. ()
Nɔ́ŋ-bò-le kō-le kō àŋ̀-le sī’ za le. woman-- we- we them- take- ‘The women—we married (lit. took) them.’
The sentence-final particle za conventionally entails role inversion (INV) of the FOCmarked referents in its scope, yielding a proposition constructed as hypothetical antecedent to which () is the reaction: (‘) ‘THEY (the women) married us (men) [not WE them]’. (Nɔŋ́ -bò ‘women’, left-dislocated grammatical object of sī’ ‘took (= married)’, is coded as topic of (‘/) and as referential link to the debate.) Of course, (‘) is not what anyone in his or her right mind would say; rather it is what the men submit the women were in fact saying by claiming for themselves the money they had earned. (‘) is the pragmatic presupposition, countered by (), which by its irrationality delegitimizes the claim. Worse, it exposes the claimants to ridicule. Due to its concise, proverb-like logic, the statement, never quoted without its pragmatic layer of particles and its terminal marker (TM), gained celebrity in Tura society in the late s. While never expressed in an explicit and irrevocable way for which its authors could be held accountable, its implied meaning was never in doubt.
. H
.................................................................................................................................. African languages, because of prejudice but also by virtue of pragmatic opacity—lack of easy mutual interpretability—may function as smokescreens against unwanted interference. President Mugabe’s speeches in Shona, and Radio Mille Collines during Rwanda’s genocide are examples of deliberately chosen language bias associated with distinct political agendas. Choice of the African language “as a medium of political discourse is a pragmatic element of meaning in its own right”, generalizes Blommaert (: ).
.. The wider scope of pragmatics: dual staging A comprehensive survey of pragmatic aspects of communication in sub-Saharan Africa would have to include modern and traditional media cultures and the way they interact, distinguishing between their urban globalized and rural community-centered manifestations (Beck and Wittmann ). The “syndrome of dominant language bias” (Bearth ) inherited in unbroken continuity from colonial times that dominates the producing side would have to be brought in relation with “duality of staging” on the interpretive side. The front stage/backstage metaphor (Cole n.d.) offers a useful frame for bringing into
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
perspective the interplay of these opposite principles, both on a macro- and on a micro-level (keeping in mind that what is called “backstage” and what “front stage” depends on the observer’s location). The persistence of backstage concepts and communicative routines perpetuated primarily through the vitality of African languages and the unbroken influence of the latter on present-day “front stage” discourse justifies, in our context, methodological precedence given to the former over the latter.
.. The interplay of language use and dual staging Dual staging recursively interacts with multiple language divides. The sociolinguistic complexity of sub-Saharan multilingual settings accounts for outranking of major languages by locally prestigious minor languages in given circumstances—the reverse of dominant language bias. A case in point is Kwenye Ukingo wa Thim, a play by Ephraim Hussein () featuring, still close to the events, the notorious Otieno case on conflicting burial rights. Besides generating an abundant literature in the fields of law, sociology, and Kenyan politics, the case spurred discourse analysis comparing its coverage in the English and in the Swahili press (McGarry ). Hussein’s play, written in Swahili, exhibits the tantalizing experience of failure to understand the issues around the burial case by members of the clan living in urban affluence and having lost touch with their origins. The problem is not loss of basic language competence—all actors supposedly are proficient speakers of Swahili—but loss of pragmatic competence in any African language, an urban problem. By contrast, when ominous arrangements for the burial in the village are interspersed with recitals in Dholuo, no decoding of it is required for apprehending their implied meaning. Reduced to its ritual function, the language of the village of the deceased incarnates the backstage pole of reference—the one which eventually prevails in the real story. Duality of ultimate reference is by no means always a matter of incomplete emancipation from the weight of an obsolete social order willfully perpetuated by exponents of a generation on its way out, as the Otieno case might suggest. In a fine-grained conversation-analytical study of HIV/AIDS education targeted at disadvantaged youth in a poor neighborhood in Nairobi, Beck (, ) analyzes the interactive use of a prosodic repair routine in Sheng, a slum variety of Swahili, for questioning the validity of the core message—abstinence from sex—which the prevention game in which they participate is supposed to convey. The counselees take advantage of the participative setting in order to legitimize themselves as source of ultimate backstage reference pitted against Western-dominated interpretive monopolies.
.. Dual staging in literary expression and public advertisement Public uses of African languages in competition with Europhone languages are not always and necessarily subversive. Writing in Duala, to mention just one example, puts one of its eminent poets, Francis Bebey (–), in a position that allows him to chide his audience without the risk of losing it, and even ensuring him a hearing far beyond the limits of Duala conversancy.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
What can communication practitioners learn from the axiom of dual ultimate reference? The rule seems to be this: in a non-routine African communication setting, seek to strike the same emotional and epistemic chord for front- and backstage-tuned audiences. An impressive proof of this theorem is the popularization of the condom likened to a Fulani herdsman’s hat in Niger, in a previously antagonistic sociocultural and religious environment (Sow ).
. F
.................................................................................................................................. As in other regions of the world, face management plays a major role in shaping verbal behavior. Face as a theoretical concept is essentially relational and reflexive (Brown and Levinson ): minimization of imposition (e.g. attenuation of orders or requests) and maximization of positive face (e.g. reverence for age and merit), keeping in mind that polishing the interlocutor’s face means to polish one’s own.
.. Strategies for mitigating social power Strategies for reducing imposition in interpersonal encounters are based on verbal role play, as in the following example gleaned from the novel Utubora mkulima (Robert : ). A woman landowner happens to meet the elderly female servant of a young city dweller who had set foot on her estate. Seeking to learn from her who the newcomer is, she starts her inquiry by verbally insinuating total inversion of their respective social status: “Please, order me to ask you . . . ” (Tafadhali ni-amru kuuliza). We discard the hypothesis of ostentatious politeness (which increases face-threatening effects that it pretends to reduce). Comparison with its English polite counterpart “Allow me to ask you . . . ” suggests that typologically different lexical and framing resources are mobilized for similar goals and must be learnt, or unlearnt, as is the case of periphrastic politeness markers (“Would you please . . . ”; Schicho ). However, honesty commands us not to exclude the alternative hypothesis that, in the case at hand, the higher degree of politeness is motivated by a perceived difference in age, illustrating positive politeness rather than minimization of imposition.
.. Conventional vs. conversational implicatures of dual directive repertories Politeness has been proposed as the main distinguishing feature motivating dual directive verb paradigms found throughout Bantu and Niger-Congo (compare Akan, Tura, Duala). According to common prejudice, the imperative form (in Swahili: short prefixless fanya x ‘do x!’, without overt reference to the actor) portrays the speaker as the one in control. By contrast, the prefix- and suffix-carrying form u-fany-e ‘you (or any person)-do-OPTATIVE x’ is counted as inherently “polite”. Dealing extensively with politeness strategies in Swahili, Schicho (: –) argues against the association of the difference between imperative
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
and subjunctive with politeness, dubbing it a “popular textbook theory”. But considering that the long form is obligatory in non-initial propositions of sequential commands, one may infer that its sequence-initial occurrence where it competes with the short imperative tends to be motivated by a pragmatic difference, which may (or must?) be equivalent to a symbolic transfer of control from the speaker to the addressee. This, however, is one way of reframing politeness in interactional categories. While the hypothesis of grammaticalization of politeness per se seems attractive because it provides a straightforward rationale for the pervasive duplication of directives throughout Niger-Congo verb form repertoires, cross-linguistic evidence in its support is far from homogeneous and far from conclusive (Nurse ; Devos and Van Olmen ). On the other hand, apparent inconsistencies may be induced by the degree to which the intimated action is believed by the speaker to lie in the self-interest of the addressee as perceived by him- or herself, in which case the straightforward imperative may be less threatening than the optative form which makes it conditional. The choice between fanya x and u-fany-e x will then co-vary with the degree of face-threatening or face-enhancement attributed to x. Coding of pragmatic differences in directive verb forms could still be either grammatically or pragmatically determined, with options ranging from firm correspondence to freely variable association.
.. Linguistic politeness as a factor of social cohesion A more fundamental difference in face management concerns the narrowly dyadic analysis of polite behavior, which appears to fall short of accounting correctly for much of African data. According to the predominant Western view, the primary purpose of verbal politeness strategies is minimization of face-threatening between individual actors. Ultimately, though rather indirectly, society should benefit from such rational behavior, assuming that what is best for the individual is best for society. The prototypical African setting is the reverse of this: individual verbal behavior is rated as polite to the extent that it benefits society: The significance of third party redress strategies may be related to the way relationships are perceived within the society, in which the speaker’s membership in a closely knit social network and her status within that network are more important than her individual wants. (Yahya-Othman : )
And: In maintaining and enhancing these relationships, the speaker has to produce appropriate linguistic behaviour, not only with reference to the person with whom she is interacting, but also with reference to others with whom both the speaker and the addressee are socially linked. (p. )
See also Nwoye ().
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. A
.................................................................................................................................. Address and greeting formulas constitute an integrated paradigm of conceptually heterogeneous components co-occurring in a pragmatically determined syntactic environment which is activated in initiating or closing an encounter, or, simply, for confirming existing social structure on a day-by-day basis. By means of greetings, society is constructed and reconstructed. A recurrent grievance against widespread land grab in rural western Ivory Coast criticizes the occupants for their failure to conform to this basic attribute of humanness: “They do not greet.”
.. Enhancing the social value of positive face In at least some languages, address terms and formulas are the preferred locus for encoding positive or negative values of face. Enhancing the interlocutor’s face is by no means limited to traits attributed to a person’s individual sphere. Individuals are honored on account of their relation to some other individual; the latter’s prominence is recognized and reflects on the former’s face, thus multiplying the added social value of an encounter or even a casual exchange of greetings. Ingeniously playing out transitivity of kinship categories, an elderly Tura woman refers to her granddaughter as “my good mother” on account of the homonymy between her (the old woman’s) mother, and the granddaughter’s deceased mother (the old woman’s daughter). The context in which this address formula occurs is clearly honorific.
.. Strategic omission of politeness markers By contrast, replacing an expected relational term by a generic, relationally underspecified term may serve as a deliberate strategy aimed at marginalization or even social ostracism. The chairwoman of a village women’s association suspected of mismanagement gets herself addressed as lóoán é ‘old-woman this-here’ instead of “my mother” or “my aunt”, as would otherwise have been the case (Vé Kouadio : ). This amounts to “bald on record” denial of face prerogatives.
.. Pragmatics as a module of language description To conclude, let us ask: Should observations of this kind be integrated into linguistic descriptions? Can a grammar which ignores them squarely be called “comprehensive”? Greetings and address formulas extrapolating kinship relations are just as much aspects of communicative competence as discriminating between two alternative directive forms. Furthermore, the recognition of language as dialogue-driven activity entails minimally that at least some attention be paid to grammatical phenomena specific to non-initial speech acts, to implicature derived through them, and to pragmatic conditions of their use.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
R Agyekum, K. (). ‘The pragmatics of DUABO “grievance imprecation”: taboo among the Akan’, Pragmatics : –. Agyekum, K. (). ‘The communicative role of silence in Akan’, Pragmatics : –. Ameka, F. K., and Breedveld, A. (). ‘Areal cultural scripts for social interaction in West African communities’, Intercultural Pragmatics /: –. Austin, J. L. (). How to Do Things with Words? Oxford: Oxford University Press. Bearth, T. (). ‘Inferential and counter-inferential markers in Swahili dialogue’, Afrikanistische Arbeitspapiere (= Swahili Forum IV) : –. Bearth, T. (). ‘Countervalue’, in J. Gippert, M. Erdal, and R. Vossen (eds.), Sentence Types and Sentence Structures (= Sprachtypologie und Universalienforschung ). Berlin: Akademie Verlag, –. Bearth, T. (). ‘Language and sustainability’, in R. M. Beck (ed.), The Role of Languages for Development in Africa: Micro and macro perspectives (= Frankfurter Afrikanistische Blätter ). Cologne: Rüdiger Köppe, –. Bearth, T., and Baya, J. (). ‘Guerre civile et résilience écologique: le cas du Parc National du Mont Sangbé à l’ouest de la Côte d’Ivoire’, Cahiers d’études et de recherches francophones/Cahiers agricultures : –. DOI:./agr... Bearth, T., Cissé, G., Jolivet, R., and Singy, P. (). HIV/AIDS Communication facing African multilingulism/Communication du VIH/SIDA face au multilinguisme africain, http://www.csrs. ch/atelierlausanne/ accessed August , . Beck, R. M. (). ‘Popular media for HIV/AIDS prevention? Comparing two comics: Kingo and the Sara communication initiative’, Journal of Modern African Studies : –. Beck, R. M. (). ‘ “Tusidanganyane—Machen wir uns doch nichts vor!” Wissensproduktion und HIV-Prävention in Nairobi (Kenia)’, Gesprächsforschung—Online-Zeitschrift zur verbalen Interaktion : –, http://www.gespraechsforschung-online.de/fileadmin/dateien/heft/ga-beck.pdf accessed August , . Beck, R. M. (). “Tusidanganyane—Voyons un peu, les gars! Production des savoirs et prévention du VIH/SIDA à Nairobi (Kenya)’, in T. Bearth, G. Cissé, R. Jolivet, and P. Singy (eds.), Communication du VIH/SIDA face au multilinguisme africain: Atelier de recherche. Université de Lausanne, – nov. , Communications no a, https://www.csrs.ch/atelierlausanne/com/a_Beck_ Handout.pdf accessed September , . Beck, R. M., and Wittmann, F. (eds.) (). African Media Cultures—Transdisciplinary Perspectives. Cologne: Rüdiger Köppe. Blommaert, J. (). ‘Ethnocoherence and the analysis of Swahili political style: steps towards a method’, Afrikanistische Arbeitspapiere : –. Brown, P., and Levinson, S. C. (). Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Chinyowa, K. (). ‘Why theatre? A theoretical view of its centrality in HIV/AIDS communication’, Trans Internet-Zeitschrift für Kulturwissenschaften July . Wien: Educult, http://www. inst.at/trans/Nr/_/chinyowa.htm accessed August , . Cole, N. L. (n.d.). ‘Goffman’s front stage and back stage behavior’, ThoughtCo., https://www. thoughtco.com/goffmans-front-stage-and-back-stage-behavior- accessed September , . Crystal, D. (). Language. Cambridge: Cambridge University Press. Devos, M., and Van Olmen, D. (). ‘Describing and explaining the variation of Bantu imperatives and prohibitives’, Studies in Language : –. Fiedler, I., and Schwarz, A. (eds.) (). The Expression of Information Structure: A documentation of its diversity across Africa. Amsterdam and Philadelphia, PA: John Benjamins.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Hussein, E. (). Kwenye Ukingo wa Thim. Nairobi: Oxford University Press. Hymes, D. (). ‘The ethnography of speaking’, in T. Gladwin and W. C. Sturtevant (eds.), Anthropology and Human Behavior. Washington, DC: Anthropological Society of Washington. Levinson, S. C. (). Pragmatics. Cambridge: Cambridge University Press. McGarry, R. G. A. (). Cross-Linguistic Discourse Analysis Model for Evaluating Interethnic Conflict in the Press. The Otieni case. Boone, NC: Parkway. Mohamed, S. A. (). Utengano. Nairobi: Longman. Mutembei, A. K. (). Ukimwi katika Fasihi ya Kiswahili, – (= AIDS in Swahili Literature). Dar es Salaam: TATAKI. Mwachofi, A. K. (). Mama ee. Nairobi: Heinemann. N’Guessan, T. S., and Cissé, G. (). ‘Équité et aspects socioculturels de la communication sur le VIH/SIDA en Côte d’Ivoire’, in T. Bearth, G. Cissé, R. Jolivet, and P. Singy (eds.), Communication du VIH/SIDA face au multilinguisme africain: Atelier de recherche. Université de Lausanne, – nov. , Communications no , https://www.csrs.ch/atelierlausanne/communication.php accessed September , . Nurse, D. (). Tense and Aspect in Bantu. Oxford: Oxford University Press. Nwosu, I. (). ‘Communication et promotion de l’environnement en Afrique’, in H. Koné and J. H. Sy (eds.), La communication pour le développement durable en Afrique. Abidjan: Presses Universitaires de Côte d’Ivoire, –. Nwoye, O. G. (). ‘Linguistic politeness and sociocultural variation of the notion of face’, Journal of Pragmatics : –. Obeng, S. G. (). ‘Verbal indirection in Akan informal discourse’, Journal of Pragmatics : –. Obeng, S. G. (). ‘Apologies in Akan discourse’, Journal of Pragmatics : –. Reynolds, K. H., and Eastman, C. M. (). ‘Morphologically based agreement in Swahili’, Studies in African Linguistics : –. Robert, Shaaban bin. (). Utubora Mkulima. London, Nairobi, and Ibadan: Evans Brothers. Schicho, W. (). ‘ “Ist es unhöflich, mit Worten sparsam zu sein?” Überlegungen zur interkulturellen Begegnung Deutsch-Swahili’, Afrikanistische Arbeitspapiere : –. Sommer, G., and Vierke, C. (eds.) (). Speech Acts and Speech Events in African Languages. Cologne: Rüdiger Köppe. Sow, S. (). ‘Foula, le chapeau de la noblesse: approches linguistiques et culturelles de la lutte contre le sida au Niger’, in T. Bearth, G. Cissé, R. Jolivet, and P. Singy (eds.), Communication du VIH/ SIDA face au multilinguisme africain: Atelier de recherche. Université de Lausanne, – nov. , Communications no a/b, https://www.csrs.ch/atelierlausanne/com/a_Sow.pdf, https:// www.csrs.ch/atelierlausanne/com/b_Sow.pdf accessed September , . Tourneux, H., Boubakary, A., and Hadidja, K. (). ‘Le “baptême” peul du VIH: Cérémonie en quatre journées’, in H. Tourneux and L. Métangmo-Tatou (eds.), Parler du sida au NordCameroun, nd rev. edn. Paris: Karthala, –. UNAIDS. (). AIDS epidemic update. December . New York: UNAIDS/WHO. Vé Kouadio, L. (). ‘Colère des femmes et silence des hommes: une analyse du discours des femmes de W. sur la gestion de leur décortiqueuse’, in T. Bearth (ed.), Dynamiques du genre: Le cas toura. Stratégies de survie en temps de crise. Abidjan: Éditions Livres Sud, –. Yahya-Othman, S. (). ‘Covering one’s social back: politeness among the Swahili’, Text : –. Yankah, K. (). The Proverb in the Context of Akan Rhetoric. Bern: Peter Lang.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. A the beginning of the third millennium, Africa is experiencing an immense growth in the use of information and communications technologies (ICT).1 According to Internet World Stats (June ), the number of Internet users in Africa is still well below that of the rest of the world,2 but mobile telephone technology, in extremely rapid expansion on the continent, seems to overcome economic and technological constraints to access the World Wide Web. Mobile-centric Internet access has until recently been more important not only in Africa but in the entire global South than in the global North; the primary way of accessing the Internet is via so-called feature phones (Deumert : ). Mobisites, websites designed for use on mobile devices, have led to an “explosion of written interaction” among South African urban teens (Walton : v). It is thus suggested that mobile phones are given priority when it comes to the development of the Internet in Africa (IRIN ). Research on digital media in Africa has tended to focus on social and economic development issues and on how different groups use the technologies, but lately, studies of the use of African languages on the Internet and in text messages have also appeared. Since electronic communication tends to be more inspired by spoken language than other written communication (e.g. Warschauer ), and as it is associated with linguistic freedom and diversity (Androutsopoulos ), African languages are increasingly used
1
ICT is used here in the narrower sense of computers, the Internet, and mobile telephony. In June , about .% of the population on the African continent were using the Internet, as compared to .% in the rest of the world (Internet World Stats ) (.% and % according to the International Telecommunication Union). 2
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
in ICT. The benefits, or the possible disadvantages, of ICT for the African languages are debated, while efforts are put into technological developments to enhance the use of these languages in electronic communication. In this regard, there are two different types of initiatives that need to be taken into account. The first type builds on standard language and (usually) a monolingual perspective: in order to make the technology more accessible, African languages are used as interface languages and for promotional purposes, and technology is developed to reach this goal. The second type is what we could call a grassroots literacy (Blommaert ) that stems from the users’ (and texters’) own needs and motivations. This literacy is often multilingual and does not necessarily follow standard orthography. Writing in African languages on the Internet and in text messages may have important consequences for these languages, especially for their value and role as written languages, since they are usually associated with oral communication.
. B ?
.................................................................................................................................. It is currently common to consider the use of African languages in ICT as a way of bridging the digital divide (e.g. Kamau ) and electronic communication is regarded as a potential channel for the revitalization of minority and dominated languages. The introduction of African languages in ICT is regarded as vital not only for the languages themselves, but also for democratic development through the spread of information and knowledge: ICT in African languages could be important in post-literacy and in dissemination and generation of knowledge . . . reliance almost exclusively on English, French, and Portuguese for the transmission of information and new knowledge puts people who are not skilled in these languages, and arguably the societies of which they are a part, at a disadvantage. The issue certainly goes further, as it is legitimate to ask what sort of future there is for languages that are not used actively in ICT. (Osborn : )
Creating Internet content in African local languages is, moreover, considered to contribute to the expansion of computer-mediated communication in sub-Saharan Africa and to make it more valuable for people: Such content would increase the relevance and value of the available information to users, enhance their literacy skills, and bridge the gap between those who are literate in English or other European languages and those who are literate in ethnic languages. (Gyamfi : )
Studies from other continents show how languages that have been oppressed by the colonial authorities are valorized on the World Wide Web (e.g. Kitalong and Kitalong ; Warschauer ). The Internet challenges the monolingual perspective on written language and makes linguistic diversity more visible (Androutsopoulos ); hence, ICT is considered as a potential vehicle for the blooming of African languages, on the condition
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
that political action is undertaken (Camara et al. ; Onguene Essono ), and/or through the public’s use of their first language (L) on the Internet and in text messages (Deumert and Masinyana ; Lexander ). More pessimistic voices (e.g. Chaudenson ; Paolillo ) claim that the Internet only increases the already existing digital divide between English (and French) on the one hand and African languages on the other: If current trends are representative, the center of gravity of the Internet will remain in North America and Europe for the foreseeable future. Africa and most of its linguistic diversity is likely to remain out of reach of the Internet. (Paolillo : )
Chaudenson (: –) fears that the computerization of African languages will lead to an exaggeratedly complex transcription, making correct spelling of the languages difficult. Because of the orthographic incoherence found in texts written in African languages, Chaudenson doubts that these data can be processed on computers. He even goes as far as to ask whether it is useful at all to put such texts online. Internet users’ practices have, however, shown that Africans find it useful to write in African languages online and that correct spelling is subordinate. There have also been technological innovations facilitating the use of African languages on the Internet, most importantly the development of Unicode.
. T
.................................................................................................................................. Many consider the localization of software and web-interfaces as necessary for the expansion of African languages on the Internet (Osborn ). The emergence of Unicode, enabling a single software product or a single website to be targeted across multiple platforms, languages, and countries, has indeed facilitated multilingualism in ICT. Efforts aiming at the translation of terminology are thus emerging from national institutions, nongovernmental organizations (NGOs) and commercial actors; software in African languages has even been produced through cooperation between companies and universities. Free software, like the navigator Mozilla Firefox and other Linux programs, has been and is being translated into different African languages through nonprofit initiatives. Swahili, the dominant sub-Saharan language on the Web (Van der Veken and de Schryver ), was the first African language used by Microsoft to present its Windows and Office programs, in (BBC News ). Other languages have followed; software in Afrikaans, Amharic, Hausa, Igbo, Rwanda, Sotho, Tigrinya, Tswana, Wolof, Xhosa, Yoruba, and Zulu is now available (Microsoft ). However, translation of terminology is not always consistent: the translations of specific lexical units differ from Linux to Microsoft (Kamau ). When using Google Search, the most-used search engine on the Web, more than thirty African languages are available as interface languages. Google is also working towards statistically based translation into a hundred African languages, but the lack of written material in many of the languages causes problems, as large databases are needed (Ledgard : ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Many international web portals are multilingual, some also in African languages: Internet news from the BBC can be read in Hausa, Rwanda/Rundi, Swahili, and Somali. A growing number of more local portals exist in different African languages; the first two portals in Swahili were launched in (BBC News ). These commercial initiatives show that there is public interest in African languages as interface languages—at least in a small percentage of them. One of the often mentioned obstacles for the use of African languages in electronic communication concerns technological difficulties related to writing systems. Since African languages, like many other languages in the world, contain diacritic and other graphic signs that are nonexistent in English, French, and other early computerized languages, adapted keyboards are needed. Keyboards developed for African languages will motivate the use of these languages on computers, making the Internet more attractive to speakers of the languages (Gyamfi ), and the design of software aimed at the standardization of such keyboards makes the actual making of the keyboard an easy task. The problem, however, is the lack of coordination and certification of the different initiatives (Schreiber ). While these initiatives build on a conception of written languages where standardization is the point of departure, digital writing does not necessarily rely on standard language. Faced with graphical challenges of African languages, writers simply replace special signs with easily available letters (e.g. Galtier ). Moreover, in countries with French as the official language, a writing regime based on the spelling rules of French is already applied for the writing of African languages, facilitating digital writing in these languages (Lexander ). The websites of local NGOs often figure in two versions: one in the local language in question and one in the former colonial language. This not only shows that the content is aimed at first language speakers, but it also points to the fact that the L speakers do not necessarily easily read the local languages, since they have often learned to read and write in the former colonial language. In a mobile novel project in South Africa, young people could choose to read the text in English or Xhosa and post comments in one of the two languages, on a mobisite. The majority of the readers downloaded the English version of the novel. However, Xhosa was also used in comments and discussion (Deumert ). A mobile telephony HIV-information service in the same country offers information in different African languages, but English is the preferred language by the overwhelming majority of users (Parker ). These examples show that use of African languages is a question not simply of accessibility, but also of the public getting used to writing their languages in electronic communications. This can only be a reality through popular use of them, through speakers of the languages acting as “uploaders” more than “downloaders” (cf. Halvorsen ).
. I
.................................................................................................................................. Languages of wider communication like Swahili, Lingala, Somali, Hausa, and Xhosa manifest themselves quantitatively on the web (Van der Veken and de Schryver ). The question, though, is whether the content in these languages is of interest for Internet users. When searching the Internet, numerous wordlists, dictionaries, and even classes
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
aimed at the teaching of an African language appear. Such pedagogical use of African languages often stems from initiatives from outside the speech community and is directed towards learners of the language. For those who cannot read or write English, French, or other common languages on the Web, surfing is of little value if the websites that exist in the language(s) they do understand are rarely updated and/or contain a lot of errors. Hence the importance of creating Internet content in African languages by L speakers. One of the most important open digital information resources is Wikipedia.org. In September , there were thirty-seven African language versions of the encyclopedia and twenty-three of these were represented in the dictionary Wiktionary. The content in these languages is often sparse, and the quality is not always deemed satisfactory.3 Deumert () argues that the Xhosa version of Wikipedia reproduces relations of power and inequality: non-native contributors produce texts intended for native speakers and thus only affirm the marginal status of Xhosa online. The Wiktionary in Xhosa contains several semantic and syntactic mistakes, errors that could not have been made by L speakers. African languages frequently appear in multilingual settings in electronic communication, either as single words or as longer stretches, when people leave messages in discussion forums, when they make contact on dating sites where the interface language is English or French, when they comment on news articles and photos, or when they blog. Often mixed codes or nonstandard versions of the languages are used—just as in spoken communication, in particular in the African cities. On social networking sites like Facebook, Africans living outside the continent use different languages to manage different relationships, in the home country, in the country of residence, and elsewhere in the world.4 Emigrants are often initiators of blogs and websites aimed at the revitalization of their home languages (Galtier ). Even these sites are multilingual; Galtier’s evaluation of a number of sites linked to Soninke identity shows that the quantity of Soninke language use varies greatly. Only a few studies have examined the actual popular use of African languages online. A comparison of linguistic aspects of Internet use among students from non-Anglophone countries found that all students used English less than their L—except African (Tanzanian) ones. It was thus concluded that the Internet had become another domain where English manifested itself as the language of the Tanzanian elite (Mafu ). Still, Swahili vocabulary has expanded with words related to ICT and even if English loans are dominant, neologisms are also being used by Tanzanians (Petzell ). Students in Tanzania, and in particular informatics students, use Swahili when working on computer and, because of this, they read and write more Swahili than before (Halvorsen ). Other studies also show that language practices change with ICT use. In North Africa, colloquial variants of Arabic in Roman script are used in electronic communication. Romanized vernacular Arabic and English make up the main languages of communication online among young professionals in Cairo, while the usual language of written communication in this group is Classical Arabic (Warschauer et al. ). In Moroccan discussion forums where French made up the interface language, French dominates the discussion, By Wikipedia’s quality index, List of Wikipedias (Wikipedia ). For a more detailed discussion of Wikipedia in African languages, see Deumert (: –). 4 The number of Facebook subscribers living in Africa has grown tremendously by .% since , and in there were about million subscribers (.% of the population, Internet World Stats, June ). 3
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
but Moroccan Arabic is used to insult, to be ironic, to joke, and to express strong personal feelings (Atifi ). Romanized vernacular Arabic is not limited to online communication, but spreads to other written communication like cards and handwritten messages to friends (Palfreyman and al Khalil ). In Sudan, poems in colloquial Arabic, using Arabic script, are exchanged in text messages (Lamoureux ). Use of African languages is not necessarily considered as enhancing democracy by the speakers themselves. The use of Krio on a mailing list set off a metalinguistic debate: it was considered by certain members as discourteous, as it excluded people from outside the speech community, and their comments led to a discussion of the role of the language on the Internet and in Sierra Leonean society (Wright ). Similar examples exist on other African websites and discussion forums (Lexander ); the Internet proves to be a space for both discussion and execution of language policies. When it comes to the written use of African languages, text messages can, however, be just as important as the Internet.
. T
.................................................................................................................................. Until Internet access becomes more common in Africa, via mobile handsets as well as computers, texting can prove to be a fruitful way of strengthening written use of the languages of the continent. The number of mobile phone subscribers per hundred inhabitants was . in (International Telecommunication Union ). Since sharing and borrowing phones is a common practice, a great proportion of the population has access to a mobile. Different empirical studies show a high degree of African languages, of codeswitching and language mixing in text messages. Quantitative measuring of language choice in texting reveals that even though English and French are the most used languages, we find African languages to a much more significant extent than in other written communication. Studies of text messages in Burkina Faso (Kibora ), Nigeria (Taiwo ), South Africa (Deumert and Masinyana ), and Senegal (Lexander ) show that African languages are used to a greater extent in the SMS than in other writing in which the texters engage. African languages are often used in multilingual messages, together with English or French, and the former colonial languages are modified according to local language practices, reflecting how these are embedded in local practices. Both English and French are used with unconventional spelling on the Internet and in text messages. While most studies (Deumert and Masinyana ; Taiwo ; Lexander ) found few examples of creative spelling of African languages in texting, more recent research (Deumert ; Dyers ; Hillewaert ) shows that unconventional abbreviations, clipping, and contractions are introduced into Xhosa text messages, in Facebook updates in Swahili, and in Senegal-based discussion forums.5 This seems to indicate that African languages are slowly being adapted to electronic communication. Moreover, Soninke web texts, Galtier () argues, initiate a standardization process of Soninke
5
KFC is, for instance, the Xhosa translation of LOL (laughing out loud), while RBT is the Wolof version of the same expression.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
orthography based on practice and not on theory. The spelling of African languages becomes the concern not only of linguists, but also of African Internet users. If people adapt languages to new media, they also adapt themselves to make use of it: people with poor literacy skills learn the necessary proceedings in order to make use of SMS (Kibora ). Texting can thus be a motivation to learn to read and write in African languages, as well as a way to develop and retain the skills acquired in literacy classes. While literacy in African languages tends to be marginalized, the use of African languages in text messages paves the way for the participation of people who are often excluded from reading and writing activities because they do not understand French or English.
. C
.................................................................................................................................. We have seen that there are two types of initiatives that promote the use of African languages in ICT. The formal initiatives driven by NGOs and commercial interests concentrate on technology development and translation of terminology, while the informal use of African languages in digital communication, often multilingual, is aimed at the expression of linguistic cultural practices—and at successful communication. In order to be more effective, the formal initiatives should take into account people’s digital literacy practices—what people do with digital literacy (cf. Barton and Hamilton’s definition of literacy practices, : ). The technology is developing so fast that it is difficult to predict what will be possible in the future. Mass use of African languages on the Internet will depend also on education and on language policies throughout the continent. It is clear, though, that through giving a written form to languages that are used more or less in oral communication only, Internet users and texters work on and with the language in new ways, according vitality and instrumental value to the writing of African languages. The Web also proves to be a constructive tool if members of a speech community decide to use it for the protection of languages whose number of speakers is critically low—a situation faced by a number of languages on the African continent.
R Androutsopoulos, J. (). ‘Bilingualism in the mass media and on the Internet’, in M. Heller (ed.), Bilingualism: A social approach. Basingstoke: PalgraveMacmillan, –. Atifi, H. (). ‘Choix linguistiques et alternance codique dans les forums diasporiques marocains’, in J. Gerbault (ed.), La langue du cyberespace: De la diversité aux normes. Paris: L’Harmattan, –. Barton, D., and Hamilton, M. (). Local Literacies: Reading and writing in one community. London and New York: Routledge. BBC News (). Surfing in Swahili, http://news.bbc.co.uk//hi/africa/.stm accessed August , . BBC News (). Microsoft Swahili speakers launch, http://news.bbc.co.uk//hi/africa/.stm accessed August , . Blommaert, J. (). Grassroots Literacy: Writing, identity and voice in Central Africa. London and New York: Routledge.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Camara, É., Ndamba, J., Nstadi, C., Rey, V., and Véronis, J. (). Traitement informatique des langues africaines: Problèmes et perspectives. CNRS, Action de Recherche Partagée AUPELF-UREF, http://horizon.documentation.ird.fr/exl-doc/pleins_textes/pleins_textes_/colloques/. pdf accessed August , . Chaudenson, R. (). ‘Langues et numérisation: français, créoles, langues africaines’, in I. Ndaywel è Nziem (ed.), Les langues africaines et créoles face à leur avenir. Paris: L’Harmattan, Institut de la Francophonie, and Agence Intergouvernementale de la Francophonie, –. Danet, B., and Herring, S. C. (eds.) (). The Multilingual Internet: Language, culture, and communication online. Oxford: Oxford University Press. Deumert, A. (). Klk cc . . . : supporting indigenous literacies in the digital space. Research report from The mLit project, http://mlit.files.wordpress.com///mlit_indigenous_literacies_adeumert_.pdf accessed August , . Deumert, A. (). Sociolinguistics and Mobile Communication. Edinburgh: Edinburgh University Press. Deumert, A., and Masinyana, S. O. (). ‘Mobile language choices—the use of English and isiXhosa in text messages (SMS): evidence from a bilingual South African sample’, English World-Wide : –. Dyers, C. (). Post-modern ‘languages’: the effects of mobile messaging by university students in three South African languages. Paper presented at the International Symposium on Bilingualism , June –, Oslo. Galtier, G. (). Les nouveaux sites Internet de la communauté soninké et la standardisation de la langue. Paper presented at the e colloque international ‘Langues et linguistique mandé’, – September, Paris. Gyamfi, A. (). ‘Closing the digital divide in sub-Saharan Africa: meeting the challenges of the information age’, Information Development : –. Halvorsen, T. A. (). ‘The ICT language in Tanzania’s higher education’, in B. Brock-Utne, B. Garbo, and G. Garbo (eds.), Language and Power: The implications of language for peace and development. Dar es Salaam: Mkuki na Nyota, –. Halvorsen, T. A. (). ‘Participation in the ICT era—implementations without humiliation, misconception and false consciousness’, International Review of Education : –. Hillewaert, S. (). ‘Writing with an accent: orthographic practice, emblems, and traces on Facebook’. Linguistic Anthropology : –. International Telecommunication Union. (). ICT Data and Statistics, http://www.itu.int/ITU-D/ ict/statistics/ accessed August , . Internet World Stats. (). Internet users in Africa, https://internetworldstats.com/stats.htm accessed August , . IRIN (). Technology: making the most of mobiles, Irinnews, United Nations’ Office for the Coordination of Humanitarian Affairs, http://www.thenewhumanitarian.org/news/// /making-most-mobiles accessed August , . Kamau, S. N. (). ‘A digital divide: Kiswahili holds the key’, Journal of Language, Technology and Entrepreneurship in Africa : –. Kibora, L. (). ‘Téléphonie mobile: l’appropriation du SMS par une “société d’oralité” ’, in M. de Bruijn, F. Nyamnjoh, and I. Brinkman (eds.), Mobile Phones: The new talking drums of everyday Africa. Bamenda (Cameroun) and Leiden: LANGAA and Africa Studies Centre, –. Kitalong, K. S., and Kitalong, T. (). ‘Complicating the tourist gaze: literacy and the Internet as catalysts for articulating a postcolonial Palauan identity’, in G. E. Hawisher and C. L. Selfe (eds.), Global Literacies and the World-Wide Web. London and New York: Routledge, –. Lamoureux, S. (). ‘Imagined connectivity, poetic text-messaging and appropriation in Sudan’, in M. Fernández-Ardèvol and A. Ros (eds.), Communication Technologies in Latin America and Africa: A multidisciplinary perspective. Barcelona: IN, –. Ledgard, J. M. (). ‘Digitale Afrika’, Aftenposten innsikt : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Lexander, K. V. (). ‘La communication médiatisée par les technologies de l’information et de la communication: la porte d’accès au domaine de l’écrit pour les langues africaines?’, in B. BrockUtne and I. Skattum (eds.), Languages and Education in Africa: A comparative and transdisciplinary analysis. Oxford: Symposium Books, –. Lexander, K. V. (). Pratiques plurilingues de l’écrit électronique: alternances codiques et choix de langue dans les SMS, les courriels et les conversations de la messagerie instantanée des étudiants de Dakar, Sénégal. Ph.D. thesis. University of Oslo. Lexander, K. V. (). ‘Texting and African languages literacy’, New Media and Society : –. Lexander, K. V., and Alcón-López, D. (forthcoming). ‘Digital language and new configurations of multilingualism: language use in a Senegal-based discussion forum’, in F. Lüpke (ed.), The Oxford Handbook of Atlantic Languages. Oxford: Oxford University Press. Mafu, S. (). ‘From the oral tradition to the information era: the case of Tanzania’, International Journal on Multicultural Societies : –, http://www.unesco.org/new/en/social-and-humansciences/resources/periodicals/diversities/past-issues/vol--no--/from-the-oral-tradition-tothe-information-era-the-case-of-tanzania/ accessed August , . Microsoft Corporation. (). ‘Language packs’, https://support.microsoft.com/en-us/help/ /language-packs accessed October , . Onguene Essono, L. M. (). ‘Langues nationales et NTIC: éclosion linguistique ou phagocytose?’, in A. Chéneau-Loquay (ed.), Mondialisation et technologies de la communication en Afrique. Paris: Karthala and Maison des Sciences de l’Homme d’Aquitaine, –. Osborn, D. Z. (). ‘African languages and information and communication technologies: literacy, access, and the future’, in J. Mugane, J. P. Hutchison, and D. A. Worman (eds.), Selected Proceedings of the th Annual Conference on African Linguistics. Somerville, MA: Cascadilla Proceedings Project, –, http://www.lingref.com/cpp/acal//paper.pdf accessed August , . Palfreyman, D., and al Khalil, M. (). ‘A funky language for teenzz to use: representing Gulf Arabic in instant messaging’, in B. Danet and S. C. Herring (eds.), –. Paolillo, J. (). ‘How much multilingualism? Language diversity on the Internet’, in B. Danet and S. C. Herring (eds.), –. Parker, M. (). Mobile language preferences by users on MXit, http://marlonparker.blogspot.com/ //mobile-language-preferences-by-users-on.html accessed August , . Petzell, M. (). ‘Expanding the Swahili vocabulary’, Africa and Asia : –. Schreiber, H. (). Presentation ‘AN’ (Africanization of software): Point Sud Development Workshop: pluridisciplinary research on language contact in West Africa, – February, Bamako. Taiwo, R. (). Code-switching/mixing in Yoruba–English Nigerian SMS text messaging. Paper presented at the International Symposium on Bilingualism , July –, Utrecht. Van der Veken, A., and de Schryver, G.-M. (). ‘Les langues africaines sur la Toile: étude des cas haoussa, somali, lingala et isixhosa’, Cahiers du Rifal : –, https://biblio.ugent.be/publication/ /file/ accessed October , . Walton, M. (). Mobile literacies and South African teens: leisure reading, writing, and MXit chatting for teens in Langa and Guguletu. Research report from The mLit project, https://mlit. files.wordpress.com///mlit_mobile_literacies_mwalton_.pdf accessed August , . Warschauer, M. (). Technology and Social Inclusion: Rethinking the digital divide. Boston and Cambridge, MA: MIT Press. Warschauer, M., El Said, G. R., and Zohry, A. (). ‘Language choice online: globalization and identity in Egypt’, in B. Danet and S. C. Herring (eds.), –. Wikipedia. (). List of Wikipedias, https://meta.wikimedia.org/wiki/List_of_Wikipedias#_articles_in_language_.other_than_main_page. accessed August , . Wright, H. K. (). ‘Email in African studies’, Convergence: The Journal of Research into New Media Technologies : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.............................................................................................................
LANGUAGE AND HISTORY .............................................................................................................
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
......................................................................................................................
, , Linguistics as a tool for historical reconstruction ......................................................................................................................
.
. I
.................................................................................................................................. T method of “Words and Things” (hereafter, WT) holds that if a word for something— let us say ‘oil palm’—may be reconstructed for an earlier form of a language let us say Proto-Bantu (PB)—then whenever and wherever people spoke PB, they knew something about oil palms (Blench : ; Bostoen : ; Ehret : –). Where the quality of evidence allows, lexical reconstructions result from applying the comparative method in historical linguistics. Principles of regular sound correspondence are used to detect and describe patterns of sound change and propose the earlier lexical forms from which those in evidence descended. Again, depending on the quality of the evidence, the distributions of meanings are mapped onto the geography of the sampled languages represented in historical language classifications as sequences of nodes. If the distributions are confined to the languages comprising a particular node, the word and meaning in question were innovated there. If a word and meaning are in languages belonging to different nodes, their innovation occurred somewhere in the area they shared. Innovations may be dated relatively—higher up the inverted tree is older; lower down is younger. Absolute dates come from glottochronology and by direct association with archaeological or paleoenvironmental records. Oil palms, for example, are present in radiocarbon-dated pollen records from Northwest-Central Africa. The languages there have reflexes of an ancient word for the tree and they descend from early nodes in the historical geography of Bantu languages (Bostoen et al. ). Glottochronology puts a clock on the patterned accumulation of random changes in vocabulary, the distribution of which across a set of related languages shows a bell curve (Ehret : –). Overlaps between inferred locations of earlier nodes in a tree and those of particular archaeological traditions allow the calibration of language divergences with the body of archaeological dates (de Luna et al. ). This need not presume speech community and material culture are isomorphic (Bostoen, chapter of this volume; Grollemund et al. ), because speakers of other languages
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
may have participated in a particular material culture. The method thus tracks a shifting universe of things driving semantic change in particular times and places. Practitioners know that cultural processes alter the array of objects people encounter, use, and talk about (Geeraerts : –) even if they focus on naming and classification in writing cultural history, bracketing thornier matters of meaning. The method has limits. Sources do not always contain enough detail to track the spatial distributions of the words and their things, reconstruct words, and locate them in a classification. The scholar must address relationships such as lexical transfer and other kinds of “contact phenomena” (Nurse : , –; Fleisch : –; Ehret : –). She must understand the effects on vocabulary of the fraught contexts of their recording: violent mercantilism, including enslavement and the purchase of slaves, imperial conquest or its many violent aftermaths, such as famine and epidemic disease, or the contentious debates about evangelization, customary law, and other dynamics of colonial rule (Irvine ; Makoni ). These issues reveal that there are more meanings and things than words; people may or may not adapt their lexicon to fit the situation (Bostoen , ; Blench ). Historians, archaeologists, and linguists tend to use WT differently. Historians track shifts and continuities in meaning or things through time and in particular places or regions, as reflections of African constructions and contests over culture, economy, and society in the past. Historians plot them in a narrative (Schoenbrun ). Most linguists focus semantic studies on a particular field of meaning (Bostoen ; Mougiama-Daouda ) rather than on the broader cultural richness of place or region in which a field of meaning existed. Reconstructing a lexicon with WT understands the repeated material practices of speaking and hearing in particular contexts across time to make meaning durable. Yet anthropological linguistic approaches to communicative practice take pragmatics and ambiguity very seriously (Hanks ), and historians write in genres other than narrative (Sewell ). The method emerged as the historical leanings of nineteenth-century lexicography, dialectology, and etymology converged in the study of onomasiology (essentially, synonymy) and semasiology (formerly interchangeable with “semantics” but, essentially, the study of polysemy; Dimmendaal ). It was largely shunted aside by structural linguistics in the twentieth century, with Meillet’s blend of social context and cognitive process being the influential exception (Geeraerts ). Together with paleoecology, archaeology, and oral traditions, the data and inferences of WT are a major source for writing about Africa’s earlier past. This chapter will focus on its applications in the field of Bantu languages, although it is alive elsewhere in Africa (Blench ; Fields-Black ; Ehret , ) and beyond (Blust ; Kirch and Green ; Heggarty ; Hill ; Ortman ).
. W, , :
.................................................................................................................................. Recent studies of the expansions of Africa’s Bantu languages have used WT to approach consensus about the broad geographical outlines, if not the timing and particular routes, of
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
that vast story (Ehret ; Klieman ; Bostoen, Grollemund, and Muluwa ; Bostoen, Clist, et al. ; Grollemund et al. ). For example, knowledge of particular species of trees, indicated by firm lexical-semantic reconstructions, correlates with paleoenvironmental evidence indicating shifts in the distribution of the varieties of forest in which such trees naturally occur (Bostoen et al. : –). In the northwestern zone of the vast region in which Bantu languages are spoken, each of three species of tree has two different names, and each pair of words has a split distribution. One distribution bisects the other, implying a relative chronology and a circumstance of renaming a species of tree well known to the speech communities ancestral to those bisected by the younger, innovating speech communities. All of them are “pioneer species”, found in regenerating forests. They mark a savannah–forest interface. The pairs of retained and innovated terms for each of the three species of tree may be mapped onto the nodes in a statistical-phylogenetic classification of Bantu languages. Those mappings may then be correlated with dated paleoenvironmental evidence that reveals savannah encroaching on rainforests, beginning c. , then creating patches of savannah inside the rainforest, some of which coalesced in north-south corridors linking savannahs on either side of the rainforest, by (Bostoen et al. ). In the initial phases of expansion Bantu speakers avoided rainforest environments in favor of the emerging corridors of familiar savannahs (Grollemund et al. ).1 The archaeological presence of pearl millet (Pennisetum glaucum) in areas today dominated by rainforest and the ability to reconstruct a root, *-cángú, with the same meaning as the most recent common ancestor of subgroups whose ancestors lived in these ancient savannah corridors (Kahlheber et al. : ) strengthens this account of expansion. These parallel conclusions reveal a great deal about the African past. Expansions involved the spread of people, rather than language shift (Bostoen et al. ). The social dynamics of Bantu speakers gaining knowledge of life in forested environments from autochthonous communities influenced a limited number of speech communities in later periods in the larger story. In the initial phases of expansion their importance unfolded in a savannah environment. These braided threads of evidence also help resolve a long-standing debate about the course of the Bantu expansions (cf. Rexová et al. ; Currie et al. ; and Grollemund et al. ). They favor the East-out-of-West model advanced by Heine et al. () and Ehret (), on largely lexicostatistical (similarity) evidence over the East versus West model advanced by Bastin et al. () and Vansina (, ), among others. But they paint the social complexity of change and continuity entailed by this long history of expansions in broad strokes. By asking about semantic fields implicated in cultural practice, politics, health, social organization, and so forth, WT can produce evidence bringing social process to tales of slow migrations, environmental change, and shifting food systems. The historian Rhiannon Stephens applies WT to semantic fields, like lineal descent systems, with sociological weight bearing on the standing of gendered individuals (Stephens ; de Luna ; Seligman ). Reconstructing kinship systems is a long-standing interest among WT practitioners (Jones and Milicic ). But its formal brilliance sometimes overshadows the gap between 1
This model’s assumptions that further research might prove incorrect include that the extent of the rainforest in was stable between and and may be used to judge if the locations of the nodes on their tree of Bantu languages during those millennia were in or out of the rainforest (Grollemund et al. : SI).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
rule and practice. As any ethnographer can tell you, kinship systems do not automatically produce particular social practices, such as property inheritance or marriage partnering. Rather, they are an “agreed formula for making political claims”. They are prone, over time, to tensions between practice and contingencies shaping the circumstantial logics in which group members hope “to expand by retaining reproductive capacity rather than by exchanging it” (MacGaffey : –). Such dynamics normally lie well beyond the purchase of WT, which cannot reconstruct the ordinary language of making claims driving such formulas. But, change in circumstantial logics leaves an imprint of “complexity and diversity” in lineal systems that WT may recover (Stephens : ). Framed by her classification of North Nyanza (NN) Bantu languages, spoken along the Ugandan littoral of Lake Victoria, Stephens has reconstructed a taxonomy of terms that indexes diverse influences for maternal kin in forming the social relations that breathe life into patrilineality (Stephens ). She reveals the contingent complexities of lineality, especially related to gender, inflecting its prescriptive practice. Two terms point to these dynamics. The first, *mwihwa/*baihwa, reconstructs regularly to Proto-North Nyanza (PNN) glossed as a kin term of reference, best rendered in English as ‘child or children of a female clanmate’. The form may be reconstructed as *jìpúá with a similar meaning— ‘man’s sister’s child’—to Proto-East Bantu (PEB), a much earlier node in a tree of the Bantu languages (Marck and Bostoen : ). PNN had other innovated terms for patricentric lineages (*ssiga) and clans (*ebika), suggesting together with *mwihwa, that growing investments in patrilineal politics turned in part on a growing importance for maternal kin ties. Their mutually constitutive relations, in play in PNN times and indexed by these bundles, generated a new term, *koiza, for maternal uncle, or mother’s brother. A perfectly good one already existed—*máá-dųmè (H) or ‘male mother’ (Marck and Bostoen : )—but it came to mean ‘male in-laws’ for PNN speakers. The innovation of *koiza implies a new context in lineal politics had emerged, in which people reformulated an older tradition of ‘male motherhood’ in favor of a new one focused on relations between a *mwihwa and a *koiza. The shift may reflect a familiar quid pro quo in instrumental politics built with lineal formulae: an ambitious *mwihwa relied on maternal kin for political success, from which the mother’s brother expected particular benefits and rights, signified metonymically by the new term, *koiza (Stephens : –). It is often presumed that lineality or dual-descent was invented or replaced only once. Given their qualities as formulae for making political claims, that presumption feels overdetermined. Stephens has excellent evidence of the phonological and tonological correspondences beneath each of her lexical reconstructions. But, in order to wring the heft of social process from kin terms of reference—‘child of female clan mate’, ‘mother’s brother’, or ‘maternal uncle’—she must resort to ethnographic descriptions, overwhelmingly from the turn of the twentieth century, treat them like packets of semantic value attached to the words, and map their distributions against the historical classification of PNN in order to seriate them. The complexity of colonial situations—such as the creation of an archly gendered customary law—framed those descriptions, imposing limits that Stephens understands. The tight scope of NN (only four languages) minimizes the effects of absent or superficial “ethnographic descriptions” of lineal politics on semantic reconstructions. At larger scales of analysis, such as PEB, those gaps are daunting (Marck and Bostoen : –).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
These examples apply classical approaches to meaning used in WT. As Fleisch has put it, changes in meaning not only should be plausible given “what is known about the mechanisms of semantic shift”; they should also have historical context. Changes in meaning should be “relevant to specific languages at certain points of time and space” (Fleisch : ). Fleisch explores both in reviewing approaches to the semantic history of a common Bantu form, *-gàndá (cl. /), which may occur with the English glosses ‘house, village, chief ’s enclosure’ in four of the five major regions in which Bantu languages are spoken today (Fleisch : –; BLR III, Main RN ). The metonymic extension from the concrete ‘house’ to the abstract ‘clan’, in particular noun classes (/ or /), is attested in a slightly narrower distribution (BLR III, Der RN ), suggesting it was a later development. The directionality fits with cognitive linguistic explanations of semantic change that privilege tangible contexts of embodied experience as sources for abstract meaning (Geeraerts : –). But that mechanism does not distinguish between different historical circumstances giving rise to the same distribution of forms, which might contradict assumptions about concrete-to-abstract directionality. For example, mapping the distributions of the reflexes of *-gàndá that refer specifically to ‘foundation pole of building’ reveals a narrower extent than that of reflexes with meanings bearing on units or features of social practice, like chiefship, village, or clan (Schoenbrun : –; Fleisch : –). Parsimony suggests that the directionality of semantic change was from abstract (more broadly distributed) to concrete (narrower distribution). The regular directionality in semantic change may be contradicted by distributional evidence mapped against historical classifications of languages. Such evidence may also contradict the assumption that an innovation occurs only once and not independently, on separate occasions (Fleisch : ). The scholar must define specific historical contexts for single (monogenetic) or repeated (polygenetic) shifts and for shifts in meaning that moved in either direction, between concrete and abstract domains. In other words, these key matters in WT are best explored as questions of social process. Did the appeal of the metonym or metaphor make it available for use in independent innovations (Fleisch’s polygenesis) that converged on the meaning ‘clan’? Or did the extension limit what mobile, interested persons could argue in instigating the formation of new social groups glossed in English as “clans” (whatever their actual constituencies or functions)? For historical semanticists, it is unnecessary to choose between the two scenarios; either or both metonymy and metaphor were in play. But, for historians, the differences are enormous. In the second scenario, the persons directing the formation of new collectivities emerge, introducing into a narrative a figure of mobility, interested in making clans (Schoenbrun ). In the first scenario, the process may be conceived of primarily as an exercise in creating political language, revealing which metaphors or metonyms stuck. Both circumstances confirm that metonymy and/or metaphor may cause paradigmatic semantic shifts. Each raises questions about semiotics by forcing historians to look for who offered the metonym/metaphor, by what means in addition to language, and in what contexts. But the second circumstance implies the emergence of a concrete network as a necessary byproduct of the features of belonging and mobility in an expanded political scale—in a word, clanship (Kodesh )—expressed by the new metonym/metaphor.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
As we have seen, gaps in the distribution and thin descriptions of semantic content often preclude a principled decision between the two cases. But in the Great Lakes region, we can choose. In the NN languages, a different lexical item—èkìkâ—displaced *-gàndá as the signifier for clanship. The stem, -ka, refers to a home or the house where people live (Stephens : ), and its extension to ‘clanship’ would appear to have followed the same logics of experience that underlay the older processes involving -gàndá. The terms are formally different, but the conditioning forces directing their extensions into new semantic domains are the same. The new term—èkìkâ—emerged as a separate, convergent extension from a central item—a building—in a concrete semantic field related to the organization of space. The directionality of semantic change was afforded by what people did in such spaces, including public healing during which social collectivities were refashioned, in part through the work of mobile, interested mediums who took advantage of the same affordances in play in the first scenario (Kodesh ; Schoenbrun ). Despite all these strengths, the bond between material culture and lexicalization is sometimes rather loose (Bostoen ). Semantically vague words do not distinguish elements of practice or circumstance. In Bostoen’s study of *-búmb- ‘to work with or fashion out of clay’ his onomasiological and semasiological analyses revealed that when the term means ‘to make pots’, it can refer to any one of many fashioning techniques involving very different bodily actions—molding, drawing, coiling, and so forth—producing pots that look different in the archaeological record (Bostoen ). Different fashioning techniques may be lexicalized with the same verb in two different speech communities or each may use a different verb to refer to the same fashioning techniques. Lexical reconstruction of those verbs will only reflect the making of pots, not the ways in which the pots were made, about which the archaeological record may have a lot to say. In other words, the distributions of things and practices will not always be matched by equivalent distributions of lexis verbalizing them. And distinctive actions in technical processes are not always verbalized with enough semantic precision to be recognizable as such, through reconstruction. Without the fieldwork observations of practice, one could not tell from studying the lexicon alone that coiled pots were made by superimposing coils to gradually build up the vessel. Thus, the properties of the languages under study, not only the sources used or the aims of scholars, limit the method. Bantu language speakers easily apply derivational suffixes to verb stems to lexicalize new referents (Bastin ), making their discontinuous distributions misleading guides to change over time. What seems a relict distribution, the holes in which practitioners of WT interpret as reflecting the replacement of older terms by newer ones (Bostoen et al. ), may actually reflect the independent, repeated, even simply temporary innovation by derivational application. Attestations may not, then, reflect inherited words but instead reflect inherited morphosemantic structures that have no necessary referential function tied to the practice or thing they name and describe. This aspect of pragmatics may elude scholars who have not mastered an African language but instead work primarily with the atomized lexis of dictionaries. Such conditions limit a central assumption of WT. Lexis does not correspond durably to cognitive and material reality across time. Changes in those realities do not generate words as the vinyl of a phonograph record rippling under its stylus produces sound (Blank ; Fleisch ; Bostoen : –).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
. N
.................................................................................................................................. Historians often want to know about intangible zones of meaning and practice implicated in material life, such as the gendered experience of architectural space. Or they want to know about dimensions of meaning and practice that exceed the material (Fleisch and Stephens ). When, for example, they point to other meanings, the material traces of the dead are not containers for a particular set of core meanings (Crossland ). Studying change and continuity in these aspects of social, affective, and intellectual life by analyzing and interpreting aspects of words’ relations to the real seems distinct from doing the same for millet, cattle, hoes, paddles, pots, and so forth. The difference can be illusory. Tangible crops, tools, and domestic animals are often signs or traces of intangible referents. Millet is a necessary component of a royal installation. An ox of a certain color must be used for sacrificial offerings. A canoe’s paddle is planted in the ground to mark the presence of a shrine. A pot is a person or a king, in gendered discourses about power and morality (David et al. ; Warnier ). We conceive of these applications of things as “extensions” from an otherwise “core” meaning at our peril. In each of the contexts of use just listed, their specific function fills up the entire context, making millet more than food, cows more than stores of wealth, paddles more than a component in a system of transportation, and pots more than vessels. For scholars to think of them in a semantic register of prototypicality, for example, they must entertain the possibility that contexts of use make each word-thing complicit in the creation of new prototypes, only tenuously linked to an “earlier” or “more widespread” or “ordinary” usage centered on agriculture, transportation, or containers. This should prevent confusing the stable mental image, prompted by hearing a word spoken, with its application, or use (Wittgenstein : e, §). Each of the categories in play in making meaning is so capacious and the scope of their relations so vast that scholars and historical actors impose their own limits on them in bringing their historical imaginations to life. Scholars must frame that terrain in cultural terms recognizable to their historical subjects. Cognitive linguistic approaches to semantics that build on conceptual metaphor (CM) theory offer a limited way to do so (Geeraerts and Cuyckens ). Embodied experience— in specific historical and cultural settings—generates much of the ground of CM, which representation bridges to language. “Some basic experiences are universal because they are motivated by biological and environmental universals, but others are constrained by architecture, material culture, and socially constructed patterns of discourse.” Statements like this essentially grant that the meaning-semantic relation is a continuum, “pervasively, though not entirely, a cultural phenomenon” (Palmer : –, ). Representation is both a mental process, in which imagination “re-presents non-present actions or events” and a material process, in which the physical properties, spatio-temporal locations, and socio-historical legacies of objects, places, and beings instigate, bringing particular actions and events no longer present into the present again (Knappett : –; Faur : –). The study of metaphor thus focuses attention on conceptualization, adding depth to Africa’s early history (Fleisch and Stephens ).
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
Complex CM contain, in the concrete parts of their basic linguistic signification, pointers to the contexts in which they were proposed. The juxtaposition of a source domain grounded in embodied experience with a more abstract target domain is the pointing event, contained within the lexical item as a compound form or as a series of semasiological additions. We saw this at work in the variety of semasiological processes linked with the lexemes *-gàndá and *-èkìkâ. It is also in play in one of the names of the python, ènzìramìre or ‘avoidance/offering swallower’, a term whose distribution includes NN languages, but not all the languages that attest *-gàndá. Thus, ènzìramìre is likely younger than *-gàndá but older than *-èkìkâ (Schoenbrun : ). Ènzìramìre is a compound term that juxtaposes the concrete event of swallowing (-mìre) with abstract ideas of avoidance, or taboo, and offerings made to spirits (ènzìra). Before it signified the literal meaning ‘python’, the juxtaposition was a form of rhetoric when first proposed, or innovated in the parlance of WT, and in the course of initial adoption by others in a community of speakers. Thus, it points to contexts of use, rhetorical or otherwise. CM and semasiological processes do not take us into contexts of use of ordinary language, but in their lexical and material iterations the juxtapositions index such contexts. CM also generates material iterations. By taking them as tokens, the ubiquitous pottery of Africa’s archaeological record (Bostoen, chapter of this volume) may be understood as material iterations of types of literal meaning. When archaeologists provided dates between and for an extraordinary set of terracotta figures discovered by press-ganged African prisoners nearly a century ago a few kilometers from the northwestern shores of Lake Victoria, arguably the location for the PNN group, they breathed new life into cultural meaning and social practice long ago (Ashley and Reid ). One of the figures in the bundle was a disembodied head with a neck-ring coiled like a constricting python, and one of the pots in the bundle had stylized pythons embossed vertically, at intervals, on its outer side. The materialized figure of the python and its lexicalized compound name were coeval. According to oral traditions and ethnographic observations from this part of the Lake Victoria region, pythons were involved in spirit possession. It seems likely, then, that these lexical and material iterations emerged in a similar context of use. That conclusion is supported by evidence internal to the data themselves and does not depend entirely on ethnographic or oral material. Critics of CMs find them to structure cognitive processes in a universalist fashion that prevents exploring why they are “only elaborated in some social worlds and not others” and why people “reproduce those phenomena” in “manipulations of the external environment?” (Fernandez ; Keane : n. ). Kövecses () answers this critique by considering cultural variation in the main foci of complex CM mapping. He shows that some conceptual or primary metaphors are found in historically unrelated language repertoires. It is not necessary to argue that the same CMs are present in all languages in order to argue that embodied experience plays a central role in linking understanding and thought through metaphor. Kövecses examines their widely divergent elaborations, as blends and extensions. But he maps too literally the fluidities, contradictions, and unsystematic character of “culture” by reference to embodied experiences of containers. Moreover, for Kövecses “culture” is not historical. It does not change in the hands of brokers or in the wake of contests. Other work builds on Kövecses but accepts cultural contingency and contest and keeps things in motion across time and space. It explores the challenges and rewards of reconfiguring public healing, politics, and clanship into a community of practice
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
aimed at constellating other communities of practice, in the expansive lake’s setting during a period of shifting seasonality and intensifying agronomic social relations (Schoenbrun ). Using WT to study CM works best with well-documented material evidence and the analysis of compound words and well-documented semasiology. It is not a panacea for deepening WT’s approach to context in making meaning. So, the principled clarity of the classic approaches to doing semantic history without documents will remain central to WT. Some uncertainties diminish when we have documents. Textual corpus linguistics in Africa has grown in the last fifteen years (de Schryver ; Kawalya et al. ). The approach promises rich rewards for WT, by revealing issues of semantic vagueness (Bostoen ) and the interface between semantics and pragmatics (de Kind et al. ). It may document degrees of elaboration of complex CM (Kövecses : –). Semantic vagueness and linguistic expressions of CM punch through the sound barrier between historical lexical semantics and semiotic process produced by the absence of textual corpora contemporary with times and places prior to the seventeenth century.
. C
.................................................................................................................................. WT has an enduring role to play in writing Africa’s history beyond times and places with literacy. Indeed, some have argued that the absence of textual corpora in the continent’s vernacular languages before c. was a blessing in disguise because it forced Africanists to develop approaches—like WT—to the historical study of diachronic semantics other than lexicography (Schadeberg : ; Blench : ). But the indeterminacies of meaning and the modular qualities of language structure limit that role, inviting both stricter application of diachronic lexical semantic methods and innovative approaches to semiotics in an environment of reconstructed, rather than recorded language. African languages are rich in lexical clues about the world in which Africans thought, felt, and worked. But the relations between clue and life are often less precise than historians might wish.
A This chapter has benefited from clarifications and cautions offered by Marcos Almeida, Kathryn de Luna, Christopher Ehret, Nicholas W. S. Smith, Rachel Taylor, and Rhiannon Stephens.
R Ashley, C., and Reid, A. (). ‘A reconsideration of the figures from Luzira’, Azania : –. Bastin, Y. (). Les relations sémantiques dans les langues bantoues. Brussels: Académie royale des sciences d’outre-mer. Bastin, Y., Coupez, A., and de Halleux, B. (). ‘Classification lexicostatistique des langues bantoues ( relevés)’, Bulletin des Séances de l’Académie Royale des Sciences d’Outre-Mer : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
Bastin, Y., Coupez, A., Mumba, E., and Schadeberg, T. C., (eds.) (). Bantu Lexical Reconstructions . Tervuren: Africa Museum, www.africamuseum.be/collections/browsecollections/humansciences/blr/ results_main accessed May , . Blank, A. (). ‘Words and concepts in time: towards diachronic cognitive onomasiology’, in R. Echardt, K. von Heusinger, and C. Schwarze (eds.), Words in Time: Diachronic Semantics from Different Points of View. Berlin: Mouton de Gruyter, –. Blench, R. (). Archaeology, Language, and the African Past. Walnut Creek, CA: AltaMira Press. Blust, R. (). ‘The prehistory of the Austronesian-speaking peoples: a view from language’, Journal of World Prehistory : –. Bostoen, K. (). ‘Vocabulary of pottery fashioning techniques in Great Lakes Bantu: a comparative onomasiological study’, in A. Akinlabi and O. Adesola (eds.), Proceedings of the th World Congress of African Languages, New Brunswick . Cologne: Rüdiger Köppe, –. Bostoen, K. (). Des mots et des pots en bantou: Une approche linguistique de l’histoire de la céramique en Afrique. Frankfurt am Main: Peter Lang. Bostoen, K. (). ‘Pearl millet in early Bantu speech communities in Central Africa: a reconsideration of the lexical evidence’, Afrika und Übersee : –. Bostoen, K. (). ‘Semantic vagueness and cross-linguistic lexical fragmentation in Bantu: impeding factors for linguistic palaeontology’, Sprache und Geschichte in Afrika : –. Bostoen, K., Clist, B., Doumenge, C., Grollemund, R., Hombert, J.-M., Muluwa, J. K., and Maley, J. (). ‘Middle to Late Holocene paleoclimatic change and the early Bantu expansion in the rain forests of Western Central Africa’, Current Anthropology : –. Bostoen, K., Grollemund, R., and Muluwa, J. K. (). ‘Climate-induced vegetation dynamics and the Bantu expansion: evidence from Bantu names for pioneer trees (Elaeis guineensis, Canarium schweinfurthii, and Musanga cecropioides)’, Comptes Rendus Geoscience : –. Crossland, Z. (). Ancestral Encounters in Highland Madagascar: Material signs and traces of the dead. New York: Cambridge University Press. Currie, T. E., Meade, A., Guillon, M., and Mace, R. (). ‘Cultural phylogeography of the Bantu languages of sub-Saharan Africa’, Proceedings of the Royal Society B , doi ./ rspb... David, N., Sterner, J., and Gavua, K. (). ‘Why pots are decorated’, Current Anthropology : –. de Kind, J., Dom, S., and de Schryver, G.-M. (). ‘Event-centrality and the pragmatics-semantics interface in Kikongo: from predication focus to progressive aspect and vice versa’, Folia Linguistica Historica : –. de Luna, K. M. (). ‘Hunting reputations: talent, individuals and community in precolonial South Central Africa’, Journal of African History : –. de Luna, K. M., Fleisher, J. B., and McIntosh, S. K. (). ‘Thinking across the African past: interdisciplinarity and early history’, African Archaeological Review : –. de Schryver, G.-M. (). ‘The compilation of electronic corpora, with special reference to the African languages’, South African Journal of Linguistics and Applied Language Studies : –. Dimmendaal, G. J. (). Historical Linguistics and the Comparative Study of African Languages. Amsterdam and Philadelphia, PA: John Benjamins. Ehret, C. (). ‘Bantu expansions: re-envisioning a central problem in early African history’, International Journal of African Historical Studies : –. Ehret, C. (). History and the Testimony of Language. Berkeley and Los Angeles, CA: University of California Press. Ehret, C. (). ‘Linguistic evidence and the origins of food production in Africa: where are we now?’, in C. J. Stevens, S. Nixon, M. A. Murray, and D. Q. Fuller (eds.), Archaeology of African Plant Use. Walnut Creek, CA: Left Coast Press, –. Faur, E. (). ‘Integral semantics and conceptual metaphor: rethinking conceptual metaphor within an integral semantics framework’, Journal of Cognitive Semiotics : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, ,
Fernandez, J. (ed.) (). Beyond Metaphor: The theory of tropes in anthropology. Stanford, CA: Stanford University Press. Fields-Black, E. (). Deep Roots: Rice farmers in West Africa and the African Diaspora. Bloomington, IN: Indiana University Press. Fleisch, A. (). ‘The reconstruction of lexical semantics in Bantu’, Sprache und Geschichte in Afrika : –. Fleisch, A., and Stephens, R. (eds.) (). Doing Conceptual History in Africa. New York: Berghahn Books. Geeraerts, D. (). Theories of Lexical Semantics. Oxford: Oxford University Press. Geeraerts, D., and Cuyckens, H. (eds.) (). The Oxford Handbook of Cognitive Linguistics. Oxford: Oxford University Press. Grollemund, R., Bradford, S., Bostoen, K., Meade, A., Venditti, C., and Pagel, M. (). ‘Bantu expansion shows that habitat alters the route and pace of human dispersals’, Proceedings of the National Academy of Sciences , : –. Grollemund, R., Bradford, S., Bostoen, K., Meade, A., Venditti, C., and Pagel, M. ‘Supporting Information’, ttps://www.pnas.org/content/suppl//// DCSupplemental, accessed September , . Hanks, W. (). Language and Communicative Practices. Boulder, CO: Westview Press. Heggarty, P. (). ‘Linguistics for archaeologists: principles, methods and the case of the Incas’, Cambridge Archaeological Journal : –. Heine, B., Hoff, H., and Vossen, R. (). ‘Neuere Ergebnisse zur Territorialgeschichte der Bantu’, in W. J. G. Möhlig, F. Rottland, and B. Heine (eds.), Zur Sprachgeschichte und Ethnohistorie in Afrika: Neue Beiträge afrikanistischer Forschungen (Festschrift Oswin Köhler). Berlin: Dietrich Reimer, –. Hill, J. H. (). ‘Proto-Uto-Aztecan as a Mesoamerican language’, Ancient Mesoamerica : –. Irvine, J. T. (). ‘Subjected words: African linguistics and the colonial encounter’, Language & Communication : –. Jones, D., and Milicic, B. (eds.) (). Kinship, Language, and Prehistory: Per Hage and the renaissance of kinship studies. Salt Lake City, UT: University of Utah Press. Kahlheber, S., Bostoen, K., and Neumann, K. (). ‘Early plant cultivation in the Central African rain forest: first millennium pearl millet from south Cameroon’, Journal of African Archaeology : –. Kawalya, D., Bostoen, K., and de Schryver, G.-M. (). ‘Diachronic semantics of the modal verb sóból- in Luganda: a corpus-driven approach’, International Journal of Corpus Linguistics : –. Keane, W. (). ‘Marked, absent, habitual: approaches to neolithic religion at Catalhöyük’, in I. Hodder (ed.), Religion in the Emergence of Civilization: Catalhöyük as a case study. Cambridge: Cambridge University Press, –. Kirch, P. V., and Green, R. C. (). Hawaiki, Ancestral Polynesia: An essay in historical anthropology. Cambridge: Cambridge University Press. Klieman, K. (). ‘The Pygmies Were Our Compass’: Bantu and Batwa in the history of West Central Africa, early times to c. C.E. Portsmouth: Heinemann. Knappett, C. (). An Archaeology of Interaction: Network perspectives on material culture and society. Oxford: Oxford University Press. Kodesh, N. (). Beyond the Royal Gaze: Clanship and public healing in Buganda. Charlottesville, VA: University of Virginia Press. Kövecses, Z. (). Metaphor in Culture: Universality and variation. Cambridge: Cambridge University Press. MacGaffey, W. (). ‘Changing representations in Central African history’, Journal of African History : –. Makoni, S. (). ‘An integrationist perspective on colonial linguistics’, Language Sciences : –.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
.
Marck, J., and Bostoen, K. (). ‘Proto-Oceanic society (Austronesian) and Proto-East Bantu society (Niger-Congo) residence, descent, and kin terms, ca. ’, in D. Jones and B. Milicic (eds.), –. Mougiama-Daouda, P. (). ‘Phonological irregularities, reconstruction and cultural vocabulary: the names of fish in the Bantu languages of the northwest (Gabon)’, Diachronica : –. Nurse, D. (). ‘The contributions of linguistics to the study of history in Africa’, Journal of African History : –. Ortman, S. (). Winds from the North: Tewa origins and historical anthropology. Salt Lake City, UT: University of Utah Press. Palmer, G. (). ‘Cognitive linguistics and anthropological linguistics’, in D. Geeraerts and H. Cuyckens (eds.), –. Rexová, K., Bastin, Y., and Frynta, D. (). ‘Cladistic analysis of Bantu languages: a new tree based on combined lexical and grammatical data’, Naturwissenschaften (): – [Electronic supplementary material: ___MOESM_ESM.pdf, ___MOESM_ESM.pdf, _ __MOESM_ESM.pdf accessed September , ]. Schadeberg, T. C. (). ‘Historical linguistics’, in D. Nurse and G. Philippson (eds.), The Bantu Languages. London and New York: Routledge, –. Schoenbrun, D. (). The Historical Reconstruction of Great Lakes Bantu Cultural Vocabulary: Etymologies and distributions. Cologne: Rüdiger Köppe. Schoenbrun, D. (). ‘Mixing, moving, making, meaning: possible futures for the distant past’, Africa Archaeological Review : –. Schoenbrun, D. (). ‘Pythons worked: constellating communities of practice with conceptual metaphor in northern Lake Victoria, ca. –’. in A. P. Roddick and A. B. Stahl (eds.), Knowledge in Motion: Constellations of learning across time and place. Tucson, AZ: University of Arizona Press, –. Seligman, A. F. (). ‘Wealth not by any other name: inland African material aesthetics in expanding commercial times, ca. th–th centuries’, International Journal of African Historical Studies : –. Sewell, W. (). Logics of History. Chicago, IL: University of Chicago Press. Stephens, R. (). ‘Lineage and society in precolonial Uganda’, Journal of African History : –. Stephens, R. (). A History of African Motherhood: The case of Uganda, –. New York: Cambridge University Press. Vansina, J. (). ‘Western Bantu expansion’, Journal of African History : –. Vansina, J. (). ‘New linguistic evidence and the Bantu expansion’, Journal of African History : –. Warnier, J.-P. (). The Pot-King: The body and technologies of power. Leiden: Brill. Wittgenstein, L. (). Philosophical Investigations, trans. and ed. by G. E. M. Anscombe. Oxford: Basil Blackwell.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. O is Africa’s mode of civilization par excellence (Derive ). In most of the continent, literacy is a recent phenomenon. This is a challenge for scholars of Africa’s human past, especially those interested in its earlier periods. It is indispensable not only to exploit alternative historical evidence, but also to develop a reliable interdisciplinary methodology. African archaeology and historical linguistics have played a leading role as complementary approaches to African history (Ehret and Posnansky ). Historical linguistics deals with the immaterial vestiges of the past in Africa’s present-day languages. Archaeology, on the other hand, unearths the material remnants of past cultures. The two disciplines offer different points of view on important issues in African history, such as migration and language expansion, subsistence economy, technological evolution, the rise of social and political complexity, the development of trade, and religion. Both deal with human behavior and culture in African communities of the past. Moreover, both share similar key concepts and terminology. Unsurprisingly, the attempts to bring together these two bodies of historical evidence have been numerous, though not always equally successful. Several authors have severely criticized the interplay between African archaeology and linguistics, blaming it for a lack of critical evaluation of underlying concepts and methods, a marked propensity for circular thinking and generating grand but highly speculative explanatory schemes (Möhlig ; Eggert ). Although this criticism is certainly not unwarranted, we should not throw out the baby with the bathwater. None of the individual disciplines is capable of solving on its own the many intricate puzzles of African history. A sound interdisciplinary collaboration between archaeologists and linguists working in Africa requires in the first place a familiarity with each other’s underlying concepts, methods, and evidence. Although historical linguistics and archaeology are two fundamentally distinct disciplines, they do share a number of principles. This chapter wants to offer an outline of these common principles and their specific interpretation in both disciplines, and to provide some thoughts on what a judicious interdisciplinary archaeo-linguistic
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
approach to the African past might look like. All issues discussed in this chapter are illustrated with examples that connect to the so-called “Bantu Expansion”. This is a textbook case of a historical conundrum that has intrigued scholars of various disciplines for many decades now. Multidisciplinary debate persists on the demographic, technological, and economic developments that facilitated this exceptional language dispersal. To be sure, this chapter is not a synthesis of multidisciplinary research on the Bantu Expansion. This was recently done elsewhere (see, e.g., Maret ; Pakendorf et al. ).
. S
.................................................................................................................................. In archaeology, stratigraphy is a fundamental principle that is underscored by the law of superposition and understood as the way in which geological and anthropogenic deposits have accumulated to form the distinct layers of an archaeological site (Orton et al. ; Gamble ). The stratigraphy at Shum Laka, for example, led to the identification of four large stratigraphical units bearing witness to , years of human occupation from the Late Pleistocene to the Late Holocene (Lavachery ; Cornelissen ). This rock shelter in the Grassfields of northwestern Cameroon is the principal archaeological site associated with the presumed Bantu homeland. The lower Late Pleistocene layers reveal a microlithic quartz industry used by people who exploited an open environment with patches of forest that did not undergo drastic changes between c. and (Cornelissen ). The upper Holocene unit, however, consisting of ocher and gray ashes, indicates significant evolution in human activities. From c. onwards, a new industry with pottery and polished macrolithic tools, such as waisted axes and blades, strongly linked to the exploitation of trees, slowly developed to become more prominent around – . The appearance of these new technologies has been attributed to the small-scale immigration of more northerly communities fleeing the climatic deterioration that plagued the Sahara and Sahel around – (Lavachery ). These immigrants may have introduced the Benue-Congo languages out of which the Bantu ancestor language(s) emerged (Bostoen ). The historical link with the Bantu Expansion becomes all the more likely if one reckons that slightly younger, but closely related technological assemblages, also associated with the exploitation of trees like Elaeis guineensis and Canarium schweinfurthii, have been recovered from several archaeological sites farther south, i.e. in Cameroon, Gabon, the two Congos, and the Central African Republic. These date from the early second millennium until the last centuries , and the younger they are, the farther south(east) they are located (cf. Maret ). Given Central Africa’s linguistic landscape, the people involved in the spread of these new lifestyles were most likely Bantu speakers. Historical linguists also rely on the concept of stratigraphy. It refers to “the layering of grammatical and lexical material in a language or dialect which reflects its historical development and past contacts between its speakers and bearers of other linguistic and cultural traditions” (Andersen : ). The lexis of a language is transmitted through time and changes through the loss of old words and the incorporation of new words. It accumulates words of different age. New words are either language-internal creations or they were adopted from another language. Loanwords are important in linguistic
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
stratigraphy, because they provide significant information on cross-cultural contacts. It is well known, for instance, that Swahili not only inherited words from Proto-Bantu (PB) or a more recent ancestor, such as Proto-Sabaki (Nurse and Hinnebusch ). It also adopted numerous loanwords from different donor languages, e.g. Arabic, English, Hindi, Malay, Malagasy, Persian, and Portuguese, resulting from different kinds of contact situations through time (Schadeberg ). At the same time, as a lingua franca since precolonial times, Swahili was a donor to many East African languages, both Bantu and other. Several East Bantu languages borrowed, for example, the Swahili words chupa and chungu to designate imported material culture, i.e. glass bottles and metal pots. These borrowings exist alongside the inherited and phonologically regular reflexes of the PB reconstructions *-cʊ́pà and *-jʊ̀ngʊ́ referring to calabashes and earthenware cooking pots (Bulkens : ; Bostoen : –). The original Swahili words ultimately stem from the same PB etymons and designated the same kind of indigenous recipients. The lexis of a language is thus a multilayered aggregate. Unlike archaeological strata, lexical layers are not subject to the law of superposition. They are never neatly superposed. There is permanent “stratigraphic contamination”, so to speak. It is the task of the historical linguist to order the successive strata. To do so, one can rely on a number of methods of relative dating, such as seriation and geographic distribution.
. S
.................................................................................................................................. Just like stratigraphy, seriation is a basic archaeological concept. In archaeology, seriation results from the combination of stratigraphy with popularity, the basic idea that new artifacts or techniques spread and become more popular before they start to disappear when similar new artifacts or techniques are invented. Seriation thus serves to order stylistic units into a relative chronological sequence (Orton et al. ; Gamble ). Pottery is one of the most commonly used artifacts for archaeological seriation. In the East African Great Lakes regions, more particularly around Lake Victoria, the earliest ceramic tradition, commonly known as Kansyore ware, has been linked to a Late Stone Age huntergatherer subsistence economy, strongly relying on aquatic resources. Its chronology ranges between the early sixth millennium and . This is exceptional, because Kansyore sites are the only known hunter-gatherer sites in East Africa associated with large quantities of ceramics before the advent of food production around – (Dale and Ashley ). In the wider area, there are several other Neolithic ceramic traditions of more recent date, i.e. later than , which are commonly called “Pastoral Neolithic” because they are associated with a lifestyle centered on herding livestock (Bower ). The oldest Early Iron Age ceramic tradition of the region is known as Urewe and dates between and (Ashley ). Urewe sites are associated with a lifestyle that differs considerably from earlier ones in that there is evidence for new practices, such as the cultivation of crops, settled village life, and metallurgy (Phillipson : ). It is commonly believed that the people who introduced this new lifestyle were Bantu speakers. The gradual eastward spread of Urewe pottery has, therefore, often been regarded as the main archaeological signature of the first spread of East Bantu languages (Lane : ). The linguistic affiliation of the producers of the above-mentioned Neolithic ceramics is less straightforward. Given the present-day
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
geolinguistic landscape, they might have spoken Cushitic (Afro-Asiatic) and/or Nilotic (NiloSaharan) languages, but this is nothing more than an educated guess. In any case, the first Bantu-speaking Urewe producers seem to have neighbored Kansyore producers during several centuries of the last millennium (Schoenbrun : ; Ashley ). In historical linguistics, seriation commonly relies on the sequential ordering of sound changes (Gregersen ; Antilla ). Every language is subject to sound changes, which can be called regular to the extent that they affect all words sharing a given phonological environment. The chronological sequencing of sound changes is primarily used for the historical classification of languages through the Neogrammarian principle of shared innovations. If closely related languages share a historical change, whether lexical, phonological, or grammatical, there is a good chance that this innovation only happened once, i.e. in their most common recent ancestor, although independent convergent change can never be entirely excluded. As a case in point, the evolution of Proto-Bantu stop consonants in front of the high vowel *u is indicative of the subclassification of the “Bantu Botatwe” (BB) languages of Zambia. In their latest common ancestor, the reflexes of *bu/*gu were still distinct from the reflex of *du, and those of *pu/*ku from that of *tu: *bu/*gu > vu, *du > zu, *pu/*ku > fu, *tu > su. This situation in still observed in Tonga and the Western BB languages (Fwe, Shanjo, Totela, Subiya). However, the Eastern BB languages (Soli, Lenje, Ila) set themselves apart by undergoing a subsequent shared innovation, i.e. the merger of places of articulation: *du > zu > vu and *tu > su > fu, hereby entirely wiping out the distinction between PB *bu/*du/*gu on the one hand and *pu/*tu/*ku on the other hand. Lenje and Soli seceded in their turn from Ila by a subsequent shared innovation, i.e. spirant devoicing: vu > fu, hereby entirely losing the contrast between PB voiced and voiceless stop consonants. In the same vein, all BB languages underwent a shared innovation in front of the PB high vowel *i, i.e. the total leveling of places of articulation. This phonological innovation goes back to their most recent common ancestor and sets them apart as a historical subgroup from neighboring Bantu languages of the Bemba and Lala-Bisa-Lamba groups, where this sound change did not occur (Bostoen ). Once one has an idea of the internal classification of a language group and the relative chronology of sound changes, seriation is also a helpful dating device for loanwords. The earlier foreign words are borrowed, the more sound changes they have in common with regularly inherited words and the better they are phonologically integrated, making it difficult to identify them. Numerous East Bantu languages share, for example, a similar word to designate pearl millet, i.e. *-bèdé. This term is in all likelihood of Nilo-Saharan origin. However, due to its overall phonological regularity, this term can be reconstructed in Proto-East Bantu. This implies that Bantu speakers had already adopted the word and the cereal to which it referred before Bantu languages started to spread over East Africa (Bostoen /).
. G
.................................................................................................................................. The geographic distribution of features yields important historical information in both archaeology and linguistics. In archaeology, a stylistic horizon stands for a group of
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
contemporaneous styles that are closely related, yet regional in character and distributed over a large area when plotted as a whole. The East African ceramic traditions discussed in section . above are such stylistic horizons. In the equatorial rainforest, more particularly the inner Congo basin, the earliest ceramics are known as the “Imbonga Horizon” (– ), named after their type site Imbonga on the Momboyo River. Its geographic distribution is centered on the left bank of the Congo River near Mbandanka and its direct tributaries: the Ruki, Ikelemba, and Lulonga. The area’s settlement history basically coincides with the penetration of the producers of this pottery. According to Wotzka (), this style would be ancestral to all subsequent archaeological ceramic traditions and even to the pots that are still produced in some villages today. They would constitute a single, historically evolved system of interrelated pottery traditions deriving from this common ancestral style. The distribution of Imbonga pottery would thus mark the initial settlement nucleus in the west-central part of the Congo basin from where several waves of eastbound upstream migrations originated in more recent periods, resulting in a gradual process of settlement expansion that was complete around . Since the area’s modern languages are all Bantu and there is no discontinuity in the archaeological sequence that could signal profound language change, the emergence of Imbonga pottery, after has been tentatively connected with the earliest immigration of Bantu speakers into the inner Congo Basin (Wotzka ). In linguistics, the method dealing with geographic distribution is linguistic geography or geolinguistics (Antilla ; Breton ; see also Lafkioui, chapter of this volume). This approach is indebted to dialectology and examines the possible relationships between language and the geocultural environment. It is used for mapping loanword diffusion routes and for determining their direction of borrowing and also as a relative chronology device. In linguistic geography, isoglosses are the equivalent of stylistic horizons in archaeology. They mark the geographic distribution of a given linguistic feature shared by a number of languages. Space is then interpreted as a function of time. The relative chronological interpretation of isoglosses is done according to certain areal norms which are not strict rules, but rather “handy rules of thumb” (Antilla : ), e.g. “the earlier form is the most scattered one, preferably occurring in the more peripheral areas” or “the more recent form occurs in a contiguous group of languages, which may be large, but is generally surrounded by languages having the older form.” A judicious historical interpretation of isoglosses requires a basic insight into the internal classification of a language family. The relative time-depth does not depend so much on the number of languages in which a feature occurs as on its distribution over distinct historical subgroups. A term that is rare but scattered amongst the North-West and East Bantu languages, for instance, is considered older than a synonym that is densely spread among West Bantu languages only. As a case in point, the most common Bantu word referring to the stiff porridge people eat as their staple dish is *-kímà. This term could be reconstructed to PB, not so much because it is widespread among present-day East and South-West Bantu languages, as because it also occurs in a few North-West Bantu languages. It originally referred to a porridge made as a mash and only later to flour porridge. The latter meaning is the outcome of a subsequent semantic innovation. This can be deduced from the fact that this meaning is solely attested among Bantu languages spoken beyond the North-West. Hence, linguistic geography is also important for diachronic semantics. Stiff porridge thus belongs to the ancestral culinary traditions of Bantu speakers, but the way they prepared it evolved in the course
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
of their expansion. Porridge made of flour is a culinary innovation, which ties in with the fact that common Bantu words designating flour such as *-tù and *-ʊ̀ngà cannot be reconstructed to PB. This is again highly relevant for the history of cereal agriculture amongst Bantu speakers (Ricquier and Bostoen ).
. A -
.................................................................................................................................. Both archaeology and historical linguistics rely on stratigraphy to reconstruct relative chronology. What about absolute dating? Archaeology achieved a major breakthrough in this regard after with the development of isotope-decay methods of which radiocarbon (14C) was the first and is still the best known to non-archaeologists. It is of great use because carbon is so widespread. Charcoal, wood, bone, fibre, shell, and seeds can all be dated. Dendrochronology, geomagnetism, thermoluminescence, electron spin resonance, and amino acid epimerization are other techniques that have different time spans and apply to different materials (Gamble : –). To be sure, these science-based techniques provide raw absolute dates which need calibration to be used as calendar dates. There are several calibration methods involving different degrees of imprecision and uncertainty, which cannot be discussed here. Nonetheless, these uncertainties would prevent nobody from considering absolute dating as a comparative advantage of archaeology over linguistics. It is not that no method was ever developed to calculate the absolute time elapsed since related languages split from a common ancestral language. Glottochronology is the bestknown (cf. Swadesh ), but also the most contested amongst linguists. This method is based on lexicostatistics, which determines the percentage of cognate words in a standard list of basic vocabulary shared by related languages. Assuming that such a basic vocabulary is more resistant to borrowing than cultural vocabulary and is replaced at a constant rate over time, Swadesh developed an arithmetic formula to convert the rate of lexical similarity between two languages in the absolute amount of time since they separated. Although Swadesh’s original device has been sophisticated in many ways, skepticism continues to prevail amongst historical linguists, many of whom categorically reject the historical significance of any quantitative study relying on such a limited and purely lexical data set. The most fundamental objection is that languages do not necessarily change at a constant rate, not even their basic lexicon. All depends on the “ecology of language evolution” (Mufwene ), i.e. the specific social, cultural, demographic, and ecological conditions in which language evolves. A related problem is the partial incompatibility between linguistics and archaeology in terms of time-depth. The scope of African archaeology reaches as deep as the earliest irrefutable evidence for human material culture originating from Ethiopia and Kenya, i.e. between . and two million years ago (Phillipson : ). The scope of African historical linguistics is much shallower, taking into account the limited time-depth of the present-day African language phyla. As far as the available data allow making an educated guess, the origins of Niger-Congo, Nilo-Saharan, and Afro-Asiatic are to be situated between , and , years ago (Blench ). Khoisan—whose genealogical unity is contested—is probably older, perhaps up to , years ago (Güldemann and Stoneking ), and
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
possibly even linked to the advent of the Homo sapiens in central and southern Africa from about , years ago (Blench ). This is a considerable discrepancy with the timedepths reachable for archaeologists, certainly if one reckons that our knowledge of the earliest language stages is fairly restricted. An additional problem is that old language families disappeared due to recent expansions. The Bantu Expansion is again a case in point. Autochthonous hunter-gatherer communities often adopted the language of the newcomers without taking on their way of life. Such is assumedly the case of the Pygmy groups living scattered throughout the Central African rainforest. None of them speaks a language which is not related to that of their agriculturist neighbors. The overall majority of Central African Pygmies speak a Bantu language, but some speak an Ubangian (NigerCongo) or a Central Sudanic language (Nilo-Saharan). However, recent molecular anthropological research has pointed out that other Central African hunter-gatherer groups not only lost their languages, but were also acculturated and assimilated into Bantu speech communities (Pakendorf et al. ). This seemingly happened without a major linguistic impact, at least not in the form of an easily discernable substrate.
. T -?
.................................................................................................................................. From the preceding it is clear that archaeology and historical linguistics share enough common ground to collaborate on certain aspects of African history, despite their particularities in terms of method, scope, and evidence. However, the question is whether it is also feasible to come to a reliable archaeo-linguistic method? How can archaeological and linguistic evidence be used together to reconstruct aspects of past cultures? And how to avoid the circular thinking which has too often characterized the interplay between archaeology and linguistics on important issues in African history? To give only one such example of circular thinking, Eggert () shows efficiently how the historical reinterpretation of Early Iron Age archaeological evidence regarding the Bantu Expansion presented in Phillipson () is strongly influenced by the historical reclassification of Bantu languages proposed by Heine (), while the same linguist relied some years later on the archaeological findings of Phillipson () to reinforce the results of his new study in Bantu language classification (Heine et al. ). Book-length studies have been dedicated to similar problems of interdisciplinary method (e.g. Ehret and Posnansky ; Blench and Spriggs ; Blench ). Within the limited scope of this chapter I can only offer some guidelines arising from my personal experiences with interdisciplinary approaches to Africa’s past. A first fundamental issue is the importance of direct collaboration between scholars of different disciplines who have a perfect command of their own body of evidence and are able to make a judicious assessment of its historical significance. To return to the Bantu Expansion, attempts to integrate linguistic and archaeological data have too long been a question of archaeologists relying on a lexicostatistically based subgroup X to link their ceramic tradition Y to an ancestral community of Bantu speakers and linguists using the same or another pottery group to trace the dispersal of a specific Bantu subgroup without a true interaction between scholars of both disciplines. Moreover, linguists and
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
archaeologists have for long time been in an unequal relationship with historians who turned their data into sweeping reconstructions of Africa’s past. These grand historical narratives—no matter how inspiring they may be—often do not stand the test of archaeological and linguistic proof. Without wanting to deprive historians of the role they have to play, certainly when historical records are in play, it is important that archaeologists and linguists really work together to put their own—independently obtained—bits and pieces together and reflect together on the historical significance of their combined data. A good example is the combination of the archaeo-botanical discovery of first millennium pearl millet finds in southern Cameroon and the fact that western Bantu pearl millet vocabulary is much older than previously thought, which made rethinking of prehistoric Central African agriculture necessary (Kahlheber et al. ; Neumann et al. ). Secondly, it is important that archaeologists and linguists mutually benefit from their specific assets. As I discussed in section . above, archaeology is blessed, for instance, with the possibility of absolute dating of which linguists can prudently take advantage to integrate the language changes they observe in an approximate chronological framework. Alternatively, linguistics is also capable of showing us historical pathways where archaeology hits its inherent limits. Where archaeology is intrinsically limited to material traces of the past, lexical reconstruction can also inform us on immaterial aspects, e.g. on how the ancestors of Bantu speakers used to conceive motherhood or power (Stephens ), or on material elements that vanished, e.g. food plants such as yams (Maniacky ) and certain wild trees (Bostoen ). These concluding remarks are not meant as a full-fledged archaeo-linguistic method, but they can be considered solid guidelines for future collaboration between linguists and archaeologists. It is only through real teamwork on specific historical cases and piecemeal engineering that we will be able to fill the large gaps in our knowledge of Africa’s past. Our understanding of early African history will be interdisciplinary or it will not be.
A I wish to thank Alexandre Livingstone-Smith and Gilles-Maurice de Schryver for their useful comments on a draft of this chapter. The usual disclaimers apply. The bibliographic references in this chapter were updated in . For a recent work on the interplay between linguistics and archaeology for the reconstruction of African history, see de Luna and Fleisher ().
R Andersen, H. (). ‘Introduction’, in H. Andersen (ed.), Language Contacts in Prehistory: Studies in stratigraphy. Amsterdam and Philadelphia, PA: John Benjamins, –. Antilla, R. (). An Introduction to Historical and Comparative Linguistics. New York: Macmillan. Ashley, C. (). ‘Towards a socialised archaeology of ceramics in Great Lakes Africa’, African Archaeological Review : –. Blench, R. (). Archaeology, Language and the African Past. Lanham, MD: Altamira Press. Blench, R., and Spriggs, M. (eds.). (). Archaeology and Language I: Theoretical and methodological orientations. London and New York: Routledge. Bostoen, K. (). Des mots et des pots en bantou: Une approche linguistique de l’histoire de la céramique en Afrique. Frankfurt am Main: Peter Lang.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Bostoen, K. (/[]). ‘Pearl millet in early Bantu speech communities in Central Africa: a reconsideration of the lexical evidence’, Afrika und Übersee : –. Bostoen, K. (). ‘Pots, words and the Bantu problem: on lexical reconstruction and early African history’, Journal of African History : –. Bostoen, K. (). ‘Shanjo and Fwe as part of Bantu Botatwe: a diachronic phonological approach’, in A. Ojo and L. Moshi (eds.), Selected Proceedings of the th Annual Conference on African Linguistics. Somerville, MA: Cascadilla Proceedings Project, –. Bostoen, K. (). ‘Wild trees in the subsistence economy of early Bantu speech communities: a historical-linguistic approach’, in D. Q. Fuller and M. A. Murray (eds.), African Flora, Past Cultures and Archaeobotany: Proceedings of the Fifth International Workshop for African Archaeobotany, London, – July, . Walnut Creek, CA: Left Coast Press. Bower, J. (). ‘The pastoral Neolithic of East Africa’, Journal of World Prehistory : –. Breton, R. J. L. (). Geolinguistics: Language Dynamics and Ethnolinguistic Geography. Ottawa: University of Ottawa Press. Bulkens, A. (). ‘Linguistic indicators for the use of calabashes in the Bantu world’, Afrikanistische Arbeitspapiere : –. Cornelissen, E. (). ‘On microlithic quartz industries at the end of the Pleistocene in Central Africa: the evidence from Shum Laka (NW Cameroon)’, African Archaeological Review : –. Dale, D., and Ashley, C. (). ‘Holocene hunter-fisher-gatherer communities: new perspectives on Kansyore using communities of western Kenya’, Azania : –. de Luna, K. M., and Fleisher, J. B. (). Speaking with Substance: Methods of language and materials in African history. Cham: Springer International Publishing. Derive, J. (). ‘L’oralité, un mode de civilisation’, in U. Baumgardt and J. Derive (eds.), Littératures orales africaines: Perspectives théoriques et méthodologiques. Paris: Éditions Karthala, –. Eggert, M. (). ‘The Bantu problem and African archaeology’, in A. Brower Stahl (ed.), African Archaeology: A critical introduction. Malden: Blackwell, –. Ehret, C., and Posnansky, M. (eds.) (). The Archaeological and Linguistic Reconstruction of African History. Berkeley, CA: University of California Press. Gamble, C. (). Archaeology, the Basics. London: Routledge. Gregersen, E. A. (). ‘Linguistic seriation as a dating device for loanwords with special reference to West Africa’, African Language Review : –. Güldemann, T., and Stoneking, M. (). ‘A historical appraisal of clicks: a linguistic and genetic population perspective’, Annual Review of Anthropology : –. Heine, B. (). ‘Zur genetischen Gliederung der Bantu-Sprachen’, Afrika und Übersee : –. Heine, B., Hoff, H., and Vossen, R. (). ‘Neuere Ergebnisse zur Territorialgeschichte der Bantu’, in W. J. G. Möhlig, F. Rottland, and B. Heine (eds.), Zur Sprachgeschichte und Ethnohistorie in Afrika: Neue Beiträge afrikanistischer Forschungen. Berlin: Dietrich Reimer, –. Kahlheber, S., Bostoen, K., and Neumann, K. (). ‘Early plant cultivation in the Central African rain forest: first millennium pearl millet from south Cameroon’, Journal of African Archaeology : –. Lane, P. (). ‘The “moving frontier” and the transition to food production in Kenya’, Azania : –. Lavachery, P. (). ‘The Holocene archaeological sequence of Shum Laka rock shelter (Grassfields, Cameroon)’, African Archaeological Review : –. Maniacky, J. (). ‘Quelques thèmes pour “igname” en bantu’, in K. Bostoen and J. Maniacky (eds.), Studies in African Comparative Linguistics, with special focus on Bantu and Mande. Tervuren: Royal Museum for Central Africa, –. Maret, P. de. (). ‘Archaeology of the Bantu expansion’, in P. Lane and P. Mitchell (eds.), The Oxford Handbook of African Archaeology. Oxford: Oxford University Press, –. Möhlig, W. J. G. (). ‘Sprachgeschichte, Kulturgeschichte und Archäologie: Die Kongruenz der Forschungsergebnisse als methodologisches Problem’, Paideuma : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Mufwene, S. S. (). The Ecology of Language Evolution. Cambridge and New York: Cambridge University Press. Neumann, K., Bostoen, K., Höhn, A., Kahlheber, S., Ngomanda, A., and Tchiengué, B. (). ‘First farmers in the Central African rainforest: a view from southern Cameroon’, Quaternary International : –. Nurse, D., and Hinnebusch, T. J. (). Swahili and Sabaki: A linguistic history (with a special addendum by Gérard Philippson). Berkeley, CA: University of California Press. Orton, C., Tyers, P., and Vince, A. (). Pottery in Archaeology. Cambridge: Cambridge University Press. Pakendorf, B., Bostoen, K., and de Filippo, C. (). ‘Molecular perspectives on the Bantu expansion: a synthesis’, Language Dynamics and Change : –. Phillipson, D. W. (). ‘The Early Iron Age in eastern and southern Africa: a critical re-appraisal’, Azania : –. Phillipson, D. W. (). African Archaeology. Cambridge and New York: Cambridge University Press. Ricquier, B., and Bostoen, K. (). Stirring up the Porridge: How early Bantu speakers prepared their cereals. Frankfurt am Main: Africa Magna. Schadeberg, T. C. (). ‘Loanwords in Swahili’, in M. Haspelmath and U. Tadmor (eds.), Loanwords in the World’s Languages: A comparative handbook. Berlin: De Gruyter Mouton, –. Schoenbrun, D. L. (). A Green Place, a Good Place: Agrarian change, gender, and social identity in the Great Lakes region to the th century. Portsmouth: Heinemann. Stephens, R. (). A History of Motherhood, Food Procurement and Politics in East-Central Uganda to the Nineteenth Century. Ph.D. thesis. Evanston, IL: Northwestern University. Swadesh, M. (). ‘Towards greater accuracy in lexicostatistic dating’, International Journal of American Linguistics : –. Wotzka, H.-P. (). Studien zur Archäologie des zentralafrikanischen Regenwaldes: Die Keramik des Inneren Zaire-Beckens und ihre Stellung im Kontext der Bantu-Expansion. Cologne: Heinrich Barth-Institut.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.............................................................................................................
LANGUAGE AND ORATURE .............................................................................................................
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. . ¨
. W A
.................................................................................................................................. A narrative art has existed since time immemorial. However, it only came into the focus of Western observers—travelers, missionaries, social workers, administrators— during the first half of the nineteenth century, a period when Europeans and Americans for the first time got into closer contact with African societies, in particular with their languages and cultures. The Western people soon realized that narration was a dominant cultural feature throughout Africa and that the quality of the tales could easily compete with verbal art in Europe, America, or Asia. Expression of the positive estimation is a collection of forty-three fables of the West African Wolof (Roger ). It has been conceived in French verses following the stylistic tradition of La Fontaine’s famous edition of Aesop’s fables in . Only a few decades after Roger’s pioneering work, collections of narrative texts had become known from all regions of Africa. They were mostly accompanied by the original texts in the African mother tongues. Western collectors were keen on the originals for scrutinizing and describing the linguistic structures and inventories of the African languages, but they used them also for practical purposes in the domains of school education and religious promulgation. Few scholars exploited the narrative texts as valuable archives for the study of general African concepts on the origin of the world, life and death, and other basic questions of humankind (Koelle ; Schlenker ; Bleek ; Steere ; Schön /; Frobenius , ; Werner ; Baumann ; Andrzejewski et al. ). Western scholars who had been living with African societies soon realized that narrating fictional texts was not a rhetoric gift of the societies in general, but due to the special skills of individual narrators, men and women. This discovery created a new perspective of looking at African narrations as oral art. Although it means a contradiction in terms, in this perspective narratives were thereupon described as oral literature. Finnegan (: ) found the key argument why African narratives are correctly considered in this way: “Oral literature is by definition dependent on a performer who formulates it in words on a specific occasion—there is no other way in which it can be realized as a literary product.”
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. . ¨
The narrators are people with a natural rhetorical talent. They have learned their repertoire of plots and the way of storytelling by watching and listening to other storytellers within their societies. Like an actor in a play, a narrator usually performs his/her art within the frame of a given narrative tradition and scenario. A special case is the institution of the so-called griot in West Africa. The term originally denotes a professional narrator, singer, musician, and chronicler of the Manding-speaking cultures, who often is employed by a specific clan to praise the clan elders and to recite the history of the clan (see, e.g., Johnson and Sisòkò ). In contrast to usual narrators in Africa, griots undergo a long-lasting apprenticeship in the entourage of a senior and well-established griot. From the Mandingspeaking cultures the institution of the professional or semi-professional narrator has spread all over West Africa. In their cultures, griots have a high reputation and the privilege of criticizing bad rulers or social abuses and defects (Dumestre ).
. G
.................................................................................................................................. In a scholarly perspective, storytelling is a highly complex phenomenon that varies from society to society and from region to region. At a first glance, each African culture appears to possess its own tradition of narration with its own preferences for themes, values, and narrative stereotypes, and with individual stylistic conventions. However, at a superregional level there are surprisingly many features that African narrative cultures share. Narrative texts are usually conceived in prose language, yet with many cultural-inherent guidelines. For instance, fables and fairy tales are introduced and concluded by specific formulas that mark their fictitious character (Neethling ). Furthermore, it is a common feature that narrative prose texts are interrupted by songs—also a kind of bound language (cf. Turay and Möhlig ). Special genres with a limited regional diffusion like praise poetry or funeral dirges tend to be completely conceived in poetic language anyway (Finnegan : –; Opland ). In a few narrative cultures, like that of the Somali and Swahili, poetic language is even dominant in all genres of spoken art (Harries ; Andrzejeweski and Lewis ; Knappert ). Therefore, at a general comparative level the genres prose and poetry appear to be of little use in defining various types of narrative texts. African languages have their own terminologies for distinguishing different genres of spoken art (Möhlig ). However, the problems in following the indigenous terminologies are twofold: first, their defining criteria often differ considerably for similar labels used in neighboring narrative cultures and, second, the African terms are incongruent with the scholarly terminology used at the worldwide meta-level. Nevertheless, in all narrative literatures of Africa a basic distinction is made between documentary narrations (reports), on the one hand, and fictional narrations, on the other. Therefore, this feature appears to be appropriate for defining different types among the majority of narrative texts. Beside this basic distinction, there is a third category of genres that the narrators sometimes consider real, sometimes fictional, or a mixture of both categories. In Western terminology these borderline cases concern in particular anecdotes (see section ... below), legends (see section ... below), and epic narrations (see section ... below).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. S :
..................................................................................................................................
.. Documentary literature Documentary literature basically refers to reports of the real world, even if these reports sometimes include matters that from a Western standpoint are considered as superstitious or fantastic. The basic rule is that in cases of doubt as to whether a text or a section of it has to be considered as real or fictitious, the decisive factor is always the self-assessment of the individual author. Many narrative cultures in Africa have a special type of oral text that concerns history. Because of their temporal coherence, Western scholars have assessed some of them as chronicles, although their value as a historical source is often doubtful (Werner ; Prins ; Freeman-Grenville ; Vansina ; Rowe ; Coupez and Kamanzi ; KleinArendt ; Fleisch and Möhlig ). There is another type of documentary literature that has often been stimulated by Western scholars, missionaries, or colonial administrators. It concerns travelogues, autobiographies, so-called ethno-texts, i.e. descriptions of special handicrafts and hunting and fishing activities, including legal concepts and practices (Gutmann ; Westermann , ; Köhler –; Geider , ).
.. Fictional narrations The genre terminology of fictional narrations used in Western folklore studies has been transferred to African narratives since Western observers have taken note of them. This refers in particular to fables, animal tales, and fairy tales. It may be questioned whether fables and animal tales are different genres at all, since in both types of narration the main protagonists are animals. Yet, from an insider’s perspective, there are at least two arguments in favor of treating them as different genres. First, a fable in comparison to an animal tale consists of a conspicuously shorter, formula-like text expressing a message that is a generally accepted maxim. In this respect fables are like proverbs that are verbally wrapped into a scenario of practical application (see also Bemile, chapter of this volume). In contrast, an ordinary animal tale consists of a significantly longer text. Like a fairy tale, it may even carry several messages. Second, the social contexts of performance are different. A fable always pursues pedagogical goals, whereas animal stories are mainly performed for entertainment.
... Fables In view of the shortness of the text, the narrative structure of a fable is comparatively straightforward, consisting of the following three to four sections (Jason ; Bremond ; Möhlig ):
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. . ¨
() Exposition: this is a short presentation of the topic, the protagonists, and the scenery. () Buildup phase: it describes in a causal sequence of sentences an offence against the maxim. () Climax: in one or two sentences the narrator formulates the consequences of the offence (e.g. punishment, damage to the offender). () Eventually an explicit formulation of the message—sometimes in the form of a proverb—may conclude the narration. In many narrative cultures, fables possess a traditional textual form, not leaving much room for the fantasy of the narrators.
... Animal tales Like fables, animal tales are characterized by stereotyped animal characters that act like human beings and often even interact with human protagonists. Specific animals are the epitomes of specific human character features. For instance, in West Africa the spider represents a trickster, i.e. a cunning personality that easily breaks social taboos, is reckless, and is often endowed with magic power. In southern Africa a similar role is played by the tortoise or in East Africa by the hare. The antagonists of the trickster often are the lion or the elephant. They represent the highest authorities of the society, namely the king or the elder of the group. They are considered powerful but stupid. Therefore, they can easily be tricked by the trickster. Another character, the hyena, stands for a stupid and voracious individual. Also, helper animals are known. They are played by a dove, gazelle, fly, or giraffe. In the moment of highest danger or difficulty of the hero protagonist, they appear in the tale and help him/her out of any misery (Bleek ; Lindblom ; Baumann ; Finnegan : –; Paulme ). The narrative structure of animal stories is not very rigid and therefore a playground for the personal fantasy of the narrators. It may contain breaks of the narrative flow, flashbacks, repetitions, and even the late introduction of new protagonists and scenarios. The main strategy of the narrators of animal stories is to keep the audience in suspense. Animal stories are conceived in prose language with the possibility of poetic insertions.
... Fairy tales The boundaries between animal stories and fairy tales are fluid. Therefore, in the terminologies of African languages they are often named by the same label. Their protagonists are usually human beings who are endowed with supernatural features. They may even have the gift of changing into the outward appearance of an animal or a ghost to pursue their criminal or strange goals (Calame-Griaule ). If they practice cannibalism either in a human or an animal appearance, they are referred to as ogres. Ghosts, wizards, sorcerers, and supernatural beings belong to the usual repertoire of fairy tales. The plots usually follow traditional types or motifs like “the daughter who wants to marry a husband of her own choice” (Görög-Karady ) or “children born from eggs” (Schmidt ). Fairy tales are usually long narrations, sometimes filling a whole evening session of several hours. An experienced storyteller invited to entertain guests at an evening party may skillfully connect several stories in combining the topics of different stories or by creating a framework story, whereby the aim of the narrator will be to increase the tension
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
of the buildup phase as slowly as possible to keep the audience in suspense up to the end of the story (cf., e.g., Turay and Möhlig : –). This genre is mainly conceived in prose but with songs inserted. In this kind of long narration the participation of the audience is almost obligatory. The audience may be encouraged by the narrator to sing the refrains of songs or to answer questions that the narrator asks them. Experienced audiences may take the initiative in heckling the narrator by cheering or provocative interruptions. In many parts of Africa until recently telling fairy tales appeared to be the most preferred entertainment for evening leisure time.
.. Borderline narrations between truth and fiction The boundaries between the real and the fantastic sphere may be blurred in one or another genre, particularly in anecdotes, legends, and epic narrations.
... Anecdotes In many African cultures short bizarre texts are narrated to amuse the audience and during a longer session to keep its attention through suspense. This kind of prose text requires not only a skillful and experienced narrator, but also a creative mind to be able to invent spontaneously new plots that have nothing to do with the plot of the main story. Their only recognizable purpose is to cheer up the listeners. If such a short text deals with personal experiences of the narrator, it may be termed anecdote. As Schmidt () in her important work demonstrates, the genre of anecdote may also fall under the category of report, since the autobiographical texts are meant to be true, although the style may be exaggerated to let it appear more witty or attractive.
... Legends Some African cultures have traditions of heroes living in ancient times where the conditions and circumstances of life did not differ much from those of today (Bascom : ). These topics are narrated in more or less standardized prose texts named legends by Western scholars. The narrative structure of these usually long texts follows a historical line from the birth of the hero up to his glorification or often tragic death (Seifert ). Some episodes within such heroic life stories may be fictional, for instance the circumstances of the hero’s birth; however, other episodes may be real history (Schmidt : ). A narrator usually marks his fairy tales as fictional. However, on other occasions the same narrator may present heroic plots in the form of true historical accounts. In such cases, it is often difficult to decide whether he/she only applies a stylistic device or whether he/she believes in the historical reality of his/her account. In this respect, legends are also borderline cases between truth and fiction.
... Epic narrations Another Western term that has been imported into Africa is epic. It refers to long historical accounts that are recited by professional narrators, so-called griots (see section . above),
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. . ¨
often with the support of traditional musical instruments and musicians. The style is predominantly poetic. Like the Western genre of the historical novel, they are a mixture of real and fictitious elements. This kind of oral art is mainly to be found in West Africa. Famous examples are the Sundjata Epos (Johnson and Sisòkò ; Cissé and Kamissoko ) and the Epic of Segu (Dumestre ). Yet epics also occur in other parts of Africa. For instance, the epic text known as the Lianja Epos belongs to the narrative culture of the Mongo of the Democratic Republic of Congo (De Rop ). Also, in Somalia and along the Swahili coast of East Africa several epic texts are known. These are ascribed to Arabian influence (Harries : –; Andrzejewski and Lewis ; Tomalcheva ; Andrzejewski ). Epic narrations, like legends, often function as media of ethnic integration, irrespective of whether the texts are perceived as true or fictional.
. M A
.................................................................................................................................. Stimulated by Western folkloristic studies (Abrahamsson ; Aarne and Thompson ), a discourse on intercultural types and motifs has also been started for narrative cultures in Africa. It has led either to motif and type indexations (Clarke ; Lambrecht ; Bremond ; Arewa ; Schmidt ; Schott ; Klipple ; Geider ) or to the portraying of typical narrative themes and characters (Evans-Pritchard ; Geider ; Schmidt , ). The French scholarly tradition in particular developed a comparative method based on the areal distribution of narrative themes like “the disobedient daughter”, “the unlike twin boys”, “the mothers eating their children”, or “the hero who created himself” (CalameGriaule ; Paulme ; Görög-Karady ). The French tradition also spread to Germany (Schott , ; Seifert ). The latest development so far is characterized by an interdisciplinary approach of comparing specific motifs and narrative themes within larger coherent areas. These studies are based on ethnological and archaeological methods (Kose and Seifert ). Other scholars focus on the performance aspect of oral art, the personalities of the narrators, and the social contexts and the strategies of narrative performances. The social contexts and the repertoires of narrative art are often explored in the historical perspective of gaining new insights into the pasts of the narrating societies (Klein-Arendt ; Geider ).
R Aarne, A., and Thompson, S. (). The Types of the Folktale, nd edn (st edn ). Helsinki: Helsingin Liikekirjapaino Oy. Abrahamsson, H. (). The Origin of Death: Studies in African mythology. Uppsala: Almkwist and Wiksell. Andrzejewski, B. W. (). ‘Somali literature’, in B. W. Andrzejewski et al. (eds.), –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Andrzejewski, B. W., and Lewis, I. M. (). Somali Poetry: An introduction. Oxford: Oxford University Press. Andrzejewski, B. W., Piłaszewicz, S., and Tyloch, W. (eds.) (). Literatures in African Languages: Theoretical issues and sample surveys. Cambridge: Cambridge University Press. Arewa, E. O. (). A Classification of Folktales of the Northern East African Cattle Area by Types. New York: Arno. Bascom, W. R. (). ‘Folklore research in Africa’, Journal of American Folklore : –. Baumann, H. (). Schöpfung und Urzeit des Menschen im Mythus der afrikanischen Völker. Berlin: Dietrich Reimer. Baumann, H. (). ‘Das Tier als Alter Ego in Afrika’, Paideuma : –. Bleek, W. H. I. (). Reynard the Fox in South Africa; or Hottentot Fables and Tales: Chiefly translated from original manuscripts in the Library of His Excellency Sir George Grey. London: Truebner. Bremond, C. (). ‘Principe d’un index des ruses’, Cahiers d’Études Africaines : –. Bremond, C. (). ‘Morphologie d’un conte africain’, Cahiers d’Études Africaines : –. Calame-Griaule, G. (). ‘L’homme-hyène dans la tradition soudanaise’, L’Homme : –. Cissé, Y. T., and Kamissoko, W. (). La grande geste du Mali des origines à la fondation de l’Empire. Paris: Karthala–Arsan. Clarke, K. (). A Motif-Index of the Folktales of Culture Area V: West Africa. Ann Arbor, MI: University Microfilms. Coupez, A., and Kamanzi, T. (). Littérature de cour au Rwanda. Oxford: Clarendon. De Rop, A. (). Lianja, l’épopée des Mongo. Brussels: Koniklijke Acadamie voor Overzeese Wetenschapen. Dumestre, G. (). La geste de Ségou. Paris: Armand Colin. Evans-Pritchard, E. E. (). The Zande Trickster. Oxford: Clarendon. Finnegan, R. (). Oral Literature in Africa. Oxford: Clarendon. Fleisch, A., and Möhlig, W. J. G. (). The Kavango Peoples in the Past: Local historiographies from northern Namibia. Cologne: Rüdiger Köppe. Freeman-Grenville, G. S. P. (). ‘Historiography of the East African coast’, Tanganyika Notes and Records : –. Frobenius, L. (). Das Zeitalter des Sonnengottes. Berlin: Dietrich Reimer. Frobenius, L. (). Kulturgeschichte Afrikas: Prolegomena zu einer historischen Gestaltlehre. Zurich: Phaidon. Geider, T. (). Die Figur des Oger in der traditionellen Literatur und Lebenswelt der Pokomo in OstKenya. vols. Cologne: Rüdiger Köppe. Geider, T. (). ‘Early Swahili travelogues’, in W. Graebner (ed.), Somoko. Popular culture in East Africa. Amsterdam and Atlanta, GA: Rodopi, –. Geider, T. (). ‘Indogene Ethnographien in der Swahili-sprachigen Dokumentarliteratur’, in T. Geider and R. Kastenholz (eds.), Sprachen und Sprachzeugnisse in Afrika: Eine Sammlung philologischer Beiträge Wilhelm J.G. Möhlig zum . Geburtstag zugeeignet. Cologne: Rüdiger Köppe, –. Geider, T. (). Motivforschung in Volkserzählungen der Kanuri (Tschadsee-Region). Ein Beitrag zur Methodenentwicklung in der Afrikanistik. Cologne: Rüdiger Köppe. Görög-Karady, V. (). Le mariage dans les contes africains: Études et anthologie. Paris: Karthala. Gutmann, B. (). Das Recht der Dschagga. Munich: C. H. Beck. Harries, L. (). Swahili Poetry. Oxford: Clarendon. Jason, H. (). ‘A model for narrative structure in oral literature, with a comment by D. Bremond’, in H. Jason and D. Segal (eds.), Patterns in Oral Literature. The Hague: Mouton, –. Johnson, J. W., and Sisòkò, F.-D. (). The Epic of Son-Jara: A West African tradition. Bloomington, IN: Indiana University Press.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. . ¨
Klein-Arendt, R. (). ‘Liongo-Fumo: Eine ostafrikanische Sagengestalt aus der Sicht der Swahili und der Pokomo’, Afrikanistische Arbeitspapiere : –. Klein-Arendt, R. (). Gesprächsstrategien im Swahili: Linguistisch-pragmatische Analysen von Dialogtexten einer Stegreiftheatergruppe. Cologne: Rüdiger Köppe. Klipple, M. A. (). African Folk Tales with Foreign Analogues, nd edn (st edn ). New York and London: Garland. Knappert, J. (). Four Centuries of Swahili Verse: A literary history and anthology. London: Heinemann. Koelle, S. W. (). African Native Literature; or Proverbs, Tales, Fables, and Historical Fragments in the Kanuri or Bornu Language. London: Church Missionary House. Köhler, O. (–). Die Welt der Kxoé-Buschleute im südlichen Afrika: Eine Selbstdarstellung in ihrer eigenen Sprache. vols. Berlin: Dietrich Reimer. Kose, E., and Seifert, M. (). ‘The history of iron working at the central Kavango: an attempt at interdisciplinary collaboration’, in W. J. G. Möhlig, O. Bubenzer, and G. Menz (eds.), Towards Interdisciplinarity: Experiences of the long-term ACACIA Project. Cologne: Rüdiger Köppe, –. Lambrecht, W. (). A Tale Type Index for Central Africa. Ph.D. thesis. Berkeley, CA: University of California. Lindblom, G. (). Kamba Folklore. Vol. : Tales of Animals with Linguistic, Ethnographical and Comparative Notes. Uppsala: Appelberg. Möhlig, W. J. G. (). ‘Überlegungen zur einer kulturinternen Analyse oraler Literatur in Afrika mit Beispielen der Dciriku vom mittleren Kavango (Namibia)’, Afrika und Übersee (): –. Möhlig, W. J. G. (). ‘The architecture of Bantu narratives: an interdisciplinary matter. Analysis of a Gciriku text’, in A. Traill, R. Vossen, and M. Biesele (eds.), The Complete Linguist: Papers in memory of Patrick J. Dickens. Cologne: Rüdiger Köppe, –. Neethling, S. J. (). The interpretation of opening–closing formulas in Xhosa folktales’, South African Journal of African Languages, Suppl. : –. Opland, J. (). Xhosa Oral Poetry: Aspect of a black South African tradition. Cambridge: Cambridge University Press. Paulme, D. (). ‘Typologie des contes africains du décepteur’, Cahiers d’Études Africaines (): –. Prins, A. H. J. (). ‘On Swahili historiography’, Swahili (): –. Roger, J.-F. (). Fables sénégalaises, recueillies dans l’ouolof et mises en vers français. Paris: Neveu, Didot, and Ponthieu [Microfiche, Paris ]. Rowe, J. A. (). ‘Myth, memoir and moral admonition: Luganda historical writing (–)’, Uganda Journal : –. Schlenker, C. F. (). A Collection of Temne Traditions, Fables and Proverbs, with an English translation, with a Temne–English vocabulary. London: Church Missionary Society. Schmidt, S. (). Katalog der Khoisan-Volkserzählungen des südlichen Afrika—Catalogue of the Khoisan Folktales of Southern Africa. Vol. : Quellen und Register—Sources and register. Vol. : Die Erzählungen—The tales. Hamburg: Helmut Buske. Schmidt, S. (). Sagen und Schwänke in Afrika: Erzählungen der Damara und Nama. Cologne: Rüdiger Köppe. Schmidt, S. (). Scherz und Ernst: Afrikaner berichten aus ihrem Leben. Cologne: Rüdiger Köppe. Schmidt, S. (). Tricksters, Monsters and Clever Girls: African folktales—texts and discussions. Cologne: Rüdiger Köppe. Schmidt, S. (). Children Born from Eggs: African magic tales—texts and discussions. Cologne: Rüdiger Köppe. Schmidt, S. (). The Forgotten Bride: International tale types in Namibia—texts and discussions. Cologne: Rüdiger Köppe. Schön, J. F. (/). Magána Hausa: Native Literature; or Proverbs, Tales, Fables and Historical Fragments in the Hausa Language. vols. London: Society for Promoting Christian Knowledge.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Schott, R. (). ‘Bericht über laufende Forschungen zur Motivanalyse afrikanischer Erzählungen im Seminar für Völkerkunde der Universität Münster’, Fabula : –. Schott, R. (). Bulsa Sunsuelima—Folktales of the Bulsa in Northern Ghana. Series S: Folktales of the Supernatural. Vol. : Tales of the Sky-God (Wen, Naawen). Münster: LIT. Seifert, M. (). Derjenige, der sich selbst erschaffen hat . . . Motivuntersuchungen zu Heldenerzählungen aus Nordnamibia und Südangola. Cologne: Rüdiger Köppe. Steere, E. (). Swahili Tales as Told by Natives of Zanzibar. London: Society for Promoting Christian Knowledge. Tomalcheva, M. (). ‘The Arabic influence on Swahili literature: a historian’s view’, Journal of African Studies (): –. Turay, A. K., and Möhlig, W. J. G. (). Temne Stories. Cologne: Rüdiger Köppe. Vansina, J. (). De la tradition orale: Essai de méthode historique. Tervuren: Musée Royal de l’Afrique Centrale. Werner, A. (). Myths and Legends of the Bantu. London and Bombay: Harap. Westermann, D. (). ‘Kindheitserinnerungen des Togonegers Bonifatius Foli: in der Ge-Mundart niedergeschrieben und übersetzt’, Mitteilungen des Seminars für Orientalische Sprachen zu Berlin : –. Westermann, D. (). Afrikaner erzählen ihr Leben. Essen: W. Girardet.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
.
. I
.................................................................................................................................. L is a system established by a people to communicate their thoughts, concepts, fears, hopes, desires, and other feelings. It is a window through which one can view the psyche of the speakers. It also serves as a vehicle of social interaction and cultural transmission. Language is basically oral and transient. It is also conventional and arbitrary. It may be supported by writing, a mnemonic, and more permanent means of communication. Peoples all over the world use language to preserve their history and their culture orally and in writing. Oral communication both predates and supersedes written communication by thousands of years, in that humans communicated orally long before they developed writing systems to communicate and to preserve verbal information or data. Several literary genres are employed in communication, e.g. narrative, drama, and poetry.1 This chapter highlights the oral rendering—orature—of these genres. It, however, focuses on the linguistic and literary aspects employed in the use of proverbs. It also indicates how the language of proverbs often exposes the psyche and ways of expressing thoughts of a people and how human beings may think in the same way but express themselves in different ways due to their historical and cultural heritage. Thus, it also conveys certain universal truths and lessons about life. However, a portion of the work tends to be eclectic, especially with regard to proverbs juxtaposed as containing concepts that depict universal truths.
. H
.................................................................................................................................. For many centuries, the speakers of different languages handed down their historical events and cultural riches through oral tradition which depended at that time on memory. Until 1
The major genres of literature are drama (which comprises tragedy, comedy, and all manner of plays), poetry (consisting of, for example, epic, lyric, ode, hymn, ballad, chant, elegy, and sonnet), and prose (referring to, e.g., short stories, narratives, and novellas). These genres are, nevertheless, interlinked, and can be mixed in a literary piece of work, especially in their oral choice of language or expressions.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
such events and riches were compiled into texts, memory was considered as the sine qua non of the vehicle of transmitting knowledge and was much cherished, as Aeschylus, Aristotle, Maurice Baring, and Basile would respectively have it: “Memory is the mother of all wisdom”, “Memory is the scribe of the soul”, “Memory is the greatest of artists and effaces from your mind what is unnecessary”, and “Memory is the cabinet of imagination, the treasury of reasons, the registry of conscience and the council chamber of thought”. Thanks to oral tradition and writing systems we have information on the ancient world and its civilizations, e.g. those of Egypt, Mesopotamia, the Hebrews, Phoenicians, Greeks, Persians, and Romans. We have further obtained knowledge about the Medieval Age, the Age of Enlightenment, and about slavery and the slave trade.2 Through the same means, knowledge has also been handed down to us on the origins of various peoples of the world and their gods,3 advances in civilization among the various peoples of the world, and the development of nation states. Even though we may now not accept totally some of the oral stories, we may have believed in many of them through certain degrees of evidence.4 Proverbs are, indeed, innovative and are not necessarily time-bound, since they are capable of linking the past with the present in order to keep past experiences alive. Nonetheless, owing to lack of space it will only be possible to discuss a few of the abovementioned cases that are most relevant to our discussion of orature with special reference to proverbs.
. C
.................................................................................................................................. This chapter defines “culture” as the whole gamut of human behavior, including entities like language, literature, religion, concepts, morals, values, festivals, family, marriage, political and legal systems, material objects, art, codes, and all manner of rites or rituals in a society. There is no gainsaying that there exist myriads of cultures in the whole world. Also, these cultures possess common patterns or universals and peculiarities or specificities. As in the case of historical heritage, these cultural aspects have been handed down by various peoples from generation to generation through oral tradition and serve as the cultural heritage for generations.
2
Slavery and slave trade were very lucrative for Europeans, especially in the nineteenth century. Certain peoples in Africa, e.g. the Dagara people, captured in their proverbs some of the atrocious activities and effects. 3 Most cultures at first believed in many gods, as exemplified by the Greeks and Romans, who had many gods representing different forces of nature and occupation, for instance, Greek religion had Zeus, Poseidon, and Athena, while Roman religion had Jupiter (god of the sky and chief god), Neptune (god of the sea), and Minerva (goddess of wisdom and the arts), etc. The existence of One God, however, was first mooted by an Egyptian pharaoh. The belief was crystalized and spread more widely by the Jews through their religion. 4 Indeed, Clement (: ), in giving an example of excavations and stories on ancient Crete, opines that “It is not wise to regard the old stories and legends of any country, however fantastic and unreal they may seem, as altogether false. They often contain a kernel of truth.”
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
. O
.................................................................................................................................. Oral tradition is the total pattern of customs, beliefs, and practices embedded in the historical and cultural heritage that were not originally written down but were transmitted orally from one generation to another. There are many forms of oral tradition, some of which are recollections or narratives of historical events, poetry, epithalamiums, folk tales, maxims, philosophical observations, sayings, proverbs, magical spells, ballads, songs, chants, and religious instructions. In general terms, oral tradition may be coterminous with oral culture or oral lore. Despite the current prevalence of radio, television, and newspapers in many cultures, which have led to the partial decline of oral tradition among several groups, many aspects of oral tradition are still manifested in a vibrant manner, especially in rural areas, through rituals, festivals, rites, ceremonies, songs, games, and regular activities of secret societies, associations, or groups.
. L
.................................................................................................................................. Apart from the definition of language given above it should be added that language can demonstrate the human capacity for acquiring and making use of complex and versatile communication systems. Language can also be used wrongly to pervert communication. Note what Socrates says about the use of false words: “False words are not only evil in themselves, but they infect the soul with evil.” Natural language is deeply entrenched in human culture. It is, therefore, not only used to disseminate and to share information but also to signify social and cultural issues, e.g. group identity, social stratification, social grooming, and also to entertain the members of the various groups. There is a great variety of linguistic diversity in the world.5 This linguistic variety is also reflected in the creation, use, and spread of proverbs. Apart from natural languages, there are artificial communication systems specifically constructed for computer programming and other artificial systems. The scientific study of language, called linguistics, began with the language analysis of Greek philosophers. In the course of this chapter due reference will be made to certain linguistic aspects that are deemed relevant to the present topic. Examples of such aspects are sound systems, grammar, semantics, registers or varieties, pragmatics, paralinguistic features (cf. Dako et al. : ), and foreignisms in a language.6
It is estimated that , languages are spoken by humans today. Some salient examples are often termed “World Languages”. The most popular ancient world languages, such as Persian, Greek, Hebrew, Hindi, Chinese, and Latin have to some extent ceded their prime positions to the current most popular modern world languages, for instance, English, French, Spanish, German, Portuguese, Swahili, and Hausa. 6 These aspects are further categorized under sound systems (phonetics, phonology); grammar (syntax, structure); lexical units, semantics (semantic space, semantic relationship); registers or varieties 5
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
. L
.................................................................................................................................. The term “literature” stems from the Latin word littera(e)—everything that is written, and refers to, e.g., newspaper articles, scientific articles in journals, magazines, books, and scripts written on cloth and even on walls. Literature has genres the major ones of which are drama, poetry, and prose (cf. n. above). These genres are, nevertheless, interlinked, and can be mixed in a literary piece of work, especially in the oral choice of language. Various aspects of these genres, e.g. mystery stories, proverbs, riddles, litotes, comparison, legends, myths, folklore, rituals, recitations, figures of speech, rhyme, lyrics, and climax, can also be found in their oral forms—the major channels of communication, that justify the term “oral literature” or “orature”.
. O
.................................................................................................................................. Oral literature or orature has been described in many ways, for instance, as folklore, folkloric literature, and traditional literature or as an aspect of any of them. In this chapter, we shall use the term “orature” in reference to all the available descriptions. The major characteristics of orature are the oral delivery of information—the spoken word—and the fact that more can be communicated orally than by writing. Also, whereas writing and reading may involve mainly three senses, i.e. sight, touch (e.g. in Braille), and feeling, orature involves all the five senses, i.e., sight, taste, hearing, touch, and smell, are employed to enhance performance. Apart from that, orature can be rightly called, as already mentioned, the most ancient, most complex, and richest channel of communication. The genres of orature intimated above are exemplified in ancient and modern art that are now documented in the works of oral narrators, for instance, in Homer’s Iliad and Odyssey, Aesop’s Fables in the fourth century , Equiano Olaudah’s Travels, Herodotus’ historical analyses, the African griots’ narratives, Arabic rāwīs, Indian Vedas, Yoruba Babalawo, Bàgr Secret Society, singing and dance groups of the Dagara people of Ghana, Burkina Faso, and Côte d’Ivoire, the Akpalu of the Ewe of Ghana, Togo, and Benin, voodos, the funeral dirge singers of certain peoples of the world, and shrines representing the gods of different natural und supernatural forces, churches, choirs of ancient and modern religions such as Christianity, Judaism, Islam, Confucianism, Hinduism, and Buddhism.
(language codes in given situations: formal, informal casual, e.g. jargon, slang, vulgarism, taboo, ideolects, sociolects, dialects/regional varieties/regional dialects, pidgins, creoles, use of language, and cultural differences); pragmatics (code-mixing, code-switching); paralinguistic features (in face-to-face interactions that assist us in following the communication that is going on; cf. Dako et al. : ); and foreignisms in a language (foreign expressions that are localized, indigenized, or nativized, loanwords, foreign words, foreign translations, and transliterations).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Some of these genres are supported or emphasized by paralinguistic features, e.g. body language. Owing to the transient nature and multiplicity of verbal art, however, it is not possible to describe all these aspects. In the ensuing pages we shall focus on proverbs.
. P
.................................................................................................................................. Generally, it is difficult but not impossible to indicate the origin of the terms used by cultures to designate “proverb”. Such terms are often ascribed to one person, one race, one ethnic group, one people, or one culture. For example, it is popularly said, as far as Europe is concerned, that Aristotle played a great role in the development of the Greek term paroimίa [paroimia] that became the source of the Latin term proverbium, and which has served as a stem for expressions in many other European languages.7 Proverb has many definitions, but a very comprehensive definition that is most relevant to our discussion here is as follows: “A proverb is a short, generally known sentence of the folk which contains wisdom, truth, morals, and traditional views in the metaphorical, fixed and memorisable form and which is handed down from generation to generation” (see Mieder : , : in Wikipedia ). As shall be shown presently, it is possible to trace back to individuals, groups, peoples, or cultures the origin, creation, spread, or popularity of certain proverbs. For instance, commoners of all walks of life, street vendors, personalities, families, and nations could be the source of the creation or spread of proverbs. Many sources of proverbs can also be found in songs, advertisements, commercials, movies, riddles, parables, literary works, and deep reflections. In order to achieve the primordial objective of this chapter of discussing proverbs in relation to language and orature, it is imperative to indicate, where possible, the language employed in proverbs, the origin, creation, collection, and spread of proverbs, types of proverbs that exist, and their structures, styles, functions, and general impact.
.. Origin, creation, collection, and spread of proverbs All peoples of our five continents have served as sources of proverbs in one way or the other. Names of personalities that are easily mentioned currently are those who have so far been committed to writing. Examples of such names are Aristotle, Cicero, Socrates, Aesop, Homer, Vergil, Seneca, Juvenal, Confucius, Julius Caesar, Augustine of Hippo, Shakespeare, Goethe, Schiller, Samuel Johnson, Descartes, Chinua Achebe, N’gugi Wa Tiongo, Wole Soyinka, Kwame Nkrumah, and a long list of other philosophers, politicians, scientists, and authors, and the Bible, the Qur’an, the Talmud, and other sacred books. Indeed, many philosophers, politicians, musicians, griots, dirgers, fabulists, oral historians, and authors have, for example, also created or are responsible for the spread of many wise sayings, maxims, and proverbs. The term “proverb” actually stems from the Latin proverbium and the term “paroemiology” stems from the Greek παροιμία (proverb). 7
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
It goes without saying that peoples who have depended solely on oral tradition must have not properly documented the sources of their proverbs, but rather popularized them and made them anonymous and a common good. Also, some wise people whose names were not appended to their creations have been cast into oblivion. For instance, the following Chinese, Ciceronian, and popular English proverbs say respectively: “The palest ink is better than the best memory”, “Epistola non erubescit” (‘A letter does not blush’), and “A good memory is not so good as a little ink”. Examples of these sources will be mentioned in the course of the analysis of selected proverbs.
.. Types of proverbs The study of proverbs, known as “paroemiology”, seeks to categorize, classify, or define the structure of proverbs. There are, indeed, many types/genres and subtypes of proverbs that can only be discussed cursorily here. Proverbs are discussed here from two points of view: form and content.
... Form of proverbs As regards form, we may categorize proverbs under their linguistic features, literary styles, and numerical and onomastic features that are used to construct the proverb. With regard to content, we may categorize proverbs into, for example, historical, religious or spiritual, social, political, economic, educational, jurisdictional, and other aspects. These categorizations are by no means iron-clad but allow for overlapping of their scopes. .... Linguistic features here refer to phonetics, phonology, morphology, syntax, semantics, and lexis. Only issues referring to the above phenomena that are considered absolutely relevant to our discussion will be discussed here. ..... Certain phonetic and phonological phenomena occur or are noticed during the creation of proverbs; examples are: vowel harmony, onomatopoeia, homophony, homonymy, synonymy, and assonance. Vowel harmony occurs or is maintained in the formulation of proverbs in many languages. This means that in a lexeme [+ATR] vowels /i e o u/ appear only with advanced vowels, whilst [–ATR] vowels occur only with retracted vowels. In cases where the central vowel sound [a] is neutral, it occurs with any of the other vowels, especially in compound words. For instance, in the following Turkish proverb we see that suffixes are formed to suit Turkish vowel harmony: Acele işe şeytan karɩşɩr (‘The devil interferes with hurried work’ > Haste makes waste), where the [+ATR] vowels /i e/ occur with each other in işe, the [–ATR] /ɪ/ harmonizes with the [–ATR] /ɪ/, whilst the neutral vowel /a/ may go with the [+ATR] /e/ or the [–ATR] /ɪ/ as in Acele and karɩşɩr. Also, in the Dagara proverb Nimie ayi bε daŋ vuoli to-kole (tuo kʋɔr) pʋɔ ɩ (‘Two eyes cannot peep through a baobab gourd at the same time’), the [+ATR] vowels /i e u o/ occur with advanced vowels and the [–ATR] vowels /ɩ ε ʋ ɔ a/ also occur together in the lexeme or free morpheme (Bemile : ) and their immediate environment: kʋɔr pʋɔ ɩ. Similar phonological processes take place in other African, Asian, and European proverbs.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
..... Onomatopoeic words and expressions which tend to describe natural sounds and ideophones which describe non-auditive onomatopoeia-like words or expressions that describe feelings or impressions about certain phenomena occur in the formulation of proverbs just as one would find them in ordinary speech. Onomatopoiea is demonstrated in the following proverbs. Dagara: Xolxol (ol’ol) tʋlʋ paalε nyagε, bin k’a ɓaanε bʋɔlε dɩrbε (‘Eating food while it’s still hot makes you satisfied, while delaying and making it cool before eating it invites more mouths in the end’) (Bemile : ), in which case the word xolxol or ol’ol describes the sound made when one struggles to cool hot food in one’s mouth. Most ideophones in Dagara (as a tone language) are expressed through tone. High and middle tones may describe something likable, smallish, or insignificant, while low tone may indicate something big, grave, or atrocious. In Japanese ideophones appear in proverbs or proverbial expressions, e.g. Nanakorobi yaoki (‘Fall seven times and stand up eight times’) which indicates the need for doing something repeatedly or intensively, i.e. [nanakorobi], in order to succeed eventually. Proverbs with onomatopoeic and ideophonic sound effects are most useful in oratory and verbal performances. ..... , , Many proverbs in the world possess words which sound in the same manner in the language but may have different meanings. For example, in this version of the English proverb “What the eye sees not, the heart does not rue”, sounds [ru:] and means ‘regret, grieve’, just as another word also sounds [ru:] but means ‘herb of grace’. In the Latin proverb Arbor mala, mala mala (‘If the tree is bad the apples, too, are bad’) the last three words are pronounced similarly (though the vowel length is different: mala (‘bad’), but māla (‘apples’) but they have two different meanings, i.e. ‘bad’ and ‘apples’ respectively. Since the two English words “rue” and “rue” and the Latin words mala, māla are written and pronounced in the same way/similarly respectively, they are not only homophones but also homographs. However, in the Spanish proverb Olivo y aceituno, todo es una (‘It is one and the same thing’), and have different phonetic realizations, namely, [olivo] and [aθeituno], but both words have the same meaning, i.e. ‘olive’. In the German proverb Kurzes Gebet, tiefe Andacht (‘A short prayer is as good as a deep devotion’) the words and are pronounced [gəˈbe:t] and [ˈanˌdaxt] respectively. Each pair of words is written and pronounced differently, i.e., they are heterographs. Each pair of words, however, has the same or a similar meaning. They are, therefore, synonyms to a large extent. ..... Elision occurs very often in proverbs. In German, we hear and read Wes’ Brot ich ess’, des’ Lied ich sing’ (‘I sing the song of person who gives me bread’ ! ‘He who pays the piper calls the tune’) in which the sounds (underlined) are elided from wessen, esse, dessen and singe. A few types of elision occur in English proverbs in the form of contraction, i.e. of “cannot” (can’t), “is” (‘s), “is not” (isn’t), “do not” (don’t), “I am” (I’m). The word “never” also experiences elision in, for instance, the English proverb “Faint heart ne’er won fair lady”. ..... Some proverbs are formulated with words in which stressed vowel sounds within words, phrases, or sentences with different consonants are repeated
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
to create internal rhyming. This phenomenon is termed “assonance” and is exemplified in the following proverbs; English: “Haste makes waste”; Dagara: Pɔg faa san’ yaa, kyɔg nuu kε bɔ pɔg (‘It’s better to keep a bad wife and look for a better one than to desert her’); and Latin: Manus manum lavat (‘One hand washes the other’). Assonance is used in proverbs, proverbial expressions, and poetic language to render them artistic, aesthetical, and beautiful, especially to the ear. ..... / Certain morphological and syntactical processes occur in verse and proverb formulation. For instance, in Dagara serial verbs occur in normal speech as well as in proverbs, e.g. Tɩεrɩ bεr sa tɩεrɩ yel (‘To think and to keep the thought to yourself is better than to think and to divulge the thought’ ! ‘One must not disclose all one’s plans’). In German, words are combined very often, e.g. Bücherwurm (‘book’ + ‘worm’ = ‘bookworm’), miteinander (with + one + other = ‘with each other’), to achieve a different meaning. Words are also combined and hyphenated in the following Dagara proverb: Bε zɔrε nɩ “Bεlε-nεbε-waar”, bε bε zɔrε “Maa-n’ʋ-waar” ε (‘They are afraid of “They-areapproaching/coming”, they are not afraid of “I-am-approaching/coming”’) to form personified nouns. Note that many personal names emanating from Dagara proverbs are mostly not hyphenated. This phenomenon, no doubt, exists in many languages all over the world. The sentence structure of proverbs may be consistent with or different from normal sentence structure. It is, however, mostly pithier than in ordinary sentence structure. This section describes a few aspects of these sentence structures. The sentence structure discussed here consists of minor/irregular sentences and major/ regular sentences. There are many types of sentences, e.g. declarations, questions, directives, commands, exclamations, admonitions, and sentences with dummy subjects. Minor/irregular sentences lack a subject and a verb or predicate. Examples: Dagara Faarɩgyɔ! (‘Keep your tongue!’); Latin Nihil sine causa (‘There is nothing without cause’). Major/regular sentences have a subject and a predicate. Needless to say, most proverbs consist of major/regular sentences, as in the following proverb translated from Zulu: ‘When you bite indiscriminately, you end up eating your own tail’. Declarative sentences are statements, e.g. Akan: Maka, maka (‘What I’ve said, I’ve said’); Latin: Roma locuta, causa finita (‘Rome has spoken; the case is closed’). There are many proverbs in many cultures which are in the form of questions. Three types of questions manifested in proverbs are discussed here: yes-no questions, whquestions, and alternative questions. Yes-no questions consist mainly of tag questions in which the questioning part comes at the end of the sentence: ‘isn’t it?’, ‘would you?’. Question by tone (from Dagara): Naŋ-sob tεr’ɩ ba[a]? (‘Has a poor person any friend?’). Other proverbs use wh-questions containing the interrogative particles ‘what’, ‘where’, ‘who’, and ‘when’; for example, Dagara: Bʋnʋ be garʋ a ŋmɩntɔŋ bε baŋ? (‘What is on the roof terrace that the sun does not know?’); Nyɩnε zir-ŋmarε pag yelmiŋa? (‘Where can you find trust in a liar?’). An example of proverbs with alternative questions is Akan Yεkɔ anaa yεba? (‘Are we going or are we coming?’). The particles ‘better’ and ‘than’ that are used in constructing comparisons are also used in creating proverbs in order to illustrate a common trait or to demonstrate how something
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
or a certain trait outweighs something else. The following Dagara proverb may serve as an example: Bʋlaŋ sa togtog (‘A little is better than nothing at all’). Exclamations and questions seek the listener’s agreement rather than an answer: Thalatta! Thalatta! (‘The sea! The sea!’; shouts of joy of Greeks when they saw the Black Sea after their failed battle against the Persian Empire in ). Dagara: Bibile bε sa baa! (‘Isn’t a child more useful than a dog!’). Users of proverbs with rhetorical questions do not require an answer. The questions rather lead to deeper reflection. For instance, Dagara: Anʋ wule bibile wɔb? (‘Who shows a child an elephant?’); Latin: Quid est veritas? (‘What is truth?’); Akan: Yεbε wu nti yεnda? (‘Because we’ll die should we not sleep?’). In proverbs that are directives/commands the subject often does not appear but is implied in the verb, e.g. German: Trau, schau, wem! (‘Trust, [but] look, who!’; one should be careful whom one trusts); English: Take it or leave it!; Dagara: Tɩεrε kε ɩrε! (‘Think before you do!’); Admonitions are usually followed by the motivation or explanation behind a command or prohibition. The Bible quotes Jesus as saying: “Judge not, and ye shall not be judged, condemn not, and ye shall not be condemned; forgive and ye shall be forgiven”, where “Judge not . . . ”, “condemn not . . . ”, and “forgive . . . ” are all admonitions, while “ye shall not be judged”, “ye shall not be condemned”, and “ye shall be forgiven”, are, respectively, motivations. The Yoruba say: Kikan l’ó ríká mu, eni à bá fi inú hàn, je aáròkiri enia (‘The person to whom you must have told your secrets tells whatever he hears to all and sundry’). Proverbs with dummy subjects, which sometimes have no intrinsic meaning but are inserted to maintain a balanced grammatical meaning, are, English: There is no smoke without fire; Dagara: A vɩεl’a, a vɩεl’a (‘It’s good, it’s good’); Ŋmε lε nɩ ŋmεlε nɩ bε tuur taa (‘It’s like and like that are following each other’).8 Many paralinguistic features are used in verbal performance, especially when proverbs are employed in oratory and political banter. Examples of such features are gestures, mimic, stress, and tones. .... Proverbs using literary skills may be further classified under paradox, metaphor, antithesis, comparison, parallels, contrasts, epigrams, alliteration, irony, personification, synecdoche, innuendo, oxymoron, hyperbole, metonyms, synonyms and antonyms, rhythm, meter, pun, rhyme, climax, and ellipsis (cf. Bemile : –).9 Proverbs have subgenres as may be exemplified here: sayings, adages, aphorisms, bywords, maxims, mottos, saws, oracular utterances, truisms, popular sayings, wise sayings or precepts, didactic sentences, pithy sayings, Wellerisms, proverbial expressions, allusions to proverbs, and anti-proverbs or “perverbs”, which also express universal truths and are easily memorable. For want of space it is not possible to discuss all these phenomena here. A few of the most striking ones will be indicated.
8
Owing to lack of space it is not possible to discuss here more lexical aspects that exist in proverbs, such as lexical gaps, collocations, context, pragmatics, deixis, contextual information, euphemisms, and litotes (see Dako et al. : –). 9 Literary styles abound and vary in the creation of proverbs (cf. Bemile ).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
..... In many proverbs the information given appears to be contradictory or even absurd, but after closer analysis some truth and sense are found in the saying. For instance, people seem to accept poverty and death, even though they hurt, or are generally rejected. Such attitudes are exemplified in the following proverbs and their equivalents. English: Poverty is no sin; German: Armut schändet nicht (‘Poverty is not shameful’); French: La pauvreté n’est pas vice (‘Poverty is no vice’); Dagara: Naŋ bε ɩ faa ɩ (‘Poverty is no evil’); Akan: Kafoɔ didi (‘[Even] debtors eat’). ..... Metaphors abound in proverbs and quotations. They are exemplified in the following translations of the well-known biblical quotation: Hebrew: Yehiy ‘or; Greek: Genēthētō phōs; Latin: Fiat lux; and English: Let there be light, where ‘or, phōs, lux, and light are metaphors that dispel darkness or ignorance. ..... The Dagara people say that: “Even if you draw lines on your body [or tattoo yourself] (in the observation of sacred rites), you still do so with decorum [and beautiful lines].” Thus, rhythm and meter are beautiful literary styles in poetry. They are applied in their poetic forms in proverbs to embellish their sound and to enhance their message. Thus, many cultures couch their proverbs rhythmically and in metric forms, e.g., in the English proverb “He who pays the piper calls the tune” we notice iambs. Also, iambs can be found in the Latin proverb Adora quod incendisti, incende quod adorasti! (‘Worship what you have burnt, burn what you have worshipped’) where we have eight stressed syllables and eight unstressed syllables. However, in the German proverb Andere Länder, andere Sitten (‘Other countries, other morals’) we have dactyls, one stressed and one unstressed syllable in each word. ..... / No sense may be made out of puns per se, but the sounds of the words are often appealing to the ear. The following may serve as examples for proverbs that are expressed in puns. German: Einbildung ist auch eine Bildung (Bildung = ‘education’, Einbildung = ‘vanity’, thus: Vanity is also education); Eine Kuh macht Muh, viele Kühe machen Mühe (Kuh = ‘cow’; Muh = ‘moo’; Kühe = ‘cows’, and Mühe = ‘strenuous work’, thus: One cow moos, many cows cause much work). In these cases, the Bildung and Einbildung are almost homophonous, just as Kuh and Muh and Kühe and Mühe are, but their meanings are different and have no sense together in the proverbs. ..... Aphorisms also occur very often in proverbs. Two examples may be cited here: Greek Gnōthi seautón! (‘Know yourself!’), and Latin Carpe diem! (‘Make use of the day! Seize the opportunity!’). ..... Proverbs often emanate from parables (or anecdotes), in which case a proverb is preceded by the parable or vice versa. Within the proverb, also, the first line may serve as an illustration of the second line. Such proverbs may be termed
10
Many European languages exhibit rhythm and meter in their proverbs. The present author is still researching into rhythm and meter in Asian and African proverbs.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
parabolic proverbs. The following proverbs may serve as examples. Dagara: Teŋzu ŋmε lε (‘So is the world’, from the parable/fable of the race between the hare and the tortoise in which a queue of the tortoises outwit the hare and win a race); Charles de Gaulle: ‘Diplomats are useful only in fair weather. As soon as it rains they drown in every drop.’ ..... / Antithetical proverbs may be described as proverbs which have ideas that are parallel or contradictory and cannot be reconciled, but which finally converge to have one effect. The following proverbs exemplify this phenomenon. Greek: Dôs moi pā stô, kai tā gân kīnāsō (‘Give me somewhere to stand, and I will move the earth’, Archimedes): Dagara: Daŋ-tεrε bε laar ε, bɩεr-tεrε n’ʋ mɩ la (‘He who has first does not laugh/rejoice, it is he who has last that laughs/rejoices’); English: He who laughs last, laughs best. ..... Repetition of expressions is very common in proverbs. Examples: Dagara: Puo zaamɩ, puo zaamɩ nɩ puo kɔbaarʋ (‘Saying “good morning” to a farm every day completes the farm work’); English: By little and little, the wolf eats up the goose; Amharic: ‘Little by little an egg will walk’; French: Petit à petit l’oiseau fait son nid (‘Little by little, the bird makes its nest’). ..... Rhyming occurs very often in proverbs, for instance in the following proverbs. Latin: Praemonitus, praemunitus (‘Forewarned is forearmed’); German: Einmal ist keinmal (‘One doesn’t count, once is no custom’). ..... Alliteration is a phenomenon which occurs in certain sounds that rhyme in, for instance, verses. Alliteration occurs in the formation of proverbs and famous quotations. This can be seen in the following examples. Dagara: Ba ba baa ba baa baabʋ (‘Their friends’ dog has not grown up like a dog’); Latin: Veni, vidi, vici (‘I came, I saw, I conquered’). ..... Dagara: Yelmiŋa-sob bε gungune i, zir-ŋmarε n’ʋ mɩ gungune (‘A truthful person does not sneak; it is a liar that sneaks’); German: Doppelt gemoppelt hält besser (‘The same thing done twice over is better’). ..... Parody occurs very often in the creation of proverbs, especially in antiproverbs, for instance: Dagara: Kur bε baŋ kε pεgr ben’ ʋ puori e (‘The tortoise does not know that it has a shell behind him’); English (anti-proverb): A friend in need is a pest indeed. ..... : “” “” Similes compare totally different concepts by using the particles “like” or “as”. They express the link between the illustration and the main point. Similes are found in many proverbs. German: Wie die Eltern, so die Kinder (‘Like the parents, so also are the children’); popular proverb: A fool in a high station is like a man on the top of a high mountain—everything appears small to him and he appears small to everybody. ..... Proverbs also use climax in expressing messages, such as in the following proverbs. Dagara: Fʋʋ wa kʋ kɔŋ fʋ nũú, ʋ mɩ yel kε nyɩ mɔ (‘If you give a leper your hand, he asks you to wrestle with him’); German: Gib ihm eine Handbreit, so nimmt er die ganze Elle (‘Give him an inch and he’ll take a mile’).
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.... () Other general features refer specifically to features that are not strictly linguistic or literary. ..... Proverbs are often created by the use of numerals, serial or ranking numbers, number of times, groups, and percentages. Examples: Chinese: ‘A hundred men may make an encampment, but it takes a woman to make a home’; Latin: Tertius gaudens (‘The third rejoicing’ ! ‘When two people fight, the third person rejoices’); English: Once bitten, twice shy; Dagara: Dεb nɩ ʋ gbεε ata, pɔg nɩ ʋ gbεε anaar (‘A man and his three times, a woman and her four times’).11 ..... In integral proverbs the concept of the last line integrates or completes the concept of the previous line. Examples of such proverbs are The Book of Proverbs (:): “The teachings of the wise are like a fountain of life; they will help you escape when your life is in danger”; Latin: Non scolae, sed vitae discimus (‘We learn not for school; we learn for our own life/living’). ..... Many proverbs express completely different thoughts under one theme. Such proverbs may be termed synthetic proverbs. For instance, under the theme “Hatred” we have some synthetic proverbs, as in The Book of Proverbs (:): “Anyone who hides hatred is a liar. Anyone who spreads gossip is a fool”. ..... Here “onomastics”12 is described as the study of personal or place names. In this case, it is said or assumed that personal or place names are used in the creation of proverbs. Examples are: Dagara: Aŋa n’a fʋ kule Bεkyile Dεr ɩ? (‘How then could you have married Bεkyile Dεr with this attitude of mind?’); English: Rome was not built in a day. In these proverbial expressions, “Bεkyile Dεr” is a personal name and “Rome” is a place name. ..... Many old proverbs have been perverted and made fun of, especially by young people and people who seem not to consider seriously the function of/or wisdom contained in proverbs. Therefore, such proverbs are sometimes called “perverted proverbs”, “anti-proverbs”, or “perverbs”. Hereinafter, we shall be using the more succint term “perverb”. Cf. the German proverb: Über Geschmack lässt sich nicht streiten (‘There’s no accounting for taste’), which is perverted to “Geschmacksache”, sagt der Affe und beißt in die Seife (‘ “A matter of taste”, says the monkey and bites into the soap’).13
Meaning: A first try that happens to fail is acceptable. A first time offence should be forgiven, but a second one will have consequences. Something that happened once might as well never have happened at all. Also, the number is a bad number in many cultures of the world. In the Dagara culture a man is allowed to try something only three times. He can, therefore, try only a third time more if he does not succeed the first or second time. A woman is, however, allowed to try four times if a first, second, or third time fails, i.e. Dεb nɩ ʋ gbεε ata, pɔg nɩ ʋ gbεε anaar. 12 This definition does not necessarily refer to onomastica, which is a term referring to, for instance, Wisdom lists in the Bible (e.g. Job and Psalm ) or any other Wisdom book on the attributes and deeds of God. 13 The term “perverb” here has been adopted from Mieder (). 11
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
... Content Content here is contrasted with form and refers to the purposes or functions of proverbs. Apart from passing knowledge and wisdom to future generations, proverbs are also used for a variety of purposes that can only be discussed in part here, for example, in order to say something urgently in a veiled way, to make discussions or conversations lively, and as a mark of a good orator. Proverbs of this nature play many functions, as exemplified by Dalfovo et al. (: ) as follows: they expose the faults so that people become aware of the faults of language; they assist people not only to be victims of the faults when committed by others, but not to commit these faults themselves; they provide moral lessons and serve as a challenge to make people think carefully and evaluate what they hear before accepting them.
Indeed, the Bible gives the purpose of the Book of Proverbs (:–) as follows: That men may know wisdom and instruction, understand words of insight, receive instruction in wise dealing, righteousness, justice and equity that prudence may be given to the simple, knowledge and discretion to the youth, the wise man also may hear and increase in learning and the man also may hear and the man of understanding acquire skill to understand a proverb and a figure the words of the wise and their riddles.
Whilst looking at the content of proverbs it would be imperative to take note of various concepts or precepts that can be termed “universal”. It is also worth noting that the same ideas, concepts, thoughts, desires, surprise, fear, acceptance, or disapproval expressed universally in different ways by people in their languages—such universal human conditions—can also be expressed through facial expressions, movement of eyebrows, lips, and other parts of the body in a similar or slightly different way. This can only or mainly be exhibited and enhanced during oral performance. For better clarification of the content of proverbs in sections ....–...., proverbs will be categorized into different functions. .... . Before discussing proverbial expressions on other issues it is relevant to see what proverbs say about orature (oracy) and writing (literacy) themselves. The following popular proverbs testify to the importance of oracy and proverbs: “In the beginning was the Word, and the Word was with God, and the Word was God . . . . And the Word was made flesh, and dwelt among us” (John :, ), “Don’t throw away the old bucket until you know the new one holds water.” In other words, it would be unwise to cast into oblivion the importance of proverbs or wise sayings, even in their oral form, in bringing us from generation to generation wisdom and truths until we can really do without the use of proverbs or wise sayings: doing without them sounds impossible at any rate.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
“Even Aesop had no remedy for superstition.” Many aspects of oral tradition are shrouded in secrecy and superstition that until now are impossible to do away with. “A teacher is better than two books.” In other words, however important writing may be it is believed that personal contact and verbal communication, i.e. oracy, still play a more important role than mere literacy. “The book gives knowledge but it is life that gives understanding (wisdom).” “The greatest clerks are not the wisest men.” This says that one does not necessarily live better or more wisely after acquiring book knowledge; one must put knowledge into practice. “A good storyteller must be able to lie a little.” This allows for versatility and creativity in oral renderings. Indeed, speakers who use proverbs when making their speeches, including verbal skills, sharpen and demonstrate their oratorical skills through the proverbs. The use of proverbs also serves to season discussions and conversations and to make them lively. To highlight the importance of the written word and content of the Qu’ran we are told that: “This Book, there is no doubt in it, is a guide to those who guard (against evil)” (Part I, Chap. , Sect. , Vs. ). The popular sayings “Books are preserved minds” and Litterae scriptae manent (‘Written letters/words endure’) also emphasize the importance of writing as a mnemonic aid to the dissemination of information to present and future generations. They also underline the necessity of committing certain information and words of guidance to writing before they are lost through the death of knowledgeable people, especially the older generation/elderly people. The following proverbs are self-explanatory: “A book is a garden carried in the pocket” (Arabian proverb). “A book that remains shut is like a block.” “A book is worthless (has little worth), yet makes the man” (Bemile, personal observation). “If you wish to learn the highest truths, start with the alphabet.” “A scholar’s ink lasts longer than a martyr’s blood” (Irish). “A good memory is not as good as a little ink.” We shall now turn our attention to other functions of proverbs, which are numerous but can only be exemplified here. .... All cultures create proverbs to commemorate certain important historical events that have deep connotations in their lives, e.g. migratory voyages, discoveries, interaction with other people, human relations, birth, death, success, failure, and wars. Examples are: Hebrew: “Who has no past, has no future”; Dagara: Bɔŋ-zagla ŋmεn’ ŋmɩnŋmɩn ʋ na wa kʋrε nɩbε a? (‘If an ordinary slave raider is that dangerous, then how will he be when he is permitted to kill people?’). Indeed, proverbs are innovative and are not necessarily time-bound, since they are capable of linking the past with the present in order to keep past experiences alive. Thus, the same proverbs may be used in an ancient and modern context, as is often the case.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
.... Different cultures have various ways of looking at or addressing themselves and other cultures. Sometimes they resort to praise or derision. Many clans and nationals are also joking partners and use proverbs in a jocular manner to express their sentiments to or about their counterparts. Proverbs are also created to conform with a people’s or country’s cultural heritage. The following proverbs may suffice as examples. They may also be derisive of other cultures/peoples. “Bargain like a gypsy and pay like a gentleman.” “It is hard to sit in Rome and fight with the Pope.” “Search knowledge be it in China!” (Arabian). “To lie about a far country is easy” (Amharic). “A lie about a foreign country that cannot be verified!” (Dagara). .... Proverbs which refer to people who talk much about the success of their enterprise but never achieve it are: Dagara: A zɩεr nɩε na nʋmε εkyε karε a aŋa? (‘Would a soup that is simmering so much be ever savory?’); Indian (Telegu): Anni vunna aaku anigi manigi vuntae aeimi laeni aaku yegriri (‘Fully served leaf [used to serve food] sticks to the ground whereas empty leaf flies over’); English: Empty vessels make the most noise. .... Dagara: Nɩ-been bε lhagr/ŋmaarε kyɩε mɔpʋla pʋɔ ɩ (‘A squirrel chased by many hunters never escapes’); Mushere: “People should walk in pairs in order to catch chicken for each other” (‘There is strength in numbers’); Xhosa: Umotu ngumtu ngabantu (‘A person is a person because of other people. No man is an island’); Amharic: “When spider webs unite, they can tie up a lion”. .... Many proverbs refer to marriage and family relationships. Some examples are: Dagara: Pɔg faa san’ yaa, kyɔg nuu kε bɔ pɔg (‘A bad wife is better than [one in] the grave; keep her and look for a good wife’); Malian: “If you want sex while traveling, travel with your wife’. Other social issues that need urgent attention and are being discussed worldwide are human rights, especially the rights of women and minority groups. For want of time and space we indicate here only issues between women and men. Certain proverbs refer, for instance, to misogynic issues and also tend to discriminate against minority groups. Misogynistic proverbs abound in most cultures. African examples are: Amharic: “Women and donkeys need a stick”; Somali: “The most dangerous thing that a man needs is a woman’; Bantu: “A woman’s clothes are the price her husband pays for peace’; Moroccan: “Like a gold ring in a pig’s snout is a woman without discretion”; Igbo: “No matter how stiff the breasts of a girl may be, they will collapse in the course of time”. In her essay on proverbial jesting Helen Yitah analyzes how Kasena women of the Upper East Region of Ghana revolt against proverbs showing negative attitudes towards women
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
by subverting and inverting misogynistic proverbs. The following proverbs are subverted and inverted accordingly: Kafena kalo na gabe kakwia to mo gabe kabia (‘It is the whip that lashes the senior wife that also lashes the junior wife’), and Kabaga na sâ àwana, ka wo sâ veaa mo (‘Once a slave accepts chains, she/he must agree to walk [in them]’) (Yitah : –). There is no doubt that there are positive proverbs in favor of women or that men are derided or discriminated against in proverbs. Some positive proverbs in regard to women may be quoted as follows: Indian: “Never strike your wife—even with a flower”; Ghana (Aggrey): “If you educate a man, you simply educate an individual, but if you educate a woman, you educate a nation”. Apart from projecting the importance of listening to and obeying parents, children play an important role in society. The following proverbs shed light on the role and attitude of children: “A child who washes his hands is fit to eat with kings” (Igbo); “A child of a snake is like a snake”; “A child behaves like its mother / The child follows his father’s way of thinking” (Mushere); Dole bε laar ʋ ma nyɩɩm ε (‘A piglet does not laugh at its mother’s protruding teeth’) (Dagara); “A child, a drunkard, and a fool tell the truth”. .... / Since all humans deal in trade, purchasing, marketing, hoarding, saving money, banking, negotiation, bargaining, or borrowing in one way or another, they also create proverbs to season their talk. The following proverbs serve as examples: “Beauty is potent but money is omnipotent”; “Banks have no heart”; “Bargain like a gypsy and pay like a gentleman” (which would now be considered offensive); Sika yε mogya (‘Money is blood’, Akan); “Know each other as if you were brothers; negotiate deals as if you were strangers to each other” (Arabian). .... // In ancient and, partly, in modern cultures, disputes were settled mainly in family circles or before a group of elders, a chief, king, or emperor. The institution of justice also created law courts to settle disputes. However, in modern times jurisdiction takes place mainly in law courts. The following proverbs demonstrate generally and specifically how cultures consider jurisdiction or lawsuits, for instance, cases in which justice is granted or denied. As shown here, some of the proverbs can be traced to their origin, but others have been so popularized that it is now difficult—even impossible or unnecessary—to trace their source. English: “The law catches the flies and lets the hornets go”; Amharic: “The poor man pleads; the judge listens”; Mushere: “Who can stop judging eggplants?” (= Nobody can be outside the law); Chilean: “Small flies are caught while the big fish continue to be at large”; Gladstone: “Justice delayed is justice denied”; “The law is an idiot, the adherents to the law are bigots, and the rest are tricksters, I think” (Bemile; personal observation). .... Arthur Schopenhauer, the German philosopher, opines that natural activities can almost compensate for the want of every kind of cultivation, but no cultivation of the mind can make up for the want of natural abilities (Chopra, p. ). Education is considered as cultivation of the mind which is absolutely necessary to supplement natural activities.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Indeed, many proverbs are regarded in the world for their high teaching value. Such proverbs emanate from or are spread by, for instance, great spiritual leaders and teachers such as Jesus Christ, Mohammed, Confucius, and Buddha as contained in religious books, the Bible, the Qur’an, and the Talmud. A popular proverb says that the “Errors of predecessors are landmarks to posterity”: It advises us to reflect deeply on the mistakes that our forefathers have committed and to try to avoid them. The following proverbs also point especially to the need for educating oneself properly: Burmese: “A man with a little learning is like the frog who thinks its puddle a great sea”; Ghana (Aggrey, as quoted in section .... above): “If you educate a man, you simply educate an individual, but if you educate a woman you educate a nation”; Guinean: “Knowledge is like a garden: if it is not cultivated, it cannot be harvested”; Moroccan: “With much knowledge, there is much sorrow; with much wisdom; there is much weeping”; Dagara: Mɔlʋ mɩ ɩag ʋ bile ɩag tu (‘As the antelope jumps, so does its kid jump’). .... We often hear people saying that they are apolitical. This may be true if one does not engage in partisan politics. It is, nonetheless, impossible to be devoid of political interest or neutral. Aristotle has this to tell us: “Man is by nature a political animal.” Proverbs are, indeed, employed very often by politicians to woo, coax, placate, castigate, advise, bamboozle, or even insult in a subtle manner their constituents or followers and opponents or detractors. The following proverbs, sayings, or quotes give different uses of proverbs in politics. As stated above, one may be able or unable to trace their origin: “For forms of government let fools contest, whatever is best administered, is the best”; “Give me good mothers and I will give you a good nation” (Napoleon); “Not many men have both good fortune and good sense (Titus)”. .... Old age is almost synonymous with wisdom in many cultures. It is, therefore, expected that old age and elderly people, especially one’s parents, are respected and are worth emulating. Their words are presumably full of wisdom and ought to be imbibed. The following popular proverbs go a long way to underline such expectations or beliefs: “What old men see while sitting, surpasses what young men see even when they climb a tree” (Mushere); “An old man’s sayings are seldom untrue” (Chinese); “An old man’s mouth may smell, but his words do not smell” (Dagara). .... For ages, religion has been looked at from different angles in the whole world. It may be considered as a means of spiritual growth or degradation, nay, an opiate of the people, depending on how it is used. The following proverbs purport to indicate how different
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
peoples look at religion or spiritual issues through proverbs. A proverb originating from Fontaine says that “Religious contention is the devil’s harvest”, and Confucius says that: “You know pretty little about this life: how can you know anything about life after death?” Let us look at proverbs from different parts of the world: Senegalese: “Even Buddhist priests of the same temple quarrel occasionally”; Burkina Faso/Gambian: “God provides for the blind vulture”; Akan: Onyame bεkyerε (‘God will provide’); Chinese: “God gives the milk, but not the pail”; Ghana: “Pray for life, and not for possessions”; Hungarian: “The Church is near, but the road is icy”; Moroccan: “By all means trust Allah, but tie your camel first”; Rwandan: “Only God generates; man only educates”. .... Some proverbs have been so much in use that they have become very popular internationally and universally. They exhibit universal precepts and may be termed universal proverbs. Examples of such proverbs are: “Everybody’s business is nobody’s business”; “Experience is the best teacher”; “An idle man’s brain is the devil’s workshop”; “He that digs a pit for others falls himself into it”; “If no one is willing to accept your point of view, try to see their point of view”. To establish universals it is also necessary to compare world proverbs. Therefore, the proverbs in sections ....–.... used in different parts of the world are juxtaposed under selected themes such as opportunity, reciprocity, prioritization, roaming or ineptitude, caution, and truth or honesty. .... / Dagara: Waab ŋmε gbʋrgbʋr nɩ waab kʋb (‘The more speedily you hit the snake the sooner you kill it’); Arabic: “Break the iron while it is hot”; Persian: “Bake the bread while the oven is hot”. .... Akan: Kɔya a agye wo no, yεn nfrε no kɔya wa (‘Do not call the forest that gave you shelter a distant one’); Yoruba: “However far a stream flows, it never forgets its origin”; Mushere: “A crab says that its mind will not forget its old house”; Xhosa: Ungalibali intake yobusi (‘Give the honey bird some of the honey that it helped you to get’);
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
Dagara: Fʋ ba wa soŋ fʋ kɔ mɔpʋla fʋʋ mɩ soŋ ʋ ʋ kɔ a ʋ mɔpʋla (‘If your friend helps you to clear your land for farming, also help him to clear his farm for farming’). .... English: “We can’t see the wood for the trees”; Arabic: “He leaves the donkey and holds onto the saddle”; Dagara: Fʋ kyεlɩ nɩ a dakɩɩr ɩaŋ ε a lee de a kyi dɩ (‘You have preoccupied yourself with the parrots and left the pigeons to eat up the millet’). .... / Akan: ɔba a ɔkyinkyini ɔnto ne ni/funu (‘The child that roams misses its parent’s/ grandparent’s burial’); Dagara: Lili-yʋɔra tambir daa n’a ʋ mɩ wa dɩ (‘The chicken that roams returns home to eat only the bitter termites left over’); Zulu: “A walking man builds no kraal”. .... Persian: “A hasty man pisses twice”; Maasai: “Rushing is not important; making sure is”; Senegalese: “Haste and hurry can only bear children with many regrets along the way”; Xhosa: Isiziba siviva ngodondolo (‘A hole in the river (isiziba) is felt with a long stick’). .... / Sanskrit: Satyameba jayte (‘Truth only wins [prevails]’); Dagara: Yelmɩŋa sa ziri (‘Truth is better than lying’); Akan/Twi: Anokore na εyε (‘The truth is good’).
. I
.................................................................................................................................. Without making any effort to understand what type of proverb is being taught or used, without committing to memory the proverb that is being taught or used, without applying the principles embedded in the creation of the proverb that one has learned, nay, studied, it is often not possible to feel the impact or effect of any proverb. It is, therefore, often necessary to reflect on a proverb in order to find out what impact its user intends to have on the addressee or hearer. How wiser our offspring might become than we ever were by reflecting on what the proverbs say and applying the precepts or teachings therein! The longevity of a proverb depends on many issues such as its linguistic and literary appeal to the language user, the linguistic structure that it possesses, all the ingredients or criteria that are used to prepare and formulate it, e.g. succinctness, pithiness, and popular usage. It also depends on the literary style, like poetic and metaphorical aspects that make the proverb linger in the memories, tongues, and hearts of the language users. Apart from that, the historical, current, and future impact/impression it has had, still has, and will have
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
in the lives of the users, the source of its creation, spread, or increase in popularity, i.e. the natural environment, and the individual, family, group, society, people, or nation and their cultural or moral values, also account for the long or short existence of a proverb. The durability of a proverb in a language also depends on the prestige and role performed by the language itself. For instance, even though the Latin language has been supposedly examined and “pronounced dead”, in modern times its influence is still glaring in many languages in the form of loanwords and expressions used in daily conversation, especially in code-mixing and code-switching with special reference to medical, theological, philosophical, legal, and oratorical language, where proverbs and other wise sayings are “proudly” employed as a mark of secrecy, scholarship, and erudition. The Greek language has also had a great impact on scientific, technological, medical, geological, mathematical, linguistic, religious or philosophical, and legal language and served in the formulation and spread of proverbs and other pithy sayings. For instance, Greek word-stems and words abound in phonetics, phonology, morphology, syntax, grammar, and the alphabets of various languages. Most of these “foreignisms” were initially expressed orally before they were committed to writing. Proverbs often create a huge impact on their users and hearers. For example, the shorter the proverb or the expression, the more impressive it is. Brevity is promoted in proverbs as is exemplified in the following proverbs: “Brevity is the soul of wit” (English); “The ideal phrase is that which is short and to the point” (Arabic); ɩεr ɩaga bε daar wure e (‘Much talking does not buy a horse’, Dagara); In der Kürze liegt die Würze (‘Brevity is the soul of wit’, German). The strength of orature lies in its resilience through the ages despite all the technological advances and progress in writing systems. Indeed, many written documents gather dust in libraries or bookshops as our present-day reading culture seems to have been relegated to the background, giving way to oral performance, i.e. orature, and thus covertly spelling the eventual doom of literacy. On the other hand, writing as a mnemonic means of communication is capable of undermining orature and illiteracy and reducing the development, training, power, and efficiency of human memory. For traditional culture and illiteracy are often frowned upon despite their richness and impact on human development. Also, very often, no sooner are things written down than they are forgotten until reference is made to them, e.g. through the use of a diary or memorandum. It is, however, possible to derive joy and satisfaction from proverbs through appreciating their poetic aspects and their use in drama, as outlined earlier in section ..., e.g. alliteration, rhymes, puns, and satire, especially in oral performance, and deciphering or unraveling the meanings and salutary messages hidden in the riddles, parables, and enigmas employed in proverb formulation. In Juvenal’s words: Difficile est saturam non scribere (‘It is difficult not to write satire’). Unraveling hidden meanings in certain proverbs in itself poses challenges to the human mind and promotes initiative in trying and succeeding to break down and reduce to ordinary speech and meaning compressed, convoluted, enigmatic, secret, sacred, and parabolic expressions in proverbs, e.g. Faarɩgyɔ! (‘Hold your tongue!’; Dagara); Favete linguis! (‘Keep silent!’) The creation of proverbs and anti-proverbs alone does point to the creativity of the human being. As stated earlier, certain criteria have to be met in the creation of different
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
.
types of proverbs. The originators of proverbs, therefore, use their creativity and set criteria to formulate proverbs that meet the required standards. The pedagogic effect of proverbs is often felt and appreciated by the speaker and the listener, especially when one listens to and accepts moral values taught and the wisdom expressed in proverbs by parents, elderly people, religious leaders, and philosophers. The scholarly, rhetorical, and oratorical use of proverbs by, for instance, lawyers, politicians, preachers, philosophers, and actors often arouses admiration, an aura of respect for the user, and reverence for the use of proverbs. In many instances, through the use of proverbs we obtain in a nutshell some knowledge about certain historical, medical, scientific, technological, linguistic, legal, and universal and specific attitudes, truths, or precepts about ourselves, God, nature and the natural environment, and the universe.
. C
.................................................................................................................................. In the foregoing discussion emphasis was laid on the relevance of language and orature with reference to proverbs. It was necessary to define language, “written” literature, and “oral” literature (orature) in general. A quick analysis of their historical and cultural implications was made. The origin, creation, collection, and use of proverbs were also mentioned. It was then imperative to specify the application of linguistic and paralinguistic features and literary styles in proverbs. Core issues like phonetic, phonological, morphological, and syntactical processes were cursorily discussed. Finally, it was of interest to indicate the impact of proverbs on their users and hearers. Some of the conclusions that could be drawn were that the language employed in formulating proverbs is different to a great extent from ordinary language, in that it is more succinct, poignant, durable, creative, popular, artistic, aesthetical, figurative, universal, and effective than ordinary language. Finally, it was, indeed, noticed that more paroemiological studies need to be done, especially in the linguistic, literary, and impact analysis of proverbs.
R Bemile, S. K. (). Dàgàrà Proverbs. (With English, French, German, Spanish and Latin equivalents.) Berlin: Dietrich Reimer. Clement, H. A. (). The Story of the Ancient World: The earliest times to the fall of Rome. London: African Universities Press in association with Harrap. Dako, K., Angsotinge, G., and Denkabe, A. (). An Introduction to Language and the Language of Literature. Legon, Accra: Sub-Saharan Publishers. Dalfovo, A. T., Beyaraaza, E. K. M., Kaboha, P., Kigongo, J. K., Mwanahewa, S. A., Wamala, E., and Zubairi ‘b Nasseem (eds.) (). The Foundations of Social Life: Ugandan philosophical studies I. Washington DC: Council for Research in Values and Philosophy. Mieder, W. (). Proverbs: A handbook. Santa Barabara, CA: Greenwood Press. Yitah, H. (). ‘Throwing stones in jest: Kasena women’s “proverbial” revolt’, Oral Tradition : –.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
F Barber, K., and de Moraes Farias, P. F. (eds.) (). Discourse and its Disguises: The interpretation of African oral texts. Birmingham: University of Birmingham Press, Centre of West African Studies. Bauman, R. (). Story, Performance, and Event: Contextual studies of oral narrative. Cambridge: Cambridge University Press. Bemile, S. K. (). ‘Multilingualism in Ghana’, in T. Bearth, W. J. G. Möhlig, B. Sottas, and E. Suter (eds.), Perspektiven afrikanistischer Forschung. X. Afrikanistentag, Zürich, .–. September . Cologne: Rüdiger Köppe, –. Bemile, S. K. (). ‘Mythology: the case of the Dagara secret society’, in E. Sutherland-Addy (ed.), Perspectives on Mythology. Accra: Woeli Pub Services for the Goethe-Institut, –. Berman, L. (). Proverb, Wit & Wisdom: A treasury of proverbs, parodies, quips, quotes, clichés, catchwords, epigrams and aphorisms. New York: Berkley. Chopra, M. (). A Book of Proverbs and Quotations. New Delhi: Lotus. Dzobo, N. K. (). African Proverbs: A guide to conduct. Vol. VIII: The Moral Value of Proverbs. Accra: Woeli Publishing. Finnegan, R. (). Oral Poetry: Its nature, significance and social context. Cambridge: Cambridge University Press. Finnegan, R. (). Oral Traditions and Verbal Arts: A guide to research practices. London: Routledge. Finnegan, R. (). Oral Literature in Africa. Oxford: Clarendon. Foley, J. M. (). The Theory of Oral Composition: History and methodology (folkloristic). Bloomington, IN: Indiana University Press. Görög-Karady, V. (ed.) (). Genres, Forms, Meanings: Essays in African oral literature. Oxford: Journal of the Anthropological Society of Oxford. Henige, D. (). ‘Oral, but oral what? The nomenclatures of orality and their implications’, Oral Tradition : –. Ong, W. J. (). Orality and Literacy: The technologizing of the word. London and New York: Routledge. Phillips, S. (n.d.). Proverbs. New Delhi: Goodwill. Tonkin, E. (). Narrating Our Pasts: The social constructions of oral history. Cambridge: Cambridge University Press. Vansina, J. (). Oral Tradition: A study in historical methodology. Chicago and London: Aldine and Routledge & Kegan Paul.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
......................................................................................................................
......................................................................................................................
. I
.................................................................................................................................. W poetry might be universal, each culture has a different understanding of what poetry is. Many African languages originally lack a cover term for “poetry” or “literature”, but rather refer to specific genres. For heuristic reasons, roughly following Jakobson’s concept of the poetic function of language (), we might highlight the concern for the creation of an elevated language, where form, sound, language structure, a select vocabulary, and content are made to harmonize beyond the casualness of everyday speech as a global concern in poetry as such and poetry in African languages more generally speaking. While Western prosodic poetic principles like quantifying metrics and rhyme are less widely attested (see, e.g., the different modes of quantifying prosody in Somali, Hausa, and Swahili genres, as well as Swahili rhyme patterns adopted from Arabic), rhythmization, recurring figures of sound and metaphorical language are widely found and can take various forms. Parallelism in tone, syntax, and semantics is a recurring—understudied— structuring principle in different poetic traditions. The concern for creating a meaningful atmosphere often goes beyond sheer linguistic concern: important genres are recited on the occasion of carefully planned performances embedded in cultural contexts and often accompanied by musical instruments. Still, genres differ tremendously in terms of medium, mode of recitation, composition, room for innovation, and audience, as well as in terms of their integration into cultural contexts, like religious rituals, ceremonies, or occasions of sheer entertainment. Thus, there is a continuum of poetic speech delivered at formalized occasions in a form of heightened language and less formalized texts or contexts, like lullabies, children’s games, work songs, and proverbs. Outside of poetry in the strict sense, various African cultures, like Twi-, Malagasy, and Swahili-speaking contexts, have been described for the estimation of discourse which makes use of poetic devices, like metaphor, allusion, and sound figures (see Banti and Giannattasio’s notion of “poetically organized discourse” ()). While this chapter deals with oral poetry, poetry in manuscript cultures, and modern and contemporary poetry in print and new performances, it is important not to consider this overview as simply chronological and to relegate oral genres or manuscript cultures to the past. To a large extent oral genres are also contemporary, not in the sense that they are
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
(only) performed the way they used to be performed one hundred years ago, but more that a number of them have proved to be and have been particularly dynamic, open to technical, thematic, and stylistic innovations, and surviving through a constant change of media. Most recently, the spoken word and hip-hop have re-explored oral traditions. Furthermore, African manuscript cultures (see section . below), but also written poetry in print which relates to oral poetry in various ways, speak counter to the continuously repeated dichotomy of oral and written literature in African languages.
. O
.................................................................................................................................. In sections ..–.. we will consider some important categories of poetry. The list is not exhaustive, but within the limited space of this chapter, it is used to categorize oral genres.
.. Praise poetry Panegyric poetry is one of the best-described poetic genres in Africa. While praise of personal achievements in war or hunting also plays a role in less centralized societies, like the Oromo or the Tuareg, the examples most often referred to are types of court poetry, celebrating the bravery and military achievements of kings and war heroes, sometimes also of Gods and spirits, as among the Hausa or Yoruba. Praise poetry traditions are found widely in the Bantu area: besides the kasàlà tradition of the Luba (Mufuta ), the panegyric genres of the Bantu-speaking groups of the Great Lakes, like the Haya (Seitel ), Hima, and Ankole (Morris ), the Southern Bantu praise poetry of the Zulu (Stuart and Cope ), Xhosa (Opland ), and Sotho (Kunene ) are certainly best described. But praise poetry is also widespread in the Horn of Africa among the Oromo and Somali (Banti a, b) as well as in West Africa, where. besides Hausa (Smith ), Manding, and Fulani praise singing, the oriki tradition of the Yoruba and the msabran at Akan courts (Andrzejewski et al. : ) are among the very best-documented cases. Praises are often recited by professional praise singers, who can become part of the king’s official entourage, like the griots, a caste of specialists in the Fulani and Manding traditions. Rather than merely celebrating the rulers, praise poems can play a regulative role in society, as, for instance, in the Zulu izibongo tradition, where the praise singer can pronounce subtle or overt criticism, or in the Hausa context, where rok’o singing does not depend on agreement between customer and singer (Smith : ). Praise songs are primarily based on praise names which people accumulate throughout their life, but which are also handed down in the lineage. Closely linked to culturally specific concepts of personhood, they often go beyond individual identity, outlive a person, but can become re-embodied (Barber ). Stylistically, praise poems are marked by a predominantly nominal style, since epithets, like the Zulu King Shaka’s praise name ‘The Ever-ready-to-meet-any-challenge’ (quoted in Finnegan : ), are basically arranged in parallel, which accounts for a distinctively elliptical style. Because of its partly formulaic character, the language register is often archaic and accompanied by a laudatory tone produced by a raise of pitch in the Zulu context or musical instruments, like drums, as, for
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
instance, in the Hausa context. Praise poetry has been adapted in various “modern” contexts to mobilize constituencies and has featured recurrently in election campaigns, as, for instance, in northern Nigeria, Somalia, and South Africa, where it has been adapted in trade union movements, the anti-apartheid struggle, and more recently in Jacob Zuma’s campaigns in (Buthelezi ). Earlier on, Xhosa and Zulu praise poems fused with Christian hymns in the nineteenth and twentieth centuries and also found their way recurrently into popular music (Buthelezi ), but also inspired written poetry (Opland ).
.. Epic-heroic poetry Epic poetry shares a lot of features with praise poetry, like the praise of a hero’s achievements and its historical orientation, but typically takes the form of a long narrative in rhythmized language. Since Finnegan’s claim that “relatively long narrative poems” basically did not exist in Africa (: ), the epic has been one of the most-discussed genres of African poetry (for a bibliography, see Barry ). In particular, discussions on orality and performance, which gained impetus for Africa in the late s, productively added new perspectives on the epic. African epics have henceforth been highlighted as multimodal performances delivered by specialized performers, like the griot, or the jeli, in the Mande tradition, often involving musical instruments (Okpewho ; Belcher ; Mulokozi ). Recurring formulas allow for considerable textual variability. A typical epic plot centered on the hero is driven by an early conflict—Liyanja breaks a spell, Sunjata is forced into exile—which after a number of episodes which make the narrative gain in epic breadth leads to a final battle, the glorious victory of the hero, and the resolution of the conflict. Epic traditions are widespread in Africa and run like a belt from Senegal to the interlacustrine area of Central Africa and are also found on Madagascar and on the Swahili coast. In West Africa, apart from Hausa and Songhai traditions, the Sunjata epic, and the Segou cycle of the Mande peoples, various Fulani epics centered around the legendary heroes Hambodedio or Silamaka or the Sufi mystic El-Hadj Omar Tall are the among the best-documented (Kesteloot and Dieng ; Belcher ). While in West Africa epics are mostly linked to political history, epic traditions in other areas, like the Jeki of the Duala in Cameroon, the epic of the Mvet and the Lianja and the Mwindo epic from the Congo basin rather “focus on the actions of a mythical or ancestral hero” (Belcher : ). In all contexts, however, epics not only record history but play an important role in constructing an ethnic or national identity. Accordingly, many epics have been re-explored in the twentieth century and adapted to new forms like the Ozidi saga which Pepper documented and also turned into a play (Clark-Bekederemo ). Other epics can be watched emerging (Belcher : –), like the epic of Shaka the Zulu. Building on oral history, it was not only turned into a Sesotho novel but Masizi Kunene also composed the “Emperor Shaka the Great: A Zulu Epic” (), a poem later published in English, and it was also adapted into TV series from the s onwards as well as into films and comics books.
.. War and hunting poetry Genres of war and hunting poetry resemble both praise and epic poetry, since they celebrate heroism and the triumph of war or the hunt or both, as in many Nilotic traditions,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
like that of the Dinka, where war is associated with cattle raiding (Finnegan : ). While the Ngoni of Tanzania, Zambia, and Malawi have a tradition of war song (imigubo), war songs are best documented for Northeast Africa, like the Oromo geerarsa (Banti a), a poem of triumph, and the Somali geeraar (Banti b), traditionally recited on horseback, which was used to challenge opponents and boost morale. In West Africa, the Akan warrior associations (asafo) led by captains use songs which glorify war and boost morale, accompanied by drums and gongs and rousing calls which create a succinct and fast rhythm and a high-spirited tension meant to intimidate the enemy (Andrzejewski et al. : ). The Mande and the Yoruba in West Africa are renowned for their hunting poetry, which is well institutionalized, as it is linked to hunters’ associations. Though the Yoruba ijala goes beyond hunting in theme, it is still locally associated with the hunters and their patron Ògún. The ijala performed by trained specialists salutes the wild animal when setting out for a hunt, but also chants the praise names of distinguished people and ancestors, making use of a formulized language in an antiphonal style of recitation often accompanied by drumming (Babalola and Adeboye : –).
.. Elegiac poetry Elegiac poetry can also be found all over Africa and takes many different forms. As Finnegan (: ) outlines, it is far less described than praise songs, since it does not have the same political relevance and tends to be performed by non-professionals. One of the best-studied examples is that of Akan dirges or assudwom, a distinct genre with an elaborate canon of forms and themes, which is chanted by women soloists on the occasion of funerals. It involves “music, sobs and tears and conjoined to bodily movement” (Nketia : ), and the women call on the deceased by praise names and refer to the traits, social status, and merits of the dead person. Poems—such as Ngoni laments, or Acholi or Yoruba dirges—also depict the personal grief of the mourner (Finnegan : ).
.. Religious poetry There is a great variety of religious poetry in Africa: “There are hymns, prayers, praises, possession songs, and oracular poetry, all with their varying conventions, content, and function in different cultures” (Finnegan : ). Furthermore, religion can often hardly be separated from social institutions, and it is thus often difficult to draw the boundaries between religious and secular genres. While magical verses like spells and incantations are not frequently found to take longer form in verse, there are various genres across Africa of supplication or intercession, which imply an idea of influencing the course of events, like rain ceremonies common in Central Africa or the demand for daily needs among the Khoisan (Finnegan : –). In mantic poetry, gnomic poetic utterances are tied to mediums, like the divining bones among the Sotho in southern Africa or the chain of seeds used in the elaborate Ifa oracular poetry of the Yoruba (Finnegan ). Furthermore, Islamic and Christian religious verse have become important genres in a number of cultures across Africa, disseminated both orally and in written form. For centuries, Arabic edification texts have been adopted in different ways into various African languages, as, for instance, the maghāzī epic of the battle of Tabuka, which has been adopted in Kanuri,
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Hausa, and Swahili (El-Miskin ). Similarly too in various Christian churches either initiated by missionaries or newly emerging churches, themes and forms of oral traditions have been particularly vital, as, for instance, in the hymns of the Church of Nazareth, which have incorporated elements of the praise song tradition of the Xhosa and Zulu in South Africa (Buthelezi ), or the Fante Methodist Church in Ghana, which re-explored Fante chorus-type lyrics (Finnegan ).
.. Lyric poetry and songs There are also various forms of short poems, which are typically sung. Traditions of poetic competitions often accompanied by dancing are widespread, for instance, in eastern Africa and have recently been extended into verbal duels among hip-hop crews (Gunderson and Barz ). Furthermore, in various traditions across the continent, stories like the Malinke talinn (Pfeiffer ) or the Gĩkũyũ ng’ano (Kabira ), are interrupted by songs which, besides embellishing the storytelling, can have mnemotechnic functions. There are a lot of rather ephemeral songs composed spontaneously, such as the Shona drinking songs (Finnegan : ). In lyrical poetry and songs, love and marriage are the most recurring topics. Apart from the elaborate Swahili and Hausa corpora of love poetry, the Somali tradition of balwo (later heello) is a striking example of romantic love poetry, marked by condensed imagery, which became particularly fashionable in towns after World War II, where special balwo parties took place (Banti b). Also in other contexts, since the second half of the twentieth century, lyric poetry has particularly increased and easily found its way into new forms of popular music, short messages, and social media, particularly in urban contexts.
. W
.................................................................................................................................. Apart from those from Egypt, there are numerous other examples of written poetic texts from Africa. Ethiopia, with its long Christian writing tradition, is certainly prominent: the first document in African written literature is Ge’ez epigraphic poems, which narrate the life of saints and martyrs and the miracles of the Virgin and also include different types of esoteric hymns (gene or qene), short witty poems marked by great obscurity of style, the use of puns, and an abundance of metaphors (Gérard : ). Furthermore, various Islamic cultures mostly in West and East Africa adapted Arabic script to write down poetry in languages like Hausa, Fulfulde, Bambara, Tamasheq, Somali, and Swahili. While there is a long tradition of composing poetry in Arabic in West Africa, particularly the Sufi brotherhoods and their movements of Islamic renewal in the nineteenth century made written Islamic poetry flourish in local languages like Hausa, Fula, and Wolof, which became the main vehicles for transporting and popularizing Islamic knowledge. Although religious verses differ in style and subject matter, this poetry, characterized by an emphasis on religious learning, ethical preoccupations, and spirituality and refined in
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
terms of style, was typically composed by learned scholars deeply committed to the spiritual betterment of their people. The Futa Jalon in present-day Guinea was an early important center of Islamic scholarship where Mohammadou Samba of Mombeya (–) is reported to have been the first to write religious verse in Fula (Gérard : ), which soon inspired other poets in the area, such as the Senegalese poet and mystic Amadou Bamba (–), founder of the muridiyya brotherhood, who composed a large number of hagiographic verses and panegyrics in Arabic and Wolof. Uthman Dan Fodio (–), an important Fulani theologian born in present-day Niger, who instigated a wave of jihads through present-day northern Nigeria, wrote jihad poems in Hausa, Arabic, and Fula. Continued by his successors Abdullah bin Muhammad (–) and Muhammad Bello (–), a huge body of both Fulani and Hausa Islamic poetry written in Arabic script came into being, mostly didactic or homiletic in nature, adopting prosodic patterns and topics of Classical Arabic verse to fit the local context and language (Hiskett ; Seydou ). In East Africa, the eighteenth and nineteenth centuries can be considered a period of “scripturalization” of poetic Swahili. Accordingly, new genres of poetry emerged, like the utenzi, a long poem typically narrating the battles of the Prophet’s followers, like the already mentioned Utenzi wa Tambuka (), or giving advice, like the Utenzi wa Mwana Kupona (), as well as the short, lyrical shairi, which reached its height in the metaphorical poetry of the Mombasan poet Muyaka bin Haji al-Ghassaniy (–) (Abdulaziz ). At the same time, local scholars also made an effort to preserve ancient Swahili oral poetry by committing it to writing, as, for instance, the songs attributed to the master poet Fumo Liyongo (Miehe et al. ). Unlike print cultures, where the written word gains prominence over the oral word, there are various ways in which literacy and orality coexist and intersect in African manuscript cultures. Although committed to manuscripts, memory and recitation often play the primary role as modes of preservation, a guarantee of authenticity, and a form of textual rendering and aesthetic appreciation.
. P L
.................................................................................................................................. In the nineteenth and the first half of the twentieth century, first Western missionaries and later also the colonial governments promoted literacy in Latin script and introduced new literary genres and new media like print and radio, which had a lasting impact on local literatures. Particularly in various Anglophone colonies, efforts were made to promote literacy through institutions like the mission school or literature bureaus by training potential new writers. Poets like the great Xhosa poet and journalist Samuel Edward Krune Mqhayi (–) played an important role in recording oral poetry, developing reference works for their own language as well as pioneer works in poetry and prose in African languages (Opland ). New media like print media and the radio offered new ways of producing, consuming, and disseminating poetry. For instance, in South Africa already at the beginning of the nineteenth century, long before books were regularly
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
published, newspapers coming from the mission press offered a platform for the publication and discussion of Xhosa poetry: the Xhosa poetess Nontsizi Mgqwetto published about ninety poems focusing on social change in the newspaper in the s (Opland , ). In Tanganyika, in the journal Mambo Leo, poets would continue their verbal duels in written form, and in the Yoruba context, the radio and later audiocassettes soon became an important way of disseminating neo-traditionalist poetry, which follows traditional divination chants (Adéèkó ). At the beginning of the twentieth century, a number of poets played with metrical conventions from European verse traditions in their written works: the Tswana poet Leetile Disang Raditladi (–) introduced rhyme in his work Sefalana sa menate (Granary of Niceties, ), the Yoruba poet Afolabi Johnson (–) imposed a measured couplet structure on Yoruba poetry, and the Malagasy poet Edouard Andrianjafintrimo (–) experimented with metrical schemes derived from French and Latin prosody. Other poets, however, explored the style of local oral genres, as, for instance, E. P. Ndhambi’s (born ) and P. E. Ntsanwisi’s (born ) collections of poetry in Tsonga, which made use of the traditional praise poem style (Marivate : ). In the Yoruba context, J. S. Ṣówándé adapted a Yoruba chanting mode in his written poetry, which departed from it thematically, as it did not promote religious or cosmic viewpoints but “addressed excesses of the local colonial administration [and] criticized the immoralities of young Christian converts” (Adéèkó : ). In South Africa, the Zulu poetry of Benedict W. Vilakazi (–), published in two volumes between and , constituted a departure from the Zulu poetic tradition of praising kings and heroes. Referring to Christian values, Vilakazi’s poetry emphasized social conflicts and protested against the inhuman conditions of mineworkers. In the Swahili context, Matthias Mnyampala (–) and Shaaban Robert (–) paved the way for a nationalist poetic tradition by reframing classical genres and addressing challenges of the modern nation state, like the provision of healthcare and education.
. P
.................................................................................................................................. While the struggle for liberation inspired many poets to use poetry as a weapon, such as the so-called Soweto poets like Wally Serote or the renowned Masizi Kunene (–), who composed critical poems in Zulu, also in other African countries the search for a (pan-) African identity and symbols of new national identities gave birth to written poetry with a nationalist, laudatory tone that continued after independence. In Twi, D. K. Abbiw published Akan Awensem (Akan Poems, ), lyrical and patriotic poems, in Somali Cabdi Muxumud Aamin (born ) praised the ideals of the Somali revolution of , and the Ndebele poetry Isidlodlo SikaMthwakazi (Mthwakazi’s Headring) edited by D. N. Ndoda () praised the liberation war, chimurenga (Chiwome : ). In the s and s, a new generation of well-educated poets like Abdilatif Abdalla (born ), Euphrase Kezilahabi (born ), and Said Ahmed Mohamed (born ) in Swahili-speaking eastern Africa, and Kgadime Matsepe (–), the ironic Pedi master poet, explored poetry as a means of criticizing their own societies, the failures to live up
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
to the ideals of independence, and social dislocations which the Acholi poet Okot p’Bitek (–) thematized in his satirical poem Wer pa Lawino (Song of Lawino, ). “Neo-traditionalists” would rather attempt to re-explore traditional modes, as, for instance, Abdilatif Abdalla who wrote his prison poetry in the classical shairi genre, or Àlàbí Ògúndépò and Ògúndáre Fónyánmu, Yoruba poets, who “are known for their inflections of chant of the Hunters’ Guild” (Adéèkó : ), while other poets like p’Bitek searched for a new poetic style. The University of Dar es Salaam became an important hub for poets like Euphrase Kezilahabi and Mugyabuso Mulokozi (born ), who experimented with free verse poetry in search of a new poetic language that would break with “traditional” Swahili canonical rules. Their works paved the way for later experiments by academic poets like the Kenyan Kithaka wa Mberia (born ), who would play with the visual presentation of poetry. In southern Africa, one of the most remarkable contemporary poets is the Zimbabwean Chirikure Chirikure (born ), a performer-poet who voices his critical poetry in an eloquent Shona style often highlighted by music, as, for instance, in his collection Hakurarwi (We Shall Not Sleep, ) (Chiwome : ). More generally speaking, performance has triggered renewed interest in poetry since the turn of the millennium. Towards the end of the twentieth century, academic poetry in African languages had lost momentum. The novel became the prime genre of intellectual literary expression; the generation of well-educated academic poets of the s and s still continued to compose critical poetry and set aesthetic standards, like Masizi Kunene’s (–) Indida yamancasakazi (The Puzzle of the Ncasakazi, ), or the poetry of the Sotho poet K. P. D. Maphalla (born ), or the Swahili writer Said Ahmed Mohamed (born ), but a new generation with a new outlook did not easily come into being, and poetry tightly linked to the book market with the school as its mainstay lacked readership and critical debate. Popular poetry, on the other hand, proved to be particularly dynamic, surviving through widely available popular media like audiocassettes, disks, Internet forums, and SMS. While popular music was an important vehicle for verse in African languages throughout the twentieth century, hip-hop and spoken word movements in particular have had a huge influence on poetry in African languages. Often Senegal and Tanzania are referred to as the most dynamic hip-hop scenes. In the s, hip-hop artists like Kool Kock of Pee Froiss, who was the first to rap in Wolof, or Mr. II, who was among the first to not merely cover American rap in Swahili, adopted features from global hip-hop culture but also explored their own poetic resources, including oral genres and musical instruments, like the kora or djembe in the case of Senegal (Moulard ). The global success of spoken word and hip-hop culture, which swept the urban centers of the Western world, also renewed interest in poetry from the s onward. Intellectual and popular poetry have come to intersect in terms of style as well as in regard to their social concern, and, in particular, poetry performances have become immensely popular. In African urban centers, poetry slams are held regularly, where hip-hoppers perform together with educated slam poets, like the Kenyan obokano player and poet Grandmaster Masese or the Tanzanian poet-performer Mrisho Mpoto. Mixed registers, heavily involving urban youth languages like Sheng in Nairobi, Tsotsitaal in Johannesburg, and Camfranglais in Yaounde, have become widely used not only by hip-hoppers but also by other performer-poets. Besides festivals, websites like badilishapoetry.com, which hosts podcasts of poetry by young highly reputed South African, Zimbabwean, and Kenyan artists, such as the Zulu poet MP Mkhize
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Da Bee, the Venda poet Dzomolavenda, and the Shona poets Memory Chirere and Madzitatiguru, as well as new and alternative avenues for publication and distribution have further fostered interest in poetry: in Nairobi, the literary journal Kwani!, launched in as a forum for new literary creativity after the sterile era of Moi, not only regularly organizes poetry slams but also publishes hip-hop verses and Sheng poetry in addition to English texts. In terms of their critical agenda and counter-culture, hip-hoppers like Ukoo Flani MauMau in Kenya or spoken word poets like the celebrated Tswana-speaking poet and scholar Kgafela oa Magogodi inscribe their poetry into the discourse of liberation associated reggae and dub poetry, Maumau warrior traditions, and anti-apartheid Soweto poetry. While in various contexts spoken word poetry and hip-hop have become profitable professional options and many successful poets and hip-hoppers take part in an increasingly global circuit of verbal art, in a number of contexts they are still risky forms of voicing protest, as, for instance, in Zimbabwe, where spoken word poetry has become a niche for expelled university students and talented poets like Biko MC and Julius Chingono.
R Abdulaziz, M. H. (). Muyaka: th Century Swahili Popular Poetry. Nairobi: Kenya Literature Bureau. Adéèkó, A. (). ‘Yoruba literature’, in S. Gikandi (ed.), –. Andrzejewski, B. W., Piłaszewicz, S., and Tyloch, W. (eds.) (). Literatures in African Languages: Theoretical issues and sample surveys. Cambridge: Cambridge University Press. Babalola, S., and Adeboye, O. (). The Content and Form of Yoruba Ijala. Oxford: Clarendon. Banti, G. (a). ‘Oromo oral literature’, in S. Uhlig (ed.) (), –. Banti, G. (b). ‘Somali oral literature’, in S. Uhlig (ed.) (), –. Banti, G., and Giannattasio, F. (). ‘Poetry’, in A. Duranti (ed.), A Companion to Linguistic Anthropology. Oxford: Basil Blackwell, –. Barber, K. (). The Anthropology of Texts, Persons and Publics: Oral and written culture in Africa and beyond. Cambridge: Cambridge University Press. Barry, A. O. (). L’épopée peule du Fuuta Jaloo: De l’éloge à l’amplification rhétorique. Paris: Karthala. Belcher, S. P. (). Epic Traditions of Africa. Bloomington and Indianapolis, IN: Indiana University Press. Buthelezi, M. (). ‘Praise, politics, performance: from Zulu izibongo to the Zionists’, in D. Attwell and D. Attridge (eds.), The Cambridge History of South African Literature. Cambridge: Cambridge University Press, –. Chiwome, E. (). ‘Shona literature’, in S. Gikandi (ed.), –. Clark-Bekederemo, J. P. (). The Ozidi Saga. Ibadan: Ibadan University Press. El-Miskin, T. (). ‘The “Kayawar” in the context of the epic tradition’, Research in African Literatures : –. Finnegan, R. (). Oral Literature in Africa. Oxford: Clarendon. Gérard, A. (). African Language Literatures: An introduction to the literary history of sub-Saharan Africa. Harlow: Longman. Gikandi, S. (). Encyclopedia of African Literature. London and New York: Routledge. Gunderson, F., and Barz, G. F. (eds.) (). Mashindano! Competitive Music Performance in East Africa. Dar es Salaam: Mkuki na Nyota. Hiskett, M. (). A History of Hausa Islamic Verse. London: SOAS.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
Jakobson, R. ( []). ‘Linguistics and poetics’, in R. Jakobson and S. Rudy (eds.), Selected Writings. Vol : Poetry of grammar and grammar of poetry. The Hague: Mouton, –. Kabira, W. M., (). Gĩkũyũ Oral Literature. Nairobi: Heinemann. Kesteloot, L., and Dieng, B. (). Les épopées d’Afrique noire. Paris: Karthala. Kunene, D. P. (). Heroic Poetry of the Basotho. Oxford: Clarendon. Marivate, C. (). ‘South African literature in Tsonga’, in L. Klein (ed.), African Literatures in the th Century: A guide. Harpenden: Oldcastle Books, –. Miehe, G., Abdalla, A., Abdulaziz, M. H., Al-Bakary, Z., Baschiera, A., Bhalo, A., Dittemer, C., Nabahany, A., Omar, Y., and Topan, F. (). Fumo Liyongo Songs: Poems attributed to Fumo Liyongo. Collected and edited by the Liyongo Working Group. Cologne: Rüdiger Köppe. Morris, H. F. (). The Heroic Recitations of the Bahima of Ankole. Oxford: Clarendon. Moulard, S. (). ‘Le rap sénégalais, un mode d’expression inédit entre oral et écrit?’, in F. VeitWild and A. Ricard (eds.), Interfaces between the Oral and the Written. Amsterdam and New York: Rodopi, –. Mufuta, P. (). Le chant kasàlà des Lubà. Paris: Julliard. Mulokozi, M. M. (). The African Epic Controversy: Historical, philosophical and aesthetic perspectives on epic poetry and performance. Dar es Salaam: Mkuki na Nyota. Nketia, J. H. K. ( []). Funeral Dirges of the Akan People. New York: Negro University Press. Okpewho, I. (). The Epic in Africa: Toward a poetics of the oral performance. New York: Columbia University Press. Opland, J. (). Xhosa Oral Poetry: Aspects of a Black South African tradition. New York: Cambridge University Press. Opland, J. (). Xhosa Poets and Poetry. Cape Town: David Philip. Opland, J. (). The Nation’s Bounty: The Xhosa poetry of Nontsizi Mgqwetho. Johannesburg: Witwatersrand University Press. Pfeiffer, K. (). Sprache und Musik in Mandinka-Erzählungen. Cologne: Rüdiger Köppe. Seitel, P. (). The Powers of Genre: Interpreting Haya oral literature. New York: Oxford University Press. Seydou, C. (). La poésie mystique peule du Mali. Paris: Karthala. Smith, M. G. (). ‘The social functions and meaning of Hausa praise-singing’, Journal of the International African Institute : –. Stuart, J., and Cope, A. T. (). Izibongo: Zulu praise-poems. Oxford: Clarendon. Uhlig, S. (ed.) (). Encyclopaedia Aethiopica. Vol. . Wiesbaden: Otto Harrassowitz.
OUP CORRECTED PROOF – FINAL, 29/2/2020, SPi
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
A I
......................................................
Aarne, A. Abalo, A. K. Y. , Abasheikh, M. I. , Abbadie, A. T. de , , Abbott, M. , Abdalla, A. Abdel-Hafiz, A. S. , Abdulaziz, M. H. , , , , , , , Abebe, A. , Abebe Gebre-Tsadik, M. , Aboh, E. O. , , , , , , , , Abrahamsson, H. Abu-Manga, A. , , , Achard, M. , Ackerman, F. , Adal, Z. Adamou, E. Addo, T. , Adeboye, O. , Adéèkó, A. , , Adelung, J. C. , Adendorff, R. Adesola, O. , , , , Adjakossi, E. Y. Adjekum, G. , Admasu, Y. Adone, D. , Afeli, A. , Ager, S. , Agyekum, K. , , Ahadzi, S. , Ahland, C. Aikhenvald, A. Y. , , , , , , , , , , Akanlig-Pare, G. , Akinlabi, A. , , , , Akinloye, O. Akinnaso, F. N. , Akwey, S. Alamin, S. , , , , , , , Al-Bakary, Z. Alcón-López, D. Alidou, H. , Alidou, O. , , , Ali Karmal Kokko, S. , al Khalil, M. , Allen, J.
Allen, J. P. , Allen Fox, R. Almkvist, H. Alsina, A. , Al-Wer, M. , , Amberber, M. , , , , , , Ameka, F. K. , , , , , , , , , , , , , , , , , , Ameur, M. , Amha, A. , , , , , , , Amiot, D. Amoako, J. , Anaagudo-Agu, O. C. , Andersen, H. , Andersen, T. , , , , , , Anderson, A. H. Anderson, B. , Anderson, E. N. , , Anderson, G. D. S. , , , , Anderson, J. Anderson, J. A. Anderson, L. A. Androutsopoulos, J. , , Andrzejewski, B. W. , , , , , , Angsotinge, G. Anonby, E. J. , Ansah, G. N. Ansaldo, U. , , Anscombe, G. E. M. Anttila, R. , , , , Anyanwu, R.-J. , Appiah, K. A. Appleyard, D. (L.) , , , , , , , , , , , , , , , , , , , , , , , , , , , Archangeli, D. , Arends, J. Arensen, J. , , , Arewa, E. O. , Armstrong, R. (G.) , , , Arnott, D. W. , Aronoff, M. Aronson, H. I. Asher, R. E. , Ashley, C. , , , , , Atchi, E. A. F. , Athanasiadou, A.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Atifi, H. , Atindogbe, G. G. Atintono, A. S. Atiya, A. S. , Attayoub, A. K. , , Attridge, D. Attwell, D. Austin, J. L. , Austin, P. (K.) , , , , , , Austin, W. Auzeanneau, M. , , Avesac, A. , Awagana, A. , , , Aycard, P. , , , , , , Ayotte, C. , Ayotte, M. , Azuonye, C. , Babaev, K. V. , , , , , Babalola, S. , Babault, S. , , , Bada, M. D. , Badat, S. Bahner, W. Bailey, A. C. Bailey, R. , Baker, C. , , , Baker, M. (C.) , , , , , , Baker, P. , , , , , , , , Bakker, P. , , Bakpa, P. , Baldauf Jr., R. B. , , Baldi, P. Bamgboṣe, A. , , , , , , , , , Banda, F. , Banti, G. , , , , , , , Baptista, M. Barasa, S. N. Barber, K. , , Baril, D. , Barnard, A. , Baroin, C. Barreteau, D. , , , Barringer, T. A. Barry, A. , Barry, A. O. , Bartels, A. Bartens, A. , , , , , , Barth, H. , , , Barton, D. Barz, G. F. , Baschiera, A. Bascom, W. R. , , , Bashir, Abeer M. A. , , , , , Basset, A. , , , , , Bassiouney, R. , ,
Bastin, Y. , , , , , , , , , , , , Batibo, H. , , , Baudelaire, H. , Bauer, F. L. , , , , Bauer, L. Bauman, R. Baumann, H. , , , , , , Baumgardt, U. Bausi, A. , , , Bavelas, J. , Bavin Woock, E. , Baya, J. , Bearth, T. , , , , , , , , , , , , , , , , , Beaudoin,-Lietz, C. Beaujard, P. , Beavon, K. , Beavon, M. , Beavon-Ham, V. , Becher, J. , , , , , Bechert, J. , Bechhaus-Gerst, M. , , , , , , , Beck, R. M. , , , , Becker, M. Beckman, J. , Bedilu Wakjira Bedou-Jondoh, E. Behnstedt, P. , Beke, C. T. , Belcher, S. P. , Bell, A. , Bembe, P. , Bemile, S. K. , , , , Bender, M. L. / L. M. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Bender, P. , Bendor-Samuel, J. T. , , , , , , , , , , , , , , , , , , , , Bengtson, J. D. Benjamin, M. Benmamoun, E. , Bennett, N. R. , Bennett, P. R. , , , , , , , , , , , , Bennett, W. Benson, C. Bentahila, A. , , , Bento, M. Benveniste, E. , Berger, P. , , , Bergman, R. , Bergman, T. G. , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Berjaoui, N. , Berlin, B. , , , , , , , Berman, L. Bernard, H. R. , Bernard, Y. , Bernini, G. Berns, M. Berry, J. , , , Bertho, J. , Bervoets, S. J. Besha, R. M. , Bessamia, R. Beukes, A.-M. , Beyaraaza, E. K. M. Beyene, T. Beyer, K. , , , , , , , , , , Beyogle, R. , Bhabha, H. K. , Bhalo, A. Bhaskararao, P. , Bhatt, R. Biber, D. , Bickel, B. , , Bickmore, L. , , , Biesele, M. , , , Bimson, K. (D.) , , , Binnick, R. I. Bird, C. , Bird, S. , , , , Birru, T. Bizimana, S. , , , Black, K. , Black, P. , , , Blackings, M. , Blaho, S. , Blank, A. , Blažek, V. , , , , , , , , , , , , , , Bleek, D. F. , , , Bleek, W. H. I. , , , , , , , , , , , Blench, R. (M.) , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Bliese, L. F. , , , Blok, D. P. , Blommaert, J. , , , , , , , , , , , , Bloomfield, L. , Blust, R. , Boas, F. , Boden, G. , Bodomo, A. , , , , Bodrogligeti, A. J. E.
Bogale, B. Bogers, K. Bokamba, E. G. , , , Bok-Bennema, R. Bollée, A. , Bolouvi, L.-P. Bombien, L. Bond, O. Bondarev, D. , Booij, G. E. Borer, H. , Borghouts, J. F. , Borowsky, T. Bosire, M. Bostoen, K. , , , , , , , , , , , , , , , , , , Botne, R. (D.) , , Boubakary, A. Bouny, P. , Bouquiaux, L. , Bourcier, A. , Boutkan, D. Boutrais, J. Bowen, J. D. , , Bower, J. , Boyd, G. Boyd, R. , , , , , , , , , , , , , , , , , Boyd, V. , Boyeldieu, P. , , , , , , , , , , , , , , Boyer, O. T. Bradfield, M. , Bradford, S. Bradshaw, M. , , Brahimi, F. , Brandstetter, C. T. , Branford, S. Brankaer, J. Braun, D. P. , Breál, M. , Bredell, A. W. , Breedveld, A. , Breedveld, J. O. , Breitbarth, A. Bremer, N. D. , , Bremond, C. , , Brentari, D. , Brenzinger, M. , , , , , , , , , , , , , , , , , , Bresnan, J. , , , Breton, R. J. L. , Bretschneider, M. Brettschneider, G. Breu, W. , Breyer, F. Bright, W. , , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Bril, I. Brinkman, I. Brock-Utne, B. , , , , , , , , , , , , Brody, M. , Brookes, H. , , , Brooks, G. E. , Brousseau, A.-M. , Brower Stahl, A. Browman, C. , Brown, K. , , Brown, P. , , , , Brown, R. M. Browne, G. M. , Bruck, A. Brugman, C. , , Brugman, J. (C.) , , Brugnatelli, V. , , Brugos, A. Brunk, K. Brusciotto, G. , , Bryan, M. A. , , , , , , , , , , , , , , , , , , , , , , , , , , , , Bryant, D. , Bubenzer, O. Buell, L. , Bugenhagen, R. D. , Bühnen, S. , , Bulkens, A. , Burgess, C. S. Burtea, B. Bußmann, H. , , , , Buthelezi, M. , , , , Butler, J. , Büttner, C. , Bynon, J. , , , , Bynon, T. Cahill, M. , , , , , , Calame-Griaule, G. , , Calhoun, C. , Calteaux, K. , Camara, É. , Campbell, F. , , Campbell, L. , , , , , , Canut, C. , Capo, H. (B. C.) , , , , , , , , , Caprile, J.-P. , , , Caquot, A. , , Cardinall, A. W. , Carleton, T. , Carlson, R. (J.) , , , , , , , , , , , Carnie, A. Caron, B. , , , ,
Carrington, J. F. , , , Carstens, V. , , , , Casali, R. (F.) , , , , , Cassidy, F. G. , Caubet, D. , , Caudwell, S. , Černý, J. , Cerulli, E. , , , , Chaker, S. , , , , , Chamberlin, J. E. , Chambers, J. K. , Chambers, M. , , Chamoreau, C. Chanda, V. M. , Chapple, D. Charette, M. Charles, M.-C. , Charpentier, J.-M. , Chaudenson, R. , Chebanne, A. , , Chebchoub, F. , Chéneau-Loquay, A. Cheng, Chin-Chuan , Cheng, L. , Chia, E. N. , , Childs, G. T. , , , , , , , , , , , , , , , , , , , , , , , , , , , , Chinyowa, K. , Chiwome, E. , , Chomsky, N. , Chopra, M. Chrisomalis, S. , Christaller, J. G. , Christiansen-Bolli, R. , Chtatou, M. , Cienki, A. Cihlar, J. E. , Cinque, G. , , Cissé, G. , , Cissé, Y. T. , Clark-Bekederemo, J. P. , Clark-Cotton, M. R. Clarke, K. , Claudi, U. , , , , , , , , , , , Claus, H. , Clement, H. A. , Clements, G. (N.) , , , , , , , , , , , , , , , , , , , , , , Clifford, J. , Clifton, J. M. , Clist, B. Cloarec-Heiss, F. , Cloete, N. Cmejrková, S. Cobbinah, A. , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Cobler, M. , Cohen, D. , , , , , , , , , Cohen, M. , , , , , , Cole, D. T. Cole, N. L. , Collins, C. , , , , , , , , , , Comfort, J. , , Comrie, B. , , , , , , , , , , , , , , , Connell, B. , , , , , , Conti Rossini, C. , , Cooper, R. L. , , Cope, A. T. , Coppola, M. Corbeil, J.-C. , Corbett, G. G. , , , , , , , , , Cornelissen, E. , Coulmas, F. , , , , , Coupez, A. , , , , Coupland, N. , Cowley, R. W. , Crabb, D. W. , , , Craig, C. G. Crane, T. M. , Crass, J. , , , , , , , , Crawhall, N. , Crazzolara, J. P. , , , Creider, C. (A.) , , , , , , , , , , , Creider, J. T. , , , Creissels, D. , , , , , , , , , , , , , , , , , , , , , , , , Croft, W. (A.) , , , , , , , , , , Crossland, Z. , Crowther, S. A. , Croze, C. Crozier, D. H. , Crum, W. E. Cruse, D. A. , Crystal, D. , , , , , Culy, C. , Cummins, J. , Curnow, T. J. , Currie, T. E. , Cuyckens, H. , , , Cyffer, N. , , , , , , , Czekanowski, J. , Daan, J. C. , , Dako, K. , , , Dalby, D. , , , , , , , , , , , , Dale, D. , Dalfovo, A. T. , Dalgish, G. M. ,
Dalphinis, M. Danet, B. , Daniels, P. (T.) , , , , , Dapper, D. , , Dard, J. , Darnell, J. C. , Das Gupta, J. , , Dasher, R. B. , , David, N. , Davies, E. E. , , Davies, I. R. L. , , , , , Davis, C. Davydov, A. V. , Dawd, A. , Deal, A.-R. de Bruijn, M. , Déchaine, R.-M. de Féral, C. , , , , , de Filippo, C. DeGraff, M. , , , , de Halleux, B. , , , de Jong, N. , , , de Kind, J. , Delafosse, M. , , , , , Dell, F. , Delplanque, A. , , , , de Luna, K. M. , , de Maret, P. , Dembski, D. G. Demisse, T. , , Demolin, D. de Moraes Farias, P. F. Dempwolff, O. , , den Besten, H. , , , , Dench, A. Denkabe, A. Denning, K. , Denny, J. P. , de Pommerol, J. , , Deprez, K. Derbyshire, D. C. Derive, J. , de Rooij, V. , de Rop, A. , Desai, Z. , , de Schryver, G.-M. , , , , , , , De Silva Jayasuriya, S. , , Destaing, E. Deuber, D. Deumert, A. , , , , , , , , Devonish, H. , de Vos, C. Devos, M. , de Wolf, P. P. , , , , , de Yaya, C. , , Dhoorre, C. S. , , Diakonoff, I. M. , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Diallo, A. , Diallo, M. , Di Carlo, P. , , Dickens, P. (J. ) , , , , , , , Dieng, B. , Diercks, M. , Diethart, J. Dieu, M. , Diffloth, G. , , Dihoff, I. (R.) , , , , , , Dil, A. S. Dimmendaal, G. J. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Dingemanse, M. , , , , , , , Di Tolla, A. M. Dittemer, C. Dixon, R. M. W. , , , , , , , , , , , , Djité, P. G. , Dobbs-Allsopp, F. W. Doke, C. M. , , , , , , , , , Dom, S. Dombrowsky-Hahn, K. , , , , , Dominicy, M. , Doneux, J. L. , , , Dooley, S. A. Doornbos, P. , Dor, J. Dorian, N. Doro Rikoto, B. , Dorvlo, K. , Doumenge, C. Downing, L. , , , , Downing, P. A. Dowty, D. R. , Drexel, A. , Dreyfus, M. , Drolc, U. , Dryer, M. S. , , , , , , , , , , , , Dube, M. , Ducroz, J.-M. , Dufková, K. , Dugas, A. Dumestre, G. , , , , , , , , Du Plessis, T. , Duran, J. J. , Duranti, A. Durie, M. Dürscheid, C. Dutcher, N.
Dwyer, D. J. , , , , , , , , Dybo, A. V. Dybo, V. A. Dyck, C. , Dyer, D. L. Dyers, C. , Dzablu-Kumah, S. W. , Dziwirek, K. Dzobo, N. K. Eastman, C. M. , Eaton, H. , , , , , Eberhard, D. M. , Eberle, A. Ebermann, E. , , Echardt, R. Edel, E. , Edgar, J. , Edwards, J. R. , Ega, É. , Egbokhare, F. O. , , , Eggert, M. , , Ehret, C. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Ehret, P. , , , Eidson, R. , Eira, C. , Eiseb, E. , Ekundayo, S. A. , Elbow, P. , El-Dash, L. , Elderkin, E. D. , , , , , , , , , , , , , , Elders, S. , , , , , , , , , , , , , , , , Elghamis, R. , Elgibali, A. El-Guzuuli, El-Shafie , Elias, P. , , , , , Elles, R. J. , Elmedlaoui, M. , , El-Miskin, T. , El Said, G. R. El-Sayed, R. , Elugbe, B. O. , Enfield, N. J. , Englund, P. , Engsheden, Å. , Enguehard, F. , Epstein, A. L. , Erben, J. Erdal, M. Erman, A. Ermisch, S. , , , , Essegbey, J. , , , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Essien, I. Evans, N. , , Evans, V. , , Evans-Pritchard, E. E. , , , Ewald, J. J. , Exter, M. Fabb, N. , , , , Fabre, A. G. , Fairclough, N. , Fallon, P. , , Faraclas, N. G. , , , Fardon, R. , Farquharson, J. T. , Fasold, R. W. , Faur, E. , Fehn, A.-M. , , , , , , , , Fennig, C. D. , , , , , , , Ferguson, C. (A.) , , , , , , , , , , , , , , Ferguson, J. , Fernandez, J. , Fernández-Ardèvol, M. Fiedler, I. , , , , , , , , Fields-Black, E. , Figueroa, E. , Filho, W. T. Finlayson, R. , , , , , , Finnegan, R. , , , , , , , , , Finney, M. A. Fischer, W. , Fishman, J. A. , , , , , , , , Fitzgerald, C. , Flavier, S. Fleisch, A. , , , , , , , , , , Fleisher, J. B. Fleishman, L. Fleming, H. (C.) , , , , , , , , , , , , , , , Flik, E. , Fodor, I. , Foley, J. M. Foley, W. A. , , Fon Sing, G. Ford, K. C. , , Förster, H. Fortune, G. , Foster, P. , Fourquet, J. , Frajzyngier, Z. , , , , , , , , , , , , , , , , , Frampton, J. , Franklin, A. L. , Freeman-Grenville, G. S. P. , , , Fricke, S. L.
Friedman, V. A. Friedner, M. Friedrich, J. , , , Frobenius, L. , Fromkin, V. , , Fronzarolli, P. Frynta, D. , Fuller, D. Q. , Furniss, G. , , , Gachathi, F. N. , , Gafaiti, S. Gafaranga, J. , , Gagnon, D. Galand, L. , , , , , Galaty, J. G. , Galtier, G. , , , , , , , Gamble, C. , , , Ganslmayr, H. Garbo, B. Garbo, G. García, O. , , , Garde, F. , Gardiner, A. , Gardner-Chloros, (B.) P. , , Garfinkel, H. , Gasparini, A. , Gasparov, M. Gates Jr., H. L. Gaudefroy-Demombynes, M. , , , Gavua, K. Gaye, D. , , Gbeto, F. Gébert, A. , Geeraerts, D. , , , , Geider, T. , , Gelderblom, D. Gensler, O. D. , , , , Gérard, A. , , Gerbault, J. , , , Gerdts, D. B. Gerhardt, L. , , , , , , , , , , Gerlach, L. , Gertner, M. H. Gerwing, J. , Giannattasio, F. , Gibbard, G. , Gick, B. , , Giesing, C. , Gikandi, S. Gil, D. Gilley, M. , Gilman, C. , Gilroy, P. , Gimba, A. M. Gingiss, P. ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Gippert, J. , , , Githinji, P. , , Githiora, C. , Givón, T. , , Gladwin, T. Glanz, C. Glaser, C. , , Glavy, J. M. , , , Gligo, H. A. Glück, H. Gnamiato, V. Goddard, C. , Goebl, H. , , Goffman, E. , Goldenberg, G. Goldsmith, J. , , , , , Goldstein, L. , Gonggryp, J. W. , Good, J. , , , , , , , , , , , , , , , , Goodchild, S. , Goodman, M. , Goodwin, C. , Gooskens, C. , Goossens, J. , , Gordon, R. G. , Gore, C. E. C. , , , , Gore, E. C. , , , , , Görög-Karady, V. , , , Gorter, D. Göschel, J. Gouaini, E. Gowlett, D. F. Goyvaerts, D. L. , , , , , , , Graebner, W. Graefe, E. , Gragg, G. (B.) , , , Granadillo, T. , Grapow, H. Gravina, R. , Gray, R. , Green, M. , , , , , , Green, R. C. , Greenberg, J. H. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Gregersen, E. A. , , , , , Grégoire, C. , , , , , Grégoire, H.-C. , Grenoble, L. A. , , , , Griefenow-Mewis, C. , , , ,
Grimes, B. F. , Grinevald, C. , , Griscom, R. , Groll, S. I. , Grollemund, R. , , , , , , , , , Grosse, R. , Grossmann, E. Grottanelli, V. L. Gruber, J. (S.) , , Guarisma, G. Guchteneire, P. Guerssel, M. , Guillon, M. Güldemann, T. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Gulfan, G. I. , Gullberg, M. , Gumperz, J. J. , Gunderson, F. , Guthrie, M. , , , , , , , , , , , , , , , , , , , , , Gutmann, B. , Gyamfi, A. , , Gyekye-Aboagye, J. , , Ha, S. Haacke, W. H. G. , , , , , , , Haberland, E. , Hachimi, A. , Hackett, P. , Hackley, R. Hadidja, K. Haeckel, E. , , Hagège, C. , Hagemeijer, T. Hagman, R. S. , , , , Haїk, I. Haiman, J. , Hair, P. E. H. , Hajek, J. , Halévy, J. , Hall, S. , Haller, B. , Halliday, M. A. K. , , , , , , Hallowes, D. P. , , , , , , , , Halpert, C. , , , Halvorsen, T. A. , , Hamdan, A. , , Hamilton, M. Hammarström, H. , , , , Hanafiou, H. S. ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Hancock, I. F. Hanks, W. , Hannig, R. , Hansford, K. , , Hantgan, A. Hardin, C. L. , Harford(Perez), C. , Harguindéguy, R. , Harkins, J. Harley, A. Harney, E. Harnischfeger, J. Harries, L. , , Harris, R. , Harris, R. J. Hartell, R. L. , Hartmann, K. , , Haruna, A. , Hasitzka-Johannes, M. Haspelmath, M. , , , , , , , , , , , , , , , , Hassan, J. , Hassan, M. , Hasznos, A. , Hau, K. , Haugen, E. , Haust, D. , Hawisher, G. E. Hawkins, J. , Hayward, R. (J.) , , , , , , , , , , , , , , , , , , , , , , , , , , , , Heath, J. , , , , , Heepe, M. , Heeringa, W. , , Heggarty, P. , , , Heikkinen, T. , Heine, B. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Heine, I. Heinrich, P. , Held, D. Heller, M. Hellwig, B. , Henderson, B. , , , , , Hendrikse, A. P. , Henige, D. Henrix, M. , Herbert, R. K. , , ,
Heritage, J. , Herring, S. C. , Herskovitz, M. J. , Hess, R. L. Hetzron, R. , , , , , , , , , , , , , , , , , , , , , , , , , , , , Heugh, K. , , , , , , , , , Heusing, G. , , , Hewson, J. Hickey, R. , , Hieda, O. , , , , , , , , , , , , , , , Highfield, A. , , Hilberth, J. , Hill, J. H. , Hillewaert, S. , Himmelmann, N. P. , , , , Hinnebusch, T. J. , , , , , , Hinrichs, L. , Hinton, L. , Hiraiwa, K. , , , Hirst, G. Hiskett, M. , Hislop, G. Hoch, J. E. , , Hodder, I. Hodge, C. T. , , , , , , Hoeft, L. , Hoff, H. , , Hoffmann, C. , , , Hoffmannová, J. Höhn, A. Holden, C. , , Hollington, A. , , , Holm, J. Holman, M. E. Holman, T. W. Holtus, G. Hombert, J.-M. , , , , , Honken, H. , Hoole, P. Hoppenbrouwers, C. , Hoppenbrouwers, G. , Hopper, P. J. , Hopson, R. K. Horn, C. Hornberger, N. H. , Houis, M. , , Hounongbe, J. de Dieu K. Hounongbe, S. Hovens, M. , , Hoymann, G. , Hristova, D. S. Hubbell, A. , Huber, M. , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Hudson, G. , , , , , , , , , , , , , , Huebner, T. Huehnergard, J. , Hume, E. Hunn, E. Hünnemeyer, F. Hunwick, J. , Hunziker, D. A. , , Hunziker, E. Hurst, E. , , , , , Huson, D. H. , Hussein, E. , Hutchison, J. P. , , Huttar, L. , , Huttenga, P. , , Hyman, L. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Hymes, D. , ,
Jirou, G. , Joffre, J. , Johanness, M. A. G. Johnson, J. H. , Johnson, J. W. , , Johnson, K. Johnson, M. , , Johnston, Sir H. H. , , , , , Jolivet, R. , Jones, D. , , Jones, P. , Jones, R. , , , Joswig, A. , , , , , , Juffermans, K. , Juillard, C. , Junge, F. , , Jungraithmayr, H. , , , , , , , , , , , , , , Junker, W. , Jurafsky, D. ,
Ibrahim, A. , Ibrahim, G. , , Ibriszimow, D. , , , , , , , , Idiatov, D. , , Iguée, A. , Ilboudo, P. T. , Imam, A. , Immerzeel, M. Inkelas, S. , Irvine, J. T. , , , , , Ivić, P.
Kabinga, M. , , , Kabira, W. M. , Kaboha, P. Kaboré, R. , Kaelin, O. Kahigi, K. Kahl, J. , , Kahlheber, S. , , , , Kaiser, D. W. , Kaji, S. , , Kakumasu, J. , Kamanzi, T. , Kamau, S. N. , , Kamei, N. , Kamissoko, W. , Kamwangamalu, N. Kan, S. Kandybowicz, J. , , , , Kanerva, J. , Kanté, S. , Kaplan, R. B. , , Karan, E. , , , , , , Karg-Gasterstädt, E. Kashina, K. , Kashoki, E. M. , Kassan, B. Kastenholz, R. , , , , , , , , , , , , , Katamba, F. , , , , , , , Kato, B. Kaufman, T. , , , , , , , , , , , , , Kawachi, K. , , , , Kawahara, S. Kawalya, D. , Kay, P. , , , , , , ,
Jacobson, R. Jaeger, M. , Jaffe, A. , Jaggar, P. , , , , Jake, J. , Jakobi, A. , , , , , , , , , , , , , Jakobson, R. , Janse, M. Jansen, J. Jansen-Winkeln, K. , Janson, T. , Jarrett, K. , Jason, H. , Jastrow, O. , Jefferson, G. , , , Jenewari, C. E. W. , , , , Jenks, P. , , Jenner, P. N. Jensen, H. , Jeppie, S. , Jerono, P. , Jesney, K.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Kayambazinthu, E. , Kaye, A. S. , , , , Kaye, J. D. Kayne, R. , Keane, W. , Kedalan, G. K. Kedrebéogo, G. Keegan, R. , Keenan, E. , Kegl, J. , Kehr, K. Keller, R. E. , Kemmer, S. , , , , Kendon, A. , Kenstowicz, M. , , , , , , , Kerr, R. , Kershner, T. L. , Kertz, L. , Kességbeu, M. A. , Kessler, B. , Kesteloot, L. , Khalil, M. K. , Khamis, C. , Khan, G. Khumalo, N. H. E. , Kibora, L. , , Kidwell, M. , Kießling, R. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Kievit, D. , Kigongo, J. K. Kihore, Y. Kikusawa, R. Kilham, H. , Kilian-Hatz, C. , , , , , , , , , , , , , , , Kim, C.-W. , , , Kim, Y. , Kimbara, I. , Kimenyi, A. , Kindell, G. E. Kinhou, S.-M. Kirch, P. V. , Kiss, K. É. Kisseberth, C. , , , , , , , Kita, S. , Kitalong, K. S. , Kitalong, T. , Kittilä, S. Ki-Zerbo, J. Klaus, D. Klein, L. Klein-Arendt, R. , ,
Kleinewillinghöfer, U. , , , , , , , , , , , , , , , , , , , , , , , Klieman, K. , Klingenheben, A. , Klipple, M. A. , Kluge, A. , Knappert, J. , Knappett, C. , Kneissler, B. Knigge, C. Knörr, J. Koch, P. , Kock, I. J. , Kocks, G. H. , Kodesh, N. , , Koelle, S. W. , , , , , , , , , , , , Koenig, M. Kogan, L. , Köhler, B. , , Köhler, O. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Kok, P. , Koné, H. König, C. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Konig, E. Koni Muluwa, J. , , , Konta, M. Koopman, H. , , , , , Koops, K. , Kootz, A. , Kortlandt, F. , Kose, E. , Koslova, N. , Kossmann, M. , , , , , , , , , , , , , , , , , , Kouadio N’Gessan, J. , , , , , , , Kouame, D. T. Kousoulis, P. Kouwenberg, S. , , , Kövecses, Z. , , , , , , Koyt, M. M. , Kra, K. A. E. , Kramer, R. , , , Kramer, R. L. , , , , , , , , Kramer, U. , Krampah, D. E. K. , Krapf, L. , Kraska-Szlenk, I. Krause, G. A. Krauss, M. , Kröll, H. , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Krönlein, J. G. , Kropp, M. E. , Kropp Dakubu, M. E. , , , , , , , , , , , , Kruskal, J. , Kube(-Barth), S. , , , , , Kubik, G. , Kuhn, T. S. , Kühnert, B. Kuku Alaki, T. , , , , , , , , , , Kula, N. , , , Kunene, D. P. , , , , , , Kuntz, J. Kurth, D. , Kusters, A. Kuteva, T. , , , , , , Kutsch Lojenga, C. , , , , , , Labahn, T. Labov, W. , Labrada, J. E. R. , LaCharité, D. , Lacroix, P. F. , , Ladefoged, P. , , , , , , , Lafkioui, M. , , , , , , , , , , Lagae, C. R. , , , , La Galy, M. W. Lakoff, G. , , , , , Lamberti, M. , , , , , , Lambrecht, W. , Lamoureux, S. , Landi, G. , Lane, P. , Lange, D. , Lanham, L. W. , , , , , , , , Lantham, R. G. , Lanza, E. , , Lapranova, L. , , Larochette, J. , , , Last, M. , , , , Lavachery, P. , Layton, B. , Lazaridis, N. Leap, W. L. Leben, W. (R.) , , , , , , Lebikaza, K. K. Lecarme, J. , , Leclère, M. Le Cur, C. , Le Cur, M. , Ledgard, J. M. , Lefebvre, C. , , , , , Legendre, G. Leger, R. ,
Legère, K. , , , , , , , , Léglise, I. , , Lehfeldt, W. Lehiste, I. , Lehmann, C. , , , Leibniz, G. W. , Lekens, B. , Lentin, J. Léonard, J. L. Le Page, R. Lepsius, (K./C.) R. , , , , , , , Lerner, G. H. le Saout, J. , Lesko, L. H. , Leslau, W. , , , , , , , , , , , , , Levinson, S. C. , , , , , , , , Lewis, I. M. , , , , Lewis, M. P. , , , , , , , , , , , , , , , , , , , , , , , , , , , , Lexander, K. V. , , , Leyew, Z. , , , , , , , , , , , , , , , , , , Li, C. N. Lichtenstein, M. H. K. , Liddicoat, A. , , , , Lieber, R. , , Liedtke, M. Lightner, T. , Lim, L. Lin, A. Y. L. Lindblom, G. , Link, C. , , Lipski, J. M. , Littig, S. , Lloyd, L. C. , Lode, L. Loesov, S. , Loh Pius Löhr, D. Lokonobei, L. L. , Longacre, R. E. , Lonnet, A. , , Lopes, A. J. , Loprieno, A. , Louali(-Raynal), N. , , , Loubier, C. , Lovegren, J. , , Lovejoy, P. E. , , Lovestrand, J. , Lowy, E. G. Lucas, C. Lucassen, D. , , , Ludolf, H. , Luffin, X. , , Lukas, J. , , , , , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Lule, D. , Lundberg, M. J. Lüpke, F. , , , , , , , , , , , , , , , , , , , , , , , , Lux, C. , , , , Lyons, J. , MacBrair, R. M. , MacDiarmid, D. N. , MacDiarmid, P. A. , MacDonald, K. , Mace, R. MacGaffey, W. , MacKaye, S. , MacLaury, R. , Maddieson, I. , , , , , , , , , Maffi, L. , , , , Mafu, S. , Mafundikwa, S. , Maganga, S. Magassouba, M. Maghway, J. B. , Maguire, W. Mahdi Hamid Muudee , Mahlalela-Thusi, B. , Mahmud, U. , , , , Maho, J. , , , , , , Makhudu, K. D. P. , Makkai, A. Makoni, S. (B.) , , Malaise, M. , Malchukov, A. , Maley, J. , , Malgoubri, P. , Manassa, C. Manessy, G. , , , , , , , , , , , , , , , , , , , , , , , , Ma Newman, R. , , , , , Manfredi, S. , , , , Manfredi, V. , , , , Maniacky, J. , Mann, M. Marçais, P. , Marchese, L. , , , , , Marck, J. , Marcus, H. G. , Maribe, T. , Marivate, C. (T. D.) , , , Marlo, M. R. , Marten, L. , , , , , , , Martin, P. W. Martin, S. E. , Martinet, A. , Martinet, J. ,
Masinyana, S. O. , , Mathes, T. K. , Matisoff, J. A. Matondo, M. , , Matras, Y. , , , , , , , , Matthews, S. , , Maurer, B. , Maurer, P. , , Mauro, T. de Mawani, A. M. Mawussi, A. F. May, S. , Mazrui, A. A. , , , Mazrui, A. M. , Mbanza, A. G. Mbense, J. T. , , Mbolifuye, F. Mbomate, J. V. , McCall, D. F. , McCarter, P. K. McCarthy, J. , , , McDaniel, C. K. , , McGarry, R. G. A. , McGill, S. , , , , McGrew, T. Mchombo, S. A. , , , , , , McIntosh, S. K. McLaughlin, F. , , , , , , , , , , , , , , McMahon, A. , , , , , , McMahon, R. , , , McNeill, (E.) D. , , McWhorter, J. , Meade, A. , , Meeussen, A. E. , , , , , , , , Meeuwis, M. Mehrotra, R. R. , Meier, G. Meillet, A. , , , Meinhof, C. , , , , , , , , , , , , , , , , , , , , , , , , Meißner, A. , Mejdell, G. , Mekonnen, A. G. Y. , Mekoulnodji, M. , Melby, A. K. Melick, C. Mendel, D. Menz, G. Mereu, L. Merolla, D. Merrifield, W. Mervis, C. B. , Meshcheriakov, A. N. Mesthrie, R. , , , , , , , , , , , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Métangmo-Tatou, L. Metekouli, Z. Mettouchi, A. , , , Metzeltin, M. Meyer, C. , Meyer, E. , Meyer, R. Meyer-Bahlburg, H. , , Michaelis, S. (M.) , , , , Micheli, I. , Mieder, W. , , Miehe, G. , , , , , , , , , , , , , , , , , , , , Mietzner, A. , , , , , , , Mikesh, P. , Milán, W. G. Miles, M. , Milicic, B. , , Militarev, A. Y. , , , , Miller, A. L. , Miller, C. , , , Miller, J. Miller-Ockhuizen, A. , Milroy, J. , , , Milroy, L. , , , , , Mitchell, A. Mitchell, L. , Mitchell, P. Mkwanʼhembo, P. Moeller, S. Mofokeng, M. , Moges, R. T. , Mohamed, S. A. , Mohanan, K. P. , Möhlig, W. J. G. , , , , , , , , , , , , , , , , , , , , , , , , , Molamu, L. , , Möller, M. , Mondanda, L. , Moñino, Y. , , , , , , , , Moore, K. E. Moore-Cantwell, C. Mooshammer, C. Moran, S. , Moravcsik, E. , Morén, B. , Moreno, M. (M.) , , , , , , , Moriggi, M. Morin, D. , , , Morolong, M. , Morris, H. F. , Morse, M. , Mosel, U. , , , Moser, R. , , Moshi, L. , , , , , Mougiama-Daouda, P. ,
Moulard, S. , Moulton, W. G. , Mourigh, K. , Mous, M. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Moxley, J. L. , , Mreta, A. Y. , Msimang, C. T. , Mudrak, O. A. Mufuta, P. , Mufwene, S. (S.) , , , , , , , , , , , , , , , , , , , Mugane, J. Mugdan, J. Mühlhäusler, P. , Mukarovsky, H. G. , , , , Mulaudzi, P. A. , Mullen Kreamer, C. Müller, C. , Müller, F. , , , , Muller, J. , Müllerová, O. Mulokozi, M. M. , Muluwa, J.-M. Mumba, E. Mumin, M. , Munganda, R. , Munkaila, M. M. Munro, P. , , Múrcia-Sànchez, C. , Murray, M. A. , Mutaka, N. , , , Mutembei, A. K. , Muysken, P. , , , , Muzale, H. R. T. , Mve, J. P. Mwachofi, A. K. , Mwanahewa, S. A. Mwinsheikhe, H. M. , Myers, S. , Myers-Scotton, C. , , , , Nabahany, A. Nabirye, M. , Nachtigal, G. , , , Naden, A. J. , , , Naden, T. , , , , , , Naїt-Zerrad, K. , , , Nakagawa, H. , , , , , , , , Namaseb, L. , , Nana Aba Amfo Nassenstein, N. , Nathan, D.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Ndamba, J. Ndayiragije, J. , Ndaywel é Nziem, I. Ndecky, A. , Ndimele, O.-M. Ndlovu, R. S. , Nédellec, B. , Neethling, S. J. , Nekvapil, J. , Nelezek, R. Nerbonne, J. , Nettle, D. , Neumann, K. , , , Neveu, F. , Nevins, A. , Newell, S. , , , , Newman, P. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Newman, T. N. , Newmeyer, F. J. Ngomanda, A. N’Guessan, T. S. , Ngwenya, A. V. , Nichols, J. , , , , , , , Nicolaї, R. , , , , , , , , , , , , , Nicole, J. , , , , Nicole, M.-C. Niggli, I. , Niggli, U. , Nikiema, E. , , Nikiéma, N. , Nikitina, T. , , Nikolaeva, T. Nixon, R. , Nixon, S. Njock, P. E. , Nketia, J. H. K. , Nkomo, Kh. Nkosi, D. Mm. , Nöldeke, T. , Noonan, M. , , , , Nooter Roberts, M. Nordbustad, F. , , , Norris, E. , Norton, R. , , , , , , , , , , Noss, P. A. , Nʼouéni, R. W. , Nougayrol, P. , , , , , , , Nstadi, C. Nsuka-Nkutsi, F. , Ntshangase, D. K. , Nuckolls, J. B. , , Nunez, J.-F. ,
Nurse, D. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Nuuh Ali, M. , Nwosu, I. , Nwoye, O. G. , Nyamnjoh, F. Nyati-Ramahobo, L. , Nyst, V. , , , Obeng, S. G. , Obiamalu, G. O. C. , Obikudo, E. Ockinga, B. G. , Odden, D. , , , , , , , , , , , , Odden, M. , Oesterreicher, W. , , O’Fahey, R. , Ogechi, N. , Ogilvie, S. Ogunleye, T. M. , Ohala, J. J. , , Ojo, A. Okombo, O. , Okpewho, I. , Oldendorp, C. G. A. , ole Mpaayei, J. T. , Oliver, R. , Olson, K. S. , , , , , Olson, R. E. Omar, Y. Ong, W. J. Ongaye, O. , Onguene Essono, L. M. , Ono, H. , Opland, J. , , , , , , Orie, O. O. , Ortman, S. , Orton, C. , , Osam, E. K. Osborn, D. Z. , , Osinde, K. , , , , Ospovat, A. Oswalt, R. L. , Oteng, F. S. , Otsuka, H. Ouane, A. Oucho, J. O. Ouhalla, J. , , , Ourso, M. Owens, J. , , , , , , , , , , , , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Oyetade, O. S. , , Oyler, D. W. , Pacchiaro, S. Page, C. , Pagel, M. , Pagliuca, W. Pakendorf, B. , , Palancar, E. L. Palayer, P. , , , , Palfreyman, D. , Pali, T. , Palm, A. , Palmer, F. R. , , , , Palmer, G. , Palmié, S. Panconcelli-Calzia, G. Pankhurst, A. Pankhurst, R. , , , Panoff, M. , Paolillo, J. , Paradis, C. , Parker, M. , Parkvall, M. , , Partmann, G. , Pasch, H. , , , , , , , , , , , , , , , , Paster, M. , Pater, J. , Patrick, P. , , Patten, A. , , , , Paulme, D. , , Payne, D. L. , , , , , , , , , , , , , Payne, T. E. , Pearsall, D. Pearson, M. , Peck, S. M. Peli, G. , Peña, J. , Pereira, C. Perrin, M. , Perrot, J. , , , Peterson, D. A. Petráček, K. , , Petrollino, S. , Petros Banksira, D. , Petzell, M. , Peust, C. , , Pfeiffer, K. , Pfennig, C. , Philippe, K. Philippson, G. , , , , , , , , , , , , , , , , , Phillips, S. Phillipson, D. W. , , , Phillipson, R.
Pichl, W. J. , Pichler, W. , Piłaszewicz, S. , , Pillet-Shore, D. , Pinker, S. , Pinsonneault, D. , Piron, P. , , Pitman, A. , Plank, F. Plisch, U.-K. , Plog, S. , Ploog, K. , , , Polotsky, H. J. , , , Pope, B. Poppe, N. Porkhomovsky, V. , Poser, W. J. , , Posnansky, M. , , Potgieter, E. F. , Potsdam, E. , , , Potts, C. , Poulos, G. , , , Pouplier, M. Powers, J. L. , Pozdniakov, K. , , , , , , , , , , , , , , Praetorius, F. , Prah, K. K. , , , , , Preston, D. R. Prichard, J. C. Prince, A. , , , Prins, A. H. J. , Prokić, J. , Prost, A. , , , , , Proyart, L. B. , Prys Jones, S. , , Puglielli, A. Pulleyblank, D. , , , , , , , , , Pullum, G. K. Purpura, A. Pustejovsky, J. Qorro, M. (A. S.) , , , , Quack, A. Quack, J. F. , Quint, N. , , , , , , , , , , , Rabenoro, I. , , Raible, W. Raith, J. , Rampton, B. , , , , Rannut, M. Raue, D. Rebuschi, G. Reh, M. , , , , , , , , , , , , , , , , , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Reichmuth, S. , Reid, A. , Reid, L. A. Reineke, B. , , , , , , , , , , Reinhard, P. , Reinisch, (S.) L. , , , , , Reintges, C. H. , Reiss, C. , Renan, E. , Renaud, P. , Renfrew, C. , , Rensink, W. G. , Retsö, J. , Rexová, K. , , , Rey, V. Reynolds, K. H. , Rhodes, R. A. , Rialland, A. , , , , , , , , , , Ricard, A. Riccitelli, J. , Rice, K. , , , Ricento, T. Richards, N. , Richardson, I. , Richter, T. S. Rickford, A. E. , Rickford, J. R. , Ricquier, B. , Ridge, S. G. , , Riebe, T. Riedel, K. , Rijkhoff, J. , , Riley, C. , , Rilly, C. , , , , , , , , , , , , Ring, J. A. , Ritchart, A. Ritter, T. , Robbers, K. , Roberge, P. T. , Robert, Sh. bin , Roberts, I. Roberts-Kohno, R. R. , Robins, R. H. , Roddick, A. P. Roger, J.-F. , Rögnavaldsson, E. , Rohde, H. , Romaine, S. , , , Rongier, J. Roper, E. M. , , , , , , , , Ros, A. Rosalie, M. Rosch, E. , Rose, S.
Rose, Sh. , , , Ross, M. D. , Rossano, F. , Rössler, O. , Rothmaler, E. , Rottland, F. , , , , , , , , , , Roulon-Doko, P. Rovenchak, A. J. M. , , Rovenchak, A. , , , Rowe, J. , Rowe, J. A. , Rowlands, E. C. , Rubagumya, C. M. , Rubin, A. D. , Rudwick, S. , , , Rudy, S. Ruhlen, M. , Rüppell, E. , Rüsch, M. , , , , Rust, F. , , , , Rutherford, J. Rzewuski, E. Saa, F. Saah, K. K. , Sacks, H. , , , , Sacks, J. , Sadembouo, E. , Sadiqi, F. , Sadock, J. M. Saeed, J. I. , , , Said, E. , Sakel, J. Salffner, S. , Sallabank, J. , Samarin, W. J. , , , , , , , , , , , , , Sambiéni, C. , , , , , Sands, B. , , , , , , , , , , Sands, B. E. , , Sankoff, D. Santandrea, S. , , , , Sapir, J. D. , , , , Sasse, H.-J. , , , , , , , , , , , Satre, R. , Satzinger, H. , , Saussure, F. de , Savà, G. , , , , Schachter, P. , Schadeberg, T. C. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Schaefer, R. ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Schaefer, R. P. , Schegloff, E. A. , , , , , , , Schenkel, W. , Schicho, W. , Schildt, J. Schladt, M. Schleicher, A. , , , Schlenker, C. F. , Schlözer, A. L. von , Schmaling, C. H. , , Schmidt, S. , , , Schmidt, W. , Schmitt, A. , Schmitt, C. Schmitz, M. Schneider, T. , Schneider-Blum, G. , , , , , , , , , , , , Schneider-Zioga, P. , Schnoebelen, T. , Schoenbrun, D. (L.) , , , , , , , , , , Schön, J. F. , Schott, R. , Schrag, B. Schrag, B. E. Schreiber, H. , , , , , , , Schrock, T. (B.) , , , , Schröder, A. , , , , Schröder, H. , Schroeder, M. , Schuchardt, H. , Schuh, R. (G.) , , , , , , , , , , , , , , , Schultz, R. Schumann, T(h). , , , Schwabe, K. Schwarz, A. , , , , , , , , , Schwarze, C. Schweinfurth, G. , Schweitzer, S. D. , Scott, E. P. , Sebba, M. , , , , , Sebeok, T. A. , , , , , , , , , , , , , , Seelig, J. M. , Segal, D. Segerer, G. , , , , , , , , , Segert, S. Séguy, J. , Seibert, U. , Seidel, F. , , , , , , , Seifart, F. , , , , Seifert, M. , , , , , , , Seitel, P. , Sekere, N. B. , Selfe, C. L.
Seligman, A. F. , Seligman, B. Z. , Sells, P. Semadegbe, J. Senft, G. , Senghas, A. Serzisko, F. , , , , , Sewell, W. , Seydou, C. , Seyfeddinipur, M. , Sezer, E. Shange, M. Sharbatov, G. S. Sharif, A. , Shay, E. , , , Shimizu, K. , , , , Shinichiro, I. Shinya, T. Shisha-Halevy, A. , Shohamy, E. , , Shopen, T. Shore J. Showalter, S. D. , Sibomana, L. , Sidnell, J. , , , , , , Siebert, R. , Siegman, A. W. Siewert, K. , , Sigurðsson, H. Á. , Silué, J. S. Silva-Corvalán, C. Silverstein, M. , Sim, R. J. Simango, R. , Simeone-Senelle, M.-C. , Simon, C. , , Simons, G. F. , , , , , , , Simpson, A. Simpson, R. S. , Singler, J. V. , , , , , Singy, P. , Sisòkò, F.-D. , , Six, V. Skattum, I. , , , , , , , Skinner, N. , Skopeteas, S. Skutnabb-Kangas, T. , , , , Slabbert, S. , , Smakman, D. , Smidt, W. , Smieja, B. Smith, M. G. , Smith, N. , , , , , Smith, N. V. , Smith Stark, T. C. Smits, H. , , , , , Smolensky, P. , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Snider, K. , , Snyman, J. W. , , , Sodahlon, K. M. Sodji, P. K. Soma Sommer, G. , , , , Sonaiya, R. , Sorin-Barreteau, L. , Sottas, B. , , , Souag, L. , , , , , , , Sow, S. , , Spagnolo, L. M. , Spear, T. , Spears, A. (K.) , Speke, J. H. , Spencer, A. , , , Spitulnik, D. , , , , Spitzmüller, J. Spolsky, B. , , , Sportiche, D. Spriggs, M. , , Spyropoulos, M. , Stahl, A. B. Stahlke, H. , , , Stallcup, K. , Stanford, R. Stankiewicz, E. , Starosta, S. Starostin, G. S. , , , , , , , , Stassen, L. , Staubs, R. Stebbins, T. , Steeman, S. , , , Steere, E. , Stein, P. Stein-Kanjora, G. , Steinthal, H. , Štekauer, P. , , Stenzel, K. Stephen, M. Stephens, R. , , , , , , , , , Sterk, J. P. , , , , , , , , , Sterner, J. Stevens, C. J. Stevenson, R. (C.) , , , , , , , , , , , , , , , Stewart, J.-M. , , , , , , , , , , , , , Stirtz, T. M. , , , , , , , Stivers, T. , , Stolz, T. , Stone, R. M. , Stoneking, M. , Storch, A. , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
Streck, B. Streck, M. P. Strecker, I. , Street, B. V. Stroomer, H. , , Struck, B. , Strümpell, F. , Stuart, J. , Stuhlmann, F. , Sturtevant, W. C. Subbarao, K. V. , Sugahara, M. Sugawara, K. , , , , Suter, E. , , , Sutherland-Addy, E. Svetlá, J. Svolacchia, M. , Swadesh, M. , , Swann, J. Sweetser, E. E. , , Sy, J. H. Sylla, K. Tabakowska, E. Tabouret-Keller, A. , , , Tadadjeu, M. , , , Tadmor, U. Taine-Cheikh, C. , , , , , , , , , , , , Taiwo, R. , Takács, G. , , , , , , , , , , , Takassi, I. Tamomo, S. , Tanaka, J. , Tasch, M. Taylor, J. R. , , Tchagbalé, Z. Tchalalao, R. Tchiemouo, N. C. Tchiengué, B. Tchitchi, T. Y. , Teferra, B. , , , , Teixeira-E-Silva, R. , , Tekleselassie, A. A. , Tellier, C. , Tembenekuzu, G. Temu, A. , ten Have, P. , , Tera, K. Tersis-Surugue, N. , Tesar, C. Teshome, D. Tessendorf, S. Theil, R. , , , Thelwall, R. , , , , Thiam, Nd. (N.)
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Thomason, S. (G.) , , , , , , , , , , , , , , , , Thompson, L. C. Thompson, S. Thonhauser, I. , Thonner, F. , Thornell, C. , , , , Thurnwald, R. Thwala, N. Timberlake, A. Tirronen, T. E. , Tishchenko, S. , Tisserant, C. , Tohoun, B. Tol, S. Tollefson, J. W. , Tomalcheva, M. , Tompkins, B. , Tonelli, L. Tonkin, E. Topan, F. Toporov, V. Torreira, F. Torrence, H. , Torrend, J. A. , Tosco, M. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Tourneux, H. , Toweett, T. , Traill, A. , , , , , , , , , , , , , , Traoré, D. , Trask, L. Traugott, E. C. , , , Treis, Y. , , , , , Trigger, B. G. , Tröbs, H. , , , , , Tropper, J. Trotman, D. V. , , Truckenbrodt, H. , Trudgill, P. , , , Tsuge, Y. , Tuchscherer, K. , , , Tucker, A. N. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Tucker, R. , Tuckey, J. K. , Tuller, L. (A.) , , , Turay, A. K. , , Turner, N. Turton, D. , , Tutshek [Tutschek], L. ,
Tyers, P. Tyloch, W. , Uhlenbeck, E. M. , Uhlig, S. , , , , , Ulbricht, H. Unseth, P. , , , , , , Ura, H. , Urdze, A. Uriat, C. Vakunta, P. W. , Valdman, A. , , Valinande, N. K. , Vallaeys, A. , van Bulck, V. , van den Avenne, C. , van den Boogert, N. , van den Eynde, K. Vanden Plas, V. H. , , van der Auwera, J. , Vanderelst, J. , , , , , , , , , , van der Hulst, H. , , , van der Veken, A. , , van der Velde, M. , van der Vliet, J. van der Wal, J. , Vandeweghe, W. van Hout, R. Vanhove, M. , , , , , , , , van Olmen, D. , Vansina, J. , , , , , , , , , , , van Urk, C. , van Zyl, J. Varro, G. Vater, J. C. , Veenstra, T. Veit-Wild, F. Vé Kouadio, L. , Velupillai, V. , Venditti, C. , Venezky, R. L. Vergote, J. , Verma, M. , Vernet, R. , Véronis, J. Versteegh, K. , , , , , , , Vetter, C. Viehweger, D. Viereck, W. Vierke, C. , Vigasin, A. Vigouroux, C. , Vince, A. Vincent, N.
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Visser, H. , Voeltz, (F. K.) E. , , , , , , , , , , , , , Voigt, R. M. , , , , , , , , , Voigt, M. Voll, R. von Heusinger, K. von Lieven, A. , von Roncador, M. , , , , , , , , von Staden, P. M. S. , Voorhoeve, J. , , Vorbichler, A. , Vossen, R. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Vroon, R. Vydrin, V. , , , , , , , , , , , Vycichl, W. , , , Waag, C. , Wagner, D. Wald, B. Waller, R. , Wallin, L. , Wallis, B. M. , Wallvik, B. Walters, J. K. , Walton, M. , Wamala, E. Wardhaugh, R. , Warnier, J.-P. , Warschauer, M. , , , Watson, J. C. E. , Watson, R. , , , Watters, J. R. , , , , , , , , , , , , , , , , Watts, S. Weakley, A. J. , , Webster, J. B. , Wedekind, C. Wedekind, K. , , , , , , , , , , , , , , Wei, Li , Weijnen, A. , Weinreich, M. , Weinreich, U. , Weiss, D. , Wellens, I. Welmers, W. E. , , , , , , , , , , , , , , , , , , , Welsby, D. Weninger, S. Werner, A. , , ,
Wescoat, M. T. , Westendorf, W. , Westermann, D. (H.) , , , , , , , , , , , , , , , , , , , , , , , , , Westphal, E. O. J. , , , , , , Wetzels, L. , Whaley, E. J. , , , White, W. M. , White-Kaba, M. , Whiteley, W. H. , , , , Wiecha, K. , Wiemann, D. Wiering, E. , Wiering, M. , Wierzbicka, A. , , , Wilkinson, E. , Williams, E. , Williamson, K. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Willis, D. Willms, A. , Wilson, J. G. , Wilson, P. , Wilson, W. A. A. , , , , , Winand, J. , , Winford, D. , , , Winkelmann, K. , , , , , , , , , , , Winkler, S. Winston, F. D. D. , , , Wissemann, H. , Wittgenstein, L. , Wittmann, F. , Woidich, M. , Wójtowisz, B. Woldemariam, H. , , Wölfel, D. J. , Wolfer, C. , Wolff, H. , Wolff, (H.) E. , , , , , , , , , , , , , , , , , , , , , Woodbury, A. C. , Woods, C. , Worman, D. A. Wotzka, H.-P. , Wright, H. K. , Wright, R. , Wurm, S. , , , Yahya-Othman, S. , , Yakpo, K. Yankah, K. , Yigezu, M. , , , , , , , , Yimam, B. Yip, M. ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Yitah, H. , Younansardaroud, H. Zabbal, Y. Zaborski, A. , , , , , , , , , , , , , , , , , , Zahan, D. , Zalizniak, A. Zawada, B. (E.) , Zec, D. , Zeller, J. , Zemanek, P. Zemp, H. , , , Zerbian, S. , Zeshan, U.
Zheltov, A. , Ziamari, K. , Zibelius-Chen, K. , Ziegelmeyer, G. , , , , Ziervogel, D. , , , , , Zima, P. , , , Zimmermann, M. , , , Zoch, U. , Zoder, Y. Zohry, A. Zsiga, E. C. Zubairi ‘b Nasseem Zuckerman, B. Zuñiga, F. Zwicky, A. M. , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
L
..............................................................
Aasá(x) , Abasuba Abbey , Abidji , , , Aburé Abyssinian Acatlan Mixtec Acheron , , Acholi , , , , , Adamawa , , , , , , , , –, , , , , , , , , , , , -Eastern, see Adamawa-Ubangi(an) -Gur , –, , -Ubangi(an) , , , –, , , , , , Adele Adjukru , Afar (also ‘Afar) , , , , , , , , , , Afitti , , , , –, Afrasian, see Afro-Asiatic Proto- Afrikaans , , , , , –, , , , Cape (Coloured) Kaapse, see Cape (Coloured) Afrikaans Orange River Afro-Asiatic (also Afroasiatic) –, , , , , , , , –, –, , , , , , , , , , , , , , , , , , , , , , –, , –, , , , , –, , , , , , , , , , , , , , , , , Proto- , Agäw (also Agaw, Agau), see Central Cushitic Proto- –, , , , South Aghem Ahanta Ahlo Aja , , , , – Akan , , , , , , , , , , , , –, , , , , , Akanic Akpafu Aku Akye ,
Alaaba Alada Alagwa , , , , , , , , Albanian Ali Alladian , Alur Amãlo Amar(r) Amdang, see Mimi American Sign Language (ASL) (A)Merico Amerindian , , Amharic , , , , , –, , , , –, , , –, –, –, , , , , , , , , , , , , , Anfillo , Angolar Animere Ankole Anufɔ (also Anufo) , Anum Anyi Anyo Anywa , Aowin Arabic , , , , , , , , , , , , , , , , , –, , , , , –, , , , , , , , , , , , –, , , , –, , , , , , –, , – Cairene Chad(ian) , Classical , , , –, , Colloquial , Creole , Egyptian , , Jewish Jijel Juba Kwarandzyey , Levantine Libyan , Magrebi Moroccan , , , Nigeria(n) , Pidgin
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Arabic (cont.) Shuwa Sinai Standard , , , , , Sudanese (also Sudanic) , , , Tripolitanian Aramaic , , , Aramanik, see Aasá(x) Arbore , , Argobba , , Ari , , -Banna , Armenian Aroid, see South Omotic Arouro ‘Asaurta Ateso, see Teso Atlantic , , , , , , , , , –, –, –, , , , , , , , , , , , , , , , –, , North(ern) Proto- South(ern) , , West, see Atlantic Attié, see Akye Atuot Austronesian , Avatime Avikam , Awi , Awjila – Awŋi (also Awngi) , , –, , , , , , , , Awutu Ayt (also Ait) Seg(h)rouchen , Baa Baale , , –, – Babanki Babuckur Bacco Bade Badwe’e B’aga , Bagirmi , , , , , Bahnar Bai , Baïnounk , , , Baiso (also Bayso) , , Baka , –, , , , , Bako Balant Bamanankan, see Bambara Bambara , , , , , , , , , , , , , , , , Bambeshi-Diddesa
Bamileke , Dschang Fe’Fe’ Bamum , Banda – Banga Bangba , Baŋgi Me , Banna , Bantoid , , , , –, , , –, , , , , Bantu , , , , , , , , , , , , , , , , , –, –, , , , , , , , , –, –, , , , , , –, , , , , , –, , , , , , , , , , , , , , , , –, , , , , , –, , , , –, –, , , , , –, , –, , , , –, –, , Common , , , , , , Southern Grassfield(s) , , Jarawan Narrow , , –, , , , , , Northwestern Southeastern Southern North Nyanza Proto- Proto- , , , , , , , , , , , , , , , , , , , , -Potou-Tano , , Proto-East , Semi- , Southern , Baoulé, see Baule Barambo , Barea , , , Bari (Adamawa) Bari (Nilotic) , , Bariba , , , , Baricanci Basaa Basila Basketo Basque Bata Batwa, see ǁX’egwi Baule , Baygo Bayot , , , , Beboid – Bedawiye (also Bedauye), see Beja Beir-Didinga, see Surmic Beiya
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Beja , –, –, , , –, , , , , Belanda Bemba , , , , , , , Standard , , Town , , , , , , Benchnon , , Bende , – Benue-Congo , , , , , –, , , , , , , , , –, , , , , –, , , , , , , , , , , , Proto- Berber , , , , , , , , , , , , , , , , , , , , –, , , , , , , –, , , –, , , , , , , – Moroccan , Central Northern , Proto- , Riffian (also Rif-Berber), see Tarifit South-West Beria –, Bert(h)a , , , –, , , , , , , Proto- Berti , Bete (also Bété) , , , Bia Biafada , Biali , , , –, , Bijogo , , Bikwin , -Jen , , Bilin , –, , , , Biraile, see Ongota Birgid Birifor Bisa , , Boa-Kula , Bobangi Bobo , Bodi Boghom Boko , Bole Bom , , , , Bomassa Bom-Kim , – Bongo-Bagirmi , , , –, , Boni , , Bono Bourbonnais, see Réunion Creole Bowiri Bozo , , Breton ,
Brong, see Bono Bua , –, , Bubi , , Buduma Buli , Bulu Bura Burak Bure Burji , , , -Geleba Burunge , , , , , , , Burushaski Bussa Bviri – Bwaka Bwamu , , , Bworo Byali, see Biali Cahi Rimi Cama, see Ebrié Cameroonian Pidgin Camfranglais , , , , –, , Camspeak, see Camfranglais Canaanite , , , Cangin , Cara , , –, , C’ara Carib Casamancese Creole Cɛfɔ , , Celtic Cerma , , , Chadic , , , , , , , , , , –, , , , , , , , –, , , , –, –, , , , , , , , , , , Proto- , , Chaga , Chaha , , Chai , , , , , Chakosi, see Anufɔ Chaldean Cham Chamba -Daka -Namshi , , Chari-Nile , , , , Chasi, see Alagwa Chasu, see Pare Chewa (also Chichewa) , , , , , , , , Chimwiini, see Mwiini Chinese , , , , , Chori , , Chumburung
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Cibak Cirecire C’Lela Coptic , –, , , , Old Sahidic Crioulo Croatian Cross River –, Cushitic , , , , , , , , , , , , , , , , –, –, –, , , –, –, –, , , , , , , , , , , , , , , , , –, Central , , –, , , –, , , –, , , –, , , , , East(ern) , , , , , , , , , , , , , , , – Highland , , , , , , Lowland , , Fringe , , Lowland North(ern) , , , , , Proto- , , , , South(ern) , , , , , , , , , , , , , , Lowland West, see Omotic Cuuramma , Daats’iin Daba Dagaare-Waale-Birifor Dagara (also Dagaare) , , , –, – Dagara-Lobr Dagbane (also Dagbani) , , Dagik , –, , , , Dahalo , , , , , , , Daju (also Dagu) , , , , , , , , , , , , of Dar Dagu of Darfur of Eref , of Lagowa , of Western Kordofan Dakoid –, Daloka, see Ngile Damara , Dan , , , , –, , , – Eastern Santa Danakil, see Afar Dangla , , East- Dangme , Datooga –, Dawrooa ,
Day , , Dazaga Defaka , , , Demotic , , Dendi , , , , Dengebu, see Dagik Dewoin Dghwede Dhaasanac (also Dhasaanach) , , Dholuo, see Luo (Nilotic) Didinga , , , , , , , , , -Murle , , , Dime , Dinik, see Afitti Dinka , , , , Diola-Fogny Dioula (also Djoula), see Jula D’iraasha (also Dirayta) , Dizi Dizoid Djandjörou, see Janjero Djenné Chiini Dogbo Dogon , , , , , , , , , –, , Dɔgɔsɛ Dogose-Ka̰a̰sa (also Kaansa-Dogose) –, Dokko Dompo , , ‘Dongo(ko) Dongola(wi) , , , Dorze Doyayo Duala , , , , Duli Dullay –, , Proto- Dutch , , Cape – Pidgin Dyimini Dyirbal Dyula, see Jula Eastern Bushman, see ǁX’egwi Eastern Sahelian, see Eastern Sudanic East Kalahari (Khoe) , , Ebira Ebrié , Edoid , Efik , , , Efutu Ega , , , Egyptian , , –, –, , , , , , Ancient , , –, , Ejagham Ekoid – El-Amira
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Elfoqaha – Eliri, see Nding Elmolo , , , Emai Emghedesie Engenni English , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , –, , –, , , , , , –, , , , , , , , , , –, , , , , –, , , , , , Fernando Po Creole, see Pichi Liberian Settler, see (A)Merico Old Pidgin , Cameroon(ian) , Kru Nigerian Pidgin , West African South African West African Creole Engsh , Eotilé Etruscan Ewe , , , , , –, , , , , , , , , , Adangbe dialect of Anlo , Peki Ewondo Pidgin , Populaire, see Pidgin Ewondo Fa d’Ambô Fali , Fanagalo (also Fanakalo) , , , Fang Fante , Farifare Feroge Figuig , , Fiome, see Gorwaa Flemish Flytaal, see Tsotsitaal Fon Fongbe (also Fon Gbe) , , , Forro Francanglais, see Camfranglais Franglais, see Camfranglais French , , , , , , , , –, , , , , , , , , , , –, , , , –, , , –, , , , , , , , , , –, , , , , , Pidgin, see Camfranglais
Fula , , , , , , , , , , , Fulani , , , , Fulfulde , , , , , , , Fur , , , , , –, , , , , Fwe Ga , , , -Dangme , , Proto- , Gã , Gaahmg , , , , , , , , , , Ga’anda Gabra Gabri Galaboid Galambu Galla, see Oromo Gamo , Gǁana , Ganda, see Luganda Ganza , Ganzo Gawwada , , Gbanyito Gbanziri , Gbaya , , –, -Bogoto Gbe –, , , Proto- – Gbeya Gbwaga Ge’ez , , , , , Geleba -Mogogodo Geme Gəmme Gen , , Gera German , , , , , , , , , , , –, High Germanic , , , , , , West Gewe Ghadames –, Ghana-Togo-Mountain , , , , G(h)imira , , Ghomara , Ghulfan Gidar Gidole, see D’iraasha Gĩkũyũ (also Gikuyu), see Kikuyu Gikyode Ginyanga Gio
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Giziga Glavda Gobu Godié Goemai Gogo Gola , , Golo , Gomara Gomba , Gonga , Gonja , , Gorwaa , , , , , Goundi Gourara Great Lakes , , Grebo , Greek , , , , , , , , Guanche Guang , Gude Guduf Guéré Gueve, see Gewe Gǀui , , , , Gula Mere Gumuz , , , , , , Gun Gungbe Gunu Gur , , , , , , , , , , , , , , , , , , –, , , , , , , , , , –, Central , , , , , , , , North(ern) , Proto- , , , Southern -Oti-Volta , Southwestern Gurage , , Gurene, see Farifare Gurma –, Gurunsi , –, , , , East Northern , Proto- Gvoko Gwama Gwandara Hadiy(y)a , , Hadza , , , , , , , , , , , , , Haitian , Hamar , Hamitic , , , –, , Harari , ,
Hassaniya Hausa , , , , , , , , , , , , , –, , , , , , , , , , , , , , , , , , , , , –, , Haya , , Hebrew , , , , , , , Hehe , , Heiban , , , , Heibanian , , , , –, , – Proto- , , –, , , , -Talodi Herero , Higi Higir Hima Hindi , , Hlonipha , –, Hona Hozo -Sezo Hungan Hwe Ibibio Icetod, see Ik Idomoid Igbo , , , –, , , , Igboid Igede , , Ijo , , , Ijoid , , , , , , , Ik(a) , , , , , , , Ila Ilwana Imdlawn Tashlhiyt Indo-European –, , , –, , , , Proto- Indoubil , Indri Ingassana, see Gaahmg Iraqw , , , , , , , –, , –, –, , , , , , , , , Irimba , Irish Gaelic Is(i)camtho, see Tsotsitaal Italic Ja̰a̰ne , –, , Jalaa , , Jamaican (Creole) , , Janjero , Japanese , , , Jarawan –, Jebel , , , , , , , , , Jen , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Jita Jɔ Jo(o)mang, see Tasomi Ju, see North(ern) Khoisan Proto- Juǀ’hoan , , , , Jukunoid , , , , , Proto- Jula , , –, , Kaachi Kaado , , Ka̰a̰sa –, Kabba , – K’abeena Kabiye , Kabuverdianu Kabyle – Kachipo, see Baale Kadallo Kadaru , Kadu, see Kadugli(an) Kadugli(an) , , , , , Kadugli-Krongo Kafa-Mocha Kaf(f)a – Kaficho Kainji , Kalenjin , , Kam, see Nyiŋom Kamba , , , Kamba(a)ta , , Kanakuru , , Kanembu , , , Kanuri , , , , , , –, –, Kara , , Karanga , Karekare Karko , , Karre Kasanga , Kasem Kasena Katcha Katla , , –, Katloid , , , , , –, Proto- -Rashadian Kavango Kebi-Benue , , , Kebu Kefoid Keliko Kemantney –, , , Kenyang , Kenyan Sign Language (KSL) Kenzi ,
Kerre Khasonke Khoe, see Central Khoisan East Kalahari Proto- West Kalahari , Khoekhoe , , , , , , Khoekhoegowab , , Khoisan –, , , , , , , –, , –, , , , , , –, , –, , , , , , , , , , , Central (South African) , , , , , –, , , , , , , , –, , , , , , , Proto- North(ern) , , , , , , , , , , , South(ern) , –, , –, , Khwe , , , –, Kibalele , , Kikongo , , , , , , , -Kitúba, see Kituba Kikuyu , , , , , Kilba Kim (Adamawa), see Kuasap Kim (Atlantic) , , , , Kimbundu , Kinande , , , , , , Kindiga, see Hadza Kinga Kingwana Kinubi, see Nubi Kinyarwanda , , , , , , Kipsikiis , Kir -Abbaian Kisetla Kisi , , , , , –, Kituba (also Kitúba) , Kivita Ko , , Koalib , , –, –, –, , , , Kobiana Kolisi , , , , , –, , , , , – Kom Koman , , , , , , , , Komba Komo , Kongo, see Kikongo Kongo-Saharan Konkomba , , Konni , Kono , Konso , -Geleba Koonzime
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Koorete Korandjé, see Kwarandzyey Kordofanian , , , , –, , –, , , , , , Proto- , –, Korean Koromfe , , , Kotoko , , , , , Kotokoli Koyra Chiini Kpatiri Kpelle , , , – Kposo Krachi, see Kaachi Krakye, see Kaachi Kresh , , , Krio , , , , Krobu Kroo, see Kru Pidgin English Kru , , , , , , , , , , , , , , , , Kua Kuasap , , Kujarge , , Kulaal Kulango , –, , Kuliak , , –, , , , , , , Proto- Kumam , , –, , , , , , Kunama , , , , , , , , , , , , , Kunfäl, see Kolisi Kusaal , Kusunda Kwa , , , , , , , , , , , , , –, , –, , –, , , , , , , , , –, , , , , , , -Kru Proto- , Kwadi , , Kw’adza , , Kwama Kwangali Kwarandzyey , , – Kyama, see Ebrié Kyerepong, see Okere Laal –, Laɓi Lafofa , , , , –, , Lama , Lamang Langi Lango , , , , Langue des Signes Franco-Africaine Laro
Larteh, see Lɛtɛ Latin , , , , , , , , , –, , , , Lefana Leggbó Lele , Lelemi , Lendu , , –, Lenje Lɛtɛ Liguri Likpe , , , , Limba , , Linda Lingala , , , , Lobiri , Logba , , Logo , , , Logol, see Lukha Logoli Lolobi Loma , , , Lomongo , Longarim Longuda –, Lɔŋto Lovale Lozi , Luba , Luganda , , Lugbara , , , , Lukha Lumun , , , , , , Lung’ie, see Principense Luo (Cameroon) Luo (Nilotic) , , , , , Luri Lushi Luwo , Luyia Lwoo Southern , Lyele , Ma Maa, see Maasai Ma’a (also Ma’á) , , , , , – Maale , Maasai , , , , , , , , , , , , –, Maba , , , , , Maban , , –, –, , , Mabia, see Gur Ma’bo , Macushi Madi (also Ma’di) , , , , , Mafa-Mefele
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Majang , , , , , , , Maji Makhuwa Makonde Makua Korovere dialect of Malagasy , , , , , , , , Malay , Mali Malo Maltese Mambiloid –, Mampruli (also Mamprule) , , , Mandara , , , Mande , , , , , , , , , , , , , , , –, –, –, , , , , , , , –, , , , , , , , , , , , , , , , Central , Eastern , – Northern , , Proto-Northwestern Proto-Southwestern South-Eastern Southern South-Western , Manding , , –, , , , , , , , , Mandinka , , Mandombe Manga Mangaya Mangbetu , , , , , , – Mani , , , , Maninka , , Manja Manjaku , Mankanya –, Mankon Manza -Gbaya -Gbea Mao , – Northern Southern Marachi Margi , , -South Marka , -Dafing , Masa Masakin, see Dagik Masalit Mascarenian Masongo Matuumbi ,
Mauritian Mauritius Creole Maxi Mayogo -Mundu Mba , , – Mbane , , , Mbanza Mbatto Mbay , , , , Sara Mbodɔmɔ (also Mbodomo) , Mbugu, see Ma’a Mbui , Mbuko , Mbukushu , Mbulu, see Iraqw Mbulungwe, see Burunge Mbum-Mangbai Mbuun Mbwaka Me’en , , Mekan Mel , Mende , , , , , , , Merarit , Merina Meroitic , , , , , , , , Meru Mesoamerican , Metyibo, see Eotilé Meyobe , Midob , Migama Mikeyir, see Shabo Mimi , , of Decorse , , of Nachtigal , , Mittu Miya , , , , Miyobe Moba , Mofu-Duvang Mofu-Gudur , Mole Mongo , , , Mon-Khmer Mono Monzombo Moore (also Mooré) , , , –, , Morisyen, see Mauritius Creole Moro , , , -Mangbetu Moru , , –, –, – -Madi , , -Mangbetu , , , Mɔyɔbɛ
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Mpiemo Mpiin Mpre (also Mpra) , Mpyemo Mubi Mumuye , , , Mundang Mundu –, Mupun , Murle , , , , , , Mursi , , Murzu Musey Musgu Mushere – Muskum Muyang Mwaghavul Mwiini , , Mzab , , Nafaanra , Naki , Nalu , Nama , , Namwanga Nancere Nandi , Nao N!aqriaxe Nara , , , , , , , , , –, , , Naro , , –, Naro-Gǁana , Nateni , , , Nawdm , , , Nawuri Nayi Ncam , Nchumuru Ndakko Ndebele , Northern Ndemli Ndi Nding , Ndogo – Ndonga Ndungale , , Nefusa – Nera, see Nara Ngaalam , Ngas , , Ngbaka , – -Gbaya Ngbandi , –, -Sango
Ngbugu Ngem Ngile , , , Ngizim , , , Ngomba Ngomvia, see Kw’adza Ngong Ngoni Nguni , , , Niger-Congo –, , , , , , , –, , , –, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , –, , , , , , , , , , North Central Proto- , , , , , , , Niger Dendi Niger-Kordofanian , , , , Niger-Saharan Nilo-Saharan –, , , , –, , –, , , , , , , , , , , , , , , , , , –, , , , , , , , , , , , , , , , , , , , –, , , , , , , , , Northeastern , Proto- Nilotic , , , , , , , –, , –, , –, , –, –, , , , , , , , , , –, , , , , , , , Eastern , , , , , , , , , , , , , , , Proto- , , Southern , , , , , , , , , Proto- Western –, , , , , , , , , Ninkare Njyem Nkonya Nobiin , , , Nouchi , –, –, , , Nsong Nubi , , Nubian , , , , , , , , , , , , –, , , , , , , , , , Darfur Dongolese, see Dongolawi Kordofan , , , , , , Nile , , , , , , Nuer , Nuni Nupe , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Nupoid , Nǀuu , , , , , Nyagbo , , , Nyalgulgule (also Njalgulgule) , Nyang Nyangiyan, see Kuliak Nyanja , , , Nyima , , Nyimang , , , , , –, , Nyiŋom , Nyun –, , Nzakara , , , Nzema Ogiek Okere Ometo , – Northwest Southeast Omo-Tana , , –, Omotic , , , , , , –, , , , , , , , , , , , North , Proto- , South , –, Western Ongota , , , , Ongotan Orig , , , , , Orma Oromo , , , , , , , , –, , , , , Oromoid – Oropom , Oti-Volta , , , –, East(ern) – West(ern) , , , Otoro , , Ouargla Oxoryok Oyda Pambia Pana , , , , – Pare , , – Päri , , , , Pastaza Quechua Pecí Pedi Pero Persian , , , Petit Mauresque Phla-Pherá Pichi Pidgin A , see Pidgin Ewondo Plateau , Podoko
Portuguese , , , , , , , , –, , , , Brazilian Creole Potou , Potou-Akanic Potou-Tano , , , Proto- , , Principense Proto-Iraqwoid Proto-Sabaki Proto-Upper Cross Pula(a)r , , Punic , Qafar, see Afar Quara , , , , Qw’adza, see Kw’adza Rashadian , , –, , , Eastern Proto- , , , Western Rendille , Rere Réunion Creole Rodriguan Creole Romance , , , , , , , Ron-Bokkos Ronga Rub, see Kuliak Rundi , , , Rwanda, see Kinyarwanda Sabaot Sabir Saboba Saharan , , , , , , , , , –, –, , , Central , East Proto- , Saho , , , -Afar , , –, , , , , Sam , Samburu , Samba-Duru – Samo Northern , Samoma , Samu , , , Samwe , Sandawe , , , , , , , , –, Sango , , , , , , , , , , , Sanskrit , Santrokofi ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Sanye, see Dahalo São-Tomense Sara , , , Mbai (also Mbay) , Seereer (also Serer) , Sehwi Semai Seme Semitic , , , , , , , , –, , , , , , , , , , , –, , , , , , , , , , Central East Ethio- (also Ethiopian Semitic) , , , , , , , , , , –, , , , , , , , , Hamito-, see Afro-Asiatic Northwest Proto- – South , Senar , Senha(d)ja de Sraïr , Senni Senufo , , , –, Proto- Serbian Sere – Seselwa, see Seychelles Creole Seychelles Creole Shabo , –, Shambaa , , Shanjo Shatt , Shay Sheko , Sheng , , , , , , , , , , , Sherbro , , Shinasha Shipki Shirumba, see Shwai Shona , , , , , , , , , , Shua Shukriyya , Shwai Sidaama , , – Sidama , – Sidamo, see Highland East Cushitic Sinyar , , – Siwa – Siwu , , , Sokna – Soli Solla, see Meyobe
Somali , , , , , , , , , , , , , , , –, , , , , , , , , , , Sign Language Songhay (also Songhai) , , , , , , , , , , , , –, , Central Northern , –, , Northwestern Proto- , , , , Soninke , –, , , , , Sotho , , , , , , , Northern Southern , , , Spanish , , , , , , Sua , Subiya Sudanic , , , , , , –, Central , , , , , , , , , , , , , , , , , , , , , , Eastern , , , , , –, , –, , , , , Northern , , – Southeastern Southern , Macro-, see Chari-Nile Northern Proto- Suma Sumerian , Sungor Supyire , , , , , , Suri , -chai Surmic (also Surma) , , , –, –, –, –, , , , , , –, , , , Northern Southern , , Southeastern , Proto- Southwestern , Susu , Swahili , , , , , , , –, , , , , , , , , , , , , , , , , –, , , , , , , , , , , , , , , –, , , , , , , – colloquial Katanga, see Lubumbashi Swahili Lubumbashi Shaba, see Lubumbashi Swahili Standard Swati (also Swazi) Syer , Syrian
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Taa Tabi, see Gaahmg Tachoni , Tadaksahak , , , – Tado Tafi Tagbo Tagdal Tagle(nnaa) , , , Tagoi , , –, , , , , Taita Dembwa Mbololo Talodi, see Tasomi Talodian , , , –, –, , Proto- , , , , , Tama , , , , , , Taman , , , , , , , –, , , , Tamashek (also Tamasheq) , Tamazight , , , Tamil Tangale Tano , , -Akan Central -Guang Western Tarifit , , , , , , , , Tasawaq , Tashelhiyt – Tasomi , , , , , , Teda -Daza , – Teen , Tegali , , , , , Tegem , , Teis-um-Danab Telegu Tem , Temein , , , , , , Temeinian , , –, , , , , , Temne , , , Tenda Tennet , , Tepes Tera Terik Teshena Teso , Teso-Turkana , , Tetserret , , , Teuso , Tigre , , Tigrinya , , –, , , , , , Tikar –
Tima , –, , , , –, , Tindiga, see Hadza Tira , , , , , Tiriki Tirma(ga) , , , , , Tiv , Tivoid – Tmessa Tocho , , , , , Togo Remnant, see Ghana-Togo-Mountain Tolona Tondi Songway Kiini Tonga Tongwe Toposa , Totela Tsamai , , Tsamakko Tshwa , Tso , Tsonga , , , Tsotsitaal , , , , –, , , , , Tsua Tswana , , , , , , , , Tuareg , –, , , , , –, , , Tubu, see Teda-Daza Tula , , , , -Longuda , -Waja –, , Tulai Tumale Tumtum, see Kadugli(an) Tupuri , , Tura –, , Turjuk , Turkana , , , , , , , , Turkic Turkish Turku , Tusian , Tuu, see South(ern) Khoisan Tuwal , , , Twi , , , , , Akuapem Asante Ubangi(an) , , , , , , , , , , , , –, , , , , , , , , , , , , Cameroun- Uduk Ugandan Sign Language Ukaan Umbundu Uncu(nwee) , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
Urubu-Kaapor Uto-Aztecan Vagla , Vai , – Vata , Vəmnəm Venda , Vhe, see Ewe Vidunda , Viemo , , , Volta-Congo , , North , , Voltaic, see Gur Waale Wací Waja , , , , , , –, Wambo (also oshiWambo) Wandala , , –, – Warata Warji Warnang, see Werni Werizoid, see Dullay Werni , , , , Weyto Win , , Wobe (also Wobé) , , Wolaitta (also Wolaytta) , , , Wolof –, , , , –, , , , , , , , , , , , , , Xamir, see Xamt’aŋa Xamt’aŋa –, –, , , , Xhosa , , , –, , , , , –, , , , , – Yaaku , –, , Yagua Yakoma Yakuba (also Yacouba) , Yala Ikom Yamna Yangara Yao Yeji Yem Yemne-Kimbi Yendang , Yeni, see Baŋgi Me
Yeyi Yoruba , , , , , , , , , , , , , , , , , , , , , , , , Ife dialect of Proto- Standard , Yoruboid Yukatec Maya Yukuben , –, –, , , , , , Yungur – Zaar Zaghawa, see Beria Zande , , , , –, –, – Zarma – Zayse , Zeem Zenaga , , –, –, , Zenata Zenatic – Zergulla Zerma Zigua Zilmamu, see Baale Zuara – Zulgo Zulu , , , , , , , , , , , , , , , , , , , , , Zumaya ǀAuni ǀGwi (also ǀGui), see Gǀui ǀXam ǂ’Amkoe ǂHaba , ǂHoan (also ǂHõã) , , , ǂKhomani ǂKx’auǁ’ẽ ǃUi , ǃXóõ , , , Eastern Lone Tree (also !Xoon) , Western !Xũ, see !Xu(u)n !Xu(u)n , , , , , Ekoka , ǁGana, see Gǁana ǁX’egwi , , –, –
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
S...................................................... I adjective , , , , , , , , , , , , , , , , , , , , , , , –, –, , , –, –, –, , , , –, –, –, , , , , , –, , , , , , , –, –, , , , –, , –, –, , , , , –, , –, , –, , Africanism , , , –, –, agreement , –, , –, , –, –, , –, , –, , , , , , , , , , , , , , , , , , –, , –, –, , , , , , , –, , –, , animacy , , , , , , , areal , , –, , , , , , –, , , , , , –, , , –, , , , , , , –, –, , , , , , , , –, , , , classification , , feature , , , , , , , contact , , , , –, , , linguistics , , , , typology , , , aspect –, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , –, –, , –, , , , , , , –, , –, , –, , , –, , , , , associative (see also possessive) , –, , , , , –, , , –, , , , , , –, –, , , , ATR harmony –, , , , , –, , , , bilingual(ism) , , , , , –, , , –, –, –, , –, , , , , , , , –, , , – borrowing , , , , , , , , , –, , , , –, , –, –, , –, , , , , , , , , –, , , , –, , , , , –, –, –, –, –, , , , , –,
case , , , , –, , , , , , , –, –, –, , , , , , –, , , –, –, , –, , , , –, , , –, –, , , , –, , , –, –, , , , , code mixing , , code-switching , , –, , , –, , –, –, , , , , –, , , , cognition , , , , , , –, –, , –, , , colonialism , , , –, , , , , , , , , –, , , , , –, –, , , , , , , , , , , , , , , , , , –, , , , complementizer –, , , , , , , , , , compound(ing) , , –, –, , , , , , , , , , , , , , , , , , , , , , , , , –, , , , , , , , , conjunction , , , , , , , , , , , , , , , , , , – consecutive (see also narrative) , , , , , –, constituent order (see also word order) –, , , –, –, , , , , , , , , , , construct case (see also possessive) , , , , , , converb , , , , , , –, , –, convergence , –, , , , , , , –, , , , , –, , , , , –, , , , , –, , , –, , coordination , , , , , , , , definite(ness) –, , , , , , , –, , –, , , , , –, –, , , , , , –, –, , , , , –, , , , deixis , , , , , dialectology , –, , –, , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
diaspora , , , , , , , –, –, , diathesis (see also voice) , , – direct speech , , –, – ejective , , , , , , –, , , , , , , , , , , , endangered (languages) , , , , , , , –, –, , , , , , , , , –, , , , –, , –, , , , , , , , evidential , , , focus , , , –, –, , , , , , , –, –, –, –, , –, , , , , , , , , , , , , , , , , , frequentative , , , , , gender , , , , , –, , –, , , , , , –, –, –, , , , –, , –, –, , , , , , , –, –, –, , –, , , –, –, , , , , –, , , –, , , –, , , , , , , , –, , , , , –, , genetic classification –, , , , , , , –, , , , , , , , , –, , , –, globalization , , –, , , , grammaticalization , , , –, , , , , , , –, , , , , , , , , , , , , , , , heritage , , , , , , , , , , , , , , , ideophones , , , , , , , , , –, implosive , , , , , , , , , , , , , , , indefinite(ness) , , , –, , , , , , , , , , , , –, –, , , , , , , indirect speech , interjection , , , , , , , , interrogative(s) (see also question) , , , –, , , –, , , , , , , , , , –, , , , –, –, , , , labial-velar , , –, , , , labio-velar ,
language contact, see areal contact policy , , , , , , , , , , , light verb , , , lingua franca , , , , , , , , , , , , , –, –, , –, , literacy , , , , –, –, –, –, –, , , –, , , –, , , logophoric(ity) , , , –, , –, , –, , , , –, , metaphor(ic) , , –, –, –, , , –, , –, , , , , , , , –, –, , metonym(ic) , –, , –, metric(al) , , , , , , , middle, see voice modality , , , , , , , mood , , , , , , , , , , , , , , , , , , , , , , –, –, , , , , , , , , mora , –, , , , multilingualism , –, , , , , , –, , , –, , , , , , , , , narrative , , , , , , , , –, , , , , –, , , , , –, , , , nasal vowels , , , , , , , , , –, –, , , –, , , , , , national language , , , , , , , , , , , , , , , negation (see also polarity) , , , , –, –, , , , , –, , , , , –, , , , , –, –, , , –, , , , –, , , – non-verbal communication –, , , , , –, , , , , , , noun class(es) –, , –, –, –, , , , , –, –, , –, –, , , , –, , –, , –, , –, , –, –, –, , , –, –, , , , , , , –, –, –, , , , , , , , –, –, –, , , , , , –, , number , , , , , , –, –, , , , , , –, , , , , , , –, , , –, , , –, , , , , –, , , , , –,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
, –, –, , , , , –, , , , , , , –, , , , , , , –, official language , –, , –, , , , , , , , , , , oral(ity) , , , , , , , , , , , , , , , , , , , , , , , , orthography , , , , , , –, –, , , , –, –, , passive, see voice person , , , , , –, , –, , –, , –, –, –, –, , , –, –, –, , , –, , –, , , , , –, –, –, , , , , –, –, , –, , –, –, –, –, , , , –, , , –, –, , , , pluractional , , , –, , , –, , , , polarity (see also negation) , –, , , , –, –, polite(ness) , , , , , , – possession , , , , , , –, , , , , , , , possessive , , , , , , , , , , , , –, , , , , –, –, , –, , –, –, –, , , , , –, , –, –, , , –, , , –, , –, , –, –, , , , , pragmatic(s) , , , , –, , , –, , , , –, , –, , , , , , , –, –, , , –, –, proverb(s) , , , , , , , –, , –, – question(s) (see also interrogative(s)) , –, –, , , , , , , , –, –, , , , , , , , , –, –, , , , , , –, , , –, –, , , reconstruction(s) , , –, , , , –, –, , –, –, –, –, , , , –, –, , , , , –, , , , –, –, , , , –, , , –, , , , , , –, , , , , , –, , , , –, , , –, , , –, , –, , –
register , –, –, , , , , , , , , , –, , , –, , , , , relative clause –, , –, , , , , , , , , –, , , , , –, , , –, , , , , , , repertoire(s) , , , , , –, –, , , , , , , , , , , revitalization , , , –, , , , script , , , , , –, , –, –, , , , –, , –, , – sequential (see also narrative) , , , , , serial verb(s) , , , , , , , , , , social network(s) , –, , , –, , speech act , –, , –, , standardization , , , , , , , , –, , , subordinate clause , , , , , , –, , , , subordination , , , –, , subsecutive (see also narrative) , syllable , –, –, –, , , –, , , –, , –, , –, , , , –, , , , , –, , , –, –, –, –, , , , –, , , , , , , –, , , , –, , , , , –, , , , –, –, , , , , tense , –, , , , , , , , , , , , , , , , , , , , , , , , –, , , , , , , , , , , –, , , –, –, –, –, –, , , –, –, , , , –, , , , , tone –, , , , –, , , , –, , , , –, , , , , –, , , , , –, , , , –, –, –, –, , , , , , , , , –, , , , –, –, , , , –, , –, –, –, , , –, , , , –, , , –, , –, –, , –, , , , , –, –, –, , –, , , , , , , –, , –, contour , –, , , , , , , ,
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
tone (cont.) downstep , –, –, –, , , , , –, , floating , , –, , , , , , , –, grammatical , , , , , , lexical , , , , , , , , melodies , – tone-bearing unit , , , upstep , , topic(alization) , , –, –, , , , , , , , , , , , –, , , , , transitivity , , –, , , ditransitive , , , , –, , , intransitive , , –, , , , –, , –, , , , , , , –, , , , , –, –, , , –, –, , –, labile , , , , transitive , , –, , , , , , , , –, , , , , , , , , , , , –, , , , , , , , , –, , , , , , –, – translanguaging ,
verb(al) extension –, , , , , , , , –, , , , –, , , , , , , , , , , , – valence , , , , , , , valency , , , voice (see also diathesis) , , –, , , , mediopassive , , middle –, , , –, , , , , , –, , , –, , passive , , –, , , , , , , , , , , , –, , , , –, , , , , –, , –, , –, , , , , word order (see also constituent order) , , , –, , –, , –, , , , , –, , , –, –, , –, , –, , , –, , , , , , –, , , , , , , , , , , , , , , , , , writing , , –, , , , –, , –, –, –, , , , –, , –, , , –, –, , , , , –, , , , –, –, –, , –
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
OXFORD HANDBOOKS IN LINGUISTICS THE OXFORD HANDBOOK OF AFRICAN AMERICAN LANGUAGE Edited by Sonja Lanehart
THE OXFORD HANDBOOK OF AFRICAN LANGUAGES Edited by Rainer Vossen and Gerrit J. Dimmendaal
THE OXFORD HANDBOOK OF APPLIED LINGUISTICS Second edition Edited by Robert B. Kaplan
THE OXFORD HANDBOOK OF ARABIC LINGUISTICS Edited by Jonathan Owens
THE OXFORD HANDBOOK OF CASE Edited by Andrej Malchukov and Andrew Spencer
THE OXFORD HANDBOOK OF CHINESE LINGUISTICS Edited by William S-Y Wang and Chaofen Sun
THE OXFORD HANDBOOK OF COGNITIVE LINGUISTICS Edited by Dirk Geeraerts and Hubert Cuyckens
THE OXFORD HANDBOOK OF COMPARATIVE SYNTAX Edited by Gugliemo Cinque and Richard S. Kayne
THE OXFORD HANDBOOK OF COMPOSITIONALITY Edited by Markus Werning, Wolfram Hinzen, and Edouard Machery
THE OXFORD HANDBOOK OF COMPOUNDING Edited by Rochelle Lieber and Pavol Štekauer
THE OXFORD HANDBOOK OF COMPUTATIONAL LINGUISTICS Edited by Ruslan Mitkov
THE OXFORD HANDBOOK OF CONSTRUCTION GRAMMAR Edited by Thomas Hoffman and Graeme Trousdale
THE OXFORD HANDBOOK OF CORPUS PHONOLOGY Edited by Jacques Durand, Ulrike Gut, and Gjert Kristoffersen
THE OXFORD HANDBOOK OF DERIVATIONAL MORPHOLOGY Edited by Rochelle Lieber and Pavol Štekauer
THE OXFORD HANDBOOK OF DEVELOPMENTAL LINGUISTICS Edited by Jeffrey Lidz, William Snyder, and Joe Pater
THE OXFORD HANDBOOK OF ENDANGERED LANGUAGES Edited by Kenneth L. Rehg and Lyle Campbell
THE OXFORD HANDBOOK OF ELLIPSIS Edited by Jeroen van Craenenbroeck and Tanja Temmerman
THE OXFORD HANDBOOK OF ENGLISH GRAMMAR Edited by Bas Aarts, Jill Bowie, and Gergana Popova
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
THE OXFORD HANDBOOK OF ERGATIVITY Edited by Jessica Coon, Diane Massam, and Lisa deMena Travis
THE OXFORD HANDBOOK OF EVENT STRUCTURE Edited by Robert Truswell
THE OXFORD HANDBOOK OF EVIDENTIALITY Edited by Alexandra Y. Aikhenvald
THE OXFORD HANDBOOK OF EXPERIMENTAL SEMANTICS AND PRAGMATICS Edited by Chris Cummins and Napoleon Katsos
THE OXFORD HANDBOOK OF GRAMMATICALIZATION Edited by Heiko Narrog and Bernd Heine
THE OXFORD HANDBOOK OF HISTORICAL PHONOLOGY Edited by Patrick Honeybone and Joseph Salmons
THE OXFORD HANDBOOK OF THE HISTORY OF ENGLISH Edited by Terttu Nevalainen and Elizabeth Closs Traugott
THE OXFORD HANDBOOK OF THE HISTORY OF LINGUISTICS Edited by Keith Allan
THE OXFORD HANDBOOK OF INFLECTION Edited by Matthew Baerman
THE OXFORD HANDBOOK OF INFORMATION STRUCTURE Edited by Caroline Féry and Shinichiro Ishihara
THE OXFORD HANDBOOK OF JAPANESE LINGUISTICS Edited by Shigeru Miyagawa and Mamoru Saito
THE OXFORD HANDBOOK OF LABORATORY PHONOLOGY Edited by Abigail C. Cohn, Cécile Fougeron, and Marie Hoffman
THE OXFORD HANDBOOK OF LANGUAGE AND LAW Edited by Peter Tiersma and Lawrence M. Solan
THE OXFORD HANDBOOK OF LANGUAGE AND SOCIETY Edited by Ofelia García, Nelson Flores, and Massimiliano Spotti
THE OXFORD HANDBOOK OF LANGUAGE ATTRITION Edited by Monika S. Schmid and Barbara Köpke
THE OXFORD HANDBOOK OF LANGUAGE CONTACT Edited by Anthony P. Grant
THE OXFORD HANDBOOK OF LANGUAGE EVOLUTION Edited by Maggie Tallerman and Kathleen Gibson
THE OXFORD HANDBOOK OF LANGUAGE POLICY AND PLANNING Edited by James W. Tollefson and Miguel Pérez-Milans
THE OXFORD HANDBOOK OF LEXICOGRAPHY Edited by Philip Durkin
THE OXFORD HANDBOOK OF LINGUISTIC ANALYSIS Second edition Edited by Bernd Heine and Heiko Narrog
OUP CORRECTED PROOF – FINAL, 2/3/2020, SPi
THE OXFORD HANDBOOK OF LINGUISTIC FIELDWORK Edited by Nicholas Thieberger
THE OXFORD HANDBOOK OF LINGUISTIC INTERFACES Edited by Gillian Ramchand and Charles Reiss
THE OXFORD HANDBOOK OF LINGUISTIC MINIMALISM Edited by Cedric Boeckx
THE OXFORD HANDBOOK OF LINGUISTIC TYPOLOGY Edited by Jae Jung Song
THE OXFORD HANDBOOK OF LYING Edited by Jörg Meibauer
THE OXFORD HANDBOOK OF MODALITY AND MOOD Edited by Jan Nuyts and Johan van der Auwera
THE OXFORD HANDBOOK OF MORPHOLOGICAL THEORY Edited by Jenny Audring and Francesca Masini
THE OXFORD HANDBOOK OF NAMES AND NAMING Edited by Carole Hough
THE OXFORD HANDBOOK OF PERSIAN LINGUISTICS Edited by Anousha Sedighi and Pouneh Shabani-Jadidi
THE OXFORD HANDBOOK OF POLYSYNTHESIS Edited by Michael Fortescue, Marianne Mithun, and Nicholas Evans
THE OXFORD HANDBOOK OF PRAGMATICS Edited by Yan Huang
THE OXFORD HANDBOOK OF REFERENCE Edited by Jeanette Gundel and Barbara Abbott
THE OXFORD HANDBOOK OF SOCIOLINGUISTICS Second Edition Edited by Robert Bayley, Richard Cameron, and Ceil Lucas
THE OXFORD HANDBOOK OF TABOO WORDS AND LANGUAGE Edited by Keith Allan
THE OXFORD HANDBOOK OF TENSE AND ASPECT Edited by Robert I. Binnick
THE OXFORD HANDBOOK OF THE WORD Edited by John R. Taylor
THE OXFORD HANDBOOK OF TRANSLATION STUDIES Edited by Kirsten Malmkjaer and Kevin Windle
THE OXFORD HANDBOOK OF UNIVERSAL GRAMMAR Edited by Ian Roberts
THE OXFORD HANDBOOK OF WORLD ENGLISHES Edited by Markku Filppula, Juhani Klemola, and Devyani Sharma