Edinburgh Handbook of Evaluative Morphology 9780748681754

Reviews and debates the latest theoretical approaches to evaluative morphology GBS_insertPreviewButtonPopup(['ISBN

220 95 27MB

English Pages 752 [750] Year 2015

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Edinburgh Handbook of Evaluative Morphology
 9780748681754

Citation preview

Edinburgh Handbook of Evaluative Morphology

Edinburgh Handbook of Evaluative Morphology Edited by Nicola Grandi and Lívia Körtvélyessy

© editorial matter and organisation, Nicola Grandi and Lívia Körtvélyessy, 2015 © the chapters their several authors, 2015 Edinburgh University Press Ltd The Tun Holyrood Road 12(2f) Jackson’s Entry Edinburgh EH8 8PJ www.euppublishing.com Typeset in 10/12 Times New Roman by Servis Filmsetting Ltd, Stockport, Cheshire, and printed and bound in Great Britain by CPI Group (UK) Ltd, Croydon CR0 4YY A CIP record for this book is available from the British Library ISBN 978 0 7486 8174 7 (hardback) ISBN 978 0 7486 7175 4 (webready PDF) ISBN 978 0 7486 8177 8 (epub) The right of Nicola Grandi and Lívia Körtvélyessy to be identified as Editor of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988, and the Copyright and Related Rights Regulations 2003 (SI No. 2498).

Contents

contents

Illustrationsix Tablesxi Abbreviationsxiii Contributorsxxi PART I   1   2   3   4   5   6   7   8   9 10 11 12 13

Introduction: why evaluative morphology? Nicola Grandi and Lívia Körtvélyessy The semantics of evaluative morphology Victor M. Prieto Evaluative morphology and pragmatics Lavinia Merlini Barbaresi Word-­formation processes in evaluative morphology Pavol Štekauer Evaluative morphology and language universals Lívia Körtvélyessy The place of evaluation within morphology Nicola Grandi Evaluative morphology and number/gender Nicola Grandi Evaluative morphology and aspect/actionality Lucia M. Tovena Evaluative morphology and sociolinguistic variation Livio Gaeta Evaluative morphology and language acquisition Wolfgang U. Dressler and Katharina Korecky-­Kröll Evaluative morphology in a diachronic perspective Katrin Mutz Evaluative morphology in sign languages Giulia Petitta, Alessio Di Renzo and Isabella Chiari Evaluative morphology in pidgins and creoles Barbara Turchetta

3 21 32 43 61 74 91 108 121 134 142 155 170

PART II 14 Evaluative morphology from a cross-­linguistic perspective: Introduction to the descriptive chapters Lívia Körtvélyessy

187

vi 15 Eurasia 15.1 Basque Xabier Artiagoitia 15.2 Catalan Elisenda Bernal 15.3 Georgian Manana Topadze Gäumann 15.4 Hungarian Ferenc Kiefer and Boglárka Németh 15.5 Israeli Hebrew Noam Faust 15.6 Ket Edward J. Vajda 15.7 Latvian Andra Kalnača 15.8 Luxembourgish Peter Gilles 15.9 Modern Greek Dimitra Melissaropoulou 15.10 Nivkh Ekaterina Gruzdeva 15.11 Persian Negar Davari Ardakani and Mahdiye Arvin 15.12 Slovak Renáta Gregová 15.13 Swedish Arne Olofsson 15.14 Tatar Fatma Şahan Güney 15.15 Telugu Pingali Sailaja 15.16 Udihe Maria Tolskaya 16 South-­East Asia and Oceania 16.1 Apma Cindy Schneider 16.2 Chinese Giorgio Francesco Arcodia 16.3 Lisu David Bradley 16.4 Muna René van den Berg 16.5 Tagalog Carl Rubino 16.6 Tibetan Camille Simon and Nathan W. Hill 16.7 Yami D. Victoria Rau and Hui-­Huan Ann Chang

Contents

195 205 216 227 238 246 253 262 269 278 287 296 306 310 321 333 341 352 361 367 375 381 389

Contents vii 17 Australia and New Guinea 17.1 Dalabon Maïa Ponsonnet and Nicholas Evans 17.2 Iatmul Gerd Jendraschek 17.3 Jingulu Rob Pensalfini 17.4 Kaurna Rob Amery 17.5 Rembarrnga Adam Saulwick 17.6 Warlpiri Margit Bowler 17.7 Yukulta and its relatives Kayardild and Lardil Erich Round 18 Africa 18.1 Berber Nicola Grandi 18.2 Classical and Moroccan Arabic Nora Arbaoui 18.3 Ewe Yvonne Agbetsoamedo and Paul Kofi Agbedor 18.4 Kɔnni Michael Cahill 18.5 Sɛlɛɛ Yvonne Agbetsoamedo and Francesca Di Garbo 18.6 Shona Rose-­Marie Déchaine, Raphaël Girard, Calisto Mudzingwa and Martina Wiltschko 18.7 Somali Nicola Lampitelli 18.8 Zulu Andrew van der Spuy and Lwazi Mjiyako 19 North America 19.1 Cabécar Guillermo González Campos 19.2 Choctaw Marcia Haag 19.3 Dena’ina Olga Lovick 19.4 Huautla Mazatec Jean Léo Léonard 19.5 Huave Maurizio Gnerre 19.6 Inuktitut Richard Compton 19.7 Plains Cree Arok Wolvengrey

401 408 416 423 430 438 448 453 461 472 480 487 496 507 515 523 529 536 542 550 559 568

viii 19.8 Slavey (Dene) and other Athabaskan languages Olga Lovick and Keren Rice 20 South America 20.1 Jaqaru Olga Birioukova and M. J. Hardman 20.2 Kwaza  Hein van der Voort 20.3 Lule Raoul Zamponi and Willem J. de Reuse 20.4 Toba Paola Cúneo 20.5 Wichí Verónica Nercesian 20.6 Yurakaré Rik van Gijn

Contents 579 595 606 616 625 634 643

References651 Language index 701 Subject index 706 Name index 716

Illustrations

Figure 2.1 Radial model of diminutives (Jurafsky 1996) 25 Figure 2.2 Radial model of diminutives (Prieto 2005) 27 Figure 2.3 Radial model of augmentatives (Prieto 2005) 28 Figure 11.1 Radial category of the diminutive (Jurafsky 1996, 542, figure 5) 145 Figure 11.2 Revised radial category of the diminutive (Jurafsky 1996, 542, figure 5) 149 Figure 11.3 Radial category of the augmentative 151 Figure 11.4 Radial category of the pejorative 152 Figure 12.1 LIS: ‘house’ + ‘big’ (Romeo 1991) 158 Figure 12.2 LIS: ‘fish’ + ‘big’ (Romeo 1991, modified) 159 Figure 12.3 LIS: (a) ‘tie’ (canonical form), (b) ‘big tie’, (c) ‘small tie’ 160 Figure 12.4 LIS: gradual modification in size of different handling patterns (a–d)160 Figure 12.5 LIS: (a) ‘street’ (canonical form), (b) ‘large street’, (c) ‘narrow street’161 Figure 12.6 LIS: (a) ‘project’ (canonical form), (b) ‘big project’ 161 Figure 12.7 BSL: (a) ‘angry’ (canonical form), (b) ‘very angry’ (Kyle and Woll 1985, 152) 162 Figure 12.8 LIS: (a) ‘champion’ (canonical form), (b) ‘big champion’, (c) ‘outstanding champion’ 162 Figure 12.9 LIS: different facial expressions with evaluative function: (a) ‘small’, (b) ‘big’ 163 Figure 12.10 LIS: (a) ‘big thing’, (b) ‘small thing’ 164 Figure 12.11 LIS: (a) ‘table’ (canonical form), (b) ‘big table’ 166 Figure 12.12 LIS: (a) ‘friend’ (canonical form), (b) ‘good friend’ 166 Figure 12.13 Auslan: (a) ‘angry’ (canonical form), (b) doubling, (c) manual stress, (d) doubling and manual stress (Johnston and Schembri 1999, 161) 167 Figure 12.14 LIS: (a) ‘true’ (canonical form), (b) ‘very true’ (simultaneous manual evaluation), (c) ‘very very true’ (doubling) 167 Figure 16.1.1 Apma’s evaluative morphology mapped onto the typology of semantic functions 350 Figure 18.5.1 Noun class pairs in Sɛlɛɛ 488 Figure 18.6.1 Lexicalisation of size and affect in Shona 498 Figure 18.6.2 Polysemy in Shona 499 Figure 18.6.3 Descriptive and evaluative functions in Shona 505 Figure 19.5.1 Morphological strategies and values expressed in Huave 551 Figure 20.6.1 System of evaluative morphology in Yurakaré 650

x

Illustrations

Map 14.1 Geographical distribution of the sample languages (Dryer and Haspelmath 2011) Map 16.1.1 Vanuatu (www.worldofmaps.net, reproduced under the Creative Commons Attribution-­Share Alike 3.0 licence) Map 16.1.2 Languages of Pentecost Island (Gray 2012, 39) Map 20.1.1 Area in Peru where Jaqaru is spoken (Google Maps)

190 341 342 595

Tables

Table 5.1 Language universals in EM from the Universals Archive 64 Table 5.2 Summary of EM universals from the Universals Archive 66 Table 14.1 The desired proportion of languages vs. the actual proportion 189 Table 15.2.1 Augmentative suffixes in Catalan 211 Table 15.2.2 Diminutive suffixes in Catalan 212 Table 15.2.3 Verbal evaluative suffixes in Catalan 213 Table 15.4.1 Excessive forms in the Hungarian National Corpus 237 Table 15.16.1 Manchu-­Tungusic evaluative suffixes and their functions 339 Table 16.1.1 Noun root reduplication in Apma 344 Table 16.1.2 Verb root reduplication in Apma 345 Table 16.1.3 Adjective root reduplication in Apma – GOOD (intensification) 345 Table 16.1.4 Bound adjective root reduplication in Apma – GOOD (intensification) 346 Table 16.3.1 Occurrence of diminutives, augmentatives and related forms in Lisu365 Table 16.4.1 Diminution in Muna 368 Table 16.4.2 Approximation with mba-­in Muna 369 Table 16.4.3 Approximation with ka-­in Muna 369 Table 16.4.4 Approximation with ka-­on taste words in Muna 370 Table 16.4.5 Minimal degree with sa-­. . . -­ha-­no in Muna 371 Table 16.4.6 Attenuation with poka-­in Muna 372 Table 16.5.1 Tagalog diminutive reduplication 377 Table 16.5.2 Tagalog reduplication of statives 377 Table 16.5.3 Tagalog attenuative reduplication with statives 377 Table 16.5.4 Tagalog attenuative reduplication with dynamic verbs 378 Table 16.5.5 The Tagalog prefix nápaka-­ 378 Table 16.5.6 Tagalog slang affixation 379 Table 16.5.7 The Tausūg prefix Caw-­ 380 Table 16.5.8 The Ibaloi prefix Ce-­ 380 Table 16.7.1 Yami evaluative morphology 399 Table 17.5.1 Morpheme slots in the verbal word in Rembarrnga 431 Table 17.5.2 Evaluative forms and semantics with (w)urlah in Rembarrnga 436 Table 17.6.1 Warlpiri skin subsection names and their diminutive forms (MB:02–11; Hoogenraad and Laughren 2012) 445 Table 18.5.1 Noun class markers and agreement targets in Sɛlɛɛ 488 Table 18.5.2 Semantics of the Sɛlɛɛ noun classes 489 Table 18.6.1 Evaluative prefixes in Shona 497 Table 18.6.2 Inflection/derivation distinction in Shona 505 Table 19.3.1 Template of the Dena’ina verb (following Tenenbaum 1978, 34) 536

xii

Tables

Table 19.4.1 Adjectival chains for evaluative morphology in Huautla Mazatec547 Table 19.7.1 Basic and diminutive vai conjunct paradigms in Cree 571 Table 20.2.1 Expression of evaluative functions in Kwaza, with reference to (examples) or [sections] 615 Table 20.4.1 Evaluative meanings in Toba 632 Table 20.4.2 Base types for evaluative morphology in Toba 632

Abbreviations

Abbreviations

1 2 3

first person second person third person

a

agent-­like argument of canonical transitive verb

abl ablative

abs absolutive

absn absentative ACBC Autonomous Community of the Basque Country acc accusative act.nom action nominal AdaSL Adamorobe Sign Language (Ghana) add additional adj adjective adjn (where n is a number) adjective prefix of class n adjvl adjectival adpart addressive particle ADS adult-­directed speech adv adverb(ial) advls adverbialiser af agent focus aff affirmative ag agent/agentive agr agreement ai animate intransitive all allative aml ameliorative an animate ant anteriority aor aorist aplt appellative app appreciation appl applicative appr apprehensive approx approximative aprl apparential art article as annexed state ASL American Sign Language

Abbreviations

xiv asp aspect

atn attention

ATR

Advanced Tongue Root

att attenuation/attenuative attr attributive

aug augmentative

augn augmented number Auslan Australian Sign Language aux auxiliary ben benefactive

Blg. Bulgarian blm blaming BSL British Sign Language BT baby talk caus causative CDS child-­directed speech CdS Corriere della Sera (Italian newspaper) cel celerative cf circumfix Ch. Chinese char characteristic cl noun class/noun class marker clf classifier cmp companion(ship) cnj conjugation cnt continuative cntv conative co compact object coll collective com comitative comm comment marker comp complementiser compar comparative compl completive con contempt cond conditional conj conjunction conn connective particle Cons consonant consec consecutive cont continuous coord coordinated cop copula CP complementiser phrase crdl cardinal Croat. Croatian CS child speech

Abbreviations xv cso cosubordinative cvb converb dat dative dav

deadjectival verbal

deadj deadjectival decl declarative def

definite deictic preposition del delimiter dem demonstrative demn (where n is a number) demonstrative of class n dep dependent verb form der derogatory der.sfx derivate suffix desid desiderative det determiner DGS German Sign Language (Deutsche Gebärdensprache) dim diminutive dim.cl diminutive class dir directional dist distal distpst distant past distr distributive dlm delimitive DM derivational morphology dna denominal adjectival dnmls denominaliser dp distant past dp determiner phrase drth direct theme (in direct–inverse system) ds different subject dss diminutive sound symbolism du dual dur durative dva deverbal adjectival deiprep

ech

echo formation

ego egophoric

ei EM

evaluative infix/interfix evaluative morphology emph emphatic end endearment Engl. English ent entity ep evaluative prefix epi epithet epist epistemic erg ergative

Abbreviations

xvi es

evaluative suffix evaluative affix evid evidential evol evolutive ex existential marker exc excessive excl exclusive eval

f,

fem.

feminine factual ff father’s father fig. figure fill filler (bound nouns) foc focus freq frequentative frust frustrative fs free state fut future fact

gen genitive gen.obj

general object

ger gerund

gnd gender Grmn. German h

higher (on the animacy scale)

hab habitual

hdg hedge HIS highly iconic structures hon honorific hor horizontal Hung. Hungarian ic

incorporation closer

idnt identity

IE Indo-­European IM inflectional morphology imp imperative impers impersonal inan inanimate-­class incep.r inceptive reduplication incess incessative inch inchoative incl inclusive incompl incompletive incp inceptive ind indicative indf indefinite ind.obj indirect object indp independent pronoun

Abbreviations xvii inf

infinitive inflection(al) ins instrumental int intensification intent intentional interj interjection intf interfix intg interrogative intr intransitive intrpt interruption ints intensifier ip instrumental prefix ipfv imperfective irr irrealis ISL Israeli Sign Language Ital. Italian iter iterative iter.r iterative reduplication infl

jus jussive kin kinship

Lat. Latin locative focus lightv light verb lim limitative link linker LIS Italian Sign Language Lith. Lithuanian loc locative log logophoric pronoun lf

m,

masc. masculine

mal malefactive man manner mark marker

mea measure M.Gr. Modern Greek mid middle voice min minimiser mlv motion light verb mn minimal number mod modifier n noun nn (where n is a number) noun of class n na animate noun neg negation, negative neg1/neg2 bipartite negator ni inanimate noun

Abbreviations

xviii nml nominal

nmls nominaliser/nominalisation nmls.sfx

nominal suffix n is a number)noun prefix of class n nom nominative NP noun phrase npst non-­past nsbj non-­subject nsit new situation ntr neuter num numeral nn (where

obj object

objn (where

obl oblique obv

n is a number) object agreement marker of class n

obviative or fourth person

opt optative

p patient-­like argument of canonical transitive verb p.c. personal communication par particle part partitive paus pausal clitic pcc personally collected corpus pej pejorative perm permanent pf patient focus pfm phrase-­final marker pfv perfective PIE Proto-­Indo-­European pk person known pl plural pla pluractional plur plurality pn personal name pol polite Port. Portuguese pos positive poss possessive pot potential aspect marker pr preposition pred predicative prf perfect prfx prefix priv privative pro pronoun proc procrastinative pro.d pro drop prog progressive

Abbreviations xix proh prohibitive prom promotion

prox proximal/proximate prs present prsp

prospective modality

prv preverb

pssor possessor pst past

pstv postverb

ptcp participle

purp purposive q

question particle/marker qualifier quot quotative qual

r realis

rdc reduction

rdp reduplication rec receptive

recp reciprocal red

reduplicated form

refl

reflexive

ref referential

reg resignation rel relative rem remote

rep reportative res resultative resp respectful rr

reflexive/reciprocal Russ. Russian

s single argument of canonical intransitive verb SAE Standard Average European sap speech act participant, first and second persons undifferentiated sbj subject sbjn (where n is a number) subject agreement marker of class n sbjv subjunctive seq sequential sfx suffix sg singular sim simultaneity SL sign language Slov. Slovene sm subject marker sngl singulative Span. Spanish sr subordinator

Abbreviations

xx ss

same subject

st stative stm stem

sub subordinate sup superlative sv

swr

stative verb switch reference

TAMM tense, aspect, mood and associated motion test testimonial thm thematic transitive verb final or theme tloc temporal locative TMA tense mood aspect tns tense top topic tr transitive ua

unit augmented number

v verb vaf

verbal affix animate intransitive verb vaif animate intransitive verb ‘final’ stem-­forming morpheme vc voluntary comitative vce voice ven veneration vii inanimate intransitive verb viif inanimate intransitive verb ‘final’ stem-­forming morpheme VL vocal language vlex lexical vowel vn verbal noun voc vocative VP verb phrase vr verbal root vta transitive animate verb vti transitive inanimate verb Vwl vowel vai

WF wh

word formation question word

Contributors

Contributors

Paul Kofi Agbedor is a senior lecturer in linguistics at the University of Ghana, Legon, and also of English at Central University College, Accra, Ghana. His research ­interests include sociolinguistics, with special interest in language planning and language policy issues; aspects of the morphology and syntax of Ewe; and Ewe sociolinguistics. Yvonne Agbetsoamedo is a lecturer in linguistics at the University of Ghana, Legon, Ghana. Her PhD dissertation, defended at Stockholm University in October 2014, concerns the description of Sɛlɛɛ, a Kwa language of the Niger-Congo family. Her research interests include typology, documentary linguistics, lexicography, syntax, morphology and phonology. Rob Amery is head of linguistics at the University of Adelaide, South Australia. He completed a PhD in 1998 on the Kaurna language and its reclamation, resulting in the publication Warrabarna Kaurna! Reclaiming an Australian Language (2000). He continues to research Kaurna and neighbouring languages in association with a sustained language revival movement Nora Arbaoui is professor of linguistics at Moulay Ismail University, Meknes, Morocco. Her research has focused on word-­ formation in the syntax–phonology interface. She is the author of the dissertation Les dix formes de l’arabe classique à l’interface syntaxe–phonologie (2010). Giorgio Francesco Arcodia is assistant professor of linguistics at the University of Milan– Bicocca, Italy. His research interests include Chinese and Japanese linguistics, language typology, theoretical morphology and grammaticalisation theory. His monograph Lexical Derivation in Mandarin Chinese was published in 2012. Xabier Artiagoitia is professor of Basque linguistics at the University of the Basque Country (UPV/EHU), Vitoria, Spain. His research interests are syntactic theory (ergativity, structure of the noun phrase, the lexicon–syntax interface) and morphology (the study of derivational morphology). Mahdiye Arvin is a researcher in general linguistics at Shahid Beheshti University, Tehran, Iran. Her field of interest is morphology, especially word-­formation. She is now working on Persian morphology and the historiography of Persian linguistics. Elisenda Bernal is a tenure-­ track lecturer at the Universitat Pompeu Fabra, Barcelona, Spain. Her research focuses on word-­formation and neology in Catalan, often from a cross-­linguistic perspective, and lexicography, with particular interest in the representation of affixes and complex words in dictionaries. Olga Birioukova is a PhD candidate in the Department of Linguistics at the University of Florida, USA. Her primary research interests include syntactic and morphological theory, case and agreement, and Jaqi languages. Her other interests include Germanic languages and contact linguistics.

xxii

Contributors

Margit Bowler is a graduate student in linguistics at the University of California, Los Angeles, USA. Her work focuses on the semantics and syntax of central Australian languages. David Bradley is professor of linguistics at La Trobe University, Australia. His research is on sociolinguistics, historical linguistics and linguistics of Tibeto-­Burman languages. He is author or editor of twenty-­four books, including Language Endangerment and Grammar of Lisu (both forthcoming), as well as five language atlases and over four hundred journal articles, chapters and reviews. Michael Cahill is a senior linguistics consultant with SIL International, USA. He worked in the Kɔnni language literacy and translation project in Ghana for several years, resulting in the volume Aspects of the Morphology and Phonology of Konni (2008). His interests include African languages and phonology, particularly the Gur family, multiply-­articulated stops, and tone systems. Hui-­Huan Ann Chang is a doctoral student in the Institute of Linguistics at the National Chung Cheng University, Taiwan. Her research has focused on Yami dictionary making and ontology building. She has helped establish the Yami online dictionary and participated in several Yami language documentation and conservation projects. Isabella Chiari is assistant professor of linguistics at the Sapienza University of Rome and research associate at the CNR-­Institute of Cognitive Sciences and Technologies, Rome, Italy. Her interests range from scientific, methodological and theoretical issues in quantitative and corpus linguistics to computational lexicography and sign language representation and its theoretical implications. Richard Compton is an Assistant Professor at the Université de Québec à Montréal, Canada. His research examines polysynthetic word-­formation, lexical categories, modification, and consonant gradation in Inuktitut. His publications include an article in Lingua (with Christine Pittman) arguing that phonological words in Inuit correspond to DP and CP phases in the syntax. Paola Cúneo is assistant researcher at the National Scientific and Technical Research Council (CONICET), Argentina. She is also a teaching assistant in linguistics at the University of Buenos Aires, Argentina. Her research has focused on a descriptive and typological approach to word-­formation and nominal classification in the Toba language (Guaycuruan, Gran Chaco, Argentina). Negar Davari Ardakani is an associate professor of general linguistics at Shahid Beheshti University, Tehran, Iran. Her research has focused on sociolinguistics: language planning, linguistic attitudes and linguistic purism. One of her most important publications is A Century of Planning for Persian (2008). Willem J. de Reuse is adjunct research professor at the University of North Texas, USA. He specialises in the description of Native American languages, and his publications include A Practical Grammar of the San Carlos Apache Language (2006), and Handbook of Descriptive Linguistic Fieldwork, with Shobhana Chelliah (2011). Rose-­Marie Déchaine is a faculty member in the Department of Linguistics at the University of British Columbia, Canada. She received her PhD from the University of Massachusetts, Amherst, USA. Drawing on data from the Algonquian and Niger-­Congo

Contributors xxiii languages, she investigates the formal typology of how syntax interfaces with phonology, morphology and semantics. Francesca Di Garbo is a postdoctoral fellow at the University of Helsinki, Finland. Her PhD dissertation, defended at Stockholm University in September 2014, is a typological study of interactions between grammatical gender and number, and grammatical gender and evaluative morphology, in the languages of Africa. Alessio Di Renzo is a technical assistant at the CNR-­Institute of Cognitive Sciences and Technologies (laboratory LaCAM and SLDS), Rome, Italy. His research interests lie in sign language linguistics, in particular in Italian Sign Language, sign ­language ­representation, and sign language evaluation for deaf/hearing children and adults. Wolfgang U. Dressler is emeritus professor of linguistics and leader of the Comparative Psycholinguistics working group at the University of Vienna, Austria, and deputy head of the Institute of Corpus Linguistics and Text Technology, Austrian Academy of Sciences. His ongoing research focuses on the theory of Natural Morphology and its applications, and on language acquisition. Nicholas Evans is professor of linguistics at the Australian National University. His research focuses on Australian and Papuan languages, typology and anthropological linguistics. His most important works are grammars of Kayardild (1995) and Bininj Gun-­wok (2010), Dying Words: Endangered Languages and What They Have to Tell Us (2010) and (with Stephen Levinson) The Myth of Language Universals (2009). Noam Faust is a teaching fellow at the Hebrew University of Jerusalem, Israel. His research focuses on the phonology, morphology and morphosyntax of Semitic languages from a universal perspective. He has published on nominal templatic phenomena in both Hebrew and Tigre, as well as on the morphosyntax of the construct state. Livio Gaeta is associate professor of German linguistics at the University of Turin, Italy. His research has focused on word-­formation and meaning, especially with regard to action nouns, the productivity of inflectional and word-­formation processes, boundary phenomena between word-­formation and syntax, cross-­linguistic research into word-­formation, grammaticalisation and language change. Rik van Gijn is a post-­doctoral researcher affiliated to the University of Zurich, Switzerland. His main research interests are the morphosyntax of South American languages, in particular the structure of complex sentences, reference tracking and morphological structure. He has done extensive fieldwork on the Bolivian isolate language Yurakaré. Peter Gilles is professor of linguistics at the Laboratory of Luxembourgish Linguistics and Literatures, University of Luxembourg. His research interests cover phonetics/phonology (especially prosody), dialectology, sociolinguistics, language in the new media and corpus linguistics, with a strong focus on the Luxembourgish language. Raphaël Girard is a PhD student at the University of British Columbia, Canada. His research focuses on the morphophonology of nasality and voicing. He is currently working on his dissertation on nasality in Chiquitano (isolate, Bolivia). Maurizio Gnerre is full professor of linguistics at the University of Naples ‘L’Orientale’, Italy. He conducts researches on Amerindian languages. His most important publications

xxiv

Contributors

include Perfil de una lengua amazonica: el Shuar (2009) and (with F. Cuturi) Aspectos léxico-­ gramaticales, estilísticos y pragmáticos del ombeayiüts (Huave de San Mateo) (forthcoming). Guillermo González Campos is associate professor and researcher in language and communication at the Atlantic Branch of the University of Costa Rica. His main research interests focus on the current sociolinguistic situation and the morphosyntactic analysis of the Cabécar language. Nicola Grandi is associate professor of linguistics at the Department of Classic Philology and Italian Studies of the University of Bologna, Italy. He received his PhD in linguistics from the University of Pavia, Italy, in 2001, with a dissertation on evaluative constructions in Mediterranean languages. His research interests cover the fields of evaluative ­morphology, word-­formation, language contact and linguistic typology. Renáta Gregová is assistant professor at the Department of British and American Studies, Pavol Jozef Šafárik University, Košice, Slovakia. Her research activities have concentrated on the morphophonological aspects of word-­formation and the structure of the syllable in linear and non-­linear phonology. Ekaterina Gruzdeva is adjunct professor and senior lecturer in general linguistics at the Department of Modern Languages, University of Helsinki, Finland. Her research interests include Paleosiberian languages, language typology, morphosyntax, sociolinguistics, language endangerment and documentation. Her most important publication is Nivkh (1998). Marcia Haag is associate professor of linguistics at the University of Oklahoma, USA. She does both theoretical and applied research in native American languages, especially Choctaw and Cherokee. She is interested in the properties of lexical roots, arbitrary complexity, primary language description, language preservation and translation of native literature. M. J. Hardman obtained her PhD from Stanford University, USA. She is an anthropological linguist at the University of Florida, USA, now emeritus. Her current interests include language and cultures, field methods, Jaqi languages, languages and gender, and language and violence. Her latest publication is Making the Invisible Visible: Gender in Language (2013). Nathan W. Hill is lecturer in Tibetan and linguistics at the School of Oriental and African Studies, University of London, England. His research focuses on Old Tibetan and Tibeto-­Burman/Sino-­Tibetan historical linguistics. His publications include A Lexicon of Tibetan Verb Stems as Reported by the Grammatical Tradition (2010) and Old Tibetan Inscriptions, co-­authored with Kazushi Iwao (2009). Gerd Jendraschek specialises in language description and linguistic typology, having published on various aspects of Turkish, Basque and the Papuan language Iatmul. He was assistant professor at the University of Regensburg, Germany, until 2012 and is currently taking part in Korean language training at Sogang University, Seoul, South Korea. Andra Kalnača is professor of Latvian linguistics at the University of Latvia, Riga, Latvia. Her research has focused on Latvian and general morphology, morphophonology, morphemics and morphosyntax. Her most important publications are the monographs Morfoloģijas stilistika (Stylistics of Morphology, 2011) and Morfēmika un morfonoloģija (Morphemics and Morphonology, 2004).

Contributors xxv Ferenc Kiefer is professor emeritus at Budapest University (ELTE), Hungary, and research professor at the Research Institute for Linguistics at the Hungarian Academy of Sciences. His research fields include morphology, mainly word-­formation, lexical semantics and the semantics–pragmatics interface. Katharina Korecky-­Kröll is a post-­doctoral assistant in the Department of Linguistics at the University of Vienna, Austria. Her main research interests are early first and second language acquisition and morphological processing in older children and adults. Lívia Körtvélyessy is associate professor at Pavol Jozef Šafárik University, Košice, Slovakia. She received her PhD in linguistics from the Slovak Academy of Sciences in 2008, and in 2013 she was awarded the habilitation degree from ELTE Budapest, Hungary. Her fields of expertise include evaluative morphology, word-­formation and linguistic typology. She is the author of a monograph (published in Slovak) on the influence of sociolinguistic factors on word-­formation and a co-­author (with Pavol Štekauer and Salvador Valera) of Word-­Formation in the World’s Languages (2012). Nicola Lampitelli is associate professor of Italian linguistics at the Université François Rabelais de Tours, France. His work focuses on the morphophonology of exponence in Romance (Italian and French) and Afroasiatic (Somali and Arabic). Jean Léo Léonard is professor of general linguistics and typology at the Paris-Sorbonne University (Paris 4) – STIH: Sens, Texte, Informatique, Histoire – (EA 4509). He has been a senior member of the Insitut Universitaire de France from 2009 to 2014 (see http://jll. smallcodes.com/home.page). His research has focused on linguistic documentation of endangered languages in France, Russia and Middle American, and on dialectology as a means to achieve fine-grained typology. Olga Lovick is an associate professor of linguistics and Dene language studies at the First Nations University of Canada in Regina, Saskatchewan. Her research focuses on the documentation of Alaskan Athabascan languages, in particular Dena’ina and Upper Tanana. Dimitra Melissaropoulou is an adjunct lecturer at the University of Patras, Greece. Her publications address issues on various aspects of morphological theory, such as evaluative morphology, the notion of head, derivational affix ordering, the synchrony and diachrony of present perfect structures, language simplification, and morphology in language contact situations. Lavinia Merlini Barbaresi was full professor (now retired) of English linguistics at the University of Pisa, Italy. Her scientific interests are in text linguistics (phenomena of complexity), morphology (description of phenomena in English and Italian), pragmatics (at all levels of the language) and semiotics (phenomena of markedness). Lwazi Mjiyako is a tutor in African languages at the University of the Witwatersrand, Johannesburg, South Africa. His research interests are African language media studies and Zulu second language teaching. He is currently working on his PhD dissertation, Making the News in isiZulu: Language Use and Representation in isiZulu News. Calisto Mudzingwa holds a PhD in linguistics from the University of South Africa, Department of African Languages. His research has focused on Bantu phonology, morphophonology and morphosyntax.

xxvi

Contributors

Katrin Mutz is senior lecturer in French and Italian linguistics at the University of Bremen, Germany. Her main topics of research are word-­formation, lexical semantics and voice, as well as language change (especially grammaticalisation) and language contact. In 2000 she wrote the monograph Die italienischen Modifikationssuffixe: Synchronie und Diachronie. Boglárka Németh is assistant lecturer in the Department of Hungarian and General Linguistics, Babeș-­Bolyai University, Cluj-­Napoca, Romania. Her research has focused on general theoretical and Hungarian-­language-­specific questions related to aspect, temporality and event structure, with special emphasis on the category of stativity. Verónica Nercesian, PhD, is a researcher at the National Scientific and Technical Research Council (CONICET) and the Linguistic Research Institute, National University of Formosa, Argentina. She teaches linguistics and lexical theory at the University of Buenos Aires, Argentina. Her current interests include Wichí dialectal variation and verbal art, and the interplay of linguistic levels. Arne Olofsson was professor of English language and linguistics at the University of Gothenburg, Sweden, until his retirement in 2010. His research focuses on syntactic variation (including grammaticalisation), word-­formation (especially the influence of Latin and Greek), lexicography (past and present) and assessment of undergraduates’ English proficiency (including test construction). Rob Pensalfini is a senior lecturer in linguistics and drama at the University of Queensland, Australia. As a linguist, his research has focused on the grammar of Australian Aboriginal languages. He has published two monographs on the Jingulu language, as well as over a dozen articles on Australian phonology, morphology, syntax and sociolinguistics, alongside several articles on the performance of Shakespeare in Australia. Giulia Petitta is a post-­doctoral researcher at the CNR-­Institute of Cognitive Sciences and Technologies, Rome, Italy. Her research interests lie in sign language linguistics, with a focus on Italian Sign Language (LIS), sign language representation, and simultaneous interpreting from LIS to spoken Italian and vice versa. Maïa Ponsonnet is a post-­ doctoral researcher at Dynamique du Langage, CNRS/ Université Lyon 2, France. Her research focuses on the linguistic emotional resources of Australian languages, in particular Dalabon (non-­Pama-­Nyungan, Gunwinyguan) and Barunga Kriol, a variety of one of the Australian creoles. Victor M. Prieto is professor of linguistics at North Greenville University, USA. His main research interest is morphopragmatics; that is, morphosyntactic issues and their effects on basic and contextual meaning. He is also interested in cognitive aspects in meaning processing. His University of Florida PhD dissertation was on semantic/pragmatic issues of Spanish evaluative morphology. D. Victoria Rau is professor in the Institute of Linguistics at the National Chung Cheng University, Taiwan. Her research interests include Austronesian linguistics, s­ ociolinguistics and applied linguistics. She has published major monographs on Yami and established three online resources for Yami language documentation. Keren Rice is university professor in the Department of Linguistics at the University of Toronto, Canada. Her research has focused primarily on the phonology and morphology of Athabaskan languages, phonological representations, morphology and fieldwork.

Contributors xxvii Erich Round received his PhD in linguistics at Yale University, USA, and now teaches at the University of Queensland, Australia. His research focuses on Australian indigenous languages, phonological and morphological theory, and advanced computational methodologies in typology and historical linguistics. His book Kayardild Morphology and Syntax appeared in 2013. Carl Rubino is a computational linguist and typologist currently working in the field of machine translation. His publications include Ilocano Dictionary and Grammar (2000), Tagalog-­English/English-­Tagalog Standard Dictionary (2002), The Encyclopedia of the World’s Major Languages: Past and Present (with Jane Garry, 2001) and Intensive Tausug (2006). His latest linguistic work involved documenting an undescribed language spoken on Leyte Island, Philippines. Fatma Şahan Güney is assistant professor of Turkic languages and linguistics at Muğla Sıtkı Koçman University, Mugla, Turkey. Her research has focused on the Tatar language, linguistics, teaching Tatar as a second language, and linguistic aspects of other Turkic languages such as Bashkir and Kazakh. Pingali Sailaja is professor in the Centre for English Language Studies, University of Hyderabad, India. Her research interests are in the areas of phonology, morphology, varieties of English, English in India, and teaching English as a second language. Her books include English Words: Structure, Formation and Literature (2004) and Indian English (2009). Adam Saulwick is a senior researcher at an Australian research organisation. His current research investigates theoretical and practical issues of high-­level information fusion, with particular focus on the nexus of computational linguistics and formal models of linguistic semantics for knowledge representation and reasoning. Cindy Schneider is a senior lecturer in linguistics at the University of New England, Armidale, Australia. She researches Austronesian and Papuan languages of the Pacific, with a focus on language description and documentation, language/dialect variation and shift, issues in literacy development, and language planning. Camille Simon is a doctoral student at the University of Paris III and the Langues et Civilisations à Traditions Orales (LACITO) laboratory, France. Her research focuses on Tibetic and Turkic morphosyntax and language contact. Pavol Štekauer is professor of English linguistics at Pavol Jozef Šafárik University, Košice, Slovakia. His research has focused on an onomasiological approach to word-­ formation, sociolinguistic aspects of word-­formation, meaning predictability of complex words, and cross-­linguistic research into word-­formation. Maria Tolskaya graduated from Moscow State University, Russia, and Harvard University, USA. She has co-­authored a descriptive grammar and text collections of Udihe (Manchu-­ Tungusic), based on extensive fieldwork. Her research includes ideophones and polarity items. Currently she participates in the electronic multimedia documentation of the endangered languages of Siberia (http://larkpie.net/siberianlanguages). Manana Topadze Gäumann is a post-­doctoral research and teaching assistant in the Department of Linguistics, University of Bern, Switzerland. Her current research interests include Caucasian languages, argument marking, morphosyntactic processes of word-­ formation, evidentiality, language variation and change.

xxviii

Contributors

Lucia M. Tovena is professor of linguistics at the University of Paris VII, France, specialising in formal semantics. She led the project that has produced both the ELICO corpus of annotated occurrences of French determiners (http://elico.linguist.univ-­paris-­diderot. fr/index.php) and a number of publications on determiners, including the volume French Determiners In and Across Time. Her research areas include negation, forms of individuation in the nominal and verbal domain, lexical aspect and pluractionality. Barbara Turchetta is professor of anthropological linguistics and sociolinguistics at the University of Tuscia, Italy. Her interests in research are on communication strategies in multilingual societies, social aspects of literacy and anthropological aspects of communication. She works as a field researcher in Africa, on both vehicular languages and Bantu languages. She is responsible for intercultural education programmes at both an Italian and an international level. Edward J. Vajda teaches Russian language and culture, linguistics, and courses on Inner Asian and Siberian peoples at Western Washington University, USA. He is involved in documenting endangered languages of the Siberian north. His research focuses on the structure and historical development of Ket and other Yeniseian languages of Central Siberia. René van den Berg is a senior linguistics consultant with SIL International, based in Papua New Guinea (PNG). His main research interests are Austronesian languages and lexicography. He has published or co-­authored grammars of Muna and Balantak (in Sulawesi) and Vitu and Lote (in PNG). Andrew van der Spuy is a senior lecturer in linguistics at the University of the Witwatersrand, Johannesburg, South Africa. His research interests are Zulu morphology and morphological theory. His publications include ‘Generating Zulu noun class morphology’ (Language Matters, 2010) and ‘The morphology of the Zulu locative’ (Transactions of the Philological Society, 2014). Hein van der Voort is associate researcher in indigenous languages at the Museu Paraense Emílio Goeldi, Belém, Brazil. He has conducted research on various undocumented languages, and on language contact in the Guaporé region, the Arctic and the Caribbean. His most important publication is A Grammar of Kwaza (2004). Martina Wiltschko is an associate professor of linguistics at the University of British Columbia, Canada. Her research has focused on grammatical categories, addressing the question of what about them is universal and what varies. She has approached this topic on the basis of a variety of unrelated languages and relative to a range of different categories. Arok Wolvengrey is professor of linguistics at First Nations University of Canada in Regina, Saskatchewan. He specialises in Cree and the Algonquian languages, concentrating on morphosyntax, discourse structure and lexicography. Notable among his ­publications are the two-­volume Cree–English dictionary nēhiýawēwin: itwēwina (2001) and his dissertation, Semantic and Pragmatic Functions in Plains Cree Syntax (2011). Raoul Zamponi is a guest researcher at the Department of Linguistics of the Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany. He has worked on little-­ known, extinct Amerindian languages and has written, inter alia, grammatical descriptions of Maipure (Arawakan) and of the isolates Betoi (Venezuela) and Waikuri (Mexico).

Part I

1 Introduction: Why Evaluative Morphology? Nicola Grandi and Lívia Körtvélyessy

1  A brief overview of the research in the field To our knowledge, the first attempt to sketch a consistent picture of so-­called evaluative morphology was made by Scalise (1984, 132–3), to whom we owe the choice of a label that has been almost unanimously accepted in the literature: affissi valutativi/evaluative affixes. It is probably not by chance that it was an Italian linguist who showed, for the first time, that seemingly different processes deserve a unique account. This is due to the well-­ known fact that Italian – like other Romance languages, such as Spanish and Portuguese, and many Slavic languages, such as Slovak – displays a rich set of evaluative affixes and other evaluative markers: these are often referred to as typical instances of these processes in the literature in the field. Scalise identifies six criteria distinguishing examples of evaluative morphology, involving both derivational and inflectional morphology:



(i) The syntactic category of the base they are attached to does not change (casa ‘house’ – casina ‘little house’). (ii) The syntactic features or the subcategorisation frame of the base (such as gender, actionality, etc.) are not changed. (iii) The semantics of the base word is changed (e.g. ragazzo ‘boy’ – ragazzino ‘little boy’; macchina ‘car’ – macchinina ‘toy car’). (iv) They allow the consecutive application of more than one rule of the same type: the result of every application is an existent word (cf. fuoco – fuocherello – ­fuocherellino ‘fire – little fire – nice little fire’). (v) They are always external with respect to other derivational suffixes and internal with respect to inflectional morphemes (e.g. contrabbandierucoli ‘small-­time smugglers’ = compound word contrabbando (‘contraband’) + derivational suffix (-­iere ag) + evaluative suffix (-­ucol(o) pej(m.sg)) + inflectional morpheme (-­i m.pl)). (vi) They can trigger specific readjustment rules, which are both transparent and idiosyncratic (for example, [tʃ] must be inserted between the base and the diminutive suffixes -­ino and -­ello if the base ends in ['one]/['ɔne]).1

According to Scalise, properties (i) and (ii) draw evaluative affixes closer to inflectional affixes; properties (iii) and (iv) render them closer to derivational affixes; whilst the last two properties seem to be peculiar to evaluative affixes. From this, he proposes the idea

Nicola Grandi and Lívia Körtvélyessy

4

of a ‘third morphology’, namely ‘evaluative morphology’ (EM): evaluative affixes would belong to an independent set of word-­formation rules placed in between derivational rules (derivational morphology or DM) and inflectional rules (inflectional morphology or IM).2 DM v EM v IM Even though Scalise’s criteria were set specifically for Italian, they drew the attention of morphologists. From then on in almost all handbooks of morphology at least a few lines have been devoted to the formation of diminutives, augmentatives, etc. And evaluative morphology has become a more and more central issue in morphological investigations. In brief, in the literature three major perspectives can be singled out. The first one concerns the formal properties of evaluative morphology, and it has long been influenced by Scalise’s view. Even if it was demonstrated (see mainly Stump 1993; Beard 1995; Bauer 1997) that the idea of this third morphology is language-­bound, and – consequently – that it cannot be viewed as a universal feature of human languages (more so as there are numerous languages that do not possess morphological diminutives, augmentatives, pejoratives and/or amelioratives), the (alleged) exceptional nature of evaluative morphology with respect to inflection and derivation has been emphasised by most of the scholars who have dealt with it. Evaluative complex words have often been placed at the periphery of morphology, taking it for granted that their behaviour is largely unpredictable and irregular. Beard (1981, 180) states that diminutive affixes are placed ‘somewhere between lexical and purely inflectional forms’; Szymanek (1988, 106–9) ascribes diminutives and augmentatives to an ‘expressive periphery’ of derivation, which does not correspond to any cognitive category; and Carstairs-­McCarthy (1992, 107) assumes that evaluative affixes pertain to ‘expressive morphology’, which is different from derivational morphology, and exceptional to principles that govern it. Moreover, diminutives, augmentatives, pejoratives and amelioratives have always been analysed as independent categories, neglecting the possible interrelations among them. In addition to this, diminutives have tacitly been assumed to be universal, and have therefore been given more prominence than other processes. To put it briefly, the debate has concentrated on the status of evaluative morphology and on its place within morphology, with the explicit aim of finding a universal collocation for diminutives, augmentatives, pejoratives and amelioratives. This issue will be dealt with by Grandi (this volume, Chapter 6), who demonstrates that the potential status of evaluative morphology heavily depends on the specific relations between inflectional and derivational morphology, which differ from one language to another and from one category to the next. Moreover, this problem cannot be tackled without considering the other two crucial perspectives on evaluative morphology, namely the semantic and the pragmatic ones. In some cases, the alleged exceptional nature of evaluative morphology changes its appearance when observed from a different point of view. According to Simon and Wiese (2011, 21), some complex constructions are often semantically underdetermined and this sometimes favours the development of seemingly exceptional patterns. Among other cases, these authors discuss the behaviour of diminutives, asserting that ‘their erratic behaviour turns out to be more systematic when viewed from a morphopragmatic perspective’ (2011, 22).

Introduction:Why Evaluative Morphology? 5 In other words, ‘the possibility of directly involving pragmatic aspects in morphology’ explains some of the seemingly unusual properties of diminutives, augmentatives, etc. What matters here, is that – as shown in most chapters of this book – evaluative affixes are less irregular than one might believe, if considered from the appropriate point of view. One of these perspectives is pragmatics, often neglected in the studies on evaluative morphology. Dressler and Merlini Barbaresi (1994, among others) focuses on this issue, considering evaluative markers as typical instances of morphopragmatics, ‘defined as the area of general pragmatic meanings of morphological rules, that is of the regular pragmatic effects produced when moving from the input to the output of a morphological rule’ (1994, 51). In brief, contrary to a widely accepted view according to which ‘pragmatics is usually seen as an antithesis to grammar’ (1994, 50), some features of evaluative morphology show that some regular morphological operations have regular pragmatic effects (1994, 54). In the opinion of Dressler and Merlini Barbaresi (1994, 55), evaluative word-­formation processes include ‘a pragmatic variable which cannot be suppressed in the description of [their] meaning’. We will deal with this issue in Section 3 below and also in Chapter 3 (by Merlini Barbaresi). The picture is further complicated by the fact that the meaning of evaluative construction usually displays a multiplicity of senses. Consider, for example, the case of the Italian suffix -­ino, which can occur in complex words expressing diminution in size (librino ‘small book’ < libro ‘book’), endearment (maritino ‘dear husband’ < marito ‘husband’), contempt (dottorino ‘untrustworthy doctor’ < dottore ‘doctor’) and singulative (zuccherino ‘sugar lump’ < zucchero ‘sugar’); it is also used in the formation of instrumental nouns (frullino ‘(small) mixer’ < frullare ‘to mix’; cf. frullatore ‘mixer’), of ethnic adjectives (tunisino ‘Tunisian’), etc. As stated by Haspelmath (2003, 211), these kinds of morphemes ‘have more abstract and general meaning and are thus more apt to be used in multiple ways than content words’. The cluster of meanings often associated with evaluative affixes has contributed to accentuate the idea of their anomalous nature. However, cross-­linguistic studies have proved that the multifunctionality of evaluative affixes is far from chaotic and irregular, both cross-­linguistically and within a single language; on the contrary, it can be organised in recurrent and regular networks of semantic (and pragmatic) relations. Jurafsky (1993; 1996) represented this network as a radial category, organised around the meaning ‘child’. The radial model can be considered a forerunner of the recent approach to multifunctionality, namely semantic maps. The main difference between the two models is that a radial category assumes a central or prototypical function and some peripheral senses whereas a semantic map does not do so. However, it must also be pointed out that, according to Haspelmath (2003, 232), these two models are compatible, since ‘the semantic-­map method is completely neutral in this respect’. Moreover, both of them aim to provide a sketch of cross-­linguistically recurrent semantic relations. In the chapters of this volume, the radial model is used to represent the multifunctionality of some evaluative classes (diminutives and augmentatives) in both a synchronic (see Prieto) and a historical perspective (see Mutz). The possible relationship between the radial approach and the semantic-­ map method is dealt with by Körtvélyessy (Chapter 14). The crucial point, here, is that the wide range of possible meanings and pragmatic values associated with most evaluative constructions are far from irregular, and contribute to a explanation of these constructions’ apparently unusual formal behaviours. To conclude, what clearly emerges is that the very nature of evaluative morphology can be explained only adopting a method which integrates the formal level with the semantic and pragmatic ones. This is precisely the method adopted in this book.

6

Nicola Grandi and Lívia Körtvélyessy

2  The state of the art Scalise’s research through which the idea of a third morphology was proposed got mixed reactions. For many years the issue concerning the place of evaluative morphology with respect to inflection and derivation has monopolised the debate, pushing aside other crucial points and the non-­affixal realisations of ‘evaluation’ (among the few exceptions: Montermini’s 1999 work on apocope, Kahane and Kahane-­Toole 1948–9 on feminine gender, Wierzbicka 1986 on reduplication, and the research on sound symbolism in evaluation; see, among others, Pentland 1975; Bauer 1996; Diffloth 1994; Gregová 2009; Gregová, Körtvélyessy and Zimmermann 2010; Körtvélyessy 2011b): a cross-­linguistic survey reveals that evaluative constructions cover a range of word-­formation strategies which is much broader than affixal morphology, including, inter alia, gender shift, compounding, reduplication, tonal variation, etc. (see also Štekauer, this volume). Stump (1993), for example, replaces Scalise’s properties of evaluative affixes with evaluative rules which, in his view, fall within the class of category-­preserving rules of derivation and compounding. By implication, their properties can be predicted and there is no need to assign the position of a third morphology to evaluative morphology rules. Beard (1995) distinguishes four types of derivational rules. One of them, labelled as expressive derivations, ‘reflects the attitude of the speaker towards the size and merit of the referent’ (1995, 175). According to Beard (1995, 163) they reflect at least five functions universally: diminutive, augmentative, pejorative, affectionate and honorific. Similarly, Bauer (2004), adopting Booij’s distinction between contextual and inherent inflection, assigns a specific position to evaluative ­morphology. This branch of research also includes, inter alia, Dressler and Merlini Barbaresi (1992), which accounts for evaluative affixes as non-­prototypical representatives of derivation, Grandi (2001b), which presents a similar approach, Nieuwenhuis (1985), Ştefănescu (1992), Fradin and Montermini (2009), Fortin (2011) and Körtvélyessy (2014). Evaluative affixes have also been studied in relation to some other morphological categories such as gender and number (see Grandi 2001a), aspect and (plur)actionality (Grandi 2008; 2009; Tovena 2010; 2011a). In addition to these theoretical studies, there are a number of publications which provide descriptions of evaluative morphology in individual languages and/or language families: Baltic languages (Ambrazas 1993), Romance languages (Hasselrot 1957; Ettinger 1974), English (Schneider 2003), Slovak (Böhmerová 2011), Spanish (Gooch 1970; Jaeggli 1980; Lázaro Mora 1999; Prieto 2005, among others), Sardinian (Grandi 2005), Italian (Rainer 1989; Napoli and Reynolds 1995; Grandi and Scalise 1999; Merlini Barbaresi 2004a, among others), Icelandic (Grönke 1955), French (Weber 1963; Hasselrot 1972; Fradin 2001, among others), Modern Greek (Mirambel 1940; Daltas 1985; Melissaropoulou 2009), German (Ott 2011), Portuguese (Rudolph 1990) and, for non-­Indo-­European languages, Swahili (Herms 1989), Akan (Appah and Amfo 2011), Hebrew (Bolozky 1994), Passamaquoddy (LeSourd 1995), Yurok (Berman 1986), Arabic (Watson 2006), Berber (Taine-­Cheikh 2002), Tibetan (Uray 1952), etc. Furthermore, there are a few studies which deal with evaluative morphology from a contrastive and areal-­typological perspective: Bakema and Geeraerts (2004), Grandi (2002), Körtvélyessy and Štekauer (2011b), Štekauer (2011), Di Garbo (2013) and, again, Körtvélyessy (2014). These works provide numerous examples from various languages which either support or rebut Scalise’s proposal. Generally speaking, a cross-­linguistic survey reveals that the main shortcoming of Scalise’s proposal is its almost exclusive link with Italian: the six

Introduction:Why Evaluative Morphology? 7 properties listed above are not attested by evaluative constructions of typologically and genealogically different languages (see, once again, Stump 1993). The idea that there is a ‘universal’ place for evaluative morphology within morphology is contradicted by data: evaluative complex words behave as derived words in some languages (e.g. Romance languages) and as inflected words in other ones (e.g. some Bantu languages). A further problem depends on the fact that hardly any of the studies on evaluative affixes provide an explicit definition of evaluative morphology: by implication, we cannot be sure that scholars are investigating the same data, or that they share the same concept of evaluative morphology. This issue will be dealt with below. To sum up, Scalise’s idea of a ‘third morphology’, besides inflection and derivation, addressed and triggered research on evaluative morphology, and, even if this idea is not confirmed by a cross-­linguistic survey of data, Scalise’s research suggested that evaluative constructions may display specific properties that distinguish them from both inflected and derived words, or, in a broader perspective, from other morphological constructions. These properties should legitimate an autonomous treatment of evaluative constructions within a theory of word formation. Almost all morphologists agree on this premise. Interestingly enough, various aspects of evaluative morphology, including its position within the system of morphology, were discussed, long before Scalise, by a representative of the onomasiological approach to word formation: the Czech linguist Dokulil (1962). In his view, evaluative morphology falls within the modificational category of word formation: modification means that the original meaning is semantically modified and that it is usually enriched. This process is not accompanied by a change of word class. Modifications can be of various nature: Dokulil (1962, 46) distinguishes four modificational onomasiological categories: (a) d iminutive onomasiological category – the concept is modified by the diminutive marker; (b)  augmentative onomasiological category – understood as a counterpart of the diminutive onomasiological category; the concept is modified by an augmentative marker. (c) derivation of female names from male names and vice versa – the concept of animate being is modified by a gender marker; (d) the young of animate beings – the concept is modified by a marker with the meaning ‘not grown-­up’. Of course, it is not only the place of evaluative morphology within the system of morphology that has been discussed. Its semantic facet comes to the forefront in Wierzbicka (1984) and especially in Jurafsky (1993; 1996), already mentioned above, which introduce a model of diminutiveness based on Lakoff’s concept of radial category, possibly with historical implications: Jurafsky concludes that the origin of the morphological diminutive goes back to the concept of ‘child’, even though cross-­linguistic research gives ample evidence that the concept of ‘child’ is by no means the only concept that can be taken into consideration. Besides the works cited above, Prieto (2005) and Fortin (2011) present a semantic-­driven analysis of evaluative morphology. As already stated in Section 1, Dressler and Merlini Barbaresi (1994, among others) provide a new perspective by pointing out that evaluative morphology is primarily located in pragmatics. In their framework, evaluative markers are typical instances of ­morphopragmatics. For a similar approach, see also Schneider (2013).

8

Nicola Grandi and Lívia Körtvélyessy

Diachronically, the origin of evaluative affixes has been investigated in both contrastive studies and individual languages. The former are represented primarily by Jurafsky’s (1993; 1996) as well as by Matisoff’s (1991) research regarding South-­East Asian languages, by Creissels’ (1999) work on some African languages and by Grandi’s study of Mediterranean languages (2002) and his wider cross-­linguistic research (2011b). As far as studies on single languages are concerned, almost all the literature deals with classical languages and with languages derived from them (this is due to the huge amount of documentation available regarding these languages, which is unparalleled in other linguistic traditions). Latin diminutives (and their Romance outcomes) are surveyed by Strodach (1933), Hakamies (1951), Butler (1971), Magni (1999), Mutz (1999; 2000) and Haverling (2011); Latin suffixes that gave rise to present-­day Romance augmentatives are analysed by Lazzeroni (1963), Gaide (1988) and Grandi (2003a). For the Greek tradition, see Petersen (1910). Last but not least, one must also consider studies regarding the acquisition of evaluative affixes (Gillis 1998; Laalo 2001; Savickienė and Dressler 2007b) and on evaluative morphology in agrammatism (Franco, Zampieri and Meneghello, 2013) which open new fields for future research.

3  Towards a definition As stated above, the idea of a ‘third morphology’, besides inflection and derivation, even if not confirmed by a cross-­linguistic survey of data, suggests that evaluative constructions display specific properties that distinguish them from both inflected and derived words, or, in a broader perspective, from other morphological constructions. Problems arise when considering the status assigned to these properties, and, consequently, to evaluative morphology. Scholars have often provided different perspectives on this issue. The inconsistencies that emerge from a survey of the literature regarding the treatment of evaluative morphology can be explained in the following manner. Firstly, as discussed above, the issue concerning the place of evaluative morphology with respect to inflection and derivation has monopolised the debate, pushing aside other crucial points and the non-­affixal realisations of ‘evaluation’. Secondly, the situation is further complicated by the fact that studies on evaluative morphology usually neglect the pragmatic dimension that, in fact, plays a crucial role in explaining the behaviour of diminutives, augmentatives, etc., as demonstrated by Dressler and Merlini Barbaresi (1994) and, more recently, by Schneider (2013).3 This state of affairs is due to the objective difficulties scholars encounter when comparing different languages from a pragmatic point of view, because of the lack of tools, corpora, databases (and even grammars), etc. that exhaustively represent all linguistic and, above all, extra-­linguistic variables involved in the pragmatic dimension of a speech act. Finally, hardly any studies on evaluative affixes provide an explicit definition of evaluative morphology: by implication, it is unclear whether scholars are investigating the same data, or whether they share the same concept of ‘evaluation’. If this is the premise, different studies on evaluative morphology and data discussed can hardly be compatible. As a consequence, in the absence of an explicit and commonly shared definition of evaluative morphology, the only criterion governing discussion and interpretation of data is scholars’ intuition: a very weak basis for a broad cross-­linguistic comparison. Therefore, the unavoidable condition for dealing with evaluative morphology from a cross-­linguistic perspective is to render explicit the criteria that will govern the selection of

Introduction:Why Evaluative Morphology? 9 data; in other words, it is necessary to provide a clear definition of evaluative morphology and of its field of inquiry. As recently stated by Haspelmath (2007; 2010), in comparative and typologically oriented surveys it is often hard to distinguish between language-­specific descriptive categories, comparative concepts, universal categories and universal functions. Haspelmath says that language comparison should be based not on structural linguistic categories (i.e. exclusively on formal parameters), but on comparative concepts. In other words, comparison should be substance-­based, since substance (unlike formal categories) is universal. The consequence is that if one focuses on morphosyntax, comparisons should be based on semantic-­functional patterns, not on formal-­structural patterns: ‘for morphosyntactic comparison to be possible, we must hold the meaning constant’ (Haspelmath 2007, 127). Categories without a semantic-­functional foundation can hardly be compared. Consistently with this premise, it is crucial to assume that different linguistic items can be considered part of the same class if they perform the same function, as this provides a cross-­linguistically valid definition of evaluative morphology. Performing the same function does not assure formal identity and is not, if considered in isolation, a sufficient condition to establish the membership of a class, despite being a necessary condition. Of course, semantic-­functional features should have formal correlates, but the very nature of these correlates falls into the set of language-­specific facts. Failure to recognise the asymmetry between these two levels has often produced weak generalisations, leading some scholars to generalise facts that, in fact, are mere tendencies in specific languages or groups of languages (cf., for example, the almost exclusive focus on the issue of the place of evaluative morphology with respect to inflection and derivation, and the consequent oblivion of non-­affixal evaluative constructions: tonal variations, compounds, ­reduplications, etc.). In order to define evaluative morphology, this approach will be followed: a ­definition of evaluative morphology will be formulated, establishing the levels in which it operates, starting from the semantic-­functional one, with the addition of the pragmatic component, which is usually neglected in typological studies. Constructions that have been labelled as evaluative in the literature can be roughly organised into the following classes, according to the function they perform (data is ­randomly selected from the chapters of the second part of the volume): • diminution in quantity and/or quality (Ewe awu-­vi ‘dress-­dim, small dress’, Slovak dom-­ček ‘house-­dim, small house’, Spanish mini-­tren ‘dim-­train, small train’, Kongo vata ‘village’ > kivata-­vata ‘small village’); • augmentation in quantity and/or quality (Berber tabhirt ‘garden.f’ > abhir ‘garden.­ aug.m’, Italian ragazz-­one ‘boy-­ aug, big boy’, French méga-­joie ‘aug-­joy, big, intense joy’); • age variation (Telugu kooti-­pilla ‘monkey-­young one.dim, young of a monkey’); • approximation/reduction/attenuation (Cabécar bätsë ́ -n­ a̱ ‘red-­att, reddish’, Catalan hor-­eta ‘hour-­dim, about an hour’, Nivkh ñeñi-­jo-­ ‘be sweet-­att, be sweetish’, Italian fischiare ‘to whistle’ > fischiettare ‘to whistle thoughtlessly, in a happy-­go-­ lucky manner’); • intensification (Italian campion-­issimo ‘champion-­ sup, the great champion’, Georgian axal(t)-­axali t’anisamos-­i ‘brand-­new clothes’); • endearment (Bulgarian brát-­če ‘brother-­dim, dear little brother’); • hypocorism (Slovak Eva > Evka);

10

Nicola Grandi and Lívia Körtvélyessy

• expression of social position (Modern Greek ypallīl(os)-­akos ‘employee-­dim, an unimportant employee’; ypallīl(os)-­ara ‘employee-­aug, an important employee’); • contempt (Italian govern-­icchio ‘government-­pej, very bad government’, Catalan cuiner-­et ‘bad cooker’, Wichí hin’u-­fwaj ‘man-­dim, little man (contempt))’; • authenticity/prototypicality (Italian burro burro ‘butter butter, real butter’, Kwaza kanwa-­tete ‘canoe-­int, real canoe’). Evaluative constructions cover a range of word-­formation strategies which is much broader than affixal morphology, including, inter alia, gender shift, compounding, reduplication, tonal variation, etc. (see also Štekauer, this volume). If the expression of diminution is considered in terms of quantity and/or quality as mentioned above, instances of suffixation, prefixation, reduplication and infixation are found. As a consequence, an attempt to define diminution consistently on the basis of formal criteria would result in language-­specific generalisations. Referring to the pragmatic/semantic functions cross-­linguistically performed by evaluative constructions as listed above, a first generalisation can be made, asserting that they can be grouped in two classes: evaluative constructions can express both descriptive (or quantitative) and qualitative evaluation. In other words, an object (or a person, an action, etc.) can be ‘evaluated’ according to both its tangible, real characteristics (its size, its shape, etc.), and the speaker’s feelings towards it. A descriptive evaluation is not influenced by personal feelings or opinions, but relies on the real and objective properties of an item. In both cases, a deviation from a standard or from a default value may be observed: Slovak domisko ‘large house’ derives from dom ‘house’, and it conveys a quantitative/­descriptive deviation from the default size of houses (see Štekauer, this volume). The standard or default value (or norm, in Schneider 2013) is culturally and/or socially determined (it is a mental representation, shared by the members of the same community). In quantitative/ descriptive evaluation there is an objective, observable, evident deviation. In qualitative evaluation the speaker perceives or feels a deviation. For example, if we compare a dachshund and a Great Dane, we can call the former cagnolino (Italian, ‘small dog’), and the latter cagnone (Italian, ‘big dog’), since they are objectively smaller and bigger than a standard dog: nobody would contradict these statements, since this ‘evaluation’ is not determined by the subjective perception of the speaker. This kind of ‘evaluation’, which is most often based on dimensional parameters, does not depend on subjective and personal judgements and is made on the basis of the influence of the extra-­linguistic context. Two speakers sharing the same conception of a standard dog will perceive a dachshund as a cagnolino, independently of the context and of their feelings. Nevertheless, someone could call his or her Great Dane cagnolino, expressing his or her affection for it. In this case, different speakers can evaluate the same object (or action, person, etc.) in different ways. This often happens when speakers do not have the same cultural background (for example, in some cultures youth is evaluated positively and old age negatively; in others, the reverse is true). Moreover, the same speaker can evaluate the same object (or action, person, etc.) in different ways if the extra-­linguistic context changes. While descriptive evaluation usually relies on the real and permanent characteristics of an item,4 qualitative evaluation is often based on temporary and variable parameters or situations: the first kind of evaluation is a description; the second one is a sort of interpretation, and often involves pragmatics as a crucial parameter (and makes a wide cross-­linguistic comparison of available data more difficult). Qualitative evaluation assumes a subjective evaluation: personal feelings or opinions and, often, the influence of extra-­linguistic context become the crucial factors. In most of these cases, there is a

Introduction:Why Evaluative Morphology? 11 pragmatic variable that cannot be omitted in the description of the meaning of the complex words. This variable, if present, is not unpredictable: some morphological processes have regular pragmatic effects; this means that some evaluative constructions are the regular formal correlate of extra-­linguistic and contextual constraints. Nevertheless, there are numerous instances of qualitative evaluation which can be accounted for by means of systemic constraints only, independently of the specific speech situation or context. Consider, among others, Slovak mam-­isko ‘mother-­aug’, expressing a pejorative evaluation, and mam-­ička ‘mother-­dim’, with an affectionate meaning: in both of these cases the qualitative evaluation is inherent in the meaning of words at the system level. Similar examples abound cross-­linguistically. This observation is reflected in Körtvélyessy’s (2014) model of evaluative formation in the form of two alternative paths of evaluative formation. To sum up, according to these perspectives the pragmatic-­semantic functions mentioned above can be grouped as follows: (1) Descriptive perspective diminution in quantity and/or quality augmentation in quantity and/or quality age variation

Qualitative perspective approximation/reduction/attenuation intensification endearment hypocorism expression of social position contempt authenticity/prototypicality

The same semantic functions can be reorganised following a different criterion if we conceive the semantic scale, according to which the evaluation takes place, as an axis with two opposite ends: positive (+) and negative (−) (the standard or default value corresponds to zero). In other words, descriptive evaluation expresses a shift towards the positive (+) end of the axis if it indicates an increase in the possession of a specific property in quantity and quality (for example, physical dimension), and a shift towards the negative (−) end of the axis if it indicates a decrease in the possession of a specific property, once again in quantity and quality. So, as for the dachshund and the Great Dane mentioned above, the former indicates a shift towards the negative pole of the axis corresponding to the semantic scale ‘physical dimension’; the latter indicates a shift towards the positive end. As for qualitative evaluation, this can be represented by a shift towards the positive end of the axis corresponding to the semantic scale ‘feeling’ if it expresses a positive feeling of the speaker towards an object, an action, a person, etc.; and towards the negative end of the axis if it expresses a negative feeling of the speaker. The following is a reorganisation of the functions mentioned above: (2) Shift towards the positive end augmentation old age intensification endearment hypocorism expression of high social position authenticity/prototypicality

Shift towards the negative end diminution young age approximation/reduction/attenuation contempt expression of low social position

12

Nicola Grandi and Lívia Körtvélyessy

Matching the two schemes in (1) and (2) produces the following table: (3) Descriptive perspective Qualitative perspective Shift towards the positive end augmentation intensification age variation endearment authenticity/prototypicality expression of social position Shift towards the negative end diminution approximation/reduction/ age variation attenuation contempt hypocorism expression of social position This table in (3) is still a list of semantic functions, and, as a consequence, it does not reach a satisfactory degree of generalisation and abstraction. In order to reach a higher level of generalisation, the most promising perspective is that of a suggestion by Wierzbicka (1989), namely considering the functions usually performed by evaluative affixes as the instantiation of typological (possibly universal) prototypes, based on semantic primitives: SMALL and BIG are representative of the descriptive side of evaluation, while the semantic primitives GOOD and BAD represent its qualitative side. Therefore, if we adopt this alternative notation, the previous table may be reformulated as follows: (4) Descriptive perspective Shift towards the positive end BIG Shift towards the negative end SMALL

Qualitative perspective GOOD BAD

Thus, typological comparison based on functional identity allows us to sketch a pattern of variation which is defined by the combination of two parameters: the two dimensions of evaluation on the one side and the two ends of the axis (corresponding to the appropriate semantic scale) on the other. In this manner, some prototypes that are supposed to be cross-­ linguistically constant and recurrent may be identified: • Prototypical diminutives indicate a shift towards the negative end on the descriptive axis. • Prototypical augmentatives indicate a shift towards the positive end on the descriptive axis. • Prototypical pejoratives indicate a shift towards the negative end on the qualitative axis. • Prototypical amelioratives indicate a shift towards the positive end on the qualitative axis. These prototypes can be organised as follows: (5) Descriptive perspective Qualitative perspective Shift towards the positive end Prototypical augmentatives Prototypical amelioratives Shift towards the negative end Prototypical diminutives Prototypical pejoratives

Introduction:Why Evaluative Morphology? 13 Needless to say, the two perspectives often overlap and a single construction can express a quantitative and a qualitative evaluation at the same time. This pattern of variation is neutral with respect to the formal behaviour of single languages: it merely indicates that each construction which expresses a shift towards the negative end on the descriptive axis can potentially be labelled as ‘diminutive’; that each construction that indicates a shift towards the positive end on the descriptive axis can potentially be labelled as ‘augmentative’; etc. Consequently, the pattern is universal, but it is not to be expected that all languages abide wholly by it or realise it in the same way. In other words, there are languages in which all the pattern is expressed fully, and languages in which parts of it are unexpressed. This explains why some languages possess a rich evaluative morphology and others do not. In other words, as stated by Körtvélyessy (2012b), some languages possess a high value of evaluative morphology saturation while others have a low value. Moreover, languages can differ with respect to the formal strategies used to express the semantic operations that complete the pattern: these formal strategies depend on the global morphological shape of each single language or group of languages (the morphological type, the preference for prefixes or suffixes, etc.5) and they therefore pertain to the level of ‘specific language facts’. In other words, typological comparison based on functional identity does not predict the exact behaviour of individual languages: it allows us to define universal prototypes for evaluative functions, without defining their concrete instantiations and their boundaries. According to the premise formulated above, a definition of ‘evaluative construction’ which is not based on language specific criteria can be formulated (see also Grandi 2002). A construction falls into the scope of evaluative morphology if it satisfies two conditions, one relating to the functional-­semantic level and the other to the formal level. The first condition indicates that a linguistic construction can be defined as evaluative if it has the function of assigning a value which is different from that of the standard or default (within the semantic scale to which it pertains) to a concept: this value usually coincides with a shift towards the negative or positive end of the scale and is assigned without resorting to any parameters of reference which are external to the concept itself. The second condition indicates that an evaluative construction must include at least the explicit expression of the standard value (by means of a linguistic form which is lexically autonomous and is recognised by the speakers of the language as an actual word) and an evaluative mark (a linguistic element specifically devoted to expressing this shift). The expression of such a standard form or concept usually coincides with the base of a synthetic construction and with the head of an analytic construction. The evaluative mark can be a suffix, a prefix, the reduplication of the lexical morpheme, a circumfix, etc. As this data confirms, functional identity does not imply formal identity.6 To conclude, it must be said that evaluation shows interesting correlations with comparison. Comparative constructions indicate a deviation, but in this case the term of comparison is overtly mentioned, and does not coincide with an abstract standard form/concept or default value, conditioned by cultural and social factors and shared by all the members of a community. For example, the Italian sentence questa è una casetta ‘this is a small house (lit. house-­dim)’ means that the house in question is small with respect to the standard size of Italian houses. The sentence la mia casa è più piccola della tua (‘my house is smaller than yours’) means that I am taking your house as a reference point; so the standard size of Italian houses is not involved at all. Both houses (mine and yours) may be small. Generally speaking, comparative constructions express a relative evaluation, while evaluative constructions express an absolute evaluation.

14

Nicola Grandi and Lívia Körtvélyessy

This definition allows us to draw clear boundaries for the category under examination, and to propose objective criteria in order to establish the membership of ambiguous cases. Consider, for example, the Italian word tunisino ‘Tunisian’. It displays a suffix, -­ino, which usually performs a diminutive function (as stated above) together with a base word which is clearly identifiable (Tunisia). However, -­ino in this form does not perform a function that can be traced back to the evaluative values mentioned above: it does not express a shift towards the negative end of a semantic scale. Therefore in this case the formal condition is satisfied, but the functional-­semantic one is not: even formal identity does not imply functional identity. A word such as pulcino ‘chick’ displays the same suffix, and is semantically consistent with the evaluative value SMALL, thus carrying a descriptive evaluation. Nevertheless the base expressing the standard cannot be identified, since the sequence pulc-­does not correspond to an actual Italian word meaning ‘hen’. So in this case, the semantic condition is fulfilled, but the formal one is not. As a consequence, even pulcino cannot be included among evaluative constructions; or, better, it cannot be mentioned as a prototypical instantiation of an evaluative construction. Broadly speaking, it belongs to a class that we may call ‘lexical diminutives’, like rabbit/bunny, cat/kitty, kitten, etc. characterised by a suppletive relation between the diminutive form and the word expressing the standard form. Finally, consider the case of Italian mangione ‘hearty eater’: a suffix usually used to express the augmentative meaning, -­one, is clearly recognizable. The base word is the verb mangiare ‘to eat’. In the meaning of mangione, although the evaluative value BIG is evident, it merges with an agentive meaning. In this case, we can place mangione among evaluative constructions, but it must be placed in a peripheral position with respect to more typical forms such as tavolone ‘big table’ < tavolo ‘table’. Therefore ‘evaluation’ can be conceived as a category with different levels of membership. Another special (even more peripheral) case should be mentioned: lexicalised diminutives, like Slovak stolička ‘chair’ < stolica ‘stool, big chair’‚ lyžička ‘teaspoon’ < lyžica ‘spoon’, etc. While the suffix is diminutive, these words identify ‘standard size’ and the objects denoted may be smaller or bigger than the standard form.7 In conclusion, the simultaneous application of the two parameters unambiguously defines the notion and the role of ‘evaluation’ in morphology. However, once two parameters are being dealt with, the possibility that just one of them is present has to be taken into consideration, and an unequivocal decision on the status of ambiguous data has to be proposed. As shown above, the violation of the functional-­semantic condition usually determines the exclusion of a construction from evaluation morphology. By contrast, the violation of the formal condition does not have the same consequences. As a result, the class of evaluative constructions can be better described in terms of different degrees of membership: the two criteria relate to its core area, or, in other terms, prototypical examples of it. Other constructions can be considered as non-­prototypical (or peripheral) members of the same class. Moreover, cases of multiple class membership cannot be excluded. In other words, a single item can be classified differently if it performs more than one evaluative function, in both a paradigmatic and a syntagmatic way. Naturally, the only case of overlapping between different values which is excluded from the classification corresponds to a vertical reading of the criteria proposed in the table in (4) above (BIG-­SMALL and GOOD-­BAD). Finally, occasional occurrences of linguistic items or constructions in atypical roles have to be taken into account. This may be the case, for example, of a diminutive which is used ironically in order to designate a particularly large entity.

Introduction:Why Evaluative Morphology? 15

4  Presentation of the book and summaries of the chapters Many research projects these days start with a simple activity – googling. Application of this method to the term ‘evaluative morphology’ brought surprisingly poor results a decade ago, even when it was supported by other keywords, e.g. diminutive, augmentative, ameliorative, pejorative, expressive morphology, appreciative suffixes, intensification, etc. Nowadays, the same search offers at least ninety web pages directly related to evaluative morphology. Undoubtedly, when Scalise introduced evaluative morphology in 1984, he could not have predicted the enormous impact that his intuition would have had on further development in the field, leading to numerous conference papers and articles.8 However, a systematic and comprehensive coverage of the multiplicity of issues related to evaluative morphology was lacking. We believe that our volume addresses this issue and therefore fills this undesirable gap. The ambitions of this handbook of evaluative morphology are therefore fourfold: (i) to map the state of the art in the field of evaluative morphology, (ii) to avoid a biased approach to evaluative morphology that presents it merely on the basis of a single theoretical framework, (iii) to map the diversity of topics by involving leading experts in the field, and (iv) to look specifically at evaluative morphology phenomena from a cross-­ linguistic perspective. In order to fulfil these objectives, the volume is divided into two parts. The first part provides the theoretical background to the second, descriptive section. The first part of the book is divided into twelve chapters (in addition to this Introduction). Chapters 2 and 3 provide a semantic and pragmatic perspective of the domain of evaluative morphology. The former is titled ‘The semantics of evaluative morphology’ and demonstrates that semantics has become one of the central topics of research into evaluative morphology. Victor M. Prieto takes Jurafsky’s (1996) radial model of diminutives as a point of departure and proposes a new, more general, semantic core for diminutives. In addition, he proposes an analogical radial model for augmentatives. Lavinia Merlini-­Barbaresi (Chapter 3) discusses pragmatic aspects of evaluative morphology, this being – with a few exceptions, mentioned in Section 2 above – an untilled field of evaluative morphology. There is no doubt that the meaning of an evaluative construction can be modified or even changed to its opposite as soon as it leaves langue, i.e. the level of the system, and that the meaning of some evaluative constructions includes a pragmatic variable. The borderline between morphology and pragmatics is fuzzy and raises a number of questions which are discussed by this author. In Chapter 4, ‘Word-­formation processes in evaluative morphology’, Pavol Štekauer provides a typological outline of various word-­formation processes used in the languages of the world. While suffixation dominates evaluative morphology, other word-­formation and inflectional processes, often neglected in the literature, are also in play. These are illustrated in this chapter, with numerous examples. Chapter 5, ‘Evaluative morphology and language universals’ by Lívia Körtvélyessy, maps language universals in evaluative morphology. Körtvélyessy examines the situation at individual levels of the language system (phonetics, phonology, morphology, syntax, lexicon, etc.) by referring to the Konstanz Universals Archive and to other resources. Furthermore, on the basis of her own research, she identifies some Euroversals, i.e., universal tendencies in evaluative morphology within Standard Average European. Nicola Grandi discusses the controversial and widely debated issue of ‘The place of evaluation within morphology’ in Chapter 6. Grandi pays special attention to the position of evaluative morphology within the inflection–derivation continuum. His arguments are primarily based on an analysis of evaluative affixes from a cross-­linguistic perspective.

16

Nicola Grandi and Lívia Körtvélyessy

Since the categories of gender and number represent another poorly studied area of evaluative morphology, considerable space is given to them in Chapter 7, ‘Evaluative morphology and number/gender’. Nicola Grandi maps the situation across various language systems and illustrates some systematic correlations between evaluative affixes on the one hand and gender markers and number categories on the other, supporting his claims by providing data from different languages. The categories of diminutiveness and augmentativeness have been traditionally related – erroneously – to the class of nouns only. Cross-­linguistic research provides us with substantial evidence that other word classes, including verbs, adjectives and adverbs, play a significant role in the overall structure of evaluative morphology. Therefore, a special chapter of the volume (Chapter 8) is devoted to ‘Evaluative morphology and aspect/actionality’. Lucia M. Tovena provides a comprehensive overview of the situation regarding this aspect of evaluative morphology, focusing in particular on the category of pluractionality. The use of evaluative morphology has always been associated with particular speech situations, specific genres, and social variables such as age, sex and social status. Livio Gaeta, in Chapter 9 on ‘Evaluative morphology and sociolinguistic variation’, draws an exhaustive picture of how evaluative morphology, social contexts and socio-­pragmatic environments can interact. Evaluative morphology is a characteristic feature of child speech. For this reason, in Chapter 10 by Wolfgang U. Dressler and Katharina Korecky-­ Kröll, ‘Evaluative ­morphology and language acquisition’, the authors take a morphopragmatic approach to evaluative morphology and account for the chronological precedence of diminutives and hypocoristics over other patterns of derivational morphology. Any linguistic issue can be studied from either a synchronic or a diachronic point of view. While preceding chapters focus primarily on the synchronic facet of evaluative morphology, Katrin Mutz (Chapter 11) approaches the field from a diachronic perspective. It is noteworthy that she provides a new panchronic model stemming from her observation that many diminutive suffixes in Indo-­European languages have developed from derivational affixes meaning ‘related to’ or ‘similar to’. Even though the final two chapters in Part I focus on the theoretical facet of evaluative morphology, they also describe a particular group of languages. The specific position of the languages under description (pidgin, creole and sign languages), along with the theoretical nature of the chapters, create a bridge between the two parts of the volume. Giulia Petitta, Alessio Di Renzo and Isabella Chiari focus in Chapter 12 on ‘Evaluative morphology in sign languages’, a previously unexplored research field. While the discussion relies on Italian Sign Language, the authors draw relevant generalisations pertaining to the lexical, morphological and prosodic levels. Barbara Turchetta (Chapter 13) draws a sketch of the development of evaluative strategies in pidgins and creoles, focusing on reduplication and repetition. As she points out, pragmatic factors seem to play a prominent role in this field. From this point of view, the position of ‘Evaluative morphology in Pidgins and Creoles’ is a highly intriguing issue. The author analyses the field by taking into account the relation between semantics and cultural perception of the world and by taking into consideration the speaker’s perception of extra-­linguistic phenomena. Part II includes descriptions of evaluative morphology in almost sixty languages of the world. This part is structured according to geographical criteria. In order to avoid repetition of well-­known facts, languages which are well documented in the field have been

Introduction:Why Evaluative Morphology? 17 excluded. These include, for example, Italian, Spanish and Russian (see Section 2 above). Nevertheless, a considerable amount of data from these languages is presented and widely discussed in the chapters in Part I. For Part II, a considerable effort was made to obtain descriptions of languages which were previously unaccounted for in this field and/or for lesser-­known languages. While we cannot claim our sample is an ideal representation of the languages of the world, it provides a new and fairly representative picture of the role of evaluative morphology worldwide. For more details on the chapters in the second part of the volume, see Körtvélyessy in Chapter 14.

5  List of languages mentioned in the volume The following list consists of languages (also including dialects and varieties) that are mentioned in the volume. It gives their areal and genetic classification. The names of listed languages are those used by the individual authors. The classification is based on The World Atlas of Language Structures (WALS; Dryer and Haspelmath 2011). If the language was not identified in WALS, the Ethnologue was used (Lewis, Simons and Fennig, 2014). Eurasia Afro-­Asiatic: Hebrew, Israeli Hebrew, Maltese Altaic: Azerbaijani, Chagatai, Even, Evenki, Gagauz, Karachay-­Balkar, Kazakh, Kirghiz, Manchu, Nanai, Negidal, Oroch, Orok, Oroqen, Solon, Tatar, Trakai Karaim, Turkish, Udihe, Ulcha Dravidian: Kannada, Malayalam, Tamil, Telugu Indo-­European: Albanian, Ancient Greek, Austrian German, Belorussian, Bengali, Breton, Bulgarian, Catalan, Croatian, Czech, Danish, Dutch, English, Faroese, French, Frisian, German, Greek, Icelandic, Irish, Italian, Kashubian, Ladin, Latin, Latvian, Lithuanian, Low German, Lower Sorbian, Luxembourgish, Macedonian, Modern Greek, Norwegian, Old Church Slavic, Ossetic, Persian, Portuguese, Romanian, Russian, Sanskrit, Sardinian, Scottish Gaelic, Serbian, Slovak, Slovene, Slovenian, Spanish, Swedish, Tat, Welsh Kartvelian: Georgian, Gurian, Khevsurian, Laz, Megrelian, Mokhevian, Pshav, Rachian, Svan Nakh-­Daghestanian: Aghul, Avar, Bezhta Northwest Caucasian: Kabardian Uralic: Enets, Estonian, Finnish, Hungarian, Khanti, Komi, Livonian, Mansi, Mari, Mordvin, Nenets, Nganasan, Sami, Selkup, Udmurt Yeniseian: Arin, Assan, Ket, Kott, Pumpokol, Yugh Yukaghir: Kolyma Yukaghir Isolates: Basque, Japanese, Korean, Nivkh Sign languages: British Sign Language, German Sign Language, Israeli Sign Language, Italian Sign Language South-­East Asia and Oceania Austro-­Asiatic: Khasi, Mang, Mnong, Vietnamese Austronesian: Agta, Anejom, Apma, Bahasa Indonesia, Balangao, Bikol, Ibaloi, Ilocano,

18

Nicola Grandi and Lívia Körtvélyessy

Javanese, Malay, Mansaka, Muna, Nelemwa, Romblomanon, Tagalog, Tausūg, Vitu, Yami Eastern Malayo-­Polynesian: Maori Hmong-­Mien: White Hmong Sino-­Tibetan: Ao, Burmese, Changsha Xiang, Chinese, Classical Tibetan, Dolakha Newar, Dumi, Hong Kong Cantonese, Lahu, Leizhou Southern Min, Lhasa Tibetan, Lisu, Mandarin, Old Tibetan, Pingding Mandarin, Shanghai Wu, Standard Mandarin Chinese, Taiwanese Hakka, Taiwanese Southern Min, Tibetan, Wenzhou Wu, Xinyi Cantonese Tai-­Kadai: Thai Pidgins and creoles: Bislama, Plantation Pidgin Fijian Australia and New Guinea Australian: Adnyamathanha, Alawa, Bāgandji, Bininj Gun-­wok, Dalabon, Djaru, Jaminjung, Jawoyn, Jingulu, Kaurna, Kayardild, Lardil, Mparntwe Arrernte, Ngalakgan, Ngandi, Ngiyambaa, Nhanda, Nukunu, Rembarrnga, Warlpiri, Warray, Yukulta Sepik: Abelam, Alamblak, Boikin, Iatmul, Manambu, Ngala, Yelogu Trans-­New Guinea: Anamuxra, Koiari Pidgins and creoles: Kriol, Tok Pisin Sign languages: Australian Sign Language Africa Afro-­Asiatic: Arabic, Berber, Classical Arabic, Colloquial Jordanian Arabic, Dangaléat, Figuig, Gen, Hamar, Hausa, Maale, Migama, Modern Standard Arabic, Moroccan Arabic, Pévé, Saho, Somali, Zenaga Bantu: Kíhehe Indo-­European: Afrikaans Kadugli: Krongo Khoisan: Khoekhoe, Nama Niger-­Congo: Aghem, Akan, Bafut, Bemba, Buli, Eton, Ewe, Fon, Fongbe, Fula, Gikuyu, Gungbe, Jóola Banjal, Kongo, Kɔnni, Lega, Lingala, Mambay, Mandingue, Noon, Nupe, Sɛlɛɛ, Shona, Sotho, Swahili, Swati, Tswana, Venda, Xhosa, Xwla-­Xweda, Zulu Nilo-­Saharan: Aja, Kanuri, Karamojong, Koyra Chiini, Maasai, Ngiti, Sara Mbay, Supyire, Turkana Pidgins and creoles: Angolar, Cameroonian Pidgin English, Juba Arabic, Kituba, Krio, Nubi, São Tomé Creole, Turku, West African Pidgin English Sign languages: Adamorobe Sign Language North America Algic: Cree, East Cree, Innu, Micmac, Nipmuck, Odawa, Ojibwe, Passamaquoddy, Plains Cree, Saulteaux, Western Swampy Cree, Woods Cree, Yurok Athabaskan (Athabascan): Dakelh (Carrier), Dena’ina Chibchan: Boruca, Bribri, Cabécar

Introduction:Why Evaluative Morphology? 19 Eskimo-­Aleut: Alaskan Inupiaq, Greenlandic, Inuktitut, Yupik Hokan: Yana Huavean: Huave Mayan: Maya, Tzeltal Mixe-­Zoque: Zoque Muskogean: Alabama, Chickasaw, Choctaw, Creek, Hitchiti, Koasati, Mikasuki Na-­Dene: Ahtna, Dane-­zaa, Dëne Sųłiné, Hupa, Koyukon, Mattole, Navajo, Slave, Slavey, Tanacross, Tłįchǫ Yatiì, Upper Tanana, Western Apache, White Mountain Apache, Witsuwit’en Otomanguean: Chocho, Huautla Mazatec, Ixcatec, Popoloca Salishan: Lushootseed, Squamish Siouan: Lakota Uto-­Aztecan: Nahuatl, Pipil Wakashan: Kwakw’ala, Nootka Pidgins and creoles: Haitian, Jamaican Creole, Martinica Creole Sign languages: American Sign Language South America Arawakan: Trinitario Aymaran: Aymara, Jaqaru, Kawki Chibchan: Bocotá, Guaymí, Malécu, Naso Guaicuruan: Kadiwéu, Mocoví, Pilagá, Toba Matacoan: Chorote, Nivacle, Wichí Mosetenan: Chimane, Mosetén Nadahup: Hup Nambikuaran: Latundê Panoan: Matsés Quechuan: Bolivian Quechua Tucanoan: Southern Barasano Tupian: Mekens, Salamãy, Sirionó, Tapiete, Yuki Isolates: Aikanã, Besiro, Guató, Kanoê, Kwaza, Lule, Movima, Vilela, Yurakaré Pidgins and creoles: Berbice Dutch Creole, French Creole, Ndyuka, Papiamentu, Saramaccan, Sranan, Suriname Creole

Notes 1. Cf. Italian balcone [bal'kone] ‘balcony’ → balconcino [balkon’tʃino] ‘small balcony’. 2. The order DM > EM > IM reproduces the linear sequence of derivation, evaluation and inflection in Italian words: campana ‘bell’ → campanile ‘bell tower’ → campanilino ‘small bell tower’ (the final -­o is the inflectional ending, expressing the categories of masculine gender and singular number). 3. See also Turchetta (this volume) for the role of pragmatics in the developments of evaluative strategies in pidgins and creoles. 4. With the possible exception of age variation. Nevertheless, the standard value is sometimes determined culturally (e.g., what is a large house in one culture may be a small house in some other; this case is not a subjective matter). 5. In other words, languages belonging to different types use different formal strategies to perform the same function.

20

Nicola Grandi and Lívia Körtvélyessy

6. Of course, the reverse can be true too. The same formal strategy can have different functions in different languages. 7. The original reason for the coining of these words was substandard size. Nowadays they are used as ‘default-­size’ names for the objects in question. 8. The first conference devoted to evaluative morphology was organised in Toulouse, in 1999: ‘La morphologie des dérivés évaluatifs’.

2  The Semantics of Evaluative Morphology Victor M. Prieto

1 Introduction Describing the meaning of evaluative affixes (evals; Scalise 1984; Stump 1993; Bauer 1997), such as diminutives (dims) and augmentatives (augs), is not straightforward. Even if seemingly basic, this is a complex task considering the vast diversity of evals’ pragmatic functions and nuances of meanings. This chapter approaches this task from a cognitive perspective. Cognitive linguistic theories give a plausible explanation of the diverse and complex network of meanings of evals. In what follows, we apply the four major principles of cognitive semantics (Evans, Bergen and Zinken 2007) in the description of evals’ meaning.

2  evals’ meaning: bodily grounded on the physical property of size and size perception Cognitive semanticists’ first principle is that conceptual structure is embodied because of the nature of our bodies: anatomical, neurological and physically constrained. We organise our reality in relation to the nature of our bounded bodies as our first human experience. The notion of ‘littleness’/‘bigness’, essential to evals, relates to size, a bounded bodily property of humans and other objects, and the physiological perception of this size via visual scanning. Thus, the starting point for the meaning of evals is physiologically grounded. Our living in the world is at least partially structured and determined by our bodies, a fact which affects cognition significantly. Our concepts and perceived reality are a function of our embodiment (Evans et al. 2007, 10). It is almost impossible to talk of phenomena or things we cannot perceive, and those we do perceive are perceived via our embodied experience. There has to be, then, an important relation and resemblance between our human minds and our body-­bounded experiences; this is embodied cognition. Human conceptual structures are embodied because the nature of human concepts results from our embodiment nature (Evans et al. 2007; Pecher and Zwaan 2005, 2). Thought has a bodily basis (Seitz 2000). When using language, we use brain areas that activate for perception and action, and we create mental embodied simulations. Meaning is more than just word definitions. For meaning, our brains function in a creative mode and construct mental worlds where we act (Berger 2012); a fundamental capacity to understand language. Our embodied rationality, then, influences what we do and what we are. Furthermore, our temporal-­spatial orientation, our interaction with objects, and the pattern of our bodily

22

Victor M. Prieto

activities shape our reality (Johnson 1987, xix). These bodily experience structures (image schemas) become abstract meanings and inference patterns. Thus, we begin with primitive concepts like ‘size’ and develop this into more abstract concepts: ‘endearment’, ‘importance’ and other evals connotations. We begin with a more quantity-­based primitive notion and expand to a more quality-­based abstract concept. Figurative extensions often come via a ‘metaphorical projection from the realm of physical bodily interactions’ and problem-­ solving (Johnson 1987, xx), naturally using much imagination (p. ix). The neurology behind language comprehension is contextual and constructive (Coulson 2006, 250–1). Coulson showed that many ‘phenomena are context dependent, including . . . depth planes [i.e. size]’ (p. 249). People activate perceptual symbols of referents (i.e. size) during language comprehension (Barsalou 1999). Language comprehenders mentally represent shape of objects (Zwaan, Stanfield and Yaxley 2002). Pecher and Zwaan show that ‘perceptual and motor representations play a role in . . . understanding language’ (2005, 2), and that language developed from perception and action. Grammar has important sensorimotor foundations, and simple perceptual processes such as visual scanning are important to sentential meaning in many instances (Langacker 1987). Coulson (2006) relates cognition to figurative language comprehension (p. 246). She argues that visual perception is active and constructive; humans use these perceptions to interact with their surroundings (p. 246). Consequently, evals’ meaning is grounded in the fundamental bodily function of visual perception, and visual perception of size, particularly; visual processing is a critical and widespread brain activity (Coulson 2006, 247). Language users then exploit optical information regarding size of objects around them to interact with their surroundings. Coulson shows that figurative language, common in evals, may be connected with brain function and complex cognitive activities (p. 247). We know that linguistic and non-­linguistic knowledge together construct meaning (Sperber and Wilson 1995). There is, then, a critical, complex interplay between visual perception processing and linguistic meaning construction. 2.1  Size: embodied conceptual primitive category We propose the category of SIZE as the major fundamental starting point for the meaning of evals. Size and size perception are critical for the organisation of our surrounding world (Schwarzkopf, Song and Rees 2010). Coulson shows that the brain strives for information on properties of objects such as size (2006, 249). It is known that size and similar properties are the most basic properties of physical objects (Descartes 1901). Bornstein and Lamb (1992) explain that infants develop a sense of the size of objects by the age of four to five months, together with binocular vision (p. 21). Humans are wired to perceive size. This early emergence of size perception seems to highlight its importance in the categorisation and organisation of the world. It may, then, be critical for our survival. That is probably the reason for a certain natural human fear of large objects. Psychiatrists describing the dissociation of macropsia (perception of objects as larger-­than-­normal items) patients explain that a gigantic surrounding environment seems hostile and aggressive to a person (Schneck 1965, 2). Fear of large objects is among the very common fears of 0–2-­year-­old infants (Child Anxiety Network 2001; Harvard Medical School 1988). Fear of large objects seems to be one of the first fears humans experience; maybe another reason humans strive for size perception. For humans, many fears are normal, useful for survival, and almost hardwired

The Semantics of Evaluative Morphology 23 or part of the chemistry of the brain (Pincus 2012, 16). Asma (2009) reports that victims startled to discover a burglar in their home usually describe the size of the intruder as much larger than they actually are, and acrophobia patients actually see the distance between themselves and the ground as larger than it really is (p. 22). Thus, the traditional concept of size embedded in evals is a concept already wired in the brain of humans for their survival and the organisation of the world. This is probably why size is a semantic primitive in Wierzbicka (1996). Semantic primes (e.g. BIG/SMALL, GOOD/BAD) are concepts innately understood that can’t be further analysed, and are often used to speak of other concepts. Because of the biological and psychological importance of size and size perception for humans, embodied cognition, and semantic universal primes such as BIG/SMALL, we argue that, synchronically, the cognitive starting point for the meaning of evals is consequently ‘big’ for augs and ‘small’ for dims. These size-­grounded concepts are primitive and embodied, and this is where all the semantic diversity of evals begins.

3  evals’ meaning: conceptually structured as size-­grounded evaluation Another guiding cognitive principle in cognitive semantics is that semantic structure is conceptual structure. Language refers to concepts rather than referents. We think of conventional meanings as concepts, even though these are not always identical (Evans et al. 2007, 11); there may be concepts or ideas in our minds that are not lexically founded (Langacker 1987). However, every conventional meaning is a concept, representationally rather than denotationally. Following this principle, we consider evals’ meanings as a conceptual structure: the conceptual structure of size evaluation. 3.1  Size evaluativeness: a conceptual structure evals are evaluative because they convey, consciously or subconsciously, a valuing of the referents or audience, according to the speaker’s judgement. For Hunston and Thompson (2000) one type of evaluation-­driven marking in language is affective (good–bad) opinion, and another is importance (subjective value in relation to degree of relevance). This makes dims, augs, pejs and amls evaluatives. These authors suggest that the parameter of importance may relate to a basic good–bad evaluation, and that identifying evaluation ‘is a question of identifying signals of comparison, subjectivity and social value . . . evaluation consists of anything which is compared to or contrasts with the norm’ (p. 13). When we perform this evaluation via bounded morphemes, we have evaluative morphology. Evaluative morphology, then, is synthetic marking of size and by extension positive/ negative emotional affect and related concepts. This implies diminution and augmentation prototypically and by association issues such as pejoration and amelioration.

3.2  evals illustrated: diminutives Let us now illustrate this concept of evaluation with dims, the most prototypical and studied eval (Körtvélyessy and Štekauer 2011b). Alonso (1937) sees dims as markers of affection or playful mood. For Náñez (1973), dims indicate a conceptual quantitative distinction regarding the magnitude of the entity referred to by the base. This often refers to the concept of smallness but recognises other important connotations. Dressler and Merlini Barbaresi (1994), looking at Italian and German particularly, describe dims as evaluative

24

Victor M. Prieto

since they ‘express an evaluation or judgment “as to value” (not “as to fact”), according to the evaluator’s intentions, perspective and standards of evaluation’ (p. 153). This is precisely what makes this morphology evaluative. dims evaluate an entity as small (in an assumed size plane) and/or (un)appreciated, and in terms of related concepts. dims are semantic/pragmatic markers of a subjective evaluation/appreciation of (often) littleness/ affection towards the referents. We conclude that size evaluativeness may originate as a human conceptual category and then may develop special significance and consequently gets coded in the language. Now, let us consider that diversity of meaning where the idea of physical size is not (overtly) present.

4  evals’ meaning: encyclopedia accessed via metaphoric linking and similar associations A major principle in cognitive semantics is that meaning representation is encyclopedic. This third guiding principle holds that semantic structure or lexical concepts do not represent neatly packaged bundles of meaning. Consequently evals are, in cognitive semantic terms, ‘points of access’ to an array of connected ideas relating to a particular concept, conceptual domain or encyclopedic meaning (Langacker 1987). We do recognise that evals have conventional meanings but these are simply ‘prompts’, in cognitive semanticists’ terms: ‘for the process of meaning construction: the “selection” of an appropriate interpretation against the context of the utterance’ (Evans et al. 2007, 12). 4.1  Diverse meanings of evals explained via radial categories We insist that an explanation for the multiplicity of meanings is extension or association from the basic size notion described above. Even evals with meanings other than size (e.g. Spanish Victor-­cito meaning ‘Victor dear, dear Victor’ rather than ‘little Victor’) are still linked to the semantic prime of SIZE via metaphoric or metonymic chaining. Lakoff and Johnson (1980) attest the importance of metaphors for human communication. We definitely recognise the ‘important role that metaphors play in shaping thoughts and ­experiences’ (Asma 2009, 13). Lakoff’s (1987) radial categories model, applied in Jurafsky (1996), will explain this. This model emerges from Rosch’s (1983) prototypes and Wittgenstein’s (1974) family resemblances. They show that some categories do not express a single concept or meaning but are characterised by related features. Resemblances are widely shared among the different nodes of semantic mappings in an overlapping fashion. Category members have an internal structure: there are typical, exemplary and even anomalous members. In evals, we see a ‘radial structure’, since there is indeed a core prime. For Rosch, a prototype functions as a cognitive reference point (i.e. size). The central subcategory (e.g. littleness for dims, as suggested here) of this network provides the basis for extending the category in new ways and for defining variations. Lakoff suggests that at the cognitive core of the formation of categories, we find image schemas and their metaphorical tokens; hence, metaphor and metonymic chaining are common useful cognitive tools in this type of semantic mapping. There may be criss-­crossing of category members with hardly any elements in common, but each must overlap with other members of the category, even if fuzzy boundaries are tolerated (Wittgenstein 1974). Such is the case of seemingly opposing connotations such as amelioration (e.g. endearment) vs. pejoration and ‘littleness’ vs. ‘very’ (in the case of

The Semantics of Evaluative Morphology 25 related–to

imitation G,M

SOCIAL GROUPS ARE FAMILIES

pets

I

I

sympathy

member

I

intimacy

L

L

M

child

I M

exactness

small type–of

CATEGORY CENTRALITY IS SIZE

affection

SEMANTICS

G

approximation

GENDER IS SIZE

small M

L,M

partitive

female M M

contempt

hedges PROPOSITIONS ARE OBJECTS

PRAGMATICS

Figure 2.1  Radial model of diminutives (Jurafsky 1996) dims), and attenuation vs. intensification (in the case of augs) in Prieto’s (2005) data. This model is based on the tendency of humans to perceive categorically and categorise on the basis of prototypes. Jurafsky (1996) tries to explain the varied and contradictory dim senses. Jurafsky and Lakoff also see the body as central in grounding interpretations of the world, including those involving power-­dominance issues. Thus, much of Jurafsky’s reasoning in linking chains for dims relies on this body–world connection; consistent with our first guiding principle above. Jurafsky’s (1996) model is a structured polysemy that clearly binds the various dim senses. Jurafsky’s diagram (Figure 2.1) indicates the sequential order of semantic associations (represented by the direction of the arrows). For example, the central position of the [child] sense and the arrow that link it to [affection] simply mean that the sense of childness (the core) is first, and then it expands to affection and so on. New nodes also may originate other ideas or connotations.

4.2  dim radial category There are basically two formal hypotheses about the basic semantics of dims from which many other variations in meaning emerge: ‘small’ and ‘child’ (Voeykova 1998). Here, we favour the more traditional perspective, ‘smallness’, or better yet ‘littleness’. Prieto (2005) presents ‘little’ as the synchronic cognitive core meaning. Even though we recognise how ‘child(ness)’ could be justified diachronically, at least in some languages, as the etymological origin of dims, the arguments below justify the cognitive need of a basic meaning of ‘littleness’. We have already established the cruciality of the concept of size for humans. Following our first guiding principle above, we support a more biological basis for dims. The perception of size is embodied and biologically wired, which is shown in visual perception activity. The perception of the concept of ‘child’ or age properties is more culturally, not cognitive, motivated and comes later in life; later than size perception. All the examples in Jurafsky (1996) and Prieto (2005), for example, link more closely and sensibly to the central concept of ‘littleness’ than to ‘childness’. Voeykova (1998) observed that the ‘smallness’ sense is very important for the child when acquiring dims in Russian. We value the arguments favouring the ‘child’ meaning,

26

Victor M. Prieto

but diachronically at least: ‘this meaning is less relevant for the acquisition of language by a child in comparison to “smallness”, since it demands a very high degree of abstraction’ (Voeykova 1998, 112). Jurafsky himself argued that dim is ‘among the grammatical primitives which seem to occur’ (near-­)universally (1996, 534). ‘Littleness’ fits more within the category of ‘grammatical primitives’ than does ‘childness’; all objects may be defined within a ‘littleness’ range, but not necessarily within a ‘childness’ range. A more primitive category serves well as a central category. Lakoff describes basic-­level status solely in terms of objects and recognises that the relatively subjective notion of ‘littleness’ is at the centre of this conceptual category of ‘diminutiveness’, in part because it has many of the characteristics and the attractiveness of basic-­level terms (fundamental in body–world connections). It is easy to use, it is the most contextually neutral term, and it is the first to enter most readers’ lexicons. Size as the core sense shows more evidence for the common tendency in semantic extension and change. dims’ radial category extends the central physical domain of size to the other non-­physical domains (a common trend in semantic associations) of gender, social power and others. This supports the evidence about the unidirectionality of semantic change from the physical to the social and conceptual domains discussed above. Jurafsky’s lambda-­abstraction examples also emerge from this core sense. For dims: ‘this process takes the original concept “small(x)”, which has the meaning “smaller than the prototypical exemplar x on the scale of size”, and lambda-­abstracting it to “lambda(y) (smaller than the prototypical exemplar x on the scale y)”’ (1996, 555). The pragmatic hedges and politeness-­marking functions he mentions are also based on this littleness concept. Furthermore, since metaphoric speech constitutes an essential element in the model ­discussed here, we need to be aware that the notion of size constitutes the base for many metaphors, since it has basic physiognomic and perceptual properties. Seitz (2001) points to key aspects of primary metaphors such as perceptual (e.g. size) and physiognomic (i.e. visual-­affective) experiences. Humans exploit perceptual features such as size when performing metaphoric thinking (Seitz 1997). Physiognomic perception has been studied well as the basis for metaphors; an extensive literature is reviewed in Seitz and Beilin (1987). Thus, size is a better base for metaphoric chaining (in our model of radial categories) than ­age-­related properties (childness) since it shows more basic perceptual primitives. Ultan (1970) and Lindauer (1988) connect size with symbolism and physiognomic stimuli. Interestingly, ‘littleness’ is the earliest attested meaning for dims. This notion dominates most links among senses (more than ‘childness’), even though it is not necessarily the most common sense in modern uses of dims. Because of all these arguments, the ‘littleness’ proposal supported here renders the radial category model shown in Figure 2.2. This model shows ‘littleness’ as the core sense of dims primarily and ‘childness’ secondarily; they are circled to signal this semantic core. All the other dim functions/senses grow out of this core sense via linking chains and constitute, in line with Dressler and Merlini Barbaresi’s (1994) position, pragmatics-­driven uses; i.e. all those outside the circled core. It may be argued then that evaluative morphology is mostly a pragmatic phenomenon. The actual linking or associations may be culture-­or language-­ specific, as are many linguistic phenomena (Sapir 1921), but the principles of metaphoric/ metonymic association and the connection to a central subcategory of [little] are more or less true across languages. Let us exemplify this metaphoric linking with the related notions of ‘child’, ‘little’, ‘dear’ and ‘not dear’. Again: ‘human[s] . . . have a natural suspicion of large creatures; small

The Semantics of Evaluative Morphology 27 [intense]

[dear]

[flirt]

2[child]

1[little]

[attenuate]

[irony]

[female]

[commiserate]

[euphemism]

[pejorative]

Figure 2.2  Radial model of diminutives (Prieto 2005)

animals and small children on the other hand can be cuddled and caressed without embarrassment or fear’ (Taylor 1995, 145). This connection between the [little] and [dear] (ameliorative) notions is thus ‘grounded in the co-­occurrence of elements within an experiential frame’ (p. 145). It is naturally embedded in humans’ perception and previous experience. Another possibility I suggest is of a more metaphorical nature. The notion of ‘littleness’ may have been transferred from the size plane to the distance plane: the more distance between two people, the less intimacy and affection between the two. Thus, if the distance (e.g. mother–child) is small, then it may reflect a high degree of affection or endearment; thus, there could be a metaphoric association between the [little] and [dear] meanings. Regarding the [child] and [little] connection, dims are normally associated with children because they are ‘little’. Thus, the connection between children and dims is a normal one observed cross-­linguistically (Jurafsky 1996; Melzi and King 2003). We typically observe in children two main features: littleness and endearment. Children are little and dear to us. According to the famous psychoanalyst Erikson (1950), generativity, embodied in the need to care for, raise and/or mentor offspring, is a crucial stage of the development of many living beings. This readiness to parent, Erikson asserts, may be viewed as naturally built in to our species. Thus, it is not absurd to think that this may be (at least) a reason for the common connection of dims with endearment. Children are little and children are dear; thus, this ‘littleness’ may be ‘dear’. This is founded on a basic logic syllogism of the type ‘if . . . then’ (Aristotle’s Prior Analytics in Smith 1989):

There are two premises and one conclusion: Premise A: If [+child] is [+little], AND . . . Premise B: If [+child] is [+dear], THEN . . . Conclusion: [+little] is [+dear] (henceforth, dim = little and/or dim = dear)

This, for some linguists, may be considered a type of semantic shift, where the word or morpheme takes on a new meaning, often related to the original one. With this approach, we can then conclude that even though there is nothing ‘little’ in the notion of ‘dear’, semantically, the common use of such affixes with the endearing function comes from the experiential association of children with features such as ‘littleness’ and ‘endearment’, logically. This is a metonymy, since there seems to be an association by context. It is empirically observable in many cultures and languages of the world that the language in children-­oriented environments is heavily characterised by such morphemes. This is probably why ‘little’ and ‘dear’ are dims meanings more commonly recorded. Let us now go to an opposite meaning of dims. Many dims actually mean [not dear] and [not little]. These are examples of irony or sarcasm, which represent the most difficult

Victor M. Prieto

28

connotations or uses to argue for in this semantic connection. It is a fact in linguistic data that dims often end up being pejs. Pharies (2002) thinks that this is the normal trajectory of Latin dims that end up becoming Spanish pejs (p. 423). According to Kruez (1996), the first and primary cues that may help signal irony are precisely the counterfactual ones. These two ideas, namely direct opposite and counterfactual, imply that when meant to be ironic, the dim actually means ‘big’ or more commonly ‘not dear or not appreciated’ or pej; the opposite of the common notions of dims (little, dear). It seems that through the agency of semantic polarisation, dims make a leap to the [−little] or [−dear] senses. Thus, semantic extension, association or metaphoric speech on the one hand and semantic polarisation on the other are links through which this polysemy spreads. This type of chaining may extend even further to social values and related concepts. For example, Prieto’s (2005) data shows that the idea of ‘littleness’ in dims may not be just physical but also in the value or appreciation of women in many modern societies. Prieto elaborates further and more clearly on this derogation-­of-­women aspect and its connection with dims. That is why Jurafsky connects the [female] meaning with the sense of [little]. Admittedly, even though the child–female connection may have cognitive and perceptual grounding, the little–female binding seems to be a more determining ­motivation today. 4.3  aug radial category Let us recall that one of Lakoff’s (1987) criteria for determining core semantics in radial category models is relation to other spatial particles or contrast sets. This criterion crucially influenced the choice of ‘littleness’ over ‘childness’ for the semantic core of dims. Due to consistency and system symmetry, then, we need to conclude that the best contrast set for the [little] dim is the [big] aug. The other core sense suggested in the literature (i.e. [child]) fails to contribute to this contrast and symmetry that the [little]–[big] contrast provides to the whole system of evals. In order for the [child] sense to account for this symmetry in the system, it would need to contrast to some abstract [−child] or [adult] sense, which has not been systematically formalised as the core sense for augs. If so, it would then create innumerable unfortunate and nonsensical interpretations. Finally, ‘big’ is the opposite extreme in the range of the primitive category of size, which, as argued before, constitutes the base for many metaphorical and inferential associations. In the same fashion as for dims, we can now see how diverse uses of augs should emerge from a common source: ‘bigness’. The following are aug functions analysed in Prieto (2005), for example: bigness, intensification, irony, pejoration, attenuation, euphemism, affection and flirtation. As a summary and introduction to this discussion, Figure 2.3 shows aug’s core sense and the semantic chains. This model is to be compared with the semantic mapping for dims above, since both follow similar principles of cores and chains. [big]

[intense]

[dear]

[flirt]

[irony] [PEJ] [attenuate]

Figure 2.3  Radial model of augmentatives (Prieto 2005)

[euphemism]

The Semantics of Evaluative Morphology 29 The intensifying sense of augs gave rise to two seemingly opposing aug functions: pej and [dear]. Both emerged precisely because of the ‘very’ meaning. As mentioned above, the quality expressed by the base (normally an adjective) augments when an aug attaches, whether it be a positive or a negative quality. This positive/negative dichotomy determines the affectionate or the pejorative function, respectively. For example, if an adjective such as ‘poor’ is augmentivised, then it may have pejorative functions because of the negativity (socially speaking) of being poor or, better, ‘very poor’. Grandi (2003b), from a dynamic typology perspective in Mediterranean languages, observed that augs express two conceptual categories: ‘big X’ and ‘one who is/makes/has X in an exaggerated way’. This notion of exaggeration may have caused pejorative connotations. On the other hand, if nouns such as ‘brother’ or ‘friend’ receive an aug without reference to size, then these can reflect affection, since they would imply something like ‘very much a brother’ or ‘very much a friend’, which are normally positive properties that inspire affection. One function of augs that seems contrary to the core sense of this suffix is that of the attenuating or diminution function: what can be labelled as the ‘diminutive aug’. There are two possible answers as to how this [attenuate/diminution] function of augs (observed in Prieto 2005 for Spanish augs: -­ón/-­azo) binds with the ‘bigness’ sense. One is that it emerges from a pej. Something that receives pejoration loses value. This value loss may have generated the interpretation that the base (for the aug attachment) loses some of its inherent value or identity. It reflects, then, an identity diminution process, which resembles the properties described above for pejorative dims. Finally, these cases of augs’ uses with ironic functions correspond to a simple antithetical process also discussed when considering the ironic force of dims. Again, the core sense for this aug function is the notion of ‘bigness’. The antithetical process goes from [+big] to [−big]. As with any other ironic process, the ironic uses of augs convey, in a sarcastic way, the idea opposite to that expressed by the linguistic form (i.e. [−big] or ‘not very’). These examples illustrate that there is no single fixed property that the dim or the aug assign to the nouns or words they suffix to, such as ‘car’ or ‘friend’. In order to understand what the speaker means, the listener must draw upon her encyclopedic knowledge relating to cars, friends, other objects, the world/society around, etc., and her knowledge relating to what it means to be ‘little’ or ‘big’. The listener then ‘constructs’ a meaning by ‘selecting’ one that is appropriate in the context of the utterance. In the case of the dim, for example, the listener needs to decide whether it is ‘dear’, or ‘little’, or pej, and so on.

5  evals’ meaning construction: conceptualisation Lastly, the fourth guiding principle for cognitive semantics is that language itself does not encode meaning. In our case, the actual suffixes (in Spanish, -­ito for dim, -­ote for aug; in English, -­ling for dim; etc.) do not encode meaning on their own. This means that -­ito means virtually nothing in itself. On this line of thought, we could even argue that the Spanish word for ‘little car’, ‘carrito’ (car-­dim), also in itself does not mean much. Instead, this is a prompt for the construction of the meaning of ‘small-­sized car’, or ‘a dear car’, or ‘a worthless car’, and so on, assuming its relevance for a communicative task. Thus, meaning is really constructed at the conceptual level. This is conceptualisation, a process in which linguistic forms such as evals are the ‘opening door’ for a ‘room’ full of conceptual operations (physical diminutivisation, endearment, pejoration, and so on in the case of dims) and the recruitment or eliciting of background knowledge. That is why

30

Victor M. Prieto

cognitive scientists such as Evans et al. (2007) argue that meaning is a ‘process rather than a discrete “thing” that can be “packaged” by language’ (p. 9). There seems to be an immediate connection of elements of the surroundings, our background knowledge, context elements and the like when we process language comprehension. Coulson (2006) affirms that we comprehend language by ‘binding elements of the discourse representation to frames in long-­term memory’ (p. 255). The space structuring cognitive model ‘attempts to characterize how speakers, given linguistic input in a particular context of use, exploit creative processes of meaning construction to form enriched meanings’ (Coulson 2006, 259). She explains that the theory of conceptual integration suggests that to comprehend language, we construct multiple ‘cognitive models and [establish] mappings based on analogy and identity in context.’ (p. 260). This model predicts that morphemes can evoke radically different information depending on the concepts activated by prior contexts, and it highlights ‘the import of imaginative processes of meaning construction’ to understand language (p. 260), which is exactly what happens with evals. Thus, meaning construction is creative and flexible (Coulson 2006). On this basis, we assert that evals are better viewed as prompting meaning construction processes, rather than solely having meaning on their own. They activate frames, promote mappings, and encourage the blending of information from different domains. For example, dims in many instances blend the more basic size domain with the social hierarchy domain, via metaphoric association. As Coulson asserts, ‘meaning construction processes are fundamentally imaginative’ (p. 263), as figurative language and fiction show. evals is another ­demonstration of that as seen in the radial category model shown here.

6 Conclusion Linguistic meaning is often constructed in a three-­stage sequential process, which could be understood as a lower-­level grounding process; not an importance hierarchy. Pragmatic meaning, contextual meaning and/or figurative language are grounded on semantic or propositional meaning. Semantic meaning is grounded on cognitive processes which normally have a neurological or biological grounding: Physiological grounding: anatomy (bodies, senses), biology (vision), neurology v Cognitive grounding: perceptual realities and mental processes v Semantic grounding: core propositional meaning; referents of linguistic forms v Pragmatic meaning: metaphors and similar processes in figurative language The major conclusion here is: humans are born with a physical body and have physical limitations such as size, as do other elements around them. This physical property turns out to be a crucial property of humans’ bodies and surrounding objects. Humans, then, out of their need to understand and organise their world and their need to survive, strive for size perception. Once size is perceived, objects are evaluated regarding size. These concepts, so important for humans, may get coded in the language in linguistic forms such as evals. Finally, this basic evaluation of size expands to a more abstract type of evaluation via analogies, metaphors and similar figurative associations. Via these metaphors, metonymies

The Semantics of Evaluative Morphology 31 and associations, ‘size’ turns out to be ‘importance’, ‘endearment’, ‘closeness’ and so on. After all, humans are not only a physical body, but a complex being with a mind, feelings, relations, needs and dreams (Ryle 1949; Murphy 2006). Entities are then evaluated regarding importance, endearment, and other related abstract parameters using evals metaphorically too.

3  Evaluative Morphology and Pragmatics Lavinia Merlini Barbaresi

The distinction between the semantic interpretation of a sentence and the pragmatic interpretation of an utterance needs to be recognized and respected. Bach (2001, 1164)

1 Introduction A large number of meanings and effects obtained by evaluative morphology in discourse can be explained only by recourse to a pragmatic interpretation. More precisely, my claim is that some productive morphological operations are autonomously responsible for effects that cannot be exhaustively explained solely by their semantics plus generic contextual conditioning. The recognition of a direct and mutual relevance of morphology and pragmatics is a relatively recent achievement, though, and still encounters difficulties. In this chapter, I will try to demonstrate how and where the two disciplines of morphology and pragmatics meet and can be profitably dealt with through a unified theoretical approach. Here follows a preliminary sample of the pragmatic meanings/effects that can be obtained by the application of various morphological rules. These are attitudinal/evaluative meanings involving speaker and addressee in their interpersonal relationships. The examples proposed, all certified, are from the web, from other publications and from a small personally collected corpus (PCC) from actual or TV talks, films and newspapers: (i) Emotion, tenderness (a mother to a child) – English: dim; reduplication (a) ho’s my lovely little girlie? (b) Who’s my boobsy-­woobsy? (PCC) (ii) Emotion, tenderness (a mother to a child) – Italian: aug Non tanta acquona fa venire un grosso pancione al mio bimbo ‘Not so much water-­aug, it gives a big tummy to my baby’ (my baby a big tummy) (Dressler and Merlini Barbaresi 1994, 439) (iii) Playful irony – French and Italian: dim Il confronto è tra le Sarkozettes e le Berlusconette ‘The comparison is between Sarkoz(y)-­dims and Berluscon(i)-­dims’ (CdS Magazine, 24 April 2008) (iv) Derogatory irony – English: dim Wifelet means little wife. It’s a nicer word than concubine or mistress. I tend to have one wifelet and Anna going at a single time . . . (Nieuwenhuis 1985, 233) (v) Euphemism – German: dim

Evaluative Morphology and Pragmatics 33 Er hat ein Gläschen über den Durst getrunken ‘He has just drunk one little glass too many’ (he is totally drunk) (dictionary.reverso.net/. . ./ein%20Gläschen%20 über). (vi) False modesty – Italian: dim Avrei anch’io una mia teorietta ‘I’d have me too a theory-­dim’ (I’d have a little theory of my own) (PCC). (vii) Ironical understatement – English: reduplication and dim I do foresee a teensy-­weensy snaggette. (Nieuwenhuis 1985, 233) (viii) Emotion, anger – English: reduplication, little-­construction Can you guys do me a teensy weensy little favor? . . . Oh, shut up! (www.fanfiction.net/Anime/Manga/Naruto) (ix) Emotion, pleasure – French: first syllable reduplication Quelle bonne sousoupe! ‘What a delicious soup’ (Fradin 2005, 174) (x) Hedged request – Spanish: dim Espera solo un minutito ‘just wait a minute-­dim’ (can you wait a little, please?) (http://foro.univision.com/t5/Escandalosas-­VIP/SECILLA-­DONDE-­ANDAS/ td-­p/298757284) (xi) Pleading – Spanish: dim Deme un pedacito de pan! ‘Give me a piece-­dim of bread!’ (Dressler and Merlini Barbaresi 1994, 244) (xii) Hedged request, jocularity – Italian: dim1 and dim2 Ci sarebbe anche da tenere il San Bernarduccio, il cagnetto ‘there should also be to take care of the Saint Bernard-­dim, the doggie’(you should also take care of . . .) (PCC) (xiii) Contempt – Italian: pej E’ il voto di una unversitucola ‘it’s a vote from a university-­pej’ (small, unimportant university) (PCC) (xiv) Sarcasm (political TV talk) – Italian: intf + dim Cosa fate voi nei vostri salott-­ar-­elli? ‘what do you do in your drawing room-­ intf-­dims?’ (. . . in your silly little parties?) (PCC) (xv) Sarcasm – Italian: dim (TV fiction: a police inspector to a thief) Il suo è un mestierino che rende! ‘yours is a job-­dim which rewards!’ (yours is quite a lucrative job) (PCC) Evaluative formations involved in the production of pragmatic meanings are here analysed and explained with recourse to the theoretical model of morphopragmatics, which will be illustrated in Section 2 below. The reference is mainly to European languages, such as Italian, a morphologically rich language, and, when possible, English, for a more general understanding. Other languages are touched upon for contrastive purposes, but with no claim to a typological approach. The central topic of my work is rather an inquiry into the exclusive relationship between morphology and pragmatics. The chapter is structured as follows: in this Introduction, I anticipate some points of discussion and give a sample of the pragmatic meanings that are involved in the strategic use of evaluative morphology; Section 2 is reserved for the presentation and discussion of a model, the theory of morphopragmatics, which, I believe, entertains a privileged relation with the property of evaluativity; in Section 3, I discuss the concept and the process of evaluation in the frame of this theory; Section 4 draws a summarising picture of the main morphological rules and mechanisms involved; and Section 5 offers some final remarks.

34

Lavinia Merlini Barbaresi

2 Morphopragmatics Morphopragmatics is a theoretical model that precisely focuses on the intersection between morphology and pragmatics. Dressler and Merlini Barbaresi elaborated it in various steps (1987; 1989a; 1992) and expanded it into a fully fledged theory in 1994, continued by the authors partly in collaboration (1999; 2001; forthcoming) and partly autonomously (Merlini Barbaresi 2001; 2002; 2004a; 2004b; forthcoming). Other early studies that contributed to extend the theory are Dressler and Kiefer (1990) on German and Hungarian excessives, and Kilani-­Schoch and Dressler (1999a) on the French -­o suffix. I refer to these studies for a detailed discussion of the bibliography concerning the emergence of a ­pragmatic awareness of pragmatic facts in morphology. In the present study, I further elaborate on the model. Morphopragmatics is an area where morphological rules directly interact with pragmatic conditions, such as speech situations, speech acts and interactional strategies. More precisely, it describes grammatical morphological operations that, given certain favourable contextual factors, are capable of systematically contributing autonomous pragmatic meanings to the speech act, that is, regular pragmatic changes that take place when moving from the input to the output of a morphological operation, be it derivational or inflectional. The morphological operation may be totally responsible for the added utterance meanings, with the word base being either neutral (dogg-­y) or contributory (dear-­ie, Spanish pequeñ-­ ito ‘small-­dim’) or even contrary (Italian gross-­ino ‘big-­dim’) to the effect pursued. The morpheme can be moved to other word bases in the utterance with no change in the overall pragmatic meaning (see the concept of ‘landing site’ in Section 2.3.3) 2.1  Demarcation of the area Morphopragmatics is distinguished from the fields of morphosemantics, lexical semantics of morphology, lexical pragmatics of morphology and pragmatics of syntactic constructions. In particular, pragmatic meanings/effects are often improperly conflated with the idiosyncratic pragmatic meaning of individual morphologically complex words, which are, in fact, the object of analysis of lexical pragmatics. This type of conflation may be incurred, for example, in such words as bun(n)-­y, a diminutive of Scottish dialectal bun (‘rabbit’s tail’), pet name for rabbit, or as tum(m)-­y, modified from stomach, where the pragmatic meanings (tenderness, playfulness) belong to the words themselves and the contextual area that they select (child environment) and not to the word-­formation operation (the added -­y diminutive). Similarly, in a slang term such as alk-­ie ‘an alcoholic’, the -­ie suffix added to the clipped base does not contribute to the derogatory meaning of the expression. It rather appears to be a mere formative or phonetic substance added. A second type of conflation may occur when the connotative meanings of morphologically complex words are foregrounded (for example, negative connotation in lexicalised derogatory star-­let, or Italian donn-­ina ‘woman-­dim’ but also ‘prostitute’). Such connotative meanings, in fact, are the realm of lexical semantics of morphology. Compare also the humorous negative connotation of words combined with a suffixoid (Mattiello 2008), such as -­holic, as in work-­o-­holic, choco-­holic, or such as -­gate, as in the family of words derived from Watergate, where -­gate is re-­analysed as a suffix and re-­used, for example, in White-­gate, Travel-­gate, Irpinia-­gate, Data-­gate and many others, all with the same negative meaning picked up from the first context of use (the Watergate political scandal).

Evaluative Morphology and Pragmatics 35 In our view and in contrast with Volek (1987), word connotations, intended as stable meanings picked up from contextual uses over time, are not seen to be relevant for a pragmatic account; they are part of the complex semantics of the word. 2.2  Connotative vs. pragmatic meanings Morphosemantic connotations are actually of great importance in a theoretical discussion of the functions of evaluatives. In the case of the diminutive, due to its ample polysemy, often paradoxically contradictory, identifying stable connotations is difficult. A good candidate would seem to be emotion/affectivity (Volek 1987; Wierzbicka 1999). But emotion is an unstable condition; it is optional and restricted to special contexts of use. Apart from the purely denotative diminutives, there are countless cases of diminutives (or evaluatives in general) exhibiting a pragmatic meaning devoid of any emotionality. This is the case, for example, of evaluatives involved in the modification of speech acts, for example hedging, as in Italian Mi presti qualche sold-­ar-­ello? ‘Can you lend me some money-­intf-­dim?’, or in Spanish Espera un minut-­ito solamente ‘just wait a minute-­dim’. Further evidence that the emotionalist hypothesis is not explanatory is the fact that precisely when emotions are very strong, for example intense love or grief, diminutives and augmentatives and also pejoratives are pragmatically inappropriate. This is due, of course, to their character of ‘fictiveness’ and ‘non-­seriousness’ (more in Sections 2.3.1.and 2.3.2), contrasting with the official formality of intense feelings. It would be socially unacceptable to condole with somebody in grief by saying ti sono vicina nel tuo dolor-­one/dolor-­accio (instead of grande dolore) ‘I feel very close to you in your grief-­aug/grief-­pej’ (instead of ‘great grief’). Emotion is also totally absent in meiosis and understatement, characterised by a sort of indifference, cool detachment, real or pretended, where the expression of emotion would be contradictory (for example, è stata una giornat-­ina niente male ‘it was a day-­dim not at all bad’ (not a bad day at all, was it?) said of an important, eventful day. Other popular hypotheses of fixed connotations indicate positive, ameliorating features, [graceful, attractive], for diminutives, and negative, pejorative features, [exaggerated, unpleasant], for augmentatives. This is actually often the case, but not necessarily so. An augmentative formation as in Italian mi guardava con i suoi occhi-­oni neri ‘she stared at me with her dark eye-­augs’ conveys a positive feeling, but referred to a different base, for example gambe ‘legs’, as in Italian ha due gamb-­one ‘she’s got two leg-­augs’ (she’s got huge legs), the augmentative contributes a negative connotation. On the other hand, a diminutive can convey unattractiveness, for example in English it’s only a short-­term boom-­let. Denotative meaning plus morphosemantic connotations, whether emotion or positive pleasantness or negative unpleasantness, cannot exhaustively explain the complex meaning structure of diminutive or augmentative formations. Morphopragmatics can fill in the gap, I think. Our configuration of a global meaning, although based on theoretically different premises, is congruent with major cognitivist approaches (see Prieto and Mutz, this volume). Following Wittgenstein, Dressler and I do not attribute meaning to semantics or to contextual pragmatics only, but see it as a global concept having both semantic and pragmatic invariant features. More precisely, we privilege a theoretical position whereby semantics is included in pragmatics (Morris 1938) in terms of meaning in context, that is, pragmatics is viewed as a superordinate of semantics. The priority of pragmatics over semantics is already implied in early pioneers of a pragmatic (ante litteram) view of evaluatives, Leo Spitzer (1921)

36

Lavinia Merlini Barbaresi

and Amado Alonso (1935). Moreover, thanks to recent studies in language acquisition (Savickienė and Dressler 2007b), it has been experimentally proved that children learn some pragmatic functions of evaluatives before being aware of their semantic meaning of smallness or bigness. Stating a priority of pragmatics may appear objectionable in relation to those languages, for example present-­day French (Dal 1997; Fradin 2005; Dressler 2010), the North Germanic languages, Finnish (but see Laalo 2001) and Estonian, which seem to use diminutives only semantically (see Olofsson’s chapter on Swedish in this voume). For example, according to Fradin (2005), the most productive French evaluative suffix, -­et, is not involved in pragmatic uses. But the preclusion might be limited to the grammatical means employed. From the surveys of many world languages in this volume, it is apparent that the grammar of evaluative morphology may differ greatly cross-­linguistically. Fradin (2005, 13) himself, on the other hand, suggests a typical French morphological formation, reduplication of the first syllable, in sou-­soupe, to render the pragmatic meaning of a German or Italian diminutive, such as Supp-­erl or zupp-­etta ‘soup-­dim’. The complex meaning structure of evaluatives is definable, I reiterate, in terms of morphosemantic denotation (dimensional smallness vs. bigness), morphosemantic ­connotations and morphopragmatic meanings characterised by fictive (see Section 2.3.1) evaluation and a context of lowered formality (non-­seriousness) (Section 2.3.2). 2.3  The tenets of the theory Dressler and I describe the field of morphopragmatics as the result of a diachronic and synchronic process of grammaticalisation of pragmatic phenomena. Specifically, we configure a level of morphologised pragmatics, where the general pragmatic meanings of morphological rules can be legitimately explained. As said above, a priority of our theoretical approach is to demonstrate the autonomy of morphological rules in conveying pragmatic meanings. It is essential, therefore, to recognise a clear separation between semantic and pragmatic meanings. 2.3.1 Fictiveness Our approach falsifies the reduction of morphopragmatics to being a mere result of general pragmatics applied to morphosemantic meaning. In contrast, we advocate a thesis whereby the denotative meaning is attributed to morphosemantics and the remainder of the meaning components to morphopragmatics. In addition to the basic semantic meaning [small], with its allosemes [unimportant] and [young] for diminutives and [big] for augmentatives, we propose for both an invariant, non-­semantic, still more basic pragmatic feature [fictive], which implies reference to the component of the speaker’s attitude in the speech event and which naturally inheres in and conforms to the fuzziness of subjective evaluations (more in Section 3). This pragmatically based feature is conceptualised as a departure from conventional, culturally accepted standards of meaning. A frame is generated of personalised values, where such standards move according to the speaker’s evaluation. For example, an evaluative remark such as Italian veramente, Giorgio è un po’ vecchi-­etto per lei ‘as a matter of fact, George is a little old-­dim for her’ (is a shade too old for her) would be a challenging assertion. The elusiveness of the speaker’s meaning, relative to the degree of departure from the standard for vecchio ‘old’ and relative to the intended pragmatic function attributed to the diminutive, whether attenuative or ironically intensifying (fuzzy ambiguity), would probably prompt a reacting question by an interlocutor (unaware of the person’s age) of the type cosa intendi per un po’ vecchi-­etto? ‘what do you mean by a little old-­dim?’

Evaluative Morphology and Pragmatics 37 Due to its fictiveness, the diminutive allows the speaker to stretch the evaluative range to a level of meaning that may fail to be immediately grasped and shared by the addressee. Evidence of the pragmatic nature of many of the meanings/effects obtained by diminutives and of their independence from semantics is the fact that many such effects can also be obtained by augmentatives, which share with diminutives the pragmatic feature [fictive], but certainly not the semantic meaning [small]. Augmentatives, that is, can be pragmatically synonymous with diminutives in many of their functions. In Italian, saying to somebody sei uno stupid-­ino ‘you’re an idiot-­dim’ would not sound dissimilar from sei uno stupid-­one ‘you’re an idiot-­aug’. In both cases, the suffix is capable of hedging the critical remark, via overtones of playfulness and tenderness added. Compare with the non-­ suffixed version sei uno stupido ‘you’re an idiot’, which would sound offensive. These pragmatic similarities between diminutive and augmentative may explain the diachronic process whereby the same suffix may have become an augmentative in Italian (-­one) and Spanish (-­on) but a diminutive in French (-­on). 2.3.2  Non-­seriousness In diminutives, the invariant morphopragmatic feature [fictive] is further specified as a character [non-­serious] (cf. Schneider’s interesting notion of sub-­normality: 2013, 144), which is at the basis of the majority of the diminutive meanings in discourse (for example, imprecision, attenuation, euphemism, but also irony, meiosis, ludic attitude and others.). The speaker evaluates the speech act and the speech situation as non-­serious, in the sense of non-­formal, non-­demanding, non-­binding, and this regulates his or her choice of an evaluative diminutive suffix as a structure conforming to his or her evaluation of the contextual premises. Its use is also judged by the speaker as downgrading his or her liability in case of some disagreement with the addressee. In our claim, non-­seriousness is the constitutive morphopragmatic feature of diminutives, but characterises augmentatives and even pejoratives as well, when these are used as pragmatic alternatives to diminutives. A well-­known example of a pejorative is Roberto Benigni’s addressing the Pope with an affectionate ah, Wojtyl-­accio! ‘Wojtyl(a)-­pej’ (ah, naughty Wojtyla!), a frequent use of the playful pejorative in the Tuscan variety of Italian; see also Gin-­ett-­accio, ‘Gino (Bartali, the famous champion)-­dim-­pej’ (naughty Gino!) Even in an apparently formal context, such as a bank deal, in the utterance adesso dovrebbe mettere una firm-­etta qui, ‘now you should put a signature-­dim here’ (just sign here), the diminutive indicates a downgraded formality of the interaction and lowers the load of the action requested. Interestingly, Robin Lakoff noted that the pragmatic diminutive can be used in many situations in which the speaker desires to minimise the impact of a statement (reported in Jurafsky 1996, 558), which is in line with her politeness sub-­maxim ‘don’t impose’. The pragmatic character [non-­serious] is linked to the semantic feature [unimportant]; nonetheless, the two types of meanings can be clearly distinguished, for example in understatements such as Italian eh, è una somm-­etta non da poco ‘well, it’s a non-­negligible sum-­dim’, where the speaker apparently contradicts the actual importance of the referent, the sum of money (litotes: non da poco ‘non-­negligible’ = ‘quite significant’) by using a diminutive. Through the diminutive, the speaker minimises the importance of the referent in contrast with the intended evaluation, but, by exploiting its pragmatic feature [non-­ serious], he actually contributes to the meiosis. A non-­serious feature added is a strategy for lowering one’s commitment to the illocutionary force. For example, a student involved in a discussion with her professor expresses

38

Lavinia Merlini Barbaresi

her disagreement by cautiously saying: (Italian) io facevo un ragionamento un po’ divers-­ ino ‘ I was having a thought a bit different-­dim’ (my line of reasoning was slightly different). The diminutive makes the student’s assertion less categorical than it would be with the simplex diverso and would allow her to withdraw easily if her opposition should become a bit ‘undiplomatic’. The student’s evaluation does not conform to clear standards and this makes her position more easily negotiable. 2.3.3  ‘Landing site’ The meaning obtained through a morphopragmatic operation extends to the entire speech act (see the notion of ‘sentence diminutive’ in Spitzer 1921), independently of the word base made relevant for the operation, in the model called ‘landing site’. A combination of sole denotative and connotative meanings, instead, would only have scope on the word base and would be unfit for characterising the communicative situation pragmatically. See, for example, Italian è una strad-­ina di campagna senza traffico ‘it’s a country lane-­dim with no traffic’, where the diminutive conveys smallness and, at most, some connotations of peace and pleasure, but does not induce any pragmatic reading. The choice of the landing site mainly depends on the lexical properties of the given word and on its pragmatic salience, with a preference for head nouns, but, often, more than one location in the sentence would be suited to this function. For example, an utterance such as Italian è una rispost-­ ina permalosa ‘it’s a resentful answer-­dim’ could be easily replaced with è una risposta permalos-­etta it’s a resentful-­dim answer’. Sometimes, suffixation could involve more than one word in the same utterance (rispost-­ina permalos-­etta ‘answer-­dim1 resentful-­dim2’), with intensified effects. This double diminution may even obtain exaggerated effects and suit, for example, an ironic environment, as in Italian manager-­ini giovani e modern-­elli (Corriere della Sera, 22 July 2010) ‘manager-­dim1s, young and modern-­dim2’.

3  Evaluation and fictiveness Evaluation, here intended as the speaker’s general pragmatic evaluative attitude, is a complex and multifaceted phenomenon. Through a diminutive, augmentative or pejorative formation, the speaker expresses a judgement ‘as to value’ (not ‘as to fact’), according to his or her intentions, perspectives and standards of evaluation. An evaluation, that is, is a mental operation assessing the value of an object or event, as more or less desirable and important in the interpreter’s view. Evaluations are inherently subjective as they express the speaker’s attitude towards the object or event. As we have said, the attitude may, but need not, be emotive. An expressed emotion, anyway, is always evaluative, as it indicates a type of stance taking. The act of evaluation is normally followed by an explicit or implicit act of approval or disapproval by either the speaker or the addressee. As reported in Caffi and Janney (1994, 328), as early as the beginning of the twentieth century, Marty (1908) distinguished between ‘emotive communication’, that is, the intentional strategic signalling of affective information (for example, evaluative dispositions, evidential commitments, volitional stances, relational orientations, degrees of emphasis), and ‘emotional communication’, which he regarded as a type of spontaneous, unintentional leakage or bursting out of emotion in speech. Evaluation would seem to belong to the first configuration, which entails volition and consciousness. Still, our evaluative formations are more elusive than evidentials and much less explicit than the language of stance taking. Seen from a morphopragmatic perspective, by using an evaluative, the speaker, as I have said, introduces an element of fictiveness in his or her evaluation, by which she or he indicates

Evaluative Morphology and Pragmatics 39 that the norms of the real world are suspended. Because of their fictiveness, evaluations are often perceived as fuzzy by the interlocutors and may become the object of negotiation more often than explicit evaluative assertions do (see vecchi-­etto in Section 2.3.1 above). Also compare the following two texts: Italian (a) cammini come una lumaca ‘you walk like a snail’ vs. (b) lumach-­ina! ‘you snail-­dim!’. In (a) the speaker makes an explicit, intentional evaluation expressing disapproval, whereas in (b) the speaker’s act of disapproval is obtained through the negative connotation of the metaphorical epithet of the base, mitigated by a mixture of affects, mainly jocularity and tender indulgence conveyed by the diminutive. How intentional, how conscious the evaluation is, it is difficult to say. The speech act of disapproval in (b), in fact, sounds like a spontaneous outburst of jocular impatience rather than a planned critical remark. Here, the evaluative suffix confers some characteristics that make the utterance also congruent with Marty’s ‘emotional communication’. Evaluatives, that is, would seem to stand in between Marty’s two configurations, as they share characteristics of both. 3.1  The evaluative process The first step of the speaker’s complex evaluative process is self-­directed and refers to his or her affective and attitudinal dispositions. These consist of continua of implied personal feelings, such as interest–disinterest, pleasure–displeasure, attachment–detachment, approval–disapproval, whose different degrees and combinations regulate the evaluative orientation, both axiologically and in terms of the affective intensity of the utterance. A second step of the speaker’s evaluation is directed at the social context of use and its relevant properties. Contextual variables too may consist of continua; for example, ­formality–informality, participant closeness–distance; but they may also refer to specific circumstances, such as nursery vs. adult and female vs. male environments, or doctor–patient or professor–student interactions, and so on. The speaker will judge contextual conditions more or less favourable for the use of evaluative suffixes. More specifically, she or he may judge the use of the suffix as more or less strategic for the purpose of his or her speech act. Expectedly, a crucial part of the evaluative process is centred on the relation between speaker and referent, the object of evaluation, and the reaction that this evaluative relation triggers. The importance of the speaker–object relationship was observed and developed by various scholars. For example, Dal’s (1997) notion of the diminutive (in -­et) as a marqueur d’appropriation is pertinent. It indicates a type of proximity between locutor and object established by the application of the rule. In Fradin’s (2005) diminutive meaning trichotomy, reference is to the sphere of the ‘locutor’s pole’ (vs. referent’s and vs. interlocutor’s). In these authors’ frameworks, the meaning of the suffix -­et at this pole is configured in semantic levels, but some of the meanings of the diminutives in -­et are not simply derivable from the denotative meaning of smallness; they are clearly at least connotative and often are capable of pragmatic functions, such as empathy and hypocorism (cf. Fradin 2005, 8, table 1). An interesting position, conflating the axiological with the size evaluation, is Inchaurralde’s (1997). He analyses the Spanish diminutives in terms of emotional proximity vs. remoteness between speaker and external entity. More precisely, in line with Ruiz de Mendoza (2000), he conceives of a diminutive effect as reducing the psychological distance from the entity and allowing the speaker to win control over it. Opening one’s personal space to a diminutivised entity is safe. And what is inside one’s space is perceived as good. By contrast, an augmentative effect may appear threatening and the entity will be kept outside of one’s personal space. These conceptions are congruent with viewing the

Lavinia Merlini Barbaresi

40

evaluative process as a private act, based on personal intentions, perspectives and standards of evaluation. At this stage, evaluation is still a cognitive operation and is connected with the speaker’s perception of the object. It acquires pragmatic relevance when it is translated into a speech act involving an addressee (cf. Fradin’s interlocutor’s pole). The evaluation then becomes a social act which triggers perlocutionary reactions. Morphopragmatic meanings are activated only with the addressee’s interpreting contribution. This emphasis on the relation to the addressee is clearly more relevant for the understanding of morphopragmatic phenomena than is Reynoso Noverón’s (2005) focus on the subjective attitudes of the speaker.

4  Morphological rules and mechanisms In a cross-­linguistic perspective, the expression of evaluative meanings can be achieved by various means, singularly or in various combinations: for example, at the level of phonology, with consonant, vowel or tone alternations and expressive palatalisation, or at the level of morphology, with suffixes, prefixes, semi-­prefixes, interfixes and more rarely infixes and circumfixes, often accompanied by gender and number changes. Extra-­ grammatical mechanisms (Dressler and Merlini Barbaresi 1994; Doleschal and Thornton 2000; Merlini Barbaresi 2004a; 2004b; Mattiello 2013), may also be involved, such as clipping and blending and reduplication. More rarely, compounding, particles and clitics may have an evaluative function. Syntactic patterns, with evaluative adjective or adverbial constructions, may add to or interfere with evaluations obtained by single constituents of the text. Prosody (e.g. in European languages of reference), combined with evaluatives, may clarify and (de)intensify meanings, but require autonomous experimental studies (Gili Fivela 2009). A good number of the evaluative means described in this list are exemplified well in the individual extra-­European languages analysed in this volume, but their pertinence to a pragmatic interpretation is difficult to establish, because of insufficient contextualisation and pragmatic explanations. Let us summarise the most important mechanisms employed in the European languages of reference:



(i) s uffixes: diminutives, augmentatives and pejoratives, as in Spanish hasta lueg-­ito, ‘good-­bye-­dim’, German Papp-­erl ‘meal-­dim’, Portuguese animal-­aço ‘animal-­aug’, Italian grass-­one ‘fat-­aug’ and Italian om-­accio man-­pej’ (ii) prefixes: as in English ironical (a couple of) mini-­ideas, French ultra rapide, Italian recent stra-­bello ‘very pretty’ (iii) interfixes: as in Italian ante-­suffixal -­ol-­ in ludic top-­ol-­one vs. serious top-­one ‘sewer rat-­aug’, or -­er, as in pazz-­er-­ell-­one ‘crazy-­intf-­dim-­aug’ (iv) reduplicatives: as in English Lizzy-­wizzy, teensy-­weensy, French joujou ‘toy’, Zizou for Zidane, sou-­soupe (‘soup’); see also Yiddish shm-­reduplicatives indicating irony, derision or scepticism, as in He’s just a baby! Baby-­shmaby. He’s already 5 years old! (v) clipped forms and blends: as in Italian clipped bibe(ron), English blend digiteria ← digital cafeteria, French clipped intell-­o ← intellectuel, Spanish blend telebobela ‘silly soap opera’ ← telenovela ‘soap opera’ plus boba ‘silly’ (vi) patterns in discourse involving word-­formation operations in speech acts (for example, diminutive in attenuated requests and mitigated orders, augmentative and elative in evaluative assertions or in rebuttals)

Evaluative Morphology and Pragmatics 41 (vii) f ormatives, prefixoids and suffixoids: as in Italian -­poli ← Tangent-­o-­poli (a financial scandal) in vall-­ett-­o-­poli ‘irregular recruitment of TV vallette (starlets)’; English -­gate ← Watergate in Enron-­gate and others (viii) analytic forms: English reduced, unstressed little or French petit →ti in Québec may express similar meanings, such as sarcasm, as in English I can’t stand your little tricks, or tenderness, as in French ti-­Jean. Also metaphorical compounds (for example, English giant-­killer, baby-­dolphin, baby-­trees, pico-­brain, über-­ brain) are often pragmatically exploited; for example, for tenderness in baby-­ dolphin, exaggeration and irony in pico-­brain, über-­brain.



4.1 Diminutives The privileged objects of a morphopragmatic description are evaluative suffixes, prototypically diminutives. According to Prieto (2005, 110), over 78 per cent of his large corpus of Spanish diminutives are used pragmatically. Diminutives are the unmarked evaluative category; that is, the presence of a productive category of augmentatives in a language implies the presence of diminutives (Dressler and Merlini Barbaresi 1994; Grandi 2011b). Diminutives are almost universally represented cross-­linguistically, or, at least, languages minimally possess the subcategory of hypocoristics, which are productively used in child/ pet/lover-­directed types of discourse. The closeness between the categories of hypocoristics and diminutives is also proved by the fact that there is a large overlap, cross-­linguistically, between their respective markers. Diminutives are the morphological mechanism that best exemplifies the variety of relevant morphopragmatic meanings. They allow the largest number of strategic uses in speech acts and speech situations, in which they obtain a range of predictable pragmatic effects, given certain circumstances (see the list of examples in Section 1). Not only emotionality in child/pet/lover-­centred speech situations, but also the ludic character of playfulness among intimates, familiarity and informality in general, sympathy and empathy and also understatement, euphemism, false modesty, irony and sarcasm, are the circumstantial factors that favour a pragmatic use of diminutives. These factors are not constitutive; they are regulative. Diminutives are also effective devices for politely hedging or downgrading requests, mitigating assertions, and increasing attractiveness in offers and invitations. 4.2 Augmentatives The pertinence of augmentatives to pragmatics is less direct and exclusive, in the sense that, at most, they confer on their bases a combination of semantic and pragmatic meaning (Merlini Barbaresi 2004a, 288). When pragmatics is predominant, both augmentatives and pejoratives may be very close to diminutives in their effects. That is, they may actually be alternative marks for signalling morphopragmatic meanings (Sections 2.3.1 and 2.3.2). In Italian, for example, and also in Spanish (Prieto 2005), augmentatives or pejoratives can convey – in addition to their semantic meanings [big] and [bad], respectively, and their intensifying effect – tenderness and jocular closeness, as in Italian il mio fratell-­one! ‘my brother-­aug’ (cf. Spanish herman-­azo!) and il mio ragazz-­accio! ‘my boy-­pej’.

42

Lavinia Merlini Barbaresi

5 Conclusion In this chapter, I have tried to be fairly wide-­ranging in discussing data and theoretical implications; nonetheless, more could have been done. After presenting and elaborating on a strong theoretical model, I would have liked to try extra validation, taking advantage of the wonderful wealth of languages analysed in this volume; but, partly because of limited space, hardly sufficient for the exposition and elaboration of the model, and partly because of the insufficient contextualisation of the data reported, I found the task too difficult and the results unsatisfactory. I am confident, though, that the model and methodology I propose can actually be ‘exported’ to other language investigations, although, elsewhere, they may yield different results in terms of meanings and formal means employed.

4  Word-­Formation Processes in Evaluative Morphology Pavol Štekauer

1 Introduction This chapter aims to provide a brief typological overview of the processes that are employed to form morphologically complex words with evaluative meaning. Since the assessment of the degree of utilisation of these processes from a cross-­linguistic perspective depends heavily on the definition of the scope of evaluative morphology (an issue on which there is no agreement among morphologists), Section 1.1 provides a sort of ‘broad definition’ of evaluative morphology as a point of departure for the typological overview presented in Section 2. Section 1.2 shows that the scope of this chapter necessarily ­overreaches the limits traditionally imposed on the field of word formation. 1.1  Scope of evaluative morphology Any discussion of word formation (WF) within evaluative morphology is preconditioned by the delineation of the scope of evaluative morphology, because it predetermines any considerations of the productivity as well as frequency of WF processes in individual languages. This question has been subject to lively debates (Scalise 1984; Stump 1993; Bauer 1997; 2004; Grandi 2005; 2011a; 2011b; Štekauer, Valera and Körtvélyessy 2012; Körtvélyessy 2012b; Grandi and Körtvélyessy, this volume). The basic reference point for my approach is the category of Quantity, including various seemingly unrelated phenomena, which, however, have a common basis: all of them can be classified as ­evaluative categories in terms of the deviation from the default value. The quantity-­founded deviations from the default value run across the fundamental cognitive categories of substance, action, quality and circumstance. By implication, it is possible to distinguish evaluatives in terms of the Quantity of substance, the Quantity of action, the Quantity of quality, and the Quantity of circumstance. If this framework of evaluative morphology is adhered to consistently, evaluative morphology encompasses, in addition to the ‘traditional’ cases of evaluative morphology, some categories of Aktionsart (frequentativeness, iterativity, intensity and distribution as well as pluractionality), attenuatives and morphologically realised forms of comparatives and superlatives; they obviously represent deviations from default values. The fact that they have mostly been ignored in research into evaluative morphology may be attributed to the persisting tradition of keeping inflectional and derivational processes apart. As stated elsewhere (e.g. Štekauer, 2014; ten Hacken, 2014; Booij 2006; Plank 1994; Scalise 1988a) cross-­linguistic data teaches us that the relation between inflection and derivation is that

44

Pavol Štekauer

of a cline. Katamba (1993, 212) has put it pregnantly when stating that one and the same category, including diminutives, may be inflectional in one language but derivational in another. Consequently, there are various inflectional processes that yield morphological evaluatives, including the comparative and the superlative degrees. Just for illustration, if Slovak domisko ‘large house’, derived from dom ‘house’, evaluates this object as quantitative deviation from the default size of house within the category of substance, then, obviously, the comparative degree of Slovak väčší ‘larger’, derived from veľký ‘large’, evaluates the size as quantitative deviation from the default size within the category of quality. Pluractional verbs express an event performed several times by a singular subject in the case of intransitive verbs and/or an event affecting multiple objects or one object multiple times; an obvious deviation from the default value of a single event pertaining to a single object: (1) Komi uj-­avaɫ/v-­ pla to.swim-­ ‘to swim (for a long time or about several swimmers)’ (Fejes, forthcoming) Sara Mbay tɪ́-­ndá pla-­hit ‘hit many times’ (Dimmendaal 2014, 595) Attenuatives are characteristic of colour terms in many languages: (2) Kazakh aq-­ša white-­ att ‘whitìsh’ (Nikolaeva 2014, 499) In Kanuri, attenuatives are the only semantic category of diminutiveness: (3) sə́lə́m-­àràm black-­ att ‘blackish’ (Hutchinson 1981, 67) Certainly, attenuatives are not restricted to the field of colours: (4) Nenets wewa-­ko silly-­ dim ‘somewhat silly’ (Wagner-­Nagy, forthcoming) Slavic and some other European languages express attenuation on verbs – in principle, by suffixation: (5) Kashubian rób-­k-­ac work-­ dim-­inf ‘to do, to work a little’ (Makuraté and Pazdjerski, p.c.) The inclusion of some Aktionsart categories in the field of evaluative morphology has significant consequences for the discussion of WF processes, because there are some

Word-­Formation Processes in Evaluative Morphology 45 l­anguages in which an Aktionsart category is the only evaluative category. This can be illustrated with an example from Nipmuck (Algonquian) where the only augmentative meaning is a distributive one (Gustafson 2000, 128): (6) pisch-­chekau peas-­ distr ‘there is an abundance of peas’ (Gustafson 2000, 129) Similarly, in Kwaza, the only augmentative morpheme is the intensifying classifier (van der Voort 2004, 461): (7)

awe nỹ-­mũ-­̕te-­ki rain big-­clf:liquid-­aug-­decl ‘it is raining heavily’

In Ladin, the only type of evaluation is the intensifying one – formed by two different prefixes, stra-­ and s-­. There are no other diminutives or augmentatives in Ladin. (8)

stra-­mudé aug-­change ‘to dazzle’ (Siller-­Runggaldier, forthcoming)

Intensification appears to have a special status in evaluative morphology because it can go in both directions from the default value, i.e. it can intensify towards smaller as well as larger quantity than the default, depending on the semantics of the basic word. In Tat, it is done by means of partial preposing reduplication: (9)

jem-­johil ints~young ‘very young’ (Authier, forthcoming)

In Maya, the only type of augmentativeness is iterativeness/frequentativeness of verbs by reduplication: (10) tin la-­k-­ets ‘I strike you with the palm’ tin la~tla-­k-­ets ‘I strike you several times with the palm’ (Tozzer 1921, 91) All in all, evaluative morphology includes not only the traditional semantic categories of diminutiveness and augmentativeness but any and all morphological expressions meeting the above-­mentioned condition of deviation from the default value.1 1.2  Evaluative formation It is generally admitted that the borderline between the derivational and the inflectional morphologies is a fuzzy one with prototypical, less typical and peripheral instances on both sides. It has also been proposed (Scalise 1984) that evaluative morphology has, at least in some languages, a special position; a fact that underlies the label ‘third morphology’ as an area of morphology situated somewhere between the two ‘classical’ morphologies. A

46

Pavol Štekauer

fairly common situation is the one occurring in Altaic languages where evaluative affixes are typically external to other derivational suffixes and internal to inflectional morphemes (Nikolaeva, forthcoming). These observations indicate that the range of processes that participate in the formation of new, morphologically complex words with evaluative meaning (henceforth ‘evaluatives’) should not and cannot be restricted to the ‘classical’ WF processes on the derivational side of the morphological cline. My sample of over 200 languages of the world provides numerous examples of a partly or fully inflectional essence of some processes underlying the formation of evaluatives. It is for this reason that, throughout this chapter, I replace the term ‘word formation’ with the term ‘evaluative formation’, meaning the formation of complex words with evaluative meaning as defined in Section 1.1 above. Since the majority of examples in Section 2 come from the derivational side of evaluative formation, Sections 1.2.1–1.2.4 illustrate the inflectional side of evaluative formation. 1.2.1  Change of inflection class The change of inflectional class, or the reservation of certain inflection classes for marking diminutives and augmentatives, respectively, appears to be the most typical and a relatively common ‘inflectional’ technique of evaluative formation: (11) Jóola Banjal e-­joba ‘dog’  pl. su-­joba (dim) ju-­joba ‘small dog’  pl. mu-­joba (aug) ga-­joba ‘big dog’ pl. ba-­joba (Creissels 2014, 571) In Aghem some nouns can be transferred to gender 11/12 in order to derive a diminutive meaning (Hyman 1979, 24). In Supyire, gender 3 is the gender of small things, and moving a root into gender 3 may have the effect of allowing it to denote a smaller object than in another gender (Carlson 1994, 105). 1.2.2  Classifiers My sample offers a few examples of employing classifiers to express evaluative meanings. For illustration, Movima makes use of the classifier -­di, whose central meaning is ‘seed, grain’: (12) awaro-­di parrot-­ dim ‘little parrot’ (Haude, p.c.). 1.2.3  Evaluative formation by clitics However rare, there do exist instances of evaluative formation by clitics. In Eton, for example, ‘the diminutive proclitic N-­ ɔ̀H-­(plural b-­ɔ̀H-­) can be put in front of any full noun in order to form a diminutive or singulative’ (Van de Velde 2008, 207): (13) Eton a. mɔ̀-­ílé dim-­tree ‘small tree’ b. bɔ̀ bílé dim-­trees ‘small trees’ < bìlé ‘trees’2 (van de Velde 2008, 207)

Word-­Formation Processes in Evaluative Morphology 47 1.2.4  Level-­ordering hypothesis There are languages in which evaluative formation violates the principles of the level-­ ordering hypothesis, in particular the rule that inflection is preceded by derivation. Thus, in Passamaquoddy, diminutive stems of one type of transitive verbs are formed from partly inflected basic stems. The Passamaquoddy inflectional suffix in such cases is repeated after the diminutive suffix whenever inflectional morphology calls for its use. Thus, for example, diminutive and plural affixes can interact and be repeated in a single stem in several possible orders. The diminutive stems are derived from plural stems, and plurals from diminutives (LeSourd 1995, 104): (14)

megghci-­ne-­he-­ə-­ss-­ohtò-­kk end-­die-­dim-­ai-­dim-­(ai) .pl-­(3)-­33an-­absn ‘the poor little ones are dead’ (LeSourd 1995, 126)

In Luxembourgish (IE), the diminutive form is formed from the plural stem: (15)

Faass ‘barrel’ – Fässchen (pl.) – Fässercher ‘keg’ (Gerald Newton, p.c.)

In Welsh, too, some diminutives can only be formed from the plural: (16)

plentyn ‘child’ – plant ‘children’ – plantos ‘little children’ (Awbery, p.c.)3

2  Typology of evaluative formation This section maps various morphological processes that take part in evaluative formation as defined in Section 1 above. The aim is to provide a concise typological overview that indicates the main tendencies as well as intriguing peculiarities of evaluative formation. It should be stressed that the extent of the employment of individual evaluative-­ formation processes in this area of morphology does not correspond to the general tendencies of WF in the languages of the world. According to the data in Štekauer et al. (2012), the most productive WF processes are suffixation (95 per cent of the languages of the world) and compounding (90 per cent) followed by reduplication (80 per cent), prefixation (72per cent) and conversion (63 per cent). While, analogically to this general situation in WF, evaluative morphology is dominated by suffixation, compounding is much less common, and conversion appears to be totally absent. Nonetheless, the range of evaluative-­ formation processes is broad in both directions from the default value. 2.1  Affixation Affixation is by far the most common strategy employed in evaluative morphology. This is thanks to suffixation, which occurs in the vast majority of the sample languages (see Section 1 for some examples). Prefixation seems to be less common and infixation is even less so. If the SAE territory is not taken into account, prefixation in my sample only occurs in Krongo and Anejom for diminutives, and in Micmac and Mosetén for augmentatives. But the capacity of prefixation has been significantly boosted recently due to the process of globalisation and the accompanying dissemination across languages of the prefixes hyper-­, super-­, maxi-­, etc. In Anejom, the prefix nalve-­forms diminutive nouns:

Pavol Štekauer

48 (17)

nalve-­eañ dim-­coconut ‘kind of coconut with small fruit’ (John Lynch, p.c.)

In Micmac, intensives and frequentatives are formed by means of the prefixes i’-­and kesi-­: (18) ketapekiey ‘I sing’ i′-­ktapekiey ‘I often sing’ kesinukway ‘I am sick’ kesi-­ksinukway ‘I am very sick’ (Hewson and Francis 1990, 180) An example of the formation of augmentatives by prefixation is taken from Scottish Gaelic: (19)



mòr-­chuairt big-­trip ‘excursion’ (William Lamb, p.c.)

Albanian provides an example of forming augmentatives of different word classes with the same prefix: (20) a. stër-­nxeh aug-­make.hot ‘to overheat’ b. stër-­madh aug-­big/large ‘enormous’ (Genesin and Matzinger, forthcoming) Since infixation is not a ‘natural’ WF process it is neither widespread nor productive in WF, and even less so in evaluative formation. There are very few languages outside the SAE territory that can illustrate this strategy of forming evaluatives. One of them is Dangaléat: (21) atype átyípépé hit ‘hit (several times)’ (Shay, p.c.) A unique situation occurs in Pingding Mandarin (Sino-­Tibetan) in which the only affix appears to be the diminutive/hypocoristic infix : (22)

xɭua flower-­dim ‘little flower’ (Blevins 2014, 144)

Blevins (2014, 145) maintains that reduplicated infixes show more frequent associations with pluractional, intensive, repetitive, iterative, distributive and augmentative/diminutive meanings than do non-­reduplicated infixes. In the SAE territory, one of the languages that make use of infixation is Spanish:

Word-­Formation Processes in Evaluative Morphology 49 (23) Spanish lejos farsfx-­adv ‘very far’ (Kornfeld, p.c.) 2.2  Prefixal–suffixal derivation This derivational process of parallel attachment to the WF base of prefix and suffix is not widespread in the world, but it does occur with verbs in a number of Slavic languages, as in (24): (24) Czech na-­modr-­al-­ý prfx-­blue-­sfx-­infl ‘partly blue’ One more example of this rare evaluative-­formation process comes from Ilocano: (25)

nag-­dakkel-­en aug-­big-­aug ‘how big!’ (Carl Rubino, p.c.)

2.3  Circumfixation Circumfixation, like any other morpheme-­splitting mechanism, is not good from the perspective of Natural Morphology. This is confirmed in WF in general and in EM in particular by the rarity of such formations. Our examples come from Kabardian (26) and Muna (27): xwa-­dajla:-­fa

(26)

cf-­foolish-­cf

(27)

sa-­wanu-­ha-­no cf-­get.up-­cf-­his ‘he can barely get up’ (van den Berg 1989, 314)

‘somewhat foolish’ (Matasović, forthcoming)

2.4 Reduplication As is generally known, reduplication can be full or partial. Both types are fairly widespread and productive in evaluative formation, and cover a whole range of semantic functions. Example (28) illustrates full reduplication with intensifying meaning: (28) Komi ɨɫɨn~ɨɫɨn far-­ rdp ‘very far’ (Fejes, forthcoming) Example (29a) illustrates partial (preposing) reduplication with the same meaning, and (29b) partial preposing reduplication with diminutivising meaning:

Pavol Štekauer

50 (29) a. Turkish b. Lushootseed

hıp~hızlı rdp-­fast ‘very fast’ (Nihan Ketrez, p.c.) ǰí~ǰəsəd rdp-­foot ‘little foot’ (Inkelas 2014, 185)

Some languages make use of both types of reduplication, as is the case of Arabana (Hercus 1994, 98): (30) a. karla~karla rdp creek-­ ‘small creek’ b. kalti~kaltiRi rdp-­hill ‘small hill, mound’ Example (31) illustrates partial preposing reduplication to express augmentativeness: (31) Gagauz kos~koca rdp-­big ‘huge’ (Menz 2014, 539) Reduplication is intriguing, inter alia, in its capacity to express both iconic and non-­iconic meanings. Iconically, reduplication is an augmentivising process because adding more form parallels a higher quantity of a substance, action, quality or circumstance, as illustrated in (32) from Alves (2014, 539): (32) Khasi khruj-­ruj-­ruj in.many-­ rdp-­rdp ‘in many great numbers’ Mang dɔ1~dɔ1 much-­ rdp ‘very much’ Mnong par~par fly-­rdp ‘fly repeatedly < /par/ ‘to fly’ Anti-­iconically, the addition of form is accompanied by reduction of quantity: (33) a. Anamuxra xya~xya white-­ rdp ‘whitish’ (Ingram 2001, 161) b. Bezhta yix~mixxo warm-­ rdp ‘half warm (not very warm)’ (Khalilov and Khalilova, forthcoming)

Word-­Formation Processes in Evaluative Morphology 51 Sometimes both iconic and anti-­iconic meanings of reduplication occur in the same language: (34) Koiari a. dubuka~dubuka black-­ rdp ‘blackish’ b. aheka aheka light light ‘very light’ (Dutton 1996, 43) Last but not least, let us mention the type which, for English, was examined by Hohenhaus (2004) under the label of identical constituent compounding. Hohenhaus concluded that the common semantic pattern for identical compounds in English is ‘an XX is a proper/ prototypical X’ for nouns and ‘XX = really/properly/extremely X’ for adjectives, adverbs and verbs. Thus, love-­love means ‘true love’ and home-­home ‘proper home’. This process with its specific semantics is not restricted to English, as demonstrated in Štekauer et al. (2012). 2.5  Prefixation of a reduplicated base Another rare process is a combination of a prefix with a reduplicated base, as instantiated by Muna. Van den Berg (1989, 295) identifies three variants of the process: (35)

Prefix -­ka-­ + reduplication: (a) simple noun base ka-­kontu-­kontu ‘small stone’< kontu ‘stone’ (b) reduplicated base ka-­kuhu-­kuhuti ‘small owl’ < kuhu-­kuhuti ‘owl’ (c) bound root ka-­bua-­bua ‘girl (age 12–15)’

2.6 Compounding While compounding appears to be – together with suffixation – the most widespread WF process, this does not apply to evaluative formation. In many languages its applicability is restricted and is bound to combinations with a special word. So, for example, in Bahasa Indonesia, for the young of animals the word anak ‘child’ is placed before the name of the animal (Müller, p.c.): (36) anak ayam dim-­hen ‘chicken’ In Malayalam, compounding relies on the word kuʈʈi ‘young’ (Mohanan, p.c.): (37) paʈʈik-­kuʈʈi dog-­ dim ‘puppy’ Nelemwa uses the noun axomoo-­‘mother’ for augmentatives of animals:

Pavol Štekauer

52 (38)

axoomo naabuc aug-­mosquito ‘big mosquito’

and the noun hulak ‘old person’ for animals or inanimates: (39)

hulaxa mwa aug-­house ‘big house’ (Bril, p.c.)

Wichí makes use of the inalienable root -­les ‘offspring’. The compound can denote the young of an animal (40a) or a little thing (40b): (40) a. hu’u-­les dim hen-­ ‘chick’ b. platu-­les plate-­ dim ‘little plate’ (Nercesian 2014, 758) It should be noted, however, that there is a fuzzy boundary between compounding and affixation in some languages. As maintained by Erschler (forthcoming), in Ossetic ‘[a] number of nominal stems function similarly to affixes and it is difficult to distinguish them from such’. (41) a. šaw-­rɐšuʁd black-­beautiful ‘very beautiful’ b. šǝrχ-­bǝn red-­bottom ‘reddish’ Another interesting type of compounding is a combination of two synonymous or semantically complementary words. The usual semantic effect is intensification: (42) Tat a. kosib-­ʕoni poor (Azerbaijani)-­poor (Hebrew) ‘very poor’ b. qoq-­mäħtel stunned-­surprised ‘very surprised’ (Authier, forthcoming) Finally, let us mention elative compounds, in which one of the constituents expresses a high degree of a property expressed by the other compound constituent; for example, Dutch Beresterk ‘strong as bear’, zielsbedroefd ‘soul-­sad = deeply sad’, aartsrivaal ‘arch rival’, bloedgang ‘very high speed’ (Hoeksema 2012).

Word-­Formation Processes in Evaluative Morphology 53 2.7  Root-­and-­pattern This WF technique, employed in Semitic languages, also finds its application in evaluative morphology: (43) Maltese a. 12vjjv3 tfajjel ‘small boy’ b. 1v22v3 qalla ‘to fry repeatedly’ (Brincat and Mifsud, forthcoming) 2.8  Sound symbolism Sound symbolism in evaluative formation is connected with the use of high front vowel and/or palatalised consonant.4 In Algonquian languages, sound symbolism commonly involves the palatalisation of *t to phonetic *[c]. This process may be accompanied by a diminutive suffix (Pentland 1975, 241): (44) Cree acihk, acihkošiš ‘little caribou’ < atihk ‘caribou’ 2.9  Non-­concatenative evaluative formation: introflection In Moseten, evaluative formation is based on the change of a stem vowel and/or its nasalisation (Sakel 2004, 100): (45)

täepäe little ‘piece’ < tipi ‘piece’

3  Additional observations This section provides some more details of evaluative formation which contribute to the picture of it as a formally very complex and diverse system. 3.1  Class-­maintaining and class-­changing derivation The vast majority of diminutives and augmentatives are formed by class-­maintaining rules. Class-­changing rules are illustrated here with Hausa (46), Icelandic (47) and Ladin (48): (46)

gamБas-­hēshḗ ‘break.off.a large.chunk-­aug’ ‘huge and muscular’ (Newman, p.c.)

(47) mót-­ingi to.meet-­ dim ‘good friend (someone who we like to meet)’ (Grönke 1955, 126)

Pavol Štekauer

54 (48)

stra-­vent-­é aug-­wind-­inf ‘to blow hard (in the context of wind)’ (Siller-­Runggaldier, forthcoming)

A rare case of class-­changing augmentivisation by reduplication is found in Jaqaru: (49)

t′usq~t′usqi rdp-­dust ‘very smoky (e.g. a room full of smoke)’ (Hardman 2000, 53)

3.2  Change of gender A change of gender is frequently employed in evaluative formation (see also Section 1.2.1 above). A derivational aspect of gender change can be seen, for example, in Slovak. Thus, chlapec ‘boy’ is masculine but chlapča ‘little boy’ is neuter; pes ‘dog’ is masculine but psíča ‘little dog’ is neuter, just like psisko ‘big dog’; and mačka ‘cat’ is feminine but mača and mačiatko, both meaning ‘kitten, little cat’, are neuter. In Zenaga, the diminutive aɣ-­əgrar-­t ‘small ram’ is feminine, derived from the masculine noun əgrər ‘ram’ (Shay 2014, 579). The basic role is played by the opposition between the masculine and the feminine genders, with the latter usually indicating diminutiveness. Prototypical cases can be found in Chadic, Cushitic and Berber languages: (50) Figuig a-­ɣənža (m.) ‘ladle’ vs. ta-­ɣənžay-­t (f.) ‘spoon’ (Shay 2014, 579). Counter-­examples can be found in the African languages Bafut, Xhosa and Hamar. In Pévé, importance or size can also be coded by use of the feminine agentive prefix i-­Shay (2014, 579): (51)

gwayʔ ‘mountain’ vs. i-­gwayʔ ‘large/important mountain’

Finally, in Migama, a change in gender in either direction results in a diminutive meaning: (52) gáàpú (m.) ‘large horn’ gáàpè (f.) ‘small horn’ dâmbá (f.) ‘mountain, mountain range’ dâmbú (m.) ‘rock’ Shay (2014, 579) 3.3  Recursive affixation in evaluative morphology Recursiveness in evaluative formation is fairly common. Lower Sorbian frequently employs a combination of two different diminutive suffixes: (53) rěp-­cyc-­k-­a beet-­ dim-­dim ‘very small beet’ (Viktor Zakar, p.c.) Lithuanian (54), Even (55) and Tapiete (56) also provide examples of a combination of two diminutive suffixes:

Word-­Formation Processes in Evaluative Morphology 55 (54) brol-­už-­ėl-­is < brol-­ùž-­is < bról-­is ‘brother’ brother-­ dim-­dim ‘little brother’ (Klimas, p.c.) (55)

olra-­ča-­kan fish-­dim-­dim ‘little fish’ (Nikolaeva 2014, 499)

(56) mbahɨru-­ra’ɨ-­mi-­re dim-­dim.pl bag-­ ‘little bags’ (González 2005, 104) An intensifying effect is obtained in Turkish by repeating the same suffix: (57) küçü(k)-­çü(k)-­çük small-­ dim-­dim ‘very very small’ (Nihan Ketrez, p.c.) In Matsés, high intensity is expressed by using two different augmentative enclitics: (58) ënapen-­quio-­shë ic-­e-­c long-­ aug-­aug be-­ npast-­ind ‘It is very, very long’ (Fleck 2003, 285) The following example of multiple recursiveness is taken from the Nain subdialect of Labrador Inuktitut (Johns 2014, 710): (59)



Kuatsâ-­pâ-­pâ-­pâ-­pâ-­pâ-­pâ-­lauk-­Kunga startle-­very-­very-­very-­very-­very-­very-­ distpst-­ind.1sg ‘I really really had a very big fright’

Recursive use of prefixes can also be found across languages: (60) Slovak and Czech pra-­pra-­starý aug-­aug-­old ‘very very old’ Finally, let us refer to a Yupik example taken from Mithun (1999, 43) which demonstrates that the order of two evaluative markers can significantly affect the meaning of an evaluative: (61) a. yug-­pag-­cuar person-­big-­little ‘little giant’ b. yug-­cuar-­pag person-­little-­big ‘big midget’

Pavol Štekauer

56 3.4  Merged evaluative suffixes

Sometimes two or even three diminutive suffixes merge into one. Wagner-­Nagy (forthcoming) illustrates such a case for Nenets with the complex diminutive suffix -­cʲeko(Ɂ)je, which includes the three diminutive suffixes -­ko, -­cʲe and -­je: (62) puxuu-­cʲeko(Ɂ)je old.woman-­ dim ‘little old woman’ In Mordvin, the two diminutive suffixes -­ka and -­ińe are combined: (63)

numolkińe ← numolo ‘rabbit’ (Maticsák, forthcoming)

3.5  Complementary distribution Sometimes, suffixation and prefixation are complementary in their scope of activity. In Micmac, prefixation is reserved for augmentatives (64b) and suffixation for diminutives (64c): (64) a. eleke’wi’skw ‘queen’ b. kji-­eleke’wi’skw ‘empress’ c. eleke’wi’skwe’j ‘princess’ (Hewson and Francis 1990, 36) A similar situation can be found in Breton: (65) a. bihan-­ik small-­ dim ‘smallish’ b. gour-­lano aug-­tide ‘high tide’ 3.6  Evaluative formation and word classes Cross-­linguistically, evaluative formation ranges over all four major word classes, but, obviously, with unequal extent and productivity and with different semantic categories employed. Each of the basic word classes is dominated by suffixation. Ample examples of evaluative nouns, verbs and adjectives are given in the preceding sections. Therefore, the following is an illustration of suffixation yielding evaluative adverbs, in particular, in Tat (with attenuative meaning) (66) and in Basque (with intensified meaning) (67), and of reduplication to form augmentatives in Gagauz (68) and the Campidanese variety of Sardinian (69): (66) yevoš-­ley / asda-­lay slowly-­ att/slowly-­att ‘rather slowly’ (Authier, forthcoming)

Word-­Formation Processes in Evaluative Morphology 57 (67) emeki-­ño softly-­ int ‘very softly’ (Hualde and Ortiz de Urbina 2003) (68)

(69)



üüsek-­üüsek high-­high ‘very highly’ (Menz, forthcoming) beni-­beni well-­well ‘very well’ (Pinto, forthcoming)

In Aghul, evaluative adverbs are derived from adjectives by a combination of reduplication and suffixation: (70)



deχi~deχi-­di quick-­quick-­state ‘very quickly’ (Maisak and Ganenkov, forthcoming)

The opposite direction of evaluative formation is offered by another Nakh-­Daghestanian language, Avar, in which adjectives can be based on adverbs combined with the participle kkarab ‘happen.ptcp’: (71) c’aħil-­kkarab dimly-­ dim ‘greyish’

t’eren-­kkarab thinly-­ dim ‘thinnish’ (Khalilov and Khalilova, forthcoming) Evaluative adverbs are also productively generated in some other languages in my sample, including many Slavic languages, Basque, Sardinian, Frisian, Dumi, Matses, Kwakw’ala, etc. The diminutivisation of pronouns is a very rare case. As pointed out by Amha (2001, 71) for Maale, the masculine demonstrative form hayí ‘this’ (which, when referring to things and non-­humans, denotes something big) can be combined with the diminutive marker -­má to express relative smallness: (72) hayi-­mmá this.m-­dim In contrast, when ha ‘this-­f/m’ (which is the base for both masculine and feminine demonstratives) is combined with the diminutive, it refers to an even smaller object. In the next example the demonstrative may be used to refer to a tiny object held in the hand: (73) ha-­mmá this-­ dim

Pavol Štekauer

58

Another example of pronoun augmentivisation comes from Agta (Austronesian): (74)

ad-­áddu ‘very many’ < addu ‘many’ (Healey 1960, 7)

3.7  Serving two masters In some languages, an evaluative marker may be applied to form complex words in two or more word classes. In Yurakaré¸ the suffix -­nñu is used with nouns, verbs and adjectives. In Supyire, the diminutive suffix -­rV produces both adjectives and nouns. The diminutive suffix -­cik in Gagauz attaches to both nouns and adjectives: (75) a. oda-­cık dim room-­ ‘small room’ b. alça-­cık low-­dim (Menz, forthcoming) In Nama, one and the same suffix serves three different word classes: nouns, verbs and adjectives: (76)

xuu-­ró’ì ‘some little thing’ ≠xarí-­ró ‘rather small’ !xóé-­ró ‘run a bit’ (Hagman 1977, 26, 33, 74)

3.8  Hybrid formations Hybrid formations of different types can be found cross-­linguistically in evaluative morphology. In Maltese, a Romance suffix is added to a Semitic element which is treated as a stem: (77) dar-­un house-­ aug ‘large house’ (Brincat and Mifsud, forthcoming) In Ossetic, the loaned Russian suffix -­ik may attach to a clipped proper noun: (78)

As-­ik (a male name) Fat-­ik (a female name) (Erschler, forthcoming)

3.9  Number of evaluative morphology markers There are languages which make use of only one diminutivising and/or augmentivising evaluative marker, like Mari, Kabardian, Hup, Bāgandji, Kolyma Yukaghir, etc. An interesting situation is observed in Dolakha Newar, in which the diminutive suffix -­cā is the only productive derivational suffix in the derivational system of the language (Genetti 2007, 93). In Udihe, the diminutive suffix -­zig’a is the most productive of all nominal derivational suffixes. Nikolaeva and Tolskaya (2001, 149) draw attention to the fact that it attaches to practically any noun with concrete semantics.

Word-­Formation Processes in Evaluative Morphology 59 On the other hand, there are languages with a multiplicity of evaluative markers (usually suffixes), like Slavic, Baltic, Romance languages and Altaic languages. As an example, Klimas (1974), based on Urbutis (1971), maintains that in Lithuanian there are nineteen regular diminutive suffixes, thirty rare diminutive suffixes and thirty-­one dialect/regional diminutive suffixes. 3.10  Special case This last point shows that the diminutive and augmentative meanings can be combined in a single evaluative. Thus, in example (79) from Croatian, which illustrates a relatively common case in some of the Slavic languages, the augmentative meaning of iterative action ‘more than default quantity – i.e. more than one jump’ combines with the diminutive meaning of low intensity ‘less than the default quantity in terms of the extent of jumping’: (79) a. skak-­ati ‘to jump’ skak-­ut-­ati ‘repeated jumping with low intensity’ b. zvižd-­ati ‘to whistle’ zvižd-­uk-­ati ‘to repeatedly whistle with low intensity’ (Mihajlović, p.c.)

4 Conclusion This typological overview, which is far from being exhaustive, provides evidence of the employment in evaluative morphology of a wide range of WF processes. In addition, it follows from the examples that the formation of evaluatives is not restricted to the field of WF itself, but there are also inflectional devices that are used for the same purpose. In general, the widely accepted assumption about the fuzzy boundary between inflection and derivation is given further support in this chapter. Nevertheless, a cross-­linguistic analysis of the data unambiguously confirms the dominance of suffixation, and the important role of reduplication. Compounding, a crucial WF process, does not play a significant role in evaluative formation, and – if employed for this purpose – it frequently relies on a single compounding constituent. From the semantic point of view, it has been shown that individual evaluative-­formation processes range over a number of semantic categories, sometimes yielding surprisingly contradictory meanings; for example, the case of iconic and anti-­iconic meaning categories produced by reduplication, or a combination of augmentativeness and diminutiveness in a single evaluative. Recursiveness of evaluative suffixes is not a unique phenomenon, and it is centred on the European territory, especially in Slavic languages. These and a number of other SAE languages are also characteristic of a (usually) large stock of evaluative affixes and of the ability to produce evaluatives of various word classes. These observations give support to the claim that evaluative morphology is a characteristic feature of SAE (with Slavic languages constituting the core linguistic area) through which this linguistic territory differs in many respects from the other languages of the world.

60

Pavol Štekauer

Notes 1. Clearly, the default value is of a relative nature because it may depend on a speech community’s culture, traditions and habits as well as, pragmatically, on a specific linguistic or situational context. 2. The origin of the diminutive proclitic is the gender 1 noun m-­ɔ̀ŋɔ́ ‘child’. 3. See Derzhanski (2005) for a discussion of diminutive plurals and plural diminutives in Bulgarian as well as some other languages of the world. 4. See Ultan (1978), Nieuwenhuis (1985), Bauer (1996), Gregová, Körtvélyessy and Zimmermann (2010) and Körtvélyessy (2012b) for a discussion of various aspects of phonetic iconicity in evaluative morphology.

5  Evaluative Morphology and Language Universals Lívia Körtvélyessy

1 Introduction The objective of this chapter is to identify correlations between evaluative morphology (EM) and language universals. First, I define the scope of EM; second, my approach to language universals is introduced; third, the overlap of language universals and EM is discussed. Finally, sources of universals in EM are reviewed, and the existing universals are discussed.

2  Scope of evaluative morphology An evaluative construction may be defined as a construction that meets two conditions: a formal one (the construction is morphologically complex) and one relating to the functional-­semantic level (the construction is a deviation from the default value). Interestingly, most of the former discussions in this field are limited to diminutives and augmentatives. It is no wonder that Bauer (1997) refers to these two categories as the core of EM. However, numerous cases of Aktionsart (for example, Arabana Distributive wantya-­wantya ‘a lot of people dying all around’; Hungarian Frequentative olvasgat ‘to read iteratively’; Maya Frequentative tin bi-­bi-­qab ‘I tap frequently with the fingers’) show that the borders of EM are, in fact, much broader. Moreover, EM overlaps with inflectional morphology too in cases that have been traditionally referred to as prototypical examples of inflection – notably, the comparison of adjectives (see van Gijn, this volume). Štolc (1958), for example, includes grading in the field of word formation. Even though the qualitative meaning of a word is not changed by grading, there is a change in quantity and a new word is derived. Moreover, Štolc finds parallels between grading and diminutivisation/augmentivisation. Both reflect a need for a quantitative change of the meaning, and they make use of the same formal means (usually suffixes). Štolc labels grading as relative diminutivisation/augmentivisation with regard to a third object. Diminutivisation/ augmentivisation proper is absolute grading. It is used to express the absolute grade of a specific quality without reference to a third object. The expressive ‘load’ of diminutivising and augmentivising suffixes is also much bigger than in the case of grading. In my approach, EM is examined against the background of the category of Quantity, and can be treated as a continuum, with prototypical cases expressing the meaning of quantity under or above a default value. This approach allows us to encompass categories such as Pluractionality (see Štekauer, Valera and Körtvélyessy 2012), as exemplified in (1):

62

Lívia Körtvélyessy

(1) Hausa tun~tùnā red~remind ‘remind many or often’ (Newman, p.c.) or Attenuation, as in (2): (2) German grün-­lich green-­ att ‘greenish’ (my example) as well as Aktionsart (Iterativity, Distributiveness, Multiplicity, Intensity, etc.) as in (3). In fact, Pluractionality is an Aktionsart, too, as in (4): (3) Slovak popiskovať iter-­whistle ‘to whistle a little bit’ (4) Maltese bas~bas to.break.wind-­to.break.wind ‘to break wind frequently’ (Brincat and Mifsud, forthcoming) In each of these examples the reference point is the default value within the cognitive categories of Substance, Action, Quality and Circumstance. Thus, tun~tùnā refers to a repeated Action or an Action bearing on more Substances; popiskova’ deviates from the default value to whistle by indicating the repetition of Action, and greenish indicates a smaller amount of Quality. The default value is language specific, influenced by many factors, such as culture, the traditions of a speech community, or a speech situation. The process of evaluation starts in extra-­linguistic reality and is motivated by the need of a speech community to express the idea of ‘less than/more than the standard quantity’ in relation to a particular object of extra-­linguistic reality. This evaluative process is reflected at the cognitive level by quantification and is expressed by the cognitive categories of Substance, Quality, Action and Circumstance. Any qualitative evaluation is based on an iconic shift that can be generalised by the metaphors SMALL IS CUTE and BIG IS NASTY. For example, Slovak vtáčatko ‘bird’ can refer not only to size, but also to qualities like tenderness, beauty or cuteness. The cognitive categories are implemented via the semantic categories of Diminutive, Augmentative, Pejorative, Ameliorative, Pluractionality, Attenuation, Intensification, Iterativity, etc., represented by concrete evaluative markers in combination with the motivating nouns, verbs, adjectives and adverbs. An evaluative construction may undergo some phonological changes (e.g. palatalisation) and is ready to enter the level of parole, where, if needed, various additional shades of emotive colouring, depending on the specific context, can be applied. This approach to EM does not, in principle, deviate from the idea presented in the Introduction to this volume (Chapter 1). The opposite is true: it makes the view broader and enables us to study language universals in EM from a wider perspective. Most of the data and conclusions in this chapter are based on the evaluation of data covering 200 languages.

Evaluative Morphology and Language Universals 63

3  The state of the art The few decades of the existence of EM1 do not seem to have been prolific in terms of the identification of universals. This is not surprising: every new field of research first concentrates on the clarification of its fundamental theoretical principles, categories and positions. In order to compensate for the lack of attention paid to the question of universals in EM, this section maps all available sources of language universals in EM. Each language universal falling within the scope of EM is briefly described and, where available, examples are provided. 3.1  The Universals Archive When searching for a language universal, the first step usually taken by linguists is to browse the Universals Archive (Plank, Filimonova, Mayer and Mayorava, n.d.). It includes the following domains: syntax, morphology, semantics, phonology, inflection, prosodic phonology, word formation, lexicon, discourse and pragmatics.2 A simple browsing test with the keyword evaluative confirms the view expressed at the beginning of this chapter: the notion of EM is too young to be included in the Universals Archive. Since EM is not the exclusive term used in reference to this linguistic field, the alternative terms expressive, extra-­grammatical and appreciative3 were used, too. A similar test was carried out with the following notions referring to EM: diminutives, augmentatives, pejoratives, amelioratives, approximation, reduction, attenuation, intensification, deintensification, endearment, hypocoristics, expression of social position, contempt, ­authenticity/ prototypicality, Aktionsart, Plurality, Frequentatives, Iterativity, Distributiveness and Multiplicity.4 Table 5.1 summarises the findings. Altogether there are three language universals in the Universals Archive that refer to EM. Let us have a closer look at them. 3.1.1  Universal 654 The source of Universal 654 is Head (1978). It is an absolute and implicational ­universal, with its domains in inflection and semantics. It expresses the correlation between the occurrence of pejorative pronominal affixes and honorific affixes: (5)

IF there are pejorative pronominal affixes, THEN there are honorific ones.

This correlation was attested in more than 100 languages. An example of an honorific p­ ronominal affix comes from Nahuatl: (6)

notē-­chīuhcā-­tzĭn my-­progenitor-­H (Karttunen 1990, 284)

No comments or counter-­examples are provided in the Archive. 3.1.2  Universal 2009 This implicational and statistical universal was proposed by Bauer (1997, 540). He ­discusses the hierarchy of word classes in reference to derivation and suggests the following implicational relations:

Domain

– – – – – – – – – – – – – – – – – – – – – – – –

Syntax

– – – – – – – – – – – – – – – – – – – – – – – –

Morphology – – – 1926 1926 654 – – – – – – – – – – – – – – – – – –

Semantics

The numbers refer to the coding of the universals in the Archive.

Expressive Extra-­grammatical Appreciative Diminutive Augmentative Pejorative Ameliorative Approximation Reduction Attenuation Intensification Deintensification Endearment Hypocoristics Social position Contempt Authenticity Prototypicality Aktionsart Plurality Frequentatives Iterativity Distributiveness Multiplicity

Notion – – – 1926 1926 – – – – – – – – – – – – – – – – – – –

Phonology

Table 5.1. Language universals in EM from the Universals Archive

– – – – – 654 – – – – – – – – – – – – – – – – – –

Inflection – – – – – – – – – – – – – – – – – – – – – – – –

Prosodic phonology – – – 2009 2009 654 – – – – – – – – – – – – – – – – – –

Word-­ formation – – – – – – – – – – – – – – – – – – – – – – – –

Lexicon – – – – – – – – – – – – – – – – – – – – – – – –

Discourse – – – – – – – – – – – – – – – – – – – – – – – –

Pragmatics

Evaluative Morphology and Language Universals 65 (7) IF augmentatives/diminutives can be formed from Determiners, THEN they can also be formed from Adverbs, Numerals, Pronouns, or Interjections; IF from Adverbs, Numerals, Pronouns, or Interjections, THEN also from Adjectives or Verbs; IF from Adjectives or Verbs, THEN from Nouns. There are numerous counter-­examples, though. For example, Kanuri, Djaru and Alawa are languages with adjectival evaluatives and no substantival evaluatives; Tat has adverbial but no adjectival diminutives; Nivkh and Vitu have only verbal evaluatives; Udihe has only verbal and adjectival augmentatives; and Kwakw’ala has adverbial and substantival augmentatives but no adjectival or verbal ones. Similar examples abound. Bauer works with a core sample of thirty languages and identifies diminutives and augmentatives as core categories of EM (1997, 537). In this way, he develops the ideas of Ettinger (1974) and Nieuwenhuis (1985). Bauer himself modifies his claim in the ­comments. He maintains that the further down the hierarchy, the less productive is evaluative derivation. At the same time he suggests that diminutive EM is more productive than ­augmentative EM and the occurrence of augmentative EM implies diminutive EM: (8) IF augmentative derivational morphology, THEN also diminutive derivational morphology. Counter-­examples, however rare, can be found. Hausa and Ilocano have only augmentatives (Körtvélyessy 2012b). 3.1.3  Universal 1926 Universal 1926 relates phonetic iconicity and diminutives and augmentatives: (9) There is an apparently universal iconic tendency in diminutives and augmentatives: diminutives tend to contain high front vowels, whereas augmentatives tend to contain high back vowels. The idea comes from Payne (1997). Generally, it is assumed that smallness is associated with front high vowels and largeness with back vowels. This assumption of phonetic ­iconicity (or sound symbolism, phonetic symbolism or phonosymbolism) engendered many experiments; for instance, Bentley and Varon (1933); Berlin (1995); Brackbill and Little (1957); Diffloth (1994); Fisher-­Jørgsen (1978); Jespersen (1922; 1933); and Sapir (1929).5 Universal 1926 claims that phonetic iconicity can be observed not only in words meaning ‘small’ or ‘big’ but also in evaluative constructions. Thus, the evaluative suffix in (10) is represented by a high vowel: (10) bird-­ie bird-­ dim Even though this universal is described as unconditional in the Universals Archive, with an absolute quality, it has been challenged by some authors.6 The Universals Archive mentions counter-­examples from Korean and Bahnaric. In their analysis of twenty-­four languages (Indo-­European, Austronesian and Niger-­Congo), Gregová, Körtvélyessy and Zimmermann (2010) call into question Universal 1926: front high vowels are typical of augmentatives rather than diminutives, while diminutive affixes are phonetically realised by central vowels.

Lívia Körtvélyessy

66 Table 5.2. Summary of EM universals from the Universals Archive Universal number 654

2009

1926

Domain

Type

Status

Quality

Basis

Counter-­ examples

Inflection Semantics Word-­ formation Word-­ formation

Implication

Achronic

Absolute

More than 100 languages

None

Implicational hierarchy

Achronic

Statistical

None

Unconditional

Achronic

Absolute?

Core sample 30, inner sample 50, broad sample 1201 Languages surveyed in Payne (1997)

Phonology Semantics

Korean, Bahnaric

Alveolar and velar consonants seem to predominate in both diminutives and augmentatives. The authors also suggest that ‘the rule of sound symbolism may be overpowered by other principles in some cases’; for example, by the principle of vowel harmony in the Hungarian language (2010, 172). The tendencies observed are of an areal rather than universal nature. 3.1.4  Universals Archive: summary Table 5.2 summarises the number, domain, type, status and quality of language universals in the Universals Archive from the field of EM. According to Table 5.1, out of 2,029 universals specified in the Universals Archive there are only three that are related to the field of EM. One of them (Universal 2009) is from the field of word formation alone, and two from the field of semantics, overlapping with either word formation and inflection (Universal 654) or phonology (Universal 1926). Universals 654 and 2009 are implicational while Universal 1926 is unconditional. All of them are achronic. Universal 654 is described as absolute. The same description is attached to Universal 1926, although with a question mark. As the discussion on this universal reveals, it is rather a statistical universal. At the same time, it is the only one out of the three for which counter-­examples can be found in the Universals Archive. 3.2  Jurafsky (1993; 1996) Another source of universals concerning EM is Jurafsky (1993; 1996). Jurafsky (1993, 423) proposes a ‘universal radial category for the diminutive construction which includes the concepts of resemblance, imitation, gender, partitives, approximation, and hedging’. He understands the sense of the word child as the historically and semantically prior sense of diminutives. Very early it was extended by the sense ‘small’. Extension of the sense ‘small’ mostly happens in the category of the diminutive. The use of diminutives to refer to children or offspring is of a universal nature. Jurafsky’s approach is both synchronic and diachronic. The diachronic senses ‘small/child’, ‘approximation’ and ‘resemblance’, as well as emotional connotations, developed the synchronic senses of affection, approximation, intensification, imitation, contempt, exactness, individuation, female gender, playfulness and metalinguistic hedges by means of various mechanisms of semantic change (metaphor, conventionalisation of inference, generalisation and lambda-­ abstraction).

Evaluative Morphology and Language Universals 67 Jurafsky supports his theory by referring to the PIE suffix *-­ko-­, which had a diminutive meaning. He aims to provide evidence that its original meaning was ‘child’. However, his argumentation was rebutted by Prieto (2005, 78) and Miller (2006, 58). Jurafsky focuses only on diminutives while augmentatives are left unnoticed. These are discussed by Matisoff (1991) in relation to the languages of South-­East Asia, in which the morpheme originally meaning ‘mother’ has grammaticalised to the augmentative, as illustrated in the following example from Thai (1991, 303): (11)

hŭa-­mɛ̂ɛ-­tiin head-­mother-­foot ‘big toe’

Unfortunately, one cannot generalise Matisoff’s observation of the historical roots of augmentatives because of the lack of cross-­linguistic data: very little attention has been paid to the semantics of augmentatives and to those areas of EM that go beyond the scope of the category of diminutiveness. 3.3  Beard (1995) Beard (1995), within the frame of his Lexeme-­Morpheme Base Morphology, distinguishes three types of derivational rules: feature value switches, expressive derivations and functional L-­derivations. His expressive derivations overlap with the concept of EM. Beard claims that expressive derivations are used to express the speaker’s opinions and feelings about the ‘size and merit’ of the denotated object and that ‘they reflect at least five functions universally: Diminutive, Augmentative, Pejorative, Affectionate and Honorific’ (1995, 163). He gives the basic features of expressive derivations: they change neither the meaning nor lexical class of the lexemes over which they operate, nor the word class – nouns are generated from nouns, verbs from verbs, etc. They are used to ‘express prejudices of the speaker as to whether the referent is smaller, larger, more likeable, or more threatening than other members of its semantic category’ (1995, 163) and are always optional and subjective. They are capable of making degrees; they are gradient. 3.4 Körtvélyessy (2012b) Körtvélyessy (2012b) examines EM in various languages of the world. It also aims to find out whether EM is a relevant criterion contributing to the identification of Standard Average European (SAE) as an independent language area. Finally, phonetic iconicity as an indispensable part of EM is experimentally studied in three typologically different languages from the sociolinguistic point of view. Methodologically, the study relies on a cross-­linguistic approach, analysing 132 world languages and 71 SAE languages. The data indicate that rather than being a universal phenomenon, EM is areally conditioned. The results also confirm the areal nature of phonetic iconicity. Besides a discussion on universals, already mentioned in this chapter, Körtvélyessy suggests a new implicational universal concerning semantic categories (see also Chapter 14, this volume). Her understanding of cognitive categories is based on Dokulilean onomasiological tradition (see Dokulil 1962). If Dokulil’s theory is applied to the EM, four ­cognitive categories can be distinguished: Quantity of Substance, Quantity of Action, Quantity of Quality and Quantity of Circumstance. These categories are illustrated by the examples below:

Lívia Körtvélyessy

68 (12)

Quantity of Substance (Arabana): kalti-­kaltiRi rdp-­hill ‘small hill’ < ‘hill’ (Hercus 1994, 98)

(13)

Quantity of Action (Belorussian): папа-­гуляць papa-­guľac freq-­walk ‘to walk a lot’ (Rudenka and Rudaya, p.c.)



(14) Quantity of Quality (Tatar): küg-­elcem att blue-­ ‘blueish’ (Miftachova, p.c.) (15) Quantity of Circumstance (Hungarian): lassacskán lassan-­cskán slowly-­ dim ‘very slowly’ (my example) From her data, Körtvélyessy draws the following universally valid conclusion: (16) IF Quantity of Quality or Quantity of Circumstance THEN Quantity of Substance and/or Quantity of Action. In addition to this universal, Körtvélyessy formulates five Euroversals. The first is the already mentioned implicational universal that relates the occurrence of augmentatives to the occurrence of diminutives in individual languages: (17)

IF augmentatives THEN diminutives.

The second Euroversal comments on the presence of EM in SAE languages: (18) Evaluative morphology (diminutives and augmentatives) is an inherent feature of the morphological systems of SAE languages – unlike the situation in the rest of the world’s languages. The third Euroversal reflects the same fact as the second, but it introduces EM saturation value:7 (19) SAE languages – unlike the rest of the world’s languages – feature an extremely high EM saturation. The next two Euroversals compare the occurrence of cognitive categories in the SAE languages and the world sample:

Evaluative Morphology and Language Universals 69 (20) The percentage of occurrence of all the cognitive categories within morphological diminutives of the SAE languages is much higher than in the other areas of the world. (21) The percentage of occurrence of all the cognitive categories within morphological augmentatives of the SAE languages (with the exception of the Quantity of Quality) is much higher than in the other areas of the world.

4 Discussion While in the previous section language universals in EM were seen from the point of view of various sources, this section takes a different point of view. Following the structure of the Universals Archive, this section examines the existence of universals by domains, with the exception of syntax and lexicon.8 4.1  Phonology, prosodic phonology This field subsumes Universal 1926, which concerns phonetic iconicity. As has already been pointed out, various analyses show (Ultan 1978; Bauer 1996; Gregová et al. 2010) that this universal is rather of an areal nature. There are languages in which diminutive and augmentative affixes comply with Universal 1926; for example: (22) Croatian učitelj-­čić teacher-­ dim ‘little teacher’ (Zrinka Jelaska, p.c.) (23) Irish éin-­ín bird-­ dim ‘little bird’ (Aidan Doyle, p.c.) (24) Javanese u-­èlèk int-­ugly ‘very ugly’ (Ignatius Supriyanto, p.c.) and there are affixes that do not comply with this Universal: (25) Bemba kaa-­sando dim-­hammer ‘little hammer’ (Alex Kasonde, p.c.) (26) Lakota wí-­la sun-­ dim ‘little sun’ (Della Bad Wound, transmitted by David Rood, p.c.) (27) Russian домище house-­ aug ‘big house’ (Viktoria Liashuk, p.c.)

Lívia Körtvélyessy

70

It should also be noted that higher tonality in diminutives versus lower tonality in augmentatives has been called into question, too. Higher tonality is mentioned by Jespersen (1933, 284): ‘The reason why the sound [i] comes to be easily associated with small, and [a, o, u] with bigger things, may be to some extent the high pitch of the vowel.’ It is also discussed by Sapir (1929), Nichols (1971) and Ultan (1978). In connection with this, Ohala (1994, 336) proposes a biological evolutionary origin of sound symbolism. He associates it with the frequency code, ‘the association of high acoustic frequency with smallness and low acoustic frequency with largeness’. However, none of these assumptions are based on the statistically significant data which would qualify them to be language universals. 4.2 Morphology There are two universals that fall within the scope of morphology. First is Universal 2009, already discussed, on the basis of Bauer’s data (see (8) above). Second, on the basis of Štekauer (this volume) and Körtvélyessy (2012b) the following universal tendency can be formulated: (28)

Evaluative constructions are derived mainly by means of suffixation.

4.3 Semantics First of all, let us mention the implicational universal ‘IF augmentatives THEN diminutives’. Even though counter-­examples are available (see above), they are rare. The universal nature of this implicational relation is also supported by Beard’s observation of five functions universally reflected by expressive derivations: diminutive, augmentative, pejorative, affectionate and honorific (Beard 1995, 163). Two fields of linguistics – morphology and semantics – overlap in Universal 654, which is based on the data from Head (1978). It conditions the presence of pejorative pronominal affixes by the presence of honorific pronominal suffixes. Thus far, there are no counter-­ examples to this universal; moreover, it is also supported by Beard’s claim based on five universal functions of expressive derivations. The absence of counte-­rexamples, however, might also be the result of the lack of attention paid to this field. The next universal, ‘IF Quantity of Quality or Quantity of Circumstance THEN Quantity of Substance and/or Quantity of Action’, has been supported statistically by data from 200 languages (Körtvélyessy 2012b). Körtvélyessy’s data has also confirmed the central claim of Jurafsky’s (1996) radial category model, namely that the frequent source of diminutiveness is the concept of child, as evidenced by languages like Bafut (Niger-­Congo), Akan (Niger-­Congo), Koyra Chiini (Nilo-­ Saharan), Khoekhoe (Khoisan), Japanese (isolate) and Ao (Sino-­ Tibetan). The concept of the young extension was found in the Malayalam concept of child. Diminutive markers with this meaning were identified, for example, in Khoekho, and another extension of the concept of child – that of seed or grain exists in Movima (isolate). However, this claim was not proved to be universal. A different type of conceptual motivation in EM can be found in connection with the category of Quantity. It can be exemplified by the diminutive suffix -­ulu in Bāgandji: (29) ηidja-­ulu mūrba-­ulu one-­ sg child-­ dim ‘one single small child’ (Hercus 1982, 81)

Evaluative Morphology and Language Universals 71 Its basic meaning is ‘singular’. On the other hand, the Austronesian Muna language makes use of the augmentivising prefix mansi-­meaning ‘more items than expected’; that is, more than the standard, default value. In Luxembourgish the singular diminutive forms have plural bases: (30)

Kätzercher ‘kitten’ < Kätzchen (pl.) < Katz ‘cat’ (Krummes, p.c.)

Jurafsky referred to his claim as one of universal tendencies, and the data above makes us limit his reference to tendencies. The semantic motivation of evaluative markers is another challenge, especially in the field of augmentatives. 4.4 Pragmatics While phonology, morphology and semantics describe the level of language system, pragmatics is connected to parole. At the level of parole, various additional shades of emotive colouring can be applied to the evaluative construction depending on the specific context. Even though it is obvious that these additional shades are more than frequent, there has been no systematic research into this issue. Generally, it is stated that diminutives and augmentatives are typical of child speech (Dressler and Merlin Barbaresi 1994) and politeness (Brown and Levinson 1978). However, a systematic study of their functions is lacking. Schneider (2013) reviews formal and semantic approaches to the typological comparison of diminutives and suggests a new approach where ‘the starting point for the analysis is neither form nor meaning but function, and especially socially motivated functions which diminutives fulfil in specific types of context’ (p. 149) He illustrates this approach with the example of the ‘modesty context’ that exists in a large number of related and unrelated languages.

5 Conclusion The basic aim of this chapter has been to give a summary of language universals in EM. To this end, various sources dealing with either language universals or EM (or ideally both) have been analysed. Even though the initial expectations were modest, the reality is even worse. In total four language universals, statistically supported by a sufficient number of data and/or without counter-­examples, are cited: (31) IF augmentatives/diminutives can be formed from Determiners, THEN they can also be formed from Adverbs, Numerals, Pronouns, or Interjections; IF from Adverbs, Numerals, Pronouns, or Interjections, THEN also from Adjectives or Verbs; IF from Adjectives or Verbs, THEN from Nouns. (32)

IF augmentatives THEN diminutives.

(33)

IF there are pejorative pronominal affixes, THEN there are honorific ones.

(34) IF Quantity of Quality or Quantity of Circumstance THEN Quantity of Substance and/or Quantity of Action. Just for illustration: the Universals Archive counts 2,029 language universals. Out of them, 789 are from the field of inflection, 157 are classified as morphological universals, 142

72

Lívia Körtvélyessy

concern semantics and 51 are matched with word formation. However, there are only 4 universals that fall within the scope of EM. This situation raises a simple question – why? There seem to be several answers to this question. Firstly, EM is not a language universal. In a sample of 203 languages, 58 languages (29 per cent) do not have any EM. By implication, the applicability of any universal in EM is restricted to the languages with EM. Secondly, the EM saturation values of individual geographical areas make it possible to order them as follows (starting with the most saturated area):9 SAE, South America, Eurasia, North America, Africa, South-­East Asia and Oceania, Australia and New Guinea (see Körtvélyessy 2012b). Consequently, EM is, in principle, an areal phenomenon, which narrows down the chances of finding universals in EM. Thirdly, the term ‘evaluative morphology’ was introduced only in 1984 (by Scalise). Even though diminutives (and rarely also augmentatives, pejoratives and honorifics) had been discussed before, the discussion did not have any unifying umbrella discipline (morphology, semantics, lexicology). EM is, however, claimed to be an independent and specific branch of morphology. Nearly thirty years of existence does not seem to have been a sufficient period of time for large-­scale and statistically supported cross-­linguistic research. Fourthly, since EM is young, its own methods of analysis, parameters that could be used for relevant comparative analysis and database are missing, and as such they are a challenge for linguists working in this field.

Acknowledgements I wish to express my gratitude to all the informants who were so kind as to supply the necessary data for this chapter: Sviatlana Rudaya and Alena Rudenka (Belorussian), Ilmira Miftachova (Tatar), Zrinka Jelaska (Croatian), Aidan Doyle (Irish), Ignatius Supriyanto (Javanese), Paul Newman (Hausa), Alex Kasonde (Bemba), Della Bad Wound (Lakota) and Viktoria Liashuk (Russian).

Notes 1. See Chapter 1, this volume. 2. The domain section classifies universals by content at different levels of generality. The domains and domain-­combinations listed are currently distinguished. Obviously, morphology is sometimes differentiated as inflection and word formation. By analogy, phonology is sometimes specified as prosodic phonology. 3. Expressive and plain morphology are distinguished by such luminaries as Zwicky and Pullum (1987); Szymanek (1988) also works with the term expressive. Extra-­grammatical is a notion used by Dressler and Karpf (1995); appreciative affixes can be found in Gràcia and Turon (2000). The accepted understanding of these terms more or less overlaps with EM. For further discussion see Körtvélyessy (2012b; 2014). 4. The notion of reduction, for example, can be identified in two universals in the domain of syntax. The first relates vowel reduction to word order and the second is about conjunction reduction in the hierarchy of ergativity. As such they do not refer to the scope of EM and were not taken into consideration. 5. For a more detailed list see Körtvélyessy (2011a). 6. See Ultan (1978), Nieuwenhuis (1985), Bauer (1996) and Štekauer et al. (2012). Ultan and Nieuwenhuis suggested that phonetic iconicity in diminutives and augmentatives is bound not only to specific vowels but also to specific consonants – diminutiveness is associated with palatal and postalveolar consonants. Ultan also believes that phonetic iconicity is not a universal ten-

Evaluative Morphology and Language Universals 73 dency, rather an areally distributed phenomenon However, Bauer (1996, 201), like Štekauer et al. (2012), does not observe ‘any universal principle of sound symbolism operating in markers of the diminutive and augmentative such that palatal articulation correlates with diminutive and not with augmentative’. 7. EM saturation is the mean of three values: word-­formation value, semantic categories value and word-­class value. The word-­formation value reflects the number of word-­formation processes and their specification. If the number of word-­formation processes is one or two, it is assigned one point; the use of three or four processes is assigned two points; etc. Extra points are added for subparameters, such as the possibility of repetitive use of evaluative affixes. The semantic category value is based on the presence of semantic categories in the EM of a language. Each semantic category is assigned one point. If a language, for example, expresses the categories of the Quantity of Substance and the Quantity of Quality, the presence of two semantic categories is assigned two points; if three categories are expressed, the score is three; etc. Word classes and the corresponding word-­class value are evaluated analogically. EM saturation is counted separately for diminutives and augmentatives where relevant. The total EM saturation is a sum total of the diminutive (dim) and the augmentative (aug) saturations. 8. The domain of syntax is omitted for two reasons. Firstly, the topic under discussion is morphology. By implication, no relevant universals are expected in this field of linguistics. Secondly, no universal reflecting a correlation between EM and syntax has been found in the literature. However, this is not to say that a correlation of this type may not be identified in future. No universals were identified for the lexicon either. 9. The division of world into language areas, with the exception of SAE, is based on WALS 2005 (Dryer and Haspelmath, n.d.). The results are taken from Körtvélyessy (2012b). Since her work focused on SAE, it was considered an individual language area. Obviously, if its territory is included in Eurasia, this language area would feature the highest EM saturation.

6  The Place of Evaluation within Morphology Nicola Grandi

1 Introduction A cross-­ linguistic survey of formal strategies that convey evaluative meanings (see Stekauer, this volume) shows that they cover a wide range of phenomena, crossing the boundaries of morphology: tonal alterations, sound changes, reduplication, affixation, compounding, etc. Of course, the cross-­linguistic frequency of these formal processes is uneven: affixation decidedly overtakes all the others. Consequently, in the literature ‘evaluative morphology’ has become synonymous with ‘evaluative affixation’, even if evaluative affixes do not cover the whole range of constructions that fall within the domain of evaluative morphology. On the basis of this premise, the aim of this chapter is not to investigate the place of all morphological strategies involved in evaluation within word formation. Since affixation is the formal device with the highest cross-­linguistic diffusion in evaluative morphology, my focus will be on evaluative affixes only. Evaluative affixation has been widely described in the literature in the field. One of the most debated issues, since Scalise (1984), concerns the place of affixes within word formation. All scholars agree in recognising the peculiar nature of evaluative affixes, and in asserting that they cannot be reduced either to inflection or to derivation. In spite of the attention devoted to this issue, the solutions proposed by scholars are all characterised by a sort of vagueness. Beard (1981, 180) states that evaluative affixes are placed ‘somewhere between lexical and purely inflectional forms’. Szymanek (1988, 106–9) ascribes evaluative affixes to an ‘expressive periphery’ of derivation, which does not correspond to any cognitive category. Carstairs-­McCarthy (1992, 107) adopts quite a similar position, asserting that evaluative affixes pertain to ‘expressive morphology’, which is different from derivational morphology, and governed by specific principles which differ from those of derivation. Scalise (1984; 1994) draws the first sketch of the specific properties of evaluative morphology:1 (1)

a. they do not change the category of the base word; b. they do not change the subcategorization frame of the base word; c. they change the semantics of the base word; d. they allow consecutive application of more than one rule of the same type and at every application the result is an existent word; e. they are external with respect to derivational suffixes and internal with respect to inflectional ones;

The Place of Evaluation within Morphology 75

f. they allow, although to a limited extent, repeated application of the same rule on adjacent cycles. (Scalise 1994, 265–6)

Since some properties are shared by evaluative and inflectional affixes (1a–b), and some other properties render evaluative affixes similar to derivational ones (1c–d), evaluative morphology cannot be fully associated with derivation or inflection. Moreover, some features seem peculiar to evaluative morphology. Consequently, Scalise concludes that the best way to explain the behaviour of evaluative affixes is to postulate the existence of a ‘third morphology’ besides inflection and derivation. This assumption has both merits and shortcomings. The main merit is the attempt to provide an exhaustive account of evaluative morphology, sketching a picture of its formal and semantic properties. As an indirect consequence, postulating the existence of a ‘third morphology’, independent from both inflection and derivation, casts doubts on the traditional view of inflection and derivation as of two independent and ‘airtight containers’. The main shortcoming of Scalise’s proposal is its almost exclusive link with Italian: a cross-­linguistic survey shows that the properties listed in (1) are not attested by evaluative constructions of typologically and genealogically different languages (see Stump 1993). Secondly, the idea that there is a ‘universal’ place for evaluative morphology within morphology gives rise to further misunderstandings, concerning the relation between form and function: most studies assume that functional identity has, as its natural consequence, formal identity. So, if two affixes in two different languages perform the same function, they must share the same formal behaviour. As a consequence, since evaluative affixes undeniably perform the same function, their place within morphology must be the same in all languages. Once again, the data contradict this assumption. The aim of this chapter is to survey the formal properties of evaluative affixes in a sample of typologically and genealogically diverse languages,2 having, in the background, the typical properties of inflection and derivation that will be used as a term of comparison. Of course, this would imply a previous analysis of the notions of inflection and derivation. As is well known, in different theoretical frameworks, different views of inflection and derivation have been proposed. For reasons of space, I cannot recap all the different positions that have emerged in the debate on this issue. I will assume a view of inflection and derivation as opposite poles of a continuum. An affix will be ascribed to them on the basis of the properties it shares with typical inflection or typical derivation. For a concise list of properties of both inflection and derivation, and for further references, see Plank (1994) and Booij (2000).

2  A survey of the properties of evaluative affixes In this section of the chapter, I will survey the properties of evaluative affixes that are cross-­linguistically more recurrent. Then I will compare them with the typical properties of inflection and derivation. 2.1  Categorial neutrality When referring to evaluative affixes, the obligatory starting point is their alleged ‘categorial neutrality’: in the literature the feature of evaluative affixes that is referred to most often is that they do not change the category of the base word. Since one of the typical functions of derivation is transcategorisation, that is, the formal strategy by means of

Nicola Grandi

76

which lexical items move from a part of speech to another, the fact that evaluative affixes are insensitive to the word class of the base has sometimes led researchers to postulate an inflectional nature for them. Basically, ‘derivation . . . is that kind of morphology that serves to create new lexemes, whereas inflection serves to create different forms of the same lexeme’ (Booij 2000, 360). The inflected forms of a single lexeme usually share the same lexical meaning, and differ in their grammatical specifications. Consequently, all of them share the same word class. According to Bauer (1997, 549), the categorial neutrality of evaluative morphology has a straightforward functional explanation: ‘a noun which is noted as being of a particular size is still a noun; a noun which is stated to be liked or disliked is still a noun; an adjective which does not apply with its full force still remains an adjective’. So we can assert that evaluative morphology has the primary function of creating different forms of a lexeme; but it differs from inflection in not being contextually determined.3 In other words, while an inflected form of a lexeme can be required by a specific syntactic context or can change the form of other words to which it relates, evaluative forms of a lexeme are usually a free choice of the speaker, as will be reported more extensively in Section 2.11. For the moment, it suffices to assert that evaluative affixes usually perform a modification function that can be compared to that of an adjective within a noun phrase: like an adjective, an evaluative affix modifies (vs. changes) the lexical meaning of its nominal head.4 Categorial neutrality is the natural consequence of this function, and cannot be compared to the similar property of inflection. To conclude this section, it is worth mentioning that evaluative affixes can ­sometimes change the word class of the base. This always happens when different semantic values merge in a single affix. This situation is illustrated by the following Italian data: (2)

mangi-­one to eat.v-­aug/pej.ag.n.m5 ‘big eater, glutton’



facil-­one easy.adj-­aug/pej.ag.n.m ‘superficial person’ In these forms, an augmentative/pejorative nuance accompanies an agentive meaning: these semantic values are expressed by the same formal exponent. It must be noted that the suffix -­one is the same used in typical augmentative formations (cf. gatto ‘cat’ > gattone ‘big cat’). 2.2  Violation of the Unitary Base Hypothesis According to Aronoff (1976, 48), the syntactic-­semantic specification of the base of a derivational affix is always unique. In other words, an affix selects words belonging only to a part of speech. Scalise (1994) proposes a modified version of this constraint, stating that verbs and nouns are incompatible within the domain of a single affix: an affix can be joined to adjectives and nouns, or to adjectives and verbs, but not to verbs and nouns. The data reveal that there are many exceptions to this restriction, even in its modified version. Incidentally, the Unitary Base Hypothesis indicates a widely attested tendency in derivational, and to a lesser extent inflectional, morphology, even though we cannot

The Place of Evaluation within Morphology 77 consider it as a categorical condition on word-­formation processes. Once again, it is quite simple to provide a functional explanation of this tendency. For example, if a suffix is the formal exponent of a process which forms, say, action nouns, the natural candidates to fill its domain will be verbs, rather than nouns and adjectives. Furthermore, if a suffix is the formal exponent of a process which leads to the formation of quality nouns, its domain will be filled by adjectives; and so on. What matters in the present analysis is that, because they are insensitive to the word class of the base, many evaluative affixes systematically violate this tendency, as the following French (3a) and Dalabon (3b from Ponsonnet and Evans, this volume) data reveal: (3) a. coffr(e)-­et chest, box.n.m-­dim.n.m ‘small chest, small box’ long-­u-­et long.adj-­intf-­dim-­adj ‘a bit long’

vol(er)-­et-­er fly(inf).v-­dim-­inf.v ‘flutter about’

b. Kanh kirdikird=wurd buka-­h-­ naHna-­n kakkak-­no. dem woman=dim 3sg>3 sg.h-­r-­look.after-­prs grand.mother-­3sg.poss ‘This young woman looks after her grandmother.’ Kardu ngorr bula-­h -­kurlh-­kurhka =wurd, maybe 1pl.incl 3pl>1-­r-­incep.r-­visit:prs=dim kardu bala-­h-­men-­yin djehneng kardu ngungurru-­kolhngu-­n wah. maybe 3pl-­r-­ideas-­say:prs looks.like maybe 1pl.incl>3:appr-­drink-­prs liquid ‘It seems that [the police patrol] is coming to pay us a bit of a visit, maybe they believe that we might be drinking alcohol.’ Since evaluative affixes usually perform a modification function (see Schneider 2013), the part of speech of the base is almost irrelevant: objects (the typical meaning of nouns), actions (the typical meaning of verbs) and properties (the typical meaning of ­adjectives; see Croft 2000a) can equally be modified, according to the appropriate semantic scale: size, intensity, etc. Therefore we can assert that the meaning of an evaluative affix is usually ‘word-­class-­proof’ in its core; and that it has ‘peripheral’ or additional features that vary according to the word class of the base. So even if the suffix -­et indicates a decrease in both the first and third forms in (3a), in the former case it is a decrease in size, while in the latter case it is a decrease in intensity. 2.3  Changes in the subcategorisation frame Categorial neutrality has often been extended to the subcategorisation frame. A multitude of data gives evidence to disprove this claim. Most of them are presented in the chapters by Grandi (Chapter 7) and Tovena in this volume; for this reason they will not be repeated here and it suffices to mention the well-­known preferential interaction between diminutives and

78

Nicola Grandi

neuter in German and in some Slavic languages, among others; or between diminutives and feminine in some Afro-­Asiatic languages, etc. Sometimes, diminutives change the value of the [± count] feature when they are attached to mass nouns. Moreover, evaluative affixes can change the inflectional class or the noun class of their bases: it is the case, for example, with some Niger-­Congo languages. These affixes can also modify the actional shape of their verbal bases, and so on and so forth. Thus the idea that evaluative affixes cannot interact with the subcategorisation frame of the base is disconfirmed by the data. 2.4  Gaps, restrictions and productivity As implicitly hinted before, the application of evaluative affixes is only occasionally constrained by word classes. This means that in a single language an affix can join words of ­different parts of speech, as shown above; and that, cross-­linguistically, the application of evaluative affixes covers almost the whole range of word classes: nouns, verbs, adjectives, adverbs, determiners, pronouns, numerals, etc. This does not imply that the application of evaluative affixes is totally free, that is ungoverned by any kind of restrictions, or that it is fully productive. Broadly speaking, one may safely state that evaluative affixes are usually more productive than derivational ones, but less productive than inflectional affixes. Their application seems constrained by a complex set of different restrictions, on two different levels: (i) restrictions delimiting their domain of application, and (ii) restrictions governing the application of the affixes within this domain. In both cases, wide cross-­linguistic variation can be observed. Because of space l­ imitations, it is not possible to draw an exhaustive sketch of all the restrictions attested in ­different languages, and therefore only the general nature of such restrictions will be considered. Further elaboration is available to readers through the numerous examples in the monographic chapters of the second part of this volume. As for the first level, for example, saying that nouns are usually the core domain of evaluative formation in all languages with a rich evaluative morphology does not mean that all nouns can be diminutivised, augmentivised, etc. The core area of evaluative ­formation in nouns is represented by bound nouns;6 or, as proposed in Grandi (2002; Chapter 7, this volume), by nouns that can be pluralised. On the other hand, verb evaluative affixes are normally found on durative, dynamic and atelic verbs (see Grandi 2009; Tovena, this volume). Similar constraints can also be found in other parts of speech; and so on. On the second level, we encounter phonological, morphological, semantic and pragmatic restrictions. These restrictions act on the domain of evaluative formation delimited by the previous constraints and often explain the choice between rival affixes. As far as these restrictions are concerned, it is worth mentioning the case of French alouette ‘skylark’, the diminutive form of which, *alouettette, does not occur because it is not possible to place the suffix -­ette on words ending with an identical or very similar string of sounds. For the same reason, the Italian word letto ‘bed’ chooses -­ino as its diminutive suffix instead of -­etto (lettino vs. *lettetto). As far as morphology is concerned, we find restrictions on the subcategorisation frame (for example, the strong preference of Modern Greek augmentatives in -­a for neuter bases) and on the internal structure of the base (evaluative formation tends to prefer simple words to complex (derived and compound) words), among others.

The Place of Evaluation within Morphology 79 Moreover, evaluative formation can rarely, if ever, be applied to the [+ learned] strata of the lexicon (in Italian, vestito ‘dress’ can be both diminutivised and augmentivised, as vestitino and vestitone respectively; its learned synonym indumento cannot: *indumentino/*indumentone). To generalise, we can conclude that evaluative formation displays gaps: not all base words that are semantically compatible with evaluative formation give rise to actual evaluative forms. It must be noted that such gaps are also present in the opposite direction: evaluative forms apparently without a base are sometimes attested. This is the case of the so-­ called diminutiva tantum, exemplified by the Dutch form in (4a) (from Bakema and Geeraerts 2004, 1046) and by Afrikaans forms in (4b) (from den Besten 2012, 281): (4) a. theekransje ‘tea party’ b. mandjie ‘basket’ boontjie ‘bean’ 2.5  Paradigms, blocking rule and prefix–suffix neutrality In the previous section, I mentioned the possible rivalry between two, or more, affixes playing the same role. In evaluative affixation, it is quite usual for a single function to have more than one formal exponent. This situation is quite frequent in inflection, which is very often organised in paradigms, but it is not absent even in derivation. The behaviour of rival affixes can be explained by means of the so-­called blocking rule: when an affix can be attached to a word, a rival, say synonymous, affix cannot be attached to the same base. In other words, two productive rules of word formation that perform the same semantic operation cannot operate on the same base words. This condition explains the distribution of the following Italian forms: (5) a. *amputamento amputazione ‘amputation’ *amministramento amministrazione ‘administration’ *ammiramento ammirazione ‘admiration’ b. ammaccamento *ammaccazione ‘denting, bruising’ accadimento *accadizione ‘event’ ampliamento *ampliazione ‘extension’ (Scalise 1994, 224) Evaluative affixes systematically violate this constraint. Many languages have rich and complex systems of evaluative affixes at their disposal, and very often synonymous affixes

Nicola Grandi

80

can be attached to the same base: in these cases, their denotative meaning is more or less the same and there can be possible pragmatic differences: (6) a. French

îl(e)-­et b. Romanian îl(e)-­ette îl(e)-­ot island-­ dim

càl-­uşél càl-­uşór càl-­út horse-­ dim

c. Portuguese

livr(o)-­eco d. Spanish livr(o)-­ete livr(o)-­inho livr(o)-­ito dim book-­

pajar(o)-­ico pajar(o)-­illo pajar(o)-­ito pajar(o)-­uelo bird-­ dim

g. Albanian

voc-­ël h. Bulgarian voc-­ërr voc-­koran voc-­rrak boy-­ dim7

brat-­lé brát-­ec brát-­ce brother-­ dim

The data unequivocally reveal that, within evaluative morphology, rivalry or competition between strategies which are functionally identical or very similar is controlled differently from typical derivational morphology and, above all, from inflection. While different inflectional affixes with the same function are most often in complementary distribution, evaluative affixes can be interchanged, usually with slight differences in the final meaning. Similar situations are attested in Niger-­Congo languages. Shona (see Déchaine, Girard, Mudzingwa and Wiltschko, this volume) has three series of diminutives, which can join the same lexical morpheme: (7) a. mù-­kómáná rù-­kómáná cl1-­boy cl11-­boy ‘boy’ ‘thin skinny boy’ b. ù-­tánhó cl1-­ladder ‘ladder’

kà-­kómáná

cl12-­boy

‘small (puny) boy’

chì-­kómáná cl7-­boy ‘small (sturdy) boy’

rù-­tánhó kà-­tánhó chì-­tánhó cl11-­ladder cl12-­ladder cl7-­ladder ‘thin scraggly ladder’ ‘small (puny) ladder’ ‘small (sturdy) ladder’

Competing evaluative class markers are attested also in Lega (see Di Garbo 2013, 123): (8)

mu-­ntu

cl1-­person

‘person’

ka-­ntu

cl11-­person

si̧ -­ntu

cl19-­person

‘small person’ ‘tiny person’

The data show that evaluative formation is not organised into paradigms in the same way inflection is: the distribution of items that fill the slots is highly restricted in inflection, while it seems quite free in evaluative affixation. In the previous examples, languages show a competition between different linguistic items that perform the same function and, crucially, that reproduce the same formal strategy. In evaluative morphology, competition goes beyond this, realising what Grandi and

The Place of Evaluation within Morphology 81 Montermini (2005) called ‘prefix–suffix neutrality’. It is well known that prefixes and suffixes can equally serve to express both inflectional and derivational categories, and that the same semantic-­functional value can be expressed, in different languages, by either strategy. For example, plural is suffixal in Italian and prefixal in Swahili; English forms agent nouns through suffixation, Malay through prefixation; etc. Thus there is a wide range of cross-­linguistic variation, but language-­internal variation is almost unattested. In other words, in a single language a derivational or inflectional class is usually expressed either by prefixes or by suffixes, but not by both of them simultaneously. Evaluative formation is again exceptional: a single language can have both evaluative prefixes and suffixes; moreover, a single evaluative function can be formally expressed both by prefixes and by suffixes even with the same base word; this is the case, for example, of the Italian forms represented in (9) (9)

appartament-­o flat-­m.sg ‘small flat’



appartament-­in-­o flat-­dim-­m.sg ‘small flat’



mini-­appartament-­o dim-­flat-­m.sg ‘small flat’



‘Prefix–suffix neutrality’ has quite a cross-­linguistic diffusion: it is widely attested in the Indo-­European languages of Europe (Romance languages, Modern Greek, etc.); it occurs in some Altaic languages (cf. Finnic diminutive prefix pikku-­and diminutive suffix -­nen), in Afro-­Asiatic (cf. Berber diminutive circumfix t___t and diminutive suffix -­ush), and, with a high frequency, in Niger-­Congo languages (where recent evaluative suffixes have emerged beside original evaluative class mark prefixes).8 2.6 Neutralisation The Lega forms in (8) display an interesting feature of evaluative formation: neutralisation. Competition between rival affixes is sometimes neutralised in specific contexts. Di Garbo (2013, 123–4) notices that in Lega both singular diminutives, class 12 ka-­and class 19 si̧ -­, pluralise shifting to class 13, the prefix of which is tu-­. Consequently their difference is neutralised in the plural. This property sometimes occurs in inflectional categories. In Alamblak, for example, the masculine third person singular ending -­r and the feminine third person singular ending -­t merge into one genderless ending in the dual, -­f, and in the plural, -­m: yima-­r person-­3sg.m ‘man’ yima-­t person-­3sg.f ‘woman’



(10)

Nicola Grandi

82



yima-­f person-­3du ‘two persons’



yima-­m person-­3pl ‘people’9

2.7  Consecutive application and recursivity Rival affixes not only are interchangeable, as observed above, but also can co-­occur serially. This feature has already been noted by Scalise (1984): evaluative affixes allow consecutive application of more than one rule of the same type. In Venda (11a) and Sɛlɛɛ (11b) (Agbetsoamedo and Di Garbo, this volume) two different formal strategies coexist in the same complex word: (11) a. khali thi-­kal-­ana cl9.clay.pot cl7-­clay.pot-­dim ‘clay pot’ ‘very small clay pot’ b. kansiɛ cl1.bird ‘bird’

ka-­kansi ɛ-­mii cl3-­bird-­dim ‘small bird’

In Muna (van den Berg, this volume), the combination of the prefixes ka-­ and poka-­ with full reduplication indicates diminution on nominal bases (12a) and attenuation on verbal bases (12b) respectively: (12) a. golu ka-­golu-­golu ‘ball’ dim-­ball-­red ‘small ball’ b. buri no-­poka-­buri-­buri ‘write’ 2sg.r.sbj-­att-­write-­red ‘just write a little’ In Udihe (Tolskaya, this volume), two diminutive suffixes can occur in the same complex word: (13) kuliga-­sa-­ziga snake-­dim-­dim ‘small snake’ In Inuktitut (Compton, this volume), ‘evaluative morphemes can be stacked together’, as illustrated in (14), ‘in which four evaluatives modify the noun meaning “computer”’: (14)





qarisaujaralaakulutuqannguaq qarisaujaq-­ralaaq-­kuluk-­tuqaq-­nnguaq computer-­small-­adorable-­old-­pretend ‘an old adorable small pretend computer’ (such as a toy computer)

The Place of Evaluation within Morphology 83 In Catalan (Bernal, this volume), forms such as those presented in (15) are widely attested: (15) rei-­et-­ó king-­dim-­dim ‘teeny king’

petit-­on-­et small-­dim-­dim ‘very very small’ They reproduce a pattern which is highly frequent in almost all Romance languages, in some Slavic languages, and in Modern Greek (see Melissaropoulou, this volume).10 Dressler and Merlini Barbaresi (1994, 99) mention some interesting examples from Italian, where consecutive application of evaluative suffixes can be described as an instance of recursivity, which has no parallel in derivation and inflection: (16) guf(o)-­in-­o-­in-­o-­in-­o howl(m.sg)-­dim-­m.sg-­dim-­m.sg-­dim-­m.sg ‘very small howl’ fett(a)-­in-­a-­in-­a-­in-­a slice(f.sg)-­dim-­f.sg-­dim-­f.sg-­dim-­f.sg ‘very small slice’11 Incidentally, it must be noted that in the recursive application of the suffix the usual (and expected) final vowel deletion does not take place. Sometimes, the final suffix of a sequence can assume the status of an autonomous word: (17) un poder(e)-­in-­o proprio in-­o art.indf.m.sg estate(m.sg)-­dim-­m.sg really dim-­m.sg ‘a really small little estate’ (Dressler and Merlini Barbaresi 1994, 102) Instances of recursivity are also attested, among others, in Shona (see Déchaine et al., this volume): (18)

chì-­chì-­nhù cl7-­cl7-­thing ‘small sturdy thing’

zvì-­zvì-­nhù cl8-­cl8-­thing ‘small sturdy things’ The data presented in this section and in Section 2.5 reveal that redundancy is a peculiar and habitual feature of evaluative morphology, both at a syntagmatic and at a paradigmatic level.

84

Nicola Grandi

2.8  Interfixes Affixes that occupy intermediate slots in sequences such those described in the previous section sometimes weaken (or lose) their semantic value, reduce their phonetic body and become interfixes. This is the case, for example, with the Modern Greek forms in (17): (19) mathit(ís)-­ar-­oúdi student(m)-­intf-­dim.ntr ‘little student’

mil(o)-­ar-­áki mela(ntr)-­intf-­dim.ntr ‘small apple’

myl(os)-­ar-­ákos mill(m)-­intf-­dim.m ‘small mill’ All suffixes (-­oúdi, -­aki, -­ákos) are preceded by the string -­ar-­, which probably originated from an independent augmentative suffix (-­ára, -­aros, -­arás; see Grandi 2002, 142–3). At present, this string makes no contribution to the whole meaning of the complex word: it is semantically empty. As Dressler and Merlini Barbaresi (1989b) maintain, the presence of interfixes is a specific feature of evaluative formation. Of course, not all of them are the outcome of previous evaluative affixes. A vast majority of these interfixes can be described as readjustment rules, either predictable or, in most cases, totally unpredictable: (20) Portuguese cas(a)-­ar-­ão house(f)-­intf-­aug.m Spanish mujer-­c-­ita woman-­intf-­dim Romanian băsmă-­l-­úţă handkerchief-­intf-­dim French lion-­c-­eau lion-­intf-­dim Dressler and Merlini Barbaresi (1989b) draw an exhaustive sketch of the properties of these linguistic items. In a cross-­linguistic perspective, they show a strong preference for suffixation against prefixation; areally, they display a patchy distribution, with the highest degree of occurrences registered in many Indo-­European languages of Europe (Romance languages, Modern Greek and, to a lesser extent, Slavic languages). 2.9  Order of affixes According to Scalise (1984; 1994) evaluative suffixes are external with respect to derivational suffixes and internal with respect to inflectional ones. This tendency is abundantly confirmed by data in a wide cross-­linguistic perspective, but some manifest exceptions are attested.

The Place of Evaluation within Morphology 85 In Plains Cree (Wolvengrey, this volume), the diminutive marker precedes some inflectional endings (21) a. niyōhtēnēn > niyōhcēnasin ni=yōhtēn+ē-­n ni=yōhtēn+a-­si-­n 1=open+thm-­sap 1=open+thm-­dim-­sap ‘I open it’ ‘I open it a bit’ b. ē-­yōhtēnamān > ē-­yōhcēnasiyān ē-­yōhtēn+am-­ān ē-­yōhtēn+a-­si-­yān comp-­open+thm-­1sg comp-­open+thm-­dim-­1sg ‘(as) I open it’ ‘(as) I open it a bit’ c. yōhtēna > yōhcēnasi yōhtēn+a yōhtēn+a-­si open+thm.imp open+thm-­dim.imp ‘open it!’ ‘open it a bit!’ In Inuktitut (Compton, this volume) evaluative markers usually occur between nominal roots and inflectional morphemes. However, it is possible for some evaluative morphemes to appear between inflectional endings: cf. -­kuluk ‘dear, adorable’ in (20): Alaana takuqqaujakuluga Alaana taku-­qqau-­ja-­kuluk-­ra Alana(abs.sg) see-­prox.pst-­decl.tr-­dear-­1sg.3 sg ‘I saw dear Alana.’ (22)

In Udihe (Tolskaya, this volume) the diminutive suffix -­ziga follows the plural suffix -­ŋku with adjectival bases: (23) a. uligdig’a-­ŋku-­ziga nice-­ pl-­dim ‘cute (plural)’ b. ŋic’a-­ŋku-­ziga small-­ pl-­dim ‘little (plural)’ Analogous formations are attested in other languages described in Chapter 7 (Grandi, this volume). 2.10  Semantic transparency ‘Another corollary of the more general and productive nature of inflection is that it is semantically more transparent than derivation’ (Booij 2000, 364). According to Bybee (1985) evaluative morphology has a high degree of productivity, and can be described as a general process since it performs a modification function. Nevertheless, evaluative forms often tend to lose their semantic transparency, to become non-­compositional, and to lexicalise. Often, lexicalisation is accompanied by formal clues. As Dressler and Merlini Barbaresi (1994) note, in Italian, gender shift is often a symptom of lexicalisation:

Nicola Grandi

86 (24) sigar-­o sigar(o)-­ett-­a cigar-­ m.sg cigar(m.sg)-­dim-­f.sg ‘cigar’ ‘cigarette’ The same holds for Catalan (Bernal, this volume), among others: (25) illa ill-­ot-­a ill-­ot island.f island.f-­aug-­f.sg island.f-­aug.m.sg ‘island’ ‘big island’ ‘islet’

Lexicalised evaluative words have a huge cross-­linguistic diffusion. Examples can be found in most chapters included in the second part of this volume. What follows is a small random sample of forms taken from them. In Wichí (Nercesian, this volume), diminutive and augmentative suffixes can create new lexemes (that is, not mere variants of a lexeme) that name a different entity from the one in the base: (26) ch’ute-­taj aug ear-­ ‘cup’ yel’a-­taj tapir-­ aug ‘horse’ In Udihe (Tolskaja, this volume) diminutivisation can give rise to a new lexical item: (27) a. zeli-­ziga ‘a fingerling, resembling bull trout’ (zeli ‘bull trout’) b. mo:teli-­ziga ‘Eurasian pygmy owl’ (mo:teli ‘Boreal owl’) c. gau-­ziga ‘a short push pole for stealing up to game’ (gau ‘push pole’) In these forms, the base word and its diminutive counterpart coexist. There are cases where the diminutive form replaces its base, acquiring its meaning and losing all evaluative nuances: (28) a. logoso ‘stag, (elk) bull’ b. jandasa ‘badger’ c. giuse ‘roe’ d. kekese ‘servant’

Negidal: lohočo:n ‘(elk) bull’ Oroch: jandaku ‘badger’ Nanai: giu ‘roe’ Oroch: keke ‘female servant’

As Tolskaja asserts, for words in (28) the original base may be understood only through related languages. Other interesting cases are found in Kwaza (van der Voort, this volume): (29) a. jerexwa=txuhũi jaguar=little ‘cat, little dog’

The Place of Evaluation within Morphology 87 Finally, it is worth mentioning the situation of Slavey (Lovick and Rice, this volume), where lexicalisation is extremely pervasive, and where evaluative affixes show a very strong tendency to change the meaning of the base substantially: (30) ’elá-­ya boat-­ dim ‘canoe’ jíye-­cho aug berry-­ ‘orange’ 2.11  Syntactic relevance ‘An important demarcation criterion often proposed in the literature is that inflection is that part of morphology that is relevant to syntax’ (Booij 2000, 38). An affix is syntactically relevant if: (i) it has syntactic consequences – for example, if it triggers agreement; (ii) it is obligatorily requested by a specific syntactic position. In the first case, the functional values of an affix extend to the syntactic context; in the second case, a word with an affix performing a specific function cannot be replaced by a word where the same function is not performed. Most often, evaluative affixes are not syntactically relevant: they are not obligatory and they do not affect syntactic structures: (31) Italian un indf.art.m.sg ‘a beautiful dog’

can-­e bell-­o dog-­m.sg beautiful-­ m.sg



un

cagn-­ol-­in-­o



un



un

bell-­o

indf.art.m.sg dog-­ intf-­dim-­m.sg beautiful-­ m.sg

‘a beautiful little dog’ indf.art.m.sg

‘a pretty beautiful dog’

can-­e bell-­in-­o dog-­m.sg beautiful-­ dim-­m.sg

cagn-­ol-­in-­o bell-­in-­o dog-­intf-­dim-­m.sg beautiful-­ dim-­m.sg ‘a pretty beautiful little dog’ indf.art.m.sg

Therefore, within a noun phrase, the occurrence of a diminutive on the head noun does not imply the occurrence of a diminutive on its modifiers (such as adjectives). But the opposite situation is also true: a diminutive suffix can be attached to an adjective without occurring on its nominal head. However, if we turn to another language, Gikuyu, we see that evaluative prefixes trigger agreement, so in this case they are syntactically relevant:12

88

Nicola Grandi

(32) a. kamũndũ karĩa kanini little:person(gen.12/13) that(agr:cl.12) little(agr:cl.12) ‘that little person’ a′. tũmũndũ tũrĩa tũnini little:people(gen.12/13) those(agr:cl.13) little(agr:cl.13) ‘those little people’ b. kamũkwa gaka kanini little:strap(gen.12/13) this(agr:cl.12) little(agr:cl.12) ‘this little strap’ b′. tũmĩkwa tũtũ tũnini little:straps(gen.12/13) these(agr:cl.13) little(agr:cl.13) ‘these little straps’ (Stump 1993, 9) If a head noun shifts towards a noun class with a diminutive function, all its modifiers acquire the corresponding agreement markers. As a consequence, contrary to what happens in Italian, not all words which contain a diminutive mark can be replaced by a word where this function is not performed. A similar situation is attested in Shona (see Déchaine et al., this volume). Nevertheless this pattern is not uniformly attested in Niger-­Congo languages. In Sɛlɛɛ (Agbetsoamedo and Di Garbo, this volume), for example, evaluative marks are never indexed on a­ djectives. Moving to a different geographical region, Slavey also exhibits some kind of agreement triggered by an evaluative mark (see Lovick and Rice, this volume). Italian and Gikuyu represent the extreme poles of a continuum in which various intermediate situations are attested. In the Italian data in (31) no kind of evaluative agreement is attested; in Gikuyu, on the contrary, an extensive evaluative agreement system operates. In other cases, some instances of partial agreement are attested. In Berber (see Grandi, Chapter 18.1, this volume), for example, diminutives are formed by shifting masculine nouns to feminine gender, whose formal exponent is the circumfix t___t. As a consequence, if an adjective modifies a diminutive noun, it must be in the feminine. Nevertheless, the agreement is limited to gender and does not involve the diminutive function; in other words, the adjective does not have a diminutive interpretation. In Berber, diminutives on nouns are expressed by an affix in which two meanings merge: feminine and diminutive. One of them spreads to the syntactic context. As a result, a non-­evaluative agreement is triggered by an evaluative-­formation process; thus, one cannot assert that diminutivisation is not syntactically relevant at all.

3 Conclusion The properties of evaluative affixes surveyed in the previous sections do not allow us to answer the question concerning the place of evaluative morphology in a univocal and universal manner, since too many language-­specific factors are at stake. As the data reveal, different languages lead to different considerations when a single property is dealt with. The case of syntactic relevance, often considered a crucial factor in distinguishing between inflection and derivation, is emblematic: there are languages where evaluative formation clearly resembles derivation (Italian), and languages where it is very close to inflection (Gikuyu). In the case of other languages it is not possible to formulate an accurate judgement.

The Place of Evaluation within Morphology 89 Therefore, the inflectional vs. derivational nature of evaluative morphology cannot be treated as a ‘yes/no question’. Nevertheless, this does not mean that some general ­comments on cross-­linguistic recurrent tendencies cannot be made on the basis of the above-­mentioned properties. Firstly, it is necessary to eradicate some misinterpretations. Categorial neutrality, a cross-­linguistic, well-­established feature of evaluative morphology, has often been considered an inflectional feature. However, as observed above, what are being dealt with are similar surface manifestations of different deep phenomena: categorial neutrality cannot be compared to inflection because it requires different grounds in order to be explained. The same holds for the alleged paradigmatic nature of evaluative morphology, once again a feature often associated with inflection. Of course, different items performing the same functions do not represent a paradigm by default. As for the case under examination, the fact that rival evaluative affixes can often be freely interchanged represents a fundamental difference as far as inflection is concerned, where the choice among different inflected forms of a lexeme is usually contextually determined. While some properties are cross-­linguistically too uneven to make reasonable generalisations about (this is the case, for example, with the changes in the subcategorisation frame or with neutralisation), others are more consistent: the presence of gaps, the role of restrictions, partial productivity and an inclination towards lexicalisation are features that characterise derivation more than inflection. Furthermore, evaluative affixes are usually syntactically irrelevant, despite some evident exceptions. In other cases one finds properties that are undoubtedly difficult to place within the dichotomy between inflection and derivation. The high degree of redundancy of evaluative morphology is without parallel both in inflection and in derivation. The consecutive application of rules of the same kind, of which recursivity is an extreme instance, does not occur either in derivation or in inflection. What emerges from this survey is that, in a mere quantitative perspective, evaluative ­formation shares more properties with derivation than with inflection. Moreover, the ­supposed analogies between inflection and evaluative affixation are often deceptive. Therefore a general preference of evaluative formation for derivation cannot be denied. In some languages, this preference is more noticeable than in others. However, the presence of a set of specific properties prevents us from fully including evaluative morphology within derivation, and suggests that the best way to describe evaluative affixes in a cross-­linguistic perspective is to adopt the position proposed by Dressler and Merlini Barbaresi (1992) in their analysis of Italian diminutives: evaluative affixes behave as non-­prototypical representatives of derivation, so their peculiar properties do not legitimate the need for ‘a proper subclass of affixation alongside derivation and inflection’ (1992, 26). This conclusion represents just a tendency among world languages. As previously discussed, situating evaluative affixes on the inflection–derivation continuum of each single language is a ‘particular language fact’, not a universal issue.13

Notes   1. In this case, there is a tacit identification of evaluative morphology with evaluative affixation.   2. The sample is basically composed of languages described in the second part of the volume, with the addition of other languages mentioned in the text.   3. The function of evaluative markers can be described in terms of ‘modification’, and is similar to the function usually performed by adjectives. In his study on diminutive formation, Schneider

90

Nicola Grandi

(2013, 138) states that ‘prototypical diminutives . . . result . . . from a process of modification, in which word class is retained and the meaning just modified’. Furthermore: ‘a standard ­description of diminutive meaning is that the meaning of the base word is essentially retained, and that the semantic component SMALL is added through the diminutive marker. This additional ­component does not change the meaning of the base word, but merely modifies it’ (p. 140).   4. In other terms, an evaluative word is usually semantically subordinate (or hyponymous) to its base word.   5. The sign ‘/’ indicates a simultaneous occurrence of different semantic values.   6. In the qualitative perspective on evaluation the situation is more complex, since pragmatic constraints exceed formal ones; see Merlini Barbaresi, this volume   7. See also Melissaropoulou, this volume, for similar data from Modern Greek.   8. According to Grandi and Montermini (2005), prefix–suffix neutrality is highly attested in VO/Pr languages.   9. In Plains Cree there is an interesting neutralisation of theme markers when a diminutive suffix is attached to a word. Cf. (21) for examples and Wolvengrey (this volume) for an in-­depth analysis. 10. In Ket an interesting simultaneous occurrence is registered. This involves processes belonging to different levels of analysis (voicing and reduplication) within a single word; see Vajda (this volume) for details. 11. This property had already been noted by Scalise (1984; 1994): evaluative suffixes allow, albeit to a limited extent, repeated application of the same rule on adjacent cycles. 12. Glosses, abbreviations, italics, translation are from the source. Note that gen is used here with the meaning of ‘gender’. 13. Fortin (2011) has criticised this view, assuming that in some languages evaluative morphology is ‘unmistakably inflectional’. The data discussed are, in some cases, identical to those presented in this chapter (for Shona, among others). Fortin’s observations do not contradict the conclusion of this chapter, the aim of which is to point out a clear cross-­linguistic tendency: even if some evaluative affixes are undeniably similar to inflection, in the vast majority of cases they tend to be closer to derivation than to inflection.

7  Evaluative Morphology and Number/Gender Nicola Grandi

1 Introduction The aim of this chapter is to focus on the complex relationship between evaluative morphology on the one hand and gender and number on the other, in order to point out and possibly explain the situations that are cross-­linguistically more frequent. Because of the huge amount of available data and the different natures of the phenomena to be discussed, I will provide a descriptive sketch of the phenomena that are cross-­linguistically most widespread. The theoretically relevant notions will be emphasised and relevant bibliographical references used in their analysis will be provided.

2  Evaluative morphology and gender Evaluative affixes often change the gender of their bases. Nevertheless, the relationship between evaluative affixes and gender is not consistent, since gender shifts often go in different directions. Because of practical restrictions, the definition of gender,1 and its assignment, cannot be considered at length. For further information on these issues, see Corbett (1991). Data will be presented according to a division of gender systems into two macro-­ classes: sex-­ based gender systems (Section 2.1) and non-­ sex-­ based gender systems (Section 2.2).2 2.1  Sex-­based gender systems In this section I will take into account only languages where there is a systematic link between an affix or a set of affixes performing an evaluative function and a specific gender value (interactions between gender and non-­affixal evaluative formation are almost absent). As a consequence, examples such as those in (1): (1) Italian tavol(o)-­in-­o table(m)-­dim-­m ‘small table’ scarp(a)-­in-­a shoe(f)-­dim-­f ‘small shoe’

Nicola Grandi

92

will not be considered, since the suffix does not have an inherent gender, as it always preserves that of the base. This is the default situation in many language families and, if we adopt a mere quantitative point of view, the mostly prevalent situation cross-­linguistically. 2.1.1  Diminutives and neuter The interaction between diminutives and neuter gender occurs frequently in many Indo-­ European languages such as German and Dutch, Modern Greek, some Slavic languages, etc: (2) a. Modern Greek lath(os)-­ák-­i mistake(m)-­dim-­ntr ‘little mistake’

gat(a)-­í cat(f)-­dim.ntr ‘kitten’ (Holton, Mackeridge and Philippaki-­Warburton 1997)

b. Bulgarian

brat-­lé brother.m-­dim.ntr ‘little brother’



kníga > knizlé book.f book.dim.ntr ‘book’ ‘booklet’ (Scatton 1984)

c. German

Mütter-­chen mother.f-­dim.ntr ‘little mother’



Händ-­chen hand.m-­dim.ntr ‘small hand’ (Ott 2011)

The data in (2) show that two formal processes are involved in the formation of neuter diminutives. In some cases, a diminutive is formed by a mere gender shift (as in the second example from Bulgarian); in other cases gender changes if a diminutive suffix is attached to the base (cf. German -­chen or Modern Greek -­áki). In order to explain this interaction, it is worth noting that the gender systems of the languages mentioned above rely mostly on formal assignment rules: gender is determined by an inherent property of the noun only in a small number of cases. Therefore it is useless to attribute the determination of gender on the basis of semantic properties in all neuter diminutives: if one wants to understand the reasons for the interaction between diminutives and neuter, it is necessary to focus solely on those nouns where this interaction can be semantically motivated.3 In this context, the starting point is represented by animate nouns. Jurafsky (1996), Magni (1999), Creissels (1999) and Grandi (2011b) demonstrate that the notion ‘young X/X cub’ plays a crucial role in the historical development of diminutive affixes in typologically and genetically different languages.4 As a consequence, the central

Evaluative Morphology and Number/Gender 93 question is why neuter gender can be assigned to animate nouns. In order to investigate this issue I will focus on South Slavic languages. For Serbian and Croatian (the same holds for other Slavic languages too; Slovak, for example), Javarek and Sudjič (1973, 13) state that ‘nouns denoting young creatures are often neuter’. So neuter gender indicates not small animals, but young ones: pups and cubs.5 According to Corbett (1991, 227–8), ‘many Indo-­ European languages assign sex-­differentiable nouns to the masculine or feminine gender as appropriate, while the young of sex-­differentiables – typically young animals which are treated as not yet sex-­differentiable – are neuter’. Neuter gender is the formal correlate of sexual indeterminacy: the sexual characteristics (shape, colour, dimension, etc.) of young animals are usually still not perceivable. This hypothesis is confirmed by Zubin and Köpcke (1986, 151): ‘the sex-­associated genders are used to identify fully differentiated taxa that have concrete imageability including overall shape and specifiable parts, while neut-­gender is used for taxa that do not, and are in this sense undifferentiated’. As it spreads over other semantic areas,6 neuter gender acquires a typical diminutive/ endearing function: (3) Serbian unuk > unuče grandson.m grandson.ntr.dim ‘grandson’ ‘little grandson’

crkva > crkviče church.f church.ntr.dim ‘church’ ‘small church’

In conclusion, since diminutive affixes very often derive from affixes which originally and primarily indicate pups and cubs, and since neuter gender, if present, is usually associated with nouns designating these specific categories, the interaction between diminutives and neuter becomes clear and can be explained unequivocally. When these affixes extend their domain application, with inanimate nouns, abstract nouns, etc. as their base, they maintain a neuter gender which, however, loses its semantic properties. The interaction between neuter gender and diminutives is not restricted to Indo-­ European languages. According to Di Garbo (2013, 127–8) it is also attested in Turkana and Karamojong, among others.7 2.1.2  Diminutives and feminine In languages with a tripartite sex-­based gender system, neuter is usually the default gender for diminutives; in a two-­part sex-­based gender system, the prevalent tendency for diminutives is to maintain the gender of the base. Nevertheless, a systematic interaction between diminutives and feminine gender is attested in some Afro-­Asiatic languages. In languages of the Semitic branch of this family, diminutives are mostly realised by changing vowels within a consonantal template (cf. Modern Standard Arabic kitāb ‘book’ > kutayyib ‘booklet’). However, diminutives can also be expressed by shifting the gender from masculine to feminine. Maltese exemplifies this situation; according to Aquilina (1959, 210) ‘diminutive form . . . is also indicated by the fem ending a attached to the masc noun’: (4) bieb-­a door.m-­dim.f ‘small door’

94

Nicola Grandi

fies-­a pickaxe.m-­dim.f ‘small pickaxe’

zappun-­a hoe.m-­dim.f ‘small hoe’8 Another interesting case, in this picture, is represented by Berber languages, belonging to a different branch of the same family, where the shift to feminine is the only strategy for diminutivisation (see also Mettouchi 1999; Grandi, Chapter 18.1, this volume). Berber diminutives are formed by means of the circumfix t___t (5a), also used in the formation of feminine nouns (5b) and of nouns of unity/singulatives (5c): (5) a. t-­a-­dar-­t dim.f-­fs.sg-­foot.m-­dim.f ‘small foot’ b. t-­a-­froukh-­t f-­fs.sg-­boy.m-­f ‘girl’ c. t-­a-­lmšmaš-­t sngl.f-­fs.sg-­(quantity of) apricots.m-­sngl.f ‘one apricot’ (Laoust 1921; Abdel-­Massih 1971) Consequently all diminutives are necessarily feminine (but not all feminine nouns allow a diminutive interpretation); moreover, only masculine nouns can be diminutivised, with the exception of those designating living beings: in this case, the pattern t-­N [+ animate]-­t refers to the corresponding female. A diminutive interpretation is sometimes possible with masculine animate nouns if there are no perceptual and exterior differences between male and female (aherdan ‘lizard’ > taherdant ‘small lizard’). According to Mettouchi (1999, 221) both the feminine and the diminutive meaning should be secondary and recent extensions of an original meaning of ‘approximation’. A quite similar pattern is attested in Iatmul (see Jendraschek, this volume), where, for inanimate nouns, feminine is assigned to express small size and masculine is either the default gender or an augmentative mark; gender is formally expressed by a complex verb agreement system: ki’viya vaali’-­ka-­di’ (mosquito bite-­prs-­3sg.m) ‘a large mosquito is stinging [me]’ vs. ki’viya vaali’-­ka-­li’ (mosquito bite-­prs-­3sg.f) ‘a small mosquito is ­stinging [me]’. 2.1.3  Augmentatives and masculine The above-­mentioned example from Iatmul introduces the topic of the relationship of ­augmentatives with masculine gender. The first data set is from Berber. As shown in the previous section, a masculine noun can be diminutivised by adding the circumfix t___t. Conversely, an inherently feminine noun can be augmentivised by deleting the initial and final t:

Evaluative Morphology and Number/Gender 95 (6)

t-­a-­bhir-­t > f-­fs.sg-­garden-­f ‘garden’



t-­a-­mar-­t > a-­mar f-­fs.sg-­beard-­f fs.sg-­beard:aug.m ‘beard’ ‘big beard’ (Abdel-­Massih 1971, 116)



a-­bhir fs.sg-­garden:aug.m ‘big garden’

As a consequence, all augmentatives are masculine, and only feminine nous can be augmentivised. As for diminutives, most animate nouns are excluded from the domain of augmentative formation: in these cases, the form without the initial t and final t designates the male. The exception is represented by nouns indicating animals with no sexual dimorphism. Di Garbo (2013, 126) describes a similar pattern from Maasai and Nama: (7) Nama ’om-­s > ’om-­i f house-­ m.aug house-­ ‘house’ ‘big house’ (8) Maasai εnk-­anásh > ɔnk-­anásh f-­sister m.aug-­sister ‘sister’ ‘very large sister’ In both cases augmentatives are formed by means of a shift from feminine to masculine; no evaluative suffix is employed.9 In other languages, a gender shift is the consequence of the attachment of an evaluative suffix. This is the case with some Indo-­European languages (Italian, Spanish, Catalan, Portuguese, Romanian and Modern Greek above all), where quite a systematic interaction between augmentatives and masculine gender is attested. In Italian, for example, 95 per cent of all augmentatives attested in one of the most important dictionaries (DISC; Sabatini and Coletti 1999) are masculine, irrespective of the gender of the base (9a); in Portuguese all forms ending in -­ão are masculine (9b), as are forms in -­ón in Spanish (9c) and -­oi in Rumanian (9d): (9) a. barc(a)-­on-­e boat(f)-­aug-­m10 ‘big boat’ b. agulh(a)-­ão needle:f-­aug.m ‘large needle’ (Allen 1941) c. mascar(a)-­ón mask(f)-­aug.m ‘big mask’ (Falco and Cecchini 1957) d. făt(a)-­oi girl(f)-­aug.m

Nicola Grandi

96 ‘big girl’ (Ştefănescu 1992)

As in the case of the relationship between neuter and diminutives mentioned above, to trace the causes of the interaction between masculine and augmentatives it is necessary to reconstruct the origin and the subsequent development of the suffixes involved: one can expect a semantic motivation for gender shift to be attested in their more ancient occurrences. All suffixes mentioned in (9) are the outcome of the Latin suffix -­(i)o, -­(i)ōnis used to form animate masculine nouns designating people with some loud and showy characteristic, whether physical or not, or people performing an action in an exaggerated way; these forms often add a pejorative nuance to this ‘agentive’ meaning: (10)

ālĕ(a)-­o dice game.n(f)-­n.pej/ag.m ‘gambling house keeper, gambler’



ment(um)-­o chin.n(ntr)-­n.pej/ag.m ‘a person with a protruding chin’

blătĕr(āre)-­o to blab, to waffle.v(inf)-­n.pej/ag.m ‘chatterbox, chatterer, blabbermouth’11

At a later stage the suffix changed its semantic value: it no longer designated a person with some showy characteristic; it indicated the characteristic itself (Lat. mento ‘a person with a protruding chin’ > ‘a protruding chin’), and acquired a plain augmentative value. The relation with agentivity and animacy vanished and masculine gender could be considered the last trace of this.12 A partially similar situation is attested in Modern Greek, where many augmentatives are masculine, irrespectively of the gender of the base:13 (11) a. dont(i)-­ás tooth(ntr)-­aug.m ‘big tooth’ b. mýt(i)-­os nose(f)-­aug.m ‘big nose’ (Sotiropoulos 1972) In (11b) the augmentative is expressed by a mere gender shift: this situation is not attested in Romance languages (at least in their standard varieties). In (11a) the masculine gender is a ‘side effect’ of an evaluative suffix, -­ás, the genesis of which closely resembles that of Latin -­(i)o, -­(i)ōnis. Its antecedent is an agentive suffix (or, better, a limited set of etymologically related suffixes) used to form masculine animate nouns designating human beings with some showy physical characteristic, or used to perform an action in an exaggerated way (for more data see Chantraine 1933):

Evaluative Morphology and Number/Gender 97 (12) Ancient Greek pōgōn-­ías beard.m-­pej/ag.m ‘bearded man’

gynaik(eîos)-­ías feminine, womanly(adj.m)-­pej/ag.m ‘effeminate man’

The same reinterpretation illustrated for data in (9) and (10) also took place in this case. Once again, masculine gender could represent the residual trace of an original animacy hallmark. 2.1.4  Augmentatives and feminine Modern Greek also attests a link between augmentatives and feminine gender, which has quite an extensive diffusion, as will be discussed at a later stage: (13) a. kefál(i)-­a head(ntr)-­aug.f ‘big head’ b. fid(i)-­ár-­a snake(ntr)-­aug-­f ‘great big snake’ (Sotiropoulos 1972; Daltas 1985) The vast majority of feminine augmentatives are derived from neuter nouns, both through a mere gender shift (13a) and through suffixation with a consequent gender shift (13b). The most reliable hypothesis on the origin of these forms (and on their semantic motivation) assumes a link between feminine augmentatives and the old neuter nominative and accusative plural ending (-­a in Ancient Greek and Latin). As is well known, this ending is, in its turn, the outcome of an Indo-­European morpheme used to form collectives (Heilmann 1963, 149). According to this hypothesis, an affix used to designate a collection of items can develop a secondary augmentative meaning (Hasselrot 1957, 156; Volpati 1955, 4). Feminine augmentatives are also attested in Romance languages (Hasselrot 1957, 135–40): (14) Spanish huert(o)-­a garden(m)-­aug.f ‘big garden’ Italian buc(o)-­a hole(m)-­aug.f ‘big hole’14 Feminine augmentatives in Romance languages on average display a low degree of productivity.15 The relationship between augmentatives and collectives will be presented at a later stage, in Section 3.1.

Nicola Grandi

98 2.2  Non-­sex-­based gender systems

In gender systems based on criteria other than sex (usually referred to in the literature as noun class systems), a noun is assigned to a given class on the basis of specific features of its referent, such as animacy, shape, size, (ir)rationality, etc. Of course, classes often only have a light or superficial semantic characterisation, and placing a noun among members of one class or another is often merely conventional. For a general overview, see Allan (1977), Craig (1986) and, once again, Corbett (1991). As a general premise, formal markers of noun classes (usually affixes) can have a primary or a secondary function. An affix that expresses a class to which a stem inherently belongs, according to its basic meaning or to its form, is primary. This stem can sometimes shift to another class, acquiring an additional meaning: the affix that expresses this class plays a secondary function. It either replaces the primary affix (replacive affixes; see Di Garbo, forthcoming) or is used beside it (additive affixes; see Di Garbo, forthcoming; see also Déchaine, Girard, Mudzingwa and Wiltschko, this volume). Usually, evaluative formation is performed by secondary affixes; this means that nouns are not classified primarily on the basis of evaluative meanings (size, ­feelings, etc.). These kinds of evaluative-­ formation processes are attested in two macro-­ areas: ­Sub-­Saharan Africa and North America. Bantu languages typically illustrate them. According to Kadima (1969), in a cross-­ linguistic perspective the following classes can have an evaluative meaning: (15)

Diminutive: 2, 7, 8, 11, 12, 13, 14, 19, 20 Augmentative: 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 20, 21, 22

Classes 12 and 13 are the most recurrent among Bantu languages and often ­exemplify the second tendency mentioned above: they usually have only secondary functions. Shona provides useful examples of this evaluative-­formation processes. The stem -­kova ‘river’ inherently belongs to class 11, rukova. In this case, class assignment is semantically motivated, since class 11 preferentially contains nouns indicating narrow, long, thin, tapered objects. If this nominal stem shifts to class 12 or 13, it acquires a diminutive meaning: kakova (cl12) ‘stream’/tukova (cl13) ‘streams’.16 The data in (16) show a sample of possible evaluative variants of the stem -­(ko)mana ‘boy’ (stem variation is not discussed because of space limitations):

(16)

Singular Plural go-­mana ma-­komana cl5-­boy cl6-­boy ‘big boy’ ‘big boys’



cl7-­boy

ci-­komana



cl12-­boy

zi-­komana

cl8-­boy

‘short and fat boy’ ‘short and fat boys’ ka-­komana ‘small, little boy’

tu-­komana

cl13-­boy

‘small, little boys’

Evaluative Morphology and Number/Gender 99



zi-­gomana cl21-­boy ‘huge boy’

ma-­zi-­gomana cl6-­cl21-­boy ‘huge boys’17

In almost all these cases, evaluative prefixes replace primary prefixes. Di Garbo (2013, 124) presents data where secondary prefixes are used alongside primary prefixes: mû-­raata

(17) Gikuyu

cl1-­friend



cl12-­cl1-­friend

‘friend’

ka-­mû-­raata

‘small friend’18

Incidentally, it is worth mentioning that in these languages there are also some interrelationships between natural gender and evaluative morphology. As shown by Creissels (1999), Grandi (2003b; 2011b) and Di Garbo (forthcoming), some African languages developed evaluative suffixes which sometimes replace and sometimes accompany the original prefixes. Two grammaticalisation paths have been reconstructed: the Proto-­Bantu word *jana ‘child’ gave rise to diminutive suffixes (-­ana, -­jana, etc.); the Proto-­Bantu word *kádi ‘woman’ gave rise to augmentative suffixes.19 As Di Garbo (forthcoming, 18–20) points out, the ‘new’ diminutive suffixes sometimes occur as feminine markers in combination with animal names and colour adjectives; and the ‘new’ augmentative suffixes can derive feminine nouns, usually from animate bases.20 Some languages spoken in North America and Canada have quite a different noun class system. In these cases, class markers are realised on verbs, as prefixes or suffixes/clitics, and classify objects by shape, type, material, etc. Dena’ina, an Athabaskan language spoken in south-­central Alaska, described by Lovick in this volume, offers some examples of this gender system. In this language, under certain circumstances, a gender prefix can be replaced by another unexpected prefix which carries an augmentative meaning (this kind of gender shift seems restricted to augmentatives).21 For example, the prefix dn-­, which usually designates heavy and dense objects, can be secondarily used to convey an augmentative meaning. Example (2) in Lovick’s chapter (this volume) reveals that the same verbal pattern, obj+gnd+gh+Ø+qun ‘to sew an object’ (in the example under discussion, the object to sew is a cover for a skin boat), can acquire two slightly different meanings depending on what fills up the gnd slot: a normal-­sized covering is indicated by the Ø-­ gender; an unusual size is expressed by the use of the dn-­gender. For an in-­depth analysis see Lovick, this volume. 2.3  Some general observations To summarise, evaluative meanings can be expressed both by a mere gender shift and by a gender change triggered by an evaluative affix. For non-­sex-­based gender systems, the former seems the only strategy available. In Niger-­Congo languages it can occasionally be reinforced by the application of an evaluative suffix, which is unrelated to gender (cf. Venda thi-­kal-­ana cl7-­clay.pot-­dim ‘very small clay pot’, where thi-­is the cl7 prefix with a secondary evaluative function; Agbetsoamedo and Di Garbo, this volume).

100

Nicola Grandi

As for sex-­based gender systems, both strategies are widely but not equally attested. When diminutives and neuter are concerned, both gender shifts and gender changes consequent on the application of an evaluative suffix are attested; nevertheless, the former prevail over the latter. Among Indo-­European languages, the link between neuter and diminutives is specific to languages spoken in Central and Eastern Europe (Germanic and Slavic languages, Albanian, Modern Greek, etc.). It is also attested in some Nilo-­Saharan languages. As for augmentatives and masculine, mere gender shifts are limited to the Near East and North Africa, with some interesting offshoots in Sub-­Saharan Africa (among others, Nilo-­Saharan and Khoisan languages). Masculine as the effect of an augmentative is attested in Indo-­European languages of Central and Western Europe; its furthest eastern occurrence seems to be found in Modern Greek. Neuter augmentatives are very rare cross-­linguistically; they occur with any regularity only in some Slavic languages (see Gregová, this volume, for some Slovak examples). Feminine is the ‘interface’ between diminutives and augmentatives, since it interacts with both of them. Jurafsky (1996, 544–7) demonstrates that this dual linking is not paradoxical, since it complies with different underlying metaphors.22 In this case, gender shifts and gender changes triggered by an evaluative affix are both attested, though with crucial differences. Where diminutives are concerned, a gender shift to feminine is the only formal strategy available; its core area coincides with North Africa and Near East (Semitic and Berber languages). When feminine is associated with augmentatives, both formal means occur, with a degree of productivity decreasing when moving from Eastern (Modern Greek) to Central Europe (Romance languages). Occasionally, it is agreement that conveys an evaluative nuance (cf. Iatmul above). The gender of an evaluative affix has a semantic motivation in a synchronic perspective only in a few cases; in most other cases, the reasons for the link between an evaluative affix and a specific gender are to be found in the history of the affix involved (see also Mutz, this volume). On the whole, a consistent picture can be drawn: from a cross-­linguistic and historical perspective, diminutives tend to be associated with morphemes that express a low degree of animacy, while augmentatives show a strong preference for morphemes originally used as markers of a high degree of animacy. In this section I have not taken into account qualitative features of evaluation. The reason for this is that possible qualitative meanings of evaluative processes involving gender shift are not different from those that do not involve it: for example, neuter diminutives usually express affection and endearment, but, of course, this is due not to neuter gender itself but to a general tendency of diminutives. Thus, the evaluative-­formation processes observed so far are not exceptional with respect to evaluative processes that do not have any relationships with gender.23

3  Evaluative morphology and number Evaluative formation interestingly interplays with number also, even if this occurs less systematically than with gender. In the following paragraphs I will provide an overview of the relationships most widespread cross-­linguistically between evaluative formation and grammatical features that can be attributed to the category of number, in its broad sense. Once again, limitations of space prevent me from providing an exhaustive definition of number and the range of its possible values. For a general indication, see Corbett (2000).

Evaluative Morphology and Number/Gender 101 3.1 Nouns Number is a grammatical category that affects most parts of speech: nouns, adjectives, pronouns, verbs, etc. Its interplay with evaluative morphology takes place primarily on nouns. The phenomena observed cluster around three focal points. 3.1.1  Cumulative affixes In African languages that adopt non-­sex-­based gender systems, as described above, evaluative affixes surface differently according to number variation: as a result, the form of the singular is always different from that of plural. While in most Indo-­European languages only the inflectional morpheme of a suffix varies when the word it is attached to changes its number value (18a), in African languages the evaluative meaning and the number value merge in a single, cumulative affix (18b). Consequently, when the number value of a word changes, the affix is replaced by a different affix: (18) a. Italian b. Shona

gatt-­in-­o gatt-­in-­i cat-­ dim-­m.sg cat-­ dim-­m.pl ‘small, little cat’ ‘small, little cats’ ka-­kova tu-­kova cl12-­river cl13-­river ‘stream’ ‘streams’

A similar situation occurs in Fula, where different number values are expressed by different suffixes: (19) stem nen ‘person’ neɗ-­ɗo person-­ sg ‘person’ yim-­ɓe24 person-­pl ‘people’ nen-­ŋgel person-­dim.sg ‘little person’ nek-­koyŋ person-­dim.pl ‘little people’



nen-­ŋga person-­aug.sg ‘big person’ nek-­ko person-­aug.pl ‘big people’ (Anderson 1992, 81)

102

Nicola Grandi

In Fula, intertwining between number and evaluative morphology is fully integrated within the noun paradigm: ‘a given Noun has a “paradigm” consisting of a singular, plural, diminutive, augmentative, diminutive plural, and augmentative plural forms . . . What is important to note about this system is the fact that the difference between a basic Noun and its diminutive is formally an exact parallel to that between the Noun and its plural’ (Anderson 1992, 80–2). 3.1.2  Unexpected plurals Derzhanski (2005) shows that diminutives sometimes have unexpected plurals: they display plurals that are not obtained by the regular pluralisation of singular diminutives. In other words, the formation of the diminutive singular and of the diminutive plural follow two parallel paths. In Breton, a singular noun, such as bag ‘boat’, gives rise to a diminutive, such as bagig, which cannot be pluralised. A diminutive plural, bagoùigoù, derives from the plural bagoù, and not from the diminutive singular. In Dakelh (also known as Carrier) too, the plural form of a diminutive singular is obtained not by pluralising it, but by diminutivising the plural of the base word (20a). The same occurs in Southern Barasano (20b): (20) a. lhi lhi-­yaz lhi-­ke lhi-­ke-­yaz dog dog-­dim dog-­pl dog-­ pl-­dim b wi wi-­aka wi-­ri wi-­ri-­aka house house-­dim house-­pl house-­ pl-­dim Another unexpected realisation of plural occurs in Portuguese diminutives, where the plural is double marked, both on the stem (by means of an allomorphic variation) and on the suffix: (21) anel-­z-­it-­o anei-­z-­it-­os ring.m.sg-­intf-­dim-­m.sg ring.m.pl-­intf-­dim-­m.pl lição-­z-­it-­a liçõe-­z-­it-­as lesson.f.sg-­intf-­dim-­f.sg lesson.f.pl-­intf-­dim-­f.pl (Vázquez Cuesta /and Da Luz 1971) In Portuguese, the double marking of the plural is limited to diminutives formed by means of the interfix -­z-­.25 These examples show that the formation of diminutive singulars and plurals is not always a result of serial processes, where the singular is the necessary premise for the plural; on the contrary, the formation of diminutive singulars and plurals is sometimes the outcomes of parallel or even independent processes.26 3.1.3  Evaluative morphology and ‘quantification’ The correlation between evaluative morphology and number that is cross-­linguistically most widespread and systematic concerns the categories of singulative and collective. The correlation between diminutives and singulatives on the one hand, and between augmentatives and collectives on the other, involves both the synchronic and the historical perspective. As shown above, some present-­day augmentative forms originally had a collective meaning. Such a polysemy is not attested in present-­day augmentative suffixes: it seems to be just a historical path.

Evaluative Morphology and Number/Gender 103 In contrast, the polysemy diminutive–singulative is cross-­linguistically very frequent from a synchronic perspective: the same strategy is often used to express the meanings ‘small X’ (with its possible affective nuances) and ‘singular X of a group of Xs’. The Berber data in (5c) typically exemplify this pattern, as do the data in (22), randomly selected from a huge database of forms:27 (22) French grêl(e)-­on hail(f)-­sngl.m ‘hailstone’ Croatian gòvedo > gòveče (Meillet and Vaillant 1952) cattle.ntr cattle.sngl.ntr ‘cattle’ ‘head of cattle’ Moroccan Arabic ngāş > ngēyş-­a (Caubet 1993) ‘pears’ pears.sngl-­f ‘one pear’ Udihe ñukte-­ziga (Tolskaya, this volume) hair-­ sngl ‘one hair’ All these bases are mass nouns (or collectives): the derived words express the individuation of a single item from the masses/groups designated by them. All affixes in (22) also occur as diminutives if the base is a countable noun. If the base word is an uncountable noun that designates a mass without internal structure (where discrete items cannot be singled out), the ‘diminutive/singulative’ word usually indicates a conventional unit of this substance (23a) or has a paucal/partitive meaning (23b): (23) a. Catalan aigü(a)-­et-­a water(f)-­dim/sngl-­f ‘small bottle of water’ Maltese ġobna > ġbejna (Borg and Azzopardi-­Alexander 1997) ‘cheese’ cheese.dim ‘cheeselet’ b. Moroccan Arabic εsəl > εsīl-­a (Caubet 1993) honey.m honey.dim/sngl-­f ‘honey’ ‘a little honey’ Modern Greek psar(i)-­ak-­i (Sifianou 1992) fish(ntr)-­dim-­ntr ‘some fish’ Sometimes, when added to mass nouns, suffixes have a purely evaluative function, usually in order to express a quality. This therefore implies that they do not have a singulative/ partitive/paucal meaning: (24) Italian

birr(a)-­ett-­a beer(f)-­dim-­f ‘a pretty good beer’

104

Nicola Grandi

Portuguese vinh-­o-­z-­inh-­o wine-­m-­intf-­dim-­m ‘pretty good wine’ (Rudolph 1990) This is a quite unexpected interpretation, and it is necessary to point out that the conversion from a mass noun to a countable noun would have taken place before diminutivisation. It is worth noting that the opposite situation is not attested: thus we do not have a singulative interpretation when a diminutive meaning is expected. In other words, the diminutive can enter into the domain of singulative, but the singulative cannot invade the field of diminutive. Therefore it may be asserted that the diminutive acts as a default value of the affixes taken into consideration, while the singulative seems to be the special or marked case. 3.1.4  Evaluative formation and plural: a possible implicational correlation As noted for (24), a diminutive interpretation is also possible when a singulative/partitive/ paucal interpretation is expected, on condition that the base noun has previously acquired the feature [+ countable]. Rainer (1989, 210–11), discussing Italian data, hypothesises that diminutives are restricted to [+ bounded] bases (in Jackendoff’s terms; see 1990; 1991). This restriction can be extended to all evaluative-­formation processes that operate at a quantitative level; with qualitative evaluation, the application of this restriction is probably less systematic.28 Jackendoff (1990, 16) makes use of this feature to give plurals, mass nouns and countable nouns a formal representation. In his framework, plural ‘encodes a multiplicity of a number of entities belonging to the same category’. As a consequence, plural is an operator that can be associated ‘to any conceptual category that admits of individuation’ (1990, 29) and only [+ bounded] entities can be pluralised. The restriction proposed by Rainer (1989) for diminutive formation clearly resembles that suggested by Jackendoff (1990) for plural formation: both of them give boundedness a central role. In this picture, the restriction on (quantitative) evaluative formation can be rephrased as follows: a noun can undergo evaluative formation if it can be pluralised. In other words, only nouns that admit plural can be morphologically evaluated. This generalisation suggests that evaluative formation implies plural. This does not mean that a single language has evaluative morphology if and only if it has plural. However, it does imply that a noun requires a plural form in order to be diminutivised, augmentivised, etc. The data in (24) confirm that uncountable nouns, and so nouns without plural, have to be recategorised in order to be evaluated.29 3.2 Verbs In the domain of verbal evaluative morphology, number often has a central place, since the evaluative counterparts of verbs can be brought back to the category of pluractionality. Pluractionality indicates that one of the participants of the verbal action or the action itself can be considered as plural. It means that the verbal action is performed by several people, or it is performed on several objects, or it is performed several times (see, among others, Dressler 1968). Sometimes, the pluractional value shades off into frequentative/iterative value: (25) Italian picchiare > picchiettare ‘to hit’ ‘to hit repeatedly’ Hebrew kafac > kifcec (Faust, this volume) ‘jump’ ‘jump around’

Evaluative Morphology and Number/Gender 105 Somali sír > dirdír (Lampitelli, this volume) ‘send’ ‘send repeatedly’ Pluractionality can be considered a way in which an action deviates from its standard/ default performance. An action can be quantitatively evaluated if it is actually performed in a way which is different from the manner it is usually carried out (for example, repeatedly, quickly, slowly, etc.; see (25)). It can be qualitatively evaluated if the deviation from the default value arises from the speaker’s feelings (for example, if an action is perceived to be performed in an unusual way). Pluractionality belongs to quantitative verbal evaluation. It must be noted that formal strategies used in verbal evaluation are often identical to those used in nominal evaluation (e.g. the Italian verbal evaluative suffix -­ettare in (25) is identical to the nominal evaluative suffix of libretto ‘booklet’ from libro ‘book’, with the obvious exception of the inflectional ending). An exhaustive picture of the interplay between pluractionality and evaluative morphology is drawn by Tovena (this volume), where the reader may find an in-­depth analysis.30

4 Conclusion In this chapter, I have focused on some seemingly problematic aspects of the behaviour of evaluative morphology. As the data demonstrate, most of these features relate to the categories of gender and number. It is hardly possible to formulate a unique and consistent framework to explain all the phenomena mentioned. As seen above, the relationships between evaluative formation and gender/number do not follow the same principles. Nevertheless, some recurrent features can be singled out. For gender, for example, a relevant part of the phenomena presented deals with animacy: diminutives and augmentatives are associated with morphological items that, respectively, originally indicated a decrease and an increase in animacy. As far as number is concerned, the notion of quantification or, better, delimitation is conducive to the explanation of some of the relationships between number values and evaluative formation.

Notes   1. ‘Gender’ also poses a crucial terminological question. In the literature, it sometimes alternates with ‘noun class’, which is perceived as more neutral, as a cover term to refer to systems of noun classification based mostly on agreement. In this chapter I will use ‘gender’ as the cover term, on the basis of tradition, since it is more recognisable in the typological literature.   2. In these sections, I will not take into account evaluative affixes or, more generally, marks on adjectives as the effect of agreement with an evaluative nominal head. In these cases, no semantic motivation triggers the use of evaluative strategies.   3. This approach requires taking the historical perspective into consideration too. In order to find a semantic motivation in the assignment of neuter gender to diminutives, one must identify nouns where gender is semantically determined from a synchronic perspective, or nouns where gender was originally assigned on the basis of semantic criteria, with a subsequent – often analogical – extension to other nouns. In this case, traces of the original semantic assignment can barely be recognised in a mere synchronic perspective.   4. See also Mutz, this volume.

106

Nicola Grandi

  5. In this case the small size is a consequence of the age, not a permanent state. Small but adult animals have masculine or feminine gender, as in Croatian mrâv (m) ‘ant’, bùva/bùha (f) ‘flea’.   6. See Jurafksy (1996) for the possible semantic extensions of the core meaning ‘young X/X cub’ (including the most obvious, ‘young is small’).   7. Significantly, in Karamojong neuter gender in animate nouns indicates offspring (Di Garbo 2013, 128).   8. See Arbaoui, this volume, for similar data from Classical and Moroccan Arabic.   9. Nama also displays an augmentative suffix, -­kàra, which sometimes combines with the gender shift to express ‘huge, extra-­large size’ (Hagman 1977, 27); Sεlεε also displays a similar word-­ formation process (Agbetsoamedo and Di Garbo, this volume). 10. The final vowel, -­e, is not associated with masculine gender. It marks the third inflectional class of Italian (-­e singular, -­i plural), which includes both masculine (e.g. fiore ‘flower’) and feminine (e.g. volpe ‘fox’) nouns. The only way to establish the gender of these nouns is through the article. The definite article for fiore is il (m.sg); that for volpe is la (f.sg). Barcone requires il. 11. Data from Lazzeroni (1963) and Gaide (1988). 12. It is worth mentioning that the ‘new’ augmentative value does not replace the original value. In most Romance languages, augmentative forms (e.g. Italian gatto ‘cat’ > gattone ‘big cat’) coexist with pejorative/agentive forms (e.g. Italian mangiare ‘to eat’ > mangione ‘big eater, glutton’). 13. See Melissaropoulou (this volume) for more data and further details on Modern Greek evaluative morphology. 14. See Kahane and Kahane-­Toole (1948–9) for details. 15. This is probably due to the fact that Romance languages – unlike Modern Greek – do not preserve neuter gender. In so-­called dialects this pattern is often more productive. 16. These examples and the following ones are from Fortune (1955). 17. See Déchaine et al. (this volume) for an exhaustive account of Shona evaluative morphology and for more data and generalisations on its interaction with gender. 18. For a general sketch of evaluative formation in African languages, see Di Garbo (2013; forthcoming). 19. Matisoff (1991) shows that similar changes took place in some South-­East Asian languages; also in this area, diminutives have been recently renewed through the grammaticalisation of the word for ‘child’; new augmentative suffixes have developed from the word for ‘female’. 20. In Sεlεε gender shift usually combines with a diminutive suffix; see Agbetsoamedo and Di Garbo (this volume). 21. This situation parallels that of Bantu languages and the distinction between primary and secondary prefixes. 22. A quite unusual link between diminutives and masculine is attested in Latvian (see Kalnača, this volume). In Italian some masculine diminutives occur, but, as Dressler and Merlini Barbaresi (1994, 96–7) note, these have often undergone a lexicalisation process. 23. For some general considerations on the interaction between gender and evaluative morphology, see Grandi (2001a). 24. Suppletive stem. 25. See also Gilles (this volume) for Luxembourgish, where the singular diminutive is sometimes formed from a plural base: Kuerf ‘basket’, pl Kierf, dim.sg Kierfchen, dim.pl Kierwercher. 26. Incidentally, the examples in (21), like those in (20), are a counter-­example to the alleged tendency of inflection to occur after derivation (cf. Greenberg’s Universal 28; Greenberg 1963). Cf. also Grandi, Chapter 6 in this volume, for a discussion of this issue.

Evaluative Morphology and Number/Gender 107 27. Among others, Jaqaru, Sεlεε, Ket, Dalabon and Wichí display a similar pattern; see Birioukova and Hardman, Agbetsoamedo and Di Garbo, Vajda, Ponsonnet and Evans, and Nercesian, all this volume, respectively. 28. There are several evaluative words derived from a [− bounded] base that are fully acceptable in specific speech situations or language varieties, such as baby talk. 29. In some languages, a full coincidence between evaluative markers and number markers is attested. In Udihe, for example, the diminutive suffix -­ziga is homophonous with the plural affix (so moxo-­ziga means both ‘cups’ and ‘small cup’; see Tolskaya, this volume). In Warlpiri, the same reduplication strategy is used in both plural and augmentative formation (Bowler, this volume). Another interesting case is that of Apma (Schneider, this volume), where the same reduplication pattern can express both plural and evaluative meanings. Incidentally, these two functions display a complementary distribution: with base forms where plural meaning occurs, evaluative meaning is not attested, and vice versa. 30. Many languages described in this volume have patterns of pluractional verb formation: see Jendraschek for Iatmul, Zamponi and de Reuse for Lule, Bernal for Catalan, Faust for Israeli Hebrew, Pensalfini for Jingulu, Lampitelli for Somali, Wolvengrey for Plains Cree, Bowler for Warlpiri, etc. For a general overview, see Grandi (2008; 2009) and Tovena (2010; 2011a).

8  Evaluative Morphology and Aspect/Actionality Lucia M. Tovena

1 Introduction 1.1  Evaluative morphology and verbs Evaluative suffixes provide a morphological device that can be exploited in the formation of verbs in some languages. Pairs of verbs like leggere ‘read’ and leggiucchiare ‘browse’ from Italian provide a first illustration. The sentences in (1) differ in that the verb is either used in the simplex form or is suffixed with what looks like an evaluative -­ucchi-­ suffix: (1) a. Ha letto il libro. have.aux.3sg read.pst.ptcp.m.sg def.art.m.sg book.m.sg ‘(she or he) read the book’ b. Ha leggi-­ucchi-­ato il libro. have.aux.3sg read-­eval-­pst.ptcp.m.sg def.art.sg book.m.sg ‘(she or he) browsed the book’ Sentence (1a) describes a reading of a book by an unspecified agent. Abstracting away from the peculiarities of each individual event, the hearer is entitled to infer that the sentence is about a canonical reading situation and that the book is fully read. By contrast, sentence (1b) specifies that the reading is done in a non-­canonical way. Leggiucchiare means to read by giving and diverting attention to a text more times within a single event and with less commitment than normal. Attention-­giving spans may also be shorter, with little stretches of text covered at a time. Culmination is less easily inferred. The situation is presented as non-­canonical because of such a form of internal multiplicity and quantitative variation. Morphologically complex verbs, such as in (1b), and reduplicated forms can be viewed as illustrating the phenomenon of pluractionality. 1.2 Pluractionality Pluractionality (Dressler 1968; Newman 1980; Cusic 1981) is the morphological expression of number inherent to verbs, different from number agreement. The plural meaning intended by pluractionality indicates that the type of event in the denotation of the verb is multiply instantiated in some way, because of action by or on more than one participant, and action temporally or spatially distributed, separately over entities or parts of one participant.

Evaluative Morphology and Aspect/Actionality 109 Cusic has identified several conceptual levels that speakers distinguish when grouping events together into pluralities. Pluralisation indicates more than one identical unit of a given level, and does not imply order among subevents (Tovena and Donazzan 2008). Pluralisation of events or larger units is called event-­external, following Cusic. Conversely, event-­internal plurality is a form of repetition within the boundary of one event that, despite its internal complexity, is viewed as a single coherent unit, with no change of participants. This type is associated with ideas of increase and decrease concerning various dimensions of the action. A similar class of verbs is called frequentative in typological studies, e.g. Finnish sataa/sadella ‘rain/rain occasionally’. Under the heading of decrease, one finds verbs describing plural actions carried out with many small phases (diminutive), with insufficient effort to produce the result (conative), with undirected effort (incassative) or with less effort than expected (tentative). Under the heading of increase/intensive, one finds verbs describing plural actions carried out with an increased number or size of parts and deviant in some respect, e.g. in appropriateness (excessive). Subtypes may combine,1 e.g. increased frequency with decreased size of units is exemplified by Sierra Nahuat reduplication kokočisneki ‘continually want to catch naps’ vs. kočisneki ‘want to sleep’ (Cusic 1981; see also Arbaoui on Arabic and Schneider on Apma, both this volume); frequency with decreased result is exemplified by Saho gemination barrar ‘flutter’ vs. barar ‘fly’; frequency with incassative is exemplified by Zoque stem reduplication plus suffixation witwitnay ‘walk aimlessly’ vs. wit ‘walk’ (Cusic 1981). Subtype combination can be viewed as a double focus on the event and on participants – primarily the involvement of agents and affectedness of patients – e.g. possibly repeated action and unachieved desired result in (2), containing Iatmul frustrative marker si’ki’k (Jendraschek 2012): (2) li’-­’at walai-­si’ki’k’-­wun 3sg.f-­dat call-­frust-­1sg ‘I tried to call her’ Modification of the event can be appreciated temporally. Suffixes such as pwali in Iatmul provide information about the duration of the situation and indicate it went on continuously over an undefined period (3) (Jendraschek 2012): (3) ki’-­vwali-­di eat-­ cnt-­3pl ‘they were eating all the time’ Conversely, the marker =mayan in Jaminjung encodes multiple repetitions of an event where the exact number of occurrences is not specified but there is a clear separation between any two of the repeated subevents (Schultze-­Berndt 2012). Productivity can vary. Romance languages use evaluative suffixation and exemplify event-­ internal plurality, e.g. Italian mordicchiare ‘bite.EM’, piovigginare ‘rain.EM’, Catalan ploviscar ‘rain.EM’, ploricar ‘cry.EM’, French mordiller ‘bite.EM’, neigeoter ‘snow.EM’ (Tovena 2011b) and Spanish chupetear ‘suck.EM’ and dormitar ‘sleep.EM’ (Rifón Sánchez 1998). However, this morphological strategy is neither consistently exploited nor extensively used within the group. For instance, French-­based creole of Mauritius exploits stem reduplication instead of affixation. Portuguese does not seem to

110

Lucia M. Tovena

use evaluative affixation productively – for instance, mordiscar ‘bite.EM’ is a rare form – and nor does Romanian. In Italian, the host of affixes productive in the nominal domain is currently shrinking to -­i/u/acchi-­ in the verbal domain (Grandi 2008). In Latin American varieties of Spanish, the suffix attaches specifically to verbal uses of the gerund, as in el viento está soplandito ‘the wind is blowing a little’ (Rainer 1996), rather than forming the full verbal paradigm. Besides the issue of productivity in a language, we must point out that not all verbs can be suffixed. The notion of aspect is essential to set out some generalisations. 1.3 Goals This chapter pursues two goals in parallel. The main aspectual notions and their relevance for a range of languages are broadly introduced. A more in-­depth study then covers aspectual properties of Italian, as representative of evaluative suffix-­taking Romance languages, and shows how evaluative suffixes on verbs alter the telicity2 of the base and reduce the event-­measuring potential of incremental themes. In the final section, we suggest that aspectual changes need not be seen as the effect of direct aspectual modifications, but may follow from the conventional role of evaluatives as degree morphology. Evaluatives convey degree modification by providing comparison functions of disequality associated with ranking, which is not the type of scale required to represent duration. The change of type of scale has more or less visible consequences depending on whether the interpretation of the verb sets the structure of the event in correspondence with another scale. Before we proceed, let’s recall positive evidence that verbs may involve evaluative suffixes, beyond simple formal similarity. Evaluative suffixation in Romance induces an effect of normalisation on inflection that generalises across categories. All pluractional verbs in Italian belong to the first conjugation class, which works as the default, although the corresponding non-­pluractional forms may belong to any class. This uniformity matches what is observed for nouns, namely that all inflect according to the most stable class when suffixed, no matter which inflectional class the base noun belongs to. The same holds for French and Spanish, mutatis mutandis, and was the case in Latin. The situation is not equally straightforward for reduplication.

2  Features for an aspectual classification 2.1 Aspect(s) Aspect has to do with the way events relate to time, primarily in terms of definition of the boundaries and of the internal structure. It is customary to distinguish two levels of aspect. One is lexical aspect, also called Aktionsart, actionality or situation aspect, and is more directly dependent on verbs’ lexical meaning; the other, called grammatical aspect or viewpoint, includes the perfective/imperfective distinction and is generally expressed by grammatical means. Both levels have temporal import. The classification proposed by Vendler (1967), with revised forms in Dowty (1979) and Smith (1991), among others, pertains to the first level and subdivides it into four types. These are: states, which are non-­dynamic, are durative and can hold of points of time, e.g. know; and the three types containing an element of change, namely activities, which are dynamic, durative and atelic, e.g. run; accomplishments, which are dynamic, durative and telic, and have a complex internal structure consisting of process and culmination, e.g.

Evaluative Morphology and Aspect/Actionality 111 write a letter; and achievements, which are dynamic, telic and instantaneous, e.g. discover. A further type is represented by semelfactives, which describe dynamic situations that are instantaneous (semel reading) or durative and atelic (processive reading), e.g. cough. States split into permanent and temporary. State and activity verbal predicates apply to all the subparts of the situation they describe; thus they are homogeneous and not associated with a natural end point. Accomplishments, achievements and semelfactives (in their semel interpretation) describe bounded situations, comprising an end point reached instantaneously or after a gradual unfolding. The second level is concerned with whether the event is viewed in its entirety, including initial and final end points, if they exist, or only partially, i.e. the perfective/imperfective dichotomy, and whether the attention focuses on the inside of the event as in progressive constructions. Evaluative suffixation is sensitive to lexical aspect in Romance, where it attaches only to dynamic and durative bases, such as verbs describing activities and accomplishments, not states and achievements. On the contrary, it attaches almost only to states in Lule (Zamponi and de Reuse, this volume). Reduplication integrated into the system of patterns of root modification in Arabic and Hebrew (Arbaoui and Faust, both this volume), or as a specific process in Somali (Lampitelli, this volume), alters lexical aspect. The former two have pluractional intensive meaning and the latter an attenuative one. Reduplication encoding grammatical (progressive) aspect is illustrated by Rembarrnga (Saulwick, this volume) and the Iatmul suffix ki’va (4) (Jendraschek 2012): (4) ki’-­ki’va ya-­di’ eat-­ intrpt(dep) come-­3sg.m ‘he was eating and coming’ (= his coming happened in the middle of his eating) The presence of different forms of reduplication may serve specialisation. For example, progressive is expressed by CV reduplication and habitual/iteration by lengthier CVC reduplication in Squamish (Bar-­el 2008) and respectively by reduced C+/a/ and lengthier C+/ ̄ah/ in Plains Cree (Wolvengrey, this volume). Apma and other languages of Vanuatu exhibit morphologically conditioned root allomorphy: a progressive prefix conditions reduplication with the secondary root as first copy, while the basic root is used for both copies in reduplication expressing iteration (Inkelas and Zoll 2005). Finally, diminutive morphology is known for contributing to domain discretisation with nouns, a phenomenon called ‘mass to count conversion’ or ‘partitive’ meaning, e.g. Italian pane/panino ‘bread/roll’ and Dutch chocolate/chocolaatje ‘chocolate/piece of chocolate’. A discretising effect in the event domain can be ascribed to perfective aspect, e.g. the preverb ts’a-­ in Georgian contributes perfective aspect and can add the meaning ‘V-­ing a little’, as in ts’asauzmeba ‘eat a little breakfast’ derived from the noun sauzme ‘breakfast’. Here, as well as in cases of action-­verb reduplication in Chinese (sometimes called delimitative aspect), e.g. kan-­kan ‘have a look’, and verb root reduplication in Kíhehe, e.g. ku-­ceenga-­ceénga ‘build a bit’ (Frampton 2009), it is a question whether these forms denote minimally sufficient instances of given event types or reduced instances. Interestingly, modification by German ein bisschen, with degree reading ‘somewhat’ and eventive reading ‘a short time’, has been used as a diagnostic (Maienborn 2005), because the eventive reading applies only to homogeneous eventualities. Last, (perfective +) delimitative + evaluative meaning can co-­occur in event nouns, such as the -­ata nominalisations in Italian (which possibly contain a mark of p­ erfective aspect and output

Lucia M. Tovena

112

double qualitative evaluation: a telic base yields augmentative mangiata ‘big eating’ and an atelic one yields diminutive guardata ‘cursory glance’), the -­azo/-­ón nouns in Spanish, e.g. ­guantazo/guantón ‘slap’ derived from guante ‘glove’, and -­da/-­aço nouns in Portuguese. 2.2  Verbal predicate aspect and nominal referents Aspectual constructions, e.g. progressive, and other factors such as durative modifiers, e.g. for two hours, and locatives indicating source/goal, contribute to aspectual interpretation at clausal level. For instance, the verb run describes an activity that is basically atelic, but sentences like those in (5) refer to delimited situations that cease after the specified distance has been covered. Notice that verbal boundedness always correlates with temporal features of the event: (5) a. He ran a kilometre. b. He ran from the door to the gate. The interdependence of aspectual factors directly related to verbs and quantity factors coming from the nominal domain has been studied in particular within formal semantics, where the dependence of clause-­level aspect on the contributions of various elements has been fully formalised. The relevance of the arguments of the verb has been emphasised by Verkuyl (1972; 1992) and Krifka (1989; 1992; 1998). The referent of a count noun that participates as a whole in a situation throughout does not set a boundary to its duration. The theme is not consumed during the event in (6) and exists before it. (6)

She patted the cat.

The situation is different when a nominal referent (typically a substance) participates in the event incrementally (Dowty 1991; Krifka 1989; 1998; Tenny 1994). A clear example is found in clauses where the extent of the theme delimits the event, insofar as it is used up (either concretely or figuratively) in parallel with the gradual unfolding of the event. The event terminates when the whole referent of the object has participated in it, as in the covering of the wall in (7). (7)

She painted the wall.

An incremental theme can also be quantitatively non-­specified – either because it is a plural nominal (8a, 8c) or because it is mass (8b) – in which cases it does not build telic predicates, but triggers iterative or continuative readings respectively. (8) a. Guests arrived for hours. b. Water poured out of the pipe for hours. c. She ate apples for months. Case can contribute aspectual information. Finnish offers an instance of such semantically conditioned structural case, as the partitive object in ammu-­i-­n karhu-­a ‘shoot-­pst-­1sg bear-­part’ yields irresultative aspect, and the accusative object ammu-­i-­n karhu-­n ‘shoot-­ pst-­1sg bear-­acc’ resultative aspect (Kiparsky 1998). Conversely, dativus ethicus in Italian

Evaluative Morphology and Aspect/Actionality 113 can contribute resultative aspect and empathy, as in mi sono mangiata la bistecca ‘I (fully) ate the steak (to my satisfaction)’. Examples (6)–(8) show that properties of the nominals and type of thematic roles are aspectually relevant. Such a relevance can be represented as the possible dependence between quantitative information in nominals and (a)telicity in events, as defined in the formal system developed by Krifka (1992; 1998). Quantisation is a form of boundedness; more precisely, a quantised predicate is such that if two entities are in its denotation, then neither can be a proper part of the other. This notion is used to identify lexical predicates that are telic, as telic events come out uniquely identified by their boundaries. Cumulativity is used to characterise atelic predicates, where a predicate is cumulative if the sum of any distinct entities in its denotation is also in its denotation. The link between properties of the participants and of the events required to represent the dependence is built via thematic roles, viewed as relations between entities and events within event semantics. Quantisation defined in the nominal domain is transferred to events, so to speak, by showing that a thematic role is a homomorphism from the event to the object whenever it has four specific properties. Two properties define the correspondence between the part structures of event and object: ‘mapping-­to-­subobject’ says that whenever the thematic relation holds for an object and an event, e.g. the object is the theme (like a roll in an eating event), then for every proper subevent of that event there is a proper part of that object that stands in that thematic relation to that subevent (e.g. every subevent of eating involves some bite of roll); ‘mapping-­to-­subevents’ says the same from the reverse viewpoint, i.e. whenever the thematic relation holds for an object and an event, then every proper subpart of the object stands in that thematic relation to some proper subpart of the event. Two properties uniquely correlate parts with wholes: ‘uniqueness-­of-­ objects’ says that whenever it holds that the thematic relation relates an object and an event, then the parts of the object are unique for the subevents; ‘uniqueness-­of-­events’ says that if an object is related to an event by a thematic relation, then no distinct object is related by the same thematic role to the same event. Taken together, these properties constitute the notion of incrementality, because sum or division done on the object produce comparative combinatorial results on the event (and vice versa) – provided we exclude situations where both object and event are atomic – and sum and division are sensitive to the relevance of the boundaries. 2.3  Testing an aspectual characterisation Saying that verbs with evaluative morphology are formed on dynamic and durative predicates in Romance means that they can have a telic base, e.g. leggere il libro, or an atelic one, e.g. vivere ‘live’. A first question is whether telicity remains the same once the suffix is attached. Example (1) indicates that derived verbs with a singular definite object are not clearly telic, contrary to what is the case with simplex forms. The telic interpretation is strongly preferred in (1a), but the reverse is the case in (1b). The change of accomplishments into atelic predicates is confirmed by traditional tests with in/per ‘in/for’ adverbial modification, since per-­adverbial modification becomes natural with no coercion effect (9), i.e. conflict resolved in favour of the meaning of the operator, whereas in-­adverbial degrades, as illustrated also by Hungarian (10), which adds a conative flavour (Kardos 2011): (9) a. Luisa L.

ha leggi-­ucchi-­at-­o eval-­pst.ptcp-­m.sg have.aux.3sg read-­

il

libr-­o

def.art.sg book-­m.sg

Lucia M. Tovena

114

per un’-­or-­a / ?? in un’or-­a. dur indf.art.f.sg-­hour-­f.sg/ in indf.art.f.sg-­hour-­f.sg ‘Luisa browsed the book for an hour/in an hour’ b. Luisa ha lett-­o il libr-­o *per L. have.aux.3sg read pst.ptcp-­m.sg def.art.m.sg book-­m.sg dur un’-­or-­a /in un’-­or-­a. indf.art.f.sg-­hour-­f.sg/in indf.art.f.sg-­hour-­f.sg ‘Luisa read the book for an hour/in an hour’ (10) János 10 perc-­ig /*10 perc alatt olvasgatott for/*10 minute under read.EM J.nom 10 minute-­ könyvet. book.acc

egy a

‘János read a little from a book for 10 minutes’

The test of entailment confirms atelicity (11): (11) a. Luisa stava leggi-­ucchi-­ando il libr-­o L. be.aux.ipfv.3sg read-­ eval-­ger def.art.m.sg book-­ m.sg quando è arrivat-­o il tren-­o when be.aux.3sg arrive.pst.ptcp-­m.sg def.art.m.sg train-­ m.sg → she browsed the book ‘Luisa was browsing the book when the train arrived’ b. Luisa stava legg-­endo il libr-­o L. be.aux.ipfv.3sg read-­ ger def.art.m.sg book-­ m.sg quando è arrivat-­o il tren-­o when be.aux.3sg arrive.pst.ptcp-­m.sg art.m.sg train-­m.sg → she read the book ‘Luisa was reading the book when the train arrived’ As for activities which are atelic, e.g. correre ‘run’, they remain atelic (12): (12) a. Si è rialzat-­o ed ha refl be.aux.3sg get.up.pst.ptcp-­m.sg and have.aux.3sg corr-­icchi-­at-­o per 5 minut-­i /*in 5 minut-­i. run-­ eval-­pst.ptcp-­m.sg dur 5 minute-­ m.pl/in 5 minute-­m.pl ‘(he) got up and run somewhat for five minutes/*in five minutes’ b. Stava corr-­icchi-­ando quando . . . → he ran somewhat be.ipfv.3sg run-­ eval-­ger when ‘he was running somewhat when . . .’ Finally, modification by degree adverbials that require intrinsically unbounded predicates is possible with both activity and accomplishment bases (13): (13) a. Ha corr-­icchi-­at-­o ancora have.aux.3sg run-­ eval-­pst.ptcp-­m.sg still po’.

un

indf.art.m.sg

Evaluative Morphology and Aspect/Actionality 115 some ‘she or he ran somewhat some more’ b. Ha mangi-­ucchi-­at-­o have.aux.3sg eat-­ eval-­pst.ptcp-­m.sg un po’. indf.art.m.sg some ‘she or he ate at the apple some more’

la

mel-­a

ancora

def.art.f.sg apple-­ f.sg still

The loss of telicity recorded by the tests can be understood using the properties characterising incrementality, since the homomorphism from objects to events in telic events with incremental theme follows from the properties of the thematic relation that mediates between event and object. Verbs with evaluative morphology may be seen to have lost some property, e.g. they no longer satisfy mapping-­to-­subevent, because it is not the case that every proper portion of the object corresponds to a part of the event, as illustrated by the conative reading. Atelicity and event-­internal pluractionality also go together in Jaminjung, where overt marking by atelic classificatory preverbs yu ‘be’ and -­ijga ‘go’, and functional overlap with progressive, match event-­internal pluractional interpretation of iterative expressions (Schultze-­Berndt 2012). 2.4  Non-­canonical events We pursue the examination of the aspectual impact of evaluative affixation on verbs by discussing valency modification and a qualitative modification of the event that signal that the event described by these verbs is perceived as non-­canonical. As we have seen, the telicity of an accomplishment is often created by the bounded nature of the entity that discharges the role of theme. Participating incrementally in the event is represented by the correspondence between the mereological structure of participants or scales associated with the event, and the event itself. All dynamic verbs are potentially associated with a scale. Such an association takes place in different ways. A specific path or scale-­denoting result phrase can be associated with an activity verb and make it telic (14): (14) a. Daniel ran to the store. b. Daniel ran himself exhausted. This proposal generalises when one considers property scales measuring an abstract dimension associated with an argument – and thereby with the event – instead of physical entities, e.g. the length of a path, and also scales associated with the event because of unexpressed indirect arguments, absolutive uses or cognate objects. In the canonical case, the unfolding of an event is measured by adjacent isomorphic transitions of the theme along a scale. The scale is related to the event by a thematic relation whereby the temporal adjacency of parts in the event corresponds to scalar adjacency. Considering languages where arguments have a strong aspectual impact, verbs modified by evaluative morphology differ from simplex forms in so far as progression of the event on the temporal dimension no longer correlates with regular progression on a scale, for instance on the scale representing the consumption of the theme. The morphological modification induces a qualitative modification of the event directly, and not by altering

Lucia M. Tovena

116

the participants. As a consequence, its use can be interpreted as indicating that one can no longer rely on information about the scale usually traversed for describing a form of progression in the canonical unfolding of the event. When is verb modification possible and why? Evaluative features of pluractionality cover the involvement of agents and the affectedness of patients. Affected patients are central in Romance verb modification, which primarily concerns verbs with an incremental theme where a property of the object provides the scale, in their simplex form. Suppose, with Rappaport Hovav (2008), that scales require that the participant whose property is measured by them be realised overtly. It follows that for verbs that lexicalise a scale through their valency, the noun phrase of which the scale is predicated must be expressed. But given that suffixation seems to disrupt the homomorphism between the structures of theme and event, and the theme ceases to provide a measure used in the definition of the event, the obligation of overt realisation is waived. In a corpus study on Italian derived verbs, Grandi (2008) has reported that verbs obtained by diminutive suffixation often exhibit a reduction in their valency. Looking at his data, we observe that the reduction results primarily from the omission of the potential incremental theme. Consider next the case where the incremental participation of a nominal referent in a situation is based on modulable or subjective factors, such as an incremental scanning of a nominal referent which is not affected itself or does not undergo a change in the course of the event. Firstly, there are unaffected theme verbs that do not lexicalise the change brought about by the action directly in the theme argument. An example of a verb with unaffected theme is ‘speak a language’. Modification indicates a general reduction in some qualitative dimension of the realisation of the event that seems to depend on world knowledge. Secondly, there are cases in which incrementality seems to be a vacuous notion because the structure of the theme can be assimilated to a point, i.e. an entity with no internal structure. In these cases, suffixation with diminutive morphology is accepted only in some contexts, independent from the issue of register. The verb leggere offers a suitable example. Reading is usually a sequential process, but can also be non-­sequential if the reading is done in one instant, for instance by a scanner, or if the object is a single letter. The suffixed form leggiucchiare is not used when the change expressed by the dynamic process is not incremental; compare the acceptable status of (15) with (16). Recall that atomic events were excluded from the formal characterisation of incrementality: (15) Lo scanner ha lett-­o il codice a barre. def.art.m.sg scanner.m.sg have.aux.3sg read.pst.ptcp def.art-­m.sg bar code.m.sg

‘the scanner read the bar code’

(16) *Lo



scanner

ha

leggi-­ucchi-­at-­o

il

def.art.m.sg scanner.m.sg have.aux.3sg read-­eval-­pst.ptcp-­m.sg def.art.m.sg

codice a barre. bar code.m.sg ‘the scanner read.EM the bar code‘

It is intriguing to observe that decrease interpretation correlates with suppression of the internal argument in Romance and other languages, whereas increase interpretation and transitivisation may share the same morphological tools, e.g. form II in Arabic, the prefix

Evaluative Morphology and Aspect/Actionality 117 for-­in Old English (Elenbaas 2007) and the /-­rra/ morpheme in Jingulu (17) (Pensalfini 1997): (17)

Arduku-­nama jaja-­rru-­mi. last (slow time) wait-­rra-­irr ‘a big mob is waiting for me’

One can understand non-­canonicity as a change from a situation describable by a simplex verb with a thematic role characterised by the properties characterising incrementality, to one describable by a derived form of that verb whose role no longer has some of these properties, or is not discharged altogether. For instance, the event described by mangiucchiare shows a form of discontinuity. Predicates for non-­canonical discontinuous events are construed with at least one theta role that lacks the property of mapping-­to-­ subobjects, by comparison with the description taken as suitable for the canonical case (Tovena 2010). Finally, non-­canonicity is recorded through association with positive or negative subjective judgements possibly ascribed to non-­participants, e.g. the suffixed Catalan predicate allargassar una historia underscores the fact that the story is unduly lengthened, and English fidget, jiggle, etc. are stylistically connoted as negative/informal verbs. Summing up, the tests for aspectual classes have revealed that affixation may induce aspectual modification of the verbal predicate, because the derived verbs are atelic when the simplex forms are not always so. Modification is also often associated with a generalised valency alteration effect. In diminutive cases, the argument that tends to be left unexpressed is the one introducing the nominal referent that participates incrementally in the event described by the simplex verb. Incrementality is crucial, because the extent of the referent of such an argument – the measure taken on it – determines the temporal measure of the event, since the event reaches its end point when the whole entity has participated. When the incremental participant is a bounded entity, the event is naturally understood as aspectually telic. Hence, the empirical data on valency lead us to the same conclusion on atelicity pointed at by the tests, in our case study. Finally, we observed that suffixation is easily available precisely when the base verb has an incremental theme and the measuring-­ out function of this argument is hampered. This last fact does not have a straightforward aspectual explanation. Rather, it points towards a feature proper to evaluatives and present across languages, which is the expression of a form of qualitative modification.

3  A specific contribution by evaluatives In this section, we will try to understand better the link between the semantic content/ syntax of evaluatives and the change in the identification of the boundaries of (single) events. This change could actually come from an altered internal constituency of the event. Typical of the class of derived verbs with diminutive interpretation is the availability of conative and incassative readings and interpretations with some sort of dispersive and fragmentative meaning or the like. The presence of these readings could be the reason why Romance pluractional verbs may be associated with frequentative verb forms productive in languages such as Finnish. This association need not rest on morphological grounds, because, for instance, English frequentatives like nibble, flicker and sparkle, which are no longer perceived as derivational, derive from old iterative affixes -­er and -­le and not from evaluative affixes. Note that Finnish also has forms built with diminutive affixes (18):

118

Lucia M. Tovena

(18) a. Lauri luki kirjan. L. read.pst.3sg book.acc ‘Lauri read a/the book’ b. Lauri lukaisi kirjan. L. read.EM.pst.3sg book.acc ‘Lauri browsed a/the book’ Evaluative morphology can express the function of modifying a value along a quantitative dimension, as well as conveying emotion and value judgement, which show in many shades of meaning. In this last section we sketch out how this association of functions may result in aspectual effects. 3.1 Gradability Evaluative morphology quantitatively alters a gradable property of the referent of the expression it modifies. In this respect, it is degree morphology that can combine with a variety of lexical classes. Adjectives may be the examples of gradable expressions cited most often (see e.g. Bolinger 1972; Kennedy and McNally 2005), and can be used in comparative constructions (19). We can detail several components in comparatives. They provide a degree predicate and a comparison function, respectively lungo and più in (19c); the entity that is being compared is understood as associated with a value relative to a standard for its class of comparison, say a group of boys, in (19a); a comparison value and a differential value can be overtly realised, e.g. respectively di quella in (19b) and due centimetri in (19c): (19) a. Gianni è più arrogante/grande. G. be.prs.3sg more arrogant/tall ‘Gianni is more arrogant/taller’ b. Quest-­a sals-­a contiene più burr-­o di this-­ f.sg sauce-­f.sg contain.prs.3sg more butter-­ m.sg than quell-­a. that-­ f.sg ‘This sauce contains more butter than that one’ c. Quest-­a cord-­a è più lung-­a di due this-­f.sg rope-­ f.sg be.prs.3sg more long-­ f.sg by two centimetri. centimetres ‘This rope is two centimetres longer’ Evaluative suffixes are also understood as expressing a comparison, made along a contextually salient dimension for entities characterised by the base predicate. They differ from other expressions of overt comparison in at least three respects: they provide a comparison function that cannot be equality (20a); the referent of the noun, for example, can be understood as associated to a value inferior/superior to a standard for its class of comparison, however this value cannot be made correspond to a precise degree, and a differential value or a ratio cannot be defined (20b); the comparison value cannot be overtly realised (20c): (20) a. Maria è piccol-­in-­a /grand-­on-­a. M. be.prs.3sg small-­ dim-­f.sg/tall-­aug-­f.sg ‘Maria is very short/very tall’

Evaluative Morphology and Aspect/Actionality 119 b. Quell-­a scatol-­a è grande il doppio/ that-­f.sg box be.prs.3sg big def.art.m.sg double/ *grand-in-a il doppio/*grand-in-a di due big-­dim-­f.sg def.artm.sg double/big-­dim-­f.sg by two centimetri. centimetres ‘That box is twice as big/twice as biggish/two centimetres bigger’ c. *Maria è piccol-­in-­a di Luisa. M. be.prs.3sg small-­ dim-­f.sg than L. ‘Maria is more shortish than Luisa’ Despite the different forms of gradability – (19a) involves a comparison between degrees of arrogance/tallness and (19b) a comparison between quantities of butter – all gradable expressions are understood as introducing scales. Related to these are the entities to which gradable expressions apply, e.g. grande is an expression that relates individuals to their degree of tallness. Scales are abstract representations of measurement, and measurements are assignments of entities to numbers in order to represent gradable properties numerically. The assignments are intended to preserve in the numbers the relationship between entities. When one entity exhibits more of a property than another, measuring associates it with a bigger number. If it is true that this rope is longer than that rope in (19c), the measure of the length of the former is a greater number. Different behaviours with respect to gradability can be captured by referring to different types of scales (Stevens 1946). ‘More’ along an ordering dimension such as height means adding a quantum that can be measured; see (19c), which contains an adjective without evaluative modification. By contrast, for qualities such as arrogance, ranking is acceptable but it seems impossible to agree on a unit of measurement and so on computing a meaningful difference. Units of measurement can be defined only if equal intervals express the same differential regardless of where they are taken from the scale, as with tallness. When the interpretation of intervals is not constant, only ranking is possible, as with arrogance. 3.2  Subjective appreciation In general, what has been said for comparatives applies to evaluative morphology. The ordering relevant to gradable predicates is preserved, because, e.g., the measure assigned to Maria in (20a) is understood as smaller than that assigned to girls in the case of piccolina and greater in that of grandona. However, (20b–c) show that evaluative morphology does not preserve relative differences among values. Rather, it takes away the possibility of interpreting specific intervals on a scale, no matter what scale is lexically associated with the adjective. This reduces the quantitative information expressible by the modified word, but creates room for the value judgement function. Turning to events, recall that duration is always measured with standard units. Evaluatives that attach stative predicates behave like adjective modifiers and have no temporal import. When evaluatives attach to dynamic predicates, a bigger quantum of eventuality may no longer correspond to a longer temporal interval, and here is where non-­canonicity really shows. The lexical content of simplex predicates specify whether aspectual, temporal and thematic information is all related. But evaluatives added onto verbs express comparisons that do not preserve relative differences among values on the event. The property describing the event becomes associated with a simple ranking, not

120

Lucia M. Tovena

with the types of scale associated with participants, which are not modified. As a result, the homomorphism between part structures of entity and event cannot be preserved. The duration of the event is not altered, but the ‘quality’ of the event covering a given time interval is modified – either diminished or intensified depending on the nature/effect of the evaluative. For example, the diminished value of reading in leggiucchiare (1b) is not precisely defined, and this interferes with the computation of a ratio with respect to other measures of the event that are left unchanged, e.g. the length of the text. The mismatch shows in various forms of non-­canonicity depending on the verb, e.g. discontinuity, conativity etc., and on the role of context and pragmatic factors.

Acknowledgements I am grateful to the editors, and to Nicola Grandi in particular, for extremely valuable help.

Notes 1. On the different word-­formation strategies available, the reader is referred to Štekauer, Valera and Körtvélyessy (2012). 2. Telic predicates express an action tending towards a goal that represents their natural end. Atelic predicates do not have to wait for a goal for their realisation, but are realised as soon as they begin.

9  Evaluative Morphology and Sociolinguistic Variation Livio Gaeta

1 Introduction Word-­formation rules have been shown to be generally sensitive to registers and text types (see Baayen 1994 on English). For instance, spoken varieties of English are said to display poorer word formation than the written language (Plag, Dalton-­Puffer and Baayen 1999). This is particularly true of those word-­formation rules whose meaning has a less concrete content, such as processes forming denominal adjectives, deadjectival or deverbal nouns, and so on. In other words, word-­formation rules displaying a more ‘grammatical’ and less ‘referential’ meaning are expected to be less used in spoken, informal varieties. At any rate, few studies have been devoted to investigating systematically the exact extent of this assumption. While Roelcke (1999, 81) observes a significant increase of nominalisations in the written variety of German for special purposes in contrast with the standard – including the spoken – varieties, Iacobini and Adinolfi (2008) find no significant differences comparing written and spoken varieties of Italian. Unfortunately, since they did not investigate any evaluative morphology (EM), their results are useless for our purposes. Plag et al. (1999) fare slightly better because – in spite of the very poor EM of English – they included in their suffix sample -­ish, which proved to be used significantly more extensively in everyday conversations than in other more controlled varieties and in the written language. Moreover, Grandi (2003a) finds that the Italian augmentative suffix -­one is more frequently used in written texts mimicking spoken varieties like comedies or in those parts of narrative texts reproducing spoken situations (typically dialogues), especially when characters of humble origin are involved. This provides support for pointing out an inverse relation: while the usage frequency of more grammatical word-­formation processes, measured in terms of both types and tokens, generally increases as long as the register and/or the text type show(s) an increasing degree of formality, the opposite is true for EM. In what follows we will try to survey the different factors influencing the usage of EM, ranging from text types to social contexts and socio-­pragmatic environments.

2  The role of sociolinguistic factors in evaluative morphology As is well known, sociolinguistic variation relates to several domains and calls into play different factors influencing the language in its concrete use and ultimately shaping the speech act. Here, we will distinguish two different perspectives. On the one hand, we will discuss the role that EM plays in contributing to the general sociolinguistic perspective of identifying the social characters of the speakers involved in a given speech situation. In

122

Livio Gaeta

this light, the traditional (especially diastratic and diaphasic) dimensions of variation will be considered. Under the diastratic heading, all truly sociological variables are considered, such as social group or class, ethnic background, sex, age and so on. Only marginal reference will be made to the diatopic dimension because that is more related to a truly dialectal variation. As for the so-­called diamesic perspective, relating to the communication channel (see Berruto 1995, 123 for discussion), it rather characterises specific linguistic varieties and only indirectly provides information on the speakers’ social attitudes and characters. For this reason, it will be mainly considered in relation to the diaphasic dimension as long as the latter is concerned with the speakers’ field of activity, their working environment and so on, besides the speakers themselves in their own personal sphere of social interaction. In this regard, the enormous increase of written modalities of social interaction recorded in the last decades thanks to the diffusion of the Internet, social networks and so on promises to increase significantly the role played by the diamesic dimension in contributing to the identification of the social characters of the speakers involved in such communicative situations. In other words, the use of certain communicative modalities and their typical linguistic features are likely to provide substantive information on the social character of the speakers using them, and this on a mass scale by virtue of the large user numbers. On the other hand, we will try to see briefly how much EM (as an instantiation of the general domain of word formation) is utilised more or less consciously by the speakers to carry out a number of operations on the speech act which reveal their social character or attitude. This opens the door for the consideration of the morphopragmatic dimension as understood by Dressler and Merlini Barbaresi (1994). 2.1  Sociolinguistic factors in the use of evaluative morphology One general question concerns the real possibility for EM to be influenced by sociolinguistic factors. In fact, it is often claimed that only some aspects of language are likely to be exposed to the influence of the (socially or pragmatically determined) extra-­linguistic context, for instance those relating to deixis (Fasold 1992, 352). In this case, contextual extra-­linguistic – for instance, spatial, perceptual and so on – factors arguably guide the way language concretely encodes reality, and conversely how the latter has an impact on the language structure. On the other hand, other aspects of language more related to the grammar intended as a description of our language competence are normally ­independent of the context, in the specific sense that language-­internal structural relations are not subject to its influence (Romaine 1984, 34). Among the latter one usually includes – besides phonology and syntax – morphology and word formation, although the latter often goes unmentioned. Clearly, this does not exclude the possibility that (evaluative) ­morphology may carry context-­related or more general social meaning, but it is generally held that morphological markers cannot be assigned a direct extra-­linguistic motivation per se, while this is less true for the encoding of, for instance, pragmatic nuances. However, the claim of a principled autonomy in the language structure is only partially true in the case of EM, because of the role played by iconicity and iconic motivation in the structuring of the acoustic signal, which has been emphasised in the literature (see Hinton, Nichols and Ohala 1994b for a survey). In fact, it has been suggested that the phonosymbolic association of phonological features such as [high], [palatal] and so on with semantic features such as [small], [familiar] and related meanings, and of phonological features such as [low], [open] and so on with meanings such as [big], [hostile] etc., especially in contrast with morphemes displaying the former traits, may have a certain socio-­cultural

Evaluative Morphology and Sociolinguistic Variation 123 background. Ohala (1994, 1997) considers this association as due to the particular form taken by the mouth during the articulation of the respective sound types. In fact, the face assumes a smile-­like expression with the mouth channel strengthened and thin during the articulation of palatal vowels, which usually display high values of F0, the so-­called fundamental frequency. The smile-­like expression may be taken to encourage a pacific, non-­aggressive interaction between speaker and hearer. Conversely, low and open vowels normally provided with low F0 values are articulated by enlarging the mouth with an apparently aggressive intent. A similar correspondence of facial expressions, possibly accompanied by concrete sounds, and type of interaction is also found in apes and other animals, which leads Ohala to lay down the hypothesis that this correspondence has provided the evolutionary basis for the development of a cultural feature. If Ohala is right – but it must be added that the evidence in support of this explanation is not watertight; cf. Diffloth (1994) – we can observe a link between the way language is structured and the external world.1 At any rate, besides this single – and debated – case, the Saussurean postulate of the conventionality of language prevents any other possible interaction. 2.1.1  The role of gender In order to understand the role played by factors of a sociolinguistic nature in the usage of EM, let us start by discussing one finding which is generally mentioned as resulting from the impact of social factors on language (see Berruto 1995, 43–4 for a discussion). On average, women are considered to employ more EM (with an attenuative function) than do male speakers. This is valid in qualitative terms, as for instance in Portuguese, in which it is conceivable for a woman to use a diminutivised form of the word for thanking, namely obrigad-­inha ‘thanks-­dim.f’, while this is not possible for a man: *obrigad-­inho ‘thanks-­ dim.m’ (see Cardona 2006, 65–7, among others). In quantitative terms, too, it has been found that, for instance, in Tzeltal women make more use of the diminutive suffix/particle ʔala as a signal of positive-­politeness endearment in conversation (Brown and Levinson 1987, 109, 251). This fact has to be seen in connection with the general tendency manifested by female speakers towards a pronounced politeness in speech situations (Lakoff 1975; Holmes 1995). However, the interpretation of this finding is quite controversial. On the one hand, this can be taken as a classic example of the effect of a socially determined factor influencing a linguistic variety, namely that spoken by women. Since biological sex or gender is a variable normally taken into consideration alongside age, social class and so on, it seems obvious to consider this as a plausible sociolinguistic factor. On the other hand, one can wonder what kind of explanation is provided by the consideration of this factor. To be sure, it is not the case that this finding explains why the girl I am talking to employs more diminutives than I normally do. In this sense, this finding only expresses a quantitative correlation of a probabilistic nature observed in a certain corpus – feminine gender ) more diminutives – which in its turn is in search of an explanation. Notice incidentally that sociolinguistic findings generally refer to a quantitative dimension, pinpointing correlations between social and linguistic features. This means that any predictive statement extracted from quantitative findings is of a probabilistic nature and cannot be interpreted deterministically. Explanations of the correlation can appeal to a number of different factors and principles which illustrate what kind of influence sociolinguistic variables can have on language. For instance, one might think that the correlation depends on the way certain cultures socialise the important tasks of transmitting the affective values of the group and of keeping a favourable communicative mood or interaction style, assigning such a burden to women.

124

Livio Gaeta

On the other hand, women might behave in this way because they want or have to stress their social role, which is – again in certain cultures – to be seen as more related to the care of interpersonal relations rather than to the mere exchange of information. Finally, one might even venture the hypothesis that women are generally rather insecure and therefore prefer to limit the reach of their claims. As can be easily imagined, this is not only strongly culture-­dependent (it is not difficult to imagine a different – matriarchal – society in which women are normally endowed with the opposite properties). What is more, the alleged explanations share the same directionality: the linguistic behaviour is seen as the reflex of a social fact, which is generally of a socio-­historical nature and is formalised in such abstract, loose (and we may add: often simplistic) terms that its explanatory power is not easy to test or to falsify. In this light, its weakness is not only constitutive (because explanations in social sciences cannot be of a strictly deductive-­nomological character; see Lass 1997, 328–9) but also external to language, because there is no real causal relationship between the usage of EM and the expression of a certain social value (and in fact, the attenuative function can be fulfilled by the employment of several linguistic means, such as honorifics, formulaic expressions of politeness and so on). This also means that it seems highly improbable to come up with sociolinguistic universals similar to those normally employed in language typology (Berruto 1995, 44). On the other hand, the explanations provided so far are of a functional nature, in the sense that they point to a teleological (or rather: teleonomical; Dressler and Merlini Barbaresi 1994, 16) motivation which explains women’s attitude as based on the social division of the society. In this light, the pronounced politeness manifested by women in their increased usage of EM responds to a precise social function (on a functional understanding of the social relations, see Mesthrie 2006, 476). In addition, EM serves this social function in perfectly adequate terms because it is able to encode a number of semantic nuances which have less to do with referential effects on the listeners’ mind but warrant their immediate, that is context-­situated, response at the level of the speech act (I will come back to this point later). An important additional factor mentioned above is identity: language is an important mean for stressing personal identity, and as such it is commonly used by any social group for its self-­statement. In this regard, it remains an open question whether women can be considered a social group on a par with others such as teenagers, old people and so on (see Edwards 2009, 132–44 for a recent survey).2 2.1.2  Age as a sociolinguistic variable Besides gender, another important factor influencing the usage of EM is undoubtedly age. This is true of any kind of child-­oriented communication which has been observed in the literature. For instance, in baby talk the role of EM is essential from the earliest stages of acquisition, in the sense that adults employ substantially more EM when talking to children (baby talk, motherese, nursery-­school teacherese and so on) and conversely that children’s average usage of EM is quite high, especially in the initial stages of language acquisition (Clark 1993, 168–9; Dressler and Korecky-­Kröll in this volume). Although diminutive formation has been observed to be among the very first morphological operations acquired by children (Ferguson 1977, 225), this is true of any sort of EM, not only that relating to the so-­called diminutivum puerile (see Staverman 1953 for the term). In fact, the child develops quite early the ability of constructing analogical formations which testify the salience of patterns of EM, exploiting for instance the contrast between diminutives and augmentatives. In this example from Italian, Matteo at the age of three extracts an augmentative form, barbone, on the basis of the false diminutive barboncino (De Marco 2002):3

Evaluative Morphology and Sociolinguistic Variation 125 (1) Matteo and his mother are walking in the street: Matteo: mamma quello è un barbone ‘Mummy that is a poddle’ Mother: perché dici che è un barbone? ‘Why do you say it is a poddle?’ Matteo: perché è un barboncino grande ‘because it is a big poddle.dim’ EM and in particular diminutive formation has even been claimed to be the only kind of word formation available to children in the very early pre-­morphological stage.4 In this light, EM plays an important role in the process of speech accommodation, which is normally understood as the process of attuning the speaker’s communicative style to the interlocutor (Giles and Powesland 1975). Since the adult and the child converge in the high usage of EM, this becomes an important expressive dimension for the child, not only to develop harmonious relations with the present audience but also to explore the world outside of the usual participants. Let us give one brief example of the baby talk typically adopted by Viennese adults to address children. In this variety, a specific ‘childish’ suffix -­i (and its variants -­li, -­tschi) is used, which has the property of leaving unaltered the property of the base (gender, word class), in contrast with other German diminutive suffixes such as -­chen and -­lein that form only neuter nouns (cf. respectively Bruder ‘brother.m’ → Brüder-­chen ‘little brother-­dim.ntr’, Tisch ‘table.m’ → Tisch-­lein ‘small table-­dim.ntr’, and so on). The childish suffix can be combined generally with nouns (Mama ‘mum’ → Mami, Papa ‘daddy’ → Papatschi) and with other word classes, as in these examples from motherese, in which the suffix is combined with adjectives and even verbs (Dressler and Merlini Barbaresi 1994, 106):5 (2) a. Guti guti. Bist doch g’scheiti g’scheiti good:ie good:ie are.2sg yet clever:ie clever:ie ‘Goody goody. You’re a nice little clever child.’ b. Kindi muß brav sitzi, soll jetzt trinki child:ie must well-­behaved sit:ie shall now drink:ie ‘My little baby must now behave, sit quiet and drink his nice milk, come on, darling.’ Notice that the suffix is generally used by adults only with items that the child already knows, or, better, that the adults believe he or she knows (Dressler and Merlini Barbaresi 1994, 105). Moreover, the suffix cannot be followed by further inflectional endings; thus the verb remains uninflected and appears in the so-­called inflective form (Teuber 1998), also typical of other language varieties such as those used in comics, chat-­lines and so on. As can be gathered from the translation, English, which has a very poor EM, displays a similar ‘childish’ suffix -­y/-­ie, enjoying a certain productivity normally restricted to child language or to situations metaphorically recreating a child’s world, especially in hypocoristics (Johnn-­y, Kat-­ie and so on), kinship terms (mumm-­y, grann-­y and so on) and other nouns such as dogg-­ie, bunn-­y, and even adjectives such as fatt-­y, short-­ie. This suffix has also been reported to be typically used by women in expressions such as: Give (me) your hand-­y! (Charleston 1960, 123). Since older people are frequently treated like infants by their caregivers (Ferguson 1977, 230), Dressler and Merlini Barbaresi (1994, 190–1)

126

Livio Gaeta

introduce the term diminutivum senile for the following Italian example, in which an adult woman reproaches an elderly man complaining about the small amount of food he has received:6 (3) Avete anche mangiato un ovino have.2pl also eaten an egg.dim ‘You’ve even had egg to eat!’ However, they observe that there seem to be culture-­specific socio-­pragmatic restrictions involved here, because the use of the diminutivum senile is allegedly blocked in similar situations in Arabic and Chinese because of great respect for old age. Nevertheless, even in these languages this restriction is effaced when speaking about very old, dependent persons in their absence. The role played by diminutives is also important because it illustrates how the primary socialisation of children takes place, by emphasising a smooth approach to our daily experience in which attenuative tuning creates the conditions for a harmonious relation with the adults’ world as a normal expectation. In this regard, it has been observed that this smooth interaction with children characterises humans and to a lesser extent apes as distinct from other species (Tomasello 1999). This attitude largely emerges within mother–child relations and has to be looked upon in connection with women’s privileged usage of diminutives discussed above, taking especially into consideration their role as a primary agent in the children’s developmental path. Notice that the crucial role played by EM in child-­centred speech situations is also important for the understanding of its usage in speech situations characterised by the non-­serious attitude of the speakers involved (see Section 2.2). In this light, it does not come as a surprise to observe that high usage of EM also characterises young people as a social group. In fact, ‘the peer group (e.g., the child’s circle of playmates within and outside the home) soon overtakes the parents as the main reference group in transmitting linguistic behavior’ (Mesthrie 2006, 476). In this connection, it has been emphasised that the high usage of evaluative expressions, besides their expressive strength, has the specific intent of signalling identity as a social group. This is, for instance, clearly the case with new varieties resulting from the multiethnic context characterising the urban development of several European capitals, in which new varieties of youth language have developed in recent years, such as Kiezdeutsch in Berlin (Wiese 2012), Rinkeby Svenska in Stockholm (Kotsinas 1998) or Multicultural London English (Kerswill 2006). These varieties present similar phenomena in particular with regard to the large use of EM, sometimes displaying specific features distinct from other language varieties because of the contact context. For instance, Multicultural London English uses the word nang to express high appreciation: Last night was so nang! The source of the term is unclear, perhaps of foreign (Bengali, Persian) origin. This gives rise to further formations, such as the reduplication nangnang referring to a ‘person who is prone to doing foolish things’ or to an ‘edible pick-­me-­up usually desired after a long night of drinking and smoking up’; the derivatives nangage ‘moment of satisfaction’, nangidge ‘something cool’, nangified ‘especially nang’; the blends nangtastic ‘really cool’, nangry ‘super angry’, nanguin based on penguin and generally used as a pet name; and so on.7 Notice the pronounced usage of extra-­grammatical techniques which aim at the effect of hyperbolisation typical of the language of youth (Augenstein 1998, 147). This can also be attained by using evaluative expressions as modifiers in compounds: nang skunk ‘really good weed’; especially

Evaluative Morphology and Sociolinguistic Variation 127 in languages where the latter are very productive, as in German: arschgeil ‘really cool, lit. ass cool’, saugenial ‘really brilliant, lit. sow genial’, and even with multiple modification: superarschviel ‘very much, lit. super ass much’, oberaffentittengeil ‘really cool, lit. over monkey hooters cool’ and so on (Androutsopoulos 1998, 117). Notice that the modifiers arsch-­, sau-­ display an utterly positive evaluation in contrast with their normal value, outside of the youth language, which has a negative connotation: arschkalt ‘very cold’, sauteuer ‘very expensive’ and so on (Androutsopoulos 1998, 110). Similar to what happens in child language, special ‘youngish’ suffixes are also used, for instance in the German shortenings Alkoholiker ‘heavy drinker, alcoholic’ → Alki, Zigarette ‘cigarette’ → Ziggi, Prolet(arier) ‘redneck’ → Prollo, depressiv ‘depressive’→ depro, and so on (see Androutsopoulos 1998, 118–24). Finally, a particular interesting case of interaction between the factors of age and gender can be drawn from Italian, in which the rich inventory of EM is especially exploited in computer-­mediated communicative contexts such as blogs, chat-­lines and so on: (4)  E poi Benafflec. Lui, che normalmente ha quest’aria da sallocchione, belloccio ma proprio sedano. Lui che beve un uischi torbato ma pensa a GenniferGarniè (la donna senza sopracciglia). Insomma, di solito è bello ma non è che mi entusiasmi. Invece in Argo è un sorcone strappamutande che manco ci posso ripensare. (http://comesefossebionda.wordpress.com/tag/cimena)

‘And then Ben Affleck. He normally looks like a real klutz (lit. a sausage.aug), pretty.dim but really silly (lit. celery). Drinking a peaty whisky while thinking about Jennifer Garner (the woman without eyebrows). In short, he’s usually pretty but I’m not really fond of him. In contrast, in Argo he’s really so cute (lit. a rat. aug.m who strips off underwear) that I can’t stop thinking about him.’

In particular, the last vivid expression nicely shows how far the expressive power of EM can go, because the base of sorcone, the feminine noun sorca ‘rat’ then ‘feminine sexual organ’, is normally used by (Roman) male speakers to indicate a showy, provoking and sexually desirable woman. In contrast, ‘[g]irls use the neologism sorcone for “charming and good looking boy”’, based in its turn on the masculine conversion sorco (Romiti 1998, 291). In other words, the usage of such a term immediately qualifies the speaker or writer as provided with the features [young] and [female].8 2.1.3  Evaluative morphology and diaphasic variation Moving to the diaphasic dimension, it has been claimed several times that spoken or informal varieties display a higher frequency of EM than written or controlled varieties (in this regard, recall Grandi’s 2003a findings concerning the Italian augmentative suffix -­one, mentioned above). Moreover, this is clearly true of all those situations in which affective relations are involved (for example, pet-­and lover-­centred speech situations), because the latter also often reproduce child-­centred speech situations (Dressler and Merlini Barbaresi 1994, 191–7). As for more controlled registers, it has been often emphasised that EM generally occurs seldom in written varieties of specialized language (Roelcke 1999, 74). However, this does not exclude the possibility that single instances of EM can be used even in controlled registers, generally for terminological reasons. Moreover, specialized language also displays vertical variation, going from more to less controlled registers. This makes it difficult to postulate categorical rules such as that suggested by Rainer (1989, 212) for Italian, which alleges that no diminutives can be formed from nouns belonging to

128

Livio Gaeta

bureaucratese or to abstract, scientific or other high-­brow lexical domains (5a), while the restriction weakens with bases used in more colloquial contexts (5b): (5) a. alunno ‘disciple’ → *alunn-­ino/*alunn-­etto ‘disciple-­dim’ allievo ‘student’ → *alliev-­ino/*alliev-­etto ‘disciple-­dim’ b. scolaro ‘disciple’ → scolar-­etto ‘disciple-­dim’ studente ‘student’ → student-­ello ‘student-­dim’ As a matter of fact, the vertical variation observed in the different Italian registers has the effect of providing a whole range of expressions displaying diminutivised forms in private texts, not necessarily only in ludic contexts (6a) but also in more controlled varieties of specialized language such as the official documents of an elementary school (6b) (examples from the Internet): (6) a.  l’anno prossimo la mia classe andrà in terza (elementare) [. . .] vorrei intraprendere con i miei alunnetti un laboratorio sulle emozioni e l’autostima (http://www.maestrasabry.it/forum/viewtopic.php?f=19&t=125) ‘next year my class will attend the third year (of the elementary school) [. . .] I would like to start with my disciples.dim a laboratory on emotions and self-­confidence’ b. La scuola è raccordata con Enti, Istituzioni e genitori degli alunnini per la realizzazione delle attività (http://csakr.calabriascuola.it/allegati%5Cnews%5 C121%5CDON%20MATTEO%20LAMANNA%20-­%20MESORACA.doc) ‘The school is linked with authorities, institutions and parents of the disciples. dim for the realization of the activities.’

The same can be observed for the Italian elative suffix -­issimo (Rainer 2003), which is predicted not to be combined with adjectives belonging to high-­brow lexical domains, for instance significativo ‘significant’; but Rainer himself finds one example of significativissimo in a literary essay, and some more on the Internet. In sum, it is probably more adequate to speak of Labovian variable rules depending on several factors such as frequency rank of the derivatives and of their bases, text genre, situational context and so on, instead of categorical restrictions. Finally, socio-­cultural factors can also have an effect: for instance, Dressler and Merlini Barbaresi (1994, 215) report the following case of diminutivum modestum drawn from the Italian academic discourse uttered by a (female) linguist in the discussion period of a large conference: Avrei anch’io una mia teorietta in proposito ‘I’d have a theory.dim myself on this point.’ These authors comment that such a situation would be much too formal to allow a diminutive in German. In other words, the reduction of social and psychological distance effected by the use of EM is strictly related to the general mechanisms underlying the so-­called interactional discourse active in a given language, which ‘has as its primary goal the establishment and maintenance of social relationship’ (Kasper 1990, 205). In this regard, Dressler and Merlini Barbaresi (1994, 216) go on to observe that, according to their informants, Italian sides with Slovene and Spanish while Czech and Basque behave rather like German. Similarly, Italian political and economic discourse tolerates a far larger use of EM in contexts in which no ludic effect is attained, as testified by examples such as an economic ripres-­ina ‘recovery-­dim’, the government’s firm-­etta ‘signature-­dim’ for one

Evaluative Morphology and Sociolinguistic Variation 129 particularly demanding decret-­one ‘decree-­aug’, the concors-­accio ‘contest-­pej’ to hire new school teachers, and so on. Finally, the importance of modulating the degree of social or psychological distance in Italian is shown by the following dialogue between a female customer (C) and the male owner (O) of a shoe shop (Dressler and Merlini Barbaresi 1994, 217): (7) C:  Lei è una persona paziente? È disposto ad avere a che fare con piedi problematici? ‘Are you a patient person? Are you willing to put up with problem feet?’ O: Vediamo questi piedini! ‘Let’s have a look at these feet.dim!’ In spite of the clearly asymmetrical customer–seller relation, requiring a certain degree of formality and social distance, the owner reacts to the non-­conventional manner adopted by the customer by using a diminutive form in order to reduce the distance while giving a ludic character to the speech situation. 2.2  Evaluative morphology in the socio-­pragmatic dimension Let us briefly turn our attention to the opposite perspective, namely what kind of ingredients EM can offer in order to be exploited felicitously in the socio-­pragmatic dimension. Here pragmatics comes necessarily into play because the perspective of language use is central for understanding the particular role of EM within a given speech situation (cf. Merlini Barbaresi in this volume). In this regard, it has already been emphasised that language contributes to personal and social identity. Terestyéni (1995) identifies three potential human character types: contact-­oriented, status-­oriented and task-­oriented people, depending on their preference for ‘the connection-­creating’ or ‘the power-­representative style of knowledge’. Among others, one type of feature distinguishing the three character types relates to the way people establish personal relationships: whether they choose friends according to their common personal experiences, or to their social status, or finally to their common (for example, professional) interests. It seems plausible to assume that contact-­oriented people are likely to make a larger use of EM than the other two.9 This is due to the relevant feature [non-­serious], which can be assigned to EM in a morphopragmatic understanding of its value as ‘a natural meaning extension of [small]’ (Dressler and Merlini Barbaresi 1994, 398; Jurafsky 1996). In this connection, Spitzer’s (1921, 201–2) pioneering remarks can be mentioned: ‘Der Spieltrieb ist eine Stimmung im Sprecher, die die Grundierung des Satzes, die “key” abgibt [. . .] Die Suffixe wirken wie Vorzeichen in der Musik, sie bestimmen die “Tonart” der menschlichen Rede.’10 The ludic perspective opened up by the use of EM makes it highly suitable for the pragmatic operations of downgrading and mitigation, facilitating social contact (Caffi 2007, 99–100). In this sense, EM markers function as personal and social indices to refer to relevant elements in the speech situation anchoring the speakers to their specific roles and responsibilities in the society (Levinson 1983, 89–94). To take one example, if it may appear obvious that EM contributes to downgrading and mitigation, it is less evident that the use of diminutives may also imply upgrading, inasmuch as the speakers involved in the speech act are compelled to respond politely to mitigated requests exactly because they are formulated in a downgraded way. Therefore, they would lose face by a refusal. An Italian speaker is well aware that a request containing EM such as Mi stamperesti queste

130

Livio Gaeta

due paginette? ‘Would you print these two pages.dim?’ is essentially more binding than its unmarked variant, because a refusal requires a stronger stance by the listener (Dressler and Merlini Barbaresi 1994, 248). On the other hand, the diminutive form can be used to reproach the addressee insofar as she or he is equated to a child, which in a predominantly masculine society with its traditional (over-­)emphasis on the notions of manliness and machismo can be felt as a face-­threatening insult. In the following example from Colloquial Jordanian Arabic (Badarneh 2010), a mother addresses her eighteen-­year-­old son, complaining of his immaturity and irresponsibility: (8) yaa xwaylid quum udrus voc xaalid.dim stand:up study.imp ‘Khalid.dim, go and study [for your exams]!’ The effect is crucially connected with the use of an allocutive form – in this case the proper name – containing EM, as will be discussed below. This example also shows that the employment of EM as a morphopragmatic marker is subject to restrictions of a socio-­ cultural and ethnographic nature, as for instance this use contrasts with the marking of the diminutive on an adult’s proper name (for example, John vs. Johnnie) in English and other languages to express affection and endearment, which is oriented towards one’s positive face (Travis 2004). In this regard, it has already been observed above that certain cultures tolerate the use of EM in certain speech situations while others do not, which also implies a learning process for new members (children and foreigners; cf. Dressler and Korecky-­Kröll in this volume). This may also lead to socio-­cultural clichés which are on one hand arbitrary and stereotypical, while on the other they serve as a powerful means of self-­identification and affirmation as a community. For instance, the abundance of EM displayed by Slavic, Baltic and Mediterranean people has been opposed as the manifestation of an ‘excessive’ and somewhat ‘childish’ attitude to British ‘soberness’ or to clarté française (Dressler and Merlini Barbaresi 1994, 409–10).11 In this connection, it does not come as a surprise to observe that EM plays a role in languages with fully fledged systems of honorifics, the latter intended as ‘resources for indexing the relative position of interlocutors, referents and bystanders either in the lexicon or the morpho-­syntax of a language’ (Brown 2011, 19). As is well known, honorifics are a central part of social deixis and pay particular attention to those factors of a sociolinguistic nature (for instance, age, wealth, occupation, family background) that determine categories of people and their interactions, and generally comprise a complex system of expressive tools going far beyond word formation (see Shibatani 2006 for a survey). For our purposes, besides the sparse reference to (largely opaque) etymologies of single honorific affixes going back to evaluative expressions, such as the all-­purpose Japanese honorific prefix o-­ used with native nouns like o-­kaban ‘bag’, o-­hanasi ‘talk’ and so on, which is associated with the meaning ‘greatness’ (Shibatani 2006, 382), a clear-­cut case of EM with an honorific value is provided by the Nahuatl diminutive suffix -­tzīn (Hill and Hill 1978). Depending on the context, this suffix can be used as a truly diminutive suffix, or to mark reverential address as a distance-­respect marker, or as a form for expressing endearment and affection. Thus, while it normally marks the relationship holding among the speech participants and can be used with any kind of lexeme including particles and adverbs (for instance, āmo-­tzīn ‘no-­dim’ and ihcōn-­tzīn ‘thus-­dim’), especially in highly reverential speech situations involving older people (‘compadres’, the godfathers), parent terms such as ‘mother’ and ‘father’ cannot be used without the diminutive suffix in everyday speech

Evaluative Morphology and Sociolinguistic Variation 131 (see for instance the reverential expression īna-­huac-­tzīn in no-­tah-­tzīn ‘with my revered father, lit. his-­with-­dim the my-­father-­dim’, in which the preposition also carries the diminutive suffix), unless they occur in insults: pūtoh mitztlācatilīh in mo-­nānah ‘son of a bitch!, lit. whore she:bore:you the your-­mother’. Notice incidentally that this easily leads to lexicalised expressions such as tlācatzīntli ‘gentleman’ from tlācatl ‘man’, zoātzīntli ‘lady’ from zoātl ‘woman’, ātzīntli ‘holy water’ from ātl ‘water’, and so on. On the other hand, a form such as no-­nān-­tzīn ‘my-­mother-­dim’ can be interpreted as a highly respectful address form: ‘my reverend mother’, or, equally well, as a more intimate ‘my dear little mother’, or as both, depending on the context. The [non-­serious] feature attributed to a speech act with the help of EM is largely supported by the particular flexibility of EM, which lends itself to several uses normally impossible with other word-­formation patterns. One example of this flexibility is the low categoriality which is testified by the selection as a landing site of unexpected parts of speech, as in the Nahuatl examples discussed above involving particles, adverbs and prepositions, or in the case of the Spanish gerund: calland-­ito ‘being quiet-­dim’, corriend-­ ito ‘running-­dim’ and so on (NGLEM 2010, 512): (9)

Debíamos de acercarnos callandito – sugería Fernando. ‘We have to get close keeping quiet – Fernando suggested.’

One further peculiar property of EM is the occurrence of so-­called interfixes, namely linking elements accompanying an evaluative morpheme and possibly adding semantic nuances which are relevant at the morphopragmatic level (cf. Merlini Barbaresi, this volume). For instance, Rainer (1993, 161–70) describes the complex system of peculiar allomorphies displayed by evaluative suffixes in Spanish, which to a certain extent also involves truly sociolinguistic variables such as urban/non-­urban origin (as in the proper name Juan-­ito vs. Juan-­cito), common/technical term (as in the feminine nouns mano ‘hand’ → man-­ita vs. moto ‘motorbike’ → mot-­ito), and so on. Finally, a particularly striking case in which EM is used to encode sociolinguistic information, exploiting its flexibility, is found in Nootka (and other North American languages such as Yana; Sapir 1915). Here, the diminutive suffix -­ˊis attached to the verb form is normally used for addressing children (10a) or for talking about children (10b) or even for speaking about oneself when addressing a child, but only to show affection at the same time (10c): (10) a. qwís-­ˊis-­tciᶜ do:so-­ dim-­2sg.imp ‘do so, little one!’ b. qwís-­ˊis-­maᶜ do:so-­ dim-­3sg.prs.ind ‘He does so’ c. waɭ̴-­cíL-­ˊis-­aH go:home-­ incp-­dim-­I ‘I’m going home, little one!’ Interestingly, the same diminutive suffix identifies social groups such as small people or dwarfs (11a), in this case accompanied by a palatalisation of all sibilants: s → ś, ts → tś

132

Livio Gaeta

and so on; eye-­impaired (not blind!) people (11b), with the addition of the lateralisation of sibilants: s → ɭ̴, ts → L and so on; hunchbacks (11c), this time with the sibilants becoming ‘thickish’ sounds pronounced with the lower jaw held in front of the upper; left-­handed people (11d), with the addition of the meaningless element -­tcHa-­ inserted after the first syllable; and so on: (11) a. hín-­tᶜ-­śiL-­ˊiś-­weˊini lightv-­come-­incp-­dim-­quot ‘he, little man, comes, they say’ b. qwíɭ̴-­ˊiɭ̴-m ­ aᶜ dim-­3sg.prs.ind do:so-­ ‘he, one-­eyed man, does so’ c. yā ́ tc ̣-­ukᶜ-­ˊic-­maᶜ walk-­ intr-­dim-­3sg.prs.ind ‘he, hunchback, is walking’ d. yā ́ ɭ̴-t­cHa-­ˊits!-­aL-­maᶜ be:there-­ ep-­dim-­now-­3sg.prs.ind ‘there now he is, poor little left-­handed chap!’ This variety of uses stresses the value of EM as a marker for expressing social relations and socio-­pragmatic connotations within a speech community. In this regard, Sapir [1915] (1991, 362) observes that ‘these distinctive forms are generally avoided when in the presence of [the referred] people, for fear of giving offence’. As explained throughout this chapter, the speaker must have access to a large number of pieces of information of a linguistic and social nature in order to become acquainted with the behavioural norms valid within a speech community. Nootka shows masterfully how their richness is perfectly covered by the large expressive potential of EM.

Notes   1. Dressler (1994) sees the reason underlying the iconic structuring of EM in its early development in small children, at a stage when no distinction is made between grammatical rules and the so-­ called extra-­grammatical (or ‘expressive’) morphology which does not obey a strict rule-­based pattern. Ohala’s hypothesis provides the phylogenetic background for Dressler’s ontogenetic interpretation. See the discussion below.   2. In this regard, cf. Cameron and Coates’ conclusion (1988, 24): ‘Explaining sex differences does not just mean explaining the usage of women, after all. It means devising methods applicable to all informants, so we can gauge the importance of sex in the complex system of intersecting social relations that support linguistic variation.’  3. Notice that barboncino is a false diminutive based in its turn on the false augmentative barbone, which actually means ‘beggar’ from barba ‘beard’, lit. ‘long beard’. Clearly, Matteo’s creation is not influenced by the derivational history of the base, which is apparently unknown to him.   4. Cf. Dressler and Merlini Barbaresi (1994, 408): ‘Diminutive formation is the first morphological rule that small children acquire [. . .], that is, it is acquired at a time when there is no distinction yet between extragrammatical morphological operations and grammatical rules.’

Evaluative Morphology and Sociolinguistic Variation 133   5. This ‘childish’ suffix differs from the its homonym -­i occurring in formations such as Studi ‘student’, Fundi ‘fundamentalist’ and so on, because it is not restricted to baby talk and generally forms nouns (Androutsopoulos 1998, 118–21).   6. Notice the use of the verbal form in the second person plural, which is normal in several varieties of Italian for addressing persons deserving a certain respect, such as older people.   7. For information see http://www.urbandictionary.com.   8. At any rate, since the Italian augmentative suffix -­one can also give rise to masculine nouns from feminine bases, such as donna ‘woman.f’ → donn-­one ‘woman-­aug.m’, sorcone can also be used, although in a less conventionalised manner, by male speakers with reference to attractive women, as in this example from the Internet: Incontrata 20 giorni fa agli imbarchi di Fiumicino. Un sorcone monumentale. . . ‘Met.f 20 days ago in the boarding area of Fiumicino airport. A monumental rat.aug.m’.   9. In this regard, recall what has been said above about women, who on average use more diminutives and are taken to exert a social function oriented towards the maintenance of social relationships, for instance solidarity (Holmes 1995, 11–16). 10. ‘The ludic instinct is a mood of the speaker’s which creates the ground of the sentence, the “key” [. . .] Suffixes [sc. diminutives] work like key signatures in music, determining the “key” of human speech.’ 11. National clichés lead to the hyper-­characterisation of certain features, often with ironic intent, by bilingual foreigners, as shown for instance by the exaggerated and inappropriate use of diminutives made by Voltaire when writing Italian (for example, bigliettini ‘tickets.dim’ for theatre tickets, while the term normally refers to a note) while he was very sparing of this in French (Folena 1983, 410). In France, as a reaction against Italian influence, the use of EM has been largely condemned since the seventeenth century, leading to a decay of its use in the standard variety (Hasselrot 1957, 216).

10  Evaluative Morphology and Language Acquisition Wolfgang U. Dressler and Katharina Korecky-­Kröll

1 Introduction Evaluative morphology is viewed here as in Merlini Barbaresi’s contribution to this volume (based on Dressler and Merlini Barbaresi 1994; 2001) in terms of morphopragmatic meanings characterised by fictive evaluation and lowered formality plus the morphosemantic denotation of size. This chapter is limited to the acquisition of the following evaluatives: hypocoristics, diminutives, augmentatives and pejoratives. Whenever pertinent, diminutives and hypocoristics are treated separately, although their meanings and forms typically overlap. They are distinguished by their functions as names (hypocoristics) vs. common nouns. Thus Engl. dadd-­y used to address one’s own father is a hypocoristic, although derived from the common noun dad; Grmn. Kasper-­l(e), although derived from the name Kaspar, is a common noun designating the main male character in a Punch and Judy show and, more in general, a male who is not to be taken seriously. Thus prefixes or prefixoids of the type mini-­, maxi-­and super-­, or attenuative suffixes as in Engl. blue-­ish, are excluded from this chapter, both for theoretical reasons (see Dressler and Merlini Barbaresi 1994) and because we do not know of any studies about their acquisition. Nor are any studies about the acquisition of elatives, such as Ital. grand-­issimo ‘very big’, known to us. But their acquisition seems to occur later than that of the evaluatives treated here. In all the languages dealt with in this chapter the evaluatives are part of non-­prototypical derivational morphology (in contrast to Bantu languages where they appear to belong to non-­prototypical inflectional morphology, whereas Déchaine, Girard, Mudzingwa and Wiltschko in this volume attribute them to mixed inflection and derivation). In our model, their pragmatic meanings are not based on, or derived from, semantic meanings, and acquisition evidence supports the priority of pragmatic meaning. Diminutives are deficient, if they have only semantic meanings, such as contemporary standard French diminutives; they express no or nearly no evaluation. Most of them are thoroughly lexicalised (Dressler 2010; Fradin 1999). The usual pragmatic meanings are expressed by mostly pejorative -­o (Kilani-­Schoch and Dressler 1999b) or reduplication. Their acquisition cannot be dealt with here, because our child language data and others we are aware of are devoid of diminutives. Thus they must be acquired later, and much later than in all the languages reported here. Neither the system of the target language nor its manifestations in corpora of adult-­ directed speech (ADS) can be the source of child language acquisition. It is (especially parental) child-­directed speech (CDS), which differs much from ADS(Ravid et al. 2008),

Evaluative Morphology and Language Acquisition 135 especially in the area of evaluatives. The input of CDS, particularly in early phases of acquisition, strongly determines the children’s output in distribution, type and token frequency and by degree of salience. Therefore longitudinal corpora of parent–child interaction must be the main source for studying the acquisition of evaluatives. Transversal group studies or diary data, which lack CDS data, are only of secondary importance. Thus the main basis for this chapter are the data of the Crosslinguistic Project on Pre-­and Protomorphology in Language Acquisition with its parallel spontaneous data collection, transcription and coding, as indicated two paragraphs below.

2  History of research on the acquisition of evaluative morphology Older research has been summarised by Ferguson (1977), who concluded (p. 224) that ‘the most prominent expressive feature of BT (baby talk) probably is the hypocoristic affix’ and that diminutives are especially frequent in CDS. Still relevant and therefore commonly cited research on the acquisition of evaluative morphology started with transversal and longitudinal investigations in Bates and Rankin (1979) on Italian diminutives (assigning them erroneously to inflection). The pioneering tests are the famous nonsense-­words ­nomination tests by Berko (1958), where the youngest Anglophone children formed from the nonsense-­word wug a diminutive wig, older ones preferred wugg-­ie, whereas only ­adolescents and adults formed wug-­let, wug-­ette, thus with stages of decreasing degree of iconicity (direct i-­iconicity > addition of iconic suffix /i/ > of suffix with little or no iconicity). The most massive longitudinal data have been collected and interpreted by Rūķe-­Draviņa (1959) in Latvian, with greater richness of diminutive suffixes in CDS and CS than in ADS. Further studies have been done on single languages on their own or within the frame of the international Crosslinguistic Project on Pre-­and Protomorphology in Language Acquisition, resulting notably in the volume edited by Savickienė and Dressler (2007b) and covering twelve languages (eight Indo-­European ones, two Finno-­Ugric, Turkish and Hebrew). Most studies have observed and, in part, tried to explain the early emergence of diminutives, e.g. Dąbrowska (2006) by the simplicity of diminutive schemas; Dressler (1994) by the appurtenance to non-­prototypical derivational morphology; many others by dependency of CS on increased use in CDS (including the order of emergence of specific allomorphs), e.g. Gillis (1997), Stephany (1997) and many contributions to Savickienė and Dressler (2007b); others by the general and ontogenetic priority of pragmatic over semantic meanings, e.g. the same sources, Dressler and Merlini Barbaresi (2001), Savickienė et al. (2009), Dressler, Lettner and Korecky-­Kröll, (2012); cf. Lepskaja (1997, 89–90). Other topics were the acquisition of diminutives as facilitators of the acquisition of noun and adjective inflection, due to the more transparent and productive inflection of diminutives in comparison with the inflection of their simplex bases (Olmsted 1994; Laalo 1998; Kempe and Brooks 2001; Kempe, Brooks and Pirott 2001; Kempe, Brooks, Mironova and Fedorova 2003; Savickienė 2001; 2003; many chapters in Savickienė and Dressler 2007b; Dressler et al. 2012), and the contribution of diminutives to the acquisition of discourse strategies (Melzi and King 2003; King and Melzi 2004). Franceschet (2013) compared Italian children and aphasics in production and processing tests, investigating the role of frequency and the distinction between potential and illegal nonce diminutives. The acquisition of augmentatives was never investigated on its own, but only and rarely in connection with diminutives. All researchers observed that augmentatives were acquired later than diminutives (Ceccherini, Bonifacio and Zocconi 1997 and Noccetti De

136

Wolfgang U. Dressler and Katharina Korecky-­Kröll

Marco, Tonelli and Dressler 2007 on Italian, Protassova and Voeikova 2007 on Russian, Thomadaki and Stephany 2007 on Greek). There are no studies on the acquisition of pejoratives.

3  Early emergence of evaluatives (particularly diminutives) Wherever diminutives and hypocoristics occur in CDS more than rarely, which appears to represent a critical mass, they are the earliest affixes of derivational morphology to emerge, simultaneously with the first inflectional affixes and with compounding (provided there is rich compounding in the CDS). This is, among the languages of our Crosslinguistic Project, the case with Croatian, Dutch, Finnish, English, German, Greek, Hebrew, Hungarian, Italian, Lithuanian, Russian and Spanish (Savickienė and Dressler 2007b). If the language in question has more than one diminutive affix, then the most productive in ADS and CDS, and those which include the putatively most iconic vowel /i/ (but cf. Bauer 1996; Gregová, Körtvélyessy and Zimmermann 2010), are the first or among the first. The latter is true for Croatian, Dutch, Greek, Italian, Lithuanian, Russian, Spanish and for the child-­centred or even child-­specific -­i suffixes of English, Hebrew, Hungarian, German and Finnish hypocoristic -­i. Only Turkish and French CDS lack such critical mass, and thus neither diminutives nor hypocoristics emerge early. On the other hand, in the two European languages which use diminutives most frequently, i.e. Dutch and Lithuanian, diminutives do not emerge earlier than in other languages, but only simultaneously with inflectional affixes. In early phases, diminutives may appear in reduced forms, especially, due to the recency effect which renders the end part of words most salient to young children, in the format of word-­final trochees, as in Ital. pall-­ina ‘ball-­dim’ and farfall-­ina ‘butterfly-­dim’ → lina, mic-­ini ‘cat-­dim.pl’ → cini. Rarely, the recency effect combines with the primacy effect, which favours the word beginning, as in Ital. panc-­ina ‘belly-­dim’ → pinna, and more deviating pall-­ina → pilla (Noccetti et al. 2007, 134). Other examples are Lithuanian hypocoristic Týtė for Rūt-­ýtė ‘Ruta-­dim’ and the frequent [‘la:di] for Grmn. Marmeláde/ Schokoláde ‘jam/chocolate’ + child-­specific suffix -­i. Allomorphic alternations may first not be mastered, as in Ital. nino, which is reduced from camion-­ino ‘truck-­dim’ without the insertion of the affricate -­c-­, which is obligatory after root-­final /on/ (adult: camion-­c-­ino). Similarly an Austrian girl did not use the opacifying umlaut in *Vogelein instead of Vögelein from Vogel ‘bird’ plus suffix -­lein (Korecky-­ Kröll and Dressler 2007, 217). Patterns of diminutive formation can be assumed to be acquired when the so-­called mini-­paradigm criterion is fulfilled, i.e. when a child produces diminutives alongside their simplex bases and also in different inflectional forms, as in Ital. biba, bib-­ina ← bimba, bimb-­ina ‘little girl(-­dim)’ or scarpa ‘shoe’, dim scarp-­etta, pl scarp-­ett-­e. The first Austrian German mini-­paradigm is Hand ‘hand’, pl Händ-­e, dim Hand-­i (Korecky-­Kröll and Dressler 2007, 218); cf. Palmović (2007, 81–3) for Croatian and Marrero, Aguirre and Albalá (2007, 263–7) for Spanish. Augmentatives, which represent, wherever they occur, a marked category in opposition to the unmarked category of diminutives (Dressler and Merlini Barbaresi 1994), emerge invariably later than diminutives: for Italian, Ceccherini et al. (1997, 161) found that the first rare examples appeared in 1;7, at least one month later than the robust emergence of diminutives, and the productive use 7 months later than that of diminutives, at 2;3. Similarly, Noccetti et al. (2007, 135, 138) in the four children studied, between 1 month

Evaluative Morphology and Language Acquisition 137 and 1 year later, found examples such as (p. 145) pall-­ona vs. pall-­ina ‘ball-­aug/dim’, maial-­one vs. maial-­ino ‘pig-­aug/dim’ (Rosa at 3;1,29). For the late emergence of augmentatives in Russian see Protassova and Voeikova (2007, 51, 62), whereas in the data of our Croatian (up to 2;8) and Greek children (up to 3;0) augmentatives do not appear at all. The only data about the acquisition of pejoratives are Italian ones communicated to us by Sabrina Noccetti: a unique imitated (thus non-­productive) token piccin-­accio ‘small-­ pej’ from Camillo at 2;7, with many diminutives from 2;0 and some augmentatives from 2;5. Pejoratives are also very rare inCDS.

4  Formal development A theoretically interesting aspect of the formal development of diminutives (none of augmentatives in the data at our disposal) is the emergence of additional diminutive suffixes, similar to what we find in the development of competing inflectional suffixes. Thus in the corpora of the four Italian children studied by Noccetti et al. (2007), all children start with the suffix -­ino, and one with -­etto also. -­etto follows soon in two other children, but late in the fourth; -­ello follows soon in two children, but later in the third and late in the fourth. Italian diminutives occur first with nouns (e.g. libr-­ino/etto ‘book-­dim’), then with adjectives (e.g. piccol-­ino ‘small-­dim’), finally with adverbs (pian-­ino ‘slowly-­dim’; Noccetti et al. 2007, 138). The same holds for the other languages investigated in Savickienė and Dressler (2007b). Some children produce certain deadjectival diminutives from the beginning, e.g. Hung. kicsi-­ke ‘small-­dim’ (Bodor and Barcza 2007, 251 vs. 250). The priority of denominal diminutives is strengthened by hypocoristics, such as Ital. mamm-­ina, Span. mam-­i, mam-­ita, Lith. mam-­ytė, Russ. mam-­at‘ka, Croat. voc mam-­ice, Greek mam-­ula, Grmn. Mam-­i, Hung. any-­u(-­ka) ‘mum-­dim’. This order of emergence and acquisition cannot be interpreted as supporting universal base preferences of diminutives but follows simply from the relative type and token frequencies of nouns, adjectives and adverbs and of the respective diminutives both in the target languages and in CDS. Diminutive cumulation and recursive diminutive suffixation may come early or late, depending on diminutive wealth in the language as materialised in CDS: relatively early in Span. chiqu-­it-­ín, chiqu-­in-­ín ‘small-­dim-­dim’, poqu-­it-­ito ‘little-­dim-­dim’ (Marrero et al. 2007, 167) and Russ. konfet-­oč-­ka ‘sweet-­dim-­dim’, kam-­us-­it-­ik ’stone-­dim-­dim-­dim (Protassova and Voeikova 2007, 60–1); later in Ital. ors-­ett-­ino ‘bear-­dim-­dim’, pezz-­ett-­ino ‘piece-­dim-­dim’ (with the more productive suffix always in second position; Noccetti et al. 2007, 134–44); different for the two Hungarian children: early any-­u-­ci, ap-­u-­ci vs. later any-­u-­ka ‘mum-­dim-­dim’, ap-­u-­ka ‘dad-­dim-­dim’ (Bodor and Barcza 2007, 250–3); and late and isolated Grmn. Kath-­i-­li, hypocoristic of Katharina (Korecky-­Kröll and Dressler 2007, 220). Interfixation of diminutives occurs rather late in Italian, e.g. pesci-­ol-­ini alongside pesc-­ ini ‘fish-­dim.pl’ (Noccetti et al. 2007, 136), and earlier in Spanish, where they are obligatory and mostly in complementary distribution with non-­interfixed diminutives (Marrero et al. 2007, 167), not optional as in Italian. Children form new but potential diminutives, such as Ital. guant-­etto instead of normal guant-­ino ‘glove-­dim’, penn-­ell-­ina ‘pen-­dim-­dim’ (Noccetti et al. 2007, 136), Russ. korov-­ jonok ‘cow-­dim’ instead of telj-­onok ‘calf-­dim’ (Protassova and Voeikova 2007, 61), but rarely ungrammatical or strange ones, such as in Italian masculine instead of feminine plurals pecor-­ini ‘sheep-­dim’, lus-­ini ‘light-­dim.pl’ (instead of luc-­ine; Noccetti et al. 2007, 135). Of course, one has to exclude from ungrammaticality child-­specific forms, which are

138

Wolfgang U. Dressler and Katharina Korecky-­Kröll

only appropriate in reference to, and in use by, small children plus in derived pragmatic contexts, i.e. in pet-­directed and lover-­directed speech (Dressler and Merlini Barbaresi 1994). Very clear examples are the child-­specific diminutive and hypocoristic suffix -­i in Hungarian, Austrian German and Hebrew, and the Greek suffix -­ina (Thomadaki and Stephany 2007, 115–17). The compositionality of diminutives leads analogically to transparent back formation from morphotactically opaque diminutives, such as Russ. jož from jož-­ik, instead of simplex jog ‘hedgehog’ (Protassova and Voeikova 2007, 62, cf. 50) or Ital. cagno from cagn-­ol-­ino instead of simplex cane ‘dog’ (Noccetti et al. 2007, 136). Back formations are also derived from (at least synchronic) non-­diminutives, such as Ital. violo from violino ‘violin’ (only diachronically a diminutive of viola) or Grmn. Müs from Muesli (Korecky-­Kröll and Dressler 2007, 222) or Lith. masc. did-­is, fem. dide from didel-­is/ė ‘big’ (Savickienė 2007, 32). Children may vary, without any meaning difference, between several actual or potential diminutive forms of the same base, such as Lith. mešk-­iukas, mešk-­utė, mešk-­ytė from mešk-­a ‘teddy-­bear’ (Savickienė 2007, 22), Greek bal-­aki, bal-­itsa from balla ‘ball’ (Thomadaki and Stephany 2007, 108–9), and Russ. kam-­ik, kam-­is-­yk, kam-­us-­ik, kam-­us-­ it-­ik from kamen ‘stone’ (Protassova and Voeikova 2007, 61). This was taken in Savickienė and Dressler 2007b, 345 and passim) as evidence against Clark’s (1993, 69–83; 1995) principle of contrast, which (derived from the structuralist principle of contrastiveness between signs) claims that children assume that form distinctions reflect functional or meaning distinctions. Often these variations appear to be just a result of children’s playfulness. We have already noted several interindividual differences between children acquiring the same language. Many of them consist in, or include, frequency differences in distribution and are due to the great (albeit indirect) impact of the input (CDS) on children’s output, as demonstrated for Greek (Thomadaki and Stephany 2007, 103–4), Lithuanian (Savickienė 2007, 19–21) and Italian (Noccetti et al. 2007, 136–41). However, frequency distributions in the input can only partially explain the order of acquisition; i.e. the most frequent diminutives in the input emerge more often than, not later than, other less frequent diminutives (Savickienė and Dressler 2007, 140). In many languages, diminutives inflect only or predominantly in productive and morphotactically transparent inflection classes, whereas this is not the case, or much less so, with their simplex bases. For example, Ital. ami[k]o ‘(male) friend’ has the unproductive and opaque plural ami[tʃ]-­i, whereas its diminutive amich-­etto has the productive and transparent plural amich-­ett-­i. Such a constellation induces Greek, Hungarian, Italian, Lithuanian and Russian children to prefer inflecting diminutives over their simplex bases, although diminutives are longer and thus demand an additional phonological effort (Savickienė and Dressler 2007b; Savickienė et al. 2009). In this way the development of diminutives facilitates and supports the development of inflection.

5  Development of meaning In the first weeks or even months of the emergence of diminutives, their meaning is difficult to seize. Often they are attested without their simplex bases, and when they appear in adjacence to their simplex base, variation between diminutive and simplex often seems to be determined by mere playfulness, similar to variation between different diminutives derived from the same base (see above, Section 4). But when meanings start to be recognisable, they are invariably pragmatic in nature. This holds both for hypocoristics and for

Evaluative Morphology and Language Acquisition 139 diminutives in the narrow sense (Dressler and Merlini Barbaresi 2001; Savickienė and Dressler 2007; Savickienė et al. 2009). Both are pragmatically appropriate to child-­centred speech situations (Dressler and Merlini Barbaresi 1994; Jurafsky 1996), especially with young children. Their spontaneous production contains high concentrations of diminutives in lexical fields where pragmatic meanings are expected, i.e. when referring to animals, body parts, food and toys, independent of whether they are small or big. Child-­specific diminutives are adequate for young children; three-­or four-­year-­old children may already object to being addressed with the child-­specific diminutives of mass nouns, such as Ital. acqu-­etta ‘water-­dim’, or their equivalent with the child-­specific i-­diminutive in Grmn. Wass-­i from Wasser ‘water’. Like the pragmatic meanings in CDS, children‘s diminutives first express endearment, intimacy, warm feelings of affection, love and kindness. They are often accompanied by prosodic indicators of tenderness and empathy (Noccetti et al. 2007, 142). Thus the Greek girl Anna uses hypocoristics and non-­standard diminutives in reference to her sisters but not in self-­reference, where it would be pragmatically inappropriate (Thomadaki and Stephany 2007, 117). Or the Lithuanian ‘caregivers and Rūta use diminutives when they speak about those who are weak, sick or helpless, i.e. when expressing sympathy for weakness’ (Savickienė 2007, 36). Soon children start to use diminutives strategically in modifying speech acts of requesting in order to facilitate obtaining the desired perlocutionary sequels. The most telling examples are represented by the contrast between diminutives used by a child in a positive interaction with the adult caretaker and the respective simplex bases in a negative interaction, as described for the Lithuanian girl Rūta by Savickienė (2007, 32–5). For a clear strategic use of a Hungarian diminutive, see Bodor and Barcza (2007, 255, reproduced in Savickienė et al. 2009, 346). When the Austrian boy Jan (aged 1;9) asked repeatedly for more Müsli but did not obtain his wish, he got angry and shouted for his *Müs, wrongly thinking that Müsli is a synchronic diminutive and thus producing an occasional back formation (Korecky-­Kröll and Dressler 2007). The semantic meaning of smallness emerges only later in acquisition. The Lithuanian, Hungarian and Austrian German acquiring children investigated in Savickienė and Dressler (2007), at least, first use consistently attributive adjectives denoting smallness for expressing smallness, which they only later express with diminutivised nouns. Earlier on diminutives of ‘small’, e.g. Lith. fem. maž-­ytė, do not signify still smaller size but pragmatic meanings attached to the adjective meaning (Savickienė 2007). Double diminutives of ‘small’, such as Span. chiqu-­it-­ín, chiqu-­in-­ín ‘small-­dim-­dim’, poqu-­it-­ito ‘little-­dim-­ dim’, also reinforce pragmatic meanings. In Italian, the semantic meaning of smallness appears when augmentatives emerge whose contrast to diminutives in meaning consists in the denotation of big size (Ceccherini et al. 1997; Noccetti et al. 2007), whereas there is a basic pragmatic similarity. The comparatively late emergence of the semantic meaning of smallness of diminutives contrasts with early development of size perception, on which Prieto (2005; this volume) bases his claim on the basic meaning of littleness for diminutives and bigness for augmentatives. The late emergence of augmentatives also contradicts Prieto’s claim. Another consequence of the priority of pragmatic meanings can be seen in the result of their lack in French, namely that the (merely semantic) diminutives occur neither in CDS nor in CS. Pragmatics plays still another role in diminutives. The well-­known stereotype, documented for many communities, that diminutives fit better with females than with males has

140

Wolfgang U. Dressler and Katharina Korecky-­Kröll

the effect that mothers often use more diminutives in CDS to girls than to boys, with the effect that these girls also use more diminutives than these boys, as attested for German by Korecky-­Kröll and Dressler (2007, 226–7), Korecky-­Kröll (2011) and Dressler et al. (2012), and for Russian by Protassova and Voeikova (2007, 50). The early emergence of pragmatic meanings of diminutives also falsifies claims by Avrutin (1999) and Ladegaard (2003) on pragmatics developing only late. However, less basic pragmatic meanings of diminutives, e.g. irony or sarcasm (Dressler and Merlini Barbaresi 1994), emerge much later: when talking about a kiss a girl gave him in kindergarten, the Austrian boy Jan (aged 5;7) formed the exaggerated ironic diminutive Kuss-­i-­busch-­i-­bull-­i-­gutt-­i ‘kiss-­dim-­kiss-­dim-­pullover-­dim-­good-­dim’ in order to overcome his embarrassment (Korecky-­Kröll and Dressler 2010, 244).

6 Conclusion Diminutives and hypocoristics are the earliest category of evaluatives and of derivational morphology at large to emerge in first language acquisition. This has nothing to do with general language typology (Savickienė and Dressler 2007; Dressler et al. 2012), and a small number of diminutives and hypocoristics in CDS is a sufficient critical mass in the input (not present in French and Turkish) for stimulating early diminutive output in early CS. Their morphological patterns emerge simultaneously with inflectional patterns and compound patterns (provided that these occur sufficiently in CDS; see Dressler, Kilani-­ Schoch and Klampfer 2003). And clearly diminutives are (much) more frequent in CDS than in ADS. This chronological precedence of diminutives over other patterns of derivational morphology is in all probability due to their adequacy of pragmatic meanings to child-­centred speech situations. Thus it is predicted that diminutives should also emerge very early and first with pragmatic meanings in Bantu languages, where they are part of inflectional morphology; a prediction yet to be tested. The chronological precedence of pragmatic over semantic meanings in CS supports the claim in Dressler and Merlini Barbaresi (1994; 2001; cf. Dressler, forthcoming; Merlini Barbaresi, this volume) that pragmatic meanings of diminutives are more basic than, or at least as basic as, semantic meanings. This fits evidence from diachrony, e.g. of the same suffixes being diminutive in some Romance languages but augmentative in others, which is only possible if the common pragmatic meaning of diminutives and augmentatives is basic. Further evidence comes from the translation of Italian augmentatives into English or German diminutives, because these target languages have only diminutives but no augmentatives, and from the typological implication that if a language has diminutives, it must also have hypocoristics, but not vice versa: hypocoristics have essentially a pragmatic meaning. For further evidence see Dressler and Merlini Barbaresi (1994) and Dressler (forthcoming). Although the evaluative character of many adjectives should make them particularly prone to diminutivisation, there is a clear noun bias in the acquisition of diminutives. But this can be explained by the earlier emergence and development of nouns than of adjectives (cf. Tribushinina et al., forthcoming). The gradual developmental expansion of diminutives in form and meaning fits the development of other morphological patterns in acquisition. The graduality of development speaks against the dual-­route model of diminutives, contrasting one diminutive default rule to all other diminutives only stored but not accountable for by rules, as proposed by Clahsen, Sonnenstuhl and Blevins (2003). The early acquisition of diminutives

Evaluative Morphology and Language Acquisition 141 has, as a consequence of the age-­of-­acquisition effect, the result that diminutives are also comparatively well preserved in aphasia: both children and aphasics process diminutives with high type or token frequency better and more rapidly than they do less frequent diminutives (Franceschet 2013; Turolla 2013) and diminutives with pragmatic meaning (Turolla 2013). But the input frequency of diminutives can only very partially explain their order of emergence in acquisition. The acquisitional precedence of i-­diminutives or diminutive suffixes containing /i/ and small children’s overall greater preference for iconic devices support the general view on an iconic preference for /i/ and palatality in diminutives, against the criticism of Bauer (1996) and Gregová et al. (2010). In languages rich in diminutives, their acquisition facilitates the development of productive and transparent inflection (Savickienė and Dressler 2007b; Savickienė et al. 2009). The main lacunae in investigations of the acquisition of evaluatives are the lack of studies in the acquisition of pejoratives, and the near-­lack of studies in the acquisition of augmentatives, in the acquisition of evaluatives in extra-­European languages, and in their later development in school children and adolescents.

11  Evaluative Morphology in a Diachronic Perspective Katrin Mutz

1 Introduction This chapter treats evaluative morphology from a diachronic angle, with a strong focus on languages with suffixal devices, due to the preponderance of this concatenative type in the languages of the world (Štekauer, Valera and Körtvélyessy 2012, 325). The theoretical perspective adopted in this chapter is a functionalist, cognitive one which draws on findings from grammaticalisation research (Heine, Claudi and Hünnemeyer 1991; Hopper and Traugott 1993) and cross-­linguistic studies (Jurafsky 1996; Matisoff 1991; Grandi 2002) as well as from historical analyses of individual languages or language families (Hasselrot 1957; Mutz 2000). After these preliminary remarks, some notes on language change in general will be given, followed by an overview of the processes typically involved in morphological change and which can also be observed in the histories of evaluative affixes. Section 3 describes, after a brief sketch of the research literature (3.1), the etymological origin and common semantic paths of changes of diminutive (3.2), augmentative (3.3) and pejorative (3.4) formatives.1 The chapter closes with a synthesis of the general trends of change observed in evaluative morphology.

2  Language change 2.1  General remarks on language change Diachronic change and synchronic variation are natural and inherent components of language which are due to the use of language for communicative purposes, be they referential, expressive or appellative ones (Bühler 1934). The (sometimes conflicting) communicative goals and intentions of speaker and hearer, pursued in linguistic interaction and using different strategies, lead to a constant negotiation of linguistic form and meaning against the background of existing norms. Such communicative interactions in the long run can lead via invisible-­hand processes (i.e. unintentionally) to language change (Keller 1984) in each micro-­and macro-­domain of language. Some components of language are less (frequently) prone to change than others: it is a commonality that components of grammar are both synchronically and diachronically more stable than lexical elements; they show less variation and change less (frequently). Grammatical change, though, surely does exist: one well-­described and well-­documented cross-­linguistic type of grammatical change is grammaticalisation (Hopper and Traugott

Evaluative Morphology in a Diachronic Perspective 143 1993). Many of the (semantic) tendencies of change detected in grammaticalisation research can also be observed in the development of evaluative affixes (see Section 3). 2.2  Language change and morphology Evaluative morphology cross-­linguistically and intra-­linguistically tends to present characteristics of both inflectional morphology and derivational morphology as well as characteristics of its own (see Grandi, this volume; Bauer 1997, 545ff; Štekauer et al. 2012, 26f.); in research literature it is therefore often represented as either non-­prototypical inflection or non-­prototypical derivation (Dressler 1989) or as a third kind of morphology, in addition to inflection and derivation (Scalise 1984; Fortin 2011). Because of this hybrid inflecto-­ derivational nature, evaluative morphology is assumed to show features of both grammatical and lexical change; i.e. the evaluative categories and the respective affixes of a given language are expected to change less frequently than the more prototypical derivational categories with their elements, but at the same time they should be less stable than more prototypical inflectional category markers within the respective language. Not all evaluative categories and their markers behave with regard to their degree of stability in the same way: diminutives tend to be more unstable than augmentatives (Grandi 2011b).2 Haspelmath (2002, 51–7) describes four different types of morphological change (valid for both derivational and inflectional morphology) which are also crucial – to a greater or lesser degree – in the development of evaluative affixes:

(i) pattern loss: e.g. in Bantu languages a recent renewal of the diminutive category is observable: older prefixal markers are replaced by more transparent suffixal diminutives (Creissels 1999); (ii) coalescence: e.g. in Ewe the autonomous lexeme ví ‘child’ developed into a diminutive suffix (Heine et al. 1991, 79–97);3 (iii) analogy: here, the subtype of analogical extension with respect to the (semantic) category of the derivational base is of particular relevance for the development of evaluative markers, e.g. in Romance languages (Mutz 2000); (iv) reanalysis: e.g. the Greek diminutive suffix -­ákion has emerged via resegmentation and (re)functionalisation (Jannaris 1897, 292). This inventory of types of morphological change must be completed by the following (non-­ morphological) diachronic processes:

(v) change in distribution and/or productivity of morphological devices: e.g. whereas in contemporary French the formation of evaluatives via suffixes is the exception (unlike in Ancient French), it is rather common in non-­standard varieties; (vi) change in the phonetic structure of formatives: e.g. Lat. -­aceus [atseus] > Ital. -­accio [atʃ:o], (vii) semantic change of affixes: e.g. Latin -­aceus ‘similar to’ > Ital. -­accio ‘ugly’. The focus in this section will be on the semantic change of evaluatives. Blank (1997) establishes eleven types of semantico-­lexical change; out of these eleven mechanisms four, following Rainer (2005), are involved in affix-­innovating reinterpretation processes (in Romance): metonymy, metaphor, semantic absorption via ellipsis, and folk etymology. The most relevant type of these semantic processes for evaluative affixes

144

Katrin Mutz

seems to be metonymy, as we will see in the next section. In particular, the metonymy-­ driven process of context-­induced reinterpretation of meaning is crucial in the emergence of evaluative meanings.4 In the following paragraphs, examples from different languages will be given to show the different paths of change which lead to the (innovating) emergence of evaluative categories or affixes, and the mechanisms of semantic change at work with already-­existing evaluative formatives.

3  Diachronic aspects of evaluative morphology 3.1 Research Whereas there is a rather large amount of research literature concerning the origin and development of evaluative affixes in the Indo-­European languages, especially in the Romance, Slavic and Germanic language families (Belič 1901; Hakamies 1951; Ettinger 1974; Mutz 2000; Grandi 2011b), other language families have been somewhat neglected in this domain (Bakema and Geeraerts 2004, 1051). This is, for example, the case for many African languages, Asian languages, Austronesian languages and Native American languages. Only a few studies on historical evaluative morphology can be found for these areas (e.g. Creissels 1999; Matisoff 1991). One reason for this research gap surely lies in the lack of written historical documents for many of these languages, which for centuries relied on oral literature and oral traditions. A description of the historical development of linguistic elements in these historically undocumented languages is only possible via reconstruction based on language comparison and synchronic polysemies, as well as on exploitation of the findings in typological research, research on language universals and grammaticalisation studies. An outstanding work in this respect is surely Jurafsky’s seminal article (1996): his model of the diminutive category, which is hypothesised as having both synchronic and diachronic cross-­linguistic validity, is based on the analysis of the semantics of the diminutive in about sixty languages. Jurafsky (1996, 536) observes that the following meanings tend to be expressed cross-­linguistically by diminutive formatives: ‘small’, ‘child/offspring’, ‘female gender’, ‘small-­type’, ‘imitation’, ‘relatedness’, ‘intensity/exactness’, ‘approximation’ and ‘individuation/partitive’. Jurafsky (1996, 561) furthermore postulates that the diminutives cross-­linguistically have equally pragmatic uses and functions: they are used in communication with children and pets, they express sympathy (or contempt), they function as metalinguistic hedges to weaken illocutionary force, and they are used as markers of irony (see Merlini Barbaresi, this volume). He describes this polyfunctionality of the diminutive in a radial category (see Figure 11.1), in which the synchronic meanings and meaning links are represented; these meaning links are also to be interpreted diachronically, as semantico-­conceptual bridges between the senses, based on the processes of metaphor, conventionalisation of inference (i.e. metonymy), generalisation and lambda-­abstraction (Jurafsky 1996, 555). The meaning ‘child’ is assumed to constitute both the synchronic central meaning of the diminutive and the etymological source from which the other meanings developed. It will be shown below that this picture drawn by Jurafsky might not be valid in all respects for all languages. For some of the proposed changes, the assumption of universal unidirectionality seems to be too strong: for example, many diminutive suffixes in Indo-­ European languages have developed from derivational affixes meaning ‘related-­to’ or ‘similar-­to’ (Magni 1999), meanings which are described by Jurafsky as being only later

Evaluative Morphology in a Diachronic Perspective 145 SEMANTICS

related-to G

imitation

exactness

small-type-of

G, M

partitive

CATEGORY CENTRALITY IS SIZE

L

I SOCIAL GROUPS ARE FAMILIES

I affection

M

small M

M I

I

approximation

GENDER IS SIZE

L

child

pets

L, M

member

female M

intimacy

M contempt

sympathy

hedges PROPOSITIONS ARE OBJECTS

PRAGMATICS G: generalisation, I: inference, M: metaphor, L: lambda-abstraction Figure 11.1  Radial category of the diminutive (Jurafsky 1996, 542, figure 5)

results of changes of diminutive elements. A revised model of the diminutive category representing the polysemy and diachronic paths of semantic change will be given in Section 3.2. A comparable research study describing the cross-­linguistic polyfunctionality and diachronic paths of change of augmentative elements does not yet exist; exceptions are the studies of Matisoff (1991) for the South-­East Asian languages and of Grandi (2002) for the Mediterranean languages. In Section 3.3 a model for the augmentative category will be presented which summarises the paths of semantic change described in this section. Substantial cross-­linguistic works on pejoratives or amelioratives from a diachronic perspective do not exist, though remarks concerning the etymology and historical development of individual evaluative affixes can be found in historical grammars and research literature. A schematised model in Section 3.4 presents the paths of semantic change leading to pejoratives described in the section. The following presentation of functional changes in evaluative morphology makes no claim to being exhaustive. Because of space restrictions, a choice had to be made to portray both the most common paths of change detectable in evaluative morphology and some more special cases of development. The focus, though, will be on the processes observed in the Romance languages, which offer a broad spectrum of different types of change.

146

Katrin Mutz

3.2  Etymological aspects and paths of change: diminutive affixes The rise and development of evaluative formatives, in most cases, can be described either as a grammaticalisation process or as a process of refunctionalisation. In the first case, elements with an autonomous syntactico-­lexical status and with referential meaning develop into evaluative clitics or evaluative affixes. In the second case elements already having an affixal status acquire, in specific contexts, an evaluative function often conserving their old meaning in other contexts, with the result of polysemous or, later on, homonymic affixes. Other types of change to be observed are, for example, resegmentation processes. As evidenced in typological research, the diminutive category is highly represented in the languages of the world and not restricted to any particular morphological language type. The preferred morphological device for diminutives cross-­linguistically is ­suffixation followed by reduplication (Štekauer et al. 2012, 325).5 In many languages, different ­synthetic and/or analytic procedures coexist; this variation shows the dynamics and ­inherent potential for change of the diminutive category. Cross-­linguistically, there seems to exist a slight tendency towards constant renewal of the diminutive category and its formatives. At least for the Indo-­European languages it has been shown that ‘diminutives seem a dynamic and unstable linguistic strategy, which, in the course of ages, has undergone a wide (cyclic?) renewal’ (Grandi 2011b, 5). The tendency to and need for renewal can be explained by the increasing loss of expressive power of the diminutives in the course of time due to exaggerated use. Another result of this semantico-­functional attrition is the widely attested existence of lexicalised diminutives (Bauer 1997, 551), e.g. German Mädchen ‘girl’, Kaninchen ‘rabbit’; Italian coltello ‘knife’, uccello ‘bird’; French soleil ‘sun’; Slovak stolička ‘chair’; Irish cáilín ‘girl’. The diminutive formatives cross-­linguistically mainly tend to arise by the following types of processes: grammaticalisation, refunctionalisation, resegmentation and merger.6 3.2.1  Where do diminutive suffixes come from? Grammaticalisation In many African and Asian languages the diminutive formatives have emerged via a process of grammaticalisation, whereby an erstwhile referential noun with the meaning ‘child’ or ‘son’ has become a diminutive element (Matisoff 1991; Heine et al. 1991, 94–7; Jurafsky 1996; Creissels 1999; Heine and Kuteva 2002, 65f.; Bakema and Geeraerts 2004, 1051). The main semantic process involved is a metonymic one: a salient, inherent property of the child, i.e. smallness, has become the central meaning of the respective element via context-­induced inferencing (Heine et al. 1991, 101), e.g.: (1) a. Ngiti: ingba ‘child’ > dim -­ngba, e.g. itsù-­ngba ‘small tree’ (Ricca 2005, 209) b. Ewe: ví ‘child’ > dim -­ví, e.g. du-­ví ‘small village’ (Heine et al. 1991, 79–97) c. Akan: ɔba ‘child’ > dim -­wa/-­ba (Appah and Amfo 2011) d. Noon: kowu ‘child’ > dim ku-­, e.g. ku-­doo ‘small stick’ (Soukka 2000, 73) e. Vietnamese: con ‘child’ > dim con, e.g. > bàn con ‘small table’ (Grandi 2011b, 20) Other examples of diminutive formatives presenting this kind of etymology are -­r and -­zi in Mandarin (Matisoff 1991, 15), -­zai in the Cantonese Yue dialect Lianjiang (Huayong and

Evaluative Morphology in a Diachronic Perspective 147 Zhe 2008), -­pil in Nahuatl (Karttunen 1990); the Proto-­Bantu nominal root *jána ‘child’ has been reconstructed as the etymological source of the diminutive marker in many South Bantu languages, e.g. in Lingala the prefixal formative mwâ-­ or in Venda the diminutive suffix -­ana (Heine and Kuteva 2002, 65f.; Creissels 1999, 30). Another cross-­linguistically attested path of grammaticalisation is the development of an autonomous adjective with the meaning ‘small’ into a (prefix(oid)al) diminutive element, e.g.: (2)

a. French petit ‘small’ > Haitian dim ti-­, e.g. ti-­chen ‘puppy’, ti-­moun ‘child’, ti-­ kiyè ‘tea spoon’ (Lefebvre 2003, 41) b. Swedish små ‘small’ > dim små-­, e.g. småbil ‘small car’, småkyrka ‘small church’ (Bakema and Geeraerts 2004, 1051) c. Ancient Greek micros ‘small’ > Modern Greek dim micro-­7

Refunctionalisation In many Indo-­European languages the inventory of the existing diminutive category has been renewed by recruiting affixes from other derivational domains. Many of the diminutive suffixes can be traced back to suffixes deriving denominal adjectives or nouns with the meaning ‘related-­to’ or ‘similar-­to’ or ‘origin-­of’ (Hasselrot 1957; Magni 1999; Mutz 2000; Grandi 2011b), e.g. (3) a. Lat. -­aster, e.g. patr-­aster ‘stepfather’ > Lat. dim -­aster, e.g. porc-­aster ‘piglet’ (Thomas 1940, 521) b. Common Slavic *-­iko > Slov. dim -­ica, e.g. kôz-­ica ‘little goat’ (Grandi 2011b, 18) c. Sanskrit -­(i)ka > dim -­ká, e.g. vamra-­ká ‘little ant’ (Lühr 2008, 229) d. PIE *-­(e)ino > Lat. -­inus, e.g. mar-­inus ‘belonging to the sea’ > Ital. dim -­ino, e.g. lett-­ino ‘small bed’ (Brugmann and Delbrück 1906, 271ff.)8 How did this refunctionalisation of the originally relational suffixes in the Indo-­European languages come about? Several plausible semantic archetypes exist. One plausible reconstruction assumes that the birth of the diminutive meaning lies in derivations indicating (patronymic) names (e.g. cognomina in Latin or proper nouns in German; Wrede 1908, 132): 9 the ‘pertaining-­to X’ meaning in this context is reinterpreted as genetic origin, i.e. ‘son/daughter of X’. A prototypical example for this development is Lat. -­inus, which became the most productive diminutive suffix in Italian: -­ino (Mutz 2000, 190–206).10 The change of meaning from ‘belonging-­to-­X’ via ‘son-­of-­X’ to ‘small-­X’ and ‘little-­X’ came about by the semanticisation of context-­induced inferences. These inferences can be based both on reinterpretational hearer strategies and on innovating speaker strategies used for expressive purposes. Another, similar reconstruction situates the origin of the diminutive meaning in the denominations of descent with animal nouns (e.g. Plank 1981, 66f.): ‘belonging-­ to-­ X’ > ‘offspring-­ of-­ X’ > ‘small-­ X’/‘little-­ X’. In each of the two scenarios the genealogical relation between animate beings being reinterpreted as a dimensional relation is crucial. The etymological origin of these IE diminutives corresponds to the one described above for the African and Asian diminutives (Grandi 2011b). After the emergence of the quantificational meaning (‘small-­X’) in the context of derivational bases denoting animate beings, an analogical expansion to other d­ erivational

148

Katrin Mutz

nominal bases with concrete meanings (e.g. artefacts) and, later on, with abstract meanings follows. This trend of change corresponds to universal unidirectional clines elaborated in grammaticalisation studies (Heine et al. 1991, 31). Language-­specifically, a further extension to non-­nominal bases can be observed (Mutz 2000) resulting in a concomitant reinterpreted diminutive meaning, e.g. an approximative sense with adjectival and verbal bases.11 Merger and/or reanalysis Another kind of renovation leading to quasi-­complex diminutive affixes is the segmental increase of an existing diminutive form. This may be by concatenation of several evaluative affixes (affix telescoping) in cases in which the first diminutive element is not expressive, productive or transparent any more (Dahl 2006, 594), e.g. German -­lein (Fleischer and Barz [1997] 2007, 179), Ancient Greek -­ídion (Haspelmath 1995, 4f.), Czech -­inka (Belič 1901, 140). Or it may be by reanalysis and resegmentation of morphological material (from the root or another affix) and subsequent analogical extension, as was the case with the following diminutive suffixes: Bulgarian -­le/-­če (Manova 2005, 237), Greek -­ákion (> Modern Greek -­aki; Jannaris 1897, 292), German -­ling (Fleischer and Barz [1997] 2007, 164), Latin -­unculus (Haspelmath 1995, 11). 3.2.2  Further semantic and functional change of diminutive affixes From quantifying meaning to qualifying meaning Diminutive affixes not only have the quantifying meanings ‘small X’, ‘young X’ etc. but can also, context-­dependently, refer to qualitative characteristics such as ‘miserable X’, ‘sweet X’.12 In the long run such context-­dependent interpretations can semanticise and lead to polysemous affixes. Italian -­uccio is such a diminutive affix which never occurs with a pure quantificational meaning but, dependent on context, expresses either an ameliorative or a pejorative meaning besides the quantifying one. Later on the quantifying meaning can get lost, with, then, only the qualifying meaning being semanticised. This diachronic semantic pathway corresponds to unidirectional universal paths of semantic change (Heine et al. 1991, 157) and can be described as the first phase in a subjectification process (Traugott 1995, 32): meanings based on the external situation become meanings based on the internal (evaluative/perceptual/cognitive) situation. From quantifying/qualifying meaning to pragmatic functions Diminutives, cross-­linguistically, are polyfunctional elements having both semantic and pragmatic meanings (Dressler and Merlini-­Barbaresi 1994); these pragmatic functions are secondary meanings (Jurafsky 1996). The semantico-­functional change can be described as a pragmaticalisation process (Günthner and Mutz 2004) in which the pragmatic functions of the diminutive arise due to the conventionalisation and semanticisation of conversational implicatures and context-­dependent inferences. In many respects the process resembles the subjectification processes described by Traugott (1995, 32). From a diachronic perspective, pragmatic functions such as mitigation, downgrading of illocutionary force or marking of politeness or honorification emerge via metonymical processes; synchronically, the links between the semantic meanings and pragmatic functions can be interpreted as metaphorical ones: e.g. the smallness of the referent is like the smallness of the request. Concluding this subsection, I propose a simplified and revisited model à la Jurafsky (Figure 11.2), which shows the network-­like panchronic category of the diminutive,

Evaluative Morphology in a Diachronic Perspective 149

SEMANTICS imitation exactness

partitive

related-to, origin-of

approximation bad

good child

female

small

affection

hedges pets

contempt intimacy

politeness PRAGMATICS

Figure 11.2  Revised radial category of the diminutive (Jurafsky 1996, 542, figure 5)

whereby the synchronic links between the senses only partially correspond to unidirectional diachronic paths.13 3.3  Etymological aspects and paths of change: augmentative affixes Augmentatives14 seem to imply the existence of diminutives, i.e. cross-­linguistic and typological studies have revealed that whereas languages having diminutives exist without the augmentative counterpart, languages with augmentatives tend to have diminutives as well (Bakema and Geeraerts 2000, 1046).15 This implicational hierarchy is also shown in language acquisition studies (Savickienė and Dressler 2007a, 343) and can be interpreted in diachronic terms: historically, diminutives emerge prior to augmentatives; augmentatives tend to be recent innovations. A case in point is the Romance languages: whereas Latin had a number of diminutives suffixes at its disposal, augmentative affixes were lacking. The Romance languages renewed the inherited diminutive inventory and developed, with the exception of French, an augmentative category (Mutz 2000). The same is true for Modern Greek vs. Ancient Greek. Grandi (2011b) sees a causal link between the renewal of the diminutive category and the emergence of an augmentative category, the latter triggering the former. As in the case of diminutive markers, augmentative affixes usually also arise either due to grammaticalisation processes or via processes of refunctionalisation. But whereas in the case of diminutives there seems to be a common cross-­linguistic preference for one specific conceptual domain as etymological source or reinterpretation context (namely ‘child’ or ‘son’), in the case of augmentatives the sources seem to be more diversified and seem to depend more on area-­specific and language-­family-­specific tendencies (Grandi 2011b).

150

Katrin Mutz

3.3.1  The semantic origin of augmentatives Grammaticalisation In many (South-­)East Asian and some African languages, grammaticalisation from a noun with the meaning ‘mother’ to an augmentative marker took place (Matisoff 1991; Jurafsky 1996; Creissels 1999; Anonby 2011), e.g. in White Hmong (niag), Malay (ibu), Vietnamese (cài), Mandingue (-­ba) and Mambay (tí).16 We have here an interesting counterpart for the grammaticalisation path from the noun ‘child’/‘son’ to a diminutive marker also observed in these languages (see Section 3.2). The semantic change can once again be explained by the conventionalisation of metonymically induced contextual implicatures as well as by contrasting analogies: ‘mother’:‘child’ – ‘big’: ‘small’. A similar path of change can be detected in many Sub-­Saharan African languages in which the augmentative element can be traced back to a noun with the meaning ‘woman’ or ‘female’ (Creissels 1999): e.g. ‘the proto-­Bantu word *kádi “woman” has given rise to a new augmentative suffix -­hadi/-­kati’ (Grandi 2011b, 20), for instance in Sotho, as in sefate-­ hadi ‘big tree’, or in Swati, as in umutsi-­kati ‘big tree’.17 Interestingly enough, nouns with the meaning ‘man’ or ‘male’ do not seem to have been grammaticalised as augmentative affixes, though there are languages with a strong correlation between masculine gender and augmentative meaning (e.g. in Maasai; Payne 1998). Still another path of grammaticalisation is the development from an autonomous adjective with the meaning ‘big’ to a (prefix(oid)al) augmentative element: the fate of the Greek augmentative macro-­is a case in point (Grandi 2002, 203). Refunctionalisation In the Indo-­European languages it is not only the main diminutive suffixes that can be traced back to refunctionalised relational suffixes. In Modern Greek and Romance the augmentative suffixes have a similar origin (Grandi 2011b). The Romance augmentative suffixes (Ital. -­one, Span. -­ón, -­azo, Port. -­ão, -­ona, -­aço) correspond to non-­augmentative Latin derivational suffixes expressing different types of relational meanings, depending on the context: in the case of Latin -­o/-­onis, the relational meaning from which the augmentative sense took its origin via a metonymy-­based reinterpretation is the individuating, possessive meaning in cognomina, e.g. Naso, Nasonis ‘man with a big nose’ > Ital. nasone ‘big nose’ (Gaide 1988). For Latin -­aceus, both the meaning of resemblance and the meaning of consistence which developed into a collective meaning seem to have been crucial for the emergence of the augmentative sense (Mutz 2000, 214–21). The step from collective meaning (‘many, much’) to augmentative meaning (‘big’) seems to have been triggered by the extension of the derivational base from a mass noun to a count noun. Derivational bases referring to countable entities which typically occur in huge masses, like *foliacea ‘foliage’, take on an intermediate position (Gamillscheg 1921). The rise of the augmentative meaning out of a collective one can be also observed in Modern Greek (Grandi 2011b, 12f.). A change from a derivational affix with locative meaning to an augmentative affix took place e.g. in Slavic languages (Grandi 2011b, 13ff.), for instance Old Church Slavic -­ište > Bulgarian det(é)-­ište ‘big child’, Macedonian čoveč-­ište ‘big man’. The cross-­linguistically used augmentative prefix super-­ (as in supermarket, supercool) also has a locative origin (the Latin spatial preposition super).

Evaluative Morphology in a Diachronic Perspective 151 3.3.2  Further semantic and functional change of augmentative affixes From quantifying meaning to qualifying meaning and beyond Like diminutives, augmentatives may also ‘develop beyond the simple indication of size in several different ways’ (Dahl 2006, 594). One strong cross-­linguistic developmental tendency of augmentative formatives is their shift to pejorative meaning (less frequently to ameliorative meaning, but see note 1) due to context-­induced reinterpretations, with an intermediate phase in which the augmentative is ambiguous or polysemous. This meaning change can be schematised as follows: augmentative meaning A > augmentative meaning A + pejorative conntotations B > polysemous formative (A/B) > pejorative meaning B (loss of augmentative meaning). Cases in point are the development of Slavic -­ište (Grandi 2002, 289f.) and the development of Latin -­aceus in Romance (Malkiel 1959; Gamillscheg 1921): in the Ibero-­Romance languages and in Old Italian the suffix has augmentative meaning, whereas in Modern French and Modern Italian the only productive meaning of the suffix is a pejorative one. The postulated semantic path seems to be a one-­way street, i.e. pejorative formatives do not seem to change into augmentative formatives. Augmentatives also tend to develop secondary pragmatic functions, and can function, for example, as exaggeration markers or upgraders of illocutionary force (Dressler and Merlini Barbaresi 1994). To conclude, I propose a panchronic meaning model of the augmentative which takes account of the previously described semantic bridges between the meanings attested cross-­ linguistically (Figure 11.3). 3.4  Etymological aspects and paths of change: pejorative affixes As shown in Sections 3.2 and 3.3, many pejorative affixes cross-­linguistically are former diminutive or augmentative elements, thereby following a general path of semantic SEMANTICS related-to resemblance consistence individuation

old

collective

mother big

good

bad

female

exaggeration contempt PRAGMATICS

Figure 11.3  Radial category of the augmentative

Katrin Mutz

152 SEMANTICS related-to

resemblance

consistence collective female small big

bad

contempt PRAGMATICS Figure 11.4  Radial category of the pejorative

change from concrete (quantifying) to more abstract (qualifying) meaning, as described e.g. in many grammaticalisation studies (Heine et al. 1991; Hopper and Traugott 1993). Depending on the context, the meaning of ‘small’ can be interpreted as ‘small’ or as ‘minor in quality’, or the meaning of ‘big’ can be interpreted as ‘coarse’. These context-­induced interpretations can conventionalise and semanticise through constant use, leading to the emergence of pejorative affixes with the original diminutive or augmentative meaning bleached out. Examples are the Italian pejorative suffix -­accio, which still had augmentative meaning in Old Italian. Pejoratives can also arise by semanticisation of negative connotations of collectives, e.g. Italian -­aglia (as in soldataglia ‘group of bad soldiers’; Ricca 2005, 203) or -­ume (as in politicume ‘group of bad politicians’; Grossmann 2004, 246), or by semantic irradiation, e.g. the fate of Latin -­aster (Mutz 2000, 179–84): Lat. patraster ‘stepfather, father-­ in-­law’, Ital. poetastro ‘bad poet’, French bellâtre ‘pretty boy (pej.)’. Like diminutives and augmentatives, pejoratives are also linked to the concept of ‘female’: in Japanese the pejorative suffix -­me was originally a noun with the meaning ‘female’ (Labrune 1999). See Figure 11.4.

4  Conclusion: trends of change in evaluative morphology Existing evaluative categories tend to be renewed in the course of time. The renovation of an existing evaluative category can be necessary because of the bleaching-­out or total loss of the evaluative meaning in existing derivations, as a result of extensive and conventionalised use (Dahl 2006, 594). The renewal (Hopper and Traugott 1993, 121f.) typically tends to be achieved by the recruitment of other forms of the respective language via ­grammaticalisation processes and/or refunctionalisation (Sections 3.2 and 3.3) or via borrowing.18 The semantic paths of change observed seem to be language-­family specific or language-­area specific, in particular concerning the augmentatives (Grandi 2003b): certain

Evaluative Morphology in a Diachronic Perspective 153 language families/areas seem to prefer certain conceptual domains as attractors of evaluative sense (e.g. in Asian languages ‘mother’ > augm, ‘child’ > dim). Cross-­linguistically, the semantic source of the diminutive and augmentative meaning tends to be anthropocentric. Many evaluative affixes in today’s languages of the world are the outcome of merger procedures, i.e. the morphological fusion of formerly separate morphological entities. The fusion can also occur between morphemes and resegmented material. In addition to the emergence of new evaluative categories and/or formatives, existing diminutive and augmentative formatives tend to undergo unidirectional meaning shifts from quantifying to qualifying meaning, thereby evolving e.g. into pejoratives. Typically, secondary pragmatic functions emerge via universal paths of semantico-­ functional change (e.g. tendencies of subjectification). The semantic changes observed are mainly metonymy-­ driven and are characterised by the semanticisation of context-­ induced inferences. The focus in this chapter has been on the cross-­linguistic tendencies of emergence and semantic change of evaluative affixes. Many aspects of the history of evaluatives have had to be neglected because of space restrictions or research gaps (e.g. only a few diachronic studies on pejoratives, non-­Indo-­European evaluative morphology or evaluative reduplication exist). Future studies should fill these research lacunae.

Acknowledgements I am very grateful for comments and help from Björn Wiemer and Martin Haase and the two editors of this handbook.

Notes   1. Since, interestingly enough, genuine (suffixal) morphological amelioratives seem to be lacking cross-­linguistically (cf. Jurafsky 1996), there will be no special paragraph devoted to ameliorative formatives. Most commonly, the attested amelioratives are semantically shifted diminutive or augmentative affixes which express ameliorative meanings in some contexts but in others still preserve their quantifying sense. An exception is the existence of the cross-­linguistically used Latin and Greek ameliorative prefixes super-­, hyper-­, mega-­ etc., which are originally spatial prepositions or dimensional adjectives, e.g. German Superabenteuer ‘great adventure’, French superfête ‘fantastic party’ (Ruf 1996; Grandi 2002, 205–8).   2. This might be due to the relative paucity of augmentative markers in comparison to diminutive ones (Livia Kortvélyessy, p.c.).   3. The process of coalescence is characteristic of grammaticalisation processes (see e.g. Traugott and Hopper 1993).   4. The innovating semantic mechanisms driven by communicative needs have to be distinguished from processes of conventionalisation and lexicalisation which are governed by social forces and conditions (Croft 2000b; Rainer 2005).   5. This is one of the reasons why we will focus on the development of diminutive suffixes. More recent studies on diachronic aspects of reduplication, though not restricted to reduplicative evaluative morphology, are gathered in issue 19 of Morphology (2009).   6. A cross-­linguistically rather uncommon type of change is to be found in Nahuatl, where the body part noun tzin with the meaning ‘anus’ gave rise to the diminutive and honorific suffix -­tzin (Stolz 1994, 79–82).   7. The prefix is a globalised one, found in many languages, predominantly in technological terms.

154

Katrin Mutz

  8. The relational PIE suffix *-­(e)ino is also the origin of diminutives in Greek, Germanic and Baltic languages (Ambrazas 1993).   9. Some scholars stress the hypocoristic value as crucial for the rise of the diminutive meaning in this context (e.g. Henzen 1965, 143). 10. Reichmuth (1956, 74): ‘Sueton [. . .] sagt: “Praeconinus vocabatur, quod pater eius praeconium fecerat”.’ This development can also be plausibly reconstructed for the Romance suffixes -­tt-­ and -­cc-­of uncertain etymological origin, because their first attestations occur with feminine proper nouns (Mutz 2000, 227). 11. This corresponds diachronically to the universal preference hierarchy with respect to the categorial nature of the derivational base (n > adj, v > adv, num, pro, interj > det; Bauer 1997, 540). 12. The description of the cross-­linguistically widely attested change from a diminutive marker to a marker of feminine gender or sex is beyond the scope of this chapter (Jurafsky 1996). 13. Whereas from a diachronic perspective the centre of the diminutive radial category seems to be the meaning ‘child’, synchronically the meaning ‘small’ is assumed to be the central one (see Prieto, this volume). This shift of focus is based on metonymy. 14. Affixoidal augmentative elements such as German Mords-­ ‘murder’ (as in Mordsgeschäft ‘big deal’), Hammer-­‘hammer’ or Bomben-­‘bomb’ will not be considered. These formations tend to have ameliorative meaning as well (Ruf 1996). 15. The Aboriginal language Nhanda seems to be an exception: it has an augmentative suffix -­barta but no morphological diminutive formative (Blevins 2001). 16. A related meaning change can be observed in Basque: the adjective tzar ‘old’ developed into an augmentative suffix -­tzar (Martin Haase p.c.; de Rijk 2008, 85), acquiring pejorative meaning as well. 17. The concept ‘female’ is linked cross-­linguistically either to the diminutive or to the augmentative sense. Cultural differences seem to play a crucial role in the divergent and contrasting developments. 18. Diminutive affixes tend to be easily borrowed cross-­linguistically (Matras 2007). Describing the role of language contact in the development of evaluative morphology is beyond the scope of this chapter (see e.g. Grandi 2002).

12  Evaluative Morphology in Sign Languages Giulia Petitta, Alessio Di Renzo and Isabella Chiari

1 Introduction: the place of morphology in the expression of evaluative functions in sign languages Addressing the question of evaluative morphology and its specificity in sign languages (SLs), whether in a comparative perspective or with a language-­specific purpose, requires the clarification of the complex and often blurred relationship among the notions of morphology, prosody and suprasegmental features in SL, a relationship far from established and clear-­cut in the literature. The aim of this chapter is to pose some questions about evaluative functions as realised in SL, with special reference to Italian Sign Language (LIS), in order to assess possible interactions among different roles morphology and prosody play in SL, and to show some tendencies in the interaction of evaluative strategies. SLs are characterised by the use of the body as a whole to convey discourse, using both manual (hand movements, hand configurations, contacts with the body/face) and non-­ manual components (mouth/tongue/cheek, eye gaze, facial expressions, eyebrows movement). The difference in modality between spoken/vocal languages (VLs) and SLs defines not only their structure but also the cognitive and formal mechanisms that are productive at discourse level (Bellugi and Fischer 1972; Emmorey 2002). When trying to apply common linguistic categories used in linguistics, mainly conceived on representations of VL, we often overgeneralise similarities and apply categories using different, often not explicit, analogies. SL descriptions have been considered to be grounded in an ‘assimilationist’ or a ‘non-­assimilationist’ perspective. While the first adopts VL-­based linguistic categories and looks at SL through these lenses, the second aims at discovering peculiarities in SLs, focusing on their specificity (Cuxac 2000; Vermeerbergen 2006). Adopting an assimilationist approach, it is possible to compare VL and SL, at the risk of over-­extending categories of the functional or formal properties of the languages without taking into account the overall picture that derives from the different modalities involved. It is nevertheless not possible to describe SL without relying on existing linguistic notions; still, it is necessary to give a full and explicit account of the use of terminology that can be misleading if not properly introduced. One of the key theoretical problems in using existing levels of analysis, typically forged for the description of VL, for SL is precisely the fact that the same functional features can be embodied in SL in linguistic layers normally not associated with that function in VL. One main issue is that of identifying the distinction between morphology and prosody (more specifically intonation), and between segmental and suprasegmental features, since the notion of segment itself is very problematic in SL research (Petitta, Di Renzo, Chiari and Rossini 2013).

156

Giulia Petitta, Alessio Di Renzo and Isabella Chiari

In order to address the issue of evaluative morphology in a semantic-­functional perspective, it is essential to agree on an explicit perspective for the use of notions shared in VL research but often used unsystematically in SL research. SL morphology, characterised by strong cross-­linguistic similarities, has been analysed as being of a double nature: sequential and simultaneous. Sequential morphology is characterised by a combination of signs (not all capable of appearing autonomously) that form a complex lexeme by being linearly concatenated. Sequential morphology in SL can ­represent complex words, compounds and affixation. Meir (2012, 83) observes that sequential morphology is not a preferred system used in SL for morphological distinctions. In simultaneous morphology ‘grammatical features are realized by altering the direction, rhythm, or path shape of the base sign’ (Aronoff, Meir and Sandler 2005, 309). Simultaneous morphology can be used both for inflection and for derivation, and seems strongly grounded in iconicity and spatial cognition. Typical examples of simultaneous morphology are verb agreement, verb aspect and classifier constructions. The preference of SL for simultaneous morphology has been grounded in production and perception factors depending on the visual-­gestural modality and on the relative slowness of manual articulators, thus compensated for by the simultaneous co-­occurrence of different morphemes (Bellugi and Fischer 1972; Emmorey 2002). The term ‘prosody’ is used in VL literature to include acoustic and perceptual ­properties of vocal sounds that are not specific to the lexical level, such as timing, stress and ­intonation, but when it comes to SL the term is also used in some cases to describe formal events that can be associated not with prosody but with functional features of intonation. Moreover, prosodic features in VL are generally associated with suprasegmentals and with properties that vary on a continuum and simultaneously to other levels, while in SL these aspects (prosody, suprasegmentality, continuity and simultaneity) interact differently with each other. In SL research, a parallel between intonation and non-­manual features of SL was proposed quite early (Woll 1981). Further on, the parallel has been questioned and reassessed. Sandler, basing her analysis on American Sign Language (ASL) and Israeli Sign Language (ISL), observes that ‘not all manual actions are lexical, and not all non-­manual articulations are prosodic. A useful working assumption is that a cue is prosodic if it corresponds to the functions of prosodic cues known from spoken language’ (2012, 69). Features considered equivalent to intonation are based on a mixed set of features of prosody, of which most are semantic, pragmatic and functional, others more specifically prosodic. We propose here to use the terms ‘prosody’ and ‘intonation’ without reference to their shared pragmatic and semantic functions in VL, since functions ordinarily performed through intonation in VL can also be expressed lexically, morphologically and syntactically. Moreover, in order to understand the specific use SLs make of different non-­manuals, we must consider the interaction with manual components and also their systematic character, their autonomy and the suprasegmental nature of the physical realisation. Thus we will be using the term ‘prosodic’ only to describe timing, rhythm and stress in SL. Comparing VL and SL also necessitates dealing with the problem of defining the role and the level to which paralinguistic features belong. In VL, paralinguistic features are defined by basing their identification and function on the channel and modality of their production, which are radically different from the ordinary phonic-­acoustic channel. In SL, there are expressions and functions of those features that are very similar to those appearing in VL (emotion expression, connotation, evaluation), but with the radical difference that they cannot be ascribed to a different modality.

Evaluative Morphology in Sign Languages 157 Even though many works have been devoted to SL morphology, no specific work has addressed the question of features and formal manifestations of evaluative functions. While some aspects that are in the domain of evaluative morphology have been discussed in the literature (mainly the formal devices used to express intensification and size variation), there has been no specific reflection on the roles, typologies and interactions of evaluation in the past, though, as we will show, a number of previous works describe specific evaluative functions from different perspectives. Some of the most relevant issues that will be discussed in this chapter relate to the identification of the main morphological means to express evaluation and their interaction in SL discourse, trying to define their place in the traditional distinction between simultaneous and sequential morphology. Moreover, SL evaluative functions pose the problem of interpreting the role played by non-­manuals and their radical complexity in SL. This complexity is due to the fact that the category of non-­manuals includes many different articulators, used for often overlapping functions in SL discourse and sometimes also overlapping with functions and formal expressions common to VL discourse. Are evaluative functions associated equally with all kinds of lexical elements that appear in SL discourse? Are there any restrictions or blocks depending on semiotic, semantic or phonological constraints? Do these functions tend to be employed differently in the case of canonical forms and classifiers? This last question brings forth the long-­debated issue about the different kind of lexical elements typical of SL, among which are special elements indicating semantic classes and often also functioning as proforms in SL discourse, often called classifiers.1 These elements can also be included in the image schema (features of the sign related to referents by iconic motivation) of a canonical form. For the purposes of this chapter, we need to observe a special kind of function of classifiers that is related to the formal properties of the handshapes and their role in canonical forms, thus including ‘size and shape specifiers’ and ‘handling classifiers’. This class of units can be called ‘interactive handshapes’, since ‘the handshape represents the hand itself as it either a) grasps, handles, or manipulates objects of certain shapes, or b) touches and strokes the surfaces, or traces the edges or other physical features of some real or imaginary object’ (Johnston and Schembri 1999, 120). We have chosen to concentrate on three main evaluative functions that are very common in SL: augmentation and diminution (of quantity or of quality) and intensification. And we have restricted our discussion to evaluative functions related to entities and properties, thus excluding the consideration of processes and actions even when they are largely present in all SLs.2

2  Main evaluative strategies in SL, with special reference to LIS If we start by asking ourselves what the ways to express the evaluative function for augmentation and diminution in SL are, we find a great variety of means, ranging from the lexical level to the morphosyntactical and to the strictly morphological. In all SLs, augmentation and diminution can be expressed by lexical sequences of canonical forms (such as ‘house’ + ‘big’, ‘rat’ + ‘small’). In these cases a specific lexical item for ‘big’ and for ‘small’ is produced in its canonical form (both manual and accompanied by non-­manual features; see Figure 12.1). We will not detail these uses further because they are not directly related to evaluative morphology. We will discuss simultaneous and sequential means used in different SLs to express evaluation, considering further whether these means can be conceived as morphological.

158

Giulia Petitta, Alessio Di Renzo and Isabella Chiari

+

(a)

(b)

Figure 12.1  LIS: (a) ‘house’ + (b) ‘big’ (Romeo 1991)

We will focus our analysis on augmentation and diminution, illustrating their relationship with intensification functions. We propose to observe evaluating functions of augmentation and diminution as governed by three main evaluative strategies concerning morphology in different ways: (i) manual evaluation: (a) manual sequential evaluation – sequence of a canonical sign and an interactive classifier; (b) manual simultaneous evaluation – modification of manual features of the sign, by variation in movement, handshape or location; (ii) non-­manual simultaneous evaluation: co-­occurrence of additional or modified non-­manual features; (iii) reduplicative evaluation (sequential and simultaneous): partial or full reduplication of the sign or a portion of the sign, with eventual variation in manual patterns. 2.1  Manual evaluation 2.1.1 Manual sequential evaluation: sequence of a canonical sign and an interactive classifier In many SLs we observe a strategy for expressing size (and quantity) that consists of a sequence of two elements, the first being the canonical form of a sign (for example ‘fish’ in LIS) followed by specific interactive handshapes (‘trace’ or ‘handling’) that indicate the large/small size (often accompanied by non-­manuals). See Figure 12.2. The status of interactive handshapes in SL is very complex, since they fulfil many functions including anaphoric reference, spatial orientation and verb agreement. In the case of size expression they play a role that can be conceived as both lexical (autonomous) and morphological. In the cases presented above, their role is, in some respects, similar to that of certain types of classifiers in VL. They are often considered bound morphemes (Supalla 1986; Zwitserlood 2012), but they also exhibit properties of lexical items and proforms. The second

Evaluative Morphology in Sign Languages 159

+

(a)

(b)

Figure 12.2  LIS: (a) ‘fish’ + (b) ‘big’ (Romeo 1991, modified)

part of the sequence (the interactive classifier) is not iconically connected with the specific image schema governing the representation of the canonical sign, though it is directly related to its referent. Although the same interactive signs can be used as proforms for anaphora in SL discourse, in these specific evaluative constructs they are bound to the immediately preceding occurrence of the canonical form of the base sign they modify, even though occasionally, when the context is very explicit, they can also follow other proforms representing the base. For this reason we favour considering the interactive classifier as a (relatively) bound morpheme that functions as an affix, belonging to sequential morphology in LIS. Affixes, such as the one shown in Figure 12.2, are recurring, bound and multiple. The same affix can be attached to different sign bases that share some features, depending on the properties of the referent. The affix for augmentation in Figure 12.2 can also be applied to innumerable LIS signs, such as ‘mouse’, ‘sandwich’ and ‘pizza slice’. Many affixes of this kind are available in SL, all of which are interactive classifiers exhibiting tracing or handling iconic features. There is no supporting evidence in any SL so far for the use of manual sequential evaluation for size in an intensification function for signs referring to concrete or abstract qualities. 2.1.2 Manual simultaneous evaluation: modification of manual features of the sign, by variation in movement, handshape or location A further type of strategy used to express augmentation/diminution in quantity or quality is through the modification of movement properties of the sign by iconically enlarging them or reducing them. Augmentation and diminution can be obtained in LIS by enlarging/reducing the handshape of the sign (Figure 12.3),3 the distance between the hands (Figure 12.5) or the movement (Figure 12.6). It is important to note that signs that are subject to this kind of simultaneous evaluation are generally not preceded by the canonical form of the sign, unless for special emphatic-­pragmatic reasons or in predicative function.

Giulia Petitta, Alessio Di Renzo and Isabella Chiari

160

(a)

(b)

(c)

Figure 12.3  -­LIS: (a) ‘tie’ (canonical form), (b) ‘big tie’, (c) ‘small tie’

(c)

(d)

(b)

(a) Figure 12.4  LIS: gradual modification in size of different handling patterns (a–d)

In Figure 12.3 we can observe the LIS sign for ‘tie’ and its augmentative and diminutive forms. Modification in the handshape is iconically related to the image schema of the base sign in its canonical form, which represents the unmarked norm for that sign. More specifically the tie is represented by the direct reference to the location and, partially, to the handling of the object, including the activity of its tying. Augmentation is achieved in LIS by modifying the handling handshape in order to express augmentation or diminution, referring to the handling pattern. The handling pattern handshape can be varied in size in a regular way that can be involved in all occurrences of a LIS sign where the pattern is used iconically (Figure 12.4). Enlargements and reductions of these kinds are only possible when the handshape is employed iconically in the canonical form of the sign. But it is also possible to use enlargement/reduction variation in the general pattern of a two-­handed sign by enlarging/ reducing the distance among the hands in respect to the realisation of the canonical form in the signing space. In the examples in Figure 12.5, the image schema refers to the object, depicting its dimension and giving a sort of (abstract) reproduction of its size. It is interesting to note that the sign for ‘street’ in ASL is identical to the LIS sign and can be modified by the same evaluative strategy.

Evaluative Morphology in Sign Languages 161

(a)

(b)

(c)

Figure 12.5  LIS: (a) ‘street’ (canonical form), (b) ‘large street’, (c) ‘narrow street’

(a)

(b)

Figure 12.6  LIS: (a) ‘project’ (canonical form), (b) ‘big project’

The same modification pattern (by enlargement/reduction of movement) can be employed in LIS to vary signs that represent abstract entities (see Figure 12.6). The image schema refers to the project design (the same sign means ‘design’ or ‘picture/make a picture’), and the movement can be enlarged, and eventually slowed down, to express the meaning ‘big project’. Other SLs also attest this kind of manual evaluation for size properties. For example, in Adamorobe SL (AdaSL) the sign to indicate a ‘small knife’ is a modification of the canonical form for ‘cutglass’ by having a bent-­B handshape and smaller movement (Nyst 2007, 152). British SL (BSL) (Sutton-­Spence and Woll 1999, 202–3) and ASL (Valli and Lucas 1992, 153–4) also report this kind of evaluative procedure. Most canonical forms are articulated on the body or on the signed space in a conventional dimension that is organised on the principle of least effort, granting the receiver perceptual saliency to detect small movements and fine-­grained differences and transitions in handshapes. Reducing the regular value of a sign can expose the signer to unusually small movements and the receiver to the need for further attention. Signers thus claim preference for lexical evaluative diminution or for non-­manual diminution. Manual simultaneous evaluation can also be employed in LIS, BSL and ASL to express evaluative functions of intensification. In ASL and Australian SL (Auslan), Johnston and Schembri (1999, 154) observe that intensification of the meaning of a sign can be achieved by increasing the amplitude of the sign (manual simultaneous evaluation) and by puffing the cheeks (non-­manual simultaneous evaluation). Figure 12.7 shows the BSL sign for ‘angry’ (a) modified to express intensification (b). While ASL tends to prefer lexical expression and manual sequential evaluation for size/ shape properties of concrete and abstract entities, it still relies heavily on manual simultaneous evaluation for intensification of properties and processes (a similar procedure is employed for aspectual variation of predicates). Both LIS and ASL tend to prefer manual simultaneous evaluation when interactive classifiers are involved in the sign, often in

Giulia Petitta, Alessio Di Renzo and Isabella Chiari

162

(a)

(b)

Figure 12.7  BSL: (a) ‘angry’ (canonical form), (b) ‘very angry’ (Kyle and Woll 1985, 152)

(a)

(b)

(c)

Figure 12.8  LIS: (a) ‘champion’ (canonical form), (b) ‘big champion’, (c) ‘outstanding champion’

c­ ombination with other evaluative types such as initial hold, manual stress and marked facial expression (mainly mouth gestures and eye movement). ASL and ISL also employ manual simultaneous evaluation reduction with the evaluative function of approximation. For example, the canonical form for ‘blue’ is rendered with shorter reduced movements in order to achieve the meaning ‘bluish’. Intensification expression in LIS, ASL and Auslan cannot be identified with formal augmentation properties, but can also be expressed by diminution properties, depending on the semantic meaning conveyed by the modified base. Modifications in manual components can express qualitative intensification that also involves subjective judgements. In particular, hand movements can be enlarged, adding multiple rotations in the middle of the trajectory (Figure 12.8). The tendency, common to ASL and LIS, seems to be toward the use of manual simultaneous evaluation when the sign is a classifier, or when the canonical form of the sign implies the use of interactive classifier elements performing the main iconic function of size/shape representation. The use of classifiers is always iconic in those signs, and since sign classifiers are generally cross-­linguistically very similar in different SLs (Zwitserlood 2012, 158) this manual evaluative strategy can produce the same morphological form in different languages (as in ASL and LIS ‘street’ in Figure 12.5). Furthermore, many SLs, among them ASL and LIS, have been reported as sharing a similar process for intensification of entities and properties, called ‘initial hold’ and con-

Evaluative Morphology in Sign Languages 163 sidered a morphological tool of intensification by Bergman (1983, 5–6). It consists in a modification in the general timing of a lexical sign, by pausing the beginning of a sign and optionally slowing its movement. However, this process seems to appear in connection with other manual simultaneous evaluation features (only for augmentation) and seems not to be sufficient to produce an intensification in quantity or quality, or as an autonomous device for evaluation. Initial hold is often accompanied by what has been called ‘manual stress’, a variation in the rhythm of the sign, mainly making movement faster and more prominent. Since both initial hold and manual stress involve rhythmic and timing properties of the signed discourse, we prefer to consider them as supporting prosodic features with an evaluative function. 2.2 Non-­manual simultaneous evaluation: co-­occurrence of additional or modified non-­ manual features A second type of evaluation can be expressed by non-­manual features which refer to simultaneous morphology using different articulators. Many SLs use facial expressions in co-­occurrence with canonical forms in order to express qualities related to the size of the referent: some facial expressions and mouth patterns are dedicated to the specific expression of size, and often support or modify the corresponding manual sign. Figure 12.9 represents the facial expression used in LIS to refer to the ‘small’–‘big’ axis. These facial expressions are obligatorily present when the canonical form represents lexical items meaning, for example, ‘big’, ‘large’, ‘fat’ and ‘small’, and in those cases they represent non-­manual properties inherent to the lexical sign. These characteristic non-­manual patterns can also occur in conjunction with other signs expressing an evaluative role modifying the meaning of the base sign form. Thus when associated with lexical items meaning ‘bigness’ or ‘smallness’ they perform an obligatory lexical function, but when associated with other signs (mostly nominal), they are not obligatory (though they are very frequent) and add an evaluative meaning to the sign. The evaluative function of non-­manuals seems to be an expression of simultaneous morphology in SL. In particular, morphological evaluation using non-­manuals in most SLs seems to be associated with mouth patterns.4 A number of SLs, among them ASL, DGS (German SL),

(a)

1

2

3

4

(b)

1

2

3

4

Figure 12.9  LIS: different facial expressions with evaluative function: (a) ‘small’: 1. mouth kiss; 2. sucked-­in cheeks; 3. closed mouth protruded forward; 4. half-­ protruding tongue; (b) ‘big’: 1. puffed cheeks; 2. teeth on lips; 3. slightly grinding teeth; 4. half-­frown mouth

164

Giulia Petitta, Alessio Di Renzo and Isabella Chiari

(a)

(b)

Figure 12.10  LIS: (a) ‘big thing’, (b) ‘small thing’

BSL and LIS, have been reported as using the same kind of mouth evaluative patterns: puffing cheeks out for augmentatives and sucking cheeks in for diminutives (for DGS, see Pfau and Quer 2010, 385–6; for BSL, Sutton-­Spence and Woll 1999, 87; for ASL, Bickford and Fraychineaud 2008, 45). AdaSL (Nyst 2007, 134) also uses puffed-­out cheeks as augmentative for size. In all these languages the mouth pattern can be associated both with canonical signs and with interactive handshapes. Though it is generally agreed that these can be seen as morphological non-­manuals, these authors do not assign them to evaluative morphology, but consider them as ‘nonmanual adjectives’ (Pfau and Quer 2010, 385; Nyst 2007, 125). In our opinion, mouth gestures seem to have an actual morphological role, and can often express evaluation autonomously from other morphological means, as we can see in Figure 12.10. The manual components of the sign remain the same, but mouth gestures express the evaluative function of ‘big’ or ‘small’. In the first one, the teeth touch the lower lips in the second, there is a small protrusion of the tongue through the lips. The same evaluative strategy can be applied also to more abstract entities in LIS, such as the canonical sign for ‘newspaper article’. We have also observed that mouth and cheek patterns in LIS are often accompanied by other facial patterns regarding the movement of the eyes and eyebrows. For example, thin eyes and eyes wide open can be associated with mouth gestures, but their semantic and pragmatic value is context-­dependent; so they are not univocally associated with the pole of augmentation or diminution. It is thus possible to identify in mouth patterns the principal non-­manual articulator related to evaluative augmentatives and diminutives, both autonomously and in contexts where these patterns co-­occur with other evaluative strategies. There are nevertheless some phonological restrictions concerning the possible use of non-­manual morphological strategies. The main restriction relies on the fact that the mouth can be involved in various kinds of functions during SL discourse: in mouthings – very common with canonical forms (though not equally frequent across SLs) – or in the expression of lexical properties of certain signs that are articulated with locations touching the mouth, or with specific mouth gestures contributing to the lexical value of the sign. Thus non-­manual evaluation can be used only with the exception of cases where the mouth is already employed with different functions. The peculiar mouth patterns identified for augmentatives and diminutives regarding size have also been reported in DGS to mean intensification of abstract properties, though this

Evaluative Morphology in Sign Languages 165 kind of intensification is often also accompanied in all SLs with other non-­manual features regarding the overall facial expression (Pfau and Quer 2010). Operations included in non-­manual simultaneous evaluation, using only non-­manual markers to express qualitative properties, are also attested in evaluative expressions related to endearment and contempt. Facial expressions usually associated with canonical forms meaning ‘disgust’, ‘refuse’ and ‘disease’ (as obligatory features of the lexical unit) can develop functional evaluative roles when co-­occurring with illustrative signs and HIS, adding evaluation in the negative pole. On the other hand, endearment requires special, dedicated non-­manuals, in particular mouth protrusion in LIS, sometimes co-­occurring with head tilt and narrow eyes. These non-­manuals occur generally with illustrative signs or size/shape classifiers, following (or referring to) canonical forms related to animated subjects. Endearment can also be achieved in LIS by adding affective non-­manuals to interactive classifiers referring to something small. For example, it can be achieved with mouth protrusion during the production of the size classifier following signs referred to animated creatures, thus combining diminutive manual sequential evaluation with affective non-­ manual simultaneous evaluation. Facial expression patterns typical of affective evaluation (raised eyebrows and corners of the mouth downward with open mouth, protruding mouth) are not used systematically; instead they are used in a way that is very similar to paralinguistic features in VL (though more constantly employed than in VL), often sharing the exact formal patterns, and are often associated also to impersonation or role shift (where the signer literally impersonates the referent he or she is referring to, and plays the actions, processes, emotions or behaviour of the character he or she is portraying, from the character’s perspective; Cuxac 2000). Furthermore, they can be produced in more complex HIS structures where they can be extended over entire HIS sequences, and are thus also used on a super-­lexical range and can hardly be assigned to the domain of morphology. 2.3 Reduplicative evaluation (sequential and simultaneous): partial or full reduplication of the sign or a portion of the sign, with eventual variation in manual patterns Reduplication occurs very frequently in SL with different linguistic functions (Kyle and Woll 1985, 150–3; Sandler and Lillo-­Martin 2006, 50; Wilbur 2009). Reduplication is widely used, for instance, for pluralisation of signs or can have aspectual properties when applied to verbs (Fischer 1973). Wilbur (2009, 331) observes that reduplication in ASL generally consists of the unreduced stressed base followed by one or more copies that are generally unstressed and gradually reduced. The use of reduplication as morphological means of intensification is also very well attested in VL and significantly common in SL. The obvious reason for the extensive use of this formal operation relies on its iconic foundation. Thus it is not very surprising that a third type of evaluation in SL is precisely the use of different forms of reduplication. More specifically, we have observed a systematic typology of evaluative reduplication in LIS consisting in partial or full reduplication of a sign (sequential reduplication). Furthermore, reduplication can be form-­preserving or form-­changing when the duplicated portion is characterised by enlargement of the movement in the reproduction of the sign, thus combining reduplication with the evaluative strategy of manual simultaneous evaluation.

166

Giulia Petitta, Alessio Di Renzo and Isabella Chiari

Reduplication is used as a tool for expressing augmentation in the dimension of a lexical unit and also as intensification. Reduplication as an augmentative strategy tends to be more associated in LIS with partial and full reduplication in which the duplicated portion is further enlarged (sequential reduplicative enlargement), while reduplication with an intensification function can be realised both as a mere repetition of the full sign and as a copy with enlargement. Reduplication as an evaluative morphological strategy can be applied to occurrences of canonical forms and to interactive classifiers. Reduplication can be also interpreted as a special case of the manual sequential evaluation strategy in the context of sequential reduplicative enlargement. Sequential reduplicative enlargement occurs when the base sign in its canonical form contains an iconic image schema referring to the whole shape of the referent (as in the LIS sign for ‘big table’; Figure 12.11b). The copied portion is a modified form of the canonical one, and can be articulated with a different handshape but with the same movement ­patterns, reduced or expanded in the signing space. This kind of reduplication belongs to sequential morphology and involves the canonical lexical morpheme followed by what can be conceived as a partially reduplicative affix. Reduplicative enlargement can be seen as a special case of manual sequential evaluation, since the interactive classifiers in the affix behave as in the former morphological operation but exhibit strong formal similarities with the image schema of the canonical form. Reduplicative enlargement in LIS cannot be applied to a sign that is already enlarged by manual simultaneous evaluation (such as LIS ‘project’ or ‘tie’), since the enlarged portion is not represented by an interactive classifier, and, as previously stated, cannot be preceded by the citation of the base canonical form. A second type of sequential reduplication can be observed in Figure 12.12 for the LIS sign ‘good friend’. This kind of reduplication seems to be typical of signs articulated in the

(a)

(b)

Figure 12.11  LIS: (a) ‘table’ (canonical form), (b) ‘big table’

(a)

(b)

Figure 12.12  LIS: (a) ‘friend’ (canonical form), (b) ‘good friend’

Evaluative Morphology in Sign Languages 167

(a)

(c)

(b)

(d)

Figure 12.13  Auslan: (a) ‘angry’ (canonical form), (b) doubling, (c) manual stress, (d) doubling and manual stress (Johnston and Schembri 1999, 161)

(a)

(b)

(c)

Figure 12.14  LIS: (a) ‘true’ (canonical form), (b) ‘very true’ (simultaneous manual evaluation), (c) ‘very very true’ (doubling)

so-­called ‘neutral space’.5 In this case the reduplication can optionally involve enlargement of the copied portion, or only the mere repetition of the canonical form. Reduplicative sequential evaluation is not attested as an affective evaluation strategy. Thus, we can identify only two functions for this typology: intensification and authenticity/prototypicality. Other qualitative and subjective evaluation functions (for example, approximation or attenuation) seem to be conveyed ordinarily in LIS by HIS elements, specific lexical items or non-­manuals. In SL literature there has also been observed a special kind of reduplication that is called ‘simultaneous reduplication’ or ‘doubling’, where the articulation of a sign is realised using both hands instead of one. Simultaneous reduplication has been observed in Auslan with an intensification function, for example in signs for ‘bad’ vs. ‘very bad’. Johnston and Schembri (1999, 161) also argue that the doubling of a sign is generally insufficient to achieve intensification unambiguously and that the majority of one-­handed signs also require non-­manual features such as facial expressions or eye widening in order to express this kind of evaluation. These authors report that intensification can be achieved by doubling, by facial expression and manual stress, and by doubling combined with facial expression and manual stress (Figure 12.13). In some cases (such as the Auslan sign for ‘bad’) signers also report the possibility of exploiting doubling alone as intensification, without additional non-­manuals. The same strategy, optionally associated with non-­manuals, is also possible in LIS, exemplified in the case of ‘true’ in Figure 12.14. The case of doubling poses some questions regarding the identification of the lexical unit(s) under scrutiny. It is in fact possible to consider the canonical form and the doubled form as two separate lexemes, or as the same lexeme that has undergone a formal modification which is regular and systematic (Meir 2012, 106). The case here is different from that of

168

Giulia Petitta, Alessio Di Renzo and Isabella Chiari

manual simultaneous evaluation because it involves a different global mechanism in the sign production (two articulators instead of one and symmetric or near-­symmetric movement), while in the former type, the overall formal representation is preserved, with an increase in some gestural features. Nevertheless, in our opinion, there are reasons to prefer the second option (same sign with evaluative morphology): intensification by doubling of one-­handed signs seems a very regular meaning (and possibly cognitive) pattern of gradual intensification, it follows a general formal pattern and, most relevantly, it is a productive means of generating intensification. This means that it can be applied to new bases having the same phonological constraints: being one-­handed, not being centred and with fixed location. Doubling is not exploited in LIS for nominal evaluation functions related to size/shape variation. This probably occurs because nominal doubling is very often used for pluralisation (Pizzuto and Corazza 1996, 190–2).

3 Conclusion As we have seen, SLs share common evaluative strategies that can be related to the lexical, morphological and prosodic levels (manual stress, initial hold). In all reported SLs these strategies seem to bear a significant cross-­linguistic similarity, probably grounded in cognitive mechanisms that are represented through common iconic evaluation strategies. As for morphology, we observed that evaluation takes place in the form of both sequential and simultaneous evaluation, and often involves their co-­occurrence in SL discourse. Even in the case of reduplication we have observed that evaluation can be expressed with both simultaneous and sequential morphological strategies. All kinds of simultaneous evaluation (manual, non-­manual, simultaneous doubling) can, in all reviewed SLs, be used both for the evaluation of size/shape and qualities and in a qualitative and more subjective perspective. As far as our data show, sequential morphology using classifier affixes is more closely related to the expression of descriptive evaluation, more specifically that related to the physical dimension, and is not frequently employed for intensification or subjective judgement. The relationship among evaluation of quantity and intensification in SL is particularly interesting, since we have observed that in simultaneous evaluation all the formal means used for augmentation and diminution in quantity can also be applied to intensification (in both directions, towards the smaller and the larger) (Štekauer, this volume). We have also observed a slight difference in the use of manual simultaneous evaluation in ASL and LIS. ASL seems less prone to its use concerning descriptive properties of physical and abstract entities than does LIS. Nevertheless there are cases of concrete objects (such as ‘box’, realised with the ASL so-­called ‘flat B’ configuration to show the sides and further bent hands to show the front and back of the box; or ‘bowl’, realised with hands curved to show the spherical shape of a bowl and with puffed-­out cheeks) that can be subject to manual evaluation by reducing or enlarging the shapes portrayed. Furthermore, there are some articulatory restrictions that involve both sequential and simultaneous morphological evaluation, mainly due to the fact that (i) an articulator is already strongly involved in the expression of lexical properties of the base sign; (ii) the peculiar movement required in manual evaluation cannot be performed with the location and handshape required; and (iii) the reduction of movement in manual evaluation would require too much effort or would not be perceptually salient enough. More specifically, we have observed that many signs that are typically modified by augmentation with manual simultaneous evaluation are not in fact modified by reduction

Evaluative Morphology in Sign Languages 169 with the same type. For example, LIS ‘project’ (Figure 12.6) can be augmented but not reduced with manual simultaneous evaluation. This asymmetry seems to be grounded in the perceptual and articulatory properties of the canonical forms that constitute the norm for further evaluative processes. In these cases, sequential evaluation or non-­manual evaluation is generally preferred. The questions that remain open are many. A more systematic consideration of affective and subjective evaluation is needed, whether in a language-­specific or a cross-­linguistic perspective, in order to establish their specificity and, crucially, whether their expression can be conceived as actually morphologic. The general framework proposed should be discussed further by taking into account predicative forms of evaluation, including aspectual expression, in order to obtain a wider picture of evaluative strategies in SLs.

Acknowledgements This chapter is the outcome of a close collaboration among the authors. Isabella Chiari is primarily responsible for Sections 1, 2.3 and 3, Giulia Petitta for Sections 2.1.1 and 2.2, and Alessio Di Renzo for Section 2.1.2. We would like to thank Paolo Rossini, Virginia Volterra and Erin Wilkinson for observations and comments that were very useful in the preparation of this chapter. We would also like to thank Barbara Pennacchi for her invaluable work on the original illustrations.

Notes 1. These elements have been differently named (highly iconic structures or HIS, size and shape specifiers, depicting handshapes) and theoretically conceived. For an overview of the different models and of the significant different in the use of the word ‘classifier’ for spoken and sign languages see Cuxac (2000); Emmorey (2003); Zwitserlood (2012). 2. In order to consider evaluative functions, we adopt an interpretative frame based on the ‘ontological’ properties of the referent of the signs and not based on the parts of speech they eventually belong to. This choice depends on the fact that the identification of the part of speech of signs is very disputed and at the present state of theoretical discussion on SL cannot be considered sufficiently set. 3. In this chapter representations of signs are provided by illustrations and SignWriting-­based representation. SignWriting is a notation system designed by Sutton (1995) and currently used by the CNR-­Institute of Cognitive Sciences and Technologies research unit as a tool for transcription and writing. 4. SLs are characterised by different uses of the mouth fulfilling significantly different functions: mouthings are labial movements that partially or fully reproduce words of VL, and mouth gestures are SL-­specific, often iconic morphemes, expressing particular functions related to the sign (for an overview, see Boyes-­Braem and Sutton-­Spence 2001). 5. The space in front of the signer’s chest is called ‘neutral space’ to distinguish it from other locations (for example, face, body).

13  Evaluative Morphology in Pidgins and Creoles Barbara Turchetta

1  Introduction: evaluative devices and world view Evaluative morphology in pidgins and creoles shows different patterns mainly due to internal developments rather than substrate influences. The two kinds of languages, far from being considered as a typological cluster, have no specific genetic affiliation and their morphological characteristics are frequently attributed to typological areas, including other genetically unrelated languages spoken in multilingual contexts of the present or of the past. In general terms, when considering evaluation and perception as two cognitive mechanisms working in any human experience of the world, it is plausible to offer a cultural interpretation of some of the morphological devices related to semantic motivations. In such a perspective, we claim that in general, evaluative devices in languages are the result of the speakers’ evaluation of physical and external phenomena, linked to human perception. The given framework of a direct evaluation of human experience in generating evaluative devices in pidgins and creoles is particularly evident, having a rather recent genesis and being strictly linked to pragmatic forces of speech. In this perspective, we will show how evaluative strategies might fall within semantic and structural categories which are definitely larger than those observable in other natural languages, where language growth and change do not undergo the terrific and fast developments which are visible in contact languages. Languages offer a wide range of morphological devices due to the evaluation of physical characteristics sorted by human beings as elements of distinction. Language categories are oriented to world views and depend on cultural assets sustaining the use of languages. This is very clear in some widely documented cases, such as in the explanation of sub-­ differentiation in several classes of prefixes for nouns in Bantu languages; most of them are attributable to cultural differentiations of physically relevant distinctions, such as human beings versus animals (related to an animacy scale), fluid substances versus solid ones, countable versus uncountable substances and so on. Evaluative morphology in itself is also strictly linked to the human perception of physical elements, social distinctions and cultural evaluations; it is connected to a speaker-­oriented process of evaluation of his or her direct experiences. Subjective commitment is probably the basis for the development of such a category in newborn languages, such as jargons and unstable pidgins, where such a semantic need is perceived and satisfied by pragmatic strategies firstly and grammatical strategies lastly, merging in a later stage of development during pidgin stabilisation. To discuss all these issues we will consider a wide but non-­extensive range of linguistic phenomena, mainly belonging to reduplication and redoubling, which can be described

Evaluative Morphology in Pidgins and Creoles 171 as evaluative strategies. We will also try to discuss some other morphosyntactic strategies which are used to transmit the semantic values of speakers’ commitment in the evaluation of phenomena and entities, as in the case of lexical reduplication and verbal serialisation. Literature devoted to pidgins and creoles distinguishes three categories of evaluation strategies responding to similar distinctions normally identified in other languages. Although these languages show a clear tendency to isolation and a lack of affixation, the distinction is made among the reduplication of lexical items, the redoubling of morphemes and the repetition of words. We will use the same terminology and discuss the feasible boundary between reduplication and redoubling in pidgin and creole languages, where lexical units may also work as grammatical morphemes depending on the phrasal context. The present contribution aims to offer an account of evaluative phenomena in this class of languages. Contact languages are a large category which is not universally accepted as such by scholars working on language interference, language contact, multilingual assets and so on. Some scholars tend to consider pidgins (both unstable and stable ones) within the same category, including second language varieties and mixed languages. In this perspective creoles display a different scale of morphological complexity, showing derivational devices. It does not matter whether pidgins and creoles are or are not included in a common linguistic category; they are widely spoken where language contact has had a deep influence, involving typologically and genetically unrelated languages for long periods, where substrates might or might not have included European languages. It is for this reason that they represent a group of languages where pragmatic strategies and ­language variation represent a vivid propulsive force in language change.

2 The typological quest: is evaluative morphology attested in pidgins and creoles? Evaluative functions in pidgins and creoles are not performed by morphological means specifically devoted to them; these languages do not respond to the widely attested tendency to express evaluation through peculiar morphological strategies. In order to sustain such a strong position, we will briefly look at the scientific debate on these languages in terms of typological characteristics. None of the typological elements traditionally chosen to distinguish pidgins and creoles from other languages would work in an effective and convincing way. Nonetheless, grouping them within a cluster necessarily implies a clear tendency to share some typological characteristics. Various scholars who have more recently been investigating the matter would not readily consider pidgins and creoles as part of a homogenous category: contact languages are far from being declared a language type. Scholars in the past have shown a common attitude to observing creoles and pidgins in a sociolinguistic (identity-­oriented) perspective, as being dependent on the lexifier languages. This attitude was common for decades after Goodman (1964) and Valdman (1978) produced their works on Caribbean creoles. The two scholars focused on the historical evidence of a strict connection between European languages and mother tongues in the pre-­colonial and colonial periods, and their goal consisted in studying substandard varieties of European languages as a language substrate of Caribbean creoles. When turning to more recent attitudes in a similar perspective, pidgins and creoles are often considered as a cluster of languages grouped together thanks to a list of relevant common factors: a non-­phylogenetic origin, a strict connection to substrate languages, a constant and fast development due to the communicative pressure of speakers in multilingual settings, and a vitality which is constantly at risk, thanks to the language ­policies

172

Barbara Turchetta

excluding them. McWhorter (1998; 2005) has made a meaningful contribution to the expansion of this kind of attitude; he has mainly focused on the definition of the category of creole languages as such being typologically distinguishable from other language types. According to his strong position (opinion), the absence of inflectional affixation, tone and rich derivational morphology distinguishes the creole type from other language types. As a matter of fact, most scholars considering the creole grammars as simple grammars had mainly worked on contact languages with a strong substrate influence of European languages. McWhorter himself worked on a database made up of a sample of eight creoles which are spoken in areas of the world where European trade and colonial policy have strongly influenced language contacts in the last centuries. Different scholars have shown that when observing other contact languages, such as creoles in Africa with African substrates, the three elements which were usually absent in the alleged creole type are in fact actually present. Klein (2006) described tone in Kituba, a creole of the Central African Republic with a Kongo substrate, and Labov (1990) and Siegel (1987) have shown how the absence of inflectional morphology does not imply language simplicity. Moreover, descriptive works on non-­European creoles have shown a strong presence of inflectional and derivational morphology, as is the case for Juba Arabic (Versteegh 1984; Avram 2011). Derivational morphology and inflectional morphology have also been used as elements to distinguish different steps within a scale of language contact development, distinguishing earlier phases and later phases where creoles and expanded pidgins are located. Bakker’s (1994) comparison between source languages and pidgins aims at demonstrating the inflectional and derivational richness of the former in comparison with the morphological reduction of the latter. A similar position was assumed by the same author (Bakker 2003) when working on the absence of evaluative morphology in pidgins in comparison with derivational strategies in creoles. Different pidgins and creoles might show similar strategies for expressing evaluative semantic content, especially in iconic pragmatic strategies occurring in speech. Such strategies might not necessarily correspond to functional categories which would be morphologically fixed in their structural systems. Iconicity is often considered as one of the most distinguishable characteristics of pidgins, especially in their first phases of evolution, where pragmatic strategies and context-­induced interpretation work constantly to disambiguate meaning (Givón 1989). Nevertheless, specialists in pidgin and creole studies have demonstrated that iconicity does not play a dominant role in pidgin and creole morphology. Affixation, conversion, reduplication and compounding dominate expanded pidgin and creole grammars, being the most relevant characteristics effectively grouping creoles into a typological class (Bakker, Daval-­Markussen, Parkvall and Plag 2011). Most of the cases we are discussing in order to show how evaluative strategies work in pidgins and creoles will show the dominant role of reduplication. Relevant studies on reduplication as a formal category in pidgins and creoles show its role in fulfilling more than simply evaluative needs (Kouwenberg 2003; Aboh, Smith and Zribi-­Hertz 2012a). Among the formal categories for evaluative morphology which have been defined (cf. Grandi and Körtvélyessy, this volume) the semantic categories most represented in pidgins and creoles are the following: quantitative in size, mass, weight; diminution and ­augmentation in a scale; intensification and attenuation of phenomena. In terms of morphological strategies for conveying evaluative contents, pidgins and creoles mainly show reduplication as the most frequent. As a matter of fact, these ­languages tend to isolation, and isolated languages generally reiterate lexical items (adjectives, adverbs, verbs) for either lexical or grammatical purposes; when duplicating lexical

Evaluative Morphology in Pidgins and Creoles 173 units, they become unpredictable in the semantic effects they create. Because of this, we will discuss the lexical use of reduplication and the redoubling of morphemes as distinct, since the authors we mention clearly underline such a distinction, though the distance between syntactical behaviour and functional meaning is not always distinguishable. The reduplication of lexical items and the redoubling of morphemes within a word context are far from being two distinct phenomena in pidgins and creoles. Consider the following example from West African Pidgin English: (1) plenti> plentiplenti a. ‘richness’ > ‘great richness’ / b. ‘very’ > ‘very much’ go> gogo a. ‘go’ > ‘wandering’ / b. fut > far fut The two examples can be interpreted as either a reduplication of lexical items or a redoubling of morphemes. The former case would offer a semantic interpretation, as given under (a), while the latter would rather offer a functional reading, as under (b), where grammatical functions are given. As highlighted by Forza (2011), reduplication in general is a formal phenomenon and involves grammatical encoding anyhow, no matter whether the origin is iconic. In fact, it has been observed that many languages have reiterations for different lexical categories in some contexts, with non-­iconic readings resulting. It is often assumed that lexical strategies imply iconicity (as in contact languages at a jargon stage) while morphological strategies fall within a grammar. To sustain non-­iconic readings of lexical strategies, Aboh, Smith and Zribi-­Hertz (2012b, 8) discuss the lexical use of reiteration for non-­iconic purposes in Maori, which is neither a pidgin nor a creole language, where there is no clear boundary between the reiteration of lexical items on the one hand, and reduplication on the other. These authors’ suggestion confirms that the distinction is weak and useless in languages with a strong tendency to isolation, whether pidgins, creoles or any other language: (2) ka mare-­mare tonu ia TMA cough-­cough-­cnt still 3sg ‘he coughed many times’ The reiteration of the verb acquires a continuative value in (2), and a similar function is predictable in many expanded pidgins, where repetition of the main verb would indicate a continuing action, an aspectual value normally brought to the same languages by aspect markers: (3) West African Pidgin English: pikin de komot-­komot laik dat boy(s) prog come out-­come out like dem ‘Boys are coming out’ (4) Plantation Pidgin Fijian (Siegel 1987, 105): vakatolu tiko yani ke, vakatolu 3 times stay dir there 3 times tiko tiko tiko tiko tiko vakavanua stay stay stay stay stay village.style ‘Three times I stayed there – I kept staying like a villager’

Barbara Turchetta

174

3  Evaluative morphology and substrate languages in pidgins and creoles When talking about pidgins and creoles of the Atlantic area, substrate transfer from Niger Congo languages is one of the most popular topics through which to establish a strict connection with the structural matrix of African languages. Niger Congo languages are considered to be responsible for the formation of early contact languages in the area during the first centuries of commercial exchanges between Europe, America and Africa. We will not discuss this relatively strong assumption, which characterised most of the many scientific works on the pidgin and creole heritage of the Atlantic area. Our aim is to discuss syntactic and morphological strategies for conveying evaluative meanings, and we will concentrate on that. Nonetheless, special attention must be given to the general attitude shown by some authors who consider iconicity as an internal factor of pidgin development, and non-­iconic constructions as part of an African matrix in many pidgins and creoles in Africa and the Americas; African substrates are normally assumed to be the core elements of a typological area defined as ‘Atlantic’ (Alleyne 1980). It must be noted that studies focused on substrate phenomena for other pidgin and creole languages without a European linguistic heritage hardly discuss iconicity or substrate languages in such terms, and tend to concentrate more on what the new grammar brings rather than on what it has inherited from other languages. If we take into account Niger-­Kordofanian languages which might have played a role in the substrate heritage of pidgin and creole languages of the Atlantic area, there are very few typological features in common. Kouwenberg and LaCharité (2003) discuss reduplication in the framework of African languages’ morphological substrate as being the origin of the development of reduplicative processes in Caribbean creoles. In their words, ‘reduplication is overwhelmingly iconic’ (Kouwenberg and LaCharité 2003, 9) as a process involved in the development of a contact language and linked to universals of language growth. On the other hand, non-­iconic reduplication should be influenced by similar processes in West African substrate languages, where it seems to be a morphological productive device. According to this perspective, iconic strategies of reduplication in pidgins and creoles should derive from inner language changes, while non-­iconic strategies would be part of substrate inheritance. Examples are given (Kouwenberg and LaCharité 2003, 9): (5) Iconic: a. Berbice Dutch Creole b. Papiamentu c. Saramaccan (6)

nangwa ‘long’ nangwa-nangwa ‘very long’ kayente ‘hot’ kayente-­kayente ‘very hot’ langa ‘long’ langa-­langa ‘very long’

Non-­iconic: a. French Creole blanch ‘white’ blanch-­blanch ‘whitish’(dim) b. Jamaican Creole laaf ‘to laugh’ laaf-­laaf ‘inclined to laughter’ c. Ndyuka bai ‘to buy’ bai-­bai ‘bought’

Non-­iconic reduplication holds an evaluative function in only some of the examples in (6) ((6c) does not express any evaluative function), whereas in the examples in (5), which are interpreted by the two authors as iconic, evaluative intent is certainly present. We would therefore conclude that iconic reduplication is semantically transparent and non-­iconic reduplication is complex, or at least visible at a more abstract level from the semantic point of view. According to Kouwenberg and LaCharité, iconic strategies (where a speaker’s

Evaluative Morphology in Pidgins and Creoles 175 evaluative intention is clear) are unmarked and attributable to universals of language, while productive non-­iconic reduplication strategies belong to a more grammatically structured level of language. When effectively comparing non-­iconic reduplication in pidgin and creole languages with West African languages, it is not easy to sustain such an affirmation. West African languages show a widely diffused pattern in reduplicating different categories of words (mainly nouns and verbs) to express regular or customary activities, intensity or other degrees in evaluation. When considering the following comparative data offered by Mutaka and Tamanji (2000, 288–9) on Cameroonian Pidgin English and a presumed substrate language (Bafut), the boundary between iconic and non-­iconic motivations seems to be vague: (7) Cameroonian Pidgin English: à dé fáin fáin 1sg cop well well ‘I am very well’ na wàkàwáká woman cop walk walk woman ‘she is a harlot’ I dón fínis kwátakwátá 3sg pst finish complete complete ‘it is completely finished’ (8) Bafut: á fíímfíî 3sg dark-­dark ‘it is extremely dark’ á sáɁánsàɁà 3sg long-­long ‘it is very long’ Sùù tɨ ɨ ntɨ ɨ S. walk-­walk ‘Suh travels a lot’ The complexity of reduplication and redoubling strategies that we can observe in pidgins and creoles of the Atlantic area is far below that observable in West African languages. The grammar of repetition reaches sophisticated levels in some of the languages normally taken as substrates of contact languages; highly structured African morphosyntactic strategies are not visible in pidgins and creoles of the same area. West Atlantic, Kwa and Kru languages of the Niger-­Kordofanian phylum (where most creolists working on substrates look for evidence) use reiteration of lexical items or morphological redoubling for aspect and focalisation, the latter being a strategy absent in pidgins and creoles. In some other languages of the same phylum, such as Nupe (a Bengue-­Congo language of Central Nigeria), reduplication has to do only with focusing (Kandybowicz 2008). Aboh et al. (2012a, 30) offer a sharp contribution in delimiting the role of substrate languages in the formation of reduplication strategies for evaluative purposes. These authors observe that reduplication in Suriname creoles derives partially from the ­substrate

176

Barbara Turchetta

i­nfluence of Gbe languages. They discuss non-­iconic reduplication in Eastern Gbe languages (Fongbe and Gungbe) and Suriname creoles (Sranan and Saramaccan) in a historical perspective; Suriname creoles and Gbe languages share a common non-­iconic reduplication pattern transforming the meaning of reduplicated items from process/event to state/result. In Suriname creoles this partially derives from substrate influence from Gbe languages (Aboh et al. 2012a, 30): (9) Saramaccan: dí kúja mandú det calabash ripe ‘The calabash ripens’ dí kúja dɛ mandú mandú det calabash cop ripe ripe ‘The calabash is ripe’ A similar semantic difference is found in Gungbe (Aboh et al. 2012a, 30): (10) Gungbe: a. àvún lɔ kú tò kɔ-­mɛ dog det die cop ground-­in ‘The dog died on the ground’ b. àvún lɔ tò kú-­kú tò kɔ-­mɛ dog det det die-­die cop ground-­in ‘The dog was lying dead on the ground’ The copula in Gungbe introduces a progressive aspect, state of affair and results. The state of affair or result in (10b) is given by the verbal reduplication.

4  Derivational evaluative morphology merging in creoles We stated in the previous section that most theories proposing the supposed ‘simplicity’ of pidgin and creole morphology concentrate on what these lack rather than on what is attested in them. More recent works on creole morphology have strongly criticised the notion of semantic transparency in contact languages, talking about ‘the semantic transparency myth’ (Kouwenberg and LaCharité 2011) as a dominant view of derivational morphology of creole languages diffused in creolistics until the end of the last century. Farquharson (2007) widely supports Kouwenberg and LaCharité’s position, taking into account more recent studies on creole morphology; his analysis of a significant samples of contact languages represents a meaningful contribution to discrediting the traditional point of view on their grammatical poverty. Creole morphology is mainly based on affixation, reduplication, compounding and zero-­ derivation. Creoles show a different balance of the four strategies to convey grammatical information, whereas pidgins might reach different levels of morphological stability, depending on their evolutionary scale and level of stabilisation. At a jargon stage, a pidgin might present more pragmatic than grammatical strategies to convey meaning, being nonetheless an effective means of communication. Affixation is a frequent strategy for the TMA systems both of creoles and of stable and expanded pidgins. Prefixation and suffixation in

Evaluative Morphology in Pidgins and Creoles 177 nouns are more frequent in creoles of the Caribbean and Pacific areas, the latter being also the area where two expanded pidgins are spoken, namely Tok Pisin and Bislama, showing a particularly rich morphology in nouns, pronouns and the TMA system. Stable and expanded pidgins, together with creoles, share common morphological features. Contiguous morphology is the most preferred, with a total lack of infixation and separate morphosyntactic constituents. Another typological characteristic of these languages is represented by a common avoidance of combinations of derivational and inflectional morphemes. The former category is definitely extensively attested in creoles, especially with nominal bases. Pidgins and creoles in fruitful sociolinguistic circumstances are vital languages and may change very fast. They develop quite impressively in grammar, especially when they represent a relevant cultural point of reference for speakers (in terms of the speakers’ identity). It has been observed that derivation is one of the first mechanisms to merge with new productive morphemes, often as the product of grammaticalisation processes. A good example is given by the new Jamaican Creole suffix -­sha attached to verbs and nouns, which is probably derived from she and now works for ‘exceeding person’ with an augmentative meaning in nouns and an iterative meaning in verbs (Farquharson 2007): (11) beg ‘to beg’ > begiisha ‘a woman who is always begging’ laaf ‘to laugh’ > laafiisha ‘a person who laughs a lot’ We have discussed the role of iconicity in lexical reduplication as being central in semantic enrichment. Redoubling is also an iconic strategy in pidgins and creoles. When denoting a semantic content, it conveys plurality, uncertain quantity or uncountable quantity. Redoubling as a word-­formation device is more frequent in non-­expanded pidgins; in stable pidgins and creoles, more strategies conveying the same functional meanings occur. When redoubling is used for large quantities it may occur as such in a language, the single form being totally absent in the lexicon. The examples in (12) show the three conditions: (12) West African Pidgin English: bonbon ( sight’

5  Augmentative and diminutive reduplication and redoubling Redoubling is the most frequent strategy registered in pidgins and creoles for evaluative functions. Literature devoted to pidgin and creole morphology (Kouwenberg 2003) distinguishes between repetition (two identical words co-­occurring) and reduplication (lexical morpheme repetition within a single word). No matter which strategy is chosen, it must be taken into account that pidgins and creoles are far from being considered inflectional languages, where evaluative morphology undergoes the same derivative constructions which are visible in European languages, partly responsible in some pidgins and creoles of substrate inheritance. Contact languages normally present isolated morphemes or crystallised forms, where grammatical morphemes within the sequence of a word are not productive. As stated before, some contact languages are particularly rich in grammatical information; the redoubling of morphemes can be effectively considered as the product of a derivational process, though not precisely coinciding with what is traditionally considered as such. As we have already pointed out, it is not easy to put a clear boundary between lexical reduplication and morphological redoubling in such languages with a strong tendency to isolation. For this reason, we will leave terminological distinctions apart; let us consider some examples discussed by different authors to show how reduplication works in pidgins and creoles to convey evaluative meaning. As stated before, pidgins and creoles cannot be considered to be part of a homogeneous category of languages and reflect the influence of different substrates, areal phenomena and interference from the native or second languages of their speakers. Because of this, we can only focus on common characteristics of the languages which share at least one of these variables. Reduplication is generally the most frequent device in many pidgins and creoles with an augmentative meaning: (15) Creole from Martinica: sé pwason-­an piti-­piti dem fish-­det small-­small ‘this fish is very small’ (16) West African Pidgin English: A laik-­am fain-­fain 1s like-­ det well-­well ‘I like it very much’

Evaluative Morphology in Pidgins and Creoles 179 A sharp differentiation was made for creole languages of the Caribbean area by Kouwenberg and La Charité (2005), who studied reduplication strategies in Caribbean creoles, mainly focusing on iconicity as an unmarked interpretation of the phenomenon, including the four common patterns: iteration, augmentation or distribution and multiplication. In their examples from Jamaican Creole (2005, 534–5): (17) Iterative tiif ‘to steal’ tiiftiif ‘to steal repeatedly’ Augmentative fek ‘to be light’ fekfek ‘to be very light’ Multiplicative saka ‘bag’ sakasaka ‘many bags’ Caribbean creoles show a clear distribution of meaning for iconic reduplication of nouns, adjectives and verbs. The three categories convey different semantic meanings, namely augmentative for adjectives (18), iteration for verbs (19) and multiplication for nouns (20). All examples are from Sranan: (18) a no pikinpikin swakiswaki man yu o toki cop neg small-­small weak-­weak man 2sg irr take fu go strey boto comp go compete boat ‘you would not use really small and weak men to compete in a boat race’ (19) ma te mi ma ben de na libi dan a ben but time 1sg mother tns cop loc life then 3sg tns yepiyepi mi helphelp 1sg ‘but when my mother was alive she would help me’ (20) mi mu lay sakasaka udu tya kom na oso 1sg obl load bag-­ bag wood carry come loc house ‘I must carry the wood home with a lot of bags’ Caribbean creoles are certainly the richest in terms of morphological and syntactic complexity; they also show a wide spectrum of sociolinguistic varieties and a wide expansion of the lexicon. Kouwenberg and LaCharité (2005) have shown too that in some Caribbean creole languages the same strategy may convey a diminutive meaning. It is the only clear counter-­ example among all the other cases where reduplication conveys an augmentative meaning, and because of that it is worth mentioning. The iconic principle of reduplication, as indicated by Lakoff and Johnson (1980, 128), works on the statement ‘more of form stands for more of content’, where the semantic redoubling would give more strength to the redoubled word. In the following example from Jamaican Creole, the semantic reinforcement seems clear: (21)

red ‘red’ > rediredi ‘strongly red’

Nevertheless, Kouwenberg and LaCharité (2005) have noticed that reduplication of colour names in Jamaican Creole might also convey a dispersive meaning, where reference is made to a non-­homogeneous dispersion of the dye. In their examples:

Barbara Turchetta

180 (22) [yellow] [black] [red] [green, unripe]

yalayala ‘yellowish’ blakiblaki ‘black-­spotted, sooty’ redredi1 ‘red-­spotted, reddish’ griinigriini ‘green-­spotted, greenish, unripe in places’

The redoubling of morphemes within a word boundary has been seen as a peculiar characteristic of creoles only by Bakker (2003). The strong position is assumed in analysing a sample of eight pidgins and creoles which all have a European language as a substrate inheritance, taking into account that the morphology of creoles shows a higher degree of derivation strategy than does that of pidgins. We have already discussed the strong tendency of pidgins and creoles towards isolation. Isolating morphemes hardly distinguish their functional role between grammatical and lexical morphology, a difference we can disambiguate only in some sentence contexts. For the sake of clarity and to justify redoubling as part of derivational processes, we will discuss some cases related to morphological repetition where reduplication works either in predicates or in nouns, either with an evaluative function or for other semantic purposes. Peter Bakker and Mikael Parkvall (2005) distinguished nineteen functional uses of reduplication in pidgins and creoles, whereby only eight of them could be attributed to an evaluative process; when distinguishing again between strict evaluation in terms of gradualness and the aspect functions of verbs, the group reduced to four (Bakker and Parkvall 2005, 512–13): Verbs:

intensive/augmentative ‘to v a lot’ [evaluation of gradualness] aspectual ‘to v continuously/regularly’ spatial distributive ‘to v in several places’ attenuating ‘to v a bit’ [evaluation of gradualness]

Nouns: distributive ‘various n’ totality ‘every n’ adj:

intensifying ‘very adj’ attenuating ‘somewhat, little adj’

[evaluation of gradualness] [evaluation of gradualness]

Consider reduplicated verbs with an iterative, habitual and distributive function, as shown by the Juba Arabic examples below. They can all be considered a case limit when merging redoubling strategies with evaluative functions only. Aspect in verbs would have an evaluative meaning, being linked to the event evaluation in terms of characteristics rather than being anchored to objective information such as the passing of time would be. Predicates work in a quite distinct way in creoles with high morphological density. Juba Arabic is one of them, being a widely diffused creole in an area where Nilo-­Saharan languages are spoken as a first or second language by speakers who fluently dominate Arabic creole. Juba Arabic (spoken in South Sudan) together with Nubi (spoken in Kenya and Uganda) and Turku (spoken in Chad and northeast Nigeria) emerged in the mid-­nineteenth century through the expansion of military occupation by the Egyptian army with troops of different linguistic origins (Miller 2003) Since then, Juba Arabic has only been exposed to any substrate influence in a limited way (Classical or Standard Arabic). From a morphological point of view, the three creoles almost totally lack productive inflectional and derivational patterns which are very common in several colloquial varieties of Arabic. This

Evaluative Morphology in Pidgins and Creoles 181 might be a reason which prompts and is compensated for by the richness of the redoubling strategy, which is very frequent in the language. A corpus of Juba Arabic discussed by Miller (2003) clearly shows the reduplication pattern in predicates and nouns, though not all occurrences have a morphological status. Augmentative (23) and distributive (24) meanings can be attributed to noun reduplication as follows (Miller 2003, 290–1): (23) grús-­ grús-­ grús ligó wén fi zamán zey dé money-­money-­money find where loc time like dem ‘where is such a big amount of money to be found at this time?’ (24)

úmon bi-­gáta-­gáta híta-­ híta-­híta 3pl TMA-­cut-­cut piece-­piece-­piece ‘they cut [the meat] in very small pieces’

Example (24) shows both verb and noun reduplication, a strategy used to intensify the distributive interpretation. Adjectives, adverbs and quantifiers show similar morphological behaviour. Modifiers  actually seem to be particularly productive in reduplication in Juba Arabic. Intensive/augmentative meaning is often conveyed, a common pattern in several Niger-­ Kordofanian and Nilo-­Saharan languages: (25) Majúb yaú sukún-­sukún-­sukún úwo bi-­kátulnás sedíd M. top hot-­hot-­hot 3s TMA-­kill people strong ‘Majub was very tough, he killed a lot of people’ Continuative connotation with intensification is also a common semantic function for redoubling. In this sense, as in the meaning shown in (25), the repetition of words does have a grammatical function, as in morpheme redoubling. The continuative is shown in the Juba Arabic example (26) (Miller 2003, 291) and in West African Pidgin English (27) (personal data): (26) Juba Arabic: Úman baadín bi-­kélem kalám de biráá-­ biráá-­ biráá 3pl after TMA-­discuss matter dem slowly-­slowly-­slowly ‘then they discuss the matter very quietly [taking their time]’ (27) West African Pidgin English: dis graun na so-­so kata-­kata dem world be like this-­ like this confusion-­ confusion ‘the world will always be a great confusion’ The redoubling of morphemes is also registered in numerals and more frequently in verbs. In the latter, a TMA prefix occurs once, followed by the reduplicated stem (Miller 2003, 293) with an intensive meaning: (28) Juba Arabic: namán beljik

wósulu fi

Réjaf

úman

dósoman-­dósoman sedíd

Barbara Turchetta

182

when Belgian arrive in R. 3pl fight-­fight ‘When the Belgians arrived in Rejaf they fought a lot’

strong

Derivational suffixes conveying evaluative meaning are rarely reported in pidgins and creoles. We have seen the strategy is present in only some Caribbean creoles. Finney (2002, 9) describes reduplication with iterative function (a common pattern in pidgins and creoles) and compound reduplication with evaluative purposes in Krio: (29) Krio: bɛn ‘bent’ bɛn bɛn ‘twisted’ čuk ‘stab’ čuk čuk ‘stab all over’ A different strategy for conveying the evaluation of events is offered by verbal serialisation. The syntax of serial verbs is quite a common pattern in pidgin and creole languages, though it conveys different semantic values from one language to another. Degree-­marking meaning is one of the four major groups of meanings defined by Muysken and Veenstra (1994, 290) for serial constructions in pidgins and creoles. Prototypes of meaning for such a category would be pass (comparative, excessive) or suffice (enough), as in the examples below: (30) São Tomé Creole: n kume pasa 1sg eat exceeding ‘I am a big eater’ (31) Angolar: tuba ka tobe pata rain hab rain exceeding ‘It always rains too much’ (32) West African Pidgin English: yu tok pas fo taim 1sg talk exceeding comp time ‘You talk passing over the time → You are a great talker’ All the examples given and discussed show the different applications of the reduplication, redoubling, repetition and serialisation of verbs in pidgins and creoles. In considering the wide application of such strategies for different semantic purposes, we would not consider evaluative constructions as a distinct natural class. Similar morphological strategies are used both for evaluative functions and for other functional purposes, such as aspect in verbs and plural in nouns. Therefore, evaluative morphology as such is hardly distinguishable as a functional and typological category in languages of this kind, being part of larger formal categories where evaluation and gradualness are but two of the several semantic functions of language units. Multifunctionality is actually and effectively a clear characteristic of languages generated in multilingual contexts. Moreover, it seems that the universal tendency to express evaluative meaning through the use of derivational morphology or any

Evaluative Morphology in Pidgins and Creoles 183 other strategy (as in pidgins and creoles with reduplication) does not necessarily imply the expression of opposing meanings, such as ‘small’ versus ‘big’, ‘bad’ versus ‘good’ and so on. It is this that distinguishes most pidgins and creoles from any other natural language.

Note 1. The difference in form from rediredi in (21) can be explained by a syllabic erosion which is frequent in reduplication.

14  Evaluative Morphology from a Cross-­Linguistic Perspective: Introduction to the Descriptive Chapters Lívia Körtvélyessy

1 Introduction The nature of the information about evaluative morphology (EM) detectable in various sources is generally twofold: either it discusses various theoretical approaches to EM or it focuses on the EM of a particular language. By analogy, this volume is divided into two parts. It starts with theoretical chapters, and its second part describes the position of EM in fifty-­two languages.1 This chapter bridges the two parts and has two basic aims. Firstly, no linguistic research is possible without a thorough theoretical background. The preceding chapters have systematically discussed various theoretical aspects of EM. By implication, they should be followed by practical application of the theory. Therefore, theoretical achievements in the field of EM are briefly mapped. Secondly, the descriptive part of the volume is introduced.

2  (Cross-­linguistic) research into evaluative morphology The parentheses in the heading of this section have a very simple explanation: cross-­ linguistic research into EM remains a relatively untilled area, for various reasons. The field of EM is quite young. If we take Scalise (1984) as its date of birth, EM is only three decades old. Consequently, its theoretical background is still ‘under construction’. However, the biggest hindrance is the lack of descriptive material. While various papers on EM in individual languages have been published during more recent years (for example, on Sardinian – Grandi 2005; Zulu – Madondo 2000; Italian – Gambino 2010; Spanish – Prieto 2005; etc.), some of them comparing two or more languages (for example, Panócová 2009; Kačmárová 2010; Di Garbo 2013; Szymanek and Derkach 2005),2 a comprehensive publication comparing EM in various languages of the world is still lacking. Information on the nature of EM in a language under investigation can be found in descriptive grammars and in papers dealing with theoretical aspects of EM (for example, the derivation vs. inflection cline, the nature of evaluative affixes, Aktionsart, gradation, etc.). EM can be studied from various perspectives: (i) the nature of processes that coin evaluative constructions (see Štekauer, this volume, Chapter 4); (ii) the semantics of evaluative constructions (see Prieto, this volume, Chapter 2); (iii) the syntactic classes that can be expressed by and/or underlie the formation of EM; (iv) morphopragmatics, primarily Dressler and Merlini Barbaresi (1994; 2001) and Schneider (2013) (see Merlini Barbaresi, this volume, Chapter 3); and

188

Lívia Körtvélyessy

(v) EM acquisition (Savickiené and Dressler 2007b). Taking into consideration that, for instance, the category of word classes varies from the cross-­linguistic point of view and that the field of word formation, including evaluative formation, has already been mapped from a cross-­linguistic perspective (see, for example, Štekauer, Valera and Körtvélyessy 2012), the field of semantics seems to be the most intriguing one; the more so because little attention has been paid to it so far – the few exceptions include Wierzbicka (1984), Jurafsky (1993), Fortin (2011) and Körtvélyessy (2012b). For a more comprehensive overview of the state of the art in the field, see the Introduction to this volume. Any of these research areas can also be studied from a cross-­linguistic perspective, an area in which research is badly needed. The paucity of cross-­linguistic studies into EM might be attributed to both the relative immaturity of the field and some methodological problems. Generally speaking, any research is anchored in a particular theory and works with a particular sample examined by a selected research method. As far as EM is concerned, several discrepancies emerge from this background. One of them is terminology. The aforementioned problem of word classes can be exemplified by the Jingulu language (Pensalfini, this volume, Chapter 17.3): ‘Jingulu words fall into three basic categories or parts of speech: verb, nominal and adverb (there is no distinct category of adjectives, and words that translate as either nouns or adjectives in English fall into the part of speech called “nominal”.’ Similarly, in Plains Cree (Wolvengrey, this volume, Chapter 19.7) there is no distinct class of adjectives, and attributive roots commonly occur as prenominal particles. For obvious reasons, the comparison of word classes that can be diminutivised or augmentivised in languages so distinct as are, for example, English, Hungarian, Slovak, Jingulu and Plains Cree seems to be rather complicated. Semantic categories do not offer a remedy, because of either the plethora of various theoretical approaches or, consequently, an absence of agreed and fixed terminology. One of the areas which has traditionally posed problems (not only in EM) is semantics. This aspect of EM can be approached from different points of view as well as by different methods, as illustrated by some earlier works (primarily Jurafsky 1993; 1996) and several chapters of this volume (Prieto, Chapter 2; Mutz, Chapter 11; Körtvélyessy, Chapter 5). The most striking impact on general thought in the field has been made by Jurafsky’s radial model, aimed at identifying a universal motivation for diminutive markers from a diachronic perspective. While highly influential, this radial model is limited in its scope (to diminutives). It is for this reason that there has been a considerable effort to extend this model to augmentatives and to complete the diachronic point of view with a synchronic perspective. This has resulted in various modifications of Jurafsky’s model, as proposed by Prieto and Mutz (this volume, Chapters 2 and 11). Moreover, Körtvélyessy (2014) has proposed the radial model of evaluative semantics based on the onomasiological method. This radial model is in a sense a hierarchical model which takes four major cognitive categories as a point of departure (Substance, Action, Quality and Circumstance). Each of these universal categories subsumes two fundamental evaluative categories, in particular, diminutives and augmentatives. These, in turn, can be realised by a range of universal semantic categories including distributiveness, intensification, attenuation etc. This model thus establishes a universal framework which is implemented to various extents (provided that language has EM). In contrast, Haspelmath’s method (2003) using semantic maps employs universal semantic features for the sake of cross-­ linguistic comparison. Its potential ­application in EM may be considered in two ways. Firstly, analogically to Körtvélyessy’s radial model of evaluative semantics, semantic maps might be used to capture the means of representation of various semantic categories in cross-­linguistic ­perspective; and secondly,

Evaluative Morphology from a Cross-­Linguistic Perspective 189 this method might be applied to individual lexical units, analogically to the semantic map for tree/wood as illustrated in Haspelmath (2003, 25).

3  Descriptive chapters: introduction The editors of the present volume believe that the data provided in the descriptive chapters may serve as an inspiration and an impulse to develop a comprehensive cross-­linguistic database of EM. In preparing this volume, much attention was paid to the selection of languages in order to obtain a representative sample. The starting point was the World Atlas of Language Structures (WALS) database version available in 2012–13, in particular its division of the world into six linguistic macro-­areas (Dryer and Haspelmath 2011). At that time, WALS covered 2,488 languages. For this volume we selected fifty-­two of them, with the aim of keeping the WALS-­based proportion of the individual macro-­areas in our sample as far as possible. Table 14.1 illustrates the desired proportion of languages vs. the actual proportion. Our original idea of a proportional language sample was, however, overpowered by more general criteria. In the first round we excluded languages that are well described in literature from the EM point of view. For this reason, for example, neither Italian nor Spanish is included in our sample. At the same time, as Körtvélyessy (2012b) proved, the Standard Average European (SAE) Sprachbund is very rich in EM in comparison to other language territories, such as Australia and New Guinea. For this reason, the actual share of Eurasian languages is much higher than the desired one. As the next step, potential authors were approached. The most important criterion was the availability vs. non-­availability of a morphologist who could potentially describe the EM of a particular language in English. More than 1,500 emails were sent, and the return rate was very low. Our primary aim was to cover not only languages with a high EM saturation but also those in which EM constructions are at the periphery of the language system. All in all, our language sample consists of fifty-­two languages representing forty-­four genera and twenty-­six language families. Map 14.1 shows their geographical distribution. 3.1 Some general observations about evaluative morphology from the cross-­linguistic point of view The nature of EM differs from language to language. The formation and the semantics of evaluative constructions are so diverse that Scalise’s suggestion of ‘third morphology’, designed for the Italian language, seems not to be universally applicable. However, there Table 14.1. The desired proportion of languages vs. the actual proportion Macro-­area

Number of languages (WALS)

Desired proportion (%)

Our sample

Actual proportion (%)

Total number Eurasia South-­East Asia and Oceania Australia and New Guinea Africa North America South America

2488 368 507 389 590 381 253

100.00 14.79 20.38 15.64 23.71 15.30 10.17

52 16 7 7 8 6 8

100.00 30.77 13.46 13.46 15.38 11.53 15.38

190

Lívia Körtvélyessy

Map 14.1.  The geographical distribution of the sample languages (Dryer and Haspelmath 2011)

are some tendencies that can be observed at the level of either morphology, phonology or semantics. In what follows, I summarise the typologically most interesting features of evaluative constructions on the basis of the descriptive chapters of this volume. 3.1.1 Morphology Suffixation is the most productive word-­formation process in general (see Štekauer et al. 2012) and it is also fully exploited in EM. It can be found, for example, in Slovak, Tibetan, Muna, Yurakaré, Tagalog, Lisu, Sεlεε, Yukulta, Zulu, Plains Cree, Warlpiri, Latvian, Ewe, Modern Greek, Nivkh, Telugu, Wichí, Tatar, Catalan and Udihe. Diminutive suffixes can be doubled as, for example, the suffix isis (-­išiš) in Plains Cree (Wolvengrey, this volume, Chapter 19.7). Suffixation can be accompanied with high tone, especially in tone languages, for example Ewe (Agbetsoamedo and Agbedor, this volume, Chapter 18.3). One suffix can often be attached to more word classes, for instance, in Dalabon (Evans and Ponsonnet, this volume, Chapter 17.1). Other affixal processes can be observed, too: prefixation in, among others, Tibetan, Tagalog, Yami, Latvian and Modern Greek; circumfixation in Yami; and infixation in Moroccan Arabic. Prefixal and suffixal derivation from verbs is found in Muna (van den Berg, this volume, Chapter 16.4). Both suffixation and prefixation can be combined with reduplication. In Muna, for instance, the prefix ka-­ is combined with full reduplication. Full root reduplication with the suffix -­(h)an is used in Tagalog (Rubino, this volume, Chapter 16.5) to express diminution. Reduplication can be used for both diminutivisation and augmentivisation. Moreover, it can be used with ideophones. While compounding is less productive in EM, it is used in, for example, Choctaw, Modern Greek, Nivkh and Telugu. Morphological processes (of very diverse naturea) which are less widespread and/or areally restricted include gender shift (Iatmul), class shift (Sεlεε), clitics (Apma, Dalabon) and echo formations (Telugu). Topics that are frequently discussed in the descriptive chapters of this volume include gender, number and the inflectional vs. derivational status of evaluative markers. These phenomena are discussed in detail in the theoretical chapters of this volume (e.g. Grandi, Chapters 6 and 7). The descriptive chapters provide further evidence by way of numerous examples. For illustration, the diminutive suffix -­ziga in Udihe follows the plural

Evaluative Morphology from a Cross-­Linguistic Perspective 191 affix in adjectives (Tolskaya, this volume, Chapter 15.16). Déchaine, Girard, Mudzingwa and Wiltschko give evidence of the mixed properties of evaluative markers in the Shona language (this volume, Chapter 18.6); in Inuktitut, some evaluative markers can appear between inflectional morphemes (Compton, this volume, Chapter 19.6); and in Bribri, intensification can be expressed by inflectional suffix (González Campos, this volume, Chapter 19.1). In Iatmul (Jendraschek, this volume, Chapter 17.2), feminine gender is assigned to small size and masculine gender expresses either the unmarked choice or large size. Interesting interrelations between plural and diminutive markers can be observed in Luxembourgish (Gilles, this volume, Chapter 15.8) as well as in Udihe (Tolskaya, this volume, Chapter 15.16). A most interesting question of EM is the range of word classes that can be diminutivised and/or augmentivised. As indicated above, in cross-­linguistic research, word classes mean treading on thin ice. While, for example, the analysis shows that adjectives are a frequently diminutivised word class in the languages of the SAE Sprachbund, there are numerous languages which do not comply with the traditional classification of word classes. This phenomenon is characteristic of Australian languages but can be observed in other geographical areas too. So, for example, in the Kɔnni language spoken in Ghana, adjectival functions are taken over by verbs (Cahill, this volume, Chapter 18.4). On the other hand, as Lovick claims (this volume, Chapter 19.3), the Dena’ina language has only a small number of uninflectable morphemes called ‘adjectives’. In a similar vein, in Nivkh, spoken on Sakhalin Island and in the Amur region of Russia (Gruzdeva, this volume, Chapter 15.10), there are no adjectives that would be morphologically distinct from verbs. Similarly, in South American Wichí (Nercesian, this volume, Chapter 20.5), there are no adjectives and adjectival meanings are expressed by verbs. These and similar facts make cross-­linguistic comparison based on word class a highly demanding business. 3.1.2 Phonology The formation of evaluative constructions is frequently accompanied by phonological changes. In Zulu (van der Spuy and Mjiyako, this volume, Chapter 18.8), for example, a sequence of complex changes can be observed. The affix -­jana is added to a noun. If the final vowel is a back vowel, it is changed to a glide /w/. Otherwise it is suppressed by elision. If the result is a combination of /C[bilabial] wj/, the /w/ is omitted. If the final vowel is a bilabial stop or nasal, it is palatalised before /j/. The process of palatalisation is also typical of the Slovak language. In the diminutive form of ‘stick’ (palica < palička), /c/ is palatalised to /č/. Similarly, the augmentative suffix -­isko leads to palatalisation of the stem-­final consonants t, d, n and l (Gregová, this volume, Chapter 15.12). In Telugu (Sailaja, this volume, Chapter 15.15), evaluation can be expressed by compounding. The word for ‘mother’ (amma), for example, is combined with other words as a second element to express higher hierarchical degree. If the first word ends in a vowel, it is deleted. At the same time, the final vowel in the new evaluative construction is lengthened. Phonological coalescence demonstrated by monophthongisation of the second element is attested in Iatmul (Jendraschek, this volume, Chapter 17.2), where kavle can be used with a pejorative meaning. If combined, for example, with gaai (house), it causes monophthongisation to ge. In Ewe (Agbetsoamedo and Agbedor, this volume, Chapter 18.3), the use of tone in ideophones plays an important role. Low tone, for example, indicates bigness or foul smell/ taste. On the contrary, high tone gives a diminutive reading to the ideophone. In the Amur dialect of Nivkh (Gruzdeva, this volume, Chapter 15.10), the suffix -­la-­has diminutive

Lívia Körtvélyessy

192

meaning if the vowel is raised to ə or i. However, further lengthening into ə̄ or ī expresses augmentative meaning. A special topic in the phonology of evaluative constructions is the issue of phonetic iconicity. As Körtvélyessy mentions in the chapter on universals in EM (this volume, Chapter 5), despite Universal 1926, phonetic iconicity in EM seems to be of an areal rather than of a universal nature. It can be identified, for instance, in Huave (Gnerre, this volume, Chapter 19.5) and Plains Cree (Wolvengrey, this volume, Chapter 19.7). 3.1.3 Semantics Many languages give support to Jurafsky’s claim that diminutive affixes originate in words for a child or offspring. In Dalabon (Evans and Ponsonnet, this volume, Chapter 17.1), for example, the evaluative marker =wurd is an enclitic developed from the noun wurd meaning ‘woman’s child’. Similarly, in Lisu (Bradley, this volume, Chapter 16.3) the nominal diminutive suffix and the nominal stem meaning ‘son, child’ are identical. The diminutive suffix -­uña in Jaqaru (Birioukova and Hardman, this volume, Chapter 20.1) developed from the root uña meaning ‘an animal offspring’. Diminution in Ewe (Agbetsoamedo and Agbedor, this volume, Chapter 18.3) is expressed by the suffix -­vi, which is etymologically related to the root ‘child’ or ‘off spring’. Obviously, not all diminutive markers come from the same semantic source. In Yukulta (Round, this volume, Chapter 17.7), the intensifying and ameliorative -­mira is semantically related to the root mira-­ meaning ‘good’, and the suffix -­mut̪ a to a root meaning ‘much, many, plenty’. In Apma (Schneider, this volume, Chapter 16.1), the evaluative marker gam and its allomorphs are used to minimise an event. This marker developed from grammaticalisation of the verb gamra ‘to just do (something)’. When discussing the origin of evaluative markers, the process of borrowing should also be mentioned. In Swedish, a language short of EM, evaluative suffixes are borrowed from French and English (Olofsson, this volume, Chapter 15.13). Jaqaru, a language spoken in the territory of Peru, borrowed the diminutive suffix -­ita/tu from Spanish (Birioukova and Hardman, this volume, Chapter 20.1). Spanish influence can be also observed in the Huave language (Gnerre, this volume, Chapter 19.5). Tolskaya (this volume, Chapter 15.16) discusses the Chinese influence on Udihe EM. The semantics of evaluative constructions is very rich. Besides the core meanings of diminutive and augmentative, evaluative units can also be used to express intensification or pejoration, as for instance in Tibetan, Muna, Yurakaré, Jaqaru, Yukulta, Plains Cree, Walpiri, Yami, Telugu and Tatar; they can attenuate the meaning as in Jaqaru, Choctaw, Apma, Walpiri, Latvian, Nivkh, Luxembourgish and Wichí. There are also a number of other evaluative meanings, such as honorifics, limitation, contempt, approximation, endearment, gradation, hypocoristics and amelioration. An interesting correlation can be observed between the meaning of evaluative constructions and kinship terms. In German, for example, compounding can be used to express augmentivisation: (1)

Groß-­veranstaltung aug-­event ‘big event’

The first element, groß, means ‘big’. The same element is used in Großvatter ‘grandfather’ to express a kinship relation. The same phenomenon can be observed in English. Grand

Evaluative Morphology from a Cross-­Linguistic Perspective 193 piano is a compound referring to an upright piano, while grandmother refers to the mother of one’s mother or father. Similarly in Choctaw (Haag, this volume, Chapter 19.2), oci is used as a diminutive marker, and at the same time it differentiates same-­sex kin (mother vs. aunt; father vs. uncle). The diminutive suffix -­vi in the Ewe language (Agbetsoamedo and Agbedor, this volume, Chapter 18.3) indicates the younger of two people in one family bearing the same name.

4 Conclusion As this chapter suggests, cross-­linguistic research into EM raises many questions. Some of these have been discussed in detail (for example, the inflectional vs. derivational nature of EM), while some have been waiting for careful analysis, for instance the position of Aktionsart or gradation in EM. We hope that the descriptive chapters will initiate a prolific and in-­depth analysis of EM from a cross-­linguistic perspective.

Notes 1. This number encompasses Yukulta (disregarding its relatives Kayardild and Lardil, discussed in Chapter 17.7) and Arabic (without the subtler distinction made in Chapter 18.2). 2. For more references see Chapter 1 of this volume.

15 Eurasia 15.1 Basque

Xabier Artiagoitia

1 Introduction In this chapter I review the most relevant aspects of Basque evaluative morphology. Genetically speaking, Basque is a language isolate and the only pre-­Indo-­European language known to exist in Western Europe. At present, there are about 700,000 native speakers distributed across three administrative territories: the Autonomous Community of the Basque Country (ACBC) and the Community of Navarre in Spain, and the regions of Lapurdi, Low Navarre and Zuberoa, the three of which together make up the western half of the Département des Pyrénées-­Atlantiques in France. Although the Basque-­speaking territory has contracted significantly throughout the centuries, the trend towards the disappearance of the language has been reversed in the last few decades and the number of speakers is now growing, thanks to officialisation and the introduction of the language in the school system in the ACBC. The legal recognition and use of Basque in the school system are, however, geographically limited in Navarre; on the French side, Basque has no official status even though the tendency to use Basque in the school system is gaining support. Typologically, Basque is a non-­rigid, well-­behaved SOV language (with postpositions, genitive–noun and noun–determiner word orders) with ergative morphology in both the nominal and the verbal agreement system. Verb agreement is fairly rich, which makes Basque a null-­subject and null-­object language. With respect to inflection, Basque is close to the agglutinative ideal of allowing clear segmentation of roots and affixes (although there are some deviations). Nominal inflection is all done by means of suffixes; for example etxe-­tik ‘from the house’, etxe-­eta-­tik ‘from the houses’. Verbal inflection makes use of both prefixes and suffixes: e-­kar(r)-­i ‘bring-­ptcp’, n-­a-­kar-­zu ‘you bring me’, d-­a-­ kar-­zu ‘you bring it’, d-­a-­kar-­tza-­zu ‘you bring them’. According to Zuazo’s (2008) widely accepted view, Basque has five dialects: Western (Biscayan in older classifications), Central (or Gipuzkoan), and Navarrese on the Spanish side of the border; and Labourdian-­Navarrese and Souletin on the French side, with a few towns in Navarre speaking the Labourdian-­Navarrese dialect. A sixth dialect, Eastern Navarrese, disappeared during the twentieth century. Despite the prestige local dialects have for their speakers, there is now a growing tendency towards convergence, thanks to the success of Standard Basque or Euskara Batua, brought about and promoted by the Royal Academy of the Basque language in the late 1960s. Unless specified otherwise, the examples given below are all from Standard Basque and I give them using its standard orthographic conventions: s, z and x stand for apico-­alveolar, dorso-­alveolar and palatal sibilants; and ll and ñ for palatal lateral and nasal respectively.

Xabier Artiagoitia

196

With respect to evaluative morphology, Basque is most interesting when it comes to diminutives, which in most cases express both diminution and endearment or affection (precision in some cases), and to emphatic reduplication. I have structured the chapter as follows: Section 2 discusses Basque diminutives; Section 3 deals with augmentatives; Section 4 tackles the issue of morpheme order between evaluative and other morphology; and Section 5 deals with emphatic reduplication.

2 Diminutives Basque has several diminutive suffixes, of which -­txo and -­xe are the most productive and widespread ones nowadays. Other diminutives include -­xka, -­xko and -­ño. I shall discuss the first two separately. I will close this section with the suffix -­tsu, also worth studying in evaluative morphology. 2.1  The diminutive -­txo This suffix has different variants (-­to, -­tto, -­txo, -­txu), of which -­txo is the standard one in present written Basque. The suffix is productive in Basque with nominal, adjectival and adverbial bases and even with quantifiers; when used with proper names and relative nouns the suffix conveys the idea of endearment: (1) n + txo a. Miren Mary ‘Mary’ b. esku, begi ‘hand’, ‘eye’ c. ama, aita ‘mother’, ‘father’

Miren-­txo Mary-­ dim ‘little/dear Mary’ esku-­txo, begi-­txo hand-­ dim eye-­dim ‘little hand’, ‘little eye’ ama-­txo, aita-­txo mother-­ dim father-­dim ‘mum’, ‘dad’

(2) adj + txo a. labur, luze ‘short’, ‘long’ b handi, txiki ‘big’, ‘small’

labur-­txo, luze-­txo short-­ dim long-­ dim ‘a little short’, ‘a little long’ handi-­txo, txiki-­txo big-­ dim small-­ dim ‘slightly big’, ‘slightly small’

(3) adv + txo a. goiz, berandu ‘early’, ‘late’ b. ondo, gaizki ‘well’, ‘badly’

goiz-­txo, berandu-­txo early-­ dim late-­ dim ‘slightly early’, ‘slightly late’ ondo-­txo, gaizki-­txo well-­ dim badly-­ dim ‘fairly well’, ‘fairly badly’

Basque 197 (4) Quantifier + txo a. asko ‘many/much’ b. nahiko ‘enough’

asko-­txo much-­ dim ‘slightly too many/much’ nahiko-­txo enough-­ dim ‘close to enough’

It is interesting to note that Basque quantifiers do not inflect for number, only determiners (articles and demonstratives) do; in this regard, the diminutive -­txo cannot be added to either articles or demonstratives: (5) a. hau *hau-­txo dim this this-­ b. gizon-­a *gizon-­a-­txo man-­ art man-­ art-­dim ‘a/the man’ ‘a/the small man’ Demonstratives are accented and can stand on their own but articles are suffixes on the rightmost element of the noun phrase: neither can take the suffix -­txo. This restriction also applies to verbs and, in principle (but see Section 4), to case endings and postpositions: (6) a. etxe-­tik, etxe-­ra *etxe-­tik-­txo, etxe-­ra-­txo house-­from house-­all house-­from-­dim house-­to-­dim ‘from, to the house’ ‘slightly from, to the house’ b. euskara-­ri, euskara-­k *euskara-­ri-­txo, euskara-­k-­txo Basque-­ dat Basque-­erg Basque-­ dat-­dim Basque-­erg-­dim ‘(to) the Basque language’ ‘slightly (to) the Basque language’ Given that case ending and postpositions, together with articles, are considered inflectional morphemes (Hualde 1991; Hualde and Ortiz de Urbina 2003), it is natural to state the restriction as a ban on attaching the diminutive -­txo to inflectional endings. The truth is, however, that -­txo cannot attach to other categories, whether they have inflectional endings or not (e.g. verbs, demonstratives). Besides, it is well known that -­txo can attach to inflected finite verb forms in headless relatives clauses: (7) daki-­da-­n-­a daki-­da-­n-­txo-­a know-­ 1sg-­comp-­art know-­ 1sg-­comp-­dim-­art ‘what I know’ ‘the little which I know’ Interestingly, the same complementiser -­(e)n is used in indirect questions; these, however, do not accept -­txo: (8) Ez daki-­t hau dauka-­zu-­n *ez dakit hau dauka-­zu-­n-­txo neg know-­1sg this have-­2sg-­comp neg know-­1sg this have-­2sg-­comp-­dim ‘I don’t know whether you have it’ ‘I don’t know slightly whether you have it’

Xabier Artiagoitia

198

This suggests that the ban on -­txo attaching to a finite verb form in (8) is motivated by the lack of a nominal base, present through an understood empty noun in (7). To summarise: (9) _____]X -­txo]X where X = n, adj, adv, quantifier In essence, -­txo is a diminutive suffix on nouns, adjectives, adverbs and quantifiers. 2.2  The suffix -­xe as emphatic and diminutive The suffix -­xe (-­txe after nasals and rotics) is generally described as an emphatic suffix for demonstrative deictics (examples in 10), for manner and time adverbs based on demonstratives (11), and for other deictic adverbs (12): (10) hau, haue-­k hau-­xe, haue-­xe-­k dim these-­dim-­pl this these-­pl this-­ ‘this’, ‘these’ ‘precisely this’, ‘precisely these’ (11) a. honela thus ‘in this way’ b. hemen, hor, han here ‘here’, ‘there’, ‘over there’

honela-­xe thus-­ dim ‘exactly in this way’ hemen-­txe, hor-­txe, han-­txe there there here-­dim there-­dim over-­there-­dim ‘right here’, ‘right there’, ‘right over there’

(12) orduan, orain orduan-­txe, orain-­txe then-­ dim now-­ dim ‘then’,‘now’ ‘right then’, ‘right now’ The label ‘emphatic’ is justified because the words above can easily function as focus, typically recognisable in Basque by being in the preverbal position (Elordieta 2001): (13) ni-­k (liburu) hau-­xe nahi dut I-­ erg book this-­dim want aux ‘I want precisely this (book)’ But, as pointed out by the Basque Academy (Euskaltzaindia 1985) and de Rijk (2008), these emphatic words need not be in focus position; they may be understood as reinforced topic phrases. Interestingly, this -­xe suffix differs quite a lot morphologically from -­txo in that, at first sight, it can follow inflectional endings: we have just seen that in the emphatic plural forms of the demonstratives, where it is sandwiched between the (plural) stem and the pluraliser (e.g. haue-­xe-­k in 10). It seems to be the case that -­xe can attach to any of the three demonstrative stems and be followed by the corresponding case ending or postposition (14), but it can also be attached to the entire demonstrative stem + case ending or postposition structure (15): (14)

Demonstrative stem + -­xe-­+ postposition: hone-­xe-­tan, horre-­xe-­tan, hare-­txe-­tan

Basque 199 this-­ dim-­loc that-­ dim-­loc that-­ dim-­loc ‘right in this’, ‘right in that’, ‘right in that over there’ Demonstrative stem + postposition + -­xe: hone-­tan-­xe, horre-­tan-­xe, hare-­tan-­txe these-­ loc-­dim those-­loc-­dim those-­loc-­dim ‘right in these’, ‘right in those’, ‘right in those over there’ (15)

According to de Rijk’s (2008) detailed description, the only case ending or postpositions that have a fixed (-­xe-­ + case ending/postposition) order are ergative, genitive and instrumental. On comparing -­txo and -­xe, it turns out that the inflectional endings that -­xe can systematically attach to are precisely demonstrative, case and postposition, categories that the suffix -­txo cannot attach to. The suffix -­xe is also attached to the verb’s participle form, to adjectives and adverbs and to quantifiers provided all these are followed by comparative morphemes (generally -­ago); in the case of the particle, -­xe is restricted to predicative uses (never in combination with the auxiliary), so it may be argued that the verb has undergone v→adj conversion: (16) Participle + -­xe: gal-­du, neka-­tu gal-­du-­xe,  neka-­tu-­xe loose-­ pctp tire-­pctp loose-­ pctp-­dim tire-­pctp-­dim ‘lost’, ‘tired’ ‘slightly lost’, ‘somewhat tired’

Adjectives + -­xe-­ + -­ago: txiki-­ago, handi-­ago txiki-­xe-­ago, handi-­xe-­ago small-­ compar bigg-­compar small-­ dim-­compar big-­dim-­compar ‘smaller’, ‘bigger’ ‘slightly smaller’, ‘slightly bigger’ (17)

Adverbs + -­xe-­ + -­ago: gero-­ago, urruti-­ago gero-­xe-­ago, urruti-­xe-­ago late-­ compar far-­ compar late-­ dim-­compar far-­ dim-­compar ‘later’, ‘farther’ ‘a little later’, ‘a little farther’ (18)

(19) Quantifiers + -­xe-­ + -­ago: gutxi-­ago, gehi-­ago gutxi-­xe-­ago, gehi-­xe-­ago little-­ compar much-­ compar little-­ dim-­compar much-­dim-­compar ‘less’, ‘more’ ‘a bit less’, ‘a bit more’ In all these cases, the meaning conveyed by the suffix is of small degree, so diminution is more salient than in the case of its emphatic value. It is worth pointing out that -­xe is far less common as a diminutive on plain adjectives than -­txo, but instances of that are reported by Camino (2008, 156), with the meaning ‘too’, for some northern dialects: (20) handi-­xe gazi-­xe big-­ dim salty-­dim ‘too big’ ‘too salty’

Xabier Artiagoitia

200 It thus seems that the rule for -­xe suffixation is as follows:

(21) _____]X -­xe]X where X = adj, adv, quantifier, dem/deictic, P, case In short, -­xe cannot attach to nouns; the rule above should also include the proviso that, when attached to the categories adj, adv and quantifier, the following suffix is generally a comparative suffix. 2.3  Other diminutives The eastern suffix -­xka is at present used in Standard Basque attached to both nominal bases and some adjectives, mostly colour adjectives: (22) a. liburu-­xka, herri-­xka, mordo-­xka dim, village-­dim, amount-­dim book-­ ‘small book’, ‘small village’, ‘small amount’ b. gorri gorrixka red red-­ dim ‘reddish’ The nuance introduced by -­xka with colour adjectives is that of approximation, not of diminution (as opposed to the corresponding -­txo form). This suffix is any case less productive than -­txo. The suffix -­sko/-­xko resembles the previous one; the variant with a palatal sibilant is taken to be a double diminutive that has arisen by affective palatalisation. The suffix -­sko as attached to nominals is present in all varieties of Basque in common nouns (e.g. idi ‘ox’ and idi-­sko ‘little ox’), but it is no longer productive in most of them. As an adjectival (and adverbial) suffix, it can be used to create diminutive forms; de Rijk (2008, 156–7) gives two dozen examples but in fact many are hardly used in Standard Basque. Here are some that are well attested in modern written Basque: (23) harro-­xko, handi-­xko, urruti-­xko proud-­ dim big-­ dim far-­ dim ‘rather proud’, ‘rather big’, ‘rather far’ Despite de Rijk’s (2008, 156) claim, most adjectives may take the suffix -­txo in Standard Basque. The suffix -­ño is found, mostly but not exclusively, in nouns in Lapourdian-­Navarrese Basque: (24) a. sagu ‘mouse’ b. haur ‘child’ c. andere ‘woman’

sagu-­ño ‘small mouse’ haur-­ño ‘little child’ andere-­ño ‘small woman, female teacher’

Basque 201 The word andereño is common in Standard Basque, as it is the standard way to refer to female teachers. 2.4  The suffix -­tsu, or how to render an approximate value The suffix -­tsu has the peculiarity of roughly meaning ‘almost, approximately’ when applied to adjectives and some adverbs: (25) a. berdin, ber ‘similar’, ‘identical’ b. orain, honela ‘now’, ‘thus’

berdin-­tsu, ber-­tsu es identical-­ es similar-­ ‘more or less similar’, ‘almost identical’ orain-­tsu, honela-­tsu now-­ es thus-­es ‘more or less now’, ‘almost this way’

It can also attach to wh-­pronouns, as in non-­tsu ‘where more or less’, and to quantifiers, as in berrogei-­tsu ‘forty more or less’.

3 Augmentatives Basque has two augmentative suffixes, which are not used very often: -­kote and -­tzar. The first is generally attached to adjectives and has a pejorative value: (26) lodi, on lodi-­kote, on-­kote fat-­ aug good-­aug ‘fat’, ‘good’ ‘big fat’, ‘too good (of a person)’ This morpheme also exists in Basque as a derivational suffix which creates group nouns from quantifiers (e.g. bi-­kote ‘couple’ < bi ‘two’; bos-­kote ‘quintet’ < bost ‘five’). The suffix -­tzar, on the other hand, is usually denominal and need not be pejorative (de Rijk 2008, 85): (27) etxe, euli, zubi etxe-­tzar, euli-­tzar, zubi-­tzar house-­ aug, fly-­aug, bridge-­aug ‘house’, ‘fly’, ‘bridge’ ‘big house’, ‘huge fly’, ‘big bridge’

4  Issues in evaluative morphemes Evaluative morphemes are external to derivational morphemes, as can be seen in the following examples with the derivational suffixes -­tasun and -­keria: (28) a. ama-­txo ama-­tasun *ama-­txo-­tasun (?) ama-­tasun-­txo mother-­ dim mother-­ sfx mother-­ dim-­sfx mother-­ sfx -­dim ‘mum’ motherhood’ ‘mumhood’ ‘little motherhood’ b. txiki-­txo txiki-­keria *txiki-­txo-­keria txiki-­keria-­txo small-­ dim small-­sfx small-­ dim-­sfx small-­sfx-­dim ‘slightly small’ ‘littleness’ ‘slight smallness’ ‘small littleness’

202

Xabier Artiagoitia

Evaluative morphemes generally exclude each other, so that we hardly find two diminutives or augmentatives together: (29) a. *etxe-­txo-­tzar, etxe-­tzar-­txo house-­ dim-­aug house-­aug-­dim ‘big small house’, ‘little big house’ b. *orain-­tsu-­txe, *orain-­txe-­tsu now-­ dim-­dim now-­ dim-­dim ‘precisely more or less now’, ‘more or less right now’ The only exception that I am aware of is the no longer productive diminutive -­ila, -­ilo, to which further diminutives or augmentatives can be added: (30) neskatila-­txo (< neska-­tila) girl-­dim young girl-­dim ‘little young girl’ ‘young girl’ The reason for this may well be that fact that words with -­ila and -­ilo are perceived as full lexical items, so that the innermost suffix is not considered a true evaluative morpheme by speakers. With respect to the order of evaluative morphemes and inflectional endings, the usual tenet that evaluative morphemes precede inflectional morphemes also holds of Basque, with the exception of -­xe (in its precision value), which can also precede postposition and case ending (e.g. 13–14). The rest of the morphemes must precede the article, singular or plural, and case endings and postpositions: (31) Diminutive/article: a. etxe-­txo-­a but *etxe-­a-­txo house-­ dim-­art house-­ art-­dim ‘the small house’ ‘little/barely the house’ Diminutive/case postpositions: b. etxe-­txo-­ra, etxe-­txo-­tik, etxe-­txo-­an house-­ dim-­all house-­ dim-­abl house-­dim-­loc ‘to the small house’, ‘from the small house’, ‘at the small house’ *etxe-­ra-­txo, *etxe-­tik-­txo, *etxe-­an-­txo house-­ all-­dim house-­ abl-­dim house-­ loc-­dim ‘slightly to the house’, ‘slightly from the house’, ‘slightly at the house’ The issue is slightly more complicated with comparative morphemes: these are most commonly attached to evaluative morphemes: (32) Diminutive/comparative: zuri-­txo-­ago, zuri-­txo-­egi, zuri-­txo-­en white-­ dim-­compar white-­ dim-­too white-­dim-­sup ‘more slightly white’, ‘too slightly white’, ‘the most slightly white’ but the reverse can also be found, although it is much less frequent:

Basque 203 (33) Comparative/diminutive: zuri-­ago-­txo, zuri-­egi-­txo, zuri-­en-­txo white-­ compar-­dim white-­too-­dim white-­ sup-­dim ‘slightly whiter’, ‘slightly too white’, ‘slightly the whitest’ The only apparent exception to this generalisation is given by the directional allative postposition, which can accept the diminutive -­txo or -­xe followed by a comparative suffix: (34) aurre-­ra, aurre-­ra-­txo-­ago, aurre-­ra-­xe-­ago forward-­ all forward-­ all-­dim-­compar forward-­ all-­dim-­compar ‘forward’, ‘slightly further ahead’, ‘slightly further ahead’ This exception does not affect just any phrase with the allative postposition, but only those with locational nouns that have become adverbialised, fixed forms.

5  Emphatic reduplication Basque has a characteristic intensificational reduplication, very common with adjectives and adverbs, and much less common with nouns and noun phrases with their corresponding case endings or postpositions. In the most common case, the reduplication consists of repeating an adjective and inflecting the second instance: (35)



handi-­handia, gorri-­gorria big-­big-­art red-­red-­art ‘very big’, ‘very, absolutely red’

This reduplication has the effect of distinguishing inflectional endings from derivational morphemes. For example, the reduplicated form of astiro ‘slowly’ or ondo ‘well’ consists of a mere repetition: (36)



astiro-­astiro, ondo-­ondo slowly-­slowly well-­well ‘very slowly’, ‘very well’

This is so because the suffixes -­ro and -­to are derivational adverb morphemes which need not take any inflectional ending. However, the partitive postposition -­(r)ik, connected historically with the ablative (de Rijk 2008), which gives rise to stative predicative adverbs when attached to nouns or adjectives, leaves the postposition outside the scope of the reduplication: (37) a. isil > isil-­ik cf. isil-­isilik quiet quiet-­ part quiet-­quiet-­part ‘quiet’ ‘quietly’ ‘very quietly’ b. gazte > gazte-­rik cf. gazte-­gazterik young young-­ part young-­young-­part ‘young’ ‘(at) young age’ ‘at a very young age’

204

Xabier Artiagoitia

Many noun phrases with postpositions, specially when they have a temporal or locational value, can be reduplicated to indicate exact time or location: atze-­atzean ‘right at the back’ (cf. atze-­an ‘at the back’), erdi-­erdian ‘right in the middle’ (cf. erdi-­an ‘in the middle’). Emphatic manner can also be conveyed by reduplicating a noun with a case ending: bihotz-­ bihotzez ‘totally with the heart, best wishes’ (cf. bihotz-­ez ‘with the heart’). Finally, I close my discussion of Basque by mentioning affective palatalisation, a sort of evaluative phonological rule (Hualde and Ortiz de Urbina 2003, 39). Affective palatalisation converts neutral words into affective words, comparable with similar words with diminutives: (38) a. sagu ‘mouse’ xagu ‘little mouse’ b. luze ‘long’ luxe ‘a little long, longish’ The usual pattern is for sibilants, stops and laterals to turn into palatal sibilants, stops and laterals and, depending on the dialect, also for /r/ to turn into [l]; word-­initial palatal sibilants may also be pronounced as affricate palatals.

15.2 Catalan

Elisenda Bernal

1 Introduction Catalan is a Western Romance language spoken by about 10 million people (Pradilla 2011), mostly in the three Spanish regions which have Catalan as their official language (Catalonia, the Balearic Islands and the Valencian Community, where it is called Valencian). Catalan is also spoken in a small part of Aragon and Murcia (Spain), Andorra, Roussillon (France) and the city of Alghero (Sardinia, Italy). Evaluative morphology in the Romance languages is seen almost exclusively in terms of evaluative suffixes, which, as is well known, do not themselves denote new words with a different lexical meaning. Instead, they modulate differences in the nuance and connotation of the base to which they are attached while affecting neither the core meaning nor the grammatical category of the root word. In this chapter, we will present a general overview of the characteristics of evaluative morphology in Catalan (Section 2), and then a more detailed analysis of the various suffixes, including augmentative (Section 3.1), diminutive (Section 3.2) and evaluative morphology in verbs (Section 3.3). Before the conclusion (Section 4), we will consider a group of affixes (both prefixes and suffixes) that are normally grouped under the heading of derivative morphology, yet which also convey evaluative properties of the bases to which they are attached (Section 3.4).

2  Primary characteristics The main characteristics of evaluative suffixes will be presented in the following sections (2.1–2.9). These features are based in part on the work of Gràcia and Turon (1998) and Pérez Saldanya, Sifre Gómez and Todolí Cervera (2004). 2.1  Grammatical category of the base Evaluative suffixes can be attached to any grammatical category with the exception of prepositions:1 (1) gat-­et cat-­ dim-­m.sg ‘little cat’

206

Elisenda Bernal



dolc-­et sweet-­ dim-­adj.sg ‘sweet’

tard-­et late-­ dim-­adv ‘late’ 2.2  Compatibility restrictions Evaluative suffixes follow certain compatibility restrictions with regard to the base to which they are attached: for example, with nouns, evaluative suffixes are normally attached to simple, non-­derived nouns (2a), although there are some exceptions (2b), but not to toponyms (2c) or collective nouns (2d): (2) a. *pensa-­ment-­et dim thinking-­ galled-­et-­a bucket-­ dim *cub-­ism-­et cubism-­ dim b. cendr-­er-­et ashtray-­ dim ‘little ashtray’ penja-­dor-­às hanger-­ aug ‘large hanger’

c. *Ferrar-­et-­a Ferrara-­ dim *Berlin-­às Berlin-­ aug d. *bestiar-­et livestock-­ dim Furthermore, it is often difficult to combine these suffixes with compounds, though it is possible to form nouns such as obre+llaun-­et-­es ‘can opener-­dim’ or para+sol-­et ‘sunshade-­ dim’. It is, however, unlikely that a speaker will say ?somia+truit-­et-­es ‘daydreamer-­dim’ or ?roda+món-­ot ‘tramp-­aug’. This difficulty is further enhanced in the case of phrasal compounds: while it is possible to find brac-­et de gitano ‘type of Swiss roll-­dim; lit. arm-­ dim of gypsy’ or ull-­ot de poll ‘callus; lit. eye-­aug of baby bird’, it seems quite unlikely that a speaker would use words such as those in (3): ?? (3) mal-­et de cap pain-­ dim-­m.sg of head ‘a slight headache’

Catalan 207 campan-­ass-­a de Gauss bell-­ aug-­f.sg of Gauss’ ‘large Gaussian curve’2 ??

As for adjectives, there are no restrictions when it comes attaching these suffixes to simple qualifying adjectives (4a), though restrictions upon relational adjectives do exist (4b): (4) a. lletj-­ot ugly-­ aug.adj ‘very ugly’ llargu-­et long-­ dim.adj ‘a bit long’ b. *policial-­et police-­ dim.adj *solar-­às solar-­ aug.adj As for adverbs, they are only combined with the diminutive suffix -­et, and even this cannot be indiscriminately attached to any adverb: for example, it is common, at least in certain contexts, to find forms such as those in (5a), yet one cannot form diminutives such as those in (5b): (5) a. d’hor-­eta soon-­ dim ‘a bit early’ lluny-­et far-­ dim ‘a bit far’ aviad-­et soon-­ dim ‘a bit soon’ de press-­et-­a fast-­ dim ‘a bit fast’ a poc a poqu-­et slowly-­ dim ‘a bit slowly’ a prop-­et near-­ dim ‘a bit close’ b. *potser-­et maybe-­ dim

Elisenda Bernal

208 *pertot-­et everywhere-­ dim *menys-­et less-­ dim 2.3  Grammatical category of base

Evaluative suffixes do not change the grammatical category of the base they modify except in certain rare cases when they are combined with verbal bases to form new nouns: (6) xiul-­et to whistle-­dim ‘whistle’

cag-­arro to shit-­aug ‘turd’



esmol-­et to sharpen-­dim ‘knife-­grinder’

2.4 Gender In the case of nouns, gender also generally remains constant (7a), though changes in gender are common in the case of nouns with lexicalised evaluative suffixes where the derivative relationship between the base and the evaluative suffix is not so obvious from a synchronic point of view (7b): (7) a. tovallolfem-­et-­afem towel-­ dim ‘small towel’ anellmasc-­àsmasc ring-­ aug ‘big ring’ b. illafem ill-­ot-­afem ill-­otmasc ‘island’ island-­ aug-­f.sg island-­ aug-­m.sg ‘big island’ ‘islet’ 2.5 Iconicity Evaluative suffixes present a certain degree of iconicity between form and meaning: diminutives are generally formed with palatal vowels i (-­í, -­ill) and e (-­et, -­ell), while augmentatives use the velar vowels o (-­ot) and u (-­uss).

Catalan 209 2.6  Inflection From a morphophonological perspective, evaluative suffixes maintain the inflection of the word from which they are derived and have a normalising tendency when the initial forms are atypical (i.e. masculine forms which deviate from the zero morpheme or ­ feminine forms other than -­a): consequently, we see carro ‘cart’, with its gender marked by -­o, reacquire the zero morpheme – usual in Catalan – when the ­diminutive is formed, resulting in carr-­et ‘cart-­dim’ and not *carreto; or bigoti ­‘moustache’, a masculine noun ending in an -­i which disappears when the diminutive (and augmentative) is formed: bigot-­et ‘moustache-­dim’, but neither *bigoti-­et nor *bigot-­et-­i (yet *globus-­às ‘balloon-­ aug’). Also, we see mà ‘hand’ and calor ‘heat’, feminine nouns that lack an inflection which nevertheless appears when evaluative nouns are formed: man-­eta ‘hand-­dim’, calor-­assa ‘heat-­aug’ (yet *sintesi-­et-­a ‘synthesis-­dim’).3 2.7  Derivative cycle Evaluative suffixes close the derivative cycle: they appear after the derivative suffixes and before the marks of inflection:4 (8)

menja-­dor-­et-­s *menja-­dor-­s-­et dining room-­dim-­pl dining room-­pl-­dim ‘small dining rooms’

2.8 Recursivity Evaluative suffixes are the only suffixes that can be recursively applied to the same base provided that two conditions are met: firstly, the suffixes must not be semantically opposed and, secondly, they must be phonologically different:5 (9) petit-­on-­et small-­ dim-­dim ‘very very small’ *man-­ass-­eta hand-­ aug-­dim *mocador-­et-­arro handkerchief-­ dim-­aug 2.9 Lexicalisation Evaluative suffixes have an extremely high capacity for lexicalisation. When this occurs, they lose the value, whether diminutive or augmentative, of the suffix: (10) cavall-­et horse-­ dim ‘easel’

210

Elisenda Bernal



bomb-­et-­a bomb-­ dim ‘light bulb’

sabat-­ot shoe-­ aug ‘dimwit’

serpent-­in-­a snake-­ dim ‘streamer’ 2.10  Semantic classification The semantic classification of evaluative suffixes tends to be based on whether they refer to size (being either augmentative or diminutive in nature) or whether they convey subjective value judgements on the part of the speaker (either expressing affection or being pejorative in nature). However, the line between size and value judgement can often be blurry, as augmentatives can sometimes have a pejorative connotation and diminutives can sometimes express affection (though diminutives can also sometimes be pejorative in nature): real-­ world use of these suffixes shows that their meaning must be considered on a case-­by-­case basis and that, with variation in intonation, the suffixes can have the opposite meaning of what would normally be expected. Here are several examples (11): (11) a. vestid-­et dress-­ dim ‘little dress’ cadir-­eta chair-­ dim ‘small chair’ b. brac-­et arm-­ dim ‘thin arm’ cord-­eta string-­ dim ‘thin string’ c. advocad-­et lawyer-­ dim ‘bad lawyer’ modist-­eta dressmaker-­ dim ‘bad dressmaker’ The examples in (11) all share the same suffix, -­et, but the specific interpretation of the suffix is not the same in each case: in (11a) we can observe the diminutive in the strict sense of the word, as it modifies the size of the bases, while in (11b) we see that the size remains unchanged and the suffix expresses evaluative degree, which in these cases affects

Catalan 211 the thickness. Finally, in (11c), we observe how adding a diminutive suffix to the name of a profession serves a pejorative function. Similarly, a veuarra can just as well indicate a powerful voice in the sense of ‘voice-­aug’ as an unpleasant voice in the sense of ‘voice-­ pej’. Consequently, it becomes clear that while it is relatively simple to establish the root meanings of evaluative suffixes, their practical connotations in real-­world use, which are very context dependent, prove resistant to systematic classification.

3  Evaluative morphology When analysing evaluative morphology in Catalan, the first challenge we come across is establishing a definitive list of affixes in the language: Fabra (1956) lists twenty-­eight infixes, nine diminutive suffixes and three augmentative suffixes; Badia i Margarit (1962) lists eight diminutive suffixes, three augmentative suffixes and one pejorative; and Cabré (1994) compiles two augmentative suffixes, nine diminutives and two pejorative suffixes. Meanwhile, the Institut d’Estudis Catalans (2002) counts three augmentatives, sixteen diminutives and one pejorative, while Cabré and Rigau (1986) focus on nine evaluative infixes. Finally, Rull (2004) notes four augmentatives (to which two superlatives are added) and two diminutives. As indicated by Gràcia and Turon (1998), following the work of Napoli and Reynolds (1995) on evaluative suffixes in Italian, it is likely that at an earlier stage of the language all evaluative suffixes were approximately equal in productivity. However, the available suffixes most often used by contemporary Catalan speakers are far fewer in number and are largely limited to three augmentative suffixes (-­arro -­arra, -­às -­assa and -­ot -­ota) and two diminutives (-­et -­eta and -­ó -­ona). We will address them in the following sections. 3.1  Augmentative suffixes As stated above, the most common augmentative suffixes in contemporary Catalan are -­arro -­arra, -­às -­assa and -­ot -­ota. The first one is combined with adjectives, and the other two with both nouns and adjectives. As stated by Rull (2004), all three suffixes can indicate affection or endearment in addition to simple augmentation. The suffixes -­às -­assa and, especially, -­ot -­ota (which, historically, had a diminutive value, as seen with ceballot ‘spring onion’) can additionally express a pejorative connotation: -­ot -­ota often indicates a measure of vulgarity (12a) though, among children, it can often adopt a more endearing value (12b):6 (12) a. poll-­ot cock-­ pej ‘huge cock’ b. panx-­ot-­a belly-­ aff ‘fat belly’ Table 15.2.1. Augmentative suffixes in Catalan Suffix

N

Examples

Adj

Examples

-­arro -­arra -­às -­assa -­ot -­ota

× × ×

dit-­arro ‘finger-­aug’ paper-­às ‘paper-­aug’ llibr-­ot ‘book-­aug’

— × ×

bon-­às ‘good-­aug’ brut-­ot ‘dirty-­aug’

Elisenda Bernal

212

Superlatives, on the other hand, prove to be different from augmentative suffixes: in Catalan (and in other Romance languages) adjectives can be regularly converted to superlatives with the addition of the suffix -­íssim -­íssima or its Latin variant, -­èrrim -­èrrima. While the two forms do not block each other out, the Latin variant is nevertheless less commonly used due to restrictions on the base with which it can be combined, namely that it must be the Latin allomorph of the base: pobr-­íssim vs. paup-­èrrim ‘poor-­sup’ (*pobr-­ èrrim). Both forms, however, can represent deviations from the norm, combining with nouns to attain an ironic or humorous effect (Rull 2004): (13)



cunyad-­íssim brother-­in-­law-­sup ‘super brother-­in-­law’



sop-­íssim-­a soup-­ sup ‘super soup’ Rull (2004, 296) also includes in the list of augmentative suffixes the suffix -­al, as seen in bass-­al ‘puddle-­aug’ or finestr-­al ‘picture window’. 3.2  Diminutive suffixes The most common and productive diminutive suffixes in contemporary Catalan are -­et -­eta and -­ó -­ona,7 both of which can be added to nouns and adjectives. The masculine suffix -­et, however, can also be attached to adverbs. Nevertheless, and as previously stated, the list of elements is quite long despite the fact that there are more restrictions governing how they can be combined and that speakers often find it difficult to modify the words in their vocabulary. Table 15.2.2. Diminutive suffixes in Catalan Suffix

N

Examples

Adj

Examples

Adv

Examples

-­et -­eta -­ó -­ona -­ell -­ella -­í -­ina -­ill -­illa -­im -­ol -­ola -­oi -­oia -­ic -­ica

× × × × × × × × ×

gat-­et ‘cat-­dim’ car-­ona ‘face-­dim’ prad-­ell ‘meadow-­dim’ corbat-­í ‘tie-­dim’ rosqu-­illa ‘nut-­dim’ plug-­im ‘rain-­dim’ estany-­ol ‘pond-­dim’ nin-­oi ‘boy-­dim’ can-­ic ‘dog-­dim’

× × — — — — — × ×

maqu-­eta ‘pretty-­dim’ petit-­ó ‘small-­dim’

× — — — — — — — —

tard-­et ‘late-­dim’

bonic-­oi ‘nice-­dim’ pobr-­ic ‘poor-­dim’

In addition to the suffixes that appear in Table 15.2.2, there are several more that are far less widespread, including: -­el·lo -­el·la (nass-­el·lo ‘nose-­dim’), -­eu -­eua (animal-­eu ‘animal-­dim’) and -­iu -­iua (cam-­iu-­a ‘leg-­dim’). Rull (2004) also includes several examples of diminutive suffixes of a more formal register, including -­ul -­ula (nòd-­ul ‘node-­ dim’), -­cul -­cula (grup-­ús-­cul ‘group-­dim’) and -­ol -­ola (folí-­ol ‘leaf-­dim’). On another note, we must also remember that there are certain cases that waver between a diminutive value and that of an entirely new meaning (therefore lending them a value similar to that of derivative suffixes). This is seen with names that indicate the young of certain animals (14a), the names of winds (14b), the names of fruits (14c), etc.:

Catalan 213 (14) a. balen-­ó whale-­ dim ‘whale calf’ b. ponent-­ol westerly wind-­dim ‘soft westerly wind’ c. ginebr-­ó juniper-­ dim ‘juniper berry’ 3.3  Evaluative morphology in verbs While the current productivity of evaluative morphology in verbs is, to say the least, doubtful, it is always associated with other meanings related to aspect, whether iterative or denoting frequency, as stated by Dardano (1978) or Grandi (2008) in the case of Italian. Along the same lines and as proposed by Rifón Sánchez (1997) in the case of Spanish, it is possible to create a semantic subcategory which includes the various morphemes that intervene in evaluative morphology in verbs and which have three primary meanings: iteration-­habitual (iter), intensity-­attenuated (att) and pejorative. In Table 15.2.3 we list the primary examples.8 Table 15.2.3. Verbal evaluative suffixes in Catalan Suffix

Examples

-­uc -­iny -­ass -­iss -­isc -­ic

menj-­uc-­ar ‘to eat-­att’ plor-­iny-­ar ‘to cry-­att/pej’ allarg-­ass-­ar ‘to stretch-­pej’ adorm-­iss-­ar-­se ‘to fall asleep-­att’ enamor-­isc-­ar ‘to fall in love-­att’ plor-­ic-­ar ‘to cry-­att/pej’

Quite frequently these suffixes are followed by -­ej-­: bes-­ot-­ej-­ar ‘to kiss-­iter/att/ cant-­uss-­ej-­ar ‘to sing-­att’, dorm-­isqu-­ej-­ar ‘to sleep-­iter/att’ – cases which must not be confused with the use of -­ejar as a verbalising suffix (here, the base is already a verb) and must instead be seen as an element that lends a negative aspect to the base, even in cases which did not previously include evaluative suffixes, as in the case of trafic-­ar ‘to traffic’ – trafiqu-­ej-­ar ‘to traffic-­pej’. Furthermore, verbal evaluative suffixes in Catalan are also accompanied by a prefix: es-­crid-­ass-­ar ‘to shout-­pej’. In all cases, the evaluated verb is a verb of the first conjugation, regardless of the conjugation of the base verb: ploure ‘to rain’ (2nd conjugation) – plov-­isqu-­ej-­ar ‘to rain-­iter/att’ (1st conjugation). pej’,

3.4  Other evaluative affixes In contrast to the suffixes which we have just presented, there are several others which more closely approximate derivative suffixes and which ultimately modify aspect. This is illustrated by the suffix -­astre -­astra, which we find in words such as those in (15):

214

Elisenda Bernal

(15) poet-­astre poet-­ pej ‘bad poet’

politic-­astre politician-­ pej ‘bad politician’ We also have the suffix -­oide, though the formal definitions found in dictionaries would seem to limit its use to indicating ‘general similarity’. Nevertheless, De Bruyne (1989) and Lázaro Mora (1999), writing about Spanish, have shown that, in specific contexts, this suffix has a tone indicating grotesque, mocking exaggeration. The suffix -­oide is found in many neologisms, as demonstrated by Cañete, Fernández and Janer (forthcoming): (16) intel·lectual-­oide pej intellectual-­ ‘pseudo-­intellectual’

romantic-­oide romantic-­ pej ‘excessively romantic’

sentimental-­oide sentimental-­ pej ‘excessively sentimental’ This same pejorative connotation can also be found in neologisms bearing the suffixes -­esc -­esca (whose meaning dictionaries seem to limit to a relational value) and -­il (whose application, according to Catalan dictionaries, is limited to a technical use in the field of chemistry):9 (17) xul-­esc cocky-­ pej ‘cocky’ caciqu-­il despot-­ pej ‘despotic’

gamberr-­il joker-­ pej ‘puckish’ Along the same lines, some articles (Bernal and Sinner 2009; Cañete et al., forthcoming; Bernal, Gené and Llopart, forthcoming) argue that certain prefixes such as mini-­, macro-­, super-­, hiper-­, ultra-­, etc. act by lending nuance and degree to the base, rather than by creating a new word, making them similar to the evaluative function of the suffixes we addressed in Sections 3.1 and 3.2. Consequently, forms such as super-­bé ‘aug/sup-­ well’,  super-­ampli ‘aug/sup-­wide’, hiper-­ric ‘aug/sup-­rich’, etc. lean more towards an augmentative and superlative value than towards their original locative value. When

Catalan 215 examining prefixes which indicate appreciation, we must also include the prefix re-­as seen in forms such as re-­bonic ‘int-­nice’ or re-­petit ‘int-­small’, in which adjectives are clearly intensified.

4 Conclusion Catalan is, without a doubt, a rich language when seen in terms of evaluative morphology: we have examples of many diverse suffixes which can modify all grammatical categories (with the exception of prepositions), though currently actual productivity seems to be limited to a few (-­et -­eta and -­ó -­ona for diminutives, and -­às -­assa and -­ot -­ota for augmentatives). Nevertheless, we must also highlight the fact that the language seeks new ways to add nuance and degree to various words by using affixes that traditionally did not play a role in evaluative morphology.

Notes 1. For the attachment of evaluative suffixes to verbs, see Section 3.3. 2. We assume, like Lloret (1996), that the evaluative morpheme is attached to the head of the compound, even though, semantically, it refers to the entire form: a renta+plat-­et-­s is not an appliance that washes small plates but rather a small appliance that washes plates. 3. However, Lloret (1996; 2013) notes that in informal contexts the productivity of the process is maintained: Lloret has documented forms such as fotfem-­assa ‘photo-­aug’, clitormasc-­et ‘clitoris-­ dim’, virusmasc-­et ‘virus-­dim’, etc. 4. Gràcia and Turon (1998, 18–19) indicate that there are certain rare cases in which a derivative suffix can be attached following the addition of an evaluative suffix: from gran-­ell ‘spot-­dim’ we obtain gran-­ell-­ut ‘spotted’. However, in cases where lexicalisation occurs, such as bot-­ ell-­a ‘bottle’ – em-­bot-­ell-­ar ‘to bottle’, or gaset-­ill-­a ‘short news article’ – gaset-­ill-­er ‘person that writes a short news article’, these authors consider that the base itself acts as an underived complex word and can therefore itself act as a base of derivation. 5. In the case of two consecutive diminutives, it seems that the general tendency is for -­et to be the final one in the sequence: carrer-­on-­et ‘street-­dim-­dim’, pobr-­iss-­et/pobr-­iss-­ó ‘poor-­dim-­dim’, yet we see pobr-­iss-­on-­et ‘poor-­dim-­dim-­dim’ (and not *carrer-­et-­ó, *pobr-­iss-­et-­ó). However, there are specific cases in which -­et cannot appear in the final position: rei-­et-­ó ‘king-­dim-­dim’, ti-­et-­ona ‘aunt-­dim-­dim’ and not *rei-­on-­et, *ti-­on-­eta. 6. For an article on the use of evaluative morphology when applied to child language within the context of fictive orality, see Bernal (2012). 7. It must be noted, however, that in most Romance languages the equivalents of suffix -­ó -­ona express augmentative value. 8. While Fabra (1956) classifies these elements as infixes when they are attached to verbal bases, we share the position of Gràcia and Turon (1998), which states that suffixes are the same as those that can be found in nominal and adjectival bases and, like them, are situated after the base and any derivative suffixes (should there be any) and before inflective morphemes. 9. All examples were taken from the database of the Observatori de Neologia at the Universitat Pompeu Fabra (http://www.iula.upf.edu/obneo).

15.3 Georgian

Manana Topadze Gäumann

1 Introduction Georgian is a Kartvelian (South Caucasian) language with about 4 million speakers. It is the official language of Georgia and is also spoken by Georgian communities in northeast Turkey, in the north Caucasus, in Azerbaijan and in Iran (Boeder 2005, 6). Georgian has about seventeen dialects belonging to an Eastern or Western group. They differ from the literary standard, which is based on the East–Central dialects. Georgian is the only written language of the Kartvelian family with its own script. It consists of a phonemic alphabet with thirty-­three characters (twenty-­eight consonants and five vowels). Stops and affricates have a three-­way opposition between voiced, voiceless aspirate and voiceless glottalised. Long consonant clusters in initial position are common. Georgian is an agglutinative language with inflectional features. It shows polypersonal verbal constructions (there are two sets of pronominal affixes which mark subject and object on the verb). The split ergative system is typical of Georgian: the subject of transitive and so-­called medial verbs in the aorist series is marked by the ergative case, while the direct object of transitive verbs and the subject of intransitive verbs require the nominative case. The noun has seven cases, and the adjective, which usually precedes the noun, agrees with it in case but not in number. There are no articles and there is no grammatical gender. According to their morphosyntactic and semantic features, verbs are generally divided into four classes: ‘transitive’, ‘intransitive’, ‘medial’ and ‘inversion’ verbs. Word order is quite free in Georgian. The typical order is SOV but other constructions are also allowed depending on contextual factors.

2  Evaluative strategies: an overview The term ‘evaluative morphology’ (EM) is not common in the traditional grammar of Georgian. Studies concerning evaluative affixes are collocated within derivational morphology and are usually limited to diminutive markers, which, in Shanidze’s (1980, 120) words, can be used ‘on the one hand, for conveying endearment and on the other, in order to express contempt’. EM in Georgian covers diminution, contempt, endearment, approximation, attenuation and intensification. Evaluative markers co-­occur mainly with nouns but also with adjectives, numerals and verbs. Georgian uses three main derivational means to form new stems: reduplication, compounding and affixation. All of them can be employed to convey evaluation.

Georgian 217 2.1  Old Georgian EM in Old Georgian is rather restricted since the written language of the fifth to the eleventh century, being the language of religious literature, is characterised by a textually rigorous style. There are two diminutive affixes, -­a and -­ak’, in use: (1) kal-­i kal-­a-­y woman-­ nom woman-­ dim-nom ‘woman’ ‘young woman’ (2) saxl-­i saxl-­ak’-­i nom house-­ dim-­nom house-­ ‘house’ ‘a little house’ (3) k’nin-­i k’nin-­ak’-­i small-­ nom small-­ dim-­nom ‘small’ ‘very small’ According to Martirosovi (1978, 113–14) the diminutive suffix -­ak’ which occurs in Old Georgian may originate from Persian loanwords such as avazak’-­i ‘robber’ (Middle Persian ávāzak), zwarak’-­i ‘sacrifice’ (Middle Persian *zōhrak), danak’-­i ‘knife’ (Middle Persian *danak) etc. Some of these loanwords have been shortened at a later stage of language development, giving rise to the parallel forms with and without the stem-­final consonant -­k’: (4) danak’-­i →  dana č’alak’-­i →  č’ala knife-­ nom knife-­.nom grove-­ nom grove.nom ‘knife’ ‘knife’ ‘grove’ ‘grove’ This process led to a semantic differentiation of double forms. The stem-­final -­ak’ of the loanwords grammaticalised as a diminutive suffix and has spread by analogy to native stems, whereas clipped forms came to be employed as semantically neutral words (Martirosovi 1978, 116–17). In Modern Georgian the suffix -­ak’ occurs mainly in lexicalised diminutives: (5) c’ign-­i → c’ign-­ak’-­i book-­ nom book-­dim-­nom ‘book’ ‘notebook’ 2.2  Modern Georgian In Standard Georgian there are a large number of evaluative affixes, but not all of them are equally productive. A diminutive value can be attributed to the following suffixes: -­a, -­ak’, -­ik’, -­ok’, -­uk’, -­an, -­in, -­on, -­un, -­al, -­il, -­at’, -­ot’, -­ut’, -­ic, -­uc, -­ič, -­uč, -­ič’, -­uč’, -­ux (Gogolashvili 2011, 235). Most of them have a dialectal origin and are not widespread in all varieties of Georgian. Some of these suffixes are fossilised in a restricted number of words and only occur once. Since the simple (monosyllabic) suffixes, with a few exceptions, are no longer productive, they must be obligatorily combined with one or more

Manana Topadze Gäumann

218

diminutive suffixes to convey the evaluative meaning. In the following examples combined suffixes are considered as one suffix. 2.3  Contempt, diminution and endearment The boundaries between quantitative and qualitative evaluation are not clear-­cut. The same suffix may encode small size, contempt or endearment, depending on the context. If the stem of a derived word is not semantically compatible with a diminutive meaning, the evaluative affix will usually acquire a pejorative value: (6) glex-­i → glex-­uč’a nom farmer-­ dim farmer-­ ‘farmer’ ‘peasant’ Contempt towards someone can also be expressed by means of compounding, as in the following examples, where the interjection vai ‘woe’ modifies the head of the compound, which acquires the meaning of ‘bad, useless, unqualified, non-­professional’:1 (7)



poet’-­i poet-­NOM ‘poet’

vai-­poet’-­i pej-­poet-­nom ‘a bad poet’

(8) mosamartle vai-­mosamartle judge.nom pej-­judge.nom ‘judge’ ‘an incompetent judge’





(9)

xelmdzghvanel-­i vai-­xelmdzghvanel-­i director-­NOM pej-­director-­nom ‘director’ ‘good-­for-­nothing director’

In Modern Georgian quantitative diminution (small size) is generally expressed by periphrastic forms (Tuskia 2004, 119): (10) saxl-­i vs. p’at’ara house-­NOM little.nom ‘house’ ‘a little house’

saxl-­i house-­NOM

Only in very few nouns can small size or young age be conveyed by the following combinations of diminutive suffixes: (11) -­ik’ + a t’omara t’oms-­ik’a sack.nom sack-­ dim ‘sack’ ‘a little sack’ (12) -­un + a čit’-­i čit’-­una bird-­ nom bird-­ dim ‘a bird’ ‘a little bird’

Georgian 219 (13) -­ ur+a

gube gub-­ura puddle.nom puddle-­ dim ‘puddle’ ‘a little puddle’

In other instances, diminutive markers bear emotional overtones and the expression of quantitative diminution is either missing or overlaps with the qualitative one. Endearment and affection are frequently expressed by the combination of two or more diminutive suffixes (-­ik’o, -­ilo, -­ik’-­on-­a, -­un-­a, -­un-­ia), as well as by the suffix -­a. These markers can occur with a limited number of nouns: (14) deda mother.nom ‘mother’ ded-­ik’o ded-­ilo ded-­ik’ona mother-­ end mother-­end mother-­end ‘mummy’ (15) tagv-­i tag-­unia mouse-­ nom mouse-­dim/end ‘mouse’ ‘little mouse’ (16) kal-­i kal-­a woman-­ nom woman-­ dim/end ‘woman’ ‘(sweet) little girl’ In rare instances, adjectives can be combined with diminutive suffixes in order to express the endearment: (17) lamaz-­i → lamaz-­una beautiful-­ nom beautiful-­ dim/end ‘beautiful’ (18) celk-­i → celk-­una mischievous-­ nom mischievous-­ dim/end ‘mischievous’ Evaluative affixes and their combinations acquire pejorative overtones when they convey a characteristic property or similarity: q’b-­ača (from q’ba ‘jaw’+ pej) ‘someone with big jaws’, kal-­ačuna (from kali ‘woman’+ pej) ‘faint-­hearted, coward man’. Diminutives are often used to create new lexical entries for technical or scientific terminology. In such instances they usually lose the evaluative meaning (Gogolashvili 2011, 238): (19) k’bil-­i → k’bil-­ana tooth-­ nom tooth-­ dim ‘tooth’ ‘cogwheel’ (20) p’ark’-­i → p’ark’-­uč’-­i sac-­ nom sac-­ dim-­nom ‘sac’ ‘ventricle’

Manana Topadze Gäumann

220

A rather marginal case of ‘diminutive meaning’ is mentioned by Pochkhua (1974, 136–7). The following compounds with xel-­(from xeli ‘hand’) as their first constituent owe their meaning to an implicature that items which you handle with one hand tend to be relatively small: xel-­čanta hand-­bag.nom ‘handbag’ (21)

(22)

(23)



(24)

xel-­džox-­i hand-­stick-­nom ‘walking stick’ xelc’ald-­i hand-­axe-­nom ‘little axe’ (to use with only one hand) xeltepš-­i hand-­plate-­nom ‘plate of small dimensions’

Such words, however, are obsolete in present-­day Georgian. Only a few of them survive in everyday speech. 2.4 Hypocorisms Some evaluative affixes, such as -­ik’o, -­ak’o, -­uca, -­una and -­unia, show a high degree of productivity when they are attached to first names in order to form hypocorisms. Furthermore, in present-­day colloquial Georgian new forms of hypocorisms of the monosyllabic pattern CVC (or less frequently CV) can be observed. They are built by dropping the final vowel of the base or the derived stem. These forms are employed to convey not only endearment, but also intimacy, familiarity, emotion or informality, and are very popular among young people: (25) Full name Hypocorism ‘New’ hypocorism a. laša laš-­ik’o laš Lasha.nom Lasha-­ dim Lasha ‘Lasha’ b. nino nin-­ik’o nin-­uca nin Nina.nom Nina-­ dim Nina-­ dim Nina ‘Nina’ In colloquial speech, endearment in hypocorisms is sometimes intensified by the plural marker -­eb: (26) Full name Hypocorism Intensified endearment a.  ek’at’erine ek’-­uča ek’-­uč-­eb-­i Catherine Catherine-­ dim Catherine-­ dim-­pl-­nom ‘Catherine’

Georgian 221 b. lia Lia ‘Lia’

l-­ik’una l-­ik’un-­eb-­i Lia-­ dim Lia-­ dim-­pl-­nom

In certain contexts, hypocorisms may convey contempt. Historically, hypocorisms indicated affiliation to a certain social class. As several authors observe (Kobakhidze 1962, 158; Topuria 1979, 102), the first names of princes and noblemen were never mentioned with diminutive forms, while peasants, by contrast, were frequently called by their diminutive names in order to mock them. 2.5  Intensification



Intensification can be conveyed occasionally by the reduplication of qualitative adjectives: (27) (28)

cxel-­cxel-­i p’ur-­i nom hot-­hot-­nom bread-­ ‘fresh bread’ axal(t)-­axal-­i t’anisamos-­i new-­new-­nom clothes-­ nom.sg ‘brand-­new clothes’

2.5.1 Superlative There are two ways of expressing superlatives: morphologically, by means of the circumfix u-­es, or by using a periphrastic form. Positive Superlative lamaz-­i u-­lamaz-­es-­i beautiful-­ nom sup-­beautiful-­sup-­nom ‘beautiful’ ‘the most beautiful’ (29)



Superlative (periphrastic form) q’vela-­ze lamaz-­i all-­on beautiful-­nom ‘the most beautiful’

In colloquial speech, superlatives can be intensified by repetition (duplication, triplication, etc.) of the prefix u-­(e.g. uu-­lamaz-­es-­i, uuu-­lamaz-­es-­i etc.). Such forms occur frequently but are deemed artificial (Gogolashvili 2011, 222). 2.6  Attenuation and approximation 2.6.1 Adjectives In adjectives, only attenuative and superlative degrees can be expressed morphologically by means of circumfixes. The comparative is obligatorily conveyed by periphrastic forms (such as the adverb upro ‘more’ + the positive degree of the adjective). Attenuation is expressed by the circumfix mo-­o, which attaches to qualitative adjectives. (30) m’cvane mo-­mc’van-­o green.nom att-­green-­att ‘green’ ‘greenish’

Manana Topadze Gäumann

222 (31) mžave mo-­mžav-­o sour.nom att-­sour-­att ‘sour’ ‘slightly sour’ (32) sulel-­i mo-­sulel-­o stupid-­ nom att-­stupid-­att ‘stupid’ ‘foolish’

In Eastern dialects, the circumfix mo-­o is sometimes replaced by c’a – o (Gogolashvili 2011, 224): (33) lurdž-­i c’a-­lurdž-­o blue.nom att-­blue-­att ‘blue’ ‘bluish’ 2.6.2 Preverbs Preverbs have a variety of functions in Georgian; they can be simple or complex (with the preverb mo-­ ‘hither’). Preverbs indicate directionality with verbs denoting motion. They can also change the basic meaning of the verb. Furthermore, preverbs are employed for the formation of future and aorist stems, where they have the function of indicating the perfective aspect. Some preverbs have also acquired additional functions in Modern Georgian. We will consider the evaluative function here. Attenuative function can be attributed to the following preverbs: še-­, c’a-­, mi-­and mo-­, since they indicate that the action expressed by the verb is performed for a short time, to a reduced extent or superficially (Aronson 1982, 440–1; Chimke 2010, 139–40). However, the expression of attenuation by means of preverbs is possible only with a restricted number of verbs. Attenuative preverbal forms occur only in the aorist or future, where the preverbs code perfective aspect and completed action (Monaselidze 2001, 24). Imperative forms are also possible, since the imperative in Georgian (second person singular and plural) coincides formally with the aorist. In the present and imperfect tenses, periphrastic forms are required to express attenuation. (34)

Neutral form da-­’c’veba prv-­lie down.fut.3sg ‘she or he will lie down’

Attenuative mi-­c’veba prv-­lie down.fut.3sg ‘she or he will lie down for a moment’

(35)

isauzmebs c’a-­isauzmebs have a breakfast.fut.3sg prv-­have a breakfast.fut.3sg ‘she or he will have breakfast’ ‘she or he will have a quick breakfast’

(36)

da-­tvreba še-­tvreba prv-­become drunk fut.3sg prv-­become drunk fut.3sg ‘she or he will become drunk’ ‘she or he will become tipsy’

(37)

ga-­atbo še-­atbo prv-­warm.aor:3sg prv-­warm.aor.3sg ‘she or he warmed something’ ‘she or he warmed something a little’

Georgian 223 Attenuative preverbal forms are frequently not contained in the dictionaries, since they are not obligatory and depend on a speaker’s choice and subjective attitude towards a statement. When Georgian attenuatives are translated into English, periphrastic forms must be used to maintain the evaluative overtones. 2.6.3  Clitics and particles Another possibility for expressing attenuation or approximation is the suffix -­savit (a clitic composed of the dative marker [-­s] + the epenthetic vowel [a-­] and the postposition [-­vit] ‘like’, ‘as’), originally a comparative postposition, which grammatcalised as an inferential marker with overtones of approximation (Topadze Gäumann 2011, 135). It can be attached to the inflected verb in order to attenuate its original meaning: (38)

davighale-­savit tired.aor.1sub-­es ‘I got like tired’ prv-­get

Approximation with numerals can be expressed by the particle -­ode (from Old Georgian oden ‘only’), which attributes an estimated value to a cardinal number: (39)

xuti five.nom ‘five’

xut-­i-­ode five-­nom-­appr ‘about five’

2.7  Georgian dialects Evaluative strategies have a high saturation in Georgian dialects in contrast to the Standard variety. Diminutives are very widespread both in Eastern (Khevsurian, Pshav, Mokhevian) and Western (Rachian, Gurian) dialects (Topuria 1979, 102; Arabuli 2009, 20). The diminutive suffix -­uk’a, which occurs in Standard Georgian as well, seems to originate from Rachian, where it conveys contempt with hypocorisms and small size with other nouns (Kobakhidze 1962, 159). In Gurian, evaluative markers can convey diminution, contempt or endearment depending on the context (Tughushi 1992, 134). In the Eastern Mountain dialects Pshav and Khevsurian, nouns and pronouns can agree with adjectives and numerals by adding a diminutive suffix (Arabuli 2009, 21): Mokhevian: (40) ert-­a tik’an-­a xq’vanda one-­ dim kid-­ dim have.pst.3sub ‘she or he had one kid (young goat)’ (41) or-­a-­i urem-­a-­i p’ur-­a-­i two-­ dim-­nom wheelbarrow-­ dim-­nom bread-­ dim-­nom ‘two wheelbarrows of wheat’ Pshavian: (42) k’arg-­unela kala-­ia good-­ dim woman-­ dim ‘good woman’

Manana Topadze Gäumann

224 Khevsurian: (43)

ai švil-­a gačenila ož-­a here child-­ dim be born.prf.3sub male-­ dim ‘a male child was born here’

Arabuli (2009, 22) points out three patterns of diminutive agreement in Eastern dialects and assumes that this kind of agreement with a diminutive marker may be considered an areal feature of Georgian Mountain dialects: (i) Adjective + noun-­dim (ii) Adjective-­ dim + noun (iii) Adjective-­ dim + noun-­dim

k’arg-­i bič’-­uk’a good-­ nom boy-­ dim k’arg-­a kal-­i good-­ dim woman-­ nom k’arg-­a kal-­a good-­ dim woman-­ dim

3  Evaluative strategies in Svan, Megrelian and Laz In the Kartvelian family, Svan shows the highest saturation of EM. Evaluative affixes (mainly diminutives) are very frequent and can be attached to both proper and common names, as well as adjectives, numerals and participles. Several authors (Chumburidze, Nizharadze and Kurdadze 2007, 75; Gazdeliani 1992, 141; among others) mention that diminutives also co-­ occur with nouns such as ‘sun’, ‘moon’, ‘month’, ‘army’, etc., without discussing which semantic value is attributed to these nouns when they occur with a diminutive suffix. In rare instances, prevalently in folklore, diminutive suffixes occur with inflected verbs or adverbs (Chumburidze et al. 2007, 75). The use of the diminutive suffixes -­ild/-­uld (and their allomorphs -­old, -­ǝld, -­il, -­el, -­al, -­ol, -­ul, -­ǝld, -­d, -­ad, -­ǝd) is very widespread, especially in the Upper Bal and Lashkh dialects of Svan (Topuria 1979, 102; Gazdeliani 1992, 141). These diminutives convey endearment, small size and approximation. They also often occur with numerals, where they express an inexact quantity (Chumburidze et al. 2007, 76): (44)

oštxû oštx-­uld four four-­ dim ‘four’ ‘approximately four’

The two other languages of the Kartvelian family, Megrelian and Laz, employ the same evaluative strategies as Georgian. In both languages there are derivational affixes, which can express diminution, contempt, intensification or endearment. The following diminutive suffixes are encountered in Megrelian and Laz: -­a, -­ia, -­ela, -­ik’a, -­ina, -­ec’a, -­ic’a, -­ut’a, -­ik’ina (Akhalaia 2006, 31). Several studies are dedicated to the formal and diachronic aspects of diminutive markers, but these lack a detailed discussion of the degree of productivity and occurrence frequency of these suffixes (Lomtatidze 1987; Akhalaia 2006; Shonia 2011; among others). The diminutive suffix -­a in Megrelian is mainly used for the formation of hypocorisms (gerge ‘George’ – gerg-­a; davit ‘David’ – davit-­a; petre ‘Peter’– pert-­a), whereas the suffix

Georgian 225 -­ia can be employed for the expression of diminution or endearment depending on the context (Shonia 2011, 211): Base noun Diminutive (45) argun-­i argun-­ia axe-­ nom axe-­ dim ‘axe’ ‘little axe’ (46) ork’ol-­i ork’ol-­ia nom jug-­ dim jug-­ ‘jug’ ‘little jug’ (47) k’ot’o k’ot’o-­ia pot.nom pot-­ dim ‘pot’ ‘little pot’ Unlike in Georgian, pejorative meaning in Megrelian can be expressed by the combination of the participial prefixes na-­/no-­ with the diminutive suffix -­ia, whereby the final vowel of the base noun is dropped for phonological reasons. The derived noun acquires the meaning of ‘old, useless, worthless’ (Shonia 2011, 212): (48) k’ot’o na-­k’ot’-­ia pot.nom pej-­pot-­pej ‘pot’ ‘old, useless pot’ (49) k’aba no-­k’ab-­ia dress.nom pej-­dress-­pej ‘dress’ ‘worthless, old dress’ In Laz, the following diminutive suffixes are used to convey small size: -­ik’a, -­ina, -­ena, -­ic’a (Shonia 2011, 212): (50) -­ik’a: xodž-­i xodž-­ik’a oxen-­ nom oxen-­dim ‘oxen’ ‘little oxen’ (51) -­ ina: oxor-­i oxor-­ina house-­ nom house-­ dim ‘house’ ‘little house’

4 Conclusion The domain of EM in Kartvelian covers diminution, contempt, endearment, attenuation, approximation and intensification. All Kartvelian languages share expression strategies such as affixation, reduplication and compounding. In Megrelian, an additional pejorative circumfix no-­/na – ia is encountered, which has no equivalent in Georgian. The occurrence of evaluative markers is restricted in Standard Georgian, but is very widespread in some dialects, as well as in Svan, where EM shows a high

226

Manana Topadze Gäumann

saturation. Further investigation is needed to explore other possible devices of EM in Kartvelian.

Acknowledgements I am grateful to Winfried Boeder for bibliographical information and for his constructive and valuable comments on the draft of this chapter. I would also like to thank Lucy Zuberbühler for correcting my English. Any remaining errors are my own responsibility.

Note 1. In such compounds vai-­ seems to be a calque of Russian gore ‘woe’ (cf. Russian gore-­kritik ‘a miserable critic, criticaster’ and its Georgian counterpart vai-­k’rit’ik’osi).

15.4 Hungarian

Ferenc Kiefer and Boglárka Németh

1 Introduction Hungarian is a Uralic language spoken by c. 13 million speakers in Central Europe. It belongs, together with Mansi and Khanti, to the Ugric branch of Uralic. In addition to Ugric the Uralic language family comprises the following branches: Finnic (Finnish, Estonian, Sami and some minority languages), Mordvin, Mari, Permic (Udmurt and Komi), and Samoyedic (Nenets, Enets, Nganasan, Selkup). Some of these languages are highly endangered. The Uralic languages are predominantly suffixing; prefixes, if any, are clearly innovations and are restricted to derivational morphology (Kiefer and Laakso, 2014). All these languages have a rich and complex derivational system, and some of the suffixes may express an evaluative meaning. With a few exceptions, evaluative morphology in Uralic is restricted to derivation.

2  Evaluative morphology In evaluative morphology the semantic and the pragmatic meaning cannot always be kept strictly apart (see the detailed discussion in Dressler and Merlini Barbaresi 1994). For example, the diminutive suffixes -­ka/-­ke, -­cska/-­cske,1 attached to nouns, rarely express just smallness: their pragmatic meaning normally overrides semantics. If one wants to express the fact that the object referred to is small, normally the adjective kis, kicsi ‘small’ is used rather than the diminutive suffix. Moreover, suffixed forms with a clear referential (semantic) meaning, such as those in (1a–c), must be considered to be separate lexical entries. With adjectival bases, such as (1d), suffixed forms are always evaluative. (1) a. tál – tál-­ka dish – dish-­ dim ‘dish’ – ‘small dish, bowl’ b. asztal – asztal-­ka table – table-­ dim ‘table’ – ‘small table’ c. szó – szó-­cska word – word-­ dim ‘word’ – ‘little word, particle’

228

Ferenc Kiefer and Boglárka Németh

d. csúnyá-­cska plain-­ dim ‘rather plain’ The frequentative suffix -­gat/-­get, which is attached to verbs, can only express pure repetition when it is attached to punctual verbs. According to our view such cases lie outside of the scope of evaluative morphology. Compare (2a) and (2b), where olvas ‘read’ is a durative and nyit ‘open’ is a punctual verb. In the first case the suffix has an attenuative meaning in addition to iterativity. (2) a. olvas-­gat read-­freq ‘read leisurely’ b. nyit-­ogat open.tr-­freq ‘open repeatedly’ A special case is represented by the excessive, which expresses the highest possible degree of a property and is expressed by the iteration of the superlative prefix, e.g. (3a-­b). (The excessive also occurs in Austrian German, see Dressler / Kiefer 1990.) (3) a.  leg-­jo-­bb sup-­good-­comp ‘best’ b. legesleg-­jo-­bb sup-­good-­comp ‘best of all’ In what follows we will have a closer look at the diminutive suffix of nouns and of adjectives, the frequentative form of verbs and the excessive. 2.1  Nouns with the diminutive suffix Nouns can form the diminutive productively either with the suffix -­ka/-­ke or with the suffix -­cska/-­cske. The latter is more productive since it applies to a wider range of nouns and obeys fewer restrictions. The diminutive suffix -­cska/-­cske cannot occur after the hypocoristic suffix -­i, while the suffix -­ka/-­ke can (e.g. (4)). The suffix -­ka/-­ke cannot be attached (i) to monosyllabic nouns (e.g. (5a)), (ii) to polysyllabic nouns ending in a stop (e.g. (5b)) and (iii) to nouns ending in a or e (e.g. (5c)). (4) Bog-­i-­ka – *Bog-­i-­cska Bog-­ dim-­dim – Bog-­dim-­dim (5) a. láb – *láb-­ka foot – foot-­ dim ‘foot’ – ‘little foot’ b. család – *család-­ka family – family-­ dim ‘family’ – ‘little family’

Hungarian 229 c. alma – *almá-­ka apple – apple-­ dim ‘apple’ – ‘little apple’ These restrictions can be explained by morphophonological regularities. Note that in Hungarian a large number of nouns end in -­ka/-­ke, e.g. (6a–b), where the use of the diminutive suffix -­cska/-­cske warrants the morphotactic transparency of the derivatives. Compare (6b) with (6c). (6) a. béka ‘frog’ b. róka ‘fox’ c. róká-­cska – *róká-­ka dim – fox-­ dim fox-­ ‘fox’ – ‘little fox’ Note furthermore that all nouns derived from monosyllabic words are lexicalised formations (e.g. (7a–d)). Consequently, they are semantically not (fully) transparent. (7) a. szál – szál-­ka thread – thread-­ dim ‘thread’ – ‘splinter’ b. tál – tál-­ka dish – dish-­ dim ‘dish’ – ‘small bowl’ c. zár – zár-­ka lock – lock-­ dim ‘lock’ – ‘cell’ d. pad – pad-­ka bench – bench-­ dim ‘bench’ – ‘small bench’ In the case of alternating stems, the suffix -­ka/-­ke always occurs with lexical bases; the suffix -­cska/-­cske, on the other hand, gets attached to the relative stem, and if the noun ends in a consonant it requires an epenthetic vowel. This explains why the suffix -­ka/-­ke does not occur with nouns ending in a or e, since these suffixes combine with the lengthened final vowel, i.e. with relative stems (Ladányi 2007, 153–65). At the same time, parallel derivations, such as the ones in (8), do occur. (8)

citrom – citrom-­ka – citrom-­ocska lemon – lemon-­ dim – lemon-­ dim ‘lemon’ – ‘little lemon’ – ‘little lemon’

Completely excluded from diminutive formation are action nouns, which are derived from verbs by means of the suffix -­ás/-­és, as in (9a–b). Such nouns refer to events rather than to objects, and events cannot take diminutives. (9) a. olvas-­ás read-­ nmls ‘reading’

Ferenc Kiefer and Boglárka Németh

230 b. főz-­és cook-­ nmls ‘cooking’

Agent nouns derived from verbs by means of the suffix -­ó/-­ő may occasionally be suffixed by the diminutive, but the meaning of the derivative involves not smallness but rather contempt or irony, as in the examples in (10a) and (10b). The same holds true for monomorphemic agent nouns, e.g. (10c). (10) a. ír-­ó – ír-­ó-­cska prs.ptcp – write-­prs.ptcp-­dim write-­ ‘writer’ – ‘petty writer’ b. jegyz-­ő – jegyz-­ő-­cske note-­ prs.ptcp – note-­prs.ptcp-­dim ‘notary’ – ‘petty notary’ c. orvos – orvos-­ocska doctor – doctor-­ dim ‘medical doctor’ – ‘petty medical doctor’ Abstract nouns derived from adjectives by means of the suffix -­ság/-­ség do not take the diminutive suffix for obvious reasons, as shown by (11). In this case we have to do with abstract notions and not with persons or objects. (11)

szép – szép-­ség – *szép-­ség-­ecske beautiful – beautiful-­ deadj – beautiful-­deadj-­dim ‘beautiful’ – ‘beauty’

As far as the meaning of the diminutive derivatives is concerned, they do not normally (if at all) express (just) smallness. In general, the pragmatic meaning of the diminutive is more important than its semantics. This meaning can be derived from its semantic meaning, the sentential context and the speech situation. Names of tiny animals are often used in a diminutive form, except in scientific discourse (e.g. (12a–b) (Kiefer 2004, 335). Names of animals, such as those in (12c–d), occur regularly in the diminutive in books meant for children, even if the animals are not small. (12) a. méh-­ecske bee-­dim ‘tiny bee’ b. egér-­ke mouse-­dim ‘small mouse’ c. tyúk-­ocska hen-­dim ‘little hen’ d. juh-­ocska sheep-­dim ‘little sheep’

Hungarian 231 Diminutive forms are also typical of baby talk (caretaker speech). The nouns concerned refer primarily to body parts (e.g. (13a)), to kinds of food that belong to the diet of a small child (e.g. (13b)) or to objects with which a small child gets in contact (e.g. (13c)). (13) a. láb-­acska foot-­dim ‘little foot’ b. tej-­ecske milk-­dim ‘little milk’ c. ágy-­acska bed-­dim ‘little bed’ The body parts of a small child are really small, but the amount of milk offered need not be little. And what is more important is that normally one would not offer milk to an adult using the diminutive form. In addition to baby talk, other typical situations in which diminutives are used are talking to a lover or to a pet. The pragmatic meaning in all these cases is endearment, affection. 2.2  Adjectives with the diminutive suffix Adjectives take the same diminutive suffixes as nouns and have exactly the same morphophonological restrictions. Corpus search (Hungarian National Corpus) shows that adjectives with negative meaning take the diminutive suffix more often than those with positive meaning. An argument can be strong or weak, but the diminutive form of erős ‘strong’ does not occur; on the other hand, gyenge ‘weak’ does appear with the diminutive suffix, as in (14a). Similarly we find examples such as those in (14b–d) in the corpus. (14) a. gyengé-­cske érv weak-­dim argument ‘(a) rather weak argument’ b. butá-­cska silly-­dim ‘little silly’ c. gyengé-­cske weak-­dim ‘weakish’ d. beteg-­ecske ill-­dim ‘little ill’ The use of diminutives in such cases may have pragmatic reasons. We do not always want to tell the truth, therefore we resort to various subduing strategies. One such strategy is to negate the positive adjective: to be not clever sounds better than to be stupid, and to be not strong sounds better than to be weak. We use the diminutive forms for the same purpose of illocutionary mitigation: the diminutive suffix subdues or tones down the negative meaning and hence the impoliteness involved.

232

Ferenc Kiefer and Boglárka Németh

To be sure, adjectives with positive meaning, too, can easily be suffixed by the diminutive suffix: we may get forms such as (15a–c), which are often used with children or young ladies (by men) and express endearment or flattery. (15) a. okos-­ka clever-­dim ‘cute little’ b. csinos-­ka pretty-­dim ‘pretty little’ c. ügyes-­ke skilful-­dim ‘skilful little’ Measure adjectives if suffixed with the diminutive suffix normally receive an evaluative meaning, as in (16a–b). (16) a. hosszú – hosszú-­cska long – long-­dim ‘long’ – ‘little longish’ b. rövid – rövid-­ke short – short-­dim ‘short’ – ‘very short’ 2.3  Verbs with the frequentative suffix In the field of evaluative morphology the research on verbal categories has been a relatively underrepresented domain, as very few studies deal with questions related to verbal evaluatives. According to Grandi (2009, 47), for example, this is due to the fact that verbal evaluatives are not as cross-­linguistically prevalent as nominal and adjectival derivatives, and, at the same time, the semantic phenomena, patterns and constraints regulating their derivation are far less homogeneous than those of nominal and adjectival derivatives. The lack of homogeneous semantic/syntactic behaviour, we may add, is to a considerable extent connected to the fact that verbal evaluatives are always embedded in the aspect and Aktionsart system of a particular language, which makes generalisation even more difficult due to the aspectual diversity of languages (for more on aspectual typology and the derivational patterns found in aspectually different languages, see Dahl 1985). Below we give a short overview of verbs with the frequentative suffix which have an evaluative meaning. In Hungarian, the only productive frequentative suffix2 is -­gat/-­get, which can be attached to verbs regardless of their feature of (in)transitivity (e.g. (17)) or the number of syllables; however, the suffix is attached to monosyllabic verbs with an epenthetic vowel (e.g. (18a–b)). (17) nyíl-­ogat – nyit-­ogat open.intr-­freq – open.tr-­freq ‘open repeatedly (intransitive)’ – ‘open repeatedly (transitive)’

Hungarian 233 (18) a. fogy – fogy-­ogat lessen – lessen-­freq ‘lessen, diminish’ – ‘lessen, diminish gradually’ b. mond – mond-­ogat say – say-­freq ‘say’ – ‘say repeatedly’ As far as morphosemantics is concerned, the derivational suffix -­gat/-­get is usually associated with iterativity (see, for example, Kenesei, Vago and Fenyvesi 1998, 303‒4), i.e. it indicates that the event denoted by the base verb is repeated in a given temporal interval. There is, however, an aspectual restriction regarding the semantics of the output: the output predicate can only have a purely iterative meaning if the base verb expresses a punctual event. In these cases the derivatives do not qualify as evaluatives since they do not imply any evaluation. Compare (19a), (19b), (19c), etc. (19) a. nyit – nyit-­ogat open.tr – open.tr-­freq ‘open’ – ‘open repeatedly’ b. üt – üt-­öget hit – hit-­freq ‘hit’ – ‘hit repeatedly’ c. csókol – csókol-­gat kiss – kiss-­freq ‘kiss’ – ‘kiss repeatedly’ At the same time there are also quite a few semantically transparent lexicalised derivatives (e.g. (20a)) as well as fully lexicalised non-­transparent ones with iterative meaning (e.g. (20b–c)). (20) a. int – int-­eget wave – wave-­freq ‘wave’ – ‘wave repeatedly’ b. sim-­ogat pet-­freq ‘pet’ c. óbé-­gat groan-­freq ‘yammer’ Derivatives whose base is a durative (process or activity) verb normally do carry a pragmatic meaning while they also express iterativity. Typically, such verbs have an attenuative meaning, i.e. they indicate the lower, reduced intensity – or in some cases even the reduced quality – of the process/activity in question. Typical examples are listed in (21a–d). (21) a. olvas – olvas-­gat read – read-­freq ‘read’ – ‘read in a leisurely way’

234

Ferenc Kiefer and Boglárka Németh

b. dolgozik – dolgoz-­gat work – work-­freq ‘work’ – ‘work slowly/half-­ heartedly’ c. sétál – sétál-­gat walk – walk-­freq ‘walk’ – ‘ramble, saunter’ d. eszik – esz-­eget eat – eat-­freq ‘eat’ – ‘eat little from time to time’ This category of derivatives is not a homogeneous one either, as there is variation regarding the presence or the degree of pejorativity and iterativity. The attenuative meaning may allow for a pejorative interpretation due to extra-­linguistic factors, such as the significance and pregnancy of the activity in question. The pejorative meaning of the derivative is frequent in cases when the activity denoted by the base verb is a professional one (see (22a–b), for example). (22) a. [az író] ír-­ogat the writer write-­freq ‘[the writer] writes slowly/poorly (repeatedly)’ b. [az orvos] gyógyít-­gat the doctor cure-­freq ‘[the doctor] cures (patients) with varying results/success’ On the other hand, these derivatives may also express endearment or sympathy in some cases, e.g. when we use them with children or sick people (depending on the pragmatic context, this is a possible interpretation in the case of the typical examples listed in (21) above). The prominence of the iterative meaning is also variable in these examples; however, this is not a pragmatic phenomenon, as the variation is due to the semantic (aspectual) features of the base verb. Additionally, the attenuative meaning is present in these cases regardless of the degree of iterativity. We should only note here that the iterative meaning is more prominent in the output structure when the activity denoted by the base verb is a non-­ homogeneous one, i.e. it has a heterogeneous phase structure. Compare (23a) and (23b), where the latter expresses iterativity more prominently than the former, which is derived from a verb with a homogeneous phase structure. (23) a. olvas-­gat read-­freq ‘read in a leisurely way’ b. ír-­ogat write-­freq ‘write repeatedly (slowly/poorly)’ As far as stative verbs are concerned, they do not take the iterative suffix productively, as shown by the ill-­formed examples in (24a–d) from Kiefer and Ladányi (2000b, 200).

Hungarian 235 (24) a. birtokol – *birtokol-­gat own – own-­freq ‘own’ b. tartalmaz – *tartalmaz-­gat contain – contain-­freq ‘contain, include’ c. gyűlöl – *gyűlöl-­get hate – hate-­freq ‘hate’ d. emlékszik – *emlékez-­get remember – remember-­freq ‘remember’ Note, however, that – though the grammaticality judgements vary – the listed ill-­formed derivatives may occur in informal speech, very often with a humorous overtone. At the same time, there are some commonly accepted derivatives as well; see for example the derivatives in (25a–d). (25) a. van – vagy-­ogat/ ?van-­ogat be – be1-­freq be3-­freq ‘there is’ – ‘is not completely well/there is a few’ b.  tud – tud-­ogat know – know-­freq ‘know’ – ‘know partially’ c. hasonlít – hasonlít-­gat resemble – resemble-­freq ‘resemble’ – ‘resemble to some extent’ d. szeret – szeret-­get love – love-­freq ‘love to some extent’ – ‘love’ In such cases the derivation only introduces an attenuative meaning, and the derivatives do not indicate iterativity since the base verb does not contain repeatable events. According to the observations presented above, the derivatives with durative base verbs are clearly evaluatives, as they involve a qualitative specification of the event expressed by the base verb and indicate the speaker’s attitude to the event in question. The derivatives with punctual base verbs, on the other hand, are simply the iterative aspectual counterparts of their base verbs. As Kiefer and Ladányi (2000b, 199) point out, the iterative and the attenuative meaning are related in a way: the repetitive nature of the event expressed by the base verb may imply at the same time the lower intensity of the event and, consequently, the reduced intensity of the process/activity expressed by the derivative. The very few derivatives with purely attenuative meaning could be the results of analogical derivation. 2.4  The excessive Hungarian has an excessive which is formed from the superlative by repeating the superlative prefix -­leg. This entails that the existence of the superlative is a precondition for the derivation of the excessive, but not all superlative forms permit such a

Ferenc Kiefer and Boglárka Németh

236

derivation. Comparison is said to belong neither to prototypical inflection nor to prototypical d ­ erivation (Dressler 1989). One decisive argument for the non-­prototypicality of comparative morphology is the lack of the existence of lexicalised forms. Note that although there are a couple of lexicalised superlative forms, they have only an adverbial use: (26) a. leg-­alá-­bb sup-­under-­comp ‘at least’ b. leg-­fel-­jebb sup-­up-­comp ‘at most’ The superlative is derived from the comparative by means of the prefix leg-­; the corresponding comparative forms are alá-­bb ‘more down’ and fel-­jebb ‘more up’. Regular superlative forms never get lexicalised, and the same is true of the excessive.3 (27) a. jó – jo-­bb – leg-­jo-­bb – leg-­es-­leg-­jo-­bb good – good-­ comp – sup-­good-­comp – exc-­sup-­good-­comp ‘good’ – ‘better’ – ‘best’ – ‘best of all’ b. nagy – nagy-­obb – leg-­nagy-­obb – leg-­es-­leg-­nagy-­obb big – big-­ comp – sup-­big-­comp – exc-­sup-­big-­comp ‘big’ – ‘bigger’ – ‘biggest’ – ‘biggest of all’ Before discussing the semantic and/or pragmatic aspects of the excessive, some morphological peculiarities should be noted. First, Hungarian is basically a suffixing language, a property it shares with other Uralic languages. The development of the superlative explains its prefixal form: the element leg-­ was originally an emphatic particle, which became a regular part of the superlative form via cliticisation by the sixteenth century (Korompay 1992, 352–3). In the case of the excessive, the element leg-­ is not simply repeated but is connected with a conjunction particle whose original form is és ‘and’ with a long vowel, but here it appears as an enclitic element.4 In principle, the prefix leg can be iterated ad libitum, but more than two repetitions are rare. Example (28) is one in which the emphatic element leg occurs three times: (28)

leg-­es-­leg-­es-­leg-­nagy-­obb exc-­exc-­sup-­big-­comp ‘biggest of all’

The number of repetitions of the particle leg-­does not affect meaning. The superlative, as is well known, expresses the highest degree of a property along some scale of values. Semantically speaking, there is nothing better than the best, nothing greater than the greatest, nothing more remote than the most remote. There is thus no semantic difference between the superlative and the excessive. In other words, the meaning difference can only be pragmatic. The use of the excessive is more typical of spoken language, as can be shown indirectly by the data in Table 15.4.1, taken from the Hungarian National Corpus, which is based on written material.5

Hungarian 237 Table 15.4.1. Excessive forms in the Hungarian National Corpus Form of the excessive prefix

Number of hits

leg-­es-­leg leg-­es-­leg-­es-­leg leg-­es-­leg-­es-­leg-­es-­leg leg-­es-­leg-­es-­leg-­es-­leg-­es-­leg

818 29 6 0

In informal conversation, however, where the degree of a given property is at stake, there seems to be no limit on how often the excessive prefix is repeated. In particular, although we find no examples of more than four repetitions in our corpus, such cases do occur in informal speech and can be produced freely.

3  Evaluative morphology in Uralic Diminutive noun derivation occurs in all Uralic languages with the exception of Finnish, which behaves in this respect like the other Scandinavian languages, which lack morphological diminutives. In general, diminutive formation seems to be an areal phenomenon: diminutives in Estonian are formed largely according to the German pattern, Livonian follows the Latvian pattern, and the Eastern Finnic languages follow the Russian pattern. Hungarian diminutives exhibit similarities with both German and Slavic. Productive augmentative derivation is only known in Samoyedic. In Nenets, for example, the augmentative suffix is -­ɁjA (e.g. (29a)) and the diminutive suffix is -­ko (e.g. (29b)). (29) a. maɁ – maɁja tent – tent-­aug ‘tent’ – ‘big tent’ b. ñawa – ñawa-­ko hare – hare-­dim ‘hare’ – ‘small hare’

Notes 1. The choice is determined by vowel harmony. 2. Other, non-­productive Hungarian frequentative suffixes are: -­kod/-­ked/-­köd (e.g. csap ‘slap’, csap-­kod ‘slap repeatedly’; lép ‘step’, -­lép-­ked ‘step repeatedly’); -­gál/-­gél (e.g. szalad ‘run’, szalad-­gál ‘run around’, nevet ‘laugh’, nevet-­gél ‘giggle’); -­dos/-­des/-­dös (e.g. lök ‘push’, lök-­dös ‘push repeatedly’; ver ‘hit’, ver-­des ‘hit repeatedly’). 3. Morphological changes such as the shortening of the vowel in jó in (27) are fully predictable and need not concern us here. 4. It should be noted that if the excessive is pronounced slowly the conjunction may appear in its original form. 5. We wish to express our gratitude to Bálint Sass, who supplied us with these data. The Hungarian National Corpus contains 187.6 million running words and 17.5 million sentences.

15.5  Israeli Hebrew

Noam Faust

1 Introduction Israeli Hebrew is morphologically a Semitic language: it is characterised by its blend of concatenative and non-­concatenative, root-­and-­pattern morphology. Of the universal evaluative patterns, only diminutives and the analogous verbal pluractionals exist; other patterns. such as augmentatives, are not attested. Diminutivisation in Israeli Hebrew is a productive process on nouns, verbs and adjectives (although to a much greater degree on nouns). Besides the productive patterns, some unproductive patterns exist. In this chapter, I will survey both productive and unproductive patterns, and point out aspects of the phenomenon that are noteworthy either from a typological perspective or because they shed light on language-­internal issues. Diminutivisation patterns can be divided into two types: templatic diminutives, in which the elements of a base are placed within a prosodically prespecified template, and suffixal diminutives. Nominal diminutivisation is predominantly suffixal, adjectives show mixed behaviour, and verbs are exclusively templatic. All templatic diminutivisation involves reduplication (that is, non-­reduplicated templates are not used).

2  Diminutivisation in the nominal system There are two non-­productive templatic patterns that derive diminutives. They involve the reduplication of either the last radical (QaTLVL, as in (1a)) or the two last radicals (QTaLTVL, as in (1b)). In both cases, the first vowel is the default /a/ and the second either /i/ or /u/. If the last two radicals are reduplicated, as in (1b), the last vowel is the last vowel of the source noun (in bold). If that vowel is an unstressed [e] in the source noun, a stressed [a] appears in the diminutives.1 Both templates host also non-­diminutives (for instance šagrir ‘ambassador’, šrafraf ‘stool’): (1)

Templatic diminutives in the nominal system:2

Noun a. kómec ‘small quantity’ nékev ‘pore’ xameš ‘five’ šéver ‘fraction’

Diminutive kamcuc nakbuv-­it xamšuš šavrir

‘very small quantity’ ‘small pore’ ‘pupil of fifth grade (pejorative)’ ‘small fraction’

Israeli Hebrew 239 xélek dag b. xatul xazir zanav kélev géver

xelkik dagig xataltul xazarzir znavnav klavlav gvarvar

‘part’ ‘fish’ ‘cat’ ‘pig’ ‘tail’ ‘dog’ ‘man’

‘particle’ ‘small fish’ ‘kitten’ ‘piglet’ ‘small tail’ ‘puppy’ ‘macho’

Given the prosodic constraints of Hebrew (strong preference for disyllabicity, initial clusters allowed), one logical possibility is not attested, namely *QTaLiL. Although these two classes are small and not productive at all, formalising them illustrates the distinctive nature of templatic, discontinuous morphology. The proposed template (C)CaCCVlexC in (2) achieves two goals: it unifies the two attested diminutive templates and rules out the unattested one. The template is built in the following manner. There are five consonantal positions, the first of which is optional. The first vocalic position is lexically occupied by /a/. The quality of the vowel in Vlex is determined for each item individually. The number of empty C-­slots determines the number of radicals to be reduplicated. In (2a), the initial, optional consonantal position is not occupied, and therefore only one position needs to be filled by copying of the last radical. The lexical templatic vowel /a/ overrides that of the base. For this lexical item, Vlex is /i/. In (2b), the initial position is occupied: this results in the copying of the last two radicals in order to satisfy the template. The first vowel of the base noun does not survive (because of the initial cluster), but the lexical vowel of the noun is linked to the last position in the template: (2)

A unified prosodic template for diminutives:

a.

b.

š | (C)

C

| x

| t

a

e

a

u

v |

e

C

r | C

[šavrir] ‘small fraction’ V lex +

C

V

| l

[xataltul] ‘kitten’

We will return to the reduplication patterns in (2) in our discussion of pluractional verbs. Besides these rare diminutives, an extremely productive diminutive process is the suffixation of -­on to masculine nouns (3a–e) and -­ón-­et to feminine nouns (3f–k). The gender of the base noun thus remains unchanged in the diminutive. (3) Diminutive -­on/-­ónim: Noun a. séfer ‘book’ b. arnak ‘purse’

Diminutive sifr-­on arnak-­on

Noam Faust

240 c. xatul d. nasix e.  kúmzic f. sir-­a g. miškéf-­et h. yalda i. yad-­i-­t j. xan-­u-­t k. dilém-­a

‘cat’ ‘prince’ ‘social gathering around campfire’ ‘boat’ ‘binoculars’ ‘girl’ ‘handle’ ‘shop’ ‘dilemma’

xatul-­on nesix-­on kumzic-­on kumzic-­on-­im.pl cf. kúmzic-­im.pl sir-­ón-­et miškaf-­ón-­et yald-­ón-­et yad-­i-­ón-­et xan-­uy-­ón-­et dilem-­ón-­et cf. dilém-­ot.pl

Some noteworthy aspects of -­on follow.3 Firstly, it contrasts with a non-­diminutive suffix -­on (xid-­on ‘quiz’ < xid-­a ‘riddle’), to which it may in principle be suffixed (xid-­on-­on ‘quiz-­dim’). This suggests that the two -­on suffixes may be one and the same suffix, attached to different levels; but this view requires further exploration. A second interesting property is related to stress. Stress is (mostly) final and mobile in native Hebrew items (xatúl – xtul-­ím ‘cat(s)’), but lexical and stable on loanwords (kúmzic – kúmzic-­im ‘campfire(s)’). Loaned suffixes such as -­íst also carry stable stress, and thus attract stress from their base, but do not lose it to subsequent suffixation: for example šékem ‘army grocery store’, šekem-­ist ‘worker in the šékem’, šekem-­íst-­im.pl’. The suffix -­on exhibits mixed behaviour. Like a lexically stressed suffix, it attracts stress from its base even when the stress of the base is lexical (3e, k). However, when pluralised, the attracted stress moves on to the plural suffix, which, as we have seen,  does not usually attract stress itself. One may conclude that -­on defines a new ­final-­stress domain, but does not itself bear lexical stress.4 The feminine form raises another question: why is the feminine suffix -­et and not the other two possible feminine suffixes -­a or -­it? In Faust (2011), I argued that -­it, but not -­a or -­et, consistently and productively marks derivation of a feminine noun from a masculine one. The feminine diminutive does not fit that description, since its gender is determined by the base noun: either -­a or (unstressed) -­et, which should be regarded as allomorphs, is expected. The selection of one or the other is a very complicated matter; in the present context, one may say that the above-­mentioned metrical strength of -­on may favour the unstressed -­et allomorph. Finally, the suffix -­on may be attached to reduplicated diminutives: klavlav-­on ‘puppy-­dim’.5 Another productive (although to a much lesser extent) diminutive suffix is -­čik, a loan from Russian. It is a pre-­stressing suffix, for example baxúr-­čik (< baxur ‘young man’), xamór-­čik (< xamor ‘donkey’), but not *kélev-­čik. This suffix can be added onto the other diminutive formations: klavláv-­čik, klavlav-­ón-­čik.

3  Adjectival diminutives Templatic diminutives are considerably more frequent and numerous in adjectives. Unlike the form of nominal templatic diminutives, that of the adjectival ones is completely predictable from the form of the base adjective: if the base has three consonants, the diminutive is QTaLTaL (4a). If the base has two consonants, it is QaLiL. In the first option, the vowel of the base is always overridden by that of the output template.

Israeli Hebrew 241 (4)

Adjectival templatic diminutives:

Noun a. kaxol ‘blue’ yarok ‘green’ xamuc ‘sour’ šamen ‘fat’ katan ‘small’ b. xam ‘hot’ kal ‘easy, light’ kar ‘cold’

Diminutive kxalxal yerarak xamacmac šmanman ktantan xamim kalil karir

‘blue-­ish’ ‘green-­ish’ ‘slightly sour’ ‘chubby’ ‘very small’ ‘warm’ ‘very easy, very light’ ‘chilly’

All of the bases have the vowel [a] as their first (or only) vowel. However, putting aside participial forms, very few adjectives in Hebrew don’t have a first vowel [a]. By analogy with the nominal template QTaLTVL, one may say that the first vowel [a] of the diminutive template is the template’s own. The second vowel must be viewed as such, because it overrides the base’s.6 The lack of QaTLiL-­type adjectives may be explained by representing the first consonantal position as non-­optional: (5)

The adjectival diminutive template: a

o

k

x

l

|

|

|

C

C

a

C

[kxaxal] ‘blue-ish’ C

a

C

As for the template of (4b), it might be analysed either as suffixation of a skeletal reduplicant, or as some defective application of the general reduplicative template in (5). All of the colour names of the type QaToL (which covers all but one colour name, setting aside loans like bež ‘beige’) have QTaLTaL versions. However, beyond this very restricted domain, the adjectival diminutive template is far from being productive, and although Israeli Hebrew has a productive process of adjective formation using participial templates, these adjectives cannot be diminutivised in the QTaLTaL template. The nominal diminutive suffixes -­on and -­čik (to a lesser extent) can be applied to adjectives of the QaTVL template and the participial templates (6a, c), but not to adjectives formed concatenatively with the suffix -­i (6b, d). The resulting item, however, rarely has all the functions of the base adjectives (the distinction between nouns and adjectives is not clear-­cut in Israeli Hebrew): for example, it sounds odd when used in a modifying function, typical of the ‘normal’ adjective (6e). (6)

Suffixation to adjectives:

a. katan-­čik šamen-­čik mušlam-­čik b. xofš-­i ‘free’ *xofš-­i-­čik

< katan ‘small’ c. xamud-­on < xamud ‘cute’ < šamen ‘fat’ regiš-­on < ragiš ‘green-­ish’ < mušlam ‘perfect’ macxik-­on < macxik ‘funny’ < xófeš ‘freedom’ d. *xofš-­i-­on ‘easy, light’ e. ??xatul šamenčik ‘fat (dim.) cat’

Noam Faust

242

4  Pluractionality in verbs Diminutive morphology in verbs is understood as pluractionality, i.e. the breaking down of a single event into many repeated occurrences (see Tovena, this volume). Israeli Hebrew has two pluractional patterns, entirely dependent on the number of radicals in the (often hypothetical) base: the partially reduplicated QTLL (7a–g) and the fully reduplicated QTQT (7h–p). While the former very rarely accepts new items with this meaning, the latter is quite productive, even with loanwords as base (7n). According to Greenberg (2010), 70 per cent of the QTQT verbs do not have a base (7o); some such cases exist in QTLL, too, although their pluractionality is not clear (širbev ‘stick out.tr’, ši(ʕ)amem ‘bore’). No template QTLTL exists in the verbal system. (7)

Two pluractional patterns:7

Related item a. laxaš ‘whisper’ b. caxak ‘laugh’ c. kiven ‘tune, direct’ d. kafac ‘jump’ e. kadur ‘ball (n.)’ f. šalat ‘remote-­control (n.)’ g. faks ‘fax’ h. likek ‘lick’ i. mišes ‘feel (with hands)’ j. pocec ‘blow up’ k. nofef ‘wave’ l. hed ‘echo (n.)’ m. daf ‘page’ n. zaping ‘zapping’ o. -­ p. -­

Pluractional verb lixšes ‘whisper repeatedly’ cixkek ‘giggle’ kivnen ‘fine-­tune’ kifcec ‘jump around’ kidrer ‘dribble’ šiltet ‘change channels repeatedly’ fikses ‘fax (not repeatedly)’ liklek ‘lick repeatedly’ mišmeš ‘feel repeatedly’ picpec ‘crush in pieces’ nifnef ‘wave repeatedly’ hidhed ‘to echo’ difdef ‘turn pages’ zipzep ‘change channels repeatedly’ hivhev ‘to flicker’ fisfes ‘miss (target, train)’

As the last example of each set illustrates, these two patterns may host verbs that are not pluractional. This is more common in QTLL (7g), because of the tendency to maintain the syllabification of the verbalised base (Bat-­El 1994a). The data in (7), and especially (7c, h–k), are very important in the description of the verbal system in general. I will now explain why. Active verbs (such as all the verbs in 7) in Israeli Hebrew are classified into one of three groups, based on their prosody and vocalisation. The groups are QaTaL, QiTeL and hiQTiL (for instance, respectively xašav ‘think’, xišev ‘calculate’ and hexšiv ‘regard, take into consideration’). What determines which root or base will participate in which of these groups is a matter of debate. However, it is a well-­ known fact that quadriradical roots and bases with more than three consonants may only be mapped to QiTeL, for only this template is fit to accommodate four consonants.8 I will therefore mark it as QiT(T)eL in what follows. All pluractional verbs appear in this template. But is it because they are pluractional and reduplicated, or because this is the only template that allows reduplication? In other words, is the reduplication applied prior to the insertion in QiT(T)eL or is it the outcome of that insertion?

Israeli Hebrew 243 Classical autosegmental morphophonology clearly opts for the second of these: reduplication in this framework is a result of template satisfaction (McCarthy 1981). This is presented in (8) for both pluractional templates, for the verbs in (7c, h): the empty templatic slots are filled by copies of the appropriate radical. (8)

Pluractional reduplication as a result of template satisfaction: a. kivnen (7c)

b. liklek (7h)

k | C

l | C

v n | | i C C

e

C

i

k | C

C

e

C

In (8), pluractional verbs are reduplicated because they are mapped into QiT(T)eL, and not because they are pluractional. However, if reduplication indeed follows from template satisfaction, two predictions are made. Firstly, for a given root there must always be one output in a given template. Secondly, biradical roots may be fitted into triradical templates and yield pluractionals, because such templates may also allow for their reduplication. As (7) shows, both predictions are wrong. kiven and kivnen (7c) both belong to QiT(T) eL, but satisfy the template in two different manners. And no pluractional verbs appear in QaTaL and hiQTiL. For reasons such as these, Goldenberg (1994) claims that reduplication operates at the root level (√lk + red => √lklk); the root then appears in QiT(T)eL because it is quadriradical. Generalising from such cases, for Goldenberg templates are different arrangement of radicals, rather than skeletal moulds into which radicals are fitted. There are thus at least three verb types in what we have called QiT(T)eL: QT(T)L, QTLL and QTQT (for example, respectively kiven, kivnen and liklek).9 While the objections and the conclusions seem largely valid, there is at least one important problem with the view of pluractionals in this analysis: QTQT and QTLL are represented as two different radical templates, whereas in fact they are two versions of the same morphological pluractional construction, conditioned by the number of radicals of the base. Moreover, template satisfaction as a process does have its advantages, as evidenced by items like fikses (7g). I would now like to propose a compromise between the two approaches, which will both express the allomorphy in QTLL~QTQT and incorporate template satisfaction. Firstly, let us consider how a non-­reduplicated QiT(T)eL verb is to be formalised. In (9), a fourth, optional C-­slot is inserted after the second C-­slot of the base. Since this infixed slot is optional, reduplication will proceed only in biradical roots (9b): (9)

Regular QiT(T)eL by infixation of an optional C-­slot: no reduplication:

a. kiven (7c) k v | | C C (C)

n | C

b. likek (7h) l k | | C C (C)

C

Noam Faust

244

Another option for the biradical roots in the scenario in (9) is QoTeT (e.g. (7j, k) above). Since these are sometimes interchangeable (pocec~picec ‘blow up’), I assume that this is a lexical choice.10 Crucially, the verbs in (9) are not pluractional. The pluractional morpheme, conversely, is realised as the suffixation of a C-­slot. In this case, four slots must be filled in both cases. As in the adjectival and nominal cases, if the base is triradical, only the last radical is reduplicated (9a); if it is biradical, both must be filled. (10)

Pluractionality realised as a suffixed C-­slot: reduplication as template satisfaction:

a. kivnen (7c) k v n | | | C C C

+CPLA

b. liklek (7h) l k | | C C C

+CPLA

As in Goldenberg’s view, the difference in derivation is occluded by the fact that all four verbs in (9) and (10) appear with the same vocalisation. However, that is a consequence of their newly acquired quadriradical nature, and not some arbitrary property. This proposal has the advantage of unifying both pluractional templates and distinguishing between these templates on the one hand and the QiT(T)eL template on the other. This is achieved in a manner that expresses both Goldenberg’s insight into the relation between radical number and vocalisation and the notion of template satisfaction, so central in the framework of prosodic morphology.

5 Conclusion In this short survey, I have presented the evaluative morphophonology of Israeli Hebrew, which is at once templatic and linear. It seems that a clear preference for linearity exists outside the verbal system. As I hope to have shown, several questions are raised by both strategies, just as both have implications for the entire system.

Notes   1. This fact can be taken to indicate that that second vowel is in fact /a/, as suggested by Falk (1996) and Faust (2011; forthcoming). Another alternative is to designate the plural stem (for example gvarim ‘men’) of these nouns as the base (see Bat-­El 2012).   2. Phonological and representational remarks: in the native vocabulary, [xC] clusters are disallowed and separated by an epenthetic [a]. [b]~[v] are morphophonologically conditioned ­allophones. Final stress is not represented in the examples.   3. Bolozky (1994) is a thorough description of the distribution of -­on, most of which is ­unsurprising. Bat-­El (1997) discusses the intriguing fact that the plurals of feminine diminutives sound odd (??sir-­on-­ot ‘boats-­dim’, cf. sir-­a, sir-­ot ‘boat(s)’). According to Bat-­El, this is due to a ­conflict between the gender of -­on and the requirement of -­ot.   4. See Bat-­El (1993) and Mel’čuk and Podolsky (1996) for some insights into how this may be implemented.   5. See De Belder, Faust and Lampitelli (2014) for the theoretical consequences of this and related facts.

Israeli Hebrew 245   6. Another view might suggest that the second [a] corresponds to the first one of the base adjectives, which is a prerequisite for the input to reduplication. There is one exception to this generalisation: xalašluš < xaluš ‘feeble’   7. See Greenberg (2010) for a thorough examination of the semantics of the QTQT pattern.   8. This might be regarded as a structural property of QiTeL or simply as the result of its having two vowels throughout its inflection. Also, notice that a few quadripartite verbs exist in hiQTiL, probably because of the form of the base: špric ‘spritz’, hišpritz ‘to spritz’.   9. Goldenberg would add at least two other radical templates: simple quadriradicals (pi(ʕ)aneaħ ‘decipher’) and augmented quadriradicals (šidreg ‘rank upwards’, related to direg ‘rank’). It is also important to add that Goldenberg denies the synchronic status of biradical roots such as √lk. 10. Izre’el (2009) suggests that this may be a case of root allomorphy conditioned by the verbal type: √pc(c) becomes √poc(c) in QiT(T)eL.

15.6 Ket

Edward J. Vajda

1 Introduction Ket is spoken by a few dozen elderly people in isolated villages near the Yenisei river in central Siberia’s Turukhansk District. General descriptions can be found in Georg (2007), Werner (1997) and Vajda (2004). The language is unrelated to the surrounding Uralic, Turkic and Tungusic families, which have provided a modest number of lexical borrowings and exerted areal influence on the morphological structure (Vajda 2009). Ket is the sole survivor of the formerly widespread Yeniseian family, along with the extinct Yugh, Kott, Assan, Arin and Pumpokol (Werner 2005). Evaluative morphology is uncommon in Ket and apparently in Yeniseian overall, though documentation of the extinct languages is uneven. Modern Ket shows only three evaluative patterns in nouns and two in adjectives. Ideophones incorporated into finite verbs convey vivid impressions of sound  or visual stimuli, and also build intensive forms. The finite verb prefixes called ‘­determiners’ occasionally convey intensity of action. Otherwise, Ket expresses size, intensity and other qualities on the phrasal level (sēˑl ke’t ‘bad man’, aqta dɯ̄ˑl ‘good child’, həna tīˑp ‘little dog’, qà oln ‘big nose’, qaddoq sēˑl ‘very bad’) or by lexical choice (uros ‘drunkard’, ləŋtəvəl ‘slob’, belɯm ‘loose woman’). As in all languages, the speaker’s subjective evaluation of the message involves deployment of emotively charged intonation, particles and interjections. The rarity of evaluative morphology may be connected with the general paucity of derivational affixes in Ket.

2  Morphological overview Ket morphology is strongly head marking with elaborate inflectional paradigms, but relatively little variety in derivational affixes (Vajda 2007, 1323–4). The nominalising suffix -­s derives nouns from all other parts of speech, including conjugated finite verb forms. Most other polymorphemic noun stems are built through root compounding. Ket  affixes that derive nouns from other nouns usually show obvious evidence of having evolved from independent noun roots, most of which still function as ­independent nouns in the language (Werner 1998, 49–58). The language’s few purely derivational affixes (that is, affixes not obviously etymologically related to roots) tend to change the form class, deriving nouns from other parts of speech, or adjectives from nouns. The examples below illustrates the most common types of Ket word forms:

Ket 247 Monomorphemic basic nouns and adjectives:1 tɯ’s tīˑk qo’n ke’t qu’s sīˑn ‘stone’ ‘snow’ ‘conifer’ ‘man’ ‘tent’ ‘old’

(2)

Compound nouns: bej-­oks ul-­tik wind-­pole water-­snow ‘mast’ ‘slush’



(1)

(3)

Derived adjectives: tɯs-­tu bej-­an rock-­with wind-­without ‘rocky’ ‘windless’

(4)

Finite verb forms: d-­b-­il-­a 1sbj-­3inan.obj-­eat ‘I ate it’





kəj-­ket hunting-­man ‘hunter’

ki’ ‘new’

kə’n ‘light, bright’

qon-­nus conifer-­tent ‘hut made from spruce branches’

kət-­tu children-­with ‘having many children’

kəd-­an children-­without ‘childless’

ku-­b-­a-­n 2sbj-­3inan.obj-­eat-­pl.sbj ‘you (pl.) eat it’

(5) Action nominals (originally derived from verb roots using suffix *-­əŋ): il-­iŋ bə̀ɣ (< *bəg-­əŋ) eat-­ act.nom find-­act.nom ‘eating’, ‘eaten’, ‘to eat’ ‘finding’, ‘found’, ‘to find’ (6)



Nominalisations made from modifiers or finite verb forms: tɯs-­tu-­s bəɣ-­s du-­b-­a-­n-­s rock-­with-­nmls find-­nmls 3m.cl.sbj-­3inan.obj-­eat-­pl.sbj-­nmls ‘something rocky’ ‘foundling’, ‘finder’ ‘the ones who are eating it’

As can be seen from these examples, the majority of Ket morphemes are lexical roots or grammatical affixes. Form-­class changing affixes such as the nominaliser -­s are more grammatical than lexical in function. For example, the adjective kə’n ‘light, bright’, when nominalised as kəns, can mean ‘something bright’, ‘someone with light skin’ or simply ‘brightness’. Action nominals that take -­s automatically acquire the function of nouns expressing any of the corresponding finite verb’s core valents. The nominalisation tàrs (made from the action nominal *tàr ‘beat’) can mean ‘one who beats’, ‘one who is beaten’ or ‘instrument used for beating’, depending on context. Most other stem-­deriving morphemes are compounded nominal roots that have acquired a more generalised function. The root -­qus (< qu’s ‘birchbark tent’) has become a suffix-­like final combining form that derives words for types of dwellings: iŋgus ‘house’ (< *ken ‘wood’), qonnus ‘lean-­to built with conifer branches’ (< qo’n ‘conifer branches or needles’), baŋgus ‘semi-­subterranean dwelling’ (< *ba’ŋ ‘earth’), etc. Some roots, such as hàŋ ‘female animal’ and ш’k ‘male animal’, have become initial combining forms: haŋsel ‘female reindeer’, haŋgaj ‘female moose’, шktip ‘male dog’, шgbən ‘male duck’. Werner (1998, 53) calls these root-­like affixes Halbaffixe (‘semi-­affixes’) to capture their intermediate status between ­compounded root and derivational affix. On a grammatical level, noun stems combine with pronominal possessive prefixes (actually special clitics) as well as with case suffixes and postpositions. Case suffixes must occur

Edward J. Vajda

248

word-­finally, while postpositions are connected to the preceding noun with possessive morphology and generally must be followed by a case suffix. Many postpositions clearly arose from possessed noun roots, while the origin of most case suffixes is unclear. The complex form in (7) illustrates the typical pattern of suffixal agglutination that has developed in Ket, probably under the influence of the surrounding, unrelated suffixal agglutinating languages: (7)

Nominal form with possessive proclitic and postposed relational morphemes: da-­hɯb-­aŋ-­na-­bal-­di-­ŋal 3f.cl.poss-­son-­pl-­poss.pl-­between-­inan.poss-­abl ‘out from between her sons’

In (7) the postposition -­bal-­ ‘between’ is etymologically related to the free noun ba’l ‘gap’. It is attached to the preceding noun by possessive morphology and in turn is followed by a case suffix expressing a more basic spatial relation such as ‘location’, ‘movement towards’, ‘movement past’ or ‘movement from’. Finally, as will be discussed further in Section 5, the Ket finite verb is strongly ­prefixing and rigidly templatic, with lexical and grammatical morphemes intertwined in linearly complex patterns. The language’s primary form class division is tripartite, with ­morphological distinctions sharply differentiating nouns, finite verbs and modifiers. Within the general class of modifiers, adjectives and adverbs are weakly differentiated morphologically (Krjukova 2005).

3  Evaluative noun morphology The relative lack of purely derivational affixes may have a causal connection with the rarity of evaluative morphology in Yeniseian languages. A few of the ‘semi-­affixes’ have taken on a marginal evaluative function in Ket. The diminutive suffix -­kit (< *gə’d ‘offspring’) productively derives nouns that denote juvenile animals, birds, fish or young plants: dūˑm ‘bird’ > dumgit ‘baby bird’, be’s ‘rabbit’ > beskit ‘baby rabbit’, to’t ‘taimen (fish)’ > totkit ‘immature taimen’, tēˑm ‘goose’ > temgit ‘gosling’, dɯ̄ˑn ‘fir tree’ > dɯngit ‘fir sapling’, etc. This suffix is primarily used to denote immature members of a species, though occasionally it serves as a hypocoristic expressing affection towards a mature specimen of the same species. Traditional Ket personal names are sometimes based on this sort of usage. One example is the name of one of our adult male informants, Pavel Sutlin, whose Ket name Dumgit translates as ‘Little Bird’. Similarly, the singulative suffix -­dis (< dēˑs ‘eye’) derives count nouns that objectively denote a small, individuated portion of a larger mass: eldis ‘single berry’, dandis ‘single blade of grass’, qondis ‘glass bead’, suldis ‘droplet of blood’, təŋdis ‘strand of (head) hair’ (< eːl ‘berries’, daːn ‘grass’, qōˑn ‘beads’, sūˑl ‘blood’, tə̄ˑŋ ‘head hair’); occasionally, it may also be used connotatively to emphasise smallness. The compound noun dapqul ‘pile, load’, which is derived with metathesis from the roots *daq ‘put’ and *pəl ‘growth, mass’, can be added after a mass noun to denote a load or portion: tikdaqpul ‘snow pile’, isdapqul ‘pile of fish’, etc. In Southern Ket, the postposed form daqpul has also been attested in a clearly evaluative function to convey an emotionally charged description of an unusually large-­sized body part. Only a few such epithets were found, however. One example is olndapqul (< oln ‘nose’ + dapqul ‘pile’), said wryly either of someone with an oversized nose or of someone who is very smart (the hunter-­ gatherer Ket believed that the organ of smell, not the brain, was the seat of intelligence and

Ket 249 reason). Another is kəjadapqul (< kəjga ~ kəja ‘head’ + dapqul ‘pile’), a word used to tease children whose head is large in proportion to the body. The noun həltisdapqul (< həltis ‘belly’ + dapqul ‘pile’) either means ‘pot belly’ (distended abdomen) or is used to ridicule an overweight person. Other evaluative suffixes (or compounded roots) may once have existed in Ket but went unrecorded.2 Nevertheless, it appears reasonable to conclude that evaluative morphology is even less characteristic of Ket than is derivational affixation.

4 Adjective morphology Ket has only a few denominal adjective suffixes, none of which are evaluative: kul-­tu ‘bearded’ (< kūˑl ‘beard’), diktu ‘resinous, gummy’ (< dīˑk ‘conifer sap’), kul-­an ‘beardless’, etc. Only one suffix modifies adjectives without changing their form class. The attenuative suffix -­la is added to adjectives to express a lesser degree of the given quality and seems to have an inherently pejorative connotation. It combines freely with adjectives that express inherently negative qualities: sel-­la ‘rather bad’; but not with adjectives such as ‘good’ that are logically incompatible with pejoration. Adding -­la to adjectives that form gradable antonymic pairs, such as həna ‘small’ and qà ‘big’, may add a pejorative connotation: həna-­la ‘rather small’ (in the sense of ‘a bit too small’), qa-­la ‘a bit big’. When added to adjectives denoting qualities such as colours that do not pair as gradable antonyms, -­la seems to be purely attenuative with no negative connotation: sulem-­la ‘reddish’, taɣɯm-­la ‘whitish’, tum-­la ‘blackish’. Though more research is warranted into how attentuative -­la is used in context, it would seem that it functions much like the Russian evaluative adjectival suffix -­ovat-­ in such words as krasnovatɨj ‘reddish’ (neutral attenuative), malovato ‘too little’ (pejorative attenuative), ploxovatɨj ‘rather bad’ (attenuated pejorative), as described in Vajda (1987, 182–3) One predicative adjective form has been attested in an evaluative form when pronounced under enthusiastic intonation. Example (8) shows the neutral way of expressing the quality, while (9) shows the same expression with heightened positive emotion: (8)

Na’n aqta-­m. bread good-­ inan.pred The bread is good (tastes good).

(9)

Na’n aqt-­iːmeː! bread good-­ eval.inan.pred The bread tastes soooo good!

It is unclear whether an evaluative morpheme has augmented the inanimate-­class predicate agreement suffix form in (9) or whether pronouncing the neutral suffix -­m as -­iːmeː is simply part of the expressive intonation. This type of lengthening has not been observed under emotive intonation with other predicate agreement suffixes: aqta-­di ‘I am good’, aqta-­gu ‘you (sg.) are good’, aqta-­du ‘he is good’, etc.

5  Finite verb structure and evaluation Ket finite verb structure consists of eight prefixal slots, six of which express subject/object agreement, tense or mood. Interdigitated between these grammatical morphemes are two

Edward J. Vajda

250

lexical prefix slots. P5 contains prefixes called ‘determiners’ that are often difficult to etymologise. The other lexical prefix slot, P7, contains an incorporated noun, adjective, directional adverb or action nominal. Verb prefixes attach to a base (a verb root, sometimes with fossilised affixes) that historically expressed the verb form’s basic lexical meaning. In most productive stem types today, however, an incorporated action nominal in P7 expresses the verb’s basic lexical meaning (Vajda 2009, 488–91). There is one suffix, which marks animate plural subject agreement: (10)

Slots in the Ket finite verb template, ordered leftmost to rightmost: P8 – subject person agreement P7 – incorporated noun, adjective, adverb or action nominal P6 – object (sometimes subject) agreement P5 – determiner (lexical morpheme) P4 – tense/mood P3 – inanimate-­class subject or object agreement P2 – tense/mood/aspect P1 – subject (or object) agreement P0 – verb base P-­1 – animate-­class plural subject agreement suffix

The two lexical prefix slots may contain morphemes involved in evaluation. The determiners in position 5, though present in all grammatical forms of a given stem, typically lack any meaning that can be separated from the meaning of the stem as a whole. For example, the verb stem meaning ‘subject sees object’ invariably contains the determiner t, which is impossible to etymologise, since no other verb stems containing the base -­oŋ exist. (11)

d-­a-­t-­ol-­oŋ 1sbj-­3m.cl.obj-­?-­pst-­see ‘she saw him’

Determiners sometimes alternate to express differences in the shape, position or moving trajectory of the stem’s transitive object or intransitive subject (Vajda 2004, 62–6). In a few instances, determiners convey repetition or intensity, as when h (areal prefix) is replaced by t to express heightened intensity. The determiner t in (13) probably arose when the ­determiner *d ‘repetition, intensity’ concatenated with the original determiner prefix h. (12)



Neutral intensity (determiner h): daesaɣulij da-­es-­a-­h-­ol-­ij 3f.cl.sbj-­into.the.open-­3m.cl.obj-­area-­pst-­speak ‘she was calling him’

(13) Heightened intensity (determiner t < h + d): daesatolij da-­es-­a-­t-­ol-­ij 3f.cl.sbj-­into.the.open-­3m.cl.obj-­area-­intense-­pst-­speak ‘she was shouting for him’

Ket 251 The determiner alternation in (12) and (13) appears to be a vestige of a now unproductive pattern, though a few other verbs add d (often becoming t in combination with another determiner) to denote repetition. The second evaluative feature of the verb morphology is quite productive and involves ideophones incorporated into P7. Ket has a large number of ideophones, created by adding the suffix -­ej ~ -­aj ~ -­oj to an onomatopoeic base or to a noun used expressively to denote a characteristic sound, shape, colour or movement. Examples (14) and (15) illustrate these forms. (14) Ideophones derived from nouns: sulej ‘turning red, blushing, reddening (of people, berries, leaves)’ (< sūˑl ‘blood’) luɣaj ‘hacking sound (of wood being chopped)’ (< lūˑɣ ‘wood chips’) kutólej ‘whistling’ (‘mouth singing’ < kūˑ ‘mouth’ + d ‘possessive’ + il ‘song’) Onomatopoeic ideophones often contain phonological traits not usually found in other native words, such as word-­initial nasals or clusters of [r] and another consonant. (15) Ideophones derived on the basis of onomatopoeia: hoɣoj ‘sound of rushing wind or roaring fire’ karsej ‘scraping on the surface of ice or snow’ kɯmarej ‘thunder (in the sky)’ lilɢej ‘crunch (of sled runners on the snow)’ nəːɣaj ‘whining, complaining’ (said of people) moraj ‘crunching in the mouth (of something dry)’ siʁej ‘chirping (of birds), squeaking (of mice)’ Many ideophones build an augmented stem denoting intense or persistent repetition byreplacing the suffix -­ej with -­tijiŋ (for noun bases) and -­tiŋej ~ -­tiŋtej (for ­onomatopoetc  bases): sultijiŋ ‘keep turning red’, ‘blushing many times’, ‘reddening of many objects’. Some ideophones have only been recorded in the intensive form: muktiŋej ‘whining, whimpering (of dogs)’, boltiŋej ‘splashing (of objects falling into water)’. Ideophones can be used as attributive modifiers: sulej eːl ‘ripening lingonberries’, sultijiŋ eːl ‘lots of ripening lingonberries’, kutólej ke’t ‘whistling man’, muktiŋej tīˑp ‘whining dog’. Productive finite verb formulas incorporate ideophones, as well. Aside from the ideophone itself, the patterns in (16) and (17) contain stem morphemes that are opaque semantically and resist glossing. (16) Expressive processual stem: IDEOPHONE-­b-­(il)-­e-­ta ‘sounds can be heard’ karej-­b-­e-­ta ‘clawing sounds can be heard’ karej-­b-­il-­e-­ta ‘clawing sounds could be heard’ (17) Expressive single action verb stem (semelfactive): IDEOPHONE-­ges ‘one sudden sound is/was heard’ (identical form in present or past tense) karej-­ges ‘a sudden scratch is/was heard (single sound)’

252

Edward J. Vajda

Both patterns are attested with dozens of incorporated ideophones, and all of the resultant forms are vividly expressive in the same way. Whether used as attributive modifiers or incorporated into finite verbs, Ket ideophones express the speaker’s subjective experience of the perceptual vividness of an event and in this way relate to evaluative morphology. The suffixes -­tijiŋ and -­tiŋej ~ -­tiŋtej used to build intensive ideophones are also peripherally evaluative.

6 Conclusion Ket has only a few evaluative morphological patterns, a fact probably related to the relatively minor role played by derivation in the language. Vajda (2007, 1323–4) noted that Ket morphology consists overwhelmingly of lexical roots and grammatical inflections and therefore does not fit into the classic typology of isolating, analytic and synthetic languages. Ket grammatical morphology is highly synthetic, while stem formation depends heavily on root compounding. Halbaffixe or ‘semi-­affixes’ predominate in noun-­from-­noun derivation. All productive verb stem creation patterns require an incorporate in the P7 slot. Some stems incorporate nouns, adjectives or adverbs, while others insert an action nominal into the P7 slot that expresses the verb’s basic lexical meaning. What little evaluative morphology does exist in Ket tends to follow these predominant patterns of root compounding and incorporation, though a few genuine evaluative suffixes are marginally used in the language.

Notes 1. Ket phonemic prosody is transcribed here as follows: v̄ˑ -­ high-­even tone on a half-­long vowel; v’ -­ abrupt rising tone ending in a laryngeal phase or optionally with full glottal closure; vː -­ rising-­ falling tone on a geminate vowel; v̀ falling tone. The symbol /ə/ represents a mid-­back rather than central unrounded vowel, realised as mid-­high [ɤ] under high-­even tone and as mid-­low [ʌ] elsewhere. The symbol /ɯ/ transcribes a high back unrounded vowel. 2. The examples of evaluative noun morphology in this section are from native Southern Ket consultant Valentina Romanenkova, who was interviewed extensively in the hopes of finding additional patterns.

15.7 Latvian

Andra Kalnača

1 Introduction In Latvian, evaluative morphology is mainly realised by means of word formation – such semantic values as SMALL/BIG and GOOD/BAD are expressed within the derivational paradigms of nouns, adjectives and verbs as diminutive/augmentative and endearing/­pejorative meanings. This is achieved by suffixation, by prefixation and also by combinations of prefixal and postfixal elements, i.e. circumfixes which attach to verbs. A similar system of evaluative morphology is found in the other Baltic language, Lithuanian (Ambrazas 1996; Urbutis 2009). The diminutive/augmentative opposition is only marginally present in the derivational system of Latvian, as neither nouns nor adjectives have true augmentative derivational affixes. This means that a derivational opposition allowing either a diminutive or an augmentative suffix to be attached to a noun is not something normally found in Latvian. At the same time there is an intricate system for expressing diminutive meanings in noun and adjective formation (see Grandi 2011b for a diachronic perspective). Nouns sometimes acquire augmentative meanings through the semantic extension MANY/MUCH > BIG. In verb formation, diminutive and augmentative meanings are intertwined with aspect (Stump 1993, 13–17; Grandi 2009) as various degrees of completeness of an action are inextricably linked to perfective aspectuality. Moreover, such meanings can be expressed in different ways: through traditional prefixation, through suffixation and even by means of circumfixes consisting of a suffix and a reflexive ending. The semantic complex GOOD/BAD in Latvian, as in many other languages, has close ties with diminutives/augmentatives (Grandi 2005). Thus diminutive nominal suffixes sometimes function as manifestations of the GOOD/BAD opposition. Contextual usage is of crucial importance here as one and the same diminutive suffix can be used to convey either positive or negative evaluation. Diminutives of feminine personal names and words denoting females sometimes acquire additional GOOD/BAD meaning when a feminine ending is replaced with a masculine ending (see Section 2.1). In Latvian, the diminutives of adjectives are generally unrelated to the GOOD/BAD distinction, while augmentative verbal circumfixes which are used to express excessive actions do imply negative evaluation. Due to space limitations the remainder of this chapter is organised as follows: descriptions of diminutive/augmentative nouns, adjectives and verbs are provided alongside commentaries on their relation to the GOOD/BAD distinction.

Andra Kalnača

254

2 Nouns 2.1 Diminutives In Latvian, diminutive nouns can be formed from nouns by means of six suffixes to indicate small size or the quality of being smaller than the referent of the base word (MLLVG 1959, 174; Kalme and Smiltniece 2001, 67–74): (1) Suffixes -­iņš (m), -­iņa (f) roka ‘hand’ -­ītis (m), -­īte (f) lācis ‘bear’ -­ulis (m), -­ule (f) vārn-­a ‘crow’ -­uks (m) suns ‘dog’ -­ēns (m) putns ‘bird’ -­elis (m), -­ele (f) roka ‘hand’

Diminutives roc-­iņ-­a hand-­ dim-­f.sg ‘little hand’ lāc-­īt-­is bear-­ dim-­m.sg ‘little bear’ vārn-­ul-­is crow-­ dim-­m.sg ‘little crow’ suņ-­uk-­s dog-­ dim-­m.sg ‘doggie’ putn-­ēn-­s bird-­ dim-­m.sg ‘birdling’ roķ-­el-­e hand-­ dim-­f.sg ‘little hand’

While the number of diminutive suffixes is considerably greater in Lithuanian (Ambrazas 1996, 87–94), diminutives in Latvian possibly exhibit more multifunctionality (Rūķe-­ Draviņa 1959). In most cases, the applicability of diminutive suffixes in Latvian depends on the property of animateness (see Grandi 2011b, 15 for typological perspective). Only the suffixes -­iņš, -­ītis and -­elis can be used to denote both animate and inanimate objects; all the rest (-­ulis, -­uks, -­ēns) normally attach to nouns denoting animate beings – humans, animals, birds, fish. All diminutive suffixes are also used to express meanings associated with the GOOD/ BAD distinction (MLLVG 1959). The most productive suffixes are -­iņš and -­ītis, as they can attach to practically any noun irrespective of its lexical meaning to form either masculine or feminine nouns as the gender of the base word requires. Their distribution, however, depends on declension (Mathiassen 1997, 54; Kalme and Smiltniece 2001, 69–70), -­īt-­ attaches to nouns ending in -­is and -­e in the singular nominative, while -­iņ-­ applies to all other declension classes (Mathiassen 1997, 42–50): (2) a. balodis ‘pigeon’ egle ‘spruce’ b. koks ‘tree’ tirgus ‘market’

balodītis (m) ‘little pigeon’ eglīte (f) ‘small spruce’ kociņš (m) ‘small tree’ tirdziņš (m) ‘small market, fair’

Latvian 255



māja ‘house’ kūts ‘cowshed’

mājiņa (f) ‘little house’ kūtiņa (f) ‘small cowshed’

Both suffixes typically express endearment (MLLVG 1959, 112–19, 130–1) through the semantic extension SMALL > ENDEARING > GOOD. However, converse (SMALL > WORTHLESS > BAD) transformations are likewise possible and are used to express irony, about something that is not rated very highly, by attaching a diminutive suffix to a noun not commonly used in its diminutive form (MLLVG 1959, 117–18): (3) deputāts (m) ‘parliamentarian’ universitāte (f) ‘university’

deputāt-­iņ-­š parliamentarian-­ dim-­m.sg ‘petty parliamentarian’ universitāt-­īt-­e university-­ dim-­f.sg ‘mediocre university’

Although the suffix -­elis can likewise be used to form both masculine and feminine diminutives, in standard Latvian it is typically associated with pejorative meaning (SMALL > BAD) (Kalme and Smiltniece 2001, 67): (4) zirgs (m) ‘horse’ māja (f) ‘house’

zirģ-­el-­is horse-­ dim-­m.sg ‘plain-­looking horse’ māj-­el-­e house-­ dim-­f.sg ‘poorly built house’

Connotationally neutral diminutives can be found in dialects and in some isolated examples such as maiss ‘sack’ – maišelis (m) ‘little sack’. Diminutives derived by means of the suffix -­ulis are almost always masculine regardless of the gender of the base word: (5) ezis (m) ‘hedgehog’ čūska (f) ‘snake’

ež-­ul-­is hedgehog-­ dim-­m.sg ‘little hedgehog’ čūsk-­ul-­is snake-­ dim-­m.sg ‘little snake’

Diminutives of feminine personal names can take both feminine and masculine endings – Ieva (f) – Ievule (f), Ievulis (M). The latter function as terms of special endearment through the semantic extension SMALL > GOOD (Kalnača 2011, 82–8). The suffix -­ēn-­is used to mark diminutives and offspring of animals, birds and fish. The derivative is always masculine regardless of the base word: (6) kaķis (m) ‘cat’

kaķ-­ēn-­s cat-­ dim-­m.sg ‘kitten’

Andra Kalnača

256 pīle (f) ‘duck’ līdak-­a (f) ‘pike’

pīl-­ēn-­s duck-­ dim-­m.sg ‘duckling’ līdac-­ēn-­s pike-­ dim-­m.sg ‘pickerel’

When attached to personal names the suffix expresses particular affection, i.e. SMALL > GOOD (Kārlis – Kārlēns). Female personal names and nouns denoting females usually become masculine – Ieva (f) – Ievēns (m), meita (f) ‘girl’ – meitēns (m) ‘girlie’. Similarly, diminutives formed by means of the suffix -­uk-­ are usually masculine regardless of the original gender: (7) lāc-­is (m) ‘bear’ meit-­a (f) ‘daughter’

lāč-­uk-­s bear-­ dim-­m.sg ‘beary’ meit-­uk-­s daughter-­ dim-­m.sg ‘darling daughter’

These diminutives mark a highly positive attitude or affection on the part of the author (Kalme and Smiltniece 2001, 74), i.e. manifest the semantic value GOOD. Finally, double diminutives involving combinations of suffixes are also possible in Latvian with the function SMALL > GOOD: (8) Suffixes -­ul-­ + -­ēn-­ varde ‘frog’ -­ēn-­ + -­iņ-­ putns ‘bird’ -­uk-­ + -­iņ-­ suns ‘dog’

Diminutives vard-­ul-­ēn-­s frog-­ dim-­dim-­m.sg ‘froglet’ putn-­ēn-­iņ-­š bird-­ dim-­dim-­m.sg ‘little birdling’ suņ-­uk-­iņ-­š dog-­ dim-­dim-­m.sg ‘little doggie’

2.2 Augmentatives Although there are no true augmentative suffixes in Latvian, the augmentative function is realised through the meaning extension MANY/MUCH > BIG. While several denominal and deverbal noun-­formation types bear a rather remote resemblance to augmentatives, none of the suffixes have denoting large quantities as their primary function. Moreover, there is nothing like a single system of base words and diminutive suffixes used to realise the derivational meaning ‘a small or a big object’. Therefore, in Latvian augmentatives can only be postulated as part of the word-­formation system with grave reservations.

Latvian 257 2.2.1  Denominal type: collective nouns All denominal collective nouns of this group denote inanimate objects and share the meaning ‘a place where there is a large quantity or number of something’; thus augmentative meaning is clearly related to the concept of location – MANY/MUCH > BIG. The set of suffixes available for this purpose (MLLVG 1959, 86, 93, 134) includes -­ājs, used to form masculine nouns regardless of the original gender, and -­iena, -­aine, used to form feminine nouns: (9) a. ledus (m) ‘ice’ – ledājs (m) ‘glacier’ egle (f) ‘spruce’ – eglājs (m) ‘spruce grove’ b. kalns (m) ‘hill’ – kalniena (also kalniene) (f) ‘highland’ priede (f) ‘pine’ – priediena (also priediene) (f) ‘pine grove’ c. krūms (m) ‘bush’ – krūmaine (f) ‘bush as a mass noun’ smilts (f) ‘sand’ – smiltaine (f) ‘sands’ Some base words can take any of the three suffixes to produce semantically identical nouns (priede ‘pine’ – priedājs, priediena, priedaine ‘pine grove’). In such cases the suffixes -­ājs, -­iena (-­iene), -­aine are in fact derivational synonyms expressing the meaning ‘a place where there is a large quantity or number of something’. 2.2.2  Deverbal type: nomina agentis Within this word-­formation type, augmentative meaning is inseparable from aspectuality. Agent nouns that refer to actions performed to a very high degree are primarily derived from perfective verbs already containing the prefix iz-­, which marks actions performed completely or in large quantities: ēst ‘to eat’ – izēst ‘to eat up’, dzert ‘to drink’ – izdzert ‘to drink up’. The suffix -­ājs (m), -­āja (f) attached to such verbs produces nomina agentis that denote persons who do something a lot (MLLVG 1959, 153–4): (10) izēst ‘to eat up’ – izēdājs ‘the one who eats up’ izdzert ‘to drink up’ – izdzērājs ‘the one who drinks up’. Such derivatives may express negative evaluation of the agent through the BIG > BAD transformation. As already pointed out at the beginning of Section 2.2, examples (9)–(10) are quite peripheral to evaluative morphology and should not be viewed as real augmentatives. Although they probably have an augmentative nuance, their primary meaning is to denote places and agents.

3 Adjectives While Latvian lacks nominal and adjectival augmentative derivational suffixes, adjectives do have an advanced paradigm of diminutives that are realised as SMALL > SMALLER or SMALL > WEAKER and can be used to scale attributes of objects against other objects. The meaning SMALL > WEAKER is applied extensively to colour names, sensations, physical and mental attributes possessed by persons or objects to a slight extent, and time and location meanings. There are comparatively few diminutives with the meaning SMALL > SMALLER and they are usually used to express physically measurable attributes of objects.

Andra Kalnača

258

All diminutive adjectives are derived from adjectives by suffixation or prefixation and can be used in masculine or feminine forms depending on the noun. Therefore in the remainder of this chapter all adjectives are listed with the masculine endings -­s, -­š which correspond to -­a in feminine forms. Unlike nominal diminutive suffixes, adjectival suffixes do not mark attitude and hence are unrelated to the GOOD/BAD distinction. 3.1  Suffixal diminutives 3.1.1  SMALL > WEAKER The suffixes -­gans (m), -­gana (f) and -­īgs (m), -­īga (f) are used to grade colours from more to less distinct: zils ‘blue’ – zilgans ‘bluish’ (see Kalme and Smiltniece 2001, 123–4). Sometimes both suffixes are used in combination. When attached to adjectives derived by means of the suffix -­gans, the suffix -­īgs indicates a very weak presence of colour: zaļš ‘green’ – zaļgans ‘greenish’ – zaļganīgs ‘slightly greenish’. Double diminutives can also be formed by attaching the suffix -­(sn)ējs (m), -­(sn)ēja (f) after the suffix -­īg-­ to indicate a very slight tinge: brūns ‘brown’ – brūnīgs ‘brownish’ – brūnīg(sn)ējs ‘slightly brownish’. The diminutive suffixes -­ens (m), -­ena (f), -­ējs (m), -­ēja (f), -­īns (m) and -­īna (f) are typically used to mark attributes possessed by persons or objects to a slight extent, as well as to moderate time and location meanings (Kalme and Smiltniece 2001, 123–4): (11) gurds ‘weary’ salds ‘sweet’ tuvs ‘near’ agrs ‘early’

gurd-­en-­s weary-­ dim-­m.sg ‘slightly weary’ sald-­en-­s sweet-­ dim-­m.sg ‘sweetish’ tuv-­ēj-­s near-­ dim-­m.sg ‘relatively near’ agr-­īn-­s early-­ dim-­m.sg ‘early as compared to others’

Some attenuative adjectives are derivational synonyms; for example, tuvs ‘near’ – tuvējs, tuvīns ‘relatively near’. 3.1.2  SMALL > SMALLER In Latvian, this type is represented by the suffix -­iņš (m), -­iņa (f) (Kalme and Smiltniece 2001, 124); for example: (12) mazs ‘little’ labs ‘good’

maz-­iņ-­š little-­ dim-­m.sg ‘teeny’ lab-­iņ-­š good-­ dim-­m.sg ‘goody’

Latvian 259 3.2  Prefixal diminutives 3.2.1  SMALL > WEAKER Prefixal diminutives are most commonly formed by means of the prefixes pa-­ and ie-­. Although semantically both prefixes are attenuative (Kalme and Smiltniece 2001, 121), the prefix ie-­normally attaches to adjectives denoting colour, while pa-­applies to all other adjectives: (13) dzeltens ‘yellow’ – iedzeltens ‘yellowish’ liels ‘big’ – paliels ‘biggish’ Some colour names allow both prefixes so that they function as derivational synonyms: zaļš ‘green’ – iezaļš, pazaļš ‘greenish’. Moreover, the prefix ie-­ can attach to colour names derived by means of the suffix -­gans to indicate a very slight tinge of colour: zils ‘blue’ – zilgans ‘bluish’ – iezilgans ‘slightly bluish’. Such derivatives are semantically synonymous with derivatives formed by means of two suffixes (cf. zaļganīgs and iezaļgans ‘slightly greenish’). 3.2.2  SMALL > SMALLER In Latvian, this type of diminutive meaning is only expressed by means of the negative prefix ne-­: liels ‘big’ – neliels ‘not big’, tāls ‘distant’ – netāls ‘non-­distant’ (MLLVG 1959, 297). The adjectives neliels and netāls actually mean ‘not very big’, ‘not very distant’, i.e. ‘less than big’ and ‘closer than distant’. This type of use of negation in expressing diminutive meanings is not typical and only applies to a small number of adjectives.

4 Verbs Both diminutive and augmentative meanings are represented in the verb-­formation system and both are linked to aspectuality. Prefixal verbal derivatives can realise perfective + diminutive and perfective + augmentative meanings. In addition to prefixation, augmentatives can also be formed by means of circumfixes, i.e. combinations of a prefix and a reflexive ending (Kalnača and Lokmane 2012). As examples of language use show, one and the same prefix, besides perfectivity, can also express different meanings denoting the extent of actions (Soida 2009). The verbal reflexive ending is not the same in the infinitive and finite forms; however, it always ends in the formative -­s, which is the last element of all such word forms (Kalnača and Lokmane 2012). Therefore from now on we will use the notation prefix + final -­s for derivational circumfixes. Suffixal iterative derivatives are associated with the augmentative meaning BIG > MANY/MUCH. Prefixal as well as suffixal derivatives that express diminutive and augmentative meanings are deverbal. 4.1 Diminutives 4.1.1  Prefixal diminutives Prefixal diminutives denote actions performed to a slight extent and can be formed by means of the prefixes pa-­, ie-­, aiz-­and uz-­(Soida 2009, 249; Horiguchi 2013):

260

Andra Kalnača

(14) staigāt ‘to walk’ – pastaigāt ‘to walk a little’ kustināt ‘to cause to move’ – iekustināt ‘to get moving’ smēķēt ‘to smoke’ – aizsmēķēt ‘to light up (a cigar)’ kliegt ‘to scream’ – uzkliegt ‘to scream out’ The prefixes ie-­, aiz-­also carry inchoative meaning as they refer to short actions performed in order to cause an action to begin (Soida 2009, 242). 4.2 Augmentatives 4.2.1  Circumfixal augmentatives Semantically, circumfixal augmentatives can be divided into two groups: actions performed to a great extent (BIG > VERY MUCH) and excessive actions (BIG > EXCESSIVE). The first type of meaning can be expressed by means of the circumfixes aiz-­+-­s, at-­+-­s, iz-­+-­s, no-­+-­s, pie-­+-­s and sa-­+-­s (Soida 2009, 253–5): (15) gulēt ‘to sleep’ – aizgulēties ‘to oversleep’ ēst ‘to eat’ – atēsties ‘to eat until the meal is no longer desired’ ceļot ‘to travel’ – izceļoties ‘to travel as much as one wishes’ brist ‘to wade’ – nobristies ‘to wade until tired’ dzert ‘to drink’ – piedzerties ‘to get drunk’ smēķēt ‘to smoke’ – sasmēķēties ‘to smoke until sated’ In some cases the circumfixes pie-­+-­s and sa-­+-­s are synonymous, for example, ēst ‘to eat’ – pieēsties, saēsties ‘to eat until sated’. The second type of meaning is normally expressed by means of the circumfix pār-­ +-­s: ēst ‘to eat’ – pārēsties ‘to overeat’, strādāt ‘to work’ – pārstrādāties ‘to overwork’ (Soida 2009, 256–7). Circumfixal augmentatives may imply negative attitude on the part of the author through the meaning extensions BIG > VERY MUCH > BAD or BIG > EXCESSIVE > BAD. 4.2.2  Suffixal augmentatives If iterativity is viewed as akin to quantity, i.e. if we presume that a repeated action is also an action that is performed a lot (Grandi 2009), it is possible to speak about suffixal augmentatives in Latvian. Iterative verbs are deverbal and can be formed by means of the suffixes -­āt, -­īt, -­ot and -­ināt: braukt ‘to drive’ – braukāt ‘to drive about’ knieb-­t ‘to pinch’ – knaibīt ‘to pinch repeatedly’ svilpt ‘to whistle’ – svilpot ‘to whistle repeatedly or continuously’ vest ‘to take (somebody somewhere)’ – vedināt ‘to take (somebody somewhere) repeatedly or continuously’ (16)

A chaotic or a pointless iterative action can be expressed by the iterative suffixes -­alēt, -­aļāt, -­uļot, -­avāt and -­elēt: (17) braukt ‘to drive’ – braukalēt ‘to drive about pointlessly’ spriest ‘to reason’ – spriedelēt ‘to indulge in speculations’

Latvian 261 Such suffixes also tend to add a pejorative meaning (BIG > MUCH > BAD) to the verbs they attach to, i.e. chaotic, pointless iterative activities are considered bad (Soida 2009, 253–7).

5 Conclusion In terms of Grandi’s diminutive/augmentative typology, Latvian basically fits in the B type: ‘presence of both diminutives and augmentatives’ (Grandi 2011b, 7). However, this is not true for all parts of speech. While nouns and adjectives have extensive diminutive systems, they lack true augmentative suffixes and prefixes and are therefore closer to the A group: ‘presence of diminutives, absence of augmentatives’. Latvian verbs, on the contrary, can be examples of the B type. As is clear from the description of the diminutive/augmentative morphology of Latvian provided above, the noun system makes use of suffixes only; diminutive adjectives can be formed by means of suffixes as well as prefixes; and verbs have the richest repertoire of diminutive/augmentative affixes, which includes prefixes, suffixes and circumfixes not commonly found in other parts of speech. However, the role of verbal suffixes in diminutive/augmentative formation is much more humble than that of their nominal and adjectival counterparts, and thus there is a certain asymmetry in the derivational use of affixes across parts of speech.

15.8 Luxembourgish

Peter Gilles

1 Introduction Luxembourgish (local language name: Lëtzebuergesch [ˈləʦəbuəjəʃ]; ISO 639–3 code ltz) is a small West Germanic language mainly spoken in the Grand Duchy of Luxembourg (500,000 inhabitants). Within a situation of societal multilingualism, Luxembourgish shares domains with German and French (official trilingualism; see Newton 1996; Horner and Weber 2008). While the first serves as the important language of everyday life and for informal literacy (especially in digital media), the latter two are used as written languages in more formal domains (administration, the workplace and so on). Luxembourgish is acquired as a first language by approximately 60 per cent of the population and spoken as an additional language by a certain number of the non-­Luxembourg residential population (Fehlen, Heinz, Peltier and Thill 2013). Historically, Luxembourgish originates from a Central Franconian dialect (i.e. Moselle Franconian) and is today considered from a sociolinguistic point of view as an ‘Ausbau language’. Due to extensive language contact with French and German (and, to a lesser extent, with English), lexical borrowing accompanied by phonological and morphological integration is quite high. The languages most closely related to Luxembourgish genealogically are German and Dutch. Like that of German, Luxembourgish morphology distinguishes three genders, but it has preserved fewer inflectional features (because of the loss of certain case distinctions, partial loss of the preterite and the present participle and so on). Umlaut and ablaut, though hardly productive any more, can indicate several morphological features (for example plural in nouns, tenses and so on), creating a complex system with numerous sub-­classes. Various word-­formation options involving prefixes, suffixes and compounding allow for the creation of new lexical items. All examples in this chapter stem from actual usage or dictionaries (Luxemburger Wörterbuch, Lëtzebuerger Online Dictionnaire) and are written in the official orthography. Luxembourgish has no fully fledged system of evaluative morphology (EM), but certain aspects of word formation can be attributed to EM. The most import feature among them is diminution. EM is thus confined to nominal word classes; verbs show no similar morphological features.

2 Diminution Within the historical West Germanic continuum, several suffixes emerged indicating diminution, and these led to a complex system in the languages concerned (see, for example,

Luxembourgish 263 Tiefenbach 1987; Fleischer and Bartz 2012, 231–5 for German varieties). As a Central Franconian variety, Luxembourgish historically lay in a contact zone between Low German and High German diminutive suffixes, that is, a suffix and an suffix, respectively. Reflexes of both suffixes can be found in the present-­day diminutive system (see Bertrang 1921; Bruch 1949; 1953). 2.1  Diminutive with and its allomorphs The most productive suffix in the present-­day language is (including its allomorphs), as it is in the genetically closest language, i.e. Standard German. The suffix originates from the extension of the Old High German suffix with a prepended (Germanic ). In contrast to Dutch or German, where these suffixes can create diminutives from nouns, adjectives and verbs, the Luxembourgish suffix is restricted to nouns only.1 In the following, I will discuss the allomorphy, gender assignment and semantic aspects of this diminutive. 2.1.1 Allomorphy The form of the diminutive morpheme shows rich allomorphy, which is dependent on the final consonant of the base noun, and which leads to the three allomorphs [ɕən], [jən] (both written as ) and [əlɕən] (written as ). In the commonest case – that is, when the base noun does not end with a (historical) velar, palato-­velar, alveolo-­palatal consonant or the alveolar fricatives [z, s] (including the affricate [ʦ]) – (pronounced [ɕən]) is attached to the base noun in the singular and in the plural (1): (1) Base noun Singular a. Päert Been Bam Club b. Kuerf Bouf Kand Pult c. Dëppen Decken Millen Uewen

Plural Päert Been Beem Clibb Kierf Bouwen Kanner Pulten Dëppen Decken Millen Iewen

Diminutive Singular Plural Päerdchen Päerdercher ‘horse’ Beenchen Beenercher ‘leg’ Beemchen Beemercher ‘tree’ Clibbchen Clibbercher ‘club’ Kierfchen Kierwercher ‘basket’ Béifchen Béiwercher ‘boy’ Këndchen Kënnercher ‘child’ Piltchen Piltercher ‘lectern’ Dëppchen Dëppercher ‘pot’ Deckchen Deckercher ‘blanket’ Millchen Millercher ‘mill’ Iefchen Iewercher ‘oven’

In addition to the suffixation of , Umlaut, that is, the fronting of a back vowel to the corresponding front vowel, takes place in the diminutive whenever possible. Considering the quite high number of monophthongs and especially diphthongs, Luxembourgish is rich in Umlaut possibilities (see Nübling 2006; Gilles and Trouvain 2013). The words in (1b) exemplify only a few possibilities of these vowel mutations. Often the plural of the base noun already shows Umlaut, as in Bam [baːm] ‘tree.sg’, Beem [beːm] ‘tree.pl’, which could lead to the assumption that the plural of the base noun serves as the input to diminution. That this is not entirely the case is reflected by the further examples in (1b): diminutives are affected by Umlaut even when the plural of the base noun is not (compare Pulten

264

Peter Gilles

(pl) with Piltchen (dim.sg)). Umlaut, then, has to be regarded, next to suffixation, as a core feature of diminution in Luxembourgish.2 The examples in (1c) further demonstrate how diminutives are formed when the base noun ends with the (unstressed) suffix . Under these circumstances, and probably for prosodic reasons, the final syllable is removed before the diminutive suffix is attached. Thus, the diminution of Dëpp-­en ‘pot’ does not lead to *Dëpp-­en-­chen but rather to Dëpp-­ chen ‘pot-­dim.sg’. The morphological aspects presented so far are identical to those of Standard German (compare, for example, German Bäum-­chen ‘tree-­dim.sg’, Töpf-­chen ‘pot-­dim.sg’ with the corresponding Luxembourgish forms Beem-­chen, Dëpp-­chen). Clear differences arise, however, regarding the plural of diminutives. While in Standard German diminutives are not marked for plural and is used throughout, Luxembourgish has its own plural suffix with the remarkable property of being a combination of both interfixation and suffixation. As can be seen in (1), the plural of these diminutives is formed with , as in, for example, Been-­ercher ‘leg-­dim.pl’. This suffix makes use of the suffix present in the language as a plural marker, for example in Bréiw-­er ‘letter-­ dim.pl’ and modifies the singular suffix accordingly. The modification from the singular to in the plural seems straightforward, as an inflectional marker is attached after a derivational marker. This, in turn, complies fully with the typological universal that inflectional applies after derivational morphology (Greenberg 1963). But the Luxembourgish plural is marked twice and a second is interfixed between base noun and diminutive suffix, leading to the following structure: ‘base noun’ + -­pl + -­dim + -­pl. It is obvious that this double marking of the plural in forms such as Béiw-­er-­ch-­er ‘boy-­dim-­pl’ violates the universal above, because inflection of the first takes place before the derivational affix is attached. The reason for this unusual double marking is discussed by Chapman (1996), assuming that the input to the plural formation of the diminutive was in fact the plural of the base noun. This means that the plural formation of the diminutive started out with base noun plurals on (Mann ‘man.sg’, Männ-­er ‘man-­pl’) and this development then spread out via analogy on further word classes. However, the interfixed plural is lost when the base noun has a trochaic prosodic structure with a final unstressed syllable, such as , or . As shown in (2), trochaic base nouns such as Kallef ‘calf’ attach only as their diminutive plural form: Källef-­cher ‘calf-­dim.pl’. Here, a general prosodic constraint may be active, aiming at a preferred dactylic prosody for diminutives in both the singular and the plural: (2) Base noun Singular Plural Kallef Kaalwer Wollef Wëllef Apel Äppel Wuerzel Wuerzelen Artikel Artikelen Fuedem Fiedem

Diminutive Singular Plural Källefchen Källefcher ‘calf’ Wëllefchen Wëllefcher ‘wolf’ Äppelchen Äppelcher ‘apple’ Wierzelchen Wierzelcher ‘root’ Artikelchen Artikelcher ‘article’ Fiedemchen Fiedemcher ‘fibre’

However, variants of these plurals exist which exceed the three-­syllable pattern of the dactyl. The examples in (3) illustrate that as well as the regular plural, as in Frä-­ ercher [fʀɛːəɕɐʀ] ‘wife-­dim.pl’, a more complex variant Frä-­cher-­cher [fʀɛːɕəɕɐʀ] can

Luxembourgish 265 be constructed through reduplication of the suffix. This pattern probably arose through hypercorrection. It is not productive any more and is restricted to only a few words: (3) Base noun Singular Plural Fra Fraen Mann Männer Spill Spiller So Soen

Diminutive (+ variant) Singular Plural Frächen Fräercher/Frächercher ‘woman’ Männchen Männercher/Männchercher ‘man’ Spillchen Spillercher/Spillchercher ‘game’ Seechen Seeërcher/Seechercher ‘legend’

When the base noun ends in an alveolar fricative, regardless of whether that is voiced or voiceless, [ɕən] changes its pronunciation to the allomorph [jən] (glide formation); see examples in (4). The plurals of these diminutives receive the regular suffix and thus retain the fricative pronunciation. Despite its phonetic consistency, this allomorphy is not reflected in the official spelling, and is used throughout: (4) Base noun Singular Plural Haus Haiser Glas Glieser Noss Nëss

Diminutive Singular Plural Haischen [ˈhɑɪsjən] Glieschen [ˈgliəsjən] Gliesercher Nësschen [ˈnəsjən] Nëssercher

’ e s uoʀh] ‘ Haisercher [ˈhɑɪzəɕɐ ʀ [ˈgliəzəɕɐ ] ’sagl‘ [ˈnəsəɕɐʀ] ’nut‘

The third and last allomorph turns up when the base noun ends either with historical velar consonants [k, ŋ, χ, ɕ] (5a) or with [ʃ, ʦ] (5b). These consonants trigger the appearance of the interfix between the base noun and the suffix : : (5) Base noun Diminutive Singular Plural Singular a. Sak Säck Säckelchen Zong Zongen Zéngelchen Duch Dicher Dichelchen Bauch Bäich Bäichelchen b. Fësch Fësch Fëschelchen Putsch Pitsch Pitschelchen Mutz Mutzen Mitzelchen Fatz Fatzen Fätzelchen

Plural Säckelcher ‘bag’ Zéngelcher ‘tongue’ Dichelcher ‘cloth’ Bäichelcher ‘stomach’ Fëschelcher ‘fish’ Pitschelcher ‘tuft’ Mitzelcher ‘cap’ Fätzelcher ‘rag’

The interfix originates from a High German diminutive suffix, which also spread into Central German during the Early Modern German period, whereas belongs to the Low German area (Tiefenbach 1987). From a language-­historical point, a word such as Zéng-­el-­chen ‘tongue-­dim-­sg’ is thus diminutivised twice.3 This general avoidance of is probably due to phonological reasons. In today’s speech the infixation of is becoming more and more optional, especially for the contexts in (5b). Note that Standard German under these circumstances also selects an allomorph other than . Here, is employed instead in most cases (Büch-­lein ‘book-­dim.sg/pl’, Bäch-­lein ‘creek-­ dim.sg/pl’).

Peter Gilles

266

Comparable to the cases of final schwa syllables discussed above, these final syllables remain present in the plural, giving rise to a further plural allomorph , as in Säck-­ el-­cher ‘bag-­dim-­sg/pl’. 2.1.2  Gender assignment The most striking feature, and one yet to be explained, concerns gender: Luxembourgish diminutives keep the gender of the base noun consistently4 This stands in great contrast to Standard German or Dutch, where all diminutives receive neuter gender by default (6). (6) m m f f

ntr

Luxembourgish Base noun Diminutive Mann m de/e Männchen Won m de/e Weenchen Fra f d’/eng Frächen Boun f d’/eng Béinchen Kand ntr de/e Këndchen

Standard German Base noun Diminutive Mann ntr das/ein Männchen Wagen ntr das/ein Wägelchen Frau ntr das/ein Frauchen Bohne ntr das/ein Böhnchen Kind ntr das/ein Kindchen

‘the/a man’ ‘the/a car’ ‘the/a woman’ ‘the/a bean’ ‘the/a child’

This gender preservation questions, of course, the category-­determining ability of the diminutive suffix in Luxembourgish. While other word-­formation suffixes affect gender (and word class), the only aspect determined by the diminutive morpheme is the selection of the plural suffix. This leads to the conclusion that the Luxembourgish diminutive morpheme cannot act as the head of the construction; see Booij (2007, 14) for a similar discussion of the Italian diminutive suffix . 2.1.3  Semantic aspects On the most general level, diminutives can be used to indicate referential smallness; for example, Äermchen ‘arm.dim.sg’ actually means ‘small arm’; and it is this semantic domain where diminution is quite productive. Furthermore, diminutives can often convey endearment and/or irony, as in Lännchen ‘country.dim.sg’ in the sense of ‘our cosy country (i.e. Luxembourg)’ or Këndchen ‘child.dim.sg’ in the sense of ‘sweetheart’. Also, several diminutives underwent lexicalisation by loosening the semantic relationship with their base nouns (7). Then either the meaning of the diminutive focuses on another semantic concept (7a), or the base noun is not in use any more and the diminutive takes over its meaning (7b). (7) Base noun a. aacht ‘eight’ Klees proper name Hiem ‘shirt’ Schaf ‘closet’ Knupp ‘bump’ b. *Lappen *Chrëschtbam *Schnéimann

Diminutive Aachtchen ‘pastry (in the form of an 8)’ Kleeschen ‘Santa Claus’ Hiemchen ‘undershirt/chemise’ Schäffchen ‘cooker’ Knippchen ‘praline’ Läppchen ‘facecloth’ Chrëschtbeemchen ‘Christmas tree’ Schnéimännchen ‘snowman’

Interestingly, the relationship with ‘size’ is lost completely in the examples in (8), where reference is made to quite big items through adjectives (grouss/gréisst ‘big(gest)’) and quantification (4 Meter ’4 metres’). Consequently, the diminutive and the base noun become more or less synonymous.

Luxembourgish 267 (8)

De gréisste Schnéimännchen ass 4 Meter héich. ‘the biggest snowman.dim is 4 metres high’ aus enger grousser eise Biddchen ‘out of a big iron tub.dim’

2.2  Diminutive with Restricted regarding the selection of the base noun, but nevertheless quite productive, is the diminutive suffix . These diminutives form hypocoristics (endearment forms). Base nouns are often kinship terms (9a), terms of intimacy (9b) or hypocoristic first names (9c). (9) a. f m ntr m m b. ntr f c. m m f m

Base noun Singular Plural Boma Bomaen Bopa Bopaen Kand Kanner Fils Filsen Jong Jongen Häerz Häerzer Popp Poppen Dominique -­ Änder -­ Stéfanie -­ Caspar -­

Diminutive Singular Plural Bomi Bomien ‘grandmother’ Bopi Bopien ‘grandfather’ Kënni Kënnien ‘child’ Fissi Fissien ‘son’ Jéngi Jéngien ‘son’ Häerzi Häerzien ‘sweetheart’ Pëppi Pëppien ‘doll’ Dëmmi -­ Ändri -­ Fanni -­ Kaschi -­

Like , Umlaut is applied consistently, and the choice of the plural suffix is determined through the diminutive and not by the word class of the base noun (compare Kann-­er ‘child-­pl’ and Kënn-­i-­en ‘child-­dim-­pl’). Note that the gender of the base noun remains the same in the diminutives.

3 Augmentation On the morphological level, augmentatives do not exist and analytical phrases are preferred instead. Sometimes, however, nouns or adjectives can be modified with prefixes such as mega-­, super-­and risen-­. While used quite frequently, these words are stylistically marked, for example megagrouss ‘big.aug’, superschéin ‘beautiful.aug’, Risebuttek ‘trouble.aug’.

4  Adjective approximation with The next suffix also does not change the word class of the base word, making it a potential candidate for EM. With the help of the suffix the semantics of certain adjectives can be rendered as approximate, in the sense of ‘somewhat, rather, quite’. Thus, gréng-­ elzeg means ‘not quite green, greenish’. (10) Base adjective Derived adjective a. blo ‘blue’ bloelzeg ‘bluish’ gro ‘grey’ groelzeg ‘greyish’ gréng ‘green’ gréngelzeg ‘greenish’

268

Peter Gilles

brong ‘brown’ brongelzeg ‘brownish’ roud ‘red’ roudelzeg ‘reddish’ wäiss ‘white’ wäisselzeg ‘whitish’ giel ‘yellow’ gielzeg ‘yellowish’ b. laang ‘long’ längelzeg ‘longish’ ronn ‘round’ ronnelzeg ‘roundish’ This type of word formation rarely spreads to other adjectives. Apart from the adjectives denoting spatial relations in (10b) and a few others, it seems that this adjective approximation is primarily restricted to colour adjectives, as (10a) shows. It is thus not productive any more. Note that in this case Standard German employs the suffix (grün-­lich ‘greenish’), which, however, is not confined to the semantics described here, but is quite polysemous and able to create various kinds of adjectives. In contrast, in having this evaluative suffix , the word formation of Luxembourgish is more nuanced.

5 Conclusion It has been the aim of this chapter to deliver a descriptive account of the EM of Luxembourgish. The chapter has focused largely on diminution. In a next analytical step, it would seem necessary, however, to amend the description with explanatory functional analyses in order to detect productive and less productive semantic patterns. This could only be done by adopting a corpus linguistic approach.

Notes 1. Besides the regular lexicon, the suffix played an important role in the formation of family names in Luxembourg and the larger Rhineland area (for example Bemtgen [ˈbeːmtɕən] literally ‘small tree’, Wildgen [ˈviltɕən] literally ‘small Will(iam)’), originally meaning ‘small one’ or ‘son of’. This word-­formation pattern was later stabilised as a family name and the meaning of the original diminutive was subsequently lost. 2. In addition, syllable final devoicing of underlying voiced obstruents takes place whenever possible: compare the plural form Kierwercher [ˈkiəvəɕɐʀ] ‘basket.dim.pl’, containing the underlying voiced obstruent [v], with the singular form Kierfchen [ˈkiəfɕən] ‘basket.dim.sg’, exhibiting the devoiced obstruent [f]. Note that final devoicing is reflected in the official orthography of Luxembourgish. 3. Several relics of the former diminution with the suffix exist today as lexicalisations: Fänd-­el ‘flag’, Fierk-­el ‘farrow’ and others have to be regarded as fossilised diminutions. 4. With the exception of Meedchen ‘girl.dim.sg’, which is – probably due to language contact with German (Mädchen) – neuter instead of feminine.

15.9  Modern Greek

Dimitra Melissaropoulou

1 Introduction Greek refers to all different varieties of the Greek language spoken in the modern era by a total of approximately 14 million to 15 million people. Nowadays, its standard form, Standard Modern Greek (SMG) is the official language of Greece, situated in northeast Europe, in the Balkan peninsula, where it is spoken by about 99 per cent of the population. It is also one of the two official languages of Cyprus (alongside Turkish), the third largest island in the Mediterranean, located east of Greece and south of Turkey. Apart from Greece and Cyprus, Greek is also spoken today in the neighbouring countries of Albania, Bulgaria and Turkey, as well as in several countries in the Black Sea area (Ukraine, Russia, Georgia and Armenia) and around the Mediterranean (southern Italy and Egypt), in some of which it is recognised as a minority language. The language is spoken too by emigrant communities in many countries in Western Europe, North America and Australia, as well as in Argentina, Brazil and other countries. Greek forms an independent branch of the Indo-­European languages. All surviving forms of Modern Greek, except for Tsakonian, are descendants of the common supra-­ regional Hellenistic Koine that was spoken in late antiquity, and constitutes an evolution of Ancient Greek (varieties). In what follows we will use the term ‘Greek’ only when making statements about SMG in general, while the term ‘dialect’ will be used to refer to varieties other than SMG. From a typological viewpoint Greek belongs to the synthetic–fusional type. Its morphology is stem based, and it is a typical example of a system with rich morphology.

2  Sketchy description of morphology As already mentioned, Greek is a morphologically rich language. Derivation and ­compounding are used in an extremely productive way in order to derive new lexemes. Greek morphology is stem based (Ralli 2005); that is, different kinds of bound stems are combined with different types of affixes of the portmanteau type, in the sense that one form may realise simultaneously more than one meaning or function, and various ­morphophonological alternations are attested. For the purposes of this chapter, we will refer mainly to derivation, and to derivational suffixation in particular, since it is through this process that evaluative morphology is strongly present in Greek. In derivational structures the stem is responsible for the basic lexical meaning. Derivational suffixes constitute, at least for Greek, the head of their constructions (Zwicky

270

Dimitra Melissaropoulou

1985; for Greek cf. Ralli 2005); they bear a specific meaning, a specific grammatical category and a specific gender value. In other words, both gender and grammatical category in derived words depend on morphological criteria, since they are assigned through percolation from the inherently marked derivational suffix to the derived structure. Lastly, the attachment of the derived stems to inflectional markers adds specific morphosyntactic features (e.g. in nominal derivation the features of case, number and inflection class are marked), completing the morphological formation. In Greek, evaluative suffixes are thought of as derivational suffixes and constitute heads of their formations (for relevant discussion see Section 4), passing their characteristics through percolation to the derived word formations (Melissaropoulou and Ralli 2008). Thus, the above-­mentioned information on derivational structures is applicable to evaluative constructions that will be discussed in the following sections. Furthermore, the productivity of evaluative suffixes, as is the case for any other derivational suffix, is strongly related to the restrictions these are subject to (for further discussion see Sections 3.1 and 3.2).

3  Evaluative constructions Greek is fairly rich in evaluative constructions. However, not all the different types of evaluation are equally productive. Evaluation is realised mainly by means of suffixation, and of prefixation or compounding to a lesser extent. Prefixation or compounding, although not as productive as suffixation, are applicable to all verbs and nouns, adjectives and adverbs, and realise both positive and negative evaluation without changing the grammatical category or other grammatical features of the base. The most typical evaluative prefixes, prefixoids or bound stems1 are psilo-­dim, mikro-­ dim, kutso-­dim, arçi-­aug, kara-­aug and poli-­aug. Some indicative examples can be seen in (1):2 (1) psilo-­kaˈpnizo dim-­to smoke.v ‘to smoke a bit’ poli-­kaˈpnizo aug-­to smoke.v ‘to smoke a lot’ psilo-­ziˈmja dim-­damage.n ‘small damage’ kara-­ziˈmja aug-­damage.n ‘big damage’ psilo-­kaˈla dim-­good.adv ‘quite good’ arçi-­maˈfjozos aug-­mobster.n ‘big mobster, chief mobster’ Furthermore, evaluative prefixes can co-­occur with the corresponding suffixes in order to maximise or mitigate both quantitative and qualitative evaluation. For example:

Modern Greek 271 (2) psilo-­xaˈz-­ulis dim-­fool.adj-­dim ‘foolish’ kara-­vlaˈx-­ara aug-­boor.n-­aug ‘(very) big boor’ In the following sections we will focus on nominal formations, since they are far more productive and interesting than any other type of evaluative construction. However, before proceeding, some important premises should be stated:

(i) I n the spirit of Grandi (this volume), in Greek the same suffix can express both qualitative and quantitative aspects of evaluation. In other words, an object (or a person, an action, etc.) can be ‘evaluated’ according both to its actual, real characteristics (its size, its shape, etc.), and to the speaker’s feelings or attitudes towards it. For example luluˈðaci < luluð-­aci ‘flower-­dim’ can be both a small and a dear, favourite or beautiful flower. However, some evaluative suffixes may differ in terms of prominence of qualitative vs. quantitative aspects of evaluation. Moreover, Greek supports the claims made by Klimaszewska (1983) that positive evaluation is more characteristic of diminutives, and negative evaluation of augmentatives. (ii) A distinction between typical and peripheral evaluative forms is useful for Greek as well. We count as typical constructions those formations that bear an evaluative reading, as discussed above, and as peripheral (or non-­prototypical) constructions those formations in which, although an evaluative reading is evident, it combines with another meaning (such as agent or instrument). For example, although the Greek suffix -­a(s) and its corresponding formations (e.g. faˈγa(s) < faγ-­a(s), ‘the person who eats to exaggeration’) bear an evaluative reading, this reading merges with the meaning of the agent. Thus, for the purposes of this chapter we will focus on those evaluate suffixes that are thought to form typical evaluative constructions. (iii) Word formations that carry an evaluative suffix (in terms of form) without identifying evaluative readings, but only default meaning and standard size (for example niˈfitsa ‘weasel’), are considered to be lexical diminutives and are excluded from the present study. In the same vein, lexicalised evaluative constructions (i.e. formations that, although they resulted from the combination of a stem with an evaluative suffix, have become semantically opaque) are also left aside. For example:3 (3)

skoˈnaci NOT dust.n.dim ‘piece of paper used in cheating at the exams’

pistoˈlaci NOT pistol.n.dim ‘hairdryer’

Dimitra Melissaropoulou

272

In the following sections, we focus on the two different facets of evaluation: suffixal diminution and augmentation. 3.1 Diminution Greek displays a wide range of diminutive suffixes, the most typical and productive of which are given in (4):4 (4) -­aci.ntr -­itsa.f -­uli(s).m -­ula.f -­uli.ntr -­opul(o).ntr -­opula.f -­aci(s).m -­ak(os).m -­utsik(οs)/-­i/-­o.adj -­aka.f -­uʎaka(s).m

pexniˈð-­aci toy.ntr-­dim ‘small toy’ kuˈkli-­itsa doll.f-­dim ‘small doll, dolly’ miˈsθ-­uli(s) salary.m-­dim ‘small salary’ γravaˈt-­ula tie.f-­dim ‘small tie’ kriˈft-­uli hide-­and-­seek.ntr-­dim ‘hide-­and-­seek’ voˈsk-­opul(ο) grazier.m-­dim ‘young male grazier’ vosk-­oˈpula grazier.f-­dim ‘young female grazier’ koˈzm-­aci(s) ‘people-­ dim’ ‘poor people’ skiˈl-­ak(os) dog.m-­dim ‘small dog, doggy’ kaˈl-­utsik(οs) good.adj-­dim ‘quite good’ maˈm-­aka mum.f-­dim ‘mummy’ beˈkr-­uʎaka(s) drunkard.m -­dim/pejor ‘drunkard’

Although the repertoire of diminutive suffixes in Greek is quite ample, they are not all equally productive. -­aci is by far the most productive suffix. It is attached to every possible base that can undergo diminution (nouns of the three grammatical genders, proper

Modern Greek 273 and common, natives or loans). -­uli(s)/-­a/-­i is extremely productive as well, attaching to both nouns and adjectives, while the three forms correspond to three different grammatical gender values realised in Greek not only in terms of realization but also in terms of base selection; that is, -­ulis selects masculine bases and -­ula feminine ones, while -­uli mostly – but not only – selects neuters.5 Crucially, the range of one suffix may include the range of the other, provided that the former has wider distribution. Moreover, while the most productive diminutive suffixes realise both quantitative and qualitative readings (e.g. -­aci, -­ulis/-­a/-­i), the most restricted ones, such as -­acis, -­akos and -­uʎakas, realise mainly or only qualitative readings (for example, -­uʎakas is mainly pejorative6). However, while SMG displays a wide range of diminutive suffixes, in the vast majority of Modern Greek dialects the diminutive suffixes in use are quite limited (Melissaropoulou 2006; 2007) and they usually flag local identity. Similarly, while high alternation is attested in the occurrence of SMG diminutive suffixes (Katramadou 2001), as shown in (5) below – a property that could be used as an argument against their derivational character7 – this is not a generalised phenomenon, since Greek dialects either display a restricted degree of alternation or do not accept it (Melissaropoulou 2006; Melissaropoulou and Ralli 2008; Ralli and Melissaropoulou 2008).8 (5) γariˈð-­ula γariˈð-­itsa γariˈð-­aki shrimp.f-­dim shrimp.f-­dim shrimp.ntr-­dim ‘small shrimp’ ‘small shrimp’ ‘small shrimp’ As regards the combinability of diminutive suffixes, SMG generally allows diminutive suffix chains (Melissaropoulou 2006; Melissaropoulou and Ralli 2008); for example: (6) -­ul-­itsa -­ul-­ak(os) -­opul-­aci

man-­uˈl-­itsa mum-­ dim-­dim ‘(sweet) mummy’ bab-­uˈl-­ak(os) dad-­ dim-­dim ‘(sweet) daddy’ lik-­opuˈl-­aci ‘wolf-­ dim-­dim ‘young or small/sweet wolf’

However, cumulation of diminutive suffixes is not unrestricted. For example, the combined suffixes cannot be identical, while the most productive suffix, -­aci, cannot be followed by other suffixes (Melissaropoulou and Ralli 2008). Lastly, with respect to the distribution of different grammatical gender values in diminution, Katramadou (2001) has argued that in SMG, masculine and feminine bases form diminutives of the same gender, along with a neuter form. For example: (7) ˈvarka.f ‘boat’ ˈðromos.m ‘road’

varˈk-­ula.f varˈk-­aci.ntr boat-­ dim ‘boat-­ dim ‘small boat’ ‘small boat’ ðroˈm-­ak(os).m ðroˈm-­aci.n road-­ dim road-­ dim ‘small road’ ‘small road’

Dimitra Melissaropoulou

274 ˈkliði.ntr ‘key’

kliˈðaci.ntr key-­ dim ‘small key’

In a similar vein, Daltas (1985), on the basis of a quantitative analysis, argues that a diminutive formation is more likely to be of neuter gender than of the other two grammatical gender values. These findings reveal that diminutive words are generally assigned the neuter gender value, verifying the claim that this is the unmarked grammatical gender value of diminutive formations for Greek (Grandi 2002). 3.2 Augmentation The formal mechanisms of augmentative constructions fall into the same categories as the diminutive ones. However, they are less widespread than the latter. The most typical ­augmentative suffixes attested in Greek can be seen in (8): (8) -­a.f -­os.m -­ar(os).m -­ara.f -­akla/ukla.f -­akla(s)/ukla(s).m -­arona.f -­u(m)ba.f

piˈrun-­a fork.ntr-­aug ‘big fork’ ˈkras-­os wine.ntr-­aug ‘large quantity of wine’ ˈγat-­ar(os) cat.f-­aug ‘big (female) cat’ krevaˈt-­ara bed.ntr-­aug ‘big bed’ koˈr-­akla daughter.f-­aug ‘big/beautiful daughter’ miˈt-­ukla nose.f-­aug ‘big/ugly nose’ ˈadr-­akla(s) man.m-­aug ‘big/handsome man’ spit-­aˈrona house.ntr-­aug ‘big house’ tsiγaˈr-­uba cigarette.ntr-­aug ‘big/fine cigarette’

The most productive of the above-­mentioned suffixes are -­a, -­ara and -­ar(os), while the most recent, -­u(m)ba, shows signs of increasing productivity. Among them, -­ara is basically the most productive augmentative suffix and is practically restriction free, since almost every noun susceptible to augmentation can attach this suffix (Daltas 1985). -­aro(s)

Modern Greek 275 attaches mainly, but not only, to [+ animate] masculine or [+ animate] non-­masculine bases, via which natural gender, i.e. sex, can be realised. For example, the feminine noun ˈγata ‘cat’, when augmentivised as ˈγataros, bears masculine grammatical gender value which also differentiates sex. In other words, ˈγataros is a big male cat. Lastly, -­a suffix combines with [− animate] neuter nouns of the most productive inflectional class (i.e. nouns ending in -­i). As in diminution, the range of augmentative suffixes in use is much more limited in Greek dialectal variation (Melissaropoulou 2009). Gender assignment in augmentative formations has been challenged in comparison with diminutive ones by emphasising the absence of neuter gender value in augmentation, accounted for on the basis of the relation between augmentatives and the characteristic [+ animate] that marks the evolution of augmentative suffixes (Grandi 2002). According to Daltas (1985, 74) ‘the probability of an augmentative form being feminine is higher than that of being masculine, whereas the probability of its being neuter is zero’. In other words the unmarked grammatical gender value for augmentation in SMG is the feminine one. Modern Greek dialects confirm the general claim that neuter gender value is not realised in the process of augmentation. However, contrary to what is thought to be the prevailing and thus characteristic value for augmentation (i.e. the feminine), what seems to prevail is the masculine gender value (for relevant discussion see Melissaropoulou 2007; 2009). As far as cumulation of augmentative suffixes is concerned, this is not a typical strategy for Greek. On the other hand, some limited alternation is attested provided that the ­selectional restrictions are satisfied. For example: (9) ˈadr-­aklas aˈdr-­akla man.m-­aug man.m-­aug ‘big man’ ‘big man’ kaθiˈʝit-­aros kaθiʝiˈt-­ara professor.m-­aug professor.m-­aug ‘great professor’ ‘great professor’ foˈn-­ara foˈn-­akla voice.f-­aug voice.f-­aug ‘big voice’ ‘big voice’ However, research on Modern Greek dialects has revealed that alternation between augmentative suffixes is not meaningless. It may express a different grade of augmentation, the stronger one being realised by a masculine augmentative suffix, offering further support for the claim that there is a profound correlation between augmentation and masculine ­grammatical gender (Melissaropoulou 2007; 2009).

4  The position of evaluative morphology in grammar Admittedly, although the position of evaluative formations within morphology has been hotly debated, it has not delivered any results of cross-­linguistic validity, since evaluative constructions are considered to be inflectional in some languages, derivational in some others and of an intermediate nature in others again. With respect to Greek, it is argued that suffixed evaluative constructions belong to the derivational domain (Melissaropoulou 2007; Melissaropoulou and Ralli 2008). Their character is supported on the basis of the

Dimitra Melissaropoulou

276

f­ollowing criteria: categorical and selectional restrictions, alternation of diminutive suffixes for the same base, and relative affixal order. As far as categorical and selectional restrictions are concerned, as already mentioned in the previous sections, bases are not freely selected by Greek evaluative suffixes. For example, the -­itsa suffix cannot attach to masculine or neuter bases: (10) *n.m -­itsa *klið-­itsa < kliði.n ‘key’ Although alternation between evaluative suffixes is relatively high with respect to the other derivational suffixes, this is limited only to the SMG form. Modern Greek dialects (e.g. Grico, Aivaliot, Pontic and Cappadocian, for which see Melissaropoulou and Ralli 2008) show a very low rate of alternation, reinforcing their derivational character and offering further support to the claim that the selection of one particular suffix can indicate social or geographical variation (Grandi 2002, 154). Lastly, with respect to their relative affixal order, contrary to what has been argued for other languages (e.g. Scalise 1988b), evaluative suffixes are not steadily peripheral with respect to derivational suffixes, but can precede other productive derivational suffixes (Melissaropoulou 2007; Melissaropoulou and Ralli 2008). Again, this offers further support to their derivational character. For example, in (13) the diminutive suffix -­uli(s) is followed by the adjectival suffix -­ik(os): (13) xoˈdr(os) xoˈdr-­uli(s) fat.adj fat. adj-­dim ‘fat’ ‘fatty’

xoˈ dr-­ul-­ik(os) fat. adj-­dim-­adjvl.sfx ‘fatty’

On the basis of their derivational status, evaluative suffixes are considered to function as structural heads of their constructions, in the same way as the other unequivocal derivational suffixes, since they (i) may impose modifications on semantic features of the base (e.g. the feature [±countable]), or cause a semantic specialization of the base; (ii) are inherently marked for a specific grammatical gender value which passes on to the derived structure; (iii) determine the inflection class of the derived structure, independently of the inflection class of the base; and (iv) are responsible for certain irregularities or gaps in the inflectional behaviour of the derived words (Melissaropoulou 2007; Melissaropoulou and Ralli 2008).

5 Conclusion In this chapter, a brief presentation of Greek evaluative constructions has been attempted, emphasising derivational suffixation, since it is through this process that evaluative morphology is strongly present in Greek. Particular emphasis was given to the range of suffixes in use, their selectional restrictions, gender assignment, alternation and combinatorial properties. Finally, the place of evaluative markers in grammar was discussed, positing that for Greek they belong to the derivational domain.

Notes 1. The status of these elements as affixes, affixoids (in the sense of Booij 2005) or pseudo-­compounds (Ralli 2005, 169–70) is not uniform and entails a careful investigation of each separate element.

Modern Greek 277 2. Verbal evaluative constructions are provided in the first person singular of the present tense, and nominal ones in the nominative case of the singular number, throughout the chapter. 3. The diminutive suffix -­aci is transcribed using the palatal /c/ since velar consonants in Greek when followed by a front vowel, /e/ or /i/, are palatalised. 4. Inflectional suffixes are given in parentheses. Absence of parentheses shows that, according to the analysis adopted (Ralli 2000), word forms coincide with stems. In examples where no specific information is provided in the gloss or the translation, qualitative evaluative readings (positive or negative) may vary depending on the context. Note that in koˈzm-­aci(s) in (4), -­aci(s) denotes sympathy or contempt. 5. Although -­ulis/-­a/-­i have three different grammatical gender values, and on this basis one could talk about one suffix that acquires the gender of the base, we recognise three different forms and talk about selection on the part of the suffix because (i) the neuter form -­uli can also select non neuter bases (e.g. Dimitra ‘female proper name’ > Dimitruli) and (ii) the suffix percolates along with gender other grammatical features to the derived form, such as membership in a specific inflection class. 6. -­uʎakas, although it carries a pejorative reading, is usually listed as a diminutive and not as an augmentative suffix, possibly due to its partial formal correspondence with -­uli. 7. High alternation could be used as an argument against the derivational status of evaluative suffixes because derivational suffixes, realising specific semantic and grammatical features and being subject to specific selectional restrictions, are generally not interchangeable. 8. On restrictions posed in the alteration between -­ula and -­itsa suffixes see Daltas (1985, 80). Note that the form γariˈð-­aki in (5) has another rather lexicalised meaning, as well referring to a particular type of crisps.

15.10 Nivkh

Ekaterina Gruzdeva

1 Introduction Nivkh (Paleosiberian) is a moribund language spoken in several dialects on Sakhalin Island and in the Amur region of Russia. This chapter is based on data for the Amur and East-­Sakhalin dialects. Nivkh is a head-­marking polysynthetic language (Mattissen 2003) with elements of morphological fusion and SOV word order. Though the nominal case system in general is quite developed, the core syntactic relations (subject, primary and secondary objects) are unmarked. The structure of a finite verb may be roughly represented as follows: object-­root-­ transitivity-­evaluation-­aspect-­causativity-­aspect/tense-­modality-­negation-­mood-­ number-­focus. The minimal verb form of a matrix clause comprises a root and a mood suffix. The verbs oppose several aspects and moods and two tenses (non-­future and future). The agreement with a subject in number is optional. There are no adjectives that would be morphologically distinct from verbs. In both dialects, the majority of 400 verbs denoting qualities carry either obligatorily or optionally the suffix -­la-­, which indicates a permanent property (Panfilov 1965, 85–8; Otaina 1978, 63). Qualitative verbs may be used in various syntactic functions as both finite and non-­finite forms. In a clause, the verb typically occurs sentence-­finally. An attribute together with a head nominal as well as a primary object together with a head verb form a polysynthetic complex that can comprise various bound morphemes,1 e.g. ţ‘in1-­t‘om2 ‘your1 fat2’ in (7) and k‘e1-­slə-­r2 ‘pulling2 a fishnet1’ in (3) respectively. A participle used in the attributive function is represented by a bare verb stem in the Amur dialect and a verb stem with the suffix -­ŋ in the East-­Sakhalin dialect. An adverbial clause is formed by a converb which bears a temporal, conditional, concessive, subjunctive, cause, purpose or manner suffix, e.g. the subjunctive converb təvənə-­ʁar ‘[I] would have come in’ in (8). A relative clause is headed by a participle that synthesises with the following head nominal, e.g. nanak1 lət2-­tāz3 ‘the ornament3 made2 by [my] sister1’ in (4). In cases of coordination, each finite verb takes the suffix -­ra, -­ta in the Amur dialect and the suffix-­ra, -­ta, -­na in the East-­Sakhalin dialect, as in (2).

2 Evaluation In Nivkh, it is possible to express both descriptive and qualitative evaluation. Both evaluative types pertain almost exclusively to verbal morphology, which is rather uncommon cross-­linguistically and breaks the hierarchy suggested in Bauer (1997, 539–41). The

Nivkh 279 h­ ierarchy implies that if in a particular language a verb can be used as the base in evaluative morphology, then a noun is also used so in that language. The word-­formation evaluative strategies in Nivkh include suffixation, reduplication, analyticisation, voicing of initial plosives, vowel lengthening and a-­raising for descriptive evaluation, as well as verb root compounding for qualitative evaluation. The strategies vary in productivity and their application depends on the dialect. The evaluative suffixes take the position after the transitivising suffix and before the TMA suffixes, which confirms their status as internal markers with respect to inflectional morphemes; see Scalise (1986, 132–3).

3  Descriptive evaluation Nivkh descriptive evaluative morphology does not indicate prototypical diminutive and augmentative meanings connected with size. Since it mostly has affective meanings of attenuation and intensification, it would be probably best described as expressive morphology; see Bauer (1997, 560–1). Being marked only on the verb, the evaluation applies not to its arguments, but to the quality, state or action referred to by that verb. Diminutivisation is more productive, but less manifold in both dialects. As for augmentivisation, in the East-­Sakhalin dialect it is also a fairly productive and uniform process, whereas in the Amur dialect it is on the contrary less productive, but involves the whole range of various markers. Repetition of the same type of marker is not possible; however, several evaluative devices may be combined within the same verb form. 3.1 Diminution Diminution, or more precisely attenuation, which expresses that a quality, state or action manifests itself to a lesser degree, incompletely, is marked in both dialects predominantly by the suffix -­jo-­ (Panfilov 1965, 80; Otaina 1978, 58–9). The diminutives derived with this suffix can appear in all syntactic functions typical of verbs. In the Amur dialect, the diminutive suffix can be attached to various verb types, with different degrees of productivity. It can be used with a majority of qualitative verbs,2 e.g. the verb stems in (1) and coordinated finite verbs with the suffixes -­jo-­ and -­ra in (2):3 (1) ñeñi-­ ‘be sweet’ → ñeñi-­jo-­ ‘be sweetish’ per-­ ‘be heavy’ → per-­jo-­ ‘be heavyish’ (2) Ñi-­n maţka-­jo-­ra vəlvələ-­ra one-­ clf.humans be.small-­ dim-­coord.3sg be.swarthy-­ coord.3sg ñi-­n pi-­l-­jo-­ra valvaγ-­jo-­ra. one-­ clf.humans be.big-­ perm-­dim-­coord.3sg be.red-­skinned-­dim-­coord.3sg ‘One is smallish and swarthy, another is biggish, a bit red-­skinned.’ (Otaina 1978, 59)

Furthermore, the diminutive suffix -­jo-­ can be hosted by some processive verbs: (3)

K‘e-­slə-­r j-­eţ-­jo-­ja. fishing.net-­pull-­cvb.man.3sg 3sg-­drag.out-­dim-­imp.2sg ‘Pulling the fishnet, drag it out a little.’ (Panfilov 1965, 80)

Ekaterina Gruzdeva

280

Diminutives are extremely common in comparative constructions, e.g. the finite verb in -­jo-­ in (4): (4) Nanak lət-­tāz hə-­tāz-­ək k‘inŋu-­jo-­d̹. elder.sister make-­ornament that-­ornament-­compar be.beautiful-­ dim-­ind ‘The ornament made by [my] elder sister is a little bit better (lit. a bit beautiful) than that ornament.’ (Otaina 1978, 60) The diminutive suffix may occasionally be used on nouns as well. Example (5) with the noun hemañχ-­jo ‘an oldish woman’ should be considered exceptional: (5) Hemañχ-­jo umgu ñi-­n hum-­d̹ . dim woman one-­ clf.humans be-­ ind old.woman-­ ‘There is one woman, an oldish woman.’ (Panfilov 1965, 80) In the East-­Sakhalin dialect, the use of the diminutive suffix seems to be restricted to verb forms derived from qualitative roots with the completive/resultative suffix -­ʁar-­/-­χar.4 The resulting verbs mean that a certain property has been not fully acquired as a result of a completed process, e.g. the verb stems in (6) and the finite verb χa-­jo-­ʁar-­d ‘[fat] is a bit bitter’ in (7): (6) piŋr-­ ‘be cheap’ → piŋr-­ʁar-­ ‘cheapen’ → piŋr-­jo-­ʁar-­ ‘cheapen a little’ polm-­ ‘be blind’ → polm-­ʁar-­ ‘become blind’ → polm-­jo-­ʁar-­ ‘become a bit blind’

(7) Ţ‘in-­t‘om χa-­jo-­ʁar-­d. you.pl-­fat be.bitter-­dim-­compl/res-­ind ‘Your fat is a little bit bitter.’ (fieldwork 2000, 7: 101) According to Otaina (1978, 60), in some rare cases in the Amur dialect diminution can also be carried out by reduplication:5 (8) . . .təvənə-­ʁar maŋg-­maŋg-­ta. come.in-­ cvb.sbjv be.shy-­be.shy-­coord.1sg ‘. . .[I] would have come in, but [I] am little bit shy.’ (Otaina 1978, 60) The raising of the vowel a to ə or i in the suffix -­la-­ of qualitative verbs also signals the diminutivisation of the verb meaning in the Amur dialect, e.g. the converb pi-­lə-­r ‘[becoming] biggish’ in (9): (9)

V-­ōla pand-­r pi-­lə-­r nan 3sg-­child grow.up-­cvb.man.3sg be.big-­ perm-­cvb.man.3sg only ūmu-­ñivx-­doχ vi-­d̹ . fight-­man-­dat go-­ind ‘Only growing up, [becoming] biggish, his son went to the enemy (lit. the fighting man).’ (Otaina 1978, 64)

Nivkh 281 Finally, in the East-­Sakhalin dialect, the attenuation of quality, state or process is marked by the analytical form with the auxiliary verb ha-­‘be so’ and lengthening of the final vowel of a lexical verb:

(10)

Mu-­goţā ha-­d. boat-­move be.so-­ind ‘[He] moved a boat a little.’ (Krejnovich 1979, 312)

Within this analytical form, the vowel of the suffix -­la-­ of a qualitative verb is typically raised to ə and lengthened: (11)

‘be red’ → paʁ-­lə-ha-­

paʁ-­la-­

‘be a little bit red’ (Krejnovich 1979, 312)

3.2 Augmentation In Nivkh, the augmentative meaning is connected with the intensity of the feature named by the corresponding verb. Augmentatives are derived mainly from qualitative verbs. In the Amur dialect, intensification of the quality can be achieved by the voicing of an initial plosive of a qualitative verb.6 This process applies to twenty verbs in the attributive function (Otaina 1978, 57), e.g. the participles in (12) and the participle duz-­la ‘very cold’ in (13), derived from tuz-­la ‘cold’: (12)

pi-­la kə-­la

‘big’ → ‘long’ →

bi-­la gə-­la

‘very big’ ‘very long’

(13) Duz-­la-­ţ‘aχ-­ŋa-­da. be.very.cold-­ perm-­water-­go.for-­imp.1pl ‘Let’s go for very cold water.’ (Panfilov 1962, 7) Reduplication of a verb stem is another frequent process employed in augmentivisation in the Amur dialect. It is applicable to fifty qualitative verbs in the attributive function (Otaina 1978, 56). In cases where a verb bears the suffix -­la-­, the whole stem is reduplicated, e.g. the participles in (14)7 and the participle q‘o-­la-­q‘o-­la ‘very rich’ in (15): (14)

ŋar-­la p‘ŋaʁ

‘fat’ → ‘young’ →

ŋa-­la-­ŋa-­la p‘ŋaχ-­p‘ŋaχ

‘very fat’ ‘very young’

(15) V-­əmək faz-­u-­ror 3sg-­mother take.off-­ tr-­cvb.ant.dist q‘o-­la-­q‘o-­la-­hās-­ku-­k‘e-­gu-­ra. be.rich-­ perm-­be.rich-­perm-­clothes-­pl-­put.on-­caus-­coord.3sg ‘His mother taking off [his clothes] put on very rich clothes [on him].’ (Otaina 1978, 56) If a verb root begins with a voiceless plosive, the latter undergoes voicing in the process of reduplication:

Ekaterina Gruzdeva

282 (16)



pi-­la ‘big’ → pi-­la-­bi-­la ‘very big’ tuz-­la ‘cold’ → tuz-­la-­duz-­la ‘very cold’ kə-­la ‘long’ → kə-­la-­gə-­la ‘very long’

Another way to intensify a quality named by the corresponding verb in both dialects is to lengthen a vowel belonging either to the root, e.g. the Amur verb stems in (17), or, alternatively, to the suffix -­la-­, e.g. the Amur participle ŋar-­lā ‘very fat’ in (18): (17)

q‘av-­ ‘be hot’ → q‘āv-­ ‘be very hot’

(18) Ŋar-­lā-­ləγi tilv-­ət-­ŋan p‘řə-­řa-­d̹. fat-­ perm-­salmon autumn-­foc-­cvb.sim/ant come-­ iter.3sg-­ind ‘The very fat salmon comes exactly in autumn.’ (Panfilov 1962, 13) One more phonological marker of augmentivisation in both dialects is the raising of the vowel a to ə or i and its further lengthening into ə- or ī respectively.8 The vowel a can be a part of the verb root, e.g. the Amur participle mə-ţka ‘very small’ in (19), which is derived from maţka-­‘be small’, or that of the suffix -­la-­, e.g. the Amur finite verb k‘e-­lī-­d̹ ‘[he] is very thin’ in (20): (19) Hoʁoř hemañχ mə-ţka-­vəñ-­ţ‘im-­ra. then.3sg old.woman be.very.small-­copper-­put.on-­coord.3sg ‘Then an old woman put on a very small copper.’ (Panfilov 1965, 83) (20)

If mu-­inə-­ke k‘e-­lī-­d̹. he die-­ desid-­cvb.dur be.thin-­perm-­ind ‘Because he is ill (lit. is going to die), [he] is very thin.’ (Otaina 1978, 66)

Besides the means described below, both dialects make use of augmentative suffixation. In the Amur dialect, the augmentative suffix -­kar-­can be taken only by six very frequent qualitative verbs indicating various parameters of an object, mostly in the attributive function (Otaina 1978, 57). All of them carry the suffix -­la-­, which remains in the augmentative in a reduced or full form: (21)

ver-­la kə-­la pi-­la to-­la u-­la ŋar-­la

‘wide’ ‘long’ ‘big’ ‘thick’ ‘high’ ‘fat’

→ → → → → →

ver-­l-­kar kə-­l-­kar pi-­l-­kar to-­l-­kar u-­l-­kar ŋar-­la-­kar

‘very wide’ ‘very long’ ‘very big’ ‘very thick’ ‘very high’ ‘very fat’

The corresponding suffix -­kař-­is irregularly used in the East-­Sakhalin dialect as well: (22)

poţur-­la ‘beautiful’



poţur-­l-­kař

‘very beautiful’

Another suffix employed in augmentivisation in the Amur dialect is -­γәt-­/-­kәt-­, which also expresses completive and distributive meanings.9 The suffix renders the intensity of quality

Nivkh 283 or action when used with verbs indicating the quality of food (‘be unripe’, ‘be raw’, ‘be sour’ etc.) or visually perceived properties (‘be dirty’, ‘be shallow’, ‘be firm’ etc.): (23) Als qala-­γәt-­ţ. berry be.unripe-­ aug/compl-­ind ‘Berries are quite unripe. ’ (Otaina 1978, 53) Apart from other augmentative markers that are used only with qualitative verbs, this suffix also attaches to stative and processive verbs, e.g. the finite verbs ţ‘iŋr-­γәt-­ta ‘[I] am suffering much’ in (24) and vi-­γәt-­ţ ‘[he] went away [finally]’ in (25): (24) Ñi pañ-­d̹ ţ‘iŋr-­γәt-­ta. I grow.up-­ind suffer-­aug/compl-­coord.1sg ‘I am growing up, [I] am suffering much.’ (Panfilov 1965, 71) (25)

If hә-­dav-­ux vi-­γәt-­ţ. she or he that-­house-­abl go-­aug/compl-­ind ‘He went [away] from that house [finally].’ (Saveljeva and Taksami 1970, 54)

In the East-­ Sakhalin dialect, augmentation is marked mainly by a very productive suffix, -­katn-­, that attaches directly to the qualitative verb root (without the suffix -­la-­). The suffix is compatible with virtually any qualitative verb10 in different syntactic functions, e.g. the verb stems in (26) and the finite verb with the suffix -­katn-­ in (27): (26)

um-­ toŋγ-­ k‘әr-­

(27)

Ҳoi-­zoŋř hor-­katn-­d. taimen-­head be.tasty-­aug-­ind ‘Taimen head is very tasty.’ (fieldwork 2000, 7: 5)



‘be angry’ → um-­katn-­ ‘be strong’ → toŋγ-­katn-­ ‘be hungry’ → k‘әr-­katn-­

‘be very angry’ ‘be very strong’ ‘be very hungry’

The augmentative markers listed above (except for the Amur suffix -­γәt-­/-­kәt-­ and the East-­Sakhalin suffix -­katn-­) may coalesce within a single verb for the purpose of further intensification of the verb meaning. The following eleven combinations of markers have been attested in the Amur dialect participles: (i) Voicing and reduplication: (28)

per-­la pi-­la

‘heavy’ → ‘big’ →

ber-­la-­ber-­la bi-­la-­bi-­la

‘very heavy’ ‘very big’

(ii) Voicing and a-­raising: (29)

pi-­la

‘big’



bi-­lə

‘very big’

(iii) Voicing and suffixation: (30)

pi-­la

‘big’



bi-­l-­kar ‘very big’

Ekaterina Gruzdeva

284 (iv) Suffixation and lengthening: (31)

to-­la

‘thick’ →

tō-­l-­kar ‘very thick’

(v) Suffixation and a-­raising: (32)

kə-­la

‘long’ →

kə-­l-­kər ‘very long’

(vi) Suffixation and reduplication: (33)

kə-­la to-­la

‘long’ → ‘thick’ →

kə-­l-­kar-­kə-­l-­kar ‘very long’ to-­l-­kar-­to-­l-­kar ‘very thick’

(vii) Voicing, a-­raising and reduplication: (34)

kə-­la

‘long’ →

gə-­lə-­gə-­lə

‘very long’

(viii) Voicing, suffixation and a-­raising: (35)

to-­la

‘thick’ →

do-­l-­kər ‘very thick’

(ix) Voicing, suffixation and reduplication: (36)

to-­la

‘thick’ →

do-­l-­kar-­do-­l-­kar ‘very thick’

(x) Suffixation, a-­raising and lengthening – either the root vowel, as in (37), or the suffix vowel, as in (38), is lengthened: (37)

pi-­la

‘big’



pī-­l-­kər ‘very big’

(38)

pi-­la

‘big’



pi-­l-­kə-r ‘very big’

(xi) Voicing, suffixation, a-­raising and lengthening – this combination gives the highest degree of intensification (Otaina 1978, 58): (39)

pi-­la

‘big’



bi-­l-­kə-r ‘very big’

4  Qualitative evaluation Qualitative evaluation makes use of verb root compounding,11 which is a common verb-­ formation strategy in Nivkh. The first root typically expresses some quality, state or process, whereas the second root adds an evaluative meaning to it. In the Amur dialect, both roots occur in the bare form, e.g. nə-­ur-­ [do-­be.good] ‘work well’, whereas in the East-­Sakhalin dialect the first root takes the suffix -­ŋ, which otherwise serves as a participial marker, e.g. mə-­ŋ-­ur-­ [hear-­ptcp-­be.good] ‘hear well’. The resulting complex verb behaves like a regular verb; it may undergo subsequent inflection and can be nominalised. It cannot, however, attach the suffix -­la-­. As can be seen from (40), the derivational pattern is semantically quite straightforward: the verb ur-­‘be good’ is used for amelioration, and the verb əki-­‘be bad’ for pejoration; cf. also East-­Sakhalin examples (41–2):

Nivkh 285 (40) a. amra-­ŋ-­ur-­ [taste-­ ptcp-­be.good] amra-­ŋ-­əki-­ [taste-­ ptcp-­be.bad]

‘taste well’ ‘taste bad’

b. t‘a-­ŋ-­ur-­ [breathe-­ ptcp-­be.good] t‘a-­ŋ-­əki-­ [breathe-­ ptcp-­be.bad]

‘breathe well’ ‘pant (lit. breathe badly)’

(41) Ţ‘χař pand-­ŋ-­ur-­d. tree grow.up-­ ptcp-­be.good-­ind ‘The tree grows up well.’ (fieldwork 1991, IV: 44) (42)

Huŋ-­iřkt amra-­ŋ-­əki-­d. that-­berry taste-­ptcp-­be.bad-­ind ‘Those berries taste bad.’ (fieldwork 2000, 7: 4)

With some verbs, positive evaluation may be achieved by using of the verb maŋg-­‘be strong’: (43)

Jaŋ p‘-­ot-­ŋ-­maŋg-­d. she or he refl-­sew-­ptcp-­be.strong-­ind ‘She sews well.’ (fieldwork 2000, 7: 32)

5 Conclusion Nivkh makes use of numerous morphophonological and morphosyntactic strategies for expressing different types of evaluation. Descriptive evaluation does not indicate the size of objects, but conveys the meanings of attenuation and intensification. This evaluative type is characterised by a large variety of word-­formation devices. By contrast, qualitative evaluation that refers to ameliorative and pejorative meanings has a single derivational pattern. A typologically interesting feature of Nivkh is a quite well-­developed mechanism of verbal evaluation with a lack of means for nominal evaluation. It should be noted, however, that descriptive evaluation pertains mostly to qualitative verbs that occur in the attributive function. Semantically, these verb forms are very close to adjectives, for which this type of evaluation is less rare cross-­linguistically.

Notes   1. In the examples, both lexical roots and grammatical morphemes of polysynthetic complexes are separated by a hyphen.   2. About fifty verbs denoting permanent qualities or a low degree of quality cannot take this suffix (Otaina 1978, 59).   3. Unless otherwise specified, the Amur examples are taken from Panfilov (1962, 1965), Otaina (1978) and Shiraishi (2006), whereas the East-­Sakhalin examples are from data collected during my fieldwork in 1989, 1991 and 2000.   4. The stative verb qo-­ ‘hurt’ may also attach the diminutive suffix -­jo-­, as in qo-­jo-­ʁar-­ ‘hurt a little’.   5. Reduplication is typically used for augmentivisation.  6. Voiced plosives usually do not occur word-­initially in the citation forms of native Nivkh words.

286

Ekaterina Gruzdeva

  7. The final root consonant may be absent from the reduplicated form, as in the first example in (14).   8. Without lengthening, the raising of a in the suffix -­la-­ expresses a diminutive meaning.   9. The corresponding East-­Sakhalin suffix -ʁar-­/-­χar apparently cannot serve as a marker of the augmentative meaning. 10. The suffix -­katn-­ also co-­occurs with the stative verb qo-­‘hurt’, as in qo-­katn-­ ‘hurt much’. 11. Or verb root serialisation in Mattissen’s terms (2003, 185–94).

15.11 Persian

Negar Davari Ardakani and Mahdiye Arvin

1 Introduction Persian (Indo-­European, Indo-­Iranian) is the national language of more than 70 million people in Iran and is also spoken in Tajikistan and Afghanistan with minor phonological, morphologic and rare syntactic differences. Old Persian was a typical inflected language, systematically cognate with its contemporaries Classical Greek and Sanskrit. By the late Middle Persian period, however the language had undergone a radical reduction of inflectional endings, thus assuming a closer resemblance to the analytical structure characteristic of New Persian. The lexical morphology, although reduced in number and variety of forms, expanded in scope and flexibility (Perry 2007, 977–9). Perry believes that Persian morphology cannot be meaningfully discussed without incidental appeal to syntax. He expresses the view that in New Persian, not only is today’s morphology yesterday’s syntax, but also today’s syntax keeps a firm hold on morphological categories and relations. Persian morphology is usually categorised into inflection and word formation. Eslami (2009, 9–13) formulates Persian inflection as follows: •  noun inflection: [noun] + [(pl. morpheme)] + (modifier clause marker)/(personal enclitics1 joined pronouns)/(indf mark (−y)) + (conjunctive enclitics) •  adjective inflection: [adj.] + (comparative morpheme)/(sup morpheme) + (conjunctive enclitics) •  verb inflection: [(negation morpheme)] + (sbjv and imp aspect morpheme)/ (continuous morpheme) + [v] + [(prf aspect)] + [(inflectional endings)] + [(personal enclitics)] • adverb inflection: [adv] + [(comparative adv morpheme)] • number inflection: [number] + [(cardinal and ordinal suffix)] Persian word formation occurs mainly under the two processes of derivation and ­compounding. Derivational morphological processes in Persian include suffixation, prefixation,2 only a single case of circumfixation and no infixation. Persian suffixes are considerably more frequent and more productive than its prefixes. As Persian is an (S)OV language, it maps morphosemantically onto a head-­final word system, thus exhibiting a marked preference for suffixation: (1)

Suffix > Prefix > Circumfix

Negar Davari Ardakani and Mahdiye Arvin

288

However, it should be noted that compounding is the main source of word formation in Persian. Different compounding patterns exist in Persian: compounding of two or more simple morphemes (free and bound) and compounding of two or more complex words. Persian compound words are structurally formed from all word categories (verbs, nouns, adjectives, adverbs and even prepositions). Persian compounds are divided into endocentric and exocentric classes. However, there are more of the former than of the latter, and among the endocentric compounds there are more head-­initial than head-­final ones (Tabataba’i 2012, 15–16). Some multi-­headed compounds are also found in Persian. Some other word-­formation processes in Persian only mentioned here are blending, back formation, truncation, acronymisation, conversion and borrowing.

2  Evaluative morphology Some concepts of evaluative morphology (EM) have been touched on in Persian grammar books through a brief mention of a few diminution, endearment and contempt affixes by Mogharrabi (1993), Sadeghi (1991–3), Keshani (1993), Tabataba’i (2003), Farshidvard (2007), Kalbasi (1992) and others. In the present study, 120,000 Persian words from the eight-­volume Sokhan Persian Dictionary3 were analysed in order to examine EM values and criteria in Persian. The first phase of data collection4 included selecting and documenting all words with some evaluative sense. In the next phase simple words were omitted and the remaining compounds, derived and very few inflected words were analysed (1,928 words in total). In the next stage, the appropriate value label for each word was selected and EM criteria were investigated. 2.1 Values The values found in Persian derived, inflected and compound words are as follows. 2.1.1  Diminution (BAD, GOOD, NEUT) Physical diminution Diminution here is delimited to physical size and age of objects (including animate and inanimate). Some diminutive markers5 change the category of the resulting word (4) and some do not (3). The resulting words normally relate to GOOD quality and most of them are used in both formal and informal contexts. (2) pestân-­ak nipple-­ dim ‘baby’s dummy’ (3)



doxtar-­bače girl-­child.dim ‘little girl’

(4) riz-­naqš small.dim-­body ‘of small body size’

Persian 289 Qualitative diminution: contempt (BAD) Diminution can indicate someone or something that is unimportant and does not deserve respect. (5) mard-­ak(e) man-­ dim.con ‘weird man’ (6)

javân-­ak young male.n-­dim.con ‘young naive boy’

(7) tefl-­aki child-­ dim.con ‘the poor little dear’ (8)

pir-­i old person-­dim.con ‘the old unpleasant person’

(9)

zer~zer-­u redundant useless talk-­con ‘a talkative person’

(10) bi-­manteq priv.mark.con-­logic ‘illogical’ (11)



dahan-­gošâd mouth-­wide.open.con ‘a talkative person who uncovers secrets to other people/running mouth’

Many words are involved in the process of producing contempt words in Persian, in both the formal and informal varieties of the language. Four suffixes, -­ak, -­aki, -­i and -­u, and one prefix, bi-­, are productive in this process. It should be mentioned that the suffix -­ak is considered to be a multi-­value suffix, used in diminution (contempt and reduction) and endearment. In addition, the prefix bi-­is an opposite marker that can produce veneration words by being added to negative bases. Reduction Reduction markers are different from the diminutive ones as they are applied not to animate bases but rather to abstract ones, producing adjectives used to describe people. An exception for the -­ak adverb marker applies to comparative adverbs and adjectives that are semantically related to time (14) and quantity (15). (12) biš-­tar-­ak a.lot.adv-­compar.mark-­rdc ‘a little bit more’

Negar Davari Ardakani and Mahdiye Arvin

290 (13) bad-­ak bad-­ rdc ‘not so bad’ (14) tӑze-­kӑr new.rdc-­work ‘beginner’

(15) kam-­kâr little.rdc-­work ‘a person who does not work enough’ 2.1.2 Augmentation Physical augmentation Augmentation markers usually apply on nouns, but the resulting compounds might be adjectives or nouns, and therefore a change of category is observed in the augmentation evaluative process. Moreover, the resulting nouns usually denote physical objects, while adjective augmentative markers which change the category of the base from noun into adjective could refer to animates or inanimates. (16) dorošt-­andâm big.aug-­body ‘stocky, big-­bodied’ Qualitative augmentation In some contexts -­ak and -­aki convey the meaning of dearness, sympathy and diminution at the same time. Therefore, it would be more accurate to allocate an extra meaning of pathos to -­ak and -­aki, as they are practically used for expressing pathos more than the other two meanings. (17) doxtar-­ak-­am girl-­ end-­mine ‘my dear (little) girl’ (18) delbar-­ak beloved/lovable-­ end ‘dear beloved’ (19) tefl-­aki child-­ end ‘the dear (little) child/person’

(20)

aziz-­jun dear-­dear.end ‘dear grandma’

Persian 291 (21)



bâbâ-­jun-­i father-­dear-­end ‘my dear dear father’

Veneration, honorific and appreciation (GOOD and formal) Appreciation, veneration and honorifics are different from endearment because they are formal. However, the three words čašmodelsir, čašmodelpâk and âqâdâyee are used in informal contexts as well. Most examples show category change. (22) salim(o)-­nafs healthy.hon-­spirit ‘spiritually healthy’ (23)

čašmodel-­pâk eye.and.heart-­pure.app ‘someone who has pure eyes and heart’

(24)

čašmodel-­sir eye.and.heart-­full.app ‘a person who is free from greed’





(25) âqâ-­dâyee great.hon-­uncle ‘the great (honoured) uncle’ Promotion (GOOD, formal) This group of evaluative markers indicates something or someone with higher rank than others. Most cases refer to animates and belong to the formal variety of Persian, although a few examples are also used in informal contexts. An opposite class of deterioration or degradation could also be set up in order to contain the low ranks and rankings. However, we did not encounter such words in Persian. There were some that could fit in the ­diminution category. (26) abar-­qodrat prom-­power ‘superpower’ (27) sar-­parastâr head.prom-­nurse ‘head nurse’ Intensification (BAD, GOOD) Evaluative markers can express also an intensification of the meaning of the base word. (28) zer~zer-­u cheap.talk~cheap.talk-­ es.int ‘a person who talks a lot’

Negar Davari Ardakani and Mahdiye Arvin

292 (29) por-­basâmad full.int-­frequency ‘frequent’ (30)

sar-­(â)~sar head-­(interfix)~head.int ‘all over (from head to toe)’

The morpheme por could attach to some bases producing negative meaning, such as porxor ‘somebody who eats a lot’. 2.2  Evaluative morphology criteria 2.2.1  Change of base semantics Persian evaluative markers undergo slight semantic changes relating to the values they bring with them to the word. This change could be predicted from the term ‘evaluative’. It is taken for granted that any relation to evaluation deals with meaning change, and as a matter of fact evaluation could be considered as a morphosemantic area of investigation. (31) zer~zer-­u pej cheap.talk~cheap.talk-­ ‘a person who speaks long and redundantly’ (32)



jenâb-­âli you-­excellent.ven ‘your majesty’

2.2.2  Successive application of two or more evaluative rules Some rare examples of double application of EM processes of different types (a prefix and a suffix/an affix and a word, an affix/a word and repetition) were observed. Sometimes the values brought to the word are not in one single value direction. For example:6 (33) biš-­tar-­ak a.lot.adv-­compar.mark-­es.dim(a.bit) ‘a bit more’ 2.2.3  Change of syntactic category of the base There are some evaluative processes and state markers in Persian which change the category of base: (34) rafiq-­bâz friend-­ int ‘a (male) person excessively attached to friends’ 2.2.4 Preservation of at least one of the morphosyntactic feature specifications of the base and change of other features In a few examples, EM processes simultaneously change the category of the base and some characteristics such as gender:

Persian 293 (35) zan-­bâre woman.f-­int ‘a (male) person with excessive desire to be with women, womaniser’

3 Conclusion 3.1  Domain of evaluative morphology EM in Persian is not limited to derivation and inflection; it is represented in the whole system of morphology including compounding. As Stump (1993) puts it, EM is a way of merging evaluation with derivation and compounding, which essentially i­ntroduces semantics into morphology. Therefore, when we talk about ­ EM, we are in the domain of morphosemantics. As a matter of fact, terms such as ­‘inflectional ­morphology’ and ­‘derivational morphology’ refer to essentially pure morphological categorisations. However, categorising the realm of morphology with reference to semantic measures such as evaluation gives rise to a semantic or, at best, a morphosemantic category which cannot be considered parallel to the other morphological categorisations. It should also be noted that the introduction of this third category of affixation opens the way for other semantic categories such as process morphology, state morphology, positive morphology, negative morphology and, more generally, mood morphology, which refers to the completeness or incompleteness of action. Scrutinising the whole realm by investigating EM uncovered a few facts about widely accepted morphological distinctions such as inflection and derivation. It seems that inflection in Persian results in new word formation or new meaning creation (33). Moreover, looking at Persian morphology in terms of evaluation uncovers some misinterpretations of Persian inflection regarding adjectives: comparative and superlative affixes are usually considered as inflectional suffixes, although they obviously convey meanings related to BIG/SMALL and BAD/GOOD. Generally the concepts of comparison, completeness/perfection (which is expressed in the inflectional morphology of languages by the grammatical label of aspect) and even in some ways mood are essentially evaluative. 3.2  Morphologic evaluative characteristics Persian morphologic evaluative characteristics are as follows: • There is a preference for applying an EM process to bases from different parts of speech (noun, adjective, verb and adverb) in the following order: (36) Verb < Adverb < Adjective < Noun • T  wo different EM processes may occur successively in the pair forms of prefix + base + suffix, base + inflectional suffix + derivational suffix, word + base + suffix, prefix + base + word, and word + base + word. • Evaluative processes may change the category of their base. • Although they usually preserve some morphosyntactic characteristics of their base, these processes change the gender reference and part of speech of some bases.

294

Negar Davari Ardakani and Mahdiye Arvin

• Evaluative processes change the meaning of their base and a hyponymy relationship usually exists between the base and the resulting word. 3.3 Values The different values found in Persian morphology could be divided into two main categories of quantitative and qualitative values. Although the two poles SMALL and BIG originally belong to the quantitative category, and similarly the two poles BAD and GOOD belong to the qualitative category, they can also be interpreted metaphorically and figuratively. On the other hand, the combination of SMALL and BIG with BAD or GOOD gives rise to new complex evaluative meanings. The values can also be looked at from the binary perspective of state versus process. The cross-­combination of the four values which stand for the probable combination of quality and quantity gives rise to four other combinatorial values. Looking at the system of values in this way may help with the reorganisation of the value concepts and their terminology accordingly. A quick count of the number of terms used for labelling the values takes us to twenty-­four terms,7 many of them synonyms or near-­synonyms which have to be defined accurately. 3.4  The most frequent processes The few studies on evaluative sound symbolism support the idea that evaluative discussion is essential in semantics and not merely a domain parallel to derivational and inflectional morphology. In fact, according to the Persian data, the most frequent and productive affixation and compounding processes were used to express contempt and form pejoratives, and not to express the small size of something, a small quantity of something or a young one.

Acknowledgements Many thanks to Dr Sofia A. Koutlaki (Quran and Hadith University, Iran) as the English editor and guideline checker of this chapter.

Notes 1. Persian enclitics are as follows: /-­i/ indefinite marker, verbal enclitics such as present forms of budan ‘to be’: -­am, -­I, -­ast, -­im, -­id, -­and, object personal pronouns, possessive pronouns and complement of preposition: -­am, -­at, -­aš, -­emân, -­etân and -­ešân (Kalbasi 1992, 32). 2. Persian has about sixty-­five suffixes and eleven prefixes (Tabataba’i 2007). 3. Total number of words: 120,000; main entries: 75,000; auxiliary entries: 45,000; documented samples: 160,000; examples: 1,000. 4. Data were collected in Shahid Beheshti University by Negar Davari Ardakani, Mahdiye Arvin, Leila Tamini, Nasrin Sabbaghan, Fateme Mahdavi, Marziye Yazdani, Maryam Mirshahnasab, Maryam Asgharnia, Zinat Sarabeigi, Mona Shahab, Parastu Mesgariyan, Akram Mirzayee and Negar ilqami, whose contributions we gratefully acknowledge. 5. Diminutive markers include -­ak, -­aki, -­če, -­i, -­ule and -­ize. Examples are: xord-­ak small-­ dim ‘very small’

Persian 295

heyvun-­aki animal-­ dim ‘the poor little dear’ bӑq-­če garden-­ dim ‘small garden’ muš-­i mouse-­ dim ‘little mouse’

zang-­ule ringing-­ dim ‘urceolate’ pӑk-­ize clean-­ dim ‘clean’ 6. biš-­tar-­ak (33) is the single example of internal inflection in Persian. Interestingly, this unique case ends with an evaluative derivational suffix. The two Persian inflectional suffixes -­tar and -­tarin are comparative and superlative markers. In this example, -­ak as a derivational suffix is abnormally placed external to an inflectional suffix, i.e. -­tar. In other words, the traditional classification of the two comparative and superlative suffixes as inflectional does not match with the definition of inflection. While they are semantically evaluative, they appear to have both inflectional and derivational morphological nature. 7. Augmentative, diminutive, endearment, affection, authenticity, admiration, politeness, modesty, honorifics, intensification, excessiveness, superlatives, contempt, disdain, pejoratives, derogation, member, serious, insignificant, gender, number, person, possibility and ability.

15.12 Slovak

Renáta Gregová

1 Introduction Slovak is an Indo-­European language belonging to the West Slavic branch of the Slavic genus. Like the other Slavic languages, Slovak is a fusional language with a relatively extensive evaluative morphology. Nevertheless, linguistic research on diminutives and augmentatives has been somehow neglected in Slovak linguistics. The first systematic account of Slovak diminutives and augmentatives was presented by Jozef Štolc (1958). Although the author mixed up standard and dialect expressions, his study can be seen as a valuable synchronic–diachronic survey that maps Slovak evaluative constructions from the time of Ľudovít Štúr1 up to the second half of the 1950s. From the beginning of the 1960s a few authors have devoted their research attention to diminutives and augmentatives (for example, Horecký 1959; 1971; Buffa 1988; Trnková 1991; Zelinková 1993; Furdík 2004). Some brief notes concerning this issue may be found in handbooks of grammar, lexicology or stylistics (for example, Dvonč et al. 1966; Ondrus, Horecký and Furdík 1980; Mistrík 1985). However, the inaccurate delimitation of the formal and semantic aspect of diminutives (and rarely augmentatives), as well as the incorrect interpretation of the morphemic and word-­formative structure of these expressions, has led to misleading conclusions and a not very precise delimitation of diminutive markers in Slovak. For example, Zelinková (1993) mentioned that -­ka is the most productive feminine diminutive suffix: jama ‘hole’ > jam-­ka ‘hole-­dim’. But the morphemic structure of the base word jama ‘hole’ is jam-­a, where jam-­ is a root morpheme and -­a is a grammatical morpheme (Sokolová, Moško, Šimon and Benko 1999). The morphemic structure of the diminutive form jamka ‘hole-­dim’ is jam-­k-­a, where jam-­is a root morpheme, -­a is a grammatical morpheme and only the segment -­k-­ is a diminutive marker. Similarly, it is said that verbal diminutives are formed by a suffix -­kať: driemať ‘to doze > driem-­kať ‘to doze-­dim’. The morphemic structure of these two words is driem-­a-­ť and driem-­k-­a-­ť, where driem-­is a root morpheme, -­a-­ is the so-­called thematic morph,2 -­ť is a form morpheme3 and the segment -­k-­ is a diminutive suffix.4 The segmentations jam-­ka and driem-­kať represent the word-­formative structure of the given words. Word-­formative structure is always binary. The segmentations jam-­k-­a and driem-­k-­a-­ť reflect the morphemic structure, which is linear (Ondruš and Sabol 1987, 167). Diminutivisation and augmentivisation in Slovak are characterised as morphological processes (Štolc 1958, 77), and thus the morphemic structure of words will be used in this chapter. This means that diminutive and augmentative suffixes will be given without grammatical endings.

Slovak 297

2 Diminutives In Slovak, diminutives can be found in the word classes of nouns, adjectives, verbs and adverbs. Diminutiveness represents one of the most important modificational onomasiological categories in the Slovak language. The modificational attribute of diminutiveness can be of a quantitative or qualitative nature (Zelinková 1993, 31): quantitative diminutives express the smallness of the object or the quality named; qualitative diminutive ­suffixes convey incorporated emotional meaning that can be either ameliorative or p­ ejorative. However, most Slovak diminutives have both semantic values, that is, quantitative (descriptive) and qualitative. For example, in ruka ‘hand’ > rúčka ‘hand-­dim’, the diminutive form of the base word ruka ‘hand’ can indicate not only a little hand, but also a pretty and/or dexterous hand (Hayeková 1956, 76). The boundary between the quantitative and qualitative character of a diminutive is usually very ambiguous and can be fully uncovered only in context (Zelinková 1993, 32). In addition to the positive emotional colouring, some diminutives have an ironic meaning. These are especially those formed from words that name intellectual professions and military ranks: for example, učiteľ ‘teacher’ > učitel-­ík ‘teacher-­dim’, or vojak ‘soldier’ > vojač-­ik ‘soldier-­dim’. The aim of this irony, which usually expresses a kind of professional mediocrity, is to derogate a certain social status. 2.1  Diminutive nouns Diminutive nouns are formed by adding a diminutive suffix to nouns of all three genders: masculine, feminine and neuter. Diminutives usually preserve the gender of the original noun (Horecký 1971, 163). 2.1.1  Masculine diminutive nouns Comparison of several sources (Hayeková 1956; Horecký 1971; Trnková 1991; Zelinková 1993; Sokolová et al. 1999; Gregová 2010) has shown that there are seven productive masculine diminutive suffixes in present-­day standard Slovak: -­ík-­/-­ik-­,5 -­ok-­, -­ček-­/-­tek-­, -­ek-­, -­k-­, -­ušk-­and -­ink-­. These suffixes can be attached to animate and inanimate masculine nouns, as well as to personal names (Zelinková 1993): (1) vláč-­ik-­Ø train-­ dim-­m.nom.sg ‘little train’ (2) líst-­ok-­Ø leaf-­ dim-­m.nom.sg ‘little leaf’ (3) dom-­ček-­Ø house-­ dim-­m.nom.sg ‘little house’

298

Renáta Gregová

(4) noš-­tek-­Ø nose-­ dim-­m.nom.sg ‘little nose’ (5) slov-­níč-­ek-­Ø word-­ der.sfx-­dim-­m.nom.sg ‘little dictionary’ (6) Marcel-­k-­o dim-­m.nom.sg Marcel-­ ‘dear Marcel’ (7) tat-­ušk-­o dad-­ dim-­m.nom.sg ‘daddy’ (8) hav-­ink-­o doggie-­ dim-­m.nom.sg ‘little doggie’ 2.1.2  Feminine diminutive nouns In standard Slovak, there are six productive feminine diminutive suffixes: -­k-­, -­ičk-­, -­ôčk-­/-­očk-­, -­enk-­/-­ienk-­, -­ink-­ and -­ušk-­, which form diminutives from all types of feminine nouns, that is from common nouns and proper nouns, and from concrete nouns and abstract nouns, as illustrated by the following examples: (9) žien-­k-­a woman-­dim-­f.nom.sg ‘little woman’ (10) Alic-­k-­a Alic-­ dim-­f.nom.sg ‘dear Alice’ (11) krav-­ičk-­a cow-­dim-­f.nom.sg ‘little cow’ (12) guľ-­ôčk-­a ball-­ dim-­f.nom.sg ‘little ball’ (13) dcér-­enk-­a daughter-­ dim-­f.nom.sg ‘little daughter’ (14) slobod-­ienk-­a freedom-­dim-­f.nom.sg ‘little freedom’

Slovak 299 (15) duš-­ink-­a soul-­dim-­f.nom.sg ‘little soul’ (16) Han-­ušk-­a Han-­dim-­f.nom.sg ‘dear Hana’ 2.1.3  Neuter diminutive nouns There are five productive neuter diminutive suffixes in the present-­day standard Slovak. The suffixes -­k-­, -­íčk-­/-­ičk-­, -­očk-­/-­ôčk-­ and -­úšk-­ form diminutives of things, tools and abstract nouns, and the suffix -­atk-­/-­iatk-­/ätk-­ is used to create diminutive forms from the nouns that denote young animals: (17) uš-­k-­o dim-­ntr.nom.sg ear-­ ‘little ear’ (18) zdrav-­íčk-­o health-­dim-­ntr.nom.sg ‘little health’ (19) kladiv-­ôčk-­o hammer-­ dim-­ntr.nom.sg ‘little hammer’ (20) vrec-­úšk-­o sack-­ dim-­ntr.nom.sg ‘little sack’ (21) kur-­iatk-­o chicken-­ dim-­ntr.nom.sg ‘young chicken/little adult chicken’ (22) hús-­atk-­o gosling-­ dim-­ntr.nom.sg ‘little gosling’ 2.2  Diminutive adjectives The most productive adjectival diminutive suffix in Slovak is -­učk-­: (23) pekn-­učk-­ý pretty-­ dim-­adj.nom.sg ‘dinky’ Rarely, the suffixes -­ičk-­and -­ink-­ may occur:

300

Renáta Gregová

(24) mal-­ičk-­ý little/small-­dim-­adj.nom.sg ‘tiny, midget’ but also: (25) mal-­ink-­ý little/small-­dim-­adj.nom.sg ‘tiny, midget’ A special type of adjectival diminutive suffix is represented by the suffixes -­av-­, -­(k)ast-­ and -­(k)ist-­, which express the very low intensity of a characteristic encoded in the basic adjective (Zelinková 1993, 39), that is, attenuation: (26) modr-­av-­ý blue-­dim-­adj.nom.sg ‘bluish’ (27) siv-­ast-­ý grey-­dim-­adj.nom.sg ‘greyish’ (28) zlat-­ist-­ý gold-­dim-­adj.nom.sg ‘golden’ 2.3  Diminutive verbs Diminutive verbs are produced by adding the suffix -­k-­ to the verbal word stem; they express the lower extent or lower intensity of an activity indicated by the word stem to which the given suffix is attached: (29) bež-­k-­ať run-­ dim-­inf ‘to run a little, to run slowly’ (30) čit-­k-­ať read-­ dim-­inf ‘to read a little’ 2.4  Diminutive adverbs Adverbial diminutives occur in the group of adverbs that express circumstances and manner. Adverbs that indicate circumstances were originally derived from nouns and/or verbs, and the diminutive marker -­k-­ joins the base word of this type in order to create a diminutive (Zelinková 1993, 40):

Slovak 301 (31) dom-­k-­ov home-­ dim-­adv ‘to one’s little lovely home’ (32) tíš-­k-­o quietly-­ dim-­adv ‘very quietly’ Adverbs that indicate manner are derived from adjectives and, as with adjectives, the diminutives are formed by adding the suffixes -­ičk-­, -­učk-­ or -­ink-­ to an adverbial base word (Zelinková 1993, 39): (33) pomal-­ičk-­y dim-­adv slow.ADV-­ ‘very slowly’ (34) hlúp-­učk-­o silly.ADV-­ dim-­adv ‘very foolishly’ (35) mál-­ink-­o little.ADV-­ dim-­adv ‘a little’

3  Primary and secondary diminutives All diminutives mentioned so far are the so-­called primary diminutives. Primary diminutiveness means that a diminutive suffix is attached to a word whose semantics are without any attribute of diminutiveness; that is, a diminutive marker is added to a neutral base word. If a diminutive suffix is joined to an already existing diminutive, this is called secondary diminutiveness6 (Trnková 1991, 91). In Slovak, unlike in the other Slavic languages (for example Czech), it is not possible to make a clear difference between primary and secondary diminutives due to the polyfunctional character of affixes (Trnková 1991, 91). Many substantives have the form of primary diminutives, but there is no evaluative aspect in their meaning. For example, ceruzka ‘pencil’ is not a diminutive, but the word žienka ‘woman-­dim’ is, although the morphemic structure of both words is identical: root morpheme – suffix k – grammatical morpheme. The -­k-­ in the word ceruzka ‘pencil’ is a derivative morpheme, and the suffix -­k-­ in žienka is a diminutive marker. The diminutive form of ceruzka is ceruz-­ôčka, where -­ôč(k)-­7 is a primary diminutive element. However, the same diminutive suffix can also function as a secondary one when added to a primary diminutive, as illustrated by the following example: guľa ‘ball’ > guľ-­ka ‘ball-­dim’ > guľ-­ ôč-­ka ‘ball-­dim-­dim’. The other factor that conceals the boundary between primary and secondary diminutive suffixes is the ability of Slovak diminutive markers to ‘chain’. In other words, Slovak diminutive morphemes show the property of consecutive application (see also Stump 1993, 4). Štolc (1958, 74) referred to the process of diminutivisation as a quantitative and/or qualitative change of meaning in an arithmetic operation that is theoretically unlimited. This means that there is a theoretically unrestricted possibility of modifying – ­quantitatively

Renáta Gregová

302

and/or qualitatively – the meaning of a given base word. The formal means of this multiple modification (or simply recursiveness) in Slovak is the repetition of a diminutive marker or the combination of several different diminutive suffixes (Štolc 1958, 74): malý ‘little’ > mal-­inký ‘little-­dim’ > mal-­il-­inký ‘little-­dim-­dim’ or mama ‘mum’ > mam-­ka ‘mum-­dim’ > mam-­ič-­ka ‘mum-­dim-­dim’ > mam-­ič-­en-­ka ‘mum-­dim-­dim-­dim’. The above-­mentioned phenomenon is closely connected with another important feature of diminutivisation in Slovak, which is the possibility of attaching different diminutive suffixes to one and the same base word; for example, dom ‘house’ > dom-­ík ‘house-­dim’ or dom-­ček ‘house-­dim’; žena ‘woman’ > žien-­ka ‘woman-­dim’ or žen-­ička ‘woman-­dim’. The semantics of these diminutives are usually identical or there is only a slight ­qualitative shift in meaning.

4  Phonological or morphophonological changes in diminutives The process of diminutivisation in Slovak is accompanied by various phonological and morphophonological changes in word stems to which a diminutive suffix is attached. The most frequent phonological process is the change of the stem-­final alveolar consonants t, d, n, l into post-­alveolar ť, ď, ň, ľ [tj, dj, nj, lj] respectively when a diminutive suffix is added: (36)

t/ť d/ď n/ň l/ľ

cesta ‘road’ > cest-­ička [cest j-­ička] ‘road-­dim’ had ‘snake’ > had-­ík [had j-­ík] ‘snake-­dim’ slon ‘elephant’ > slon-­ík [slon j-­ík] ‘elephant-­dim’ metla ‘broom’ > metl-­ička [metl j-­ička] ‘broom-­dim’

As to the morphophonological processes, the alternations of consonants that differ by more than one distinctive feature occur at the end of the base word. This type of alternation also includes the replacement of one consonant with a two-­consonant cluster (for details see Sabol 1989, 340). The alternating consonants are as follows (see also Horecký 1971; Zelinková 1993): (37) k/č žiak ‘pupil’ > žiač-­ik ‘pupil-­dim’ h/ž stuha ‘ribbon’ > stuž-­ka ‘ribbon-­dim’ c/č zajac ‘rabbit’ > zajač-­ik ‘rabbit-­dim’ ch/š kožuch ‘fur coat’ > kožúš-­ok ‘fur coat-­dim’ h/žť krčah ‘mug’ > krčiažt-­ek ‘mug-­dim’ ch/šť mech ‘sack’ > mešt-­ek ‘sack-­dim’ s/šť nos ‘nose’ > nošt-­ek ‘nose-­dim’ z/žť obraz ‘picture’ > obražt-­ek ‘picture-­dim’ ž/žť koža ‘skin’ > kožt-­ička ‘skin-­dim’ žk/žť slúžka ‘housemaid’ > slúžt-­ička ‘housemaid-­dim’ šk/šť líška ‘fox’ > líšt-­ička ‘fox-­dim’ sk/šť pysk ‘lips’ > pyšt-­ek ‘lips-­dim’ zk/žť bozk ‘kiss’> božt-­ek ‘kiss-­dim’ st/šť prst ‘finger’ > pršt-­ek ‘finger-­dim’ Alternations of consonants that differ by one distinctive feature only can also be found. These are the alternations:

Slovak 303 (38)

ň/n ť/t

kuchyňa ‘kitchen’ > kuchyn-­ka ‘kitchen-­dim’ smrť ‘death’ > smrt-­ka ‘death-­dim’

Quantitative vocalic alternations when vowels alternating in the word stem differ by their length, or by length and the distinctive glide feature (for details see Sabol 1989), are also quite numerous in the word stems of diminutives. These are the alternations of a short monophthong and a long monophthong, for example: (39) a/á u/ú

hlas ‘voice’ > hlás-­ok ‘voice-­dim’ ruka ‘hand’ > rúč-­ka ‘hand-­dim’

or the alternations of a long monophthong and a short monophthong, for example: (40) í/i

čítať ‘to read’ > čit-­kať ‘to read-­dim’

or the alternations of a short monophthong and a diphthong, for example: (41) e/ie

žena ‘woman’ > žien-­ka ‘woman-­dim’

or the alternations of a diphthong and a short monophthong, for example: (42) ô/o

kôň ‘horse’ > kon-­ík ‘horse-­dim’

5 Augmentatives In standard Slovak, augmentatives can be found only in the word classes of nouns and adjectives. 5.1  Augmentative nouns There is only one augmentative suffix for nouns of all genders: masculine, feminine and neuter. It is the suffix -­isk-­o. This means that all augmentatives are of neuter gender, despite the gender of the base word, as illustrated in the following examples: (43) dub-­isk-­o oak.m-­aug-­ntr.nom.sg ‘big oak’ (44) žen-­isk-­o woman.f-­aug-­ntr.nom.sg ‘big woman’ (45) hniezd-­isk-­o nest.ntr-­aug-­ntr.nom.sg ‘big nest’ The augmentative marker -­isk-­o triggers no alternations in the word stems to which it is attached; however, the vowel -­i-­ from this suffix causes the palatalisation of the

304

Renáta Gregová

stem-­final consonants t, d, n, l, which are then pronounced as [tj, dj, nj, lj] respectively. For example, the sound form of the above-­mentioned example hniezd-­isko ‘nest-­aug’ is [dj] – [hnjiezdjisko]. 5.2  Augmentative adjectives In Slovak, there are two augmentative adjective suffixes, -­ánsk-­ and -­izn-­. For example: (46) širok-­ánsk-­y/široč-­izn-­ý aug-­adj.nom.sg wide-­ ‘extremely wide’ The augmentative marker -­izn-­ induces the alternation of the stem-­final velar consonant [k] into a palato-­alveolar č [ʧ], as exemplified by the adjective široký ‘wide’ > široč-­izný ‘wide-­aug’.

6 Conclusion Comparison of evaluative morphology in Slovak to that of the other Slavic languages (see Gregová 2010) has shown that all Slavic languages have rich evaluative morphology. In all these languages, the category of diminutiveness is much more common than the category of augmentativeness. The evaluative marker is usually a suffix. Diminution affects nouns, verbs, adjectives and adverbs. Nominal diminutive suffixes prevail over verbal, adjectival and adverbial diminutive elements. Augmentation is the domain of nouns, less frequently adjectives and rarely verbs and adverbs. It can be said that, as in Slovak, the evaluative (and especially diminutive) morphological process in all Slavic languages is a rather complex phenomenon that enables a diminutive marker to be joined to an already existing diminutive form. Diminutive markers trigger miscellaneous phonemic and morphophonemic changes in base words.

Notes 1. Ľudovít Štúr (1815–56) was a Slovak politician, writer and national thinker. He was one of the leading personalities of the Slovak national movement in the nineteenth century. He is also the codifier of the Slovak language (1843) that has developed into contemporary standard Slovak. 2. A thematic morph in Slovak is a vocalic or an empty element that has no meaning, and its function is to supplement or to extend the verbal stem (Ondruš and Sabol 1987, 166). 3. A form morpheme in Slovak is a special type of verbal morpheme. The form morpheme -­ť expresses the infinitive (Ondruš and Sabol 1987, 166). 4. The segment -­k-­ seems to be an infix. However, infixes are inserted inside their base (Bauer 2003), and this segment is attached after the base. In Slovak, a suffix is delimited as ‘a morph that occurs after a root morph and before a grammatical morph’ (Sokolová et al. 1999, 56). 5. In standard Slovak, the so-­called Rhythmic Law or the neutralisation of quantity says that there cannot be two long syllables one immediately after the other (see Sabol 1989). This is a progressive type of sound change, which means that when there is a long syllable at the end of the base word to which the suffix -­ík should be attached, the vowel in the suffix becomes short; for example, somár ‘donkey’ > somár-­ik ‘donkey-­dim’. The variations -­ôčk-­/-­očk-­, -­enk-­/-­ienk-­, -­íčk-­/-­ičk-­, -­očk-­/-­ôčk-­ and -­atk-­/-­iatko-­/-­ätk-­ (see Sections 2.1.2 and 2.1.3) are

Slovak 305 given according to the same principle and by the other phonological rules of the Slovak language. 6. This phenomenon is also known as a double diminutivisation or a double diminutive. However, Slavic linguists prefer the terms ‘secondary diminutiveness’ and ‘secondary diminutive’ (see, for example, Dunn and Kairov 2009). 7. The full form of this diminutive element is -­ôčk-­(see Section 2.1.2); however, the consonant -­k-­ becomes redundant in the process of secondary diminutivisation, which is why it is omitted.

15.13 Swedish

Arne Olofsson

1 Introduction Evaluative morphology defined as suffixation has virtually no place in present-­day Swedish word formation. The core of the treatment to follow here cannot but be very brief. However, in order to give some idea of how Swedish handles the semantics involved, the means actually in use will be exemplified and, to some extent, discussed under the heading of ‘Alternatives to suffixation’ (Section 6).

2  General status Swedish is one of the Scandinavian or Nordic languages and belongs to the North Germanic group of the Indo-­European family. In Sweden, there are well over nine million speakers of the language, and some 300,000 residents of Finland regard Swedish as their first language. Its closest relatives are Norwegian and Danish. Swedish and Norwegian exhibit a fairly high degree of mutual intelligibility in general, whereas Swedish and Danish are less mutually intelligible, particularly as regards the spoken language. Films and TV programmes in Danish and Norwegian are regularly subtitled in full when shown in Sweden; instructions and information on products for the Scandinavian market, on the other hand, often stop at giving some content words in more languages than one in a running text. Icelandic and Faroese, descendants of Norwegian, belong to the same group but are too different from Swedish for mutual comprehension without special training.

3  Selected typological characteristics Present-­day Swedish has features from different types. In nouns, there are no case distinctions other than the genitive, which is productively formed by an -­s suffix placed after any other inflectional suffixes. There can thus be an agglutinative chain of three inflectional suffixes after the stem of the noun: bil-­ar-­na-­s (stem ‘car’ + pl + def + gen). Present-­day Swedish has overt definite and (for count nouns in the singular) indefinite articles. The definite one takes the form of a suffix, but in noun phrases with a premodifier, definiteness is signalled twice, not only with the suffix but also with a determiner: den stora bil-­en ‘def big car-­def’. For more details, see SAG (1999) and Holmes and Hinchliffe (2008). As for verbs, most Swedish main verbs form their preterite and their past participle by means of suffixes, but Swedish still has its fair share of ablaut formations, both the classic types (for instance binda – band – bunden for the principal parts of the verb meaning ‘bind’)

Swedish 307 and mixtures. There are about a dozen modal/temporal auxiliaries, but in some cases there is a thin line between auxiliaries and main verbs because of ongoing grammaticalisation.

4  Some general aspects of the vocabulary The influx of loanwords reflects variations in cultural influence: Christian terminology based on Latin and Greek (early Middle Ages), Low German (late Middle Ages), High German (sixteenth century), French (seventeenth and eighteenth centuries) and English (nineteenth century and onwards). All through the centuries, there has been a trickle of Latin and Greek words imported via other languages. Loans include prefixes and suffixes, some of which have relevance for evaluative morphology, as will be demonstrated below. In word formation, Swedish shares with Danish, Norwegian and German the ability to form long and complex solid compounds. Like these three, but unlike English, where compounds may move from open to solid on the basis of currency, Swedish has no open compounds in the written standard form of the language. For further general information, see Malmgren (1994), Thorell (1981) and Söderbergh (1971).

5  Evaluative morphology In comparison with, for instance, Romance and Slavic languages, Swedish is extremely poor when it comes to expressing sizes and attitudes by means of suffixes. In present-­day Swedish, there are no suffixes that unambiguously create augmentatives, diminutives or hypocorisms. In some textbooks on Swedish word formation, these concepts are therefore not treated at all. 5.1 Diminutives Historically, some words for animals now ending in -­ing or -­ling were formed as diminutives in Old Swedish, for instance geting ‘wasp’ from get ‘goat’ and gässling ‘gosling’ from gås ‘goose’, mentioned by Söderbergh (1971); for these and other formations from the past, see Wessén (1971, 64). In present-­day Swedish, however, the suffix in question has practically no productivity in that function. The strongest candidate for a diminutive suffix in present-­day Swedish would be -­ett, given as the only example by Thorell (1981, 116) and described as ‘the closest Swedish gets to a diminutive suffix’ in the wording (my translation) of Malmgren (1994, 59). It is separable in some loanwords from French or Italian thanks to the fact that both the diminutive and the derivational base have been borrowed. A clear-­cut example is statyett, which is indeed just a small staty (statue). Slightly different examples are sandal – sandalett and opera – operett. As pointed out by, for instance, Dahl (2009, 218), diminutives tend to become lexicalised and thus different in that the suffix, although in a sense separable, does not (even in the source languages) uniquely indicate ‘reduced physical size’. A sandalett is more delicately designed, more elegant and more suitable for festive occasions, not smaller as a whole than a sandal. An operett is not ­necessarily shorter than an opera but distinguished by comic and light-­hearted content. The computer term diskett is a nice illustration of the non-­productivity of -­ett. The word is best seen as borrowed (from English) as a whole, because before the digital era, Swedish disk already had four very different senses (‘counter/bar’, ‘kitchenware just washed or about to be washed’, ‘intervertebral cartilage layer’ and ‘central part of a composite flower’). After borrowing the computer term disk, Swedish thus had five homonymous potential

308

Arne Olofsson

­derivational bases for a form ending in -­ett, which would then have been difficult to interpret. In English, the -­ette suffix is successfully used productively in a few words, the best-­known example of which is probably kitchenette, but in Swedish, there is no evidence that -­ett would be widely understood to signal diminution when added to an existing noun. 5.2 Augmentatives As for augmentatives, Italian -­one in the form it has taken in Swedish when borrowed via French (-­ong, with the spelling -­ng conventionally employed in Swedish to approximate French nasalisation) occurs in a number of loanwords (for instance balkong, kartong, salong), but in none of them can the suffix be meaningfully isolated as augmentative-­ forming, although the derivational bases are words in their own right. And it is definitely impossible to create a Swedish augmentative through the addition of -­ong to a native, i.e. non-­borrowed, noun. 5.3 Hypocorisms As with augmentatives and diminutives, the shortage of suffixes for hypocorisms is striking. There are some candidates (see below), but it is virtually impossible to draw the line between true hypocorisms and simple slang-­like changes that aim to signal informality, familiarity or intimacy.

6  Alternatives to suffixation In the virtual absence of evaluative suffixes, the semantics of augmentation and diminution is handled by prefixation, sometimes bordering on and originating in compounding. For augmentation, an extensive range of Greek/Latin prefixes is available and can be used productively, with borrowed and native derivational bases (cf. Ellegård and Olofsson 1995, 59, 63, 102, 112). Examples are hypersnabb (‘-­fast’), makromolekyl (‘-­molecule’), megastjärna (‘-­star’), maxikjol (‘-­skirt’), supermakt (‘-­power’). Some of the instances are partial or full calques from English. The native set, which is typically developed from compounding, which, as mentioned above, is always of the solid type in Swedish, contains adjectives which may become part of lexicalisation processes in contrast with their use as modifiers in noun phrases. The prototypical example is stor ‘big/great/large’. A stormarknad is a hypermarket, whereas a stor marknad is any big fair or marketplace. A storkök is a place for large-­scale cooking for a hospital or a school, whereas a stort kök is any spacious kitchen. Elements that are derived from nouns are also used, for instance jätte- (‘giant’), which can be prefixed to nouns, e.g. jättestad (‘-­city’), to adjectives, e.g. jätte­arg (‘-­angry’), and adverbs, e.g. jättefort (‘-­fast’). That jätte in such usage has undergone a change from a referring noun to a reinforcing prefix is nicely demonstrated by the fact that it can reinforce the adjective liten (‘small’). Similarly, bamse, ‘big, heavy and strong person or animal’, can be readily prefixed to nouns, adjectives and adverbs, particularly in children’s speech. For diminution, some Greek/Latin prefixes can be used productively, for instance mikro, as in mikrofilm, and mini, as in minibuss (cf. Ellegård and Olofsson 1995, 63). The Swedish adjective liten ‘small’ can also be used, but never in its uninflected form. A lillsemester, alternatively called minisemester (semester in Swedish is, surprisingly from an etymological point of view, the term for ‘holiday weeks with pay’), is a two-­or three-­day trip,

Swedish 309 t­ypically abroad. The plural of liten is heteroclitic and has the form små, and that form is used as a prefix regardless of the grammatical number of the noun. As with prefixed ­augmentatives, lexicalisation comes into play: A småhus is a one-­family home and actual size enters only marginally into the technical definition, which is rather based on the contrast with a block of flats. Less lexicalised are, for instance, småsten (‘-­stone’), småbarn (‘-­child’) and småfågel (‘-­bird’). For hypocorisms, there is no obvious set of affixes available. In children’s speech and parents’ baby talk, a foot (Swedish fot) may be referred to as fossing, with the otherwise generally unproductive diminutive suffix mentioned above combined with a change of the stem. For first names, there is a non-­segmental morphological pattern for such familiarising change towards a canonical disyllabic trochaic form: Bo becomes Bosse (to his friends) and Kristina becomes Kina, Stina or Tina (to hers). These changed forms are summarily described as hypocorisms by Holmes and Hinchliffe (2008, 213). The same thing tends to happen with terms for school subjects and activities: matematik becomes matte, and geografi becomes (or, rather, used to become) jogge. Another pattern, which works without regard to whether the noun is proper or common and whether it denotes a person or a thing, cuts off the end of a long noun, often a compound, and replaces the end with a suffix -­is (often described – without semantic requirements – as diminutive-­forming, e.g. in Josefsson 1997, 156–9, and in Wikipedia). Etymologically it is a Latin word ending taken over jestingly centuries ago in students’ slang. As to its function, it is reminiscent of English -­y/-­ie, which occupies a similar borderline position between diminutives and hypocorisms: aunty/auntie, birdie. Particularly in or via children’s speech, Swedish daghem (literally ‘day-­home’, i.e. ‘daycare centre’) has become dagis and vaktmästare ‘(school) janitor’ has become vaktis. The form grattis is nowadays far more frequent, at least in informal communication, than gratulation/er/ ‘congratulation/s/’. The suffix is also used to derive familiarising nouns from adjectives: kändis ‘a celebrity’ (känd is a past participle meaning ‘known’) and even from verbs: snackis ‘event currently much talked about’ (snacka is an informal word for ‘talk’). However, strict criteria for hypocorisms are not met by this or any other suffix.

15.14 Tatar

Fatma Şahan Güney

1 Introduction Tatar, also known as ‘Volga Tatar’ and ‘Kazan Tatar’, is one of the Kipchak languages of the Turkic family. The total number of speakers of Tatar is about 5.5 million to 6 million worldwide, most of them living in Tatarstan, Bashkortostan and Western Siberia. Tatar consists of three dialects, namely the western, the central and the Siberian. The literary language is based on the central dialect in terms of its phonetic features but draws most of its morphological features from the western dialect. From a typological point of view, Turkic languages, Tatar included, are syntactically left-­branching and have an SOV order. Modifiers of any type precede the head of the construction. No gender exists in Turkic and there is no agreement between the numeral and its head. Turkic languages are agglutinative, characterised by well-­developed vowel harmony and have a synthetic character. Inflectional and derivational processes are carried out mostly by bound morphemes. Besides its highly synthetic character, many linguistic categories, such as aspect and modality, are expressed through periphrastic constructions, which have not become synthetic forms yet. ‘Turkic languages, basically, lack declensional and conjugational categories, classes, irregular verbs, and suppletive forms’ (Johanson 1998, 36).

2  General picture The evaluative morphology of Tatar is rather sketchy, evaluative suffixes being not very productive and the distribution of them among word classes and various semantic groups being far from regular. Evaluative morphemes are of a suffixal type, like the other bound morphemes in Tatar, and cluster mostly around diminutive and caressive functions. The application of evaluative morphology across all word categories is restricted. Evaluative morphemes of diminution, affection, caressing, intensification, deintensification and other types are considered to be derivational, and they belong to the inventory of derivational suffixes since they always precede the inflectional ones. Augmentatives are nearly absent in Turkic languages, Tatar included. Excessiveness in a quantitative or qualitative feature of singular or plural form of an object, person or action is mainly expressed by repetition of the qualifying or quantifying modifiers preceding them. The duration, frequency, intensity or manner of an action expressed by a verb form may be augmented by repeated forms of adverbs. As for adjectives and quantifiers, intensification by reduplicating the first syllable of the base form and changing the last

Tatar 311 consonant of the reduplicated syllable is well attested in Tatar and other Turkic languages. Deintensive or attenuative forms of adjectives, on the other hand, particularly those relating to colour, are formed morphologically by attaching specific deintensifying suffixes, such as -­su, -­KIlt.1 In Tatar, there are no specific suffixes for emotional remarks expressing disregard, disdain or neglect, and no morphemes to augment, enlarge, increase or magnify a certain property of an entity. One way to convey a sense of disdain is to reduplicate and juxtapose two elements, the second of which is the mere sound reflection of the first one: malay-­şalay ‘boy and so on’. No suffixal forms of contempt, irony, sarcasm, belittlement, insult or a scornful attitude exist in Tatar, and there are no pejoratives in suffixal form.

3 Nouns There are some evaluative markers in Tatar, applying exclusively to nouns, which may not only indicate the smallness of an item in size and/or age, but also express positive emotions of the speaker towards a person, animal or thing conveyed by the base form. Among them, -­kAy and -­kAş carry out diminutive, endearing and caressive functions (Ganiyev 1998, 279): (1) bala-­kay child-­ dim ‘dear little child’ -­kAy and -­kAş are reported to be attached only to animate nouns and some names of body parts2 (Valiullina 1997, 57): (2) a. sandugaç-­kay nightingale-­ dim ‘dear little nightingale’ b. baş-­kay head-­ dim ‘dear head’ Terminating vowels are dropped before the suffix: (3) eŋ-­key < eni-­key mother-­ dim ‘dear mother’ The last consonant of the suffix may be truncated, when attached to the most frequently used kinship terms: (4) et-­ke father-­ dim ‘dear father’ The suffixes -­kAy and -­kAş, in some word forms -­kAç, attach mostly to kinship terms, and are employed when addressing someone in an affectionate way:

Fatma Şahan Güney

312 (5) a. baba-­kay grandfather-­ dim ‘dear grandfather’ b. tugan-­kay brother/sister-­ dim ‘dear brother or sister’

Caressive forms and endearments, which are used when addressing someone in an affectionate way, are frequently formed by attaching -­kAy to words conveying precious or valued things: (6) a. altın-­kay-­ım dim-­1poss gold-­ ‘my precious’ b. göl-­key-­ĭm rose-­ dim-­1poss ‘my dear rose’ As is evident from the above examples, the first person possessive pronoun -­(I)m is frequently attached to the end of the word form in order to strengthen the sense of affection. When attached to generic names, such as kız ‘girl’, göl ‘flower’ and yoldız ‘star’, the suffix -­kAy may indicate only diminution and/or endearment if there is no appellative context: (7)

göl-­key flower-­dim ‘floret’

One another suffix used as a caressive and appellative is -­(I)y. Like -­kAy, it is attached to kinship terms, generally to the ones expressing closer relationships, in order to convey feelings of endearment and affection. Such forms are often used when addressing someone who is younger or older than the speaker, as in aga-­y ‘dear brother, uncle’, a caressive form of aga ‘brother, uncle’. It may be used as an informal title for older men and as a polite form when addressing them directly. Similarly baba-­y ‘old man, elderly gentleman’ is the way to address informally an elderly man, especially a friend of the family. The form apa-­y, on the other hand, is used when addressing women older than the speaker. Some other forms of informal address carrying the suffix -­(I)y are as follows: (8) a. göl-­iy-­ĭm rose-­ dim-­1poss ‘my dear rose’ b. cingi < cing(e)-­y wife of the elder brother or uncle-­dim ‘dear sister-­in-­law’ It needs to be noted that the suffix -­(I)y is not detachable from the most frequently used forms of kinship terms, such as eniy ‘mother’, etiy ‘father’, abıy ‘elder brother’, babay ‘grandfather’ and ebiy ‘grandmother’, which are the standard forms for the concepts they convey, and no sense of evaluation is evident in their meaning.

Tatar 313 A widespread diminutive suffix across all Turkic languages, as well as Tatar, is -­çIk, which may also carry a caressive function: (9) Tatar: yoldız-­çık star-­ dim ‘little star’ The suffix -­çIk may also indicate that the item diminished in size or age is pitied by the speaker, even with no sense of diminution: (10) a. koş-­çık bird-­ dim ‘poor little bird’ b. hatın-­çık woman-­ dim ‘poor woman’ The more archaic and less frequent suffix in Tatar ‘which gives the noun diminutive meaning’ (Ganiyev 1998, 265) is -­tIk: (11) bülĭm-­tĭk division-­ dim ‘small section’ Some word forms suffixed with -­çIk or -­tIk in Tatar have already become lexicalised, and some of them do not reveal any diminutive meaning. Even though such words convey root-­ related meanings, some of them act as independent conceptual units: (12) a. kap-­çık sack-­ dim ‘small bag’ b. buvın-­tık joint-­ dim ‘joint, limb of an insect’ Diminutive -­ çIk is widely used among Turkic languages, though not evenly as productive: (13) Kazakh (Öner 1998, 48): tüyir-­şik sphere-­ dim ‘small sphere’ (14) Kirghiz (Öner 1998, 48): köl-­çük lake-­ dim ‘small lake’

Fatma Şahan Güney

314 (15) Karachay-­Balkar (Berta 1998, 305): narat-­çık pine-­ dim ‘small pine’ (16) Azerbaijani (Schönig 1998, 251): uşaġ-­çıġ child-­ dim ‘little child’

In Turkish, -­çIk is a productive suffix which may attach to nouns (17a), adjectives (17b) and adverbs (17c), carrying out different semantic functions: (17) a. kulübe-­cik dim hut-­ ‘little hut’ b. ince-­cik thin-­ dim ‘very thin’ c. yavaşça-­cık slowly-­ dim ‘very slowly’ Newly formed scientific terms with a diminutive sense often carry -­çIk in Turkish: (18) karın-­cık abdomen-­ dim ‘tummy; ventricule’ Caressive forms and endearments may also be constructed by attaching -­çIk and first person possessive suffix -­(I)m mostly to kinship terms: (19) anne-­ciğ-­im mother-­ dim-­1poss ‘my dear mother’ Repeated instances of diminutive forms may also occur in Turkish: (20) küçü(k)-­cü(k)-­cük small-­ dim-­dim ‘very very small’ Among the non-­productive evaluative suffixes of diminutive and caressive function are -­çAk, and -­çA, both occurring with only a limited number of words in Tatar. -­çAk and -­çA both have diminutive and caressive functions coupled with the sense of pity:



(21)

kilĭn-­çek daughter-­in-­law-­dim ‘dear poor daughter-­in-­law’

Tatar 315 -­çAk has had a diminutive function in both historical and modern times in Turkic with a very limited usage: (22) Kazakh: ini-­şek brother-­ dim ‘dear little brother’ (23) Kirghiz: ini-­çek ‘dear little brother’ (24)

Trakai Karaim (Berta 1998, 305): til-­çek dim tongue-­ ‘little tongue’

(25) Chagatai: çukur-­çak cavity-­ dim ‘small cavity’ Diminutive -­çAk, which may also express pity, mostly attaches to animate names in Turkish and occurs in a few words: (26) yavru-­cak child-­ dim ‘poor child’ In such forms, the sense of diminution, endearment and pity may be conveyed all together. The diminutive suffix -­çA is also used in many other Turkic languages besides Tatar: (27) Tatar: yözak-­ça lock-­ dim ‘small lock’ The suffix -­çA is reported to be the most common diminutive suffix in Kazakh (Kirchner 1998, 322), though languages of the Kipchak and other groups have used the form with a limited number of words: (28) Kazakh: balık-­şa fish-­dim ‘small fish’ (29) Azerbaijani: mėydan-­ça

316

Fatma Şahan Güney

square-­ dim ‘little square’ The diminutive and the caressive suffix -­çAğIz, an extended form of -­çAk, is more productive in Turkish: (30) çocuk-­cağız child-­ dim ‘poor child’ Even though there are no augmentatives in Tatar, a few morphemes like -­bay, -­lAç and -­tey seem to be used with a very limited number of words: (31) a. sakal-­bay aug beard-­ ‘beardie’ b. mıyık-­laç moustache-­ aug ‘beardie’ c. sakal-­tey beard-­ aug ‘beardie’

4  Adjectives and adverbs Intensification of adjectives, specifically those relating to colour, and adverbs in Tatar is carried out either by using intensifying particles such as döm and ör, as in ör yaŋa ‘brand new’, döm kara ‘completely black’, or by reduplicating the first syllable of an adjective: (32) a. kap~kara int~black ‘pitch black’ b. ap~arık int~skinny ‘very skinny’ Both intensifying particles and reduplicated parts are placed just before the adjective. Deintensification in the same category, mainly with regard to colour adjectives, on the other hand, is carried out by attaching some specific deintensifying suffixes to the word base. The most frequently used suffixes of deintensification are -­su, -­KIlt, -­şIl and -­sIl: (33) a. zeŋger-­su purple-­ att ‘somewhat purple’ b. yeş-­kĭlt green-­ att ‘greenish’

Tatar 317 c. kük-­sĭl blue-­ att ‘bluish’ A few formatives such as -­sIman and -­DAy attach to nouns and express deintensification, reduction and approximation in quality or quantity, with more of a sense of similarity: (34) a. bala-­sıman child-­ att ‘like a child; childish’ b. yĭfek-­tey silk-­ att ‘silky’ Among its other semantic functions, GInA3 in Tatar realises the attenuation and diminution of adjectives and adverbs, as in yomşák4 kına ‘quite soft(ly)’, tiz gĭne ‘at once’. There are a few words, however, where word forms which have the suffixal form of GInA are lexicalised and have their own independent meaning:5 (35) kĭç(ĭ)-­kĭné small-­ dim ‘small’ GInA also functions as a diminutive and caressive suffix, which attaches to a limited number of nouns, as in: (36) bala-­gına-­m child-­ dim-­1poss ‘my little child’ The suffixal variant of GInA also has the form -­gınay/gĭney, a complex form consisting of the old diminutive suffix -­gına/gĭne plus -­(I)y, which is another diminutive and caressive suffix: (37) bil-­gĭney waist-­ dim ‘my dear waist’ Attenuation of adjectives is carried out in Tatar by the suffix -­rAk, originally a comparative suffix where two items are compared in terms of their quality or quantity. However, ‘when there are no items to compare with, then the suffix -­rAk gives the adjective a meaning that the quality the adjective expresses is not more but less than the standard’ (Hisamova 2006, 120): (38) kıska-­rak short-­ att ‘rather short’

318

Fatma Şahan Güney

It may even attach to adverbs in order to express the meaning ‘a little bit’: (39) tĭgĭley-­rek there-­ att ‘a little bit there’ Similarly, the attenuative function is fulfilled by the suffixes -­çAk and -­çA in Turkish, as in:6 (40) a. büyü(k)-­cek att big-­ ‘somewhat big’ b. büyük-­çe big-­ att ‘somewhat big’ On the other hand, -­çA intensifies or augments the meaning of adverbs:7 (41) hızlı-­ca rapid/rapidly-­ int ‘very rapidly’ The propensity or excessiveness of some quality within an object or pertaining to an animate being are expressed by the denominal adjectival suffixes -­çIl, -­çAn, and the deverbal adjectival suffix -­KIr. -­çAn is quite a productive suffix which expresses propensity: (42) a. yerdem-­çĭl help-­ dna(aug) ‘responsive’ b. süz-­çen word-­ dna(aug) ‘talkative’ c. siz-­gĭr sense-­ dva(aug) ‘sensible, perceptive’ In Kirghiz, ‘a frequent suffix in neologisms is -­çIl, denoting propensity to something’ (Kirchner 1998, 347): (43) ulut-­çul nation-­ dna(aug) ‘nationalist’ The suffix -­çIl has similar functions in other Turkic languages: (44) Kazakh: şay-­şıl tea-­ dna(aug) ‘fond of tea’

Tatar 319 (45) Turkish: ben-­cil me-­ dna(aug) ‘selfish’

5 Verbs If the fulfilment of an action is less than the standard, or it does not meet the expectation of the speaker, some specific suffixes are added to the verb base. These suffixes are attached to the base form preceding all inflectional suffixes. The lesser degree of the quality of an action is expressed by the denominal verb formative -­lAn attached to words of a nominal or adjectival type: (46) divana-­lan-­u dav(att) mad-­ ‘to become a little wild’ There are a number of deverbal verbal suffixes in Tatar used in expressing the lesser degree or partial fulfilment of an action. Such deverbal verb formatives, such as -­(I)ştIr, -­KAlA8 and -­(I)nkIrA, can be considered as evaluative since, besides any other nuances, they all indicate that the action expressed by the verb is carried out not to its full extent but to a lesser degree: (47) a. ukı-­ştır-­u read-­ att-­vn ‘to read randomly or from time to time’ b. suk-­kala-­w plough-­ att-­vn ‘to plough for a while or from time to time’ c. üs-­ĭnkĭre-­w grow-­ att-­vn ‘to grow a little’ Today, the function of such deverbal suffixes as -­(I)ştIr, -­KAlA and -­(I)nkIrA is fulfilled mainly by the compound verb forms constructed with the postverbs al-­ ‘to take’ and töş-­ ‘to descend’, which are both productive forms in expressing a lesser degree of an action: (48) a. kızdır-­ıp al-­u fry-­ cvb pstv-­vn ‘to fry a little’ b. tuŋ-­a töş-­ü freeze-­ cvb pstv-­vn ‘to freeze slightly’

6 Conclusion The evaluative morphology of Tatar is related mostly to nouns and adjectives, though a limited number of suffixes with evaluative functions do occur with verbs, and a few of

320

Fatma Şahan Güney

them with adverbs as well. The semantic categories of evaluative morphology in Tatar, and in fact in many Turkic languages, are concentrated mainly around diminutive and attenuative functions. Among the evaluative suffixes of the nominal category, -­çIk and -­kAy are the most productive in Tatar, having both diminutive and caressive functions. -­kAy attaches only to the names of animate beings, while -­çIk may occur with both animate and inanimate nouns. The evaluative morphology pertaining to the category of adjectives is mainly related to diminutive and attenuative functions. The intensification of adjectives is carried out mainly by reduplication, while deintensification in the same category is fulfilled by some specialised suffixes used mainly with colour adjectives, and the productive adjectival suffix -­rAk, which also has a comparable function in Tatar. GInA, originally a clitic in Tatar, is also used as an evaluative suffix, which may occur with nouns, adjectives and adverbs. Verbs have fewer evaluative morphemes, namely -­(I)ştIr, -­KAlA and -­(I)nkIrA, the first two being quite productive. In our opinion, they all have evaluative value, particularly an attenuative one, since the verb form affixed with any of them means ‘to do something a little bit, for a while or randomly’. In terms of the formal properties of the evaluative morphemes in Tatar, they are mainly of a suffixal type and they always precede the inflectional morphology.

Notes 1. The dotless and capitalised ‘I’ used in the representation of individual suffixes signals both front and back variants of middle unrounded reduced vowels, represented respectvively as ĭ and ı. Other capitalised letters also represent both front and back allaphones of the suffix vowel or consonant. 2. Valiullina states that there are some registers belonging the central dialect where -­kAy may also attach to object names: utın-­kay, balta-­kay (Valiullina 1997, 57). 3. This particle has four allamorphic variations, namely gına, gĭne, kına and kĭne, and may attach to various syntactic phrases and even sentences in order to fulfil different semantic functions. It is not a stress-­bearing unit, so it is written separately in Tatar orthography. 4. Accented vowels show the position of pitch-­accent. 5. In such forms, pitch-­accent is placed on the last syllable of the word form, as would be expected in the regular process of word formation. 6. As (40) shows, the word-­final -­k drops before -­çAk and -­çIk in Turkish: yumuşa(k)-­cık ‘very soft’, çabu(k)-­cak ‘very speedily’. 7. ‘Accent has a distinctive function in cases such as güzel-­cé “rather nice”, gü’zélce “nicely”’ (Csató and Johanson 1998, 207). 8. The suffix -­KAlA has the allamorphs -­kala, -­kele, -­gala and -­gele.

15.15 Telugu

Pingali Sailaja

1 Introduction Telugu is one of the four major Dravidian languages spoken primarily in the state of Andhra Pradesh in the south of India. There are around twenty-­six Dravidian languages, classified as Southern, South-­Central, Central and Northern groups. Telugu belongs to the South-­Central group of languages and is the most prominent one in this group. The other three major literary languages of Dravidian are Tamil, Malayalam and Kannada, all belonging to the Southern group (Krishnamurti 2003, 19). According to the 2001 Census of India, there were then 73,002,856 speakers of Telugu in India. And there are 74,049,000 speakers of Telugu across the world (Lewis, Simons and Fennig, 2014). There are four regional dialects of Telugu – northern, comprising the Telangana region; southern, comprising Rayalaseema and the coastal districts of Nellore and Prakasam; eastern, comprising Visakhapatnam, Vijayanagaram, Srikakulam; and central coastal districts, comprising Guntur, Krishna, and East and West Godavari. Standard Telugu as known today comes from the elite groups of the central coastal regions. Social dialects emerge on the basis of the education levels of the users of the language (Krishnamurti 2003, 23). In linguistic aspects, Telugu is different from the more prevalent Indo-­ European languages of the Indian subcontinent. Many aspects of the morphology of Telugu are common to all Dravidian languages. Like all other Dravidian languages, Telugu is an SVO language and in morphological type is an agglutinative language. There is only suffixation and no prefixation or infixation. Any prefixes that might be apparent are historically borrowed from Sanskrit. Compounding is very productive and so are other morphological ­constructions such as reduplication and echo formation.

2  Evaluative morphology Telugu is most unlike the neighbouring Indo-­European language family in which evaluative morphology (EM) is considerable, particularly in suffixation. Telugu matches the other Dravidian languages in possessing EM to a moderate degree. Evaluative formation is realised mostly in morphological processes such as compounding, affixation, reduplication and echo formation. It is also exemplified as a phonological process. These are illustrated here. Apart from some instances of diminution, several other semantic aspects are realised in Telugu evaluative constructions. The details of EM are presented below according to morphological type of construction, followed by the phonological exemplification.

322

Pingali Sailaja

2.1 Compounding Compounding is a very productive process in Telugu. Both sub-­compounds and co-­ compounds exist. Sub-­compounds are right headed. EM is seen in constructions in which some specified words are compounded with other words to give an evaluative construction. Diminution is realised with the word pilla ‘girl, young one’:1 (1) pilla ‘girl’, ‘young one’: a. kukka-­pilla dog-­young one.dim ‘puppy’ b. pilli-­pilla cat-­young one.dim ‘kitten’ c. kooti-­pilla monkey-­young one.dim ‘young of a monkey’ d. pičika-­pilla sparrow-­young one.dim ‘young of a sparrow’ e. kaaki-­pilla crow-­young one.dim ‘young of a crow’ This process is quite productive. However, if there is a lexical item that already has the sense of ‘a young one’, such as duuɖa ‘calf’, pilla is not used. So, for a calf, one is unlikely to hear aaʋu pilla ‘young of a cow’, where pilla is compounded with aaʋu ‘cow’, although it would not be considered to be incorrect. When referring to human beings, evaluative meaning is expressed through compounding, once again with words that are fixed. The words used are amma ‘mother’ and ayya ‘father’. Although the word amma literally means ‘mother’, it is also used as a respectful address form for a stranger or a person one is not familiar with or for a person of a higher hierarchical status. Telugu has a lexical and post-­lexical rule of deleting a word-­final vowel if the following word begins with a vowel. This phonological change is exemplified in the right-­hand column in (2). (2) a. siita-­amma siitamma Sita-­mother.end.,resp ‘dear Sita, respected Sita’ b. ǰyooti-­amma ǰyootamma Jyoti-­mother.end.,resp ‘dear Jyoti, respected Jyoti’ c. durga-­amma durgamma Durga-­mother.end.,resp ‘dear Durga, respected Durga’ d. ǰaabili-­amma ǰaabilamma moon-­mother.end.,resp ‘dear moon, respected moon’

Telugu 323 The words in (2), except for (2d), are used as endearments with children or loved ones. In parallel, these compounds are also respectful referents for women who are of a higher status. Diminution is expressed in a slightly different manner in these cases. The word amma does not carry the sense of diminution inherently. Since these words are used for children, the sense of diminution is conveyed by extension from the endearment. Only pragmatic factors make the sense of diminution available. In (2d), the moon is personified as a woman, and it can be compounded with amma and used more as a respectful reference. In fact, a further nuance of meaning is available here. More than respect, the sense conveyed is one of closeness or proximity. Phonologically, when the words are used to call or address a person, the final vowel is lengthened. Thus, the address forms would be: (3) a. siit-­amma-­a Sita-­mother.end.,resp-­voc ‘O Sita’ b. durg-­amma-­a Durga-­mother.end.,resp-­voc ‘O Durga’ Other than names, the same word, amma, can be compounded with adjectives that already express the quality of diminution. Phonologically, when the first vowel is high and front /i/ and the second vowel is back and low /a/, assimilation takes place and a mid vowel /e/ is realised. (4) a. čiʈʈi-­amma čiʈʈemma little-­mother.end ‘dear little girl’ b. bulli-­amma bullemma tiny-­mother.end ‘dear little girl’ c. buǰǰi-­amma buǰǰemma small-­mother.end ‘dear small girl’ The resultant words are nouns that are used to refer to children and are endearments. The sense of diminution is carried by the adjective with which amma is compounded. In these cases, the respectful form does not arise. Often, these words become the pet names of individuals. Once again, if a child is being called, then the final vowel is lengthened. Just as amma ‘mother’ is used to refer to women, ayya ‘father’ is used for men.2 However, ayya is ‘father’ only in some dialects. In many dialects the common word for father is naanna. Other words for father are abba and appa. Nevertheless, ayya is a respectful form of address or reference in many dialects. It behaves just like the word amma to create new compounds in the masculine gender. In (5) it is compounded with names and in (6) it is compounded with adjectives that already indicate diminution. (5) ayya: a. šiʋa-­ayya šiʋayya Siva-­father.end.,resp ‘dear Siva, respected Siva’

Pingali Sailaja

324 b. raama-­ayya raamayya Rama-­father.end.,resp ‘dear Rama, respected Rama’ (6) a. činna-­ayya small-­father.end ‘dear small boy’ b. bulli-­ayya tiny-­father.end ‘dear little boy’

činnayya bullayya

The examples in (5) are endearments and carry a sense of exaltation or respect. Just like words with amma, the words in (5) are used for children, and the sense of diminution is carried by them as a consequence of such usage. The word ayya does not inherently carry a sense of diminution. In (6) the words carry the sense of diminution and endearment. The masculine forms are fewer in number than the feminine forms. When used to call someone, the final vowel is lengthened: šiʋayyaa, raamayyaa, and so on (Krishnamurti and Gwynn 1985, 305). Other words that can be similarly used in compounds are talli ‘mother’ and tanɖri ‘father’. These words are of a higher style than amma, naanna and ayya. The meaning derived is that of endearment. As a consequence of their being compounded with an adjective indicating the quality of diminution, the sense of diminution is also available. (7) talli ‘mother’: a. buǰǰi-­talli small-­mother.end ‘dear little girl’ b. činna-­talli small-­mother.end ‘dear little girl’ (8) tanɖri ‘father’: a. buǰǰi-­tanɖri small-­father.end ‘dear little boy’ b. činna-­tanɖri small-­father.end ‘dear little boy’ c. čiʈʈi-­tanɖri tiny-­father.end ‘dear little boy’ The words talli and tanɖri cannot be compounded with names, so the following are unacceptable: (9) a. *siita-­talli Sita-­mother.end ‘dear little Sita’

Telugu 325 b. *durga-­talli Durga-­mother.end ‘dear little Durga’ c. *raama-­tanɖri Rama-­father.end ‘dear little Rama’ However, compounds containing names of people, and amma or ayya as illustrated in (2) and (5), can further take talli and tanɖri: (10) a. siita-­amma-­talli siitammatalli Sita-­mother.end.,resp-­mother.end.,resp ‘dear Sita, respected Sita’ b. durga-­amma-­talli durgammatalli Durga-­mother.end.,resp-­mother.end.,resp ‘dear Durga, respected Durga’ c. raama-­ayya-­tanɖri rammayyatanɖri Rama-­father.end.,resp-­father.end.,resp ‘dear Rama, respected Rama’ d. šiʋ-­ayya-­tanɖri Siva-­father.end.,resp-­father.end.,resp ‘dear Siva, respected Siva’ These forms carry an extra sense of affection or respect. As before, these are used mostly with children. All names that are compounded with amma, ayya, talli and tanɖri are names of different deities, so the compounded words can also be the names of gods and goddesses, and it is acceptable to address and refer to the gods using such expressions. As in all previous cases, the words have final vowel lengthening when used to call someone. Intensification is also possible through compounding by using the adjective pačči ‘raw’. The effect of such compounding is to create a word that intensifies the meaning of the word with which it is compounded. (11) a. pačči-­abaddham raw-­lie b. pačči-­dongatanam raw-­robbery

paččabaddham

‘gross lie’

pačči-­dongatanam

‘gross robbery’

2.2 Derivation There are relatively few productive derivational affixes in Telugu that carry an evaluative meaning. There is a gender distinction in those that exist. The masculine affix is -­gaaɖu and the feminine is -­di. Both carry a pejorative sense. There are some differences in the manner in which they function. The masculine suffix -­gaaɖu attaches to masculine names of people and to occupations of individuals. It is more productive than the feminine form.3

326

Pingali Sailaja

(12) -­gaaɖu: a. raaǰu-­gaaɖu Raju-­ pej.3m ‘Raju fellow’ b. raʋi-­gaaɖu Ravi-­ pej.3m ‘Ravi fellow’ c. pliiɖar-­gaaɖu pleader-­ pej.3m ‘pleader fellow’ d. laayar-­gaaɖu lawyer-­ pej.3m ‘lawyer fellow’ e. daakʈar-­gaaɖu doctor-­ pej.3m ‘doctor fellow’ The exception to the general derogatory use of this suffix is the word andagaaɖu ‘handsome man’.4 This suffix can also attach to some adjectives with a pejorative sense. (13) a. rauɖi-­ gaaɖu rowdy-­ pej.,end.3m ‘rowdy fellow’ b. moddu-­gaaɖu dull-­ pej.,end.3m ‘dull fellow; dullard’ While the default semantics of words derived through -­gaaɖu is that of a pejorative nature, it is possible to use the words with adjectives affectionately – rauɖigaaɖu, moddugaaɖu – referring to someone who is very close and generally a friend or a child. The feminine equivalent of this suffix is -­di.5 It cannot attach to names. So, expressions such as *giita-­di are absent, where Gita is a name. -­di attaches sparingly to a few words, usually occupations, such as the following. (14) a. kalekʈar-­di collector-­ pej.3f ‘(that) collector woman (creature)’ b. daakʈar-­di doctor-­ pej.3f ‘(that) doctor woman (creature)’ In these cases too, the sense is pejorative. So a sentence such as the following refers to the doctor derogatorily. (15) daakʈar-­di ʋačč-­indi doctor-­ pej.3f come-­ pst.sg.3f ‘That doctor woman has come.’

Telugu 327 There is no evidence of any kind to claim that these suffixes are inflections. The change in meaning that is brought about suggests that these must be treated as derivational affixes. 2.3 Reduplication There are several forms of reduplication in Telugu and in Dravidian. In terms of morphological constructions, reduplication may also be characterised as a type of compounding. In general, reduplicated forms convey the distributive or intensive sense. Reduplicative adjectives conveying a distributive or intensive sense to the nouns they modify are i­llustrated in (16). (16) a. činna~činna small~rdp.distr.int ‘little-­little, small-­small’ b. pedda~pedda big~rdp.distr.int ‘big-­big’ c. manči~manči good~rdp.distr.int ‘good-­good’ d. ʋeeɖi~ʋeeɖi hot~rdp.distr.int ‘hot-­hot’ e. duura duura far~rdp.distr.int ‘distant-­distant’ The quality expressed by the adjective is meant to apply distributively to all the items indicated by the noun. At the same time, intensified meaning is also possible. (17) a. činna~činna ʋastuʋulu small~rdp.distr.int objects ‘many small objects, tiny objects’ b. ʋeeɖi~ʋeeɖi annam hot~rdp.distr.int rice ‘very hot rice’ Since rice is non-­count, a distributive sense is not possible, and only intensification is possible. When these reduplicated forms are used with an adverbial suffix, -­gaa, the sense is that of intensification. (18) a. ʋaɖa-­lu ʋeeɖi~ʋeeɖi-­gaa unn-­aayi snack-­ pl hot~rdp.distr.int-­adv be-­ prs.pl.3ntr ‘The vadas (snacks) are very hot.’ b. paalu tiyya~tiyya-­gaa unn-­aayi milk sweet~rdp.distr.int-­adv be-­ prs.pl.3ntr ‘The milk is very sweet.’

328

Pingali Sailaja

A number of adverbs derived by the addition of -­gaa to reduplicated adjectives exemplify intensification. (19) a. tʋara~tʋara-­gaa quick~rdp.distr.int-­adv ‘very quickly’ b. mella~mella-­gaa gentle~rdp.distr.int-­adv ‘very gently’ c. metta~metta-­gaa soft~rdp.distr.int-­adv ‘very softly’ d. pedda~pedda-­gaa big~rdp.distr.int-­adv ‘very hugely/loudly’ e. masaka~masaka-­gaa blurred~rdp.distr.int-­adv ‘very blurred’ Interestingly, the first part can be added iteratively, so the following are possible: (20) a. tʋara~tʋara~tʋara-­gaa quick~rdp.distr.int~rdp.distr.int-­adv ‘very, very quickly’ b. mella~mella~mella-­gaa gentle~rdp.distr.int~rdp.distr.int-­adv ‘very, very gently’ c. metta~metta~metta-­gaa soft~rdp.distr.int~rdp.distr.int-­adv ‘very, very softly’ Although these structures are possible, they are rare and are more literary in nature. The iterativity indicates added dimensions of the meaning of the original word. Some non-­finite participles, affirmative and negative, are identified by Krishnamurti and Gwynn (1985, 186) for Telugu. Of these, the past participle, the durative and the negative durative can be reduplicated to indicate continuous action, thereby signalling intensification. The past participle is reduplicated, giving the sense of intensification. The form remains a participle even after reduplication. (21) a. čees-­i~čees-­i do-­ pst.ptcp~do-­pst.ptcp ‘having worked a lot’ b. čepp-­i~čeppi say/tell-­ pst.ptcp~say/tell-­pst.ptcp ‘having said a lot’ c. čuus-­i~čuus-­i see/wait-­pst.ptcp~see/wait-­pst.ptcp ‘having seen/waited a lot’

Telugu 329 d. tin-­i~tin-­i eat-­ pst.ptcp~eat-­pst.ptcp ‘having eaten a lot’ These words are used in sentences as follows: (22) a. čuus-­i~čuus-­i aanandinč-­aam see-­ pst.ptcp~see-­pst.ptcp enjoy-­ pst.1pl.incl ‘We saw (it) a lot and enjoyed it. b. ʋin-­i~ʋin-­i ʋisuku ʋačči-­ndi exasperation come-­ pst.sg.3f hear-­pst.ptcp~ hear-­pst.ptcp ‘We got exasperated listening (to it) too much.’ The durative indicates continuous action. (23) a. čuus-­tuu~čuus-­tuu paɖa-­taaɖu see-­ dur~see-­dur fall-­ fut.sg.3m ‘While watching he will fall.’ b. ʋin-­ʈuu~ʋin-­ʈuu raas-­tunaaɖu hear-­ dur~hear-­dur write-­prs.prog.sg.3m ‘He is writing while listening (to something).’ Similarly, the negative durative -­aka forms also can be reduplicated. (24) a. raay-­aka~raay-­aka write-­ neg.dur~write-­neg.dur ‘not having written for a long time’ b. raa-­ka~raa-­ka come-­neg.dur~come-­neg.dur ‘not having come for a long time’ All of these reduplicated verb forms exemplify continuous action and thereby intensification. 2.4  Echo formation Echo formation in Dravidian appends a nonsense syllable gi-­. The vowel length in this ­syllable is determined by the vowel length of the word to which it attaches. These words add the meaning of ‘and such things’ to the original meaning. Morphologically, echo formation may be considered to be a type of compounding. Echo formation is a highly productive process in Telugu and other Dravidian languages. These words carry a wide range of meanings. They can be derogatory, or simply dismissive. They usually do not carry just the literal sense of ‘other such things’. Once again, the pragmatics determines the exact nuance of meaning. (25) a. puli~gili tiger~ech ‘tiger and such things (pejorative/dismissive)’

330

Pingali Sailaja

b. paalu~giilu milk~ech ‘milk and such things (pejorative/dismissive)’ c. čaduʋu~giduʋu education~ech ‘education and such things’ abbaayi-­l-­aki čaduʋu~giduʋu anaʋasaram boy-­ pl-­dat education~ech unnecessary ‘Education and such (useless) things are unnecessary for boys.’ In (25c), the echo form is used dismissively. Generally, nouns participate in echo formation. People can be referred to derogatorily using echo formation, where the name is repeated with the nonsense syllable. (26) a. goopi~giipi Gopi~ech ‘Gopi (pejorative)’ b. meeri~giiri Mary~ech ‘Mary (pejorative)’ In these words, the individual is referred to pejoratively. Depending on the context, these constructions can also mean ‘Gopi and other boys’, or ‘Gopi and his group’, all in a dismissive sense. 2.5  Phonological features There are further phonological effects that express intensification. When adjectives are used in an exaggerated form, the original word can take a different phonological form. In (27), the first column illustrates the original adjective, and the second is a form that expresses an intensified sense of it. Depending on the context, these words may express sarcasm.6 (27) a. goppa goppha ‘great’ ‘very great’ b. pedda peddha ‘big’ ‘very big’ c. tiyya tiyyha ‘sweet’ ‘very sweet’ d. pulla pullha ‘sour’ ‘very sour’ e. čeedu č heedu ‘bitter’ ‘very bitter’ f. kaaram khaaram ‘hot’ ‘very hot’ Aspiration is added at some point in the word. Generally, when a word has a geminate consonant, aspiration is attached to that sound. In fact, the geminate consonant also tends

Telugu 331 to be further lengthened. When there is no gemination in a word, aspiration tends to be on the first consonant in the word. Interestingly, aspiration also appears with sounds that do not take aspiration phonologically (as with y and l) in these constructions. Words that carry a sense of smallness can also have further diminution added to them. In such cases, the geminate consonant is lengthened: (28) a. činna ‘small’ b. buǰǰi ‘tiny’

činnna ‘tiny’ buǰǰǰi ‘very tiny’

Over and above the senses illustrated so far with adjectives, some words when exaggerated exhibit vowel length. (29) a. čiikaʈ i čiiikaʈi ‘dark/darkness’ ‘very dark’ b. loopala looopala ‘inside’ ‘deep inside’ In these cases, the extra vowel length is phonetic in nature and is not phonological; yet additional meaning is conveyed.

3 Conclusion In conclusion, we may say that the senses that that are exemplified in EM in Telugu are those of diminution, endearment, respect, intensification and pejorative meaning. The same structure may exemplify different meanings. Also, pragmatic considerations make several nuances of meaning available. A semantic grouping emerges in this context. Diminution, endearment and respect go together, whereas pejorative and dismissive meanings form one semantic set, with the same structure exemplifying one or more of these meanings.

Acknowledgements I am grateful to Professor V. Prakasam for a discussion on the nuances of the data.

Notes 1. The compound kooti-­pilla in (1c) can also function as a sub-­compound with the meaning ‘girl who looks/behaves like a monkey’. 2. Although the word amma means ‘mother’ and is feminine, it can be used as an endearment for men in some contexts. Thus, the following can also be used while talking to a man: Emi ayindi amma? What happened mother.end ‘What happened, dear?’ 3. There is a general honorific suffix in Telugu, -­gaaru, which is gender neutral and attaches to proper names and to the profession of the person who is being referred to.

332

Pingali Sailaja

4. The word anda-­gatte is the feminine equivalent of anda-­gaadu. The suffix -­gatte is a variation of -­katte, which is seen in words such as celi-­katte ‘female friend of royalty’. 5. There are two separate functions of -­di. One is illustrated in (5). -­di is also a neuter and feminine form used in predicative positions attaching to adjectives in sentences that do not have a copula. aa pilla cinna-­di. that girl small-­ sg.3f ‘That girl is small.’



The masculine equivalent of this is: aa pillaaɖu cinna-­ʋaaɖu. that boy small-­he.sg.3m ‘That boy is small.’ 6. The word kharma ‘bad destiny’ in Telugu is probably derived historically through this process of adding aspiration to words. It was probably added to karma ‘destiny’, which does not have any negative connotations. Many speakers of Telugu are unaware of the neutral word altogether.

15.16 Udihe

Maria Tolskaya

1 Introduction Udihe and other Southern Tungusic languages have a limited number of evaluatives and use them very sparingly, compared to their Northern relatives (e.g. Evenki). This discrepancy can be attributed to the fact that in Northern Tungusic languages, attaching evaluative suffixes brings about unification of the diverse nominal bases and may facilitate language acquisition, while in Udihe, a language with simpler morphology, there is no comparable diminutive advantage. 1.1  Genetic affiliation and sociolinguistic status Udihe (Udeghe) is a Manchu-­Tungusic language spoken by fewer than fifty people in the Russian Far East. The Manchu-­Tungusic language family includes about twelve languages in Central and Eastern Siberia, the Russian Far East, Sakhalin and northern China. There is no agreement about the composition of the Manchu-­Tungusic language family or its genetic relationship with the putative Altaic macro-­family. Tsintsius (1949, 35) classifies Evenki, Even, Solon and Negidal as Northern Tungusic languages, and groups Manchu together with Southern Tungusic languages. Doerfer (1978) considers Oroch and Udihe as part of a central transitional branch, while Kormushin (1998, 11) argues for treating Oroch and Udihe as originally Northern Tungusic languages that were considerably influenced by Southern Tungusic ones. Whaley, Grenoble and Li (1999) question the applicability of the family tree model to Tungusic (particularly, Northwestern Tungusic) languages, and view them as a set of language-­dialect continua. At the present time, Udihe exists in two dialects, the Southern and the Northern. The Southern dialect was influenced by the intensive Udihe–Chinese contacts in the nineteenth century, and Janhunen (1999, 32) considers Udihe ‘the most Sinicised Amur Tungusic idiom’. Udihe has been also undergoing drastic structural changes due to contacts with Russian. Despite recent attempts of language revitalisation, the current sociolinguistic situation is deplorable: Udihe is used only by elderly people, and its transmission to children has stopped. 1.2  Linguistic profile Phonologically, Udihe is characterised by phonemic vowel length and laryngealisation, but in present-­day pronunciation, laryngealisation is almost absent. Udihe vowel harmony is

334

Maria Tolskaya

based on rounding and height distinctions, while some suffixes (including the diminutive suffix -­ziga) are disharmonic. Overall, Tungusic languages employ agglutination with some elements of fusion. In Udihe, each derivational and inflectional suffix has a fairly consistent shape and function. The major lexical classes are distinguished morphologically and syntactically, although nouns and adjectives share some derivational suffixes. There is no gender distinction. Plurality is not obligatorily expressed morphologically on nouns. Udihe is a head-­final, head-­marking language, with the basic SOV constituent order largely affected by information structure.

2  Evaluative morphology The inventory of evaluative morphemes in Udihe is impoverished and consists of two diminutive suffixes: the productive disharmonic suffix -­ziga, and the semi-­productive suffix -­sA (-­s’A) that obeys vowel harmony; both of them are compatible with nouns and adjectives. This section examines Udihe diminutive suffixes in terms of their descriptive/expressive content, and looks closely at the contexts where they occur or, on the contrary, are disfavoured. 2.1  The diminutive suffix -­ziga/-­zig’a The earlier Udihe linguistic materials (e.g. Schneider 1936) show that the diminutive suffix had laryngealisation: -­zig’a, but it is not laryngealised in the current pronunciation. This loss has certain consequences for stress assignment: the diminutive suffix -­ziga does not receive stress automatically, since in Udihe primary stress falls on the rightmost long or laryngealised vowel, and otherwise the final vowel is stressed. Moreover, the diminutive suffix -­ziga has become homophonous with the plural affix; thus moxo-­ziga ‘a small cup’/‘cups’ is ambiguous between a diminutive and plural reading. The two homophonous affixes may have a different origin. The Udihe plural affix -­ziga corresponds to the Nanai, Oroch and Ulcha collective plural affixes (Tsintsius 1949, 255), and is probably borrowed from Chinese (jĭ ge ‘a few, several’). Boldyrev (1987, 165) suggests that the diminutive -­ziga goes back to the Proto-­Tungusic -­*tkAn, but there seems to be not enough evidence for this in other Tungusic languages. 2.1.1  The diminutive suffix -­ziga/-­zig’a in nouns The suffix -­ziga is fairly productive – that is, consultants can build diminutive forms from any noun whose semantics allows for such formation. Nevertheless, it is used very rarely. Although cross-­linguistically diminutives tend to express affection, the primary descriptive content of the nominal suffix -­ziga is the referent’s small size: (1) a. ni:-­ziga ‘little man (in a folk tale); grass effigy’ (ni: ‘man’) b. ugda-­ziga ‘small canoe; toy canoe’ (ugda ‘canoe’) c. ugbe-­ziga ‘small river rapid’ (ugbe ‘river rapid’) This suffix does not evoke any phonological changes, and the base noun is usually obvious, but there are a few words of obscure origin: (2) a. kasanziga ‘puppy’ (cf. Evenki kača ‘puppy’) b. o:nzige ‘deer’ < Proto-­Tungusic *oron ‘reindeer’ (Kormushin 1998, 275) c. aziga ‘girl’ (cf. Evenki asa:tka:n ‘girl’)

Udihe 335 When these ‘diminutiva tantum’ words are pluralised, the suffix -­ziga can occur twice, e.g. a-­ziga-­ziga ‘girls’, although diminutives with -­ziga rarely attach the optional homophonous plural marker. The diminutive suffix -­ziga is not employed as a means for loanword adaptation, but is quite compatible with Russian borrowings, e.g. tractori-­zig’e ‘toy-­ tractor’ (Schneider 1936, 73). A word with the diminutive suffix -­ziga can always be paraphrased as a noun phrase with the adjective ŋic’a ‘small’, but occasionally diminutivisation gives rise to a new lexical item: (3) a. zeli-­ziga ‘a fingerling, resembling bull trout’ (zeli ‘bull trout’) b. mo:teli-­ziga ‘Eurasian pygmy owl’ (mo:teli ‘Boreal owl’) (Kormushin 1998, 249) c. gau-­ziga ‘a short push pole for stealing up to game’ (gau ‘push pole’)

The compatibility of -­ziga with different semantic nominal classes does not obey the animacy hierarchy of ‘Proper name > Human > Animal > Inanimate’; rather, it seems to reverse this order. The suffix -­ziga does not occur in addresses with kinship terms, where it is blocked by suppletive vocative forms, and its hypocoristic use is not attested either. In our corpus of data, the nouns that most frequently attach this suffix are inanimate: zugdi ‘house’, kawa ‘bark house’ and xokto ‘trail’.1 There is no predilection for affective use of diminutives in pet-­related speech, and in’ai-­ ziga usually denotes a small size dog, not a ‘cute doggie’. Diminutives with -­ziga can refer to young animals, but here they compete with numerous age-­specific terms, such as bakana ‘one-­year-­old bear cub’ and sü: ‘one-­year old boar’. An alternative way to form young animal designations is the possessive noun phrase with the head site ‘kid, child’: (4) a. zugu site-­ni b. kuti site-­ni otter kid-­3sg tiger kid-­3sg ‘young otter’ ‘tiger cub’ Affective diminutivisation of uncountable mass nouns that is typical of motherese Russian (such as moločko ) is not attested in Udihe. With mass nouns, the only possible readings of the diminutive suffix -­ziga are singularity and individuation (5a, b), or small quantity, as in (5c). (5) a. ñukte-­ziga b. samikta-­ziga c. mo:-­ziga

‘a hair’ (ñukte ‘hair’) ‘an eyelash’ (samikta ‘eyelashes’) ‘small tree; small stick; piece of wood’ (mo: ‘tree; wood’)

Rare instances of expressive usage of -­ziga with the evocation of endearment rely on context, intonation or the adjective ŋic’a ‘small’ in the ‘emphatic’ form: ŋic’a-­i ‘little, tiny’ (Nikolaeva and Tolskaya 2001, 176–7). Examples with some additional notes are provided in (6): (6) a. ŋic’a-­i kawa-­ziga b. ŋic’a-­i ba:ta-­ziga small-­ emph bark.house-­dim little-­emph boy-­dim ‘a little bark house’ ‘a little boy’

Maria Tolskaya

336

c. Ta: kä-­la ña ic’e-­i olondo-­ziga bie. there near-­ loc again small-­ emph ginseng-­ dim be.prs.hab ‘There is another little ginseng root nearby.’ (Affection is obvious here. The subsequent narration goes like this: ‘Baby, where did you grow from? I should search for more; your mother must be sleeping nearby.’) d. Uti zugu-­ziga mine-­we ise:n-­zi: kxxxx amäde:-­ni. this otter-­ dim me-­ acc see.pst.ptcp-­ins.ss hiss back.pst-­3sg ‘When this young otter saw me, it started hissing and moved back.’ (The narrator’s affection was manifested in her intonation and the admission: ‘It was so cute. I love the young wildlife.’) This brief overview of the semantic categories expressed by the nominal diminutive suffix -­ziga suggests that it does not bear any inherent affective connotations. 2.1.2  The diminutive suffix -­ziga in adjectives The suffix -­ziga can also attach to some non-­compound qualitative adjectives, but this does not lead to attenuation of the feature expressed by the adjective; rather, a diminutive reading is assigned to the noun. The diminutive suffix -­ziga in adjectives (but not in nouns) is often preceded by the emphatic marker -­i, for example uligdig’a-­i-­ziga ‘cute’. Most frequently, the suffix -­ziga attaches to parametric adjectives with a base denoting a low degree of a measurable feature, as in (7), and is preceded by the fossilised diminutive suffix -­s’A/-­c’A, which will be addressed in section 2.2.2. (7) a. xaŋus’a-­ziga b. ŋic’a-­i-­ziga c. nemnec’e-­ziga d. enimes’e-­i-­ziga e. da:s’a-­i-­ziga

‘very shallow; small and shallow’ ‘very little, tiny’ ‘very thin; small and thin’ ‘very light; small and light’ ‘very near’

(xaŋus’a ‘shallow’) (ŋic’a ‘small’) (nemnec’e ‘thin’) (enimes’e ‘light’) (da:s’a ‘near’)

This forms double diminutives, and each of them can be emphasised in its turn, thus intensifying diminution as in (8): (8) nientile:-­ni uti konz’o-­ziga-­i ŋic’a-­i-­ziga-­i-­de. open.pst-­3sg that chest-­ dim-­refl small-­ emph-­dim-­emph-­foc ‘She opened her (jewellery) box, a small box.’ In adjectives, the diminutive suffix -­ziga follows the plural suffix -­ŋku: (9) a. uligdig’a-­ŋku-­ziga b. ŋic’a-­ŋku-­ziga nice-­ pl-­dim small-­pl-­dim ‘cute (plural)’ ‘little (plural)’ The examples discussed show that with adjectives, the suffix -­ziga is more selective and is more likely to express affection (especially when accompanied by the emphatic form), while with nouns it has a wider distribution and a purely diminutive meaning.

Udihe 337 2.2  The diminutive suffix -­s’A/-­sA/-­c’A The second morphological way in which Udihe indicates a referent’s diminution is the suffix -­s’A/-­sA/-­c’A. There is considerable variation from speaker to speaker with respect to vowel laryngealisation in this suffix, which influences the stress pattern (laryngealised allomorphs are always stressed, according to the stress assignment rules). In the modern Udihe language, these variants have basically the same meaning, although they seem to go back to two different Proto-­Tungusic affixes (-­sA < *čA:n, and -­s’A < *čA:kA:n). 2.2.1  The diminutive suffix -­s’A/-­sA with nouns This suffix is semi-­productive and occurs mostly in designations of household items, some animals (often females) and persons, as well as landscape features. It is not compatible with first names or recent borrowings. Affective meaning is not typical of nouns derived with this suffix. For the words cited in (10), the base stem is present in the modern Udihe language, while for examples in (11) the base is recoverable only from related languages, and the primary diminutive meaning has bleached. (10) a. mama-­sa b. mafa-­sa c. joxo-­so d. buga-­sa e. kada-­sa f. ŋamakta-­sa

‘wife, old woman’ mama ‘older woman from the same kin’ ‘old man’ mafa ‘old man, husband’ ‘pot, small kettle’ joxo ‘caldron’ ‘small island’ buga ‘island’ ‘small cliff’ kada ‘cliff’ ‘small mosquito’ ŋamakta ‘mosquito’

(11) a. logoso ‘stag, (elk) bull’ b. jandasa ‘badger’ c. giuse ‘roe’ d. kekese ‘servant’ e. xaŋusa ‘small river shallows’ f. bä:sa ‘brook, river’

Negidal lohočo:n ‘(elk) bull’ Oroch jandaku ‘badger’ Nanai giu ‘roe’ Oroch keke ‘female servant’ Oroch xaŋu ‘river shallows’ Evenki bira-­ka:n ‘brook’

Occasionally the two diminutive suffixes co-­occur, as in kuliga-­sa-­ziga

2.2.2  The diminutive suffix -­s’A/-­c’A with adjectives The diminutive suffix -­s’A typical of nouns occurs with adjectives as well, where it has one more variant: -­c’A. This suffix is non-­productive with adjectives in present-­day Udihe, but it can be found in about a dozen words whose base already includes the semantic component ‘small’. In addition to the double diminutives introduced in (7), further examples are given in (12): (12) a. emus’e ‘alone, lonely’ b. neptes’e ‘low’ c. umac’a ‘short’ d. wac’a ‘few, not numerous’

(omo ‘one’) (nepte-­ligi ‘flat’) (Oroch u:mi ‘short’) (Oroch ujaka: ‘few, not numerous’)

338

Maria Tolskaya

Although the recursive use of affixes is not characteristic of Udihe non-­ideophonic adjectives, the diminutive suffix -­c’a can attach to the word ŋic’a ‘small’: ŋic’a-­c’a ‘very, very small’. There are also other affixes of deadjectival derivation in Udihe (Nikolaeva and Tolskaya 2001, 185–8). However, the semantic modifications that they bring to the base concern the degree of a feature (approximation, attenuation or intensification) rather than evaluation proper, and therefore they will be not discussed here.

3  The morphological and expressive status of diminutives The previous sections have shown that the status of the two diminutive suffixes in question may be different. Their semantics and irrelevance to the syntax suggest their derivational character, but the unusual position of the diminutive suffix -­ziga after the plural affix -­ŋku in adjectives seems to contradict Greenberg’s Universal 28: derivational affixes are expected to appear closer to the root than inflectional ones. Taking into account their mutual order in adjectives (-­s’a-­ziga) and dissimilar behaviour with respect to vowel harmony, as well as different distribution (the suffix -­s’A/-­c’A has an idiosyncratic compatibility, while -­ziga is very undiscriminating), we can conclude that the suffix -­s’A/-­c’A has more cohesion with the base and shows more canonical derivational features than -­ziga, a less prototypical representative of derivational morphology. If we turn to the expressive dimension, the occurrence of diminutive suffixes is mostly determined semantically, and their affective connotations, if any, are usually supported by a special emphatic form and intonation. The speakers of Udihe do not employ diminutives as a communicative ‘positive politeness’ strategy, or to mark endearment and empathy. Expressive functions are taken over by other linguistic means, for example by ideophones, which comprise a large part of the lexicon in Udihe (Tolskaya 2012). Endearment can be expressed lexically (e.g. by means of the uninflecting postpositional word baja-­ni in addresses: aziga baja-­ni ‘dear girl’). Attenuation, singularity or a small amount of the affected object are typically rendered by the semelfactive verbal suffix -­ndA ‘a little bit’, for example alasi-­nde ‘wait a little’.

4  Evaluatives from the comparative and language acquisition perspective Despite a high degree of structural similarity in Tungusic languages, they differ with respect to both diversity of their evaluative inventory and their saturation of evaluatives. Udihe and other Southern Tungusic languages have lost most of the rich Tungusic ­evaluative morphology and rarely employ the remaining suffixes, while the languages of the Northern group (especially Evenki and Even) abound in diminutive, augmentative, pejorative and ameliorative suffixes that can be attached to members of various lexical categories, including first names and personal pronouns, verbs and particles. The frequency of diminutives in Udihe folk tales, narratives, songs and riddles is also considerably lower than in these genres in Evenki and Even. Among the factors that influence the evaluative abundance or impoverishment may be the morphosyntactic type of the language. Thus, Evenki and Even have developed more traits of fusion in their overall agglutinative morphology, where the case suffixes show morphophonemic alternations, depending upon the final consonant of the base (Tsintsius 1947); therefore, attaching a diminutive suffix results in the unification of case paradigms.

Udihe 339 In Udihe, the nominal morphology is quite transparent with few exceptions, and the diminutive suffix -­ziga does not further simplify morphological or phonological rules. The diminutive impoverishment in Udihe may have the same source as in Turkish, where diminutivisation reportedly does not facilitate language acquisition (Ketrez and Aksu-­Koç 2007). The distribution of evaluative suffixes roughly corresponds to the major division between the Northern and Southern Manchu-­ Tungusic languages, as illustrated in Table 15.16.1.

Northern Manchu-­Tungusic

Table 15.16.1. Manchu-­Tungusic evaluative suffixes and their functions Language

Morpheme

Function

Example

Evenki

-­čan

diminutive, pejorative

Even

-­kAn

diminutive, affection

biraka:-­čan ‘brook’ asi-­čan ‘vicious woman’ tolgoki-­kan ‘small sled’

-­tkAn

diminutive

beje-­tken ‘boy’

-­kAkun

augmentative, endearment

-­pčAne

augmentative

-­nde:/-­ndä -­mija -­mi

augmentative

-­kun

augmentative

mo:ty-­kakun ‘huge elk’ si:-­ke:ku:n ‘you, poor  thing’ bira-­pčane ‘what a large  river’ beje-­nde: ‘huge man’ bira-­mija ‘large river’ gule-­mi ‘old shabby  house’ beje-­kun ‘huge man’

-­kAt -­vlA

endearment (with  verbs) pejorative (with verbs)

a:sin-­kat-­kel ‘sleep, my  dear’ pektyre-­vle-­ ‘fire a gun  badly’

-­kAn

diminutive

kuŋa-­kan ‘little child’

-­kAjA

augmentative

dü-­kaja ‘large house’

-­jAkAn

pejorative (decrepit)

-­mkAr

augmentative/  ameliorative pejorative (decrepit)

adal-­jakan ‘old shabby  net’ dü-­mkar ‘what a large  house’ unta-­mija ‘shabby shoes’

-­mijA -­ndä Negidal

Oroqen

pejorative

augmentative/  endearment

-­hA:jA:/-­ augmentative kA:jA: -­hA:n/-­kA:n diminutive -­nžA

augmentative

-­kAn, -­hAn -­tʃən

diminutive/  endearment

Reference

Vasilevich  (1940) Nedjalkov  (1997)

Tsintsius  (1947)

ama-­ndä ‘dear father’ beje-­ha:ja: ‘huge man’ beja-­ha:n/ beja-­ka:n  ‘brook’ beje-­nže ‘huge man’ bəyə-­kən ‘dear/little  person’ bira-­han ‘small river’

Tsintsius  (1982)

Whaley and  Li (1998)

Maria Tolskaya

340 Table 15.16.1. (continued) Language

Southern Manchu-­Tungusic

Udihe

Oroch

Nanai

Morpheme

Ulcha

Manchu

Example

Reference Schneider  (1936) Nikolaeva and Tolskaya (2001)

-­s’A/-­c’A

diminutive

joxo-­so ‘small kettle’

-­ziga/-­zig’a

diminutive

ni:-­ziga ‘little man’ xaŋus’a-­ziga ‘very  shallow’

-­kA(n)

diminutive

ga:-­ka:(n) ‘small twig’

-­čkA:(n)

diminutive

hi:kke-­čke:(n) ‘boy’

-­kA(n)

diminutive/ endearment diminutive

inda-­kan ‘puppy’

-­čA(n) Orok

Function

mama-­čan ‘old woman’

-­kA

diminutive

nemde-­ke ‘very narrow’

-­tA

diminutive

duku-­ta ‘little house’

-­ka:(n)

diminutive

maŋgu-­ka:(n) ‘brook’

-­ča(n)

diminutive

-­kA:ku:t

augmentative (intensification)

maŋgu-­ča(n) ‘small  river’ ade-­ka:ku:t-­če-­mi: ‘fast asleep’

-­gAn, -­kAn

diminutive, attenuation

bira-­gan ‘small river’ jilesi-­ken ‘a little  forward’

Avrorin and  Boldyrev (2001) Avrorin  (1959) Petrova  (1967)

Sunik (1985)

Gorelova  (2002)

Note 1. Along with the regular diminutive xokto-­ziga ‘small trail’ < xokto ‘trail’, the same meaning can be expressed sound-­symbolically: xoktüö: ‘tracks of small animals’ (Schneider 1936, 81). This seems to be the only example of magnitude sound symbolism in the conventional Udihe lexicon.

16  South-­East Asia and Oceania 16.1 Apma

Cindy Schneider

1 Introduction 1.1  Background Apma is an Oceanic language of the Austronesian family.1It is actively spoken by 8,000 people in central Pentecost, Vanuatu, which makes it a relatively large language for this linguistically diverse country. Map 16.1.1 shows Pentecost’s location within the island nation, and Map 16.1.2 shows Apma’s place within Pentecost Island.

Map 16.1.1  Vanuatu (www.worldofmaps.net, reproduced under the Creative Commons Attribution-­ Share Alike 3.0 licence)

342

Cindy Schneider

Map 16.1.2.  The languages of Pentecost Island (Gray 2012, 39)

1.2  Typological profile Apma has a (Cons) Vwl (Cons) syllable structure; (usually) penultimate stress; ten vowels (five short vowels and their elongated counterparts) and seventeen consonants. Nominative–accusative relations are flagged by a fairly rigid SV/AVP constituent order. Apma sentences also grammatically encode passivisation, which is an uncommon feature amongst Oceanic languages. Pronominal distinctions are inclusive/exclusive and singular/dual/plural; there is no gender distinction. Nouns are directly or indirectly possessed, with four indirect possession categories and an ‘associative construction’ encoding non-­controlling relationships between NPs. Apma nouns take little morphology. Some bound nouns require pronominal affixation, but most nouns are free. The class of adjectives falls into two main categories. The ‘type 1’ adjectives can take a nominalising suffix. The ‘type 2’ adjectives formally resemble verbs but, unlike verbs, have a simple morphology. The verb, with the VP that contains it, is the most morphologically complex syntactic category. It is made up of up to three phonological words: (1) an optional preverbal complex; (2) a required head; and (3) an optional postverbal complex. Each is morphologically complex. The preverbal complex contains a subject pronoun which hosts an aspect/modality marker, followed by optional morphemes, including

Apma 343 the clitics gam ‘minimiser (min)’ and mu ‘additional (add)’ (Section 2.2). The verb head itself takes morphology, including the partitive marker te (Section 2.3). Finally, overt direct object NPs appear within the postverbal complex and can host their own dependent morphemes.

2  Evaluative morphology Evaluative morphology (EM) in any given language must meet certain formal and semantic conditions (Grandi 2005, 1). Crucially, it requires an identifiable form that carries one of the semantic values of BIG/SMALL/GOOD/BAD. This form must relate morphologically to a recognisable, lexically autonomous base form. Semantically, the meaning of the base changes to accommodate the addition of the evaluative form. EM plays a small but productive role in Apma. Nouns, verbs and adjectives are reduplicated, and their formal by-­products express modified concepts of either diminution/ distribution, augmentation/extension or intensity (Section 2.1). The VP can optionally incorporate minimising and/or augmenting morphemes into its structure (Section 2.2). And the partitive marker, which as a clitic moves between nouns and verbs, has an evaluative function when attached to verbs (Section 2.3). The evaluative morphemes discussed herein are non-­referential. Instead, they have an attributive function whereby they signal a hyponymy relation between a base form and its derivative. 2.1 Reduplication Reduplication is a productive process used on noun, verb and adjective roots to express a variety of evaluative concepts. Except for the bound roots in Table 16.1.4, where reduplication occurs from left to right, all other roots reduplicate from right to left, and the process is bimoraic.2 That is, either two consecutive light (Cons Vwl or Vwl Cons) syllables are reduplicated, or one heavy one (Cons Vwl Cons or Cons Vwl:) is. 2.1.1  Reduplication of noun roots Reduplication of the noun root leads to one of two semantic results. It can either augment the quantity denoted in the noun’s base form whilst leaving its size intact (BIG), or conversely, it can pluralise the base form while at the same time minimising and dispersing it (SMALL). Bauer (1997, 545) argues that plural marking is actually not inflectional in many languages. Indeed, in Apma, reduplication is only a secondary strategy for expressing plurality. Pluralisation is instead productively encoded when nii ‘plural (pl)’ follows the noun head. In cases where the noun is both reduplicated and also followed by nii, the effect of the pluralisation seems to be more pronounced. That is, the notion of ‘plenty’ factors in, as implied by ‘many trees’ or ‘lots of wood’ in Table 16.1.1 below. The examples listed as ‘SMALL’ in Table 16.1.1 illustrate how noun root reduplication can ‘scatter’ or disperse the noun, with the reduplicated referents being reduced in size in comparison to the base. Tsibi~tsibikte ‘ancestors’ in Table 16.1.1 could, arguably, be viewed as a sort of attenuation/diminution on the base form, tsibikte ‘grandparent’. 2.1.2  Reduplication of verb roots Verb root reduplication has two major functions: to augment, extend, continue or habitualise the action or event, on the basis of an objective criterial norm (BIG), or to intensify

Cindy Schneider

344 Table 16.1.1. Noun root reduplication in Apma Semantic effect of reduplication

Base form

Meaning

Reduplicated form

Meaning

Augmentation of quantity (BIG)

apma butsuka leut maasup

‘something’ ‘tree’ ‘something’ ‘dust’

apm[a]~apma nii butsu~butsuka nii le~leut nii maa~maasup

malikte

‘old thing’

mal~mali

nguduka nutsun

ngudu~nguduka nii nutsu~nutsun

wakte

‘wood’ ‘(his or her)  child’ ‘strips of  pandanus leaf’ ‘thin one’

‘many things’ ‘many trees’ ‘many kinds of things’ ‘clear area (contains only  dust)’ ‘old things lying  around’ ‘lots of wood’ ‘(his or her) children’

huralan masen mwel songo tsibikte

‘a walk’ ‘space, gap’ ‘cycad tree’ ‘dirt’ ‘grandparent’

rumu

Diminution; dispersal (SMALL) ?

rumu~rumuk

‘strips of pandanus leaf  bound together’

wak~wakte

‘thin ones’

hural~huralan nii mase~maseut mwel~mwel songo~songo tsibi~tsibikte

‘many short walks’ ‘small plots of land’ ‘a small cycad tree’ ‘bits of dirt’ ‘ancestors’

the portrayal of the event (GOOD). Table 16.1.2 groups the verbs into the two semantic categories. An interesting counter-­example to the intensification function is mwap ‘burn’ > mwap~mwap ‘be hot’, which appears to be a deintensification. 2.1.3  Reduplication of adjective roots Reduplication of adjectives causes their meaning to intensify; this is a reliable outcome, as shown in Table 16.1.3. Some type 2 adjectives occur as bound roots, and reduplication proceeds from left to right, rather than vice versa (Table 16.1.4). However the semantic effect is the same. 2.1.4  Reduplication: summary Reduplication on noun, verb and adjective roots creates evaluative meanings such as diminution/dispersal, augmentation and intensity. Other syntactic effects are plural marking on nouns, and aspectual marking on verbs. However, there are regular and productive alternatives for marking these grammatical categories, and so it ­reasonable to assert that the primary function of reduplication is derivational, not inflectional. 2.2  ‘Minimisation’/‘addition’ morphemes on the verb Two morphemes in the VP encode ‘minimisation’ (Section 2.2.1) and ‘addition’ (Section 2.2.2). They seem inflectional in that they have optional reserved slots within the

Apma 345 Table 16.1.2. Verb root reduplication in Apma Semantic effect of reduplication

Base form

Meaning

Reduplicated form

Meaning

Augmentation / extension / habitualisation / continuation (BIG)

aldiro bwal dahkuru deng dibwiri

‘visit’ ‘fight’ ‘follow’ ‘cry’ ‘touch’

al~aldiro bwal~bwal dah~tahkuru deng~teng dib~tsibwiri

dooni gabarani huu lep lum min ras riak sal

‘want’ ‘throw out’ ‘learn’ ‘give’ ‘punch’ ‘drink’ ‘be sick’ ‘remain’ ‘glide, float  along’ ‘snatch, grab,  pull’ ‘kill’ ‘come out’

doo~tooni gab~kabarani huu~huu lep~lep lum~lum min~min ras~ras ria~riak sal~sal

‘visit regularly’ ‘continue fighting’ ‘follow everywhere’ ‘cry all the time’ ‘go around touching  everything’ ‘always want’ ‘spread around’ ‘practise’ ‘exchange mats’ ‘fight with fists’ ‘drown’ ‘be sick all the time’ ‘try hard but fail’ ‘float (up and down)’

san~sania

‘shake’

va~vamte va~vaut

‘break’ ‘lower’ ‘arrive’ ‘not exist’ ‘burn’ ‘be dry’ ‘be quiet’ ‘shine strongly’ ‘squeeze’

wat~wat wih~wih wuk~wukngak bulo~bulong mwap~mwap raga~ragah roo~rong sil~sil

‘keep killing’ ‘come out one after the  other’ ‘break into pieces’ ‘duck up and down’ ‘arrive little by little’ ‘not exist anywhere’ ‘be hot’ ‘be very dry’ ‘be very quiet’ ‘blaze down’

wuu~wuuri

‘squeeze hard’

san vamte vaut

Intensification (GOOD)

wat wih wukngak bulong mwap ragah roonga sil wuuri

Table 16.1.3. Adjective root reduplication in Apma – GOOD (intensification) Base form

Meaning

Reduplicated form

Meaning

biri solo nitsu gabis watnede temee

‘small’ ‘long, tall’ ‘small, childlike’ ‘good’ ‘different’ ‘black’

biri~biri sol~sol nitsu~tsu gab~kabis wat~watnede mee~meesik

‘very small’ ‘very long, very tall’ ‘very small, very childlike’ ‘really good’ ‘very different’ ‘very black’

VP a­ rchitecture (normally attaching to the preverbal complex, but sometimes to the verb head itself, depending upon verbal phonotactics). However, their function is derivational, being either to minimise or to augment the base meaning of the verb root. Gam is common and productive, while mu is much rarer.

Cindy Schneider

346

Table 16.1.4. Bound adjective root reduplication in Apma – GOOD (intensification) Base form

Meaning

Reduplicated form

Meaning

-­mkan -­mrang -­mret -­msee

‘be sharp’ ‘be lazy’ ‘be worn out’ ‘be cut’

-­mkan~kan -­mrang~rang -­mret~ret -­msee~see

‘be very sharp’ ‘be very lazy’ ‘be very worn out’ ‘be cut all over the place’

2.2.1  Gam ‘minimiser (min)’ Gam (with allomorphs gam=, ga=, =gam and =ga), is a grammaticalisation of the verb gamra ‘to just do (something)’. It functions to minimise an event. In (1), an elderly woman has just explained the intricacies of pre-­Western cooking. She then contrasts this with the relative ease of using a saucepan, using gam ‘minimiser (min)’ (shown underlined) to emphasise her point: (1)



BAD (contempt): Ta=t=ba=m lel-­i=nga, 1pl=pfv=neg1=add do-­ tr=neg2



gema ah kaamat di, bi go ah ra=t di 1pl.excl.indp rel 1pl.excl.pfv live and one rel 3pl=pfv live

te=gen go=ah 3sg.pfv=like one=prox

niaha, rel

baawo, ani kaa=m=gam iusum sosban nante=ah. before but 1pl.excl=ipfv=min use saucepan nothing.more=prox ‘We don’t do it any more, like us who lived before, and those who were alive before, but we just use a saucepan, nothing more.’ Gam is also used with events that are about to occur, in order to downplay their duration or importance (2): (2) BAD (attenuation): Na=n=ga van ne-­gil-­i taro 1sg=irr=min go link-­dig-­tr taro

katsil, bi three and

na=n=ga mulma. 1sg=irr=min come.back ‘I’ll just go and dig three taros, and then I’ll come right back.’ Gam may express a ‘near miss’; the speaker in (3) uses it to convey how the coconut crab has just beaten his competitor in a race: (3) BAD (near miss): Te=gam dok ne-­van ut go, ba 3sg.pfv=min stay link-­go place other comm

mwe=gen go=ah, 3sg.ipfv=like one=prox

wakatsiwas=ah te sadok=te. coconut.crab=prox 3sg.pfv sit=compl ‘When he got to the other place, this coconut crab was already sitting there.’ (Lit.:

Apma 347 ‘He was just going to the other place, like this, this coconut crab was already sitting there.’) Gam can also be used to express the very recent past. In (4), the speaker is correcting a talk he has just finished. Gam functions more as a tense marker here than as an evaluative morpheme: (4) NOT EVALUATIVE (very recent past): Go=ah niah na=m=gam vep one=prox rel 1sg=ipfv=min say ba comm

tei te foc 3sg.pfv

nehu, tei Liwusvet, comp foc L.

ba=i=te Liwusvet=nga, neg1=be=part L.=neg2

ani tei atsi=ah niaha, tei no-­n kadadago. but foc someone=prox rel foc clf-­3sg.poss guard ‘What I was just saying, that it was Liwusvet, but it wasn’t Liwusvet, but this person was his guard.’ Example (4) demonstrates how gam ‘minimiser (min)’ has a spectrum of functions, some of which fall outside the scope of EM. 2.2.2  Mu ‘additional (add)’ Mu appends a sense of ‘addition’ to the base meaning of the verb. In (5), mu is used twice. In the first instance it relates to sak ‘go up’ to indicate additional distance, that is, ‘go up a little more’. In the second instance, mu is associated with wutihi ‘find’ to convey the meaning ‘find another’: (5) BIG (augmentation): Bi ra=m=ru vep, ‘Ta=n mu=ru sak and 3pl=ipfv=du say 1pl=irr add=du go.up

si, a.little

na bih ta=n mu=ru wutihi te=go.’ 1sg ipfv.think 1pl.incl=irr add=du find part=one ‘And the two of them said, “Let’s go a little further, I think we’ll find another one.”’ Example (6) follows the preface: Once there was a man named Bulemamkan. He had ten wives. The morpheme mu ‘additional (add)’, which modifies the verb di ‘stay, exist’, conveys that a new character in the story ‘also exists’: (6)

BIG (augmentation): Tei lego, bi foc before and



rel man

ah

dalmwa

atsi havin te=mu di, this woman 3sg.pfv=add stay

nii

ra=t hural ne-­tka-­i. 3pl=pfv walk link-­carry-­tr ‘Once there was also this woman who slept with lots of men.’ pl

2.2.3  Gam and mu in relation to each other Within the preverbal complex, gam and mu are ordered consecutively. Co-­occurrence of the two morphemes is possible, but rare, and there appears to be no interaction between

Cindy Schneider

348

them; each simply derives the root in its own predictable way. This is shown in (7), where each contribute its respective ‘minimising’ and ‘additional’ semantics to the meaning of the verb. (7) BAD=BIG: Entorah niah ta=m while rel 1pl.incl=ipfv

ka-­da

ga=mu dahki min=add put.salt.on

le~leut

nii nong, . . . pl now ‘When we just salt our different foods again now, . . .’

clf=1pl.incl.poss various~thing

2.3 Partitive te Partitive te is a clitic that can attach to either the noun or the verb (Schneider 2008). As a nominal marker, it signifies ‘part of whole’ semantics and non-­specificity/indefinite quantity in the NP. However, we are not concerned here with its nominal function. Partitive te also attaches to the VP. In appropriate contexts it is used as a morphosyntactic marker within the verb complex, where it changes the meaning of the base. In affirmative statements, it signifies that the action of the verb is not fully executed, or that it is carried out in an unconfident or non-­rigorous manner. Pragmatically, it conveys deferentiality and politeness. On the other hand, partitive marking makes negative statements stronger than they otherwise would be, and emphasises that an event is absolutely and completely unfulfilled. Like gam ‘minimiser (min)’ and mu ‘additional (add)’, partitive te occupies an optional reserved slot within the VP. However there are periphrastic alternatives to the partitive and, on balance, its function is more derivational than inflectional. 2.3.1  Partitive marking in the affirmative The partitive conveys a sense of ‘partial execution’ or ‘attempt’ at completing an event, whether the verb is transitive or intransitive. In (8), siba ‘peel’ acquires a sense of non-­completion when partitive te is encliticised. The repetition of siba further suggests that the agent is engaged in not a single decisive action, but a series of tentative ones. The final instantiation of siba ‘peel’ is not marked for partitive, because the speaker anticipates that the action will come to an end: (8) BAD (approximation/partial execution): Nema siba=te ba, nema siba=te ba, 3sg.prsp peel=part comm 3sg.prsp peel=part comm Mabonmwel, nema siba i biri bu. M. 3sg.prsp peel ins small knife ‘She’ll peel it, she’ll peel it, Mabonmwel, she’ll peel it with the small knife.’ In (9), an attempt to break is marked with the partitive (9a), whereas a successful execution of the verb instead takes transitive marking (9b): (9) a. BAD (approximation/partial execution): Kaa=ga mu=bma ne-­bwah=te.

Apma 349

2pl=min add=come link-­break=part ‘Now you guys come and try to break it.’

b. Mwa=bwah-­a, ra=mwa bwah-­a vet nong. 3sg.ipfv=break-­tr 3pl=ipfv break-­ tr stone this ‘He breaks it, they break this stone.’ In (10), te ‘partitive (part)’ attaches to the verb as an indication that the activity, vep ‘talk’, may be executed less than fully: (10) BAD (reduction/partial execution/lower social position): Ani na=n vep=te nge teweb. but 1sg=irr talk=part just a.little.bit ‘But I’ll just talk a little bit.’ The speaker’s pragmatic message expresses ‘negative politeness’: he wants to avoid imposing upon his listener by taking up too much time. 2.3.2  Partitive marking in the negative While partitive in the affirmative codes an event as partially or tentatively operative, partitive marking on negated verbs has the converse effect: it suggests that the action is not somewhat, but completely, unfulfilled. In (11) a mother, fearful for her life, is pleading with her vengeful son, swearing that she had never abandoned him. She uses te ‘partitive (part)’ with mkoo ‘abandon’ to emphasise this fact: (11)

GOOD (intensification): Nutsu-­k tewot, ko=bma, na=t ba=mkoo child-­ 1sg.poss beloved 2sg=come 1sg=pfv neg1=abandon te=nga i kik! part=neg2 ind.obj 2sg.obj ‘My beloved child, come here, I never abandoned you at all!’

The following pair of examples contrasts the unmarked non-­partitive in (12a) with its more definite, fully unexecuted partitive counterpart in (12b): (12) a. Na=t=ba git-­a=nga. 1sg=pfv=neg1 find-­tr=neg2 ‘I can’t find it (but it might be somewhere else).’ b.

GOOD (intensification): Na=t=ba git-­a te=nga. 1sg=pfv=neg1 find-­tr part=neg2 ‘I can’t find it anywhere (and I’ve finished looking).’

3 Summary EM has a limited but productive role in Apma. In Figure 16.1.1, the derivational semantic functions of Apma EM are mapped onto the general typological representation of

Cindy Schneider

350

EM as presented in the introductory chapter to this volume. We can see that Apma’s BIG and SMALL categories conform to Nieuwenhuis’ proposed hierarchy for BIG ­(augmentatives) and SMALL (diminutatives) in EM: Noun > Adjective/Verb > Adverb/ Numeral/Pronoun/Interjection > Determiner (Nieuwenhuis 1985, 221, as cited by Bauer 1997, 540). For example, since verb roots take EM in Apma, then we would also expect the same from nouns (because nouns are higher on the hierarchy). This is indeed the case. While Figure 16.1.1 presents Apma EM in absolute terms, the examples in Section 2 demonstrate that there are varying degrees of ‘goodness of fit’ within the EM paradigm. For example, noun and verb root reduplication have a partially syntactic function (pluralisation of nouns, and aspectual distinctions on verbs). The other evaluative morphemes (gam ‘minimiser (min)’, mu ‘additional (add)’ and te ‘partitive (part)’) have optional but reserved slots within the VP. This makes them appear superficially inflectional. Yet as unique identifiable forms, all assign meaningful BIG/SMALL/GOOD/BAD values to their hosts and thus modify standard baseline meanings in (usually) predictable ways. Therefore all meet the criteria for being evaluative morphemes. Descriptive perspective

Qualitative perspective

big

good

Positive Noun root reduplication

Augmentation of Verb root quantity reduplication

Verb root reduplication

Augmentation/ extension/ habitualisation/ continuation

VP clitic mu ‘additional (add)’

Augmentation

Negative

Adjective root reduplication

Intensification

VP clitic te ‘partitive (part)’ negative

Intensification

small

Noun root reduplication

Diminution / dispersal

Intensification

bad

VP clitic gam ‘minimiser (min)’

Contempt/ attenuation/ near miss

VP clitic te ‘partitive (part)’ affirmative

Approximation/ partial execution/ reduction/lower social position

Figure 16.1.1.  Apma’s evaluative morphology mapped onto the typology of semantic functions

Apma 351

Acknowledgements Thanks to the many Apma-­speaking people of central Pentecost who worked with me between 2003 and 2006, and who provided me with the data used in this chapter.

Notes 1. Spelt ‘Abma’ in earlier publications by Schneider. 2. Some of the data on reduplication is taken from Gray and Buletangsuu Temwakon (2012).

16.2 Chinese

Giorgio Francesco Arcodia

1 Introduction In many European languages, ‘Chinese’ (French chinois, German Chinesisch, etc.) is commonly used to refer to Standard Mandarin Chinese (henceforth: SMC), i.e. the national language of China;1 in English, this variety is also referred to as ‘Mandarin’ or ‘Mandarin Chinese’. However, ‘Chinese’ could be used to refer to any language (or ‘dialect’) belonging to the Sinitic family of languages, and ‘Mandarin’ is also the name of the largest branch of Sinitic (in terms of number of speakers); thus, it may be used to indicate either SMC or Mandarin dialects as a group. In what follows, we will use the term ‘Chinese’ only when making statements about Sinitic in general or about earlier stages of development of the language; in keeping with the tradition, we shall use the term ‘dialect’ to refer to varieties other than SMC, also because, with the (partial) exception of Cantonese, they lack a standard form (i.e. they have the status of ‘dialects’ in relation to SMC). Varieties other than SMC are indicated with the name of the main centre where they are spoken, followed by the dialect group (e.g. ‘Shanghai Wu’). Chinese, or Sinitic, is a branch of the Sino-­Tibetan language family; the relation between Sinitic and Tibeto-­Burman has been firmly established on the basis of shared cognates, regular sound correspondences and (reconstructed) derivational morphology (LaPolla 2001). However, Sinitic shares many prominent typological characteristics with the non-­Sinitic languages of East and South-­East Asia: analytic and isolating morphology, with the loss of (old) derivational processes (see Section 2), overwhelmingly monosyllabic morphemes, verb-­medial word order (vs. verb-­final for the majority of Tibeto-­Burman languages), complex classifier and tone systems, etc. This is not wholly uniform throughout Sinitic: as may be expected, varieties to the south are closer to the ‘South-­East Asian type’, with stronger isolating tendencies, whereas languages to the north resemble their Altaic neighbours to some extent, with simpler tone systems, fewer classifiers and agglutinating tendencies, and a preference for suffixation (Comrie 2008). Having evolved from Mandarin dialects, SMC is representative of the ‘Northern type’.

2  Morphology: an overview Chinese has virtually no inflection, that is, no obligatory expression of tense, aspect, gender, case, etc., except for number in personal pronouns (SMC 我 wǒ2 ‘I’ vs. 我們 wǒ-­men ‘we’), although this is arguably not a fully grammaticalised category in some dialects (Yue 2003).

Chinese 353 Compounding is by far the most common word-­formation process (Lin 2001), involving the combination of lexical morphemes (free and/or bound), as in SMC: (1) a. 網球 wǎng-­qiú net-­ball ‘tennis’ b. 吹風機 chuī-­fēng-­jī blow-­wind-­machine ‘blowdryer’ Another morphological technique which is very common in Sinitic is reduplication, which may involve nouns, verbs, adjectives, adverbs and classifiers. Reduplication in Sinitic has a rather broad range of meanings, both within and across dialects, including e.g. universal quantification (for classifiers), perfective aspect, repetition of action (for verbs), and many others. Many patterns of reduplication are evaluative in nature, such as Taiwanese Hakka3 驚驚 giang24giang24 ‘somewhat afraid’, vs. 驚 giang24 ‘afraid’ (Lai 2006, 491). Moreover, segmental and/or suprasegmental (tonal) changes in the lexical root may have a morphological function, but this is arguably never found in SMC. There seems to be a general consensus on the reconstruction of a number of subsyllabic derivational affixes for Old Chinese;4 none of these survived into the modern dialects, albeit some vestiges may be found (e.g. in Jin dialects; Sagart 2004). The kind of morphology which flourished in the evolution of Chinese involves the agglutination of monosyllabic morphemes; since in most cases there is no formal differentiation between lexical morphemes and would-­be affixes, it is unclear whether Chinese has productive derivational processes in its present historical stage. For instance, is the (SMC) morpheme 學 xué ‘study, -­logy’ a derivational suffix in words such as: (2) a. 動物學 dòngwù-­xué animal-­study ‘zoology’ b. 語言學 yǔyán-­xué language-­study ‘linguistics’ given that 學 xué can be used as a free form, as in (3)? (3) 我學日語 wǒ xué Rìyǔ 1sg study Japanese ‘I study Japanese’ The issue is, needless to say, far beyond the scope of the present chapter (for an overview, see Arcodia 2012).

354

Giorgio Francesco Arcodia

3  Evaluative morphology Chinese is fairly rich in evaluative morphology, although not all such constructions are still productive. Since a thorough presentation of this domain of word formation in Sinitic languages would require a book-­length study, to say the least, here we can describe in some detail only the situation of SMC, the most studied and researched variety of Chinese; data from other dialects will be also presented, but in an unsystematic fashion. We shall discuss separately concatenative and non-­linear morphology, i.e. reduplication and changes in the root, namely ‘ablaut’ (segmental) and tone change (Bickel and Nichols 2007). 3.1  Evaluative affixes Most of the oldest (syllabic) derivational affixes of Chinese seem to have been evaluative in nature (see Wang 1989, 5–14). The prefix 阿-­ a-­ has been used at least since the fifth century ce as a marker of endearment, and may be added to single-­syllable personal names and kinship terms (Lin 2001, 62): (4) a. 阿寶 ā-­bǎo end-­Bao ‘Bao’ b. 阿姨 ā-­yí end-­maternal.aunt ‘auntie’ This prefix is not used very often in SMC, and it is more common in Southern dialects (Wang 1989). In Hong Kong Cantonese, it is normally prefixed to first names, kinship terms (for elder relatives) and surnames (Matthews and Yip 2011, 429): (5) a. 阿陳 a-­Chán end-­Chan ‘Mr Chan’ b. 阿黃 a-­Wóng end-­Wong ‘Mr Wong’ Note that, when prefixed with 阿-­ a-­, surnames undergo tone change (e.g. Wòhng > Wóng). The prefix 老-­ lǎo-­‘old’ also has a very long history, as it has been in use since the Tang Dynasty (618–907 ce), attaching to personal names, kinship terms and names of animals: (6) a. 老王 lǎo-­Wáng old-­Wang ‘old Wang’

Chinese 355 b. 老兄 lǎo-­xiōng old-­elder.brother ‘brother, old chap’ c. 老虎 lǎo-­hú old-­tiger ‘tiger’ At present, the only productive usage of 老-­ lǎo-­ is the first one, in which the prefix is added to surnames to convey ‘a vague sense of seniority, familiarity, affection or colloquialism’ (Lin 2001, 61); the other cases exemplified here are now fully lexicalised. Chinese also makes use of a prefix 小-­ xiǎo-­ ‘young, small’, which is also attached to personal names: (7)



小李 xiǎo-­Lǐ young-­Li ‘young Li’

Needless to say, this is a less respectful form of address than 老-­ lǎo-­ (Li and Thompson 1981, 37). Two suffixes which certainly deserve to be mentioned here are -­子 -­zi and -­兒 -­r. The morpheme 子 zǐ, bearing the third tone in SMC, has a broad range of meanings, including ‘child’, ‘son’, ‘seed’, etc.; it is found in complex words such as:



(8)

子女 zǐ-­nǚ son-­daughter ‘sons and daughters’

In its use as a bound right-­hand constituent, however, it is toneless (i.e. in the neutral tone): (9) 例子 lì-­zi example ‘example’ Thus, it is one of the very few cases in which affixal status may be argued for (also) on the basis of formal criteria, namely fixed position, bound status and phonetic (suprasegmental) reduction. In its modern usage, -­子 -­zi is generally considered to be a ‘dummy affix’ (Lin 2001, 81), which adds no lexical meaning to the morpheme it attaches to, as appears to be the case in the examples provided above. However, -­子 -­zi originally also had a diminutive value, which is still apparent in a few modern words such as 刀子 dāozi ‘small knife, pocketknife’ (and 大刀子 dà dāozi ‘big (small) knife’ would sound odd; Pan, Ye and Han 2004, 88). Although -­子 -­zi derived words are commonly encountered in SMC, from the synchronic point of view this suffix has, at best, very limited productivity (see Nishimoto 2003).

356

Giorgio Francesco Arcodia

The suffix -­兒 -­r also derives from a noun ‘child’ (SMC ér), arguably the commonest source for diminutives cross-­linguistically (Jurafsky 1996), and is the only non-­syllabic affix of SMC, thus numbering also among the handful of affixes which may be identified on formal criteria. This suffix can already be found in sixth-­century texts, in which it is used to form nicknames; this is argued to be the basis on which it developed its diminutive usage (Wang 1989, 12). Nowadays, -­兒 -­r is still fairly productive; the original diminutive meaning, with a sense of endearment, is preserved more often for words derived with -­兒 -­r than for those derived with -­子 -­zi; compare: (10) 老頭兒 lǎo-­tóu-­r nmls-­end old-­ ‘old man’ (11) 老頭子 lǎo-­tóu-­zi old-­ nmls-­pej? ‘old fogey/codger’ Note also that many words derived with -­子 -­zi and -­兒 -­r which have at present no connection with evaluative (diminutive/endearment) meaning seem to be the product of the typical semantic extensions of CHILD diminutives (Cominetti 2010).5 The use of -­兒 -­r can also be found outside the Mandarin group. In Wenzhou Wu it can convey smallness and endearment, but also pejorative meaning (Zhengzhang 2008, 127–36); thus, its evaluative meaning is much more evident than in SMC. In Central and Southern dialects, cognates of -­兒 -­r are not very common, but there are other suffixes deriving from a noun ‘child/son’, as Hong Kong Cantonese -­仔 -­jái, Leizhou Southern Min -­仔 -­kia42 or Changsha Xiang -­崽 -­tsai41 (Wu 2005, 10, 88): (12) 書仔 syū-­jái book-­ dim ‘booklet’ (Matthews and Yip 2011, 46) (13) 樹仔 ts’iu33-­kia42 tree-­ dim ‘small tree’ (Cai 1993, 120) Again, for these morphs the evaluative flavour is stronger. Leizhou Southern Min -­仔 -­kia42 is particularly interesting, as it can also be attached to verbs, to indicate short duration of action, and in this usage it loses its tone: (14) 鼻仔 pi34-­kia smell-­ att ‘smell a bit’ (Cai 1993, 122)

Chinese 357 3.2  Non-­linear evaluative morphology: reduplication, ablaut, tone change Reduplication of nouns is not productive in SMC, being limited to a few items, yielding meanings in the area of universal quantification: (15)

人人 rén-­rén person-­person ‘everyone’ Most common kinship terms are reduplicated forms, such as 姐姐 jiě-­jie ‘elder sister’, 哥哥 gē-­ge ‘elder brother’; these are fully lexicalised (note the neutral tone of the second syllable), and probably originate from child language (Ōta 1987). Note that Chinese parents often address their children with a reduplicated form of (the last syllable of) their name, as e.g. 亮亮 Liàng-­Liang ‘Liang’, arguably ­yielding ­endearment (Lin 2001, 71–2; cf. 娃娃 wá-­wa ‘child-­child = baby, doll’). This  kind of (weakly) ­diminutive/endearment reduplication is also found in child language and baby talk in Hong Kong Cantonese (Matthews and Yip 2011). In Shanghai Wu, both ­diminutive  and ­universal quantification reduplication of nouns are found (Zhu 2006, 58): (16)

花花 hò-­ho flower-­flower ‘small flower’

(17)

家家 kà-­ka family-­family ‘every family’.



Reduplication of adjectives typically conveys vividness and/or intensification, as in SMC 高高 gāo-­gāo ‘(very) tall’. In Taiwanese Southern Min, monosyllabic adjectives are reduplicated for attenuation, whereas they are triplicated for vividness/intensification (Tsao 2004, 287): (18)

紅紅 âng-­âng red-­red ‘reddish’



(19)

紅紅紅 âng-­âng-­âng red-­red-­red ‘red’ (vivid description; cf. the Hakka example in Section 2)

For disyllabic adjectives, ‘diminishing’ reduplication takes the form ABAB, whereas ‘increasing’ reduplication is AABB (Tsao 2004, 296–7):

Giorgio Francesco Arcodia

358 (20)

(21)



老實老實 láusit-­láusit honest-­honest ‘kind of honest’ 老老實實 láu-­láu-­sit-­sit honest-­honest ‘very honest’

There are many functions for verbal reduplication, as hinted above. As to evaluative meaning, in many (if not most) varieties the reduplication of verbs indicates the so-­called ‘delimitative aspect’, i.e. ‘doing an action “a little bit”, or for a short period of time’ (Li and Thompson 1981, 232): (22)

(23)



看看 kàn-­kan look-­look ‘have a look’ (SMC) 幫助幫助 pòngzu-­pòngzu help-­help ‘help (a bit)’ (Shanghai Wu; Zhu 2006, 86)

However, when a compound verb made of coordinate constituents is reduplicated, it generally follows the AABB pattern and indicates intensification/iteration, just as seen above for adjectives, rather than attenuation, as e.g. in SMC:



(24)

跑跑跳跳 pǎo-­pǎo-­tiào-­tiào run-­run-­jump-­jump ‘run about, run and jump in a vivacious way’6

Evaluative meaning may also be expressed through modifications in the lexical morpheme, either at the segmental level (ablaut), at the suprasegmental level (tone change) or both, although this is not found in SMC.7 In Xinyi Cantonese, for instance, the diminutive form of nouns is built by adding a nasal coda and raising the pitch at the end of the tonal contour (Luo 1987, 200–1; tone values after the change are not indicated in the source): (25) a. 豬 tsy53 ‘(big) pig’ b. tsyn ‘small pig’ Note that here the ‘base’ form indicates not normal size, but BIG; with terms of address, the base form indicates someone who belongs to the elder generation, whereas the ‘changed’

Chinese 359 form is used with someone who belongs to the younger generation. This pattern of ablaut/ tone change has a fairly consistent SMALL meaning for most word classes in Xinyi (Luo 1987, 203–9): (26) a. 食飯 sek22-­fan11 eat-­rice ‘eat’ b. seŋ-­fan11 eat.att-­rice ‘eat a bit’ (27) a. 讀多書 tok22 tɔ53 sy53 read more book ‘read more books’ b. tok22 tɔn sy53 read more.att book ‘read a bit more books’ (28) a. 有時讀書 jɐu23si23 tok22 sy53 sometimes read book ‘read books sometimes’ b. jɐu23sin tok22 sy53 sometimes.att read book ‘read books occasionally, sporadically’

4 Conclusion In this chapter, we have shown that modern Sinitic languages are quite rich in evaluative morphology, both concatenative and non-­linear, although this may be not so evident if only SMC is considered. As to the semantic domains of evaluative morphology, SMALL appears to be by far the most developed; note that, while the development MOTHER > AUG is very widespread in East and South-­East Asia, it is probably absent in Sinitic (cf. Matisoff 1991). It seems that SMALL is most often associated with GOOD, whereas ­pejorative meaning is not very common.

Notes 1. Here ‘China’ is meant to include both the People’s Republic of China and the Republic of China (Taiwan), which are de facto controlled by different governments. SMC is also a co-­official language of the Republic of Singapore. 2. Traditional Chinese characters have been used as a default throughout the chapter. The Pinyin system is used for the ­transliteration of SMC; for all other varieties, we use the transcriptions provided by the sources. Tones are indicated either as diacritics or as superscript numbers, depending on the romanisation system.

360

Giorgio Francesco Arcodia

3. ‘Taiwanese Hakka’ is used here loosely as a cover term for the Hakka dialects spoken in Taiwan. The transcription of this example represents the Siyen (四縣, SMC Sìxiàn) variety. 4. In the broadest sense, the language from the earliest written records until the third century ce. 5. In Beijing casual speech, -­兒 -­r is also sometimes used just as a substitute for other syllables, as in 今兒 jīnr for 今天 jīntiān ‘today’. 6. Note that, whereas adjectival reduplication and the reduplication of coordinated verbs are morphological in nature, other patterns of verb reduplication could also be understood as syntactic phenomena, given e.g. that aspect particles may appear between base and reduplicant. 7. Note, however, that in SMC if -­兒 -­r is added to a closed syllable, the coda is replaced by -­兒 -­r, as e.g. in 熊 xiǒng [çyŋ] ‘bear’ > 熊兒 xiǒngr [çyə̃r] (Sun 2006, 38).

16.3 Lisu

David Bradley

1 Introduction Lisu (ISO 639–3 lis) is spoken by over a million people in south-­western China, north-­ eastern Burma/Myanmar, northern Thailand and north-­eastern India. Like most other Tibeto-­Burman languages, it is verb-­final with mainly postposed marking; for more information, see Bradley (2003; forthcoming). Like all other Burmic languages, it has no verbal person or number agreement morphology; tense–aspect, modality, epistemicity and so on are marked by various postverbal elements. All of these apart from a few clause-­final markers are phonologically independent words; modals are also syntactically independent words. The only pre-­head verbal elements are negation markers. Nominal case marking is also post-­head; core case marking is noun phrase final, non-­obligatory and fairly infrequent. There is also no number marking on any Lisu nominal forms, other than second and third person animate pronouns, which may have the plural suffix -­wà. Lisu forms are cited in a transcription reflecting the orthography, which represents a central dialect;1 other dialects have slightly different forms, but exactly the same structures for nominal diminutive and augmentative suffixes; there are also deverbal stative verb2 diminutives and augmentatives used as postnominal modifiers, and diminutive, augmentative and intensifier adverbial nominals which precede the verb. All Lisu examples are cited from a corpus of narrative, conversational and ritual texts, representing the three main dialects.

2  Nominal suffixes Lisu diminutive and augmentative nominal forms have a suffix on the head of the nominal nucleus. Other elements (demonstrative, numeral plus classifier, case marker, topic marker) may follow within the noun phrase. The nominal diminutive suffix is -­zà, which is identical to the nominal stem meaning ‘son/child’. The nominal augmentative suffix is -­dà ma, which is a bound nominal suffix meaning ‘a big one’; there is a homophonous verb dà ‘fetch water’, and ama is a verb nominaliser suffix, but there is no obvious semantic connection between fetching water and being big, and there is no medial a in the augmentative suffix. Most nominal stems have one syllable, but most nominal nuclei have two or sometimes three or more syllables (see examples below); thus most nominal nucleus forms are compounds. Some nominal stems are bound; there are many bound nominal forms which occur only in the second slot in a compound noun, some which occur only in the first slot in a compound noun, and a few which may occur in both. Many nominal heads with a

362

David Bradley

bound lexical second element contain a dummy first syllable. There is a very frequent but non-­productive first syllable a with various tones conditioned by the tone of the following syllable, as seen in ascending generation kin terms like a pa ‘grandfather’, in animal and vegetable terms such as á gù ‘dove’ and à pu ‘cucumber’; in question words such as a li ‘which?’ and grammaticalised question words derived form stative verbs such as a mja ‘how many?’ from the verb mjà ‘many’; in temporal nominals such as à so ‘a short time ago’ as well as in the diminutive, augmentative and intensifier adverbial nominals discussed below; and in many other very frequent nouns such as á thà ‘knife’; for examples in Southern Lisu, see Bradley et al. (2006, 322–39), and in Northern Lisu, see Bradley (1994, 1–11). The other very frequent and completely productive Lisu nominal prefix is jí (Northern dialect é, Southern dialect í); this occurs before almost any bound nominal stem; a bound nominal stem which occurs in second position may occur as an independent head noun, with this prefix as the first syllable, as in jí dzi ‘tree, stem’; otherwise such bound nominal head stems occur after a bound lexical form such as thò-­‘pine’ as in thò dzi ‘pine tree’; the bound forms such as thò-­and -­dzi do not occur independently.3 There is a further slot in the head nominal for a stative verb modifier, following the nominal elements, such as lǽ ‘umarried’, mò ‘old’ and næ ‘black’, as in zà mɯ̀ lǽ ‘unmarried daughter’, xò sá mò ‘old Buddhist monk’ and á gù næ ‘black dove’, resulting in a three-­syllable head noun. As discussed below, the stative verbs wù ‘big’ and ʐo ‘small’ may occur with augmentative and diminutive meanings in this slot. The stative verb diminutive or augmentative modifier does not occur followed by a nominal diminutive or augmentative suffix; *NP -­wù dàma and *NP -­ ʐo zà are ungrammatical; however, the reverse order -­zà -­ʐo does occur, as seen in (3c) below; but -­dà ma -­wù is unattested. The nominal stem zà is extremely frequent as a head noun, and occurs very frequently as the first syllable in two-­syllable compounds such as zà nø ‘child’, zà mɯ̀ ‘daughter’ and zà mɯ ‘wife’; and less often as the second element in two-­syllable compounds as in sá zà ‘son of a sibling of opposite gender from ego’, as well as in initial and final slots in longer compound forms such as zà mɯ zà ‘woman’. In song language,4 it may also occur as a single syllable nominal form. Its core meaning is ‘son’, also extended to ‘child’. This stem may occur with the prefix jí-­ in the meaning ‘son/child’. The formative prefix jí-­ is homophonous with the third person animate pronoun, so a form such as jí dzi can also mean ‘his or her tree’, and likewise jí zà can also mean ‘his or her son/child’. A bound nominal stem may occur with any preceding pronoun as a possessive; this pronoun provides the first nominal element in a compound form, as in ŋa zà ‘my son/child’ (lit. ‘I son/child’). The form zà is also extended to refer to groups of humans of both genders, or to humanity in general. It may appear as a final syllable in a nominal compound, as in tsho pá zà ‘man’, or ‘woman’ as seen above; tsho ‘person’ is a bound form like zà which may occur first or second in a two-­syllable nominal form, as in là tsho ‘people’, tsho mò ‘elders’ and so on; pá is the less frequent of two male suffixes. The extended-­meaning zà is also found as the first and third or second and fourth syllables in four-­syllable compounds such as zà ȿʅ́ zà nø ‘sibling children’; ȿʅ́ here is a bound form usually used as the classifier for groups of siblings (Bradley 2001); and tsho zà wà zà ‘people, humanity’; here wà is a bound form in words for groups of humans, also seen as the plural suffix on second and third person pronouns and as the numeral classifier for humans in Southern Lisu and in some subdialects of Central Lisu. The use of this nominal stem as a diminutive suffix -­zà is quite frequent and very productive; almost any nominal form can have it added, typically producing a nominal head with three syllables such as á thà zà ‘small knife’. In the Southern dialect, the diminutive can also have the reduced form à or æ̀ , but speakers still regard these as derived from zà, and l­iterate

Lisu 363 speakers will usually write it this way, even if they pronounce it as à or æ̀ . Some bound nominal stems occur only with a following zà diminutive suffix, as in tɕhɿ zà ‘string/thread’. In addition, there are a few verbs which have diminutive -­zà as a second syllable, such as ɕǿ zà ‘to pity’, but the vast majority of verbs are one syllable and cannot occur with a following -­zà. (1) a. á ȿʅ̀ ɲi -­zà ŋa la bæ -­ a? what day dim be come say q ‘What little day did you say has come?’ -­zà tȿho je læ. b. á tí tɕhi a.bit thread dim turn away urge ‘Please let’s turn the thread away.’ c. bjà ga ɕǿ -­zà lè bæ -­a lo. bee chase pity dim entirely say decl true ‘(We) truly say that we really pity those who chase bees.’ By contrast, the augmentative -­dà ma is bound and may only occur in the last slot in the nominal head, and must either follow a lexical head nominal, typically of two or more syllables, or the formative dummy prefix jí-­. It is somewhat less frequent than the diminutive use of -­zà, but is completely productive. The form jí dà ma may have its expected meaning ‘a big one’, but is much more often lexicalised to mean ‘God’ (a traditional sky god, also an alternative term for the Christian God) as in (2b). (2) a. là ma -­dà ma thø ma tǽ a li ŋa gɯ ɀu tsɿ‾ ŋo. tiger aug this nmls obj how be also take cause fut ‘This is how (we) will get this big tiger caught.’ b. ʐò jí-­ -­dà ma by tu læ. we nmls aug roar stand urge ‘Let’s roar out and stand up for God.’

3  Stative verb modifiers Lisu also uses its dimensional stative verb wù ‘big’ as an augmentative and ʐo ‘small’ as a diminutive, using the same structure as for other combinations of noun plus stative verb modifier; these compounds may also have the more specific meaning ‘eldest’ and ‘youngest’, especially when following a kinship term. The verbal forms may also be reduplicated to produce an adverbial form; the verbal form wù occurs too as a nominal classifier in the meaning ‘a big one’ after a numeral, and there is a bound nominal form -­wu ‘size’ derived from it. Compared to -­dà ma, the augmentative use of the verb -­wù is somewhat more frequent, especially in song language; however, the diminutive -­zà is much more frequent than the modifier use of the verb -­ ʐo as a diminutive. Example (3c) shows the possible combination of -­zà -­ ʐo mà-­ dʐo. (3) a. mý -­wù mà-­ dʐo -­ȿʅ̀ country aug neg exist thing neg exist ‘There is nothing that the big country does not have.’

364

David Bradley

b. bjà -­ʐo we -­zà xwă -­a. bee dim guest son speak decl ‘The small bee speaks to the guests.’ c. tɕhi bɯ -­dà ma tɕhi bɯ -­zà -­ɀo kwa -­be ŋa -­à mi. banjo aug banjo dim small loc abl be but ‘However, it is from big banjos and small banjos.’ It is possible that a stylistic preference for parallelism in structure may be a factor in why (3c) is possible; this has a pair of four-­syllable nominals, rather than one with three and one with four syllables.

4  Adverbial nominals To express augmentation or diminution of verbal state or action, an adverbial nominal derived from forms nominalised by the prefix a-­ (discussed above) is used. The augmentative is à mjɑ̌ ‘very, a lot’ and the diminutive is á tí ‘a bit’; these normally occur immediately preceding the negative (if any) and then the verb. They are a separate constituent forming a separate word, and are thus unlike verbal constituents which follow the verb. The augmentative is derived from the very frequent dimensional stative verb mjà ‘many’ with the nominalising prefix a and followed by a fused extra syllable á;5 there is also a very frequent deverbal bound nominal form -­mja ‘quantity’; for discussion of grammaticalisation of this and other similar dimensional stative verbs such as wù ‘big’ in Lisu and related languages, see Bradley (1995). The diminutive has the prefix a plus the bound stative verb form -­tí ‘sole, only’, the latter developed from the numeral thì ‘one’ by the addition of a prefix which deaspirates the initial and changes the tone; while this does not occur as an independent verb, it is fairly frequent as a postnominal modifer; it also occurs as a pre-­numeral modifier meaning ‘another’ and as the numeral ‘one’ when added to a round number, as in tshi tí ‘eleven’ (lit. ‘ten one’). The diminutive á tí very frequently carries a bleached meaning of politeness rather than its core meaning of ‘a bit’, especially in imperatives; when it has this bleached meaning, its syntactic position is less fixed and it may occur closer to the beginning of the clause, preceding various other nominal elements. See (1b) above for an example of á tí in sentence-­initial position indicating politeness. The adverbial form á tí may be followed by the nominal diminutive -­zà as in (4b), which further mitigates it. However, à mjɑ̌ dà ma does not occur. ŋa. (4) a. nu á tí lé thà-­ ɕø̌ you a.bit only neg.imp touch be ‘Don’t touch it at all!’ b. á tí -­zà gò le ɲa ɣá kwa tɤ̌ -­o. a.bit dim twist change cond rock loc jump pst ‘When (it) twisted a little bit, (he) jumped on the rock.’ c. à mjǎ ká kɤ ŋo. a.lot stab prf fut ‘(He) will have stabbed (him) many times.’

Lisu 365 There is also an extremely frequent general intensifier form à khɯ́: it occurs seventy-­two times in the corpus, much more frequently than á tí and à mjɑ̌. This is derived from the a-­ prefix plus the verb khɯ́ ‘exceed/go beyond’. Like à mjɑ̌ and á tí, this normally occurs in an immediately preverbal slot; a clause may only contain one of these three adverbial nominal forms, i.e. they do not co-­occur. xa ȿʅ̀ . (5) á mò à khɯ́ dzɤ̀ horse int ride good thing ‘The horse is a very good thing to ride.’

5 Conclusion Table 16.3.1 shows the overall frequency of diminutive and augmentative forms as a proportion of all tokens of the identical forms in the corpus. As Table 16.3.1 shows, postnominal modifier use of stative verbs wù and ɀo as ­diminutive or augmentative is more frequent than their use as main verbs, while nominal use of zà is much more frequent than its use as a diminutive. By far the most frequent use of the syllable /ɀo/ is as the classifier for humans, perhaps derived from the verbal form. As we have seen, the diminutive and augmentative nominal suffix forms of Lisu are asymmetrical: though both are bound suffixes, they have a different number of syllables, and their syntactic status is different: -­dà ma only occurs bound as an augmentative, while the diminutive -­zà is much more frequent in its nominal head use, and has many extended nominal uses, and the augmentative has no current nominal or verbal source and no other uses. It appears that the development of these two Lisu suffixes is a recent phenomenon. The diminutive follows a widespread semantic pathway by grammaticalising a form for ‘son’; but closely related languages such as Lahu and Burmese grammaticalise different forms.6 Lisu productively adds stative verbs as postnominal modifiers; this includes the stative verb for ‘big’ in augmentative and other meanings, and the stative verb for ‘small’ in diminutive and other meanings; this is parallel to many other stative verbs which are ­productively used in the same way. While there is no derivational morphology for indicating augmentation or diminution of state or action attached to verbs in Lisu, there are of course separate adverbial forms which can express this meaning; these precede the verb. In Lisu, these adverbials are a sub-­class of nominals, and are derived by prefixing a nominal prefix a-­ to the source form. Table 16.3.1. Occurrence of diminutives, augmentatives and related forms in Lisu DIM/AUG

Noun

-­zà -­dà ma -­ɀo -­wù á tí

34 30 7 42 17

189 ‘son/child’

à mjă

4

Verb

-­wu ‘size’

4 ‘small’ 29 ‘big’ -­tí ‘sole/only’

-­mja ‘quantity’

mjà ‘many’

Other

-­ɀo ‘human clf’ 1 clf ‘a big one’ tí numeral ‘another/one’

% DIM/AUG 15.2 100 63.6 58.3 (100) (100)

366

David Bradley

Notes 1. The transcription used here represents tones with superscript diacritics: acute accent for a high tone, grave accent for a low tone, haček for a rising tone and mid tone unmarked, as in Bradley with Hope, Fish and Bradley (2006). The mid and low tone both occur either with creaky phonation or with normal phonation; creaky phonation is indicated by underlining the vowel, while normal phonation is unmarked. 2. Note that Lisu and all other related and most unrelated languages in East and mainland South-­East Asia have no adjective form class; adjectival meanings are expressed by stative verbs which function syntactically like other verbs. 3. Many stative and some other verbs can also be nominalised into abstract nouns by adding the same prefix: jí næ ‘blackness’ from næ ‘black’, jí dɀʅ́ ‘a rule’ from dɀʅ́ ‘to rule’, and so on. 4. Song language is stylistically archaic, and is composed of pairs of syntactically parallel seven-­ syllable lines of four syllables plus three syllables; the length and structural restrictions on lines in song language mean that nominal forms are often shorter than in spoken language. 5. The Southern Lisu form is /à mjáà/. 6. Lahu grammaticalises ɛ́ ‘baby’ into a diminutive, while Burmese grammaticalises kh əlè ‘child’.

16.4 Muna

René van den Berg

1 Introduction The Muna language (ISO 639–3 code mnb) is spoken on the island of Muna and some smaller neighbouring islands in the province of Southeast Sulawesi, Indonesia, by an estimated 300,000 speakers.1 Muna is an Austronesian language belonging to the Celebic macro-­group within Western Malayo-­Polynesian (Mead 2003). Major publications relating to Muna are van den Berg (1989; 2004) and van den Berg with La Ode Sidu (1996). All Muna words in this chapter are written in the standard orthography.2 Muna morphology is mostly agglutinative with a rich and diverse array of forms. Prefixation and suffixation are prevalent, but there is also infixation, circumfixation, reduplication (with three subtypes: partial, full and supernumerary), and various combinations of these categories. Less common are compounding, incorporation and abbreviation. The major word classes are verbs and nouns, both of which have extensive morphological possibilities, especially verbs. Verbs are classified into three declension classes (a-­, ae-­and ao-­class, based on the 1sg subject form) and are inflected for subject, direct object and indirect object. In addition, verbs are morphosyntactically grouped into a further three classes: stative (corresponding to adjectives), dynamic (intransitive) and transitive. Closed classes, most of which have few morphological possibilities, are pronouns, numerals and classifiers, quantifiers, adverbs, prepositions, conjunctions, particles and interjections. Inflectional morphology centres on four person–number sets: subject prefixes, direct object suffixes, indirect object suffixes and possessor/agent suffixes (on nouns and on passive participles). TMA morphology is limited to irrealis -­um-­, perfective -­mo and reduplication for continuous action. Verbal derivational morphology is broad, including a variety of valency-­changing affixes (among others causative fo-­1 and feka-­, reciprocal po-­, detransitivising fo-­2, applicative -­ghoo, transitivising -­Ci, requestive fe-­, accidental passive ti-­) and nominalisations (ka-­, -­ha, as well as the circumfix ka-­ .  .  . -­ha). Many affixes, however, appear to be category-­neutral vis-­à-­vis the inflection/derivation distinction, such as ta-­‘only’, -­e ‘vocative’, pe-­‘approximately’ and paka-­‘when first’.

2  Evaluative categories Muna has a relatively rich saturation of evaluative morphology, covering diminution, contempt, approximation, attenuation and intensification. Most of these function productively in the language. In the following subsections each of these five evaluative categories and their

René van den Berg

368

­ orphological exponents are dealt with in detail. Phonology does not seem to play an indem pendent role in evaluative morphology, and so the emphasis will be on formal ­expression and semantics. 2.1 Diminution The combination of the prefix ka-­ and full reduplication (defined as disyllabic and prefixing) on a nominal base normally indicates diminution. Examples are shown in Table 16.4.1. A number of points should be noted about this formation. Firstly, what counts as a ‘small’ token of a particular entity is of course subject to cultural and personal norms. For example, since Muna is relatively flat with only low hills, any elevation in the landscape may count as a ka-­kabha-­kabhawo ‘hill’. Secondly, many formations have developed specific lexicalised meanings. Examples include ina ‘mother’ > ka-­ina-­ina ‘foster mother, wet nurse’; lima ‘hand, arm’ > ka-­lima-­lima (i) ‘handrail’, (ii) ‘pickpocket’, and mata ‘eye’ > ka-­mata-­mata ‘kind of edible shellfish’. In some of these cases a diminutive meaning is still available given the right context. For instance, when admiring a newborn baby, one could comment on its ka-­lima-­lima ‘little hands’ and ka-­mata-­mata ‘little eyes’, overriding their more usual lexicalised meanings. Thirdly, diminutive formation appears to be mainly productive on nouns referring to concrete, specific and tangible entities. Generic nouns such as kalalambu ‘toy’ and kadadi ‘animal’ do not take the diminutive, and neither do natural objects such as gholeo ‘sun’, wula ‘moon’, kawea ‘wind’ and ifi ‘fire’. 2.2 Contempt The combination of the prefix ka-­ and full reduplication on a nominal base may also indicate contempt, but this is a marginal procedure, limited to two elicited examples (discovered while investigating diminution): guru ‘teacher’ > ka-­guru-­guru ‘non-­professional teacher, poorly performing teacher’ and pahulo ‘hunter’ > ka-­pahu-­pahulo ‘poorly performing hunter’. In both cases the formation is applied to a noun referring to a profession, indicating a lack of ability and conveying contempt. 2.3 Approximation Various morphological procedures exist in Muna to signal approximation. The target of the approximation is either a stative verb (translated as an adjective) or a numeral phrase. Table 16.4.1. Diminution in Muna Base form

Meaning

Reduplicated form

Meaning

golu kabhawo kabhera kontu lia raha robhine sau

ball mountain piece (cut off) stone cave, hole platform (for drying) woman, female tree, stick

ka-­golu-­golu ka-­kabha-­kabhawo ka-­kabhe-­kabhera ka-­kontu-­kontu ka-­lia-­lia ka-­raha-­raha ka-­robhi-­robhine ka-­sau-­sau

small ball small mountain, hill small piece (cut off) small stone, pebble small cave, small hole small platform small (and young) female small tree, small stick

Muna 369 Table 16.4.2. Approximation with mba-­in Muna Base form

Meaning

Reduplicated form

Meaning

dea hali lalesa nifi taa wanta

red difficult wide thin good long

no-­mba-­dea-­dea no-­mba-­hali-­hali no-­mba-­lale-­lalesa no-­mba-­nifi-­nifi no-­mba-­taa-­taa no-­mba-­wanta-­wanta

it is rather red, it is reddish it is rather difficult it is rather wide it is rather thin it is rather good, it is OK it is rather long

Table 16.4.3. Approximation with ka-­in Muna Base form

Meaning

Reduplicated form

Meaning

kodoho lolu lui luntu pongke rombu

far stupid tough slow deaf fat

no-­ka-­kodo-­kodoho no-­ka-­lolu-­lolu no-­ka-­lui-­lui no-­ka-­luntu-­luntu no-­ka-­pongke-­pongke no-­ka-­rombu-­rombu

rather far rather stupid rather tough rather slow rather deaf rather fat

2.3.1  mba-­+ reduplication The combination of the prefix mba-­and full reduplication on a stative verbal base carries an approximative meaning, indicating that a certain degree of the state of affairs holds, but that it cannot be properly called that state. English equivalents are ‘somewhat’, ‘rather’, ‘quite’ and ‘–ish’. Table 16.4.2 shows several examples of this productive process. (The prefix no-­marks 3rd person singular realis subject.) In some people’s speech mba-­ is in free variation with ma-­. This ma-­ is also the allomorph found when the stative verbal base contains a prenasalised consonant: mpona ‘long (of time)’ > no-­ma-­mpona-­mpona ‘rather long (of time)’. An example of this formation in a clause is shown in (1): (1)

Do-­po-­kapi-­kapihi welo daoa no-­mba-­hali-­hali rampano 1/3pl.r-­recp-­rdp-­search in market 3sg.r-­approx-­rdp-­difficult because no-­bhari mie. 3sg.r-­many person ‘Looking for each other in the market is rather difficult because there are so many people.’

2.3.2  ka-­+ reduplication A second avenue to expressing approximation is by combining the prefix ka-­ with full reduplication. This procedure is limited to fewer than ten stative verbs, some of which are shown in Table 16.4.3. Approximation with ka-­appears to be limited to stative verbs referring to notions with a negative or undesirable meaning component. With many neutral or positive stative verbs, this formation is impossible: bhala ‘big’ (*no-­ka-­bhala-­bhala), rubu ‘small’, maho ‘near’, kesa ‘beautiful’, but also the inherently negative dai ‘broken; ugly’. On the basis of kadoho ‘far’, both no-­ka-­kodo-­kodoho and no-­mba-­kodo-­kodoho are possible, both meaning

René van den Berg

370 Table 16.4.4. Approximation with ka-­on taste words in Muna Base form

Meaning

Reduplicated form

Meaning

kolo

sour

rather sour

meko

sweet

tembe

insipid, tasteless

no-­ka-­kolo-­kolo ~ no-­ka-­olo-­olo no-­ka-­meko-­meko ~ no-­ka-­eko-­eko no-­ka-­embe-­embe

rather sweet rather insipid, rather tasteless

‘rather far’, without any clear difference in meaning. The negative meaning overtones of ka-­could be related to various other ka-­+ reduplication derivations referring to negative or undesirable character traits. In most cases the simple base does not exist. Examples include no-­ka-­dhoro-­dhoro ‘unmannered, arrogant’ (*dhoro); no-­ka-­hao-­hao ‘greedy’ (*hao) and no-­ka-­duso-­duso ‘hasty, reckless’ (*duso). Three stative verbs referring to taste sensations display the unusual feature of (optional) initial consonant apheresis, as shown in Table 16.4.4. 2.3.3 -­hi The suffix -­hi is a multi-­purpose affix. In addition to indicating plurality and diversity on nouns and verbs, it may also signal that the state or event is part of a larger series or that the action was done at leisure, without specific purpose. For example, do-­fumaa-­hi (3pl.r-­eat-­plur) may mean (i) ‘they all eat’, (ii) ‘they eat (among other things)’ and (iii) ‘they eat (at their leisure)’. With a few stative verbs, it appears that -­hi may also have an approximating meaning, as in ne-­taa-­hi ‘it is rather good, it is OK’ (from taa ‘good’) and no-­rubu-­hi ‘it is rather small’ (from rubu ‘small’). Since the ‘part of a larger series’ meaning appears to be the most accessible, some language consultants do not accept the meaning ‘rather’ in these cases, and most textual examples are in fact ambiguous. For example, a bird is described as ne-­ ngkonu-­hi parewa-­no (3sg.r-­round-­plur body-­3sg.poss), which is either ‘its body is rather round’ or ‘its body is round (among other things)’, in the context of a longer description. The use of mba-­+ reduplication would disambiguate in this case. 2.3.4  sa-­. . . ha-­no A rather special formation on the basis of stative verbs is the combination of the prefix sa-­ and the suffix -­ha, followed by a possessive suffix, which is almost always 3sg -­no, though other person–number combinations are also possible. Allomorphs of sa-­are sae-­(with ae-­ verbs) and sao-­(with ao-­verbs). The resulting formation is not so much an approximation, but rather means that a state X has been reached to a minimal degree in the perception of the speaker. English translation equivalents are ‘only just X’, ‘just X enough’, ‘(just) barely X’, ‘just about X’, ‘not really X’, ‘almost not X’. Table 16.4.5 illustrates some examples. The following show clausal examples: (2) Tee aini sao-­meko-­ha-­no. tea this minimal-­sweet-­degree-­3sg.poss ‘This tea is barely sweet enough.’ (3) Ne-­fanahi oe sao-­pana-­ha-­no. 3sg.r-­heat water minimal-­hot-­degree-­3sg.poss ‘She heated the water until it was just about hot (but not boiling).’

Muna 371 Table 16.4.5. Minimal degree with sa-­. . . -­ha-­no in Muna Base form

Meaning

Affixed form

Meaning

bhie lowu meko ntalea pana pande

heavy drunk sweet light (not dark) hot able, clever

sao-­bhie-­ha-­no sao-­lowu-­ha-­no sao-­meko-­ha-­no sae-­ntalea-­ha-­no sao-­pana-­ha-­no sa-­pande-­ha-­no

only just heavy, not really heavy just about drunk, barely drunk only just sweet, not too sweet just barely light, not really light just barely hot, not really hot just barely able, not really clever

(4) No-­ghuse sao-­bhie-­ha-­no. 3sg.r-­rain minimal-­heavy-­degree-­3sg.poss ‘It is raining, but not really that heavy.’ (Lit. ‘It rains just about heavy.’) It seems helpful to think of this construction in terms of grading along a scale of 1–10, where 5 is a fail and 6 a pass. If a cup of tea is considered sao-­meko-­ha-­no, this means it just narrowly scores a 6 for being sweet. If, however, it is no-­mba-­meko-­meko ‘rather sweet’, its score is around 7–8. In some of the illustrative examples there appears to be a temporal meaning element, translated as ‘until’. It appears that (at least historically) this formation is related to the prefix sa-­, signalling a temporal connection as in the sentence sa-­rato-­ku, no-­kala-­mo (when-­arrive-­1sg.poss 3sg.r-­go-­prf) ‘as soon as I arrived, he went away’. In other examples this temporal element seems to have disappeared and only the resulting state is in focus. This sa-­ .  .  . -­ha-­no formation is also possible on intransitive verbs, as in sa-­toolea-­ha-­no (minimal-­contain-­degree-­3sg.poss) ‘just barely able to contain, just the right size’, and sa-­wanu-­ha-­ku (minimal-­get.up-­degree-­1sg.poss) ‘I am just able to get up (but I cannot really walk).’ This usage is rather rare and mostly occurs in elicited material. In these cases there is the nuance of ability. Finally, on the basis of the stative verb taa ‘good’, the derivation sae-­taa-­ha-­no ‘just about good’ has acquired the specialised meaning ‘medium, around the middle of the scale, average’, when followed by a nominalised stative verb. An example is: (5) Nunsu-­no sae-­taa-­ha-­no ka-­bhala-­no beak-­ 3sg.poss minimal-­good-­degree-­3sg.poss nmls-­big-­3sg.poss ‘Its beak is of average size.’ (Lit. ‘Its beak, its bigness is just about good.’) 2.3.5  pe-­ The prefix pe-­ signals approximation with numbers. It is found with so-­called measure phrases, that is, phrases consisting of a prefixed numeral and either a classifier or a measure noun. It also occurs with numeral verbs. The meaning of pe-­in each case is ‘approximately, about’, and as illustrated in (6), pe-­ must co-­occur with the irrealis prefix na-­ in measure phrases, or with the irrealis set of plural subject markers in numeral verbs, as illustrated in (7). (6) a. na-­pe-­se-­dhamu irr-­approx-­one-­hour

‘about one hour’

René van den Berg

372

b. na-­pe-­lima-­wula-­mo ‘about five months ago’ irr-­approx-­five-­moon-­prf c. na-­pe-­tolu-­folu ue ‘about 60 cm’ (lit. ‘about 30 knucklebones’) irr-­approx-­three-­ten knucklebone d. na-­pe-­raa-­moghono ghulu ‘about two hundred (animals)’ irr-­approx-­two-­hundred body (7) a. da-­pe-­tolu-­fulu ‘there are about thirty of us/them’ 1pl.incl/3pl.irr-­approx-­three-­ten b. da-­pe-­raa-­moghono ‘there are about two hundred of us/them’ 1pl.incl/3pl.irr-­approx-­two-­hundred 2.3.6 Reduplication Reduplication in Muna is a very rich and rewarding area for research, as it has several forms and many meanings, some of which overlap. Reduplication typically indicates continuous or durative aspect, intensification, distribution, plurality (versus duality) or personification. Occasionally, full reduplication may have an approximating meaning, as shown by the following pair: rambi se-­mata (lit. ‘o’clock one-­eye’) ‘one o’clock’ and rambi-­rambi se-­mata (lit. ‘rdp-­o’clock one-­eye’) ‘around one o’clock’. 2.4  Attenuation and reduction 2.4.1  poka-­and reduplication The combination of the prefix poka-­ with full reduplication of the verbal base (either intransitive or transitive) is a productive process in which the added meaning component of the resulting verb is that of downplaying the seriousness and importance of the activity. The action is not really performed in the strict sense of the word, but it is done ‘just a bit’, ‘just for fun’, ‘not properly and not seriously’, ‘as if one is learning to do it’, ‘according to one’s ability (and therefore not really well)’, ‘somewhat randomly’. Occasionally the meaning can even shade into ‘pretend to’. Providing adequate translation examples out of context is not easy, and the examples in Table 16.4.6 give rough equivalents only. These derivations retain the transitivity of the original verbal base. Examples in clauses (with some comments) show the range of usage. (8)

O

art

tobho ka-­fegholei miina dagger nmls-­sharpen.with.magic neg

nae-­mbali 3sg.irr-­can

Table 16.4.6. Attenuation with poka-­in Muna Base form

Meaning

Reduplicated form

Meaning

buri fumaa kadiu

write eat bathe

no-­poka-­buri-­buri no-­poka-­fuma-­fumaa no-­poka-­kadi-­kadiu

linda toda

dance notch

no-­poka-­linda-­linda no-­poka-­toda-­toda

just write a little eat just a little, eat a little for fun take a bath, but not properly; just have fun playing with water dance a little, dance just for fun just make some notches (not properly)

Muna 373

do-­poka-­buna-­buna-­ane; amba-­do kamokula-­hi 3pl.r-­not.seriously-­rdp-­pull.out-­3sg.ind.obj word-­3pl.poss elder-­ pl do-­marangkuni-­ane. 3pl.r-­hepatitis-­3sg.ind.obj ‘A dagger sharpened with magic, you cannot just pull it out (from its sheath) for fun or for no particular reason; the elders say that (if you do that) you will get hepatitis.’

(9)



No-­bisara-­mo tomi ‘Insoba poka-­lengka-­lengka-­kanau 3sg.r-­speak-­prf finch try not.seriously-­rdp-­open-­1sg.ind.obj do-­po-­bisara-­ghoo deki.’ 1pl.incl.r-­recp-­speak-­purp first ‘The finch said (to the monkey), “Just open (your mouth) a little bit for me, so that we can talk first.”’ (The finch is trapped inside the monkey’s mouth and is hoping to escape. A normal request would be lengka-­kanau ‘open for me’. By using poka-­ he is downplaying the importance of his request.)

(10) Aini-­ha-­e-­mo, a-­fo-­suli-­e-­mo ini-­a ka-­buri-­nto-­a, this-­ loc-­3sg.obj-­prf 1sg.r-­caus-­return-­3sg.obj-­prf this-­ paus nmls-­write-­2sg.pol-­paus a-­poka-­bhori-­bhori-­hi-­e dua we lalo 1sg.r-­not.seriously-­rdp-­scribble-­pl-­3sg.obj also loc inside ampa ka-­pooli-­ku. until nmls-­able-­1sg.poss ‘Here it is, I hereby return your writings (to you), I have just scribbled a little bit inside, to the extent of my ability.’ (From a letter in which the author, a respected and knowledgeable community leader, returned draft pages of a Muna dictionary with his comments written in. The use of poka-­, as well as his choice of bhori ‘scribble’ instead of the more usual buri ‘write’, is a polite way of downplaying his contribution.)

2.4.2  -­hi Possibly the ‘leisurely’ meaning of -­hi on verbs can also be viewed as a case of attenuation (see also Section 2.3.3). In such cases the suffix signals that the action is performed in a relaxed and leisurely way, often somewhat aimlessly and without a specific purpose. Typical examples include de-­mpali-­mpali-­hi ‘they take a stroll, at their leisure, without going anywhere’, from mpali-­mpali ‘stroll’ (usually reduplicated) and ne-­ngko-­ngkora-­hi ‘she or he is sitting down, in a relaxed way, without any particular purpose’, from ngkora ‘sit’, with partial reduplication. 2.5  Intensification Intensification is typically signalled by full reduplication, but only with nouns, as illustrated in (11). (With verbs, full reduplication usually signals continuous action.) (11) a. ai-­ai ‘youngest sibling’ rdp-­younger.sibling (ai ‘younger sibling’) b. ko-­se-­ghulu-­ghulu-­ha-­e ‘all over his whole body’ whole-­one-­rdp-­body-­loc-­3sg.obj (ko-­se-­ghulu-­ha-­e ‘the whole body’)

374

René van den Berg

c. miina bhe lambu-­lambu ‘there is not a single house’ neg with rdp-­house (miina bhe lambu ‘there is no house’) d. no-­ko-­kiri-­kiri ‘it is full of thorns’ 3sg.r-­have-­rdp-­thorn (no-­ko-­kiri ‘it has thorns’) e. we wunta-­wunta ‘right in the middle’ loc rdp-­middle (we wunta ‘in the middle’) Partial reduplication (defined as monosyllabic reduplication with optional lengthening of the reduplicated vowel) on stative verbs may also signal intensification, as with no-­bha-­ bhala ‘very big’, no-­ghi-­ghito ‘very black’ and no-­pu-­pute ‘very white’.3 This pattern is rather limited in the standard variety of Muna, but in the southern dialect this is much more common. However, only the consonant is reduplicated in that case. The vowel is a lengthened copy of the verb-­class vowel o or e: no-­bhoo-­bhala ‘it is very big’, no-­doo-­dai ‘it is very bad’, no-­koo-­kuni ‘it is very yellow’, ne-­tee-­taa ‘it is very good’.

3 Conclusion Evaluative morphology figures prominently in Muna, though the semantic nuances are not always easy to pin down and in a few cases (such as -­hi), one can argue over whether a particular formation should be classified as evaluative morphology.

Acknowledgements I want to acknowledge some of the Muna friends who helped me to learn their beautiful language between 1985 and 1994: †Nilus Larangka, †Hanafi, †La Ode Abdul Fattah, †La Ada and Syahruddin. More recently, Mainuru Hado, La Habi, La Indasa and Lawa have been very helpful.

Notes 1. The exact number of Muna speakers is difficult to come by, as several thousand Muna speakers are found in the provincial capital of Kendari, but nobody knows their exact number. Also, since the population is shifting to Indonesian, there is an increasing number of ethnic Muna people with a limited or only a passive knowledge of the language. 2. Most symbols have their expected value, but is an implosive /ɓ/, a voiced dental plosive /ḓ/ and a voiced uvular fricative /ʁ/. 3. There are also two known cases of partial reduplication with an approximative meaning: rubu ‘small’ > no-­ru-­rubu ‘it is rather small’ and mpona ‘long (of time)’ > no-­mpo-­mpona ‘it is rather long’.

16.5 Tagalog

Carl Rubino

1 Introduction Tagalog is a native language of Southern Luzon, Philippines, which includes the capital city Manila and the fertile agricultural lands of south-­central Luzon. Tagalog is also spoken natively with dialectal variation on the islands of Lubang and Marinduque, with communities on Mindoro, Palawan and Mindanao as well. Since 1937, a standardised form of Tagalog has served as the national language Pilipino for the highly multilingual archipelago, replacing the regional lingua francas and colonial languages Spanish and English as the preferred means of intercultural communication for the majority of the Philippine population. The language retains its Philippine-­type Western Austronesian structure with minimal morphosyntactic interference from its colonial languages Spanish and English, most notably in the syntactic comparative construction and some subordination strategies. It is a predicate-­initial language, with a canonical word order reflecting agent before patient, and it has a highly prefixing morphology. Like its sister indigenous Philippine languages, Tagalog has complex verbal morphology which reflects aspect, mode and voice (orientation). Orientation encodes the semantic relationship between the verb and its nominative argument via prefixes, suffixes, infixes and combinations thereof. Tagalog words can be categorised into two main classes: open-­class roots that may take inflectional and/or derivational affixation, and a closed class of non-­inflecting forms such as particles, enclitics and case-­marking articles. There is also a pervasive particle (na after consonants, -­ng after n or vowels) used to link related constituents (heads and attributes) and called a ligature. All open-­class roots may predicate, given the appropriate morphology, and all verbs may be morphologically nominalised. Evaluative morphology (EM) in Tagalog is derivational and stems from two sources: borrowed Spanish diminutive suffixes and native morphological processes. I will detail the Spanish borrowings and native morphological processes that contribute to EM first, and then exhibit current trends in EM processes used in modern Tagalog slang, concluding with a sampling of formally and functionally distinct EM in sister Philippine languages.

2  Spanish borrowings The highly productive Spanish diminutive suffixes -­ito (masculine) and -­ita (feminine) appear in the Tagalog language, mostly with lexicalised nouns that were borrowed as whole units with their affixes, for example plato ‘plate’ > platito ‘little plate, saucer’; baso ‘glass’ > basito ‘small glass’; botelya ‘bottle’ > botelyita ‘little bottle’; bangko ‘bench’ >

376

Carl Rubino

bangkito ‘stool, small bench’; kutsara ‘spoon’ > kutsarita ‘teaspoon’; bola ‘ball’ > bolita ‘pellet, ball bearing’; papel ‘paper’ > papelita ‘slip of paper’ > papelito ‘small slip of paper; paper package for powdered drugs’; guapo ‘good looking’ > guapito ‘really cute; good-­looking young man’. These derivatives in -­ito/-­ita may be commonly applied to Hispanicised names, mostly to reflect endearment, for example Paco > Paquito, Carlos > Carlitos, Perla > Perlita. Tagalog is a language which does not reflect gender grammatically. The third person pronouns are gender neutral and the only agreement conventions that currently exist in the language stem from the large influx of gender-­carrying adjectives from Spanish, for example tsino ‘Chinese (Spanish borrowing of chino)’ > tsinito ‘little Chinese man’, tsinita ‘little Chinese woman’. Although most of the -­ito/-­ita diminutives entered the language as fully fledged lexical items, this suffix may be used with some native roots, carrying the diminutive sense it had in its original language, for example baklâ ‘gay’ > baklita ‘little gay boy’, bago ‘new’ > bagito ‘greenhorn, little newby’. Like the -­ito/-­ita suffix, the Spanish suffix -­illo/-­illa (Tagalog -­ilyo/-­ilya) survives in modern Tagalog. Batchelor and San José (2010, 452) categorise this diminutive suffix in Spanish as a possible carrier of a pejorative overtone, especially in Latin America where its use is ‘almost exclusively pejorative’. One could classify Tagalog words taking this suffix in a cline of morphological complexity – at one end of the spectrum, words that were borrowed wholesale where Tagalog speakers do not immediately recognise the diminutive morphology, and at the other end, words where the diminutive morphology serves a semantic function. Words that were borrowed wholesale and are most probably psychologically monomorphemic to Tagalog speakers include pasilyo ‘hall’, gantsilyo ‘crochet’, tripilya ‘tripe soup’, gatilyo ‘trigger’, mansanilya ‘chamomile’ and karilyo ‘shadow play’. Words that were probably borrowed wholesale but whose simplex forms still exist in Tagalog include baratilyo ‘bargain, sale’, gusanilyo ‘auger, small worm’, lomilyo ‘tenderloin’, bombilya ‘light bulb’, makinilya ‘typewriter’, kusinilya ‘gas stove’, palilyo ‘toothpick’ and karetilya ‘push cart’. And finally, there are words where the diminutive affix is immediately recognised by Tagalog speakers. Some of these exist in Spanish, such as platilyo ‘small plate’, bigotilyo ‘thin moustache’, abogadilyo ‘paralegal, nearly a lawyer’, kursilyo ‘short course of study’, sapatilya ‘heeled slipper’ and bintanilya ‘ticket window’; and some of these words take the Spanish diminutive suffix with native roots: bago ‘new’ > bagotilyo ‘little newby, novice’; binatà ‘bachelor’ > binatilyo ‘teenager’.

3  Native evaluative morphology Native EM strategies include prefixation, reduplication, and reduplication with suffixation. Full root reduplication with the suffix -­(h)an can be employed to create diminutive nouns or to express imitation (Schachter and Otanes 1972, 100). See Table 16.5.1. With stative verb roots where agentivity is involved to reach the resulting state (morphologically encoded with stress shift), disyllabic reduplication is employed to intensify. Disyllabic roots are repeated, but with trisyllabic roots, the first reduplicand is disyllabic and ending in an open syllable (Schachter and Otanes 1972, 234). See Table 16.5.2. With stative verb roots that express inherent qualities, this same type of reduplication is used, but for attenuation. Schachter and Otanes (1972, 236) describe this as ‘adjectival moderation’. This is used with both ma-­and unaffixed adjectives. See Table 16.5.3. When these same stative verbs are doubled with an intervening linker (na after

Tagalog 377 Table 16.5.1. Tagalog diminutive reduplication Base form

Meaning

Reduplicated form

Meaning

bulaklak bahay pari baril kabayo tao

flower house priest gun horse person

bulaklak-­bulaklakan bahay-­bahayan pari-­parian baril-­barilan kabayo-­kabayuhan tau-­tauhan

artificial flower toy house, play house fake priest toy gun toy horse, horse game, seahorse toy human figure, puppet, effigy

Table 16.5.2. Tagalog reduplication of statives Base form

Meaning

Reduplicated form

Meaning

bútas butás Punit punít hiwaláy Baliktád

hole with holes tear torn separate(d) reverse(d)

butas-­butás

thoroughly holey

punit-­punít

thoroughly torn

hiwá-­hiwaláy balí-­baliktád

thoroughly scattered topsy-­turvy

Table 16.5.3. Tagalog attenuative reduplication with statives Base form

Meaning

Reduplicated form

Meaning

malinis masarap matalino masalitâ bago buhaghag sanáy tahimik sariwâ

clean tasty intelligent talkative new porous experienced quiet fresh

malinis-­linis masarap-­sarap matali-­talino masali-­salitâ bagu-­bato buha-­buhaghág sanay-­sanáy tahi-­tahimik sari-­sariwâ

rather clean rather tasty rather intelligent rather talkative rather new rather porous rather experienced rather quiet rather fresh

­consonants or -­ng after vowels), the meaning is that of intensification, not attenuation. In example (1) this use is illustrated with two adjectives borrowed from the colonial languages, English and Spanish (source: twicsy.com/i/7QELpd). (1)

May

ex

shades o wala, guwapo-­ng guwapo at shades or neg.ex handsome-­ link handsome and

hot na hot si Daniel Padilla. hot link hot art Daniel Padilla ‘With or without shades, Daniel Padilla is really handsome and very hot.’ Disyllabic reduplication may also be used to attenuate actions (Schachter and Otanes 1972, 340–1). The examples in Table 16.5.4 portray this function with both agent-­oriented (mag-­ prefix) and patient-­oriented (-­in suffix) verbs. The prefix nápaka-­ is also used to intensify. It most commonly appears with statives

Carl Rubino

378 Table 16.5.4. Tagalog attenuative reduplication with dynamic verbs Base form

Meaning

Affixed form

Meaning

mag-­lakád mag-­hanáp mag-­isíp linis-­in

walk search think clean

maglakád-­lakád maghanáp-­hanáp mag-­isíp-­isíp linis-­linis-­in

do a little walking do a little searching do a little thinking clean a little

Table 16.5.5. The Tagalog prefix nápaka-­ Base form

Meaning

Prefixed form

Meaning

ma-­gandá ma-­bangó mahál maliít

beautiful fragrant expensive small

nápaka-­gandá nápaka-­bangó nápaka-­mahál nápaka-­liít

quite beautiful, very beautiful really fragrant really expensive really small

which take the prefix ma-­ (Table 16.5.5). A noteworthy feature of this prefix is that the single argument predicate produced cannot take a nominative argument but only a genitive one. This unique syntax closely mirrors that of the nominative morphology found in Tagalog exclamatives, using the prefixes pagka-­or ka-­. (2)

int-­happy

Nápaka-­ligaya ko noon. 1s.gen then.pst ‘I was really happy then.’ [Intensive with genitive actor]

(3)

Pagka-­bait-­bait ni Mrs Santos! nmls-­kind-­kind gen Mrs Santos ‘Mrs Santos is so kind!’ [Exclamative with genitive actor] (Schachter and Otanes 1972, 282)

(4) Iwas-­an mo lang yung sobra-­ng pagka-­bait mo. tr 2s.gen just dist over-­ link nmls-­kind 2s.gen Avoid-­ ‘Just avoid being excessively nice.’ [Nominative use with same syntax] Unlike in many other Philippine languages where affixation or reduplication is employed to express comparative degree, Tagalog has calqued a Spanish syntactic form, borrowing the Spanish word más ‘more’ placed before the adjective. However, superlative degree is formed morphologically with the prefix pinaka-­: ganda ‘beauty’> maganda ‘beautiful’ > mas maganda ‘more beautiful’ > pinakamaganda ‘most beautiful’. Tagalog young people are also known to derive new word forms using what is called ‘baliktad’ speech, or syllable reversal. Since most Tagalog roots are disyllabic, this does not pose a challenge for spontaneous creation and immediate recognition; for instance pangit ‘ugly’, ngitpa ‘pretty ugly’. Some of these reversals take additional morphology and have gained acceptance in their newly lexicalised forms, for example ma-­baho (ST-­stink) ‘smelly’ > hoba > hobabs ‘really stinky person’, guwapo ‘good looking’ > pogi ‘handsome, cute.’ In addition, some recently coined slang terms have employed non-­productive affixes (Zorc and San Miguel 1991). See Table 16.5.6.

Tagalog 379 Table 16.5.6. Tagalog slang affixation Suffix

Use

Examples

-­y

augmentative pejorative attenuative

abno ‘abnormal, mentally retarded’ bading ‘gay’ kastila ‘Spanish’

abnóy ‘retard’ badinglets ‘little gay boy’ kastilaloy ‘damn Spaniard’

puke ‘vagina’ binata’ ‘bachelor’

pukengkay ‘specific vagina’ Binats ‘highly eligible bachelor’

bigote ‘moustache’ puta ‘prostitute’

bigots ‘cool moustache’ putatsing ‘whore, immoral woman’

-­lets -­Coy -­Cay

-­s

-­tsing

pejorative

discourse variation

pejorative

4  Evaluative morphology in other Philippine languages EM can be found in many forms in Philippine languages quite distinct from what exists in Tagalog. I will provide a short sample of what can be found in the hundreds of languages across the archipelago. In the Cordilleran language of Balangao spoken in Northern Luzon, the complex prefix pagaCCV-­ is used with colours to attenuate them: ditak ‘red’ > pagaddiditak ‘reddish’; pokaw ‘white’ > pagappopokaw ‘whitish’; lanhi ‘green’ > pagallalanhi ‘greenish’. The infix -­in-­ combined with CVC reduplication forms diminutives and pejoratives: balon ‘packed lunch’ > binabalon ‘inadequate lunch’; balli ‘typhoon’ > binaballi ‘wee typhoon’; Paola ‘name’ > Penapaola ‘derogatory ways of Paola’ (Shetler 1976, 104–6). In the Mansaka language of Mindanao, the circumfix gi-­ -­an is used to augment or ameliorate, for example otaw ‘man’ > gyotawan ‘famous man’, dakora ‘large’ > gidakoraan ‘largest’ (Svelmoe and Svelmoe 1974, 17). Romblomanon, a Visayan language spoken in Romblon province north of Panay Island, employs two different strategies to intensify states. The prefix ka-­ with full reduplication of the root signals intense degree: basa’ ‘wet’ > ka-­basa’basa’ ‘very wet’; huga’ ‘difficult’ > ka-­huga’huga’ ‘very difficult’. Some roots can take the (ma)CV-­ prefix to encode veracity, intensifying the speaker’s conviction: punu’ ‘full’ > mapupunu’ ‘really full, truly full, clearly full’ (Newell 2006, 66–7). The Ilocano language of Northern Luzon uses the prefix nag-­ with the enclitic -­(e)n or the complex prefix nakaCVC-­to intensify states: na-­lukmég ‘ST-­fat’ > naglukmegen ‘really fat, surprisingly fat’ > nakaluklukmég ‘really fat’. To moderate the degree of statives, the prefix paN-­ is employed with the suffix -­en: lukmeg ‘fat (root)’ > panglukmegen ‘rather stout’; kusipet ‘almond-­eyed’ > pangusipeten ‘slightly squinty’; dakkel ‘big’ > panakkelen ‘rather large’. Initial closed syllable reduplication is employed to form comparatives and the ka-­ -­an circumfix serves to create superlatives: na-­pintas (ST-­beauty) ‘beautiful’ > napinpintas ‘more beautiful’ > kapintasan ‘most beautiful’ (Rubino 2000, lvi).

Carl Rubino

380 Table 16.5.7. The Tausūg prefix Caw-­ Base form

Meaning

Prefixed form

Meaning

mag-­dagan utang mag-­panayam mag-­taymanghud mag-­’itung

ag-­run debt ag-­play recp-­sibling ag-­count

magdawdaganan magkaaw’utang magpawpanayam magtawtaymanghud mag’aw’itung

run all over get into a lot of debt play around entire family keep counting

Table 16.5.8. The Ibaloi prefix Ce-­ Base form

Meaning

Prefixed form

Meaning

kabajo aso baley too damisaan baro

horse dog house person table clothes

kekebajo aaso bebedey tetoo dedamisaan bebecho

horse figurine, horse statuette toy dog play house doll toy table toy clothes

In the Tausūg language of Jolo Island, the prefix Caw-­, where C represents the initial consonant of the root, forms distributives where many arguments are affected or involved. However, with single arguments, the affix has a durative or intensifying function. It forms collectives with kin terms (Rubino 2006, 275–7). See Table 16.5.7. The Ibaloi language of Northern Luzon employs a reduplicative prefix Ce-­ to denote diminutive imitation or pretence, akin to the use of full reduplication in Tagalog. This prefix may occur with concomitant initial root vowel changes or medial consonant mutations (Ruffolo 2004, 104). See Table 16.5.8. In the Bikol language of South-­eastern Luzon, full reduplication may be used to express diminutives as well as plurals and intensives, for example mahal ‘expensive’ > mahal-­ mahal ‘somewhat expensive’ or ‘very expensive’. Another prefix, Curu-­, is used to express both augmentation of events (plural, iteration, continuative, distributive) and attenuation (diminution and imitation), but not intensive augmentation: balyo ‘change’ > burubalyo ‘keep on changing’; banggi ‘night’ > burubanggi ‘every night; karabasa ‘pumpkin’ > kurukarabasa ‘small pumpkin’; but barato ‘cheap’ > burubarato ‘somewhat cheap’, not ‘very cheap’ (Mattes 2006).

5 Conclusion Derivational EM in the Tagalog language derives from two distinct origins with two distinct formal characteristics: Spanish suffix borrowing and indigenous Austronesian reduplication and/or prefixation. These evaluative processes are used to attenuate, forming diminutives with related semantic extensions. Augmentative morphology used to refer to specific entities, although present in the colonial language Spanish, does not exist in Tagalog. However, there are strategies to intensify actions and states with native morphemes. The inventories of EM in Philippine languages vary from language to language in both form and function. Hence, much more can be learned in this realm from sister languages, many of which remain undocumented.

16.6 Tibetan

Camille Simon and Nathan W. Hill

1 Introduction Most researchers see Tibetan as a member of a language family which also includes Burmese and Chinese; this family is known by names including ‘Tibeto-­Burman’, ‘Sino-­ Tibetan’ and ‘Trans-­Himalayan’, of which the last is the most neutral and accurate (cf. van Driem 2012). In 650, Tibetan was reduced to writing as an administrative exigency of running the Old Tibetan empire; the earliest extant documents date from a century later (Hill 2010, 110–12). Tibetan linguistic history is conventionally divided between Old Tibetan (eleventh century and earlier) and Classical Tibetan (later texts). Tibetan boasts a vast literature with a wide variety of genres, and the family of Tibetic languages spoken today is comparable in size and diversity to the Romance languages (Tournadre 2008, 282–3). Lhasa Tibetan is the language spoken in the city of Lhasa, the cultural and historical centre of the Tibetan-­speaking area; this dialect is closely affiliated with the other Central Tibetic languages (Ü-­kä, Dbus-­skad).1 Different morphosyntactic processes are attested in Lhasa Tibetan to express evaluative semantics.2 In addition to diminutive and intensive formations, Lhasa Tibetan also has honorifics, which express an evaluation by the speaker.

2 Diminutives Diminutives are mostly expressed by suffixation. Only one reduplicative structure is used as a diminutive. 2.1  Suffixation 2.1.1  The suffix -­ḥu In a formation that is no longer productive in Lhasa Tibetan, the noun bu ‘son, child’ suffixed to a noun forms a diminutive (e.g. bum-­pa ‘water pot’, bum-­bu ‘small water pot’, cf. Uray 1952, 185). The diminutive can add an unpredicted element of meaning (e.g. rlig-­pa ‘testicles’, rlig-­bu ‘scrotum’, cf. Uray 1952, 185). In some cases the non-­diminutive form is unattested or the diminutive derives from a verb stem (e.g. √dril ‘turn’, dril-­bu ‘bell’, √tor ‘throw’, thor-­bu ‘fragments, miscellanea’3). After open syllables the -­b-­ weakens to -­ḥ-­ (pronounced [ɣ] in Old Tibetan), and the vowel -­a-­ ablauts to -­e-­, e.g. spra ‘ape’, spreḥu ‘monkey, rta ‘horse’, rteḥu ‘colt, mare’ (cf. Uray 1952, 186).4 Often the suffix assimilates to the preceding final consonant, as in

382

Camille Simon and Nathan W. Hill

lug-­gu ‘lamb’ (cf. lug ‘sheep’) (Uray 1952, 185–6). However, this orthographic practice does not reflect phonological or phonetic gemination in Lhasa Tibetan. The word mde-­ḥu ‘bullet’ (cf. mdaḥ ‘arrow’), as an early modern technology, serves as a terminus post quem for the productiveness of this formation. 2.1.2  The suffix -­tsam A reflex of the Classical Tibetan clitic -­tsam,5 pronounced [ts] in Lhasa, occurs in the word tog.tsam ‘a bit’ and in a diminutive construction, in which it is postponed to verbs, adjectives or numerals. Suffixed to adjectives, -­tsam weakens the strength of the adjective. (1) ང-­ར་ དུས་ཚོད་ མང་-­ཙམ་ དགོས་-­ཀྱི་འདུག ṅa-­r dus.tshod maṅ-­tsam dgos-­kyi.ḥdug lot-­dim need-­prs.test 1sg-­dat time ‘I need a bit more time.’ With numerals, this suffix tends to convey an idea of approximation, more than a diminutive; it is often understood as ‘almost num’, or ‘a bit more than num’. (2) ཁོང་ མ་-­ཐུག་-­ནས་ ལོ་ བཅུ་-­ཙམ་ khoṅ ma-­thug-­nas lo bcu-­tsam 3sg.hon.abs neg-­meet-­conn year ten-­dim ‘I have not met him for about ten years.’

ཕྱིན་-­སོང་། phyin-­soṅ/ go-­pst.test

When suffixed to verbs, this derivation triggers a modification of the part of speech: to be used as a verb, the derived stem must be reverbalised with the help of the light verb byed ‘to do’. (3) ང-­འི་ རོགས་པ-­འི་ ཡིག་ཚད་ ṅa-­ḥi rogs.pa-­ḥi yig.tshad 1sg-­gen friend-­gen exam.abs ‘I took a look at my friend’s exam.’

ལྟ་-­ཙམ་ lta-­tsam look-­dim

བྱས་པ་ཡིན། byas-­pa.yin lightv-­pst.ego.intent

Related to its diminutive function, -­tsam is also grammaticalised as an aspectual marker; suffixed to a verb it means ‘to have just V’. (4) ཁ་ལག་ བཟས་-­ཚར་-­ཙམ་-­ཡིན། kha.lag bzas-­tshar-­tsam-­yin meal.abs eat-­finish-­asp-­pst.ego.intent ‘I have just eaten.’ 2.1.3  The suffix -­ḥdra Two structures which contain the morpheme ḥdra ‘to be similar’ or the adjective ḥdra. po ‘similar’ convey a diminutive meaning that can be paraphrased as ‘kind of NP’. The first structure is derived by the following morphosyntactic rules: NP → NP-­ḥdra; the second structure has the two alternative forms N → N-­GEN-­bzo.ḥdra and N → N-­GEN-­bzo-­ḥdra-­po. The indefinite marker cig ‘a’ frequently follows the noun phrase that results from these processes. In natural speech, the syllable is phonologi-

Tibetan 383 cally realised /ra/ or /rä/ (instead of the regular /ɖa/), but most speakers still recognise its etymology. Morphosyntactically, NP-­ḥdra modifies the whole noun phrase to which it is postposed. It is best described as a noun phrase enclitic.6 (5) ཁོ་ མི་ སྐྱོ་པོ་-­འདྲ་-­ཅིག་ ཡོད་-­པ་འདྲ། kho mi skyo.po-­ḥdra-­cig yod-­pa.ḥdra 3.sg.abs person poor-­dim-­indf cop-­epist ‘It seems that he was kind of poor.’ (Hoshi et al. 1981–90) The second derivation, which associates the verb bzo ‘to make’ with the verb ḥdra or the adjective ḥdra-­po, has the same structure as a postposition: it is only postposed to nouns, to which it is linked by the genitive marker. (6) ཁོ་རང་ ཐུགས་སྤྲོ-­འི་ བཟོ་འདྲ་-­ཅིག་-­ལ་ཡཱ་ སྐད་གཏང་-­གར་བྱས། kho.raṅ thugs.spro-­ḥi bzo.ḥdra-­cig-­la.yā skad.btaṅ-­gar-­byas dim-­indf-­dat invite-­conn-­do 3sg.hon.abs party-­gen ‘[They] decided to invite him to a kind of party.’ (Hoshi et al. 1981–90) Both structures are fully productive. 2.2 Reduplication The only reduplication strategy that conveys a diminutive meaning applies to both adjectives (7) and verbs (8), but not to nouns. The derivation has the form adj/v-­la ma adj/v. Its meaning can be paraphrased by ‘not very adj’ or ‘to half-­V’/‘not to V whole-­heartedly’/‘to almost V’. (7) གསལ་-­ལ་ མ་-­གསལ། gsal-­la ma-­gsal clear-­conn neg-­clear ‘not very clear’ (8) པཱ་-­ལགས་ གཤེ་གཤེ་-­བཏང་-­བྱུང་། pā-­lags gśe.gśe-­btaṅ-­byuṅ/ father-­ hon scolding-­lightv-­pst.ego.rec



ཡིན་ནའང་ ཉན་-­ལ་ མ་-­ཉན་-­བྱས་པ་ཡིན། yin.na.ḥaṅ ñan-­la ma-­ñan-­byas-­pa.yin/ but listen-­conn neg-­listen-­lightv-­pst.ego.intent ‘Father scolded me, but I only half-­listened.’

3 Intensives Lhasa Tibetan has a number of intensive derivations formed through suffixation and reduplication.

Camille Simon and Nathan W. Hill

384 3.1  Suffixation

Two intensive formations are expressed through suffixes, namely adj-­drag(s) and n/adj-­ tsha.po. Suffixation with -­drag(s) is productive, but suffixation with tsha.po is, at least partially, lexicalised. 3.1.1  The suffix -­drag(s) The suffix -­drag(s) is used to derive excessive meaning of adjectives regularly. Etymologically, it results from the grammaticalisation of the adjective drag.po ‘ferocious, violent’. The two spellings -­drag and -­drags are both accepted. (9)

མོ་ཊ་ མང་-­དྲགས་-­བཞག Mo.ṭa maṅ-­drags-­bźag car lot-­too.much-­prf.evid ‘There are too many cars.’

The modification of the range of syntactic functions accessible to the lexical item is made clear, here, by the absence of copula, which would otherwise be necessary for the expression of predicative adjectives. Here, the adjective suffixed by -­drags is directly followed by the verbal aspecto-­modal suffix -­bźag. 3.1.2  The suffix -­tsha.po Synchronically, tsha.po is an independent adjective ‘hot’. It may also combine with a nominal stem (N tsha.po) to form a derived adjective. In such compounds, tsha.po regularly alternates with chen.po ‘big’, which conveys a more literary register. In this ­combination tsha.po (or chen.po) maintains its aspiration and tone; thus, phonologically in this context too tsha.po is treated simply as an adjective following a noun. However, syntactically and semantically the result is a single adjective that bears no relationship to ‘hot’, and for this reason it is tempting to analyse -­tsha.po as a derivational suffix. Most of these adjectives have a pejorative meaning. (10) རྙོག་གྲ → རྙོག་གྲ་ ཚ་པོ། rñog.gra → rñog.gra tsha-­po problem → problem hot-­pos

→ ‘problematic’

For a few adjectives this suffix derives an intensive equivalent:7 (11) ལྗིད་-­པོ་ ljid-­po heavy-­pos

→ ལྗིད་-­པོ་ ཚ་པོ། → ljid-­po tsha-­po → heavy-­ pos hot-­pos



‘(very) heavy’

Because the structure is partly lexicalised, the intensive meaning is sometimes weak. 3.2 Reduplication Since reduplication is iconically linked to intensive meaning, it comes as no surprise that this meaning is mainly expressed through reduplication. Reduplication seems fully productive, although it might be restricted to a particular semantically motivated subgroup of lexical items.

Tibetan 385 3.2.1  Syntactic reduplication Simple reduplication of adjectives, verbs and numerals conveys an intensive meaning. (12)

ག་ལེ་ ག་ལེ-­ར་ གཉིད་ཁུག་-­གི་འདུག ga.le ga.le-­r gñid.khug-­gi.ḥdug slow slow-­ adv fall.asleep-­imp.test ‘[I] fall asleep very slowly.’

This meaning is not always evident, as in the following example: (13) ཁོ་ གཉིས་-­གཉིས་ སྐད་ཆ བཤད་-­ཀྱི་འདུག kho gñis-­gñis skad.cha bśad-­kyi-­ḥdug two-­two.abs speech.abs speak-­imp.test 3sg ‘The two of them are talking.’ The reduplication of a verb or a part of the verb phrase conveys a meaning that can be interpreted as an evaluative or an aspectual value (intensification or quantification of the process; see also Vittrant and Robin 2007, 5–6).



(14)

ཁ་ ཕྱེ་-­གྲབས་ ཕྱེ་-­གྲབས་ kha phye-­grabs phye-­grabs mouth open-­about open-­about ‘[He] is just about to open [it].’

བྱེད་-­ཀྱི་འདུག bye-­kyi.ḥdug lightv-­imp.test

The reduplication of an interrogative in a question invites the interlocutor to answer with an enumeration: (15) ནང་-­ལ་ སུ་-­སུ་ ཡོད། naṅ-­la su-­su yod/ inside-­ dat who-­who exist.ego ‘Who lives at your home?’ (Lit.: ‘Who [and] who lives at [your] home?’) Or, if the interrogative is used as an indefinite, with the nominalisation of the verb: (16)

གན་རྒྱ་ དང་ ཡི་གེ་ ག་རེ་ ག་རེ་ བཞག་-­དགོས་-­ཡོད་-­ནའང་ གཞག་-­གི་ཡིན། gan.rgya daṅ yi.ge ga.re ga.re bźag-­dgos-­yod-­naḥaṅ bźag-­gi.yin contract and letter what what put-­must-­cop-­but put-­fut.ego

‘I will write whatever contract and letter are necessary/I will write as much contracts and letters as necessary.’ (Lit.: ‘What and what contract and letter I should make, I will make it.’)8 (Hoshi et al. 1981–90) 3.2.2  Reduplication of adjectives and verbs after an interrogative This construction applies equally to adjectives and verbs. Its meaning is: ‘As ADJ/V as possible’.9 (17) a. གང་-­མང་-­མང་ gan.-­man.-­man.

Camille Simon and Nathan W. Hill

386 intg-­lot-­lot ‘as much as possible’ b. གང་-­ཡག་-­ཡག gan.-­yag-­yag intg-­good-­good ‘as good as possible’ c. གང་-­ཐུབ་-­ཐུབ་ gan.-­thub-­thub intg-­can-­can ‘as much as possible’

This form, if not totally lexicalised, is restricted to a small number of stems in Lhasa Tibetan. 3.2.3  Reduplication before rkyaṅ This construction, meaning ‘completely ADJ’, is particularly frequently used with colour adjectives, but is also attested with other adjectives, such as: ཅིག་-­པ་ གཅིག་-­རྐྱང gcig-­pa gcig-­rkyaṅ (same-­pos same-­single) ‘exactly the same’, གསར་-­པ་ གསར་-­རྐྱང་ gsar-­pa gsar-­rkyaṅ (new-­pos new-­single) ‘very new’, དཀར་-­པོ་ དཀར་-­རྐྱང་ dkar-­po dkar-­rkyaṅ (white-­ pos white-­single) ‘completely white’. (18) མོ་ སྨྱོན་-­མ་ སྨྱོན་-­རྐྱང་ mo smyon-­ma smyon-­rkyaṅ 3sg.f crazy-­pos crazy-­single ‘She is completely crazy!’

རེད་བཞག red-­bźag cop-­evid

3.2.4  Other complex reduplication structures Other complex reduplication structures conveying an intensive meaning are attested, some of them being fully lexicalised, as in (20). Other forms are more productive, such as the structure neg-­V dgu-­V. Described in Vittrant and Robin (2007, 14), it conveys an intensive and pejorative meaning and can be paraphrased as: ‘to V too much and without care’. (19) མི་ འདི-­ས་ མི་-­དྲན་ དགུ་-­དྲན་ བྱེད་-­ཀྱི་ཡོད་རེད། mi ḥdi-­s mi-­dran dgu-­dran byed-­kyi.yod.red man dem-­erg neg -­remember nine-­remember lightv-­imp.fact ‘His thoughts are all over the place./He thinks in an excessive and desultory fashion.’ (Vittrant and Robin 2007, 14) In (20a), the adjectival stem draṅ is repeated, once with a positive suffix and then with the suffixed verb bźag ‘to put’. In (20b), the verb byuṅ ‘to appear’ is repeated, once preceded by the interrogative particle and then by the adjective ‘a lot’. (20) a. དྲང་-­པོ་ draṅ-­po frank-­pos ‘really frank’

དྲང་-­བཞག draṅ-­bźag frank-­put

b. གང་-­བྱུང་ gaṅ-­byuṅ

མང་-­བྱུང་ maṅ-­byuṅ

Tibetan 387 intg-­appear ‘any old how’

lot-­appear

4  Honorifics Lhasa Tibetan has a pervasive honorific system, in which an honorific equivalent is available for much of the vocabulary. A small proportion of the honorific lexicon is purely lexical and thus arbitrary, but others words are morphologically analysable, although it is difficult to evaluate their synchronic productivity. Examples cited in this section are taken from Rdo.dgon Gsaṅ.bdag Rdo.rje and Ross (2002) and Mélac, Robin and Simon (2014). A precise inventory of the attested morphological processes of honorific lexicon derivation can be found in Kitamura (1975, 68–74). As shown by DeLancey (1998), morphologically constructed honorific nouns and adjectives consist of compounds, in which the first constituent has a categorisation property over the second. In such compounds, the first constituent is an unanalysable honorific root of a noun or verb, related to the semantic domain of the non-­honorific root. The second constituent is the non-­honorific root and carries the semantic information. A possible nominal suffix is deleted in the honorific form: thab ‘stove’ → gsol-­thab (eat.hon-­stove) ‘stove. hon’, las ka (work-­ nmls.sfx) ‘work’ → phyag-­las (hand.hon -­ work) ‘work.hon’, par ‘picture’ → sku-­par (body.hon -­picture) ‘picture.hon’, lam-­ka (road-­nmls.sfx) ‘road’ → phebs-­lam (go.hon -­road) ‘road.hon’. If the non-­honorific word is already a compound, its first component is replaced by an honorific categoriser morpheme. Often the first component is replaced by its honorific equivalent: sems-­khral (mind-­tax) ‘worry’ → thugs-­khral (mind.hon -­tax) ‘worry.hon’. The process is also generalised to other types of bisyllabic nouns, containing no ­categoriser  in the non-­honorific form: chu-­tshod (water-­measure) ‘clock’ à phyag-­tshod (hand.hon -­measure) ‘clock.hon’, yi.tsi ‘soap’ (< Ch. 胰子yízi) → phyag-­tsi (hand.hon -­?) ‘soap.hon’. The fact that this morphological rule operates on the loanword ‘soap’ indicates that this formation was still productive at the time of the loan. An alternative strategy is available for marking honorific verbs. Most Lhasa Tibetan verbs are multisyllabic, composed of a predicative noun and a light verb. Honorific verbs are constructed with the honorific counterpart of the predicative noun and the honorific equivalent of the light verb. This formation is fully productive in Lhasa dialect. (21) ལན་ བརྒྱབ་ lan rgyab Answer lightv ‘to answer’

→ ལྗགས་-­ལན་ སྐྱོན་ → ljags-­lan skyon answer.hon lightv.hon → ‘to answer.hon’

For verbs (full or light) that lack a lexical honorific equivalent, the honorific light verb gnaṅ is appended to the verb. (22) a.  འདམ་ → འདམ་-­གནང་ ḥdam → ḥdam-­gnaṅ /dam/ /dam-­nāng/ ‘to.choose’ → ‘to.choose.hon’ b. སྐད་བཏང་ → སྐད་བཏང་གནང skad-­btaṅ → skad-­btaṅ-­gnaṅ

388

Camille Simon and Nathan W. Hill

/kǟ’-­tāng/ /kǟ’-­tāng-­nāng/ voice-­ lightv voice-­ lightv-­hon ‘to.invite’ → ‘to.invite.hon’ In this case, since both the verb and the light verb carry a tone, gnaṅ is not a suffix, but rather this is a serial verb construction.

5 Conclusion Three categories of evaluative meaning are expressed by different morphological means in Lhasa Tibetan: diminutive, intensive and honorific. Reduplication mainly serves to express intensive meanings, although one type of reduplication conveys diminutive or pejorative meaning. Prefixation is limited to the construction of honorific nouns, whereas suffixation is attested for all three types of meaning. Certain constructions not only express evaluative values, but also modify the range of the syntactic functions of the stem; this is especially frequent when the stem is a verb.

Notes 1. Dbus-­skad is a transliteration of the Tibetan word for the dialect we are discussing here. Ü-­kä is an attempt to reproduce something like the pronunciation of this word in the dialect itself. In IPA it would be [ykɛʔ]. Among the many treatments of Lhasa Tibetan, Kitamura (1977) provides a good short grammar; Hoshi (1988) is a good longer treatment. There is no accurate short treatment of Classical Tibetan; Schwieger (2006) is the best lengthy treatment. 2. In order to restrict the current discussion to morphological processes, a formation is described only if it meets at least two of the following criteria: (i) the item resulting from the application of the process is one phonological word (containing only one aspiration and one tone); (ii) its meaning is not strictly compositional; (iii) the process alters the part of speech or range of syntactic functions of the input form; (iv) at least one element of the derived form is phonologically different from the expected outcome (e.g. /ra/ rather than /ɖa/); and (v) at least one element of the derived form cannot be used synchronically as an independent linguistic unit. 3. The symbol √ precedes an uninflected verbal root. 4. In the word byiḥu ‘little bird’ (cf. bya ‘bird’) the ablaut is unpredictably to -­i-­ rather than -­e-­. 5. The behaviour of -­tsam in Classical Tibetan is not well enough understood to allow us to suggest whether or not it should be considered evaluative morphology. 6. Note that the same verb ḥdra ‘to be similar’, shown in (5), has also been grammaticalised as a component of an epistemic marker (Vokurková 2008). 7. Although the standard orthography gives ljid-­po in (11), the positive adjectival suffix in this and other words is pronounced /ko’/. What dictates the selection of /po/ versus /ko’/ in Lhasa Tibetan is not well understood; perhaps it can be linked to spellings such as sa-­ḥon for sa-­bon ‘seed’ and śo-­ge for śo-­be ‘lie, falsehood’, which also give evidence of interchange among -­b-­, -­ḥ-­and -­g-­ as early as Old Tibetan (cf. Hill 2011). 8. For other examples of simple reduplication of verbs with an aspectual (iterative, continuative or exhaustive) meaning, see Vittrant and Robin (2007, 6). For more on Tibetan reduplication in general see Uray (1954) and Vollmann (2009). 9. Etymologically, gaṅ is an interrogative pronoun meaning ‘what’. In its interrogative functions, it is replaced by ga in modern Lhasa Tibetan.

16.7 Yami

D. Victoria Rau and Hui-­Huan Ann Chang

1 Introduction This chapter describes evaluative morphology (EM) in Yami, a Philippine Bashiic language in the Austronesian family spoken by 3,000 speakers on Orchid Island. Previous studies in Austronesian languages have provided rich data on gestalt symbolism (Blust 1988, 59ff; 2009, 355) and lexical iconicity (Rau 1994). Studies on Yami have also laid the foundation for a description of the grammar (Rau and Dong 2006) with detailed analysis of reduplication (Rau and Dong 2005) and iconicity in the tense–aspect–modality system (Rau 2005). Since the mid-­1990s, this language has been well documented, including in a Yami–Chinese–English dictionary (Rau, Dong and Chang with Rau and Rau 2012) and on three websites established between 2005 and 2009 on Yami digital archives, Yami e-­learning and a Yami online dictionary. Based on the rich corpus of the Yami language and the assistance of Yami language consultant Maa-­neu Dong, this chapter aims to give a general picture of Yami EM. The chapter begins with a description of reduplication, followed by augmentatives and diminutives, and appreciatives and depreciatives. A complete table of Yami EM is presented in the conclusion.

2  Reduplication in evaluative morphology Reduplication is the most common and productive morphological process in Austronesian languages (Himmelmann 2005). In Yami, this phenomenon occurs primarily by partial reduplication (CV-­ rdp) and full reduplication (CVCV-­ rdp) without coda to indicate ­plurality, repetition/continuation, intensification, comparison, diminution, attenuation and ­imitation/fakeness (Rau and Dong 2005). As (1) shows, Yami reduplication can be used primarily for the formation of augmentatives (1a–k) and diminutives (1l–o). A metaphorical extension from size to value can be further developed. (1) Yami reduplication of augmentatives and diminutives: a. aro-­aro rdp-­many ‘more’ b. api-­apia rdp-­good ‘better’

D. Victoria Rau and Hui-­Huan Ann Chang

390 c. ma-­te-­téneng sv-­rdp-­smart ‘all smart’ d. ma-­te-­teneng sv-­rdp-­smart ‘better at, know more about’ e. ma-­te-­te-­teneng sv-­rdp-­rdp-­smart ‘understand a lot’ f. rak-­rako rdp-­big ‘older’ g. ara-­rako rdp-­big ‘all big, biggish’ h. mi-­lavi-­lavi af-rdp-­cry ‘keep crying, fussing’ i. mi-­ci-­cipa af-rdp-­spit ‘keep spitting phlegm’ j. ma-­ra-­rahet sv-rdp-­bad ‘all bad’ k. ma-­rahe-­rahet sv-­rdp-­bad ‘worse’ l. ma-­veve-­veveh sv-­rdp-­short ‘shorter’ m. ali-­likey rdp-­small ‘all small, smallish’ n. alik-­likey rdp-­small ‘smaller’ o. ava-­avang ‘rdp-­boat’ ‘toy boat’

Semantically, the derived forms are downgraded or upgraded in their dimensional structure. Take alikey ‘small’ in (2), for example. Through partial reduplication, alilikey ‘all small, smallish’ indicates either plurality or diminutive, as in (3). (2)

ya likey rana o ayob ko. AUX small already NOM clothes 1SG.GEN ‘My clothes have shrunk.’ (Lit.: ‘My clothes are small.’)

Yami 391 (3)

ali-­likey o among. RDP-­small NOM fish ‘All the fish are very small.’

A shift in domain among the data can be caused by a metaphorical transfer. The transition from small/big to partitive (1c, 1j, 1m) can be seen as a transfer from SIZE to AMOUNT, and the transition from small to continuity (1h and 1i) can be seen as a transfer from SIZE to DURATION (Jurafsky 1996). The reduplicated form of avavang ‘toy boat’ in (1o), analysed as an imitation of the natural object avang ‘warship, large boat, boat’, marks a transition from SIZE to DEPRECIATION.

3  Augmentative and diminutive affixation There are no typical augmentatives and diminutives such as ‘small X’ or ‘big X’ in Yami. Besides reduplication, diminutives and augmentatives can be expressed by affixation. The following examples in Sections 3.1–3.3 with affixation encode a primary meaning of augmentatives, while the examples of approximatives in Sections 3.4–3.5 only develop a secondary meaning of quantitative evaluation via a metaphorical extension from SIZE to intensification or attenuation. 3.1  Augmentative intensifier marker tey-­ ‘very’ The intensifier marker tey-­can apply to gradable stative verbs with semantic x to generate a meaning of very x, as in (4). (4) Augmentative intensifier formation of tey-­: a. tey-­rako aug-­big ‘very big’ b. tey-­likey aug-­small ‘very small’ c. tey-­apia aug-­good ‘very good’ d. tey-­ma-­rahet aug-­sv-­bad ‘very bad’ e. tey-­ma-­oyat aug-­sv-­strong ‘very diligent’ f. tey-­ma-­lma aug-­sv-­lazy ‘very lazy’ Compared with malma ‘lazy’ in (5), the augmentative teymalma ‘very lazy’ in (6) represents an intensification of the semantic force.

D. Victoria Rau and Hui-­Huan Ann Chang

392 (5)

ya

ma-­lma o mehakay ito. SV-­lazy nom male that ‘That man is lazy.’

(6)

ka tey-­ma-­lma 2sg.nom aug-­sv-­lazy ‘You are really lazy.’

aux

ya. this

3.2  Superlative augmentative ni-­. . . na ‘most’ Yami superlatives are formed by affixing ni-­ to the stem, followed by the possessive/ genitive marker na, as in niapia na ‘the best’. The intensification process presents a gradable quality, as shown in (7). (7) Superlative augmentative formation of ni-­. . . na: a. ni-­anaro na sup-­long 3sg.gen ‘the longest’ b. ni-­ma-­veveh na sup-­sv-­short 3sg.gen ‘the shortest’ c. ni-­alikey na sup-­small 3sg.gen ‘the smallest’ d. ni-­apia na sup-­good 3sg.gen ‘the best’ e. ni-­ma-­teneng na sup-­sv-­smart 3sg.gen While tey-­‘very’ in (8) intensifies semantic force, ni-­. . . na ‘most’ in (9) presents an even more centralised semantic meaning of the good taste of the fish soup. (8) tey-­apia a yop-­en o asoy no ivay ya. aug-­good link drink-­pf nom soup gen FISH.NAME this ‘This ivay fish soup is very good to drink.’ (9) ni-­apia na a yop-­en sup-­good 3sg.gen link drink-­pf ‘This ivay fish soup is the best to drink.’

o

asoy no

ivay

ya.

nom soup gen FISH.NAME this

3.3  Distributive marker mika-­‘gradually, increasingly’ The distributive marker mika-­ can refer to ‘gradually, increasingly’ to mark an action of temporal contiguity, as exemplified in (10). (10) Distributivity formation of mika-­: a. mika-­kala-­kalat

Yami 393 distr-­rdp-­climb ‘climb in droves’ b. mika-­kda-­kdas distr-­rdp-­cut.on.the.skin ‘scratch the skin open’ c. mika-­‘i-­’iteng distr-­rdp-­drip ‘sweat dripping one after another’ d. mika-­cio-­ciok distr-­rdp-­new.sprout ‘sprout one after another’ In (11), the motion event (Rau, Wang and Chang 2012) of mikalakalat ‘climb upward together’ can be interpreted as temporal simultaneity. However, in (12), mikakalakalat ‘climb upward one after another’ expresses the sequential event/action with temporal contiguity. (11) mi ta mi-­kala-­kalat go 1pl.nom.incl af-­rdp-­climb ‘Let us go climb trees in the mountains.’

do

loc

toko-­tokon. rdp-­mountain

(12) to sia rana mika-­kala-­kalat o tao aux 3pl.nom already distr-­rdp-­climb nom people a tanang do ma-­ka-­karang a tokon a. link go.up loc sv-­rdp-­high link mountain par ‘The people climbed to the higher mountain in droves.’ 3.4  Approximative -­en ‘become or do such and such easily’ The following examples are by no means prototypical augmentatives or diminutives; however, their approximative affixations have developed a secondary meaning of quantitative evaluation via metaphorical extension. When the approximative -­en is attached to a stem, the semantics denoted by the stem becomes augmented. The augmented force signifying habitualness, however, can be either a positive or a negative evaluation. The examples in (13) are intensified.1 (13) Approximative -­en ‘become or do such and such easily’: a. ma‘e-­’en-­en sv-­rdp-­cold-­approx ‘tend to feel cold easily’ b. ma-­it-­itkeh-­en af-­rdp-­sleep-­approx ‘have a habit of falling asleep’ c. ma-­na-­nakaw-­en maN-­ta-­takaw-­en af-­rdp-­steal-­approx ‘have a habit of stealing’ d. ma-­ngaa-­ngaa-­ap-­en maN-­aa-­aa-­ap-­en

D. Victoria Rau and Hui-­Huan Ann Chang

394 af-­rdp-­rdp-­take-­approx ‘love to go fishing’ e. ma-­m‘i-­m’ing-­en af-­rdp-­laugh-­approx ‘love to laugh’ f. ma-­sara-­saray-­in sv-­rdp-­happy-­approx ‘tend to be optimistic’ g. maka-­boa-­boang-­en af-­rdp-­familiar-­approx ‘adapt to a group easily’

In (14), manakaw is an action of stealing, whereas mananakawen in (15) becomes a habit of stealing. (14) ya ma-­nakaw 3sg.nom af-­steal ‘He stole money.’ (15)

so

nizpi.

obl money

ma-­na-­nakaw-­en so pza-­pzat-­an af-­rdp-­steal-­APPROX obl rdp-­protect-­nmls ‘(He) often steals other people’s things.’

no

tao.

gen person

Similarly, maboang in (16) is a state of being close to someone, while makaboaboangen in (17) is a habit of getting along well with others. ya

jiaken o anak ni wari. 1sg.loc nom child gen younger.brother ‘My younger brother’s children are very close to me.’

(16)

aux

(17)

aux

ya

ma-­boang

sv-­familiar

maka-­boa-­boang-­en

af-­rdp-­familiar-­approx

si

nom

‘Masaray gets along easily with others.’

Masaray. pn

3.5  Approximative marker mala-­‘smell or look a little like’ The approximative marker mala-­ can refer to either ‘smell a little like’, as in (18a–d), or ‘look a little like’, as in (18e–g), to mark an approximation. As mala-­is added to the stem, a metaphorical extension from SIZE to EVALUATION is developed. (18) Approximation formation of mala-­: a. mala-­inaw approx-­fish.smell ‘smell a little like fish’ b. mala-­pe-­pseng approx-­rdp-­burned ‘smell a little like something is burned’ c. mala-­vo-­vongtot

Yami 395 approx-­rdp-­rancid ‘smell a little like rotten food’ d. mala-­ta-­taci approx-­rdp-­urine ‘smell a little like urine’ e. mala-­círing approx-­word ‘look talkative’ f. mala-­váyo approx-­new ‘look young’ g. mala-­ngóngoy approx-­mouth ‘look mouth protruding’ Some expressions of hedges, such as sort of in English, are used to weaken a speaker’s assurance of a proposition (Lakoff 1972). In Yami, when the diminutive morpheme mala-­is attached to a stem, the semantics denoted by the stem becomes attenuated. Compared with (19), the prefix mala-­ in (20) adjusts the category membership property ‘burned’ by making central members of a category less central, i.e. ‘smell a little like burned smell’. (19)

Ya ma-­pseng o kanen. aux sv-­burned nom rice ‘The rice is burned.’

(20)

mo ma-­‘angno o ya mala-­pe~pseng ang? 2sg.gen pf.able-­smell nom aux approx-­rdp~burned q ‘Do you smell that burned smell?’

4  Appreciatives and depreciatives Among the descriptive/quantitative EM discussed above, apart from mala-­ ‘taste, smell or look a little like’, which marks attenuation, the rest all encode intensification. Although SIZE may be the primary meaning in EM, its metaphorical extension can be either positive or negative. This section presents the attitudinal/qualitative dimension of Yami EM. Qualitative EM is associated with the Yami attitude marker ka-­(Rau and Dong 2010), which only encodes depreciatives. Once again, the following examples are not primarily evaluative but have developed a secondary meaning as depreciatives in the process. 4.1  Depreciative marker mapaka-­‘pretend to be’ The prefix mapaka-­‘pretend to be’ denotes the feature of ‘marginality’ or ‘imitation’, as in (21). The negative evaluation is not about the action itself, but about the person who fakes the action.

396

D. Victoria Rau and Hui-­Huan Ann Chang

(21) Depreciative formation of mapaka-­: a. mapaka-­‘it-­’itkeh pretend-­ rdp-­sleep ‘pretend to sleep’ b. mapaka-­deh-­dehdeh pretend-­rdp-­foreigner ‘disguise as foreigners’ c. mapakey-­ngey-­ngen pretend-­rdp-­sick ‘pretend to be sick’ d. mapaka-­tok-­toklay pretend-­rdp-­limp ‘pretend to limp’ e. mapaka-­tele-­teleh pretend-­rdp-­deaf ‘pretend to be deaf’ While keyngeyngen ‘hurt’ in (22) is a statement of the physical condition, mapakeyngeyngen ‘faked sickness’ in (23) implies a negative evaluation of the person who fakes it. (22) pan-­pandan na do key-­ngey-­ngen na rdp-­limit 3sg.gen loc int-­rdp-­hurt 3sg.gen no ngepen na ni akay a. gen tooth 3sg.gen gen grandfather par ‘My grandfather’s teeth really hurt.’ (23) ma-­sekeh a ma-­ngay do gako am, SV-­unwilling link af-­go loc school par to mapakey-­ngey-­ngen a. aux pretend-­rdp-­hurt par ‘He didn’t want to go to school, so he faked illness.’ 4.2  Depreciative delimitation ka-­ ‘only’ The delimitive marker ka-­‘only’ can be attached to a stem to convey limited and negative action, as in (24).2 (24) Depreciative formation of limited action ka-­: a. ka-­’ak-­akaw dlm-­rdp-­plough ‘only farm’ b. ka-­nake-­nakem dlm-­rdp-­think ‘only want to’ c. ka-­pa-­nangdang ka-­paN-­dangdang dlm-­caus-­warm ‘only get warm by the fire’

Yami 397 d. ka-­raing dlm-­kill.fish ‘only kill fish’ e. ka-­sa dlm-­one ‘only one, small amount and unimportant’ f. ka-­itkeh dlm-­sleep ‘only sleep’ Compared with panangdang ‘get warm’ in (26), kapanangdang ‘only get warm near the fire’ in (27) is viewed negatively as implying a person who does nothing except sitting there getting heat. (25) pa-­nangdang jito. caus-­get.warm.near.the.fire there ‘Come warm yourself by the fire.’ (26) ka-­pa-­nangdang mo koimo a, dlm-­caus-­get.warm.near.the.fire 2sg.gen someone par ji angay ma-­ngayo a? neg go af-­pick.up.firewood par ‘Do you only sit there getting warm and not go pick up firewood?’ 4.3  Depreciative intensification ka-­. . . na ‘very’ Intensification ka-­‘very’ is prefixed to a reduplicated verb stem, followed by a third person singular genitive pronoun na to indicate a negative evaluation of a great amount (for quantity) or a great degree (for attribute) as in (28). (27) Depreciative intensification ka-­. . . na: a. ka-­singa-­singat na int-­rdp-­expensive 3sg.gen ‘so expensive’ b. ka-­owy-­oyat na ka-­oya-­oyat na int-­rdp-­strong 3sg.gen ‘so strong’ c. ka-­iley-­lamnay na int-­rdp-­take.it.easy 3sg.gen ‘so laid back’ d. key-­ngey-­ngen na int-­rdp-­hurt 3sg.gen ‘so painful’ e. ka-­key-­key-­kai na int-­rdp-­rdp-­fast 3sg.gen ‘so quickly’

D. Victoria Rau and Hui-­Huan Ann Chang

398

Pragmatically, the intensification ka-­RDP. . . na expresses a speaker’s attitude towards certain states or events. Compared to the augmentative intensifier marker tey-­‘very’, discussed above in Section 3.1, depreciative intensification ka-­RDP. . . na is used to express a complaining attitude. In contrast to the description of vegetables being expensive in (28), the speaker’s negative evaluation of the prices in (29) implies ‘if it were not so expensive, I would buy it’. (28)

ya

ma-­singat

aux sv-­expensive

o

napa.

nom vegetable

‘Vegetables are expensive.’

(29) ka-­singa-­singat na? tosia, ji int-­rdp-­expensive 3sg.gen don’t.want.it neg ‘That expensive? Never mind then, I’m not buying it.’

ko nazang-­i. 1sg.gen buy-­lf

4.4  Depreciative blaming attitude marker ka-­‘it serves you right’ The blaming marker ka-­ can be used to find fault with someone who did something that she or he should not have done, as in (30–2). It is not the action itself that is being evaluated, but rather someone who is involved in the action that is being criticised. A detailed discussion of the special morphosyntax of this evaluative marker can be found in Rau and Dong (2010). (30)

ka-­bakbak mo jia? blm-­hit 2sg.gen this.loc ‘Who told you to hit him?’

(31)

ka-­panci mo nia? blm-­call 2sg.gen this.gen ‘Who told you to call him?’

(32) ka-­panmama mo, man-­ngo ori, blm-­chew 2sg.gen af-­what that ya mika-­kte-­kteb so ngepen ang. aux distr-­rdp-­break obl tooth par ‘Look at you. Why eat betel nut? Your teeth have all fallen out one after another from eating betel nut.’

5 Conclusion EM in Yami is marked by both reduplication and affixation, as summarised in Table 16.7.1. There is no apparent prototypical EM to mark terms such as ‘big table’ or ‘small chair’ in many Indo-­European languages. However, the discussion of Yami examples in this chapter has increased our understanding of the process of the metaphorical extension from SIZE to EVALUATION and further development of evaluation as a secondary meaning in the language. The only quantitative EM in Yami consists in augmentative tey-­‘very’, superlative ni-­. . . na ‘the Xest’ and distributive mika-­‘increasingly, one after another.’

Yami 399 Table 16.7.1. Yami evaluative morphology Approximation (secondary)

Quantitative perspective BIG, augmentatives (primary)

Qualitative perspective BAD, pejoratives (secondary)

BIG, augmentatives:  ma/maN-­maka-­RDP-­ BASE-­en ‘become or do such and such easily’ SMALL, diminutives:  mala-­(RDP)-­BASE ‘taste, smell or look a little like’

tey-­(ma-­)-­BASE ‘very’ ni-­(ma-­)-­BASE na ‘the most’ mika-­RDP-­BASE ‘gradually, increasingly’

mapaka-­RDP-­BASE ‘pretend to’ ka-­(RDP)-­BASE ‘only’ ka-­RDP-­BASE na ‘very’ ka-­BASE ‘it serves you right’

The two approximatives -­en ‘do such and such easily’ and mala-­ ‘smell, or look like’ have developed a secondary meaning as augmentatives and diminutives, respectively. The processes of intensification and attenuation adjust the category membership property of the state/action/event to a less central position. Even though quantitative perspectives are independent from qualitative perspectives, BIG and BAD seem to be more frequently encoded than GOOD and SMALL in Yami. Incidentally, this conclusion also corresponds to the preponderance of negative emotion in Yami (Rau, Wu and Yang, forthcoming). Finally, all the affixes interpreted as evaluative in the qualitative perspectives convey negative attitudes towards the person who does the action; for example, mapaka-­ ‘pretend to be such and such’, delimitive ka-­‘do nothing except such and such’, depreciative intensification ka-­‘very. . . (contrary to the speaker’s expectation)’ and blaming ka-­‘it serves you right’. As these affixes evaluate negatively not the action itself but rather the person who does the action, these evaluative meanings are only secondary via metaphorical extension. Notwithstanding secondary meaning, they are by no means peripheral or trivial in the Yami system. The blaming prefix ka-­has its own morphosyntax, different from the ergative case marking in the regular Philippine focus system. A detailed discussion of the morphosyntax of ka-­is certainly beyond the scope of this chapter and awaits another paper.

Notes 1. The second line in (13c) indicates that the root is takaw and that N-­ of the maN-­prefix is a homorganic nasal, which assimilates to the place of articulation of the beginning segment of the root. The second line in (13d) indicates that the root is ap and that N-­becomes ng-­when it precedes a root which starts with a vowel. 2. The second line in (24c) indicates that the root is dangdang and that N-­ is a homorganic nasal which assimilates to the place of articulation of the beginning segment of the root.

17  Australia and New Guinea 17.1 Dalabon

Maïa Ponsonnet and Nicholas Evans

1 Introduction 1.1 Context Dalabon is a polysynthetic language of Northern Australia, with only half a dozen remaining speakers. It belongs to the non-­Pama-­Nyungan Gunwinyguan family. There is as yet no comprehensive reference grammar, but information on various aspects of the language can be found in Evans, Brown and Corbett (2001), Evans and Merlan (2003), Evans, Merlan and Tukumba (2004), Ross (2011) and Ponsonnet (2015) (see Cutfield’s 2011, 22 literature review). There are no augmentative devices in Dalabon. Diminutives, on the other hand, are frequent in emotional speech, but have not previously been reported for the language. One reason is that they do not occur with equal frequency in all contexts, and it was the deployment of methods designed to elicit emotion-­laden speech in Dalabon as part of the first author’s doctoral thesis (Ponsonnet 2014) which brought a much higher proportion of diminutive use – specifically the showing of three emotionally charged films about Aboriginal stories (Ten Canoes, Samson and Delilah and Rabbit-­Proof Fence), for which commentary was sought. 1.2  Diminutive =wurd =Wurd is an enclitic derived from the noun wurd ‘woman’s child’, whose reduplicated form wurdurd1 means ‘child’. This matches the etymology postulated by Jurafsky (1996) for diminutives in various languages. Like many diminutives worldwide, Dalabon =wurd has three uses: denoting small objects, adding emotional connotations, and pragmatic functions, particularly interactional softening (which we will not present here for reasons of space). Section 2 summarises the respective distribution of these values of the Dalabon diminutives. Section 3 discusses denotational (scalar) senses, and Section 4 emotional meanings.

2  Distributional profile =Wurd can attach to most word classes. Nouns are the most common host, especially nouns referring to persons, and less frequently animals or inanimates. =Wurd is also not uncommon on verbs. It occurs, much more rarely, on adjectives, adverbs, numerals and demonstratives. We leave these marginal cases aside to focus on the diminutives with nouns and verbs.

Maïa Ponsonnet and Nicholas Evans

402

The meanings and functions of =wurd distribute as follows: • on nouns and adjectives: low on scale of: ○ age (‘young’) on nouns referring to animates; ○ size (‘small’) on count nouns referring to inanimates, adjectives or verbs; ○ quantity (‘a small amount’) on mass nouns; ○ intensity/completeness (event), with verbs and (at least some) adjectives; • on all word classes where diminutives are attested with some frequency: emotional meanings: ○ compassion and positive empathy, approval; ○ intimacy and daily routines; • on nouns referring to inanimates and on verbs, where it may have a softening function, when reporting a negative event or mitigating an apology or criticism.

3  Denotational senses Denotational uses of =wurd are frequent with nouns, and specify that the denotatum is low on a certain scale, whether of age, size or intensity. It is most commonly encliticised to nouns referring to animates, specifically those referring to persons. These may be kin or subsection terms,2 occasionally proper names or other descriptive nouns, but the most frequent configuration is for =wurd to appear on ‘human categories’. By this we mean nouns denoting classes of persons based on criteria of age and/or gender, e.g. biyi ‘man/male person’, kirdikird ‘woman/female person’, yawurrinj ‘young man’, wurdurd ‘child’. 3.1  Categories not defined by age When a human category is defined by criteria other than age, the diminutive specifies that the referent is a member of this category at the lower range of age, i.e. as young considering the age range of the category. This applies to categories defined by gender (biyi ‘male’; kirdikird ‘female’ (1)), but also to ‘subsections’, which denote classes of persons based on a socio-­centric kinship grid. Animals can pattern like non-­age-­based human categories: =wurd applies to young animals. (1)

30037/2007 – 14’ (LB) [Narr]3 [When women get pregnant]: Da-­h-­yidjnja-­n wurdurd, ka-­h-­yin. [. . .] 2sg>3-­r-­have-­prs child 3sg-­r-­say.prs

Kardu kirdikird=wurd, kardu o biyi=wurd. maybe woman=dim maybe or man=dim ‘You have a child’, she says. ‘Maybe a baby girl or maybe a baby boy.’

=Wurd never refers to adults of small size, only to young individuals. Diminutives are unattested with nouns denoting older persons, such as nakohbanj ‘old man’. Diminutives on human terms thus operate on an age scale. Note that in the example above the speaker could have achieved reference to the same individuals without using a diminutive – the diminutive just makes the sentence more precise.

Dalabon 403 3.2  Categories defined by age =Wurd can apply to life-­stage nouns, i.e. human categories defined by age, provided that these categories themselves denote a relatively early life-­phase, e.g. yawurrinj ‘young man’ (2), yawk ‘young woman’ or wurdurd ‘child’. (2)





20120710b_003_MT 064 [TC]: Kanh kirdikird-­ngong bula-­h-­dja-­na-­ng dem woman-­many 3pl>3-­r-­foc-­see-­pfv kanunh yawurrinj=wurd ka-­ye-­yu nahda. dem young.man=dim 3sg-­sub-­lie.prs there ‘The women have seen the young man who’s over there.’

As with diminutives cross-­ linguistically, this use is frequent with wurdurd ‘child’. Wurdurd=wurd is a common alternative to wurdurd, used to refer to children of any age up to adolescence. With age-­based human categories the diminutive does not restrict the denotation to a younger member of the category: in (2) it does not mean ‘young for a young man’. Rather, it confirms that the denotatum is young, on an absolute scale. That the denotational sense ‘young’ is still present is shown by the incompatibility of =wurd with nouns denoting older persons, like nakohbanj ‘old man’. Nonetheless, with age-­defined categories the denotational input of diminutives is redundant, since youth is already part of the meaning of the host noun. 3.3  Nouns referring to inanimates When =wurd is encliticised to nouns referring to inanimates, it indicates that the referent is low on a scale of size (count nouns, (3)), or quantity (mass nouns, (4)) – not age. Terms for animals can also attract this sense. However, with inanimate nouns the scalar sense is not obligatory: =wurd can also be found on nouns denoting large things in positive emotional contexts, such as in (5). (3)



20111208_001_MT 006 [ContEl]: Bad-­dulum-­no=wurd kanidjah stone-­hill-­fill=dim there ‘There is a small stone hill there.’

ka-­h-­di. 3sg-­r-­stand/be.prs

(4) 20110530_001_MT 26 [Narr] [The speaker wishes to justify not giving money to a relative who had made a request]: Nunh kanh bad=wurd bula-­h-­ngabbu-­n ngey-­karn nga-­h-­dja-­koh-­nam-­urru-­n. dem dem stone=dim 3pl>1-­r-­give-­prs 1sg-­emph 1sg-­r-­foc-­eyes/gaze-­put-­refl-­prs ‘Then they give me just a bit of money and I manage by myself.’ (5)

20120705b_006_MT 54 [RPF] [Three children being chased hide together under a large blanket]: Bulu ka-­h-­barrkb-­ong kardu blankid=wurd. 3sg>3-­r-­cover-­pfv maybe blanket=dim 3pl ‘It’s like covering them, the good old blanket.’

Maïa Ponsonnet and Nicholas Evans

404 3.4 Verbs

=Wurd can also encliticise to verbs, after the TMA inflection.4 The only other clitics that occur in this position are possessive/oblique pronouns, which encode participants that are either possessors of incorporated nouns or not projected by the argument structure of the verb.5 Most occurrences of =wurd on verbs may be interpreted under one of the following scalar senses: • one of the arguments is young6 (6); • the event is incomplete (do x incompletely, (7)), or less intense. (6)

20120705b_004_MT 025 [RPF]: Woywoy bulu ka-­h-­naHna-­n=wurd bulu ka-­h-­djukko-­djukko-­n. interj.compassion 3pl 3sg>3-­r-­look.after-prs=dim 3pl 3sg>3-­r-­iter.r-­wash-­prs ‘Oh, she looks after [the children], she gives them a shower.’

(7) 20110530_004_MT 57 [ConvEl] [A police car looped towards us but turned away before reaching us]: Kardu ngorr bula-­h-­kurlh-­kurhka=wurd, maybe 1pl.incl 3pl>1-­r-­incep.r-­visit.prs=dim

kardu bala-­h-­men-­yin djehneng kardu ngungurru-­kolhngu-­n wah. maybe 3pl-­r-­ideas-­say.prs looks.like maybe appr.1pl.incl>-­drink-­prs liquid ‘It seems that [the police patrol] is coming to pay us a bit of a visit, maybe they believe that we might be drinking alcohol.’

However, since other readings are available when =wurd attaches to verbs (see below), it would be possible to analyse (6) and (7) in a way that obviates the scalar readings, e.g. by interpreting the =wurd in (6) as expressing approval for the subject’s care for the children (see Section 4.2). Note, though, that such ambiguities are precisely what one would expect in contexts furnishing a semantic bridge for the transition from diminutive to compassion, namely talking lovingly about activities involving children. In fact, there are no clear cases of verb+dim attracting scalar interpretations outside emotional contexts. By contrast, as will be show in Section 4.2, some occurrences fall under none of the senses above and must be explained solely with respect to an emotional context.

4  Emotional senses The above account suggests that denotational contributions of =wurd are not always very salient. They are obligatory for animate nouns (no spontaneous use of diminutives when the referent is not young), but also often redundant. With inanimate nouns, the denotational sense is not obligatory, and on verbs it is scantily attested: diminutives do not occur independent of an emotional or pragmatic contribution, and some examples attract no denotational interpretation at all. In fact, with all word classes, the use of diminutives is primarily determined by the emotional colouring of the context, which may be of one of the three types described below, or by some softening effects (which we cannot discuss here for reasons of space).

Dalabon 405 4.1 Endearment The expression of endearment or sympathy is a common use of diminutives cross-­ linguistically. In Dalabon, this connotation became particularly clear when it was possible to compare a significant number of diminutive uses in commensurable contexts, namely when two speakers were asked to comment on films. They used far more diminutives on animate nouns referring to young characters they felt sympathy for than when referring to the young characters that they did not approve of. 4.2 Compassion As described by Myers (1986, 113–17) for the Pintupi (Central Australia), compassion is a morally central, socially structuring value among Aboriginal groups. This is also the case among the Dalabon. Diminutives are abundantly used to express compassion, i.e. feeling bad because something bad happens to someone else. In (8), the speaker expresses her compassion for the hero of one of the films, a mature man of impressive stature, when he collapses after being badly wounded. Here the diminutive is placed at the end of the verb, thus avoiding a diminutive on a noun referring to someone who is not young. Note the revealing Kriol translation given by the speaker, Maggie Tukumba: bobala imin boldan, lit. ‘poor fellow he fell down’. The word bobala, from English ‘poor fellow’, is widely used in north Australian creole to express compassion in the same way as =wurd here. (8)



20120713a_002_MT 174 [TC]: Wa:h ka-­h-­rakka-­ng=wurd. interj 3sg-­r-­fall-­pfv=dim ‘Oh, he fell over poor fellow.’

At other times, speakers use diminutives to express satisfaction when something good happens to someone. =Wurd also express ‘secondary compassion’ – speakers’ approval and satisfaction when they witness someone else being compassionate, for instance people sharing with, or taking care of, someone else. In (9), the speaker comments on the heroine of one of the films taking care of her grandmother. The speaker was not fond of this protagonist, but did use a diminutive to refer to her in this particular occasion. (9)



20120719a_001_MT 205 [SD]: Kanh kirdikird=wurd buka-­h-­naHna-­n kakkak-­no. dem woman=dim 3sg>3sg.h-­r-­look.after-­prs grandmother-­3sg.poss ‘This young woman looks after her grandmother.’

These occurrences are important, because they show that speakers use =wurd not simply to express their own emotions, but also to express their desires and assessment as to how people should relate to each other emotionally. In these cases, diminutives express their moral appraisal of emotions rather than simply their own emotions. 4.3  Intimacy and familiarity Diminutives can also depict intimacy or familiar events. Dalabon diminutives occur when speakers witness an intimate scene of someone’s daily life, typically when it corresponds

Maïa Ponsonnet and Nicholas Evans

406

to a routine that is familiar to all, and that the speaker can thus identify with: making fire (10), waking up in the morning, combing one’s hair. (10)



20120720_003_LB 075 [SD]: Mimal ka-­h-­marnu-­ruru-­n.  .  . Kanh kirdikird=wurd. . . fire 3sg>3-­r-­ben-­burn.rdp-­prs dem woman=dim ‘Her fire is burning for her . . . The young woman . . .’

Diminutives, or other expressive features such as intonation, are also used when referring to a well-­known animal species, or when noting the familiar sound of an identified bird singing in the morning.

5 Conclusion The clustering of diminutive and evaluative meanings found in Dalabon is not unusual cross-­linguistically. Once one looks in more detail, however, more culture-­specific factors emerge, such as the range of evaluative semantics spanning affection for small things (e.g. children), endearment generally, compassion for those suffering but equally satisfaction at something good happening to someone else, and the expression of approval for compassion witnessed in others. In other words, the evaluative uses span a wide range of empathetic contexts, holding both between the speaker and some element of the represented event, and between the speaker and someone in the represented event who is themselves exhibiting empathy, respect or compassion. Perhaps surprisingly for cultures where the expression of empathy and compassion has been reported to be a core value, there has been little research so far on diminutives, evaluatives and other comparable expressions in other Australian languages. Notable exceptions are Wilkins (1989, 358), who reports a special ‘pity’ suffix -­penhe for Mparntwe Arrernte; Donaldson (1980, 194–5), who describes a verbal derivational suffix -­guwa-­y in Ngiyambaa which ‘indicates that the event referred to by the verb to which it is attached is emotionally affecting’ in a wide range of ways from pity through outrage to remonstration, apology and affection; Evans (2003, 473–9), who reports that in Bininj Gun-­wok, incorporation of the noun root yaw ‘(woman’s) child’ into verbs can express the smallness of one of the event participants; and Merlan (1983, 66), stating that Ngalakgan has a special ‘compassion prefix’ marking the speaker’s empathy for the absolutive argument of the verb, roughly along the lines expressed by bobala in Kriol and poor feller in Aboriginal English. The Ngiyambaa and Ngalakgan verbal affixes, in particular, are semantically close to the emotional uses of Dalabon =wurd, though in contrast to Dalabon neither Donaldson for Ngiyambaa nor Merlan for Ngalakgan suggests any link either to diminutives or to an etymology related to ‘child’. It thus appears that, although the use of verbal markers for expressing generalised empathy is not unusual in Australian languages, diminutives and words for ‘child’ are only one pathway by which this meaning gets grammaticalised. It must be stressed, though, that more thorough descriptive attention to this semantic dimension is needed for Australian languages, and it may well be that diminutives have often slipped under the radar for the reasons described in the introduction, namely that it is only in a limited set of contexts that speakers express emotion by morphological means.

Dalabon 407

Acknowledgements We would like to thank our Dalabon teachers, in particular Maggie Tukumba, †David Karlbuma, †Lily Bennett and Queenie Brennan, for their insightful discussions of this material. Evans gratefully acknowledges the assistance of the Australian Research Council (Discovery Project DP0878126: Language and Social Cognition) and Ponsonnet the support of the Endangered Documentation Program (IGS0125).

Notes 1. Wurdurd is derived from the fully reduplicated wurd-­wurd. 2. Subsections, or ‘skin names’, are a set of sixteen names (eight male, eight female) assigned to people by birth, on the basis of their parents’ subsection. Skin names determine someone’s ­position in a socio-­centric ‘kin-­space’. They are a very common way to address or refer to people in Dalabon. 3. The codes refer to recording labels, allowing for the location of the files in the AIATSIS ­audio-­visual archive or in the Endangered Languages Archive (ELDP). Abbreviations between brackets refer to types of data: [ConEl]: contextualised elicitation; [ConvEl]: conversation during elicitation; [Narr]: narrative; [RPF]: comment on Rabbit-­Proof Fence; [SD]: comment on Samson and Delilah; [TC]: comment on Ten Canoes. 4. Though we lack the space for a thorough comparison, it is worth noting that the semantic effects found here overlap only partially with those reported for Italian verbal evaluatives by Grandi (2009). 5. E.g. Kardû ko ngayh-­mele-­monwoyan=ngan ‘I better make up my swag’, where encliticised =ngan ‘my’ modifies the incorporated nominal mele-­‘swag, bedding’. 6. Presumably ‘small’ if the argument was inanimate, but there is no example.

17.2 Iatmul

Gerd Jendraschek

1 Introduction Iatmul (ISO 639–3 ian), a so-­called Papuan language, is spoken by about 40,000 people in the East Sepik Province of Papua New Guinea, along and close to a stretch of the Sepik River, about 80 kilometres inland from the northern coast of the island of New Guinea. The only genetic affiliation beyond doubt is that as a Ndu language. The Ndu family contains at least six different languages: Abelam, Boikin (Boiken), Iatmul, Manambu and Yelogu (Kaunga), as well as Ngala, which constitutes an exclave to the west of the otherwise continuous Ndu area (Aikhenvald 2008, 21). Foley (1986, 242) lists Sawos as a seventh Ndu language; this area is probably best described as a transition zone characterised by a dialect continuum connecting Iatmul with Northern Ndu, i.e. dialects of Boikin and Abelam. Previous descriptions of Iatmul morphosyntax are Jendraschek (2009a; 2009b; 2009c; 2011) and Staalsen (1965; 1972; n.d.); some of Staalsen’s data were reproduced and interpreted in Foley (1986). Probably the first account of Iatmul is a list of words in Zöller (1891), even though the term ‘Iatmul’ would not be used before Bateson (1932). Iatmul villages are trilingual, with everybody being fluent in Tok Pisin, almost everybody speaking Iatmul (although to varying degrees), and a minority having learnt English as a third language. Competence in Iatmul correlates to a large extent with age. It is the native language of almost all adults, but relegated to second language status among children, who could be labelled semi-­speakers.

2  Typological profile Iatmul is moderately agglutinative, with about eighty suffixes and five prefixes. While irregular inflection is negligible, roots may change form due to various phonological processes, and polymorphemic structures can be phonetically coalescent to an extent that makes segmentation difficult. Grammatical relations are marked on the dependents, with the exception of the subject relation, which is cross-­referenced on the verb. The only two major word classes are nouns and verbs. These are clearly distinguished by the morphology they can take, and there is little derivation across these two classes. Deverbal stem-­forming nominalisations mostly involve semi-­affixes that can be traced back to free nouns. Some deverbal nominalisations involve reduplication. While there are no morphological devices to verbalise nouns, coalescence between nouns and verbs yields lexicalised complex predicates.

Iatmul 409

3  Evaluative morphology 3.1  Gender allocation While Iatmul has no derivational affixes for diminutives or augmentatives, the two-­way gender system is routinely used to express quantitative evaluation. Whereas nouns for referents high on the empathy hierarchy (notably humans) have inherent gender, less empathy-­inducing (notably inanimate) referents receive gender on the basis of their (perceived) size. Only nouns without inherent gender allow speakers to allocate a gender, depending on their characterisation of the referent. Note that some nouns without natural gender do nevertheless have a fixed gender in Iatmul. This is notably the case for natural phenomena: nya ‘sun’ has masculine gender, whereas bap ‘moon’ has feminine gender. Morphologically, Iatmul nouns have covert gender, i.e. gender is marked not on the noun itself, but on its modifiers and in subject–verb agreement. In example (1), the relevant noun is nyaan ‘child’, which can refer to both male and female referents. (1)

kan

nyaan wega yi-­di’ child market go-­3sg.m ‘this (male) child went to the market’ dem1.sg.m

In (1), the gender of nyaan is signalled by the masculine form kan of the demonstrative, and gender agreement on the verb. In (2), the demonstrative is in its feminine form kat, and the agreement suffix -­li’ on the verb also signals that the subject referent is female/feminine. (2)

kat

nyaan wega yi-­li’ child market go-­3sg.f ‘this (female) child went to the market’ dem1.sg.f

While animate referents with natural gender (i.e. sex) can be combined only with the agreement marker that corresponds to that gender, referents without natural gender can be allocated either gender. In this case, small size is expressed by assigning feminine gender, while masculine gender correlates with large size, or represents the unmarked choice. One morphosyntactic manifestation of gender allocation is the combination with a demonstrative. As illustrated in (1) and (2), the other way of signalling gender is verb agreement. However, while the agreement markers in (1) and (2) signalled the natural gender of nyaan ‘child’, gender allocation in (3a) is not triggered by natural gender (only female mosquitoes bite!), but has a purely evaluative function. (3) a. ki’viya vaali’-­ka-­di’ mosquito bite-­ prs-­3sg.m ‘a (large) mosquito is stinging [me]’ b. ki’viya vaali’-­ka-­li’ mosquito bite-­ prs-­3sg.f ‘a small mosquito is stinging [me]’ By choosing the – biologically inappropriate – masculine gender, the referent of ki’viya ‘mosquito’ is characterised as ‘large’, or rather ‘not small’, as the masculine is the

Gerd Jendraschek

410

f­ unctionally unmarked gender one would choose if no characterisation were intended. The feminine gender in (3b) is the marked choice, correlating with perceived smallness of the referent. For a better understanding of the pragmatics of evaluative gender allocation, we shall next see some textual examples where the evaluation applies to abstract referents. In (4)– (6), the relevant nominal expressions are saaki ‘story’, vi’wavaak ‘what I have seen’, vaak ‘topic; fact’ and wapuchapuk ‘story’. In (4), the speaker marks the noun saaki ‘story’ as feminine (via kat, a feminine form of the demonstrative), thereby expressing that the story is rather short. (4)

kat saaki [. . .] ba’gu-­kat yi-­b-­a saaki dem1.sg.f story dance-­ dat go-­2/3du-­sr story ‘this story is the story of the two going dancing’

Allocation of feminine gender fulfils a similar diminutive function in (5), where the evaluated expression is wa’tnana vaak ‘the crocodiles’ topic’ or, to give a freer translation, ‘the story about the crocodiles’. Note that in ap=maak ‘that topic’ the demonstrative cliticises to its host, thereby triggering the phonetic change of the latter from vaak to maak. (5)

ap=maak

dem2.sg.f=topic

apmâ

ak

dem2.sg.f

wa’t-­na-­na crocodile-­gen-­gen

vaak topic

yi-­ka

gi’li’-­ka-­li’ finish-­prs-­3sg.f ‘that topic, that crocodiles’ topic comes to an end like that’ dem2:adv go-­ dep

The narrators of examples (4)–(5) added a diminutive evaluation of the stories they were telling me by marking them as feminine, perhaps as a sign of modesty. In contrast, example (6) shows that masculine gender is equally possible. As explained above, choice of masculine gender does not necessarily imply that this story is perceived as particularly long by the narrator, masculine being the unmarked member of the gender opposition. (6)

Wun-­a wapuchapuk 1sg-­gen story ‘my story ends like that’

apmâ dem2:adv

gi’li’-­ka-­di’. finish-­prs-­3sg.m

Finally, examples (7)–(11) exemplify the evaluative use of gender with nouns of various semantic classes; in each case, the feminine (b) versions are comparable to diminutives. (7) a.  an da dem2.sg.m thing ‘that (big) thing’ b. at da dem2.sg.f thing ‘that small thing’ (8) a. kan laavu dem1.sg.m banana ‘this (big) banana’

Iatmul 411 b. kak laavu dem1.sg.f banana ‘this small banana’ (9) a. ankwi ji’vwa dem2.sg.m.nmls:rem work ‘that (hard) work’ b. akkwi ji’vwa dem2.sg.f.nmls:rem work ‘that simple work’ (10) a. ankwi mi dem2.sg.m.nmls:rem wood ‘that (big) tree’ b. akkwi mi dem2.sg.f.nmls:rem wood ‘that small tree’ (11) a. ankwi gaai dem2.sg.m.nmls:rem house ‘that (big, long) house’ b. akkwi gaai dem2.sg.f.nmls:rem house ‘that small, short house’ In these elicited examples, demonstratives have been included in order to produce an overt marking of the allocated gender. Iatmul demonstratives are morphologically quite complex, with initial segments representing the deictic degree (proximal, distal, anaphoric or discourse-­relevant) and further segments indicating remoteness or degree of nominalness; these distinctions are irrelevant for the evaluative function of gender, so I will forgo a detailed analysis here. For the masculine (a) versions of (7)–(11), the modifiers in the English translation are in parentheses to signal that the evaluative interpretation (as an augmentative) is not obligatory, masculine gender also functioning as the unmarked member of the opposition. The Iatmul examples so far have shown that quantitative evaluation can be expressed through nominal inflection, and also via agreement on the predicate. Next we will take a look at adjective–noun compounds and discuss to what extent they should be analysed as examples of evaluative morphology. 3.2  Adjective–noun compounds Certain adjectives expressing evaluation form compounds with nouns. Insofar as these are not ad hoc compounds, but the result of conventionalised processes, the operation has derivational characteristics, and suggests that these adjectives function as evaluative semi-­prefixes. This concerns the adjectives maak ‘small’ and kavle ‘bad’; no convincing examples could be found for morphological use of their antonyms ni’ma ‘big’ and apma ‘good’. This is in line with the typological tendency for diminutives to be more frequent than augmentatives (Grandi 2011b, 7).

412

Gerd Jendraschek

While the lexical use of these adjectives is not relevant for a study of evaluative morphology in Iatmul, there are both phonological and semantic criteria that distinguish their lexical use from their use as semi-­prefixes. In (12), the most obvious difference is phonological: whereas the morphologically free adjective in (12a) is also phonologically unbound, the semi-­prefixal use triggers phonological assimilation at the morpheme boundary. At the same time, the contraction to a single phonological word leads to vowel ­shortening in the first syllable of the second element. (12) a. maak vaala small canoe ‘a small canoe’ b. maap-­mala dim-­canoe ‘small type of canoe’ In addition to the phonological consequences of adjective–noun coalescence, example (13) also illustrates the semantic difference between the two uses. While (13a) is an instance of lexical ad hoc modification with compositional semantics, the compound/derivation in (13b) has been lexicalised with corresponding semantic narrowing. (13) a. maak nyaan small person ‘a small person’ b. maat-­nyan dim-­person ‘child’ As for pejorative evaluation with kavle ‘bad’, the example that best illustrates the semi-­ prefixal use is (14b). The phonological coalescence manifests itself here through monophthongisation in the second element. (14) a. kavle gaai bad house ‘a bad house’ b. kavle-­ge pej-­house ‘toilet’; ‘mourning house’ Note, however, that the (b) version has been lexicalised with two unrelated meanings, ‘toilet’ and ‘mourning house’. The meaning as ‘mourning house’, that is the house where the family of the deceased lives and receives guests during the weeks following the funeral, probably goes back to the expression kavle yi ‘to die’, literally ‘go bad’; furthermore, this ‘mourning house’ has nothing pejorative about it. Therefore, used with this meaning kavlege may be a genuine compound, rather than a case of evaluative derivation. Further examples of pejorative meaning specialisation are kavle-­gu ‘home brew’ < ‘bad water’ and kavle-­yaki ‘marijuana’ < ‘bad tobacco’.

Iatmul 413 3.3  Event-­specifiers Since verbs typically express the dynamic component of a situation, they may host derivations that express nuances with respect to the dynamic nature of that situation. In other words, such derivations specify how the event unfolds. The term ‘event-­specifier’ for this type of derivational morpheme has been adopted from van Breugel’s grammatical description of Atong (van Breugel 2014, 376). Iatmul has at least nine such event-­specifying derivations. Some of them are relevant for a description of evaluative morphology in Iatmul insofar as they quantify the event. This quantification may concern the temporal extent of the situation, or the entities involved in the event. The event-­specifier -­pwali~vwali expresses that an event went on continuously over an undefined period. (15) ki’-­vwali-­di cont-­3pl eat-­ ‘they were eating all the time’ The suffix -­gi’li’ ‘do all’ (derived from the verb gi’li’ ‘finish’) is attached to a verb to quantify the subject (16) or object referents (17). (16) kwalabi ki’-­gi’li’-­laa ya-­di mango eat-­ finish-­consec come-­ 3pl ‘all of them ate mangoes and came’ (17) kwalabi ki’-­gi’li’-­di mi’na di-­kak mango eat-­ finish-­3pl as.soon.as 3pl-­dat ‘after theyx had eaten all the mangoes, theyy beat them’

viya-­di hit-­3pl

Next, there is -­yaki ‘do completely’, referring to an event that was carried out to completion (18). This contrasts with -­tabi’ ‘do incompletely’ (19). (18) ki’-­yaki-­di eat-­ compl-­3pl ‘they ate all there was’ (19) ji’vwa kut-­tabi’-­li’-­di ya-­di’ work do-­ incompl-­ipfv-­3pl come-­3sg ‘they had not completed the task when he arrived’ We may finally mention the frustrative derivations here. A frustrative ‘indicates that the action was done to no avail – that is, the desired result was not achieved’ (Aikhenvald 2008, 293). Iatmul has two frustrative markers: -­tavi~lavi ‘try to do’ or ‘be about to do (but then abandon)’, ‘not do as expected’, glossed as ‘conative’; and -­si’ki’k ‘do in vain’. (20) ki’ki’da ki’-­lavi-­di food eat-­ cntv-­3pl ‘they could not finish the food (although they tried)’

Gerd Jendraschek

414 (21) li’-­kat walai-­si’ki’k’-­wun 3sg.f-­dat call-­ frust-­1sg ‘I tried to call her’ (= I called but she did not hear me)

The event-­specifier suffixes in Iatmul have often been misanalysed as ‘temporal suffixes for dependent verbs’ (Foley 1986, 182, following Staalsen 1972, 52; same in Roberts 1997, 111) as they routinely occur in clause linking. They are, however, not restricted to same-­subject clause linking, but can also appear in different-­subject clause linkage (22), and even in verb–noun compounds such as that in (23). (22) wun-­a ki’ki’da ki’-­labi’-­wun ki’-­di’ incompl-­1sg] [eat-­ 3sg.m] [1sg-­gen food eat-­ ‘(as) I did not finish my meal, he ate it’ (23) ki’-­labi’-­yaki eat-­ incompl-­tobacco ‘cigarette butt’ 3.4 Reduplication Reduplication of adjectives does not have a regular semantic outcome. In some cases, the reduplication produces an intensifying effect, as in gwaigwai ‘very close’, derived from simple gwa’i ‘close’ (with a glottal stop, which is lost in the reduplication due to phonological simplification). A similar case is the reduplication (here with onset alternation) of kavle ‘bad’ to kavlesavle ‘shame on . . . ’, a common curse expression. In other cases, we observe a downgrading or approximating effect similar to -­ish in English. Compare examples (24) and (25). (24)

kan

(25)

kan

laavuga

dem1.sg.m paper

‘this book is big’

ni’man’-­a big-­3sg

laavuga ni’man ni’man’-­a paper big big-­3sg ‘this book is biggish’ (actually not really big, but too big to be small) dem1.sg.m

Outside evaluative contexts, reduplication in Iatmul is further attested in the following: • ideophones, where it expresses an iterative component of the situation; • the formation of reciprocal expressions, such as the adverb awaksawak ‘each other’, derived from awak ‘in (re)turn’, or paaku saaku (kasa) ‘(to play) hide and seek’ from paaku ‘to hide’; • distributive adverbials derived from nouns: kava ‘place’ > kavakava ‘everywhere’; • distributive numerals: ki’ta ‘one’ > ki’ta ki’ta ‘one each; one by one’; • emphatic (contrastive) personal pronouns: wun ‘I’ > wuno wunan ‘I, however’ (i.e. in contrast to others); • deverbal nominalisations: wa ‘to say’ > wawa ‘pronunciation’; vi’ ‘to see’ > vi’vi’ ‘perspective, interpretation’.

Iatmul 415

4  Areal comparison A very similar system of gender allocation has been described for the genetically related and neighbouring language Manambu, where gender assignment is based on shape, size, quantity, and extent or intensity (Aikhenvald 2008, 123). As in Iatmul, gender is covert, but a large house can be marked as masculine and a small house as feminine via agreement. An important difference from Iatmul, however, is that in Manambu it is female gender which is formally and functionally unmarked (Aikhenvald 2008, 124); the question as to which system is more conservative and which due to innovation would go beyond the scope of this chapter. Manambu gender alternation produces effects that can go beyond size and shape: a feminine night refers to the beginning of darkness, contrasting with a masculine night, referring to that part of the night when it is completely dark. Mosquitoes are assigned masculine gender if occurring in huge quantities (Aikhenvald 2008, 119–20). The use of feminine gender for shorter stories corresponds to what we have seen for Iatmul.

5 Conclusion We have seen that gender is only weakly grammaticalised in Iatmul, so that semantic and pragmatic factors play an important role in gender assignment, and allow speakers to express subjective quantitative evaluation of a referent by assigning one of the two genders. The fact that another Ndu language, Manambu, has a very similar gender system suggests that this is an areal phenomenon. We have further seen the development of the adjectives for ‘small’ and ‘bad’ into evaluative semi-­prefixes, used as a tool for lexical innovation. Evaluative event-­specifiers derive verbs quantifying the extent of the event or the involvement or affectedness of participants. Finally, reduplication is used for various purposes, one of which is intensive or approximative evaluation.

17.3 Jingulu

Rob Pensalfini

1 Introduction Jingulu, a language of the Barkly Tableland of Australia’s Northern Territory (Chadwick 1975; Pensalfini 2003), has some morphological properties that are atypical of the region. One relevant to the current discussion is the high degree of morphological fluidity it shows, in which case markers may appear on verbs as part of a system of switch reference (Austin 1981), and tense markers may appear on nouns to indicate spatial deixis (Pensalfini, 2014). Of particular interest to this volume is the morpheme -­kaji, which has been inadequately described as an adverbialiser (Pensalfini 2003, 73–6), but which under further scrutiny is best described as an evaluative morpheme. While it does not perform the most prototypically evaluative functions such as diminution, augmentation or pejoration, it does evaluate the referent against a normative ideal (Jingulu, typically of languages of the region, does not have diminutive, augmentative or comparative morphology). Like evaluative morphology in other languages, it does not neatly fit the derivational/inflectional dichotomy. While -­kaji could indeed be argued to derive adverbs from other elements in its prototypical usage, we will see that at its core the morpheme evaluates, which may result in a category shift on the part of the element bearing the morpheme, but does not necessarily do so. Unlike a true derivational morpheme, -­kaji is not restricted in terms of the parts of speech to which it can attach. Furthermore, it can attach to both inflected and uninflected elements. Finally, once -­kaji attaches to an element, the resulting word may either function as an adverb or be further inflected in accordance with the part of speech to which its root belongs.

2  Parts of speech Like many Australian languages, Jingulu words fall into three basic categories or parts of speech: verb, nominal and adverb (there is no distinct category of adjectives, and words that translate as either nouns or adjectives in English fall into the part of speech called ‘nominal’). It should be noted that parts of speech are here defined in terms of syntactic distribution and morphological selection, in other words in formal terms. Consider the following sentences: (1) a. Ngaaku birri-­wunya-­ana-­miki marluka-­yili-­rni. later visit-­ 3du-­1obj-­came old.man(m)-­du-­erg ‘Those two old men came to see me later.’

Jingulu 417 b. Wilinja ya-­jiyimi. countryman(m) 3sg-­come ‘Our countryman is coming.’ Note that word order within Jingulu clauses is completely free. The three words in (1a) are an adverb, a verb and a nominal respectively, while (1b) consists of a nominal followed by a verb. Each of these parts of speech has its own formal properties, the most relevant of which are discussed below. A complete description of these parts of speech and how they interact can be found in Pensalfini (2003). 2.1 Nominals Nominal words typically consist of a root (which may include a gender marker – Jingulu has four genders), followed by a number suffix and a case marker, as required, in that order. Case marking may further be followed by optional discourse and deictic morphology, but that does not enter into the current discussion. The word marluka-­yili-­rni in (1a) provides a clear example of Jingulu nominal structure. The root marluka is masculine (and ends in the typical masculine -­a).1 This is followed by the marker of dual number -­yili (Jingulu distinguishes dual from plural by distinct morphology, and leaves singular nominal unmarked). Finally, the nominal is inflected with the ergative case marker -­rni, which appears on nominals functioning as the subject of a transitive clause (Jingulu leaves subjects of intransitive clauses and objects unmarked for case). In (1b) the nominal word is wilinja, which, being singular and an intransitive subject (absolutive case), bears neither overt number nor overt case morphology. It does end in the typical masculine -­a. Words which translate into English as substantive nouns, adjectives (descriptive or predicative), demonstratives and pronouns all behave in more or less this fashion and are all examples of the nominal class. 2.2 Verbs Verbs show complex inflection in Jingulu, and typically take one of two forms. In (1a) the verb birri-­wunya-­ana-­miki consists of a root birri-­, followed by subject then object agreement markers, and a final morpheme which encodes some subset of tense, aspect, mood and associated motion (TAMM). Subject agreement encodes the person and number of the subject (so wunya-­on this word tells us that the subject is third person and dual), while object agreement encodes person only (hence -­ana-­is a first person object, but does not specify number) – number is either inferred by context (as in the case of (1a)) or else (optionally) specified by the use of a free object pronoun. Third person singular subject agreement, and third person object agreement, are typically null (unmarked), with one important exception, discussed below. The final morpheme -­miki encodes both past tense and movement away from the speaker or salient referent. Pensalfini (2003) calls this final element the light verb, because it is the only obligatory part of the verb. Sentence (1b) demonstrates a verb, ya-­jiyimi, which contains no root, consisting only of a subject agreement marker (ya-­for third person singular) and the element -­jiyimi which encodes both present tense and motion towards the speaker or salient referent. The basic notions come, go, do and be in fact cannot be expressed with an initial root,

Rob Pensalfini

418

but only using agreement with a light verb (final element). In these cases, third person singular subject agreement is overtly expressed with the morpheme ya-­, instead of being null as it usually is, presumably because light verbs are formally affixes and cannot stand alone without either a root or an agreement marker. 2.3 Adverbs Adverbs, such as ngaaku in (1a), are formally distinguished by their inability to take inflection of any sort. They do not inflect for number, gender, case or tense. They can occur anywhere in a clause, but have a tendency to be clause-­peripheral if they modify the entire proposition, or adjacent to the verb if they modify the verb alone. Adverbs can be used predicatively, in which case they require a verb to occur in the clause with them: (2) Ngaaku ya-­yi. later 3sg-­fut ‘It will be/get later.’ In (2) the tense inflection cannot attach directly to the adverb ngaaku, but instead must head its own verbal word, triggering overt third person singular subject agreement.

3  -­kaji deriving adverbs from other lexical items The morpheme -­kaji typically carries a sense that the word’s referent is an ideal or superior example of its type. While Chadwick (1975, 41) describes it as an ‘emphatic’, its meaning is not always, or not even typically, to emphasise the word to which it attaches. Its precise meaning varies depending on the context, but Pensalfini (2003) glosses it as through because it typically translates into English as ‘thoroughly, properly, right through, directly, very much so’ or something similar, as all the examples in this section demonstrate. Pensalfini (2003, 73) considers it to be an adverbialiser because, in its prototypical use, it appears to change the part of speech of its host from a co-­verbal root to an adverb: (3) a.  Mamambiyaku diyinu maja-­mi ila-­kaji soft bloodwood.gum get-­ irr put-­through ‘The soft bloodwood gum is got to put directly onto sores.’

mandarra-­ngka. sore-­all

b. Nyamini-­rna ngaliminymini nginaniki diyim-­kaji ya-­ju ngini-­rni jangki dem(f)-­foc bat this(f) fly-­through 3sg-­do dem.ntr-­foc high.up ‘The bat is flying all around above us.’

c. Nyinuwa jarrkaja-­nga-­rruku durdurdbi-­kaji this.way run-­ 1sg-­went beat-­through ‘I went for a run and now my heart is thumping.’

nga-­nu jingirdi-­rni. 1sg-­did heart-­foc

d. Kurrubardu ngirrma-­nga-­yi, mujiya-­kaji nga-­ka. boomerang make-­ 1sg-­fut forget-­through 1sg-­pst.hab ‘I want to make a boomerang but I’ve completely forgotten how.’

Jingulu 419 In all of the sentences in (3), -­kaji is attached to an element that is usually a verbal root, which is to say that these morphemes would usually be affixed to agreement and TAMM markers (light verbs). For example, the notion ‘I’ve forgotten (how)’ in the final clause of (3d) would usually be expressed as mujiya-­nga-­ka. The addition of -­kaji to the root creates a separate word. This clause could also have been Nga-­ka mujiya-­kaji. Other criteria such as stress and vowel harmony (not discussed here but see Pensalfini 2002 and 2003 for details) also show that -­kaji creates a new word boundary in these cases. The semantic effect of -­kaji is to add the sense of having completely forgotten, as opposed to having (merely) forgotten. In every instance in which it occurs, -­kaji performs an evaluative function. In (3a) ‘put’ becomes ‘directly put, put right on’; in (3b) ‘fly’ becomes ‘fly around, really fly’; in (3c) ‘beat’ becomes ‘thump, beat hard, really beat’; and in (3d) ‘forget’ becomes ‘completely forget’. The morpheme does not affect the aspectual interpretation of the clause, which is essentially determined by the TAMM marker on the verb. Sentence (3b) provides another clear example of how -­kaji creates a new word. Note that the verbal element in this sentence, ya-­ju, involves overt marking of the third person singular subject, which is only found in verbs with no initial root, as discussed in Section 2. All of the examples thus far suggest that -­kaji is a typical derivational affix, attaching to stems of one class (verbal roots) to derive another (adverbs), and in each case the result is an adverb with relatively predictable semantics. However, derivational affixes are typically selective about the parts of speech to which they will attach. As the examples in (4) demonstrate, however, -­kaji attaches to nominal and adverbial stems as well as to verbal ones: (4) a.  Mankiya-­nu dibij-­kaji ya-­rruku. sit-­ 3sg-­did outside-­through 3sg-­went ‘She sat here and he went right outside.’ b. Jama-­rni-­rni yaba, yaba-­kaji that-­ erg-­foc youth youth-­through ‘That’s a youth, a proper young man.’

jama-­rni-­ma. that-­erg-­emph

c. Ambaya-­nga-­ju Jingulu-­kaji. speak-­ 1sg-­do Jingulu-­through ‘I’ll talk pure Jingulu right through.’ d. Ngunu buba miji-­yirri! dem.ntr fire take-­go.imp ‘Get some firewood! Why, precisely?’

Nyambana-­kaji? why-­through

In (4a), -­kaji is suffixed to an adverb. In (4b–c), it is suffixed to a substantive nominal (noun). In (4d), it is suffixed to a question word, arguably an adverb. This suffix does not adhere to the sorts of selectional restrictions typical of derivational morphology.

4  -­kaji attached to inflected words It is has long been observed that inflection occurs ‘outside of’ derivation with respect to the root. That is, if a language has only suffixes, then derivational suffixes will precede inflectional ones. Within the standard lexicalist view, this is because derivation is a process that

Rob Pensalfini

420

occurs within the lexicon, prior to the insertion of lexical items into syntactic structures, while inflection reflects syntactic relations. If -­kaji were truly a derivational morpheme, in this view, we should not find it suffixed to elements which are already inflected. However, as the examples in (5)–(6) demonstrate, -­kaji can attach to words that are already fully inflected nominals or verbs. (5) a. Nyamina-­rni-­nu bardakurri-­rni-­kaji dem.f-­foc-­did good-­f-­through ‘She’s really well now.’

ya-­ju. 3sg-­do

b. Nginduwa ya-­jiyimi bininja-­rni ya-­jiyimi man-­foc 3sg-­come this_way 3sg-­come ‘Men are coming this way now for a proper fight!’

bunbaku-­rna-­rni-­kaji! fight-­dat-­foc-­through

c. Ngiji-­wunyu-­nu ngiji-­warna-­nu ngirraku ngaya ngunu-­mbili-­kaji. see-­3du-­did see-­3sgsubj.1obj-­did 1pl.excl.acc 1sg.nom dem.ntr-­loc-­through ‘Those two saw me and him right over there.’

d. Nyamba-­arndi-­kaji nya-­rriyi-­rni? what-­ ins-­through 2sg-­will.go-­foc ‘How exactly will you go?’ In (5a), -­kaji attaches to a nominal which has been inflected with feminine gender suffix to show agreement with the preceding demonstrative. In (5b–d) it appears on elements which bear semantic case markers (which function much like prepositions in English): the dative and focus marking in (5b), the locative in (5c) and the instrumental in (5d). In all of these cases, the element to which -­kaji attaches is a fully inflected nominal, and the inflection itself is determined through its relationship to other elements in the clause, and therefore is not derivational or lexical in the traditional sense. (6) a. Imbiyi-­mindu-­wardi: Jingulu, Kuwarrangu, Warlmanku imbiyi-­mindu-­wardi-­kaji. speak-­ 1du.incl-­hab Jingulu Kuwarrangu Warlmanpa speak-­ 1du.incl-­hab-­through ‘So now we speak: Jingulu, Kuwarrangu and Warlmanpa – we speak them all now.’ b. Ambaya-­ardi-­kaji amanjamanja-­kaji, nginda-­kina-­mbili-­kina. speak-­ hab-­through children-­through that(m)-­neg.indf-­loc-­neg.indf ‘No children here speak it at all now.’ c. Ngawu-­nu maja-­nya-­yi-­kaji, nyamirni-­kaji ngawu-­nu maja-­nya-­yi, home-­did get-­2sg-­fut-­through 2sg.erg-­through home-­did get-­2sg-­fut, Kirnbininku-­nu. Kirnbininku-­did ‘You will come to know this here camp properly now, you’ll really get it, this here Kirnbininku.’

The sentences in (6) show -­kaji attached to fully inflected verbs, following agreement and TAMM marking. As these words are fully inflected for syntactic properties of the clause, such as tense and agreement, a derivational suffix should not be able to attach to them.

Jingulu 421

5  -­kaji followed by further inflection As explained in Section 2 (Section 2.3 in particular), adverbs in Jingulu do not inflect. If -­kaji truly derives adverbs, then no inflectional suffix should be able to attach to the derived form (this does not apply to discourse or emphatic morphology, which is occasionally found on adverbs). However, the examples in this section all show words ‘derived’ by -­kaji that then take the range of inflections appropriate to the stem as if it had not had -­kaji attached to it. (7) a. Jamaniki-­la-­wa ya-­marri marlarluka-­kaji-­mbili-­rni, Waramunga-­rni this(m)-­pl-­will_go 3sg-­distpst old.man(rdp)-­through-­loc-­foc Waramungu-­ erg ngini-­ngka-­rni. that(ntr)-­all-­foc ‘These Warumungu people came right here to where our old folk were.’ b. Karrijba-­kaji-­ka ya-­jiyimi jani yu-­garn? road-­through-­pst.hab 3sg-­come q 2sg-­go ‘Who’s that coming down that there road you were driving along?’

The morpheme -­kaji appears between a nominal stem and its locative case marker in (7a), and between a nominal stem and a deictic marker in (7b).

6 Conclusion The Jingulu morpheme -­kaji is an evaluative morpheme, marking the referent of its host as a proper or ideal instantiation of its type. Previously treated either on semantic grounds as an emphatic marker (Chadwick 1975) or on formal grounds as a category-­changing derivational morpheme (Pensalfini 2003), it resists these traditional categorisations. As an evaluative morpheme, its distribution is determined principally by semantic scope, and not by morphosyntactic function. It does not show the selectional restrictions typical of derivational morphemes, and can occur outside of inflection. Furthermore, it can occur inside of inflection, without altering the selectional properties of the stems to which they attach. There is another morpheme, -­nama, in Jingulu which shows the same formal properties. It sometimes appears to derive adverbs, and can occur either inside or outside of inflection. (8) a.  Jaburra-­nama nga-­rruku. before-­time 1sg-­went ‘I already went.’ b. Ajajika-­nama mindu-­wa? how.far-­time 1du.incl-­will.go ‘For how long are we going?’ c. Jama-­rni-­rni jawularri-­nama. that-­ erg-­foc young.man-­time ‘He’s still a young uninitiated man.’

422

Rob Pensalfini

d. Marlarluka-­rni-­mbili-­nama ya-­marri. old.man(rdp)-­foc-­loc-­time 3sg-­distpst ‘In the days of the old folk they did.’ Like the evaluative -­kaji, this morpheme also seems to exert semantic scope, but its interpretation is not evaluative. It rather introduces the notion of time and can translate as ‘in the time of, at the time, while, during’ Unlike -­kaji, the morpheme -­nama seems to have an aspectual affect. These two morphemes therefore seem to form a formal class, having the same distribution (which is unlike any other Jingulu morpheme) and similar scope selection, but performing different functions – one evaluative, one aspectual.

Note 1. Jingulu encodes gender in a way that is reminiscent of Romance languages. Each of the four genders has a characteristic ending (masculine -­a, feminine -­irni, vegetable -­imi and neuter -­u or consonant-­final). However, there are numerous exceptions – not all masculine nominals end in -­a, and not all nominals ending in -­a are masculine. Therefore roots and their gender endings are glossed as a single morpheme here and in other literature on Jingulu.

17.4 Kaurna

Rob Amery

1 Introduction Kaurna, the language indigenous to Adelaide, the capital city of the state of South Australia, lay dormant for more than a century, but has been undergoing a sustained revival since the 1990s (Amery 2000). Apart from some place names and a handful of common words still in use in Nunga1 English, the Kaurna language is known only from written records, almost all of which were recorded in the mid-­nineteenth century by German missionaries Christian Teichelmann, Clamor Schürmann and Samuel Klose and a number of other British, German and French observers. There are only two very short word lists and a few additional words recorded in the early twentieth century (Bates 1919; Black 1920; Tindale, n.d.). Teichelmann and Schürmann (1840), henceforth referred to as T&S, produced a twenty-­four-­page sketch grammar, a vocabulary of 2,000 words and a compilation of 200 sentence examples with English translations. Following this publication, Teichelmann continued working on the language and compiled a ninety-­nine-­page handwritten manuscript (Teichelmann 1857) and a three-­page paper on Kaurna verbs (Teichelmann 1858), the originals of which are held in Sir George Grey’s collection in the South African Public Library, Cape Town. Teichelmann (1857), henceforth referred to as TMs, has somewhat more complex entries than T&S (see Simpson 1992). Kaurna is a typical Pama-­Nyungan2 language and is thus almost exclusively suffixing, with a good range of case suffixes, both core and peripheral. Tense and aspect are also marked by a range of suffixes and occasionally appear as suppletive forms. Derivation too is achieved through suffixation or reduplication and in a number of instances by a combination of suffixation and reduplication. Compounding in nouns and verbs is common as well. Pronominal enclitics, as reduced forms of independent pronouns, are commonly used. Prefixation is almost non-­existent in Kaurna, though as we shall see, limited prefixation is perhaps beginning to emerge from reduced compounds, particularly with the initial element tuku-­‘small’.3 As in other Pama-­Nyungan languages, word order is relatively free in Kaurna, with every possible permutation – SV, VS, SOV, OSV, SVO, OVS, VSO and VOS – having been recorded in the historical corpus, just as we would expect from our knowledge of related languages. Word order within the noun phrase also appears to be free, with meyunna pulyunna (people + black) and pulyunna meyunna (black + people) ‘Indigenous/ Aboriginal people’ having been recorded. Furthermore, as in some other Pama-­Nyungan languages, noun phrases may be discontinuous, with case-­marking or number suffixes determining the composition of the phrase, as in (1):

Rob Amery

424 (1)

Kaurna (T&S 1840, 16): Nantu-­rla tuta-­ngga maie-­ndi Horse-­du grass-­loc graze-­prs ‘Both the horses are grazing.’

purla. 3du

For the purposes of this chapter I will draw almost all observations from T&S and TMs. Other nineteenth-­century sources are primarily word lists, often poorly recorded, and the few sources which include limited sentence material in Williams (1840), Koeler (1842), Wyatt (1879) and Stephens (1889) are drawn from a pidgin Kaurna and are not representative of the true Kaurna language; nor do they reveal anything about evaluative morphology (EM). Only the twentieth-­century source (Black 1920) includes some simple sentences from the true Kaurna language, but again, nothing that illuminates the topic under consideration here. There are no sound recordings of Kaurna as it was spoken in the nineteenth century before it became dormant. As a result there is an element of guesswork in interpreting the historical records (see Amery 2013). For the purposes of this chapter, there are some difficulties in determining the exact meaning of lexemes and suffixes. There are a limited number of examples to draw from and we simply cannot go and elicit further examples or seek clarification. We are forced to work with what we have. Fortunately we can turn to related languages for inspiration. Adnyamathanha is still spoken to some extent in the Flinders Ranges, some 400–500 km to the north of Adelaide, and is reasonably closely related to Kaurna, with many cognates readily identifiable. Given the nature of the Kaurna corpus, how much EM would we expect to have been documented, even if it were there used productively in the language? Whilst there are many semantic adjectives4 documented in the vocabulary, very few actually appear in the many hundreds of sentence examples included in T&S and TMs. Of the 200 sentence examples in T&S (1840, 65–72), there are at the most just four semantic adjectives used to modify a noun: gadla tauara (fire + great) ‘large fire’ (T&S 1840, 70), gadla ngarnda (fire + much) ‘large fire’ (T&S 1840, 68), burkanna turlanna (old.man-­pl + angry-­pl) ‘warlike men’ (T&S 1840, 66), and turlabutto meyu (angry-­full.of + man) ‘angry man’ (T&S 1840, 70). In addition, the compounds pulyunna meyu ‘black man’ and pinde meyu ‘white man’ appear in a number of sentence examples. However, many more such compounds or phrases, such as warra wilta (language + hard) ‘correct language’ and yukuna warra (crooked + language) ‘untrue or false statement; offensive language’ are cited within the vocabulary. There are no phrases in either T&S or TMs where the noun is modified by tukutya ‘little’, though there are some compounds (discussed later). I find no examples where two or more semantic adjectives are used in the same sentence. Yet we would certainly expect this to be possible and even frequently used, within the language. Some more marginal constructions in languages, such as contrastive focus reduplication in English (Ghomeshi, Jackendoff, Rosen and Russell 2004), may be seldom encountered in a sizeable corpus. There could easily be additional undocumented morphology in a poorly documented language, such as Kaurna. Still, the corpus, such as it is, does reveal some important insights.

2  Evaluative morphology There is little evidence of EM in Kaurna, not just because the documentation is somewhat limited, but because it was probably never there. Rather, it appears that Kaurna is lexically rich in terms of differentiating on the basis of age, size and gender, though gender is not

Kaurna 425 grammaticalised.5 There are totally distinct lexemes yakkana ‘older sister’, yunga ‘older brother’ and panyappi ‘younger sibling’ and even distinct terms such as kuttari ‘deceased older sister’. Similarly for animal terms, the male, female and offspring are often denoted by totally different lexemes, as in tarnda ‘male red kangaroo’, kurlo ‘female red kangaroo’, nantu ‘male grey kangaroo’, wauwe ‘female grey kangaroo’, kurtaka ‘young kangaroo’ but idla ‘joey’ (young kangaroo). Many other animal species have distinct terms for their young. This is a typical pattern found in Aboriginal languages. There is just one instance documented by Wyatt (1879) which, taken alone, suggests that reduplication might have a diminutive function. He lists korre ‘male emu’ and koore koore ‘young emu’. Whilst there are numerous animals, birds and insects encoded with reduplicated lexemes, there are no other instances at all like this. Given Wyatt’s somewhat different spelling (korre vs. koore) perhaps they are in fact unrelated. The phonemic representation /kardi/ ‘emu’ is known from T&S, other sources and the closely related Nukunu language (Hercus 1992). The term for ‘young emu’ is more probably /kuri kuri/, with a mere chance resemblance between Wyatt’s spelling of these two words. Tukkutya is the word for ‘little’ in Kaurna. It is glossed in T&S as ‘small; little; child; infant’, though there are separate words ngarto ‘child’ and wakwakko ‘child; offspring’. As discussed earlier (note 4), there is probably no morphological or syntactic distinction between nouns and adjectives in Kaurna (as in many other Aboriginal languages) and indeed there are examples in the Kaurna corpus where tukkutya takes a nominal case suffix, as in (2):6 (2) Kaurna (T&S 1840, 24): Ngai ningka palta ngaityo tokuty(a)-­urlo. 1sg[abs] nearly throw[pst] 1sg.poss little-­poss ‘My little one has almost thrown at me.’ There are a number of examples where a reduced form of tukutya is prefixed to or compounded with other roots to derive terms, as in tukuangki (from tukutya ‘small’ + ngangki ‘female’). Tukuangki does not mean ‘small woman’ as we might expect, but rather ‘mother of many children; prolific woman’. Another term, tukupartapartana (from tukutya ‘small’ + parta ‘intercourse’ + partana ‘many; much’), has the same meaning ‘mother of many children’. Tukupitina ‘having no children; barren’ derives from tukkutya ‘small’ + pi (meaning unknown) + -­tina ‘without’ (the privative suffix). Similarly, tukkupurlaitya ‘having only two children’ derives from tukkutya ‘small’ plus purlaitye ‘two’. The last term recorded, tukkuparka, refers to a ‘grown-­up female; grown-­up daughter’ (the meaning of the second element, parka, is unknown). In none of these cases does the prefix tukku-­ / tuku-­/ refer to something small, but rather it refers to the mother of little ones or perhaps a woman capable of bearing children. So the use of tuku-­ ‘small’ in these examples plays no role in EM. In the reclaimed or revived Kaurna language, however, the prefixed compound element tuku-­has been used with a diminutive meaning in tukurru ‘cup’ (from tuku-­ ‘small’ + kurru ‘container’), tukammi ‘saucer’ (from tuku ‘small’+ tammiammi ‘plate’), tukuwingkura ‘microwave (oven)’ (from tuku-­‘small’ + wingkura ‘wave’), tukindo ‘hour’ (from tuku ‘small’ + tindo ‘day; time’),7 but it would appear that this usage is not a feature of the traditional language, and as these forms are lexicalised they are not a feature of EM.

Rob Amery

426

This all adds up to a complete absence of any kind of diminutive morphology in Kaurna, though there are restrictive -­ndi /-­nti/ and -­anda /-­anta/ suffixes. The restrictive -­ndi suffix is discussed underneath the pronoun chart by T&S (1840, 8) with examples such as Ngattondi wappeota ‘I alone (or myself) will do it’, whilst the vocabulary lists kumandi ‘only one’. In examples (3) and (4) it occurs attached to numerals: (3)

Kaurna (T&S 1840, 68): Warti-­ngga ngai wandi kuma-­rlukko-­ndi. between-­loc 1sg[abs] slept one-­time-­only ‘Between here and there I slept only once.’

(4)

Kaurna (T&S 1840, 69): Paini-­ngga purlaitye-­ndi meyu-­rla tittappi. two-­only man-­du hang[pst] long.ago-­loc ‘Formerly, only two men have been hanged.’ Kaurna does, however, appear to have an augmentative suffix /-­intyarla ~ -­inyarla/. T&S (1840, 7) make the following observation under the subheading ‘Comparison’: Hitherto, only one degree of comparison is known which may be viewed as comparative or superlative; it is formed by adding the termination -­intyerla, or -­inyerla, to the positive; as, karra, high, karraintyerla, higher, or very high; yakki, deep, yakkintyerla, deeper. The reduplicative form imparts intensity to the original meaning, or probably places it in the superlative.

Whilst T&S view this suffix as a comparative, there are no examples in the corpus of the form ‘X is Y-­er than Z.’ I think it more likely that /-­intyarla ~ -­inyarla/ could well be a true augmentative and not contextually determined as comparison is. In the sketch grammar, this -­intyerla /-­intyarla/ ~ -­inyerla /-­inyarla/ suffix is listed under adjectives. However, in the vocabulary there is an example of this suffix being affixed to a verb, though T&S (1840, 35) seem to force it into their adjective word class. Note: Padlondi v.n. to die; v.a. to desire; wish; long; as mai padlonend’ai, I desire food; kuri padlond’ai, I wish to have a dance. Padlonintyerla, adj. comp. strongly desiring; longing.

There are several other morphemes with an augmentative kind of meaning, though as a reviewer points out, these are not primarily augmentatives since the evaluation is of the person or thing (head noun), not the property qualifying them to which the suffix is attached. So any augmentative meaning here is a secondary nuance. Note the suffixes -­butto /-­purtu/ ‘full of’ and -­partanna /-­partana/ ‘many; much; all’: Butto, an affix, denoting that a person or thing is in a great measure possessed of what the word to which it is affixed signifies; as turlabutto, wrathful; angry; yertabutto, full of earth; dirty. (T&S 1840, 4) Partanna, adj. many; much; all; an affix denoting that a person is to a great extent, or in a high degree, possessed of what the word to which it is affixed signifies; as mantapartanna, full of lies; lying etcetera. (T&S 1840, 38)

Kaurna 427 Whilst Kaurna has no morphology for ‘good’ or ‘bad’ as such, it does have suffixes -­mpi ‘desirous of’, -­ marngutta ‘desirous of getting something’ (probably derived from marngu ‘envy’ + -­ta ‘fut’) and -­wādli /-­waadli/ ‘disliking’. Note the following examples: Maiimpi ‘in want of food’ (from mai ‘food’) Paru-­marngutta ‘desirous of getting meat’ (from paru ‘meat’) Mettillittillawādli ‘disliking/hating thieves’ (from mettillittilla ‘thief’) Puiyowādli ‘disliking smoke/cigarettes’ (from puiyo ‘smoke’) Marrawādli ‘liberal; generous’ (from marra ‘hand’) Maiwādli ‘liberal in giving food; generous’ (from mai ‘food’) Wādli is used both as an adjective preceding the noun, as in wādli marra ‘stingy’, and as a suffix, as above, where it has the opposite meaning. If these suffixes, -­mpi, -­marngutta and – wādli, really are making a judgement about the desirability of the object to which they are attached then these are instances of EM. Reduplication is used in many languages to encode diminutives or augmentatives. Reduplication is common in Aboriginal languages and Kaurna is no exception. In this language, reduplication takes many forms and serves many functions. Amongst the functions of reduplication in Kaurna verbs is that of durative, iterative or habitual notions. Teichelmann (1858, 2) claims that ‘intensive or iterative verbs [are] formed by redoubling the root of the verb’, and several examples appear in TMs: kumba kumbandi, to go away forever; whereas the simplex leaves the possibility of returning, the reduplication expresses alway intensity of the simplex (TMs 1857, 23) kutteni, again, once more; kutte kutteni, repeatedly (TMs 1857, 27) metti mettindi, intens. to steal. (TMs 1857, 37)

It is unclear whether this reduplication is derivational, producing a new lexeme with a related meaning, or whether reduplication here is truly an instance of EM. It will be forever impossible to tell given the paucity of examples, the lack of context and the inability to check with native speakers. So reduplication in verbs appears to have an augmentative function, whilst there is no hint of a diminutive in Kaurna verbs. There are also examples in the Kaurna corpus of reduplication of time words having an augmentative function. Consider the following time words: bukki ‘formerly; a long time ago’ bukkibukki ‘intensively used; a very long time ago’ The two augmentatives, one encoded with the suffix -­intya or -­intyarlo /-­intyarlu/ and the other by reduplication, seem to work together to achieve an even greater degree of augmentation. Teichelmann (1857) provides the following: painingga8 ‘some time ago; formerly; in time past; long ago’ paininggaintyarlo9 ‘more in the past than painingga; of a remote time; ancient’ painipainingga ‘more distant than paininggaintyarlo’

428

Rob Amery

A number of time words are documented with this augmentative form, as in: tarkarrintyerlo ‘a still longer time hence’ (tarkarri ‘in future, at a later period; a long time hence’) kurlaintyerlo ‘later; latest’ (kurla ‘afterwards’) bukkintyerlo ‘of a very remote time’ (presumably a more remote time than bukki)

3 Conclusion The Kaurna language from the Adelaide Plains in South Australia appears to have little by way of EM. Evaluative notions are encoded lexically in the main. The records of the Kaurna language point to the existence of a productive augmentative suffix, as well as reduplication being used on time words and verbs with an augmentative function. Diminutives seem to be totally absent, though there is a restrictive suffix used mostly on numerals in the extant examples. There is no morphology for ‘good’ or ‘bad’ per se, though there are suffixes for notions of ‘liking’, ‘desiring’ and ‘disliking’. EM is an important aspect of morphology for consideration and investigation. Had we undertaken a more thorough investigation of the Kaurna corpus for the existence of diminutive notions some years ago, we probably would not have developed new compounds for new terms that have used tuku ‘small’ with a diminutive meaning in these compounds, as such usage appears to run counter to the semantics of compounding with tuku ‘small’ in the historical corpus.

Notes 1. Nunga [nʌŋgə] is the term that Aboriginal people in southern South Australia use to refer to themselves, and Nunga English is the distinctive variety of Aboriginal English spoken in this region. The term Nunga seems to originate from the far west of the state, where nyangga means ‘Aboriginal person’. 2. Pama-­Nyungan languages are spread across most of Australia. The name derives from nyungar ‘man; person’ at the south-­western extremity and pama ‘man; person’ in Cape York at the other extremity. 3. Throughout this chapter, historical spellings appear in italics. Reconstructed phonemic representation appears in non-­italics between slash brackets / / where required. A near-­phonemic revised spelling system was adopted in 2010 but will not be used within this chapter. 4. Many Australian Aboriginal languages make no formal distinction between adjectives and nouns (see Dixon 1982), and this is probably also true of Kaurna, where semantic adjectives appear to take nominal case suffixes. 5. Kaurna pronouns make distinctions between sg, du and pl and a range of case forms, but there are no gender distinctions. Note pa ’3sg’ (that is, ‘he, she, it’). 6. Here we see tukkutya ‘little’ taking the ergative suffix and behaving like any other nominal (see note 4). 7. Tukindo ‘hour’ was devised in 1996 during the course of Amery’s PhD study of Kaurna language revival (Amery 1996, 12). The terms tukurru ‘cup’, tukammi ‘saucer’ and tukuwingkura ‘microwave’ were introduced into Reclaimed Kaurna during the course of teaching the language during the 1990s and early 2000s. They certainly appear in the pilot edition of the Kaurna Learner’s Guide (Amery and Simpson et al. 2007, 64)

Kaurna 429 8. Whilst -­ngga /-­ngka/ is the locative suffix, its appearance in painingga appears to have been incorporated into the root, though undoubtedly it was an ancient temporal marker here. *Paini is not documented and painingga appears to be a simple counterpart of bukki with parallel morphology. 9. The suffix -­rlo /-­rlu/ is the productive temporal marker and is homophonous with the ergative and instrumental suffixes, a pattern which is common in Aboriginal languages.

17.5 Rembarrnga

Adam Saulwick

1 Introduction Rembarrnga1 attests morphological processes which are used to express some evaluation, namely diminution, prototypicality, intensification and amelioration. Other evaluative functions (such as augmentation) are expressed with analytic constructions or regular polysynthetic morphological verbal-­word-­formation processes. The language does not have a subsystem of its morphology dedicated to evaluation (Scalise 1984). Rather, it makes use of regular language-­specific morphosyntactic strategies to express various types of evaluation. These include free lexemes (not discussed here, as analytic forms are beyond the scope of this volume), verb stems with inflectional affixes, and adverbials (both incorporated and free). Rembarrnga is a highly endangered, polysynthetic language located in central Arnhem Land, in the Northern Territory of Australia. The number of speakers is perhaps less than fifty. Linguo-­genetically it has been identified as a member of the eastern branch of the Gunwinyguan (GN) family (Non-­Pama-­Nyungan, Australian), more closely related to Ngandi and Ngalakgan than to the central GN languages (Bininj Gun-­wok or Dalabon) or the western GN languages (Warray and Jawoyn) (Evans 2003). This chapter is structured as follows. Section 2 gives a brief introduction to Rembarrnga grammar. Section 3 describes evaluative strategies within the morphology which cover diminution, approximation, augmentation, prototypicality and intensification. There is also a section (3.5) on positive affect, which is included here as it can be analysed as ameliorative shift (see Introduction to this volume). Section 4 summarises the findings.

2  Brief description of grammar Typologically, Rembarrnga attests polysynthetic verbal words with templatic structure and agglutinating morphology. It is head marking for person, number, and tense, mood and aspect (TMA), and dependent marking for case. Rembarrnga is prototypical of a GN language in the sense that it displays a high degree of morphosyntactic features common to GN languages. These include complex verb morphology with obligatory pronominal argument prefix slots for up to two arguments, the incorporation of nominals,2 verbs and adverbials into verbal words, and lexicalised noun– verb compounding. There are also a number of valence-­altering processes effected through the verbal morphology, including benefactive and comitative applicative prefixes and causative and reflexive/reciprocal suffixes. Major TMA distinctions are encoded through inflectional suffixes. Nominal incorporation appears to be somewhat more productive in

Rembarrnga 431

−3 −2 −1

0

1

2

3

4

5

6

7

8

9

Incorporated nominal 1

Verb stem(s) 1 prepound(s)

Stem

Thematic, causativizer, tranzitivizer

Stem

Reflexive/reciprocal

Progressive

TMA

Number

Particle

Case

Comitative applicative

−4

Incorporated adverb

Number

Incorporated adverb

Subordinate

−8 −7 −6 −5

Number

Pronominal subject/object

−9

Benefactive applicative

Table 17.5.1. Morpheme slots in the verbal word in Rembarrnga

Rembarrnga than in some other GN languages, as it permits incorporation of a wide range of nominals constrained by an animacy hierarchy. External to the verbal word, Rembarrnga morphosyntax attests ergative/absolutive aligned case marking with local and a small set of other cases, including temporal and proprietive cases. There is also residual noun class morphology. Word order has a discourse function rather than signalling grammatical relations. For detailed descriptions of the grammar see McKay (1975; 2011) and Saulwick (2003a). The polysynthetic verbal word is the nucleus of sentential expressions. This is constructed of a verb root with obligatory and optional morphemes. Obligatory components include the verb stem, pronominal prefixes and TMA suffixes. Table 17.5.1 illustrates the template with obligatory slots shaded in grey. For instance, example (12) below shows only obligatory slots −9, 0 and 6 filled. A plus symbol ‘+’ in slots −2 and 0 indicates that these slots may be internally complex, and can be filled by more than one element of the type designated. Obligatory pronominal argument prefixes register up to two arguments. Verb stems can be identified for their valency on the basis of whether they take monovalent or bivalent prefixes.3 Minimally, a verbal word is as exemplified in (12) below, where the stem is the 1-­place predicate nu ‘sit/be’.4

3  Evaluative strategies I adopt the notion that ‘[e]valuation is a mental process during which objects of extra-­ linguistic reality are assessed from the point of view of quantity (big vs. small) and quality (good, bad, nice, nasty etc.)’ (Körtvélyessy, forthcoming). Körtvélyessy points out the highly subjective nature of evaluation. In this brief survey of the evaluative in Rembarrnga I focus on speakers’ assertions or interrogation of subjective judgement with regard to size (diminution, augmentation) and quality (prototypicality, intensification and positive evaluation). Grandi and Körtvélyessy (this volume) articulate two definitional criteria of an ‘evaluative construction’: one semantic (assignment of a non-­standard value) and one formal (the explicit expression of the standard – via a lexically autonomous form – and an evaluative mark). From the point of view of Grandi’s (2011b) quadripartite typology, Rembarrnga falls into category Type A, as it attests diminutive and the absence of compliant augmentative morphology as defined by Grandi and Körtvélyessy. Nonetheless, the data presented here suggest that evaluative morphology in Rembarrnga

Adam Saulwick

432

is not a distinct morphological category and, as we will see, the language’s polysynthetic genius causes a categorisation problem under a strict interpretation of the definitional criteria. 3.1 Diminution Rembarrnga has two diminutives, ganya(ng)h and dakku, both of which prototypically express diminution in size.5 Although the former can occasionally occur as an independent form, it is typically a bound morpheme, as in (1), which shows the core diminutive sense:6 (1) Nyuwarr-­ganyah barr-­ma-­nginy barr-­giya-­bimbu-­niny. Perchlet-­ dim 3a>3-­pick.up-­pst.hab 3a>3-­nose-­draw-­pst.hab ‘They used to draw (a mark) on the nose of the little perchlets.’ (S2–92–060) Its denotation also extends to intensification. For instance, adding ganya(ng)h to a lexical diminutive ‘short’ intensifies the diminution: (2)



Nga-­rdarl-­miny garra gorloh-­na-­ganyangh. 1>3-­ remove.bark-­pst.pfv up be.short-­ nml-­dim ‘I removed a really short piece of bark (that was) up high.’ (S2–92–078)

Ganya(ng)h can also express diminution in quantity with additional negative emotive evaluation. In (3), suffixed to jarng ‘bad’, it not only expresses ‘insufficient’ quantity but also carries the speaker’s strong disapproval: (3)

Mœlak mani ralk yirri-­mangœ~ma-­nginy; jarng-­na-­ganyangh neg money big rel1a>3-­rdp-­get-­pst.cont bad-­ nml-­dim ‘We weren’t getting much money; a miserly amount.’ (S2–102–053 & 054)

Ganya(ng)h does not occur frequently in my corpus. However, this may simply be a data anomaly due to context dependent frequency, as observed by Ponsonnet and Evans (this volume) for Dalabon. The word dakku has a range of senses depending on discourse context and morphosyntactic environment. Unlike ganya(ng)h it is not a morphological diminutive and so is beyond the scope of this volume. However, it occurs frequently in discourse and so is briefly discussed here when occurring in a morphological derivation. As a nominal (suffixed with the nominal marker -­na) it typically means (female or male) ‘child’.7 When denominalised it can express a diminution in quantity, as in (4). (4) ø-­mærmæ-­dakku-­na-­wæ-­tti-­ny. pst.3-­eye-­small-­nml-­dnmls-­rr-­pst.pfv ‘He squinted.’ or ‘He made his eyes smaller.’ (Lit. ‘He made himself eye-­ smaller.’) (S4–16–038) This extension aligns with the well-­worn cross-­linguistic pathway (Jurafsky 1996) attesting to terms for ‘child’ as a source for diminutives. Dakku can be incorporated into stative stems taking a temporal locative suffix to predicate a stage of life:

Rembarrnga 433 (5) Murdi ngan-­bak-­yi-­nœh-­mœrn nga-­dakku-­ni-­ji nginda-­gah .  .  . FF 3>1-­ ben-­say-­stm-­pst.rem 1-­child-­sv.pst-­tloc 1crdl.pro-­all ‘When I was a child/young/little, my father’s father said to me . . .’ (S2–31–015a) As pointed out in note 6, dakku can be construed as either ‘child’ or a ‘small thing’. When denoting the former it may be diminutivised with -­ganya(ng)h, as in (6). Da-­yappa-­ngœnœ ø-­ngarriny-­bordoh-­mœ-­tti-­ny-­yarrpparrah 2>3-­sister-­1mn.dat.pro pst3-­hand-­put-­stm-­rr-­pst.pfv-­1ua.dat.pro dakku-­ganyah-­ngœnœ-­hgan. child-­ dim-­1mn.dat.pro-­dat ‘My sister put her hand on me for my little child.’ (S2–8–052) (6)

3.2 Augmentation Augmentation in size, number and quality is expressed with two independent nominals: ralk and garli. Ralk appears to be specialised to size, as in (7), while garli can refer to both size and intensity, as in (16) below. (7) Nœnda nga-­walang-­ralk-­miyanœ-­yimany. that 1-­ seq-­big-­inch.stm-­prog.pst ‘Then I was getting bigger.’ (S3–60–058) Their analytic forms prohibit further discussion here. However, it is worth stating that the conceptual category of augmentation is expressed via Rembarrnga’s regular polysynthetic morphological processes, such as inchoativisation of the stem, as in (7). 3.3 Prototypicality Prototypicality can be expressed with a bound morpheme -­jininy, glossed ‘real’ (8), and the incorporated adverbial gakku-­, glossed ‘properly’ or ‘completely’ (9): Len yarr-­ritj-­ø len-­jininy gurbæt-­gah bamurru-­hgah. ground.sugarbag 1a>3-­seek-­npst gnd.sugarbag-­real antbed-­all rotten.log-­ all ‘We look around for real ground honey in an antbed in a rotten log.’ (S1–114–065) (8)

(9) Bœnda ga-­du-­ru, wurru there npst3-­stand.intr-­npst but mœlak ge-­gakku-­yarrart-­rta many. neg sbjv.npst3-­properly-­grow.up-­fut not.yet ‘It is there but it has not yet properly grown up.’ (S3–68–206) Neither of these morphemes occurs as an independent word.

Adam Saulwick

434 3.4  Intensification

Reduplication of verb stems is predominantly used to encode progressive aspect. Reduplication of nominals can express augmentation in number, e.g. dakku~dakku ‘children’. However, reduplication of the diminutive serves to intensify: (10) Mærmæ-­na ø-­yorlop-­miny garra-­wala seed-­ nml pst3-­come.out-­pst.pfv on.top-­ abl walang nanh dakku-­na-­ganya~ganya. nml-­dim-­int and this small-­ ‘The seed came out from the top (of the tree) and all the teeny-­weeny (saplings sprouted).’ (S2–92–094) Reduplication of adjectivals can express augmentation in quality, as in (11). (11) Nattœ nganapparru boy urlah~urlah, ngarr-­gurrhwarr-­a. that buffalo body good-­ int 12>3-­shoot-­fut ‘That buffalo’s got a really healthy body, we’ll shoot him’ (e.g. ‘really good body’). (S4–97–035) 3.5 Amelioration Rembarrnga has a nominal with the form (w)urlah, glossed ‘good’, which is used to express a speaker’s generalised positive evaluation of an entity or a proposition. Although, under a strict interpretation of Grandi and Körtvélyessy’s definition of evaluative morphology, it could be considered a marginal case, it is discussed here, as speakers use it to encode a subjective judgement, representing ‘a shift towards the positive end of the qualitative axis’ (see Grandi and Körtvélyessy, this volume). It can be used adjectivally, predicatively and as a temporal modifier. In (12), the external positive evaluator (w)urlah ‘good’ modifies the verb’s only argument. (12) Wurlah ga-­nu-­ra? good npst3-­sit-­npst ‘Is she well?’ (S4–138–027) With 2-­ place predicates, such as dom ‘drink’ in (13), the modifier (w)urlah has scope over the object argument. In translation this is rendered as a type of secondary predication: (13)



Woh, da-­dom-­ma, urlah. yes 2>3-­drink-­fut good ‘Yes, you (can) drink it, (it’s) OK.’ (S2–130–016)

Example (14) shows the verbal word with the nominal bim ‘painting’ incorporated into the transitive stem bu ‘hit’. The bivalent pronominal prefix nga-­ encodes 1st person singular acting on 3rd person singular.8 Reduplication of (w)urlah intensifies the evaluation. Note the ambiguity between adjectival and adverbial interpretations, for example, ‘do a good x’ and ‘do x well’. This may arise from the close semantic relationship between

Rembarrnga 435 these senses, where a good or bad product may be expected to result from a well or badly executed process. (14) Urlah~urlah nga-­walang-­bim-­bu-­ni-­yimany. good-­ int 1>3-­ seq-­painting-­hit-­stm-­prog ‘Then I was painting really well.’ (Lit. ‘I was hitting it really good.’) (S3–60–016) In (15) the adverbial sense is more clearly expressed. (15)

Wurlah yarr-­jarda-­ø. good 1a>3-­hand.twirl.firestick-­npst ‘(Then) we could twirl it well.’ (S2–83–031)

We have seen in (12)–(13) (w)urlah external to the verbal morphology with scope over morphologically bound internally referenced arguments. With the inchoative suffix -­miny, it forms a predicate expressing the speaker’s subjective assertion of ‘doing good’; a derived ameliorative predicate. In (16) it occurs with the nominal garli ‘big’; a more literal translation could be ‘they began to do good (with) their big (power)’. (16) Barr-­wurlah-­miny-­pparrah garli-­barrpparrah. 3augn>3-­good-­inch.pst.pfv-­ua big -­3ua.poss.pro ‘It was time for the two of them to use their special power.’ (S1–131–028) When suffixed to (w)urlah, the causativiser -­wa derives a positive evaluation predicate. The context of the discussion in (17) was Rembarrnga people’s attitudes towards adornment scars. The causative sense ‘make something good’ is used to assert the speaker’s subjective judgement of the visual appeal of an adornment scar. (17) Borlitj-­yih barr-­wurlah-­wa-­niny bi barr-­wurlah-­waniny. adornment.scar-­ ins 3>3-­good-­caus-­pst person 3>3-­good-­caus-­pst ‘An adornment scar made a person (look) attractive.’ (S3–101–007) (W)urlah can also take a pronominal prefix to index an entity. In this construction, the speaker offers (18) or seeks (19) an evaluation of personal well-­being: (18)

Malak ngi-­urlah nga-­ngawa-­ttœ-­n. neg 1rel-­good 1-­listen-­rr-­pst.pfv ‘I don’t feel well, I’m crook.’ (S2–130–002)

(19)

Muju nginy-­urlah? intg.tag 2-­good ‘Are you all right?’ (S2–31–020)

With a temporal suffix -­gun, the evaluator can be used as a temporal satellite: (20)



Nga-­ni bœnda-­ja urlah-­gun. 1-­sit.sv.npst there-­ loc good-­last ‘I stayed there for a good while.’ (S1–60–036)

Adam Saulwick

436 Table 17.5.2. Evaluative forms and semantics with (w)urlah in Rembarrnga Form

Ameliorative semantics

Translation

Example

Bare external form 1Pronominal prefix External pred. scope 1Inchoativiser 1Causativiser 1Temporal

aml(x)

‘something good’ ‘you good’ ‘x do y well’ ‘start to do something well’ ‘make something good’ ‘stay for (in)appropriate time’

(12) (19) (14) (16) (17) (20)

aml(x)

aml(pred(x,

y))

(x, y)) (x, y)) aml(time_of(x)), pred (x) aml(inch.aml.pred

aml(caus.aml.pred

Table 17.5.2 gives an overview of the morphosyntactic forms with corresponding semantics as discussed above. In sum, a speaker’s subjective positive evaluation may be expressed through various morphological means. Semantically, it may be rendered as a 1-­place predication and additionally with scope over an argument of another predicate. Inchoative and causative suffixes derive second-­order predicates expressing amelioration. With a temporal marker, speakers may express their subjective attitude with respect to ‘appropriate’ duration. Thus, (w)urlah functions as a generalised marker expressing a shift towards the positive end of the qualitative perspective which, depending on the morphosyntactic context, has adjectival, second-­order predicative or temporal senses. Given Rembarrnga’s polysynthetic nature, the data described put (w)urlah squarely within the language’s morphology. However, this poses a problem for the second formal condition of Grandi and Körtvélyessy’s evaluative construction definition (see Introduction) in that there is no explicit expression of the standard against which the evaluative mark can be compared.9 Instead, regular polysynthetic verbal-­word-­formation processes are used. Perhaps an alternative is to reformulate this condition to allow for language-­internal criteria to identify what falls within morphology.

4 Conclusion In this brief survey of evaluation in Rembarrnga I have shown that the language makes use of a variety of strategies to express certain evaluative senses, namely diminution, augmentation, prototypicality, intensification and amelioration. In Rembarrnga there is no distinct category of the morphology dedicated to evaluation. However, counter to the claim of Scalise (1984) that language has a separate evaluative component, the strategies described in this chapter make use of Rembarrnga’s regular morphosyntactic processes also utilised for other functions. This is analogous to the analysis of Bauer (1997) for other languages. Furthermore, Stump (1993) proposed a type of middle evaluative morphological process between word-­formation rules and inflectional rules. The Rembarrnga data exemplified do not neatly align with Stump’s model, as the evaluative strategies described here are dispersed across a range of morphosyntactic processes. Formally, some strategies make use of specialised suffixes, or lexemes functioning as free modifiers, some with scope over arguments internally referenced within the verbal word, while others use verb stems with derivational affixes or incorporated nominals, as well as adverbial and temporal elements. In terms of the list of semantic functions performed by evaluative morphology as articulated by Grandi and Körtvélyessy (see Introduction to this volume), Rembarrnga attests

Rembarrnga 437 diminution, prototypicality, intensification and amelioration. We have not seen examples of morphological processes expressing age variation, approximation, endearment, hypocoristics or social position. Augmentation is expressed through analytic constructions or regular polysynthetic morphology, thereby posing a possible classification conundrum.

Notes 1. In Ethnologue (Lewis, Simons and Fennig 2014) the language is spelled Rembarunga, but in the standard orthography used in contemporary printed material (see Saulwick 2003b) and in communities where the language is spoken or referenced, it is spelled Rembarrnga. 2. As with many Australian languages, nouns and adjectives form a superclass of ‘nominals’. 3. In parts of the pronominal prefix paradigm, mono-­ and bivalent forms are homophonous. However, the corresponding 2nd person singular forms are not and are therefore used as a diagnostic for verb stem valence (see Saulwick 2003a; 2007). 4. Exceptions to minimal verbal morpheme structure are commands, e.g. dom-­ma! drink-­ imp ‘Drink!’ 5. Both may occur as free-­standing nominal modifiers, as in (i) and (ii). Numbers in parentheses starting ‘(S-­’ refer to location in the corpus text. (i) Bi ganyangh, ganyangh bi. man dim dim man ‘A little man.’ (S2–101–048) (ii) Dakku bu-­yih barr-­bu-­wa walang ga-­ru-­n. child other-­erg 3l>3-­hit-­pst.pfv seq npst3-­cry-­npst ‘Another child hit her/him and (now) (s)he is crying.’ (S2–68–002) 6. I use the symbol ‘>’ to represent ‘acting on’ relation in the bivalent pronominal prefix gloss. Note that lower animates, such as perchlets in (1), do not trigger number agreement. 7. It can be used as a free adjectival nominal modifier meaning ‘small’, ‘little’ or ‘minor’. 8. There are no examples in my corpus of (w)urlah having scope over the pronominally referenced argument with a 3-­place predicate. So scope assignment for these predicates is unknown but likely to be based on context. For instance, (iii) shows adjectival scope over the theme which is not pronominally referenced. (iii)

Nanh-­ma yemetj ngan-­dehwa-­ø. this-­foc wrong 3>1-­give-­pst ‘He gave me the wrong picture.’ (S4–82–046)

9. Indeed, there are a number of other morphemes in Rembarrnga that could be analysed as expressing an evaluative function but are not discussed as they do not meet the second criterion.

17.6 Warlpiri

Margit Bowler

1 Introduction Warlpiri is a Pama-­Nyungan language spoken by approximately 3,000 people in central Australia.1 It has been studied by a number of linguists as a result of its typologically unusual features, including a highly flexible word order, discontinuous constituents and extensive pro-­drop. Warlpiri has no grammatical gender, and is a strictly suffixing language. Warlpiri is split-­ergative; its flexible word order is licensed through the use of ergative-­absolutive case markers and nominative-­accusative subject/non-­subject agreement clitics. Several authors (e.g. Bittner and Hale 1995; Simpson 1991) argue that Warlpiri distinguishes between only two major lexical categories: nouns and verbs. Adjectival expressions are subsumed within the nominal category and host case marking and other nominal morphology. Warlpiri verbs in turn consist of approximately 120 roots, which can combine with an open class of adverb-­like ‘preverbs’ or ‘coverbs’ (Nash 1982). Warlpiri nominals can occur as full arguments of the predicate (their expected ‘nominal’ usage) or combine with other nominals (their predicational, ‘adjectival’ usage). Bittner and Hale (1995) rank the Warlpiri nominals from the most likely to express the arguments of the main predicate to the most likely to occur as secondary predicates. In this chapter, I will follow Bittner and Hale in making reference to Warlpiri common nouns (the most likely to occur as full arguments of verbs) and to more adjectival expressions of cardinality, quality, psychological state and so on (the most likely to occur as secondary predicates). Warlpiri evaluative morphemes are nominal suffixes, and generally precede case marking:2 (1) a. pirli-­pardu-­rla rock-­ dim-­loc b. *pirli-­ngka-­pardu rock-­ loc-­dim ‘on a little rock’ (MB:01–10) Many evaluative morphemes are now considered ‘hard language’, or vocabulary used only by older speakers of the language. Many younger speakers do not use all of the ­following morphemes any more, and not all younger speakers associate these morphemes with their evaluative usage.

Warlpiri 439

2  Augmentative morphology I will begin with a discussion of Warlpiri augmentative morphology, since augmentatives are more common in Warlpiri than diminutives. Warlpiri possesses a number of augmentative evaluative suffixes. Several of these evaluative suffixes are derived from nouns. 2.1  -­nyayirni -­Nyayirni is the most commonly used Warlpiri augmentative suffix, used frequently by speakers of all ages. It typically expresses a shift along the quantitative scale. However, -­nyayirni can also be used to express that an item is proper, true or prototypical (for that item). The reading that arises when -­nyayirni suffixes onto a nominal is dependent on whether it is a common noun or an adjectival nominal. It is far more common for -­nyayirni to co-­occur with adjectival nominals. This use of -­nyayirni is very well attested. (2)

lalypa-­nyayirni flat-­aug ‘very flat’ (KRN, PRN)

(3) ngurrju-­nyayirni good-­ aug ‘very good’ (WDP 2000) The combination of -­nyayirni and a common noun refers to a prototypical, ‘real’, ‘true’ form of the noun. Jurafsky argues that ‘the metaphor “category centrality is size” links central or prototypical members of a category to large size, and peripheral or ­marginal members of a category to small size’ (Jurafsky 1996, 548). The usage of Warlpiri augmentative morphology to indicate prototypicality follows from this; cf. the usage of augmentatives to express prototypicality in Spanish and Navajo (Jurafsky 1996, 549). (4) ngapa-­nyayirni water-­ aug ‘fresh water’ (MB:01–10) (5) warna-­nyayirni snake-­ aug ‘poisonous snake’ (WDP 2000) The interpretation of -­nyayirni in these contexts requires some contextual knowledge about Warlpiri culture and the central Australian environment. In central Australia, the characteristic of being poisonous or dangerous is a highly salient feature of many indigenous snakes. Warna-­nyayirni ‘snake-­aug’ therefore results in the reading ‘poisonous snake’, the prototypical form of the central Australian snake. When -­nyayirni combines with numerals, it denotes ‘exactness’. This is another ­realisation of the ‘prototypical’ readings given above.

Margit Bowler

440 (6)

rdakapala–nyayirni five-­aug ‘exactly five’ (MB:02–11)

Dixon (2002) argues that many Australian languages distinguish between ‘generic’ and ‘specific’ semantic classes. -­Nyayirni is only able to combine with generic common nouns; it cannot combine with specific nouns describing small semantic ranges, e.g. *yankirri-­nyayirni ‘emu-­ aug’, *marlu-­nyayirni ‘kangaroo-­ aug’. This feature of -­nyayirni could therefore provide a useful diagnostic as to whether a noun is generic or specific. 2.2 -­mpayi The augmentative suffix -­mpayi is rarely used by younger speakers today. This suffix is used specifically to quantitatively evaluate adjectival nominals describing the cardinal directions, ‘up’, ‘down’ and ‘tall’. (7) karlu-­mpayi aug west-­ ‘very far west’ (WDP 2000) (8) kirrirdi-­mpayi tall-­ aug ‘very tall’ (WDP 2000) -­Mpayi cannot combine with common nouns, or with adjectival nominals falling outside these semantic categories. This suffix is probably related to the directional enclitic =mpa ‘past’/‘across’. 2.3 -­katu When the augmentative morpheme -­katu combines with an adjectival nominal, it generally presupposes comparison between two entities (Laughren 1986). However, -­katu can also be used to express purely quantitative evaluation. Usage of -­katu is additionally frequently accompanied by a speaker’s positive attitude towards the item or quality being evaluated. (9) Yinya=ji yungka ngurrju-­katu yalyuyalyu turawuju=ju, Napangardi. there=top give.imp good-­aug red trouser=top Napangardi ‘Give that red pair of trousers to me, Napangardi, which is better.’ (WDP 2000) (10)

Pilykirr-­pakaka warlu yarrmalyka-­katu! dead.wood-­chop firewood worn.out-­aug ‘Chop off some wood that is really dry and will burn quickly.’ (WDP 2000)

When -­katu combines with a common noun, it can have a comparative reading ‘more’ in reference to the quantity of the noun. This is analogous to its comparative usage with adjectival nominals.

Warlpiri 441 (11)

 arlu ngula nyurdi-­manta yungu=rlupa M purrami karlarra warlu kangaroo that shoulder.imp aux.comp=1pl.sbj.incl cook.npst west firewood -­katu-­rla. -­ aug-­loc

‘Put that kangaroo on your shoulders so we can cook it over to the west where there is more firewood.’ (Napanangka, Y 1988)

-­Katu can also express positive qualitative evaluation. It can combine with a more heterogeneous class of words in these contexts, including pronouns (as in (12)). (12) Nyuntu-­katu-­rlu=ju jikily-­kijika pina-­katu-­rlu. aug-­erg=top do.properly.imp knowledgeable-­ aug-­erg 2-­ ‘It’s better if you deal with this paper as you know more about it.’ (WDP 2000) (13) Jilja-­katu-­rla yi=rlipa=jana wajili-­pinyi. sandhill-­ aug-­loc aux.comp=1pl.sbj.incl=3pl.nsbj chase.quickly.npst ‘We’d be better off chasing them in the sandhills.’ (H59:7.154–5) -­Katu can also be used to pick out an entity that the speaker has never seen before, or to pick out a unique entity from a set. This use of -­katu may be more frequent among eastern Warlpiri speakers.3 Example (15) shows the range of meanings that can arise from a single usage of -­katu. All of these readings are accompanied by positive ­qualitative evaluation; in central Australia, shadiness is a positive quality for trees to possess. (14)

nyampu-­katu this -­ aug ‘a unique/new item to the speaker’ (MB:02–11)

(15) wurrkali-­katu-­kurra bloodwood-­ aug-­all ‘to the shadier bloodwood tree’ ‘to the shady bloodwood tree’ ‘to the only bloodwood tree with shade’ (MB:02–11) This morpheme is probably historically related to the nominal katu ‘high up’. Many younger Warlpiri speakers do not use this evaluative morpheme, or understand its evaluative usage. Instead, -­katu is used only to mean ‘high up’. 2.4 -­karrikarri The evaluative suffix -­karrikarri presupposes comparison when it combines with ­adjectival nominals (Laughren 1986). This suffix does not combine with common nouns. (16) wiri-­karrikarri big-­ aug ‘older’ (Nakamarra, tape-­184)

Margit Bowler

442 (17) mata-­karrikarri slow-­ aug ‘slower’ (darby)

This evaluative suffix is probably related to the quantificational nominal karrikarri ‘large amount’, which can form discontinuous constituents with common nouns. 2.5  Augmentatives in combination with negation Both -­karrikarri and -­nyayirni can combine compositionally with the privative suffix -­wangu ‘without’.

(18)

pirriya-­wangu-­karrikarri cold-­without-­aug ‘a little less cold’ (David Nash, p.c.)

(19) yurnmi-­nyayirni-­wangu ripe-­ aug-­without ‘not completely ripe’ (WDP 2000) -­Nyayirni in combination with -­wangu can also result in the reading ‘special’ or ‘sacred’, possibly through negation of the ‘prototypical’ reading arising from -­nyayirni. (20) yimi-­nyayirni-­wangu story-­ aug-­without ‘a sacred story’/‘not the archetype of a story’ (Warlpiri Bible, trans. Swartz) 2.6  Augmentatives: a summary Warlpiri augmentative morphemes are primarily used to express evaluation along the quantitative scale. The suffixes -­katu and -­karrikarri can express comparison, and -­nyayirni can be used to indicate ‘prototypicality’ when combining with common nouns (cf. Jurafsky’s 1996 metaphor ‘category centrality is size’). Only the augmentative suffix -­katu can be used to express positive qualitative evaluation. The augmentative suffixes are restricted in their ability to combine with one another. -­Nyayirni cannot co-­occur with -­karrikarri, and -­mpayi cannot co-­occur with -­katu. When -­nyayirni co-­occurs with -­katu, it can either precede or follow the suffix. When -­nyayirni co-­occurs with -­mpayi, -­mpayi must precede -­nyayirni.

3  Diminutive morphology 3.1  -­pardu The diminutive suffix -­pardu is used frequently by Warlpiri speakers of all ages. This suffix is used for both quantitative and qualitative evaluation. Like -­nyayirni, -­pardu interacts with the semantics of the noun that it combines with. When -­pardu combines with a common noun, the resulting interpretation is a decrease in size of the nominal. However,

Warlpiri 443 many speakers prefer to use the adjectival nominal wita ‘small’ rather than -­pardu to evaluate the size of objects. (21) jurlpu-­pardu bird-­ dim ‘little bird’ (Hale59/60.FieldNotes) (22) jarntu-­pardu dog-­ dim ‘little dog’ (MB:01–10) When -­pardu combines with an adjectival noun expressing psychological state, quality or quantity, the resulting interpretation is a decrease in intensity of the properties associated with that noun. (23) wiri-­pardu big-­ dim ‘biggish’ (PPJ, 10/85) (24) wantiki-­pardu wide-­ dim ‘rather broad’ (ML 85) The diminutive -­pardu can also combine with temporal expressions. This is used to express approximation. (25) yangkurra-­pardu-­kari yesterday-­ dim-­other ‘a couple of days ago’ (MB:02–11) (26) ngaka-­pardu-­kari later-­ dim-­other ‘soon’/‘next time’ (MB:02–11) These uses of -­pardu support Jurafsky’s 1996 observation that diminutive morphology can indicate approximation, or the marginality of an item’s membership within a semantic class. This is also the source of the ‘attenuation’ reading that arises when -­pardu combines with adjectival nominals. -­Pardu is able to express both pejorative and affectionate readings when it is used for qualitative evaluation, though it is more common to use -­pardu to express sympathy or affection. This usage of -­pardu is restricted to humans. Warlpiri speakers ­generally reject this ‘affectionate’ usage towards inanimate objects or animals, even pets. (27) yapa-­pardu person-­ dim ‘friend’/‘family member’ (MB:02–11)

444

Margit Bowler

(28)  Kaparli-­pardu ngajuku-­palangu . . . nganimpaku-­palangu kaparli-­pardu grandmother-­dim parent ancestor grandmother-­ dim wurlkumanu. old.woman ‘My old maternal grandmother . . . our dear old grandmother.’ (WDP 2000) -­Pardu can also have a pejorative meaning; however, this usage is not as common. (29) Nyuntu=ju=npa yukurdu-­pardu, wati-­wangu, yaparranji-­pardu. uninitiated.male-­dim man-­without child-­dim 2=top=2sg.sbj ‘You’re just a kid, not an initiated man, just a child.’ (WDP 2000) 3.2 -­pawu -­Pawu is a form of -­pardu conditioned by register. This diminutive morpheme is used when speaking to or about young children, and is a prominent feature of Warlpiri ‘baby talk’. Warlpiri baby talk is not the speech of Warlpiri children; instead, it is a distinct register used by adults when speaking to children. Warlpiri baby talk attempts to emulate children’s speech and is used to make it easier for children to understand older Warlpiri speakers. Features of the baby talk register are outlined in Laughren (1984). The use of -­pawu does not necessarily imply that the speaker is conveying affection or sympathy towards the evaluated item. Instead, -­pawu is used to express affection towards the addressee, typically a child. In addition to this qualitative function, it may be used for quantitative evaluation, as in (30)–(31). (30) yuwarli-­pawu house-­ dim ‘little house’/‘animal’s nest’ (MB:03–08) (31) wita-­pawu small-­ dim ‘small item’/‘small child/baby’ (MB:03–08) The term wita-­pawu ‘small-­dim’ has been lexicalised by many Warlpiri speakers to refer simply to a small item. This term is used frequently outside the Warlpiri baby talk register, and does not necessarily express any affection on the part of the speaker towards the item being evaluated. 3.3  Diminutive skin name forms Warlpiri culture has eight subsections, divided into two matrimoieties and two patrimoieties. These subsections are used to describe kin relations between individuals. Siblings share the same subsection name (‘skin name’), and skin names within a matrimoiety repeat every four generations. Each subsection has two distinct name forms; female skin names begin with N (e.g. Napangardi) and male skin names begin with J (e.g. Japangardi). Warlpiri therefore has a total of sixteen unique skin names, one male and one female name for each subsection. Warlpiri speakers frequently address one another by their skin name, rather than by their given name.

Warlpiri 445 Table 17.6.1. Warlpiri skin subsection names and their diminutive forms (MB:02–11; Hoogenraad and Laughren 2012) Basic Warlpiri skin name

Diminutive skin name form(s)

Nakamarra Jakamarra Nampijinpa Jampijinpa Nangala Jangala Napaljarri Japaljarri Napanangka Japanangka Napangardi Japangardi Napurrurla Jupurrurla Nungarrayi Jungarrayi

Nakarra, Wajala Jakarra Ngampardi, Ngampijakurdu, Nampiini Jampirlka, Jampiiti Nangali, Ngangkarla Jangkarli Ngalyirri Japalyi Ngamana, Nangka Janama Ngampayardi, Napangayi Jangari, Japayardi Napurru, Ngampurla Jupurru, Jurlama, Julayi Ngampukurlu, Nungayi Jukurtayi

The diminutive suffix -­pardu can be suffixed directly onto skin names. However, each Warlpiri skin name also has a unique diminutive form or forms associated with it, used to express sympathy or affection. Table 17.6.1 is not an exhaustive list of all of the possible skin name diminutives, since it is possible to create innovative diminutive forms in the course of conversation. Hoogenraad and Laughren (2012) provide a list of Warlpiri skin name diminutive forms. These skin name forms exhibit several cross-­ linguistic features of diminutives described by Jurafsky (1996), including the frequent use of fronted (labial) consonants (as in Wajala), high vowels (as in Nangali and Japalyi) and long vowels (as in Nampiini and Jampiiti). 3.4  Diminutives: a summary Warlpiri has a single diminutive morpheme with two separate forms conditioned by register. This evaluative morpheme -­pardu/-­pawu is associated with both quantitative and qualitative evaluation; it can be used to express both affectionate and (less commonly) pejorative attitudes. Cross-­linguistically, diminutives are often derived from the word for ‘child’ (Jurafsky 1996). -­Pardu, however, is not related to the Warlpiri word for child (kurdu), nor is it related to ‘child’ in any neighbouring languages.

4  Reduplication as evaluation Reduplication of adjectival nominals can express both augmentation and diminution along the quantitative scale SMALL ↔ BIG; cf. Grandi (2011b). However, reduplication of the evaluative suffixes is not possible, e.g. *-­nyayirni-­nyayirni, *-­katu-­katu and so on.

Margit Bowler

446 4.1  Reduplication as augmentation

Reduplication has a number of different functions in Warlpiri, including expressing plural number, repetition of an action, extended duration of an event, or distribution across a large spatial area. Reduplication of an adjectival nominal can indicate intensification of the properties associated with that nominal. (32)



wiri-­wiri big-­big ‘very big’ (KRN)

4.2  Reduplication as diminution Reduplication of an adjectival nominal can also be used to express attenuation of the properties associated with the nominal. (33)



ngapa wiri-­wiri water big-­big ‘a biggish water(hole)’ (HN:1160)

5 Conclusion Warlpiri evaluative morphemes are strictly suffixing, bound morphemes that typically combine only with nominals. This chapter has shown that Warlpiri evaluative morphemes are sensitive to the semantics of the nominals that they combine with. In particular, evaluative morphemes are sensitive to the distinction between common nouns and ‘adjectival’ nominals describing quantity, quality, psychological state and so on. This distinction is particularly relevant for the augmentative morphemes -­mpayi and -­nyayirni; the former can combine only with a subset of adjectival nominals, and the latter gives rise to very different meanings depending on the semantics of the nominal it combines with. Reduplication can also express both augmentation and diminution along the quantitative scale. Grandi (2011b, 7) notes that augmentatives are cross-­linguistically less common than diminutives. Warlpiri is unusual in this respect in that it has significantly more augmentatives than diminutives. However, many younger speakers no longer use all of the evaluative morphemes listed in this chapter. Several evaluative morphemes (e.g. -­katu, -­karrikarri) are related to Warlpiri nominals. Many younger Warlpiri speakers associate these morphemes with their non-­evaluative meanings, and no longer associate them with their evaluative interpretations.

Acknowledgements I would like to thank the Australian-­American Fulbright Commission for their support of my research in Australia. I would also like to thank Russell Schuh at UCLA for his advice on this project. My deepest thanks go to the Warlpiri speakers in Yuendumu who taught me about their language.

Warlpiri 447

Notes 1. The data in this paper comes from a combination of other linguists’ field notes on Warlpiri (Ken Hale, Mary Laughren, Stephen Swartz, Paddy Patrick Jangala, Jane Simpson and David Nash, among others), including a 2000 draft of the Warlpiri Dictionary Project (cited in this chapter as WDP 2000). The examples in this chapter taken from the Warlpiri corpus appear with citations referring to their location within the corpus. Additional data comes from my own fieldwork in Yuendumu, Northern Territory (March–August 2012) (cited in this chapter as MB). 2. Choice of the -­rla versus -­ngka form of the locative case marker, as in (1b), is conditioned by the number of syllables in the word to which the locative suffix attaches. 3. The speakers who provided me with this data all speak the central/western Warlpiri dialect. They reported that they would be able to understand this usage of -­katu as ‘unique;’ however, they additionally asserted that this use of -­katu would be more common in the eastern Warlpiri dialect. I have not yet confirmed this with speakers of the eastern Warlpiri dialect.

17.7  Yukulta and its Relatives Kayardild and Lardil

Erich Round

1 Introduction Yukulta, also known as Ganggalida, is an extinct member of the Southern branch of the non-­Pama-­Nyungan, Tangkic language family of north-­western Queensland, Australia. Key sources are a master’s dissertation and related sketch grammar by Sandra Keen (1972; 1983) plus around twenty hours of Keen’s field recordings, deposited at the Australian Institute of Aboriginal and Torres Straight Islander Studies (AIATSIS) in Canberra, Australia. These recordings are now undergoing transcription by the current author, and provide significant additional information about the language; however, since the recordings consist overwhelmingly of elicited sentence translations from English into Yukulta by a single speaker, it is very likely that some of Yukulta’s grammatical and lexical resources went unrecorded. Yukulta’s rich nominal and verbal inflectional system is strictly suffixing and dependent marking. Because inflection appears only at the end of stems, one can reliably distinguish polymorphemic stems, such as kuɲa-­mira-­ in (1a), from syntactic phrases built on multiple stems, such as kuɲa-­and mira-­in (1b). (1) Yukulta: a. kuɲa-­mira-­n̪t̪ a b. kuɲa-­n̪t̪ a mira-­n̪ t̪ a small-­ int-­dat small-­ dat good-­ dat ‘very small (dat)’ ‘small, good (dat)’ The inflectional system distinguishes just two word classes: nominals (including all words in an NP) and verbals, and thus for example there is no inflectional distinction between nouns and adjectives. However, it is possible to distinguish syntactically between sub-­ classes of nominals due to the strict word order in noun phrases, set out in (2).1 (2)

Word order of the Yukulta noun phrase: determiner/possessive NP – number – adjective(s) – head noun – modifier(s)2

Outside of the NP, word order is highly flexible. Most main clauses host a second-­position clitic complex which marks features of core arguments and up to one oblique, as well as tense, (ir)reality and transitivity. Subjects and objects follow a complex mix of ergative–absolutive and nominative–accusative alignment in main clauses, and subordinate clauses can be marked with up to two layers of ‘complementising case’ (Dench and Evans 1988) which indicate tense/aspect and co-­referentiality relationships between the matrix and subordinate clauses.

Yukulta and its Relatives Kayardild and Lardil 449 Since a practical orthography has not been established for Yukulta, forms will be transcribed phonemically, and in order to aid comparison between languages, phonemic transcriptions will also be used for Kayardild and Lardil. Words are presented in their citation forms, which can include overt, default inflection and the effects of phonological rules (on which see Hale 1973; Round 2009; 2011).

2  Evaluative morphology Yukulta possesses relatively little evaluative morphology (EM) that we know of, all of which is nominal or adjectival. Comparisons with Yukulta’s relatives Kayardild and Lardil are offered in Section 3. Kayardild possesses nominal, adjectival and verbal EM; Lardil possesses only verbal EM. 2.1  Intensifying and ameliorative -­mira on adjectival and nominal bases The nominal suffix -­mira is related etymologically to the root mira-­‘good’. Its most abundantly attested evaluative use, however, is not as an ameliorative marker with the sense ‘good’, but as an intensifying adjectival marker, ‘very’, as in (3). (3) Yukulta: a. kalkan-­ta ‘sick-­abs’ b. kuɲa-­ɻa ‘small-­abs’ c. paʈaŋu ‘big(abs)’ d. t̪ ali-­ja ‘laden-­abs’ e. wara-­ɻa ‘far-­ abs’

kalkan-­mira kuɲa-­mira paʈaŋu-­mira t̪ ali-­mira wara-­mira

‘very sick (abs)’ ‘very small (abs)’ ‘very big (abs)’ ‘heavily laden (abs)’ ‘very far (abs)’

In (4a), -­mira attaches to a nominal base ‘old man’ and yields an intensifying nominal stem ‘very old man’. In (4b) it attaches to the nominal base pijan-­‘swimmer’ (a nominalisation of pijac-­ ‘to swim’) and yields an ameliorative nominal stem ‘good swimmer’. These are the sole attestations I have of -­mira attaching to nominal bases. (4) Yukulta: a. palalaɲi ‘old man (abs)’ b. pijan-­ta ‘swimmer-­ abs’

palalaɲi-­mira ‘very old man (abs)’ pijan-­mira ‘good swimmer (abs)’

2.2 Intensifying -­mut̪a on one adjectival base The suffix -­mut̪ a is etymologically related to the root mut̪ a-­‘much, many, plenty’. Example (5) is the sole attestation of evaluative -­mut̪ a, where it attached to an adjectival stem and conveys an intensifying meaning. (5) Yukulta: kun̪ a-­ɻa

‘small-­abs’

kun̪ a-­mut̪ a

‘very small (abs)’

2.3 Ameliorative -­pata on kinship nouns Keen’s recordings are mostly of one speaker, Alice Gilbert. However, in a short recording session, Sydney Webber makes the sole attested use of -­paʈa, a suffix recognisable from

Erich Round

450

Kayardild (cf. Section3.1) as an ameliorative marker on kin terms. Alice Gilbert did not use -­paʈa in her many tokens of kin terms, and the suffix may have been dialectally restricted. In (6) the marker -­paʈa displaces the meaningless kinship suffix -­cu of ŋari-­cu ‘father’s mother’. (6) Yukulta: ŋari-­cu

‘father’s mother-­kin (abs)’

ŋari-­paʈa

‘dear granny (abs)’

2.4  A note on non-­evaluative forms containing -­mira and -­mut̪a In addition to their evaluative uses in Sections 2.1 and 2.2, the elements -­mira and -­mut̪ a also appear in Yukulta compounds, as in (7a,b). (7) Yukulta: a. maral-­ta b. cul-­ta

‘ear; hearing-­abs’ maral-­mira ‘bone-­abs’ cul-­mut̪ a

‘having good hearing (abs)’ ‘(of fish) bony (abs)’

In cases such as these, -­mira and -­mut̪ a should be interpreted not as evaluative suffixes, but as roots with their usual meanings, respectively ‘good’ and ‘much, many, plenty’. Like other n +adj compounds in the language, the complex stems in (7a, b) are adjectival and have the sense ‘having [n] which is [adj]’. By way of contrast, if -­mira and -­mut̪ a were acting as evaluative suffixes in (7a, b), then the complex stems should be nominal, just like the nominal roots to which -­mira and -­mut̪ a attach.

3  Comparison with related languages Yukulta is compared with its Southern Tangkic sister, Kayardild, in Section 3.1 and with its Northern Tangkic cousin, Lardil, in Section 3.2. 3.1  Kayardild (Southern Tangkic) Like Yukulta, Kayardild belongs to the Southern branch of Tangkic. Major sources on Kayardild are a comprehensive grammar by Evans (1995), formal analyses of phonology, morphology and syntax by Round (2009; 2013), and field recordings by Stephen Wurm in 1960, by Evans from the 1980s to the present, and by Round from 2005–2007. Like Yukulta, Kayardild employs the suffix -­mira as an intensifying evaluative ‘very’ on adjectival stems (Evans 1995, 194), as in (8). (8) Kayardild: a. cuŋarpa ‘big (nom)’ cuŋarpa-­mira ‘very big (nom)’ b. palar-­a ‘white-­nom’ palar-­mira ‘very white (nom)’ c. muŋuru-­wa ‘knowledgable-­nom’ muŋuru-­mira ‘very knowledgeable (nom)’ A related evaluative construction in Kayardild is X-­mira-­ X-­, in which the stem X appears twice, once suffixed with -­mira and once, in the next word, as a simple stem:3 (9) Kayardild: a. mutha-­a

‘much-­nom’

mut̪ a-­mira mut̪ a-­a

‘very much’ (nom)

Yukulta and its Relatives Kayardild and Lardil 451 b. caʈarka ‘Crow (nom)’ (proper name) c. ʈaŋka-­a ‘man-­nom’ d. ŋicin-­ta ’1sg.poss-­nom’

caʈarka-­mira caʈarka ‘truly Crow’ (nom) ʈaŋka-­mira ʈaŋka-­a ŋicin-­mira ŋicin-­ta

‘proper man’ (nom) ‘very much mine’ (nom)

In the X-­mira-­ X-­ construction the stem X may be nominal (9b, c), in which case its intensifying meaning appears to be ‘very much (an) X’, comparable to the use of -­mira on a noun stem in Yukulta’s palalaɲi-­mira ‘very old man’ in (4) above. In (9d) the stem X is a possessive pronominal stem. Although Kayardild, like Yukulta, possess a root mut̪ a-­ ‘much, many, plenty’, it does not employ -­mut̪ a as an intensifying suffix. Like Yukulta, Kayardild uses the ameliorative suffix -­paʈa on kin terms (10), where it optionally replaces the kin suffixes -­cu and -­t̪ u on those stems which have either (10b, c); and some -­paʈa forms, as in (10d), select exceptional stems (Evans 1995, 192). (10) Kayardild: a. kacakaca b. marka-­t̪ u-­wa c. papi-­cu-­wa d. t̪ apu-­cu-­wa

‘daddy (nom)’ father’s sister-­kin-­nom’ ‘father’s mother-­kin-­nom’ ‘elder brother-­kin-­nom’

kacakaca-­paʈa marka-­(t̪ u-­)paʈa papi-­(cu-­)paʈa t̪ aput̪ a-­paʈa

‘dear daddy (nom)’ ‘dear auntie (nom)’ ‘dear granny (nom)’ ‘dear brother (nom)’

An evaluative construction found in Kayardild but not in Yukulta is the verbal pejorative formed with the prefix piʈin-­, shown in (11).4 The form piʈin-­has a constant semantic value ‘fail to do properly’ (Evans 1995, 295–6). It bears a clear resemblance to piʈi-­‘bad’, though the etymological source of the final n is not known. (11) Kayardild: a. ŋutic-­a b. cinkac-­a c. ŋalamat̪ -­a

‘throw-­actual’ piʈin-­ŋutic-­a ‘mis-­throw-­actual’ ‘follow-­actual’ piʈin-­cinkac-­a ‘fail to correctly-­follow-­actual’ ‘marry-­actual’ piʈin-­ŋalamat̪ -­a ‘wrongly-­marry-­actual’

3.2  Lardil (Northern Tangkic) Lardil is the sole member of the Northern branch of the Tangkic language family. Major sources are Klokeid (1976) and a sketch grammar and dictionary (Ngakulmungan Kangka Leman 1996), both of which are largely based on empirical fieldwork by Ken Hale in 1960. Lardil has no equivalent of the evaluative suffixes described for Yukulta above. Its root kupa-­ ‘good’ does appear in an intensifying ‘very’ construction X-­ kupa-­ X-­, which closely parallels the X-­mira-­ X-­ of Kayardild both in form and in semantics; however, in Lardil the construction is not morphological, but phrasal, consisting of three separate words. Clear evidence for this is that each of the three words undergoes phonological processes which apply only word-­finally. In (12), for example, final /ʈa/ deletes from under­ lying /kuɳʈakuɳʈa/, and final /a/ in underlying /kupa/ raises to surface u.5 (12) Lardil: kuɳʈakuɳ

‘dry(nom)’

kuɳʈakuɳ kupu kuɳʈakuɳ ‘completely dry’ (underlyingform:/kuɳʈakuɳʈakupakuɳʈakuɳʈa/)

Erich Round

452

Though Lardil has no nominal or adjectival evaluatives, it does possess one minor evaluative construction, which is verbal. Four of the eight monosyllabic verbal roots in Lardil undergo reduplication which is semantically attenuative, with the sense ‘do X lightly’, as shown in (13).6 (13) Lardil: a. pet̪ a b. ɻat̪ a c. ɳet̪ a d. mat̪ a

‘bite’ (underlying: /pec/) ‘spear’ (/ɻac/) ‘hit; kill’ (/ɳec/) ‘seize; grab’ (/mac/)

peecpe ɻaala ɳeene maaɲma

‘nibble’ (/peec-­pec/) ‘jab lightly’ (/ɻaac-­ɻac/) ‘tap softly’ (/ɳeec-­ɳec/) ‘twitch’ (/maac-­mac/)

The association between monosyllabic reduplication and the ‘do X lightly’ meaning is difficult to attribute merely to chance, since the ‘do X lightly’ meaning is not associated with the reduplication of longer verbal roots in Lardil, or with any verbal reduplications in Yukulta or Kayardild, despite the existence of several dozen reduplicated verbal stems. To this extent, Lardil appears to be a counter-­example to the proposed implicational universal (Nieuwenhuis 1985) that a language will have verbal EM only if it has nominal EM.

Notes 1. Keen (1972; 1983) makes additional word-­class distinctions, such as ‘time words, locational words’, based on semantic criteria. 2. The final ‘modifier’ position may be any of determiner, number or adjective. 3. Evans (1995, 195) transcribes an example of X-­mira-­ X-­ as a single word; however, additional examples indicate that the correct segmentation is as two words – for example, in (9a, c), the second word behaves as a disyllable and takes the overt nominative suffix -­a; this would not appear if the X-­mira X-­ form were a single, six-­syllable word. 4. Evans (1995, 296) mentions a verbal prefix cil-­/cul-­ ‘do intensely’, but lexemes containing this element exhibit highly idiosyncratic semantics, such as cilpuɻut̪ -­a ‘stay in one place; keep doing something-­actual’, cf. puɻut̪ -­a ‘gather-­actual’. 5. On a-­raising in Lardil, see Round (2011, 340–1). 6. In monosyllabic verbal reduplication, the vowel of the first copy is lengthened, shown here as aa and ee. On the presence of final, underlying /c/ in Lardil verb stems see Klokeid (1976) and Round (2011, 336–7).

18 Africa 18.1 Berber

Nicola Grandi

1 Introduction 1.1  Berber languages The label ‘Berber languages’ conventionally indicates a group of genetically related and structurally similar languages spoken in North Africa (mainly in Morocco and Algeria); smaller groups of speakers are found in Libya, Tunisia, Mali and Niger. Berber languages constitute a branch of the Afro-­Asiatic family. Because of the similarities among these languages, many scholars speak of a single Berber language, so in the following paragraphs I will refer to Berber languages using the singular ‘Berber’.1 Since the level of standardisation of Berber languages is on average quite low, and since no common Berber alphabet exists today, differences in written forms of these languages are quite frequent. This explains why the same form may sometimes appear with different graphic representations in the following sections. Most data are accompanied by an indication of the source grammar, so the reader can easily see which dialect I am referring to. 1.2  Evaluative formation Within the studies on evaluative morphology, Berber deserves particular attention since ‘evaluative words’ are formed by means of two typologically unusual processes: a circumfix (t___t) and a process of subtractive morphology.2 Moreover, Berber evaluative constructions systematically interact with the categories of gender and number. Although this kind of interaction is not surprising when considering evaluative morphology (cf. Grandi, this volume, Chapter 7), in the case of Berber there is a systematic relationship between feminine gender and diminutives on the one hand, and masculine gender and augmentatives on the other. This relationship also affects the input of these word-­formation processes, constraining their domain of application. Before analysing the data, it is necessary to give a concise sketch of Berber nominal morphology.

2  Concise sketch of nominal morphology Berber nouns can be inflected for three classes: gender, number and state. As for inflectional morphology, adjectives behave like nouns (Kossmann 1997, 119). The existence of an autonomous class of adjectives in Berber is sometimes questioned by scholars.

Nicola Grandi

454 2.1 Gender

The Berber gender system is based on the opposition between masculine and feminine; formal assignment largely prevails over semantic assignment: (1) a. Masculine gender is assigned to nouns starting with: a-­ assif ‘river’ i-­ ifri ‘cave’ ou-­ ouchche ‘jackal’ wa-­ wagerzam ‘leopard’ b. Feminine gender is assigned to nouns starting (and often ending) with t: taxamt ‘tent’, tigemmi ‘house’. Semantic assignment concerns only nouns designating living beings, where the distinction between male and female is evident (so where sexual dimorphism is present). In these cases, ‘le féminin se forme du nom masculin par la préfixation et la suffixation d’un t’ (Laoust 1921, 3): (2)

a-­froukh

>

t-­a-­froukh-­t

fs.sg-­boy.m f-­fs.sg-­boy-­f

‘boy’ ‘girl’ a-­founas > t-­a-­founas-­t fs.sg-­ox.m f-­fs.sg-­ox-­f ‘cow’ (Laoust 1921, 3)

2.2 Number Berber displays two number values: singular and plural. Usually, only the latter is marked. The plural has three forms, according to the type of noun. The first type is the most regular, and is usually called ‘external plural’: it consists in changing the initial vowel of the noun, and in adding the suffix -­n (3a). The second type, known as the ‘internal plural’ or ‘broken plural’, involves a change both in the initial vowel and in the vowel(s) of the root (or lexical morpheme; (3b)). The third is the ‘mixed plural’: it combines changes in the initial vowel, changes in radical vowel(s) and the suffix -­n (3c): (3) a. a-­rgaz > fs.sg-­man ‘man’ b. a-­gadir > fs.sg-­wall ‘wall’ c. a-­fous > fs.sg-­hand ‘hand’

i-­rgaz-­en fs.pl-­man-­pl ‘men’ i-­goudar fs.pl-­wall.pl ‘walls’ i-­fass-­en fs.pl-­hand.pl-­pl ‘hands’

Other forms fall into the category of number, and interact interestingly with evaluative morphology. Berber has a class of collective and mass nouns, singular in form, but plural in their meaning; almost all these nous are masculine:

Berber 455 (4) aman ‘water’ zaou ‘hair’ azMur ‘(quantity of) olive trees’ zz‡udeyya ‘(quantity of) carrots’ These nouns cannot be pluralised. A noun of unity (a sort of singulative form) can be derived from a mass/collective noun by means of the circumfix t___t,3 already mentioned in (1) and (2) with regard to feminine nouns; as will be shown later, the same strategy is used in diminutive formation: (5)

t-­a-­qizzu-­t (Aspinion 1953, 66) sngl.f-­fs.sg-­(quantity of) carrots-­sngl.f ‘one carrot’ t-­a-­lmšmaš-­t (Abdel-­Massih 1971, 118) sngl.f-­fs.sg-­(quantity of) apricots-­sngl.f ‘one apricot’ t-­a-­zemmur-­t (de Vinciennes and Dallet 1960, 71) sngl.f-­fs.sg-­(quantity of) olive trees-­sngl.f ‘one olive tree’

As a matter of fact, all nouns of unity are feminine. They can be pluralised, but their plural form is not equivalent to the mass/collective base noun: the latter indicates an unbound and undetermined quantity of items, the former a multiplicity of single items.4 2.3 State Berber nouns can also be inflected for state. Berber distinguishes between free state (unmarked) and annexed state (also called construct state). Formal means to realise the annexed state vary from one dialect to another: the prevalent strategies are a change in the initial vowel, a loss of the initial vowel, an addition of a semi-­vowel word-­initially or, in some cases, no change at all: (6) Free state Annexed state adrar udrar ‘mountain’ tamurt tmurt ‘country’ asif wasif ‘river’ taddart taddart ‘village’ Nouns in the annexed state usually occur as postverbal subjects (as in (7a), where I maintain the gloss and the translation of the source grammar), as complements of a preposition or in noun complement constructions. (7) a. yus-­ed u-­rgaz a-­meqqran venu-­est as.sg-­homme as.sg-­grand ‘est venu un homme grand’ (‘a big man came’)

Nicola Grandi

456 b. zrigh a-­qchich a-­mevvyan j’ai.vu fs.sg-­garçon fs.sg-­petit ‘j’ai vu un petit garçon’ (‘I saw a small boy’) (Schoen 1959, 22)

3  Evaluative constructions In the following sections I will focus on nominal evaluative formation, which is far more interesting than the verbal and adjectival. 3.1 Diminutives Berber diminutives are formed by means of the circumfix t___t, also used in the formation of feminine nouns (2) and of nouns of unity (5): (8)

t-­a-­dar-­t dim.f-­fs.sg-­foot-­dim.f ‘small foot’ t-­a-­ncuc-­t dim.f-­fs.sg-­lip-­dim.f ‘small lip’ t-­a-­zrtil-­t dim.f-­fs.sg-­carpet-­dim.f ‘small carpet’ t-­a-­qbu-­t dim.f-­fs.sg-­hole-­dim.f ‘small hole’ t-­a-­zrtil-­t dim.f-­fs.sg-­mat-­dim.f ‘small mat’

(Aspinion 1953, 13) (Kossmann 1997, 112) (Abdel-­Massih 1971, 128) (Bentolila 1981, 408) (Abdel-­Massih 1971, 128)

All diminutives are necessarily feminine. The interaction between evaluative constructions and gender is not surprising, since it involves many typologically different and genetically unrelated languages. Nevertheless, as far as Berber is concerned, the relation between evaluative formation and gender seems more systematic than elsewhere. In Berber we observe a total overlap between diminutives and feminine gender, since the circumfix t___t is the only formal strategy to form both feminine nouns5 and diminutives. As a consequence, it is necessary to understand where the boundary between the two categories can be placed, because if we can assert that all diminutives are feminine, the reverse is not true: not all feminine nouns allow a diminutive interpretation. The starting point in order to deal with this issue is represented by nouns that theoretically admit both interpretations, namely animate nouns (the class of nouns where gender is semantically assigned):6 (9) a. t-­a-­fruq-­t dim.f-­fs.sg-­bird-­dim.f ‘little bird’

(Basset and Picard 1948, 19)

Berber 457 b. t-­a-­herdan-­t f/dim.f-­fs.sg-­lizard-­f/dim.f ‘small lizard’ or ‘female lizard’ t-­a-­slem-­t f/dim.f-­fs.sg-­fish-­f/dim.f ‘small fish’ or ‘female fish’ c. t-­a-­founas-­t f-­fs.sg-­ox-­f ‘cow’ t-­a-­froukh-­t f-­fs.sg-­boy-­f ‘girl’ t-­a-­srdun-­t f-­fs.sg-­male mule-­f ‘female mule’ t-­a-­ydi-­t f-­fs.sg-­dog-­f ‘bitch’ t-­a-­tbir-­t f-­fs.sg-­pigeon-­f ‘female pigeon’

(Kossmann 1997, 113) (Kossmann 1997, 113) (Laoust 1921, 3) (Laoust 1921, 3) (Bentolila 1981, 407) (Bentolila 1981, 407) (Kossmann 1997, 112)

These data reveal that the pattern t-­N [+ animate]-­t is interpreted as feminine in most cases (examples such as those in (9c) are considerably more frequent than those in (9a) and (9b)). The interpretation is unequivocally diminutive only in (9a). In (9b) both interpretations are possible. The distribution of the two meanings can be explained as follows: masculine nouns designating living beings usually do not have a ­morphological d ­ iminutive form; the pattern t-­N [+ animate]-­t refers to the corresponding female. A diminutive interpretation is sometimes possible with masculine nouns indicating reptiles, fish, etc. (cf. 9b) or a species (cf. 9a) (see also Kossmann 1997). These tendencies can be summarised by stating that diminutives can be formed from animate nouns on condition that they designate living beings without sexual dimorphism, so if there are no perceptual and exterior differences between male and female. If sexual dimorphism is present, then the pattern t-­N [+ animate]-­t always designates the female.7 This generalisation poses a clear-­cut boundary between the two possible interpretations, and has a relevant theoretical consequence: in Berber, animate nouns can hardly be ‘evaluated’. This contradicts all supposed ‘evaluability hierarchies’ that agree in assigning the highest positions to animate nouns.8 Moreover, the overlap between feminine gender and diminutives has the consequence of excluding feminine nouns from the domain of diminutive formation: of course, they cannot be prefixed and suffixed with a further t. So, nouns such as tamart ‘beard’ and tagant ‘forest’ cannot have a morphological diminutive. To conclude, we can state that in Berber the domain of diminutive formation is restricted to masculine inanimate and countable nouns: the circumfix t___t can have a diminutive interpretation only in this case. Animate nouns are excluded from diminutive formation, because the circumfix t___t designates the female; uncountable nouns are excluded because the circumfix t___t

Nicola Grandi

458

d­ esignates the noun of unity; feminine nouns are excluded because they cannot bear the circumfix t___t.9 3.2 Augmentatives Augmentative formation also presents many problematic and puzzling aspects. The process used to form augmentatives is the opposite of the one used to form diminutives: ‘certain feminine nouns give augmentatives by a process that is the reverse of diminutive formation’ (Abdel-­Massih 1971, 116). Therefore, if, on the one hand, only masculine nouns (with the exception of most animate nouns and uncountable nouns) can be diminutivised, then, on the other hand, only feminine nouns can be ­augmentivised: in ­augmentative formation the initial t and final t (if present) of a feminine noun are ‘deleted’; so we have a typologically unusual instance of subtractive morphology: (10)





t-­a-­bhir-­t > a-­bhir (Abdel-­Massih 1971, 116) f-­fs.sg-­garden-­f fs.sg-­garden.aug.m ‘garden’ ‘big garden’ t-­a-­mar-­t > a-­mar (Abdel-­Massih 1971, 116) f-­fs.sg-­beard-­f fs.sg-­beard.aug.m ‘beard’ ‘big beard’ t-­a-­gan-­t > a-­gan (Aspinion 1953, 14) f-­fs.sg-­forest-­f fs.sg-­forest.aug.m ‘forest’ ‘big forest’

As a consequence, all augmentatives are masculine. Once again, most animate nouns are excluded from the domain of augmentative formation: in these cases, the form without the initial t and final t designates the male. As for diminutives, the exception is represented by nouns indicating animals with no sexual dimorphism: tacettuft ‘ant’ > acettuf ‘big ant’ (Kossmann 1997, 113). Both diminutives and augmentatives can be reinterpreted as pejoratives (Mettouchi 1999, 219). 3.1.1  Problematic data Even though in almost all cases it is clear which word is the base and which word is the evaluative form (since just one of them indicates the default value), in some instances a direction in the process of derivation can barely be established: (11)

t-­a-­qenda‡-­t

>

a-­qendar

f-­fs.sg-­weight-­f fs.sg-­weight.m

’50 kilos of weight’

(Kossmann 1997, 113)

’100 kilos of weight’

In these cases, it is not clear whether the masculine form is the augmentative of the feminine one or vice versa: ‘[s]ouvent, le féminin désigne quelque chose de plus petit que le masculin sans qu’on puisse dire que l’un est diminutif ou que l’autre est augmentative’ (Kossmann 1997, 113). So, sometimes, ‘il dépend du mot si la forme féminine est diminutive ou la forme masculine est augmentative’ (Kossmann 1997, 112).

Berber 459

4 Conclusion It must first be remarked that quite an unusual polysemy seems to emerge: a single formal means, namely the circumfix t___t, is used to form feminine nouns, nouns of unity and diminutives. What is relevant for a theory of word formation is that the only way to predict the semantic role of the affix is to know the meaning of the base. In other words, the meaning of the circumfix t-­.  .  .-­t cannot be specified in isolation from the morphological structure it is part of. As shown above, in all cases the circumfix is the only formal strategy available. This determines, as a consequence, a sort of complementary distribution of the three interpretations: if feminine interpretation is possible, then diminutive and singulative interpretations are ruled out, etc. It is quite difficult to understand the origin of this unusual polysemy, since written documentation of the ancient states of Berber languages is rare. Mettouchi (1999, 220–1) suggests that the feminine is a secondary and recent extension of an original ‘marqueur de différenciation’; on the basis of this consideration, the diminutive and singulative meanings would be more ancient than the feminine interpretation.10 Moreover, the interaction of evaluative-­formation strategies with the category of gender is more systematic than in other languages. This interaction involves both input and output of evaluative-­formation processes. As for the input, gender is the main restriction on these word-­formation processes: as stated above, only masculine nouns (but not all masculine nouns) can have a diminutive form; only feminine nouns (but not all feminine nouns) can have an augmentative form. As far as output is concerned, evaluative-­formation processes always change the gender of the base: all diminutives are feminine, all augmentatives are masculine. Moreover, diminutives and augmentatives exclude one another: a single base word cannot have both a diminutive form and an augmentative one. Diminutives exclude augmentatives and, vice versa, augmentatives exclude diminutives. Consequently, in Berber we can have a ‘normal (or standard) beard’, tamart; a ‘big beard’, amar, but not a ‘small beard’; and we can have a ‘normal foot’, adar, a ‘small foot’, tadart; but not a big foot! As for the domain of evaluative formation, animate nouns are excluded, with a few exceptions. This is quite an unusual situation in a cross-­linguistic perspective. Finally, if an attempt is made to place evaluative formation within Berber morphology, some difficulties are encountered. The use of the circumfix t___t with adjectival bases is no doubt inflectional. With nouns, it of course designates a referent other than the masculine. However, it also has syntactic relevance, since adjectives agree with the head noun in gender and number, and verbs agree with the subject in gender.11 So this circumfix presents characteristics which render it very close to inflection, even if it cannot be numbered among typical inflectional processes. Since all diminutives are formed by a gender shift (from masculine to feminine; cf. Grandi, this volume, Chapter 7, for a general sketch of the interaction between gender and evaluative formation), they also have syntactic relevance. If an adjective modifies a diminutive noun, it takes the feminine gender. Nevertheless, the agreement is limited to gender; in other words, it does not hold that the adjective must have a diminutive nuance too. The same holds for augmentative formation. Evaluative forms have syntactic relevance, since they trigger partial agreement: modifiers agree with their nominal heads for gender, which in turn is a consequence of evaluative formation. If we consider syntactic relevance as a crucial parameter in establishing the derivational vs. inflectional nature of a word-­formation process (cf. Grandi, this volume, Chapter 6), then we cannot ascribe Berber evaluative formation only to derivation.

460

Nicola Grandi

Notes   1. I will refer especially to Tamazight and Kabyle varieties.   2. The field of evaluative formation is wider (Vycichl 1961); in this chapter I will focus only on the two processes just mentioned, which seem to be the more productive. As a consequence, this chapter does not aim to draw an exhaustive picture of Berber evaluative constructions.   3. ‘The noun of unity is formed from a collective noun by the same process used [. . .] for the formation of diminutives’ (Abdel-­Massih 1971, 117).   4. Not all phenomena mentioned in this paragraph share the same status. While the formal ­strategies used to form plural from singular are clearly an instance of inherent inflection, the formation of nouns of unity is close to derivation.   5. There are also cases in which the feminine is formed by a prefix t-­, that is to say without a ­simultaneous suffixation with -­t.  6. f/dim.f in (9b) means that both feminine and diminutive interpretations are possible.   7. It is interesting to note that the Kabyle words for child, agrud, and for baby, ilufan, are both masculine. This indicates that masculine seems to be the default gender in Berber.   8. This position also holds from a historical perspective, since actual diminutive affixes often derive from affixes originally used to designate pups, cubs and young animals in general (Matisoff 1991; Grandi 2011b; Mutz, this volume).   9. There are a few examples which have an ambiguous interpretation. For example, taddouarit is a feminine noun (without a masculine counterpart) with a possible diminutive nuance: ‘jolie maison’ (Laoust 1921, 2). 10. ‘Nous pensons que cet affixe “t” est primitivement un marqueur de différenciation avec hiérarchisation du répéré par rapport au repère. Il se pourrait qu’il soit d’abord apparu dans les contextes partitifs ou diminutifs’ (Mettouchi 1999, 221). 11. Aqšiš amštuh̟ i-­r̟ uh̟ ‘le petit garçon, il est parti’ vs. tqšišt tamštuht̟ t-­r̟ uh̟ ‘la petite fille, elle est partie’ (Mettouchi 1999, 217).

18.2  Classical and Moroccan Arabic

Nora Arbaoui

1 Introduction Arabic belongs to the Semitic language group. It is a non-­concatenative language in which word patterns form the basis of derivation. Three categories of evaluative morphology are used in Arabic: diminutive, pluractional verbs and augmentative. Arabic has three diminutive forms: fuʕayl, fuʕayʕil and fuʕayʕiil (wherein stand for any consonant). The first form is based on trilaterals (bases with three root consonants), the second on quadrilaterals (sometimes encountered with quinquilaterals, too). The third form is unique to quinquilaterals. These patterns are productive and systematic. They apply to both nouns and adjectives, as well as some adverbs. Pluractional verbs have only one pattern: faʕʕal. There are also augmentatives, derived specifically from active participles, which take on one of four patterns for trilaterals: faʕil, faʕiil, faʕuul, faʕʕaal, and one for quadrilaterals: mifʕaal. In what follows, I first propose a description of the process of diminutive formation in Classical Arabic, and then discuss pluractional verbs and augmentatives. I then conduct a brief comparison with Moroccan Arabic.

2  Diminutive formation in Classical Arabic 2.1  Diminutive formation from trilateral bases Diminutives in Classical Arabic are formed by the insertion of y after the second consonant. This insertion is accompanied by a vocalisation that is specific for diminutives. Whatever the vocalisation of the base – , , or – the diminutive always has the vocalisation . (1) Diminutive kalb ‘dog’ kulayb ‘doggie’ jisr ‘bridge’ jusayr ‘small bridge’ jabal ‘mountain’ jubayl ‘hill’ juħr ‘cave’ juħayr ‘small cave’ kabiir ‘big’ ʔaswad ‘black’

kubayr suwayd

‘slightly big’ ‘barely black’

Nora Arbaoui

462 baʕd ‘after’ buʕayd qabl ‘before’ qubayl

‘just after’ ‘just before’

Interestingly, prefixes and infixes used in the formation of adjectives disappear in the diminutive; only the root consonants are preserved:1 (2) a.

C

b. a

a

u

a

|

|

|

|

V

C V

V

|

C

V C V C V

C

[C V]

C

|

|

|

|

|

|

|

s

w

d

s

w

y

d

V

To summarise, in order to derive a diminutive from a trilateral noun or adjective, it is enough to insert the root consonants, as in the template in (3): (3) C

u

a

|

|

V

C V

[C V] C

|

|

|

|

C1

C2

y

C3

V

2.2  Diminutive formation from quadrilateral bases (4) Diminutive a. dirham ‘dirham’ durayhim qimaṭr ‘drawer’ qumayṭir jundub ‘grasshopper’ junaydib jaʕfar ‘proper name’ juʕayfir

‘just one dirham’ ‘small drawer’ ‘small grasshopper’ ‘proper name’

b. masjid ‘mosque’ kitaab ‘book’

‘small mosque’ ‘small book’

musayjid kutayyib

The nouns in (4a) are true quadrilaterals, i.e. nouns in which all four consonants are radicals. In contrast, the nouns in (4b) are themselves derived from trilateral bases, and have a consonantal or vocalic augment. In this last case, the diminutive form includes the consonantal augment, but gets rid of the vocalic one together with the vocalisation of the base. Indeed, as for trilateral bases, the quadrilateral diminutive has its own vocalisation, while still inserting y after the first consonant:

Classical and Moroccan Arabic 463 (5) a. C

i

a

|

|

V

C

V

C

V

C V

b. u

a

i

|

|

| [C V]

C V C V

C

V

C V

|

|

|

|

|

|

|

|

|

d

r

h

m

d

r

y

h

m

The base nouns have different vocalisations (, and ). Nevertheless, their diminutives carry the same vocalisation . In sum, in order to form the diminutive of a quadrilateral, the consonants of the base – be they radicals or augments – are placed in the template in (6): (6)

u

C

a

|

|

V

C V

|

|

C1

C2

i | [C V]

C

| y

V

C V

|

|

C3

C4

2.3  Diminutive formation from quinquilateral bases (7) a. safarjal ‘quince’

Diminutive sufayrij

‘small quince’

b. mudaħraj ‘rolled up’

duħayrij/duħayriij

‘rolled up slowly’

c. munṭaliq ‘going’

muṭayliq/muṭayliiq

‘going slowly’

d. sulṭaan ‘sultan’ sulayṭiin sakraan ‘drunk person’ sukayriin

‘little sultan’ ‘slightly drunk person’

Diminutives of quinquilaterals are also formed with a y after the second consonant and a vocalisation resembling that of quadrilateral diminutives . That said, four different cases are present. First, all base consonants are radical, in which case the last consonant is simply suppressed: (8) a.

b. a

a

a

u

a

i

|

|

|

|

|

|

C V C

V

| s

| f

C

V C V

C V

C

V C V

[C V]

C V C V

|

|

|

|

|

|

|

r

j

l

s

f

y

r

| j

Nora Arbaoui

464

Second, the quinquilateral includes an augment (m below), in which case it is that augment which is suppressed:

(9) a.

b. u

a

a

u

a

i

|

|

|

|

|

|

V

[C V] C

V

C V

C V

C V C V C V C V

C V C

|

|

|

|

|

|

|

|

|

|

m

d

ħ

r

j

d

ħ

y

r

j

Third, the quinquilateral form has two consonantal augments, in which case one of the augments is suppressed: (10) a.

b. u

a

i

u

a

i

|

|

|

|

|

|

C V

C

|

|

m

n

C V C V [C V] C

V C V C V C V |

|

|

|

l

q

m

|

V C

|

|

|

y

l

q

V

In this latter case, the suppression may or may not be accompanied by lengthening of the vowel i. Fourth, the quinquilateral base includes two augments, one of which is manifested as a long vowel; in this case, the diminutive preserves all the consonants of the base, as well as the length of the base’s vocalic augment:

(11) a.

u

b.

a

|

/

\

C V C V C V C V C V |

|

s

l

|

u

a

|

|

C V C V

i /

\

[C V] C V C V C V

|

|

|

|

n

s

l

y

|

| n

2.4  Diminutive formation from words with more than five elements (12)

mustaḍrib ‘become thick (honey)’ muḍayrib/muḍayriib ‘become a little thick’ ʕandaliib ‘nightingale’ ʕunaydil/ʕunaydiil ‘small nightingale’

In bases with more than five elements, the base includes either three augments, in which case two are suppressed in the diminutive (13); or five consonants and a long vowel, in

Classical and Moroccan Arabic 465 which case the last consonant is suppressed and the length of the vowel is either ­shortened or preserved. The quality of the diminutive vowel is unsurprisingly i (14): (13) a.

b. u

a

i

u

a

i

|

|

|

|

|

|

C V C V C V C V C V CV

C V

|

|

|

m

s

t

|

|

|

|

r

b

m

(14) a.

C V |

[C V] C V C V |

|

|

y

r

b

b. a

a

|

|

i /

\

u

a

i

|

|

|

C V C V C V C V C V C V

C V C V

|

|

|

|

|

|

n

d

l

b

[C V]

C V C V

|

|

|

|

n

y

d

l

If so, whatever the size of the base, in order to form the diminutive, one has to insert: • • • •

y after C2; a vowel u in V1; a vowel a in V2; a vowel i in V4 if the diminutive form has more than four consonants, including y:

In order to satisfy this template, if the nominal base includes more than four consonants, the following adjustments are made: (15)

u |

a |

C V

C V

i | [C V] C V C V | y

• if all consonants are radical, suppress the last one; • if one of the consonants is an augment, suppress the augment; • if there are three radicals and two consonantal augments, suppress one augment; • if there are three radicals, one consonantal augment and one vocalic augment in the form of a long vowel, preserve the consonant and the CV position of that long vowel; • if there are three radicals and three consonantal augments, suppress two augments;

Nora Arbaoui

466

• if there are five radicals and a long vowel, suppress the last consonant and the position of the long vowel. To summarise, the diminutive template always reduces the number of consonants of the base to four by suppressing radicals or augments. 2.5  Feminine diminutives Feminine diminutives employ the same strategy as masculine ones: insertion of y in C3 and vocalisation: (16) kalba ‘female dog’ safarjala ‘one quince’ ħamraaʔ ‘red’ ħublaa ‘pregnant’

Diminutive kulayba sufayrija ħumayraaʔ ħubaylaa

‘female doggie’ ‘one small quince’ ‘barely red’ ‘recently pregnant’

It is noteworthy that for those feminine bases on which there is no segmental gender marking, such gender marking does appear on the diminutive in the form of a suffix -­a: (17) Diminutive naʕl ‘sandal’ nuʕayla qadam ‘foot’ qudayma

‘small sandal’ ‘small foot’

3  Pluractional verbs in Classical Arabic Pluractionality in verbs denotes the repetition of an action; in some cases, the repetition denoted ends up having a diminutive meaning, in the sense that the event is broken up into smaller segments. Arabic has one pluractional pattern, faʕʕala: (18) qaṭaʕa ‘cut’ faraq ‘separate’

Pluractional verb qaṭṭaʕa ‘slice’ farraq ‘disperse’

This pattern regularly hosts causative verbs with the pattern [Argument1 make Argument 2 do something]. If Argument 2 does not exist, the verb takes on an intensive meaning (Arbaoui 2010). Instead of having multiple arguments, the verb has multiple instances, thus endowing it with intensive, pluractional sematics. The pattern of pluractional verbs contains a CV position after C2 which allows this consonant to geminate: (19) a.

b. a

a

a

a

|

|

|

|

C V C V C V

C V [CV] C V C V

|

|

|

|

f

r

q

f

\

/ r

| q

Classical and Moroccan Arabic 467

4  Augmentative formation from trilateral bases in Classical Arabic Only participial forms have augmentative variants. Principally, the participial form derived from trilateral verbs is faaʕil. However, to indicate an intensive meaning, one of four patterns may be used: faʕil, faʕiil, faʕuul and faʕʕaal: (20) faaṭin ‘cautious’ ħaaðir ‘prudent’

Augmentative faṭin ‘very cautious’ ħaðir ‘extremely prudent’

saamiʕ ‘hearing’ samiiʕ ʕaalim ‘knowledgeable’ ʕaliim

kaatim ‘secretive’ šaakir ‘thankful’

‘good hearing’ ‘very knowledgeable’

katuum šakuur

‘very secretive’ ‘very thankful’

ʕaalim ‘knowledgeable’ ʕallaam faatik ‘devastating’ fattaak

‘highly knowledgeable’ ‘extremely devastating’

With verbs of measure IV ʔafʕal, i.e. a causative verbal form, the participial form is basically mufʕil and the augmentative one is mifʕaal: (21) muqdim ‘intrepid’ muʕṭii ‘generous’ muʕiin ‘helpful’

Augmentative miqdaam ‘very intrepid’ miʕṭaaʔ ‘extremely generous’ miʕwaan ‘very helpful’

5  Comparison with Moroccan Arabic 5.1 Diminutives In Moroccan Arabic, too, diminutives involve an infix y after the second consonant. However, unlike in Classical Arabic, this y is geminated, and the vocalisation is absent. In general, Moroccan Arabic does not preserve the vowels of Classical Arabic:2 (22) sbəʕ ‘lion’ məšš ‘cat’ kas ‘glass’

Diminutive sbəyyəʕ mšəyyəš kwəyyəs

‘small lion’ ‘small cat’ ‘small glass’

To illustrate, consider the noun sbəʕ and its diminutive sbəyyəʕ: (23) a.

b. |

C

V

C

|

|

s

b

V

| C V |

C

V

C

|

|

s

b

V

| [C V C V ] \

/ y

C V |

Nora Arbaoui

468

Unlike in Classical Arabic, where the same template is used for masculine and feminine diminutives, the feminine diminutives of Moroccan Arabic include not a geminate yy but rather a vowel i. This vowel is always placed in the same position, namely after the second consonant: (24) bəlʁa ‘slippers’ dəmʕa ‘tear’ məšša ‘female cat’

Diminutive bliʁa dmiʕa mšiša

‘small slippers’ ‘just one tear’ ‘small female cat’

Thus the examination of trilaterals identifies two realisations of the diminutive: yy if the base is masculine and i if it is feminine. It is noteworthy that for those feminine bases on which there is no segmental gender marking, such gender marking does appear on the diminutive, in the form of a suffix -­a: (25) dar ‘house’ bənt ‘girl’

Diminutive dwira bnita

‘small house’ ‘little girl’

Now consider quadrilaterals: (26) Diminutive məqbəḍ ‘hairgrip’ mqibiḍ ṭəbṣil ‘plate’ ṭbiṣil sərwal ‘trousers’ sriwil ṣəbbaṭ ‘shoe’ ṣbibiṭ səllum ‘ladder’ slilim

‘small hairgrip’ ‘small plate’ ‘short trousers’ ‘small shoe’ ‘small ladder’

In (26), a vowel i surfaces, and not yy. This is so despite the fact that the forms are not feminine, as demonstrated by the lack of suffix in both base and diminutive. Moreover, a vowel i 3 appears after the third consonant of the diminutive, regardless of the quality of the base vowel in that position:

(27) a.

b. a |

/

i \

/

i \ [CV ]

/

\

C V C V C V

C V C V C V C V C V

C V C V

|

|

|

|

|

|

|

|

s

r

w

l

s

r

w

l

The same applies to quadrilateral feminine bases: (28) Diminutive ṣəndala ‘sandal’ ṣnidila bərraka ‘shack’ bririka

‘small or fine sandal’ ‘small shack’

Classical and Moroccan Arabic 469 To illustrate, consider the noun ṣendala and its diminutive ṣnidila: (29) a.

a

b.

a

i

i

a

/ \ / \ / \ C V C V [CV] C V C V C V+C V | | | | s n d l

| / \ / \ C V C V C V C V C V+ C V | | | | s n d l

To summarise, the diminutive in Moroccan Arabic is formed on masculine trilateral bases by the insertion of a geminate yy after the second consonant; feminine trilateral bases, as well as all quadrilateral bases, involve the insertion of a vowel i in the same position. In these last two cases, the diminutive includes a second vowel which belongs either to the quadrilateral or to the feminine morpheme. Adjectives exhibit a different pattern. Alongside yy insertion, there are two other strategies. Some adjectives form diminutives by reduplication of the second consonant, in which case a vowel i appears between the two copies; others form their ­diminutives by inserting a geminate ww (this option is reserved to bases with vowels other than ə): (30) Diminutive ṣʁir ‘small’ ṣʁəwwər/ṣrəyyər/ṣʁiʁər kbir ‘big’ kbəwwer/kbəyyer/kbiber

‘slightly small’ ‘little big’



‘subtly soft’ ‘barely red’

rṭəb ‘soft’ ħmər ‘red’

rṭiṭəb ħmimər

To illustrate, consider ṣʁir and its diminutive ṣʁəwwer/ṣʁəyyər:

(31) a.

b. y i

C | ṣ

V

/ \ V C V C V | r

C | ʁ

C | ṣ

ə | V C V | ʁ

ə / \ | [C V] C V C V | r w

or rṭəb and its diminutive rṭiṭəb: (32) a.

i

b. | C | r

V

C |

V

/ C V

C V C

|

|

b

r

|

V

\

|

[C V] C V C V |

| b

Nora Arbaoui

470

To summarise, Moroccan Arabic shares with Classical Arabic the infixation of an element y or i in the formation of diminutives, depending on morphophonological considerations, but it is distinct in that it has a specific pattern for adjectives. 5.2  Pluractional verbs Moroccan Arabic has two patterns for pluractional verbs: fəʕʕəl, which is equivalent to faʕʕal in Classical Arabic, and fəʕləl, which is specific to Moroccan Arabic: (33) Pluractional verb qṭəʕ ‘cut’ qəṭṭəʕ ‘slice’ frəq ‘separate’ fərrəq ‘disperse’ mal ‘incline’ məlməl ‘move’ šəmm ‘smell’ šəmšəm ‘sniff’ 5.3 Augmentatives Moroccan Arabic has two augmentative patterns for participial forms, fəʕaal and məfʕal. Unlike in Classical Arabic, either pattern can be used with trilateral verbs. It should be noted that Moroccan Arabic has lost measure IV and mostly uses the imperfective form instead of the active participle: (34) ta-­yəkdəb ‘he lies’ ta-­ynəmm ‘he prattles’

Augmentative kəddab nəmmam

‘pathological liar’ ‘extreme prattler’



məʁyar məškak

‘extremely jealous’ ‘extremely suspicious’

ta-­yʁiir ta-­yšəkk

‘he is jealous’ ‘he is suspcious’

6 Conclusion The formation of diminutives in Arabic takes into consideration the size of the base. In Classical Arabic, the template of the diminutive depends on whether the base is trilateral, quadrilateral or quinquilateral. In Moroccan Arabic, the diminutive template depends on whether the base has a full vowel (a vowel other than ə), regardless of whether that vowel is internal to the base or a feminine suffix. Some adjectives in Moroccan Arabic form their diminutives by reduplication, an option which is not available in Classical Arabic. The cases surveyed above present conflicting arguments with respect to a view that derives them directly from the root. That the vocalisation of the base is suppressed and adjectives lose their prefixes argues for such a view; on the other hand, the fact that the derived form may contain augments and (if possible) preserve original vowel length indicates that the surface form of the base must be referred to. The augmentative in Classical Arabic reveals an interesting phenomenon: the participial base faaʕil has four corresponding augmentative forms, namely faʕil, faʕiil, faʕuul and faʕʕaal. This implies that the augmentative form is derived from a root, not from the participle.

Classical and Moroccan Arabic 471

Notes 1. In this chapter, I implement the CVCV model as proposed in Lowenstamm (1996). 2. The vowel in, for example, sbəyyəʕ in (22) is transcribed as the epenthetic [ə], even though its phonetic realisation is closer to a short [i]. But this is a predictable phonological fact: the quality of epenthetic vowels before [y] is always a short [i]. 3. Surface vowels with the quality [i] are underlyingly long in Morrocan Arabic (Lowenstamm 1990): this is why they are depicted as occupying two positions in the template in (27).

18.3 Ewe

Yvonne Agbetsoamedo and Paul Kofi Agbedor

1 Introduction 1.1  Language classification and sociolinguistic status Ewe is a major dialect cluster of Gbe or Tadoid of the Volta-­Niger family together with Fon, Gen, Aja and Xwla-­Xweda (Duthie 1996). It is spoken in the south-­eastern part of the Volta Region of Ghana and in some parts of Southern Togo.1 Ewe is a Kwa language from the Niger-­Congo family. It is used as a language of instruction in most basic schools in the Volta Region and is also studied as a subject up to university level. Ewe serves as a lingua franca for some communities, such as the Balɛɛ community whose native language is Sɛlɛɛ. 1.2  Major typological features Ewe is an isolating language with some level of agglutination. Words in the language may be formed by the processes of reduplication, affixation and compounding. As in most Kwa languages, there are also ideophones in Ewe – marked words that depict sensory imagery (Dingemanse 2011, 25). Borrowing is another means of increasing the vocabulary of the language. Ewe is a tone language with three level tones: low, mid and high. Falling and rising tones also occur as a result of historical developments whereby segmental components of a word undergo erosion but the tone assigned to such components is not lost. Thus, floating tones move to adjacent units to form a falling or a rising tone. Example (1) below illustrates a rising tone. Note that all the illustrations in this chapter are constructed examples. (1)

è á yi à? cf. ǎ-­yi-­à? 2sg pot go Q 2sg.pot-­go-­q Will you go? Will you go?

Ewe is not a morphologically rich language. However, some grammatical categories are coded by morphological means. Aspectual and modal categories in the verbal domain are marked by prefixes and suffixes. Nouns, on the other hand, have the non-­high vocalic prefixes à-­ or è-­, both of which are remnants of proto-­Niger-­Congo noun class markers. Number distinction is differentiated by the suffixation of the plural marker -­wó. The language is predominantly an SVO language, but it is not uncommon to find seman-

Ewe 473 tically or pragmatically marked sentences with SOV, OSV or OVS word orders as well. Ewe has serial verb constructions. One important feature of the language is its logophoric marking. Logophoricity, according to Huang (2000), refers to the phenomenon whereby the ‘perspective’ of an internal protagonist of a sentence or discourse, as opposed to that of the current, external speaker, is being reported by some morphological and/or syntactic means. The Ewe logophoric pronouns yè (sing.) and yèwo (pl.) function as subjects of an embedded clause introduced by the complementiser bé; they are always cliticised to the verb. Compare (2a) and (2b). (2) a. ayiyi be yè-­lè log-­be spider comp ‘Spider1 says he1 is sick’

dɔ lé-­m sickness catch-­prog

b. ayiyi be é-­lè spider comp 3sg-­be ‘Spider1 says he2 is sick’

dɔ lé-­m sickness catch-­prog

2  Morphological evaluative markers In this section we examine the diminutive and augmentative morphemes in Ewe. The various meanings and functions of each morpheme are discussed. The suffixes -­vi and -­i are used for the encoding of diminutives and the suffix -­gã for augmentatives. These suffixes will be discussed in Sections 2.1, 2.2 and 2.3 respectively. 2.1  The diminutive suffix -­vi In Ewe, the most common way of expressing diminutive is by the use of the diminutive suffix -­vi. This suffix comes from the word for ‘child’ or ‘offspring’, vi (Heine, Claudi and Hünnemeyer 1991). This phenomenon is attested in many African and Asian languages (Creissels 1999; Greenberg 1959; Jurafsky 1996; Matisoff 1991). The noun vi exists as an independent lexical item in the language, and this is illustrated in (3) and (4) below: (3)

Ama dzì vi èvè Ama bear child two ‘Ama has given birth to two children’

(4)

Kofi nyé vi bùame Kofi be child respectful ‘Kofi is a respectful child’

The noun vi ‘child’ has a mid tone, while the diminutive suffix -­vi has a high tone. The suffix -­vi has a wide range of meanings including ‘small’, ‘young’, ‘membership/citizenship’ and ‘insignificance/triviality’. These different meanings express quantitative evaluation. 2.1.1 Small The basic meaning related to the diminutive is ‘small’ (see Jurafsky 1996). Dressler and Merlini Barbaresi (1994) state the meanings of diminutives as ‘smallness’ as they distinguish between morphosemantic and morphopragmatic meanings of diminutives.

Yvonne Agbetsoamedo and Paul Kofi Agbedor

474

According to these authors, the morphosemantic denotation of the diminutive is ‘smallness’, while its morphopragmatic connotation involves evaluative features such as ‘endearment and affectiveness’. The most basic meaning carried by the diminutive suffix -­vi is ‘small’. This is shown in (5): (5) a. xɔ-­vi b. kplɔ̃-­vi c. awu-­vi d. ame-­vi e. edu-­vi

‘small house’ ‘small table’ ‘small dress’ ‘smallish person’ ‘small town’

2.1.2 Young Another meaning associated with the diminutive suffix -­vi is ‘young’. For example, the Ewe word avu ‘dog’ when suffixed with the diminutive marker yields avu-­vi, which refers to a young dog or puppy. This applies to all animal offspring. Other examples include those shown in (6): (6) a. ŋutsu-­ví b. nyɔnu-­ví dim woman-­ dim man-­ ‘young man/boy’ ‘young woman/girl’ It should be noted here that the notions of ‘small’ and ‘young’ are related. For things that grow, the young are usually small in size as well. As they grow, their size increases, under normal circumstances. The same suffix added to a name indicates the younger of two people bearing the name (usually in the same family). For example, a female child born on Saturday is named Ama. When another female child is born later, but also on a Saturday, she is named Ama-­vi ‘younger Ama’, to distinguish her from the older one. This process applies to all the day names, as illustrated with the following pairs of male vs. female names. (7)

Male Kɔdzo Kɔdzo-­vi Kɔbla Kɔbla-­vi Kɔku Kɔku-­vi Yaw Yaw-­vi Kɔmi Kɔmi-­vi

Female Adzo Adzo-­vi Abla Abla-­vi Aku Aku-­vi Awo Awo-­vi Ama Ama-­vi

Interestingly, in Akan, a closely related language, the diminutive suffix -­ba/-­wa can only derive female names from male names (Appah and Amfo 2011). This same notion of ‘young’ is extended to plants to indicate young plants or seedlings: (8)

atadi-­vi agbitsa-­vi mango-­vi paya-­vi

‘pepper seedling’ ‘aubergine seedling’ ‘mango seedling’ ‘avocado seedling’

To this effect, small size in the fruits of these plants will be expressed using the adjective sue ‘small’ to avoid any ambiguity. For example, a SMALL-­SIZE mango fruit will be mango sue.

Ewe 475 2.1.3 Membership/Citizenship The notion of membership is part of the polysemy of the suffix -­vi, which is not diminutive by default. In this regard, the suffix is attached to nouns referring to names of an organisation, ethnic group, town or country, as exemplified in (9): (9) a. Ewe-­vi b. Ghana-­vi c. dume-­vi d. hlɔ-­vi

‘a member of the Ewe ethnic group’; ‘a citizen of Ghana’ ‘a citizen’ ‘a member of a clan’

2.1.4  Insignificance/Triviality The diminutive suffix -­vi can be added to a base to indicate insignificance or triviality. For example, it may be used in certain contexts to refer to ideas, things or events that are of little significance, as illustrated in (10): (10) a. nu-­vi thing-­ dim ‘little thing/insignificant thing’ b. nya-­vi case-­ dim ‘trivial case’ 2.2  The diminutive suffix -­i Diminutivised adjectives are formed by the suffixation of -­i to, mostly, the dimensional adjectives ‘big’, ‘long’, ‘wide’ and ‘spherical’. The underlying form of the diminutive suffix is -­i but it has different realisations depending on the final vowel of the adjective. It is realised as -­i when the final vowel is /u/ or /i/, -­e after /o/ or /e/ and -­ɛ after /ɔ/ or /ɛ/. When the vowel of the stem is the low open central vowel /a/, the suffix fuses with it to become -­ɛ (Ameka 2012). The suffix is accompanied by a high tone and adds a diminutive meaning to the original meaning of the adjective. For example, the adjective lɔ̀bɔ̀ɔ̀ with a low tone means ‘long and big’. But when the tone changes to high, together with a high tone suffix -­i (i.e. lɔ́bɔ́ɛ/ lɔ́bóe), the adjective means ‘long and small’. Other examples include those in (11): (11) a. gòrò b. gbàdzà

‘round and big’ ‘wide and big’

góró-­e gbádzɛ́

‘round and small’ ‘wide and small’

The non-­derived dimensional adjectives are inherently associated both with dimension and with size. This has interesting implications for their combinability with diminutivised and non-­diminutivised nouns. Consider the examples in (12): (12) a. ati lɔ̀bɔ̀ɔ̀ stick long ‘big/thick long stick’ b. ati lɔ́bɔ́ɛ stick long.dim ‘thin long stick’

Yvonne Agbetsoamedo and Paul Kofi Agbedor

476 c. *ati-­vi lɔ̀bɔ̀ɔ̀ stick-­ dim long ‘small long stick’ d. ati-­vi lɔ́bɔ́ɛ stick-­ dim long.dim ‘small thin and long stick’

The non-­derived nouns in (12a) and (12b) co-­occur with both diminutivised and non-­ diminutivised adjectives. However, when the noun is diminutivised, it can only combine with the diminutivised adjective as in (12d). Example (12c) is unacceptable because the inherent meaning of the adjective does not correspond to the meaning of the diminutive suffix marked on the noun. As a consequence, the use of these adjectives and their diminutivised forms does not require any other modification of the noun because it is already explicit in the adjectives themselves. Nonetheless, when both the noun and the adjective are diminutivised, the diminutive meaning of the entire construction is intensified. The adjective gã ‘big’ plus the suffix -­i becomes gɛ̃ ‘big-­dim’. Interestingly, the diminutivised ‘big’ does not mean ‘SMALL-­SIZE’ but, rather, a gradable evaluation of the ­referent in terms of attitude, as in (13): (13)

e-­be ye-­nye ame-­gã gake e-­zu ame-­gɛ̃ 3sg-­say log-­be person-­aug but 3sg-­become person-­big.dim ‘He says he is the boss but he has become less of a boss’.

This diminutivised adjective is only used with human referents and always has pejorative connotations. In a way, one does not want to refer to the referent as small, nor is one ­interested in calling the referent big (in terms of size). Since the use of the diminutivised ‘big’ is pejorative, the referent is not directly referred to but called behind his or her back. We found one adjective in the language which seems to have been derived in the same way as those discussed above. This adjective has a diminutive reading without an augmentative counterpart. The word kákɛ́ indicates a small part of a whole or a small quantity and might have been derived from the word kaka ‘taking a piece of’. Note the high tone, which marks diminutive on adjectives: (14) mɔlu kákɛ́ aɖe ko-­e le rice small indf only-­ foc be ‘only a little quantity of rice is left’

loo

adpart

In certain pragmatic contexts, kákɛ́ can also be used sarcastically or can express a negative attitude towards the interlocutor or referent. 2.3  Augmentative constructions Unlike most Bantu languages, which mark augmentation morphologically, augmentative morphemes seem to be lacking among members of the Kwa group of the Niger-­Congo language family, especially those spoken in Ghana. Augmentatives represent a marked category as opposed to the unmarked category of diminutives. It is not surprising, therefore, that the use of the adjective ‘big’ gã as an augmentative morpheme-­suffix in Ewe is highly

Ewe 477 restricted to human references. The adjective gã ‘big’ can be said to have grammaticalised into a suffix with only human referents. However, gã seem to have grammaticalised with the names of certain towns in the Volta Region of Ghana, such as Angloga and Anfoega. Non-­human referents will use the adjective form and the resultant meaning is ‘BIG-­SIZE’, as illustrated in (15): (15) me-­ʄle eʋu 1sg-­buy car ‘I bought a big car’

gã big

aɖe indf

The basic meaning of the augmentative -­gã is ‘big’, but depending on the context in which the adjective is used, it may have several other connotations, which include: grown up, boss or leader, senior (in age or occupation) and high status or important personality. These meanings are outlined in the following subsections. 2.3.1  Grown up Just as the diminutive suffix indicates youth, the augmentative suffix marks a person as grown up or as an adult. The following examples refer to male and female adults respectively: (16) a. ŋutsu-­gã aug male-­ ‘male adult’ b. nyɔnu-­gã female-­ aug ‘female adult’ 2.3.2 Boss/Leader The augmentative -­gã also carries the meaning ‘boss or leader’. A person would most often introduce his or her boss or leader as stated in (17): (17)

nye ame-­gã enye 1sg.poss person-­ aug 3sg-­be ‘This is my boss’

sia dem

Compare the utterances in (13) and (17). The speaker cannot say (13) directly to her boss, unless of course she wants to lose her job or position. On the other hand, the utterance in (17) may be uttered whether or not the referent is present. 2.3.3 Seniority The suffix -­gã also connotes seniority, either in age or in occupation. To indicate seniority in age, the augmentative suffix is affixed to a person’s name to distinguish him or her from a younger person of the same name, as also observed for the diminutive suffix -­vi in (7). We have such names as: (18) a. Kɔdzo-­gã name-­ aug ‘older Kodzo’(lit. ‘big Kodzo’)

Yvonne Agbetsoamedo and Paul Kofi Agbedor

478 b. fo-­gã elder.brother-­ aug ‘older of two elder brothers’

Seniority in occupation is also indicated with the augmentative suffix -­gã. (19) a. nufiala-­gã teacher-­ aug ‘senior teacher’ b. gadzikpɔ la-­gã aug accountant-­ ‘senior accountant’ c. lɔya-­gã lawyer-­ aug ‘senior lawyer/experienced lawyer’ 2.3.4  High status The augmentative suffix also carries the expression of ‘high status’ or ‘important personality’. It is normally suffixed to the word ame ‘person’. This is shown in (20). (20)

wo-­kpe ame-­gã gbogbo aɖe-­wo 3pl-­invite person-­aug plenty indf-­pl ‘A lot of important personalities were invited’

3  Ideophones and the expression of evaluation Ideophones are words that depict sensory imagery (Dingemanse 2011, 25). Ewe ideophonic words can fall into any syntactic class of the language: nominal, adjectival, intensifier, verbal, adverbial and interjection (Ameka 2001). Tone plays a very important role in the interpretation of ideophones in Ewe. A low tone on the ideophone gives it an augmentative reading: in other words, a low tone indicates that the entity that the ideophone describes is either ‘big’ in size or unpleasant in taste or smell. In the same way, an ideophone marked with high tone describes an entity as ‘small’ in size or pleasant in taste or smell. For example, the adjective gblɔ̀yì ‘loose and dangling’ is mostly reduplicated to form an adverb, to describe a manner in which a person walks. Consider examples (21a) and (21b): (21) a. Kofi le zɔ̀zɔ̀-­m gblɔ̀yìgblɔ̀yì Kofi be walk-­prog sluggish.sluggish ‘Kofi is walking sluggishly’ b. Kofi le zɔ̀zɔ̀-­m gblɔ́yígblɔ́yí Kofi be walk-­prog sluggish.sluggish ‘Kofi is walking sluggishly’ Both sentences describe an event taking place in a similar manner (i.e. Kofi is walking sluggishly). The only difference between the two sentences is the change in tone. In (20a),

Ewe 479 the ideophonic adverb is marked with a low tone. This depicts ‘Kofi’ as a big person while on the other hand in (20b), ‘Kofi’ is perceived as a small person, compared to the ‘Kofi’ in (20a). Most ideophones in the language behave in similar ways, in the sense that altering the tone pattern brings about different interpretations of the entity described by the ideophone. However, it is important to state here that not all ideophones have a lexical stem from which they are derived, and thus we cannot say that the ‘stem-­less’ ideopheones technically belong to the field of evaluative morphology. Nonetheless, it is interesting to highlight the role tone plays in expressing the different evaluative meanings of the same word.

4 Conclusion In this chapter, we have briefly described evaluative morphology in Ewe. We discussed two ways of expressing evaluative information, namely diminutive and augmentation. The diminutive suffix -­vi is attached to nouns, and has the basic meaning of ‘smallness’. Apart from this basic meaning, -­vi also carries meanings such as ‘young’ (both of animals and plants), ‘insignificance/triviality’ and ‘membership’ of an organisation or nation/town. The suffix -­vi takes its source from the noun vi, which means child or offspring. The other diminutive suffix identified is a high tone -­i, which attaches to an adjective (especially dimensional adjectives), and, together with a high tone on the base, indicates ‘smallness’. Augmentation is marked mainly by the grammaticalised form of the word for big, gã. Just like -­vi, -­gã carries other meanings or connotations apart from its basic meaning of ‘bigness’. These include ‘grown up’, ‘boss/leader’, ‘seniority’ and ‘high status/important personality’. Another area of focus in relation to evaluation is the use of ideophones, which imitate the sound, shape, taste etc. of things and people. Depending on the degree of the size, shape or taste of the entity being described, an ideophone can have a low tone or a high tone. A low tone marks the entity being described as ’big’ in size and shape or unpleasant or disagreeable in taste or smell. A high tone on the ideophone marks the entity being described as either ’small’ in size and shape or pleasant and agreeable in taste and smell. As noted earlier, Ewe does not have a very rich morphology, or for that matter evaluative morphology. Nevertheless, that does not make it a completely isolating language either. Some grammatical categories are marked morphologically, and this includes evaluation. Of course, evaluation does not involve diminutives and augmentatives only. Further research is needed to find out whether other evaluative markings, such as endearment, intensification, attenuation and affection, are also morphologically marked in the language.

Note 1. Where geographic varieties exist, this chapter refers to Ewe as spoken in Ghana.

Kɔnni

18.4 K nni

Michael Cahill

1 Introduction In this chapter, I present data from Kɔnni [kma], a Gur language from northern Ghana which exhibits interesting patterns for the diminutive, especially when compared to its closest genetic relative Buli [bwu]. The diminutive suffix /-­biŋ/ actually consists of two forms which differ both in tone and in which nouns they can suffix to. Synchronically, /-­biŋ/ is not cognate with the present-­day lexeme for ‘child’ in Kɔnni, but one of the two forms is cognate with ‘child’ in Buli. One of the significant issues in evaluative morphology is its inflectional vs. derivational nature. As we will see below, the Kɔnni diminutive is unambiguously derivational. Evaluative morphology deals with morphologically complex forms, but many ‘evaluative’ functions in Kɔnni are treated in the syntax rather than the morphology, and we present a few examples of such before turning to morphology. In most comparative statements, a verb, tɪyaŋ ‘to pass/surpass’, is used in a serial verb construction, and is combined with a verb which expresses a quality. This is consistent with the general pattern in Kɔnni and many other African languages that adjectival functions are often expressed as verbs, for example dʊnsɪ ‘to be heavy’. A literal rendering ‘X is Q passes Y’ denotes that ‘X has more of a particular quality Q than Y does.’ The verbs below are glossed with English verbal morphology, even when this does not reflect the Kɔnni, for example dɪ̀lá, as the present tense of the verb ‘to be big’ is simply glossed ‘be.big’. (1) ‘Be more X than Y is’: a. náá!bʊ́ dɪ́lá tɪ̀yàŋ kpɪ́áŋ cow.the be.big pass chicken ‘the cow is bigger than chicken’ b. ǹ dʊ́nsɪ́ tɪ̀yàŋ ŋ̀ hɔ̀wwá I be.heavy pass my wife ‘I’m heavier than my wife’ c. ŋmínné tùgùré tɪ̀yàŋ chɪ́ɪ́ŋ sun.the be.bright pass moon ‘the sun is brighter than moon’ d. gbàáŋ yá yɪ́!ɪ́ŋ à tɪ́yàŋ náá!gɪ́ŋ dog have intelligence and pass cow ‘dog is smarter than cow’

Kɔnni 481 Note that in (1d) there is an alternative construction, ‘X has noun Q and passes Y’, distinct from the others. In this chapter, I focus on the diminutive -­biŋ in Kɔnni, which is a suffix and thus belongs to the domain of morphology rather than the syntactically based data above.1 To put this morpheme in context, some background on Kɔnni nominal morphology is required first, and the remainder of this section is devoted to that. 1.1  Minimal nouns and noun classes A minimal noun in Kɔnni contains a noun stem and a singular or plural suffix (with the exception of nouns of Class 5, which have a zero suffix). (2) díí-­ŋ díí-­rí dí-­è dí-­é-­!hé forehead-­ sg forehead-­ art forehead-­ pl forehead-­ pl-­art ‘forehead’ ‘the forehead’ ‘foreheads’ ‘the foreheads’ The suffix -­ŋ́ may be regarded as an indefinite singular suffix; it appears on a noun in citation form.2 To make it definite, it is replaced by a definite suffix. To make it plural, it is replaced by the plural suffix. To make it both definite and plural, both the definite and the definite plural suffixes are appended. The specific forms of the suffixes vary by noun class, to which we now turn. The noun class system of Kɔnni (numbered arbitrarily) is indicated by the definite article and plural suffixes, as in (3): (3) Nouns Singular Class 1 /-­ ŋ́/ bee síébíŋ

Sg. def. /-­rÍ/ síébírí

Plural /-­A/ síébíè

Pl. + Def. /-­A-­hÁ/ síébíé!hé

Class 2 /-­ ŋ́́/ courtyard gbàáŋ

/-­kÚ/ gbààkʊ́

/-­tÍ/ gbààtɪ́

/-­tÍ-­tÍ/ gbààtɪ́tɪ́

Class 3 /-­ ŋ́/ man dèmbíŋ

/-­kÁ/ dèmbìké

/-­sÍ/ dèmbìsí

/-­sÍ-­sÍ/ dèmbìsísí

Class 4 /-­ ŋ́́/ meat nɔ̌ŋ

/-­bÚ/ nɔ̀mbʊ́

/-­tÍ/ nɔ̀ntɪ́

/-­tÍ-­tÍ/ nɔ̀ntɪ́tɪ́

Class 5 /-­Ø/ child bʊ̀á

/-­wÁ/ Irreg. Irreg. bàllɪ́lɪ́ bʊ̀áwá bàllɪ́

Mixed (singular definite from one class, plural from a different one) goat bɪ̀ɪ́ŋ bɪ̀ɪ̀kʊ́ bìé bìèhé Capitalised vowels in the suffixes indicate that that vowel harmonises in ATR value with the root, so /Í/, for example, is realised as either [í] or [ɪ́]. All nominal suffixes are High-­ toned, with the exception of Noun Class 1 plural /-­A/, which exhibits polar tone, a tone opposite to the preceding stem’s tone.3

Michael Cahill

482 1.2  Adjectival forms

Adjectives are attached as suffixes to the noun stem. To say ‘thin’, for example, you must say ‘thin thing’, and others are exemplified below. The forms for ‘thing’ in (4) exhibit the allomorphs /jà-­/ for singular and /ɲìŋ-­/ for plural. (4)

‘X thing’ jà-­kʊ̀ʊ́ŋ jà-­bɪ̀áŋ jè-­bíŋ jà-­kpɪ́!ɪ́ŋ jà-­yɛ̀ɛ̀lɪ́ŋ

‘the X thing’ jà-­kʊ̀ʊ̀-­rɪ́ jà-­bɪ̀à-­kʊ́ jè-­bì-­ké jà-­kpɪ́ɪ́-­!ká jà-­yɛ̀ɛ̀lɪ̀-­ká

‘X things’ ɲìŋ-­kʊ̀-­rá ɲìm-­bɪ̀à-­tɪ́ ɲìm-­bì-­sí ɲìŋ-­kpɪ́ɪ́m-­à ɲìn-­yɛ̀ɛ̀l-­á

‘the X things’ ɲìŋ-­kʊ̀-­rá-­há ɲìm-­bɪ̀à-­tɪ́-­tɪ́ ɲìm-­bì-­sí-­sí ɲìŋ-­kpɪ́ɪ́m-­!á-­há ɲìn-­yɛ̀ɛ̀l-­á-­há

Gloss Class (Sg./Pl.) ‘old’ Class 1 ‘bad’ Class 2 ‘small’ Class 3 ‘big’ Class 3/1 ‘white’ Class 3/1

Adjectival morphemes are suffixed to the noun stem and occur before the nominal suffix. It is important to note that the simple noun ‘thing’ jàáŋ belongs to Noun Class 4, as shown by the suffix of the definite singular form jààbʊ́ ‘the thing’. However, the forms in the singular definite column vary, and show that the adjectival lexeme determines the suffix and thus the noun class of the entire word. Two, and rarely three, adjectives may be affixed to a noun stem, as in (5): (5) Two and three adjectives in Kɔnni: a. tígí-­yɛ́ɛ́lɪ́-­!bíŋ house-­white-­small ‘small white house’ b. jà-­kʊ̀-­yɛ̀ɛ̀lɪ̀-­chʊ̀ʊ̀sɪ́!kɪ́ŋ thing-­old-­white-­spoiled ‘old white spoiled thing’ A phonological diagnostic is also in order here. Spreading of high tones is common in Kɔnni. The direction of high-­tone spreading within noun–adjective complexes is from left to right, from the noun stem into the adjective. This is unlike the left-­to-­right high  spread  in almost all other environments in Kɔnni, in which an ­underlying  HLH ­configuration yields surface H!HH, as seen in several cases in (2)–(4). In the specific case of noun–adjective, however, this direction of spreading is reversed. (6) H-­spread from noun into adjective: a. tígí-­yɛ́ɛ́!lɪ́ŋ house-­white ‘white house’ (cf. jà-­yɛ̀ɛ̀lɪ́ŋ ‘white thing’) b. jɔ́rɔ́-­dʊ́n!sɪ́ŋ ladder-­heavy ‘heavy ladder’ (cf. jà-­dʊ̀nsɪ́ŋ ‘heavy thing’) The underlying and surface forms of jɔ́rɔ́-­dʊ́n!sɪ́ŋ ‘heavy ladder’ are given in (7), with underlying tones above the syllables they are sponsored by.

Kɔnni 483

(7) H-spread from noun into adjective: H H-L H | |

|

|

H H L H | |

\

| ‘heavy ladder’

To sum up, the adjective in Kɔnni is suffixed to the noun stem, and controls the noun class of the entire derived noun.

2  The diminutive -­biŋ The adjectival form -­biŋ, translated ‘small’,4 and which attaches to noun stems, has two forms which differ phonetically only in tone and, as discussed below, have different historical sources.5 Note the difference in tonal patterns in (8a) and (8b).6 The singular definite suffix and the plural suffix show that all of these forms, both in (a) and in (b), belong to Noun Class 3. (8) ‘Small’ – two forms: a. L /-­bì-­/: ‘small cleared area’ ‘small cutlass’ ‘small father’ ‘small hat’ ‘small lizard’ ‘small shea tree’ ‘small thing’ ‘small tin can’ ‘small woman’ ‘finger’

b.

HL /-­bí ‘-­/: ‘small broom’ ‘small chicken’ ‘small dog’ ‘small farm’ ‘small horse’ ‘small river’ ‘small (junior) wife’ ‘small arm’

Sg.

Sg. def.

Plural

Noun stem

kɔ́gí-­!bíŋ kàrɛ́ntìè-­bǐŋ chòm-­bǐŋ kàgbàà-­bǐŋ gʊ̀ráá-­!bíŋ té!bíŋ jè-­bǐŋ kɔ́ŋkòm-­bǐŋ hɔ̀gù-­bǐŋ nú!bíŋ

kàgbàà-­bìké jè-­bìké hɔ̀gù-­bìké nú-­!bíké

chòm-­bìsí kàgbàà-­bìsí gʊ̀ráá-­!bísí té!-­bísí ɲìm-­bìsí kɔ́ŋkòm-­bìsí hɔ̀gù-­bìsí nú!bísí

/kɔ́g-­/ /kàrɛ́ntɪ̀à-­/ /chɔ̀N-­/ /kàgbàà-­/ /gʊ̀ráá’-­/ /táá-­/ /jàà-­, ɲìN-­/ /kɔ́ŋkɔ̀N-­/ /hɔ̀gʊ̀-­/ /núú-­/

súm-­bí!ŋ́ kpìè-­bí!ŋ́ gbè-­bí!ŋ́ kúó-­bí!ŋ́ dùùm-­bí!ŋ́ múgú-­bí!ŋ́ hɔ̀gù-­bí!ŋ́ nú-­bí!ŋ́

kpìè-­bí!ké gbè-­bí!ké kúó-­bí!ké múgú-­bí!ké nú-­bí!ké

súm-­bísì kpìè-­bísì gbè-­bísì kúó-­bísì múgú-­bísì nú-­bísì

/súN-­/ /kpɪ̀à-­/ /gbàà-­/ /kʊ́á-­/ /dùùN-­/ /mʊ́gʊ́-­/ /hɔ̀gʊ̀-­/ /núú-­/

Since the final high tone of the words is due to suffixal -­ŋ́ as in citation forms of nouns in (3), the forms in (a) can be analysed as having an L tone for the -­bi-­stem, and those in (b) as having am HL tone for the -­bi-­stem. The (a) forms are more common, but (b) forms are far from rare. There is no correlation between the phonological shape of a noun and which form of ‘small’ it takes. Changes in vowels between the ‘small’ form and the underlying noun stem are due to leftward ATR spread from the adjective ‘small’. Example (9) illustrates the tonal arrangements of the two forms of -­biŋ:

Michael Cahill

484

Note that both low-­toned and high-­toned noun stems occur with both forms of ‘small’. (9) a. Tonal derivations for L-toned /-bi-/, after L-and H-toned noun stems: L LH

L LH

|

|

| |

H L H

\/

|

HLH

| |

|

|

(b) Tonal derivation for HL-toned /-bi-/ (similar in both L-and H-toned noun stems): H HL H |

|

H HLH

|

|

\ /

Note also that the plural forms in (8b) have the low-­toned form of the plural suffix -­sì, while the forms in (8a) have the high-­toned form -­sí, though the plural forms will not be further discussed here. Besides the difference in tones, these forms are shown to be distinct by other lines of evidence. A noun stem will select a particular form of ‘small’ to conjoin with; matching a noun with the other form will be incorrect. One of my Kɔnni language consultants made a mistake once in pronouncing the wrong form, *kɔ̀b-­bí!ŋ́, rather than the correct kɔ̀b-­bǐŋ ‘small bone’. He immediately corrected himself. Also, there are pairs of words which differ only in the form of ‘small’ which is chosen, but this choice makes a difference in meaning. Two examples occur in (8). First is hɔ̀gù-­bǐŋ, which is the normal construction for ‘small woman’, and hɔ̀gù-­bí!ŋ́, which is a specific relational term ‘small (junior) wife’ for a junior wife in a polygamous marriage. Similarly, nú!bíŋ ‘finger’ includes the L form of ‘small’, while nú-­bí!ŋ́ ‘small arm’ includes the HL form, as in (10). (10) Tonal derivations for nú!bíŋ ‘finger’ and nú-bí!ŋ´ ‘small arm’: H L H |

|

|

H LH →

nu-bi-ŋ

arm-smallL-ART

|

/

nu-biŋ

H HL H |

|

|

H →

nu-bi-ŋ

|

HLH | /

nu-biŋ

arm-smallHL-ART

Buli is the language most closely related to Kɔnni (Naden 1989; Lewis, Simons and Fennig 2014); the two languages form a subgroup of their own. The historical connection with Buli illumines the different Kɔnni forms. Buli has cognates for both the ‘small’ forms in Kɔnni, but in Buli they also exhibit segmental differences. The first pair of forms, which we will connect to (8a), are biik (def. biika, pl. bisa), and bili or bilik (def. bilika, pl. bilisa).7

Kɔnni 485 (11) Buli forms (from Kröger 1992; cf. (8a)): ‘small’ ‘small thing’ ‘the small thing’ a. -­biik jàà-­bìīk jàà-­bììk-­ā small thing-­small thing-­small-­def b. -­bili jàà-­bìlīk jàà-­bìlìkā small thing-­small thing-­small-­def Kröger notes that the latter (11b) is used only with a restricted set of nouns (for example pò-­bìlī ‘junior wife’). In Buli, then, there are two similar forms, of which the first is more general. Tonally, these both have a rising tone (low to mid) when following a low-­toned noun, and in that way resemble the forms in (8a) above.8 A reasonable hypothesis is that Proto-­Buli-­Kɔnni had both the forms in (11). Kɔnni merged these two adjectives (11a–b) into the L-­toned -­bi-­, giving the pattern in (8a). Besides this, there is another Buli candidate to serve as a model for the initially high-­ toned Kɔnni forms in (8b). This is the high-­toned bíík ‘child’ (def. bííká, pl. bísá). Cross-­ linguistically, ‘child’ is often adapted to mean ‘small’ (Jurafsky 1996). Kɔnni, as the more innovative language of the two (Cahill 1997), in all probability adopted the noun bíík ‘child’ as an adjectival form with prototypical meaning ‘small’, giving the forms in (8b). The word for ‘child’ in present-­day Kɔnni, bʊ̀á (with irregular plural bàllɪ́), appears to be not cognate at all with the Buli forms for ‘child’. Besides relating Kɔnni ‘small’ to Buli ‘child’, there is an additional connection within Kɔnni to yet another similar-­sounding lexical item. This is the noun bíŋ ‘seed’, which combines with many nouns, for example ɲʊ̀ʊ̀-­bíŋ ‘small intestine’ (lit. ‘stomach-­seed’, plural ɲʊ̀ʊ̀-­bíè), gàmbíŋ ‘front rib’ (lit. ‘rib-­seed’, plural gàm-­bíè. It is shown to be a different lexeme from the ‘small’ forms discussed above by three lines of evidence. First, ‘seed’ occurs as an independent noun, as in (12). Second, as seen from the plural form, ‘seed’ is in Noun Class 1, not Noun Class 3, as ‘small’ is. Third, the tone is unambiguously high, not low or high-­low, as the forms of ‘seed’ are. (12) Seed: Singular Sing. def. Plural Kɔnni bíŋ bínní bíè Buli bírí bíní bíé

Pl. + Def. bíé!hé

As seen from the cognate Buli forms, this lexeme ‘seed’ is robust in the sub-­family. It is definitely a noun, so that ‘small intestine’ and ‘front rib’ are not noun–adjective complexes, but rather compound nouns. This is also shown by the lack of ATR spread from the bíŋ into the preceding noun. (13) a. Spreading of ATR from adjective into noun stem: ! kpíé-­ bíŋ ‘small chicken’ (cf. kpɪ̀áŋ ‘chicken’) kúó-­bí!ké ‘the small farm’ (cf. kʊ́áŋ ‘farm’) tèm-­bí!ké ‘the small stone’ (cf. tǎŋ ‘stone’) bʊ̀ntùù-­bǐŋ ‘small toad’ (cf. bʊ̀ntʊ̀ʊ̀kpɪ́!ɪ́ŋ ‘big toad’) kɔ́ŋkòm-­bǐŋ ‘small tinv can’ (cf. kɔ́ŋkɔ̀ŋ ‘tin can’)

b. No spreading of ATR from noun 2 into the first noun stem: ɲʊ̀ʊ̀-­bíŋ ‘small intestine’ (cf. ɲʊ̀ʊ́ŋ ‘stomach’)

Michael Cahill

486

gàm-­bíŋ ǹ nɪ̀mbʊ́á!bíŋ ǹ táá!bíŋ

‘front rib’ (cf. gànɪ́ŋ ‘rib’) ‘my youngest sibling’ (cf. ǹ nɪ̀mbʊ́à ‘my younger sibling’) ‘my younger sister’ (cf. ǹ táà ‘my sister’)

This ATR behaviour is one way to tell the difference between bíŋ ‘seed’ and -­biŋ ‘small’. The latter, as an adjective, spreads its ATR value to a preceding noun in (13a), and the former, as a noun within a compound, does not spread its ATR in (13b).

3 Conclusion The Kɔnni morphemes -­biŋ ‘small’ fit into the adjectival position in Kɔnni. A striking feature of these is that there are two adjectives which are almost identical in form and function, but can be distinguished by details of their tonal behaviour, by their semantic collocations – that is, which nouns they are allowed to attach to – and by their historical antecedents when compared to Buli. A further complexity is the similar-­looking noun biŋ ‘seed’, which also appears to be at times loosely translatable as ‘small’, but again is quite distinct in its syntactic, morphological and phonological behaviour. Syntactically, it functions as an independent noun. Morphologically, it belongs to Noun Class 1. Phonologically, it does not spread the feature ATR into a previous noun stem. An application from Kɔnni to other languages, therefore, is that what may appear to be surface identity and a single morpheme may, upon closer examination, turn out to be a much more complex situation.

Notes 1. Another type of diminutive in Kɔnni uses the free word bɪ̀tá ‘small’. This lexeme has a wide semantic range, as in chììsé bɪ̀tá ‘wait a little’ and víí !bálí bɪ̀tá-­!bɪ́tá ‘say again slowly’ (or ‘more quietly’). 2. What I have labelled as the singular definite suffix here corresponds to the noun class suffix itself in other Gur languages. Proto-­Buli-­Kɔnni has undergone a restructuring of the noun class system, and Kɔnni uniquely has merged almost all singular indefinite forms into the suffix -­ŋ́ (Cahill 1997). 3. Various consonantal and vocalic epenthesis and deletion processes are to be seen in the data here, but are not relevant to the issues in this chapter. For further treatment of the polar tone, see Cahill (2004), and for a fuller treatment of other phonological processes, see Cahill (2007). 4. The Kɔnni diminutive ‘small’ is the focus of this chapter. It can be contrasted with the augmentative, the adjective ‘big’, as in (4), but there is little to say about the augmentative other than the bare fact of its existence. 5. Synchronically, one might be tempted to call these allomorphs. But as seen further in this section, there is no synchronic phonological conditioning, there are cases when substituting one form for the other changes the meaning, and which form is chosen for a particular noun stem is solely dependent on the noun. The relationship between the two forms deserves more investigation. 6. Only data actually attested is listed in (8a, b), although in most cases it is clear what the missing form would be expected to be. 7. Besides high and low tones, Buli also has a mid tone, represented by a macron, as in ā. In many cases this corresponds to a Kɔnni high tone. 8. The rising tone and internal morphology in Buli bɪ̀īk and Kɔnni bǐŋ are synchronically different, but have a common predecessor.

18.5  Sεlεε

Yvonne Agbetsoamedo and Francesca Di Garbo

1 Introduction 1.1  The language and the community Sɛlɛɛ is a Niger-­Congo language of the Kwa subgroup. Traditionally, it is classified as one of the fourteen Ghana-­Togo Mountain (henceforth GTM) languages which, to date, are only sparsely described. The status of GTM languages within Kwa is questioned by some scholars (see e.g. Blench 2009), who claim that they should be considered as an independent group within Niger-­Congo. Sɛlɛɛ is spoken by the people of Santrokofi, a small community located in the Volta Region of Ghana. They refer to themselves as Balɛɛ. According to the census report from the Ghana Institute of Linguistics, Literacy and Bible Translation (GILLBT), in 2002 the population of the Balɛɛ amounted to about 11,300 people, with a growth rate of 3.5 per cent. The community is largely multilingual, and the major second language spoken in the Santrokofi area and used as a lingua franca is Ewe. Some Balɛɛ also speak Akan and English. The data for this study were collected during two field trips to the language community conducted by Yvonne Agbetsoamedo between 2010 and 2013. The first trip took place in August 2010: data collection on that occasion was restricted to diminutives. The results from a preliminary investigation into diminutive constructions in Sɛlɛɛ are presented in Agbetsoamedo (2011). The second trip took place in January 2013 and was aimed at collecting information on the encoding of augmentatives. On both occasions, data collection was conducted by combining elicitation sessions and focus group discussions, where speakers were asked to discuss various ways to refer to animate and inanimate entities with respect to evaluation of quantity and quality. The chapter is structured as follows: an overview of nominal morphology in Sɛlɛɛ is provided in Section 2, whereas the evaluative constructions available in the language are discussed in detail in Section 3. A discussion of the findings in the light of the theoretical framework of the present volume is found in Section 4.

2  Nominal morphology: an overview Sɛlɛɛ is a predominantly agglutinative language. One of the most relevant characteristics of Sɛlɛɛ nominal morphology is its noun class system. This is based on prefixes, which usually carry low or mid tone.1 Noun class prefixes are in most cases overtly coded

Yvonne Agbetsoamedo and Francesca Di Garbo

488

Table 18.5.1. Noun class markers and agreement targets in Sɛlɛɛ Noun class

Prefix

Demonstrative

Numeral

Interrogative pronoun

1 2 3 4 5 6 7 8

o-­/ɔ-­/ϕ ba-­ ka-­ se-­/sɛ-­/si-­ di-­/li-­/ni-­/le-­/lɛ-­ n-­ ko-­/kɔ-­/ku-­ a-­

wɔ ba ka se le be ko ya

o-­ ba-­ ka-­ e-­ ni-­ n-­ ko-­/kɔ-­/ku-­ a-­

ɔ-­ ba-­ ka-­ sɛ-­ lɛ-­ m-­ kɔ-­ a-­

Figure 18.5.1  Noun class pairs in Sɛlɛɛ

on nouns and carry agreement with the following targets: demonstratives, numerals and interrogative pronouns. There is no noun class agreement with adjectives and verbs. The individual noun classes are either singular or plural. In line with the Niger-­Congo tradition, the individual noun classes are paired according to the type of number value that they convey. Thus, a noun class pair consists of one singular class and one plural class. This is referred to as gender. Table 18.5.1 illustrates the noun classes and agreement targets of Sɛlɛɛ. Singular–plural pairs are illustrated in Figure 18.5.1. Cardinal numbers are used for the numbering of the individual classes: the singular classes are odd-­numbered whereas the plural classes are even-­numbered. The only exception is class 7, which functions as the singular class of gender 7/8, and the plural class of gender 3/7 (see Figure 18.5.1). In Figure 18.5.1, the solid lines represent the most frequent genders whereas the dashed lines stand for the less frequent ones. Following Corbett’s (1991) terminology, we refer to the least frequent genders as inquorate genders. Noun class semantics in Sɛlɛɛ is generally not transparent. However, certain patterns can be identified. For example, gender 1/2 is mostly used with human nouns and is also the default gender of all borrowings. Moreover, nominalised verbs are assigned to class 1 and nominalised adjectives to class 5.

Sεlεε 489 Table 18.5.2. Semantics of the Sɛlɛɛ noun classes Gender

Semantics

1/2 5/8 7/8 3/6 1/4 3/7 7/6 1/8

Human terms (identity, kinship), some animals, borrowed words Things and food with circular, cylindrical or concave shape. Animal offspring Farm and farm-­related concepts Most external body parts, mass nouns Domain of some human experience, some plants (edible and non-­edible) The noun for ‘fish’, diminutives Limbs: hand and leg The nouns for ‘running stone’ (stone traditionally used for grinding) and ‘corn’

The semantic underpinnings of noun class systems can be depicted when looking at mechanisms of noun class shifts, whereby nouns change their gender when more marked construals of their referents are intended (for a general discussion on the semantics and function of noun class systems across the languages of the world see Aikhenvald 2003; Corbett 1991). In Sɛlɛɛ, the only productive type of noun class shift is connected with the formation of diminutive nouns; this will be illustrated in detail in Section 3.1.1. Table 18.5.2 shows some of the semantic domains associated with each noun class pair in Sɛlɛɛ. Another strategy of nominal derivation is suffixation. The nominal suffixes of Sɛlɛɛ are also strictly connected with the formation of diminutives and are discussed in detail in the next section.

3  Evaluative constructions Evaluative constructions in Sɛlɛɛ can be both morphological and analytical. However, morphological evaluative markers exist only in the case of diminutives. There are no augmentative morphemes, but augmentative-­like meanings are always expressed analytically. 3.1  Diminutive constructions 3.1.1  Diminutive morphemes There are two constructional patterns for the morphological encoding of diminutives in Sɛlɛɛ: • gender shift + suffixation; • suffixation Genders 5/8 and 3/7 (see Table 18.5.1 and Figure 18.5.1) are very productive in the formation of diminutives.2 The diminutive suffixes are: -­bi, -­mii, -­ɛ and -­nyi. In our corpus, the shift to gender 5/8 can combine only with the suffix -­bi. This combination is very productive either as a means of expressing small size or, when the base noun refers to an animal, as a strategy for encoding offspring. The two uses are illustrated in (1) and (2). (1) a. bloblo b. lɛ-­bloblo-­bi cl1.bread cl5-­bread-­dim ‘bread’ ‘small bread/slice of bread’

490

Yvonne Agbetsoamedo and Francesca Di Garbo

(2) a. ɔ-­kla b. lɛ-­kla-­bi cl1-­cat cl5-­cat-­dim ‘cat’ ‘kitten’ There are also a few cases where the suffix -­bi co-­occurs with the shift to gender 3/7. This is shown in (3). (3) a. tie b. ka-­tie-­bi cl1.goat cl3-­goat-­dim ‘goat’ ‘tiny goat/lean goat’ The suffixes -­mii and -­ɛ can co-­occur only with the shift to gender 3/7. This is illustrated in (4) and (5). (4) a. kansiɛ b. ka-­kansiɛ-­mii cl1-­bird cl3-­bird-­dim ‘bird’ ‘small bird’ (5) a. lɛ-­pɔtrɔ b. ka-­pɔtrɔ-­ɛ cl5-­toad cl3-­toad-­dim ‘toad’ ‘small toad’ This strategy seems to be more frequent with animal names where it encodes small size, as opposed to the strategy exemplified in (2) which encodes young age. In our corpus, the suffix -­nyi is attested with only two nouns: oti ‘person’ and kɔnɛɛ ‘hand’. In both cases, the occurrence of the suffix is associated with the shift to gender 3/7. The resulting meaning is derogatory, as shown in (6) and (7). (6) a.  o-­ti b. ka-­ti-­nyi cl1-­person cl3-­person-­dim ‘person’ ‘tiny person’ (7) a. kɔ-­nɛɛ b. ka-­nɛɛ-­nyi cl7-­hand/arm cl3-­hand/arm-­dim ‘hand/arm’ ‘tiny hand/arm’ The suffix -­bi is the only diminutive suffix which can occur on its own, that is without any shift in gender. This pattern is attested both with nouns assigned by default to gender 5/8, and with nouns assigned to other genders, as illustrated in (8) and (9).

Sεlεε 491 (8) a. di-­bulaa b. di-­bulaa-­bi cl5-­onion cl5-­onion-­dim ‘onion’ ‘small onion’ (9) a. o-­suɔtɔ b. o-­suɔtɔ-­bi cl1-­man cl1-­man-­dim ‘man’ ‘boy’ Among the morphological diminutives discussed in this section, those involving -­bi (both the pattern ‘gender shift + suffixation’ and simple suffixation) are the most productive. There is also reason to believe that -­bi is the only semantically transparent diminutive marker in Sɛlɛɛ. This observation will be accounted for in Section 3.1.3. 3.1.2  Semantics and usages of the diminutive morphemes Diminutive constructions involving -­bi have some peculiar usages. For example, when the constructions ‘shift to gender 5/8 + -­bi’ and ‘-­bi’ are used with uncountable nouns, they function as quantifiers. The resulting meaning is that of ‘a small quantity of X’ or ‘a part of X’, as exemplified in (10). (10) a. ka-­mɔɔ b. a-­mɔɔ-­bi cl3-­rice.with.chaff cl8-­rice.with.chaff-­dim ‘rice with chaff’ ‘small quantity of rice with chaff’ This is cross-­linguistically a very common aspect of the semantics and functions of diminutive markers (Jurafsky 1996). There also exist some sorts of lexicalised diminutive forms, where the presence of -­bi does not correspond to a transparent diminutive semantics. These are exemplified in (11) and (12): (11) lɛ-­wɛntrɛ-­bi cf. a-­wɛntɛ cl5-­star-­dim cl8-­moon ‘star’ ‘moon’ (12) lɛ-­kpɛɛ-­bi cf. kɔ-­kpaa cl5-­leg-­dim cl7-­leg ‘toe’ ‘leg’ In (11), the word for ‘star’ differs from the word for ‘moon’ not only in the presence of the suffix -­bi. Besides the stem and the fossilised diminutive, in fact, there is a morpheme -­rɛ-­ whose semantics and function are impossible for us to pin down. Similarly, there seems to be some kind of vowel alternation between the stem for ‘toe’ and that for ‘leg’.

Yvonne Agbetsoamedo and Francesca Di Garbo

492

3.1.3  Origins of -­bi in the context of diminutive morphemes The diminutive suffix -­bi originated from the noun obi ‘child’, which belongs to gender 1/2. The noun still exists in the language as an independent lexical item: (13)

o-­bi ko-­fiɛ cl1-­child 3pk.rp-­sick ‘a child is sick’

When juxtaposed to other nouns, obi is used to express notions such as ‘human offspring’ or ‘membership’. This is illustrated in (14) and (15): (14) ɔ-­ka o-­bi cl1-­chief cl1-­child ‘prince/princess’ (15)

o-­kpoo cl1-­town ‘citizen’

o-­bi

cl1-­child

Thus, the origin of the diminutive suffix -­bi conforms to the most common grammaticalisation path for diminutive constructions across the languages of the world (Jurafsky 1996). The data suggest that the use of -­bi as a diminutive marker was originally restricted to animal nouns and to the encoding of ‘young age’. Later on, the suffix got extended to the expression of ‘smallness’ with inanimate entities. This is also cross-­linguistically very common. The origin of the other three suffixes associated with diminutive formation in Sɛlɛɛ, namely -­mii, -­ɛ and -­nyi, is not clear. What we observed in our corpus is that, as discussed in Section 3.1.2, these suffixes are restricted in use, since they occur only to express small size. This, together with the fact that the three suffixes are generally semantically more opaque and less productive than -­bi, leads us to suggest that the latter suffix might be diachronically more recent. As illustrated in Section 3.1.1, the use of the diminutive suffixes in Sɛlɛɛ is almost always combined with gender or class shift. None of these genders can, however, express diminution independently from suffixation. This brings us to an interesting consideration on the nature of Sɛlɛɛ diminutive morphology in the context of other Niger-­Congo languages. In most Bantu languages, for example, the encoding of diminutives is performed by means of gender shifts that are not supported by additional suffixation. Some genders are semantically inherently associated with evaluative meanings, such as gender ka/tu in Shona, which is almost exclusively used to form diminutives (see Déchaine, Girard, Mudzingwa and Wiltschko, this volume). In some of the south-­eastern Bantu languages, such as Venda, Tswana and Zulu, diminutive suffixes have developed from the Proto-­Bantu word for ‘child’ *jana (Creissels 1999). In some languages, these suffixes have replaced gender shifts in the formation of diminutives. In other languages, gender shift and suffixation may co-­occur on one and the same noun, very similarly to what is found in Sɛlɛɛ. For example, in the Bantu language Venda, diminutives are encoded by: (i) gender shift to

Sεlεε 493 gender 7/8 or 20/8; (ii) the suffix -­ana; (iii) a combination of gender shift and suffixation. When only gender shift is used, the resulting meaning is ‘small size’. When gender shift and suffixation co-­occur, the resulting meaning is ‘very small, tiny size’. Examples (16a), (16b) and (16c) illustrate this contrast: (16) Venda (Niger-­Congo, Bantu) (Poulos 1990, 87): a. khali b. thi-­kali c. thi-­kal-­ana cl9-­clay.pot cl7-­clay.pot cl7-­clay.pot-­dim ‘clay pot’ ‘small clay pot’ ‘very small clay pot’ From such comparative evidence, it is plausible to postulate a similar pattern of development for the diminutive morphemes of Sɛlɛɛ. Gender shifts might have been the original strategy for the encoding of diminutives in the language. This was later on combined with a set of diminutive suffixes. Interestingly, the most recent and productive among these suffixes, -­bi, is used without any gender shift when the diminutive of inanimate nouns is encoded. 3.1.4  Analytical diminutives In Sɛlɛɛ, the ideophonic adjectives3 meaning ‘small’, namely pìtìpìtì/pítípítí, bììbìì/bííbíí, wɛ̀kyɛ̀wɛ̀kyɛ̀/ wɛ́kyɛ́wɛ́kyɛ́, kɛ̀ɛ̀kɛ́ɛ̀ and kafíí, function as regular noun modifiers. However, there are cases where combinations of the type ‘SMALL + noun’ also convey evaluative meanings. We refer to the latter constructions as analytical diminutives. The reason that leads us to classify such constructions as an instance of evaluation rests on a peculiar pattern of interaction between the adjectives for ‘small’ and tone. Tonal alternation is used to express the speakers’ positive or negative attitudes towards the referents, and adds to the default meaning of the adjectives. High tone is associated with derogatory meanings whereas low tone expresses a positive attitude. Such usage of tonal alternation is restricted to the adjectives meaning ‘small’, and is not found anywhere else in the language. It is important to mention that what we call ‘analytical diminutives’ are constructions which do not involve morphology at all, and thus are peripheral to the domain of evaluative morphology. In fact, the adjectives mentioned above are stems in Sɛlɛɛ: words like *piti or *wɛkyɛ etc. are ungrammatical. The only arguable case might be biibii, since we know that obi is the word for ‘child’. The evaluative meanings encoded by analytical diminutives are attenuation, endearment and contempt. Example (17) illustrates the use of analytical diminutives in a typical ‘offer-­context’: (17) am a-­pìtìpìtì nwu ko nin-­te mi lɛ-­ta fɔ nɛ. 1sg cl8-­small def only 3pl-­lie 1sg 1sg.rp-­give 2sg part ‘I have only something small to offer you.’ (lit. ‘Only something small lies, I give you.’) The speaker attenuates the amount of food that he is presenting to one of the chiefs of the area during a festivity. The use of the adjective pìtìpìtì with low tone does not diminish the referent quantitatively but rather does so qualitatively. It is a strategy used to avoid boasting and showing off about the present. Example (18) illustrates a case in which the adjective meaning ‘small’– the same as in (17) – carries high tone. The resulting meaning is pejorative:

Yvonne Agbetsoamedo and Francesca Di Garbo

494 (18)

a-­mankani-­bi pítípítí ya-­mmle fa-­a-­tɛɛ nɛ? cl8-­cocoyam-­dim small cl8-­dem 2sg-­dp-­give.me it ‘Did you have to give me the very small cocoyams?’

3.1.5  Strategies for the pragmatic reinforcing of diminutives Some of the diminutive markers in Sɛlɛɛ, both morphological and analytical, can be graded by means of phonological strategies such as vowel lengthening and reduplication. These have the effect of intensifying the pragmatic force of the diminutive marker. Speakers seem to choose between the intensified and non-­intensified form depending on their proximity to the addressee and on the utterance context. The intensified forms of the diminutive suffix -­bi, namely -­biibii and -­bii, are exemplified in (19) and (20): (19) a. lɛ-­mɔɔ-­bi b. lɛ-­mɔɔ-­bii cl5-­rice.with.chaff-­dim cl5-­rice.with.chaff-­dim ‘a grain of rice’ ‘a tiny grain of rice’ (20) a. di-­si-­bi b. di-­si-­biibii cl5-­head-­dim cl5-­head-­dim ‘a small head’ ‘a very small head’ The most common way of expressing intensification with analytical diminutives is by reduplication. The adjective meaning ‘small’ is reduplicated as many times as the speaker deems it appropriate. For example, the adjective wɛ́kyɛ́wɛ́kyɛ́ can be reduplicated as wɛ́kyɛ́wɛ́kyɛ́wɛ́kyɛ́wɛ́kyɛ́ to mean ‘very very very tiny or small’. Speakers report on the fact that the reduplicated forms of biibii are pragmatically stronger than the reduplicated forms of wɛ́kyɛ́wɛ́kyɛ́ and pitipiti. All adjectives for ‘small’ conform to the reduplication pattern except kafii. When they want to intensify an analytical diminutive with kafii, speakers combine it with the synonymous adjective biibii. So far, the occurrences of intensified analytical diminutives in our corpus express only negative attitudes. It will be interesting to find out whether the intensification function of the analytical diminutives can also express positive attitudes towards both the diminutivised referents and the interlocutor. 3.2  Augmentative-­like constructions There is no grammaticalised strategy for the encoding of augmentatives in Sɛlɛɛ. The property words kplɛ ‘big’ and muɔ ‘to be big, to grow’, besides referring to the size of a referent, are sometimes used by speakers to express appreciation or derogation. Tonal alternation does not play any role in the manipulation of the meaning of kplɛ and muɔ. Thus, the use of these words cannot be regarded as an analytical evaluative construction as defined in Section 3.1.4.

Sεlεε 495

4 Conclusion In this chapter, the inventory of the evaluative constructions of Sɛlɛɛ has been described. We have shown that there exists a fundamental asymmetry in the way in which evaluation is encoded in the language. Sɛlɛɛ has a rich inventory of diminutive constructions which are both morphological and analytical. In contrast, there are no morphological (or even analytical) augmentatives in the language. This asymmetry reflects the general cross-­ linguistic pattern whereby diminutives are the most frequent and least marked component of the grammatical domain of evaluation.

Notes 1. There are three level tones in Sɛlɛɛ: low, mid and high. They function both grammatically and lexically. 2. Many nouns are assigned to gender 5/8, whereas only one noun is assigned by default to gender 3/7: kakpaku/kɔkpaku ‘fish’. 3. In Sɛlɛɛ, most property words are ideophonic. Ideophonic adjectives have reduplicative stems (e.g kɛɛkɛɛ ‘small’, klekle ‘holy’) or stems that are characterised by the presence of long vowels (buu ‘wet’, sɔɔ ‘dry’).

Rose-­Marie Déchaine, Raphaël Girard, Calisto Mudzingwa, Martina Wiltschko

18.6 Shona

Rose-­Marie Déchaine, Raphaël Girard, Calisto Mudzingwa and Martina Wiltschko

1 Introduction Shona, a southern Bantu language (Guthrie Zone S10) spoken in Zimbabwe, Zambia, Botswana and Mozambique, is SVO and agglutinative, and has a two-­tone (high, low) system. Five dialects are recognised (Karanga, Korekore, Manyika, Ndau and Zezuru); this discussion is based on Karanga. Following Fortin (2011), we take evaluatives to be expressives (Potts 2007) whose interpretation is independent of descriptive content. As expressives, Shona evaluatives show the following properties: (i) they are predicated of the utterance situation; (ii) they present speaker-­oriented evaluation; (iii) they are not paraphrasable; (iv) they achieve their content by being uttered; and (v) they are sensitive to context of use. Evaluatives are from a set of N-­class prefixes that mark grammatical class; this holds of all Bantu languages (Maho 1999). Table 18.6.1 lists Shona evaluative prefixes, which define three paradigms: augmentative, diminutive and honorific. (Bantu scholarship often ignores honorifics.) Evaluatives range over size and affect, which we model as scalar properties with one of three values (Fortin 2011): normal/neutral (0), positive (>0) or negative ( naag-­ó (suffixation and gender change) woman.f.sg.abs woman.m.abs-­pl ‘woman’ ‘women’ b. albaáb => albaabb-­ó (suffixation and gender change) door.m.sg.abs door.f.abs-­pl ‘door’ ‘doors’ c. mádax => madáx (prosodic shift and gender change) head.m.sg.abs head.f.pl.abs ‘head’ ‘heads’ d. míis => miis~s (reduplication) table.m.sg.abs table.m~pl.abs ‘table’ ‘tables’ Adjectives, in turn, do not belong to a clearly defined category in Somali. Items such as yár ‘small’ and wéyn ‘big’ are best interpreted as state verbs displaying a particular defective paradigm (Andrzejewski 1969; Banti 1988b; 2001; Mansuur 1988, 95–6; Mous 2012). The adjectival paradigm is formed of an adjectival root and the inflected forms of the reduced paradigm of the verb yahay ‘to be’. A reduced paradigm is characterised by reduced distinctions in subject marking. Reduced present forms are identical to the root, whereas past forms display distinct inflectional endings.6 Saeed (1993, 185–202) claims that there are two sorts of adjectives: ‘basic adjectives’ (a small number), such as yár ‘small’ and wéyn ‘big’, and those formed from nouns and verbs by addition of lexical suffixes, such as caan-­sán ‘famous’ (cf. cáan ‘fame’), wanaag-­sán ‘good’ (cf. wanáag ‘goodness’) and jar-­án ‘chopped’ (cf. jár ‘to break’).7 All types follow the head noun and form a relative clause: (3) Adjectives: a. shalay baa rí-­dii yesterday foc goat-­ def.f.sg.nom cad(d)-­áyd la qash-­ay. be.white-­ pst.3sg.f impers kill-­ pst.3sg.f ‘Yesterday, the white goat (the goat which was white) was killed.’ (Banti 1988b, 209)

Somali 509 b. áqal-­ka cád-­Ø house-­ def.m.sg.abs be.white-­ prs ‘the white house (a house which is white)’ (Saeed 1993, 189) In (3a), the adjective is inflected at the 3F past form, whereas in (3b) the same adjective takes the reduced present form. As a general feature of Lowland East Cushitic, NPs are head-­initial and the unmarked clause order is SOV. In addition, (Northern) Somali main clauses display obligatory focus marking, as shown in (4):8 (4) Obligatory focus marking: a. Gaadhí cusúb-­ Ø b-­àan car.m.sg.abs be.new-­ prs foc-­pro.1sg sóo iibsad-­ay deiprep buy-­ pst.1sg ‘I bought a new car’ b. Nín-­kii gaadhí-­gíi cusb-­áa man-­ def.m.sg.abs car-­ def.m.sg.abs be.new-­ pst.3sg.m b-­úu wat-­ay foc-­pro.3sg.m drive-­ pst.3sg.m ‘the man drove the new car’ (Saeed 1993, 222–3) As in (3a) above, the adjective cusúb ‘new’ is inflected in the past and is syntactically rendered by the following relative clause: ‘the car which was new (at time of driving)’ (see 4b). Adjectives are pluralised only through reduplication.9 The reduplicated plural is formed by copying the first consonant, first vowel and second consonant (a CVC-­sequence) to the left of the stem. Only the last vowel bears the high tone.10 A few examples are shown in (5): (5) Pluralisation of adjectives: a. cád => cad~cád ‘white’ white pl~white ‘white’ ‘white’ (plural agreement) b. cusúb => cus~cusúb ‘new’ new pl~new ‘new’ ‘new’ (plural agreement) c. fudúd => fud~fudúd ‘easy, light’ easy pl~easy ‘easy’ ‘easy’ (plural agreement) d. yár => yar~yár ‘small’ small pl~small ‘small’ ‘small’ (plural agreement) e. wéyn => waa~wéyn ‘big’ big pl~big ‘big’ ‘big’ (plural agreement)

510

Nicola Lampitelli

Finally, plural agreement is not mandatory in adjectives, as the examples in (6) show: (6) Plural agreement in adjectives: a. nimán-­ka wanaag-­sán man.m.pl.abs-­det.abs to.be.good-­ adj ‘good men’ b. nimán-­ka wan~wanaag-­sán man.m.pl.abs-­det.abs pl~to.be.good-­adj ‘good men’ (Saeed 1993, 195) In the next section, I present evaluative morphology in nouns.

3  Evaluative morphology in nouns As mentioned in Section 1, evaluative morphology in nouns consists of a syntactic ­configuration involving a noun followed by an adjective. Such analytic constructions are beyond the scope of this chapter. For this reason I will show only two examples, which best illustrate the entire typology of evaluation in nouns: (7) Quantitative evaluation in nouns: a. gúri yár house.m.sg.abs small ‘a small house’ b. gúri wéyn house.m.sg.abs big ‘a big house’ These constructions are extremely productive and are used to express quantitative evaluation. More precisely, yár-­constructions express diminution and, in a few cases, age variation, whereas wéyn-­constructions are used for augmentation although not for age variation (cf. sán yár/wéyn ‘a small/big nose’, búug yár/wéyn ‘a small/big book’, wíil/ínan yár ‘a young boy’, etc.).11 The syntactic construction N + yár/wéyn can have a qualitative interpretation, although this seems to happen in a very few cases: (8) Qualitative evaluation in nouns: a. wáx yár instant.m.sg.abs small ‘about an instant’ b. daqiíq yár minute.f.sg.abs small ‘about a minute’ These examples correspond roughly to Spanish hor-­it-­a ‘hour-­ dim-­f’, moment-­it-­o ‘moment-­dim-­m’, and so on. To conclude, Somali has a productive quantitative strategy on nouns involving two functions: diminutivisation and augmentation. Instead of using typical evaluative morphological processes, Somali quantitative evaluation is formed analytically by a noun followed

Somali 511 either by the adjective yár ‘small’ or by the adjective wéyn ‘big’. The adjectives yár and wéyn behave as shown in Section 2. In the next section, I deal with evaluative morphology in verbs and adjectives.

4  Evaluative morphology in verbs and adjectives Verbs and adjectives display a specific process for evaluative morphology. Unlike nouns, both verbs and adjectives undergo reduplication to express qualitative evaluation. The use of yár ‘small’ and wéyn ‘big’ as we have shown for nouns is excluded. An example of each category is shown in (9): (9) Evaluation on verbs and adjectives: a. V: cún~cun (cf. cún ‘to eat’) att~eat ‘to nibble’ b. A: cad~cád (cf. cád ‘white’) att~white ‘whitish’ In (9a), the reduplicated verb form indicates a repeated action of eating. This action involves small portions of food. In other words, forms such as that in (9a) have a pluractional interpretation (Cusic 1981; Tovena 2011c). In a similar way, (9b) expresses an attenuated state of the property ‘to be white’. Note that the reduplication pattern shown in (9) works like that used to pluralise adjectives; cf. (5) above.12 As a matter of fact, Andrzejewski (1969, 64–6) affirms that the reduplicated form ‘conveys the notion of plurality’ involving states or activities in which: (10)

Reduplication patterns: a. one subject is involved more than once; b. two or more subjects are involved once or more than once each, consecutively or simultaneously. (Andrzejewski 1969, 66)

In other words, Andrzejewski’s definition of reduplication meanings does not include examples such as (9b) but only pluractionals. That said, plural and diminutive are formally identical in adjectives.13 Context disambiguates between interpretations. Let me go back to the data. Other examples of evaluated verbal forms are proposed below (cf. Saeed 1999, 49–50): (11) Pluractional reduplicated verbs: a. bood~bóod (cf. bóod ‘to jump’) att~jump ‘to jump repeatedly (small little jumps)’ b. dír~dir (cf. dír ‘to send’) att~send ‘to send repeatedly’ c. jár~jar (cf. jár ‘to cut off’) att~to cut off ‘to mince’

512

Nicola Lampitelli

The verbal forms in (11), like pluractional forms cross-­linguistically, indicate approximation, reduction and/or attenuation of the meaning expressed by each basic verb. Evaluation in verbs is restricted to pluractional interpretations although it seems a productive morphological strategy. As shown in (9), the same morphophonological process, reduplication, occurs in adjectives, too. In the same way, it is used to express approximation, reduction and/or attenuation of the meaning the basic adjective refers to. This applies generally to adjectives expressing a colour and those expressing either a discrete quality or state, such as yár ‘small’ and jarán ‘chopped’, respectively: (12) Approximation, reduction and attenuation in adjectives: a. yar~yár (cf. yár ‘small’) att~small ‘tiny, thin’ b. cas~cás (cf. cás ‘red’) att~red ‘reddish’ c. jar~jar-­án (cf. jarán ‘chopped’) att~to.cut.off-­adj ‘chopped up’ The reduplicated form yaryár in (12a) can be used to express evaluation of nouns denoting humans, such as gabádh ‘girl’, wíil ‘boy’, etc.. As a consequence, the following contrast arises: (13) a. gabádh yar~yár girl.f.sg.abs att~small ‘a (very) thin girl’ (cf. Italian una ragazza magra ‘a thin girl’) b. gabádh yár girl.f.sg.abs small ‘a small/young girl’ (cf. Italian una ragazz-­in-­a ‘a girl-­dim-­f’) Forms such as that in (12b) seem productive. Finally, observe the form in (12c), jarjarán ‘chopped up’. It is built on jár ‘to cut off’. Interestingly, both the verb and the adjective have a corresponding evaluative form (11c vs. 12c), repeated in (14) for the sake of clarity: (14) a. jár~jar (cf. jár ‘to cut off’) att~to.cut.off ‘to mince’ b. jar~jar-­án (cf. jarán ‘chopped’) att~to.cut.off-­adj ‘chopped up’ Both forms in (14) are built on the root jár ‘to cut off’. In (14a), the root is reduplicated and gives rise to a pluractional verbal form, járjar ‘to mince’. In (14b), instead, the root is suffixed with -­án (the adjective-­deriving suffix; see Section 2 above). The adjective jarán ‘chopped’ is derived. This form is in turn reduplicated, thus forming an evaluated

Somali 513 adjectival form, jarjarán ‘chopped up’. In other words, evaluation is performed through the same morphophonological process and involves a similar qualitative nuance in both examples. Before concluding, an important clarification deserves attention. It concerns the reduplication process. Consider the following fact: the evaluative strategies I described above differ from each other with respect to category. Nouns use a syntactic construction, whereas verbs and adjectives use a morphological one. Interestingly, adjectives do not behave like nouns. This situation is an additional argument in favour of the hypothesis according to which adjectives are state verbs, as discussed in Section 2. Adjectives and verbs are the same category in Somali (cf. Andrzejewski 1969).14

5 Conclusion In this short survey, I have shown the general pattern of evaluation in Somali. This process primarily involves nouns, although not through a morphological strategy. Rather, we have seen that evaluation on nouns takes place using the adjectives yár ‘small’ and wéyn ‘big’ as modifiers. Besides nouns, evaluation concerns verbs and adjectives. Due to the particular status of the latter category in Somali, verbs and adjectives behave in the same manner with respect to evaluative morphology. Reduplication expresses approximation, reduction and/ or attenuation. We have observed that this process is quite productive, as almost all weak verbs can be reduplicated. To conclude, Somali displays typical evaluative processes only in verbs and adjectives, which belong to the same syntactic category.

Acknowledgements I thank Houssein Assoweh and Moubarak Ahmed, friends and colleagues from Djibouti, for their help in finding data on evaluative morphology. I thank Giorgio Banti for his valuable and interesting comments. All errors are mine.

Notes   1. I adopt official Somali spelling based on the Latin alphabet. The following specific conventions apply: c [ʕ], q [q], sh [ʃ], kh [x], dh [ɖ], x [ħ],’ [ʔ], y [j] and j which is generally pronounced [ʤ]. High tones are marked by accents.   2. For a definition of ‘evaluative construction’, see Grandi and Körtvélyessy, this volume.   3. Some words in different categories are distinguished by tone alone, e.g. kú ‘in, on’ (pr) vs. ku ‘you’ (pro), lá ‘with’ (pr) vs. la ‘one’ (pro) (examples from Saeed 1993, 23). For verbs, see Andrzejewski (1956) and Saeed (1993).   4. Recent work on spoken Djibouti Somali (Northern Somali branch) shows that speakers generalise the use of suffixation in pluralisation. For instance, given míis ‘table’, possible plural forms are miis-­yáal ‘table-­pl’ or miis-­ó ‘table-­pl’ instead of the standard reduplicated form miisás ‘tables’ (cf. Lampitelli 2012).   5. Reduplication, (cf. 2d), consists in copying the last consonant of the stem to the right of it. Then the vowel /a/ is inserted between the two identical consonants, as the following additional examples show: áf vs. afáf ‘tongue(s)’, wiíl vs. wiilál ‘boy(s)’, qóys vs. qoysás ‘famil-­y/ies’, tuúg vs. tuugág ‘thie-­f/ves’, xeér vs. xeerár ‘traditional law(s)’.

514

Nicola Lampitelli

  6. The complete paradigms of two adjectival verbs, adág ‘strong’ and fiicán ‘good’, are illustrated in Puglielli and Mansuur (2012, 931–3).   7. Andrzejewski (1969, 56–64) calls such lexical suffixes ‘root extensions’ and claims that basic adjectives are followed by a zero suffix. In other words, Andrzejewski does not divide these items into two separate groups.   8. Cf. Banti (2011), Mansuur (1988) and Mous (2012) for detailed discussions on particular features of Somali and its similarities to other Cushitic languages.   9. Reduplication in adjectives does not work like reduplication in nouns: cf. míis vs. miisás (2d). See Andrzejewski (1969, 64–6) for more details on how reduplication works in the adjectival paradigm. In the examples in (5), there is an irregular form: waa~wéyn /pl~big/ (5e). 10. Each V, N and ADJ bears one and only one accentual tone (Hyman 1981). As a consequence, in reduplicated forms such as yar~yár /pl~small/ ‘small’, the accentual tone is not reduplicated. See Andrzejewski (1956; 1964; 1969) for the distribution of the accentual tone in verbs, nouns and adjectives, respectively. 11. According to the data provided by my informants, a lexical interpretation of the analytic construction N + ADJ can occur in a few cases, such as: (i) a. macalgád yár spoon.f.sg.abs small ‘teaspoon’ b. macalgád wéyn spoon.f.sg.abs big ‘tablespoon’ c. búug yár book.m.sg.abs small ‘notebook’ c. wadd-­ó wéyn road-­ f.sg.abs big ‘highway’ 12. In reduplicated verb forms, the high tone falls on the penultimate vowel and not on the last one. 13. Andrzejewski (1969, 66) does not mention the diminutive interpretation of reduplication in adjectives, but he does assert that adjectives share such a morphological process with so-­called weak verbs. The overwhelming majority of Somali verbs are weak verbs (only five strong verbs exist in the language, one of these being the above-­mentioned yahay ‘to be’; Banti 2001). 14. Similarly and unsurprisingly, adverbs do not exist as a separate category in Somali. In fact, Saeed (1993, 6 and 277ff.) claims that ‘the grammatical function adverbial is performed by a range of categories, principally nouns and noun phrases’. In other words, in Somali, nouns are modified by verbs (adjectival function) and verbs by nouns (adverbial function).

18.8 Zulu

Andrew van der Spuy and Lwazi Mjiyako

1 Introduction Zulu is a member of the Bantu family of languages of sub-­equatorial Africa. Typically for a language of this family, it is highly agglutinative, with an SVO word order. Phonologically it has a basic (C)V syllable structure, though (N)C(w) syllable-­initial consonant clusters are permitted, as in the second syllable of i-­ngwe ‘leopard’. There is a simple five-­vowel system, but the language has an unusually large number of consonants. As is true of all Bantu languages (Nurse and Philippson 2003a, 8), the nouns are divided into classes, each class commanding a distinctive set of agreement prefixes (Canonici 1990). The classes are marked by (usually) distinctive prefix sequences, consisting of the so-­called ‘initial vowel’ (which functions as a default determiner; van der Spuy 2001), and the noun class prefix proper, for example u-­m-­fana cl1 ‘boy’; a-­ba-­fana cl2 ‘boys’; a-­ma-­ fu cl6 ‘clouds; i-­si-­tsha cl7 ‘dish’, i-­zi-­tsha cl10 ‘dishes’. In some classes the two prefixes have fused into a single prefix, for example i-­fu cl5 ‘cloud’; u-­fudu cl11 ‘tortoise’. (The class numbering system, devised by Bleek 1869, is used for all Bantu languages, thus cl7 in Zulu corresponds historically to cl7 in Swahili; although there is no cl8 in Zulu, a cl8 is found in Swahili and many other Bantu languages.) It is probable that in origin these classes had a semantic import (Denny and Creider 1976); but, with the exception of classes 1 and 2, where every noun has the feature [+human], there is no class in modern Zulu where all the members have a discernible common meaning. Furthermore, [+human] nouns occur in classes other than cl1. Unlike other Bantu languages, for example Shona, Zulu has not retained classes 12 and 13, the diminutive classes: (1)

ka-­bhuku cl13 ‘little book’ vs. bhuku cl5 ‘book’ (Dale 1968, 8, 16)

Nor does Zulu have any class with an augmentative meaning, as, for example, Swahili has: (2)

ji-­tu cl 5 ‘giant’ vs. m-­tu cl1 ‘person’ (Perrott 1951, 31)

Rather, it marks the diminutive and augmentative forms of nouns by means of suffixes. The language makes considerable use of agreement morphology, to an extent that Carstens (2011) has termed ‘hyperagreement’. Agreement is evidenced on determiners, adjectives, possessives (for example, i-­gama la-­mi cl5 ‘my name’, but i-­si-­bongo sa-­mi cl7 ‘my surname’), auxiliary verbs and main verbs. Verbs are marked for a range of tenses,

516

Andrew van der Spuy and Lwazi Mjiyako

aspects and moods and also for polarity; in addition, there is a complex set of subject agreement markers. Besides first and second person agreement, there is agreement for each of the twelve noun classes, which may change according to mood, for example the prefixes in u-­hamb-­a ‘he or she goes’ (cl1, indicative), a-­hamb-­e ‘he or she may go’ (cl1, subjunctive) and e-­hamb-­a ‘while he or she goes’ (cl1, participial). Transitive verbs may also incorporate object pronouns, which, if third person, show noun class agreement, for example: (3) a. ng-­a-­li-­bon-­a 1sg-­pst-­obj5-­see-­vaf ‘I saw it’ (cl5) b. ng-­a-­yi-­bon-­a 1sg-­pst-­obj9-­see-­vaf ‘I saw it’ (cl9)

2  Evaluative morphology As is perhaps to be expected in a language so rich in morphology, Zulu has much evaluative morphology (EM). Various kinds of evaluation are found in nouns, adjectives, verbs and ideophones. In all instances, the EM is derivational rather than inflectional. Strangely, despite this affinity for EM, Zulu fails to make use of a device which other Bantu languages use very effectively, namely conveying an evaluative meaning by changing the class of the noun, as in the Shona and Swahili examples in (1) and (2). There remains some trace of the use of class to convey evaluation: when [+human] nouns occur in classes other than cl1, they often have a meaning which conveys disparagement, praise or simply exceptional status. For example, a number of [+human] nouns in cl5 have negative connotations: i-­thatha ‘forward person’, i-­zimu ‘cannibal’, i-­bhimbi ‘one who sings out of tune’, i-­gwala ‘coward’ (Poulos and Msimang 1998, 51). Some nouns of cl9 convey the idea of expertise: i-­n-­kombi ‘marksman’, i-­n-­gibeli ‘expert horseman’, i-­n-­hlambi ‘expert swimmer’ (Poulos and Msimang 1998, 64). However, these nouns do not meet Grandi and Kortvélyessy’s (this volume) criterion that a morphological change should accompany an evaluative shift in meaning. Rather, the evaluative meaning is part of the basic meaning of the noun. The noun stems do not occur in a ‘neutral’ class (say, cl1), with a neutral meaning: there are no forms like *u-­m-­bhimbi ‘singer’. A rare exception is cl1 u-­m-­gibeli ‘rider’, as opposed to cl9 i-­n-­gibeli ‘expert rider’; but this pairing is not productive: there are no pairs *u-­m-­khombi ‘shooter’ – i-­n-­ kombi ‘marksman’; *u-­m-­hlambi ‘swimmer’ – i-­n-­hlambi ‘expert swimmer’. In some instances, this may be because the pairs are blocked by nouns with similar forms but divergent meanings, such as u-­m-­hlambi cl3 ‘flock’ (not ‘swimmer’), but even where there is no such blocking, the expected pairs do not occur. Zulu, then, unlike many other Bantu languages, does not use its noun class system for the purposes of evaluation. The EM that is found in Zulu consists of a number of derivational affixes (for example /jana/ diminutive, -­kazi augmentative, -­isis-­intensive), as well as reduplication and a special use of the relative form. The morphology will be discussed under the headings of these devices, rather than grouped as ‘EM of nouns’, ‘EM of verbs’, etc. This is because some of these devices (specifically the diminutive and augmentative suffixes and reduplication) apply to more than one lexical category.

3  The diminutive suffix /jana/ Although Zulu has no diminutive noun class, it does have a diminutive form, and this is completely productive. It is formed by adding the suffix /jana/ to the noun or adjective. It

Zulu 517 can be used to express simple diminution in size, or, when used with a plural noun, diminution in quantity: a-­ma-­sokis-­ana ‘a few socks’ < a-­ma-­sokisi ‘socks’. (The consonant /j/ – orthographic – is elided between a consonant and a vowel by a general phonological rule, which Khumalo 1987 calls ‘y-­deletion’.) The diminutive can be used with a pejorative meaning, as in: (4)

lowo m-­fany-­ana that n1-­boy-­dim ‘that good-­for-­nothing boy’ (Poulos and Msimang 1998, 103)

It also has an ameliorative sense, as in: (5)

Bheka nje a-­ka-­se-­mu-­hle look just neg-­sbj1-­now-­adj1-­handsome ‘Look how handsome my dear little boy is.’

u-­m-­fany-­ana wami. det-­n1-­boy-­dim my

The ameliorative sense appears to be confined to [+human] nouns: meanings like i-­n-­j-­ana *‘dear little dog’ and i-­n-­cwaj-­ana *‘sweet little book’ are not found. The suffix /jana/ presumably derives from the Proto-­Bantu noun root *yánà ‘child’ (Guthrie 1971, 144). 3.1  The phonology of the diminutive The diminutive has a somewhat complex phonology. In most reference grammars, the diminutive suffix is referred to as -­ana (e.g. Malcolm 1960; Doke 1973; Poulos and Msimang 1998, 102). The reason for this is that the /j/ at the beginning of the affix never surfaces; however, the affix has a palatalising effect on certain sets of preceding consonants, and therefore Khumalo (1987) analyses it as /jana/. His analysis will be accepted here. The formation of the diminutive can be expressed as the following sequence of steps: (6)

a. When the suffix /jana/ is added to a noun, the final vowel of the noun stem is changed to a glide /w/ if it is a back vowel; otherwise it is elided altogether. b. If this results in a sequence /C[bilabial] wj/, the /w/ is elided. (In Khumalo’s account, the /w/ first assimilates to /j/ and is then elided; but we see no justification for positing this extra step.) c. A bilabial stop or nasal is changed to a palatal consonant before /j/ (‘labial palatalization’; Khumalo 1987, 164ff.). d. In some varieties an alveolar stop or nasal is changed to a palatal before /(w)j/ (‘alveolar palatalization’; Khumalo 1987, 179–80). e. The /j/ is subsequently elided by y-­deletion, as described above.

The following examples illustrate these steps, indicated there as (a), (b), etc. The examples are given in standard orthography, as this is phonologically very transparent; the diminutive suffix is written as . (Some of the examples are based on those in Khumalo 1987.)

(7)

u-­bu-­so ‘face’→ (a) u-­bu-­sw + yana → (e) u-­bu-­sw-­ana a-­ma-­sokisi ‘socks’ → (a) a-­ma-­sokis + yana → (e) a-­ma-­sokis-­ana

Andrew van der Spuy and Lwazi Mjiyako

518







i-­phupho ‘dream’ → (a) i-­phuphw + yana → (b) i-­phuph-­yana → (c) i-­phush-­yana → (e) i-­phush-­ana i-­n-­dlebe ‘ear’ → (a) i-­n-­dleb + yana → (c) i-­n-­dletsh-­yana → (e) i-­n-­dletsh-­ana i-­si-­godi ‘district’ → (a) i-­si-­god + yana → (d) i-­si-­goj-­yana → (e) i-­si-­goj-­ana or: → (e) i-­si-­god-­ana i-­moto ‘car’ → (a) i-­motw + yana → (d) i-­motshw-­yana → (e) i-­motshw-­ana or: → (e) i-­motw-­ana

Steps (6a), (6b) and (6d) apply only in the diminutive and in no other part of the grammar. Khumalo (1987) claims that (6c) applies not only in the diminutive, but also in the passive and the locative; however, the evidence for an underlying /j/ in these last two constructions is very tenuous. Labial palatalisation in the passive and locative is also analysable as a process of dissimilation before /w/. 3.2  Other uses of the diminutive In some circumstances the diminutive suffix occurs in a reduplicated form -­yany-­ana. This reduplicated suffix signifies ‘worthlessness, ugliness’, or, by contrast, ‘pity, sympathy’, for example a-­ba-­fany-­any-­ana det-­n2-­boy-­dim-­dim ‘worthless boys’, i-­n-­gany-­any-­ana det-­ n9-­child-­dim-­dim ‘poor little child’ (Poulos and Msimang 1998, 108). (Orthographic represents phonemic /ɲ/.) Adjectives may also take the diminutive suffix /jana/: the diminutive form of adjectives has the meaning of ‘slightly, somewhat’, for example -­fushane + -­yana → -­fushany-­ana ‘somewhat short’, -­dala + -­yana → -­dadl-­ana ‘somewhat old’. (The change /l/ → /ɮ/ – orthographic → – is confined to the diminutive forms of adjectives.) If the adjective stem is monosyllabic, an alternative suffix -­nyana is preferred: -­hle + nyana → -­hle-­nyana ‘somewhat good’. The diminutive form has acquired the additional function of indicating the comparative degree, as in u-­m-­fana o-­m-­dadl-­ana det-­n1-­boy rel-­adj1-­old-­dim ‘an older boy’, i-­n-­ duku e-­n-­de-­nyana det-­n9-­stick rel-­adj9-­long-­dim ‘a longer stick’. The suffix sequence -­az-­ane signifies ‘female and young’, as in i-­si-­bhuz-­az-­ane ‘young she-­goat’ < i-­m-­buzi ‘goat’ (Poulos and Msimang 1998, 109). This suffix sequence has a pejorative meaning in the word u-­m-­faz-­az-­ane ‘good-­for-­nothing woman’ (Poulos and Msimang 1998, 109); however, this is not productive.

4  The augmentative suffix -­kazi The augmentative of nouns and adjectives is formed by adding the suffix -­kazi. This is not as productive as the formation of the diminutive, partly because there is an identical suffix -­kazi ‘feminine’ (derived from Proto-­Bantu *kádì̦ ‘woman, female’ (Guthrie 1971, 129); the augmentative presumably derives from this). The augmentative suffix cannot be added to nouns where there is an established feminine form in -­kazi, thus i-­m-­vu-­kazi cl9 ‘ewe’ (not ‘big sheep’); i-­n-­ja-­kazi cl9 ‘female dog’ (not ‘big dog’). But it is found with other animate nouns, for example u-­m-­ntu-­kazi cl1 ‘big person’, including nouns that have a feminine meaning, such as i-­n-­tombi-­kazi cl9 ‘big girl’. It is also found with inanimates, for example i-­bhodwe-­kazi cl5 ‘big pot’ (Poulos and Msimang 1998, 110). It is not totally productive: a form like ?i-­bhilidi-­kazi cl5 ‘big building’ is regarded as odd, and it cannot be used with mass nouns: *u-­bisi-­kazi cl11 ‘much milk’.

Zulu 519 The augmentative can have a pejorative meaning, for example u-­m-­lomo-­kazi cl3 ‘big ugly mouth’ (Poulos and Msimang 1998, 110). This appears to be confined to nouns referring to body parts. In Zulu both the diminutive and the augmentative have pejorative associations. They can both be used descriptively, but only the diminutive can have an ameliorative meaning. Adjectives may also take the augmentative suffix -­kazi. The augmentative form of adjectives is intensive in meaning, for example -­khulu-­kazi ‘very big’, -­fushane-­kazi ‘very short’ (Poulos and Msimang 1998, 144). However, this is not productive; for example, *-­dala-­kazi ‘very old’ is not acceptable. The augmentative has acquired the additional function of indicating the superlative degree with the adjective -­khulu ‘big’, as in i-­moto e-­n-­kulu-­kazi det.n9-­car rel-­adj9-­big-­ aug ‘the biggest car’. However, it is not found with other adjectives.

5  The evaluative relative Zulu has an unusual construction which may be termed the ‘evaluative relative’. It has been described in detail by Wilkes (1990). First the prefix ngu-­ ~ yi-­is attached to a noun. This makes the noun into a predicative form: ng-­u-­m-­fana ‘it is a boy’, y-­i-­n-­gane ‘it is a child’. (The vowels of the prefix ngu-­ ~ yi-­are deleted before another vowel by  phonological rule.) Then the prefix sequence o-­ku-­is attached; this means ‘that which’: o-­ku-­ng-­u-­m-­fana ‘that which is a boy’, o-­ku-­y-­i-­ngane ‘that which is a child’. O-­ku is a cl15 prefix sequence: this class is used by default for vague or indefinite ‘things’, as in: (8)

Lokho o-­ku-­sh-­ilo dem15 rel.2sg-­obj15-­say-­prf ‘what you said makes sense’

ku-­n-­e-­n-­gqondo sbj15-­have-­det-­n9-­sense

The evaluative relative construction conveys amelioration, or disparagement/distancing, depending on context; for example, (9)

O-­ku-­y-­i-­n-­gane

rel-­sbj15-­be-­det-­n9-­child

ku-­ya-­khal-­a

sbj15-­pfm-­cry-­vaf

‘the poor little child is crying’ or ‘that damn child is crying’.

6 Reduplication Another way of marking evaluation is reduplication. This is found regularly in nouns, verbs and ideophones. 6.1  Reduplication in nouns One (non-­productive) kind of reduplicated noun has the meaning ‘real, typical’. In the few examples that exist, only the first syllable of the stem is reduplicated:





(10)

i-­nsizwa-­nsi ‘ a real young man’ < i-­nsizwa i-­ntombi-­nto ‘a real young woman’ < i-­ntombi u-­mu-­ntu-­ntu ‘ a real/typical person’ < u-­mu-­ntu (Poulos and Msimang 1998, 85)

Andrew van der Spuy and Lwazi Mjiyako

520

In the more productive form of reduplication, the whole stem is reduplicated if it has only two syllables, for example i-­mi-­fula-­fula ‘many different rivers’; i-­zi-­godi-­godi ‘many different valleys’ (Poulos and Msimang 1998, 85). Only the first two syllables of the stem are reduplicated if it has more than two syllables: i-­mi-­vivi-­vivinyo cl4 ‘many tests’. If the stem consists of a single syllable, a ‘stabiliser’ affix (Poulos and Msimang 1998, 202) -­yi-­ is inserted between the repetitions: i-­mi-­zi-­yi-­zi cl4 ‘many villages’. A slightly different form of noun reduplication can be seen in the name a-­ba-­fana-­ba-­ fana det-­n2-­boy-­n2-­boy ‘boys-­boys’, the nickname of South Africa’s national football team. This appears to be a unique example of reduplication incorporating the noun prefix, and is not used everywhere: in Johannesburg, the reduplicated stem has been relexicalised as a cl5 noun i-­bafana-­bafana. The evaluation associated with this particular reduplication is affection or admiration. 6.2  Reduplication in verbs The reduplicative form of the verb is sometimes known as the ‘diminutive’ (Doke 1973, 148ff.). According to Doke (1973), it means ‘do X a little’. But it has other meanings, depending on context. It can mean ‘do X repetitively’ (Poulos and Msimang 1998, 202). According to Cook (2013, 1), it can also mean ‘do X without much skill’; this usage appears to be primarily self-­deprecating, for example ngi-­ya-­cul-­a-­cul-­a ‘I’m just singing a bit’, and is thus connected to the diminutive meaning of the form. Phonologically, it follows the same rules as reduplication of the noun, and is based on the simplest form of the verb, for example hamb-­a ‘go’, phek-­a ‘cook’. This simplest form is bimorphemic, the suffix -­a being the default verb suffix: there are no monomorphemic verb forms in Zulu. The reduplications are hamb-­a-­hamb-­a and phek-­a-­phek-­a. In other aspects or moods or in the negative, the final vowel (but not the affix -­a-­in the middle of the reduplication) can be replaced by other suffixes, for example -­hamb-­a-­hamb-­ile ‘has walked (red)’, -­hamb-­a-­hamb-­e ‘may walk (red)’, -­hamb-­a-­hamb-­i ‘does not walk (red)’ (see Cook 2013). The following are further examples of verbal reduplication, showing stems that do not conform to the bisyllabic pattern: giji-­gijim-­a ‘run (red)’; gigi-­gigithek-­a ‘giggle (red)’, lw-­a-­yi-­lw-­a ‘fight (red)’, m-­a-­yi-­m-­a ‘stand (red)’. If the basic stem of the verb begins with a vowel, then a is prefixed to the second part of the reduplication: al-­a-­y-­al-­a ‘refuse (red)’, esa-­y-­esab-­a ‘fear (red)’. This is presumably the ‘stabiliser’ affix yi-­, by analogy with the stabiliser prefix used with monosyllabic stems. However, Cook (2013) associates it with the verbal morpheme -­ya-­which indicates that the verb is used phrase-­ finally (van der Spuy 1993; Buell 2005). 6.3  Reduplication in ideophones Ideophones are a class of word very frequent in Zulu. They can be compared to ­onomatopoeic words, but they evoke a wider range of ideas than just sound, including motion, colour, smell and actions. Reduplication of an ideophone can sometimes refer to a repetition of the action; for example: (11)

W-­a-­thi bhaja-­bhaja a-­me-­hlo. (Fivaz 1963, 146) sbj1-­pst-­go shift!-­shift! det-­n6-­eye ‘His eyes went back and forth shiftily.’

But it can also signify intensity:

Zulu 521 (12) (a) li-­bomvu klubhu-­klubhu sbj5-­red red!-­red! ‘it is very very red’; (b)  U-­bhanana u-­mnandi u-­thi ncamu-­ncamu. det.n3-­banana sbj3-­tasty sbj3-­go tasty!-­tasty! ‘The bananas are very very tasty.’ (Poulos and Msimang 1998, 413)

7  The intensive form of verbs Zulu has a very productive set of verbal suffixes (the ‘extensions’; cf. Schadeberg 2003) which modify the basic meaning of the verb. One such extension is -­isis-­, presumably a reduplication of the causative extension -­is-­(Poulos and Msimang 1998, 199). It has an intensive meaning, and is thus evaluative. Examples are phuz-­isis-­a ‘drink heavily’; gijim-­ isis-­a ‘run very hard’; zam-­isis-­a ‘try very hard’.

8  The reflexive/applied form of verbs In this construction, the reflexive object prefix zi-­is attached to the basic verb stem, and the so-­called ‘applied’ extension -­el-­ (Doke 1935, 52–3) is suffixed to it. This extension has a benefactive meaning: thus theng-­a means ‘buy’ and takes a single object with the semantic role of theme; theng-­el-­a means ‘buy (something) for (someone)’and takes two objects with the semantic roles of beneficiary and theme. The zi-­+ -­el-­construction has a literal meaning of ‘do X for oneself’, but has the figurative or contextual meaning ‘just do X, do X for no particular purpose’: ngi-­zi-­fund-­el-­a 1sg-­refl-­read-­appl-­vaf ‘I am reading for myself/I am just reading’.

9  The clitic particle -­nje This clitic can be added to most parts of speech, and has the meaning of ‘just, a mere’; thus it is evaluative. Examples are Uyacula=nje ‘he or she is just singing’, Ungumfana=nje ‘he is just a boy’, Umusha=nje ‘he or she is just young’.

10 Conclusion It has been seen that Zulu has a wide range of EM. It is set apart from other Bantu languages in that it does not use its noun class system for evaluation. Diminutive, augmentative and reduplicative forms have much the same range of meanings as found in other languages, though it is noteworthy that both the diminutive and the augmentative have pejorative meanings, and both of them are acquiring the function of degree of comparison. The relative evaluative construction, for example okuyinja ‘that thing which is a dog’, pejorative or ameliorative depending on context, is perhaps the most unusual evaluative form, compared to other languages.

Acknowledgements The authors would like to thank Sandra Linton for her kind help in editing the manuscript.

19  North America 19.1 Cabécar

Guillermo González Campos

1 Introduction Cabécar is one of five indigenous languages still spoken in Costa Rica.1 Currently, it is the language with the largest number of speakers in this country.2 In the national census of 2011, 16,985 persons identified themselves as speakers of an indigenous language. Approximately 85 per cent of all of these speak Cabécar as their mother tongue.3 Like other indigenous languages still alive in Costa Rica, Cabécar is a Chibchan language. It is classified by Constenla (2008) as belonging to the West Isthmian group with Bribri, Naso and Boruca. The Chibchan family is the largest among the indigenous genealogical groups in southern Central America.4 Constenla (2002, 2005) has demonstrated that Chibchan is part of a larger group with Misumalpan and the two Lencan languages, which he named the ‘Lenmichian Micro-­phylum’. Cabécar has two well-­differentiated dialects: the northern one spoken in Chirripó and Valle de La Estrella, and the southern one spoken in Talamanca and Ujarrás. The differences between these dialects are mostly in phonology. Northern Cabécar has a vowel sound [ɤ] and aspirated consonants which are not present in the southern variety of the language. The main typological features of this language5 are the following: •  It is an ergative language. The canonical word order in a transitive clause is erg-­abs-­verb. • The possessor precedes the possessed, and descriptive adjectives follow the noun. • Cabécar uses postpositions, not prepositions. • It is a tonal language, with two different tones: low and high. Regarding the morphological features, Cabécar is an inflectional language that depends entirely upon suffixation. The following word categories are inflected: nouns, adjectives, verbs, some adverbs and numeral quantifiers. There is no grammatical gender marker. Number is distinguished by the plural marker -­wá on human nouns and a few domestic and farm animals (dogs, cats, hens, horses and the like). In the verbal morphology, tense, aspect and mood are marked by suffixes and clitics. With respect to orthography, Cabécar uses the Latin alphabet, based on its employment in Spanish. The letters have the same phonetic value as in Spanish with the exception of vowels with umlauts, ä [ɤ], ë [ɪ] and ö [ʊ], the consonants l [ɺ] and y [ʤ], and the aspirated consonants that are represented by digraphs (kj, tj and pj). There are two diacritical signs:

524

Guillermo González Campos

the macron below (_) marks nasality in the vowels, and the acute accent (´) is used for marking the high tone.

2  Cabécar evaluative morphology In Cabécar, as in other indigenous languages of Costa Rica, evaluative morphology is very poor. This is due to the fact that evaluation is done largely using syntactic strategies. For example, with nouns and verbs, diminution or augmentation is expressed through adjectival and adverbial modification. Evaluative morphology is found only in adjectives and a few adverbs. It is related to the property of specifying various degrees, which are measured or estimated by the speaker’s subjective experience. Specifically, there are two concepts involved in this process: attenuation and intensification. Attenuation is expressed by an inflectional suffix -­na̱ that indicates a decrease in the quality of the property expressed by the adjective, as (1) illustrates: (1) a. bätsë ́ -n­ a̱ att red-­ ‘reddish’ b. dalá-­na̱ yellow-­ att ‘yellowish’ c. tse̱ ́ -­na̱ cool-­ att ‘fairly cool’ This suffix is also added to some adverbs (especially time adverbs) for the same purpose. It is seen in the words in (2): (2) a. aléji̱ a̱ -­na̱ recently-­ att ‘fairly recently’ b. bájakjë ́ -n­ a̱ long.time.ago-­ att ‘a short time ago’ There are some adjectives that must necessarily take this suffix. In the examples given in (1), the root of the adjective may be free. When this happens, the last vowel of adjective is usually lengthened.6 It occurs in this way in bätsë ́ ë́ ‘red’, daláá ‘yellow’ and tse̱ ́ e̱ ́ ‘cool’. However, as indicated earlier, the presence of -­na̱ is obligatory in some cases. It was found that this was the case for dokóna̱ ‘ugly’ and dolóna̱ ‘black’. Neither *dokóó nor *dolóó is grammatically possible. In these words, the meaning of the suffix -­na̱ is not ‘attenuation’; it seems to mark the characteristic expressed by the adjective as a stable property. In other cases, the addition of the suffix causes small changes of meaning. For example, spa̱ ́ a̱ ́ means ‘green’, but spa̱ ́ na̱ is always translated as ‘new’, never as ‘greenish’. There are almost no similar examples of this. Accordingly, it cannot be seen as a derivational use of -­na̱ .

Cabécar 525 In Cabécar, the attenuation of adjectives is also formed by reduplication. The most common is the full reduplication that generates equivalent expressions to those formed with -­na̱ , as (3) illustrates: (3) a. bolo~bolo sweet~att ‘tasteless’ b. dala~dala yellow~att ‘yellowish’ In this case, too, the morphological process can be observed in some adverbs: (4) a. bétjí~bétjí quickly~att ‘fairly quickly’ Partial reduplication also occurs in Cabécar. It involves the repetition of the last syllable of the word. However, this type of reduplication, which has the same meaning as full reduplication, is far less frequent. Two examples are listed in (5): (5) a. bätsë ́ ~tsë́ red~att ‘reddish’ b. doló~ló black~att ‘blackish’ As shown, in Cabécar, reduplication has non-­iconic meaning, since the addition of the form signifies the diminution in quality of the property denoted by the word. Intensification, the other evaluative notion expressed in Cabécar through morphological strategies, is formed by adding an enclitic morpheme. There are two clitics used for this purpose: rama̱ /rrama̱ 7 and si̱ ́ . The first is added to an adjective only to intensify its meaning: (6) a. báá=rrama̱ beautiful=int ‘very beautiful’ b. dalá=rrama̱ yellow=int ‘very yellow’ c. siö ́ =rrama̱ blue=int ‘very blue’ It is quite clear that rama̱ /rrama̱ was an independent word in former times, but today it is phonologically bound to adjectives. The main evidence for this is that the sounds [r] and

526

Guillermo González Campos

[ɾ] cannot appear at the beginning of a word. In this position, this phoneme is always manifested as [d], but the form *dama̱ has never been recorded. The other clitic with intensive significance is si̱ ́ . When it is added to adjective stems, it produces the same effect as rama̱ /rrama̱ .8 So in principle, they are synonymous forms, as (7) illustrates: (7) a. báá=si̱ ́ beautiful=int ‘very beautiful’ b. daláá=si̱ ́ yellow=int ‘very yellow’ Nevertheless, this clitic can be applied to noun stems, unlike rama̱ /rrama̱ , which can only modify adjectives. When this happens, it adds a sense of ‘authenticity’ or ‘prototypicality’ to the noun that it modifies: (8) a. chíchi=si̱ ́ dog=int ‘an authentic dog’ b. ju=si̱ ́ house=int ‘an authentic house (a traditional Cabécar house)’ A few nouns formed with si̱ ́ have the peculiarity of designating a specific member of the set expressed by the lexical root. This is because the specific member is regarded as the best specimen of its class. It is, then, a semantic change, as in (9): (9) a. ni̱ ma̱ =si̱ ́ fish=int ‘bobo fish (Joturus pichardi)’ b. chi̱ mo̱ =si̱ ́ banana=int ‘black banana (Musa sp.)’ c. tkäbë ́ =si̱ ́ snake=int ‘terciopelo (Bothrops asper)’ As (9a) illustrates, the word ni̱ ma̱ si̱ ́ does not designate ‘an authentic fish’, but rather a particular species of fish: the bobo. Obviously, the cause of this is that the Cabécar people very much like to eat that fish. The situation is somewhat similar to the example given in (9b) and (9c). It cannot be said that derivational morphology is present in these cases, because they are specific, not widespread. Rather, these are examples of individual changes of meanings. In conclusion, Cabécar evaluative morphology is almost limited to adjectives and involves only two concepts: (i) attenuation formed by suffixation and reduplication, and (ii) intensification formed by adding clitics.

Cabécar 527

3  Evaluative morphology in languages related to Cabécar In southern Central America languages, a scarce evaluative morphology is not exclusive to Cabécar. The Chibchan languages of this area usually do not express evaluation by means of morphological strategies. The most common way of forming diminutives, augmentatives pejoratives, amelioratives and so on is by using adjectival or adverbial modification. In Bribri, the closest language genetically to Cabécar, there is a system very similar to that described above. The evaluative morphology is limited almost entirely to adjectives and is restricted to processes of attenuation and intensification. Full and partial reduplication is used, as in Cabécar, to construct the attenuated form of adjectives. Likewise, intensification is expressed by adding an inflectional suffix -­ë. In addition, Bribri has a diminutive suffix -­ala/-­la that can be added to nouns to express endearment. For example, na̱ u̱ ́ la means ‘dear little uncle’ (na̱ u̱ ̀ is ‘maternal uncle’).9 A similar situation occurs in Boruca, a language whose last fluent native speakers died in 2004. Boruca has no suffix to indicate diminution, augmentation or age variation in nouns. All these are formed using post-­nominal modifiers. However, adjectives can be modified by three different suffixes to mark attenuation and intensification.10 The opposite happens in Térraba, a variety of Naso now extinct in Costa Rica. In this dialect, attenuation and intensification are formed by adding to the adjective an adverbial modifier. However, diminution and augmentation in the nouns are marked by derivative suffixes. The diminutive suffix is -­hua/-­hu̱ a̱ and can be added to nouns, adjectives and some adverbs. Moreover there is the suffix -­quë ́ o from which are derived augmentative forms, but its frequency is very low and it can be applied only to nouns. Like Cabécar and Bribri, Térraba has full reduplication as a word-­formation strategy. But reduplication usually has no special significance except that it forms the stems. Only in a few cases is it used to express attenuation. An example of this is c’ríc ‘warm’, whose attenuative form is c’ríc’ríc ‘fairly warm’.11 Finally, other languages of the area have no evaluative morphology at all. This is the case of Malécu (or Guatuso), another language spoken in Costa Rica which is classified as part of the Votic branch of the Chibchan family. In Malécu, all evaluative markers are modifiers that function as phonologically independent words.12 Therefore, it can be concluded that evaluative morphology has not been well developed in the indigenous languages of Costa Rica.

Notes

The editors are aware of the fact that some of the languages discussed in this chapter belong to the South American macro-­area. The classification of languages into macro-­areas is based on WALS and since the focus of the discussion is Cabécar (a North American language), the chapter has been placed in the section on the North American languages.   1. The other languages are Bribri, Guatuso or Malécu, Guaymí or Ngäbe, and Bocotá or Buglere.   2. Actually, Guaymí or Ngäbe is the language with the most native speakers in the region. However, the majority of the Guaymí people live in Panamá. In Costa Rica, the 2011 census recorded a population of 3,654 people who spoke Guaymí.   3. The data are from INEC (2011).   4. Currently, the principal survey of the Chibchan languages is Constenla (2012).  5. For more details on the grammatical structure of Cabécar, see Verhoeven (2012). See also Margery (1989), still the main source for the study of Cabécar.

528

Guillermo González Campos

  6. This is particularly frequent in Northern Cabécar.   7. The alternation between [r] and [ɾ] is allophonic in Cabécar. Both sounds are allophones of the phoneme /d/.   8. A few informants indicate that rama̱ /rrama̱ intensifies the meaning of adjectives to a lesser degree than si̱ ́ . On the other hand, there is a dialectal variation in the usage of one form or the other. The clitic si̱ ́ is preferred by speakers of the northern dialect, while in the south, the ­speakers prefer to use the other clitic.   9. For more information about this language, see Constenla, Elizondo and Pereira (1998). 10. For more details, see Quesada Pacheco (1995). 11. On Térraba, see Constenla (2007). 12. On this, see the grammar of Malécu prepared by Constenla (1998).

19.2 Choctaw

Marcia Haag

1 Introduction Choctaw is a Muskogean language of the south-­eastern portion of North America, now the state of Mississippi in the United States. In 1830, the Choctaws were forcibly removed to the Plains region of the USA, now the state of Oklahoma. Most speakers now reside in one of those areas, and are affiliated with the Mississippi Band of Choctaws or Choctaw Nation of Oklahoma. Choctaw’s closest sister language is Chickasaw, followed by Alabama and Koasati, then by Hitchiti (virtually extinct) and Mikasuki, and is most distantly related to Creek (Muskoki). There are perhaps 2,000 fluent and semi-­fluent speakers. Structurally, Choctaw’s major features are that it has argument selection based on semantic rather than grammatical role (it is in the group of active/stative languages); it has an extensive switch-­reference system; it has ample tense, aspect and modal marking on the clausal verb; and it has a highly articulated definiteness system. It has robust SOV word order. Morphologically, Choctaw’s derivational morphology is quite sparse, but the inflectional morphology is dense at both clause and word level. I will use IPA rather than vernacular orthographies in this chapter. 1.1  Derivational morphology Choctaw is interesting linguistically for its lack of affixes that signal lexical category. The most basic way for lexemes to be related is through a verb-­to-­noun derivation wherein the derived noun is not distinct morphologically from the verb. The derivative is related to the verb through the verb’s argument structure. An intransitive verb with an agent subject produces an agent noun, while an intransitive verb with a patient subject produces a resultative noun. A transitive verb produces a noun that can be either an agent or a result, since either semantic role may be selected (Haag, forthcoming). This can be seen in the examples in (1). (1)

Intransitive verb, agent subject: baliili ‘run’ (verb) or ‘runner’ (agent) Intransitive verb, patient subject: baʃa ‘be cut’ (verb) or ‘a cut’ (result) Transitive verb: holaabi ‘lie’ (verb) or ‘liar’ (agent) or ‘a lie’ (result)

An important word-­ formation process in Choctaw is compounding. Compounds are made with all lexical categories. The distinction between adjoined modifiers and

530

Marcia Haag

c­ ompounds is made phonologically and through semantic ­differences: phrasal ­modific ation is  ­rule-­governed while compounds undergo semantic drift. 1.2  Inflectional morphology The inflectional set of morphemes is far larger and more complicated than the derivational set. The two basic groups of inflections centre on whether an expression is an argument or a predicate, irrespective of the lexical categories of its heads. A further consideration is whether a predicative expression is modifying (syntactically adjoined) or is the main predicate. In (2), the subject argument marker at and its many compounds, e.g. kat, is used in nominal subjects and sentential subjects, and to mark switch-­reference in dependent clauses. (This is not an exhaustive list of its uses.) The adverbial clause is also marked for same subject with hoʃ. (2) Ohojo-­ at imponna-­hoʃ aiokli-­kat ikhaana-­h. sbj clever adv.ss beautiful-­comp:ss know-­prs woman-­ ‘The woman, being clever, knows that she is beautiful.’ Other important inflectional morphology involves stem-­modifying operations such as reduplication of and infixation into the verbal stem. The most prominent examples involve the aspect system. In (3), the marker of iterative aspect is the infix /h/ plus the reduplicated and nasalised vowel of the penultimate syllable of the verb stem. (3) pisa ‘look at’ pih̃isa ‘keep looking at’ holaabi ‘lie’ holahãbi ‘keep on lying’

2  Evaluative morphology Choctaw has two productive evaluative morphemes, oʃi and holba. Both are examples of diminutives. Choctaw does not appear to have any augmentative morphological formatives, although there are a very few frozen forms that contain a morpheme that signals ‘importance’. Both oʃi and holba are conservative in their semantic fields, and both form derivations through compounding. Bound forms can thus be analysed as bound words rather than affixes. As is common, they both have allomorphy that is not dependent on general phonological rules in the language. Both markers are restrictive in the kinds of lexical categories to which they may attach. Oʃi is strictly attached to nouns, while holba attaches to nouns and adjectives. While oʃi is a classic diminutive, holba has uses that exemplify some of the more abstract uses of evaluative morphology. Choctaw evaluatives do not extend to a speaker’s evaluation of something as ‘good/ bad’, as occurs in many languages. Thus, they do not have pejorative and affectionate functions. Since there are no longer any large Choctaw speech communities, there is no documentation about evolution of the diminutives to include these uses.

Choctaw 531 2.1  Oʃi: semantic fields and extensions The morpheme oʃi is ubiquitous in Choctaw. Semantically, oʃi is related to the notion of ‘offspring’, and indeed exists in a free form which means ‘one’s son’. Its overarching sense is relatedness in the sense of having been generated from a prototype, so while oʃi derivatives generally signal a more diminutive size than the prototype, far more important is the sense that the oʃi derivative was created in some image of the base form. This is in keeping with the supposition that the core metaphor for diminutives is ‘child’, as discussed in Jurafsky (1996). Oʃi does not have extended uses with lexemes of categories other than noun. The following are examples of one common type oʃi derivative, and the most transparent, namely the relation of an animal and its young. (4) tʃokfi ‘sheep’ → katos ‘cat’ → nita ‘bear’ → akãka ‘chicken’ → ʃokatti ‘frog’ →

tʃokfoʃi ‘lamb’ katosoʃi ‘kitten’ nitoʃi ‘bear cub’ akãkoʃi ‘chicken egg’ ʃokattoʃi ‘frog spawn’

Note that the oʃi derivatives involving egg-­laying animals refer to the eggs and not the younger animal. A second semantic category of oʃi derivatives is one involving things that are similar to a prototype, and generally smaller. The metaphor here is that the derivative marks a quality that is less than the quality marked by the prototype, and imitative of it. (5) piini ‘boat’ (any) → iʃtfotooha ‘auger’→ book ‘river’ → osaapa ‘field’ → takkon ‘peach’ →

piinoʃi ‘skiff; yawl; small boat’ iʃtfotoohoʃi ‘small borer’ bookoʃi ‘creek’ osaapoʃi ‘garden’ takkonloʃi ‘plum’

In this class, we see that the last two examples, ‘garden’ and ‘plum’, have a newly lexicalised meaning that departs from smaller size. In the case of takkonloʃi we also have a distinct allomorphy (loʃi) that is not required by the phonology. A third semantic class is made of lexicalised words whose metaphors are further abstracted from the sense of generation from a prototype. Here, we encounter partitive uses of the diminutive, which Jurafsky (1996, 555) analyses as ‘selecting a unit from a larger, more articulated mass’. (6) abooha ‘room’ → ibbak ‘hand’ → okla ‘people’ →

aboohoʃi ‘room with a toilet’ ibbakoʃi ‘finger’ okloʃi ‘tribe’

In this class, we have the euphemism aboohoʃi, not just a ‘small room’ but a special one among the other rooms of a house. Ibbakoʃi ‘finger’ explicitly gives the sense of an organ that is physically generated from the hand – a part of it, but hardly an imitation of it. Okloʃi

Marcia Haag

532

‘tribe’ again explicitly suggests the generation of a group of people from a larger group, to whom the smaller group belongs. Oʃi is not used simply to depict small size. The word for ‘baby’ is made from ‘very small’ and ‘child’, not with oʃi. Hattak oʃi must mean ‘the man’s son’ not ‘small man’. (7) alla ‘child’ + õssi ‘very small’ → hattak ‘man’ + iskitini ‘little’ →

allõssi ‘baby’ hattak iskitini ‘little people, elves’

2.2  Holba: semantic fields and extensions This diminutive marker is used in the more abstract senses of deviation from a prototype. Holba may be used predicatively to mean ‘like’ or ‘similar to’, as in: (8) Hattak yammat ãki holba my.father like man dem.sbj ‘That man is like my father.’

-­ h. prs

But it is used far more often in compounds of both nouns and adjectives, and in these uses it takes on evaluative properties. Holba has two allomorphs, oba and olba, which are always bound. Holba does not originate as a metaphor based on ‘child’, as does oʃi. Rather, it participates in an important metaphor that has been analysed as belonging to the realm of the diminutive, that of centrality and marginality (Jurafsky 1996, 548–54). In Jurafsky’s discussion, there are several aspects to this metaphor, each of which is manifest in holba derivations. In the first sense, the metaphor is ‘marginal is small’. Among Choctaw regional dialects, holba may alternate with oʃi since the latter can signal the evaluation of a quality as being less than that of a prototype. So in (9) the central concept ‘pumpkin’ is ­marginalised by oʃi in some dialects and by holba is others to create words that signify smaller and different species of this vegetable family. Importantly, having both forms allows speakers greater leeway in word formation: now Choctaws can have two kinds of vegetables that can be distinguished from each other and from the central concept. (9) issito ‘pumpkin’ issito holba ‘squash’ (some dialects) issito holba ‘cucumber’ issitoʃi ‘squash’ (other dialects) The second sense is characterised as ‘generalisation’ or ‘bleaching’ (Jurafsky 1996, 553). Here, increased abstraction results in a sense of being related to a central concept. Jurafsky attributes English -­ish to this usage, pointing out that ‘(t)he resulting sense has completely left the original source domain of size’. The default English translation of holba is with -­ish. I will include the related bound word -­like, which matches the morphological status of holba more closely.

Choctaw 533 (10) ohoyo ‘woman’ → ohoyo holba ‘womanish’ tʃokfi ‘sheep’ → tʃokfoba ‘sheep-­like’ oktʃako ‘blue’ → oktʃakolba ‘bluish purple’ This is the productive use of this marker: speakers may evaluate something as ofi holba ‘doggish’ or alla holba ‘child-­like’ in the same way that a Spanish speaker may choose to produce zapatito ‘shoe + dim’ → ‘little shoe’. The listener will ­interpret the meaning of the derivative on the basis of the standard use of the diminutive rule. Another abstraction of the centrality metaphor is that of ‘approximation’. In his ‘lambda-­abstraction-­specification’ model, Jurafsky argues that a quality can be marked at a lower point on a scale whose prototypical exemplar is the central concept (1996, 554–5). It is then ‘less’. Note that the central concept no longer has to have anything to do with size, due to replacement of the original metaphor (predicate) ‘size’ with some variable, which will serve as a second-­order predicate. In the examples in (11), the holba derivative describes an approximation of a central category. These are derivatives that will generally be lexicalised. (11) piʃoktʃi ‘milk’ → hattak ‘man’ → issi ‘deer’ → lakna ‘yellow’ → haksi ‘drunk; duped’ →

piʃoktʃi holba ‘milky’ hattak holba ‘statue; image of a man’ issoba ‘horse’ laknoba ‘drab’ haksolba ‘simple-­minded’

From the examples above, the word for ‘horse’ issoba can be analysed as ‘an approximation of a deer’, certainly not ‘small deer’ or ‘deer-­like’. The word for ‘drab’ laknoba is a description of a quality that is an inferior – lesser – example of ‘yellow’. Someone who is haksolba ‘simple-­minded’ is an approximation of a person not in full command of his wits, as a drunk or out-­of-­control or duped person might be, yet clearly in a condition differentiable from any of these well-­defined states. Choctaw speakers do not use holba for intensification. They use lexical items such as tʃohmi ‘sort of; somewhat’ and fiihna ‘very; quite’, adjoined phrasally in syntax, to do this. (12)

abiika tʃohmi ‘somewhat sick’ falaja fiihna ‘very long’

2.3  Morphological status of oʃi and holba While most of the evaluative morphs studied have been affixes (Körtvélyessy, forthcoming), neither of the Choctaw forms will pass phonological tests for affixes, even though oʃi is virtually always bound, as are the allomorphs of holba. Two phonological rules are in play here. First is the rule of rhythmic lengthening common to Muskogean languages, wherein, in a string of at least two light syllables (CV), the second and every other even syllable will be lengthened. In Choctaw, a final syllable is extra-­metrical. The presence of a heavy syllable blocks the rule.

Marcia Haag

534 (13) /pi.sah/ ‘he looks’ /pi.sa.lih/ ‘I look’ Rule application: rhythmic lengthening of second light syllable. /pi.saa.lih/ (14) /hak.loh/ ‘she listens’ /hak.lo.lih/ ‘I listen’ Rule does not apply: CVC /hak/ blocks series of CV syllables.

A second general phonological rule is vowel elision, wherein, at word boundaries with vowel clash, the final vowel of the first word is optionally elided. (15) /iti ani/ → /itani/ ‘oak tree’ Rhythmic lengthening does not apply across word boundaries: */itaani/. In oʃi derivatives, we do not find rhythmic lengthening in syllables that would be eligible; hence we have compounding with vowel elision, not affixation. (16) /nita/ ‘bear’ → /nitoʃi/ ‘bear cub’ * /nitooʃi/ There are fewer instances of holba derivatives where the rhythmic lengthening rule might apply because the full form prevents the vowel elision rule from applying. The common derivatives that use the vowel-­initial allomorphs happen to have (C)VC blocking. Since Choctaw is an endangered language, fewer new words are being generated with these allomorphs. (17) /is.so.ba/ ‘horse’ /lak.no.ba/ ‘drab’ Rhythmic lengthening is blocked.

VC + CV + (CV) CVC + CV + (CV)

2.4  Status of evaluatives in sister languages There is a small amount of evidence that the Creek and Koasati languages have extended uses for diminutives (Jurafsky 1996; Munro 1988), and, at least in Creek, for augmentatives (Martin 2010). Creek uses the cognate oci as a true diminutive marker, and also has some of the same semantic categories as Choctaw. In addition, oci is used to differentiate same-­sex kin (mother versus aunt; father versus uncle). Creek and Koasati also have a diminutive based on cognates of ossi ‘small’. The Koasati word for ‘son’ is actually made of both the oʃi and ossi cognates: ocoosi (literally) ‘little son’. Creek uses the osi variant in verbal constructions to refer to a diminutive subject, as in (18). (18) Opaan-­os-­too-­s. dim-­aux -­decl dance-­ ‘(A baby) is dancing.’ (Munro 1988, 541)

Choctaw 535 Munro states that the diminutive -­os-­ may sometimes have word-­formation properties (so that, in her example, see may become glance) under appropriate pragmatic conditions, but that generally this use is ‘triggered’ by a diminutive subject (1988, 542). Chickasaw seems to be as conservative as Choctaw in its uses of its oʃi cognate, oʃiʔ, but does not appear to use holba as an evaluative, although it has this word as a lexeme.

19.3 Dena’ina

Olga Lovick

1 Introduction Dena’ina (also spelled Tanaina) is an Athabascan language spoken in south-­central Alaska in the area surrounding the Cook Inlet. Today, there are fewer than sixty fluent speakers of Dena’ina (own estimate). There are four dialects which differ phonologically, morphologically and lexically (see also Kari 1975a). Dena’ina belongs to the Northern Athabascan language group, a sub-­family of about twenty languages spanning much of western North America from north-­western Alaska to the Hudson Bay in the east and to central British Columbia, Alberta and northern Saskatchewan and Manitoba in the south. The Athabascan languages are renowned for their complex polysynthetic morphology, which is often represented by a templatic structure (see Rice 2000 for a comparative discussion of Athabascan morphology). Verbs have anywhere between thirty-­five (Koyukon, following Jetté and Jones 2000, 759f.) and seven morpheme ‘slots’ (Navajo, following Faltz 1998, 10), depending on the language and, crucially, on the level of detail in the analysis – the difference between Navajo and Koyukon is largely due to the latter. For the Inland dialect of Dena’ina, Tenenbaum (1978, 34) proposes nineteen prefix slots, one stem slot and one suffix/enclitic slot. This template is shown in Table 19.3.1, where inflectional morphemes are shaded grey; unshaded morphemes are derivational or lexical (see discussion below). Not all slots need to be filled in any given verb form. The disjunct boundary is a word-­internal boundary for certain morphophonological processes (for detail see Kari 1975b). The following description is based heavily on Tenenbaum (1978), the only comprehensive study of Dena’ina morphology to date. She (1978, 31) draws a sharp distinction

8

7

6

5

4

3

2

1

0

−1

Negative

Thematic

Semelfactive

Mode

Perfective

Subject pronoun

Classifier

Stem

Enclitics

Inceptive

Conative/Semitransitive

Thematic

Deictic pronouns

13 12 11 10 9

Object pronouns

# (Disjunct boundary)

Distributive plural

Incorporate

Iterative

Adverbial/derivational

Postposition

Object of postposition

19 18 17 16 15 14 #

Gender

Table 19.3.1. Template of the Dena’ina verb (following Tenenbaum 1978, 34)

Dena’ina 537 between inflectional, derivational and ‘thematic’ (lexical) verb prefixes.1 The basic lexical unit of the Dena’ina verb is the verb theme, which is composed of the stem plus one or more lexical prefixes (which may be a zero morpheme). The verb theme is the locus of inflectional morphology. Nouns in Athabascan languages tend to be less complex as they only can be inflected for possession. Nouns referring to humans and dogs may take a plural enclitic. Nouns are frequently of deverbal origin (i.e., they consist of a fully inflected verb form plus a nominalising enclitic). Compounding is a common word-­formation strategy. No other word classes show inflectional morphology. The remainder of this chapter is structured as follows: the two evaluative strategies in Dena’ina are discussed in Section 2. Section 2.1 explains evaluative functions of the gender system while Section 2.2 looks at several enclitics. Section 3 contains a brief comparison to several other Northern Athabascan languages. A brief summary and discussion of evaluative strategies in Dena’ina and other Athabascan languages are offered in Section 4.

2  Evaluative strategies 2.1  Gender prefixes In addition to the well-­discussed classificatory verb system in Athabascan (see for example Basso 1990 for Western Apache), several of the Northern Alaskan languages have a set of ‘gender’ prefixes that function as a basic nominal classification system; they are found in the qualifier zone of the verb, although their exact position is dependent on language and even dialect (see also Kari 1975a, 51; Hargus 2000). Tenenbaum (1978, 132–44) demonstrates the complex interaction of prefixes and classificatory stems in assigning nouns to classes. To state this generally, the verb stems classify objects by shape, while the gender prefixes classify them by type, material or size: Although an exact meaning for these prefixes is difficult to state, general meanings can be suggested: d-­class nouns are usually stiff, made of wood (if the stem allows) or in the case of [noun] class 4, fabric objects, d-­means made of skin. N-­class nouns are usually small and round, or made of flour. Dn-­class nouns are always heavy and dense. (Tenenbaum 1978, 133f.)

Tenenbaum (1978, 133) also mentions a Ø-­prefix which is used for all other objects, including humans and animals. An example of the evaluative function expressed by gender markers can be found in the following discourse example, where the Mouse People tell a man how to cook fish eggs:2 (1) Mary Hobson, Mouse story, in Lovick (2007): (a) 61. Jit gheli dudlik shla va tuk’tenghiłtl’eł. jit gheli dudlik shla v-­a little.bit really bucket little 3obj-­into tu-­k’-­te-­n-­gh-­n-­ł-­tl’eł. water-­ indf.obj-­incp-­gnd-­fut-­2sg.sbj-­clf-­pour ‘Just put a tiny bit [of fish eggs] into the bucket.’ (b) 62. Daghiłtey va tuk’dnghiłtl’idilay! d-­Ø-­gh-­n-­ł-­tey v-­a adjvl-­3sg.sbj-­cnj-­ntr-­clf-­strong 3obj-­into tu-­k’-­dn-­gh-­n-­ł-­tl’it-­ilay water-­ indf.obj-­gnd-­cnj-­2sg.sbj-­clf-­pour-­neg.imp ‘Don’t put too much into it!’

538

Olga Lovick

Lines 61 and 62 both contain the verb theme tu#obj+gnd+Ł+tl’it ‘to pour something (of a particular gender) into a liquid’.3 They differ in mode/aspect inflection (the inceptive morpheme, the conjugation marker and the stem form in 61 indicate future inflection, whereas the lack of inceptive, the conjugation marker, the stem form and the enclitic indicate a negative imperfective imperative in 62). More relevant here is the change in gender. Fish eggs are typically marked by n-­gender (Tenenbaum 1978, 138–40) as shown in 61. In 62, however, when the man is warned against putting in too much, the dn-­gender marker for heavy and dense objects is used. This change from n-­ to dn-­ thus fulfils an augmentative function. (The same dn-­gender marker is used again later in the story when the man, of course, cooks too many fish eggs.) A similar example is shown in (2):4 (2) Augmentative use of dn-­: a. Pete (1977, 11): Shunkda ghunhdi baydalgi qeyeghayqun. sh-­un-­kda ghun=hdi baydalgi 1sg.pssor-­mother-­pej dem=top skin.boat qey-­Ø-­gh-­z-­Ø-­qun 3pl.sbj>3sg.obj-­gnd-­qual-­cnj-­clf-­sew ‘My mother sewed up [the skins for] the skin boat.’ b. Tenenbaum (2006, 94): . . . niłkinqeydnazqun . . . nił-­ki-­n-­qey-­dn-­gh-­z-­Ø-­qun recp-­end.to.end-­distr-­3pl.sbj>3sg.obj-­gnd-­qual-­cnj-­clf-­sew ‘. . . they sewed [the enormous skins] together end to end . . .’ Both verb forms are derived from the same theme, obj+gnd+gh+Ø+qun ‘to sew object’, and both of them refer to the act of sewing up the skin covering for a boat. The (normal-­ sized) covering for the skin boat in (2a) takes the Ø-­gender, whereas the unusual size of the skin covering in (2b) is indicated by the use of the dn-­gender. A quick search of about twenty Dena’ina texts and of the lexical stem file by James Kari (n.d.) revealed that this use of dn-­is fairly productive in a number of verb themes: a particular object takes an inherent Ø-­, n-­ or d-­ gender prefix if it is of normal size, but takes a dn-­ gender prefix if it is unusually large or heavy. Dn-­thus has clear augmentative function. The search also showed that dn-­ is the only gender prefix with this function; it is impossible to use any of the other gender prefixes with, say, a diminutive meaning. This suggests that augmentation is a secondary function of the gender marker dn-­, and that the other gender markers do not have evaluative meaning components. Considerably more productive is another strategy, discussed in the next section. 2.2  ‘Adjectives’, enclitics and suffixes Although most property concepts are expressed by stative verbs, Dena’ina has a small number of uninflectable morphemes often called ‘adjectives’. Lovick (2005, 32) mentions four adjectives: kda ~ tda ‘old, shabby’, gguya ~ ggwa ‘little’, shla ‘little’ and ka’a ‘big’; Kari (n.d.) also lists cheh ~ chegh ‘big’. Often, these adjectives occur as free morphemes modifying a noun phrase (3a); but they also are used as evaluative enclitics (3b) capable of attaching to a number of different lexical categories, as shown in (3b, 4a–b). Note that the meaning of shla ‘little, dim’ is different when used as an enclitic from its use as an adjective.

Dena’ina 539 (3) Adjectival vs. enclitic use of shla ‘little, dim’: a. Pete (2003, 167): Tommy tukda shla ghunhdi, shq’ay daghiłkegh. Tommy tu-­kda shla ghun=hdi T. father-­ pej little dem=top sh-­q’ay d-­Ø-­gh-­n-­ł-­kegh 1sg.obj-­short.of adjvl-­3sg.sbj-­cnj-­ntr-­clf-­big ‘As for Tommy’s father, [he was] a small one, he was shorter than me.’ b. Pete (1977, 13): Tukdashla nunul’anni! tu-­kda=shla nu-­n-­gh-­l-­’an-­ni pej=dim iter-­qual-­opt-­clf-­look-­opt father-­ ‘Let her look at [her] dear father again!’ (4) Combinatorial freedom of -­kda: a. Pete (1989, 3): łik’akda ghin łik’a=kda ghin dog=pej dem ‘that dear old dog’ b. Pete (1977, 21): bingha qenq’a ghu ts’itl’uhqinił’uhkda b-­ingha qenq’a ghu 3sg.pssor-­older.brother house there ts’i-­tl’uh-­q-­n-­i-­ł-­’uh=kda straight-­underside-­areal-­cnj-­pfv-­clf-­co.is.in.position=pej ‘the end of his older brother’s house was sticking out over the bank shabbily’ An interesting point about (4) is that =kda ~ tda seems to have opposite meanings: in (4a), it expresses affection, whereas in (4b) it adds a pejorative meaning component. Both functions of =kda are common in Dena’ina. Some examples of the pejorative function are given in (5).5 (5) Pejorative readings of -­kda; all examples from Kari (n.d.) a. nuntaldaqkda nu-­n-­te-­z-­esh-­l-­daq=kda iter-­gnd-­incp-­cnj-­1sg.sbj-­clf-­sleep=pej ‘it’s no good I overslept’ (the meaning of nuntaldaq is ‘I overslept’) b. ghinikda ghini=kda dem=pej ‘the no-­good one’ (nickname of the brown bear) c. nelnikda nelni=kda steambath=pej ‘shabby steambath’

540

Olga Lovick

As exemplified in (2a) and (3), many kinship terms contain =kda. In these forms, =kda is obligatory. Some examples are shown in (6). (6) =kda ~ tda in kinship terms; all examples from Kari (n.d.): a. shunkda sh-­un-­kda 1sg.pssor-­mother-­pej ‘my mother’ b. shtukda sh-­tu-­kda 1sg.pssor-­father-­pej ‘my father’ c. shchitda sh-­chi-­tda 1sg.pssor-­grandmother-­pej ‘my grandmother’ d. shchitdatda sh-­chi-­tda=tda 1sg.pssor-­grandmother-­pej=pej ‘my dear grandmother’ A possible explanation is offered by Jetté (1907) for the lower dialect of Koyukon. He observes that the depreciative morpheme kāla (k’aale in the modern orthography used in Jetté and Jones 2000, 308) is ‘require[d] to be used when one mentions one’s own relatives’ (Jetté 1907, 9) to express humility in the speaker. It is likely that the Dena’ina forms have a similar origin. Since the pejorative morpheme is obligatory in this language (*shun ‘mother’, *shtu ‘father’ etc. are impossible forms), and because of the semantic shift that gave rise to forms such as (6d), it is now better analysed as a suffix rather than an enclitic in kinship terms. Note also that two evaluative morphemes can be used together, as shown in (3b) above and again in (6d). In (6d), the same evaluative morpheme occurs twice: once as part of the stem for ‘grandmother’ and again with an endearment function, resulting in the more affectionate meaning ‘grandma’. This demonstrates not only the tight fusion of -­tda with the root chi ‘grandmother’ to form the stem chitda, but also the continued productivity of the evaluative morphemes.

3  Comparison to other Athabascan languages The gender system discussed in Section 2.1. is productive only in the Alaskan Athabascan languages and becomes less so as one travels eastward. Holton (2000, 237ff.) discusses ‘adjectival class prefixes’ in Tanacross which he states are cognate with the gender prefixes in Ahtna (following Kari 1990) and, by extension, with those in Dena’ina. His discussion is not very detailed, however, and suggests that the gender system in Tanacross has become less productive. Certainly this is true for the eastern neighbour of Tanacross, Upper Tanana (author’s own field notes), where gender assignment has become very opaque and ­non-­productive. Rice (1989, 608), discussing the d-­gender marker in the Canadian Athabascan language Slavey, points out that it is ‘generally not productive’. In his account of gender prefixes in Dëne Sųłiné, Cook (2004, 175) states that they occur mainly in stative themes but that the gender contrasts are vestigial even there, and that sometimes the occur-

Dena’ina 541 rence of a gender marker cannot be explained semantically. Similar statements can be found in descriptions of other languages. Evaluative enclitics or suffixes similar to those described in Dena’ina are covered in detail in Lovick and Rice (this volume). While the gender system becomes less productive in the eastern Alaskan and Canadian Athabascan languages, the enclitic system remains productive across the Northern Athabascan languages.

4  Conclusion: evaluative morphology in Dena’ina and beyond It appears that Athabascan languages have two morphological strategies for the expression of evaluation. In languages with a fully productive gender system, the dn-­ gender prefix can be used to indicate unusually large size. Secondly, each language has a small set of evaluative suffixes. More on these suffixes can be found in Lovick and Rice (this volume).

Notes 1. In Tenenbaum’s (1978, 32f.) terminology, derivational prefixes are productive, while thematic/ lexical ones are not. 2. Examples are presented as follows: the source of the data is followed by an orthographic transcription. Numbers in the transcription line refer to position of the line in the original narrative, if applicable. The next lines contain a morpheme breakdown indicating underlying morphemes and a morpheme gloss. Morphophonemic processes are not explained in this chapter; the interested reader is referred to Tenenbaum’s (1978) detailed study. The final line contains a free translation. 3. All obligatory lexical and derivational prefixes are listed left to right followed by the (imperfective) stem. Disjunct prefixes are separated by #, conjunct prefixes by +. obj indicates that the verb theme is transitive, gnd that it requires a gender marker. 4. The use of the pejorative in the form for ‘mother’ in (2a) and in the form for ‘father’ in (3a) below will be explained in the discussion of the examples in (6). 5. Some readers may notice that in (5a), the gender prefix precedes the inceptive, whereas it follows the inceptive in (1a) and in the template in Table 19.3.1. This is one of the morphological dialect differences addressed in Kari (1975a, 51); the form in (5a) must be from the Outer Inlet dialect, although this is not indicated in Kari (n.d.).

19.4  Huautla Mazatec

Jean Léo Léonard

1 Introduction Mazatec, an Eastern Otomanguean language spoken in the Papaloapam Basin in the north-­western part of the state of Oaxaca in Mexico, has the highest number of speakers (220,000) among the Popolocan sub-­family (which also includes Popoloca, Chocho and Ixcatec, all endangered languages according to UNESCO criteria). Mazatec dialects can be divided into three groups: the Highlands (Huautla, San Jerónimo, San Lucas, San Mateo, San Pedro, San Antonio, San Lorenzo, San Francisco Huehuetlán), the Lowlands (Jalapa, San Pedro Ixcatlán, San Miguel Soyaltepec) and the Valley or Cañada dialects (among which are Mazatlán and Chiquihuitlán).1 This chapter will deal mostly with the Huautla dialect (96°50 W, 18°15 N).2 Huautla Mazatec has no dominant word order (Dryer and Haspelmath 2011; see also Agee 1993, for the same statement on the San Jerónimo Tecoatl dialect) and is highly head-­ marking3 – adpositions may even be incorporated into the verb. Nominal inflection is limited to possessive suffixes, especially for inalienable nouns (see Pike 1948, 105), whereas verb inflection seems extremely complex at first sight, with stative-­active preverbs and a high degree of suppletive allomorphy (the so-­called conflative paradigms; see Jamieson 1982 and Léonard and Kihm 2010 and 2012 for a more parsimonious description). Huautla Mazatec has four level tones: for example, á is marked high or H; à is mid-­high or h; a is mid, or M; and a is low, or L. Contrast is well illustrated by the minimal pair tsé ‘small’ and tse ‘big’, and at least five contour tones, that is, HM, HL, LM, Lh, ML (for example, áa is high-­mid, or HM). While famous for its very intricate consonant inventory, Mazatec can nevertheless be described more parsimoniously as having a basic consonant inventory of no more than eight contoids (t, ts, č, k, š, x, w realised as b, y) enriched by a voice quality correlation (with the features spread and constricted glottis).4 The vowel system is pentavocalic, with oral and nasal vowels, like most Otomanguean languages. Modern spelling goes according to Spanish conventions: < y, x, ch > stand for palatal segments; (y, š, č), < j > and < ’ > stand for breathiness and creakiness; < n > is used for nasal vowels (in, un, en, on, an); and < z > reads as a retroflex affricate.

2  Evaluative morphology and lexical structure As evaluative morphology (EM) in Mazatec happens to highlight particularly well the continuum between the lexicon and syntax, from word-­formation processes to phrasal

Huautla Mazatec 543 p­ rojections of compounds, we deem it necessary to give a few hints of cycles of word formation in Huautla Mazatec, focusing on adjectives. Root compounding of the type specifier + head is the basic constructional process in Mazatec, with a strong trend for compound units to coalesce, surfacing in synthetic forms. Indeed, in further cycles of compounding, patterns may differ from this basic scheme (root + adj, root + n, etc., as in the lexicon for vegetals, where root complementation frequently occurs). Roots are polyvalent, surfacing into lexical categories through strings of the types (clf + n): they acquire their lexical category through specification – usually by prefixation or cliticisation of specifiers – or complementation, through enclisis. In nouns, a former system of classifiers can be traced back to Proto-­Popolocan (Veerman-­ Leichsenring 2004), for example nise ‘bird’ < noun specifier ni-­ + see ‘sing’. In adjectives, tendencies towards prefixal derivation (that is, with Proto-­Mazatec *x-­; in Huautla, j-­5) can be observed, though many have opacified synchronically. Not only stem but also root allomorphy is quite common in Mazatec dialects: for example, in Kirk’s cognate sets (1966) compare ntí ‘small’ with Ayautla ’chi, ki’nti, Soyaltepec chìí, Jalapa kjìntí and Huautla ’ntí. The palatalised forms ’chi,6 xixi (Ixcatlán) and schi,‘ixi (San Jerónimo) translate as ‘tiny’ (Span. chico, chiquito). The k(j)i-­ prefix stands for a predicative specifier (‘being small’). For the notion ‘big’, Huautla, Jiotes and Soyaltepec have tse, je, Chiquihuitlán xje, chanka and Jalapa chánkà, kitse ‘full’. In elders’ speech, the proclitic ntí ‘small, tiny’ is frequently used as a respectful term of reference (see Cowan 1965, 88 for a similar remark: ’ntík’èn ‘respected dead’), as in mythological narratives, where all elements of the myth tend to be mentioned using a diminutive proclitic.7 Here, diminutive proclisis, instead of diminutive-­attributive enclisis, as shown in (1) and (2) below, can be accounted for by discursive constraints, resorting more to syntax than to lexical compounding. Adjectives in the lexicon appear as single roots, as in san ‘bitter’, chàn ‘brown’, xi ‘sweet’, tàja ‘hard’ and za’ai ‘flat’ (‘hard’ and ‘flat’ appear with rearticulated vowels – a complex nucleus either reaspirated, that is, spread VhV, spelled -­aja-­, or interrupted, that is, constricted V’V); or with a plain stative classifier, as in sase ‘green’; or with an empty property specifier, as in i-­nda ‘mild’ and i-­so ‘blue’. Some adjectival forms result from reduplication, as in tjotjo ‘tasty’, and most adjectives may be expanded with a predicative enclitic, as in naská=kji ‘beautiful’ and ch’ao=kji ‘ugly’. Deadjectivisation requires class inflection preverbs (see Jamieson 1982) such as jkuè ‘rough’ > (jè) sí-­jkuè ‘she or he rasps, makes smooth’, za’ai ‘flat’ > sí-­za’aijòn ‘flattens, makes flat’, with Jamieson’s inflectional class 8 (Jamieson 1982, 149), which we categorise here as a causative (tsi-­in Chiquihuitlán, sí-­in Huautla). Gradation implies adverbal constructions with nguisa (in its modern spelling): Huautla ntaa ‘good’ > nkìsantaa ‘better’ (in Kirk’s notation), Chiquihuitlán ntaja ‘good’ > tuxantaja ‘better’ (Kirk 1966).These paradigms give some hints of the intricate patterns of interaction between inflection and the lexicon in adjectival patterns in Mazatec. A glance at these phenomena has been necessary before handling facts relevant to EM, in which adjectival forms play a crucial role.

3  Basic morphosyntactic patterns for evaluative morphology Basic patterns for EM are given in (1) and (2): interestingly, (1a) conveys more of a diminutive meaning than an attributive denotation, as ‘small hen’ would be denoted straightforwardly by a taxonomic compound, as in (1b).

544

Jean Léo Léonard

(1) a. xo’nda ndí hen small ‘chicken’, Span. pollito b. xo’nda nise hen bird ‘small hen’, Span. gallina chiquita That is, a truly small hen would be qualified as a ‘bird hen’ (1b), whereas the compounding strategy for ‘chicken’ requires the attributive enclitic pattern in (1a). In (1c), a few unmarked diminutive patterns are enumerated, which are quite close to mere attributive constructions – as opposed to more marked diminutive patterns, available in (3) and (5). (1) c. tsoti ndí girl small ‘little girl’, Span. niñita natsé ndí rabbit small ‘little rabbit’, Span. conejito ndi’a ndí house small ‘small house’, Span. casita Augmentative patterns are given in (2a), but turn out to be quite similar to mere attributive constructions as well, whereas data in (2b) hints at further evaluative expansions, as with ndí chito ndí ‘small little cat’, with bilateral compounding, and at pejorative connotations conveyed by superlative cliticisation with =la, in (2c). (2) a. natsé je rabbit big ‘big rabbit’ ndi’a je house big ‘big house’ b. chito ‘cat’ chito ndí cat small ‘small cat’ ndí chito ndí small cat small ‘small little cat’ c. ti ’mbá boy fat ‘a big/fat boy’ tsoti ’mbá girl fat ‘a big/fat girl’ ti ’mbá=la

Huautla Mazatec 545 boy fat=sup ‘quite a big/fat boy’ tsoti ’mbá=la girl fat=sup ‘quite a big/fat girl’ Although EM in Mazatec is conveyed mostly by free compounds of the n + adj type (see Section 2 above), it does not rely merely on syntax. This chapter aims to show that if Mazatec does not display a rich inventory of evaluative markers of the affixal type, EM should nevertheless be handled (as in other languages of the Otomanguean stock) as a most relevant standpoint in understanding subtle construction patterns deeply embedded in the syntactic typology of lexical patterns of compounding.

4  Evaluative morphology and the lexicon-­syntax continuum In order to enhance the relevance of diminutive-­attributive chains such as (3a) and of cyclic tautological strings such as those in (3b–d), we shall compare these analytic – that is to say, over-­analytic – patterns with synthetic inflectional strategies elsewhere in the language. To do so, we will take a look at verb inflection allomorphy through the stem for ‘grind’ in Section 4.1, which happens to be the same root as that of yo’bi, a synthetic nominal compound. In Sections 4.2.1 and 4.2.2 we will enter more into the details of complex analytic strings, both for diminutives and for augmentatives. A lexeme such as yo’bi ‘cloud’ can be analysed as |yu| ‘dough’ + |wi| ‘go’ (n + mlv) which can enter into a predicative pattern (Span. nuve pequeña) as in (3a), with further possible expansions of the pleonastic string patterns (3b–d): (3) a. yo’bi ndí cloud small ‘small cloud’, Span. nubecita b. yo’bi ndí tsé cloud small little ‘small little cloud’ c. yo’bi ndí tsé ndí tsé, cloud small little small little ‘small little small cloud’ d. yo’bi ndí tsé ndí tsé ndí. cloud small little small little small ‘too much of a small little small cloud’ Therefore, chains of embedded patterns of the type XYZ(Y). . ., (which we have already called ‘tautological strings’ in the introduction to this section) where X = N, Y = adj 1, Z = adj 2 – for example, X = yo’bi, Y = ndí and Z = tsé – can be used for diminutive and superlative expansions.8 4.1  Conditions for root inflectional allomorphy Compare a nominal form, such as the one in (4a), analysed as clf-­root, with inflectional paradigms for ‘grind’ in (4b). In the former, the same polyvalent root yo ‘dough’, ­combines

546

Jean Léo Léonard

with a nominal specifier na-­, as opposed to the noun yo’bi ‘cloud’, which patterns with a lexical complementiser (n root yo + v root bi = compound noun). In (4b), the nominal root (n root) yo is associated with mlv b’a-­‘carry’, belonging to one of the main preverb classes, which displays intense stem allomorphy, as do all preverbs with a labial onset in Mazatec. The examples in (4b) show the extreme border of the cycle of root coalescence in Mazatec stems, as opposed to the other extreme of morphosyntactic polarity, that is, the epithetic, diminutive and superlative cycles presented above in (1)–(3): (4) a. na-­’yo clf-­ dough ‘dough’ Span. masa b. (jè) b’a-­u (3.pro.d) mlv.hab-­dough ‘she or he grinds’ (an) b’-­o-­a (1sg.pro.d) mlv.hab-­dough-­1sg ‘I grind’ (ji) yu-­i (2sg.pro.d) dough-­2sg ‘you grind’ (jòn) yo-­ò (2pl.pro.d) dough-­2pl ‘you (folks) grind’ (jin) yu-­ijin (1pl.excl.pro.d) dough-­1pl. excl. ‘you (folks) grind’ (ñá) yo-­à (1pl.incl.pro.d) dough-­1pl. incl ‘we grind’ 4.2  Tautological chains for evaluative morphology 4.2.1 Diminutive Evaluative forms in (1)–(3) show enhanced analytic patterns, whereas nominal forms such as yo’bi ‘cloud’, nise ‘bird’ and natsé ‘rabbit’ display enhanced synthetic strategies through prefixal derivative patterns as in (4a), while verbal forms display allomorphic stems in which the specifier (that is, the preverb) coalesces with the root |yu| ‘dough’, as in (4b). However, at a deeper lexical level, all these forms are compounds. EM enhances this basic trend in Mazatec, as it does for most Otomanguean languages. The fact that morphosyntactic patterns for EM in Table 19.4.1 are obtained through compounding or, stated more properly, through recursive permutation (rows 1–3) or expanded cyclical phrases (rows 4–6), for a language where stem allomorphy strongly surfaces in inflectional patterns, in spite of a very low rate of derivation at the lexical level – except by prefixal specifiers as in (4a) – suggests that EM processes may ­highlight basic typological properties of a language that would otherwise appear less overt. This Mazatec data exemplifies how EM may work to display the lexico-­syntactic interplay – that is, as a prism. At least, in this language, EM enhances both the basic compounding trend and recursivity, as summarised [or] schematised in (5a):

Huautla Mazatec 547 Table 19.4.1. Adjectival chains for evaluative morphology in Huautla Mazatec 1 2 3 4 5 6 7 8

(5)

Y

X

ndí ndí ndí ndí ndí ndí

jno jno jno jno jno jno jno jno

n pred dim pred dim

1 sup 2 sup 3 sup 4 sup

Y

Z

Y

Z

Y

ndí ndí ndí ndí ndí ndí

tsé tsé tsé tsé

ndí ndí ndí

tsé tsé

ndí

‘cornfield’ ‘small cornfield’ ‘small cornfield’ ‘very tiny cornfield’, ‘most tiny cornfield’ ‘quite a very tiny cornfield’ ‘quite a piece of a very tiny cornfield’ ‘too much of a very tiny cornfield’

a. X = n, Y = pred, X YZ = epi, X YZY = dim pred, X YZ YZ Y = sup

4.2.2 Augmentative Augmentative expansions follow the same YZ pattern of pleonastic or tautological strings: the sets of augmentative-­predicative patterns given in (5b) combine in (5c) to give analytic, cyclic expansions of the diminutive-­attributive matrix. The phrase in (5d), which expands the evaluative chain, could be interpreted as a superlative of the syntactic type, rather than of the morphological type shown in (2c), as no enclisis of the =la superlative is here required, all while conveying the same evaluative connotation at the phrasal level instead of at the word level: (5) b. jno tse jno ng’a cornfield big/huge ‘a big cornfield’ c. jno ng’a tse cornfield huge big ‘quite a big cornfield’ (Span. una milpa regrande/una milpa grandecita) d jno ng’a tse ng’a cornfield huge big huge ‘too much of a big cornfield’ (Span. (una) milpa demasiado grande) In the same way as processes of lexical formation in Mazatec resort to specification (noun classifiers for noun stems, as in (4a), mlv for verbal stems, as in (4b)) in the first cycle, which could be called the stem derivation cycle, the second cycle resorts more to complementisation. As we saw in Section 2, tautological strings as in Table 19.4.1 or (5c–d) fall into the range of the later cycle.

5  Hints at predicative morphosyntax As this language relies more on phrasal strings to convey finer-­grained magnitudes of adjectival properties, not only would it be hardly possible to obtain diminutive forms in Mazatec, but predication itself may also require circumclitic predicative forms, as in (6a). This process neatly recalls that of the pleonastic strings typical of Mazatec EM, and may confirm the utility of EM for typological purposes.

Jean Léo Léonard

548 (6) a. naská=kji (ndí) najmá xo=kji nice=pred (small) bean dem=pred

Compound expressions of diminutive and augmentative phrases may enter into intricate patterns of admirative constructions, as in (7), with an adverbial proclitic intensifier to and lexicalised forms of a relativised predicative operator (xkia < rel xi, pred k(j)i, phatic =’a): to xkia. (6) b. (to) kjoan-­ská=ni abstract-­foolish=cop (int) ‘it’s a joke’ (7)

to

int

xkia chito=ni rel.pred cat=cop

to xkia ndí chito=ni int rel.pred small cat=cop ‘what a nice lovely cat!’ (Span. gato bonachón), to xkia ndí chjota=ni int rel.pred small person=cop ‘what a nice person!’ Enclitic =ni is called ‘instrumental’ by Cowan (1965), but could be glossed more p­ roperly as a predicative particle, or even more properly as a copula, akin to cleft sentence patterns such as presentative c’est in French.

6 Conclusion Mazatec turns out to be an exemplar language as far as EM and its implementation in the lexicon and grammar are concerned, especially through the process of pleonastic chains resorting to lexical morphology of compounds. Though Mazatec can be considered as highly synthetic and, to some extent, as a language of the fusional type (see examples in (4b) above), its deep structure depends more on compounding strategies than on anything else. A complex array of processes linked with EM points to cyclic tautological compounding as a basic property of the language, which deserves to be taken into account within the framework of EM as part of general morphology.

Acknowledgements We thank the Institut Universitaire de France for providing the conditions and fundings for fieldwork in fall 2012 and summer 2013, within the framework of the ALMaz  – a Mazatec geolinguistic project included in a broader endeavour, the MAmP project (2009–14); see Léonard (2010) and Léonard and Kihm (2012). We also thank the Labex EFL, strand 7, operation EM2 (Cross-­Mediated Elicitation of Endangered Languages) for additional support in data processing. We are very grateful to Marijana Petrović (Lacito, CNRS, France) and to Bien Dobui (University Paris 3) for revising the initial version of this paper.

Huautla Mazatec 549

Notes 1. Cf. Léonard, Dell’Aquila and Gaillard-­Corvaglia (2012) for a revised classification. See Gudschinsky (1958) and seminal studies by Kirk (1970) or Casad (1974, 46–9) for classification of Mazatec dialects according to mutual intelligibility. On the Chiquihuitlán dialect in particular, see Jamieson (1988). An outstanding description of Huautla Mazatec morphology can be found in Pike (1948, 95–165). 2. Most of the Huautla data presented here have been elicited by the author during fieldwork sessions in September 2012, with Clementina Elodia Cerqueda García, school teacher from the Naxó Kafè school (CMPIO or Plan Piloto), and in July and August 2013, with Juan Casimiro Nava, Mazatec linguist and school teacher, translator of the official Mazatec handbooks for primary schools in the Huautla region. Both speakers were about fifty years old. Elicitation of evaluative morphology (EM) was eased by the fact that Spanish – the local Dachsprache or ‘roof language’ – provides a wide array of diminutives and augmentatives, as shown in translations of English glosses. 3. On this notion, crucial for typological linguistics, see Nichols (1986) and Lehmann (2005). 4. See Golston and Kehrein (1998), as opposed to an earlier analysis in terms of consonant cluster sets in Pike and Pike (1947). 5. See Gudschinsky (1958, 473). 6. Chi for ‘small’ is also found in Huautla, in lexicalised evaluative forms, such as ’ndí chile’a ‘my sweet baby’, literally ‘small mouth (of mine)’; final -­a stands here as a reduced form of tso’a ‘mouth’. 7. From our own fieldwork in San Andrés Hidalgo, with Leonardo Martínez Moreno (Chano Moreno) and with Fernando Ernesto Juárez García (FEJG) and Agua de Canoa (Huautla de Jiménez), both interviewed in autumn 2011. For example, in FEJG’s narrative about the Chikon Tokoxo: ndí chjota chínganà = dim people ancient ‘our respected elders’, nga fìsíchjèele ndí cho ‘(so) that they came to steal his small animals ’ – of the Chikon Tokoxo. 8. One of our informants, Juan Casimiro Nava, while not rejecting patterns such as (3a–d), deems them slightly ‘far-­fetched’, though he agrees they could be used in literary Mazatec for stylistic purposes, for instance in poetry. These remarkable analytic and cyclic chains of adjectives available for EM in Huautla Mazatec therefore may be considered as standing at the borderline between structural and stylistic resources, instead of currently functional patterns. Nevertheless, our main informant, Clementina E. Cerqueda, does confirm these patterns, strongly asserting that they are in use in her own network of speakers. This recalls how much diaphasic variation reminds of dialectal (diatopic variation), though ‘diaphasic isoglosses’ within society may be much harder to fathom than diatopic isoglosses within a bounded territory.

19.5 Huave

Maurizio Gnerre

1 Introduction Huave, a linguistic isolate sharing many Mesoamerican areal features, is spoken by approximately 17,000 people settled along the extreme south-­eastern Pacific coast of the state of Oaxaca, Mexico. Four main Huave variants are recognised, the most vital being that of San Mateo del Mar, where speakers (approximately 12,000 people) refer to themselves as ikoots (1pl ‘we all’) and call their language Ombeayiüts ‘our (maximum inclusion) mouth’. Some main features of Ombeayiüts are: • (i) a rich system of phonological ‘harmony’ characterised mostly by palatal/ non-­ palatal oppositions, encompassing most morphological-­ derivational strategies; • (ii) a large number (perhaps over 800) of verbal-­nominal-­adjectival roots, while all the other roots are word-­class specific (i.e. nominal, adjectival, pronominal, adverbial, numeral, etc.); • (iii) a moderately fusional morphology including total or partial verbal-­nominal-­ adjectival root reduplications, prefixes, suffixes, two infixes and two clitics; • (iv) a verbal system built mostly on aspectual values, with an important role played by the realis/irrealis opposition (as in example (17)); verbs listed below are quoted in the 3sg. realis form, with glosses provided in the English infinitive form); • (v) a basic VOS/VSO grammatical order (as in example (6)), with mod-­n (as in example (2)), and a genitive construction of the ’3poss-­X Y’ (for ‘Y’s X’) type (as in example (1)). Speakers of Ombeayiüts take advantage of roughly four main strategies to codify quantitative (QUAN.: MORE/LESS, BIG/SMALL), and qualitative (QUAL.: GOOD/BAD, ENDEARMENT/NO ENDEARMENT, POLITE/IMPOLITE) evaluations and representations. While in several cases the semantic/functional distinction between quantitative and qualitative evaluation is straightforward, in others it is not, and these evaluations are not so clear cut, and not easy to sort out. The four evaluative strategies are grounded on some dimensions found in three of the five features listed above, and pertain basically to two types: root-­internal and root-­ external. Strategies of the first type are grounded on feature (i), the phonological one (mostly on the palatal/non-­palatal opposition), and on feature (iii), the morphological one

Huave 551 (mostly on total or partial root reduplication). Strategies of the second type are grounded on another dimension of feature (iii), i.e. on affixation (prefixation and suffixation only), and on feature (v), the syntactic-­typological one, with the implementation of independent adverbial modifiers. This last strategy is likely to be a reflex of contact with Spanish, not evident when Huave quantifiers xowüy and xiyay, both ‘very, much’, are implemented, but fully evident when Spanish loan más is (frequently) used. The four strategies can combine. Strategies and their values (QUAL. and QUAN. only) are summarised in Figure 19.5.1. Root-internal

Root-external Reduplications

Q U A L.

Syntactic

Affixes (Prefixes/Suffixes)

Calques(?)

Loans

cheter/tsotor -rang-rang, -xom-xom

Q U A N.

nadam-­dam

an-ka-; -V(V)y

xowüy

más

Figure 19.5.1.  Morphological strategies and values expressed in Huave

2  Morphological evaluative strategies The strategies listed above will be described and exemplified in order, i.e. moving from root-­internal (phonological) strategies (Section 2.1) to root-­external ones (Section 2.2, implementing reduplications and affixes), and to syntactic ones (Section 2.3, use of independent adverbs). 2.1  Root-­internal strategies 2.1.1  Phonological alternations, with or without reduplications The focus of the following examples is on both vowel (V) and vowel and consonant (either V+C and C+V) palatal/non-­palatal alternations. We will list examples of both types that can be also interpreted as cases of phonetic iconicity. Both types occur with (partial or full) root reduplications, which is perhaps the most frequent morphological strategy implemented in evaluations. Furthermore, in several examples a very frequent suffix of Ombeayiüts, -­V(V  y, occurs, playing an almost central role in evaluative morphology. V alternations Examples include: • /oe: i/: a-­toel: a-­til: a-­toel ‘to gore, to toss’, but, with a human agent: ‘to punch, to blow (with a fist)’, but: a-­til ‘to sting, to pierce through’. • /oo: ee/: a-­moog: a-­meeg; a-­moog ‘to put a large container with its open side down (for instance, a canoe)’, but: a-­meeg ‘same meaning referring to a small container

Maurizio Gnerre

552

(for instance, a cup)’. As should be clear, in this, as well as in several other cases which follow, it is the object of the verb that is to be evaluated. • /ooe: iie/: a-­kooen: a-­kiien; a-­kooen ‘to fold a blanket, a sheet, a curtain’: a-­kiien ‘to fold a napkin, a handkerchief, but also a corn leaf to wrap a tamal of corn paste’. V alternations with root reduplications Examples include: • /ü: i/: a-­wüch-­wüch: a-­wich-­wich-­ay; a-­wüch-­wüch ‘to swing, to rock, something large, such as a circus swing’; a-­wich-­wich-­ay ‘same meaning for a small swing, or a hammock’: (1)

a-­wich-­wich-­ay mi jow nine 3-­swing-­swing-­ints 3.poss hammock child ‘she or he swings the child’s hammock’

• /ea: i/: a-­teap-­teap: a-­tip-­tip-­eay; a-­teap-­teap ‘wavy movement of a sail, while it is set up’; a-­tip-­tip-­eay ‘wavy movement when the wind is weak (and the sail is not small)’. Both in this and in the above alternation, reduplications are iconic of movements being repeated. V+C or C+V alternations Examples include: •  /os: ex/:1 a-­lojngos-­üy: a-­lejnguex-­ay; a-­lojngos-­üy ‘the land is uneven, with dunes and hills’; a-­lejnguex-­ay ‘same meaning, but referring to small dunes or hills’. • /ots: ech/: adj, tokots: tekech; tokots oleaj ‘short person (impolite); also referring to trousers or water-­pipes’; tekech ‘short person (polite); also referring to several objects such as a rope, or a thread’. This is an example of the polite/impolite evaluations codified by other forms presented here, as under /tso/:/che/ and /tsü/:/chi/ below. • /rrar:rir/: a-­ndarrar: a-­ndirir; a-­ndarrar, ‘to lay out, something large, such as fishing nets’; a-­ndirir ‘to spread(as a small bird’s wings)’. • /sa: xi/: adj, pasak: pixik; pasak, ‘ruffled, referring to human hairs and to large branchy trees’; pixik ‘(a plant) with many leaves’: (2)

ne-­pixik xa-­mbaj adj-­leaf 1.poss-­basil ‘my basil is leafy’

• /tso: che/: tsoto-­m: chete-­m; tsoto-­m ‘sits down (in an impolite or improper way)’; chete-­m ‘sits down (properly)’. A mother can tell her child: (3) chete-­r najneaj! sit-­ imp.2 well ‘sit properly!’

Huave 553 • /tsü: chi/: a-­tsüp: a-­chip, achip-­ip (with partial reduplication) ‘glossy, slippery’; na-­tsüp, to refer to slippery, muddy ground; moderately impolite when referring to a bald man: (4)

na-­tsüp mi-­omal Chico adj-­glossy 3.poss-­head Chico ‘Chico’s head is glossy’

•  /tsa: chi/: a-­tsajlas ‘to spread, to stretch, to open (as a bird’s wings)’: a-­chijlix, an opposition mostly in endearment, as the second form is used for a flower ‘to blossom’ or for a child opening his or her arms: (5)



a-­chijlix owix a nine det child 3-­open arm ‘the child opens his or her arms’

V+C or C+V alternations with partial root reduplications Examples include: • /ots: ech/: ich: a-­kots-­ots: a-­kech-­ech: a-­kich-­ich; a-­kots-­ots ‘to catch, to get hold, with strength, or something heavy’; a-­kech-­ech ‘to take something not too heavy’; a-­kich-­ich ‘to take something light or take from the tip’. V+C or C+V alternations with full root reduplications Examples include: • /ats: ich/: a-­lats-­lats: a-­lich-­lich; a-­lats ‘to spread, to lay out, a rope, used also for electric cables’, a-­lich ‘to spread a thread’. • /ants: inch/: a-­rrants-­ats: a-­rinch-­ich: a-­rrants-­ats, ‘to itch, to tingle, also strong electric shock’; a-­rinch-­ich ‘to tickle, also slight electric shock’; ne-­rinch-­ich to be ticklish’. 2.2  Root-­external strategies 2.2.1 Reduplications The reduplicated verb forms shown above (in (1), and verb forms under /ü/:/i/, /ea/: /i/ and /os/: /ex/) are followed by an emphasising (or intensifying (int) suffix -­V(V)y (ay, üy, eay, etc.). While their meanings express repetitions or reiterations of movements or physical shapes (such as ‘dunes’ or ‘hills’), several reduplicated forms followed by the same suffix carry meanings quite divergent from their default forms, lexicalised mostly on the QUAL. (GOOD/BAD) parameter, frequently codifying moral negative, or even stigmatised, evaluations. In some cases, however, these evaluations are only mildly negative, or rather ironic or joking, as is the case for the reduplicated form of a-­rang-­ ‘she or he does’: a-­rang-­ rang-­üy ‘she or he boasts’, said, for instance, of somebody strutting about in a new dress, or showing off with a superior attitude. Other similar lexicalisations express a range of ironic meanings, as in the adjective derived from a-­ndeak-­ ‘to speak’, ne-­ndeak ‘gossiper, overstater’, which when reduplicated conveys an even worse meaning: ne-­ndeak-­ndeak-­ üy-­an ‘somebody who talks about invented, overstated, or false contents, and even talks

Maurizio Gnerre

554

to himself or herself’. From the root -­lomb/ -­lemb ‘to stand up’ (a ‘positional’ in the grammatical terminology used for Mesoamerican languages), a transitive form is derived (from lomb-­) through the causative affix -­ch: a-­lomb-­och (causative) ‘to raise, build up, but also to calm, to pacify’. However, when the anti-­passive infix -­j-­ is inserted, and the int suffix -­V(V)y follows, some forms with strongly negative meanings are derived, as: a-­lombo-­j-­ ch-­ay, ‘to boast, to think oneself’, and the adjective ne-­lomb-­oj-­ch-­ay ‘vain, conceited’. These forms when (partially) reduplicated assume an even more negative meaning: a-­lombo-­j-­mbo-­j-­ch-­ay ‘to be a very conceited person’. Some other roots, however, when reduplicated and suffixed with -­V(V)y, convey a strongly pejorative or stigmatised evaluation of somebody’s behaviour. Some examples are: a-­xom ‘to find’: a-­xom-­xom-­üy ‘to take advantage (of something, or somebody)’, as in: (6)

a-­xom-­xom-­üy a naxey ma-­waiich a nine det child 3-­find-­find-­int det man 3.irr-­deceive ‘the man took advantage of the child, deceiving him’

Somewhat similar examples are provided by roots which, when reduplicated and suffixed with -­V(V)y, are lexicalised with a negative meaning, such as: a-­xot ‘to hide’ > ne-­xot-­ xot-­üy ‘somebody who always hides, revealing a guilty conscience’; a-­sok, ‘to enter, to put’ > ne-­sok-­sok-­üy (nipilan) ‘(a person) who enters or walks wherever she or he can’, assuming the pejorative meaning of ‘thief’; a-­jüy ‘to walk’ > a-­jüy-­jüy (najtaj),’(a woman) who goes out (around) at any time’, assuming the meaning of ‘prostitute’. These evaluative semantic drifts could be explained, at least in some cases, by the local conceptualisation of the motivations for repeating the same action many times. So if -­rang, ‘to do’ could be understood as doing something only once, the insistence on ‘doing and doing’ the same action several times can be interpreted as a ‘performance’, and therefore a way for somebody to show off. Another example is provided by -­xom ‘to find’, a meaning conveying the value of a single and ‘chance’ occurrence; when reduplicated and emphasised, the meaning of ‘chance’ is lost, and the semantic drift leads to a ‘search’, possibly an intentional one, which, in a society without any ‘capitalistic’ or ‘entrepreneurial’ spirit, could reveal bad intentions. 2.2.2  Affixes Some uses of -­V(V)y with verbal-­nominal-­adjectival roots have been exemplified, but the same suffix also has other functions among the evaluative strategies, such as that of ‘emphasiser’ or ‘intensifier’ of a specific property or a quality. In Gnerre and Cuturi (2014) we identify approximately ten functions for it, displayed with most word classes. Here I will exemplify only three of these classes: adjectives, numerals and adpositions. Adjectives With adjectives, -­V(V)y often codifies ‘superlatives’, as in raan ‘white’: ra-­j-­n-­üy ‘very white’ < *raan-­üy (with the phonological process VV>Vj, quite common in Ombeayiüts) and in ne-­jinin ‘glossy, polished’: ne-­ji-­nin-­eay ‘very glossy’. (7) peats pich na-­weat-­üy at nej peats titiüm tortilla cotton adj-­tasty-­int as 3 tortilla bean ‘the tortilla (made with seeds) of cotton is very tasty, as is the one made with beans’

Huave 555 Some alternative strategies, part of the derivative morphology of the language, are available to speakers to codify superlatives from default adjectives. One of these is provided by reduplications, such as a-­rraj ‘to shine’ > ne-­rraj ‘shiny, light’ > ne-­rra-­j-­rraj ‘very shiny, very transparent’. Numerals With numerals, -­V(V)y is lexicalised in a few (perhaps only two: ijpüw ‘two’ and nomb, one of the forms for ‘one’) cases, producing derived forms with evaluative (both negative) meanings quite divergent from the plain numeral meanings. In (8), for example, ijpüw-­eay assumes the meaning of ‘very few’:2 (8) xa-­noj ta-­ndok ijpüw-­eay tixem 1.poss-­husband 3.pst-­to fish num.two-­int shrimp ‘my husband caught very few shrimps’ In the derivation from nomb ‘one’ we find added to -­V(V)y (-­üy) the clitic -­an absolutive, producing nomb-­üy-­an, literally ‘only one’, but actually meaning ‘unique’, in a negative evaluation such as ‘silly, stupid’; for example: (9)

nde ndoj sa-­jaw i-­ngün ler-­nomb-­üy-­an since when 1.pst-­see 2-­get drunk compl.2-­num.one-­int-­abs ‘since I saw you get drunk, you have became very silly’

Adpositions The adpositions (adp) tiül ‘in, inside’ and wüx ‘on, over’ are modified by -­V(V)y (til-­üy and wüx-­ay, acquiring different meanings: Cuturi and Gnerre, forthcoming). The second modified adposition, followed by the abs -­an, becomes wüx-­áy-­an, acquiring a positive evaluative meaning as ‘at the right point, good’: (10)

wüx-­ay-­an

arrar xe-­pow hot 1.poss-­oven ‘my oven is hot at the right temperature’ adp.loc-­int-­abs

An example of -­an carrying an evaluative negative meaning is provided by a frequently utilised Spanish–Ombeayiüts syntagm, puro ombeáy-­an ‘only mouth’, referring to somebody who speaks without thinking, or without care in what she or he says. Parallel usages with a negative evaluation are exemplified by a-­piüng ‘to say, to state’ > a-­piüng-­an ‘to say (only)’, i.e. ‘to lie’. 2.2.3 Augmentatives When preceding verbal forms, an-­carries augmentative values (getting close to those carried by -­V(V)y). With many roots, an-­occurs followed by -­ka-­, forming a complex prefix an-­ka-­, which provides an augmentative (aug) semantic value, as in an-­ka-­kants, ‘very red’ < na-­kants ‘red’. From a-­rinch-­ich ‘to tickle’ the form an-­ka-­rinch-­ich is derived: (11)

an-­ka-­rinch-­ich xe-­owix kos sajna-­n itch 1.poss-­arm because swollen.1.ptcp ‘my arm itches very much, because it is swollen’ aug-­to

Maurizio Gnerre

556

Another, not very frequent, augmentative prefix is pa-­, usually producing forms carrying negative evaluations, as in a-­jaw ‘to watch, to observe’ > pa-­jaw ‘telltale’. 2.2.4  Superlatives and comparatives Beyond the devices codifying superlatives discussed above, there are a few lexical items which intrinsically carry superlative and comparative values, the outcome of lexicalised morphological processes. Among the superlatives are chingüy, where the -­V(V)y ending is added to a form *ching-­that we cannot compare with other forms, and meaning ‘the minor, last one’: (12) aaga nine nench kiaj nej chingüy tiül na-­kwal-­aran det child boy dem 3 the minor adp.loc nmls-­offspring-­gen ‘this little boy is the last of the offspring’ Comparative constructions of inequality are not so frequent in Ombeayiüts discourse, and perhaps before contact with Spanish these were not present in the language, although some rare lexical hints of intrinsically comparative forms (and constructions) can be found. One of these, kicheech, possibly derived from kich (both an adjective, ‘small’, and a noun, ‘child’), carries the intrinsic comparative value ‘minor’. This lexical item is preserved to refer to ‘Christmas’: ‘Kicheech Pasca’ ‘minor feast’, as opposed to Nadam Pasca ‘big feast’, i.e. Easter (Pasca < Span. Pascua). Kicheech, however, can play a full comparative role when followed by qué: (13) xe-­kwal kicheech qué aaga nine nench kiaj 1.poss-­son minor than det child boy dem ‘my son is younger than this child’ To codify evaluations of equality, the adv at is frequently used in many constructions and combinations (as seen in example (7)), such as at-­an ‘equal’, several other forms, and a verbalised form with personal prefix; for example: (14)



s-­at ik 1-­equal 2 ‘I am like you’

3  Syntactic strategies Additional strategies for both superlative and comparative constructions available in the language are not morphological but syntactic, as they are made possible through independent intensifying adverbs, probably a calque (‘cast’) from Spanish muy. The adv xowüy ‘very’, however, is derived from the root xey-­: xow-­, (a-­xey-­ay ‘to multiply’) and is related to the adj xeyay ‘much, many’. So, once again, to form such an adv/adj the same m ­ orphological derivation seen above is implemented. In its grammatical/discursive implementation, xowüy precedes adjectives which are morphologically modified, usually through a prefix la-­, as in na-­kants ‘red’: xowüy la-­kants ‘very red’, and na-­jneaj ‘good’: xowüy la-­jneaj ‘very good’. With some adjectives a personal prefix can also be used, as in (to quote from the clever rabbit in a famous tale):

Huave 557 (15)

Nde ma-­tsamb xik, kos xowüy s-­apateay! neg irr-­eat 1.obj because very 1-­small ‘Don’t eat me, because I am very small!’

Here s-­apateay is a form of the adj n-­apateay ‘small’, preceded by the first person prefix s-­. On the other hand, the comparative constructions made possible through the Spanish loanwords más ‘more’ and que ‘than’ are fully syntactic, and do not require any morphological modifications of the adj. Several key function words (see Suárez 1983, 135–7) have entered Ombeayiüts, as well as several other Mesoamerican languages, since the early contact with Spanish. These loanwords allow speakers to produce fully comparative constructions, where más is used as an adverb, as in Spanish, even omitting the second term of the comparative construction; for example: (16)

más ne-­kwiür xe-­kawüy more fast nmls-­run 1.poss-­horse ‘my horse is faster . . .’ (Stairs Kreger and Scharfe de Stairs 1981)

A form in Ombeayiüts that sometimes is used to introduce the second term of the comparison after más is kikwantes, probably an old Spanish syntagmatic loan of obscure origin, fully integrated, however, into the language, and carrying a meaning/function such as ‘than’: (17) más na-­jneaj sa-­na-­rang kiakwantes ikona more adj-­good 1.fut-­irr-­do than 2 pl ‘I will do better than you all’ It should be noted that más preceding an adjective, as in (18), does not require modification, as was the case in the superlative constructions already shown with adjectives preceded by xowüy. Frequent comparative constructions in present-­day Ombeayiüts take full advantage of both más and que, as in: (18) más na-­jneaj kawüy que burr more adj-­good horse than donkey ‘the horse is better than the donkey’

4 Conclusion In Ombeayiüts a set of morphological strategies is implemented to codify quantitative and qualitative evaluations. These extend from the root-­internal V and V+C (or C+V) palatal/ non-­palatal oppositions, to roots reduplications, to the implementation of suffixes (such as -­V(V)y) and prefixes (such as an-­ka). These can modify not only verbs and adjectives, but also some numerals and adpositions. An additional strategy, fully root-­external, and limited to the codification of superlatives, is the use of an independent adv, likely to be a calque from Spanish, which affects adjectival morphology. Finally, to codify most comparative constructions, two loans from Spanish were introduced.

558

Maurizio Gnerre

Notes 1. Symbols used, following current Ombeayiüts orthography, are as follows: ü: high back unrounded vowel [ï], j: [x], VV: double or long vowel [V:]. 2. This usage could be interpreted as a parallel with, if not a calque from, Sp. ‘dositos’ (two-­ Diminutive affix), carrying the same meaning of ‘very few’.

19.6 Inuktitut

Richard Compton

1 Introduction Inuktitut belongs to the Inuit sub-­branch of the Eskimo-­Aleut language family and is spoken in the Canadian Arctic. Inuktitut is part of a dialect continuum extending across the Arctic from Alaskan Inupiaq in the west to Greenlandic dialects in the east. Although the term ‘Inuit’ is typically used to refer to the Inuit people, it is also used in the literature to refer to the dialect continuum as a whole. Though much of the description of evaluatives below will extend to all of Inuit, it will focus primarily on data from the South Baffin subdialect of Inuktitut.1 Inuktitut is a highly polysynthetic language in which verbal complexes can contain an array of optional elements corresponding to modals, adverbials and restructuring verbs, in addition to obligatory elements such as tense, mood and agreement, as illustrated in (1) and (2):2 (1)

(2)



Sinikannigasuariaqaqtuq. sinik-­kanniq-­gasuaq-­giaqaq-­tu-­q sleep-­again-­try-­should-­decl.intr-­3sg ‘He or she should try to sleep again.’ Niringaannguarunnaqtuq. niri-­ngaaq-­nnguaq-­gunnaq-­tu-­q eat-­instead-­pretend-­can-­decl.intr-­3sg ‘He or she could pretend to eat instead.’

Elements of the same type can also be applied recursively (see Fortescue 1980), as shown in (3), where the affixal restructuring verbs -­nnguaq ‘pretend’, -­gasuaq ‘try’ and -­guma ‘want’ all combine with the verb niri ‘eat’. (3)



Nirinnguagasuarumajuq. niri-­nnguaq-­gasuaq-­guma-­ju-­q eat-­pretend-­try-­want-­decl.intr-­3sg ‘He or she wants to try to pretend to eat.’

Furthermore, a closed class of affixal verbs such as -­liuq ‘make’, -­qaq ‘have’ and the copula -­u trigger obligatory noun incorporation of their objects (see Johns 2007; the example in (6) is from Legislative Assembly of Nunavut 2002):3

Richard Compton

560 (4)

(5)



Igluliutuinnarumajunga. iglu-­liuq-­tuinnaq-­guma-­ju-­nga house-­make-­only-­want-­decl.intr-­1sg ‘I just want to make houses/igloos.’ Uqalimaarviralaaqaqtugut. uqalimaarvik-­ralaaq-­qaq-­tu-­gut library-­small-­have-­decl.intr-­1pl ‘We have a small library.’

ivviulauqpuq ivvi-­u-­lauq-­pu-­q you(sg)-­cop-­dist.pst-­ind.intr-­3sg ‘it was you’ (6)

Together these properties yield the possibility of highly complex polysynthetic words that correspond to entire clauses in more analytic languages. Except for one prefix, ta-­, which is confined to the demonstrative system, the language is exclusively suffixing. The language lacks productive compounding.4 No two freely occurring nouns or verbs can combine directly into a single word. While noun incorporation in other languages in which incorporation is optional could be argued to be compounding, in Inuktitut incorporation is categorical in nature; a subset of verbs obligatorily incorporate their objects, while for all other verbs incorporation is impossible.5 This suggests that incorporation involves a syntactic head selecting its complement. Furthermore, unlike what we might expect of compounds, incorporated nominals are referential (as evidenced by such incorporated elements as proper names, pronouns, wh-­words and demonstratives) and can also introduce new discourse referents. For these reasons Sadock (1980) and Johns (2007) argue that incorporation in Inuit languages is a syntactic process. The language has an ergative-­absolutive case system, but there is extensive use of an anti-­ passive construction (whose use and prevalence varies between dialects; see Johns 2001).

2  Evaluative morphology A relatively large number of Inuktitut morphemes meet the criteria for evaluative morphology outlined in this volume – namely that of (i) ‘assigning a value which is different from that of the standard or default (within the semantic scale to which it pertains) to a concept’ and (ii) ‘includ[ing] at least the explicit expression of the standard value [. . .] and an evaluative mark’ (Grandi and Körtvélyessy, this volume). For instance, the suffixes -­ralaaq ‘small’ and -­jjuaq ‘big’ can combine with the noun nanuq ‘polar bear’ to express non-­standard size: (7) a. nanuralaaq nanuq-­ralaaq polar.bear-­small ‘a little polar bear’ b. nanurjjuaq nanuq-­jjuaq polar.bear-­big ‘a big polar bear’

Inuktitut 561 2.1 Position Evaluatives have usually been analysed as derivational morphemes in the Eskimoan literature (Fortescue 1980 for West Greenlandic; de Reuse 1994 for Siberian Yupik) since they typically occur between nominal roots and inflectional morphology such as number, case and possessor marking, as illustrated in (8)–(10). (8)

(9)

(10)



nanuralaat nanuq-­ralaaq-­t polar.bear-­small-­abs.pl ‘little polar bears’ Nanurjjuap takujanga nanuralaaq. nanuq-­jjuaq-­p taku-­ja-­nga nanuq-­ralaaq polar.bear-­big-­erg.sg see-­decl.tr-­3sg.3sg polar.bear-­small(abs.sg) ‘The big polar bear sees the little polar bear.’ ulugjjuara ulu-­jjuaq-­ra ulu-­big-­1sg.poss.abs.sg ‘my big ulu (a traditional woman’s knife)’

However, it is possible for at least some evaluative morphemes to appear between ­inflectional morphemes. For example, while in (11) the morpheme -­kuluk ‘dear, adorable’ ­combines directly with the argument it modifies (in this case a proper name), it is also possible for it to appear inside the verbal complex between the declarative mood marker and agreement, as in (12), without any discernible affect on meaning (that is to say, it is still Alana who is dear – not the event or action of seeing). (11)



Alaanakuluk takuqqaujara. Alaana-­kuluk taku-­qqau-­ja-­ra Alana-­dear(abs.sg) see-­ prox.pst-­decl.tr-­1sg.3sg ‘I saw dear Alana.’

Alaana takuqqaujakuluga. Alaana taku-­qqau-­ja-­kuluk-­ra Alana(abs.sg) see-­ prox.pst-­decl.tr-­dear-­1sg.3sg ‘I saw dear Alana.’

(12)

In addition, a number of evaluative morphemes can be employed adverbially. For example, the morphemes -­jjuaq ‘big’ and -­tsiaq ‘good’ can also modify verbs: (13) a. iglujjuaq iglu-­jjuaq house-­big ‘a big house’ b. Nirijjuaqtuq. niri-­jjuaq-­tu-­q

Richard Compton

562

eat-­big-­decl.intr-­3sg ‘He or she ate a lot.’

(14) a. qimmitsiaq qimmiq-­tsiaq dog-­good ‘a good dog’ b. Sinitsiaqtuq. sinik-­tsiaq-­tu-­q sleep-­good-­decl.intr-­3sg ‘He or she is sleeping well.’ In (13) and (14), where an evaluative modifies the event denoted by a verb, the evaluative morpheme appears between the root and inflection. However, in cases where an evaluative modifies an entire proposition, it can also appear outside inflection and express a ­speaker-­oriented meaning (Compton 2012). For instance, -­jjuaq ‘big’ also possesses a negative connotation when modifying names, as in (15), and this negative meaning emerges when it appears at the edge of a verbal complex outside inflection with the meaning of ‘unfortunately’, as shown in (16). (15)



Mialijjuaq Miali-­jjuaq Mary-­big ‘(that) big/pestering Mary’

Tikittujjuaq. tikit-­tu-­q-­jjuaq arrive-­ decl.intr-­3sg-­big ‘Unfortunately, he or she arrived.’ (16)

A further way in which evaluatives can appear outside inflection is in modifying demonstrative pronouns, which are portmanteaus for case and number: unakuluk una-­kuluk this.one.abs.sg-­adorable ‘this lovely one’ (17)

In sum, while the more common linear ordering of evaluatives between roots and inflectional morphology conforms to an analysis in which evaluatives are derivational in nature, the possibility of placing certain evaluatives between inflectional morphemes, as well as outside inflectional suffixes (at the right edge of words), is unexpected in derivational morphology. I return to the topic of the status of these morphemes in Section 2.5. 2.2  Stacking and variable order Evaluative morphemes can be stacked together, as illustrated in the form in (18), in which four evaluatives modify the noun meaning ‘computer’:

Inuktitut 563 (18)



qarisaujaralaakulutuqannguaq qarisaujaq-­ralaaq-­kuluk-­tuqaq-­nnguaq computer-­small-­adorable-­old-­pretend ‘an old adorable small pretend computer’ (such as a toy computer)

Evaluatives often also exhibit variable ordering with respect to each other. In (19) the morphemes -­ralaaq ‘small’ and -­nnguaq ‘pretend’ may appear in either order, while in (20) -­tsiavaq ‘good’ and -­kuluk ‘dear, adorable’ can appear in either order. (19) a. igluralaannguaq iglu-­ralaaq-­nnguaq house-­small-­pretend b. iglunnguaralaaq iglu-­nnguaq-­ralaaq house-­pretend-­small ‘a small pretend house’ (20) a. iglutsiavakuluk iglu-­tsiavaq-­kuluk house-­good-­adorable b. iglukulutsiavaq iglu-­kuluk-­tsiavaq house-­adorable-­good ‘a good adorable house’ 2.3  Evaluatives in noun incorporation While incorporated nominals typically lack inflection for number, case and possession, they can be modified by evaluative morphemes. In (21) the evaluative -­jjuaq ‘big’ appears between the noun iglu ‘house’ and the incorporating verb -­liuq ‘make’, while in (22) the evaluatives -­tsiavaq ‘good’ and -­nnguaq ‘pretend’ appear between iglu and -­qaq ‘have’. (21)





(22)



Iglujjualiurumajunga. iglu-­jjuaq-­liuq-­guma-­ju-­nga house-­big-­make-­want-­decl.intr-­1sg ‘I want to build a big house.’ Iglutsiavannguaqaqtuq. iglu-­tsiavaq-­nnguaq-­qaq-­tu-­q house-­good-­pretend-­have-­decl.intr-­3sg ‘He or she has a nice pretend/toy house.’

2.4 Productivity Modification by evaluatives is highly productive, with ungrammatical combinations being due to semantic/pragmatic considerations. For instance, while it was possible to use -­kuluk to modify pusi ‘cat’ in (23), the same construction with natsiq ‘harbour seal’ in (24) was

Richard Compton

564

considered awkward because seal meat forms a part of the traditional diet and thus seals are typically not described as cute or adorable. (23)



pusikuluk pusi-­kuluk cat-­adorable ‘an adorable cat’

(24) #natsikuluk natsiq-­kuluk harbour.seal-­adorable ‘an adorable harbour seal’ Even proper names, which we might expect to be more resistant to modification, are compatible with a variety of evaluative morphemes:6 (25)

(26)

(27)



Juusipiruluk anijuq. Juusipi-­ruluk ani-­ju-­q J.-­darn/poor(abs.sg) go.out-­ decl.intr-­3sg ‘That darn Juusipi went out.’ Juusipijjuaq nirijuq. Juusipi-­jjuaq niri-­ju-­q J.-­big(abs.sg) eat-­decl.intr-­3sg ‘Big Juusipi is eating.’ Juusipiralaaq nirijuq. Juusipi-­ralaaq niri-­ju-­q J.-­small(abs.sg) eat-­ decl.intr-­3sg ‘Little Juusipi is eating.’

Furthermore, neologisms such as uqalimaarvik ‘library’ and qarisaujaq ‘computer’ are compatible with evaluatives, as illustrated in (5) and (18) above, as are borrowed words such as tiivii ‘television/TV’: (28)





Tiiviikuluktaaqqaujunga. tiivii-­kuluk-­taaq-­qqau-­ju-­nga television-­adorable-­get-­prox.pst-­decl.intr-­1sg ‘I bought a nice little television.’

Such examples provide additional evidence of the near full productivity of evaluative morphemes in Inuktitut. 2.5  Categorial status While non-­inflectional morphemes appearing inside complex words, including evaluatives, have traditionally been categorised as derivational morphology in the literature on Eskimoan, recent work has observed that unlike derivational morphology in more analytic

Inuktitut 565 languages, these morphemes exhibit a number of properties typically ascribed to syntax. For instance, de Reuse (2009) argues that many such morphemes in the related Siberian Yupik language are fully productive, can be applied recursively, exhibit variable ordering within words and can interact with the syntax outside the word. These facts lead de Reuse to propose a third type of morphology, distinct from derivation and inflection, which he calls ‘productive non-­inflectional concatenation’. Fundamentally, his argument is that this type of morphology in Yupik behaves, in many important respects, like syntax. Observing the same degree of parallelism between morphology and syntax in Inuktitut, Compton and Pittman (2010) employ the framework of Distributed Morphology (Halle and Marantz 1993; Harley and Noyer 1999) to propose that Inuit phonological words are actually syntactically phrasal; nominal words are actually full DPs and verbal complexes are actually full CPs (stripped of their DP arguments). Following in this vein, Compton (2012) argues that the evaluative morphemes examined herein are in fact adjectives and adverbs. Taking yet another approach within Distributed Morphology, Cook and Johns (2009) argue that evaluatives and other morphemes appearing between the root and inflection are functional heads in syntactic structure, following Cinque’s (1999) cartographic proposals regarding adverbial functional heads. 2.6  Meaning range of evaluatives Evaluatives in Inuktitut encompass a wide range of meanings. Using the terminology employed by Grandi and Körtvélyessy (this volume), I have divided a number of these morphemes into the semantic fields shown in (29)–(35), although this list is not exhaustive and does not include adverbial readings. (The example in (31c) is from Spalding 1998, Aivilik dialect.) (29) Augmentation: -­jjuaq, -­kallak and -­kutaaq: a. natsirjjuaq natsiq-­jjuaq harbour.seal-­big ‘a big harbour seal’ b. qimmikallak qimmiq-­kallak dog-­big/fat ‘a big fat dog’ c. saakutaara saa-­kutaaq-­ga table-­long-­1sg.poss ‘my long table’ (30) Diminution: -­ralaaq: kuuralaaq kuuk-­ralaaq river-­small ‘a small river; a stream’ (31) Age variation: -­tuqaq, -­viniq and -­taasaaq: a. aanniaviktuqaq aanniavik-­tuqaq

566 hospital-­old ‘an old hospital’ b. igluviniq iglu-­viniq house-­old/former ‘an old/former house’ c. iglutaasaaq iglu-­taasaaq house-­new ‘a new house’ (32) Intensification: -­vijjuaq: qimmituqavijjuaq qimmiq-­tuqaq-­vijjuaq dog-­old-­really ‘a really old dog’ (33) Endearment: -­kuluk, -­ngai, -­tsiaq and -­tsiavaq: a. qimmikuluk qimmiq-­kuluk dog-­adorable ‘a precious little puppy’ b. Sailanngai Saila-­nngai Saila-­ dim ‘Saila’ (diminutive/hypocoristic form of proper name) c. qimmitsiaq qimmiq-­tsiaq dog-­good ‘a good dog’ d. iglutsiavaq iglu-­tsiavaq house-­good ‘a well-­built house’ (34) Authenticity: -­nnguaq: nanunnguaq nanuq-­nnguaq polar.bear-­pretend ‘a pretend polar bear; a polar bear carving’ (35) Contempt: -­ruluk and -­aluk: a. qiturngannguarulu(ng)a qiturngannguaq-­ruluk-­(ng)a doll-­darned/poor-­3sg.poss.abs.sg ‘his or her poor little doll’ b. innaaluk inna-­aluk

Richard Compton

Inuktitut 567 that.one.abs.sg-­darned ‘that darned person/one’ Some of these morphemes exhibit meanings corresponding to additional semantic functions. For instance, in addition to its positive descriptive meaning, -­jjuaq also possesses a negative qualitative meaning.

3 Conclusion While this chapter has examined the large inventory of evaluative morphemes in Inuktitut, a similar range of evaluatives is found in other dialects within the Inuit dialect continuum, such as Greenlandic dialects, as well as in languages of the Yupik branch of Eskimo-­Aleut. Comparing Inuktitut to West Greenlandic, for instance, an analogous range of evaluatives exists, although individual morphemes often express different meanings. For example, the morpheme -­nnguaq, which means ‘pretend, imitation, fake’ in Inuktitut, means ‘small, dear’ in West Greenlandic (Fortescue 1980). Conversely, dialects along the continuum also exhibit synonymous evaluatives with distinct phonological forms. Although evaluatives in Eskimoan languages were originally classified as derivational, evidence regarding the syntactic nature of polysynthetic words in these languages suggests alternative treatments, such as analysing evaluatives as adjectives and adverbs (Compton 2012), functional heads (Cook and Johns 2009) or a third type of morphology alongside inflection and derivation (de Reuse 2009; see also Johns 2014).

Notes 1. According to Dorais (2003), much of the variation between dialects is of a phonological and lexical nature. However, there are also a number of morphosyntactic differences in areas such as mood, case and tense marking. 2. Certain suffixes trigger deletion, lenition or coalescence at morpheme boundaries. This is a general phenomenon and not specific to evaluatives. I use the standard Romanised orthography for Inuktitut, which corresponds to a broad IPA transcription except that = [ŋ],  = [ŋŋ], = [ddʒ], = ʁ and = ɣ. 3. Inuktitut distinguishes between distant past (before today) and recent past (earlier today), for which I have used the term ‘proximate past’ (see Hayashi 2011). 4. Sadock (2003) lists several compounds in West Greenlandic, but all appear to be loan translations such as siniffik pooq ‘sleeping bag’. 5. Sadock (1980, 11) notes one exceptional verb in West Greenlandic which appears to be optionally enclitic. 6. As observed by Hayashi and Spreng (2005), achievement verbs without tense marking are interpreted as recent past. Consequently, in the absence of past tense marking, an achievement verb like ani ‘go out’ is typically interpreted as recent past, while an activity verb like niri ‘eat’ is interpreted as progressive. Hayashi and Spreng also demonstrate that unlike West Greenlandic, which has been argued to be tenseless, Inuktitut obligatorily marks tense.

19.7  Plains Cree (Algonquian)

Arok Wolvengrey

1 Introduction Plains Cree is one member of the extensive language continuum known as Cree-­ Montagnais-­Naskapi which stretches from the eastern coasts of Quebec and Labrador westwards to the Rockies. Within this large range, the Plains Cree or ‘y’ dialect is the westernmost, located in south-­central Saskatchewan and central and northern Alberta, with outlying communities in the regions of Moberly Lake (British Columbia), Fort Smith (Northwest Territories) and Rocky Boy (Montana). The territory covered by Plains Cree is large enough to allow for the differentiation of subdialects, the largest division being between southern Plains Cree (nēhiyawēwin, which retains /e:/ and /i:/ as distinct phonemes) and northern Plains Cree (nīhiyawīwin, which, like Woods Cree, has merged /e:/ and /i:/ to /i:/) (cf. Algonquian Linguistic Atlas, n.d.; Wolvengrey 2011, 4–9). Though much of the data discussed in this chapter will apply to the entire Cree language continuum, the specific forms cited come primarily from southern Plains Cree as spoken in Saskatchewan. The terms ‘Cree’ and ‘Plains Cree’ will be used fairly interchangeably throughout, unless a specific dialect is cited for contrast or added content. Cree and its sister languages in the Algonquian family are well known for their polysynthetic nature and thus very rich morphology, particularly in the structure of the verb. However, even extensive verbal templates such as those of Wolfart (1973) and Bakker (2006) do not generally indicate much in the way of evaluative categories, largely due to the usual restriction of these templates to inflection. Although evaluation can be syntactically indicated, this is also limited by the absence of a distinct class of adjectives in Cree. Evaluation does, however, manifest itself in the form of both root and affixal morphemes across the three main lexical word classes: noun, verb and particle.

2  Evaluative categories A number of evaluative categories can be found within Cree morphology, but space permits only a thorough discussion of two main categories. In Section 2.1, we will find that the most prevalent evaluative category is the diminutive, which is clearly delineated through all lexical word classes. Particularly among nouns, the diminutive overlaps to a certain extent with both ameliorative and pejorative meaning, neither of which is found morphologically differentiated from the diminutive in Cree. In contrast to diminutives, augmentatives are marked quite differently and are more diverse in their patterning. Section 2.2 will highlight augmentatives and the overlapping category of the frequentative.

Plains Cree 569 2.1 Diminutives The most prevalent evaluative category in Plains Cree morphology is the diminutive, which is marked through all lexical classes. 2.1.1 Nominal Nouns in Cree, and Algonquian languages in general, are divided into two grammatical genders, animate and inanimate, and this noun classification system is pervasive within Cree grammar. It is interesting to note, therefore, that the basic diminutive suffix, -­is (in Plains, Woods and Western Swampy Cree, but -­iš [ɪʃ] in eastern dialects), can be added to virtually all nouns regardless of gender, as with the regular nouns given in (1): (1)

a. Animate noun (na): b. Inanimate noun (ni):

asikan ‘sock’ > asikanis maskisin ‘shoe’ > maskisinis

‘little sock’ ‘little shoe’

Morphophonological rules operating at the derivational boundary result in several apparent allomorphs of -­is in Plains Cree: -­os, -­s and -­V:s. A doubled form of the suffix, -­isis (-­išiš), can also be used. While this is perhaps more prevalent with animate nouns indicating small creatures (2a), it can also appear on inanimate nouns (2b), and relative degrees of smallness can be indicated by using both -­is and -­isis (2c):1 (2) a. mahihkanisis mahihkan-­isis wolf-­ dim ‘wolf cub’ b. ayiwinisis ayiwinis-­is article.of.clothing-­ dim ‘small article of clothing’ c. sākahikan > sākahikanis > sākahikanisis sākahikan sākahikan-­is sākahikan-­is-­is lake lake-­ dim lake-­ dim-­dim ‘lake’ ‘small lake’ ‘pond’ One additional indicator of the diminutive is the sound symbolic shift of /t/ to /c/ ([ts] in Plains Cree, [tʃ] in most other Cree dialects). This occurs quite consistently, with any /t/ in the noun stem being diminutivised, and can even affect phones in prefixes and proclitics (3): (3)

nitastotin > nicascocinis nit=astotin nit=astotin-­is 1=hat 1=hat-­ dim ‘my hat’ ‘my cap, my small hat’

In the eastern dialects which have an additional phonemic distinction between /s/ and /ʃ/, a second sound symbolic shift from /s/ to /ʃ/ can also occur (Pentland 1975, 242). The meaning of diminutive nouns is usually a straightforward indication of smallness in comparison to the norm. However, the diminutive can also indicate endearment (or amelioration), as in certain kinship terms (4a–b) and nicknames (4c–d):

Arok Wolvengrey

570

(4) a. nōhtāwiy > nōhcāwīs n-­ohtāwiy n-­ohtāwiy-­is 1-­father 1-­father-­dim ‘my father’ ‘my uncle (father’s brother)’ b. nītim > nīcimos n-­ītimw n-­ītimw-­is 1-­cross.cousin.of.opposite.gender 1-­cross.cousin.of.opposite.gender-­dim ‘my cross-­cousin’ ‘my sweetheart’ c. mahihkan > mahihkanis mahihkan mahihkan-­is dim wolf wolf-­ ‘wolf’ ‘little wolf; dear Wolf’ d. /kitimāk-­/ > kitimāk-­ vr ‘poor; pitiful’

kicimākis kitimāk-­is vr-­dim ‘poor thing; cutie’

Though it is not as common, diminutives can be used pejoratively, as with the diminutive form in (5): (5) mōniyāw > mōniyās mōniyāw mōniyāw-­is whiteman whiteman-­ dim ‘whiteman’ ‘(darned) whiteman’ Cree does not have a pejorative form distinct from the diminutive, though some Algonquian languages such as Ojibwe do. This is illustrated here through the Saulteaux dialect, which is also spoken in Saskatchewan and is most closely associated with Plains Cree. Saulteaux diminutives (-­ēns and variants) are given in (6a–b), while pejoratives (-­(w)ihš), indicating anger or disgust, are given in (7a–b) (Cote 1984, 21): (6) a. cīmān > cīmānēns cīmān cīmān-­ēns boat boat-­ dim ‘boat’ ‘little boat’ b. inini > ininīns inini inini-­(ē)ns man man-­ dim ‘man’ ‘little man’ (7) a. cīmān > cīmānihš cīmān cīmān-­ihš boat boat-­ pej ‘boat’ ‘darned boat’

Plains Cree 571 b. inini > ininiwihš inini inini-­(w)ihš man man-­ pej ‘man’ ‘darned man’ Some speakers of the Odawa dialect of Ojibwe actually make a three-­way distinction between the diminutive (-­ens), contemptive (-­enh) and pejorative (-­ish) (cf. Valentine 2001, 185–93). 2.1.2 Verbal Diminutive patterns are also pervasive in Plains Cree verb paradigms, where the nominal -­is is paralleled by the verbal diminutive -­si. This suffix productively derives diminutive verb stems which indicate a reduced or smaller amount of activity: (8) a. nipā > nipāsi nipā nipā-­si dim sleep sleep-­ ‘sleep!’ ‘sleep a bit, nap!’ b. atoskē > acoskēsi atoskē atoskē-­si work work-­ dim ‘work!’ ‘work a bit!’ Interestingly, diminutive verbs in Innu (Montagnais) are used to indicate a smaller actor/ subject, while either actor or action can be diminutivised in East Cree (Marguerite MacKenzie, p.c.). In contrast, diminutive verb forms of this type do not occur at all in Saulteaux (Ojibwe), where the parallel construction (built on -­hsi or -­hsī) constitutes a negative mode. Cree and Algonquian verbs are generally divided into four main verb classes on the basis of the twin criteria of transitivity and the animacy of the participants. This yields four basic stem classes: (i) vii or inanimate intransitive verbs; (ii) vai or animate intransitive verbs; (iii) vti or transitive inanimate verbs; and (iv) vta or transitive animate verbs. Diminutives of vai stems (as in (8) above) are most commonly attested, but all four stem classes can be so marked. The paradigms in Table 19.7.1 illustrate the basic and diminutive vai paradigms in the conjunct order. Table 19.7.1. Basic and diminutive vai conjunct paradigms in Cree Basic vai conjunct 1sg 2sg 1pl.excl 1pl.incl 2pl 3sg 3pl 3obv

Diminutive vai conjunct

comp

Stem

Person

ē-­ ē-­ ē-­ ē-­ ē-­ ē-­ ē-­ ē-­

nipā nipā nipā nipā nipā nipā nipā nipā

-­yān -­yan -­yāhk -­yahk -­yēk -­t -­cik -­yit

1sg 2sg 1pl.excl 1pl.incl 2pl 3sg 3pl 3obv

comp

Stem

dim

Person

ē-­ ē-­ ē-­ ē-­ ē-­ ē-­ ē-­ ē-­

nipā nipā nipā nipā nipā nipā nipā nipā

-­si -­si -­si -­si -­si -­si -­si -­si

-­yān -­yan -­yāhk -­yahk -­yēk -­t -­cik -­yit

572

Arok Wolvengrey

Although the diminutive morpheme is highlighted in a separate column in Table 19.7.1, all evidence suggests that -­si is a vai derivational suffix, albeit a very productive one. The same can be said for the more limited vii stem class, where the diminutive is more likely to appear as -­sin (9a) or, alternatively, show further complications due to stem shape, as in (9b), where it could be seen synchronically as an infix -­si-­ preceding a stem-­final /n/. (9) a. kinwāw > kinwāsin kinw-­ā-­w kinw-­ā-­sin-­Ø be.long-­ viif-­3sg.inan be.long-­ viif-­dim-­3sg.inan ‘it is long’ ‘it is somewhat long’ b. kimiwan > kimiwasin kimiw-­an-­Ø kimiw-­an-­Ø rain-­ viif-­3sg.inan rain-­ viif-­3sg.inan ‘it is raining’ ‘it is drizzling’ Transitive verbs add yet further complications. Transitive inanimate (vti) stems are augmented by a thematic suffix which displays considerable allomorphy, dependent on its stem and/or its position within the paradigms. Furthermore, vti stems can appear in a number of sub-­classes, of which only the first (vti1) still partially distinguishes this verb class from vai stems in Plains Cree. The examples in (10) illustrate the variation of theme markers present in the vti1 independent order (10a), conjunct order (10b) and imperative order (10c) respectively, but also the neutralisation to the -­a theme in all diminutive forms, suggesting the derivation of an invariant diminutive stem (e.g. yōhcēnasi-­): (10) a. niyōhtēnēn > niyōhcēnasin ni=yōhtēn+ē-­n ni=yōhtēn+a-­si-­n 1=open+thm-­sap 1=open+thm-­dim-­sap ‘I open it’ ‘I open it a bit’ b. ē-­yōhtēnamān > ē-­yōhcēnasiyān ē-­yōhtēn+am-­ān ē-­yōhtēn+a-­si-­yān comp-­open+thm-­1sg comp-­open+thm-­dim-­1sg ‘(as) I open it’ ‘(as) I open it a bit’ c. yōhtēna > yōhcēnasi yōhtēn+a yōhtēn+a-­si open+thm.imp open+thm-­dim.imp ‘open it!’ ‘open it a bit!’ Finally, the place of the diminutive within the most complex vta paradigms must be mentioned. Only here does -­si appear to occur within the inflection, following the vta theme suffixes (e.g. direct (drth) in (11)) which, unlike in vti stems, have not fused with the stem. (11) a. nisīkinamawāw > nisīkinamawāsiw ni=sīkinamaw-­ā-­w ni=sīkinamaw-­ā-­si-­w 1=pour.for-­ drth-­3sg 1=pour.for-­ drth-­dim-­3sg ‘I pour it for him or her’ ‘I pour a bit of it for him or her’

Plains Cree 573 b. sīkinamawin > sīkinamawisin sīkinamaw-­i-­n sīkinamaw-­i-­si-­n pour.for-­ drth-­sap.imp pour.for-­ drth-­dim-­sap.imp ‘pour it for me!’ ‘pour a bit of it for me!’ The full extent of the interaction between the infamous Algonquian direct–inverse system and the diminutive in Plains Cree remains to be explored. 2.1.3 Root/Particle Free, indeclinable particles serve many functions in Plains Cree, and many can be diminutivised. Furthermore, the sound symbolic shift of /t/ to /c/ can occasionally be used even without a diminutive suffix, as when the common greeting in (12) is directed to a child: (12)

tānisi > cānisi tān-­isi tān-­isi what-­manner what-­manner ‘how; hello’ ‘hello’ [said to a child]

More commonly, a suffix is used, though more variation appears among particles. Quantifiers generally take the basic nominal -­is (or -­īs) as in (13). Quantifiers specifying small amounts sometimes occur without a non-­diminutive counterpart (13c–d): (13) a. mihcēt > mihcēcis mihcēt mihcēt-­is many many-­ dim ‘many’ ‘fairly many’ b. mistahi > miscahīs mistahi mistahi-­is much much-­ dim ‘much; a lot’ ‘quite a bit’

c. -­-­

cikawāsis *cikawa(w)-­isis ??-­ dim ‘few’



d. (/apis(t)-­/ >) ‘small’

apisīs apisi-­is small-­ dim ‘a little, a little bit’

Other particles, such as temporals (14), locatives (15) and manner adverbials (16), are commonly formed with -­ēs, -­īs or -­ēsīs, while even less regular derivations can occur (17): (14)

kayās > kayāsēs kayās kayās-­ēs long.ago long.ago-­ dim ‘long ago’ ‘quite some time ago’

Arok Wolvengrey

574 (15)

wāhyaw > wāhyawēs wāhyaw wāhyaw-­ēs far.away far.away-­ dim ‘far away’ ‘quite a way’

(16) sōhki > sōhkēsīs sōhki sōhk-­ēsīs hard hard-­ dim ‘hard; fast’ ‘a little faster’ (17)

awasitē > awasicahēs awas-­itē awas-­it-­ahēs away-­there away-­there-­dim ‘further on’ ‘a little further’

Occasionally, as with nouns, three degrees can be distinguished, as in (18): (18)

ayiwāk > ayiwākēs > ayiwāk ayiwāk-­ēs more more-­ dim ‘more’ ‘a little more’

ayiwākēsīs ayiwāk-­ēs-­īs more-­ dim-­dim ‘a tiny bit more’

In addition to diminutive meaning, at least one such particle appears to function primarily as a pejorative: (19) māmāsīs (*māmās?)-­īs (*poor)-­dim ‘poorly done; any old way’ As stated earlier, Cree has no distinct class of adjectives. However, attributive roots commonly occur as prenominal particles. Nouns diminutivised in this way usually also include a diminutive suffix (20): (20) /apis(t)-­/ ~ apisci-­~ apisi-­ ‘small’ a. niska > apisci-­niskis nisk-­a apisci-­nisk-­is goose-­ anim.sg small-­goose-­dim ‘goose’ ‘laughing goose; small goose’ b. pāskisikan > apisci-­pāskisikanis pāskisikan apisci-­pāskisikan-­is gun small-­gun-­dim ‘gun’ ‘.22 rifle, small gun’ Verbs may also be marked in this way, though more common are verbs built with the bound root (Algonquian ‘initials’) as the sole lexical element (21) or joined with at least one other bound root (Algonquian ‘medials’) (22), both stem types being completed by ‘finals’ (e.g. viif or vaif).

Plains Cree 575 (21) apisāsin apis-­ā-­sin-­Ø small-­ viif-­dim-­3sg.inan ‘it is small’ (22)

apiscicāpiw apist-­icāp-­i-­w small-­eye-­vaif-­3sg ‘she or he has small eyes’

2.2  Augmentatives and frequentatives By contrast with the pervasive diminutive constructions, augmentatives are not as common or as consistently rendered in Cree. Though various roots can form the bases for both nominal and verbal constructions, it is the verbal frequentative that most closely resembles the main diminutive constructions surveyed above. 2.2.1 Nominal Augmentative or frequentative morphology on nouns comes in three main forms: prefixed particles (see Section 2.2.3 below), attributive root morphemes compounded with nominal roots, and deverbal forms of the verbal frequentative -­ski. Three main roots can be used to indicate augmentation: /mis(t)-­/ ‘big’, /kiht-­/ ‘great, large’ and /mahk-­/ ‘big’. The last of these is largely restricted to use with body parts and, when joined with an appropriate nominal root, can stand alone to refer to a larger than normal body part or a person or creature characterised by such a body part (23a), or be further derived to verbal usage (23b) (cf. Okimāsis 2004, 14): (23) a. mahkistikwān mahk-­[i]stikwān big-­head ‘large head’/‘person with a large head’ b. mahkistikwānēw mahk-­[i]stikwān-­ē-­w big-­head-­vaif-­3sg ‘she or he has a large head’ In contrast, many nouns end in what appears to be a nominalising suffix -­sk, indicating the frequentative or habitual (24): (24)

kimotisk ‘thief’ (cf. vai kimoti-­‘steal’) mātosk ‘crybaby’ (cf. vai māto-­‘cry’) nōtinikēsk ‘fighter, boxer’ (cf. vai nōtinikē-­‘fight’)

Synchronically these appear to be deverbal constructions clipped from verbal frequentatives, discussed immediately below.

576

Arok Wolvengrey

2.2.2 Verbal The verbal frequentative is formed by the suffix -­ski, paralleling diminutive -­si, except that the frequentative is almost exclusively limited to occurrence with vai stems. The examples in (25) parallel some of the nominal examples in (24) above, and illustrate the derivation of the corresponding nouns: (25) a. mātow > mātoskiw > mātosk māto-­w māto-­ski-­w māto-­ski cry-­ 3sg cry-­ hab-­3sg cry-­ hab ‘she or he cries’ ‘she or he cries all the time’ ‘crybaby’ b. nōtinikēw > nōtinikēskiw > nōtinikēsk nōtin-­ikē-­w nōtinikē-­ski-­w nōtinikē-­ski fight-­gen.obj-­3sg fight-­hab-­3sg fight-­hab ‘she or he fights’ ‘she or he fights all the time’ ‘fighter; boxer’ Though less common, the frequentative suffix can be accompanied by reduplication, as in (26): (26) kiyāskiw > ka-­kiyāskiskiw > kakiyāskisk kiyāski-­w Ca~kiyāski-­ski-­w kakiyāskiski lie-­ 3sg rdp~lie-­hab-­3sg continually.lie.habitually ‘she or he lies’ ‘she or he continually lies’ ‘liar’ Reduplication in Plains Cree primarily functions to mark verbal aspect in two main patterns. ‘Light reduplication’ (marked by copying the initial consonant plus /a/; i.e. Ca-­) indicates an ongoing action, while ‘heavy reduplication’ (marked by copying the initial consonant plus /āh/; i.e. Cāh-­) indicates repetitive or distributive action (Ahenakew and Wolfart 1983). Thus, heavy reduplication in particular overlaps in meaning with the frequentative: (27) kiyāskiw > kāh-­kiyāskiw kiyāski-­w Cāh~kiyāski-­w lie-­ 3sg rdp~lie-­3sg ‘she or he lies’ ‘she or he lies again and again’ 2.2.3 Root/Particle Both reduplication and root morphemes are the primary means of marking ­augmentation among Cree particles. The examples in (28) illustrate the reduplication of particles themselves, with the resultant meaning of augmentation and/or repetition or distribution: (28) a. kinwēs > kāh-­kinwēs kinwēs Cāh-­kinwēs long.time rdp-­long.time ‘for a long time’ ‘for a very long time’

Plains Cree 577 b. mēskoc > māh-­mēskoc mēskoc Cāh-­mēskoc instead rdp-­instead ‘instead, in exchange’ ‘each in turn’ c. pītos > pāh-­pītos pītos Cāh-­pītos different rdp-­different ‘different’ ‘very different; each different’ Paralleling the nominal structures already surveyed, root morphemes like /mis(t)-­/ (29) and /kiht-­/ (30) are frequently turned into (a) prenominal, (b) preverbal or even (c) pre-­ particle augmentatives: (29) a. sikāk > misi-­sikāk sikākw misi-­sikākw skunk big-­skunk ‘skunk’ ‘Big Skunk’ [legendary figure] b. kitow > kito-­w call.out-­3sg ‘she or he calls out’

misi-­kitow misi-­kito-­w big-­call.out-­3sg ‘she or he calls out loudly; it thunders’

c. ayiwāk > misi-­ayiwāk ayiwāk misi-­ayiwāk more big-­more ‘more’ ‘much more’ (30) a. okimāw > kihci-­okimāw okimāw kihci-­okimāw chief great-­chief ‘chief’ ‘king’ b. itwēw > kihci-­itwēw itwē-­w kihci-­itwē-­w say.so-­ 3sg great-­say.so-­3sg ‘she or he says so’ ‘she or he swears an oath’ c. mitātahtomitanaw > kihci-­mitātahtomitanaw mitātahtomitanaw kihci-­mitātahtomitanaw one.hundred great-­one.hundred ‘100’ ‘1,000’

3 Conclusion The main evaluative categories in Plains Cree are marked in four distinct ways. A suffix and consonant symbolism function together to mark the diminutive, and this spills over to ameliorative and/or pejorative meaning as well. In contrast, more lexical or reduplicative means are typically used for augmentation. However, the frequentative matches the semantics of augmentation with the suffixal strategy of diminutives.

578

Arok Wolvengrey

Note 1. Examples are glossed in four lines in order to provide the Cree data in both its standard Roman orthography (line 1) and a morphemic analysis (line 2), along with morpheme glosses (line 3) and an idiomatic English translation (line 4). Though this may occasionally lead to redundancy in the glossing, it is done for the benefit of Plains Cree readers as we promote literacy in the language.

19.8  Slavey (Dene) and other Athabaskan Languages

Olga Lovick and Keren Rice

1 Introduction Athabaskan languages are spoken in three geographical regions of North America. Best known are the Navajo and Apache varieties, spoken in the south-­west of the United States. There are also Athabaskan languages in Alaska and in parts of Canada including the Northwest Territories, the Yukon Territory, Manitoba, Saskatchewan, Alberta and British Columbia. In addition, a group of Athabaskan languages is/was spoken on the Pacific Coast, in parts of northern California and Oregon. See, for example, Hargus (2011) for an annotated bibliography. Athabaskan languages are characterised by rich verbal morphology. The languages are largely prefixing, although the morphology under discussion in this chapter is suffixing. We begin with a discussion of evaluative morphology in Slavey (often called Dene; Northwest Territories, Yukon Territory, Alberta, British Columbia) and then compare it with several others in the family.

2  Slavey (Dene) evaluative morphology What we call Slavey is composed of a number of different varieties which differ in terms of phonology, morphology, syntax and lexicon. The words included in this section are drawn from different varieties. Sample augmentative and diminutive forms are given in (1). Most Slavey data is from Rice (1989), with some from Moore and Wheelock (1990) and the Dene K’ę́ę́ T’ahsíi Ts’uuzi Gha Edįtł’éh (Nadaatł’éh)/South Slavey Topical Dictionary (1993).1 (1) a. Augmentative: cho/sho/ro/ho/o: ’elá-­sho boat-­ aug ‘barge’ ’ek’élu-­cho road-­ aug ‘highway’ tthe-­cho rock-­ aug ‘boulder’

580

Olga Lovick and Keren Rice

b. Diminutive: yaa/ya/zhaa/zha/aa/ah/a: sah-­yaa bear-­ dim ‘bear cub’ ’elá-­ya boat-­ dim ‘canoe’ The examples in (1) illustrate the evaluative morphemes attached to nouns. They can also attach to verbs:2 (2) a. Augmentative: nechá-­cho it.is.big-­ aug ‘it is really big’ łek’á-­cho it.is.fat-­ aug ‘it is really fat’ nedé-­o it.is.long-­ aug ‘it is really long’ b. Diminutive: łekǫ-­a it.is.sweet-­ dim ‘it is a little sweet’ aetsili-­a it.is.small-­ dim ‘it is really small’ They also attach to words which are difficult to classify as to category: (3) Augmentative: máhsi-­cho thank.you-­ aug ‘thank you very much’ dádí-­o no-­ aug ‘no way!’ In the following, we discuss these items’ properties. 2.1 Semantics The terms ‘augmentative’ and ‘diminutive’ cover a range of meanings, typical of such ­elements cross-­linguistically. The augmentative is usually translated as ‘big’ and the diminutive as either ‘little’ or ‘young’:

Slavey (Dene) and other Athabaskan Languages 581 (4) a. Augmentative: Size: kǫ́-­cho fire-­aug ‘big fire’ elíi-­ro sickness-­ aug ‘big sickness’ det’oni-­cho aug bird-­ ‘large birds’ b. Diminutive: Size: e-­tthí-­tth’en-é-­a indf.poss-­head-­bone-­poss-­dim ‘small skull’ chǫ-­a bird-­ dim ‘little bird’ łígí-­a little-­ dim ‘little one’ Age: sah-­ya bear-­ dim ‘bear cub’ tsá-­ya beaver-­ dim ‘young beaver’ ’įts’é-­ya moose-­ dim ‘moose calf’ The evaluative item might not contribute much, simply reinforcing size or age: (5) Diminutive: ts’ído-­a child-­ dim ‘child, children’ However, while the evaluative morphemes can modify the stem, indicating age and size, they frequently contribute a more substantial change in meaning. The evaluative element may identify a particular species or class that is contained within the class named by the noun that they are attached to; Rice (2012, 56) refers to these as as part for whole metonymy. Examples are given in (6)–(9):3 (6) Augmentative: det’one-­cho bird-­ aug ‘bald eagle’

582

Olga Lovick and Keren Rice

bedzi-­cho caribou-­ aug ‘bull caribou’ deh-­cho river-­ aug ‘Mackenzie River’ dene-­cho person-­ aug ‘giant’ (7) jíye-­cho aug berry-­ ‘orange’ ’ǫ́k’a-­cho whiskey.jack-­ aug ‘kingfisher’ The examples in (6) and (7) involve augmentation. However, the evaluative form does not simply identify something that is larger than some standard, but rather the ­prototypical largest of the category identified by the noun (6). The prototypically largest can be the male of the species (e.g. bull caribou);4 it can also be a different species that is similar to but larger than the one named by the noun on its own (7). The augmentative is also used to indicate a long span of time, as in (8). (8) xa-­o gháré ’įhbé-­o gháré winter-­ aug through summer-­aug through ‘all winter and summer’ Parallel to the augmentative, the diminutive can identify a species that is smaller than the one named by the noun. (9) Diminutive: nǫ́da-­ya lynx-­ dim ‘domestic cat’ ’elá-­ya boat-­ dim ‘canoe’ łu-­a spoon-­ dim ‘teaspoon’ In the words in (6)–(9), there is lexicalisation, with a new meaning created that is not predictable from the sum of the parts. The diminutive can indicate endearment and is frequently used with kinship terms, especially if the possessor is first person:

Slavey (Dene) and other Athabaskan Languages 583 (10) ts’ezhoni-­a old.woman-­ dim ‘little old lady’ shį-­a song-­ dim ‘little ditty’ bebí-­ya baby-­ dim ‘little baby, pitiful little baby’ (11) -­díá-­a dim younger.sister-­ ‘dear little sister’ -­tú-­a daughter-­ dim ‘little daughter’ The augmentative is common in place names.5 (12) náįlį-­cho water.falls-­ aug ‘Alexander Falls’ tł’á-­o-­cho bay-­ areal-­aug ‘Big Bay’ mįe-­go-­cho lake-­ areal-­aug ‘Round Lake’ Some nouns always occur with evaluative morphology, with lexicalised meaning.6 (13) dega-­ho/dagao white-­ aug ‘swan’ tłį-­cho dog-­ aug ‘horse’ A second diminutive, tséle, indicates that the modified noun is small in size compared to some standard. It does not appear to show the degree of lexicalisation that the other diminutive does. (14) deh-­tséle river-­ dim ‘creek’ The forms with tséle in (15) perhaps show lexicalisation. We do not know whether the constellation in question was identified as such prior to contact (15a), or whether this term is based on an English term. In (15b), it is possible that the coyote was considered to be a small wolf.

584

Olga Lovick and Keren Rice

(15) a. yíhda-­tséle dipper(constellation)-­ dim ‘little dipper’ b. bele-­tséle wolf-­ dim ‘coyote’ Given the range of possible meanings and the strong tendency to lexicalisation, Slavey evaluative morphology is best characterised as derivational. It does not have a regular, predictable effect on the meaning of a particular noun. This is true not only in the cases where there is a clear shift in meaning, but even when it references size or age. While in terms of semantics these items are derivational, nevertheless they have properties that suggest that they originated as separate words, and these are the topic of Section 2.2. See Rice (2014) for additional discussion. 2.2  Word-­like properties of Slavey evaluative morphology 2.2.1  Order of elements Slavey and other Athabaskan languages have a possessive suffix that can occur in the alienable possessive construction. In Slavey, it has the form -­é (sometimes just a tone). The possessive suffix generally follows post-­nominal modifiers, as in (16). (16) a. tu ‘water’ wele ‘hot’ tu-­wele ‘soup’ -­tuwel-­é ‘soup (possessed)’ b. ke ‘shoe(s)’ ch’įle ‘pointed’ ke-­ch’įle ‘high heel shoes’ -­ke-­ch’įl-­é ‘high heel shoes (possessed)’ The diminutive and augmentative follow the possessive suffix, as in (17).7 (17) a. ’ah ‘snowshoe’ -­’ah-­é ‘snowshoes (possessed)’ ’ah-­cho ‘hunting snowshoes’ -­ah-­é-­cho ‘hunting snowshoes (possessed)’ b. łuh ‘spoon’ -­luz-­é ‘spoon (possessed)’ łuh-­ah ‘small spoon, teaspoon’ -­lúz-­ah ‘small spoon, teaspoon (possessed)’ Wilhelm and Saxon (2010), in work on the possessive suffix in Tłįchǫ Yatiì and Dene Sųłiné, argue that this suffix indicates syntactic licensing of a nominal, with its form and occurrence at least partially lexical. If the possessive suffix is syntactic/inflectional, it would be expected to occur outside the modifiers, as in (16). However, with the evaluative suffixes (17), it occurs inside them, unexpectedly if the evaluative morphemes were part of the word.

Slavey (Dene) and other Athabaskan Languages 585 The position of the possessive suffix provides insight into forms that might be considered to be ‘truly’ lexicalised. While in general in the nouns that we have checked the order is noun–possessive suffix–evaluative item, with some nouns this is not the case. The word for ‘horse’ is based on the word ‘dog’ followed by the augmentative, and the possessive suffix is the final item. (18)

tłį ‘dog’ tłį̜-­cho ‘horse’ -­tłį-­cho-­é ‘horse (possessed)’

The possessed form of ‘grizzly bear’ has variable forms. (19)

sah ‘bear’ sah-­cho ‘grizzly bear’ -­sah-­cho-­é ~ -­zah-­é-­cho ‘grizzly bear (possessed)’

The external position of the evaluative items in most cases suggests their origin as separate words. Another suffix, the plural ke, follows the evaluative affixes. (20) dene ‘person’ dene-­cho ‘giant’ dene-­cho-­ke ‘giants’ sah ‘bear’ sah-­ya ‘bear cub’ sah-­ya-­ke ‘bear cubs’ While the possessive suffix is considered to be inflectional, the plural is generally considered to be derivational, primarily because it is not obligatory (e.g. Rice 1989). 2.2.2 Agreement Agreement also suggests that the evaluative items originated as separate words. Athabaskan languages have a class of ‘areal’ nouns referring to an area, place or situation. An areal prefix occurs when an areal noun is subject of an intransitive, direct object of a transitive, object of a postposition, or a possessor, as in the following examples, using the areal noun kǫ́é ‘house’ compared with the non-­areal noun tene ‘pail, bucket’. The areal has the form go-­. (21) a. Agreement with areal subject: tene we’ǫ ‘the pail is located’ kǫ́é gó’ǫ ‘the house is located’ b. Agreement with areal direct object: tene whehtsį ‘she or he made a pail’ kǫ́é góhtsį ‘she or he made a house’ Areal agreement also occurs when an areal noun is modified by an evaluative morpheme or by another modifier. In the final example of (22), the areal has the reduced form o-­. (22) Agreement of evaluative element with areal noun: kǫ́é go-­cho house areal-­aug ‘town, society, big building’

586

Olga Lovick and Keren Rice

náendí k’é go-­cho 3.buy place areal-­aug ‘mall’ mįe go-­cho lake areal-­aug ‘Round Lake’ (literally ‘big lake’) tł’á-­o-­cho bay-­ areal-­aug ‘Big Bay’ While the meanings are not predictable, the morphology again betrays a phrasal origin.

3  Other languages Evaluative morphology is found in languages across the family; in this section we briefly survey several languages. 3.1  Dene Sųłiné Dene Sųłiné (Manitoba, Saskatchewan, Alberta, Northwest Territories) is closely related to Slavey. Rice (2012, 57) gives many evaluative forms in Dene Sųłiné. (23) a. Augmentative chogh: łue-­chogh fish-­aug ‘whale’ jíze-­chogh western.jay-­ aug ‘hawk’ teli-­chogh container-­ aug ‘barrel’ b. Diminutive aze: chiz-­aze lynx-­ dim ‘cat’ jí-­aze berry-­ dim ‘raisin’ łest’édh-­aze bread, bannock-­dim ‘cookie’ c. Diminutive tsele: des tsele river dim ‘creek’ nunie tsele wolf dim ‘coyote’

Slavey (Dene) and other Athabaskan Languages 587 Rice identifies another use of the diminutive: to ‘single out individuals who are of different generations than those referred to be the non-­derived stems, but not necessarily descending generations. Nevertheless, the derived forms do suggest an especially close relationship to ego’ (Rice 2012, 32). (24) Diminutives with kinship terms (Rice 2012, 32–3): Term Literal sunaghaze (2.obj-­foc ‘I give you just a little tiny piece of unploughed land.’ (an almost jokingly small piece) (18) Juma-­n-­ps mam-­m yap al-­shu-­cha-­w mach-­kna. 2-­of-­also mother-­2.poss field buy-­sbj.sub-­lim-­foc irrigate-­ rem ‘Your mother also, when just buying the field, used to irrigate.’ (Hardman 2000, 23) The limitative suffix in Jaqaru does not semantically equal the diminutive one and cannot be replaced by the latter without altering the meaning. The data in (17) were accompanied by a consultant’s facial expression of nothingness, and it would lose its connotation of insignificance and almost of a joke if the limitative suffix were to be taken out. This never occurs with the diminutive suffix: no matter how small the size of the referent implied, at no time is scorn or a connotation of insignificance present in the semantics of such a word/sentence. 2.3 Augmentation Augmentation of root/stem semantics of quality/quantity in Jaqaru is expressed through the morphological process of reduplication. It is very productive and applies to any possible root category, nominal or verbal:

Jaqaru 599 (19)

utxitx-­utxitx warmi-­wa small-­small woman-­foc ‘tiny women’ (a group of)

(20)

ñaw-­ñawi key-­key ‘lots of keys’ (as on a key ring)



(21)



was-­wasa walk-­walk ‘walk around everywhere’

When a nominal root is being reduplicated, it results in a noun, most of the time translated into English with an adjective, as in (19). Quite often the reduplicated nominal root turns into a toponym, referring to a specific place: (22)

qap-­qapu ‘a place with the plants out of which spindle is made’ < qapu ‘spindle’

(23)

juncx’-­juncx’u ‘hot springs’ < juncx’u ‘hot’

(24)

qal-­qala ‘a place, field full of rocks’ < qala ‘rock, stone’

While for some nominal roots the process of reduplication simply augments the root semantics, resulting in a noun generally translated in English as ‘very X’, this is not an across-­the-­board phenomenon. Compare the data in (25)–(26) with (27)–(30): (25)

(26)

(27)



t’usq-­t’usqi dust-­dust ‘a very smoky, dusty place’ jask-­jaski bland.easy-­bland.easy ‘very bland, very easy’ janq’-­janq’u white-­white ‘not so white, losing its colour’

(28)

tz’irar-­tz’irara black-­black ‘medium black’

(29)

cxaw-­cxaw ujara raw-­raw corn ‘medium-­cooked/done corn’







(30)

ch’am-­ch’ama night-­night ‘not very dark, late dusk’

Olga Birioukova and M. J. Hardman

600

What is obvious from the data above is that the root semantics determines the meaning of the reduplicated construction. Perhaps this has something to do with the ability of the root meaning to be measured and evaluated. At this point, this is one of the questions for future research and investigation. When a verbal root is being reduplicated, the meaning is somewhat predictable and is conveyed into English with such modifying phrases as ‘all the time, continuously’, as in (31)–(33):



pijch-­pijchi sweep-­sweep ‘sweep all the time, continuously, all day long’

(32)

payq-­payqa eat liquid foods-­eat liquid foods ‘drink and drink (something) all day long, continuously’

(33)

jaq-­jaqu throw-­throw ‘throw many times, continuously’

(31)





The phenomenon of reduplication for the purposes of augmentation is very productive and works the same way even with verbs borrowed from Spanish: (34)



salt-­salta (Span. saltar ‘to jump’) jump-­jump ‘jump all the time’

It should be noted that not only verbal roots participate in reduplication, but also verbal stems which consist of a verbal root and a derivational suffix of modification. Such derivations have probably been lexicalised together with that suffix and therefore the reduplication of the whole stem is permissible in Jaqaru: (35)



taj-­nuq-­taj-­nuqu step-­on-­step-­on ‘walk step by step’ (Hardman 2000, 53)

Here -­nuqu is a derivational suffix of motion modification, a very productive suffix encountered with many other verbs in Jaqaru. In (35) the whole stem tajnuqu participates in reduplication, which is explicable if the verbal root and the motion modifier were lexicalised as one unit. Phonologically, reduplication is somewhat easy and predictable, at least at the word level: the final vowel of the root is deleted and then the root is reduplicated with the final vowel present.6 However, in a sentence, more vowels could be subject to deletion depending on the morphosyntactic rules of the language. Compare (36) with (26): (36)

Jask-­jask-­ch-­″-­w pal-­w-­i-­qa. bland-­bland-­lim-­obj-­foc eat-­ pfv-­3-­top ‘She, he or they ate just a lot of bland food.’

Jaqaru 601 2.4  Intensification In Jaqaru there exists a special intensifier suffix, -­y. This suffix is added onto a root just before reduplication takes place: (37) jaski-­y-­jaski bland.easy-­ int-­bland.easy ‘very bland, very easy’ (38)

yacx″i-­y-­yacx″i hard.difficult-­int-­hard.difficult ‘very hard, very difficult’

Unlike the reduplication process, the intensifier -­y can only be added onto nominal roots, which change their syntactic category to become adjectives as a result of this process. These are the only ‘adjectives in Jaqaru [which] may occur before or after the noun modified without any restriction’ (Hardman 2000, 54): (39) ch’ama-­y-­ch’am uta-­wa dark-­int-­dark house-­foc ‘very dark house’ Constructions with the intensifier are always translated into English as ‘very X’. Interestingly enough, roots that would resist such interpretation through augmentation do take the intensifier and have no trouble in augmenting the semantics of the nominal root as adjectives. Compare (40)–(43) with (27)–(30): (40) janhq’u-­y-­janhq’u white-­int-­white ‘very white’ (41) tz’irara-­y-­tz’irara black-­int-­black ‘very black’ (42) cxawa-­y-­cxawa raw-­int-­raw ‘very raw’ (43) ch’ama-­y-­ch’ama night-­ int-­night ‘very dark night, moonless night’ For verbal roots there exists a separate intensifier, -­ch. Phonologically, it seems that both intensifiers do not require a deletion of the final vowel of a root, though ­examples with more than three syllables in a root have not been yet encountered.

Olga Birioukova and M. J. Hardman

602 (44) jayra-­ch-­jayra dance-­int-­dance ‘to dance untiringly’ (Hardman 2000, 54)

It should be noted that the meaning of a verbal construction with the intensifier -­ch does not equal the meaning of the reduplicated root/stem, though the semantics is very close. Examples (45)–(46) illustrate this; we repeat (21) and (33) for convenience: (21)



was-­wasa walk-­walk ‘walk around everywhere’

(45) wasa-­ch-­wasa walk-­int-­walk ‘walk constantly at intervals’ (33)



jaq-­jaqu throw-­throw ‘throw many times, continuously’

(46) jaqa-­ch-­jaqa throw-­int-­throw ‘throw constantly at times, intervals’ This concludes the outline of the use of evaluative suffixes and morphological processes of evaluation in Jaqaru.

3  Evaluative morphology of Aymara and Kawki 3.1 Diminutivisation Diminutivisation through suffixation is shared by all of the Jaqi languages. In Aymara it is done with -­qallu, originating from a nominal root qallu ‘young animal’ (Hardman 2001): (47) iwis-­qallu sheep-­ dim ‘baby sheep, lamb’ (Hardman 2001) (48) tari-­qallu cloth-­dim ‘small carrying cloth’ (Hardman 2001) There is also a suffix borrowed from Spanish, -­itu, used when referring to the referent’s size. Often this is part of a whole borrowing from Spanish, as in (50). This suffix is homophonous with the 3.sbj>1.obj agreement morpheme and, therefore, can cause some ambiguous and humorous situations (Hardman 2001).

Jaqaru 603 (49) chuym-­itu heart-­ dim ‘little heart’ (50)

chik-­itu < Span. chiquito ‘little boy’ (Hardman 2001)

The Kawki diminutive suffix is the same as in Jaqaru, -­uña, and the distribution of this suffix is very similar to that of the Jaqaru diminutive, that is, it appears directly after the nominal root: (51) patx-­uña-­ta terrace-­dim-­dir ‘from the little terrace’ (Lyra Language Editor, n.d.) (52) kapr-­uñ-­nh goat-­dim-­1.poss ‘my little goat(s)’ (Lyra Language Editor, n.d.) 3.2 Limitation As was mentioned earlier, the limitative suffix goes back to Proto-­Jaqi, and is therefore found in Aymara and Kawki as well. The phonological shape of this suffix is the only aspect that differentiates the limitative in all three languages; the semantics is the same. In Aymara it is -­ki: (53)

jank’a-­k soon, fast-­lim ‘just soon, just fast’ (Lyra Language Editor, n.d.)

(54) Iki-­ña-­ni-­ki-­ ã-­t-­wa. sleep-­nom-­poss-­lim-­idnt-­1.poss-­foc ‘We have just blankets.’ (Lyra Language Editor, n.d.) In Kawki the suffix is -­cha, just as in Jaqaru. The order of the limitative is also the same as in Jaqaru – it follows a diminutive: (55) qaq-­uña-­ch-­na rock.hill-­dim-­lim-­dir ‘just in the littlle hill’ (Lyra Language Editor, n.d.) 3.3 Augmentation Reduplication is also present in Aymara for the purpose of augmenting the semantics of the root/stem. Example (56) demonstrates it even with a borrowed Spanish root and an Aymara directional suffix reduplicated as one unit:7 (56) altu-­r-­altu-­. < Span. alta/o ‘high’ high-­ dir-­high-­dir ‘higher and higher’ (Hardman 2001)

604

Olga Birioukova and M. J. Hardman

(57) k’ach″a-­t-­k’ach″a.t slow(ly)-­dir-­slow(ly)-­dir ‘one by one, very slow(ly)’(Hardman 2001) Cases of reduplication in Kawki are yet to be discovered. 3.4  Intensification Intensification has been encountered only in Kawki and the data for it are quite limited, though it does exist. It is very similar to intensification in Jaqaru: (58) jallk ″a-­y-­jallk″a sudden(ly)-­int-­sudden(ly) ‘very sudden(ly)’ (Lyra Language Editor, n.d.)

4 Conclusion In Jaqaru, as in the other two Jaqi languages, evaluative morphology is expressed through the use of suffixes which perform functions of diminution, limitation/reduction, augmentation and intensification. Despite some of the phonological variations, the shared morphological processes are very much alike. All three languages also have a diminutive suffix borrowed from Spanish which occurs in Jaqi words exactly where the native suffix would normally be used. There exists neither pejorative nor any type of comparative (hierarchical) suffixes in these languages, because ranking and comparison are not part of any Jaqi language grammar or of a Jaqi mindset. Thus, these languages have their native, original ways only to evaluate referents, but not to compare or judge. When a ranked comparison is needed, ‘this is done with Spanish borrowing or by implications’ (Hardman 2000, 24).

Acknowledgements We would like to express our special gratitude and thanks to our consultant and a native speaker of the Jaqaru language, Dr Dimas Bautista Iturrizaga. The research and data for this chapter have become possible due to his unwaning energy and perpetual interest in researching of his native language. Jilatxi!

Notes 1. WALS (Dryer and Haspelmath 2011) classifies Jaqaru as belonging to the Aymaran family and Ethnologue (Lewis, Simons and Fennig 2014) does not offer any classification; however, there exists historical evidence that Jaqaru, Kawki and Aymara are descended from one Proto-­ Jaqi family (Hardman and Mosley 1986). It is a common misconception to consider Jaqaru an Aymaran language, thus referring to the former as a dialect of Aymara. Jaqaru, Kawki and Aymara are three extant languages belonging to one linguistic family. Their membership was determined through the sharing of certain fundamental features of structures which distinguish this family from other neighbouring ones. The term ‘Jaqi’ will be used throughout the chapter to refer to the linguistic family of the Jaqaru as well as Aymara and Kawki languages. Aymara is also spoken in Bolivia and Chile.

Jaqaru 605 2. Hardman, p.c., March 2012. 3. Hardman, p.c., January 2013. 4. -­mashi is a very peculiar suffix. Semantically it means ‘companion’ and, hence, it is glossed as cmp throughout the chapter. When -­mashi gets added on to a root with several suffixes already present, it triggers a new cycle of nominal suffix order, thus behaving more like a root itself and turning the whole derivation into a compound noun. However, it is impossible to have mashi by itself as a root – all the attempts to use it independently with nominal suffixes were rejected by the native speaker. 5. The term ‘limitative’ seems to fit the semantics of this suffix the best, and therefore the abbreviation lim is used for this morpheme throughout the chapter. 6. This applies to roots with two or three vowels. It has not been tested with roots that have four or more syllables. 7. These data are translated with adjectives in the comparative form in English. It is important to remember that the Jaqi languages do not have ranking, and therefore in no way does the Aymara phrase in (56) aim to compare the heights of two or more referents.

20.2 Kwaza

Hein van der Voort

1 Introduction Kwaza is a highly endangered language isolate of the Brazilian Amazon with twenty-­five remaining speakers. First contacts with Western culture must have taken place about a century ago. Nowadays, the traditional ways of life are remembered only by the elderly. The traditional neighbours of the Kwaza are the Aikanã (isolate), Kanoê (isolate), Latundê (Nambikuara) and Salamãy (Tupian). Kwaza was described by van der Voort (2004). It displays similarities with neighbouring languages, due to areal diffusion throughout the Guaporé–Mamoré region (Crevels and van der Voort 2008). The parts of speech are verbs, nouns, adverbs and particles. Attributive modification is expressed by nominal juxtaposition. Word order is variable, but head-­final orders are predominant. In complex sentences, the main clause tends to be sentence-­final. Main word stress tends to fall on the last syllable of the stem, before the inflectional complex. Kwaza is a morphologically complex suffixing language. Mood, person and case are inflectional. A minimal sentence consists of a verb root and subject and mood inflections. Noun morphology is simple, with restricted case marking and no number. Kwaza derivational morphology is used for word formation, tense, modality, aspect, valency, direction, classification and other categories. Reduplication has various functions. Evaluative morphology tends to be derivational. Grandi and Körtvélyessy (this volume) propose a universal definition classifying evaluative expressions against axes of description and quality. Bauer´s (1997) typological overview of evaluative morphology includes observations from a handful of South American languages. In this chapter, I will present a unified description of evaluative morphology in Kwaza. Kwaza possesses a certain variety of evaluative expressions, employing different kinds of strategies, including grammatical constructions, lexical items and combinations of both. However, Kwaza is relatively poor in evaluative morphology, and most morphological evaluative expressions are of a rather marginal nature. This chapter will focus on morphological constructions. All Kwaza data result from my fieldwork.

2  Expression of basic evaluative notions Wierzbicka´s (1999, 36) semantic primitives BIG, SMALL, GOOD and BAD are expressed in Kwaza by verb roots in analytic constructions. Some of them may occur in compounds and some occur in morphologically elaborate constructions that express less

Kwaza 607 prototypically evaluative notions. Here I will briefly discuss the expression of the diminutive, augmentative, ameliorative and pejorative. The diminutive in Kwaza is in principle expressed by the adverb txuhũi ‘small, little’, which can also be used as a verb root: ‘be small/little’ or ‘do little’. Furthermore, it is attested in compounds. Examples (1) and (2) illustrate compounding with the verb wai-­ ‘be good’ and the noun jerexwa ‘jaguar’, respectively: (1) Wai=txuhũi-­tse. good=little-­ decl ‘It got a little better.’/‘It tasted somewhat good.’ (2) jerexwa=txuhũi jaguar=little ‘cat, little dog’ Diminutives as in (2) are not very productive. It is possible that the derivational nominal classifier -­tohoi ‘child’, as in jerexwa-­tohoi ‘jaguar kitten’, originates historically from txuhũi. The augmentative is expressed only analytically, by the verbs nỹ-­ and txi-­‘to be big’, which incorporate a nominal classifier, representing the object of evaluation. The ameliorative is also expressed by a verb root, in this case the verb wai-­‘be good’, which usually incorporates a classifier, adding the connotation ‘be beautiful’. These expressions will not be discussed further here, but they are included in Table 20.2.1 below, which is based on Grandi and Körtvélyessy´s classification. With regard to the pejorative sense, this is considered in Kwaza as an antonym of the ameliorative. As in the neighbouring isolate language Kanoê (Bacelar and da Silva Júnior 2003), antonyms of quality verbs are often created through litotes, that is, through morphological negation of the opposite. In Kwaza, the pejorative is also expressed through litotes: wai-­ ‘be good’ vs. wai-­he-­ ‘be bad’. Although this is not a clear case of evaluative morphology, it does represent an evaluative meaning expressed with the help of morphological means.

3  Specific evaluative expressions In contrast to the basic evaluatives, several specific and sometimes rather marginal evaluative notions are expressed in Kwaza by morphology. Here, I will present these evaluative expressions by morpheme. In Section 4, Table 20.2.1 summarises evaluative expressions by function. 3.1  Expletive use of -­nãi The adverbial nominaliser -­nãi creates infinitive-­like nouns that may function as complements or arguments. It is also used to derive adverbs. With evaluative verb roots it seems to form a fixed combination that is often used with an expletive sense. Example (3) features the verb root wai-­‘be good’, which in combination with -­nãi is used as an adverb and as a sentential particle, expressing scrutiny: (3) Wai-­nãi tãtãi-­ra. good-­ nmls pound-­ imp ‘Beat it well!’ (connotation: ‘Don’t mess around, but hit it right!’)

Hein van der Voort

608

In (3), the bound morpheme -­nãi adds an expletive connotation to an already evaluative sense. The verb root kukui-­‘hurt’ itself is also used as an expletive particle meaning ‘Wow!’, ‘Damn(ed)!’ or ‘Yecch!’ When combined with -­nãi it often expresses disapproval, surprise or amazement: (4) Kukui-­nãi ja-­he-­xa-­hỹ-­tsy. hurt-­ nmls eat-­neg-­2-­nmls-­reg ‘My! You eat nothing!’ 3.2  Intensifying -­te(te) Intensification is expressed in various ways, both lexically and morphologically. The intensifying suffix -­tete, or its rarer non-­repeated allomorph -­te, is applied to verbs, nouns and adverbs, with different connotations. It does not change or determine the grammatical category of the word. When applied to verbs it has an augmentative effect, as (5) and (6) show: (5)

Awe nỹ-­mũ-­te-­ki. clf:liquid-­int-­decl rain big-­ ‘It is raining heavily.’

(6) Kukui-­tete-­ki. hurt-­ int-­decl ‘It hurts a lot.’ When applied to adverbs, this suffix has an intensifying effect, as (7) and (8) show: (7)

Txarwa-­te hãrã-­ki. first-­int stop-­ decl ‘He just finished a second ago.’

(8) Ti-­rjỹ-­tete-­xa-­hỹ-­re. what-­ dir:area-­int-­2-­nmls-­intg ‘Where exactly did you (say it was)?’ When applied to nouns, it expresses authenticity: (9) kanwa-­tete canoe-­ int ‘real canoe’ The intensifying morpheme has apparently undergone areal diffusion, since the neighbouring isolate Kanoê has a similar morpheme with a similar meaning, -­kete-­ ‘really’ (Bacelar 2004). Furthermore it is doubtlessly related to similar morphemes meaning ‘very, really, truly’ etc. in many Tupi languages, for example -­te(te) in Mekens (Galucio 2001).

Kwaza 609 3.3  Intensive habitual -­tjarjỹ-­ The verbal habitual morpheme -­tjarjỹ-­has an intensive connotation, ‘always much’: (10)

Awe-­tjarjỹ-­ki. rain-­much-­decl ‘It is raining too much.’ (context: it has been raining every day)

The intensive habitual apparently contains a fossilised instance of the directional morpheme -­rjỹ-­‘area’, adding the sense of ‘all around’ to the habitual meaning. 3.4  Intensive use of nominaliser -­hỹ Kwaza morphology is recursive and unmarked verbalisation is very productive. Consequently, nominalised verbs can in turn be (zero-­)verbalised again and then nominalised and so on. In the following examples, the nominaliser -­hỹ may be added before verbalisation in order to form a construction that is reminiscent of clefting, creating focus on one of the arguments of the (derived) verb: (11)

Axy nỹ-­xy-­hỹ-­ki. clf:house-­nmls-­decl house big-­ ‘The house is a big one.’

Although the literal meaning of (12) is ‘I fried pieces of manioc’, the nominaliser -­hỹ-­ creates intensive focus on ‘pieces of manioc’: (12) Jo-­kanε txitxi-­dy-­da-­hỹ-­ki. manioc-­ clf:oblong fry-­ caus-­1sg-­nmls-­decl ‘They are pieces of manioc which I fried.’ This intensifying use of -­hỹ may have a comparative and superlative connotation: (13) Nỹ-­hỹ-­hỹ-­ki. big-­ nmls-­nmls-­decl ‘It is bigger./It is the biggest.’ Note that recursion of the nominaliser in (13) is not an instance of reduplication. The first occurrence of the nominaliser -­hỹ is part of the root and can be replaced by a specific classifier: (14) Nỹ-­ro-­hỹ-­ki. big-­ clf:cup-­nmls-­decl ‘The cup is bigger/biggest.’ Verbs derived from nouns through verbalisation can be nominalised again, and this use of the nominaliser -­hỹ-­ has an intensifying or focalising effect. Note the first word of (15):

Hein van der Voort

610

(15) Tswa-­wa-­hỹ hay=dy-­wa-­rati erewe-­here=jã-­re. man-­ indf.sbj-­nmls cut=cut-­ indf.sbj-­clf:man write-­ aprl=be-­intg ‘That very man who cut (the mortar) he paints it, you see.’ The noun tswa ‘man’ has apparently been verbalised, as witnessed by the indefinite subject cross-­reference morpheme -­wa-­. The resulting predicate means ‘Someone is a man’ or ‘They are men’, which is then nominalised by -­hỹ, yielding a noun meaning ‘The one who is a man’ or ‘Those who are men’. 3.5  Intensifying reduplication Reduplication of verb roots and adverbs may have an intensifying meaning, in addition to other semantic effects: (16) Ỹ-­hỹ-­ko awe~awe-­tse. nmls-­ins rain~rain-­ decl this-­ ‘Here it always rains a lot.’ Note that this operation usually has a descriptive perspective, although it can also have a qualitative perspective: (17) Watxi~watxi ãwỹi-­ra. true~true see-­ imp ‘Aim right!’/‘Look well!’ Reduplication of (usually) the first syllable of verbal and adverbial roots can also have an intensifying effect: (18)

Ca~cari-­ki duture. red~shoot-­decl pig ‘He killed many pigs.’

(19)

Txa~txarwa-­te onε-­ki. red~first-­int arrive-­ decl ‘He arrived suddenly.’

3.6  Attenuative and endearing reduplication Full reduplication of verb roots may also have an attenuated effect, although this is not very productive. The verb root kuri-­‘to stop, be quiet’ is recognisable in the fixed verb root kurikuri-­‘do a little later’, ‘do in a minute’.1 Perhaps the attenuative is more productive with partial reduplication. Whereas full reduplication of the verb root kahε-­ ‘to bite’ has an intensifying effect, kahε~kahε-­ ‘keep on biting (ferociously)’, reduplication of its first syllable produces an attenuated and perhaps endearing meaning:

(20)

Jerexwa-­tohoi ka~kahε-­nỹ-­ki. jaguar-­small red-­bite-­refl-­decl ‘The little dog is biting (playfully).’

Kwaza 611 The same was apparently attested with the verb root rãmã~rãmã-­‘to rattle’, where reduplication of the last syllable seems to have an attenuative effect: rãmã~mã-­‘to tremble’. The endearing example (20) may imply that not all attenuation should be regarded as a shift to the negative end of the qualitative perspective. 3.7 Lamentative -­ʔwỹtε-­ Kwaza has a special lamentative suffix -­ʔwỹtε-­, which refers only to animate beings. Its meanings range from ‘poor’ and ‘what a pity’ to ‘the late’. It can be applied to both verb roots and verb stems, and to nouns.

(21) (22)



Ereri-­ʔwỹtε-­ki. grow-­pity-­decl ‘It is a pity (the doggie) grew so much.’ djy-­ʔwỹtε joxu ynỹ-­hỹ-­le brother-­pity Joxu call-­nmls-­only ‘only (my) regretted brother the late one called Joxu’

There might be an etymological relationship with the verb ãwãte-­‘be sad’. 3.8  Evaluative uses of the interrogative mood As seen in many of the examples, clauses normally end with a person marker followed by a mood marker. The interrogative is one of the main clause mood inflections. Apart from its canonical interrogative use, the interrogative is also used for emphasis. Here I will illustrate some evaluative applications of the emphatic interrogative. The most transparent of these is a common interrogative expressing indignation:2 (23)



Pε͂rε͂jã=dyitxirjỹ-­xa-­re. speak=noise.make-­2-­intg ‘How noisily are you conversing!’

The interrogative mood is also used evaluatively in fixed expressions involving particular lexical or morphological elements. A common expletive construction combines the interrogative mood with the particle kukui ‘Wow!/Yecch!/Damn(ed)!’: (24) Kukui ñãsi-­he-­da-­re. wow! hear-­ neg-­1sg-­intg ‘Damn, I can’t hear anything!’ Another fixed expression involving the interrogative conveys relief. It is based in part on non-­canonical use of the procrastinative aspectual suffix -­t(j)ara-­, which normally postpones the event to the near future. When followed by the nominaliser -­hỹ-­ and the cosubordinative mood marker -­ta, the morpheme -­t(j)ara-­ expresses relief that something was the case:

Hein van der Voort

612 (25)



Tẽitja a-­da-­tara-­hỹ-­ta. alone exist-­1sg-­proc-­nmls-­cso ‘Luckily, I was alone.’

When combined with the first person inclusive pronoun txana, the negative morpheme -­he-­ and the interrogative mood -­re, -­t(j)ara-­ indicates relief that something was not the case: (26) Txanã awe-­he-­tjara-­re. we.incl rain-­ neg-­proc-­intg ‘Good that it did not rain!’ In another construction, regret or compassion is expressed through the verb root daki-­ ‘it’s a pity’ in a subordinate clause preceding an interrogative main clause: (27) Daki-­tja tsẽi-­re. pity-­ cso tear-­intg ‘A pity it (the cloth) tore.’ Example (27) does not represent a question sentence, so one would expect the main clause to be in the declarative mood (-­ki or -­tse), instead of the interrogative (-­re). However, the interrogative mood marker is used evaluatively here, lending intensity to the expression of regret by the subordinate verb. The interrogative form of the appellative conjectural -­herejã-­ is often attached to nouns or verbs in order to draw the attention of the hearer, or even to express surprise or indignation about the necessity of mentioning something that should be self-­evident to the hearer: (28) Mã-­herejã-­re. mother-­ aplt-­intg ‘It is mother, you silly!’ (29)

A~anũ-­herejã-­re. red~plant-­aplt-­intg ‘He is planting, don’t you see?’

Again, the interrogative mood marker is used evaluatively here, putting emphasis on the entire expression. 3.9 Expletive -­rai-­ ‘damned!’ The morpheme -­rai ‘damned!’ has basically an exclamative function. Although it can be used in a neutral way, it usually has a strongly expletive connotation. However, this often occurs in an ironic context of faking indignation. The morpheme is mostly applied to verb stems, between the person cross-­reference and mood markers, which is unusual for derivational morphemes:

(30)

Ãwãka-­da-­rai-­re. heron-­1sg-­damn-­intg ‘Am I a bloody heron?’ (that is, ‘Should I eat all this fish?’)

Kwaza 613 (31) Kwaza-­dy-­nãi waihỹ=wara areta-­da-­rai-­he-­ki. Kwaza-­ pos-­nmls nice=but know-­1sg-­damn-­neg-­decl ‘Kwaza is beautiful but I don’t understand it at all.’ As observed by Bauer (1997, 558), it is not uncommon for ‘expressive morphology’ to be less subject to a fixed position and to appear in alternative morphological patterns. This can also be observed with the Kwaza expletive. In answer to the ironical question posed to a male consultant whether he thought his brother-­in-­law was pretty, he provided two alternative answers: (32)



Waihỹ-­rai-­re. nice-­damn-­intg ‘Pretty no way!’

(33) Waihỹ-­he-­rai. nice-­ neg-­damn ‘Pretty no way!’ Example (34) is a possible response to a compliment about tasty soup that the speaker made. In this context the expletive alternative in fact expresses modesty: (34) Wai-­he-­rai. good-­ neg-­damn ‘It does not taste good.’ It may have been noted that -­rai occurs often in emphatically used interrogative clauses. Note that, if it is combined with the negative morpheme -­he-­, it will not have a negative sense: (35) Kui-­he-­rai-­re. drink-­ neg-­damn-­intg ‘Did he drink! Jesus!’ However, when the negative is absent the expletive construction may have a negative sense: (36) Kui-­rydy-­hỹ-­rai-­re. drink-­ irr-­nmls-­damn-­intg ‘He isn’t going to drink, no way!’ (37) Atsile-­rydy-­hỹ-­rai-­re. heavy-­ irr-­nmls-­damn-­intg ‘Ah, that doesn’t look heavy at all!’ (38) Kukui-­dy-­da-­rai-­re=jã-­si nãi-­xa-­re. hurt-­ caus-­1sg-­damn-­intg=be-­swr like-­2-­intg ‘It doesn’t hurt, no!’ (speaker does not want to admit he has pain)

Hein van der Voort

614

Example (39) represents a fixed expression, in which -­rai behaves as if it were a mood inflection, followed by a cosubordinative clause based on the verb root wai-­‘good’: (39)



Karari-­da-­rai thirst-­1sg-­damn ‘I’m so thirsty!’

wai-­ta. good-­cso

In example (40), -­rai is attached to a noun, which is then apparently zero-­verbalised in the interrogative mood: (40)



ñũ-­rai-­re faeces-­damn-­intg ‘It isn’t dung, pal!’ (context: the eater was afraid to taste liquorice)

In all the examples in this section, the evaluative morphological element -­rai occurs in – for derivational morphemes – unusual positions, playing an important role in frequently used evaluative expressions. It may be etymologically related to the verb root raihỹ-­ ‘to ruin’, and/or to the root of nouns such as raiʔỹhỹ ‘animal, thing’, raiʔỹnũ ‘poison’.

4  Expression of evaluative functions In Sections 2 and 3 we have seen the different uses of the evaluative morphemes of Kwaza. Table 20.2.1, based on Grandi and Körtvélyessy (this volume), represents a summary of the different evaluative expressions organised by function. The majority of evaluative functions in Kwaza are expressed by derivational suffixes. Several expressions combine with the interrogative mood. Furthermore certain evaluative morphemes have a wide range of use. The suffix -­rai, the interrogative mood and reduplication are used in three of the four types of evaluative notions, covering the positive end of the descriptive perspective and the entire range of the qualitative perspective. The fourth type, the negative end of the descriptive perspective, is expressed by the adverb txuhũi, although sometimes through compounding.

5 Conclusion Kwaza employs morphological strategies to express many evaluative functions. The distribution of these forms and functions over the four-­way typology based on semantic primitives is uneven and the prototypical evaluative functions are rarely expressed morphologically. Evaluative morphology in Kwaza tends to be derivational, although inflectional morphology is involved in several expressions. Most evaluative morphology attaches to verbal roots or stems, and does not change the grammatical category of the constituent (except adverbial nominaliser -­nãi and some evaluative uses of nominaliser -­hỹ). Some expressions, such as litotes and the form -­tete-­, are shared with neighbouring languages, due to areal diffusion. No conspicuous phonological adaptations were observed.

Acknowledgements This chapter is based on data collected during my fieldwork in Rondônia. I am grateful to my teachers of Kwaza: Kyikãu Mãdε, Teteru, Towenwe, Turuwε, Wari and Waruwaru

Kwaza 615 Table 20.2.1. Expression of evaluative functions in Kwaza, with reference to (examples) or [sections] Descriptive perspective Shift towards positive end

[BIG] [2] Augmentation (39) (24) (16) (5)-­(6)

nỹ-­, txi-­ v-­rai wai-­ kukui-­intg v reduplicate v-­tete

Qualitative perspective [GOOD] [2] wai-­ Intensification (38)–(39) v-­rai (11)–(15) v-­hỹ (10) v-­tjarjỹ (7)–(8) adv-­tete (17)–(19) adv/v reduplicate Authenticity Relief Endearment

Shift towards negative end

[SMALL] [2] Diminution (1)–(2)

txuhũi v/n compound

[BAD] [2] Attenuation

(9)

n-­tete

(25)–(26) (txana) v-­tara-­intg (21)–(22) (27) (20)

n/v-­ʔwỹtε-­

daki-­

v reduplicate

wai-­he-­

(20) v reduplicate Contempt/Indignation [3.1] v-­nãi (scrutiny) (30)–(37, 40) v-­rai (23) -­intg v-­herejã-­intg (28)–(29) kukuinãi (disapproval) (4)

for their patience, and to the inhabitants of the Tubarão-­Latundê indigenous reserve for their hospitality. The research was funded by the Netherlands Organisation for Scientific Research (NWO) and the Amazon Research Foundation (FAPESPA, programa PPDOC).

Notes 1. This form is conspicuously similar to the adverb kori(kori) ‘(very) soon’ in Tupi-­ Guarani languages. 2. This may not be unusual cross-­linguistically: note how the English translation of the first example also involves an interrogative expression.

20.3 Lule

Raoul Zamponi and Willem J. de Reuse

1 Introduction Lule was once spoken by a semi-­nomadic people known under the same name who lived in the Gran Chaco between the Pilcomayo river and the Andean foothills of north-­western Argentina. It is a dialect of an extinct language, also called Lule (or, less commonly, Lule-­Tonocoté), also spoken in that area by other groups: the Ysistiné, the Toquistiné, the Oristiné and, probably, the Tonocoté. This last group, part of the sedentary indigenous population of the present-­day provinces of Tucumán and Santiago del Estero, is treated as linguistically separate from the Lule by early colonial sources such as the Relaciones geográficas de Indias (Jiménez de la Espada 1965, I: 390–6; II: 78–85). They are, however, connected linguistically to the Lule by the Sardinian Jesuit missionary Antonio Maccioni (1688–1753) in his grammar and vocabulary of Lule (proper) as shown in the title of his work: Arte y vocabulario de la lengua lule y tonocoté (Machoni 1732). It is likely that the Tonocoté, who suffered constant attacks from the Lule, had shifted to the Lule language by Maccioni’s time. Mentions of Lule and its co-­dialects disappear from the historical record after the expulsion of the Jesuits in 1776 (Viegas Barros 2001, 14). At the time of renewed interest in the indigenous languages of Argentina, under the impetus of scholars such as Lafone y Quevedo (1894), Calandrelli (1896) and Cabrera (1910), it became clear that no Lule speakers could be found. We may surmise that the Lule language lost its last speakers sometime in the course of the early nineteenth century. The external relationships of the Lule language are controversial. The most widely accepted affiliation is with the moribund Vilela of the Bermejo river basin further east1 to form the Lule–Vilela family. The possibility cannot be excluded, however, that the shared vocabulary of the two languages (Viegas Barros 2001) is entirely due to contact, rather than common ancestry, and that they are geographically close isolates (Zamponi 2008, li–lv). More work on the precise historical relationship between Lule and Vilela is therefore a strong desideratum. The above-­mentioned work by Maccioni, first published in Madrid in 1732, is our primary source for the Lule language.2 The picture of Lule that emerges from Maccioni’s Arte is that of an agglutinating language with a rather simple and basically suffixal morphology. The language appears to have open classes of nouns, verbs (which includes most of the words that are semantically adjectives) and adverbs, and closed classes of pronouns, demonstratives, postpositions and particles (including discourse markers and interjections). Nominal inflectional categories include number (singular vs. plural) and possession.

Lule 617 Verbal inflectional categories include person (of the subject), number, tense and mood. The derivational morphology is somewhat richer. The language has a set of ‘transpositional’ affixes with a specific word-­class-­changing function and a wider set of ‘semantically rich’ affixes (with a concrete meaning) including nine instrumental prefixes (which is quite unusual for the Chaco area, and actually for South America as a whole) and a set of morphemes expressing evaluative values.

2  The system of evaluative morphology Eight evaluative suffixes (or morphemes that we may consider as such3) are attested in Lule, as well as a polysemic reduplicative process which expresses, among other things, attenuation and speed. Most of these suffixes and the reduplicative process are not restricted to a specific word class, being attested with verbs and nouns, and sometimes also with adverbs. In the following subsections, each of the affixes and the reduplicative process will be examined as far as semantics and uses are concerned. The forms and the examples are drawn from Maccioni’s Arte, and will be presented in a tentative phonemic transcription based on the analysis of Maccioni’s orthography of Lule proposed in Zamponi (2008).4 2.1  -­eks ‘celerative’ Used with activity verbs denoting activities that can be carried out at variable speed, the suffix -­eks signals (just like the above-­mentioned reduplicative process; see Section 2.9. and also compare examples (1) and (51)) that a given activity is realised quickly: (1) ne-­eks-­ (2) is-­eks-­ come-­ cel urinate-­ cel ‘come quickly’ (D:132) ‘urinate quickly’ (D:100) (3) apes-­eks-­ go.down-­ cel ‘go down quickly’ (D:24) 2.2  -­etsi ‘ameliorative’ The suffix -­etsi, a grammaticalisation of the verb etsi-­ ‘be good, healthy’, appears in the Arte with two state verbs. It conveys an ameliorative meaning: (4) un-­etsi-­ (5) akem-­etsi taste-­ aml give.off.a.smell-­ aml ‘have a good taste’ (D:118) ‘give off a good smell’ (D:114–15) It also occurs with state and activity verbs, producing state verbs with the meaning ‘be good to v’ or ‘worthy of v-­ing’. Interestingly, in these forms, -­etsi does not appear adjacent to the stem of the base verb, but on the right periphery, after the (inflectional) subject marker: (6)



kai-­p-­etsi (7) eat-­3sg-­aml ‘it is good to eat’ (G:54)

amaitsi-­p-­etsi love-­3sg-­aml ‘(it is) worthy of love, lovable’ (G:70)

618

Raoul Zamponi and Willem J. de Reuse

unuk-­p-­etsi (9) yet-­ip-­etsi loathe-­3sg-­aml dance-­3sg-­aml ‘it is worthy of hate, contemptible’ ‘it is a good time to dance’ (G:55) (G:70) (8)

The stem with which the suffix -­etsi occurs may also include a nominalising ­reduplicative process (see also Section 2.9. for more on reduplication), as seen in (10) and (11): (10) ti•ti-­p-­etsi (11) do•nmls-­3sg-­aml ‘it is good to do’ (G:55)

eutita ka[•ka]i-­p-­etsi now eat[•nmls]-­3sg-­aml ‘now is a good time to eat’ (G:55)

Followed by the intensive suffix -­ike (Section 2.6.), -­etsi is also used with nominal stems. Added to the noun ekis ‘dream’, the suffix sequence of -­etsi plus -­ike indicates that the referent of the noun is good: (12) ekisetsike |ekis-­etsi-­ike| dream-­ aml-­int ‘good dream’ (D:123) The same sequence added to the noun ilɁe ‘elder sister’ indicates that the referent of the noun is a blood relative: (13) ilɁetsikep |ilɁe-­etsi-­ike-­p| elder.sister-­ aml-­int-­3sg ‘his or her elder sister (by blood)’ (G:93) 2.3 -­eyu ‘pejorative’ Contrasting with -­etsi, the suffix -­eyu is pejorative, and is a grammaticalisation of the verb eyu-­‘become ill’. It is attested with two state verbs (14)–(15) and one noun (16): (14) un-­eyu-­ (15) akem-­eyu-­ taste-­ pej give.off.a.smell-­ pej ‘have a bad taste’ (D:13) ‘give off a bad smell’ (D:99) (16) ekis-­eyu dream-­ pej ‘nightmare, bad dream’ (D:123) 2.4 -­(i)pan ‘derogatory’ The suffix -­pan (postvocalic variant) or -­ipan (postconsonantal variant) has a derogatory connotation and is attested with nouns referring to human beings only:5 (17) kumue-­pan woman-­ der ‘prostitute’ (D:112)

Lule 619 It is also attested with two verbal stems, producing two terms for ‘slave’: (18) met-­ipan (19) wenek-­ipan take-­ der buy-­der ‘slave’ (D:36) ‘bought slave’ (D:36) 2.5 -­ikeps ‘augmentative’ The suffix -­ikeps is augmentative when used with nouns denoting inanimate referents: (20)



epu-­kwe-­ikeps lightning-­son-­aug ‘flood’ (epukwe ‘rain, shower’) (D:55)

With a kinship term, it indicates that the referent of the noun is a blood relative: (21) pe•pe-­ikeps (22) yaɁa-­ikeps-­ts father•rdp-­aug brother-­ aug-­1sg ‘my father (by blood)’ (G:92) ‘my brother (by blood)’ (G:93) It is also used with state verbs. In this case, it has an intensifying meaning: (23) etsi-­ikeps-­ |etsi-­ikeps-­| (24) leteps-­ikeps-­ be.good-­ aug be.damp-­ aug ‘be very good’ (D:94) ‘be very damp’ (D:78) (25) ewikeps-­|ewi-­ikeps-­| (26) ap’u-­ikeps-­ be.a.lot-­ aug be.stuffed-­ aug ‘be very much’ (D:94) ‘be very stuffed’ (D:72) Finally, -­ikeps also appears to intensify the adverb mia ‘only’ producing a verb meaning ‘be just only’. The 3sg present indicative form of this verb, in turn, also has an adverbial use attested by (27). This verb can also occur as an adverb with the adverbialising suffix -­le, as in (28): (27) mia-­ikeps-­p (28) mia-­ikeps-­le only-­ aug-­3sg only-­ aug-­advls ‘just only’ (D:124) ‘truly’ (G:86) 2.6  -­ike ‘intensive’ The suffix -­ike occurs with state verbs with the effect of intensifying their meaning. The difference between -­ike and -­ikeps is not always clear. Pairs of stems attested with both -­ikeps and -­ike are in examples (23)‒(31), (25)‒(35) and (28)‒(40): (29) waleks-­ike-­ (30) amaitsike-­|amaitsi-­ike-­| know-­ int love-­ int ‘know a lot, be very wise’ (G:63) ‘desire’ (D:53)

Raoul Zamponi and Willem J. de Reuse

620

(31) etsike-­|etsi-­ike-­| (32) lokots-­ike-­ be.good-­ int be.soft-­ int ‘be very good’ (D:94) ‘be very soft (to the touch)’ (D:27) (33) yeun-­ike-­ (34) kelots-­ike-­ be.happy-­ int be.clear-­ int ‘be very happy’ (G:86) ‘be clean’ (D:9) (35) ewike-­ |ewi-­ike-­| int be.a.lot-­ ‘be very much’ (D:94)

(36) wele-­ike-­ be.second-­ int ‘be the last, be final’ (D:108)

The forms in (37)–(38) are 3sg indicative present forms, used adverbially, of verbs derived by lokots-­ ‘be mild’ by means of -­ike and of the above-­mentioned reduplicative process, here expressing attenuation (see Section 2.9). Of these forms, that in (38) is unique in that it contains this reduplicative process twice, once on lokots-­and once on -­ike. The combined semantic import of -­ike and of attenuative reduplication, both in (37) and in (38), cannot be derived from Maccioni’s glossing, but whatever it might turn out to be, it is definitely expressive: (37) loko[•ko]ts-­ike-­p (38) loko[•ko]ts-­ike[•ke]-­p be.mild[•att]-­int-­3sg be.mild[•att]-­int[•att]-­ 3sg ‘mildly’ (G:86) ‘mildly’ (G:86) The corpus also contains a kinship term noun formed with -­ike. Like -­ikeps in (21) and (22), -­ike indicates that the referent of the stem noun is a blood relative: (39) melu-­ike-­p younger.sister-­ int-­3sg ‘her younger sister (of a woman, by blood)’ (G:93) In addition, like -­ikeps in (27) and (28), the suffix -­ike can intensify the adverb mia ‘only’ resulting in a verb whose 3sg indicative present form is also used adverbially (40): (40) mia-­ike-­p only-­ int-­3sg ‘certainly’ (D:47) Finally, (41) is a state verb derived by means of -­ike from a stem pili(-­) that may be the noun pili ‘wing; feather’ or the verb pili-­‘fly’: (41)

pilike-­|pili-­ike-­| wing/feather/fly-­int ‘be light (in weight)’ (D:83)

Lule 621 2.7 -­ewi ‘intensive’ A further intensive suffix is -­ewi, a grammaticalisation of the verb ewi-­ ‘be many’. It is attested in the following state verb: (42) pei-­ewi-­ be.wide-­ int ‘be very wide’ (D:14) 2.8  -­ke ‘approximative’ The approximative suffix -­ke means ‘about, more or less’. It is attested with two adverbs: (43) and (44): (43) eutita-­ke (44) ustse-­ke=ma approx how-­ approx=intg now-­ ‘a little time ago’ (D:8) ‘how, approximately?’, ‘how, more or less?’ (D:40) 2.9  Reduplication with attenuative, celerative and other meanings There exists a reduplicative process which, like the reduplicative nominaliser mentioned in connection with examples (10)‒(11) above, copies the (second) consonant of the onset and the (first) vowel of the nucleus of syllable to which it is joined. However, unlike the reduplicative nominaliser, it is placed after the first syllable of the polysyllabic stems with which it occurs. It can be observed in our corpus as a suffix •CV after monosyllabic stems composed of a consonant and a vowel ((47), (48) and (51)) and bisyllabic stems ending in a CV syllable ((56) and (57b)); as an infix [•CV] within stems consisting in a single ­syllable provided with onset and coda (45), or whose nucleus coincides with a diphthong (49); or within stems formed of at least three syllables the second of which is CV ((58)‒ (60)); and within other polysyllabic stems following a syllable with an onset ((46), (50) and (52)‒(55)). This reduplication is a polysemic process which has five evaluative meanings connected to the aspectual classes of the base verbs: • attenuative, when applied to state verbs (45)–(48) and activity verbs denoting events which can be carried out with more or less intensity (49)–(50): (45) lu[•lu]ps-­ (46) ake[•ke]l-­ be.hot[•att] be.hot[•att] ‘be lukewarm’ ‘be lukewarm’ (47) p’o•p’o-­ (48) si•si-­ be.white/be.dry•att for.there.to.be•att ‘be whitish, be almost dry’ (D:5, 76, 82) ‘be there in a small measure’ (G:61, 73)

622

Raoul Zamponi and Willem J. de Reuse

(49) ka[•ka]i-­ (50) olo[•lo]m-­ eat[•att] speak[•att] ‘eat a little’ (G:65) ‘speak a little’ (G:65) • celerative, when applied to activity verbs referring to activities that can be carried out with more or less speed: (51) ne•ne-­ (52) come•cel ‘come quickly’ (D:132)

yep-­ku[•ku]i-­ ip-­move[•cel] ‘move quickly’ (D:94)

• iterative (or multiplicative), with verbs of achievement: (53) oko[•ko]ts-­ (54) tako[•ko]ts-­ embrace[•iter] lean.out[•iter] ‘lean out many times’ (D:21) ‘embrace many times’ (D:2) • incessative, with activity verbs denoting events that can be carried out for an indeterminate length of time: (55) takyu[•yu]Ɂu-­ look.for[•incess] ‘look for incessantly’ (D:29) • evolutive, with, apparently, incrementative verbs: (56) wato•to-­ recover/live•evol ‘convalesce’ (D:41) The various meanings expressed by this reduplicative process can be regarded as forming a continuum. This continuum is attenuative at one end and celerative at the other, and can be represented as follows: ‘attenuative’ (‘a bit’) > ‘evolutive’ (‘a few/ little at a time’) > ‘iterative’ (‘several times’) > ‘incessative’ (‘incessantly’) > ‘celerative’ (‘quickly’). This reduplication is also used with relational nouns (denoting spatial relations) and adverbs. The meaning conveyed is attenuative with relational nouns (57b), to be compared with (57a) and with the time adverb eutilem ‘before’ (58). The meaning is iterative with the time adverbs epile ‘afterwards’ (59) and tsiketole ‘when?’ (60): (57) a. moke=le le distant.location=loc foc ‘I will go far’ (G:77) b. moke•ke=ma le distant.location•att=loc foc ‘I will go a little way’ (G:77)

ka-­n-­s go-­fut-­1sg ka-­n-­s go-­fut-­1sg

Lule 623 (58) euti[•ti]lem (59) epi[•pi]le before[•att] afterwards[•iter] ‘a little before’ (G:83) ‘from time to time’ (G:84, V:85) (60) tsike[•ke]tole=ma when[•iter]=intg ‘how often? (G:82)

3 Conclusion In this final section we will deal with two issues: the formal devices used in Lule evaluative morphology and the question of the relationship between evaluative morphology and word classes. We have seen that Lule evaluative morphology uses two formal strategies: suffixation and reduplication. As far as suffixes are concerned, we described one celerative suffix (Section 2.1), one ameliorative suffix (Section 2.2), one pejorative suffix (Section 2.3), one derogatory suffix (Section 2.4), one augmentative suffix (Section 2.5), two intensive suffixes (Sections 2.6 and 2.7) and one approximative suffix (Section 2.8). This is quite an impressive array of evaluative suffixes, but as indicated in note 3, it is not impossible that the ameliorative, the pejorative and the augmentative should be analysed as separate words rather than as suffixes. If one chose to do so, the number of suffixes would be reduced to one celerative suffix, one derogatory suffix, two intensive suffixes and one approximative suffix, which is a less impressive system. As far as reduplication is concerned, we have described it as a polysemic process with five evaluative meanings: attenuative, celerative, iterative, incessative and evolutive (Section 2.9.). Interestingly, some of the evaluative suffixes also have at least one non-­evaluative meaning. Added to a kinship term, the ameliorative, augmentative and one intensive suffix (-­ike) indicate that the referent of base noun is a blood relative (13), (21), (22) and (39). The ameliorative may form with state and activity verbs derivates that express the meaning ‘be good to v’ or ‘worthy of v-­ing’ (6)‒(11). Finally, it is also noteworthy that the derogatory suffix acts as a deverbal nominaliser in (18) and (19). We now turn to the relationship between Lule evaluative morphology and the word classes it attaches to. The celerative attaches to verbs. The ameliorative, the pejorative and the derogatory attach to nouns and verbs. The augmentative attaches to nouns, state verbs and one adverb. One intensive suffix (-ike) attaches to state verbs, one adverb, one noun plus an element that could be a further noun or an activity verb. The second intensive suffix (-ewi) attaches to state verbs. The approximative attaches to adverbs. The evaluative reduplication is found with verbs of different aspectual classes plus nouns and adverbs. Because of the small number of attestations of each sort of evaluative morpheme, it is hard to draw definite conclusions about the relationship between evaluative morphology and word classes. It does appear that evaluative morphology favours verbs and particularly state verbs more than nouns. It is also more common with adverbs than with nouns.

Notes 1. A language that at present has only one or two surviving semi-­speakers (Golluscio and González 2008).

624

Raoul Zamponi and Willem J. de Reuse

2. The remaining documentation of Lule consists of a short vocabulary by the Catalan Jesuit ­missionary José Ferragut (Gilij 1782, 364–6), a text of the Lord’s Prayer (Hervás y Panduro 1787, 102–3), not identical to that in Machoni (1732, Catecismo part, 1) and a few words in eighteenth-­ century Jesuit reports. 3. In our philological analysis of Maccioni’s grammar, the issue of word boundaries needs to be addressed. Maccioni’s written word spaces are one criterion by which we can decide whether he considered a form to be one or more words. Unfortunately, this criterion is weakened by the fact that Maccioni is not very consistent in his writing of word spaces, and in such cases it is difficult to determine whether a form consists of one or of several phonological words. This issue is of course relevant in determining which evaluative strategies are part of morphology, and which are separate words and therefore not relevant to a description of evaluative morphology. As a rule of thumb, we considered those evaluative elements which were almost always written together with the preceding word or stem to be suffixes (that is morphology), and we considered those evaluative elements which were almost always written separate from the preceding word or stem to be separate words (that is, relevant to the syntax rather than to the morphology). This rule of thumb is not very satisfactory, as shown by the qualification ‘almost always’.   The presumed affixal forms with the suffix written separately are the originals in (5) aquém écy, (6) caip eci, (16) ecquys eyú and (21) pepè yqueps. 4. Examples from the grammar portion of Maccioni’s Arte will be indicated by the abbreviation G, in parentheses, after their translation, while those from the Spanish–Lule dictionary portion or Vocabulario are indicated by the abbreviation D. The abbreviations G and D will then be followed by the number of the page in which the example occurs. Note that D is paginated separately from G. 5. Example (17) is to be compared to Italian donnaccia ‘prostitute’, from donna ‘woman’.

20.4 Toba

Paola Cúneo

1 Introduction The Toba language (qom lʔaqtaqa)1 belongs to the Guaicuruan linguistic family, that is, the same family as Pilagá, Mocoví and Kadiwéu. The Toba group is made up of approximately 70,000 people,2 most of whom dwell in the Gran Chaco area, in the provinces of Chaco, Formosa and Salta (Argentina), south-­eastern Bolivia, and a region close to El Cerrito district (Paraguay). As a result of migration, the Toba people also dwell in permanent settlements close to big cities in Argentina. In addition, there are dialectal differences (mainly phonological and lexical) that speakers themselves are aware of. In the provinces of Chaco and Formosa, Messineo (1991) identifies four major ethno-­dialectal areas, which are partly consistent with factions self-­identified as groups: Dapigemlʔek (Northwest), NoʔolgaGanaq (Mid North), LʔañaGashek (Mid South) and Takshek (Southeast). From a sociolinguistic perspective, Toba is being replaced by Spanish in certain communicational environments, particularly in urban contexts, and also because speakers are socially and economically marginalised. The following are some of the typological characteristics of the Toba language relevant for the present chapter. Firstly, Toba is characterised by morphological complexity, both in nouns and in verbs. It has a tendency towards polysynthesis and agglutination, and it shows a rich derivational morphology. This language makes a distinction between alienable and inalienable possessed nouns, and attributive possession is encoded following a head-­marking pattern. Additionally, nouns are inflected for number and gender, and noun gender is an inherent lexical feature. Compound nouns give rise to phonologically, morphologically and semantically autonomous units. Verb morphology shows the features of an active–inactive language, involving three sets of dependent pronominal markers: active, middle and inactive (Messineo 2003). Toba verbs do not inflect for tense or mood but show aspectual distinctions as suffixes expressing direction and position. While this language is characterised by the absence of a defined lexical category of adjective, it is possible to identify two ‘intermediate’ categories between nouns and verbs, encoding the qualities or attributes of their referents (Messineo 2003, 112–13). As regards the order of constituents, the basic, though flexible, word order is AVO and VS. Like the rest of the Guaicuruan languages, Toba has a closed system of six demonstrative classifiers placed before nouns and combining features of configuration (form and position) with deixis (proximity, motion and absence) (Klein 1979; Messineo 2003). In Toba, evaluative morphology (EM) involving diminution and augmentation can perform both descriptive (quantitative) and qualitative functions, according to definitions

Paola Cúneo

626

posed from a cross-­linguistic perspective (cf. Grandi 2005; Grandi and Körtvélyessy, this volume). Highly productive evaluative morphemes are the following: -­ole(k), -­aeʔ and -­lyale(k), for diminution, and -­(l)o:m, -­naq/-­naGa, -­(d)ay(k) and -­ltaʔa/-­lateʔe, for augmentation. They can be applied to nouns, pronouns, determiners and, to a lesser extent, verbs and adverbs. Moreover, evaluative derivation in Toba is subject to lexicalisation and can create new words. Previous papers comprise a comprehensive analysis of EM in Toba (Cúneo 2013b; 2014).3 Toba data come mostly from the Dapigemlʔek dialect, and were personally collected doing fieldwork (2002 to date) with speakers dwelling in the province of Chaco and the district of Derqui (Buenos Aires), in Argentina. Data from other dialectal areas are specified in the examples.

2  Evaluative morphology 2.1  Physical and spatial dimension (size): -­ole(k) and -­aeʔ dim In Toba, small size is expressed by means of the diminutive suffix -­ole(k) attached, most frequently and productively, to nominal bases. However, this suffix can also be attached to personal pronouns, ‘adjectives’ and even, though exceptionally, verbs and adverbs. In contrast, the suffix -­taeʔ is typically attached to demonstrative classifiers. The -­ole (f)/-­olek (m) suffix makes it possible to express physical dimension (small size) in animate and inanimate nouns, including Spanish loanwords (2) and compound nouns (3): (1)

n-­qayk-­ole-­k

pos.indf-­head-­dim-­m

‘little head’

(2) kaay-­ole-­k horse-­dim-­m ‘small or short horse’ (3) kizoq-­l-­tel-­ole tiger-­poss3-­ear-­dim ‘little butterfly’ (lit.: ‘tiger ear-­dim’) As shown above, nouns made diminutive through -­ole(k) are inflected for gender and the marked value is the masculine (-­k). In these cases, gender is transparent to the base gender. As regards number, the form of the diminutive plural masculine suffix is -­olqa,4 while the feminine form remains unchanged (-­ole). The diminutive -­ole(k) is also appended to derivative attributive nouns (4) and descriptive verbs (5).5 It is attached too, although in a limited way, to adverbs of time in order to express a short period of time (6), and to a lesser extent, to verbal bases (7). Examples (4)–(7) illustrate the ubiquity and malleability of this diminutive suffix: (4) choGo-­day-­k-­ole-­k receive.compassion-­attr-­m-­dim-­m ‘poor . . .’ (expression of pity)

Toba 627 (5) tokchi-­gi-­ole a-­na qadol red/raw-­loc-­dim f-­dem catfish ‘the catfish [is] a little raw’ (= the fish is small) (6) kadaʔak-­ole-­k ñi-­mateec for.a.while-­dim-­m 1-­rest ‘I’m going to rest for a (little) while’ (7) koʔollaGa seʔeso ʔetta-­igi-­ole-­k adv dem 3.light.up-­loc-­dim-­m ‘then that [the star] was lit a little’ (mythical story on the morning star, Takshek variety, Eastern Chaco and Formosa) In contrast, the suffix -­taeʔ is attached to demonstrative classifiers (and derived forms) functioning as determiners. While the demonstrative bears the marking, the diminutive is interpreted as applying to the head of the noun phrase, that is, the noun (cf. ‘transference of diminutivisation’; Bauer 1997, 554–7):6 (8)

zi-­taeʔ qaʔem dem:hor-­dim lake ‘that little lake’

(9)

dem:prox-­dim

neʔena-­taeʔ

‘your little house’

ʔad-­ma

poss2-­house

Example (10) shows joint and contextualised occurrence of both types of suffixes described above: (10)

so-­m dem:dist-­foc

yaʔaxayki-­ole-­k old-­dim-­m

y-­aGaa-­tak 3a-­call-­prog



i-­yaGa-­ta-­ike seʔeso-­taeʔ pyoq l-­ʔenaxat Boca 3-­call-­dur-­desid dem:dist-­dim dog poss3-­name Boca ‘the little elderly man was calling that little dog called Boca’

As shown in (10), evaluative suffixes do not trigger agreement with other elements in the syntactic construction: the use of a diminutive in the head or modifiers of the noun phrase does not involve the use of the diminutive in other elements. 2.2 Individuation: -­ole(k) dim Diminutivisation through the suffix -­ole(k) may function as an individuation mechanism. When dealing with abstract or mass nouns, the suffix may alter the semantics of the base and change the noun into a concrete or count noun. For example, an abstract (11a) or mass (12a) noun (see also (21) below) may result in a concrete or count noun following application of the diminutive (11b) and (12b).

Paola Cúneo

628 (11) a. qom l-­ʔaqtak Toba poss3-­word/language ‘Toba language’ b. qom l-­ʔaqtak-­ole-­k Toba poss3-­word/language-­dim-­m ‘Toba little word’ (= with few letters) (12) a. zi toGo dem:hor sand ‘the sand’ b. zi toG-­ole-­k dem:hor sand-­dim-­m ‘the little sandy/sand field’ 2.3  Age variation -­lyale(k)

Lexical forms l-­yale-­k ‘poss3-­son-­m’/l-­yale ‘poss3-­daughter.f’, which function as independent lexemes in this language, can also encode age variation: ‘young X’. They are the preferred resource when forming compound nouns referring to animals, and they convey the meaning ‘offspring’: (13) a. ketaq ‘goat’ b. ketaq-­lyale/lyalek ‘young she-­goat/young he-­goat’ In (13), grammatical gender is semantically assigned on the basis of biological gender, as is usually the case with names referring to mammals. 2.4  Intensification: -­(l)o:m The -­(l)o:m suffix has an emphatic and expressive function. This suffix is attached both to adverbs expressing time and to nominal bases: qome-­lo:m

(14)

adv:later-­int

(15)

saishet da d-­oɁoche-­ta naɁaGa-­lo:m neg comp 3-­sleep-­dur day-­int ‘he cannot sleep all day long!’

‘some time later, much later’

(16) qaɁañ-­ol-­o:m young.woman-­dim-­int ‘very young woman’ Example (16) shows that evaluative suffixes with different meanings (dim, aug) may be successively applied. Interaction among different evaluative suffixes is also observed in other very common words referring to culturally prominent social relations, such as ‘young man or woman’ or ‘elderly man or woman’. In these cases, the augmentative -­o:m

Toba 629 involves an intensification value and is applied externally with respect to diminutive suffix. However, it is hard to determine whether these sequences are productive or whether the last suffix is merely attached to a lexicalised base. For example: (17) yaʔaGayki-­ol-­o:m old-­dim-­aug ‘extremely old, very old’ 2.5  Affective dimension: endearment and contempt Apart from the physical and spatial dimensions, the -­ole(k) suffix also expresses affective values, and it may occasionally be difficult to distinguish the descriptive use (small size) form the qualitative function (expression of affection):7 (18) laʔ ya-­qay-­ole-­k greeting poss1-­brother-­dim-­m ‘hello, my (little) brother’ (affective; the brother is not necessarily young or small) When the diminutive is attached to personal pronouns, it has a prominently expressive function (in these cases, interpretation as physical dimension is not acceptable). Gender in the diminutive is referential in this case, based on the biological sex of the speakers. For example: (19)

ayem-­ole:: pro1-­dim ‘poor me!’ (*‘little/small I’)

(20)

ʔam-­ole-­k pro2-­dim-­m ‘poor you!’ (*‘little/small you’)

According to this affective function, the diminutive category allows for configuration of certain speech acts, such as greeting (18) and expressing pity (19) – (20), as well as requests (21) or exclamations of surprise (22): (21) soroʔ-­ole woʔo ka anarin-­ole-­k aunt-­dim ex dem flour-­dim-­m ‘aunty, do you have a little flour?’ (22)

ahhh .  .  . y-­ap-­ole::-­k interj poss1-­grandfather-­dim-­m ‘ahhh . . . my grandpa’ (when the speaker has not seen him for a long time)

In addition, the diminutive may mark or create discursive contexts, particularly in expressive genres making the speaker/performer visible. This happens in advice (nqataGak; see Messineo 2003, 198) (23), or songs (lʔonek) (24), significantly reinforcing the personal and affective component. Examples (23) and (24) are adapted from Messineo and Dell’Arciprete (2005, 56, 59).

Paola Cúneo

630 (23)

yal-­ole ʔan-­lo-­shigem poss1-­son-­dim 2-­look-­dir:up ‘little daughter, pay attention’

(24) ekoʔ nala dapigem-­ole-­k seems.like fruit north-­ dim-­m ‘it seems like [I’m chewing] fruit in the little north’ (said with affection and nostalgia, when remembering the place of origin) As regards the expression of contempt, the -­naq (m)/-­naGa (f)8 and -­day(k) suffixes are often used to connote pejorative meanings or jokes, and even insults.9 They are attached to noun bases denoting body parts in order to form attributive nouns. The gender of the suffix corresponds to the gender of the (human) referent, as shown in (25). Attributive nouns formed by this suffix lose the inalienable possession marker. (25) a. la-­piaʔ poss3-­foot ‘his or her foot’ b. piaʔ-­naq / piaʔ-­naG-­a foot-­aug.m foot-­aug-­f ‘big-­footed man’ ‘big-­footed woman’ c. piaʔa-­day-­k / piaʔa-­day foot-­attr-­m foot-­attr.f ‘big-­footed man’ ‘big-­footed woman’ In (25)b, c), an evaluation on the part of the speaker is conveyed. The -­day(k) suffix may also be attached to verbal bases and can highlight certain actions or habits which the culture considers to be negative. In (26), ʔochaGay(k) ‘sleepyhead’ means a person who sleeps all day (maybe ‘damaged’ as a result of shamanism or cultural taboo). (26) ʔoch-­aG-­ay sleep-­nmls-­attr ‘sleepyhead woman’ 2.6  Approximation and non-­prototypicality: -­ltaʔa/-­lateʔe and -­lyale(k) Finally, we will explore the function of kinship terms as approximatives. The terms l-­taʔa poss3-­ father ‘his or her father’ and l-­ateʔe poss3-­mother ‘his or her mother’ encode notions of augmentation and similarity, resemblance or approximation, when involved in the formation of compound nouns – especially useful for ethnozoological vocabulary. In addition to approximation regarding size (27), terms of kinship form nouns referring to species dangerous to man – such as the coral snake, which is poisonous and lethal (28) (example from Messineo and Cúneo 2011, 153) – and referring to species which have recently been introduced into the culture (29): (27)

l-­qona-­lateʔe

poss3-­toe-­aug.f

‘his or her big toe’

Toba 631 (28) wizik-­lateʔe (fake).coral.snake-­aug.f ‘(real) coral snake’ (Micrurus pyrrhocryptus) (29) teʔesaq-­ltaʔa bee-­aug.m ‘domestic (queen) bee’ (Apis mellifera, European type) The function of terms of kinship is to categorise an entity in association with another entity which is conceptualised as prototypical or more widely known. The compound noun formed by these terms may share similar shape features with the prototype (referred to by the base) and be slightly different from it in certain aspects (for example, behaviour or habitat in the case of animals). Apart from this notion of similarity, there is the idea of something ‘extra’ resulting from the augmentative notion (extra size, extra power, extra colour, extra danger; and extra-­exotic, rare or unknown). In this sense, these terms combine approximation (or similarity in relation to a prototype) with augmentation.10 Likewise, the meaning of the term of kinship -­lyale(k) ‘his or her son or daughter’ mentioned above can be extended to ‘a small type of’ or ‘smaller-­sized species’ for ethnobiological names: (30) a. ʔeleʔ ‘parrot’ b. ʔeleʔ-­lyale ‘Amazon parrot’ (31) a. qadol b. qadol-­lyale

‘catfish’ (Pimelodus sp.) ‘type of fish’ (undetermined)

The -­lyale(k) proximative function contrasts with the expression of age variation (meaning ‘offspring’) when assigning grammatical gender. In this case (see (13)), gender is assigned on the basis of biological gender, while in examples (30)–(31) the diminutive lexeme maintains the gender of the base.

3 Conclusion Forms documented in Toba involve two major semantic functions recognised for EM in languages: descriptive and qualitative evaluation. Table 20.4.1 summarises the semantic-­ functional domains of EM in Toba. Some diminutive and augmentative suffixes are ­multifunctional, as they may convey more than one evaluative meaning. Concerning the bases they are attached to, evaluative forms in Toba are most commonly and productively applied to noun bases. However, they can also be applied to nouns, attributive nouns and verbs, pronouns, determiners and, less frequently, verbs and adverbs. Table 20.4.2 shows evaluative forms in Toba, based on Bauer’s hierarchy (1997).

Paola Cúneo

632 Table 20.4.1. Evaluative meanings in Toba Dimension

Form and meaning

Physical and spatial dimension (size)

-­ole(k) and -­aeʔ ‘physically, spatially, bodily small X’ -­lyale(k) ‘young X’, ‘offspring’ -­(l)o:m ‘very X’ -­ole(k) ‘a small piece of X’ (partitive) -­ltaɁa/-­lateɁe and -­alo ‘similar to X but with a certain attribute to a high degree’ -­lyale(k) ‘a small type of X’

Temporal dimension:   Age variation   Intensification Individuation Approximation and non-­prototypicality

-­ole(k) ‘dear X’ -­naq/-­naGa and -­day(k) ‘pejorative’, ‘one who is/has/makes X to a high degree’

Affective dimension:   Endearment and contempt

Table 20.4.2. Base types for evaluative morphology in Toba Noun

.

-­ole(k)

Adjective, Verb -­ole(k)

-­lyale(k) -­(l)o:m -­naq/-­naGa -­day(k) -­ltaʔa –lateʔe

.

Adverb, Pronoun

Determiner

.

-­ole(k)

-­taeʔ

-­(l)o:m -­day(k)

Notes  1. The phoneme (consonant) inventory in the Toba language and orthographic graphemes [in square brackets] used are provided below (based on Messineo 2003, 36). Labial Alveolar Plosive Fricative Nasal Tap Lateral Glide

p m w

t s

d n ɾ [r/d] l

Palatal tʃ [ch] ʃ [sh]

Velar ʒ [z] ɲ [ñ]

k

g

Uvular q

ɢ

Glottal ʔ

ʎ [ll] y

  2. According to INDEC (2004–5), only about 58 per cent of the people who claim to be first generation members and/or descendants of the Toba group can speak and/or understand the indigenous language.

Toba 633   3. Research on this subject was part of a postdoctoral scholarship (2012–14) by the CONICET (National Scientific and Technical Research Council), titled Morfología evaluativa: aspectos morfosintácticos, semánticos y discursivos de los aumentativos y los diminutivos en toba (guaycurú), directed by Dr Cristina Messineo.  4. -­qa is added as expected for noun bases marked by the suffix -­k, indicating masculine gender; cf. Carpio (2007).   5. According to Messineo (2003, 112–13), the two intermediate categories between nouns and verbs are attributive nouns and descriptive verbs. The former, like all nouns, inflect for gender and number, although unlike nouns, they are not marked for the possessor. They may act as noun modifiers, heads of noun phrases, and predicatives. In contrast, descriptive verbs admit verbal suffixes (durative, locative, directional) but do not carry personal prefixes. For the predicative function, they occupy the canonical position preceding the noun phrase (according to the VS order for intransitives).   6. See also examples of the diminutive as applied to descriptive verbs, such as tokchigiole ‘a little raw’ (= the little catfish is raw).   7. According to our data, the affective interpretation seems to be the privileged function in a dialectal area of the north-­western region. The high frequency of use of diminutive forms characterises – according to the speakers themselves – the group’s ‘way of speaking’. However, a dialectal study is required to shed light on this matter.   8. To a lesser extent, we have documented the -­saq (m) and -­saGa (f) allomorphs with the same value.   9. Although the -­naq/-­naGa suffix appears to have a more clearly pejorative meaning than -­(d) ay(k), the semantic or pragmatic contrast between the two suffixes is still unclear. 10. A similar function is performed by the synchronically unproductive suffix -­alo (see Cúneo 2013a).

20.5 Wichi

Verónica Nercesian

1 Introduction Wichi belongs to the Mataguayan (also known as Matacoan) language family, together with Chorote, Maka and Nivacle, and is spoken in the lowlands of the Gran Chaco region1 in South America. The geographical distribution of the Wichi language comprises two countries: Argentina (in the provinces of Formosa, Salta and Chaco along the Bermejo and Pilcomayo rivers) and Bolivia (along the banks of the Pilcomayo river from Yacuiba in the south to Villamontes in the north, in the Tarija department). There are two wide diatopic (geographical) varieties of Wichi called pilcomayeño and bermejeño, according to whether the speakers live by the Pilcomayo or the Bermejo rivers. In addition, two sub-­varieties have been identified within each of the two above-­mentioned dialects, depending on whether the speakers live upstream or downstream (Gerzenstein 1992; Nercesian 2011). The number of native speakers is estimated at 28,676 (INDEC 2004/5) in Argentina and 4,115 (ORCAWETA 2011, 18) in Bolivia, amounting to approximately 32,790. From a typological point of view, Wichi is a head-­marking, agglutinating, polysynthetic language. Arguments are indexed on the verb following a nominative–accusative alignment in the first and second person and a neutral alignment in the third person (zero for both subject and object). Nominal phrases are optional, but when they are overt, the basic word order is SVO. Wichi is not a case-­marked language and does not have adpositions. It has applicatives, causatives, and noun incorporation allowing possessor raising and classificatory incorporation. The major word classes are nouns, verbs and adverbs. Nouns are divided into alienable and inalienable classes; there is no gender. Properties, attributes and states are expressed by a group of intransitive verbs. Minor classes are pronouns, interrogative particles, illocutionary particles, conjunctions, quantifiers and numerals. In Wichi, word classes are flexible. Nouns and verbs share some grammatical categories such as tense and iterative aspect, pronominal subject prefixes, negative affixes and evaluative morphology (EM).

2  Evaluative morphology From a cross-­linguistic point of view, EM can basically express both descriptive (quantitative) and qualitative evaluation (Wierzbicka 1980; Stump 1993; Grandi 2005; Grandi and Körtvélyessy, this volume; Körtvélyessy 2012a). In Wichi, evaluative constructions perform both functions and use three highly productive suffixes: -­fwaj diminutive, -­les

Wichi 635 diminutive class and -­taj augmentative. They apply to nouns, verbs and some adverbs and numerals. Augmentative and diminutive derivation do not change the word class and can even create new words referring to entities that are completely different from those referred to by the base word. Evaluative constructions follow the same stress pattern as basic, derived and compounded words; there are no special phonological changes. As to their position in the morphological structure, evaluative suffixes appear after derivatives (1)–(2) but before inflectional morphology (3)–(4) and demonstrative suffixes (5): (1)

to-­tichunha-­yaj-­taj indf.poss-­think-­nmls-­aug ‘big thinking’, ‘great idea’

(2) a-­lon-­ek-­fwaj 2poss-­kill-­nmls-­dim ‘my little killing’ (it is little because of the size of the animal that was killed) (3) afwenche-­fwa(j)-­s bird-­ dim-­pl ‘little birds’ (4) afwenche-­ta(j)-­s bird-­ aug-­pl ‘big birds’ (5) atsinha-­fwaj=tsu woman-­ dim=dem.all ‘that little woman (who is going away)’ Nouns derived through EM pluralise in the same way as nouns ending in /-­χ/, spelled j; that is, by replacing the uvular fricative with the plural suffix /-­s/, spelled s. 2.1  Descriptive perspective Descriptive evaluation is based on the real characteristics and objective properties of an entity and expressed by attaching the augmentative -­taj and diminutives -­fwaj and -­les. The same suffixes are used for nouns that refer to animate, inanimate, human or non-­human entities (6)–(11). There are no semantic limitations for the use of evaluative suffixes depending on the semantic field of the base word: (6) asinoj-­fwaj dog-­ dim ‘small dog’ (7) asinoj-­taj dog-­ aug ‘big dog’

Verónica Nercesian

636 (8) hin’u-­fwaj man-­ dim ‘small man’ (9) hin’u-­taj man-­ aug ‘big man’ (10) tente-­fwaj dim stone-­ ‘small stone’ (11) tente-­taj stone-­ aug ‘big stone’

Diminutive and augmentative constructions evaluate the size of an item according to a culturally established standard size, even if the noun refers to an animate entity. When the augmentative suffix is combined, for example, with the word lhos ‘son’, it evaluates the size of the person, but not necessarily his age. In examples (8) and (9), the diminutive and the augmentative constructions evaluate the size of the ‘man’ quantitatively. Likewise, in example (12), the ‘woman’s son’ which the speaker refers to is a child. Thus, the augmentative evaluates the size of the son according to the standard size of children. Nonetheless, since the size of children is generally age-­dependent, the evaluative may correspond to and indicate age variation. A ‘big child’ may refer to a child becoming a ‘young boy’.2 (12) che ya-­w’en to-­w’ey wemek-­fwaj o ropa usada conj.irr 1sbj.pl.excl-­have indf.poss-­clothes 3sbj:be.old-­dim or used clothes desarmá to-­yen-­to-­kuset-­lhos-­(t)aj-­a.  .  . 3sbj:unpick indf.sbj-­make-­indf.poss-­pants-­son-­aug-­ic ‘if we have old or second-­hand clothes we unstitch them to make big-­son-­pants’ (explanation about the sewing workshop, by EG3) Similarly, in example (13), the augmentative evaluates the size of the baby iguanas as compared to the standard size: (13)

T’eynlo-­k tim alhe-­les-­(t)a(j)-­s. 3sbj:want-­desid 3sbj:swallow iguana-­babies-­aug-­pl ‘He [Takfwaj] wants to swallow the big baby iguanas.’ (mythical story, by AG)

Human age variation is lexically expressed as follows: han’ofwaj ‘baby’ or ‘child’, mamse ‘young boy’, lhetsha ‘young girl’, hin’u ‘(adult) man’ and atsinha ‘(adult) woman’. Conversely, age variation among animals is formed by combining the word lhos ‘son’ or the root les ‘offspring’ with the base word that refers to the animal. The meaning of lhos and les in the compound is extended to the generic sense of ‘baby or cub’, shown in (14)–(15). The inalienable noun is added to the right of the base following the same word order that genitive phrases show: possessor–possessed. The compounding with les can also be applied to words naming plants to refer to ‘sprouts’ (16):

Wichi 637 (14)

tsun’ataj lhos sheep 3poss:baby ‘lamb’, ‘young sheep’

(15)

kaila-­les goat-­babies ‘baby goats’





(16)

hal’o-­les plant/tree-­babies ‘sprout’, ‘bud’

There is an interesting semantic and functional extension of les ‘offspring’ towards ‘small size’ in combination with nouns referring to non-­animate entities (17)–(19). The diminutive meaning of the compound differs from the one expressed by derivation through -­fwaj, since the former refers to the whole class/group of that entity, while the latter refers to the size of a single item within the class. In example (17), nisojles means ‘class of shoes that are small’; in (18), noyhoyles is ‘class of ropes that are thin’; and tenteles means ‘class of stones that are small’ in (19). Note that the same word can be both derived through -­fwaj (10) and combined with les (19): (17) nisoj-­les shoe-­ dim.cl ‘small shoes’ (18) noyhoy-­les rope:pl-­dim.cl ‘thin ropes’ (19) tente-­les stone-­ dim.cl ‘small stones’ That the semantic and functional extension was to express ‘the small class/group of something’ could explain the use of the plural form of the word lhos ‘son’ with a generic sense. Moreover, as a category, ‘son’ refers to the younger generation, which is smaller in age than other members of the same class that are taken as a reference, that is, ‘parents’. The strategy of using the word ‘son’ or ‘offspring’ to express evaluation seems to be widespread within the family. Apart from Wichi, Chorote and Nivacle also use this formally similar inalienable word for evaluative purposes (Chorote: -­les/-­lis; Nivacle: -­xɬas). Lastly, diminutive and augmentative suffixes also create new words that name a different entity from the one named by the base. Furthermore, the ‘small’/‘big’ size evaluation is inherent to the meaning of the word. Examples (20)–(21) are items created from an existent word, whereas (22)–(24) come from a root that does not exist on its own: (20) ch’ute-­taj ear-­ aug ‘cup’

Verónica Nercesian

638 (21) yel’a-­taj tapir-­ aug ‘horse’ (22)

han’ofwaj ‘baby’, ‘child’

(23)

kalaktaj ‘white-­necked heron’

(24)

kow-­kowtaj ‘dark-­billed cuckoo’



Examples (23) and (24) are ideophones: (23) combines onomatopoeia and augmentative derivation, and (24) combines onomatopeia, reduplication and augmentative derivation. In addition, there are complex bases formed by the comparative adverb hote ‘like’ and the diminutive -­fwaj to create cardinal numbers (25), which traditionally count up to five. Higher numbers are expressed by the quantifier nitokw ‘many’: (25) hotefwajhi ‘one’ hotefwasa ‘two’ hotefwasi ‘three’ Synchronically, these forms have been replaced by the Spanish numeral system and do not constitute a clear-­cut word class. Given that these numerals are obsolete and outdated in the language, it is quite difficult to break them into morphemes. However, we can recognise the comparative adverb in combination with the diminutive suffix in the singular form for number ‘one’ and in the plural for numbers ‘two’ and ‘three’. 2.2  Qualitative perspective Wichi resorts to the same suffixes -­fwaj dim and -­ taj aug to express qualitative evaluation, that is, the subjective evaluation of an item on the basis of the speaker’s personal feelings or opinions and on the context. From a qualitative perspective, the diminutive can express attenuation, contempt and endearment, and the augmentative, intensification. 2.2.1  Attenuation/Intensification Actions and properties expressed by verbs can be attenuated or intensified through the diminutive and the augmentative, respectively. The meaning conveyed by evaluative suffixes deriving verbs is close to that prototypically conveyed by adverbs. They can attenuate the speed of a movement (26) or the intentionality and volition of the agent (27): (26) Asinoj yik-­fwaj. dog 3sbj:go.away-­dim ‘The dog goes/went away slowly.’

Wichi 639 (27)

N’-­t’ischey-­fwaj. 1sbj-­laugh-­dim ‘I laugh/laughed unintentionally.’

The diminutive deriving an action verb can express little/lack of intentionality (27) or little/lack of volition, indecision or unwillingness (28), rather than reduced speed (26). In that case, slower speed of movement would be a consequence of lack of volition. In (28), this manner of walking could be a pejorative way of referring to the agent in some contexts, since the word means that the person walks in a way that arouses pity or compassion: (28)

W’ilek-­fwaj. 3sbj:walk-­dim ‘She or he walks in a way that arouses pity or compassion’.

When the diminutive combines with state verbs, it is the intensity of the property/state denoted by the verb that is attenuated (29)–(32): (29) nech’e to-­tchema to-­w’ey toj la-­wemek-­fwaj just 1sbj.pl.incl-­catch indf.poss-­clothes conj 3sbj:be.old-­dim ‘we just gather little old clothes’ (explanation about the sewing workshop, by EG) (30)

nom=pej la-­walay 3sbj:come=iter 3poss-­axe ‘new axes kept coming’

nech’a yik new

lhetek ichot-­fwaj 3poss:head 3sbj:be.red-­dim ‘and the [Takfwaj’s] head was reddish’ wit ifwaj=hi Takfwaj wej conj 3sbj:ax=loc Takfwaj tail ‘because they axed his tail’ (mythical story, by AG) (31) akoj t’i ha-­t’en-­hi-­fwaj t’i 3sbj:be.tasty liquid neg.r-­hard/thick-­neg.r-­dim liquid ‘the bee’s liquid (the wax) is very tasty and not very thick’ (explanation about gathering rubito honey, by AG) (32) toj iwuytsu ha-­ts’in-­hi-­fwaj wit ileche toj fwitsaj conj thus neg.r-­be.rowdy-­neg.r-­dim conj 3sbj:stop conj 3sbj:be.aggressive ‘thus, [the bees] become less rowdy and stop being aggressive’ (explanation about gathering moro-­moro honey, by AM) The diminutive attenuates the intensity of the property attributed to the single argument, tow’ey ‘clothes’ (29), and lhetek ‘his or her head’ (30). Examples (31)–(32) illustrate the negative form of verbs derived through the diminutive. The negative form of t’en ‘to be hard/thick’ means the opposite, ‘to be soft, flexible/fluid, diluted’, whereas the negative form of the evaluative construction in (31) means ‘not very hard/thick’, meaning that it is gradually less ‘hard/thick’ than ‘almost hard/thick’ but ‘harder/thicker’ than ‘soft/fluid’.

Verónica Nercesian

640

Likewise, while the negative form of ts’in ‘to be rowdy’ means the opposite, ‘to be quiet, calm’, the negative form of the evaluative construction hats’inhifwaj in (32) means ‘not very rowdy’, ‘almost quiet and calm’. Thus, the negative form does not neutralise the semantic function of attenuation of the diminutive but downgrades it without eliminating the attribute completely. Although less frequent than the use of the diminutive to express attenuation, the augmentative is used with the semantic function of intensification, as (33) illustrates: (33)

to-­fwtaj is lawut indf.sbj-­say 3sbj:be.good beehive.hole ‘it is said that they are good beehive holes’

lawut wefw-­taj wit pitaj-­taj beehive.hole 3sbj:be.big-­aug conj 3sbj:be.long-­aug ‘they are very big and long’ (explanation about gathering llana honey, by AG) The augmentative intensifies the degree of the property attributed to the single argument: wefw ‘to be big’ > wefwtaj ‘to be very big’ and pitaj ‘to be long’ > pitajtaj ‘to be very long’. As shown in example (33), the qualitative intensification of the property implies a shift towards the positive end of the positive–negative axis on a semantic scale. The ‘beehive holes’ are considered good because they are big and long. This, however, depends on the connotation of the property. The attenuation of the attribute ‘rowdy’ is considered a good thing in (32), since it implies that the bees are not so aggressive. In fact, it seems that, in cases where EM combines with state verbs, the negative or positive connotation is carried by the verb rather than by the evaluative form. Hence, if a property or attribute is considered positive in a particular case, its intensification will be positive whereas its attenuation will be considered negative. If a property or attribute is considered negative, its attenuation will be positive and its intensification, negative. A similar intensification meaning is shown by adverbs derived through the diminutive (34)–(35): (34) Sak’alh toj ti-­pak nech’e-­fwaj fabric conj indf.sbj-­dye recently-­dim ‘The fabric that was very recently dyed is hers.’ (35) ¡Ufw, ufw, ch’ayu, ais, interj interj 3sbj:be.hot voc ‘Uf, uf, it burns, hey, it burns!’

la-­kho. 3poss-­belonging

chayu! 3sbj:be.hot

elh yu(k)=pej: ‘¡Yaj-­nepho, la-­pelh(a)j-­hila-­k paj-­fwaj!’ other 3sbj:say=iter proh-­leave 2sbj-­be.white-­fut-­desid then-­ dim ‘and the other repeated once again: “Do not leave! You are going to be white promptly!”’ (mythical story, by AG) The diminutive in combination with the adverb nech’e ‘recently’ in (34) means ‘very recently’, ‘a very short time ago’; it intensifies temporal proximity. Likewise, the evaluative construction pajfwaj ‘promptly, immediately’ in (35) expresses temporal proximity from the adverb paj ‘then’. The use of the diminutive instead of the augmentative in these cases could be explained by the fact that the evaluative construction changes

Wichi 641 towards a shorter temporal distance from an event occurring at a particular moment in time. Another use of the diminutive for semantic intensification is the evaluative construction formed from the cardinal number unu ‘one’ (36). As mentioned above, Wichi cardinal numbers have been completely replaced by Spanish loanwords (Vidal and Nercesian 2009) that are phonologically and morphologically integrated into the recipient language: (36)

Hiw’en unu-­fwaj atsetaj. 3sbj:have one-­ dim orange ‘She or he has only one orange.’

The EM intensifies the idea of ‘singleness’ and extremely reduced quantity. Of course, this is not a descriptive evaluation since there is no smaller amount than one item in this case. 2.2.2 Contempt/Endearment The evaluative construction with the diminutive can also express both contempt and endearment. The contempt or endearment interpretation depends on context, pragmatic functions and cultural concepts. For example, it is unlikely that a Wichi speaker would interpret the evaluative construction illustrated in (37) in a contemptuous way, because Wichi old people deserve respect. Similarly, mothers may refer to their children from a standard or qualitative perspective (38), but it is extremely rarely that they would imply contempt. On the other hand, it would be possible to use contempt when referring to a man (39), in addition to expressing small size, as shown in (8). Again, the descriptive or qualitative perspective of the evaluation depends on the context: (37)

n’-­choti-­fwaj 1poss-­grandfather-­dim ‘my dear little grandfather’ (endearment)

(38)

n’-­lhose-­fwaj 1poss-­daughter-­dim ‘my dear little daughter’ (endearment)

(39) hin’u-­fwaj man-­ dim ‘little man’ (contempt)

3 Conclusion Wichi EM uses three highly productive suffixes that form evaluative constructions from nouns, verbs, adverbs and numerals: -­fwaj diminutive, -­les diminutive class and -­taj ­augmentative. These are multifunctional suffixes that basically express both descriptive (quantitative) and qualitative evaluation. On the one hand, diminutive and augmentative constructions evaluate the size of an item according to a culturally established standard size. By a semantic and functional extension, the root les ‘offspring’ refers to the whole small-­size class of an entity. Additionally, evaluative suffixes create new words naming a new concept/entity whose ‘small’/‘big’ size is

642

Verónica Nercesian

inherent to the word meaning. From a qualitative perspective, in combination with nouns, verbs, adverbs or numerals, the diminutive can express attenuation, contempt and endearment, and the augmentative, intensification. Negative evaluative constructions graduate the intensity of a property/attribute/state between the two positive–negative axis extremes.

Notes 1. Chaco is a multilingual region, where forty native peoples live nowadays speaking at least twenty-­ nine languages (Golluscio and Vidal 2009/10, 5ff.). Apart from Wichi and the other Mataguayan languages, seven linguistic families (Tupian, Guaicuruan, Lule-­Vilela, Arawakan, Zamucoan and Enlhet-­Enenlhet or Maskoy) and the two isolated languages Besiro or Chiquitano and Guató (this last one considered extinct) are spoken in this area. 2. The unitalicised words in (12) are Spanish loanwords without full phonological integration. 3. Initials refer to the name of the native speaker that performed the utterances.

20.6 Yurakaré

Rik van Gijn

1 Introduction Yurakaré (also Yuracaré, Yurújare) is a small, isolate language, spoken in central Bolivia. The number of speakers can only be estimated, and lies between 2,500 and 3,500. Attempts have been made to connect Yurakaré with larger genealogical groups, but none of these attempts have had lasting success, and in fact most classifications regard Yurakaré as an isolate. The language is endangered, as the younger generation acquires at best a passive knowledge of it in the vast majority of communities, and it is giving way to Spanish as the preferred language of communication. The Yurakaré people live in small settlements along one of the many rivers of the area, and practise small-­scale farming, fishing and hunting. They are surrounded by other ethnic groups, such as the Trinitario (Arawak) to the north, the Chimane (Mosetenan) to the north-­ west, Bolivian Quechua (Quechuan) to the south and south-­west, and the Tupí-­Guaraní languages Yuki and Sirionó to the east. The use of Spanish is omnipresent both within and outside the Yurakaré community. Yurakaré is a polysynthetic, agglutinating (both prefixing and suffixing) and ­head-­marking language. Subjects and direct objects are indexed on the verb, and full NPs c­ an, but need not be expressed overtly. The alignment pattern of Yurakaré is nominative–accusative, determined only by the form and position of the pronominal affixes: direct objects are prefixed, subjects are suffixed. A system of applicative markers can increase the type and number of head-­marked arguments. The language distinguishes the following lexical classes: nouns, verbs, adjectives, adverbs, ideophones, interjections and enclitic particles. Nouns and adjectives can be used predicatively without any additional derivational marking or the use of copula. More detailed information can be found in Van Gijn (2006).

2  Evaluative morphology Different kinds of evaluations in Yurakaré are marked by the suffixes -­nñu diminutive, -­mashi medial degree, -­lë augmentative, and prefixed partial reduplication intensification. Although the markers can all be described under the heading of evaluative morphology, they form a heterogeneous group in the sense that they do not form a morphosyntactic system or paradigm: the hosts they combine with are not entirely equivalent, and they occupy different positions in the morphological template. They can also combine with each other in some circumstances. In spite of these differences, their functions sometimes overlap. I describe each of the markers in the following subsections.

Rik van Gijn

644 2.1 Diminutive -­nñu The diminutive marker -­nñu /ɲ:u/ on nouns marks small size:1

(1) a. Anakkimaya ajachanñu! ana-­kka-­ima-­ø=ya a-­katcha-­nñu dem-­mea-­coll-­3=rep 3sg.p-­axe-­dim ‘This size, his little axe was.’ b. Lëttima katcha, latijsha amashtunñu anakkima. lëtta-­ima katcha latijsha a-­machitu-­nñu ana-­kka-­ima coll axe then 3sg.p-­machete-­dim dem-­mea-­coll one-­ ‘He had an axe, and his little machete was of this size.’ This marker can also appear on property words or – less frequently – on event words expressing processes, where it can refer either to the size of the referent about which the property or action is predicated, as in (2a), or to the degree to which the predicate applies, as in (2b): (2) a. Todito anu shuyulënñu aalparatunñu, bënama animal bëshënñu. todito anu shuyulë-­nñu a-­alparatu-­nñu bë-­nama all like.this beautiful-­ dim 3sg.p-­apparatus-­dim atn-­be.like.this animal bëshë-­nñu animal thing-­ dim ‘It has all these beautiful little adornments (apparatus), like these little animal things.’ b. Ottonñunaja kummëw! otto-­nñu-­ø=w=naja kummë=w go.out-­dim-­3=pl=nsit tree=pl ‘The trees have come out a little.’ (the tips come out above the subsiding water) The suffix is also marginally used to encode endearment: (3)



Nij amakkay balitu, birikinñu? nij ama-­kka=y bali-­tu biriki-­nñu neg wh-­mea=loc go.pl-­1pl.s Brigida-­ dim ‘We did not go far, did we, little Brigida?

The suffix -­nñu is related to a number of roots related to suckling, small child, such as -­ñu-­‘child’ (obligatorily possessed), iñuma ‘procreate’, ñuñu ‘suckle’, ñuña ‘breastfeed’ and ñuñuta ‘nipple’. 2.2 Augmentative -­lë and -­ilë The suffix -­lë is found on modifiers, and augments what is expressed in the root it applies to. It is used more or less productively in the context of comparative constructions, even though it is not required there:2 (4)

Sëja pëpëli mëjsha. së=ja pëpë-­lë-­y

më=jsha

Yurakaré 645

1sg.pro=thm old-­ aug-­1sg.s ‘I am older than you.’

2sg.pro=abl

It is also used in non-­ comparative constructions, as in (5), where it simply indicates that what is expressed in the property or manner expression applies to a high degree: (5) a. Kamishëlë werejwerejweja, tanaya latiji kummëy wowore aposo. kamishë-­lë werejwerejwe-­ø=ja tana-­ø=ya latiji kummë=y wowore aug untie-­3=ss tie-­3=rep then tree=loc snake slowly-­ a-­poso 3sg.p-­beard ‘He very slowly untied himself, and tied the snake’s whiskers (beard) to the tree.’ b. Apiiw amalishtawti, kuratalë tuntaya libushajtiya. a-­pii=w amala-­ishta-­ø=w=ti kurata-­lë tunta=ya 3sg.p-­older.brother=pl come-­ fut-­3=pl=ds fast-­ aug tie=ss.irr li-­busha-­jti-­ø=ya del-­lay-­hab-­3=rep ‘When his older brothers were about to arrive, she very quickly wrapped him and put him back.’ In combination with spatially used demonstrative adverbs, the suffix -­lë indicates a location near the place the demonstrative refers to: (6) a. Lanalëmashi apumë na wororilë atchi. l-­ana-­lë-­mashi a-­pummë na wororilë ati=chi ref-­this-­aug-­hdg ipfv-­sing.prog dem great.tinamou.(bird) dem-­dir ‘Close by, a tinamou bird was singing there.’ b. Latijsha lanalëchi atattajsha kuniriya latiji. latijsha l-­ana-­lë=chi a-­tatta=jsha ku-­niri-­ø=ya latiji after.that ref-­dem-­aug=dir 3sg.p-­back=abl 3sg.vc-­speak-­3=rep then ‘Then, very close to her, from behind, she spoke to her.’ In combination with temporal vocabulary, this suffix indicates proximity in time to the moment expressed in the adverb: (7) a. numannumali numannuma-­lë=y dawn-­ aug=loc ‘in the early morning’ (around dawn) b. shëli shëy-­lë=y yesterday-­ aug=loc ‘the other day’ (not too long before yesterday) The marker also appears on verbs (in dependent clauses only), where it indicates that the action was just completed when another event started. This seems to relate directly to the proximity in time found as a function of lë on temporal adverbs:

646

Rik van Gijn

(8)

La dujnë kokkolëja, kumalaya abashti. la dujnë kokko-­lë-­ø=ja ku-­mala-­ø=ya a-­bashti then meat throw.away.pl-­aug-­3=ss 3sg.vc-­go.sg-­3=rep 3sg.p-­wife ‘Immediately after he threw away the meat, he followed his wife.’

The element -­lë occurs on a number of modifiers as a (semi-­)lexicalised element, for instance on ñuñujulë ‘small’, shudyulë ‘beautiful’, tishilë ‘now’ and willë ‘far’. The common denominator of all these different uses of -­ lë seems to be that it marks p­ roximity to some standard. This proximity can be interpreted spatially, temporally or qualitatively, depending on the nature of the root it applies to and the construction it appears in. The default interpretation for the use of the suffix with quality-­denoting roots is to i­ncrement the value expressed in the root. Nevertheless, in combination with the ­diminutive -­nñu, either on the same element (9a) or on another element (9b), the interpretation of -­lë can be guided towards smallness: (9) a. Anajalimanñu kapuchuya! ana-­kka-­lë-­ima-­nñu ka-­puchu-­ø=ya mea-­aug-­coll-­dim 3sg.coll-­escape-­3=rep this-­ ‘Only [a piece] this tiny survived, was overlooked (escaped).’ b. Anajalimaya lacha na aturumanñu. ana-­kka-­lë-­ima-­ø=ya lacha na a-­turuma-­nñu dem-­mea-­aug-­coll-­3=rep too dem 3sg.p-­arrow.type-­dim ‘She also had a small arrow, only this size.’ A seemingly related suffix -­ilë is marginally used on verbs to derive a person for whom the event expressed in the root applies excessively: tiyaylë tiya-­ilë eat-­ char ‘glutton’ (10)

2.3  Hedging -­mash(i) The suffix -­mashi appears on different kinds of predicates with the meaning ‘hedged’, referring to the limited extent to which the quality or event expressed in the root applies: (11) a. Kusuti lipuwamashiw latiji. kusuti li-­puwa-­mashi-­ø=w latiji maybe del-­drunk-­hdg-­3=pl then ‘Maybe they became a little drunk then.’ b. Kamalamashijti ush na shoja. ka-­mala-­mashi-­jti ushta na shoja 3sg-­go.sg-­hdg-­hab before dem daughter ‘ She carried (was pregnant) only a little, this woman.’ (that is: it wasn’t too visible) c. Lëmmuy tidala bintamashi. lëmmuy ti-­dala binta-­mashi-­ø

Yurakaré 647 only 1sg-­head hurt-­hdg-­3 ‘My head just hurts a little.’

In combination with verbs of cognition, such as knowing, realising, forgetting, etc., -­mashi indicates that the cognitive access is somewhat blurred: (12) a. Tëpshë malatumba? Ati nish lëdojo tëylemashi. tëtë-­bëshë ma-­la-­tumba-­ø ati nish lëdojo të-­yle-­mashi what-­ ent 3sg-­mal-­drop-­3 dem neg much 1sg-­be.known-­hdg ‘What was it that he dropped on them? I don’t know too much about this. b. Wëshëtëmashi latiji. wëshë-­ta-­mashi latiji feel-­ mid-­hdg then ‘Then it started to dawn on him.’ (lit.: ‘He realised it somewhat then.’) The suffix -­mash is found as an alternative to -­mashi, often (but not necessarily; see (14b)) forming adverbs rather than verbs: Nachimash malaya timbushucham. nachi-­mash mala=ya ti-­n-­bushu-­cha-­m there-­ hdg go.sg=ss.irr 1sg-­ben-­lie.down.sg-­jus-­2sg.s ‘Move a little further and wait for me lying down.’ (13)

The same variant of the marker is found in imperatives: (14) a. Tintelemashma! ti-­n-­itele-­mash-­ma 1sg-­ben-­let.go-­hdg-­imp.sg ‘Give me some slack.’ (let me go a little) b. Malamashma nakkamashi! mala-­mash-­ma na-­kka-­mashi go.sg-­hdg-­imp.sg dem-­mea-­hdg ‘Move over a little more that way.’ Combinations with -­nñu are possible, though they do not seem to be common. The precise effect of the combination is unclear: Wuriwurishinñumashi. wuriwurishi-­nñu-­mashi black-­dim-­hdg ‘It is somewhat black.’ (15)

Combinations with -­lë occur more often. In these, -­mashi hedges the effect of -­lë (see also (6a) above): (16) Pipikka, kusu dyindyilmashibëla. Pipi-­kka-­ø kusu dyindyi-­lë-­mashi-­ø=bëla Pipi-­ mea-­3 maybe small-­aug-­hdg-­3=cnt ‘He was the size of Pipi, or maybe still a bit smaller.’

Rik van Gijn

648 2.4  Intensifying reduplication

The final morphological expression of evaluation marks intensity, and is formally very different from the ones discussed above. Firstly, intensity is expressed by (partial) reduplication rather than by an affix, and secondly, the reduplicated form attaches to the left edge of the root rather than the right edge. The basic reduplicative process for forming intensity prefixes is to copy the first two moras of the stem, except the final coda (if present in the original), which is replaced by /h/, spelled j: (17) a. meyeye ‘disobedient’ >> meyej~meyeye ‘very, really disobedient’ b. binta ‘strong’ >> bij~binta ‘very strong’ Historical processes seem to have led to a number of exceptions to this general pattern of reduplication. It is beyond the scope of this chapter to discuss the deviations (for more information, see Van Gijn 2014). Suffice it to say that intensity reduplication always involves final [h], at least one syllable and at most two, and strong stress (in the case of disyllabic reduplication always on the rightmost syllable). In terms of semantics, the effects of the intensity reduplication are diverse, according to the semantics of the root it applies to. In the most straightforward case, when it applies to a root that expresses a quality X, it simply indicates that the quality applies to a high degree (‘very X’), as can be seen in (17), including the superlative interpretation:

(18)

Atijti mapi shujshunñe. ati-­jti ma-­pi shuj~shunñe dem-­hab 3pl-­older.brother.of.man ints~big ‘That was their biggest (oldest) brother.’

Words expressing processes or achievements are interpreted as being very close to their point of culmination when they carry the intensity prefix: (19) a. Shamajshamishtuychi. shamaj~shama-­ishta-­ø=w=ya=chi ints~die.pl-­fut-­3=pl=rep=frust ‘They really would have died.’ b. shëwëjshëwishtati shëwëj~shëw-­ishta-­ø=ti ints~become.dark-­fut-­3=ds ‘when it was really getting dark’ c. Malajmalawitaya. malaj~ma-­la-­wita-­ø=ya ints~3pl-­mal-­arrive.sg-­3=rep ‘They really had the urge [to urinate].’ (lit.: ‘[The urine] really arrived with them.’) Intensity reduplication is less commonly found on action words (in fact I have found it only on movement verbs), where its effect is that the action is performed to a great extent:

Yurakaré 649 (20)

Majmala. maj~mala-­ø inst~go.sg-­3 ‘He went really far.’

The intensity prefix takes on a more interactional meaning when applied to question words. In these circumstances it seems to express a strong curiosity to know the precise answer to the question, often in combination with the frustrative marker =chi: (21)

Ajami shetachi? aj~ama=y sheta-­ø=chi ints~wh=loc lost-­3=frust ‘Where exactly did he get lost?’

It is marginally possible to apply the intensity reduplication to nouns, as long as they can be interpreted as something gradable: (22)



majmapa maj~ma-­pa ints~3pl-­younger.brother.of.man ‘their youngest brother’

Combinations with other evaluative markers are rare. I have found no combination of the intensity marker with the limited degree marker -­mashi. The only instance I have found of a combination of the intensity marker with augmentative -­lë is in the lexicalised expression idojolë, which has idojo as its basic root, meaning ‘surpass’. With the augmentative, it indicates ‘much, many’; with the intensity reduplication, this basic meaning is exaggerated: (23) latijsha idojidojolë majmatataja latijsha idoj~idojo-­lë maj~matata-­ø=ja after.that ints~surpass-­aug ints~big-­3=ss ‘then, when he was really, really big’ A combination of the intensity marker and the diminutive -­nñu is possible in instances such as the one in (24), which also contains a lexicalised instance of -­lë, but this combination is very rare. In this case, it is difficult to say which of the markers has scope over the other: (24)

ñuj~ñuñujulë-­nñu ints~small-­dim ‘very small’

3 Conclusion The field of evaluative morphology in Yurakaré consists of four markers that do not form a natural morphological system: the reduplicated intensity marker attaches to the left edge of the stem, and the three evaluative suffixes occupy different positions in the morphological template, corresponding to their scope.

Rik van Gijn

650 Standard

Relative

Absolute

Decrease

-nñu

-mashi

Increase

Incremental

Maximal

-lë

Reduplication

Figure 20.6.1  System of evaluative morphology in Yurakaré

(25)

ROOT -­lë -­nñu -­mashi

Moreover, they do not attach to the same kind of roots in equal measures. The diminutive -­nñu has a preponderance with nouns, although it can also be found on adjectives and verbs; the other markers have a preference for verbs and adjectives. In terms of function, the markers sometimes overlap, but their different functions can be schematically represented as in Figure 20.6.1, which makes the differences between the markers clear.

Acknowledgements I am grateful to the speakers of Yurakaré for patiently teaching me their language. I am furthermore thankful to a number of funding agencies that have supported the documentation of Yurakaré over the years: Prof. Muysken’s Spinoza Program ‘Lexicon and Syntax’ (NWO Netherlands) and the Volkswagen DoBeS foundation.

Notes 1. Examples are glossed in four lines in order to provide the Yurakaré data in both its standard Roman orthography (line 1) and a morphemic analysis (line 2), along with morpheme glosses (line 3) and an idiomatic English translation (line 4). 2. Comparative constructions do not seem to be highly grammaticalised in Yurakaré, as there is quite some variation. One of the points of variation is the presence of the augmentative marker on the predicate, which can be left out if it is sufficiently clear that a comparison is made. Alternatively, the intensifying prefix can be used, or no evaluative modification at all.

References

References

Abdel-­Massih, E. T. (1971), A Reference Grammar of Tamazight (Middle Atlas Berber), Ann Arbor: Center for Near Eastern and North Africa Studies, University of Michigan. Aboh, E. O., Smith, N. and Zribi-­Hertz, A. (eds) (2012a), The Morphosyntax of Reiteration in Creole and Non Creole Languages, Amsterdam: Benjamins. Aboh, E. O., Smith, N. and Zribi-­Hertz, A. (2012b), ‘Reduplication beyond the word level’, in E. O. Aboh, N. Smith and A. Zribi-­Hertz (eds), The Morphosyntax of Reiteration in Creole and Non Creole Languages, Amsterdam: Benjamins, pp. 1–26. Agbetsoamedo, Y. (2011), ‘The Sɛlɛɛ diminutive -­bi: origin and function’, paper presented at the 12th International Pragmatics Conference, Manchester, 2–8 July. Agee, M. (1993), ‘Fronting in San Jeronimo Mazatec’, SIL Mexico Papers, 10, pp. 29–37. Ahenakew, F. and Wolfart, H. C. (1983), ‘Productive reduplication in Plains Cree’, in W. Cowan (ed.), Actes du Quatorzième Congrès des Algonquinistes, Ottawa: Carleton University, pp. 369–77. Aikhenvald, A. Y. (2003), Classifiers: A Typology of Noun Categorization Devices, Oxford: Oxford University Press. Aikhenvald, A. Y. (2008), The Manambu Language of East Sepik, Papua New Guinea, Oxford: Oxford University Press. Akhalaia, P. (2006), ‘ანთროპონიმთა კნინობით-­ალერსობითი ფორმები მეგრულში’ [Diminutive forms and forms of endearment in Megrelian anthroponyms], saenatmecniero dziebani, 23, pp. 31–6. Algonquian Linguistic Atlas (n.d.) http://www.atlas-­ling.ca. Allan, K. (1977), ‘Classifiers’, Language, 53: 2, pp. 285–311. Allen, J. H. D., Jr. (1941), ‘Portuguese word-­formation with suffixes’, Language Monographs, 17: 2. Alleyne, M. (1980), Comparative Afro-­American, Ann Arbor: Karoma. Alonso, A. (1935), ‘Noción, emoción, acción y fantasía en los diminutivos’, Volkstum Kultur Romanen, 8, pp. 104–26 (also in A. Alonso (1951), Estudios lingüísticos: Temas españoles, Madrid: Gredos, pp. 161–89). Alonso, A. (1937), El artículo y el diminutivo, Santiago de Chile: Universidad de Chile. Alves, M. (2014), ‘Austro-­Asiatic’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 520–44. Ambrazas, S. (1993), ‘On the development of diminutives in the Baltic languages’, Linguistica Baltica, 2, pp. 47–67. Ambrazas, V. (ed.) (1996), Dabartinės lietuvių kalbos gramatika, Vilnius: Mokslo ir enciklopedijų leidykla. Ameka, F. K. (2001), ‘Ideophones and the nature of the adjective word class in Ewe’,

652

References

in F. K. E. Voeltz and C. Kilian-­Hatz (eds), Ideophones, Amsterdam: Benjamins, pp. 25–48. Ameka, F. K. (2012), Ewe: Its Grammatical Constructions and Illocutionary Devices, Munich: Lincom Europa. Amery, R. (1996), ‘Kaurna numbers, maths and the quantification of time’. unpublished paper, available from University of Adelaide (also in R. Amery (1998), Warrabarna Kaurna! Reclaiming Aboriginal Languages from Written Historical Sources: A Kaurna Case Study, 2 vols., University of Adelaide: PhD dissertation, vol. II, pp. 64–77). Amery, R. (1998), Warrabarna Kaurna! Reclaiming Aboriginal Languages from Written Historical Sources: A Kaurna Case Study, 2 vols., University of Adelaide: PhD dissertation. Amery, R. (2000), Warrabarna Kaurna! Reclaiming an Australian Language, Lisse: Swets & Zeitlinger. Amery, R. (2013), ‘A matter of interpretation: language planning for a sleeping language, Kaurna, the language of the Adelaide Plains, South Australia’, Language Problems and Language Planning, 37: 2, pp. 101–24. Amery, R. and Simpson, J. with Wanganeen, T., Buckskin, J., Goldsmith, J. and Kaurna Warra Pintyandi (2007), Kulluru Marni Ngattaitya – Sounds Good to Me: A Kaurna Learner’s Guide, draft prototype version, Kaurna Warra Pintyandi, Adelaide (final version: R. Amery and J. Simpson with Kaurna Warra Pintyanthi (2013), Kulurdu Marni Ngathaitya! Sounds Good to me! A Kaurna Learner’s Guide, Kent Town: Wakefield Press). Amha, A. (2001), The Maale Language, Leiden: PhD dissertation. Anderson, S. R. (1992), A-­Morphous Morphology, Cambridge: Cambridge University Press. Androutsopoulos, J. K. (1998), Deutsche Jugendsprache: Untersuchungen zu ihren Strukturen und Funktionen, Frankfurt: Lang. Androutsopoulos, J. K. and Scholz, A. (eds) (1998), Jugendsprache – langue des jeunes – Youth Language, Frankfurt: Lang. Andrzejewski, B. W. (1956), ‘Accentual patterns in verbal forms in the Isaaq dialect of Somali’, Bulletin of the School of Oriental and African Studies, 18: 1, pp. 103–29. Andrzejewski, B. W. (1964), The Declensions of Somali Nouns, London: School of Oriental and African Studies. Andrzejewski, B. W. (1969), ‘Some observations on hybrid verbs in Somali’, African Language Studies, 10, pp. 47–89. Andrzejewski, B. W. (1979), The Case System in Somali, London: School of Oriental and African Studies. Anonby, E. J. (2011), A Grammar of Mambay, an Adamawa Language of Chad and Cameroon, Cologne: Rüdiger Köppe. Anvari, H. (2003), Farhang-­e Bozorg-­e Sokhan, 8 vols.,Tehran: Sokhan Press. Appah, C. K. I. and Amfo, N. A. N. (2011), ‘The morphopragmatics of the diminutive morpheme (-­ba /-­wa) in Akan’, Lexis 6: Diminutives and Augmentatives in the Languages of the World, pp. 85–103, http://lexis.univ-­lyon3.fr/IMG/pdf/Lexis_6. pdf. Aquilina, J. (1959), The Structure of Maltese, Malta: Royal University of Malta. Arabuli, A. (2009), ‘კნინობითის ერთი მოულოდნელი სინტაქსური უნარისათვის ქართული ენის დიალექტებში’ [Towards one unexpected syntagmatic ability

References 653 of diminution in the Georgian dialects], iberiul-­k’avk’asiuri enatmecniereba, 37, pp. 20–4. Arbaoui, N. (2010), Les dix formes de l’arabe classique à l’interface phonologie– syntaxe: pour une déconstruction du gabarit, University of Paris VII: PhD ­dissertation. Arcodia, G. F. (2012), Lexical Derivation in Mandarin Chinese, Taipei: Crane. Arends, J., Muysken, P. and Smith, N. (eds) (1994), Pidgins and Creoles: An Introduction, Amsterdam: Benjamins. Aronoff, M. (1976), Word Formation in Generative Grammar, Cambridge, MA: MIT Press. Aronoff, M., Meir, I. and Sandler, W. (2005), ‘The paradox of sign language morphology’, Language, 81: 2, pp. 301–44. Aronson, H. (1982), Georgian: A Reading Grammar, Columbus, OH: Slavica. Asma, S. (2009), On Monsters: An Unnatural History of Our Worst Fears, Oxford: Oxford University Press. Aspinion, R. (1953), Apprenons le berbère: initiation aux dialectes Chleuhs, Rabat: Éditions Félix Moncho. Augenstein, S. (1998), Funktionen von Jugendsprache: Studien zu verschiedenen Gesprächstypen des Dialogs Jugendlicher mit Erwachsenen, Tübingen: Niemeyer. Austin, P. (1981), ‘Switch-­reference in Australia’, Language, 57: 2, pp. 309–34. Authier, G. (forthcoming), ‘Tat’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Avram, A. A. (2011), ‘Pseudo-­reduplication, reduplication and repetition: the case of Arabic-­lexified pidgins and creoles’, Revue Roumanine de Linguistique-­Romanian, 3, pp. 225–56. Avrorin, V. A. (1959), Grammatika nanajskogo jazyka. Vol. 1, Moscow: Akademija Nauk. Avrorin, V. A. and Boldyrev, B. V. (2001), Grammatika orochskogo jazyka, Novosibirsk: RAN. Avrutin, S. (1999), Development of the Syntax–Discourse Interface, Dordrecht: Kluwer. Baayen, R. H. (1994), ‘Derivational productivity and text typology’, Journal of Quantitative Linguistics, 1, pp. 16–34. Baayen, R. H. and Schreuder, R. (eds) (2003), Morphological Structure in Language Processing, Berlin: De Gruyter. Bacelar, L. (2004), Gramática da língua Kanoê, Katholieke Universiteit Nijmegen: PhD dissertation. Bacelar, L. and da Silva Júnior, A. R. (2003), ‘Tipologia da negação em Kanoê’, Signótica: Revista do Mestrado em Letras e Lingüística, 15: 2, pp. 237–47. Bach, K. (2001), ‘Semantically speaking’, in I. Kenesei and R. M. Harnish (eds), Perspectives on Semantics, Pragmatics, and Discourse, Amsterdam: Benjamins, pp. 147–70. Badarneh, M. A. (2010), ‘The pragmatics of diminutives in colloquial Jordanian Arabic’, Journal of Pragmatics, 42: 1, pp. 153–67. Badia i Margarit, A. M. (1962), Gramática catalana, 2 vols., Madrid: Gredos. Bakema, P. and Geeraerts, D. (2004), ‘Diminution and augmentation’, in G. Booij, C. Lehmann, J. Mugdan and S. Skopeteas with W. Kesselheim (eds), Morphology: An International Handbook on Inflection and Word-­Formation. Vol. 2, Berlin: De Gruyter, pp. 1045–52.

654

References

Bakker, P. (1994), ‘Pidgins’, in J. Arends, P. Muysken and N. Smith (eds), Pidgins and Creoles: An Introduction, Amsterdam: Benjamins, pp. 25–39. Bakker, P. (2003), ‘The absence of reduplication in pidgins’, in S. Kouwenberg (ed.), Twice as Meaningful: Reduplication in Pidgins, Creoles and Other Contact Languages, London: Battlebridge, pp. 37–46. Bakker, P. (2006), ‘Algonquian verb structure: Plains Cree’, in G. Rowicka and E. Carlin (eds), What’s in a Verb?, Utrecht: LOT, pp. 3–27. Bakker, P., Daval-­Markussen, A., Parkvall, M. and Plag, I. (2011), ‘Creoles are typologically distinct from non-­creoles’, Journal of Pidgin and Creole Languages, 26: 1, pp. 5–42. Bakker, P. and Parkvall, M. (2005), ‘Reduplication in pidgins and creoles’, in B. Hurch (ed.), Studies on Reduplication, Berlin: De Gruyter, pp. 511–32. Banti, G. (1988a), ‘Two Cushitic systems: Somali and Oromo Nouns’, in H. van der Hulst and N. Smith (eds), Autosegmental Studies on Pitch Accent, Dordrecht: Foris, pp. 11–50. Banti, G. (1988b), ‘“Adjectives” in East Cushitic’, in M. Bechhaus-­Gerst and S. Fritz (eds), Cushitic – Omotic: Papers from the International Symposium on Cushitic and Omotic Languages, Cologne, 6–9 January 1986, Hamburg: Helmut Buske, pp. 205–59. Banti, G. (2001), ‘New perspectives on the Cushitic verbal system’, Proceedings of the Twenty-­Seventh Annual Meeting of the Berkeley Linguistics Society: Special Session on Afroasiatic Languages, pp. 1–48. Banti, G. (2011), ‘Somali language’, in S. Uhlig (ed.), Encyclopaedia Aethiopica. Vol. 4, Wiesbaden: Harrassowitz, pp. 693a–696b. Bar-­el, L. (2008), ‘Verbal number and aspect in Skwxwú7mesh’, Recherches Linguistiques de Vincennes, 37, pp. 31–54. Barsalou, L. (1999), ‘Perceptual symbol systems’, Behavioral and Brain Sciences, 22, pp. 577–660. Basset, A. and Picard, A. (1948), Éléments de grammaire berbère (Kabilie-­Irjen), Alger: Édition ‘La Typo-­Litho’ et Jules Carbonel Réunis. Basso, K. H. (1990), Western Apache Language and Culture: Essays in Linguistic Anthropology, Tucson: University of Arizona Press. Batchelor, R. E. and San José, M. Á. (2010), A Reference Grammar of Spanish, Cambridge: Cambridge University Press. Bat-­El, O. (1993), ‘Parasitic metrification in the Modern Hebrew stress system’, Linguistic Review, 10, pp. 189–210. Bat-­El, O. (1994a), ‘Stem modification and cluster transfer in Modern Hebrew’, Natural Language & Linguistic Theory, 12: 4, pp. 571–96. Bat-­ El, O. (1994b), ‘The optimal acronym word in Hebrew’, in P. Koskinen (ed.), Proceedings of the 1994 Annual Conference of the Canadian Linguistic Association, Toronto: Toronto Working Papers in Linguistics, pp. 23–37. Bat-­El, O. (1997), ‘On the visibility of word internal morphological features’, Linguistics, 35, pp. 289–316. Bat-­El, O. (2012), ‘Prosodic alternations in Modern Hebrew segholates’, in M. Muchnik and T. Sadan (eds), Studies in Modern Hebrew and Jewish Languages, Jerusalem: Carmel, pp. 116–29. Bates, D. (1919), Typescripts, Correspondence, Photographs etc., 11 vols., Folio 6/III/5k, University of Adelaide: Barr Smith Library.

References 655 Bates, E. and Rankin, J. (1979), ‘Morphological development in Italian: connotation and denotation’, Journal of Child Language, 6, pp. 29–52. Bateson, G. (1932), ‘Social structure of the Iatmul people of the Sepik River’, Oceania, 2: 3, pp. 246–89. Bauer, L. (1996), ‘No phonetic iconicity in evaluative morphology’, Studia Linguistica, 50: 2, pp. 189–206. Bauer, L. (1997), ‘Evaluative morphology: in search of universals’, Studies in Language, 21: 3, pp. 533–75. Bauer, L. (2003), Introducing Linguistic Morphology, Washington, DC: Georgetown University Press. Bauer, L. (2004), ‘The function of word-­formation and the inflection–derivation distinction’, in H. Aertsen, M. Hannay and R. Lyall (eds), Words in their ­ Places: A Festschrift for J. Lachlan Mackenzie, Amsterdam: Vrije Universiteit, pp. 283–92. Beard, R. (1981), The Indo-­European Lexicon: A Full Syncronic Theory, Amsterdam: North Holland. Beard, R. (1995), Lexeme-­Morpheme Base Morphology, Alabama: SUNY Press. Beckwith, C. (1993), The Tibetan Empire in Central Asia, Princeton: Princeton University Press. Belič, A. (1901), ‘Zur Entwicklungsgeschichte der slavischen Deminutiv-­ und Amplificativsuffixe’, Archiv für slavische Philologie, 23, pp. 134–206. Bellugi, U. and Fischer, S. D. (1972), ‘A comparison of sign language and spoken language’, Cognition, 1: 2, pp. 173–200. Bentley, M. and Varon, E. J. (1933), ‘An accessory study of phonetic symbolism’, American Journal of Psychology, 45: 1, pp. 76–86. Bentolila, F. (1981), Grammaire fonctionnelle d’un parler berbère: Aït Segrouchen d’Oum Jeniba (Maroc), Paris: Selaf. Berger, B. (2012), Louder than Words: The New Science of How the Mind Makes Meaning, New York: Basic Books. Bergman, B. (1983), ‘Verbs and adjectives: some morphological processes in Swedish Sign language’, in J. Kyle and B. Woll. (eds), Language in Sign: An International Perspective in Sign Languages, London: Croom Helm, pp. 3–9. Berko, J. (1958), ‘The child’s learning of English morphology’, Word, 14, pp. 150–77. Berlin, B. (1995), ‘Huambisa sound symbolism’, in L. Hinton, J. Nichols and J. Ohala (eds), Sound Symbolism, Cambridge: Cambridge University Press, pp. 76–93. Berman, H. (1986), ‘A note on the Yurok diminutive’, International Journal of American Linguistics, 52: 4, pp. 419–22. Bernal, E. (2012), ‘The translation of fictive orality and diastratic variation: appreciative derivation’, in J. Brumme and A. Espunya (eds), The Translation of Fictive Dialogue, Amsterdam: Rodopi, pp. 155–64. Bernal, E., Cabré, M. T., and Freixa, J. (eds) (forthcoming), La neologia lèxica catalana, Barcelona: Institut d’Estudis Catalans. Bernal, E., Gené, M. and Llopart, E. (forthcoming), ‘La neologia per prefixació’, in E. Bernal, M. T. Cabré and J. Freixa (eds), La neologia lèxica catalana, Barcelona: Institut d’Estudis Catalans. Bernal, E. and Sinner, C. (2009), ‘Al seu rotllo: aproximació al llenguatge juvenil català’, Zeitschrift für Katalanistik, 22, pp. 7–36. Berruto, G. (1995), Fondamenti di sociolinguistica, Bari: Laterza.

656

References

Berta, A. (1998), ‘West Kipchak Languages’, in L. Johanson and É. Á. Csató (eds), Turkic Languages, London: Routledge, pp. 301–17. Bertrang, A. (1921), Grammatik der Areler Mundart, Brussels: Lamertin et Hayez. Besten, H. den (2012), Roots of Afrikaans: Selected Writings of Hans den Besten, ed. T. von der Wouden, Amsterdam: Benjamins. Bickel, B. and Nichols, J. (2007), ‘Inflectional morphology’, in T. Shopen (ed.),  Language Typology and Syntactic Description. Vol. III: Grammatical Categories and the Lexicon, 2nd edn, Cambridge: Cambridge University Press, pp. 169–240. Bickford, J. A. and Fraychineaud, K. (2008), ‘Mouth morphemes in ASL: a closer look’, in R. Müller de Quadros (ed.), Sign Languages: Spinning and Unraveling the Past, Present and Future, Petrópolis: Editora Arara Azul, pp. 32–47. Bittner, M. and Hale, K. (1995), ‘Remarks on definiteness in Warlpiri’, in E. Bach, E. Jelinek, A. Kratzer and B. H. Partee (eds), Quantification in Natural Languages, Dordrecht: Kluwer, pp. 81–107. Black, J. M. (1920), ‘Vocabularies of four South Australian languages – Adelaide, Narrunga, Kukata, and Narrinyeri – with special reference to their speech sounds’, Transactions of the Royal Society of South Australia, 44, pp. 76–93. Blank, A. (1997), Prinzipien des lexikalischen Bedeutungswandels am Beispiel der romanischen Sprachen, Tübingen: Niemeyer. Bleek, W. H. I. (1869), A Comparative Grammar of South African Languages. Part 2: The Concord, Cape Town: Juta and Trübner. Blench, R. (2009), ‘Do the Ghana-­Togo Mountain languages constitute a genetic group?’, Journal of West African Languages, 36: 1/2, pp. 19–36. Blevins, J. (2001), Nhanda: An Aboriginal Language of Western Australia, Honolulu: University of Hawaii Press. Blevins, J. (2014), ‘Concatenative derivation: infixation’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 138–53. Blust, R. (1988), Austronesian Root Theory: An Essay on the Limits of Morphology, Amsterdam: Benjamins. Blust, R. (2009), The Austronesian Languages, Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, Australian National University. Bodor, P. and Barcza, V. (2007), ‘Acquisition of diminutives in Hungarian’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­ Linguistic Perspective, Amsterdam: Benjamins, pp. 231–61. Boeder, W. (2005), ‘The South Caucasian languages’, Lingua, 115, pp. 5–89. Böhmerová, A. (2011). ‘Suffixal diminutives and augmentatives in Slovak: a systemic view with some cross-­linguistic considerations’, Lexis 6: Diminutives and Augmentatives in the Languages of the World, pp. 59–84, http://lexis.univ-­lyon3.fr/ IMG/pdf/Lexis_6.pdf. Boldyrev, B. V. (1987), Slovoobrazovanie imen sushchestvitel’nykh v tunguso-­man’chzhurs kikh jazykakh v sravnitel’no-­istoricheskom osveshchenii, Novosibirsk: Nauka. Bolinger, D. (1972), Degree Words, The Hague: Mouton. Bolozky, S. (1994), ‘On the formation of diminutives in Modern Hebrew morphology’, Hebrew Studies, 35, pp. 47–62. Booij, G. (2000), ‘Inflection and derivation’, in G. Booij, C. Lehman and J. Mugdan with W. Kesselheim and S. Skopeteas (eds), Morphologie, Morphology: Ein interna-

References 657 tionales Handbuch zur Flexion und Wortbildung/An International Handbook on Inflection and Word Formation. Vol. 1, Berlin: De Gruyter, pp. 360–9. Booij, G. (2005), The Grammar of Words: An Introduction to Linguistic Morphology, Oxford: Oxford University Press. Booij, G. (2006), ‘Inflection and derivation’, in K. Brown (ed.), Encyclopedia of Language and Linguistics, 2nd edn, Oxford: Elsevier, pp. 654–61. Booij, G. (2007), The Grammar of Words: An Introduction to Linguistic Morphology, rev. edn, Oxford: Oxford University Press. Booij, G., Guevara, E., Ralli, A., Sgroi, S. C. and Scalise, S. (eds) (2005), Morphology and Linguistic Typology: On-­ Line Proceedings of the Fourth Mediterranean Morphology Meeting (MMM4), Catania, 21–23 September 2003, University of Bologna, http://www3.lingue.unibo.it/mmm2/?page_id=158. Borg, A. and Azzopardi-­Alexander, M. (1997), Maltese, London: Routledge. Bornstein, M. and Lamb, M. (1992), Developmental Psychology, 3rd edn, New York: Lawrence Erlbaum. Boyes-­Braem, P. and Sutton-­Spence, R. (eds) (2001), The Hands are the Head of the Mouth, Hamburg: Signum Press. Brackbill, Y. and Little, K. B. (1957), ‘Factors determining the guessing of meanings of foreign words’, Journal of Abnormal and Social Psychology, 54: 3, pp. 312–18. Bradley, D. (1994), A Dictionary of the Northern Dialect of Lisu (China and Southeast Asia), Canberra: Department of Linguistics, Research School of Pacific and Asian Studies, Australian National University. Bradley, D. (1995), ‘Grammaticalisation of extent in Mran-­Ni’, Linguistics of the Tibeto-­ Burman Area, 18: 1, pp. 1–28. Bradley, D. (2001), ‘Counting the family: family group classifiers in Yi Branch languages’, Anthropological Linguistics, 43: 1, pp. 1–17. Bradley, D. (2003), ‘Lisu’, in G. Thurgood and R. J. LaPolla (eds), The Sino-­Tibetan Languages, London: Routledge, pp. 222–35. Bradley, D. (forthcoming), Grammar of Lisu, Berlin: De Gruyter. Bradley, D., with E. R. Hope, J. Fish and M. Bradley (2006), Southern Lisu Dictionary, STEDT Monograph Series 4, Berkeley: University of California. Brincat, J. and Mifsud, M. (forthcoming), ‘Maltese’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Brown, K. (ed.) (2006), Encyclopedia of Language and Linguistics, 2nd edn, Oxford: Elsevier. Brown, L. (2011), Korean Honorifics and Politeness in Second Language Learning, Amsterdam: Benjamins. Brown, P. and Levinson, S. C. (1978), Politeness: Some Universals in Language Usage, Cambridge: Cambridge University Press. Brown, P. and Levinson, S. C. (1987), Politeness: Some Universals in Language Usage, 2nd edn, Cambridge: Cambridge University Press. Bruch, R. (1949), ‘Zur Diminutivbildung im Luxemburgischen’, Revue Trimestrielle d’Études Linguistiques, Folkloriques et Toponymiques, 31/2, pp. 169–79. Bruch, R. (1953), Grundlegung einer Geschichte des Luxemburgischen, Luxemburg: Linden. Brugmann, K. and Delbrück, B. (1906 [1916]), Grundriss der vergleichenden Grammatik der indoeuropäischen Sprachen. Band 2, Strasbourg: Trübner.

658

References

Bruyne, J. de (1989), ‘Antolojoide’, Boletín de la Real Academia Española, 69, pp. 91–130. Buell, L. (2005), Issues in Zulu Verbal Morphosyntax, University of California, Los Angeles: PhD dissertation. Buffa, F. (1988), ‘O deminutívach v poľštine a slovenčine’, Studia Linguistica Polono-­ Slovaca, 1, pp. 41–7. Bühler, K. (1934), Sprachtheorie: Die Darstellungsfunktion der Sprache, Jena: Fischer. Butler, J. L. (1971), Latin -­īnus, -­īna, -­ĭnus and -­ĭneus, Berkeley: University of California Press. Bybee, J. L. (1985), Morphology: A Study of the Relation Between Meaning and Form, Amsterdam: Benjamins. Cabré, M. T. (1994), A l’entorn de la paraula, Valencia: Universitat de València. Cabré, M. T. and Rigau, G. (1986), Lexicologia i semàntica, Barcelona: Enciclopèdia Catalana. Cabrera, P. (1910), Ensayos sobre etnologia Argentina. Tomo 1: Los Lules, Cordoba: F. Domenici. Caffi, C. (2007), Mitigation, Amsterdam: Elsevier. Caffi, C. and Janney, R. W. (1994), ‘Toward a pragmatics of emotive communication’, Journal of Pragmatics, 22, pp. 325–73. Cahill, M. (1997), Towards Reconstruction of the Noun Class System of Proto-­Buli-­ Kᴐnni, unpublished MS, Ohio State University. Cahill, M. (2004), ‘Tonal polarity in Kɔnni nouns’, Studies in African Linguistics, 33: 1, pp. 1–33. Cahill, M. (2007), Aspects of the Morphology and Phonology of Kɔnni, Dallas: SIL. Cai, Y. (1993), 海康方言誌 [The Haikang Dialect], Guangzhou: Zhongshan Daxue Chubanshe. Calandrelli, M. (1896), ‘Filología americana: Lule y Tonocoté’, Biblioteca, 1: 1, pp. 261–76. Cameron, D. and Coates, J. (1988), ‘Some problems in the sociolinguistic explanation of sex differences’, in J. Coates and D. Cameron (eds), Women and their Speech Communities, London: Longman, pp. 13–26. Camino, I. (2008), ‘Nafarroa Behereko Euskara Zaharra’, Anuario del Seminario de Filología Vasca don Julio de Urquijo–International Journal of Basque Linguistics and Philology, 42: 1, pp. 101–70. Cañete, P., Fernández, S. and Janer, F. (forthcoming), ‘La neologia per composició culta’, in E. Bernal, M. T. Cabré and J. Freixa (eds), La neologia lèxica catalana, Barcelona: Institut d’Estudis Catalans. Canonici, N. N. (1990), ‘Subclasses of Zulu nouns’, South African Journal of African Languages, 10: 2, pp. 52–8. Cardona, G. R. (2006), Introduzione all’etnolinguistica, Turin: UTET. Carlson, R. (1994), A Grammar of Supyire, Berlin: Mouton. Carpio, M. B. (2007), ‘Número y categorías afines en la lengua toba (familia guaycurú, Argentina)’, in A. Fernández Garay and M. Malvestitti (eds), Estudios lingüísticos y sociolingüísticos de las lenguas indígenas sudamericanas, Santa Rosa: Universidad de La Pampa, pp. 13–27. Carstairs-­McCarthy, A. (1992), Current Morphology, London: Routledge. Carstens, V. (2011), ‘Hyperactivity and hyperagreement in Bantu’, Lingua, 121: 5, pp. 721–41.

References 659 Casad, E. (1974), Dialect Intelligibility Testing, University of Oklahoma: SIL. Caubet, D. (1993), L’arabe marocain. Tome I: Phonologie et morphosyntaxe, Paris: Éditions Peeters. Ceccherini, M., Bonifacio, S. and Zocconi, E. (1997), ‘Acquisition of diminutives in Italian’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­Linguistic Perspective, Amsterdam: Benjamins, pp. 157–63. Census of India (n.d.), http://www.censusindia.gov.in/Census_Data_2001/Census_Data_ Online/Language/Statement4.aspx. Chadwick, N. (1975), A Descriptive Grammar of the Djingili Language, Canberra: Australian Institute of Aboriginal Studies. Chantraine, P. (1933), La formation des noms en grec ancien, Paris: Champion. Chapman, C. (1996), ‘Diminutive plural infixation and the “West Franconian” problem’, in J. Fisiak (ed.), Linguistic Reconstruction and Typology, Berlin: De Gruyter, pp. 71–88. Chappell, H. (ed.) (2004), Chinese Grammar: Synchronic and Diachronic Perspectives, Oxford: Oxford University Press. Charleston, B. M. (1960), Studies on the Emotional and Affective Means of Expression in English, Bern: Francke. Child Anxiety Network (2001), http://www.childanxiety.net/Fears_Phobias_Anxiety.htm. Chimke, H. (2010), ზმნისწინთა სემანტიკური ნიუანსები ახალ სალიტერატურო ქართულში [Semantic Nuances of Preverbs in Modern Literary Georgian], Akaki Tsereteli State University, Kutaisi, Georgia: PhD dissertation, http://www.atsu.edu. ge/geo/sadoqtoro%20disertacia/disertacia_chimke.pdf. Chumburidze, Z., Nizharadze, L. and Kurdadze, R. (2007), სვანური ენა [Svan Language], Tbilisi: Petiti. Cinque, G. (1999), Adverbs and Functional Heads: A Cross-­ Linguistic Perspective, Oxford: Oxford University Press. Clahsen, H., Sonnenstuhl, I. and Blevins, J. P. (2003), ‘Derivational morphology in the German mental lexicon: A dual mechanism account’, in R. H. Baayen and R. Schreuder (eds), Morphological Structure in Language Processing, Berlin: De Gruyter, pp. 125–55. Clark, E. V. (1993), The Lexicon in Acquisition, Cambridge: Cambridge University Press. Clark, E. V. (1995), ‘Later lexical development and word formation’, in P. Fletcher and B. MacWhinney (eds), The Handbook of Child Language, Oxford: Blackwell, pp. 393–412. Cominetti, F. (2010), Sviluppo di lessico bisillabico e morfologia derivazionale in cinese: i suffissi -­zi e -­r, MA dissertation, University of Pavia. Compton, R. (2012), The Syntax and Semantics of Modification in Inuktitut: Adjectives and Adverbs in a Polysynthetic Language, University of Toronto: PhD dissertation. Compton, R. and Pittman, C. (2010), ‘Word-­formation by phase in Inuit’, Lingua, 120: 9, pp. 2167–92. Comrie, B. (2008), ‘The areal typology of Chinese: between North and Southeast Asia’, in R. Djamouri, B. Meisterernst and R. Sybesma (eds), Chinese Linguistics in Leipzig, Paris: EHESS/CRLAO, pp. 1–21. Comrie, B., Golluscio, L., González, H. and Vidal, A. (2010), ‘El Chaco como área lingüística’, in Z. Estrada and R. Arzapalo (eds), Estudios de lenguas amerindias 2: Lenguas indígenas, Hermosillo: Universidad de Sonora, pp. 85‒131. Constenla, A. (1998), Gramática de la lengua guatusa, Heredia: Editorial de la Universidad Nacional.

660

References

Constenla, A. (2002), ‘Acerca de la relación genealógica entre las lenguas lencas y las lenguas misumalpas’, Revista de Filología y Lingüística de la Universidad de Costa Rica, 28: 1, pp. 189–206. Constenla, A. (2005), ‘¿Existe relación genealógica entre las lenguas misumalpas y las chibchenses?’, Estudios de Lingüística Chibcha, 24, pp. 7–85. Constenla, A. (2007), La lengua de Térraba: esbozo fonológico y gramatical, y pequeño diccionario, San José: Editorial de la Universidad de Costa Rica. Constenla, A. (2008), ‘Estado actual de la subclasificación de las lenguas chibchenses y de la reconstrucción fonológica y gramatical del protochibchense’, Estudios de Lingüística Chibcha, 27, pp. 117–35. Constenla, A. (2012), ‘Chibchan languages’, in L. Campbell and V. Grondona (eds), The Indigenous Languages of South America: A Comprehensive Guide, Berlin: De Gruyter. Constenla, A., Elizondo, F. and Pereira, F. (1998), Curso básico de bribri, San José: Editorial de la Universidad de Costa Rica. Cook, C. and Johns, A. (2009), ‘Determining the semantics of Inuktitut postbases’, in M.-­A. Mahieu and N. Tersis (eds), Variations on Polysynthesis, Amsterdam: Benjamins, pp. 149–70. Cook, E.-­D. (2004), ‘A grammar of Dëne Sųłiné (Chipewyan)’, Algonquian and Iroquoian Linguistics, Special Athabaskan Number, Memoir 17, Winnipeg: Algonquian and Iroquoian Linguistics. Cook, T. (2013), ‘Explaining the final vowel mismatch in Zulu reduplication’, University of Pennsylvania Working Papers in Linguistics, 19: 1, pp. 40–50, http://repository. upenn.edu/pwpl/vol19/iss1/6. Corbett G. (1991), Gender, Cambridge: Cambridge University Press. Corbett, G. (2000), Number, Cambridge: Cambridge University Press. Cote, M. (1984), Nahkawēwin: Saulteaux (Ojibway Dialect of the Plains), Regina: Saskatchewan Indian Federated College. Coulson, S. (2006), ‘Constructing meaning’, Metaphor and Symbol, 21: 4, pp. 245–66. Cowan, G. (1965), Some Aspects of the Lexical Structure of a Mazatec Historical Text, Mexico: SIL. Craig, C. (ed.) (1986), Noun Classes and Categorization, Amsterdam: Benjamins. Creissels, D. (1999), ‘Origine et évolution des diminutifs et augmentatifs dans quelques langues africaines’, Silexicales, 2, pp. 29–35. Creissels, D. (2014), ‘Niger-­Congo’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 558–72. Crevels, M. and van der Voort, H. (2008), ‘The Guaporé-­Mamoré region as a linguistic area’, in P. Muysken (ed.), From Linguistic Areas to Areal Linguistics, Amsterdam: Benjamins, pp. 151–79. Croft, W. (2000a), ‘Parts of speech as language universals and as language-­particular categories’, in P. Vogel and B. Comrie (eds), Approaches to the Typology of Word Classes, Berlin: De Gruyter, pp. 65–102. Croft, W. (2000b), Explaining Language Change, London: Longman. Csató, É. Á. and Johanson, L. (1998), ‘Turkish’, in L. Johanson and É. Á. Csató (eds), Turkic Languages, London: Routledge, pp. 203–35. Cúneo, P. (2013a), Formación de palabras y clasificación nominal en el léxico etnobiológico en toba (guaycurú), Munich: Lincom Europa. Cúneo, P. (2013b), ‘Diminutivo en toba (guaycurú): forma y función’, paper presented at

References 661 the III Encuentro de Lenguas Indígenas Americanas (ELIA), Universidad Nacional de Río Negro, San Carlos de Bariloche, 15–17 May. Cúneo, P. (2014), ‘Augmentative in Toba (Guaycuruan): form and function’, in S. Danielsen, K. Hannß and F. Zúñiga (eds), Word Formation in South American Languages, Amsterdam: Benjamins, pp. 51–77. Cusic, D. (1981), Verbal Plurality and Aspect, Stanford University: PhD dissertation. Cutfield, S. (2011), Demonstratives in Dalabon, A Language of South-­Western Arnhem Land, Monash University: PhD dissertation. Cuturi, F. and Gnerre, M., (forthcoming), ‘Adposiciones espacial-­direccionales y temporal-­ aspectuales en Ombeayiüts (Huave de San Mateo del Mar) y sus funciones adverbiales’, in A. López Cruz and M. Swanton (eds), Memorias del Coloquio Francisco Belmar: las lenguas otomangues y oaxaqueñas ante el siglo XXI, Oaxaca: IEEPO/ INALI/UABJO/Fundación Alfredo Harp Helú. Cuxac, C. (2000), La langue des signes française (LSF): les voies de l’iconicité, Faits de Langues, n. 15–16, Special Issue, Paris: Ophrys Dąbrowska, E. (2006), ‘Low-­level schemas or general rules? The role of diminutives in the acquisition of Polish case inflections’, Language Sciences, 28, pp. 120–35. Dahl, Ö. (1985), Tense and Aspect Systems, Oxford: Blackwell. Dahl, Ö. (2006), ‘Diminutives and augmentatives’, in K. Brown (ed.), Encyclopedia of Language and Linguistics, 2nd edn, Oxford: Elsevier, pp. 594–5. Dahl, Ö. (2009), ‘Diminutives and augmentatives’, in K. Allan (ed.), Concise Encyclopedia of Semantics, Oxford: Elsevier, pp. 218–19. Dal, G. (1997), Grammaire du suffixe -­et(te), Paris: Didier Érudition. Dale, D. (1968), Shona Companion, Gweru: Mambo Press. Daltas, P. (1985), ‘Some patterns of variability in the use of diminutive and augmentative suffixes in spoken Modern Greek Koine (MGK)’, Glossologia, 4, pp. 63–88. Danielsen, S., Hannß, K. and Zúñiga, F. (eds) (forthcoming), Word Formation in South American Languages, Amsterdam: Benjamins. Dardano, M. (1978), La formazione delle parole nell’italiano di oggi, Rome: Bulzoni. De Belder, M., Faust, N. and Lampitelli, N. (2014), ‘On a low and a high diminutive: Evidence from Italian and Hebrew’, in A. Alexiadou, H. Borer and F. Schäfer (eds), The Syntax of Roots and the Roots of Syntax, Oxford: Oxford University Press, pp. 149–53. De Marco, A. (2002), ‘The development of diminutives in Italian: input and acquisition’, in M. D. Voeikova and W. U. Dressler (eds), Pre-­and Protomorphology: Early Phases of Morphological Development in Nouns and Verbs, Munich: Lincom Europa, pp. 133–51. de Reuse, W. J. (1994), Siberian Yupik Eskimo: The Language and its Contacts with Chukchi, Salt Lake City: University of Utah Press. de Reuse, W. J. (2009), ‘Polysynthesis as a typological feature: an attempt at a characterization from Eskimo and Athabascan perspectives’, in M.-­A. Mahieu and N. Tersis (eds), Variations on Polysynthesis, Amsterdam: Benjamins, pp. 19–34. de Reuse, W. J. with P. Goode (2006), A Practical Grammar of the San Carlos Apache Language, Munich: Lincom Europa. de Vinciennes, L. and Dallet, J. M. (1960), Initiation a la langue berbère (Kabylie), Fort-­ National: Fichier de Documentation Berbère. Déchaine, R.-­ M., Girard, R. and Mudzingwa, C. (forthcoming), Shona Consonant Mutation: Implications for the Phonology–Phonetics and Phonology–Syntax Interface, University of British Columbia. Déchaine, R.-­M., Girard, R., Mudzingwa, C. and Wiltschko, M. (2014), ‘The ­internal

662

References

syntax of Shona class prefixes’, Language Sciences, special issue on grammatical gender, guest eds M. Laporcaro, T. Paciaroni and A. Thornton, 43, pp. 18–46. DeLancey, S. (1998). ‘Semantic categorization in Tibetan honorific nouns’, Anthopological Linguistics, 40, pp. 109–23. Dench, A. and Evans, N. (1988), ‘Multiple case-­ marking in Australian languages’, Australian Journal of Linguistics, 8, pp. 1–47. Dene K’ę́ę́ T’ahsíi Ts’uuzi Gha Edįtł’éh (Nadaatł’éh)/South Slavey Topical Dictionary (1993), 2nd edn, Fort Simpson, Northwest Territories: Teaching and Learning Centre, Dehcho Divisional Board of Education. Denny, J. P. and Creider, C. A. (1976), ‘The semantics of noun classes in Proto-­Bantu’, Studies in African Linguistics, 7, pp. 1–30. Derzhanski, I. (2005), ‘On diminutive plurals and plural diminutives’, in G. Booij, E. Guevara, A. Ralli, S. Sgroi and S. Scalise (eds), Morphology and Linguistic Typology: On-­Line Proceedings of the Fourth Mediterranean Morphology Meeting (MMM4), Catania, 21–23 September 2003, University of Bologna, pp. 73–89, http://www3.lingue.unibo.it/mmm2/?page_id=158 and http://morbo.lingue.unibo.it/ mmm/mmm-­proc/MMM4/073–089-­Derzhanski-­MMM4.pdf. Descartes, R. (1901), Meditations on the First Philosophy in Which the Existence of God, and the Real Distinction of Mind and Body, are Demonstrated (English translation of (1641) Meditationes de prima philosophia, in qua Dei existentia et animæ immortalitas demonstrator), http://philos.wright.edu/DesCartes/MedE.html. Di Garbo, F. (2013), ‘Evaluative morphology and noun classification: a cross-­linguistic study of Africa’, SKASE Journal of Theoretical Linguistics, 10: 1, pp. 114–36, http:// www.skase.sk/Volumes/JTL22/pdf_doc/07.pdf. Di Garbo, F. (forthcoming), Diminutives, Noun Classification and Quantification: A Case Study of Bantu and Beyond, unpublished MS. Diffloth, G. (1994), ‘i: big, a: small’, in L. Hinton, J. Nichols and J. Ohala (eds), Sound Symbolism, Cambridge: Cambridge University Press, pp. 107–14. Dimmendaal, G. J. (2014), ‘Derivation in Nilo-­Saharan’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 591–608. Dingemanse, M. (2011), The Meaning and Use of Ideophones in Siwu, Nijmegen, Radboud University: PhD dissertation, http://thesis.ideophone.org. Dixon, R. M. W. (1982), Where Have All the Adjectives Gone? And Other Essays in Semantics and Syntax, Berlin: De Gruyter. Dixon, R. M. W. (2002), Australian Languages, Cambridge: Cambridge University Press. Doerfer, G. (1978), ‘Classification problems of Tungus’, in M. Weiers (ed.), Beiträge zur nordasiatischen Kulturgeschichte, Wiesbaden: Harrassowitz, pp. 1–26. Doke, C. M. (1935), Bantu Linguistic Terminology, London: Longman. Doke, C. M. (1973), Textbook of Zulu Grammar, Cape Town: Longman. Dokulil, M. (1962), Tvoření slov v češtině: teorie odvozování slov, Prague: Nakladatelství československé akademie věd. Doleschal, U. and Thornton, A. M. (eds) (2000), Extragrammatical and Marginal Morphology, Munich: Lincom Europa. Donaldson, T. (1980), Ngiyambaa: The Language of the Wanggaybuwan, Cambridge: Cambridge University Press. Dorais, L.-­J. (2003), Inuit Uqausiqatigiit – Inuit Languages and Dialects, 2nd edn., Iqaluit: Nunavut Arctic College.

References 663 Dowty, D. (1979), Word Meaning and Montague Grammar, Dordrecht: Reidel. Dowty, D. (1991), ‘Thematic proto-­ roles and argument selection’, Language, 67, pp. ­547–619. Dressler, W. U. (1968), Studien zur verbalen Pluralität, Vienna: Österreichische Akademie der Wissenschaft, Phil-­Hist. Dressler, W. U. (1989), ‘Prototypical differences between inflection and derivation’, Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung, 42, pp. 3–10. Dressler, W. U. (1994), ‘Evidence from first phases of morphology acquisition for linguistic theory: extragrammatic morphology and diminutives’, Acta Linguistica Hafniensia, 27, pp. 93–105. Dressler, W. U. (ed.) (1997), Studies in Pre-­and Protomorphology, Vienna: Österreichischen Akademie der Wissenschaften. Dressler, W. U. (2010), ‘Morphologie dynamique et statique des diminutifs’, Mémoires de la Société de Linguistique de Paris, 17, pp. 141–54. Dressler, W. U. (forthcoming), ‘Myths and facts about diminutives in synchrony, diachrony and child language’, Yearbook of the University of Athens, Episimi Loghi (Πρυτανεία Θ. Πελεγρίνη). Dressler, W. U. and Karpf, A. (1995), ‘The theoretical relevance of pre-­and protomorphology in language acquisition’, Yearbook of Morphology, Dordrecht: Kluwer, pp. 99–122. Dressler, W. U., Kastovsky, D., Pfeiffer, O. E. and Rainer, F. (eds) (2005), Morphology and its Demarcations, Amsterdam: Benjamins. Dressler, W. U. and Kiefer, F. (1990), ‘Austro-­ Hungarian morphopragmatics’, in W. U. Dressler, H. Luschützky, O. Pfeiffer and J. Rennison (eds), Contemporary Morphology, Berlin: De Gruyter, pp. 69–77. Dressler, W. U., Kilani-­Schoch, M. and Klampfer, S. (2003), ‘How does a child detect morphology? Evidence from production’, in R. H. Baayen and R. Schreuder (eds), Morphological Structure in Language Processing, Berlin: De Gruyter, pp. 391–425. Dressler, W. U., Lettner, L. E. and Korecky-­Kröll, K. (2012), ‘Acquisition of German diminutive formation and compounding in a comparative perspective: evidence for typology and the role of frequency’, in F. Kiefer, M. Ladányi and P. Siptár (eds), Current Issues in Morphological Theory: (Ir)regularity, Analogy and Frequency, Amsterdam: Benjamins, pp. 237–64. Dressler, W. U. and Merlini Barbaresi, L. (1987), ‘Elements of morphopragmatics’, in J. Verschueren (ed.), Levels of Linguistic Adaptation, Amsterdam: Benjamins, pp. 33–51. Dressler, W. U. and Merlini Barbaresi, L. (1989a), ‘Grammaticalizzazione morfopragmatica: teoria e tipologia, con particolare riguardo ai diminutivi nell’italiano, tedesco e inglese’, Quaderni del Dipartimento di Linguistica e Letterature Comparate (Università di Bergamo), 5, pp. 233–52. Dressler, W. U. and Merlini Barbaresi, L. (1989b), ‘Interfissi e non-­interfissi autosuffissali nell’italiano, spagnolo e inglese’, in F. Foresti, E. Rizzi and P. Benedini (eds), L’italiano tra le lingue romanze: Atti del XX Congresso della Società di Linguistica Italiana, Rome: Bulzoni, pp. 243–52. Dressler, W. U. and Merlini Barbaresi, L. (1992), ‘Italian diminutives as non-­prototypical word formation’, in L. Tonelli and W. U. Dressler (eds), Natural Morphology: Perspectives for the Nineties, Padua: Unipress, pp. 21–30.

664

References

Dressler, W. U. and Merlini Barbaresi, L. (1994), Morphopragmatics: Diminutives and Intensifiers in Italian, German and Other Languages, Berlin: De Gruyter. Dressler, W. U. and Merlini Barbaresi, L. (1999), ‘Morphopragmatics’, in J. Verschueren, J. Blommaert and J.-­ O. Östman (eds), Handbook of Pragmatics, Amsterdam: Benjamins, pp. 1–14. Dressler, W. U. and Merlini Barbaresi, L. (2001), ‘Morphopragmatics of diminutives and augmentatives: on the priority of pragmatics over semantics’, in I. Kenesei and R. M. Harnish (eds), Perspectives on Semantics, Pragmatics, and Discourse, Amsterdam: Benjamins, pp. 43–58. Dressler, W. U. and Merlini Barbaresi, L. (forthcoming), ‘Pragmatics and morphology: morphopragmatics’, in Y. Huang (ed.), The Oxford Handbook of Pragmatics, Oxford: Oxford University Press. Dryer, M. S. and Haspelmath, M. (eds) (2011), The World Atlas of Language Structures (WALS) [CD-­ROM]. Oxford: Oxford University Press. Dryer, M. S. and Haspelmath, M. (n.d.), The World Atlas of Language Structures Online, Leipzig: Max Planck Institute for Evolutionary Anthropology, http://wals.info. Dunn, J. A. and Kairov, S. (2009), Modern Russian Grammar, Abingdon: Routledge. Duthie, A. S. (1996), Introducing Ewe Linguistics Patterns, Accra: Ghana Universities Press. Dutton, T. E. (1996), Koiari, Munich: Lincom Europa. Dvonč, L., Horák, G., Miko, F., Mistrík, J., Oravec, J., Ružička, J. and Urbančok, M. (1966), Morfológia slovenského jazyka, Bratislava: Slovenské pedagogické nakladateľstvo. Edwards, J. (2009), Language and Identity, Cambridge: Cambridge University Press. Elenbaas, M. B. (2007), The Synchronic and Diachronic Syntax of the English Verb– Particle Combination, Radboud University: PhD dissertation. Ellegård, A. and Olofsson, A. (1995), Klassiska ord, Malmö: CWK Gleerups Utbildningscentrum AB. Elordieta, A. (2001), Verb Movement and Constituent Permutation in Basque, Utrecht: LOT. Emmorey, K. (2002), Language, Cognition, and the Brain: Insights from Sign Language Research, Mahwah, NJ: Lawrence Erlbaum. Emmorey, K. (ed.) (2003), Perspectives on Classifier Constructions in Sign Languages, Mahwah, NJ: Lawrence Erlbaum. Erikson, E. (1950), Childhood and Society, New York: Norton. Erschler, D. (forthcoming), ‘Ossetic’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Eslami, M. (2009), ‘Sakhtar-­e Tasrifi-­ye Kalame dar Zaban-­e Farsi’, Journal of Tabriz University Human Sciences and Literature Faculty, 211, pp. 2–13. Ettinger, S. (1974), Form und Funktion in der Wortbildung: Die Diminutiv-­und Augmentativmodifikation im Lateinischen, Deutschen und Romanischen. Ein kritischer Forschungsbericht 1900–1970, Tübingen: Narr. Euskaltzaindia [Basque Academy] (1985), Euskal Gramatika: Lehen Urratsak – I, Bilbao: Euskaltzaindia. Evans, N. (1995), A Grammar of Kayardild: With Historical-­Comparative Notes on Tangkic, Berlin: De Gruyter. Evans, N. (2003), Bininj Gun-­Wok: A Pan-­Dialectal Grammar of Mayali, Kunwinjku and Kune, Canberra: Pacific Linguistics.

References 665 Evans, N., Brown, D. and Corbett, G. (2001), ‘Dalabon pronominal prefixes and the typology of syncretism: a network morphology analysis’, Yearbook of Morphology 2000, pp. 103–72. Evans, N. and Merlan, F. (2003), ‘Dalabon verb conjugation’, in N. Evans (ed.), The Non-­Pama-­Nyungan Languages of Northern Australia: Comparative Studies of the Continent’s Most Linguistically Complex Region, Canberra: Pacific Linguistics, pp. 268–83. Evans, N., Merlan, F. and Tukumba, M. (2004), A First Dictionary of Dalabon, Maningrida: Maningrida Arts and Culture, Bawinanga Aboriginal Corporation. Evans, V., Bergen, B. K. and Zinken, J. (2007), ‘The cognitive linguistics enterprise: an overview’, in V. Evans, B. K. Bergen and J. Zinken (eds), The Cognitive Linguistics Reader, London: Equinox, pp. 1–60. Fabra, P. (1956), Gramàtica catalana, Barcelona: Teide. Falco, A. and Cecchini, M. (1957), La lingua spagnola, Rome: Cremonese. Falk, Y. (1996), ‘Metrical structure in Modern Hebrew’, Hebrew Linguistics, 40, pp. 5–19. Faltz, L. M. (1998), The Navajo Verb: A Grammar for Students and Scholars, Albuquerque: University of New Mexico Press. Farquharson, J. T. (2007), ‘Typology and grammar: creole morphology revisited’, in U. Ansaldo, S. Matthews and L. Lim (eds) (2007), Deconstructing Creole, Amsterdam: Benjamins, pp. 21–38. Farshidvard, K. (2007), Tarkib-­o Eshteghagh dar Zaban-­e Farsi, Tehran: Zavar Press. Fasold, R. (1992), ‘Sociolinguistics in linguistics’, in K. Bolton and H. Kwok (eds), Sociolinguistics Today: International Perspective, London: Routledge, pp. 351–5. Faust, N. (2011), La morphologie nominale de l’hébreu modern: études en morpho-­ syntaxe, University of Paris VII: PhD dissertation. Faust, N. (2014), ‘Underlying vowel length in Modern Hebrew: the many realizations of the vowel /a/’, Brill’s Annual of Afro Asiatic Language and Linguistics 6, pp. 156–83. Fehlen, F., Heinz, A., Peltier, F. and Thill, G. (2013), Umgangssprachen, Luxembourg: STATEC, http://www.statistiques.public.lu/fr/publications/ series/rp2011/2013/13– 13-­langues/index.html. Fejes, L. (forthcoming), ‘Permic’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Ferguson, C. A. (1977), ‘Baby talk as a simplified register’, in C. E. Snow and C. A. Ferguson (eds), Talking to Children, Cambridge: Cambridge University Press, pp. 219–35. Finney, M. A. (2002), ‘Compounding and reduplication in Krio’, Journal of West African Languages, 29: 2, pp. 23–34. Fischer, S. D. (1973), ‘Two processes of reduplication in the American Sign Language’, Foundations of Language, 9: 4, pp. 469–80. Fisher-­Jørgsen, E. (1978), ‘On the universal character of phonetic symbolism with special reference to vowels’, Studia Linguistica, 32, pp. 80–90. Fivaz, D. (1963), Some Aspects of the Ideophone in Zulu, Hartford Seminary Foundation, Connecticut: MA dissertation. Fleck, D. W. (2003), A Grammar of Matses, Rice University: PhD dissertation, http://hdl. handle.net/1911/18526.

666

References

Fleischer, W. and Barz, I. (1997, 2007, 2012), Wortbildung der deutschen Gegenwartssprache, Tübingen: Niemeyer (1st edn; later edns (2007), Tübingen: Niemeyer; (2012), Berlin: De Gruyter). Folena, G. (1983), L’italiano in Europa: esperienze linguistiche del Settecento, Turin: Einaudi. Foley, W. A. (1986), The Papuan Languages of New Guinea, Cambridge: Cambridge University Press. Fortescue, M. (1980), ‘Affix ordering in West Greenlandic derivational processes’, International Journal of American Linguistics, 46: 4, pp. 259–78. Fortin, A. (2011), The Morphology and Semantics of Expressive Affixes, University of Oxford: PhD dissertation, http://ora.ox.ac.uk/objects/uuid:88a23d7c-­c229–49af-­ 9fc9–2cb35fce9d54. Fortune, A. (1955), An Analytical Grammar of Shona, London: Longmans, Green. Fortune, G. (1984), Shona Grammatical Constructions. Vol. 1, 3rd edn, Harare: Mercury Press. Forza, F. (2011), ‘Doubling as a sign of morphology: a typological perspective’, Journal of Universal Language, 12: 2, pp. 7–44. Fradin, B. (1999), ‘La suffixation en -­ET est-­elle evaluative?’, Silexicales, 2, pp. 69–82. Fradin, B. (2001), ‘À propos du suffixe -­ET’, Le Français Moderne, LXIX: 1, pp. 86–98. Fradin, B. (2005), ‘Problemi semantici in morfologia derivazionale’, in A. M. Thornton and M. Grossmann (eds), La formazione delle parole, Rome: Bulzoni, pp. 1–34. Fradin, B. and Montermini, F. (2009), ‘La morphologie évaluative’, in B. Fradin, F. Kerleroux and M. Plénat (eds), Aperçus de morphologie du français, Saint-­Denis: Puv, pp. 231–66. Frampton, J. (2009), Distributed Reduplication, Cambridge, MA: MIT Press. Franceschet, R. (2013), The Process of Diminutivization in Patients With Language Impairments and Children Without Any Language Impairment: A Comparative Study, MA thesis, University of Padua. Franco, L., Zampieri, E. and Meneghello, F. (2013), ‘Evaluating evaluative morphology in agrammatism: a case study in Italian’, Lingue e Linguaggio, 2, pp. 191–222. Furdík, J. (2004), Slovenská slovotvorba, Prešov: Náuka. Gaide, F. (1988), Les substantifs masculins latins en . . .(i)o, . . .(i)onis, Louvain: Éditions Peeters. Galucio, A. V. (2001), The Morphosyntax of Mekens (Tupi), University of Chicago: PhD dissertation. Gambino, M. (2010), Italian Evaluative Morphology: A Syntactic Approach, http://paduaresearch.unipd.it/2964. Gamillscheg, E. (1921), ‘Grundzüge der galloromanischen Wortbildung’, in E. Gamillscheg and L. Spitzer (eds), Beiträge zur Romanischen Wortbildungslehre, Geneva: Olschki, pp. 1–80. Gamillscheg, E. and Spitzer, L. (eds) (1921), Beiträge zur Romanischen Wortbildungslehre, Geneva: Olschki. Ganiyev, F. (1998), ‘Süz Yasalışı’, in M. Z. Zakiyev, F. Ganiyev and K. Z. Zinnatullina (eds), Tatar Grammatikası I, Moscow: Insan/Kazan: Fikĭr, pp. 201–466. Gazdeliani, E. (1992), ‘კნინობითობის ხმოვნის სიგრძისა და შედგენილობისათვის სვანურში’ [About length and structure of diminutive vowel in Svan], iberiul-­ k’avk’asiuri enatmecniereba, 31, pp. 141–57. Genesin, M. and Matzinger, J. (forthcoming), ‘Albanian’, in P. O. Müller, I. Ohnheiser,

References 667 S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Genetti, C. (2007), A Grammar of Dolakha Newar, Berlin: De Gruyter. Georg, S. (2007), A Descriptive Grammar of Ket. Part I: Introduction, Phonology and Morphology, Folkestone: Global Oriental. Gerzenstein, A. (1992), ‘Una variedad oriental del mataco’, Hacia una Nueva Carta Étnica del Gran Chaco, 4, pp. 67‒79. Ghomeshi, J., Jackendoff, R., Rosen, N. and Russell, K. (2004), ‘Contrastive focus reduplication in English (the salad-­salad paper)’, Natural Language & Linguistic Theory, 22, pp. 307–57. Giles, H. and Powesland, P. F. (1975), Speech Style and Social Evaluation, London: Academic Press. Gili Fivela, B. (2009), ‘Modificare la melodia per attenuare o rafforzare: scelte fonologiche o modulazioni fonetiche?’, in B. Gili Fivela and C. Bazzanella (eds), Fenomeni di intensità nell’italiano parlato, Florence: Franco Cesati Editore, pp. 149–63. Gilij, F. S. (1782), Saggio di storia americana. Vol. 3: Della religione, e delle lingue degli Orinochesi, e di altri Americani, Rome: Luigi Perego. Gilles, P. and Trouvain, J. (2013), ‘Luxembourgish’, Journal of the International Phonetic Association, 43: 1, pp. 67–74. Gillis, S. (1997), ‘The acquisition of diminutives in Dutch’, in W. U. Dressler (ed.), Studies in Pre-­and Protomorphology, Vienna: Österreichischen Akademie der Wissenschaften, pp. 165–79. Gillis, S. (ed.) (1998), ‘Studies in the acquisition of number and diminutive marking’, Antwerp Papers in Linguistics 95, http://uahost.uantwerpen.be/apil/apil95/diminutives1.pdf. Givón, T. (1989), Mind, Code and Context: Essays in Pragmatics, Hillsdale, NJ: Lawrence Erlbaum. Gnerre, M. and Cuturi, F. (2014), ‘Afijos fusionales, dedicados y portmanteau en Ombeayiüts’, in J.-­L. Leonard (ed.), Patterns in Meso-­American Morphology, Paris: M. Houdiard, pp. 234–62. Gogolashvili, G. (ed.) (2011), თანამედროვე ქართული ენის მორფოლოგია [Morphology of Present-­ Day Georgian Language], Tbilisi: Arnold Chikobava Institute of Linguistics. Goldenberg, G. (1994), ‘Principles of Semitic word-­ structure’, in G. Goldenberg and S. Raz (eds), Semitic and Cushitic Studies, Wiesbaden: Harrassowitz, pp. 29–64. Golla, V. (ed.) (1996a), Hupa Language Dictionary/Na:tinixwe Mixine:whe’, 2nd edn, Hoopa, CA: Hoopa Valley Tribal Council. Golla, V. (1996b), ‘Hupa’, in I. Goddard (ed.). Handbook of North American Indian. Vol. 17, Washington, DC: Smithsonian Institution, pp. 365–89. Golluscio, L. and González, H. A. (2008), ‘Contact, attrition and shift in two Chaco languages: the case of Tapiete and Vilela’, in K. D. Harrison, D. S. Rood and A. Dwyer (eds), Lessons from Documented Endangered Languages, Amsterdam: Benjamins, pp. 196–294. Golluscio, L. and Vidal, A. (2009/10), ‘Recorrido sobre las lenguas del Chaco y los aporte a la investigación lingüística’, Amerindia, 33/4, pp. 3‒40. Golston, C. and Kehrein, W. (1998), ‘Mazatec onsets and nuclei’, International Journal of American Linguistics, 64: 4, pp. 311–37.

668

References

González, H. A. (2005), A Grammar of Tapiete (Tupi-­Guarani), University of Pittsburgh: PhD dissertation. Gooch, A. (1970), Diminutive, Augmentative and Pejorative Suffixes in Modern Spanish, Oxford: Pergamon Press. Goodman, M. F. (1964), A Comparative Study of Creole French Dialects, The Hague: Mouton. Gorelova, L. M. (2002), Manchu Grammar, Leiden: Brill. Gràcia, L. and Turon, L. (1998), ‘Els sufixos apreciatius: una descripció del problema i una proposta d’anàlisi semántica’, Estudi General, 17, pp. 11–35. Gràcia, L. and Turon, L. (2000), ‘On appreciative suffixes’, Acta Linguistica Hungarica, 47: 1–4, pp. 231–47. Grandi, N. (2001a), ‘Su alcune presunte anomalie della morfologia valutativa: il rapporto con il genere ed il numero’, Archivio Glottologico Italiano, 1, pp. 25–56. Grandi, N. (2001b), ‘I suffissi valutativi tra derivazione e flessione: uno studio interlinguistico’, Archivio Glottologico Italiano, 2, pp. 129–73. Grandi, N. (2002), Morfologie in contatto: le costruzioni valutative nelle lingue del Mediterraneo, Milan: Franco Angeli Grandi, N. (2003a), ‘Mutamenti innovativi e conservativi nella morfologia valutativa dell’italiano: origine, sviluppo e diffusione del suffisso accrescitivo -­one’, in N. Maraschio and T. Poggi Salani (eds), Italia linguistica anno mille–Italia linguistica anno duemila: Atti del XXXIV Congresso Internazionale di Studi della Società di Linguistica Italiana, Rome: Bulzoni, pp. 243–58. Grandi, N. (2003b), ‘Matrici tipologiche vs. tendenze areali nel mutamento morfologico: la genesi della morfologia valutativa in prospettiva interlinguistica’, Lingue e Linguaggio, 3, pp. 105–45. Grandi, N. (2005), ‘Sardinian evaluative morphology in typological perspective’, in I. Putzu (ed.), Sardinian in Typological Perspective, Bochum: Brockmeyer University Press, pp. 188–209. Grandi, N. (2008), I verbi deverbali suffissati in italiano: dai dizionari al Web, Cesena: Caissa Italia. Grandi, N. (2009), ‘Restrictions on Italian verbal evaluative suffixes: the role of aspect and actionality’, York Papers in Linguistics, 2: 10, pp. 46–66. Grandi, N. (2011a), ‘Evaluative affixes between inflection and derivation: a typological survey’, paper presented at the 44th Annual Meeting of the Societas Linguistica Europaea, Universidad de la Rioja, Logrono, 10 September. Grandi, N. (2011b), ‘Renewal and innovation in the emergence of Indo-­European evaluative morphology’, Lexis 6: Diminutives and Augmentatives in the Languages of the World, pp. 5–25, http://lexis.univ-­lyon3.fr/IMG/pdf/Lexis_6.pdf. Grandi, N. and Montermini, F. (2005), ‘Prefix–suffix neutrality in evaluative morphology’, in G. Booij, E. Guevara, A. Ralli, S. Scalise and S. C. Sgroi (eds), Morphology and Linguistic Typology: On-­ Line Proceedings of the Fourth Mediterranean Morphology Meeting (MMM4), Catania, 21–23 September 2003, University of Bologna, pp. 143–56, http://www3.lingue.unibo.it/mmm2/?page_id=158 and http:// mmm.lingue.unibo.it/mmm-­proc/MMM4/143–156-­Grandi-­Montermini-­MMM4.pdf. Grandi, N. and Scalise, S. (1999), ‘Les règles d’altération nominale en italien’, Silexicales, 2, pp. 83–93. Gray, A. (2012), The Languages of Pentecost Island, Ashford: Manples (British Friends of Vanuatu).

References 669 Gray, A. and Buletangsuu Temwakon, P. (2012), Bongmehee: Ih nii mini lemtakte nii li dalekte ah Apma. Ol nem mo pikja blong ol samting long Apma lanwis/An Illustrated Dictionary of Apma Language, unpublished MS. Greenberg, J. H. (1959), ‘Africa as a linguistic area’, in W. R. Bascom and M. J. Herskovits (eds), Continuity and Change in African Cultures, Chicago: University of Chicago Press. Greenberg, J. H. (1963), ‘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, pp. 73–113. Greenberg, J. H., Ferguson, C. A. and Moravcsik, E. A. (eds) (1978), Universals of Human Language, 4 vols., Stanford: Stanford University Press. Greenberg, Y. (2010), ‘Event internal pluractionality in Modern Hebrew: a semantic analysis of one verbal reduplication pattern’, Brill’s Annual of Afroasiatic Languages and Linguistics, 2, pp. 119–64. Gregová, R. (2009), ‘On phonetic iconicity in evaluative morphology of West Slavic languages’, Sučasni doslidžennja z inozemnoji filolohiji, 7, pp. 90–3. Gregová, R. (2010), ‘Evaluative morphology and (mor)phonological changes in diminutives of Slavic languages’, Sučasni doslidžennja z inozemnoji filolohiji, 8, pp. 118–21. Gregová, R., Körtvélyessy, L. and Zimmermann, J. (2010), ‘Phonetic iconicity in the evaluative morphology of a sample of Indo-­European, Niger-­Congo and Austronesian languages’, Word Structure, 3: 2, pp. 156–80. Grönke, U. (1955), Die Diminutiva des Isländischen, Göttingen: PhD dissertation. Grossmann, M. (2004), ‘Nomi collettivi’, in M. Grossmann and F. Rainer (eds), La formazione delle parole in italiano, Tübingen: Niemeyer, pp. 244–52. Grossmann, M. and Rainer, F. (eds) (2004), La formazione delle parole in italiano, Tübingen: Niemeyer. Gudschinsky, S. C. (1958), ‘Mazatec dialect history’, Language, 34, pp. 469–81. Günthner, S. and Mutz, K. (2004), ‘Grammaticalization vs. pragmaticalization? The development of pragmatic markers in German and Italian’, in W. Bisang, N. Himmelmann and B. Wiemer (eds), What Makes Grammaticalization? A Look from its Fringes and its Components, Berlin: De Gruyter, pp. 77–107. Gustafson, H. S. (2000), A Grammar of the Nipmuck Language, Winnipeg: University of Manitoba. Guthrie, M. (1971), Comparative Bantu. Vol. 2, Farnborough: Gregg International. Haag, M. (forthcoming), ‘What determines the constraints on the relationships between roots and lexical categories?’, in V. Vapnarsky and E. Veneziano (eds), Lexical Polycategoriality: Cross-­Linguistic, Cross-­Theoretical, and Language Acquisition Appoaches, Amsterdam: Benjamins. Haas, M. (1972), ‘The expression of the diminutive’, in M. E. Smith (ed.), Studies in Linguistics in Honor of G. L. Trager, The Hague: Mouton, pp. 148–52. Hagman, R. S. (1977), Nama Hottentot Grammar, Bloomington: Indiana University. Hakamies, R. (1951), Étude sur l’origine et l’évolution du diminutif latin et sa survie dans les langues romanes, Helsinki: Suomalaisen Kirjallisuuden Seuran Kirjapainon Oy. Hale, K. (1973), ‘Deep–surface canonical disparities in relation to analysis and change: an Australian example’, in T. A. Sebeok (ed.), Current Trends in Linguistics. Vol. XI: Diachronic, Areal and Typological Linguistics, The Hague: De Gruyter, pp. 401–58.

670

References

Halle, M. and Marantz, A. (1993), ‘Distributed Morphology and the pieces of inflection’, in K. Hale and S. J. Keyser (eds), The View from Building 20, Cambridge, MA: MIT Press, pp. 111–76. Hardman, M. J. (1978), ‘Jaqi: the linguistic family’, International Journal of American Linguistics, 44: 2, pp. 146–53. Hardman, M. J. (2000), Jaqaru, Munich: Lincom Europa. Hardman, M. J. (2001), Aymara: Studies in Native American Linguistics 35, Munich: Lincom Europa. Hardman, M. J. and Mosley, M. E. (1986), ‘Historical linguistics and archaeology’, paper presented at American Anthropological Association annual meeting, Washington, DC, November. Hargus, S. (2000), ‘The qualifier prefixes in Yukon Deg Xinag (Ingalik)’, International Journal of American Linguistics, 66: 1, pp. 1–21. Hargus, S. (2007), Witsuwit’en Grammar: Phonetics, Phonology and Morphology, Vancouver: UBC Press. Hargus, S. (2011), Athabaskan Languages, Oxford: Oxford Bibliographies. Harley, H. and Noyer, R. (1999), ‘State-­of-­the-­article: Distributed Morphology’, GLOT, 4: 4, pp. 3–9. Harvard Medical School (1988), ‘Childhood fears and anxieties’, Harvard Medical School Mental Health Letter, http://www.accg.net/childfear.htm. Haspelmath, M. (1995), ‘The growth of affixes in morphological reanalysis’, Yearbook of Morphology 1994, pp. 1–29. Haspelmath, M. (2002), Understanding Morphology, London: Arnold. Haspelmath, M. (2003), ‘The geometry of grammatical meaning: semantic maps and cross-­ linguistic comparison’, in M. Tomasello (ed.), The New Psychology of Language. Vol. 2, Mahwah, NJ: Lawrence Erlbaum, pp. 211–42. Haspelmath, M. (2007), ‘Pre-­established categories don’t exist: consequences for language description and typology’, Linguistic Typology, 11: 1, pp. 119–32. Haspelmath, M. (2010), ‘Comparative concepts and descriptive categories in cross-­ linguistic studies’, Language, 86: 3, pp. 663–87. Haspelmath, M. and Tadmor, U. (eds) (2009), Loanwords in the World’s Languages: A Comparative Handbook of Loanword Typology, Berlin: De Gruyter. Hasselrot, B. (1957), Études sur la formation diminutive dans le langues romanes, Uppsala: Uppsala Universitets Arsskrift; Wiesbaden: Harrassowitz. Hasselrot, B. (1972), Étude sur la vitalité de la formation diminutive française au XXe siècle, Uppsala: Acta Universitatis Uppsaliensis. Haverling, G. (2011), ‘La suffixation “diminutive” du latin préclassique et classique au latin tardif’, in M. Fruyt and O. Spevak (eds), La quantification en latin, Paris: L’Harmattan, pp. 225–57. Hayashi, M. (2011), The Structure of Multiple Tenses in Inuktitut, University of Toronto: PhD dissertation. Hayashi, M. and Spreng, B. (2005), ‘Is Inuktitut tenseless?’, in C. Gurski (ed.), Proceedings of the 2005 Canadian Linguistics Association Annual Conference,  pp.  1–12, http://ling.uwo.ca/publications/CLA-­ACL/CLA-­ACL2005. htm. Hayeková, M. (1956), ‘Deminutíva v slovníku, ich štylistické označenie a využitie’, Slovenská reč 1–2, XXI, pp. 76–9. Head, B. F. (1978), ‘Respect degrees in pronominal reference’, in J. H. Greenberg,

References 671 C. A. Ferguson and E. A. Moravcsik (eds), Universals of Human Language, 4 vols., Stanford: Stanford University Press, vol. 3, pp. 151–212. Healey, P. M. (1960), An Agta Grammar, Manila: Institute of National Language and SIL. Heilmann, L. (1963), Grammatica storica della lingua greca, Turin: SEI. Heine, B., Claudi, U. and Hünnemeyer, F. (1991), Grammaticalization: Conceptual Framework, Chicago: University of Chicago Press. Heine, B. and Kuteva, T. (2002), World Lexicon of Grammaticalization, Cambridge: Cambridge University Press. Henzen, W. (1965), Deutsche Wortbildung. 3: Durchgesehene und ergänzte Auflage, Tübingen: Niemeyer. Hercus, L. A. (1982), The Bāgandji Language, Canberra: Australian National University. Hercus, L. A. (1992), A Nukunu Dictionary, Canberra: Australian Institute of Aboriginal and Torres Strait Islander Studies. Hercus, L. A. (1994), A Grammar of the Arabana-­Wangkangurru Language, Lake Eyre Basin, South Australia, Canberra: Australian National University. Herms, I. (1989), ‘Zur Bildung von Diminutiv-­und Augmentativformen im Swahili’, Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung, 42, pp. 738–47. Hervás y Panduro, L. (1787), Idea dell’Universo. Vol. 21: Saggio pratico delle lingue, Cesena: Gregorio Biasini. Hewson, J. I. and Francis, B. (1990), The Micmac Grammar of Father Pacifique, Winnipeg: University of Manitoba. Hill, J. H. and Hill, K. C. (1978), ‘Honorific usage in Modern Nahuatl: the expression of social distance and respect in the Nahuatl of the Malinche Volcano’, Language, 54: 1, pp. 123–55. Hill, N. W. (2010), ‘An overview of Old Tibetan synchronic phonology’, Transactions of the Philological Society, 108: 2, pp. 110–25. Hill, N. W. (2011), ‘Alternances entre ḥ et b en tibétain ancien et dans les langues tibétaines modernes’, Revue d’Études Tibétaines, 20, pp. 115–22. Himmelmann, N. (2005), ‘The Austronesian languages of Asia and Madagascar: typological characteristics’, in K. A. Adelaar and N. Himmelmann (eds), The Austronesian Languages of Asia and Madagascar, London: Routledge, pp. 110–81. Hinton, L., Nichols, J. and Ohala, J. (eds) (1994a), Sound Symbolism, Cambridge: Cambridge University Press. Hinton, L., Nichols, J. and Ohala, J. (1994b), ‘Introduction: sound-­symbolic processes’, in L. Hinton, J. Nichols and J. Ohala (eds), Sound Symbolism, Cambridge: Cambridge University Press, pp. 1–12. Hisamova, F. M. (2006), Tatar Tĭlĭ Morfologiyesĭ, Kazan: Megarif. Hoeksema, J. (2012), ‘Elative compounds in Dutch: types and historical development’, in G. Oebel (ed.), Crosslinguistic Comparison of Intensified Adjectives and Adverbs. Vol. 2, Hamburg: Dr. Kovač, pp. 97–142. Hohenhaus, P. (2004), ‘Identical constituent compounding: a corpus-­based study’, Folia Linguistica, 38: 3–4, pp. 297–332. Holmes, J. (1995), Women, Men and Politeness, London: Longman. Holmes, P. and Hinchliffe, I. (2008), Swedish: An Essential Grammar, 2nd edn, London: Routledge. Holton, D., Mackeridge, P. and Philippaki-­Warburton, I. (1997), Greek: A Comprehensive Grammar of the Modern Language, London: Routledge.

672

References

Holton, G. (2000), The Phonology and Morphology of the Tanacross Athabaskan Language, University of California Santa Barbara: PhD dissertation. Hoogenraad, R. and Laughren, M. (2012), Warlpiri Picture Dictionary, Alice Springs: Institute for Aboriginal Development Press. Hopper, P. and Traugott, E. C. (1993), Grammaticalization, 2nd edn, Oxford: Oxford University Press. Horecký, J. (1959), Slovotvorná sústava slovenčiny, Bratislava: Vydavateľstvo Slovenskej akadémie Vied. Horecký, J. (1971), Slovenská lexikológia. I: Tvorenie slov, Bratislava: Slovenské pedagogické nakladateľstvo. Horiguchi, D. (2013), ‘Priedēkļi kā internacionālismu “nacionalizācijas” līdzekļi’, in A. Kalnača, I. Lokmane (eds), Valoda: nozīme un forma. 3: Teorija un metodoloģija latviešu valodniecībā, Riga: LU Akadēmiskais apgāds. Horner, K. and Weber, J.-­J. (2008), ‘The language situation in Luxembourg’, Current Issues in Language Planning, 9: 1, pp. 69–128. Hoshi, M. (1988), 現代チベット語文法(ラサ方言)[Gendai Chibetto-­go bunpō (Rasahyōgen)], Tokyo: Yunesuko Higashi Ajia Bunka Kenkyū Sentā. Hoshi, M. et al. (1981–90), Texts of Tibetan Folktales, Tokyo: Toyo Bunko. Hualde, J. I. (1991), Basque Phonology, London: Routledge. Hualde, J. I. and Ortiz de Urbina, J. (eds) (2003), A Grammar of Basque, Berlin: De Gruyter. Huang, Y. (2000), Anaphora: A Cross-­Linguistic Study, Oxford: Oxford University Press. Huayong, L. and Zhe, M. (2008), ‘The meaning morphemes “son” and diminutive form in Lianjiang dialect of Yue group’, Linguistic Sciences 7: 6, pp. 626–35. Hunston, S. and Thompson, G. (eds) (2000), Evaluation in Text: Authorial Stance and the Construction of Discourse, Oxford: Oxford University Press. Hurch, B. (ed.) (2005), Studies on Reduplication, Berlin: De Gruyter. Hutchinson, J. P. (1981), The Kanuri Language: A Reference Grammar, Madison: University of Wisconsin. Hyman, L. M. (ed.) (1979), Aghem Grammatical Structure, Los Angeles: University of South California. Hyman, L. M. (1981), ‘Tonal accent in Somali’, Studies in African Linguistics, 12: 2, pp. 169–201. Iacobini, C. and Adinolfi, A. (2008), ‘La derivazione nel parlato dell’italiano messa a confronto con quella dello scritto’, in M. Pettorino, A. Giannini, M. Vallone and R. Savy (eds), La comunicazione parlata: Atti del Congresso Internazionale (Napoli, 24–25 febbraio 2006), Naples: Liguori, pp. 494–512. Inchaurralde, C. (1997), ‘Space, reference, and emotional involvement’, in S. Niemeier and R. Dirven (eds), The Language of Emotions, Amsterdam: Benjamins, pp. 135–54. INDEC (2004–5), Encuesta complementaria de pueblos indígenas, Instituto Nacional de Estadísticas y Censos, http://www.indec.gov.ar/webcenso/ECPI/index_ecpi.asp Indigenous Language Dictionary Project (2008), Yami (Tao) Dictionary Project, http:// yamibow.cs.pu.edu.tw/index_en.htm. INEC (2011), X Censo Nacional de Población y VI de Vivienda: resultados generales, San José: Instituto Nacional de Estadística y Censos. Ingram, A. (2001), Anamuxra, Sydney: University of Sydney.

References 673 Inkelas, S. (2014), ‘Reduplication’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 169–89. Inkelas, S. and Zoll, C. (2005), Reduplication: Doubling in Morphology, Cambridge: Cambridge University Press. Institut d’Estudis Catalans (2002), Gramàtica de la llengua catalana, Barcelona: Institut d’Estudis Catalans. Preliminary version, http:// http://www.iecat.net/institucio/­ seccions/filologica/gramatica. Izre’el, S. (2009), ‘Constructive constructions: Semitic verbal morphology and beyond’, in G. Goldenberg and A. Shisha-­ Halevy (eds), Egyptian, Semitic and General Grammar: Studies in Memory of H. J. Polotsky, Jerusalem: Israel Academy of Sciences and Humanities, pp. 106–30. Jackendoff, R. (1990), Semantic Structures, Cambridge, MA: MIT Press. Jackendoff, R. (1991), ‘Parts and boundaries’. in B. Levin and S. Pinker (eds), Lexical and Conceptual Semantics, Oxford: Blackwell, pp. 9–45. Jaeggli, O. (1980), ‘Spanish diminutives’, in F. H. Nuessel (ed.), Contemporary Studies in Romance Languages, Bloomington: Indiana University Linguistics Club, pp. 142–58. Jamieson, C. (1982), ‘Conflated subsystems marking person and aspect in Chiquihuitlán Mazatec verb’, Internationl Journal of American Linguistics, 48, pp. 139–67. Jamieson, C. (1988), Gramática mazateca del Municipio de Chuiquihuitlan, Oaxaca, Mexico: Instituto Lingüístico de Verano. Janhunen, J. (1999), ‘Tonogenesis in Northeast Asia: Udeghe as a tone language’, Journal de la Société Finno-­Ougrienne, 88, pp. 31–8. Jannaris, A. N. (1897), An Historical Greek Grammar, London: Macmillan. Javarek, V. and Sudjič, M. (1973), Teach Yourself Books: Serbo-­Croat, London: Hodder. Jendraschek, G. (2009a), ‘Switch-­reference constructions in Iatmul: forms, functions, and development’, Lingua, 119, pp. 1316–39. Jendraschek, G. (2009b), ‘Origin and development of the Iatmul focus construction: subordination, desubordination, resubordination’, Folia Linguistica, 43: 2, pp. 345–90. Jendraschek, G. (2009c), ‘Clause linkage in a language without coordination: the adjoined clause in Iatmul’, in J. Helmbrecht, Y.-­M. Shin, S. Skopeteas and E. Verhoeven (eds), Form and Function in Language Research: Papers in Honour of Christian Lehmann, Berlin: De Gruyter, pp. 139–50. Jendraschek, G. (2011), ‘Questions on transitivity: Iatmul and beyond’, in F. Kratochvíl, A. Coupe and R. J. LaPolla (eds), Studies in Transitivity: Insights from Language Documentation, special issue of Studies in Language, 35: 3, pp. 555–87. Jendraschek, G. (2012), A Grammar of Iatmul, Habilitationsschrift, Universität Regensburg. Jespersen, O. (1922), Language: Its Nature, Development and Origin, London: Allen and Unwin. Jespersen, O. (1933), ‘Symbolic value of the vowel i’, in O. Jespersen, Linguistica: Selected Papers in English, French and German, Copenhagen: Levin and Munksgaard, pp. 283–303. Jetté, J. (1907), On the Language of the Ten’a: Part 3, MS KO898J1907a, Fairbanks: Alaska Native Language Archive. Jetté, J. and Jones, E. (2000), Koyukon Athabaskan Dictionary, Fairbanks: Alaska Native Language Center. Jiménez de la Espada, M. (ed.) (1965), Relaciones geográficas de Indias: Perú, 3 vols., Madrid: Atlas.

674

References

Johanson, L. (1998), ‘The structure of Turkic’, in L. Johanson and É. Á. Csató (eds), Turkic Languages, London: Routledge, pp. 30–66. Johanson, L. and Csató, É. Á. (eds) (1998), Turkic Languages, London: Routledge. Johns, A. (2001), ‘An inclination towards accusative’, Lingustica Atlantica, 21, pp. 127–44. Johns, A. (2007), ‘Restricting noun incorporation: root movement’, Natural Language & Linguistic Theory, 25, pp. 535–76. Johns, A. (2014), ‘Eskimo-­Aleut’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 702–23. Johnson, M. (1987), The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason, Chicago: University of Chicago Press. Johnston, T. and Schembri, A. (1999), ‘On defining lexeme in a signed language’, Sign Language & Linguistics, 2: 2, pp. 115–85. Josefsson, G. (1997), On the Principles of Word Formation in Swedish, Lund: Lund University Press. Jurafsky, D. (1993), ‘Universals in the semantics of the diminutive’, Proceedings of the Annual Meeting of the Berkeley Linguistics Society, 19: 1, pp. 423–36. Jurafsky, D. (1996), ‘Universal tendencies in the semantics of the diminutive’, Language, 72: 3, pp. 533–78. Kačmárová, A. (2010), ‘A short survey of diminutives in Slovak and English’, in A. Kačmárová (ed.), English Matters, pp. 17–21, http://www.pulib.sk/elpub2/FHPV/ Kacmarova1/index.html Kadima, M. (1969), Le système des classes en bantou, Leuven: Vander. Kahane, H. and Kahane-­Toole, R. (1948–9), ‘The augmentative feminine in the Romance languages’, Romance Philology, 2, pp. 135–75. Kalbasi, I. (1992), Sakht-­e Eshteghaghi-­ye Farsi-­ye Emruz, Tehran: Institute of Human Sciences and Cultural Studies. Kalme, V. and Smiltniece, G. (2001), Latviešu literārās valodas morfoloģija un vārddarināšana, Liepāja: LiePA. Kalnača, A. (2011), Morfoloģijas stilistika, Riga: LU Akadēmiskais apgāds. Kalnača, A. and Lokmane, I. (2012), ‘Semantics and distribution of Latvian reflexive verbs’, in A. Usonienė, N. Nau and I. Dabašinskienė (eds), Multiple Perspectives in Linguistic Research on Baltic Languages, Newcastle upon Tyne: Cambridge Scholars, pp. 231–59. Kandybowicz, J. (2008), The Grammar of Repetition: Nupe Grammar at the Syntax– Phonology Interface, Amsterdam: Benjamins. Kardos, E. (2011), Toward a Scalar Semantic Analysis of Telicity in Hungarian, University of Debrecen: PhD dissertation. Kari, J. (1975a), ‘A classification of Tanaina dialects’, Anthropological Papers of the University of Alaska, 17: 2, pp. 49–53. Kari, J. (1975b), ‘The disjunct boundary in the Navajo and Tanaina verb prefix complexes’, International Journal of American Linguistics, 41: 4, pp. 330–45. Kari, J. (1990), Ahtna Athabaskan Dictionary, Fairbanks: Alaska Native Language Center. Kari, J. (2007), Dena’ina Topical Dictionary, Fairbanks: Alaska Native Language Center. Kari, J. (n.d.), Dena’ina Lexical File, Lexware document, unpublished MS. Karttunen, F. (1990), ‘Conventions of polite speech in Nahuatl’, Estudios di Cultura Nahuatl 20, pp.  281–96, http://www.historicas.unam.mx/publicaciones/revistas/ nahuatl/pdf/ecn20/344.pdf.

References 675 Kasper, G. (1990), ‘Linguistic politeness: current research issues’, Journal of Pragmatics, 14: 2, pp. 193–218. Katamba, F. (1993), Morphology, London: Macmillan. Kaye, A. (ed.) (2007), Morphologies of Asia and Africa, Winona Lake, IN: Eisenbrauns. Keen, S. (1972), A Description of the Yukulta Language: An Australian Aboriginal Language of North-­West Queensland, Monash University: MA dissertation. Keen, S. (1983), ‘Yukulta’, in R. M. W. Dixon and B. J. Blake (eds), Handbook of Australian Languages, Canberra: Australian National University Press, pp. 190–304. Keller, R. (1984), ‘Bemerkungen zur Theorie des sprachlichen Wandels’, Zeitschrift für Germanistische Linguistik, 12, pp. 63–81. Kempe, V. and Brooks, P. J. (2001), ‘The role of diminutives in Russian gender learning: Can child-­directed speech facilitate the acquisition of inflectional morphology?’, Language Learning, 51, pp. 221–56. Kempe, V., Brooks, P. J., Mironova, N. and Fedorova, O. (2003), ‘Diminutivization supports gender acquisition in Russian children’, Journal of Child Language, 30, pp. 1–15. Kempe, V., Brooks, P. J. and Pirott, L. (2001), ‘How can child-­directed speech facilitate the acquisition of morphology?’, in M. Almgren, A. Barrena, M.-­J. Ezeizabarrena, I. Idiazabal and B. MacWhinney (eds), Research on Child Language Acquisition: Proceedings of the 8th Conference of the International Association for the Study of Child Language, Somerville, MA: Cascadilla, pp. 1237–47. Kenesei, I., Vago, R. M. and Fenyvesi, A. (1998), Hungarian, London: Routledge. Kennedy, C. and McNally, L. (2005), ‘Scale structure, degree modification and the semantic typology of gradable predicates’, Language, 81, pp. 345–81.
 Kerswill, P. (2006), ‘Migration and language’, in U. Ammon, N. Dittmar, K. J. Mattheier and P. Trudgill (eds), Sociolinguistics/Soziolinguistik, 2nd edn, Berlin: De Gruyter, vol. 3.3, pp. 2271–85. Keshani, K. (1993), Eshteghagh-­ e Pasvandi dar Zaban-­ e Farsi-­ ye Emruz, Tehran: Markaz-­e Nashr-­e Daneshgahi (Iran University Press). Ketrez, F. N. and Aksu-­Koç, A. (2007), ‘The (scarcity of) diminutives in Turkish child ­language’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­Linguistic Perspective, Amsterdam: Benjamins, pp. 279–93. Khalilov, M. and Khalilova, Z. (forthcoming), ‘Bezhta’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Khumalo, J. S. M. (1987), An Autosegmental Account of Zulu Phonology, University of the Witwatersrand, Johannesburg: PhD dissertation. Kiefer, F. (ed.) (2000), Strukturális magyar nyelvtan. 3: Morfológia, Budapest: Hungarian Academy of Sciences. Kiefer, F. (2004), ‘Morphopragmatic phenomena in Hungarian’, Acta Linguistica Hungarica, 51, pp. 325–49. Kiefer, F. and Laakso, J. (2014), ‘Uralic derivational morphology’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 473–92. Kiefer, F. and Ladányi, M. (2000a), ‘A szóképzés’, in F. Kiefer (ed.), Strukturális magyar nyelvtan. 3: Morfológia, Budapest: Hungarian Academy of Sciences, pp. 137–64.

676

References

Kiefer, F. and Ladányi, M. (2000b), ‘Morfoszintaktikailag semleges szóképzés’, in F. Kiefer (ed.), Strukturális magyar nyelvtan. 3: Morfológia, Budapest: Hungarian Academy of Sciences, pp. 165–214. Kilani-­Schoch, M. and Dressler, W. U. (1999a), ‘Morphopragmatique interactionelle: les formations en -­o du francais bronché’, in L. Tonelli and W. U. Dressler (eds), Natural Morphology: Perspectives for the Nineties, Padua: Unipress, pp. 31–52. Kilani-­Schoch, M. and Dressler, W. U. (1999b), ‘Perspective morphopragmatique sur les formations en -­o du français branché’, in I. Mel’čuk (ed.), Dictionnaire explicatif et combinatoire du français contemporain: recherches lexico-­ sémantiques IV, Montreal: Les Presses de l’Université de Montréal, pp. 55–66. Kilani-­Schoch, M. and Dressler, W. U. (2005), Morphologie naturelle et flexion du verbe français, Tübingen: Narr. King, K. and Melzi, G. (2004), ‘Intimacy, imitation and language learning: Spanish diminutives in mother–child conversation’, First Language, 24, pp. 241–61. Kiparsky, P. (1998), ‘Partitive case and aspect’, in M. Butt and W. Geuder (eds), The Projection of Arguments, Stanford: CSLI, pp. 265–307. Kirchner, M. (1998), ‘Kazakh and Karakalpak’, in L. Johanson and É. Á. Csató (eds), Turkic Languages, London: Routledge, pp. 319–32. Kirk, P. L. (1966), Proto-­Mazatec Phonology, University of Washington: PhD d­ issertation. Kirk, P. L. (1970), ‘Dialect intelligibility testing: the Mazatec study’, International Journal of American Linguistics, 36: 3, pp. 205–11. Kitamura, H. (1975), ‘The honorifics in Tibetan’, Acta Asiatica, 29, pp. 56–74. Kitamura, H. (1977), Tibetan: Lhasa Dialect, Tokyo: Asia Africa Gengo Bunka Kenkyūjo, Tokyo Gaikokugo Daigaku. Klein, H. (1979), ‘Noun classifiers in Toba’, in M. Mathiot (ed.), Ethnolinguistics: Boas, Sapir and Whorf Revisited, The Hague: Mouton, pp. 85–95. Klein, T. (2006), ‘Segmental typology of African creole languages: examining uniformity, simplification and simplicity’, in F. A. Olaoba and M. A. Pemberton (eds), Selected Proceedings of the 36th Annual Conference on African Linguistics, Somerville, MA: Cascadilla Proceedings Project, pp. 42–50. Klimas, A. (1974), ‘Studies on word-­formation in Lithuanian’, Lituanus 20: 3, http://www. lituanus.org. Klimaszewska, Z. (1983), Diminutive und augmentative Ausdrucksmöglichkeitendes Niederländischen, Deutschen und Polnischen: Eine konfrontative Darstellung, Wrocław: Wydawnictwo Polskiej Akademii Nauk. Klokeid, T. J. (1976), Topics in Lardil Grammar, MIT: PhD dissertation. Kobakhidze, A. (1962), ‘კნინობითის წარმოება რაჭულში’ [The formation of diminutives in Rachian], iberiul-­k’avk’asiuri enatmecniereba, 13, pp. 157–61. Koeler, H. (1842), ‘Einige Notizen über die Eingeborenen an der Ostküste des St. Vincent’s Golfs in Süd-­Australien, 1837 und 1838’, Monatsberichte ueber die Verhandlungen der Gesellschaft fuer Erdkunde zu Berlin, 3, pp. 42–7. Korecky-­Kröll, K. (2011), Der Erwerb der Nominalmorphologie bei zwei Wiener Kindern: Eine Untersuchung im Rahmen der Natürlichkeitstheorie, University of Vienna: PhD dissertation. Korecky-­Kröll, K. and Dressler, W. U. (2007), ‘Diminutives and hypocoristics in Austrian German (AG)’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­Linguistic Perspective, Amsterdam: Benjamins, pp. 207–30. Korecky-­Kröll, K. and Dressler, W. U. (2010), ‘Diminutiverb bei zwei Wiener Kindern’,

References 677 in H. Bergmann, M. M. Glauninger, E. Wandl-­Vogt and S. Winterstein (eds), Fokus Dialekt: Analysieren – Dokumentieren – Kommunizieren. Festschrift für Ingeborg Geyer zum 60. Geburtstag, Hildesheim: Olms, pp. 239–50. Kormushin, I. V. (1998), Udyxejskij jazyk, Moscow: Nauka. Korompay, K. (1992), ‘A névszójelezés’, in L. Benkő (ed.), A Magyar nyelv történeti nyelvtana, vol. 2.1, Budapest: Hungarian Academy of Sciences, pp. 321–54. Körtvélyessy, L. (2011a), ‘Phonetic iconicity: lost in universality’, Bulletin of the Transilvania University of Braşov, 4: 53, pp. 137–44. Körtvélyessy, L. (2011b), ‘A cross-­linguistic research into phonetic iconicity’, Lexis 6: Diminutives and Augmentatives in the Languages of the World, pp. 27–39, http:// lexis.univ-­lyon3.fr/IMG/pdf/Lexis_6.pdf. Körtvélyessy, L. (2012a), ‘Evaluative morphology from cross-­ linguistic perspective’, paper presented at the Universals and Typology in Word-­Formation II conference, Košice, 26–8 August. Körtvélyessy, L. (2012b), Evaluative Morphology from a Cross-­Linguistic Perspective, Budapest, ELTE: Habilitationsschrift. Körtvélyessy, L. (2014), ‘Evaluative derivation’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 296–316. Körtvélyessy, L. and Štekauer, P. (2011a), ‘Introduction’, Lexis 6: Diminutives and Augmentatives in the Languages of the World, p. 3. http://lexis.univ-­lyon3.fr/IMG/ pdf/Lexis_6.pdf. Körtvélyessy, L. and Štekauer, P. (eds) (2011b), Diminutives and Augmentatives in the Languages of the World, Lexis 6, special issue, http://lexis.univ-­lyon3.fr/IMG/pdf/ Lexis_6.pdf. Kossmann, M. G. (1997), Grammaire du parler berbère de Figuig (Maroc oriental), Paris: Éditions Peeters. Kotsinas, U.-­ B. (1998), ‘Language contact in Rinkeby, an immigrant suburb’, in J. K. Androutsopoulos and A. Scholz (eds), Jugendsprache – langue des jeunes – Youth Language, Frankfurt: Lang, pp. 125–48. Kouwenberg, S. (ed.) (2003), Twice as Meaningful: Reduplication in Pidgins, Creoles and Other Contact Languages, London: Battlebridge. Kouwenberg, S. and LaCharité, D. (2003), ‘The meanings of “more of the same”: iconicity in reduplication and the evidence for substrate transfer in the genesis of Caribbean Creole languages’, in S. Kouwenberg (ed.), Twice as Meaningful: Reduplication in Pidgins, Creoles and Other Contact Languages, London: Battlebridge, pp. 7–18. Kouwenberg, S. and LaCharité, D. (2005), ‘Less is more: evidence from diminutive reduplication in Caribbean Creole languages’, in B. Hurch (ed.), Studies on Reduplication, Berlin: De Gruyter, pp. 533–46. Kouwenberg, S. and LaCharité, D. (2011), ‘The typology of Caribbean Creole reduplication’, Journal of Pidgin and Creole Languages, 26: 1, pp. 194–218. Kráľ, Á. and Sabol, J. (1989), Fonetika a fonológia, Bratislava: Slovenské pedagogické nakladateľstvo. Krejnovich, E. A. (1979), ‘Nivxskij jazyk’, in Jazyki Azii i Afriki III: Jazyki drevnej perednej Azii (nesemitskie). Iberijsko-­kavkazskie jazyki. Paleoaziatskie jazyki, Moscow: Nauka, pp. 295–329. Krifka, M. (1989), ‘Nominal reference, temporal constitution and quantification in event

678

References

semantics’, in R. Bartsch, J. van Benthem and P. van Emde Boas (eds), Semantics and Contextual Expression, Dordrecht: Foris, pp. 75–115. Krifka, M. (1992), ‘Thematic relations as links between nominal reference and temporal constitution’, in I. Sag and A. Szabolcsi (eds), Lexical Matters, Stanford: CSLI. pp. 29–53. Krifka, M. (1998), ‘The origins of telicity’, in S. Rothstein (ed.), Events and Grammar, Dordrecht: Kluwer, pp. 197–235. Krishnamurti, Bh. (2003), The Dravidian Languages, Cambridge: Cambridge University Press. Krishnamurti, Bh. and Gwynn, J. P. L. (1985), A Grammar of Modern Telugu, Delhi: Oxford University Press. Krjukova, E. (2005), Klass slov-­opredelitelej v enisejskikh jazykakh, Tomsk: Candidate degree dissertation. Kröger, F. (1992), Buli-­English Dictionary, Münster: Lit. Kruez, R. (1996), ‘The use of verbal irony: cues and constraints’, in J. Mio and A. Katz (eds), Metaphor: Implications and Applications, Mahwah, NJ: Lawrence Erlbaum, pp. 23–38. Kyle, J. and Woll, B. (eds) (1985), Sign Language: The Study of Deaf People and their Language, Cambridge: Cambridge University Press. Laalo, K. (1998), ‘Diminutives in Finnish child-­directed and child speech’, in S. Gillis (ed.), ‘Studies in the acquisition of number and diminutive marking’, Antwerp Papers in Linguistics 95, pp. 138–48, http://uahost.uantwerpen.be/apil/apil95/diminutives1.pdf. Laalo, K. (2001), ‘Diminutives in Finnish child-­directed and child speech: morphopragmatic and morphophonemic aspects’, Psychology of Language and Communication, 5: 2, pp. 72–80. Labov, W. (1990), ‘On the adequacy of natural languages: the development of tense’, in J. Singler (ed.), Pidgin and Creole Tense–Mood–Aspect Systems, Amsterdam: Benjamins, pp. 1–58. Labrune, L. (1999), ‘Les dérivés évaluatifs du japonais et l’émergence de la marque’, Silexicales, 2, pp. 107–16. Ladányi, M. (2007), Produktivitás és analógia a szóképzésben: elvek és esetek, Budapest: Tinta Könyvkiadó. Ladegaard, H. J. (2003), ‘Politeness in young children’s speech: context, peer group influence and pragmatic competence’, Journal of Pragmatics, 36, pp. 2003–22. Lafone y Quevedo, S. A. (1894), ‘Los Lules’, Boletín del Instituto Geográfico Argentino, 15, pp. 185–246. Lai, H.-­L. (2006), ‘Iconic coding of conceptualization: Hakka reduplicative constructions’, Language and Linguistics, 7: 2, pp. 483–99. Lakoff, G. (1972), ‘Hedges: a study in meaning criteria and the logic of fuzzy concepts’, Chicago Linguistic Society Papers, 8, pp. 183–228. Lakoff, G. (1987), Women, Fire, and Dangerous Things: What Categories Reveal About the Mind, Chicago: University of Chicago Press. Lakoff, G. and Johnson, M. (1980), Metaphors We Live By, Chicago: University of Chicago Press. Lakoff, R. (1975), Language and Woman’s Place, New York: Harper and Row. Lampitelli, N. (2012), ‘La flexion nominale en somali de Djibouti: Constatations empiriques et implications théoriques’, paper presented at Xuska 40 guurada farsoomaliida

References 679 (Celebrations for the 40th anniversary of Somali writing), Université de Djibouti, Djibouti, 17–22 December. Langacker, R. (1987), Foundations of Cognitive Grammar. Vol. 1, Stanford: Stanford University Press. Laoust, B. (1921), Cours de berbère marocain, Paris: A. Challamel Éditeur. LaPolla, R. J. (2001), ‘The role of migration and language contact in the development of the Sino-­Tibetan language family’, in R. M. W. Dixon and A. Y. Aikhenvald (eds), Areal Diffusion and Genetic Inheritance: Case Studies in Language Change, Oxford: Oxford University Press, pp. 225–54. Lass, R. (1997), Historical Linguistics and Language Change, Cambridge: Cambridge University Press. Laughren, M. (1984), ‘Warlpiri baby talk’, Australian Journal of Linguistics, 4: 1, pp. 73–88. Laughren, M. (1986), Warlpiri Maths Resource Material, Yuendumu: Bilingual Resources Development Unit. Lázaro Mora, F. (1999), ‘La derivación apreciativa’, in I. Bosque and V. Demonte (eds),  Gramática descriptiva de la lengua española, Madrid: Espasa, vol. 3, pp. 4645–82. Lazzeroni, R. (1963), ‘Per la storia dei derivati in -­on-­nelle lingue classiche’, Studi e Saggi Linguistici, 3, pp. 1–48. Lefebvre, C. (2003), ‘The emergence of productive morphology in creole languages: the case of Haitian Creole’, Yearbook of Morphology 2002, pp. 35–81. Legislative Assembly of Nunavut (2002), Debates (Hansard), 22 February. Lehmann, C. (2005), ‘Typologie d’une langue sans cas: le maya yucatèque’, Travaux du SELF, 10, pp. 101–14. Léonard, J. L. (2010), ‘Enquêtes exploratoires pour l’ALMaz (Atlas LingüísticoMazateco): elicitation croisée, entre typologie et codification d’une langue otomangue’, Géolinguistique, 12, pp. 59–109. Léonard, J. L., Dell’Aquila, V. and Gaillard-­Corvaglia, A. (2012), ‘The ALMaz (Atlas Lingüístico Mazateco): from geolinguistic data processing to typological traits’, STUF, 65: 1, pp. 78–94. Léonard, J. L. and Kihm, A. (2010), ‘Verb inflection in Chiquihuitlán Mazatec: a fragment and a PFM approach’, in S. Müller (ed.), Proceedings of the HPSG10 Conference, Stanford: CSLI, pp.  288–306, http://csli-­ publications.stanford.edu/HPSG/2010/ leonard-­kihm.pdf. Léonard, J. L. and Kihm, A. (2012), ‘Classes flexionnelles du mazatec et diasystème: empirisme critique et formalisation’, Bulletin de la Société de Linguistique de Paris, CVII: 1, pp. 379–446. Lepskaja, N. I. (1997), Jazyk rebjonka: ontogenez rečevoj kommunikacii, Moscow: Lomonosov University. LeSourd, P. (1995), ‘Diminutive verb forms in Passamaquoddy’, International Journal of American Linguistics, 61: 1, pp. 103–34. Levinson, S. C. (1983), Pragmatics, Cambridge: Cambridge University Press. Lewis, M. P., Simons, G. F. and Fennig, D. C. (2014), Ethnologue: Languages of the world, seventeenth edition, Dallas: SIL. Online version: http://www.ethnologue.com. Li, C. N. and Thompson, S. A. (1981), Mandarin Chinese: A Functional Reference Grammar, Berkeley: University of California Press. Li, F.-­K. (1930), Mattole: An Athabaskan Language, Chicago: University of Chicago Press.

680

References

Lieber, R. and Štekauer, P. (eds) (forthcoming), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press. Lin, H. (2001), A Grammar of Mandarin Chinese, Munich: Lincom Europa. Lindauer, M. S. (1988), ‘Size and distance perception of the physiognomic stimulus taketa’, Bulletin of the Psychonomic Society, 26: 3, pp. 217–20. Lloret, M. R. (1996), ‘Els diminutius i les marques de gènere’, in Estudis de lingüística i filologia oferts a Antoni M. Badia i Margarit, Barcelona: Publicacions de l’Abadia de Montserrat, vol. 3, pp. 63–76. Lloret, M. R. (2013), ‘Diferències en les terminacions flexives dels elements nominals’, in E. Clua and M. R. Lloret (eds), Qüestions de morfologia i lexicologia catalanes: Volum d’homentage a Joaquim Viaplana, Alacant: Institut Interuniversitari de Filologia Valenciana, pp. 245–78. LOD (2007–), Lëtzebuerger Online Dictionnaire, http://www.lod.lu/lod. Lomtatidze, K. (1987), ‘ნათესაობის ზოგი სახელის წარმოებისა და კნინობითობის -­ია სუფიქსის საკითხისათვის ქართულში’ [Issues on the formation of some kinship terms and the diminutive suffix -­ia in Georgian], iberiul-­k’avk’asiuri enatmecniereba, 26, pp. 68–73. Longobardi, G. (1994), ‘Reference and proper names: a theory of N-­movement in syntax and Logical Form’, Linguistic Inquiry, 25: 4, pp. 609–65. Lovick, O. (2005), Agentivity and Participant Marking: A Text-­Based Study, Universität zu Köln: PhD dissertation. Lovick, O. (2007), Dena’ina Texts, MS TI004L2007, Fairbanks: Alaska Native Language Archive. Lowenstamm, J. (1990), ‘Vocalic length and syllable structure in Semitic’, in A. S. Kaye (ed.), Semitic Studies in Honor of Wolf Leslau on the Occasion of his 85th Birthday, Wiesbaden: Harrassowitz, pp. 949–65. Lowenstamm, J. (1996), ‘CV as the only syllable type’, in J. Durand and B. Laks (eds), Current Trends in Phonology: Models and Methods, Salford: ESR, pp. 419–41. Lühr, R. (ed.) (2008), Nominale Wortbildung des Indogermanischen in Grundzügen: Die Wortbildungsmuster ausgewählter indogermansicher Einzelsprachen. Band 2, Hamburg: Dr. Kovač. Luo, K. (1987), 信宜方言誌 [The Xinyi Dialect], Guangzhou: Zhongshan Daxue Chubanshe. LWB [= Tockert, J., Palgen, H., Meyers, J., Ludovicy, E., Bruch, R., Dumont, J., Rinnen, H., Hess, J., Koenig, L. et al.] (1950–77), Luxemburger Wörterbuch, Luxemburg: Linden. Lyra Language Editor (n.d.), http://test.aymara.ufl.edu/LyraEditor/languageeditor.html. Machoni [Maccioni], A. (1732), Arte y vocabulario de la lengua lule y tonocoté, Madrid: Herederos de Juan García Infanzon. Madondo, L. (2000), Some Aspects of Evaluative Morphology in Zulu, http://uzspace. uzulu.ac.za/bitstream/handle/10530/647/Some%20Aspects%20of%20evaluative%20morphology%20in%20Zulu.%20L.M.M.S.%20Madondo.pdf?sequence=1. Magni, E. (1999), ‘Sémantique du diminutif et relations d’appartenance’, Silexicales, 2, pp. 139–48. Mahieu, M.-­A. and Tersis, N. (eds) (2009), Variations on Polysynthesis, Amsterdam: Benjamins. Maho, J. F. (1999), A Comparative Study of Bantu Noun Classes, University of Göteborg: PhD dissertation.

References 681 Maienborn, C. (2005), ‘On the limits of the Davidsonian approach: the case of copula sentences’, Theoretical Linguistics, 31, pp. 275–316. Maisak, T. and Ganenkov, D. (forthcoming), ‘Aghul’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Malcolm, D. M. (1960), A New Zulu Manual, Cape Town: Longman. Malkiel, Y. (1959), ‘The two sources of the Hispanic suffix -­azo, -­aco’, Language, 35, pp. 139–48. Malmgren, S.-­G. (1994), Svensk lexikologi, Lund: Studentlitteratur. Manova, S. (2005), ‘Derivation versus inflection in three inflecting languages’, in W. U. Dressler, D. Kastovsky, O. E. Pfeiffer and F. Rainer (eds), Morphology and its Demarcations, Amsterdam: Benjamins, pp. 233–52. Mansuur, A. (ed.) (1988), Le lingue Cuscitiche e il Somalo, Rome: MAE, Dipartimento per la Cooperazione allo Sviluppo. Margery, E. (1989), Diccionario cabécar-­español español-­cabécar, San José: Editorial de la Universidad de Costa Rica. Marrero, V., Aguirre, C. and Albalá, M. J. (2007), ‘The acquisition of diminutives in Spanish’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­Linguistic Perspective, Amsterdam: Benjamins, pp. 155–81. Martin, J. (2010), A Grammar of the Creek Language, Lincoln: University of Nebraska Press. Martinet, A. (1960), Éléments de linguistique générale, Paris: Armand Colin. Martirosovi, A. (1978), ‘კნინობითი სახელები ძველ ქართულში’ [Diminutives in Old Georgian], iberiul-­k’avk’asiuri enatmecniereba, 20, pp. 108–20. Marty, A. (1908), Untersuchungen zur Grundlegung der allgemeinen Gramatik und Sprachphilosophie, Halle: Niemeyer. Matasović, R. (forthcoming), ‘Kabardi’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Mathiassen, T. (1997), A Short Grammar of Latvian, Columbus, OH: Slavica. Maticsák, S. (forthcoming), ‘Mordvinic’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Matisoff, J. A. (1991), ‘The mother of all morphemes: augmentatives and diminutives in areal and universal perspectives’, in M. Ratliff and E. Schiller (eds), Papers from the first Annual Meeting of the Southeast Asian Linguistics Society, Tempe: Arizona State University, pp. 293–349. Matras, Y. (2007), ‘The borrowability of structural categories’, in Y. Matras and J. Sakel (eds), Grammatical Borrowing in Cross-­Linguistic Perspective, Berlin: De Gruyter, pp. 31–73. Mattes, V. (2006), ‘One form – opposite meanings? Full reduplication in Bikol’, paper presented at the Tenth International Conference on Austronesian Linguistics, Puerto Princesa City, Palawan, Philippines, 17–20 January, http://www-­ 01.sil.org/asia/ philippines/ical/papers/mattes-­diminutive%20and%20Augmentative%20Bikol.pdf. Matthews, S. and Yip, V. (2011), Cantonese: A Comprehensive Grammar, 2nd edn, London: Routledge. Mattiello, E. (2008), ‘From sexgate to vallettopoli: contrasting English and Italian combining forms’, in C. Vergaro (ed.), Conversarii: Dynamics of Language Contact in the Twenty-­First Century. Vol. 2, Perugia: Guerra, pp. 177–90.

682

References

Mattiello, E. (2013), Extra-­Grammatical Morphology in English, Berlin: De Gruyter. Mattissen, J. (2003), Dependent–Head Synthesis in Nivkh: A Contribution to a Typology of Polysynthesis, Amsterdam: Benjamins. McCarthy, J. J. (1981), ‘A prosodic theory of nonconcatenative morphology’, Linguistic Inquiry, 12, pp. 373–418. McKay, G. R. (1975), Rembarnga: A Language of Central Arnhem Land, Australian National University: PhD dissertation. McKay, G. R. (2011), Rembarnga, a Language of Central Arnhem Land, Munich: Lincom Europa. McWhorter, J. (1998), ‘Identifying the creole prototype: vindicating a typological class’, Language, 74: 4, pp. 788–818. McWhorter, J. (2005), Defining Creole, Oxford: Oxford University Press. Mead, D. (2003), ‘Evidence for a Celebic supergroup’, in J. Lynch (ed.), Issues in Austronesian Historical Phonology, Canberra: Pacific Linguistics, pp. 115–41. Meillet, A. and Vaillant, A. (1952), Grammaire de la langue Serbo-­ Croate, Paris: É. Champion. Meir, I. (2012), ‘Word classes and word formation’, in R. Pfau, M. Steinbach and B. Woll (eds), Sign Language: An International Handbook, Berlin: De Gruyter, pp. 77–112. Mélac, E., Robin, F. and Simon, C. (2014), Vocabulaire thématique de langue parlée français-­tibétain, Paris: L’Asiathèque. Mel’čuk, I. and Podolsky, B. (1996), ‘Stress in Modern Hebrew nominal inflection’, Theoretical Linguistics, 22, pp. 155–94. Melissaropoulou, D. (2006), Η μορφολογική διαδικασία του υποκορισμού στη διαλεκτική ποικιλία της Ελληνικής: Σύγκριση με την ΚΝΕ [Diminution in Greek Dialectal Variation: Comparison with Standard Modern Greek], University of Patras: MA dissertation. Melissaropoulou, D. (2007), Μορφολογική περιγραφή και ανάλυση του Μικρασιατικού ιδιώματος Κυδωνιών και Μοσχονησίων [Word Formation in the Asia Minor Dialect of Kydonies and Moschonisia (Aivaliot)], University of Patras: PhD dissertation. Melissaropoulou, D. (2009), ‘Augmentation vs. diminution in Greek dialectal variation: an optimal system’, in G. Boyé, F. Montermini and J. Tseng (eds), Selected Proceedings of the 6th Décembrettes: Morphology in Bordeaux, Somerville, MA: Cascadilla, pp. 125–37. Melissaropoulou, D. and Ralli, A. (2008), ‘Headedness in diminutive formation: evidence from Modern Greek and its dialectal variation’, Acta Linguistica Hungarica, 55, pp. 183–204. Melzi, G. and King, K. (2003), ‘Spanish diminutives in mother–child conversations’, Journal of Child Language, 30, pp. 281–304. Menz, A. (forthcoming), ‘Gagauz’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Merlan, F. (1983), Ngalakan Grammar, Texts and Vocabulary, Canberra: Pacific Linguistics. Merlini Barbaresi, L. (2001), ‘The pragmatics of the “diminutive” -­y/ie suffix in English’, in C. Schane-­ Wolles, J. Rennyson and F. Neubarth (eds), Naturally!, Turin: Rosenberg and Sellier, pp. 315–26. Merlini Barbaresi, L. (2002), ‘Morfopragmatica: due paradigmi morfologici a confronto’,

References 683 in G. Porcelli, M. L. Maggioni and P. Tornaghi (eds), Due codici a confronto, Brescia: La Scuola, pp. 109–39. Merlini Barbaresi, L. (2004a), ‘Alterazione’, in M. Grossmann and F. Rainer (eds), La formazione delle parole in italiano, Tübingen: Niemeyer, pp. 264–92. Merlini Barbaresi, L. (2004b), ‘Aggettivi deaggettivali’, in M. Grossmann and F. Rainer (eds), La formazione delle parole in italiano, Tübingen: Niemeyer, pp. 444–50. Merlini Barbaresi, L. (forthcoming), ‘The pragmatics of word-­formation’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Messineo, C. (1991), ‘Variantes dialectales del complejo lingüístico toba’, in J. Braunstein (ed.), Hacia una nueva carta étnica del Gran Chaco II, Las Lomitas: Centro del Hombre Antiguo Chaqueño, pp. 12–22. Messineo, C. (2003), Lengua toba (guaycurú): aspectos gramaticales y discursivos, Munich: Lincom Europa. Messineo, C. and Cúneo, P. (2011), ‘Ethnobiological classification in two indigenous languages of the Gran Chaco region: Toba (Guaycuruan) and Maká (Mataguayan)’, Anthropological Linguistics, 53: 2, pp. 132–69. Messineo, C. and Dell’Arciprete, A. (eds) (2005), Lo’onatacpi na qom derquil’ecpi: ­materiales del taller de lengua y cultra toba, Buenos Aires: Ministerio de Educación, Ciencia y Tecnología. Mesthrie, R. (2006), ‘Society and language: overview’, in K. Brown (ed.), Encyclopedia of Language and Linguistics, 2nd edn, Oxford: Elsevier, pp. 472–84. Mettouchi, A. (1999), ‘Le “t” n’est-­il qu’une marque de féminin en berbère (kabyle)?’, Faits de Languages, 14, pp. 217–25. Miller, C. (2003), ‘Reduplication in Arabic-­based contact languages’, in S. Kouwneberg (ed.), Twice as Meaningful: Reduplication in Pidgins, Creoles and Other Contact Languages, London: Battlebridge, pp. 289–300. Miller, G. (2006), Latin Suffixal Derivatives in English and their Indo-­European Ancestry, Oxford: Oxford University Press. Mirambel, A. (1940), ‘Note sur le suffixe diminutif -­áki en grec moderne’, Bulletin de la Société de Linguistique de Paris, 41, pp. 89–91. Mistrík, J. (1985), Štylistika, Bratislava: Slovenské pedagogické nakladateľstvo. Mithun, M. (1999), The Native Languages of North America, Cambridge: Cambridge University Press. Mogharrabi, M. (1993), Tarkib dar Zaban-­e Farsi, Tehran: Toos Press. Monaselidze, M. (2001), ზმნისწინი საშუალი გვარის ზმნებთან [Preverbs with Medial Verbs], Tbilisi: tbilisis universit’et’is gamomcemloba. Montermini, F. (1999), ‘L’apocope en italien en tant que procédé de morphologie évaluative’, Silexicales, 2, pp. 149–59. Moore, P. and Wheelock, A. (1990), Wolverine Myths and Visions: Dene Traditions from Northern Alberta, Lincoln: University of Nebraska Press. Morris, C. W. (1938), Foundations of the Theory of Signs, Chicago: University of Chicago Press. Mous, M. (2012), ‘Cushitic’, in Z. Frajzyngier and E. Shay (eds), The Afroasiatic Languages, Cambridge: Cambridge University Press, pp. 342–422. Müller, P. O., Ohnheiser, I., Olsen, S. and Rainer, F. (eds) (forthcoming), Word-­ Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter.

684

References

Munro, P. (1988), ‘Diminutive syntax’, in W. Shipley (ed.), In Honor of Mary Haas, The Hague: Mouton, pp. 539–55. Murphy, N. (2006), Bodies and Souls, or Spirited Bodies?, Cambridge: Cambridge University Press. MLLVG (Mūsdienu latviešu literārās valodas gramatika. I: Fonētika un morfoloģija) (1959), Riga: LPSR ZA izdevniecība. Mutaka, N. and Tamanji, P. N. (2000), An Introduction to African Linguistics, Munich: Lincom Europa. Mutz, K. (1999), ‘Les dérivés évaluatifs de l’italien: comment la diachronie explique la synchronie’, Silexicales, 2, pp. 161–7. Mutz, K. (2000), Die italienischen Modifikationssuffixe: Synchronie und Diachronie, Frankfurt: Peter Lang. Muysken, P. and Veenstra, T. (1994), ‘Serial verbs’, in J. Arends, P. Muysken and N. Smith (eds), Pidgins and Creoles: An Introduction, Amsterdam: Benjamins, pp. 289–302. Myers, F. R. (1986), Pintupi Country, Pintupi Self: Sentiment, Place and Politics among Western Desert Aborigines, Canberra: Australian Institute of Aboriginal Studies/ Washington, DC: Smithsonian Institution Press. Naden, T. (1989), ‘Gur’, in J. Bendor-­Samuel (ed.), The Niger-­Congo Languages, New York: SIL/University Press of America. pp. 140–68. Náñez, E. (1973), El diminutivo: historia y funciones en el español clásico y moderno, Madrid: Gredos. Napoli, D. J. and Reynolds, B. (1995), ‘Evaluative affixes in Italian’, Yearbook of Morphology 1994, pp. 151–78. Nash, D. (1982), ‘Warlpiri verb roots and preverbs’, in S. Swartz (ed.), Papers in Warlpiri Grammar: In Memory of Lothar Jagst, Darwin: SIL, pp. 165–216. Nedjalkov, I. V. (1997), Evenki, London: Routledge. Nercesian, V. (2011), Gramática del wichí, una lengua chaqueña: interacción fonología– morfología–sintaxis en el léxico, Universidad de Buenos Aires: PhD dissertation. Nercesian, V. (2014), ‘Matacoan’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 743–66. Newell, L. E (2006), Romblomanon Dictionary, Manila: Linguistic Society of the Philippines. Newman, P. (1980), The Classification of Chadic within Afroasiatic, Leiden: Universitaire Pers Leiden. Newton, G. (1996), Luxembourg and Lëtzebuergesch: Language and Communication at the Crossroads of Europe, Oxford: Oxford Universty Press. Ngakulmungan Kangka Leman (1996), Lardil Dictionary, Gununa: Mornington Shire Council. NGLEM (2010), Nueva gramática de la lengua española: manual, Madrid: Asociación de Academias de la Lengua Española. Nichols, J. (1971), ‘Diminutive consonant symbolism in western North America’, Language, 47: 4, pp. 826–48. Nichols, J. (1986), ‘Head-­marking and dependent-­marking grammar’, Language, 62: 1, pp. 56–84. Nieuwenhuis, P. (1985), Diminutives, University of Edinburgh: PhD dissertation. Nikolaeva, I. (2014), ‘Altaic’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 493–508. Nikolaeva, I. and Tolskaya, M. (2001), A Grammar of Udihe, Berlin: De Gruyter.

References 685 Nishimoto, E. (2003), ‘Measuring and comparing the productivity of Mandarin Chinese suffixes’, Computational Linguistics and Chinese Language Processing, 8: 1, pp. 49–76. Noccetti, S., De Marco, A., Tonelli, L. and Dressler, W. U. (2007), ‘The role of diminutives in the acquisition of Italian morphology’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­Linguistic Perspective, Amsterdam: Benjamins, pp. 125–53. Nübling, D. (2006), ‘Zur Entstehung und Struktur ungebändigter Allomorphie: Pluralbildungsverfahren im Luxemburgischen’, in C. Moulin and D. Nübling (eds), Perspektiven einer linguistischen Luxemburgistik: Studien zu Diachronie und Synchronie, Heidelberg: Winter, pp. 107–25. Nurse, D. and Philippson, G. (2003a), ‘Introduction’, in D. Nurse and G. Philippson (eds), The Bantu Languages, London: Routledge, pp. 1–12. Nurse, D. and Philippson, G. (eds) (2003b), The Bantu Languages, London: Routledge. Nyst, V. A. S. (2007), A Descriptive Analysis of Adamorobe Sign Language (Ghana), Utrecht: LOT. Ohala, J. J. (1994), ‘The frequency code underlies the sound-­symbolic use of voice pitch’, in L. Hinton, J. Nichols and J. Ohala (eds), Sound Symbolism, Cambridge: Cambridge University Press, pp. 325–47. Ohala, J. J. (1997), ‘Sound symbolism’, in Proceedings of the 4th Seoul International Conference on Linguistics (SICOL), Seoul: Linguistic Society of Korea, pp. 98–103. Okimāsis, J. (2004), Cree: Language of the Plains, Regina: Canadian Plains Research Center. Olmsted, H. (1994), ‘Diminutive morphology of Russian children: a simplified subset of nominal declension in language acquisition’, in C. E. Gribble, R. A. Rothstein, E. C. Haber, H. M. Olmsted, R. Szulkin and C. E. Townsend (eds), Alexander Lipson: In Memoriam, Columbus, OH: Slavica, pp. 165–207. Ondrus, P., Horecký, J. and Furdík, J. (1980), Súčasný slovenský spisovný jazyk: Lexikológia, Bratislava: Slovenské pedagogické nakladateľstvo. Ondruš, P. and Sabol, J. (1987), Úvod do štúdia jazykov, Bratislava: Slovenské pedago­ gické nakladateľstvo. Öner, M. (1998), Bugünkü Kıpçak Türkçesi, Ankara: Türk Dil Kurumu Yayınları. ORCAWETA (Organización Capitanía Weenhayek y Tapiete) (2011), ‘Resultados del censo Weenhayek’, Plan estratégico de desarrollo del pueblo weenhayek 2011–2015, http://www.scribd.com/doc/144167874/Plan-­Estrategico-­de-­Desarrollo-­Pueblo-­ Weenhayek-­2 011–2015-­O RCAWETA-­C entro-­E studios-­y -­P royectos-­S RL-­B G-­ Bolivia. Ōta, T. (1987), 中國語歷史文法 [A Historical Grammar of Modern Chinese], Beijing: Beijing Daxue Chubanshe. Otaina, G. A. (1978), Kachestvennye glagoly v nivxskom jazyke, Moscow: Nauka. Ott, D. (2011), ‘Diminutive-­formation in German: spelling out the classifier analysis’, German Linguistics, 14, pp. 1–46. Palmović, M. (2007), ‘The acquisition of diminutives in Croatian’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­ Linguistic Perspective, Amsterdam: Benjamins, pp. 73–88. Pan, W., Ye, B. and Han, Y. (2004), 漢語的構詞法研究 [Research on Word Formation in Chinese], Shanghai: Huadong Shifan Daxue Chubanshe. Panfilov, V. Z. (1962), Grammatika nivxskogo jazyka. Tom 1, Moscow; Leningrad: Nauka. Panfilov, V. Z. (1965), Grammatika nivxskogo jazyka. Tom 2, Moscow; Leningrad: Nauka.

686

References

Panocová, R. (2009), ‘On sound symbolism in evaluative morphology of East Slavic ­languages’, Sučasni doslidžennja z inozemnoji filolohiji, 7, pp. 324–8. Panocová, R. (2011), ‘Evaluative suffixes in Slavic’, Bulletin of the Transilvanian University of Brasov Series IV: Philology and Cultural Studies, 53: 1, pp. 175–82. Parsons-­Yazzie, E. and Speas, M. (2008), Dine Bizaad Binahoo’aah/Rediscovering the Navajo Language, Flagstaff: Salina Bookshelf. Payne, D. (1998), ‘Maasai gender in typological perspective’, Studies in African Linguistics, 27: 2, pp. 159–75. Payne, T. E. (1997), Describing Morphosyntax: A Guide For Field Linguists, Cambridge: Cambridge University Press. Pecher, D. and Zwaan, R. (2005), ‘Introduction to grounding cognition’, in D. Pecher and R. Zwaan (eds), Grounding Cognition: The Role of Perception and Action in Memory, Language and Thinking, Cambridge: Cambridge University Press, pp. 1–7. Pensalfini, R. (1997), Jingulu Grammar, Dictionary, and Texts, MIT: PhD dissertation. Pensalfini, R. (2002), ‘Vowel harmony in Jingulu’, Lingua, 112, pp. 561–86. Pensalfini, R. (2003), A Grammar of Jingulu: An Aboriginal Language of the Northern Territory, Canberra: Pacific Linguistics. Pensalfini, R. (2014), ‘Verbs as spatial deixis markers in Jingulu’, in D. Guillemin, R. Pensalfini and M. Turpin (eds), Language Description Informed by Theory, Amsterdam: Benjamins, pp. 123–52. Pentland, D. H. (1975), ‘Diminutive consonant symbolism in Algonquian’, in W. Cowan (ed.), Papers of the Sixth Algonquian Conference, 1974, Ottawa: National Museums of Canada, pp. 237–52. Pérez Saldanya, M., Sifre Gómez, M. and Todolí Cervera, J. (2004), Morfologia catalana, Barcelona: Editorial UOC. Perrott, D. V. (1951), Teach Yourself Books: Swahili, London: Hodder. Perry, J. R. (2007), ‘Morphology of Persian’, in A. S. Kaye (ed.), Morphologies of Asia and Africa, Winona Lake, IN: Eisenbrauns, vol. 2, pp. 975–1019. Pete, S. (1977), Diqelas Tukda: Transcribed and Edited by James Kari, Fairbanks: Alaska Native Language Center. Pete, S. (1989), ‘The hunting dog: a Dena’ina tale of subsistence values’, trans. B. Pete, J. Kari and J. Fall, Alaska Fish & Game, November–December, pp. 3–7. Pete, S. (2003), ’1919 trip to Susitna station’, in J. Kari and J. A. Fall (eds), Shem Pete’s Alaska: The Territory of the Upper Cook Inlet Dena’ina, Fairbanks: University of Alaska Press, pp. 160–8. Petersen, W. (1910), Greek Diminutives in -­ion, Weimar: R. Wagner Sohn. Petitta, G., Di Renzo, A., Chiari, I. and Rossini, P. (2013), ‘Sign language representation: new approaches to the study of Italian Sign Language (LIS)’, in L. Meurant, A. Sinte, M. Van Herreweghe and M. Vermeerbergen (eds), Sign Language Research, Uses and Practices: Crossing Views on Theoretical and Applied Sign Language Linguistics, Berlin: De Gruyter/Ishara Press, pp. 137–58. Petrova, T. I. (1967), Jazyk orokov ul’ta, Leningrad: Nauka. Pfau, R. and Quer, J. (2010), ‘Nonmanuals: their grammatical and prosodic roles’, in D. Brentari (ed.), Sign Languages, Cambridge: Cambridge University Press, pp. 381–402. Pfau, R., Steinbach, M. and Woll, B. (eds) (2012), Sign Language: An International Handbook, Berlin: De Gruyter.

References 687 Pharies, D. (2002), Diccionario etimológico de los sufijos españoles y de otros elementos finales, Madrid: Gredos. Pike, K. (1948), Tone Languages: A Technique for Determining the Number and Types of Pitch Contrasts in a Language, with Studies in Tonemic Substitution and Fusion, Ann Arbor: University of Michingan Press. Pike, K. and Pike, E. (1947), ‘Immediate constituents of Mazatec syllables’, International Journal of American Linguistics, 13, pp. 78–91. Pincus, D. (2012), Growing Up Brave: Expert Strategies for Helping Your Child Overcome Fear, Stress, and Anxiety, New York: Little, Brown. Pinto, I. (forthcoming), ‘Sardinian’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Pizzuto, E. and Corazza, S. (1996), ‘Noun morphology in Italian Sign Language (LIS)’, Lingua, 98, pp. 169–96. Plag, I., Dalton-­Puffer, C. and Baayen, R. H. (1999), ‘Morphological productivity across speech and writing’, Journal of English Language and Linguistics, 3: 2, pp. 209–28. Plank, F. (1981), Morphologische (Ir-­)Regularitäten: Aspekte der Wortstrukturtheorie, Tübingen: Narr. Plank, F. (1994), ‘Inflection and derivation’, in R. E. Asher and J. M. Y. Simpson (eds), Encyclopedia of Language and Linguistics. Vol. 3, Oxford: Pergamon Press, pp. 1671–8. Plank, F., Filimonova, E., Mayer, T. and Mayorava, T. (n.d.), The Universals Archive, http://typo.uni-­konstanz.de/archive. Pochkhua, B. (1974), ქართული ენის ლექსიკოლოგია [Lexicology of Georgian Language], Tbilisi: tbilisis universit’et’is gamomcemloba. Ponsonnet, M. (2014), The Language of Emotions: The Case of Dalabon (Australia), Amsterdam: Benjamins. Ponsonnet, M. (2015), ‘Nominal subclasses in Dalabon (South Western Arnhem Land)’. Australian Journal of Linguistics, 35: 1, pp. 1–52. Potts, C. (2007), ‘The expressive dimension’, Theoretical Linguistics, 33, pp. 165–97. Poulos, G. (1990), A Linguistic Analysis of Venda, Pretoria: Via Afrika. Poulos, G. and Msimang, C. T. (1998), A Linguistic Analysis of Zulu, Cape Town: Via Afrika. Pradilla, M. À. (ed.) (2011), Informe sobre la situació de la llengua catalana, Barcelona: Observatori de la Llengua Catalana. Prieto, V. (2005), Spanish Evaluative Morphology: Pragmatic, Sociolinguistic, and Semantic Issues, University of Florida: PhD dissertation, http://etd.fcla.edu/UF/ UFE0010940/prieto_v.pdf. Protassova, E. and Voeikova, E. (2007), ‘Diminutives in Russian at the early stages of acquisition’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­Linguistic Perspective, Amsterdam: Benjamins, pp. 43–72. Puglielli, A. (1984a), ‘La derivazione nominale in somalo’, in A. Puglielli (ed.), Aspetti morfologici, lessicali e della focalizzazione, Rome: MAE, Dipartimento per la Cooperazione allo Sviluppo, pp. 1–52. Puglielli, A. (ed.) (1984b), Aspetti morfologici, lessicali e della focalizzazione, Rome: MAE, Dipartimento per la Cooperazione allo Sviluppo. Puglielli, A. and Mansuur, M. (eds) (2012), Qamuuska Af-­Soomaaliga, Rome: UniTrePress. Puglielli, A. and Siyaad, C. M. (1984), ‘La flessione del nome’, in A. Puglielli (ed.),

688

References

Aspetti morfologici, lessicali e della focalizzazione, Rome: MAE, Dipartimento per la Cooperazione allo Sviluppo, pp. 53–112. Quesada Pacheco, M. A. (1995), Hablemos boruca (Chá din̈ div tégat tegrá), San José: Ministerio de Educación Pública. Rainer, F. (1989), ‘Appunti sui diminutivi italiani in -­etto e -­ino’, in M. Berretta, P. Molinelli and A. Valentini (eds), Parallela 4. Morfologia: Atti del V Incontro Italo-­Austriaco della Società di Linguistica Italiana a Bergamo, 2–4 ottobre 1989/Morphologie: Akten des 5. Österreichisch-­ Italienischen Linguistentreffens in Bergamo, 2.-­ 4. Oktober 1989, Tübingen: Narr, pp. 207–18. Rainer, F. (1993), Spanische Wortbildungslehre, Tübingen: Niemeyer. Rainer, F. (1996), ‘Inflection inside derivation’, Yearbook of Morphology 1995, pp. 83–91. Rainer, F. (2003), ‘Studying restrictions on patterns of word-­formation by means of the Internet’, Italian Journal of Linguistics, 15: 1, pp. 131–40. Rainer, F. (2005), ‘Semantic change in word formation’, Linguistics, 43: 2, pp. 415–41. Rainer, F. (2010), Carmens Erwerb der deutschen Wortbildung, Vienna: Österreichischen Akademie der Wissenschaften. Ralli, A. (2000), ‘Α feature-­based analysis of Greek nominal inflection’, Glossolojia, 11–12, pp. 201–27. Ralli, Α. (2005), Mορφολογία [Morphology], Αthens: Patakis. Ralli, A. and Melissaropoulou, D. (2008), ‘Ο Υποκορισμός στη Διαλεκτική Ποικιλία της Κοινής Νεοελληνικής’ [Diminution in Modern Greek dialectal variation], in K. Dimadis (ed.), Proceedings of the 3rd European Conference of the European Society of Modern Greek Studies (Bucharest, 2–4 June 2006), Athens: Ellinika Grammata, pp. 569–82. Rappaport Hovav, M. (2008), ‘Lexicalised meaning and the internal temporal structure of events’, in S. Rothstein (ed.), Theoretical and Crosslinguistic Approaches to the Semantics of Aspect, Amsterdam: Benjamins, pp. 13–42. Rau, D. V. (1994), ‘Lexical iconicity in Tagalog’, Philippine Journal of Linguistics, 24: 2, pp. 17–28. Rau, D. V. (2005), ‘Iconicity, tense, aspect, and mood morphology in Yami’, Concentric: Studies in Linguistics, 31: 1, pp. 65–94. Rau, D. V. and Dong, M.-­ N. (2005), ‘Yami reduplication’, Concentric: Studies in Linguistics, 31: 2, pp. 57–87. Rau, D. V. and Dong, M.-­N. (2006), Yami Texts with Reference Grammar and Dictionary: Language and Linguistics, Taipei: Institute of Linguistics, Academia Sinica. Rau, D. V. and Dong, M.-­N. (2010), ‘Serves you right: ka-­ as an attitudinal marker in Yami’, paper presented at the 4th Conference on Language, Discourse, and Cognition. (CLDC 2010), 1–2 May, Taipei: National Taiwan University. Rau, D. V., Wang, C.-­ C. and Chang, H.-­ H. (2012), ‘Investigating motion events in Austronesian languages’, Oceanic Linguistics, 51: 1, pp. 1–17. Rau, V., Dong, M.-­N. and Chang, H.-­H with D. E. Rau and G. A. Rau (eds) (2012), Yami (Tao) Dictionary, Taipei: National Taiwan University Press. Rau, V., Wu, Y.-­H. and Yang, M.-­C. (forthcoming), ‘A corpus approach to Yami emotion’, in E. Zeitoun, S. Teng and J. Wu (eds), A Festschrift to Lillian Huang, Taipei: Academia Sinica. Ravid, D., Dressler, W. U., Nir-­Sagiv, B., Korecky-­Kröll, K., Souman, A., Rehfeldt, K., Laaha, L., Bertl, J., Basbøll, H. and Gillis, S. (2008), ‘Core morphology in child

References 689 directed speech: cross-­linguistic corpus analyses of noun plurals’, in H. Behrens (ed.), Corpora in Language Acquisition Research, Amsterdam: Benjamins, pp. 25–60. Rdo.dgon Gsaṅ.bdag Rdo.rje, L. and Ross, A. (2002) Źe.saḥi lag.deb blo.gsar dgaḥ.ston, Lhasa: Bod.ljoṅs mi.dmaṅs dpe.skrun.khaṅ. Reichard, G. (1951), Navaho Grammar, New York: J. J. Augustin. Reichmuth, W. (1956), Die lateinischen Gentilicia und ihre Beziehungen zu den römischen Individualnamen, Universität Zürich: Phd dissertation. Reynoso Noverón, J. (2005), ‘Procesos de gramaticalización por subjetivización: el uso del diminutivo en español’, in D. Eddington (ed.), Selected Proceedings of the 7th Hispanic Linguistics Symposium, Somerville, MA: Cascadilla, pp. 79–86. Ricca, D. (2005), ‘Cumulative exponence involving derivation: some patterns for an uncommon phenomenon’, in W. U. Dressler, D. Kastovsky, O. E. Pfeiffer and F. Rainer (eds), Morphology and its Demarcations, Amsterdam: Benjamins, pp. 197–213. Rice, K. (1989), A Grammar of Slave, Berlin: De Gruyter. Rice, K. (2000), Morpheme Order and Semantic Scope, Amsterdam: Benjamins. Rice, K. (2014), ‘Athabaskan’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 669–701. Rice, S. (2012), ‘“Our language is very literal”: figurative expression in Dene Sųłiné (Athapaskan)’, in A. Idström and E. Piirainen (eds), Endangered Metaphors, Amsterdam: Benjamins, pp. 21–75. Rifón Sánchez, A. (1997), Pautas semánticas para la formación de verbos en español mediante sufijación, Santiago de Compostela: Universidade de Santiago de Compostela. Rifón Sánchez, A. (1998), ‘La derivación verbal apreciativa en español’, ELUA, 12, pp. 211–26. de Rijk, R. P. G. (2008), Standard Basque: A Progressive Grammar, Cambridge, MA: MIT Press. Roberts, J. R. (1997), ‘Switch-­reference in Papua New Guinea: a preliminary survey’, in A. Pawley (ed.), Papers in Papuan Linguistics, Canberra: Research School of Pacific and Asian Studies, Australian National University, pp. 101–241. Roelcke, T. (1999), Fachsprachen, Berlin: Schmidt. Romaine, S. (1984), ‘The status of sociological models and categories in explaining language variation’, Linguistische Berichte, 90, pp. 25–38. Romeo, O. (1991), Dizionario dei segni: la lingua dei segni in 1400 immagini, Bologna: Zanichelli. Romiti, S. (1998), ‘The mural writings of the young in Rome’, in J. K. Androutsopoulos and A. Scholz (eds), Jugendsprache – langue des jeunes – Youth Language, Frankfurt: Lang, pp. 280–304. Rosch, E. (1983), ‘Prototype classification and logical classification: the two systems’, in E. Scholnick (ed.), New Trends in Conceptual Representation: Challenges to Piaget’s Theory?, Hillsdale, NJ: Lawrence Erlbaum, pp. 73–86. Ross, B. B. (2011), Prosody and Grammar in Dalabon and Kayardild, University of Melbourne: PhD dissertation. Round, E. R. (2009), Kayardild Morphology, Phonology and Morphosyntax, Yale University: PhD dissertation. Round, E. R. (2011), ‘Word final phonology in Lardil: implications of an expanded data set’, Australian Journal of Linguistics, 31, pp. 327–50. Round, E. R. (2013), Kayardild Morphology and Syntax, Oxford: Oxford University Press.

690

References

Rubino, C. (2000), Ilocano Dictionary and Grammar, Honolulu: University of Hawai’i Press. Rubino, C. (2006), Intensive Tausug: A Pedagogical Grammar of the Language of Jolo, Philippines, Springfield: Dunwoody Press. Rudolph, E. (1990), ‘Portuguese diminutives as special indicators of emotions’, Grazer Linguistische Studien, 33/4, pp. 253–66. Ruf, B. (1996), Augmentativbildungen mit Lehnpräfixen, Heidelberg: Winter. Ruffolo, R. (2004), Topics in the Morpho-­ Syntax of Ibaloy, Northern Philippines, Australian National University: PhD dissertation. Ruiz de Mendoza Ibáñez, F. J (2000), ‘El modelo cognitivo idealizado de tamaño y la formación de aumentativos y diminutivos en español’, Revista Española de Lingüística Aplicada, 1, pp. 355–74. Rūķe-­Draviņa, V. (1959), Diminutive im Lettischen, Lund: Håkan Ohlsoons Boktryckeri. Rull, X. (2004), La formació de mots: qüestions de normativa, Vic: Eumo. Ryle, G. (1949), The Concept of Mind, Chicago: University of Chicago Press. Sabatini, F. and Coletti, V. (1999) (eds), DISC: Dizionario italiano Sabatini Coletti, Florence, Giunti. Sabol, J. (1989), Syntetická fonologická teória, Bratislava: Jazykovedný ústav Ľudovíta Štúra Slovenskej akadémie vied. Sadeghi, A. A. (1991–3), ‘Shiveha va Emkanat-­e Vajesazi dar Zaban-­e Farsi-­ye Moaser (1–12)’, in Majale-­ye Nashr-­e Danesh, vols. 64–73, 74–7, 79–80, Tehran: Markaz-­e Nashr-­e Daneshgahi. Sadock, J. M. (1980), ‘Noun incorporation in Greenlandic: a case of syntactic word formation’, Language, 56, pp. 300–19. Sadock, J. M. (2003), A Grammar of Kalaallisut (West Greenlandic Inuttut), Munich: Lincom Europa. Saeed, J. I. (1993), Somali Reference Grammar, 2nd rev. edn, Kensington, MD: Dunwoody Press. Saeed, J. I. (1999), Somali, Amsterdam: Benjamins. SAG (Svenska Akademiens Grammatik) (1999), Stockholm: Norstedts Ordbok. Sagart, L. (2004), ‘Vestiges of archaic Chinese derivational affixes in Modern Chinese dialects’, in H. Chappell (ed.), Chinese Grammar: Synchronic and Diachronic Perspectives, Oxford: Oxford University Press, pp. 123–42. Sakel, J. (2004), A Grammar of Mosetén, Berlin: De Gruyter. Sandler, W. (2012), ‘Visual prosody’, in R. Pfau, M. Steinbach and B. Woll (eds), Sign Language: An International Handbook, Berlin: De Gruyter, pp. 55–76. Sandler, W. and Lillo-­ Martin, D. (2006). Sign Language and Linguistic Universals, Cambridge: Cambridge University Press. Sapir, E. (1915), ‘Abnormal types of speech in Nootka’, Geological Survey, Memoir 62, Anthropological Series n. 5, Ottawa: Government Printing Bureau (reprinted in V. Golla (ed.) (1991), The Collected Works of Edward Sapir. Vol. VI: American Indian Languages 2, Berlin: De Gruyter, pp. 357–79). Sapir, E. (1921), Language: An Introduction to the Study of Speech, New York: Harcourt, Brace. Sapir, E. (1929), ‘A study in experimental symbolism’, Journal of Experimental Psychology, 12, pp. 225–39. Saulwick, A. (2003a), Aspects of the Verb in Rembarrnga, a Polysynthetic Language of Northern Australia: Grammatical Description, Texts and Dictionary, University of Melbourne: PhD dissertation.

References 691 Saulwick, A. (2003b), A First Dictionary of Rembarrnga, Maningrida: Maningrida Arts and Culture, Bawinanga Aboriginal Corporation. Saulwick, A. (2007) ‘Rembarrnga polysynthesis in cross-­ linguistic perspective’, in M. Miestamo and B. Wälchli (eds), New Challenges in Typology, Berlin: De Gruyter. Saveljeva, V. N. and Taksami, C. M. (1970), Nivxsko-­russkij slovar’, Moscow: Sovetskaja enciklopedija. Savickienė, I. (2001), ‘The role of diminutives in Lithuanian child language acquisition’, Linguistica Baltica, 9, pp. 109–18. Savickienė, I. (2003), The Acquisition of Lithuanian Noun Morphology, Vienna: Österreichischen Akademie der Wissenschaften. Savickienė, I. (2007), ‘Form and meaning of diminutives in Lithuanian child language’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­ Linguistic Perspective, Amsterdam: Benjamins, pp. 13–41. Savickienė, I. and Dressler, W. U. (2007a), ‘Conclusions’, in I. Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­Linguistic Perspective, Amsterdam: Benjamins, pp. 343–9. Savickienė, I. and Dressler, W. U. (eds) (2007b), The Acquisition of Diminutives: A Cross-­ Linguistic Perspective, Amsterdam: Benjamins. Savickienė, I., Dressler, W. U., Barcza, V., Bodor, P., Ketrez, N., Korecky-­Kröll, K., Palmović, M., Stephany, U. and Thomadaki, E. (2009), ‘Diminutives as pioneers of derivational and inflectional development: a cross-­linguistic perspective’, in M. Bertuccelli Papi, A. Bertacca and S. Bruti (eds), Threads in the Complex Fabric of Language: Linguistic and Literary Studies in Honour of Lavinia Merlini Barbaresi, Ghezzano: Felici Editore, pp. 339–54. Scalise, S. (1984), Generative Morphology, Dordrecht: Foris. Scalise, S. (1986), Generative Morphology, 2nd edn, Dordrecht: Foris. Scalise, S. (1988a), ‘Inflection and derivation’, Linguistics, 26, pp. 561–81. Scalise, S. (1988b), ‘The notion of “head” in morphology’, Yearbook of Morphology 1, pp. 229–45. Scalise, S. (1994), Morfologia, Bologna: Il Mulino. Scatton, E. A. (1984), A Reference Grammar of Modern Bulgarian, Columbus, OH: Slavica. Schachter, P. and Otanes, F. T. (1972), Tagalog Reference Grammar, Berkeley: University of California Press. Schadeberg, T. C. (2003), ‘Derivation’, in D. Nurse and G. Philippson (eds), The Bantu Languages, London: Routledge, pp. 71–89. Schneck, J. M. (1965), ‘Macropsia’, American Journal of Psychiatry, 121, pp. 1123–4. Schneider, C. (2008), ‘The partitive marker in Abma’, Anthropological Linguistics, 50: 2, pp. 148–73. Schneider, E. R. (1936), Kratkii udeisko-­russkii slovar, Leningrad: Uchpedgiz. Schneider, K. P. (2003), Diminutives in English, Tübingen: Niemeyer. Schneider, K. P. (2013), ‘The truth about diminutives, and how we can find it: some theoretical and methodological considerations’, SKASE Journal of Theoretical Linguistics, 10: 1, http://www.skase.sk/Volumes/JTL22/pdf_doc/08.pdf. Schoen, P. (1959), Initiation à la langue kabyle, Alger: Centre d’Études Régionales de Kabylie. Schönig, C. (1998), ‘Azerbaijanian’, in L. Johanson and É. Á. Csató (eds), Turkic Languages, London: Routledge, pp. 248–60.

692

References

Schultze-­Berndt, E. (2012), ‘Pluractional posing as progressive’, Australian Journal of Linguistics, 32, pp. 7–39. Schwarzkopf, D., Song, C. and Rees, G. (2010), ‘The surface area of human V1 predicts the subjective experience of object size’, Nature Neuroscience, 14, pp. 28–30. Schwieger, P. (2006), Handbuch zur Grammatik der klassischen tibetischen Schriftsprache, Halle (Saale): International Institute for Tibetan and Buddhist Studies. Seitz, J. A. (1997), ‘Metaphor, symbolic play, and logical thought in early childhood’, Genetic, Social, and General Psychology Monographs, 123: 4, pp. 373–91. Seitz, J. A. (2000), ‘The bodily basis of thought’, New Ideas in Psychology: An International Journal of Innovative Theory in Psychology, 18: 1, pp. 23–40. Seitz, J. A. (2001), ‘A cognitive-­perceptual analysis of projective tests’, Perceptual and Motor Skills, 93, pp. 505–22. Seitz, J. A. and Beilin, H. (1987), ‘The development of comprehension of physiognomic metaphor in photographs’, British Journal of Developmental Psychology, 5, pp. 321–31. Shanidze, A. (1980), ქართული ენის გრამატიკის საფუძვლები, [Foundations of Georgian Grammar], vol. 3. Tbilisi: sakartvelos ssr mecnierebata ak’ademia. Shay, E. (2014), ‘Afro-­Asiatic’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 573–90. Shetler, J. (1976), Notes on Balangao Grammar, Huntington Beach: SIL. Shibatani, M. (2006), ‘Honorifics’, in K. Brown (ed.), Encyclopedia of Language and Linguistics, 2nd edn, Oxford: Elsevier, pp. 381–390. Shiraishi, H. (2006), Topics in Nivkh Phonology, University of Groningen: Groningen Dissertations in Linguistics 61. Shonia, A. (2011), ‘კნინობითი სახელები მეგრულ-­ლაზურში’ [Diminutives in Megrelian and Laz], saenatmecniero dziebani, 31, pp. 209–14. Siegel, J. (1987), Language Contact in a Plantation Environment: A Sociolinguistic History of Fiji, Cambridge: Cambridge University Press. Sifianou, M. (1992), ‘The use of diminutives in expressing politeness: Modern Greek versus English’, Journal of Pragmatics, 17, pp. 155–73. Siller-­Runggaldier, H. (forthcoming), ‘Ladin’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Simon, H. J. and Wiese, H. (2011), ‘What are exceptions? And what can be done about them?’, in H. J. Simon and H. Wiese (eds), Expecting the Unexpected: Exceptions in Grammar, Berlin: De Gruyter, pp. 3–31. Simpson, J. (1991), Warlpiri Morpho-­Syntax: A Lexicalist Approach, Dordrecht: Kluwer. Simpson, J. (1992), ‘Notes on a manuscript dictionary of Kaurna’, in T. Dutton, M. Ross and D. Tryon, (eds), The Language Game: Papers in Memory of Donald C. Laycock, Canberra: Australian National University, pp. 409–15. Smith, C. (1991), The Parameter of Aspect, Dordrecht: Kluwer. Smith, R. (1989), Aristotle: Prior Analytics, Indianapolis: Hackett. Söderbergh, R. (1971), Svensk ordbildning, 2nd edn, Stockholm: Läromedelsförlagen. Soida, E. (2009), Vārddarināšana, Riga: LU Akadēmiskais apgāds. Sokolová, M., Moško, G., Šimon, F. and Benko, V. (1999), Morfematický slovník slovenčiny, Prešov: Náuka. Sotiropoulos, D. (1972), Noun Morphology of Modern Demotic Greek, The Hague: Mouton.

References 693 Soukka, M. (2000), A Descriptive Grammar of Noon, a Cangin Language of Senegal, Munich: Lincom Europa. Spalding, A. (1998), Inuktitut: A Multi-­Dialectal Outline Dictionary, Iqaluit: Nunavut Arctic College. Sperber, D. and Wilson, D. (1995), Relevance: Communication and Cognition, Oxford: Blackwell. Spitzer, L. (1921), ‘Das Suffix -­one im Romanischen’, in E. Gamillscheg and L. Spitzer (eds), Beiträge zur Romanischen Wortbildungslehre, Geneva: Olschki, pp. 183–205. Staalsen, P. (1965), Iatmul Grammar Sketch, MS, Ukarumpa: SIL. Staalsen, P. (1972), ‘Clause relationships in Iatmul’, Pacific Linguistics. Series A – Occasional Papers, 31, pp. 45–69. Staalsen, P. (n.d.), Iatmul Verbs, MS, Ukarumpa: SIL. Stairs Kreger, G. A. and Scharfe de Stairs, E. F. (1981), Diccionario Huave, Mexico: ILV. Staverman, W. H. (1953), ‘Diminutivitis Neerlandica’, De Gids, 116: 7, pp. 407–19. Ştefănescu, I. (1992), ‘On diminutive suffixes’, Folia Linguistica, 26, pp. 339–56. Štekauer, P. (1995), An Encyclopaedia of English Linguistics, Prešov: Slovacontact. Štekauer, P. (2011), ‘On some issues of diminutives from a cross-­linguistic perspective’, in S. D. Schmid, P. Gévaudan, W. Mihatsch and W. Waltereit (eds), Rahmen des Sprechens, Narr: Tubingen, pp. 425–35. Štekauer, P. (forthcoming), ‘The delimitation of derivation and inflection’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Štekauer, P., Valera, S. and Körtvélyessy, L. (2012), Word-­Formation in the World’s Languages: A Typological Survey, Cambridge: Cambridge University Press. Stephany, U. (1997), ‘Diminutives in early Greek’, in W. U. Dressler (ed.), Studies in Pre-­ and Protomorphology, Vienna: Österreichischen Akademie der Wissenschaften, pp. 145–56. Stephens, E. (1889), ‘The Aborigines of Australia: being personal recollections of those tribes which once inhabited the Adelaide plains of South Australia’, read before the Royal Society of New South Wales, 2 October. Stevens, S. S. (1946), ‘On the theory of scales of measurement’, Science, 103, pp. 677–80. Štolc, J. (1958), ‘K morfológii deminutív v slovenčine’, Jazykovedné Štúdie, III, pp. 19–81. Stolz, T. (1994), Grammatikalisierung und Metaphorisierung, Bochum: Brockmeyer. Strodach, G. K. (1933), ‘Latin diminutives in -­ello/a and -­illo/a’, Language Monographs, 14. Stump, G. T. (1993), ‘How peculiar is evaluative morphology?’, Journal of Linguistics, 29, pp. 1–36. Suárez, A. (1983), Mesoamerican Indian Languages, Cambridge: Cambridge University Press. Sun, C. (2006), Chinese: A Linguistic Introduction, Cambridge: Cambridge University Press. Sunik, O. P. (1985), Ul’chskij jazyk, Leningrad: Nauka. Supalla, T. (1986), ‘The classifier system in American sign language’, in C. Craig (ed.), Noun Classes and Categorization, Amsterdam: Benjamins, pp. 181–214. Sutton, V. (1995), Lessons in SignWriting: Textbook and Workbook, La Jolla: Deaf Action Committee for Sign Writing. Sutton-­Spence, R. and Woll, B. (1999), The Linguistics of British Sign Language: An Introduction, Cambridge: Cambridge University Press.

694

References

Svelmoe, G. and Svelmoe, T. (1974), Notes on Mansaka Grammar, Huntington Beach: SIL. Szymanek, B. (1988), Categories and Categorization in Morphology, Lublin: Redakcja Wydawnictw Katolickiego Uniwersytetu Lubelskiego. Szymanek, B. and Derkach, T. (2005), ‘Constraints on the derivation of double diminutives in Polish and Ukrainian’, Studies in Polish Linguistics, 2, pp. 93–112. Tabataba’i, A. (2003), Esm va Sefat-­e Morakkab dar Zaban-­e Farsi, Tehran: Markaz-­e Nashr-­e Daneshgahi. Tabataba’i, A. (2007), ‘Sarf-­e Zaban-­e Farsi’, Bukhara Journal of Culture and Art, 63, pp. 227–41. Tabataba’i, A. (2012), ‘Moghayeseye Radeshenakhti-­ye Esm-­e Morakkab dar Zaban-­e Farsi va Englisi’, Journal of Comparative Linguistics, 1, pp. 1–17. Taine-­ Cheikh, C. (2002), ‘Morphologie et morphogenèse des diminutifs en zenaga (berbère de Mauritanie)’, in K. Nait-­Zerrad (ed.), Articles de linguistique berbère: mémorial Werner Vycichl, Paris: L’Harmattan, pp. 427–54. Taylor, J. (1995), Linguistic Categorization: Prototypes in Linguistic Theory, New York: Oxford University Press. Teichelmann, C. G. (1857), Dictionary of the Adelaide Dialect, MS 4vo. pp. 99 (with double columns), No. 59, Bleek’s Catalogue of Sir George Grey’s Library dealing with Australian languages, South African Public Library. Teichelmann, C. G. (1858), Of the Verb, MS 8vo. pp. 3. No. 57, Bleek’s Catalogue of Sir George Grey’s Library dealing with Australian languages, South African Public Library. Teichelmann, C. G. and Schürmann, C. W. (1840), Outlines of a Grammar, Vocabulary, and Phraseology, of the Aboriginal Language of South Australia, Spoken by the Natives In and for Some Distance Around Adelaide, Adelaide: Published by the authors at the native location. ten Hacken, P. (2014), ‘Delineating derivation and inflection’, in R. Lieber and P. Štekauer (eds), The Oxford Handbook of Derivational Morphology, Oxford: Oxford University Press, pp. 10–25. Tenenbaum, J. M. (1978), Morphology and Semantics of the Dena’ina Verb, Columbia University: PhD dissertation. Tenenbaum, J. M. (ed.) (2006), Dena’ina Sukdu’a: Traditional Stories of the Tanaina Athabaskans, Fairbanks: Alaska Native Language Center. Tenny, C. (1994), Aspectual Roles and the Syntax–Semantics Interface, Dordrecht: Kluwer. Terestyéni, T. (1995), ‘Styles of knowledge and greetings habits in Hungarian’, International Journal of the Sociology of Language, 111, pp. 47–55. Teuber, O. (1998), ‘Fasel beschreib erwähn: Der Inflektiv als Wortform des Deutschen’, Germanistische Linguistik, 141–2, pp. 6–26. Thomadaki, E. and Stephany, U. (2007), ‘Diminutives in Greek child language’, in I.  Savickienė and W. U. Dressler (eds), The Acquisition of Diminutives: A Cross-­ Linguistic Perspective, Amsterdam: Benjamins, pp. 89–123. Thomas, F. (1940), ‘Le suffixe latin -­aster/-­astrum’, Revue des Études Anciennes, 42, pp. 520–8. Thorell, O. (1981), Svensk ordbildningslära, Stockholm: Esselte Studium. Thurgood, G. and LaPolla, R. J. (eds) (2003), The Sino-­Tibetan Languages, London: Routledge. Tiefenbach, H. (1987), ‘-­chen und -­lein: Überlegungen zu Problemen des sprachgeog-

References 695 raphischen Befundes und seiner sprachhistorischer Deutung. Mit fünf Karten’, Zeitschrift für Dialektologie und Linguistik, 54: 1, pp. 2–27. Tindale, N. B. (n.d.), Assorted Papers: Kaurna Vocabulary Card File, Adelaide: South Australian Museum, Tindale Collection. Tolskaya, M. (2012), ‘The morphology of Udihe ideophones’, in A. L. Malchukov and L. J. Whaley (eds), Recent Advances in Tungusic Linguistics, Wiesbaden: Harrassowitz, pp. 89–122. Tomasello, M. (1999), The Cultural Origins of Human Cognition, Cambridge, MA: Harvard University Press. Topadze Gäumann, M. (2011), ‘The expression of evidentiality between lexicon and grammar: a case study from Georgian’, Linguistic Discovery, 9: 2, pp. 122–38. Topuria, V. (1979), ‘კნინობითი ქართველურ ენებში’ [Diminutives in Kartvelian languages], in V. Topuria, Šromebi, Tbilisi: mecniereba, pp. 102–31. Tournadre, N. (2008), ‘Arguments against the concept of “conjunct”/“disjunct” in Tibetan’, in B. Huber et al. (eds), Chomolangma, Demawend und Kasbek: Festschrift für Roland Bielmeier zu seinem 65. Geburtstag, Halle (Saale): International Institut for Tibetan and Buddhist Studies, vol. 1, pp. 281–308. Tovena, L. M. (2010), ‘Pluractionality and the unity of the event’, in M. Aloni and K. Schulz (eds), Amsterdam Colloquium 2009 LNAI 6042, Heidelberg: Springer, pp. 465–73. Tovena, L. M. (2011a), ‘When small is many in the event domain’, Lexis 6: Diminutives and Augmentatives in the Languages of the World, pp. 41–58, http://lexis.univ-­ lyon3.fr/IMG/pdf/Lexis_6.pdf. Tovena, L. M. (2011b), ‘The imperfect measure of internally plural events’, in G. Bezhanishvili, S. Lobner, V. Marra and F. Richter (eds), Logic, Language and Computation: Eighth Tbilisi Symposium, Berlin: Springer, pp. 301–21. Tovena, L. M. (2011c), ‘Issues in the formation of verbs by evaluative suffixation’, in G. Massariello Merzagora and S. Dal Maso (eds), I luoghi della traduzione: le interfacce. Atti del XLIII Congresso Internazionale di Studi della Società di Linguistica Italiana, vol. 2, Rome: Bulzoni, pp. 913–25. Tovena, L. M. and Donazzan, M. (2008), ‘On ways of repeating’, Recherches Linguistiques de Vincennes, 37, pp. 85–112. Tozzer, A. M. (1921), A Maya Grammar, Cambridge, MA: Harvard University. Traugott, E. C. (1995), ‘Subjectification in grammaticalization’, in D. Stein and S. Wright (eds), Subjectivity and Subjectivisation: Linguistic Perspectives, Cambridge: Cambridge University Press, pp. 31–54. Travis, C. E. (2004), ‘The ethnopragmatics of the diminutive in conversational Colombian Spanish’, Intercultural Pragmatics, 1: 2, pp. 249–74. Tribushinina, E., van den Bergh, H., Kilani-­Schoch, M., Aksu-­Koç, A., Dabašinskienė, I., Hrzica, G., Korecky-­Kröll, K., Noccetti, S. and Dressler, W. U. (forthcoming), ‘The role of explicit contrast in adjective acquisition: a cross-­linguistic longitudinal study of adjective production in spontaneous child speech and parental input’, First Language. Trnková, J. (1991), ‘O substantívnej deminutívnej derivácii v slovenčine’, Slovenská reč, 56: 2, pp. 87–95. Tsao, F. (2004), ‘Semantics and syntax of verbal and adjectival reduplication in Mandarin and Taiwanese Southern Min’, in H. Chappell (ed.), Chinese Grammar: Synchronic and Diachronic Perspectives, Oxford: Oxford University Press, pp. 285–308.

696

References

Tsintsius, V. I. (1947), Ocherk grammatiki evenskogo lamutskogo jazyka, Leningrad: Uchpedgiz. Tsintsius, V. I. (1949), Sravnitel’naja fonetika tunguso-­ man’chzhurskikh jazykov, Leningrad: Uchpedgiz. Tsintsius, V. I. (1982), Negidal’skij jazyk, Leningrad: Nauka. Tughushi, Z. (1992), ‘კნინობითი სახელები გურულში’ [Diminutives in Gurian], iberiul-­k’avk’asiuri enatmecniereba, 31, pp. 132–40. Tupe-­Perú: Marka, Aysha y Qullqa (n.d.), http://txupi.wordpress.com/tupe-­jaqaru. Turolla, C. (2013), ‘Puoi aspettare un minutino?’ Aspetti semantici, pragmatici e neurolinguistici nella morfologia valutativa della lingua italiana, University of Padua: Master’s thesis. Tuskia, M. (2004), სახელთა აფიქსური წარმოება ახალ სალიტერატურო ქართულსა და დიალექტებში [Nominal Affixation in Modern Literary Georgian and in Dialects], Tbilisi: tbilisis universit’et’is gamomcemloba. Ultan, R. (1970), ‘Size–sound symbolism’, Stanford University Working Papers on Language Universals, 3, pp. 1–31. Ultan, R. (1978), ‘Size–sound symbolism’, in J. H. Greenberg, C. A. Ferguson and E. A. Moravcsik (eds), Universals of Human Language, 4 vols., Stanford: Stanford University Press, vol. 3, pp. 525–68. Uray, G. (1952), ‘A Tibetan diminutive suffix’, Acta Orientalia Academiae Scientiarum Hungaricae, 2, pp. 182–220. Uray, G. (1954), ‘Duplication, germination and triplication in Tibetan’, Acta Orientalia Academiae Scientiarum Hungaricae, 4: 1–3, pp. 177–256. Urbutis, V. (1971), ‘Garsų kaita ir žodžių daryba’, Kalbotyra, 23: 1, pp. 71–8. Urbutis, V. (2009), Žodžių darybos teorija, Vilnius: Mokslo ir enciklopedijų leidybos institutas. Vajda, E. (1987), ‘Semantic constraints on formal patterns in Russian grammar’, International Journal of Slavic Linguistics and Poetics, 35–6, pp. 181–97. Vajda, E. (2004), Ket, Munich: Lincom. Vajda, E. (2007), ‘Ket morphology’, in A. S. Kaye (ed.), Morphologies of Asia and Africa, Winona Lake, IN: Eisenbrauns, pp. 1277–325. Vajda, E. (2009), ‘Loanwords in Ket’, in M. Haspelmath and U. Tadmor (eds), Loanwords in the World’s Languages: A Comparative Handbook of Loanword Typology, Berlin: De Gruyter, pp. 471–94. Valdman, A. (1978), Le créole: structure, statut et origine, Paris: Klincksieck. Valentine, J. R. (2001), Nishnaabemwin Grammar, Toronto: University of Toronto Press. Valiullina, Z. M. (1997), ‘Imiya Sush’estvitel’noe’, in M. Z. Zakiyev, F. A. Ganiyev and K. Z. Zinnatullina (eds), Tatarskaya Grammatika II: Morfologiya, Kazan: Akademiya Nauk Tatarstana, pp. 19–63. Valli, C. and Lucas, C. (1992), Linguistics of American Sign Language, Washington, DC: Gallaudet University. van Breugel, S. (2014), A Grammar of Atong, Leiden: Brill. Van de Velde, M. L.O. (2008), A Grammar of Eton, Berlin: De Gruyter. van den Berg, R. (1989), A Grammar of the Muna Language, Dordrecht: Foris. van den Berg, R. (2004), ‘Notes on the southern Muna dialect’, in J. Bowden and N. P. Himmelmann (eds), Papers in Austronesian Subgrouping and Dialectology, Canberra: Pacific Linguistics, pp. 129–69.

References 697 van den Berg, R., with La Ode Sidu (1996), Muna-­English Dictionary, Leiden: KITLV Press. Van der Spuy, A. (1993), ‘Dislocated noun phrases in Nguni’, Lingua, 90: 4, pp. 335–55. Van der Spuy, A. (2001), Grammatical Structure and Zulu Morphology, University of the Witwatersrand: PhD dissertation. Van der Voort, H. (2004), A Grammar of Kwaza, Berlin: De Gruyter. van Driem, G. (2012), ‘The Trans-­Himalayan phylum and its implications for population prehistory’, Communication on Contemporary Anthropology, 5, pp. 135–42. van Gijn, R. (2006), A Grammar of Yurakaré, Nijmegen, Radboud University: PhD dissertation. van Gijn, R. (2014), ‘Reduplication in Yurakaré’, in S. Danielsen, K. Hannß and F. Zúñiga (eds), Word Formation in South American Languages, Amsterdam: Benjamins, pp. 143–61. Vasilevich, G. M. (1940), Ocherk grammatiki evenkijskogo tungusskogo jazyka, Leningrad: Uchpedgiz. Vázquez Cuesta, P. and Da Luz, M. A. M. (1971), Gramática da língua portuguesa, Lisbon: Edições 70. Veerman-­Leichsenring, A. (2004), ‘Popolocan noun classifiers: a reconstruction’, International Journal of American Linguistics, 70: 4, pp. 416–51. Vendler, Z. (1967), Linguistics in Philosophy, Ithaca, NY: Cornell University Press. Verhoeven, E. (2012), ‘Cabécar: a Chibchan language of Costa Rica’, in J. Sakel and T. Stolz (eds), Amerindiana: Neue Perspektiven auf die indigenen Sprachen Amerikas, Berlin: Akademie. Verkuyl, H. J. (1972), On the Compositional Nature of the Aspect, Dordrecht: Reidel. Verkuyl, H. J. (1992), A Theory of Aspectuality, Cambridge: Cambridge University Press. Vermeerbergen, M. (2006), ‘Past and current trends in sign language research’, Language & Communication, 26, pp. 168–92. Versteegh, K. (1984), Pidginization and Creolization: The Case of Arabic, Amsterdam: Benjamins. Vidal, A. and Nercesian, V. (2009), ‘Loanwords in Wichí, a Mataco-­Mataguayan language of Argentina’, in M. Haspelmath and U. Tadmor (eds), Loanwords in the World’s Languages: A Comparative Handbook of Loanword Typology, Berlin: De Gruyter, pp. 1015‒34. Viegas Barros, J. P. (2001), Evidencia del parentesco de las lenguas lule y vilela, Santa Fe: Subsecretaria de Cultura de la Provincia de Santa Fe. Visser, M. (2008), ‘Definiteness and specificity in the isiXhosa determiner phrase’, South African Journal of African Languages, 28: 1, pp. 11–29. Vittrant, A. and Robin, F. (2007), ‘Réduplication dans les langues tibeto-­ birmanes: l’exemple du birman et du tibétain’, Faits de Langue, 29, pp. 77–99. Voeykova, M. D. (1998), ‘Acquisition of diminutives by a Russian child: preliminary observations in connection with the early adjectives’, in S. Gillis (ed.), Studies in the Acquisition of Number and Diminutive Marking, Antwerp Papers in Linguistics 95, pp. 97–113. Vokurková, Z. (2008), Epistemic Modalities in Spoken Standard Tibetan, University of Paris VIII: PhD dissertation. Volek, B. (1987), Emotive Signs in Language and Semantic Functioning of Derived Nouns in Russian, Amsterdam: Benjamins.

698

References

Vollmann, R. (2009), ‘Reduplication in Tibetan’, Grazer Linguistische Studien, 71, pp. 115–34. Volpati, C. (1955), ‘Coppie di nomi di due generi’, Lingua Nostra, 16, pp. 2–5. Vycichl, W. (1961), ‘Diminutiv und Augmentativ im Berberischen’, ZDMG, 111: 2, pp. 243–53. Wagner-­Nagy, B. (forthcoming), ‘Nenets’, in P. O. Müller, I. Ohnheiser, S. Olsen and F. Rainer (eds), Word-­Formation: An International Handbook of the Languages of Europe, Berlin: De Gruyter. Wang, L. (1989), 漢語語法史 [A History of Chinese Grammar], Beijing: Shangwu Yinshuguan. Watson, J. C. E. (2006), ‘Arabic morphology: diminutive verbs and diminutive nouns in San’ani Arabic’, Morphology, 16, pp. 189–204. Weber, M. (1963), Contributions à l’étude du diminutif en français moderne, Université de Zürich: PhD dissertation. Werner, H. (1997), Die ketische Sprache, Wiesbaden: Harrassowitz. Werner, H. (1998), Probleme der Wortbildung in den Jenissej-­Sprachen, Wiesbaden: Harrassowitz. Werner, H. (2005), Die Jenissej-­Sprachen des 18. Jahrhunderts, Wiesbaden: Harrassowitz. Wessén, E. (1971), Svensk språkhistoria. 2: Ordbildningslära, 5th edn, Stockholm: Almqvist & Wiksell. Whaley, L. J., Grenoble, L. A. and Li, F. (1999), ‘Revisiting Tungusic classification from the bottom up: a comparison of Evenki and Oroqen’, Language, 75: 2, pp. 286– 321. Whaley, L. J. and Li, F. (1998) ‘The suffix -­kan in Oroqen’, Studies in Language, 22: 2, pp. 447–71. Wierzbicka, A. (1980), Lingua Mentalis: The Semantics of Natural Language, Sydney: Academic. Wierzbicka, A. (1984), ‘Diminutives and depreciatives: semantic representation for derivational categories’, Quaderni di Semantica: Rivista Internazionale di Semantica Teorica e Applicata, 5: 1, pp. 123–30. Wierzbicka, A. (1986), ‘Italian reduplication: cross-­cultural pragmatics and illocutionary semantics’, Linguistics, 24, pp. 287–315. Wierzbicka, A. (1989), ‘Semantic primitives and lexical universals’, Quaderni di Semantica, 10: 1, pp. 103–21. Wierzbicka, A. (1996), Semantics: Primes and Universals, Oxford: Oxford University Press. Wierzbicka, A. (1999), Emotions across Languages and Cultures: Diversity and Universals, Cambridge: Cambridge University Press. Wiese, H. (2012), Kiezdeutsch: Ein neuer Dialekt entsteht, Munich: Beck. Wilbur, R. B. (2009), ‘Productive reduplication in a fundamentally monosyllabic language’, Language Sciences, 31: 2, pp. 325–42. Wilhelm, A. and Saxon, L. (2010), ‘The “possessed noun suffix” and “possession” in two Northern Athabaskan languages’, invited talk, Workshop on Nominal Dependents, Banff Park Lodge, Banff, 8–9 May. Wilkes, A. (1990), ‘Comments on the so-­called indefinite copulative relative in Zulu’, South African Journal of African Languages, 10: 1, pp. 34–9. Wilkins, D. (1989), Mparntwe Arrernte (Aranda): Studies in the Structure and Semantics of Grammar, Australian National University: PhD dissertation.

References 699 Williams, W. (1840), ‘The language of the natives of South Australia’, South Australian Colonist, 1: 19, pp. 295–6. Wittgenstein, L. (1921), Tractatus Logico-­Philosophicus, London: Routledge & Kegan Paul. Wittgenstein, L. (1974), Philosophical Grammar (trans. A. Kenny), Oxford: Blackwell. Wolfart, H. C. (1973), Plains Cree: A Grammatical Study, Philadelphia: American Philosophical Society. Woll, B. (1981), ‘Question structure in British Sign Language’, in B. Woll, K. Kyle and M. Deuchar (eds), Perspectives on British Sign Language and Deafness, London: Croom Helm, pp. 136–49. Wolvengrey, A. (2011), Semantic and Pragmatic Functions in Plains Cree Syntax, Utrecht: LOT. Wrede, F. (1908), ‘Die Diminutiva im Deutschen’, Deutsche Dialektographie, 1, pp. 71–144. Wu, Y. (2005), A Synchronic and Diachronic Study of the Chinese Xiang Dialects, Berlin: De Gruyter. Wyatt, W. (1879), ‘Some account of the manners and superstitions of the Adelaide and Encounter Bay tribes’, in J. D. Woods (ed.), The Native Tribes of South Australia, Adelaide: Government Printer, pp. 157–81. Yami projects (n.d., a), Digital Archiving: Yami Language Documentation, http://yamiproject.cs.pu.edu.tw/yam. Yami projects (n.d., b), Yami Language Learning Centre, http://yamiproject.cs.pu.edu.tw/ elearn. Young, R. W. and Morgan, W., Sr. (1987), The Navajo Language: A Grammar and Colloquial Dictionary, rev. edn, Albuquerque: University of New Mexico Press. Young, R. W., Morgan, W., Sr. with Midgette, S. (1992), Analytical Lexicon of Navajo, Albuquerque: University of New Mexico Press. Yue, A. O. (2003), ‘Chinese dialects: grammar’, in G. Thurgood and R. J. LaPolla (eds), The Sino-­Tibetan Languages, London: Routledge, pp. 84–125. Zamponi, R. (2008), ‘Sulla fonologia e la rappresentazione ortografica del lule’, in R. Badini, T. Deonette and S. Pineider (eds), A. Maccioni: Arte y vocabulario de la lengua lule y tonocoté, Cagliari: CUEC/Centro di Studi Filologici Sardi, pp. xxii– lviii. Zelinková, K. (1993), ‘Sústava slovenských deminutív’, in Zborník k 11. medzinárodnému zjazdu slavistov, Nitra: Vysoká škola pedagogická, pp. 28–44. Zhengzhang, S. (2008), 溫州方言誌 [The Wenzhou Dialect], Beijing: Yuwen Chubanshe. Zhu, X. (2006), A Grammar of Shanghai Wu, Munich: Lincom Europa. Zöller, H. (1891), Deutsch-­Neuguinea und meine Ersteigung des Finisterre-­Gebirges, Stuttgart: Union Deutsche Verlagsgesellschaft. Zorc, D. R. and San Miguel, R. (1991), Tagalog Slang Dictionary, Kensington, MD: Dunwoody Press. Zuazo, K. (2008), Euskalkiak: Euskararen dialektoak, San Sebastian: Elkar. Zubin D. A. and Köpcke K. M. (1986), ‘Gender and folk taxonomy: the indexical relation between grammatical and lexical categorization’, in C. Craig (ed.), Noun Classes and Categorization, Amsterdam: Benjamins, pp. 139–80. Zwaan, R. A., Stanfield, R. A. and Yaxley R. H. (2002), ‘Language comprehenders mentally represent the shape of objects’, Psychological Science, 13: 2, pp. 168–71. Zwicky, A. (1985), ‘Heads’, Journal of Linguistics, 21, pp. 1–29.

700

References

Zwicky, A. and Pullum, G. (1987), ‘Plain morphology and expressive morphology’, Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics Society, pp. 330–40. Zwitserlood, I. (2012), ‘Classifiers’, in R. Pfau, M. Steinbach and B. Woll (eds), Sign Language: An International Handbook, Berlin: De Gruyter, pp. 158–85.

Language index

Language index

Abelam, 408 Adamorobe Sign Language see Sign Language: Adamorobe Adnyamathanha, 424 Afrikaans, 79 Aghem, 46 Aghul, 57 Agta, 58 Ahtna, 540, 588 Aikanã, 606 Aja, 472 Akan, 6, 70, 146, 474, 487 Alabama, 529 Alamblak, 81 Alaskan Inupiaq, 559 Alawa, 65 Albanian, 48, 80, 100 American Sign Language see Sign Language: American Anamuxra, 50 Ancient Greek see Greek: Ancient Anejom, 47 Angolar, 182 Ao, 70 Apache Western, 537 White Mountain, 591 Apma, 107, 109, 111, 190, 192, 341–51 Arabana, 50, 61, 68 Arabic, 6, 109, 111, 116, 126, 180, 193n Classical, 106, 180, 461–71 Colloquial Jordanian, 130 Modern Standard, 93, 180 Moroccan, 103, 106n, 190, 461–71 Juba, 172, 180, 181 Arin, 246 Assan, 246 Australian Sign Language see Sign Language: Australian Austrian German see German: Austrian Avar, 57 Aymara, 595, 602, 603, 604, 605 Azerbaijani, 52, 314, 315 Bafut, 54, 70, 175 Bāgandji, 58, 70 Bahasa Indonesia, 51 Balangao, 379 Basque, 56, 57, 128, 154n, 195–204 Beaver, 587 Belorussian, 68, 72 Bemba, 69, 72

Bengali, 126 Berber, 6, 9, 54, 81, 88, 94, 100, 103, 453–60 Berbice Dutch Creole see Creole: Berbice Dutch Besiro, 642n Bezhta, 50 Bikol, 380 Bininj Gun-wok, 406, 430 Bislama, 177 Bocotá, 527n Boikin, 408 Bolivian Quechua, 643 Boruca, 523, 527 Breton, 56, 102 Bribri, 191, 523, 527 British Sign Language see Sign Language: British Bulgarian, 9, 60n, 80, 92 Buli, 480, 484, 485, 486 Burmese, 365, 366n, 381 Cabécar, 9, 523–8 Cameroonian Pidgin English see Pidgin English: Cameroonian Cantonese, 146, 352 Hong Kong, 354, 356, 357 Xinyi, 358 Catalan, 9, 10, 83, 86, 95, 103, 107n, 109, 117, 190, 205–15, 623n Chagatai, 315 Changsha Xiang, 356 Chickasaw, 529, 535 Chimane, 643 Chinese, 111, 126, 192, 334, 376, 381, 352–60 Standard Mandarin, 352–60 Chocho, 542 Choctaw, 190, 192, 193, 529–35 Chorote, 634, 637 Classical Arabic see Arabic: Classical Classical Tibetan see Tibetan: Classical Colloquial Jordanian Arabic see Arabic: Colloquial Jordanian Cree, 53, 568–78 East, 571 Plains, 85, 90n, 107n, 111, 188, 190, 192, 568–78 Western Swampy, 569 Woods, 568 Creek, 529, 534 Creole Berbice Dutch, 174 French, 174 Jamaican, 174, 177, 179

Language index

702 Creole (cont.) Martinica, 178 São Tomé, 182 Suriname, 175, 176 Croatian, 59, 69, 72, 93, 103, 106n, 136, 137 Czech, 49, 55, 128, 148, 301 Dakelh, 102 Dalabon, 77, 107n, 190, 192, 401–7, 430, 432 Dangaléat, 48 Danish, 306, 307 Dena’ina, 90, 191, 536–41, 588 Dëne Sųłiné, 540, 584, 586–7 Djaru, 65 Dolakha Newar, 58 Dumi, 57 Dutch, 52, 79, 92, 111, 136, 262, 263, 266 East Cree see Cree: East Enets, 44, 227 English, 29, 32, 33, 35, 40, 41, 51, 81, 117, 121, 125, 126, 130, 136, 140, 188, 189, 192, 223, 262, 294, 307, 308, 309, 352, 369, 370, 375, 377, 395, 405, 406, 408, 411, 414, 416, 417, 418, 420, 423, 424, 448, 480, 487, 532, 549n, 550, 578n, 583, 597, 599–601, 605n, 615n, 650n Estonian, 36, 227, 237 Eton, 46 Even, 54, 333, 338, 339 Evenki, 333, 334, 337–9 Ewe, 9, 143, 146, 190–3, 472–9, 487 Faroese, 306 Figuig, 54 Finnish, 36, 109, 112, 117, 136, 227, 237 Fon, 472 Fongbe, 176 French, 6, 9, 32–4, 36, 37, 40, 41, 77, 78, 80, 84, 103, 109, 110, 133n, 134, 136, 139, 140, 143, 146, 147, 149, 151–3n, 174, 192, 195, 262, 307, 308, 352, 548 French Creole see Creole: French Frisian, 57 Fula, 101, 102 Gagauz, 50, 56, 58 Gen, 472 Georgian, 9, 111, 216–26 German, 6, 23, 32, 34, 36, 40, 62, 78, 92, 111, 121, 125, 127, 128, 136, 138–40, 144, 146–8, 153n, 154n, 192, 237, 262–4, 266, 268, 307, 352 Austrian, 136, 138, 139, 228 High, 263, 265, 307 Low, 263, 307 German Sign Language see Sign Language: German Gikuyu, 87, 99 Greek, 8, 136–9, 143, 148, 150, 153n, 154n, 307, 308 Ancient, 97, 147–9 Modern, 6, 10, 78, 81, 83, 84, 90n, 92, 95–7, 100, 103, 106n, 147–50, 190, 269–77 Greenlandic, 559, 561, 567

Guató, 642n Guaymí, 527n Gungbe, 176 Gurian, 223 Haitian, 147 Hamar, 54 Hausa, 53, 62, 65, 238–45 Hebrew, 6, 52, 104, 107, 111, 135, 136, 138 Israeli, 238–45 High German see German: High Hitchiti, 529 Hong Kong Cantonese see Cantonese: Hong Kong Huautla Mazatec, 542–9 Huave, 550–8 Hungarian, 34, 61, 66, 68, 113, 136–9, 188, 227–37 Hup, 58 Hupa, 592 Iatmul, 94, 100, 107n, 109, 111, 190, 191, 408–15 Ibaloi, 380 Icelandic, 6, 53, 306 Ilocano, 49, 65, 379 Innu, 571 Inuktitut, 55, 82, 85, 191, 559–67 Irish, 69, 72, 146 Israeli Hebrew see Hebrew: Israeli Israeli Sign Language see Sign Language: Israeli Italian, 3–6, 9, 10, 13, 14, 16, 17, 19n, 23, 32–41, 75, 76, 78, 79, 81, 83, 85, 87–9, 91, 95, 97, 101, 103–6n, 108, 109–12, 116, 121, 124, 126–9, 133n, 135–40, 146–8, 151, 152, 155, 187, 189, 211, 213, 266, 307, 308, 407n, 507, 512, 624n Italian Sign Language see Sign Language: Italian Ixcatec, 542 Jamaican Creole see Creole: Jamaican Jaminjung, 109, 115 Japanese, 70, 130, 152 Jaqaru, 54, 107n, 192, 595–608 Javanese, 69, 72 Jawoyn, 430 Jingulu, 107n, 117, 188, 416–22 Jóola Banjal, 46 Juba Arabic see Arabic: Juba Kabardian, 49, 58 Kadiwéu, 625 Kannada, 321 Kanoê, 606–8 Kanuri, 44, 65 Karachay-Balkar, 314 Karamojong, 93, 106n Kashubian, 44 Kaurna, 423–9 Kawki, 595, 602–4 Kayardild, 193n, 448–52 Kazakh, 44, 313, 315, 318 Kɔnni, 191, 480–6 Ket, 90, 107, 246–52 Khanti, 227 Khasi, 50

Language index 703 Khevsurian, 223, 224 Khoekhoe, 70 Kíhehe, 111 Kirghiz, 313, 315, 318 Kituba, 172 Koasati, 529, 534 Koiari, 51 Kolyma Yukaghir, 58 Komi, 44, 49, 227 Kongo, 9, 172 Korean, 65, 66 Kott, 246 Koyra Chiini, 70 Koyukon, 536, 540 Krio, 182 Kriol, 405, 406 Krongo, 47 Kwakw’ala, 57, 65 Kwaza, 10, 45, 86, 606–15 Ladin, 45, 53 Lahu, 365, 366n Lakota, 69, 72 Lardil, 18, 193, 448–52 Latin, 8, 28, 96, 97, 110, 143, 147–53n, 212, 307–9 Latundê, 606, 615 Latvian, 106n, 135, 190, 192, 237, 253–61 Laz, 224, 225 Lega, 80, 81 Leizhou Southern Min see Southern Min: Leizhou Lhasa Tibetan see Tibetan: Lhasa Lingala, 147 Lisu, 190, 192, 361–6 Lithuanian, 54, 59, 136, 138, 139, 253, 254 Livonian, 237 Low German see German: Low Lower Sorbian, 54 Lule, 107n, 111, 616–24, 642n Lushootseed, 50 Luxembourgish, 47, 71, 106n, 191, 192, 262–8 Maale, 57 Maasai, 95, 150 Macedonian, 150 Malay, 81, 150 Malayalam, 51, 70, 321 Malécu, 527, 528n Maltese, 53, 58, 62, 93, 103 Mambay, 150 Manambu, 408, 415 Manchu, 333, 340 Mandarin, 146, 352, 356 Pingding, 48 Mandingue, 150 Mang, 50 Mansaka, 379 Mansi, 227 Maori, 173 Mari, 58, 227 Martinica Creole see Creole: Martinica Matsés, 55, 57 Mattole, 592 Maya, 45, 61

Megrelian, 224, 225 Mekens, 608 Micmac, 47, 48, 56 Migama, 54 Mikasuki, 529 Mnong, 550 Mocoví, 625 Modern Greek see Greek: Modern Modern Standard Arabic see Arabic: Modern Standard Mokhevian, 223 Mordvin, 56, 227 Moroccan Arabic see Arabic: Moroccan Mosetén, 47, 53 Movima, 46, 70 Mparntwe Arrernte, 406 Muna, 49, 51, 71, 82, 190, 192, 367–74 Nahuatl, 63, 130, 131, 147, 153n Nama, 58, 95, 106n Nanai, 86, 334, 337, 340 Naso, 523, 527 Navajo, 439, 536, 579, 590 Ndyuka, 174 Negidal, 86, 333, 337, 339 Nelemwa, 51 Nenets, 44, 56, 227, 237 Ngala, 408, 445 Ngalakgan, 406, 430 Nganasan, 227 Ngandi, 430 Ngiti, 146 Ngiyambaa, 406 Nhanda, 154n Nipmuck, 45 Nivacle, 634, 637 Nivkh, 9, 65, 190–2, 278–86 Noon, 146 Nootka, 131, 132 Norwegian, 306, 307 Nubi, 180 Nukunu, 425 Nupe, 175 Odawa, 571 Ojibwe, 570, 571 Old Church Slavonic, 150 Old Tibetan see Tibetan: Old Oroch, 86, 333, 334, 337, 338 Orok, 340 Oroqen, 339 Ossetic, 52, 58 Papiamentu, 174 Passamaquoddy, 6, 47 Persian, 126, 217, 287–95 Pévé, 54 Pidgin English Cameroonian, 175 West African, 173, 177, 178, 181, 182 Pilagá, 625 Pingding Mandarin see Mandarin: Pingding Plains Cree see Cree: Plains

Language index

704 Plantation Pidgin Fijian, 173 Popoloca, 542 Portuguese, 3, 6, 40, 80, 84, 95, 102, 104, 109, 112, 123 Pshav, 223 Pumpokol, 246 Rachian, 223 Rembarrnga, 111, 430–7 Romanian, 80, 84, 95, 110 Romblomanon, 379 Russian, 17, 25, 58, 69, 72, 136–8, 140, 226n, 237, 240, 249, 333, 335 Saho, 109 Salamãy, 606 Sami, 227 Sanskrit, 147, 287, 321 São Tomé Creole see Creole: São Tomé Sara Mbay, 44 Saramaccan, 174, 176 Sardinian, 6, 56, 57, 187 Saulteaux, 570, 571 Scottish Gaelic, 48 Sɛlɛɛ, 82, 88, 472, 487–95 Selkup, 227 Serbian, 93 Shanghai Wu see Wu: Shanghai Shona, 80, 83, 88, 90n, 98, 101, 106n, 191, 492, 496–506, 515, 516 Sign Language Adamorobe, 161 American, 156 Australian, 161 British, 161 German, 163 Israeli, 156 Italian, 16, 17, 37, 155 Sirionó, 643 Slave, 541 Slavey, 87, 88, 540, 579–93 Slovak, 3, 6, 9–11, 14, 44, 54, 55, 62, 93, 100, 146, 188, 190, 191, 296–305 Slovene, 128 Solon, 333 Somali, 105, 107n, 111, 507–14 Sotho, 150 Southern Barasano, 102 Southern Min Leizhou, 356 Taiwanese, 357 Spanish, 3, 6, 9, 17, 28, 29, 33–5, 37, 39–41, 48, 49, 80, 84, 95, 97, 109, 110, 112, 128, 131, 136–8, 189, 192, 195, 205, 213, 214, 375–80, 439, 510, 523, 533, 542, 549n, 551, 555–7, 597, 600, 602–4, 624n–6, 638, 641, 642n, 643 Squamish, 111 Sranan, 176, 179 Standard Mandarin Chinese see Chinese: Standard Mandarin Supyire, 46, 58 Suriname Creole see Creole: Suriname Svan, 224, 225

Swahili, 6, 81, 515, 516 Swati, 150 Swedish, 36, 147, 192, 306–9 Tagalog, 190, 375–80 Taiwanese Hakka, 353, 357, 360n Taiwanese Southern Min see Southern Min: Taiwanese Tamil, 321 Tanacross, 540 Tapiete, 54 Tat, 45, 56 Tatar, 68, 72, 190, 192, 310–20 Tausūg, 380 Telugu, 9, 190–2, 321–32 Thai, 67 Tibetan, 6, 190, 192, 352, 381–8 Classical, 381, 382, 388n Lhasa, 381–3, 386–8 Old, 381, 388n Tłįchǫ Yatiì, 584 Toba, 625–33 Tok Pisin, 177, 408 Trakai Karaim, 315 Trinitario, 643 Tswana, 492 Turkana, 93 Turkish, 50, 55, 135, 136, 140, 269, 314–16, 318–20n, 339 Turku, 180 Tzeltal, 123 Udihe, 58, 65, 82, 85, 86, 103, 107n, 190–2, 333–40 Udmurt, 227 Ulcha, 334, 340 Upper Tanana, 540, 589 Venda, 82, 99, 147, 492, 493 Vietnamese, 146, 150 Vilela, 616, 642n Vitu, 65 Warlpiri, 107n, 190, 438–47 Warray, 430 Welsh, 47 Wenzhou Wu see Wu: Wenzhou West African Pidgin English see Pidgin English: West African Western Apache see Apache: Western Western Swampy Cree see Cree: Western Swampy White Hmong, 150 White Mountain Apache see Apache: White Mountain Wichí, 10, 52, 86, 107n, 190, 191, 192, 634–42 Witsuwit’en, 589 Woods Cree see Cree: Woods Wu Shanghai, 352, 357, 358 Wenzhou, 356 Xhosa, 54 Xinyi Cantonese see Cantonese: Xinyi Xwla-Xweda, 190, 192, 472

Language index 705 Yami, 389–400 Yana, 131 Yelogu, 408 Yugh, 246 Yuki, 643 Yukulta, 190, 192, 193n, 448–52 Yupik, 561, 565, 567

Yurakaré, 58, 190, 192, 643–50 Yurok, 6 Zenaga, 54 Zoque, 109 Zulu, 187, 190, 191, 492, 515–22

Subject index

Subject index

ablaut, 262, 306, 354, 357–9, 381, 388 accomplishment, 110, 111, 113–15 achievement, 111, 567n, 622, 648 acquisition, 16, 36, 134, 135, 137–41, 149, 333, 338, 339 of augmentatives, 135, 141 of diminutives, 134, 135, 140, 141 of elatives, 134 of evaluative affixes see acquisition of evaluatives of evaluative morphology see acquisition of evaluatives of evaluatives, 8, 134, 135, 141, 188 first language, 26, 124, 134, 135 of pejoratives, 136, 137, 141 second language, 171, 178, 180, 408, 487 actional see actionality actionality, 3, 6, 16, 43, 44, 61–4, 78, 109, 110, 111, 187, 193, 232 active see voice: active activity verb, 111, 112, 114, 115, 233–5, 300, 349, 567, 571, 617, 621–3 address form, 131, 322, 323 addressive particle, 476 admiration, 295, 520, 548 admirative see admiration adpart see addressive particle af see focus: agent aff see affirmative affectionate, 6, 11, 29, 37, 67, 70, 311, 312, 326, 443, 445, 530, 540 affective, 23, 26, 35, 38, 39, 103, 123, 127, 165, 167, 169, 200, 204, 279, 335–8, 474, 629, 632, 633n palatalisation, 200, 204 affectivity see affective affirmative, 211, 328, 348–50 ag see agentive age, 9–12, 16, 19, 22, 25, 26, 106n, 122–4, 126, 127, 130, 141, 218, 288, 311, 313, 335, 402, 403, 408, 424, 437, 477, 490, 492, 510, 527, 565, 581, 584, 593, 596, 628, 631, 632, 636, 637 age variation see age agent see agentive focus see focus: agent agentive, 3, 14, 54, 76, 96, 97, 106n, 376, 380, 630, 636, 639, 640 agentivity see agentive agglutinating morphology see agglutinative morphology agglutination see agglutinative morphology agglutinative morphology, 195, 216, 248, 306, 310, 321, 334, 338, 352, 353, 367, 408, 430, 472, 487, 496, 515, 616, 625, 634, 643 Aktionsart see actionality

see agreement agreement, 43, 87, 88, 94, 100, 105n, 108, 156, 158, 195, 224, 249, 250, 279, 310, 333, 361, 376, 409, 411, 415, 417–20, 437, 438, 459, 488, 500, 504, 505, 509, 510, 515, 516, 559, 561, 585, 589, 590, 593, 596, 602, 627 diminutive see diminutive: agreement evaluative see evaluative: agreement gender see gender: agreement inflectional see inflectional: agreement noun class see noun class agreement amelioration see ameliorative ameliorative, 4, 12, 15, 23, 24, 27, 62–4, 145, 148, 151, 153n, 154n, 192, 284, 285, 338, 339, 430, 434–7, 449–51, 502, 503, 517, 519, 521, 527, 568, 569, 577, 607, 617, 618, 623 circumfix, 379 marker, 449, 450 meaning, 151, 153n, 154n, 519, 617 prefix, 153n suffix, 338, 451, 623 aml see ameliorative analytical diminutive see diminutive: analytical animacy see animate animate, 7, 92, 93–100, 105, 106n, 147, 165, 170, 250, 254, 275, 288–91, 297, 311, 335, 351, 318, 320, 361, 362, 402, 404, 405, 409, 431, 437, 456–9, 487, 518, 569, 571, 611, 626, 635, 636 animateness see animate aplt see appellative app see appreciative appellative, 142, 312, 612 appreciation see appreciative appreciative, 15, 25, 28, 63, 64, 72, 119, 126, 215, 291, 389, 395, 494, approx see approximative approximation see approximative approximative, 9, 11, 12, 25, 63, 64, 66, 94, 144, 145, 148, 149, 162, 167, 192, 200, 201, 211, 213, 216, 221, 223–5, 262, 267–9, 308, 317, 338, 348, 350, 367–72, 374n, 382, 391, 393–5, 399, 414, 415, 430, 437, 438, 443, 507, 512, 513, 523, 533, 550, 554, 621, 623, 625, 630–2, 634 asp see aspect aspect, 6, 16, 110, 112, 115, 117–19, 156, 162, 165, 169, 173, 175, 180, 182, 213, 232, 233, 234, 250, 253, 257, 259, 278, 287, 293, 310, 350, 352, 360n, 361, 375, 382, 384, 385, 389, 417, 419, 422, 423, 430, 448, 472, 516, 520, 538, 550, 576, 606, 617, 621, 623, 625 continuative, 109, 112, 173, 181, 380, 388n, 647 continuous, 328, 329, 367, 372, 373, 413, 432 agr

Subject index 707 delimitative, 111, 358 imperfective, 110, 111, 114, 287, 346–9, 413, 470, 538, 541n, 596, 598, 645 irresultative, 112 iterative see iterative verb lexical see actionality marker, 173, 342, 344, 382, 523, 529, 530 perfective, 110, 111, 222, 253, 257, 259, 346, 347, 349, 353, 367, 403, 405, 432–5, 437, 536, 539, 600 potential, 82, 472 procrastinative, 611, 612 progressive, 111, 112, 115, 173, 176, 304n, 329, 431, 433–5, 473, 478, 567n, 627, 645 resultative, 112, 113, 280 aspectual see aspect atelic verb, 78, 110–15, 117, 120n atelicity verb see atelic verb att see attenuative attenuation see attenuative attenuative, 9, 11, 12, 25, 27–9, 36, 37, 40, 43, 44, 56, 62–4, 68, 82, 84, 111, 123, 124, 126, 134, 167, 172, 180, 188, 192, 213, 216, 221–3, 225, 228, 233–5, 249, 258, 259, 279, 281, 285, 300, 311, 316–20, 336, 338, 340, 343, 346, 350, 356–9, 367, 372, 373, 376–80, 389, 391, 395, 399, 443, 446, 452, 479, 493, 507, 511–13, 524–7, 597, 610, 611, 615, 617, 620–3, 638–40, 642 circumfix, 221, 222 reduplication see reduplication: attenuative aug see augmentative augmentation see augmentative augmentative, 4–6, 8–16, 21, 28, 29, 32, 35–41, 45–56, 58, 59, 61–72, 72n, 73n, 76, 78, 84, 86, 87, 94–101, 102, 105, 106n, 107n, 112, 118, 121, 124, 127, 129, 132n, 133n, 134–7, 139–43, 145, 149–53, 153n, 154n, 160, 164, 166, 177–81, 188, 191, 192, 196, 201, 202, 205, 206–12, 214, 215, 215n, 237, 238, 253, 256, 257, 259–61, 267, 270, 271, 274, 275, 277n, 279, 281–3, 286n, 290, 295n, 296, 303, 304, 307–10, 316, 318, 319, 338–40, 350, 359, 361–5, 379, 380, 389, 391–3, 398, 399–401, 409, 411, 416, 426–8, 431, 439–42, 445, 453, 458, 461, 467, 470, 473, 476–9, 486n, 487, 489, 494–7, 499–503, 505, 506, 515, 516, 518, 519, 521, 527, 530, 534, 538, 544, 545, 547–9n, 555, 556, 568, 575, 577, 579, 580–93, 593n, 607, 608, 619, 623, 628–31, 635–8, 640–7, 649, 650 acquisition see acquisition: of augmentatives circumfix, 253, 259–61, 379 construction, 274, 476, 636, 641 enclitic, 55 honorific, 503 intensifier, 391, 398, 640 mark see augmentative marker marker, 7, 94, 150, 153n, 283, 290, 303, 304, 650n meaning, 14, 45, 59, 97, 99, 150–3, 177–9, 181, 192, 253, 257, 259, 279, 281, 286n, 426, 515, 567 onomasiological category, 6 prefix, 150, 556

suffix, 84, 86, 99, 102, 106n, 121, 127, 133n, 150, 154n, 191, 201, 211, 212, 228, 232, 237, 253, 256, 261, 274, 275, 277n, 280, 282, 296, 303, 361, 362, 426, 428, 439, 440, 442, 477, 478, 516, 518, 519, 576, 588, 598, 603, 611, 623, 631, 636, 637 superlative, 392 value, 96, 106n, 215n, 294, 555 augmentativeness see augmentative augmentivisation, 29, 50, 54, 58, 61, 71, 78, 79, 94, 95, 104, 188, 190–2, 275, 279, 281, 282, 285n, 296, 458 authenticity, 10–12, 63, 64, 167, 295n, 526, 566, 608, 615 baby talk, 107n, 124, 125, 133n–40, 231, 309, 335, 357, 444 backformation, 138, 139, 288 ben see benefactive benefactive, 430, 431, 521, 647 bigness, 21, 28, 29, 36, 139, 163, 191, 371, 479 bipartite negator see negative blaming, 398, 399 blm see blaming blocking, 79, 126, 157, 212, 335, 501, 503, 505, 516 bt see baby talk caressive, 310–14, 316, 317, 320 categorial neutrality, 75–7, 89 CDS see baby talk cel see celerative celerative, 617, 621–3 cf see circumfix child directed speech see baby talk child language acquisition see acquisition: first language child speech, 16, 71, 135, 139, 140 childness, 25, 26, 28 circumclitic see clitic circumfix ameliorative see ameliorative: circumfix attenuative see attenuative: circumfix augmentative see augmentative: circumfix derivational see derivational: circumfix diminutive see diminutive: circumfix evaluative see evaluative: circumfix feminine see feminine: circumfix pejorative see pejorative: circumfix singulative see singulative: circumfix superlative see superlative: circumfix circumfixation see circumfix cl see noun class: marker classifier, 45, 46, 156–9, 161, 165, 166, 168, 169, 279, 280, 347, 348, 352, 353, 361–3, 365, 367, 371, 536–9, 543, 545–7, 607–10, 625, 627 clf see classifier cliticisation see clitic clitic, 40, 46, 99, 190, 223, 236, 247, 248, 287, 294n, 309, 320, 343, 348, 350, 375, 379, 382, 383, 401–4, 407, 410, 423, 438, 440, 448, 473, 521, 523, 525, 526, 528n, 536–41, 543, 544, 547, 548, 550, 555, 567n, 569, 588, 589, 592, 643 augmentative see augmentative: enclitic

Subject index

708 clitic (cont.) diminutive see diminutive: proclitic evaluative see evaluative: clitic pausal, 373 cnt see aspect: continuative coalescence, 143, 153n, 191, 408, 412, 546, 567n cognitive category, 4, 24, 43, 62, 67–9, 74, 188 grounding, 28, 30 linguistics, 21 semantics, 21, 23, 24, 29 coll see collective collective, 97, 102, 103, 150–2, 206, 257, 334, 380, 454, 455, 460n, 596, 598, 644, 646 compar see comparative comparative, 13, 23, 43, 44, 61, 75, 110, 118, 119, 182, 199, 200, 202, 203, 221, 223, 236, 280, 287, 289, 292, 293, 295n, 317, 378, 379, 416, 426, 518, 556, 604, 605n, 609, 638, 644, 645 concepts, 9 construction, 13, 118, 280, 375, 556, 557, 644, 650n comparison see comparative compassion, 402, 404–6, 612, 639 compound, 3, 52, 78, 140, 177, 182, 193, 215n, 218, 247, 248, 288, 309, 319, 331, 336, 358, 361, 362, 387, 412, 425, 485, 486, 543, 545, 546, 548, 588, 605n, 615, 625, 626, 628, 630, 631, 636, 637 compounding, 6, 10, 40, 47, 51, 52, 59, 74, 136, 172, 176, 177, 190–2, 216, 218, 225, 246, 252, 262, 269, 270, 279, 284, 287, 288, 293, 294, 308, 321, 322, 325, 327, 329, 353, 367, 423, 428, 430, 472, 529, 530, 534, 537, 543–6, 548, 560, 607, 614, 636 con see contempt cont see aspect: continuous contact, 126, 129, 154n, 170–6, 178, 262, 263, 268n, 333, 375, 551, 556, 557, 583, 593n, 606, 616 contempt, 5, 10–12, 25, 33, 63, 64, 66, 144, 145, 149, 151, 152, 165, 192, 216, 218, 221, 223–5, 230, 277n, 288, 289, 294, 295n, 311, 346, 350, 367, 368, 493, 566, 571, 615, 629, 630, 632, 638, 641, 642 continuative see aspect: continuative continuous see aspect: continuous conversion, 47, 104, 111, 127, 172, 175, 176, 178, 199, 288 count noun see countable noun countable noun, 78, 103, 104, 111, 112, 150, 170, 248, 276, 306, 402, 403, 457, 627 CS see child speech deintensification, 63, 64, 310, 316, 317, 320, 344 delimitative see aspect: delimitative depreciation see depreciative depreciative, 389, 391, 395–9, 540 der see derogatory derivation see derivational: morphology derivational affix, 3, 16, 76, 144, 150, 224, 246–9, 253, 264, 325, 327, 338, 353, 354, 375, 409, 419, 436, 516 base, 143, 147, 150, 154n, 307, 308 category, 81, 143



circumfix, 259, 287 class, 81 marker, 264 morpheme, 177, 201, 203, 413, 416, 420, 421, 561, 612, 614 morphology, 3, 4, 6–9, 15, 16, 19n, 34, 43, 45–7, 59, 63, 65, 74, 75, 76, 79, 80, 83, 85, 88–90n, 106n, 134–6, 140, 143, 156, 172, 176, 177, 182, 187, 215n, 216, 227, 232, 235, 237, 240, 244, 252, 264, 269, 287, 294, 295n, 325, 338, 352, 365, 367, 382, 383, 387, 408, 412, 413, 419, 423, 458–60n, 461, 489, 496, 504, 505, 507, 526, 529, 530, 546, 555, 556, 562, 564, 565, 567, 572, 573, 576, 600, 605n, 606, 617, 625 prefix, 537, 541n, 543 process, 43, 49, 178, 180, 287, 293, 310, 352, 353 rule, 4, 6, 67 suffix, 3, 46, 49, 58, 74, 84, 150, 182, 190, 201, 233, 257, 269, 270, 276, 277n, 293, 295n, 310, 334, 353, 384, 406, 419, 420, 572, 600, 614 suffixation see derivational: suffix synonyms, 257–9 system, 58, 227 derogation see derogatory derogatory, 28, 32, 34, 295n, 298, 326, 329, 330, 379, 490, 493, 494, 618, 623 desid see desiderative desiderative, 282, 627, 636, 640 dim see diminutive dim.cl see diminutive class diminution see diminutive diminutive, 3–16, 21, 23, 25–9, 32–41, 44–8, 51–60n, 61–72n, 73, 73n, 77, 78, 80–9, 89n, 90n, 91–6, 99–106n, 109, 111, 112, 116–19, 123–32n, 133n–49, 151–4n, 157, 158–62, 164, 165, 168, 172, 174, 190–3, 196–213, 215, 215n–21, 223–5, 227–32, 237–42, 244n, 253–9, 263–8n, 270–7n, 279, 280, 288–92, 294n, 295n, 296–305n, 307–17, 320–4, 331, 333–40, 343, 344, 350, 356, 358, 361–6n, 367, 368, 376, 380–3, 388–90, 393, 395, 399, 401–6, 409–12, 416, 425–8, 430–4, 436–9, 442–6, 453, 455–60n, 461, 473–6, 479, 480, 483, 486n, 487, 489–99, 502, 505–7, 510–12, 514, 515, 517–19, 521, 524, 525, 527, 530, 531, 533–5, 539, 545–9n, 565, 566, 568–75, 577, 579–84, 586–93n, 596–8, 602–4, 607, 615, 625–31, 633n–42, 644, 646, 647, 649, 650 acquisition see acquisition: of diminutives affix, 4, 47, 69, 92, 93, 117, 136, 145, 148, 154n, 192, 217, 261, 376, 391, 516 affixation see diminutive: affix agreement, 224 analytical, 493, 494 circumfix, 81, 88, 94, 456–8 class, 98, 502, 503, 504, 515, 516, 637, 640, 641 compound, 637 construction, 66, 382, 487, 489, 491, 492, 495, 506, 544, 575, 636, 641 diminutivum modestum, 128 diminutivum puerile, 124 diminutivum senile, 126 double, 199, 305n, 336

Subject index 709 enclisis, 543 false, 132n fossilised, 491 impoverishment, 339 infix, 10, 48, 265, 379, 467, 470, 572 lexical, 432 lexicalised, 14, 217 mark see diminutive marker marker, 7, 57, 70, 85, 88, 90n, 146, 149, 154n, 188, 191–3, 216, 224, 256, 288, 294, 296, 300, 302, 304, 474, 491, 492, 494, 532, 534, 644 meaning, 37, 39, 46, 48, 54, 59, 90n, 94, 98, 104, 106n, 138, 147, 148, 150, 154n, 179, 191, 218, 220, 253, 259, 279, 286n, 297, 301, 313, 336, 356, 362, 368, 376, 382, 383 388, 425, 466, 475, 476, 491, 538, 543, 574 onomasiological category, 7, 8 particle, 123 pattern, 544, 571 plural, 60n, 102, 264 prefix, 81, 259, 364 primary diminutive, 301 proclisis see diminutive: proclitic proclitic, 46, 60n, 543 reading see diminutive: meaning reduplication, 178, 240, 357, 377, 379, 381, 391, 427, 434, 469, 470, 494, 520 schema, 135 secondary, 301, 305n sense see diminutive: meaning sentence, 38 series, 502 sound symbolism see sound symbolism stem, 47 suffix, 16, 37, 47, 53, 54, 56, 58, 59, 65, 70, 78, 81, 82, 85–7, 90n, 92, 99, 107n, 116, 123, 125, 130, 131, 135, 137, 141, 143, 144, 146, 147, 149, 150, 153, 153n, 190, 192, 193, 196–8, 207, 210–12, 217–19, 223–5, 227–32, 237, 240, 241, 248, 253–6, 258, 263–7, 271–3, 276, 277n, 279, 280, 285n, 296–302, 304, 307, 309, 313, 315–17, 334–6, 338, 361–3, 365, 375, 376, 442, 443, 445, 473–7, 479–81, 489, 490, 492–4, 516–18, 527, 569, 573, 574, 576, 596–8, 603, 604, 626, 629, 635, 637, 638 suffixation see diminutive: suffix template, 239–41, 461–70 value, 211, 212, 217, 355 vowel, 465 diminutiveness see diminutive diminutivisation, 29, 39, 49, 57, 58, 61, 78, 79, 86, 88, 94, 102, 104, 123, 128, 139, 140, 188, 190, 191, 238, 241, 265, 279, 280, 296, 301, 302, 305n, 335, 339, 433, 458, 475, 476, 494, 502, 510, 569, 571, 573, 574, 596, 602, 627 dismissive, 329–31 dm see derivational: morphology double diminutive see diminutive: double dss see sound symbolism du see number: dual dual see number: dual dur see durative verb duration see durative verb

durative verb, 78, 109–14, 119, 120, 228, 233, 235, 282, 310, 328, 329, 346, 356, 372, 380, 391, 427, 436, 446, 627, 628, 633n dynamic verb, 78, 110, 111, 113, 115, 116, 119, 367, 378, 413 see echo-formation echo-formation, 190, 321, 329, 330, 380 elative, 40, 52, 128 EM acquisition see acquisition: of evaluative morphology EM universal see universal EM embodied see embodiment embodiment, 21–3, 25, 27, 155 emotion see emotional emotional, 23, 32, 33, 35, 38, 39, 41, 66, 118, 128, 156, 165, 219, 220, 248, 249, 297, 311, 399, 401–6, 597 empathy, 39, 41, 113, 139, 338, 402, 406, 409 emph see emphatic emphatic, 159, 198, 199, 203, 204, 236, 335, 336, 338, 403, 414, 418, 419, 421, 611, 613, 628 reduplication see reduplication: emphatic enclitic see clitic endearing, 5, 9, 11, 12, 22, 24, 27, 29, 31, 63, 64, 93, 100, 123, 130, 139, 165, 192, 196, 211, 216, 218–20, 223–5, 231, 232, 234, 253, 255, 266, 267, 288, 289, 291, 295, 311, 312, 314, 315, 323, 324, 331, 335, 338–40, 354, 356, 357, 376, 405, 406, 437, 474, 479, 493, 527, 540, 550, 553, 566, 569, 582, 596, 610, 611, 615, 629, 632, 638, 641, 642, 644 endearment see endearing ep see evaluative: prefix epenthetic vowel see vowel: epenthetic es see evaluative: suffix euphemism, 27, 28, 32, 37, 41, 531 Euroversal, 15, 68 eval see evaluative: affix evaluative affix, 3–8, 12, 15, 16, 21, 23, 46, 59, 73–80, 82, 84, 87–90n, 91, 99–101, 105n, 108, 110, 113–17, 142–6, 148, 153, 187, 213, 215n–20, 224, 249, 354, 499, 585, 588 affixation see evaluative: affix agreement, 87, 88, 504, 505, 627 category, 41, 43, 45, 143, 144, 152, 153, 188, 367, 568, 569, 577 chain, 547 circumfix, 13, 40, 49, 190, 253, 453 class, 5, 80, 81 clitic, 46, 146, 192, 538, 541 complex word, 4, 7 concept, 343, 428, 521, 606, 607, 614 construct see evaluative: construction construction, 5–10, 13–15, 40, 61, 62, 65, 70, 71, 75, 159, 182, 187, 189, 190–2, 270, 271, 275–7n, 296, 321, 322, 431, 436, 450–3, 456, 460n, 487, 489, 494, 495, 505, 507, 513n, 634, 635, 639–42 content see evaluative: meaning degree, 210 derivation, 65, 412, 626 ech

Subject index

710 evaluative (cont.) device, 170, 279 enclitic see evaluative: clitic expression, 126, 130, 165, 606, 607, 614 feature, 116, 251, 474 formation, 11, 33, 38, 45–9, 51, 53, 54, 56, 57, 78–81, 84, 88, 89, 91, 98, 100, 104–6n, 143, 188, 275, 321, 453, 456, 459, 460n function, 13, 14, 40, 81, 91, 99, 103, 146, 155, 157, 161, 162, 164, 169n, 171, 174, 177, 178, 180, 182, 214, 222, 248, 319, 343, 409, 411, 416, 419, 430, 437, 480, 496, 499, 504, 505, 537, 614, 615 infix, 40, 47, 48, 190, 211, 215n mark see evaluative: marker marker, 3, 5, 7, 13, 55, 58, 59, 62, 71, 80, 81, 85, 88, 89n, 107n, 143, 190–2, 216, 223, 225, 276, 291, 292, 304, 311, 398, 431, 436, 473, 479, 489, 527, 545, 560, 589, 649 meaning, 32, 40, 43, 46, 74, 86, 98–101, 107n, 111, 143, 152, 153, 163, 172, 174, 178, 180, 182, 188, 192, 218, 219, 227, 232, 271, 277n, 284, 288, 294, 322, 325, 344, 356, 358, 381, 388, 399, 406, 436, 446, 479, 492, 493, 516, 538, 555, 607, 608, 621, 623, 631, 632 modification, 119, 650 morpheme, 82, 85, 131, 201, 202, 215n, 249, 310, 320, 334, 343, 347, 350, 416, 421, 438, 441, 445, 446, 530, 533, 540, 561–5, 567, 580, 581, 584, 585, 588, 590, 614, 623, 626 notion see evaluative: concept nuance see evaluative: meaning paradigm, 499 pattern, 164, 238, 246 prefix, 81, 87, 99, 132, 270, 496, 497, 504 process, 5, 39, 40, 62, 88, 98, 100, 104, 169, 290, 292, 293, 380, 436, 459, 513 reduplication, 153, 165, 389, 623 relative construction, 519, 521 role, 165 rule, 6, 204, 292 semantics see evaluative: meaning semi-prefix, 411, 415 sense see evaluative: meaning sound symbolism see sound symbolism strategy, 16, 19n, 155, 157, 158, 160, 162–6, 168–72, 216, 223, 224, 279, 430, 431, 436, 459, 507, 513, 537, 550, 551, 554, 624n suffix, 3, 36, 37, 39, 41, 56, 59, 65, 81, 83, 84, 90n, 95, 96, 99, 100, 105, 108–11, 118, 131, 143, 192, 201, 205, 208–11, 215n, 223, 249, 252, 268, 270, 271, 276, 277n, 279, 291, 292, 295, 308, 310, 314, 319, 320, 333, 339, 439, 441, 442, 445, 450, 451, 541, 584, 592, 596, 602, 617, 623, 627, 628, 635, 638, 641, 649 suffixation see evaluative: suffix type, 278, 285 universal see universal value, 14, 320, 617 exc see excessive excessive, 34, 109, 130, 182, 214, 228, 235–7, 253, 260, 293, 295n, 310, 318, 384 excessiveness see excessive

expressive derivation, 6, 67, 70, 74 feature, 135, 406 function, 338, 628, 629 intonation, 249 morphology, 4, 15, 72n, 74, 132n, 279, 613 palatalisation, 40 see gender: feminine false diminutive see diminutive: false fem see gender: feminine female gender see gender: feminine feminine circumfix, 459 gender see gender: feminine fictive, 35–8, 134, 215n fictiveness see fictive first language acquisition see acquisition: first language focus, 278, 371, 609 agent, 390, 393, 394, 396–9 contrastive, 424 marking, 420, 509 intensive, 609 locative, 398 patient 392, 395 system, 399 fossilised diminutive see diminutive: fossilised freq see frequentative verb frequentative verb, 43, 45, 48, 61, 63, 64, 68, 104, 109, 117, 228, 232–5, 237n, 310, 568, 575–7 frequentativeness see frequentative verb frust see frustrative frustrative, 109, 413, 414, 648, 649 full reduplication see reduplication: full fusional morphology, 269, 296, 548, 550 f

gender, 3, 6, 16, 25, 46, 54, 65, 66, 88, 90n–6, 99, 100, 105, 105n, 106n, 124, 125, 127, 145, 190, 208, 209, 239, 240, 244n, 254–7, 262, 266, 267, 270, 273, 277n, 292, 293, 295n, 297, 303, 310, 325, 331, 334, 342, 352, 362, 376, 402, 409–11, 415, 417, 418, 422n, 424, 428n, 453, 456, 459, 460n, 466, 488–93, 495n, 503, 507, 508, 536–9, 541n, 569, 570, 625, 626, 629, 630, 631, 633n, 634 agreement, 409 allocation, 409, 410, 411, 415 animate see animate assignment, 91, 92, 98, 105n, 263, 266, 275, 276, 415, 456, 540 biological see gender: natural change see gender shift feminine, 6, 9, 54, 57, 75, 78, 81–4, 86, 88, 91–5, 97, 99, 100, 102, 103, 106n, 109, 114, 115, 118, 119, 123, 127, 128, 131, 133n, 137–9, 154n, 191, 207–9, 215n, 239, 240, 242, 244n, 253–8, 266–8n, 272–5, 293, 296–9, 303, 324–6, 330, 331n, 332, 337, 375, 386, 390, 391, 394, 395, 397, 409–11, 414, 415, 418, 420, 422n, 453–60n, 466, 468–70, 507, 508, 510, 512, 514, 518, 571, 572, 574, 575, 626–8, 630, 631, 633n grammatical, 216, 272–7n, 438, 628, 631 inanimate see inanimate inquorate, 488

Subject index 711 marker, 7, 16, 417, 466, 468, 523, 537, 538, 540, 541n marking see gender: marker masculine, 3, 9, 19n, 54, 57, 76, 77, 81–4, 86–8, 91–7, 100–4, 106n, 108, 111, 113–16, 118, 119, 123, 125, 127, 130, 133n, 137, 138, 150, 191, 205, 206, 208, 209, 212, 215n, 239, 240, 253–8, 263, 266, 267, 272–6, 297, 298, 303, 323–5, 332, 375, 409–11, 414–17, 420–2n, 453, 454, 457–60n, 466, 468, 469, 507–10, 514n, 626–9, 631, 633n natural, 99, 275, 409, 628, 631 neuter, 54, 78, 84, 92, 93, 96, 97, 100, 103, 105n, 106n, 125, 266–8, 272–7n, 297, 299, 303, 332n, 418–22n, 537, 539 prefix, 99, 537, 538, 540, 541n shift, 6, 10, 40, 54, 85, 91–3, 95–7, 99, 100, 106n, 190, 459, 489, 491–3 system, 91–3, 98, 99–101, 409, 454, 537, 540, 541 value, 91, 270, 272–7n gnd see gender grammatical aspect see aspect gender see gender: grammatical grammaticalisation, 36, 67, 99, 106n, 142, 144, 146–50, 152, 153, 177, 192, 217, 307, 346, 352, 362, 364–6n, 382, 384, 388, 406, 415, 425, 477, 479, 492, 494, 617, 618, 621, 650 see habitual habitual, 111, 180, 182, 213, 336, 343, 345, 350, 393, 418, 420, 421, 427, 432, 500, 504, 546, 575, 576, 609, 645, 646, 648 habitualness see habitual high tone see pitch accent hon see honorific honorific, 6, 63, 67, 70, 71, 72, 124, 130, 148, 153n, 192, 291, 295n, 331, 381–3, 387, 388, 496, 497, 499, 503, 504 augmentative see augmentative: honorific honorification see honorific hypocorism see hypocoristic hypocoristic, 9, 11, 12, 16, 39, 41, 48, 63, 64, 125, 134, 136–40, 192, 220, 221, 223, 224, 228, 267, 307–9, 335, 437, 566 affix, 135 infix, 48 suffix, 136, 138, 248 value, 154n hab

ideophone, 190, 191, 246, 251, 252, 338, 414, 472, 478, 479, 493, 495n, 516, 519, 520, 638, 643 im see inflectional: morphology imperfective see aspect: imperfective implicational universal see universal: implicational inan see inanimate inanimate, 52, 93, 94, 248–50, 254, 257, 288, 290, 297, 320, 335, 401–4, 407n, 409, 443, 457, 487, 492, 493, 518, 569, 571, 572, 575, 591, 619, 626, 635 incp see inceptive incep.r see inceptive: reduplication inceptive, 131, 132, 536–9, 541n reduplication, 77, 404

incorporation, 252, 367, 406, 430, 431, 559, 560, 563, 634 individuation, 66, 103, 104, 144, 151, 335, 627, 632 infix, 177, 243, 287, 304n, 321, 367, 375, 462, 550, 554, 621 diminutive see diminutive: infix evaluative see evaluative: infix hypocoristic see hypochoristic: infix inflectional see inflectional: infix iterative see iterative infix pejorative see pejorative: infix reduplicated see reduplicated infix infixation see infix infl see inflectional: morphology inflection see inflectional: morphology inflectional affix, 3, 75, 78, 80, 136, 172, 375, 430 affixation see inflectional: affix agreement, 504 allomorphy, 545 category, 81, 143, 616, 617 class, 46, 78, 81, 106n, 110, 138, 270, 275–7, 543 ending see inflectional: suffix feature, 89, 216, 262, 508 infix, 530 marker, 264, 270, 617 morpheme, 3, 46, 85, 101, 177, 191, 197, 202, 279, 530, 536, 561, 562, 564 morphology, 3, 4, 6–9, 15, 19n, 43, 45, 47, 49, 59, 61, 63, 64, 66, 71, 72n, 74–6, 79, 80, 83, 85, 87–90n, 106n, 110, 134, 135, 138, 140, 141, 143, 156, 172, 195, 209, 236, 245n, 252, 264, 284, 287, 293–5n, 320, 352, 367, 404, 408, 411, 416–21, 449, 453, 459, 460n, 496, 504, 505, 507, 529, 530, 537, 538, 542, 543, 561–3, 565, 567, 568, 572, 606, 614, 635 paradigm, 246, 545 pattern, 140, 180, 546 prefix, 537 process, 15, 43, 44, 46, 310, 459 rule, 4, 436 suffix, 19, 47, 85, 105, 125, 137, 191, 197–9, 202, 203, 277n, 287, 293, 295n, 306, 319, 327, 334, 421, 430, 508, 524, 527, 562 system, 448, 508 inquorate gender see gender: inquorate intensification see intensive intensifier, 45, 552, 648, 649 intensifying see intensive intensity see intensive intensive, 9, 11, 12, 15, 25, 28, 29, 36, 40, 41, 43, 45, 48, 49, 52, 55, 56, 59, 62–4, 66, 77, 109, 111, 120, 144, 157–9, 161–8, 172, 175, 180, 181, 188, 191, 192, 203, 213, 215, 216, 220, 221, 224, 225, 233, 235, 246, 250–2, 279, 281–5, 291, 295n, 300, 310, 311, 316–18, 320, 325, 327–31, 333, 336, 338, 340, 343–6, 349, 350, 356–8, 361, 365, 367, 372–4, 376–81, 383–6, 388, 389, 391–3, 395, 397–9, 402, 414, 415, 426, 427, 430–4, 436, 437, 443, 446, 449–51, 466, 467, 476, 478, 479, 494, 516, 519–21, 524, 525–8n, 533, 548, 553, 554, 556, 566, 593n, 601, 602, 604, 608–10, 612, 615, 618, 619–21, 623, 628, 629, 632, 638–43, 648–50

Subject index

712 interference see contact interfix, 33, 35, 40, 77, 84, 87, 102, 104, 131, 137, 264, 265, 292 interfixation see interfix intf see interfix intimacy, 25, 27, 139, 145, 149, 220, 267, 308, 402, 405 ints see intensifier ipfv see aspect: imperfective ironic, 27–9, 32, 37, 38, 40, 41, 133n, 140, 144, 212, 230, 255, 266, 297, 311, 553, 612 irony see ironic irr see irrealis irrealis, 117, 179, 346, 347, 349, 367, 371, 372, 418, 550, 554, 557, 613, 636, 645, 647 irresultative see aspect: irresultative isolating morphology, 180, 252, 352, 472, 479 iter see iterative verb iteration see iterative iterative infix, 530 reduplication see reduplication: iterative verb, 43, 45, 48, 59, 62–4, 104, 112, 115, 177, 179, 180, 213, 228, 233–5, 259–61, 282, 388n, 404, 427, 503, 530, 539, 622, 623, 634, 639, 640 iterativeness see iterative verb iterativity see iterative verb iter.r see reduplication: iterative see kinship kinship, 125, 192, 193, 267, 311, 312, 314, 335, 337, 354, 357, 362, 363, 380, 402, 407n, 444, 449–51, 489, 504, 534, 540, 569, 582, 587, 591, 619, 620, 623, 630, 631

kin

language acquisition see acquisition language contact see contact language documentation see documentation language interference see contact language universal see universal language level-ordering hypothesis, 46, 47 lexical aspect see actionality diminutive see diminutive: lexical reduplication see reduplication: lexical vowel see vowel: lexical lexicalisation, 85, 87, 89, 106n, 116, 131, 134, 146, 153n, 208, 209, 215n, 217, 229, 233, 236, 266, 268n, 271, 277n, 307–9, 313, 317, 355, 357, 363, 368, 375, 378, 384, 386, 408, 412, 419, 425, 430, 444, 491, 498, 531, 533, 548, 549n, 553–6, 582–5, 588, 593, 593n, 600, 626, 629, 646, 649 lexicalised diminutive see diminutive: lexicalised lf see focus: locative lim see limitative suffix limitative suffix, 596–8, 600, 603, 605n littleness, 21, 24–8, 139, loanword, 217, 240, 242, 307, 308, 335, 387, 557, 626, 641, 642n locative focus see focus: locative log see logophoric marking logophoric marking, 473, 476

m see gender: masculine mal see malefactive

malefactive, 647, 648 manual morphology, 155–9, 161–9 masc see gender: masculine masculine see gender: masculine mass noun see uncountable noun meliorative see ameliorative metaphor, 22, 24, 26–8, 30, 31, 39, 41, 62, 66, 100, 125, 143–5, 148, 177, 294, 389, 391, 393–5, 398, 399, 439, 442, 531, 533, 591 metaphorical see metaphor metonymic see metonymy metonymy, 24, 26, 27, 30, 143, 144, 146, 148, 150, 153, 154n, 581 mid voice see voice: middle middle voice see voice: middle minimal degree, 370, 371 mitigation, 129, 148, 231 modificational category, 7, 297 mood, 23, 123, 133n, 249, 250, 278, 293, 417, 430, 516, 520, 523, 559, 561, 567n, 606, 611, 612, 614, 617, 625 morpheme order, 196, 588, 593 morphopragmatic see morphopragmatics morphopragmatics, 4, 5, 7, 16, 33–8, 40, 41, 122, 129–31, 134, 187, 473, 474 motherese see baby talk natural gender see gender: natural neuter see gender non-canonicity, 117, 119, 120 non-iconic reduplication see reduplication: non-iconic non-linear morphology, 354, 357, 359 non-manual morphology, 155–7, 161, 163–5, 167, 168 noun animate see animate countable see countable noun inanimate see inanimate mass see mass noun noun class, 81, 88, 98, 99, 105, 431, 481–3, 485–9, 492, 498, 499–503, 506, 515, 516, 519–21 agreement, 488, 504, 516 marker, 88, 99, 247, 248, 250, 472, 501–3, 515, 637 prefix, 487, 488, 496, 498, 500, 501, 503, 505, 515 suffix, 486 ntr see gender: neuter number dual, 81, 342, 347, 416, 417, 424, 426, 428 singular, 3, 19, 70, 71, 77, 81, 83, 85–7, 94, 95, 98, 101–3, 106n, 108, 113–16, 118, 119, 202, 205, 206, 208, 221, 222, 249, 254–6, 258, 263–8n, 277, 297–300, 303, 304, 306, 310, 326, 329, 332, 342, 369, 383, 397, 409–11, 414, 417–19, 428, 434, 437n, 454–8, 460n, 481–3, 485, 486n, 488, 497, 498, 500, 502, 503, 507–10, 512, 514n, 560–2, 564, 566, 567, 574, 616, 638, 646–9 paucal, 103, 104 plural, 3, 46, 47, 55, 60n, 71, 81, 85, 97, 98, 101, 102, 104, 106n, 107n–9, 112, 114, 133n, 136–8, 182, 190, 191, 198, 202, 209, 220, 222, 240,

Subject index 713 244n, 247, 248, 250, 262–8n, 280, 281, 287, 306, 309, 310, 327, 330, 334–6, 338, 342–4, 347, 348, 361, 362, 371, 373, 380, 417, 421, 424, 428n, 446, 454, 455, 460n, 472, 473, 478, 481–5, 488, 497, 498, 500, 502, 503, 507–11, 513n, 514n, 517, 523, 536, 537, 557, 561, 585, 588, 589, 616, 626, 635–9, 644–6, 648 onomasiological category, 7, 188, 297 palatalisation affective see affective: palatalisation expressive see expressive: palatalisation part see partitive partial reduplication see reduplication: partial partitive, 25, 66, 103, 104, 111, 112, 144, 145, 149, 203, 343, 348–50, 391, 493, 531, 632 passive see voice: passive patient focus see focus: patient paucal see number: paucal paus see clitic: pausal pausal clitic see clitic: pausal pej see pejorative pejoration see pejorative pejorative, 3, 4, 6, 10–12, 15, 23, 24, 27–9, 33, 35, 37, 38, 40, 41, 62–4, 67, 70, 72, 76, 96, 129, 134, 136, 137, 142, 145, 148, 151–3, 154n, 191, 192, 201, 210, 211, 213, 214, 218, 219, 225, 234, 238, 249, 253, 255, 261, 273, 277n, 284, 285, 292, 294, 295n, 297, 311, 325, 326, 329–31, 338, 339, 356, 359, 376, 379, 384, 386, 388, 399, 412, 416, 443–5, 451, 458, 476, 493, 496, 500–3, 517–19, 521, 527, 530, 538–41n, 544, 554, 568, 570, 571, 574, 577, 604, 607, 618, 623, 630, 632, 633n, 639, 640 acquisition see acquisition: of pejorative affix, 151, 152 agentive meaning, 76, 96, 97, 106 circumfix, 225 infix, 379 meaning, 148, 151, 154n, 191, 225, 234, 253, 255, 261, 285, 331, 356, 359, 384, 386, 388, 412, 444, 517–19, 521, 539, 554, 568, 577, 630, 633n pronominal affix, 63, 70, 71 suffix, 152, 211, 623 perception, 10, 16, 21–3, 25–7, 30, 40, 139, 156, 170, 370 perfective see aspect: perfect pf see focus: patient pfv see aspect: perfective phonetic iconicity, 60, 65, 67, 69, 72n, 122, 135, 192, 208, 551 pitch accent, 320n, 507 pl see number: plural pleonastic, 545, 547, 548 pluractional see pluractionality pluractional reduplication see reduplication: pluractional pluractionality, 6, 16, 43, 44, 48, 61, 62, 104, 105, 107n, 108, 110, 111, 115–17, 238, 239, 242–4, 461, 466, 470, 511, 512 plural see number: plural

pluralisation pattern see number: plural see politeness polite see politeness politeness, 26, 37, 41, 71, 123, 124, 129, 148, 149, 231, 295, 312, 338, 348, 349, 364, 373, 550, 552, 553 polysemous see polysemy polysemy, 25, 28, 35, 102, 144–6, 148, 151, 268, 459, 475, 496, 498, 499, 617, 621, 623 polysynthesis see polysynthetic morphology polysynthetic morphology, 278, 285n, 401, 430–3, 436, 437, 536, 559, 560, 567, 568, 625, 634, 643 pot see aspect: potential potential see aspect: potential pragmatic see pragmatics pragmaticalisation, 148 pragmatics, 4, 5, 7, 8–11, 15, 16, 19, 20n, 21, 24–6, 30, 32–41, 60, 63, 64, 71, 78, 80, 90n, 120, 121, 126, 129, 132, 134, 135, 138–41, 144, 145, 148, 149, 151–3, 156, 159, 164, 170–2, 176, 227, 230, 231, 233, 234, 236, 323, 329, 331, 348, 349, 398, 401, 404, 410, 415, 473, 476, 494, 535, 563, 633n, 641 prefix ameliorative see ameliorative: prefix augmentative see augmentative: prefix derivational see derivational: prefix diminutive see diminutive: prefix evaluative see evaluative: prefix gender see gender: prefix inflectional see inflectional: prefix noun class see noun class: prefix privative see privative: prefix prefix–suffix neutrality, 79, 81, 90n prefixal diminutive see diminutive: prefix prefixation see prefix pretence, 380 primary diminutive see diminutive: primary priv see privative privative prefix, 289 suffix, 425, 442 proc see aspect: procrastinative processive verb, 111, 279, 283 proclitic see clitic procrastinative see aspect: procrastinative productivity, 43, 56, 78, 85, 89, 97, 109, 110, 125, 143, 211, 213, 215, 215n, 220, 224, 270, 274, 279, 307, 355, 387, 540, 563, 564 prog see aspect: progressive progressive see aspect: progressive proh see prohibitive prohibitive, 640 prosodic constraint, 239, 264 features, 156, 163 morphology, 244 phonology, 64, 69, 72n shift, 508 structure, 264 template, 239 prosody, 40, 155, 156, 242, 252n, 264 pol

714 prototypicality, 10–12, 63, 64, 167, 430, 431, 433, 436, 437, 439, 442, 526, 630, 632 punctual verb, 228, 233, 235 purp see purposive purposive, 373 quadrilateral bases, 461–3, 468–70 qualitative verb, 278–83, 285 Quantity of Action, 43, 50, 67, 68, 70 71 of Circumstance, 43, 50, 67, 68, 70, 71 of Quality, 43, 50, 67–71, 73n of Substance, 43, 50, 67, 68, 70, 71, 73n quinquilateral bases, 461, 463, 464, 470 quot see quotative quotative, 132 see realis radial category, 5, 7, 15, 24–8, 30, 66, 70, 144, 145, 149, 151, 152, 154n, 188 rdc see reduction rdp see reduplication realis, 77, 82, 369–71, 373, 374, 402–5, 550, 639 reanalysis, 143, 148 recursion see recursive recursive, 54, 55, 59, 82, 83, 89, 137, 209, 302, 338, 546, 559, 565, 609 recursiveness see recursive recursivity see recursive red see reduplication redoubling see reduplication reduction, 9, 11, 12, 36, 50, 63, 64, 116, 128, 160–2, 168, 169, 289, 290, 317, 349, 350, 372, 507, 512, 513, 597, 604 reduplicated base, 51 diminutive see diminutive: reduplication infix, 48 verb see verb: reduplicated reduplication, 6, 9, 10, 13, 16, 32, 33, 36, 40, 45, 47, 49–51, 54, 56, 57, 59, 62, 68, 74, 82, 90n, 107n, 109–11, 126, 134, 146, 153, 153n, 158, 165, 166–8, 170–83, 183n, 190, 203, 216, 221, 225, 228, 236–8, 242–5n, 265, 279–81, 283–5n, 286n, 292, 302, 320, 321, 327, 328, 343–6, 350, 351n, 353, 354, 357, 358, 360n, 367–70, 372–4, 374n, 376–80, 383–6, 388, 388n–91, 393–9, 404, 406, 408, 414, 415, 421–5, 427, 428, 432, 434, 445, 452, 452n, 469, 470, 472, 478, 494, 507–9, 511–13, 513n, 514n, 516, 519–21, 525–7, 530, 543, 550–3, 555, 557, 576, 577, 598–601, 603, 604, 606, 609–11, 614, 617–23, 638, 643, 648–50 attenuative, 378 diminutive see diminutive: reduplication emphatic, 196, 203 evaluative see evaluative: reduplication full, 49, 82, 158, 165, 166, 190, 368, 369, 372, 373, 379, 380, 389, 525, 527, 610 iconic, 179 inceptive see inceptive: reduplication intensificational, 203 iterative, 48, 179, 180, 182, 328, 388n, 404, 427 lexical, 171, 177, 178 r

Subject index non-iconic, 174–6 partial, 373, 374, 389, 390, 525, 527, 553, 610, 643, 648 pattern, 107n, 181, 239 pluractional, 243 verbal see verb reduplicated rep see reportative reportative, 644–4, 648 repetition see reduplication res see aspect: resultative resp see respectful respectful, 131, 322–5, 355, 543 resultative see aspect: resultative rhythmic lengthening, 533, 534 SAE see Standard Average European sarcasm, 27, 29, 33, 41, 140, 311, 330 scepticism, 40 second language acquisition see acquisition: second language secondary diminutive see diminutive: secondary seniority, 355, 477–9 sentence diminutive see diminutive: sentence serial construction see serial verb serial verb, 171, 182, 286n, 388, 473, 480 sg see number: singular singular see number: singular singulative, 5, 46, 94, 102–4, 248, 455, 459 circumfix, 455 size, 5, 6, 10, 13, 14, 20n–31, 39, 44, 54, 62, 67, 76, 77, 94, 98, 99, 106n, 109, 134, 139, 145, 151, 157–65, 168, 169, 172, 191, 210, 210, 218, 223–5, 246, 248, 254, 266, 271, 279, 285, 288, 294, 307, 309, 311, 313, 334, 335, 343, 358, 363, 365, 371, 381, 389, 391, 394, 395, 398, 402, 403, 409, 415, 424, 431–3, 439, 442, 443, 465, 470, 474–6, 478, 479, 489, 490, 492–4, 496, 498–500, 502, 503, 505, 506, 517, 531–3, 537, 538, 560, 581, 583, 584, 593, 596, 598, 602, 626, 629–32, 635–7, 641, 644, 646, 647 slang, 34, 308, 309, 375, 378 smallness, 23, 25, 26, 36, 38, 39, 57, 65, 70, 139, 146, 148, 163, 227, 230, 248, 266, 297, 311, 331, 356, 406, 410, 473, 474, 479, 492, 569, 593n, 646 sngl see singulative social group, 25, 122, 124, 126, 131, 145 sound symbolism, 6, 53, 65–7, 69, 70, 72n, 73n, 192, 294, 340, 551, 569, 573 speaker-oriented, 170, 496, 562 speech accommodation, 125 st see stative verb Standard Average European, 15, 47, 48, 59, 67–9, 72, 73n, 189, 191 stative classifier, 543 preverb, 542 theme, 541 verb, 119, 203, 234, 283, 285n, 286n, 361–5, 366n, 367–8, 374, 376–9, 390–6, 398, 432, 433, 435, 529, 538, 596 suffix ameliorative see ameliorative: suffix augmentative see augmentative: suffix

Subject index 715 derivational see derivational: suffix diminutive see diminutive: suffix evaluative see evaluative: suffix hypocoristic see hypocoristic: suffix inflectional see inflectional: suffix limitative see limitative suffix noun class see noun class: suffix pejorative see pejorative: suffix privative see privative: suffix suffixal diminutive see diminutive: suffix sup see superlative superlative, 43, 44, 202, 203, 211, 212, 214, 221, 228, 235, 236, 287, 293, 295n, 378, 379, 392, 398, 426, 519, 544–7, 554–7, 609, 648 augmentative see augmentative: superlative circumfix, 221, 379 suprasegmentality, 155, 156, 353, 355, 358 sv see stative verb telic verb, 110–13, 115, 117, 120n telicity see telic verb template, 93, 238, 239–44, 245n, 250, 462, 463, 465, 466, 468, 470, 471n, 536, 541n, 568, 643, 649 diminutive see diminutive: template templatic diminutive see diminutive: template tense, 179, 249–51, 277n, 278, 347, 352, 361, 389, 416–18, 420, 423, 430, 448, 480, 523, 529, 559, 567n, 606, 617, 625, 634 TMA, 176, 177, 181, 279, 367, 404, 430, 431 tns see tense tonality, 70 tone, 40, 172, 190, 191, 214, 231, 252, 252n, 352, 354–9, 359n, 362, 364, 366n, 384, 388, 388n, 472, 473, 475, 476, 478–86n, 487, 493, 495n, 496, 503, 505, 507–9, 513n, 514n, 523, 524, 542, 584, 593n transcategorisation see conversion trilateral bases, 461, 462, 467–70 triviality, 473, 475, 479 uncountable noun, 78, 103, 104, 111, 112, 150, 139, 170, 177, 248, 257, 335, 402, 403, 454, 455, 457, 458, 489, 491, 518, 627 universal category, 9, 188 EM, 61–4, 66, 69, 71, 72, 73n, 192, 238 function, 9, 70 implicational, 63, 67, 68, 70, 452 language, 15, 61, 63, 64, 66, 70–2, 72n, 73n, 106, 144, 174, 175, 338 morphological, 71

pattern, 13, 238 prime, 23 prototypes, 13 sociolinguistic, 124 statistical, 63, 66 tendency, 15, 71, 182 universal quantification, 353, 357 Universals Archive, 15, 63–6, 69, 71 see veneration veneration, 289, 291, 292 verb activity see activity verb atelic see atelic verb durative see durative verb dynamic see dynamic verb frequentative see frequentative verb iterative see iterative: verb pluractional see pluractionality punctual see punctual verb qualitative see qualitative verb reduplicated, 176, 180, 329, 358, 374, 397, 452, 452n, 511, 514, 520, 553 serial see serial verb verbal reduplication see verb reduplicated verbal serialisation see serial verb voc see vocative vocative, 130, 137, 323, 335, 367, 640 voice, 375, 497, 498, 503, 504 active, 242, 625 middle, 625, 647 passive, 367, 518 voicing, 90n, 279, 281, 283, 284, 503, 587 ven

see word formation word formation, 4–7, 10, 15, 34, 40, 43–9, 51, 53, 59, 61, 63, 64, 66, 72, 72n, 73n, 74, 77, 79, 106n, 120n, 121, 122, 125, 130, 131, 177, 188, 190, 253, 256, 257, 262, 266, 268, 268n, 270, 271, 279, 285, 287, 288, 293, 306, 307, 320n, 353, 354, 430, 436, 453, 459, 527, 529, 532, 535, 537, 542, 606 word order, 72n, 195, 216, 278, 352, 375, 417, 423, 431, 438, 448, 473, 515, 523, 529, 542, 596, 606, 625, 634, 636 wf

youth language, 126, 127 zero-derivation see conversion

Name index

Name index

Abdel-Massih, E. T., 94, 95, 455, 456, 458, 460n Aboh, E. O., 172, 173, 175, 176 Adinolfi, A., 121 Agbetsoamedo, Y., 82, 88, 99, 106n, 107n, 190–3, 487 Agee, M., 542 Aguirre, C., 136 Ahenakew, F., 576 Aikhenvald, A. Y., 408, 413, 415, 489 Akhalaia, P., 224 Aksu-Koç, A., 339 Albalá, M. J., 136 Allan, K., 98 Allen, J. H. D., Jr., 95 Alleyne, M., 174 Alonso, A., 23, 36 Alves, M., 50 Ambrazas, S., 6, 154n, 254 Ambrazas, V., 253, 254 Ameka, F. K., 475, 478 Amery, R., 423, 424, 428n Amfo, N. A. N., 6, 146, 474 Amha, A., 57 Anderson, S. R., 101, 102 Androutsopoulos, J. K., 127, 133n Andrzejewski, B. W., 507, 508, 511, 513, 513n, 514n Anonby, E. J., 150 Appah, C. K. I., 6, 146, 474 Aquilina, J., 93 Arabuli, A., 223, 224 Arbaoui, N., 106n, 109, 111, 466 Arcodia, G. F., 353 Aronoff, M., 76, 156 Aronson, H., 222 Asma, S., 23, 24 Aspinion, R., 455, 456, 458 Augenstein, S., 126 Austin, P., 416 Authier, G., 45, 52, 56 Avram, A. A., 172 Avrorin, V. A., 340 Avrutin, S., 140 Awbery, G., 47 Azzopardi-Alexander, M., 103 Baayen, R. H., 121 Bacelar, L., 607, 608 Bach, K., 32 Badarneh, M. A., 130 Badia i Margarit, A. M., 211 Bakker, P., 172, 180, 568

Banti, G., 507, 508, 513, 514n Barcza, V., 137, 139 Bar-el, L., 111 Barsalou, L., 22 Barz, I., 148, 263 Basset, A., 456 Basso, K. H., 537 Batchelor, R. E., 376 Bat-El, O., 242, 244n Bates, D., 135, 423 Bates, E., 135 Bateson, G., 408 Bauer, L., 4, 6, 21, 43, 60n, 61, 63, 65, 69, 70, 72n, 73n, 76, 136, 141, 143, 146, 154n, 278, 279, 304n, 343, 350, 436, 606, 613, 627, 631 Beard, R., 4, 6, 67, 70, 74 Beilin, H., 26 Belič, A., 144, 148 Bellugi, U., 155, 156 Benko, V., 296 Bentley, M., 65 Bentolila, F., 456, 457 Bergen, B. K., 21 Berger, B., 21 Bergman, B., 163 Berko, J., 135 Berlin, B., 65, Berman, H., 6 Bernal, E., 83, 86, 107n, 214, 215n Berruto, G., 122–4 Berta, A., 314, 315 Bertrang, A., 263 Besten, H. den, 79 Bickel, B., 354 Bickford, J. A., 164 Birioukova, O., 107, 192 Bittner, M., 468 Black, J. M., 423, 424 Blank, A., 143 Bleek, W. H. I., 515 Blench, R., 487 Blevins, J., 48, 154n Blevins, J. P., 140 Blust, R., 389 Bodor, P., 137, 139 Boeder, W., 216, 226 Böhmerová, A., 6 Boldyrev, B. V., 334, 340 Bolinger, D., 118 Bolozky, S., 6, 244n Bonifacio, S., 135 Booij, G., 6, 43, 75, 76, 85, 87, 266, 276

Name index 717 Borg, A., 103 Bornstein, M., 22 Boyes-Braem, P., 169 Brackbill, Y., 65 Bradley, D., 192, 361, 362, 364, 366n Bradley, M., 366 Brill, I., 52 Brincat, J., 53, 58, 62 Brooks, P. J., 135 Brown, D., 401 Brown, L., 130 Brown, P., 71, 123 Bruch, R., 263 Brugmann, K., 147 Bruyne, J. de, 214 Buell, L., 520 Buffa, F., 269 Bühler, K., 142 Buletangsuu Temwakon, P., 351 Butler, J. L., 8 Bybee, J. L., 85 Cabré, M. T., 211, 616 Cabrera, P., 616 Caffi, C., 38, 129 Cahill, M., 191, 485, 486n Cai, Y., 356 Calandrelli, M., 616 Cameron, D., 132 Camino, I., 199 Campos, G., 191 Cañete, P., 214 Canonici, N. N., 515 Cardona, G. R., 123 Carlson, R., 46 Carpio, M. B., 633n Carstairs-McCarthy, A., 4, 74 Carstens, V., 515 Casad, E., 549 Caubet, D., 103 Ceccherini, M., 135, 136, 139 Cecchini, M., 95 Chadwick, N., 416, 418, 421 Chang, H.-H., 389, 393 Chantraine, P., 96 Chapman, C., 264 Charleston, B. M., 125 Chiari, I., 16, 155 Chimke, H., 222 Chumburidze, Z., 224 Cinque, G., 565 Clahsen, H., 140 Clark, E. V., 124, 138 Claudi, U., 142, 473 Coates, J., 132 Coletti, V., 95 Cominetti, F., 356 Compton, R., 82, 85, 191, 562, 565, 567 Comrie, B., 352 Constenla, A., 523, 527, 528 Cook, C., 565, 567 Cook, E.-D., 540 Cook, T., 520

Corazza, S., 168 Corbett, G., 91, 93, 98, 100, 401, 488, 489 Cote, M., 570 Coulson, S., 22, 30 Cowan, G., 543, 548 Craig, C., 98 Creider, C. A., 515 Creissels, D., 8, 46, 92, 99, 143, 144, 146, 147, 150, 473, 492 Crevels, M., 606 Croft, W., 77, 153 Csató, É. Á., 320 Cúneo, P., 626, 630, 633n Cusic, D., 108, 109, 511 Cutfield, S., 401 Cuturi, F., 554, 555 Cuxac, C., 155, 165, 169 Da Luz, M. A. M., 102 da Silva Júnior, A. R., 607 Dąbrowska, E., 135 Dahl, Ö., 148, 151, 152, 232, 307 Dal, G., 36, 39 Dale, D., 515 Dallet, J. M., 455 Daltas, P., 6, 97, 274, 275, 277 Dalton-Puffer, C., 121 Dardano, M., 213 Daval-Markussen, A., 172, 180 De Belder, M., 244n De Marco, A., 124, 135 de Reuse, W. J., 561, 564, 567, 591 de Rijk, R. P. G., 154n, 198–201, 203 de Vinciennes, L., 455 Déchaine, R.-M., 80, 83, 88, 98, 106n, 134, 191, 492, 499, 501, 503, 505 DeLancey, S., 387 Delbrück, B., 147 Dell’Aquila, V., 549n Dell’Arciprete, A., 629 Della Bad Wound, 69 Dench, A., 448 Denny, J. P., 515 Derkach, T., 187 Derzhanski, I., 60n, 102 Descartes, R., 22 Di Garbo, F., 6, 80–2, 88, 93, 95, 98, 99, 106n, 107n, 187 Di Renzo, A., 16, 155 Diffloth, G., 6, 65, 123 Dimmendaal, G. J., 44 Dingemanse, M., 472, 478 Dixon, R. M. W., 428n, 440 Doerfer, G., 333 Doke, C. M., 517, 520, 521 Dokulil, M., 7, 67 Doleschal, U., 40 Donaldson, T., 406 Donazzan, M., 109 Dong, M.-N., 389, 393, 395, 398 Dorais, L.-J., 567n Dowty, D., 110, 112 Doyle, A., 69

718 Dressler, W. U., 5, 6, 8, 16, 23, 26, 32–6, 40, 41, 71, 72n, 83–5, 89, 104, 106n, 108, 122, 124, 125, 127–30, 132n, 134–41n, 143, 148, 149, 151, 187, 188, 227, 228, 236, 473 Dryer, M. S., 17, 73n, 189, 190, 542, 604n Dunn, J. A., 305n Duthie, A. S., 472 Dutton, T. E., 51 Dvonč, L., 296 Edwards, J., 124 Elenbaas, M. B., 117 Elizondo, F., 528 Ellegård, A., 308 Elordieta, A., 198 Emmorey, K., 155, 156, 169 Erikson, E., 27 Erschler, D., 52, 58 Eslami, M., 287 Ettinger, S., 6, 65, 144 Evans, N., 107n, 190, 192, 401, 406, 407, 430, 432, 448, 450–2n Evans, V., 21, 23, 24, 30 Fabra, P., 211, 215n Falco, A., 95 Falk, Y., 244n Faltz, L. M., 536 Farquharson, J. T., 176, 177 Farshidvard, K., 288 Fasold, R., 122 Faust, N., 104, 107n, 111, 240, 244n Fedorova, O., 135 Fehlen, F., 262 Fejes, L., 44, 49 Fennig, D. C., 17, 321, 437n, 484, 604n Fenyvesi, A., 233 Ferguson, C. A., 124, 125, 135 Fernández, S., 214 Filimonova, E., 63 Finney, M. A., 182 Fischer, S. D., 155, 156, 173 Fisher-Jørgsen, E., 65 Fivaz, D., 520 Fleck, D. W., 55 Fleischer, W., 148, 263 Folena, G., 133n Foley, W. A., 408, 414 Fortescue, M., 559, 561, 567 Fortin, A., 6, 7, 90n, 143, 188, 496 Fortune, A., 106 Fortune, G., 498, 501 Forza, F., 173 Fradin, B., 6, 33, 36, 39, 40, 134 Frampton, J., 111 Franceschet, R., 135, 141 Francis, B., 48, 56 Franco, L., 8 Fraychineaud, K., 164 Furdík, J., 296 Gaeta, L., 16 Gaide, F., 8, 106n, 150

Name index Gaillard-Corvaglia, A., 549n Galucio, A. V., 608 Gambino, M., 187 Gamillscheg, E., 150, 151 Ganenkov, D., 57 Ganiyev, F., 311, 313 Gazdeliani, E., 224 Gené, M., 214 Genesin, M., 48 Genetti, C., 58 Georg, S., 246 Gerzenstein, A., 634 Ghomeshi, J., 424 Giles, H., 125 Gili Fivela, B., 40 Gilij, F. S., 624n Gilles, P., 106n, 191, 263 Gillis, S., 8, 135 Girard, R., 80, 98, 134, 191, 492, 499 Givón, T., 172 Gnerre, M., 192, 554, 555 Gogolashvili, G., 217, 219, 221, 222 Goldenberg, G., 243–5n Golla, V., 592 Golluscio, L., 623, 642 Golston, C., 549 González, H. A., 55, 623n Gooch, A., 6 Goode, P., 591 Goodman, M. F., 171 Gorelova, L. M., 340 Gràcia, L., 72n, 205, 211, 215n Grandi, N., 4, 6, 8, 13, 15, 16, 29, 41, 43, 77, 78, 80, 84, 85, 88, 89n, 92, 94, 98, 106n, 107n, 110, 116, 120, 121, 127, 142–7, 149–53n, 154n, 172, 187, 190, 213, 232, 253, 245, 260, 261, 271, 274–6, 343, 407n, 411, 431, 434, 436, 445, 453, 459, 460n, 513n, 516, 560, 565, 606, 607, 614, 625, 634 Gray, A., 342, 351 Greenberg, J. H., 106n, 264, 338, 473 Greenberg, Y., 242, 245n Gregová, R., 6, 60n, 65, 69, 100, 136, 141, 191, 297, 304 Grenoble, L. A., 333 Grönke, U., 6, 53 Grossmann, M., 152 Gruzdeva, E., 191 Gudschinsky, S. C., 549n Günthner, S., 148 Gustafson, H. S., 45 Guthrie, M., 496, 517, 518 Gwynn, J. P. L, 324, 328 Haag, M., 193, 529 Haase, M., 154 Hagman, R. S., 58, 106n Hakamies, R., 8, 144 Hale, K., 438, 443, 447n, 449, 451 Halle, M., 565 Han, Y., 355 Hardman, M. J., 54, 107n, 192, 595–8, 600–5n, 606n

Name index 719 Hargus, S., 537, 579, 589, 590 Harley, H., 565 Haspelmath, M., 5, 9, 17, 73n, 143, 148, 188–90, 542, 604n Hasselrot, B., 6, 97, 133n, 142, 147 Haude, K., 46 Haverling, G., 8 Hayashi, M., 567n Hayeková, M., 297 Head, B. F., 63, 70 Healey, P. M., 58 Heilmann, L., 97 Heine, B., 142, 143, 146–8, 152, 473 Heinz, A., 262 Henzen, W., 154n Hercus, L. A., 68, 70, 425 Herms, I., 6 Hervás y Panduro, L., 624 Hewson, J. I., 48, 56 Hill, J. H., 130 Hill, K. C., 130 Hill, N. W., 381, 388n Himmelmann, N., 389 Hinchliffe, I., 306, 309 Hinton, L., 122 Hisamova, F. M., 317 Hoeksema, J., 52 Hohenhaus, P., 51 Holmes, J., 123, 133n Holmes, P., 306, 309 Holton, D., 92 Holton, G., 540 Hoogenraad, R., 445 Hope, E. R., 366 Hopper, P., 142, 152, 153n Horecký, J., 296, 297, 302 Horiguchi, D., 259 Horner, K., 262 Hoshi, M., 383, 385, 388n Hualde, J. I., 57, 197, 204 Huang, Y., 473 Huayong, L., 146 Hunnemeyer, F., 142, 473 Hunston, S., 23 Hutchinson, J. P., 44 Hyman, L. M., 46, 507, 514n Iacobini, C., 121 Inchaurralde, C., 39 Ingram, A., 50 Inkelas, S., 50, 111 Izre’el, S., 245 Jackendoff, R., 104, 424 Jaeggli, O., 6 Jamieson, C., 542, 543, 549n Janer, F., 214 Janhunen, J., 333 Jannaris, A. N., 143, 148 Janney, R. W., 38 Javarek, V., 93 Jelaska, Z., 69 Jendraschek, G., 94, 107n, 109, 111, 191, 408

Jespersen, O., 65, 70 Jetté, J., 536, 540 Jiménez de la Espada, M., 616 Johanson, L., 310, 320 Johns, A., 55, 559, 560, 565, 567 Johnson, M., 22, 24, 179 Johnston, T., 157, 161, 167 Jones, E., 536, 540 Josefsson, G., 309 Jurafsky, D., 5, 7, 8, 15, 24–8, 37, 66–7, 70, 71, 92, 100, 129, 139, 142, 144–6, 148–50, 154n, 188, 192, 356, 391, 401, 432, 439, 443, 445, 473, 485, 491, 492, 531–4 Kačmárová, A., 187 Kadima, M., 98, 501 Kahane, H., 6, 106n Kahane-Toole, R., 6, 106n Kairov, S., 305n Kalbasi, I., 288, 294 Kalme, V., 254–6, 258, 259 Kalnača, A., 106n, 255, 259 Kandybowicz, J., 175 Kardos, E., 113 Kari, J., 536, 537–41n, 588 Karpf, A., 72n Karttunen, F., 63, 147 Kasonde, A., 69 Kasper, G., 128 Katamba, F., 44 Keen, S., 448, 449, 452n Kehrein, W., 549 Keller, R., 142 Kempe, V., 135 Kenesei, I., 233 Kennedy, C., 118 Kerswill, P., 126 Keshani, K., 288 Ketrez, F. N., 50, 55, 339 Khalilov, M., 50, 57 Khalilova, Z., 50, 57 Khumalo, J. S. M., 517, 518, Kiefer, F., 34, 227, 228, 230, 234, 235 Kihm, A., 542, 548 Kilani-Schoch, M., 34, 134, 140 King, K., 27, 135 Kiparsky, P., 112 Kirchner, M., 315, 318 Kirk, P. L., 543, 549n Kitamura, H., 387, 388n Klampfer, S., 140 Klein, H., 625 Klein, T., 172 Klimas, A., 55, 59 Klimaszewska, Z., 271 Klokeid, T. J., 451, 452n Kobakhidze, A., 221, 223 Koeler, H., 424 Köpcke K. M., 93 Korecky-Kröll, K., 16, 124, 130, 135–40 Kormushin, I. V., 333–5 Kornfeld, L. M., 49 Korompay, K., 236

Name index

720 Körtvélyessy, L., 5, 6, 11, 13, 15, 17, 23, 43, 60n, 61, 65, 67–9, 70, 72, 72n, 73n, 120n, 136, 142, 153n, 172, 188, 189, 192, 431, 434, 436, 513n, 516, 533, 560, 565, 606, 607, 614, 626, 634 Kossmann, M. G., 453, 456–8 Kotsinas, U.-B., 126 Kouwenberg, S., 172, 174, 176, 178, 179 Krejnovich, E. A., 281 Krifka, M., 112, 113 Krishnamurti, B., 321, 324, 328 Krjukova, E., 248 Kröger, F., 485 Kruez, R., 28 Krummes, C., 71 Kurdadze, R., 224 Kuteva, T., 146, 147 Kyle, J., 162, 165 La Ode Sidu, 367 Laakso, J., 227 Laalo, K., 8, 36, 135 Labov, W., 128, 172 Labrune, L., 152 LaCharité, D., 174, 176, 179 Ladányi, M., 229, 234, 235 Ladegaard, H. J., 140 Lafone y Quevedo, S. A., 616 Lai, H.-L., 353 Lakoff, G., 7, 24–6, 28, 179, 395 Lakoff, R., 37, 123 Lamb, M., 22 Lamb, W., 48 Lampitelli, N., 105, 107n, 111, 244n, 513n Langacker, R., 22–4 Laoust, B., 94, 454, 457, 460n LaPolla, R. J., 352 Lass, R., 124 Laughren, M., 440, 441, 444, 445, 447n Lázaro Mora, F., 6, 214 Lazzeroni, R., 8, 106n Lefebvre, C., 147 Lehmann, C., 549n Léonard, J. L., 542, 548, 549n Lepskaja, N. I., 135 LeSourd, P., 6, 47 Lettner, L. E., 135 Levinson, S. C., 71, 123 Lewis, M. P., 17, 321, 437n, 484, 604n Li, C. N., 355, 358 Li, F., 333 Li, F.-K., 592 Liashuk, V., 69 Lillo-Martin, D., 165 Lin, H., 353–5, 357 Lindauer, M. S., 26 Little, K. B., 65 Llopart, E., 214 Lloret, M. R., 215n Lokmane, I., 259 Lomtatidze, K., 224 Longobardi, G., 503 Lovick, O., 87, 88, 99, 191, 537, 538, 541, 587–9 Lowenstamm, J., 471n

Lucas, C., 161 Lühr, R., 147 Luo, K., 358, 359 Lynch, J., 48 McCarthy, J. J., 4, 74, 243 Machoni [Maccioni], A., 616, 624n McKay, G. R., 431 MacKenzie, M., 571 Mackeridge, P., 92 McNally, L., 118 McWhorter, J., 172 Madondo, L., 187 Magni, E., 8, 92, 144, 147 Maho, J. F., 496, 501, 504 Maienborn, C., 111 Maisak, T., 57 Makuraté, H., 44 Malcolm, D. M., 517 Malkiel, Y., 151 Malmgren, S.-G., 307 Manova, S., 148 Mansuur, A., 508, 514n Marantz, A., 565 Margery, E., 527n Marrero, V., 136, 137 Martin, J., 534 Martinet, A., 499 Martirosovi, A., 217 Marty, A., 38, 39 Matasović, R., 49 Mathiassen, T., 254 Maticsák, S., 56 Matisoff, J. A., 8, 67, 106n, 142, 144–6, 150, 359, 460n, 473 Matras, Y., 154 Mattes, V., 380 Matthews, S., 354, 356, 357 Mattiello, E., 34, 40 Mattissen, J., 278, 286n Matzinger, J., 48 Mayer, T., 63 Mayorava, T., 63 Mead, D., 367 Meillet, A., 103 Meir, I., 156, 167 Mélac, E., 387 Mel’čuk, I., 244n Melissaropoulou, D., 6, 83, 90n, 106n, 270, 273, 275, 276 Melzi, G., 27, 135 Meneghello, F., 8 Menz, A., 50, 57, 58 Merlan, F., 401, 406 Merlini Barbaresi, L., 5–8, 15, 23, 26, 32–4, 40, 41, 83–5, 89, 90n, 106n, 122, 124, 125, 127–32n, 134–6, 138–40, 144, 148, 151, 187, 227, 473 Messineo, C., 625, 629, 630, 632, 633n Mesthrie, R., 124, 126 Mettouchi, A., 94, 458–60n Midgette, S., 590 Mifsud, M., 53, 58, 62

Name index 721 Miftachova, I., 68 Mihajlović, I., 59 Miller, C., 180, 181 Miller, G., 67 Mirambel, A., 6 Mironova, N., 135 Mistrík, J., 296 Mithun, M., 55 Mjiyako, L., 191 Mogharrabi, M., 288 Mohanan, K.P., 51 Monaselidze, M., 222 Montermini, F., 6, 81, 90n Moore, P., 579 Morgan, W., Sr., 590 Morris, C. W., 35 Moško, G., 296 Mosley, M. E., 604 Mous, M., 508, 514n Msimang, C. T., 516–21 Mudzingwa, C., 80, 98, 134, 191, 492, 499 Munro, P., 534, 535 Murphy, N., 31 Mutaka, N., 175 Mutz, K., 5, 8, 16, 35, 100, 105n, 142–4, 147–50, 152, 154n, 188 Muysken, P., 182, 630 Myers, F. R., 405 Naden, T., 484 Náñez, E., 23 Napoli, D. J., 6, 211 Nash, D., 438, 442, 447n Nedjalkov, I. V., 339 Nercesian, V., 52, 86, 107n, 191, 634, 641 Newell, L. E., 379 Newman, P., 53, 62, 108 Newton, G., 47, 262 Ngakulmungan Kangka Leman, 451 Nichols, J., 70, 122, 354, 549n Nieuwenhuis, P., 6, 32, 33, 60n, 65, 72n, 350, 452 Nikolaeva, I., 44, 46, 55, 58, 335, 338, 340 Nishimoto, E., 355 Nizharadze, L., 224 Noccetti, S., 135–9 Noyer, R., 565 Nübling, D., 263 Nurse, D., 515 Nyst, V. A. S., 161, 164 Ohala, J. J., 70, 122, 123, 132n Okimāsis, J., 575 Olmsted, H., 135 Olofsson, A., 36, 192, 308 Ondrus, P., 296 Ondruš, P., 296, 304 Öner, M., 313 Ortiz de Urbina, J., 57, 197, 204 Ōta, T., 357 Otaina, G. A., 278–85n Otanes, F. T., 376–8 Ott, D., 6, 92

Palmović, M., 136 Pan, W., 355 Panfilov, V. Z., 278–83, 285n Panocová, R., 187 Parkvall, M., 172, 180 Parsons-Yazzie, E., 591 Payne, D., 150 Payne, T. E., 65, 66 Pazdjerski, D.-V., 44 Pecher, D., 21, 22 Peltier, F., 262 Pensalfini, R., 107n, 117, 188, 416–19, 421 Pentland, D. H., 6, 53, 569 Pereira, F., 528 Pérez Saldanya, M., 205 Perrott, D. V., 515 Perry, J. R., 287 Pete, S., 538, 539 Petersen, W., 8 Petitta, G., 16, 155, 169 Petrova, T. I., 340 Pfau, R., 164, 165 Pharies, D., 28 Philippaki-Warburton, I., 92 Philippson, G., 515 Picard, A., 456 Pike, E., 549n Pike, K., 542, 549n Pincus, D., 23 Pinto, I., 57 Pirott, L., 135 Pittman, C., 565 Pizzuto, E., 168 Plag, I., 121, 172 Plank, F., 43, 63, 75, 147 Pochkhua, B., 220 Podolsky, B., 244n Ponsonnet, M., 107n, 190, 192, 401, 407, 432 Potts, C., 496 Poulos, G., 493, 516–21 Powesland, P. F., 125 Prakasam, V., 331 Prieto, V., 5–7, 15, 25, 27–9, 35, 41, 67, 139, 154n, 187, 188 Protassova, E., 136–8, 140 Puglielli, A., 507, 508, 514n Pullum, G., 72n Quer, J., 164, 165 Quesada Pacheco, M. A., 528n Rainer, F., 6, 104, 110, 127, 128, 131, 143 Ralli, A., 269, 270, 273, 275, 276 Rankin, J., 135 Rappaport Hovav, M., 116 Rau, D. V., 389, 393, 395, 398, 399 Ravid, D., 134 Rdo.dgon Gsaṅ.bdag Rdo.rje, L., 387 Rees, G., 22 Reichard, G., 590 Reichmuth, W., 154n Reynolds, B., 6, 211 Reynoso Noverón, J., 40

Name index

722 Ricca, D., 146, 152 Rice, K., 87, 88, 536, 540, 579, 584, 585 Rice, S., 581, 586, 587 Rifón Sánchez, A., 108, 213 Rigau, G., 211 Roberts, J. R., 414 Robin, F., 385–8n Roelcke, T., 121, 127 Romaine, S., 122 Romeo, O., 158, 159 Romiti, S., 127 Rosch, E., 24 Rosen, N., 424 Ross, A., 387 Ross, B. B., 387, 401 Rossini, P., 155 Round, E. R., 449, 450, 452n Rubino, C., 49, 190, 379, 380 Rudaya, S., 68 Rudenka, A., 68 Rudolph, E., 6, 104 Ruf, B., 153n, 154n Ruffolo, R., 380 Ruiz de Mendoza Ibáñez, F. J., 39 Rūķe-Draviņa, V., 135, 254 Rull, X., 211, 212 Russell, K., 424 Ryle, G., 31 Sabatini, F., 95 Sabol, J., 187n, 296, 302, 303 Sadeghi, A. A., 288 Sadock, J. M., 560, 567n Saeed, J. I., 508–11, 513n, 514n Sagart, L., 353 Sailaja, P., 191 Sakel, J., 53 San José, M. Á., 376 San Miguel, R., 378 Sandler, W., 156, 165 Sapir, E., 26, 65, 70, 131, 132 Saulwick, A., 111, 431, 437n Saveljeva, V. N., 283 Savickienė, I., 8, 36, 135–41, 149, 188 Saxon, L., 584 Scalise, S., 3, 4, 6, 7, 15, 21, 43, 45, 72, 74–6, 79, 82, 84, 90n, 143, 187, 189, 276, 279, 430, 436 Scatton, E. A., 92 Schachter, P., 376–8 Schadeberg, T. C., 521 Scharfe de Stairs, E. F., 557 Schembri, A., 157, 161, 167 Schneck, J. M., 22 Schneider, C., 109, 192, 348, 351n Schneider, E. R., 334, 335, 340 Schneider, K. P., 6–8, 10, 37, 71, 77, 89n, 107n, 187 Schoen, P., 456 Schönig, C., 314 Schultze-Berndt, E., 109, 115 Schürmann, C. W., 423 Schwarzkopf, D., 22 Schwieger, P., 388n Seitz, J. A., 21, 26

Shanidze, A., 216 Shay, E., 48, 54 Shetler, J., 379 Shibatani, M., 130 Shiraishi, H., 285n Shonia, A., 224, 225 Siegel, J., 172, 173 Sifianou, M., 103 Sifre Gómez, M., 205 Siller-Runggaldier, H., 45, 54 Simon, C., 387 Šimon, F., 296 Simon, H. J., 4 Simons, F. G., 17, 321, 437n, 484, 604n Simpson, J., 423, 428n, 438, 447n Sinner, C., 214 Siyaad, C. M., 507, 508 Smiltniece, G., 254–6, 258, 259 Smith, C., 110 Smith, N., 172, 173 Smith, R., 27 Söderbergh, R., 307 Soida, E., 259–61 Sokolová, M., 296, 297, 304n Song, C., 22 Sonnenstuhl, I., 140 Sotiropoulos, D., 96, 97 Soukka, M., 146 Spalding, A., 565 Speas, M., 591 Sperber, D., 22 Spitzer, L., 35, 38, 129 Spreng, B., 567n Staalsen, P., 408, 414 Stairs Kreger, G. A., 557 Stanfield, R. A., 22 Staverman, W. H., 124 Ştefănescu, I., 6, 96 Štekauer, P., 6, 10, 15, 23, 43, 47, 51, 61, 70, 72n, 73n, 74, 120n, 142, 143, 146, 168, 187, 188, 190 Stephany, U., 135, 136, 138, 139 Stephens, E., 424 Stevens, S. S., 119 Štolc, J., 61, 296, 301, 302 Stolz, T., 153n Strodach, G. K., 8 Stump, G. T., 4, 6, 7, 21, 43, 75, 88, 253, 293, 301, 436, 634 Suárez, A., 557 Sudjič, M., 93 Sun, C., 360n Sunik, O. P., 340 Supalla, T., 158 Supriyanto, I., 69 Sutton, V., 169 Sutton-Spence, R., 161, 164, 169 Svelmoe, G., 379 Svelmoe, T., 379 Szymanek, B., 4, 72n, 74, 187 Tabataba’i, A., 288, 294n Taine-Cheikh, C., 6

Name index 723 Taksami, C. M., 283 Tamanji, P. N., 175 Taylor, J., 27 Teichelmann, C. G., 423, 427 ten Hacken, P., 43 Tenenbaum, J. M., 536, 537, 538, 541n Tenny, C., 112 Terestyéni, T., 129 Teuber, O., 125 Thill, G., 262 Thomadaki, E., 136, 138, 139 Thomas, F., 147 Thompson, G., 23 Thompson, S. A., 355, 358 Thorell, O., 307 Thornton, A. M., 40 Tiefenbach, H., 263, 265 Tindale, N. B., 423 Todolí Cervera, J., 205 Tolskaya, M., 58, 82, 85, 103, 107n, 191, 192, 335, 338, 340 Tomasello, M., 126 Tonelli, L., 135, 136 Topadze Gäumann, M., 111, 223 Topuria, V., 221, 223, 224 Tournadre, N., 381 Tovena, L. M., 6, 16, 77, 78, 105, 107n, 109, 117, 242, 511 Tozzer, A. M., 45 Traugott, E. C., 142, 148, 152, 153n Travis, C. E., 130 Tribushinina, E., 140 Trnková, J., 296, 297, 301 Trouvain, J., 263 Tsao, F., 357 Tsintsius, V. I., 333, 334, 338, 339 Tughushi, Z., 223 Tukumba, M., 401, 405, 407 Turolla, C., 141 Turon, L., 72, 205, 211, 215 Tuskia, M., 218 Ultan, R., 26, 60n, 69, 70, 72n Uray, G., 381, 382, 388n Urbutis, V., 59, 253 Vago, R. M., 233 Vaillant, A., 103 Vajda, E., 90n, 107n, 246, 249, 250, 252 Valdman, A., 171 Valentine, J. R., 571 Valera, S., 43, 61, 120n, 142, 188 Valiullina, Z. M., 311, 320n Valli, C., 161 van Breugel, S., 413 Van de Velde, M. L. O., 46 van den Berg, R., 49, 51, 82, 190, 367 Van der Spuy, A., 191, 515, 520 Van der Voort, H., 45, 86, 606 van Driem, G., 381 van Gijn, R., 61, 643, 648 Varon, E. J., 65 Vasilevich, G. M., 339

Vázquez Cuesta, P., 102 Veenstra, T., 182 Veerman-Leichsenring, A., 543 Vendler, Z., 110 Verhoeven, E., 527n Verkuyl, H. J., 112 Vermeerbergen, M., 155 Versteegh, K., 172 Vidal, A., 641, 642 Viegas Barros, J. P., 616 Visser, M., 503 Vittrant, A., 385, 386, 388n Voeikova, E., 136–8, 140 Vokurková, Z., 388n Volek, B., 35 Vollmann, R., 388n Volpati, C., 97 Vycichl, W., 460n Wagner-Nagy, B., 44, 56 Wang, C.-C., 393 Wang, L., 354, 356 Watson, J. C. E., 6 Weber, J.-J., 262 Weber, M., 6 Werner, H., 246, 247 Wessén, E., 307 Whaley, L. J., 333, 339 Wheelock, A., 579 Wiemer, B., 153 Wierzbicka, A., 6, 7, 12, 23, 35, 188, 606, 634 Wiese, H., 4 Wilbur, R. B., 165 Wilhelm, A., 584 Wilkes, A., 519 Wilkins, D., 406 Williams, W., 424 Wilson, D., 22 Wiltschko, M., 80, 98, 134, 191, 492, 499 Wittgenstein, L., 24, 35 Wolfart, H. C., 568, 576 Woll, B., 161, 164 Wolvengrey, A., 85, 90n, 107n, 111, 188, 190, 192, 568 Wrede, F., 147 Wu, Y., 356 Wu, Y.-H., 399 Wyatt, W., 424, 425 Yang, M.-C., 399 Yaxley R. H., 22 Ye, B., 355 Yip, V., 354, 356, 357 Young, R. W., 590, 591 Yue, A. O., 352 Zakar, V., 54 Zampieri, E., 8 Zamponi, R., 107n, 111, 616, 617 Zelinková, K., 296, 297, 300–2 Zhe, M., 147 Zhengzhang, S., 356 Zhu, X., 357, 358

Name index

724 Zimmermann, J., 6, 60n, 65, 136 Zinken, J., 21 Zocconi, E., 135 Zoll, C., 111 Zöller, H., 408 Zorc, D. R., 378

Zribi-Hertz, A., 172, 173 Zuazo, K., 195 Zubin D. A., 93 Zwaan, R. A., 21, 22 Zwicky, A., 72n, 269 Zwitserlood, I., 158, 162, 169n