An Ethnographic Grammar of the Eipo Language 3496026596

Citation preview

?( ZUR EINFUHRUNG

In dieser Schriftenreihe werden die Ergebnisse innerhalb des Schwerpunktprogramms (SPP) der Deutschen Forschungsgemeinschaft (DFG) »Interdisziplinare Erforschung von Mensch, Kultur und Umwelt im zentralen Hochland von West-Irian (Neuguinea)« veroffentlicht. Die Plane fur dieses umfassende Projekt wurden seit dem Friihjahr 1972 in der Abteilung Siidsee des Berliner Mu­ seums fur Volkerkunde, Staatliche Museen PreuBischer Kulturbesitz, erarbeitet. In der Folge konnte ein Forscherteam aus Vertretern der fachlich zustandigen biowissenschaftlichen, geisteswissenschaftlichen und geowissenschaftlichen Disziplinen gewonnen werden. Auf entsprechenden Antrag richtete die DFG am 25. Oktober 1973 das SPP ein. Das Untersuchungsgebiet dieses SPP liegt im schwer zuganglichen zentralen Bergland von Neuguinea, im Osten der indonesischen Provinz Irian Jaya. Diese Region (zwischen den Siedlungsgebieten der Yali im Westen und der Ok im Osten) war bis in die Gegenwart noch nicht erforscht, die Bevolkerung weitgehend isoliert gegeniiber der weltumfassenden westlichen Ziviiisation. Die ErschlieBung dieses Gebietes durch Missionare und indonesische Administration, der unvermeidbare Kontakt zur modernen AuBenwelt, war indessen vorauszusehen. Die Zielsetzung dieses Schwerpunktprogramms ist die moglichst ausfuhrliche und systematische Dokumentadon iiber die Bevolkerung, ihre Kultur und ihren natlirlichen Lebensraum. Das Projekt bedingt von der DFG mit Fordermitteln unterstlitzte Einzeluntersuchungen durch eine Anzahl Projekdeiter und deren Mitarbeiter in partnerschaftlicher Kooperation. Folgende Spezialdisziplinen sind in das Programm integriert: Physische Anthropologie, Zahnmedizin, Allgemeinmedizin, Tropenmedizin, Ethnopsychologie, Humanethologie, Anthropogeographie, Linguistik, Ethnologie, Wirtschaftsethnologie, Ethnosoziologie, Ethnomedizin, Musikethnologie, Archaologie, Geologie, Photogeologie, Palaontologie, Geomorphologie, Seismologie, Flydrologie, Bodenkunde, Klimatologie, Botanik, Kulturpflanzenkunde, Zoologie und Haustierkunde. Zusatzlich wurde eine wissenschaftliche Filmdokumentation eingeplant. Ein KoordinierungsausschuB mit Vertretern der Bio-, Geistes- und Geowissenschaften stimmt die einzelnen Spezialuntersuchungen aufeinander und auf das zentrale Thema ab. In enger interdisziplinarer Zusammenarbeit der verschiedenen Spezialisten soil die intensive Erforschung des Untersuchungsgebietes und der dort lebenden Bevolkerung in wechselseitiger Erhellung der mannigfachen Aspekte ermoglicht werden. Die Untersuchungen begannen 1974, nachdem die erste Expeditionsgruppe am 22. Mai wahrend einer Lufterkundung mittels der Missionary Aviation Fellowship (MAF) das von den Piloten »X-Valley« genannte Hochtal als besonders geeignet fur die Feldforschung erkannt hatte. In diesem Tal, dem Eipomek-Tal (4°25' —4°27's. Br. und 140°00'—140°05' ostl. L.), ist seitdem eine groBere Zahl von Einzelprojekten realisiert worden. 1975/76 wurden das Famek-Tal (westliches Nebental) und das Bime-Tal

.. : : v

(im Osten) in die Forschungen einbezogen. Erkundungen nach Siiden (Larye) und Norden (Kosarek, IdenburgFluB), Osten (Tanime, Ok Bap) und Westen (Nalca) dehnen die Forschungsplane auf das weitere Untersuchungs­ gebiet aus. Dieses ist infolge der bisherigen Ergebnisse als der Mek-Siedlungsraum (Mek-Sprachfamilie) zu verstehen. Das intensivst erforschte Gebiet ist das slidliche Eipo­ mek-Tal. Ein von den Einheimischen »Eipomek« benannter FluB, der im Zentralmassiv, im Jayawijaya-Gebirge (ehemals Oranje-Gebirge), entspringt, durchzieht das Tal von Siiden nach Norden. Das Tal ist von Bergziigen begrenzt, die zum Teil liber 3000 m hoch sind und mit sehr steilen Hangen nordwarts verlaufen. Mehrere Hohenriicken ragen auch wie Barrieren weiter im Norden in das zerkliiftete Hochtal hinein. Viel Regen, Nebel und auch intensive Sonnenstrahlung bestimmen das wechselvolle Wetter. Die Einheimischen, die Eipo, sind eine typische, alteingesessene Neuguinea-Berglandbevolkerung (»Bergpapua«), Kleinwiichsige (Pygmaen) mit weniger als 150 cm KorpergroBe, hellbrauner bis dunkelbrauner Haut und schwarzlichem Kraushaar. Diese Bevolkerung des siidlichen Eipomek-Tales ist sprachlich und kulturell einheitlich. Indessen siedeln die Eipo in Dorfschaften (etwa 10-30 Rundhiitten), die sozial, politisch und wirtschaftlich selbstandig sind. Ihre »Haufendorfer« stehen auf Kuppen, Bergnasen und Felsvorspriingen in 1600-2000 m Hohe. Die Siedlungen bestehen aus Familienhlitten (fur Frauen, Madchen und kleine Jungen), mindestens einem Mannerhaus (auch fur die initiierten Knaben) und einer Frauenhiitte (vor allem als Aufenthaltsort zur Zeit der Menstruation). Die Bevolkerung der Dorfer (jeweils etwa 35—140 Eipo) ist in patrilinearer Deszendenz bestimmten, weit liber die Region verstreuten exogamen Sippen verbunden. Die Mitglieder der jeweiligen Mannerhaus-Gruppierung bilden eine in gemeinsamem Interesse handelnde und kultisch integrierte Gemeinschaft, sind aber innerhalb dieser Bindungen relativ selbstandig. Jeweils einzelne Manner mittleren Alters flihren kraft ihrer Personlichkeit in gewissen Grenzen. Kultzentren sind die (sakralen) Mannerhauser. Die Glaubensvorstellungen grlinden sich auf bedeutende Ahnengeister und auch auf Totengeister und Naturgeister. Unterhalb der Region des primaren Regenwaldes liegen die Garten der Eipo auf den Hangen und auf dem Talgrund, zwischen ausgedehnten Brachlandflachen mit Sekundarvegetation. Als »neolithische Pflanzer« ziehen die Eipo in ihren Garten vor allem SliBkartoffeln (Ipomoea batatas) und Taro (Colocasia esculenta). Sie terrassieren daflir auch steile Hange. Die zum Teil sehr ausgedehnten Terrassenanlagen werden allein mit dem schlichten Grab stock und mit den bloBen Handen erarbeitet. —Schwein und Hund sind die Haustiere. Die materielle Ausrlistung der ihrer rauhen Umwelt gut angepaBten Eipo ist karg. Doch selbst in ihrem schlichten Bedarf sind sie nicht vollkommen autark, sondern auf einige Materialimporte aus Nachbarregionen angewiesen. Die Planung zur Erforschung des gesamten Untersu­ chungsgebietes ist in einem »Generalexpose« detailliert beschrieben. Im Museum fur Volkerkunde in Berlin wur­ de ein Dokumentations-Archiv begrlindet, um die Feld1

forschung karteimaBig zu erfassen, mit Duplikaten aller Foto-, Film- und Tonaufnahmen. Alle fiir das Thema des SPP relevanten Dokumentationen und Untersuchungsergebnisse sollen in dieser Schriftenreihe publiziert werden. Die Autoren sind jeweils die Projektleiter der Einzeluntersuchungen sowie Autorengruppierungen fur verschiedene interdisziplinare Themen. Es ist vor allem eine Publikation der Fakten, eine Darlegung der Befunde, so daB die eingeschlossenen Folgerungen nachvollzogen werden konnten. Unter Beriicksichtigung der Situation, daB dieses die vielleicht letzte Chance war, einen verbliebenen »weiBen Fleck« auf unserer Erde systematisch weitestgehend zu erforschen, sollte vor allem das Befundmaterial publiziert werden, denn im Wandel

der Theorien sind die Fakten konstant (soweit naturgemaB die jeweilige Erkenntnis ausreichend war). Diese monographieartigen (oder auch nur artikelartigen) Publikationen sind als Einzellieferungen erhaltlich und erscheinen in unregelmaBiger Folge in deutscher Sprache mit Resiimees in Englisch (bzw. Indonesisch). Zum AbschluB des gesamten Unternehmens, d. h. nach etiichen Jahren, werden die Einzelpublikationen in einer Serie von Banden zusammengefaBt, gegliedert in die Sektionen I. Umwelt, II. Mensch und III. Kultur. Ein abschlieBender Band soil eine zusammenfassende Gesamtiibersicht vermitteln und einen von alien Projektteilnehmern erarbeiteten Generalindex enthalten. Gerd Koch

AN ETHNOGRAPHIC GRAMMAR OF THE EIPO LANGUAGE spoken in the central mountains o f Irian Jaya (West New Guinea), Indonesia by Volker Heeschen (Starnberg)

23. Beitrag zur Schriftenreihe

MENSCH, KULTUR UND UMWELT IM ZENTRALEN BERGLAND VON WEST-NEUGUINEA

DIETRICH REIMER VERLAG BERLIN

Die Deutsche Bibliothek —CIP-Kurztitelaufnahme Heeschen, Yolker: An ethnographic grammar of the Eipo language spoken in the central mountains of Irian Jaya (West New Guinea), Indonesia / by Volker Heeschen. —Berlin : Reimer, 1998 (Schriftenreihe Mensch, Kultur und Umwelt im zentralen Bergland von West-Neuguinea ; Beitr. 23) ISBN 3-496-02659-6

Gedruckt mit Unterstiitzung der Deutschen Forschungsgemeinschaft

Max-Planck-lnstitut fur evolutionare Anthropologie Bibliothek

ISBN 3-496-02659-6 © 1998 by DIETRICH REIMER VERLAG BERLIN Dr. Friedrich Kaufmann, Unter den Eichen 57, 12203 Berlin Alle Rechte vorbehalten, Vervielfaltigungen jeder Art sind nicht gestattet. Printed in Germany Layout: Olaf Holy Herstellung: Reiter Druck, Berlin

CONTENTS

1. 1.1 1.2 1.3

PRELIMINARIES................................................ 9 Preface: dictionary, texts, and grammar............ 9 Acknowledgements.............................................10 Abbreviations and symbols ...............................11

PART ONE Prolegomena to a grammar of the Eipo language . 13 2. THE EIPO AND THEIR LANGUAGE . . 15 2.1 The history of research and the state of the a r t ......................................................15 2.2 Dialects and languages ...................................... 16 2.3 The Eipo and the Eipo language.....................21 2.4 The vocabulary ................................................... 25 2.5 The ethnography of speaking ......................... 30 3. 3.1 3.2 3.3 3.3.1 3.3.2 3.3.3 3.3.4 4. 4.1 4.1.1 4.1.2 4.1.3 4.2 5. 5.1 5.2

THEORY: PHILOLOGY AND “INJUNCTIVE GRAMMAR” ....................... 36 Field work: data and th eo ries........................... 36 Philology .............................................................. 41 “Injunctive grammar” ........................................ 46 Observations ........................................................46 Further examples.................................................49 Deixis, ellipsis, expansibility..............................54 Basic units. Problems ........................................ 63 THE POSITION OF EIPO WITHIN THE MEIC LANGUAGES AND THE OTHER PAPUAN LANGUAGES ................................72 Remarks on the history of the Mek languages ................... 72 Remarks on phonological developments of the Mek languages ........................................ 72 Remarks on the morphology of the Mek languages ..................................................... 75 The Mek languages and the other Papuan languages................................................. 79 Some remarks on typology................................. 85 TALKING ABOUT LANGUAGE...............95 Language awareness: the gift of talking about language..................................... . 95 Talks about language with Eipo and Yalenang.............................................................. 102

PART TWO A grammar of the Eipo language ............................115 6. GUIDE TO PRONUNCIATION AND SKETCH OF PHONOLOGY..................... 117 6.1 Segmental phonemes .......................................117 6.1.1 Phoneme inventory...........................................117 6.1.2 Description of phonemes, allophones, and distribution................................................. 118 6.1.2.1 Consonants ..........................................................118 6.1.2.2 Vowels ..................................................................122 6.1.3 Consonant sequences ...................................... 125 6.1.4 Contrastive sets ......................................... 126 6.1.5 Padding: sounds outside the system ............ 128 6.2 Syllables and suprasegmental features.........128

6.2.1 6.2.2 6.3 6.3.1 6.3.2 6.4

Syllable types and the phonological word . . 128 Suprasegmentals ................................................ 131 Resum e................................................................ 136 Some phonological processes......................... 136 Characteristics of Eipo phonology...............138 Transcribed te x t................................................. 139

7. 7.1 7.1.1 7.1.2 7.1.3 7.1.4 7.1.4.1 7.1.4.2 7.1.4.3

WORDCLASSES. MORPHOSYNTAX .. 141 Words requiring clarification ..........................142 Spatial deictics ....................................................143 Pronouns ............................................................ 148 Interrogatives......................................................153 Connectives and postpositions .....................156 Copulative coordinators................................... 156 Correlatives ........................................................160 Connectives co-ordinating theme and rheme, adversatives ...........................................161 Agentive, means, source. Postpositions and connectives................................................. 168 Spatial relations. Postpositions and connectives..........................................................172 Connectives denoting temporal relations . . 179 Comparative constructions...............................191 Reason, purpose, recipient. Postpositions and connectives................................................. 192 Relativizers........................................................... 195 Basic units: words .............................................197 N ouns...................................................................197 Remarks on the semantic classification of nouns.............................................................. 197 Remarks on the expression of gender and number ........................................................200 Noun formation ................................................203 Derivation ...........................................................203 Compounds.........................................................207 Adjectives............................................................210 Adjective formation ......................................... 212 Numeral adjectives ........................................... 213 Adverbs .............................................................. 215 Ad-verbs...............................................................215 Negation, affirmation, initiators......................221 Verbs.....................................................................223 The stem ..............................................................224 Verb formation ....................................... 229 Adjunct formations........................................... 229 Compound verbs. Lexicalization....................231 Compound verbs. Grammaticalization . . . . 238 Tense-aspect and mood suffixes ...................246 Object pronouns................................................255 Suffixes for tense, mood, and person-number ................................................. 257 Infinite verbal fo rm s......................................... 259

7.1.4.4 7.1.4.5 7.1.4.6 7.1.4.7 7.1.4.8 7.1.4.9 7.2 7.2.1 7.2.1.1. 7.2.1.2 7.2.1.3 7.2.1.3.1 7.2.1.3.2 7.2.2 7.2.2.1 1.2.22 7.2.3 7.2.3.1 1.23.2 7.2.4 7.2.4.1 7.2.4.2 7.2.4.2.1 7.2.4.2.2 7.2.4.2.3 7.2.4.3 7.2.4.4 7.2.4.5 7.2.4.6 8. 8.1 8.2

SYN TAX............................................................265 Retrospective glance at first observations and morphology .......................... 265 Nouns, adjectives, and adverbs in expanded structures ........................................ 265 5

1

8.2.1 8.2.1.1 8.2.1.2 8.2.1.3 8.2.1.4 8.2.1.5 8.2.2 8.2.2.1 8.2.2.2 8.3 8.3.1 8.3.2 8.3.2.1 8.3.2.2 8.3.2.3 8.3.2.4 8.3.2.5 8.4 8.4.1 8.4.2 8.4.3 8.4.3.1 8.4.3.2 8.5 8.5.1 8.5.2 8.5.2.1 8.5.2.2 8.5.2.3 8.5.2.4 8.5.3

Nouns as heads: general remarks and non-standard attributes.................................... 265 Possessive noun phrases...................................268 Nouns and spatial deictics .............................. 269 Nouns and possessive pronouns....................271 Nouns and adjectives ....................................... 274 Nouns and relative clau ses.............................. 276 The limits of expansion...................................277 Noun serialization..............................................277 Yerbless clauses and predicate structures . . 280 Clauses with finite v e rb s...................................284 Wordorder ..........................................................286 Clause structure................................................. 288 Major constituents and case-marking ........... 288 Thematization .................................................... 298 Forms of distribution....................................... 303 Forms of summarizing..................................... 308 Transitivity and valency increas e ....................311 Clause-combining.............................................316 Clause-chaining................................................. 319 From co-ordination to subordination: adverbial clauses ............................................... 326 Subordination ................................................... 334 Object (and subject) clauses............................. 334 Relative clauses and nominalization................336 Remarks on discourse...................................... 343 Isolated sentences and discourse...................343 Resume of previous findings related to discourse ........................................................345 Characteristics of some discourse types . . . 345 Subject-tracking and switch reference............348 Object-tracking ...................................................349 Linkage.................................... 351 L’envoi: ideal structure and real “texts” . . . . 353

PART THREE Texts from various Mek languages, Summary, Bibliography, In d ex.................. ............................ 357 9. TE X TS................................................................359 9.1 Yaber (1976, Limirye, Bime valley, Bime language): Calab kwemdina - Narrative of the creation of C alab ................................ 359

6

9.2

9.3

9.4

9.5 9.6

9.7

9.8

9.9

Kabrob Tengket (1992, Londini, formerly Munggona, Eipomek, Tani dialect): Tengketnang kwemdina —Narrative of the creation of the Tengket clan ................. 361 Lublub Salawala (1992, Dubokon, Tanime valley, Tani dialect): Salawalanang kwemdina —Narrative of the creation of the Salawala clan .........................................364 Lekwoleb (1976, Dingerkon, Eipomek, Larye dialect): Youke dolamik yupe — Narrative of how they built the men’s house .................................................................. 370 Yakob (1981, Dingerkon, Eipomek, Eipo): Yulye yupe —Narrative of the Y u lye............ 372 Yakob (1992, Londini, formerly Dingerkon, Eipo): Bolmedin-nang kwemdina —Narrative of the creation of the Bolmedin clan ...................................... 373 Turunang Kalimna (1992, Lumdakla, Famek, Fa dialect): Kalimna-nang kwemdina —Narrative of the creation of the Kalimna clan ........................................ 374 Turunang Kalimna (1992, Lumdakla, Famek, Fa dialect): Medial-nang kwemdina - Narrative of the creation of the Medial c la n .............................................377 Namin Senadin (1983, Nohomas, In valley, Yale language): Senadin-nang mangkahioka —Narrative of the creation of the Senadin c la n ...........................................378

10.

SUMMARY........................................................381

11.

BIBLIOGRAPHY ...........................................383

12.

INDEX ..............................................................403

MAPS AND TABLES

Map 1 Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Table 8 Table 9 Table 10 Table 11 Table 12 Table 13 Table 14 Table 15 Table 16 Table 17 Table 18 Table 19 Table 20 Table 21 Table 22 Table 23 Table 24 Table 25 Table 26 Table 27 Table 28

The Mek language a r e a ......................................17 Dialects and languages in the Mek are a........19 Percentages of shared vocabulary Yale, Okbab, K orapun................................................. 19 Percentages of shared vocabulary in the centre of the Mek languagea area ...................20 System of plosives in the eastern Mek languages ..................................................... 72 System of plosives in Eipo and Yale................72 Eipo pitch and Yale low to n e .......................... 74 Eipo pitch and Yale high tone ........................ 75 Free pronouns in several Mek languages . . . . 75 Spatial deixis in several Mek languages........75 Interrogative pronouns in several Mek languages..................................................... 76 Stem-final consonants of Mek verbs..............76 Tense-aspect suffixes in several Mek languages ..................................................... 76 Object pronouns.................................................. 77 Person-number suffixes in several Mek languages ..................................................... 77 Connectives in several Mek languages........... 78 Telefol and Eipo kinship terms ......................81 Angguruk and Mek kinship term s..................81 Nalum - Eipo correspondences......................81 Telefol - Eipo correspondences......................81 Angguruk - Eipo - Yale correspondences . . 82 Stem correspondences Angguruk - Eipo - Y a le.................................... 82 Verb formulas Ok, Dani, M e k ........................83 First person plural suffixes.............................. 84 Phoneme inventory. Consonants and vow els...................................................................117 Distinctive features ......................................... 118 Eipo -e and -ye in other Mek languages . . . . 122 Diphthongs ...................................................... 124 Consonant sequences ..................................... 125

Table 29 Suprasegmentals of 'bukne “I have mixed” ..................................................132 Table 30 Suprasegmentals of buJdne “I have sat down” ............................................. 132 Table 31 Suprasegmentals of 1bukmale “he is mixing” ......................................................133 Table 32 Suprasegmentals of butdmal “he is sitting” ......................................................133 Table 33 Suprasegmentals in accented and non-accented syllables...................................... 133 Table 34 Suprasegmentals of meta “fat” .......................133 Table 35 Suprasegmentals of 1koubmase “I was planting” ................................................. 134 Table 36 Suprasegmentals of koubmase “I was crossing (a river)” .................................. 134 Table 37 Suprasegmentals of 'dibne “I have eaten” . . . 135 Table 38 Suprasegmentals of dilamle “he eats”............135 Table 39 Noun derivation by - a k ................................... 203 Table 40 Noun derivation by -dam ................................. 203 Table 41 Noun derivation by -ta ....................... 204 Table 42 Noun derivation by -we (labialized consonant)...................................... 204 Table 43 Verb stem changes............................................228 Table 44 Expanded and non-expanded verb stems .. 228 Table 45 Tense-aspect and tense-mood-person-number suffixes............ 247 Table 46 Meanings of tense-aspect suffixes................252 Table 47 Tense-mood-person-number suffixes........257 Table 48 Transitivity nuclei and forms of summarizing................................ 315 Table 49 Features of non-final clau ses........................ 318 Table 50 Clause-chaining types and possibilities . . . . 322 Table 51 Consecutio temporum and consecutio aspectuum............................................................329 Table 52 Consecutio temporum in conditional clauses.................................................................. 331 Table 53 Means and functions of relative clauses . . . 336

7

1.

PRELIMINARIES

/. / Preface: dictionary, texts, and grammar Alles laBt sich besser schreiben Darum laB die schlechtere Fassung stehn. [....] Es kommt nie auf die Satze an. Nur das Werk allein zahlt. Friedrich Diirrenmatt

When writing a grammar the fieldworker apparently leaves the culture of the people whose language he is studying out of account. While for the compiling of a dictionary and the collection of texts it is indispensable to enter into the culturally specific conceptions and ways of thinking, the mere grammatical structure of a language can be con­ ceived independently from the culture of the speakers, unless the linguist refers to theories connecting both, grammar and culture, so that either certain ways of think­ ing or “Weltbilder” can be correlated with certain gram­ matical structures, or the former can even form or mould the latter. Even if this correlation or the prescriptive force of grammar is not regarded as a topic within the range of a modern linguist’s preoccupations and is accordingly left aside, as will be done by the author of this study, one may always find enough reasons to reflect upon the correlation of anthropological work and linguistic descriptions and explanations. “Linguistics without anthropology is sterile, anthropology without linguistics is blind” (Hockett 1973:675): there are systematic ways of connecting anthropology and linguis­ tics. They lead from Lazarus and Steinthal via Wundt and Cassirer, Boas and Sapir to what cognitive anthropology and nascent ethnosemiotics have to offer today: an under­ standing of, and an insight into, foreign thinking. How­ ever, this is not a method leading to a grammar. The con­ necting path is rather a detour, as far as a grammar of the Eipo language is concerned. It was difficult to abstract from the work on the dictionary, from translating, from the work with the informants, the situation of the speak­ ers and the context of speaking. To this extent, these stud­ ies are also an attempt at making the difficulty of abstract­ ing visible in the grammar. Walking on the by-roads and making detours means for the grammarian, firsdy, that he remains conscious of the many connecting links leading from the dictionary to the grammar, secondly, that each sentence and each structure do not exist by themselves, but have to be traced first in transient speech acts, thirdly, that dictionary-making and translating teach him that each discovery in the field of grammar is irrevocably linked to the interpretation and translation of foreign speech, and, finally, that his own process of learning and the work with the informants are a sort of game, in which the learner tries to understand something that is very difficult to understand from his own point of view, be it that of a certain linguistic theory or simply that of his own mother tongue, and in which the teachers learn to think about their own language. In a word: the grammar has to show how the discoveries were made. It is philological, close to utterings and speech and their transience and imponderabilities, but also close

to the difficulties which the learners as well as the teachers had with their language. It works with terms free, as far as possible, from hypotheses and theories, but at the same time it forms a theory out of its methods, which means nothing other than that the construction of knowledge and the “making” of grammatical structures in speaking become the object of its investigation. The chapters preceding the Eipo grammar proper espe­ cially serve this aim. The introduction is concerned with the anthropological context and with the contexts and situ­ ations of speaking. Some remarks about the quantity and the variability of the vocabulary of a small ethnic commu­ nity also belong here. Beside this, it gives a survey of the research that has been done among the Eipo themselves and their neighbours up to now. The chapter on theory offers positive arguments for all that was said here about the difficulties of deliverance from these contexts; speak­ ing and the structures realized while speaking are con­ ceived as a process of continuous clarification and as a bundle of strategies necessary for the construction of longer sequences of speech. The chapter about the his­ tory together with excursions into comparative linguistics and into the typology tries to treat the characteristics of the Eipo language in detail; we will have to ask the ques­ tion whether these characteristics are a historical inheri­ tance of the Papuan languages, whether they can be explained from the point of view of typology or whether they are of a more essential nature, i.e. whether they can be regarded as solutions for that process of clarification or for the above-mentioned strategies. The chapter about language awareness, less pretentiously called “speaking about language”, tries to record which meanings and structures the Eipo speakers were aware of before an inquisitive linguist called their attention to them, and it tries to evoke the factors which, during the “speaking about language”, made the speakers aware of other mean­ ings and characteristics of their language. This chapter also has consequences for theoretical assumptions and views regarding language history. The grammar itself offers a relatively conservative analysis of the phonology and of the word and morpheme classes; in the syntactic part I try to be serious about regarding grammar as a pro­ cess of increasing clarification and as a strategy for the construction of longer units of speech. The limited size of Eipo noun phrases and clauses as well as the fact that case-marked and thematized noun phrase constituents of the clause become capable of being isolated and movable leads back to the idea of building-blocks or basic units which, on the one hand, form independent utterances and, on the other hand, are pieced together in discourse. The grammar makes use of the results of the interviews about language; it remains close to the data by revealing the way leading to the results. The texts illustrate the unity and diversity of the Mek languages as well as the grammar working in connected discourse. Though the three parts and the chapters of this ethnographic grammar relate to each other, the individual parts as well as single chapters can be read separately. The plan of this ethnographic grammar brings about some repetitions. —The long list of references mirrors the wide and diverging interests; as a pure list of references to a grammar its length would have to be considered as unseemly. Part One was finished in 9

i

June 1991, Part Two and Three in July 1995. I have not tried to discuss relevant anthropological and linguistic works published after 1991 or 1995. Some ideas devel­ oped in this grammar and examples illustrating these ideas have been published in shorter articles (concerning chap­ ter 2.5 cf. Heeschen, forthcoming a, concerning chapter 3.3 cf. Heeschen 1994c, concerning chapter 4 cf. Hee­ schen 1992c, concerning chapter 5. cf. Heeschen, forth­ coming b). Anyone who writes more or less explicitly follows models and examples. As a result of this another difficulty for the writing of this study arose. With regard to theory the models for the linguist were Wundt (1911—1912), Biihler (1934), Hormann (1978) and Levelt (1989), all of them authors of ethnopsychological or psycholinguistic works or studies concerned with language history or compari­ son: however, they do not offer a model for the compre­ hensive grammar of a single language. This study would have to be split up into a series of texts, experiments and typological works. From this I refrained, because I wanted the study to fit into the imposing series of Papuan gram­ mars already existing. In a grammar of the Yale language, which I hope to present subsequently, I shall, however, introduce data from ontogenesis and from extensive tests and interviews, so that the origin of grammatical struc­ tures and not so much the laws of a “langue” will be brought into focus. In addition to the studies mentioned above there were examples concerned with other types of languages than the Papuan ones. Among others there is the introduction to the Greek language by Poeschel (1975) which is readable as well as pleasant to read; it is also an introduction to the history of culture and literature, and a lot of grammatical examples illustrate the context in which the utterances were made. Last not least, there is the grammar of Yidiq by Dixon (1977b); he follows the model of a generative transformational grammar, though not very closely, but he investigates each detail promising information of historical and ethnological importance. Both, Poeschel and Dixon, never lose sight of the reader. These models have encouraged me to characterize the samples also with regard to style, contextual dependence and their cultural significance. The grammatical samples are regarded as speech acts from a living culture. They reflect the fact that the informants were not just a means of eliciting grammatical structures. Informants laugh, tell, narrate, report, and chat. In other words: they refer to cul­ turally specific and meaningful events. Of course, all in­ stances taken from the published texts (Heeschen 1990b) were discussed with the narrators, native grammarians, informants, or friends. The so-called free translations are still close to the Eipo original: sometimes this results in rather unconventional usages of English words and gram­ mar. - Most of the grammars of Papuan languages set to work using the method of description supplied by the tagmemic model. The more extensive ones are exemplary for the very reason that their methods of description are often eclectic, that they have to consider the requirements of the complex morphological part of the languages or that they discuss whether this or that model adequately describes the specific data or surface structures (cf. Brom­ ley 1981, Bruce 1984, Foley 1991, Haiman 1980, W Seiler 1985).

The student who wants to take his first steps in the Eipo language should carefully read the guide to pronunciation (especially chapters 6.1.1 — 6.1.2.2, 6.2.2, and 6.4); he should cursorily pick up the basic forms and functions of the “Words requiring clarification” (chapters 7.1.2 7.1.4.9); he must learn the formula of the verb (chapter 7.2.4), the basic meaning of the tense-aspect suffixes (chapter 7.2.4.3), form and function of the object pro­ nouns (chapter 7.2.4.4), the suffixes for tense, mood, and person-number (chapter 7.2.4.5), and the forms of the infinite verbs (chapter 7.2.4.6); he should know the prin­ ciples of verb compounding (chapters 7.2.4.2.2-7.2.4.2.3); and he should read the chapters on clause structure (8.3.1 —8.3.2.5). The student who wants to make himself more familiar with the Eipo language in addition to the in­ formation contained in these chapters should begin with simple and clear texts (Heeschen 1990b): texts no. 2 “Sie kamen in der Erde”, no. 8 “Geschichte vom Yulye”, no. 19 “Vom Waldomanye”, no. 46 “Wie sie das Eglu Mannerhaus bauten”, no. 62 “Die Nimdamkone-Frau”, no. 105 “Wie sie Engebner verheirateten” and no. 131-132 “Marchen”.1 If the learner has read these texts, he may cast a more or less quick glance at the other chap­ ters in this work. —The texts from Bime (in the Bime lan­ guage), from the Tani and Fa valley dialects, from the Larye dialect of the Una language, and from Kosarek or Inlom (in the Yale language) rather serve the purposes of language comparison. The material on which these studies are based is, first of all, the published texts or the tape recordings preceding these (Heeschen 1990b) and the myths I collected in 1992 (cf. chapters 9.2 —9.3 and 9.6. - 9.8). In addition, there are texts which I did not publish; the total extent of the mate­ rial may cover about 800 pages in A-4 format. Over and above this, I used: explanations regarding words and texts, talks about grammar (about ten hours’ tape recordings, cf. Heeschen 1984a and chapter 5), spontaneous conversa­ tions of the inhabitants of the hamlet of Dingerkon (about seven hours’ tape recordings) and around 15 note­ books containing words, idioms and sentences from the first months after contact (during these first months taperecorders were not used in agreement with the anthropol­ ogists, because the Eipos’ reaction to them was unknown. As nothing had been published about the language and the first verbal contacts and collections of data were nec­ essarily unilingual, the linguist had no other choice but to listen and take notes). The tape-recordings regarding the published texts and the talks about grammar will be accessible in the Tonarchiv of the Forschungsstelle fur Humanethologie in der Max-Planck-Gesellschaft.

I

j

|

1.2 Acknowledgements The narrators, their characteristics and special gifts have been portrayed (Heeschen 1990b). Some of the narrators

1 No. 2 “They came in the earth”, no. 8 “Story o f the Yulye”, no. 19 “Story o f the Waldomanye”, no. 46 “How they built the Eglu men’s house”, no. 62 “The Nimdamkone woman”, no. 105 “How they married Engebner”, and no. 13 1-13 2 “Fairy tales”

10 i

also helped with the transcriptions, the explanations of texts, and the talks about grammar. These were mainly (the Christian names used today are added in round brackets): Melase, Buk (Yakob), Bingde (Laik), Walebyan (Filipus) and Kwengkweng (Enus). With the Yalenang these were: Silas Dibul, Levi Dibul, Musa Dibul, and Imael Amoldoman. I express my special gratitude to all of them, but also to the inhabitants of Dingerkon, who by allowing me to enter into their lives and by including me in their social network, encouraged me to an uncontrolled and quick learning of their language. By taking notes of all the utterances accompanying exchange, of all the exhortations, apprehensions and admonitions to behave as the Eipo expected me to, I obtained my first data and results. I stayed in the field during the following periods: 1. with the Eipo (and their eastern neighbours, the Bime people): June - July 1974, October 1974 — April 1975, March - August 1976, June 1979, Octo­ ber —December 1979, and for periods of up to two months in 1981,1986, 1989, 1992,1995, and 1997. 2. with the Yalenang (the Eipos5 neighbours in the west of the Mek language area): December 1978 January 1979, April 1979 —November 1981, Sep­ tember —December 1983, and for periods of up to two months in 1986, 1989, 1992, and 1994. At the mission station of Kosarek in the Yale area, I also worked with guests and informants from Eipomek. Research 1974-1976 was sponsored by LIPI and from 1978 up to now by the Protestant Church of Irian Jaya. Since 1994 I revisited the Eipo and the Yalenang as a tourist, and no research was realized.

Institute of Linguistics, Irian Jaya Branch, for several flights during the time from 1981—1994; the Vereinigte Evangelische Mission, Wuppertal, who employed me from 1978—1981; and the Protestant Church of Irian Jaya (GKI), who invited me as staff member of the Vereinigte Evangelische Mission and without whose tolerance and hospitality research in Irian Jaya would have been impos­ sible after 1976. I am also indebted to Lieselotte Schiefer and Susanne Bur­ ger from the Institut fur Phonetik und Kommunikationswissenschaft of Munich University and to Claus Heeschen, Max-Planck-Institut fur Psycholinguistik, Nijmegen, for revising my phonetic transcriptions. The German version of Part One (Heeschen 1991) was submitted to the Technical University of Berlin as “HabiHtationsschrift” (postdoctoral thesis). G. Hincha, B. Nothofer, R. Posner, W A. L. Stokhof, G. Senft and S. A. Wurm examined and reviewed this version. I thank my wife Helga Heeschen, who translated the thesis into English. I am indebted to G. Senft who examined and reviewed (the English version of) Part One and Part Two and to A. Pawley, G. Reesink, C. L. Voorhoeve and Tsuyoshi Takizawa for critical remarks and helpful comments on parts of the grammar (Part Two) or on prepublished parts of this work related to Part One and Part Two (cf. above chapter 1.1). I thank Ruth Schubert and Kerstin Kazzazi, both of them trained linguists and native speakers of English, for correcting my English. —It goes without say­ ing that the usual disclaimers apply. 1.3 A.bbreviations and symbols —

These stays were made possible by the assistance of many persons and institutions. I express my gratitude to the fol­ lowing persons: Klaus Helfrich and Gerd Koch, the initia­ tors of the project “Mensch, Kultur und Umwelt im zentralen Bergland von West-Neuguinea”, Ursula FarHollender of the Deutsche Forschungsgemeinschaft for their assistance and advice, the protestant ministers Klaus Reuter and Herbert Schekatz, who invited me to co-oper­ ate with the Vereinigte Evangelische Mission and who unrestrictedly supported my scientific interests, the many members of the staff and friends from the Protestant Church of Irian Jaya, among them Pin Sawen, Tomas Yoteni, Bernd Dimara, August Kafiar, and Willem Pusop, and finally Richard Steinbring and his family from the Summer Institute of Linguistics, who helped me with the organization of my visits to Kosarek in 1989 and 1992. Further I would like to thank the following institutions: the Deutsche Forschungsgemeinschaft, who granted the funds for my first two stays with the Eipo, for my visit to the Yalenang in 1983 and for two trips to Eipomek and Kosarek and adjacent areas in 1992; the Max-PlanckInstitut fur Psycholinguistik in Nijmegen and its director, WJ.M. Levelt, for his advice and support; I. Eibl-Eibesfeldt and the Fors chungs stelle fur Humanethologie in der Max-Planck-Gesellschaft, who granted the funds for my stays in 1986, 1989, 1994, and 1995; the Indonesian authorities, especially LIPI (Indonesian Institute of Sci­ ences); the Missionary Aviation Fellowship; the Summer

/

[] ()

,0 1, 2, 3 *

1. In Eipo texts the hyphen indicates the break between bound morphemes (not the boundary between two syllables!). This hyphen is repeated in the glosses. 2. In Eipo texts the hyphen connects postpositions and connectives as well as nominalizing -nye with the preceding word. indicates different kinds of grammatical information concerning one morpheme or one unanalysed word in the text. Two or more underlined words in the text corre­ spond to one word in the gloss. Square brackets with an appropriate sub­ script in the glosses indicate syntactic func­ tions. Round brackets in the free translations con­ tain additional information concerning either constituents missing in the Eipo text or comments or explanations related to the culture of the Eipo. The signs for “plus”, “minus”, or “zero” as used in charts indicate whether a feature is present, absent, or irrelevant. indicate first, second, or third person. 1. The asterisk marks reconstructed pho­ nemes, morphemes, or words. 2. In the index the asterisk marks words or entries as found either in the diet, or in Heeschen (1990b): additional information concerning 11

form, meaning, or function is given in this grammar. New Eipo words not part of the diet, are also marked by this asterisk in the index. p The question-mark as used in the glosses or free translations signals that the immediately preceding translation, gloss, remark, or indi­ cation is doubtful. ad. adjunct adj. adjective adv. adverb, adverbial (phrase) adversative advers. att. attenuative (manner of action) attributive attrib. C consonant caus. causative cf. compare col. collective comparison comp. con. connective d. dual derivational suffix der. dictionary (= Heeschen & Schiefenhovel diet. 1983) different subject (in the next clause) ds. dur. durative, repetitive E. Eipo emphasis, emphatic (ally) emph. exhaust. exhaustive (manner of action) fut. I, II, III immediate, near, and far future hab. habituative (for -/^-suffix of the verb) hortative-deliberative, optative hort. I. instrument (phrase) idiom. idiomatically) imper. imperative mode inc. inchoative (manner of action) incomp. incomplete (manner of action) inf. infinite (verbal form) int. interrogative (marker) intrans. intransitive io. indirect object (abbreviation used in the glosses) L. dialect of Larye or Una

12

lit. med. met. n. o.

literal (ly) sentence medial verb metaphorical (ly) noun, nominalizing 1. one, 2. direct object (abbreviation used in the glosses) O (direct) object (abbreviation used in wordorder formulas) opt. optative pa. I, II, III today’s past, near past, remote past perf. perfective (for -^-suffix of the verb) pi. plural poss. possessive post. postposition present tense pres. quotative particle qrec. reciprocal refl. reflexive relative, relativizer rel. repetitive rep. resultative res. subject s. subject (abbreviation used in wordorder for­ S mulas) sc. scene, background singular sgsomeone s.o. simultaneous sim. same subject (in next clause) ss. sth. something subordination, subordinate subord. successive sue. theme t. dialect of the Tani valley T. transitive trans. verb(al) V. 1. vowel, 2. verb (abbreviation used in wordV order formulas) verbal adjunct v. ad. verbal noun vn. verbal root v.root verbal stem v.stem Yale language Y.

PART ONE Prolegomena to a grammar o f the Eipo language

2.

THE EIPO AND THEIR LANGUAGE

2 / The history o f research and the state o f the art The history of research in Papuan languages has been reported reliably by Laycock (1975), Voorhoeve (1975b), Wurm (1982) and Foley (1986), whose studies also con­ tain a great number of references to unusual views as well as to early descriptive works. At the end of the nineteenth century the fact that two types of languages had to be distinguished, the Austronesian and the non-Austronesian, i. e. the Papuan languages, was generally accepted. During the first half of the twentieth century a great number of monographs appeared, among them studies which even today represent the only source for the lan­ guage documented there, and which offer reliable materi­ al for general linguistic studies. Pilhofer (1933) described two characteristics of the Kate language, which not until four decades later were discussed as typological features of Papuan languages and which since then have enjoyed the growing interest of general linguists: the “Satzinnenformen” (sentence-medial verb forms, cf. Wurm 1982:83-85) and the “Wechselformen” (switch-refer­ ence, cf. Haiman & Munro [eds.] 1983). Dempwolff’s description of Yabem (1939) was the basis for studies on verb serialization (Bradshaw 1978b, Bisang 1986); it is true that Yabem is an Austronesian language, but verb serialization is also to be found in several Papuan lan­ guages (cf. Foley 1986). For the time after World War II two developments can be recorded. 1. Thanks to the work of individual scholars (e.g. Anceaux, Cowan, Drabbe), thanks to the work of the Summer Institute of Linguistics, which has been represented in Papua New Guinea since the mid1950s and in Irian Jaya since 1973, and, last but not least, thanks to the growing interest of individual linguists in the Papuan languages (e.g. Foley 1991, Haiman 1980, W Seiler 1985, Voorhoeve 1965) the number of grammatical descriptions of single lan­ guages has grown continually and the reliable docu­ mentation of the vocabulary of single languages has also increased (cf. Keysser 1925, Flierl & Strauss [eds.] 1977). 2. The studies of Cowan, Capell and Wurm fur­ nished more and more proof of the genetic affilia­ tion of individual Papuan language families and phylums. Up to the fifties, the term “Papuan language” was only used in a negative sense, in so far as the Papuan languages as a whole were set in opposition to the Austronesian or Australian languages; but the term was in no way meant to indicate a relationship of the Papuan languages among each other (Wurm 1982:2). The research done by Wurm and his col­ leagues, among others Dutton, Laycock, McElhanon and Voorhoeve, almost all of the Australian Nation­ al University, Canberra, culminated in the remark­ able effort to subsume the incredible number of 750 Papuan languages, representing about 20% of all the known languages of the world, in the classificatory system of five greater and several smaller language

phylums (McElhanon & Voorhoeve 1970, Wurm [ed.] 1975, Wurm 1982). Language phylum here cor­ responds to language family, a well-known term also applied to the Indo-European languages. Concern­ ing this, Foley (1986:13—14) remarks: The historical and genetic interrelationships of the Papuan lan­ guages “remain hypotheses and conjectures, because detailed comparative work according to the rigorous methods of comparative linguistics [...] has yet to prove these links. They are bold hypotheses, indeed, and future research may prove Wurm right.” Foley himself retracts the division into “phylums” and speaks of only about 60 language families, the term being used here as it is known from expressions like “Romance language family”. In spite of such criticism, the heuristic value of Wurm’s and his colleagues7 classificatory efforts will probably remain; curiously, the division into families veils the fact that some language families are certainly more closely related to each other than others, and that some of them, with the same degree of certainty, are not related at all. Foley’s own efforts at a historical recon­ struction do not altogether contradict Wurm’s results (cf. chapter 4). Most of the grammars published up to now employ the model of tagmemics in their method of description, which is to say that research was guided by the presenta­ tion of data in whatever way and not by the theoretical problems of general linguistics. Only a few add methodo­ logical and theoretical considerations to their descriptions. Haiman (1980, cf. p.LII), it is true, describes phonology and syntax within the framework of generative transfor­ mational grammar, but the longest chapter by far, follow­ ing the “nature” of the Hua language, is dedicated to mor­ phology. Moreover, the author believes in a principle of iconicity: the structure of a language often reflects reality, and this is probably meant to put a check on an all too great discrepancy between surface and basic structures. Bromley (1981:12—16) finds the stratificational grammar helpful, but in general reverts to models making it pos­ sible to take semantic relationships and characteristics of discourse into account. He is mainly concerned with the relations between events and the participants in these events. Bruce (1984:3) mainly follows the model of tag­ memics, but the grammar “contains comments on certain theoretical questions where the structure of Alambak is particularly relevant”. Thus he deviates from the straight­ forwardness of a single method of description and fol­ lows diverse “theoretical approaches”. W. Seiler (1985:4), finally, would very much like to write a “theory-neutral” grammar, if this were possible: like all other authors he attaches great importance to a clean presentation of data. On the other hand, especially Seiler’s study has been enriched with excursions concerning the theory and the universality of descriptive, but theory-implying, terms. It stands to reason that these remarks concerning the state of the technique of description in no way represent an evaluation; there are excellent, readable tagmemic gram­ mars (cf. Foreman 1974), and the grammars characterized above serve as models, especially in view of their proxim15

ity to the data, their eclecticism and their theoretical restraint, to which any later writer of a grammar has to revert. The immense classificatory efforts and the number of descriptive grammars have, indeed, hardly been taken notice of by general linguistics up to now Neither in stud­ ies concerning the history of language, nor in considera­ tions concerning the adequacy of description, nor in typo­ logical and generally theoretical dicussions has adequate attention to Papuan languages been paid. It is the great merit of Foley (1986) to present these languages to the generally interested linguistic public. The treasures which the Papuan languages offer to soci­ olinguistics and to the ethnography of communication have often not even been located. This laboratory, in which small and smallest speech-communities mix, lan­ guages disappear, speakers of the same language differen­ tiate themselves from each other, speakers of different languages confederate and form marriage alliances, and in which multilingualism is a characteristic of man’s social nature, is only known to specialists. Only he who takes into account the cultural contexts of speaking and the rules of verbal interaction in the small communities with­ in a hamlet and in the “intercultural” communication between two such residential groups will perceive that “sprachbiinde”, typological features, borrowing, mixing of languages and characteristics of spoken language reach into and form the seemingly fixed system of a single grammar. All this still awaits research. 2.2 Dialects and languages In 1911 a Dutch military expedition camped near Mount Goliath (see map) for two days. One of the members, de Kock (1912), brought a word-list containing 97 words back with him. Since that time these languages had been known under the name of Goliath languages. Schiefenhovel rechristened the group of languages and, analogous to the Ok languages, in which ok means “river, water”, and due to the non-existence of a local term, called it after Eipo mek “river” (eastern dialects me, Yale mak). Another finding justifies this incidentally given name: in almost all myths of the Mek people the mek tree (Papuacedrus torricellensis) plays a significant part. Man is thought to be either created from splinters of that tree or accompanied by planks or shields made from the mek tree; Eipo mek “water, river” and Eipo as well as Bime and Yale mek “tree” are beyond doubt homonyms (cf. chapters 9.1. and 9.9 and Heeschen 1994a and b). Bromley (1966-7:299 and 1973:15—16) completed de Kock’s isolated findings: wordlists were added from Nalca, Korapun (where in the sixties the missions began to work), from the area of the upper Ok Cop and from a place called Wanan, which is still mentioned in Voorhoeve (1975a) and which is said to be located not far north of Angguruk, but which I have never been able to find. On the basis of this first informa­ tion, the Mek languages, then still called Goliath lan­ guages, were included in the Trans-New Guinea Phylum, the largest of altogether five larger and several smaller Papuan language phylums (McElhanon & Voorhoeve 1970, Voorhoeve 1975). 16

The linguistic and grammatical spade-work in the Mek area was done by WM, and J.E. Rule, R. Cutting (1972) and, together with the Rules, G. and J. Fowler (1972). Copies of these manuscripts were not available when we began research in Eipomek. With Louwerse, however, who worked in Langda, an exchange of written provision­ al results took place (cf. Heeschen 1975 and 1978a, Hee­ schen Sc Schiefenhovel 1983, Louwerse 1988, Heeschen 1990b). Together with my own studies in Kosarek (cf. Heeschen 1985 and 1992a and c), and with the paper by Elinor Young (1986) as well as the studies of Andrew Sc Anne Sims as well as Anne Sims and Linda Jones of the Summer Institute of Linguistics (1982 and 1986 on the Ketengban language in Omban) findings from all over the Mek area are now available. Since 1985 another team of the Summer Institute of Linguistics has been working in the Mek area, namely J. Sc R. Steinbring, in Kosarek. The results of our research induced Wurm (1982:187) to attribute to the Mek languages “provisionally the status of a sub-phylum” within the Trans-New Guinea Phylum. Bromley (1973:16) rightly concluded in spite of scanty data that the Goliath family clearly represented “a group of closely related languages sharing 50% or more basic vocabulary with each other”. This was confirmed com­ pletely by Heeschen (1978), who in the study mentioned found the smallest percentage to be 59%. It applies to the relation between Kosarek and Sirkai (situated somewhere north of Eipomek, but no longer marked on the map, because of the uncertainty of the exact location). When checking the Kosarek vocabulary gathered in 1976 I found several mistakes. As a result of the corrections the degree of relationship increased up to about 10% each. The Summer Institute of Linguistics evidently assumes far smaller figures, so that the close relationship of the Mek languages was internally criticised (Jan Godschalk, oral communication). The data of the Summer Institute of Linguistics are not available to me (cf. chapter 4, Silzer Sc Heikkinen 1984 and Silzer & Heikkinen Clouse 1991). On the basis of the calculation of 1978 and of the practi­ cal work with the informants in the field (i.e. with Eipo informants in Larye and in all other areas north of the central range, and with informants from Kosarek in Larye, Sela, Korapun south of the central range, and in all areas starting from Bime to the west to Nipsan and the Obahak valley north of the central range) the following groups of dialects and language areas emerge.2

2 The map was drawn by Carsten Riihlemann after Helmcke (1983) and according to the indications o f the author. The positions of Kosarek and Konae (Helmcke: Kono) have been corrected. All information irrelevant to linguistic work has been left out.

Map 1 The Mek language area

Legend • and names in capital letters designate mission stations or names by which the relevant area or the language spoken in that area is known to the Indonesian administration. Names in parentheses () designate languages. Names in small print designate rivers. Underlined names designate mountains. + designates mountain peaks. Names in bold type designate language families. The position of a (mission or government) station is identical with that of the airstrip, not with the posi­ tion of hamlets. In some cases the name and the position of a station are identical with the site of a central hamlet in a valley, for example Bime with the hamlet Turue, Tanime with the hamlet Dubokon and Eipomek with the hamlet Londinin. Where no other names are known I will use the name of the river or that of the valley for the language as well as the people: Bime people or Bime language or Sesom dia­ lect (of the inhabitants of the Sesom valley). The natives speak of themselves as clan members, for example Malyonang “people of the Malyo clan”, as inhabitants of a ham­

let, for example Dingerkon-nang “people of the hamlet of Dingerkon” and as inhabitants of a valley, for example Inlom-nang “people of the In valley” or Eipe-fum-nang‘ptop\z of the Eipo valley”. Names used in the literature are: Kimyal for the language of the Korapun people, named after one of the important clans of the Erok valley, Kim­ yal; Hmanggona for the language of the Nalca people; Una for the language of the people from Larye or Langda, named after the interrogative pronoun una “what”; Ketengban for the language in the area around Omban and Okbab; Yale (language) for the Kosarek people or the inhabitants of the In valley. The Eipo speak of the people in the east (from Tanime to Okbap) as the ware-nang “ware people” and of their language as wareyupe “ware language”; here again the interrogative pronoun ware “what” was used as a name. The expressions which the Eipo use for themselves are lik-nang “lik people” or likyupe “lik lan­ guage”, from lik,, which means “not to want, not to like”, serving as a badge. In other respects east and west serve as identification: in contrast to the speakers of the Angguruk language, who speak a totally different language belonging to the Dani language family, the language of the Kosarek people is called the yale yubu, meaning “language of the east” (the term Yale used here is not identical with the 17

name Yali used in the literature for the people of Angguruk and their language). And in opposition to the speakers of Ok languages in the east of the Mek area the inhabi­ tants between Omban and Okbab see themselves as speakers of the ketengban language, “the language of the west”. Accordingly, the Eipo, seen from Tanime, are ketengban-tam-nang “people from the west-side”, or, seen from the Fa valley, they a r yale-tam-nang “people from the east-side” or also iknabye-nang “people from the higher mountain areas”. In addition to that, as far as we know, the terms Eipe-nang “Eipo people” and Eipe yupe “Eipo language” had been in use among the Eipo before white contact; “Eipo” became the current spelling in such a short time that it was impossible to reintroduce the correct form “Eipe” (with labialized /p/, cf. chapters 4 and 6). There are two reasons against mapping language boundar­ ies, firstly the phenomenon of dialect-chaining, which is well-known among the Papuan languages, secondly the smallness of the settled areas. 1. Dialect-chaining means that the inhabitants of valley A are able to understand the dialects of the valleys B and C (situated either to the east or west), but not that of valley D; the inhabitants of B, how­ ever, still understand the dialect of D. For A, D is a language, for B and C it is still a dialect. Thus Ketengban is a foreign language for the Eipo, but not for the Tanime people, who still understand Ketengban, although the Tanime people are in con­ tact with the Eipo, and although the people living at the centre of the chain, in this case the Tanime, Bime and Kamume peoples, find themselves sur­ rounded by a dialect area stretching from Eipomek to Okbap. 2. The territory of the Eipo, for example, stretches for a maximum of 15 km along the Eipo river, in the east-west direction the farthest distance is 2.5 km. The extent of Kosarek, Nipsan, Tanime and Bime is about the same, Nalca and Larye are slightly larger (on the map these areas correspond to a space of 5 to 2 mm). Between the inhabited areas there are immense stretches of uninhabited country used only for hunting and gathering wood, nuts etc. Between Kona and the Sesom valley, for instance, there is a totally uninhabited, 25 km long high valley. Between Korapun and the Sesom Valley or between Larye and Eipomek there are 30 km, or 20 km of high mountain regions unfit for habitation, and south of Eipomek and Tanime there are 100 km2 of almost untouched mountain forest. Language or dialect boundaries would fake a structure that does not exist. In addition to that, it happens time and again that speak­ ers of one language settle in another language or dialect area, either because of marriage relations or trading-part­ nerships or because of banishment from their own com­ munity or conflicts with their clan members. During the past decade nearly all young men from the hamlet of Nohomas in Kosarek got their wives from Nipsan, and in recent times all young men of the Kisamlu clan in Eipo­ 18

mek have been married to girls from Larye. - Language and dialect boundaries become irrelevant in everyday interaction and communication or they are built up con­ sciously, for example when the Eipo set up an ideological barrier beween themselves and the inhabitants of the val­ leys in the west, whom they regard as their traditional enemies. It is a fact, though, that the languages change perceptibly whenever one goes to the east or to the west from one valley into the next or whenever one crosses the central range in either direction, and that the languages change only slightly whenever one walks along a valley either from north to south or vice versa. Instead of mapping boundaries I shall present the follow­ ing diagram. A plus sign (+) means that the speakers still understand each other, so that their languages, according to the definition, must be regarded as dialects. A minus sign (-) signifies that the speakers do not understand each other, so that they are members of different language groups. A question-mark means that either the speakers were not sure of their judgement or that I could not verify the relationship of the dialects or languages. Two ques­ tion-marks signify that it took the speakers several days to get used to the language in order to be able to understand it. The results of the lexicostatistic calculations of 1978 have also been taken into account. The classification offered by Silzer and Heikkinen (1984: map 9, cf. Silzer & Heikkinen Clouse 1991) contradicts my findings mainly in the following points: 1. The Yale language (Kosarek), Nipsan, Konae, Bu valley and Sesom valley doubtless form only one dia­ lect area. 2. The map does not take into account that the Una and the Tanime language have by far the highest per­ centage of shared vocabulary (95%) and that Eipo, the Tanime language, and the Una language, judged by their vocabulary, form one dialect area (Eipo-Una 93%, Eipo-Tanime 93%). 3. The dialect of Tanime, which morphologically is clearly a member of the eastern or ware dialects, is no longer understood in Okbab. Nevertheless, all dia­ lects east of Eipomek have been subsumed under the Ketengban language. 4. The area south of the central range and east of Larye-Langda has not been considered at all, although linguistically it is no terra incognita (see Heeschen 1978 regarding the dialect of Kinome and Louwerse, e.g. 1988 and before that in sketches).

Table 1 Dialects and languages in the Mek area Y a le

Yale

N ibsan

K o n a e N alca

Eipom ek

T a n i­ me

B im e

K am u ­ O m - O kb ap me ban

K o ra ­ pun

S e la

U na

K in o m e

+

+

-

-

-

-

-

-

7

-

-

+

-

-

-

-

-

-

-

-

7

-

-

+

11

-

-

-

-

-

-

7

_

-

-

-

-

-

-

-

11

7

-

-

+

??

11

-

-

-

-

+

+

+

+

11

-

-

+

+

+

+

+

-

-

+

+

+

+

-

-

+

+

-

-

?

?

-

-

+

-

+

-

-

-

-

Nibsan

+

K onae

+

+

Nalca

-

-

+

Eipo

-

-

11

-

T anim e

-

-

-

-

+

Bime

-

-

-

-

11

+

K am um e

-

-

-

-

11

+

+

+

-

-

O m ban

-

-

-

-

-

+

+

+

Okbap

-

-

-

-

-

11

+

+

K o ra p u n

-

-

-

11

-

-

-

-

-

-

Sela

7

?

?

1

-

-

-

-

-

-

Una

-

-

-

-

+

+

+

+

?

+

-

-

K inom e

-

-

-

-

+

+

+

+

?

-

-

-

In the table above three clearly defined language areas can be made out without much difficulty: 1. The entire East including the Southeast with Kinome; in this area only the relationship between the Okbab language and the ware language is uncertain. According to the criterion of mutual understanding they are different languages, according to the criterion of shared basic vocabulary (85%) they are merely dia­ lects. This problem will be treated further below. 2. The Northwest from Konae up to the Yalenang. 3. The Southwest with Sela and Korapun. The double question-mark characterizing the relation be­ tween the dialects of the Northwest and the Sela language may be premature. In any case, it mirrors the fact that both Sela, although situated further to the east than Korapun, and the Yale language have a phoneme system which has developed furthest from a still to be reconstructed ProtoMek, cf. the loss of -s- between vowels in the following line:

a)

Bime Eipo Sela Korapun Yale potong fotong hong vsvng hong “hair, fur, feathers”. The linguistic diversity of the outermost corners of the Mek area is also to be found in the percentage of the shared vocabulary (after Heeschen 1978:9): Table 2 Percentages of shared vocabulary Yale, Okbab, Korapun Yale Yale Okbab Korapun

66 78

Okbab 66

Korapun 78 63

+

+

+ +

For the Papuan languages Wurm’s opinion holds good that one speaks of dialects of a language “if their percent­ ages of shared basic vocabulary items are mostly above 81%, and as separate languages if these percentages are mostly below 78% and rarely above 81%” (Wurm 1982:70). If one proceeds from the centre, i.e. the Eipo language, things are less certain. First, in table 1 the double ques­ tion-marks strike the eye, especially the one indicating the relation of Eipo speakers to Konae. The Eipo speakers claim to be unable to understand the more adjacent Nalca language, but to have no difficulty understanding the Konae language spoken further to the west. In my opin­ ion the reason is that intonation and accentuation in Nalca are quite different. On the whole, it is not the amount of shared vocabulary which decides whether a language or dialect is understood or not, but rather suprasegmental characteristics, different particles connecting clauses and inflexional morphemes of verbs which are pro­ nounced differently, but are etymologically related. This is the case especially for the relation of Eipo and Bime as well as of Eipo and Nalca, because, according to the logic of lexicostatistics, Bime-Eipo-Nalca would have to form one language area. On the other hand, the same factors that prevent understanding here do not set up a barrier between Eipo, Tanime, and Larye. I shall here present the first person dual of today’s future (cf. chapter 7.2.4.5) for the illustration of the above-mentioned differ­ ences: (2) Larye (Una) Eipo____ Tanime bin-kwo-nim bi-nu-nam bin-ko-nim go-fut.I-ld.

go-fut.I-ld.

go-fut.I-ld.

The two of us shall go.

63 19

1

In spite of these differences Eipo, Larye (Una), and Tanime form one language area according to the criterion of mutual understanding as well as that of the percentage of shared cognates. But what is still more important is the fact that the inhabitants of the three valleys communicate with each other. There are many trading-partnerships and marriage relationships. In all of the three valleys the three “dialects” are spoken. Many of the children in Eipomek have mothers from Tanime or Larye, learn their mother’s dialect and also learn the Eipo language in the communi­ cation with their fathers and Eipo-speaking age-mates. The rhythm of acquisition of the first dialect from one generation to the next should follow that of marriage rela­ tionships; however, this has never been documented. For all these reasons the percentages of shared cognates do not present a correct picture: on the one hand they do not mirror the real differences in intonation, connectives and inflexional morphemes, on the other hand they do not reflect the existing network of communication. Table 3 Percentages of shared vocabulary in the centre of the Mek language area Nalca Nalca Eipomek Una Tanime Bime

77 72 72 65

Eipomek 77 93 93 89

Una 72 93 95 90

Tanime 72 93 95

Bime 65 89 90 93

93

With some caution a fourth language area can be singled out in addition to the three already mentioned, namely that of the Eipo language. Here one must always bear in mind, firstly, that the demarcation to the west (Nalca) and to the east (Bime) holds good under the assumption of a missing network of communication, of different suprasegmental characteristics, and of different connectives and inflexional morphemes of verbs, but less on account of the criterion of shared cognates, and, secondly, that in the language area of Eipo-Una-Tanime we indeed find a high percentage of shared cognates and a network of communication, but that especially this centre of the Mek area reveals phonological and grammatical differences making the assumption of language boundaries more like­ ly than that of a dialect area. In the western as well as in the eastern Mek area there are relatively large and homogeneous dialect areas, the centre of diversity, however, is situated in the more closely limit­ ed area between Larye, Eipomek, Tanime, and perhaps also Nalca. According to the rules of dialect geography the west and the entire east must be supposed to have been the scene of relatively recent migrations, proceeding from the centre as their starting-point. This general outline becomes more distinct when other than dialect-geographical facts and assumptions are added. 1. In the myth of the origin of the Kisamlunang, probably the oldest clan in Eipomek, it is said that the ancestor came along the Ca, Be, and Era rivers; I have identified the Ca river with the Ok Cop (cf. map). Nearly all the mountains which gave birth to the clans resident in Eipomek are situated between 20

Larye and Eipomek on the one hand and the Speelmann and Oranje mountains on the other hand. Louwerse, it is true, reports that there are Kisamlu people in Larye and Langda, too, and that they trace their origin from Eipomek (oral communication). But surely there is a difference between the first coming of the clans in, as it were, prehistorical times, as is recounted in the myths, and the tales of origin of lineages reporting “historical” migrations, schisms, expulsions, and movements of individuals caused by marriage and conflict; thus, in Eipomek there are many members of the Malyo clan, whose “historical” place of origin is doubtless in Larye, but who represent its “prehistorical” coming as a route along the rivers in the southeast (along the Kino, Era, and Mo river). After the first settlement there were doubtless many migrations: unfortunately, none of those conducting research in the Mek area except myself have documented myths of origin, so that the distinction I have made between the first coming and the later migrations has still to be inves­ tigated more closely.3 2. The Yulye, the ancestor of a clan from Diduemna and Nalca, took a route from Tanime via Eipomek to the west. In the myths the Yulye is often repre­ sented as a tree; in this shape his course can be fol­ lowed along the Yamek valley up to the Obahak valley, even to Angguruk, but not to the Yalenang (cf. chapter 9.5). 3. The Yalenye, ancestor of one of the predominant clans of the east, comes from, and was born by, the Lim mountain, which is situated in the centre of the Mek area. 4. Nearly all the clans of the Yalenang mention Mhemlo in the myths as the starting-point of their migrations to the west; Mhemlo is situated in the Nalca valley. 5. Nearly all the clans of the Yalenang speak of the Lim or Kamur mountain as their ancestors’ place of origin and birth (cf. map). 6. From Eipomek up to the Yalenang there are a few clans with identical names in each valley, for example the Nabyal clan, a sub-clan of the Kisamlu clan, and the Midin clan (cf. also the name “Kimyal” for the languages of Korapun and Nalca: in the myth of origin of the Kisamlu clan Kim was the first human being born to the ancestor by the pig; in y a l we find the Eipo word yala, yale, y a l “seedling, variety, clan”). 7. Furthermore, in the southeast and the northwest there are several clans unknown in the centre and in the east, a fact in favour of the migration theory. These clans are, for example: Nibsan, Lilbul, Soo, Yalak, Dabla, Dibul.

3 C f chapters 9.3, 9.6, and 9.7: the people first move to the west and then turn eastwards.

8. The above-mentioned common features of the Sela and Yale languages probably result from the splitting-up of a party moving westward beyond Mhemlo along the Yamek river, one group partly turning south, partly moving further westward; it was followed by a second wave settling Korapun and pushing parts of the first group either further to the west or forcing the southward-bound group again to the east. 9. That such a circular motion actually took place I also deduce from the fact that somewhere between Sela and Sainme there is the greatest break in the dialect chain of the Mek area. Sela, which together with Yale is the most innovative dialect, in the east borders on the phonologically most conservative dialect, namely Bidabuk, a place said to be situated three days’ walk east of Sela. Compare the mainten­ ance and development of intervocalic /t/ in the following line: (3.) Bidabuk Bime Eipo Yale Sela bitik bisi bisik pik pik “path, direction, goa1”. The circular motion also leads back to the probable centre of diversity and supports our dialect-geo­ graphical considerations. Even today the outer borders of the Mek languages can only be marked correctly in parts. The eastern border is probably formed by the ranges between Kloof and Sobger, between Okbab and Ok Bi and the Juliana mountains. The western borders are the lower Yaholi, the ranges be­ tween Kosarek and Yaholi respectively between Obahak and the upper Yaholi. The inhabitants of the Obahak and the Solo valleys are partly bilingual, speaking a Dani lan­ guage in addition to a Mek dialect. In the North and South the borders are lost in the far less densely populated flat land. The conditions are probably similar to those we found in 1976 30 km south of the Idenburg at about 140° east longitude. Speakers of totally different languages still keep close contact with speakers of one Mek dialect, and women of the Mek area were married to (or abducted by) them. The low density of population makes the marking of exact boundaries impossible. For this it would be nec­ essary to have a map showing even the small hamlets. North of Keleka there are about five hamlets with one men’s house each and a maximum of three to five family huts. During my flights to Kosarek and during some heli­ copter surveys I could make out neither settlements nor gardens beyond these hamlets. 23

by still more evidence besides the above-mentioned oral tradition and besides the dialect-geographical circum­ stances. In the sacred nets of their men’s houses the Eipo also keep sago branches; sago grows up to an altitude of 1000 m, the Eipo, as has been proved, have never seen the plant themselves, and sacred objects are not exchanged by trading partners. Taro is the food for special occasions, especially for feasts and rites; the sweet potato, however, is the staple food, but plays no part in religious tradition. The sweet potato (Ipomea batatas) was introduced to New Guinea 400 years ago, but it reached the highlands later than that, and the remote part of the Mek area maybe only 250 to 100 years ago. Only then a better colonization of the mountain valleys and growth of the population was possible. The myths tell of the joyful moment when men came together and increased in numbers. Thus, for the last one or two centuries, we may take a cer­ tain continuity of colonization and a general identity of settlers for granted. Helmcke (1983) compared aerial pho­ tographs of 1945 with those he took himself in 1976. In some cases clearings can be made out on the old photo­ graphs exactly in the same place at which they are found today. “Zusammenfassend laBt sich mit ausreichender Sicherheit feststellen, daB die vorgelegte Interpreta­ tion der amerikanischen Trimetrogon-Luftbilder von 1945 ergab, daB alle Taler, die zur Zeit der Gelandeaktivitaten des SPP (1974-1976) besiedelt wurden, auch zum Zeitpunkt der Aufnahmen bereits besie­ delt waren.”4 Helmcke has the general impression “daB die Bewohner dieser Taler zwischen 1945 und 1976 [...] nur relativ geringe Veranderungen der gerodeten Bereiche verursachten”5 (Helmcke 1983:39). The non-religious traditions of the Eipo, for example tales of wars, reach back to the times of the grandfathers of the now middle-aged or age­ ing men (and women?), that is back to the years around 1935; these tales do not provide any indications of too dramatic changes within the structure of colonization and population either. Even the generation of first settlers had to rely on other clans or sibs: marriage partners have to belong to different clans, children are members of their fathers’ clans; conse­ quently, we find patrilineal, exogamic, and usually virilocal marriage rules. The actual distribution of landed property does not mirror the claims of the first settlers. The clans settling there later also have landed property. The posses­ sion of garden areas, hunting grounds, and mountain regions is usually handed down to the sons in the patrilin­ eal line; but in consequence of the fact that bachelors can adopt orphans and leave them their property, land can pass to other clans. In addition to that, unused grounds

The Eipo and the Eipo language.

We have already mentioned that the Kisamlu clan and its subclans, the Nabyalnang and the Omkaraknenang, were the first to take possession of the Eipo valley; they prob­ ably came from the southeast of the Mek area. The as­ sumption that they originally came from the lower alti­ tudes north and south of the central range is supported

4 “It may be concluded with sufficient certainty that the interpreta­ tion presented here o f the American Trimetrogon aerial photo­ graphs o f 1945 yielded the result that all the valleys which were settled during the field activities o f the SPP (1974-1976) had al­ ready been settled at the time the photographs were taken.” 5 [...] that between 1945 and 1976 [...] the inhabitants o f these val­ leys caused only relatively minor changes in the cleared areas.”

21

can be left in usufruct to those who wish to evade con­ flicts in neighbouring valleys or who follow their married sisters. Women acquire usufruct by marriage, but only if the yoldane, literally “the fucking each other”, is good, which means that the relationship of the partners must be good and that they must have children. Thus, every insig­ nificant corner and every ever so remote mountain region is somebody’s claim, either in usufruct or as possession. In this context the hamlet as a community of settlers justifies its existence, and that independendy of any claim deduced from the first occupation and setdement of the land: on the one hand, there are members of different clans resid­ ing there, with quite different rights, but these members can intermarry; on the other hand, the property reverts to the hamlet on the extinction of a clan, even if there should be members of this clan living in remoter hamlets. - Clans that came from elsewhere and have not deposited any sacred objects from the times of origin in the men’s house, usually possessing none of their own, are granted the status of melimnenang “people who help” or “helpers”, in the performance of the rites. The settled areas and the gardens are situated at an altitude of 1300 to 2000m. The hamlets consist of up to 30 round huts, the number of inhabitants varying between 60 and 200 per hamlet. The Eipo’s chief occupation is the making and tending of gardens, in which they grow their staple food, the sweet potato (kwaning Ipomoea batatas), and, in addition, taro {am: Colocasia esculenta), sugar cane (kuy: Saccharum officinarum), bananas (kwalye: Musa paradisiaca), “pitpit” (pace: Saccharum edule) and edible, cultivated grass (teyang Setaria palmifolia). In the mountain forest the Pandanus conoideus is used. The women hunt and collect frogs, lizards, insects and larvae. The hunting of marsupials and birds, especially of a pigeon species {yalma), is of secondary importance from the alimentary point of view, although hunting plays an important part in Eipo thinking; all the rites and also the great dance and exchange feasts are preceded by hunting-periods lasting several weeks. As domestic animals there are dogs, used for hunting, and pigs, serving for exchange and for the maintenance of alliances and trading-partnerships; in all the rites pig fat is indispens­ able either for rubbing into the skin or as a sacrificial gift to the ancestors. The material culture of the Eipo is extremely poor. The most important tools are stone-adzes, stone-knives, and digging-sticks. As ornaments they have bird’s feathers, especially bird-of-paradise feather decorations, chains of pig, dog or marsupial teeth, nassa shell headbands and various bangles. Bows and arrows are weapons in war and for hunting. In the neighbouring valleys the culture is al­ ready richer, in Tanime and Diduemna people have large, painted door boards (shields) and in the north there are drums. In the egalitarian and segmentary community there is no formal, political leadership, but each hamlet has one or more big-men, who, by means of exchange transactions, win a greater circle of helpers and followers, and who are distinguished by their diligence, liberality, oratory power, by taking over the responsibility in war and by their knowledge in all ritual religious activities. The big-men take the lead in most public affairs, although this is not so by necessity. It is they who may summon others to the 22

boys’ initiation, to the building of the men’s house and to the making of big gardens. The Eipo very carefully watch their youngsters and young men for the attributes charac­ terizing a big-man and try to find out who has the stuff for a big-man in him. The Eipo are connected with the Larye area in the south and with Tanime in the east through intermarriage and trading-partnerships. From the upper Heime valley they (and all the other valleys in the west up to Angguruk) get the rough-outs for the stone-adzes. Great dance and ex­ change feasts stabilize and keep up those relationships; they are the highlights in an Eipo’s life and offer the opportunity to get rid of the surplus of pigs by serving them to the guests —for in New Guinea one does not eat one’s own pigs. The Eipos’ relations to the Famek valley in the west were hostile. The inhabitants of the Fa valley were their traditional enemies. The Eipo use a special term for this w a r - ^ / - ; for enmity within their own val­ ley, however, they use the term abala “armed conflict among normally allied parties”. The ritual and religious activities of the Eipo concentrate on a repetition of the founding-acts of their culture: the appearance of the ancestors, each of them born by a spe­ cial mountain, and the appearance of the cultural goods, first of all the men’s house, offering the first men an orderly shelter in the unfinished nature and in the midst of the elemental fertility of female powers; the men’s house is the only guarantee for the continuance of the other goods, such as the cultivated plants, the pigs and the for­ mulas of healing and blessing. The founding-acts of the culture also include the tearing-off of moss from trees and rocks, the making of valleys and of cleavages enab­ ling the.water to flow down, for example the making of the river beds and the hardening of the still soft and shak­ ing ground. Piety means being close to one’s ancestor and warding off all the strong forces emanating from other ancestors and their wives. The relics of the ancestors and of their first female partners are kept in the men’s house, each of which belongs to a single clan. Such partners, as it were ritually banned counterparts of the male world of the ancestors, are the totems of the clans: the pig, the dog, the sun and the moon or the first sacred digging-stick.6 Further subjects of religious care are the growth of the nutritional plants and the prospering of the boys. The means by which knowledge is passed on are rites and cere­ monies, not verbal teaching. The names of the ancestors, animals, plants, and implements are tabooed and sacred, not so much the acts themselves. According to Schiefenhovel (1988:73) the total population of six hamlets in the upper Eipomek valley (Munggona, Kwarelala, Dingerkon, Imarin, Mumyerumde and Kabcedama) was approximately 440 (in August 1975). In Dingerkon, where I lived and worked, I counted 52 in­ habitants, excluding babies and infants in April 1975 (Schiefenhovel: 60). The hamlets of Moknerkon with about 50 and of Talim with about 120 inhabitants must be added, so that there is a total number of about 600. Of these at best two thirds spoke the Eipo dialect, the others either spoke the Larye dialect (unci) or that of Tanime 6 Cf. chapters 9.2, 9.3, 9.6, and 9.7.

(ware). In Dingerkon, with its high proportion of men belonging to the Malyo clan (whose origin is Larye) and of men with strong ties to the Tanime valley (where they possess gardens), the proportion of Eipo speakers was dramatically smaller. Of 31 male speakers 10 spoke Eipo, of 21 female speakers there were 14 who spoke Eipo. As a child learns its mother's dialect, the relation will be changed in favour of Eipo within the generation growing up now. This can already be noticed among the male speakers: of ten Eipo speakers seven were young men under the age of 20. Two male Eipo speakers (Bingde and Melingde) and a female one (Berekto, meanwhile married to a ware speaker, Bobuk, and, in 1989, mother of three children) were the offspring of the marriage of Keblob (Larye speaker) and Danto (Eipo speaking woman from Munggona). After very careful calculation, instead of 400 Eipo speakers (two thirds of the population), only 300 (half of the population, when proceeding from the situ­ ation in Dingerkon) remain. If, however, the Eipo speaking women in Larye or Tanime, their children, and the inhabi­ tants of the tiny hamlets in the lower Eipomek valley, Simsimwe and Kererimna are included, one again reaches the figure of approximately 400. There are far more numerous speech communities in New Guinea (with up to 60,000 speakers) and some that are a lot smaller (with less than 100 speakers). But if we identify the residential unit (the hamlet or a cluster of neighbouring hamlets) with the community of social interaction, and if we take my calculations with regard to the number of languages, their speakers, and of the num­ ber of hamlets per dialect community as a basis, we arrive at a figure between 400 and 500, which represents the sta­ tistical average for New Guinea. Speech community qua social community does not exceed 500. This figure is interesting for two reasons: firstly, one arrives at similar figures for Australian languages. At precontact times there were approximately 200 000 aboriginal inhabitants, 600 tribes and 200 languages. The number of speakers per language, accordingly, is 1000 and the number of speakers per tribe and closer social community is 333 (Blake 1981:4-6, Dixon 1980:18). Secondly, according to Forge (1972), egalitarian societies from 450 members upwards tend to split in cases of conflict, i.e. these groupings lack the institutions enabling them to settle conflicts and to keep the community together. I mention these figures be­ cause, on the one hand, we more or less instinctively tend to think of speech communities in dimensions of several thousand or even millions of speakers. On the other hand, these figures are of great consequence for our con­ ceptions of norm, of the synchronic stability of a lan­ guage, and of the smooth transition of a language from one condition to another. Offhand it is to be supposed that the normative forces are less influential, and that individuals, for example eloquent big-men, or groups, for example men's house communities, are of far greater importance. Languages probably change more easily and quickly. Traditional models of diachronic development and hypotheses of typology must fail, because psycholinguistic and sociolinguistic factors, i.e. the laws of language production, of understanding, of acquisition and of the conditions of speaking in tiny communities have to be given a fuller consideration. According to my knowledge,

these correlations have not yet been studied systematically. There is evidence in the literature, however, focusing on a better investigation of those factors. Lithgow (1973, cf. Conrad & Dye 1975) found out that the Muyuw language, spoken on the island of Woodlark, changed 16% of its basic vocabulary within the lifetime of two generations: this change approaches the rate marking the division into two languages. Pawley (1974) assumes that smaller communities permit a faster change of the grammatical strucure. On the other hand, Bradshaw (1978a:45) believes that processes of borrowing and assimilation are to be expected more frequently in bigger speech communities than in the isolated and fragmented communities of precontact times. On the one hand, the mutual assimilation of two languages via pidginization and Papuanization within two or three generations has been documented (Thurston 1982 and 1987), on the other hand, Lincoln remarks (1976) that speakers avoid such an assimilation. The divergent findings allow no general con­ clusion. One would presume that, first of all, there should be a difference between the change within a speech com­ munity and a change caused by multilingualism of speak­ ers, and by exchange and assimilation of two languages; but unfortunately, this cannot be proved because of the smallness of the speech communities. The findings in Eipomek are not unequivocal, either. As mentioned above, three dialects are spoken in the valley. Speakers of one dialect have a passive understanding of the others, but neither can I prove a mixing of dialects nor the active use of a dialect except of the speaker’s “mother tongue”. Words like deja for Eipo mana “marsupi­ al” are immediately identified as belonging to the Tanime dialect, but never actively used. All the same, the formulas of healing and blessing as well as the texts of songs and dancing-songs are usually not in Eipo, the former are mostly in the older form of the Tanime dialect, the latter in the Tanime as well as the Larye dialect. When an Eipo reports the words of a Tanime or a Larye speaker in a nar­ rative, he also uses the corresponding dialect. On the other hand, there are examples of extensive borrowing between the Angguruk language, which is a Dani lan­ guage, and the Yale language (cf. chapter 4). Although processes like quick changes, borrowings and assimilation cannot be found in the dialect system or are still unknown to me, they are, all the same, to be met with in the speech and style of individuals. Here the topic of the smallness of the speech community and its normative power again comes up. However, the findings of other studies should not be put aside, because they give an idea of what is generally to be expected in a small speech com­ munity. The following remarks really require a knowledge of the grammar; they are inserted here, however, in order to cau­ tion the reader against all the rules in the grammatical part. When it is stated: ‘the Eipo language has the rule x’, the reader should bear in mind what has been stated here about the smallness of the speech community and its norms. The rule should properly read: the speaker y or the group of men from such or such a clan or the group of co-initiates say “x” under certain circumstances, for example in exhor­ tative speeches or in fairy tales. Some examples: 23

1. The vocabulary used in fairy tales and in songs and in the speeches of big-men is often intentionally enigmatic. Thus the narrator of a fairy tale instead of the commonly used “walk” may say “stumble, dash, disappear” or instead of “distribute”, “place into the water” etc. Information is also a riddle, which has to be solved by the hearer. 2. Normally the bound morphemes for the deictic particles like ei- “above” are compounded with -tarn in order to form free morphemes or words: eitam “above”. In songs -de is often used instead of -tam, some speakers use -tanik, and without any percep­ tible regularity, one can often hear -uk, so that there are three additional forms: eide, eitanik and eiyuk. 3. It is often not predictable whether, when com­ pounding two verbs, the first terminates in one of the two possible consonants, -b- or -k-, whether these consonants are replaced by an -/-stem or whether a derivational affix harmonizing with the vowel of the stem syllable plus -/-stem is added. The compound of j ^ - “fall” and kwe'teb- “make a move­ ment away from oneself” is sakkm'teb- as well as sakalkwe'teb- “dash off”, the compound of melib“help” and yan- “come” is melilyan- and melibyan“come and help” (cf. chapter 7.2.4.1). 4. The speakers have the choice of several possibil­ ities when linking sentences or clauses. Bingde loves the following construction: (4, cf. chapter 8.5.2) “Bi-lulum!” winyab-ik. Winyabne dare, aik be-(y)uk. go-2pL/hort.

say-3pl./pa.III say/vn.

t./ds.

hut go-3sg./pa.III

They said they should go, after that he went into his hut. Instead of winyabne dare Melase prefers: winyabik ora “they said and then”; Buk consistently uses a third possibility in many of his utterances: instead of win­ yabik dare he says winyabukabiye “after having said”. Instead of taking up the verb “say” in the second sentence the constructions at the beginning of the second sentence can also finish the first sentence, so that it becomes part of the chain:

c?)

Sik-de: “Bi-lulum!” winyabne dare, aik be-(y)uk. they-but

go-2pl./hort.

say/vn.

t./ds.

hut

go-3sg./pa.III

But when they said: “Go!” he went into his hut. Such chaining is typical of Bingde, Babyal or Bowungde; Melase sticks to the first possibility. Whether the subject change from the first to the second clause is marked, is again a question of individual preferences or of the speaker’s mood. (6) Winyab-ik ora, aik be-(y)uk. say-3pl./pa.III

when

hut

go-3sg./pa.III

When they said this, he went into his hut. Winyab-ik-buk,

aik be-(y)uk.

say-3pl./pa.III-when/ds.

hut

go-3sg./pa.III

When they said this, he went into his hut.

The construction in (6) only stresses the successive­ ness, the one in (5) the antagonism and the change of subject, the one with winyabne dare very probably includes the change of subject. These characteriza­ tions, however, suggest a regularity which is only conditionally valid, not generally for all speakers at all times, but only for the individual speaker and even then only for one text. Whether the sentences are chained in the manner described above or not is equally a characteristic feature of an individual speaker’s style. 5. The possibility of chaining and the number of noun phrases in a sentence depend on each other. Whether the speaker uses a form of summarizing or a form of distribution, is quite often a matter of style or individual preference (cf. chapters 3.3.1, 8.3.2.3, and 8.3.2.4). Compare: (? )

Yulye arye

motokwe kweb-reib-uk.

Yulye

mountain

s.

scoop-put-3sg./pa.III

The Yulye created the mountains. Yulye ya-lam-uk Yulye

ora, motokwe kweb-reib-uk.

come-hab.-3sg./pa.III when mountain

scoop-put-3sg./pa.III

The Yulye came and with that (along his way) he created the mountains. One has to bear in mind that the second way of con­ struction is often only used to avoid a sentence with two nominal phrases and a possibly necessary sub­ ject-marking in the first sentence. Here also the speakers have a choice. Bowungde, for instance, and Bingde use the second possibility, which is character­ istic of quick speakers and narrative texts. A system­ atic speaker, however, like Kwengkweng, who con­ sciously chooses his words in the presence of a foreign, European hearer, uses the second possibility as well. —Short sentences and the “and-then-style” are typical of Melase, long chains of Bingde and Bowungde, and long chains with subordinate clauses of Kwengkweng and Babyal. The grammar of Eipo is a field of possibilities. The influ­ ence of individual speakers must be rated higher than in our speech community. The Eipo speaker is not confronted with a society of millions and its normative possibilities through writing, school and mass media, but possibly only with a group of co-initiates or a men’s house community with not more than 15 members: some of them, and also women and children, even speak another dialect. An individual’s style can easily serve as an example in such a small group, but it can just as well be quickly forgotten. There is a quicker and higher turn-over of rare possibil­ ities used by only a limited number of speakers. Thus, the term “Eipo language” has to be used carefully. Virtually wedged between two larger dialect communities, with its stock of speakers changing quickly and dramati­ cally from one generation to the next, and exposed to the individual features and preferences of its speakers, Eipo is perhaps only a passing phenomenon in the history of the Mek languages. When, according to Laycock (1982), “lin­ guistic diversity” is also based on the conscious choice of

the speakers of Melanesian languages, a grammar has to study diversity as well as the conditions enabling a choice. - Some of these attempts at characterizing Eipo will again be found in the next two chapters. 2.4

The vocabulary

The dictionary and the supplement to the dictionary con­ tain 5682 main entries and 933 sub-entries (Heeschen Sc Schiefenhovel 1983, Heeschen 1990b); the difference be­ tween main and sub-entries can be seen in the following examples: main entries are for example using “eye”, duk “egg, seed”, uwula “secret”, dob- “take”, kanye “spirit, soul”, tenibren- “consider”. Under these main entries ap­ pear as sub-entries asing duk “eye-ball”, uwula dob- “steal”, kanye tenibren- “think, ponder about”. Most of the sub­ entries consist of such compounds and idioms. An entry like (8) na kanye I

mind

this is what I think, this is my concern was also regarded and inserted as a sub-entry because of its idiomatical features. The absolute figures, however, are irrelevant except when one compares them to other lexicostatistic works, points out the existence of “open” and “closed” word classes, makes at least an attempt at show­ ing the problems of such a classification and tries to char­ acterize the vocabulary of a small community in view of semantic structures as well as by attempting to determine the frequency of certain words and groups of words in common speech or in certain ways of usage. The quantity of the vocabulary can be compared with the following results of lexicographic works. 1. In comparison with the number of words in the dictionaries of languages of western civilization, which contain up to 500,000 entries and which store up and accumulate the speaking of many genera­ tions, trades and professions, poets and writers as well as dialect communities, the small number can­ not be a surprise. The dictionaries of Papuan lan­ guages contain between 2000 and about 8000 entries: Awa 2900 (R. & A. Loving 1975), Ekagi 4000 (Steltenpool 1969), Enga 5200 (Lang 1973), Fore 3600 (Scott 1980), Kapauku 2000 (Doble 1960), Kate 8000 (Flierl & Strauss 1977), Kewa 4300 (K.J. & J. Franklin 1978), Selepet 3200 (K.A. & N.A. McElhanon 1970), Telefol 4400 (P. Sc A. Healey 1977), Toaripi 4000 (Brown 1968), Wahgi 3200 (Ramsey 1975, vgl. Thurston 1987: 61—64). The numbers are only partially comparable because of the different methods of separating main- and sub­ entries. The most comprehensive dictionary, that of the Kate language, is based on the work of mission­ aries and linguists since 1886. The language became a vehicle of missionary work and schooling in other language areas as well, borrowing a number of words from Tok Pisin, one of the lingua francas in Papua New Guinea.

2. Dixon (1980:2) guesses: “Australian languages probably each have a vocabu­ lary consisting of at least 10,000 lexical words [...] This is about the size of the working vocabulary for the average citizen in any country.” It was mentioned above that in Australia one prob­ ably finds conditions of language diversity and of numbers of speakers similar to those in Papua New Guinea. According to my calculations dictionaries of Australian Aboriginal languages do not contain more entries than dictionaries of Papuan languages.7 3. The publishing house Klett issues a well-known series “Grund- und Aufbauwortschatz” for most of the European languages; the basic vocabulary con­ sists of 2000 items, the vocabulary for the advanced learner of 2500 additional items. When mastering the basic vocabulary, a reader is able to understand 85% of a text, when mastering the enlarged vocabu­ lary even between 90 and 95% of an unknown text (Giovannelli 1977). The 4500 words and idioms must be related to the approximately 50,000 items contained in the dictionary of the Italian language, for example. - The dictionary of the Yale language was almost complete in 1985. In 1986 and 1989 I again worked on the dictionary and on new texts in Kosarek; in 1986 I found 30 new words, in 1989 I discovered only ten new words in the new texts. 4. The vocabulary of a German child at the age of 5,8 - 6,0 years consists of about 6000 words (Augst, Bauer Sc Stein 1977). This comparison is in no way meant to imply that the vocabulary of a non-accultured people could be called childish or primitive: but the still restricted environment of a child and the number of objects it perceives and manipulates can probably be compared to that of a community whose scope covers the area of a four hours’ walk and whose material culture is extremely poor. 5. Pidgin speakers do not need more than 1500 words (Hall, quoted by Miihlhausler 1974:73), and the sign languages of the Australians consist of up to about 1400 signs (Kendon 1988:112). From all this one may cautiously draw the conclusion that the dictionary mirrors the vocabulary adequately as far as quantity is concerned. What is lacking are mainly the names of mountains, stretches of open country, and rivers; it also does not contain all the names of plants and animals. But when considering the dialect conditions, the turn-over of the vocabulary from one generation to the next, the borrowings, and the expressions used in sacred formulas and songs the number of words - the basic and enlarged vocabulary - used by an Eipo in daily life is probably far smaller than 6000.

7 E. g. Yir-Yoront: 3700 (Alpher 1991), Western Desert language: 3200 (Douglas 1988, Pintupi: 8100 (K. C. & L. E. Hansen 1992), Walmajari: 4300 (E. Richards & J. Hudson 1990).

25

1. It has been stated already that intensive borrow­ ing occurs sometimes in the border areas of the Mek languages in the east and west. Occasionally words also travel from the periphery to the centre. In the Fa valley, situated next to the Eipo valley in the west, people use the word sabalho “ceremonial shield” from Angguruk sapalhe instead of the Eipo kelabye or Yale kelabi. The quickness of borrowing and the tak­ ing-over of great parts of the vocabulary makes the attempt at classifying languages an unwonted and difficult undertaking (cf. Dutton 1982, Thurston 1982 and 1987). Borrowings and systematic relexification, as may be applicable to the relations between languages cannot be transferred to groups of dia­ lects without close examination. The Eipo carefully keep separate the three dialects spoken in their val­ ley. There is only one field in which other dialects enrich Eipo. In songs and dancing-songs the Eipo use the technique of parallelismus membrorum. There is a great need for new words. One item of the figure of speech is often taken from a neighbouring dialect or imitated from it, first in the pronunciation of the neighbouring dialect, then in analogy to Eipo pho­ nology. Thus Tani botamubre in the following line of a song

(9)

Inas botamub-re, Kelas bokarik-de Inas

circle-3sg./pa.I

Kelas

go/to/and/fro-3sg./pa.I

Inas, she has circled, Kelas, he has walked to and fro may one day become an Eipo-proper boltanguble in comments, interpretations and reproduction in songs of their own. (The verbs in the song allude to the clumsy way of walking of the Europeans. The verb is a compound of bob- “carry”, tangub- “turn” and the suffix T. -re respectively E. -de for 3sg.). Many of the Eipo words included in the 6000 items are such artificial constructions existing only in song for the time being. 2. Another domain in which there is more than one expression for one thing is the technique of giving names. The Eipo distinguish between the dibe si, the “true name”, xhefurume si, the “name indicating the very essence or origin”, and the limne si, the “sham name”. In addition there are the names called “codes” in our dictionary and the sacred, tabooed names or secret names. The “true name” is the one used in daily communication. Semantic or formal characteristics can only be ascertained for the smaller part of the other names. Attempts at classifica­ tion and generalization, puns, and circumlocutions characterize the more transparent ones among those names that are not “true names”. The easiest case is that of the “sham name”; it imitates the sound of the true name, like Baye for Babyal., a per­ sonal name, where Baye is also suggestive of bay “forest”. Thus, the singer of a love song can say that this or that woman goes into the forest, mean­ ing she has an affair with Babyal. Examples for the furume si are : y o dina “an insect species”, literally translated “wood or tree food”, which means that

the insects live on and upon wood and trees; mure “red soil from which red colour is won”, for the true name sisilya aiming more at the shades of colour and at the function - sisilya is used for paint­ ing and decoration; and yanduana “cricket species”, literally “the coming”, instead of the true name binduana “the going away”, which is probably a pun: usually crickets jump off when somebody comes along, but they must have “come” from some­ where, i.e. must have a place of origin. Most of the code names are circumlocutions: menekelbana “the going on ahead” foryan “foot”,yongona “the softening” for mok “rain”, thus “rain softening the earth”, aiktam teknanye “the one which stands beside the house” foryurye “cordyline {Cordyline terminalis)”\the cordyline is ritually planted beside the men’s house in order to guarantee the solidness of the earth. These names are used by the Eipo in spe­ cial, tabooed areas, for instance near the dwellingplaces of the ancestors and of the spirits of the dead, where it is forbidden to call things or persons by their names openly. The Eipo learn the sacred names of mountains, ancestors, animals, cultivated plants, parts of the body, and tools during initiation; women and those who have not yet been initiated are not allowed to hear these names. Only a few of the secret names can be explained: balum cang “something long that is uncovered” for penis; sarab, cf. sarabsarabana “the cracking (noise made by the cassowary when breaking through the bush)” for kwitma “cassowary”; moma fmfinya “the one decorat­ ed with tendril-like drawings” for kwit ateka “sacred, central post of the initiation hut”; maybe this post, at some time, was not only wrapped with green leaves but also painted and decorated; salsalape “(a crowd of) men” for keting “sun”; the sun, conceived of as male, is the totem of the Tengket clan. The meaning of the secret names is often doubtful, to be sure; but it is certain that their phonology as well as that of many words used in sacred formulas points to older layers of the language or to the east­ ern dialects. Thus, the names of mountains and those of ancestors conceived of as female often have the suffix -ner. Limner “Lim mountain”, Wisalner “ancestress of an extinct clan”. Ner in the dia­ lects east of Eipomek is the word for “woman”, Eipo kil In the sacred name sakuryang for “(wild) game, fire-tongs” the syllable-final -r- also points to eastern dialects (cf.above the verb forms botamubre and boltanguble). All these names are part of the vocabulary, but they are rarely used and cause the dictionary to swell to a size veiling the effectiveness and the smallness of the “basic vocabulary”. 3. Many words only belong to individuals, groups, one generation or one hamlet. Only the initiates know the sacred or secret names. Only the women who compose verses and songs understand all the words and expressions used in their songs. Of pas­ sages that are difficult to understand men say:

( 10) Kilape

sik

tenine.

woman/pl.

they

think/vn.

treeless mountain tops” and telibak “slope at the be­ ginning of the timber-line”. In everyday speech it is:

This is what women invent. In the upper Eipo valley “throw” is generally donok-, only in the hamlet of Talim nonok- is used. From one speaker only, Babesikna, I heard -tanik instead of the normally used -tam in the deictic particles, for example: (li)

_

ei-tanik . up-side

instead of ei-tam up-side

up there. The newly initiated boys agree upon calling certain things or events by different names; this is how the code-names like aiktam teknanye for the normally usedyurye “cordyline” come into being (cf. above). When with the Yalenang, I got a verb desek- “stand up from a meal”; the meaning could be deduced from the components de- “eat” and sek- “stand”, the special meaning, however, “stand up from a meal, showing an approaching person that nothing has been left over” was known only to some boys whose fathers were brothers. All of them belonged to the same men’s house. Word taboos are known from other Papuan lan­ guages and especially from the Australian aboriginals. No one was allowed to use the name of a deceased and the words for things he possessed or which were associated with him. They fell into oblivion (cf. Holzknecht 1988, Kendon 1988:89, 447—451, Simons 1982). Of the Eipo I only know that they avoid touching upon old things or stirring up stories of the lives of the deceased. The old people no longer sing or tell stories, too many of their tales or songs would have to allude to the deceased or to events of the past. The Eipo explicitly say: ( 12)

Kam-nang-tam yupe teleb. life-people-side

speech

good

The speeches and tales of those who live or are young are good. It must be assumed that a considerable part of the vocabulary changes in the course of generations. When looking more closely at the special oratory forms like speeches of big-men, songs, and fairy tales one discovers where the change will probably take place. There is a striking difference beween the broad field of meaning of some verbs and nouns in daily speech and the marked, deliberately new and elaborate expressions for exactly the same things in those special oratory forms. “Slope” is kin in com­ mon speech, only in the songs two different expres­ sions are used: lokobak “slope immediately below the

(13) Kum yan-ma 1. wind

come-dur.-3sg./pres.

The wind is blowing. In the songs “the wind ruffles, tousles, rumbles and trembles”: bungbung ab-? tingling a b m i m i ah-, burwur ab-. Bin- “go” normally covers the following meanings: “go, fly, wind, run, rush off”. Only in songs and fairy tales we find newly coined and more pre­ cise idioms, which the story-teller strews into his text as a kind of riddle, the meaning of which must be guessed by the hearers, who appreciate it as some­ thing new and interesting. The hero of a fairy tale does not “walk”: (14) Lare kin

ei-foub-lob-le.

Lare

up-stumble-incomp.-3sg./pa.I

slope

He has stumbled down the slope of the Lare river from above. “Stumble” here does not even stand for “walk in an unsteady way”, but for the unusual haste with which the hero returns to his hamlet after his adventure. In tales (for example of war and strife) we also find more precise expressions: (15) Kobolkoukwe dak-de

yangan-m-uk.

spread/legs/ad.(?)

come-dur.-3sg./pa.III

beat-inf./sim.

He came spreading his legs in jumping and beating them together (showing guileless joy). Lelang keil, taram ou-faka,

a-faka

Lelang

here-split/stem

leg

breast

down-split/stem

yanga-(y)uk. come-3sg./pa.III

Lelang scraped the earth with chest and legs (i.e. he crawled along the ground). This deliberately new and at first sight often dark preciseness must not be confused with a quite dif­ ferent way of precise description of activities or events, which we are unaccustomed to. Two verbs with a general meaning placed together in a series give a more specific description of an action than the single verb which, for example, the English or German language provides for the same action. The Eipo language does not only say that something falls; it also obligatorily adds the position from which something or someone falls: (16, cf. chapter 7.2A2.2) Tek-uke sak-ma-1. stand-inf./sue.

fall-dur.-3sg. /pres.

He falls from a standing position. Bongob-uke sak-ma-1. lie-inf./suc.

fall-dur.-3sg./pres.

He falls from a lying position. 27

While in (16) the preciseness is a common grammat­ ical feature, in (14 and 15) it is a distinguishing attrib­ ute of a special style. The words and idioms from the sphere of songs and fairy tales probably make up the part of the vocabulary that changes quickly. My only proof for this is that my informants, when explaining songs and fairy tales, often stopped, dis­ cussed the matter among themselves and only hesi­ tatingly agreed upon a meaning, adding that the expression was either dit si, i.e. belonged to the realm of songs, or that it was nukna yupe, i.e. belonged to the language of fairy tales. Lithgow (1973) reports a case of very rapid change. He compared the language of the older, middle and young generations, after noticing that the young often had difficulties in understanding the language of the older generation or sometimes did not even understand them at all. He draws the conclusion: “In Muyuw there has been a 13% change in cognate count and significant grammatical changes in 50 years or less. This is a more rapid change than that which is assumed in most lexicostatisdcal studies.” (Lithgow 1973:108, cf. Thurston 1987). Lithgow’s figures must be seen against the back­ ground of the early theories of lexicostatistics: it was assumed that about 15 —25% of the basic vocabu­ lary was changed within a period of one thousand years and that the number of shared cognates deter­ mining the boundary between languages, on the one hand, and dialects, on the other, was 81% (cf. Gudschinsky 1956). This means that the renewal of the vocabulary takes place more than twenty times as fast as had been originally assumed, and this also means that new languages can arise in a period of 50 to 100 years instead of one thousand. Thurston (1987:65-67) believes that the main rea­ son for rapid change and renewal is the smallness of speech communities; he arrives at the same conclu­ sions as the author (Heeschen 1992a). Small speech communities have no norm-giving institutions. An old expression is not kept alive by being “registered” in dictionaries or traditional texts, so that it can occa­ sionally be revived by some speakers. Nor are there any means which present to the speaker ready-made directions for use deriving from la langue as fa it social. On the other hand, the acts of speech do not have to assert themselves against thousands of speakers in order to become part of the language. The step from the individual act of speaking to a momentary and transient norm is a short one. Each speaker can at once exercise his influence upon the norm, which may be accepted for the time being in the restricted circle of his relatives, his men’s house community or among his trading-partners. But the fact that each speaker in these small circles can act as an innovator results in a quick disappearance of newly created words. Creation as well as disappearance follow each other more quickly than in larger speech commu­ nities. On the one hand the following statement is true: “Families and small networks of close friends rapidly generate words and idioms based on com­

mon experience, peculiar to themselves and difficult for outsiders to interprete correctly” (Thurston: 1987:66). On the other hand, it is equally certain that these small groups are unable to establish and keep up norms, when women are brought in from an­ other, equally small community or when it seems necessary to form alliances with yesterday’s enemies. 4. Finally, the dictionary contains mistakes, which, when corrected, reduce the number of items. Single mistakes like the following, however, play only an insignificant part. (17) ngat tokwe hard

soil

clay is probably identical with kat tokwe, simply “hard soil or soil of a solid quality”.8 Nimirye in nimirye kelapey in the dictionary rendered by “(female) lover” is probably a compound of nimi “man” and nirye “all”; the correct translation would be “woman belonging to many men”. Deye “that which is the reason for sth., cause, interior cavity” has not been recognized in the following examples: (18) Memerei

le-lam-ak.

invocation

speak-hab.-3pl./pres.

They invoke the ancestors. Mereye

don-ma-1.

labour

take-dur.-3sg./pres.

She is in labour. In the first instance of (18) we have a compound of mem “taboo” and deye, in the second of me “child” and deye. The free translation should be: “They utter the first and tabooed prayer” and “she feels the be­ ginning and the cause of the child”. Beside such errors there are also systematical ones. Between the first and second stem of a verbal com­ pound a derivational affix, consisting of the vowel of the first base and -/-, can appear, for example -alin sakalkweteb- “drop sth., (in fairytales for) rush off”, which again consists of sak- “fall” and kweteb“make a movement away from oneself”. Beside this there is a verbal compound without a derivational affix sak(k)weteb~. There are quite a number of such pairs and the compilers have tried to find different meanings for the different forms. The effort was in vain. The verb with a derivational suffix only follows euphonic rules, preventing a direct succession of heavy syllables or of two stressed syllables. Beside dongob- “put (sth.) on top o f”, deib- “put” and dob“take” the second stem of a verbal compound can also be lob- “release, let go”. Almost all the com­ pounds were inserted in the dictionary: again the compilers tried to find different meanings in all

8 I checked this in 1997; speakers insisted on a difference between kat and ngat.

cases. They succeeded only in some, for example deiblob- (with deib- as the first stem) “to put some­ thing negligently, to mislay”. The composition with these verbs as the second stem has often been grammaticalized already. It serves to indicate aspect, man­ ner of action, and a higher degree of transitivity. These compounds more properly belong to the chapter on inflexional morphology than in the dic­ tionary. If one removed one or two of each of these compounds, which can be formed with 400 primary verbs, the dictionary would at once be deprived of 400 to 800 entries. There are several reasons against fixing the vocabulary of a small speech community at a number of five to ten thousand words. I estimate the “basic vocabulary” of the Eipo language at 800, the “basic plus 'advanced’ vocabu­ lary” at two to three thousand. The absolute figures would not play such an important part, were it not for the provi­ sion of comparative material. It is much more essential that considerations on the quantity of the vocabulary serve as a means of keeping in mind the social conditions of speaking in small communities. Here also, even if on a "scale” growing smaller and smaller, the vocabulary is compiled of old expressions, idio- and sociolects as well as terminologies and artificial languages; it would be erro­ neous to believe that the entire quantity of words —small as it is in comparison to European languages - would have to form part of the commonly used vocabulary. The most remarkable feature is probably the high fluctuation: new words come up as quickly as in larger speech commu­ nities, each tale and each song uses three to five new words, according to my estimation, and at least as many new idioms. But the means of “registering”, the chances that something new is set up as a norm and is handed down over a long period and in many groups, are a lot smaller because of the absence of institutions. Although only the grammar can provide a justification for the manner in which the vocabulary is allotted to the dif­ ferent word classes, I shall here anticipate some of the arguments, because this leads over to the characterization of the dictionary with regard to content. Of the 5600 words 2207 are nouns and 2600 verbs. Of the nouns roughly one third, i.e. 863, are place-names and proper names as well as names of plants and animals. Of the verbs 431 are primary ones, i.e. they are not compounds or at least not recognizable as compounds. Of the com­ pounds there are many with bin- “ der Linguist, der eine konkrete Sprache beschreibt, sogar dann, wenn er keine deutliche Vorstellung von einer typologischen Untersuchung hat, sich auf seine typologische Erfahrung stiitzt, die natiirlich auch eine Form von Typologisierung ist.” (Uspenskij Sc Zivov 1976: 140-141)12

6. Most grammars following one of the models of generative transformational grammar were pub­ lished in the seventies. In the eighties field research­ ers rather joined the empirically-led discussion about universals and typology (cf. the work of Derbyshire 1985, who makes use of the studies on universals of wordorder, and W Seiler 1985, who aims at writing a theory-neutral grammar and who, when considering subject- and case-marking in the Imonda language, with the help of results from other grammars arrives at generalizations and well-founded conclusions of his own; cf. chapter 2.1).

The statement about the excessive explicitness of many “exotic” languages may serve as a typical example. Foley, at the end of his chapter “Verbal semantics in Papuan lan­ guages” (Foley 1986:128), touches the subject (taking up the tenor of so many anecdotal accounts):

7. The so-called “reference grammars” were written with the aim of avoiding any kind of discussion on theory: theory-neutral representations with an inven­ tory of merely descriptive terms, sharpened through the description of many languages and through the study of typology (cf. for instance Sohn 1975 or the various grammars in Dixon & Blake [eds.] 1979— 1991). Seen from the point of view of the history of science, tagmemics take over the role of the “refer­ ence grammars” in the description of Papuan lan­ guages; field researchers adopt the descriptive terms of tagmemics without claiming to promote the dis­ cussion on theory. The investigation of Papuan lan­ guages has yielded one single theoretical study (Bee 1973c).

Of course, such features only attract attention if one pro­ ceeds from one’s mother tongue or from one’s own, almost pre-scientific typological experience. This, how­ ever, does not reveal the achievement and the function of that explicitness. I should say that the fact of this verbal construction exists prior to all theory, it is only a matter of finding it, and this finding is linked to the primary process of translating from the foreign language into one’s own; I shall return to this in the following chapter. Theories do not create this fact, but they make it comparable with other facts and gradually release the process of discovering and translating from the accidentalness of “typological experience”. They are methods of objectivizing or of the construction of data. Field research should not and can­ not deny the “Stadium der ersten Verbliifftheit”13 (Gabe-

On the one hand, the discussion on theory in modern lin­ guistics compels the field researcher to deal carefully with the data, a process also contributing to his experience; on the other hand, this discussion should be a downright shock to him, because what for him are facts or his data does not always count in the light of scientific theory. One does not have to go so far as Itkonen (quoted by Finke 1979:58), who altogether doubts the status of lin­ guistics as an empirical science, but a naive conception of facts is exposed to a multitude of attacks. Only theory establishes what is to be considered as grammatical data,

“A final point concerns the concreteness and explicit­ ness of verbal descriptions in Papuan in comparison to English and other European languages. This is most apparent in the serial verb constructions, in which, for example, 'bring5 is rendered by take come, 'feel5by holdperceive, or 'tumble5by come descend hitP

12 “[...] when compiling concrete descriptions o f languages the lin­ guist in principle only takes down what seems to be important to him from the point o f view o f typology. Accordingly it is con­ ceivable that the linguist who describes a concrete language relies on his typological experience, which is also a kind o f typological system, even if he has no clear conception o f a typological inves­ tigation.” 13 “phase o f first bewilderment”. Such a phase was experienced by Sasse (1988) and Pawley (1993). Sasse worked on the Iroquois Cayuga language and Pawley on the Papuan Kalam language. Both languages are “strange” seen from the point o f view o f our “typological experience”. Sasse was led to postulate an “imma-

39

lentz 1984:83), the first moment of perceiving a strange phenomenon. However, the field researcher also has to be able to be astonished time and again at what theorists do to his data. Dixon’s wish for many “restatements” finds its deeper justification in these phases of research which obviously have to be different ones. The conception of what data are, however, has undergone a change. The field researcher proceeds from what he hears, and writes down, from what he can publish as a text and what can be exem­ plified by texts. In case of doubt the demand for exact knowledge of the language he wants to describe has prior­ ity. “Dieser Aufgabe”, namely that of describing a lan­ guage “ist nur der gewachsen, der die Sprache praktisch beherrscht: das wissenschaftliche Kennen, das Erkennen und Beurteilen setzt ein Konnen voraus; die wissenschaftliche Darstellung wird nichts Anderes sein, als eine sachgemasse Erklarung dieses Konnens [...] Somit wird die Grammatik zur Selbstschilderung, zur Selbstbespiegelung und Selbstanalyse” (Gabelentz 1984: 82)14. Modern linguistics acts as if the subjective factors in the investigation of a language did not exist. Ideally, it obtains data from the intuition or the judgement of the speakers about the correctness of sentences. The theoretical lin­ guist is to be found far away from all typological experi­ ence, but in the midst of intuitive self-contemplation about sentences of his own language. Grammar is today the competent (“sachgemasse”) representation of the compe­ tence (“Konnen”) of one’s own language. The native speakers, who surround the field researcher, are not men­ tioned in linguistic theory (cf. various contributions in Coulmas 1981). A modern grammar is not supposed to consist in the objectification of what the field researcher learns from the native speakers and of what, in certain situ­ ations, is an adequate utterance in a speech community, but in the self analysis of those to whom the grammarian is listening. This, as field researchers admit, is a hopeless undertaking. In dealing with real native speakers one quickly finds out that the ability to give one’s opinion on sentences is an individual faculty and is closely connected to the speaker’s education. The native speaker as a crea­ tion of modern linguistics turns out to be the result of centuries of reflection about language and as a partner in the established discipline of linguistics (cf. Heeschen 1984a and 1990c). Field research has no use for the con­ ception of data of modern linguistics. Nevertheless it is legitimate to investigate not only the subjective process of discovery on the part of the gram­ marian, but also the manner in which informants —native speakers —join in the language and question games of the field researcher, and in which discoveries are made by those to whom the questions are put. In doing this the aim is not to teach the informants what the linguist is looking for (cf. Hale 1965, Kibrik 1977:14, Wong 1975:50); it is essential to find out, rather, what the speak­ er works at in his language, which are the problems his language confronts him with, and where he searches for innovation, improvement and more precise ways of expression. For each utterance is a compromise between 40

what one has in mind to express and what has actually been shaped by means of language. This is, to begin with, a separate field of investigation; it would be the analysis of the pre-scientific metalinguistic awareness and its func­ tions (Heger 1970), and the data acquired by the field researcher would be “secondary reflections of primary performance” (Levelt 1972:22). The field of investigation would be the counterpart of the typological experience of the grammarian, except that the native speaker does not want to gain a standpoint outside his language, but one, as it were, above his language, suited for solving the prob­ lems of planning and controlling utterances. The idea of linguistic autonomy has been called in ques­ tion by field research in many ways: 1. Necessity and status of theories when describing individual languages are quite uncertain. 2. Linguistic theory and field research have devel­ oped different conceptions of what should belong to the field of data. If the field researcher still makes use of examples and texts, he must have tried to interpret them: this means, however, that he has to turn his special attention to the ethnological milieu. 3. The subjective point of view of the person de­ scribing a language has hardly ever been considered. The typological experience of the field researcher penetrates into each act of searching for, and the dis­ covery of, data. The “phase of first amazement” guides the acquisition of data from the very beginning. 4. The degree of mastery of the foreign language has never been taken into consideration. The self analysis resulting from guided acquisition of a second lan­ guage, however, helps to distinguish between theoryneutral and theory-dependent statements. Statements are transformed into results of processes of learning; they lose their validity in a theory of language which is not at the same time a theory of speaking, because the grammarian in the capacity of a learner will be able to say something about “usus” (Paul), but not about competence. 5. If field research gives up the idea of linguistic autonomy and associates with psycholinguistics, the subjective relation of the native speakers to their lan­ guage becomes a separate object of research: not the forms of utterances, but the laws of their formation come into focus. It is essential to find out those

nent-typological analysis” in order to do justice to the peculiar features o f Cayuga, on the one hand, and to make use o f modern typology and universal grammar, on the other hand. Pawley was led to rethinking the relationships between idiomatic competence, the lexicon, and formal or “autonomous” syntax. Cf. also chap­ ters 3.34,3.3.2, and 8.3.2.3. 14 “Only he who masters the language in practice is equal to this task”, namely that o f describing a language: “scientific knowl­ edge, cognition and judgement presuppose competency; the scientific representation will be nothing but a correct explanation o f this competency. [...] thus grammar becomes a description, contemplation and analysis o f one’s own self.”

layers or structures in language which are accessible to judgement and which may be deliberately planned. Finke (1979:59—61) very sensibly distinguishes between observation and experience. If an ethnologist or a mis­ sionary, who, let us say, have no special linguistic training, or if a linguist without special theoretical interest make their observations in the field and without great difficul­ ties classify this form as a substantive or that as a verb, it is easy to understand that observations also depend on theo­ ry. Terms like “substantive” or “verb” make sense only within the statements made by linguistics, but in addition to that there must also be observations without the for­ mation of linguistic terms. In ontogenesis as well as in the unguided acquisition of a second language, promoted and demanded by the sociolinguistic situation in New Guinea, the learner observes, categorizes, and generalizes. The “theory” directing these activities would have to coincide with the manner in which man comprehends the world, processes the sensory data (which are already categorically pre-shaped) and organizes his experiences in chains of signs. If we understand field research as a conversation, here the researcher making an effort at self-analysis, there the informant, who is able to reach a meta-linguistic standpoint, the field researcher has the opportunity to examine the schemata which exist prior to all dependence on linguistic theories and which influence observations of a linguistic nature. The question of whether the observa­ tions necessarily depend upon a more or less explicit lin­ guistic theory will here be left unanswered. The observations are distinguished by Finke (1979:61) from the experiences: “Erfahrungen sind theoretisch organisierte Systeme von Beobachtungen.”15 If field researchers insist upon the value of their data and leave the “restatements” to the theorists, the difference between observation and experience is certainly at the root of the empirical attitude. Field researchers are close to observation and the theoretical organization of the observations may be conventional, at least without any ambition for deduction and formalization. The “restaters” attach great importance to correct theoretical organization of the data. The theorists restrict the visual angle to the organizability of the data (and may for some moments indulge in the dream of linguistic autonomy). The field worker witnesses how his data are assembled and on which non-linguistic factors their understanding, their interpretation, and their organizability depend. These fac­ tors are: culture and environment of the community whose language is being investigated, the “typological experience” of the linguist, the informants’ ability to learn and their pre-scientific metalinguistic knowledge, the dia­ logue between researcher and informants, and, finally, “coagulated” into examples and texts, the utterances, which unperturbed by any idea of linguistic autonomy mirror the culture and the environment of the speakers. Field research as an empirical science does not begin with linguistic data, but with exactly the same matter which is studied by philologists. I do not believe that this empiri­

cal basis of field research merely belongs into a chapter entitled “discovery procedures”, but that it is the basis for any construction of data and is systematically part of linguistics, insofar as it intends to and must reveal the theoretical organization of observations. 3.2 Philology The reader who has followed the literature since 1957 and who has even cast a glance into this or that grammar of an “exotic” language, will be surprised to find the key-word “philology” in a chapter which tries to define field research in its relation to linguistic theory. But the differ­ ence between an autonomous linguistic science and one that regards language as embedded in culture, sociology, and into the principles of speaking reappears immediately. The philologist enlarges all the questions which the field researcher or the descriptive linguist, interested in ethno­ logical problems, does not exclude from his examinations by the dimension of history; the field researcher elicits the meaning of speech acts just died away, which he neverthe­ less has to present in writing, be it as examples in his grammar, be it as an independent compilation of texts; the philologist interprets old texts, the language of which has either completely perished or, as an earlier stage in the development of the language, is no longer understood without commentary. It is his task to restore the cultural and social environment so that what appears as a work removed from time can again be understood as a “speech act” (regarding the key-word philology in general cf. Horstmann 1989). In the opinion of Anttila (1973:177) linguists in contrast to philologists occupy an extreme position, insofar as they “die Ansicht vertreten, daB man alle Zusammenhange von Sprache mit Nicht-Sprachlichem ausklammern konne”16. This is exactly what the philologist does not do, “denn er studiert die Sprache in der Gesamtheit ihrer kulturellen Beziige und bezieht dabei alle moglichen Verbindungen zur Lebensweise der Menschen ein (Brauche, Kunst- und Gebrauchsgegenstande usw.). Er benutzt also Sprache und Linguistik in Verbindung mit Geschichte, Archaologie, Volkskunde, Religionswissenschaften usw.”17 The attempt to establish linguistics as an autonomous sci­ ence with an object exclusively its own is at the same time an attempt to free oneself from the traditional connection with philology and to cast aside the links connecting it with the auxiliary disciplines of philology (Jager 1987:198—200, cf. C. Heeschen 1972). The contrary pro-

15 “Experiences are theoretically organized systems o f observa­ tions.” 16 “[...] hold the view that all connections between language and non-linguistic matters can be disregarded.” 17 [...] for he studies language in the totality o f its cultural referenc­ es including all possible relations to the way o f life o f the people (customs, objects o f art and o f daily use etc.). That is, he uses lan­ guage and linguistics in connection with history, archaeology, folklore, science o f comparative religion etc.”

41

cess, namely the interweaving of an autonomous linguistic science with other disciplines, took place during the last three decades without taking philology into consideration. “Was Philologen zu alien Zeiten getan haben, dringt erst jetzt als eine theoretische Notwendigkeit auch in die synchronische Linguistik ein. Durch Soziolinguistik, anthropologische Linguistik, Ethnolinguistik und Pschycholinguistik ist der kulturelle Bereich Teil der Sprachtheorie geworden. In anderen Worten, diese Disziplinen sind synchronische Philologie.” (Anttila 1973:184-185)18 Although Ehlich (1981) along with Anttila and Jager attacks the idea of linguistic autonomy with the help of the insights of philology, he defines the differences between field research, philology and autonomous linguistics more exactly. The field researcher who starts to work at a lan­ guage unknown to him uses the classical type of “native speaker”: “This type of native speaker bridges the gap to an alien world [...] he opens the doors to the linguist’s destination, the treasuries of unknown languages and cultures” (Ehlich 1981:154). The second type of “native speaker”, according to Ehlich, is the linguist himself: he himself answers for “the funda­ mental capacities and competences that are in play when people talk”. The philologist cannot call on either of the two useful helpers for assistance. He has to deal with “de­ tached texts”, i.e. with texts for whose understanding he is not aided by the momentary situation in which speaking takes place or by the present cultural setting. Ehlich de­ fines the philologist as a secondary hearer, that is a reader: “We may say, then, that the philologist is a contingent hearer/reader of a text, the term 'contingent’ being taken in its exact philosophical sense.” (Ehlich 1981:159) In order to overcome the distance between himself and the text the philologist proceeds in two different ways: in the way of the field researcher who undertakes to elicit the meaning of a speech act with the help of other disci­ plines, and in that of the linguistic “native speaker”, who proceeds introspectively and can rely on his competence, the difference being that the philologist has to take great pains to build up his competence with the help of those other auxiliary disciplines. In the practice of research the differences between the philologist and the field worker are obscured. On the one hand, the real “native speaker” cannot explain and clarify all the things that interest the linguist, so that even in the foreign speech community he continues to depend on assumptions and intuitions. On the other hand, a lot of work has to be done by the field researcher far away from all the informants; not only does he have to construe a grammar out of the chaos of the notes taken by him, but also the translation and interpretation of the texts he has gathered is often not different from the work of the phi­ lologist; in the case of the philologist the temporal gap is wide, in the case of the field researcher it is only small. 42

But the field worker as well as the philologist finds help by looking into a botanical work, into books on comparative religion, or into studies about kinship systems. At home he is assisted by an arsenal of libraries, in the field by the questions he asks the informants in ever varying form and growing despair, and often he does not even understand the answers. Whenever the author was asked after his periods of field work which of the linguistic theories had been most useful to him, he could only point to the uni­ versity courses on Middle High German or on the Iguvine Tablets. All the authors quoted or mentioned here agree that there is no direct approach to the data, whether they are texts from past times or utterances by speakers of “exotic” lan­ guages. Such an approach would only exist if the differ­ ence between the general laws of cognition, which at the same time are considered as a prerequisite for each act of cognition, and the acts of cognition and the utterances, which are historically conditioned and which differ from one speech community and from one individual to the next, were evened out. This, however, would presuppose a timeless point of view, independent of what has hap­ pened to develop as a result of a historical process. Each translation and interpretation of a foreign text or of an utterance, taken down as example, comment, or text can only proceed hypothetically: they presuppose meaning, devise a general conception guiding the comprehension of single sentences and words in order to either confirm or dismiss the presupposed meaning; this process can be repeated again and again and all available aid is made use of. It is a process of approximation which may never end, in fact the hermeneutic cycle. The philological process of understanding, in which sub­ ject and object are not conceived to be of timeless validity, but as a dialogue between speaker and hearer, each of them “caught in the web” of their own subjectivity, was first and most thoroughly thought through by Schleiermacher, the translator of Plato and the interpreter of the Holy Scripture. Cognizing and understanding are insepa­ rable from speech. As early as 1804, —even before Wil­ helm von Humboldt’s great works — Schleiermacher points to the identity of speaking and cognizing. All the sciences, except mathematics and the experimental study of nature, depend on language. Cognizing and under­ standing, however, are not only bound to single languages or, in Schleiermacher’s words, to certain “Sprachkreise”, but in particular cases to certain epochs or even only to single speech acts during the life of a single individual or author. The problem of understanding not only becomes salient when speaker and hearer, who speak different lan­ guages and who may even belong to different ages, want to understand each other, but even between two individu­ als speaking the same language and belonging to the same culture. Every understanding is an act in which differences are evened out. Seen from the ability of humans to speak

18 “What philologists have always done is only now penetrating into synchronic linguistics as a theoretical necessity. Through sociolin­ guistics, anthropological linguistics, ethnolinguistics and psycho­ linguistics the cultural sphere has become part o f linguistic ihen­ ry. In other words, these disciplines are synchronic philology.”

to, and understand, his fellows this restrictedness of the initial position, the unalterable difference, is no absolute obstacle. Thinking and understanding proceed from con­ crete, ordinary events and coincidences and remain insep­ arable from speech; pure thinking and the advanced acts of understanding only gradually liberate themselves from being involved in what has accidentally developed and from the subjectivity of speaker and hearer or author and reader. The irrationality of linguistic differences, however, cannot be abolished by logical constructions or rationalized bv making the plurality of languages itself the subject of a theory of cognition. In the same way, the difference between two individuals cannot be evened out by what they have in common through culture and language. The annihi­ lation of differences, indeed, is inherent in the conversing of humans, in fact, in dialectics; but this annihilation can only be conceived as a historical process and as a neverending approximation. In the end Schleiermacher has to return to an ontological principle embracing all mankind in order to be able to set up a concept of truth. Leaving aside the special epistemological problems, one can claim for philology that it is the art and the science of annihilating the difference between a text and its reader or between speaker and hearer. For this process of annihilation Ehlich coined the very fitting expression “de-alienation”. It can be applied to the fieldworker as well as to the philologist. In view of the tasks of the field researcher I would like to point out three things from the general course of Schleiermacher?s considerations, from the annihilation of differences or from “de-alienation”. 1. The cyclic nature of knowledge, beginning with the understanding of a foreign utterance, growing more certain through the reconstruction of the gen­ eral setting and returning to that utterance with greater certainty, is repeated in the work of the descriptive grammarian. Each act of speaking is, first of all, individual, appropriate for certain situations, and embedded in the special methods of achieving understanding within a culture. In Schleiermacher’s words (1977:78): “Objektivieren wir uns die Sprache, so finden wir, daB alle Akte des Redens nur eine Art sind, wie die Sprache in ihrer eigentiimlichen Natur zum Vorschein kommt, und jeder Einzelne nur ein Ort ist, in dem die Sprache erscheint, wie wir denn bei bedeutenden Schriftstellern unsere Aufmerksamkeit auf ihre Sprache richten und bei ihnen eine Verschiedenheit des Stils sehen. - Ebenso ist jede Rede immer nur zu verstehen aus dem ganzen Leben, dem sie angehort, d.h., da jede Rede nur als Lebensmoment des Redenden in der Bedingtheit aller seiner Lebensmomente erkennbar ist, und dies nur aus der Gesamtheit seiner Umgebungen, wodurch seine Entwicklung und sein Fortbestehen bestimmt werden, so ist jeder Redende nur verstehbar durch seine Nationalitat und sein Zeitalter.”19 The field researcher has to remain aware of the fact that all utterances are only such life-moments. This momentary character of speech developing into lan­ guage is discernible in what pragma-, ethno- and

psycholinguistics undertake to do. In the author’s observations concerning the individuality of Eipo speakers and concerning the vocabulary of single groups or speakers such life-moments are recorded. In the speech community of the Yalenang I wit­ nessed great differences of oratorial style; in a single paragraph-like sentence reminding one of Thucy­ dides, a big-man could state the most complicated facts, allude at the same time to myths, characterize the matter represented by him either as the opinion of others or of his own and ironically comment or anticipate the audience’s reactions. Nothing of this mastery is to be found in the children’s tales, nothing in most fairy tales told by women, nor in the tales of war recounted by young men. Senft (1990) calls attention to the elaborate, tradition-conscious style of speech of members of the ancient clans of the Trobriand Islands; in contrast to these, speakers of more recent migration waves adopt new develop­ ments. The language appears differently in each speaker and the field researcher would have to be sure of a whole series of individual grammars be­ fore levelling out a grammar from utterances and speech acts. In philology, indeed, more than in lin­ guistics, such individual grammars are known (cf. Spitzer 1961). Studies of style are of equal impor­ tance to linguistics, “denn bekanntlich ist die Allgemeinsprache nichts als ein Durchschnitt von Individualsprachen, die Grammatikalisierung verschiedener Sprechakte [...] nihil est in syntaxi quod non fuerit in stylo. Syntax, ja Grammatik sind nichts als gefrorene Stilistik [...]” (Spitzer 1961: II, 516-517).20 Philologists looking into the life-moments of a lan­ guage have discovered in the so-called vernacular languages what linguists would like to restrict to “exotic” languages. The tendency to distribute two noun phrases to two clauses is a typical feature of colloquial speech. “We’ll go after sunset” becomes “we’ll go when the sun’s gone down”. In vernacular languages abstracts are avoided; instead of “spring” people say “when the trees are budding” or “when the cuckoo calls” (Maurer 1933). The splitting of a verb like “bring” into “take” and “come” and, gener-

19 “When regarding language as an object we find that all acts o f speaking are only one manner in which language comes to light in its proper nature and that each individual is only a place in which language appears, just as when reading eminent authors we turn our attention to their language and find out their differences o f style. In the same way each utterance can only be understood from the totality o f the life to which it belongs, i.e. as each utter­ ance is perceivable only as the speaker’s moment o f life, with the same contingency as all moments o f his life, and this only from the totality o f his surroundings, which determine his develop­ ment and his further progress, thus each speaker can be under­ stood only through his nationality and the age he lives in.” 20 “[...] for, as is well-known, common language is nothing but the mean average o f individual languages, the grammaticalization of dif­ ferent speech acts [...] nihil est in syntaxi quod non fuerit in stylo. Syntax, even grammar, is nothing but the frozen art of style [..

43

ally speaking, the excessive precision of some speech communities (cf. chapter 3.3.1) is called enumerative way of speaking by Havers (1931); it is not only found in a grammaticized form in some few lan­ guages, but it is a characteristic feature of vernacular languages as well as of “primitive” languages, as they were then called. This is also a point suggesting that grammar is nothing but a mean average of individu­ al languages, and it is even possible that this average has to be found not only from a multitude of idio­ lects of a greater speech community, but from similar styles which are encountered in different languages. Just like the language of poets, which was investi­ gated by Spitzer, vernacular languages have a style of their own, which may turn into grammar. Bloom­ field (1927) must also have been aware of such calculation of a mean average from the sum of indi­ vidual styles and life-moments of a language when examining the linguistic faculties of his Menomini informants. One of them loved archaic con­ structions, the other spoke a rich Menomini abound­ ing in idiomatic expressions, and a third had a poor vocabulary and spoke in ill-formed sentences as soon as more than everyday conversation was demanded. In order to differentiate between good and bad Menomini the grammarian would have to comment on nearly every grammatical construction and very many words of the lexicon. From this and from the above mentioned examples we can learn what a long way it is from the life-moments of a lan­ guage to generalizations in the form of a grammar.

word-lists and dictionaries, are seldom presented to the scientific public. Schaeder (1987:48—52) remarks upon the low status of the compilation of dictionar­ ies and the theory of lexicography within the differ­ ent linguistic theories. The collection of texts seems to be of equally low status. Hardly any of the recent­ ly published grammars of Papuan languages has more than only “sample texts” (a small survey con­ cerning this problem in Heeschen 1990b: 10—11), With the absence of text-collections the art of elicit­ ing the facts is interrupted, the linguist conceals his philological work, and utterances and speech imme­ diately become data in a grammar. There are, indeed, enough text theories, but, of course, they say noth­ ing about the integration of the text-collector’s work into the grammatical description of a language ac­ cording to traditional standards. That it is necessary to think about this integration is sometimes shown in controversies. Senft (1986: 132—53) published three texts in his Kilivila gram­ mar. Lithgow remarks about this (1988:148): “I wonder if the three texts given at the end of the grammar section would be accepted by Kiriwina speakers as Biga Bwena (“Good Language”). In real life situations you can communicate your meaning with the key words of a sentence, letting the hearer fill in the grammatical gaps. I suspect that those who know their language well would want to edit or expand these texts before presentation in book. We would not present fragmentary, disjointed utterances as examples of our language.”

2. Given to the field researcher are the utterances of the informants. The linguist turns them into exam­ ples and texts. Disregarding the fact that even texts are full of individual characteristics (cf. under 1), we find that the texts are a representation of utterances (H. Seiler 1970:92), that is, abstractions from the lifemoments of a language; further steps of abstraction follow. Seiler (1970:92) comprehends the production of texts as the first step of a

If we assume the texts really to be “fragmentary”, a fact that would have to be examined, Lithgow’s posi­ tion, in view of what has been said in this chapter, is absurd. All philological questions remain unan­ swered. The question of how far the “Representation von Gesprochenem” (representation of spoken lan­ guage) may be emended remains open. Up to now it has been the generally accepted practice to correct only evident mistakes and to leave out, of course, laughing and clearing one’s throat. Lithgow’s criti­ cism presupposes data manipulated according to un­ clear guidelines, not to say sentences from a manual which the speakers of Kilivila themselves do not understand. That the speakers’ way of talking is “fragmentary” is a very interesting discovery for the philologist, and it is equally interesting to find out that they are able to edit and emend their “texts”. Unfortunately these discoveries, which are them­ selves worth looking into, would be swept under the carpet, if the way from the utterance to the grammat­ ical presentation could not be reconstructed. What follows from this criticism is that collections of texts are a step in the construction of data. They must not mirror anything but the life-moments of a language.

“linguistischen Beschreibung einer Sprache [...]. Als einen weiteren Teil der linguistischen Beschreibungsarbeit sehe ich die Ubersetzung solcher Texte in eine dem wissenschaftlichen Publikum zugangliche Sprache an. Dazu gesellt sich das, was eine nach generativen Prinzipien angefertigte Grammatik fiber die Struktur einer Sprache auszusagen hat.”21 If one adds that this step-by-step procedure must not be thought of as separate phases in the study of a language, but that from the utterance via its repre­ sentation and translation one has to deduct continously by anticipation and reversion, one would have to agree with Seiler unreservedly. The single steps depend upon each other, and grammar, in the hermeneutic cycle, has again and again to make sure of its first data in order to avoid being regarded as an abstraction of abstractions. All the same, Seiler’s conception of a linguist’s task is too comprehensive compared with the practice of research. Even the preconditions of translating,

21 “[• ••] linguistic description o f a language [...]. As a further step in the task o f linguistic description I regard the translation o f such texts into a language accessible to the scientific public. To this is added what a grammar written according to generative principles has to say about the structure of a language.”

3. In Seiler’s three tasks of linguistic description, translation occupies the middle position. One has again to return to Schleiermacher in order to gain an impression of the unity of philological and grammatical work. Each word and each sentence which the field researcher refers to in his descrip­ tions is implicitly the result of the art of translating. The field researcher, whose aim it is to pin down his results, is bound to the process of understand­ ing foreign life-moments, referring to foreign envir­ onments and strange “belief systems”, that is to the never-ending process of translating and ad­ equation. This part of the construction of data, which is also a transition from observation to experi­ ence, is hardly taken notice of by linguists. It is philosophers like Quine (1960) or linguistic-philo­ logical “outsiders” like Wandruszka (1984) who, from this part of the task of description, doubt the validity of abstract grammatical statements (cf. ref­ erences in Lewandowski 1980:111, 1022—1023, Gar­ cia et al. 1987). Ethnologists like Needham (1972) and Idallpike (1979) have presented contributions full of data valuable to the linguist; however, they are less interested in the connection between field­ work and translation than in considering data in the light of possible similarities between speakers of different languages. On the side of the science of translation, opinions range from the fundamental impossibility of trans­ lating to the belief that at least the sense of a text can be adequately rendered. However that may be, one is again caught in the hermeneutic cycle: in order to understand the contents the translator has to know the meaning of single words, phrases and sentences, in order to understand these he tries to guess at and outline the basic idea of the whole (Schleiermacher’s “Divination”). Quite in accor­ dance with Schleiermacher, there remains a differ­ ence, according to Quine (1960) there is a funda­ mental vagueness between original and translation; if description is conceived of as a “translation” of observation into experience, the difference is also to be found, but here it is the difference between the language “observed” and the meta-language (cf. Schnelle 1973). These general considerations can be completed by remarks about definition and paraphrase. The field researcher, in the course of his work, has to rely upon the explanation of new words and constructions through what is already known to him. These explanations yield data of a strange double nature. They are original meta­ language as well as new primary data. This double charac­ ter refers to the hermeneutic cycle in finding the data. Understanding does not begin with translation but with the reflex of the primary data in definition and para­ phrase. Here are two examples: The Yalenang love to answer a question about the meaning of a word in a cer­ tain way; when I, for instance, asked for yin “bow”, the answer could be something like this:

■ (28, Yale) Mana ol-bi-dob-ne sembe le-dob. cuscus

hunt-go-inf.-n.

for

speak-inf.

This is something you say when you go hunting mar­ supials. The Yalenang’s reasoning runs along the line of the func­ tions of things. The knowledge of this way of thinking is important to the field researcher. Definitions of this kind, on the one hand, are primary data, if this talking about linguistic matters is regarded in the light of the ethnogra­ phy of speaking or if the field researcher investigates the speakers’ faculties of classification from the psychological point of view. On the other hand, the definitions teach the field researcher to proceed with great care when evalu­ ating synonyms (or also antonyms and appellatives) in further attempts at understanding. The Yale speakers place together yin “bow” and mana “marsupial” or yubu “speech” and ok “laughing” and explain one through the other. The first two words belong to the functional com­ plex of hunting, the other two to that of conversation: first people talk, then they laugh in a friendly manner. The “bow” is not defined as a tool and placed beside other artefacts, “speech” is not placed beside other human vocal utterings like “screaming, singing or calling”. Definition and paraphrase are connected to ways of understanding specific to the culture, and the researcher discovers the general meaning only from the continuous interplay be­ tween primary utterances and meta-language. The second example leads us to the Eipo language. (29, cf. chapters 7.1.4.4, 8.3.1, 8.3.2.1) Betengde arye mape fetereb-ma-1. Betengde

s.

boy/pl.

explain-dur.-3sg./pres.

Betengde explains it to his boys. When asked for the meaning of arye, the informants pre­ sented a second sentence: (30) Betengde, el-ce, el mape fetereb-ma-1. Betengde

he-but

he

boy/pl.

explain-dur.-3sg./pres.

It is Betengde who explains it to his boys. In the second sentence the subject-indicating arye is left out, instead of this we find the subject in the front posi­ tion and a pronoun referring to it. To begin with, the field researcher regards the second sentence as a meta-linguistic assessment of the first sentence. What he takes for “meta-language”, however, is only a paraphrase in the eyes of the informant, both sentences are of equal value, the paraphrase can be either a “bestimmte kanonische Formulierung”22 (Ungeheuer 1969:224) of the semantics of the first sentence or a new sentence formed ad hoc for the purpose of explaining the first sentence (cf. L.R. & H. Gleitman 1970). In order to understand the second sen­ tence the linguist has to revert to the first one. The second sentence as an utterance is a primary datum of observa­ tion as well as the first one, at the same time it is “meta­ language” for the linguist as well as for the “native speak22 “a certain canonical formulation”.

45

I

er”; one also has to bear in mind that the informant mere­ ly aims at explaining the meanings of the sentences to the linguist, not at stylistically well-formed sentences. The lin­ guist, of course, has to consider the pragmatics of the question game as well as the psychology of the infor­ mants. A long, subjective process of data construction precedes the decision as to which meaning should be attributed to a word and what the exact wording of the rule of subject-marking may be. The translation is ulti­ mately influenced by whether there are any culture-specif­ ic semantic clusters of characteristics or whether there are language-specific paraphrase relations between two sen­ tences. Of course the linguist can deduce and expect from an SOV-language like Eipo that the subject of a transitive or ditransitive sentence is marked by a postposition; in this case he probably does not do justice to the second of the above-mentioned sentences and maybe, like Lithgow, he even believes its form to be fragmentary. The philolo­ gist, who looks into the life-moments of a language, holds back his judgement and formulates rules which are to stand the test of the data from translation, definition, par­ aphrase, individual languages, styles of speaking and the process of understanding. The fact that field research is close to philology, the retro­ spective view of Schleiermacher’s hermeneutics, the methods of data construction and the manifold ways of systematizing observation into experience lead to the question of whether the whole of linguistic description, understanding, representation of utterances, translation and the system of general statements, called grammar, can make linguistic theory more transparent or can contribute something to the formation of theory. In other words: can grammar follow the lines of its own process of un­ derstanding and those of mutual understanding with the informants, on the one hand, and can it, on the other hand, represent the differences between observed and descriptive language or between utterances and meta-Ianguage? If these two questions can be answered in the affirmative, this would generally mean that the linguist does not with­ draw too much from those life-moments, namely the speech acts, and that he first, beyond and apart from pre­ set categories and conceptions, listens to the pragmatics of understanding and speech. This means in particular: 1. Understanding is not simply brought about by the fact that two speakers belong to the same speech community. Understanding is always the result of a process of understanding; it is a general process, which on the one hand makes use of means specific to language and on the other hand forms these means. For this reason grammar must represent speech as a process of increasing clarification, it is a strategy of achieving understanding but it does not guarantee understanding. It takes into consideration “die standige Wiederherstellung einer strukturell gestorten und auch immer prekar bleibenden Verstandigung”23 (Jager 1987:203). 2. Within the process of increasing clarification and within the strategies of understanding, grammar provides heuristic means. The concepts it uses for the purpose of description should be non-essentia46

listic and, as far as possible, free of hypotheses. They should reflect the supposition that grammar itself is in a continual process of development, as it is always merely the momentary result of individual speaking and has to stand the test of permanence in success­ ful understanding. Further, they should mirror something of the incompleteness suggested by the unalterable difference between speaker and hearer, between text and reader, and between observed and described language. 3. With regard to the speakers of the language under investigation their work on the language should be documented: in individuality of speech, in style, in definitions, paraphrases, and in the fields of grammar which are either open to intentional manipulation or can be made accessible to conscious judgement in meta-linguistic processes of under­ standing. 4. The field researcher should also consider the construction of his knowledge, Gabelentz’s “Selbstbespiegelung” or Ehlich’s “post-native” competence. This competence will, however, prove itself less in the abstractions of grammar than in acts of under­ standing, of translation of utterances and texts. The consciousness of the difference between one’s own “typological experience” or the meta-language used by the researcher, on the one hand, and the observa­ tions, on the other hand, is also part of this compe­ tence. 33

“Injunctive” Grammar

33.1 Observations The Yaghan tongue - and by inference all language - proceeds as a system of navigation. Named things are fixed points, aligned or compared, which allow the speaker to plot the next move. Bruce Chatwin

The following considerations are meant to be theory inso­ far as they describe and systematize the first observations in a foreign language- or speech-community: perhaps they are only attempts at neutralizing the “cognitive disso­ nance”, which arose when the observations were judged by current and accepted models and ways of talking about language. The dissonance consisted in two things: 1. The question I was often asked coming home after a field trip, that of how communication with speakers of a language of which one did not know anything beforehand comes about - a situation of which many introductions into linguistics dream could only be answered with hints and with descrip­ tions of my life there. At first, communication took place without language. The arrival, the erection of a 23 “[...] the continual re-establishment o f a communication, which is structurally disturbed and always remains precarious.”

; ;

tent or a hut, gift and return gift, the walk through the village, sitting down in the men’s house, taking part in meals, and warding off importunity and greed were embedded into what was self-evident: into laughing, anger, expectation, uneasiness and fear. In terms of ethology: communication was non­ verbal and took place in the field of tension between approximation and flight, between agonal and synagonal acting and behaviour. It was as if co-opera­ tion in daily life and non-verbal communication in a vis-a-vis situation suspended the necessity of an additional understanding by verbal means. When language was added, the scene, the actions and the actors were known already. Exposition and plan could be named, but my memory cannot separate the two different phases, so that language keeps appearing only as a transcending of the moment and as a naming of what had long been known. 2. The second kind of dissonance can be explained with the help of a short text, which is given here in an abbreviated version: (31) Bongto dob-uk. Bongto

Babyal lukunde do-be-(y)uk.

take-3sg./pa.III Babyal

night

take-go-3sg./pa.III

Yakumanye kilape

Babyal do-be-(y)uk.

Yakumanye

Babyal

woman/pl.

take-go-3sg./pa.III

Lukenyan kalik-il-bin-m-uk.

Kalik-le

night

sneak-inf./sim.

sneak-der.-go-dur.-3sg./pa.III

ya-(yuk). come-3sg./pa.III

Berek-uk

dil-kalik-il-be-(y)uk. Lape asik dam

dawn-3sg./pa.III

see-sneak-der.-go-3sg./pa.III

Lape

bin-m-uk-une,

Barwe fu

kona

go-dur.-3sg./pa.III-sc.

Barwe

top

headwater

hamlet close

mar-ya-(y)uk. wound-come-3sg./pa.III

Ketinge noitam mar-ya-(y)uk,

or-asik,

sun

across-hamlet

middle

wound-come-3sg./pa.III

Lape asik

dam. Or-asik

dam

ya-(y)uk.

Lape

close

close

come-3sg./pa.III

hamlet

across-hamlet

Ali bisik ol-dan-m-ik-ine,

Fenimde

Ali

Fenimde

way

fight-refl.-dur.-3pl./pa.III-sc.

marab-uk.

Fenimde-ak deikye noitam

wound-3sg./pa.III

Fenimde-at

crown

middle

tinge marab-uk. cuirass

wound-3sg./pa.III

He abducted Bongto. Babyal abducted and fetched her at night. Babyal abducted and fetched the wife of Yakumanye. At night he sneaked up. He came and sneaked up. In the morning he looked around, and sneaked up. He walked close to the Lape hamlet, and there he came from the mountain top above the headwater region of the Barwe river in order to wound and to fight. At noon he came in order to wound and to fight, that was near the hamlet over

there, near the Lape hamlet. He came from the vicinity of the hamlet over there. At the Ali path they quarrelled and fought, and there he wounded Fenimde. He wounded Fenimde right through the upper part of the cuirass. The style of this narration is remarkable insofar as it does not sum up the whole action in one period and as it con­ tains short sentences instead of connecting them with con­ junctions; the sentences are only connected by -ine and -une. The style resembles that of written narrations, in which the Eipo also change periods into short sentences.24 Even this inconspicuous text presupposes knowledge of some circumstances which no native narrator would explicitly mention. Why Babyal abducted the young, beautiful Bongto and that the abduction of a woman is a reason for war even among normally allied hamlets need not be mentioned. The hearers also know the time when such an abduction usually takes place: in the early morn­ ing, when a woman goes to a place at some distance from the hamlet to relieve herself. If the abduction is noticed, there will be a lot of screaming, the men rush to their weapons and follow the adversary or go to meet him. The way of fighting is always the same: the parties stand fac­ ing each other at a safe distance, they abuse each other, only some of the fighters leave their own ranks, dash towards the enemy, shoot their arrows, and again with­ draw to a safe distance. The one who holds a higher posi­ tion has the advantage. Babyal had to make a considerable detour in order to reach this position. What counts and what is narrated in detail are the hits and wounds. Bongto, Yakumanye and Fenimde belong to the hamlet of Dingerkon, Babyal and his followers to the hamlet of Munggona. Thus the plural in oldanmik is quite understandable. The singular from binmuk-une on is clear, too: Babyal, by his abduction of Bongto, provided the reason for the war; as the originator, he, in place of all his comrades-inarms, is held responsible for all the hits and misses of his party, positively as well as negatively. On the other hand, the verbal forms like dobuk do not reveal whether the pro­ nominal subject contained in the verbal form is feminine or masculine; neither the free pronoun ely nor the forms of the third person singular of the verb differentiate as to gender. The narrator, Terban, is an inhabitant of Dingerkon and tells his tale there, so that all statements of place and all deictic references have to be seen from the place where the story is told, including the deictic verbs “come” and “go”. The suffixes of the personal names reveal the gender, -to only occurs in female names, -de and -yal and -anye only in male names. And finally another peculiarity: in narratives the Eipo like to use different per­ sonal names; these may be names from the time before the initiation or secret names, which they obtained during or after their initiation. Thus, Yakumanye is called Tinteningde in daily life.

24 Cf. texts no 7 and 8 in Heeschen 1990b:48—50. Two versions o f a narrative are to be found there, the first in a written version, the other in oral transmission. Text (31) was narrated by Terban in 1975. At that time no Eipo was able to write. Only in 1978 had the first Eipo been taught writing.

47

It is not surprising that this text appears to be fragmentary to the foreign listener. One has the feeling that one can only grasp few links of a long, hidden chain. Other texts do not even introduce the subject subsequendy, they merely use “over there” and “up there” without defining it more precisely. To find out who took, abducted or wounded whom often requires tiresome inquiries. If one wants to understand, one has to observe and reconstruct. This applies still more to utterances that are not part of a narrative. It is impossible to collect examples for a con­ struction like: ane.

go/away-3sg./pa.I

sc./this/one

It has gone, this one here. if one has not observed that the speaker saw a bird, tried to catch it and that the bird flew off: the interpreter can now rack his brains in order to find out, whether ane is something like a relative particle or whether it implies a second sentence, which might run thus: (33) Gum ob-ne. hit-lsg./pa.I

I have not hit it. In the first case the correct translation would be: “This one, this is the one that has flown off,” in the second case: “It has flown off (it is in the state of having flown off and I did not hit it)”. In the first case ane would really be a kind of relativizing or nominalizing particle, in the second case the clause ending in ane would describe the background of a second and more important state­ ment, which, paradoxically, need not be uttered in the vis-a-vis situation. Perception and the logic of the con­ text of action guide the formation of grammatical cate­ gories. For the sake of completeness I would like to point out that even the above-mentioned, very simple text is not without its problems of translation. Deikye means “crown (of the head), middle, ridge-shaped parts of the body”. Had Fenimde been hit at the crown, the passage would run: (34) Fenimde deikye noitam-ak Fenimde

crown

middle-at.

In that case the mentioning of the cuirass is incompre­ hensible, as the cuirass protects only breast and back, the neck being protected by old netbags, which are wound around it. If deikye referred to tinge, the passage would read: (35) tinge deikye noitam cuirass

crown

middle

right through the (bulging part of the) cuirass. Thus the translator can only guess that the series was formed with a view to greater precision: “He wounded Fenimde, in the middle, at or through the cuirass”. 48

(36) ^ Yakumanye kilape Yakumanye

Babyal do-be-(y)uk.

woman/pl. Babyal

take-go-3sg./pa.III

Babyal abducted the wife of Yakumanye. In the part of the text left out here a regular SOV-sentence occurs:

(32) Bindob-le

not

Among the typological features the linguist could hope to find in the Eipo language as a Papuan language are SOVwordorder and ergativity. In the only sentence containing two complete noun phrases, which are not implicitly taken up again by pronouns, the rule of the wordorder has been violated:

j ; j 1

;

|

(37) Munggona-nang Dingerkon-nang marab-ik. Munggona-people

Dingerkon-people

wound-3pl./pa.III

The people of Munggona wounded the people of Dingerkon. In the text we find a principle of distribution of the noun phrases, instead: explicit subject and direct object are dis­ tributed to different sentences. The first sentence —Bongto dobuk —introduces the direct object, the second —Babyal lukunde dobeyuk —the subject (agent). The third sentence sums up the first two sentences. The following sentence is also to be understood as an example of this principle of distribution: (38) Ali bisik ol-dan-m-ik-ine, Ali

way

Fenimde marab-uk.

fight-refl.-dur.-3pl./pa.III-sc. Fenimde

wound-3sg./pa.III

On the way called Ali they fought against each other, and there he wounded Fenimde. Usually a sentence with two noun phrases, as will be dem­ onstrated later, is divided into such series of clauses. With­ out this distribution over several clauses it is to be com­ prehended as the summary of the preceding sentences. Munggona-nang Dingerkon-nang marabik is such a summary, too. After the enumeration of the days of fighting and of the injuries the narrative goes on as follows: (39) Mab-m-ik-ye,

j berek-uk,

tubto fe-ya-(y)ik.

sleep-dur.-3pl./pa.III-and dawn-3sg./pa.III again

Ol-an-m-ik.

shoot-come-3pl./pa.lll

Tukto ton marab-ik.

fight-come-dur.-3pl./pa.III Tukto

one

wound-3pl./pa.III

Mungona-nang Dingerkon-nang marab-ik. Munggona-people

Dingerkon-people

wound-3pl./ pa.III

The next morning the hostilities continued. They fought. They also wounded Tukto. The Munggona people wounded the Dingerkon people. Tukto is a man from Dingerkon, and from the deictic “come” in feyayik “they came aiming” it is evident that the 3pl. of the verbs refers to the people of Munggona, so that the last sentence, indeed, only mentions known actors. If the last sentence were not a summary, it would probably, but not necessarily read as follows:

j ;

(40) Munggona-nang fe-yan-m-ik-ye-ine, Munggona-people

' (44) Fenimde marab-uk.

shoot-come-dur.-3pl./pa.III-and-sc.

wound-3pl./pa.III

The people of Munggona waged war, and in doing so they wounded people from Dingerkon. The principle of distribution is complemented by that of continuous clarification. Both principles cannot always be stricdy separated. I propose that the term distribution should be used whenever core elements like subject and direct object are distributed to different sentences and that one should speak of clarification whenever peripheral elements or adverb phrases are distributed or whenever the verb itself is clarified. The relation of the first three sentences to each other can be described as one of increasing precision and clarifica­ tion. The second sentence subsequently introduces the subject and states the time, the third mentions whose wife was abducted, because this is important for the formation of parties in the ensuing war. Further on in the text place and time are clarified in a series: (41) Barwe fu Barwe

kona mar-ya-(y)uk.

headwaters top

wound-come-3sg./pa.III

Ketinge sun

noitam mar-ya- (y) uk. middle

wound-come-3sg./pa.III

He came from the mountain top above the headwater region of the Barwe river, in order to wound and to fight. At noon he came in order to wound and fight... What is also typical is the lining up of the verbs and verbal compounds for the purpose of clarification, for example in the first five sentences: (42) dobuk “he took”, from dob- “take”, dobeyuk “he stole or fetched” from dob- “take” and bin- “go”, kalikilbinmuk “he walked creeping, he lurked”, from kalik- “bend, creep up” and bin- “go”, dilkalekilbeyuk “he looked out and crept up”, from dib(ren-) “see”, kalik- “bend, creep up” and bin- “go”. What is typical here is that a simple verb, i. e. one that is not a compound, occurs first and is then taken up again and semantically clarified in the compounds. Also typical is the subsequent clarification of deictic refer­ ences: (43) or-asik,

Lape asik

across-hamlet Lape

hamlet

Fenimde-ak marab-uk.

wound-3sg./pa.III Fenimde-at

wound-3sg./pa.III

He wounded Fenimde, it was Fenimde whom he wounded.

Dingerkon-nang marab-ik. Dingerkon-people

Fenimde

dam close

from the hamlet over there, from (a site) close to the Lape hamlet. In the following series there is another subsequent expla­ nation:

In this example the direct object is additionally marked by the postposition -ak “in, at” in order to avoid any mis­ understanding. The first sentence could also mean “Fenimde wounded (s.o.)”. Other dialects and languages of the Mek language family spoken by the neighbours of the Eipo have been charac­ terized as ergative languages (cf. Louwerse 1988:108-9, cf. chapters 4.2, 7.1.4.4 and 8.3.2.1). In Eipo sentences can read as follows: (45) Babyal arye Bongto dob-uk. Babyal

s.

Bongto

take-3sg./pa.III

Babyal abducted Bongto. Bongto fi

mab-m-uk.

Bongto

sleep-dur.-3sg./pa.III

very

Bongto was asleep. The subject —the term is merely used formally here for the noun phrase agreeing with the person-number suffixes of the verb —is marked as the agent in the transitive sentence, the subject of an intransitive sentence and the direct object (patient) of a transitive sentence are not marked. The short text shows that no subject is marked in the role of the agent and this in spite of the fact that in two sentences the subject as the agent as well as the direct object as the patient have the feature “human” and that case-marking, if not obligatory, should be desirable. Instead, in one case even the direct object is marked (Fenimde-ak), so that from this text one could conclude that Eipo is a nominative lan­ guage, and not an ergative one. 3.3.2 Further examples The second kind of dissonance was caused, to put it nega­ tively, by the strange incompleteness of the sentences, to put it positively, by the principles of distribution and clar­ ification. Before passing from these observations to theo­ retical considerations, more examples must be given. For further examples concerning subject marking and the dis­ tribution of more than one noun phrase over several clauses I partly rely on results already published (Heeschen 1990a and 1994c). The restricted governing faculty of the verb corresponds to the limited means of agreement and of connection: the relation of other parts of the sentence to and among each other is, to a great extent, regulated by semantics, not to say by the objects (cf. Biihler further below). One can apply the concepts of cohesion, which was quoted when regarding contextual units above the level of a sentence, to the sentence itself. The means of creating cohesion within a single sentence or clause are lacking. —But at this point here I would like to stress the fact that such negative characterization of Eipo is actually inadmissible; it is the result of the meta-language used by the linguist and the 49

philologist (cf. chapter 3.2). We will see that, as a result of lacking cohesion within the sentence, the sentence itself becomes a unit open to pragmatical construction and manipulation, which is no longer determined deductively by a theory, but by utterance and speech. Accordingly, the second kind of dissonance will be dissolved by the fact that the indefinite and vague is the determining feature. The context of speech structures the sentence and the sequence of sentences. The advantage of the (for the present) negative characterization is that, in accordance with the remarks about the work of the philologist made further above, the point of view of the observer and his learning of the language as well as the construction of the data do not escape attention. The lacking cohesion and dependence on the situation outside the realm of language or on speech are first revealed in what one could call the telegraphic style of many of the Eipos’ utterances. This style, however, is completely grammatical and in no way less valid than grammatically more complicated utterances; it is to be found in the language of elegant orators as well as in everyday disputations, which, strangely enough, in other respects present the most complex Eipo, more complex even than some of the speeches of the big-men or stories told by eloquent orators playing with the language. The following is an example of this telegraphic style: (46) Sisin-nang name/vn.-people

arye: “Gum biknaryuk, mape bikne-arye-uk s.

not

know/vn.-from-only

kuwul ouk-ye

ate, kilape

childish

be/in£(?)-and

for

kwit

dika

boy/pl.

aik-uk, mane lik,

woman/pl. hut-only

cuscus

dislike

na sek-nam-ne, mape bo-ya-nam-ne,”

initiation beginning I

cut-fut.III-lsg.

boy/pl. carry-come-fut.III-lsg.

winya-lam-ak.

ur-tek-lam-ak.

Kwaning yongone ton, dol

possibly-stand-hab.-3pl./pres.

sweet/potato cook/vn.

and

kwene ton, yo

kwaknine ton walwal.

take/vn.

bundle/vn.

and

wood

and

weeds

ignorance

They pretended they did not know anything about the sharpening of stone-adzes, of the making of gar­ dens and about the cutting of wood, and that is why they seemed to be only standing around. Nor did they know anything about the harvesting of sweet po­ tatoes, of weeding and of shouldering bundles of wood. (At the time of the building of the sacred men’s house all profane activities are tabooed, the men pretend not to know anything about them and leave the necessary daily tasks to the women or to helpers from other hamlets.) What is “missing” here is the fixing of the events to a cer­ tain time. Even the finite form of the verb in teklamak is only an indefinite present. Pronouns, nouns attached to them in an attributive function, so to speak, verbal nouns, as well as nouns likewise attached to these in an attributive manner constitute the whole grammar. The function of this style becomes clear in the following example. (48) Metek me, metek ob

ara, el yupe gum, el

small

t.

child small

be/inf.

he

speech not

he

engene tonok, el mum engen-ma-lye-ora, cry/vn.

unique

he

breast

cry-dur.-3sg./med.-when

mum

ate engen-ma-le tenen, mum

breast

for

dib-lye

cry-dur.-3sg./pres.

think/vn.

breast

dara, wine u-lam-lye-ine,

eat-3sg./med. t.

thus

be-hab.-3sg./med.-sc.

wik kankana big

kib-lye-ora,

ore “tete” a-lam-le.

become-3sg./med.-when

then

tete

between

make-hab.-3sg./pres.

say-hab.-3pl./pres.

The big-men say: “They are ignorant, the boys are still childish and uninitiated, they have only been in the women’s huts, they do not like cuscus (and they do not go hunting), that is why I am going to begin the initiation, I am going to bring the boys (into the men’s house).” (Through the initiation the boys are led away from the world of the mothers and women and are introduced into the men’s house.) The connection between kelape aik-uk, mane lik with the rest of the sentence is completely unmarked. Nothing is said about the fact that the boys live and eat in the women’s huts, about the danger they are believed to be in through contact with the women, about the eating of unconsecrated food and about hunting. The telegraphic style dispenses with other pieces of information we expect to get with each sentence: (4?) Sik-de ya they-but

50

ketene

stone/adze

sharpen/vn.

walwal, wa

wine

ignorance

make/vn.

walwal, yo

werene walwal ate

ignorance

cut/vn.

wood

ignorance

for

garden

The small child, as long as it is still small, does not talk, it only cries, it cries for the breast (of its mother), but when it has been fed, and when this has lasted some time, and when it has grown big (acquired tallness), it says “mummy”. The style adding pronouns, nouns, and verbal nouns in the manner of attributes serves the description of the back­ ground, it presents the fixed points in an unfinished version, as it were; by and by these are defined in more detail and established in time and place and in relation towards the other participants. Reference to time and space coincidentally establishes the definiteness of noun phrases. Lacking cohesion, the term here being used with regard to the relation of the parts constituting a sentence, can be found in many combinations of words: what will become an attribute, an object, an adverb or an adverbial phrase in the translation is not to be identified in isolated and quick­ ly spoken utterances; in slow speech, pauses and intona­ tion hint at the structure of speech, in allegro forms they are omitted. The insufficient situatedness of single utter­ ances described above corresponds to the lacking faculty of connection of single words and parts of the sentence to the whole. The following utterance:

(49) Mok aik bi-na-(a)b. rain

hut

' (54) Saboka teleb ub-lye-ora,

go-fut.III-lpl.

tobacco

can be translated by, either, “there is rain (or: it is raining), let us go into the hut” or, secondly, “on account of the rain (or: as it is raining), let us go into the hut”. The rela­ tion of mok to the rest of the utterance remains uncer­ tain. Systematic uncertainty is to be found in those cases in which two nouns have to be attributed to one verb or in which one does not know whether an “adjective” belongs to a verb, as an adverb, or whether, as an attrib­ ute, it modifies a noun; there is systematic uncertainty, compare: (5°) Ninye asik

be-(y)uk.

man

go-3sg./pa.III

hamlet

The man went to the hamlet. (Or:) He went to a hamlet of humans (and not to a hamlet of spirits). It is uncertain whether we have a compound ninye asik “hamlet of humans” or two noun phrases. Other uncer­ tainties turn up in the subsequent constructions: (51) Aik bope fidina hut

warmth

gum.

take/up/vn. not

Our hut does not take up warmth. (Or:) There is no taking up (in) the warm hut. Saboka teleb dib-ma-1. tobacco

good

eat-dur.-3sg./pres.

He is smoking good tobacco. (Or:) He is smoking well (or with relish). Closer examination teaches us that such “sentences” are summaries of previous acts of clarification and distribu­ tion, just as the transitive sentence Munggona-nang Dingerkon-nang marabik “the people from Munggona wounded the people from Dingerkon” (37) was to be understood as a summary of the sentences previous to this statement. In the first case (53) ninye “man” would already have been defined, and it could be assumed that it was known, for example: (52) Ya

areb-ma-lum

ninye asik

stone/adze give-dur.-2sg./pa.III man

hamlet

be-(y)uk. go-3sg./pa.III

The man you gave the adze to went into the hamlet. In the case that ninye asik is a compound the “complete” sentence would perhaps read: (53) Yalenye, basam fatan Yalenye

pig

ate, ninye asik be-(y)uk. man

hamlet go-3sg./pa.III

When the Yalenye (the figure of an ancestor) had a craving for pork, he went to the hamlets of humans. The second sentence of (51) is composed of the follow­ ing parts which clarify the status of teleb either as an attrib­ ute or as an adverb:

dib-ma-1.

be-3sg./med.-when

eat-dur.-3sg./pres.

The tobacco is good, and he smokes it Saboka dob-lye-ora,

teleb dib-ma-1.

tobacco

good

take-3sg./med.-when

eat-dur.-3sg./pres.

He has taken tobacco, and he smokes it with relish. In the first instance of (51) things are different. Although in the abbreviated form the status of bope “warm” is systematically uncertain, we know from the cultural cir­ cumstances that bope belongs to the verb as an adverb (or as an object). The “taking up of warmth” or, to imitate the construction, the “making warm” of earth, men’s house and everything that is to thrive is one of the found­ ing acts of the ancestor: the ancestor is asked for a cyclic repetition of this act, compare: (55) Nun tokwe bope fidib-nam-se-lam. us

earth

warm

take/up-fut.III-us-2sg.

You will warm up our soil. The interpretation of a construction directly depends upon something that is not part of the utterance this con­ struction is embedded in, on previous acts of clarification and distribution. The summarizing of series of words and clauses does not exact grammatical expedients, as there are case suffixes or means of agreement, which would make the utterance understandable by itself. The inner structure or cohesion of such a summary depends on the discourse, not on the morphology of single word classes. The gram­ mar of Eipo can be conceived of as a process in the course of which an attempt is made to summarize the acts of clarification and distribution and to make sentences out of utterances, sentences no longer totally dependent on discourse and context, but more or less autonomous. In one case such summarizing does not succeed at all, namely in serialized attributive and in predicative con­ struction. Feldman (1986:130) coined the expression “nominal serialization” for the phenomenon whereby many Papuan languages, instead of “the short, quiet man”, say “short man, quiet man”. A single head is repeated when it is specified by two attributes. A specify­ ing element is also repeated when it refers to different heads. A special case of this serializing is that the relative clause repeats rather than deletes the head of the superor­ dinate clause. Thus the Eipo say: (56, cf. chapter 8.2.2.1) kil teleb, kil mikib woman

desire/vn. for

good

good

woman

strong

a good and strong woman. The following is an example of repetition of the head as well as of the construction of which it is a part. While the head must be repeated, it is optional which parts of the surrounding construction are taken over: (57, cf. chapter 8.2.2.1) Yuk motokwe bi-nam-ne dara, fere motokwe other

mountain

go-fut.III-vn.

t.

far

mountain

51

I

bi-nam-ne dara, sik kayape-ak: “Angun-de go-fut.III-vn.

t./ds.

they friend/pl.-at

you-but

bob-se-lulum!” winya-lam-ik. carry-us-2pl./hort.

say-hab.-3pl./pa.III

With regard to carrying into other areas far off, there they said to their friends: “You shall carry it for us!” (This passage is about the sacred digging stick, which must not be dropped or shaken, so that those who carry it often take turns to prevent this.) Perhaps there is still another case of unusual structuring or of the impossiblity of summarizing: when two clauses are to be connected. Thurman (1975:342) distinguishes between two kinds of connection, “true chaining”, which, according to him, occurs only in the languages of the highlands of New Guinea, “where it serves to chain to­ gether a sequence of clauses by predicting the subject of the next clause”; in this “true chaining” the clauses mostly have the same subject (excepting the case of subject change), the same object and often also the same adverbi­ al phrases. “On the other hand, the other type of chaining, for which we prefer Loos’ term linkage, operates on the sentence and even the paragraph levels to link sen­ tences or paragraphs together, usually by repetition of at least part of the previous clause”. The first and the second sentence in the next instance are connected by exactly what Thurman calls “linkage”. Quite often the temporal relation between the two sentences is a general one; against a background something new happens. The consecutio temporum, however, may be much more com­ plicated and connected with a change of place, a change of subject or with the sequence of actually incompatible actions. The construction arising then might be called “linkage chaining”. There is no generalizing or summariz­ ing “and then” or “and there”. The connecting clause is “adjusted” in single steps to the following one; the first step may be to state the temporal setting, the second the scene of action. The connecting particles are attached to the verb as a repeated “head”. The step-by-step “adjusting” especially reveals itself, when no particles, but different tense and aspect forms of the connecting verb are used. (58, cf. chapter 8.5.2.4) Mote seb-uka ab-ik-ye, dance

asik bolub-m-ak.

dance-inf./suc. make-3pl./med.-and hamlet gather-dur.-3pl./pres.

Asik bolub-m-ak-ye,

bolub-uk-ab-ik-irye,

hamlet gather-dur.-3pl./pres.-and gather-inf./suc.-make-3pl./med.-and

ni bolub-ik-ye

dara, wine: “Areb-na-(a)b!”

all gather-3pl./med.-and

t.

now

(59) Menten-m-ik

dare, el ketengbantamu-lam-uk

miss-dur.-3pl./pa.III

t./ds.

it

west

j

I

I ;

s: i

;

be-hab.-3sg./pa.M

ora, ayam ketengbantam u-lam-uk

ora,

t

when ayam

when

-

west

be-hab.-3sg./pa.III

yin dob-uka ob-uk.

\

bow

i

take-inf./sue.

kill-3sg./pa.III

They (the men) missed him (the Yulye, an ancestral figure worrying men), and it lived in the west, the ayaffl rat lived in the west, it killed him with bow and arrow.

give-fut.III-lpl.

winya-lam-ak. say-hab.-3pl./pres.

Having danced they gather in the hamlet. And then they say: “Let us give!” (In order to prepare a dance and exchange feast the men hunt, gather in the for­ est, dance, come together in the hamlet and prepare the exchange transactions.) 52

In the second sentence the “gathering” is first presented as an action that is still going on and not yet completed; this is indicated by the aspect-suffix -ma- after the verbal stem bolub- “gather”. After that this action is presented as completed, so that a new action can follow in succession; the suffix -irye as well as the special verb form used here point to the fact that the relation between the “having gathered” and the subsequent summons is a causative or, more cautiously expressed, an obligatory one. Those who have gathered and those who summon need not necessarily be the same persons, so that the narrator puts in another break: the “gathering” is presented as a complex ed action in the verbal form bolubik-ye, the -ma suffix is missing, at the same time a sentence-medial form of the verb is used, that is, a form which can only conclude non­ sentence-final clauses; the conjunction dare, however, places the following action in opposition to the first action, so that in spite of the obligatory connection the acting subjects need not be identical. In contrast to what I said concerning attributive and pred­ icative constructions, “linkage chaining” is a favourite, but in no way an obligatory construction. In (58) the connec­ tion could have been brought about by a single dare. The choice of the connective construction is often a matter of personal style (cf. chapter 2.3). The gradual adjusting to the subsequent sentence fits as a special case into the pro­ cess of continuing clarification, which makes “autono­ mous” sentences with case suffixes and with more than one noun phrase appear as derived and summarized. Many instances cannot be systematized, reference to num­ ber, gender, time, space, or agents being slowly clarified. Thus sik betinye “the two of them” often precedes the naming of the two participants, or a deictic reference like u-bisik “the way down there” is often clarified by more special indications, for example through dabotam “the low­ lands” or, again, through proper names. Clarification and expansion become systematic, when pronominal and spatial deixis are “semantically filled up”. Here is an example of the clarification of pronominal deixis:

The following is an example of the expansion of spatial deixis which is at first rather general. m Yo laswe ou-deib-uk shield

outiba, kwemdine yo laswg

down-put-3sg./pa.III down/there creation

Marikle-nang une-tam, ortam Marikle-people

be/vn.-side

shield

dei-am-uk-ak

across/there put-perf.-3sg./pa.l U-at

\ :

tokne deib-uk,

ukwe a-ub-ma-l-tok

landslide put-3sg./pa.III fire

u-lam-uk-ak

win

here-be-dur.-3sg./pres.-rel.

ob-ra-nin

bin-am-dik.

be-hab.-3sg./pa.III-at fight-refl.-inf./sim. go-perf.-3d./pa.III

He (the ancestor) put the (sacred protecting) shield down below, at the side, where the Marikla live, there, where the landslide began, there he had put it, there, where now the fire flares up (in the gardens), it was there, and there the two of them went along quarrel­ ling (two ancestors quarrelling about the shield). The principle of clarification and that of distribution overlap when an object (or any other part of the sentence) has to be defined by an apposition: (6p Talim-nang aik dob-nam-ak-ye, Dokuramduweik Talim-people

hut

take-fut.III-3pl.-and

creation

take-fut.III-3pl.-and

The Talim people, in order to build a hut, in order to build the sacred Dokuramduweik men’s house of the times of origin [...] Generally one can say that the things which are to be more closely defined are transferred to a second sentence, the construction of which is identical with the foregoing one, which, however, can leave out certain parts of the sentence (mostly the subject) or replace them by anaphora and which weaves the more closely defining parts into the construction now “relieved” of some parts of the sen­ tence. The principle of clarification is grammaticized in the prin­ ciple of distribution. Subject and object, i.e. the core ele­ ments, are the parts most frequendy distributed: (62) Basam arye wine deib-uk pig

s.

now

put-3sg./pa.III

obora, Sunum then

arye bebea lik-lam-le.

Sunum

The herdswoman of the wild pigs scrapes the sky. The herdswoman of the wild pigs scrapes it with flashes of lightning (by striking and whipping with a cord. That is how the flashes of lightning come into being in the lowlands where there are wild pigs). Basam arye tokwe, basam arye Binal kon pig

s.

earth

put-3sg./pa.III

s.

Binal

top

deib-m-uk-ine

basam luk-donok-uk.

put-dur.-3sg./pa.III-sc.

pig

open-throw-3sg./pa.III

The pig set (created) the (fertile) earth, the pig set it at the top of the Binal mountain, the pig turned it over (broke it up to make it fertile).

(64) Basam si

ora, winilye fetereb-ki-n.

pig

and

name

three

explain-you-lsg./pa.I

I have told you and explained to you three names of pigs (of the times of creation). Methods of summarizing especially include the extraction of the subject and its resumption by means of a pronoun as well as the thematizing (and the simultaneous left dis­ location) of a noun phrase or of an actant; extraction and thematizing, of course, also demand richer means of intonational structuring (cf. chapters 8.3.2) Only at the end of this integrative process, that is, the summarizing of serialized or otherwise loosely connected units into a clause, do “autonomous” sentences with more than one or two actants appear. Here the role of these actants is often marked by suffixes or postpositions.

The pig gave birth, it gave birth to Sunum (the first man of the Kisamlu clan).

garden

Dokuramduweik del-i-lam-ak.

assign-hab.-3pl./pres.

Tengket-nang arye

put-der.-hab.-3pl./pres. Tengket-people

pig

(65) Wa gum ublinye bite ub-lye-nye

deibuk.

Dokuramduweik

lightning scrape-hab.-3sg./pres.

The combining and integration of such loose, serialized structures begins with the insertion of co-ordinating par­ ticles which maybe only help to mark the phrases with marked intonation, compare

Dokuramduweik

mem kwemdina dob-nam-ak-ye, [...] taboo

kil

tend/vn. woman s.

s.

del-i-lam-ak.

not

ninye arye

be-3sg./med.-n. concerning man

s.

nuk-lam-ak. The men attribute gardens to the one who does not own any.

put-der.-hab.-3pk/pres.

Sal arye kil

ate, el makal-nang ate yin

They founded the Dokuramduweik, the people of the Tengket clan founded it.

man

for

s.

woman

he

brother-people

for

bow

are-lam-le. give-hab.-3sg./pres.

Peripheral cases such as “indirect object” and instrumental or adverbial phrases, too, can be distributed (see chapter 83.2.3). Two examples follow which are less systematic than those in which core elements are distributed. In both of them an adverbial noun phrase is inserted later: (63) Yoa win

kil

wild

woman

tend/vn.

im lik-lam-le. sky

scrape-hab.-3sg./pres.

Yoa

The man gives a bow to (her) brothers as (a pay­ ment) for the woman). A-kwatema arye ninye nakina talel-yanga-lam-le. here-snake

s.

man

sickness

grasp-come-hab.-3sg./pres.

This snake (the totem of a clan) causes illness to the people.

wild

53

p

On the one hand, such acts of summarizing cause the rise of grammatical structures working against strange inco­ herence of series strung together which were described above, and create coherence within a sentence; on the oth­ er hand, they are not only acts of summarizing in the grammatical sense advocated here, but real summaries of what is known already and what has already been told. The examples (64) and (65) are some of the few clear cases of sentences with three noun phrases found in the Eipo texts (Heeschen 1990b); these examples and others, quite similar, but less clear, all appear in additions to, and explanations and paraphrases of already given textual material. They really summarize, and their grammatical structure represents a possibility of the Eipo language and is, as it were, the sum of preceding distributed units. If one looks only at the marking of the subject, its marking in (62) and (63) must be contrasted to all the cases where there is no marking. One must also keep in mind that the marking is not obligatory, as the texts are summaries, so that the roles of the different actants should be clear. In several approaches I have tried to show what the prin­ ciples of distribution and of additional clarification con­ sist in. The strategy of the Eipo speaker evidently consists in introducing no more than one noun plus one qualifying element or no more than one verb plus one actant in one clause; from the point of view of grammar, this corre­ sponds to the principle of distribution, and, from the point of view of discourse, it corresponds to the principle of not adding more than one piece of information per clause. The construction of more complex sentences was understood as a possibility which exists literally at the end of a speech or text. The way leading there consists, firstly, of series of attributes, appositions and predicates and, secondly, of the summarizing of nominal and verbal series. Serialization for the sake of clarification, on the one hand, and clarifying means of cohesion within a sen­ tence, on the other hand, are strategies complementing and excluding each other. The following sections will focus on examining from the theoretical point of view what continuous clarification means for the concept of the sentence and for the formation of linguistic concepts, what the basic units consist of and what concept of grammar is suited for the description of processes and not of rules. 3.3.3 Deixis, ellipsis, expansibility The beginning of each theory is characterized by asking questions and by being astonished. Variability in history, morphological complexity, universals of wordorder or syntax permitting an indefinite number of sentences with restricted means were such beginnings. This study starts with the astonishment caused by the discovery that speak­ ing is always embedded in the circumstances of communi­ cating with and understanding each other, circumstances which make the individual speech act appear only as a link in a long chain of non-verbal action and knowledge. From this it becomes evident why, in my opinion, some gram­ matical forms are characteristic, namely those which, depending on the conversational situation, the necessity for information and the ascertainment of having been 54

understood, permit clarifying sequences and series. The following arguments are to be understood as an attempt at systematizing the initial astonishment mentioned above. One may ask the general question of what the use of lan­ guage as a “tool” actually is. Evidently, language is little suited for describing an unknown path in such a manner that we can find it without difficulty; it is not easy to get an exact idea of an unknown face by a mere verbal description. It is probably still more difficult to fabricate a tool or build a hut only with the help of verbal instruc­ tion. But language is admirably suitable for referring to well-known paths and faces and for directing activities. Speaking always refers to a known, common sphere of action and perception. The mere naming possesses a force directing action and regulating behaviour: name the path, and one knows that today it will be difficult to use, because the torrents of last night’s rain have washed away planks and bridges; name the man, and people laugh, because he has a strange way of speaking and because he is stingy; say that you will today go and fetch the leaves of a certain tree, and people know that you are going to thatch the roof of your hut. Speaking focuses on a tiny part of what is known or is presupposed to be known for the moment. Language integrates an enormous amount of the data from environment and social intercourse. But the data come from a common sphere of action and percep­ tion, and thus mere naming already has communicating functions in that it appeals to this common ground, makes people adjust to communal action, separates the known from what is foreign, deals with social and cognitive dis­ sonances and hands down the image (the ideology) of the spatial and social organization of a society against the course and change of times (cf. Heeschen 1988). Some other, more specific observations may be added and fitted into this general picture. 1. More than 80% of the information pouring in upon man from his environment reach him through the visual channel. According to Mehrabian (1972) the visual channel contributes 55% to the effective­ ness of a message, the “vocal” (i.e. among others paralinguistic and intonation features) 38% and the verbal 7%. The percentages may be doubtful, but the statement that the non-verbal channel offers more information has been confirmed by research (cf. Wallbott 1990:86—87). In the case of requests and demands non-verbal behaviour is of crucial sig­ nificance (cf. Laucht Sc Herrmann 1978, WinterhoffSpurk Sc Grabowski-Gellert 1985). When flirting, an interaction which one can hardly imagine without verbal utterances, the male partner expresses himself verbally, but the female partner controls the situation by non-verbal means (Grammer, work in progress). But there are also counter results (Wallbott 1990: 87-89), from which I conclude that the importance of the verbal channel increases when the tasks grow more complex, for instance, when a deception has to be discovered. Speaking is a means of focusing, and more complex tasks turn attention to the probability and the truthfulness of the verbal utterance, so to speak, to the history and deep structure of the verb­ al action.

2. The role of language in social interaction is uncer­ tain; the tasks it can take on in this respect in a cer­ tain society have to be learnt from ontogenesis. Pri­ marily, language is a system of reference to space, to names of localities and living beings, to movements in this space, it is an organon processing and inte­ grating information, which draws up a model of the environment. If one can talk of innate characteris­ tics in the ontogenesis of language acquisition or if it is permissible to conclude from the “Funktionslust” that a genetic programme exists, the “urge” to know the names of things must be one of them. Children (and also grown-ups) are content if they can name an unknown object or an unknown per­ son, as if by this act alone a more thorough cogni­ tion could be reached; thus one gains a label for things and persons with which it is possible to oper­ ate, play and deal in the absence of the real objects. Even the first vocalizations of children are directed to objects, less frequently to the partners in interac­ tion (Freedle & Lewis 1977). The same kind of prec­ edence is to be found in the inability of small chil­ dren to talk to others about things that are removed from the situation of the speakers in time and space (Trevarthen Sc Hubley 1978); Keenan and Schieffelin (1975) prove that the conversations of young chil­ dren refer to objects that can be perceived. Bergmann (1990:206) proposes “that in every kind of discourse there operates a basic principle which I shall call the principle of 'local sensitivity7and which can be described as the structural tendency built into every topic talk to turn to local matters77, “local mat­ ters77being objects and processes in the field of per­ ception. 3. What we know from the history of evolution sup­ ports our opinion about speaking as removed and distanced from interaction and as object-dominated. New patterns of behaviour —and speaking is a pat­ tern of behaviour —are pace-makers of evolution; they are enforced by moving to new biotopes and changing exploitation of resources. This is support­ ed by Jerison’s (1976:101) opinion that speaking and language developed as instruments of “mapping77 new biotopes and of the optimal exploitation of new or plentiful resources. According to this, lan­ guage is primarily a system referring to space, to names of localities and living beings, to movement in this space; in congruence with the evolution of all the systems of communication it is, first and fore­ most, an organ processing and integrating informa­ tion and creating a model of the environment. Jerison says that even today we need language for drawing up directions for acting - plans for move­ ment in space —and for telling stories. Communica­ tion, according to him, has developed as a side-effect of evolution. “We need language more to tell stories than to direct actions. In the telling we create mental images in our listeners that might normally be produced only by the memory of events as recorded and integrated by the sensory and perceptual systems of the brain77.

When one tries to sum up the opinions, observations and theories quoted above, a paradox arises first. The ability of man to experience and perceive something with his senses and to move in a known environment without language or only with fragments of language cannot be expressed by any language and lies beyond the sphere controlled by grammar, insofar as it is intended to be the realization and the expression of those experiences. On the other hand, language reaches its full functionality and its adaptive advantages —to use a biological argument —only when, beyond any experience and beyond any interactional con­ text, it serves as a mnemotechnical tool and only simulates reality. If, with Jerison, we content ourselves with the gen­ eral remark that we need language in order to tell stories, that vital functionality of language is not immediately evi­ dent. The ethnological literature knows enough examples which show why the selectional advantages of language consist in the simulative mode (Reynolds 1976) of speak­ ing. Australian myths enumerate paths and water holes of areas even at a great distance; in an emergency, for instance in times of drought, the myths help to guide the destitute to the saving water holes, even if they have never walked there before (Birdsell 1979). Myths are also reports about the times of occupation and settlement of the land, and the exact tradition establishes the rights of ownership and helps to settle quarrels on this matter. Tales or reports keep a record of exchange and gifts in former generations so that, again in times of need, later generations may expect help from the descendants of old exchange and trading-partners they have never seen before. Stories and songs describe exemplary liberality or castigate cases of greed or miserliness, they impart values and concepts of values and they stabilize ideology and identity. The paradox dissolves, if speech acts, focusing on some­ thing within the environment immediately to be experi­ enced with the senses, and complete “speech products77 are not contrasted without any reference to each other, but are regarded each as the final result of a series of pos­ sibilities. In the transition from the one to the other pos­ sibility Biihler (1934:366—367) sees “einen Befreiungsschritt, der vielleicht einmal im Werdegang der Menschensprache zu den entscheidendsten gehorte. Wir vermogen ihn zwar nicht historisch zu rekonstruieren, wozu so gut wie jeder Anhalt in der Linguistik von heute fehlt, konnen ihn aber systematisch bestimmen als die Befreiung, soweit sie geht und moglich geworden ist, von den Situationshilfen; es ist der Ubergang vom wesentlich empraktischen Sprechen zu weitgehend synsemantisch selbstandigen (selbstversorgten) Sprachprodukten.7725

25 “[...] a liberating step that was perhaps one o f the most decisive ones in the development o f language. Though for lack of practi­ cally any basis in modern linguistics we cannot reconstruct it his­ torically, we can determine it systematically as the liberation from situational clues, to the extent that this is possible; it is the transi­ tion from essentially empractical speech to language products that are largely synsemantically independent (or self-supporting).” (Buhler, translated by Goodwin, 1990:418).

The transition consists in the fact that “mit der Befreiung aus den Umstanden der Sprechsituation eine neue Verankerung stattfindet [...]; es ist das Symbolfeld der Sprache, worin die neue Fixierung erfolgt” (Buhler 1934:367).26 Buhler, it is true, in the passage quoted above, transfers the liberating step into the remote past, but the essence of his “theory of language” consists precisely in proving the psychological reality of the transitions or liberating steps. It must be considered, however, that Buhler gained his concept of language from the history of the Indo-Euro­ pean languages; in all of these languages the liberating step was performed long ago, insofar as it consists in pro­ viding the means of granting the formation of an autono­ mous sentence without a deictic field, for instance by case systems which “anchor” and fix the role of the noun phrase in the sentence. In phases of language acquisition, in reduced (de-grammaticized) and reconstructing (re­ grammaticized) languages as well as in certain types of languages, which, for the sake of brevity, we shall classify as belonging to the agglutinative and isolating types, the systems and categories that are to enable a complete sym­ bolic field are probably still coming into being. In other words, case systems are used pragmatically and flexibly; the utterance which one day may become a ditransitive sentence is built up successively from smaller units. The thesis advocated in this study is that in the principles of additional clarification and distribution and in the princi­ ple of —structurally possible —summarizing of smaller units into sentences of increasing complexity some part of the speaker’s work becomes visible, a work which con­ sists in transferring an utterance still depending on the deictic field into an autonomous structure, namely the symbolic field of language. The usefulness of language as a tool accordingly proves itself in the known field of perception and action; it only gains its full functionality, however, in the discourse detached from what is here and now. Grammar will have to be understood as a stategy, or better, as a method, in the original sense of the word, of mediating between pri­ mary usefulness and full functionality. To speak in con­ crete terms and to return to our observations: the gram­ mar must be able to show which problems the speaker will have to solve, if he intends to present a rudimentary text like (31) in summarizing sentences. The autonomous sentence, however, which, following Buhler, we can call sentence without a deictic field, is ulti­ mately an illusion. The fact that language refers to space and to movements in space and thus always reverts to the deictic mode seems to be established by the history of evolution. The Broca region in the brain of the higher pri­ mates corresponds to a region directing and co-ordinating sequences of movements (cf. Dingwall 1979, Spuhler 1977). For Lorenz (1973:175) thinking is movement in a simulative mode, in his own words it is “probeweises und nur im Gehirn sich abspielendes Handeln im vorgestellten Raum”.27 Speaker and hearer concentrate their full atten­ tion on what is represented by the symbols of language, namely the objects and movements. “Man ist dort bei den Dingen, von denen gesprochen wird, und laBt die konstruktive oder rekonstruierende innere Tatigkeit zum guten Teil vom Gegen56

stand selbst, den man schon kennt oder soweit et durch den Text bereits angelegt und aufgebaut ist, gesteuert werden.”28 (Buhler 1934:171) Speech acts focus on one object, on one action and never on all objects, their relations and on all the sequences of actions belonging to a “Gesamtvorstellung”29 in the sense of Wundt. For this reason each linguistic reference to a cir­ cumstance is open, speech “laBt allenthalben Spielraume offen”30 (Buhler 1934:171). Buhler (1934:172) even speaks of the “prinzipiellen Offenheit sprachlicher Fassungen von Gegenstanden und Sachverhalten”31. Speech acts only obtain sense and meaning to the same extent to which they merely help to point to a number of objects and circum­ stances and to reconstruct a relation between them. “Die stoffliche Steuerung des Sprachdenkens ist ein Phanomen, welches mit einigen anderen Tatsachen zusammen den wichtigen Satz zu beweisen gestattet, daB das Andeuten, welches der zeigende Finger vollbringt, nicht nur die Leistung der Zeigworter charakterisiert, sondern weit dariiber hinaus auch im Funktionsbereich der Begriffsworter zu finden ist und zu den Struktureigenheiten der menschlichen Sprache gehort.”32 (Buhler 1934:172, cf. Hormann 1978:455) Accordingly, the pre-condition for the understanding of a speech act is exact knowledge of the social and spatial environment. Non-verbal communication forms the sense of an utterance, symbolic systems, “belief systems” and models divide among themselves the production of meaningful utterances. The speaker always “means” this whole context, the single speech act focuses on a tiny part of the whole. All these pre-conditions seem to permit the conclusion that the speaker passively receives sense and meaning. Prior to each utterance, however, there is the ability of man to comprehend something as a sign. Semi­ otic activity operates in dealing with objects, in the multi­ functionality of tools, in play, in rites, ceremonies, decora26 “[...] with the liberation from the circumstances o f speech a new anchoring takes place [...], the new fixation takes place in the symbolic field o f language.” (Buhler, translated by Goodwin, 1990:419). 27 “[...] acting on probation taking place only in the brain in an ima­ gined space.” 28 “One is then located at the things being talked about, and in one's constructive or reconstructive inner activity one follows for the most part the object itself, whether it is an object that one already knows or an object that has been sketched and constructed by the text.” (Buhler, translated by Goodwin, 1990:194). 29 “total conception”. 30 “leaves a latitude for operation open” (Buhler, translated by Good­ win, 1990:194). 31 “the fundamental openness o f the linguistic rendering o f objects and state o f affairs” (Buhler, translated by Goodwin, 1990:194-5). 32 “Together with a few other facts, the phenomenon o f the materi­ al guidance o f speech thought goes to prove the important prop­ osition that the sort o f indication given by the pointing finger is not only characteristic o f the work o f the deictic words, but can also be encountered far beyond that in the functional domain of the conceptual words, and indeed that it is one o f the structural properties o f human language.” (Buhler, translated by Goodwin, 1990:195).

don, ornaments, and in the significance of facial expres­ sion and gestures. In space, whose objects man relates to himself, to his history and identity and in social interaction all things become signs. A tool like the adze, which man obtains by trading and with which he manipulates his environment, becomes the symbol of individual wealth and communal welfare; with an expression like “he goes to pick the leaves of the trees’7, which is poor from a lin­ guistic point of view, the Eipo point to a complicated cere­ mony of fertility: at the end of the building of the men’s house the big-men walk through the countryside cutting off the leaves of the lower branches of certain trees, which then grow less in breadth and more in height; this again is a metaphor of fertility and growth. The little boys, too, whose growth is a cause for continuous concern “are to shoot up”. Simple linguistic references like “the boys step over the threshold of the men’s house, the men close the men’s house with banana leaves” show the poverty of the language in the face of the richness of meanings, built up by rites and ceremonies: the boys, who are torn away from their mothers in order to be initiated, experience their rebirth in the world of males (cf. chapter 2.4). The mere naming of objects and statements with a literal meaning are fictions. Semiotic activity is always ahead of the single word and of linguistic form. It is understand­ able that semiotics refers to these connections. According to Eco semiotic activity tends “alle Aspekte der Kultur und des sozialen lebens als Zeichen zu begreifen und gerade die Gegenstande mit einzuschlieBen”33 (Eco 1977: 42-43). The psycholinguist Hormann is not far from the position of semiotics. He reckons with the “von Cassirer betonten Fahigkeit zum Als-SymbolAuffassen [...]. Was uns hier als Phanomen der Bedeutungstransparenz entgegentritt —daB wir im Laut, in der Geste der Hand und durch sie das wahrnehmen, was sie bedeuten, namlich den GruB oder die Drohung oder den Hinweis —, ist als Prin^ip in der Wahrnehmung jedenfalls der hoheren Lebewesen immer enthalten; Bruner hat es einmal mit dem nicht ganz gliicklichen Ausdruck ‘going beyond the information given’ benannt. Tatsachlich schaffen wir Information, indem wir fiber die [...] Sinnesdaten hinausgehen.”34 (Hormann 1978:18) One only has to add what our examples and our train of thought leads to, — and Hormann does so too in the course of his argumentation —namely that the meaning of each utterance always goes beyond the linguistic form. .We create meaning in taking the scanty linguistic material appearing in the sound as an opportunity to transfer that which has been referred to into one of the above men­ tioned contexts. Thus the acts of speaking and of understanding have to be regarded as “die Konkretisierung eines allgemein schon vor der gehorten AuBerung vorhandenen SinnHorizonts”35 (Hormann 1978:395). A little later Hor­ mann says (1978:398) “daB die ursprfingliche Funktion der Sprachverwendung in einer Werbesserung und Erganzung’ der auch schon nicht- und vorsprachlich geleisteten Orientierung besteht”36. For the linguist these state­ ments would, indeed, be quite irrelevant, if he did not find

the pre- and non-linguistic orientation or the references to the semantic horizon in the language. The first parts of a specifying and clarifying series achieve the linguistic orien­ tation. Bfihler’s deictic field presents the pattern with the help of which the symbolic field is filled. It is important to keep in mind that there is no absolute contrast between the two fields. Both are mediated in language in manifold ways. In the observations and examples (31) there are three instances of the filling-up and expanding of a deictic expression. Especially (43) and (60) are characteristic of the Eipo speakers. Language makes use of a topomnestic method: in the same manner as a cavalry regiment in start­ ing position only needs the direction “towards the end of the wood”, for the first part of a specifying and clarifying series it is enough to name a “landmark” or a “direction” for the following utterance, except that the orientation is not achieved in real space, but in the model which speaker and hearer possess of the things and persons. The first general reference occurs, as it were, still outside the “sen­ tence” that is going to follow. The deictic field is becom­ ing identical with the mechanisms of thematizing. The symbolic field closes inasmuch as the theme is “situated” in space and time (cf. 48, where this “situating” is illus­ trated). This situating, however, no longer takes place in real space, but in the imagined one. Thematizing is noth­ ing but generalized deixis, a topomnestic pointing to something by means of language, whose primary function was locating. By the more exact “situating” of what has been pointed to in a general manner the agreement between theme and the person-number affixes of the verb arises, so that manifold affinities exist between theme and subject. The subject noun phrase is an anticipatory expansion and clarification of the verbal person (-number) suffixes. Conversely, with respect to number, these suf­ fixes clarify the noun phrase (cf. chapter 7.2.1.2). If any form of extraction is understood to be necessary to start the original or topomnestic pointing to something, the “situating” by means of agreement will be necessary in order to capture and “anchor” what has been pointed to in the “sentence” as a symbolic field. Of course the relations between deixis, reference and naming are well-known. But modern linguistics starts from the sentence as a complete symbolic field; for this reason pronouns, for instance, can appear as the result of trans­ formational deletions. Data from ontogenesis suggest an opposite relationship: an anticipating or repeating agree33 “[. ••] comprehend all aspects o f culture and o f social life as signs and expressly to include the objects.” 34 “[...] ability stressed by Cassirer to comprehend something as a symbol [...]. What we here encounter as a phenomenon o f semantic transparency - that in the sound, in the gesture o f the hand and through them we perceive what they mean, namely the greeting or the menace or the hint —, is as a principle always inher­ ent in the perception o f at least the higher living creatures; for this Bruner once used the slightly infelicitous expression ‘going beyond the information given’. Indeed, we create information by going beyond the data conceived by the senses.” 35 “[. ••] the materialization o f a semantic horizon which is generally already in existence prior to the utterance heard.” 36 “[...] that the original function o f using language consists in an ‘improvement and a complementing’ o f the orientation already achieved at a non- and pre-linguistic stage.’ ”

57

ment brought about by deictic elements —either by “pro­ nouns” or by words of local deixis —seems to be funda­ mental. For C. and W Stern (1965) deixis is the startingpoint for the development of the descriptive function of language. In the course of language acquisition, first deictic acts, then deictic expressions are replaced by syn­ tactic constructions, which, in my opinion, are expansions or clarifications of the generalized topomnestic deixis and thematizing.37 Constructions like “he sees him, the man the boy” are learnt earlier than those like “the man sees the boy”. A proposal by Hudson (1984) is in agreement with the data from ontogenesis; he interprets determiners as heads. According to this, the pronoun would be the basic form in he, (the) man, while in the terms used in this study, the nominal expression, added as an apposition, would have to be interpreted as clarification and expansion. While the official typology regards the forms of agree­ ment as an equivalent of case systems, that is as construc­ tions permitting the sentence as a complete symbolic field, practical linguists are aware of the fact that agreement as a means of sentence construction is fast becoming ineffi­ cient (cf. Foley 1986, Whitehead 1981—9182). Agreement does not lead to the construction of longer, self-support­ ing sentences; we find the same short-windedness and the same clarifying serialization as in our examples. In other words: agreement does not lead to a hierarchically struc­ tured sentence, but is to be conceived as a linear, clarifying series, in which the nominal expressions are placed in a loose ad-position to the deictic expression and may be considered as expansions of it. The general principle that preferably no more than one piece of new information should be given in one sentence qua utterance is probably also applicable to languages working with agreement; this means nothing but that thematizing in the form of extrac­ tion and repeating agreement of situating morphemes are sufficient for the construction of the clause. Considering what common experiences precede each act of speaking and hearing, and considering that language is an instrument suitable for focussing on something in a known space of perception and action, each utterance is of necessity incomplete, not to say elliptical. When restricting the terms “incompleteness” and “ellipsis” to language, all the clauses preceding a final “summary” are incomplete if judged by the forms of summarizing. The phenomenon of “ellipsis” only exists when the sentence as a complete symbolic field or, following our observa­ tions, summarizing sentences, are taken as the starting point. Sentences constructed according to the principles of clarification or of distribution are incomplete or ellipti­ cal by definition. For clarification is an open process. However, seen pragmatically, it ends quickly and precisely at the point where the speaker assumes that the hearer has understood him; and distribution prevents the arising of autonomous sentences or clauses: what is being uttered by and by in linear sequence is not hierarchically inserted into the following clauses. The principle of distribution runs counter to the fundamental role attributed to “apprehen­ sion”. “Agreement” has the function “that part of the apprehension of a linguistic object is to keep it con­ stant”. The same linguistic object, Lehmann goes on, could appear in different places of a sentence. The fact that the object is always the same is expressed in natural 58

languages “by means of agreement in grammatical cate­ gories, i.e. by categorizing, mentioning the categories of, the apprehended object” (C. Lehmann 1982:233). Our general premisses permit us to say that the constancy of the objects by means of agreement or such like is no longer necessary, as the objects are constantly present in the space of perception or are kept in memory. Linguistic presence and constancy of the objects introduced in speech are important wherever the linguist examines sum­ marizing sentences, but they are the final product of utterances in which the linguistic means are either not yet in existence or are still coming into being.38 When perusing the literature on ellipsis it soon becomes clear that the aim of these introductory remarks is not so much to judge utterances from the perspective of com­ plete symbolic fields (or: deductive theories), but to understand the principles of speech as a starting-point for developments which mediate between deictic field and symbolic field or which show the pragmatical foundation of processes of categorizing, not of the finished categor­ ies themselves. Ortner (1987) has made an extensive contribution to the history of the concept of ellipsis, and important contribu­ tions concerning the theory are to be found in the anthol­ ogy edited by Meyer-Hermann & Rieser (1985). In a sim­ plified version the discussion can be summed up as follows: on the one hand there are those who regard ellip­ ses as deducted from complete sentences and take them for an attribute of the parole\ on the other hand there is a group willing to accept ellipses also as a fundamental fact of language and who attribute to it a place not only in the theory of performance but in the theory of language it­ self. The first group is characterized by Ortner (1987:136) as follows: “Es wird also zwischen Ellipsen und Vollsatzen eine Ubereinstimmung im propositionalen [...] Satzgehalt angenommen, wobei noch in den fortgeschrittensten Konzeptionen daran festgehalten wird, daB Elliptizitat eine sekundare Eigenschaft —“defectus dictionis!” - ist, also eine Eigenschaft, die aus Satzen, die unter onomasiologischem (Benennungs-)Aspekt eigentlich vollstandig waren, Ellipsen macht, um damit besondere kommunikative Aufgaben zu realisieren.”39 37 Cf. Bowerman 1985, Miller 1976, Moerk 1977; concerning deixis Harweg 1990, Lyons 1977, Sennholz 1985; concerning congru­ ence - here concerning the special case o f cross-reference - Foley 1986, Keenan 1987, C. Lehmann 1983 and 1985; concerning interesting cases of congruence in Papuan languages (including the “situating” morphemes attached to the verb) c f Anceaux 1965, Foley 1986, Foreman 1974, Lang 1973, Peckham 1982, Whitehead 1981-1982. 38 C f Bromley (1981), who describes a language having only extremely few means at its disposal for keeping constant the lin­ guistic objects in the sense o f apprehension. 39 “Accordingly, one assumes an agreement in the propositional [...] content o f a sentence between ellipses and complete sentences, while even the most advanced conceptions adhere to the idea that ellipticity is a secondary attribute - “defectus dictionis” - , that is an attribute turning sentences which, from the onomasiological (terminological) point o f view, were actually complete, into ellip­ ses in order to realize special communicative tasks.”

Wherever it is judged that something is missing in a sen­ tence, for example the subject in utterances of small chil­ dren or in the telegraphic style of aphasiacs (Claus Heeschen, oral communication) or that something is deleted without leaving a trace, we are among the first group. It must be mentioned in their favour that the linguistic form of ellipses is strikingly regular - for example the omission of the subject. In the same way as it is difficult to describe the grammar of a three-year-old child in positive terms, that is without recourse to the language of grown-ups and without judgements about defectivity, it is difficult to free oneself from the concept of ellipsis, even if one takes it for a fundamental feature of language. Only if the lan­ guage of the grown-ups or the “complete sentences” are understood as summaries, can one give up the concept of ellipsis: the linguistic form of ellipsis consists of basic units which the speaker can put together und which in the course of being put together make visible the “construc­ tion” of grammar during speech. I shall return to this. Within the second group there are three approaches which I would like to mention. 1. Strohner & Rickheit (1985) introduce the concept of functional completeness into the discussion. They contrast structural analyses, which, for the sake of brevity, we would like to define here as conven­ tional grammatical analyses, with functional analyses. Grammatical analyses may ask whether a sentence is grammatically complete, functional analyses examine whether an utterance serves its purpose in the course of speech or in a dialogue and is understood or can be understood. Speech and dialogue do not aim at the production of structurally complete sen­ tences, but at achieving understanding in a co-opera­ tive effort. If this aim is achieved, the utterance is functionally complete. Strohner & Rickheit (1985:257) state that: “So einleuchtend aufgrund der bis jetzt angestellten Uberlegungen die Erganzung der Vollstandigkeitsbeurteilung nach strukturellen Kriterien durch funktionale Kriterien auch sein mag, so erschreckend kann es jedoch auch sein, feststellen zu miissen, sich damit den festen Boden strukturalistischer Sprachanalysen unter den FiiBen weggezogen zu haben.”40 I should say that the shock need not be so great, if it is taken into consideration that forms and systems of communication have developed in the service of functions. Structural completeness is certainly a by­ product of evolution. It requires explanation and deduction, and the “incomplete” linguistic structures developing directly in certain functional circles are the primary ones. Biihler’s opinion that language develops by becoming more and more able to do without situational help supplies the general context for this functional aspect of language. The concep­ tion of summarizing presented here may be a first step towards understanding structure as a result of functionally adequate types of utterances; this con­ ception, however, has been demonstrated only with very specific linguistic material.

2. Parallel to Biihler’s abstractive relevance of the linguistic sign, Ortner (1987:35) speaks of a princi­ ple of abstractive paraphrasability. According to him it is significant for the discussion on ellipsis, because “auf Grund der angenommenen semantischen Gleichwertigkeit von Ellipsen und Vollsatzen erstere gerne als Ableitungen aus letzteren erklart werden. Tatsachlich sind Ellipsen und Vollsatze, aber auch Appositionsstrukturen und annahernd aquivalente Relativsatzstrukturen, Wortbildungskonstruktionen und annahernd aquivalente Papraphrasierungen, Infinitivgruppen und annahernd aquivalente Nebensatze usw. Mittel, mit denen ahnliche oder identische Bilder von Sachverhalten erzeugt werden konnen [ ...].”41. If I understand Ortner correctly, he interprets the types of construction, including ellipses and com­ plete sentences, as global signs possessing a content, a meaning or referring to the circumstances of a case. The exact reference is not affected, if the global sign changes its form or if the signifier only realizes components of the global sign. In that case the equivalence would be safeguarded by the compo­ nents meaning the entire facts in an abstractively relevant manner. In Wundt’s version: if the sentence is the successive unfolding of a total conception, then according to the principle of abstractive paraphrasability, some few single steps will be sufficient to still produce the total conception. With the help of Ortner’s principle the paradox that speaking functions best as a means of focussing in a known space, but that language finds its real functionality in the simulative mode, can be explained with the help of the theory of signs. In addition, it is a means of interrupting the endless process of clarification and of examining the sentences with forms of distribu­ tion in order to find out to what extent the clauses are paraphrases and thus do not require the means of agreement and cohesion. Paraphrasability in any case presupposes the co-ordinative linking of clarifi­ cations and sentences with distribution. The periods of Eipo, then, mainly consist of one instance of thematizing to which series of appositions and of attributing clauses of equal order are attached. .

40 “Plausible as it may be to complement the judgement concerning completeness with regard to structural criteria by functional crite­ ria on the basis o f what has been considered hitherto, it can indeed be alarming to realize that one has cut the firm ground of structural linguistic analyses from under one’s own feet.” 41 “[...] on the strength o f the assumed semantic equality o f ellipses and complete sentences the former are often explained as deduc­ tions from the latter. Actually, ellipses and complete sentences, but also appositional structures and approximately equivalent rel­ ative clause structures, constructions o f word formation and approximately equivalent paraphrases, groups o f infinitives and approximately equivalent clauses and so on are means by which similar or identical images o f facts can be created [...]”.

59

3. The study carried out by Kindt (1985), is in my opinion the most important contribution to the dis­ cussion on ellipses. He is the only one to consider ellipses seriously as autonomous structures and makes them the starting-point of a new syntax model. It would be very desirable to examine in detail the correspondence of his theses, which he extracted from German language material, with evi­ dence from this study, which is based on quite differ­ ent material. I shall return to this in the grammar proper. Kindt (1985:168) criticizes the hypothesis of reduction and of defectiveness, according to which ellipses must always be seen in relation to complete sentences or to self-supporting symbolic fields. He supports his critical assessment with two arguments: “Erstens sollte sich m. E. die Formulierung von Syn­ thes egrammatiken moglichst an den Konstruktionsprinzipien von Sprachproduktion orientieren; diesbeziiglich nehme ich aber an, daB Sprecher bei der Produktion elliptischer AuBerungen nicht so vorgehen, daB sie zunachst (gedanklich) einen vollstandigen Satz konzipieren und anschlieBend kontextuell redundante Elemente dieses Satzes wieder tilgen. Zweitens halte ich es fur zweckmaBig, wenn der generis che Stellenwert sprachlicher Einheiten in Grammatiken reflektiert wird; unter diesem Aspekt ist zu bedenken, daB elliptische Sprachformen moglicherweise zum Teil die primaren For men im Spracherwerb sind und daB sich die sogenannte grammatische Vollstandigkeit erst aus erhohten Formulierungsforderungen ergibt.”42 From a host of non-verbal-verbal interactions, questionand-answer games, fragments, and finally more complex, but still incomplete sentence structures arise —Biihler’s “language islands” —, which in accordance with the func­ tional completeness (cf. point 1) and the increased “requirements of formulation” necessarily lead to enlargement and clarification. Kindt proposes that the corresponding rules of production should be those of expansion; they guarantee, “daB Syntaxmodelle dem Prinzip der stiickweisen Bearbeitung sprachlicher AuBerungen in Produktion wie Rezeption Rechnung tragen”43 (Kindt 1985:192). (Kindt, who evidently is more of a theorist, overlooks that his basic assumptions could be supported in many ways by results of psycholinguistic research and by studies of comparative linguistics.) Different types of linking enable expansions.44 This directly leads us back to our sentences with distribu­ tion. Let us regard the sentences in chapter 3.3.2 and 8.3.2.3: the clauses can each be linked by a kind of “and”, insofar as one disregards the consecutio temporum, the differ­ ent sentence-medial forms of the verb and the aspect forms of the verb. Each clause is “defective” in relation to another one. One verb preferably attaches only one argu­ ment to itself, so that the linkage is not only a coincidental expansion of a construction which happened to turn out too small, but a necessary means of linking fundamental units of speech. From the general preconditions of speech, but also from the types of utterances which have already been described 60

in linguistic terms the conclusion can be drawn that posi­ tive principles of speech and of speech formation and not reduced or defective linguistic forms must be the startingpoints for final observations. If step-by-step construc­ tions, basic units which somehow are not yet tangible, are so important and if these units have to be specified, dis­ tributed to different clauses and expanded, it must impli­ citly be evident that complete or summarizing sentences only arise casually. The sentence is a by-product of the processes of speech formation. The units which have become tangible could not have been deduced from a starting-symbol “S”. Thus the interesting question arises; whether, for the moment, one could not do entirely with­ out the concept of the sentence, for evidently the pro­ cesses of sentence formation are much more interesting than the casually arising final product. The linguistic researcher proceeds from basic units, from blocks which, when connected, form a sort of sentence; he sees the pro­ cesses of clarification, distribution and linkage leading to longer units of utterances, and the most important sub­ jects of examination are the manner of linking and the problems speakers and hearers have in concluding the acts of understanding. The sentence is a field of possibilities and, as we do not need a starting-symbol, its form is open

42 “Firstly, in my opinion, the formulation o f synthesis grammars should, as far as possible, be oriented towards the constructional principles o f speech- production; with regard to this, however, I assume that speakers, when producing elliptical utterances, do not proceed in such a manner that they first (in their minds) draft a complete sentence and subsequendy delete contextually redun­ dant elements o f this sentence again. Secondly I believe that the genetic status o f linguistic units should be reflected in grammars; from this point o f view one has to consider that elliptical forms o f speech are possibly in part the primary forms in the aquisitiori o f language and that so-called grammatical completeness only results from increased requirements o f formulation.” 43 “[...] that models o f syntax take into consideration the principle o f the piecemeal fashioning o f linguistic utterances in production as well as reception.” 44 Cf. Paul 1960:138: beside the copulative connection o f subject and predicate “jede andere Art der Satzerweiterung” is brought about “dadurch, dass das Verhaltnis von Subject und Pradikat mehrmals auftritt. Wir konnen zwei Grundformen des auf diese Weise erweiterten Satzes unterscheiden. Die erste besteht darkij dass zwei Subjekte zu einem Pradikate oder zwei Pradikate zu einem Subjekte treten. Ist dabei das Verhaltnis der beiden Sub­ jekte zu dem gemeinsamen Pradikate oder das der beiden Pradi­ kate zu dem gemeinsamen Subjekte vollig gleich, so lasst sich ein solcher dreigliedriger Satz ohne wesentliche Veranderung des Sinnes mit einem zweigliedrigen vertauschen, dessen eines Glied eine kopulative Verbindung ist ” English translation: beside the copulative connection o f subject and predicate “each kind o f enlargement o f a sentence is brought about by the repeated appearance o f subject and predicate. We can distinguish two basic forms o f the sentence enlarged in this way. The first consists in two subjects joining one predicate oi two predicates one subject. If, in this process, the relation o f the two subjects to their common predicate or that o f the two predi­ cates to their common subject is completely identical, such a tri­ partite sentence can, without an essential change o f the meaning be exchanged for a bipartite sentence, one component o f which is a copulative co-ordination.”

and, in principle, a limit is not in view and theoretically not necessary.45 Such a way of looking at things promotes the compilation and descriptive inventorization of characteristics. Not all the characteristics marking what is to become a sentence are constantly and necessarily assembled. The falling into­ nation towards the end is regarded as one of the unalter­ able characteristics of the sentence; in 1989 I transcribed a text in the Yale language in which the speaker hastened on with rising intonation whenever I expected a full stop. This is permitted in the Yale language, because the pre­ ceding sentence can without any formal characteristics be linked as a relative clause to the following subject or to the form of extraction at the beginning of the following sen­ tence. The narrator, a big-man and, according to the judg­ ment of the listeners, one who “closes our ears and makes them ring” (someone who talks people over and pushes aside their own ideas), grammaticizes the apo-koinou con­ struction described by Paul (1960:96—97); he does not say “1 went to my father. He was in the garden”, but “I went to my father was in the garden”. I only mention this in order to show that when characteristics are collected a new theory need not be invented each time one hits upon something unexpected, provided it is not classed as be­ longing to transient parole. Linguistic research has followed many tracks leading, in my opinion, to the thesis advocated here. Considerations and results pointing in this direction are not only to be found in research on dialogue (cf. Goodwin 1986), in “systemic linguistics” or in the so-called “Prague School”, but also in the works of widely different authors who can­ not be attributed to any particular school. For Coseriu (1974) the system of language is a field of possibilities, norm as a temporary, realized fixation is a dynamic explo­ ration of those possibilities. Lehmann (1995) lays down the parameters of grammaticalization and in doing this demonstrates what H. Seiler (1986:1) means, when he presents his model of language as a process, as against the conception of language being a ‘formal’ or ‘abstract object’ ”; in Seiler’s opinion language is the result of a per­ manent process of construction. For Aitchison (1989, cf. 1987) the development of language at any stage is a “set of options”, of which only few are ruled out at the cross­ roads of the history of language. “However, a number of converging factors lead speakers to take certain recurrent routes. An overall aim in future research, then, must be to predict and explain the preferred pathways of language evolu­ tion.” (Aitchison 1989:170; cf. Ehlich 1982) With respect to “developing” or “emergent grammars” I would like to mention the “Anatomie der Syntax” by Doerfer (1973), which unfortunately has hardly been tak­ en notice of: grammar is understood as an increasing structuring of the simplest blocks of information; our example (31) can, indeed, —and in accordance with the subsequent discussion —be regarded as a sequence of such blocks. Doerfer reaches his views by the examination of a great number of languages, among them also “exot­ ic” languages. In my opinion Doerfer’s assumptions are closely related to the ideas found in works on grammati­

calization.46 If it is true that “structure, or regularity, comes out of discourse and is shaped by discourse as much as it shapes discourse in an on-going process” (Hopper 1987:142), then one may ask the question whether the concepts which follow from the notions of “language evolution” or “emergent grammar” should be applied to the sentence, i. e. to “S”. Indeed, some researchers argue against a theory-independent notion of “S”. Pawley (1987:357, cf. chapters 7.2A2.2, 8.4.1 - 8.4.3.2) says that serial verb constructions (and one may add: clause­ chaining) “show an attempt to crunch several underlying clauses into one surface structure”. Hence, the theorydependent dichotomy between deep and surface structure can be neglected: “It needs to be said that ‘clause’ is not a well-defined entity in either language. What we have, in fact, is more like a scale of constructions ranging from phrases to prototypical clauses to sequences of pro­ totypical clauses. In between are various construc­ tion types that are something less than or something more than the prototypical clause structure (Pawley 1987:357) The analysis of Ik texts led Serzisko (1992) to look for a means of segmenting speech and establishing linguistic units. Pauses proved to be the only reliable means. The analysis resulted in the establishment of N + V + connec­ tive sequences, of apposition, parenthesis, and after­ thought units; however: “Eine Konsequenz dieses Vorgehens war jedoch, daB die urspriinglich angenommene Einheit Satz nicht mehr als solche zu erkennen war.”47 (Serzisko 1992:3) Instead of “S” Serzisko found units of speech similar to those discussed in this grammar: units which can do with­ out much lookahead or backtracking and which conform to a one-lexical-noun-phrase constraint, i. e. to “clauses” which give only one piece of new information. —The playful proposal, then, to do away with “S” is not “out of the way” for linguists who are willing to try out new avenues. One can ask to what extent such approaches repeat the position of Paul, for whom language as an object of lin­ guistic research was only accessible through historical studies, and for whom, in Humboldt’s words, language only found its final destination in the individual. The cru­ cial question would be whether the linguist was willing to examine the above-mentioned apd-koinou construction as an option within the system of a language or whether he 45 I presented some o f these considerations at a symposium “Mutu­ alities in Dialogue” o f the research team “Dynamics o f Dia­ logue”, Werner-Reimers-Stiftung, Bad Homburg, March 13th 16th, 1991. The linguist Per Linell said, cum grano sails that most linguists would think I was crazy. Science also consists o f playing and trying out. 46 Cf. Heine, Claudi & Hiinnemeyer 1991, Hopper 1987, Hopper & Traugott 1993, C. Lehmann 1982. 47 “This procedure resulted in no longer being able to recognize the originally assumed unit sentence”

61

only regards it as a matter of individual style or as belong­ ing to the field of performance (cf. chapter 2.3). The gen­ eral question emerging behind this specific one is whether all the views of language as a process, a construction or as a “set of options” hold good only in language comparison or only in the history of a language over more than one generation, or whether they are also relevant synchronically, i.e. for single speakers or restricted social groupings: whether we, by comparison and by examining history, abstractingly state the formation of categories such as “sentence” or “subject” or “transitivity” and then assign to them prescriptive power over individual speech or whether the formation, the incompleteness and the grammaticalization of these categories are also active synchronically in the utterances of individual speakers. When looking at the implicative relations of typology, for exam­ ple those between case marking, agreement, pronominal co-reference, and verb serialization, one enters the field of language comparison; yet, one cannot comprehend the pragmatics and the optionality of single means in one sin­ gle language or with one single speaker. At any rate, the authors and studies quoted above are not working against the opinions which were developed in chapters 3.1 and 3.2 and were foremost in the discussion of the examples. The linguist has to follow the lifemoments of a language, he is required to comprehend the construction of his knowledge as well as grammar as a process. In these circular procedures he needs concep­ tions that are non-essentialistic and free of hypotheses. This chapter, from the examples and the discussion of sentence formation in Eipo to the concept of the sen­ tence, is a demonstration of that method and of, I hope, a careful formation of a concept. It possibly culminates in the “crazy” idea that the concept of the sentence can alto­ gether be dispensed with. As a theoretical concept it has proved unneccessary. As an “injunctive” concept it will be retained. I have proposed the term “injunctive grammar” (Heeschen 1990a) as a general framework for a grammar col­ lecting data and employing philological methods, for a description which is aware of the fact that it proceeds from the limited point of view of a meta-language and of fragmentary theoretical approaches and which then exam­ ines individual cases in order to abstract and generalize again with their help (cf. Sasse 1988, Pawley 1993). I have formed the adjective from “injunction” (from Latin iniungere, “add”) which is, according to Hassenstein (1976:367), “ein sachbezogener (= deskriptiver) Begriff der beschreibenden Naturwissenschaften, der wegen der Beschaffenheit des von ihm zu reprasentierenden Gegenstandsbereichs nicht die Kriterien einer Defi­ nition erfullen kann. Insbesondere gilt dies: 1. wenn sich ein zur Begriffsbestimmung heranzuziehendes Kennzeichen gegen den Rand des vom Begriff reprasentierten Sachbereichs hin allmahlich, flieBend verliert und nicht sprunghaft verschwindet [...]; 2. wenn fur die Begriffsbestimmung mehrere, unabhangig voneinander variierende konstitutive Merkmale vorliegen, aus denen bei dem gegebenen Stand der Fors chung kein einzelnes mit sachlicher Be62

grundung als alleiniges differenzierendes Merkma] herausgehoben werden kann.”48 “Injunctions” have been used in general biology and in ethology; Hassenstein (1976:367), a biologist and etholo­ gist himself, names as “Beispiele fur Begriffe, die derzeitig nur als Injunktionen zur hypothesenfreien Deskription dienen konnen [...]: Individuum, Pflanze und Tier, gesund und krank, Domestikation, Art (Spezies), Leben; angeborenes und erlerntes Verhalten, Spielverhalten, Drohen, Balz, Revierverhalten”.49 One would like to jump at once from these examples to those of linguistics: word-formation and syntax; subject, agent, theme; transitive and intransitive; langue and parole; competence and performance; learning and innate. But first it must become clear that “injunctions” in lin­ guistics are opposed to essentialistic definitions. Linguists ask for the essence of a phenomenon. Although Altmann & Lehfeldt (1973:21) critizise this kind of concept forma­ tion merely in relation to problems of language typology, we can nevertheless agree with them, because especially classification and comparison make the fragility of essen­ tialistic concepts visible (and because the description of a foreign language always implies the question of compar­ ability and the describing meta-language). “Kennen wir das Wesen eines Dinges oder einer Klasse, so kennen wir implizit auch all die Eigenschaften, die aus dem Wesen folgen.”50 According to this, classification (and concept formation) is only possible p er genus et differentiam “[...] wenn wir das Wesen eines Dinges kennen. Nicht sinnvoll ist es, durch logische Division zu klassifizieren, wenn wir die Beziehungen zwischen Essenz und den aus ihr herleitbaren Eigenschaften nicht angeben konnen, weil uns die Essenz nicht bekannt ist [.. .].”51 (Altmann & Leh­ feldt 1973:22) 48 “’Injunktion’ “is “a factual (= descriptive) term from descriptive natural sciences, which because o f the nature o f the objective field which it is to represent cannot fulfil the criteria o f a defini­ tion. This is especially the case 1. if a characteristic mark essential to the definition gradually melts away towards the edge of the field represented by the concept instead o f abrupdy disappearing 2. if for the definition several constituting characteristics varying independendy from each other are available, from which with the given state o f the art none can be singled out with objec­ tive reasons as the only distinguishing feature.” 49 “[...] examples for concepts which, for the moment, can only serve as injunctions for a description free o f hypotheses [...]: individual, plant and animal, healthy and ill, domestication, spe­ cies, life; innate and learned behaviour, play, threatening, court­ ship, territoriality.” Cf. regarding the topic o f “indistinct boundar­ ies” o f words in colloquial speech Hassenstein 1979. 50 “If we know the essence o f a thing or o f a class, we also implicit­ ly know all the characteristics following from the essence.” 51 “[...] if we know the essence of a thing. It does not make sense to classify by logical division, if we cannot state the relationship between essence and the characteristics which can be derived from it, because we do not know the essence [...].”

This abstinence, it is true, applies especially to the natural sciences, but in linguistics also one is far from knowing all the characteristics of language “aus denen sich alle anderen in irgendeinem Sinne herleiten lieBen”52 (Altmann & Lehfeldt 1973:22—23) in description and classifi­ cation.53 There are enough studies favourable to a dissolving of defined concepts which occur within traditions or theories. They are all from the pen of authors who have either intensively studied single non-Indo-European languages or whose aim is to make observations about a great num­ ber of languages. I myself have tried to relativize the con­ cept of the subject (qua agent plus corresponding case­ marking) and that of the ditransitive sentence (Heeschen 1990a). The principle of distribution and the possibility of summarizing which promotes case-marking, are general­ izations from the material presented here and in the above mentioned contribution. All the different pragmatically caused possibilities of structuring inherent in one single language were, long before my study, explored by Keenan (1975), by means of language comparison, for the subject, and by Hopper & Thompson (1980) and Lazard (1984) for the concept of transitivity. Keenan isolates 30 charac­ teristics for “subject”; the characteristics may be of a semantic, pragmatic or syntactic nature. In a given lan­ guage a noun phrase is a subject insofar as it possesses these characteristics to a greater extent than other noun phrases or more than for instance “a prime number” pos­ sesses these characteristics. According to Keenan (1975:312) subject is a “cluster concept” or, like many concepts in the social sciences, a “multi-factor concept”. The proximity to “injunction” suggests itself. Hopper & Thompson describe the defining properties of transitivity. It develops along the parameters of “telicity” and “punc­ tuality” of the verb, the conscious volitionality of the agent and the degree of affectedness of the object. According to them, transitivity is ultimately a discoursephenomenon and cannot be treated adequately on the level of the sentence (cf. Du Bois 1987). The grammar of Eipo will show that increasing transitivity is to be understood as clarification (for instance if the verb at the transition from a 1st to a 2nd person, which is actively given in the field of action and perception, increasingly gains the faculty to attach an object) and that “subjecthood” depends on the degree of transitivity and that, due to this dependence, it also has to represent a clarifying step (cf. chapter 8.3.2.5). The scalability of characteristics and their increasing clus­ tering, in dependence on speech, and thus on methods of clarification, has been proved by a number of authors, who have examined the marking of the agent in ergative languages or generally the pragmatics of case-marking (including wordorder).54 The universality of certain cate­ gories or their verification in theory have hardly ever been directly attacked. Dixon (1965:177) speculates about whether, without the tradition of classical antiquity, there would ever have been anything like the verb or the noun. The category of the indirect object has been questioned by Whitehead (1981—1982) with good reasons (cf. chapter 3.3.2 on the principle of distribution, Faltz 1978, W Seiler 1985), and at least for one language the category of the indirect object is considered unneccessary (Gil 1984).

When considering the preconditions of each speech act, preconditions from the history of evolution, from psy­ cholinguistics and from its dependence on discourse, and when considering the results of single descriptions and the results within linguistics, the boldness with which real utterances have been used here as a basis for a theory might appear as a timid following of long established opinions or, at most, as rushing ahead in a predetermined direction. However, while in other descriptive works the logic of the material often asserts itself against theorydependent concepts, it might occur here that the “logic of being produced” of the utterances which have been examined and will be examined by us permits of no con­ cepts adequate to description: the danger already became evident in the fact that, on the one hand, the concepts of clarification, distribution and expansion and, on the other hand, the sentence, which, as it consists of smaller, more basic units, does not permit of internal delimitations and of boundaries towards more comprehensive units, made us speak of something that had not yet been introduced into the discussion, namely of the basic units. In addition, just like in research on child language, seemingly incom­ plete categories, categories in the process of formation, have to be examined and consequendy the descriptive concepts neccessarily contain fewer characteristics in Hassenstein’s (and Keenan’s or Hopper & Thompson’s) sense, than is suggested by use. 3.3.4 Basic units. Problems Basic units or building-blocks are independent pieces of information in on-going conversation, in connected dis­ course, or in “texts”. With respect of formal syntax they are ellipses and not yet autonomous or self-sufficient sen­ tences, with respect of the principle of dependence of the symbolic field on the deictic field they are autonomous and independent. Whatever can be segmented by means of pauses (cf. Serzisko 1992) without being in need of further clarification is a basic unit. Discourse and its “increased requirements of formulation” assembles these units in clauses and sentences and transfers them from the deictic field to the symbolic field. In the language of the Eipo we find as basic units: 1. The words. By this we generally understand all that the lexicographer can list alphabetically in a pre­ liminary wordlist, when he listens to native speakers’ 52 “[. ••] from which all the others may be derived in some sense.” 53 For a good idea o f essentialistic definitions the reader should study the quotations under the headwords “adjective” and “ad­ verb” in Knobloch (1961 ff). Regarding this c f the corresponding definitions in BuBmann (1990) which enumerate characteristics in a more technical manner. It would also be interesting to follow up the essentialistic definitions o f the sentence concept which even penetrate into logical, deductive theories. Further literature regarding the headword Sat% “sentence” in BuBmann (1990). 54 Cf. chapter 3.3.1 und 3.3.2; cf. Anderson & Wade 1988, Bechert 1977, Bossong 1985, Bromley 1981, Derbyshire 1985, Feldman 1986, Foley 1986:95, 10 8 -110 , 171, Li & Lang 1979, Ochs 1982, Potts & James 1988, Sasse 1981, Schieffelin 1981 und 1985, Scott 1986, Whitehead 1981-1982.

63

talk. The possibility of whether by itself it is a suffi­ cient and comprehensible answer to a question may serve as a simple test for whether something is a word or not. As a second test one may regard the possibility whether the linguist, eliciting, establishing meanings, and searching forms, can talk with the informants about what seems to be a word. This test turns the attention to two discrepancies. Firstly: while informants have no difficulties whatsoever in talking about words and in occasionally enumerating and defining words with pleasure, linguists do not agree at all about what should be regarded as a word (cf. Heeschen 1984a, 1990c, Kramsky 1969, Mel’cuk 1976, H.J. Seiler 1964). Secondly: most languages have no word for “word”; in the Yale language, for instance, yubu and in Eipo ju p e equally mean “speech, speaking, word, utterance, sound” (cf. Heeschen 1978b). “Particles” (conjunctions, connec­ tives etc.), however, have no meaning and, according to the informants’ judgement, are ju p e teikyonok “merely and only sound”. Thus, words which only have a pragmatic or syntactic meaning (cf. Schmid 1970) do not stand out from the “Sinn-Horizont” of an utterance and are not specified. The fifth chapter will show that informants are, nevertheless, able to talk about these particles sensibly or that they quick­ ly learn to talk about them. The fact that the criteria “syntactic, semantic, pragmatic, autonomous”, which Schmid (1970) has recourse to for the pur­ pose of classifying the parts of speech, need not all be fulfilled in order to be able to talk about words in the metalinguistic discourse, can only disconcert anyone who wants to attribute a theory-dependent status to the word and does not take it for an “injunction”. 2. All words and constructions which by virtue of the speech-initiating act of topomnestic pointing can in some way be taken out of the proper sentence structure, either by left dislocation or by thematizing, are to be regarded as basic units. I would like to emphasize that “topomnestic”, although taken over from Biihler, but not restricted to Biihler’s deictic acts in a certain area, here has a wider meaning in accordance with the variety of meaning of Greek topos, which can also mean “theme, subject, topic”, especially of speech. Basic units of this kind are especially deictic particles and any kind of adverbial phrases as well as “subjects”. This compilation again shows clearly why the subjects were first taken out of the sentence structure proper. They are not to be counted among the core elements of a sentence; according to this opinion, agreement with the verb is not a means of cohesion, but a means of “ad-posi­ tioning” with the help of anticipatory or recipient deixis. 3. A special case of all these kinds of general topo­ mnestic pointing are all those forms which require clarification only in principle or in all probability. Thus the pronoun el “they” in example (59) had to be specified by ayam “rat”, or-asik “in the hamlet over there” from example (43), however, did not have to

be defined by Lape asik dam “near the Lape hamlet”. These cases are important because, in contrast with the opinions of many linguists, they identify the method of anticipatory deixis with that of recipient deixis and make it appear as the normal case. Antici­ patory deixis very well demonstrates the general movement of each utterance from the general to the particular, from pro-semantic reference to replace­ ment by words specifying the meaning, it shows clar­ ification as a filling of semantic gaps. Examples as found in chapter 3.3.2 and 8.2.2.1 are also important because in anticipation of the grammar of Eipo they can show the grammaticalization of the clarifying series: a clause beginning with anticipatory deixis indicates that it will have a subject different from that of the preceding clause. 4. All constructions of the type xy, which have to be repeated, when either x or y is varied, are to be regarded as basic units. This holds true especially, when the head is modified by more than one attrib­ ute or when one attribute belongs to more than one head. Generally, the principle of repetition applies to all noun phrases, consequently; basic units are all nouns plus one defining element. 5. All constructions consisting of the series ± pronoun + noun ± verbal noun are to be regarded as basic units (cf. examples 47-48). 6. All verbs, either as verbal nouns or in inflected form, which attach a noun, are to be considered as basic units. Which part this noun would have to play in clauses with forms of distribution or in the pos­ sible forms of summarizing will remain open. These basic units are independent insofar as they are not related to a concept of the sentence or are not to be gen erated from a starting symbol “S”. The units are primary and basic, just like the building bricks of a construction set, which can be put together more or less arbitrarily. The linear sequence of such units is held together either through the identity of the objects or events in the real field of action and perception or through the acts of topomnestic pointing in the imagined space. However, such a unit can take over quite a different function, if it appears together with other units, just as the building stone in the finished house is no longer a roughly hewn block, but has the function either of forming part of a whole wall or of being the foundation for the roof-beams. In accordance with our preconditions, however, we are not seeking for the essence of something, but for the function and for the acts of assigning function. In the same measure as the construction (of the sentence) is completed, the function of the basic units is differentiated and is specified. Incidentally, by the serialization of the units named under the first and the last point, especially when adding the function of anticipatory deixis named under point three, a great, not to say an indefinite, number of sentences in Eipo may be constructed. Indeed, most of the Eipo sentences have one of the structures resulting from this combination.

In spite of this, the grammar resulting from the basic units is rudimentary. The reality of the units results from a series of observations made in most of the different lin­ guistic disciplines. The smaller the units are and the more they coincide with the word as a unit, the closer they are to what (cf. chapter 3.3.3) we came to regard above as fundamental to lan­ guage: to the general faculty of comprehending some­ thing as a symbol. Language signs are a special case of general semiotic activity. If there is anything like a biolog­ ical programme for human speech, it consists first of all in the fundamental semiotic activity and next in the “urge” to know the names of persons and things. There are no reasons for the linking of language signs, i.e. for syntax, for which learning and social interaction would not give a sufficient explanation. Conjunctions, sentencemedial forms of the verb and the whole complex of “deixis between the clauses”, which can combine a text like (31) into one single period, are learnt.55 I have repeatedly talked about the function of language as a detour (Heeschen 1988); speakers reach their aims by the representa­ tion of facts which are just as foreign to the inner states, the wishes, problems and emotions of the speaker as they are outside the immediate interactional context: it is then that language reaches its full functionality. The transition from the first phase of language acquisition, which is still connected with primary semiotic activity and the funda­ mental faculty of comprehending something as a symbol, to the second phase, in which the “representational func­ tion of speech” and therefore the more complex syntac­ tic forms of language develop, coincides with the step from the circle of the primary social partners into the world of the playmates and into that of the “fathers and other strangers” (Gleason 1975), in which the detours, which language permits, must be used. Foppa and Gar­ den & Brown (quoted in Heeschen 1988:219) ask, why the child continues to learn speaking, if the first incom­ plete means together with those of non-verbal communi­ cation work so excellently in social interaction with the first partners. Gleason (1975:293) tried to give a first answer to this interesting and provoking question, which nevertheless did not give a new impulse to research: “[...] children have to talk to their fathers and other strangers, and these people are not tuned to them in the warm, sen­ sitive way their mothers are”. We consider the question as well as the answer as evidence for the evaluation of basic units “close to word and sign” and of more developed forms of speech, which use detours and demand repre­ sentation. Sampson uses empirical arguments against Chomsky’s innateness hypothesis. In doing this he follows the con­ siderations of the biologist and evolutionary theorist Simon. In a world undergoing constant change, in which organs with a long evolutionary history take over new functions and in which old functions are taken over by new organs, in which “evolution” potters around in trial and error and occasionally finds astonishingly provisional solutions, the living systems, whose partial systems are hierarchically structured, have a better chance of survival in selection and adaptation. For Simon (Sampson 1978:191) it is easy

“to grasp the concept that, in a universe which develops by blind piecemeal trial and error rather than in conformity to a preordained plan, a complex ensemble has a far better chance of being translated into reality if it can be built up in many stages, each of which has independent survival-potential, than if it has no stable constituents intermediate in size between the primitive elements and the whole.” Let us leave Sampson insofar as he turns against Chomsky’s conception of language as consisting of a fin­ ished, autonomous bioprogramme which Chomsky, how­ ever, considers to be inadequate from the point of view of evolutionary history. If there were no intermediate con­ stituents between the state and the individual, the individ­ ual would be lost, if the state perished. If there were no intermediate forms of organization between the faculty of articulation and the finished sentence, speaking would be lost together with the possibility of forming sentences. Even if in this study the linearity of utterances is occa­ sionally stressed, language is doubtlessly hierarchically structured. Language, into which the basic units are hier­ archically integrated, continues to function as a system which took over the “representational function” and it continues to serve as a means of communication in social interaction even, or already, if the partial systems break down or have not yet been developed (cf. further below). The basic units produce proof of their “survival poten­ tial”, even if people do not talk “in sentences” or if the problems of sentence-internal cohesion have not yet been solved (further biological arguments are to be found in Hormann 1978:41—49). There are two primary, fundamental linkages in language (modification and government) coinciding with two of the above-mentioned basic units, firstly noun + modifier and, secondly, noun + verb. For reasons to be explained further on I tend to replace “subject” or “object noun phrase” by the general term noun. Keenan sets up a “meaning-form dependency prin­ ciple”. Language comparison yields that wherever the construction noun + adjective appears and wherever a form of agreement between the two constituents prevails, it is the adjective that agrees with the noun. In addition there is semantic agreement: if we take a series of nouns and a series of adjectives and form them into construc­ tions of the kind mentioned above, one can see at once that the meaning of the adjectives depends on the mean­ ing of the noun serving as argument. Keenan calls “adjec­ tive” a functional category, so that the general conclusion is that functional categories vary with the choice of the argument and that their exact interpretation depends on this argument. There is, indeed, no agreement between noun and adjective in the Eipo language and we can only find out from the context, whether the construction is an 55 C f Piattelli-Palmarini (ed.) 1980, Slobin (ed.) 1985, Bowerman 1988.

65

attributive or a predicative one or whether it it is one simi­ lar to a relative clause (cf. chapters 7.2.2 and 8.2.1.4), but the privileged morphological and semantic agreement between the two wordclasses, which was pointed out by Keenan, has to be regarded as grammaticalization or clar­ ification of a particularly frequent linkage, and it is of no importance to us whether the linkage favours the one or the other type of construction. Of course, the linkage noun + verb forms part of the entire recent discussions about agreement and govern­ ment, on the one hand, and about the typology of wordorder and its implicative relations to other grammatical phenomena, on the other hand.56 In W. Lehmann’s (1986:6) opinion “the forces governing the meaningful relationships in the syntactic patterning known as order” are fundamental in the search for “primes in syntax”; for him these forces are either agreement or government. If one leaves aside “cross-reference”, government has “the dominant role in the clause”. The problem of govern­ ment is, of course, closely connected with that of transi­ tivity. If here we simply take over the results of the dis­ cussion on typology and of that on modification and government and only generally refer to it, it still remains worthwhile to discover the close linkage of object and verb in some marginal areas. 1. According to Sasse (1984) the “noun incorpora­ tion” in some languages accentuates monolithic blocks consisting of a verb and a usually direct object qua patient. In that case the sentence consists of two parts, a first part formed by a noun phrase functioning as “topic”, and a second formed by a verb as “com­ ment”, into which a direct object is incorporated. 2. Classificatory verbs (Lang 1975) are a further sign of the basic unit noun + verb. Wurm (1982:81), reverting to Lang’s studies, distinguishes two types of such verbs, “existential verbs, and pro-verbs in predications”. The first type consists of verbs which have to be translated by the auxiliary “be”. Each of these verbs, however, occurs only in connection with a certain, semantically defined group of nouns, for example one of the verbs only with big, tall or mas­ culine objects or beings, the other only with small, round or feminine objects or beings. In the Mek lan­ guages this type occurs rudimentarily, at the most. Thus, the Eipo can express the existence of non­ human beings and objects outside the hamlet by the verb mab- “sleep, rest, stay”, for example: ( 66) Kwitma utam

ma-lam-le.

cassowary

sleep-hab.-3sg./pres.

down/there

The cassowary lives or stays down there. (Or:) The cassowary is a being from down there (in the low­ lands). For humans, however, the verb ub- “live,be” is used:

The pro-verbs form a second type of classificatory verbs. “The predications consist of an adjunct, usually a noun, which has a specific meaning and a pro-verb whose meaning is more general, and the combina­ tion adjunct + pro-verb functions as verb-phrase.” (Wurm 1982:82) The first and the second type are complementary to each other. This second type is not very distinctive in the Mek languages either, at any rate the verb ab“do, say” preferably combines with concrete move­ ments or sounds, for example: (68, cf. chapter 7.2.4.2.1) birbir atrembling/ad.

make

tremble, have a temperature. And the verb ub- “be, live” preferably combines with words describing inner states, for example: (69) tulum ubsmoke

be

be furious. Besides, there is a great number of essential seman­ tic relations (“wesenhafte Bedeutungsbeziehungen”, Porzig 1934): a verb only combines with some few, often only with one single object, for example57: (70) maneng webgrass

cut

cut grass, work in the garden, yupe

leb-

language

speak

speak, mek kwebwater

scoop

scoop. Interestingly enough, all the types are to be regarded as “semantic” figurae etymologicae (internal objects): they block the extension of transitivity of the sen­ tence. On the one hand, the constructions are so closely connected that they often yield new idiomatic meanings, on the other hand, the nouns have retained so much from their role as object that the verbs cannot govern any other object. 3. With this we again return to the principle of speech, according to which one clause should never contain more than one piece of new information, and this means in concrete grammatical terms that, if possible, a verb should not govern more than one noun phrase. I have demonstrated that a beautiful sentence like “Fater gave Melase an adze” never

(67) Ningke Dingerkon u-lam-le. Ningke

Dingerkon

be-hab.-3sg./pres.

Ningke lives in Dingerkon. (Or:)Ningke is one (of the inhabitants) of Dingerkon.

56 Cf. Hawkins 1983. Cf. Derbyshire 1985, Foley 1986, Hopper & Thompson (eds.) 1982, C. Lehmann 1983 and 1985, Li (ed.) 1977, Plank (Ed.) 1984, Porzig 1934. 57 Cf. chapters 7.2.1.1, 7.2.4.2.1, 8.3.2.1, 8.3.2.5.

occurs in the reality of speech (Heeschen 1990a: 140-142). It is necessarily cut into pieces and “de­ composed”, positively speaking, the noun phrases are distributed among different clauses. Most of the grammars of Papuan languages, indeed, deal with ditransitive sentences, but in all the examples at least one argument, —often even two —, are replaced by ; pronouns.58 This, however, means nothing other than that the sentence cannot be understood with­ out additional information from the field of percep­ tion or from the preceding speech. It would, at most, offer a scheme in the sense of Biihler, which would have to be specified and expanded and in which the noun phrases performed would only be loosely con­ nected to the verb, in the relation of co-reference, in forms of extraction or in chains of appositions. Such chains are described by Lithgow (1976); Whitehead (1981-1982) turned against the universality of the indirect object, because it does not fall within the canonical cases of government in the Papuan lan­ guages. Bisang (1986) remarks that the Yabem lan­ guage has a strong tendency to limit the number of actants per verb to one or at most two. Lawrence (1972) describes a similar tendency in Oksapmin, and according to him, this language restricts the number of information units per sentence as much as possible. W Seiler (1985) only takes S and O for nominal core elements of the Imonda language. The examinations of Du Bois (1987) agree with this: the statistical analysis of Sacapultec Maya texts demon­ strates that the agent is often introduced in intransi­ tive sentences and that then the unit O - V contains the new information, so that 46.9% of the transitive sentences have no “lexical argument” at all, 50.3% of these sentences have only one argument, and only 2.8% have two arguments. Du Bois (1987:819) sets up a “one lexical argument constraint”. According to him, the restriction that one sentence should not introduce more than one piece of new information, i.e. not more than one “lexical argument”, also applies to some Austronesian languages, which were examined from the point of view of the discourse grammarian. Perhaps there is still another source for that princi­ ple of speech and for Du Bois’ “constraint”. Those clauses in Eipo which have no finite verb and which probably are ahead of the clause that finishes the whole period with a finite verb often contain socalled “Satzinnenformen”59 of the verb (cf. chapter 7.2.4.5); possessing something of a nominal charac­ ter they are best translated by a participle. Verbal nouns, as they occur in the examples 47—48, can also take the place of medial verb forms. If, however, one regarded these verbal forms as nouns, the rela­ tion of the objects, arguments or actants to the “verb” would not be one of government, but one of modification: the crying of a child, the cutting of the wood. One can better appreciate some of the series of clarification, when starting from the nominal character of the finite verb in the 3rd person singu­ lar, compare:

' (71 )

Yo

ob-ma-1.

wood

hit-dur.-3sg./pres.

There is a hitting of wood. Yo

ob-ma-ni-1.

wood

hit-dur.-me-3sg./pres.

There is a hitting of wood with reference to me. Lelengde arye ob-ma-ni-1. Lelengde

s.

hit-dur.-me-3sg./pres.

There is a hitting by Lelengde with reference to me. Yo

arye ob-ma-ni-1.

wood

i.

hit-dur.-me-3sg./pres.

There is a hitting by the wood with reference to me. Lelengde arye yo

dob-uka,

yo

arye

Lelengde

take-inf./suc.

wood

i.

s.

wood

ob-ma-ni-1. hit-dur.-3sg./ pres.

Lelengde hits me with a piece of wood. In this series it becomes clear that transitivity in­ creases in the same measure as more than one noun phrase is introduced; subject qua agent and instru­ ment exclude each other, “source, agent, instru­ ment” only become differentiated in sentences with distribution (cf. Clark & Carpenter 1989 and chap­ ters 4.2 and 8.3.2.3); agreement between subject and verb is brought about in the same measure as the feature “animate” is introduced. However that may be: the nominal character of the medial verb forms could to a certain extent be responsible for the fact that the expansion of the clauses is restricted. Accordingly, the same limitations would apply to the non-sentence-final clauses as to the basic units men­ tioned above under point four and five; a maximum basic unit of this kind would be his wood-cutting. I know of no discourse studies justifying the limited­ ness of this type. Wundt (1911-1912, part 2:342— 346), however, already talked of the attributive way of thinking (“attributive Gedankenform”). He found it realized in the languages of the Bushmen (i.e. the San languages). His literal translation of a story is very similar to the interlinear version of the Eipo texts; at first glance there is indeed a similarity be­ tween the languages, except for the fact that Eipo can conclude a sentence with finite verbs possessing suf­ fixes for tense, person, and number: “Busches Mann-hier da gehend laufen-zu WeiBem, er WeiBer gebend-hin Tabak, er da gehend rauchen [.. .]”60 (Wundt 1911-1912, part 2:343). 58 Cf. Collier & Gregersen 1985:155, Geary 1977:66, Lawrence 1971:119, Murane 1974:134, Ross with Paol 1978:8, Wells 1979:67. 59 The term was coined by Pilhofer 1933:35; “medial verb form” is the term used in the English literature. 60 “Of-bush man-here there going run-to white-man, he white-man giving-away tobacco, he there going smoke [..

67

1 The verbal nouns attach themselves attributively to the subject or to the anaphoric pronoun. Possibly some readers may think this remote example is anec­ dotal, which, considering Wundt’s reliabilty and astuteness, it is not. But, as far as I know, the modifi­ cation generally and the voice of the verb with regard of the phenomenon of verb serialization and clause chaining have not received the same attention as, for instance, the transitivity of sentences with verbs. Therefore with the term “attributive series” I will have to come back to the problem when treating verb serialization and clause-chaining in the gram­ mar. For psycholinguistics the clause is the unit to which plan­ ning, production, and understanding are directed. How­ ever, the actual “processing” accompanying speech passes off according to an “incremental mode”, as Kempen and Hoenkamp call it; Levelt (1989:28) goes back to these authors and to Wundt and describes the “mode of pro­ cessing” as follows: “It combines serial and parallel processing in the fol­ lowing way: Each fragment of information will have to be processed in stages, going from the conceiving of messages to articulation. Still, all processing com­ ponents can work in parallel, albeit on different frag­ ments. If the fragments are small (i.e., if the compo­ nents require little lookahead), incremental pro­ cessing is efficient, producing fluent speech without unintended interruptions. That it is sufficient for a processing component to be triggered into activity by only a minimal fragment of characteristic input was called ‘Wundt’ s principle’.”61 From this it must be concluded that the speaker plans smaller units before he sees the finished sentence “in his mind’s eye”. Consequently, a relatively autonomous status has also to be attributed to our basic units. One condition of “incremental processing” is particularly interesting: The “fragments” have to be sufficiently small. Levelt writes later (1989:125): “The requirement of incrementality [...] implies that a surface structure is, by and large, generated ‘from left to right’ as successive fragments of the message become available. Wundt’s principle requires that the generation of surface structure occur without much lookahead or backtracking, so that each surface unit produced can immediately be processed by the Pho­ nological Encoder.” The clause, indeed, remains the unit, but the speaker must be able to plan it without too much backtracking or look­ ahead. It will be permitted to say that the technique of clarification and of the attributive series, that the principle of distribution, which naturally works with few new quan­ tities each time, and that the principle of speech of intro­ ducing only one piece of new information per clause completely agree with this condition. I know of no efforts to establish a connection between results from the de­ scription of languages and the findings of research on speech production (cf. Grosjean 1983). 68

Another clue comes from a quite different direction. Kohler & Altmann (1986) cite the constructional unit of linguistic systems as an example of a “synergetic linguis­ tics”. Constructional units are texts, sentences, words, syl­ lables and sounds. The size is regulated according to Menzerath’s law as applied to the sentence: “Bevor der Satz fertig erzeugt ist, regelt sich nach seiner zukiinftigen GroBe die GroBe der Teilsatzlangen: je groBer der Satz ist, desto kiirzer sind die Teilsatzlangen.”62 (Kohler & Altmann 1986:259) We intended, however, to leave the boundary of the sen­ tence indefinite, and especially for the Papuan languages it was considered whether, in view of the grammatical means of verb-serialization and switch reference, which permit of almost endless sentences, the paragraph and the text were not the fundamental unit (Longacre 1972a and b). If the sentence is open and if the units of speech operating above the level of the sentence are long, short­ ness of the clauses and smallness of the basic units, func­ tioning as clauses in speech, arise as if automatically. The forms of summarizing also make sense: if the text or the speech approaches its end, the remaining constructional unit shrinks, and accordingly the sentences or clauses may become longer.63 As Du Bois (1987) starts from the text, their constraint in the relation of the constructional units among each other is also determined. Psycholinguistics as well as mathematical linguistics tend to attribute a certain autonomy to the basic units; they do not determine at all in which manner the construction of larger units has to be performed and into what kind of linkage, i.e. cohesion, the basic units will enter with them. It has to be empha­ sized that the abandoning of a theory-dependent sentence concept need not neccessarily be connected with the “incremental processing” and the mutual dependence of the constructional units. My theory, however, that the cat­ egories are differentiated only in the course of speech and the fact that I understand grammar as a process of “pro­ ducing” such categories in speech, does depend on the autonomy of the basic units. A wide sphere opens up when one brings together the results of research on language acquisition, on the uncon­ trolled acquisition of a second language, and of studies concerning speech contact, simplification and telegra-

61 Cf. Wundt 19 11-19 12 , part 2:251-252. 62 “Before the sentence has been produced completely, the length of the clauses is regulated according to its future length: the longer the sentences the shorter the clauses.” 63 In a series o f remarkable papers Fenk & Fenk-Oczlon (1993), Fenk-Oczlon (1983) and Fenk-Oczlon & Fenk (1994, 1995) have proved the cross-linguistic validity o f Menzerath’s law. For exam­ ple, the number of syllables is limited to 5 - 9 per simple sen­ tence, a fact which “corresponds to tirne-related limits, which might be operative at the level o f syllable perception (and produc­ tion)” (Fenk & Fenk-Oczlon 1993). In Eipo this number would correspond to a unit which consists o f a noun, a more or less complex verbal form and a connective. Generally, Menzeraths law “serves the ‘constant’ and ‘economic’ flow o f linguistic infor­ mation, avoiding an overcharge as well as a waste o f cognitive resources” (Fenk & Fenk-Oczlon 1993:23-24).

j

1 « ; j j 1

j ;

phese.64 Generally it may be assumed that the same basic units operate in the processes of decomposition and of reconstruction. I am going to pick out only a few individu­ al observations. The utterances of small children without reference to subject and of telegraphese have already been mentioned. Children first comprehend transitive verbs as intransitive ones, i.e. they interpret the noun placed in front of the verb uniformly as theme. Givon (quoted by Klein 1984:94) distinguishes between the pragmatic and the syntactic mode in syntax: both modes are differently developed in individual languages. The syntactic mode pre­ vails in written language, the pragmatic one in oral speech (and of course in the many speech communities without writing). We have met with two characteristics of the modes in examples and in the discussion: in the syntactic mode subject-predicate structures and subordination pre­ vailed, in pragmatic speech, however, theme-rheme struc­ tures and a slight grammatical linkage were predominant (cf. the principle of distribution; Du Bois 1987, Sasse 1984, Heeschen 1990a). The principle of distribution and of the successive building up, or breaking up, of longer utterances is to be found in child language: Bier trinken, Mann trinken “beer drink, man drink” instead of *der Mann trinkt Bier “the man drinks beer”, or Thea spulen, Thea auf “Thea wash, Thea up” instead of *Thea willaufspulen “Thea wants to wash up” (cf. Weir 1970:81-83, J. G. & P. A. De Villiers 1985:59—62). It is to be found in many languages with verb serialization, and this again seems to be charac­ teristic of areas with a great variety of languages and with speech contacts, which make us assume processes of assimilation and of decomposition (Papuan languages, Niger-Congo languages, cf. chapter 4.2). This, then, is the evidence for basic units, which do not correspond to what a ‘clause’ is, but which roughly corre­ spond to utterances having the size of an expanded noun phrase or that of an inflected verb governing one noun phrase and followed by a connective, and which can be produced without much lookahead or backtracking. The point is that all these structures are not to be regard­ ed as defective or elliptical. They are genuine speech struc­ tures, whose successive compounding and whose manipu­ lation for the purpose of longer utterances only give rise to something like a syntactic mode: the sentence is the result of many developments, not something which initiates a process without any precondition. - Chapters 2.3 to 2.5 described, among other things, the sociological conditions making decomposition and restructuring pos­ sible. The fourth chapter will prove, among other things, that such processes may have been possible in history, and the fifth chapter will lift a corner of the curtain behind which the manipulative processes of construction, the linkage of basic units, take place. But first the problems have to be mentioned which an “injunctive grammar” meets and which, if solved, would lift it above a wellmeant opinion and a reasonable approach. The nature of the problem may be sketched as follows. Plank (1987) draws attention to an old distinction, accord­ ing to which there are languages which mark grammatical relationships only pragmatically and as desired. Thus, to stick to our examples, in Eipo the subject qua agent is °nly marked under special circumstances, and the direct °bject is only rarely marked, for instance when the subject

and the object both have the feature “human” and when the speaker pleases to leave the subject unmarked. “In its other manifestation functional relation cod­ ing is strictly grammaticized, hence does not allow for variation according to the functional needs of particular nouns in particular uses”. (Plank 1987:177) The obligatory use of case suffixes and tense morphemes is well-known to us from the Indo-European languages. The distribution of these two groups, according to Plank, coincides with familiar typological differences. “Analytic languages and languages with agglutinative nominal morphology prefer optional marking, viz. economy of use, whereas languages with cumulative (‘inflective’) nominal morphology tend to econo­ mize systematically, by means of selective neutraliza­ tion and methodical assignment of nouns to declen­ sion classes.” (Plank 1987:178) This distinction confronts us with the problem of either being able to draw up an “injunctive grammar” only for a group of languages and to suspend it for the second group, or of acting against the rough, but accepted and plausible distinction by trying to find evidence for prag­ matical procedures and optionality also in the second group. Our own preconditions are partly in favour of the first way: the concrete speech act, the life moment of a language, should always be the starting point for general­ izations, gaining reliability by philological and hermeneuti­ cal methods. If the distinction into two groups remains, our arguments will fade on the long way to a general theo­ ry; this, however, is not detrimental to the validity of the circular arguments, as from their preconditions provision­ alness is intrinsic to them. Individualizing descriptions of languages (cf. chapter 3.2) accord with this procedure, seeking a medium position between a single fact and a general theory. Certain trends in linguistics, but certainly those in cognitive anthropology and in semiotics, support the assumption that descriptions which gain their concep­ tions in dealing with the material itself should be permit­ ted. Schegloff, Jefferson & Sacks (1977) positively invite linguists to investigate in which way individual languages deal with their problems. Lightfoot (quoted by Van Valin 1987:395) demands that the linguist should first restrict himself to quickly accessible “every day data”; only later, when the description is growing more complete, “we shall need to seek out certain kinds of less readily available facts relevant to the refinements”. What is demanded by Lightfoot is regarded as dangerous by Anderson: “For those working on ‘exotic’ languages, indeed, the primary analytic difficulty is often to locate areas 64 Birdsong 1989, Boeder 1987, Denison 1988, J. G. & P. A. De Vil­ liers 1985, Klein 1984, Klein & Dittmar 1979, Lee 1987, Mills 1985, Miihlhausler 1974, Schieffelin 1985, Sinclair-de-Zwart 1973, Slobin 1982 and 1985, Thurston 1982 and 1987, Ureland 1988, Ureland (ed.) 1985, Weir 1970, White 1989, Wode 1984, Wurm 1986.

69

of structure that are central to the language in its own terms, rather than confining their attention to those areas that can be conveniently accommodated within some antecedently given theoretical perspec­ tive” (Anderson 1984:21) Anderson demands that it should at least be guaranteed that description and abstraction fit into the framework neccessary to the discussion of typologically very differ­ ent languages in order to prevent the arising of a theory of only one special language. Against this one could argue that many of the theories current today were formed on the basis of a particular language type, namely that of English and of a few European languages and that the lin­ guist has no nobler task than to widen the perspective arising from his ethnocentric-linguistic viewpoint. It must be added that we indeed try to build in many more facts than would be admitted by the normal linguist thinking along the dichotomies prevalent since Saussure, namely data from speech and not only from language. Thus it may be concluded that the procedure consisting in individual­ izing and in the compilation of features can be justified against the scruples stated above in the same measure as it is compatible with the laws of speaking, the production of speech, and as, at least in this way, it can lay claim to cov­ ering a field of data larger than that accepted by systemic linguistics. This enlarged field of data, however, also gives rise to an objection to the too rigorous opposition of the two lan­ guage types which Plank had in mind. Let us assume that the categories of the non-agglutinative type are not sub­ ject to pragmatics and optionality. In that case the history of these languages teaches us that these categories are, firstly, multifunctional and, secondly, partly exposed to very fast change. If speaking meets with an already exist­ ing, morphologically marked category and has to use it as a “norm”, the form becomes an instrument of new func­ tions. The speakers do not work on the form, but on its semantics. In the breakdown of the case systems of some Indo-European languages or language families or in the changes of the wordorder (cf. Li [ed.] 1975b and 1977) certainly some of the same processes of decomposition and of restructuring take effect as are at the basis of the real acts of clarification, distribution and attribution of information structures. Du Bois’ (1987) “constraint” sure­ ly also applies to naively told texts of languages with fixed categories (cf. Havers 1931, Maurer 1933). The findings of psycholinguistics result from studies which mostly dealt with the speakers of Indo-European languages, and Buhler gained a great many of his insights from, and test­ ed them on, Brugmann’s (and Delbriick’s) language-his­ torical works. Even if most of the forms of extraction can be connected with the system of a language, one may ask, whether they should not also be regarded as attempts to go beyond this system, as it were, to make some blocks or “basic units” of the masonry of the system suitable for pragmatical use in the speech act (cf. Altmann 1981, Bossong 1981). —The second way out of Plank’s dilemma leads back to the “principles of language history”: what does not take effect in the individual minds (“Einzelgeister”) and their speech acts cannot find expression in lan­ guage history. The systematic connection between speech 70

act and history is indeed plausible in our material, but it is by no means proven. Independent of Plank’s contrasting of two language types, another objection would arise against an “injunctive grammar”, if we turned those languages which make a pragmatic and optional use of their grammatical means themselves into a system, whose features would include pragmatics and optionality, but which would be prescribed to the speaker and to the speech act in the same way in which “use” and “norm” would have to avail themselves of the means provided by an “economy of use” in con­ formity to the rules and to Durkheim’s law. Such systems would be stable. Their development would escape the clutches of individual speech acts. One can, indeed, to a certain extent find instances for the regular transition of systems with verb serialization to those with case-marking.65 The change would correspond to what Sapir (1921:150) called “drift”: “Language is not merely something that is spread out in space, as it were —a series of reflections in individual minds of one and the same timeless pic­ ture. Language moves down time in a current of its own making. It has a drift.” Lass (1987:162—165) describes the difficulties Sapir had with this concept. If Sapir, on the one hand, adhered to the idea that language is composed of acts of usage and formed by usage, “drift” on the other hand can not be quite autonomous, independent of usage and without a psychic substrate, even if the “drift” takes place in the course of generations and thus beyond the activity of sin­ gle speakers. If one admits the autonomy of systems with pragmatics and optionality and of “drifts”, there remains the task of looking for the psychic substrate. The possibil­ ity of a grammar as a process continues to exist, because its aim changes only insofar as it no longer examines the process itself, but rather the process of interaction be­ tween the prescriptive system and the creativity of the speakers. In view of the enormous number of system grammars it is worthwhile sounding out the space left to the creativity of the speaker. I believe that pragmatics and optionality appear in the linking of basic units into larger forms and that, in the course of an increasing distance from the basis, controlled, more or less conscious manip­ ulation can gain the upper hand (cf. chapter 3.2 and 5). It is not the general invalidity of the objection that is deci­ sive for the fact that it may be rejected, but the possibility that it can only relate to partial systems of a language. By this, however, it is only partly refuted. A final argument against “injunction” may follow from the fact that our perception is categorical (cf. Lorenz 1973).66 Applied to language: something is either good or bad, a person is either “one of us” or a stranger, some-

65 Cf. among others Bruce 1986 and 1988, Hyman 1975a, Li & Thompson 1973, Lord 1982, W. Seiler 1986a and b. 66 As deixis and the replacement of deictic expressions by specifying series as well as the filling o f gaps called ellipses were not unim­ portant in our argumentation, I would like to draw the readers attention to Auer (1988), who separates phenomena o f deixis and those o f speaking in the contex t o f the situation.

thing is either “a heap” or “a few grains” (cf. Hassenstein 1979, and only in the history of music does one speak of ‘‘a mighty handful”), something either reaches the top of the accessibility hierarchy and becomes a subject or not. \Xre do not perceive continua, and even a complex field of transitivity does not permit of a corresponding number of constructions, but makes a choice between a few contrast­ ing sentences necessary. At the bottom of this there arises the question for the describing linguist, whether the lan­ guage he is going to describe, and the conceptions he uses in describing, are to be isomorphic, like a faithful copy or a model. If language proceeds categorically and takes “leaps” in the formation of its categories, “injunctions” are a means of approaching this formation; language and the language of observation must not proceed according to the same principle, because it is necessary to differen­ tiate and to overcome the —in the original sense of the word - “preconceived” opinion. The conceptions and characterizations depending on theory lack this capacity of adequation and lie athwart the procedures of speech production; the describing linguist has to change the theo­ ry in order to adapt himself to the data (cf. chapter 3.1). Many linguistic theories regard language as an autono­ mous object reserved to the linguist. “Injunctions” make it possible to receive the features for a grammatical phe­ nomenon from other disciplines. They are helpful, be­ cause they do not ask what a certain category may be, but try to retain the transfer to a category and to study the gain in features. The describing linguist can argue about the status of a category, just as an educated person has no diffi­ culty in finding intermediate stages between good or bad or “one of us” and “strange”. A philological and individ­ ualizing procedure operating with “injunctions” represents an attempt to mediate between the possibly categorically perceived facts and the necessarily precoined language of the observer; it is at the same time the paradoxical attempt of a meta-language, which merely intends to describe and not to subsume. Theory dissolves into argumentation, it is method, a statement from which, following Garvin, we started at the beginning of this chapter. Let us again demonstrate the long, sometimes rather airy course of this third chapter with the help of a concrete utterance. A child says: da, an meinem \’U-trinken-Becher; da ist eine Wespe “there, at my to-drink-mug, there is a wasp”. The fieldworker notes down a genuine speech act and gropes in the dark, because the grammar of the sentence does not correspond to the image he has already formed of the language. The philologist is interested in the utter­ ance, because the style possibly represents the grammar of

the future. We forbid the theorist to speak of a unique act of performance or of defectiveness for the reasons offered by the fieldworker and by the philologist. We would rather induce him to ponder about the utterance so that he may discover the principle of clarification, of the filling-up of deictic expressions with semantically weighti­ er words or groups of words, and the principle of distri­ bution. We would also encourage him to abstain from explaining the construction in “zu-trinken-Becher” by complicated deletions or other derivations, but propose to him to reread the chapter about the basic syntactic rela­ tions in the “Prinzipien der Sprachgeschichte” by Paul (1960). The objection that there is no consistent theory of a synchronic state of language does not count, as we are generally interested in the laws of speech production and less in fabricated objects: what counts is what happens in the sports ground, not the rules of the game in a book. The status of “an meinem zu-trinken Becher” as sentence or clause remains to be seen; the describing linguist has to find more features before he can finally determine this status. It is part of the description to find the themerheme structure, the gesture of topomnestic deixis, it is also part of this to start searching for a subject and object, maybe in vain. The questions of why the clauses were not “summarized”, why and which smaller units, basic units, were used, and by which means cohesion is achieved lead into theory. The discovery that we are dealing with spoken language with its —compared to written language —pecu­ liarities does not get us any further, as all language was once speaking. The descriptive linguist and the theorist have to choose, whether they can and want either to dis­ miss the utterance as an individual case or to discover in it the general laws of the “production” of the grammar in speech. If they choose the first possibility, the fieldworker will be robbed of his efforts, and the philologist will have no choice but to study a hapax legomenon. Sometimes during my time of field research and especial­ ly during the many months of working on the edition of the Eipo texts, I had the impression that the language of the Eipo consisted entirely of utterances like those of a child. The fact that the texts could be translated and over­ all had a general significance, even if the meaning was sometimes a surprising one, permits of the circular argu­ ment that the translator interpreted the individual cases correctly: the understanding was a success. The linguist can resume the thread here: he will not only study “what was heard or read once”; what he heard has a significance in the context of the disciplines dealing with language and speaking.

71

4.

THE POSITION OF EIPO WITHIN THE MEK LANGUAGES AND THE OTHER PAPUAN LANGUAGES

The two following chapters do not represent exhaustive contributions to language history, to typology, or to the psychology of language awareness; the material would go beyond the scope of the prolegomena to a grammar of Eipo. The chapters only resume the ideas we discussed above, when reasons turned up for the fast changeability of the Papuan languages, the proximity of these languages to speech, and for a grammar being composed while speaking. Here these reasons are regarded from a particu­ lar perspective. In addition to this, the chapters will present some of the material which has its place neither in the grammar proper, nor in collections of texts. 4.1

Remarks on the history o f the Mek languages

History here means two things: firstly, we shall point out some differences between the Mek languages and, at the same time, probable developments and, secondly, we shall examine the possibility of proving or at least of making plausible the genetic affiliation of the Mek languages to the languages of the Trans-New Guinea Phylum, especially with the Ok languages bordering on the Mek area in the east.

Here the first and second vowel should be identical (cf. further below and Heeschen 1992a). The sequence regu­ larly contracts to unvoiced aspirated stops, but not all instances of /p/ and /k/ can be traced back to that sequence. Besides that /p/ develops into /f/ for which no longer the sequence can said to be underlying. Varia­ tion, contraction, and allophones indicate the direction of historical development. The reduction of the system also holds good for Korapun and Sela. I assume that proceed­ ing from east to west one finds an increase in the number of sound changes that have taken place. The changes of eastern /p/ and /t/, which may correspond to proto-Mek */p and */t/ examplify this increase and its direction. Original /p/ already changes in the east, compare: (72) Bime Larye Tanime

hit-dur.-l sg./pres.

I am hitting. Eipo and all western dialects lose this /p/, for example: (73) Eipo

I am hitting Yale

Table 4 System of plosives in the eastern Mek languages p b

t d

k g

This system could be similar to the system of proto-Mek plosives. Proceeding from Larye and Eipomek the systems become more and more “defective”, unstable or are reduced to one series of plosives, for example: Table 5 System of plosives in Eipo and Yale Eipo (p) t b d

k g

Yale

k -

p b

d

Eipo has retained /p/ only as a labialized /p/ in the coda position, the opposition between /k/ and /g/ is restricted to two in no way certain examples with /g/ at the onset of a syllable. In the Yale language /b/ and /d/ are mostly realized as voiceless, and /p/ and /k/ may be interpreted as a sequence of b or k + vowel + h + vowel. 72

ob-ma-n hit-dur.-l sg./pres.

4.1.1 Remarks on phonological developments o f the Mek languages 67 In the east of the Mek area (from Tanime to Okbab) we find two series of plosives, a voiced and a voiceless one, for example in Bime:

pob-ma-n [p p (eastern Mek) > pf, ph (easterncentral Mek) > f, h (central-southern Mek) > 0 (central and west­ ern Mek). Eastern and proto-Mek /t/ first changes via /s/ to /h/. The change first takes place between vowels, for example: (74) Bime Eipo

ambotum “yesterday, tomorrow” “hamlet” atei ambosum “yesterday, tomorrow” asik “hamlet”.

67 If not stated otherwise, all information is based on my own stud­ ies. These, however, were compared with Fowler & M. and J. Rule 1972, Louwerse 1978, W M. and J. E. Rule & Cutting 1972, A. & A. Sims 1982; c f Heeschen 1978a, 1985, 1992a, Heeschen & Schiefenhovel 1983. From Sela there is a “Field report No. 1. Phonology” by Jan Godschalk in my possession, from Koropun a “Kimyal-English dictionary” by Elinor Young. Both authors have not released their manuscripts yet, so that I have to use my own, more scarce data or have to revert to oral communications by Jan Godschalk and Myron Bromley. In this context the question of which allophones the plosives have can be left out o f account; thus all the dialects from Bime to Eipomek, pardy even Yale, have implosive and pre-nasalized variants.

In the west the change also takes hold of the position at the onset of a syllable and in the coda position, (75) Eipo

tong kat song kas

Yale

“smell” “hard” “smell” “hard”.

/t/ between vowels, which had become /s/, turns to /h/ in the west, compare: (76) Eipo Obahak Yale

asik fotong hik hong

“hamlet” “hair, fur” “hamlet” “hair, fur”,

According to this, the loss of the initial vowel must have come about after the change into /h/. Evidently as a reflex of this vowel occasionally voiceless vowels appear with the effect that in the Yale language the realization of the subsequent /h/ may be very different, in an individual case voiced as well as voiceless or breathy. The develop­ ment proceeds further in the west, /h/ between vowels disappears and causes the rising of new, voiceless, aspirat­ ed plosives [ph, th, kh], compare: (77) Eipo

ambosum “yesterday, tomorrow” basam “pig” dib-ting “they would eat” Fa valley baham “pig” Yale ambohum, ampum “yesterday, tomorrow” baham, pam “pig” de-heng, teng “they would eat”. Again the eastern dialects seem to be most conservative: proto-Mek *t > t (eastern Mek) > t-, -t, -s- (central Mek, Eipo) > s-, -s, -h-, -0- (western Mek, Yale).

while for new /t/ I have only found two examples hither­ to (cf. Fleeschen 1992a: 18—25). /k/ between vowels, especially labialized /k/, in the west still undergoes another, even more dramatic change: it dis­ appears and leaves behind as traces pharyngealized and laryngealized vowels, compare: (80) Eipo Nalca Keleka Eipo Yale

/b/ and /d/ are retained throughout the entire Mek area, /g/ occasionally vanishes in a mysterious manner or becomes /k/, for example: (81) Bime Eipo Yale

kape habo

“stone knife” “stone knife”, realized as ['xabd] or ['habo].

The development of a new aspirated /k/, which does not turn into /h/ as old /k/ does, probably arises from stronger friction, as described above, together with subse­ quent syllable contraction, compare: (79) Eipo Yale

ketek“to decline” kheik-, keheik- “to be hard or declining”.

The intermediate stages for this development are unknown to me. In the Yale language new /p/ has developed to /f/ since 1986, often one already hears Jam for baham, pam “pig”, and new or old /k/ alternates with either /h/ or [kx],

bob-ma-r bob-ma-1 bo-lam-la carry-dur.-3sg./pres.

he is carrying Bime, Eipo, Yale dou “dry (of weather)” Bime egin “heavy” Eipo, Yale ikin “heavy”. The change from /r/ to /!/ in the coda position (and between vowels) in the forms of “he carries” is the rule. Going from east to west, from Tanime (and all eastern dialects) to Eipo (and all western dialects), the loss of syl­ lable-initial /k/ in some important, frequently used words is not regular, for example: as still shown in the eastern Mek dialects and lan­ dob-ma-rak dob-nam-dak dob-rak dob-dak guages, merged in Eipo.119 Clusters were reduced, cf. Eipo take-dur.-3d./pres. take-fut.III-3d. take-3d./pa.I take-3d./pa.I jin , Bime gyin “bow”. In Bime I noted implosive allo[dopmafakd [dop'namdak1] [dobtiraE] [dop'daE] phones of /g/. A few remaining instances of Eipo /g/ the two of them the two of them the two of them the two of them are heard in these words: are taking will take have taken have taken. (226) /gum/ [gum] “not” However, there are also reasons for keeping the two /angun/ [^an'guin] “you (all)” sounds separate: /bagema/ [bage'ma] “necklace of orchid fibres”. 1. The difference is established in the writing system, and the Eipo like to make that difference. /f/ . is a voiceless, labiodental fricative. It occurs in syllable 2. According to their judgements /r/ in syllable codas onsets: is phonemically different from /d/. Speakers expect de-sonorization in the coda, that is /d/ > /t/, and (227) not sonorization. Final /d/ seems to be “unnatural”. /fito/ [fi'to] “earth-oven” /finfinya [fin'fi-njn] “tendril of a climber, 3. There is no variation between /d/ and the alveo­ antennae of insects”. lar vibrant. 4. The voiced plosive is a good marker for conso­ nantal strength in syllable onsets, while the alveolar flap signals the voiced release of a preceding conso­ nant, thus elegantly signalling the speakers’ attempts at creating the ideal syllable structure CV. 120

118 {d} is the dual morpheme, different tenses are expressed by other morphs, see chapters 7.2.4.3 and 7.2.4.5. 119 Cf. Bime gat “hard” geteb'-“to bundle” keteb'-“to sharpen” Eipo kat “hard” ki'tib- “to wrap” keteb'- “to sharpen”.

/(/ certainly comes from former */p/ (cf. chapter 4.1.1). The Bime language still shows [(J)] to which, occasionally, an affricate corresponds in Eipo: (228) Bime/potong/ Eipo/fotong/

[ -k*tek-ke-ak

■ (313) dob-nu-(y)ak >

te(k)-ke-ak

stand-you-3pl./ pa .1

they will take.

stand-you-3pl./ pa.I

*tek.ke.ak te.ke.ak they have shown up for you, (or:) they have stood around you.

The -^-sterns of verbs may be assimilated and deleted, when followed by a suffix beginning with /m/ onset:

> sa(k)-kwefteb-

*sak-kwe'teb-

(314) dob-m(a)-uk

fall-move/away

fall-move/away

sa.kwe.’teb*sak.kwe.'tebto depart hastily.135 (310) -1.1- > -1*kwel-il-lam-ak

he was taking.

> kwel-iQ-lam-ak create-caus.-hab.-3pl/pres.

*kwe.’lil.la.mak they create.

kwe.’li(l).la.mak

one-n.

*ton.nye someone.

to(n).nye

The last example also illustrates the regular deletion of the vowel of the -^-suffix, when the following tense-personnumber suffix starts with a vowel. Progressive assimilation occurs in (315) k > m >

ng ng

/_____n /_____k.

For example:

> to(n)-nye

one-n.

> *dom.muk > domuk

take-dur.-3sg./pa.III

create-caus.-hab.-3pL/pres.

(311) -n.n- > -n*ton-nye

> dobnyak

take-fut.I-3pl.

talekna > talengna “construction of the ^te-roof-type” bob-nam-ke-ab > bobnangkeab

*yan-nam-ne

> ya(n)-nam-ne

come-fut.III-1 sg.

come-fut.III-lsg

*yan.nam.ne I will come.

ya.nam.ne

carry-fut.III-you-1 pi.

we will carry for you The vowels of the object pronouns -ne-y -ke-y -se- can be assimilated to the vowel of the following suffix.

The vowels of the object pronouns -ne-y -ke-y -se-, that of the connective -ye “and”, and word-final vowels of utter­ ance-medial verb forms can be deleted, when they are followed by a connective or another verb form starting with a vowel peak: the sequence .CV VC. is avoided, for example:

e>

u

/____high back vowel.

For example: (316) kwoteb-nu-lulum heal-me-2pl./hort.

(312) bob-se-uk

you shall heal me

> bobsuk

carry-us-3sg./pa.III

deib-su-mun-ye

he carried for us

put-us-2pl./med.-and

fera-arye

and you having put it down for us.

> ferarye

away-by

Assimilation to -lul obligatory:

because of being far way fak-dongob-uka ab-uk-ye split-inc.-inf./sue.

> fakdongobukabukye

the third singular optative, is

(317) deib-su-lul [deipsyM ]

make-3sg./pa.III-and

after he had opened dib-nye-ora

[1y 1],

put-us-3sg./hort.

> dibnyora

he might put down for us.

eat-1 sg./med.-and

(I) having eaten dib-ma-n-ye-ane

> dibmanyane

The connective y e “and” is reduced to -i- in interconsonantal position, for instance: (318) dob-m(a)-uk-ye-buk -

eat-dur.-l sg./pres.-and-sc.

while I am eating.

> dobmukibuk

take-dur.-3sg./pa.III-and-when/ds.

A special type of deletion probably occurs in the sequence of suffixes for the immediate future tense (cf. chapter 7.2.4.3 and 7.2.4.5):

he was taking, but then another ... 135

Another way o f reducing the geminate is the derivational suffix:

sakalkm'teb- / sa.kal.kwe.'teb-/.

137

Regressive assimilation determines the choice of the vowel of the derivational suffix: it is the vowel of the pre­ ceding verb stem, and, if that vowel is a diphthong, it is the weaker member of that diphthong. (319) sak-al-kwe'tebfall-der.-move/away

to depart hastily dob-ol-bintake-der.-go

to follow. The connectives ane “while” and -irye “and” can be assimi­ lated to the preceding vowel of the verb suffix. (320) dob-m(a)-uk-une

into one morpheme what may have consisted of two mor­ phemes. While the morpheme boundaries, that is the onset of the first syllable and the coda of the final syl­ lable, are suited for maximal contrastiveness, the internal structure of these morphemes tends to be less suited for contrasting single segments. We observed that front vowels tend to be more “fronted” and that back vowels tend to be more “backed” when they occur in open syllables in pre-accent position. The same process occurs in the CV.CV or CV.CVC structures mentioned above. The two vowels join in that process fus­ ing two syllables by means of vowel harmony and sonor­ ization of intervocal consonants136: (322) /si'sibmal/

[si'sipmal]

“he is saying names, he is telling (a story)” /murub’nab/ [murup'na:p] “we will make peace”.

take-dur.-3sg./pa.III-sc.

while he was taking dob-ma-(u)pe-urye

> dobmaburye

take-dur.-lpl./pa.III-and

and when we were taking. Fusional assimilation occurs in the fusion of the future suffix -na- followed by the first plural present-future suffix -ab. This fusion leads to the only long vowel in Eipo pho­ nology: (321) buk-na-ab [mbok'na:p]

(323) /dipe/ ['dypwe] “soot” /luna/ ['lvnn] “the calling in chorus” /babese/ ['J3nj3oso] “cricket species”. 6.3.2 Characteristics o f Eipo phonology

> buk’nab

sit-fut.III-lpl.

we will sit down. Variation between intervocal -k-, -ng~, and -m- on the one hand and between intervocal -n- and -r- on the other hand is common, but I think -m- is typically Tani, while Eipo prefers -ng-, cf. Eipo tangup-, Tani tamub- “to turn”, -r- was heard in Munggona, while I heard -n- in Dingerkon, com­ pare suruk- and sunuk- “to bend”. By and large, variation between monophthongs and diphthongs is also due to dialect variation, compare Munggona weik, Dingerkon (and Tani and Bime) wik “big”. However, inter-hamlet and inter-valley variation of inter­ vocal consonants should call our attention to the fact that in phonological words having the structures CV.CV or CV.CVC the contrastive potential of the intervocal consonants is somewhat diminished or even vanishing, /t/ and /n/ are frequently flapped, /k/ is fricativized, frequently /n/ and /r/ seem to be interchangeable. Besides that, intervocal /l/, /s/, and /m/ occur. The two vowels seem to assimilate to each other, compare berek“to dawn”, batak- “to stay behind”, kunuk- “to get dark”. In such words one can no longer recognize isolatable morphemes. One begins to ask oneself whether, for example, hire “black magic” and kite “paraphernalia”, on the one hand, and kelib-, kilib- “to wipe”, kerik- “to scratch off”, kifdik- “to scratch”, kidikf- “to go to the side, to get lost”, and kisik- “to put or push aside”, on the other hand, originally had some parts in common: sonorization and vowel harmony obscure the derivational history and fuse 138

In speech vowels tend to be centralized and consonants tend to be fricativized or sonorized (see also chapter 6.4), for example:

Word-initial onsets are simple (almost no clusters) and stable. Word-final codas lose their consonantal strength: strongly voiced release turns them into CV structures. 1 believe that consonant clusters, most of them occurring in the onset of word-final syllables, develop into an addi­ tional means of signalling the end of a phonological word. Word-medial onsets tend to be sonorized, if they occur within a morpheme and if they do not demarcate morpheme breaks. The opposition between voiced and voiceless plosives is reduced to that between /d/ and /t/. In the coda the main functional load goes to /k/ and /b/, the latter one being realized as [p]. Eipo has a five-vowel system. A seven-vowel system may result from the overlap between more fronted and more raised /i/ and /e/ and between more backed and more raised /u/ and /o/. Words which do not belong to the class of verbs may have up to six or seven syllables, but most of them usually have one, two, or three syllables. Verb forms may have up to twelve syllables, but most of them usually have two to four syllables. The accent is realized through pitch and intensity. Length is dissociated from the means of giving phonemic promi­ nence to a syllable, but it is a means of giving non-phonemic prosodic prominence to pre-accent, post-accent,

136 Regressive assimilation determining the vowel o f the derivation­ al suffix (cf. above in this chapter, example [319]) could be a spe­ cial case o f vowel harmony; the intervocal consonant is sonor­ ized, for example /k/ in (319/ is realized as [x].

and word-medial and word-final accented syllables. The place of the accent is either + initial or - initial. In the lat­ ter case it moves to the end of the phonological word. CVC syllables may stop this movement: they attract the accent. The accent is partially phonemic. It forms no clear instances of minimal pairs in the domain of words which do not belong to the class of verbs. It forms minimal pairs in the domain of verbs. Tones can be said to be allophones of the accent, if two-syllable words become mono-syllabic. This may heighten the awareness of speak­ ers for tonal structures. Intonation may override the accent patterns. I think that one feature dominates the phonology of Eipo: wherever possible, speakers attempt to create CV syllable structures and to create equally weighted syllables. The means of balancing the weightedness of syllables are as follows: compensatory lengthening of non-accented syllables, fronting and backing —decentralizing —vowels in open syllables and in pre-accent position, placing the accent and, optionally, length on syllables, whose peak is a central vowel, and shortening accented syllables. Words which do not belong to the domain of verbs can easily avoid CVC syllables by opening these closed syllables. The verbal morphology requires a certain number of CVC sequences (verb stems and typical suffixes like -lam-, -mal etc. have the structure CVC). Consequendy, verbs cannot move the accent to the more weighted syllable, because several “more weighted syllables” make their appearance. This may have led to assigning a fixed place to the accent in the domain of verbs. The dominance of CV syllables, or, at least, the perma­ nent attempts at creating such syllables, conforms to what is known of syllable-timed languages (cf. Auer & Uhmann 1988).137 Other features are more or less in line with the syllable-timed type. Firsdy, accented and non-accented syl­ lables show no different structures. Secondly, the syllable boundaries are well demarcated, as far as word-initial onsets and word-final codas are concerned, but boundar­ ies within a morpheme are not so well demarcated. Third­ ly, the vowel system is stable; some connectives, the deriva­ tional suffix, and, when the intervocal consonant is sonorized, the vowels in CV.CVC structures show initial stages of vowel harmony. Fourth, when word-final vowels precede connectives, vowels are deleted in order to avoid the sequence of Vpeak (qua coda) and Vpeak (qua onset) and to create CV sequences. Fifth, the accent is not a means of grammar, but solely one of differentiating lex­ emes. The Eipo data support the opinion that the type of isochronism of a language dominates, or, at least, influ­ ences characteristic features of the segmental and suprasegmental system and that the type of isochronism is not an epiphenomenon which merely follows from these fea­ tures.

6A ■Transcribed text The following text is the beginning of a narrative which was published in full length in Heeschen (1990b:220-223). The passage is presented four times: 1. in the orthography proposed in chapter 6.1.1 and used throughout this work (note the use of the hyphen, which indicates morpheme breaks, and that of brackets, which enclose deleted segments or seg­ ments to which no gloss is given), 2. in a rather broad phonetic transcription above which the intonation is iconically indicated, 3. in glosses which translate the lexemes and analyse the meaning of the bound morphemes, and 4. in a free translation. Kwengkweng (Enus) Nabyal: Bol korun-nang yangan-m-ik Sik-de Langde bisik dara, V_

___ r

~

~~~\_

'sikde 'laqdo j3i'sik

jaqan'mik

they-but Langde

come-dur.-3pl./pa.III t.

way

Langde arye-ub-uka

a-ya-(y)ik

dara,

.laqd9

ar^u^u.Ka

aja'jek

dam

Langde

from-be-inf. / sue. here-come-3pl./pa.III t./ds.

“tokwe-fum-nang yakan-m-ak,”

winyab-uka-ab

to%olfumlnar)]

ja'KanmaxJ^jajloKap

earth-in-people

come-dur.-3pl./pres. say-inf./suc.-with

kote

2.

dara

sak-a-n-m-ik.

ko're

sakanmik

rock

fall-make-rep.-dur.-3pL/pa.III

Kote sak-uka ab-ik-ye,

sik asik-nang are

ko're ,saKoka'j3ikb

stiti a'siknaq aro

rock

they

fall-inf./suc. ma.ke-3pl./pa.III-and

basam ob-uka ab-ik-ye, ba'sam ofto'%a-f5iki9 pig

sik walwal arye _ __/ \ tjik wal'wal affi>

kill-inf./suc. make-3pl./pa.III-and they

imbik __ ?im'bik

a'renmik

raw

give-rep.-dur.-3ph/pa.III

hamlet-people t.

ignorance

by

are-n-m-ik.

137 Louwerse (1978:75, cf. Steinbring 1989) writes: “The phonologi­ cal word in Una has features of word-timed rhythm. Both when in isolation and within a sentence there is a tendency for words to take the same length o f time.” My data are not in line with this suggestion. The duration o f two syllable words is constantly around 400 milliseconds, that o f three syllable words between 600 and 800 milliseconds.

139

ipp^

3.

Imbik

are-n-m-ike,

ukwe daline

gum,

*Wbik a'renmiqo

'ukwo da'lino

qom

raw

fire

give-rep.-dur.-3pl./pa.III

imbik are wik kobrob, sik-de

kwaning-toto

imbik 'am JJi'kofterop’’ 'sikdo

kwa'mrjtoto

raw

sweet/potato-pl.

t.

big created

they-but

bace-toto

bo-ban-m-ik.

lmba?tb't5to

bo'^a-nmik

pitpit-pl.

4.

cook/put/vn. not

3. They did not cook it in fire, the raw things were numer­ ous, they brought sweet potatoes and pitpit.

dare, sik-de saboka

bo'jPa-nmiktaRQ

'sitb ab, parang

and

ab, berase

l?ikanapn qha'J3akapn paraqap"

ab

'“ bira-saap"

4. After bringing it, they (the white people) gave tobacco, fish, steel-axes, bush-knives, and rice, and having done so, they (our people) said: “Rice, that is the primordial form of the east (that is, of nassa headbands, the most valued decoration), it is tabooed in this world here, it is the prim­ ordial form of the east, being created for the foundation of this world here, we cannot eat it.” Having said that, they went and threw it away.

“berase are yalye dei,”

5. Having thrown it away, they said: “As to the bushknives, they are the east’s foundation of our world here, it is forbidden for us.” And they threw them away.

a.reb^d'ar9

mbi'ra-saro

jaAo/ei

6. They did the same with the steel-axes.

give/vn.

rice

east

fish

and steel/axe and bush/knife and rice

arebne

dare, t./ds.

winya-n-m-ik,

t.

“a-motokwe

and

ground

mem, a-motokwe

win'janmek

'amotokw^maem

'arnotok^

say-rep.-dur.-3pl./pa. I ll

here-mountain

here-mountain

kwemdina

taboo

kobrob yalye dei,

nun-de dine mtncto 'dins

Xwem'dine qhofkogjarorei foundation

5.

ab,

sa^uKBp1

carry-go-dur.-3pl./pa.III t./ds. they-but tobacco

ikan ab, kapak

1. They (the white people) came from Langde, and having come from there, they came here, and they (the people here) said: “The people from inside the earth are coming,” and they rushed down the rocks. 2. Having rushed down, the people of these hamlets here, having killed pigs, they gave raw (pork to the white people), because they did not know (because they did not know how to handle the strangers).

carry-go-dur.-3pl./pa.III

Bo-ban-m-ik

(The arrival of) the White people

created

east

ground we-poss. eat/vn.

mem,”

winyab-uka donok-ol-ban-m-ik.138

maem

nja(3xKU

dono'xolJJanmik

taboo

say-inf./suc.

throw-der.-go-dur.-3pl./pa.III

Donok-ol-ban-m-ike, “parang are a-motokwe donoxol'J3anmiko

pa'raq

ara'modokwo

throw-der.-go-dur.-3pl./pa.III bush/knife

t.

here-mountain

kwemdina yalye nun mem,” winyab-uka-ab 'qPemdinajata

nun maem pin'jaPuKa'p^

foundation

we

east

taboo

say-inf./sue.-with

bai-donok-o-n-m-ik.

baidAno,%onmek] outside-throw-be-rep.-dur.-3pl./pa.III

6.

Yalea

wini-n-m-ik.

ja'le-a

Jpi'nmmik

steel/axe

do/thus-rep.-dur.-3pl/pa.III 138 Note in this sentence the treatment o f /p/ in the Indonesia11 loanwords kapak “steel-axe” and parang “bush-knife”.

140

7.

WORDCLASSES. MORPHOSYNTAX

The more than 5 600 words or lexemes of the main entries in the Eipo dictionary (Heeschen & Schiefenhovel 1983) and in the additions (Heeschen 1990b:401-4) are distribut­ ed among the following wordclasses (cf. chapter 3.4): nouns adjectives verbs adverbs postpositions and connectives personal pronouns interrogative pronouns and adverbs spatial deictics roots interjections unclassified

2 515 183 2 600 78 64 9 18 32 28 13 155

Detailed semantic, morphological, syntactic, and pragmat­ ic characteristics of each wordclass will be given in the following chapters. As may be guessed from the sheer number of nouns and verbs, only these two classes are productive in the Eipo language, all other classes being closed and only consisting of a limited number. In present day Eipo there is not much word-formation, by means of which a word may be shifted form one class to another. Historically, such processes of word-formation must have been common, as is testified by some word families. Verbs and nouns were apparently formed out of noun-like roots by means of reduplication, derivation and compounding (see chapters 7.2.1.3.1 and 7.2.1.3.2), for instance: fu 4'headwaters’’ is probably connected with fub- "to fall out, to let fall, to fall apart”, f uk- "to straighten” (the areas of the headwaters are flat),^^ - “to break openy\fum "hole, opening, valley”,furume "the essen­ tial part, the basis, the origin”,furwe "drop, splash”. kwa "sth. spread or forked, fork” is probably con­ nected with kwab'- "to go away, to spread, to untie”, 'kwak- “to divide, to squeeze apart, to copulate”, kwakman "bifurcation”, kwak'- "to get cracked”, kwangnab- "to adopt, to take as one’s own child”, kwanik- "to curl”, kwakwa "butterfly species”, kwalye "banana”, kwanim "left” (also in the sense of awk­ ward), kwaning "sweet potato”, kware "shy, timid” (cf. kwab'-), kwaya "bend, curvature”, kwate "vagina”, kwatema "snake species, rainbow”. In "sth. even, flat, down or low” is probably connect­ ed with lub'- "to rub in, to smoothen”, Ink- "to hol­ low out, to open” (and by means of that "to lay bare”), lulu “surface, sth. which has a flat surface”, lumub- "to string sth., to spread out”, lumun "in one row”. Synchronically by far the larger part of new words is formed by verb compounding (see chapters 7.2.4.2.1 and 7.2.4.2.2) and the simple conjoining of two nouns as in asik kata "village ground”, from asik "hamlet” and kata "place” (see chapter 7 .2.1 .3.2). Besides that, new objects

are easily designated by juxtaposing a noun and a verbal noun as in mourn barimna “brassiere”, from mourn “breast (of woman)” and barib- “to support”. If we take into account only derivation, adverbs as well as nouns and verbs are not a closed, though rather a semi-productive wordclass. In the class of adverbs I have lumped together sentential adverbs, locationals, and temporals. A good number of adverbs are typically derived from a noun to which a particle or a postposition is suffixed, for example baytam “in the forest, outside”, from bai “forest” and -tam “side, at the side o f”. Such formations fade into syntax (cf. chapter 7.1.4.6). Within the class of nouns about 300 verbal nouns can be found. Though all verbs can form a verbal noun, only these 300 are listed in the dictionary. These not only assume syntactic functions as a kind of sentence-medial verb (see chapter 7.2.4.6), but their use has become fixed outside these limited syntactic functions and they occur in constructions typical of a noun. Such nouns are, for example, koine “the planting”, from ’koub- “to plant”, kwemdina “myth, creation, tale of origin”, from kwebreib“to originate from, create”, nakina “sickness”, from nakin“to be sick”. They occur in positions which are usually filled by human agents or nouns referring to objects: (324)

Nakina

taleb-ma-ni-1.

[sickness] s

seize-dur.-me-3sg./pres.

Sickness seizes me. (Or:) I am sick. Kwemdina [creation] Q

fetereb-nam-ki-n. explain-fut. Ill-you-1 sg./pres.

I will explain the tale of origin to you. Within the class of verbs about 450 words of doubtful classification are hidden. I will call them adjuncts (see chapter 7.2.4.2.1). They typically consist of a reduplicated root or stem, but other forms similar to verbal nouns or participles also show up. They co-occur with the verbs ab“to say, to make” and ub- “to be, to become”, for example (325) Birbir

an-ma-ni-1.

shivering

make-dur.-me-3sg./pres.

I shiver. (Or:) I have fever. 4 Adjuncts” seem to be an open class. They could. be

attributed to the class of nouns, but inevitably they would form a mixed subclass, thus I decided to leave their status open. In my overall classification they form part of the class of verbs, being the carriers of meaning in what seem to be auxiliary verb constructions. In this grammar interjections will not be treated separate­ ly: some of them are entirely outside the system (cf. chap­ ter 6.1.5), the remainder can be subsumed under the class of sentential adverbs (see chapter 7.2.3). With the exception of postpositions and connectives (and the bound morphemes of the spatial deictics) all words can form a single utterance: (326) Make. bird

That is a bird. 141

good

(327) kam sibman

That is good. (Or: That is well done.)

life

Teleb.

Yateanye? who

Who? (Or: Who is coming, who has done this? etc.) Na. I That is me. (Or. That is mine, I have done it. etc.) These words are basic units of longer utterances and dis­ course (see chapter 3.3.4). We do not assume that utter­ ances as found in (326) are elliptical. All utterances, not only those given above, rely on the universe of gestures, social interaction, and on the whole context of situation: these are means by which the sense is mediated. Grammar is the strategy by which, step by step and according to the principle of additional clarification, those single basic units lose their dependency on that universe and are built into the more or less autonomous structures which we call sentences. Speakers are highly aware of words (see chap­ ter 5), furthermore, words can be isolated from coherent utterances and are subject to metalinguistic discourse. The bound morphemes of the verb are subject to quick com­ ments, semantic clarifications, corrections, self-correc­ tions, and paraphrases, and thus betray a high level of awareness. Speakers also isolate postpositions and con­ nectives. The psychological reality of this wordclass can be deduced from the speaker’s ability to construct mean­ ingful examples around them and to switch from con­ densed grammatical structures to meaningfully elaborated examples (see chapter 5.2). If we presume that a general symbolic activity of human beings, their “drive” to assign meaning to objects and to the constellation of objects, underlies and, seen from an evolutionary point of view, precedes connected language, the phenomenon the linguist has to look at is the question of why and how symbols are conjoined and serialized. Certainly, some wordclasses correspond to the world we live in and to the way we perceive that world. They relate to living beings, objects, states, and events. Other wordclasses serve the function of what makes language unique compared to other systems of symbols: the possibility of conjoining symbols. Nouns and verbs belong to the first group, postpositions, connectives and the bound mor­ phemes following the verbstem to the second. Between these two groups, adjectives and adverbs may assume an intermediate and oscillating position. They form single utterances, although their names suggest that they typical­ ly occur in positions governed by another wordclass. The problems of wordclasses seem to arise not from an over­ all classification as given at the beginning of this chapter, but from subclassification and overlap between single classes. These problems are language-specific, and the problems of the Eipo language arise from the fact that some words, particularly those classified as adjectives and adverbs, occur in just one idiomatic expression or in just one construction. Sometimes only the weak criterion of distribution leads to classification. Thus sibman “creating the base or permanence” only occurs —as far as I know — in the expression 142

creating/permanence

creating a long or lasting life. The position of sibman is that of an adjective in predica­ tive function. If the word is derived either from 'sib- “to harvest sth. unripe” or sib'- “to keep ready for use” by means of the suffix -man, it should be a noun. Classifica­ tion is dificult or even impossible, because only few words show this pattern of formation and because the expres­ sion is only found in stories related to ideas about life and death, which prevent the interpreter from moving from concrete meanings to more abstract ones. Such problems as well as overlap, subclassification, and the possibility of unproved conversion are inherently given with language Classification, then, is a means of summarizing some characteristic features without any theoretical claims. In the chapters 7.1.1 to 7.2.3 I will treat wordclasses whose common feature is that they are not inflected. As I regard words as the basic units of speech, I will not treat separately some bound morphemes which co-occur with these wordclasses either in derivational or seemingly syn­ tactic processes. That is to say, I will not assign to them a separate descriptive level. I will start this part with the description of those classes which are closed and which only contain a small number of items, but which are, sta­ tistically and in respect of syntax, the most frequent ones in discourse. The last section (7.2.4) of this chapter is devoted to the verb and its inflection. Certainly inflected verbs, too, are basic units, but their capacity of forming sentences is known from previous grammars. 7. / Words requiring clarification Reference and extension of the words and morphemes treated in the following chapters (7.1.1—7.1.4) systemati­ cally depend on, and receive their semantic interpretation from, either the speech situation itself or the linguistic context: spatial deictics, pronouns, and interrogatives are most often clarified by proper nouns and, to a lesser extent, by adjectives, adverbs, and verbs. They refer to the fixed and known (or actually not yet known) points of the geographical and social surroundings, either to the scene or to the agents. Connectives and postpositions relate utter­ ances to each other and represent condensed automatized grammatical structures, which, as means of interclausal deixis, are subject to semantic clarifications in the same way as the true deictic categories are. However, they are not basic units as defined in chapter 3.3.4; they are related to basic units, if they can be systematically paraphrased and clarified by lexical means. The semantics of nouns, adjectives, adverbs, and verbs is the meta-language of all these junctional means (cf. chapter 5.1). All of them are systematic abbreviations, that is to say, the essence of grammar, and they replace extended utterances in which the speakers name mountains, rivers, objects located in space, human agents, and events as well as actions. Gram­ mar relates and condenses these acts of naming.

7, /. 1 Spatial deictics We have stressed and will do so again (chapters 2.4 and 7.2.1.1) that the Eipo orient themselves in their mountain­ ous environment by an incredibly dense network of names for mountains, hills, slopes, rivers, and plains, and by distin­ guishing four deictic points of reference relative to ego’s position. Eipo speakers mainly use the spatial deictics as a condensed and abbreviating structure in face-to-face-communication: here the deictics are accompanied by a point­ ing gesture. During acts of orientation in the actually per­ ceived space and in their transposition into stories, that is, in reported speech, spatial deictics are common and fre­ quently constructed with verbs of movement and transac­ tion. Both systems, that of names and that of deictics, rely on each other. In case of misunderstanding the speakers easily fall back on the more or less intrinsic systems of ref­ erence. Clarification and expansion insert the deictics into the textual structure, a process during which the deictics assume anaphoric function. In striking contrast to the closely related Yale language, the Eipo deictics do not develop into a system of determiners or into a system of demonstratives, almost obligatorily constructed with nouns (cf. chapters 4.1.2 and 9.9). Anaphoric use remains rudi­ mentary. This is the basic set of the spatial deictics: aeiou-, uor-, er-

here up there down there across there.

bai

a-ba-lam-se.

here-sit-hab.-lsg./pa.III outside from/here-gohab.-lsg./pa.III

I lived in this hut, I was going from here into the forest. Marikle-nang lukenyan or-yan-ma-se-ak, night

from/across-come-dur.-us3pl./pres.

a-mab-ma-lam-buk. here-sleep-dur.-2sg./pres.-when/ds.

During the night the Marikle people come to us from across (the valley) there, while you are asleep here. El Talim motokwe u-bin he Talim

mountain

up-way

u-bisik

ya-nam-ne.

go-fut.III-3sg./med. down-way come-fut.III-lsg.

If she (a spirit woman) goes on the way up there, I will come along the way down there. Din a-gum ub-lye-ora,

wa

ba-lam-ak.

food here-not be-3sg./med.-then garden go-hab.-3pl./pres.

If there is no food here, they will go into the gar­ dens. In “texts”, as “Deixis am Phantasma”, and prefixed to nouns which are not suitable for being accompanied by a gesture, the deictics quickly serve other functions, but note that these uses are neither frequent nor obligatory, a“here” prevails in linkage as well as in demonstrative and anaphoric uses. The system is reduced, because precise directionality seems to be incompatible with the uses which tie the deictics into the textual structure. A-is found in linkage constructions, which, indeed, represent a spe­ cial case of anaphoric use: Imde asik

woman

Nimdamkone

(328) Aik a-bu-lam-se,

Marikle-people

(329) Ei-bisik bi-nam-lye,

(330) Nimdamkone kil

These bound morphemes most frequently combine with verbs, nouns, postpositions and predicativizing suffixes. They combine with finite and infinite sentence-medial verbs, constantly referring either to actually perceived or to previously mentioned space, irrespective of the tenses of the verb, slightly adjusting their meaning according to whether the verbs are stative or directional, for example:

hut

The deictics are prefixed above all to nouns, occasionally to adjectives and adverbs, for example a-motokwe “the mountain here, our living-area”, er-asik “the hamlet across there”, a-teleb “this here is good”, a-yuk “in another way here”. Some more extended utterances are as follows:

Imde

hamlet

Talim asik

fanab-m-uk-e

start-dur.-3sg./pa.III-and Talim

hamlet

yan-m-uk.

A--yan-m--uk-e,

come-dur.-3sg./pa.III

here-come-dur.-3sg./pa.III-and

Dingerkon ya- (y) uk. Dingerkon

come-3sg./pa.III

The Nimdamkone woman started from the Imde hamlet and came to the Talim hamlet. Having come here (having arrived there in Talim), she came and arrived at Dingerkon. Anaphoric and demonstrative uses are necessarily mix­ ed, because the deictics cannot do without the nouns (or at least one noun forming part of the utterance) they refer to. This is evident in concluding remarks. Yupe “speech” and kil “woman” in the following utterances must be repeated and cannot be replaced by a pronoun. Thus, the hearers of a speech may say: (331) Nun a-yupe

gekeb-nam-ab.

we

hear-fut.III-lpl.

here-speech

We will bear in mind this speech.

o-lul.

down/there-go/vn. be-3sg/hort.

Perhaps, he may have gone and still lives down there in the area of the Talim hamlet.

Or having talked about a woman, the speaker may con­ clude as follows:

Ou-bobre

ba-min!”

(332) A-kil

ara, sisin

kil.

down/there-carry/inf./sim.

go-2sg./imper.

here-woman

t.

woman

name/vn.

Go and carry it down there! 143

As to this woman, she is a big woman (or leading woman or one who speaks or says the right things). The transposition into a place other than that of the speech situation, that is the “Deixis am Phantasma” in imagined spaces, is illustrated in the following passage. Far down in the#lowlands, which the Eipo do not know, a man watches a spirit woman and tries to hide himself: (333) A-kil

ara, a-yanga-lam-lye-ak-da,

here-woman

t.

here-come-hab.-3sg./med.-at-but

a-tek-am-lul.

“A-ei-am-ki-n!”

here-stand-per f.-3sg./hor t.

here-see-perf.-you-l sg./pa.I

winyab-lul. say-3sg./hort.

As to this woman here, she may have come to the place where he might have been standing. “I have seen you here (or there)!” she may have said. The demonstrative use prevails in the following passage which resumes previously given explanations: (334) A-motokwe ara ninye mem,

a-ninye wiliba,

here-mountain

here-man border

t.

man

forbidden

a-kilape,

sal

yangane mem,

wiliba

here-woman/pl.

man

come/vn.

border

ate

a-yupe

ara mem!

because

here-speech

t.

forbidden

forbidden

As to this mountain, it is forbidden for man, this here is the border of man (beyond which no human beings are supposed to live), it is forbidden for these men and women (whom we have talked about), be­ cause it is the border, it is forbidden to even mention this talk! In comparisons the Eipo speakers quickly point to some­ thing taken out of the immediate, perceived environment in order to explain constellations of objects or of partici­ pants which are presumed not to be understood by the hearer, it is a “demonstratio ad oculos” of things to be ima­ gined: (335) Aike irikna a-ub-ma-le-to-ak,

ou-tonun

hut

down-as

edge

here-be-dur.-3sg./pres.-as-at

li-am-ik-ye-ak

aik dike139

put/into-perf.-3pl./med.-and-at

hut

food

ou-deli-lam-ak. down-put-hab.-3pl./ pres.

They put away the food at one edge of the hut, at a place which is similar to this one here (the speaker points to something), in a similar way they have put down there (things into a stringbag). Some idiomatic expressions should be mentioned: ou-tenen kanye “base way of thinking, lit. down thought mind”, outoto “base, mean”, ei-monok- “to be silent or motionless up 144

there (in the face), to veil one’s feelings”, u-kolaib- “to beat down there, to castrate”. The Eipo language prefixes d- to the basic set of deictics. This prefix intensifies the meaning, indicates longer dis­ tances or sharper contrast. This d- also shows up in some pairs of conjunctions: are and dare, ora and dora, arora and darora (see chapter 7.1.4). In the Yale language it regularly forms the plural of the deictics, cf. nimi ane “this man here” and nimi dane “these men here” (cf. Heeschen 1992a:15—16). The prefix set is as follows: da-

here (in a wider area around the speaker and hearer, here and there) deivery far up there (across the mountains) douvery far down there dor-, der- very far across there (across the ridges in the next valley). Dei- and dou- of this set quite often take the suffixes -da, -de. This suffix seems to be the same as contrastive -de (see chapter 7.1.4). Deide and doude are more like directional adverbs, since they are not bound to a following noun or verb. The woman of (333) enters the scene as follows: (336) Doude

“Ibuke” winyab-ra-nin yangan-ma-1.

very/far/down Ibuke

say-refl.-inf./sim. come-dur.3sg./pres.

Somewhere very far down there she is coming, say­ ing to herself: “Ibuke (calling the names of pigs)”. The southern side of the central range, the area of Larye, is always dou- or doude “very far across there”, and the far away hunting-grounds of the Eipo are dei- or deida “very far up there”. The feature “contrast” shows up in this example: (337) An yuk

asik

you

hamlet

alone

a-ub-na-lyam, here-be-fut.II-2sg./hort.

der-motokwe

bi-nam-ab.

very/far/across/there-mountain

go-fut.III-lpl.

nun-da we-but

You alone should stay in this hamlet here, but we will go to the mountain very far across there. While the features “very far away” and “contrast” are compatible with dei-, dou-, and der-, this is not so with da- “here”, because what is not here, should be some­ where up, down, or across there; da-, then, assumes a variety of meanings “here and there, from here to there, at the periphery of something that is close to the deictic centre”.140 Very often it is prefixed to verbs as in Yale motokwe da-binmaminye babye! “even if you go to the eastern mountains from here to there”. What seems to represent pure deictic use, assumes another meaning in paraphrases of the informants. Perhaps the translation of the third of the following examples comes close to the original meaning of da-: 139 dika or kwit dika is the string o f foods connecting the novices with the times o f creation. 140 For some idiomatic uses cf. Heeschen 1982:86-87

(338) Da-buk-nin

bin-ma-n.

here/there-sit-inf./sim.

go-dur.-lsg./pres.

The woman seized something down there, the men seized something here, and the novices seized some­ thing up there.

I am going, sitting here and there. (Or: I am restless, I am exhausted.) Wasumurye-nang Lim kon da-u-lam-ik. Wasumurye-people

Lim

top

here/there-be-hab.3pl./pa.III

The Wasumurye people lived here and there (that is: permanently or for a long time) on top of the Lim mountain. El asik he

hamlet

noitam da-une middle

Sometimes it is very important to know the relative posi­ tion of participants and the spatial situatedness of events. The -/^-formations play a crucial role in describing such constellations.141 Two men have approached two women in order to have sexual intercourse with them; the women effectively defend themselves (note the construction with outam, which is isolated from the utterance): (341) Kilape

gum.

here-be/vn. not

He does not live precisely here in the middle of the hamlet.

“yob-si-lye-ak

we

tenen,

yo

sal

think/vn. man

The most common postposition attached to the basic set of deictics and to dei- and dou- of the derived set is -tam “side, at the side of, to the side o f”. The compounds are free morphemes qua words and could be classified as directional adverbs (cf. chapter 7.2.3), were it not for the fact that they also appear in ad-nominal (adjective-like) position and in predicative constructions. Their meaning, however, is strictly tied to the afore-mentioned sets of deictics, and, indeed, no difference between, for example, a- and atam “here” or dei- and deitam “very far up there” can be found. Eitam ninye “the people of (the area) up there” illustrates attributive construction, and in bisik are atam “as to the way, it is here” the deictic is in predicate position. The compounds can be accompanied by a point­ ing gesture and many uses seem to be restricted to utter­ ances, which are isolated or, so to say, not yet woven into the textual structure, for example: (339) atam betinye-nang, eitam

betinye-nang

here

two-people

two-people

up/there

here, there are two of them, up there, there are two of them. They would become syntactically integrated into whole utterances, if the preceding example were turned into betinye-nang a-yanmarak, betinye-nang ei-yanmarak or into betin­ ye-nang atamyanmarak, betinye-nang eitamyanmarak “two of them come here, and two of them come from above”. In the last paraphrase the compounds take the position of directional adverbs, which finally becomes their privileged position in “texts”. One further example may be suffi­ cient, which at the same time shows that the use of these deictics in a “text” is indicative of a vivid style, still in need of clarifications or of comparisons taken from the imme­ diate environment and still subject to turning the paratactic clauses into a chain by means of infinite verb forms and conjunctions: (340) Kilape

outam

talebrongob-ik, sal

woman/pl.

down/there

seize-3pl./pa.III

atam

man here

talebrongob-ik, kwit-nang eitam

taleb-ik.

seize-3pl./pa.III

seize-3pl./pa.III

initiation-pl.

up/there

nun song”

woman/pl. copulate-us-3sg./med.-at/sub.

outam,

kilape

down/there

woman/pl. tree

weleb-uka

lin

lab-reib-dik.

climb-inf./sue.

urine

urinate-put-3d./pa.III

dislike

dub top

The women, thinking: “We detest that one (of them) may have sexual intercourse with us,” these women climbed up to the top of a tree and urinated, upon the men down there. Other compounds are much less frequently used, are re­ stricted to single speakers or certain communicative gen­ res (cf. chapters 2.3 and 2.4), and do not establish other meanings. Instead of -tam as the last component, we occa­ sionally find -tanik “side, to the side of (obsolete, form of the eastern Tani dialect”, (--)ton “one, and, well”, (~)ak “at”, and in songs (-ajbasak and -de, -da. Only once I heard asenun “here-like, as if being (already) here, easily”142. The morpheme -to-, found in all comparative conjunctions (see chapter 7.1.4.7), is suffixed to the deictics, for example Katau eito “like the Katau mountain up there”. A mor­ pheme -(y)ubuka, probably from ub- “to be” (cf. aryubuke in chapter 7.1.4.5), is rarely attached to the primary set of deictics, its meaning being “from”, for example: (342) Katau eiyubuka

yakan-m-ak.

Katau

come-dur.-3pl./pres.

up/there/from

They come from the Katau mountain up there. The basic set of deictics also takes the predicative —or predicativizing suffixes -te or, most often, -teba, for example ateba “here (it is)”.143 The compounds either con­ stitute single utterances, for example, the answer to the question where something is may be orteba “it is across there”, or, in “texts”, they are isolated from the remainder 141 For extended examples see Heeschen 1990b:208, line 8 -17, and 368, line 24-29. 142 On -se- see below in this chapter. 143 I doubt whether I was right in stating that the “past tense o f -teba: -tebuk yields, together with the four deictic particles, an­ other set o f expressions adding the feature ‘prevlously men­ tioned in discourse’ to the spatial references” (Heeschen 1982:85). Except for atebuk, no unelicited examples are at hand. Atebuk itself is the conjunction ate or ateb, to which -uk “only” is added. Yet it is likely that the deictic term developed into that conjunction, cf. chapter 7.2.3.1.

145

of the utterance, possibly accompanied by a pointing gesture, or they take the sentence-final position, the pre­ ceding construction probably being subordinated to them. In that position they resume the reference to location in the preceding utterance, giving a shade of emphatic meaning to the utterance as a whole. Their main use, how­ ever, is in on-going face-to-face-communication, during which pointing gestures occur, much being understood from the context of situation. The first and second exam­ ples attempt to show the isolated position, the third illus­ trates the sentence-final position, which is almost exclu­ sively taken by ateba. This construction is typical of songs. The -/^Az-formations often repeat, or “agree” with, the deictic reference in the preceding utterance: (343) Outeba,

Dabotam yo basam u-lam-le.

down/there lowlands

wild

pig

be-hab.-3sg./pres.

Down there, the wild boars live in the lowlands. Or-yanga-lam-le,

orte. across/there

It is across there, he usually comes from across there. Dingerkon din

are-lam-ik-ye

ateba.

Dingerkon

give-hab.-3pl./pa.III-and

here/it/is

This is how they gave food to the Dingerkon hamlet. The -/