The Tungusic Languages 9781317542797, 1317542797

Table of contents :
The Tungusic Languages
Contents
List of Tables
List of Contributors
Preface and Acknowledgements
Technical Notes
Abbreviations
Chart of the Manchu Script
Language Map
1. Tungusic as a language family
2. Early Far Eastern sources on Tungusic
3. Early Western sources on Tungusic
4. Proto-Tungusic
5. Jurchen
6. Written Manchu
7. Siberian Ewenki
8. Orochen
9. Solon
10. Neghidal
11. Ewen
12. Oroch
13. Udihe
14. Nanai
15. Ulcha
16. Uilta
17. Spoken Manchu
18. Sibe
19. Sociolinguistic aspects of Tungusic
20. Tungusic in time and space
Index

Citation preview

THE TUNGUSIC LANGUAGES The Tungusic Languages is a survey of Tungusic, a language family which is seriously endangered today, but which at the time of its maximum spread was present all over Northeast Asia. This volume offers a systematic succession of separate chapters on all the individual Tungusic languages, as well as a number of additional chapters containing contextual information on the language family as a whole, its background and current state, as well as its history of research and documentation. Manchu and its mediaeval ancestor Jurchen are important historical literary languages discussed in this volume, while the other Tungusic languages, around a dozen altogether, have always been spoken by small, local, though in some cases territorially widespread, populations engaged in traditional subsistence activities of the Eurasian taiga and steppe zones and the North Pacific coast. All contributors to this volume are well-known specialists on their specific topics, and, importantly, all the authors of the chapters dealing with modern languages have personal experience of linguistic field work among Tungusic speakers. This volume will be informative for scholars and students specialising in the languages and peoples of Northeast Asia, and will also be of interest to those engaged with linguistic typology, cultural anthropology, and ethnic history who wish to obtain information on the Tungusic languages. Alexander Vovin (†), Directeur d’études, École des hautes études en sciences sociales (EHESS), Paris, France. José Andrés Alonso de la Fuente, Associate Professor, Institute of Linguistics, Translation Studies and Hungarian Studies, Jagiellonian University, Kraków, Poland. Juha Janhunen, Professor Emeritus of East Asian Languages and Cultures, University of Helsinki, Finland.

ROUTLEDGE LANGUAGE FAMILY SERIES Each volume in this series contains an in-depth account of the members of some of the world’s most important language families. Written by experts in each language, these accessible accounts provide detailed linguistic analysis and description. The contents are carefully structured to cover the natural system of classification: phonology, morphology, syntax, lexis, semantics, dialectology, and sociolinguistics. Every volume contains extensive bibliographies for each language, a detailed index and tables, and maps and examples from the languages to demonstrate the linguistic features being described. The consistent format allows comparative study, not only between the languages in each volume, but also across all the volumes in the series. The Germanic Languages Edited by Ekkehard Konig and Johan van der Auwera The Turkic Languages, 2nd Edition Edited by Éva Csató and Lars Johanson The Indo-Aryan Languages Edited by George Cardona and Dhanesh K. Jain The Austronesian Languages of Asia and Madagascar Edited by Alexander Adelaar and Nikolaus Himmelmann The Tai-Kadai Languages Edited by Anthony V. N. Diller, Jerold A. Edmondson and Yongxian Luo The Iranian Languages Edited by Gernot Windfuhr The Khoesan Languages Edited by Rainer Vossen The Munda Languages Edited by Gregory D.S. Anderson The Celtic Languages, 2nd Edition Edited by Martin J. Ball and Nicole Müller The Sino-Tibetan Languages, 2nd Edition Edited by Graham Thurgood and Randy J. Lapolla The Indo-European Languages, 2nd Edition Edited by Mate Kapović

The Mayan Languages Edited by Judith L. Aissen, Nora C. England, and Roberto Zavala Maldonado The Languages of Japan and Korea Edited by Nicolas Tranter The Mongolic Languages Edited by Juha Janhunen The Oceanic Languages Edited by John Lynch, Malcolm Ross and Terry Crowley Language Isolates Edited by Lyle Campbell The Bantu Languages, 2nd Edition Edited by Mark Van de Velde, Koen Bostoen, Derek Nurse, and Gérard Philippson The Semitic Languages, 2nd Edition Edited by John Huehnergard and Na‘ama Pat-El The Dravidian Languages, 2nd Edition Edited by Sandford B. Steever The Uralic Languages, 2nd Edition Edited by Daniel Abondolo and Riitta-Liisa Valijärvi The Tungusic Languages Edited by Alexander Vovin, José Andrés Alonso de la Fuente, and Juha Janhunen

Other titles in the series can be found at https://www.routledge.com/Routledge-LanguageFamily-Series/book-series/SE0091

THE TUNGUSIC LANGUAGES Edited by Alexander Vovin, José Andrés Alonso de la Fuente, and Juha Janhunen

First published 2024 by Routledge 4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Third Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 selection and editorial matter, Alexander Vovin, José Andrés Alonso de la Fuente, and Juha Janhunen; individual chapters, the contributors The right of Alexander Vovin, José Andrés Alonso de la Fuente, and Juha Janhunen to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Vovin, Alexander, editor. | Alonso de la Fuente, José Andrés, editor. | Janhunen, Juha, 1952– editor. Title: The Tungusic languages / edited by Alexander Vovin, José Andrés Alonso de la Fuente, and Juha Janhunen. Description: Abingdon, Oxon ; New York, NY : Routledge, 2023. | Series: Routledge language family | Includes bibliographical references and index. Identifiers: LCCN 2023004787 (print) | LCCN 2023004788 (ebook) | ISBN 9781138845039 (hardback) | ISBN 9781032463940 (paperback) | ISBN 9781315728391 (ebook) Subjects: LCSH: Tungus-Manchu languages. | LCGFT: Essays. Classification: LCC PL450 .T86 2023 (print) | LCC PL450 (ebook) | DDC 494/.1—dc23/eng/20230501 LC record available at https://lccn.loc.gov/2023004787 LC ebook record available at https://lccn.loc.gov/2023004788 ISBN: 978-1-138-84503-9 (hbk) ISBN: 978-1-032-46394-0 (pbk) ISBN: 978-1-315-72839-1 (ebk) DOI: 10.4324/9781315728391 Typeset in Times New Roman by Apex CoVantage, LLC

CONTENTS List of tables

xiii

List of contributors

xvi

Preface and acknowledgements Technical notes

xviii xx

List of abbreviations

xxii

Chart of the Manchu script

xxiv

Language map

xxvii

 1 Tungusic as a language family Juha Janhunen Ethnic nomenclature 2 Data and sources 4 Internal taxonomy 7 Typological profile 10 Grammatical framework 14 References and further reading 15

1

 2 Early Far Eastern sources on Tungusic Alexander Vovin Chinese sources 19 Korean sources 21 Japanese sources 23 References and further reading 24

19

 3 Early Western sources on Tungusic José Andrés Alonso de la Fuente Types of sources 27 Eighteenth century 28 Nineteenth century 31 References and further reading 33

26

 4 Proto-Tungusic Juha Janhunen Data and sources 36 Vowel system 37 Vowel harmony 41 Consonant system 43 Morpheme structure 44 Word formation 47 Number and case 52

35

vi Contents

Numerals 55 Pronouns 57 Person marking 60 Verbal stems 63 Verbal forms 65 Syntax 70 Lexical structure 71 References and further reading 72  5 Jurchen Daniel Kane and Marc Hideo Miyake Data and sources 77 Notational conventions 79 The writing system 79 Origin of the Jurchen script 84 Segmental structure 87 Nominal forms 93 Pronouns 95 Numerals 95 Verbal forms 96 Lexicon and language contacts 98 References and further reading 100

76

 6 Written Manchu Alexander Vovin Data and sources 104 Segmental structure 106 Phonotactics and morphophonology 109 Word formation 111 Number and case 116 Numerals 120 Pronouns 121 Participles 123 Converbs 125 Complex verbal forms 128 Imperatives 130 Syntax 131 Lexicon and language contacts 132 List of primary sources 134 References and further reading 134

103

 7 Siberian Ewenki Juha Janhunen Taxonomic status 140 Dialectal division 142 Data and sources 143 Segmental structure 146 Phonotactics and morphophonology 148 Word formation 150 Number and case 154 Numerals 158 Pronouns 160

139

Contents vii

Person marking 163 Verbal forms 166 Verbal functions 171 Syntax 173 Lexicon and language contacts 176 References and further reading 179  8 Orochen Lindsay J. Whaley Data and sources 185 Segmental structure 185 Phonotactics and morphophonology 187 Word formation 188 Number and case 190 Adjectives 192 Numerals 193 Pronouns 194 Person marking 195 Other word classes 196 Verbal morphology 197 Participles 198 Imperatives 200 Other modal forms 201 Converbs 201 Syntax 203 Lexicon and language contacts 204 References and further reading 205

184

 9 Solon Sangyub Baek Data and sources 207 Segmental structure 208 Phonotactics and morphophonology 210 Word formation 211 Number and case 213 Numerals 216 Pronouns 217 Person marking 218 Verbal morphology 220 Participles 220 Imperatives 222 Converbs 223 Phrase structure 226 Sentence types 227 Lexicon and language contacts 229 References and further reading 230

206

10 Neghidal Sofia Oskolskaya Data and sources 235 Segmental structure 236 Phonotactics and morphophonology 238

234

viii Contents

Word formation 240 Number and case 242 Numerals 245 Pronouns 246 Person marking 248 Verbal morphology 249 Participles 250 Finite tense and aspect 251 Imperatives 252 Other modal forms 253 Converbs 255 Syntax 257 Lexicon and language contacts 257 References and further reading 258 11 Ewen Andrej Malchukov Dialectal division 261 Data and sources 262 Segmental structure 262 Phonotactics and morphophonology 264 Word structure and word classes 266 Word formation 266 Number and case 268 Adjectives 272 Numerals 272 Pronouns 273 Person marking 274 Other word classes 275 Verbal morphology 276 Participles 278 Finite tense and mood 279 Converbs 280 Phrase structure 282 Sentence types 283 Passive and causative 285 Complex sentences 286 Lexicon and language contacts 288 References and further reading 289

260

12 Oroch Shinjiro Kazama Taxonomic status 295 Data and sources 297 Segmental phonemes 298 Phonotactics and morphophonology 299 Word formation 300 Number and case 301 Adjectives 306 Numerals 307

294

Contents ix

Pronouns 308 Person marking 310 Other word classes 311 Verbal morphology 312 Voice and aspect 313 Participles 315 Imperatives 317 Other modal forms 318 Converbs 319 Syntax 321 Lexicon and language contacts 323 References and further reading 324 13 Udihe Elena Perekhvalskaya and Juha Janhunen Data and sources 327 Segmental structure 329 Phonotactics and morphophonology 332 Word formation 333 Number and case 336 Adjectives 340 Numerals 341 Pronouns 342 Person marking 345 Other word classes 346 Verbal morphology 347 Voice and aspect 349 Participles 350 Tense and mood 352 Converbs 354 Complex predicates 356 Syntax 357 Lexicon and language contacts 359 References and further reading 359

326

14 Nanai Shinjiro Kazama Data and sources 365 Segmental structure 367 Phonotactics and morphophonology 369 Word formation 370 Number and case 372 Adjectives 377 Numerals 378 Pronouns 378 Person marking 380 Other word classes 381 Verbal morphology 383 Voice and aspect 385 Participles 386

364

x Contents

Finite indicative forms 389 Imperatives 391 Other modal forms 391 Converbs 393 Syntax 395 Lexicon and language contacts 398 References and further reading 399 15 Ulcha Shinjiro Kazama Taxonomic status 408 Data and sources 409 Segmental structure 410 Phonotactics and morphophonology 411 Word formation 412 Number and case 413 Adjectives 416 Numerals 417 Pronouns 417 Person marking 419 Other word classes 419 Verbal morphology 420 Voice and aspect 422 Participles 423 Finite indicative forms 424 Imperatives 425 Other modal forms 426 Converbs 428 Syntax 429 Lexicon and language contacts 432 References and further reading 433

407

16 Uilta Toshiro Tsumagari and Yoshiko Yamada Data and sources 437 Segmental structure 438 Phonotactics and morphophonology 440 Word formation 441 Number and case 443 Numerals 444 Pronouns 445 Person marking 446 Verbal morphology 448 Participles 448 Finite indicative forms 449 Imperatives 450 Converbs 450 Phrase structure 451 Sentence types 452 Complex sentences 454

436

Contents xi

Lexicon and language contacts 455 References and further reading 457 17 Spoken Manchu Veronika Zikmundová and Gao Wa Dialectal division 464 Data and sources 465 Segmental phonemes 466 Phonotactics and morphophonology 468 Word formation 470 Number and case 470 Numerals 472 Pronouns 473 Verbal forms 474 Complex predicates 475 Sentence types 476 Complex sentences 477 Lexicon and language contacts 479 References and further reading 480

463

18 Sibe Veronika Zikmundová Data and sources 484 Segmental structure 485 Morphophonology and phonotactics 487 Word formation 488 Number and case 489 Numerals 490 Pronouns 491 Verbal forms 492 Complex predicates 495 Syntax 496 Lexicon and language contacts 498 References and further reading 498

483

19 Sociolinguistic aspects of Tungusic Nadezhda Mamontova Demographic background 501 Legislational status 503 Data and sources 503 Ewenki as a supraregional language 504 Ewen in northeastern Siberia 506 The Amur-Sakhalin region 506 Tungusic languages in education 508 Tungusic languages in the public sphere 512 Future prospects 513 References and further reading 514

501

20 Tungusic in time and space Juha Janhunen Data and sources 518 External relationships 520 Areal position 525

517

xii Contents

Protohistorical setting 527 Stages of expansion 529 Sic transit gloria 531 References and further reading 532 Index

538

TABLES 1.1 1.2 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 5.1 5.2 5.3 5.4 5.5 5.6 5.7 6.1 6.2 6.3 6.4 6.5 6.6 6.7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 8.1 8.2 8.3 8.4 8.5 8.6 8.7

Names of the Tungusic languages Classification of the Tungusic languages Proto-Tungusic vowels Tungusic vowel correspondences Proto-Tungusic vowel harmony Vowels in non-initial syllables Proto-Tungusic consonants Proto-Tungusic consonant clusters Proto-Tungusic personal pronouns Proto-Tungusic person markers Components in Jurchen characters Strokes in Jurchen characters Variation in Jurchen characters Syllables in Ming Jurchen readings Jurchen vowels Jurchen consonants Jurchen case markers Manchu vowels Manchu consonants Manchu vowel harmony Manchu case markers Manchu personal pronouns Manchu converb markers Manchu complex verbal forms Siberian Ewenki short vowels Siberian Ewenki long vowels Siberian Ewenki consonants Siberian Ewenki vowel combinations Siberian Ewenki case markers Siberian Ewenki personal pronouns Siberian Ewenki person markers Siberian Ewenki finite functions Orochen vowels Orochen consonants Orochen vowel harmony Orochen case markers Orochen personal pronouns Orochen person markers Orochen imperative forms

3 9 38 39 41 42 43 45 58 61 81 82 82 83 88 89 94 107 108 110 117 121 125 128 146 147 147 149 157 161 164 171 186 186 187 191 194 195 200

xiv Tables

8.8 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 10.1 10.2 10.3 10.4 10.5 10.6 10.7 11.1 11.2 11.3 11.4 11.5 11.6 12.1 12.2 12.3 12.4 12.5 12.6 12.7 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 14.1 14.2 14.3 14.4 14.5 14.6 14.7 14.8 15.1 15.2 15.3

Orochen converb markers Solon vowels Solon consonants Solon vowel harmony Solon case markers Solon personal pronouns Solon person markers Solon imperative forms Solon converb markers Neghidal short vowels Neghidal long vowels Neghidal consonants Neghidal case markers Neghidal personal pronouns Neghidal person markers Neghidal converb markers Eastern Ewen vowels Ewen consonants Ewen case markers Ewen personal pronouns Ewen person markers Ewen converb markers Oroch vowels Oroch consonants Oroch case markers Oroch personal pronouns Oroch person markers Oroch participles Oroch converb markers Udihe short vowels Udihe long vowels Udihe consonants Udihe case markers Udihe personal pronouns Udihe person markers Udihe participles Udihe converb markers Nanai vowels Nanai consonants Nanai case markers Nanai personal pronouns Nanai person markers Nanai participles Nanai imperative forms Nanai converb markers Ulcha vowels Ulcha consonants Ulcha case markers

202 208 209 210 214 217 219 222 223 236 237 238 242 246 248 255 263 264 268 273 274 280 298 298 302 308 310 315 319 330 330 331 337 342 345 350 354 367 368 373 378 381 387 391 393 410 411 414

Tables xv

15.4 15.5 15.6 15.7 15.8 16.1 16.2 16.3 16.4 16.5 16.6 16.7 17.1 17.2 17.3 17.4 17.5 18.1 18.2 18.3 18.4 18.5 19.1 19.2 19.3

Ulcha personal pronouns Ulcha person markers Ulcha participles Ulcha imperative forms Ulcha converb markers Uilta vowels Uilta consonants Uilta vowel harmony Uilta case markers Uilta personal pronouns Uilta person markers Uilta converb markers Spoken Manchu vowels Spoken Manchu consonants Spoken Manchu case markers Spoken Manchu personal pronouns Spoken Manchu verbal forms Sibe vowels Sibe consonants Sibe case markers Sibe personal pronouns Sibe verbal forms Decline of the ethnic languages Ethnic languages at school Language teaching by the region

418 419 423 426 428 439 439 441 443 445 446 450 466 467 471 473 475 485 486 490 491 494 502 508 511

CONTRIBUTORS José Andrés Alonso de la Fuente Associate Professor, Institute of Oriental Studies, Faculty of Philology Jagellonian University, Kraków, Poland Sangyub Baek Associate Professor, Center for International Relations Muroran Institute of Technology, Muroran, Japan Gao Wa Postdoctoral Researcher Central University for Nationalities, Peking, PR China Juha Janhunen Professor Emeritus of East Asian Languages and Cultures Department of Languages, University of Helsinki, Helsinki, Finland Daniel Kane (1948–2021) (1948–2021) Professor of Chinese, Department of International Studies Macquarie University, Sydney, Australia Shinjiro Kazama Professor of Linguistics, Graduate School of Global Studies Tokyo University of Foreign Studies, Fuchu, Japan Andrej Malchukov Außerplanmäßiger Professor of Linguistics, Department of English and Linguistics Johannes Gutenberg University, Mainz, Germany Nadezhda Mamontova Newton International Fellow, School of Geography, Earth and Environmental Sciences, University of Birmingham, UK Visiting scholar, Cultural History and European and World History, University of Turku, Finland Marc Hideo Miyake Independent Researcher ‘Aiea, Hawai‘i, USA Sofia Oskolskaya Researcher, Department of the Languages of Russia, Institute for Linguistic Studies Russian Academy of Sciences, St. Petersburg, Russia

Contributors xvii

Elena Perekhvalskaya researcher, Department of the Languages of russia, institute for Linguistic studies institute for Linguistic studies, rAs, st. Petersburg, russia Toshiro Tsumagari (1951–2020) Professor Emeritus, Hokkaido university, sapporo Director, Hokkaido Museum of northern Peoples, Abashiri, Japan Alexander Vovin (1961–2022) Directeur d’études, Centre nationale de la recherche scientifique (Cnrs) École des hautes études en sciences sociales (EHEss), Paris, France Lindsay J. Whaley Professor of Linguistics and Classics, Department of Linguistics Dartmouth College, Hanover, new Hampshire, usA Yoshiko Yamada Associate Professor, College of Liberal Arts Muroran institute of technology, Muroran, Japan Veronika Zikmundová Deputy Head, institute of Asian studies, Faculty of Arts Charles university, Prague, Czechia

PREFACE AND ACKNOWLEDGEMENTS This volume is a survey of Tungusic, a language family which is seriously endangered today, but which at the time of its maximum spread just a couple of centuries ago was present all over Northeast Asia, with the speakers of one particular Tungusic language, Manchu, even ruling over several neighbouring regions, including, most importantly, China. Manchu and its mediaeval ancestor Jurchen are also important historical literary languages, while the other Tungusic languages, around a dozen altogether, have always been spoken by small local, though in some cases territorially widespread, populations engaged in traditional subsistence activities of the Eurasian taiga and steppe zones and the North Pacific coast. Following the example of other volumes in the Routledge series on language families, including, in particular, those on Turkic (1998, 2021) and Mongolic (2003), this volume offers a systematic succession of separate chapters on all the individual Tungusic languages, as well as a number of additional chapters containing contextual information on the language family as a whole, its background and current state, as well as its history of research and documentation. All contributors to this volume are well-known specialists on their specific topics, and, importantly, all the authors of the chapters dealing with modern languages have personal experience of linguistic field work among Tungusic speakers. Technically, all chapters on the individual languages follow a roughly uniform template, starting with an ethnolinguistic introduction and a history of documentation, and proceeding through phonology, morphophonology, phonotactics, as well as derivational and inflectional morphology to a brief description of syntax, lexicon, and external contacts, as well as a list of references. The general emphasis in all chapters is on the synchronic structural description of the target language against a diachronic and comparative background. Since this is the first comprehensive handbook of Tungusic in English, it has been an important task to create and establish a generally acceptable standard of notation and terminology. In this respect, the chapters have been unified during the editorial process. In spite of the unified format of the descriptive chapters, the authors have had the freedom to focus on the topics they consider important. For instance, although the general focus of most chapters is on phonology and morphology, some chapters offer relatively more information on syntax. The ordering of the data in the individual chapters also varies to some extent, reflecting the preferences of the authors and the specific characteristics of each individual language. The amount of diachronic and comparative information likewise varies, though such information has also been added to some chapters during the editorial process as far as required by the general framework. The editors are convinced that diachrony is an essential clue to understanding synchrony, and it is a particularly important tool when the facts concerning an entire language family are to be summarized.

Preface and acknowledgements xix

This book has been very long in the making. Originally commissioned to Alexander (Sasha) Vovin soon after 2003, a first attempt to collect manuscripts from authors was completed around 2010. However, this attempt did not result in publication, and most of the texts written then were ultimately published elsewhere, or have been rewritten and included in the present volume in the form of updated versions. Due to his involvement in other projects, Sasha Vovin subsequently invited José Andrés Alonso de la Fuente to be a coeditor and the active contact person coordinating the correspondence with the authors. However, in late 2020, the two initial editors asked Juha Janhunen to complete the actual editing work along lines similar to the volume on Mongolic, which he had edited earlier. Due to the prolonged editorial process, many changes have taken place in the original plan, including the choice of the topics, the ordering of the chapters, and even the selection of the authors. Most importantly, and most sadly, three of the original authors have passed away over the years. Most recently, Sasha Vovin (1961–2022) himself died only several months before the completion of the project, while Tsumagari Toshiro (1951– 2020) and Daniel Kane (1948–2021) also passed away without having seen the final versions of their texts. However, the chapters of both Kane and Tsumagari have been completed by their coauthors Marc Hideo Miyake and Yamada Yoshiko, respectively. As far as possible, all chapters have been brought up to date by the authors and the editors. As editors we express our thanks to all authors for their valuable contributions and cooperation, as well as for their patience. Many other colleagues have also provided us comments and answered our questions in the process of our work, including, but not limited to Uwe Bläsing (Leiden), Tom Eriksson (Vilnius), Stefan Georg (Bonn), Michael Knüppel (Göttingen), Li Baowen (Peking), Alfred F. Majewicz (Poznań), Matti Miestamo (Helsinki), I. N. Novgorodov (Yakutsk), Mehmet Ölmez (Istanbul), William Rozycki (Bloomington), Fresco Sam-Sin (Leiden), Marek Stachowski (Cracow), Giovanni Stary (Venice), and Jerzy Tulisow (Warsaw). Our thanks go also to the Routledge staff Andrea Hartill and Iola Ashby for their patience, understanding and smooth cooperation. The editorial input of Juha Janhunen has been supported by the Laboratory Program for Korean Studies through the Ministry of Education of the Republic of Korea and Korean Studies Promotion Service of the Academy of Korean Studies (AKS-2016-LAB-2250004). 10 December 2022 José Andrés Alonso de la Fuente (Cracow, Poland) Juha Janhunen (Helsinki, Finland)

TECHNICAL NOTES Material from spoken languages is transcribed in the present volume in a unified system based on the phonemic application of the basic Roman letters, complemented by a selection of special symbols and diacritically modified letters. The general principle is that one letter corresponds to one phonemic segment. Distinctive long vowels, diphthongs, and vowel sequences, as well as consonant geminates, are written as sequences of two letters. Otherwise, sequential notation is used only in a few special cases which are open to alternative phonemic analyses (such as the different phonation types of vowels in Udihe). The correspondence of the letters to their phonetic values is approximate and varies from language to language. For a more exact description of the underlying sounds in each given language, the International Phonetic Alphabet (IPA) is applied when necessary. Most Tungusic languages have a relatively simple consonant system organized in terms of four places of articulation: labial, dental, palatal, and velar, and comprising up to four nasals written as m n ñ ƞ, two sets of stops phonemically characterized as “strong” vs. “weak” and written as p t c k vs. b d j g, two fricatives written as s (dental sibilant) and x (velar to laryngeal), two glides written as w (labial, in some idioms with a dentilabial realization) and y (palatal), and two liquids written as l (lateral) and r (vibrant/trill). The palatal stops c j are in most languages pronounced as affricates (in some idioms also with dental values). Additional consonants present only in a few languages include f (labial fricative with both bilabial and dentilabial realizations), š (distinctive palatal sibilant), q ɢ (distinctive postvelar or uvular stops), as well as h (distinctive postvelar to laryngeal fricative or spirant). The vowel systems show more variation. In the simplest form, as attested in several Tungusic idioms, the system comprises two rounded back vowels written as o u (midhigh vs. high), two unrounded central or back vowels written as a e (low vs. mid-high), and one unrounded front vowel written as i (high). It is important to note that the letter e is here used for a central or back vowel quality, corresponding to the phonetic definition of “schwa” [ə]. The corresponding front vowel [e] is written, when necessary, as é or iĕ. Other vowels, occurring as distinctive segments in several languages include ʊ ɪ (lowered counterparts of u i) and ɵ (a higher or centralized counterpart of o), as well as ä ö ü (palatal counterparts of a o u). In reconstructions, the letter ï (the velar counterpart of i) is also used. As far as written languages using a non-Roman script are concerned, established transcription or transliteration systems are used for Written Manchu (Möllendorff-Norman), Written Mongol (Balk-Janhunen), Mandarin Chinese (Pinyin, mainly without tones), Japanese (Hepburn), and Korean (either McCune-Reischauer or Revised Romanization, depending on the author), with a consideration of individual preferences in the case of personal names. For the Manchu script, see also the separate chart in this volume. For Russian, a mixed system (with c ch sh shh zh kh for ц ч ш щ ж х and è e ë ya yu y i i iï ’ ” for э е ë я ю ы и й ии ь ъ) is used. All bibliographical references are, however, provided

Technical notes xxi

in the original script, as are the Chinese characters (mainly in unsimplified form) for relevant terms and names in the chapter texts. Note that many modern Tungusic languages have one or several written standards based on either Cyrillic or Roman letters. Any linguistic data from these languages is, however, quoted in the unified notation as used in this volume. Language data, including reconstructions, Romanizations from Chinese, Japanese, Korean, and Russian, as well as tentative “readings” from non-alphabetic scripts (as in the case of Jurchen written in Chinese characters) are quoted in italics, while Romanized Manchu and Mongol are quoted in boldface. Reconstructions are marked by the asterisk (*), and “readings” are marked by the dagger (†). In some cases, the radical symbol (√) is used to indicate lexical roots. Since the transcription of the spoken languages is intended to be basically phonemic, phonemic slashes (/ /) are used only in specific cases when required by the discourse. Phonetic data in IPA are given in square brackets ([ ]), while angle brackets (〈 〉) indicate graphic segments (letters) and sequences. The less vs. greater than signs (< >) stand for the direction of diachronic evolution, and the arrows (← →) show the direction of borrowing. The hyphen (-) indicates morpheme boundaries, the equal sign (=) boundaries of clitics, the plus sign (+) word boundaries in compound words, and the ampersand (&) the boundary between reduplicated elements. Glosses are indicated by the single quotes (‘ ’). In some cases connective segments or etymological morpheme boundaries are indicated by the period mark (.), while the colon (:) surrounded by a space on both sides indicates paradigmatic relationships. The ethnic nomenclature has been unified throughout the volume, as explained in Chapter 1. For grammatical terminology, the principle has been to use, in the first place, terms already established in Tungusic studies, but since many features lack an established term, especially in English, this volume also uses a number of innovative terms, especially for features for which the Tungusic languages themselves exhibit no uniform picture (as in the systems of converbs). Some new terms have also been adopted from mainstream linguistics. All terms are included in the List of abbreviations. For historical datings, the abbreviations BZ (Before Zero = BC) and AZ (After Zero = AD) are used.

ABBREVIATIONS 〈1〉 〈2〉 〈3〉 〈abl〉 〈acc〉 〈add〉 〈adj〉 〈adn〉 〈adv〉 〈aff〉 〈ag〉 〈al〉 〈alt〉 〈and〉 〈ant〉 〈anticaus〉 〈aor〉 〈appr〉 〈assoc〉 〈augm〉 〈aux〉 〈caus〉 〈cl〉 〈coll〉 〈com〉 〈comp〉 〈compl〉 〈conc〉 〈cond〉 〈conf〉 〈conj〉 〈conn〉 〈conneg〉 〈cont〉 〈contr〉 〈coop〉 〈cop〉 〈corr〉 〈csec〉

first person second person third person ablative accusative additive adjectival adnominal adversative affirmative agentive alienable alternative andative anterior anticausative aorist apprehensive associative augmentative auxiliary causative clitic collective comitative comparative completive concessive conditional confirmative conjunctive connective connegative continuative contrastive cooperative copula corrogative consecutive

〈ctemp〉 〈ctxt〉 〈cv〉 〈cx〉 〈dat〉 〈deb〉 〈del〉 〈der〉 〈des〉 〈desid〉 〈dest〉 〈dim〉 〈dir〉 〈disc〉 〈disj〉 〈dist〉 〈distr〉 〈dn〉 〈dv〉 〈dx〉 〈e〉 〈elat〉 〈emph〉 〈equ〉 〈evid〉 〈excl〉 〈exist〉 〈expr〉 〈ext〉 〈fem〉 〈fin〉 〈foc〉 〈fut〉 〈hab〉 〈hort〉 〈hyp〉 〈gen〉 〈immed〉 〈imp〉

contemporal contextual converb case suffix dative debitive delative derivative designative desiderative destinative diminutive directive discursive disjunctive distal distributive denominal deverbal derivational suffix epenthetic elative emphatic equative evidential exclusive existential expressive extension feminine finite focus futur(itiv)e habitive hortative hypothetical genitive immediative imperative

Abbreviations xxiii

〈impers〉 〈imprf〉 〈inch〉 〈incl〉 〈indef〉 〈infin〉 〈instr〉 〈intens〉 〈intent〉 〈interj〉 〈interr〉 〈iter〉 〈itr〉 〈lat〉 〈lim〉 〈loc〉 〈masc〉 〈med〉 〈mir〉 〈mod〉 〈moder〉 〈mom〉 〈mult〉 〈multid〉 〈n〉 〈nec〉 〈neg〉 〈nmlz〉 〈nondum〉 〈nx〉 〈obl〉 〈obv〉 〈ocx〉 〈onom〉 〈opt〉 〈ord〉 〈p〉 〈part〉 〈pass〉 〈pej〉 〈perm〉 〈pers〉 〈pl〉 〈pluprf〉

impersonal imperfective inchoative inclusive indefinite infinitive instrumental intensive intentive interjection interrogative iterative intransitive lative limitative locative masculine medial mirative modal moderative momentative multiplicative multidirectional noun necessitative negative nominalization nondumitive number suffix oblique obviative oblique coaffix onomatopoetic optative ordinal person partitive passive pejorative permissive personal plural pluperfect

〈plur〉 〈poss〉 〈pret〉 〈prev〉 〈prf〉 〈prob〉 〈progr〉 〈prol〉 〈propr〉 〈prosp〉 〈prox〉 〈prs〉 〈pst〉 〈ptcp〉 〈ptcl〉 〈purp〉 〈px〉 〈recipr〉 〈redupl〉 〈ref〉 〈refl〉 〈rem〉 〈res〉 〈restr〉 〈rev〉 〈rx〉 〈sem〉 〈sep〉 〈sg〉 〈sim〉 〈sing〉 〈stat〉 〈sup〉 〈superl〉 〈term〉 〈top〉 〈tr〉 〈transl〉 〈v〉 〈vblz〉 〈voc〉 〈vol〉 〈volit〉 〈vx〉

pluritative possessive pretentive preventive perfective probabilitive progressive prolative proprietive prospective proximal present past participle particle purposive possessive suffix reciprocal reduplication reference reflexive remote resultative restrictive reversive reflexive suffix semelfactive separative singular similative singulative stative supine superlative terminative topic transitive translative verb verbalization vocative voluntative volitive personal ending

CHART OF THE MANCHU SCRIPT In this volume, the letters of the Manchu alphabet are Romanized according to the system of Jerry Norman (e.g. Norman 2013: xv–xxiv), which, in its turn, is a slightly modified version of the system introduced by Paul Georg von Möllendorff (1892). There have been many proposals of Romanization in the past, but all of them involve problems in the details, not only because of the constant confusion between the principles of transcription vs. transliteration, but also because of the presence of special conventions used when writing foreign words in the Manchu script (cf. Osterkamp 2016). The chart below covers the letters used to represent the sounds of the inherited Tungusic vocabulary of Manchu but only a basic set of letters devised for the transcription of Chinese and other loanwords. Apart from the system of Norman, the chart includes the Romanizations used in Hauer (2007) and Hu Zengyi (ed.) (1994: 894–905), complemented by the Cyrillic transcription of Zakharov (1875: 3–14). For more comprehensive comparative tables, cf. Stary (2000: xvi–xvii) and, especially, Lie (1972: 72–77).

Hauer

Hu Zengyi

Zakharov

ᠠ᠊ ᠪ

a

a

a

а

b

b

b

б

ᠴ

c

c

ch

ч (ц)

ᡱ

cy

c’

chy

чы

d

d

d

д

e

e

e

э

f

f

f

ф

g

g

g

г, г̄

g’

g’

gg

г̄

h

h

h

х, х̄

h’

h’

hh

)-(

i

i

i

и, i, й

ᡩ ᡩ᠋᠊ ᡝ᠊

Norman

ᡶ ᡶ᠋ ᡤ᠊ ᡬ᠊ ᡥ᠊ ᡭ᠊ ᡳ᠊

Chart of the Manchu script xxv Hauer

Hu Zengyi

Zakharov

ᠵ᠊

j

j

zh

чж

ᡷ᠊

jy

’j

zhy

чжи

ᡰ᠊

ž

ǰ

rr

ж

k

k

k

к, к̄

ᠺ᠊

k’

k’

kk

к̄

ᠯ᠊

l

l

l

л

ᠮ᠊

m

m

m

м

n

n

n

н

o

o

o

о

p

p

p

п

r

r

r

р

s

s

s, s’

с

sy

s

sy

сы

š

š

sh

ш

t

t

t

т

u

u

u

у

ū

û, ô

uu

у, ӯ

w

w

w

в

y

y

y

)-(

dz

z

z

цз

ts

z’

c

ц

ng

ng

ng

-нг, -н-

ᡴ᠊ ᡴ

Norman

ᠨ᠊ ᠣ᠊ ᡦ ᡵ᠊ ᠰ᠊ ᠰᡟ ᡧ᠊ ᡩ ᡩ᠋᠊ ᡠ᠊ ᡡ᠊ ᠸ᠊ ᠶ᠊ ᡯ᠊ ᡮ᠊ ᠇ᡴ᠌᠊ ᠇ᠭ᠌

Note that in the system of Hu Zengyi 〈s’〉 is used to disambiguate the sequence 〈s〉 cum 〈h〉, i.e., [s.x ~ s.χ] from 〈sh〉 [ʂ], as, for example, in 〈tas’ha〉 ‘tiger’ (for tasha, Zakharov тасха). In Zaxarov’s system, 〈я ѣ iō ю〉 mark sequences with a palatal element (iya iye iyo iyū). The two lines on some consonant letters in the chart indicate combinations with a following “male” or “female” vowel (d k t), or occurrence in initial vs. medial vs. final position (f ng). In fact, due to the history and typology of the script, several other letters have also separate graphic forms for the initial, medial, and final, as well as, in some cases, absolute position, of which the above chart includes only the initial forms, with the exception of the letter sy, which, for technical reasons, is given in its complex initial-final form.

xxvi Chart of the Manchu script

REFERENCES Corff, Oliver & Kyoko Maezono & Wolfgang Lipp & Dorjpalam Dorj & Görööchin Gerelmaa​ & Aysima Mirsultan & Réka Stüber & Byambajav Töwshintögs & Xieyan Li (eds.) (2013) Auf kaiserlichen Befehl erstelltes Wörterbuch des Manjurischen in fünf Sprachen—“Fünfsprachenspiegel”, vols. 1–2 + indices 1–5, Wiesbaden: Harrassowitz. Hauer, Erich (2007) Handwörterbuch der Mandschusprache, 2., durchgesehene und erweiterte Auflage herausgegeben von Oliver Corff, Wiesbaden: Harrassowitz Verlag. Hu, Zengyi 胡增益 (ed.) (1994)『新满汉大词典』[A Comprehensive Manchu-Chinese Dictionary], 烏魯木齊 [Urumqi]: 新疆人民出版社. Lie, Hiu (1972) Die Mandschu Sprachkunde in Korea, Indiana University Uralic and Altaic Series 114, Bloomington, Indiana: Indiana University, Research Institute for Inner Asian Studies. Möllendorff, P[aul] G[eorg] von (1892) A Manchu Grammar with Analyzed Texts, Shanghai: American Presbyterian Mission Press. Norman, Jerry (2013) with the assistance of Keith Dede and David Prager Branner, A Comprehensive Manchu-English Dictionary, Cambridge (Massachusetts) and London: Harvard University Asia Center. Osterkamp, Sven (2016) Review of: Corff & al (2013), Asiatische Studien—Études Asiatiques 70 (3): 942–965, Berlin. Stary, Giovanni (2000) A Dictionary of Manchu Names: A Name Index to the Manchu Version of the “Complete Genealogies of the Manchu Clans and Families of the Eight Banners”, Aetas Manjurica 8, Wiesbaden: Otto Harrassowitz. Zakhrov, Ivan [Иванъ Захаровъ] (1875) Полный маньчжурско-русскiй словарь [A complete Manchu-Russian dictionary], Санктпетербургъ [St. Petersburg]: Типографiя Императорской Академiи Наукъ.

LANGUAGE MAP This map shows the locations and maximal territories of the Tungusic languages at a time level corresponding broadly to the ethnographic present of the 17th to 19th centuries. The languages are indicated by numbers corresponding to the chapters in this volume. For more detailed maps, indicating individual localities and/or clans and tribal-level groups, see, in particular, Dolgikh (1960 map), Vasilevich (1958 map), and Service (1978 map).

11

11

7 7 7

7 8

16

10 15 9

9 9

7

14 12 13

17 18 9

7 Siberian Ewenki

10 Neghidal

13 Udihe

16 Uilta

8 Orochen

11 Ewen

14 Nanai

17 Manchu

9 Solon

12 Oroch

15 Ulcha

18 Sibe

xxviii Language map Note that the western expansion of Ewenki towards the Ob basin and the northern and northeastern expansion of Ewen towards the Anadyr’ basin and the Arctic coast was still in progress during this period, while the expansion of Yakut was gradually hollowing out the territories of both Ewenki and Ewen in the Lena basin, which also marked the approximate mutual boundary of Ewenki and Ewen in the north. Solon underwent translocations from the Middle Amur basin to the Nonni basin in the late 17th century and further to the Hailar basin in the early 18th century, while still another translocation took it, together with Sibe, to Jungaria in 1764. The presence of Ewen on central Kamchatka and Ewenki on central Sakhalin dates back only to the 19th century. In the course of the 20th century, the territories of all Tungusic languages have shrunk to a network of points marking the settlements where the remaining speakers live. As a case of exception, broader expanses of space are still occupied locally by the northernmost reindeer herding Ewenki and Ewen speakers, but even in these cases the bulk of the population lives in multiethnic fixed settlements.

REFERENCES Dolgikh, B. O. [Б. О. Долгих] (1960) Родовой и племенной состав народов Сибири в XVII в. [Clan and tribe composition of the peoples of Siberia in the 17th c.], Труды Института этнографии им. Н. Н. Милухо-Маклая, Новая серия, том LV, Москва [Moscow]: Издательство Академии Наук СССР. Service, Robert T. (1978) ‘Notes on a map of the Tungusic peoples and their dialects’, in: Gerhard Doerfer & Michael Weiers (eds.), Beiträge zur nordasiatischen Kulturgeschichte, Tungusica 1: 52–65, Wiesbaden: Otto Harrassowitz. Vasilevich, G. M. [Г. М. Василевич] (1958) Эвенкийско-русский словарь [Ewenki-Russian dictionary], Mосква [Moscow]: Государственное издательство иностранных и национальных словарей.

CHAPTER 1

TUNGUSIC AS A LANGUAGE FAMILY Juha Janhunen

Tungusic is a medium-sized language-family with some 15 different languages distributed over a vast territory in Northeast Asia, including both Siberia and Manchuria, and today divided between the Russian Federation and the People’s Republic of China, giving it a transnational status. In terms of territorial spread Tungusic is, however, a “large” language family, comparable with the other “large” language families of northern and central Eurasia, such as Mongolic, Turkic, and Uralic. In terms of demographic weight, again, Tungusic is a “small” language family, for the total number of people speaking Tungusic languages today is hardly more than 30,000, probably even less. As a language family, Tungusic is, in fact, seriously endangered, and many of the individual Tungusic languages are on the verge of extinction, with some recently documented varieties already extinct. The political status of Tungusic has, however, varied in the course of time, and the presence of both Russia and China in Northeast Asia is of a relatively recent date. Most importantly, one particular Tungusic language, Manchu, has a history the like of which few other languages can show, for it was not only the regionally dominant language of Manchuria, spoken by a considerable population, but also the official first language of the Manchu Empire of the Qing (1616–1911), which comprised not only Manchuria, but also all the neighbouring regions, including Mongolia, Tibet, East Turkestan, and China. Manchu enjoyed a brief period of renewed symbolic status during the short-lived empire of Manchukuo (1932–1945), but subsequently it has been degraded to just one of the 54 “minority languages” of the People’s Republic, sharing the status of the other Tungusic languages, which have always been spoken by smaller and politically less influential populations. On the Russian side, the speakers of the Tungusic languages are conventionally included in the category of the “Small Peoples of the Far North” (narody Krainego Severa, Sibiri i Dal’nego Vostoka). As far as their traditional modes of subsistence are concerned, the Tungusic speakers represent several regionally diversified adaptations, comprising the settled small-scale agriculturalists of central Manchuria, the settled or semisettled fishing and hunting populations of the Amur basin in northern Manchuria, the nomadic hunters and fishermen of the Ussuri Taiga in eastern Manchuria, the Mongolian-type pastoral nomads of Transbaikalia, and the nomadic reindeer herders living dispersed over most parts of Siberia, with small local communities also on the island of Sakhalin and the peninsula of Kamchatka. A diaspora group of Manchu speakers, officially known as the Sibe, live since the 18th

DOI: 10.4324/9781315728391-1

2 Juha Janhunen

century in Jungaria, northern Sinkiang (Xinjiang), at the modern border towards Kazakhstan, where they share the lifestyle of the other local populations. The difference in the economic adaptations correlates with differences in material and social culture, as well as with the genetic composition of the different Tungusic speaking populations. Obviously, many of the local groups today or historically speaking Tungusic languages, represent traces of other ethnic groups which were secondarily, and in some cases very recently, absorbed into the Tungusic speaking sphere. There are, however, several cultural features that are shared by all Tungusic speaking populations. Most importantly, all Tungusic speakers are traditionally adherents of the North Asian type of shamanism, a complex world view which may or may not be classified as a “religion”. Not surprisingly, the term shaman was also transmitted to international usage from, or via, the Tungusic languages. By comparison, other major religions, including Buddhism, Christianity, and Islam have gained little foothold among Tungusic speakers, though officially Tibetan Buddhism was one of the favoured religions of the Manchu Empire, and Russian Orthodox Christianity has been introduced to most Tungusic speaking groups in Siberia. ETHNIC NOMENCLATURE The names used to refer to the individual Tungusic languages are in most cases based on the corresponding ethnonyms used either by their speakers or by their neighbours. The ethnic nomenclature is, however, a complicated issue and involves many sources of confusion. The English term “Tungusic”, as used here, is itself a relatively new name for the language family, though analogous names have been current since, at least, several decades, in German (tungusische Sprachen) and French (langues toungouses). The standard term in Russian usage is still today “Tungus-Manchu” (tunguso-man’chzhurskie yazyki), while in English the form “Manchu-Tungus” has occasionally been used. Such double names suggest a basic dichotomy of the language family into “Manchu” and “Tungus”, a probably premature conclusion which, in any case, remains to be confirmed. The term “Tungusic” has obvious advantages, one of which is its formal parallelism with “Mongolic” and “Turkic”. It happens that the ethnonyms “Tungus” and “Manchu” are also both of a relatively recent origin. “Manchu” (manju) was the term consciously introduced, and perhaps artificially coined, by the Manchu ruling elite around the time of their dynastic rise (1635), until which time Manchu speakers had identified themselves as “Jurchen” (jušen). “Tungus”, on the other hand, was the term adopted by the Russians in the late 16th century from the local Samoyedic speakers in the Lower Yenisei region in reference to the Siberian “Tungus”, today known as Ewenki and Ewen. In both Russian and European usage, the term “Tungus” has later been applied also to the other Tungusic speaking peoples, with the exception of the Manchu, who are normally separated from the “Tungus”, though not because of linguistic considerations, but because of their distinct demographic, cultural, and political status. With the advance of the Russians towards the Pacific coast, starting in the mid 17th century, and especially after the acquisition by Russia of the areas north of the Amur (1858) and east of the Ussuri (1860), specific names came to be applied to the Tungusic speaking groups in this region, subsequently known as the Russian Far East. The names were typically adopted from the immediately neighbouring groups, both Tungusic and non-Tungusic. For some groups, several parallel names were applied. However, after

Tungusic as a language family 3

the introduction of Soviet rule, most of these “old” names were regarded as politically incorrect and were replaced by “new” names, often based on the endonyms used by the indigenous groups themselves, or simply on the appellative nouns meaning ‘man, person’ in their languages. For this reason, most Tungusic peoples of Siberia and the Russian Far East are known by two names, one of which is “old”, while the other one is “new”. Both names have been used, and are still occasionally used in western sources on Tungusic, though in most cases the “new” names have gradually gained ground also in international Tungusic studies. Further variation is caused by the ways how the “new” names are handled in Russian and other languages, depending on both morphological and orthographical issues, including problems concerned with Romanization. The Russian terms have normally three different stem forms, one used in the masculine singular (referring to a male member of the ethnic group) and the generic plural form, the second in the feminine singular form (referring to a female member of the ethnic group), and the third in the corresponding adjectival form. In some cases, there is variation in how the basic masculine singular form is constructed. The correspondence with the actual native shapes of the ethnonyms is often approximate, and the Russian term is typically based on only one dialectal variant of the indigenous ethnonym. On the Chinese side, the ethnonyms used for the Tungusic speaking groups are partly based on adaptations of the traditional Manchu usage, but the orthographic rendering of the Chinese terms in Latin letters varies depending on the system of Romanization, and also on their degree of adaptation to the rules of Chinese syllable structure. Some ethnonyms are used on both sides of the Sino-Russian border, while in other cases different names are used for a single ethnic group on the two sides of the border. To avoid additional confusion, the names of the Tungusic languages in the present volume are based on a standard created with a parallel consideration of both the Russian and the Chinese name forms (Table 1.1). In some cases, it has been judged more appropriate to use a form closer to the actual indigenous ethnonym, rather than a direct adaption of the Russian or the Chinese name form. TABLE 1.1 NAMES OF THE TUNGUSIC LANGUAGES 1

2

3

4

5

Ewen

èvén

èvény

èvénka

èvénskii

6

7

8

Lamut

Ewenki

èvénk

èvénki

èvenkíika

èvenkíiskii

eve

Ewenke

Tungus

evn

Nanai

nanáec

nanáicy

nanáika

nanáiskii

Neghidal

negidálec

negidál’cy

negidálka

negidal´skii

Hezhe/n

Gold

gld neg

Oroch

óroch

órochi

órochka

órochskii

oac

Orochen

orochón

orochóny

orochónka

orochónskii

Elunchun

orh

Solon

solón

solóny

solónka

solónskii

Ewenke

eve

Udihe

udègéec

udègéicy

udègéika

udègéiskii

Taz

ude

Ulcha

ul’ch

úl’chi

ul’chánka

úl’chskii

Mangun

ulc

Uilta

uíl’ta

uíl’ta

uil’tínka

uil’tínskii

Orok

oaa

Name forms: 1. Recommended standardized name, 2. Russian masculine singular, 3. Russian plural, 4. Russian feminine singular, 5. Russian adjectival form (masculine singular), 6. Chinese name (Pinyin), 7. “old” name, 8. ISO 639–3 language code.

4 Juha Janhunen

It may be pointed out that many of the Russian ethnonyms, both “old” and “new”, are artificial creations and have only a diffuse and arbitrary connection with the endonyms and exonyms of the ethnic groups concerned. Similar issues are observed on the Chinese side. Also, as may be seen, the ISO codes for the Tungusic languages are randomly based on either the “old” or the “new” names, and some languages, like Solon and (Siberian) Ewenki, are lumped together under one single code. Confusion is also caused by the similarity of several ethnonyms, notably Ewenki and Ewen, Ulcha and Uilta, as well as Orochen, Oroch, and Orok. A practical consequence of the ethnonymic similarities is that census data, and the accompanying data concerning the numbers of speakers, are not always reliable. In addition to the above-listed Tungusic languages, whose speakers are recognized as distinct ethnic groups, there are three further varieties, technically known as Arman, Kili, and Kilen, which are also best classified as separate languages, though they are often included in the context of Ewen (Arman) and Nanai (Kili and Kilen), so also in the present volume. Arman may be characterized as an exceptionally archaic variety of Ewen and therefore transitional towards Ewenki, while Kili and Kilen are “mixed” languages, combining Nanai morphology with an Ewenki or (early) Oroch-Udihe type of basic vocabulary. Although extremely interesting in the taxonomic framework, all these varieties are insufficiently documented. Arman and Kili, which were spoken on the Russian side, are recently extinct, while Kilen on the Chinese side is still reported to have a few speakers. DATA AND SOURCES Apart from Manchu, the Tungusic languages became first known to international scholarship via the publications of the German naturalists in Russian service in the 18th century. In his comparative lexicon P. S. Pallas (1786–1789) provides lexical data from up to nine varieties of “Tungus”, that is, Ewenki and Ewen, as well as Manchu. Thanks to its wide areal distribution Ewenki was the natural object of many of the early attempts of documentation. Material for the first grammatical description of (Khamnigan) Ewenki was collected by the Finnish field linguist M. A. Castrén (1856), as published by Anton Schiefner, who also published several other early collections of data. In the latter half of the 19th century, vocabularies and short grammatical descriptions became available on Oroch and Nanai. Ever since its beginnings, the field of Tungusic studies has evolved in the context of German and Russian scholarship. The Russian tradition of Tungusic language studies, as summarized by Gorcevskaya (1959), was initially focused on Manchu, but it soon expanded to comprise also several other Tungusic languages, including, in particular, Ewenki, Ewen, Nanai, and Oroch. Some written materials of religious and educational content were also published in Ewenki, Ewen, and Nanai using a Cyrillic-based ortography. The authors of many of the early publications were not professional linguists, with the notable exception of A. O. Ivanovskii, whose data on Solon (1894) retain their value on a dialectal variety that still awaits a more comprehensive documentation. Without going deeper into the individual languages, the posthumously published lexical work of S. M. Shirokogoroff (1944), later re-organized and republished by Doerfer and Knüppel (2004), should also be mentioned since it covers several varieties of both Ewenki and Orochen from a pan-dialectal perspective. During the Soviet period, the documentation of the individual Tungusic languages was initially stimulated by the need to develop orthographies and literary languages for the

Tungusic as a language family 5

purpose of raising literacy and increasing the political consciousness of the indigenous peoples. While many of the activists of the initial period ended up as schoolteachers, or perished during the purges of the 1930s, some became established scholars. After WW2 and until the 1980s, there was a “Golden Age” dominated by several eminent Soviet Tungusologists, each of whom was typically specialized on one “major” Tungusic language, including G. M. Vasilevich on Ewenki, K. A. Novikova on Ewen, and V. A. Avrorin on Nanai. Avrorin also worked on Oroch and authored a grammar of Written Manchu, published posthumously (2000). There were, of course, also several other scholars, many of whom worked on a number of more “minor” Tungusic varieties, including L. D. Rishes on Arman, T. I. Petrova on Uilta, E. P. Lebedeva (together with V. A. Avrorin) on Oroch, and L. I. Sem on the dialects of Nanai. Two of the Soviet Tungusologists of the “Golden Age” rose above the others in that they initiated the comparative and diachronic study of Tungusic. V. I. Cincius, who initially worked on Ewen, but later also on Ewenki and Neghidal, authored the first ever “comparative phonetics” of the Tungusic languages (1949). Later, she became the initiator and editor-in-chief of the “comparative dictionary” of the Tungusic languages (1975– 1977), which still remains the only work of its kind. O. P. Sunik, whose synchronic work concerned, in particular, Ulcha and Kili, authored comparative monographs on syntax (1947) and morphology (1962, 1982). The works of Cincius and Sunik operate with the entire range of Tungusic languages, including Manchu, as well as with a broad spectre of their dialectal varieties. It may be noted that Russian publications on Tungusic normally apply a Cyrillic-based quasi-phonemic transcription. A more advanced project aiming at creating a unified phonetic transcription for all the Tungusic languages was introduced by Novikova (1961). This transcription, which had both Cyrillic and Roman-based options, was, however, never adopted for general use. Brief systematic descriptions of all Tungusic languages (with the exception of Manchu), as authored by Russian Tungusologists of the Soviet period, are also present in the collective volume published under the editorship of P. Ya. Skorik in the series “Languages of the Soviet Union” (1968). A similar volume, published in the post-Soviet period, but still involving many Soviet-period authors, was edited by V. M. Alpatov & al. in the series “Languages of the World” (1997). In general, the post-Soviet period meant initially a decline in the Tungusological tradition of Russia. During Soviet times, Tungusic studies had been concentrated in the “Altaic Sector” (Sektor altaiskikh yazykov) of the Leningrad Branch of the Institute of Linguistics of the Soviet Academy of Sciences, but due to a series of administrative reforms the “Sector” was closed and its remaining staff dispersed to other entities and institutions. Even so, several of the younger scholars with a background in the “Altaic Sector” have continued their Tungusological activity also later. At the same time, new regional centres of Tungusic studies have developed in places like Novosibirsk and Blagoveshhensk. The German tradition of Tungusology was in the 20th century continued mainly in the context of “Altaic” studies, though many of the German Altaists have actually been adherents of the Anti-Altaist point of view. After WW2 the field was taken up in Germany by Johannes Benzing, who, on the basis of Soviet primary sources, authored a descriptive grammar of “Lamut” (1955), followed by a “comparative grammar” of the Tungusic languages (1956), which up to the present day remains the principal reference tool on the topic. While arriving on many points at results similar to those of Cincius (1949), Benzing added important theoretical insights and established a base for a Western taxonomical and terminological tradition concerning the grammatical categories of Tungusic.

6 Juha Janhunen

Other eminent German scholars who started working on Tungusic in the postwar period include the Pro-Altaist K. H. Menges, who contributed a large general survey of the Tungusic languages. combining both synchrony and diachrony, to the Tungusological volume published under the general editorship of Berthold Spuler (1968), and the Anti-Altaist Gerhard Doerfer, who, among other things, published several detail studies (1973, 1975, 1984) concerning the taxonomic position of Kili (Kur-Urmi Nanai). Doerfer also worked broadly on the external connections of Tungusic and initiated a project on “North Asian cultural history”, the main result of which was an important collection of papers under the joint editorship of himself and Michael Weiers (1978). The volume also contains a map of the Tungusic languages, compiled and commented on by Robert G. Service (1978). Another important language of Tungusic studies is Japanese. Modern Tungusology in Japan was founded by Ikegami Jirō, who initially focused on Uilta, a language whose speakers lived partly under Japanese administration on South Sakhalin (Karafuto) in 1905–1945, from where they were evacuated to Hokkaido after WW2. Ikegami did, however, work also on Manchu, as well as on general Tungusology. His collected papers, most of which are in Japanese, were published in two volumes (1999, 2001). His student Tsumagari Toshirō continued along similar lines of research, but added a component of field work in China and the Russian Far East. He also contributed important papers on issues connected with the transnational position and endangered status of the Tungusic languages in Russia and China (1996, 1997). Stimulated by opportunities of fresh field work, a new generation of Tungusologists has subsequently grown up in Japan. A channel for the field materials is offered by the series “Publications on Tungus Languages and Cultures” (ツングース言語文化論集), in which some 70 volumes have appeared so far, many of them by Kazama Shinjirō. Compared with Japan, Tungusology in China has had a late start. Apart from Manchu studies, in which field several major reference works, including dictionaries, have been produced, the Tungusic languages were included in the series “Grammatical Sketches of the Languages of the Minority Nationalities of China” (中國少數民族語言簡志叢 書), in which volumes on Solon, Orochen, “Hezhe”, and Sibe were published, all in 1986. Importantly, “Hezhe” in this context refers to Kilen, an otherwise little documented language on which the only other general source is the “Hezhen” grammar by Zhang & Zhang & Dai (1989), published exceptionally in English. Otherwise, Chinese scholarship has mainly produced a limited number of relatively unsophisticated collections of primary data from the Tungusic languages spoken in China. Linguistic papers on specific grammatical topics have also been published, especially in the journal “Minority Languages” (民族語文), but they all tend to have the same basic problem in that they are detached from the international framework of Tungusic studies and exhibit an ignorance of the results already achieved elsewhere. English has not been a major language of Tungusic studies until very recently. An early pioneer in this respect was the Latvian linguist and folklorist Peter Schmidt (Pēteris Šmits), who published lexical data on Neghidal, Ulcha, and Oroch, collected in the early 20th century, in a series of papers (1923ab, 1928) which still retain their value. Apart from occasional papers on specific topics, the Tungusic languages have also been discussed in a number of general introductions to the “Altaic” languages, including those by Nicholas Poppe (1965), Igor de Rachewiltz & Volker Rybatzki (2010), and Václav Blažek & al. (2019). This line of research is being continued today in the framework of “Transeurasian”, as displayed broadly in the collective volume edited by Martine

Tungusic as a language family 7

Robbeets and Alexander Savelyev (2020). The volume contains, among other things, grammatical sketches of three sample Tungusic languages, Ewen (Pakendorf & Aralova 2020), Nanai (Oskolskaya 2020), and Sibe (Jang 2020). During the early stages of the field, the principal task of Tungusic studies was documentation. Today, when many Tungusic languages are kept alive by a rapidly diminishing number of aged semi-speakers, documentation no longer serves linguistic purposes, though it can accompany the efforts aimed at supporting language survival and revitalization where it is still possible. One of the last major documentation projects was initiated by Kim Juwon in Korea, with the aim of recording at least one speaker of each Tungusic (as well as Mongolic and Turkic) language. The results of this extensive enterprise, involving a large number of younger scholars, are being published since 2008 in the “Altaic Language Series” of Seoul National University. Volumes on Ewen, Nanai, and Spoken Manchu (as well on a number of Turkic and Mongolic languages) have already appeared in the series. Fortunately, most Tungusic languages are well documented, and the extant material offers possibilities for many new types of linguistic work. One of the fields that have not yet been exhausted is areal typology. Among the first to take up this challenge was Li Bing (1996) in his discussion concerning the types and origin of vowel harmony in Tungusic. The first attempt at a typologically oriented description of an entire Tungusic language was the Ewenki grammar by Igor Nedjalkov (1997), followed by the exceptionally voluminous description of Udihe by Irina Nikolaeva & Maria Tolskaya (2001). Although not without problems, these grammars still serve as the main tools on Tungusic for general typologists, who do not have access to the vast literature published in the traditional languages of Tungusology. Specific typological topics have also been taken up by Andrej Malchukov (1996, 2000) and Andreas Hölzl (2015, 2018, 2020). The increasingly international status of Tungusic studies was first publicized in a conference in Bonn (2000) and its proceedings, edited by Carsten Näher & al. (2002–2004). Two other volumes charting the new perspectives of the field include those edited by Malchukov & Whaley (2012) and Hölzl & Payne (2022). As far as diachrony is concerned, one of the principal issues on which different opinions have been expressed concerns the internal taxonomy of the Tungusic languages. This issue has been touched upon in all general presentations of Tungusic, as well as in many works dealing with the individual languages. Some of the most important specific papers on the issue include Ikegami (1974), Doerfer (1978), Vovin (1993), Whaley & Grenoble & Li (1999), Georg (2004), as well as, most recently, Whaley & Oskolskaya (2020). The languages whose taxonomic position has been viewed as particularly problematic include Manchu, Ewen, and Orochen. Even so, it is possible to arrive at a relatively uncontroversial and generally acceptable classification, as discussed below. INTERNAL TAXONOMY All classifications proposed for the Tungusic languages operate with the model of a branching family tree, though in the details, especially with regard to closely related varieties, the model of areal grouping has also been used. It cannot be denied that some boundaries between the individual Tungusic languages are transitional, though others are not. The basic question is, therefore, where the big faultlines within the family go. This, again, depends on how they are defined. Most classifications proposed so far are based on a quantitative analysis of a selected set of features, which can be phonological,

8 Juha Janhunen

morphological, and/or lexical. In a less mechanistic framework we should, however, take a critical look at the taxonomic parameters, especially in the sense that not all differences and similarities can have the same taxonomic weight. This is because some features are more primary than others, and many features involve secondary parallel developments conditioned by areal contact. There is a consensus on that the Tungusic language family comprises four clearcut and internally coherent groups (as first explicitly pointed out by Ikegami). The four groups, which in the Japanese tradition of Tungusology are often referred to as groups I, II, III, and IV, will here be termed Ewenic, Orochic, Nanaic, and Jurchenic. The Ewenic group comprises, above all, Ewenki, Ewen, and Solon, but also Arman, Neghidal, and Orochen. The Orochic group, which has also been called “Udegheic”, comprises Oroch and Udihe, two languages both of which have historically been identified as “Oroch”. The Nanai group comprises Nanai proper, Ulcha, and Uilta, as well as the two “mixed” languages Kili and Kilen. Finally, the Jurchenic group, which has also been called “Manchuric”, comprises Manchu and Sibe, as well as their mediaeval ancestor Jurchen. In this context, it is irrelevant whether or not the recorded forms of Jurchen are directly ancestral to the written and spoken forms of Manchu and Sibe, for the Jurchen-Manchu lineage has all the time involved internal dialectal variation, due to which the dialectal base of the normative forms has varied. Geographically, the four groups form a somewhat unbalanced system. Ewenic covers most of Siberia between the Ob-Yenisei region and the Pacific coast, extending to the Arctic coast in the north, as well as to the northern parts of Manchuria and Mongolia in the south. Jurchenic has likewise a wide distribution, covering historically much of southern and central Manchuria, but also parts of northern and western Manchuria. In addition, the secondary diaspora group of the Jurchenic (Manchu) speaking Sibe was formed in the 18th century in Jungaria, accompanied by a smaller group of Ewenic (Solon) speakers, known as the Ongkor Solon. By contrast, Nanaic is restricted to a rather well-defined territory in the basin of the Lower Amur and its tributaries, including the lower courses of the Sungari and Ussuri, while Orochic occupies the immediately adjacent mountain region of Sikhote-Alin to the east, located between the Lower Amur main basin and the Pacific coast. There is less consensus on how the four groups are organized with regard to each other in the framework of a binary family tree. Most classifications proposed so far operate with the concept of two principal branches, identified as “Northern Tungusic” and “Southern Tungusic”. There is also agreement on that Ewenic belongs to the context of Northern Tungusic, while Jurchenic is Southern Tungusic. For Orochic and Nanaic, with or without Kili and Kilen, however, three different classifications have been proposed. The traditional view is that they form a separate “Amur Tungusic” branch, also termed “Central Tungusic” (Doerfer) or “East Tungusic” (Vovin). This classification is supported by the many structural features which these languages have in common, and which distinguish them from both Ewenic and Jurchenic. Alternatively, Amur Tungusic has been viewed as a separate sub-branch of Southern Tungusic, leaving Ewenic alone in the Northern Tungusic branch (so, for instance, de Rachewiltz & Rybatzki). The third option is that Amur Tungusic is divided between the two principal branches, with Orochic going together with Ewenic into the Northern Tungusic branch and Nanaic together with Jurchenic into the Southern Tungusic branch (Georg). The division of Amur Tungusic between the Northern and Southern Tungusic main branches, which is the view adopted here, is supported by the distribution of several

Tungusic as a language family 9

primary innovations, especially in the phonology. Thus, for instance, Proto-Tungusic *ü has a dual representation in the modern Tungusic languages, appearing as (*)i in Ewenic and Orochic and as (*)u in Nanaic and Jurchenic. Similarly, Proto-Tungusic initial *p is represented as (*)p in Nanaic and Jurchenic vs. (*)x in Ewenic and Orochic. Both of these, as well as several other differences involve very early innovations in one or the other branch, or in both, and should therefore be considered as taxonomically diagnostic, whereas, for instance, the morphological features shared by Orochic and Nanaic must be connected with secondary areal interaction. It may be concluded that Amur Tungusic retains its relevance as an areal concept, but not as a primary genetic node in the family tree (Table 1.2). TABLE 1.2 CLASSIFICATION OF THE TUNGUSIC LANGUAGES Northern Tungusic

 

Ewenic

Siberian

Ewen

 

Arman Neghidal Siberian Ewenki Khamnigan Ewenki Manchurian

Urulyunggui Borzya

Orochen Solon diaspora Amur Tungusic

Orochic

Ongkor Solon Oroch Udihe

Southern Tungusic

Nanaic

Nanai Ulcha Uilta mixed

Kili Kilen

Jurchenic

mediaeval

Jurchen

modern

Manchu

diaspora

Sibe

Within each of the four branches there are important differences between the languages with regard to their degree of innovativeness. This is particularly clear in the Orochic branch, where Udihe may be understood as a highly innovative form of Oroch, with which it is linked via transitional dialects. Similarly, in the Nanaic branch, Nanai is in some respects the most conservative member, while Ulcha and Uilta share a number of secondary innovations absent in Nanai, and Uilta contains an additional set of innovations peculiar to it alone. The Ewenic branch, which has more members, has a somewhat more complex structure, in that it is divided into two major sub-branches, which may be termed Manchurian Ewenic and Siberian Ewenic. Siberian Ewenic comprises, most importantly, Siberian Ewenki, but also Ewen and Neghidal, which may be seen as more innovative members of the sub-branch. Ewen is linked with Siberian Ewenki via Arman, a transitional idiom that is best understood as a conservative variety of Ewen.

10 Juha Janhunen

Manchurian Ewenic, on the other hand, comprises Solon and Orochen, of which the latter shows transitional features towards Siberian Ewenki. Occasional attempts at revising the internal structure of Ewenic by assigning special taxonomic properties to either Ewen (Vovin) or Orochen (Whaley & Grenoble & Li) have not yielded generally acceptable results. An important transitional position within Ewenic is, however, formed by Khamnigan Ewenki, which is simply the complex of Ewenic dialects spoken by the Khamnigan, whose other ethnic language is Khamnigan Mongol. It happens that the Khamnigan speak two different varieties of Ewenic, one of which, known as the Borzya (or Man’kovo) variety, is a conservative type of Manchurian Ewenic, while the other, known as the Urulyunggui (or Urul’ga) variety, belongs to the context of Siberian Ewenic. Even so, the two varieties of Khamnigan Ewenki share also a number of common features, which must be connected with their mutual areal interaction. When it comes to innovativeness, Manchu, and Jurchenic as a whole, is traditionally viewed as the most innovative form of Tungusic. By a mechanic count of distinctive properties (Doerfer) Manchu is certainly the most “different” Tungusic language. The main characteristic of Manchu, as compared with all other Tungusic languages, is its preference for analytic solutions instead of the more synthetic structures present in the other three groups, including also Nanaic. It remains a matter of some dispute as to how “original” the synthetic morphology of Nanaic, Orochic, and Ewenic is, but at least many of the synthetic features present in them and absent in Manchu, such as, for instance, the system of person markers, are clearly primary and have been lost in Jurchenic. This means that the innovativeness of Manchu is mainly due to losses that have simplified the structure of the language in almost all respects, including phonology, phonotactics, morphology, and morphosyntax. The taxonomic position of the two “mixed” languages, Kili and Kilen, remains also disputable. Earlier classified as Nanai “dialects”, they are today generally recognized as separate languages which combine Nanai morphology and morphosyntax with a Northern Tungusic basic vocabulary. Moreover, while the Northern Tungusic lexical component in Kili (as first identified by Sunik) is of a Manchurian Ewenic type, the Kilen lexicon contains some features that would seem to be more compatible with Orochic (as argued by Kazama). In this situation, it remains a matter of opinion whether these languages are to be understood as varieties of relexified Nanai or restructured Northern Tungusic. Altogether, although relatively much research has been done on, especially, Kili, the taxonomy and mutual relationships of Kili and Kilen are a field that still remains to be studied in much more detail. TYPOLOGICAL PROFILE All Tungusic languages are unambiguous members of the Transeurasian typological belt traditionally known as “Ural-Altaic”, or, in a more restricted framework, as “Altaic”. Tungusic is also geographically adjacent to all the other language families belonging to this belt, that is, Uralic (in the west), Turkic (originally in the southwest), Mongolic (in the south), Koreanic (in the southeast), and Japonic (historically on Sakhalin). Bilingualism in either Turkic (Yakut) or Mongolic (Daghur, Khamnigan Mongol, mainstream Mongolian) has been characteristic of certain Tungusic speaking groups for centuries. Therefore, it is not surprising that most of the typological features observed in Tungusic are also present in the other “Ural-Altaic” languages. There are, however, differences in

Tungusic as a language family 11

the details, and some features are areally transitional both within Tungusic and within the “Ural-Altaic” complex. The following is a list of some of the most prevalent typological features present in Tungusic. The list proceeds from segmental phonology and phonotactics (1–4) to morphology (5–8) and syntax (9–12). As the discussion is focused on structural properties, no actual language data are quoted in this context. (1) Consonants: All Tungusic languages have a relatively simple consonant system comprising most typically 18 consonants, and distinguishing between four places of articulation (labial, dental, palatal, velar), combined with six to seven manners of articulation (nasals, weak stops, strong stops, fricatives, glides, and liquids). The stops (weak vs. strong) are differentiated according to either voice (voiced vs. unvoiced) or aspiration (unaspirated vs. aspirated), or both. The system of liquids comprises a lateral (l) and a vibrant/trill (r), but there is a clear areal tendency to reduce the role of the latter in the languages spoken in the Amur region. The system of fricatives comprises typically a dental sibilant (s) and a velar continuant (x), but some languages have additional segments in the labial and palatal sets. Other additional consonants include a set of distinctive postvelars, present marginally in Manchu, but phonetically also in several other Tungusic languages in the Amur region. (2) Consonant phonotactics: In initial (#C) and final (C#) positions only single consonants are permitted, and medial clusters can contain only two segments (C1C2), with the restrictions concerning their combinations varying from language to language. Geminate consonants (C1C1) are generally rare, and tend to be simplified if they arise at morpheme boundaries. A few languages, notably Ewen, Solon, and Oroch, have, however, secondary geminates produced by processes of vowel elision and/ or consonant assimilations. Of the individual consonants the vibrant (r) does not appear in initial position, although the lateral (l) does. Most Tungusic languages have also an initial velar nasal (ƞ), a Siberian areal feature absent in most other languages of the “Ural-Altaic” belt. The phonotactic status of the velar continuant (x), often realized as a laryngeal spirant, varies depending on its origin: in the individual languages, it can represent an “original” initial consonant of the same type (*x), an initial strong labial stop (*p), or an initial or medial dental sibilant (*s). (3) Vowels: The system of vowels is more variable and comprises variously from 5 (in Manchu) to 9 (in Solon) distinctive short vowels, organized in 2–4 levels of articulation. Most Tungusic languages, with the exception of Manchu, have also a varying number of long vowels, which may be analysed as vowel sequences (V1V1). Diphthongoid sequences of two non-identical vowels (V1V2) are present especially in Amur Tungusic. Phonational distinctions (breathy vs. creaky voice), reminiscent of tones, are present in Udihe, but otherwise suprasegmental features (other than vowel length) are not distinctive in Tungusic. (4) Vowel phonotactics: The distribution of vowel qualities is regulated by a progressive vowel harmony, which in most Tungusic languages operates on the vertical (highlow) axis, with or without an accompanying tongue root opposition (advanced vs. retracted tongue root). In a few marginal idioms (western Ewen), however, vowel harmony operates on the horizontal (palatal-velar) axis, opposing front vowels to back vowels. Most Tungusic languages have also traces of a labial harmony, due to which the non-high unrounded back vowel (a) cannot be combined with the corresponding rounded vowel (o). Both types of vowel harmony affect both word roots

12 Juha Janhunen

(5)

(6)

(7)

(8)

and suffixes. Originally phonetic phenomena, they have gained a phonological status due to secondary developments, such as the neutralization of the opposition between certain vowel qualities. These developments differ from language to language. Parts of speech: There is a clear morphological and syntactic contrast between two word classes, which may be identified as nominals and verb(al)s. In this context, nominals comprise the subclasses of nouns, spatials (spatial nouns), adjectives, numerals, and pronouns, all of which have slight morphological idiosyncracies. In addition, there is a small class of invariables with no productive morphology. In terms of root structure, nominals and verb(al)s are similar, with most roots being bisyllabic vowel stems ((C)VC(C)V), but there are also monosyllabic and bisyllabic consonant stems ((C)(VC)VC), though the set of possible final consonants in them is limited to only a few possible segments. In some languages, complications to this system have been created by the secondary loss of final vowels, while in others the system has been simplified leaving only vowel and nasal stems. A small number of roots, functioning as pronouns and other auxiliaries, are monosyllabic with no final consonant ((C)V). Trisyllabic and longer roots are secondary and may be analysed as derivatives. Morphological strategies: Grammatical and derivational functions are expressed by suffixes, attached to the preceding word root in a fixed order. The coherence of the suffixally marked word is enhanced by vowel harmony, and the morpheme boundary can involve segmental alternations of both the root and the suffix, rendering the sequence less agglutinative and more fusional. The degree of fusional morphology varies, however, being maximal in the northeast (Ewen) and minimal in the southeast (Manchu). There is also a similar areal transition in the general amount of morphology, which increases when moving from Jurchenic via Amur Tungusic to Ewenic. All Tungusic languages have also postclitics, attached to word forms as phonologically dependent but semantically independent elements. Nominal morphology: The categories of nominal inflectional morphology comprise number, case, and possession. Number marking tends to be obligatory in the Siberian sphere, but less so in the Manchurian sphere. The number of suffixally marked cases varies from 6 (in Manchu and Sibe) to 11 (in Ewen) or 12–13 (in Solon and Siberian Ewenki), though the exact number is always a matter of interpretation. For object marking, most Tungusic languages, except Jurchenic, have two cases (accusative vs. partitive/​designative). For locational expressions a three-way contrast between locative (where?), dative/lative (whither?), and ablative (whence?) cases is used, though the distinction between the locative and dative/lative functions is occasionally fuzzy. Most Tungusic languages lack a true adnominal genitive case, though some languages, especially in the Manchurian sphere, have developed a secondary genitive case from what is originally a possessive form (a derivational feature). Personal possession is marked by possessive suffixes of pronominal origin, except in Jurchenic, where only analytic means are used. In the first person plural, a distinction is made between the inclusive and exclusive functions, also in the Jurchenic pronouns, but not in Nanaic (with the exception of Kili and Kilen). A separate set of suffixes expresses reflexive possession. Most Tungusic languages, except Jurchenic, have also a suffixal marker of alienability. Verbal morphology: The typical verbal paradigm in Tungusic comprises four types of forms: finite indicative forms, finite imperative forms, participles, and converbs. Participles are morphologically nominals, which is why they can take the nominal

Tungusic as a language family 13

type of suffixal marking for number, case and possession. The finite indicative forms take also person marking, except in Jurchenic, but the set of person markers is different from the possessive suffixes. The imperative forms do not generally take suffixal person marking, but they tend to be used in person-specific functions. Most converbs are morphologically invariable, but some can take plural marking of the nominal type. Apart from actual converbs, there is a secondary category of “quasiconverbs”, which are morphologically transparent case forms of participles. Many functional categories, including voice, mood, and aspect are expressed by derivational suffixes, located between the verbal root and the inflectional markers. (9) Phrase structure: The word order in both nominal and verbal phrases is strictly head-final. In sequences of a numeral and a noun, both members can function as head nouns, though more often the numeral functions as a modifier. In the languages with no adnominal genitive, nominal phrases are head-marked, with the nominal or pronominal possessor indicated on the head noun by a possessive suffix. In the languages with both possessive suffixes and an adnominal genitive, double marking can be present, and the genitival attribute precedes an adjectival modifier (GAN). Verbal phrases are, by contrast, dependent-marked, in that the object stands in the accusative (definite object) or partitive/designative (indefinite or designative object) case. Adverbials are also marked by the relevant case suffixes. (10) Sentence structure: In an independent head clause, all Tungusic languages favour the verb-final basic word order (SOV), though deviations from this are not rare especially in the Siberian sphere, where they may or may not be connected with recent Russian interference. The form of the finite predicate is chosen from the sets of finite indicative and imperative forms, but, importantly, participles can also be used in the finite function with person marking of the possessive type. In practice, there is a tendency observed in all Tungusic languages to replace the finite indicative forms with participles, which, then, function as finite tense forms for the present, past, and (in many languages) future tenses. The finitization of participles, in general typology also known by the somewhat misleading and confusing terms “verbalization” and “insubordination”, has in some Tungusic languages led to the synchronic presence of two sets of finitely used forms, in which case the original finite set has tended to develop secondary evidential functions (as in Nanai). In other languages, the original finite set has receded or disappeared altogether (as in Oroch). (11) Complex sentences: There are no primary conjunctions in Tungusic, but subordinate clauses are linked to the head clause with the help of converbs and quasiconverbs, which express various temporal, conditional, or causal relations. In a complex sentence, the predicate of only the last (main) clause stands in a finite form, which can also be a finitely used participle. Depending on their reference, converbs and quasiconverbs can be divided into the conjunct (same-subject) and disjunct (different-subject) types, though some converbs can be inherently ambivalent. In quasiconverbs this distinction is expressed by the opposition between possessive (disjunct) and reflexive (conjunct) suffixes. (12) Sentence types: For the equative and existential functions, most Tungusic languages use a fully conjugated copula-existential verb, which, however, can be absent in a temporally and personally unmarked nominal sentence. Negation is primarily expressed by a fully conjugated negation verb, which precedes an invariant connegative form of the semantic main verb. In many Tungusic languages this system has undergone restructuring either by the suffixalization of the conjugated negation

14 Juha Janhunen

verb to the connegative form of the main verb or by the reduction of the morphology of the negation verb and the creation of a number of invariant negation particles. Existential negation is mainly expressed by a number of language-specific privative nouns (‘absent, absence’), which can also be suffixalized. Interrogation in polar questions is expressed by interrogative particles, which are typically (post)cliticized to the finite verb. Some Tungusic languages have a separate corrogative particle, used in non-polar (content) questions containing lexical question words. For other discursive functions, a varying number of discourse particles of the (post)clitical type are used. GRAMMATICAL FRAMEWORK The grammatical description of the Tungusic languages is firmly anchored in the research tradition of comparative Altaic studies. This tradition also determines to a large extent the terminology used to describe the grammatical features present in the Tungusic languages. One characteristic of this terminology is that it is generally oriented from form to function, rather than vice versa. This means that a single form, or its cognate representations shared by several languages, is labelled uniformly even though its functions may vary from language to language. This principle also applies within a single language: even if a form has several synchronic functions, it still retains its formal identity and can conveniently always be referred to by the same term. There are, however, exceptions, for in some cases it is motivated to use different terms when referring to the different functions of a single form. Participles are perhaps the most typical example of multifunctional forms. The forms termed “participles” in Altaic studies are actually polyfunctional nominalized verbs. They can be used in the “prototypical” participial function as adnominal modifiers before a head noun, but they can also be used as independent head nouns indicating either an actor (actor nouns) or an action (action nouns). In the latter function, especially when combined with case suffixes, they serve as “quasiconverbs”, that is, as predicates of subordinate clauses. Because of the possibility of finitization they can, however, also have the role of finite predicates. This multiplicity of functions calls for terminological differentiation. When used as nominalized verbs, participles are often assumed to indicate a primarily aspectual opposition, which is why the two basic participial forms are traditionally identified by the terms “imperfective” vs “perfective”. However, when used as finite predicates, these forms tend to assume temporal functions and are better referred to by the terms “present” vs. “past”. Tense and aspect as functional categories are in these cases, as always, difficult to keep apart. Many Tungusic languages have also a diachronically secondary “futuritive” participle, which, when used in a finite function, may be identified as the “future” tense marker. More rarely do the Tungusic languages offer examples of monofunctional nominalizations, but when such forms are present (as in Manchu and Sibe), they are best referred to as “infinitives”, implying nominalized verbs that are only, or primarily, used as independent action nouns. Another multifunctional verbal form in Tungusic is the so-called “aorist”. At a deep diachronic level, this is another participial form, which still has the function of a participle in a few languages (as in Manchu and Solon). More generally, this form has, however, become the marker of a temporally unmarked finite paradigm. In many Tungusic languages, this “aorist” paradigm constitutes the basic “unmarked” set of finite forms, as opposed to the temporally differentiated forms based on finitely used participles. For

Tungusic as a language family 15

this reason, the remaining examples of the participial use of the aorist may be referred to as the “aorist participle”. This form has, however, also another use, in that, in a negative clause, it is the form that the main verb takes when combined with the conjugated forms of the negation verb. In this function, the aorist participle is best identified terminologically as the “connegative”. Many terminological problems are connected with the converbs. Unlike the participles and the aorist paradigm, which are relatively uniform in all Tungusic languages, there is much more formal variation in the systems of converbs, and many languages possess language-specific converbs, often recent innovations, that have no direct parallels elsewhere. Nevertheless, it seems reasonable to apply unified names for those converbal forms that are present in two or more Tungusic languages. It may be noted that one of the converbs present in most Tungusic languages is based on the aorist participle and is therefore, on formal grounds, best referred to as the “aorist converb”, even though several other names have been used for it in the past. Altogether, the terminology used for the converbs is traditionally less well established and more impressionistic than for most other morphological categories, which is why the creation of a new unified terminology requires an innovative approach. The same is true of many of the modal and aspectual forms which are expressed derivationally, but which, nevertheless, can have cognates in several Tungusic languages. When it comes to nominal morphology the traditional terminology is better established. For most nominal forms it is reasonable to use maximally simple traditional appellations, although these names do not always cover all of their actual functions. The typical case paradigm of a Northern Tungusic language comprises, for instance, the unmarked nominative, as well as the suffixally marked accusative, partitive, dative, ablative, locative, prolative, directive, and instrumental cases, to which a varying number of more specific secondary cases can be added. The functions behind these labels can, however, be multiple. The dative, for instance, apart from the “prototypical” dative function of indicating the recipient (“indirect object”), has also a wide range of locative and temporal functions, many of which overlap to a varying extent with those of other cases, such as the locative and the directive. In this, as well as in other similar examples, the name of a form should best be understood as a simplified and conventionalized cover term for all the functions that the form has. REFERENCES AND FURTHER READING Alpatov, V. M. [В. М. Алпатов] & I. V. Kormushin [И. В. Кормушин] & G. C. Pyurbeev [Г. Ц. Пюрбеев] & O. I. Romanova [О. И. Романова] (eds.) (1997) Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки мира (Languages of the World) [3]: 267–284, Москва [Moscow]: “Индрик”. Avrorin, V. A. [В. А. Аврорин] (2000) Грамматика маньчжурского письменного языка [A grammar of Written Manchu], Санкт-Петербург [St. Petersburg]: “Наука”. Benzing, Johannes (1955) Lamutische Grammatik, mit Bibliographie, Sprachproben und Glossar, Akademie der Wissenschaften und der Literatur, Veröffentlichungen der Orientalischen Kommission 6, Wiesbaden: Franz Steiner Verlag. Benzing, Johannes (1956) Die tungusischen Sprachen: Versuch einer vergleichenden Grammatik, Akademie der Wissenschaften und der Literatur in Mainz, Abhandlungen der geistes-und sozialwissenschaftlichen Klasse 1955 (11), Wiesbaden: Franz Steiner Verlag.

16 Juha Janhunen

Blažek, Václav, in collaboration with Michal Schwarz and Ondřej Srba (2019) Altaic Languages: History of Research, Survey, Classification and a Sketch of Comparative Grammar, Brno: Masaryk University Press. Castrén, M. Alexander (1856) Grundzüge einer tungusischen Sprachlehre nebst kurzem Wörterverzeichniss, herausgegeben von Anton Schiefner, Nordische Reisen und Forschungen 9, St. Petersburg: Kaiserliche Akademie der Wissenschaften. Cincius, V. I. [В. И. Цинциус] (1949) Сравнительная фонетика тунгусоманьчжурских языков [Comparative phonetics of the Tungusic languages], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство Министерства просвещения РСФСР. Cincius, V. I. [В. И. Цинциус] (ed.) (1975–1977) Сравнительный словарь тунгусоманьчжурских языков: Материалы к этимологическому словарю [Comparative diactionary of the Tungusic languages: Materials for an etymological dictionary], vols. 1–2, Ленинград [Leningrad]: “Наука”. Doerfer, Gerhard (1973) ‘Das Kur-Urmiische und seine verwandten’, Zentralasiatische Studien 7: 567–599, Bonn. Doerfer, Gerhard (1975) ‘Ist Kur-Urmiisch ein nanaischer Dialekt?’, Ural-Altaische Jahrbücher 47: 51–63, Wiesbaden. Doerfer, Gerhard (1978) ‘Classification problems of Tungus’, in: Doerfer & Weiers (eds.) (1978), 1–26. Doerfer, Gerhard (1984) ‘Die Körperteilbezeichnungen des Kili (Ein Beitrag zur Frage der dialektologischen Klassifikation)’, Nyelvtudományi Közlemények 86: 238–246, Budapest. Doerfer, Gerhard & Michael Knüppel (2004) Etymologisch-ethnologisches Wörterbuch tungusischer Dialekte (vornehmlich der Mandschurei), Hildesheim: G. Olms. Doerfer, Gerhard & Michael Weiers (eds.) (1978) Beiträge zur nordasiatischen Kultur­ geschichte, Tungusica 1, Wiesbaden: Otto Harrassowitz. Georg, Stefan (2004) ‘Unreclassifying Tungusic’, in: Naeher & Stary & Weiers (eds.) (2004), 45–57. Gorcevskaya, V. A. [В. А. Горцевская] (1959) Очерк истории изучения тунгусоманьчжурских языков [A brief history of the study of the Tungusic languages], Серия “История отечественного языкознания” [History of linguistics in Russia] 2, Ленинград [Leningrad]: Государственное учебно-педагогическое издательство Министерства просвещения РСФСР. Hölzl, Andreas (2015) ‘A typology of negation in Tungusic’, Studies in Language 39 (1): 117–157. Hölzl, Andreas (2018) ‘Constructionalization areas: The case of negation in Manchu’, in: Evie Coussé & Peter Andersson & Joel Olofsson (eds.) Grammaticalization Meets Construction Grammar, Constructional Approaches to Language 21: 241–276, Amsterdam: John Benjamins. Hölzl, Andreas (2020) ‘The complexification of Tungusic interrogative systems’, in: Dankmar Enke & Larry Hyman & Johanna Nichols & Guido Seiler & Thilo Weber (eds.) Language Change for the Worse, Chapter 8: i–xxxi, Berlin: Language Science Press. Hölzl, Andreas & Thomas E. Payne (ed.) (2022) Tungusic Languages: Past and Present, Studies in Diversity Linguistics 32, Berlin: Language Science Press. Ikegami, Jiro (1974) ‘Versuch einer Klassifikation der tungusischen Sprachen’, in: Georg Hazai & Peter Zieme (eds.), Sprache, Geschichte und Kultur der Altaischen Völker,

Tungusic as a language family 17

Protokollband der XII. Tagung der Permanent International Altaistic Conference 1969 in Berlin, Schriften zur Geschichte und Kultur des Alten Orients 5: 271–272, Berlin: Akademie Verlag. Also published in Ikegami (2001), 395–396. Ikegami, Jirō 池上二良 (1999)『満洲語研究』[Researches on the Manchu Language], 東京 [Tokyo]: 汲古書院. Ikegami, Jirō 池上二良 (2001)『ツングース語研究』[Researches on the Tungus Languages], 東京 [Tokyo]: 汲古書院. Ivanovskii, A. O. [А. О. Ивановскiй] (1894) Mandjurica I: Образцы солонскaго и дахурскaго языковъ [Manjurica I: Specimens of the Solon and Daghur languages], С.-Петербургъ [St. Petersburg]: Типографiя Императорской Академiи наукъ. Reprint: Debter—Deb-ther—Debtelin: Materials for Central Asiatic and Altaic Studies 2, Budapest: Akadémiai Kiadó (1982). Jang, Taeho (2020) ‘Xibe and the Manchuric languages’, in: Robbeets & Savelyev (eds.) (2020), 269–287. Li, Bing (1996) Tungusic Vowel Harmony: Description and Analysis, Academisch Proefschrift, Universiteit van Amsterdam, Holland Institute of Generative Linguistics (HIL) Dissertations 18, Amsterdam. Malchukov, Andrej L. (1996) ‘Internal relative clauses in Tungusic languages in a synchronic and a diachronic perspective’, Sprachtypologie und Universalienforschung 49 (4): 358–362. Malchukov, Andrej L. (2000) ‘Perfect, evidentiality and related categories in Tungusic languages’, in: Lars Johanson & Bo Utas (eds.), Evidentials: Turkic, Iranian and Neighbouring languages, 441–469, Empirical Approaches to Language Typology, Berlin: De Gruyter Mouton. Malchukov, Andrej L. & Lindsay J. Whaley (eds.) (2012) Recent Advances in Tungusic Linguistics, Turcologica 89, Wiesbaden: Harrassowitz Verlag. Naeher, Carsten & Giovanni Stary & Michael Weiers (eds.) (2002–2004) Proceedings of the First International Conference on Manchu-Tungus Studies (Bonn 2000), vol. 1: Trends in Manchu and Tungus Studies, Tunguso-Sibirica 8 (2002), vol. 2. Trends in Tungusic and Siberian Linguistics, Tunguso-Sibirica 9 (2004), Wiesbaden: Harrassowitz Verlag. Nedjalkov, Igor (1997) Evenki, Routledge Descriptive Grammars, London and New York: Routledge. Nikolaeva, Irina & Maria Tolskaya (2001) A Grammar of Udihe, Mouton Grammar Library 22, Berlin: Mouton de Gruyter. Novikova, K. A. [К. А. Новикова] (1961) Проект единой фонетической транскрипции для тунгусо-маньчжурcких языков [A project on a unified phonetic transcription for the Tungusic languages], Москва [Moscow] & Ленинград [Leninrgad]: Академия Наук СССР, Институт языкознания. Oskolskaya, Sofia (2020) ‘Nanai and the Southern Tungusic languages’, in: Robbeets & Savelyev (eds.) (2020), 305–320. Pakendorf, Brigitte & Natalia Aralova (2020) ‘Even and the Northern Tungusic languages’, in: Robbeets & Savelyev (eds.) (2020), 288–304. Pallas, P. S. (1786–1789) Linguarum totius orbis vocabularia comparativa / Сравнительные словари всѣхъ языковъ и нарѣчiй, vols. 1–2, Petropoli / Въ Санктпетербургѣ: Тypis Iohannis Carolii Schnoor. Poppe, Nicholas (1965) Introduction to Altaic Linguistics, Ural-Altaische Bibliothek 14, Wiesbaden: Otto Harrassowitz.

18 Juha Janhunen

de Rachewiltz, Igor & Volker Rybatzki, with the collaboration of Hung Chin-fu (2010) Introduction to Altaic Philology: Turkic, Mongolian, Manchu, Handbuch der Orientalistik VIII.20, Leiden: Brill. Robbeets, Martine & Alexander Savelyev (eds.), (2020) The Oxford Guide to the Trans­ eurasian Languages, Oxford Guides to the World’s Languages, Oxford: Oxford University Press. Schmidt, P[eter] (1923a) ‘The language of the Neghidals’, Acta Universitatis Latviensis 5: 3–38, Riga. Schmidt, P[eter] (1923b) ‘The language of the Olchas’, Acta Universitatis Latviensis 8: 229–288, Riga. Schmidt, P[eter] (1928) ‘The language of the Oroches’, Acta Universitatis Latviensis 17: 17–62, Riga. Shirokogoroff, S. M. (1944) A Tungus Dictionary: Tungus-Russian and Russian Tungus, Photogravured from the manuscripts, Shinobu Iwamura (ed.), Tokyo. Service, Robert T. (1978) ‘Notes on a map of the Tungusic peoples and their dialects’, in: Doerfer & Weiers (eds.) (1978), 52–65. Skorik, P. Ya. [П. Я. Скорик] (editor-in-chief) & al., Монгольские, тунгусоманьчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5, Ленинград [Leningrad]: “Наука”. Spuler, B[erthold] (editor-in-chief) (1968), Altaistik: Tungusologie, mit Beiträgen von W[alter] Fuchs & Ivan A. Lopatin & Karl H. Menges & Denis Sinor, Handbuch der Orientalistik I.V.3, Leiden: E. J. Brill. Sunik, O. P. [О. П. Суник] (1947) Очерки по синтаксису тунгусо-маньчжурских языков (Essays on the syntax of the Tungusic languages), Ленинград [Leningrad]: Государственное учебно-педагогическое издательство Министерства просвещения РСФСР. Sunik, O. P. [О. П. Суник] (1962) Глагол в тунгусо-маньчжурских языках: Морфологическая структура и система форм глагольного слова (The verb in the Tungusic languages: The morphological structure and system of forms of the verbal word), Москва [Moscow] & Ленинград [Leningrad]: Издательство Академии Наук СССР. Sunik, O. P. [О. П. Суник] (1982) Существительное в тунгусо-маньчжурских языках: В сравнении с другими алтайскими языками [The noun in the Tungusic languages: In comparison with the other Altaic languages], Ленинград [Leningrad]: “Наука”. Tsumagari, Toshirō 津曲敏郎 (1996)「中国・ロシアのツングース諸語」[The Tungusic languages in China and Russia],『言語研究』Journal of the Linguistic Society of Japan 110: 177–191, 京都 [Kyoto]. Tsumagari, Toshirō (1997) ‘Linguistic diversity and national borders of Tungusic’, in: Hiroshi Shoji & Juha Janhunen (eds.), Northern Minority Languages: Problems of Survival, Senri Ethnological Studies 44: 175–186, Osaka: National Museum of Ethnology. Vovin, Alexander (1993) ‘Towards a new classification of Tungusic languages’, Eurasian Studies Yearbook 65: 99–113, Berlin & London: Eurolingua. Whaley, Lindsay J. & Lenore A. Grenoble & Fengxiang Li (1999) ‘Revisiting Tungusic classification from the bottom up: A comparison of Evenki and Oroqen’, Language 75 (2): 286–321. Whaley, Lindsay J. & Sofia Oskolskaya (2020) ‘The classification of the Tungusic languages’, in: Robbeets & Savelyev (eds.), (2020), 81–91. Zhang, Yanchang [張彥昌] & Zhang Xi [張晰] & Dai Shuyan [戴淑艳] (1989) The Hezhen Language [赫哲语], 長春 [Changchun]: 吉林大學出版社 [Jilin University Press].

CHAPTER 2

EARLY FAR EASTERN SOURCES ON TUNGUSIC Alexander Vovin

The earliest information pertaining to the Tungusic languages and their speakers is contained in Chinese sources starting with the 1st millennium BZ. The early information is, however, invariably circumstantial, hidden behind obscure ethnonyms and brief mentions concerning the mutual intelligibility or non-intelligibility of the local non-Chinese languages, as spoken to the northeast of China. Substantial data start becoming available only after the rise of the Jurchen of the Jin dynasty (1115–1234), a period from which native materials are also preserved in the Jurchen language and script. From that time on, sources written in the Chinese language offer a gradually growing amount of information on Jurchen-Manchu. This information culminates in the large handbooks, dictionaries, and encyclopedias issued during the Qing dynasty (1644–1911), often in bilingual Chinese-Manchu editions. The situation is similar for Korea, where, as in China, the primary interest was directed to Jurchen-Manchu, a language for which practical tools were needed for the purposes of translation and interpretation. Most of the extant tools were compiled in the 18th century. An additional object of documentation in Japan was offered by the Tungusic languages spoken on Sakhalin and the Lower Amur, a region to which Japanese traders and explorers travelled regularly since the beginnings of the Matsumae rule of Ezo (1590–1871). Below is a more detailed discussion of the early Far Eastern sources on Tungusic, as arranged according to their origin (China, Korea, and Japan). CHINESE SOURCES Unlike Turkic and Mongolic, Tungusic languages appear in Chinese sources quite late. This reflects the fact that the Chinese provided data in their dynastic histories either on their immediate neighbours (Xiongnu, Xianbei, Ruanruan, Tuyuhun, Turks, Japanese, Koreans, and others), or on more remote regions that were crucial for Chinese foreign policy, such as various polities in the Western Regions (Eastern and Western Turkestan). Apparently, until quite late, Tungusic peoples and languages belonged to neither of these categories, a circumstance that provides some support to the point of view that the homeland of Tungusic was located relatively far in central or northern Manchuria. It is sometimes claimed that the Yìlóu (挹婁, Late Han Chinese *Ɂɨp-l[i]o[B], Early Middle Chinese *Ɂjəp-lju[B] ~ *Ɂjəp-ləu[B]), briefly mentioned in volume 85 of the Later Han History (後漢書 Hòu Hàn Shū) were Tungusic (Pulleyblank 1983: 446), but the linguistic

DOI: 10.4324/9781315728391-2

20 Alexander Vovin

evidence for such a conclusion is lacking. The same can be said about the Mòhé (靺鞨, Late Han Chinese *mɑt-gɑt, Early Middle Chinese *mwât-ɣât) tribes, which have also been assumed to have been Tungusic speaking. There are many scattered Jurchen words in Chinese transcription in various sources starting from the Song period (960–1279), which were carefully studied in an excellent monograph by Sun Bojun (2004). The first coherent linguistic source produced by the Chinese on a Tungusic language is, however, a short vocabulary of Jurchen, Explanation of the Language of the Jin State (金國語解 Jīn guó yǔ jiě), found as an appendix after volume 135 of the History of Jin (金史 Jīn shǐ), which was compiled in 1343, that is, during the Yuan period (1271–1367). This source presents Jurchen in Chinese transcription without citing the Jurchen script. Its usefulness is limited, as approximately half of it is taken up by titles and proper nouns. However, in the remaining portion we find some usual Jurchen words, for example, 阿鄰 †alin ‘mountain’. The interesting aspect of this vocabulary is that it also lists some Jurchen words not found in other sources, for example 兀典 †ucen (Early Mandarin *ǔ-tjɛ̌n) ‘bright star’ vs. usual Jurchen 兀失哈 †ušiha (Kane 1989: 136) ~ 斡失哈 †ošiha ‘star’ (Kiyose 1977: 97), or 合喜 †hahi (Early Mandarin *xɔ́-xǐ) ‘puppy’ vs. usual Jurchen †indahu (Kane 1989: 216) ~ †indahun ‘dog’ (Kiyose 1977: 105). While †ucen ‘bright star’ has no explanation and can be a corrupted form, †hahi ‘puppy’ looks like a loan from the Old Korean predecessor of Middle Korean †kàhí ‘dog’. It should be noted that care must be taken when using the edition of this vocabulary in the Imperially Commissioned Explanation of Liao, Jin, and Yuan Words (欽定遼金元 三史語解 Qīndìng Liáo Jīn Yuán sānshǐ yǔ jiě) from the 46th year of Qianlong (乾隆) reign (1782), as this particular edition introduced numerous mistakes in the process of ‘correcting’ Jurchen data according to Manchu. During the following Ming period (1368–1644), two vocabularies of Jurchen were compiled: the Sino-Jurchen Vocabulary of the Bureau of Translators (女眞館譯語 Nǚzhēn guǎn yìyǔ, NZGYY), and the Sino-Jurchen Vocabulary of the Bureau of Interpreters (女 眞譯語 Nǚzhēn yìyǔ, NZYY). The former was first published by Grube (1896), and later, much more recently, by Kiyose (1977). It was also used by Jin Qizong in his dictionary of Jurchen (1984). The second was published by Kane (1989). The difference between the two vocabularies is that while NZGYY represents the written Jurchen language, the NZYY provides a unique glimpse into the spoken Jurchen of the time. Also, while NZYY does not include any Jurchen script, it is considerably bigger than NZGYY. With the advent of the Qing dynasty in 1644, there was a surge in the publication of Manchu dictionaries, grammars, textbooks and readers. All these sources are in fact semi-native, and not just Chinese. The main goals of those sources, especially of those published from the 1750s onward, was to help the ethnic Manchu, who were rapidly losing their native language, to retain and improve their proficiency in Manchu, as well as to help ethnic Chinese to learn the language. With few exceptions, most of these materials remain unpublished in modern times and exist only in manuscript or xylograph form. Among dictionaries the three most important ones are: Dà Qīng quán shū (DQQS 大 清全書) / Daicing gurun i yooni bithe ‘A Complete Book of the Great Qing’ (1683), the first dictionary of the Manchu language, in two volumes comprising fifteen fascicles; Qīng wén huì shū (QWHS, 清文彙書) / Manju isabuha bithe ‘A Classified Book of the Qing language’ (1750), in one volume with twelve fascicles; and Manju gisun i buleku bithe (MGBB) ‘A Mirror of the Manchu Language’, a thematic Manchu-Chinese dictionary, compiled during the second half of the Kangxi (康熙) reign (1662–1722). The last dictionary was subsequently revised and enlarged in 1772 and republished under the

Early Far Eastern sources on Tungusic 21

new title Yù zhì zēng dìng Qīng wén jiàn (御製増訂清文鑑) / Han i araha nonggime toktobuha Manju gisun i buleku bithe ‘A Revised and Enlarged Mirror of the Manchu Language Written by the Emperor’ (HANTMGBB), which remains the largest of all Manchu-Chinese dictionaries. In the following years of the Qianlong reign it was further enlarged by adding first a Mongolian, then a Tibetan, and finally a Uighur section. This latest pentaglot version, known as Wǔ tǐ Qīng wén jiàn (五体清文鑑) / Sunja gisun i hergen kamciha Manju gisun i buleku bithe ‘A Mirror of the Manchu Language Placed Together in the Letters of Five Languages’ (WTQWJ), is probably the best known of all Qing period dictionaries. It was reprinted in a typeset edition by the Nationalities Press (民族出版社 Mínzú Chūbǎnshè) first in 1957 and then again in the 1990s in three huge volumes. Later it has also been republished with five indexes by Corff et al. (2013). Among grammars of Manchu, Mǎn Hàn zǐ Qīng wén qǐ méng (満漢字清文啓蒙) / Manju Nikan hergen i Cing wen ki meng bithe ‘A Manual of Chinese and Manchu Letters’ (MHZQWQM) may be mentioned. Below is a chronological list of the Chinese sources discussed above (abbreviation, Chinese name in characters, Chinese name in transcription / Manchu name, date of compilation or publication): HHS JS NZGYY NZYY DQQS MGBB MHZQWQM QWHS YZDQWJ WTQWJ

後漢書 Hòu Hàn Shū, 5th century AZ 金史 Jīn shǐ, 1343 女眞館譯語 Nǚzhēn guǎn yìyǔ, Ming period after 1407 女眞譯語 Nǚzhēn yìyǔ, late Ming period 大清全書 Dà Qīng quán shū / Daicing gurun i yooni bithe, 1683 Manju gisun i buleku bithe, between 1692 and 1722 満漢字清文啓蒙 Mǎn Hàn zǐ Qīng wén qǐ méng / Manju Nikan hergen i Cing wen ki meng bithe, 1730. 清文彙書 Qīng wén huì shū / Manju isabuha bithe, 1750 御製増訂清文鑑 Yù zhì zēng dìng Qīng wén jiàn / Han i araha nonggime toktobuha Manju gisun i buleku bithe, 1772 五体清文鑑 Wŭ tǐ Qīng wén jiàn / Sunja gisun i hergen kamciha Manju gisun i buleku bithe, between 1772 and 1796

KOREAN SOURCES Like the Chinese, the Koreans used to be engaged in the study of Jurchen and Manchu, but not of the other Tungusic languages. The history of these studies is described by Hiu Lie (1972), who, however, focuses on Manchu, offering only a very brief account of Jurchen language studies. Fortunately, a much earlier study by Ogura Shinpei contains a succinct account of Jurchen language studies in premodern Korea (1940: 660–670), complemented by a brief but highly informative description of Manchu language studies as well (1940: 603–634). More recently, Seong Baeg-in (1970) has published a description of one of the earliest Manchu inscriptions, Samjŏndobi (三田渡碑) ‘Samjŏndo inscription’ (SJDB), which is now standing in the yard of a residential area in southern Seoul in the open air, fully exposed to the elements of weather (as a result of this “preservation” policy, the inscription is gradually becoming almost illegible). There are also many other contributions by Korean scholars regarding Jurchen and Manchu studies, among which those by Kang Sin-hang (1966) and Song Ki-Joong (1981–1982, 2000, 2001: 156–180) are the most important and representative.

22 Alexander Vovin

During the Koryŏ dynasty (918–1392) the study of the Jurchen language was chiefly conducted in Tongmungwan (通文館), the official Bureau of Interpreters and Translators of the state. This Bureau prepared and published fourteen different readers in Jurchen (Ogura 1940: 668–669). Unfortunately, not a single one of them is extant today. Should even one of them turn out to have survived, it would become a significant addition to the extremely limited corpus of Jurchen texts. The earliest extant Jurchen inscription in Korea is Kyŏngwŏnbi (慶源碑) ‘Kyǒngwǒn inscription’ (KW), which was originally located in Korea’s North Hamgyŏng province (which at the time when it was made was not Korean but Jurchen territory), but which was moved in 1918 (during Japanese colonial rule) to the National Museum of Korea. Apart from this, the earliest Korean text Yongbi ŏch’ŏnga (龍飛御天歌) ‘Song of Dragons Flying in Heaven’ (YŎ), written in the national alphabet Han’gŭl (한글) in 1445, preserves several Jurchen words in Korean transcription in the commentary. Unfortunately, practically all of them are either place names or proper nouns. Manchu studies in Korean fared much better in terms of preservation of sources. Out of multiple readers and dictionaries compiled during the Yi dynasty (1392–1910), a majority is still extant. Dictionaries from this period include: Tongmun yuhae (TY, 同文類解), ‘Explanation of Identical Words’ in two volumes; Han-Ch’ŏng munkam (HCM, 漢淸文 鑑) ‘A Mirror of Chinese and Manchu Words’ (HCM), also known as Han-Han-Ch’ŏng munkam (HHCM, 韓漢淸文鑑) ‘A Mirror of Korean, Chinese and Manchu Words’ in fifteen volumes, with a facsimile edition by Min Yŏng-gyu (1956); Pangŏn cipsŏk (PC, 方 言輯釋) ‘A Collection and Explanation of Dialects’ in four volumes; and Samhak yŏg.ŏ (SY, 三學譯語) ‘The Trilingual Translations’ (SY) in six volumes. The Pangŏn cipsŏk and the Samhak yŏg.ŏ also include Mongolian and Japanese. As far as readers are concerned, the largest and most important ones are Sam.yŏk ch’onghae (SC, 三譯總解) ‘A General Explanation of Trilingual Translations’ in ten volumes and Ch’ŏng.ŏ Nogŏltae Ŏnhae (CNŎ, 淸語老乞大諺解) ‘A Manchu Language Laojida’ in eight volumes, later published as Ch’ŏng.ŏ Nogoltae Sinsŏk (CNSS, 淸語老乞大新釋) ‘A New Commentary on the Manchu Language Laojida’ (CNS) also in eight volumes, as published in a facsimile edition with an index of Korean word forms by Chŏng Kwang (1998, cf. also Tsumagari 1977–1978). Smaller readers of Manchu include So-a ron (SAR, 小兒論) ‘Small Children’s Discourse’ and P’alse-a (PSA, 八歳兒) ‘Eight-year-old Children’, each of which has only one volume. It is difficult to overestimate the importance of the Korean eighteenth-century materials on Manchu, for in them we have foreign transcriptions of Manchu in a language that uses an alphabetic system of writing. Thus, the most important information provided by these transcriptions is on the history of Manchu phonology and phonetics throughout the eighteenth century. It has to be added that while some pioneering research has been done by Korean scholars, much of the research carried out so far has focused on the history of Korean language, rather than of Manchu. Therefore, these materials remain a virtually untouched gold mine for Manchu studies, awaiting a systematic and detailed analysis. The Korean sources discussed above are listed below in chronological order (abbreviation, Korean name in Chinese characters, Korean name in transcription, date of compilation or publication): KW YǑ

慶源碑 Kyŏngwŏnbi, between 1138 and 1153 龍飛御天歌 Yongbi ŏch’ŏnga, 1445

Early Far Eastern sources on Tungusic 23

SJDB CNǑ SAR TY SC CNS HCM PSA PC SY

三田渡碑 Samjŏndobi, 1639 淸語老乞大諺解 Ch’ŏng.ŏ Nogŏltae Ŏnhae, 1703 小兒論 So-a ron, 1720 (?), later edition 1774 同文類解 Tongmun yuhae, 1748 三譯總解 Sam.yŏk ch’onghae, 1703, second revised edition 1774 淸語老乞大新釋 Ch’ŏng.ŏ Nogoltae Sinsŏk, 1765 漢清文鑑 Han-Ch’ŏng munkam, 1775 (?) 八歳兒 P’alse-a, 1777 方言輯釋 Pangŏn cipsŏk, 1778 三學譯語 Samhak yŏg.ŏ, 1789

JAPANESE SOURCES The Japanese chronicle Azuma kagami (AK, 吾妻鏡), ‘Azuma Mirror’ preserves the earliest known foreign transmission of a Jurchen text in original script, †GURun-ni KADAgun ‘bona fide of the country’ (Kiyose 1991: 364–370). Apart from this early curiosity, and unlike the situation in China and Korea, there was never any systematic study of either Jurchen or Manchu in premodern Japan. This, however, is well compensated by a number of unique 18th to 19th-century Japanese sources on the Amur Tungusic languages Uilta (Orok) and Ulcha (Santan). These sources include Karafuto nishi okuchi risū sho (カラフト西奥地理数書) ‘A Record of a Survey of the Far Western District of Sakhalin’ (KNORS); Hokuchi risū torishirabe sho (北地 理数取調書) ‘A Report of a Survey of the Northern Land’ (HRTS); Ezo zōshi (蝦夷 草紙) ‘Essays on Ezo’ (EZ); Karafuto-jima zakki (瓦喇弗吐島雑記) ‘Various Notes on Sakhalin’ (KZ); Karafuto-jima chiri Santan Rosia-koku kikigaki (唐太島地理三 靼魯斎亜國聞書) ‘Records of Sakhalin, Santan, and Russia’ (KCSRK); Ezo fudoki (Seki) (蝦夷風土記―関) ‘Ezo Gazeteers (Seki)’ (EF-S); Ezo fudoki (Niiyama) (蝦 夷風土記―新山); ‘Ezo Gazeteers (Niiyama)’ (EF-N); Karafuto zakki (Nakamura) (唐太雑記―中村) ‘Various Notes on Sakhalin’ (KZ-N); Hen’yō bunkai zukō (辺要 分界図考) ‘A Study on the Geography of Important Outlying Areas’ (HBZ); Ezo go shū (蝦夷語集) ‘A Collection of Ezo Words’ (EGS) by Uehara Kumajirō (上原熊次 郎), mainly containing data on Ainu; Shindan Karafuto kaitōki (新談カラフト廻島 記), ‘Newly Discussed Records on Going around Sakhalin’ (SKK), containing only names of trade items; Karafuto shima shinsetsu (哈喇土島新説) ‘A New Account of Sakhalin Island’ (KSS); Karafuto nikki (唐太日記) ‘Sakhalin diary’ (KN), containing only numerals; and Santan goi (山丹語彙) ‘A Santan Vocabulary’ (SG). There is also a vocabulary collected by Mogami Tokunai (最上徳内) and contained in the manuscript titled “Einige Nachrichten über Krafto und Sandan, ein Auszug aus dem Tagebuch meines alten Freundes Mogami Tok’nai auf seinen Reisen dahien” by Ph. Fr. von Siebold. These sources were studied in great detail by Ikegami Jirō, the doyen of Tungusic studies in modern Japan (Ikegami 2002abc). Unfortunately, in spite of their relatively large volume, the value of these early materials on Uilta and Ulcha is somewhat reduced by the fact that they are all written in the Japanese katakana syllabic alphabet. Due to the syllabic nature of the transcriptions and to the fact that Japanese has much simpler phonotactics than any Tungusic language, most of the transcriptions are imprecise and sometimes simply misleading. For this reason, the Japanese transcriptions should only be used in conjunction with the more recent data available in the Cyrillic or Latin alphabets.

24 Alexander Vovin

The Japanese sources discussed above are listed below in chronological order (abbreviation, Japanese name in original script, Japanese name in transcription, date of compilation or publication): AK EF-N EZ KZ KCSRK

吾妻鏡 Azuma kagami, 1266 蝦夷風土記―新山 Ezo fudoki (Niiyama), 1789 蝦夷草紙 Ezo zōshi, 1790 瓦喇弗吐島雑記 Karafuto-jima zakki, 1790 唐太島地理三靼魯斎亜國聞書 Karafuto-jima chiri Santan Rosia-koku kikigaki, 1792 EF-S 蝦夷風土記―関 Ezo fudoki (Seki), 1792 KZ-N 唐太雑記 Karafuto zakki (Nakamura), 1801 HBZ 辺要分界図考 Hen’yō bunkai zukō, 1804 EGS 蝦夷語集 Ezo go shū, no date, but probably before 1811 HRTS 北地理数取調書 Hokuchi risū torishirabe sho, mid 19th c. KNORS カラフト西奥地理数書 Karafuto nishi okuchi risū sho, mid 19th c. KN 唐太日 Karafuto nikki, 1854 KSS 哈喇土島新説 Karafuto shima shinsetsu, 1854 SKK 新談カラフト廻島記 Shindan Karafuto kaitōki, 1854 SG 山丹語彙 Santan goi, compiled between 1882 and 1886, but probably containing data collected earlier REFERENCES AND FURTHER READING Chŏng, Kwang 鄭光 (ed.) (1998)『淸語老乞大新釋』[A new explanation of the Manchu language Laojida], Seoul: Taehaksa. Corff, Oliver & Kyoko Maezono & Wolfgang Lipp & Dorjpalam Dorj & Görööchin Gerelmaa & Aysima Mirsultan & Réka Stüber & Byambajav Töwshintögs & Xieyan Li (eds.) 2013, Auf kaiserlichen Befehl erstelltes Wörterbuch des Manjurischen in fünf Sprachen—„Fünfsprachenspiegel”, vols. 1–6, Wiesbaden: Harrassowitz. Grube, Wilhelm (1896), Die Sprache und Schrift der Jučen, Leipzig: Kommissions-Verlag von O[tto] Harrassowitz. Ikegami, Jirō 池上二良 (2002a)「十九世紀なかごろのオロッコ語彙－サンタン 語・ギリャク語を含む」[A Vocabulary of mid-nineteenth century Orok—Including Santan and Ghilyak words], in:『ツングース・満州語資料訳解』[Translation and a commentary to Tungusic and Manchu language materials], 154–260, 札幌 [Sapporo]: 北海道大學出版会. Ikegami, Jirō 池上二良 (2002b)「サンタンことば集」[A collection of Santan (words], in:『ツングース・満州語資料訳解』[Translation and a Commentary to Tungusic and Manchu Language Materials], 261–321, 札幌 [Sapporo]: 北海道大学出版会. Ikegami, Jirō 池上二良 (2002c)「サンタンことば拾遺」[An additional collection of Santan words], in:『ツングース・満州語資料訳解』[Translation and a Commentary to Tungusic and Manchu Language Materials], 322–334, 札幌 [Sapporo]: 北海 道大学出版会. Jin, Qizong 金啓孮 (1984)『女真文辞典』[A dictionary of the Jurchen language], 北 京 [Peking]: 文物出版社. Kane, Daniel (1989) The Sino-Jurchen Vocabulary of the Bureau of Interpreters, Indiana University Uralic and Altaic Series 153, Bloomington, Indiana: Indiana University, Research Institute for Inner Asian Studies.

Early Far Eastern sources on Tungusic 25

Kang, Sin-hang 妻信沆 (1966)「李朝中期以後의 譯者에 對한 考察」[A commentary on the translators from the mid-Yi dynasty period],『成均館大學校論文集』11. Kiyose, Gisaburō N. (1977) A Study of the Jurchen Language and Script, Kyoto: Hōritsubunka-sha. Kiyose, Gisaburō N. (1991) ‘The significance of the new Kitan and Jurchen materials’, in:『日本語學とアルタイ語學』[Japanese Linguistics and Altaic Linguistics], 359– 377, Tokyo: Meiji shoin. Lie, Hiu (1972) Die Mandschu-Sprachkunde in Korea, Indiana University Uralic and Altaic Series 114, Bloomington: Indiana University Press. Min, Yŏng-gyu 閔泳珪 (ed.) (1956)『韓漢淸文鑑』[A mirror of Korean, Chinese and Manchu words], vol. 1: Facsimile, vol. 2: Index, Seoul: 延世大學東方學研究所. Ogura, Shinpei 小倉進平 (1940)『朝鮮語學史』[A history of linguistics in Korea], 東 京[Tokyo]: 刀江書院. Pulleyblank, Edwin G. (1983) ‘Chinese and their neighbors in prehistoric and early historic times’, in: David N. Keightley (ed.), The Origins of Chinese Civilization, 411– 466, Berkeley, Los Angeles: University of California Press. Seong Baeg-in 成百仁 (1970) 三田渡碑滿洲文 [The Manchu inscription of Samjŏndobi], Tong’a munhwa 東亞文化9: 117–148 [Seoul]. Song, Ki-Joong (1981–1982) ‘The study of foreign languages in the Yi dynasty (1392– 1910)’, I–III, Journal of Social Sciences and Humanities 54: 1–45, 55: 1–63, 56: 1–57, Seoul: The Korean Research Center. Song, Ki-Joong 宋基中 (2000) 「朝鮮時代女眞學/清學」[The study of Jurchen/Manchu in the Chosŏn dynasty], Altai Hakpo『알타이학보』10: 105–119, [Seoul]. Song, Ki-Joong (2001) The Study of Foreign Languages in the Chosŏn Dynasty (1392– 1910), Korean Studies Dissertation Series No. 1, Seoul: Jimoondang International. Sun, Bojun 孙伯君 (2004)『金代女真語』[The Jurchen Language of the Jin Period], 瀋 陽 [Shenyang]: 遼寧出版社. Tsumagari, Toshirō 津曲敏朗 (1977–1978)「清語老乞大の研究：満州語研究のため の一資料(1), (2)」[A study of Laojida in the Manchu language: Materials for the study of Manchu],『札幌商科大学・札幌短期大学論集（人文編）』21: 211– 248, 22: 161–169, 札幌 [Sapporo].

CHAPTER 3

EARLY WESTERN SOURCES ON TUNGUSIC José Andrés Alonso de la Fuente

There is an abundance of Tungusic language material hidden in early western sources dating from the 17th century to the first part of the 19th century. With few exceptions, most of them await to be properly edited according to current philological standards. This task must be approached in tandem with similar initiatives aimed at publishing the even more important unpublished sources from the second part of the 19th century onwards. For illustrative purposes, A. A. Burykin (2006: 52–53) mentions several dozens of unpublished recordings from the early 20th century stored at only a single institution in Russia, the Phonogrammic Archives of the Institute of Literature of the Russian Academy of Sciences. The present account will focus exclusively on Western sources from the late 17th to the early 19th century. These sources are the result of endeavours initiated and accomplished for the most part by French, Dutch, or German travellers and scholars active in various roles in Imperial Russia. For obvious reasons, the resulting language data usually cover a number of Northern Tungusic lects (mostly Ewenki and Ewen) and/or Manchu. In the following, the focus will be on non-Manchuric sources, for the history of early Manchu studies in Europe is relatively well known (cf. e.g. Söderblom Saarela 2020). The selection of sources included in the list below is inevitably somewhat subjective. For instance, Rüdiger (item #12), whose only contribution, some might argue, was to spread further Strahlenberg’s (item #5) mistaken identification of Manchu with Chinese, is included since his publication illustrates in no ambiguous terms the many pitfalls of working with early sources. To be more precise, the 26-item list below includes the most important works on Tungusic produced in the period of time that goes from 1692, the year in which Nicolaes Witsen’s (1641–1717) travelogue was first published, to 1856, when the language data collected by M. A. Castrén (1813–1852) saw finally the day of light. After this period the scene of Tungusic studies was long dominated by Russian (including Soviet) publications. The production of new materials or even the re-edition of previously recorded data outside of Russia comes to a stop during the second half of the 19th century, with only a handful of exceptions, including, notably, the publication in Paris of the collections of Gerhard von Maydell and Alexander Czekanowski by Anton Schiefner (1874, 1877), who had earlier been instrumental in making Castrén’s materials available to the scholarly world. The value of Castrén’s work as a terminus post quem marking the starting point of the scientific study of the Tungusic languages hardly requires justification. Cincius (1975: xxv) refers to Castrén’s data (item #26 in the list below) as the only early

DOI: 10.4324/9781315728391-3

Early Western sources on Tungusic 27

contribution worth being included among the materials of her Tungusic comparative dictionary. There were, of course, plenty of reasons behind this chronological limit: for one thing, the comparative dictionary was primarily based on new field materials collected during the Soviet period, but also, many early western sources were not easily accessible at the time. However, such limitations are not actual any more, which is why it is relevant to take a new look at these sources and to see what they can offer for historical and comparative studies of Tungusic. TYPES OF SOURCES The items described below can be classified variously according to their origin, form, or status of publication. With regard to their origin, sources can be either primary or secondary. The former type contains material which, generally speaking, was collected in situ either by the author or by a second party, whereas the latter type involves the reproduction of previously published materials, with or without acknowledgement to the source(s), and often in the form of compilation from several different sources, both primary and secondary. Errors of various kinds, including misunderstandings, misreadings, and typos, abound in many sources of this type, and what appear to be conscious modifications, especially with regard to spelling practices, are sometimes introduced without explanation. For example, Klaproth’s (item #22) re-use of Messerschmidt’s (item #4) data shows numerous discrepancies, as in the first five numerals, which are given by Klaproth as 〈mûkoon, djuhr, Ilánn, degénn, tóŋa〉, while Messerschmidt writes 〈Mukónn, Djuhr, Ilánn, Degénn, Tónga〉. One of the first tasks of the philologist is to identify and account for these and other such anomalies. As for the form of the sources, they can be divided into two types: explicit and non-explicit. The explicit type comprises sources that aim at offering language data in a systematic arrangement, for instance, in the form of a wordlist. Such sources are typically the result of organized linguistic fieldwork. Non-explicit sources, by contrast, contain language data, often individual technical terms or random phrases, reflecting casual observations, and scattered throughout the body of a text, or also in the appendices to a text dealing with topics other than language in the first place. Sources of the latter type require the researcher to “recollect” the data and present them in a more transparent (explicit) manner. With regard to their status of publication, sources are either published or unpublished. The list below contains mainly materials that have been published, either by the original author soon after the date of collection or later by other authors. Publication in this context does not, however, mean the presence of a proper critical edition corresponding to modern philological standards, which is available in only a few cases. Needless to say, there is the possibility that there exist hitherto unknown unpublished sources preserved in manuscript form in various archives, especially in Russia. Even so, it is unlikely that any major early collection of Tungusic data would have remained unknown up to the present day. Below, the items are grouped on a rough chronological basis into eighteenth-century and nineteenth-century materials. Before Castrén there was little difference in the motives and approaches of the persons doing the documentation, but even so the eighteenth century, dominated by the era of Enlightenment, was characterized by an endeavour to discover the diversity of the world, while the nineteenth century, moving towards the era of

28 José Andrés Alonso de la Fuente

Romanticism, was more focused on the historical connections and origins of phenomena, including languages, a focus that functioned as the ultimate stimulation also for Castrén. The main goal of the present list is to offer a starting point for future research. It is to be hoped that the ready availability of some of these materials in digitalized form on the Internet today will result in them being properly edited and analysed in the future. Bibliographic information aside, the list below includes information about (i) the type and amount of language material contained in each given source, (ii) any subsequent reproduction of the language material in other publications, indicated by the arrow (→), as well as (iii) any modern philological analysis done on the basis of the source. EIGHTEENTH CENTURY 1. 1692: Nicolaes Witsen, Noord en Oost Tartarye, Ofte bondigh ontwerp Van eenige dier landen en volken, zo als voormaels bekent zyn geweest, vols. 1–2. Amsterdam, 1692; 2nd ed. Amsterdam: François Halma, 1705; reprint Amsterdam: M. Schalekamp, 1785. (References below are provided according to the 2nd edition of 1705.) • •

Vol. 1, between p. 6 and p. 7: Manchu fragment. Vol. 2, p. 654: Ewenki (Lord’s Prayer, absent in the 1st ed. of 1692). → → → → →

•

Chamberlain, Oratio dominica . . . , p. 14 (item #2). Leibniz, Collectanea etymologica . . . , Pars 2, pp. 374–375 (item #3). Schultze, Orientalisch- und Occidentalischer . . . , p. 108 (item #8). Adelung, Mithridates . . . , vol. 4, pp. 223–225 (item #20). Marcel, Oratio dominica . . . , p. 31 (item #17).

Vol. 2, p. 678: Ewen numerals.

Remarks: general descriptions claim that this work contains “about four dozen” words in Ewenki and Ewen (see, e.g., Gorcevskaya 1959: 6; Clark 1981–1982: 70–71 mentions only “Ewen”), though there are no wordlists as such. The language data (Ewenki, Ewen) were collected at some point between 1666 and 1677. For the analysis of the Ewenki, Ewen and Manchu data, see Pevnov (2018a, 2018b) and Golvers (2018). 2. 1715: Joanne Chamberlaynio [John Chamberlain], Oratio dominica in diversas omnium fere gentium linguas versa. Amstelædami [Amsterdam]: Typis Guilielmi & Davidis Goerei. Remarks: see item #1. 3. 1717: Godofr. Gvilielmi Leibnitii [Gottfried Wilhelm Leibniz], Collectanea etymologica illustrationi linguarum, veteris celticae, germanicae, gallicae aliarumque inservientia, 1–3. Hanoveræ: Sumptibus Nicolai Foersteri. Remarks: see item #1. 4. 1720–1727: Daniel Gottlieb Messerschmidt, Forschungsreise durch Sibirien 1720–1727, Quellen und Studien zur Geschichte Osteuropas [Eduard Winter, Georg Uschmann, Günther Jarosch (eds.)], vols. 1–5. Berlin: Akademie-Verlag, 1962–1977. •

Vol. 2: pp. 72–73 (“Orotong-Tungusicae”), 89–90 (“Tungusische Sprache”). → Klaproth, Asia polyglotta, pp. 286–288 (item #22).

Early Western sources on Tungusic 29

→ Balbi, Atlas ethnographique, p. 37 (as “de la Tungouska-Inférieure”) (item #23). Remarks: According to Messerschmidt’s travel diary, one Ewenki helped him to write down the language samples (c. 160 words), which were gathered between June 17 and July 11, 1723. 5. 1730: Philipp Johann von Strahlenberg, Das Nord- und Östliche Theil von Europa und Asia. Stockholm: in Verlegung des Autoris. •

sub “Harmonia linguarum”: “Mantischeou” = Chinese → Pallas, Linguarum totius orbis vocabularia . . . , #170 in Appendix (item #14). → Rüdiger, Grundriß . . . , pp. 94–95, §195: “Die Sprache der Mantschu” (item #12). “Tongusi-konni” = (Spoken) Manchu “Tungusi-sabatschi alias Lamuti” = Ewen “Tongusi-oleni” = Ewenki → Schultze, Orientalisch- und Occidentalischer . . . , pp. 212–219 (item #8). → Hervás, Aritmetica delle nazioni . . . , pp. 151–152 [331–332, 336] (item #13).

Remarks: See Krueger (1975). Clark (1977–1978) discusses the (Spoken) Manchu data. 6. 1736–1742: Gerhard Friedrich Müller, Nachrichten über Völker Sibiriens (1736– 1742). Manuscript, only partially published. Remarks: The Southern Ewenki data from “husshing” dialects (c. 140 words), scattered across Müller’s writings, are collected in Helimski & Katz (2003: 237–240). 7. 1742: Jakob Johann Lindenau, Beschreibung der Peschie Tungusen, oder so genannte[n] Lamuten zu Ochot, 1742. Manuscript, only partially published. Remarks: Lindenau’s report contains an Arman-German glossary (550 words). Lindenau refers to the Arman as “Tungusy Udskogo Ostroga”). See Titova (1983: 53–70, Russian translation, and 71–76, Arman-German glossary). 8. 1748: Benjamin Schultze, Orientalisch- und Occidentalischer Sprachmeister. Leipzig: zu finden bey Christian Friedrich Gessnern (2nd ed. 1769). p. 204: Ewen (“Lamutische”) numerals → Hervás, Aritmetica delle nazioni . . . , pp. 151–152 [333] (item #13). Remarks: See items ##1, 5. It is unclear what the provenance of Schultze’s Ewen numerals was. They differ significantly from the Ewen numerals found in other early sources (see the comparative table below). It is open to speculation whether Schultze took them from Witsen; he could have “corrected” and modified them to his liking. Witsen confuses ‘12’ and ‘13’, but this is corrected in Schulze. Strahlenberg’s ‘2’ is actually ‘7’, while his 〈Dagalkun〉 is not even a numeral, cf. 〈tagalkun〉

30 José Andrés Alonso de la Fuente

in Hervás, who carries on with the semantic confusion. However, there are other differences that cannot be explained so easily, cf. e.g. Witsen 〈Dziakondzian〉 ‘15’ vs. Schulze 〈Dongandzian〉, Witsen 〈Muginadzian〉 ‘30’ vs. Schultze 〈Muginadziandzialakan〉, etc. Fischer

   

Witsen

Messerschmidt

Strahlenberg

Schultze

Hervás

I

IIII

1

Omun.

Mukónn.

Omokon

Omun

omum.

úmmukon

omon

2

Dzur.

Djuhr.

Nudan

Dzur

dzur.

dsjur

dschur

3

Ilan.

Ilánn.

Dagalkun

Jean

jean.

illán

ilàn

4

Dagan.

Degénn.

Ullan

Dagan

dagan.

díggin

dyg’yn

5

Tongan.

Tóņga.

Degen

Dongan

dongan. túngja

tongna

9. 1747–1768: Johann Eberhard Fischer, Vocabularium continens trecenta vocabula triginta quator gentium, maxima ex parte Sibiricarum. Manuscript, only partially published. • • •

Section I: Tungus River, column 9. Section III: Manchu, column 3. Section IIII: Nerchinsk & Selenga Ewenki, column 2. → → → →

von Schlözer, Allgemeine Nordische Geschichte, p. 418 (item #10). Georgi, Bemerkungen einer Reise . . . , vol. 1, pp. 268–271 (item #11). Klaproth, Reise in den Kaukasus . . . , vol. 2, pp. 553–555 (item #19). Klaproth, Asia polyglotta, Sprachatlas, column XXXXII (item #22).

Remarks: Appr. 300 words, including numerals, in (at least) four lects. The language data were collected around 1725. For facsimile reproductions, see Doerfer (1965: 47–69, 96–117, 119–139) and Gulya (1995: 27–28, 192). 10. 1771: August Ludwig von Schlözer, Allgemeine Nordische Geschichte. Halle: bey Johann Justinus Gebauer. Remarks: see item #9. 11. 1775: Johann Gottlieb Georgi, Bemerkungen einer Reise im Rußischen Reich im Jahre 1772 [in den Jahren 1773 und 1774], vols. 1–2. St. Petersburg: Gedruckt bey der Kayserl. Akademie der Wisseschaften. Remarks: see item #9. 12. 1782: Johann Christian Christoph Rüdiger, Grundriß einer Geschichte der menschlichen Sprache nach allen bisher bekannten Mund- und Schriftarten mit Proben und Bücherkenntniß. Leipzig: bey P. G. Kummer. Remarks: see item #5. 13. 1786: Lorenzo Hervás, Aritmetica delle nazioni e divisione del tempo fra l’orientali. Cesena: Per Gregorio Biasini all’ Insegna di Pallade. Remarks: see item #5.

Early Western sources on Tungusic 31

14. 1786–1789: [Peter Simon Pallas,] Linguarum totius orbis vocabularia comparativa, vols. 1–2. Petropoli: Typis Iohannis Caroli Schnoor. Facsimile reprint: Hamburg: Helmut Buske Verlag, 1977–1978. → Klaproth, Verzeichniss . . . , pp. 72–89 (Pallas’ columns 138–145, 163) (item #21). → Klaproth, Asia Polyglotta, Sprachatlas, column XXXXII (item #22). → Balbi, Atlas ethnographique, Sixième tableau (Pallas’ columns 138–146) (item #23). Remarks: The Tungusic lects included in Pallas (columns 138–146 cum 163, and 143–151 cum 159 in the appendix about the numerals; note that column 170 refers to Strahlenberg’s “Chinese”, see item #5) represent only three varieties: Ewenki (along with Chapogir), Ewen, and Manchu. The popularity of this compilation is well known. Eichhorn (1807: 400–403), who discusses Daghur and other non-Tungusic materials, ignores that #146 in Pallas corresponds also to Tungusic. 15. 1790: Jean Baptiste Barthélemy de Lesseps, Journal historique du voyage de M. de Lesseps, [. . .]; Depuis l’instant où il a quitté les frégates Françoises au port SaintPierre & Saint-Paul du Kamtschatka, jusqu’à son arrivée en France, le 17 octobre 1788, vols. 1–2. Paris: de l’imprimerie royale. •

Vol. 2, pp. 356–375: Appr. 500 Ewen (“Lamoute”) words in a comparative table which includes French, Russian, Itelmen, Koryak, and Chukchee as well. → Sauer, An Account . . . , Appendix no. I [pp. 1–8] (item #16).

Remarks: see item #16. NINETEENTH CENTURY 16. 1802: Martin Sauer, An Account of a Geographical and Astronomical Expedition to the Northern Parts of Russia, by Commodore Joseph Billings. London: Printed by A. Strahan. Remarks: see item #15. The label “Tungoose” is somewhat misleading, for the language represented in Sauer’s wordlist is Ewen, rather than Ewenki. This is also obvious, for we know that Sauer’s original source was de Lesseps (item #15). Thus, Sauer has Fish = 〈Olra〉, Sun = 〈Nultian〉, Earth = 〈Tor〉, Sea = 〈Nam〉, which correspond to “Standard” Ewen olra, ñɵɵlten, tuur, nam, cf. “Standard” Ewenki ollo, dilacaa, dunne, laamu, respectively. For whatever reason, Sauer opted not to profit from the Ewen wordlist collected by another member of the expedition, surgeon major Michael Robeck, as reproduced in Sarychev’s account (see item #18). Thus, de Lesseps has 〈Kh-Elonn〉 ‘nostrils’ = Sauer 〈K-elon〉 vs. Sarychev 〈Будолъ〉, de Lessep’s 〈Golbani〉 ‘night’ goes with Sauer’s 〈Golban〉 vs. Sarychev 〈Долбу〉, etc. However, in copying de Lesseps’ data, Sauer made numerous mistakes. For example, instead of de Lesseps’ 〈Nioungann〉 ‘6’ Sauer has 〈Kilkok〉 (cf. Sarychev’s 〈Юнгенъ〉), which actually corresponds to de Lesseps’ “Kamtschadale” (that is, Itelmen) 〈killk-okk〉, de Lesseps’ 〈Nong annioubeï〉 ‘he’ becomes Sauer’s 〈Nongenatsche〉, which corresponds to de Lesseps’ 〈Nong ann achi〉 ‘she’, etc.

32 José Andrés Alonso de la Fuente

17. 1805: Jean Joseph Marcel, Oratio Dominica CL linguis versa, et propriis cujusque linguae characteribus plerumque expressa, Parisiis: Typis Imperialibus. • •

p. 25: Tataro-Mantchou. p. 31: Tungusiane (Lord’s Prayer).

Remarks: see item #1. Marcel informs himself that he had access to Manchu through the missionaries in Peking and L. Langlès in 1790. 18. 1811: G. A. Sarychev [Г. А. Сарычевъ], Путешествiе капитана Биллингса Чрезъ Чукотскую землю отъ Берингова пролива до Нижнеколымскаго острога, [. . .] в 1791 году; Санктпетербургь [St. Petersburg]: Въ Морской Типографiи. •

pp. 93–102: Appr. 300 (Kolyma) Ewen words.

Remarks: Gorcevskaja (1959: 8). Although Sarychev, Sauer and Robeck were members of the same expedition, Sarychev published Robeck’s Ewen wordlist, whereas Sauer (item #16) opted to copy de Lesseps’ language data (item #15). 19. 1812–1814: Julius Heinrich von Klaproth, Reise in den Kaukasus und nach Georgien unternommen in den Jahren 1807 und 1808, vols. 1–2. Halle und Berlin: in den Buchhandlungen des Hallischen Waisenhauses. Remarks: see item #9. 20. [1806–]1817: Johann Christoph Adelung, Mithridates oder allgemeine Sprachenkunde, vols. 1–4. Berlin: In der Vossischen Buchhandlung. Remarks: see item #1. 21. 1822: Julius Heinrich von Klaproth, Verzeichniss der chinesischen und mandschuischen Bücher und Handschriften der Königlichen Bibliothek zu Berlin. Paris: In der Königlichen Druckerei. Remarks: see item #14. 22. 1823: Julius Heinrich von Klaproth, Asia polyglotta. Paris: J. M. Eberhart, 2nd ed. 1831. • pp. 286–300: “Tungusen”. • Sprachatlas, column XXXXII. Remarks: see items ##4, 9, 14. 23. 1828: Adrien Balbi, Atlas ethnographique du globe, ou classification des peoples anciens et modernes d’après leurs langues. Paris: chez Rey et Gravier. • • •

pp. 8 and 37: “Famille toungouse” sub “Groupe des langues tartares”: column 119 Mandchoue. column 120 Toungouse (Jeniseisk, Tchapogires, Mangaseja, Nertschunsk, Bargusin, Angara-Supérieure, Jakouzk, Okhotsk, Lamoutes, Tungouska-Inférieure).

Remarks: See items ##4, 14. 24. 1832: Philipp F. B. von Siebold, ‘Verhandeling over de afkomst der Japanners’, Verhandelingen van het Bataviaasch Genootschap van Kunsten en Wetenschappen 13: 183–275. •

after p. 275: comparative tables with language data from Chinese (three dialects), Korean, Japanese, Manchu and “Sandan”, and Sakhalin and Hokkaido Ainu.

Early Western sources on Tungusic 33

Remarks: de Charencey (1902) combines language data from von Siebold, Furet (1860), and de Sabir (1861). 25. 1848: G. Adolf Erman, Reise um die Erde durch Nord-Asien und die beiden Oceane in den Jahren 1828, 1829 und 1830, Abt. 1, Bd. 3: Die Ochozker Küste, das Ochozker Meer und die Reisen auf Kamtschatka im Jahre 1829. Berlin: verlegt bei G. Reimer. •

pp. 59–64: Appr. 100 words and the numerals in Aldan Ewenki and Manchu.

Remarks: The Ewenki data were collected in 1829. On p. 58, Erman acknowledges Wilhelm Schott (1802–1889) for the Manchu data. 26. 1856: M. Alexander Castrén, Grundzüge einer tungusischen Sprachlehre nebst kurzem Wörterverzeichniss, Herausgegeben von Anton Schiefner, Nordische Reisen und Forschungen 9, St. Petersburg: Kaiserliche Akademie der Wissenschaften. • • •

pp. 71–96: Appr. 1500 Ewenki words in the Nerchinsk, Man’kovo, and Urul’ga (historical) dialects. pp. 119–136: comparative tables with materials collected by A. F. Middendorff (1843–1845), G. I. Spasskii (1810, c. 130 words), and G. Gerstfeldt (1855). pp. 137–139: sentences collected by A. F. Middendorff.

Remarks: The language data was collected mainly in 1848, and posthumously published by Anton Schiefner, together with language samples collected by others (see description above). See Adam (1873), which is a French translation (grammar only). REFERENCES AND FURTHER READING Adam, Lucien (1873) ‘Grammaire de la langue Tongouse’, Revue de linguistique et de philologie comparée 6 (2): 129–182, (3): 217–238, Paris. Burykin, A. A. [AA. Бурыкин] (2006) ‘Материалы по языкам и фольклору тунгусоманьчжурских народов в коллекциях фонограммархива Института русской литературы РАН’ [Tungusic language and folklore materials at the Phonogrammic Archives of the Institute of Russian Literature of the Russian Academy of Sciences], in: Alessandra Pozzi & Juha Antero Janhunen & Michael Weiers (eds), Tumen jalafun jecen akû: Manchu Studies in Honour of Giovanni Stary, Tunguso-Sibirica 20: 43–54, Wiesbaden: Harrassowitz. Cincius, V. I. [В. И. Цинциус] (ed.) (1975–1977) Сравнительный словарь тунгусо-​ маньчжурских языков: Материалы к этимологическому словарю [Comparative diactionary of the Tungusic languages: Materials for an etymological dictionary], vols. 1–2, Ленинград [Leningrad]: “Наука”. de Charencey [Hyacinthe] (1902) ‘Sur quelques dialects est-altaïques’, Journal asiatique 19: 150–153, Paris. Clark, L[arry] V. (1977–1978) ‘Some Spoken Manchu words in Strahlenberg (1730)’, Manchu Studies Newsletter 1–2: 91–95, Bloomington. Clark, L[arry]. V. (1981–1982) ‘Manchu-Tungus Lexicography’, Manchu Studies Newsletter 4: 45–90, Bloomington. Also published in: Walravens (2006: 113–171). Doerfer, Gerhard (1965) Ältere westeuropäische Quellen zur kalmückischen Sprachgeschichte (Witsen 1692 bis Zwick 1827), Asiatische Forschungen 18, Wiesbaden: Harrassowitz Verlag. Eichhorn, Johann Gottfried (1807) Geschichte der neuern Sprachenkunde, Göttingen: Vandenhoek und Ruprecht.

34 José Andrés Alonso de la Fuente

Furet, P[ère Louis] (1860) Letters a M. Léon de Rosny sur l’Archipel Japonais et la Tartarie Orientale, Paris: Maisonneuve & Cie. Golvers, Noël (2018) ‘Verbiest’s Manchu fragment’, in: Naarden & van Brederode & de Graaf & al. (eds.) (2018), 411–420. Gorcevskaya, V. A. [В. А. Горцевская] (1959) Очерк истории изучения тунгусоманьчжурских языков [A brief history of the study of the Tungusic languages], Серия “История отечественного языкознания” [History of linguistics in Russia] 2, Ленинград [Leningrad]: Государственное учебно-педагогическое издательство Министерства просвещения РСФСР. Gulya, János (1995) Johann Eberhard Fischer: Vocabularium Sibiricum (1747): Der etymologisch-vergleichende Anteil, Frankfurt am Mein: Peter Lang. Helimski, Eugen & Hartmut Katz (2003) Gerhard Friedrich Mueller: Nachrichten über Völker Sibiriens (1736–1742), Hamburg: Universität Hamburg. Krueger, J[ohn] R. (1975) ‘Preface to the re-printing’, in: Ph. Johann von Strahlenberg, Das Nord- und Ostliche Theil von Europa und Asia, Studia Uralo-Altaica VIII: [1–11]. Szeged: Universitas Szegediensis de Attila József Nominata. Naarden, Bruno & Tom van Brederode & Tjeerd de Graaf & Wim Honselaar & Janine Jager & Cecilia Odé & Lisa van Schaik & Nicoline van der Sijs (eds.) (2018), The Fascination with Inner Eurasian Languages in the 17th Century. The Amsterdam mayor Nicolaas Witsen and his collection of ‘Tartarian’ glossaries and scripts, Amsterdam: Pegasus. Pevnov, Alexander (2018a) ‘Linguistic interpretation of the Lord’s Prayer in Evenki (Tungus)’, in: Naarden & al. (eds.) (2018), 385–400. Pevnov, Alexander (2018b) ‘Even (Lamut) Numerals’, in: Naarden & van Brederode & de Graaf & al. (eds.) (2018), 400–409. de Sabir, C[onstantin] (1861) Le fleuve Amoûr: Histoire, géographie, ethnographie, Paris: Georges Kugelmann. Schiefner, Anton (1874) ‘Baron Gerhard von Maydell’s tungusische Sprachproben’, Mélanges asiatiques 7: 359–377, Paris. Schiefner, Anton (1877) ‘Alexander Czekanowski’s tungusisches Wörterverzeichnis’, Mélanges asiatiques 8: 356–416, Paris. Söderblom Saarela, Mårten (2020) The Early Modern Travels of Manchu: A Script and Its Study in East Asia and Europe, Encounters with Asia, Philadelphia: University of Pennsylvania Press. Titova, Z. D. [З. Д. Титова] (1983) Я. И. Линденау: Описание народов Сибири (первая половина XVIII в.): Историко-этнографические материалы о народах Сибири и Северо-Востока [Jakob Ivanovich Lindenau: Description of the peoples of Siberia (first half of the 18th century): Historical and ethnographic materials about the peoples of Siberia and the Northeast], Magadan [Магадан]: Магаданское книжное издательство. Walravens, Harmut (ed.) (2006) Bibliographies of Mongolian, Manchu-Tungus, and Tibetan Dictionaries, Compiled by Larry V. Clark, John R. Krueger, Manfred Taube, Hartmut Walravens, Michael K. Walter, Orientalistik Bibliographien und Dokumentationen, Wiesbaden: Harrassowitz Verlag.

CHAPTER 4

PROTO-TUNGUSIC Juha Janhunen

Since there is no doubt about the mutual relationship of the Tungusic languages in the context of a coherent language family, the principles of comparative linguistics presuppose that there once existed a protolanguage from which all the modern forms of Tungusic descend. This protolanguage, termed Proto-Tungusic, may be assumed to have constituted a complete natural language, spoken by a concrete population, the Proto-Tungusic speech community. Like all natural languages, it must have contained internal variation of both the regional and the social type, but most traces of the former variation have been extinguished with time, and for the purposes of diachronic reconstruction Proto-Tungusic may be viewed as a more or less uniform language, which was gradually diversified due to divergent innovative developments involving both the introduction of secondary (new) features (positive innovations) and the loss of original (old) features (negative innovations) in the subsequent branches, subbranches, individual languages, and dialectal varieties of Tungusic. The divergent developments that have taken place after the breakup of Proto-Tungusic have inevitably obscured some of its properties, which is why our knowledge of it will always remain incomplete. Even so, thanks to the relatively large number of the modern Tungusic varieties and their relatively close mutual affinity, the picture that can be formed of Proto-Tungusic is reasonably detailed and, as it seems, reliable. By tracking the shared innovations in the branches and subbranches, we can also to some extent reconstruct a number of intermediate protolanguages, corresponding to the internal taxonomy of the language family. However, while the Ewenic, Orochic, Nanaic, and Jurchenic branches all show an unambiguous set of branch-specific innovations, the features that distinguish Northern Tungusic and Southern Tungusic from each other are fewer in number. This suggests that the initial binary division of Proto-Tungusic into a northern and a southern branch was rather soon followed by the further split of these branches into the four subbranches, each of which has a considerable time depth of independent evolution. As it is, the structural and material properties of Proto-Tungusic can be reconstructed in considerable detail as far as both phonology and morphology are concerned. The syntax of the protolanguage can also be approached on the basis of the reconstructed morphological categories, as well as by comparing the typological features of the modern languages. In the lexicon, it is possible to establish a large corpus of items inherited from the protolanguage, though the relatively close mutual relatedness of the Tungusic languages may occasionally make it difficult to distinguish inherited lexical items from secondary borrowings between the modern branches and languages. This is also true of morphological elements, which can have an areal history and modern distribution distinct from that of lexical items, an extreme example being offered by the “mixed” languages Kili and Kilen. DOI: 10.4324/9781315728391-4

36 Juha Janhunen

DATA AND SOURCES The two principal published sources on Proto-Tungusic until the present day are the classic comparative analyses of V. I. Cincius (1949) and Johannes Benzing (1956), of which the former deals mainly with phonology, while the latter offers also a detailed discussion of derivational and inflectional morphology. A somewhat idiosyncratic reconstruction of the Proto-Tungusic phoneme system was presented by S. A. Starostin (1991: 3–39) in connection with external comparisons in an “Altaic” framework. For verbal morphology and morphosyntax, the comparative treatises of O. P. Sunik (1947, 1962, 1982) retain their value. Information on both phonology and morphology is also given in the two general surveys authored by Sunik (1968, 1997), while B. V. Boldyrev (1987) deals with nominal derivation. Compared with the synchronic publications on the individual Tungusic languages, specific papers on diachrony and reconstruction in the framework of comparative linguistics are not particularly numerous. As far as phonology is concerned, much of the discussion has circulated around the issue of vowel harmony, a topic first taken up by Joshua Ard (1981) and later thoroughly analysed by Li Bing (1996). The main issue has concerned the type of vowel harmony not only in Tungusic, but also in several neighbouring language families, including Mongolic, Koreanic, and others. Arguments in favour of a primary tongue-root harmony (TRH) in the languages of Northeast Asia, including Tungusic, have been brought forward by Joseph & Ko & Whitman (2014, 2020), but refuted by Barrere & Janhunen (2019). Other papers connected with the history of vowels include those by Doerfer (1967) on the vowels of non-initial syllables, Doerfer (1978) and Hesche (1978) on the vowel *ö, as well as by Ryzhkov-Shukumine (2020) on long vowels. In the realm of consonants, some important topics related to the diachrony of Jurchen-Manchu have been discussed by Norman (1977), as well as Vovin (1997) and Näher (1999). In the field of comparative and diachronic morphology, important papers were authored by Ikegami Jirō (1985, 1995). Particularly many contributions, mainly in a typological framework, have been made by Kazama Shinjirō, who has published, among other things, on the partitive and designative cases (1997, 2012), alienability (2008a), personal pronouns (2007a, 2008b), causative (2004) and passive (2009), clitics (2007b), as well as historical phonology (2016). He has also studied the general diachronic status of the Nanaic branch (2008), as well as, on the lexical side, “terms of direction” in Tungusic (1997b). A typologically innovative contribution with a view on the prospects of internal reconstruction and the issue concerning analytic vs. synthetic morphology is the monograph of Alonso de la Fuente (2011), accompanied by other morphological papers by him (2012, 2018). The lexical resources of the Tungusic languages, though without reconstructions, are collected in the comparative dictionary published under the editorship of V. I. Cincius (1975–1977). Otherwise, diachronic studies on the lexicon of the Tungusic languages have mainly been concerned with external comparisons with other language families, as well as with the etymology of individual lexical items, topics connected with the areal context and chronological setting of Tungusic. Lexical data have also been used as the basis for ethnographical conclusions, as exemplified by the study of Uilta lexicon by L. I. Missonova (2013). Another line of lexical studies involves the statistical analysis of vocabulary for the purposes of internal taxonomy, as in Whaley & Oskolskaya (2020). The most recent contributions to this field are Novgorodov (2021) and Novgorodov & Zaksor (2021).

Proto-Tungusic 37

VOWEL SYSTEM There is a broad consensus on that the Proto-Tungusic vowel system comprised 8 distinct segments, though a system with 7 vowels has also been proposed (as by Starostin). Opinions differ more concerning the phonetic qualities and types of oppositions of the vowels. The conventional view is that Proto-Tungusic had a system organized in terms of oppositions based on height (high vs. non-high), frontness (front vs. back), and rounding (rounded vs. unrounded). Thus, there were the two rounded back vowels *u *o [u o], the two rounded front vowels *ü *ö [y ø], the two unrounded front vowels *i *e [i e/æ], and the two unrounded back vowels, *ï *a [ɨ a/ɑ]. The vowels *u *ü *i *ï were high, while the vowels *o *ö *e *a were non-high. In the non-high category, the vowels *o *ö are often specified as phonetically mid-high, while the vowel *a is specified as low. The vowel *e (also written *ä) is variously specified as either mid-high [e] or low [æ]. A similar system has been reconstructed for the neighbouring Turkic and Mongolic languages, but synchronically it is widely attested only in Turkic, while in Mongolic and Tungusic it is present only in a few varieties spoken in the immediate neighbourhood of Turkic: Western Mongolic (Oirat) and the western dialects of Ewen. The possibility lies close at hand that the vowel system in these western varieties of Mongolic and Tungusic has been influenced, or enhanced, by Turkic (in the case of Ewen by Yakut). The rest of the Mongolic and Tungusic languages show synchronically a system where the values of the vowels are “rotated”. In the rotated system, the back vowels have lowered and in some cases “pharyngealized” values, while the front vowels are centralized or velarized. In its most perfect form such a system is observed synchronically in the eastern dialects of Ewen, but closely similar systems have been reported also from Neghidal, Orochen, Solon, and Ulcha, though these languages show, at least dialectally, a tendency to lose some distinctions. In the other Tungusic languages, the paradigm has been simplified due to mergers which have reduced the number of distinct vowels to 6, as in Nanai, Uilta, and Oroch, or 5, as in Siberian Ewenki and Manchu. It may be noted that the six-vowel systems are not identical: Of the original vowels, Nanai has lost the distinctions *u vs. *o and *ü vs. *ö, Uilta the distinctions *u vs. *ü and *ï vs. *i, and Oroch the distinctions *ü vs. *i and *ï vs. *i. A common feature for all Tungusic languages with the exception of the western dialects of Ewen is that the non-high unrounded vowel *e is pronounced with the centralized or velarized value [ə], rather than with the fronted and possibly lower values [e ɛ æ]. The vowels with the phonetic feature of “pharyngealization” are normally assumed to be pronounced with a Retracted Tongue Root position (RTR), while the vowels with no “pharyngealization” would be pronounced with a “normal” or Advanced Tongue Root position (ATR). While this seems, indeed, to be true of the prototypical case of Ewen, the phonetic oppositions in the other languages with a “rotated” vowel system are based on simple differences of height, meaning that the system has four distinct levels of height, comprising the two high vowels *ü *i = [u i], the two mid-high vowels *u *ï = [ʊ ɪ], the two mid-low vowels *ö *e = [ɵ/o ə], and the two low vowels *a *o = [a/ɑ o/ɔ]. We are, therefore, dealing with a basic dichotomy between two types of vowel oppositions: horizontal (velar vs. palatal) or vertical (higher vs. lower). The difference between these two types should not be exaggerated, for as long as the number of vowels in the para­ digm remains the same, it is a question of phonetic details, which, moreover, can be transitional.

38 Juha Janhunen

In view of the broader areal context, there is little doubt that vowel “rotation” was an innovation in Tungusic. We do not know, however, when, exactly, this innovation started its spread. Tentatively it may be assumed that, although the original vowel oppositions were of the horizontal type, the tendency to verticalize the system may have been incipient already in Proto-Tungusic (Table 4.1). Even so, the process was completed only in the Post-Proto-Tungusic period, in which connection all languages underwent a number of restructuring processes involving, most typically, the loss of some distinctions. TABLE 4.1 PROTO-TUNGUSIC VOWELS *ü

 

*i

*u

*ï *ö

*o

*e *a

Restricting the discussion here to the initial syllable, the representation of the vowels in the modern languages (Table 4.2) suggests that the original system was still preserved more or less intact in the protoforms of Ewenic and Nanaic, while in Orochic and Jurchenic it had underwent simplifications before the breakup of these branches. •

•

•

•

The vowels *a *e *o *i are preserved with no loss of distinctivity in all modern languages, with only some phonetic variation in the qualities of a = [a ɑ], e = [ə] ~ [e ɛ æ], and o = [o ɔ]. Examples: *nada/n ‘seven’ > (*)nada/n in all modern languages, *sele ‘iron’ > (*)sele in all modern languages (> Ewenki dialectal xele, Ewen xel), *bola ‘autumn’ > (*)bolo in all modern languages (> Ewen bol, Manchu bolori), *mirkü- ‘to crawl’ > Northern Tungusic (*)mirki- (> Solon milki-, Oroch mikki-, Udihe miki-) = Southern Tungusic (*)mirku- (> Nanai miku-, Ulcha micu-, Uilta mitu-, Manchu micu-). The vowel *ü is exceptional in that it is not preserved as such in any modern Tungusic language. Its reconstruction is, however, necessitated by the correspondence of Northern Tungusic *i to Southern Tungusic *u, as in *tüƞe/n ‘breast’ > Northern Tungusic *tiƞe/n = Southern Tungusic *tuƞge/n (Manchu tunggen). The vowel *u is preserved as a distinct segment, with the rotated quality ʊ, in Ewen, Orochen, and Solon, as well as, dialectally and/or in traces, in Neghidal, Oroch, and Ulcha. In Nanai, it has merged with *o > o, a development observed as a secondary tendency also in Neghidal, Oroch, and Ulcha, while in the remaining languages it is represented as the indistinct high rounded vowel u. Examples: *turkïï ‘sledge’ > Ewen tʊrkɪ, Oroch tʊkki, Ulcha tʊcɪ, Nanai tokɪ vs. Udihe tuxi, Uilta tuci; *dulï/n ‘middle, center’ > Solon dʊlɪ/n, Neghidal dolɪ/n, Ulcha dʊlɪ/n, Nanai dolɪ/n vs. Ewenki duli/n, Manchu dulin. The vowel *ï is likewise preserved as a distinct segment, with the rotated quality ɪ, in Ewen, Orochen, and Solon, as well as, dialectally and/or in traces, in Neghidal, Ulcha, and Nanai. In the other languages it has merged with *i, as in *dïlï ‘head’ > Ewenki dil, Oroch dili = Udihe dili, Manchu jili (‘the base of the horn on deer’) vs. Ewen dɪl, Orochen dɪlɪ = Solon dɪlɪ = Ulcha dɪlɪ, Nanai jɪlɪ. It may be noted that ɪ has a slightly different distribution in the modern languages than ʊ, in that the former is present in Nanai but not in Oroch, while the latter is present in Oroch but not in Nanai.

Proto-Tungusic 39

•

The vowel *ö is preserved as a distinct segment, with the rotated quality ɵ, in Ewen and Uilta. Traces of this vowel are also observed in Neghidal and Ulcha, though in these languages it has tended to merge with either o or u. In Oroch and Udihe it has merged with o, while in the remaining languages it has merged with u, as in *xölökii ‘squirrel’ > Ewen ɵliki, Uilta xɵlɵ, Neghidal ɵlukii ~ ɵlɵxii, Ulcha xolo ~ xulu vs. Oroch oloki, Udihe oloxi vs. Ewenki ulukii, Orochen uluki, Solon uluhu, Nanai xulu, Manchu ulhu. TABLE 4.2 TUNGUSIC VOWEL CORRESPONDENCES

 

*a

*e

*o

*ö

*u

*ü

*i

*ï

Ewen

a

e

o

ɵ

ʊ

i

i

ɪ

Neghidal Orochen Solon Ewenki

u

Oroch Udihe Uilta

Manchu

i

u ɵ

Ulcha Nanai

u ʊ u ʊ

u

ɪ

o u

i

Note: The synchronic vowel qualities are shown as transcribed in the present volume. The shaded areas indicate Post-Proto-Tungusic paradigmatic mergers.

From the point of view of vowel rotation it is important to note that the eight-vowel systems synchronically observed in several Northern Tungusic languages, that is, Ewen, Orochen, and Solon, as well as, marginally, in Neghidal, are not direct developments of the original Proto-Tungusic system. The merger of *ü [y] with *i [i] in Northern Tungusic reduced the set of rounded vowels to the reflexes of *u *o and *ö. As the values of *u *o were lowered and possibly pharyngealized (as in Ewen), *ö remained their only counterpart on the non-lowered (and non-pharyngealized) side. Therefore, to make the system complete, a new distinction had to be created between the synchronic values u [u] and ɵ [ɵ]. In Orochen and Solon, where *ö seems to be regularly represented as u, a new ɵ was introduced mainly in innovative lexicon, borrowings, and by sporadic developments. In Ewen, which (like Uilta) preserves *ö > ɵ as a distinct segment, a new u was created by splitting the reflexes of *ö between the values [u] and [ɵ]. In spite of considerable work done on the background of this split (Doerfer and Hesche), the details remain unclear, though it is obvious that both contextual and dialectal factors must have been involved. One contextual factor that seems to have favoured the development *ö > u is the presence of a following syllable-final consonant (either primary or secondary), as in *ölse ‘meat’ > Ewen ulre = Ewenki (*)ulre, Neghidal ulle, Orochen ule, Solon ulde, Oroch ukte, Ulcha ulse. In addition to the basic short vowels, Proto-Tungusic also had the corresponding long vowels. It is a matter of interpretation whether these should be understood as

40 Juha Janhunen

monosegmental long monophthongs, or as biphonemic sequences of two identical segments, that is, as “double vowels” (V1V1). The qualitative development of these long vowel elements follows the patterns observed for the short vowels. The feature of vowel length is, however, rather unstable in all Tungusic languages, and a word attested with a long vowel in one language or dialect may appear with a short vowel in another. Even so, cases of inherited vowel length are present in all major branches with the exception of Jurchenic. Among the individual languages, irregular shortening is particularly frequently observed in Udihe and Kilen, cf. e.g. *ƞaala ‘hand’ > (Ewenic:) Ewen ƞaal, Ewenki ƞaala, Neghidal ƞaala ~ ƞala, Solon naala, (Orochic:) Oroch ƞaala, (Nanaic:) Uilta ƞaala = Ulcha ƞaala = Nanai ƞaala vs. Udihe ƞala = Kilen ƞala, Manchu gala; *ƞeele- ‘to be afraid’ > (Ewenic:) Ewen ƞeel-, Ewenki ƞeele- = Neghidal ƞeele-, Solon neele-, (Orochic:) Oroch ƞeele-, (Nanaic:) Uilta ƞeele- = Ulcha ƞeele- = Nanai ƞeele- vs. Udihe ƞele-, Manchu gele-. The background of the long or double vowels is unknown. It would be tempting to assume that they derive by contraction from sequences of two vowels originally separated by a consonant, perhaps *x (*VxV), but there is no evidence confirming this assumption. They also do not seem to be conditioned by any other contextual, such as prosodic, factors. Their analysis as sequences of two segments, as preferred here, is, however, supported by the fact that there were also “diphthongoid” sequences of two non-identical vowels (V1V2). The system of these sequences is still open to alternative interpretations, but two uncontroversially established “diphthongoids” are *ïa and *ie. Of these, *ie is reliably attested only in a single item, yielding variously the diphthongoid ie [je:] or the monophthong ii [i:] in the modern languages: *niekii ‘duck’ > Ewen nieki, Ewenki niikii, Solon niixii, Nanai nie-cien ~ nii-cien, Manchu niyehe. By contrast, *ïa is frequent and yields the synchronic diphthongoids ɪa [jæ:] or ie [je:], e.g. *mïaba/n ‘heart’ > Ewen mɪawa/n [mjæ:wən], Ewenki miewa/n [mje:wan] = Neghidal miewa/n, Solon miega/n, Oroch mɪawa/n, Udihe määwa/n, Uilta miewa/n, Nanai mɪawa/n, Manchu niyaman (< *miyaman). In some languages, notably Solon and Uilta, the (harmonically neutral) diphthongoid ie has triggered the introduction of an analogous secondary short vowel, iĕ = é [e], which, however, remains synchronically rare and marginal. The diphthongoids *ïa and *ie could possibly also be reconstructed as *ïya and *iye, that is, as containing an intervocalic palatal glide (note that the glide in Manchu examples like niyehe involves a purely orthographic convention). This assumption is not contradicted by the presence of metathesis in the items *xï(y)a(-) ~ *xa(y)ï(-) ‘what?’ > Ewenic & Orochic *ïa- = Manchu ya (ya) vs. Nanaic *xaï-, and *sï(y)a- ~ *sa(y)ï- ‘to bite’ > Ewenic & Orochic & Nanaic *sïa- vs. Jurchenic *sai-, though in these cases the original sequence might also have been *a(y)ï, rather than *ï(y)a. There are also analogous sequences with the labial vowel *u, which may have been combined with an intervocalic labial glide *w, as in *ju(w)a/n ‘ten’ > Nanaic & Jurchenic *jua/n vs. Ewenic & Orochic *jaa/n. An exceptional correspondence is present in Nanaic *au-‘to sleep’ vs. Ewenic & Orochic *aa- = Jurchenic (Manchu) a-. It is particularly difficult to verify in the etymological material the sequences *ü(w)e ~ *ö(w)e and *ü(y)i, which have occasionally been postulated for Proto-Tungusic, for in most cases, at least, it is sufficient to reconstruct the long vowels *öö and *üü, respectively, although some of their reflexes in the modern languages involve diphthongoids, as in *jöör ‘two’ > Ewenic & Orochic *juur vs. Nanaic & Jurchenic *juer, *müüki ‘snake’ > Ewenic & Orochic *miiki vs. Nanaic & Jurchenic *muiki (> Manchu meihe). A unique correspondence set is involved by the items for ‘downstream’: Northern Tungusic *ƞïa- (> Ewenki ƞie-) vs. Nanaic waï- ~ ƞoï- vs. Manchu

Proto-Tungusic 41

wa- (in wala = LOC *wa-la ‘underneath, west side’), which, supposing that they go back to a single protoform, would suggest the presence of a labial element either before or inside the sequence *ïa, i.e. *ƞwïa- ~ *ƞïwa-, from which the Southern Tungusic data could be derived by assuming a metathesis and assimilative rounding of the vowel *a, i.e. > *ƞwaï- > *ƞoï- (= Nanaic ƞoï-) ~ *waï- (= Nanaic waï- > Manchu wa-). VOWEL HARMONY All Proto-Tungusic vowel qualities were present also in non-initial syllables, but in this position their distribution was governed by vowel harmony. The members of the vowel paradigm were divided into two classes, traditionally known as the “male” (陽 yang) vs. “female” (陰 yin) vowels, sometimes also identified as “hard” vs. “soft” vowels, respectively. The male vowels were *a *ï *o *u, corrresponding to the original back (or lower) vowels, while the female vowels were *e *i *ö *ü, corresponding to the original front (or higher) vowels. The corresponding long vowels, as well as the diphthongoids *ïa and *ie, were divided accordingly (Table 4.3). TABLE 4.3 PROTO-TUNGUSIC VOWEL HARMONY “male”

*u *uu

*o *oo

*a *aa

*ï *ï ï

*ïa

“female”

*ü *üü

*ö *öö

*e *ee

*i *ii

*ie

The harmonic class of a root was determined by the vowel of the initial syllable, from which the harmony proceeded progressively to the following syllables of both the root and any suffixes and clitics. With some variation in the details, this system is still present in all the modern Tungusic languages. Due to the neutralizing effect of vowel harmony, non-initial syllables could have only four distinctive vowel qualities, which may conveniently be analysed as archiphonemic *A *I *O *U. Vowel length was distinctive also in the non-initial syllables, but compared with the initial syllable length was even more unstable in this position, except in suffixes, where long vowels are somewhat more systematically preserved, especially in the Ewenic branch. The sequence *A(y)I in non-initial syllables had merged with the long vowel *II already in Proto-Tungusic, but it can be reconstructed in several suffixes in which the final element -(y)I was a separate suffixal morpheme, as in the imperfective participle in *-rII < *-rA-(y)I, based on the aorist participle in *-rA. While the fact that Proto-Tungusic had vowel harmony is generally recognized, opinions differ as to what type of harmony the protolanguage had. The issue is connected with the structure of the reconstructed vowel system. Those who favour the assumption that the vowels were organized along the palatal-velar axis assume that the language had a horizontal palatal-velar harmony (PVH), while those who operate with tongue-root oppositions assume that the language had a vertical tongue-root harmony (TRH). There is no definitive answer to this dispute, especially as both parameters may have been involved, and the issue is of little relevance as long as the size of the paradigm remained unchanged. However, there are both systemic and areal facts that favour the conclusion that the original system was of the palatal-velar type, although it may have been undergoing rotation already at the level of the protolanguage. It is particularly relevant to notice that the vowel *ü has a dual representation in the modern languages, in that it has merged with *u in

42 Juha Janhunen

Southern Tungusic against *i in Northern Tungusic. Both mergers are natural to explain if we assume that the original phonetic value of the vowel was [y], and not [u]. In this connection it may be noted that there was also some restructuring of the vowel qualities of non-initial syllables, in that the vowel *U merged with the reflexes of *O in Southern Tungusic and with the reflexes of *I in Northern Tungusic. The total picture may be summarized as follows (Table 4.4): TABLE 4.4 VOWELS IN NON-INITIAL SYLLABLES  

*A

*O

*U

*I

Ewen

A

U

I

I

Neghidal Orochen Solon Ewenki Oroch Udihe Uilta

U

Ulcha Nanai Manchu *A = *a *e, *O = *o *ö, *U = *u *ü, *I = *ï i. The shaded areas indicate Post-Proto-Tungusic paradigmatic mergers.

•

•

•

•

The vowel *A = *a *e is retained in all branches, as in *baka- ‘to find’ > Ewenic *baka- (> Solon baxa-), Orochic *baka- > *ba’a- (> Oroch baa-), Nanaic *baka(> Nanai baa-), Jurchenic *baka- (> Manchu baha-); *xöre ‘mountain, forest’ > Ewenic *ure (> Neghidal uye), Orochic *u(r)e (> Oroch uwe, Udihe we), Nanaic *xöre/n (> Nanai xure/n), Jurchenic *ure (> Manchu wehe = we-he ‘stone’). The vowel *O = *o *ö is represented as a rounded *U (U/O) in all branches, as in *japko/n ‘eight’ > Ewenic *japku/n (= Ewenki japku/n), Orochic *jakpu/n (> Oroch jappu/n), Nanaic *jakpu/n (> Nanai jakpo/n), Jurchenic *jakpu/n (> Manchu jakūn); *pemö/n ‘lip(s)’ > Ewenic *xemu/n (= Ewenki xemu/n), Orochic *xemu/n (= Oroch xemu/n), Nanaic *pemu/n (= Nanai pemu/n), Jurchenic *pemu/n (> Manchu femen). The vowel *U = *u *ü is retained as a labial *U (U/O) in Southern Tungusic but is represented as *I in Northern Tungusic, as in *adulï ‘net’ > Ewenic *adïlï (> Ewenki adil), Orochic *adïlï (> Udihe adili), Nanaic *adulï (> Nanai adolɪ → Oroch adʊli); *xedü/n ‘wind’ > Ewenic *edi/n (= Ewenki edi/n), Orochic *edi/n (= Oroch edi/n), Nanaic *xedu/n (= Nanai xedu/n), Jurchenic *edu/n (= Manchu edun). The vowel *I = *ï *i is retained in all branches, as in *jabï ‘boat’ > Ewenic *jabï (> Ewenki jaw), Orochic *jabï (> Oroch jawi), Nanaic *jabï (> Nanai jayɪ → Manchu jaya); *beri ‘bow’ > Ewenic *beri (> Ewenki ber), Orochic *beri > *beyi (= Oroch beyi), Nanaic *beri > *böri > *buri (= Nanai buri), Jurchenic *beri (~ *böri > *buri > *beri) (= Manchu beri).

Proto-Tungusic 43

It may be noted that those modern languages that have lost the paradigmatic distinction between the vowels *ï vs. *i (or *ü) and/or *u vs. *ü (or *ö), nevertheless tend to retain this distinction at the lexical (deep) level, in that items containing only these vowels are still synchronically divided into two harmonic classes, as in Ewenki dil ‘head’ (< *dïlï) : ACC dil-wa vs. il ‘sinew’ (< *xüli) : ACC il-we, juu ‘dwelling’ (< *juug) : ACC juu-wa vs. juur ‘two’ (< *jöör) : ACC juur-we. Apart from the basic dichotomy between the “male” and “female” vowels, all Tungusic languages show also traces of what seems to have been a phonetic tendency to a labial harmony in Proto-Tungusic. This tendency was triggered only by the short (single) “male” vowel *o in the combination *o-a, in which the low vowel *a is represented as [o], as in *ora/n ‘(domestic) reindeer’ > *oro/n > Ewenic *oro/n (= Ewenki oro/n > Neghidal oyo/n), Orochic *oro/n (> Udihe oro/n), Nanaic *oro/n (= Nanai oro/n), Jurchenic *oro/n (= Manchu oron). The phonological status of this feature varies, however, and in some languages (as in Ewenki) it could synchronically still be analysed as phonologically irrelevant, while in other languages (as in Nanai) it has become distinctive because of other changes in the vowel system (such as the development *u > o in the combination *u-a, which, then, yields synchronically o-a without labial harmony). CONSONANT SYSTEM The consonant paradigm of Proto-Tungusic can be relatively unambiguously reconstructed as having distinguished between four places of articulation: the labials *m *b *p *w, the dentals *n *d *t *s *l *r, the palatals *ñ *j *c *y, and the velars *ƞ *g *k *x, as well as seven manners of articulation: the nasals *m *n *ñ *ƞ, the weak stops *b *d *j *g, the strong stops *p *t *c *k, the fricatives *s *x, the glides *w *y, and two liquids, the lateral *l, and the vibrant (trill) *r (Table 4.5). The labials, including the glide *w, were apparently realized as bilabials, the dentals mainly as alveolars, while the velars may have had both front velar and back velar (post-velar or uvular) allophones. The dental fricative *s was realized as a sibilant, while the velar fricative *x may have had both velar and laryngeal realizations. The phonetic distinction between the weak and strong stops may have involved voice (voiced vs. voiceless), aspiration (unaspirated vs. aspirated), or both. The palatal stops *c *j probably involved affrication with a sibilant release, which may have varied between the palato-alveolar and alveolo-palatal positions. All these features, and the segmental paradigm as a whole, are preserved with surprisingly little variation in the modern Tungusic languages. TABLE 4.5 PROTO-TUNGUSIC CONSONANTS *m

*n

*ñ

*ƞ

*b

*d

*j

*g

*t

*c

*k

*p

*s *w

*x *y

*l *r

44 Juha Janhunen

The principal changes that have occurred are the loss of the original velar fricative *x and the positionally conditioned origination of a new velar fricative with very similar phonetic characteristics from former *p *k and/or *s in Northern Tungusic and Jurchenic. In so far as these changes led to the loss of *p, a new segment of the same type has been introduced in secondary vocabulary (loanwords and descriptive items). Other new segments, filling the gaps in the system in some modern languages, include the labial (phonetically dentilabial) fricative f [f] in Jurchenic and Bikin Nanai, as well as the palatal sibilant š [ʃ ɕ] in Jurchenic. Also, a separate set of post-velar consonants, especially the stops [q ɢ] and the fricative [χ], but also the nasal [ɴ], have been formed Jurchenic and Amur Tungusic, though they seem to have reached a phonemic status only marginally, and only in Jurchenic. Positional changes have also affected the phonotactic properties, but not the paradigmatic status, of several other consonants, including, in particular, the palatal and velar nasals *ñ *ƞ, the palatal stops *c *j, the glides *w *y, and the liquids *l *r. The diagnostic differences between the four branches with regard to the segments *x *p in initial position may be illustrated as follows: •

•

The original (primary) velar (laryngeal) fricative *x is preserved only in Nanaic, while in Ewenic, Orochic, and Jurchenic it has been lost without a trace, as in *xörge ‘heavy’ > Ewenic *urge (= Ewenki urge), Orochic *urge > *ugge (= Oroch ugge), Nanaic *xurge (> Nanai xuyge > Ulcha xuje- > Uilta xude), Jurchenic *urge (> Manchu ujen = uje/n). However, the loss development in Jurchenic must have taken place separately from the similar development in Northern Tungusic. In the case of Jurchenic, the loss may have been stimulated by the areal influence of Mongolic. The original strong labial stop *p is preserved as such only in Nanaic (in Bikin Nanai > f), while in Ewenic and Orochic it is represented as a secondary *x and in Jurchenic (Manchu) as f, as in *pesi/n ‘shaft’ > Ewenic *xesi/n (= Ewenki xesi/n > esi/n), Orochic *xesi/n (= Oroch xesi/n), Nanaic *pesi/n (= Nanai pesi/n), Jurchenic *pesi/n > *fesi/n (= Manchu fesin). As may be seen, the development *p > *x took place in the Northern Tungusic protolanguage of Ewenic and Orochic, while the development *p > f in Jurchenic represents a rather recent innovation (spread on an areal basis to Bikin Nanai). Occasionally, the regular picture is confused by interbranch borrowings, as in *pokta ‘textile, dress’, which yields regularly Ewenki xokto (> okto) and Nanai-Ulcha-Uilta pokto, but also irregularly Oroch-Udihe pokto (borrowed from Nanai) and Uilta xokto (borrowed from Ewenki).

MORPHEME STRUCTURE Underived roots in Proto-Tungusic were either monosyllabic of the type (C)V(V)(C) or bisyllabic of the type (C)V(V)(C)CV(V), meaning that consonant clusters (CC) could be present only in medial intervocalic position, while long vowels (VV), though not always easy to verify, were permitted in all syllables. A vocalic anlaut (#V) could also be assumed to have been preceded by a zero consonant, (#ØV) which could technically be analysed as an additional member of the consonant paradigm (an approach not adopted here). Longer structures, either bisyllabic with a final consonant or polysyllabic with additional syllables, were diachronically, though not always synchronically, derivatives. Grammatically, both basic roots and derived stems were divided into two main inflectable classes, nominals and verb(al)s, with an additional, but much smaller, class,

Proto-Tungusic 45

being formed by non-inflectable lexemes or “invariables”. There was no primary structural difference between these classes, but it seems that only nominal stems could occur as free morphemes, while verbal stems were always accompanied by suffixes, as is still the case in most of the modern Tungusic languages. A special group of nominal stems was formed by those ending in the so-called unstable n (/n), which had inflectional properties of its own and which could be absent before derivational suffixes. Monosyllabic roots of the type (C)V, with a short (single) vowel, functioned typically as auxiliaries, as in the personal pronouns, e.g. 1SG *bi : 2SG *si, the copula root *bi-, and the negation verb *e-, but also in the non-auxiliary basic verbs *bö- ‘to die’, *di- ‘to come’, and *ga- ‘to take’. It is possible that the personal pronouns belonging to this type, when used in isolation, were pronounced with a long (double) vowel, as is normally the case in the modern languages, though lexically they may still have contained a short vowel. Otherwise, the type (C)VV, with a long vowel, was present in both nominal and verbal roots, e.g. *paa ‘liver’ (= Nanai paa, elsewhere *paa-kun), *saa- ‘to know’, *moo ‘tree’, *oo- ‘to make’, *möö ‘water’, *ñöö- ‘to exit’ (> Ewenki yuu-), *ii- ‘to enter’, *ƞüü ‘who?’. This type also comprises items with the diphthongoid *ïa, e.g. *xïa ‘what?’ : *xïa- ‘to do what?’, *gïa ‘other, comrade’. Monosyllabic consonant roots of the types (C)VC and (C)VVC are conspicuously rare in the comparative database, suggesting that they represent an archaic structure that was disappearing already in Proto-Tungusic. Only a few final consonants are attested in these roots, notably *g *p *l *r *n, as in *juug ‘dwelling’, *jep- ‘to eat’, *maal- ‘to die out’ (only in Nanaic), *daar ‘fathom’, *sïa/n ‘ear’, *göön- (~ *ƞöön-)‘to say’. With the monosyllabic roots constituting only a small minority, the bulk of all nominal and verbal lexemes were bisyllabic, though in some of them internal reconstruction allows to identify monosyllabic primary roots, as in *xolsa ‘fish’ = *xol-sa, *ölse ‘meat’ = *öl-se (with the class marker *-sA). The suffixal elements that could be added to the bisyllabic lexemes could be composed of a single consonant (-C), an open syllable (-CV), or a combination of both (-CCV). The consonants that were permitted in the role of the first component of a medial cluster were restricted to only two archiphonemic stops, *B *G, the three nasals *m *n *ƞ, and the two liquids *l *r, while the position of the second component was normally occupied by any of the stop obstruents, or also the sibilant *s. Altogether, some 30+ consonant clusters can be reconstructed (Table 4.6).

TABLE 4.6 PROTO-TUNGUSIC CONSONANT CLUSTERS  

*p

*b

*t

*d

*B

*pt

*bd

*G

*kt

*gd

*m

*mt

*md

*n

*mb

*nt

*nd

*ƞ

*ƞt

*ƞd

*lt

*ld

*r

*rp

*rb

*l

*lp

*lb

*c

*j

*kc

*gj

*k

*g

*s *ps *ks

*mk

*mg

*ms *ns

*ƞk

*ƞs

*rc

*rk

*rg

*lc

*lk

*lg

*ls

Horizontal lines: first component, vertical columns: second component. Neutralized segments: *B = *b *p, *G = *g *k.

46 Juha Janhunen

There are several systematic gaps in this system. For instance, no geminates (C1C1) can be reconstructed for Proto-Tungusic. Also, the two archiphonemic obstruents *B *G occurring in the role of the first component of clusters seem to have been represented phonetically as either weak [b g] or strong [p t] depending on the quality of the following segment, which was also an obstruent. Since there was no distinction between the weak and strong segments in this position, they could in all cases be identified as *b *g, assuming that these were the less marked members of the pairs. However, the tradition of Tungusic studies favours in these cases the phonetic approach, for which reason the clusters *pt *ps *kt *kc *ks are here written with the letters

, while the clusters *bd *gd *gj are written with the letters . Somewhat unfortunately, this convention is different from the one that prevails in Mongolic studies, where analogous clusters are traditionally always analysed as containing the weak segments *b *g (as well as *d) as the first component. Some of the reconstructed clusters seem to have been very marginal, being attested in only a few items. This is, for instance, the case with *gj *kc, attested only in *sugjansa ‘fish’ > ‘salmon’ and *gökcee/n ‘pike’, both with some irregularities in the correspondences. There are also a few clusters that could possibly be added to the list. It may be noted that all of the above-listed clusters contain an obstruent as one or both of the two components. There are, however, etymons which might originally have contained a cluster of two sonorants, a potential example being offered by the items for ‘tongue’: Ewenic *ilƞi (= Ewenki ilƞi ~ inƞi ~ iƞi ~ inni), Orochic *iƞƞi (= Oroch iƞƞi ~ iƞi), Nanaic *xilƞu > *silƞu (> Nanai siƞmu ~ sirmu, Ulcha siñu), Jurchenic *ileƞu (> Manchu ilenggu), all of which could be traced back to the reconstruction *xilƞü, unless the word was originally trisyllabic *xileƞü. Alternative reconstructions can also be proposed for the items for ‘mouth’: Ewenic *amƞa (= Ewenki amƞa ~ amma), Orochic *aƞma (= Udihe aƞma > Oroch amma), Nanaic *aƞma (= Nanai aƞma ~ amga), Jurchenic *aƞ(g)a (> Manchu angga), which could back to *amƞa (with a regular metathesis in Orochic and Nanaic), but also to *amga. In general, weak stops following a nasal show a dichotomy in their development, in that they are preserved as stops in Southern Tungusic, but nasalised in Northern Tungusic, as in *kamdo/n ‘glue’ > Ewenic & Orochic *kamnu/n (= Ewenki & Oroch kamnun), Nanaic *kamdu/n (= Nanai kamdo/n), Jurchenic (irregular) *amdu/n (= Manchu amdun). In the case of the homorganic cluster *nd the nasalization led to a geminate, which was subsequently simplified to the single segment *n in Northern Tungusic, as in *ƞïnda : *ƞïnda-kun ‘dog’ > Ewenic *ƞïna-kï/n (> Ewenki ƞinakin > inakin ~ ninakin), Orochic *ïnakï/n (> Oroch inaki/n), Nanaic *ƞïnda (> Nanai ɪnda), Jurchenic *ïndakun (> Manchu indahūn). There seem to be no roots with the labial cluster *mb, but the analogous nasalization development in this cluster is observed at morpheme noundaries, as in the personal marker 1SG *-n+bi > *-m-bI > Ewenic *-mI > -m vs. Nanaic *-mbI, e.g. *ƞene- ‘to go’ : Ewenki 1SG ƞene-m (< *ƞene-mbi) vs. Nanai ene-e-mbi (< *ƞene-re-mbi). This means that the distinction between the clusters *nd *mb against the medial nasals *n *m was lost in Northern Tungusic. It could be assumed that a similar loss of distinction would have taken place between the velar cluster *ƞg and the medial nasal *ƞ, but the comparative database shows no evidence of this distinction at the level of Proto-Tungusic. Therefore, the correspondence of Northern Tungusic *ƞ to Southern Tungusic *ƞg is conventionally derived from *ƞ, assuming that Southern Tungusic underwent a clusterization development *ƞ > *ƞg, as in *soƞa- ‘to cry’ > Ewenic *soƞo- (= Ewenki soƞo- > Ewen xoƞ-), Orochic *soƞo- (= Oroch soƞo-), Nanaic *soƞgo- (= Nanai soƞgo-), Jurchenic *soƞgo- (= Manchu

Proto-Tungusic 47

songgo-). This may have been an areal innovation, since a similar clusterization development is known from Mongolic. Apart from the rules governing the combination of consonants to clusters, there were several other phonotactic restrictions affecting the distribution of the individual consonant phonemes. For instance, the vibrant *r did not appear in initial position, while the fricative *x did not occur medially or finally, or in clusters. The glides *w *y had a limited distribution, with the labial glide *w being distinctive against the weak labial stop *b only in initial position, notably in the item *waa- ‘to kill’, while the palatal glide *y is reliably attested only in medial intervocalic position, as in *pöye ‘wound’. The palatal nasal *ñ, which is well attested in initial position, as in *ñaa- ‘to rot’, is difficult to verify in other positions except in the idiosyncratic clusters *ƞñ ñƞ, as in *ñeƞñe ‘spring(time)’ (attested as such in all branches, e.g. Ewenki ñeƞñe, Manchu niyengniyeri), *añƞa (?) ‘year’ (with several irregularities in the correspondences, e.g. Ewenki añƞa-nii, Manchu aniya). Except for vowel length, no suprasegmental distinctions can be reconstructed for Proto-Tungusic. The progressive nature of vowel harmony supports the assumption that there was a non-distinctive initial stress. Initial stress is also suggested by the fact that several Tungusic languages have undergone processes of final vowel loss, either universally (loss of all short vowels in final position), as in Ewen, or selectively (loss of certain vowel qualities), as in Ewenki (which has lost the high vowels *ï *i < *I *U). WORD FORMATION In the suffixal string following underived word roots, the first slot was taken by derivational suffixes, followed by inflectional markers, as well as clitics. All suffixal elements were affected by vowel harmony, which, in addition to the non-distinctive initial stress, contributed to the coherence of the phonetic word. Most suffixes, both derivational and inflectional, were differentiated according to the part of speech. In the class of nominals, some derivational suffixes were specific to subclasses with adjectival, numeral, or pronominal semantics, although inflectionally all of these subclasses followed the pattern of nouns. The derivational suffixes of both nominals and verbs had both lexical and grammatical functions, and could also operate between the parts of speech, converting nominals to verbs and vice versa. Most Tungusic languages also have a limited number of examples of zero derivation, or nomina-verba, some of which can be traced back to Proto-Tungusic, e.g. *tügde ‘rain; to rain’ (attested in all branches except Jurchenic). From the point of view of their effect on the part of speech of the underlying root, the derivational suffixes can be divided into those producing denominal nouns, deverbal nouns, denominal verbs, and deverbal verbs. (1) Denominal nouns involve a number of lexical and/or grammatical categories which retain their productivity in all, or most of the modern Tungusic languages, as well as a couple of diachronically and typologically important groups of forms which had lost their productivity already before the Proto-Tungusic level, but which remain sufficiently transparent to be recoverable by the methods of internal reconstruction. •

Most importantly, Proto-Tungusic had traces of two class markers, *-sA, which referred to homogeneous masses (uncountables), including liquids, and *-tA, which referred to individualizable objects (countables). Both elements are often combined with a preceding velar stop segment *-k- (= *-G-), which

48 Juha Janhunen

•

•

•

apparently represented a separate element, as in *see-k-se ‘blood’ (> Ewenki seekse), *sile-k-se ‘dew’ (> Ewenki silekse), *xüü-k-te ‘tooth’ (> Ewenki iikte), *xoosï-k-ta ‘star’ (> Ewenki oosikta), but *-sA, in particular, is also attested in combination with other consonants, some of which may simply have been original root-final segments, while others may have had an obscured derivational function, as in *ö.l-se ‘meat’ (> Ewenki ulle), *xïma.n-sa ‘snow’ (> Ewenki imanna), *gïra.m-sa ‘bone’ (> Ewenki giramna). As a specific semantic development, the combination *-k-sA, when attached to names of animals, has received the meaning ‘skin, fur’ or ‘meat’ (of the animal concerned), as in Ewenki sulakii ‘fox’ : sulakii-ksa ‘fox skin; fox meat’, Nanai mapa ‘bear’ : mapa-ksa ‘bear skin’. There are several unexplained details concerning the structure and function of the Proto-Tungusic class markers. Among other things, from the point of view of the internal taxonomy of the Tungusic languages it is important to note that the Jurchenic counterpart of the suffixal combinations *-k-sA, *-n.sA-, *-m.sA is *-ƞgi. In this respect, Jurchenic differs not only from Northern Tungusic, but also from Nanaic, as in Manchu senggi ‘blood’ vs. Nanai sekse, Manchu giranggi ‘bone’ vs. Ulcha gɪramsa. Of grammatical importance is the marker of alienable possession, reconstructable as *-ƞU-, as in Nanai PX1SG jɪlɪ-ɪ ‘my (own) head’ (< *dïlï-bï) vs. AL-PX1SG jɪlɪ-ƞgo-ɪ ‘the head (e.g. of an animal) that belongs to me’ (< *dïlïƞu-bï). This could also be classified as an inflectional suffix, since it normally occupies a position between (other) derivational suffixes (proper) and a following case suffix. It can, however, also precede the proprietive suffix, as in Nanai PROPR jɪlɪ-ko ‘having (one’s own) head’ : AL-PROPR jɪlɪ-ƞgo-ko ‘having (an alien) head’. The marker of alienability is synchronically attested in this function in all branches except Jurchenic. It is, however, possible that a cognate of this marker is also present in Jurchenic, where it can be identified with the Manchu element -ƞge (Written Manchu -ngge, Spoken Manchu and Sibe -ƞe), which functions as a “nominalizer”, transforming participles to “infinitives”, that is, independent head nouns, as in Sibe yawe- ‘to go’ : PTCP.AOR yawe-re ‘going, one who goes’ : INFIN yawe-re-ƞe ‘the act of going’. Another form probably connected with the marker of alienable possession is the possessive marker *-ƞII < ? *-ƞU-(y)I, which basically marks the possessor form of nominals (‘of’, ‘belonging to’), as in Ewenki beye ‘person’ : POSS beye-ƞii ‘of the person, belonging to the person’. When used as a nominal predicate or as an independent head noun, this form can take additional markers for number and/or case, but in the adnominal position it can function very much like a genitival case form. It is also present in possessive pronouns of the type Ewenki 1SG bii : POSS min-ƞii ‘my, mine’. Such usages have led to its transformation to a true nominal case marker in several Ewenic varieties spoken in the Manchurian sphere, notably Solon, Orochen, and Khamnigan Ewenki. It is possible that the Manchu genitive marker -i (Written Manchu -i ~ -ni) also diachronically represents this same element, although the phonetic correspondence (loss of medial *-ƞ-) is not regular. A different aspect of possession, relating to the entity that is possessed, is expressed by the proprietive forms, e.g. *asïï woman, wife’ > Ewenki asii : PROPR asii-cii ‘having a wife, married’ vs. Nanai asɪ : PROPR asɪ-ko id. The suffixes marking this function show an exceptional and taxonomically

Proto-Tungusic 49

•

interesting set of segmental correspondences, for which reason it is not clear whether they have a single origin or not: Ewenki-Neghidal -cII ~ -lkAAn, Orochen -cI ( *-c- in Ulcha, and *-lk- > -l- in Uilta cannot be considered regular. The Orochic forms go back to *-kI, which, in view of the Nanaic data, could also reflect *-lkU, with the regular developments *U > I in non-initial syllables (in Northern Tungusic) and *-lk- > -k- (in Orochic, later > -x- in Udihe). The consonantism of *-lkU is compatible with Ewenic *-lkAAn, which means that Ewenic *-cII must be a different element, though it may be contained in Ewen -pcI ~ -pcUn (for plural proprietives). In view of all this, the possibility cannot be ruled out that Proto-Tungusic had a suffix with the cluster *-lk- (plus a vowel) for marking the proprietive function. However, this suffix does not seem to have a cognate in Jurchenic, where the proprietive function is expressed by the suffix -ƞgA, e.g. Manchu gerbu ‘name’ : PROPR gerbungge (gerbu-ƞge) ‘having a name, famous’, cf. Nanai gerbu-ku vs. Ewenki gerbii-cii id. Most other types of denominal nouns in Proto-Tungusic involved a higher degree of lexicalization with no grammatical functions. Among the more widely attested elements are the diminutive suffixes *-cAA/n (in Ewenic, Orochic, Nanaic) and *-kAA/n (in all branches), *bïra ‘river’ : DIM *bïra-kaa/n (> Ewenki bira-kaa/n), *puta ‘bag’ : DIM *puta-caa/n (> Nanai puta-ca/n). Close physical association was expressed by the suffix *-ptO/n, as in *sïa/n ‘ear’ : *sïa-pto/n ‘earflaps’ (> Ewenki sie-ptu/n, Nanai sɪa-pto/n). Several suffixes derived nominals with adjectival meanings, which, again, could also be used as nouns, an example being *-mA for materials, as in *sele ‘iron’ : *sele-me ‘(made of) iron; sword’ (= sele-me, in all branches). Spatials, numerals, and pronouns also had a number of specific derivational patterns, many of which survive in several modern languages.

(2) Deverbal nouns may be divided into two types: lexical and grammatical. The former are full nouns with nominal morphology and nominal syntax, while the latter combine nominal morphology with verbal syntax. Strictly speaking, both types involve nominalization, that is, the representation of verbal roots as nominals, but in a more narrow sense only the latter type, that is, “participles”, as understood in the Tungusic context, correspond to the definition of grammatical nominalization. •

Suffixes producing lexical nominalizations in Proto-Tungusic included, in particular, those for locative and instrumental nouns (nomina loci et instrumenti). Some of these suffixes seem to have yielded derivatives that were lexicalized already in the protolanguage, cf. e.g. *-(ƞ)kUU, as in *tege- ‘to sit’ : *tegeƞküü ‘place to sit; chair’ (> Ewenki tege-ƞkii, Nanai tee-ƞku, Manchu teku); *-pOO/n, as in *pupu- > *puu- ‘to saw’ : *puu-poo/n ‘saw’ (> Ewenki xuuwuu/n, Nanai poo-po/n, Manchu fufun). Another category, still productive in the modern languages, is formed by the actor nouns (nomina actoris), for which the Proto-Tungusic suffix *-mdII can be reconstructed, yielding Ewenic *-mnII, Orochic *-mdi (in view of the consonantism, probably borrowed from Nanaic), Nanaic -mdI > Nanai -mjɪ-, and Jurchenic -mji > Manchu -msi, as in Ewenki elge- ‘to lead’ : elge-mnii ‘guide’, Oroch saa- ‘to know’ : saa-mdi

50 Juha Janhunen

•

‘knowledgeable person’, Nanai aloosɪ- ‘to teach’ : aloosɪ-mjɪ ‘teacher’, Manchu taci- ‘to study’ : taci-msi ‘student’. Participles are distinguished from other deverbal nouns not only by their verbal syntax, but also by their connection with temporal and/or aspectual distinctions, for which reason they are best discussed in the context of verbal inflection.

(3) Denominal verbs are perhaps also best divided into two types, which may be identified as general vs. specific verbalizations. Both groups, but especially the latter, involve lexicalized examples, although several suffixes remain productive in the modern languages. • •

The two commonly attested general verbalizers are *-lAA- and *-dAA-, as in *möö ‘water’ : *möö-lee- ‘to fetch water’ (> Ewenki muu-lee-, Nanai mue-le-), *jola ‘stone’ : *jola-daa- ‘to throw’ (> Ewen jol-da-, Uilta jolo-do-). Specific verbalizers with a technically verified Proto-Tungusic background include *-cI- for instrumentals (‘to use as something’), as in *okta ‘medicinal herb’ : *okta-cï- ‘to cure with medicine’ (> Udihe okto-si-, Ulcha okto-cɪ-); *-mAA- for captatives (‘to catch something’), as in *xölö- ‘squirrel’ : *xölömöö- ‘to (go) hunt(ing) for squirrels’ (> Ewenki ulu-mee-, Uilta *xɵlɵ-mɵ- > xolo-mo- ~ xulu-me-); *-mO(.sI-) for odoratives (‘to smell of something’), as in *ölse ‘meat’ : *ölse-mö.si- ‘to smell of meat’ (> Ewenki ulle-mu-, Uilta ulise-mu.si-). Several other types, potentially also deriving from ProtoTungusic, can be reconstructed at the levels of Northern and Southern Tungusic.

(4) Deverbal verbs in Tungusic are a heterogeneous group of derivational forms which, like deverbal nouns, can express both lexical and grammatical functions. The lexical functions are typically connected with aspectual differences and Aktionsart, but also modality, making them often difficult to characterize in exact terms diachronically. The grammatical functions, by contrast, are more focused on the expression of voice, which, then, was a derivational feature in Proto-Tungusic, as it still is in the modern Tungusic languages. It may be noted that most of the suffixes of this group—as also of the other groups—are not, or are only rudimentarily, attested in Jurchenic, which may mean either that they originated as suffixal elements only after the separation of Jurchenic from the rest of Tungusic (implying that Jurchenic was the first branch to break off), or that they were secondarily lost in Jurchenic (in connection with the general reduction of the morphology in this branch). •

Among the voice markers that can be traced back to Proto-Tungusic are the medial suffix *-p-, as in *jalo- ‘to fill’ : MED *jalo-p- ‘to be filled’ (> Ewenki jalu-p-, Nanai jalo-p-) and the causative-passive suffix *-bU-, as in *ii- ‘to go’ : CAUS-PASS *ii-bü- ‘to put/take/carry in’ (> Ewenki ii-w-, Nanai ii-wu-), *göön- ‘to say’ : CAUS-PASS *gööm-bü- ‘to make somebody say; to be said’ (> Udihe gu.mu-, Nanai um-bu-, Manchu guwe-mbu- ‘to cause a sound to be made’). The latter suffix has served as the base for the complex causative marker *-bU-kAAn-, as in *ice- ‘to see’ : CAUS *ice-bü-keen- ‘to show’ (> Ewenki ice-w-‌keen-, Nanai ice-ween-). There are also two reciprocal-cooperative suffixes that have a wide distribution: *-ldUU- (with an irregular reflex in Nanaic), as in *aya- ‘to be reconciled’ : COOP *aya-ldUU- ‘to be mutually reconciled’ (> Ewenki aya-ldii-, Nanai aya-lto-), and *-mAAcI-, as in *soorï- ‘to fight’ : RECIPR *soorï-maacï- ‘to fight each other’ (> Oroch

Proto-Tungusic 51

•

•

• •

soori-maaci-, Nanai sorɪ-maacɪ-, cf. also Ewen coor-‘to hit’ : coor-mat- ‘to hit each other’). The most important marker of the aspectual type is *-jA-, whose original function seems to have been to express progressive action, as still in, for instance, Ewenki dukuu- ‘to write’ : PROGR dukuu-ja- ‘to be writing’. Although technically a derivational suffix, this element had received already in Proto-Tungusic the grammatical notion of future tense, as observed in Ewenic, Orochic, and Nanaic, cf. e.g. *waa- ‘to kill’ : PROGR = FUT *waa-ja- ‘to be going to kill’ : AOR 2PL *waa-ja-ra-su ‘you will kill’ (> Ewenki waa-ja-ra-s, Nanai waa-ja-a-su). In combination with the finitely used futuritive participle marker *-ƞAA, the suffix *-jA- has also yielded the secondary Northern Tungusic complex future marker *-jA-ƞAA-, as in FUT 2PL *waa-ja-ƞaa-sun ‘you will kill’ (= Ewenki waa-jaƞaa-sun). The futuritive function of *-jA- is, however, not attested in Jurchenic, where this element is present only in the composition of a few petrified forms, as in Manchu gai- ‘to take’: gaija- id., ‘to receive’ and possibly bude- ‘to die’ : buce- id. Suffixes producing forms of the Aktionart type include *-cI- for continuatives (‘to do continuously’), e.g. *tege- ‘to sit down’ : CONT *tege-ci- ‘to sit, to be seated’ (> Ewenki tege-t-, Manchu tece-), *-ktA- for distributives (‘to do several times and/or in several locations’), as in *gelee- ‘to request, to search’ : DISTR *gelee-kte- ‘to search’ (= Ewenki gelee-kte-, Nanai gelee-kte-); *-nA.sI- for (re-) iteratives (‘to do repeatedly’), as in ITER *gelee-ne- ‘to request (repeatedly)’ (= Ewenki gelee-ne- ‘to woo’, Uilta gele-ne-); *-lU for inchoatives (‘to begin’), as in *tügde- ‘to rain’ : INCH *tügde-lü- ‘to start raining’ (> Udihe tigde-li-, Nanai tugde-lu-); *-rgUU- for reversives (‘to do again in the opposite direction’), as in *böö- ‘to give’ : REV *böö-rgüü- ‘to give back, to return’ (> Ewenki buurgii-, Oroch buu-gi-, Nanai buu-gu-, Ulcha buu-ju-), and *-sIn- for semelfactives, as in (Ewenic) SEM *aa-sïn- ‘to fall asleep’ (> Ewenki aa-sin-, cf. Nanai ap-sɪn- id.). Modality is clearly involved in the desiderative suffix *-mO.sI-, as in *jep- ‘to eat’ : DESID *jep-mö.si- > *jem-mö.si- ‘to want to eat’ (> Ewenki jem-mu-, Nanai je-mu.si-). A special group of deverbal derivatives is formed by the andatives (‘to go to do something’). The marker of this feature shows irregular correspondences, but the original form may have been *-ƞnA-, as still attested in Oroch alongside with -nA-, e.g. Oroch ice- ‘to see’ : AND ice-ƞne- ~ ice-ne- ‘to go to see’. In Nanaic this element is represented as -ƞdA- ~ -ndA- in Ewenic as *-nAA-, and in Jurchenic as *-nA-, as in Ulcha waa- ‘to kill’ : AND waa-ƞda- ‘to go to kill’, Nanai tee- ‘to sit’ : tee-nde- ‘to go to sit’, Ewenki dukuu- ‘to write’ : AND dukuu-naa- ‘to go to write’, Manchu buta- ‘to hunt, to fish’ : AND buta-na- ‘to go to catch game or fish’. There is no doubt that all of these variants represent suffixalized traces of the independent verb *+ƞene- ‘to go’, which in this construction seems to have been attached to a preceding unmarked verbal stem. As serialized sequences of unmarked verbs are otherwise not typical of Tungusic, it is possible that the andative construction also originally contained an overt converb marker, which had been lost before the Proto-Tungusic stage. Another possibility is that the language had earlier been more analytic and had possessed the ability to use plain (zero-marked) verbal stems in a converbal function.

52 Juha Janhunen

NUMBER AND CASE All nominals (N) in Proto-Tungusic, with or without derivational suffixes (DX), could take suffixal markers for number (NX) and case (CX), in this order (N-DX-NX-CX). The unmarked nominal stem indicated primarily the singular number, although, under certain conditions, it may also have indicated a generic plural, as is still the case in the Tungusic languages of the Manchurian sphere. Normally, however, the plural number of countable entities was marked with a plural suffix, which was independent of the semantics of the underlying nominal, but morphologically differentiated into two allomorphs, *-l, which was added to vowel stems, and *-r, which replaced the final nasal of the stems ending in the unstable nasal */n, as still in, for instance, modern Siberian Ewenki, cf. bira ‘river’ : PL bira-l, muri/n ‘horse’ : PL muri-r. The small number of stems that ended in a consonant other than */n seem to have taken the plural marker *-l with the connective vowel *I, as in Ewenki (dialectal) juug ‘house’ : PL juug.i-l. Although no longer productive in most of the modern Tungusic languages, the consonantal suffixes *-l : *-r were probably the “original” plural markers in Proto-Tungusic, and may have been so in Pre-Proto-Tungusic, as well. Traces of both *-l and *-r can be found also in those modern languages that synchronically use more complex plural markers. The suffix *-r is, for instance, likely to be present in the numeral *jöör = PL *jöö-r ‘two’, which retains the final *r in all languages except those that change it regularly to *l (as in Neghidal juul) or lose it altogether (as in Manchu juwe). The background of the allomorphy between *-l and *-r remains unknown, but it is likely that it was not directly conditioned by phonetic factors. More possibly, the unstable */n may have originated as a separate derivational suffix, perhaps even a class marker, in which case *-r would have been the corresponding plural marker. This possibility is, however, also problematic, since the nominals ending in */n do not seem to form any semantically coherent “class”. Outside of Ewenki and Ewen, the common plural marker is *-sA-l, which obviously contains the class marker *-sA- in combination with the primary plural marker *-l. The form *-sA-l is attested in Neghidal (-sAl), Orochen (-sAl), Solon (-sAl), Oroch (> -sA/g), Nanai (-sAl), Ulcha (> -sAlI), Uilta (-sAl), and Manchu (> -sA). In view of its distribution, this marker must also go back to Proto-Tungusic. It is, however, somewhat surprising that it contains the element *-sA-, which as a derivational class marker refers to homogenenous masses (uncountables). Assuming that the element is the same, it must once have had the function of a more general plural marker. This is also suggested by its presence in the item *ïa-sa ‘eye’ (> Ewenki iesa), of which the plural form *ïa-sa-l is used in an undifferentiated singular/plural function in Southern Tungusic (Nanai nasal, Manchu yasa ‘eye/s’). The plural function of *-sA seems also to be present in a number of other relict forms, such as Ewen jʊʊ ‘house’ : COLL jʊʊ-sag ‘village’ (< *juu-sa-g). The other obscured class marker *-tA, used for individualizable entities (countables) is also attested in combination with the plural marker *-l, yielding the complex plural marker *-tA-l. This combination (with the regular loss of the final *-l) is, however, reliably attested only in Jurchenic, where it is present in a number of kinship terms, as in Manchu sarga-ta ‘women, wives’ (SG sargan), deo-te ‘younger brothers’ (SG deo). These forms may have a connection with several irregular plurals of kinship terms in Northern Tungusic, such as Ewenki am-tii-l (< ? *am-ta-yï-l) ‘fathers’ (SG ami/n). A possible trace of *-tA as a plural marker is also present in the item *pö(-)te ‘child’ (> Ewenki xute), which itself has the irregular plural form *pö(-)ri-l (perhaps a separate etymon).

Proto-Tungusic 53

The plain nominal stem, as well as the corresponding plural-marked forms, functioned as the basic unmarked case form, conventionally identified as the nominative or “indefinite” case. Other cases were marked by endings, which followed the plural markers, if present. The case forms reconstructable for Proto-Tungusic, like their reflexes and analogues in the modern Tungusic languages, may be divided into three functional categories, comprising grammatical, local, and modal cases. This division is, however, not strict, as the local and modal cases could also have grammatical functions, while the grammatical cases, such as the unmarked nominative, could occasionally also have local or modal functions. As the language had no genitive case (proper), all marked cases modified a verbal headword. Just how many cases can be traced back to Proto-Tungusic is a matter of interpretation that depends on how the distributional and formal criteria are evaluated. •

•

For strictly grammatical functions, two cases can be reconstructed, the accusative and the partitive-designative. The accusative, which may have been accompanied by the notion of definiteness and/or specificness, though this is difficult to verify with certainty, indicated the direct object and was marked by the suffix *-bA, which has regular reflexes in all branches and modern languages: Ewenic *-wA ~ *-mA ~ *-bA, Orochic *-A ~ *-mA ~ *-bA, Nanaic *-wA ~ *-yA ~ *-bA, Manchu -be (= -be, with no vowel harmony). It is possible that the suffix originally had the shape *-pA, as still attested in several modern languages in a number of spatial and pronominal forms, where the accusative is used in a locative and/or prolative function, as in Nanai juli- ‘front (part)’ : ACC juli-pe ~ julie-pe ‘ahead’. The partitive-designative case is both functionally and formally more problematic. In absolute use, with no further marking, it indicates a partial object (‘some of’), while in combination with possessive marking it has a designative function (‘for the benefit of somebody’). Both functions are observed in Ewenic, where this case is also used in combination with existential negation, while in Orochic and Nanaic only the designative function is attested. On the formal side, the suffix appears as -yA(-) ~ -A(-) in Ewenki and Orochen-Solon, -yA(-) ~ -ñA(-) ~ -A(-) in Neghidal, -gA- in Ewen, -yA- ~ -A- in Oroch, -nA- in Udihe, -gO- in Nanai, and -jU- in Ulcha. In Uilta, the designative case is syncretized in a complex way with the dative and instrumental cases, while in Manchu it is absent. Although there is no doubt that all these elements belong functionally together, there is no single reconstruction that can explain the formal diversity observed in the modern languages. Assuming that there have been several irregular developments, the shape *-gO, suggested by the Nanai-Ulcha data, could be taken as a possible candidate for the (Pre-)Proto-Tungusic form of the suffix. The principal local cases in the Tungusic languages represent two chronological levels, primary and secondary, both of which make a basic distinction between the dative-locative and ablative functions. On the primary level, the dative-locative was marked by the suffix *-lAA, techically termed the “locative”, and the ablative by the suffix *-kI. On the secondary level, a new system of local cases was formed by the grammaticalization of the spatial *+doo ‘inside’, which itself came to function as a dative-locative, technically termed the “dative”, but which was also used as a coaffix before the primary markers in the triplet DAT *-dOO : LOC *-dOO-lAA : ABL *-dOO-kI. This triplet constitutes still synchronically the core of the local case system of the Northern Tungusic languages, especially of consonant stems, as in Ewenki jaw ‘boat’ : DAT jaw-duu : LOC jaw-du-laa : ABL jaw-du(u)-k, while vowel

54 Juha Janhunen

•

•

•

stems take the primary locative marker, as in Ewenki moo ‘tree’ : DAT moo-duu : LOC moo-laa : ABL moo-du(u)k. This system is also preserved in Uilta, as in ute ‘door’ : DAT ute-du : LOC ute-le : ABL ute-duu, while in Nanai and Ulcha the ablative in *-dOO-kI seems to have been lost. Jurchenic preserves only the dative, represented in Manchu as -de (= -de, with no vowel harmony, but with traces of vowel harmony in Jurchen), though the primary locative is also present in Jurchenic in petrified spatials, as in *do-la ‘inside’ (> Manchu do.lo), which has regular cognates in all branches of Tungusic: Ewenic *doo-laa (= Ewenki doo-laa), Orochic *doolaa (> Oroch doo-loo), Nanaic *doo-la (= Nanai doo-la). It is, however, possible that the primary ablative marker *-kI survives in the otherwise unexplained Manchu ablative marker -ci (-ci), and perhaps also in the Ulcha-Uilta prolative marker -ki, as in uce ‘door’ : PROL uce-ki, and its Uilta cognate -(k)kie-. However this may be, the primary ablative marker is present in the aorist converb forms, in which *-kI follows the aorist participle, as in Ewenki ga- ‘to take’ : CV.AOR-2SG ga-da-k.i-s ‘if/when you take’. Most Tungusic languages also have one or more additional local cases, including, in particular, the prolative and directive, which express specific spatial relations. The primary prolative (prosecutive) marker (‘along, by way of’) may be reconstructed as *-lII, as attested, alongside with the complex marker *-dOO-lII, in the synchronic case paradigms of Ewenic and Orochic, as in Ewenki bira ‘river’ : PROL bira-lii ‘along the river’. The form *-lII may be traced back to earlier *-lA-(y)I, which is obviously based on the locative marker *-lAA. As a productive case the prolative is not attested in Southern Tungusic, but it is present in the Nanaic pronominal forms *xaa-lïï ‘when?’ (> Nanai xaa-lɪ) and *tee-lii ‘then’ (> Nanai tee-li), which have cognates in Northern Tungusic (> Ewenki a-lii, tee-lii). The directive (‘in the direction of’) likewise has a Proto-Tungusic background and was originally marked by the primary suffixes *-sI for spatials and *-tI for regular nouns, as still attested in Southern Tungusic while in Northern Tungusic the expanded forms *-sI-kII and *-tI-kII are used, as in (spatial) *xama-sï ‘backwards’ (> Nanai xama-sɪ, Manchu amasi) : *xama-sï-kïï (> Ewenki ama-skii) vs. (regular nouns) Ulcha xewe/n ‘lake’ : DIR xewen-ti, Nanai ogda ‘boat’ : DIR ogda-ci (< *-tI), Ewenki moo ‘tree’ : DIR moo-tkii, jaw ‘boat’ : DIR jaw-tikii. An idiosyncratic feature of the primary directive forms is that they can take the diminutive suffix *-kAA/n, as in Manchu DIR dosi (do-si) ‘towards the inside’ : DIM dosikan (do-si-kan) ‘a little more towards the inside’. Moreover, in Northern Tungusic, the diminutive suffix precedes the element *-kII, as in Ewenki ure ‘mountain’ : PL-DIR ure-l-tikii ‘towards the mountains’ : DIM ure-l-ti-kee-kii ‘a little bit more towards the mountains’. In Uilta, the directive endings are *-sAi (for spatials) and *-tAi (for regular nouns), which might suggest that the element *-kII had originally the shape *-kA(y)I or *-kI(y)A. In Ewenic, this element is attested also alone in the directive function, as in Ewenki solo-kii ‘upstream’. All of this suggests that *-kII represents the trace of some independent spatial noun that was suffixalized only shortly before the Proto-Tungusic stage. As for the difference between the forms *-sI vs. *-tI, its underlying reason remains unknown. A primary modal case is the instrumental in *-jI (‘by the means of’), which is attested in Ewenic, Orochic, and Nanaic, as in Ewenki ƞaala ‘hand’ : INSTR ƞaala-t (< *ƞaala-ji), Oroch tada ‘arrow’ : INSTR tada-ji, Nanai ogda ‘boat’ : INSTR ogda-jɪ. In Jurchenic, the instrumental function is expressed by the genitive, as in

Proto-Tungusic 55

•

Manchu jilgan ‘voice’ : GEN jilgan-i (jilgan-i) ‘of the voice; with a voice’. It is difficult to say whether this is due to functional or formal syncretism, but the possibility remains that the Jurchenic genitive ending *-i diachronically represents a merger of the original possessive (*-ƞII) and instrumental (*-jI) markers. Moreover, the instrumental case also has the function of a separative, suggesting an original syncretism between modal and local functions. This is particularly obvious in Nanai and Ulcha, which lack the secondary ablative in *-dOO-ki, leaving the separative function to be expressed by the instrumental, as in Ulcha xagdʊn ‘house’ : INSTR xagdʊn-jɪ ‘from the house’. In Nanai, this function can also be expressed by the instrumental form of the suffixal spatial -jIA(-) ‘side’, yielding the complex separative case marker -jIA-jI, as in Nanai jayɪ ‘boat’ : INSTR jayɪ-jɪ ~ SEP jayɪ-jɪajɪ ‘from the boat’. Very probably, the spatial -jIA(-) is a cognate of the element *-gII- in the Northern Tungusic elative suffix *-gII-jI, which also contains the instrumental marker *-jI. In both Southern and Northern Tungusic, the instrumental is attested in several contexts that would normally be expected to require a separative case, including expressions of the type ‘to be made of’, ‘to be afraid of’, and the denotion of the standard of comparison (‘than’). It is difficult to determine, which of the two functions, instrumental or separative, was originally the basic one. Especially the Northern Tungusic languages have a varying number of additional case forms, many of which are language-specific and, therefore, probably recent. Of particular interest are the secondary local cases in Ewenki, marked by the complex endings *-kI-lAA and *-kI-lII and conventionally identified as the “directive-locative” and “directive-prolative”, respectively. The element *-kImay or may not be identical with the element *-kII in the complex directive markers *-sI-kII : *-tI-kII. A different element with the same form, but apparently identical with the primary ablative marker *-kI, is present in the Solon delative in -(dU-)lAA-xI, as in Solon muu ‘water’ : DEL muu-leexi ‘from the water’. It is possible that the delative complex was formed simply by analogy from the equation DAT -dU : ABL dU-xI = LAT -dU-)lA/A : DEL -dU-)lAA-xI. Another possibility, though perhaps less likely, would be that the primary locative marker *-lAA was also a former spatial which, like *+dOO, was suffixalized and could take the primary ablative marker *-kI. Of interest are also a few spatial case forms in Northern Tungusic in which the element *-dOO has merged with a preceding nasal *n, implying that the development *nd > *n has here been active at the suffix boundary, as in Udihe xegi- ‘lower part’ : DAT xegi-nu < *pergii-n-döö.

NUMERALS Proto-Tungusic had a regular decimal system with etymologically separate roots for each basic numeral. Morphologically, all numerals were nouns, though with occasional derivational idiosyncracies. The items for the digits can be reconstructed as: 1 *emö/n, 2 *jöö-r, 3 ïla/n, 4 *dügi/n, 5 *tuñga, 6 ñöƞö/n, 7 nada/n, 8 *japko/n, 9 *xüyegü/n, 10 *ju(w)a/n. With the exception of 2 *jöö-r, which is probably a plural form from the unattested singular *jöö/n, and 5 *tuñga, which is a vowel stem, all these items end in an unstable /n. In fact, in certain uses and forms, all numerals can be used in the plural, as in the Ewenki multiplicatives tunƞa-l ‘five times’, jaa-r ‘ten times’. Formal problems of reconstruction are presented by the item for ‘nine’, which appears in at least two sets of forms in the modern languages. Nanai-Ulcha-‌Uilta xuyu/n, Manchu

56 Juha Janhunen

uyun, and Ewen uyu/n may be explained from a form like *xöyöö/n, though the Southern Tungusic data could also be derived from *xüyü/n or *xüyüü/n, while Ewenki-Solon yegi/n, Orochen yeyi/n, and Udihe yeyi/n presuppose a shape like *yegü/n. All of these forms can, at least technically, be bridged by the trisyllabic reconstruction *xüyegü/n (Benzing), which is also suggested by Neghidal iyegi/n and Jurchen †uye(’)un. We may, thus, assume that Northern Tungusic, with the exception of Neghidal and Ewen, lost segments at the beginning, i.e. *xüyegü/n > *yegü/n, while Southern Tungusic and Ewen, in at least two separate developments, lost segments in the middle and possibly underwent vowel assimilations, i.e. *xüyegü/n > *xüyeü/n ~ *xüyüü/n ~ *xöyöö/n. Oroch xuyu/n must, however, be a borrowing from Nanaic, replacing the expected form *yeyi/n (as in Udihe). The reconstruction of the item for ‘five’ is also uncertain in the details, for the internal cluster shows a unique set of correspondences: Nanai toɪƞga, Ulcha tʊnja > Uilta tunda, Manchu sunja < Jurchen †šunja < *tunja, Oroch tʊƞa ~ Udihe tʊƞa/n, Ewenki tunƞa ~ Ewen tʊnƞa/n, Neghidal toñƞa, Orochen tʊƞƞa, Solon tʊƞa ~ toƞa. Note that in Ewen and Udihe the item has been restructured into a nasal stem. The rest of the diversity would seem to be explained by assuming that the original cluster was *ñg, but the status of the syllable-final palatal nasal is problematic in view of the general phonotactic framework of Proto-Tungusic. The second member of the cluster was, in any case, probably the velar stop *g and not the velar nasal *ƞ, for the correspondences differ from those of *añƞa ‘year’. For the higher decades, Proto-Tungusic may originally have applied the multiplicative principle, as still in, for instance, Ewenki: 20 juur+jaa-r, 30 ilan+jaa-r, etc., though in some languages and dialects the derivational principle is also attested, as in Orochen 50 tʊƞƞa-ƞɪ, 60 ñuƞun-ƞi, etc. The compounds for the decades have in many languages undergone irregular simplifications, as in Manchu 60 ninju (nin-ju < *niƞgun+juwa/n), 70 nadanju (nadan-ju < *nadan+‌juwa/n). An exceptional item is Manchu 50 susai (? su-sai, based on 5 sunja), which has been borrowed into Nanaic, yielding Nanai sosɪɪ, Ulcha sʊsaɪ. For the decades 20–40 many Tungusic idioms use items borrowed from Mongolic: 20 *xorï/n (in all branches), 30 *gutï/n (in all branches), 40 *deki (in all branches except Orochic). The correspondences and distribution of these items suggest that they were borrowed already before the Proto-Tungusic stage, and the source language was Pre-Proto-‌Mongolic. It cannot, however, be ruled out that the borrowing took place from Para-Mongolic to Pre-Proto-Jurchenic, in which case the data in the other Tungusic languages would represent later, though still relatively early borrowings from Jurchenic. In some cases, languages that have been in later contact with Mongolic have replaced the old borrowings with newer ones, as in Khamnigan Ewenki (Urulyunggui) 30 guci/n (← Khamnigan Mongol guci/n), (both dialects), 40 duci/n (← Khamnigan Mongol duci/n); note also Orochen 90 yeree/n (← Mongolian yere/n). It is impossible to reconstruct any uniform principle for the formation of the intermediate numerals, though the additive principle, as in Ewenki 11 jaan umun, 12 jaan juur, 21 juur+jaa-r umun, may have been applied. More importantly, the entire set for the teens from 11 to 19 was borrowed by Jurchen from Para-Mongolic, yielding 11 †omšo, 12 †jirxon, 13 †gorxon, 14 †durxun, 15 †tobuxon, 16 †nilxun, 17 †daluxon, 18 †ñoxun, 19 †oñoxon. Of this set Manchu preserves only omšon ‘eleventh (month)’, jorhon ~ jorgon ‘twelfth (month)’, and tofohon ‘fifteen, fifteenth (day), (day of) full moon’, of which, because of their calendrical significance, the items for ‘eleven’ and ‘twelve’ were transmitted also to Solon, as in Solon ʊnsun bie ‘November’ and jʊrgʊn bie ‘December’.

Proto-Tungusic 57

As for the powers of ten, Tungusic has two native items for ‘hundred’, neither of which can technically be traced back to Proto-Tungusic because of their limited distribution. Northern Tungusic has 100 *ñamaa : *ñamaa-jïï, as attested in Ewenki ñamaa : ñamaajii, Ewen ñama < *ñamaa, Orochen ñamaajɪ, Solon namaaji, and Neghidal ñama, while Southern Tungusic uses the Manchu item 100 tanggu (taƞgu), as attested in Nanai taƞgo, Ulcha taƞgʊ, Uilta taƞgu, Oroch taƞgʊ, Udihe taƞgu, but also, alongside with *ñamaajï, in Neghidal taƞgʊ, Orochen taƞgʊ, Solon taƞgʊ. The Manchu item itself represents the nominal use of the verb *taƞu- ‘to count’, attested in all branches, as in Ewenic *taƞï- (> Ewenki -taƞi-), *taƞï- (> Oroch taƞi-), Nanaic *ta(w)u- (= Nanai tao-, with an apparently irregular loss of medial *ƞ). For the higher powers of ten, virtually all Tungusic languages show reflexes of 1000 *mïƞa/n and 10.000 *töme/n, both borrowed from Mongolic. While it is technically possible to trace these items to the Proto-Tungusic level, their status in the individual languages was probably enhanced by later contacts and internal borrowing. Much of the derivational morphology of numerals is specific to the individual branches and languages. It is nevertheless likely that the suffixes forming ordinals, Ewenic *-II < *V-(y)I vs. Orochic and Nanaic *-IA(-), represent (or contain) a common element, as in Ewenki il-ii ‘third’ < *ïl-ïï < *ïla-(y)ï vs. Nanai ɪl-ɪa- id. Typically, the ordinals are marked by the third person plural possessive suffix, which in these cases is present as a relict even in Manchu, as in ORD-PX3SG Ewenki il-ii-tin ‘third’, Nanai ɪl-ɪa-cɪa id., Manchu ila-ci (ila-ci) id. The formation of the ordinals has, however, undergone many innovations that make exact reconstructions difficult. Easier to reconstruct is the suffix forming distributives, which is *-tAl in all branches, identical with the complex plural marker *-tA-l, as in Ewenki juu-tel ‘by the twos’ = Manchu juwe-te (juwe-te < *jöö-te-l). Another complex suffix, attested in all branches except Jurchenic, and containing, at least superficially, the collective plural marker *-sA, is *-l-sA for counting periods of days, as in Ewenki ila-lla ~ ila-lda ‘(for) three days’ = Nanai ɪla-lta id. The element *-l- is, however, not the regular plural marker, since it is attested also in *jöö-lse ‘(for) two days’ > Ewenki juu-lle ~ juu-lde = Nanai jue-lte. PRONOUNS On both formal and functional grounds the pronominal words that can be reconstructed for Proto-Tungusic may be divided into personal, demonstrative, interrogative, and reflexive pronouns. Most of the pronominal roots belong to the maximally simple type of auxiliary morphemes with the structure *(C)V. In absolute use, these may have had a tendency to be lengthened to *(C)VV, as in the modern Tungusic languages. Also, most of the actual free forms involved extensions by suffixes to *(C)V-C(-) or *(C)V-CV(-). The personal pronouns of the first and second persons form a matrix in which the first person is indicated by the consonant *b and the second person by the consonant *s, while the distinction between the numbers is indicated by the vowels *i for the singular and *ö for the plural. In the oblique stem, before case suffixes, a “pronominal n” is added, before which the first person consonant is represented as *m. Of the actual case forms, Northern Tungusic allows the reconstruction of a “partitive” in *-A, used to mark the object, while all branches have reflexes of an accusative in *-bA, also used in the object function. In Northern Tungusic, the accusative forms are based on the partitive stem, while in Southern Tungusic the accusative marker seems to be attached directly to the nasal stem of the pronoun. All branches also have traces of a derivational possessive form in *-ƞII,

58 Juha Janhunen

expressing the possessor in predicative position, while both Jurchenic and several Ewenic and Nanaic languages suggest the additional presence of an actual adnominal genitive in *-I (Table 4.7). TABLE 4.7 PROTO-TUNGUSIC PERSONAL PRONOUNS  

NOM SG PL

OBL

PART

ACC

GEN

POSS

1

*bi

*min-

*min-e

*min(-e)-be

*min-i

*min-ƞii

2

*si

*sin-

*sin-e

*sin(-e)-be

*sin-i

*sin-ƞii

*bö

*mön-

*mön-e

*mön(-e)-be

*mön-i

*mön-ƞii

*sö

*sön-

*sön-e

*sön(-e)-be

*sön-i

*sön-ƞii

1 2

EXCL

Due to the secondary analogy of the oblique stem, the first person singular pronoun is represented in Nanai as mii. This must be a recent development, as Ulcha and Uilta have still bii. In the plural pronouns, the vowel of the basic forms has been a matter of some dispute in the past, and reconstructions such as *böö : *söö, *böe : *söe, and *büü : *süü have also been proposed. The correspondences are, indeed, slightly exceptional, showing Manchu be (< *buwe) : suwe and Nanai bue : sue (< *böö : *söö) against buu : suu in all other languages and branches. These may be compared with, for instance, *möö ‘water’, which yields mɵɵ in Ewen and Ulcha-‌Uilta against mue in Nanai and muu elsewhere (in Manchu only in the complex form muke = mu-ke id.). However, the variation is probably best explained by assuming that the lengthening of the original short monosyllables of the type *(C)V took place separately in the different branches and languages. The primary difference between items such as *bö : *sö and *möö was one of vowel length. In Ewenic, Orochic, and Jurchenic the first person plural pronoun is normally used in the exclusive function (‘we without you’), as opposed to a separate inclusive pronoun (‘I/we with you’), represented variously as Ewenki bit ~ mit ~ mut, Ewen but ~ mut, Neghidal bit, Solon-Orochen miti, Oroch biti, Udihe miti ~ minti, Manchu muse (muse). The variation in the form of the inclusive pronoun, as well as its absence in Nanaic, suggests that it could be a secondary feature, innovated separately in the different branches and even languages, possibly at least partly under the influence of Mongolic (where the inclusive pronoun is also a relatively recent feature). Structurally, all the variant forms seem to be combinations of a first person pronoun with a second person pronoun, but the combinations differ depending on whether the underlying pronouns are singular or plural, and whether the first component appears in the basic form or in the oblique stem. On the hand, the Manchu inclusive pronoun might also be analysed as a plural in *-sA-l, i.e. < *mu/n-se-l. Most curiously, however, in the Ewenic and Orochic forms, the second component is *-ti (*bi+*ti : *mi/n+*ti ~ *bö+*ti : *mö/n+*ti), which is different from the regular form of the second person singular pronoun *si. This may mean that the original form of the second person pronouns contained an initial *t, i.e. 2SG *ti : 2PL *tö, though the development to 2SG *si : 2PL *sö must have taken place in Pre-Proto-Tungusic and would, in any case, have been irregular. Incidentally, traces of *+ti are also present in the system of person markers. If this is so, the inclusive pronouns in Ewenic and Orochic must also be old, though the observed variation in the forms suggests that the system has undergone secondary evolution. Actual pronouns for the third person are present only in Jurchenic, as in Manchu 3SG i : GEN in-i, 3PL ce : GEN cen-i. Although absent in the other branches, these

Proto-Tungusic 59

items are likely to be old and may be traced back to Proto-Tungusic 3SG *i : 3PL *ti, of which the oblique stems, *+i/n- : *+ti/n- are attested as suffixal person markers in the other branches. For the pronominal function the other branches use the dummy nominal stem *nuga/n- > Ewenic *nuga/n- ~ *nuƞa/n- ‘he, she’ (= Ewenki nuƞa/n-), Orochic *nuƞa/n- (> Udihe nua/n-), Nanaic *noa/n- (> Ulcha naa/n-, Uilta noo/n-, Nanai ñoa/n- with an apparently secondary palatal nasal). In inflected forms several patterns are observed. The original pattern is probably preserved in Ewenic, where the dummy stem takes the regular plural marker *-r, and all marked case forms take the third person possessive suffixes, either singular or plural, as in Ewenki NOM nuƞan : ACC-PX3SG nuƞan-ma-n : PL nuƞa-r- : PL-PX3PL nuƞa-r-tin : PL-ACC-‌PX3PL nuƞa-r-wa-tin. In Nanaic, the difference between the singular and plural is expressed only by the possessive suffixes, as in Nanai NOM PX3SG ñoa-ni : ACC ñoam-ba-nɪ : PL ñoan-ci : ACC-PX3PL ñoam-ba-cɪ. In Orochic, the possessively marked plural has been reanalysed as a new base form for the plural declension, as in Oroch PX3SG nʊƞa-ni : ACC-PX3SG nʊƞa-ma-ni : PL nʊƞanti : ACC nʊƞanti-wa. The principal reflexive pronoun may be reconstructed as *mee/n ‘(one)self’ : OBL *mee/n- : PL *mee-r, attested in this function in all branches except Jurchenic. The plain nasal stem of this pronoun, in Orochic and Nanaic also *meene > mene, is used adnominally to indicate reflexive possession, ‘(one’s) own’, as well as in the emphatic function, ‘(by) (one)self’. In the latter function, possessive suffixes can be added, as in Ewenki PX1SG meen-mi ‘(I) myself’, while in other functions the reflexive suffixes (for singular and plural) are used. Case marking is absent in the object position, as in Ewenki (object forms) RX meen-mii ‘(one)self’: PL-RX-PL mee-r-we-r ‘(them)selves’, though in Nanaic the accusative marker is used with a plural referent, as in Nanai SG RX mee-pi : PL ACC-​RX.PL mee-pe-eri. In the other cases the regular nominal case markers are used, as in Nanai DAT-RX meen-du-yi : DAT-RX.PL meen-du-eri. In all branches, the reflexive pronoun can be reduplicated yielding complexes with a distributive meaning like mee&mee-, meen&meen, mene&mene ‘each one’. Importantly, the reduplicated complex is also attested in Manchu, meni&meni > meimeni ‘each, every, individually’, confirming that the reflexive stem is of a Proto-Tungusic origin. Otherwise Manchu uses the noun beye ‘body’ < *beye ‘man, person, body’ in the function of a dummy reflexive pronoun, a feature ultimately borrowed from Mongolic and attested also elsewhere in Tungusic. The demonstrative pronouns reflect a basic dichtonomy between the proximal and distal functions, as expressed by the roots *e- (> e- ~ ee-) ‘this’ vs. *ta- (> ta- ~ taa-) ‘that’, never attested without suffixes. The independent singular forms in Proto-Tungusic may be reconstructed as *e-ri vs. *ta-rï, with regular reflexes in all branches except that Nanai and Ulcha have *e-ri vs. *te-ri, while Manchu has ere (e-re) vs. tere (te-re). The original plural forms may have been *e-ri-l vs. *ta-rï-l, as still in Ewenic, or also *e-rise-l vs. *ta-rï-sa-l, as in Uilta, while Manchu has ese (e-se) vs. tese (te-se). Both the plain roots and the extended forms are attested with case markers, but the total paradigm in the modern languages is often incomplete and mixed. Moreover, many case forms of the demonstratives have become lexicalized in spatial or other functions, as in Ewenki ACC er-we ‘this much’ vs. tar-wa ‘that much’, DAT e-duu ‘here’ vs. ta-duu ‘there’ : INSTR e-r.i-d ‘for this reason’ vs. ta-r.i-d ‘for that reason’. In spatial functions, various derivational extensions are also used, among which the element *-bA- ~ *-bAA- > -wA- : -wAAis attested in all branches, as in Ewenki DIR e-we-skii ‘in this direction’, LOC ta-waa-laa ‘over there’. There is an obvious connection with the noun *buga ‘place’, in the modern languages also represented as bua ~ boo ~ baa, Manchu ba (ba), as is also suggested by

60 Juha Janhunen

the Manchu spatial nouns uba ‘this place; here’ (< *e+ba) vs. tuba ‘that place; there’ (< *te+ba). A separate temporal item present in all branches except Jurchenic and based on the proximal pronoun is *e-sii ‘now’, possibly a directive form in *-sI-. Another spatial root traceable back to Proto-‌Tungusic is *caa- ‘the one further away’, as in Oroch DAT caa-du ‘behind’, Manchu LOC cala (ca-la) ‘over there (on the other side)’. The two basic interrogative pronouns, present in all branches, are *xï(y)a(-) (> Ewenic and Orochic *ïa-, Jurchenic *ya-) ~ *xa(y)ï(-) (> Nanaic *xaï-) ‘what?’ and *ƞüü (also reconstructed as *ƞüi) ‘who?’. The former is a pro-nomen-verbum also attested as a verb, as in Ewenki ie- = Nanaic *xaï- ‘to do what?’, while in pronominal use with case markers it often forms lexicalized expressions which correlate with the corresponding demonstratives, as in Nanai DAT ei-du ‘here’ (< *e-ri-döö) vs. ta-do ‘there’ (< *ta-doo) vs. xaɪ-do ‘where?’. Probably connected with *xa(y)ï(-) (= perhaps *xa-yï) is the more simple root *xa- ~ *xaa-, which is attested in several lexicalized items, including *xa-duu ‘how many?’ (> Ewenki a-dii, Nanai xa-do), *xa-so/n ‘how much?’ (> Ewenki a-su/n, Ulcha xa-sʊ/n), PROL *xa-lïï ~ *xaa-lïï ‘when?’ (> Ewenki a-lii, Nanai xaa-li), DIR *xa+baa‌sï(-kïï) ‘in which direction?’ (> Ewenki a-wa-skii, Oroch a-waa-si, Nanai xa-osɪ, Manchu absi). Also possibly related is the root *xo- ~ *xoo- : *xoo-nï ‘how?’ (> Ewenki oo-n, Nanai xo-ni) : *xoo-kïï ‘how many?’ (> Ewenki oo-kii, only in Northern Tungusic) : *xookïn ‘when?’ (> Ewenki oo-kin, only in Ewenic). A separate root, attested in Northern Tungusic is *ii(-) : *ii-r(i) ‘which one?’ : DAT *ii-döö ‘where?’ : LOC *ii-lee ‘where?’, etc. PERSON MARKING Proto-Tungusic had two sets of suffixal person markers. One of the sets—the verbal predicative endings (VX)—marked the subject (actor/agent) of finite verbal predicates, while the other set—the possessive suffixes (PX)—marked primarily the possessor of nominal headwords, but secondarily also the subject (actor/agent) of non-finite verbal forms in both finite and non-finite use. Both systems of person markers differentiated between three persons in the singular and plural, as well as between the exclusive and inclusive functions of the first person plural. The system of possessive suffixes was complemented by a separate set of reflexive suffixes (RX), which indicated reflexive possession with regard to a singular or plural referent. Because of the presence of suffixal person marking, which was apparently obligatory in the protolanguage, the use of explicit pronouns in the functions of subject or adnominal possessor was probably superfluous in neutral contexts. Therefore, it may be assumed that the basic (nominative) and adnominal (possessive) forms of the personal pronouns were mainly used to express emphasis. The situation has, however, evolved, and several modern Tungusic languages make frequent use of double marking (independent pronoun + person marking). In the possessive construction this has been accompanied by the development of the originally derivational possessive form into an actual genitive case form. Person marking of both the predicative and the possessive type is absent in Jurchenic, where only analytic (dependent-‌marked) constructions are possible. The absence of person marking must, however, be due to the secondary loss of this feature, probably under the influence of Mongolic, but possibly also of Chinese. The fact that person marking must have originated in Pre-Proto-Tungusic is confirmed by the very forms of the person markers. All person markers are based on enclitically used pronominal forms, which were initially postposited, then cliticized, and finally suffixalized to the preceding nominal or

Proto-Tungusic 61

verbal base. The systems of predicative and possessive person markers seem to have been partly overlapping already in the protolanguage, and further confusion in the individual branches and languages has been created by the evolution of the system of finitely used verbs, as well as by positionally conditioned phonological changes in the forms of the suffixes. As a result, some languages, with Oroch as the extreme example, have virtually eliminated the original distinction between predicative and possessive markers. The original system is better preserved in Ewenic, which, with some modifications, can serve as a tentative basis for the Proto-Tungusic reconstruction of the system (Table 4.8). TABLE 4.8 PROTO-TUNGUSIC PERSON MARKERS  

 

AOR

VX

PX

SG

1

 

*-n- > *-m-

*-bI

*-bI

2

*-n-

*-sI

*-sI

3

*-RA-

*-nI

*-nI

*-RA-

*-bO

*-bOn

RX PL

1

*-bII EXCL INCL

*-pU

*-tI

2

*-sO

*-sOn

3

-Ø

*-tIn

RX

*-bArI

Segmental alternations: *A= *a *e, *O = *o *ö, *I = *ï *i, *-RA- = the aorist formative *-rA- (> *-tA- ~ *-dA-) ~ -dA- ~ *-sI-, depending on the verb class and stem type.

In this system, the difference between the predicative and possessive sets is evident in the first person plural exclusive and second person plural forms, which in the predicative set reflect the basic forms of the pronouns PL1.EXCL *+bö : PL2 *+sö, while the corresponding possessive forms are based on the oblique stem, or possibly the adnominal genitive forms, of these same pronouns, i.e. EXCL.OBL(-ADN) *+bön(-i) : *+sön(-i). It may be noted that the first person plural exclusive possessive suffix has an initial *b (> Ewenki -wun), and not *m, as in the corresponding oblique stem of the independent pronoun (> Ewenki mun-). This could mean that the grammaticalization of this suffix took place before development *b > *m, as conditioned by the following syllable-final nasal. However, it is also possible that the suffix was restructured under the analogy of the corresponding predicative suffix, in which the initial *b (> Ewenki -w) is the expected representation. The third person possessive suffixes are also based on the original oblique stems (or adnominal possessive forms) of the corresponding pronouns, i.e. SG3PX *-nI (> Ewenki -n) < *+in(-i) : PL3PX *-tIn < *+tin(-i), though the final nasal of the third person plural suffix is preserved only in Ewenic (> Ewenki -tin), while the other branches have *-tI. The distribution of the third person predicative suffixes seems to have undergone restructuring, for which reason it is possible that the Ewenic system does not represent the original state. In Ewenic, the third person singular predicative suffix is identical with (and probably generalized from) the corresponding possessive suffix, while the third

62 Juha Janhunen

person plural has no suffix at all (-Ø = zero marked). In several other languages, as in Nanai, it is the third person singular that is unmarked, while the third person plural is marked by the regular plural marker *-l, otherwise attested in the nominal declension. It may be noted that Jurchenic (Manchu) also preserves the third person possessive suffixes in at least two relict forms, in that SG3PX *-nI is present in the third person imperative suffix -kini (< *-gI-nI), while PL3PX *-tI is present in the suffix for ordinal numerals -ci (< *-tIn). In the first and second person singular, the two sets of person markers are identical in Ewenic, corresponding directly to the personal pronouns SG1 *+bi : SG2 *+si. The actual difference between the two sets is that while these suffixes in possessive use are attached directly to the nominal (or nominalized verbal) stem, in finite use they require the presence of a special finite (aorist) marker *-n-, which, then, merges with the person marker and yields the complex markers SG1 *-m-bI (> Ewenki *-mI > -m) : SG2 *-n-sI (> Ewenki -nri ~ -ndi ~ -ni). The complex suffix *-m-bI is also present in Nanaic (> Nanai -m.bI), but unlike Ewenic, Nanaic shows *-tI in the second person. This may mean that the pronoun underlying the Nanaic marker had the shape SG2 *+ti, as is also suggested by the Northern Tungusic first person plural inclusive pronoun *bi-ti ~ *min-ti, etc. This, again, suggests that the Ewenic system may for this detail be more innovative. The most difficult to analyse diachronically are the first person plural inclusive markers, which may tentatively be reconstructed as *-pU for the predicative and *-tI for the possessive set. The possessive suffix PX1PL.INCL *-tI is only attested in Ewenic (> -t) and Orochic (> -ti), which is why the vowel *I is technically ambiguous and could represent either *I or *U at an earlier level. There is, however, an obvious connection with the corresponding inclusive pronoun, suggesting that the suffix was formed, perhaps only in Northern Tungusic, by the development *+biti > *-tI. By contrast, the predicative suffix VX1PL.INCL *-pU, as attested in Ewenic (> -p) and Orochic (> -pi), has a cognate in the Nanaic non-differentiated first person plural predicative suffix *-pU (> Nanai -pO, Ulcha -pU, Uilta -pu). There are several possible explanations for this situation, but if we assume that the absence of the exclusive/inclusive distinction in Nanaic, as opposed to all other branches, is secondary, then the suffix *-pU must originally have had the inclusive function also in Nanaic, though it was later generalized to the exclusive function, as well, replacing the suffix VX1PL.EXCL *-bO. If this is so, then *-pU must also be a reflex of a corresponding independent inclusive pronoun. The problem here is that it is difficult to relate the form *-pU to the pronoun *biti, as attested in Northern Tungusic. More likely, the pronoun underlying *-pU had a different composition. However this may have been, the suffixalization of the inclusive marker for the predicative set is likely to have taken place at an earlier time level than that of the corresponding possessive suffix. The reflexive marker is also connected with the corresponding pronoun. This is especially evident from the plural form *-bArI (> Ewenic *-bAr, Orochic *-bAyI, Nanaic *-bArI), which clearly contains the regular nominal plural marker *-r- and may be compared with the reflexive pronoun *mee/n : PL *mee-r, suggesting that the pronoun had originally the shape *bee- : SG *bee/n : PL *bee-r, in which the initial *b was secondarily nasalized under the assimilative impact of the final nasal of the singular form, i.e. *bee/n > *mee/n. The plural reflexive suffix contains, however, the additional vowel *I, which must represent a separate suffixal element, i.e. *-bA-r-I. This same suffixal element is present in the singular form *-bII (> Ewenic *-bii, Orochic *-bI, Nanaic *-bI), which apparently goes back to earlier *-bA-(y)I. The function of the element *I remains unknown.

Proto-Tungusic 63

VERBAL STEMS An important formal difference between nominals and verb(al)s in Proto-Tungusic seems to have been that the former, in their basic form (nominative), could occur as free morphemes with no suffix, while the latter were always marked by one or more suffixes, which determined their status with regard to the four categories of verbal forms: (i) imperatives, (ii) finite indicative forms, (iii) participles, and (iv) converbs. The only synchronic exception from this seems to be the basic second-person imperative form in Manchu, which for most verbs is identical with the plain stem. This is, however, likely to be a relatively recent innovation in Manchu, based on structural influence from Mongolic. Apart from the morphological category of each given verbal form, which was determined by the syntactic environment, verbal morphology involved features that were conditioned lexically and/or phonologically. The lexical factor determined the conjugational class of the verb, while the phonological structure of the stem caused morphophonological variation at the boundary of the stem and the suffix (as well as within the suffix with regard to the requirements of vowel harmony). In terms of phonological structure, verbal stems were divided into those ending in a vowel (vowel stems) and those ending in a consonant (consonant stems). As in the nominals, consonant stems were rare and represented a type that must have been receding already in Proto-Tungusic. The consonants attested as final segments of verbal stems included *g *p *l *n, of which only *p and *n had a somewhat higher frequency because of their occurrence in the composition of derivational suffixes: *p in the medial suffix *-p-, and *n in the semelfactive suffix *-sIn-. The morphological differences between the individual verbal lexemes are mainly evident in one particular feature, the formation of the so-called aorist stem, which involves a secondary stem that served in the protolanguage as the base for the finite indicative paradigm, but also for other related functions. The aorist stem was formed by the aorist marker, whose form depended on both lexical and phonological factors. From this point of view, verbs may be divided into the following types: (1) General vowel stems: These took the aorist marker *-rA(-), e.g. *ice- ‘to see’ : AOR *ice-re(-), *waa- ‘to kill’ : AOR *waa-ra(-). In the languages spoken in the Amur basin, the form of this marker has been affected by the tendency of eliminating the vibrant *r in the intervocalic position, yielding variously -yA(-) (as in Neghidal and Oroch) or -Ø- (zero, as in Udihe), though secondary alternations have arisen due to the occasional preservation of *r in certain positions, yielding *-A(-) ~ *-rA(-) (as in Nanai). Also, the loss of stem-‌final high vowels and the formation of secondary consonant stems in Ewenic has been followed by consonantal assimilations due to which the aorist marker can secondarily also appear in the shapes *-lA(-) ~ *-nA(-) ~ *-dA(-) ~ *-tA(-). (2) General consonant stems: These were also originally marked by *-rA(-), in which, however, the vibrant had developed to a stop in the postconsonantal position, yielding, already in the protolanguage, either *-tA(-) after the strong obstruent *p or *-dA(-) after the sonorants *l *n, e.g. *jep- ‘to eat’ : AOR *jep-te(-) (> Manchu jete), *göön- ~ *ƞöön- ‘to say’ : AOR *göön-de(-) ~ *ƞöön-de(-). In Northern Tungusic, the cluster *-n-d- underwent the regular change to *-n-, yielding the synchronic aorist marker -A(-) for this stem type, as in Ewenki gun- ‘to say’ : AOR gun-e(-) vs. Nanai un- : AOR un-de(-). Two verbs that probably also originally belonged to this group are *naag- ‘to hit (a goal)’ and *neeg- ‘to put’, in which the final consonant seems to

64 Juha Janhunen

have been the weak obstruent *g, suggesting that *g also triggered the development *-rA(-) > *-dA(-). In the modern languages these verbs are vowel stems taking either *-rA(-) or *-dA(-) in the aorist, as in Ewenki naa-ra(-) vs. Ewen naa-da(-), or the marker *-dA(-) has been incorporated into the stem, as in Udihe nagda- < *nag-da‘to hit (a goal)’. In this connection the status of the stem-final obstruent *p deserves some attention. It is curious that only the two obstruents *p (strong) and *g (weak) are attested in stem-final position, while, for instance, *b (weak) and *k (strong) are not. Most probably, this is simply because of the residual character of primary consonant stems, with stems ending in other consonants having been lost. The fact that the final segment of *jep- ‘to eat’ was, indeed, *p (strong) and not *b (weak) is confirmed not only by the allomorph *-tA(-), conditioned by the preceding strong obstruent, as in Khamnigan Ewenki (Borzya) AOR.1SG *jep-te-m, but also by a few examples in which this stem seems to be handled like a vowel stem, as in Khamnigan Ewenki (Urulyunggui) AOR.1SG jep.i-m (< *jep.i-mi), Manchu IMP jefu (jefu < *jep-u). It should be noted that although *p was distinctive in stem-final position, it followed the rules of consonant phonotactics and lost its distinctivity with regard to *b (= *B) in syllable-final position, which is why it is widely represented as *b > w in the modern languages, as in Siberian Ewenki AOR.1SG jew.u-m (< *jeb.i-mi). Against this background, it could be speculated whether the exceptional correspondence in Nanaic *au- vs. Ewenic & Orochic *aa- ‘to sleep’, especially in view of the Nanaic form SEM ap-sɪn- ‘to fall asleep’, should not be explained by assuming the presence of an original stem-final *b, i.e. *ab-. (3) Dynamic verbs: This is a small and closed group of vowel-stem verbs which instead of *-rA(-) took the marker *-dA(-) in the aorist, e.g. *oo- ‘to become’ : AOR *oo-da(-), *bö- ‘to die’ : AOR bö-de(-). All verbs belonging to this class appear to have a semantic profile connected with dynamic movement or change of state, the other items being *di- ‘to come’ and *ga- ‘to take’. In view of this shared semantics, it is possible that the element *-dA- had originally a related semantic function. It could also be speculated, though it cannot be proved, that these verbs were originally consonant stems in *d, an otherwise unattested stem type, after which the vibrant of the regular aorist marker *-rA(-) could have been lost, i.e. *-d-rA(-) > *-d-dA(-) > *-dA(-). In the modern languages, this stem type has tended to be marginalized, but traces of it are preserved in all branches, including Jurchenic, where the element *-dA- has become incorporated into the stem, as in Manchu ji- ‘to come’ : AOR jidere (= jide-re < *ji-de-re, containing also the productive aorist marker -re). It may be noted that the verbs *naag- ‘to hit a goal’ and *neeg- ‘to put’, though often placed in this class, should probably be understood as regular consonant stems in *g. (4) Static verbs: A somewhat larger but also closed group, these verbs comprise both vowel stems and consonant stems and share a semantic reference to static activity, including static properties. For this group, the aorist was marked by *-sI(-), e.g. *gïl- ‘to be cold’ : AOR *gïl-sï(-), *deg- ‘to fly’ : AOR *deG-si(-). In the modern languages, this class has largely been eliminated by transferring the items to the regular vowel or consonant stems. All languages preserve, however, at least in traces, two auxiliary items belonging to this class: the copula *bi- ‘to be’ : AOR bi-si(-) and the negation verb *e- : AOR *e-si(-), though, again, there are examples of the marker *-sI- to have become incorporated into the stem, as in Manchu bi- ‘to be’ : AOR bisire (= bisi-re < *bi-si-re, with the additional aorist marker -re). Obviously,

Proto-Tungusic 65

the element *-sI- was originally a suffix indicating static activity. It is, however, unclear why this suffix did not take the regular aorist marker *-rA-. It could also be speculated that these verbs were originally consonant stems in *s, another otherwise unattested stem type, after which the regular aorist marker *-rA(-) could have been lost, i.e. *-s-rA(-) > *-s-A(-), though, again, we do not know why the aorist stem vowel in this case is *I and not *A. VERBAL FORMS The four categories of verbal forms were all marked by category-specific suffixes and suffixal sequences, as summarized below: (i)

Imperatives constituted in Proto-Tungusic a separate group of heterogeneous finite forms for various types of requests and commands. Apart from imperatives proper, directed at the second person, these forms comprised “voluntatives”, directed at the first person, as well as “permissives”, directed at the third person. Although differentiated by the person, not all imperative forms contained an actual person marker. Also, the systems in the modern languages are not fully congruent. Even so, it is possible to reconstruct an “imperative” marker *-gI-, which is present in the forms IMP1SG *-gI-tA (> Ewenki -ktA, Nanai -gI-tA, with an unidentified final element *-tA) : 3SG -gI-nI (> Ewenki -gin ~ -ƞin, Nanai -gInI, Manchu -kini = -ki-ni, with PX3SG *-nI) : 1PL.INCL *-gI-pU (> Nanai -gI-pu, with VX1PL.INCL *-pU vs. Ewenki -gAA-r ~ -gAA-t) : 3PL *-gI-tI/n (> Ewenki -k-tin ~ -ƞi-tin, Nanai -gI-cI, with PX3PL *-ti/n). For the basic second person imperative, Northern Tungusic has 2SG *-kAl (> Ewenki -kAl, Oroch -kA) : 2PL *-kAl-sO (> Ewenki -kAl-su, Oroch -kA-sU), while Southern Tungusic has the possibly innovative type based on the aorist stem and extended by the vowel *-U (> Nanai -O, Manchu -O). It may be noted that Manchu still retains suffixally marked imperative forms for a few monosyllabic verbs, as in bi- ‘to be’ : IMP bisu (< *bi-s-u), je- ‘to eat’ : IMP jefu (< *jep-u), although the regular second-person imperative for most verbs is unmarked. (ii) The system of finite indicative forms was based on the aorist stem, expanded by the person markers of the predicative type. Importantly, however, the first and second person singular forms, as suggested by Northern Tungusic, took the marker *-nagainst the regular aorist marker *-rA- in all other persons, as in Ewenki waa- ‘to kill’ : AOR 1SG waa-m < *waa-n-bï : 2SG waa-n-di ~ waa-n-ni < *waa-n-sï : 3SG waa-ra-n < *waa-ra-nï, etc. This means that the elements *-n- and *-rA- filled similar functions at some prehistorical stage of the language, and it is even possible that the forms with *-n- represent the more “original” type of aorist formation, later replaced by *-rA- in all but the two most frequently used personal forms. In the other branches, the system has undergone more restructuring, but a trace of *-n- is present also in the Nanaic first person singular marker *-m.bI < *-n-bI, as in Nanai omɪ- ‘to drink’ : AOR 1SG omɪ-a-m.bɪ < *umï-ra-n-bï. On the other hand, there is no trace of *-n- in the apparently otherwise archaic Nanaic second person singular marker *-tI < *+ti, as in Nanai 2SG omɪ-a-cɪ < *umï-ra-tï, meaning that the reconstruction of the second person singular aorist form remains ambiguous. Assuming that the use of *-n- as an aorist marker in the first and second person singular is an original feature inherited from Proto-Tungusic, it is unclear how the stems taking the aorist markers *-dA(-) or *-sI(-) were treated for this detail in

66 Juha Janhunen

the protolanguage. In Ewenic, both markers are normally present throughout the personal paradigm, but they can be dialectally absent in the first and second person singular, as in Ewenki ga- ‘to take’ : AOR 1SG ga-m (< *ga-n-bï) ~ ga-da-m (< *ga-da-n-bï), bi- ‘to be’ : AOR 2SG bi-nni (< *bi-n-si) ~ bi-si-nni (< *bi-si-n-si). In these cases the person markers 1SG *-n-bI : 2SG *-n-sI imply the presence of the element *-n- also after *-dA(-) and *-sI(-), suggesting that all the forms concerned have undergone some secondary restructuring. As far as the function of the aorist paradigm is concerned, it may originally, as the conventional term implies, have denoted a temporally unmarked type of finite predication, whose actual temporal reference depended on the inherent aspectual content of the verb, with past tense reference being more typical for perfective verbs against present tense reference for imperfective verbs, as is still the synchronic situation in Ewenic. In the other branches, with the increasingly frequent use of secondary finite paradigms based on participles, the aorist has tended to become marginalized. As a result, it has either been totally or almost totally lost, as in Orochic, or it has developed evidential functions, as in Nanaic, where it conveys connotations of visual evidence, mirativity, or assertion. (iii) In the absence of temporal and aspectual distinctions in the aorist paradigm, the principal means for expressing tense and aspect in Proto-Tungusic was offered by participles. At the same time, the aorist stem itself also functioned as a basic participle, indifferent to temporal and aspectual distinctions. •

•

The aorist participle was apparently originally identical with the aorist stem, which, thus, could function as a nominalized verbal form in both adnominal and independent uses, as well as, with the person markers of the predicative type, in the role of the base of the aorist finite paradigm. This means that the aorist finite paradigm is actually not finite, but simply another—older—set of finite forms based on nominalization. However, in the modern languages the nominalizing function of the aorist participle is attested in productive use only in Solon-Orochen and Manchu-‌Sibe, which have generalized the suffix *-rA also after the aorist markers *-dA- and *-sI-, as in Solon ga- ‘to take’ : AOR ga-da- : AOR-PTCP.AOR ga-da-r/a, Spoken Manchu bi- ‘to be’ : AOR *bisi- : PTCP.AOR bisi-le (< *bi-si-re). At the same time, the non-finite use of the aorist participle is well preserved in all branches and languages in the function of the connegative form, that is, the invariable non-finite form that accompanies the inflected forms of the negation verb *e-, as in Ewenki e-si-m buu-re ‘I don’t give’ (NEG-‌AOR-‌‌VX1SG give-CONNEG), Sibe eme yawe-re ‘don’t go!’ (PROHIB go-‌CONNEG). The verbs taking the aorist markers *-dA(-) or *-sI(-) show these markers also in connegative use, as in Ewenki e-si-m bi-si ‘I am not’ (NEG-‌AOR-‌‌VX1SG be-CONNEG), although there is, again, a tendency to add the historically pleonastic suffix *-rA(-), as in Ulcha eji di-de ~ eji di-de-re ‘do not come!’ (PROHIB come-CONNEG-CONNEG). The aorist participle also served as the basis for what is conventionally known as the imperfective participle, marked by the suffix *-RII = *-rII (> *-tII ~ *-dII), *-dII, *-sII, depending on the verb class, e.g. *saa- ‘to know’ : PTCP. IMPRF *saa-rïï (> Ewenki saa-rii, Ulcha saa-rɪɪ > saa-rɪ), *ga- ‘to take’ : PRTCP.IMPRF *ga-dïï (> Ewenki ga-dii, Ulcha ga-dɪ/ɪ), *bi- ‘to be’ : PRTCP. IMPRF *bi-sii (> Ewenki bi-sii, in other branches replaced by restructured

Proto-Tungusic 67

•

•

forms). Formally, the marker *-RII goes back to *-RA-(y)I, which must be a derivative in *-(y)I of the aorist participle. Attested in all branches except Jurchenic, the imperfective participle is widely used, in combination with the person markers of the possessive set, to form a finite paradigm with a present tense reference, as in Nanai ice- ‘to see’ : PTCP.IMPRF ice-i (< *ice-rii) : PRS 1SG ice-i-yi (< *ice-rii-bi) : 2SG ice-i-si (< *ice-rii-si) : 3SG ice-i-ni (< *icerii-ni), etc. As a participle, it is attested in the roles of actor noun, action noun, and adnominal modifier. As the perfective counterpart of the imperfective participle, several different forms are attested. The productive perfective participle marker in Ewenic is *-cAA. This marker yields nominalizations with a full range of participial functions, including finite use to indicate past tense, as in Neghidal baxa- ‘to find’ : PTCP.PRF baxa-caa : PST 1SG baxa-caa-w, etc. In the other branches, *-cAA is attested only in more marginalized uses. In Uilta (> -tA-) it is included in the composition of a complex past evidential marker, as in PST EVID 1SG-EXT ƞene-te-mi-ge (< *ƞene-cee-mi+): 2SG-EXT ƞene-te-​si-ge (< *ƞene-cee-si+) : 3SG.EXT ƞene-te-e (< * ƞene-cee+) etc., while in Nanai (> -cA-) it is used as the so-called preterite marker for a small selection of verbs, including ga- ‘to take’, ji- ‘to come’, o- ‘to become’, and bi- ‘to be’, as in ga- ‘to take’ : PRT 1SG ga-ca-yɪ (< *ga-caa-bï) : 2SG ga-ca-sɪ (< *ga-caa-sï) : 3SG ga-ca (< *ga-caa), etc. In the preterite function, which also involves evidentiality, other verbs in Nanai take the marker *-kA-, as in omɪ- ‘to drink’ : PRT 1SG omɪ-ka-yɪ : 2SG omɪ-ka-sɪ : 3SG omɪ-ka, etc., which is also attested in Udihe, as in jawa- ‘to take’ : PRT-1SG jawa-’a-i (< *jaba-ka-bï). The original form and status of the marker *-kA- remain unclear, though it is possibly a cognate of the Ewenic probabilitative participle in *-rkA-, which also has a past tense reference, as in Ewenki ga- ‘to take’ : PTCP.PROB ga-rka ‘(one) who has probably taken’. If this is so, *-cAA and *-(r)kA were both originally participle markers of the perfective type and/or past tense range, and both seem to have had the potential of acquiring evidential connotations. Another set of perfective/past tense participles is present in Nanaic and Orochic, in which three seemingly different suffixes, distributed depending on the verb class and stem type, are used for this function: -xA/n for general vowel stems, -kI/n for general consonant stems, and -cI/n for the verbs ga- ‘to take’, ji- ‘to come’, o- ‘to become’, and bi- ‘to be’, as in Nanai saa- ‘to know’ : PTCP.PRF saa-xa/n, xool- ‘to sound’ : PTCP.PRF xool-kɪ/n, bi- ‘to be’ : PTCP. PRF bi-ci/n. All these forms can be used in both non-finite and finite functions, with the finite paradigm taking the person markers of the possessive set except in the first person singular, which has -mbI (~ m-bI), as in Nanai sɪa- ‘to eat’ : PTCP.PRF sɪa-xa/n : PST 1SG sɪa-xa-mbɪ (~ sɪa-xam-bɪ) : 2SG sɪa-xa-sɪ : 3SG sɪa-xa-nɪ, etc. The suffix -xA/n has a cognate in the Manchu perfective participle and past tense marker in -xA (-ha ~ -he ~ -ho) ~ -kA, as in sa- ‘to know’ : PTCP.PRF sa-ha (sa-xa). As a possible protoform of this suffix, *-ksAhas been proposed (Benzing), in which case this would be a cognate of the Ewenic anterior converb in *-ksA/A > -ksA/A ~ -xA/A, which itself is originally a nominalized form, taking also the plural marker, as in Ewenki ice- ‘to see’ : CV.ANT ice-kse : PL ice-kse-l ‘after having seen’. Even so, in the absence of confirming evidence, it cannot be taken for certain that the development *ks >

68 Juha Janhunen

•

x would have been regular in all the languages and dialects concerned, leaving open the possibility of secondary contact influence. Future tense reference is expressed in all branches of Tungusic with the exception of Jurchenic by two alternative means: the regular aorist paradigm of the progressive form in -jA- (the so-called simple future), or the finitely used futuritive participle in *-ƞAA(-). These means are also combined by adding the futuritive participle marker to the progressive form, yielding the compound suffix *-jA-ƞAA(-) (the so-called complex future). The plain use of the futuritive participle has, however, developed modal connotations, as in the Nanai optative, in which it is combined with the particle =tAnII, as in sɪa- ‘to eat’ : OPT 2SG sɪa-ƞa-sɪ=tanɪɪ ‘please eat!’ (< *sïa-ƞa-sï=). Therefore, the future participle and the corresponding finite paradigm normally occur in the complex form *-jA-ƞAA(-), which in several languages and dialects is represented as -jA-gAA(-) > -jAA(-), as in Oroch eme- ‘to come’ : PTCP.FUT eme-jee (< *eme-je-ƞee) : FUT 1SG eme-jee-mi ~ eme-jee-yi, etc.

(iv) Converbs are a particularly diversified category of verbal forms in Tungusic, for which reason they present considerable challenges for both synchronic and diachronic analysis. This is especially because in each given language new members have been constantly entering the system of converbs from the category of quasiconverbs, that is, converbally used case forms of nominalized verbs. This is also the reason why the borderline between converbs and participles is often difficult to draw. Although basically invariable, many converbs retain some morphology as a memory of their origin. At the same time, the phonetic development of converb markers often shows unexplained irregularities. •

•

The most widely attested and functionally most neutral converbal form of the conjunct (same-subject) type, here termed the “connective” converb (‘when, while’), is marked by the suffix *-mII (> Ewenic -mII ~ -mI, Orochic -mI, Nanaic -mI, Manchu -me). Because of its frequency, this converb has occasionally been used as the dictionary form of verbs, which is probably the reason why it has also been called the “infinitive”. In Manchu, this form, in combination with the auxiliary *bi- ‘to be’, has yielded the complex present tense marker -mbi (< *-me+‌bi). Several languages also have the corresponding plural form *-mArI (> Neghidal -mAy, Oroch -mAyi, Udihe -mei ~ -mu, NanaiUlcha -mArI ~ -mAArI, Uilta -mAri). As can be seen, there is a formal analogy with the reflexive markers SG *-bII : PL *-bArI, which allows the converb markers to be reconstructed as SG *-mA-(y)i : PL *-mA-r-I. It may be assumed that the converb was formed by the suffixation of the reflexive markers to a verbal base extended by a nominalizing element containing a nasal. A formal analogy of the connective converb is the “consecutive” converb (‘after’), marked by the suffix SG *-pII (> Nanai-Ulcha -pI, Manchu -fi) : PL *-pAri (> Nanai-Ulcha -pArI). In Uilta, the exceptional extended forms SG -p.ie : PL -pi.ssA.A are used, which seem to be parallelled by Oroch SG -wi : PL -wi-sA, perhaps containing the plural suffix *-sA-l. Although absent in Ewenic, this converb obviously dates back to Proto-Tungusic, where it originally had the compound markers SG *-pA-(y)I : PL *-pA-r-I, based on the suffixation of the reflexive markers to a nominalizing element containing the consonant *p.

Proto-Tungusic 69

•

•

•

•

The disjunct (different-subject) counterpart of the connective and consecutive converbs with a general temporal-conditional reference was formed by the “aorist” converb (‘when, if, after’), marked by the complex suffix *-RA-kI-, to which the possessive suffixes were attached to mark the person. This is a quasiconverbal combination of the aorist participle with the primary ablative case marker *-kI. The form of the aorist marker in the complex is determined by the class and stem type of the verb, as in Ewenki ga- ‘to take’ : CV.AOR-2SG ga-da-k.i-s (< *ga-da-‌kï-sï), bi- ‘to be’ : CV.AOR-3PL bi-si-‌k.i-tin (< *bi-siki-tin). The aorist converb is attested in Northern Tungusic (Ewenki-Orochen -RAk.i-, Solon -k.kI-, Neghidal -yA-ki-, Ewen -RA-k-, Oroch -(RA)-ki-), as well as in Uilta (-RA-(y)i-), but not in Nanai or Ulcha. In Nanai and Ulcha there is, however, a related aorist-based form with the marker *-RAA, here termed the “anterior” converb, as in Nanai japa- ‘to take’ : CV.ANT japa-raa. With verbs taking *-dA(-) in the aorist this marker is synchronically added to the aorist stem, as in Nanai ga- ‘to take’ : AOR ga-da : CV.ANT ga-da-raa. There is a possibility that *-RAA is actually a direct reflex of *-RA-kI, with the regular loss of the intervocalic *k. It should be noted, however, that *-RAA is only used in the conjunct function and is not marked for person. A further complication is that *-RAA is also attested in Oroch (-RAA ~ -AA) and Neghidal (-yAAn). These could be morphological borrowings from Nanai-Ulcha, but the Neghidal form suggests the presence of a final nasal, perhaps identical with the singular third person possessive suffix *-n, which is absent in the other languages. Another person-marked form is the “purposive” converb or “supine” in -dAA> -dA- (‘in order to’), which can take both the possessive suffixes for disjunct use and the reflexive suffixes for conjunct use, as in Ewenki baka- ‘to find’ : CV.PURP PX1SG baka-daa-w : RX baka-daa-wii. Attested in all Ewenic languages, this form is also used as the base for the finite paradigm of the so-called distal or remote future imperative. It is unclear whether there is a formal connection, and what exactly, with the Ulcha converb in -bdA-, which also has a purposive function. A certainly related form is, however, the terminative converb in *-dAAlAA(-) > *-dAlAA(-) (‘until’), which is attested both in Ewenic (-dAlAA) and Nanaic (Nanai-Ulcha -dAlA, Uilta -dAlAA). This is again an obvious quasiconverb composed of a nominalized base in *-dAA and the locative case marker *-lAA, to which in some languages (as in Ewen) person markers can also be attached. A similar function in Manchu is filled by the suffix -tOlO (-tala ~ -tele ~ -tolo), which, however, is a borrowing from Mongolic and totally unrelated to the marker attested elsewhere in Tungusic. A functionally specialized form is the “contemporal” converb in *-ƞAsII-, which can be combined with both possessive and reflexive markers and refers to simultaneous past action (‘when in the past’), as attested in Ewenki (-ƞAsii-), Neghidal (-ƞAsA-), Ewen (-ƞsI-), Oroch (-ƞAsA-), and Uilta (-ƞAssie-). This form may also be based on a nominalization, but like the purposive converb in *-dAA- it does not seem to contain a case marker, leaving open the question concerning its exact diachronic structure. There are several other converbs with a more restricted distribution and often with somewhat diffuse or variable formal and functional profiles, suggesting a recently completed or still ongoing process of grammaticalization. The Ewenic languages share, for instance, a “successive” converb (‘after’) in *-mdIn : PL -mdI-r

70 Juha Janhunen

(> Ewenki -mnin : -mni-r, Neghidal -mdin, Ewen -mnIn vs. Orochen -mnA), a “simultaneous” converb (‘while’) in *-nAkAn : PL *-nAkA-r (> Orochen -nAkAn, Neghidal -nAkAn vs. Ewen -nIkAn : -nIkA-r), and a “limitative” converb (‘until’) in *-k(V)nAAn (> Orochen -ƞnA/n, Neghidal -knAAn, Ewen -kAn). An apparently very basic suffix expressing the function of a simultaneous converb is Ewenki -nA : PL -nA-l, with more marginal attestations elsewhere in Ewenic. As the plural forms show, many of these converbs are actually nominalizations which were originally used as finite predicates, but which were secondarily transferred to the role of converbs. For this reason they contain no case marking, as would be typical of quasiconverbs. An important form shared by the Nanaic languages is the conditional converb with a marker perhaps reconstructable as *-gU-cA- (> Nanai -OcIA-, Ulcha -wUcA-, Uilta -KutA-). This may have an indirect connection with the Manchu conditional converb in -ci (-ci), which is otherwise difficult to relate to the data of the other Tungusic languages. SYNTAX Although it is notoriously difficult to reconstruct the syntax of any prehistorical language, it may be assumed that most of the syntactic features shared synchronically by the modern Tungusic languages were already present in Proto-Tungusic. The relatively large amount of reconstructable shared morphology also allows conclusions to be made concerning, especially, morphosyntax. The most important division line within Tungusic in this respect goes between Jurchenic and the other branches, with Jurchenic being relatively more analytic and the other branches relatively more synthetic as far as the expression of grammatical relations is concerned. It is possible that the role of synthesis has grown after the Proto-Tungusic stage in some branches, notably Ewenic, but, more obviously, synthetic constructions have been lost in Jurchenic, leaving only occasional traces. Following is a brief list of some of the syntactic properties that may be assumed to have been present in Proto-Tungusic: •

•

•

•

Word order: The basic word order is likely to have followed the general “Ural-Altaic” verb-final (SOV) pattern, with the nominal phrase also organized on the head-final principle (GAN). At the same time, the presence of morphological means for case marking and person marking may have allowed some freedom in this respect. Agreement: The principal agreement feature was that the subject person was indexed in the predicate by the predicative person markers, making the overt use of the personal pronouns unnecessary in most situations. Similarly, the person of the possessor was indexed on the nominal headword by the possessive and reflexive suffixes. Some modern languages of the Ewenic branch also show occasional agreement in case and/or number between a nominal headword and its modifiers, but this is not likely to have been a feature of the protolanguage. Copula: In equative sentences, the use of the copula *bi- may have been facultative in situations unmarked for tense, aspect, mood, or person, while otherwise the copula was used. The same basic copula was apparently also used in locative and existential sentences, in which its presence was probably obligatory in all situations. Finitization: There are no traces of unmarked finite predicates. All verbal forms used in the finite function seem to have been based on nominalizations. The nominalized

Proto-Tungusic 71

•

•

•

•

predicates represented, however, two separate chronological stages, in that the aorist participle(s) had undergone finitization earlier than the other nominalized forms, as is evident from the two sets of person markers that can be reconstructed for Proto-Tungusic. The other nominalized forms, which in finite use are combined with the possessive suffixes may suggest that the relationship of the subject to the predicate was conceived of as “possessive”. Even so, the overt subject of finite predicates was always unmarked. Compound predicates: The finitized nominalizations could also be seen as nominal predicates linked to the subject without a copula. If a further temporal, aspectual, or modal specification was required, compound forms could be built with the proper forms of the copula *bi-, as is still common in the modern Tungusic languages. Negation: For the negation of both finite and non-finite predicates, the negation verb *e- was used, with the semantic main verb placed in the invariant connegative form, identical with the aorist participle. Originally the conjugated negation verb preceded the main verb, though secondarily the order has been reversed in some modern languages (as in Nanai), with the negation verb being transferred into a clitic or suffix. For prohibition, the idiosyncratic form *e-ji was used. Existential negation may have taken place with the help of negative existential nouns, though the forms attested in this function in the modern languages vary and cannot necessarily be traced back to Proto-Tungusic. Interrogation: Polar questions were probably marked by a clitic, most typically placed on the predicate. Oroch, Nanai-Ulcha, and Manchu suggest a clitic with the shape *=nV/V (> Oroch =nuu, Nanai-Ulcha =nOO, Sibe =na), while Ewenki-Solon, Udihe, and Uilta suggest *-gV/V (> Ewenki-Ewen =GU, Solon =gI/I, Udihe =ye’u, Uilta =i ~ =yi). Some modern languages (like Ulcha) also have a special “corrogative” clitic marking non-polar (content) questions, but this feature is unlikely to derive from the protolanguage. Complex sentences: Clause chaining took place with the help of converbs, including quasiconverbs and converbally used participles, which marked subordinate clauses. The finite aorist paradigm and finitized participles were only used in the final clause of the sequence which ended the complex sentence.

LEXICAL STRUCTURE Due to their relatively close mutual relationship, it is not surprising that the Tungusic languages show considerable agreement in their lexical resources. No comprehensive study concerning the proportions of shared lexicon in all the individual Tungusic languages has been made, but it is obvious that the rate of shared items in the basic lexicon is generally very high within each branch and only slightly lower between the branches. For instance, according to preliminary calculations (Novgorodov), based on the Leipzig-Jakarta List, more than 90 per cent of the basic lexicon of Ewen have cognates in the other Ewenic languages, while around 60 per cent are shared with one or more other branches of Tungusic. The shared items typically cover important subsets such as spatials, numerals, pronouns, and other auxiliaries, but also a large number of basic nominal and verbal roots. In this context it may be noted that the validity of any calculations made about the lexicon shared by the Tungusic languages is potentially decreased by the possibility of secondary mutual borrowing between the individual branches and languages, as well as by the presence of parallel borrowings from third sources, especially from Mongolic

72 Juha Janhunen

and Yakut. At the same time, it is relevant to observe that even some items that have been dismissed as mismatches in the past are actually more or less regular reflexes of Proto-Tungusic originals, two examples being Ewen mɪa/n ‘ten’, which derives by dissimilation from *ñïa/n < *jaa/n < *ju(w)a/n and is a direct cognate of Ewenki jaa/n id., and Manchu niyaman (= ñama/n) ‘heart’, which is also a dissimilative development of *mïama/n and a cognate of Ewenki miewa/n, both deriving from the original shape *mï(y)aba/n. As far as cultural vocabulary is concerned, there is considerable regional diversity between the speakers of different adaptations, specialized variously to environments such as steppe, forest, river, and seaside, and subsistence strategies such as cattle breeding, hunting, fishing, sea mammal hunting, and even, in the Manchurian sphere, gardening and small-scale agriculture. There is no doubt that isoglosses in cultural vocabulary are a potential source of information concerning the mutual relations of the Tungusic languages. To take just one example, the item for ‘domestic reindeer’ may technically be reconstructed as Proto-Tungusic *ora/n, as attested in Ewenki (oro/n), Neghidal (oyo/n), Ewen (ora/n), Oroch (oro), Udihe (oro ~ olo), Nanai (oro/n), Ulcha (oro/n), and Manchu (oron), but apparently not in Orochen or Solon. However, considering that only the speakers of Ewenki, Ewen and Neghidal are known to have been actively engaged in reindeer breeding, it is possible that the speakers of the other languages have adopted this word by internal borrowing. At the same time, the speakers of Uilta, whose economic adaptation does comprise reindeer breeding, seem to have lost this word, replacing it by ulaa < *ulaa/g, an obscure item otherwise attested only in Oroch (ula), though possibly connected with Manchu ulha ‘domestic animal’ (ulha < *ulaga). All of this is relevant to the reconstruction of the history of reindeer breeding, which was certainly not yet practised by the speakers of Proto-Tungusic. Most importantly, cultural vocabulary, when properly analysed and reconstructed, has relevance to the determination of the location of the original linguistic homeland of Tungusic and the subsequent routes of language spread, as well as the resulting contacts with other languages and language families. REFERENCES AND FURTHER READING Alonso de la Fuente, José Andrés (2011) Tense, Voice and Aktionsart in Tungusic: Another Case of “Analysis to Synthesis?”, Tunguso-Sibirica 31, Wiesbaden: Harrassowitz Verlag. Alonso de la Fuente, José Andrés (2012) ‘The first imperative of Tungusic’, Studia Universitatis Iagellonicae Cracoviensis 129: 7–34, Kraków. Alonso de la Fuente, José Andrés (2018) ‘Past tenses, diminutives and expressive palatalization: Typology and the limits of internal reconstruction in Tungusic’, in: Ákos Bertalan Apatóczky & Christopher P. Atwood & Béla Kempf (eds.), Philology of the Grasslands: Essays in Mongolic, Turkic, and Tungusic Studies, 112–137, Leiden: Brill. Ard, Joshua (1981) ‘A sketch of vowel harmony in the Tungus languages’, in: Bernard Comrie (ed.), Studies in the Languages of the USSR, Cambridge Language Surveys, 23–43, Cambridge: Cambridge University Press. Barrere, Ian & Juha Janhunen (2019) ‘Mongolian vowel harmony in a Eurasian context’, International Journal of Eurasian Linguistics 1 (1): 46–77, Leiden. Benzing, Johannes (1956) Die tungusischen Sprachen: Versuch einer vergleichenden Grammatik, Akademie der Wissenschaften und der Literatur in Mainz, Abhandlungen

Proto-Tungusic 73

der geistes- und sozialwissenschaftlichen Klasse 1955 (11), Wiesbaden: Franz Steiner Verlag. Boldyrev, B V. [Б. В. Болдырев] (1987) Словообразование имен существительных в тунгусо-маньчжурских языках в сравнительно-историческом освещении [The derivation of nouns in the Tungusic languages in a comparative-historical framework], Новосибирск [Novosibirsk]: “Наука”. Cincius, V. I. [В. И. Цинциус] (1949) Сравнительная фонетика тунгусо-маньчжурских языков [Comparative phonetics of the Tungusic languages], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство Министерства просвещения РСФСР. Cincius, V. I. [В. И. Цинциус] (ed.) (1975–1977) Сравнительный словарь тунгусо-​ маньчжурских языков: Материалы к этимологическому словарю [Comparative diactionary of the Tungusic languages: Materials for an etymological dictionary], vols. 1–2, Ленинград [Leningrad]: “Наука”. Doerfer, Gerhard (1967) ‘Urtungusisch *-ï , *-i, *-u, *-ü im Evenkischen’, Zeitschrift der Deutschen morgenländischen Gesellschaft 117 (1): 103–118, Wiesbaden. Doerfer, Gerhard (1978) ‘Urtungusisch ö’, in: Gerhard Doerfer & Michael Weiers (eds.), Beiträge zur nordasiatischen Kulturgeschichte, Tungusica 1: 66–116, Wiesbaden: Otto Harrassowitz. Hesche, Wolfgang (1978) ‘Urtungusisch *ö in erster Silbe (= *ö1) im Lamutischen’, in: Gerhard Doerfer & Michael Weiers (eds.), Beiträge zur nordasiatischen Kulturgeschichte, Tungusica 1: 117–125, Wiesbaden: Otto Harrassowitz. Ikegami, Jirō (1985) ‘The category of person in Tungus: Its representation in the indicative forms of verbs’, 『言語研究』 Journal of the Linguistic Society of Japan 88: 86–96, 京都 [Kyoto]. Revised version published in: Ikegami Jirō 池上二良 (2001) 『ツングース語研究』   [Researches on the Tungus Language], 369–379, 東京 [Tokyo]: 汲古書院. Ikegami, Jirō (1995) ‘The element -n in the indicative forms of verbs in Tungus languages’, 『言語研究』 Journal of the Linguistic Society of Japan 107: 1–15, 京都 [Kyoto]. Revised version published in: Ikegami Jirō 池上二良 (2001) 『ツングース 語研究』 [Researches on the Tungus Language], 380–394, 東京 [Tokyo]: 汲古書院. Joseph, Andrew & Ko Seyongyong & John Whitman (2020) ‘A comparative approach to the vowel systems and harmonies in the Transeurasian languages and beyond’, in: Martine Robbeets & Alexander Savelyev (eds.), The Oxford Guide to the Transeurasian Languages, Oxford Guides to the World’s Languages, 486–508, Oxford: Oxford University Press. Kazama, Shinjirō 風間伸次郎 (1997a) 「ツングース諸語における「部分格」」 [The partitive case in the Tungusic languages],『環北太平洋の言語』Languages of the North Pacific Rim 3: 103–120, Kyoto: Graduate School of Letters, Kyoto University. Kazama, Shinjirō 風間伸次郎 (1997b)「ツングース語の方位名称について」[The terms of direction in the Tungus languages],『北海道立北方民族博物館研究紀要』 Bulletin of the Hokkaido Museum of Northern Peoples 6: 113–124, 網走 [Abashiri]: 北海道立北方民族博物館. Kazama, Shinjirō (2004) ‘On the “causative” forms in Tungus languages’, in: Carsten Naeher (ed.) (2004) Proceedings of the First International Conference on Manchu-Tungus Studies (Bonn 2000), vol. 2. Trends in Tungusic and Siberian Linguistics, Tunguso-Sibirica 9 (2004), Wiesbaden: Harrassowitz Verlag, 73–92.

74 Juha Janhunen

Kazama, Shinjirō 風間伸次郎 (2007a)「ツングース諸語の三人称代名詞につい て」[Third person pronouns in Tungusic languages], 『語学教育フォーラム』 13: 173–184, 東京 [Tokyo]: 大東文化大学語学教育研究所. Kazama, Shinjirō 風間伸次郎 (2007b)「ナーナイ語とウデヘ語の付属語につい て」[Clitics in Nanai and Udihe], 『アジア・アフリカの言語と言語学』2: 49–83, 東京 [Tokyo]: 東京外国語大学アジア・アフリカ言語文化研究所. Kazama, Shinjirō (2008a) ‘Alienable possession suffixes in Tungusic languages’, Linguistic Typology of the North, 1: 125–140, Tokyo: Tokyo University of Foreign Studies, Research Institute for Languages and Cultures of Asia and Africa (ILCAA). Kazama, Shinjirō (2008b) ‘‘Third’ person pronouns in Udihe and Nanai’, Linguistic Typology of the North, 1: 141–153, Tokyo: Tokyo University of Foreign Studies, Research Institute for Languages and Cultures of Asia and Africa (ILCAA). Kazama, Shinjiro (2008c) ‘The diachronic development of group III of Tungusic languages’, Linguistic Typology of the North, 1: 103–124, Tokyo: Tokyo University of Foreign Studies, Research Institute for Languages and Cultures of Asia and Africa (ILCAA). Kazama, Shinjirō 風間伸次郎 (2009)「ツングース諸語の受身」 [Passive in Tungusic languages], 『語学研究所論集』14: 81–96, 東京 [Tokyo]: 東京外国語大学語 学研究所. Kazama, Shinjirō (2012) ‘Designative case in Tungusic languages’, in: Andrej L. Malchukov & Lindsay J. Whaley (eds.), Recent Advances in Tungusic Linguistics, Turcologica 89: 123–152, Wiesbaden: Harrassowitz Verlag. Kazama, Shinjirō (2016)「ツングース諸語において祖形 -ks- が仮定される音対応 について」[On the phonetic correspondences of *-ks- in Tungusic languages], in:『 語学研究所論集』[Journal of the Institute of Language Research], 21: 1–14, 東京 [Tokyo]: 東京外国語大学 [Tokyo University of Foreign Studies]. Ko, Seongyeon & Andrew Joseph & John Whitman (2014) ‘Comparative consequences of the tongue root harmony analysis for proto-Tungusic, proto-Mongolic, and proto-Korean’, in: Martine Robbeets & Walter Bisang (eds.), Paradigm Change: In the Transeurasian Languages and Beyond, Studies in Language Companion Series 161: 141–176, Amsterdam: Benjamins. Li, Bing (1996) Tungusic Vowel Harmony: Description and Analysis, Academisch Proefschrift, Universiteit van Amsterdam, Holland Institute of Generative Linguistics (HIL) Dissertations 18, Amsterdam. Missonova, L. I. [Л. И. Миссонова] (2013) Лексика уйльта как историко-​ этнографический источник [Uilta lexicon as a source on history and ethnography], Москва [Moscow]: “Наука”. Näher, Carsten (1999) ‘Der urtungusische stimmlose velare Plosiv im Mandschu’, Journal de la Société Finno-Ougrienne 88: 113–130, Helsinki. Norman, Jerry (1977) ‘The evolution of Proto-Tungusic *t to Manchu s’, Central Asiatic Journal 21 (3/4): 229–233. Novgorodov, I. N. [Иннокентий Николаевич Новгородов] (2021) ‘Устойчивый словарный фонд в генетической классификации на примере восточного наречия эвенского языка’ [Stable vocabulary in genetic classification, on the example of the eastern dialect of Ewen], Crede Experto 2 (29), Irkutsk: http://ce.if-mstuca.ru. Novgorodov, I. N. [Иннокентий Николаевич Новгородов] & L. Zh. Zaksor [Любовь Жоржевна Заксор] (2021) ‘Нанайский язык в генетической классификации’ [The

Proto-Tungusic 75

Nanai language in the genetic classification], Crede Experto 3 (30), Irkutsk: http:// ce.if-mstuca.ru. Ryzhkov-Shukumine, Alexandre (2020) ‘Long vowels in Proto-Tungusic’, International Journal of Eurasian Linguistics 2 (1): 70–93, Leiden. Starostin, S. A. [С. А. Старостин] (1991) Алтайская проблема и происхождение японского языка [The Altaic problem and the origin of the Japanese language]. Москва (Moscow): “Наука”. Sunik, O. P. [О. П. Суник] (1947) Очерки по синтаксису тунгусо-маньчжурских языков [Essays on the syntax of the Tungusic languages], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство Министерства просвещения РСФСР. Sunik, O. P. [О. П. Суник] (1962) Глагол в тунгусо-маньчжурских языках: Морфологическая структура и система форм глагольного слова [The verb in the Tungusic languages: The morphological structure and system of forms of the verbal word], Москва [Moscow] & Ленинград [Leningrad]: Издательство Академии Наук СССР. Sunik, O. P. [О. П. Суник] (1968) ‘Тунгусо-маньчжурские языки: Введение’ [The Tungusic languages: Introduction], in: Монгольские, тунгусо-маньчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 53–67, Ленинград [Leningrad]: “Наука”. Sunik, O. P. [О. П. Суник] (1982) Существительное в тунгусо-маньчжурских языках: В сравнении с другими алтайскими языками [The noun in the Tungusic languages: In comparison with the other Altaic languages], Ленинград [Leningrad]: “Наука”. Sunik, O. P. [О. П. Суник] (1997) ‘Тунгусо-маньчжурские языки’ [The Tungusic languages], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 153–162, Москва [Moscow]: “Индрик”. Vovin, Alexander (1997) ‘Voiceless velars in Manchu’, Journal de la Société FinnoOugrienne 87: 263–280, Helsinki.

CHAPTER 5

JURCHEN Daniel Kane and Marc Hideo Miyake

Jurchen is the Tungusic language once spoken by the historical Jurchen, a mediaeval ethnic group in southern Manchuria who in 1115 established a regional empire known in Chinese as the Great Jin (大金 Da Jin) dynasty, which fell as a result of Mongol invasions in 1234. As a Manchurian polity, the Jin dynasty succeeded the Liao (遼) dynasty (907– 1125), which had been founded by the Para-Mongolic-speaking Khitan, and which had comprised also substantial parts of the Jurchen ethnic territory. As a language, Jurchen may be seen as a more or less direct ancestor of Manchu, and the ethnonym “Jurchen”, in the Manchu form jušen, was still used as the endonym of the historical Manchu until the introduction of the new name “Manchu” in 1635. In the shape 女眞 Nüzhen ~ Ruzhen *[ʐytʂin], or alternatively, allegedly for reasons of secondary naming taboo, 女直 Nüzhi ~ Ruzhi, it was also the name by which the Jurchen were known to the contemporary Chinese. The modern international spelling “Jurchen” (also Jürchen) is based on the Mongolic form *jürcin ~ *jörcin (PL jürci-d ~ jörci-d). As a regional ethnonym, “Jurchen” may or may not be connected with several earlier ethnic and tribal names of the Manchurian sphere, including Sushen (肅愼), known from Chinese sources since the latter half of the last millennium BZ, and Chaoxian (朝鮮), attested since the last centuries BZ and later used to refer to Korea and the Koreans. The Jurchen ethnic territory was originally centered on the Liaodong region, extending also to the northern parts of the Korean peninsula in the south and the Sungari basin in the north, but at the height of its power the Jin empire comprised large parts of northern China, as well, with capital cities spread between today’s Acheng (Upper), Liaoyang (Eastern), Daming (Northern), Datong (Western), Peking (Central), and Kaifeng (Southern). Demographically, the ethnic Jurchen constituted only a small minority, perhaps some 10 per cent—though still several million people—of the total population of the Jin empire. However, acculturation, ethnic mixing, and migrations led to the rapid decline of the Jurchen ethnicity and language, leaving intact only the northern core population from whose context the Manchu subsequently emerged. On the Korean Peninsula, the expansion of the Korean language and ethnicity gradually pushed the southern boundary of Jurchen northwards, a process completed only during the Joseon (朝鮮) dynasty of Korea (1392–1910). During the Ming dynasty of China (1368–1644), the northern groups of the Jurchen, which were only secondarily involved in the consolidation of the Manchu, were known as “Wild” or “Savage” Jurchen (野人女眞 Yeren Nüzhen). At the time of the arrival of the Russians on the Amur, a group of these Jurchen, identified as “Jucher”, lived, together with the linguistic ancestors of the Daghur and Solon, in the Middle Amur basin, from where they were later transferred to central Manchuria. The pre-dynastic Jurchen had no script, but the educated classes were literate in Khitan and Chinese, and it cannot be ruled out that a local derivative of the Chinese script, which DOI: 10.4324/9781315728391-5

Jurchen 77

underlay the Khitan “Large Script”, was occasionally also used to write Jurchen. Possibly based on this tradition, the Jurchen founding emperor ordered in 1120 that a script be devised for writing Jurchen. This script came to be known as the Jurchen “Large Script” (女眞大字 Nüzhen dazi), followed in 1138 by a variety termed the Jurchen “Small Script” (女眞小字 Nüzhen xiaozi). During the initial period of the Jin dynasty a significant number of books, including translations from Chinese, were written and published in Jurchen. Textbooks were produced and schools established, but rapid Sinicisation led to a decline in interest. A dozen or so of the books published in Jurchen were still extant during the Yuan period (1279–1368), but none seem to have survived into the Ming. Even so, the Jurchen script remained in limited marginal use until the adoption of the Mongol script in 1599. Unfortunately, unlike the Khitan, the Jurchen did not use their native language in personal epitaphs, for which reason only a small number of memorial stelae in Jurchen have been preserved. In addition, important samples of the Jurchen language and script are present in the bilingual Sino-Jurchen vocabularies prepared by the “Bureau of Translators” (四夷館 Siyiguan) and the “Bureau of Interpreters” (會同館 Huitongguan) of the Ming administration. DATA AND SOURCES Jurchen is the first recorded Tungusic language, but interpretation of the records is very problematic. The exact meaning of the terms “Large Script” and “Small Script” is unclear. There is only one style of Jurchen characters in the extant inscriptions, which is a mixture of logographic and syllabic characters. The inscriptions are very few, and have not been well preserved. As of today, only 12 major inscriptions dating from the period between 1138 and 1413 are known, as well as a small number of shorter texts and graffiti written or engraved on rocks, seals, mirrors, travel passes, or other objects, occasionally in a cursive variety of the script. Two inscriptions that are fairly long are the “Jin Victory Stele” (大金 得勝陀頌碑 Da Jin desheng tuosong bei) of 1185, as discussed by Tamura Jitsuzō (1976) and Liu Fengzhu & Yu Baolin (1981), and the “Presented Scholars Stele” (女眞進士題名 碑 Nüzhen jinshi timing bei) of 1224, both of which, despite their eroded condition, can provide some information on Jin Jurchen syntax. There are also Chinese versions available of these two inscriptions. Another important inscription, the only one from the Ming period, is the trilingual Jurchen-Mongol-Chinese “Tyr Stele” of 1413, originally erected at the Jurchen outpost of Nurgan on the Lower Amur to mark the site of a temple. After some earlier studies by, among others, Osada Natsuki (1958) and Louis Ligeti (1961), an up-to-date publication of it was prepared by Golovachev & al. (2011), who also analyse the parallel Chinese and Mongol texts. The inscriptions would not be even partially decipherable without the Sino-Jurchen Vocabulary of the Bureau of Translators (女眞館譯語 Nüzhen guan yiyu), conventionally dated from the Yongle (永樂) period (1403–1435). Its entries are given in Jurchen script and Chinese transcription (which must be interpreted in the light of Ming Chinese phonology), with the meaning added in Chinese. It was first published by Wilhelm Grube (1896), who listed the Jurchen entries, translated the Chinese glosses, and transcribed the Chinese phonetic renderings of the Jurchen characters. The publication by Grube has later been superseded by Kiyose (1977) and Daoerji & Hexige (1983). However, although this is in many respects our most important source for the Jurchen language and script, it must be used with great caution. Daoerji & Hexige (1983: 253) estimate that, of the 917 entries, no less than 230 must be amended in one way or another. The vocabulary

78 Daniel Kane and Marc Hideo Miyake

itself clearly has many layers, the earliest dating from the Jin, augmented from time to time with words and expressions from other texts, some of which may have been from genuine Jurchen documents or inscriptions, but many Jurchen terms are simply calques from Chinese and could not have been derived from genuine sources. The Chinese characters used to transcribe Jurchen provide only an approximate reading of the Jurchen characters. It is not known where the phonetic annotations in the vocabulary came from, but it seems that the underlying language is a form of Ming Jurchen, quite similar to the genuine spoken Jurchen recorded in Chinese transcription in the vocabulary of the Bureau of Interpreters. The vocabulary of the Bureau of Translators contains also dated memorials, of which the first one refers to events from 1458 to 1526. It is not known why these word lists and texts were compiled. The most plausible explanation is that bureaucratic procedure required that communications from various non-Chinese tributaries be written in the native language and script. As knowledge of the Jurchen script had apparently become rare among the Jurchen during the Ming period, the task of providing native language translations of the petitions was left to Chinese scribes who simply translated the Chinese texts, word by word, according to the vocabulary lists compiled by previous translators, with no regard for Jurchen word order, nominal and verbal endings being used at random. The resulting texts, which are attached to the glossary, cannot shed any light on Jurchen morphology or syntax. The Vocabulary of the Bureau of Interpreters (女眞譯語 Nüzhen yiyu) was published by Kane (1989), as reviewed by Janhunen (1991) and indexed by Näher (1999). It does not contain any Jurchen characters, but over a thousand words and expressions from spoken Jurchen in Chinese transcription. It is part of a series of such vocabularies of various non-Chinese languages, known as Hua-Yi Yiyu (華夷譯語), and dated to the late fifteenth century. Their purpose was presumably to help with basic communication with tributary delegations. Like the vocabulary of the Bureau of Translators, it paid little attention to Jurchen grammar: for instance, there is not a single accusative ending in the whole work. It is clear, however, that this vocabulary is closer to genuine late Ming Jurchen than the ungrammatical entries of the Bureau of Translators. Several sheets of paper with Jurchen characters in cursive style written on them were discovered in the “Forest of Stelae” (碑林 Beilin) in Xi’an in 1973. They contain many graphs not in the Ming vocabulary, and several in a slightly different form, but give no indication to their pronunciation or meaning. Other information on Jurchen is available in various Chinese texts, as published by Herbert Franke (1997ab, 2000). Most importantly, the official dynastic history of the Jin (金史 Jin shi), written in Chinese, contains a rather extensive glossary of personal names and official titles. The only comprehensive modern lexical source on Jurchen is the dictionary compiled by Jin Qizong (1984), who also published many other important contributions concerning the Jurchen language and script, including a joint study with Jin Guangping (1980). Other studies on the Jurchen language and script, as well as on the individual inscriptions. include those by Sun Bojun (2004) and Aisin Gioro Ulhicun (2002, 2009b), also in coauthorship with Yoshimoto Michimasa (2017). Studies on more specific topics include Kiyose (1997, 1999) on phonology, Burykin (1999) on morphology, Ligeti (1960), Janhunen (1993), Rozycki (1994), and Vovin (2006) on lexical borrowings, as well as Jin (1979), Kiyose (1984), Krippes (1988), Janhunen (1994), Aisin Gioro Ulhicun (2009a), and Vovin (2012) on the script. The taxonomic position of Jurchen and Manchu is discussed in Kiyose (1998) and Vovin (2006). Concerning the ethnonym “Jurchen”, see Janhunen (2004).

Jurchen 79

NOTATIONAL CONVENTIONS To deal with the various types of information available on Jurchen, which often correlate with linguistic differences, it is relevant to distinguish the data (i) by the period, (ii) by the underlying script, and (iii) by the source. To indicate the chronological difference between the data from the Jin and Ming periods, the letters D (= Dynastic Jin) and M (= Ming) are used below. To indicate the type of script from which a given piece of data derives, the letters J (= Jurchen script) and C (= Chinese script) are used. To indicate the difference between the two Ming-period vocabularies of Jurchen, the letters A (= the Bureau of Translators) and B (= the Bureau of Interpreters) are used. This means that Jurchen items will typically be identified by up to three letters, corresponding to the parameters of period, script, and source. Since the phonological reconstruction of individual Jurchen graphs, particularly the logograms, is problematic, the data from sources written in the Jurchen script will be presented below in a tentative transliteration, with logograms written in capital letters and phonograms (or syllabograms) written in lower case letters, and with the boundaries between characters indicated by the hyphen. Data based on Chinese transcriptions of Jurchen items will be given in quasiphonemic “readings”, marked by the dagger sign (†), as deduced from the phonetic values of the Chinese characters and interpreted in the light of attested Manchu cognates. Manchu is quoted in the conventional Romanization (in boldface). THE WRITING SYSTEM Although the Jin shi states that the Jurchen had two scripts, the Large (1120) and the Small (1138), all material written in Jurchen is written in a single script (hereafter, “the Jurchen script”) with a few potential exceptions. It is not clear whether the Jurchen script was the Large Script, the Small Script, or even whether two distinct scripts existed. Normally such a basic issue should be dealt with first, but instead the structure and usage of the Jurchen script will be outlined to provide the foundation for a discussion of the questions of the Large and Small Scripts and their origins. The Jurchen script is undoubtedly “Sinoform”: similar to the Chinese script in appearance while not actually being Chinese. Like Chinese characters, Jurchen characters each fill a square-shaped space and were written in columns from right to left, unlike the later right-to-left vertical script for Jurchen’s Manchu successor. A few Jurchen characters are completely identical to Chinese characters, but the type of correspondence varies: •

•

In some cases, an identical form correlates with an identical meaning but has different readings in the two languages, e.g.  ‘one’ = Jurchen EMU (emu) vs. Jin Chinese *iʔ,  ‘two’ = Jurchen JO (juwe) vs. Jin Chinese *ži,  ‘day’ = Jurchen INENGI (inenggi) vs. Jin Chinese *žiʔ.  ‘moon, month’ = Jurchen BIGA (biya) vs. Jin Chinese *ƞüe. In other cases, there is no correlation of either meaning or sound. Often, such characters function as phonograms in Jurchen, e.g.  ai vs. Jin Chinese *biƞ [third Heavenly Stem],  mei vs. Jin Chinese *yiw ‘right’,  je vs. Jin Chinese ši ‘arrow’,  uru vs. Jin Chinese *šiʔ ‘to lose’,  li vs. Jin Chinese *xiw ‘to rest’, but there are also examples of logographic values, e.g.  GORXON ‘thirteen’ vs. Jin Chinese *kew ‘mouth’.

80 Daniel Kane and Marc Hideo Miyake

Most of the above characters (with the exceptions of ) also have exact lookalikes in the Khitan Large Script and/or Khitan Small Script. Many more Jurchen characters look like Chinese or Khitan Large Script characters with an altered or additional stroke, often a dot. For instance, the Jurchen lookalikes of Chinese and Khitan Large Script ‘one’ and ‘two’ have variants with bent horizontal strokes (), the exact lookalikes of Chinese and Khitan Large Script ‘day’ and ‘month’ have dotted variants (), and  ILAN ‘three’ (ilan) looks like Chinese and Khitan Large Script 三 ‘three’ with a fourth stroke added to the lower right corner. Still other Jurchen characters do not have any obvious Chinese or Khitan counterparts: for instance,  MEDE ‘sea’ (mederi) does not resemble Jin Chinese 海 *xai ‘sea’ or any other Chinese or Khitan Large Script character. Some Jurchen characters have clear Khitan Large Script cognates not shared with Chinese: thus, for instance, Jurchen  ANYA ‘year’ (aniya) and Khitan Large Script  AI ‘year’ are more similar to each other than either is to Jin Chinese 年 *nien ‘year’. All Jurchen characters, regardless of the degree of their similarity to Chinese or Khitan characters, were constructed out of a finite number of stroke types. Nearly all stroke types in Jurchen characters can also be found in Chinese or Khitan characters. For instance, every stroke in  MEDE ‘sea’ is shared with the Chinese and Khitan scripts, even though the way those ten strokes are combined is unique to Jurchen. A rare exception to that rule is an un-Chinese and un-Khitan stroke resembling a left-hand parenthesis appearing in place of the stroke ㇓ as a left-hand component in the characters of the mid-12th century Kyŏngwŏn (慶源) inscription located in what is now North Korea. Jurchen strokes form components which often appear in multiple characters. Since no native analyses of Jurchen characters comparable to those for Chinese or even Tangut characters survive, it is difficult to determine what is a component in the Jurchen script. The following criteria for Jurchen components are admittedly arbitrary and influenced by Chinese conventions. Although they probably do not precisely match the lost native criteria, they will still facilitate a tentative description of Jurchen character structure. •

•

•

•

Large standalone strokes, such as, for instance, ㇓in fifty characters including ​ ​ and others, are generally considered components. Exceptions are recurring combinations of large nonadjacent strokes: for instance, half the characters with ㇟ on the right also contain ㇓ on the left, so the combination ㇓+ ㇟ may be a single component that flanks other components. Hence  BEGI [rank corresponding to Manchu beile] may be analysed as consisting of two components, ㇓+ ㇟ and ⺩. Contiguous and/or intersecting strokes are generally considered to be components: for instance, ⺩is a single component of four strokes rather than a combination of 一 with 土 or 工, etc. In a small set of exceptions, the upper left portion of one component overlaps with the bottom of the component above it: thus,  la is a combination of  and  with a dot. Small standalone strokes such as dots (actually very short strokes like ㇔ rather than circles) are not considered components even though they are not contiguous: e.g., the ホ of characters such as  is a single component with two small standalone strokes, and  are one-component characters despite their dots. The high frequency of dots in the Jurchen script distinguishes it from the Chinese script. About a fifth of all Jurchen characters contain dots. Out of 198 characters on the first face of the Kyŏngwŏn inscription, 59 have dots, including three out of the six characters   AMBA-an ANCU-un GURU-un for ‘Great Jin State’.

Jurchen 81

If dots and other short strokes are excluded as components, Jurchen characters contain one to four components (Table 5.1). TABLE 5.1 COMPONENTS IN JURCHEN CHARACTERS Components

Characters

%

1

582

40

Layout

Examples

2

404

28

⿱

⬚

 

193

13

⿰



81

6

⿸



24

2

⿵



15

1

⿹



14

je- ‘to eat’ : IMP jefu, which retains the final consonant in the imperative, and *di- : AOR di-de- > ji- ‘to come’ : IMP ji-o ~ j-u, which takes the imperative marker directly to the vowel stem. The latter verb is also etymologically present in gaji- ‘to bring’ : IMP gaj-u, from gai- ~ ga- ‘to take’ (as in AND ga-na- ‘to go to take’) + ji- ‘to come’. In the sentence, the basic imperative functions as a finite predicate implying a command directed at the second person, both singular and plural, as in [MYK 1.6–8] ere bira-be wasi-me gene ‘go down this river!’ (this river-ACC descend-CV.CONN go), cf. also [41]. The corresponding negative imperatives are formed by the prohibitive particle ume combined with the aorist participle of the following verb, which in this case functions as a connegative, as in [NSB 3] ume inenggi goida-ra ‘do not delay [for the whole] day’ (PROHIB day delay-CONNEG), [LJB 1: 11b.1–2] arki nure ume omi-re ‘do not drink distilled liquor [or] rice wine’ (distilled.liquor rice.wine PROHIB drink-CONNEG).

Written Manchu 131

Apart from the basic imperative, a command, request, wish, or invitation to action can be expressed by the voluntative (“subjective optative”) in -ki, directed at the first person, as well as by the permissive (“objective optative”) in -kini, directed at the second or third person (and containing a reflex of the third person singular possessive suffix), as in [48–49]: [48]

muse tuba-de dedu-me 1pl.incl that.place-dat spend.the.night-cv.conn ‘Let us go and spend the night there!’ [LGD 1: 218]

[49]

jui son

sim-be

dasa-me banji-kini rule-cv. live-perm conn ‘Son, so that you would live ruling the disorderly country under heaven, 2sg-acc

abka heaven

facuhūn disorderly

yo-ki go-vol

gurun-be country-acc

se-me banji-bu-habi say-cv.conn live-caus-prf I gave birth to you.’ [MYK 1.6] Another modal form of the imperative sphere is the optative (also: “subjunctive”) in -cina, which expresses a wish (‘would’) or obligation (‘should’), as in [TM 7] bi inu bithe hūla-ki se-cina ‘I, too, would say that I would like to read books’ (1SG also book read-VOL say-OPT). The negative counterpart of the optative is the apprehensive (also: “dubitative”) in -rahū, a form based on the aorist participle (containing the unidentified element -hū), which expresses a wish that something should not happen (‘so that it should not’), as in [50]: [50]

taidzu sure beile bata-be sabu-rahū se-me [title] wise [title] enemy-acc perceive-appr say-cv.conn ‘So that he would not see the enemy, the wise Taizu Beile . . . ’ [MYK 2.96]

SYNTAX Manchu syntax follows the general patterns observed in the other Tungusic languages, as well as in the neighbouring Mongolic languages. The basic word order is strictly verb-final (SOV), the only elements that can follow a finite predicate being a few emphatic particles, notably kai, as well as clitics like the interrogative marker =o. In the sentence, the indirect object normally precedes the direct object, and in both the nominal and the verbal phrase a modifier always precedes the headword. There are no prepositions or prefixes, only postpositions and suffixes. Also, there is no verbal or nominal agreement in number, case, or person. Adjectival nominals take case marking only when used as independent headwords. There is a strong tendency for pro-drop and subject-drop, though their conditions and exact mechanisms remain unstudied. There are also a number of conjunctional elements introducing dependent clauses in Manchu. The most frequent among them include anggala ‘rather than’, dabala ‘besides’, jakade ‘because’, onggolo ‘before’, manggi ‘after’, amala ‘after’. These are typically

132 Alexander Vovin

petrified nominal forms of generic or spatial nominals, often containing the local case suffixes of LOC -lO and DAT -de, combined with a participial form of the preceding verb, as in [LJB 2: 72b.6–7] elcin amba-sa ji-dere jakade ‘because there came ambassadors and ministers’ (ambassador official-PL come-PTCP.AOR because = thing-DAT), [MLD I.1] juhe tuhe-re onggolo ‘before the ice melts’ (ice melt-PCTP.AOR before), [LJB 2: 66.7– 67.1] hanci o-ho manggi ‘after [we] have come close’ (close become-PTCP.PRF after). A typologically rare detail is the postverbal “nondumitive” negator unde ‘not yet’, which requires the connegative form of the preceding verb, as in [51]: [51]

goro distant

ba-i place-gen

cooha army

wacihiya-me isi-nji-re unde finish-cv. reach-vennondum conn conneg ‘While the troops from the distant place had not yet completely arrived . . . ’ [MYK 5.364]

LEXICON AND LANGUAGE CONTACTS As a Tungusic language, Manchu has a special position, in that the role of the lexicon inherited from Proto-Tungusic is relatively less significant than in the other Tungusic languages. By contrast, Manchu has borrowed rather extensively from several languages, including, in particular, the neighbouring Mongolic languages, as well as Chinese, but also from various forms and stages of Khitanic and Koreanic, as well as, via literary contacts, Tibetan and Sanskrit. Manchu also contains many items shared with Ghilyak, some of which reflect ancient contacts between Jurchenic and Amuric. Unfortunately, the influence of languages such as Khitan and Old Korean is difficult to appreciate to the full extent because of the extreme paucity of source materials on these languages. Also, Manchu contains many words of unknown origin, and although some of them may turn out to be borrowings from Khitanic or Koreanic, this does certainly not apply to all of them. Below, the lexical layers of Manchu are examined in some more detail. Examples of basic vocabulary items inherited from Proto-Tungusic, and attested also in the Ewenic and Nanaic branches, include ‘blood’ se.nggi (< *see-), ‘bone’ girang.gi (< *gïra.m-), ‘to die’ bu.ce- (< *bö.d-), ‘ear’ šan (< *sïan), ‘eye’ yasa (< PL *ïa-sa-l), ‘to give’ bu- (< *böö-), ‘hand’ gala (< *ƞaala), ‘fire’ tuwa (< *toga), ‘wind’ edun (< *xedün), ‘water’ mu.ke (< *möö). An item shared only with Nanaic is ‘earth’ na (< *naa). Items with possible Korean parallels include biyoran ‘cliff’ ≈ Korean pyerang < Middle Korean pìrèy ~ pìryèy < *pira(ng), cecere- ‘to press tightly’ ≈ Middle Korean cìcírɨ́- ~ cìcɨ́r- < *cìcírɨ́-, fulehe ‘root’ ≈ Middle Korean pùrhúy < *purɨkuy, nitan ‘weak, faded’ ≈ Middle Korean nyèth- < *nitV-h, se- ‘to say’ ≈ Middle Korean hʌ- ~ hʌy- < Old Korean †hyʌ- < *hyʌ-, singkeyen ‘chilly’ ≈ Middle Korean sìk- : sìk-ɨ́n < Old Korean †sik-un < *sink-ɨn. While some of these may involve accidental similarities, there are several cultural words which show a more obvious connection, though the direction and route of borrowing is not always clear: fatan ‘comb of the loom’ ≈ Middle Korean pʌ̀tʌ́y < *pata(n)-i, fisen ‘seed’ ≈ Middle Korean psí < *pisi, fucihi ‘Buddha’ ≈ Old Korean †putɨki > Middle Korean pùthyè, fushe- ‘to fan’ ≈ Middle Korean pùch-. A cultural item shared on a wider areal basis is misun ‘bean paste’ ≈ Middle Korean myècú < Old Korean †micu < *micu ≈ Japanese miso (味噌 = Chinese weizeng).

Written Manchu 133

Mongolic loanwords in Manchu reflect prolonged contact extending from Pre-Proto-Mongolic to Modern Mongolian. Many Mongolic items are also present in the other Tungusic languages, but they are often due to parallel borrowing, or to secondary borrowing via Jurchen-Manchu. Also, many of the items have wider connections, especially in Turkic, including Pre-Proto-Bulgharic, reflecting the flow of influences from Turkic to Mongolic to Tungusic. Some random examples of Mongolic items in Manchu include akta ‘gelding’ ← *agta, baturu ‘hero’ ← *baatur < *baxatur < *bagatur (with wider Eurasian connections), borcila- ‘to hang up to dry’ ← *borca-la- (also with wider parallels), boro ‘gray’ ← *boro < *bora (from Pre-Proto-Bulgharic), cira ‘face’ ← *cira : *cira-i, dabagan ‘mountain pass’ ← *daba-gan > *daba-xan, doholon ‘lame’ ← *doko-loƞ < *doka-lang, falanggū ‘palm of the hand’ ← *palaƞku > Middle Mongol †xalaƞku, gindana ‘prison’ ← *gindan, hergen ‘title’ ← *kerge-m, hūntaha ‘goblet’ ← *kundaka > *kundaga/n, ikiri ‘twin/s’ ← *ikire (from Pre-ProtoBulgharic), kirsa ‘steppe fox’ ← *kirsa, sain ‘good’ ← *sain < *saxïn < *sagï/n (from Pre-Proto-Bulgharic). There are certainly also many Para-Mongolic elements in Manchu received during the interaction of the Jurchen with the Khitan and other Khitanic-speaking tribal groups. However, it is often difficult to distinguish the Para-Mongolic and Proto-Mongolic lineages in the lexical material, and some items, such as, for instance, niohe ‘wolf’ : niyahan ‘puppy’, could also have been received from the common Pre-Proto-Mongolic ancestor of Mongolic and Khitan(ic), cf. Mongolic *noka-i < *ñoka ‘dog’. It is reasonable to assume that many of the administrative terms used in Manchu were borrowed from Khitan, e.g. doro ‘ceremony’, gurun ‘country, state’. Also, some of the Turkic items that in Manchu have a shape different from that attested in Mongolic may have been transmitted by Khitan, cf. e.g. eihen ‘donkey’ vs. Mongolic *elji-ke/n (from Pre-Proto-Bulgharic). Quite certainly Khitanic is the Jurchen set for the teens, of which 11 omšon, 12 jorhon ~ jorgon, 15 tofohon, and 16 niolhun are preserved in Manchu. The Chinese loanwords in Manchu also reflect several chronological layers, and they are a particularly important source regarding the phonological history of Mandarin. Examples include boobai ‘treasure’ ← *bawbay 寶貝 bǎobèi, giyang ‘river’ ← *gyang 江 jiāng, hūwangdi ‘emperor’ ← *hwangdi 皇帝 huángdì, laba ‘trumpet’ ← *laba 喇叭 lǎba, lan ‘orchid’ ← *lan 蘭 lán, lang ‘young man’ ← *lang 郎 láng, lian dzi ‘lotus seed’ ← *lyanzi 蓮子 liánzi, loo ‘prison’ ← *law 牢 láo, niyang ‘girl’ ← *nyang 娘 niáng, wang ‘prince’ ← *wang 王 wáng, yamun ‘government office’ ← *yamun 衙 門 yámen, yangse ‘kind’ ← *yangzi 様子 yàngzi. Finally, there are the words of unknown origin. Items attested also in Jurchen include aga ‘rain’, uncehen ‘tail’, haha ‘man’, hehe ‘woman’, nimaha ‘fish’, jugūn ‘road’, while items attested only in Manchu include uheri ‘all’, bulukan ‘warm’, sukū ‘bark, skin’, šahūrun ‘cold’, tuwa- ‘to look’, yonggan ‘sand’. Quite probably, some of these will turn out to have Tungusic etymologies, while others may be borrowings from Khitan, or also from other extinct languages of Manchuria. It should also be borne in mind that phonetic developments in Manchu have occasionally rendered inherited words less transparent. For instance, wehe ‘stone, rock’, which has occasionally been classified as an item of unknown origin, is actually a regular cognate of Ewenki urekeen ‘hill’, from Proto-Tungusic *xöre-kee/n.

134 Alexander Vovin

LIST OF PRIMARY SOURCES LGD LJB MLD MYK NSB SJDB TM

Manju gisun Loo gi da, 18th c., quoted after Tsumagari (1977–1978) Lakcaha jecen-de takūraha ba-be ejehe bithe, 1723 Manzhou laodang (Manchu Old Archives), Taizong reign, 1628–1643 Manju yargiyan kooli, 1781 quoted after Imanishi (1992) Nišan saman-i bithe (The Tale of the Nišan Shamaness), 19th c. Samjŏndo bi (Samjŏndo Inscription), 1632 AD Tanggū meyen (A Hundred Chapters), 19th c.

REFERENCES AND FURTHER READING Adam, Lucien (1873), Grammaire de la langue mandchou, Paris: Librérie-Éditeur J. Maisonneuve et cie. Aisin Gioro, Ulhicun 愛新覺羅烏拉熙春 (1985)『滿語讀本』Manju gisun i hūlara bithe [A textbook of the Manchu language], 呼和浩特 [Hohhot]: 内蒙古民族出版社. Alonso de la Fuente, José Andrés (2010–2011) ‘Written Manchu uju ‘head’’, Ural-Altaische Jahrbücher, N. F. 24: 204–226. Alonso de la Fuente, José Andrés (2012) ‘Written Manchu talman ‘fog, mist’ and the Tungusic liquids’, Zentralasiatische Studien 41: 107–135. Alonso de la Fuente, José Andrés (2013a) ‘Written Manchu akjan ‘thunder’’, Acta Orientalia Academiae Scientiarum Hungaricae 66 (1): 59–68. Alonso de la Fuente, José Andrés (2013b) ‘Written Manchu guwafu ‘crutch; pole’ and gûwa ‘(an)other’, International Journal of Diachronic Linguistics and Linguistic Reconstruction 10: 27–54. Alonso de la Fuente, José Andrés (2014) ‘Written Manchu ferguwecuke ‘astonishing’ and the oldest causative formations in Tungusic’, in Bareja Starzyńska & Agata Rogala & Filip Majkowski (eds), A Window Onto the Other: Contributions on the Study of the Mongolian, Turkic, and Manchu-Tungusic Peoples, Languages, and Cultures Dedicated to Jerzy Tulisow on the Occasion of His Seventieth Birthday, 105–123, Warsaw: Elipsa. Alonso de la Fuente, José Andrés (2017a) ‘From converb to classifier? On the etymology of Literary Manchu nofi’, in: Michal Németh & Barbara Podolak & Mateusz Urban (eds.), Essays in the History of Languages and Linguistics Dedicated to Marek Stachowski on the Occasion of his 60th Birthday, 57–80, Kraków: Księgarnia Akademicka. Alonso de la Fuente, José Andrés (2017b) El libro de Nishán la Saman [The tale of the Nišan shamaness], Madrid: Ediciones Complutense. Alonso de la Fuente, José Andrés & Guillaume Jacques (2018) ‘Associated motion in Manchu in typological perspective’, Language and Linguistics 19 (4): 501–524. An, Shuangcheng 安雙成 (ed.) (1993)『满漢大詞典』(A Comprehensive Manchu-Chinese Dictionary), 瀋陽 Shenyang: 遼寧人民出版社. Avrorin, V. A. [В. А. Аврорин] (2000) Грамматика маньчжурского письменного языка [A grammar of Written Manchu], Санкт-Петербург [St. Petersburg]: “Наука”. Chuang, Chi-fa 莊吉發 (1984)『雍正朝漢合璧奏習校注』[An annotated edition of memorials from Yongzheng’s reign], vol. 1, 台北 [Taipei]: 文史哲出版社. Chuang, Chi-fa 莊吉發 (2012)『清代準噶爾史料初編』[A first collection of historical materials on the Jungars in the Qing period], 台北 [Taipei]: 文史哲出版社. Chuang, Chi-fa 莊吉發 (2014a)『清語老乞大譯註』[Laojida in Manchu translated and annotated], 台北 [Taipei]: 文史哲出版社.

Written Manchu 135

Chuang, Chi-fa 莊吉發 (2014b)『滿漢異域録校注』[An annotated edition of the Manchu and Chinese “Records of a Mission to Remote Frontiers”], 台北 [Taipei]: 文史哲 出版社. von der Gabelentz, H[ans] Conon [de la] (1832) Élémens de la grammaire mandchoue, Altenbourg: Comptoir de la littérature. von der Gabelentz, Hans Conon (1864) Sse-schu, Schu-king, Schi-king in Mandschuischer Uebersetzung mit einem Mandschu-deutschem Wörterbuch, Zweites Heft: Wörterbuch, Abhandlungen für die Kunde des Morgenlandes 3.2, Leipzig: Brockhaus. Gimm, Martin (1968) Die chinesische Anthologie Wen-Hsüan in mandjurischer Teilübersetzung einer Leningrader und einer Kölner Handschrift, Wiesbaden: Franz Steiner Verlag. Gimm, Martin (1969) Die Kaiserliche Ku-wen-Anthologie von 1685/6, Ku-wen YüanChien, in mandjurischer Übersetzung, vols. 1–2, Wiesbaden: Otto Harrassowitz. Gorelova, Liliya M. (2002) Manchu Grammar, Handbuch der Orientalistik, Section 8 (Uralic and Central Asian Studies), vol.7, Leiden: Brill. Haenisch, Erich (1959) Dokumente aus dem Jahre 1788 zur Vorgeschichte des Gorka-Krieges, Bayerische Akademie der Wissenschaften, philosophisch-historische Klasse, Abhandlungen, Neue Folge 49, München: Verlag der Bayerischen Akademie der Wissenschaften. Haenisch, Erich (1961), Mandschu-Grammatik mit Lesestücken, Lehrbücher für das Studium der orientalischen Sprachen 6, Leipzig: VEB Verlag Enzyklopädie. Haenisch, Erich (1965) Bericht von einer chinesischen Gesandschaft nach Annam im Jahre 1668/69: Eine Mandschu-Handschrift aus dem Pekinger Palastmuseum, Bayerische Akademie der Wissenschaften, philosophisch-historische Klasse, Abhandlungen, Neue Folge 60, München: Verlag der Bayerischen Akademie der Wissenschaften. Haenisch, Erich (1970) Historische Mandschutexte, aus dem Nachlaß mit Anmerkungen herausgegeben von Michael Weiers, Asiatische Forschungen 29, Wiesbaden: Otto Harrassowitz. Haneda, Tōru 羽田亨 (1937)『満和辭典』[A Manchu-Japanese dictionary], 京都 [Kyoto]: 京都帝國大學満蒙調査會. Haneda, Tōru 羽田亨 (1972)『満和辭典』[A Manchu-Japanese dictionary], [Reprint], 東京 [Tokyo]: 国書刊行会. Harlez, C[harles] de (1884) Manuel de la langue mandchoue: Grammaire, anthologie & lexique. Paris: Maisonneuve frères & C. Leclerc. Hauer, Erich (1952–1955) Handwörterbuch der Mandschusprache, vols. 1–3, Wiesbaden: Otto Harrassowitz. Hauer, Erich (1991) ‘Abriß der manjurischen Grammatik: Ein nachgelassener Text,’ [published by] M[artin] G[imm], in: Martin Gimm & Giovanni Stary & Michael Weiers (eds.), Klassische, moderne und bibliographische Studien zur Mandschuforschung, Aetas Manjurica 2: 117–146. Wiesbaden: Otto Harrassowitz. Hauer, Erich (2007) Handwörterbuch der Mandschusprache, 2., durchgesehene und erweiterte Auflage herausgegeben von Oliver Corff, Wiesbaden: Harrassowitz Verlag. Hayata, Teruhiro 早田輝洋 (1998)『満文金瓶梅訳注：序―第十四回』[Translation and commentary to the Manchu version of Jing Ping Mei, from the introduction to chapter 14], 東京 [Tokyo]: 第一書房.

136 Alexander Vovin

Hayata, Teruhiro 早田輝洋 (2012)「満州語のn ~ Ø 交替の史的概観」[A historical overview of the alternation of n ~ Ø in Manchu], Altai Hakpo『알타이학보』22: 93–110, Seoul. Hayata, Teruhiro 早田輝洋 (2015)「満州語の多様な形態素-nggeの文法的位置づ け」[The grammatical status of -ngge morphemes in Manchu],『言語研究』Journal of the Linguistic Society of Japan 148: 33–60, 東京 [Tokyo]. Honggiri, Baktan 張博 (2006) Baktan’i emgi Manju gisunbe sumimalaki! Decode Manchu Language with Baktan! s.l.: Baktan ‘BKT’ Studio (limited edition). Hu, Zengyi 胡增益 (ed.) (1994)『新满汉大词典』[A New Comprehensive Manchu-Chinese Dictionary], 烏魯木齊 Urumqi: 新疆人民出版社. Ikegami, Jirō 池上二良 (1999)『満洲語研究』[Researches on the Manchu language], 東京 [Tokyo]: 汲古書院. Imanishi, Shunjū 今西春秋 (1964)『校注異域録』[Tulišen’s I-yü-lu revised and annotated], 天理 [Tenri]: 天理時報社. Imanishi, Shunjū 今西春秋 (1992)『満和蒙和対訳満洲実録』[The Manchu Veritable Records in Manchu and Mongolian with a Japanese translation], 東京 [Tokyo]: 刀水書房. Ivanovskii, A. O. [А. О. Ивановскiй] (1893–1895) Маньчжурская хрестоматiя (A Manchu reader), vols. 1–2, Санктпетербургъ [St. Petersburg]. Ji, Yonghai 季永海 & Liu Jingxian 劉景宪 & Qu Liusheng 屈六生 (1986)『满语语 法』[A Grammar of the Manchu language), 北京 [Peking]: 民族出版社. Kawachi, Yoshihiro 河内良弘 (1996)『満州語文語文典』[A grammar of the Written Manchu language], 京都 [Kyoto]: 京都大学学術出版会. (Reprinted in 2014.) Klaproth, Julius von (1828) Chrestomathie Mandchou: Ou, recueil de textes Mandchou, destiné aux personnes qui veulent s’occuper de l’étude de cette langue, Paris. Imprimerie royale. Kŭrteva, Svetla [Светла Къртева] (1991) Учебник по манджурски език [A textbook of the Manchu language], София [Sofia]: УИ “Св. Климент Охридски”. Li, Gertraude Roth (2000) Manchu: A Textbook for Reading Documents, Honolulu: University of Hawai‘i Press. Möllendorff, P[aul] G[eorg] von (1892) A Manchu Grammar with Analyzed Texts, Shanghai: American Presbyterian Mission Press. Naeher, Carsten (2004) ‘A note on vowel harmony in Manchu’, in: Carsten Naeher & Giovanni Stary & Michael Weiers (eds.) (2004) Proceedings of the First International Conference on Manchu-Tungus Studies (Bonn 2000), vol. 2. Trends in Tungusic and Siberian Linguistics, Tunguso-Sibirica 9: 129–136, Wiesbaden: Harrassowitz Verlag. Norman, Jerry L. (1965) ‘A grammatical sketch of Manchu’, Phi Theta Annual: Papers of the Honor Society in Oriental Languages of the University of California, vol. 9: 1–43, Berkeley: University of California Library. Norman, Jerry (1978) A Concise Manchu-English Lexicon, Seattle and London: University of Washington Press. Norman, Jerry (2013) with the assistance of Keith Dede and David Prager Branner, A Comprehensive Manchu-English Dictionary, Cambridge (Massachusetts) and London: Harvard University Asia Center. Nowak, Margaret & Stephen Durrant (1977) The Tale of the Nišan Shamaness: A Manchu Folk Epic, Publications on Asia of the Institute on Comparative and Foreign Area Studies 31, Seattle: University of Washington Press. Orlov, A. [М.] [A. М. Орловъ] (1873) Грамматика маньчжурскaго языка [A Grammar of the Manchu language], Санкт-Петербургъ [St. Petersburg]: Типографiя Императорской Академiи Наукъ.

Written Manchu 137

Özalan, Uluhan (2021) Mançuca Nișan Şaman hikâyesi (Nisan Šaman i Bithe), Giriş— Şekil Bilgisi—Çeviri—Metin—Sözlük) [Introduction, Morphology, Translation, Text, Glossary], Ankara: Türk kültürünü araştırma enstitüsü. Pang, Tatiana A. & Giovanni Stary (1998) New Light on Manchu Historiography and Literature: The Discovery of Three New Documents in Old Manchu Script, Wiesbaden: Harrassowitz Verlag. Pashkov, B. K. [Б. К. Пашков] (1963) Маньчжурский язык [The Manchu language], Москва [Moscow]: Издательство восточной литературы. Pozdneev A. M. [A. M. Позднѣевъ] (1904) Опытъ собиранiя образцовъ маньчжурской литературы [A collection of samples of Manchu literature], Владивостокъ [Vladivostok], [originally published in:] Извѣстiя Восточнаго Института 3 (1902): 1–32, 37–70, 81–120, 121–168, 169–216, 4 (1903): 271–272, 272–360, 375–406. Rozycki, William (1981) A Reverse Index of Manchu, Indiana University Uralic and Altaic Series 140, Bloomington: Indiana University, Research Center for Inner Asian Studies. Seong, Baeg-in 成百仁 (1999)『만주어와 알타이어학 연구』[Study of the Manchu language and Altaic linguistics], Seoul: Taehaksa. Sinor, Denis (1949) ‘Le verbe mandjou’, Bulletin de la Société de Linguistique de Paris 45: 146–156. Sinor, Denis (1968) ‘La langue mandjoue’, in: Altaistik: Tungusologie, Handbuch der Orientalistik I.V.3: 257–280, Leiden: E. J. Brill. Stary, Giovanni (1985) Three Unedited Manuscripts of the Manchu Epic Tale “Nišan saman-i bithe”: Facsimile Edition with Transcription and Introduction, Wiesbaden: Kommissionsverlag Otto Harrassowitz. Stary, Giovanni (2000) A Dictionary of Manchu Names: A Name Index to the Manchu Version of the “Complete Genealogies of the Manchu Clans and Families of the Eight Banners”, Aetas Manjurica 8, Wiesbaden: Otto Harrassowitz. Stary, Giovanni (1990–2003) Manchu Studies: An International Bibliography, vols. 1–4, Wiesbaden: Harrassowitz Verlag. Stary, Giovanni (2005) “What’s Where” in Manchu Literature, Aetas Manjurica 11, Wiesbaden: Harrassowitz Verlag. Sunik, O. P. [О. П. Суник] (1997) ‘Маньчжурский язык’ [The Manchu language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 162–173, Москва [Moscow]: “Индрик”. Tsumagari, Toshirō 津曲敏朗 (1977–1978)「清語老乞大の研究：満州語研究のため の一資料(1), (2)」[A study of Laojida in the Manchu language: Materials for the study of Manchu],『札幌商科大学・札幌短期大学論集（人文編）』21: 211– 248, 22: 161–169, 札幌 [Sapporo]. Tsumagari, Toshirō 津曲敏朗 (2000) ‘The Manchu first person plural pronouns: with reference to some neighboring languages”, Altai Hakpo『알타이학보』11: 99–110, Seoul. Tsumagari, Toshirō 津曲敏朗 (2001)「満州語動詞語尾-ciの末用法と-cinaについ て」[The Manchu sentence-final converb in -ci with reference to the origin of -cina], Altai Hakpo『알타이학보』10: 139–150, Seoul. Tsumagari, Toshirō 津曲敏朗 (2002)『満州語入門２０講座』[An introduction to the Manchu language in twenty lessons], 東京 [Tokyo]: 大學書林. Tulisow, Jerzy (2000) Język mandżurski [The Manchu language], Języki Azji i Afryki, Warszawa [Warsaw]: Wydawnictwo Akademickie Dialog.

138 Alexander Vovin

Tulisow, Jerzy (2004) Krótki kurs języka mandżurskiego [A brief course of the Manchu language], Warszawa [Warsaw]: Wydawnictwo Akademickie Dialog. Volkova, M. P. [М. П. Волкова] (1961) Нишань самани битхэ (предание о нишанской шаманке) [The tale of the Nišan shamaness], Mосква [Moscow]: Издательство восточной литературы. Walravens, Hartmut & Martin Gimm (1978) Deutsch-mandjurisches Wörtverzeichnis (nach H. C. von der Gabelentz Mandschu-Deutschem Wörterbuch), Sinologica Coloniensia 4, Wiesbaden: Franz Steiner Verlag GmbH. Yamamoto, Kengo 山本謙吾 (1955)「満州語文語形態論」[The morphology of the Manchu written language], in: Ichikawa Sanki 市河三喜 & Hattori Shirō 服部四郎 (eds.),『世界言語概説』 [Introduction to the languages of the world], vol. 2: 489– 538, 東京 [Tokyo]: 研究社. Zakharov, Ivan [Иванъ Захаровъ] (1875) Полный маньчжурско-русскiй словарь [A complete Manchu-Russian dictionary], Санктпетербургъ [St. Petersburg]: Типографiя Императорской Академiи Наукъ. Zakharov, Ivan [Иванъ Захаровъ] (1879) Грамматика маньчжурскaго языка [A Grammar of the Manchu language], Санктпетербургъ [St. Petersburg]: Типографiя Императорской Академiи Наукъ.

CHAPTER 7

SIBERIAN EWENKI Juha Janhunen

By Siberian Ewenki we understand the mainstream varieties of the Ewenki language, the principal representative of the Ewenic branch of the Tungusic family. Most speakers of Siberian Ewenki live scattered in Siberia and the Russian Far East, where their ethnic territory extends from the eastern tributaries of the Ob in the west to the island of Sakhalin in the east, and from the Baikal region in the south to the lower courses of the Yenisei and Lena in the north. Administratively, most Ewenki live today in the Krasnoyarsk, Zabaikal’sk, Khabarovsk, and Primorsk Krais, the Tomsk, Irkutsk, Amur and Sakhalin Oblasts, as well as in the Sakha and Buryat Republics (Yakutia and Buryatia) of the Russian Federation. Between 1930 and 2007 the Ewenki enjoyed titular autonomy within the so-called Ewenki Autonomous District (Èvenkiiskii avtonomnyi okrug), also known as “Ewenkia” (Èvenkiya), located in the northwestern part of Krasnoyarsk Krai. In Russia, the speakers of Ewenki are traditionally classified as one of the “Small Indigenous Peoples of the Far North, Siberia and the Far East” (korennye malochislennye narody Severa, Sibiri i Dal´nego Vostoka), among whom they form the largest group by population size. According to the last Soviet census (1989), the number of Ewenki was 29,975, of whom, however, only 29.5 percent spoke Ewenki. A more recent census (2010) gives the figure of 38,396 individuals, of whom only c. 10 percent, that is, c. 4,800 people speak the Ewenki language. The low proportion of native language speakers suggests that the language is seriously endangered, with most of the speakers being far beyond reproductive age. Even so, there are still children learning Ewenki in a few localities in southern Yakutia, in the Amur region, and possibly in the former Ewenki Autonomous District. Before the Soviet period, the Ewenki used to be known as the “Tungus” par excellence, Russian tungús (PL tungúsy), a name adopted for them by the Russians in the early 17th century from the Samoyedic Tundra Nenets, the western neighbours of the Ewenki in the Lower Yenisei region, who call the Ewenki by the ethnonym tuƞkoq : PL tuƞkos°-q. Ethnographically and ecologically, the Tungus used to be divided into four groups: Reindeer Tungus (olennye tungusy), Horse (Equestrian) Tungus (konnye tungusy), Dog Tungus (sobach’i tungusy), and Foot (Pedestrian) Tungus (peshie tungusy). Most of the Siberian Ewenki speakers used to fall in the category of Reindeer Tungus, though the population of the Horse Tungus was also considerable. In the late 17th to early 18th centuries, the period of the Russian expansion towards the Pacific coast, the single largest Ewenki group was formed by the Horse Tungus of Transbaikalia. Before the colonial period, all “Tungus” groups were nomadic, with hunting, fishing, gathering and small-scale cattle breeding as their occupations. The situation changed, especially, with the forced settlement policies of the Soviet period.

DOI: 10.4324/9781315728391-7

140 Juha Janhunen

The ethnonym Ewenki (also Evenk, Evenki, Russian MASC èvénk : PL èvénki : FEM èvenkíika : ADJ èvenkíiskii), replaced the older name “Tungus” during the Soviet period and derives from the endonym ewenkii [əβənki:] ~ eweƞkii [əβəŋki:], as used by most Ewenki speakers. This name is derived from the simple root ewe- : ewen (< *ebe/n), an ethnonym possibly connected with the proximal pronoun e- ‘this’ and also used by the Ewen (Lamut). The corresponding adjectival form is ewe-dii, from which the instrumental case form ADJ INSTR ewe-dii-ji > ewe-dii-t is used in the meaning ‘(speaking) in the Ewenki language’. Other ethnonyms used for the Ewenki and related groups include Khamnigan and Orochen (on which see below), as well as Kili (as used in the Amur region and on Sakhalin). TAXONOMIC STATUS The separation of Siberian Ewenki from the rest of the Ewenic languages and dialects involves several problems. Many of these problems are connected with the fact that several varieties of Ewenic closely related to Siberian Ewenki are spoken in the Chinese parts of Manchuria south of the Amur main course. In the broad sense, all of these varieties belong to the context of the Ewenki language. However, due to ethnohistorical, ecological, political, and administrative differences it is today customary to distinguish, apart from “regular” Siberian Ewenki and Ewen, at least four other Ewenki-related ethnolinguistic entities, which are: (1) Neghidal, an Ewenic idiom spoken in the Amgun’ basin of the Lower Amur region of the Russian Far East by a population which ethnohistorically represents a combination of the Ewenki language with a local cultural base similar to that of the neighbouring “Palaeosiberian” Ghilyak (Nivkh) speakers. In the Russian taxonomy, the Neghidal (Russian negidal’cy) are classified as a separate ethnic group with an ethnic language of their own. Linguistically, Neghidal may in most respects be seen as an innovative form of Ewenki. In the present volume Neghidal is treated in a separate chapter. (2) Solon, a complex of Ewenic idioms spoken by a number of regional populations in Manchuria, historically originating from the Middle Amur basin but today scattered in the Nonni (Nenjiang) and Hui-Imin-Hailar basins. The Solon, who use the endonym eweƞki, may be seen as an aberrant Ewenki group that has developed under the influence of the Mongolic Daghur speakers, as well as, indirectly, of the Jurchen-Manchu. Linguistically, Solon is in some respects innovative and is not readily intelligible to speakers of other Ewenic idioms, including regular Siberian Ewenki. A special subethnic group of Solon is formed by the Ongkor Solon (today linguistically extinct), who since the mid 18th century have lived in the Ili region of East Turkestan. In the present volume Solon is treated in a separate chapter. (3) Khamnigan Ewenki, a complex of Ewenic idioms historically spoken by the “Horse Tungus” of Transbaikalia, including adjacent parts of Mongolia and Manchuria, but today preserved only in the Mergel basin of Hulun Buir League (City), northern Inner Mongolia, China. The Khamnigan varieties of Ewenki have developed in close interaction with Khamnigan Mongol, a Mongolic language in which the Khamnigan are ethnically bilingual. The Khamnigan speak two taxonomically different varieties of Ewenki, each of which is transmitted within separate clans, with Khamnigan Mongol serving as the lingua franca of the population.

Siberian Ewenki 141

(4) Orochen, a complex of Ewenic idioms spoken by a number of local populations on the modern Chinese side of the Upper and Middle Amur basin. Most Orochen live historically along minor rivers flowing into the Amur from the Khingan Ranges and identify themselves according to these rivers by names such as Kumarchen, Selpechen, Birarchen, and others. The name Orochen, from Ewenki orocien (Chinese Pinyin Oroqen), also attested as Oronchon, from Manchu oroncon (Mandarin Elunchun), is conventionally associated with the Ewenki word oron ‘domestic reindeer’ and is also used by some Siberian Ewenki groups on the modern Russian side of the border, though the Orochen south of the Amur have never had reindeer. Linguistically, the Ewenic varieties spoken by the Orochen in China are separated from Siberian Ewenki by both some archaisms and, above all, several regional innovations, which is why they may be classified as an independent, though internally heterogeneous, complex of dialects. In the present volume they are treated in a separate chapter. Due to a mistake made in the ethnic taxonomy, the Ewenic groups in China are arbitrarily divided into two distinct “minority nationalities”, officially known as the Orochen and Ewenki. In this context, the “Ewenki” (Ewenke) comprise all the groups that are not “Orochen” (Elunchun), that is, the Solon and Khamnigan, to which a third group, known as the “Yakut” (Yakute) is added. The “Yakut”, historically also known as the Manchurian Reindeer Tungus, represent a small 19th-century migrant group of reindeer herders from the Siberian side and speak (or used to speak) a regular Siberian Ewenki dialect, immediately intelligible to Ewenki speakers on the Russian side of the border. From a diachronic perspective, the diversity of the different forms of Ewenic may be understood somewhat differently. If we look at two apparently early and, hence, primary phonological innovations, we can see that Ewenic shows a dichotomy between two branches, which may be termed Siberian Ewenki and Manchurian Ewenki. Manchurian Ewenki in this connection comprises, in particular, Solon. • •

A primary innovation that took place in Siberian Ewenki, but not in Manchurian Ewenki, was the loss of the final high vowels *i *ï after a single consonant, as in *dïlï ‘head’ > Siberian Ewenki dil vs. Solon dɪlɪ (> later dialectally dɪl). A primary innovation that took place in Manchurian Ewenki, but not in Siberian Ewenki, was the raising of initial *ie [e:] into ii [i:] as in *iesa : PL *iesa-l ‘eyes’ > Siberian Ewenki iesal vs. Solon iisal.

On both of these points Neghidal goes together with Siberian Ewenki, as does Ewen. This type also comprises one of the two dialects of Khamnigan Ewenki, the so-called Uru­ lyunggui (or Urul’ga) dialect, while the other Khamnigan Ewenki dialect, the so-called Borzya (or Man’kovo) dialect goes together with Solon, though it lacks the other innovations characteristic of Solon. This suggests that the Borzya dialect represents a conservative form of Manchurian Ewenki, while Neghidal and Ewen are innovative forms of Siberian Ewenki. Orochen seems to be based on an early form of Manchurian Ewenki, but it has secondary influences from Siberian Ewenki. This general picture is confirmed by a number of important lexical shibboleths. For instance, the Common Tungusic word for ‘sun’, *sigöön, is preserved in Solon and the Borzya dialect of Khamnigan Ewenki. In most forms of Siberian Ewenki it has been replaced by *dïlacaa, an innovation shared by Orochen and the Urulyunggui dialect of

142 Juha Janhunen

Khamnigan Ewenki, though, interestingly, not by Neghidal. The Common Tungusic word for ‘to drink’, *umï-, has regular reflexes in most forms of Ewenic, but in Solon, Orochen and the Borzya dialect of Khamnigan Ewenki it has been replaced by the innovative shape *ïmï- (> ɪmɪ- > imi-). It may be noted that Orochen and Neghidal occupy in these cases an ambivalent position. DIALECTAL DIVISION As a whole, the Ewenic branch of Tungusic is surprisingly homogeneous, especially considering the huge territory it occupies almost all over Siberia and Manchuria. This suggests that the language spread rapidly and relatively late from a compact homeland, which was probably located in the Middle Amur basin, the region where the separation of the Siberian and Manchurian branches must also have taken place. Apparently, Ewenki spread by both migration and diffusion and was adopted by several originally non-Tungusic local populations in various parts of Siberia. The original differences between these populations are still visible in material and spiritual culture, as well as in physical anthropology. However, substratal features in the language are conspicuously few, suggesting that Ewenki spread as a prestige medium, which easily replaced the older local languages. Even so, Ewenki does show some internal variation even if we do not consider the more aberrant Ewenic varieties like Neghidal and Solon. Focusing on Siberian Ewenki in the narrow sense, meaning the language spoken by the Ewenki in Russia, but considering also some features of Khamnigan Ewenki and Orochen, we may distinguish between at least three principal groups of dialects: southern, eastern, and northern. The diagnostic feature most often used in the classification of Ewenki dialects is the distribution and representation of the laryngeal fricative x [h]. This sound can stand for two original segments in Ewenki: either Proto-Tungusic initial *p, which was spirantized to *x already before Proto-Ewenic but later dialectally lost, or Proto-Tungusic initial or medial *s, which was preserved in Proto-Ewenic but later dialectally and positionally desibilized (debuccalized) to x. In initial position Proto-Tungusic also had a primary *x, but it is systematically represented as zero in all forms of Ewenic. The representation of *s, in particular, is conventionally used as a shibboleth distinguishing the three principal groups of Ewenki dialects: (1) In the southern dialects, geographically extending from the Sym and Stony Tunguska basins in the west to the Baikal region in the east, *s is preserved as a sibilant both initially and between vowels. The varieties conventionally classified as belonging to this group include the Stony Tunguska (podkamenno-tungusskii), Nepa (nepskii), Sym (symskii), Tokma-Upper Lena (tokminsko-verkholenskii), North Baikal (severobaikal’skii), and Vitim-Nercha (vitimo-nerchinskii) dialects. On a phonetic basis, depending on whether *s is pronounced as a dental s [s] or a palato-alveolar sh [ʃ], these dialects are further divided into the so-called hissing (sekaiusshie) and hushing (shekaiushhie) varieties, with the Stony Tunguska, Nepa and Vitim-Nercha dialects belonging to the former and the Sym, Tokma-Upper Lena and North Baikal dialects belonging to the latter group. Solon, Neghidal, Orochen and Khamnigan Ewenki also belong to the former group. The phonetic difference in the pronunciation of the sibilant is itself linguistically irrelevant, but it is occasionally reflected in language contacts, notably in the Russian term shamán ‘shaman’, which was borrowed from the hushing form [ʃama:n], as opposed to the hissing form [sama:n].

Siberian Ewenki 143

(2) In the eastern dialects, geographically extending from the Vitim and Olekma basins in the west to the Amur basin and Sakhalin in the east, *s is preserved as a sibilant in initial position but represented as x [x ~ h] in intervocalic position, as in *asii ‘woman’ > axii [ahi:]. The varieties conventionally classified as belonging to this group include the Barguzin (barguzinskii), Vitim-Olekma (vitimo-olekminskii), Upper Aldan-Zeya (verkhnealdansko-zeiskii), Uchur-Zeya (uchursko-zeiskii), Ayan-Maya (ayano-maiskii), Chumikan (chumikanskii), Selemdzha-Bureya-Urmi (selemdzhinsko-bureisko-urmiiskii), and Sakhalin (sakhalinskii) dialects. (3) In the northern dialects, geographically extending from the Lower Yenisei basin to the upper courses of the Vilyui and Olenek, *s is preserved as a sibilant in medial position but represented as a velar to laryngeal x [x ~ h] in initial position, as in *sulakii ‘fox’ > xulakii [hulaki:]. The varieties conventionally classified as belonging to this group include the Ilimpiya (ilimpiiskii) and Yerbogachen (erbogachenskii) dialects. Ewen, with the exception of Arman, also belongs to this type, making it, from this point of view, another member of the same continuum. However, in some subdialects of both Ewenki and Ewen *s is represented as x [h] both initially and medially. Altogether, the dialectological picture is rather complicated and may in some cases reflect highly localized and chronologically shallow developments. Areally, the desibilization (debuccalization) of *s [s] to x [x ~ h] is a widespread phenomenon in Siberia, attested, with varying positional conditions, also in Turkic (Yakut and Dolgan), Mongolic (Buryat and Khamnigan Mongol) and Samoyedic (Forest Nenets, Selkup dialectal). The representation of *s in postconsonantal position is also of dialectological relevance. Most importantly, in the position after the homorganic sonorants *l *n, Proto-Tungusic *s developed into Proto-Ewenic *r [r ~ dr] (perhaps still an allophone of /s/), which is preserved only in a few marginal Ewenki dialects in the west (Sym) and east (AyanMaya), while in several other dialects in the north (Ilimpiya), center (North Baikal) and east (Uchur-Zeya) it has yielded d. In most dialects of Siberian Ewenki the clusters ld nd have undergone full assimilation to the geminates ll nn, which, in turn, can be simplified to monophonemic l n. All these representations are also attested in the other Ewenic idioms: r in (mainstream) Ewen, d in Solon and Khamnigan Ewenki, and full assimilation in Neghidal and Orochen. The development of Proto-Ewenic (*p >) *x (> Ø) shows a considerably less clear dialectological picture. It seems that this segment was preserved long into the history of the individual forms of Ewenic until it started disappearing independently in several local dialects. *x is systematically preserved in Ewen and Neghidal, and systematically lost in Solon and Orochen. In Siberian Ewenki it is widely preserved, but lost, especially, in the Sym, North Baikal, Barguzin and Sakhalin dialects, as in *xekuu ‘hot’ > xekuu ~ ekuu. It is also lost in the Urulyungui dialect of Khamnigan Ewenki, but preserved, or only facultatively lost, in the Borzya dialect. Phonemically, *x > x yields the same sound as *s > x, meaning that the two sources of x complement each other in the synchronic system of phonotactics. DATA AND SOURCES Apart from Manchu, Ewenki is by far the best known and most thoroughly documented Tungusic language. It was also the first Tungusic language that the Russians encountered

144 Juha Janhunen

in Siberia, and from which scattered data started appearing in Western sources. P. S. Pallas (1786–1789) lists vocabulary items from six dialects of Siberian Ewenki proper, as spoken in the administrative districts of Yeniseisk, Mangazeya, Barguzin, and Yakutsk, as well as on the Upper Angara and in the Tunguska region (“Chapogir”) (n:os 139–143, 146). In addition, there are items from Nerchinsk, representing a Borzya type of dialect of Khamnigan Ewenki (n:o 138), and two varieties of Ewen (n:os 144–145). The Ewenki dialects of the Khamnigan came to play a particularly important role in the history of Ewenki studies, for they were the object of the first systematic grammatical description and lexical documentation of Ewenki by M. A. Castrén (1856). Although Castrén had during his travels in Siberia come into contact with many other dialects, his actual fieldwork on Ewenki took place around Nerchinsk, where he documented both varieties of Khamnigan Ewenki (in his terms Urul’ga and Man’kovo). Castrén’s work, as published by Anton Schiefner, remained long the only grammatical source on Ewenki. A Russian translation of his grammar was published decades later by E. I. Titov (1926), who added a corpus of fresh lexical materials from several dialects spoken in the Baikal region. Another early scholar to document the lexical resources of Ewenki was the geologist Alexander Czekanowski, whose collection was also published by Schiefner (1881). Of far greater importance was, however, the work of S. M. Shirokogoroff, whose extensive analysis of the social anthropology (1933) and shamanism (1935) of the “Northern Tungus” focuses on the Ewenki and Orochen groups of Transbaikalia and Manchuria and contains unique dialectological and onomastic information. Shirokogoroff’s lexical materials became available only posthumously (1944) and have more recently been republished as a critical edition by Gerhard Doerfer and Michael Knüppel (2004). Of particular value are Shirokogoroff’s data on several otherwise understudied dialects of the Siberian-Manchurian interspace, such as Birarchen, as discussed by Doerfer (1983). After the October Revolution, much of the research done on Ewenki in Russia was focused on the practical goal of creating and maintaining a normative language and native literacy for the ethnic Ewenki, from whom the Neghidal and Ewen were separated as distinct groups. In this context, the choice of the base for the literary language became an important issue. The choice fell initially on the Nepa dialect of the southern group, but was later transferred to the Stony Tunguska dialect. From the geographical point of view this was an unfortunate decision, as was immediately noted by Shirokogoroff (1938), but phonetically these dialects had the advantage of preserving the sibilant *s in all positions. The leading authority on Ewenki during the Soviet period came to be G. M. Vasilevich, who documented the Ewenki lexicon in a series of bilingual dictionaries (1934a, 1940a, 1948a, 1958, posthumously 2005). She published also a textbook (1934b), a grammar (1940b), and collections of folklore (1936, 1966). Other bilingual dictionaries include those by Gorcevskaya & Kolesnikova & Konstantinova (1958), Kolesnikova & Konstantinova (1960), Kolesnikova (1983), and Boldyrev (1988, 1994), while grammatical descriptions have been published by Konstantinova (1964, 1968) and Cincius (posthumously 1997). Ewenki syntax is described on a monographic level in Kolesnikova (1966), complemented by Rudnickaya (2019) for the modern spoken language. Of other grammatical works the monographs by Gorelova (1979) on aspect, Khasanova (1983) on imperatives, and Brodskaya (1988) on complex sentences may be mentioned. Varlamova (1986) is a monograph on Ewenki phraseology.

Siberian Ewenki 145

Modern synchronic descriptions of Ewenki in the Western languages are still scarce. The first major summary, including diachronic notes, was compiled on the basis of secondary sources by K. H. Menges (1968). A more comprehensive description with a typological orientation, based both on primary and secondary sources, was attempted by Igor Nedjalkov (1997), but the result is somewhat unsatisfactory because of the misleading transcription and the user-hostile format of the series; see, however, the more specialized papers on Ewenki morphology by him (Nedjalkov 1994, 1995), also with V. P. Nedjalkov (1988ab). A user-friendly, though concise, general grammar is Bulatova & Grenoble (1999). Janhunen (1991) is focused on Khamnigan Ewenki, on which see also Doerfer (1985b). Work on Ewenki dialectology, begun by Schiefner (1859), was continued during the early Soviet period with a focus on the Barguzin dialect of Transbaikalia by Nicholas Poppe (1927) and Elbeg-Dorji Rinchino, as published by Kotwicz (1953). An early general survey of all dialects of Siberian Ewenki was published by Vasilevich (1948b). The Ewenki dialects of Yakutia and adjacent regions were extensively documented by Romanova & Myreeva (1962, 1964), who also published a dialectological dictionary (1968). Andreeva (1988) is a phonetic study of another dialect spoken in Yakutia. The two volumes by Bulatova (1987, 1999) offer a documentation of the dialects of the Amur region and Sakhalin. The dialects of the Amur Ewenki are also in the focus of the phonetic research of Morozova & al. (e.g. 2019), who have additionally studied issues of linguistic interference and code switching, cf. e.g. Procukovich & Ivanashko & Morozova (2018). The general situation of language maintenance among the Siberian Ewenki is discussed by Atknine (1997). Ethnonymic issues are discussed in Helimski & Janhunen (1990) and Pevnov (2020). Much of the Russian work on Ewenki both during and after the Soviet period has been of a documentary and/or practical nature. A wider areal perspective was, however, introduced by Romanova & Myreeva & Barashkov (1968) in their analysis of the interaction of Ewenki and Yakut. In a similar framework, though with a more advanced comparative apparatus, Khabtagaeva (2010, 2017) has analysed the contacts of Ewenki with Buryat and Khamnigan Mongol. A comprehensive modern analysis of all Ewenki dialects still remains to be done, but for the purposes of historical and areal comparisons with Mongolic the pioneering work of Doerfer (1985a) retains its value as an unsurpassed milestone. Apart from the many collections of Ewenki folklore, as published by Vasilevich (see above), M. G. Voskoboinikov (1958, 1960, 1967), Myreeva (1990, 2013), Romanova & Myreeva (1971), and Varlamova & Myreeva (2014), accompanied by an onomastic catalogue by Varlamova & al. (2018), a considerable amount of textual material on Ewenki has appeared in educational and literary contexts. The Ewenki literary language functioned initially (1931–1937) in the so-called Unified Northern Alphabet (Edinyi severnyi alfavit), based on Roman letters, but it was re-established after the war on a Cyrillic basis, as first extensively described by Konstantinova & Lebedeva (1953). In the postwar period several Ewenki intellectuals, with Alitet Nemtushkin (1939–2006) as the leading personality, have distinguished themselves as professional writers. One of the largest samples of Ewenki running text is the recently published “Bible for Children”, translated by Bulatova & Myreeva (2011). Ewenki textbooks for schools have a history dating back to the earliest years of the literary language. It has to be noted, however, that the language in these publications represents a partially artificial norm and bears particularly many traces of interference with Russian.

146 Juha Janhunen

SEGMENTAL STRUCTURE The prewar Roman-based orthography of Siberian Ewenki was more or less phonemic and used one letter for each phoneme of the language. Distinctions that could not be written with the resources of the standard Roman alphabet were expressed by using special symbols and diacritically modified letters, which were also employed with similar phonetic values for the other “northern” languages of the Soviet Union. The postwar Cyrillic orthography, as still used today, is, by contrast, language specific and involves conventions based on an interpretation of Ewenki sounds in terms of their Russian phonetic and graphemic equivalents. Many of these conventions are in contradiction with the segmental structure of Ewenki, which means that Ewenki phonology cannot be adequately described with the Cyrillic orthographical values or their Romanizations (as is done by Nedjalkov). In the present description, the basic Roman letters are used to represent the segments and sequences of segments of Ewenki with no direct relation to the conventions applied in the Cyrillic orthography. Regular Siberian Ewenki has a relatively simple system of segmental phonemes with 5 short and 6 long vowels, as well as 18 consonants. The short vowels are organized in an almost prototypical triangle (Table 7.1), in which the rounded back vowels u [u] and o [o ~ ɔ], the unrounded low back vowel a [a ~ ɑ], and the unrounded high front vowel i [i] are within their expected “normal” ranges of variation. The mid-high unrounded vowel e is, however, centralized and may be described as having the “reduced” value [ə], though in the Transbaikalian dialects, especially in Khamnigan Ewenki, the latter can undergo rounding yielding the value [ɵ]. This means that, synchronically, Ewenki shows the asymmetrical situation of having only one front (palatal) vowel against four back (velar) vowels. TABLE 7.1 SIBERIAN EWENKI SHORT VOWELS u o

 

i

e a

Historical and comparative evidence suggests that this simple system is the result of recent mergers. Proto-Ewenic had a system of 7 to 8 distinct vowels, as still synchronically preserved in Solon, Orochen, Neghidal, and Ewen. In this larger system, the modern Ewenki high vowels u and i were represented by the pairs [ʊ] vs. [u] and [ɪ] vs. [i], respectively, with the lower members [ʊ ɪ] standing for the Proto-Tungusic back vowels *u *ï and the higher members [u i] standing for the Proto-Tungusic front vowels *ö *i. It seems that, in Ewenki, the distinction [ɪ] vs. [i] was lost first, while the distinction [ʊ] vs. [u] has survived until very recently and is still present in at least some forms of Khamnigan Ewenki. The distinction between two kinds of u, the one velar and the other centralized, is systematically indicated in many older descriptions of Ewenki (notably Castrén and Vasilevich), though in most cases it is a question of combinatory variants conditioned by the rules of vowel harmony. In most forms of modern Siberian Ewenki, however, the distinction seems to have lost its phonemic role, and it is also not indicated in the orthography of the Ewenki written language. The long vowels, here written with the digraphs uu oo aa ii ee, have qualities more or less identical to their short equivalents. The paradigm of the long vowels has, however, one additional member, here written with the digraph ie, pronounced as a long mid-high

Siberian Ewenki 147

unrounded front vowel [e:]. This vowel, which represents the Proto-Tungusic “diphthong” conventionally reconstructed as *ia or *ïa, repairs the symmetry of the paradigm and raises the total number of distinctive qualities to six (Table 7.2). This also places Ewenki in the company of several other Northeast Asian languages whose paradigm of long vowels is larger than that of the corresponding short vowels. Technically, this situation could mean that the long vowels are less marked and, hence, more basic, than their short counterparts. TABLE 7.2 SIBERIAN EWENKI LONG VOWELS uu

 

ii

oo

ee

ie

aa

From the point of view of segmental phonology, the status of the long vowels in Ewenki is somewhat ambiguous. Although it would be convenient to analyse them as sequences of two separate vowel segments, with ie as a diphthongoid sequence, there is no synchronic evidence suggesting that the components of these sequences can function as separate elements with, for instance, independent morphological roles. In this respect, Ewenki differs from the neighbouring Mongolic languages, especially Khamnigan Mongol, in which long vowels can often be morphologically segmented. Also, vowel length in Ewenki shows occasional dialectal and positional lability, in that a long vowel in one dialect or one position can be represented as a short vowel in another dialect or another position. Even so, in most cases vowel length is stable, allowing it to be viewed as an unambiguous lexically determined property of those morphemes in which it is present. The consonant system of Siberian Ewenki (Table 7.3) retains the basic features of Proto-Tungusic and comprises 18 members, representing four distinct places of articulation: labial, dental, palatal, and velar. In terms of manner of articulation, the consonants are divided into nasals (m n ñ ƞ). weak stops (b d j g), strong stops (p t c k), fricatives (s x), glides (w y), and liquids (l r). TABLE 7.3 SIBERIAN EWENKI CONSONANTS m

n

ñ

ƞ

b

d

j

g

t

c

k

p

s w

x y

l r

The nasals, glides and liquids are inherently voiced, while the strong stops and fricatives are inherently unvoiced. The weak stops are also typically pronounced with voice in Siberian Ewenki, but in dialects spoken in the vicinity of Mongolic they can lose voicing, in which case the strong stops are pronounced with aspiration. The palatal stops c j [c ɟ] can also be realized as palato-alveolar or alveolo-palatal affricates with a sibilant release [tʃ dʒ] ~ [tɕ dʑ]. It may be noted that, except for the factor of voicing, the fricatives and

148 Juha Janhunen

glides are complementary and form a set corresponding to the four places of articulation (w s y x), with the velar fricative x being most often realized as a laryngeal [h]. As far as liquids are concerned, Ewenki is a two-liquid language, with both a lateral l [l] and a vibrant (trill) r [r]. There are several combinatory phenomena that affect the distribution of the consonants. Most importantly, only one set of stops occurs in syllable-final position, with the additional exclusion of the palatals. Assuming that the weak stops are less marked than the strong stops, all syllable-final stops could be identified as representing the weak set b d g, e.g. gugda [gugda] = /gugda/ ‘high’ vs. jikte [ɟiktə] = /jigte/ ‘berry’. However, for reasons of convention, this analysis will not be adopted here. The segments g b undergo postvocalic spirantization, which in the case of g is phonetic, e.g. buga [buɣa] ‘earth’, but which in the case of b is probably better understood as phonemic, e.g. *jabï ‘boat’ > jaw [ɟaβ] ~ [ɟaw]. A contrast between the segments b and w is present in word-initial position, e.g. baa- ‘to be unwilling’ vs. waa- ‘to kill’, while in postvocalic position the contrast is, in the framework adopted here, between p and w, e.g. ƞene- ‘to go’ : AOR 1PL.INCL ƞene-re-p vs. EXCL ƞene-re-w. From the diachronic point of view it may to be noted that due to the development of *p to *x in Pre-Proto-Ewenic, the synchronically attested strong initial p is a relatively new phoneme that has gained ground in loanwords and descriptive vocabulary mainly after the Proto-Ewenic period. The velar fricative x, on the other hand, can also represent an original *s, and due to dialectal differences in the developments *x > Ø and *s > x, its frequency and distribution in the lexicon can vary considerably. The vowels i ii ie tend to have a palatalizing effect on a preceding consonant, e.g. biega [bje:ɣa] ‘moon’. This effect is not present in the dental stops t d, however, allowing them to remain phonetically distinct from the corresponding palatal stops (affricates) c j. To enhance this distinction, the vowels i ii can have slightly velarized allophones after t d, as in timanii ‘tomorrow’, digin ‘four’, a combinatory property which has resulted in the confusing convention of i being written as a “back-i” (the letter 〈ы〉 ery) in the Cyrillic orthography. This velarized allophone of i should, however, not be mistaken as a trace of an original Proto-Ewenic velar *ï, for it is observed also in cases of an original palatal *i (< *i & *ü). PHONOTACTICS AND MORPHOPHONOLOGY Phonotactically and morphotactically, Ewenki remains a fairly typical “Ural-Altaic” language with no initial or final consonant clusters, with no lexically distinctive prosodic properties (except vowel length), with a well-developed suffixal morphology, and with a progressive vowel harmony enhancing the internal coherence of phonological words. Stress is basically located on the initial syllable, though long vowels in non-initial syllables are often accompanied by additional expiratory prominence, and the final syllable can be marked by a rise in pitch. An areally conditioned “Ural-Altaic” feature is the absence of the vibrant r in initial position, though initial l is well attested, e.g. lukii ‘arrow’. Like several other languages in the region, Siberian Ewenki retains the velar nasal ƞ in initial position, e.g. ƞoo ‘smell’. Unlike the mainstream dialects, Khamnigan Ewenki has, however, replaced initial ƞ with n (as in Solon), e.g. *ƞii ‘who’ > nii. In some words (with an internal nasal), ƞ can also be lost dialectally in Khamnigan Ewenki, e.g. *ƞene- ‘to go’ > (Borzya) nene- vs. (Urulyunggui) ene-. There are no major differences in the phonological appearance of different classes of words, except that a few deictic and auxiliary roots can have the simple structure (C)

Siberian Ewenki 149

V(-), e.g. 1SG bi ‘I’, AUX NEG e- ‘not’. Regular lexemes are functionally divided into nominals and verbals, of which only nominals can be used alone, while verbals always require the presence of one or more suffixes. Most nominals and verbals have the bisyllabic structure (C)V(V)(C)CV(-), e.g. buga ‘earth’, garpa- ‘to shoot’, but there are also monosyllabic vowel stems of the type (C)VV(-), e.g. jaa ‘kin’, yuu- ‘to come out’. Monosyllabic consonant stems of the type (C)V(V)C(-) are mainly secondary and have been formed by the loss of the final vowel i (< *ï *i). The original vowel is still present in the paradigm when phonotactically required, and synchronically it could be analysed as a non-lexical epenthetic connective vowel, as in *dili (*dïlï) > dil ‘head’ : PX1SG dil.i-w (< *dïlï-bï) ‘my head’. After the labial glide w (< *b) the connective vowel is u, e.g. waa‘to kill’ : PASS waa-w- ‘to be killed’ : AOR 1SG waa-w.u-m. There are, however, also a few examples of primary monosyllables with a final consonant, e.g. guun- ~ gun- ‘to say’ (< *göön-), Khamnigan Ewenki (Borzya) juug [dʑu:k] ‘dwelling’ vs. regular Siberian Ewenki juu (< *juug). A special class of lexemes is formed by those ending in the primary nasal /n, which represents the original Proto-Tungusic final nasal *n. This class comprises both nominals and verbals, some of which are monosyllabic, like guun- ~ gun- ‘to say’, while the majority are bisyllabic, in which case the final nasal may also be understood as an obscured derivational element, as in oron = oro/n = oro(-)n ‘reindeer’. In many cases, the derivational relationship is synchronically transparent, as in uku- ‘to suck’ : uku-n ‘female breast’. This type also comprises several examples of ambivalent nomina-verba, that is, stems that can function both as nominals and verbals, e.g. beyuu- ‘to hunt’ : beyuu-n- ‘to hunt’ : beyuu-n ‘deer’. Nomina-verba in Ewenki are relatively rare, but cases like tigde ‘rain’ : tigde- ‘to rain’ show that vowel stems can also be comprised by this phenomenon. Vowel harmony in Siberian Ewenki represents a direct development of the primary Proto-Tungusic palato-velar harmony combined with a secondary Proto-Ewenic labial harmony. Due to the rotation of the vowel system, the palato-velar oppositions initially developed into verticalized pairs of higher (originally palatal) and lower (originally velar) members, as still synchronically observed in Solon, Orochen, Neghidal, and Ewen. The subsequent mergers of the high vowels have, however, simplified the system, leaving only the three non-high vowels a o e as harmonically active, while the two high vowels u i are harmonically inactive or neutral. The harmonically active vowels are mutually exclusive, but all of them can be combined with the neutral vowels, except that labial harmony permits o in a non-initial syllable only if the preceding syllable also has o. This means that a non-high vowel of the initial syllable can only be followed by an identical non-high vowel, or by a high vowel, while a high vowel of the initial syllable can be followed by either another high vowel or an unrounded non-high vowel. Considering only the short vowels of the first two syllables, there are altogether 17 permitted vowel combinations (Table 7.4). TABLE 7.4 SIBERIAN EWENKI VOWEL COMBINATIONS 1st o

2nd syllable o

a u i e

a

u e

i

150 Juha Janhunen

The long vowels have a basically similar harmonic status, with the exception that oo of the initial syllable is combined with a (and not o) in the following syllable, e.g. ooran ‘rapids’. Also, in many Siberian Ewenki dialects, but not, for instance, in Khamnigan Ewenki, the non-high vowel following aa is pronounced with the “reduced” quality [ə], making it phonetically similar to e, as in ƞaala [ŋɑ:lə] ‘hand’. It happens that this feature is incorporated into the normative orthography of Ewenki, but even so, there is no reason to regard it as phonologically relevant, for the low vowels a and e do not contrast in this position. It may be noted that there is no similar “reduction” of a after ie, which otherwise follows the harmonic pattern of aa, e.g. mieta [mje:ta] ‘head skin (of deer)’. The loss of harmonic oppositions in the high vowels has had important consequences for the synchronic functioning of vowel harmony in Ewenki. The harmonically alternating non-high vowels a e, whose distribution in non-initial syllables was originally determined by the rules of vowel harmony, have become lexically fully distinctive when following u or i, as in turga ‘encounter’ vs. turge ‘support’, ila- ‘to set on fire’ vs. ile- ‘to lick (of a dog)’. If we assume that the harmonic status of a word is determined by the vowel of the initial syllable, then we have to postulate an abstract deep level on which Ewenki still retains the etymologically conditioned distinction between two kinds of u and two kinds of i. This analysis is supported by the fact that roots containing only u or i are harmonically ambiguous and can take either a aa or e ee in suffixal syllables, e.g. surii ‘pike perch’ : ACC surii-wa, ulukii ‘squirrel’ : ACC ulukii-we. In such cases, the harmonic status of the root must be assumed to be lexically determined. This raises the question concerning the phonological status of labial harmony in Ewenki. The non-high rounded vowels o oo can occur in non-initial syllables only after a short o in the previous syllable, e.g. olo- ‘to be frightened’ : CAUS olo-pkoon- ‘to frighten’. A word can contain several syllables with o, but only if the first syllable also has o, e.g. onokto ‘textile’ : ADJ onokto-mo ‘(made of) textile’. In this environment, neither a nor e is permitted. When a neutral vowel intervenes in a word with o in the first syllable, labial harmony stops working and a appears instead of o, e.g. omolgii ‘young man’ : ACC omolgii-wa ‘boy’. All of this suggests that o in non-initial syllables is simply an allophone of a, meaning that labial harmony is merely a phonetic phenomenon. However, following the established convention, labial harmony will be incorporated in the notation used for Ewenki in the present description. Apart from vowel harmony, Ewenki has rather few segmental alternations. Most importantly, the diachronic loss of the high vowel *i in non-initial syllables, combined with the rules of phonotactics, which do not permit syllable-final consonant clusters, condition the alternation of i with zero (Ø) not only at the end of stems, as in dil (< *dïlï) ‘head’ : PX1SG dil.i-w (< *dïlï-bï) ‘my head’, but also within several suffixal morphemes. At the same time, the rules of consonant distribution, which neutralize all syllable-final stops, condition alternations like t : j, as in xunaat (< *xunaajï) ‘daughter’ : PL xunaaj.i-l. Still other morphophonological alternations are caused by the behaviour of suffix-initial w (< *b) d r s in cases like ACC -wA : -mA and AOR -rA- : -nA- (nasalization after a stem-final nasal), DAT -duu : -tuu (strengthening of weak obstruents after a stem-final obstruent), AOR -rA : -dA : -tA : -lA (loss of vibration after a stem final consonant), and PX2SG -si : -ni : -li : -ri (progressive assimilation to the quality of the preceding consonant). WORD FORMATION Ewenki is a language with a relatively high degree of morphological synthesis, realized by attaching both derivational and inflectional suffixes to nominal and verbal roots. The

Siberian Ewenki 151

principal difference between nominals and verb(al)s, apart from their inherently different semantic and syntactic roles, is that they take different sets of morphological markers. Nominals in this context include also the special classes of spatials, numerals, pronouns, and adjectives. By definition, derivational suffixes form secondary stems, which can take the same inflectional suffixes as primary lexical stems. Considering the division of regular lexemes into nominals and verb(al)s, derivational suffixes can be divided into four types depending on whether they derive (1) denominal nominals, (2) deverbal nominals, (3) denominal verbs, or (4) deverbal verbs. The derivation of verbs from nominals may also be termed “verbalization”, while the derivation of nominals from verbs may be termed “nominalization” (in the basic meaning of the term). The level of productivity in the different categories of derivation varies greatly, but, in general, derivational suffixes are often not fully productive and tend to yield lexicalized meanings. A special category of nominalized verbs is formed by the participles (as discussed separately further below), which are typically fully productive and retain a number of verbal features in their syntactic and semantic behaviour. (1) Denominal nominals: Most suffixes deriving nominals from nominals produce only a limited number of lexicalized derivatives, though the degree of their productivity varies considerably. Since they do not change the grammatical status of the underlying nominal root, they are mainly of a lexicological interest. Examples: • • • • • • • • •

-gAAn for inhabitants of a region, e.g. bira ‘river’ : bira-gaan ‘river dweller’, kamnii ‘mountain foot, narrow lowland’ : kamnii-gaan ‘mountain-foot dweller’ (from which Khamnigan); -gii-dAA for orientational nouns from spatial roots, e.g. julee ‘front’ : julee-giidee ‘front side’, ƞie ‘lower part’ : ƞie-gii-daa ‘lower course’ (from which PL ƞie-gii-daa-l ‘Neghidal’); -kAAn for general diminutives, e.g. bira ‘river’ : DIM bira-kaan ‘small river’, beye ‘person’ : DIM beye-keen ‘doll’; -kuun ~ -kAA-kuun for augmentatives, e.g. beye ‘person’ : AUGM beye-kuun ‘big/real person’, mooti ‘elk’ : AUGM mooti-kaa-kuun ‘big elk’; -ƞAsA for deceased relatives, e.g. akii ‘elder brother’ : akii-ƞasa ‘late elder brother’, enii ‘mother’ : enii-ƞese ‘late mother’; -ruk ~ -luk for holders of objects, e.g. ulle ‘meat’ : ulle-ruk ‘vessel for meat’, tewu(u)l ‘load’ : tewu(u)l-luk ‘bag for keeping foodstuff’; -sAAk for conglomerations of objects, e.g. guule ‘house’ : guule-seek ‘village’, ure ‘mountain’ : ure-seek ‘mountainous locality’; -sik for items of clothing, e.g. tuge ‘winter’ : tuge-sik ‘winter clothes’, samaan ‘shaman’ : samaa-sik ‘shaman’s clothing’; -t-kAAn for diminutives of animate beings, e.g. beye ‘person, man’ : DIM beyetkeen ‘boy’, ollo ‘fish’ : DIM ollo-tkoon ‘small fish’.

Ewenki also retains many examples of obscured Proto-Tungusic denominal derivatives, which include collective countables in -k.tA, e.g. oosi-kta ‘star/s’, ñuri-kte ‘hair/s’, as well as uncountables (terms for homogeneous masses) in -k.sA, e.g. jaliksa ‘saliva’, sile-kse ‘dew’. The suffix -k.sA can also have productive uses when attached to animal names, e.g. sulakii ‘fox’ : sulakii-ksa ‘fox as substance: fox meat, fox skin’.

152 Juha Janhunen

(2) Deverbal nominals: Apart from the participles, which remain verb(al)s as far as their arguments are concerned, there are several other more or less productive types of deverbal nominals. These are synchronically full nominals with no verbal features. Examples: • • • • • • • •

-ktA for results of action, e.g. gira- ‘to step’ : gira-kta ‘step’, guun- ‘to say’ : guun.i-kte ~ guu-kte ‘expressed idea’; -jAAk for locations of former actions, e.g. bi- ‘to be’ : bi-jeek ‘a place where one was before’, kiƞlee- ‘to ski’ : kiƞlee-jeek ‘a place where somebody has gone on skis, ski track’; -kiit for general actions or locations of action, e.g. in- ‘to live’ : in-kiit ‘life’, ayii- ‘to cure’ : ayiit-kiit ‘hospital’; -lAAn for skilful performers of action, e.g. ikee- ‘to sing’ : ikee-leen ‘skilful singer’, waa- ‘to kill, to hunt’ : waa-laan ‘good hunter’; -mAktA for performers of recently completed actions, e.g. baldi- ‘to be born’ : baldi-makta ‘newborn’, yuu- ‘to come out’ : yuu-mekte ‘a bear that has just come out of its den’; -mnii ~ -mnie for performers of actions, e.g. elge- ‘to lead’ : elge-mnii ‘guide’, but also for female members of an ethnic group, e.g. luca ‘Russian’ : luca-mnii ‘Russian woman’; -ptin for results or instruments of action, e.g. dal- ‘to close’ : dal.i-ptin ‘lid, cover’, xaaw- ‘to add’ : xaaw.u-ptin ‘addition’; -wuun for results or instruments of action, e.g. taƞ- ‘to count’ : taƞ.i-wuun ‘count, number’, soko- ‘to scoop’ : soko-wuun ‘scoop’.

All of the above-listed suffixes are of a complex origin and, hence, diachronically secondary. There are also a couple of formally simple and apparently primary suffixes of this type, both of which yield deverbal nouns with a diffuse general meaning: • •

-n, e.g. iñe- ‘to laugh’ : iñe-n ‘laughter, joke’, ƞene- ‘to go’ : ƞene-n ‘going, movevement, track’; -r, e.g. ew- ‘to descend’ : ew.u-r (also ew.u-rii) ‘descent, slope’, jogo- (also jogo-r-) ‘to be in need’ : jogo-r ‘need’.

(3) Denominal verbs: This category comprises a variety of more or less lexicalized items, most of which express a factitive, captative, or instrumental relationship to the underlying nominal stem. Examples: • • • • •

-dAA- for instrumentals, e.g. jal ‘thought’ : jal-daa- ‘to think’, ngaala ‘hand’ : ngaala-daa- ‘to grasp (with hands)’; -lA- ~ -lAA- ~ -lii- for captatives, e.g. jikte ‘berry’ : jikte-le- ~ jikte-lee- ‘to pick berries’, muu- ‘water’ : muu-lee- ‘to bring water’, ñute ‘resin (of larchtree)’ : ñute-lii- ‘to collect resin’; -lAA- for instrumentals, e.g. asii ‘woman’ : asii-laa- ‘to marry a woman’, buse ‘belt’ : buse-lee- ‘to put a belt on’, but also translatives, e.g. xegdi ‘big’ : xegdi-lee- ‘to grow big’; -mA- ~ -mAA- ~ -mAAt- ~ -mii- for captatives, especially of hunting and fishing, e.g. ollo ‘fish’ : ollo-mo- ~ ollo-moo- ~ ollo-moot- ~ ollo-mii- ‘to fish’, ulu-kii ‘squirrel’ : ulu-mee- ‘to hunt for squirrels’; -mu- for odoratives (‘having a certain smell’), e.g. ulle ‘meat’ : ulle-mu- ‘to smell like meat’, cuuka ‘grass’ : cuuka-mu- ‘to smell like grass’;

Siberian Ewenki 153

• • • •



-ƞ- for factitives, e.g. jaw ‘boat’ : jaw.i-ƞ- ‘to make a boat’, sirba ‘fish soup’ : sirba-ƞ- ‘to cook fish soup’; -tAA- for possessives (‘having something’), e.g. juu ‘dwelling’ : juu-taa- ‘to inhabit (a dwelling)’, gerbii ‘name’ : gerbii-tee- ‘to bear a name’; -ti- for consumatives (‘eating or drinking something’), e.g. ulle ‘meat’ : ulle-tii‘to eat meat’, cay- ‘tea’ : cay-tii- ‘to drink tea’; -w.u- after vowels ~ -mu- (< *-bO-) after nasals for translatives and factitives, e.g. xegdi ‘big’ : xegdi-w- ‘to grow big’, sienkaan ‘ear mark’ : sienkaan-mu- ‘to mark the ear (of a reindeer)’.

There is also a suffix which functions as a more general verbalizer with no specific meaning: •

-l-, e.g. gowost ‘nail’ : gowost.i-l- ‘to nail’, xekuu ‘warm’ : xekuu-l- ‘to warm’, umu-kta ‘egg’ : umu-l- ‘to lay eggs’.

(4) Deverbal verbs: Compared with the other categories of derivational forms, deverbal verbs are more intimately connected with grammaticalized functions, including modal and aspectual distinctions, as well as voice. Some of these forms have a relatively high degree of productivity. Suffixes indicating modal and aspectual distinctions include: • • • • • • • • • •

-ktA- for distributives (repeated action in time and space), e.g. soƞo- ‘to cry’ : soƞo-kto- ‘to cry all the time’, tuksa- ‘to run’ : tuksa-kta- ‘to run around’; -jA- after sonorants ~ -cA- after obstruents for ongoing action (progressive aspect), e.g. dukuu- ‘to write’ : dukuu-ja- ‘to be writing’, icet- ‘to watch’ : icet-ce- ‘to be watching’; -l- for inchoative action, e.g. jew- ~ jep- ‘to eat’ : jew.u-l- ‘to start eating’, songo- ‘to cry’ : songo-l- ‘to begin to cry’; -mAlcA- for momentaneous action, e.g. ga- ‘to take’ : ga-malca- ‘to grasp suddenly’, yuu- ‘to come out’ : yuu-melce- ‘to come out suddenly’; -mkA- for tentative action, e.g. arca- ‘to meet’ : arca-mka- ‘to go in order to meet’, tet- ‘to put on (clothes)’ : tet.i-mke- ‘to try on’; -mu- for desideratives (‘wanting to do’), e.g. aa- ‘to sleep’ : DESID aa-mu- ‘to want to sleep’, jew- ~ jep- ‘to eat’ : DESID jem-mu- ‘to want to eat’; -nAA- for andatives (‘going to do’), e.g. dukuu- ‘to write’ : AND dukuu-naa- ‘to go to write’, tege- ‘to sit down’ : AND tege-nee- ‘to go to sit down’; -ƞnA- for habitual action, e.g. bi- ‘to be’ : bi-ƞne- ‘to be habitually’, minee- ‘to cut’ : minee-ƞne- ‘to cut habitually’; -sin- for semelfactive action, e.g. aa- ‘to sleep’ : SEM aa-sin- ‘to fall to sleep’, tuksa- ‘to run’ : SEM tuksa-sin- ‘to run away’; -t- : -c.i- (< *-cI-) for continuous action, e.g. ice- ‘to see’ : ice-t- ‘to watch’, tege‘to sit down’ : tege-t- ‘to sit’.

Suffixally indicated voice distinctions in Ewenki correspond to the causative, reflexive, passive, reciprocal, and collective functions, some of which are formally interconnected: • •

-pkAAn- for causatives, e.g. ii- ‘to go’ : CAUS ii-pkeen- ‘to force to go’, baka‘to find’ : CAUS baka-pkaan- ‘to force to find’; -ldi- ~ -ldii- for cooperatives, e.g. baka- ‘to find’ : COOP baka-ldi- ‘to see each other, to meet’, ƞene- ‘to go’ : COOP ƞene-ldi- ‘to go together’;

154 Juha Janhunen

• • •

-mAAt- : -mAAc.i- for reciprocals, e.g. guun- ‘to say’ : RECIPR guu-meet‘to say to each other, to make an agreement’, ƞoorca- ‘to wrestle’ : RECIPR ƞoorca-maat- ‘to wrestle with each other’; -rgA- for anticausatives, e.g. sugca- ‘to break’ : ANTICAUS sukca-rga- ‘to be broken’, eti- ‘to tear’ : ANTICAUS eti-rge- ‘to be torn’; -w.u- after vowels ~ -mu- (< *-bO-) after nasals for causatives from intransitives, e.g. eme- ‘to come’ : CAUS eme-w- ‘to bring’, and reflexive-passives from transitives, e.g. sokor- ‘to loose’ : REFL sokor.i-w- ‘to be lost’, waa- ‘to kill’ : PASS waa-w- ‘to be killed’, CAUS eme-w- ‘to bring’ : CAUS-PASS emew.u-w- ‘to be brought’.

It has to be added that although most derivational suffixes are specific to either nominals or verb(al)s, some suffixes are ambivalent and can be attached to both classes of words, an example being -w.u- ~ -mu-, which forms both factitives and translatives from nouns and causatives and reflexives from verbs. Even so, in most cases the derivational relationship is unambiguous. For instance, the verb xulda-w- ‘to cover oneself’ must be based on the verbal stem xulda- ‘to cover’, rather than the homonymous nominal stem xulda ‘cover’. The high degree of lexicalization of the Ewenki derivational forms is also observed in certain specific classes of words. Spatials, numerals, and adjectival nominals all have specific patterns of derivation, most of which are non-productive. Adjectival nominals, for instance, have often idiosyncratic sets of correlative derivational forms, as in xula-ma ‘red’ : xula-r.ii-n ‘red’ : xula-r.ga- ‘to be(come) red’ : xula-taa- ‘to be red (the sky)’. There is, however, a productive set of suffixally marked adjectival forms which, with some reservations, can be called forms of comparison. The suffixes are -dimAr ~ -timAr ~ -tmAr for the comparative and -dig ~ -diguu ~ -tiguu ~ -tkuu for the superlative, e.g. ƞoonim ‘long’ : COMP ƞoonim-dimar ‘longer’ : SUPERL ƞoonim-diguu ‘longest’, aya ‘good’ : COMP aya-tmar ‘better’ : aya-dig ~ aya-tkuu ‘best’. In both the comparative and the superlative construction, the base of comparison (‘than’, ‘of all’) stands in the ablative case. NUMBER AND CASE Siberian Ewenki has a rather typical North Asian nominal paradigm with an elaborate system of suffixally marked case forms and a likewise suffixally marked plural. Unlike the situation in several other languages further to the south and east, including also some varieties of Tungusic, plural marking in Siberian Ewenki is more or less obligatory, which means that it may be identified as belonging to the realm of inflectional morphology, rather than derivation. In the chain of morphemes, the plural marker (PL) follows derivational suffixes (DX) and precedes case endings (CX), to which possessive and reflexive suffixes (PX/RX) as well as clitics (CL) can also be added (N-DX-PL-CX-PX/RX=CL). Like many other languages with obligatory plural marking, Siberian Ewenki has basically a single plural marker, which is -l, added as such to stems ending in a vowel, e.g. bira ‘river’ : PL bira-l, and with the mediation of the connective vowel i to stems ending in a consonant, e.g. bur (< *buri) ‘island’ : PL bur.i-l (< *buri-l). The marker -l in combination with the connective vowel i is also used on the remaining primary consonant stems, notably juu ~ juug(-) ‘dwelling’ : dialectal PL juug.i-l. However, stems ending in the nasal /n take in the plural the marker -r, which replaces the final nasal, e.g. muri/n ‘horse’ : PL

Siberian Ewenki 155

muri-r, though by analogy the marker -l can dialectally also be used in these cases, i.e., PL muri-l. The distribution of the markers -l and -r derives from Proto-Tungusic and is not phonologically motivated in the modern language (nor in Proto-Tungusic), which means that the marker -r cannot be viewed as a simple allomorph of -l. Rather, there are two classes of nouns: those ending in n, which take the plural marker -r, and those ending in any other segment, which take the plural marker -l. From this point of view, the final n of nominal stems can be synchronically analysed as a separate morphological element, which, though not a singular marker, implies the singular number and has normally no other lexical meaning. In several Ewenic idioms, notably Solon, Orochen and Neghidal, as well as in the other Tungusic languages of the Manchurian region, the plural is also marked by the complex suffix (*)-sA-l, which contains the plural marker -l in combination with the primary collective marker (*)-sA. In regular Siberian Ewenki, -sA-l is attested only in a few lexicalized examples denoting relatives, e.g. akii ~ akiin ‘elder brother’ : COLL-PL akii.n.a-sa-l ‘elder brothers’. Kinship terms also have other types of irregular plural forms, including amiin ‘father’ : PL am-t.ii-l and eniin ‘mother’ : PL en-t.ii-l, as well as akiin ‘elder brother’ : PL ak-n.ii-l, ekiin ‘elder sister’ : PL ek-n.ii-l, nekun ‘younger sibling’ : PL nek-n.ii-l. The case system of Siberian Ewenki is relatively rich, with some 10 to 15 separate cases, depending on the synchronic analysis, and with some variation between the dialects (Table 7.5). From the functional point of view, the cases may be divided into three categories: (1) grammatical, (2) local, and (3) modal cases. As in other languages of the “Ural-Altaic” type, the argument structure in Ewenki follows the nominative-accusative principle, with an unmarked nominative for the subject and a marked accusative for the direct object. Unlike several adjacent languages, however, Ewenki does not use the unmarked nominative to indicate an indefinite, unspecific or generic direct object. Instead, for this purpose a special partitive (“indefinite accusative”) case is used, which also expresses partial objects. Another peculiarity is that Siberian Ewenki has no actual genitive case. (1) The two suffixally marked grammatical cases are, consequently, the accusative and the partitive, both of which function as cases of the direct object (definite or total vs. indefinite, generic or partial). •

•

The accusative is marked by the suffix *-bA, realized as -wA after vowels and non-nasal consonants, e.g. bira ‘river’ : ACC bira-wa, moo ‘tree’ : PL moo-l : PL-ACC moo-l-wa, and as -mA after nasals, e.g. beyuun ‘deer’ : ACC beyuun-me. Given the essentially allophonic relationship of medial w [β] and b [b], the suffix can also have the shape -bA after non-nasal sonorants. At least dialectally, this can be strengthened to -pA after obstruents, e.g. aamut ‘lake’ : ACC aamut-pa, while after the velar stop k an assimilation to -kA is also possible, e.g. inmek ‘pack’ : ACC inmek-pe ~ inmek-ke. The partitive suffix has the form -yA after original vowel stems, including stems that have lost the final vowel, e.g. bira ‘river’ : PART bira-ya, aamut (< *aamuji) ‘lake’ : PART aamut-ya (< *aamuji-ya), and the form -A after original consonant stems, notably stems ending in the nasal n, e.g. oron ‘reindeer’ : PART oron-o. The variant -A is also used after the plural markers -l and -r, e.g. PL-PART bira-l-a ‘rivers’, oro-r-o ‘reindeer’.

156 Juha Janhunen

(2) The local cases, which may also have occasional grammatical functions, form a rather heterogeneous system of six to eight different forms, most of which have cognates also outside of the Ewenic branch. Functionally, four different types of spatial relation are indicated: (a) location ‘at’ or movement ‘to’, (b) movement ‘from’, (c) movement ‘towards’, and (d) movement ‘along’ or ‘through’. Formally, some of the case markers are primary or “simple”, while others are secondary or “derived”. Several of the secondary suffixes are composed of a coaffix in combination with a primary case marker. The cases have established names, which are also used below. •

•

•

•

•



Two cases involve the element -du ~ -duu (< *-dOO), which alone marks the dative (‘at’, ‘to’) and in the extended form -duk : -duki- ~ -duuk : -duuki(< *-dOO-kI) the ablative (‘from’), e.g. beye ‘person’ : DAT beye-duu : ABL beye-duuk. After stem-final obstruents, the variants -tuu : -tuk ~ -tuuk are used, e.g. aamut ‘lake’ : DAT aamut-tuu : ABL aamut-tuuk, though the phonological status of this alternation remains somewhat unclear. After stems ending in w (= b), the variants -duu : -duk ~ -duuk are used in the normative language e.g. jaw ‘boat’ : DAT jaw-duu : ABL jaw-duuk. Two cases involve the element -l-, which in the form -lAA marks the locative and in the form -lii the prolative case, e.g. moo ‘tree’ : LOC moo-laa : PROL moo-lii. These suffixes are dialectally also taken by stems ending in a consonant, e.g. aamut ‘lake’ : LOC aamut-laa, but in the normative language the coaffixal element -du- ~ -tu-, both synchronically and diachronically identical with the dative case marker, is inserted between the stem and the actual case suffix, e.g. DAT-LOC aamut-tu-laa : DAT-PROL aamut-tu-lii. Two cases involve the coaffixal element -k- (< *-kI-) in combination with the locative and prolative markers -lAA : -lii. This coaffixal element is conventionally identified as adding a “directive” function, yielding the complex “directive-locative” and “directive-prolative” endings DIR-LOC -k-lAA and DIR-PROL -k-lii, respectively, e.g. aamut ‘lake’ : DIR-LOC aamuj.i-k-lAA : DIR-PROL aamuj.ik-lii. These same complex endings are also added to the plural forms in -l- and -r- with the help of the connective vowel i, e.g. oron ‘reindeer’ : PL oro-r : PL-DIR-LOC oro-r.i-k-lAA. The directive-locative and directive-prolative cases are both relatively rarely used and can even be absent dialectally, in which case they are replaced by the more basic endings -lAA and -lii, respectively. The actual directive case (‘towards’) is marked by the suffix -tkii after vowels and -tikii (< *-tIkII) after consonants, e.g. moo ‘tree’ : DIR moo-tkii, jaw ‘boat’ : DIR jaw-tikii. The postconsonantal suffix variant is also used after plural forms, e.g. oron ‘reindeer’ : PL-DIR oro-r-tikii. The elative case is marked by the suffix -giit : -giiji- ~ -git : -giji- (< *-gIIjI), after obstruents -kiit : -kiiji- ~ -kit : -kiji-, e.g. bira ‘river’ : ELAT bira-giit, aamut ‘lake’ : ELAT aamut-kiit. The elative marker consists diachronically of the spatial formative -gii- combined with the instrumental suffix *-ji > -t, which may explain why the elative implies a more diffuse place of origin (‘from the direction of’) than the ablative (‘from’).

The principal functional issue in the Ewenki system of local cases concerns the difference between the dative (also termed “dative-locative”) in -duu and the locative in -lAA. Obviously, in some contexts both cases can be used more or less synonymously. Even so—and in spite of the somewhat misleading terminology—a static location (‘where?’) is typically expressed by the dative, while the locative has a

Siberian Ewenki 157

dynamic reference (‘whither?’), e.g. aamut ‘lake’ : DAT aamut-tuu ‘on the lake’ : LOC aamut-laa ‘to the lake’. At the same time, the locative tends to be more limited to concrete locational (spatial) situations, while the dative also has a number of abstract, non-locational or grammatical uses, including temporal (‘when?’) and benefactive (‘for whom?’). Perhaps most importantly, the dative can express the agent of a passivized verb (‘by whom?’), but also the recipient (‘to whom?). (3) The category of modal cases comprises two functional entities, only one of which, the instrumental, has a uniform shape in all Ewenki dialects, while the other, the comitative, is represented by varying forms. •

•

The instrumental is marked by the suffix -ji after original final consonants and -t (< *-ji) after vowels, e.g. kalan ‘kettle’ : INSTR kalan-ji, ƞaala ‘hand’ : INSTR ƞaala-t. Stems that have secondarily lost the final vowel *i restore the vowel stem and take the suffix with the help of the connective vowel i, e.g. adil ‘net’ : INSTR adil.i-t (< *adïlï-jï). The comitative is marked by the suffix -nun, in the eastern dialects also -ñun (< *-ñuun), e.g. girkii ‘friend’ : COM girkii-nun ~ girkii-ñun. In several dialects throughout the Ewenki language area, however, this suffix is replaced by COM2 -gAlii, e.g. asii ‘wife’ : COM2-RX asii-galii-wii, after obstruents -kAlii, e.g. xunaat ‘girl’ : COM2 xunaat-kalii, and after nasals -ƞalii, e.g. akiin ‘elder brother’ : COM2 akiin-ƞalii.

The basic functional difference between the instrumental and the comitative is that the former typically refers to a passive instrument (‘by what?’), while the latter refers to an active participant (‘with whom?’). There are, however, occasional deviations from this pattern. TABLE 7.5 SIBERIAN EWENKI CASE MARKERS coaffix

DAT

case

V

ACC

-BA

PART

-yA

DAT

-Duu

ABL

-Duuk : -Duuki-

LOC

-lAA

LOC PROL

*Ci

*C

/n

-A

-Du-lAA -lii

DAT

PROL

DIR

LOC

-k-lAA

DIR

PROL

-k-lii

DIR

-tkii

ELAT

-Giid : -Giiji-

INSTR

-t : -ji-

COM

-Nun

COM2

-GAlii

-Du-lii

-tikii -ji -ƞAlii

Stem types: V = vowel stems, *C original consonant stems, *Ci secondary consonant stems, /n = primary nasal stems. Segmental alternations: A = a o e, B = w b p, D = d t, G = g k, N = n ñ.

158 Juha Janhunen

Two other forms need to be discussed in this connection, though their status with regard to the category of case remains controversial: •

•

Normative Ewenki descriptions often maintain that the language has a “genitive” case formed by the possessive marker POSS -ƞii ~ -ƞi. Indeed, this marker expresses ownership (‘whose?’) in independent and predicative positions and fills “genitival” functions in a loose sense. Grammatically, however, it is ambivalent, in that it can be attached to the plural ending like case suffixes, but it can also take a plural ending like derivational forms, e.g. kuƞakaan ‘child’ : PL kuƞakaa-r : PL-POSS kuƞakaa-r-ƞii ‘(the one thing) of the children’ : PL-POSS-PL kuƞakaa-r-ƞii-l ‘(the many things) of the children’. Moreover, it can take other case suffixes, e.g. atirkaan ‘old woman’ : POSS atirkaan-ƞii ‘the one belonging to the old woman’ : POSS-ACC atirkaan-ƞii-wa. Even so, its definition as a genitive case marker is probably correct for some forms of Ewenki, including, at least, Khamnigan Evenki. A somewhat similar situation is connected with the proprietive markers -cii ~ -cie and -lkAAn, which form proprietive nominals, e.g. xute ‘child’ : PROPR xute-cii ~ xute-l-keen ‘having children, with children’. These forms can modify both a nominal and a verbal headword. However, although rather close in function to the comitative case, they are not case forms in the strict sense, since they can function as independent nominal stems and take number and case suffixes, e.g. xeƞke ‘speech’ : PROPR xeƞke-cii ‘one having (the ability of) speech’ : PROPR-PL xeƞke-cii-l ‘those having (the ability of) speech’, gugda ‘height’ : PROPR gugda-cii ‘one having height, high’ : PROPR-ACC gugda-cii-wa.

The possessive and proprietive forms represent two sides of the possessive relationship: while the possessive form indicates who is the possessor, the proprietive forms indicate what is possessed. Importantly, although both categories are formally derivational, they are fully productive and therefore close to inflectional morphology. NUMERALS The numerals in Ewenki are nominal words which normally function as modifiers before a nominal headword, though in some constructions, and in independent use, the numeral may also act as the headword. The numerals follow a decimal system with roots inherited from Proto-Tungusic, and with separate items indicating the digits from 1 to 10, as well as the lower powers of 10. Khamnigan Ewenki also has specific numerals for the decades from 20 to 40. With few exceptions, the roots of the cardinal numerals serve as bases for ordinals, as well as for several other types of numeral derivatives. In independent use, as well as dialectally in adnominal position, the numerals can take case suffixes. Plural forms from numerals are mainly used in special functions (as discussed below). The nominal following a numeral other than ‘one’ takes, however, normally the plural form, e.g. ulukii ‘squirrel’ : tunƞa ulukii-l ‘five squirrels’ (five squirrel-PL). The independent forms of the cardinal numerals for the basic digits are: 1 umun ~ umuun : umn- (< *umun), 2 juur, 3 ilan, 4 digin, 5 tunƞa ~ Khamnigan Ewenki tunna, 6 ñuƞun ~ Khamnigan Ewenki nuƞun, 7 nadan, 8 japkun, 9 yegin, 10 jaan. With the exception of 2 juur and 5 tunƞa ~ tunna, all of these are nasal stems in /n. The final nasal can be absent before derivational elements, sometimes together with the preceding vowel. In some uses, and in specific derivational forms, the final nasal can also be replaced by r,

Siberian Ewenki 159

which, then, represents the plural marker, as is etymologically the case also with 2 juur = PL juu-r. The numeral 1 umun ~ umuun is dialectally used in the diminutive form umukeen ~ umuu-keen. For the decades, Ewenki uses a dialectally varying combination of lexical, synthetic and analytic means. In Khamnigan Ewenki, the lower decades are expressed by the independent lexemes 20 orin (both dialects), 30 gutin (Borzya) ~ gucin (Urulyunggui), 40 ducin (both dialects), borrowed from Mongolic, with orin and gutin representing an archaic layer and gucin and ducin a modern layer of borrowings. The subsequent decades are formed in Khamnigan Ewenki by the suffix *-ƞii > -n-nii, added to the stems of the basic digits: 50 tunƞan-nii, 60 nuƞun-nii, 70 nadan-nii, 80 japkun-nii, 90 yegin-nii. Regular Siberian Ewenki, however, uses the analytic forms 20 juur+jaa-r, 30 ilan+jaa-r, 40 digin+jaa-r, 50 tunƞa+jaa-r, 60 ñuƞun+jaa-r, 70 nadan+jaa-r, 80 japkun+jaa-r, 90 yegin+jaa-r, in which the plural form jaa-r functions as the nominal headword for the basic digits. There is also dialectal variation in the formation of the intermediate numerals. The most widely spread pattern is based on the simple additive principle, in which the numeral for the decade is followed by the numerals for the digit, e.g. 11 jaan umun, 12 jaan juur, 21 juur+jaa-r umun, etc. Alternatively, for the numerals of the second decade (“teens”), the pattern 11 jaan-duuk umun, 12 jaan-duuk juur, etc., with ABL jaan-duuk ‘from ten’, is used. In a few dialects of the southern group a third pattern, involving the petrified suffix -jeleke, from INSTR -ji + xeleke ‘remainder, extra’, is attested, e.g. 11 umun-jeleke, 12 juur-jeleke, etc. For the powers of ten, Ewenki uses the native Tungusic word 100 ñamaa : ñamaaji ~ ñamaajii, Khamnigan Ewenki namaaji, as well as the borrowed items 1000 miƞan ~ miƞgan (from Mongolic) ~ 1000 tisica (from Russian) ~ tixinca (from Russian via Yakut), depending on the dialect. In some dialects in close contact with Mongolic, as in Khamnigan Ewenki, the Mongolic item 10 000 tumen is also known. The ordinal numerals from 3 to 9 are formed by the suffix -ii, or also -ii-pti, added to the numeral root without the final nasal and the preceding vowel: il-ii(-pti) ‘third’, dig-ii(pti) ‘fourth’, tunƞi-ii(-pti) ‘fifth’, ñuƞ-ii(-pti) ‘sixth’, nad-ii(-pti) ‘seventh’, japk-ii(-pti) ‘eighth’, yeg-ii(-pti) ‘ninth’. Exceptionally, however, jaa-gii(-pti) ‘tenth’ is marked by the synchronically irregular suffix form -g.ii(-pti). In complex ordinals, only the last member of the sequence takes the ordinal marker, e.g. tunƞa+jaar dig-ii(-pti) ‘fifty-fourth’. For the first two digits, Khamnigan Ewenki has umu-tkii ‘first’ and juu-kii ‘second’, while in regular Siberian Ewenki a variety of suppletive stems are used: ñoo-guu ~ ñoo-wuu ‘first’ < ‘foremost’ (from ñoo- ‘to go ahead, to be in advance’), julee-guu ‘first’ < ‘foremost’ (from julee ‘front’), nono-pti ‘first’ < ‘earlier’ (from nonon ‘beginning’), elekees.i-pti ‘first’ (from elekees ‘at the beginning’), pierway ‘first’ (from Russian), gie(-pti) ‘second’ < ‘other’. In the intermediate numerals starting with the second decade, the forms umukee-gii and juu-gii are used, e.g. jaan umukee-gii ‘eleventh’, jaan juu-gii ‘twelfth’, juur+jaar juu-gii ‘twenty second’. In actual use, the ordinals are normally followed by the third person plural possessive suffix, e.g. PX3PL il-ii-tin ‘the third (of them)’. Khamnigan Ewenki has the option of expressing ordinals analytically with the help of the Mongolian loanword dugaar ‘number’, e.g. umun dugaar ‘first’, etc. A special category of numerals is formed by the collectives, which incorporate a suffixal element indicating the class of the nominal. These elements, which have analogies in other languages of the Far East, function as formatives for synthetic counters and include:

160 Juha Janhunen

• -ii ~ -nii for humans, e.g. juur-ii ‘two together’, ila-nii ‘three together’, digi-nii ‘four together’, tunƞa-nii ‘five together’; • -n.nu ~ -n.du (< *-n-su) ~ -r-du (with the plural marker *-r-) for households, e.g. ila-nnu ~ ila-ndu ~ ila-r-du ‘three households’, but exceptionally -pu in juur-pu ‘two households’; • -musA ~ muxA for places and directions, e.g. ilan-musa ‘three places/directions’, digin-muse ‘four places/directions’; • -llA ~ -ldA (< * -l-sA) for days, e.g. juu-lle ~ juu-lde ‘two days’, ila-lla ~ ila-lda ‘three days’; • -gdA : -rAgdA (< *-r-A.gdA, apparently also containing the plural marker -r-) for individualized entities, including humans and animals, e.g. juu-regde ‘two entities’, ila-ragda ‘three entities’, tunƞa-gda ~ tunƞa-ragda ‘five entities’, ñuƞu-regde ‘six entities’. The same function is expressed by the suffix -ƞnA, e.g. ila-ƞna ‘three entities’, digi-ƞne ‘four entities’ tunƞa-ƞna ‘five entities’, but exceptionally -rA in ñuƞu-re ‘six entities’. The elements -gdA and -ƞnA are synchronically distinct, but they may be diachronically connected, which would explain their functional similarity and paradigmatic interaction. Other derivatives based on numerals include multiplicatives in -AA ~ rAA (< *-r-AA, with the plural marker *-r-), e.g. umn-ee ‘once’, juur-ee ‘twice’, ila-raa ‘three times’, distributives in -tAl, e.g. umu-tel ‘one at a time’, juu-tel ‘by the twos’, ila-tal ‘by the threes’, as well as, at least in Khamnigan Ewenki, approximatives in -kin, e.g. juu-kin ‘about two’, ila-kin ‘about three’, tunƞa-kin ‘about five’. Structured entities composed of a given number of parts (“tuples”) are formed by -mAAn, which can also be attached to the plural form in -r of the numerals, e.g. umun-meen ‘single’, ilan-maan ~ ila-r-maan ‘triple’, ñuƞun-meen ‘six-fold’. Proprietive forms in -cii from the numerals are used to indicate age, e.g. PROPR juu-cii ‘two years old’, PROPR ila-cii ‘three years old’, PROPR jaa-cii ‘ten years old’. Several numeral forms can take the diminutive suffix -kAAn, which can convey a limitative notion, e.g. COLL-DIM juur-ii-keen ‘only two of them’. A few functions can also be expressed by regular morphological forms of the numeral paradigm. The plural forms of the basic numerals are dialectally used as multiplicatives, e.g. PL tunƞa-l ‘five times’, PL jaa-r ‘ten times’, or also as collectives, e.g. ila-r ‘the three of them’. The instrumental form of ordinal numerals, combined with the reflexive suffixes SG -wii : PL -wA-r ~ -wA-l expresses the time of completion of an action, e.g. ORD-INSTR-RX-PL tunƞ-ii-t-pA-l ‘on the fifth (day)’. PRONOUNS The basic pronominal roots in Siberian Ewenki may conveniently be divided into the conventional categories of personal, demonstrative, and interrogative pronouns, as well as a reflexive pronoun. All of these are more or less regular developments of the corresponding Proto-Tungusic elements. The basic personal pronouns comprise only the items for the first and second persons singular and plural. The pronominal roots form a set in which the initial consonants b : s indicate the person and the etymologically palatal vowels i : u the number. The roots are lexically monosyllables with a short vowel (CV), but in absolute usage the vowel is lengthened (CVV). The oblique stems end in a “pronominal n” (CVn-), to which the case endings are added. Also, the nasal conditions the change of b to m in the first person

Siberian Ewenki 161

items. The case paradigm follows that of regular nominals with the additional feature that the function of the object form can be expressed either by the partitive in -e or, alternatively, by the complex suffix -e-we, containing also the accusative ending (Table 7.6). The possessive forms in -ƞii, dialectally -nii, function in adnominal position as “genitives”, but formally they are possessive pronouns which can also be inflected in number and case. Prototypically, the possessor is expressed by the possessive suffixes, which, like the predicative personal endings are ultimately, and in some cases still transparently, connected with the personal pronouns. TABLE 7.6 SIBERIAN EWENKI PERSONAL PRONOUNS   SG

1

PL

1

2 2

EXCL

NOM

OBL

PART

PART-ACC

POSS

bii

min-

min-e

min-e-we

min-ƞii

sii

sin-

sin-e

sin-e-we

sin-ƞii

buu

mun-

mun-e

mun-e-we

mun-ƞii

suu

sun-

sun-e

sun-e-we

sun-ƞii

The system is complicated by the fact that the first person plural pronoun buu : munis used in an exclusive function only (‘we without you’). The corresponding inclusive function (‘I/we with you’) is expressed by the item EXCL mit ~ mut : (< miti, as in the Borzya dialect of Khamnigan Ewenki, dialectally also bit) : OBL mit.i- ~ mut.i- : ACC mit-pe ~ mut-pe : POSS mit-ƞii ~ mut-ƞii. This can be combined with the numeral for 2, yielding mit+juur ‘we two’, which may be seen as an incipient dual first person pronoun. The function of the third person pronoun is filled by the root nuƞan, Khamnigan Ewenki nugan, which is morphologically a regular nominal, though synchronically it only has the pronominal function and normally refers to humans only (for non-humans, including animals, the demonstrative pronouns in combination with a suitable noun are used). In the singular form nuƞan ‘he, she’ the final nasal has a dual status, in that it is morphologically a part of the stem, but it also conveys the notion of the third person singular possessive suffix -n, which appears in the oblique forms after the case markers, e.g. ACC-PX3SG nuƞan-ma-n : DAT-PX3SG nuƞan-duu-n. In the corresponding plural forms the final nasal is replaced by the plural suffix -r, followed by the third person plural possessive suffix -tin, which in the oblique forms follows the case markers, e.g. PL-PX3PL nuƞa-r-tin ‘they’ : PL-ACC-PX3PL nuƞa-r-wa-tin : PL-DAT-PX3PL nuga-r-duu-tin. The personal pronouns can take a limited number of derivational suffixes and postclitics, such as the diminutive suffix in -kAAn and the augmentative suffix in -kuun ~ -kAA-kuun, e.g. 1SG-DIM bi-keen ‘I, the little one’ : 1SG-AUGM bi-kuun ~ bi-kee-kun ‘I, the big one’. Similative and equative functions are expressed by the elements -gAcin ~ -kAcin ~ ‌‑ƞAcin ‘similar to’ and -diin ~ -din ‘of the same size as’, e.g. 1SG-SIM min-ƞecin ‘like me’, 1PL.INCL-SIM mit-kecin ‘like you and me/us’, 3P-PL-SIM-PX3PL nuƞa-rgacin-tin ‘like they’, 1SG-EQU min-diin ‘of the same size as me ~ equally tall as I’. The reflexive pronoun has the root SG meen : PL mee-r, which in its reflexive function is always combined with the reflexive suffixes SG -wii ~ ‑wi : PL -wA-r, dialectally -wAl, with the suffix-initial alternation w (b) ~ p ~ m. The basic form, with no case suffix, is used in the function of a direct object, with different forms referring to a singular and a plural subject: RX SG meen-mii ‘(one)self’ : PL-RX-PL mee-r-we-r ‘(them)selves’. In the oblique forms, the case markers, including the possessive marker, are placed before the

162 Juha Janhunen

reflexive suffixes, e.g. DAT-RX.SG meen-duu-wii : PL-ABL-RX-PL mee-r-duuk-pe-r : POSS-RX.SG meen-ƞii-wii : POSS-PL-RX.SG meen-ƞii-l-wii : PL-POSS-RX-PL meer-ƞii-we-r, etc. The root SG meen : PL mee-r can also be used in an intensive (emphatic) function, in which case it takes the possessive suffixes, e.g. PX1SG meen-mi ‘(I) myself’, similarly: PX2SG meen-ni, PX3SG meen.i-n, PL-PX1PL.EXCL mee-r-wun, PL-PX1PL. INCL mee-r-ti, PL-PX2PL mee-r-sun ~ mee-r-run, PL-PX3PL mee-r-tin, also in oblique forms, e.g. DAT-PX1SG meen-duu-w ‘to myself’ : PL-DAT-PX3PL mee-r-duu-tin ‘to themselves’ : POSS-PX-1SG meen-ƞii-w ‘my own’ : PL-POSS-PX3PL mee-r-ƞii-tin ‘their own’, etc. Alternatively, the derivationally expanded forms DER SG meen-ee-keen ‘self’ : PL meen-ee-kee-r ‘selves’ can be used indifferently in all persons. For the reciprocal function, the elements mee- : meen : PL mee-r are reduplicated in a variety of combinations, including mee&mee-, meen&mee-r-, mee-r&mee-r-, and mee&mee-gii-, e.g. DX-PL-RX-PL mee-mee-gii-l-we-r ‘(they) with regard to each other’. The demonstrative roots are e- for proximal and ta- for distal reference. In absolute use, the derived forms er ~ eri ‘this’ : PL er.i-l ‘these’ and tar ~ tari ‘that’ : PL tar.i-l ‘those’ are used. These can be used both as independent headwords and as modifiers to other nominals. In the singular case forms, many of which function as lexicalized adverbs, both the simple roots e- vs. ta- and the derived stems er- vs. tar- occur: • •

The simple roots e- vs. ta- are attested in DAT e-duu vs. ta-duu : ABL e-duuk vs. ta-duuk : LOC e-lee vs. ta-laa : PROL e-lii vs. ta-lii, all of which have a general spatial and/or temporal reference; The derived stems er- vs. tar- are attested in PART er-e vs. tar-a : PART-ACC er-e-we vs. tar-a-wa (used as the default object form) : DIR-LOC er.i-k-lee vs. tar.ik-laa : DIR-PROL er.i-k-lii vs. tar.i-k-lii : DIR er-tikii vs. tar-tikii : ELAT er-giit vs. tar-giit : INSTR er.i-t vs. tar.i-t (mainly with a causal function: ‘for this/that reason’).

The demonstratives also yield a number of correlative derivatives, including the diminutives er-keen ‘this one’ vs. tar-kaan ‘that one’ (with an affectionate connotation), the similatives er-gecin ‘like this’ vs. tar-gacin ‘like that’, the equatives er-diin ‘as big as this’ vs. tar-diin ‘as big as that’, the limitatives er.i-gde ‘only this’ vs. tar.i-gda ‘only that’, the quantitatives er-be (er-we) ‘this much’ vs. tar-ba (tar-wa) ‘that much’, and the emphatic forms er=mel ‘this very’ vs. tar=mal ‘that very’. Specific spatial stems are formed by the elements -w- : -wA- : -wAA- (< *-bAA-), e.g. e-w-gii ‘this place’ : INSTR e-w-gii-t = ELAT e-w-giit ‘from here’, DIR e-we-skii ~ e-w-gii-skii ‘in this direction’, e-w-gii-dee ‘this side’ : DAT e-w-gii-dee-duu ‘here’, e-w-gii-geen ‘inhabitant of this region’, etc. vs. ta-w-gii- ~ ta-waa-gii- ‘that place’ : LOC ta-waa-laa ~ ta-waa-gii-laa ‘over there’, ta-waa-gii-daa ‘that side’ : LOC ta-waa-gii-daa-laa ‘there, on that side’, etc. An idiosyncratic derivative with an etymological connection with the root e- is present in e-sii ~ e-xii ‘now’. The interrogative pronouns comprise ƞii ‘who?’ (with reference to persons only) : PL ƞii-l, as well as the roots ie-, ii-, oo- and a- for a variety of other question words. •

The root ie- is a pro-nomen-verbum. As a verb it means ‘to do what?’, e.g. PROGRAOR-3SG ie-ja-ra-n ‘what is s/he doing?’ : PTCP.PRF-PX2SG ie-caa-s ‘what did you do?’ : CV.PURP ie-daa ~ -PX3SG ie-daa-n ‘why?’. As a nominal it functions as a general question word in the meaning ‘what?’. Derivatives include (denominal:) ie-ma, ie-gda ‘what kind of?’, (possibly deverbal:) ie-wuu ~ ie-wuu-na ‘why?’.

Siberian Ewenki 163

•

•

•

More commonly, the expanded stem ie-kuun ~ ie-kun ‘what?’ : PL ie-kuu-r ~ ie-ku-r, with a full nominal paradigm and with derivatives like ie-kuu-di ‘what kind of?’ : ie-kuu-ma ‘made of what?’, is used. The root ii- is attested in the local case forms DAT ii-duu ‘where?’ : ABL ii-duuk ‘from where?’ : LOC ii-lee ‘where?’ : PROL ii-lii ‘which way?’, as well as in the derivatives ii-r ‘which one?’ : ii-ree ‘where?’, ii-di ~ ii-dik ~ ii-diguu ‘which one?’. It may be noted that the root ii- is phonetically close to ie-, and the two can dialectally merge due to the development *ie > ii, as in Manchurian Ewenki. Even so, the two roots are etymologically distinct, and both are present in the language as separate items, as can also be seen from the fact that they take different vowels in suffixal forms (aa vs. ee), in that ie- is harmonically velar, while ii- is palatal. The root oo- is attested in the forms oo-n ‘how?’, oo-kin ‘when?’ : ACC oo-kin-ma ‘how long?’ : DAT oo-kin-duu ‘in the course of how long a time?’, DIR-LOC oo-kindu-laa ~ DIR PROL oo-kin-du-lii ‘until/by what time?’, oo-kii ‘how many?’ : DAT oo-kii-duu ~ PROL oo-kii-lii ‘for how much?’ : INSTR oo-kii-t ‘how?’, oo-kii-raa ‘how many times?’, oo-kii-lla ~ oo-kii-lda ‘how many days?’. Related items are also oo-kii-r ‘a long time ago’ : oo-kii-pti ‘ancient’. The root a- (< Proto-Tungusic *xa-) is attested in the forms a-sun ~ a-xun ‘how much?’ (of uncountables), a-dii ‘how many?’ (of countables) : INSTR a-dii-t ‘by how much?’, PROPR a-dii-cii ‘how many years old?’, a-dii-raa ‘how many times?’, adii-lla ~ adii-lda ‘how many days?’, a-w- ‘what place?’ : INSTR a-w-giit ‘from where?’ : DIR a-wa-skii ‘whither?’, a-w-guu ~ a-wa-guu ‘which?’, a-wa-dii ‘what kind of?’.

In combination with the clitic =wAl ~ =mAl the interrogatives yield a set of indefinite pronouns, e.g. ngii=wel ‘somebody’, iekuun=mal : PL iekuu-r=wal ‘something’, ie-kuu-ma=wal ‘some kind of’, ii-duu-wel ‘somewhere’, oo-kii=wal ‘some number of’, oo-kii-duu=wal ‘at some time’, a-sun=mal ‘some amount of’, a-dii=wal ‘some number of’, a-wa-dii=wal ‘some kind of’, etc. The same function can be expressed by the clitic =dAA, which, however, is more often used to mark connegation (combination with negation), e.g. ngii=dee ‘somebody/(not) anybody’. Another connegative clitic, used in the eastern dialects, is =kAAt, e.g. iekun=kaat ‘(not) anything’, ii-duu=keet ‘(not) anywhere’. PERSON MARKING Siberian Ewenki preserves the two separate sets of suffixal person markers inherited from Proto-Tungusic (and lost only in Jurchenic). The two sets are conventionally known as the predicative personal endings and the possessive suffixes, respectively (Table 7.7). The personal endings (VX) express the actor (subject) of a verbal predicate and are based on the basic (nominative) forms of the personal pronouns. The possessive suffixes (PX) express prototypically the owner (possessor) of a nominal headword and are based on the oblique stems of the personal pronouns. Importantly, there are markers also for the third person, both singular and plural, although no original third person pronouns are preserved in the language. The system is complicated by the circumstance that the possessive suffixes can also be attached to a verbal predicate if the latter involves the finite use of a nominalized verb, normally a participial form in the context of the phenomenon known as “verbalization”, though better termed “finitization” (or “re-verbalization”). Unlike some other Siberian languages, however, Siberian Ewenki

164 Juha Janhunen

does not attach personal endings to actual nominals without a copular verb. Apart from the actual possessive (“personal-possessive”) suffixes, there is also a set of reflexive (“reflexive-possessive”) suffixes (RX). TABLE 7.7 SIBERIAN EWENKI PERSON MARKERS  

VX V

SG

*V

*C

/n

1

-m

-w

-Bi

-mi

2

-n-Si

-s

-Si

-ni

-Bi/i

mi/i

3

-n

RX PL

PX /n

1

-n -wi/i

EXCL

-w

-Bun

INCL

-p

-t

2

-s

-Sun

3

-Ø

-tin ~ -ti-r

RX

-wAr

-mun -ti -nun -BAr

-mAr

Stem types: (*)V = (original) vowel stems (including secondary consonant stems), *C original consonant stems, /n = primary nasal stems. Segmental alternations: B = w b p, S = d l r.

The interaction of the person markers with the preceding segments involves a number of relatively simple morphophonological alternations. For instance, the w (b) of the first person markers is nasalized under the impact of a preceding *n, e.g. oron ‘reindeer’ : PX1PL.EXCL oron-mun (< *oron-bun). Also, the initial *s of the second person markers is in postconsonantal position represented by *d, which can dialectally be further assimilated to n l r, depending on the preceding consonant, e.g. PX2SG oron-ni (< *oron-dï < *oron-sï) : PL-PX2PL oro-r-run (< *oro-r-dun < *oro-r-sun), adil ‘net’ : PL-PX2PL adil.i-l-lun (< *adïlï-l-dun < *adïlï-l-sun). The final *i of the singular first and second person markers and the plural first person inclusive marker is in regular Siberian Ewenki preserved only after stems ending in an original final consonant, but lost after vowel stems, e.g. juu ‘dwelling’ : PX1SG juu-w (< *juu-bï) : 2SG juu-s (< *juu-sï) : 1PL.INCL juu-t (< *juu-tï). Markers consisting of a single consonant are linked to a preceding consonant by the connective vowel i, e.g. xalgan ‘foot’ : PX3SG xalgan.i-n, unta ‘shoe’ : PL unta-l : PL-PX3SG unta-l.i-n. The same types of allomorphic variation are observed when the possessive suffixes are added to marked case forms after the case suffixes (N-NX-CX-PX), e.g. nekun ‘younger sibling’ : COMPX1SG nekun-nun-mi, kalan ‘kettle’ : PL-ACC-PX2SG kala-r-wa-s, ƞaala ‘hand’ : ABL-PX3SG ƞaala-duuk.i-n. The predicative personal endings have in practice limited use, in that they almost solely occur in the so-called “aorist” (and related) forms of the verb, which, however, are very frequent. The aorist paradigm is exceptional in that the person markers are normally attached to the stem with the help of an aorist marker, which is *-n- for the first and second persons singular and *-rA- for the rest of the paradigm. For this reason, the singular first and second person endings always occur in the nasalized forms VX1SG -m

Siberian Ewenki 165

(< *-n-bI) : 2SG -nni ~ -ndi (< *-n-sI). Another peculiarity of the “aorist” paradigm is that the third person plural has a zero marker, meaning that it is the least marked member of the paradigm, while the third person singular has the ending -n, identical with the corresponding possessive suffix. Not surprisingly, the two sets of person markers can be dialectally confused. One source of confusion lies in the fact that the plural personal endings 1PL.EXCL -w : 2PL -s are homonymous with the singular possessive suffixes 1SG -w : 2SG -s. To avoid this homonymy, the plural personal endings have in some dialects, as in Khamnigan Ewenki, been replaced by the corresponding possessive suffixes, a feature which can dialectally be extended to the first person inclusive forms. e.g. baka- ‘to find’ : AOR-1PL.INCL baka-ra-p ~ (Urulyunggui) baka-ra-t. The opposite generalization, involving the use of the plural first person inclusive personal ending in the function of the corresponding possessive suffix, is also attested in some of the southern dialects of Siberian Ewenki, e.g. inmek ‘pack bag’ : PX1PL.INCL inmek.i-t ~ (Stony Tunguska) inmek.i-p. In the same dialects the first person singular possessive suffix has the form -m for all stem types, e.g. oron ‘reindeer’ : PX1SG oro-m, purta ‘knife’ : PX1SG purta-m vs. “standard” oron-mi, purta-w. It may be added that Khamnigan Ewenki has generalized the sequence n.i-n from the third person singular form of nasal stems also to other consonant stems, e.g. xukur ‘bovine’ : PX3SG xukur-nin. Khamnigan Ewenki (Urulyunggui) also occasionally uses the innovative “plural” form -ti-r of the plural third person possessive suffix, generally represented as -tin (= ti-n) in most other Ewenki dialects. For reflexive possession, the reflexive suffixes SG -wii ~ -wi : PL ‑wA-r, dialectally -wA-l, with the suffix-initial alternation w (b) ~ p ~ m, are used, e.g. girkii ‘friend’ : (singular subject and singular reference) RX girkii-wii ‘one’s own friend’ : (singular subject and plural reference) PL-RX girkii-l-wii ‘one’s own friends’ : (plural subject and singular reference) RX-PL girkii-we-r ‘people’s own friend’ : (plural subject and plural reference) PL-RX-PL girkii-l-we-r ‘people’s own friends’. The basic reflexive forms, with no case marking, imply object position, while other case forms are marked by the corresponding case suffixes (N-NX-CX-RX), e.g. juu ‘dwelling, home’ : LOC-RX juu-laa-wii ‘to one’s own home’, guule ‘house’ : PL-ABL-RX-PL guule-l-duuk-pe-r ‘from people’s own houses’. Possessive person marking is also connected with the phenomenon of alienability. In Siberian Ewenki, alienable (indirect) possession is marked by the suffix -ƞ- ~ -ƞ.i- (< *-ƞI- < *-ƞU-), always followed by a possessive or a reflexive suffix, e.g. dil ‘head’ : PX1SG dil.i-w ‘my head’ : AL-PX1SG dil.i-ƞ.i-w ‘the head of an animal belonging to me’, ulle ‘meat’ : AL-RX-PL ulle-ƞ-me-r ‘the meat of an animal belonging to them(selves)’. In practice, inalienable (direct) possession is possible with items denoting relatives, friends, domestic animals, body parts, native places, personal effects, or other types of inherent relationship. When no such relationship is involved, the possession is considered alienable and requires the use of the alienability marker, cf. e.g. oron ‘reindeer’ : PX1SG oron-mi ‘my (domestic) reindeer (which I own)’ vs. ulukii ‘squirrel’ : AL-PX1SG ulukii-ƞ.i-w ‘my (wild) squirrel (which I have caught)’. In the dialects there are, however, deviations from this simple basic pattern. Depending on the type of possession, the alienability marker can also be attached to pronominal stems, e.g. er ‘this’ : AL-PX1SG er.i-ƞ.i-w ‘this my possession’.

166 Juha Janhunen

VERBAL FORMS In accordance with the general Tungusic pattern, Siberian Ewenki exhibits four categories of verbal forms: (i) imperatives, (ii) indicatives, (iii) converbs, and (iv) participles. Each of these categories is characterized by a unique set of formal and functional properties. (i) Imperatives are inherently monofunctional finite forms with a deontic modal content. In Siberian Ewenki they form a personal paradigm which, however, is only partially based on the use of the regular finite person markers (VX). The modal content is conveyed by an imperative marker, which is placed between the verbal stem and the person marker. The imperative marker has two apparently unrelated shapes: •

•

The basic second person direct imperative is marked by the suffix ‑kAl for the singular and -kAl-lu ~ -kAl-du for the plural, e.g. baka- ‘to find’ : IMP.2SG baka-kal ‘(thou) find!’ : 2PL baka-kal-lu ‘(you) find!’, guun- ~ guu- ‘to say’ : IMP.SG2 guu-kel ‘(thou) say!’ : 2PL guu-kel-lu ‘(you) say!’. As can be seen, the singular form has no actual person marker (-Ø), while the plural form contains the suffixalized second person plural pronoun *-su (> -du > ‌‑lu), which is also the source of the personal ending VX2PL -s, as used in the aorist paradigm. In Khamnigan Ewenki, the plural form has the ending -du.i, containing the dialectally restricted innovative vowel sequence ui, e.g. ga- ‘to take’ : IMP 2SG ga-kal : 2PL ga-kal-dui. In the other persons the imperative marker is -g- ~ -k- (< *-gI-) ~ after nasal stems ‑ƞ-, to which additional elements are added to indicate the person. The third person forms take simply the corresponding possessive suffix, yielding -g.i-n ~ -ƞ.i-n for the singular and -k-tin ~ -ƞi-tin for the plural, e.g. baka- ‘to find’ : IMP-3SG baka-g.i-n ‘may s/he find!’ : 3PL baka-k-tin ‘may they find!’, aasin- ‘to sleep’ : IMP-3SG aasin-ƞ.i-n ‘may s/he fall asleep’ : 2PL aasin-ƞi-tin ‘may they fall asleep’. The first person singular takes the idiosyncratic element -tA, yielding -k-tA (< *-gI-tA), e.g. baka-ktA ‘let me find!’, which in combination with the possessive suffix PX1PL.EXCL -wun can also be used as the base for the plural first person exclusive form, e.g. IMP-1PL.EXCL baka-ktA-wun ‘let us (without you) find!’. Dialectally, the plural form can omit the element -tA-, e.g. baka-k-pun ~ baka-w-wun. The plural first person inclusive form is marked variously by the complex endings -g-AA-t ~ -g-AA-r, containing either the corresponding possessive suffix PX1PL.INCL -t or the plural marker -r, e.g. IMP-1PL.INCL baka-g-aa-t ~ baka-g-aa-r ‘let us (with you) find!’.

(ii) Indicatives are likewise inherently monofunctional finite forms, which can also convey temporal, aspectual and/or modal connotations. They are, however, not used in the function of imperatives (proper). All indicatives in Siberian Ewenki follow the so-called aorist pattern of conjugation, which typically involves the use of the aorist markers *-n- for the first and second persons singular and *-rA- for the rest of the paradigm, added either directly to the plain verbal root or to a derivationally expanded secondary stem before the predicative personal endings (VX). There are also a few exceptional verbs which take the stem extensions -dA- or -si- in the aorist paradigm. Altogether, the aorist stem formation is the single most complicated detail of Ewenki morphology and morphophonology. •

The most simple picture is presented by stems ending synchronically in a vowel, in which case there are no complications at the juncture of the stem and

Siberian Ewenki 167

•

•

•

•

the suffixal complex, except that the aorist marker *-n- merges with the singular first person ending *-bI, e.g. waa- ‘to kill’ : AOR 1SG waa-m : 2SG waa-n-di ~ waa-n-ni : 3SG waa-ra-n : 1PL.EXCL waa-ra-w : 1PL.INCL waa-ra-p : 2PL waa-ra-s : 3PL waa-ra. Secondary consonant stems, which in open syllables have lost a final *i, follow a basically similar pattern, in that they retain the vowel stem before the aorist marker *-n-, e.g. luk- : luk.i- ‘to untie’ : AOR 1SG luk.i-m : 2SG luk.i-n-di ~ luk.i-n-ni. In the rest of the paradigm, the aorist marker *-rA- is attached to the consonant stem and is represented by the variants ‑dA- (after sonorants) ~ -lA (normatively after l) ~ rA (normatively after r) ~ -nA- (normatively after nasals) ~ ‑tA- (after obstruents), e.g. (x)orol- : (x)orol.i- ‘to turn around’ ITR : AOR3SG (x)orol-do-n ~ (x)oro-l-lo-n, sokor- : sokor.i- ‘to lose’ : AOR-3SG sokordo-n ~ sokor-ro-n, um- : um.i- ‘to drink’ : AOR 3SG um-da-n ~ um-na-n, is- : is.i- ‘to reach’ : AOR-3SG is-ta-n. Dialectally, however, the basic vowel stem, without the aorist marker, is also attested in the plural first person inclusive and second person forms, e.g. il- : il.i- ‘to stop’ : AOR-1PL.INCL il-la-p ~ il.i-p, 2PL il-la-s ~ il.i-s. The only systematically surviving type of primary consonant stems is formed by nasal stems, ending in an original *n. In the aorist paradigm of these stems the marker *-rA- is represented as -A-, a feature inherited from Pre-Proto-Ewenic, while in the singular first and second persons the connective vowel i is used, a situation synchronically attested in, at least, the Urulyunggui dialect of Khamnigan Ewenki, e.g. guun- ~ gun- ‘to say’ : AOR 1SG gun.i-m : 2SG gun.i-ndi : 3SG gun-e-n : 1PL.EXCL gun-e-bun : 1PL.INCL gun-e-t : 2PL gun-e-sun : 3PL gun-e. The prevailing pattern in the Siberian dialects is, however, that the vowel -A- has been generalized also to the singular first and second person forms, e.g. AOR-1SG gun-e-m, similarly aasin- ‘to fall asleep’ : AOR-1SG aasin-a-m. Original obstruent stems are synchronically represented by the item jep- ‘to eat’, which in the singular first and second person forms has variously the stem jew.u-, as in most Siberian dialects, or jep.i-, as in the Urulyunggui dialect of Khamnigan Ewenki, e.g. AOR.1SG jew.u-m ~ jep.i-m. In the rest of the paradigm, the aorist marker -rA- is represented by the regular post-obstruent variant -tA-, e.g. AOR-3SG jep-te-n. Dialectally, however, the marker -tA- has been generalized also to the singular first and second person, e.g. AOR-1SG jepte-m. There are also a few verbal stems ending in g, notably deg- ‘to fly’ and xig- ‘to skin’, for which the aorist stems jeg-re- and xig-re-, with the unexpected cluster gr (instead of -k-t-), are attested in the Siberian dialects, cf. also the nominalization deg-ii ‘bird’ (< *‘flying’). As synchronic exceptions, Siberian Ewenki has traces of three verbs which, instead of the marker *-rA-, form the aorist paradigm with the element *-dA-, a feature inherited from Proto-Tungusic. The three verbs are bu- : bu-de- ‘to die’, ga- : ga-da- ‘to take’, and oo- : oo-da- ‘to become’. In a few varieties, the element -dA- is absent in the singular first and second person forms, but present in the rest of the paradigm, a situation attested in, at least, the Urulyunggui dialect of Khamnigan Ewenki, e.g. AOR 1SG ga-m : 2SG ga-ndi : 3SG ga-da-n : 1PL.EXCL ga-da-bun : 1PL.INCL ga-da-p : 2PL ga-da-sun : 3PL ga-da. In most varieties, however, the paradigm has been levelled either by generalizing the element -dA- to the singular first and second person forms, e.g. AOR 1SG

168 Juha Janhunen

•



ga-da-m : 2SG ga-da-nni, or by replacing *-dA- by *‑rA- in the rest of the paradigm, e.g. AOR- 3SG ga-ra-n. Finally, there are two verbs, both auxiliaries, which form the aorist paradigm with the element -si-. The two verbs are the copula-existential bi- : bi-si- and the negation verb e- : e-si-. The element -si- is present in all forms of the aorist paradigm, e.g. AOR 1SG e-si-m : 2SG e-si-nni : 3SG e-si-n : 1PL.EXCL e-si-w : 1PL.INCL e-si-p : 2PL e-si-s : 3PL e-si. The verb bi- can, however, dialectally also be inflected like a vowel stem: AOR 1SG bi-m : 2SG bi-nni : 3SG bi-re-n : 1PL.EXCL bi-re-w : 1PL.INCL bi-re-p ~ bi-p : 2PL bi-re-s ~ bi-s : 3PL bi-re.

Apart from the finite use of the personal forms of the aorist paradigm, the plain aorist stem is used in the negation construction as an invariant connegative form, combined with conjugated forms of the negation verb, e.g. e-si-m buu-re ‘I don’t give’ (NEGAOR-1SG give-CONNEG), similarly e-si-m gun-e ‘I don’t say’ (say-CONNEG), e-si-m ga-da ‘I don’t take’ (take-CONNEG), e-si-m bi-si ‘I am not’ (COP-CONNEG). Dialectally, at least in Khamnigan Ewenki, the conjugated negation verb can also be postposited and attached as a postclitic or suffix to the aorist form of the semantic main verb, e.g. e-si-m buu-re → buu-re+e-si-m > buu-r=esi-m > buu-re-si-m (give-CONNEG-AOR-1SG).

(iii) Participles (as understood here) are polyfunctional nominalizations of verbs, which can have both finite and non-finite uses. Thanks to their polyfunctionality they can appear variously in the roles of dependent adnominal modifiers, independent nominal headwords, or finite predicates. In the latter role they may or may not be accompanied by a copular verb. As finite predicates without a copula, participles interact with the aorist type of finite forms, with the only difference being that participles prototypically take the possessive suffixes as person markers (PX), while the aorist is combined with the predicative personal endings (VX). Like converbs, participles can convey temporal-aspectual and modal meanings. •

•

The basic temporally and aspectually unmarked participle, conventionally identified as the imperfective participle, is formed by the suffixes -rii (with variants), -ii, -dii, and ‑sii, based on the aorist marker, and sharing the latter’s morphological and morphophonological patterns, e.g. eme- ‘to come’ : PTCP.IMPRF eme-rii, jew- ~ jep- ‘to eat’ : PTCP.IMPRF jep-tii, guun- ~ gun- ‘to say’ : PTCP.IMPRF gun-ii, ga- ‘to take’ : PTCP.IMPRF (*)ga-dii ~ ga-rii, bi- ‘to be’ : PTCP.IMPRF bi-sii. This form occurs in all the various functions typical of Ewenki participles, e.g. saa- ‘to know’ : PTCP.IMPRF saa-rii (adnominally) ‘knowing, who knows’ ~ (as an independent actor noun) ‘one who knows’ ~ (as an action noun) ‘the fact of knowing’, also with person marking, e.g. PX1SG saa-rii-w (agentive use) ‘one whom I know/knew’ ~ (finite use) ‘I am (the) one who knows, I know/knew’. Combined with the dative marker -duu, this participle yields a quasiconverbal form of simultaneous action, e.g. girku- ‘to walk’ : PTCP.IMPRF-DAT-RX-PL girku-rii-duu-we-l ‘while walking’. Also, it is often combined with the derivational marker -jA- ~ -cA- for ongoing action (progressive aspect) in the complex form in -jA-rii, which functions as an imperfective or progressive participle, e.g. soƞo- ‘to cry’ : PROGR-PTCP.IMPRF soƞo-jo-rii ‘(one) who is crying’. Completed action is expressed by the perfective participle, marked by the suffix -cAA ~ -cA-, added directly to the verbal stem, e.g. dukuu- ‘to write’ : PTCP.

Siberian Ewenki 169

•

•

PRF dukuu-caa ‘(one) who has written’, guun- ~ gun- ‘to say’ : PTCP.PRF gun-cee ‘(one) who has said’, bu- ‘to die’ : PTCP.PRF bu-cee ‘(one) who has died, dead’, bi- ‘to be’ : PTCP.PRF bi-cee ‘(one) who has been’. This form has a functional range analogous to that of the imperfective participle, and in combination with case suffixes it yields quasiconverbal forms referring to completed or anterior action, e.g. eme- ‘to come’ : PTCP.PRF-LOC-PX3SG eme-ce-lee-n ‘after he had come’. Future action is expressed by the futuritive participle, marked by the suffix -ƞAA ~ -ƞA-, synchronically always added to the derivational marker -jA- ~ -cAfor ongoing action (progressive aspect), yielding the complex suffix -jA-ƞAA ~ -cA-ƞAA, e.g. waa- ‘to kill’ : PROGR-PTCP.FUT waa-ja-ƞaa ‘(one) who will kill’. In Khamnigan Ewenki, the suffix has further yielded the fully synthetic future marker ‑jAgAA(-) > -jAA(-), e.g. guun- ~ gun- ‘to say’ : PTCP.FUT gunjee ‘(one) who will say’ ~ ‘s/he will say’. Other nominalizations are formed by the suffixes -pkii ~ ‑wkii (habitive), e.g. xawal- ‘to work’ : PTCP.HAB xawal.i-pkii ‘(one) who usually works’, -mAcin (debitive), e.g. bele- ‘to help’ : PTCP.DEB bele-mecin ‘(one) who has to help’, -nA (resultative, often with a passive meaning, especially when formed from a transitive verb), e.g. dukuu- ‘to write’ : PTCP.RES dukuu-na ‘what has been written’, -ƞAAt : -ngaat.i- (optative), e.g. suruu- ‘to leave’ ITR : PTCP.OPT suruu-ƞeet ‘(one) who would like to leave’, ‌‌‑ƞkii (iterative), e.g. eme- ‘to come’ : PTCP.ITER eme-ƞkii ‘(one) who frequently comes/has come’, and ‑rkA(probabilitative), e.g. ga- ‘to take’ : PTCP.PROB ga-rka ‘(one) who has probably taken’. A somewhat exceptional form in this context is the one in ‑p.kAA ~ -w.kAA ~ -mu.kAA (necessitative, with a passive meaning and etymologically containing the reflexive-passive marker -p- ~ ‑w- ~ -mu-), which is normally used only as an independent head noun, e.g. ƞene- ‘to go’ : PTCP.NEC ƞenepkee ‘(the fact that it is) necessary to go’. A complete list of all nominalizations would be even longer. There is, however, considerable variation as to which forms are actively used in each given dialect, and in which functions.

(iv) Converbs are monofunctional non-finite forms, which serve as predicates of subordinated clauses, and which, at the same time, function as adverbial modifiers to a following verb in the context of a complex sentence. Converbs typically convey temporal-aspectual or modal meanings, by which they link the subordinated clause to the main clause and the verbal headword. Many converbal markers in Siberian Ewenki can take marking for both number (plural) and person (reflexive and possessive). The person markers reflect the referential status of the converb with regard to the subject of the main clause in accordance with the division into the conjunct (same-subject), disjunct (different-subject) and ambivalent types. •

The conjunct converbs may be regarded as the most basic type both functionally and formally. When referring to a plural subject (any person) they normally take plural marking. The three principal forms of this category may be termed the simultaneous, anterior and (general) connective converb. The simultaneous converb is marked by the suffix -nA and expresses simultaneous uncompleted action. It is often, though not necessarily, formed from a secondary stem containing the derivational marker -jA- ~ -cA- for ongoing action (progressive aspect), e.g. bi- ‘to be’ : CV.SIMULT bi-ne : PL bi-ne-l ‘while being’, ngene- ‘to go’ :

170 Juha Janhunen

•

•

PROGR-CV.SIMULT ƞene-je-ne : PL ƞene-je-ne-l ‘while going’. The anterior converb is marked by the suffix -ksA ~ -ksAA, dialectally also -xA ~ -xAA, and expresses anterior completed action, e.g. ice- ‘to see’ : CV.ANT ice-kse : PL icekse-l ‘after having seen’. The connective converb is marked by the suffix -mii and expresses actions that have a conditional (‘if’) and/or temporal (‘when’) relation to the main action, e.g. ga- ‘to take’ : CV.CONN ga-mii : PL ga-mii-l ‘if/ when (. . .) take’. Apart from its converbal function, this form has several types of additional uses making it similar to “infinitives” in European languages, for which reason it is conventionally used as the dictionary form for Ewenki verbs. The principal disjunct converb is marked by the suffix -k.i- added to the aorist stem, yielding the complexes -rA-k.i- (with variants), -A-k.i-, -dA-k.i-, and -si-k.i-, to which possessive suffixes are attached to indicate the actor person. Because of its connection with the aorist, this converb may be termed the “aorist converb”. Moreover, the element -k.i- derives from the primary ablative suffix *-kI, as still present in the regular ablative marker -du-k ~ duu-k(i), which means that the aorist stem is here used as a nominalization, and the entire construction is structurally quasiconverbal (case form of a nominalized verb). Functionally, the aorist converb corresponds closely to the conjunct connective converb, except that it always implies a different subject, e.g. nono- ‘to begin’ TR : CV-AOR-1SG nono-ro-k.i-w ‘when I began’, ice- ‘to see’ : SEM ice-sin‘to glance’ : CV.AOR-3SG ice-sin-e-k.i-n ‘when s/he glanced’, ga- ‘to take’ : CV.AOR-2SG ga-da-k.i-s ‘if/when you take’, bi- ‘to be’ : CV.AOR-3PL bi-sik.i-tin ~ bi-re-k.i-tin ‘if/when they are’. Many of the ambivalent converbs are of a transparent quasiconverbal origin, in that they contain a nominalizing element accompanied by a nominal case suffix. They are therefore better discussed in the context of participles. Some of them are, however, based on participles no longer in active use as nominalizations. Most importantly, the element -dAA- ~ -tAA- forms a set of purposive converbs in combination with the possessive suffixes, e.g. baka- ‘to find’ : CV-PURPPX1SG baka-daa-w ‘so that I would find’, PX3PL baka-daa-tin ‘so that they would find’. When combined with the reflexive suffixes, this form functions as a final converb or “supine”, e.g. CV.PURP-RX baka-daa-wii : PL baka-daawa-r ‘in order to find’. The “supine” is also used in a finite function to express the so-called “remote future imperative” of the second person. An apparently related form is the terminative converb in -dAlAA- = -dA-lAA- (< *-dAA-lAA, with CX LOC -lAA), e.g. CV.TERM-1SG baka-dalaa-w ‘until I find’.

In general, converbs in Siberian Ewenki are a diffuse and dialectally variable category whose limits, especially with regard to quasiconverbal forms, are not always easy to define. Apart from the major forms listed above, there are several minor forms which may also be classified as converbs. The suffixes of these forms are typically complex and contain a variety of both derivational and inflectional elements. For instance, the complexes -m-nA-k : PL ‑m-nA-k.i-l (possibly from *-mV- + CV.AOR *-rA-k.i-) and -m-ni.n ~ ‑mmi.n : PL -m-ni-r ~ -mmi-r express succession (‘immediately after’) in conjunct constructions, while the complex -ktA-wA- (< DX -ktA- + CX ACC -wA- + PX/RX) functions as a converb of anteriority (‘immediately before’) in both conjunct and disjunct constructions. A very special form is the so-called “converb of simultaneous (past) action” in -ƞA-sii-, combined with either the possessive or the reflexive suffixes, and normally used with a head clause containing reference to the past tense. This converb, here termed

Siberian Ewenki 171

“contemporal” is particularly often formed from the copula-existential bi- ‘to be’, e.g. CV.CTEMP-PX3SG bi-ƞesii-n ‘at the time when s/he was’ : RX bi-ƞesii-wii ‘while being (in the past)’. It is also attested in combination with the progressive marker -jA- ~ -cA-, e.g. ƞene- ‘to go’ : PROGR-CV.CTEMP-RX ƞene-je-ƞesii-wii ‘while going’. The special status of this form is contained in the fact that it is formally different both from the basic conjunct converbs, which do not take person marking, and from the disjunct converbs, including quasiconverbs, which do take person marking, but which also contain an element of nominalization in combination with case marking. Formally, the suffix -ƞAsii- = -ƞA-sii- is reminiscent of participle markers, especially of the aorist participle in -sii, as formed from the copula-existential bi-. However, the element -ƞAsii- is never attested in an actual participial function. VERBAL FUNCTIONS While the Ewenki verbal system can be relatively neatly described in terms of four formal categories, as discussed above, it is much more difficult to present a holistic picture of verbal functions, especially as far as non-imperative finite predicates are concerned. There is no simple answer to questions like how many moods, tenses or aspectual categories Ewenki has, for each of these categories can be manifested on three different levels: derivation, inflection, and syntax. This suggests that many of the forms and structures are recently grammaticalized and not yet integrated into a single morphologically coherent paradigm (Table 7.8). TABLE 7.8 SIBERIAN EWENKI FINITE FUNCTIONS  

 

DX

DX

INDEF

AOR

PROGR

PAST

-JA-JA-

INCH

-JA-

FUT

-JA-

PRF

MOD

-JA-

PX +

+

+

-RA+

+ +

-l.i-

+

-RA-ƞAA-

+

+

+

-cAA-

+

+

+

-cAA-

+

+

+

ITER

-ƞkii-

RES

-nA-

+

+

SUBJ

COP

+

-pkii-

PROGR

PROGR

VX

-Rii-

HAB FUT

NX

-RAIMPRF

PRS

PTCP

+

-cAA-

+

+

+

DEB

-mAci.n-

+

+

+

PROB

-rkA-

+

+

+

OPT

-ƞAAt.i-

NEC

-BkAA-

-m-

+ +

Selected finite forms, as listed according to their basic temporal reference or modal functions, including their combinations with number and person markers and copula. Segmental alternations: A = a o e, B = w p m, R = r d l n t or zero (Ø).

172 Juha Janhunen

The basic division lies between the forms using the aorist type of finite conjugation with predicative personal endings (VX) and those based on finitized participles normally taking the possessive suffixes (PX) for person marking. Some temporal, aspectual and/or modal functions are conveyed by derivational suffixes, combined with the aorist type of conjugation, while others are conveyed by participle markers, which require the possessive type of person marking. However, since participles are nominalized forms, a conflict arises between the personal paradigm and nominal number marking (NX). As a result, some finitized participles can in the third person forms take either the corresponding possessive suffixes or the nominal number markers (zero for the singular vs. -l/-r for the plural), e.g. baka- ‘to find’ : PTCP.PRF-PX3PL baka-caa-tin ~ PL baka-caa-l. Finally, some finitized participles can also be used like regular nouns in analytic constructions with a copula (COP), normally bi- ‘to be’, in which case the participle can take optional number marking, while person is marked on the copula, e.g. PTCP.PRF (PL) baka-caa(-l) + COP-AOR-2PL bi-si-sun. As far as tense is concerned, the aorist paradigm is essentially indefinite, or “universal”, and conveys only a generic reference to a “non-future” time range, e.g. saa- ‘to know’ : AOR 1SG saa-m ‘I know’, bi- ‘to be’ : AOR 1SG bi-si-m ‘I am’, though, when formed from verbs with an inherent perfective aspectual content, it normally denotes recent past, e.g. baka- ‘to find’ : AOR-3SG baka-ra-n ‘s/he has just found’. When unambiguous reference to the actual present tense range has to be indicated, the derivational suffix -jA- ~ -cA- for ongoing action (progressive aspect) is added, e.g. PROGR-AOR3SG baka-ja-ra-n ‘s/he finds, s/he is finding’, also in the copula-existential, e.g. PROGRAOR-1SG bi-je-m ‘I am being’. The forms with the suffix -jA- ~ -cA- can, however, also refer to the future (the so-called “first future tense”), in which case the singular third person, the plural first person inclusive, and the plural second person forms exceptionally lack the aorist marker before the personal endings, e.g. FUT 3SG baka-ja-n : 1PL.INCL baka-ja-p : 2PL baka-ja-s. Reference to an immediate future (“the second future tense”) can also be expressed by linking the progressive suffix -jA- with the inchoative suffix -l- : -l.i-, e.g. buru- ‘to fall’ : PROGR-INCH-AOR.1SG buru-je-l.i-m ‘I am just about to fall’. The most widespread way of future marking (“the third future tense”) is, however, the use of the complex suffix -jA-ƞAA-, based on the futuritive participle, e.g. PROGR-PTCP. FUT-1PL.INCL baka-ja-ƞaa-t ‘we (with you) shall find’. An indefinite present tense can, in principle, also be expressed by the finitized imperfective participle, but more often this form refers to a recent past, e.g. eme- ‘to come’ : PTCP.IMPRF-PX2SG eme-rii-s ‘you just came’, dialectally also to an immediate future, e.g. saa- ‘to know’ : PTCP.IMPRF-PX1SG saa-rii-w ‘I shall find out’. Simultaneous action, often in reference to the present tense range, can be marked by the habitive participle marker, which is normally used in combination with a copula, e.g. dukuu- ‘to write’ : PTCP.HAB dukuu-wkii + AOR-1SG bi-si-m ‘I am habitually writing’, while the finite uses of the iterative participle have mainly a past tense reference, e.g. beyukte- ‘to hunt’ : PTCP.ITER-PX1SG beyukte-ƞkii-w ‘I used to hunt’. The principal past tense paradigm is based on the perfective participle, which can also be combined with a copula to yield complex analytical past tense forms, e.g. waa- ‘to kill’ : PTCP.PRF-PX2PL waa-caasun ‘you killed’ : waa-caa-l bi-cee-sun ‘you had killed’ (kill-PTCP.PRF-PL COP-PTCP. PRF-PX2PL). A corresponding imperfective or progressive past tense is formed by adding the derivational suffix -jA- ~ -cA-, e.g. geleekte- ‘to search’ : PROGR-PTCP.PRFPX3PL geleekte-je-cee-tin ‘they were searching’. Another form marginally attested as a past tense marker (“the second past tense”) is the resultative participle, especially when

Siberian Ewenki 173

formed from intransitive verbs, e.g. xukti- ‘to run away’ : PTCP.RES-PX3SG xukti-ne-n ‘s/he ran away’. The same form without any personal markers is also used in the function of a second person imperative, in which case the plural is expressed by the nominal plural marker -l, e.g. PTCP RES eme-ne ‘(thou) come!’ : PL eme-ne-l ‘(you) come!’. Additional modal functions are expressed by the relevant participles in finitized use: -mAcin for the debitive mood, e.g. oo- ‘to do’ : PTCP.DEB-PX1SG oo-macin-mi ‘I have to do’, -rkA- for the probabilitative mood, e.g. eme- ‘to come’ : PTCP.PROB-PX3PL eme-rketin ‘they have probably come’, -ƞAAt.i- for the optative mood, e.g. baka- ‘to find’ : PTCP. OPT-PX1SG baka-ƞaat.i-w ‘I would like to find’, and -wkAA- ~ -pkAA- ~ -mukAA- for the necessitative mood, e.g. guun- ~ gun- ‘to say’ : PTCP NEC gun-mukee ‘it is necessary to say’. The perfective participle, when combined with the suffix -m(u)- for desideratives, yields the complex marker -m-cAA-, which expresses the subjunctive (conditional) mood, e.g. tewlesin- ‘to go picking berries’ : SUBJ-PX1PL.EXCL tewlesin.i-mcee-wun ‘we would go picking berries’. There are also a few deverbal derivational elements forming defective paradigms with grammaticalized modal functions. For unknown reasons, these elements are combined with a mixed set of personal markers incorporating features of both the aorist paradigm and the possessive suffixes, but they cannot be nominalized or converbalized. Examples are the suffixes -nAA-, expressing uncertainty (‘perhaps’), and -rgu-, expressing doubt (‘I wonder if’), both of which follow the regular aorist paradigm in the singular first and second persons, as well as in the plural first person exclusive and third person forms, but omit the aorist marker -rA- in the other forms, which can also have variable person markers (VX and PX) e.g. ana- ‘to push’ : 1SG ana-rgu-m : 2SG ana-rgu-nni : 3SG ana-rgu-n : 1PL.EXCL ana-rgu-ra-w ~ ana-rgura-wun : 1PL.INCL ana-rgu-ra-p ~ ana-rgu-t : 2PL ana-rgu-s ~ ana-rgu-sun : 3PL ana-rgu-ra. Another exceptional form is marked by the suffix -ƞnaa-, conventionally characterized as a “remote future imperative” (“let me later”, etc.), forming a “paradigm” together with the reflexive forms of the “supine”, e.g. tet- : tet.i- ‘to dress on’ : 1SG tet.i-ƞnee-m : RX = 2SG tet-tee-wii : 3SG tet.i-ƞnee-n : 1PL.EXCL tet.i-ƞneewun : 1PL.INCL tet.i-ƞnee-p : RX-PL = 2PL tet-tee-we-r : PL3 tet.i-ƞnee-tin. SYNTAX Although Ewenki as a whole follows relatively consistently the syntactic patterns congruent with “Ural-Altaic” typology, there are features that distinguish it from its southern neighbours, including not only the various Mongolic idioms, but also Manchu. Some of these features are particularly evident in Siberian Ewenki, while others are also shared by the other varieties of Ewenic. For instance, the order of constituents in phrases, clauses and complex sentences in Siberian Ewenki is considerably less strictly regulated than in several neighbouring languages with an otherwise similar typology. Although the basic rule of word order requires all adverbal modifiers to be placed before the verbal headword (SOV), there are frequent deviations from this rule, as in [1] (examples adapted from Kolesnikova 1966): [1] emu-keen beye xuw.u-l-le-n guluwuun-ma campfire-acc one-dim person blow-inch-aor-3sg ‘One person started to blow up the campfire.’

174 Juha Janhunen

Similarly, while adnominal modifiers normally precede the nominal headword (GAN), participial modifiers, in particular, are often placed after the headword, as in [2]: [2] kuƞa-kaa-r-wa eme-cee-l-we child-dim-pl-acc come-ptcp.prf-pl-acc ‘They brought the children who had arrived.’

[. . .] eme-w-re come-caus-aor.3pl

One factor that makes this relative freedom of word order possible is that an adnominal modifier normally agrees with the nominal headword in number and case, as in [3]: [3] jikte-l-we taalu-ma-l-duu tige-l-duu berry-pl-acc birchbark-dx-pl-dat vessel-pl-dat ‘They poured the berries into birchbark vessels.’

uƞkulbu-re pour-aor.3pl

The rules of agreement are, however, complex and dialectally variable. For instance, full agreement for both number and case between a nominal headword and a numeral modifier is typical only of the southern group of dialects, while the other dialects tend to have agreement only for case. An even more interesting type of non-agreement is present in examples like [4], where only the “modifier” takes case marking while only the “head noun” takes number marking. Since this is particularly common with numerals, the non-agreement probably signals that the relationship between the constituents is actually ambivalent. [4] etirkeen juur-we xinukii-l eme-w-re-n old.man two-acc hazel.grouse-pl come-caus-aor-3sg ‘The old man brought two hazel grouse.’ The lack of a proper genitive case means that possessive relationships have to be expressed mainly with the help of the possessive suffixes. In the nominal possessive construction, the unmarked possessor stands normally before the possessively marked head noun, e.g. mooti ulle-n ‘the meat of an/the elk’ (elk meat-PX3SG), xargii ulukii-l.i-n ‘the squirrels of the forest’ (forest squirrel-PL-PX3SG). Exceptionally, however, the possessively marked possessor can also follow the nominal headword, etirkeen adil.i-n ~ adil.i-n etirkeen ‘the old man’s net’ (old. man net-PX3SG ~ net-PX3SG old.man), juu-laa-tin ewenkii-l ‘to the dwelling of the Ewenki people’ (dwelling-LOC-PX3PL Ewenki-PL). Another use of the possessive construction is with the spatials (spatial nominals), often called “postpositions”, e.g. togo daga-duu-n ‘by the fire’ (fire side-DAT-PX3SG). The spatials, many of which are also used as regular nouns, indicate such specific spatial or temporal circumstances as are not expressed by the case suffixes, e.g. amar ‘back part’ > ‘behind’, doo ‘inner part’ > ‘inside’, dulin ‘center part’ > ‘in the middle of’, julee ‘fore part’ > ‘in front of’, oldon ‘side’ > ‘beside’, oyo ‘upper part’ > ‘above’, xergii ‘lower part’ > ‘below’, etc. The presence of two separate cases, the accusative and the partitive, for the differential marking of the direct object also has syntactic consequences. While the accusative marks a definite, specific and/or total object, as in [1] and [2], the partitive typically indicates an indefinite, generic, and/or partial object, as in [5]. [5] miti kete-ye 1pl.incl much-part ‘We caught a lot of fish.’

ollo-yo fish-part

waa-ra-p kill-aor-1pl.incl

Siberian Ewenki 175

More interestingly, the partitive, when combined with a suffix of personal or reflexive possession, can be used in a designative function, implying that something is intended for some particular purpose, as in [6]. This special use of the partitive, which has a parallel in Ewen, is possible since the inherent reference of this case (generic, indefinite) is otherwise not combinable with the reference of person marking (specific, definite). [6] bi sin-duu buu-je-m xute-(w)ii 1sg 2sg-dat give-progr-aor.1sg child-rx ‘I will give you my daughter to your wife.’

asii-ya-s woman-part-px2sg

The partitive case is also combined with existential negation, which itself is expressed by the privative noun aacin ‘absent, lacking’ : PL aaci-r, e.g. mudan-a aacin ‘there is no end, endless’ (end-PART absent), irgii-ye aacin ‘has no tail, tailless’ (tail-PART absent). As far as verbal negation is concerned, it is systematically expressed with the help of the negative auxiliary (negation verb) e- : AOR e-si- ‘(do) not’. All verbal predicates, including imperative forms, participles, and converbs, are negated by the corresponding forms of the negative auxiliary in combination with the connegative form (= aorist stem) of the main verb, e.g. baka- ‘to find’ : AOR = CONNEG baka-ra : NEG-IMP.1SG e-kte baka-ra : NEG-FUT-PX3SG e-jeƞee-n baka-ra : NEG-SUBJ-PX2SG e-mcee-s baka-ra, etc. However, the negative auxiliary does not take productive derivational suffixes, which are, therefore, attached to the main verb, e.g. PROGR baka-ja- : e-cee-n baka-ja-ra ‘s/he was not finding’ (NEG-PTCP.PRF-PX3SG find-PROGR-CONNEG). The negative auxiliary also has a few exceptional forms and stems, including e-cin : PL e-ci-r (participial past, with a possessive paradigm) and e-ji (second person singular prohibitive), whose dialectal status varies. A lexicalized derivative of e- with the implication of an emphatic future is EMPH e-tee- ‘to be unwilling (to do something)’, which is also combined with the connegative form of the main verb, e.g. e-tee-m gun-e ‘I do not want to say’ (NEGEMPH-AOR.1SG say-CONNEG). In equative sentences, nominal predicates are relatively seldom used without a copula, though occasional examples are attested for all types of nominals, including nouns, adjectives, numerals, and pronouns. Much more commonly, a copula is used, normally a finite form of bi- : AOR bi-si- ‘to be’, e.g. aya ‘good, fine, well’ : AOR 1SG aya bi-si-m ‘I am fine’. The copula is always necessary in the cases when specific temporal, aspectual, or modal meanings have to be expressed, e.g. iƞinipcu bi-cee-n ‘(it) was cold’ (cold COP-PTCP.PRF-3SG), xuƞtu bi-pkii ‘(it) is usually different’ (different COP-PTCP. HAB), and it can also accompany existential negation, e.g. aacin bi-cee-n ‘there was not’ (NEG.EXIST COP-PTCP.PRF-3SG) : aaci-r bi-cee-tin ‘there were not’ (NEG.EXIST-PL COP-PTCP.PRF-3PL). The copula itself is negated regularly by the negative auxiliary e-, e.g. e-si-n bi-si ‘(it) is not’ (NEG-AOR-3SG COP-CONNEG) : e-cee-n bi-si ‘(it) was not’ (NEG-PTCP.PRF-3SG COP-CONNEG). Another copular verb is oo- : AOR oo-da- ‘to become’, e.g. xegde-l oo-da-s ‘you (all) have grown big’ (big-PL becomeAOR-2PL), also with existential negation, e.g. AOR 3SG aacin oo-da-n ‘was lost’ (NEG. EXIST become-AOR-3SG). The copula oo- is also used in combination with the comparative forms in -dimAr ~ -timAr ~ -tmAr, e.g. oo-da-n bayi-tmar ‘(s/he) became richer’ (become-AOR-3SG rich-COMP). Apart from the classes of nominals and verb(al)s, both of which are characterized by a rich inflectional and derivational morphology, Ewenki has a small class of words that may be classified as “invariables”. Many of these are actually petrified non-productive adverbial forms of regular nominals, including numerals, pronouns and spatials, but others are

176 Juha Janhunen

etymologically isolated and may also represent borrowings from other languages, e.g. ñaan ~ ñaani ‘again’ (native), arai ‘hardly’ (from Mongolic), mene ~ meene ‘in vain’ (from Yakut). There is also a number of semi-independent postclitics, whose status as bound elements is confirmed by their harmonical behaviour, e.g. =kA, as used to express admonition in combination with the first person plural inclusive future forms, e.g. oo- ‘to do’ : oo-ja-p=ka ‘let us do!’, suruu- ‘to leave’ ITR : suruu-je-p=ke ‘let us leave!’. Of syntactic relevance are, in particular, the clitics =gu ~ =ku ~ =ƞu for interrogation (polar questions), e.g. 1SG bi : INTERR bii=gu ‘(do you mean) me?’, aya bi-si-ndi=gu ‘are you fine?” = ‘how are you?’ (good COP-AOR-2SG=INTERR), and =dAA ~ =tAA for additive emphasis (‘also, even, again’), e.g. 1SG bii=dee ‘even me’. When repeated, the latter two elements are also used in the function of connectors (‘either or’, ‘both and’), e.g. eme-cee=gu e‑cee=gu ‘did s/he come or not?’ (come-PTCP.PRF=INTERR NEG-PTCP. PRF=INTERR), aamut=taa bira=daa ‘both lakes and rivers’ (lake=ADD river=ADD). There are no actual conjunctions in Ewenki. LEXICON AND LANGUAGE CONTACTS Due to the relatively shallow depth and coherent structure of the Tungusic language family, Siberian Ewenki shares a considerable part of its lexicon with the other Tungusic languages, including, especially, the other varieties of Ewenic. Even after the dissolving of Proto-Ewenic, lexical items have travelled freely between the individual Ewenic languages, which ultimately spread to cover almost all parts of Northeast Asia. Considering the fact that Siberian Ewenki must have absorbed and extinguished a substantial number of earlier Pre-Tungusic languages of this macroscopic region, the amount of recognizable substratal lexicon is surprisingly small. This may, however, be an illusion, caused only by the fact that no systematic substratal lexical study of Ewenki has ever been carried out. It may be assumed that substratal lexical items are particularly abundant in certain semantic fields, including, for instance, botanical and topographical terminology and toponymy. What is, however, well known is that Ewenki has interacted lexically with all of its historically attested neighbours, which include Mongolic in the south, Amuric in the east, Kolymic in the northeast, Turkic in the north, as well as Samoyedic in the west and northwest. In most cases, the interaction has started at a local level, but some lexical items have subsequently moved over wide distances, reaching dialects spoken far away from the original contact area. In most cases, Ewenki has been the receiving partner in the contact relationship, but occasionally Ewenki lexical items have also been transmitted to the neighbouring languages. Obviously, the number and distribution of loanwords received by Ewenki reflects directly the date and intensity of the contact: the greater the number and the wider the distribution of a layer of loanwords is, the older and/or more intensive the contact has been. The following is a brief survey of the non-Tungusic lexicon in Ewenki, as arranged according to the contact language families and the corresponding regions: (1) Mongolic: Contacts between Mongolic and Tungusic date back to times before the protolanguages, and Ewenki preserves its share of the older layers of Mongolic loanwords. The interaction continued between Proto-Mongolic and Proto-Ewenic, which were apparently contemporary languages, both dated to around 1000 AZ, and both spoken in the Amur basin—Mongolic in the Amur source region (Onon, Argun) and Ewenic in the Middle Amur (and Zeya) basin. The distribution of the Mongolic

Siberian Ewenki 177

elements in Ewenki (and the rest of Tungusic), the number of which rises to several hundred items, shows that the degree of their geographical diffusion varies in a more or less direct relation to the distance from Mongolic (as shown in detail by Doerfer 1985). Thus, the dialects spoken in the western (Cis-Baikalian) part of the Ewenki language area have considerably fewer Mongolic elements than the dialects spoken in the eastern (Trans-Baikalian) part. While most of these elements reflect a Proto-Ewenic time level, contacts between Mongolic and Ewenki have continued locally until modern times in two locations: the Baikal region with Buryat, and the Amur source region with Khamnigan Mongol as the Mongolic partner. The interaction between Khamnigan Mongol and Khamnigan Ewenki is particularly illustrative from the diachronic point of view, for it shows how easily inherited lexicon can be replaced by loanwords under conditions of systematic ethnic bilingualism. For instance, Khamnigan Ewenki uses Mongolic loanwords for concepts like unsi- ‘to read’ and bici- ‘to write’, while the actual Siberian Ewenki dialects use the native words taƞ.i- ‘to read’ and dukuu- ‘to write’. A similar relationship of intensive contact exists between Solon and Daghur. (2) Amuric: The language family with which Tungusic has contacted since ancient times in the Sungari-Amur basin is Amuric, today represented by the Nivkh and Nighvng languages, collectively known as Ghilyak. While most of the interaction between Amuric and Tungusic has concerned the Jurchenic and Nanaic branches, Ewenic also came into direct contact with Amuric in the Amgun’ basin on the Lower Amur, where a mixture of Ewenki speakers and a local Nivkh population resulted in the formation of the Neghidal ethnic group. Occasional Amuric lexical items can also have spread to the actual Siberian Ewenki dialects, a notable, though somewhat problematic example being Ewenki loƞjama ‘the first quarter of the moon’, which seems to be based on the Amuric basic vocabulary item loƞ ‘moon’. For unknown reasons this item is attested as far west as the Stony Tunguska region. (3) Kolymic: The Kolymic family is today represented by the two Yukaghir languages Odul and Wadul, also known as Forest and Tundra Yukaghir, respectively. In historical times these languages have contacted mainly with Ewen (as well as with Yakut), but there are reasons to assume that the Kolymic family has moved to its modern location from the south, where its former territory may have been covered also by Siberian Ewenki. Ewenki has some lexical items with analogues in Yukaghir, an example being Ewenki yenee ~ yeƞee ~ eƞne ~ eƞñe ‘river’, which may be compared with Yukaghir enuƞ ‘river’, as well as with the name of the river Yana in northeastern Siberia. Unless this item is part of a general Siberian substratum, it is likely to involve an early borrowing from Kolymic to Ewenki, especially as it has no credible cognates in the other Tungusic languages. (4) Turkic: Although there is no evidence of early contacts between Turkic and Tungusic, the two language families met in central Siberia in the late Middle Ages, when Yakut spread northwards along the Lena basin, reaching areas previously occupied by Ewenic speakers. As a result, the Ewenki and Ewen dialects spoken in the neighbourhood of Yakut have borrowed a large number of Yakut lexical items representing all areas of life. In many cases, these items show unmistakable Yakut phonetic and/or semantic features, as in Ewenki (eastern dialects) tie ‘taiga’ ← Yakut tïa id. < Proto-Turkic *tag ‘mountain’, saas ‘age’ ← Yakut saas id. < Proto-Turkic *yaaš ‘year (of age)’. Since, however, Yakut has also contacted with Mongolic, many Mongolic items have reached Ewenki via Yakut, e.g. emte- ‘to cure’ ← Yakut

178 Juha Janhunen

em-tee- id., based on the root (*)em ‘medicine’, a widespread Central Eurasian cultural item of unknown origin, well attested also in Turkic, but transmitted to Yakut from Mongolic. A special position in the history of the Ewenki-Yakut contacts is occupied by the Dolgan, a recently (18th to 19th cc.) formed ethnic group in the Taimyr region, composed mainly of former Ewenki speakers who switched over to a dialectal variety of Yakut. (5) Samoyedic: The contact between Ewenki speakers and the Samoyeds dates back to a time immediately preceding the arrival of the first Russians in northwestern Siberia in the early 16th century. As an early token of the contact, the Common Samoyedic hydronym *yen(tə)si ‘Yenisei’ was transmitted to Ewenki in the form *yense-gii > yendegii ‘large river, Yenisei’. The principal Samoyedic groups with which Ewenki has been in direct contact include the Nganasan, Forest and Tundra Enets, Forest and Tundra Nenets, and Selkup (various dialectal groups). As a result, the northwestern dialects of Ewenki have come to share a number of regional lexical items with Samoyedic. In many cases it may be a question of local substratal lexicon, whose ultimate origin remains obscure, as in Ewenki caatira ~ caatara ‘polar fox’ = Nganasan satërë, Forest Enets sedoro < *sätərə. In the immediate neighbourhood of Samoyedic, especially in the Sym basin, Ewenki has also been in a potential contact with the Ob-Ugric (Khanty-Mansi) and Yeniseic (Ket-Yugh) languages, but the confirmed lexical traces of these contacts are minimal. In the most recent period, Ewenki has been under growing influence of the two colonial languages of Northeast Asia, Russian and Mandarin Chinese. Although contacts with Russian started as early as the 16th century, it was only during the Soviet period that virtually all speakers of Siberian Ewenki became bilingual in Russian, a development that has resulted in the loss of speakers by language shift and the increasingly endangered status of the Ewenki language. The impact of Chinese is even more recent but equally massive, though it affects only the Ewenic varieties spoken in China. While earlier Russian loanwords in Ewenki were often transmitted by Yakut, Siberian Ewenki today borrows freely from Russian, especially in the fields of social, administrative and technical vocabulary. In the Ewenki Cyrillic orthography, Russian elements follow the Russian normative spelling, which makes it difficult to estimate their degree of nativization at the level of the spoken language. Chinese loanwords that have spread to Siberian Ewenki are mainly connected with premodern cultural innovations, and many of them have been transmitted via Mongolic, as in Ewenki ganca ~ ganja ~ gansa ← Mongolic gaƞsa ← Mandarin ganzi 桿子 ‘tobacco pipe’. In general, lexical borrowings, even from the colonial languages, are not particularly abundant in Siberian Ewenki, and they do not pose a risk to language survival. What has become a risk is, however, the increasingly common phenomenon of code switching to Russian, which is leading to an intergenerational disruption in Ewenki language transmission. Compared with the regions where Russian is the dominant language of the immediate environment, the situation is considerably better in parts of Yakutia, where Yakut acts as a buffer between Ewenki and Russian under conditions of widespread trilingualism. A similar role of buffer language is traditionally played by Daghur for the Solon speakers in Manchuria, and by Khamnigan Mongol for the speakers of Khamnigan Ewenki, while the Orochen speakers are today directly exposed to the Chinese state language and, hence, more vulnerable to the effects of rapid linguistic assimilation.

Siberian Ewenki 179

REFERENCES AND FURTHER READING Andreeva, T. E. [Т. Е. Андреева] (1988) Звуковой строй томмотского говора эвенкийского языка: Экспериментально-фонетическое исследование [The sound structure of the Tommot dialect of Ewenki: An experimental-phonetic study], Новосибирск [Novosibirsk]: “Наука”. Atknine, Victor (1997) ‘The Evenki language from the Yenisei to Sakhalin’ in: Hiroshi Shoji & Juha Janhunen (eds.), Northern Minority Languages: Problems of Survival, Senri Ethnological Studies 44: 109–121, Osaka: National Museum of Ethnology. Boldyrev, B. V. [Б. В. Болдырев] (1988) Русско-эвенкийский словарь—Лӯчады-эведы турēрук [Russian-Ewenki dictionary], Москва [Moscow]: “Русский язык”. Boldyrev, B. V. [Б. В. Болдырев] (1994) Русско-эвенкийский словарь [Russian-Ewenki dictionary], Новосибриск [Novosibirsk]: Институт Языкознания РАН. Brodskaya, L. M. [Л. М. Бродская] (1988) Сложноподчиненные предложения в эвенкийском языке [Complex sentences in the Ewenki language], Новосибирск [Novosibirsk]: “Наука”. Bulatova, N. Ya. [Н. Я. Булатова] (1987) Говоры эвенков Амурской области [The Ewenki dialects of the Amur Oblast], Ленинград [Leningrad]: “Наука”. Bulatova, N. Ya. [Н. Я. Булатова] (1999) Язык сахалинских эвенков [The language of the Sakhalin Ewenki], Санкт-Петербург [St. Petersburg]: “Бесконфликтный Север”. Bulatova, Nadezhda Ja. & Lenore A. Grenoble (1999) Evenki, Languages of the World Materials 141, Munich: LINCOM Europa. Bulatova, N. Ya. [Н. Я. Булатова] & A. N. Myreeva [А. Н. Мыреева] (translators) (2011) Библия куӈакāрдȳ [Bible for children], Москва [Moscow]: Библиява хуӈтут тȳpэˉ рди дукудярӣ Институт. Castrén, M. Alexander (1856) Grundzüge einer tungusischen Sprachlehre nebst kurzem Wörterverzeichniss, Herausgegeben von Anton Schiefner, Nordische Reisen und Forschungen 9, St. Petersburg: Kaiserliche Akademie der Wissenschaften. Cincius, V. I. [В. И. Цинциус] (1997) ‘Эвенкийский язык’ [The Ewenki language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 267–284, Москва [Moscow]: “Индрик”. Doerfer, Gerhard (1983) ‘Das Birare’, Journal de la Société Finno-Ougrienne 78: 1–19, Helsinki. Doerfer, Gerhard (1985a) Mongolo-Tungusica, Tungusica, Band 3, Wiesbaden: Otto Harrassowitz. Doerfer, Gerhard (1985b) ‘Das Kamniganische’,『北方文化研究』Bulletin of the Institute for the Study of North Eurasian Cultures 17: 69–75, Sapporo. Doerfer, Gerhard & Michael Knüppel (2004) Etymologisch-ethnologisches Wörterbuch tungusischer Dialekte (vornehmlich der Mandschurei), Hildesheim: G. Olms. Gorcevskaya, V. A. [В. А. Горцевская] & V. D. Kolesnikova [В. Д. Колесникова] & O. A. Konstantinova [О. А. Константинова] (1958) Эвенкийско-русский словарь [Ewenki-Russian dictionary], Ленинград [Leningrad]: Государственное учебнопедагогическое издательство. Gorelova, L. M. [Л. М. Горелова] (1979) Категория вида в эвенкийском языке [The category of aspect in the Ewenki language], Новосибирск [Novosibirsk]: “Наука”. Helimski, Eugene & Juha Janhunen (1990) ‘Once more on the ethnonym “Tungus”’, In: Gedenkschrift für Irén N. Sebestyén, Specimina Sibirica 3: 67–72, Quinqueecclesiae (Pécs).

180 Juha Janhunen

Janhunen, Juha (1991) Material on Manchurian Khamnigan Evenki, Castrenianumin toimitteita 40, Helsinki: Finno-Ugrian Society. Khabtagaeva, Bayarma (2010) ‘Mongolic elements in Barguzin Evenki’, Acta Orientalia Academiae Scientiarum Hungaricae 63 (1): 9–25, Budapest. Khabtagaeva, Bayarma (2017) The Ewenki dialects of Buryatia and their relationship to Khamnigan Mongol, Tunguso-Sibirica 41, Wiesbaden: Harrassowitz Verlag. Khasanova, M. M. [М. М. Xасанова] (1986) Повелительное наклонение в эвенкийском языке [The imperative in the Ewenki language], Ленинград [Leningrad]: “Наука”. Kolesnikova, V. D. [В. Д. Колесникова] (1966) Синтаксис эвенкийского языка [Syntax of the Ewenki language], Москва [Moscow] & Ленинград [Leningrad]: “Наука”. Kolesnikova, V. D. [В. Д. Колесникова] (1983) Словарь эвенкийско-русский и русскоэвенкийский [Ewenki-Russian and Russian-Ewenki dictionary], Ленинград [Leningrad]: “Просвещение”, (2nd ed. 1989). Kolesnikova, V. D. [В. Д. Колесникова] & O. A. Konstantinova [О. А. Константинова] (1960) Русско-эвенкийский словарь [Russian-Ewenki dictionary], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Konstantinova, O A. [О. А. Константинова] (1964) Эвенкийский язык: Фонетика, морфология [The Ewenki language: Phonetics and morphology], Москва [Moscow] & Ленинград [Leningrad]: “Наука”. Konstantinova, O A. [О. А. Константинова] (1968) ‘Эвенкийский язык’ [The Ewenki language], in: Монгольские, тунгусо-маньчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 68–87, Ленинград [Leningrad]: “Наука” Konstantinova, O A. [О. А. Константинова] & E. P. Lebedeva [Е. П. Лебедева] (1953) Эвенкийский язык: Учебное пособие для педагогических училищ [The Ewenki language: A study tool for pedagogical colleges], Москва [Moscow] & Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Kotwicz, Władysław (1953) ‘Le dialecte tongous de Bargouzine: Matériaux recueillis par D. Rinčino’, Rocznik Orientalistyczny 16: 315–326, Kraków. Menges, K. H. (1968). ‘Das Ewenki’, in: Altaistik: Tungusologie, Handbuch der Orientalistik I.V.3: 42–92, Leiden: E. J. Brill. Morozova, O. N. [О. Н. Морозова] & S. V. Androsova [С. В. Андросова] & S. V. Kolesnikov [С. В. Колесников] (2019) ‘Длительность гласных в эвенкийском языке’ [Vowel quantity in the Ewenki language], Теоретическая и прикладная лингвистика 5 (1): 88–100, Благовещенск [Blagoveshhensk]. Myreeva, A. N. [А. Н. Мыреева] (ed.) (1990) Эвенкийские героические сказания [Ewenki epic tales], Памятники фольклора народов Сибири и Дальнего Востока 1, Новосибирск [Novosibirsk]: “Наука”. Myreeva, A. N. [А. Н. Мыреева] (2013) Дулин буга Торгандунин—Торгандун Среднего мира [Torgandun of the Middle World], G. I. Varlamova [Г. И. Варламова] (ed.), Новосибирск [Novosibirsk]: “Наука”. Nedjalkov, Igor (1994) ‘Evenki’, in: Peter Kahrel & René Van den Berg (eds.), Typological Studies in Negation, Typological Studies in Language 29: 1–34, Amsterdam: John Benjamins. Nedjalkov, Igor (1995) ‘Converbs in Evenki’, in: Martin Haspelmath & Ekkehard König (eds.), Converbs in Cross-linguistic Perspective, Empirical Approaches to Language Typology 13: 441–463, Berlin: Mouton de Gruyter. Nedjalkov, Igor (1997) Evenki, Routledge Descriptive Grammars, London and New York: Routledge.

Siberian Ewenki 181

Nedjalkov, Vladimir P. & Igor V. Nedjalkov (1988a) ‘Resultative, stative, passive, perfect in Evenki’, in: Vladimir P. Nedjalkov (ed.), Typology of Resultative Constructions, Typological Studies in Language 12: 241–257, Amsterdam: John Benjamins. Nedjalkov, Vladimir P. & Igor V. Nedjalkov (1988b) ‘Meanings of tense forms in Evenki’, Lingua Posnaniensis 29: 87–100, Poznań. Pallas, P. S. (1786–1789) Linguarum totius orbis vocabularia comparativa / Сравнительные словари всѣхъ языковъ и нарѣчiй, vols. 1–2, Petropoli / Вь Санктпетербургѣ: Тypis Iohannis Carolii Schnoor. Pevnov, A. M. [A. M. Певнов] (2020) ‘О происхождении самоназваний эвенков’ [On the origin of some Evenki endoethnonyms], Acta Linguistica Petropolitana 16 (3): 744–755, Санкт-Петербург [St. Petersburg]. Poppe, N. N. [Н. Н. Поппе] (1927) Материалы для исследования тунгусского языка: Наречие баргузинских тунгусов [Materials for the study of the Tungus language: The dialect of the Barguzin Tungus], Ленинград [Leningrad]: Издательство Академии Наук СССР. Procukovich, E. A. [Е. А. Процукович] & Yu. P. Ivanashko [Ю. П. Иванашко] & O. N. Morozova [О. Н. Морозова] (2018) ‘Переключение кодов в эвенкийском языке’ [Code switching in the Ewenki language], Теоретическая и прикладная лингвистика 4 (1): 58–68, Благовещенск [Blagoveshhensk]. Romanova, A. V. [А. В. Романова] & A. N. Myreeva [А. Н. Мыреева] (1962) Очерки токкинского и томмотского диалектов [Outlines of the Tokko and Tommot dialects], Москва [Moscow] & Ленинград [Leningrad]: “Наука”. Romanova, A. V. [А. В. Романова] & A. N. Myreeva [А. Н. Мыреева] (1964) Очерки учурского, майского и тоттинского говоров [Outlines of the Uchur, Maya and Totto dialect], Москва [Moscow] & Ленинград [Leningrad]: “Наука”. Romanova, A. V. [А. В. Романова] & A. N. Myreeva [А. Н. Мыреева] (1968) Диалектологический словарь эвенкийского языка: Материалы говоров эвенков Якутии [A dialectological dictionary of the Ewenki language: Materials on the dialects of the Ewenki of Yakutia], Ленинград [Leningrad]: “Наука”. Romanova, A. V. [А. В. Романова] & A. N. Myreeva [А. Н. Мыреева] (1971) Фольклор эвенков Якутии [Folklore of the Ewenki of Yakutia], Ленинград [Leningrad]: “Наука”. Romanova, A. V. [А. В. Романова] & A. N. Myreeva [А. Н. Мыреева] & P. P. Barashkov [П. П. Барашков] (1968) Взаимовлияние эвенкийского и якутского языков [Interaction of the Ewenki and Yakut languages], Ленинград [Leningrad]: “Наука”. Rudnickaya, E. L. [Е. Л. Рудницкая] (2019) Общая характеристика морфосинтаксиса устного эвенкийского языка начала XXI века [General characteristics of the morphosyntax of spoken Ewenki at the beginning of the 21st century], Санкт-Петербург [St. Petersburg]: Нестор-История. Schiefner, Anton (1859) ‘Beiträge zur Kenntniss der tungusischen Mundarten’, Bulletin de la Classe historico-philologique de l’Académie impériale des sciences de St.-Pétersbourg 16: 563–580. Schiefner, Anton (1881) ‘Alexander Czekanowski’s tungusisches Wörterverzeichniss’, Mélanges Asiatiques 8: 335–416, Saint-Pétersbourg. Shirokogoroff, S. M. (1933) Social Organization of the Northern Tungus, with Introductory Chapters Concerning Geographical Distribution and History of These Groups, Shanghai: The Commercial Press Ltd. Shirokogoroff, S. M. (1935) Psychomental Complex of the Tungus, London: Kegan Paul, Trench, Trubner & Co., Ltd.

182 Juha Janhunen

Shirokogoroff, S. M. (1944) A Tungus Dictionary: Tungus-Russian and Russian Tungus, Photogravured from the manuscripts, Shinobu Iwamura (ed.), Tokyo. Shirokogoroff, S. M. (1938) ‘Tungus literary language’, Introductory notes by Inoue Kōichi, Asian Folklore Studies 50: 35–66, Nagoya: Nanzan University (1991). Titov, E. I. [Е. И. Титов] (1926) Тунгусско-русский словарь с приложением книги М. А. Кастрена Основы изучения тунгусского языка [Tungus-Russian dictionary, with the book of M. A. Castrén “Elements of the study of the Tungus language”], Иркутск [Irkutsk]: Издание Читинского Краевого Государственного Музея им. А. К. Кузнецова. Varlamova, G. I. [Г. И. Варламова] (1986) Фразеологизмы в эвенкийском языке [Ewenki phraseology], Новосибирск [Novosibirsk]: “Наука”. Varlamova, G. I. [Г. И. Варламова] & A. N. Myreeva [А. Н. Мыреева] (eds.) (2014) Обрядовая поэзия и песни эвенков [Ewenki ritual poetry and songs], Памятники фольклора народов Сибири и Дальнего Востока 32, Новосибирск [Novosibirsk]: “Наука”. Varlamova, G. I. [Г. И. Варламова] & A. N. Varlamov [А. Н. Варламов] & N. E. Zakharova [Н. Е. Захарова] & M. P. Yakovleva [М. П. Яковлева] (2018) Имена собственные персонажей эвенкийскоцо эпоса: Словарь указатель [Proper names of the figures in Ewenki epics: A dictionary guide], Носибирск [Novosibirsk]: Издательство СО РАН. Vasilevich, G. M. [Г. М. Василевич] (1934a) Эвенкийско-русский (тунгусско-русский) диалектологический словарь [Ewenki-Russian (Tungus-Russian) dialectological dictionary], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Vasilevich, G. M. [Г. М. Василевич] (1934b) Учебник эвенкийского (тунгусского) языка [A textbook of the Ewenki (Tungus) language], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Vasilevich, G. M. [Г. М. Василевич] (1936) Материалы по эвенкийскому (тунгусскому) фольклору [Materials on Ewenki (Tungus) folklore], Вып. 1, Ленинград [Leningrad]: Издательство Института народов Севера ЦИК СССР им. П. Г. Симодовича. Vasilevich, G. M. [Г. М. Василевич] (1940a) Эвенкийско-русский (тунгусскорусский) словарь [Ewenki-Russian (Tungus-Russian) dictionary], Москва [Moscow]: Государственное из[дательст]во иностранных и национальных словарей. Vasilevich, G. M. [Г. М. Василевич] (1940b) Очерк грамматики эвенкийского (тунгусского) языка [Outline of Ewenki (Tungus) grammar], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Vasilevich, G. M. [Г. М. Василевич] (1948a) Русско-эвенкийский (русско-тунгусский) словарь [Russian-Ewenki (Russian-Tungus) dictionary], Mосква [Moscow]: Государственное издательство иностранных и национальных словарей. Vasilevich, G. M. [Г. М. Василевич] (1948b) Очерки диалектов эвенкийского (тунгусского) языка [Outlines of Ewenki (Tungus) dialects]. Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Vasilevich, G. M. [Г. М. Василевич] (1958) Эвенкийско-русский словарь [EwenkiRussian dictionary], Mосква [Moscow]: Государственное издательство иностранных и национальных словарей. Vasilevich, G. M. [Г. М. Василевич] (1966) Исторический фольклор эвенков: Сказания и предания [Historical folklore of the Ewenki: Tales and legends], Москва [Moscow] & Ленинград [Leningrad]: “Наука”.

Siberian Ewenki 183

Vasilevich, G. M. [Г. М. Василевич] (2005) Русско-эвенкийский словарь [RussianEwenki dictionary], Издание 2-е, переработанное, vols. 1–2, Санкт-Петербург [St. Petersburg]: “Просвещение”, online: http://evengus.ru/slovari/rus-evk/ Voskoboinikov, M. G. [М. Г. Воскобойников] (1958) Фольклор эвенков Бурятии [Folklore of the Ewenki of Buryatia], Улан-Удэ [Ulan-Ude]: Бурятское книжное издательство. Voskoboinikov, M. G. [М. Г. Воскобойников] (1960) Эвенкийский фольклор: Учебное пособие для педагогических училищ [Ewenki folklore: A textbook for pedagogical colleges], Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Voskoboinikov, M. G. [М. Г. Воскобойников] (1967) Фольклор эвенков Прибайкалья [Folklore of the Baikalian Ewenki], Улан-Удэ [Ulan-Ude]: Бурятское книжное издательство.

CHAPTER 8

OROCHEN Lindsay J. Whaley

The term Orochen (officially Pinyinized as Oroqen) refers, in the technical sense, to the Ewenic dialects spoken in northern Manchuria in the region of the Lesser and Greater Khingan Ranges (Chinese Xing’an Ling 興安嶺) on the Chinese side of the Upper and Middle Amur basin. In terms of the modern administrative division, clusters of Orochen speakers are dispersed across six townships in Heilongjiang Province and Hulun Buir City Prefecture, Inner Mongolia. In the context of the latter, the Orochen are the titular ethnic group of the Orochen Autonomous Banner (Elunchun zu Zizhiqi 鄂倫春族自治旗) with the center in the town of Alihe. Somewhat misleadingly, the ethnonym Orochen or Orochon (Russian orochën or orochón) is also used for the nearby groups of Siberian Ewenki speakers on the Russian side of the Amur, and occasional confusion takes place also with regard to the non-Ewenic ethnicities of Oroch (Russian óroch : PL órochi) and Orok (Russian órok : PL óroki). The ethnonym Orochen is used, in dialectal forms such as orocɪan ~ orosɪan ~ arʊsɪan, as the endonym of several Orochen groups. It was also used as an exonym by the Manchu in the shape oroncon (from which Chinese Elunchun 鄂倫春). However, several local groups today classified as “Orochen”, including, in particular, the Birarchen, Selpechen, and Kumarchen in the Lesser Khingan region, Heilongjiang, identify themselves by the names of the local rivers whose basins they inhabit. The name Orochen ~ Oronchon itself is of a disputed origin: most probably it is based on Tungusic oron ‘reindeer’ (< *ora/n > Manchu oron), but it has also been connected with Tungusic (x)oro(n) ‘mountain top’ (< *xoran > Ewenki dialectal oron) and Manchu oron ‘place’ (from Mongolic). In historical times the Orochen speakers have not had reindeer, and some groups have specifically identified themselves as mʊrcɪan ‘horse people’ (from mʊrɪn ‘horse’). Orochen is part of the complex dialect chain of Ewenic idioms which also comprises Solon, Khamnigan Ewenki, Neghidal, Siberian Ewenki, and Ewen. It should be noted that the division of all these idioms as separate entities is based as much on geography and the labels placed upon their speakers as it is on specific delineating linguistic criteria. That said, there are significant enough differences to treat these varieties as distinct languages. In China, the Orochen speakers are for both linguistic and, perhaps primarily, cultural and historical reasons, recognized as a distinct ethnic group (Elunchun zu 鄂倫 春族), while the speakers of Solon and Khamnigan Ewenki, together with the “Yakut Ewenki” or Manchurian Reindeer Tungus, are artificially classified as another ethnic group (Ewenke zu 鄂溫克族). Orochen itself falls along a dialect chain but, based on variation in the lexicon, sound correspondences, and morphological isoglosses, the language can usefully be divided into four dialects: western, central, northeastern and southeastern. Though these dialects are all mutually intelligible, speakers of the western dialect and the northeastern dialect DOI: 10.4324/9781315728391-8

Orochen 185

indicate some difficulties in understanding one another. The western dialect is geographically most proximate to Solon, and unsurprisingly, it uniquely shares some features with Solon. The same can be said of the northeastern dialect and Siberian Ewenki. This chapter describes the central dialect, as spoken in the Orochen Autonomous Banner. Until the early 1950s the Orochen lived primarily as nomadic subsistence hunters and fishers. After a forced settlement program they were concentrated into a small number of multiethnic villages. As a result there has been steady language shift towards monolingualism in Mandarin Chinese such that fluent speakers today are almost exclusively in their 60’s or older. The shift was precipitated by a fairly abrupt change in lifestyle to small-scale agriculture, a large influx of Mandarin speakers, and an educational system that is predominantly executed in Mandarin. The social disruption of the Cultural Revolution also played no small role in effecting language shift. Officially, the Orochen population has been growing, reaching ca. 8,700 individuals (2010), of whom some 45 per cent live in the Orochen Autonomous Banner and more than 50 per cent in Heilongjiang. The number of fluent speakers is no more than several hundred. There is, however, a significant number of semi-speakers with varying degrees of proficiency in the language. DATA AND SOURCES The data that form the main basis of this chapter were collected by Lindsay J. Whaley and Fengxiang Li over the course of eight years (1996–2005) during which they carried out six fieldwork trips in Inner Mongolia and Heilongjiang. These data are a combination of elicited word lists, sentences, and texts. Orochen is among the least studied Tungusic languages. In the early part of the 20th century, a relatively large database on the lexicon of several “Orochen” dialects, but no systematic grammatical information, was collected by the Russian ethnographer S. M. Shirokogoroff, as published by Doerfer & Knüppel (2004). On the basis of these materials Doerfer (1983) authored a diachronic study of the Birarchen dialect. An apparently related manuscript of materials for an “Orochen” dictionary by Shirokogoroff is in the process of being published, as reported by Lyudmila Missonova (2020). A problem with Shirokogoroff’s materials is, however, that they contain data mainly from the Siberian Ewenki dialects spoken on the Russian side and do not necessarily give a comprehensive picture of the Orochen dialects (proper) spoken on the Chinese side. For the central dialect of actual Orochen (proper), the monographs by Hu Zengyi (1986, 2001) remain the most complete description of the language to date, complemented by Sa Xirong (1981) and Zhang & Li & Zhang (1989). Other works containing dialectological information include Janhunen & Xu & Hou (1989), Whaley & Grenoble & Li (1999), and Whaley & Li (2000a). Li (1996) is a study of vowel harmony in several Tungusic languages. He includes valuable information on the differences among Orochen varieties. Other studies devoted to specific issues are Whaley & Li (1998, 2000b, 2012) and Lulich & Whaley (2012). SEGMENTAL STRUCTURE There are 18 native vowel phonemes in Orochen, of which 8 are short and 10 are long (Table 8.1). With rather minor phonetic differences, the system is similar to that attested in the eastern dialects of Ewen. Diachronically, it represents a verticalized version of the original Proto-Tungusic system.

186 Lindsay J. Whaley TABLE 8.1 OROCHEN VOWELS u uu

 

i ii

ʊ ʊʊ

ɪ ɪɪ

ɵ ɵɵ

e ee

ie

o oo

a aa

ɪa

As can be seen, the qualitative inventory follows a pattern that divides the short vowels into the four rounded back vowels u ʊ ɵ o (< *ü *u *ö *o), the two unrounded central vowels e a (< *e *a), and the two unrounded front vowels i ɪ (< *i *ï). In terms of tongue height, the vowels are organized into two sets, comprising the higher vowels u ɵ e i and the lower vowels ʊ o a ɪ. Apart from tongue height, the difference between these two sets may also be understood as involving tenseness (tense vs. lax) and tongue root position (advanced tongue root vs. retracted tongue root), with the higher vowels being pronounced more tense and with advanced tongue root (+ATR), while the lower vowels are pronounced more lax and with retracted tongue root −ATR or RTR). This means that the central vowel e is pronounced, as also in most other Tungusic languages, with the centralized quality [ə], while the rounded vowel o is pronounced with the relatively low quality [ɔ]. The system of long vowels (VV) includes additionally the segments here written ie ɪa, which represent the original Proto-Tungusic diphthongoid sequences *ie *ïa. The vowel ie is pronounced, as in Siberian Ewenki and Neghidal, as a long front monophthong [e:], while its lower counterpart ɪa still retains a palatal onglide and is pronounced [jɛ:]. A word-initial oo is pronounced with labial onglide, e.g. oo- [wɔ:-] ‘to do’. The consonant system (Table 8.2) also retains the Proto-Tungusic pattern and comprises 18 segments, divided into four places of articulation: labials (m b p w), dentals or alveolars (n d t s l r), palatals (ñ j c y), and velars (ƞ g k x), as well as six manners of articulation: nasals (m n ñ ƞ), weak stops (b d j g), strong stops (p t c k), fricatives (s x), glides (w y), a lateral (l), and a rhotic (r). The opposition between the weak and strong stops involves basically voice (weak = voiced vs. strong = unvoiced), but it may also involve aspiration (weak = plain or unspirated vs. strong = aspirated). The palatal stops are pronounced with palato-alveolar or alveolo-palatal sibilant affrication: c [tʃ] ~ [tɕ] vs. j [dʒ] ~ [dʑ]. The rhotic (r) is pronounced as a flap [ɾ].

TABLE 8.2 OROCHEN CONSONANTS m

n

ñ

ƞ

b

d

j

g

t

c

k

p

s w

x y

l r

Orochen 187

Several of the consonants have positional variants. Intervocalic k shows a tendency to lenite to [x] ~ [h], e.g. keeke ‘cat’ = [kə:kə] > [kə:xə] ~ [kə:hə]. Intervocalic g also weakens, though there is some variation as to whether the lenition results in [ɣ] or whether it elides and results in a long vowel, e.g. togo ‘fire’ = [tɔgɔ] > [tɔɣɔ] ~ [tɔ:]. The sibilant s is pronounced as a “hissing” dental or alveolar [s] before non-front vowels, and as a “hushing” palato-alveolar or alveolo-palatal [ʃ] ~ [ɕ] before front vowels, e.g. 2PL suu = [su:] vs. 2SG sii = [ʃi:] ~ [ɕi:]. Intervocalic s also shows a tendency to be desibilized to x [h] ~ [x], e.g. asɪn- ‘to sleep’ = [aʃɪn-] > [ahɪn-] ~ [axɪn-]. This means that the velar fricative x is in Orochen basically an intervocalic variant of both k and s, though it can also appear in recent Chinese loanwords. The Ewenic *x (< Proto-Tungusic *p), still preserved in Neghidal and several Siberian Ewenki dialects, has been lost in Orochen, as also in Solon, e.g. ʊnajɪ ‘daughter’ (< *xunaajï). As a marginal phoneme in Chinese loanwords, Orochen has also the bilabial fricative f [φ], e.g. feytin ‘airplane’ ← Chinese feiji 飛機. PHONOTACTICS AND MORPHOPHONOLOGY The syllabic template for Orochen roots is (C)V(V)(C)(C), thus one finds V (e-mun ‘one’), VV (oo- ‘to do’), CV (be-ye ‘man’), CVV (moo ‘tree’), VC (ɪm- ‘to drink’), VVC (oon ‘how?’), CVC (ber ‘bow’), CVVC (juur ‘two’). In all positions, but especially in non-initial syllables, vowel length (VV) is conspicuously unstable (V/V). When complex codas arise, the first of the two consonants is always the palatal glide y (uyl- ‘to pick up’). The consonants b d g j c ñ are only found in onsets. Some suffixes, e.g. the suffix of the terminative converb in -mdUr, permit CCV structure. Sequences of vowels other than the long vowels are not permitted. The upper limit of syllables in a root is four, though this is rare. In multi-morphemic words, potential violations of these phonotactic rules arise, and they are resolved by inserting an epenthetic vowel, which can be either A (= a e), e.g. alaat- ‘to wait’ :CV.TERM alaat.a-mdʊr or I (= i ɪ), e.g. mʊrɪn ‘horse’ : PX3SG mʊrɪn.ɪ-n, or also, after a labial consonant, U (= u ʊ), e.g. jeb- ‘to eat’ : AOR 1SG jeb.u-m. A special stem type is formed by words ending in the primary unstable nasal /n, which can be absent before derivational suffixes, e.g. mʊrɪn = mʊrɪ/n ‘horse’ : mʊrɪ-ksa ‘horse hide’. Exceptionally, other consonants can also be lost in derivation, e.g. ekur ‘cow’ : eku-kse ‘cow hide’. Suffix-initial alternations are conditioned by the interaction of stem-final and suffix-initial consonants. For instance, the accusative case marker -BA, which has the allomorph -wA after stems ending in a sonorant, has the allomorph -mA after stems ending in a nasal, and the allomorph -pA after stems ending in an obstruent. Also, in verbal inflection, special historically conditioned rules govern the formation of the aorist stem. The vowels in non-initial syllables follow the rules of vowel harmony, which operates in a left to right direction. The harmonic system may be described in terms of tongue height and/or tongue root position (Table 8.3). The vowels are divided into two categories, higher (+ATR) and lower (−ATR), and the choice of suffixal vowels is determined by the harmonic status of the vowel of the final syllable of the stem. In the speech of some elderly speakers rudiments of an erstwhile rounding (labial) harmony can also be found, but in general rounding harmony has been lost in Orochen. TABLE 8.3 OROCHEN VOWEL HARMONY +ATR

u uu

ɵ ɵɵ

e ee

i ii

ie

−ATR

ʊ ʊʊ

o oo

a aa

ɪ ɪɪ

ɪa

188 Lindsay J. Whaley

Word level stress (or prominence) is strictly word final in Orochen. Vowels in a stressed syllable are produced in a slightly higher position (with a higher pitch) than their non-stressed counterparts and with greater duration. At the level of more complex utterances, declarative sentences and content questions use a slightly falling intonation, while polar questions use a sharply rising intonation at the end of the sentences. Imperatives employ a sharply falling intonation. WORD FORMATION The basic distinction between nominals and verb(al)s is evident in both inflectional and derivational morphology. In the following, the principal derivational patterns are listed for denominal and deverbal nominals, as well as for denominal and deverbal verbs. Deverbal verbs with grammaticalized functions connected with voice, mood, aspect, and Aktionsart, as well as participles, which are nominalized verbs with verbal properties, are dealt with in connection with verbal morphology. The derivational patterns of adjectives, numerals, and pronouns are also discussed in the relevant sections below. (1) Denominal nouns. The following suffixes are highly productive and can be affixed to any semantically suitable nominal root: -kAn [diminutives indicating small physical size or an endearing quality], e.g. moo ‘tree, wood’ : moo-kan ‘stick’, kʊmaka ‘deer’ : kʊmaka-kan ‘small deer, lovely deer’; -kAncA/A [contempt], e.g. koto ‘knife’ : koto-konco ‘that useless knife’, owon ‘cake, bun’ : owon-konco ‘that crappy cake’; -ksA [hide/skin of land animals], e.g. ekur ‘cow’ : eku-ksa ‘cow hide’, torokɪ ‘boar’ : torokɪ-ksa ‘boar hide’; -mñA/A [augmentatives indicating large physical size, intensity or a negative quality], e.g. ɪlɪksa ‘mucus’ : ɪlɪksa-mña ‘lots of (disgusting) mucus’, atikkacan ‘old woman’ : atikkaca-mña ‘large and old woman, very old woman’; -tIEp [suffixed to a time expression to refer to the fullness of that time], e.g. aƞƞanɪ ‘year’ : aƞƞanɪ-tɪap ‘the whole year’, bolo ‘autumn’ : bolo-tɪap ‘the whole autumn’; -cIEn [affixed to place names or ethnic names to create a collective noun for a group of people], e.g. gan ‘the Gan River’ : gan-cɪan ‘people from the Gan River area’, moƞgol ‘Mongol(ian)’ : moƞgol-cɪan ‘Mongol people’. The suffixes -kAn, -kAncA/A, and -mñA/A can also be attached to adjectives, numerals and pronouns. The following suffixes are non-productive and are found only on a limited number of noun roots: -ldUwUn [containers], e.g. awʊn ‘hat’ : awʊn-dʊwʊn ‘hat container’, imukse ‘oil’ : imukse-lduwun ‘oil container’, inme ‘needle’ : inme-lduwun ‘needle container’; muu ‘water’ : muu-lduwun ‘water container’; -mAAkAAn [overabundance], e.g. beyun ‘moose’ : beyun-meekeen ‘too many moose’, olo ‘fish’ : olomookoon ‘too many fish’, ponto ‘deer antler’ : ponto-mookoon ‘too many antlers’, sagdɪ ‘old (person)’ : sagdɪ-maakaan ‘too many old people’; -ptOk ~ -mkAk [like the productive suffix -tIEp attached to a time expression to refer to the fullness of that time], e.g. aƞƞanɪ ‘year’ : aƞƞanɪ-ptok ‘the whole year’, sikse ‘evening’ : sikse-ptɵk ‘the whole evening’, bɪaga ‘month’ : bɪaga-mkak ‘the whole month’ (not acceptable to most speakers), dolbo ‘night’ : dolbo-mkok ‘the whole night’, erde ‘morning’ : erde-mkek ‘the whole morning’ (not acceptable to most speakers), iniyi ‘day’ : iniyi-mkek ‘the whole day’ (not acceptable to most speakers); -ptIn [something falling short of an ideal], e.g. ɪgdɪƞkɪ ‘comb’ : ɪgdɪƞkɪ-ptɪn ‘poor comb’, mʊrɪn ‘horse’ : mʊrɪ-ptɪn ‘horse-like thing’, tupige ‘fruit’ : tupige-ptin ‘unripe fruit’; -ptUn

Orochen 189

[coverings], e.g. iike ‘pot’ : iike-ptun ‘bag for carrying the cooking utensils’, sɪan ‘ear’ : sɪa-ptʊn ‘ear muff’, urke ‘door’ : urke-ptun ‘curtain made for the door’, ʊñakan ‘ring finger’ : ʊñaka-ptʊn ‘ring’, ʊrʊwʊn ‘thumb’ : ʊrʊwʊ-ptʊn ‘thumb ring’; -rAgdA [unaccompanied entities], e.g. asɪ ‘woman’ : asɪ-ragda ‘unaccompanied woman’, mʊrɪn ‘horse’ : mʊrɪ-ragda ‘loner horse’, ñɪray ‘male’ : ñɪray-ragda ‘only male’; -rAn [with some kinship terms expressing non-biological relationship], e.g. amin ‘father’ : ami-ran ‘step-father’, ekin ‘sister’ : eki-ren ‘step sister’, enin ‘mother’ : eni-ren ‘step-mother’, ʊnaji : ʊnaji-ran ‘adopted daughter’, ute : ute-ren ‘adopted son’; -rUk [containers], e.g. damga ‘tobacco’ : damga-rak ‘container for cigarettes or tobacco’, imukse ‘oil’ : imukse-ruk ‘oil container’, inme ‘needle’ : inmeruk ‘needle container’; -cIn [agentives, also on some verbal roots], e.g. adʊn ‘group of horses’ : adʊ-cɪn ‘horseherder’; -wU [types of paths], e.g. okto ‘path’ : okto-wʊ ‘human trodden road’, tergen ‘vehicle’ : terge-wu ‘path made by a vehicle’; -wUn [instruments, more commonly on verbal roots], e.g. uliye ‘antler’ : uliye-wun ‘horn made from an antler for making deer calls’. (2) Deverbal nominals. The following are two productive nominalizing suffixes that attach to verbal roots: -mAl [resultant state, also used to derive adjectives from a verbal base], e.g. jirɵ- ‘to move across a surface’ : jirɵ-mel ‘mark left by movement across a surface’, uru- ‘to gather’ : uru-mel ‘heap’; -ƞkI [instrument used to carry out an action], e.g. jik- ‘to chop’ : jik.i-ƞki ‘cutting board’, erke- ‘to bundle, to wrap’ : erke-ƞki ‘rope used to bundle stuff’, gʊljɪ- ‘to latch’ : gʊljɪ-ƞkɪ ‘latch’, ɪgdɪ- ‘to comb’ : ɪgdɪ-ƞkɪ ‘comb’. The following are non-productive suffixes deriving nouns from verbs: -gAn [resultatives], e.g. bodo- ‘to calculate, to think’ : bodo-gon ‘strategy, imagination, thought’, kʊr- ‘to pen up’ : kʊr-gan ‘animal pen’, mada- ‘to expand’ : mada-gan ‘interest from money’, ñaro- ‘to paint’ : ñaro-gan ‘picture’, tokto- ‘to decide’ : tokto-gan ‘decision’; -lgA [nouns with an unpredictable semantics], e.g. bodo- ‘to calculate, to think’ : bodo-lgo ‘plan’, delki- ‘to split’ : delki-lge ‘method of splitting’, janda- ‘to sing’ : janda-lga ‘song’, edel- ‘to cross a river’ : edel-ge ‘place to cross a river’, eyi- ‘to play’ : eyi-lge ‘toys’, waa- ‘to kill’ : waa-lga ‘animals to be killed, method of slaughter’; -r [nouns with an unpredictable semantics], e.g. bʊdʊ- ‘to paint’ : bʊdʊ-r ‘dye’, esu- ‘to sweep’ : esu-r ‘broom’, too- ‘put in, to fill’ : too-r ‘big bullet, sound of a big bullet’; -ptIn [as with nominal roots, something that falls short of an ideal, but in some cases also with no obvious negative implication], e.g. erke‘to wrap round’ : erke-ptin ‘rope/yarn which is poor for wrapping’, gʊljɪ- ‘to lock, to latch’ : gʊljɪ-ptɪn ‘latch in a bad condition’, dɪaƞna- ‘to place a mat underneath’ : dɪaƞna-ptɪn ‘mat’, lipki- ‘to clog, to stop up’ : lipki-ptin ‘stopper’; -wUn [instrument used to carry out an action], e.g. ɪgdɪ- ‘to comb’ : ɪgdɪ-wʊn ‘comb (for humans)’, kadɪ- ‘cut’ : kadi-wʊn ‘knife, sickle’, kuci- ‘to wrap around’ : kuci-wun ‘string’, tukti- ‘to go up’ : tukti-wun ‘ladder’, cergu- ‘to make rope’ : cergu-wun ‘rope, tool for making rope’. (3) Denominal verbs. There are two productive verbalizing derivational strategies based on noun roots: -lAA- [general verbalizer], e.g. adʊn ‘herd of horses’ : adʊ-laa- ‘to herd horses’, muu ‘water’ : muu-lee- ‘to fetch water’; -mAA- [affixed to animal nouns to indicate hunting], e.g. kʊmaka ‘deer’ : kʊmaka-maa- ‘to hunt deer’, olo ‘fish’ : olo-moo- ‘to fish’. There are also some non-productive derivations of verbs from nouns: -dA- [to use an object for a typical use], e.g. jisu ‘colour’ : jisu-de- ‘to dye’, leke ‘whetstone’ :

190 Lindsay J. Whaley

leke-de- ‘to sharpen’; -sI- [indicating occupation during a given time period], e.g. bolo ‘autumn’ : bolo-sɪ- ‘to spend the autumn’, jugu ‘summer’ : jugu-si- ‘to spend the summer’; -tU [applied to animal names to indicate abundance or infestation], e.g. kumke ‘louse’ : kumke-tu- ‘to have lice’, uluki ‘squirrel’ : uluki-tu- ‘to have squirrels’. Verbs can also be derived from adjectival nominals using the following suffixes: (productive) -sIE-, e.g. aya ‘good’ : aya-sɪa- ‘to praise’, nicukun ‘small’ : nicukun-cie- ‘to look down upon’; (non-productive) -lgI-, eku ‘hot’ : eku-lgi- ‘to warm’, gɪlɪ ‘cool’ : gɪlɪ-lgɪ- ‘to cool’. (4) Deverbal verbs: Apart from the deverbal verbs with grammaticalized functions, there are a few non-productive suffixes that derive verbs from verb roots: -rgI- [reversives for movement back], e.g. eme- ‘to come’ : REV eme-rgi- ‘to return’, bu- ‘to give’ : REV bu-rgi- ‘to return’; -w- ~ -U- [causatives from stative verbs], e.g. eme- ‘to come’ : CAUS em-u- ‘to bring back’; ƞene- ‘to go’ : CAUS ƞen-u- ‘to bring’, si- ‘to be dark’ : CAUS si-w- ‘to extinguish’. NUMBER AND CASE Orochen retains the Ewenic basic nominal template in which the nominal root, followed by derivational suffixes, receives markers for number, case, and possession, in this order (N-DX-NX-CX-PX). However, unlike Siberian Ewenki, Neghidal and Ewen, but like Solon, Orochen uses plural marking relatively seldom, allowing the basic form of nominals to function as an unmarked generic plural. If speakers wish to make plurality explicit, the complex suffix -sAl can be employed, e.g. moo ‘tree’ : PL moo-sal. The simple suffix -l is retained in pronominal and verbal inflection, though dialectally it can also be used on nouns, e.g. beye ‘man, person’ : PL beye-l. The suffix *-r of primary nasal stems seems to be lost and can be replaced by -l, e.g. mʊrɪn ‘horse’ : PL mʊrɪ-l. With kinship terms the suffix -nAr (from Mongolic) can also be used, e.g. naacʊ ‘maternal uncle’ : PL naacʊ-nar. The case system comprises the unmarked nominative and 12 suffixally marked cases (Table 8.4), of which eight are attested with closely similar forms and functions in Siberian Ewenki and Neghidal: the accusative in -wA (after sonorants) ~ -pA (after obstruents) ~ -mA (after nasals), the partitive in -yA (after regular stems) ~ -A (after primary nasal stems), the dative in -dU ~ -tU, the ablative in -dU-k ~ -tU-k, the locative in -lA (after vowel stems) ~ -dU-lA ~ -tU-lA (after consonant stems), the prolative in -lI (after vowel stems) ~ -dU-lI ~ -tU-lI (after consonant stems), the directive in -kkI (after vowel stems) ~ tIkI (after consonant stems), and the instrumental in -jI ~ -cI. In addition, the possessive form in -ƞI, which in Siberian Ewenki and Neghidal is a derivational category, functions in Orochen as a true nominal case, as also in Solon. The three remaining cases are the delative in -lA-k ~ -dU-lA-k ~ -tU-lA-k, the lative in -kk-Aki ~ -tIk-Aki, and the elative in -yIjI-k ~ -IjI-k ~ -ƞIjI-k, also simplified to -IjI-k for all stem types (with the loss of stem-final vowels before the suffix). As may be seen, the delative, shared with Solon, is based on the locative, expanded by the element -k. The formal relationship between the locative and delative is analogous to that between the dative and ablative. The same element -k is also present in the elative, which otherwise corresponds to the elative in *-gII-jI in Siberian Ewenki and Ewen. Note that vowel length in several case suffixes is variable, in that the vowel A in the locative, delative, and elative, and the vowel I in the genitive and prolative, can also be long.

Orochen 191 TABLE 8.4 OROCHEN CASE MARKERS  

V

C

/n

POSS

-ƞI

ACC

-wA

PART

-yA

-BA

-mA

DAT

-DU

ABL

-DUk

LOC

-lA

-DUlA

DEL

-lAk

-DUlAk

PROL

-lI

-DUlI

DIR

-kkI

-tIkI

LAT

-kkAkI

-tIkAkI

INSTR

-jI

-CI

ELAT

-yIjIk

-IjIk

-A

-ƞIjIk

Stem types: V = vowel stems, C = consonant stems, /n = primary nasal stems. Segmental alternations: A = a e o ɵ, U = ʊ u, I = ɪ i, B = w m p, C = j c, D = d t.

•

•

•

•

The basic grammatical cases include the nominative, accusative, partitive, and possessive (genitive). The nominative is used for subjects, while the direct object is marked with either the accusative for definite objects or the partitive for indefinite objects, as in Siberian Ewenki. Less commonly, an indefinite object can also be unmarked. The possessive case indicates phrase-level possession, e.g. asɪ-ƞɪ ɪgdɪƞkɪ ‘the woman’s comb’ (woman-POSS comb). A noun preceded by a modifier in the possessive case may also take possessive suffixes. The dative has a wide range of uses. It can indicate recipients, benefactives, location, e.g. ikece-ken-du ‘in a small pot’ (pot-DIM-DAT), point in time, e.g. tarɪ erin-du ‘at that time’ (that time-DAT), and making comparisons, e.g. gʊgda min-du ‘taller than me’ (tall 1SG-DAT). The dative is also used to mark the agent in a passive clause, e.g. emun bey-mñaa-du saa-w-caa-n ‘it was known by a bad person’ (one personAUGM-DAT know-PASS-PST-3SG). The separative counterpart of the dative is the ablative, which indicates motion away from a specific point, e.g. ure-duk ‘from the mountain’ (mountain-ABL). By comparison, the elative indicates motion away from a reference point, but without a specified starting point, e.g. ur-ijik ‘away from the mountain’ (mountain-ELAT). The locative is used to express a destination, e.g. elgi-le ‘onto the earth’ (earthLOC). The corresponding separative case is the delative, which expresses the source of an event or action, e.g. ure-lek ‘out from the mountain’ (mountain-DEL). The prolative, which belongs to the same formal set, indicates a reference point along which an action occurs, e.g. ure-li ‘alongside the mountain’ (mountain-PROL). The difference between the directive and the lative is similar to that between the ablative and elative, in that the directive indicates motion towards a specific location, conceptualized as a point, e.g. ure-kki ‘to the mountain’ (mountain-DIR), whereas the lative is less specific and indicates motion towards an unspecified location that

192 Lindsay J. Whaley

•

is within the vicinity of a reference point, e.g. ure-kkeki ‘towards the mountain’ (mountain-LAT). The principal function of the instrumental is to express the instrument or means by which an action is accomplished, as in [1]:

[1] mergen beye amɪ-ƞɪ nɪptca-jɪ-n hunter person father-poss gun-instr-px3sg ‘The hunter shot a monster with his father’s gun.’

sɪkʊr monster

waa-ca kill-pst

When attached to adjectival nominals, the instrumental expresses manner, e.g. ayɪ-jɪ icere-n ‘s/he looked carefully’ (good-INSTR see-AOR-3SG). It can also function in the role of a comitative and express accompaniment, e.g. noonɪn mʊrɪn-jɪ aasɪn-a-n ‘he and his horse slept’ (3SG horse-INSTR sleep-AOR-3SG), or equivalence, as in [2]: [2] oƞto gek mʊrɪn-jɪ emun nadal other mare horse-instr one same ‘(She) was the same as the other mares.’

bi-ki-n be-aor-3sg

ADJECTIVES Adjectives in Orochen are placed before the noun they modify and do not show agreement with the noun. When used as independent head nouns, they take the normal nominal number and case markers, e.g. sagdɪ ‘old’ : sagdɪ-sal-dʊ ‘to the old ones’ (old-PL-DAT). Adjectives have, however, derivational patterns that are only partially shared with nouns. Adjectives can be derived from nouns, adjectives, and verbs. (1) Adjectives derived from nouns. The productive suffixes deriving adjectival nominals from nouns are the following: -ktUrA [possession of a large quantity of something], e.g. ɪƞakta ‘hair’ : ɪƞakta-ktʊra ‘hairy’, ure ‘mountain’ : ure-kture ‘mountainous’; -mAAkAAn [intensifier, also used with nouns], e.g. beyun ‘moose’ : beyu-meekeen ‘full of moose’, calban ‘white birch’ : calban-maakaan ‘full of white birch’; -mAAn [enjoyment of consuming something], e.g. arakɪ ‘liquor’ : arakɪ-maan ‘liquor-loving’, jepte ‘food’ : jepte-meen ‘food-loving’; -m(U)nAn [having the smell of something], e.g. beye ‘person’ : beye-munen ‘smelling like a human’, olo ‘fish’ : olo-mnon ‘fishy smelling’ (there is variation among speakers about whether the suffix is one or two syllables); -cI [proprietives], e.g. beye ‘person’ : PROPR beye-ci ‘pregnant’, muu ‘water’ : PROPR muu-ci ‘having water’. (2) Adjectives derived from adjectives. This group includes both suffixes that are specifically used on adjectival bases and others that can also be attached to nominal stems: -dA [intensifier], e.g. baraan ‘a lot’ : baraan-da ‘excessive(ly much)’, goro ‘far’ : goro-do ‘very far’; -kA [with color terms, expressing a dotted quality], e.g. bagda- : bagda-rɪn ‘white’ : bagda-ka ‘white-dotted’, kiruke ‘grey’ : kiruke-ke ‘white-dotted’; -kAn [intensifier, identical with the suffix for diminutives, as used on nouns and numerals], e.g. ayama ‘beautiful’ : ayama-kan ‘dear and beautiful’, eku ‘hot’ : eku-ken ‘rather hot’; -kIE [used on color terms as a hedge], e.g. kiruke ‘grey’ : kiruke-kie ‘greyish’, sɪlan ‘blue’ : sɪlan-kɪa ‘bluish’; -lbIE/n [used with color terms to specify the quality of lightness of hue], e.g. koƞna- : koƞna-rɪn ‘black’ : kɔƞna-lbɪa ‘light black’, sɪƞa- : sɪƞa-rɪn ‘yellow’ : sɪƞa-lbɪa (some speakers pronounce the suffix

Orochen 193

with a final n); -lI [intensifier], e.g. gʊgda ‘tall’ : gʊgda-lɪ ‘very tall’, ʊla- : ʊlarɪn ‘red’ : ʊla-lɪ ‘very red’ > ‘someone who has had too much to drink’; -mñA/A [increased intensity, identical with the augmentative suffix used on nouns], e.g. eku ‘hot’ : eku-mñe ‘very hot’, bulgu ‘fat’ : bulgu-mñe ‘very fat’; -cIrA [attenuation], e.g. yoldo ‘lean’ : yoldo-cire ‘somewhat lean’, daa ‘near’ : daa-cɪra ‘a bit nearer’. Orochen also uses partial reduplication to express the increased intensity on some color terms. This is the only instance of prefixation in the language, e.g. bagdarɪn ‘white’ : bag & bagdarɪn ‘very white’; kara ‘dark’ : ka-p & kara ‘very dark’. This pattern, absent in Siberian Ewenki, but well known from Mongolic and Turkic, is also attested in Solon and apparently involves Mongolic influence on Manchurian Ewenic. (3) Adjectives derived from verb roots. The following are productive suffixes of this category: -lAn [enjoyment of an activity], e.g. janda- ‘to sing’ : janda-lan ‘enjoying singing’; soƞo- ‘to cry’ : soƞo-lon ‘enjoying crying’; -mA [tendency], e.g. aƞka- ‘to be thirsty’ : aƞka-ma ‘tending to be thirsty’, soƞo- ‘to cry’ : soƞo-mo ‘tending to cry’; -mAk [having the property of immediately doing or completing an action], e.g. jeb‘to eat’ : jeb.u-mek ‘eating right away’, ɪlan- ‘to stand up’ : ɪla-mak ‘standing right up’; -mAl [resultant state], e.g. akkɪ- ‘to sew’ : akkɪ-mal ‘sewn’, uruu- ‘to gather’: uruu-mel ‘heap’; -mUgdI [possessing a certain property, used also on nouns to derive adjectives], e.g. iñe- ‘to laugh’ : iñe-mugdi ‘laughable’, panca- ‘to be angry’ : panca-mʊgdɪ ‘full of anger’; -ptAn [ability to be the object of an action], e.g. jeb- ‘to eat’ : jeb.u-pten ‘edible’, ɪgdɪ- ‘to comb’ : ɪgdɪ-ptan ‘combable’. In addition there are two unproductive suffixes: -mAktA [having the property of something that has just occurred], e.g. eme- ‘to come’ : eme-mekte ‘having just come’, dalga- ‘to roast’ : dalga-makta ‘having just been roasted’ (this suffix is actually productive in the northeastern dialect); -ktA ~ -ktU ~ -tU [prone towards an activity], e.g. aamɪ- ‘to be sleepy’ : aamɪ-ktʊ ‘prone to sleeping’, gerbe- ‘to work’ : gerbe-ktu ‘one who likes to work’, iñe- ‘to laugh’ : iñe-kte ‘one who likes to laugh’, kʊlaka- ‘to steal’ : kʊlaka-tʊ ‘prone to theft’, soƞo- ‘to cry’ : soƞo-to ‘one who likes to cry’. NUMERALS The lexically independent cardinal numerals in Orochen are, for the digits: 1 umun ~ emun, 2 juur, 3 ɪlan, 4 diyin, 5 tʊƞƞa, 6 ñuƞun, 7 nadan, 8 japkʊn, 9 yeyin, 10 jaan; for the decades: 20 ʊrɪn, 30 gʊtɪn, 40 deki, 90 yereen; and for the higher numerals 100 ñamaajɪ, 1000 mɪƞga ‘thousand’, 10,000 tume/e. Among these, the items for 1 to 10 and 100 are native Tungusic, while the other items are loanwords from Mongolic. The system as a whole is similar to Solon, which is the only other Ewenic language that has cognates for the items 40 deki (an old borrowing from Pre-Proto-Mongolic, possibly transmitted by Manchu) and 90 yereen (a relatively recent borrowing from Mongolic). The item 30 gʊtɪn has a cognate in the Borzya dialect of Khamnigan Ewenki, the item 20 ʊrɪn also in Neghidal. It may be noted that Orochen, like Solon, retains the Ewenic item 100 ñamaajɪ, which in Neghidal has been replaced by the Manchu loanword taƞgu. All other cardinal numerals are created from these elements. The numerals for the decades 50 to 80 are formed by the suffix -ƞI, attached to the items for the corresponding digits: 50 tʊƞƞa-ƞɪ, 60 ñuƞun-ƞi, 70 nadan-ƞɪ, 80 japkʊn-ƞɪ. Other numerals are formed by the additive and multiplicative principles, e.g. 11 jaan umun, 12 jaan juur, 21 ʊrin umun, 191 umun ñamaajɪ yereen umun, 90,000 yeyin tumee, etc.

194 Lindsay J. Whaley

The cardinal numerals are used for counting, as nominal modifiers, e.g. tʊƞƞa mʊrɪn ‘five horses’, or as substantives, that is, independent head nouns. The substantive use of a numeral is, however, unusual, except in derived forms. Numerals can be modified by two words, adɪɪ ‘approximately, several’ (originally < ‘how many?’) and top ‘exactly’. These modifiers are placed before the numeral. The numeral roots can take derivational suffixes, some of which are specific to numerals, while others are also used with other nominals. Ordinal numerals are formed with the suffix -ki, e.g. emun-ki ‘first’, jaan juur-ki ‘twelfth’. Before most other derivational suffixes, the final consonant of the numeral stem is dropped, as in the distributives in -tAl, e.g. jaa-tal ‘ten each, ten by ten, for ten hours’, and the diminutives (in the meaning ‘only’) in -kAn, e.g. umu-ken ‘only one’. The nasal is retained, however, in the collective numerals with an additional meaning of contempt, formed by the suffix -kAncA/A (also affixed to nouns and pronouns), e.g. ɪlan-kancaa ‘those three scoundrels’. Of greatest interest are the suffixal counters attached to numeral roots to express the number of countable entities. Several of these are recorded from Orochen: -lA for the number of days, e.g. juu-le ‘two days’, -mAn for the number of prongs on a moose or deer’s antlers (originally used for any repeated entities or “tuples”), e.g. ɪla-man ‘threepronged’, -nA for the number of animals comprising a group, e.g. ɪla-na ‘(a group of) three animals’, -ƞnA for the number of animals not necessarily gathered in a group, e.g. uri-ƞne ‘twenty animals’, -rA for the number of repetitions (‘times’), e.g. juu-re ‘twice’, ɪla-ra ‘three times’. Some of these forms are no longer productive in all dialects, and typically speakers accept the use of many suffixes only with small numbers. The method by which fractions are expressed is to place ubu-ƞi, which is the possessive form of ubu ‘portion’, after the denominator, which is expressed by a cardinal number, and before the numerator, which is also a cardinal number, e.g. juur ubu-ƞi emun ~ juur ubu-ƞi emun-in (one-PX3SG) ‘one-half’ (literally: ‘one of two portions’). The use of the possessive suffix on the numerator is optional. PRONOUNS The Orochen basic personal pronouns (Table 8.5) retain the original Ewenic system more or less intact in virtually the same form as attested in Siberian Ewenki. The non-nominative forms are based on an oblique stem in -n, and the object form involves a combination of the partitive and accusative markers. Unlike Siberian Ewenki, however, Orochen uses the possessive form as a true case form. TABLE 8.5 OROCHEN PERSONAL PRONOUNS   SG PL

NOM

OBL

PART-ACC

POSS

1

bii

min-

min-e-we

min-ƞi

2

sii

sin-

sin-e-we

sin-ƞi

buu

mun-

mun-e-we

mun-ƞi

suu

sun-

sun-e-we

sun-ƞi

1 2

EXCL

The first person plural pronoun buu : OBL mun- has an exclusive function, while the corresponding inclusive pronoun is miti : OBL miti-. The third person pronoun has the stem noon-, simplified from the dialectally attested full form nʊgan-. In the singular the third person possessive suffix is added in all cases: NOM-PX3SG noon.i-n : OBL noon-CXPX3SG, e.g. ABL noon-dʊk.ɪ-n, while in the plural the original plural stem noo-r- is used:

Orochen 195

NOM-PL-PX3PL noo-r-tɪn : OBL noo-r-CX-PX3PL, e.g. INSTR noo-r-jɪ-tɪn. This pronoun is a rare example in which the plural marker -r is preserved in Orochen. The reflexive pronoun is meen-. This pronoun carries a distinct semantic role from its antecedent in a clause. It is inflected for case with the regular case markers, followed by the reflexive suffix -BI for singular and -BAl for plural reference, e.g. bii meen-du-wi jepte olo-m ‘I cooked a meal for myself’ (1SG REFL-DAT-RX meal cook-AOR1SG). In object position no accusative marking is present, e.g. balɪ beye meen-mel aysɪlaa-ca ‘blind people helped themselves’ (blind man REFL-RX.PL help-PST). In emphatic use, when the reflexive pronoun and its antecedent have the same semantic role in the clause, the derived form meen-eke is used, e.g. noonɪn meen-eke jep-te ‘he himself is eating’ ~ ‘he is eating by himself’ (3SG REFL-DX eat-AOR). The demonstrative pronouns are (proximal) eri ‘this’ : PL eri-l ‘these’ vs. (distal) tarɪ ‘that’ : PL tarɪ-l ‘those’. The demonstratives can be used either independently, with the regular case inflection, or as adnominal modifiers, with no agreement with the head noun. Also, since the language has no other means, such as articles, to indicate definiteness, demonstratives are commonly used in this role when the context is not clear. The basic interrogative pronouns are nii ‘who?’ and ɪkʊn ‘what?’. The root ɪ- (< *ïa-) is also present in ɪ-da ‘why?’ and ɪa-ma ‘which?’. Other interrogative words are based on the roots i- (< *ii-), e.g. DAT i-du ‘whither?’ : LOC i-le ‘where?’; o- (< *oo-), e.g. oo-kɪ ‘how many?’ : oo-n ‘how?’; and aa- (< *xa-), e.g. PROL aa-lɪ ‘when?’. In interrogative sentences, the question words remain in situ, e.g. sii nii-we ice-ce-y ‘whom did you see?’ (2SG who see-PST-2SG), suu tarɪ juu-we oon ɪlɪw-ra-y ‘how are you constructing that house?’ (2PL that house-ACC how build-AOR-2PL). Pronouns can take a few derivational suffixes. For instance, the suffix -kAncA/A can be used on personal and demonstrative pronouns to express contempt, e.g. 2SG sii : sii-kencee ‘you bad boy/girl’, PL tarɪ-l ‘those’ : tarɪ-l-kancaa ‘those scoundrels’. PERSON MARKING Orochen preserves the general Ewenic pattern of two sets of person markers with some idiosyncracies (Table 8.6). A unique development is the replacement of the suffix (*)-s in TABLE 8.6 OROCHEN PERSON MARKERS        SG

PX  CV

1

-I-

-w

-m

2

-I-

-y

-ni

3

-I-

-n

RX PL

VX 

C

1

PTCP

-Ø

AOR

-n

-BI EXCL

-BUn

INCL

-tIr

2

-sUn

3

-tIn

RX

-BAl

-w -r

-p -y

-l

-Ø

Stem types: C = consonant stems (with a connective vowel), V = vowel stems. Segmental alternations: A = a e o ɵ, U = ʊ u, I = ɪ i, B = w m p.

196 Lindsay J. Whaley

both the second person singular (*-sI) of the possessive set (PX) and the second person plural (*-sU < *-sO) of the predicative set (VX) by -y. As in the other Ewenic languages, the predicative set is used with the aorist type of finite paradigm, while the possessive set marks nominal possession, but also verbal predication when the predicate is a nominalized verb. A secondary difference has arisen between the nominal and verbal uses of the possessive set in the first person plural inclusive forms, which are -tIr in nominal vs. -r in verbal use. In the predicative use of the possessive set, the third person forms are replaced by the nominal number markers, zero for the singular and -l for the plural. The allomorphy of the person markers also follows the general Ewenic patterns. Thus, consonant stems require the presence of the connective vowel -I- before monosegmental suffixes, while nasal stems require the suffix-initial nasal m and obstruent stems the obstruent p instead of w (< *b), as used after vowels. Possession in the Orochen noun phrase has normally both head and dependent marking. The possessor is marked with the possessive (genitive) case in -ƞI and the possessee is marked with a possessive suffix that indexes the person and number of the possessor, e.g. min-ƞi mʊrɪn-ɪw ‘my horse’ (1SG-POSS horse-PX1SG). The reflexive suffixes are used when the possessor is coreferent with the subject of a clause. In principle, the reflexive suffixes should correlate with the reflexive pronoun meen, but there are examples of personal pronouns being used instead, e.g. sebu noon-ƞɪ-n mukiren-mi mʊktʊrga-ca ‘the teacher snapped his own stick’ (teacher 3SG-POSS-PX3SG stick-RX snap-PST). OTHER WORD CLASSES Manner and means are mostly expressed by regular nouns or adjectives marked with the instrumental case. However, many manner adverbs can be derived from adjectives using the suffix -gdi, e.g. ñama ‘warm’ : ñama-gdi ‘warmly’, jutii ‘sweet’ : jutii-gdi ‘sweetly’. Manner adverbs are typically placed directly before the verb. Spatial adverbs are etymologically identifiable as spatial nouns that bear case marking and/or special affixes, e.g. uy/e- ‘upper side, top’ : LOC uy-le ‘above’ : DATPX3SG uye-du-n ‘over’ : DIR uy-kki ‘upwards’ : DER uy-de ‘higher’; ergi- ‘lower side, bottom’ : LOC ergi-le ‘below’ : DIR ergi-kki ‘downwards’ : DER ergi-de ‘lower’; ama/y- ‘back side’ : LOC amay-la ‘behind’ : DIR ama-kki ‘backwards’ : DER amargɪ-da ‘behind’. However, there are enough gaps in the overall set of these words and enough irregularity to suggest that they are frozen forms. Spatial adverbs are generally placed before the verb. They are also used in the role of spatial postpositions after nouns and pronouns marked by the possessive case, in which position they can take possessive suffixes, e.g. kʊtʊn-ƞɪ amay-laa-n ‘behind the town’ (town-POSS back. side-LOC-PX3SG). While the spatials are inflectable nouns, there is also a small set of true postpositions in Orochen. These are synchronically invariable, e.g. jalɪn ‘on behalf of’, ƞacɪn ‘like, similar’ (abstracted from the originally suffixal element SIM *-gAcIn with the same meaning), jirge ‘as, equivalent to’, and japka ‘as far as’. The noun preceding a postposition remains in the basic form, e.g. beye jalɪn ‘on behalf of the man’. Temporal adverbs include a heterogenic collection of forms such as teƞƞe ‘last year’, sodʊyɪ ‘always’, sootɪ ‘often’, telin ‘just now’, and ñoowʊdʊ ‘earlier’. Temporal adverbs usually occur sentence initially, though they are sometimes placed after the subject.

Orochen 197

Degree adverbs include forms such as manɪ ‘very’ and bayla ‘a little bit’. Degree adverbs are placed before the entity that they modify. Polar questions are formed by placing the particle ɪa after a verbal form. Somewhat unusually, this particle can also take person markers to make the question disjunctive, cf. e.g. bii ƞene-m ɪa ‘did I go?’ (1SG go-AOR1SG INTERR) vs. bii ƞene-m ɪa-m ‘did I go or not’ (1SG go-AOR1SG INTERR-AOR.1SG). This is possible because of the dual origin of ɪa(-) both as a pronoun (‘what?’) and as a pro-verb (‘to do what?’), as attested in the other Ewenic varieties. However, the use of this item as a question particle seems to be unique to Orochen. A polar question that anticipates an affirmative response uses the particle antɪ, e.g. noonɪn beyu-ce antɪ ‘he hunted, didn’t he?’ (3SG hunt-PST INTERR). There is also a set of sentence-final evidential particles, including oo [doubt], be [assumption], ke [certainty], je [reluctancy], and eri [surprise], the last one of which is apparently identical with the proximal pronoun eri ‘this’. While these function as independent invariant words, there are also a few delimiting enclitic particles which are placed directly after the constituent that they are delimiting, e.g. =lA ‘only’, =dAt ‘almost’. As clitics, these particles follow the rules of vowel harmony, as determined by the word to which they are attached, e.g. bii awarʊ-m ‘I am starving’ : bii awarʊ-m=dat ‘I am almost starving’ (1SG starve-AOR1SG=PTCL). VERBAL MORPHOLOGY As in other Tungusic languages, the verbal forms in Orochen can be divided into finite aorist, (finite) imperative, non-finite converbal, and polyfunctional (finite and non-finite) participial forms. The actual finite paradigm, which is inflected in person and number, is built from a mixture of finite aorist and finitely used participial forms. The formation of the aorist stem involves lexically and phonologically conditioned allomorphy shared with the other Ewenic languages. Orochen retains the distinction between four types of verbs, which form the aorist stem by the elements (1) -rA-, e.g. eme- ‘to come’ : AOR eme-re-, (2) -A-, e.g. un- ‘to say’ : AOR un-e-, (3) -dA-, e.g. oo- ‘to become’ : AOR oo-da-, or (4) -sI-, e.g. bi- ‘to be’ : AOR bi-si-. Due to a recent dialectal confusion, apparently caused by the phonetic change of intervocalic s to x, the aorist marker in type (2) can also appear as -ki- in the copula-existential bi- ‘to be’ : AOR bi-si- > bi-xi- > bi-ki- and the negation verb e- ‘not to do’ : AOR e-si- > e-xi- > e-ki-. It is unclear whether any other verbs are conjugated according to this pattern. Type (2), which is characteristic of the primary nasal stems in -n, comprises only a few derivational elements. As in the other Ewenic languages, the aorist marker -rA- of type (1), which remains the only synchronically productive type, has several variants depending on the final segment of the verbal stem: -rA- after vowels and the rhotic r, e.g. ir- ‘to cook’ ITR : AOR ir-re-; -lA- after the lateral l, e.g. edel- ‘to cross a river’ : AOR edel-le-; -nA- after the nasal m, e.g. ɪm- ‘to drink’ : AOR ɪm-na-; and -tA- after the obstruents p t k, e.g. jeb- ‘to eat’ : AOR jep-te-. After vowel stems, the aorist marker -rA- is absent in the first and second persons singular, e.g. janda- ‘to sing’ : AOR 1SG janda-m : 2SG janda-nɪ : 3SG janda-ra-n, etc. The aorist stem also functions as the invariant connegative form combined with the fully conjugated forms of the negation verb, e.g. noonɪn e-ki-n janda-ra ‘s/he does not

198 Lindsay J. Whaley

sing’ (3SG NEG-AOR-3SG sing-CONNEG), bii e-ce-w jep-te ‘I did not eat’ (1SG NEGPST-1SG eat-CONNEG). Moreover, in Orochen, as also in Solon, the aorist stem can be used adnominally in a participial function. The categories of voice and Aktionsart are expressed derivationally. The category of voice is represented by the passive in -w-, the reciprocal in -mAt-, the cooperative in -ldI-, and the causative in -wkAAn- ~ -wkAn- ~ -kwAn- (with the form varying by speaker). Since these are derivational categories, their suffixes are followed by the regular finite and non-finite inflectional elements, e.g. janda- ‘to sing’ : janda-ldɪ-ra ‘they sang together’ (singCOOP-AOR3PL). When combined with each other the passive suffix normally precedes the causative suffix in the verbal template. In the causative construction, the causee is usually placed in the accusative case, but, when indefinite, it can also be in the partitive case or remain unmarked. In those instances where the causee can be construed as a recipient, the dative case can also be used, as in [3]: [3] bii min-ƞi edi-du-w jep-kwen-e-m 1sg 1sg-poss husband-dat-px1sg eat-caus-aor-1sg ‘I am making my husband eat.’ ~ ‘I am permitting my husband to eat.’ There is a robust set of derivational Aktionsart markers which clarify the nature of the event being described by the verb: -jA- [progressive], -ksI- [desiderative], -ktA- [distributive], -l- [inceptive], -lIsIn- [inceptive], -mUn- [desiderative], -nA[intentional = andative], -t- [deliberative], -sIn- [inceptive], -tI- [iterative], -mAlcA[momentaneous]. Of note is the fact that there are two suffixes classified here as “desiderative” and three classified as “inceptive”. To date, not enough text data has been collected to establish the differences among these suffixes. Speaker intuitions suggest that -mUn- expresses a higher degree of intensity than -ksI- for the desiderative suffixes. It is possible that the three inceptive suffixes might be usefully re-classified as “inceptive”, “ingressive”, and “inchoative”, respectively, but not enough data has been collected to establish this. On the other hand, if a unified terminology were to be sought with the other Ewenic languages, including, Siberian Ewenki in the first place, the suffix -l- could also be labelled as “inchoative” and the suffix -sIn- as “semelfactive”. Obviously, the functional differences are very minute and may vary from dialect to dialect and from speaker to speaker, and they may also involve lexicalized irregularities. It is possible to use combinations of Aktionsart suffixes, e.g. dooldɪ-mʊ-na-ra-n ‘he goes to spy, he wants to go to hear’ (hear-DESID-INTENT-AOR-3SG), but there are also restrictions on the ordering of the combinations: for example, the previous example is unacceptable with the Aktionsart suffixes reversed (*dooldɪ-na-mʊn-a-n). However, there is not sufficient data to determine what all the restrictions on ordering are, or whether they are predictable on the basis of semantic scope considerations. PARTICIPLES As compared with Siberian Ewenki, the Orochen system of participles, like that of Solon, is somewhat reduced. Most importantly, the Common Tungusic imperfective participle in *-rII does not seem to be attested in active use in Orochen, though it is preserved in lexicalized relics, such as saa-rɪ ‘smart’ (= PTCP.PRS *saa-rïï ‘one who

Orochen 199

knows’, from the verb saa- ‘to know’). Instead, the plain aorist stem in -rA (with variants) is used in the same function, e.g. (example from Hu 2001) kakara tuuree-re erin oo-caa ‘there came the time of the cock crowing’ (cock crow-PTCP.AOR time becomePST). The participial use of the basic aorist stem, which is shared by Solon, but which is not attested in Siberian Ewenki, looks like an archaic feature, since the aorist marker is ultimately also a nominalizing suffix, though it has lost its participial uses in the other Ewenic languages. Apart from the aorist participle, which normally has present tense reference, Orochen has the perfective (past) participle in -cA/A and the habitive participle in -ƞkI/I. In addition, there is another perfective (past) participle in -ptIlA, which, like its counterpart in Neghidal, refers to a remote past and may originally have had an impersonal reference. Dialectally, a frequentative participle in -jAAƞAAcI is also attested. Participles have four primary functions. First, they are used for the purposes of relativization. In relative constructions, the head noun agrees with the participle, using the possessive suffixes, e.g. eri bii je-ƞki-w jepte ‘the food I used to eat’ (this 1SG eat-PTCP. HAB-PX1SG food). No person marking is, however, added to the remote past participle in -ptIlA, perhaps because of its impersonal connotation, e.g. noortɪn ice-ptile juu ‘the house that they had (already) seen’ (3PL see-PTCL.REM house). Relative clauses use a gapping strategy, that is, the anaphor of the relativized noun is not indicated in the relative clause. There are no relative pronouns used. Second, participles can be used as finite predicates (predicative nominals), in which use they take person markers of the possessive type in the first and second persons, and regular nominal number marking in the third person (Table 8.6). This is the most common use of participles, and, in particular, the perfective participle in -cA/A functions as the default past tense form of the finite paradigm, e.g. bii sin-ji ƞene-ldi-ce-w ‘I went together with you’ (1SG 2SG-INSTR go-COOP-PST-1SG). By contrast, the finite indicative present tense in Orochen is normally expressed by the aorist. However, the aorist is also used to express the past tense in texts for backgrounded clauses, where the paste tense marker -cA/A is used on foregrounded clauses. The aorist can also be used for near future reference if the context makes this clear, for instance, by use of a temporal adverb. Third, participles are used in conjunction with the copulas bi- ‘to be’ and oo- ‘to become’ in the formation of analytic tenses. In the analytic tenses, a past tense form of one of these two verbs is placed after the participle with no subject agreement. The participle indexes agreement with the subject. An aorist participle with oo-ca (become-PST) indicates present habitual; a habitual participle with bi-ce (be-PST) indicates present habitual; and a perfective participle with bi-ce indicates pluperfect, as in [4]: [4] bii un-e-m tarɪ-l beye 1sg say-aor-1sg that-pl man ‘I said that those men had eaten.’

jep-ce-l eat-ptcp.prf-pl

bi-ce be-pst

In view of the above-mentioned roles of participles, it has to be added that Orochen has, at least historically, one other participial marker, namely, that of the futuritive (future) participle in -jA/A(-), a cognate of the identical marker in -jAA (< *-jA-ƞAA) in Siberian Ewenki. It seems that this element survives in Orochen only in the role of a future tense marker -jA/A- (< *-jAA-) of the finite conjugation, e.g. sɪlba- ‘to tell’ : sɪlba-ja-w ‘I will tell’ (tell-FUT-1SG). The fact that the future tense is originally based on a participle is

200 Lindsay J. Whaley

confirmed by its personal paradigm, which employs the possessive type of person markers. Therefore, the future tense forms remain clearly distinct from the progressive forms in -jA- (< *-jA-), which are conjugated according to the aorist pattern. It is possible that the above-mentioned frequentative participle in -jAA-ƞAA-cI is also connected with the future marker. IMPERATIVES Orochen retains, with some modifications, the Common Ewenic paradigm of imperative forms for all persons. This basic imperative set may also be identified as “proximal”, in that it is used to issue commands and wishes that are expected to be carried out immediately or in the near future. A corresponding “distal” set, which lacks a first person singular form, is built on the element -DA- (the purposive converb marker, also known as the “supine”) and is used to issue commands and wishes that will be carried out after another activity is completed, or in the indefinite future (Table 8.7). TABLE 8.7 OROCHEN IMPERATIVE FORMS        SG

PL

1

 

PROX 

C

IMP

-I-

-ktA

DIST  PX/VX

SUP

PX

n/a

n/a

2

-kAl

-DA-

-y

3

-yi-

-n

-DA-

-n

EXCL

-yi-

-wUn

-DA-

-wUn

INCL

1

-GA-

-l/e

-DA-

-(tI)r

2

-GAl-

-U

-DA-

-sUn

3

-yi-

-tIn

-DA-

-tIn

Stem types: C = consonant stems (with a connective vowel). Segmental alternations: A = a e o ɵ, U = ʊ u, I = ɪ i, G = g k ƞ.

Both sets take person marking of basically the possessive type, but the “proximal” set has several idiosyncracies, which have close parallels in Siberian Ewenki. Thus, in the third person singular and plural, as well as in the first person plural exclusive form, the imperative is marked by the element -yI- (< *-gI-), while the first person plural inclusive form is marked by the suffix -GA-l, in which the final -l ~ -lA represents the plural marker of the nominal declension (cf. Siberian Ewenki IMP.1PL.INCL -gAA-r). The second person plural form in -GA-lU = -GAl-U is an irregular development of the earlier form with a plural person marker of the predicative type (< *-kAl-sU), with the initial consonant replaced in analogy of the first person plural inclusive form (*k > g). There is dialectal and individual variation in the forms of the “proximal” suffixes. For the second person the suffixes SG -jI-y : PL -jI-sUn, for the first person plural exclusive the suffix -ktA-wUn, and for the first person plural inclusive the suffix -ƞnA-r, have also been recorded (Hu 1986, 2001). Functionally, only second person imperatives (imperatives proper) have imperative force, while first person imperatives (voluntatives or optatives) have a hortatory reference, e.g. miti-we budi-yi-wun ‘let us undress!’ (1PL. INCL-ACC undress-IMP.PROX-1PL.

Orochen 201

INCL). Third person imperatives (permissives) have the force of a wish for an outcome that is only indirectly in control of the speaker, e.g. noonɪn tɪmana beyu-de-n ‘let him hunt tomorrow!’ (3SG tomorrow hunt-IMP.DIST-3SG). As can be seen, the agent of the verb is placed in the accusative case for the “proximal” imperative. This is true for the first and third persons. The addressee of second person imperatives is rarely expressed, but if it is, it is in the nominative. For all “distal” commands, the agent of the verb is in the nominative. OTHER MODAL FORMS Apart from the imperative there are several other modal contents that are expressed with a variety of grammatical and lexical means. These comprise the subjunctive mood, as well as constructions indicating obligation, necessity, possibility, and ability. •

The subjunctive (conditional) mood is formed with the suffix -sI-, which, when no other tense marking is present, is placed on the verb in the protasis between the aorist tense marker and a person marker of the possessive type, as in [5]:

[5] tari-l ɪna-ra-sɪ-tɪn buu 1pl.excl that-pl arrive-aor-subj-3pl ‘If those people arrive, we will eat together.’

emun-du one-dat

jep-te-wun eat-aor-1pl.excl

The past conditional is formed in the same way, except that the protasis verb uses the past tense in -cA/A-. It should be noted that in the northeastern dialect, there is a distinct suffix for the past conditional, -pkI-. The future conditional is also formed in the same way, except that the protasis verb uses the future tense in -jA/A-, or the aorist if the context makes it clear that the speaker is uttering a future condition. In all these constructions, the element -sI-, dialectally also -sI-k-, resembles functionally a case suffix. •

•

•

Obligation is expressed by putting the enclitic particle =mAl on the verb, e.g. bii aa-ja-w=mal ‘I should stop’ (1SG stop-FUT-1SG=MOD). Necessity is indicated by a figura etymologica in which the first instance of the verb is in a participial dative case form, e.g. bii aa-ra-dʊ aa-m ‘I must stop’ (1SG stop-PTCP.AOR-DAT stop-AOR.1SG). Possibility is expressed by placing the word bijen ‘might, possibly’, originally the third person singular future tense form of the verb bi- ‘to be’, after the verb, e.g. noonɪn tɪmana eme-re-n bijen ‘he might come tomorrow’ (3SG tomorrow comeAOR-3SG might). Ability is expressed by the verb ete- ‘to be able’ in combination with the connective converb, e.g. umun iniyi eme-rgi-mi ete-ni ‘one day you will return’ (one day comeREV-CV.CONN be.able-AOR2SG).

CONVERBS Orochen has a dozen productive converb markers (Table 8.8), which may be divided into the conjunct (same-subject), disjunct (different-subject), and ambivalent types. Most of the conjunct converbs are formally invariant, while the disjunct and ambivalent converbs can take person agreement suffixes to indicate the subject.

202 Lindsay J. Whaley TABLE 8.8 OROCHEN CONVERB MARKERS  

 

PX/RX

SS

CONN

-mI

+

ANT

-ksA

+

SUCC

-mnA

+ +

SIM

-nAkAn

DUR

-GAy

IMMED

-JIkIn

DS

+ +

+

+

POST

-mdAk

+

TERM

-mdUr

+

LIM

-ƞnA/n

+

+

+

PURP

-DA-

+

+

+

AOR

-RAk-

+

+

ANT2

-cAlA

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e o ɵ, U = ʊ u, I = ɪ i, D = d t, G = g k, J = j c, RA = the aorist marker depending on the conjugational class.

Many of the converbal suffixes have formal and functional counterparts in the other Ewenic languages and may be named accordingly, though their functions show differences from language to language. Thus, Orochen retains the general connective converb in -mI (default for conjunct use), the two roughly synonymous anterior and successive converbs in -ksA and -mnA (‘after’), the simultaneous converb in -nAkAn (‘while’), the durative converb in -Gay (‘while’), the immediative converb in -JikIn, the purposive converb (“supine”) in -dA/A (‘in order to’), and the aorist converb in -RAk- (default subordination for disjunct use). In addition, there are the posterior converb in -mdAk (conjunct), and the roughly synonymous terminative and limitative (‘until’) converbs in -mdUr (conjunct) and -ƞnA/n (ambivalent), respectively. The quasiconverbal form in PTCP.PRF-LOC -cA-lA may also be viewed as a grammaticalized entity in the function of a second anterior converb. The connective converb in -mI functions as an “infinitive”, e.g. mergen buyu-mi ete-ce ‘the hunter was able to hunt’ (hunter hunt-CV.CONN be.able-PST). However, it can also be used when the context makes the temporal relation between the converb and the main verb clear. When a temporal relationship has to be specified, other forms, such as the anterior converb in -ksA or the posterior converb in -mdAk, are used, e.g. juu-le-wi ii-kse too nʊlɪ-ca ‘after she went into her room, she lit a fire’ (house-LOC-RX enter-CV.ANT fire light-PST), tarɪ muuliƞki-we joo-cɪ-mdak ƞene-re-n ‘he grabbed the water bucket and went’ (water.bucket-ACC grab-DUR-CV.POST go-AOR-3SG). In disjunct constructions, the different subject is indicated by possessive suffixes, as in [6], while reflexive suffixes are used if the subject is the same as that of the main verb, though some speakers leave the converb uninflected in the latter case. [6] bii teeti-we arʊn on-ƞna-n 1sg clothes-acc clean become-cv.lim ‘I washed the clothes until they were clean.’

sɪlkɪ-ca-w wash-pst-1sg

Orochen 203

Apart from the quasiconverbal form in -cAlA- = -cA-lA- (< *-cAA-lAA) the local case forms of other participles, including the aorist participle in -rA-, can also be used in converbal functions, in which case they take personal or reflexive marking, as in [7]: [7] bii jepte-ye oolo-ro-du-wi 1sg food-part cook-ptcp.aor-dat-rx ‘While I was cooking, I burned my hand.’

ƞaala-wɪ hand-rx

dalga-ca-w burn-pst-1sg

SYNTAX Like other Tungusic languages, Orochen is an SOV language. Compared with Siberian Ewenki, its basic constituent order is even more rigid. An exception from the canonic word order is formed by complement clauses, which follow the verb, e.g. bii saa-m noonɪn caƞa-ra-n ‘I know that he is singing’. It may be noted that no complementizer is used, but the complement clause remains finite. The typical word order beyond the basic constituents is: temporal/spatial—subject— direct object—indirect object/obliques—manner—adverbial—verb. Typically, Orochen employs parataxis rather than using complex sentences, and adverbial subordination is usually accomplished through the use of participles and converbs. However, there are a handful of subordinating conjunctions such as tuksen, which introduces a result clause, and jalɪn, which introduces a concession clause, as in [8]. The subordinating conjunctions occur at the end of the clause. [8] tarɪ ƞanakɪ-mña kʊmaka-wa ñamña-ca that dog-augm deer-acc chase-pst ‘Although that nasty dog chased the deer,

jalɪn although

e-ce jo-ro neg-pst catch-conneg it did not catch it.’ When participles and converbs are used for adverbial subordination, typically only one occurs with the main verb, rather than strings of non-finite forms. In addition to finite verbs, adjectives and nouns can serve as the main predicate. In this role, they are optionally accompanied by the copular verb bi- ‘to be’. Parataxis is used in place of coordinating clauses. Phrase-level coordination employs the clitic =dA. The clitic can be placed on all conjuncts or only on the final conjunct, e.g. noonɪn gʊgda(=da) aya-makan(=da) saarɪ=da ‘she is tall, beautiful and smart’ (3SG tall=EMPH good-AUGM=EMPH smart=EMPH). Clause level disjunction uses oomal . . . oomal ‘either or’. The conjuncts are placed after the predicate. Information that is recoverable from the first clause is not repeated in the second, as in [9]: [9] sii tɪmana eme-rgi-re-y ɪa 2sg tomorrow come-rev-aor-2sg interr ‘Are your returning tomorrow or the next day?’

oomal disj

timinin next.day

oomal disj

In noun phrases, modifiers occur before the head nouns. Though there is variability permitted among modifiers, the preferred order of modifiers is: quantifier—

204 Lindsay J. Whaley

demonstrative—possessor—number—adjective—noun. Relative clauses are placed before the noun. They use a gapping strategy for the relativized noun. Subject, object and oblique relatives are all permitted. The verb of a relative clause is in a participial form, as in [10]: [10] eri buu uge-ƞki-wun ride-ptcp.hab-px1pl.excl this 1pl.excl ‘The horse we used to ride was beautiful.’

mʊrɪn aya-makan horse good-augm

bi-ce be-pst

LEXICON AND LANGUAGE CONTACTS Orochen remains poorly integrated in comparative Tungusic studies, and this concerns also its lexical resources. There exists no dictionary of any form of Orochen, not to mention a comprehensive lexical survey of all dialects. The linguistically unsophisticated materials recorded by S. M. Shirokogoroff (Doerfer & Knüppel 2004) can only to an insufficient extent amend the situation. There is a danger that the language, which is already seriously endangered, will become extinct before its lexical resources are fully documented. The last stages of documentation are also facing the problem of declining language skills among the remaining speakers and semi-speakers. This is particularly obvious when it comes to the phonological shape of lexical items, which already in the extant materials shows uncontrolled variation for many details, such as, for instance, vowel length. What can be said on the basis of today’s knowledge is that Orochen is lexically very close to Siberian Ewenki. It shares some innovations, such as the specific shape of the item *ïm- ‘to drink’ (< *umï-), with Solon and the Borzya dialect of Khamnigan Ewenki, the most distinct representatives of the Manchurian Ewenki branch, and it also has some Mongolic loanwords shared with Solon, notably the numeral 90 yereen, possibly borrowed from Daghur via Solon, rather than directly from Mongolian (proper). On the other hand, it also shares diagnostic items with Siberian Ewenki, such as dɪlaca ‘sun’ vs. Solon sigun. Phonologically Orochen is considerably more conservative than Solon, preserving, for instance, the velar nasal in initial position, as in ƞaala ‘hand’ vs. Solon naal/a. The conclusion is probably correct that the Orochen dialects have developed from an early form of Manchurian Ewenki with a heavy secondary influence from Siberian Ewenki. Unlike the Solon, who for generations have been bilingual in Daghur and, later, in Mongolian (proper), and many groups of Siberian Ewenki with an equally long tradition of bilingualism in Yakut or Buryat, the Orochen have no similar history of bilingualism in any regional language. The knowledge of Manchu was apparently never widespread among the Orochen, and the colonial impact of Russian and, in the Manchukuo period (1931–1945), Japanese also remained insignificant, though some items, such as cay ‘tea’, may have reached the Orochen via multiple ways: from Manchu, Mongolian, Daghur, or Russian. Orochen has also several relatively old borrowings from Chinese, e.g. poo ‘rifle’ (Chinese 砲 pao), probably transmitted by Mongolic. Intensive influence from Chinese began only in the 1950s, and it soon reached a massive scale, with unlimited borrowing and code switching. The ongoing language shift has apparently been facilitated by the lack of intermediate languages and a corresponding tradition of multilingualism, which in the case of Solon have contributed to the better resistance of the ethnic language.

Orochen 205

REFERENCES AND FURTHER READING Doerfer, Gerhard (1983) ‘Das Birare’, Journal de la Société Finno-Ougrienne 78: 1–19, Helsinki. Doerfer, Gerhard & Michael Knüppel (2004) Etymologisch-ethnologisches Wörterbuch tungusischer Dialekte (vornehmlich der Mandschurei), Hildesheim: G. Olms. Hu, Zengyi 胡增益 (1986)『鄂倫春語簡志』[A sketch of the Orochen language],『中 國少數民族語言簡志叢書』[Languages of the minority nationalities of China], 北京 [Peking]: 民族出版社. Hu, Zengyi 胡增益 (2001)『鄂倫春語研究』[Study on the Orochen language], 北京 [Peking]: 民族出版社. Janhunen, Juha & Xu Jingxue & Hou Yucheng (1989) ‘The Orochen in Xinsheng’, Journal de la Société Finno-Ougrienne 82: 145–169, Helsinki. Li, Bing (1996) Tungusic Vowel Harmony: Description and Analysis, Academisch Proefschrift, Universiteit van Amsterdam, Holland Institute of Generative Linguistics (HIL) Dissertations 18, Amsterdam. Lulich, Steven & Lindsay J. Whaley (2012) ‘An acoustic phonetic study of Oroqen vowels’, in: in: Malchukov & Whaley (eds.) (2012), 61–80. Malchukov, Andrej L. & Lindsay J. Whaley (eds.) (2012) Recent Advances in Tungusic Linguistics, Turcologica 89, Wiesbaden: Harrassowitz Verlag. Missonova, Lyudmila (2020) ‘Habet sua fata manuscripta’, in: Anna Król & Wioletta Laskowska-Smoczyńska (eds.), Bronisław Piłsudski w stulecie śmierci: W stronę niepodległej: materiały pokonferencyjne/Bronisław Piłsudski on the centennial of his death: Towards an independent homeland: proceedings of the conference, Kraków: Manggha Museum of Japanese Art and Technology, 194–207. Sa Xirong 薩希榮 (1981)『簡明漢語鄂倫春語對照讀本』[Comparative textbook of Chinese and Orochen], 北京 [Peking]: 民族出版社. Whaley, Lindsay J. & Lenore A. Grenoble & Fengxiang Li (1999) ‘Revisiting Tungusic classification from the bottom up: A comparison of Evenki and Oroqen’, Language 75 (2): 286–321. Whaley, Lindsay J. & Fengxiang Li (1998). ‘The suffix -kan in Oroqen’, Studies in Language 22 (2): 447–471). Whaley, Lindsay J. & Fengxiang Li (2000a) ‘Oroqen dialects’, Central Asiatic Journal 44 (1): 105–130. Whaley, Lindsay J. & Fengxiang Li (2000b) ‘Emphatic reduplication in Oroqen and its Altaic context’, Linguistics 38 (2): 355–372. Whaley, Lindsay J. & Fengxiang Li (2012) ‘The grammaticalization cycle of causatives in Oroqen dialects’, in: Malchukov & Whaley (eds.) (2012), 167–180. Zhang, Yanchang [張彥昌] & Li Bing [李兵] & Zhang Xi [張晰] (1989), The Oroqen Language [鄂倫春语], 長春 [Changchun]:吉林大學出版社 [Jilin University Press].

CHAPTER 9

SOLON Sangyub Baek

Solon is the Ewenic language of the Solon (Suolun Ewenke 索倫鄂溫克), mostly residing in Hulun Buir City Prefecture, Inner Mongolia, as well as in the Nonni (Nenjiang) basin in Heilongjiang Province, China. Officially, the Solon are included under the cover term “Ewenke” (鄂溫克), which also comprises the “Yakut” (Yakute Ewenke 雅庫特鄂溫克), whose ethnic language is Siberian Ewenki, and the “Tungus Ewenke” (Tonggusi Ewenke 通古斯鄂溫克), who are bilingual speakers of Khamnigan Mongol and two varieties of Khamnigan Ewenki. The ethnonym “Ewenke” was adopted in 1957, when all these groups were artificially united to form the “Ewenke” nationality (Ewenke zu 鄂溫克族), as opposed to the Orochen (Elunchun zu 鄂倫春族). Nonetheless, it has to be stressed that the Ewenic languages spoken by the three “Ewenke” groups are very distinct. Among them, Solon, like Orochen, is a language spoken only inside the current borders of PR China. According to the Chinese national census of 1982 and 1990, the population of the “Ewenke” nationality was some 19,000 and 26,000 individuals, respectively. A more recent figure is c. 31,000 (2010). Since the “Yakut” and the Khamnigan make together hardly more than 3,000 individuals, the number of ethnic Solon must today be close to 30,000. Of them, c. 12,000 (2020), live in the Ewenke Autonomous Banner (Ewenke zu zizhiqi 鄂溫克族自治旗) in the southern part of Hailar District, Hulun Buir, where the Solon form close to 10 per cent of the local population, while the rest are mainly dispersed elsewhere in Hulun Buir. The Solon in Heilongjiang have been reported to number c. 2,700 (2000). Native language maintenance has been traditionally good among the Solon, and, although the current situation is unknown, the number of fluent speakers today must still be well over 10,000 individuals, including children. This makes Solon the most viable Tungusic language today, and the only one that still has significant generational transmission. There is also a diaspora group of a few hundred individuals, known as the Ongkor Solon, in Jungaria, northern Sinkiang (Xinjiang), where, however, the last native speaker died in the 1990s. The geographical dispersal of the Solon corresponds to a difference between the three main dialect groups of the language, which may be termed Nonni Solon, Hailar Solon, and Ongkor Solon. The Solon refer to themselves by the ethnonym eweƞki (~ eweƞxi) or also eweƞki aɪman ‘Ewenki people’. The exonym “Solon” (Mongolian soloon, Chinese suolun 索倫) is, according to a generally accepted explanation, based on Ewenki sologon ‘upstream inhabitant’, from solo- ‘to go upstream’. This exonym probably dates back to the time when the Solon were formed as a distinct group of Ewenic speakers. However, the same appellation has also been applied to the Mongolic-speaking Daghur, with whom the Solon have historically lived in a symbiotic relationship. Culturally, the Solon have for some centuries been characterized by a lifestyle similar to that of the Daghur and Manchu, with DOI: 10.4324/9781315728391-9

Solon 207

hunting, fishing, and small-scale agriculture, complemented by elements of Mongolian type pastoral nomadism. Solon remains a virtually unwritten language. This is partly because the Solon have a long history of multilingualism, not only in Daghur, but also in Manchu and Mongolian, as well as, later, in Russian, Japanese, and Chinese. For this reason, the Solon speakers have used other languages, especially Manchu and Mongolian, but today increasingly often Chinese, for all written communication, without losing their own spoken language. However, occasional attempts have been made in the past to write Solon in the Manchu and Mongol scripts. DATA AND SOURCES Research on Solon was initiated by Russian scholars, joined by Chinese and Japanese since the 1980s. The first to document the language was A. O. Ivanovskii (1894), who recorded a word list and 25 sample sentences from the Nonni dialect. A more comprehensive documentation of the phonology and morphology of the Hailar dialect, comprising also a lexicon of c. 2,000 words and a collection of texts, was published by Nicholas Poppe (1931). Based on the materials of Ivanovskii and Poppe, a grammatical sketch of Solon was authored by V. I. Cincius (1997), while Viktor Aktnin (1986) made a specific study of the semantics of some converbs in Solon. The Ongkor Solon dialect was also documented already in the early 20th century by G. J. Ramstedt and F. V. Muromski, but their collections were published only decades later by Pentti Aalto (1976–1977) and Stanisław Kałużyński (1971). The linguistic extinction of Ongkor Solon was reported in Bai Lan & Janhunen (1992). The first Chinese description of Solon was by Hu Zengyi (1984), which served as a source for the introduction to Tungusic languages in Korean by Kim Juwon (1986). There soon followed the more comprehensive grammar by Hu & Chaoke (1986). Chaoke (Chog), a native Solon linguist, has subsequently been particularly active in publishing, in both Chinese and Japanese, reference works and research papers, including two grammatical descriptions (1995, 2009), an extensive monograph on morphophonemics and nominal morphology (2003), two other morphological monographs (2017ab), and a semantically arranged dictionary of some 4,300 items (1991), supplemented by an index compiled by Tsumagari Toshirō (1993). He was also the principal editor of a collection of 42 folklore texts in IPA transcription and a translation into Mongolian (Chaoke & al. 1988). Another native Solon scholar is Dao’erji (Dorji), who is the author of a Solon-Chinese dictionary (1998) with more than 17,000 words, as well as of a Solon-Mongolian dictionary (2014) with more than 25,000 words. Japanese work on Solon has been accelerated by the opportunities to conduct fieldwork in China since the 1980s. As a matter of fact, the first Japanese publication on Solon dates back to Kamimakise (1940), in which a short grammar and a comparative list of words in Japanese, Solon, and Mongolian are included. Lie (1978) is a further study based on these materials. The first modern Japanese work on Solon is a short overview by Tsumagari (1989) based on secondary sources. From the 1990s Japanese scholars have collaborated with native Solon scholars and consultants. This collaboration has resulted in a collection of Solon basic sentences by Chaoke & Tsumagari & Kazama (1991), a basic vocabulary of Solon and other “Ewenke” idioms spoken in China by Chaoke & Tsumagari (1995), and a practical textbook of Solon in Japanese by Chaoke & Nakajima (2005). Another basic vocabulary including more than 2,000 words with abundant examples was published by Kazama & Tuoya (2011), who also published two collections of folkloric texts (2007, 2008).

208 Sangyub Baek

On the basis of his field materials Kazama has also published a sketch grammar and collections of texts with ethnological notes (2005, 2018ab, 2020, 2022), as well as several grammatical studies dealing with aspect, voice, and modality (2013), transitivity (2014), complex sentences (2015), and information structure (2016). He has also studied the Mongolian influence on Solon (2010). However, the only reliable description of Solon in English remains the compact grammar by Tsumagari (2009), which also contains diachronic and comparative information. Most recently, Baek Sangyub has used Solon data in a number of typologically oriented papers dealing with correlatives (2012ab), converbs (2014), conditionals (2015), person markers (2016), and auxiliaries (2017) in the Tungusic languages. Solon, like Orochen, remains relatively poorly integrated into comparative Tungusic studies. Apart from the general sources on the Tungusic languages, very little specific work has been done on the dialectology and diachrony of Solon. In particular, until the present day the Nonni dialect remains little investigated. A specific topic that would require more research is the lexical and structural interference of Solon with Daghur and Manchu. An initial step in this field has been made by Bayarma Khabtagaeva (2012). The present chapter is focused on the Hailar dialect. The sources of the sentence examples are marked as follows: C = Chaoke (1995), D = Daorji (1998), HC = Hu & Chaoke (1986), K = Kazama (2005, 2013, 2014), KT = Kazama & Tuoya (2007, 2008, 2011), P = Poppe (1931), and T = Tsumagari (2009). SEGMENTAL STRUCTURE There have been varying opinions in the past concerning the size and structure of the Solon vowel system. In the present description it is assumed that the system comprises 9 short and 9 long vowels, organized phonetically into five levels opening (Table 9.1). Basically, this is a verticalized version of the Proto-Tungusic system, similar to that attested in Orochen and Ewen. Thus, there are four rounded vowel qualities, comprising two pairs, of which the higher members u [u] and ɵ [o] represent the original front vowels *ü and *ö, respectively, while the lower members ʊ [ʊ] and o [ɔ] represent the original back vowels *u and *o, respectively. Synchronically, the opposition between u ɵ and ʊ o involves tongue height, but it may also be described in terms of tongue root position, with u ɵ being pronounced with advanced tongue root (ATR) and ʊ o with retracted tongue root (RTR). A similar distinction exists in the unrounded vowels i e vs. ɪ a, from original *i *e vs. *ï *a, though in these cases, the phonetic differences involve more clearly only tongue height, with the vowel ɪ [ɪ] being pronounced significantly lower than i [i], while the vowel e [ə] is a centralized vowel higher than a [a ~ ɑ]. In addition, there is the unrounded front quality [e ~ ɛ], here written ie, which originally occurs only as a long vowel (from the diphthongoid sequence *ïa), but there are reasons to assume that it has synchronically developed a short counterpart, here written iĕ. Thus, Solon may be described as having an equal number of short and long vowels. TABLE 9.1 SOLON VOWELS u uu

i ii

ʊ ʊʊ

ɪ ɪɪ ɵ ɵɵ

e ee

o oo

iĕ ie a aa

Solon 209

Vowel length in Solon is, in principle, distinctive in all syllables, and it seems to be somewhat more stable than in the other Ewenic languages, especially as compared with Orochen and Neghidal. The quantitative correlation may be illustrated by minimal pairs such as axɪn ‘elder brother’ vs. aaxɪn ‘liver’, bu- ‘to die’ vs. buu- ‘to give’, orʊ- ‘to gather’ vs. orʊʊ- ‘to pull out’, sele ‘iron’ vs. selee ‘rest’. The Solon consonant system follows the typical Tungusic pattern with 17 segments, organized in terms of four places and seven distinctive manners of articulation (Table 9.2). Unlike most other Ewenic idioms, Solon lacks a distinctive palatal nasal. TABLE 9.2 SOLON CONSONANTS m

n

 

ƞ

b

d

j

g

p

t

c

k

s w

x y

l r

The opposition between the strong and weak stops p t c k vs. b d j g involves voice (in the weak set), but it can also involve aspiration (in the strong set). The palatal stops c j are pronounced as affricates of the palato-alveolar or alveolo-palatal types [tʃ dʒ] ~ [tɕ dʑ]. The velar obstruents k g x can have front and back velar allophones depending on the vowel environment. Also, the weak velar stop tends to be realized as the fricative [ɣ ~ ʁ] in the intervocalic and word-final positions, e.g. bʊgʊ [bʊʁʊ] ‘deer’, mugun [muɣun] ‘silver’, tiinug [ti:nuɣ] ‘yesterday’. The sibilant s [s] has the “hushing” allophone [ʃ ~ ɕ] when preceding the palatal vowels i ii ɪ ɪɪ iĕ ie, cf. e.g. 2SG sii [ʃi: ~ ɕi:] vs. 2PL suu [su:]. The unmarked nasal n in final position, can merge with the preceding vowel to yield a nasalized vowel, i.e. Vn > [Ṽ], e.g. ɪlan [ɪlã] ‘three’, mʊrɪn [mʊrĩ] ‘horse’, sien [ʃeẽ ~ ɕeẽ] ‘ear’. Diachronically, the consonant system shows several features distinguishing Solon from the other Ewenic languages. Some of these features involve developments shared with Daghur. Thus, the original palatal stop (affricate) c is normally represented as s in the word-initial and intervocalic positions, e.g. saagʊ ‘opposite’ < *caaguu > Siberian Ewenki caagu/u ‘next’, ise- ‘to see’ < *ice- > Siberian Ewenki ice- id., also in Mongolic loanwords, e.g. sagaan ‘white (of cattle)’ (← Mongolic cagaan ‘white’). There is, however, some variation in word-initial position, e.g. saalban ~ caalban ‘birchtree’ < *caalban > Siberian Ewenki caalban id., and c is also attested in loanwords, e.g. cilin [mythical unicorn] (← Chinese qilin 麒麟), in consonant clusters, and in the geminate cc, e.g. etcil = eccil ‘these’. A similar loss of stop articulation has taken place in *k, represented as x [x ~ h], e.g. xaltax/a ‘one of a pair’ < *kaltaka > Siberian Ewenki kaltaka ‘one half’, though there is some variation in original front-vocalic words, e.g. xeexe ~ keeke ‘cat’ < *keeke > Orochen keeke, and k is attested in loanwords, e.g. kumun ‘music’ (← Manchu kumun id.). The original *x (< *p) is lost in Solon, as also in Orochen, e.g. ʊnaajɪ ‘daughter’ < *xunaajï > Khamnigan Ewenki (Borzya) xunaaji.

210 Sangyub Baek

Another feature, shared with Khamnigan Ewenki, is the elimination of the velar nasal in word-initial position, where *ƞ > n, e.g. naal/a ‘hand’ < *ƞaala > Siberian Ewenki ƞaala. However, in Solon, the velar nasal has also tended to be lost in intervocalic position, where *ƞ > g, e.g. ineg [inəɣ] ‘day’ < *ineƞi > Siberian Ewenki ineƞi. A distinctive ƞ is nevertheless retained in nasal environments, e.g. niƞun ‘six’ < *ñuƞun > Siberian Ewenki ñuƞun, tʊƞa ‘five’ < *tunƞa > Siberian Ewenki tunƞa. In recent loanwords, mainly from Chinese, several marginal consonant phonemes can also occur. However, in nativized loanwords the foreign sounds have been replaced by native phonemes. For instance, Chinese f is represented as p, as in pen ‘flour’ (← Chinese fen 粉), peƞxan ‘phoenix’ (← Chinese fenghuang 鳳凰). PHONOTACTICS AND MORPHOPHONOLOGY Solon remains true to the suffixing agglutinative type. Hence, each derivational and inflectional morpheme is added to nominal or verbal stems one by one in a fixed order, and the boundary of the stems and suffixes tends to be clear, as in eggeel ‘ox’ : PL-ACC-PX3 eggeel-sel-we-n ‘his/her/their oxen’, naƞɪɪ- ‘to open’ : PASS-PROGRAOR-3SG naƞɪɪ-wʊ-jɪ-ra-n ‘(it) is being opened’. There are, however, two phonotactic constraints that increase the internal coherence of inflected words: vowel harmony and consonant assimilations. Vowel harmony (Table 9.3) in Solon is of the vertical type and can also be described in terms of tongue root position (±ATR), with the higher qualities u ɵ e i being pronounced with advanced (+ATR) and the lower qualities ʊ o a ɪ with retracted (−ATR = RTR) tongue root. The vowel ie and its short counterpart iĕ are harmonically ambivalent. Vowel harmony applies progressively both in word roots and in complex morpheme strings, meaning that the harmonic status of a word is determined by the vowel of the initial syllable. Exceptions from vowel harmony can, however, occur in loanwords. Another phenomenon that potentially reduces the role of vowel harmony in independent word roots is the historically very recent tendency of reduction and loss of vowels in non-initial syllables. This tendency is apparently stimulated by the influence of Mongolic languages, and it affects all vowel qualities, e.g. naal/a ‘hand’ = naala ~ naal : naal-, ug/e ‘word’ = uge ~ ug : ug-, dɪl/ɪ ‘head’ = dɪlɪ ~ dɪl : dɪl-. The loss of vowels has obvious implications on consonant phonotactics, and it can ultimately also affect the duration of the long vowels in non-initial syllables. TABLE 9.3 SOLON VOWEL HARMONY +ATR

u uu

ɵ ɵɵ

e ee

i ii

 

−ATR

ʊ ʊʊ

o oo

a aa

ɪ ɪɪ

iĕ ie

In addition to the vertical vowel harmony, Solon shows occasional signs of labial harmony, in that unrounded vowels can become rounded when following a syllable with rounded vowels. The phenomenon normally concerns the low vowels a aa e ee, which become o oo ɵ ɵɵ after a syllable also containing o oo ɵ ɵɵ, as in the accusative marker -wAA-, e.g. ACC-PX3 xoƞko-woo-nɪ ‘his/her/their bell’, but there are also examples of the high vowels i ii ɪ ɪɪ being represented as u uu ʊ ʊʊ after a syllable containing u uu ʊ ʊʊ, as in the derivational suffix -sIn, e.g. xʊnɪ-sʊn ‘shepherd’ (note that in this example there

Solon 211

is also an intermediate syllable with the unrounded vowel ɪ). The functioning of the labial harmony is, however, unpredictable, and the data contain sporadic variation, as in the past tense (perfective participle) marker -sAA, e.g. PST oo-saa ~ oo-soo ‘(it) became’. In this situation, the status of labial harmony in Solon remains unclarified for the time being. Consonant assimilation at suffixal boundaries is of the regressive type and is in Solon of a much wider scope than in the other Ewenic languages, affecting both nasals and obstruents. A preconsonantal nasal takes the position of the following consonant, e.g. mʊrɪn ‘horse’ : ACC mʊrɪm-ba, though exceptions are encountered in loanwords, e.g. waƞ ‘king’ : DAT waƞ-dʊ (← Chinese wang 王). Stem-final non-nasal consonants also undergo assimilation, e.g. juug ‘house’ : GEN juun-ni : ACC juub-be : DAT juud-du, jip- (~ jep-) ‘to eat’ : AOR jit-te- : PST jic-cee (or jit-cee, depending on the phonemic analysis). These assimilations reflect the historical tendency to homogenize all consonant clusters containing an archiphonemic stop (B D G) or a liquid (l r) as the first component and a stop or a nasal as the second component not only at suffixal boundaries, but also within word roots and suffixal morphemes, e.g. saddɪɪ ‘old’ < *saGdïï > Siberian Ewenki sagdii id., ɪƞa-tta ‘hair’ < *ïnƞa-Gta > Siberian Ewenki inƞa-kta id., ugge-ddi ‘heavy’ < *örge-Gdi > Siberian Ewenki urge-gdi id., ukke ‘door’ < *örke > Siberian Ewenki urke id. As a result, Solon has a complete set of intervocalic geminate stops (pp bb tt dd cc jj kk gg), which, moreover, occur very frequently in speech. The gemination trend in Solon is, however, historically very recent, and can still be dialectally absent, and there seems to be variation also in individual lexical items, e.g. mergen ~ meggen ‘skilful’. Moreover, there are indications that, at least in some items, the syllable-final stops preceding another stop have participated in the Daghur rhotacism (*b *d *g > r), e.g. saccɪ- ~ sarcɪ- ‘to split’ < *caBcï- > Orochen capcɪ- (← Mongolic *cabci-), sarmɪtta ~ sarmɪlta ‘eyebrows’ < *sarmɪrta < *sarmï-Gta > Siberian Ewenki sarmi-kta. This may mean either that Solon at some stage in the past underwent a similar rhotacism, or that the words suggesting a rhotacist development passed via Daghur before (re)entering Solon. As far as prosodic features are concerned, the issue has not been studied, but it is probably correct to conclude that Solon has a fixed non-distinctive expiratory stress on the first syllable of a phonological word. It has also been proposed (Tsumagari) that there is a high pitch on the penultimate “mora” of a word, that is, on the penultimate syllable of words ending in a short vowel (CV#), or on the last syllable of words ending in a long vowel (VV#) or a consonant (VC#). This pattern, which would be similar to that observed in Daghur, would contribute to the coherence of the structural identity of a word. WORD FORMATION Solon retains the morphological distinction between nominals and verb(al)s, with regular nouns, spatial nouns, adjectives, numerals and pronouns belonging to the nominals. Ambivalent stems of the nomen-verbum type occur mainly in the realm of items pertaining to natural and physical phenomena, e.g. ɪmanda(-) ‘snow’ : ‘to snow’, edin(-) ‘wind’ : ‘to blow (of wind)’, ɪlaan(-) ‘light’ : ‘to glitter, to shine’, as in HC ɪmanda ɪmanda-jɪ-ra-n ‘it is snowing’ (snow snow-PROGR-AOR-VX3), KT exuuddi edin edin-ji-re-n ‘hot wind is blowing’ (hot wind blow-PROGR-AOR-VX3). Other examples of nomina-verba (zero derivation) are only occasional, e.g. xʊlaxa (> xʊlax) ‘thief’ : xʊlaxa- (> xʊlxa-) ‘to steal’ and are conditioned by various diachronic factors. Adjectival nominals are distinguished from regular nouns primarily by syntactic position and some derivational idiosyncracies,

212 Sangyub Baek

but not by inflectional morphology. An adjective can also function as an independent head noun, e.g. nisuxun ‘small’ ~ ‘small one’, as in HC miti nisuxun-sel-we-ni tiĕwɪ-gaarɪ ‘let us gather the small ones’ (1PL.INCL small-PL-ACC-PX3 gather-IMP.1PL.INCL). Derivational morphemes are specialized to being used with either nominal or verbal roots. Denominal nominals and verbs, as well as deverbal nominals, with the exception of participles, involve lexical operations, while deverbal verbs convey grammaticalized functions connected with differences of voice, mood, aspect, and Aktionsart. Adjectives, numerals, and pronouns have also specific derivational patterns. The following is a selection of some of the most frequent derivational forms based on nominal and verbal roots: (1a) Denominal nouns: -sIn [persons engaged in a certain profession or activity], e.g. iixe ‘pot’: iixe-sin ‘cook’, xʊnɪn ‘sheep’ : xʊnɪ-sʊn ‘shepherd’; -xsA [fur of animals], e.g. ɪmagan ‘goat’ : ɪmaga-xsa ‘goat fur’, tuulge ‘wolf’, tuulge-xse ‘wolf fur’; -ttUn [hard coverings], e.g. erugun ‘thumb’ : erugu-ttun ‘thumb protector made of bone for archery’, ʊnaxan ‘finger’ : ʊnaxa-ttʊn ‘ring’; -ldUr [soft coverings], e.g. nɪxama ‘neck’ : nɪxama-ldʊr ‘scarf’, sien ‘ear/s’ : siel-dʊr ‘earmuff/s’; -xAn [lexicalized diminutives], e.g. os-xon ‘fish’, ʊna-xan ‘finger’, us-xen ‘knife’, xolɪɪ-xan ‘insect, worm’. Solon also retains lexicalized Proto-Tungusic denominal derivatives like the collective countables in -ttA (< *-G.tA), e.g. nuu-tte ‘(human) hair’, ɪƞa-tta ‘hair’, ii-tte ‘tooth’, oosɪ-tta ‘star’, ʊsɪ-tta ‘nail’. The original nominal bases of these items have been lost. (1b) Denominal adjectives: The most important productive suffix deriving adjectives from nouns is the proprietive suffix -sI (< *-cII), e.g. anta/n ‘taste’ : PROPR anta-sɪ ‘tasty’, sanaa ‘compassion’ : PROPR sanaa-sɪ ‘compassionate’, unugu ‘cow’ : PROPR unugu-si ‘having a cow, with cows’. Another productive suffix is -mAAn [having a preference for or habit of eating or drinking something], e.g. osxon ‘fish’ : osxon-moon ‘fond of eating fish’, ulde ‘meat’ : ulde-meen ‘fond of eating meat’. There is also an array of evaluative suffixes, including -xAn, -xAya, -kkAn, and -sIlA, indicating the degree of intensity, e.g. aya ‘good’ : aya-xan ‘rather good’, aya-xaya ‘almost good’, aya-kkan ‘very good’, gʊdda ‘tall’ : gʊdda-sɪla ‘rather tall’. Solon, like Orochen, also employs the Mongolic device of reduplicating the initial sequence (C)V of adjectival roots, complemented by a syllable-final -b ~ -m, e.g. aya ‘good’ : a-b & aya ‘very good’, gɪltarɪn ‘white’ : gɪ-b & gɪltarɪn ‘very white’, xonnorɪn ‘black’ : xo-b & xonnorɪn ‘very black’. (2) Deverbal nominals: Most items in this category denote instruments, but there are also suffixes for the results, methods or other aspects of action. • Instruments: -ŋxI, e.g. sɵxɵ- ‘to ladle’ : sɵxɵ-ŋxi ‘ladle’, ɪla- ‘to burn’ : ɪla-ŋxɪ ‘fuel’; -ŋxU, e.g. saxsɪ-‘to hoe’ : saxsɪ-ŋxʊ ‘hoe’; -(U)r [also materials], e.g. malt/a- ‘to harrow’ : malt-ʊr ‘harrow’, bʊdʊ- ‘to dye’ : bʊdʊ-r ‘dye’; -UUn, e.g. ɪddɪ- ‘to comb’ : ɪdd-ʊʊn ‘comb’, tixxe- ‘to nail’, tixx-uun ‘nail’, xadɪ- ‘to cut’ : xad-ʊʊn ‘cutter’; -sUn, e.g. sɪna- ‘to wedge’ : sɪna-sʊn ‘wedge’, xadɪ- ‘cut’, xadɪsʊn ‘board’. • Others: -n [results], e.g. giela- ‘to separate’ : giela-n ‘room’, tɵtɵgɵ- ‘to decide’ : tɵtɵgɵ-n ‘restriction’; [results] -gAn, e.g. tarɪ- ‘to plant’ : tara-gan ‘field’, bodo‘to think’, bodo-gon ‘method, strategy’; -lgA [methods], e.g. sɪlxɪ- ‘to wash’ : sɪlxɪ-lga ‘the way how to wash’, naƞɪɪ- ‘to open’ : naƞɪɪ-lga ‘the way how to

Solon 213

open’; -ssA [consequences], e.g. ɪla- ‘to blow one’s nose’ : ɪla-ssa ‘nasal mucus’, saaƞʊ- ‘to be frosty’ : saaƞʊ-ssa ‘frost’; -ttAlA [something remaining after an action is performed], e.g. jip- ‘to eat’ : jit-ttele ‘left-overs of food’. (3) Denominal verbs: The basic verbalizing suffix attached to nouns is -lAA-, which implies a subject or object relationship with the underlying noun, e.g. doo ‘song’ : doo-loo- ‘to sing’, xʊdaa ‘marriage’ : xʊdaa-laa- ‘to marry’, ɪlgaa ‘flower’, ɪlgaalaa- ‘to bloom’. Another suffix is -dAA-, which implies an instrumental relationship with the underlying noun, e.g. mɪɪsan ‘gun’ : mɪɪsan-daa- ‘to shoot’, ʊxxʊn ‘rope’, ʊxxʊn-daa- ‘bundle’. (4) Deverbal verbs: Solon has four derivationally marked voice forms: the passives in -wU/U- ~ -gUU-, e.g. jawa- ‘to catch’ : PASS jawa-wʊ- > jaw-ʊʊ- ‘to be caught’, mandaa- ‘to hit’ : PASS mandaa-wʊ- ‘to be hit’, sɪlba/a- ‘to instruct’ : PASS sɪlbaa-gʊʊ- ‘to be instructed’; the causatives in -xAAn-, e.g. CAUS jawa-xaan- ‘to make someone catch’, CAUS mandaa-xaan- ‘to make someone hit’; the cooperatives (sociatives) in -ldI-, e.g. RECIPR mandaa-ldɪ- ‘to hit each other’; and the reciprocals in -mAAsI-, e.g. taa- ‘to pull’ : RECIPR taa-maasɪ- ‘to pull together’. The passive suffix can also be used to produce causatives from intransitive verbs, e.g. ii- ‘to enter’ : CAUS ii-guu- ‘to let in’, yuu- ‘to go out’ : CAUS yuu-guu- ‘to let out’. The functional difference between reciprocal and cooperative forms is fuzzy, and there are examples of the cooperative suffix being used in a reciprocal function, e.g. K tac-cɪl mandaa-ldɪɪ-sa/a ‘they hit one another’ (that-PL hit-COOP-PST). Modal and aspectual forms include the desideratives in -mUUn-, e.g. jip- ‘to eat’ : DESID ji-muun- ‘to want to eat’; the progressives in -jI-, e.g. jɪnjɪ- ‘to say’ : PROGR jɪnjɪ-jɪ- ʻto be in the action of saying’, soƞo- ‘to cry’ : PROGR soƞo-jɪ- ʻto be in the action of crying’; the statives in -sI-, e.g. ili- ‘to stand’ : STAT ili-si- ʻto be in the state of standing’, xulee- ‘to lie down’ : STAT xulee-si- ʻto be in the state of lying’; the reversives in -rgI-, e.g. eme- ‘to come’ : REV eme-rgi- ‘to come back’, buu- ‘to give’ : REV buu-rgi‘to return’; and the andatives in -nAA-, e.g. jawa- ‘to catch’ : AND jawa-naa- ‘to go to catch’, beyuusi- ‘to hunt’ : AND beyuusi-nee- ‘to go to hunt’. The synthetic progressive aspect marker -jI- can also be replaced by the analytic progressive construction involving the main verb in the form of the connective converb in -m/I in combination with the copula-existential bi- ‘to be’. It may be noted that the inchoative marker *-l-, common in the other Ewenic languages, including Orochen, is in Solon preserved only in a few lexicalized traces, including dege-li- ‘to fly’ and gugge-l- ‘to move’. Instead, an analytic construction with the verb exili- ‘to start’ is used. In this connection it may also be mentioned that the subjunctive mood in -mcA- (< *mU-cAA-), as known from Siberian Ewenki, Neghidal, and Ewen, is not attested in Solon. NUMBER AND CASE Nominals take marking for number (plural), case, and person (possessor), in this order (N-NX-CX-PX). A special type of nominal stems is formed by those ending in the primary nasal /n, which can be dropped before suffixes. The plural marker is -sAl (< *-sA-l), used without regard to animacy, e.g. ukkeexe/n ‘boy’ : PL ukkeexee-sel, xʊnɪ/n ‘sheep’ : PL xʊnɪ-sal, yeem/e ‘thing’ : PL yeem-sel, jolo ‘stone’ : PL jolo-sol. Nevertheless, it has to be emphasized that this plural marker has a rather emphatic connotation, and its use

214 Sangyub Baek

is not obligatory. In particular, it is omitted when the nominal is modified by a numeral or a quantifier, e.g. juur ukkeexen ‘two boys’ (two boy), baraan maƞɪɪ ‘a lot of devils’ (many devil). The simple plural markers *-l and *r are attested only in a few relics, e.g. ur ‘seed’ : PL uri(-)l ‘children’. In previous studies, varying opinions have been expressed concerning the number of cases in Solon, ranging from 11 to 14. Here, it will be assumed that the system has 13 cases. including the unmarked basic form, or nominative. The suffixally marked cases are: genitive, accusative, partitive, dative, ablative, locative, delative, prolative, directive, instrumental, elative, and comitative (Table 9.4). The delative is a specifically Manchurian Ewenic case, shared with Orochen, while the comitative is specific to Solon. TABLE 9.4 SOLON CASE MARKERS  

V

GEN

-nI/I

C

ACC

-w/A : -wAA-

-bA : -bAA-

PART

-y/A

-A

DAT

-dU

ABL

-dUxI

LOC

-(dU-)lA/A

-dUlA/A

DEL

-(dU-)lAAxI

-dUlAAxI

PROL

-(dU-)lI/I

-dUlI/I

DIR

-t(I)xI

-tIxI

INSTR

-jI

ELAT

-gIIjI

COM

-gIlII

/n -(n)I/I -mA : -mAA-

Stem types: V = vowel stems, C = consonant stems, /n = primary nasal stems. Segmental alternations: A = a e o ɵ, U = ʊ u, I = ɪ i.

The case markers show only minor allomorphic variation, connected with vowel harmony as well as the formal distinction between stems ending in a vowel (V), a consonant (C), or the primary nasal /n. There are rather few original consonant stems, the type example being juug ‘house’ (< *juug, with a change of the harmonic status in Solon), but even it tends to lose the final consonant. There are, however, relatively many stems ending in the liquids l r, including the numeral juur ‘two’ (originally a plural in *-r). Secondary consonant stems, formed due to the loss of final vowels, are treated like vowel stems. •

The nominative functions as the case of the subject and the nominal predicate, e.g. KT nelki eme-cci namad oo-soo ‘as spring came, it became warm’ (spring come-CV. ANT warm become-PST), cf. also KT [1]:

[1] muu mandɪɪ eme-cci juu ga-wʊ-saa water great come-cv.ant house take-pass-pst ‘As there came a lot of water, the house was taken (by it).’

Solon 215

•

•

Like in Orochen, the nominative can, however, also be used to indicate an indefinite or unspecific object, e.g. KT min-du norɪ buu-xe ‘give me an arrow!’ (1SG-DAT arrow give-IMP.2SG). The exact rules of object marking remain unclarified, but the lack of marking on object is without a doubt due to Mongolic influence. The regular object cases are the accusative in -w/A ~ -bA for a definite and the partitive (“indefinite accusative”) in -y/A ~ -A for an indefinite object, cf. e.g. K tarɪ tayyaa dilxum-be waa-saa ‘he killed that fly’ (that that.ATTR fly-ACC kill-PST) vs. HC sii orotto-yo xadɪ-xa ‘cut (some) grass!’ (2SG grass-PART cut-IMP.2SG). Before the third person possessive suffix -nI the accusative marker has the allomorph -wAA~ -bAA-, as also in Orochen. The partitive marker has also been recorded in the form -I = -yI, e.g. K bii emun umutta-[y]ɪ edde-su[-u] ‘I broke an egg’ (1SG one egg-PART break-PST-1SG). The case here glossed as “genitive” (GEN) reflects the original possessive form in *-ƞII, which in Solon, even more so than in the other Ewenic languages, functions as a true genitive case, marking the adnominal possessor on both nouns and pronouns, e.g. ukkeexen-ii unegel ‘the boy’s story’ (boy-GEN story), miti-nii si[g]un ‘our sun’ (1PL. INCL-GEN sun). As in the other languages of the region, it can also mark the subject in nominal clauses, e.g. KT sin-ii jɪnjɪ-jɪ-r-sɪ jʊxʊ-ra-n ‘what you are saying is right’ (2SG-GEN say-PROGR-PTCP.AOR-PX2SG be.right-AOR-VX3), cf. also K [2]:

[2] bii tar-nɪɪ er ineg eme-b-be-n 1sg that-gen this day come-ptcp.aor-acc-px3 ‘I know that he is coming today.’ •

•

•

saa-m know-aor.1sg

The dative in -dU marks an arrival point, e.g. tayyaa uriilen-du ‘to that village’ (that. ATTR village-DAT), a point of time, e.g. tarɪ u[y]i-du ‘at that time’ (that time-DAT), a static location, e.g. KT jʊxaal-dʊ-n=kɪ jɪggaa-mɪ am[ɪ]dɪr-saa ‘he lived happily at the summer camp’ (summer.camp-DAT-PX3=TOP be.happy-CV.CONN live-PST), or a beneficiary, e.g. KT ukkeexen-du-wi juu oo-m buu-see ‘he made a house for his son’ (boy-DAT-RX house make-CV.CONN give-PST). In existential and possessive constructions it indicates location or possessor, e.g. KT nimeer uriilen-du beye bi-si-n ‘there are people in the neighbouring village’ (neighbour village-DAT person be-AOR-VX3), KT min-du ukkeexen nexun aasɪn ‘I do not have a younger brother’ (1SG-DAT boy younger.brother NEG.EXIST). With participles, the dative forms quasiconverbal constructions with a temporal function, e.g. C mʊrɪn ʊgʊ-d-dʊ-wɪ aya-jɪ xise-xe ‘when riding a horse, be more careful!’ (horse ride-PTCP.AOR-DAT-RX good-INSTR be.careful-IMP.2SG), K asɪ edi juur-i/i am[ɪ]dɪ-r-dʊ-wɪ ‘when a wife and a husband live together’ (wife husband two-COLL live-PTCP.AOR-DAT-RX). The ablative in -dUxI is the separative counterpart of the dative and has a corresponding range of multiple functions, including the indication of source, e.g. tayyaa mʊrɪn-duxɪ ‘from that horse’ (that.ATTR horse-ABL), starting point, e.g. beejin-duxi xaylar jaxxa ‘from Peking to Hailar’ (Peking-ABL Hailar until), and the standard of comparison, e.g. HC exur-ni ulde-nin xʊnɪn-nɪ ulde-duxi anta-sɪ ‘beef is tastier than mutton’ (cow-GEN meat-PX3 sheep-GEN meat-ABL taste-PROPR). The locative in -lA/A ~ -dUlA/A refers to a location where the action is performed, e.g. HC ju[u]g-dulee esi enixen emu-xxen aasɪn-jɪ-ra-n ‘(there is) a woman sleeping alone in the house now’ (house-LOC now woman one-LIM sleep-PROGRAOR-VX3), K derem-bi emun jolo-lo-n nee-see ‘(he) put his cow breast on a stone’

216 Sangyub Baek

•

•

•

•

(breast-RX one stone-LOC-PX3 put-PST). Compared with the dative, which refers to an exact location or point of arrival, the locative has a more general reference. The prolative in -lI/I ~ -dUlI/I refers correspondingly to movement along a surface or through something, e.g. tayyaa tugguu-lii ‘along that road’ (that road-PROL), sʊnta bʊga-lɪɪ ‘through a deep place’ (deep place-PROL). The delative in -lAAxI ~ -dUlAAxI is the separative counterpart of the locative and denotes the area from where an action takes place, e.g. HC muu-leexi exur emeji-re-n ‘the cow is coming from the water area’ (water-DEL cow come-PROGRAOR-VX3). By comparison, the elative in -gIIjI refers to the inside of a source, e.g. HC bii xiena-gɪ(ɪ)jɪ malta-sʊ-ʊ ‘I dug (it) out from the pile of grass’ (1SG grassELAT dig-PST-1SG). The directive in -tIxI ~ -txI expresses the direction of action (‘towards’), e.g. HC esxe tarɪ ure-tixi ul-see ‘(my) uncle went towards that mountain’ (uncle that mountain-DIR go-PST), K asɪ-tɪxɪ-wɪ gun-ji-re-n ‘he is speaking to(wards) his wife’ (wife-DIR-RX say-PROGR-AOR-VX3). In some cases the directive seems to have a separative function, e.g. doo-txɪ osxon jaw/a- ‘to catch fish from a river’ (river-DIR fish catch). Such examples may, however, involve a confusion with the phonetically similar ablative suffix -dUxI. The instrumental in -jI expresses instruments, means, and materials, e.g. bɵɵsɵ ‘cloth’ : INSTR bɵɵsɵ-ji ‘of cloth’, mʊrɪn ‘horse’ : INSTR mʊrɪn-jɪ ‘by horse’, sel/e ‘iron’ : INSTR sel-ji ‘of iron’. In addition, it has a comitative function, e.g. bii emme-ji ‘I and/with mother’ (1SG mother-INSTR). Moreover, it forms modal adverbs from adjectival nominals, e.g. aya ‘good’ : INSTR aya-jɪ ‘well’, nandaaxan ‘beautiful’ : INSTR nandaaxan-jɪ ‘beautifully’, nagan ‘slow’ : nagan-jɪ ‘slowly’. In the latter function it resembles a derivational suffix. The status of the comitative in -gIlII is unclear, and the origin of the comitative marker is unknown. It is mentioned in older literature (Poppe), e.g. asɪɪ ‘wife’ : COM asɪɪ-gɪlɪɪ ‘with (one’s) wife’, juug ‘house’ : juug-gilii ‘with a house’, where the suffix -tie, borrowed from Mongolic -tie ~ -tai [-thɛ:], is also mentioned as being used in Solon, e.g. nɪnaxɪn-tie-wɪ ‘with his (own) dog’ (dog-COM-RX). However, modern sources do not confirm these data. Possibly, the comitative has been fully replaced the instrumental case.

NUMERALS The basic cardinal numerals are: 1 emun, 2 juur, 3 ɪlan, 4 digin, 5 tʊƞa ~ toƞa, 6 niƞun, 7 nadan, 8 jakkʊn ~ jaxʊn, 9 yegin, 10 jaan. Old Mongolic loanwords (probably transmitted by Manchu) are used for 20 orɪn, 30 gotɪn (< *gʊtɪn, by the analogy of orɪn), 40 dexi, while a more recent Mongolic loanword is used for 90 yereen. The other decades are formed by the suffix -ƞƞie, i.e., 50 tʊƞƞie, 60 niƞuƞƞie, 70 nadaƞƞie, 80 jaxʊƞƞie, though older sources (Poppe) also show the suffix -r-nII, i.e., 50 toƞa-rnɪɪ, 60 niƞu-rnii, 70 nada-rnɪɪ, 80 jakkʊ-rnɪɪ. For the powers of ten, the native Tungusic item 100 namaajɪ and the Mongolic borrowings (possibly transmitted by Manchu) 1000 (emun) mɪƞgan ~ mɪƞan and 10000 (emun) tumun are used. The intermediate numerals are formed by the additive principle, e.g. 11 jaan emun, 12 jaan juur. As a whole, the system is identical with that attested in Orochen. As a diachronic curiosity, Solon preserves a couple of Para-Mongolic numerals for the teens, transmitted via Jurchen-Manchu, in the names of the months. In general, months are numbered by the regular numerals, e.g. ɪlan bie ‘March’ (three month), but ʊnsun bie

Solon 217

‘November’ and jʊrgʊn bie ‘December’ retain reflexes of the Para-Mongolic numerals for ‘eleven’ and ‘twelve’. A special item is also anɪ bie ‘January’ (year month). Ordinals are formed by the suffix -sI, e.g. emu-si ‘first’, juu-si ‘second’, ɪla-sɪ ‘third’, etc., in older sources (Poppe) also by the direct borrowings from Manchu in -cI, e.g. ɪla-cɪ ‘third’, duyu-ci ‘fourth’. For the first two ordinals, the Manchu borrowings uju ‘first’ and jaɪ ‘second’ have been recorded. Alternatively, the Mongolic ordinal formative dugeer (in this form, apparently treated as an independent word) can also be used, e.g. emun dugeer ‘first’, juur dugeer ‘second’, ɪlan dugeer ‘third’. Other numeral derivatives include the collectives in -I/I or -nI/I ~ -nnie, e.g. juur-ii ‘two together’, ɪla-nɪɪ ~ ɪla-nnie ‘three together’; the distributives in -tAl ~ -tUl, e.g. emu-tul ‘one each’, juu-tul ‘two each’, ɪla-tal ‘three each’; the limitatives in -xxAn, e.g emu-xxen ‘only one’ > ‘alone’, ɪla-xxan ‘only three’; and the multiplicatives in -rAA, e.g. ɪla-raa ‘three times’. The multiplicative function can also be expressed analytically with the words mʊdan or (from Mongolic) ʊdaa ‘time’, e.g. ɪlan mʊdan ~ ɪlan ʊdaa ‘three times’. PRONOUNS Pronouns in Solon may be divided into the standard four categories of personal, demonstrative, reflexive and interrogative pronouns. The basic personal pronouns (Table 9.5) follow the pattern shared by the other Ewenic languages, with different stems for the basic and the oblique forms. The object forms contain the markers of both the partitive and the accusative, while the genitives may represent mergers of the original short pronominal genitives in *-i and the possessive forms in *-ƞII. The use of the basic forms in the subject position, as well as of the genitive forms in adnominal position, is not obligatory, as the pronominal reference can also be expressed by the suffixal person markers. TABLE 9.5 SOLON PERSONAL PRONOUNS   SG PL

NOM

OBL

PART-ACC

GEN

1

bii

min-

min-e-we

min-ii

2

sii

sin-

sin-e-we

sin-ii

buu

mun-

mun-e-we

mun-ii

suu

sun-

sun-e-we

sun-ii

1 2

EXCL

For the first person plural inclusive function the Common Ewenic stem miti (< *miti) : OBL miti- : ACC miti-we, etc., is used. In the third person, the pronoun noga- (< *ñuƞa-) is attested in old sources for Nonni Solon (Ivanovskii), with regular forms such as ACC-PX3 nogan-ma-n : DAT-PX3 nogan-dʊ-n : PL noga-r-, but it seems to have been lost in modern Hailar Solon, which uses the demonstrative pronouns, especially tar/ɪ ‘that’, also in reference to persons, often in combination with a suitable regular noun, e.g. tar ukkeexen ‘that boy’ > ‘he’ : PL tar ukkeexee-sel ‘those boys’ > ‘they (males)’, tar ʊnaajɪ ‘that girl’ > ‘she’ : PL tar ʊnie-sal ‘those girls’ > ‘they (females)’. Among the Ewenic languages, this feature is unique to Solon and is obviously due to the influence of the neighbouring Mongolian (proper) language, as well as Manchu. It may be noted that Daghur originally does have a third person pronoun, but it has been replaced by the demonstratives in Hailar Daghur.

218 Sangyub Baek

The reflexive stem is mee/n-, used independently in the form meeni ‘oneself’. The oblique forms take the reflexive suffix -wi ~ -bi after the case marker, e.g. DAT-RX meendu-wi : LOC-RX meen-dulee-wi, etc., except that the object form remains unmarked for case, i.e., RX meem-bi. The possessive form is attested both with and without the reflexive suffix, i.e., meen-ii ~ meen-ii-wi ‘one’s own’. In the emphatic function, the form meeƞ-xen ‘(by) oneself’ is used. Unlike the other Ewenic languages, modern Hailar Solon uses these forms indistinctly for both singular and plural reference. The reciprocal function is formed by reduplicating the reflexive pronoun: meeni & meeni ‘(with regard to) each other’. The demonstrative pronouns are based on the opposition of the roots e- (proximal) vs. ta- (distal). The basic free forms are eri vs. tarɪ, which can also be used independently in reference to persons, e.g. KT eri bi-kki samaa-txɪ mandɪ neele-re-n ‘as for him, he is very much afraid of a shaman’ (this be-CV.AOR shaman-DIR very be.afraid-AOR-VX3), KT tarɪ ʊʊtaa-ccɪ yuu-se ‘he went out in a hurry’ (that hurry-CV.ANT go.out-PST), but also adnominally, e.g. eri beye ‘this person’ > ‘s/he’, tarɪ sukkel ‘that devil’. The final vowel can be dropped especially in adnominal position, e.g. er ineg ‘this day’, tar bʊga ‘that place’. There are also the expanded forms eye/e ~ eyye/e vs. taya/a ~ tayya/a, often used adnominally, e.g. eyyee comoo ‘this cup’, tayyaa bitege ‘that book’, but also independently. The plural forms are variously er-sel vs. tar-sal or ec-cil vs. tac-cɪl, or also elur ~ ellur ‘these’ (< eri ʊlʊr) vs. talʊr ~ tallʊr ‘those’ (< tarɪ ʊlʊr), based on combinations incorporating the noun ʊlʊr ‘people’ (from Daghur). On the demonstrative pronouns are based the demonstrative verbs too- ‘to do like that’ (< ta- + oo- ‘to do’) : ottoo- < et-too‘to do like this’ (< er + t-oo-), which, again, yield lexicalized forms like CV.CONN too-mɪ ‘doing like that’ > ‘in that way’ > ‘so’. The basic interrogative pronouns are ii- (< *ïa-) : ɪɪ ‘what?’ : ɪ-xʊn > ʊxʊn > oxon ‘what?’ and nii (< *ƞii) ‘who?’. Other interrogative roots include ii- ~ i- (< *ii-), which yields local adverbs like DAT i-du ‘where?’ : ABL i-duxi ‘from where?’ : LOC i-lee ‘where?’ : PROL i-lii ‘which way?’ : DEL i-leexi ‘from which direction?’ and the secondary stem ii-r ~ i-r ‘which?’ : GEN in-ni : ACC ib-be : PART iy-ye : DIR it-tixi ‘which way?’ : INSTR ij-ji ‘how?’ : ELAT ig-giji ‘from where?’ : DER ig-gu/u ‘which one?’ : DER it-tuu ‘how?’; as well as a- (< *xa-) and oo- (< *oo-), which yield a heterogeneous selection of question words like a-gʊʊ > a-wʊʊ ~ a-wʊ ‘who?’ : PL a-wʊ-sal : a-dɪɪ ‘how many?’ : aa-l ‘when?’ and oo-xɪ/ɪ ‘how much?’ : oo-xɪ-dʊ ‘when?’ : oo-xɪ-jɪ ‘for how much’ : oo-ndɪɪ ‘what kind of?’, verbalized as oo-n- ‘to do how?’. An isolated item is yoondaa ~ yoodaa ~ yoodon ‘why?’, borrowed from Daghur yoo/n ‘what’ : DAT-RX yoon-d-aa ‘why?’. Another Mongolic borrowing from the same root is yeem/e ‘something’ > ‘thing’. Other indefinite pronouns are formed from the interrogatives by clitics such as =xAt/I, e.g. oo-ndɪɪ=xatɪ ‘some kind of’. PERSON MARKING The system of person marking in Solon is somewhat simplified as compared with the other Ewenic languages (Table 9.6). The basic dichotomy between possessive suffixes (PX) and predicative personal endings (VX) is preserved only in the singular first and second persons, while in the other persons suffixes of the possessive set are used invariably for all purposes. There is, however, some secondary differentiation between the nominal and participial paradigms. Also, Solon has lost the distinction between the singular and plural forms in the third person, though the reflexive set (RX) still retains this distinction. Participles in finite use with a third person subject bear no marking for person or number.

Solon 219 TABLE 9.6 SOLON PERSON MARKERS    

PX    V

PTCP

AOR

1

-bI

~ -wI

-U

-mI

2

-CI

-s/I

-bI

-wI

/n SG

RX PL

3

VX

1

EXCL

-mUn

INCL

C

-tI

-t/I

2

-CUn

-sUn

RX

-bAl/I

-wAl/I

-In/I

-nIn/I

-(nI-)n/I

-ndI

-ttI

-ø

-n

Stem types: C = consonant stems, V = vowel stems, /n primary nasal stems. Segmental alternations: A = a e o ɵ, U = ʊ u, I = ɪ i, C = c s.

There are several details concerning the person markers that are not confirmed by reliable and/or unambiguous information. For instance, older sources (Poppe) claim that the person markers do not follow vowel harmony, while according to more recent information harmonic variation seems to be present. In the second person possessive suffixes, the initial s was apparently earlier regularly represented as c in postconsonantal position, but modern sources no longer confirm this alternation and use s also after consonants. The vowel of the plural reflexive suffix -bAl/I ~ -wAl/I has also been reported (Hu & Chaoke) as I, e.g., -bIl ~ -wIl. The first person singular possessive suffix -bI has the variant -wI, but also -wAyA, after oblique case markers, except in the accusative, which has -yA. In the nominal phrase the possessive suffixes indicate the possessor, e.g. mʊrɪn-sɪ ‘your horse’ (horse-PX2SG), dɪl-nɪn ‘his/her/its/their head’ (head-PX3). An important feature of Solon, as compared with all other Ewenic languages, including Orochen, is that it lacks the category (and marker) of alienable possession. Instead, there is a tendency to indicate alienability by omitting the possessive suffixes after the possessively used genitival forms of the personal pronouns, e.g. min-ii jaxa ‘my baggage’ (1SG-GEN baggage), min-ii biteg ‘my book’ (1SG-GEN book). In cases of unalienable possession, the possessive suffixes are used more regularly, e.g. sin-i/i aba-sɪ ‘your father’ (2SG-GEN father-PX2SG). The reflexive forms refer to the subject. The basic reflexive forms, unmarked for case, imply object position, as in KT [3]: [3] eyyee guren-ii waƞ-du bikki mʊrɪm-bɪ this.attr country-gen king-dat top horse-rx ‘(He) gave his horse to the king of this country.’

buu-see give-pst

When the subject is in the plural, the corresponding plural reflexive suffix is used, e.g. HC tallʊr meeni juug-du-wil nenu-see ‘they returned to their own house’ (those REFL house-DAT-RX.PL return-PST). As in the other Ewenic languages, the predicative personal endings (VX) in Solon are used in the aorist type of finite conjugation, which includes also all the derivational forms referring to the categories of voice, mood, aspect, and Aktionsart. The actual finite

220 Sangyub Baek

paradigm also comprises forms based on participles, which in the finite function take person markers of the possessive type, though only in the first and second persons. VERBAL MORPHOLOGY As in other Ewenic languages, the verbal forms in Solon may be divided into finite indicative, finite imperative, non-finite converbal, and polyfunctional participial forms. In the finite template the order of the constituents is: stem-voice/aspect-tense/mood-person/ number. The difference between finite indicative and finitely used participial forms is, however, less clear than in the other Ewenic languages, for the system of participles is substantially reduced, and the finite aorist marker also functions as a participle, as in Orochen. The principal difference between the aorist and finitely used participial forms concerns the person markers, but even these have merged in several persons. As far as the formation of the aorist stem is concerned, Solon retains the four original types, marked by (1) -rA- after vowels and -tA- after consonants, e.g. eme- ‘to come’ : AOR eme-re-, jip- ‘to eat’ : AOR jit-te-; (2) -A- after original nasal stems, e.g. gun- ‘to say’ : AOR gun-e-; (3) -dA- in a few lexicalized examples, e.g. bu- ‘to die’ : AOR bu-de-, oo- ‘to become’ : AOR oo-do-, and (4) -sI- in the copula-existential bi- ‘to be’ : AOR bi-si- and the negation verb e- : AOR e-si. In type (1), the aorist marker -rA- is present only in the third person, while in the other persons the person markers are added directly to the stem, e.g. jawa- ‘to catch’: AOR 1SG jawa-mɪ : 2SG jawa-ndɪ : 3SG & PL jawa-ra-n : 1PL EXCL jawa-mʊn : 1PL INCL jawa-ttɪ : 2PL jawa-sʊn. In the other stem types the aorist marker is retained throughout the paradigm, e.g. aasɪn- ‘to sleep’ : AOR 1SG aasɪn-a-mɪ : 2SG aasɪn-a-ndɪ : 3SG & PL aasɪn-a-n : 1PL EXCL aasɪn-a-mʊn : 1PL INCL aasɪn-a-ttɪ : 2PL aasɪn-a-sʊn. In the predicative position the aorist functions as a non-past tense form with a present or future reference, e.g. HC sii awʊ-ya gele-ndi ‘whom are you looking for?’ (2SG who-PART search-AOR.2SG), KT tɪmaasɪn ineg sin-dule/e nene-m=iĕ ‘tomorrow I will definitely go to your place’ (tomorrow day 2SG-LOC go-AOR.1SG=EMPH). This is different from the other Ewenic languages, in which the aorist functions as a non-future tense, marking either present or past actions. Solon also lacks the finite (“first”) future tense, as attested in, for instance, Siberian Ewenki. However, present-tense reference can be emphasized by using the progressive marker -jI-, e.g. HC aba enie juur-i/i jaanjɪmaasɪ-jɪ-ra-n ‘father and mother are talking with each other’ (father mother two-COLL talk-RECIPR-PROGR-AOR-VX3). The aorist stem is also used as the invariable connegative form, combined with the fully conjugated forms of the negation verb e-, e.g. e-si-m saa-r/a ‘I do not know’ (NEGAOR-1SG know-CONNEG), e-si-n oo-do ‘it does not become’ > ‘it is not possible’ (NEG-AOR-VX3 become-CONNEG, a structural borrowing from Mongolian). PARTICIPLES Solon retains only three participial forms in active use. They may be identified as the aorist, perfective, and futuritive participles. Their temporal references are, correspondingly, present, past, and future. •

The aorist participle is marked by the suffix -r/A, e.g. jawa- ‘to catch’ : PTCP.AOR jawa-ra ‘one who will catch’ : PROGR-PTCP.AOR jawa-jɪ-ra ‘one who is catching’. Among the Ewenic languages, such participial use of the aorist is shared only

Solon 221

•

by Orochen. The suffix -r/A is also added to verbs that form the actual aorist paradigm with other suffixes, e.g. jip- ‘to eat’ : AOR jit-te- : PTCP.AOR jit-te-r/e, ga- ‘to take’ : AOR ga-da- : AOR-PTCP.AOR ga-da-r/a, bi- ‘to be’ : AOR bi-si- : PTCP. AOR bi-si-r/e. When expanded by case suffixes, the aorist participle yields quasiconverbal forms, which could also be viewed as fully grammaticalized converbs, e.g. ise- ‘to see’ : PTCP.AOR-DAT-PX3 ise-r-du-ni ‘when he is looking’, jawa- ‘to catch’ : PTCP.AOR-INSTR jawa-r-jɪ ‘in order to catch’. Older sources (Poppe) mention also the imperfective participle in -rI (< *-rII), based on the aorist, e.g. aasɪn- ‘to sleep’ : PROGR-PTCP.IMPRF aasɪn-ja-rɪ ‘one who is sleeping’. This suffix is also added to the verbs forming the aorist by the markers -A- or -dA-, e.g. ga- ‘to take’ : AOR-PTCP.IMPRF ga-da-rɪ. However, the imperfective participle seems to have disappeared from active usage in modern Hailar Solon. The perfective participle is marked by the suffix -sAA, after consonants -cAA (< *-cAA), e.g. waa- ‘to kill’ : PTCP.PRF waa-saa ‘one who has killed’, buu- ‘to give’ : PTCP.PRF buu-see ‘one who has given’, (primary nasal stem) gun- ‘to say’ : PTCP.PRF gun-cee ‘one who has said’, (secondary consonant stem) ul/i- ‘to go’ : ul-see ~ ul-cee ‘one who has gone’. In predicative position, person markers of the possessive type are used, with the additional feature that the first person singular marker *-w merges with the participial suffix, yielding the form -sU(-)U, e.g. bi- ‘to be’ : PTCP.PRF-1SG bi-su-u ‘I was’. The perfective participle is frequently used in the function of an adnominal modifier, e.g. KT min-ii xudaasɪ ga-saa uƞke-wi ‘the scarf that I bought dearly’ (1SG-GEN expensive take-PTCP.PRF scarf-RX). In finite use it is the standard form for expressing past tense, e.g. HC tallʊr uner bayan-jɪ tege-see ‘they lived really richly’ (they really rich-INSTR live-PST). With the accusative marker it can also be used to mark a nominalized complement sentence, as in K [4]:

[4] tayyaa that

eyyee this

tukke-w bag-acc

xoccoo-tɪxɪ market-dir

ga-saa-wa-n take-ptcp.prf-acc-px3

bii saa-jɪ-m=iĕ 1sg know-progr-aor1sg=emph ‘I know that s/he bought this bag at the market place.’ •

The futuritive participle is marked by the suffix -jAgA/A ~ -jIgA/A (< *-jA-ƞAA), e.g. nenu- ‘to return’ : PTCP.FUT nenu-jege/e ‘one who will return’. In predicative position, combined with person markers of the possessive type, this form marks the finite future tense, e.g. jawʊʊ-jagaa-n ‘he will be caught’ (be.caught-FUT-PX3). As in the case of the perfective participle, the first person singular marker *-w merges with the participle suffix, yielding the synthetic form -jUgU(-)U, e.g. jip- ‘to eat’ : FUT-1SG jij-jugu-u ‘I will eat’.

The perfective and futuritive participles can be combined with the copula-existential bi- ‘to be’ to form periphrastic constructions indicating a remote past (pluperfect), e.g. uli-cee bi-see ‘he had gone’ (go-PST be-PST), or a subjunctive, e.g. uli-jege bi-see ‘he would have gone’ (go-FUT be-PST). As in Orochen, the auxiliary bi-see functions in

222 Sangyub Baek

these constructions as an invariant tense marker, while the person markers are added to the main verb, e.g. eme-su-u bi-see ‘I had come’ (come-PST-1SG be-PST). IMPERATIVES Solon, like Orochen, makes a systematic difference between “proximal” or “present” and “distal” or “future” imperative forms. Both types of imperative have a full personal paradigm, except that no distinction is made for number in the third person (Table 9.7). TABLE 9.7 SOLON IMPERATIVE FORMS PROX

DIST

IMP SG

1 2

PL

1 2

3

PX/VX

C

V

-gAtI

~ -ktI

-XA/l

SUP

PX/RX

-DUU-

-nI

-DAA-

-wI

EXCL

-gAtI

-mUn

-DAA-

-mUn

INCL

-GAA-

-rI

-DAA-

-wAlI

-XAl-

-dUn

-DAA-

-sUn

-gI-

-nI

-DAA-

-nI

Stem types: C = consonant stems, V = vowel stems. Segmental alternations: A = a e o ɵ, U = ʊ u, I = ɪ i, D = d t, G = g k, X = x k.

Formally, the “proximal” imperatives represent the original Ewenic imperative paradigm, while the “distal” imperatives are personal forms of the purposive converb or “supine”. Functionally, the “proximal” imperatives refer to actions expected to occur in the immediate future, e.g. teg/e- ‘to sit’ : IMP 2SG tege-xel ~ tek-kel : 2PL tekkel-dun ‘sit down!’, HC mʊrɪn-dʊxɪ-wɪ ewu-xe ‘descend from your horse!’ (horse-ABL-RX descend-IMP.2SG), while the “distal” imperatives involve commands that are to be fulfilled in a more remote future, e.g. K ɪlan xonoo-dʊ nene-dee-wi ‘go (to that place) in three days!’ (three day.and. night-DAT go-SUP-RX). Also, the type of modality implied by the imperative forms varies from person to person. For instance, the first person plural inclusive forms signify a volition or an invitation to action, e.g. tege- ‘to sit’ : SUP-1PL.INCL tege-dee-weli ‘let us sit down (later)’, K mit ug-geer=ie ‘let us go!’ (1PL.INCL go-IMP.1PL.INCL=EMPH). Looking in more detail at the personal forms, we can see that the second person imperatives in SG -XA/l : -Xal-dU derive directly from their Proto-Ewenic sources *-kAl : *-kAl-sU > Siberian Ewenki -kAl : -kAl-dU, with the minor difference that the final l of the marker tends to be lost in the singular form. The third person imperative in -gI-nI also corresponds to the source form *-gI-, followed by the corresponding singular possessive suffix. In the first person, the plural inclusive form in -GAA-rI corresponds to Siberian Ewenki -gAA-r, and it cannot be ruled out that the Solon form preserves the original final vowel of this form. As in Siberian Ewenki, the first person singular and the first person plural exclusive forms are based on one and the same element, with the plural form marked by the corresponding possessive suffix, though the segmental correspondence of this element -GAtI(-) with its counterparts in the other Ewenic languages, based on the form *-GtA(-), is not regular.

Solon 223

In the distal paradigm, the third person form, as well as the first person exclusive and second person plural forms, are marked by the regular possessive suffixes, while the second person singular and first person plural inclusive forms are marked by the singular and plural reflexive suffixes, respectively. This is different from Siberian Ewenki, which, like also Ewen and Neghidal, uses the plural reflexive form of the “supine” in reference to the second person. The first person singular form in -DUU-nI involves several innovations: it incorporates the corresponding possessive suffix *-w, which has merged with the vowel of the “supine” marker, i.e. -DUU- < *-DAA-w, while it also contains the third person possessive suffix, apparently analogically generalized from the third person form. CONVERBS The system of converbs in Solon, like that of participles, is relatively reduced, but the converbal forms have a high frequency. With regard to their scope of reference, the converbs can be divided into the conjunct (same-subject) and ambivalent types. In disjunct usage, the latter take person markers of the possessive type, while in conjunct usage they either take reflexive suffixes or remain unmarked. In modern Hailar Solon, there are six converbal forms in active use (Table 9.8). TABLE 9.8 SOLON CONVERB MARKERS  

 

CONN

-m/I

C

PX

SS

DS

+

ANT

-ccI ~ -ssI

COMPL

-tAAn/I

+ +

DUR

-gII

+

AOR

-kkI-

IMMED

-jIxI- : -jIxAA-

-cI

-kI-

+

+

+

+

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct), C = consonant stems. Segmental alternations: A = a e o ɵ, I = ɪ i.

Superficially, at least, the converb markers are rather different from those used in the other Ewenic languages, and many converbs known from the other Ewenic languages are not used in Solon. For instance, the purposive converb in -DAA- is in Solon only preserved in the paradigm of the “distal” imperative forms, but not as a converb. •

The connective converb in -mI ~ -m, with formal cognates in all Tungusic languages, is basically used to connect two simultaneous events, e.g. KT ta-dʊ doo-loo-mɪ ili-siji-re-n ‘while singing (he) is standing there’ ~ ‘(he) stands there and sings’ (that-DAT song-VBLZ-CV.CONN stand-STAT-PROGR-AOR-VX3). However, unlike its cognates in the other Ewenic languages, the connective converb in Solon does not indicate conditional or temporal subordination, though, depending on the context, it can convey an implication of anteriority, e.g. KT sie baxa-m ɪm-saa-sʊƞ=gie ‘did you find the tea and drink it?’ (tea find-CV.CONN drink-PST-2PL=INTERR.EMPH), or causality, as in KT [5]:

224 Sangyub Baek

[5] biega jiggin ʊdan e-mi tʊkkʊ-r month about rain neg-cv.conn fall-conneg ‘As it has not rained for about a month, tʊsxaa-sal-nɪɪ jit-te-r orooto-nɪn calf-pl-gen eat-aor-ptcp.aor grass-px3 the grass for the calves to eat has withered.’

olgo-soo wither-pst

The connective converb can also be understood as indicating purpose, e.g. alban xaa-m uli-re-n ‘(he) goes to fulfil (his) official duty’ (official.duty fulfil-CV.CONN go-AOR-VX3). In fact, in combination with the andative suffix -nAA-, the connective converb forms the suffixal complex -nAA-mI, which in Chinese sources has also been analysed as a separate purposive converb, e.g. KT juur-i=kki uxum-be-n dal-naa-mı emesee ‘the two came to lick the milk’ (two-COLL=TOP milk-ACC-PX3 lick-AND-CV. CONN come-PST). When reduplicated, the connective converb can convey a pluritative, continuative, multiplicative, or emphatic meaning, e.g. KT juur nɪnɪxɪn-ɪn=kɪ unsii-m unsii-m unsii-m unsii-m ul-ji-re-n ‘his two dogs are going sniffing and sniffing’ (two dog-PX3=TOP sniff-CV.CONN sniff-CV.CONN sniff-CV.CONN sniff-CV.CONN go-PROGR-AOR-VX3). •

•

The anterior converb in -ccI, also -ssI and in earlier sources (Poppe) -kcI, is an irregular cognate of the form in *-ksA with a similar function in Siberian Ewenki and Orochen. It is used to link closely related events taking place in a time sequence, e.g. KT beye-wi sɪkkɪ-ccɪ eme-xe ‘come after taking a bath!’ (body-RX wash-CV. ANT come-IMP.2SG). It can often be understood as a causal connector, e.g. HC tayya mʊrɪn-dʊxɪ tixi-ssi beldir-wi xonso-soo ‘he fell down from the horse and broke his leg’ (that horse-ABL fall-CV.ANT leg-RX break-PST). It also forms the two lexicalized forms oo-ccɪ and bi-cci, based on the copular verbs oo- ‘to become’ and bi- ‘to be’ and functioning as coordinative connectors (‘and’), e.g. KT jaxʊn ʊnaajɪ oo-ccɪ ɪlan nɪnɪxɪn-ɪɪ uligir ‘the story of the eight girls and three dogs’ (eight girl become-CV.ANT three dog-GEN story), K lʊʊt oo-ccɪ sɪwʊn gurun ‘Russia and Japan’ (monster become-CV.ANT sun country), HC imin doo-nɪ muu-nin arʊkkʊn bi-cci anta-sɪ ‘the water of the Imin river is clean and tasty’ (Imin river-GEN water-PX3 clear be-CV.ANT taste-PROPR). The forms oo-ccɪ and bi-cci are based on Mongolic models. The completive converb in -tAAn/I is unique to Solon. It expresses the completion of an action and is functionally very close to the connective converb, e.g. KT tegeteen muu ɪmɪ-jɪ-ra-n ‘having sat down, (he) is drinking water’ (sit-CV.COMPL water drink-PROGR-AOR-VX3). In addition, in combination with the copula-existential bi- ‘to be’, it indicates duration of a completed state, as in KT [6]:

[6] ukk[e]-wi e-si-n naƞɪɪ-m door-rx neg-aor-vx3 open-cv.conn ‘(She) did not let open her door gʊljɪ-taan bi-ji-see close-cv.compl be-progr-pst but was keeping it closed.’

buu-r give-conneg

Solon 225

•

•

The durative converb in -gII is probably a cognate of the similar converb in -Gay in Orochen. It refers to a continued or repeated action and is often used with reduplication, e.g. HC tarɪ arɪxɪ ɪmɪ-gɪɪ ɪmɪ-gɪɪ sotto-soo ‘drinking and drinking, he got drunk’ (that liquor drink-CV.DUR drink-CV.DUR get.drunk-PST). It can also be combined with the limitative particle =l, e.g. KT ug-gii=l ug-gii=l ug-gii=l ug-gii=l emun nadan axɪn nexun-ii uriilen-d[u]-n ɪsɪ-naa-saa ‘he went and went and arrived at a village of seven brothers’ (go-CV.DUR=LIM go-CV.DUR=LIM go-CV.DUR=LIM go-CV. DUR=LIM one seven elder.brother younger.brother village-DAT reach-AND-PST). The aorist converb in -kkI ~ -kI (< *-rA-kI) takes person markers of the possessive type to indicate a non-coreferential subject, e.g. HC sii uli-kki-si bii uli-mi ʻif you go, I goʼ (2SG go-CV.AOR-2SG 1SG go-AOR.1SG). The aorist converb can, however, also be used in a corefential function, in which case it is normally unmarked, e.g. K edduu-lee-kki uril-nii uge-we-n dooldɪɪ-xoldo=nie ‘when you grow up, listen to children’s words!’ (big-VBLZ-CV.AOR children-GEN word-ACC-PX3 listen-IMP.2PL=PTCL). In practice, the aorist converb functions synchronically as a marker of general conditional and temporal subordination without regard to referentiality. This is possible since the functions of the connective converb in -m/I, which originally served as the corefential counterpart of the aorist converb, have become more restricted in Solon, leaving the aorist converb as the only marker of true unreal conditionals. Moreover, the aorist converb bi-kki of the copula-existential bi- ‘to be’ has received the role of a conditional marker, as in KT [7]:

[7] bii tarɪ erin-du tayyaa bayta-w saa-saa 1sg that time-dat that matter-acc know-ptcp.prf ‘If I had known it at that time, doroo ays[ʊ]-laa-mɪ nen-cee immediately help-vbzl-cv.conn go-ptcp.prf I would immediately have gone to help.’

bi-see bi-kki be-pst be-cv.aor

bi-su-[u] be-pst-1sg

In addition, the form bi-kki, or also RX bi-kki-wi, has been grammaticalized to a topic particle, e.g. K sii bikki aal=xa[t] min-e-w xasɪlaa-ndɪ ‘as for you, you always embarrass me’ (2SG TOP when=PTCL 1SG-PART-ACC embarrass-AOR.2SG). The topic particle is often used in constrastive constructions, as in KT tondo bey bikki gujee-me-cci, xʊnjaan bey bikki sisiri-wuu-re-n ‘an honest person is loved, while a cunning person is hated’ (honest person TOP love-PASS-CV.ANT, cunning person TOP hate-PASS-AOR-VX3). The uses of bi-kki reflect the models of similar constructions in Mongolic. •

The immediative converb in -jIxI- : -jIxAA- expresses an action that immediately precedes another action (‘as soon as’). Of the two allomorphs, -jIxI- is used before in the third person form -jIxI-n/I, e.g. P gurgel-buu-jixi-ni ta-dʊkɪ-nɪ ay xɵɵm yuure-n ‘as soon as (the dog) had shaken (the bell), there appeared good food’ (moveCAUS-CV.IMMED-PX3 that-ABL-PX3 good food come.out-AOR-VX3). This form is also attested in the fixed phrase gun-e-xen gun-jixi-ni ‘as soon as (s/he) had said (that) . . . ’ (say-AOR-DX say-CV.IMMED-PX3). The allomorph -jIxAAis used before the other person markers, e.g. HC bii uxxe-du ii-jixee-weye ‘as soon as I had entered by the door’ (1SG door-DAT enter-CV.IMMED-PX1SG), HC sii eme-jixee-si ‘as soon as you come’ (2SG come-CV.IMMED-PX2SG).

226 Sangyub Baek

PHRASE STRUCTURE In the nominal phrase, the modifiers precede the headword with no agreement in case or number, e.g. nandaaxan ukkeexee-sel ‘beautiful boys’ (beautiful boy-PL), aya xʊnɪsal-wa ‘the good sheep’ (good sheep-PL-ACC). This is different from Siberian Ewenki, which, in several dialects, shows a strong tendency to agreement. •

•

Possessive phrases use three alternative strategies: dependent marking, head marking, and double marking. Most commonly, possession is marked by the genitive case on the modifier, e.g. min-ii gebbi ‘my name’ (1SG-GEN name), ʊdan-ɪɪ muu ‘rainwater’ (rain-GEN water). However, the nominal headword can also take the corresponding possessive suffixes, e.g. min-ii axɪm-bɪ ‘my elder brother’ (1SG-GEN elder.brother-PX1SG), moo-nɪɪ naccɪ-nɪn ‘the leaves of the tree’ (tree-GEN leaf-PX3), tayyaa ukkeexen-ii dɪl-nɪn ‘that boy’s head’ (that boyGEN head-PX3), but with a pronominal modifier, head marking alone is also sufficient, e.g. dɪl-nɪn ‘his/her/its/their head/s’ (head-PX3). In this connection it may be noted that in spatial constructions, involving a postpositionally used spatial nominal after a regular noun, no marking is normally present on either member, e.g. xoggo oron-dʊ ‘on top of the cupboard’ (cupboard top-DAT), addar doo-loo ‘inside the box’ (box inside-LOC). There are, however, occasional exceptions from this. Apart from spatials, other postpositions are rare, but an example is jaarɪn ‘for (the sake of)’, which requires the genitive in the preceding nominal, e.g. KT enin-ci sin-ii jaarɪn bey-du e-ji-re-n nene-r=xun ‘for your sake your mother is not remarrying’ (mother-PX2SG 2SG-GEN for person-DAT NEG-PROGR-AOR-VX3 go-CONNEG=PTCL). When combined with participles, this postposition forms a purposive construction, as in D [8]:

[8] bitege-y tatɪ-r-nɪ book-part learn-ptcp.aor-gen ‘In order to study

jaarɪn in.order.to

mogo-r jogo-r-jɪ e-si-n be.difficult-ptcp.aor suffer-ptcp.aor-instr neg-aor-vx3 (one) is not afraid of difficulties and suffering.’

neeli-r be.afraid-conneg

In addition, also in combination with participles, the postposition jaarɪn can be used as a concessive connector, e.g. KT eyyee comoo tixi-see jaarɪn e-si-n edduu-re ‘although this cup fell, it did not break’ (this cup fall-PTCP.PRF for NEG-AOR-VX3 break-CONNEG). The postposition jaarɪn itself can also be used nominally in the meaning ‘replacement’, e.g. jaarɪn-nɪ beye ‘deputy’ (replacement-GEN person). •

Relative clauses with a participial form as the predicate also function as adnominal modifiers placed before the headword of a nominal phrase. No person markers are normally used on the participial predicate. Elements that can be relativized include both core arguments and peripheral adjuncts, e.g. (subject relativization) tayyaa uriilen-du eme-see beye ‘the person who came to that village’ (that village-DAT come-PTCP.PRF person), (object relativization) min-ii ga-saa uƞke ‘the scarf that I bought’ (1SG-GEN take-PTCP.PRF scarf), (adjunct relativization) emun uriilen-du

Solon 227

•

ɪsɪ-naa-r uy-du-n ‘at the time when (one) arrives at a village’ (one village-DAT reach-AND-PTCP.AOR time-DAT-PX3). A special type of verbal phrase is formed by the combination of the connective converb of a lexical main verb with the fully conjugated forms of a set of auxiliaries, which contribute various temporal, aspectual, and modal contents. The verbs used in such constructions include exili- ‘to begin’ [inchoative], e.g. oo-m exili- ‘to start making’ (make-CV.CONN start); ete- ‘to finish’ [completive], e.g. oo-m ete‘to finish making’ (make-CV.CONN finish); ete- ‘to be capable’ [capability], e.g. taƞɪ-m ete- ‘to know how to count’ (count-CV.CONN be.capable); bax/a-‘to find’ [intuition], e.g. bodo-m bax/a- ‘to come to think’ (think-CV.CONN find); bi- ‘to be’ [progressive], e.g. iggi-m bi- ‘to be raising’ (raise-CV.CONN be); oo- ‘to become’ [possibility or permission], e.g. ii-m oo- ‘to be allowed to enter’ (enter-CV.CONN become); is/e- ‘to see’ [attemptive], e.g. antaa-la-m is/e- ‘to taste’ (taste-VBLZ-CV. CONN see); buu- ‘to give’ [benefactive] sɪbbaa-m buu- ‘to show (for somebody)’ (show-CV.CONN give); ga- ‘to take’ [perfective], e.g. aysʊ-laa-m ga- ‘to save’ (help-VBLZ-CV.CONN take). Occasionally, the perfective converb can also be used in such constructions, e.g. nee- ‘to put’ [resultative], e.g. jic-ci nee- ‘to eat (completely)’ (eat-CV.ANT put).

SENTENCE TYPES Compared with Siberian Ewenki, Solon, like Orochen, conforms to a rather strict head-final word order, with the verb as the final constituent of both finite and non-finite clauses (SOV). The argument structure follows the nominative-accusative pattern, but an overt subject can be absent in a sentence, since it can be expressed morphologically by the person markers. The following is a survey of some of the most important syntactic features of Solon, as classified according to sentence types: •

•

•

Equative sentences with nominal predicates do not require a copula in the unmarked present tense, e.g. (examples from Tsumagari 2009) bii malsɪn ‘I am a cattle herder’ (1SG cattle.herder), eri juu arʊkkʊn ‘this house is clean’ (this house clean). For other temporal and/or modal contents, the corresponding forms of the copula bi- are used, e.g. bii malsɪn bi-su-u ‘I was a cattle herder’ (1SG cattle.herder be-PST-1SG), eri juu arʊkkʊn bi-see ‘this house was clean’ (this house clean be-PST). If comparison is involved, the point of reference is marked by the ablative case, e.g. HC eri beye axɪn-dʊxɪ gʊdda-sɪla ‘he is a bit taller than (his) elder brother’ (that person elder. brother tall-DX). Existential sentences require the presence of the copula bi- also in the present tense. Existential location is normally expressed by the dative case, e.g. KT tayyaa uriilen-du ɪlan juu bi-si-n ‘there are three houses in that village’ (that village-DAT three house be-AOR-VX3), The same construction can be used to express possession, but possession can also be expressed by the proprietive suffix attached to nominals. Interestingly, as in several Mongolian varieties, a pleonastic construction involving both the dative case and the proprietive suffix is also attested, e.g. KT tayyaa bey-du emun ukkeexen juur ʊnaajɪ-sɪ ‘that person has one son and two daughters’ (that person-DAT one son two daughter-PROPR). Negative sentences use three different strategies depending on whether the sentence is equative, existential/possessive, or verbal. Nominal predicates are negated by the

228 Sangyub Baek

•

•

•

element ɵntɵ ~ ɵntu ‘other, not the one’, e.g. KT bii een-cin ɵntɵ ‘I am not a doctor’ (1SG medicine-DX NEG.COP), D eri biteg min-i ɵntu ‘this book is not mine’ (this book 1SG-GEN NEG.COP). Existential and possessive sentences employ the negative existential noun aasɪn ‘absent, absence’, e.g. HC doo jaxxa teggu-ye aasɪn ‘there is no road to the river’ (river until road-PART NEG.EXIST). As in Siberian Ewenki, existential negation requires the use of the partitive case, cf. also e.g. HC tan-nagan eru sana[a]-ya aasɪn ‘there is no such bad thought’ (that-DX bad thoughtPART NEG.EXIST). Verbal negation takes place with the negation verb e- : AOR e-si-, which has a full verbal paradigm and is combined with the connegative form of the main verb, e-si-m eme-re ‘I do not come’ (NEG-AOR-1SG come-CONNEG), P e-jixi-ni jek-te ‘as soon as (he) had not eaten’ (NEG-CV.IMMED-PX3 eat-CONNEG), also in more complex constructions, e.g. KT togo-jɪ ugii-m e-si-n oo-do ‘one should not play with fire’ (fire-INSTR play-CV.CONN NEG-AOR-VX3 become-CONNEG). The prohibitive (second person imperative) form of the negation verb is e-ji, e.g. eji bodo-ro ‘do not think!’ (PROHIB think-CONNEG). Passive and causative sentences are primarily distinguished by the derivational voice markers attached to the verbal base before inflectional suffixes. The agent of passive predicates stands in the dative case, e.g. HC xʊnɪn tuulge-du jawa-wʊsaa ‘the sheep was caught by the wolf’ (sheep wolf-DAT catch-PASS-PST), KT sii aal=xa[t] bey-du sisi-wu-r bey ‘you are a person always hated by others’ (2SG when=PTCL person-DAT hate-PASS-PTCP.AOR person). In a causative construction, a non-coreferential causee stands in the accusative or instrumental cases, e.g. KT tayyaa kino[o] beye-w soƞoo-xoon-o-n ‘that movie makes people cry’ (that movie person-ACC cry-CAUS-AOR-VX3), KT sii tarɪ-jɪ biteg-wi eeri-xeen-ke ‘make him read your book!’ (2SG that-INSTR book-RX read-CAUS-IMP.2SG). A coreferential causee is marked by the reflexive suffixes, e.g. KT juur ukkeexem-bi uli-xeen-ce ‘he made his two sons go’ (two son-RX go-CAUS-PST). Interrogative sentences can be classified into polar and content questions. Polar questions are marked by the interrogative clitic =gI/I, e.g. T eri uxur aya=gɪɪ ‘is this ox good?’ (this ox good=INTERR), KT mugun ga-da-ndɪ=gɪ ‘do you need money?’ (silver take-AOR-2SG=INTERR). For content questions, interrogative pronouns and adverbs are used in the syntactic position of the constituent that is the target of the question, e.g. HC ʊxʊn-a ga-da-ndɪ ‘what do you (want to) buy?’ (whatPART take-AOR-2SG), HC ib-be ga-saa-sɪ ‘which one did you take?’ (which-ACC take-PST-2SG). Complement sentences are indicated by the quotative particle guƞken ‘saying’, which diachronically represents a grammaticalized reflex of the otherwise lost simultaneous converb in *-nAkAn from the verb gun- ‘to say’. The particle guƞken is placed after the finite forms of the predicates in subordinated questions, e.g. K sii i-lee nene-ndi guƞken min-ii exim-be=l aƞʊʊ-saa ʻmy elder sister asked: where do you go?ʼ (2SG which-LOC go-AOR.2SG QUOT 1SG-GEN elder.sister-PX1SG=PTCL ask-PST). In combination with the connective converb it is also used in constructions expressing intention or wish, e.g. T kinoo isi-mi guƞken joon-jɪ-mɪ ‘I think I will watch a movie’ (movie see-CV.CONN QUOT think-PROGR-AOR.1SG), KT mʊrɪn ʊgʊ-m guƞken e-ji-m bodo-ro ‘I do not think that I will ride a horse’ (horse ride-CV.CONN QUOT NEG-PROGR-AOR.1SG think-CONNEG).

Solon 229

•

Correlative sentences involve pairs of sentences containing an interrogative word and linked together by the aorist converb, e.g. i-lee bi-kki i-lee latta-ra-n ‘it sticks where it is’ (which-LOC be-CV.AOR which-LOC stick-AOR-VX3), awʊʊ-nɪn teree-kki awʊʊn-dʊ-n satan buu-m=iĕ ‘I will give candy to whom wins’ (who-PX3 win-CV.AOR who-DAT-PX3 candy give-AOR.1SG=EMPH), ooxɪ xɪndakkoƞ bi-kki-wi ooxɪ aya ‘the cheaper it is, the better’ (how cheap be-CV.AOR-RX how good). It has been suggested (Tsumagari) that such constructions are based on a Chinese model.

LEXICON AND LANGUAGE CONTACTS Historically, until the mid 17th century, Solon was spoken in northern Manchuria in the Zeya basin north of the Middle Amur. The symbiosis of the Solon with the Daghur was formed in this region, where also a Jurchen-Manchu-speaking population, subsequently known by the Daghur name “Jucher”, was living. In connection with the Russian expansion in Siberia and the Far East, all these populations were moved by the Manchu government to the Nonni basin in central Manchuria, where Daghur and Solon speaking groups survive until today. The modern Solon and Daghur groups in Hulun Buir were formed after another major translocation in the early part of the 18th century, followed by the formation of the diaspora group of Ongkor Solon in Jungaria in the 1760s. As a result, the Solon came to comprise three regionally separate populations, each of which has its own history of language contacts. The most important contact partner of Solon has for centuries been Daghur, which continues to be the second language of a large part of Solon speakers. The Daghur, by contrast, have traditionally not been bilingual in Solon, but, rather, in Manchu, which is why many of the apparent Manchu elements in Solon are likely to have been transmitted by Daghur. Even so, Daghur has many loanwords from Solon, including also some basic items such as degii ‘bird’. Solon has also influenced Daghur grammar: for instance, the system of two imperative sets (“proximal” and “distal”) has been copied to Daghur. Daghur lexical borrowings in Solon often contain specific Daghur features, including rhotacism and occasional matathesis, e.g. Solon sarpa ‘chopsticks’ ← Daghur sarpa < *sagpa < *sabka. It is, however, possible that Solon also participated in some of the phonetic developments of Daghur. In many cases, the historically very recent regressive assimilation trend in Hailar Solon has obscured the diagnostic features of Daghur loanwords, as in modern Hailar Solon edde- ‘to break’ < earlier Hailar Solon (Poppe) erde- ← Daghur erde- < *ebde-. The impact of Mongolian (proper) on Solon, mainly in the form of the Khorchin dialect, has grown stronger only in the Hailar region, where Solon children until recently have been divided between Mongolian and Chinese-medium schools. Most Hailar Solon are today fluent also in Mongolian. As a result, many modern cultural concepts, including also religious terms related to Buddhism, have been adopted from Mongolian, e.g. ajɪl ‘work, profession’ ← Mongolian ajɪl, sʊrgʊʊl ‘school’ ← Mongolian sʊrgʊʊl, xʊwɪlgan ‘reincarnation’ ← Mongolian xʊbilgaan. The borrowings also comprise various types of particles and auxiliary words, e.g. law ‘certainly’ ← Mongolian lab, daxɪn ‘again’ ← Mongolian CV. MOD daxi-n ‘repeatedly’, xamgɪɪn ‘most’ ← Mongolian GEN xamg-iin ‘of all’, xerbe ‘if’ ← Mongolian xerbee. Besides, some bound morphemes, such as the proprietive suffix -tie ← Mongolian -tai and the limitative clitic =l ← Mongolian =l, have also been borrowed.

230 Sangyub Baek

Of the colonial languages, Russian used to be the most important in the past, and the older cultural vocabulary contains several Russian borrowings, e.g. xiliep ‘bread’ ← Russian xleb, including items of international lexicon transmitted by Russian, e.g. masɪɪn ‘machine, car’ ← Russian mashína, xarandaa ‘pencil’ ← Russian karandásh, though it cannot be ruled out that many of these items have passed to Solon via Mongolian. During the Manchukuo period (1931–1945), Japanese was also strongly present in the region, and many Solon children were trained in Japanese schools. The possible impact of the Japanese language on Solon remains, however, unstudied. More recently, the impact of Chinese has become overwhelming, but in general the Chinese loanwords are phonetically adapted to the language, e.g. ceesee ‘toilet’ ← Chinese cesuo 廁所, maasan ‘immediately’ ← Chinese mashang 馬上, and they can also take derivational suffixes, e.g. gambɪɪ ‘pen’ ← Chinese gangbi 鋼筆 : VBLZ gambɪɪ-daa- ‘to write (with a pen)’. Some of the older Chinese loanwords, like soƞko ‘window’ ← Chinese chuanghu 窗戶, have probably been transmitted by Mongolian. For Ongkor Solon, a new Jungarian contact network was formed, involving Turkic languages, especially Kazakh, but also Daghur, Manchu, and Russian. The linguistic impact of this contact network remains to be studied. The extant information on Ongkor Solon suggests that the language was under a particularly strong impact of local Manchu (Sibe), with not only lexical, but also grammatical interference as a result. REFERENCES AND FURTHER READING Aalto, Pentti (1976–1977) ‘G. J. Ramstedts Onkor-solonisches Wörterverzeichnis’, Rocznik Orientalistyczny 38: 31–41, 39 (1): 55–84, Warszawa. Atknin, V. D. [В. Д. Аткнин] (1986). ‘Семантика солонского деепричастия на -tani’ [The semantics of the Solon converb in -tani], Актуальные вопросы языков народностей Севера [Current problems of the languages of the peoples of the North], 104–108, Якутск [Yakutsk]: Якутский филиал СО АН СССР. Baek, Sangyub 白尚燁 (2012a)「ツングース諸語のWH相関構文の分布に対する類 型的考察」[A typological contemplation on the distribution of WH correlatives in the Tungusic languages],『北方言語研究』Northern Language Studies 2: 163–181, 札 幌 [Sapporo]. Baek, Sangyub (2012b) ‘An areal-typological view on the distribution of WH correlatives in the Tungusic languages’, Altai Hakpo『알타이학보』22: 29–62, Seoul. Baek, Sangyub 白尚燁 (2014)「ツングース諸語における副動詞語尾-miの相違に ついて」[Functional differences of the converb in -mi in Tungusic],『北方言語研 究』 Northern Language Studies 4: 85–110, 札幌 [Sapporo]. Baek, Sangyub 白尚燁 (2015)「ツングース諸語における条件文の相違について」 [Differences of conditionals in the Tungusic languages],『北方言語研究』Northern Language Studies 5: 129–152, 札幌 [Sapporo]. Baek, Sangyub 白尚燁 (2016)「地域類型論的観点から見たツングース諸語の定動 詞における3人称標示: 数の対立を中心に [Third person marking on finite indicative forms from the perspective of areal typology],『北方言語研究』Northern Language Studies 6: 53–71, 札幌 [Sapporo]. Baek, Sangyub 白尚燁 (2017)「地域言語学的観点から見たツングース諸語の補助 動詞」[Auxiliary verbs in Tungusic from the perspective of areal linguistics],『北方 言語研究』Northern Language Studies 8: 59–79, 札幌 [Sapporo]. Bai Lan & Juha Janhunen (1992) ‘On the present state of the Ongkor Solon’, Journal de la Société Finno-Ougrienne 84: 225–229, Helsinki.

Solon 231

Chaoke 朝克 (1991)『エウンキ語基礎語彙集』 [Ewenke basic vocabulary], 東京 [Tokyo]: 東京外国語大学アジア・アフリカ言語文化研究所. Chaoke 朝克 (1995)『鄂温克语研究』 [Researches on the Ewenke language], 北京 [Peking]: 民族出版社. Chaoke 朝克 (2003)『エウェンキ語形態音韻論および名詞形態論』[Morphophonemics and noun morphology of the Ewenke language], 東京 [Tokyo]: 東京外国語大学 アジア・アフリカ言語文化研究所. Chaoke 朝克 (2009)『鄂温克语参考语法』[Reference grammar of the Ewenke Language], 北京 [Peking]: 中国社会科学出版社. Chaoke 朝克 (2017a)『鄂温克语动词形态论』 [Verb morphology of the Ewenke Language], 北京 [Peking]: 中国社会科学出版社. Chaoke 朝克 (2017b)『鄂温克语名词形态论』 [Noun morphology of the Ewenke Language], 北京 [Peking]: 中国社会科学出版社. Chaoke 朝克 et al. (1988)『鄂温克民间故事』[Ewenke folktales], 海拉尔 [Hailar]: 内 蒙古文化出版社. Chaoke 朝克 & Nakajima Motoki 中島幹起 (2005) 『エウェンキ語への招待』 [Invitation into Ewenki language], 東京 [Tokyo]: 大学書林. Chaoke 朝克 & Tsumagari Toshirō 津曲敏郎 (1995)『鄂温克語三方言対照基礎語彙 集』 [A basic vocabulary of three Ewenki (Ewenke) dialects], 小樽 [Otaru]: 小樽商 科大学言語センター. Chaoke 朝克 & Tsumagari Toshirō 津曲敏郎 & Kazama Shinjiro 風間伸次郎 (1991),『 ソロン語基本例文集』[Solon basic sentences], 札幌 [Sapporo]: 北海道大学文 学部. Cincius, V. I. [В. И. Цинциус] (1997) ‘Солонский язык’ [The Solon language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 226–236, Москва [Moscow]: “Индрик”. Dao’erji, Du. [Dular Dorji] 杜・道尔基 (1998)『鄂汉词典』 [An Ewenke-Chinese dictionary], 海拉尔 [Hailar]: 内蒙古文化出版社. Dao’erji, Du. [Dular Dorji] 杜・道尔基 (2014)『鄂蒙词典』[An Ewenke-Mongolian dictionary], 北京 [Peking]: 民族出版社. Hu, Zengyi 胡増益 (1984)「鄂温克語概况」[A brief description of the Ewenke language],『民族语文』1: 60–74, 北京 [Peking]. Hu, Zengyi 胡増益 & Chaoke 朝克 (1986)『鄂温克語简志』[A sketch of the Ewenke language],『中國少數民族語言簡志叢書』[Languages of the minority nationalities of China], 北京 [Peking]: 民族出版社. Ivanovskii, A. O. [А. О. Ивановскiй] (1894) Mandjurica I: Образцы солонскaго и дахурскaго языковъ [Manjurica I: Specimens of the Solon and Daghur languages], С.-Петербургъ [St. Petersburg]: Типографiя Императорской Академiи наукъ. Reprint: Debter—Deb-ther—Debtelin: Materials for Central Asiatic and Altaic Studies 2, Budapest: Akadémiai Kiadó (1982). Kałużyński, Stanisław (1971) ‘Solonisches Wörterverzeichnis nach F. V. Muromskis handschriftlichen Sprachaufzeichnungen’, Rocznik Orientalistyczny 34 (1): 39–77, 34 (2): 15–56, Warszawa. Kamimakise, Saburō 上牧瀬三郎 (1940)『ソロン族の社会』[The society of the Solon people], 東京 [Tokyo]: 生活社. Kazama, Shinjirō 風間伸次郎 (1997)「ツングース諸語における「部分格」[“Partitive” in the Tungusic languages],『環北太平洋の言語』Languages of the North Pacific Rim 3: 103–120, 札幌 [Sapporo].

232 Sangyub Baek

Kazama, Shinjirō 風間伸次郎 (2005)「ソロン語口語コーパスとその分析」 [Corpus of the Solon language and its analysis],『言語情報学研究報告』8: 11–43, 東 京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2010)「ソロン語におけるモンゴル語の影響 ― 言 語接触の一事例として」[Mongolian influence on the Solon language],『語学教育 フォーラム』24: 163–183, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2013)「ソロン語のアスペクト、ボイス、モダリテ ィ」 [Aspect, voice, and modality in Solon],『語学研究所論集』18: 409–423, 東 京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2014)「《データ：他動性》ソロン語」 [Transitivity in Solon], 『語学研究所論集』19: 331–339, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2015) 「《データ：（連用修飾的）複文》ソロ ン語」 [Complex sentences in Solon],『語学研究所論集』20: 215–224, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2016)「《データ：情報構造と名詞述語文》ソロン 語」 [Information structure and nominal predicate sentences in Solon],『語学研究所 論集』 21: 249–257, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2018a)「ソロン語」[Solon language], in: Li Linjing 李林静 & Yamakoshi Yasuhiro 山越康裕 & Kogura Norikazu 児倉徳和 (eds.),『 中国北方危機言語のドキュメンテーション』, 101–159, 東京 [Tokyo]: 三元 社. Kazama, Shinjirō 風間伸次郎 (2018b)『ソロンの文化と生活 1 A』 [Solon culture and living 1 A], 府中 [Fuchu]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 (2020)『ソロンの文化と生活 2』[Solon culture and living 2], 府中 [Fuchu]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 (2022)『ソロンの文化と生活 3』[Solon culture and living 3], 府中 [Fuchu]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 & Tuoya トヤー (2007) 『ソロンの民話と伝説1』 [Solon folk tales and legends 1], 札幌 [Sapporo]: 北海道大学大学院文学研究科. Kazama, Shinjirō 風間伸次郎 & Tuoya トヤー (2008) 『ソロンの民話と伝説 2』 [Solon folktales and legends 2], 東京 [Tokyo]: 東京外国語大学アジア・アフリカ 言語文化研究所. Kazama, Shinjirō 風間伸次郎 & Tuoyaトヤー (2011)『ソロン語基礎語彙』[Basic vocabulary of Solon], 東京 [Tokyo]: 東京外国語大学アジア・アフリカ言語文化 研究所. Khabtagaeva, Bayarma (2012) ‘Dagur elements in Solon Evenki’, Acta Orientalia Academiae Scientiarum Hungaricae 65 (3): 335–346, Budapest. Kim Juwon 金周源 (1986)「퉁구스제어」[Tungusic languages], in: Kim Bang-han 金芳漢 & Kim Juwon 金周源 & Chong Chemun 鄭堤文,『몽골어와 퉁구스어』 [Mongolic and Tungusic languages]: 149–279, 서울 [Seoul]: 민음사. Lie, Hiu (1978) ‘Solonisches material aus dem Huin-gol’, in: Gerhard Doerfer & Michael Weiers (eds.), Beiträge zur nordasiatischen Kulturgeschichte, Tungusica 1: 126–178, Wiesbaden: Otto Harrassowitz. Poppe, N. N. [Н. Н. Поппе] (1931) Материалы по солонскому языку [Materials on the Solon language], Материалы Комиссии по исследованию Монгольской и Тувинской Народных Республик и Бурят-монгольской АССС 14, Ленинград: Издательство Академии наук СССР.

Solon 233

Tsumagari, Toshirō 津曲敏郎 (1989)「ソロン語」[Solon language]. in: Kamei Takashi 亀井孝 & Kōno Rokurō 河野六郎 & Chino Eiichi 千野栄一 (eds.),『言語学大辞 典』2: 522–523, 東京 [Tokyo]: 三省堂. Tsumagari, Toshirō 津曲敏郎 (ed.) (1993)『朝克著『エウェンキ語基礎語彙集』索 引』[A Solon (Ewenke) Index to the Classified Vocabulary by D. O. Chaoke, with English Equivalents], 小樽 [Otaru]: 小樽商科大学言語センター. Tsumagari, Toshirō 津曲敏郎 (1996)「中国・ロシアのツングース諸語」[The Tungusic languages in China and Russia],『言語研究』Journal of the Linguistic Society of Japan 110: 177–191, 京都 [Kyoto]. Tsumagari, Toshirō (1997) ‘Linguistic diversity and national borders of Tungusic’, in: Hiroshi Shoji & Juha Janhunen (eds.), Northern Minority Languages: Problems of Survival, Senri Ethnological Studies 44: 175–186, Osaka: National Museum of Ethnology. Tsumagari, Toshirō (2009) ‘A sketch of Solon grammar’ [ソロン語の文法概略],『北方 人文研究』Journal of the Center for Northern Humanities 2: 1–21, 札幌 [Sapporo].

CHAPTER 10

NEGHIDAL Sofia Oskolskaya

Neghidal (Negidal) is a critically endangered Tungusic language of the Ewenic group, closely related to Siberian Ewenki. The speakers of Neghidal, today officially recognized as a distinct ethnic group (Russian SG MASC negidálec : PL negidál’cy), inhabit the Amgun’ river basin in Khabarovsk Krai of the Russian Far East. In the 17th century the Neghidal population is estimated to have comprised no more than 390 individuals (Dolgikh 1960). In the middle of the 19th century the population was 331 (Boshnyak 1859), while according to the 1897 Russian census it was 399 (Patkanov 1912). In the early Soviet period (1926–1927) K. M. Myl’nikova and V. I. Cincius counted 371 individuals, after which the population has gradually grown to its current size around 500. The official number according to the census of 2010 was 513 individuals. Thus, although the Neghidal population is rather small, it has been stable for the last several centuries. On the basis of a number of ethnic and linguistic differences the Neghidal can be divided into two groups, known as the Upper and Lower Neghidal. The Upper Neghidal inhabit the Middle Amgun’ basin (the villages of Vladimirovka and im. Poliny Osipenko in the district of im. Poliny Osipenko), while the Lower Neghidal live in the Lower Amgun’ basin and in the adjoining Lower Amur region (the villages of Tyr and Beloglinka of the Ulcha district, and the city of Nikolayevsk-on-Amur). The dialect of the Lower Neghidal is recently extinct, and the speakers of the Upper Neghidal dialect are also very few, possibly only three (2020). Linguistically, the Upper Neghidal dialect is closer to Siberian Ewenki than the Lower Neghidal dialect. Neghidal does not have a literary language, nor an official writing system, and there are no classes of or in Neghidal at school. In spite of historical multilingualism in the region, the dominant language of all Neghidal today is Russian. The name Neghidal is of Ewenki origin and is based on Ewenki ƞiegiidaal = ƞiegii-daa-l, from ƞie ‘lower part, hill foot, river bank’ + the local derivative complex -giidaa- + the plural suffix -l, i.e., ‘those living further down’. This name was applied by the taiga-dwelling nomadic Ewenki to the river-oriented and more settled Neghidal. The Neghidal themselves have several descriptive self-appellations, including naa beye-nin ~ naa bee-n : PL naa beye-sel-tin ~ naa bee-sel-tin ‘people of (the local) land’ (land person-PX3SG : -PL-PX3PL), ɪlkan beye-nin ‘real people’ (real person-PX3SG), and emƞun beye-nin ‘Amgun’ people’ (Amgun’ person-PX3SG). The Russian inhabitants of the Lower Amur basin call the Neghidal by the name gilyák : PL gilyakí, i.e., “Ghilyak”. The confusion is due to the fact that a considerable part of the Neghidal, especially the Lower Neghidal group, are actually former Ghilyak (Nivkh) speakers who shifted over to an Ewenic language. Culturally, the Neghidal are very close to the neighbouring Amur Ghilyak (Nivkh). DOI: 10.4324/9781315728391-10

Neghidal 235

DATA AND SOURCES The first information about the inhabitants of the Amgun’ basin reached Russia in the 17th century, and the term “Neghidal” was also used quite early in Russian documents (Vdovin 1953), though the Neghidal territory was incorporated in the Russian Empire only in 1858. The Baltic German naturalist Alexander Theodor von Middendorff (1847–1875) was the first scientist to visit the Lower Amur region during his Siberian expedition 1843–1844, and he was also the first to single out the Neghidal as an ethnic group different from other Tungusic tribes. His works do not contain much information on the Neghidal language, however, only a list of words, but no linguistic description. Another Baltic German naturalist Leopold von Schrenck undertook an expedition to the Amur region in 1853–1856. In his publications (Schrenck 1881, in Russian 1883) he gave a short description of the Neghidal. He concluded that the Neghidal are ethnically and genetically related to the (Nanai-speaking) Samagir, but linguistically closer to the Ulcha. The first professional linguistic study of Neghidal was published by Peter Schmidt (1923), based on materials collected upon his request by Benedek Baráthosi-Balogh, K. D. Loginovskii, and P. Krastin. Schmidt’s study contains comments on lexicon and phonetics, as well as a Neghidal-English vocabulary of c. 1200 words. The first grammatical notes on Neghidal were published soon afterwards by Leo (Lev) Sternberg, whose ethnographic study (Sternberg 1933: 529–552) contains a few simple sentences and morphological paradigms, as well as an ethnographic vocabulary of c. 600 words. In the 1920s and 1930s, many scholars visited the Amgun’ region. Linguistic work was initiated by K. M. Myl’nikova and V. I. Cincius, who collected also material on ethnography and folklore. Some results of their research were published in an article (Myl’nikova & Cincius 1931), which contains, among other things, information on Neghidal clans, as well as a grammatical sketch of the Neghidal language. The latter was the first reasonably detailed description of Neghidal grammar, and the authors pay also attention to dialectal differences and to the comparison of Neghidal with the other Tungusic languages. Myl’nikova and Cincius were the first who distinguished the Upper and Lower groups of the Neghidal and established the corresponding dialectal difference. V. D. Kolesnikova and O. A. Konstantinova visited the Lower Neghidal in 1961. Their material was included in the “Comparative dictionary of the Tungusic languages” published under the general editorship of V. I. Cincius (1975–1977). They also authored a grammatical sketch (Kolesnikova & Konstantinova 1968), which contains information on Neghidal phonetics and grammar, as well as a short sample text. Cincius also continued research on the Neghidal language and culture. Among her results is an article on Neghidal riddles (1957), a posthumous grammatical sketch (1997), and a monograph titled “the Neghidal language” (1982), which still remains one of the most profound works on Neghidal. It includes an introduction with sociolinguistic and ethnographic information, a sketch of Neghidal phonetics and grammar, 40 texts in Neghidal with translation into Russian and meta-information, and a Neghidal-Russian vocabulary. During the period 1981–2000 Neghidal was intensively documented and studied by M. M. Khasanova and A. M. Pevnov. They published the results of their work in several articles (Pevnov & Khasanova 1994, 2006, Pevnov 1998, 2007) and in a book (Khasanova & Pevnov 2003). The latter is a collection of Neghidal myths and mythological stories. The texts are presented in both dialects of Neghidal, accompanied by meta-information and a

236 Sofia Oskolskaya

translation into Russian. Notes on the texts, summaries of the plots in English, as well as information about the Neghidal people, folklore and language (especially morphological characteristics) in Russian and English, are also provided. Shinjirō Kazama made linguistic field trips to the Lower Neghidal in 1994 and 1998 and published a collection of his field data in a book (Kazama 2002) which contains five texts in Neghidal with word-by-word translation into Japanese, summaries of the plots in Japanese, English and Russian, a list of basic sentences elicitated from the informants (the sentences are given in Russian with a translation into Neghidal and Japanese), and a general outline of Neghidal phonetics and grammar. Linguists who have more recently visited the area of the Neghidal include E. Yu. Kalinina, V. Yu. Gusev, and S. Yu. Toldova. As a result, E. Yu. Kalinina (2008) published an article devoted to a specific phonetic problem concerning vowel harmony in Neghidal. Her data are based on material from the Upper Neghidal dialect, because, as it turned out, no speakers of Lower Neghidal could be located. The information concerning the recent extinction of the Lower Neghidal dialect was confirmed by her colleagues, who visited the Neghidal villages on the Lower Amur in 2005 without finding any surviving speakers. Since 2012, another documentation project on Neghidal has been active with the support of the Endangered Languages Documentation Programme (ELDP), as reported by Brigitte Pakendorf and Natalia Aralova (2018). The present chapter is entirely based on the above-mentioned published works, texts and dictionaries, especially Schmidt (1923), Myl’nikova & Cincius (1931), Kolesnikova & Konstantinova (1968), Cincius (1982, 1997), Kazama (2002), Khasanova & Pevnov (2003), Pevnov & Khasanova (2006), and Kalinina (2008). SEGMENTAL STRUCTURE The Neghidal vowel system is not dialectally uniform. The original system, somewhat better preserved in the now extinct Lower dialect, seems to have comprised 8 short vowels, which correspond to the Proto-Tungusic system, as also observed in a closely similar form in Ewen, Solon, and Orochen. The vowels, here written as u ʊ ɵ o i ɪ e a, may be divided into four pairs: the high rounded back vowels u ʊ, the mid to low rounded back vowels ɵ o, the high unrounded front vowels i ɪ, and the mid to low unrounded central vowels e a. In each pair, one member is higher than the other, meaning that the vowels can also be divided into a higher row comprising the segments u ɵ i e and a lower row comprising the segments ʊ o ɪ a (Table 10.1). The vowels of the higher row are originally palatal (*ü *ö *i *e), while the vowels of the lower row are originally velar (*u *o *ï *a). Phonetically, this means that the unrounded back vowel e of the higher row is pronounced with the quality [ə], as in most other Tungusic languages, though after the palatal consonants c j ñ y it can be realized with the slightly higher quality [ɘ], e.g. ceƞgele [ʧɘŋgələ] ‘ball (as of yarn)’. TABLE 10.1 NEGHIDAL SHORT VOWELS u

 

ʊ

i ɪ

ɵ

e

o

a

Neghidal 237

The synchronic status of the vowels ʊ [ʊ] (< *u) and ɪ [ɪ ~ e] (< *i) is under question, and at least in the Upper dialect they seem to be absent as distinct segments, meaning that the vowel system in the Upper dialect has been reduced to comprise only the 6 segments u ɵ o i e a. However, items containing an original *u often show o in both dialects, as in oja ‘trace’ (< *uja), or also a variation between u and o, as in toñƞa ~ tuñƞa ‘five’ (< * tuñƞa), suggesting that the vowel may actually be ʊ, or, at least, that it retained the distinct quality of /ʊ/ until very recently. The variation may also be due to the deterioration of the language skills of the last speakers. The situation is similar for original *ï, which is reflected in a variation between [i] and [e], as in dil [dil] ~ [del] ‘head’ (< *dïlï), nikan [nikan] ~ [nekan] ‘Chinese’ (< *nïkan), suggesting that the vowel is still, or was until recently, a distinctive /ɪ/. The status of the vowel ɵ [o] (< *ö), in opposition to u [u] (< *ü) and o [ɔ] (< *o), is also problematic. The normal representation of original *ö seems to be u in both dialects, as in (Upper) lukii ~ (Lower) luxii ‘arrow’ (< *lökii), but there are also cases of variation between o and u, as in oko- ~ uku- ‘to suck’ (< *ökö-), suggesting that /ɵ/ is still, or was until recently, a distinct segment. On the other hand, there is a tendency of rounding the vowel e [ə] towards the quality [ɵ] after labial consonants, as in [bɵjə] ‘man, person’ (< *beye), which points to the possibility that [ɵ] is merely a positional variant of e. Recent research (Kalinina) suggests that the two vowels may contrast in some contexts, as in [ɵlɵʧɵɵs] ‘you are scared’ vs. [ələʧɵɵs] ‘you gorged yourself’, but the data is controversial and can be disputed, and it is likely that at least some varieties of Neghidal have only 5 short vowels: u o i e a, a system similar to Siberian Ewenki. Like other forms of Ewenic, Neghidal retains well the Proto-Tungusic quantitative correlation of vowels. The system of long vowels is basically identical with that of the short vowels, but with the addition of the mid-high unrounded front vowel quality [e:], here written as ie, which represents the Proto-Tungusic diphthongoid sequence *ïa, as in sien [se:n] ‘ear’ (< *sïan). The resulting system of 9 long vowels (Table 10.2) is, however, only theoretical, for it seems that several distinctions have been eliminated in the modern dialects, though probably very recently. Thus, the original *uu seems to be represented systematically as oo, as in joo ‘house’ (< *juu/g), while the original *öö is represented as uu, as in guun- ‘to say’ (< *göön-). Since there is no opposition between ɵɵ and ee, the latter vowel can take the more rounded value [ɵ:], but will here be written phonemically as ee. Also, there seem to be no cases of a distinctive ɪɪ, which means that the actual system of long vowels in both dialects probably comprises only 6 entities: uu oo aa ee ie ii. Qualitatively, this system is identical with that attested in the neighbouring related (Nanai) and unrelated (Ghilyak) languages. The Neghidal system of short vowels was probably also developing in this direction, with ɪ [ɪ ~ e] functioning as the short counterpart of the long ie [e:], but the ongoing processes of language obsolescence left this development uncompleted. TABLE 10.2 NEGHIDAL LONG VOWELS uu

 

ii

ɵɵ

ee

ie

oo

aa

ʊʊ

ɪɪ

The Neghidal consonant system comprises 18 phonemes (Table 10.3), which, according to the place of articulation are: the labials m b p w, the dentals n d t s l r, the palatals ñ

238 Sofia Oskolskaya

j c y, and the velars ƞ g k x. By the manner of articulation the system comprises: the nasals m n ñ ƞ, the weak (voiced) stops b d j k, the strong (voiceless) stops p t c k, the (voiceless) fricatives s x, the glides w y, the lateral l, and the rhotic r. The phonetic realizations of the consonants, as well as the system as a whole, are very similar to those attested in the other Ewenic languages. However, especially in the Lower dialect, the velar obstruents g k x can (could) have the rather pronounced back-velar or uvular allophones [ɢ q χ] in words containing original back vowels. Some data suggest that these allophones may have become phonemic in a few items, as in etiyeqqen ‘old man’. Also, the weak velar stop g can have the fricative realizations [ʁ ɣ] in intervocalic position. TABLE 10.3 NEGHIDAL CONSONANTS m

n

ñ

ƞ

 

b

d

j

g

[ɢ]

p

t

c

k

[q]

x

[χ]

s w

y l r

Unlike the situation in most other Tungusic languages, the rhotic r, realized as a trill, is a marginal phoneme, occurring mostly in recent loanwords or ideophones. This is because the original Proto-Tungusic *r was systematically eliminated in Neghidal. The principal development was *r > y, as in *oron ‘reindeer’ > oyon, but in consonant clusters *r could yield also other results, as in *gerbi ‘name’ > gelbi, *örke ‘door’ > (Upper) [uʧkə] ~ (Lower) uyke. Synchronically, r can appear as a variant pronunciation of l in the cluster lk ~ rk, at least in the item elkejin ~ erkejin ‘slowly’. For unknown reasons, r is also present in the native Neghidal word dirkeen ~ dɪrkaan ‘fly’. The velar fricative x, which may also have the laryngeal realization [h], represents in initial position the secondary Ewenic *x (< *p), as in xemun ‘lip/s’ (< *pemön). Unlike Ewen and most dialects of Ewenki and Orochen, Neghidal shows no tendency of debuccalizing *s to x. However, the Lower dialect normally fricativizes postvocalic *k to x, as in (Upper) akin ‘elder brother’ : PL ak-nii-l vs. (Lower) axin : PL ax-nii-l. The latter development, which in intervocalic position has a parallel in Orochen, may probably be viewed as a phonetic tendency with not necessarily a fully phonemic status. As in Siberian Ewenki, the weak and strong series of stops p t k vs. b d g in Neghidal do not contrast in syllable-final position. Therefore, assuming that it is the weak series that is less marked, all syllable-final stops could also be understood as “archiphonemic” b d g, e.g. [ogda] ‘boat’ = /ogda/ vs. [okto] ‘medicine’ = /ogto/. The palatal stops c j do not originally occur in syllable final positions, but secondary examples like *örke ‘door’ > [uʧkə] could, then, also be phonemized accordingly as /ujke/. For reasons of consistency and phonetic transparence this interpretation will not be followed here. PHONOTACTICS AND MORPHOPHONOLOGY Syllable types in Neghidal include: V (e-din ‘wind’), VV (aa-mi ‘sleep’), VC (in-me ‘needle’), VVC (iil ‘bowstring’), CV (di-gin ‘four’), CVV (moo ‘tree’), CVC (det ‘swamp’), CVVC (ƞies-kii ‘downhill’). No more than one consonant can occur in word-initial and

Neghidal 239

word-final position. In medial position clusters of two consonants are permitted. Suffixes with an initial cluster have allomorphs that allow them to be attached to a stem-final consonant, e.g. joo ‘house’ : DIR joo-tkii : PL joo-l : PL-DIR joo-l-tixii. All consonants except r can occur in word-initial position, while the final position is compatible only with the stops p t k, the fricatives s x (the latter in the Lower dialect instead of k), the glides w y, the lateral l, the nasals m n ƞ (but not ñ), including the primary stem-final nasal /n (< *-n). Some suffix-initial consonants undergo assimilations and adaptations, many of which are morpheme-specific. For instance, a suffix-initial w, as in the accusative marker -wA, is represented variously as -pA (after voiceless obstruents), -mA (after nasals), and -bA (facultatively after the lateral l), e.g. ACC joo-wa ‘house’, is-pa ‘larchtree’, edin-me ‘wind’, PL joo-l-ba ~ joo-l-wa ‘houses’. This alternation is partly connected with the ambiguous relationship between the segments b and w in various positions, as also known from other Ewenic languages. Also, as in other Ewenic languages, the initial consonant of the aorist marker -yA- (< *-rA) is progressively assimilated by a preceding stem-final consonant, e.g. ɪl- ‘to stand up’ : AOR.3PL ɪl-la. In the Lower dialect the sequence -tki can undergo progressive assimilation to -tci (= -cci), e.g. joo ‘house’ : DIR joo-tkii > joo-tcii. Vowel length is in Neghidal, as in the other Ewenic languages, somewhat unstable in non-initial syllables. Although long vowels have a phonemic status in all syllables, they are often realized as short in non-initial syllables, though they may nevertheless leave a prosodic trace and are often perceived as manifestations of a stress accent, which itself is non-distinctive in Neghidal. In the present description, vowel length is indicated as it is recorded in the sources, which means that a single word or suffixal morpheme can occasionally show different vowel lengths in different contexts. The reasons underlying this variation may be both articulational and perceptional, and they may also involve phonological and lexical factors, in that some long vowels, or the long vowels of some lexical items, may be more liable to undergo shortening than others. The distribution of vowels within words is governed by the rules of vowel harmony. The vowels are divided into three harmonic groups on the basis of height: the higher (originally palatal) vowels e ee ~ ɵɵ and possibly ɵ (here written as e ee), the lower (originally velar) vowels a aa o oo ie and possibly ɪ ʊ ʊʊ (if distinct from i u uu), and the neutral vowels i ii u uu (assuming that they do not contrast with ɪ ɪɪ ʊ ʊʊ). A word can contain only higher vowels, e.g. beye ‘man, person’ : ACC beye-we, or only lower vowels, e.g. joo ‘house’ : ACC joo-wa. Neutral vowels can be combined with both higher and lower vowels, but they are regressively lowered when preceding a syllable with a low vowel, a tendency that can occasionally produce a paradigmatic alternation within a single morpheme, though only at the phonetic level, e.g. is [is] ‘larchtree’ : ACC is-pa [ɪspa]. Labial harmony, that is, the progressive assimilation of a aa to o oo in non-initial syllables under the impact of a short o of the initial syllable, is not productive in Neghidal, though there are some apparently lexicalized stems that follow labial harmony and even require it in suffixes, e.g. oyon ‘reindeer’ (< *oron < *oran) : ACC oyon-mo. At the same time, there are examples of stems that synchronically break the rules of labial harmony. Although labial harmony may altogether be analysed as a non-phonemic phenomenon also in other Ewenic idioms, its disruption in Neghidal is probably connected with the neutralization of the distinction between *u and *o, for which reason the sequence o-a (< *u-a) is synchronically fully possible, as in oman ‘marrow’ (< *uman). At morpheme boundaries a connective vowel, normally i (after labials u), is added after stem-final consonants to avoid complex clusters, but also with no obvious

240 Sofia Oskolskaya

phonotactic reason in some morphological contexts, e.g. laaƞ ‘trap’ : DIR laaƞ.i-tkii : PX1SG laaƞ.i-w : PX3PL laaƞ.i-tin. An i or ii can also be added after a final consonant of a word if the following word begins with a consonant, e.g. bii sin xute-s bi-si-m kuuyanu > bii sini xute-sii bi-si-mi kuuyanu ‘I am your son Kuuyanu’ (1SG 2SG.GEN childPX2SG be-AOR-1SG Kuuyanu) (Pevnov & Khasanova 2006). The exact background of this phenomenon remains to be clarified, but it may also be diachronically conditioned. Another sandhi phenomenon at word boundaries concerns the deletion of the first member of a vowel cluster, e.g. amaxa ule-wee-n > amax_ule-wee-n ‘bear meat’ (bear meat-ACC-PX3SG). WORD FORMATION Like other Ewenic languages, Neghidal makes a morphological difference between two main parts of speech, nominals and verb(al)s. Nominals comprise also adjectives, numerals and pronouns. All morphology is based on suffixation of a relatively transparent agglutinative type. Suffixes containing harmonically active vowels follow the rules of vowel harmony: A = a ~ o [ɔ] ~ e [ə ~ ɵ], AA = aa ~ oo [ɔ:] ~ ee [ɵ:]. Derivational suffixes serve, on the one hand, to modify the semantics of the underlying base lexeme, and, on the other, to convert words from one part of speech to another, as listed below. (1a) Denominal nouns: -xAAyAA, -njA [objects of large size], e.g. beye ‘man, person’ : beye-xeeyee, beye-nje ‘big man’; ‑xAAn ~ ‑kAAn, ‑xkAAn ~ ‑kkAAn [objects of small size], e.g. bɪya ‘river’ : bɪya-xkaan ‘small river’; ‑xsA ~ -ksA [animal skin], e.g. monoxaan ‘hare’ : monoxaa-xsa ‘hare skin’; ‑g [groups of homogeneous objects], e.g. jagda ‘pine tree’ : jagda-g ‘pine forest’; ‑(n)kAAn [inhabitants of a place], e.g. emƞun ‘Amgun’ river’ : emƞu-nkeen ‘an inhabitant of the Amgun’ river basin’; ‑pun ~ ‑ptun [objects placed on a surface], e.g. sitkii ‘wall’ : sitkii-pun ‘wall carpet’: -mAAn [keen on something], e.g. ayaxii ‘liquor’ : ayaxii-maan ‘drunkard’. Another (de)nominal suffix is -ktA for collective nouns, including items of vegetation, insects, body parts, etc., but normally lacking a synchronically productive base, e.g. xoko-kto ‘honeyberry’, ƞanma-kta ‘mosquitoes’. (1b) Denominal adjectives: ‑mA [material], e.g. sele ‘iron’ : sele-me ‘of iron’; -kto ~ -ktu [characteristic feature], e.g. kumke ‘louse’ : kumke-ktu ‘having many lice’; -tmAA [increased intensity, comparative grade] aya ‘good’ : aya-tmaa ‘better’; -lAA [decreased intensity, moderative grade], e.g. ƞonom ‘long’ : ƞonum-la ‘longish’. (2) Deverbal nouns: ‑wun [instrument or result of an action], e.g. egdi- ‘to comb one’s hair’ : egdi-wun ‘comb’; ‑ƞki [instrument or place of an action], e.g. waa- ‘to kill’ : waaƞki ‘the place where an animal is/was killed during hunting’; -xiit ~ -kiit [place of action], e.g. muulii- ‘to carry water’ : muulii-xiit ‘place from where water is taken, an ice-hole’; -ptin [type of dress or cover], e.g. uxil- ‘to wrap’ : uxil.i-ptin ‘wrap’. This group also comprises one item deriving nominals with a mainly adjectival meaning: -pti [reference to a temporal frame], e.g. sikse- : sikse-gi- ‘to grow dark’ : siksegi-pti ‘(taking place in the) evening’. (3) Denominal verbs: There are only a few stems that function as both nouns and verbs with zero derivation, e.g. sikse ‘evening’ : sikse- ‘to grow dark’. The general suffix converting nominals to verbs is -lAA-, e.g. muu ‘water’ : muu-lee- ‘to fetch water’. Instrumental verbs are formed by the suffix -dAA, e.g. asaxii ‘wing’ : INSTR asaxii-daa- ‘to flap (with) wings’.

Neghidal 241

(4) Deverbal verbs: As in other Tungusic languages, deverbal verbs are formed by several productive markers with grammaticalized functions connected with voice, mood, and aspect (including Aktionsart). •

•

•

Voice markers include: ‑w- ~ ‑mu- [passive], e.g. ice- ‘to see’ : PASS ice-w- ‘to be seen’; -p- [anticausative], e.g. emee- ‘to leave’ : ANTICAUS emee-p- ‘to stay, to remain’; -mAAc- [reciprocal], e.g. waa- ‘to kill’ : RECIPR waa-maac‘to kill each other’; -ldi- [cooperative], e.g. baxa- ‘to find’ : COOP baxa-ldi‘to meet each other’; -wkAAn- ~ -mkAAn-, -wkAAc- ~ -mkAAc- [causative], e.g. ice- ‘to see’ : CAUS ice-wkeen- ‘to show’. Aspectual markers include: -jA- [progressive or imperfective], e.g. bi- ‘to be’ : PROGR bi-je- ‘to live (for a long time)’; -c- ~ -ci- ~ -t- [resultative], e.g. tege‘to sit down’ : tege-c- ‘to sit’; -ygi- ~ -gi- ~ -dgi- [reversive or restitutive], e.g. ice- ‘to see’ : REV ice-ygi- ‘to see again’; -l- [inchoative], e.g. jep- ‘to eat’ : INCH jep.u-l- ‘to begin to eat’; -sin- [semelfactive or instantaneous], e.g. kɪk‘to bite (several times)’ : SEM kɪxi-sin- ‘to bite (once)’; -ktA- ~ -ksA- [distributive], e.g. cɪxa- ‘to cut’ : DISTR cɪxa-kta- ‘to cut several objects’; -s- ~ -si- [iterative], e.g. ITER cɪxa-s- ‘to cut something in several places’; -wAAc- ~ -pAAc- ~ -mAAc- [habitive], e.g. bi- ‘to be’ : HAB bi-weec- ‘to visit from time to time’; -nAk- [pluritative], e.g. ga- ‘to take, to buy’ : PLUR ga-nak- ‘to buy every object’. Modal markers include: -mu- : -mu-si/n- [desiderative], e.g. om- ‘to drink, to smoke’ : DESID om-musi- ‘to be eager to drink/smoke’; -tce- [volitive], e.g. jawa- ‘to take’ : VOLIT jawa-tce- ‘to want to take’; -nAA- [andative or intentive], e.g. waa- ‘to kill’ : AND waa-naa- ‘to go (with the intention) to kill’.

Many derivational markers can combine with each other, and at least some of the verbal suffixes can freely change their positions in a word without a semantic change. For instance, the reversive marker -(y)-gi- can either precede or follow the semelfactive marker -sin-, e.g. ƞenew- ‘to carry’ : REV-SEM-PST ƞenew-gi-sin-cee ‘s/he carried’ vs. giexu- ‘to go back’ : SEM-REV-PST-3PL giexu-sin-gi-caa-tin ‘they went back’. There is also a number of nominal suffixes which are highly productive and can be attached to various types of nominals, including pronouns, but, which, nevertheless remain outside of the basic categories of inflectional morphology. Thus, they have an intermediate status between derivation and inflection. They include, in particular, the proprietive (comitative) markers -lkAAn, -cii, -lgAli, -nAAn, -nA, e.g. asii ‘woman’ : PROPR asii-lkaan ‘having a wife, married’; the similative marker -gAcin ~ -kAcin ~ ‑ƞAcin, e.g. beye ‘man, person’ : SIM beye-gecin ‘like a man’; the contrastive (emphatic) marker -nmAA ~ -tmAA, e.g. axi ‘elder brother’ : CONTR axi-nmaa ‘the elder brother (as opposed to the younger brother)’; and the limitative marker -gdA, e.g. dɪl ‘head’ : LIM dɪli-gda ‘only a head’. This group also comprises the marker of alienable possession ‑ƞ-, e.g. dɪli-ƞ- ‘(loose) head (belonging to somebody else, e.g. a killed animal)’, as well as the suffix ‑gida ~ ‑y(i)da ~ ‑ƞida ~ ‑lida, which indicates spatial proximity, e.g. boga ‘place’ : boga-yda ‘neighbourhood, adjacency’. In the nominal morpheme string, all suffixes of this group occupy a position before the markers of number, case, and possession. Although compounding is uncharacteristic of Neghidal, there are some examples of conventionalized nominal phrases like asii+xute ‘woman+child’ > ‘daughter’. Some of these phrases may also be copies from the Amur Tungusic languages.

242 Sofia Oskolskaya

NUMBER AND CASE As in the other Ewenic languages, the nominal stem in Neghidal takes inflectional suffixes in the order number—case—possession (-NX-CX-PX), e.g. iesa ‘eye’ : PL-ACCPX1SG iesa-l-ba-w. Any derivational suffixes precede number marking. Plural marking in Neghidal is somewhat less obligatory than in Ewenki, for which reason the unmarked singular form can occasionally also have a plural reference, even when combined with a plural form of the verb, as in boga beye-nin bi-jee-tin ‘maybe they were sky people’ (sky person-PX3SG be-FUT-3PL). Otherwise, plural is marked by the Common Ewenic suffixes -l, e.g. sagdi ‘old (man)’ : PL sagdi-l ‘old men’ and -sAl, especially for nouns denoting humans, e.g. samaan ‘shaman’ : PL samaa-sal. A few kinship terms take the lexicalized suffix -nil, which incorporates the final n of the stem, e.g. nexun ‘younger sibling’ : PL nexni-l = nex-nil ~ neƞ-nil. The form am-til ‘fathers’ > ‘parents’, from amin ‘father’, could also be viewed as a lexicalized example of an associative plural in -til. The most important difference between Neghidal and the other Ewenic languages is the absence of the plural marker -r, which is replaced by the regular -l ~ -sAl, e.g. oyon ‘reindeer’ : PL oyo-l ~ oyo-sol. The loss of -r as a plural suffix is apparently connected with the general elimination of the rhotic r as a phoneme in Neghidal. The case system in Neghidal is somewhat reduced as compared with Ewenki and Ewen, and includes the unmarked nominative and 8 suffixally marked cases: accusative, partitive, dative, ablative, locative, prolative, directive, and instrumental. The case suffixes show some allomorphic variation depending on the stem type (Table 10.4). Some special allomorphs of the case markers are used in the possessive declension. Thus, the ablative marker -Dukkey (< *-Duuki-) has the variant -Dukki- when combined with the possessive suffixes. Also, the accusative marker -wA has the variant -wu- in ACC-PX1SG -wu-w and the variant -wAA- in ACC-PX3SG -wA-n ~ -wAA-n. Additional variation is connected with the unstable character of vowel length in non-initial syllables. The directive case has the suffix variants -tti/i > -tci/i and -ttixi/i in the Lower dialect. Diachronically, all case markers are basically identical with those attested in Siberian Ewenki. The functions of the cases are also similar to those in the other Ewenic languages: •

The unmarked nominative is used in the subject position, e.g. emen xonat bi-cee-n ‘there lived a girl’ (one girl be-PST-3SG), as well as, in combination with a copula, TABLE 10.4 NEGHIDAL CASE MARKERS  

Vwyl

pts

k

mƞ

ACC

-wA

-pA

-kA

-mA

PART

-yA

DAT

-Du/u

ABL

-Dukkey : -Dukki-

LOC

-lA/A

-(Du-)lA/A

-du-lA/A

PROL

-li/i

-(Du-)li/i

-du-li/i

DIR

-tki/i

-itki/i

-tiki/i

-ñA

/n

PL -bA

-A

INSTR -Ci Stem types: V = vowel stems, /n = primary nasal stems, PL = plural stems. Segmental alternations: A = a o e, D = d t, C = j c.

Neghidal 243

•

•

•

in the role of a nominal predicate, e.g. egdeƞe beye oo-caa-n ‘s/he became an adult person’ (big person become-PST-3SG). The accusative (“definite accusative”) is the primary way for marking a direct object. It can also occur with intransitive verbs to indicate place or time, e.g. xuyeen-me=dee xul-li ‘(one) walking in the forest’ (forest-ACC=PTCL walk-PTCP. IMPRF), juu-le-we bi-cee-le-y ‘after having lived for two days’ (two-DAY-ACC be-PTCP.PRF-LOC-RX). The partitive (“indefinite accusative”) is used to mark an indefinite, unspecific, or partial object, e.g. olo-yo jep-mu-l-ce-w ‘I got the desire to eat some fish’ (fishPART eat-DESID-INCH-PST-1SG). As in Ewenki, the partitive, when combined with the possessive or reflexive suffixes, can also be used in a designative function, e.g. aya asi-ya-s baka-jiƞa-w ‘I will find a good wife for you’ (good woman-PARTPX2SG find-FUT-1SG). However, unlike in Ewenki, the negative existential noun aacin is normally not combined with the partitive in a connegative function, but with the nominative, e.g. amta aacin ‘there is no taste, tasteless’. An isolated example like idgi-ye-sel aacin oo-caa-tin ‘they lost (their) tails’ (tail-PART-PL NEG.EXIST become-PST-3PL) would suggest the connegative use of the partitive, but the irregular order of the suffixes (CX-NX) remains to be explained, and the whole construction may be ungrammatical. The dative has both locative and temporal functions, e.g. caalbukkaan-duu ‘(they lived) at Chalbukkan’, anƞan-du ‘(it happened) in (a certain) year’ (year-DAT). It can also express recipient, as in [1]:

[1] asi aya koto-wo eyexi-du=da buu-ye-n woman good knife-acc frog-dat=ptcl give-aor-3sg ‘The woman gave a good knife to the frog.’ In the passive construction the dative marks the agent, e.g. ugduli-du waa-w-caa-n ‘(he) was killed by a huge bear’ (big.bear-DAT kill-PASS-PST-3SG). It can also express the cause of an action, e.g. beye-l-duu-tin tagda-caa-w ‘I became angry at the people’ (person-PL-DAT-PX3PL become.angry-PST-1SG), or also the result of an action, as in [2]: [2] ule-ye-n ule-ye-n egdiƞe coƞdoxo-du ule-cee dig-aor-3sg dig-aor-3sg big pit-dat dig-pst ‘He digged and digged until he finally digged out a big pit.’ •

The ablative indicates the source of an action, often with motion verbs, e.g. togo-dukkey yuu-ye-n ‘(he) came out of the fire’ (fire-ABL come.out-AOR-3SG). It also has temporal uses, e.g. ta-dukkey ‘since that time’ (that-ABL). In the comparative construction, the ablative marks the standard of comparison, as in [3]:

[3] dolin buga beye-l-dukki-n maƞga beye aacin middle land person-pl-abl-px3sg strong person neg.exist ‘There are no people stronger than the people of the Middle Land.’ The ablative can also indicate source material, e.g. pasi-dukkey o(o)-da ‘(they) were formed (out of) pieces’ (piece-ABL become-AOR.3PL). Another function of the ablative is to indicate the cause of a situation, as in [4]:

244 Sofia Oskolskaya

[4] ayimkan xala-nin jeekte omu-dukki-n bayan-ca-tin Ayimkan clan-px3sg wasp nest-abl-px3sg become.rich-pst-3pl ‘The Ayimkan clan became rich because of a wasps’ nest.’ •

•

•

•

•

•

The locative expresses both dynamic movement (to) and static location (at), e.g. sin-dulee eme-w-je-w ‘I will bring it to you’ (2SG-LOC come-CAUS-FUT-1SG), xieña-nin bi-si-n ɪmana-laa ‘the trace is on the snow’ (trace-PX3SG be-AOR-3SG snow-LOC). It can also indicate the goal of motion, e.g. axota-laa ƞene-sin-e-n ‘he went hunting’ (hunting-LOC go-SEM-AOR-3SG), or the source of information, e.g. saman-dula saa-pila ‘(it) was learnt from a shaman’ (shaman-LOC know-PTCP. PRF.REM). Also, it often implies a close contact with an object, e.g. dɪl-laa-y iktecee-n ‘he knocked on his (own) head’ (head-LOC-RX knock-PST-3SG). The prolative indicates motion through or along something, e.g. miixi tuy-li xul-le-n ‘the snake moves by land’ (snake land-PROL move-AOR-3SG). It may also denote a period of time that lasts after an earlier-mentioned reference point, or, in reference to trading transactions, the price of a purchase. The instrumental typically indicates instrument, e.g. ogda ‘boat’ : INSTR ogda-ji ‘by boat’, but it also has a comitative function, e.g. amin ‘father’ : INSTR amin-ji ‘with (one’s) father’. In the latter function, when repeated, it can mark coordination, e.g. etixen-ji atixan-ji bi-ce-tin ‘there lived an old man and an old woman’ (old.man-INSTR old.woman-INSTR be-PST-3PL). It can also denote the source of emotional reaction, e.g. ie-ji ƞeele-l-cee-sun ‘what are you afraid of?’ (what-INSTR be.afraid-INCH-PST-2PL), а use that reflects its historical origin as a separative case marker. Adjectival nominals in the instrumental have an adverbial function and express manner, e.g. tuygen ‘quick’ : INSTR tuygen-ji ‘quickly’. In this function the instrumental marker has often been interpreted as a derivational suffix. The directive indicates the direction of movement, e.g. omu-ttixi-y ƞuni-je-n ‘(it) returned to its nest’ (nest-DIR-RX return-AOR-3SG). It also denotes the addressee of a speech action, e.g. oñoco gun-ce-n siƞeye-ttixi ‘the chipmunk said to the mouse’ (chipmunk say-PST-3SG mouse-DIR). In some examples the directive indicates cause, e.g. eyeekte sawgi-tki-n e-ce dooldi-ya ‘(he) did not hear (it) because of the noise of the stream’ (stream noise-DIR-PX3SG NEG-PST hear-CONNEG). In the Lower dialect, the directive marker is often attested in combination with the postpositional spatial element -gida/a- (< -gII-daa-) ‘place, side’, e.g. joo ‘house’ : joogida-tki ‘to(wards) the house’. In this usage, the complex suffix -gida-tki seems to have been taking over the role of the directive in its concrete local function, while the simple suffix -tki/i came to be used in the more abstract functions of the directive, as when denoting the addressee of a speech action. Neghidal also has the “genitival” suffix ‑ƞi, which marks the possessor in predicative position, e.g. ey ƞɪnaxin beye-ƞi ‘this dog belongs to the man’ (this dog man-POSS). As in the other Ewenic languages, the grammatical status of this suffix remains ambiguous: its productivity suggests that it is not a mere derivational suffix, but its morphosyntactic behaviour renders it different from the actual case markers. Another form sometimes mentioned in connection with the case paradigm is the so-called “vocative”, which is formed by the pattern -in → -A/y from the kinship terms amin ‘father’ : VOC ama ~ amay; eñin ‘mother’ : VOC eñe ~ eñey; axin ‘elder brother’ : VOC axa ~ axay; exin ‘elder sister’ : VOC exe ~ exey; note also the aberrant formation in nexun ‘younger sibling’ : VOC nexu. The vocative forms

Neghidal 245

are used when addressing persons, and syntactically they form independent clauses, e.g. eñ-ey, min-du e-xel soƞo-yo ‘mother, do not cry because of me!’ (mother-VOC 1SG-DAT NEG-IMP.2SG cry-CONNEG). The vocative can hardly be classified as a real case since it is restricted to a closed set of nominals, and also since vocative forms are not used as a part of fully-formed sentences. NUMERALS Numerals in Neghidal are nominal words with a regular nominal morphology, but with some syntactic idiosyncracies. In particular, they do not take modifiers, and, as in other Ewenic languages, their relationship to an accompanying noun is somewhat ambiguous. The basic cardinal numerals are: 1 eme/n, 2 juul, 3 ɪla/n, 4 digi/n > dii/n, 5 toñƞa, 6 ñuƞu/n, 7 nada/n, 8 japku/n, 9 iyegi/n ~ yegi/n, 10 jaa/n, all of which are in a more or less regular relationship to their counterparts in, for instance, Siberian Ewenki. It may be noted, however, that in the numeral 2 PL juu-l (< *jöö-r) the original plural marker *-r has been replaced by -l. The numeral 1 emen often occurs in the function resembling that of an indefinite article, e.g. emen asatka-kkan bi-ce-n ‘there was a girl’ (one girlDIM be-PST-3SG), but it is not fully grammaticalized in this function and its use is not obligatory. Of the decades only 20 oyi/n (< *x/orïn) has a special root (ultimately borrowed from Mongolic, but perhaps secondarily transmitted to Neghidal via Manchu), while the other decades are formed by the multiplicative principle, e.g. 30 = 3 × 10 ɪlan jaan, with SG jaan against PL *jaa-r in the other Ewenic languages. For the powers of ten, 100 taƞgu (from Manchu) and 1000 mɪƞga/n (from Mongolic via Manchu) are used. The intermediate numerals for 11–19 and from 31 onwards are formed by the additive principle, e.g. 11 = 10 + 1 jaan emen, while the numerals for 21–29 require the use of the ablative form oyin-dukkey (twenty-ABL), e.g. 21 oyin-dukkey emen. Ordinals are formed from the cardinals by the suffix -(g)ii-tin, with the loss of the final nasal and the preceding vowel of the numeral stems, e.g. dig-ii-tin ‘fourth’, nad-iitin ‘seventh’, ja-gii-tin ‘tenth’. The final element -tin may be identified as the possessive suffix of the plural third person. For the first two ordinals, the suppletive forms ñogu(-tin) ~ ñowu(-tin) ‘foremost, first’ and gie-tin ~ gɪnma ~ gɪtma ‘other, second’ are used. Collective numerals are formed from the cardinals by the suffix -nii, e.g. ɪla-nii ‘three together’. For the counting of collective groups of humans, the suffix complex -ni-ƞA-sAl, with the plural marker -sAl, is used, e.g. nada-ni-ƞa-sal ‘all the seven’. The corresponding forms of 2 juu-l are juu-yii ‘two together’ (< *juu-r-ii) and juu-yi-ƞesel ‘both of the two’. For counting animals, the special collective forms 2 ju-yi-gde, 3 ɪla-ƞna, 4 di-ƞne, 5 toñƞa-y-agda, 6 ñuƞi-ye-gde, 7 nada-ƞna, 8 japku-ƞna, 9 yegi-ƞne, and 10 jaa-yi-gda are used. Forms based on 1 emen include eme-xin ‘alone’ and eme-xkin ‘one and only’. Other numeral classes include the approximatives in -mAl, e.g. ñuƞun-mal ‘about six’; the distributives in -tAl, e.g. eme-tel ‘one by one’, juu-tel ‘by the two’; and the limitatives in SG -xkAAn : PL -xkAA-l ~ -kAxAA-l, e.g. eme-xkeen ‘only one’, ɪla-xkaa-l ‘only three’, dik-kexee-l ‘only four’, iyex-kexee-l ~ yex-kexee-l ‘only nine’. A distributive meaning can also be expressed by reduplicating the numeral, with the reduplicated part in the instrumental case, e.g. juul juul-ji ‘by the two’ (two two-INSTR). Neghidal also preserves the suffixal counter for ‘days’ (also ‘points in a game’) -lA (< *-lsA), e.g. juu-le ‘two days’, ɪla-la ‘three days’. These forms can take the prolative case marker, with or without an intermediate plural marker, to indicate the corresponding

246 Sofia Oskolskaya

duration of time, e.g. juu-le-l-duli ‘in two days’ (two-DAY-PL-PROL), ɪla-la-li ‘in three days’ (three-DAY-PROL). Other numeral derivatives include the forms for ‘times’ in -nAA ~ -yAA, e.g. emn-ee ‘once’, digi-yee ‘four times’, and ‘layers’ (“tuples”) in -nmAAn, e.g. ɪlan-maan ‘three layers, three-fold’. These forms can also take the diminutive suffix -kkAAn in the meaning ‘only’, e.g. em-nee-xeen ‘only once’, ɪla-nmaa-kkaan ‘only three layers’. Repetitive numerals (‘for how manieth time?’) are formed by the complex suffix -(g)iwgim(i), e.g. ɪl-iwgim(i) ‘for the third time’. The corresponding forms for the first two digits are ñogut or elexes ‘for the first time’ and gie-wgim(i) ‘for the second time’. PRONOUNS The Neghidal basic personal pronouns (Table 10.5), including the rules of their stem allomorphy, are essentially identical to those in Siberian Ewenki, with the minor exception that the object form is always expressed by a combination of the partitive and accusative markers. The normal oblique stem of all basic pronouns ends in n, but the directive case can also be formed from the vowel stem in e, identical with the partitive form, e.g. 1SG-DIR min-tikii ~ mine-tkii. The first person plural pronoun bu/u : OBL mun/e- is used in the exclusive function, while the corresponding inclusive pronoun in bit : OBL bit- in the Upper dialect and bitte ~ butte : OBL bitte- ~ butte- in the Lower dialect. In difference from the other Ewenic languages, the plural oblique stems 1PL. EXCL mune- : 2PL sune- can also take the plural suffix -l, e.g. 2PL-PL-ACC sune-l-be ‘you all’ (in object position). TABLE 10.5 NEGHIDAL PERSONAL PRONOUNS   SG PL

NOM

OBL

PART-ACC

GEN

POSS

1

bi/i

min/e-

min-e-we

min

min-ƞi

2

si/i

sin/e-

sin-e-we

sin

sin-ƞi

bu/u

mun/e-

mun-e-we

mun

mun-ƞi

su/u

sun/e-

sun-e-we

sun

sun-ƞi

1 2

EXCL

In the function of adnominal genitives (GEN), the Lower dialect uses the oblique stems 1SG min ‘my’ : 2SG sin ‘thy’ : 1PL.EXCL mun ‘our’ : 2PL sun ‘your’, e.g. icexen min sien-mu-w ‘look at my ears!’ (see-IMP.2PL 1SG.GEN ear-ACC-PX1SG). In the Upper dialect, but occasionally also in the Lower dialect, the basic nominative forms are used in the same function, e.g. si iesa-s ‘your eye(s)’ (2SG eye-PX2SG). All personal pronouns also have the special possessive form in -ƞi, based on the oblique stem, e.g. 1SG-POSS min-ƞi ‘mine’, which is used in predicative position. In the third person, the stem noƞa- (< *nuƞa- < *ñuƞa-) is used, yielding SG noƞa-n : PL noƞa-ti-l ~ noƞa-l-ti-l. The plural forms obviously represent a reanalysis of the more original structure PL-PX3PL *nuƞa-r-tin, as still present in Siberian Ewenki. The singular stem noƞan- also functions as the oblique base, to which the case markers, always accompanied by the corresponding possessive suffix, are added, e.g. ACC-PX3SG noƞan-maa-n : DAT-PX3SG noƞan-duu-n : ABL-PX3SG noƞan-dukki-n : LOC-PX3SG noƞan-dulaa-n, etc. The possessive marker -ƞi- also takes the possessive suffix in an analogous way: 3SG-POSS-PX3SG noƞan-ƞi-n ‘his’. In the plural, the oblique forms are based on the plural stem noƞa-l-, which also requires the corresponding possessive suffix

Neghidal 247

after the case markers, e.g. 3P-PL-ACC-PX3PL noƞa-l-ba-tin : DAT-PX3PL noƞa-l-duutin : ABL-PX3PL noƞa-l-dukki-tin : LOC-PX3PL noƞa-l-dulaa-tin, etc. The reflexive pronoun is meen : OBL meen- ‘(one)self’. The basic form expresses an adnominal possessor, e.g. meen koto-yi muu-tki jolo-do-co-n ‘he threw his (own) knife into the water’ (REFL knife-RX water-DIR stone-VBLZ-PST-3SG). In the function of an object, the basic form is combined with the singular and plural markers of reflexive possession, but with no accusative case marker, i.e., REFL-RX meen-mi : REFL-RX.PL meen-mey. In the other cases the reflexive markers are added after the case markers, e.g. meen-dulee-wey ‘to themselves’ (REFL-LOC-RX.PL). The forms of predicative possession follow the same pattern, e.g. meen-ƞi-yi ‘his/her own’ (REFL-POSS-RX). There is also the emphatic form meene-xeen ‘(one)self’, which is used indifferently in both numbers. The two basic demonstrative pronouns are ey ‘this’ (< *eri) : PL eyi-l ‘these’ vs. tay ‘that’ (< *tarï) : PL tayi-l ‘those’. The same forms are used as oblique stems in the directive and instrumental cases, i.e., DIR ey-tixii vs. tay-tixii : INSTR ey-ji vs. tay-ji, while the accusative has both the simple forms ACC ey-we vs. tay-wa, mainly in the Lower dialect, and the complex forms e-yee-we vs. ta-yaa-wa, which would seem to contain the partitive marker -yA-. Exceptionally, the demonstrative pronouns also have the special elative forms ELAT ey-giit ~ ey-mel-ji vs. tay-giit ~ tay-mal-ji, although the elative has otherwise been lost as a productive case in Neghidal. The other case forms are based on the simple stems (with a long vowel) ee- vs. taa-, i.e., DAT ee-duu vs. taa-duu : ABL ee-dukkey vs. taa-dukkey : LOC ee-lee vs. taa-laa : PROL ee-lii vs. taa-lii. The case forms of the demonstrative pronouns are normally used as adverbs, e.g. taa-dukkey ice-t-ce-n ‘s/he was watching from there’ (that-ABL see-RES-PST-3SG), or as independent headwords, e.g. tay-ji ƞeele-l-ce-n ‘he was frightened by that one (a devil)’ (that-INSTR be.afraid-INCH-PST-3SG). In the adnominal position, no agreement markers are required, e.g. tay kutuwe-we ice-cee-n ‘s/he saw that toad’ (that toadACC see-PST-3SG). The basic interrogative pronouns are ƞii ~ ñii ~ nii ‘who?’ and ie : ie-kun ~ ie-xun ‘what?’, also verbally ie- ‘to do what?’ : CV.PURP ie-daa ~ RX ie-daa-y ‘why?’. The corresponding plurals are ƞii-ye ~ ñii-ye ~ nii-ye and ie-l : ieku-l ~ iexu-l, respectively. The oblique forms follow the nominal paradigm, e.g. ACC ƞii-we, ie-wa, iekun-ma, except that the pronoun ƞii ~ ñii ~ nii has no partitive. Indefinite pronouns are derived from interrogative pronouns by adding the clitic particles =dAA, =wAl, =kAA ~ =xAA, e.g. nii=dee ‘somebody’. The interrogatives are also used in a connegative function when combined with negation, e.g. ie-wa=da e-si-n baxa-ya ‘s/he did not find anything’ (what-ACC=PTCL.CONNEG NEG-AOR-3SG find-CONNEG). Other interrogative items include oo- : oo-n ~ oo-ni ~ oo-ni-gda ‘how?’ : oo-xin ‘when?’ : oo-xin=daa ‘always’, a- : a-sun ‘how much?’ : a-dii ‘how many?’, ii- : DAT ii-duu ‘where’ : LOC ii-lee ‘where?’ : DIR ii-tkii ‘which way?’. The root ii- ‘what?’ is also attested in subordinated clauses, as in [5]: [5] e-si-n ice-w-we ii-tki degi-l-cee-wee-n neg-aor-3sg see-pass-conneg what-dir fly-inch-ptcp.prf-acc-px3sg ‘One could not see where (it) flew.’ Quantifiers include (Lower) copal and (Upper) opkaci/n ‘all, whole’, as well as bayan ‘many’ and xaanin ‘some’. They take regular nominal suffixes, as in [6]:

248 Sofia Oskolskaya

[6] sune-l-be bi copal-ba-sun ilbe-je-w ilbe-je-w pl2-pl-acc 1sg  all-acc-px2pl  drive.away-fut-1sg  drive.away-fut-1sg ‘I will drive away you all.’ PERSON MARKING Neghidal preserves the distinction between two sets of person markers: possessive suffixes (PX) and predicative personal endings (VX). Compared with Siberian Ewenki, the two sets are, however, less distinct in Neghidal, for the possessive suffixes have replaced the predicative endings in the second person singular and plural and the first person plural exclusive forms, leaving only the first person singular and the first person plural inclusive forms distinct from the corresponding possessive forms (Table 10.6).

TABLE 10.6 NEGHIDAL PERSON MARKERS   SG

PX

 

 

VX

C

m

CV

/n

PL

AOR

1

-i-

-uu-

-w

-m

-bi

-m

2

-i-

-uu-

-s/i

-si

3

-i-

-uu-

-nin : CX-n

RX PL

 

1

-n

-y/ii

-mii

-bii -bun

EXCL

-i-

-uu-

-Wun

-mun

INCL

-i-

-uu-

-t

-ti

2

-i-

-uu-

-sun

3

-i-

-uu-

-tin

RX

-s/i

-WAy

-wun -p -sun -Ø

-mAy

-bAy

Stem types: C = consonant stems (with a connective vowel), V = vowel stems, /n = primary nasal stems, PL = plural stems. Segmental alternations: A = a e, W = w p k.

The allomorphy of the possessive suffixes depends mainly on the stem type. The possessive suffix of the third person singular has, however, also the morphologically conditioned “long” allomorph -nin for the basic form, while the oblique forms, in which the possessive suffix follows the case markers (CX), retain the original “short” suffix -n, identical with the third person predicative ending, cf. e.g. amaxa xuyi-l-nin ‘the bear’s children’ (bear childPL-PX3SG) vs. omki xuyi-l-ba-n ‘the flying squirrel’s children’ (flying.squirrel child-PLACC-PX3SG). Another special development is that diachronically secondary consonant stems ending in the labial nasal m take all possessive suffixes, but not the reflexive suffixes, by the connective vowel -uu-, e.g. tem ‘raft’ : PX 1SG tem-uu-(w) : 2SG tem.uu-s : 3SG tem.​ uu-nin etc., while other consonant stems require the connective vowel -i-, e.g. laaƞ ‘trap’ : PX 1SG laaƞ.i-w : 2SG laaƞ.i-s : 3SG laaƞ.i-nin etc. The connective vowel can also be absent in the plural first person exclusive and second person forms if the stem ends in an obstruent, e.g. jepkit ‘food’ : PX 1PL.EXCL jepkit-pun : 2PL jepkit-sun. The system of person markers also comprises the reflexive suffixes (RX) for singular and plural reference, as used to indicate reflexive possession, but also, in some verbal

Neghidal 249

forms, the (conjunct) subject. The reflexive case paradigm lacks the accusative, since the basic reflexive forms automatically imply reference to an object, e.g. jon-gi-ye-n xuyi-l-bi ‘she is thinking about her children’ (recollect-REV-AOR-3SG child-PL-RX). The suffix of alienable possession ‑ƞ- is always combined with the possessive or reflexive suffixes, e.g. abgu-ca-tin ule-ƞ-ƞey emen tige-tki ‘they removed their meat (from fire and put it) into a bowl’ (remove.from.fire-PST-3PL meat-AL-RX.PL one bowl-DIR). VERBAL MORPHOLOGY The system of verbal forms in Neghidal offers challenges similar to those met in other Ewenic languages, in that the interrelationship between forms and functions is not fully crystallized, and from the functional point of view similar roles are filled by forms of several different categories. Even so, the basic formal division between finite predicative forms, nominalized forms or participles, subordinative forms or converbs, and imperatives, is clear. In this system, participles have a crucial position, for they are inherently multifunctional and can, therefore, enter the spheres of finite forms, converbs, and even imperatives, without losing their principal ability to function as nominalizations. In the finite function, the difference between participles and finite forms proper is also signalled by the different personal markers they take, in that the subject of a participial predicate is indicated by the possessive suffixes, while finite forms proper take the actual predicative personal endings. In Neghidal, this difference is somewhat obscured because of the tendency of the two sets of personal markers to merge, but the difference is still retained in a few persons. From the morphological point of view, the most complicated category is the finite forms proper, based on the aorist stem of the verb. The aorist itself, when combined with the personal markers, functions as an unmarked general tense of the finite conjugation, but it also provides the stem for several other forms functioning as participles or (quasi) converbs, or also as the connegative form. Neghidal preserves the four conjugational types of aorist formation: (1) original vowel stems, including also secondary consonant stems (due to the loss of the final vowel), which form the aorist by the element -yA- (< *-rA-), absent in the first and second persons singular, (2) original nasal stems, including a few types of derived stems, which form the aorist by the element -A- (< *-A- < *-dA- < *-rA-), added to the final nasal of the stem, (3) a limited class of special stems which form the aorist by the element -dA- (< *-dA-), and (4) a likewise limited class of stems which form the aorist by the element -si- (< *-sI-), e.g. (1) ana- ‘to push’ : AOR ana-ya-, (2) guun- ‘to say’ : AOR guun-e-, (3) ga- ‘to take’ : AOR ga-da-, (4) bi- ‘to be’ : AOR bi-si-. The full aorist personal paradigm may be illustrated as follows: ana- ‘to push’ : 1SG ana-m : 2SG ana-s/i : 3SG ana-ya-n : 1PL.EXCL ana-ya-wun : 1PL.INCL ana-ya-p : 2PL ana-ya-sun : 3PL ana-ya. After secondary consonant stems, the initial y of the aorist marker is assimilated by the preceding stem-final consonant, e.g. ayaw- ‘to love’ : AOR-3SG ayaw-wa-n, atal- ‘to take off’ : AOR-3SG atal-la-n, xig- ‘to skin’ : AOR-3SG xig-ge-n, jep- ‘to eat’ : AOR3SG jep-pe-n, aak- ‘to dock’ : AOR-3SG aak-ka-n, ɪs- ‘reach’ : AOR-3SG ɪs-sa-n, etc. Verbal negation is formed by the fully conjugated negation verb e- : AOR e-si-, which precedes the invariant connegative form of the notional verb. The connegative is identical with the aorist stem, which also functions as the third person plural finite form of the aorist paradigm, e.g. waa- ‘to kill’ : AOR waa-ya- : CONNEG waa-ya, as in e-si-m waa-ya ‘I did not kill’ (NEG-AOR-1SG kill-CONNEG).

250 Sofia Oskolskaya

PARTICIPLES Neghidal has six basic participial forms, which may be divided in terms of voice into two types, personal/active and impersonal/passive, and in terms of temporal-aspectual reference into three types past/perfective, present/imperfective, and future/futuritive. In addition, there is a number of other forms that can be used in a participial function. •

•

• •

The personal present/imperfective participle (glossed PTCP.IMPRF) is ultimately based on the aorist stem and is marked by the suffix -yii ~ -yi (< *-rII), with the same initial alternations as in the aorist marker. The corresponding impersonal participle incorporates the reflexive-passive suffix -w-, yielding the complex suffix -w-wii ~ -w-wi (< *-bO-rII). An analogous relationship exists between the personal past/perfective participle (glossed PTCP.PRF) in -cA/A (< *-cAA) and its impersonal counterpart in -w-cA/A (< *-bO-cAA). There is also another impersonal past participle marked by the suffix -plA ~ -pilA, which refers to a remote past (glossed PTCP.REM). The morphological background of the latter marker is unclear, but it might contain the anticausative (medial) element -p-. The personal futuritive participle is marked by the suffix -jA(ƞ)A ~ -jA (< *-jA-ƞAA), while the corresponding impersonal participle bears the marker -pcAn, which, again, seems to contain the anticausative (medial) element -p-. Other participle-like forms of the personal type, with a modally and/or aspectually modified meaning, include those in -wkii ~ -wki (habitual action), -mAktA (recently completed action), and ‑ƞAAt (potential action).

Thanks to their multifunctionality, participles can occur variously as adnominal modifiers, independent headwords, or finite predicates. As adnominal modifiers their status with regard to argument structure is influenced by the presence or absence of possessive suffixes. When used without possessive suffixes, they are oriented towards the subject e.g. waa-caa amban ‘the demon that killed (somebody)’ (kill-PTCP.PRF demon). By contrast, when combined with possessive suffixes, they are oriented towards the object (including an oblique object), e.g. exi-m gun-cee-nin amban ‘the devil that my elder sister spoke about’ (elder.sister-PX1SG say-PTCP.PRF-PX3SG devil). In the latter case, depending on the type of reference (conjunct or disjunct), they can also take the reflexive suffixes, e.g. xulegi eme-cee-yi oja-li-y ƞunu-sin-e-n ‘Xulegi returned the (same) way he had come’ (Xulegi come-PTCP.PRF-RX trace-PROL-RX return-SEM-AOR-3SG). The impersonal participles are normally oriented towards the object, e.g. waa- ‘to kill’ : PASS-PTCP.IMPRF waa-w-wii ‘(one) that is being killed’ : PRF waa-w-caa ‘(one) that has been killed’ : PRF.REM waa-pla ‘(one) that was killed long ago’. When used as finite predicates in the role of tense forms (glossed PRS vs. PST vs. FUT), participles take person markers of the possessive type, with the minor exception that the third person singular form takes the “short” variant -n of the possessive suffix, while regular nouns would take the “long” variant -nin. However, in the third person, the possessive suffixes can also be replaced by the corresponding singular and plural forms of the nominal declension, meaning that the singular remains unmarked (-Ø), while in the plural the suffixes -l or, more rarely, -sAl can be used, e.g. iƞñekte-we tesi-l-ca-l ‘(they) began to collect bird cherries’ (bird.cherry-ACC collect-INCH-PST-PL). The present impersonal participle is often used to express obligation, e.g. togo-wo ɪlan-mii

Neghidal 251

tamiwun-ji ɪla-ci-w-wi ‘if you make fire you should keep it on with a poker’ (fire-ACC make.fire-CV.CONN poker-INSTR make.fire-RES-PASS-PRS). Occasionally, an impersonal form can also be combined with an object, e.g. teli-we nee-kte-yi-we-n ice-w-wi ‘it is seen that (he) brought up dried fish’ (dried.fish-ACC put-DISTR-PTCP.IMPRF-ACCPX3SG see-PASS-PRS). The temporal and modal reference of the participles can be somewhat unstable. For instance, the present participle can sometimes refer to the past, e.g. ii-lee=dee ƞene-yi-nin e-si-n ice-w-we ‘one cannot see where he went’ (what-LOC=PTCL go-PTCP.IMPRFPX3SG NEG-AOR-3SG see-PASS-CONNEG). The habitual participle, which normally implies a habitually repeated action, can also refer to a single action, e.g. ƞene-wki ƞenewki ‘(it) is going and going’ (go-PTCP.HAB go-PTCP.HAB). The future impersonal participle, which is rather rarely used, occurs mainly in questions, e.g. e-du oon bi-pcen ‘how (should one) live here (now)?’ (this-DAT how be-PTCP.FUT.IMPERS). It is possible that the element -cA- in the marker of this participle is identical with the deverbal suffix -JA-, which in Neghidal is used to express an interrogative near future tense (Khasanova & Pevnov). As independent headwords, participles can take case markers and function as quasiconverbs, often with a temporal meaning (‘when’, ‘after’). The case forms used in these constructions include the dative, locative, ablative, and instrumental, e.g. talu-wa ile-yi-du-n ‘when (s/he) was stripping birchbark’ (birchbark-ACC strip-PTCP.IMPRF-DAT-PX3SG), yuy-gi-ce-lee-n ‘after (s/he) had gone out’ (go.out-REV-PTCP.PRF-LOC-PX3SG), emece-dukki-wey ‘after they had come’ (come-PTCP.PRF-ABL-RX.PL), tige-we jawa-yi-ji-n ‘when he took a bowl’ (bowl-ACC take-PTCP.IMPRF-INSTR-PX3SG). The negation of participial forms takes place with the help of the participial forms of the negation verb e- in combination with the connegative form of the notional verb. The personal past/perfective participle of the negation verb is formed regularly as e-cee and can be used also in a finite function, e.g. e-cee-n ana-ya ‘(s/he) did not push’ (NEGPST-PX3SG push-CONNEG). The corresponding present/imperfective participle of the negation verb has the special form e-ti/i, based on the derived stem e-te-, which has basically a future tense reference, as in si min-e-(w)e e-te-s waa-ya ‘you will not kill me’ (2SG 1SG-PART-ACC NEG-FUT-2SG kill-CONNEG). The participial form e-ti/i can be used both independently and with case suffixes, e.g. tay beye e-ti ƞeele-ye bi-cee-n ‘that man was (one who was) not afraid’ (that man NEG-PTCP.IMPRF be.afraid-CONNEG be-PST-3SG), e-ti-ji-w saa-ya gelbi-s=ku ‘(do you think that) I do not know your name?’ (NEG-PTCP.IMPRF-INSTR-PX1SG know-CONNEG name-PX2SG=INTERR). FINITE TENSE AND ASPECT The Neghidal system of finite indicative forms is composed of a mixture of aorist forms and finitely used participles, some of which can also be used in combination with an auxiliary. If analysed functionally in terms of tense, the system may be understood as comprising the aorist, which functions as a general unmarked non-future tense, participle-based past and future tenses, a derivationally formed near future tense, and two analytic forms functioning as the perfect and pluperfect tenses. •

The aorist has a varying temporal reference depending on the lexical semantics (aspectual or actional characteristics) of the verb. When formed from static and atelic verbs, it normally refers to the present, e.g. beyun xuklee-ye-n ‘the elk is lying’ (elk

252 Sofia Oskolskaya

•

•

•

lie-AOR-3SG), ii-lee ƞene-s ‘where are you going?’ (what-LOC go-AOR.2SG). When formed from telic verbs, it normally refers to the past, e.g. bi beyun-me waa-m ‘I (have) killed an elk’ (1SG elk-ACC kill-AOR.1SG). The temporal reference can also be influenced by derivational suffixes. For instance, inchoative verbs in -l- and semelfactive verbs in -sin- always have a past reference in the aorist, e.g. tay etixen enu-l-le-n ‘that old man fell ill’ (that old.man be.ill-INCH-AOR-3SG), while distributive derivatives in -ktA/A- always have a present reference, e.g. iedaa tay ñitku-l-be waa-kta-(y)a-sun ‘why do you kill those small (animals)?’ (why that small-PL-ACC kill-DISTR-AOR-2PL). In some cases the interpretation depends on the situation, e.g. ie-wa tolki-ci-m ‘what did I (just) dream about?’ (what-ACC dream-RES-AOR.1SG). The past tense is based on the past/perfective participle in -cA/A. The reference range of this form is wide: it can refer both to recent events and to a remote past, e.g. baxa-caa-w ‘I (just) found (it)’ (find-PST-1SG), gojo jule-du . . . bayan beye bi-ceetin ‘long time ago there lived many people’ (long.time fore.part-DAT many person be-PST-3PL). The future tense is based on the futuritive participle in -jA/A. It is used in reference to the future with no specific restrictions, e.g. sun-e-we jepu-l-jaa-n ‘s/he will eat you’ (2PL-PART-ACC eat-INCH-FUT-3SG). Reference to a near future can, however, also be expressed by the regular aorist forms of inchoative verbs in -l-, e.g. amban min-e-we jepu-l-le-n ‘the devil is about to eat me’ (devil 1SG-PARTACC eat-INCH-AOR-3SG). In the other tenses, the inchoative function is retained, e.g. asi tay olo-l-bo tii jep.u-l-ce-n ‘the woman began to eat those fishes like this’ (woman that fish-PL-ACC so eat-INCH-PST-3SG). Reference to a near future can also be expressed by the progressive (imperfective) marker -jA-, though this usage seems to be confined to interrogative sentences, while in affirmative sentences the regular future in -jA/A- is used, cf. e.g. si ii-lee ƞene-je-s ‘where are you going?’ ~ ‘where will you go?’ (2SG what-LOC go-PROGR-AOR.2SG) vs. bi muu-tki ƞenejee-w ‘I will go to the water’ (1SG water-DIR go-FUT-1SG). The difference between the elements -jA- and -jA/A- has, however, been difficult to retain because of the instability of vowel length in non-initial syllables, which has caused some confusion in the conjugational patterns. The analytic tense forms of perfect and pluperfect are based on the combination of the aorist and past tense forms of the copula-existential bi- ‘to be’ with a following past participle of the notional verb. In these constructions, the participle takes no personal marking, but in the plural it takes the regular nominal plural marker in -l, e.g. waa- ‘to kill’ : PRF 1SG waa-caa bi-si-m ‘I have killed’ : 1PL waa-caa-l bi-siwun : 3PL waa-caa-l (bi-si), PLUPRF 1SG waa-caa bi-cee-w ‘I had killed’ : 1 PL waa-caa-l bi-cee-wun : 3PL waa-caa-l bi-cee-tin. These forms indicate a past action the result of which continues to be relevant at the time of reference (present for the perfect and past for the pluperfect). Similar analytic forms can also be based on the futuritive participle, though they have an additional modal (deontic) implication, e.g. jawa- ‘to take’ PTCP.FUT jawa-jaa : 1SG AOR jawa-jaa bi-si-m ‘I will have to take’ : PST jawa-jaa bi-cee-w ‘I had to take’.

IMPERATIVES The imperatives form a heterogenic paradigm of finite forms with a modal function of the imperative type. The forms are differentiated according to person, but actual person

Neghidal 253

markers are present in them only marginally. Although a personal pronoun can be present, it is not an obligatory part of the imperative sentence. •

•

•

The neutral imperative marker for the second person singular is -kAl ~ -xAl. It is used in commands or requests, e.g. si egdiƞe muulek-ke jawa-xal ‘take a big bucket!’ (2SG big bucket-ACC take-IMP.2SG). The corresponding plural marker is -kA-sun ~ -xA-sun ~ -xAn, all reflecting original form *-kAl-sun (IMP-2PL). In prohibition the imperative is marked on the negation verb: SG e-kel : PL e-kesun, dialectally (Ust’-Amgun’) also u-xesun, e.g. u-xesun ƞene-(y)e ‘don’t go!’. In the function of a “future” imperative for the second person, the “supine” form in -dAA-, with the singular and plural reflexive markers can also be used, e.g. esi eme-ygi-jixi-n si gundee-y ‘now when (it) comes tell (it)!’ (now come-REV-CV.AOR-PX3SG 2SG saySUP-RX). The “future” imperative refers to a remote action that is supposed to take place after some other event in the future. The first person imperatives, which may functionally also be identified as “optatives” or “voluntatives”, are marked by the suffixes -ktA ~ -xtA for the singular and ‑ƞAy ~ ‑gAy ~ ‑ƞAn for the plural. In the singular the form indicates volition of the speaker concerning his own actions, e.g. ja-l-bi yuu-wkee-kel bi taƞi-kta ‘bring out your relatives (and) I will count them!’ (relative-PL-RX go.out-CAUS-IMP.2SG 1SG count-OPT.1SG). The corresponding plural is used in the inclusive sense and expresses invitation to a joint action, e.g. ƞene-sin-ƞey ‘let us go!’ (go-SEMVOL.1PL.INCL). The third person imperatives, which may functionally be identified as “permissives”, are marked by the element -ƞi- ~ -gi-, to which the possessive suffixes of the third person singular and plural are added, yielding SG -ƞi-n ~ -gi-n : PL -ƞi-tin ~ -gi-tin. These forms denote wish or permission, e.g. tay bumaska-wa xute-du-w bu-ygi-ƞi-n ‘let him/her give that money back to my son!’ (that money-ACC child-DAT-PX1SG give-REV-PERM-3SG).

OTHER MODAL FORMS Apart from the imperatives, which involve a closed set of forms with functions distributed according to the subject person, Neghidal has several other forms with a modal content, which are morphologically situated at the intersection of inflection and derivation. Basically, it is a question of derivational forms of either the verbal or the nominal type, but the use of these forms in the function of finite predicates involves exceptions especially as regards the choice of the person markers. The principal forms of this category include what may be identified as the subjunctive, potential, and preventive moods. •

The subjunctive mood is formed by the suffix -mcA(-), which takes the possessive markers in the first and second person, while the third person has zero marking in the singular and the plural marker -l in the plural. This type of morphological behaviour is also known from participles, meaning that -mcA(-) could be analysed as a participial suffix. The subjunctive occurs with either a present or a past reference. In the latter case, the suffix with person and number marking is attached to the auxiliary bi- ‘to be’, preceded by the past participle of the notional verb. The subjunctive is typically used in apodosis of counterfactual conditional sentences, as in [7]:

254 Sofia Oskolskaya

[7] atika-kkaan aawun-maa-n teti-w-weki-sun old.woman-dim  hat-acc-px3sg  put.on-caus-cv.aor-px2pl ‘If you put the old woman’s hat on me, ayama-mat yamka-mca-w very.good-dx.mod perform.rite-subj-px1sg I would perform the shamanistic rites better.’ •

The potential mood is formed by the suffix -nA/A(-), which takes the personal endings of the aorist type, meaning that the form is actually a deverbal verb. The aorist marker -yA- appears, however, only in the third person plural form, e.g. ana- ‘to push’ : POT 1SG ana-naa-m : 2SG ana-naa-s : 3SG ana-naa-n : 1PL.EXCL ananaa-wun : 1PL.INCL ana-naa-p : 2PL ana-naa-sun : 3P ana-naa-ya. This form indicates that a situation is likely to happen. The personal forms of the potential mood are normally accompanied by the clitics =jAxA ~ =cAxA or =mAl, which stress the notion of likelihood, as in [8]:

[8] ey-gecin beyiƞee-ji degel-be-sun aw-wixi-tin this-like fur-instr face-acc-px2pl wash-cv.aor-px3pl ‘If they wash your face with this kind of fur, ɪs-na-n=ǯaxa reach-pot-3sg=ptcl it is probably enough.’ •

•

The preventive mood functions as an expression of warning (‘look out!’) and is formed by the suffix -jA-, which is identical with the deverbal suffix for progressive (imperfective) verbs, as also used to indicate the near future in the system of tense forms. Like the potential mood, the preventive mood follows the personal paradigm of the aorist set, but with the aorist marker -yA- used only in the third person plural, to which, exceptionally, the nominal plural marker is also attached, e.g. waa- ‘to kill’ : waaja-ya-l=bel ‘look out!—they will kill!’ (kill-PREV-AOR-PL=PTCL). In the second person the additional derivational suffixes -xkO- or -lkA- are added before the person markers, e.g. waa-ja-xkO-sun ‘look out!—you will kill!’ (kill-PREV-DX-2PL), min-duk(k)e(y) tepekeelee-w-je-lke-s ‘look out!—you will be shot by me’ (1SG-ABL shoot.arrow-PASS-PREV-DX-2SG). Modal notions can also be expressed by analytic constructions. Thus, along with the potential mood a possible situation can be expressed by a construction which includes a participle of a notional verb in combination with the auxiliary bi- ‘to be’ in the form of the future tense in -jA/A-, e.g. tayasal beyiƞee-we waa-yi bi-je-tin ‘they will probably kill the animal’ (those.people animal-ACC kill-PTCP.IMPRF be-FUT-3PL), aya bi-ce bi-je-n tetcenƞe-nin ‘her clothes were probably beautiful’ (good be-PTCP.PRF be-FUT-3SG clothes-PX3SG). The inevitability of a situation is expressed by the combination of the futuritive participle of a notional verb with the auxiliary bi- ‘to be’, e.g. bi=le iexun=daa aacin-ji ew-gi-je bi-si-m ‘I have to return without anything’ (1SG=PTCL what=PTCL.CONNEG NEG.EXIST-INSTR go.down-REV-PTCP.FUT be-AOR-1SG).

Neghidal 255

CONVERBS Neghidal has about ten productive converbal forms, which may be divided into the conjunct (same-subject), disjunct (different-subject), and ambivalent types. In general, the converbal markers are morphologically invariant, but a few have different forms depending on the number and/or person of the subject (Table 10.7). Most of the converbs have formal and functional parallels in the other Ewenic languages. TABLE 10.7 NEGHIDAL CONVERB MARKERS  

 

CONN

-mi/i

SUCC

-mdin

PL

PX/RX

SS

DS

+

-mAy

+

SIM

-nAkAn

+

ANT1

-yAAn

+

ANT2

-xAAnAm

LIM

-knAAn

+

+

+

TERM

-DAlA/A

+

+

+

CTEMP

-ƞAsA

+

+

+

PURP

-dA-

+

+

+

AOR

-yAki-

+

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e, D = d t.

•

The principal conjunct converb is the general connective converb in -mi/i, which is used in protasis of conditional and temporal sentences, e.g. si etixen-me ice-mi e-xel waa-ya ‘if you see a bear, do not kill it!’ (2SG bear-ACC see-CV.CONN NEGIMP.2SG kill-CONNEG). It can also occur in counterfactual conditional sentences, as in [9]:

[9] ti ñexe-je-we-s saa-caa bi-mii so behave-ptcp.fut-acc-px2sg know-ptcp.prf be-cv.cond ‘If I had known that you are going to behave like that,

e-mcee-w xelbu-ye neg-subj-px1sg take.along-conneg I would not have taken you with me.’

The connective converb can also express simultaneous action, e.g. ta-lii e-lii toksa-kta-mi ileƞ moo-wa baxa-ca-n ‘while s/he was running here and there, s/he found dry wood’ (that-PROL this-PROL run-DISTR-CV.CONN dry wood-ACC find-PST-3SG). For some verbs, such as ete- ‘to finish’, the converbs in -mi/i serve as dependents of matrix predicates, e.g. solaki kamnu-mi ete-yeen gun-e-n ‘when the fox had finished gluing it said . . . ’ (fox glue-CV.CONN finish-CV.ANT say-AOR-3SG). Although normally invariant in form, the suffix -mi/i, which etymologically contains the reflexive suffix of

256 Sofia Oskolskaya

singular reference, retains the corresponding plural form -mAy in occasional use in the Upper dialect, e.g. sagdi oo-may ‘when they became old’ (old become-CV.CONN.PL). •

•

•

Following, as far as possible, the terminology established for the other Ewenic languages, the other conjunct converbs may be identified as the successive converb in -mdin, dialectally also -mnin ~ (Lower) -mdixAAn, the simultaneous converb in (Upper) -nAkAn ~ (Lower) -nAxAn, and the anterior converbs in -yAAn and (Lower) -xAAnam. The successive converb indicates a succession of events, e.g. ƞene-mdixeen ice-ce ‘he had been walking (for a time) when he saw’ (go-CV.SUCC see-PST). The simultaneous converb indicates an action taking place in parallel with that of the main clause, e.g. xute-yi gelbi-wee-n gelbi-t-nexen gun-ce-n ‘calling her child by name she said’ (child-RX name-ACC-PX3SG name-VBLZ-CV.SIM say-PST-3SG). The temporal reference can, however, in some cases be also to an anterior action, e.g. egdiƞe jeekte-xeeye-l oo-naxaan degi-l-ce-l ‘big wasps emerged and flew’ (big wasp-AUGM-PL become-CV.SIM fly-INCH-PST-PL). Very similar implications of anteriority are also expressed by the actual anterior converbs, e.g. waa-yaan telge-ce-tin ‘they killed (it) and cut (it) up’ (kill-CV.ANT1 cut.up-PST-3PL), aya koto-l-bo jawa-xanam beyun-me telge-naa-ygiƞay ‘let us take good knives and go to cut up the elk!’ (good knife-PL-ACC take-CV. ANT2 elk-ACC cut.up-AND-REV-VOL.1PL). The background of the anterior converb in -yAAn, with cognates in Nanai-Ulcha and Oroch, remains without a definitive explanation, though the suffix clearly contains the aorist marker -yA- (< *-rA-). The most frequent disjunct converb is the aorist converb in (Upper) -yAki- ~ (Lower) -yixi. These markers correspond functionally to the aorist converb in *-rA-kI- in the other Ewenic languages, though the form -yixi (~ ? -yii-xi) has also been analysed (by Cincius) as containing the imperfective participle in -yi/i (< *-rII). The latter analysis would seem to be corroborated by the fact that the element -ki- (originally a local case suffix) is attested in Lower Neghidal also in combination with the perfective participle, e.g. waa- ‘to kill’ : waa-caa-xi-n ‘when I had killed’ (kill-PTCP. PRF-CX-PX3SG), though the combination waa-caa-yiixi-n (kill-PTCP.PRF-CV. AOR-PX3SG) has also been recorded from the language. Such forms are impossible to verify today, but the suffix complex -cAA-xi-PX is well attested in the extant text materials, e.g. sagji be(y)e gun-cee-xi-tin teƞ o-ygi-ca ‘what the old people had said became true’ (old man say-PTCP.PRF-CX-PX3PL precisely become-REV-PST). Except for its switch-referential status, the aorist converb is functionally very similar to the connective converb in -mi/i, in that it also expresses a general subordination with both conditional and temporal implications, e.g. ie-ma=da moo-la do-yixi-n moo kalga-m-ma-n ‘when (it) alights on any tree, the tree is broken’ (what-DX= PTCL tree-LOC alight-CV.AOR-PX3SG tree break-PASS-AOR-3SG). Other disjunct converbs include the limitative converb in -knAn, the terminative converb in -dAlA/A-, the contemporal converb in ‑ƞAsA-, and the purposive converb in -dAA-. The limitative and terminative converbs express an action that takes place before reaching a limit, e.g. tii xeteke-s-se-n . . . lam megdin.i-n dagama-l.i-knan ‘so (it) jumped until the sea brink approached’ (so jump-ITER-AOR-3SG sea brinkPX3SG approach-INCH-CV.LIM), i tii sagdan-dala-y iin-ca-n ‘and so (s/he) lived until old age’ (and so get.old-CV.TERM-RX live-PST-3SG). The contemporal converb expresses simultaneous action in the past, e.g. elan siwun bi-ƞese-n jegde-ce-n tuy ‘when there were three suns, the ground was burning’ (three sun be-CV.CTEMPPX3SG burn-PST-3SG ground). Finally, the purposive converb (the “supine”) expresses the purpose or goal of an action, e.g. jaa-l-bi altat-ca-n ete-dee-tin ‘s/he

Neghidal 257

asked her relatives to stop’ (relative-PL-RX ask-PST-3SG stop-CV.PURP-PX3PL). When used in a conjunct construction, the purposive converb takes the reflexive suffixes, e.g. beye-l ice-ce-tin teje-wee-n saa-daa-way ‘people looked in order to learn the truth’ (person-PL see-PST-3PL truth-ACC-PX3SG know-CV.PURP-RX.PL). Converbs can have negative forms. In the negative construction the converb markers and possessive and/or reflexive suffixes are attached to the negation verb, e.g. e-dee-tin kusi-ye ‘so that they would not fight’ (NEG-CV.PURP-PX3PL fight-CONNEG). SYNTAX Neghidal shares the general syntactic typology of the other Ewenic languages, for instance, the nominative-accusative alignment of the argument structure, the basic verb-final word-order (SOV) of the sentence, the headfinal structure of the nominal phrase (GAN), and the use of postpositional spatial nominals to indicate such more complex spatial and temporal relationships as are not expressed by the case forms. However, as in Siberian Ewenki, deviations from the rules of word order are relatively common. In complex sentences, the subordinate clause, expressed by non-finite verb forms (converbs and quasiconverbs), normally precedes the main clause. No native conjunctions are used, but some roles of conjunctions are filled by enclitic particles like =dAA, which, when repeated, can have a coordinative function for both clauses and nominal phrases, e.g. eyexi=dee cex oo-ca, beyun=dee xukti-l-dee-y cex batkari-ca ‘(both) the frog got ready and the elk was (also) preparing to run’ (frog=PTCL ready become-PST elk=PTCL run-INCH-CV.PURP-RX ready prepare-PST). Polar questions are marked by the interrogative clitic =gu ~ =ku ~ =ƞu ~ =wu, e.g. cɪwkaaƞ-ƞa-s jeb-je-m=ƞu ‘may I eat your little bird?’ (little.bird-AL-PX2SG eat-PROGR-AOR.1SG=INTERR). LEXICON AND LANGUAGE CONTACTS Neghidal has three layers of Tungusic lexicon. The basic layer involves the vocabulary deriving from Proto-Tungusic and shared by all Tungusic languages. This layer comprises a large number of words pertaining to the basic vocabulary, including grammatical elements, pronouns, numerals, kinship terms, terms for body parts, natural conditions, plants and animals, various verbal functions, and others, but there are also words for technological concepts, like metals, tools and weapons, dwellings, etc. This layer contains a number of Mongolic elements transmitted to Proto-Tungusic. A second layer of Tungusic lexicon comprises the vocabulary shared with the other Northern Tungusic, especially Ewenic, languages, and, in particular, with Siberian Ewenki, which is by all criteria the closest relative of Neghidal. Since Neghidal culture includes a component connected with taiga hunting and reindeer breeding, there are many specific terms pertaining to these areas of activity that are shared with Siberian Ewenki. Some of these items are also borrowings from third languages. A third layer of Tungusic lexicon comprises words shared on an areal basis with the other Tungusic languages spoken in the Lower Amur basin, including, historically, also Jurchen-Manchu. Several items with no further etymology were borrowed very locally to Neghidal from Ulcha, e.g. (Upper) aapi ~ (Lower) xapi ‘swim bladder (of fish)’ ← Ulcha xapa id. Manchu loanwords were also transmitted to Neghidal mainly via Nanai and Ulcha, and phonetic criteria often allow them to be distinguished from the inherited

258 Sofia Oskolskaya

lexicon. For instance, the numeral 100 taƞgu was probably borrowed from Nanai-Ulcha taƞgu, which, in turn, is based on Manchu tanggu [taƞgu] id. The Manchu item is ultimately a reflex of the Proto-Tungusic stem *taƞu ‘count’ : *taƞu- ‘to count’, which is regularly represented in Neghidal as taƞ- : taƞi- ‘to count, to read’. External non-Tungusic sources of lexicon include, most importantly, Amur Nivkh, which was apparently the substrate language for, at least, Lower Neghidal. The lexical traces of Nivkh in Neghidal are still insufficiently studied, but they probably include items pertaining to such local activities as river fishing, dog-breeding, and bear cult. In many cases, the exact phonological relationship remains problematic, cf. e.g. Lower Neghidal kiwte ‘fish-hook’ (for sturgeon) ← Amur Nivkh k’ite = Sakhalin Nighvng k’itƞ (< *ki-teƞ) ‘harpoon’ (for sturgeon). There are also a few items borrowed from Yakut, e.g. (Upper) akari ‘stupid’ ← Yakut akaarï id., (Lower) ogus ‘bull’ ← Yakut ogus id. Russian borrowings started entering Neghidal in the 18th to 19th centuries, and some of them have undergone considerable phonetic and semantic adaptations, e.g. Russian bumázhka ‘a small piece of paper’ → Neghidal bumaska ~ guumaska ~ maaska ‘money’. Often, the etymological origin of Neghidal words remains unclear. There are, for instance, a large number of terms for the local types of fish, many of which are shared with the neighbouring languages, while others are not. Some Neghidal items with an apparently Tungusic structure, e.g. apkakta = apka-kta ‘oak’, are also unique and have no known cognates in the other Tungusic languages. REFERENCES AND FURTHER READING Boshnyak, N. K. [Н. К. Бошнякъ] (1859) ‘Экспедицiи въ Приамурскомъ краѣ’ [Expeditions in the Amur region], Морской сборникъ 39 (2): 323–342, Санкт-Петербургъ [St. Petersburg]. Cincius, V. I. [В. И. Цинциус] (1957) ‘Загадки негидальцев’, Ученые записки ЛГПИ 132: 215–226, Ленинград [Leningrad]. Cincius, V. I. [В. И. Цинциус] (1982) Негидальский язык: Исследования и материалы [The Neghidal language: Studies and materials], Ленинград [Leningrad]: “Наука”. Cincius, V. I. [В. И. Цинциус] (1997) ‘Негидальский язык’ [The Neghidal language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 188–201, Москва [Moscow]: “Индрик”. Cincius, V. I. [В. И. Цинциус] (ed.) (1975–1977) Сравнительный словарь тунгусоманьчжурских языков: Материалы к этимологическому словарю [Comparative diactionary of the Tungusic languages: Materials for an etymological dictionary], vols. 1–2, Ленинград [Leningrad]: “Наука”. Dolgikh, B. O. [Б. О. Долгих] (1960) Родовой и племенной состав народов Сибири в XVII в. [The clan and tribal composition of the peoples of Siberia in the 17th c.], Труды института этнографии им. Н. Н. Миклухо-Маклая, Новая серия, vol. LV, Москва [Moscow]: Издательство АН СССР. Kalinina, E. Yu. [Е. Ю. Калинина] (2008) ‘Этюд о гармонии гласных в негидальском языке, или негласные презумпции о гласных звуках’ [A study of vowel harmony in the Neghidal language], in: A. V. Arkhipov [А. В. Архипов] et al. (eds.), Фонетика и нефонетика, К 70-летию Сандро В. Кодзасова, Языки славянских культур, 272– 282, Москва [Moscow]. Kazama, Shinjirō 風間伸次郎 (2002)『ネギダール語テキストと文法概説』[Neghidal Texts and Grammar],『ツングース言語文化論集』[Publications on Tungus

Neghidal 259

Languages and Cultures] 19, ELPR Publications Series A2–021, 吹田 [Suita]: 大阪 学院大学情報部. Khasanova, M. M. & A. M. Pevnov (2003) Myths and Tales of the Neghidals, ELPR Publications Series A2–024, Osaka. Kolesnikova, V. D. [В. Д. Колесникова] & O. A. Konstantinova [О. А. Константинова] (1968) Негидальский язык [The Neghidal language], in: Монгольские, тунгусоманьчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 109–128, Ленинград [Leningrad]: “Наука”. v. Middendorff, A. Th. (1847–1875) Reise in den äussersten Norden und Osten Sibiriens während der Jahre 1843 und 1844, St. Petersburg: Kaiserliche Akademie der Wissenschaften. Myl’nikova, K. M. [К. М. Мыльникова] & V. I. Cincius [В. И. Цинциус] (1931) Материалы по негидальскому языку, in: Тунгусский сборбик 1: 107–218, Ленинград [Leningrad]: Издательство АН СССР. Pakendorf, Brigitte & Natalia Aralova (2018) ‘The endangered state of Negidal: A field report’, Language Documentation & Conservation 12: 1–14, Honolulu. Patkanov, S. K. [С. К. Паткановъ] (1912) Статистическiя данныя показывающiя племенной составъ населенiя Сибири, языкъ и роды инородцевъ [Statistical data showing the tribal composition of the indigenous population of Siberia, their languages and clans], vol. III, С-Петербургъ [St. Petersburg]. Pevnov, A. M. [А. М. Певнов] (1998) ‘Рефлексы вибранта в негидальском на фоне родственных ему языков’ [The reflexes of the vibrant in Neghidal in comparison with related languages], Journal de la Société Finno-Ougrienne 85: 125–147, Helsinki. Pevnov, A. M. [А. М. Певнов] (2007) ‘Глагольное основоизменение в негидальском языке’ [Verbal derivation in Neghidal], Acta Linguistica Petropolitana 3 (3): 191–244, Санкт-Петербург [St. Petersburg]: Нестор-История. Pevnov, A. M. [А. М. Певнов] & M. M. Khasanova [М. М. Хасанова] (1994) ‘Негидальский язык’ [The Neghidal language], in: V. P. Neroznak [В. П. Нерознак] (ed.), Языки народов России: Красная книга, Энциклопедический словарьсправочник [Red book of the languages of the peoples of Russia: An encyclopaedic dictionary], 128–133, Москва [Moscow]: “Аcademia”. Pevnov, A. M. [А. М. Певнов] & M. M. Khasanova [М. М. Хасанова] (2006) ‘Негидальцы и негидальский язык’ [The Neghidal people and language], Acta Linguistica Petropolitana 2 (1): 447–542, Санкт-Петербург [St. Petersburg]: “Наука”. Schmidt, P[eter] (1923) ‘The language of the Neghidals’, Acta Universitatis Latviensis 5: 3–38, Riga. v. Schrenck, Leopold (1881) Die Völker des Amur-Landes, Reisen und Forschungen im Amur-Lande in den Jahren 1854–1856, vol. 3, St. Petersburg: Kaiserliche Akademie der Wissenschaften. Sternberg, Leo [Л. Я. Штернберг] (1933) Гиляки, орочи, гольды, негидальцы, айны: Статьи и материалы [The Ghilyak, Oroch, Gold, Neghidal, Ainu: Papers and materials]. Хабаровск [Khabarovsk]: Дальгиз. Vdovin, I. S. [И. С. Вдовин] (1953) ‘Историко-этнографические сведения о негидальцах середины XVIII столетия’ [Historical and ethnographic data on the Neghidal in the middle of the 18th century], Ученые записки ЛГУ, Серия факультета народов Севера 2: 208–214, Ленинград [Leningrad].

CHAPTER 11

EWEN Andrej Malchukov

Ewen (Even) is the language of the Ewen people (Russian èvén : PL èvény), a Northern Tungusic (Ewenic) minority group conventionally known as the Lamut (Russian lamút : PL lamúty), an ethnonym based on the Tungusic name of the Sea of Okhotsk (*laamo ‘sea’). The name Ewen, currently accepted in Russian specialist literature, is also the most widespread endonym used by the Ewen. According to the census of 2010 the total population of the Ewen was 21,830 individuals. Despite their small number the Ewen are spread across a huge area in north-eastern Siberia. The majority of the Ewen, 14,914 individuals, live currently within the Yakut (Sakha) Republic, or Yakutia, while 2,555 live in Magadan Oblast’, 1,104 in Khabarovsk Krai, and 1,848 on Kamchatka (Kamchatskii Krai) of the Russian Federation. Smaller groups of Ewen are scattered across the Chukchee and Koryak Autonomous Districts. According to the 2010 census statistics only 25 % of ethnic Ewen (5,656 individuals) spoke Ewen as their mother tongue. At present, the majority of Ewen are bilingual, that is, they speak Russian, as well. Trilingual Ewen, who also speak Yakut, are largely resident in Yakutia. The language retention rate for the ethnic Ewen resident in Yakutia is even lower than average. The rates are higher among elderly individuals, among groups engaged in traditional activities, e.g. reindeer-breeding, and in places of concentration of Ewen population. In communities with a significant Ewen population, the Ewen language is used as a means of instruction in the preliminary grade of school and is taught as a subject in primary school. At higher levels of education Ewen is taught at some teachers’ colleges, at the Institute of the Peoples of the Far North (Institut narodov Severa) of the Herzen Pedagogical University in St.-Petersburg, and at the Department of Northern Philology of the North-Eastern Federal University in Yakutsk. Ewen has a written form which is used quite extensively in publication, including both textbooks for primary school and fiction (both original works and translations from Russian). The writing system was introduced in the early 1930s. Originally based on the Latin alphabet, it was transferred to Cyrillic in 1937. In 1958 several special letters, as for the phonemes /ɵ/ and /ƞ/, were added to the Ewen alphabet. It should be noted, however, that in view of dialect disparity the adopted writing system is not equally suitable for speakers of all dialects. Therefore, in the 1980s there was an attempt to reform the Ewen writing system in order to make it more similar to that of Yakut and, consequently, more comprehensible for speakers of the western dialects. The proposed changes, however, have gained only restricted currency. Standard “literary” Ewen is based on the eastern dialects, in particular, the Ola dialect spoken in Magadan Oblast’. As an oral language, Standard Ewen has not yet achieved recognition as a means of communication among speakers of

DOI: 10.4324/9781315728391-11

Ewen 261

the different dialects. It is less intelligible for speakers of the western dialects and its use is largely restricted to the written form of the language. DIALECTAL DIVISION There are about a dozen distinct dialects of Ewen mentioned in the specialist literature. Apart from the extinct Arman variety, today often classified as a separate language, there are two main dialect groups, largely distinguished on phonetic grounds: eastern and western. Speakers of the western dialects are mainly resident in Yakutia, while speakers of the eastern dialects live in the Magadan and Khabarovsk regions, on Kamchatka, as well as in some enclaves in the north-eastern parts of Yakutia. The eastern dialects can be subdivided into the central-eastern (in particular, the Ola dialect, but including also most other varieties spoken in the Magadan region), north-eastern (the Kamchatka dialect) and south-eastern (the Okhotsk dialect) groups. The western dialects include the western proper (in particular, the Sakkyryr dialect), north-western (in particular, the Allaikha dialect) and mid-western (in particular, the Indigirka dialect) groups. The eastern dialects are characterized by the use of the fricative s [s] in word-medial and word-final positions, while the western dialects display a pharyngeal x [h] in all phonetic positions, as in eastern asɪ vs. western axɪ ‘woman’. The pharyngeal [h] is sometimes attributed to the influence of Yakut (Menges). Other features that are indisputably due to interference from Yakut are the labial assimilation of open vowels in non-initial syllables, as in eastern oran vs. western oron, the diphthongization of long vowels, as in eastern jɵɵr vs. western juɵr [ʤyør] ‘two’, and the articulation of the word-final palatal c and j as alveolar t and d, respectively, as in eastern uj vs. western ud ‘trace’. These latter features are less characteristic for some mid-western dialects, which are in this respect intermediate between the western proper and the eastern dialects. In the grammar, the (mid-)western dialects are characterized by the lack of possessive pronouns (personal pronouns are used instead), and the loss or irregularity of modifier agreement, which may also be attributed to interference from Yakut. The south-eastern dialects show interference from the neighbouring Ewenki dialects; in particular, they merge the front and back variants of the high vowels and regularize verbal agreement paradigms according to Ewenki patterns, e.g. AOR-3SG maa-ra-n ‘he killed’ instead of AOR.3SG maa-n. The extinct Arman variety, which used to be spoken on the river Arman (Arman’) by a group of settled Ewen (called mene by the nomadic Ewen), stands out in that it retained a number of archaic features (still found in Ewenki), such as the word-initial consonants s-, l-, n-, j-. Another peculiar characteristic of Arman (which may be due to a shift in stress patterns) is the metathetic restructuring of the word-final syllables to avoid a closed syllable in word-final position, as in regular ewen ‘Ewen’ vs. Arman ewne. In the grammar, Arman is less distinctive, as many of its morphological peculiarities that are not phonologically conditioned are attested in some other dialects, as well. Among exceptions one may note the absence of the basic imperative forms in Arman. Some other dialects are mostly distinguished on lexical grounds: thus, the north-western dialects reveal numerous borrowings from Yukaghir, and the north-eastern dialects from Koryak and Chukchee. Generally, it is difficult to provide a definitive list of Ewen dialects, since—as in many other cases where dialectal differentiation is to a large extent contact-induced—we are dealing with a dialect continuum, with a number of intermediate idioms showing different degrees of conventionalization.

262 Andrej Malchukov

DATA AND SOURCES The study of Ewen was initiated in the middle of the 19th century, largely thanks to Anton Schiefner (1859, 1874), who published language materials collected by earlier expeditions. The first grammar of Ewen was written by V. G. Bogoraz in the late 1890s, but it appeared much later (Bogoraz 1931). In 1947 a classic description by V. I. Cincius (1947) was published, giving the first systematic account of the phonetics and morphology of Ewen. Another fundamental work, which has a marked diachronic bias, is Johannes Benzing’s Lamutische Grammatik (Benzing 1955). The description of the Ola dialect by K. A. Novikova (Novikova 1960, 1980) remains the most detailed treatment of any Ewen idiom so far (see also Novikova 1997). In the subsequent years there appeared a number of works dealing with various aspects of the Ewen language, including phonetics and phonology (Sotavalta 1978, Hesche 1978, Lebedev 1978, Burykin 1986), the graphic system of the written language (Novikova 1958, Novikova & Lebedev 1980, Burykin 2004), morphology (Robbek 1982, 1984, 1992, Burykin 1984, Malchukov 1993, 1995, 2001), syntax (Gladkova 1991, Malchukov 1999, 2002), as well as lexicon (Petrov 1991, 1997, Danilova 1991). Dialectal studies include the monographic descriptions by the native Ewen scholars V. D. Lebedev (1978, 1982), V. A. Robbek (1989), and X. I. Dutkin (1995), as well as the earlier study by L. D. Rishes (1947). Sociolinguistic issues are addressed by Burykin (2004), while language contacts are discussed by Malchukov & Petrov (1994) and Malchukov (2003). Bilingual Russian-Ewen and Ewen-Russian dictionaries were published in the 1950s by V. I. Cincius and L. D. Rishes (1952, 1957). Their comprehensive Russian-Ewen dictionary contains also dialectal data. It later served as the basis for the Ewen-Russian-German dictionary of Doerfer & Scheinhardt & Hesche (1980). Over the last few decades the field of Ewen studies has grown still more international, with notable contributions by Japanese and Korean linguists to documentary and descriptive studies (Kazama 2003, 2019, Kim Juwon 2011). Progress has also been observed in the fields of phonetics (Aralova 2015), morphology (Sharina 2001), syntax and typology (Malchukov 2008, 2020a), dialect studies (Kuz’mina 2010, Sharina & Kuz’mina 2018, Kazama 2015, 2016–2020 2022, Matić 2020), language contacts (Malchukov 2006, Pakendorf 2013), as well as folklore and language documentation (Burykin 2001, Kazama 2009, 2013, Matić & Lavrillier 2013, Malchukov 2020b). A general sketch is offered by Pakendorf & Aralova (2020). In Russia, which remains central in the field of Ewen studies, research on Ewen is currently conducted at the Institute for Linguistic Research of the Russian Academy of Sciences and the Herzen Pedagogical University in St. Petersburg, as well as at the Department of Northern Philology of the North-Eastern Federal University in Yakutsk. SEGMENTAL STRUCTURE There are 18 vowel phonemes in Ewen, which fall into groups of “hard” (originally back) and “soft” (originally front) vowels, distributed in accordance with the vowel harmony rules. The vowels are further organized in terms of length and roundedness. The “hard” vowels are: (unrounded short) ɪ a, (unrounded long) ɪɪ aa ɪa, (rounded short) ʊ o, (rounded long) ʊʊ oo. The “soft” vowels are: (unrounded short) i e, (unrounded long) ii ee ie, (rounded short) u ɵ, (rounded long) uu ɵɵ. Thus, Ewen has preserved the original Proto-Tungusic vowel system, which has been considerably simplified in Ewenki. However,

Ewen 263

although this system of 18 phonemes is valid for most dialects, the phonetic realizations of the vowel qualities vary. The original distinction between the “hard” and “soft” vowels followed the palato-velar pattern, in that the “soft” vowels i u ɵ (< *i *ü *ö) were fronted as compared with the corresponding “hard” vowels ɪ ʊ o (< *ï *u *o). However, the eastern and, in particular, the north-eastern dialects have verticalized the distinctions within the pairs, yielding four distinctive levels of height (Table 11.1).

TABLE 11.1 EASTERN EWEN VOWELS u uu

 

ʊ ʊʊ

i ii ɪ ɪɪ

ɵ ɵɵ

e ee

ie

o oo

a aa

ɪa

This system was represented more adequately in the original Latin orthography of the 1930s, where ɪ was written as ⟨e⟩, and ɵ (= *ö) was (less consistently) rendered by ⟨y⟩, that is, it was not distinguished from u (also < *ö). Some other eastern dialects, in particular, the Ola dialect, have reinterpreted the opposition between the “hard” and “soft” series as one of pharyngealized (“hard”) vs. non-pharyngealized (“soft”) vowels (as described by Novikova), where pharyngealization may be related to tongue root retraction. On the other hand, the western dialects have preserved and even reinforced under Yakut influence the palato-velar basis of the opposition: in these dialects, the “soft” u has optionally the fully fronted value ü [y], as in untun ~ üntün ‘shaman’s drum’, while the “hard” ɪ is pronounced as a central vowel (hence the suggestions to represent it by Russian ⟨ы⟩ in the writings of Novikova and Lebedev). This discrepancy in the interpretation of the vowel values in the different dialects makes it difficult to give a unified representation of the vowel system of Ewen. The distinction between short and long vowels is phonemic in Ewen, although it is not represented in the current orthography. The vowels ɪa (< *ïa) and ie (< *ie) are phonetically represented as rising diphthongs; with respect to vowel length they are close to long vowels, but they have no short counterparts. In non-initial syllables short a and e undergo reduction, and in many dialects (especially eastern) they may further merge into a neutral [ə], though this positional neutralization does not have paradigmatic consequences. The high vowels i ii ɪ ɪɪ have a more open articulation when preceded by dorsal [t] and [d], and are rendered in the current orthography by ⟨ы⟩ in this position, but by ⟨и⟩ elsewhere. The tendency for vowel reduction in non-initial syllables has also resulted in the elision of final short vowels of polysyllabic words (which are preserved in Ewenki), cf. e.g. tog ‘fire’ < *toga, bak- ‘to find’ < *baka-, ƞen- ‘to go’ < *ƞene-, miir ‘shoulder’ < *miire, anƞan ‘year’ < *anƞanɪ. Note that this reduction of final vowels has yielded “secondary” stems with a stable stem-final n, which contrast with the diachronically primary stems with the primary unstable (*)/n. The consonant system is more homogeneous across Ewen dialects. It comprises 17 consonant phonemes, representing phonetically six places (bilabial, alveolar, palato-alveolar, velar, uvular, and laryngeal) and seven manners of articulation (nasals, weak/voiced stops, strong/voiceless stops, fricatives, glides, a lateral, and a rhotic) (Table 11.2).

264 Andrej Malchukov TABLE 11.2 EWEN CONSONANTS m

n

ñ

ƞ

b p

d

j

g

t

c

k

  [q]

[h]

s w

 

y

[γ]

l r

Concerning the articulation of the consonant phonemes, the following should be noted: The strong alveolar stop t is slightly aspirated. The palatal stops c and j are realized as affricates. The fricative s has an alveolar and a pharyngeal allophone, i.e., s [s] and x [h], whose distribution is subject to dialectal variation; furthermore, in certain dialects (e.g. the Omolon dialect, as described by Bogoraz) s has a “hushing” articulation. The strong velar stop k is realized as a uvular [q] in “hard-vowel” words and as a velar [k] elsewhere. The weak velar g is realized as a stop word-initially and after consonants, and as a fricative [γ] elsewhere. In the current Cyrillic orthography c is rendered by ⟨ч⟩, while j is rendered by ⟨д⟩ in combination with the “soft sign” ⟨ь⟩ in word-final position or before a consonant, as in ⟨удь⟩ = uj ‘trace’, and in combination with the “iotated” vowel letters ⟨я е ё и ю⟩ before a vowel, as in ⟨дю⟩ = jʊʊ ‘house’. The same conventions, derived from the rules of Russian orthography, apply for rendering the palatal nasal ñ. The Cyrillic orthography and the speech of fully bilingual individuals contain also marginal phonemes found only in recent borrowings from Russian. In dialectal speech they are substituted by native phonemes similar in articulation. Thus, [f] ⟨ф⟩ is rendered by p, as in pabrɪka ‘plant’ ← Russian fábrika, [z] ⟨з⟩ by c or x [h], as in coolʊta ‘gold’ ← Russian zóloto, [ʒ] ⟨ж⟩ by [h], as in xʊrnaal ‘magazine’ ← Russian zhurnál, [ts] ⟨ц⟩ by c or [h], as in cena ‘price’ ← Russian cená, [ʃ] ⟨ш⟩ by s, as in iskɵɵle ‘school’ ← Russian shkóla, and [χ] ⟨x⟩ by k, as in kolkoos ← Russian kolxóz. Thus, as in the case of the vowel system, the consonant system of Ewen is on the whole conservative; the innovations concern not the inventory of consonant phonemes, but, rather, their distribution and phonetic realizations. The most important diachronic innovation was the shift of word-initial consonants (in all varieties except Arman): *ldeveloped to n-, as in nɵki ‘arrow’ < *lökii > Ewenki lukii, *s- to [h-], as in xel ‘iron’ < *sele > Ewenki sele, *y- to ñ-, as in ñɵɵ- ‘to leave’ < *yöö- > Ewenki yuu-, and *w- to m-, as in maa- ‘to kill’ < *waa- > Ewenki waa-. PHONOTACTICS AND MORPHOPHONOLOGY There are no special restrictions, except for the vowel harmony rules, on the distribution of vowels. Consonant clusters (of at most two consonants) occur only in medial position. Due to the aforementioned consonant shifts, the fricative s, the glides w y, and the lateral l do not occur word-initially (except for the Arman variety). The rhotic r is also excluded from the initial position (as already in Proto-Tungusic). In the eastern dialects the distribution of the pharyngeal [h] is restricted to the word-initial position, whereas in the mid-western dialects [s] occurs only before palato-alveolar stops.

Ewen 265

The major phenomenon affecting the distribution of vowels in Ewen is vowel harmony. Vowel harmony, which has a different phonetic basis in the different dialects, implies restrictions on the co-occurrence of “hard” (originally back) and “soft” (originally front) vowels within a word form. Since the vowel quality (“hard” versus “soft”) of a suffix is determined by the vowels of the stem, each suffix has two distinct forms, one for use with “hard-vowel” stems and the other for use with “soft-vowel” stems. Compare, for instance, the form of the locative case marker -LA, which appears as -la when attached to the “hard-vowel” stem moo ‘wood’ : LOC moo-la vs. -le when attached to the “softvowel” stem mɵɵ ‘water’ : LOC mɵɵ-le. Labial assimilation is more characteristic of the western dialects, but it is sporadically attested elsewhere. It may be noted that labial assimilation, unlike the palatal harmony, is not always determined by the (rounded) value of the stem vowel, for it can also operate regressively, as in the case of umun < emun (< *eme-wu-n) ‘he brought’, as attested in the Ola dialect. Consonant alternations are conditioned by assimilation processes. Progressive assimilation occurs more regularly and applies, in particular, in the following cases: If a voiceless stem-final consonant is followed by a suffix beginning with a voiced consonant, the latter becomes voiceless, as in the dative marker -DU, e.g. jʊʊ ‘house’ : DAT jʊʊ-dʊ vs. okaat ‘river’ DAT okaat-tʊ. If a stem-final non-nasal alveolar consonant is followed by a verbal suffix beginning with r, the latter undergoes complete assimilation, as in the aorist marker -RA(-), e.g. 1SG nan-ra-m ‘I sent’, xat-ta-m ‘I call’, gɪd-da-m ‘I spear’, ɪs-sa-m ‘(I) reached’. If a suffix-initial w is followed by a long vowel, it is assimilated by the preceding stem-final consonant, while the vowel of the suffix becomes rounded, as in the iterative marker -WAAC-, e.g. oo-waat- ‘to do usually’, xɵr-rɵɵt- ‘to go usually’, bakkoot- ‘to find usually’, bel-lɵɵt- ‘to help usually’, etc. If a stem-final nasal is followed by the suffix-initial l, the latter undergoes nasalization, as in the locative case marker -LA, e.g. nam ‘sea’ : LOC nam-na vs. okaat ‘river’ : LOC okaat-la. Further, certain assimilation patterns vary across dialects. Thus, in most eastern dialects r can be fully assimilated by the preceding sonants l n, while the mid-western dialects show only partial assimilation with d as the result, e.g. (eastern) nanra ~ nanna vs. (western) nanda ‘hide, fur’, (eastern) olra ~ olla vs. (western) olda ‘fish’. Apocope is restricted to certain morphological and/or lexical contexts. Thus, a stem-final unstable /n is deleted when followed by a suffix beginning with p b j k or a sonant, cf. e.g. the base form oran ‘reindeer’ with the forms PL ora-r, DIR.LOC ora-kla, DER ora-pcɪ ‘with plenty of reindeer’, etc. Further, the unstable /n merges with a following g and w (< *b) into ƞ and m, respectively (note that the preceding vowel is here unstable, as well), e.g. oran ‘reindeer’ : ELAT or-ƞɪc (< *oran-gïïjï) : ACC ora-m (< *oran-ba). Stems with a stable final n, which historically is a result of the reduction and loss of short final vowels, are not affected by these alternations, cf. e.g. anƞan ‘year’ (< *anƞanɪ) : ELAT anƞan-gɪc : ACC anƞan-ʊ. On the other hand, if a stem-final primary nasal /n is followed by a suffix-initial r, the latter undergoes elision in certain verbal forms, as in the aorist stem, e.g. AOR gɵɵn-e-m ‘I said’ vs. AOR em-re-m ‘I came’. This latter feature, shared by Ewenki, goes back to, at least, Proto-Ewenic, while most of the other phenomena discussed represent developments specific the Ewen only. Epenthetic vowels a e i, historically derived from word-final vowels which underwent loss elsewhere, are used to avoid consonant clustering. Normally, they occur at morpheme boundaries, as in okaat ‘river’ : PL okaat-a-l, bey ‘man’ : PL bey-i-l. However, in a few cases these vowels can also “appear” in diachronically complex suffixes, as in the directive case marker -tkI ~ tAkI, e.g. jʊʊ ‘house’: DIR jʊʊ-tkɪ, bey ‘man’ : DIR bey-teki.

266 Andrej Malchukov

Ewen stress has not been experimentally studied and its characteristics (place, type) remain controversial. All authors seem to agree that stress is attracted to heavy syllables, meaning syllables closed by a consonant or syllables containing a long vowel or a diphthong. Some authors state that the primary stress is on the last syllable unless attracted to a heavy first syllable (Novikova), while others define the stress pattern as an expiratory stress on the first syllable with a possible pitch accent on the last syllable (Benzing, Menges). The latter view is more consistent with the observed reduction of short vowels in non-initial syllables. In any case, stress is not a phonemic property in Ewen. WORD STRUCTURE AND WORD CLASSES Morphologically, Ewen is an agglutinating language, using exclusively suffixation. There are, however, certain deviations from strict agglutination. Thus, nouns and verbs fall into distinct inflection and conjugation types, which are only partially phonologically conditioned. Moreover, in the case of verbal conjugation, segmentation into categories of tense/mood and agreement is problematic. Apart from agglutinating suffixation, Ewen makes use of the following morphological processes: (1) vowel and consonant alternations in stems, as in cases involving the alternation of the stem‑final consonant in order to express the aspect and valency of onomatopoetic verbs, e.g. bʊta.l- ‘to break’ TR, bʊta.r‘to break’ ITR, bʊta.k- ‘to be breaking’ TR, bʊta.m- ‘to be breaking’ ITR (such examples could also be viewed as sets of correlative derivatives from an otherwise unattested root); (2) periphrastic expressions, as in the periphrastic expression of negation in verbs; (3) reduplication, as in the derivation of onomatopoetics; (4) compounding, as in the formation of complex numerals, e.g. digen-mɪar ‘forty’ from digen ‘four’ and PL mɪa-r ‘ten’; and (5) conversion, as in the derivation of temporal adverbs from the corresponding nouns, e.g. xiisecen ‘evening’ > ‘in the evening’. The primary distinction that is relevant for the Ewen parts-of-speech system is that between nominals and verbs. The nominal parts of speech include nouns (proper), adjectives (which are not fully differentiated from nouns), pronouns, numerals, and adverbs. Adverbs are normally nominal or denominal words and have partially retained traces of nominal inflection (as in the case of locational adverbs). Verbs are represented by both finite and non-finite (participial and converbal) forms. There is also a category of syntactic words, represented by the closed classes of enclitic particles and postpositions. Finally, interjections, together with onomatopoetics, constitute a class of their own. WORD FORMATION Derivation and inflection are not strictly differentiated, in that there are numerous fully productive markers which do not ostensibly change the lexical meaning of the stem and are not obligatory, as in the case of evaluative suffixes on nouns and aspectual markers on verbs. In the following we take a restrictive approach to inflection and will regard all categories that are not obligatorily expressed as derivational. In the case of verbs, the expression of a category on the negative auxiliary, rather than on the lexical verb, can be taken as diagnostic for the inflectional status of a morphological element. The most frequently occurring derivational suffixes of nominals (nouns and adjectives) and verbs, with examples, are listed below (N and V stand for the nominal and verbal stem feeding the derivation): (1a) Denominal nouns: -nkAn [inhabitant of N], e.g. gorod ‘town’ : goroda-nkan ‘town inhabitant’; -rUk [receptacles for N], e.g. mɵɵ ‘water’ : mɵɵ-ruk ‘bucket’; -ƞAAt

Ewen 267

[material for N], e.g. moomɪ ‘boat’ : moomɪ-ƞaat ‘material for a boat’; -s [fur/hide of N], e.g. ɵliki ‘squirrel’ : ɵliki-s ‘squirrel fur’; -g, -kA-g, -sA-g [collective], e.g. jʊʊ ‘house’ : jʊʊ-sag ‘village’. Some other denominal suffixes, mostly with an evaluative and referential function, are functionally similar to inflectional. They include, in particular, the following forms: -kAn, -cAn [diminutive], e.g. jʊʊ ‘house’ : jʊʊ-kan ‘little house’; -kAAyA, -ñjA, -mkAAr [augmentative], e.g. bey ‘man’ : bey-i-mkeer ‘giant’; -mdAs [similative], e.g. jʊʊ ‘house’ : jʊʊ-mdas ‘(looking) like a house’; -dmAr [contrastive-emphatic], e.g. jʊʊ-dmar ‘this very house’; -nrAg [restrictive], e.g. jʊʊ-nrag ‘only this house’; -tAn [distributive], e.g. jʊʊ-tan ‘every house’; -ƞ- [alienable possession], e.g. PX1SG jʊʊ-w ‘my house’ : AL-PX1SG jʊʊ-ƞ-ʊ ‘my temporary house’. (1b) Denominal adjectives: -dI, [characteristic of N], e.g. asɪ ‘woman’ : asɪ-dɪ (teti) ‘women’s (coat)’; -mdA [related to N/designated for N], e.g. ʊnta ‘shoe/s’ : ʊntamda pabrɪka ‘shoe-factory’; -lkAn [proprietive], e.g. jʊʊ ‘house’ : jʊʊ-lkan ‘having a house’; -lA [connegative proprietive], in combination with the prepositional negative particle ac, e.g. jʊʊ ‘house’ : ac jʊʊ-la ‘without a house, homeless’; -gdA [having a big N], e.g. miir ‘shoulder(s)’ : miire-gde ‘broad-shouldered’; -pcI ~ -pcUn [plural proprietive], e.g. oron ‘reindeer’ : oro-pcʊn ‘(one) having many reindeer’. The possessive derivatives (proprietive and others) are primarily used as attributes referring to a property, but they may also refer to the possessor of a property. (2) Deverbal nouns (nominals): -n, -nmAy [action nominal], e.g. taƞ- ‘to read’ : taƞa-n, taƞa-nmay ‘reading’; -JAk, -CAk [place for V-ing], e.g. begde- ‘to cure’ : begde-cek ‘hospital’; -mkIn [way to a place for V-ing], e.g. buyus- ‘to hunt’ : buyuse-mkin ‘way to a hunting place’; -ƞkA, -ldIwUn [instrument], e.g. taƞ- ‘to read’ : taƞa-ƞka, taƞaldɪwʊn ‘book’; -mƞa, -j [agentive], e.g. taƞ- ‘to read’ : taƞa-mƞa ‘reader’, hotʊ- ‘to lead’ : hotʊ-j ‘one who shows the way’; -lAn [good at V-ing], e.g. buyus- ‘to hunt’ : buyuse-len ‘a good hunter’; -mAt [(recent) resultative], e.g. balda- ‘to be born’ : balda-mat ‘baby; newly born’; -tlA [(distant) resultative], e.g. jeb- ‘to eat’ : jebe-tle ‘what was eaten; left-overs of food’. The last three markers are mostly used attributively, and for some dialects have been qualified as participles. (3) Denominal verbs: -ltA-, -g- [produce N], e.g. ʊnta ‘shoe(s)’ : ʊnta-lta- ‘to make shoes’, xootara/n ‘road’ : xootara-g- ‘to make a road’; -tI- [consume N], e.g. kam ‘yukola [dried fish]’ : kam-tɪ- ‘to eat yukola’; -dA-, -rA-, -lA- [to act with N], e.g. tʊrkɪ ‘sled’ : tʊrkɪ-da- ‘to go with a sled’, gɪd ‘spear’ : gɪd-la- ‘to strike with a spear’; -tA- [to have N (as a kin)] akan ‘elder brother’ : akan-ta- ‘to have an elder brother’; -lAC- [to have/use as N], e.g. jʊʊ ‘house’ : jʊʊ-lat- ‘to have a house’; -lbA- [to acquire N], e.g. oron ‘reindeer’ : oro-lba- ‘to acquire reindeer’; -lA- [to go to fetch N], e.g. bokat ‘nut(s)’ : bokat-la- ‘to go to fetch nuts’; -mA- [to go to fetch N/hunt for N], e.g. ɵl- : ɵliki ‘squirrel’ : ɵl-me- ‘to go to hunt for squirrels’. (4) Deverbal verbs: These can have a modal, aspectual or valency-changing function. Modal categories: -m- [desiderative], e.g. xɵr- ‘to go (away)’: xɵre-m‑ ‘to want to go’; -scI- [conative], e.g. xɵre-sci‑ ‘to try to go’; -nA- [purposive-directional = andative], e.g. ic- ‘to see’ : it-ne‑ ‘to go to see’. Aspectual categories: -D‑ ~ -JId- [progressive], e.g. xɵr- ‘to go (away)’ : xɵre-d- ‘to be going’; -jAAn [durative], e.g. xɵre-jeen‑ ‘to go for a long time’); -l- [inchoative], e.g. xɵre-l- ‘to begin to go’; -C(I)‑ [resultative], e.g. ɪl‑ ‘to stand up’ : ɪla-t‑ ‘to stand’, also [imperfective], e.g. bɵɵ‑ ‘to give’ : bɵɵ-t‑ ‘to be giving, to give several times, to distribute’; -sAn—~ -s- [momentative] when derived from telic, e.g. xɵr- ‘to go (away)’ : xɵre-sn‑ ‘to go for a while’, [inchoative] or [delimitative] from atelic, e.g. xukle‑ ‘to sleep’ : xukle-sne- : AOR.3SG xukle-sne-n

268 Andrej Malchukov

‘(s/he) slept for a while/went to sleep’; -kAC- [distributive] when derived from telic, e.g. kɵke‑ ‘to die’ : kɵke-ket‑ ‘to die one after another’, [multiplicative or sepetive] from atelic, e.g. xoƞ- ‘to cry’ : xoƞ-kat‑ ‘to whimper’; -WAAC- [iterative], e.g. xɵr‘to go (away)’ : xɵr-rɵɵt-te-n‑ ‘to go usually’; -G(A)rA- [habitual], primarily in reference to past, e.g. xɵr-ger‑ ‘used to go’. Voice/valency-changing categories: -B- [medio-passive], e.g. aaƞa- ‘to open’ : aaƞa-b- ‘to be open’; -W‑ [adversative passive], e.g. maa- ‘to kill’ : maa-w- ‘to be killed’, ʊdal- ‘to rain’ : ʊdala-w- ‘to be caught by rain’; -WkAn- [causative], e.g. xɵr- ‘to go (away)’ : xɵr-uken- ‘to make/ let go’, ic- ‘to see’ : ic-uken- ‘to make/let see, to show’; -mAC- [reciprocal], e.g. coor-‘to hit’ : coor-mat- ‘to hit each other’; -ldA- [cooperative], e.g. xɵr- ‘to go (away)’ : xɵre-lde- ‘to go together’, less frequently [reciprocal], e.g. bak- ‘to find’ : baka-lda- ‘to meet’. NUMBER AND CASE Nouns inflect, in this order, for number, case, and possession (either personal or reflexive), e.g. jʊʊ ‘house’ : PL-LOC-PX3PL jʊʊ-l-dʊla-tan ‘in their houses’. For the category of number, Ewen makes a distinction between the singular, which is unmarked, and the plural, which is marked by the suffixes -l (default) or -r (for nouns with the stem-final unstable /n, which elides here), e.g. jʊʊ ‘house’ : PL jʊʊ-l ‘houses’, xɪrkan ‘knife’ : PL xɪrka-r ‘knives’. A restricted number of nouns (largely, kinship terms) take the special (historically complex) plural markers -sAl, -tAl, -nIl, e.g. aman ‘father’ : PL am-tɪl ‘fathers; parents’, ewen ‘Ewen’ : PL ewe-sel ‘the Ewen people’. Plural markers are restricted to referential-specific NPs. Also, in combination with numerals a noun may stand in singular if inanimate and/or indefinite. The category of case involves in all dialects the unmarked nominative case and 11 suffixally marked cases: accusative, designative, dative, locative, ablative, elative, prolative, directive, instrumental, comitative, and equative (Table 11.3).

TABLE 11.3 EWEN CASE MARKERS   ACC DES DAT ABL LOC PROL DIR ELAT INSTR COM EQU

V/C -w : -wA-gA-dU -dUk : -dUk-U-lA -lI -tkI -gIc : -gIj-I-c : -j-I-ñUn -gcIn

PL -bU

-jɪ

/n -m : -mA-ƞAn-dU n-dUk : n-dUk-Un-dUlA n-dUlI n-tAkI -ƞIc : -ƞIj-I-ñ : -ñ-I-

-gAcIn

-ƞcIn

-dUlA -dUlI -tAkI

Stem types: V/C = default vowel and consonant stems, /n = primary nasal stems, PL = plural stems. Segmental alternations: A = a e, U = ʊ u, I = ɪ i.

Ewen 269

Depending on the form of the case suffixes, three declension types of nouns (and nominals, in general) can be distinguished: (1) nouns with a stem-final vowel or consonant, with the exception of the primary /n, (2) nouns ending in the primary (unstable) /n, and (3) nouns incorporating the plural suffixes -l or -r. Of these, type (1) may be viewed as the default class, which comprises the majority of all nominal stems. Sample paradigms: jʊʊ ‘house’ (default stem) : ACC jʊʊ-w : -wa- : DES jʊʊ-ga- : DAT jʊʊ-dʊ : ABL jʊʊ-dʊk : -dʊk-ʊ- : LOC jʊʊ-la : PROL jʊʊ-lɪ : DIR jʊʊ-tkɪ : ELAT jʊʊ-gɪc : -gɪj-ɪ- : INSTR jʊʊ-c : -j-ɪ- : COM jʊʊ-ñʊn : EQU jʊʊ-gcɪn. xɪrka/n ‘knife’ (nasal stem) : ACC xɪrka-m : -ma- : DES xɪrka-ƞA- : DAT xɪrkan-dʊ : ABL xɪrkan-dʊk : -dʊk-ʊ- : LOC xɪrkan-dʊla : PROL xɪrkan-dʊlɪ : DIR xɪrkan-takɪ : ELAT xɪrka-ƞɪc : -ƞɪj-ɪ- : INSTR xɪrka-ñ : -ñ-ɪ- : COM xɪrka-ñʊn : EQU xɪrka-ƞcɪn. jʊʊ-l ‘houses’ (plural stem) : ACC jʊʊ-l-bʊ : DES jʊʊ-l-ga- : DAT jʊʊ-l-dʊ : ABL jʊʊ-ldʊk : -dʊk-ʊ- : LOC jʊʊ-l-dʊla : PROL jʊʊ-l-dʊlɪ : DIR jʊʊ-l-takɪ : ELAT jʊʊ-l-gɪc : -gɪj-ɪ- : INSTR jʊʊ-l-jɪ : COM jʊʊ-l-ñʊn : EQU jʊʊ-l-gacɪn. The nominative, accusative, designative, and comitative may be classified as “syntactic” cases, while the other cases are “semantic”, that is, the former, unlike the latter, can mark the subject and/or the direct object functions. Exceptionally, the designative case, which historically corresponds to the partitive case in Ewenki, is only used in combination with the markers of possession (either personal or reflexive), e.g. jʊʊ ‘house’ : DESPX3SG jʊʊ-ga-n : DES-RX jʊʊ-ga-y. •

The nominative, which is formally unmarked (and not indicated in glosses), is used to mark subjects, but also direct objects when followed by the reflexive possessive marker, as in [1].

[1] etiken or-mɪ xepken-ni old.man reindeer-rx  catch-aor.3sg ‘The old man caught his (own) reindeer.’ •

When no reflexive possessive marker is present, the direct object is marked by the accusative, as in [2]. The accusative is also used to indicate temporal location/duration, e.g. tugeni-w ‘during the winter’ (winter-ACC), ere-w ineƞi-w ‘today’ (this-ACC day-ACC).

[2] etiken oro-m xepken-ni old.man reindeer-acc catch-aor.3sg ‘The old man caught the reindeer.’ •

The designative case in Ewen is more similar to the corresponding form in Nanai than to the Ewenki partitive, which also has designative uses with possessive markers, but which basically indicates an indefinite or unspecific object (without possessive markers). Thus, the designative case in Ewen performs a double function, in that it marks the direct object and, simultaneously, assigns the beneficiary function to its possessor, as in [3]:

[3] bi bey or-ƞa-n emu-re-m 1sg  man reindeer-des-px3sg bring-aor-1sg ‘I brought a reindeer for the man.’

270 Andrej Malchukov

Some other uses of the designative case include the marking of a created object, as in [4], or of an object complement, as in [5]: [4] bey  jʊʊ-ga-y   oo-n man house-des-rx make-aor.3sg ‘The man made a house for himself.’ [5] ere-w asatka-m atɪka-ƞ-ga-y  ga-lɪ this-acc girl-acc wife-al-des-rx take-imp2sg ‘Take this girl as your wife!’ Exceptionally, the designative case can also mark subjects of intransitive (mostly existential) predicates, such as, for instance, the negative existential acca, e.g. xut-ke-ku acca ‘I have no children’ (child-DES-PX1SG NEG.EXIST). This usage reminds of Ewenki, where the partitive is also used in combination with the negative existential noun. •

The comitative normally marks the dependent constituent within a complex subject NP, but it can be also used to mark a comitative object; note the variation in the verbal agreement, as in [6]:

[6] etiken atɪka-ñʊn bi-weet-te ~ bi-weet-te-n old.man old.woman-com be-iter-aor.3pl ~ be-aor.3sg ‘There lived an old man with an old woman.’ •

The dative is used to indicate a wide range of functions. Animate nouns take the dative to indicate the addressee (but not the addressee of speech, which is in the directive), as in [7]:

[7] bi  tʊrkɪ-w  xurken-du bɵɵ-ri-w 1sg sled-acc youth-dat give-pst-1sg ‘I gave the sled to the youth.’ Inanimate nouns may take the dative case-marking to indicate static location. In this function, the dative competes with the locative, as in [8]: [8] bey jʊʊ-dʊ ~ jʊʊ-la  bi-s-ni man house-dat ~ -loc be-aor-3sg ‘The man is in the house.’ As compared to Ewenki, the dative is more restricted in the locative use: it is mostly used to indicate a temporary location and normally combines with animate subjects. It can also perform a temporal function, attaching to nouns with a temporal reference, e.g. tugeni-du ‘in winter’ (winter-DAT). Finally, it can mark an agent of certain derived constructions, such as causative and passive. •

The locative serves as a general marker of static location. Unlike its counterpart in Ewenki, it is only restrictedly used in the (al)lative function, mostly in combination with goal-oriented motion verbs like em- ‘to come’, mʊcʊ- ‘to return’, ɪs- ‘to reach’.

Ewen 271

•

Used with temporal nouns, the locative performs a temporal function, being synonymous with the temporal use of the dative and prolative cases. The prolative, which is formally connected with the locative, is used to indicate motion through space or along a route, e.g. okaat-lɪ ‘along a river’ (river-PROL), or also a period of time, e.g. jɵɵr caas-lɪ ‘in two hours’ (two hour-PROL). Another local case is the directive, which indicates movement towards a goal in both a locative and a temporal sense. Additionally, verbs of perception, e.g. ic- ‘to look’, assign the directive to the object of perception, whereas verbs of speech, e.g. gɵɵn- ‘to say’, assign it to the addressee of speech. The semantic difference between the directive and the locative is that the former indicates only direction, while the latter carries an additional implication that the goal would be reached in the course of motion, as in [9]:

[9] jʊʊ-tkɪ-y  vs. jʊʊ-la-y mʊcʊ-rɪ-n house-dir-rx house-loc-rx return-pst-3sg ‘He was returning towards his home.’ vs. ‘He returned to his home.’ •

•

•

Of the two cases with a separative meaning, the ablative indicates the source of motion, e.g. ɪsag-dʊk mʊcʊ-rɪ-n ‘s/he returned from the forest’ (forest-ABL returnPST-3SG), or the source argument co-occurring with verbs of taking, e.g. ga- ‘to take’. With animate nouns it can also indicate the source (the causer) of an event/ process, while with inanimate (and, predominantly, non-count) nouns it marks material. In the comparative construction the ablative is used to mark the object of comparison. The elative is functionally similar to the ablative, in that it also denotes movement away from a source, but it lacks the implication of an actual source point, e.g. okaat-kɪc em-re-n ‘came from (the direction of) the river’ (river-ELAT comeAOR-3SG). Thus, it forms a separative counterpart of the directive case. The instrumental is used to denote an instrument or means (in particular, means of transportation), e.g. tʊrkɪ-c em-ri-w ‘I came by sled’ (sled-INSTR come-PST-1SG). In some dialects the instrumental extends its use to indicate material, as well, competing with the ablative in this function. Verbs of emotion assign the instrumental to the source of the emotional reaction, e.g. nakata-c ƞeel-re-m ‘I am afraid of the bear’ (bear-INSTR fear-AOR-1SG). The equative marks an “object of equation”, as in [10]. In the specialist literature, the equative is not always considered to be a true case, but it qualifies as a case on morphological grounds: it occurs in the case slot in a word form and is incompatible with other case-markers.

bey-gecin tɵɵre-l-re-n [10] ʊcaka-n saddle.reindeer-px3sg m an-equ sp eak-inch-aor-3sg ‘His saddle reindeer started speaking like a human.’ •

In the eastern dialects two other cases are found: the directive-locative in -k-lA and the directive-prolative in -k-lI, both of which have formal and functional counterparts in Ewenki. The former indicates a spatial or temporal limit, e.g. okaata-k-la ‘up to the river’ (river-DIR-LOC), xiseeci-k-le ‘until evening’ (evening-DIR-LOC), while the latter is similar in function to the prolative but carries an additional implication of objects coming in close contact, e.g. okaata-k-lɪ tut-te-n ‘(s/he) runs just along

272 Andrej Malchukov

•

the river’ (river-DIR-PROL run-AOR-3SG). In the western dialects a corresponding spatial/temporal limit is expressed through a periphrastic construction involving the postposition ɪstala ‘up to’. An additional element reminiscent of a case marker is the “genitival” emphatic possessive marker -ƞI, used under inversion or in predicative position, as in [11]:

[11] erek oran etike-ƞi this reindeer old.man-poss ‘This reindeer is the old man’s ~ belongs to the old man.’ ADJECTIVES There is no clear-cut distinction between nouns and adjectives in Ewen, although certain derivational suffixes are specific for adjectives. Morphologically, adjectives (adjectival nominals) are similar to nouns, in that they can inflect (under agreement) for number and case, e.g. gʊda-l-dʊla jʊʊ-l-dʊla ‘in (the) big houses’ (big-PL-LOC house-PL-LOC). On the other hand, with regard to their syntactic distribution adjectives are, effectively, indistinguishable from non-count nouns. That is, both can be used both in an argument and a modifier position. For example, the noun xel means ‘iron’, but also ‘of iron’, e.g. xel gɪd ‘iron spear’, while nood means ‘beautiful’, but also ‘beauty’ when followed by a possessive suffix, e.g. nood asatkan ‘(a) beautiful girl’ vs. asatkan nooda-n ‘(the) girl’s beauty’. Adjectives expressing core adjectival meanings (such as size and color) can take intensity markers, such as -mAkAn, -mkAr, -dmAr for high intensity and -mrIn, -sUkAn for reduced (moderate) intensity, e.g. xʊlaña ‘red’ : INTEN xʊlaña-makan ‘intensive red’ : MODER xulaña-sukan ‘reddish’. There is no special comparative form of an adjective: the base form combines meanings of both the positive and the comparative degree. The latter meaning is obtained within the comparative construction, where the standard of comparison stands in the ablative case and the adjective optionally takes an intensity marker, as in [12]: eƞi-dmer [12] ƞeeluki ƞɪn-duk wolf dog-abl strong-intens ‘The wolf is stronger than the dog.’ The meaning of the superlative degree can be expressed either morphologically through the use of the appropriate possessive ending in combination with the alienable possessive marker, or lexico-syntactically by the ablative forms of the quantifier PX3SG cele-n ‘all’ in the position of the standard of comparison, as in [13]: gʊda-ƞa-tan ~ cele-duku-n gʊd [13] erek oran this reindeer high-al-px3pl all-abl-px3sg high ‘This reindeer is the highest (of them all).’ NUMERALS Numerals inflect for case (some classes of numerals for number and possession, as well) and are mostly used as attributive modifiers. Five major classes of numeral forms can be distinguished: cardinal, approximate cardinal, ordinal, distributive, and multiplicative.

Ewen 273

The simple cardinal numerals are: 1 ɵmen, 2 jɵɵr, 3 ɪlan, 4 digen, 5 tʊnƞan, 6 ñuƞen, 7 nadan, 8 japkan, 9 uyun, 10 mɪan, 100 ñama. Tens are derived by compounding, involving the plural form of the numeral mɪan ‘ten’ as the second member of the compound: ɪlan-mɪa-r ‘thirty’ (three-ten-PL). The numerals above ten as well as hundreds are formed periphrastically, combining the corresponding names for hundreds, tens and digits, e.g. 211 jɵɵr ñama mɪan ɵmen (2 × 100 + 10 + 1). In the dialects the numerals above ten are also expressed through another perihphrastic construction making use of the “postpositional nouns” xulek ‘extra part’ or ɵy-du-n ‘above’ (upper.part-DAT-PX3SG), e.g. 11 mɪan ɵme-ñ xulek (ten one-INSTR above) ~ mɪan omen ɵydun (ten one above). The cardinals serve as the base for the derivation of the other classes of numerals: the approximate cardinals in -klI, e.g. mɪa-klɪ ‘about ten’, the ordinals in -gI‑~ -I‑ followed by a possessive ending, e.g. jɵɵr-i-ten ‘the second’ (two-ORD-PX3PL), the distributives in -tAl, e.g. ɵme-tel ‘one by one’, and the “tuples” in -rmAn, e.g. ɪla-rman ‘three-fold, of three components’. PRONOUNS There are about a dozen distinct pronominal classes in Ewen. On both derivational and functional grounds these classes may be divided into five major groups: personal, reflexive, demonstrative, interrogative, and indefinite. Among the basic personal pronouns (Table 11.4) a distinction is made between personal proper and possessive personal pronouns. TABLE 11.4 EWEN PERSONAL PRONOUNS   SG

1 2

PL

1 2

EXCL

NOM

OBL

ACC

GEN

POSS

bii

min-

min-u

min

min-ƞi

xii

xin-

xin-u

xin

sin-ƞi

buu

mun-

mun-u

mun

mun-ƞi

xuu

xun-

xun-u

xun

sun-ƞi

The complete set of personal pronouns is: 1SG bii : OBL min-, 2 SG xii : OBL xin-, 3SG 3SG noƞan (= noƞa-PX3SG) : OBL noƞ-CX-PX3SG (in the accusative, elative, comitative, and equative cases) ~ noƞan-CX-PX3SG (in the other cases), 1PL EXCL buu : OBL mun-, 1PL INCL mut : OBL mut-, 2PL xuu : OBL xun-, 3PL noƞartan (= noƞa-PLPX3PL) : OBL noƞar-CX-tan. In the western dialects the exclusive form buu is in disuse, or is merged with the corresponding inclusive form mut to yield but ~ büt, while Arman retained only the exclusive pronoun in the form buuye. For logical reasons, the paradigms of the personal pronouns lack the designative case. The possessive function is expressed by two sets of forms: simple (GEN) and emphatic (POSS). The simple forms are synchronically homophonous with the oblique stems of the personal pronouns, i.e., GEN 1SG min ‘my’ : 2SG xin ‘thy’ : 1PL EXCL mun ‘our’ : 2PL xun ‘your’, while the emphatic forms incorporate the possessive marker -ƞI, i.e., POSS 1SG min-ƞi ‘mine’, 2SG xin-ƞi ‘thine’ : 1PL EXCL mun-ƞi ‘ours’ : 2PL xun-ƞi ‘yours’. The simple possessive forms represent heritage from Proto-Tungusic (SG1 *min-i : SG2 *sin-i : PL1.EXCL *mön-i : PL2 *sön-i) and correspond to similar forms in Neghidal and Southern Tungusic. The

274 Andrej Malchukov

accusative forms in -u apparently represent diachronically the same type of “partitive-accusative” complexes in *-A-wA as are observed in Siberian Ewenki. The reflexive pronoun is SG meen ‘self, one’s own’ : PL mee-r, whose oblique forms are always accompanied by the corresponding reflexive suffixes, e.g. SG DAT-RX meend-i ‘to oneself’ : PL-DAT-RX.PL mee-r-d-ur. Like the personal pronouns, the reflexive pronoun has no designative case form, but it also lacks the accusative case, for the basic reflexive stem in combination with the reflexive suffixes inherently implies object position, i.e., SG RX meen-i ‘oneself’ (as object) : PL-RX.PL mee-r-bur ‘oneselves’ (as object). Emphatic personal (logophoric) pronouns are derived from the reflexive pronouns by adding the possessive suffixes, e.g. SG PX1SG meen-mu ‘I myself’. The corresponding possessive pronoun contains the possessive marker -ƞI, i.e., REFL-POSS meen-ƞi ‘one’s own’. A reciprocal pronoun is formed from the plural form mee-r by the suffix -teke(n)-, combined with the plural reflexive suffix, i.e., PL-RECIPR-RX.PL meer-tek-mer ‘each other’. The demonstrative pronouns are based on the roots (proximal) e- vs. (distal) ta-, which are used as such in the local cases, e.g. DAT e-du : ABL e-duk : LOC e-le : PROL e-li, while the other case forms are based on the extended stems erek (= e-re-k) ‘this’ : OBL ere- vs. tarak (= ta-ra-k) ‘that’ : OBL tara-. The corresponding qualitative demonstratives are er-rɵɵcin ‘like this’ vs. tar-roocɪn ‘like that’. The class of interrogative pronouns includes the basic (pro-)verb ɪa- ‘to do what?’ and its (pro)nominal derivative ɪa-k ‘what?’, as well as ƞii ‘who?’, ir-rɵɵcin ‘what kind of?’ > ‘what color’, a-dɪ ‘how many?’, and a-sʊn ‘how big?’. Indefinite pronouns are derived from the interrogatives by means of the enclitic particles =Wul ~ =(G)Ul for non-referential indefinites, e.g. ƞii=wul ‘anybody’, as well as =WuttA ~ =(G)UttA and =dA for referential indefinites, e.g. ƞii=wutte, ƞii=de ‘somebody’. The pronouns in =dA are also used in a connegative function when in the scope of negation, cf. e.g. ƞii=de e-s-ni em-re ‘nobody came’ (who=CONNEG NEG-AOR-2SG come-CONNEG) vs. ƞii=wutte e-s-ni em-re ‘somebody has not come’ (who=INDEF NEG-AOR-3SG come-CONNEG). PERSON MARKING Like other Ewenic languages, Ewen has two sets of person markers: personal possessive markers (PX) and subject agreement markers (VX). Both sets distinguish three persons in singular and plural, with the additional distinction between the inclusive and exclusive forms in the first person plural. There is also a set of singular and plural reflexive possessive markers (RX), used under coreferentiality of the possessor (Table 11.4).

TABLE 11.5 EWEN PERSON MARKERS   SG

VX

PX

AOR

C

V

/n

PL

SUBJ

1

-m

-U

-w

-mU

-bU

-U

2

-nrI

-A-

-s

-sI

3

-n(I)

-A-

-n

-nI

-I

-y

-mI

RX

-s -Ø -bI

Ewen 275   PL

1

VX

PX

AOR

C

EXCL

-U

-WUn

INCL

-p

-A-

2

-s

-sAn

3

-Ø

-tAn

RX

V -t

-Ur -wUr

/n

PL

SUBJ

-mUn

-bUn

-l-bUn

-tI

-l-tI -l-sAn -l

-mUr

-bUr

Stem types: V = vowel stems, C = consonant stems, /n = primary nasal stems, PL = plural stems. Segmental alternations: A = a e, U = ʊ u, I = ɪ i.

The possessive and reflexive markers show allomorphic variation corresponding to the declension types of nouns, with distinct forms used for regular stems (default), stems ending in an unstable /n, and plural forms. If a possessive relation is regarded as temporary or conventional, the noun takes apart from the possessive endings the alienable possession marker -ƞ-, attached before the number marker, e.g. ulre ‘meat’ : PX1SG ulre-w ‘my flesh’ vs. AL-PX1SG ulre-ƞ-u ‘my meat (that I cook, eat, etc.)’, xut ‘child’ : PX1SG xut-u ‘my (own) child’ vs. AL-PX1SG xute-ƞ-u ‘my child (who actually belongs to somebody else)’. In this connection, it may be noted that the first person singular and plural exclusive possessive markers have the shapes -kU : -kUn when following the designative case marker, e.g. DES-PX1SG xutke-ku. The same shapes are also used in the “supine”, e.g. em- ‘to come’ : SUP-PX1SG em-de-ku ‘let me come!’. The presence of the extra consonant segment k in these cases is synchronically unmotivated. While nominals can only take the personal and reflexive possessive markers, finitely used verbs can be combined with both the personal possessive markers and the subject agreement markers. In practice, the subject agreement markers are restricted to the aorist type of inflection, used in the future and non-future tenses, while most other verbal forms used as finite predicates, notably those of the past tense and the hypothetical mood, but also, for instance, the “supine”, take the (default set of the) personal possessive markers. This situation, also known as “verbalization” or “insubordination”, is due to the historical circumstance that the possessively marked finite predicates are actually based on polyfunctional nominalizations of verbs. Note also that a somewhat special set of personal person markers is used in the forms of the preventive and subjunctive (SUBJ) moods. In the latter set the plural forms contain the nominal plural marker -l, while the third person singular form is unmarked for person. OTHER WORD CLASSES Most adverbially used words in Ewen are actually nominals. Many time adverbs are simply homophonous with the basic forms of the corresponding nouns, e.g. mɵntelse ‘early autumn’ > ‘in early autumn’, bajɪkar ‘morning’ > ‘in the morning’. Modal adverbs are mainly formed by adding the instrumental case marker to the corresponding adjective, e.g. ay ‘good’ : INSTR ay-ɪ-c ‘well’, or to a noun, e.g. gɪama/n ‘friendship’ : INSTR gɪama-ñ ‘(in a) friendly (manner)’. Quantifying adverbs are derived from numerals by

276 Andrej Malchukov

suffixes like -RakAn for multiplicatives (iteratives) and -RijUr for collectives, e.g. ɪlan ‘three’ : MULT ɪl‑rakan ‘three times’ : COLL ɪl-rɪjʊr ‘three together’. Locative relations are expressed by spatial nominals (relative nouns), e.g. doo ‘inner part; inside’, xerde ‘lower part; underneath’. Spatial nominals take the regular nominal case markers, but they normally have a defective paradigm restricted to the local cases, and they can also have irregular or idiosyncratic forms, cf. e.g. LOC xer-gi-le ‘underneath’, ELAT xer‑gic ‘from underneath’, DIR (irregular) xes-seki ‘down’, xer-gi-lte ‘one under another’. Functionally, some spatial nominals combine a locative and a temporal meaning, e.g. DIR ama-skɪ ‘back(wards); back in time, ago’, LOC jul-le ‘ahead; in the future’. Spatial nominals are also used as the formal heads of possessive-like postpositional constructions in which they take the corresponding possessive endings to indicate the person and number of the lexical noun, e.g. mugdeken xerde-du-n ‘under the stump’ (stump lower.part‑DAT-PX3SG). There are also a few postpositionally used adverbs based on lexicalized verbal forms, e.g. ɪs- ‘to reach’ : CV.TERM ɪs-tala, ereel- ‘to go around’ : NMLZ ereel-i ‘around’. These assign the required nominal case to the preceding noun, e.g. jʊʊ-la ɪstala ‘up to the house’ (house-LOC until), jʊʊ-w ereli ‘around the house’ (house-ACC around). Another group of verb-based forms are the resultative adverbs in -s, derived from a restricted number of onomatopoetic “destructive” verbs, e.g. teker- ‘to break off’ ITR : teke-s ‘to pieces’. Degree adverbs differ from other adverbs in that they are morphologically non-derived, e.g. xoo ‘very’, asʊkʊt ‘almost’. Even so, they are nominals and can take agreement suffixes, e.g. PL xoo-l. This leaves only interjections and onomatopoetics as a class of truly invariable independent words, e.g. erey ‘oh!’, kiree ‘fie!’, gele ‘come on!’, co&co [quieting an untamed reindeer], tag&tag [rattling sound], keeku&keeku [imitating the cuckoo]. All particles, except for the prenominal negative particle ac ‘without’, used in the negative proprietive construction, e.g. ac jʊʊ-la ‘without a house, homeless’, as well as ñaan ‘again, also’, are enclitics. They share features with both syntactic words and suffixes. Thus, they are similar to the latter in that they are subject to vowel harmony and assimilatory processes. The most frequent enclitics are: =dA emphatic, contrastive, coordinating; =k ~ =(k)kA emphatic (often with imperatives); =sI contrastive-adversative; =tIt concessive (often with the connective converb); =tAkAn restrictive; =GU interrogative; =WUl ~ =WUttA indefinite. VERBAL MORPHOLOGY Apart from person/number, the finite verb inflects for tense/mood. Additionally, verbs, both finite and non-finite, can take aspect and voice suffixes. The usual ordering of the markers of the verbal categories is: voice-aspect–tense/mood–person/number, e.g. maa-mac-coot-tɪ-tan ‘they used to kill each other’ (kill-RECIPR-HABIT-PST-3PL). Categories of aspect and voice are not paradigmatic, that is, several markers of these categories can co-occur in a word-form. The order of these markers is partially conventional (as described by Robbek for aspectual markers), but partially determined by differences in the semantic scope of the suffixes, cf. e.g. it-ne-wken- ‘to make (somebody) go and see’ (see-AND-CAUS-) vs. ic-uke-ne- ‘to go (in order) to make (somebody) see’ (see-CAUS-AND-). Verbs can be divided into four conjugational classes depending on the form of the non-future tense (aorist) in -RA-, as well as some related verbal forms with an initial R

Ewen 277

(*r). (1) The vast majority of verbs takes this suffix for the non-future tense in the form -RA-, although the initial R can be assimilated by the preceding consonant, cf. e.g. em- ‘to come’ : AOR em-re-, tet- ‘to dress’ ITR : AOR tet-te-, od- ‘to finish’ : AOR od-da-. This class of regular verbs is contrasted with three classes of irregular verbs: (2) Verbs with the stem-final primary /n take the suffix -A- in the aorist, e.g. gɵɵn- ‘to say’ : AOR gɵɵn-e-, xepken- ‘to catch’ : AOR xepken-e-. (3) A limited number of change-of-state verbs takes the marker -d(A)- in the aorist, e.g. oo- ‘to become’ : AOR oo-da-, ga- ‘to take’ : AOR ga-da-, nee- ‘to put’ : AOR nee-de-. Finally, (4) certain stative verbs, as well as derived statives taking the desiderative form in -m-, take the marker -s(A)‑ in the aorist, e.g. bi- ‘to be’ : AOR bi-se-, e- [negation verb] : AOR e-se-, xɵr- ‘to go (away)’ ITR : DES xɵre-m‘to want to go’ : AOR xɵre-m-se-. All these four classes go back to Proto-Tungusic and are shared by the other Ewenic languages. There are a few differences between the conjugational classes as to which forms of the personal paradigm actually contain the aorist marker. Thus, the markers -s(A)‑ and -d(A)- are present in all forms of the paradigm, while the marker -RA- is absent (unlike in Ewenki) in the second-person forms of both the singular and the plural, and also in the third person singular of vowel stems. For reasons of historical phonology (the loss of short vowels in final position), the marker -A- is physically absent in the third person plural. Following are examples of the aorist paradigms of the four conjugational classes: (1) xaa- ‘to know’ : 1SG xaa-ra-m : 2SG xaa-nrɪ : 3SG xaa-n : 1PL EXCL xaa-r-ʊ : 1PL INCL xaa-ra-p : 2PL xaa-s : 3PL xaa-r. (2) gɵɵn- ‘to say’ : 1SG gɵɵn-e-m : 2SG gɵɵn-e-nri : 3SG gɵɵn-ni : 1PL EXCL gɵɵn-u : 1PL INCL gɵɵn-e-p : 2PL gɵɵn-e-s : 3PL gɵɵn. (3) oo- ‘to become’ : 1SG oo-da-m : 2SG oo-da-nrɪ : 3SG oo-d-nɪ : 1PL EXCL oo-d-ʊ : 1PL INCL oo-da-p : 2PL oo-da-s : 3PL oo-d. (4) bi- ‘to be’ : 1SG bi-se-m : 2SG bi-se-nrɪ : 3SG bi-s-ni : 1PL EXCL bi-s-u : 1PL INCL bi-se-p : 2PL bi-se-s : 3PL bi-s. As may be seen, the third person plural of the aorist paradigm contains no actual person marker, but only the basic aorist stem. An exception is formed by the negation verb e- : e-se-, which takes the corresponding possessive suffix in this form, i.e., AORPX3PL e-s-ten. The unmarked aorist stem is also used as an invariable connegative form in combination with the fully conjugated negation verb, yielding a periphrastic negative paradigm, e.g. NEG AOR + CONNEG 1SG e-se-m xaa-r : 2SG e-se-nri xaa-r : 3SG e-s-ni xaa-r : 1PL EXCL e-s-u xaa-r : 1PL INCL e-se-p xaa-r : 2PL e-se-s xaa-r : 3PL e-s-ten xaa-r. While the construction with the basic negation verb e- : e-se- is also attested in most other Tungusic languages, Ewen has also a specific “modal connegative” in -ƞA, which is used in combination with other verbs with an inherently negative meaning, e.g. baa- ‘to be unwilling’, turku- ‘to be unable’. In the “modal connegative” construction, the head verb precedes the fully conjugated negation verb, e.g. xɵr- ‘to go (away)’ : xɵr-ƞe baa-n ‘he does not want to go’ (go-CONNEG.MOD be-unwilling-SG3), badʊ-ƞa turku-n ‘he cannot ride’ (ride-CONNEG.MOD be.unable-SG3). The other forms based on the aorist stem and showing the same four conjugational classes include the imperfective participle in -RI-, e.g. em-ri ‘coming’, gɵɵn-i ‘saying’, oo-di ‘becoming’, bi-si ‘being’; the anterior converb in -RIjI (derived from the imperfective participle), e.g. em-riji ‘after coming’, gɵɵn-iji ‘after saying’, oo-dɪjɪ ‘after

278 Andrej Malchukov

becoming’, bi-siji ‘having been’; and the aorist (conditional) converb in -RAk-, e.g. 1SG em-rek-u ‘if I come’, gɵɵn-ek-u ‘if I say’, oo-dak-ʊ ‘if I become’, bi-sek-u ‘if I were’. Apart from the finite forms (proper), which follow the aorist type of conjugation, and the connegative forms, which may be seen as invariable nominalizations or “infinitives”, the Ewen system of verbal morphology comprises the non-finite categories of participles and converbs. The distinction between finite and non-finite forms is, however, not clear-cut, as non-finite forms (in particular, participles), tend to be reanalysed as finite when used as final predicates. Therefore, the actual system of verbal forms in finite function involves a mixture of finite and non-finite elements. The formal distinction between forms of finite and non-finite origin is signalled by the different types of person marking they take (VX vs. PX). PARTICIPLES Participles are verbal forms combining both verbal and nominal features. Morphologically, participles are similar to verbs in that they inflect for voice, aspect, and negation, and convey temporal and modal meanings. Used in nominal positions participles indicate relative tense. On the other hand, participles are similar to nominals in that they display number agreement, as well as inflect for case and possession. In the sentence, participles can act as modifiers, independent head nouns, and as finite predicates. There are four distinct participial forms that occur in all dialects of Ewen: the imperfective, perfective, necessitative, and hypothetical participles. The imperfective participle is marked by the suffix -RI ~ -I ~ -dI ~ -sI, depending on the conjugational class of the verb, and is aspectually ambivalent. When derived from telic verbs, it indicates recent anteriority with regard to the primary event, whereas when derived from atelic verbs, it refers to simultaneity with the primary event, e.g. em- ‘to come’ : PTCP.IMPRF em-ri ‘(one) who came’, gɪrka- ‘to walk’ : PTCP.IMPRF gɪrka-rɪ ‘(one) who walks’. The perfective participle takes the marker -cA and refers to completed actions, e.g. PTCP.PRF em-ce ‘(one) who has come’. The necessitative participle in -nnA and the hypothetical participle in -Jiƞe have a modal reference, e.g. PTCP.NEC eme-nne ‘(one) who should come’ : PTCP.HYP em-jiƞe ‘(one) who may come’. In the dialects some other verbal forms have also been qualified as participles, such as, for instance, the forms in -dAƞ, -mAt, -tlA, -nA in the eastern dialects. However, many of these forms may alternatively be analysed as regular deverbal nouns, as they do not consistently manifest verbal features (e.g. the possibility to combine with the negative auxiliary). As in the other Tungusic languages, the participles in Ewen perform a wide range of functions, being used as attributive modifiers, e.g. em-ce bey ‘the man who has come’ (come-PTCP.PRF man), as sentential arguments/adjuncts, e.g. bii em-ce-we-n xaa-ra-m ‘I know that s/he came’ (1SG come-PTCP.PRF-ACC-PX3SG know-AOR-1SG), or as finite predicates, e.g. bii em-ce bi-se-m ‘I have (already) come’ (1SG come-PTCP.PRF be-AOR-1SG). When used as finite predicates the participles either combine with a copula in a nominal fashion (as in the example above), or take directly the agreement endings of the possessive type. The latter is particularly characteristic of the “modal” (necessitative and hypothetical) participles, whose finite occurrences have sometimes been qualified as special moods (Novikova), e.g. bii eme-nne-w ‘I must come’ (1SG come-PTCP. NEC-1SG), bii em-jiƞe-w ‘I may/will come’ (1SG come-PTCP.HYP-1SG). Furthermore, in the eastern dialects the imperfective participle in predicative use has been reanalysed as an actual finite past tense form (imperfect).

Ewen 279

FINITE TENSE AND MOOD The markers of tense and mood are located in the same slot in the verbal template of Ewen. Five moods may be distinguished: imperative, preventive, hypothetical, subjunctive, and indicative. The otherwise unmarked indicative mood incorporates the category of tense, with a distinction between future, non-future, and past, as discussed below. The imperative mood has three basic markers, cumulatively expressing person and number of the addressee: -LI for the second person singular imperative, -LI-lrA ~ -lrA for the second person plural imperative, and -GAr for the first person plural inclusive invitation, e.g. xɵr- ‘to go (away)’ : IMP.2SG xɵr-li ‘go!’ : IMP.2PL xɵr-lilre ‘go!’ : IMP.1PL. INCL xɵr-ger ‘let’s go!’. The rest of the heteroclitic imperative (voluntative-hortative) paradigm is filled by the “supine” function of the purposive converb in -dA- in combination with the personal possessive endings, e.g. taƞ- ‘to read’ : SUP-PX1SG taƞ-da-w ‘let me read!’ : 3SG taƞ-da-n ‘let him read!’. The “supine” in combination with the reflexive possessive endings is also used as a remote imperative/optative form for the second person, e.g. SG taƞ-da-y : PL taƞ-da-wʊr ‘read (later)!’. Although such uses of the purposive converb can be explained through ellipsis of the main verb in the imperative form, synchronically they are better analysed as distinct from converbs. Note in particular, that the reflexive forms can combine with the personal pronouns, e.g. em- ‘to come’ : 2SG xii em-de-y ‘you come (later)!’, and that the first person plural exclusive form in -dA-kUn is used as inclusive e.g. taƞ- ‘to read’ : 1PL taƞ-da-kʊn ‘let’s read’! There is also a “polite imperative” form in -ƞA-, available only in the second person, e.g. SG taƞa-ƞa-nrɪ : PL taƞa-ƞa-s ‘please read!’ The gaps in the polite imperative paradigms are filled through the use of the accusative case form of the hypothetical participle, e.g. PTCP.HYP-ACCPX3SG taƞ-dɪƞa.wa-n ‘(please) let him read!’ The latter forms are similar to those based on the purposive converb in that they also often express a delayed imperative meaning (‘later’). The preventive (“preventive-hypothetical”) mood in -JIk- (in dialects also in -k-) has a function resembling the “dubitative” forms in Mongolian, e.g. tik- ‘to fall’ : PREV tikcik ‘(look out:) s/he may fall’ : PX2SG tik-cik-es ‘mind your step (or you may fall)’. The hypothetical mood in -mnA‑ implies possibility or probability, e.g. xɵr- ‘to go (away)’ : HYP-PX3SG xɵre-mne-n ‘s/he will probably go’. The subjunctive mood in -mc(I) is used in an optative function or as an apodosis of counterfactual conditionals, as in [14]: [14]  deetle-y bi-seke-n   xin-teki dege-le-mc-u  wing-rx be-cv.aor-3sg 2sg-dir  fly-inch-subj-1sg  ‘If I had wings I would fly to you.’ As noted above, modal contents can also be expressed by participles, for which reason the list of verbal moods in some grammatical descriptions of Ewen is extended to include also predicatively used participial forms. The indicative mood is represented by two basic tense forms present in all dialects: the future in -JI‑ and the non-future or aorist in -RA- ~ -sA~ -dA- ~ -A-. While the future implies an unambiguous reference to future actions, e.g. em- ‘to come’ : FUT-1SG em-ji-m ‘I shall come’, the interpretation of the non-future form depends on the actionality of the verb, and also on the presence of aspectual markers with a perfectivizing or imperfectivizing function. When derived from telic verbs or from verbs including perfectivizing suffixes such as the momentative (semelfactive) form in -sAn-, the non-future form refers

280 Andrej Malchukov

to a recent past, e.g. AOR-1SG em-re-m ‘I have (just) come’, xukle- ‘to sleep’ : MOMAOR-3SG xukle-sn-e-n ‘(s/he) (just) fell asleep/slept for a while’. As in Ewenki, with telic (perfective) verbs reference to the actual present requires the use of the progressive aspect marker -J- ~ -JID-, cf. e.g. AOR-3SG em-re-n ‘(s/he) came’ vs. PROGR-AOR3SG eme-d-de-n ‘(s/he) is coming’, MOM-PROGR-AOR-3SG xukle-sen-jid-de-n ‘(s/he) is falling asleep’. However, when derived from atelic verbs, the non-future form refers to the present, e.g. AOR-3SG xukle-n ‘(s/he) sleeps’, ɪlac- ‘to stand’ : AOR-3SG ɪlat-ta-n ‘(s/he) stands’. This is different from Ewenki, where aorist forms, when not supported by the progressive marker, more rarely have a present tense reference. Apparently, the progressive aspect is less grammaticalized in Ewen and less integrated into the temporal system than in Ewenki. The expression of the past tense differs in the dialects: the eastern dialects make use of the past form in -RI ~ -I-~ -dI- ~ -sI-, derived from the imperfective participle, e.g. em- ‘to come’ : PST-1SG em-ri-w ‘I came’, while the western dialects express the past through the use of the periphrastic perfect in PTCP.PRF -cA + the copula bi-, e.g. em-ce bi-xe-ndi ‘you (already) came’ (come-PTCP.PRF be-AOR-2SG). Note that the perfective participle is less grammaticalized in this use than in Ewenki, as it needs copula support and cannot normally take the agreement suffixes directly (with the exception of the negation verb). In the mid-western dialects, where both past tense forms occur, they are opposed as imperfect and perfect, the latter also having evidential (‘apparently’) or mirative connotations, e.g. omƞa-ca bi-xe-m ‘(it turns out that) I forgot’ (forget-PTCP.PRF be-AOR-1SG). The negation verb e- has some idiosyncratic temporal and modal forms. Thus, it has a future paradigm in -te-, e.g. e-te-m xaa-r ‘I shall not know’ (NEG-FUT-1SG know-CONNEG), a related imperfective participle in -ti, e.g. ƞii=de e-ti-n xaa-r ‘unknown to anybody’ (who= CONNEG NEG-PTCP.IMPRF-PX3SG know-CONNEG), and a past tense in -c‑ ~ -cI‑, e.g. e-c-u xaa-r ‘I did not know’ (NEG-PST-1SG know-CONNEG). It also has the specific second person prohibitive form SG e-ji : PL e-ji-lre. CONVERBS The Ewen system of converbs comprises eight productive forms (Table 11.6), which represent three morphological classes: (1) converbs that do not inflect at all, (2) converbs that inflect for number but not person, and (3) converbs that inflect for both number and person. In class (2), plural number (PL) is indicated either by the regular plural marker of nominal declension, or by the plural form of the reflexive marker (RX, when the corresponding singular form also contains the reflexive marker), while in class (3) number and person are indicated by the markers of personal possession (PX). With regard to referentiality (coreferential vs. switch-referential) the converbs may be divided into conjunct or same-subject (SS), disjunct or different-subject (DS), and ambivalent (SS/DS) types. TABLE 11.6 EWEN CONVERB MARKERS  

 

PL

PX/RX

SS

CONN

-mI

+

SUCC

-mnIn

+

LIM

-kAn

SIM

-nIkAn

+ -nIkA-r

+

DS

(+)

Ewen 281  

 

PL

ANT

-RIj-I

-RIj-Ur

PX/RX

SS

DS

+

CTEMP

-ƞsI-

+

+

+

PURP

-dA-

+

+

+

AOR

-RAk-

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e, U = ʊ u, I = ɪ i, RA = the aorist marker depending on the conjugational class.

•

•

•

The group of non-inflecting converbs includes three forms: the general connective converb in -mI, the successive converb in -mnIn, and the limitative converb in -kAn. These are all used in same-subject constructions only, though with some exceptions for the limitative converb. The connective converb can convey conditional, temporal, or causal meanings. When used in combination with the emphatic clitics =dA and =tIt, it can also have a concessive meaning. Within a temporal clause, depending on the telicity of the subordinate verb, it implies either simultaneity or temporal precedence, cf. e.g. mʊcʊ-mɪ tog-ʊ it-te-n ‘when he returned, he saw a fire’ (return-CV.CONN fire-ACC see-AOR-3SG) vs. mʊcʊ-d-mɪ tog-ʊ it-te-n ‘when he was returning, he saw a fire’ (return-PROGR-CV.CONN fire-ACC see-AOR-3SG). The successive converb indicates that the subordinate event immediately succeeds or interrupts the main clause event, e.g. eyen-dide-mnin xor-ra-n ‘he was floating (down the river) when he (suddenly) got stuck’ (float-PROG-CV.SUCC get.stuckAOR-3SG). The limitative converb indicates a temporal limit or consequence of the main event, e.g. gurgewci-n heci-ken ‘s/he worked until s/he got tired’ (workAOR.3SG get.tired-CV.LIM). The group of converbs inflecting for number (of the subject), but not for person, includes two forms: the simultaneous converb in -nIkAn, indicating simultaneity, and the anterior converb in -RijI, indicating anteriority of action. From the formal point of view the simultaneous converb is a deverbal noun, which is why it has the regular nominal plural form in -nIkA-r, e.g. nulge- ‘to nomadize’ : CV.SIM nulge-niken ‘as (I, you, s/he) nomadize(s)’ : PL nulge-nike-r ‘as (we, you, they) nomadize’. The anterior converb involves, however, a quasiconverbal construction based on the instrumental case form of the imperfective participle in -RI combined with the reflexive marker (PTCP.IMPRF-INSTR-RX), which is why the singular and plural forms are distinguished by the corresponding forms of the reflexive marker, e.g. em- ‘to come’ : CV.ANT SG em-nij-i : PL em-nij-ur ‘having come’. Finally, the group of converbs inflecting for both person and number includes the contemporal converb in -ƞsI-, the purposive converb in -dA-, and the aorist converb in -RAk‑ ~ -Ak‑ ~ -dAk‑ ~ -sAk-, depending on the conjugational class of the verb. The contemporal converb refers to a process/state in the past including the time span of the main event, e.g. kʊƞa bi-ƞsi-n ‘when s/he was a child’ (child be-CV.CTEMPPX3SG). The purposive converb, also known as the “supine”, indicates purpose, e.g. gɵɵ-li taƞ-da-n ‘tell him/her to read’ (say-IMP.2SG read-CV.PURP-PX3SG). The aorist converb in -RAk‑ ~ -Ak‑ ~ -dAk‑ ~ -sAk-, based on the same element as the aorist marker -RA‑ ~ -A‑ ~ -dA‑ ~ -sA-, has conditional-temporal meanings similar

282 Andrej Malchukov

•

to those of the general connective converb in -mi. Although functionally similar, the aorist converb differs from the connective converb with regard to referentiality (conjunct vs. disjunct reference), cf. e.g. em-mi gɵɵn-ji-m ‘if/when I come I shall tell’ (come-CV.CONN say-FUT-1SG) vs. em-rek.e-s gɵɵn-ji-m ‘if/when you come I shall tell’ (come-CV.COND-PX2SG say-FUT-1SG). Apart from the limitative converb in -kAn, Ewen has traces of another converb with a similar function, containing the “terminative” marker -d(A)lA-, which, unlike -kAn, can take person marking. The active use of -d(A)lA- is, however, restricted to constructions containing the negation verb, e.g. e-dle-s em-re ‘until you come’ (NEG-CV.TERM-2SG come-CONNEG). The same form is also present in the lexicalized postposition ɪs-tala ‘until’.

PHRASE STRUCTURE As regards word order, Ewen is a typical “Ural-Altaic” language, that is, a consistent head-final language on both phrase and clause level. Within a nominal phrase, the head normally follows its modifier(s), both attributive and possessive. An attributive modifier agrees in case and number with its head: anƞamta-l-dʊla jʊʊl-dʊla ‘in (the) new houses’ (new-PL-LOC house-PL-LOC). In the western dialects, however, agreement is less regular and may even be absent in prenominal modifiers. The modifier position can be occupied by different classes of attributive nominals—adjectives, participles, numerals, demonstrative pronouns, as well as non-count nouns, e.g. moo-l nʊƞa-l ‘wooden bows’ (wood-PL bow-PL), and degree adverbs, e.g. hoo-l eƞeye-l bileke-l ‘very rich settlements’ (very-PL rich-PL settlement-PL). If a modifier is focused, it is postposed, in which case it obligatorily shows agreement in case in all dialects. Attributive constructions with focused attributes can also dialectally pattern as possessive with the attribute as the formal head of the construction, a phenomenon known as attribute raising, as in [15]: bey-u → bey xoonɪ-wa-n bakalda-rɪ-w [15] xoonɪ-w mighty-acc man-acc  m  an m  ighty-acc-px3s meet-pst-1sg ‘I met a mighty man.’ → ‘It was a mighty man whom I met.’ Unlike in attributive constructions, dependency in possessive constructions is headmarked, that is, the head noun inflects for the person and number of the possessor, e.g. etiken ora-r-nɪ ‘the old man’s (many) reindeer’ (old.man reindeer-PL-PX3SG). If the possessor is expressed by a personal pronoun, the dependency is double-marked as the pronoun takes the possessive form (identical with the oblique stem), e.g. min ora-r-bʊ ‘my (many) reindeer’ (1SG.GEN reindeer-PL-1PXSG). As noted above, postpositional phrases are patterned as possessive constructions, headed by a spatial nominal (relative noun), e.g. xɪakɪta ɵyde-le-n ‘on the top of the tree’ (tree top-LOC-PX3SG). If the possessor is focused, it is marked by the emphatic possessive marker -ƞI, which is postposed and carries an appropriate case to agree with the head, as in [16]: xi-ƞi-l-bu ga-d-nɪ [16] oro-r-bʊ-s reindeer-acc-px2sg 2sg-poss-pl-acc take-aor-3sg ‘It was your (many) reindeer that he took.’ Within a verb phrase a verb governs the case of its arguments. Ewen makes a distinction between two types of government: direct and indirect (postpositional). In the latter

Ewen 283

type the case marker is assigned to the postposition rather than to the relevant argument, cf. e.g. jʊʊ-la teget‑ ‘to sit in a/the house’ (house-LOC sit-) vs. jʊʊ doo-la-n teget‑ ‘to sit in(side of) a/the house’ (house inside-LOC sit-). The list of (lexical) governors includes, apart from finite and non-finite verbs, also deverbal nouns, e.g. oro-m kɵsci-mƞe ‘reindeer-breeder’ (reindeer-ACC breed-AG), and adverbs, e.g. okaat-ʊ holaakɪ ‘upwards the river’ (river-ACC upwards). SENTENCE TYPES Within the clause the basic word order is SOV. This order is more strictly adhered to than, for instance, in Ewenki, and any deviations from the basic word order are conditioned discourse-pragmatically. The SVO structure is attested when the verb, rather than the whole verb phrase, is rhematic or contrastive, and it is frequent in imperative sentences, e.g. dolcɪ-lra min-u ‘listen to me!’ (listen-IMP.2SG 1SG-ACC). The order VS(O) is mostly found in thetic-presentational contexts, e.g. bid-de etiken atɪka-ñʊn ‘there lived an old man and an old woman’ (live-AOR.3PL old.man old.woman-COM). Ewen, like the other Tungusic languages, follows the (nominative‑)accusative pattern: the subject is in the unmarked nominative case, the object in the marked accusative or designative cases. In combination with the marker of reflexive possession (RX), the object is unmarked, e.g. jʊʊ-y it-te-n ‘(s/he) saw his/her (own) house’ (house-RX seeAOR-3SG). The only deviation from the nominative pattern is the use of the designative case to mark the intransitive subject of certain “unaccusative” predicates (of appearance/ existence), as attested in the eastern dialects, e.g. (examples from Novikova 1980) kʊmaƞ-ga-n xie-n ‘a seal appeared (for him)’ (seal-AL-DES-PX3SG appear-AOR.3SG), jʊʊga-kʊ acca ‘I have no house’ (house-DES-PX1SG NEG.EXIST). Also, some dialects, in particular, the south-eastern, following a pattern well known from Turkic and Mongolic, show traits of topic-prominence, in that a topical subject, especially when contrastive, is often marked by the particle bimi = bi-mi ‘if/when it is’ (be-CV.CONN), e.g. ƞeeluki bimi gɵɵn-ni ‘(as for) the wolf (it) said . . . ’ (wolf TOP say-PST). The subject is obligatorily lacking in impersonal constructions formed by medial (medio-passive) forms of verbs, as well as by certain types of verbal nouns with a modal function, e.g. ere-w okaat-ʊ daw-kɪc ‘this river can be crossed’ (this-ACC river-ACC cross-NMLZ). Note also that a pronominal subject may always be dropped, which means that Ewen is a “pro-drop” language. The predicate shows agreement with the subject in person and number. Within verbal clauses agreement is expressed on the verb by means of the person markers (Table 11.5). In nominal clauses the copula bi‑ : bi-si- ‘to be’ can be absent in the third person of the aorist; otherwise it inflects for person and number, whereas the nominal predicate inflects only for number, e.g. oro-r berge-l bi-si-ten ‘the reindeer were fat’ (reindeer-PL fat-PL be-PST-3PL). Only exceptionally, within certain types of exclamative clauses, can a nominal predicate take the person agreement endings (of the possessive type) with a copula, e.g. xii ɪak eƞi-s ‘how strong you are!’ ~ ‘what strength you have!’ (2SG what strong-PX2SG). Participial predicates may follow either a nominal pattern, combining with a copula, or, depending on the participle, the verbal pattern, taking agreement endings directly. In either case they tend to be reanalysed as finite tense/mood forms, as is also evident from the rise of the synthetic imperfect (in -RI-) in the eastern dialects and the periphrastic perfect (in -cA-) in the western dialects. In the following, some specific syntactic features connected with possessive, existential, and interrogative constructions are discussed in more detail.

284 Andrej Malchukov

•

•

•

On the clausal level possession can be encoded in a number of ways. First, it can be expressed by the proprietive form in -lkAn or the negative proprietive construction in ac + -lA in predicate position, e.g. tar bey jʊʊ-lkan ‘that man has a house’ (that man house-PROPR) vs. tar bey ac jʊʊ-la ‘that man does not have a house’ (that man NEG.EXIST house-PROPR.CONNEG). Second, it can be modelled on the locative-existential pattern with the possessor marked with the locative case and the possessed item in the subject position, e.g. bey-le jʊʊ bi-s-ni ‘the man has a house’ (man-LOC house be-AOR-3SG). Finally, as in the Turkic languages, it can be expressed through a construction with a thematic possessor (which may be omitted but is cross-referenced by possessive agreement) within the subject phrase, e.g. (bey) jʊʊ-n bi-s-ni ‘the man has a house’ (man) (house-PX3SG be-AOR-3SG). Locative existential sentences differ from locative proper sentences both in word order and case marking: in the former the locative NP stands in the locative case, while in the latter it is mostly in the dative, cf. e.g. jʊʊ-la bey bi-s-ni ‘there is a man (somebody) in the house’ (house-LOC man be-AOR-3SG) vs. bey jʊʊ-dʊ bi-s-ni ‘the man is in the house’ (man house-DAT be-AOR-3SG). Another difference between locative existential and locative proper sentences concerns their negative counterparts. In existential sentences negation is expressed through the negative existential noun acca ‘absent, not existent’, e.g. jʊʊ-la bey acca ‘there is no man (nobody) in the house’ (house-LOC man NEG.EXIST). This pattern also carries over to possessive-existential constructions, e.g. (existential possession) noƞan-dʊla jʊʊ acca (3SG-LOC house NEG.EXIST) ~ (thematic possessor) noƞan jʊʊ-n acca (3SG house-PX3SG NEG.EXIST) ‘s/he does not have a house’. In locative sentences, by contrast, negation is expressed by the negated form of the copula-existential bi- : bi-si- ‘to be’, e.g. bey jʊʊ-dʊ e-s-ni bi-s ‘the man is not in the house’ (man house-LOC NEG-AOR-3SG be-CONNEG). The same construction is used in equative clauses to negate nominal predicates, e.g. erek bey e-s-ni buyuse-mƞe bi-s ‘this man is not a hunter’ (this man NEG-AOR-3SG hunt-AG be-CONNEG). In Ewen, unlike in Ewenki, the formation of interrogative sentences usually does not involve fronting of the interrogative pronoun, e.g. ak-mʊ ile ƞen-re-n ‘where did my (elder) brother go?’ (brother-PX1SG where go-AOR-3SG). Fronting of an interrogative pronoun may, however, be used for emphasis, as in rhetorical questions, e.g. ile ak-mʊ ƞen-re-n ‘where did my (elder) brother go?’ (where brother-PX1SG go-AOR3SG). Polar and alternative questions are formed by adding the enclitic particle =GU to the predicate, or to both predicates in an alternative question, e.g. min-u xaa-s=ku e-se-s=ku ‘do you know me or not’ (1SG-ACC know-AOR.​2PL=INTERR NEGAOR-2PL=INTERR). Interestingly, interrogative pronouns are not restricted to matrix clauses. This also holds for the interrogative pro-verb ɪa- ‘to do what’, which can be used in non-finite verbal forms, as in [17]:

[17] ɪa-rak-ʊ min-u nei-nri do.what-cv.aor-1sg 1sg-acc scold-aor.2sg ‘what have I done (so wrong) that you scold me?’ Many of such non-finite forms have been lexicalized, forming a parallel to, for instance, similar lexicalizations in the Mongolic languages, e.g. ɪa- ‘to do what?’ : CV.CONN ɪa-mɪ ‘why?’ : CV.PURP-RX ɪa-da-y ‘what for?’, etc.

Ewen 285

PASSIVE AND CAUSATIVE Ewen has preserved the distinction (lost in Ewenki) between the two passive forms in -B- (*p) and -W- (*b), which are referred to here as medio-passive (MED.PASS) or medial and adversative-passive (ADV.PASS), respectively. The medio-passive (medial), is mostly derived from transitives and expresses an objective-resultative or anti-causative function, e.g. urke aaƞa-p-ta-n ‘the door is open/opened’ (door openMED.PASS-AOR-3SG). When derived from intransitives it yields an impersonal construction with a potential meaning, as in [18]. In both constructions the agent is left unexpressed. ƞene-p-te ac tʊrkɪ-la-c [18] e-ti neg-ptcp.imprf go-med.pass-conneg neg.exist sled-propr.conneg-instr ‘It is impossible to go (there) without a sled.’ The adversative-passive is of the type familiar from Japanese and denotes an action that is unfavourable for the (derived) subject. When derived from transitives, the derived subject most often corresponds to the object of the base verb, while the agent (if present) is marked by the dative, as in [19]. In other cases the derived subject corresponds not to the underlying object but, rather, to its possessor. In such cases the construction remains transitive, as in [20]: nugde-du maa-w-ra-n [19] etiken old.man bear-dat kill-adv.pass-aor-3sg ‘The old man was killed by a bear.’ [20] etiken nugde-du gɪa-y old.man bear-dat friend-rx ‘The old man had his friend killed by a bear.’

maa-w-ra-n kill-adv.pass-aor-3sg

Moreover, the adversative passive can be formed from intransitive verbs, as well. in which case the underlying subject can appear either in the dative or in the accusative case, or in a reflexive form replacing the accusative, e.g. etiken or-mɪ xɵre-w-re-n ‘the old man lost his reindeer’ (literally: ‘had his reindeer gone’) (old.man reindeer-RX go.away-ADV. PASS-AOR-3SG). Causative proper forms are derived from both intransitives and transitives with the help of the marker -WkAn-. Causatives derived from intransitives take the causee in the accusative, e.g. hut-u em-ukene-m ‘I made/let the child come’ (child-ACC come-CAUSAOR.1SG). Causatives derived from transitives can take the causee either in the dative or, less frequently, in the accusative, as in [21]: kajjak-tʊ ~ kajjak-ʊ miine-w kool-ʊkan [21] ewe-sel Ewen-pl Kajjak-dat Kajjak-acc wine-acc drink-caus.aor.3pl ‘The Ewen made/let Kajjak drink the wine.’ If the causee is in the dative the causative form is open to both the factitive (‘to make’) and the permissive (‘to let’) interpretation, while if it stands in the accusative only the factitive interpretation is possible.

286 Andrej Malchukov

COMPLEX SENTENCES Coordination in phrases and clauses has no overt marking (conjuncts are simply juxtaposed), though it can also be expressed by the conjunctional clitics =dA (in declarative sentences) and =GU (in interrogative sentences). Otherwise conjunctional adverbs like ñaan ‘again, also’ > ‘and’ and tadʊk = ta-dʊk (that-ABL) ‘then’ > ‘and’ can be used to conjoin clauses; some dialects use these connectors to conjoin noun phrases, as well. Conjunctional converbs, such as tiemi = tie-mi (take-CV.COND) ‘taking’ > ‘therefore’, ɪamɪda = ɪa-mɪ=da (do.what-CV.COND=INDEF) ‘whatever doing’ > ‘however’, and tarɪc = tar-ɪc (that-INSTR) ‘in that way’ > ‘then’, are also used as analogues of complex constructions with adverbial clauses in other languages, but they do not qualify as subordinate by standard syntactic criteria, in that none of the clauses introduced by a conjunction permits center-embedding into another clause. Ewen is similar to other “Ural-Altaic” languages, in that it makes use of non-finite constructions instead of finite subordinate clauses. Complement clauses and relative clauses are formed by participles, while adverbial clauses employ both converbs (proper) and quasiconverbs (participles in oblique cases). The subject of non-finite subordinate clauses is in the unmarked (nominative) case if lexical, or in the possessive (including reflexive possessive) form if pronominal, following the pattern of possessor in a possessive construction. Objects and adjuncts (if any) are retained in the form required by the sentence structure. •

•

Complement clauses frequently occur with verbs of cognition, perception, emotion and speech, e.g. xin em-ce-wu-s it-ti-w/xaa-rɪ-w/teleƞ-ri-w ‘I saw/knew/told that you came’ (2SG.GEN come-PTCP.PRF-ACC-PX2SG see/know/tell-PST-1SG). Indirect questions are formed by the same type of participial complements in combination with the interrogative pronouns (fulfilling a relative function), e.g. xin idʊk em-cewu-s ulgim-ri-w ‘I asked where you came from’ (2SG.GEN whence come-PTCP. PRF-ACC-PX2SG ask-PST-1SG). Relative clauses (RC) are expressed by participles, which normally precede the head noun, though they can also follow the head noun when focused or used non-restrictively. In a subject RC, where the head noun is coreferential to the participial subject, the participle is uninflected for person, though it can take case and number agreement, e.g. buyu-m maa-ca etiken ‘the old man who killed the (wild) reindeer’ (wild.reindeer-ACC kill-PTCP.PRF old.man). In an object RC, where the head noun is coreferential to one of the verbal complements, the participle takes the (possessive) agreement markers to agree with the subordinate subject, e.g. etiken maa-ca-n buyun ‘the (wild) reindeer that the old man killed’ (old.man kill-PTCP.PRF-PX3SG wild.reindeer). Relativization of the possessor follows the same pattern as relativization of the head of NP; note, in particular, the absence of agreement in the participle when the possessor of the subject is relativized, e.g. atɪka-ƞa-n kɵke-ce etiken ‘the old man whose wife has died’ (old.woman-AL-PX3SG die-PTCP.PRF old-man). Apart from the prenominal position, RCs can also be internally headed, in which case the notional head retains its position in the clause, as well as its case-marking, as in [22]:

buyu-m maa-ca-wa-n emu-ri-w [22] etiken old.man wild.reindeer-acc kill-ptcp.prf-acc-px3sg bring-pst-1sg ‘I brought the reindeer that the old man killed.’

Ewen 287

Internal RCs have a structure identical to that of complement clauses, and in all probability they developed from a reanalysis of complement clauses or accusative-marked adverbial clauses. Unlike prenominal RCs, internal RCs are mostly non-restrictive (appositive). •

Adverbial clauses are formed by converbs or quasiconverbs. The latter tend to be grammaticalized as they increase in frequency and get more specialized in meaning. Thus, the form in -RIjI : PL -RIjUr (= -RI-j-I : -RI-j-ur), which historically involves a combination of the imperfective participle with the instrumental case, marked for reflexive possession, may synchronically already be analysed as a true converbal form. Forms that still remain at the level of quasiconverbs include the anterior form in PTCP.PRF-LOC -CA-lA-, the concessive form in PTCP.IMPRF-DIR.PROL -RI-klI-, and the purposive form in PTCP.HYP-ACC -JIƞA-wA-. The syntactically important distinction between conjunct (same-subject) and disjunct (different-subject) adverbial clauses is signalled through the use of the reflexive possessive vs. personal possessive endings on the participles, cf. e.g. em-ce-le-y ulgim-re-n ‘after coming back s/he asked’ (come-PTCP.PRF-RX ask-AOR-3SG) vs. em-ce-le-n ulgim-re-n ‘after s/he came back, s/he (another person) asked’ (come-PTCP.PRF-PX3SG ask-AOR3SG). The same distinction obtains in other types of subordinate clauses, that is, in complement clauses and relative clauses, cf. e.g. gɵɵn-ce-y omƞa-rɪ-n ‘s/he forgot what/that s/he (him/herself) had said’ (say-PTCP.PRF-RX forget-PST-3SG) vs. gɵɵn-ce-we-n omƞa-rɪ-n s/he forgot what/that s/he (another person) had said’ (sayPTCP.PRF-ACC-PX3SG forget-PST-3SG). Note also that some conjunct converbs, including even the synchronically invariable connective converb in -mI (in Arman SG -m-I : PL -m-Ur) contain historically the reflexive marker. The only converb that without changing its form can occur in both same-subject and different-subject constructions is the limitative converb in -kAn, cf. e.g. kuusi-r kɵke-ken ‘they fought until they died’ (fight-AOR.3PL die-CV.LIM) vs. buler-bu nɵkle-n kɵke-ken ‘he shot at the enemies until they died’ (enemy-PL-ACC shoot-AOR.3SG die-CV.LIM). In general, such “variable-subject” converbs, which are common in, for instance, some Turkic languages, are not typical of Ewen.

Converbs can also occur in sequences. Clause-chaining is especially typical for the anterior converb in -RIjI, which is often used in a narrative function, as in [23]: [23] xukle-riji teg-rek-u bɵɵdel-u en-se-n sleep-cv.ant sit-cv.aor-1sg leg-px1sg hurt-aor-3sg ‘As I, after sleeping, sat up, my legs hurt.’ More interestingly, apart from converbs, which permit multiple embedding, participles can also embed converbs, as in [24]: [24] adal-ɪ bak-rɪjɪ xɵr-ce-we-n gɵɵ-weet-te net-rx find-cv.ant go.away-ptcp.prf-acc-px3sg say-iter-aor.3pl ‘They said that he left after he found his net’. Thus, non-finite forms are not different from finite with respect to converbal embedding, which can be taken as an argument for the clausal status of constructions with

288 Andrej Malchukov

non-finite forms. This is also confirmed by their use with anaphors. Anaphors (reflexives, possessive reflexives, and reciprocals) have clause-bound antecedents, that is, they refer to a subordinate subject in a non-finite clause, rather than to the main clause subject, as in [25]: [25] ak-mʊ jʊʊ-la-y mʊcʊ-raka-n eydu ukcen-re-m elder.brother-px1sg house-loc-rx return-cv.aor-3sg everything tell-aor-1sg ‘When my elder brother returned to his (own) house, I told (him) everything’. LEXICON AND LANGUAGE CONTACTS As expected, Ewen shares most of its basic vocabulary with the other Ewenic languages. Arman has many special words not found in other Ewen varieties, even in the basic vocabulary, e.g. Arman ƞees ‘nose’, but on the whole it usually goes with Ewen rather than with Ewenki in its lexical choices (Doerfer), cf. e.g. Ewen ñɵɵlten = Arman yɵɵlteƞ ‘sun’ vs. Ewenki dilacaa. Early loans from Mongolic are also shared with other Tungusic languages and may be traced back to the level of the two protolanguages, e.g. mʊran ‘horse’ < *murïn ← Mongolic *murï/n (~ *morï/n), emgun ‘saddle’ < *emegön ← Mongolic *emegöl (> *emexel > emeel), but there are also some later loans from the Post-Proto-Mongolic period, e.g. bool ‘slave’ ← Mongolic *bool (< *boxol < *bogol). Many Mongolic items have penetrated Ewen via Yakut. Borrowings from the neighbouring Palaeosiberian languages are dialectally restricted: loans from Chukchee and Koryak mostly occur in the north-eastern dialects, e.g. maaya ‘supply of food’ ← Chukchee maymay (= may&may), atwat ‘boat’ ← Chukchee atwaat, while loans from Yukaghir are found in the north-western dialects, e.g. nalɪma ‘sled’ ← Forest Yukaghir nalime, oondɪ ‘water’ ← Forest Yukaghir ondi, kɵtlen ‘little’ ← Tundra Yukaghir kɵyleƞ. The north-western dialects have also borrowed some kinship terms from Tundra Yukaghir, e.g. emje ‘younger brother/sister’ (Dutkin), which, in turn, has received numerous loans from Ewen, including, in particular, terminology of reindeer breeding, but also kinship terms. The south-eastern (Okhotsk) dialects have borrowed many Ewenki words, e.g. ookɪ ‘how many?’, nakɪta ‘bear’, including also some grammatical morphemes, e.g. the third-person imperative forms in -gAn ← Ewenki -gin. On the other hand, those Ewenki dialects (e.g. Tokkо) that are in contact with Ewen, also reveal signs of structural interference. The most important source of borrowing, apart from Russian, is Yakut. As expected, massive borrowing is attested in the western dialects, which show interference from Yakut in all domains of grammar and vocabulary. However, Yakut loans have penetrated through interdialectal borrowing to a number of eastern dialects, as well, and they are also found in Arman. The Yakut lexical influence is not restricted to cultural vocabulary, such as, e.g. mʊƞka ‘fishing net’ ← Yakut muƞxa (ultimately from Samoyedic), tʊʊx ‘salt’ ← Yakut tuus, but it also comprises lexicon pertaining to fauna and flora, e.g. koobak ‘hare’ ← Yakut koobax, as well as kinship terms, e.g. abaga ‘grandfather’ ← Yakut abaga (ultimately from Mongolic). Apart from nouns, adjectives, e.g. akaarɪ ‘stupid’ ← Yakut akaarɪ, adverbs, e.g. xootorʊ ‘soon’ ← Yakut sootoru, as well as particles and discourse markers, e.g. bagar ‘maybe’ ← Yakut bagar, have also been borrowed. Verbs are borrowed in the stem form, e.g. tala- ‘to rob’ ← Yakut talaa-, xata- ‘to be able’ ← Yakut

Ewen 289

sataa-, etc. Some grammatical markers have been borrowed, as well; thus, for instance, the western dialects form ordinal numerals with the formative -s ← Yakut -s, e.g. ɪlɪ-s ‘third’. Although the first loanwords from Russian date back to the 17th century, lexical borrowing from Russian, in particular, in the fields of technical and political vocabulary, became especially extensive after the October Revolution. Early loans underwent radical phonological adaptation, e.g. ʊrbaak ‘shirt’ ← Russian rubáxa, tebeeki ‘tobacco’ ← Russian tabák, pacɪɪba ‘thank you’ ← Russian spasíbo. Some items may have been mediated by Yakut, e.g. oxok ‘stove’ ← Yakut oxoq ← Russian ochág. Recent loans usually preserve the original form more closely, at least in the written language. Russian adjectives are borrowed with a stem-final -Ay, e.g. rayonnay ‘regional’ ← Russian raiónnyi, while verbs are borrowed in the second person singular imperative form, e.g. zwonɪ- ‘to telephone’ : AOR-3SG zwonɪ-ra-n ← Russian IMP.2SG zvoní. Russian interference is also obvious in the syntax. REFERENCES AND FURTHER READING Aralova, Natalia (2015) Vowel Harmony in Two Even Dialects: Production and Perception, PhD Thesis, University of Amsterdam. Benzing, Johannes (1955) Lamutische Grammatik, mit Bibliographie, Sprachproben und Glossar, Akademie der Wissenschaften und der Literatur, Veröffentlichungen der Orientalischen Kommission 6, Wiesbaden: Franz Steiner Verlag. Bogoraz, V. G. [В. Г. Богораз] (1931) ‘Материалы по ламутскому языку’ [Materials on the Lamut language], in: Тунгусский сборник 1: 1–106, Ленинград [Leningrad]: Издательство АН СССР. Burykin, A. A. [А. А. Бурыкин] (1983) Категория переходности-непереходности глагола в эвенском языке [The category of transitivity-intransitivity in the Ewen language], Диссертация на соискание ученой степени кандидата наук [PhD thesis], ЛО ИЯ АН СССР. Burykin, A. A. [А. А. Бурыкин] (1986) ‘К фонологической интерпретации некоторых явлений в вокализме диалектов эвенского языка’ [On the phonological interpretation of the Ewen vowel system], in: Фонетика языков Сибири и сопредельных регионов, 57–63, Новосибирск [Novosibirsk]. Burykin, A. A. [А. А. Бурыкин] (1997) ‘Письменная и устная форма эвенского языка: диалектная структурa и фукнкциональный статус диалектов’ [The written and oral forms of the Ewen language: the dialectal structure and the functional status of dialects], in: D. M. Nasilov [Д. М. Насилов] (ed.), Малочисленные народы Севера, Сибири и Дальнего Востока, 54–88, Санкт-Петербург [St. Petersburg]: “Наука”. Burykin, A. A. [А. А. Бурыкин] (2001) Малые жанры эвенского фольклора [Minor genres of Ewen folklore], Санкт-Петербург: “Петербургское Востоковедение”. Burykin, A. A. [А. А. Бурыкин] (2004). Язык малочисленного народа в его письменной форме (на материале эвенского языка) [The language of a minority people in its written form (on materials from the Ewen language)], Санкт-Петербург [St. Petersburg]: “Петербургское Востоковедение”. Cincius, V. I. [В. И. Цинциус] (1947) Очерк грамматики эвенского (ламутского) языка [An outline of Ewen (Lamut) grammar], Ленинград [Leningrad]: Учпедгиз.

290 Andrej Malchukov

Cincius, V. I. [В. И. Цинциус] & L. D. Rishes [Л. Д. Ришес] (1952) Русско-эвенский словарь [Russian-Ewen dictionary], Москва [Moscow]: Издательство иностранных и национальныйх словарей. Cincius, V. I. [В. И. Цинциус] & L. D. Rishes [Л. Д. Ришес] (1957) Эвенско-русский словарь [Ewen-Russian dictionary], Ленинград [Leningrad]: Учпедгиз. Danilova, A. A. [А. А. Данилова] (1991) Бытовая лексика эвенского языка [Everyday vocabulary of the Ewen language], Якутск [Yakutsk]: ЯНЦ СО РАН. Doerfer, Gerhard & Wolfram Hesche & Helmut Scheinhardt (1980) Lamutisches Wörterbuch, Wiesbaden: Otto Harrassowitz. Dutkin, X. I. [Х. И. Дуткин] (1995) Аллаиховский говор эвенов Якутии [The Allaikha dialect of the Yakutian Ewen], Санкт-Петербург [St. Petersburg]: “Наука”. Gladkova, N. I. [Н. И. Гладкова] (1991) ‘Синтаксис’ [Syntax], in: K. A. Novikova [К. А. Новикова] & al. (1991), 202–297. Hesche, Wolfram (1978) ‘Urtungusisch *ö in erster Silbe (= *ö1) im Lamutischen’, in: Gerhard Doerfer & Michael Weiers (eds.), Beiträge zur Nordasiatischen Kulturgeschichte, Tungusica 1: 117–125, Wiesbaden: Otto Harrassowitz. Kazama, Shinjirō 風間伸次郎 (2003)『エウェン語 テキストと文法概説』[Ewen texts and grammar],『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 23, 吹田 [Suita]: 大阪学院大学情報部. Kazama, Shinjirō 風間伸次郎 (2009)『エウェン語 テキスト 2AB』[Ewen Texts 2AB],『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 45AB, 東京 [Tokyo]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 (2013)『エウェンの民話』[Ewen folktales],『ツン グース言語文化論集』[Publications on Tungus Languages and Cultures] 56, 東京 [Tokyo]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 (2015)「エウェン語ブィストラヤ方言の概説とテキ スト」[A sketch and texts on the Bystraya dialect of Ewen],『北方言語研究』Northern Language Studies 5: 83–128, 札幌 [Sapporo]. Kazama, Shinjirō 風間伸次郎 (2016–2020)『エウェン語ブィストラヤ方言テキ スト１』[Texts on the Bystraya dialect of Ewen in Kamchatka], [1/2016, 2/2018, 3/2020] 東京 [Tokyo]: 東京外国語大学 =『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 62, 65, 66. Kazama, Shinjirō 風間伸次郎 (2019)「《データ：「情報表示の諸要素」》エウ ェン語」[Information structure in Ewen],『語学研究所論集』22: 183–188, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2022)『エウェン語アルカ方言の研究』[A study of the Arka dialect of Ewen],『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 69, 東京 [Tokyo]: 東京外国語大学. Kim, Juwon (2011) A Grammar of Ewen, Altaic Languages Series 06, Seoul: Seoul National University Press. Kuz’mina, R. P. [Р. П. Кузьмина] (2010) Язык ламунхинских эвенов [The language of the Lamunkhin Ewen], Новосибирск [Novosibirsk]: “Наука”. Lebedev, V. D. [В. Д. Лебедев] (1978) Язык эвенов Якутии [The language of the Yakutian Ewen], Ленинград [Leningrad]: “Наука”. Lebedev, V. D. [В. Д. Лебедев] (1982) Охотский диалект эвенского языка [The Okhotan dialect of the Ewen language], Ленинград [Leningrad]: “Наука”.

Ewen 291

Levin, V. I. [В. И. Левин] (1935) Самоучитель эвенского языка [Teach-yourself textbook of the Ewen language], Ленинград [Leningrad]: Угпедгиз. Malchukov, A. L. [А. Л. Мальчуков] (1989) Структура простого глагольного предложения в эвенском языке [The structure of the simple verbal sentence in the Ewen language], Диссертация на соискание ученой степени кандидата наук [PhD thesis], ЛО ИЯ АН СССР. Malchukov, Andrej L. (1993) ‘Adversative constructions in Even in relation to Passive and permissive’, in: Bernard Comrie & Maria Polinsky (eds.), Causatives and Transitivity, Amsterdam: Benjamins. Malchukov, Andrej L. (1995) Even, Languages of the World/Materials 12, München: LINCOM Europa. Malchukov, A. L. [А. Л. Мальчуков] (1998) ‘Условные конструкции в эвенском языке’ [Conditional constructions in the Ewen language], in: V. S. Khrakovskii [В. С. Храковский] (ed.), Типология условных конструквий, 453–477, Санкт-Петербург [St. Petersburg]: “Наука”. Malchukov, A. L. [А. Л. Мальчуков] (1999) Структура простого предложения в эвенском языке: структурные и содержательные аспекты [The structure of the simple sentence in the Ewen language: Structural and semantic aspects], СанктПетербург [St. Petersburg]: “Наука”. Malchukov, Andrej L. (2001) ‘Imperative constructions in Even’, in: V. S. Xrakovskij (ed) Typology of Imperative Constructions, 159–180, München: LINCOM Europa. Malchukov, A. L. [А. Л. Мальчуков] (2002) Синтаксис эвенского языка: структурные, семантические, коммуникативные аспекты [Syntax of the Ewen language: structural, semantic and discourse aspects], Диссертация на соискание ученой степени доктора наук [Habilitation thesis], Санкт-Петербург [St. Petersburg]: ИЛИ РАН. Malchukov, Andrej L. [А. Л. Мальчуков] (2003) ‘Russian interference in Tungusic languages in an areal-typological perspective’, in P. S. Ureland (ed.), Convergence and divergence of European languages, Studies in Eurolinguistics 1: 235–251, Berlin: LOGOS. Malchukov [Mal’čukov], Andrej (2006) ‘Yakut interference in North-Tungusic languages’, in: Henrik Boeschoeten & Lars Johanson (eds.), Turkic Languages in Contact, 122–138, Wiesbaden: Otto Harrassowitz. Malchukov, A. L. [А. Л. Мальчуков] (2008) Синтаксис эвенского языка: структурные, семантические, коммуникативные аспекты [Syntax of Ewen: structural, semantic, and discourse aspects], Санкт-Петербург [St. Petersburg]: “Наука”. Malchukov, Andrej (2020a) ‘Grammaticalization in Ewen (North-Tungusic) in a comparative perspective’, in: Walter Bisang & Andrej Malchukov (eds.), Grammaticalization Scenarios: Cross-Linguistic Variation and Universal Tendencies, vol. 1: Grammaticalization Scenarios from Europe and Asia, Comparative Handbooks of Linguistics 4.1: 399–432, Berlin: De Gruyter Mouton. Malchukov, Andrej (2020b) Metele: An Ewen Epic Tale by Darja Mixajlovna Osenina, Seoul: SNU Press. Malchukov, A. L. [А. Л. Мальчуков] & A. A. Petrov [А. А. Петров] (1994) ‘Русскоэвенские языковые связи’ [Russian-Ewen linguistic contacts], in: В. М. Панькин (ed.) Контактологический энциклопедический словарь-справочник 1: 229–239, Москва [Moscow]: Аз.

292 Andrej Malchukov

Matić, Dejan (2020) ‘Reflexives, reciprocals, and intensifiers in Ewen: An exercise in Tungusic dialectology’, International Journal of Eurasian Linguistics 2 (2): 244–285. Matić, Dejan & Alexandra Lavrillier (eds.) (2013) Эвенские нимканы Дарьи Михайловны Осениной [Even tales of Dar’iia Mikhailovna Osenina], Fürstenberg/ Havel: Kulturstiftung Sibirien. Novikova, K. A. [К. А. Новикова] (1958) Эвенский фольклор [Ewen folklore], Магадан [Magadan] Магаданское книжное издательство. Novikova, K. A. [К. А. Новикова] (1960) Очерки диалектов эвенского язык: Ольский говор [Outlines of Ewen dialects: The Ola dialect] I, Ленинград [Leningrad]: “Наука”. Novikova, K. A. [К. А. Новикова] (1980) Очерки диалектов эвенского язык: Ольский говор [Outlines of Ewen dialects: The Ola dialect] II, Ленинград [Leningrad]: “Наука”. Novikova, K. A. [К. А. Новикова] (1997) ‘Эвенский язык’ [The Ewen language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 284–304, Москва [Moscow]: “Индрик”. Novikova, K. A. [К. А. Новикова] & V. D. Lebedev [В. Д. Лебедев] (1980) Правила орфографии эвенского языка [Orthography rules of Ewen], Якутск [Yakutsk]: Якутский филиал СО АН СССР. Pakendorf, Brigitte (2009) ‘Intensive contact and the copying of paradigms: An Ėven dialect in contact with Sakha (Yakut)’, Journal of Language Contact, VARIA 2: 85–110. Pakendorf, Brigitte & Natalia Aralova (2020) ‘Even and the Northern Tungusic languages’, in: Martine Robbeets & Alexander Savelyev (eds.), The Oxford Guide to the Transeurasian Languages, 288–304, Oxford Guides to the World’s Languages, Oxford: Oxford University Press. Petrov, A. A. [А. А. Петров] (1991) Лексика духовной культуры эвенов [Vocabulary of the spiritual culture of the Ewen], Ленинград [Leningrad]: РГПУ. Petrov, A. A. [А. А. Петров] (1997) Лексика духовной культуры тунгусов (эвенки, эвены, негидальцы, солоны) [Vocabulary of the spiritual culture of the Tungus (Ewenki, Ewen, Neghidal, Solon], Санкт-Петербург [St. Petersburg]: Образование. Rishes, L. D. [Л. Д. Ришес] (1947) Арманский диалект эвенского языка (Очерк грамматики, тексты и словарь) [The Arman dialect of Ewen: A grammatical outline, texts and vocabulary], Диссертация на соискание ученой степени кандидата наук [PhD thesis], Ленинград [Leningrad]: ЛО ИЛ АН СССР. Robbek, V. A. [В. А. Роббек] (1982) Виды глагола в эвенском языке [Verbal aspect in the Ewen language], Ленинград [Leningrad]: “Наука”. Robbek, V. A. [В. А. Роббек] (1984) Категория залоговости в эвенском языке [The category of voice in the Ewen language], Ленинград [Leningrad]: “Наука”. Robbek, V. A. [В. А. Роббек] (1989) Язык эвенов Березовки [The language of the Berezovka Ewen], Ленинград [Leningrad]: “Наука”. Robbek, V. A. [В. А. Роббек] (1992) Грамматические категории эвенского глагола [The grammatical categories of the Ewen verb], Санкт-Петербург [St. Petersburg]: “Наука”. Schiefner, A[nton] (1859) ‘Beiträge zur Kenntniss der Tungusischen Mundarten’, in: Mélanges Asiatiques 3: 682–746, St. Petersbourg. Schiefner, A[nton] (1874) ‘Baron Gerhard von Maydell’s Tungusische Sprachproben’, in: Mélanges Asiatiques 7: 323–377, St. Petersbourg.

Ewen 293

Sharina, S. I. [С. И. Шарина] (2001) Персональность и посессивность в эвенском языке [Category of person and possession in Ewen], Новосибирск [Novosibirsk]: “Наука”. Sharina, S. I. [С. И. Шарина] & R. P. Kuz’mina [Р. П. Кузьмина] (2018) Нижнеколымский говор эвенского языка [The Nizhnekolymsk dialect of Ewen], Новосибирск: “Наука”. Sotavalta, Arvo (1978) Westlamutische Materialien, bearbeitet und herausgegeben von Harry Halén, Mémoires de la Société Finno-Ougrienne 168, Helsinki: Suomalais-Ugrilainen Seura.

CHAPTER 12

OROCH Shinjiro Kazama

Oroch (Orochi) is a Tungusic language of the Orochic group that was spoken until recently by a small population in the northern part of the Sikhote Alin Mountains, also known as the Ussuri Taiga, between the Lower Amur basin and the Strait of Tartary. The Oroch used to conduct a semi-settled way of life, based mainly on river fishing and hunting, but including also sea mammal hunting on the coast, with several more or less fixed settlements composed of summer and winter camps and located along local rivers, including the Tumnin, Khadi (Khadia), and Koppi, flowing into the Strait of Tartary, as well as the Khungari (Gur), flowing into the Amur. The region has been part of the Russian Far East since 1860. Initially administered in the context of the Maritime Province (Primor’e), it was in 1948 transferred to the neighbouring Khabarovsk Krai, where it is today divided between the raions of Komsomol’sk-na-Amure and Vanino. The most important urban center in the region is the port of Sovetskaya Gavan’ (until 1923 Imperatorskaya Gavan’) at the Strait of Tartary. The ethnonym Oroch (Russian MASC óroch : PL órochi : FEM órochka) was first applied to the Oroch by the Russian explorers arriving in the region in the 1850s. However, it seems not to have functioned as an original endonym of the Oroch, though in the modern language it is registered as oroci : PL oroci-sa : VBLZ oroci-da- ‘to speak in Oroch’. Most probably it is based on a confusion with the term “Orochen”, which has been used for several Ewenki-speaking groups in the neighbourhood, and which, by analogy, has also been extended to the Oroch. Conventionally, these appellations are assumed to be derived from the Tungusic word (*)oron ‘domestic reindeer’, implying that the Oroch were originally reindeer breeders. This is, however, unlikely, for there is no trace of reindeer herding in Oroch culture, and Oroch, together with Uilta (Orok), is one the two Tungusic languages that instead of (*)oron use a different term, ulaa (< *ulaa/r), for ‘domestic reindeer’. The true original endonym of the Oroch is naani < naa+nii ‘local people’, shared with the Nanai and Ulcha, with whom the Oroch have traditionally been in close contact. The southern neighbours of the Oroch are the Udihe, whose language is taxonomically the closest relative of Oroch. Typically, in early Russian accounts of the indigenous population of the Ussuri Taiga, no clear distinction was made between the Oroch and Udihe, and the ethnonym Oroch, or “Orochen”, was used indifferently for both ethnic groups, and, in fact, more often for the Udihe than for the Oroch (proper). However, in spite of the fact that Oroch and Udihe share both their primary taxonomic status and a number of secondary features with each other, it has been reported that the Oroch speakers used to have difficulties in understanding Udihe, while they were able to communicate with the Nanai and Ulcha, although the latter speak languages of a different Tungusic subgroup. This is apparently due to the many secondary and highly DOI: 10.4324/9781315728391-12

Oroch 295

idiosyncratic innovations present in Udihe, which have distanced it from both Oroch and the other Tungusic languages. The size of the Oroch ethnic population has in recent decades varied between 400 and 500 individuals. The exact number is difficult to establish because of the confusion with the “Orok” (Uilta) and “Orochen” (Ewenki) in the census data, but the number of Oroch in Khabarovsk Krai has been given as 426 (2002) and 441 (2010), of whom 8 individuals (2010) are supposed to have listed Oroch as their mother tongue. This information is, however, likely to be incorrect, for according to more reliable data the last speaker of Oroch died in 2008. The last speakers lived in two villages, Us’ka-Orochskaya and Datta, the former in the basin and the latter at the mouth of the river Tumnin, with some speakers also living in the nearby urban centers. By the present day (2022), however, not even semi-speakers are left. The language has never been officially taught at schools, but in recent years some practical tools, including a dictionary have been published, allowing samples of the language to be used occasionally for the purposes of ethnic revivalism. Oroch was traditionally divided into three dialects, corresponding geographically to the Tumnin, Khungari, and Khadi (with Koppi) basins. The Khadi dialect had several features reminiscent of Udihe, while the Tumnin and Khungari dialects were interacting with Nanai and Ulcha. The last speakers represented the Tumnin dialect, which has also served as the basis for the attempts at using the language in writing. TAXONOMIC STATUS Although both Oroch and Udihe are on geographical grounds conventionally classified as “Amur Tungusic” languages, it is generally recognized that they form a taxonomic group of their own and are by several primary innovations actually linked to the Ewenic languages, rather than to the areally closer Nanaic group. The following are some of the phonetic innovations that Oroch and Udihe share with Ewenic: •

• • •

The merger of *ü in the initial syllable and *U = *u & ü in non-initial syllables with *i, e.g. Oroch & Udihe tigde ‘rain’ = Ewenki tigde < *tügde > Nanai tugde; Oroch & Udihe taƞi- ‘to count, to read’ = Ewenki taƞ.i- < *taƞu- > Nanai taon-. An exceptional representation is present in at least one Oroch item, which must be a borrowing from Nanai: Oroch adʊli ‘fishing net’ ← Nanai adolɪ < *adulï > Ewenki adil.i = Udihe adili. The loss of initial *x and the subsequent development of initial *p to a new *x, e.g. Oroch & Udihe edin ‘wind’ = Ewenki edin < *xedün > Nanai xedun; Oroch & Udihe xokto ‘path’ = Ewenki (*)xokto < *pokto > Nanai pokto. The weakening of medial *p to *b > w, e.g. Oroch & Udihe jawa- ‘to take, to catch’ = Ewenki jawa- < *japa- > Nanai japa-. The progressive nasalization of the weak stops *b *d to *m *n after nasal consonants and the subsequent simplification of the arising nasal geminates, e.g. Oroch inaki ‘dog’ = Udihe ina’i = Ewenki ƞinakin < *ƞïnda-kïn > Nanai ɪnda; Oroch kamnʊ ‘glue’ = Udihe kamnu = Ewenki kamnu/n < *kamdo/n > Nanai kamdo/n. A related feature is the preservation of medial intervocalic *ƞ as opposed to the cluster *ƞg, e.g. Oroch & Udihe saƞa ‘hole’ = Ewenki saƞaar < *saƞaa/r > Nanai saƞgar. In one item Udihe, but not Oroch, shows the Nanai type of representation, implying a secondary borrowing from Nanai: Oroch toƞno ‘right, straight’ = Ewenki toƞno < *toƞdo > Nanai toƞdo → Udihe toƞdo.

296 Shinjiro Kazama

•

•

•

•

Among the innovations shared specifically by Oroch and Udihe a prominent place is occupied by the developments connected with the reduction of the phonemic load of the liquids *l and *r. A tendency to eliminate the vibrant *r, is also observed in Neghidal and, to a lesser extent, in the Nanai group, perhaps reflecting a common substratum. In clusters with a stop as the second component, the lateral *l developed into G = g k, e.g. Oroch doogdi- ‘to hear’ = Udihe dogdi- < *dooldï- = Ewenki dooldii- = Nanai dooljɪ-; Oroch kakta- ‘to break’ = Udihe kakta’a ‘half’ < *kalta- = Ewenic (*)kalta: (*)kalta-ka = Nanai kalta- : kalta-xa/n. As a secondary development, the cluster *lb > *gb underwent further assimilation to bb in Oroch, e.g. Oroch dobbo ‘night’ < *dogbo = Udihe dogbo < *dolbo = Ewenki dolbo = Nanai dolbo, while in Udihe the geminates *gg < *lg and *kk < *lk were simplified to g and k, respectively, e.g. Oroch diggan-a- ‘to say’ = Udihe digan-a- < *dïlgan- = Ewenki dilgan ‘voice’ = Nanai jɪlgan; Oroch sikki- ‘to wash’ = Udihe siki- < *sïlkï- = Ewenki silki- < *sïlku> Nanai sɪlko-. Under similar conditions, the vibrant *r underwent a development to G = g k, which means that its opposition with regard to *l was positionally cancelled. The accompanying assimilation and simplification developments were also analogous to those observed with *l, cf. e.g. Oroch gebbi ‘name’ < *gegbi = Udihe gegbi < *gerbii = Ewenki gerbii < *gerbüü > Nanai gerbu; Oroch еgdе ‘magic trick’ = Udihe egde ‘means’ < *erde/n = Nanai erde/n ‘means, trick’; Oroch iggi ‘tail’ = Udihe igi Ewenki irgi < *xürgü > Nanai xuigu; Oroch gappa- ‘to shoot’ < *gakpa- = Udihe gakpa- < *garpa- = Ewenki garpa- = Nanai garpa-; Oroch ukke ‘door’ = Udihe uke < *urke = Ewenki urke < *örke > Nanai uike. Before nasals *r may be assumed to have undergone an initial development to *g, which was nasalized to *ƞ. In this position, *r was nasalized also in Ewenic, but to the dental *n, cf. e.g. Oroch imme ‘needle’ < *iƞme = Udihe iƞme vs. Ewenki (*)inme < *irme < *xürme > Nanai xurme. In Udihe, the geminate ƞƞ arising from the cluster *rƞ is simplified to ƞ, e.g. Oroch xeƞƞe/n ‘knee’ = Udihe xeƞe < *xeƞƞe/n vs. Ewenki xenƞe/n < *xerƞe/n = Kilen xerƞ/en < *perƞe/n > Nanai peiƞe/n. In medial intervocalic position, *r is lost, yielding a contracted long vowel, e.g. Oroch & Udihe daama ‘(small of the) back’ < *darama = Ewenki (*)darama > Neghidal dayama; Oroch & Udihe goo ‘distant’ < *goro = Ewenki (*)goro > Neghidal goyo; Oroch uwee/n ‘mountain’ = Udihe ue/n < *xuree/n > Nanai xuree/n vs. Neghidal uyee/n.

Other developments shared by Oroch and Udihe involve mainly cases of assimilation, metathesis, or simplification of consonant clusters of various types: •

•

Assimilation with regard to the place of articulation is observed in the clusters *pt > (regressive) Oroch tt > (simplification) Udihe t and *mk > (progressive) *mp, e.g. (regressive) Oroch datta ~ data ‘estuary’ = Udihe data < *dapta ~ *dapto- > Ewenki daptu/n, cf. Nanai (irregular) daa; (progressive) Oroch & Udihe simpi- ‘to cough’ < *simki- > Kilen simki- = Ewenki simki- vs. Nanai (metathesis) siƞbi-si-. In the clusters *pk *bg Oroch and Udihe show a labial-velar metathesis shared on an areal basis with the Nanaic group, except with Kilen and Bikin Nanai. The resulting clusters have additionally undergone regressive assimilation in Oroch, e.g. jappʊ/n

Oroch 297

•

‘eight’ < *jakpʊ/n = Udihe jakpu/n = Nanai jakpo/n < *japku/n = Ewenki japku/n; Oroch subbu ‘fish skin’ < *sugbu = Udihe sugbu = Nanai sogbo < *subgu = Bikin Nanai sobgo. A basically similar development is shown by the clusters *mƞ and *mg, which seem to have merged to *mƞ in all languages except possibly in Nanai, e.g. Oroch amma ‘mouth’ < *aƞma = Udihe aƞma = Nanai aƞma < *amƞa = Ewenki amƞa; Oroch ommo- ‘to forget’ < *oƞmo- = Udihe oƞmo- < *omƞo- = Ewenki omƞovs. Nanai oƞbo- < < ?*omgo-. The intervocalic sibilant *s is preserved in mainstream (Tumnin) Oroch, as well as in Kilen, but is represented as x (h) in the Khadi dialect, as in Udihe. This also concerns the cluster *ns, which yields s ~ x, e.g. Khadi Oroch naxa ‘skin’ < *nasa = mainstream Oroch nasa = Kilen nasa < *nansa > Ewen (*)nanra vs. Nanai nanta. In the cluster *ms both components are preserved intact in Oroch and Udihe, as well as in Kilen, e.g. Oroch & Udihe giamsa ‘bone’ = Kilen giamsa < *gïramsa > Ewen (*)gɪramra. However, the cluster *ls yields different reflexes in Oroch vs. Udihe, with Oroch showing the development *ls > *lt > kt and Udihe the development *ls > *lAs > lAh, e.g. Oroch xʊkta ‘blanket’ vs. Udihe xulaha < *xulasa < *xulsa = Kilen xulsa > Ewen (*)xulra < *pulsa > Nanai polta.

DATA AND SOURCES The ethnocultural context of the Oroch is described in Lar’kin (1964) and Turaev & al. (2001), as well as, with texts and a glossary, by S. V. Bereznickii (1999). Linguistically, however, Oroch remains one of the least documented Tungusic idioms, and due to its recent extinction the situation is impossible to change. The language was first documented lexically in a concise vocabulary of 45 words under the false label “Ghilyak” (Ghilack or Ghailack) by the British naval surgeon John M. Tronson (1859). Other early vocabularies were published under the label “Orochen” by Prokopii Protodiakonov (1888) and Sergei Leontovich (1896), the latter containing also grammatical notes. There followed several other glossaries in ethnographic works, but the first linguistically reliable documentation of Oroch lexicon was published by Peter Schmidt (1928), based on materials collected in 1908. At about the same time, the Russian explorer V. A. Arsen’ev also collected lexical data from “Oroch”, but his materials, published much later (2008) by A. Kh. Girfanova and N. L. Sukhachev, are more focused on Udihe, which he at the time did not distinguish from Oroch. The first linguistic analysis of Oroch morphology, with an accompanying discussion of the position of Oroch among the Tungusic languages, was authored by V. I. Cincius (1949). Most intensively, Oroch was studied by V. A. Avrorin and E. P. Lebedeva, who published two volumes of folkloric materials, containing folktales, legends, riddles, and a dictionary with c. 5,500 entries (1966, 1978). They also authored a grammatical sketch (1968), later followed by another one by Lebedeva (1997). A more substantial grammatical description was published by B. V. Boldyrev on the basis of Avrorin’s materials (2001). More recently, fieldwork with the last speakers was done by Kazama, who published a report consisting of a basic vocabulary and 22 texts (1996), including audio recordings (available online). The latest fieldwork report by V. V. Baranova and K. A. Maslinskii (2014), based on information from 2001 and 2010, confirms the extinction of the language. Even so, materials for practical use, including a phrasebook and a picture dictionary, have been published by G. S. Abramova (1994, 2002). A larger dictionary for potential school use was published by Girfanova (2007).

298 Shinjiro Kazama

Issues pertaining to the history and taxonomy of Oroch, especially its relationship to Udihe, have been discussed by Avrorin (1979), Kazama (2009), and Alonso de la Fuente (2014). Alonso de la Fuente (2017) has also analysed the lexical data in Tronson’s glossary in view of their relevance to diachronic and dialectological studies. Ko Seongyeon (2012) is a synchronic analysis of the Oroch vowel system. SEGMENTAL PHONEMES Oroch has six qualitatively distinctive vowel phonemes, comprising the three rounded back vowels u ʊ o [u ʊ ɔ], the two unrounded central vowels e a [ə a], and the single front vowel i [i] (Table 12.1). The phonetic difference between ʊ and u is most clearly observed in the initial syllable, e.g. ʊgda [ʊgda] ‘boat’ vs. ugde [ugdə] ‘band’. In non-initial syllables the distinction tends to be lost, but the independent status of ʊ is confirmed by its occurrence in words containing no other vowels, e.g. bʊsʊ [bʊsʊ] ‘cloth’ and sʊnjʊ- [sʊndʑʊ-] ‘to select’. TABLE 12.1 OROCH VOWELS u

 

ʊ

e

o

a

i

The same qualities occur also as long (double) vowels: uu ʊʊ oo ee aa ii. Additionally, there is phonetically a long low front vowel [æ:], which has been analysed as a separate long vowel /ä/, but which is possibly better understood as the phonetic manifestation of the diphthongoid sequence ia, an interpretation followed here. Other vowel sequences include (ending in i) ui ʊi oi ei ai, (ending in u ʊ) eu aʊ iu, (beginning with u ʊ) ʊa ue. The sequence ia can also form the triphtongoids iaʊ iai. The status of the vowel sequences remains, however, open to alternative interpretations, especially with regard to the role of ʊ and ia in them. The consonant system is of the Common Tungusic type and comprises 18 phonemes (Table 12.2), divided between four places of articulation: labial, dental, palatal, and velar; and seven manners of articulation: nasals, weak stops, strong stops, fricatives, glides, a lateral, and a vibrant. The palatal stops j c are pronounced as affricates with a sibilant release. When preceding the vowel i, the sibilant s [s] is realized as alveolo-palatal [ɕ]. The opposition between the weak and strong stops is mainly based on voice (in the weak set). In the neighbourhood of the vowels ʊ o a the velar stops have back-velar or uvular realizations, e.g. biakaa/n [biaqa:] ‘river’. TABLE 12.2 OROCH CONSONANTS m

n

ñ

ƞ

b

d

j

g

t

c

k

p

s w

x y

l r

Oroch 299

Like Ewen, Solon, and Uilta, Oroch has conspicuously many geminates. Gemination can, in principle, concern all stops and nasals, but the most commonly attested examples are the stops bb gg pp cc kk and the nasals mm nn ƞƞ. Geminates are in most cases results of assimilation in original clusters, e.g. dobbo ‘night’ (< *dolbo), jappʊ/n ‘eight’ (< *jakpun < *japkun), amma ‘mouth’ (< *aƞma < *amƞa). PHONOTACTICS AND MORPHOPHONOLOGY The normal syllable structure of Oroch may be schematized as (C)V(V)(C). Single vowels can contrast with long vowels and diphthongs in all positions, except in monosyllables ending in a vowel, which are always pronounced long, i.e. (C)VV, e.g. 1SG bii. Exceptionally, however, the structure (C)VVV(C) is also observed, in which case the three-vowel sequence normally forms a triphthong starting with the sequence ia, e.g. miaʊca/n ‘gun’ (ultimately from Chinese niaoqiang 鳥鎗). The exceptional phonotactic status of the triphthongs could also be taken as an argument for treating ia as a separate vowel phoneme, i.e. /ä/. The vowels are arranged into two harmonic groups, the lower or “hard” vowels ʊ o a and the higher or “soft” vowels u e, with i remaining neutral. The “hard” vowels represent the original back qualities *u *o *a, while the “soft” vowels represent the original front qualities *ü *ö *e, but synchronically the distinction may also be analysed as involving tongue root position (retracted vs. advanced). A phonological word can only contain either “hard” or “soft” vowels, as well as the neutral vowel i, and the choice of vowels in suffixes is determined by the harmonic status of the root. It may be noted that roots containing the vowel ʊ take “hard” vowels in suffixes, e.g. bʊsʊ ‘cloth’ : ACC bʊsʊ-wa, sʊnjʊ- ‘to select’ : PTCP.PRF sʊnjʊ-xa/n. Roots containing the neutral vowel i are harmonically ambiguous and can also take “hard” vowels in the suffixes, e.g. isi- ‘to reach’ : PTCP.PRF isi-xan. This means that Oroch until recently had a distinction between i (< *i) and *ɪ (< *ï) and retains this distinction still synchronically at the lexical deep level. The sequence ia is normally combined with “hard” vowels, e.g. nia [ɲæ:] ‘man, person’ : PL nia-sa. Oroch also has a relatively prominent labial harmony, in that suffixes and clitics with the variants a e (< *A) normally take o after roots containing the vowel o, including the double vowel oo, e.g. omoo/n ‘lake’ : ACC omoo-mo. Labial harmony could possibly be viewed as a merely phonetic phenomenon, since Oroch, unlike, for instance, Nanai, retains the distinction between o and ʊ, but the presence of synchronic exceptions from the harmonic rule probably signals a partially lexicalized status. Also, the domain of the labial harmony does not extend across an original high vowel, including the neutral vowel i, cf. e.g. jolo ‘stone’ : VBLZ jolo-do- ‘to throw (a stone)’ vs. oroci ‘Oroch’ : VBLZ oroci-da- ‘to speak (in) Oroch’. Among the consonants, the vibrant r does not occur in word-initial position. In medial position, the original *r has largely been eliminated by deletive and assimilative processes, though synchronically r has been restored in loanwords, e.g. eru/n ‘time’ (apparently borrowed from Ulcha). The lateral l and the glides w y are also relatively rare segments in initial position, but their general frequency is raised by their use in suffixes. Morphophonological alternations of consonants concern the stem-final and suffixinitial positions. Stem-finally, the unstable /n is lost in absolute position, but it appears as a homorganic nasal segment m n (ñ) ƞ before suffixes beginning with a stop. A somewhat similar behaviour is shown by the stem-final g = /g, which is absent in absolute position

300 SHINJIRO KAZAMA

but appears as b g k ƞ before suffixes. Historically, g represents in these cases the merger of *g and *l. Suffix-initially, the glide w alternates with b after consonants, while in combination with a stem-final /n it is represented as m. Oroch prosody remains an unexplored field, but it is probably safe to say that the primary stress normally falls on the initial syllable, which also determines the harmonic status of the word, including suffixes and clitics. WORD FORMATION Nouns and verbs, or nominals and verbals, remain morphologically distinct and take also different sets of derivational suffixes. Nominals can also be transformed derivationally to verbals, and vice versa. (1) Denominal nouns: Many suffixes deriving denominal nouns are not fully productive and are only attested in lexicalized items. Also, some suffixes have a rather vague function, while others are very narrowly specialized. •

•

•

•

Semi-productive suffixes include -ptU/n, which indicates objects loosely connected with the meaning of the base noun, e.g. isa/g ‘eye(s)’ : isa-ptʊ/n ‘glasses’, xaña/n ‘shadow, spirit’ : xaña-ptʊ/n ‘mirror’, and -ksA, which denotes hides and skins of animals e.g. beyu/n ‘elk’ : beyu-kse ‘elk hide’, xeete ‘seal’ : xeete-kse ‘seal skin’. The suffix -mAgdA indicates nets used for catching specific kinds of fish, e.g. dawa ‘chum salmon (Oncorhynchus keta)’ : dawa-magda ‘chum salmon net’, peende ‘grayling (Thymallus arcticus)’ : peende-megde ‘grayling net’. An unproductive suffix shared with Ewenic, but absent in Nanaic, is -ki, as used in a few animal names, e.g. olo-ki ‘squirrel’, sʊla-ki ‘fox’, cf. also sewe/n ‘helping spirit of a shaman’ : sewe-ki ‘wooden idol (representing a spirit)’. The suffix -ƞkA/n forms nouns denoting inhabitants, e.g. namʊ ‘sea’ : namʊƞka/n ‘seaside dweller’. Most Oroch clan names, including modern surnames, are derived from river names by this suffix, e.g. Akur [river] : Akʊ-ƞka/n [Akunka family], Kopi [river] : Kopi-ƞka/n [Kopi family]. Evaluative suffixes include -kA(A)/n for diminutives, e.g. adʊli ‘net’ : DIM adʊli-ka/n, xite ‘child’ : DIM xite-ke/n ‘baby’, and -misA for pejoratives, e.g. amba/n ‘devil’ : PEJ amba-misa ‘contemptible devil’, tʊksa ‘bastard’ : tʊksa-misa ‘contemptible bastard’. No augmentatives have been attested in Oroch. In this connection it may be mentioned that kinship terms show irregular stem alternations, with one stem variant functioning as a reference form (normally with the possessive suffixes) and the other as an address form or “vocative”, e.g. amin- ‘father’ : VOC amaa, enin- ‘mother’ : VOC eñee; akin- ‘elder brother’ : akaa, ekin- ‘elder sister’ : VOC ekee ~ egee ~ exee, neƞu- ~ neu- ‘younger sibling’ : VOC nokoo. Similar stem variation is known from other Tungusic languages. The “vocative” stem is also used in some derivational forms, as in the terms denoting deceased relatives, formed by the suffix -ƞAsA, e.g. amin- : ama- ‘father’ : ama-ƞasa ‘late father’, enin- : eñe- ‘mother’ : eñe-ƞese ‘late mother’.

(2) Deverbal nouns: Apart from participles, which retain verbal functions, there are only a couple of common types of nouns derived from verbs:

Oroch 301

•

•

Actor nouns are formed by the suffix -mdi, e.g. gaʊ- ‘to punt’ : gaʊ-mdi ‘punter’, saa- ‘to know’ : saa-mdi ‘knowledgeable person’, etuu- ‘to watch’ : etuu-mdi ‘watch, guard’, tuƞsere- ‘to translate’ : tuƞsere-mdi ‘translator’. Another, less productive actor noun suffix is -ƞga ~ -ƞgʊ, e.g. agga-la- ‘to deceive’ : aggaƞga ‘lier, crafty man’, baatʊ- ‘to hunt’ : baatʊ-ƞgʊ ‘hunter’. The suffix -ƞki forms nouns denoting the place or instrument of action, e.g. muule‘to pick up water’ : muule-ƞki ‘bucket’, niru- ‘to write’ : niru-ƞki ‘pen, pencil’.

(3) Denominal verbs: Oroch retains a few Common Tungusic nomina-verba, e.g. tigde ‘rain’ : tigde- ‘to rain’. Normally, verbs are derived from nouns by suffixes. Many of the derived forms are lexicalized in specific meanings. • • •

The principal verbalizing suffix is -lA-, which has a wide range of uses, e.g. kokto ‘gloves’ : kokto-lo- ‘to wear gloves’, yoso ‘lock’ : yoso-lo- ‘to lock’, asa ‘woman, wife’ : asa-la- ‘to marry (of men)’. Another verbalizer is -dA-, e.g. jolo- ‘stone’ : jolo-do- ‘to throw’, kiaka ‘Udihe’ : kiaka-da- ‘to speak (in) Udihe’. Captative verbs are formed by the suffix -mA- e.g. beyu/n ‘elk’ : beyu-me- ‘to hunt for elks’, dawa ‘chum salmon (Oncorhynchus keta)’ : dawa-ma- ‘to catch chum salmon’. Translative verbs are derived from adjectival nominals by the suffix -nA-, e.g. masi ‘strong’ : masi-na- ‘to become strong’, okki/n ‘bad’ : okki-na- ‘to become bad’, sakai/n ‘black’ : sakai-na- ‘to become black’.

(4) Deverbal verbs: Most types of deverbal verbs are grammaticalized into various functions connected with voice, aspect, or Aktionsart. Following are a couple of other types involving a more clearly identifiable lexical meaning: •

•

Andatives are formed by the suffix -nA- ~ -ƞnA-, e.g. ice- ‘to see’ : AND ice-ne~ ice-ƞne- ‘to go to see’, jep- ‘to eat’ : AND jepu-ne- ‘to go to eat’. Sentence examples: bii aaki-mi beyu-me waa-na-xa-ni ‘my elder brother went to hunt for elks’ (1SG elder.brother-​PX1SG elk-ACC kill-AND-PST-3SG), jawi-wi doolo-ni ʊʊ-ƞna-xa-ni ‘s/he went to get on the boat’ (boat-RX inside-LOC-PX3SG get.on-AND-PST-3SG). The transitive correlates of motion verbs are formed by the primary causativepassive suffix -bU- ~ -wU- ~ -U-, e.g. ii- ‘to go in’ : CAUS ii-wu- ‘to put/bring in’, cʊca- ‘to run away’ : CAUS cʊca-ʊ- ‘to release’.

NUMBER AND CASE Both animate and non-animate nouns can take the plural suffix -sA/g (< *-sA-l), but most of the actual examples involve nouns denoting humans, e.g. andaxa ‘guest’ : PL andaxa-sag. Some words have irregular lexicalized plural forms, e.g. xite ‘child’ : PL xii/g ‘children’. A special plural marker present only in Oroch and Udihe is -ntA, as used in asa ~ asa/n ‘woman’ : PL asa-nta ‘women’ and xuse ‘male’ : PL xuse-nte ‘men’. Associative plurality is marked by the suffix -nAA, e.g. amin- ‘father’ : PL ASSOC ami-naa ‘father and others, parents’, akin- ‘elder brother’ : PL ASSOC aki-naa. A similar function can also be expressed by the suffix -yA, e.g. ama-ya ‘father and others’, sʊlaanjiga-ya ‘Sulanjiga people [an Udihe clan] and others’. Collective plurality is expressed by the suffix -jikA (= Udihe -jigA = Nanai -jOAn < *-jUkan), e.g. xikke ‘young’ : COLL xikke-jike ‘youth’, pasi ‘piece’ : COLL pasi-jika ‘pieces’.

302 Shinjiro Kazama

Adjectival nominals, demonstrative pronouns, and participles can form a collective plural by the suffix -gAtU < +getu ‘people’, e.g. sagdi ‘old’ : COLL sagdi-gatʊ ‘old people’. Occasionally, this suffix can take the plural marker -sA/g, cf. e.g. taa-dʊ xikke-getu-se ʊgda-wa oo-i-ti ‘the young men are making a boat there’ (that-DAT young-COLL-PL boat-ACC make-PRS-3PL). Examples with participles: bolo-xo-gatʊ beyu-me seenju-we waa-xa-ti bi-ci-ti ‘the people who used to hunt in the autumn caught elks and char’ (become.autumn-​PTCP.PRF-COLL elk-ACC char-ACC kill-PST-3PL be-PST-3PL), axa-nta meneji-​xe-getu gebbeƞku-we ga-ci-ti ‘the women who stayed there got some berries’ (woman-PL stay-​PTCP.PRF-COLL berry-ACC get-PST-3PL). The Oroch case paradigm comprises the unmarked nominative and eight suffixally marked cases: accusative, designative, dative, ablative, locative, prolative, directive, and instrumental (Table 12.3.). The case markers show some allomorphic variation depending on whether they are added to stems ending in a vowel (V), the primary nasal /n, or the consonant /g (< *g & *l). The segments /n and /g are unstable or “hidden”, in that they do not appear prepausally, but only before suffixes, in which position /g is also represented as b and k, e.g. isa/g ‘eye(s)’ : DAT isag-du : ACC isab-ba : DIR isak-ti. TABLE 12.3 OROCH CASE MARKERS  

V

/n

/g

ACC

-A

-mA

b-bA

DES

-yA ~ -lA-

-nA- ~ -lA-

g-lA-

DAT

-dU

n-dU

g-dU

ABL

-dUi

n-dUi

g-dUi

LOC

-lA

n-dU-lA

g-dU-lA

PROL

-li

n-dU-li

g-dU-li

DIR

-ti

n-ti

k-ti

INSTR

-ji

n-ji

g-ji

Stem types: V = vowel stems, /n = nasal stems, /g = stems ending in an unstable g. Segmental alternations: A = a e o, U = ʊ u.

•

•

The nominative signals the subject of both intransitive and transitive verbs, e.g. eñee, ama eme-i-ni ‘mother! father is coming’ (mother.VOC father come-PRS-3SG). It is also the case of the nominal predicate, e.g. sii esi bii asi-mi ‘you are my wife now’ (2SG now 1SG wife-PX1SG), taa-dʊi namʊ-ƞka-sa jaƞgia o-ci-ti ‘then the Namunka clan became leaders’ (that-ABL sea-DX-PL leader become-PST-3PL). Before a noun marked for possession the nominative indicates the possessor, e.g. sama untu-ni ‘(the) shaman’s drum’. The accusative marks the direct object, e.g. sʊlaki xokto-wo-ni ice-xe-ni ‘s/he saw the traces of a fox’ (fox path-ACC-PX3SG see-PST-3SG). Rarely, accusative marking can also be absent with no functional difference. With verbs of motion the accusative is functionally close to the prolative and indicates the route of movement, e.g. taa-dʊi ʊli-wa eye-ki ƞene-xe-ni ‘then s/he went downstream along the river’ (thatABL river-ACC downstream-DIR go-PST-3SG). The accusative can also express the progress of time, e.g. taa-dʊ adi aƞƞani-wa bi-mi buk-ki-ti ‘then, some years passed and they died’ (that-DAT how.many year-ACC be-CV.CONN die-PST-3PL).

Oroch 303

•

•

•

The verb isi- ‘to reach, to arrive’ can be used transitively with an object in the accusative, e.g. megge omoo-mo isi-xa-ni ‘the hero reached the lake’ (hero lake-ACC reach-PST-3SG), but it can also be used intransitively with the goal in the locative or directive case. The accusative is also attested as marking the goal of motion with the verb eme- ‘to come’, e.g. tooyo mapaaca megge-me laƞ eme-xe-ni ‘then grandfather came near the hero’ (then grandfather hero-ACC near come-PST-3SG). In emotional expressions, adjectives can take the accusative marker, e.g. annayia xoto-mo, saagdi=kia ‘oh, what a big town it is!’ (INTERJ town-ACC big.EMPH=PTCL). When used in combination with numerals, the accusative can also function as the marker of approximate number, e.g. tʊƞa taƞgʊ daab-ba bi-i ʊsi-wa jawa-xa-ni ‘s/he took a leash about five hundred fathoms long’ (five hundred fathom-ACC be-PTCP.PRS leash-ACC take-PST-3SG). The designative is used only with the possessive or reflexive markers. This case is mainly required by “effective verbs”, whose object represents the result of the action, “something designated for someone”, e.g. jʊba-la-ni oo-xo-ni ‘s/he made a house for her/him’ (house-​DES-PX3SG make-PST-3SG), je(w)u-le-ni buu-xe-ni ‘s/ he gave her/him some food’ (food-DES-​PX3SG give-PST-3SG). Occasionally, the designative can co-occur with the accusative in a single clause, e.g. jaa-la-yi juu nia-wa baa-xa-mi ‘I found two men to become my friends’ (friend-DES-RX two man-ACC find-PST-1SG). The dative marks the indirect object and the location or time of an activity, e.g. tii-du xite-ke-du ʊpa-wa buu-xe-ni ‘then s/he gave some flour to the small child’ (thatDAT child-DIM-DAT flour-ACC give-PST-3SG), meene meene jʊʊg-dʊ-wi bi-ci-ti ‘they were in each other’s houses’ (REFL REFL house-DAT-RX be-PST-3PL), buu bagdi-xa-mʊ lʊca jʊba-dʊ-ni ‘we were living in a Russian house’ (1PL.EXCL live-PST-1PL.EXCL Russian house-DAT-PX3SG), ineƞi tokon-dʊ-ni bii ami-mi eme-jee-ni ‘at midday my father will come’ (day middle-DAT-PX3SG 1SG fatherPX1SG come-FUT-3SG). With verbs like nee- ‘to put’, lowo- ‘to hang’, tee- ‘to sit’ the dative expresses the location where the subject or object is after the action. With verbs like japa- ‘to take, to catch’ the dative implies a part of an entity (e.g. a body part) when the contact continues for some time, e.g. tii nia-wa ƞaala-dʊ-ni jawaxa-ni ‘s/he took the man by the arm’ (that man-ACC hand-DAT-​PX3SG catch-PST3SG). Nouns denoting persons experiencing an emotion or feeling are marked by the dative, e.g. xite-du aa-mʊsi maƞga ‘the child is very sleepy’ (child-DAT sleep-DESID. PRS hard). In the causative construction, the causee stands in the dative, e.g. min-du umukse-we e-ci-ni jeu-kune-ye ‘s/he did not make me eat the fat!’ (1SG-DAT fatACC NEG-PST-3SG eat-CAUS-CONNEG). When attached to participles, the dative yields quasiconverbal forms functioning as predicates of subordinate clauses with a temporal meaning (‘when’, ‘after’), e.g. tuu xite-ji sori-i-dʊ-ni ami-ni eme-xe-ni ‘so when s/he was fighting with the child his/her father came back’ (thus child-INSTR fight-PTCP.PRS-DAT-PX3SG fatherPX3SG come-PST-3SG), noƞo-ni jʊʊk-ti ii-gi-i-du-yi guƞ-ki-ni ‘when s/he entered the house, s/he said’ (3P-PX3SG house-DIR go.in-REV-PTCP.PRS-DAT-RX say-PST3SG), eeki paatila-xan-dʊ-ni siƞe meene jʊʊk-ti ƞei-xe-ni ‘after the frog had hit (it), the mouse went home’ (frog hit-PTCP.PRF-DAT-​PX3SG mouse REFL house-DIR go-PST-3SG). The ablative indicates the starting point in place or time, e.g. bii xoƞto baa-dʊi emegi-i ‘I have come from another place’ (1SG other place-ABL come-REV-PRS). It

304 Shinjiro Kazama

•

•

•

•

also marks the standard of comparison, e.g. ei moo tii moo-dʊi gogda ‘this tree is taller than that tree’ (this tree that tree-ABL tall). The locative indicates location at a restricted place, including, in some cases, an exact point within an entity, e.g. bii mapa dili-la-ni miaʊca-la-xa-mi ‘I shot the bear in the head’ (1SG bear head-LOC-PX3SG gun-VBLZ-PST-1SG). It can also refer to the source or context of an object or individual, e.g. gaam nia-sag-dʊla ila nia-wa sʊnjʊ-xa-ti ‘from among all the men they chose three’ (all man-​PL-LOC three manACC choose-PST-3PL). In some uses it expresses locational limit, e.g. nukte-ni=dee naa-la isi-gi-xa-ni ‘her hair (was so long that it) reached the earth’ (hair-​PX3SG=​ PTCL land-LOC reach-REV-PST-3SG). In a more abstract use, the locative can also refer to the cause of a situation, e.g. suu okkin-dʊla-sʊ bii jaab-bi gaam waa-xa-ti ‘they killed all of my fellows because of your mistake’ (2PL bad-LOC-PX2PL 1SG friend-PX1SG all kill-PST-3PL). The prolative indicates the place through which a motion goes or an activity is carried out, e.g. saƞa-li ice-ci-i-ni ‘s/he is looking through the hole’ (hole-PROL seeDUR-​PRS-3SG), taa-dʊ puji paawa-li ice-xe-ni ‘then the heroine looked through the window’ (that-DAT heroine window-PROL see-PST-3SG), tukki xokto-li-ni ƞenexe-ni ‘s/he went along the tracks of the sleigh’ (sleigh path-PROL-PX3SG goPST-3SG). The directive indicates the direction of a motion or action, e.g. jʊba-ti ii-xe-ni ‘s/ he went into the house’ (house-DIR go.in-PST-3SG), omo sʊlaki namʊ-ti tutu-xe-ni ‘a fox ran towards the sea’ (one fox sea-DIR run-PST-3SG). It can also be used with verbs of perception, e.g. namʊ-ti ice-i-ni ‘s/he is looking at/towards the sea’ (sea-DIR see-PRS-3SG). With verbs of saying and emotional reactions the directive denotes the addressee, e.g. xitu-ti gune-i-ni ‘s/he says to the child’ (child-DIR sayPRS-3SG), taa-dʊ edi-ni asan-ti-wi tagda-xa-ni ‘then the husband got angry with his wife’ (that-DAT husband-PX3SG wife-DIR-RX get.angry-​PST-​3SG). When added to participles, the directive forms a quasiconverbal construction indicating a temporal limit, e.g. nʊƞa-ni sikse=dee xosikta-ti odi-gi-mi bʊta-xa-ni ‘s/he used to fish until the stars stopped twinkling’ (3P-PX3SG evening=PTCL star-DIR stopREV-CV. C ​ ONN fish-PST-3SG). The instrumental is a multifunctional case. Basically, it indicates an instrument or a material, e.g. xʊnnʊi-wa tada-ji gappa-​ci-xa-mi ‘I used to shoot grouse with an arrow’ (grouse-​ACC arrow-INSTR shoot-DUR-​PST-1SG). It has, however, also the function of a comitative, linking coordinated and/or reciprocal actors, e.g. pujin-ji meggen-ji tuu tee-ci-i-ti ‘the heroine and the hero were sitting like that’ (heroine-​INSTR hero-INSTR so sit-DUR-PRS-3PL), tei mapa-ji meggen-ji tuu sori-maci-li-xa-ti ‘the grandfather and the hero began to fight’ (that grandfather-INSTR hero-INSTR fight-RECIPR-INCH-PST-3PL). When attached to adjectival nominals, the instrumental forms modal adverbs, e.g. aya-ji bi-jee-pi.e ‘let us live well’ (good-INSTR be-FUT-1PL.INCL​.EXT). The instrumental is also used with some verbs of emotion, e.g. bii sʊlaki-ji e-si-mi ƞeele-ye ‘I am not afraid of the fox’ (1SG fox-INSTR NEG-PRS-1SG be.afraid-CONNEG), taa-ji ilaji-mi e-ci-ni eme-gi-ye ‘s/he did not come back because s/he is ashamed of it’ (that-INSTR be.ashamed-CV.CONN NEG-PST-3SG come-​REV-CONNEG). Some descriptions of Oroch (Avrorin & Lebedeva) mention also a distinct separative case in -jiji, e.g. ʊgda : SEP ʊgda-jiji ‘from the boat’, but this feature is connected with Nanai influence and rarely used in Oroch.

Oroch 305

The instrumental marker can also be attached to verbs with the help of the otherwise non-productive nominalizing suffix -n-, yielding a quasiconverbal form in -n-ji-, e.g. tada-ji gappa-n-ji waa-xa-ni ‘he killed it by shooting with an arrow’ (arrow-INSTR shoot-NMLZ-​INSTR kill-PST-3SG). •

•

The comitative function can also be expressed by the suffix -mUnA ‘together with’ (cf. Nanai -mOlIA), e.g. asa-mʊna daʊ-gi-xa-ti baggi-si ‘he crossed the river with his wife’ (wife-COM cross-REV-PST-3PL other.side-DIR), suu juu nii-kee eni-mune bi-i-si=nuu ‘do you live with your mother?’ (2PL two person-DIM mother-COM be-PRS-2SG=INTERR). Like most other Tungusic languages, Oroch has no adnominally used genitive case, the possessive relationship being expressed by the basic form (nominative), combined with a possessively marked head noun. In predicative position, the Common Tungusic possessive marker -ƞi ~ -ƞʊ is used, e.g. ei jʊbba iskolo-ƞi ‘this house is the school’s ~ belongs to the school’ (this house school-POSS). Such possessive forms function as new nominal bases and can themselves be marked for number and case.

A special group of nouns is formed by the spatials (locational nouns), used as local/ temporal adverbs, as well as in the role of postpositional headwords of constructions in which they define the local or temporal position of an adnominal noun. As nouns, they can take marking for case and possession, but their paradigm is defective and involves differences as compared with regular nouns. One of the exceptional features of spatials is that they take the locative case, and not the dative, for expressing static location, e.g. taa-dʊ sii moo xeggi-le-ni xiri-ja-si ‘you should hide under the tree there’ (that-DAT 2SG tree underside-LOC-PX3SG hide-FUT-2SG), buu julie-le-mu egdi nia ƞene-xe-ti ‘many people went (there) before us’ (1PL.EXCL frontside-LOC-PX1PL.EXCL many man go-PST-3PL), sʊlaki lop tii-xe-ni baaca oyo-lo-ni ‘the fox fell to the island with a bang’ (fox ONOM fall-PST-3SG island upperside-LOC-​PX3SG). The locative can also convey a temporal meaning, as in [1]: [1] cʊca-xan amaa-la-ni run-away-ptcp.prf after-loc-px3sg ‘After he had run away gogdi asa-ma-ni xite-we-ni child-acc-px3sg Nanai woman-acc-px3sg the Nanai people killed them all: his wife and his children.’

gaam all

waa-xa kill-pst

Other features of spatials include the special directive in -si and prolative in -ki, and the use of the accusative also in a prolative function. All spatial paradigms are lexicalized and often contain a combination of regular and special forms, cf. e.g. jule- ‘front side’ : DIR jule-si ‘forwards’ : PROL jule-ki ‘ahead’, amaa- ‘back side’ : DIR amaa-si ‘backwards’ : PROL amaa-li-PX ‘behind’, ui- ‘upper side’ : LOC ui-le ‘above’, DIR ui-si ‘upwards’ : PROL ui-ki ‘over’, xeggi ‘underside’ : DIR xeggi-si ‘downwards’ : ACC xeggi-me ‘under’. Places and directions are typically determined on the basis of their relative position to rivers: DIR soloo-ki ‘up the river’, DIR eyee-ki ‘down the river’, LOC dii-le : DIR dii-si ‘mountainside, far from the river’, LOC ƞia-la : DIR ƞia-si ‘riverside, near the river’.

306 Shinjiro Kazama

Apart from the actual spatials, there is a category of secondary spatials, that is, locational nouns that follow the nominal declension. This group includes, for instance: doo‘inside’, agdan- ‘between’, baggia- ‘opposite side’, jappa- ‘beside’, kiya- ‘edge, end’, ogdon- ‘side’, and others. Secondary spatials are also formed by the suffix -ji ‘side’, e.g. baggi-ji ‘opposite side’, maaƞgʊ-ji ‘the Amur river side’ (maaƞgʊ ‘Amur’). ADJECTIVES Adjectives in Oroch are nominal words with some derivational and syntactic idiosyncracies. They can also occur as independent headwords, equal to nouns, in which function they can take number and case markers, e.g. egdi ‘many’ ~ ‘much’ (adjectival quantifier) : egdi-we saa-xa-ti ‘they knew a lot of things’ (many-ACC know-PST-3PL). •

•

•

The most important and fully productive adjectival category is formed by the proprietive adjectives, formed from nouns denoting relatives, body parts, parts of entities, portable things, etc., and marked by the suffix -ki (= Udihe -xi = Nanai -kO < *-lkU). Proprietive adjectives are the most common means for the expression of possession in both adnominal and predicative position, e.g. noƞon-ti seeƞgi-ki o-ci xite-ki o-ci ‘they got both relatives and children’ (3P-PX3PL relative-PROPR become-​PST child-PROPR become-PST), gida-ki nia miaʊcaƞ-ki nia-ti guƞ-ki-ni ‘the man with a spear said to the man with a gun’ (spear-PROPR man gun-PROPR man-DIR say-PST-3SG), seekse-ki asa-nta-wa laki-la=daa e-si-ti eme-kun-e ‘they do not let menstruating women come near’ (blood-PROPR woman-PL-ACC near-LOC=PTCL NEG-​PRS-3PL come-CAUS-CONNEG). When the noun marked with the proprietive suffix has a nominal modifier, the latter stands in the instrumental case, e.g. gaamdi aisin-ji=maki xajʊƞ-ki ‘(she) has entirely golden clothes’ (all gold-INSTR=​PTCL clothes-​PROPR), aaki-ni sʊlaki-ji tekte-ki, neƞu-ni cibje-ji tekte-ki ‘the elder brother wears the clothes of a fox, while the younger brother wears the clothes of a lynx’ (elder.​brother-​PX3SG fox-INSTR clothes-PROPR younger.brother lynx-INSTR clothes-​PROPR), omo moo-​dʊ nadan-ji xite-ki umuki bi-ci-n(i) ‘there lived a flying squirrel with seven children on a tree’ (one tree-DAT seven-INSTR child-PROPR flying.squirrel be-PST-3SG). The corresponding privative marker is -lAci < DES -lA + *aci (cf. Udihe NEG. EXIST anci ‘absence, absent’), as in [2b]: omo mapaaca caagjan-ji wayaƞ-ki, gagda-ni waya-laci ‘one of the bears had a white pattern, the other had no pattern’ (one bear white-INSTR pattern-​PROPR other-PX3SG pattern-PRIV), inaki digga-laci o-ci-ni ‘the dog stopped barking’ (dog voice-PRIV become-PST-3SG). Privation can also be expressed analytically by the privative noun anaa, e.g. nʊƞan-ti jʊʊ anaa ooni bagdi-jaƞa-ti ‘how will they live without a house?’ (3P-PX3PL house PRIV how liveFUT-3PL), bii ami=daa anaa eni=dee anaa ‘I have neither a father nor a mother’ (1SG father=PTCL PRIV mother=PTCL PRIV). Derivationally formed degree adjectives include the moderatives in -ƞkʊ (= Udihe -aƞku), e.g. xite-ti sagdi-ƞkʊ o-ci-ti ‘their children became fairly old’ (child-PX3PL old-​MODER become-PST-3PL), and -lAA, e.g. egdi-lee waa-xa-ni ‘s/he caught rather many’ (many-MODER kill-PST-3SG), yaawʊ tugge-lee-ni xuli-kte-xe-si=a ‘why did you come back so quickly?’ (why quick-MODER-​PX3SG go.around-​DISTR-​ PST-​2SG=CORROG), nʊƞan-ti aya-laa bi-i jolo baa-xa ‘they found a rather good

Oroch 307

•

stone’ (3P-​PX3PL good-MODER be-PTCP.PRS stone find-PST). The contrastiveemphatic suffix -dUmA can be used on both nouns and adjectival nominals, e.g. ekin-dume gune-i-ni neƞun-ti ‘the elder sister said to the younger sister’ (elder.sister-CONTR say-PRS-3SG younger.sister-DIR), miaʊcaƞ-ki-dʊma miaʊca-la-ica-i-ni ‘the one with a gun intended to shoot’ (gun-PROPR-CONTR gun-​VBLZ-​INTENTPRS-3SG). Other denominal derivational suffixes attested on adjectives include -mA [material], e.g. ookto-mo jʊʊ ‘a house made of grass’ (grass-DX house), bʊsʊ-ma paawa ‘a window made of cloth’ (cloth-DX window), sʊlaki-ksa-ma aapʊ ‘a hat made of fox fur’ (fox-DX-DX hat); -si [possessing a property], e.g. amta ‘taste’ : amta-si ‘tasty’, maƞa ‘hard(ness)’ : maƞa-si ‘hard’, xeku ‘heat’ : xeku-si ‘hot’, [with no synchronic function] ñama ~ ñama-si ‘warm’, xakti ~ xakti-si ‘dark’; -pti [temporal relation], e.g. sikse-pti andaxa ‘a guest who came in the evening’ (evening-DX guest), dobbo-pti sʊgjasa ‘fish caught in the night’ (night-DX fish), ainaa-pti teeluƞu ‘ancient story’ (long.ago-DX tale).

NUMERALS The numerals of the first decade are: 1 omo/n, 2 juu/g, 3 ila/n, 4 dii/n, 5 tʊƞa ~ tʊƞa/n, 6 ñuƞu/n, 7 nada/n, 8 jappʊ/n, 9 xuyu/n, 10 jaa/n. For the lower decades, the lexically separate items 20 oi/n and 30 gʊti/n (ultimately from Mongolic) are used, while the numerals for the subsequent decades are formed suffixally by -jAA/n (< 10 +jaa/n): 40 diin-jaa/n, 50 tʊƞan-jaa/n, 60 ñuƞun-jee/n, 70 nadan-jaa/n, 80 jappʊn-jaa/n, 90 xuyun-jee/n. For the powers of ten, 100 (omo) taƞgʊ (from Manchu) and 1000 (omo) miƞga (from Mongolic via Manchu) are used. All numerals with the exception of 2 juu/g, 5 tʊƞa, and 100 taƞgʊ are historically nasal stems, containing lexically the primary nasal /n, which appears in inflected forms. The item 5 tʊƞa has, however, also the nasal stem tʊƞan-, formed analogically after the other numerals, and its accusative form is tʊƞa-ma, cf. e.g. 3 ila/n : ACC ila-ma, 4 dii/n : ACC dii-ma. This feature of the item for ‘five’ is shared with Udihe, suggesting that the analogical transformation to nasal stems took place in the common ancestor of Oroch and Udihe. The numeral 9 xuyu/n deserves attention. For this item, Oroch conforms with Nanaic (Nanai-Ulcha-Uilta xuyu/n) and Ewen (uyun), while Udihe ye(y)i conforms with mainstream Ewenic (Ewenki-Neghidal-Solon yegin, Orochen yeyin). The two forms are not immediately compatible, suggesting that they derive from two intermediate protoforms, *xöyöö/n and *yegü/n, respectively. Considering also the fact that Oroch, like Udihe and Ewenic, has regularly lost initial *x, it must be concluded that Oroch 9 xuyu/n is a borrowing from Nanaic (either Nanai or Ulcha). The borrowing must have taken place relatively recently, since it is not shared with Udihe. In this process Oroch lost the inherited numeral for ‘nine’. Grammatically, numerals behave like nouns and can take case marking, e.g. ila-ma aaƞa-i-dʊ-bbayi gaam mute-xe-ti ‘staying there three days, they could finish everything’ (three-ACC stay-PTCP.PRS-DAT-RX.PL all be.able-PST-3PL). Ordinal numerals are formed by the suffixes -ƞi ~ -(w)ii-ƞi ~ -(w)ia-gi, e.g. omo-ƞi ‘first’, juu-ƞi ~ juu.w-ii ~ juu.w-ia ~ juu.w-ia-gi ‘second’, ila-ƞi ~ il-ia-gi ‘third’, etc. The forms are highly variable and rarely used, mainly in reference to the order of children in a family, and often with possessive suffixes. For ‘first’, the suppletive items ñaʊki ~ ñooki and adama ~ ʊdama ‘firstborn’ : ORD adama-ƞi ~ ʊdama-ƞi are also used, as in [2]:

308 Shinjiro Kazama

[2] nʊƞan.ti-dʊ ila xite bi-ci-ni: adama-ƞi-ti 3p.px3pl-dat three child be-pst-3sg firstborn-ord-px3pl ‘They had three children: the first one was a boy, juu-wii.ƞi-ti asa xite, ila-wii.ƞi-ti ñaa two-ord-px3pl female child three-ord-px3pl again the second was a girl, and the third was again a boy.’

xuse male

xuse male

xite, child

xite child

Collective numerals are marked by -mi.sA ~ -ƞA.sA (containing the plural marker -sA), e.g. juu-mi.se ~ juu-ƞe.se ‘the two of them, both’, ila-mi.sa ~ ila-ƞa.sa ‘the three of them’. The diminutive suffix -kA/n ~ -kkA is used on numerals to express limitation (‘only’), e.g. omo-kko/n ‘only one’, juu-kke/n ‘only two’, ila-kka/n ‘only three’. This suffix can also be attached to a noun accompanying the numeral, e.g. eni-ji juu nii-ke bagdi-i-mʊ ‘I live with my mother, only the two of us’ (mother-INSTR two person-DIM live-PRS-1PL.EXCL). The Common Tungusic counter for days is attested in Oroch in the form -ktA (< *-ltA), e.g. ila-kta ‘three days’, dii-kta ‘four days’. A special form is emne ‘once’. PRONOUNS The Oroch basic personal pronouns are 1SG bii : 2SG sii : 1PL EXCL buu : 2PL suu, with the oblique stems 1SG min- : 2SG sin- : 1PL EXCL mun- : 2PL sun- (Table 12.4).

TABLE 12.4 OROCH PERSONAL PRONOUNS SG PL

 

NOM

OBL

ACC

POSS

1

bii

min-

mine-we

mini-ƞi

2

sii

sin-

sine-we

sini-ƞi

buu

mun-

mune-we

muñu-ƞu

suu

sun-

sune-we

suñu-ƞu

1 2

EXCL

The accusative forms are of the same type as in Ewenic, based on the designative-like vowel stems 1SG min.e- : 2SG sin.e- : 1PL EXCL mun.e- : 2PL sun.e-, to which the regular accusative marker -we is added. The system comprises the possessive forms 1SG mini-ƞi : 2 SG sini-ƞi : 1PL EXCL muñu-ƞu (< *muni-ƞi) : 2PL suñu-ƞu (< *suni-ƞi), which express the possessor in predicative position, e.g. ei inaki sini-ƞi ‘this dog is yours’ (this dog 2SG-POSS), but which otherwise are nominal bases that can also take other case suffixes. Oroch, like Udihe, has also a separate pronoun for the first person inclusive function: 1PL INCL biti : ACC biti-we : OBL biti- : POSS biti-ƞi. For the third person the stem nʊƞa/n- ~ noƞo/n- is used. In the singular, it takes the possessive suffix PX3SG -ni after the case markers: NOM nʊƞa-ni : ACC nʊƞa-ma-ni : DAT nʊƞan-dʊ-ni : LOC nʊƞan-dʊla-ni, etc. In the plural, however, the form nʊƞan-ti ~ noƞon-ti, which historically contains the third person plural possessive suffix -ti, has been reanalysed as a new nominal stem: NOM nʊƞanti : ACC nʊƞanti-wa : DAT nʊƞanti-dʊ : LOC nʊƞanti-la, etc., a feature shared with Udihe and Kilen. The corresponding

Oroch 309

possessive forms are SG nʊƞa-ƞi-ni : PL nʊƞanti-ƞi. The functional properties of the third person pronouns may, however, require further consideration, for more often than as true third person pronouns they are used in an obviative function, referring to a newly introduced (“fourth person”) actor. Reference to actual third person actors is more typically made with the help of demonstrative pronouns, often in combination with generic nouns like nia ‘man, person’. There is also an indefinite pronoun with the shape xaa/n- : PX3SG xaa-ni : OBL xaa-CX-ni ‘some, someone’, e.g. tias egdi, xaa-ni saagdi-ƞkʊ, xaa-ni ƞiici-ƞku deili-i-ti ‘very many (birds) were flying, some-PX3SG of them big, some-PX3SG of them small’ (very many some big-DX some small-DX fly-PRS-3PL), taanti-wa baa-(y)i-ti, oroci-sa jawa-i-ti xaa-wa-ni ‘the Oroch found such things and took some of them’ (such-ACC find-PRS-3PL Oroch-PL take-PRS-3PL some-ACC-PX3SG). The reflexive pronoun is mee/n-, which in inflected forms takes the marker of reflexive possession -yi. The accusative is exceptional, REFL-ACC-RX meeb-be-yi, and an extension element -be- ~ -bu-, similar to the accusative marker, is also present after monosyllabic case markers: DAT-EXT-RX meen-du-bu-yi : ABL-EXT-RX meen-dui-be-yi : LOC-RX meen-​dule-yi : PROL-RX meen-duli-yi : DIR-EXT-RX meen-ti-be-yi : INSTREXT-RX meen-ji-be-yi. In independent adverbial (emphatic) use the form meene is used, e.g. nʊƞanti meene eme-xe-ti ‘they came (by) themselves’ (3PL REFL come-PST-3PL). The form meen(e) is also used adnominally in a possessive function, and when reduplicated it expresses reciprocality, e.g. nʊƞanti meene meene jʊʊg-dʊ-wi bi-ci-ti ‘they were in each other’s house’ (3PL REFL REFL house-DAT-RX be-PST-3PL). The possessive form is meeni-ƞi, and although basically not used adnominally, it is nevertheless attested also in this position, functioning as a genitive, e.g. noƞo-ni ƞei-xe-ni meeni-ƞi naa-ti-wi ‘s/he left for her/his own homeland’ (3P-PX3SG leave-PST-3SG REFL-POSS land-DIR-RX). The demonstrative pronouns are ei : OBL ee- ‘this’ (proximal) : ACC ee-we : OBL eevs. tei ~ tii : ACC taa-wa : OBL taa-. The oblique case forms are mainly used in adverbial functions and can in the directive and instrumental be replaced by the corresponding pronominal spatials ewe- vs. tawa- : DIR ee-ti ~ ewe-si vs. taa-ti ~ tawa-si : INSTR ee-ji ~ ewe-ji vs. taa-ji ~ tawa-ji. The root taa- also yields the adjectival pronoun taa-nti ‘such’. The collective form COLL tei-getu ~ tii-getu ‘those people’ can be used as a replacement for the third person plural pronoun, as in [3]: [3] tii-getu baya nia that-coll rich man ‘They were rich people,

bi-ci-ti, be-pst-3pl

xoƞto-getu=dee taa-wa other-pl=ptcl that-acc and others also heard it.’

doogdi-xa-ti hear-pst-3pl

The basic interrogative pronouns are yaa ~ yaʊ ‘what?’ : ACC yaa-wa : OBL yaʊand nii ‘who?’ : OBL nii-. Some of the oblique case forms of yaʊ- can be replaced by forms based on the root i- ~ ii- : DAT i-du : LOC ii-le : PROL ii-li. Other forms based on the root yaa- are DAT yaa-dʊ ‘how?’, yaa-wʊ ‘why?’. This root also functions as a pro-verb, e.g. suu yaa-(y)i-sʊ ‘what are you doing?’ (2PL do.what-PRS2PL), tii nia yaa-xa-ni ‘what did he do?’ (that man do.what-PST-3SG), suke yaa-xa-ni ‘where did the axe go?’ (axe do.what-PST-3SG), yaa-na-​xa-si ‘what did you come

310 Shinjiro Kazama

for?’ (do.what-AND-PST-2SG). Other interrogative words include adi ‘how many?’, aʊki ‘which?’, anti ‘what kind of?’, aali ‘when?’, and awaasi ‘whither?’. The element aƞi(-), shared with Udihe, is used as a filler, allowing the speaker to gain time when s/he is choosing the next word, e.g. mapacaa aƞi-wa aƞi-xa-ni, sʊgjasa-wa loo-xo-ni ‘grandfather, well, what did he do, he hung the fish’ (grandfather what-ACC do.whatPST-3SG fish-ACC hang-PST-3SG). PERSON MARKING Oroch retains the Common Tungusic system of possessive (PX) and reflexive (RX) suffixes (Table 12.5), as used on nominals to express possession. The distinction with regard to verbal personal endings (VX) has, however, been almost totally lost, since non-participial finite predicates of the aorist type have been replaced by participial ones, and finitely used participial predicates are typically marked by the possessive suffixes. TABLE 12.5 OROCH PERSON MARKERS PX SG

PL

VX

V

/n

/g

1

-wi ~ -(y)i

-mi

b-bi

-mi ~ -wi

2

-si

n-si

k-si

-si

ƞ-ni

-ni ~ -N-

-mi

b-bi

3

-ni

RX

-wi ~ -(y)i

1

EXCL INCL

-mU

b-bU

-mU ~ -wU

-pi

m-pi

k-pi

-pi

-sU

n-su

k-su

-sU

3

-ti

n-ti

k-ti

-ti

RX

-(b)bA(y)i

2

Stem types: V = vowel stems, /n = nasal stems, /g stems ending in and unstable g. Segmental alternations: A = a e o, U = ʊ u, N = n m ƞ.

In practice, however, there is a minor difference in the use of several person markers as combined with nominals and finitely used nominalized verbs. If a participial suffix ends in a “hidden” /n, this segment is not “restored” before person markers, except that the first person singular takes the suffix -mi, e.g. saa- ‘to know’ : PTCP.PRF = PST saaxa/n : PST-1SG saa-xa-mi : PST-2SG saa-xa-si : PST-1PL.INCL saa-xa-pi : PST-3PL saa-xa-ti. By contrast, in the corresponding possessive forms of actual nominals, the nasal is present as a separate segment before person markers beginning with an obstruent, e.g. inaki/n ‘dog’ : PX1SG inaki-mi : PX2SG inakin-si : PX1PL.EXCL inaki-mʊ : PX1PL.INCL inakim-pi : PX3PL inakin-ti. This suggests that the ending -mi in the first person singular is synchronically not conditioned by the stem-final nasal of the participle marker. When used on finite predicates, the person markers agree with the person and number of the subject. There is, however, a tendency to use the third person singular marker also with plural subjects.

Oroch 311

In the reflexive set, the basic form unmarked for case is used in the object function (“unmarked accusative”), e.g. eni-ni nada xite-wi je-(w)ukuƞ-ki-ni ‘mother fed her seven children’ (mother-PX3SG seven child-RX eat-CAUS-PST-3SG). The following is an exceptional example of the combination of the accusative marker with the reflexive suffix: bʊa-wa-yi isi-gi-xa-ti ‘they reached their homeland’ (land-ACC-RX reach-REV-PST-3PL). The possessive suffixes can also be combined with the marker of alienable possession, which is -ƞi-. The alienability suffix is taken by nouns denoting, in the following descending order: occupation or social status > plants, wild animals, fish, birds, foods > body parts, heavenly bodies and natural phenomena. Spatials, terms for commodities, and nouns denoting integral parts of entities never take this suffix. Examples: sama-ƞi-bei e-ci-ti ice-gi-(y)e ‘they did not see their (own) shaman’ (shaman-AL-RX.PL NEG-PST-PL3 seeREV-​CONNEG), siƞe eeki ukte-ƞi-we-ni gaam jep-pi-ti ‘the mouse ate all the meat of the frog’ (mouse frog meat-AL-ACC-PX3SG all eat-PST-3PL), sʊlaki ila mʊda xuli-kte-xe-ni tii umuki moo-ƞi-ti-ni ‘the fox came three times to the tree of the flying squirrel’ (fox three time go.​around-DISTR-PST-3SG that flying-squirrel tree-AL-DIR-PX3SG). OTHER WORD CLASSES Compared with nominals and verb(al)s, the number of other word classes, falling under the general category of invariables, is insignificant. These include various types of adverbs, a few postpositions, as well as interjections. Many of the adverbs and postpositions are of a nominal or verbal origin, and some retain features of nominal declension. Degree adverbs include baadi ‘more’ and cuu ‘most’, e.g. inaki baadi ayakta o-ci-ti ‘the dogs got more angry’ (dog more angry become-PST-3PL), noƞo-ni cuu sagdi jaƞgia oci-​gi-xa-ni bi-ci ‘he had become the greatest chieftain’ (P3-PX3SG most big chieftain become-​REV-PST-3SG be-PST). Other intensifying words include xeme&xeme ‘very’ and anii ~ anaa ‘very’, e.g. tei gasa xeme&xeme sagdi gasa ‘that waterfowl is a very big one’ (that waterfowl very&very big waterfowl), buu ʊgda-mʊ anii ƞocko-ni ‘our boat is very small’ (1PL.EXCL boat-PX1PL.EXCL very small). The nominal maƞga ‘hard, strong’ ~ ‘hardness, strength’ can also be used in a similar function, e.g. maƞga kusuƞ-ki, maƞga egdeƞ-ki nia ‘(he is a) very powerful and very cunning man’ (hard power-PROPR hard trickPROPR man). The nominal kete ‘big’ yields the intensifying adverb kette ‘very’, which in negative clauses has the meaning ‘(not) so’, e.g. nʊƞa-ni kette gʊgda e-si-ni bi-ye ‘s/he is not so tall’ (3P-PX3SG so tall NEG-PRS-3SG be-CONNEG). Quantitative adverbs include gaam ~ gaamdi ‘all’ and cʊpali ‘all’ (a Lower Amur item also attested in Ulcha and Uilta), e.g. bii gaam xeete-we taƞi-jaƞa-i ‘I will count all the seals’ (1SG all seal-ACC count-FUT-1SG), sʊgjasa-wa cʊpali jep-pi-ti ‘they ate all the fish’ (fish-ACC all eat-PST-3PL). These can also be used adnominally, in which usage they should probably be understood as nominals, e.g. nʊƞanti dobbo gaamdi nia-wa waa-kta-​xa-ti ‘during the night they killed all the men’ (3PL night all man-ACC kill-​DISTR-PST-3PL), cʊpali xite eme-xe-ti ‘all the children came’ (all child come-PST3PL). Another nominal word is xacin ‘sort, kind’, which yields the reduplicated phrase xacin&xacin ‘all kinds of’, e.g. nʊƞanti xacin&xacin gebbeƞku-we baggi-xa-ti ‘they gathered all kinds of berries’ (3PL kind&kind berry-ACC gather-PST-3PL). A quantifier with only adverbial uses seems to be ñaƞga ‘a little’, e.g. xite-ke ñaƞga sagdi o-ci-ni ‘the child grew a little bigger’ (child-DIM a.little big become-PST-3SG). The limitative item

312 Shinjiro Kazama

bai ‘merely, simply, in vain, for free’ (← Chinese bai 白) is also used only adverbially, e.g. e-ci-si waa-ya, bai paati-xa-si ‘you did not kill him, you only hit him’ (NEG-PST2SG kill-CONNEG only hit-PST-2SG). There are also a few temporal adverbs, including elee ‘already, soon’, e.g. elee laki eme-i-ni ‘s/he is already coming near’ (already near come-PRS-3SG), and ñaa ‘(once) more, again’, e.g. ñaa tigde tigde-li-xe-ni ‘it started raining again’ (again rain rain-INCH-PST-3SG). An aspectual adverb always used in combination with a past tense form of the predicate verb is xaʊlia ‘at last, barely’, e.g. xaʊlia jeb-jeƞe-ni egdi o-ci-ni ‘at last his/her food has increased’ (at.last eat-PTCP.FUT-PX3SG many become-PST-3SG). In addition to the spatials, there are only a few other postpositionally used words. The “postposition” oron- ‘(in) place (of)’ is a regular noun and takes marking for both case and person, e.g. ami-mi oron-dʊ-ni too-wo ilaci-xa-ni ‘instead of his father he struck the fire’ (father-RX place-DAT-PX3SG fire-ACC strike-PST-3SG). The word baiti ‘towards’ looks like a directive form in -ti and takes person marking, e.g. gaam buu baiti-mʊ ice-gi-xe-ti ‘everybody looked towards us’ (all 1PL.EXCL towards-PX1PL. EXCL see-REV-PST-3PL). An item of a verbal origin is acapti ‘face to face with’ (from aca- ‘to meet’), which requires the accusative form of the preceding noun, e.g. ʊgda-wa xooƞko-wo-ni eye-we acapti oo-kʊƞ-ki-ni ‘s/he turned the boat so that its prow pointed against the stream of the river’ (boat-ACC prow-ACC-PX3SG stream-ACC against make-CAUS-PST-3SG). The following interjections have been registered in Oroch: gee ‘well, now, come on!’, maa ‘here!’ (take it!), anaya ~ annayia ‘ah, wow!’. VERBAL MORPHOLOGY Oroch shows a tendency to simplify verbal morphology as far as both the stem structure and the system of suffixes are concerned. Even so, traces of the original system with four major stem-types can still be discerned: (1) vowel stems, which originally took the aorist marker *-rA-, (2) consonant stems, which took the secondary aorist marker variant *-DA- < *-rA-, including nasal stems, after which the aorist marker was reduced to *-A-, (3) “change-of-state verbs”, which took the aorist marker *-dA-, and (4) a limited number of derived and underived stems which took the aorist marker *-sI-. Unlike the Nanaic languages, Oroch (like also Udihe) has not developed a morphological distinction between short-vowel stems and long-vowel stems. Since the aorist has been lost as a finite category, the differences between the stem types are most clearly observed in the connegative form, based on the aorist stem and used in combination with the negation verb. The vowel stems also differ from all other stem types in the formation of the perfective participle (finite past tense). (1) Stems ending in a vowel. This is the synchronic default type of verbs in Oroch. For this type, the aorist stem, as used in the function of connegative, is formed by the suffix -yA (< *-rA) and the perfective participle by the suffix -xA/n, e.g. ƞene- ‘to go’ : CONNEG ƞene-ye, saa- ‘to know’ : PTCP.PRF saa-xa/n. (2a) Stems ending in the non-nasal consonants p and g. The only commonly used basic verbal stem ending in p is jep- ‘to eat’, which forms the aorist stem by the suffix -tA (< *-rA) and the perfective participle by the suffix (*)-ki/n > (by metathesis and assimilation) -pi/n, i.e., CONNEG jep-te : PTCP.PRF jep-pi/n (< *jek-pin
-(U)kUn(A)- (for vowel stems) ~ -mUkAn(A)- (for nasal stems) ~ -buken(e)- (for the verb bu- ‘to die’). The function of the causative varies widely between ‘to make/force someone do’ and ‘to let/allow someone do’, e.g. yaʊ laʊji-ma-ni ee-le eme-ukəƞ-ki-ni ‘why did he send such a slave here?’ (what slave-ACC-PX3SG this-LOC come-CAUS-PST-3SG), gee,

314 Shinjiro Kazama

•

•

• •

•

mine-we aya-ji bagdi-kʊn-jaa-si ‘well, you will let me live well’ (INTERJ 1SG-ACC good-INSTR live-CAUS-FUT-2SG), eñee, ice-kuƞ-ke ‘mother, let me see (it)’ (mother.VOC see-CAUS-IMP), nʊƞa-ni mine-we je-ukene-​i-ni ‘s/he gave me some food’ (3P-PX3SG 1SG-ACC eat-CAUS-PRS-3SG), buu sune-we e-jee-mu ii-kune-ye ‘we shall not let you come in’ (1PL.EXCL 2PL-ACC NEG-FUT-1PL.​EXCL go.inCAUS-CONNEG). The causative can also be used to prevent switch-reference in complex sentences, e.g. laki-laa eme-uken-ee gappa-xa-ni ‘he shot it when it came near’ (near-LOC come-CAUS-CV.ANT shoot-​PST-3SG). The reciprocal voice (mutual action) is marked by the suffix -mAci-, e.g. tenee ñooki noƞonti xuwen-du baaci-maci-xa-ti ‘they met for the first time in the forest’ (just former 3PL forest-DAT meet-RECIPR-PST-3PL), baaci-xan-ji-bbai noƞonti meeni-ƞi xajʊ-bbai juweci-​meci-li-xe-ti ‘after they had met, they began to exchange their things with each other’(meet-PTCP.PRF-​INSTR-RX.PL 3PL REFL-POSS thingRX.PL change-RECIPR-INCH-​PST-3PL). The medial (reflexive) voice is marked by the suffix -p- ~ -ptA- ( Nanai -lO-), e.g. ñaa bude-li-xe-ti ‘they started dying again’ (again die-INCH-PST-3PL). The stative aspect (static action) is marked by the suffix -si-, e.g. bii dili-wi enu-sii-ni ‘I have a headache’ (1SG head-PX1SG ache-STAT-PRS-3SG), aya-ji tee-si-jee-si ‘you should be sitting well’ (good-INSTR sit-STAT-FUT-2SG), seekse eye-si-i-ni ‘blood is flowing’ (blood flow-STAT-​PRS-3SG). The durative aspect (prolonged action) is marked by the suffix -ci-, e.g. bii asanti-mʊ yaa-wa ice-ci-i-su ‘why are you looking at my wife?’ (1SG wife-DIR-1PL. EXCL what-ACC see-​DUR-​PRS-2PL), noƞonti ukte-we-ni miina-ci-i-ti ‘they are shredding his meat’ (3PL meat-​ACC-​3SG shred-DUR-PRS-3PL). In a few lexicalized items progressive action is expressed by the suffix -cA-, e.g. bii jawa-ca-i-wi ‘I am holding it’ (1SG catch-PROGR-​PRS-1SG), cf. also [4]:

[4] nʊƞa-ni juu mapa xite-we-ni 3p-px3sg two bear child-acc-px3sg ‘He brought the two bear cubs home

jʊʊk-ti house-dir

gai-xa-ni, bring-pst-3sg

jʊʊg-dʊ uyu-ce-xe-ni house-dat tie.up-progr-pst-3sg and kept them tied up at home.’ •

The distributive aspect (action repeated separately) is marked by the suffix -ktA-, e.g. xebu-​kte-xe-ni ‘s/he brought various things’ (bring-DISTR-PST-3SG), beyu-me egdi-we waa-kta-​xa-ni ‘he hunted many elks’ (elk-ACC many-ACC kill-DISTR-​PST-​3SG),

OROCH 315

•

•

goo asa-kta-xa-ni ‘s/he chased it far away’ (far chase-DISTR-PST-3SG). A lexicalized meaning is present in gele- ‘to want’ : DISTR gele-kte- ‘to seek, to look for’. The reversive aspect (repetitive-reversive action) is marked by the suffix -gi(< *-rgU- > Kilen -rgi-, Nanai -gO-), e.g. skola-ji cʊca-gi-xa-mi ‘I escaped from school’ (school-INSTR escape-REV-PST-1SG), jʊʊk-ti eme-gi-xe-mu ‘we came back home’ (house-DIR come-REV-PST-1PL.EXCL), tii ƞaacaki deili-gi-xe-ni ‘the heron flew back again’ (that heron fly-REV-PST-3SG). The verbally used stems expressing the coming of times or seasons usually take this suffix, e.g. dobbo-gi-xa-ni ‘night fell’ (to.become.night-REV-PST-3SG). The multidirectional aspect (repeated action in time or space) is marked by the suffix -wAci-. This suffix is often used with verbs of motion, e.g. xite-ni tada-ji ewi-i-ni, tutuweci-i-ni ‘the child plays with a bow and runs around’ (child-PX3SG bow-INSTR play-PRS-3SG run-MULTID-PRS-3SG), ooli kiaca tuu deili-weci-xe-ti ‘the raven and the eagle flew around in this way’ (raven eagle thus fly-MULTID-PST-3PL).

PARTICIPLES Oroch has three basic participial forms, which may be identified as the imperfective, perfective, and futuritive participles. These can all be used both as nominals (in adnominal position or as independent headwords) and as finite verbs. In the latter function they serve to mark three finite tenses: present, past, and future, respectively. The imperfective and perfective participles have also negative counterparts, formed by the suffixation of the negation verb -A- (< *+e-) and its corresponding participial forms to the connegative form of the base verb. Additionally, a distinction is made between personal (active) and impersonal (passive) participles, but this distinction is only valid for the imperfective participle. TABLE 12.6 OROCH PARTICIPLES PERS

IMPERS

V

n

C

-(y)i

-A.i

-DA.i

PRF

-xA/n

-ki/n

-Ki/n

FUT

-jAƞA

IMPRF

NEG-IMPRF

-A-si

-D.A-si

NEG-PRF

-A-cin

-D.A-cin

IMPRF

-(w)U.yi

-mU.yi

-bU.yi

NEG-IMPRF

-w.Asi

-mU.Asi

-bU.Asi

Stem types: V = vowel stems, n = nasal stems, C = non-nasal consonant stems. Segmental alternations: A = a e o, U = ʊ u, D = d t, K = k p.

The participle markers show some variation for the different stem types (Table 12.6). For vowel stems, the imperfective participle is formed by the suffix -(y)i (< *-rII) and the perfective participle by the suffix -xA/n (also present in Nanaic), e.g. eme- ‘to come’ : PTCP.IMPRF eme-i : PTCP.PRF eme-xe/n. For consonant stems the imperfective participle is always formed from a secondary vowel stem based on the aorist, while the perfective participle takes the suffix -ki/n ~ -pi/n, aag- ‘to dock’ : PTCP.IMPRF aagda-i : PTCP.PRF

316 Shinjiro Kazama

aak-ki/n, jep- ‘to eat’ : PTCP.IMPRF jepte-i : PTCP.PRF jep-pi/n, gun- ‘to say’ : PTCP.​ IMPRF gune-i : PTCP.PRF guƞ-ki/n. The verbs ga- ‘to take’, o- ‘to become’, bu- ‘to die’, nee- ‘to put’, and bi- ‘to be’ have the participial forms PTCP.IMPRF gada-i, odo-i, bude-i, neede-i, bi-i and PTCP.PRF ga-ci/n, o-ci/n, buk-ki/n, neek-ki/n ~ neekki-xe/n, bi-ci/n, respectively. The negation verb e- has the forms PTCP.IMPRF e-si and PTCP.PRF e-ci/n, which also yield the synthetic postverbal negative participle markers. The futuritive participle has the suffix -jAƞA for all verb types, e.g. PTCP.FUT emejeƞe, jeb-jeƞe, aakki-jaƞa (based on the perfective participle), gun-jeƞe, and also ga-jaƞa, o-joƞo, bu-jeƞe, nee-jeƞe, bi-jeƞe, e-jeƞe. The impersonal participles may be analysed as sequences composed of the primary passive marker -bU- ~ -mU- ~ -(w)U- and the participle markers -yi and -A-si. The allomorphy of the passive marker causes differences in the form of the complex suffix, yielding -(w)U-yi : NEG -w-Asi for vowel stems, -bU-yi : NEG -bU-Asi for non-nasal consonant stems, and -mU-yi : NEG -mU-Asi for nasal stems, e.g. PASS-PTCP.IMPRF eme-u-yi, jeb-bu-yi, gum-mu-yi : PASS-​NEG-PTCP.IMPRF eme-w-e-si, jeb-bu-e-si, gum-mu-e-si. The special monosyllabic stems have the forms PASS-PTCP.IMPRF ga-ʊ-yi, o-ʊ-yi, bub-bu-yi, ne-u-yi, bi-u-yi : PASS-​NEG-PTCP.IMPRF ga-w-a-si, o-wo-si, bub-bu-e-si, nee-w-e-si, bi-w-e-si. The negation verb e- does not appear to have an impersonal participle. When used as finite predicates the personal participles take person marking, though not obligatorily. Sample paradigm of the verb saa- ‘to know’: PTCP.IMPRF saa-(y)i : PRS 1SG saa-(y)i-wi : 2SG saa-(y)i-si : 3SG saa-(y)i-ni : 1PL. EXCL saa-(y)i-mʊ : 1PL.INCL saa-(y)i-pi : 2PL saa-(y)i-sʊ : 3PL saa-(y)i-ti, PTCP.PRF saa-xa/n : PST 1SG saa-xa-mi : 2SG saa-xa-si : 3SG saa-xa-ni : 1PL.EXCL saa-xa-mʊ : 1PL.INCL saa-xa-pi : 2PL saa-xa-sʊ : 3PL saa-xa-ti, PTCP.FUT saa-jaƞa : FUT 1SG saa-jaƞa-yi : 2SG saa-jaƞa-si : 3SG saa-jaƞa-​ni : 1PL. EXCL saa-jaƞa-mʊ ~ saa-jaƞa-wʊ : 1PL.INCL saa-jaƞa-pi : 2PL saa-​jaƞa-sʊ : 3PL saa-​jaƞa-ti. In addition to the finitely used futuritive participle marker -jAƞA-, the indicative future tense can be marked by the suffix -jAA-, e.g. xoŋto bʊa-ti ŋei-jee-pi ‘let us leave here and go to another place!’ (other land-DIR leave-FUT-1PL.INCL). The full paradigm, as exemplified by the verb eme- ‘to come’, goes as follows: FUT1SG eme-jee-mi ~ eme-jee-yi : 2SG eme-jee-si : 3SG eme-jee-ni : 1PL.EXCL emejee-mu : 1PL.INCL eme-jee-pi : 2PL eme-jee-su : 3PL eme-jee-ti. It should be noted that the suffixes -jAƞA- and -jAA- in finite use are not free variants: they take partly different person markers (in the first person singular), and they may also have involved a functional difference connected with information structure. This possibility should be verified by a study of the relevant corpus of examples. The negation verb e- takes 1SG -mi also in the present tense, e.g. PRS 1SG e-si-mi : 2SG e-si-si, etc., PST 1SG e-ci-mi : 2SG e-ci-si, etc. This pattern is repeated by the finitely used negative participles, e.g. mute- ‘to be able’ : PRS 1SG mute-e-si-mi : 2SG mute-esi-si, etc. For the future tense, the negation verb has the participial form PTCP.FUT e-jeƞe and the indicative future form e-jee-, e.g. FUT 1SG e-jee-yi : 2SG e-jee-si, etc., as well as another set of forms based on the exceptional stem FUT a-ta- : FUT 1SG a-ta-mi : 2SG

Oroch 317

a-ta-si : 3SG a-ta-ni : 1PL.EXCL a-ta-mʊ : 1PL.INCL. a-ta-pi : 2PL a-ta-sʊ : 3PL a-ta-ti. The latter form, which has a cognate in Udihe, is apparently more original. When a participle is used as an adnominal modifier, the head noun may be the actor, the object, the place of action, or some other circumstance connected with the action, e.g. jobo-i mapaaca mamaaca meene jʊʊg-dʊ-yi bagdi-xa-ti ‘there lived a poor old couple in their house’ (be.poor-PTCP.PRS grandfather grandmother REFL house-DAT-RX livePST-3PL), jali-dʊ oggo-xo ukte bi-i-ni ‘there is some dried meat in the shed’ (shed-DAT dry-PTCP.PRF meat be-​PRS-3SG), agdiʊ-jaƞa agdʊ-la maimaaki ‘there is no news that may be told’ (tell-PTCP.​FUT news-DES NEG.EXIST). When used as an independent head noun in the function of a sentential argument, a participle can take case marking and possessive suffixes, e.g. tee-gi-i-ni laki ‘(the place where) he is sitting (is) near’ (sit.down-REV-PTCP.IMPRF-PX3SG near), bii guƞ-ki-mi teƞ bi-i-ni ‘what I said is true’ (1SG say-PTCP.PRF-PX1SG true be-PRS-3SG), nʊƞa-ni juu asa eme-i-​we-​ni ice-xe-ni ‘he saw two women come’ (3P-PX3SG two woman come-PTCP. IMPRF-ACC-​PX3SG see-PST-3SG), mapa buk-ki-me-ni ice-xen-ji megge agda-xa-ni ‘the hero rejoiced when he saw that the bear had died (bear die-PTCP.PRF-ACC-PX3SG see-PTCP.PRF-​INSTR hero rejoice-PST-3SG), jeb-jeƞe-ti=dee manaa-p-pi-ni ‘they have run out of food’ (eat-PTCP.​FUT-PX3PL=PTCL finish-MED-PST-3SG). The impersonal participles, which do not take person marking, have a wide range of meanings and may express (1) proper action (‘what should be done’), e.g. ooni ƞene-uyi ‘how should I/we/one go’ (how go-IMPERS.PRS), (2) possibility (‘what can be done’), e.g. egdi-we oo-wi egdi-we waa-wʊyi ‘if we make many (snares), we will catch many (animals)’ (many-​ACC make-CV.CSEC kill-IMPERS.PRS), yaʊ-ji=daa waa-wasi amba ‘there is no way to kill the devil’ (how-INSTR=​PTCL kill-IMPERS.PRS.NEG devil), (3) passive action (‘what is done’), e.g. anda-wa xukkun-ji beici-uyi ‘musk deer are hunted by snares’ (musk.deer snare-INSTR hunt-IMPERS.PRS), naa baa-wasi ‘the earth is not found’ (earth find-IMPERS.​PRS.NEG), or (4) purpose, e.g. bunin-ti ƞene-uyi xokto ‘the road for going to hell’ (hell-DIR go-IMPERS.PTCP.IMPRF path). In addition to the main set of participles, Oroch has the synchronically isolated participlelike form in -cA, which corresponds to the perfective participle in |*-cA/A in several other Tungusic languages. In Oroch this form is marginal and has a low productivity. Always based on transitive verbs, it has a passive resultative meaning (glossed as PTCP.RES), although no overt passive marker is present, and it is mainly used nominally as a sentential argument or as an adnominal modifier, e.g. je(w)u iko-co=doo bi-i-ni, tala=daa bi-i-ni ‘there is some cooked food and also some sliced raw fish’ (food cook-​PTCP.RES=​PTCL sliced.raw.-​fish=PTCL be-PRS-​3SG), ila-ca moo xukeline-li-xe-ni ‘the lit firewood began to roll’ (light-PTCP.RES tree roll-INCH-PST-3SG). Unlike the other participles, this form cannot take arguments and cannot be used as a finite predicate. IMPERATIVES The imperative paradigm in Oroch is highly reduced and comprises only second-person forms for singular and plural. Unlike several other Tungusic languages, Oroch does not have a separate set of future (“distal”) imperative forms. The basic singular imperative is marked by the suffix -kA (< *-kAl), which has also the variants -gA ~ -wA ~ -pA depending on the stem type. The variant -kA is used on nasal stems, e.g. diggan- ‘to speak’ : IMP diggaƞ-ka, but also on vowel stems in variation with -gA, e.g. tee- ‘to sit down’ : IMP tee-ke ~ tee-ge. The variant -pA is taken by consonant

318 Shinjiro Kazama

stems in p, e.g. jep- ‘to eat’ : jep-pe (< *jek-pe < *jep-ke). In most examples, vowel stems take the variant -wA, e.g. bii xute-mi odo-wo ‘become my child!’ (1SG child-PX1SG become-IMP), ñaƞga=daa ʊmi-wa ‘drink even a little!’ (a.little=PTCL drink-​IMP). The corresponding plural form is -kA-sU ~ -gA-sU ~ -wA-sU ~ -pA-sU, e.g. esi=gde waaxa-mʊ, tala-wa-sʊ ‘now we caught it, slice it!’ (now=PTCL kill-PST-1PL.EXCL slice-​ IMP-2PL). The imperative sentence can also contain the second-person pronoun as an overt subject, e.g. sii ñuu-we baa-si, dookci-na-ga ‘go out to listen!’ (2SG go.out-IMP outside-DIR listen-AND-IMP), suu=dee teeyeci oo-wo-sʊ ‘you also make it that way! (2PL=PTCL that.way make-IMP-2PL). Prohibition is expressed by the prohibitive form e-ji of the negation verb e-, combined with the connegative form of the following main verb, e.g. mine-we eji waa-ya ‘don’t kill me!’ (1SG-ACC PROHIB kill-CONNEG), sagdi uli-ti eji ƞene-ye ‘do not go to the big river! (big river-DIR PROHIB go-CONNEG). In the second-person plural, the person marker can be added to the connegative form, e.g. suu eji cʊca-a-sʊ ‘do not run away!’ (2PL PROHIB run.away-CONNEG-2PL). OTHER MODAL FORMS In addition to the imperative, there are also a few other modal contents that are expressed inflectionally, derivationally, or analytically. The relevant categories include the subjunctive, intentive, and desiderative. •

The subjunctive mood is marked by the suffix -mcA-, complemented by person markers of the possessive type, e.g. eme- ‘to come’ : SUBJ 1SG eme-​mce-yi : 2SG eme-mce-si : 3SG eme-​mce-​ni : 1PL.EXCL eme-mce-mu : 1PL.INCL ememce-pi : 2PL eme-mce-su : 3PL eme-​mce-ti. It may be noted that the subjunctive marker is not a deverbal derivational suffix, since it has no other paradigmatic forms and it does not take participial markers in finite use. In fact, it could synchronically also be viewed as a true finite form, though diachronically it actually contains a participle marker (< *-mO-cAA-). The subjunctive expresses a counterfactual situation, as used in the main clause of a complex sentence, e.g. tigde e-siki-ni bi-ye bii eme-mce-yi ‘if it did not rain, I would come’ (rain NEG-CV. AOR-PX3SG be-CONNEG 1SG come-SUBJ-1SG). The same content can also be expressed analytically by combining the person-marked futuritive participle of the main verb with the past tense form (perfective participle) of the auxiliary bi- ‘to be’, e.g. tigde e-siki-ni bi-ye bii eme-jeƞe-yi bi-ci ‘if it did not rain, I would come’ (rain NEG-CV.AOR-PX3SG be-CONNEG 1SG come-PCTP.​FUT-​PX1SG be-PST), cf. also [5]:

[5] adʊli-kaaƞ-ki bi-ci net-dim-propr be-ptcp.prf ‘If I had had a small net,

bi-mi be-cv.conn

ei noso-wo waa-jaƞa-yi this sable-acc kill-ptcp.fut-px1sg I would have caught this sable.’

bi-ci=gini be-pst=ptcl

Oroch 319

•

•

Intentives are marked by the derivational suffix -icA- (< *-kIcA-), e.g. ili-ica-xa-ni ‘he intended to stand up’ (stand-INTENT-PST-3SG), bii miaʊca-la-ica-xa-mi ‘I intended to shoot (it)’ (1SG gun-VBLZ-INTENT-PST-1SG). As an exception, the verb jep- ‘to eat’ has the form INTENT jep-pice- (< *jek-pice- < *jep-kice-). Desideratives are formed by the derivational suffix -mUsi-, e.g. bii je-musi-i-wi ‘I am hungry’ (1SG eat-DESID-PRS-1SG), bii doo-wi ƞeele-musi-i-ni ‘in my mind I am afraid’ (1SG inside-PX1SG be.afraid.INTENT-PRS-3SG).

CONVERBS Oroch has eight converbal forms, here termed the connective, consecutive, anterior, successive, contextual, contemporal, purposive, and aorist converbs. With regard to their coreferentiality with the subject of the main clause, the converbs may be divided into the conjunct (same-subject), disjunct (different-subject), and ambivalent types (Table 12.7). In disjunct use, the subject person is typically marked by the possessive suffixes, while some converbs can also take the reflexive suffixes to indicate conjunct use. Two of the conjunct converbs have special plural forms. TABLE 12.7 OROCH CONVERB MARKERS  

 

PL

CONN

-mi ~ -m=

-m-Ayi

PX

RX

SS +

CSEC

-wi

-wi-sA

+

ANT

-(R)AA

+

SUCC

-mdiA

+

DS

+

CNTX

-(k)icAA-

+

CTEMP

-ƞAsA-

+

+

+

+ +

PURP

-(A)lA(k)A-

+

+

+

+

AOR

-(RA)ki-

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e, R = aorist stem in d t y.

• The connective converb for parallel action is marked by the suffix -mi : PL -mayi (< *-mArI), which contains historically the reflexive suffix. The singular suffix can be extended to -miA, e.g. ƞene-mie, ƞene-mie, nia egdi-we-ni baa-xa-ni ‘she went and went and found many men’ (go-CV.CONN.EXT go-CV.CONN.EXT man manyACC-PX3SG find-PST-3SG). • The consecutive converb in -wi (< *-pi) ~ EXT -wie has cognates in Nanaic and Jurchenic, but not in Udihe. The plural form -wi-sA resembles that of Uilta (-pi-ssAA), but not that of Nanai and Ulcha (-pAArI), which contains the plural reflexive marker (-ArI). This converb indicates temporal or conditional subordination, e.g. bii juu-kte ila-kta bi-wi ƞei-jeƞe-yi ‘I will leave in two or three days’ (1SG two-day three-day be-CV.CSEC leave-FUT-1SG), noso xokto-wo-ni baa-wisa xokto-li-ni ʊja-wʊyi ‘if you find the track of a sable, you should follow the track’ (sable path-ACC-PX3SG find-CV.CSEC.PL path-PROL-​PX3SG follow-IMPERS.PRS).

320 Shinjiro Kazama

•

•

•

•

•

•

The anterior converb for an action followed by another action is marked by the aorist-​ based suffix -yAA (< *-rAA), with the variants -AA ~ -dAA ~ -tAA depending on the stem type, e.g. tada-yi jawa-yaa gappa-xa-ni ‘he took his arrow and shot’ (arrow-RX take-CV.ANT shoot-PST-3SG), timaaki tee-gi-yee jep-tee (~ jepte-yee) guuline-xe-ni ‘in the morning he got up, ate, and departed’ (morning get.​up-​CV.ANT eat-CV.ANT depart-PST-3SG), otoƞgo-yi baika taan-aa (~ taana-yaa) tukti-​xe-ni jʊʊk-ti baiti-ni ‘he pulled his dugout a little and ascended towards his house’ (dugout-RX a.little pull-CV. ANT ascend-​PST-​3SG house-DIR towards-PX3SG). The successive converb in -mdiA is functionally similar to the anterior converb, but it normally implies a more prolonged action followed by another action, e.g. bagdi-mdia xuwe-me=dee xuli-kte-li-xe-ti ‘they lived and lived, and then they started going (hunting to the forest)’ (live-CV.SUCC forest-ACC=PTCL go.around-DISTRINCH-PST-3PL). This converb is often combined with the auxiliary bi- ‘to be’, e.g. tuu ewi-mdie bi-mi dobbo-gi-xa-ni ‘while s/he was playing like that, the night fell’ (thus play-CV.SUCC be-CV.CONN become.night-​REV-PST-​3SG). The contextual converb in -(k)icAA- (= Nanai -OcIA-) indicates temporal or conditional subordination in disjunct constructions, e.g. jʊa-gi-icaa-ni sʊgjasa-wa beici-jee-ti ‘when summer comes, they will fish’ (become.summer-REV-CV.CNTXPX3SG fish-ACC chase-FUT-3PL), ami-si beyu-me waa-(y)icaa-ni imukse-we-ni jeb-jeƞe-si ‘when your father kills an elk, you will eat its fat’ (father-PX2SG elkACC kill-CV.CTXT-PX3SG fat-ACC-PX3SG eat-FUT-2SG). The suffix variant -kicAA is attested in bu- ‘to die’ : CV.CNTX bu-kicee-. The contemporal converb in -ƞAsA- (= Udihe -ƞAhA/n-) expresses simultaneous action in the past. In conjunct use it takes reflexive marking, e.g. bii ee-du bagdi-xa-mi ƞiici bi-ƞese-yi ‘I used to live here when I was small’ (1SG this-DAT live-PST-1SG small be-CV.​CTEMP-RX), buu ƞiici bi-ƞese-bbei yaʊ=daa xaci ewin-ji ewi-xe-mu ‘when we were small, we used to play with every kind of toy’ (1PL.EXCL small be-CV. CTEMP-RX.PL what=PTCL various toy-INSTR play-PST-1PL.EXCL). The purposive converb in -lAA- ~ -AlAA-, in the Khadi dialect -lAkA- (= Udihe -lAgA-) indicates the purpose of action, e.g. aya-ji xuli-lee-wei kesi gele-i-ti ‘they prayed for luck in order to have success in hunting’ (good-INSTR go.around-CV.PURP-RX.PL luck want-​PRS-3PL), sʊgjasa-wa jekke beici-jee-si=me buu je(w)u-leke-mu ‘you will always catch enough fish for us to eat’ (fish-ACC always catch-FUT-2SG=PTCL 1PL.EXCL eat-CV.PURP-​PX1PL.EXCL). The origin of this converb is unclear, but it is possible that the element -lA- is identical with the designative case marker -lA-. The aorist converb is composed of the aorist marker, which varies depending on the stem type, and the element -ki- (< *-kI-), complemented by person marking of the possessive type. After vowel stems, the overt aorist marker can also be absent. This converb is always used in a disjunct function and is not combined with reflexive suffixes. It has a wide meaning referring variously to simultaneous (‘when’), anterior (‘after’), or conditional (‘if’) action, e.g. teelumu-we gele-eki-si bii teelumuci-jeƞe-yi ‘if you want a story, I will tell (it)’ (story-ACC want-CV.AOR-PX2SG 1SG story-VBLZ-FUT-1SG), sii jixa-wa buu-ki-si bii kino-ti ƞene-​jeƞe-yi bi-ci ‘if you gave me money, I would go to the cinema’ (2SG money-ACC give-​CV.AORPX2SG 1SG cinema-DIR go-FUT-PX1SG be-PST), sii bu-deki-si=dee saa-jaƞa-yi ‘if you die, I will know it’ (2SG die-CV.AOR-PX2SG=PTCL know-FUT-1SG), timai tee-gi-eki-ti, ice-eki-ti, neƞun-ti asa-ni maaki ‘when they got up and looked in the

Oroch 321

morning, the younger brother’s wife was not there’ (morning get.up-REV-CV.AORPX3PL see-CV.AOR-​PX3PL younger.brother-PX3PL wife-PX3SG NEG.EXIST). The aorist converb of the negation verb is e-siki-, while the auxiliary bi- ‘to be’ yields bi-eki-, e.g. suu asa-ma e-siki-s[u] buu-ye sune-we waa-jaƞa-yi ‘if you do not give me the wife, I will kill you all’ (2PL wife-ACC NEG-CV.AOR-​2PL give-CONNEG 2PL-ACC kill-FUT-1SG), tuu ƞene-mdie bi-eki-ni nia xokto-ni baadia egdi o-d[o]i-ni ‘as she went on, there were more tracks of people’ (thus go-CV.SUCC be-​CV.AOR-PX3SG man path-3SG more many become-PRS-3SG). SYNTAX Oroch is a verb-final (SOV) and head-final language, though occasional deviations from these patterns are observed. There is no agreement between modifier and modified, but a numeral or other quantifier, when placed after a noun, can take accusative marking, e.g. tuu oo-wi sʊgjasa-wa egdi-we waa-jaƞa-sʊ ‘doing so, you will catch a lot of fish’ (thus do-CV.CONN fish-ACC many-ACC kill-FUT-2SG). In such constructions, the noun preceding the quantifier can also remain unmarked, e.g. taa-dʊ nia xokto-ni egdiwe-ni baa-xa-ni ‘s/he found many tracks of people there’ (that-DAT man path-PX3SG many-ACC-PX3SG find-PST-3SG). Occasionally, an unmarked noun can follow a marked quantifier, e.g. nʊƞa-ni om taƞgʊ-wa xʊkka tule-xe-ni ‘he set one hundred snares’ (3P-PX3SG one hundred-ACC snare set-​PST-​3SG). Apparently, in such clauses we are not dealing with a modifier-modified relationship, but with the juxtaposition of two separate objects, of which one is a quantifier. Other syntactic features characteristic of Oroch include the following: •

•

Topic is often, though not obligatorily, marked by the contrastive clitic =gdA(AnA), e.g. cuu neƞu-ti=gde meneji-xe-ni ‘the youngest of them remained’ (most younger. sibling-​PX3PL​=TOP remain-PST-3SG), sʊlaki=gda taa-dʊ=gdaana meici-i-ni ‘the fox then thought (about it)’ (fox=TOP that-DAT=TOP think-PRS-3SG), tei amaala-ni=gdaana ñaa sʊlaki eme-​gi-​xe-ni ‘after that, the fox came again’ (that backside-LOC-PX3SG=TOP again fox come-​REV-PST-3SG). Two other clitics with a similar function are =tAni and =lA, e.g. taa-dʊ=tani tei mapaca eu-xe-ni ‘then that grandfather went down towards the river’ (that-DAT=TOP that grandfather go.towards.river-PST-3SG), esi=le eji ƞeele-ye ‘now do not be afraid!’ (now=TOP PROHIB be.afraid-CONNEG). No copula is required in equative sentences with a present-tense reference and with a third-person subject. In other tenses and persons, the copula bi- ‘to be’ is used, e.g. noƞonti xite-ti ilan xuse-me xite bi-ci ‘their children were three boys’ (3PL childPX3PL three male-DX child be-PST), bii sii xite-si bi-i-wi ‘I am your child’ (1SG 2SG child-PX2SG be-PRS-1SG), ii-li=dee ƞene-m=dee mute-wesi bi-ci-ni ‘it was impossible to go anywhere’ (what-PROL=​PTCL go-CV.CONN=PTCL be.able-PTCP. IMPERS-NEG be-PST-3SG). The futuritive participle bijee = PTCP.FUT bi-jee and its irregular variants binjee ~ bienjee are lexicalized as modal particles in the meaning ‘probably’, e.g. aak-si jʊʊg-dʊ bi-i-ni bijee ‘your elder brother is probably at home’ (elder-brother-PX2SG house-DAT be-PRS-3SG PROB), meneji-xe jaak-ti=la gaamdi buk-ki-ti bijee ‘their remaining friends probably all died’ (remain-PTCP.​ PRF friend-PX3PL=​TOP all die-PST-3PL PROB), xoƞto mapa bijee meici-xe-ni ‘he

322 Shinjiro Kazama

•

•

•

•

thought that it had probably been a different old man’ (other grandfather PROB think-PST-3SG), biti baaci-jaa-pi=daa bienjee ‘we shall probably see again’ (1PL. INCL meet-FUT-1P.INCL=​PTCL PROB). Another modal particle based on the copula bi- is biiskee ~ bieskee < bi-e-si=kee ‘isn’t it?’ (bi-NEG-​PRS=​PTCL), which expresses the speaker’s confirmation or cross-​questioning, e.g. bii sin-ti guƞ-ki-mi biiskee ‘I said (it) to you, didn’t I?’ (1SG 2SG-DIR say-PST-1SG CONF), min-du miaʊca bi-i-ni bieskee ‘I have a gun, can’t you see it?’ (1SG-​DAT gun be-PRS-3SG CONF). Other auxiliaries include mute- ‘to be able, to be possible’ and kaadawasi ‘it is unclear, it is not known’ = IMPERS.PRS kaada-wasi (from an otherwise nonproductive verbal base), both of which are combined with the connective converb of the semantic main verb, e.g. mute-i-si=nuu miaʊca-la-mi ‘can you shoot with the gun?’ (be.able-PRS-2SG=​INTERR gun-VBLZ-CV.CONN), eme-gi-mi=dee kaadawasi ‘it is not clear whether s/he will come or not’ (come-REV-CV.CONN=PTCL unclear). Negation is expressed at several layers of the language. The negation of a verbal predicate normally takes place analytically by combining the conjugated negation verb e- with the connegative form of a following main verb, e.g. e-si-mi saa-ya ‘I do not know’ (NEG-​PRS-1SG know-CONNEG), uke-we e-ci-ti kata-ya ‘they did not tie the door’ (door-ACC NEG-PST-3PL tie-CONNEG), ee-du e-jee-pi bi-ye ‘we will not be here’ (this-DAT NEG-FUT-​1PL.INCL be-CONNEG). For the negation of existence or possession the negative existential noun maaki ~ maimaaki is used, e.g. nuƞen-ti asa-ni maaki ‘their younger brother has no wife’ (younger.brother-​PX3PL woman-PX3SG NEG.EXIST). Privation can also be expressed by the privative noun anaa ‘not having’ or the privative suffix -lAci ‘without’. Interrogation in polar questions is expressed by the clitic =nuu, e.g. sii=dee eme-i-​ si=nuu ‘are you also coming?’ (2SG=PTCL come-PRS-2SG-INTERR), sii baa-ti-si ƞene-mi goo=nʊʊ ‘is it far to go to you?’ (2SG place-DIR-PX2SG go-CV.CONN far=INTERR), toomdia megge ice-xe-ni, tʊksʊ=nʊʊ, tamna=nʊʊ ‘then the hero saw something like a cloud or fog around him’ (then hero see-PST-3SG cloud=INTERR fog=INTERR). When attached to a question word, this clitic implies indefiniteness, e.g. mapa awaa-si=nʊʊ awaa-si=nʊʊ xuli-kte-i-ni ‘the bear is wandering around’ (grandfather where-DIR=INTERR where-​DIR=INTERR go.around-DISTR-PRS-3SG). Normally, in content questions containing a question word, the corrogative clitic =kA ~ =(g)A is used, e.g. ii-le ƞene-i-si=ge ‘where are you going?’ (what-LOC go-PRS2SG=CORR), yaʊ agdʊ-ni bi-jee-n[i]=ke ‘what kind of news are there?’ (what newsPX3SG be-FUT-3SG=CORR). Emphasis of various types is expressed by the clitic =dAA ‘also, even’, e.g. mindu=daa buu-we ‘give me, too!’ (1SG-DAT=PTCL give-IMP). When attached to question words, this clitic gets a connegative meaning, e.g. yaʊ=daa xokto-ni=doo maaki ‘there was no road of any kind’ (what=PTCL road-PX3SG=PTCL NEG.EXIST), awaa-si=daa eji ƞene-ye ‘do not go anywhere!’ (where-DIR=PTCL PROHIB goCONNEG). The word yaʊ=daa ‘what indeed?’ can also function as an emphatic particle, e.g. buu eje-mu yaʊ=daa aliasi maƞga megge ‘our master is an incredibly strong man’ (1PL.EXCL master-PX1PL.EXCL what=PTCL incredible strong hero). In clause-final position =dAA has a concessive function, e.g. ʊnta-ya-yi oo-mi tumnee-si=dee min-du aya ‘even if she cannot make shoes, it is fine with me’ (shoe-DES-RX make-CV.CONN know.how-NEG-PRS=PTCP 1SG-DAT good).

Oroch 323

•

Other enclitic particles include =bAki ‘still’, e.g. megge tee-gi-mi ui-le=beki jawaxa-ni ‘the hero got up and caught (it when it was still) above (in the air)’ (hero get. up-​REV-​CV.CONN upper.side-LOC=PTCL catch-PST-3SG); =mAAki ‘at last, completely’ (= Udihe -mei), e.g. taa-dʊ=maaki jawa-ci-na-xa-ni ‘at last, she went to catch (it)’ (that-DAT=PTCL catch-DUR-AND-PST-3SG), gaaki jʊʊƞ-ni gaaki sañani=maaki bi-ci-ni ‘in the house of the crows there was nothing but crow droppings’ (crow house-PX3SG crow dropping-​PX3SG=​PTCL be-PST-3SG); =kiA [exclamation on adjectival nominals], e.g. annayia, xatala-ma, aya=kia ‘oh, girl, how beautiful!’ (INTERJ girl-ACC good=PTCL); =gini [emphasis of predicate] (= Udihe =gini, Nanai =gOAni), e.g. tuu bi-eki-ni=le aya=gini ‘if it is so, it is fine’ (thus be-CV.AOR-​ PX3SG=TOP good=PTCL), buu xala-mʊ egdi=gini ‘our relatives are many’ (1PL. EXCL relative-PX1PL.EXCL many=PTCL); =mA [emphasis of predicate], e.g. ceje, aya xajʊ bi-i-ni=me ‘true, it is a good thing’ (true good thing be-PRS-3SG=PTCL), sii meneji-ke, bii ƞen(e)-jee-m=me ‘you stay here and I will go’ (2SG remain-IMP1SG go-FUT-CV.CONN=PTCL), emne edi-le-xe-si bi-jee-si=me ‘once you have married, you will live (as you are)’ (once husband-VBLZ-PST-2SG be-FUT-2SG=PTCL).

LEXICON AND LANGUAGE CONTACTS The boundary of Oroch towards Udihe is somewhat fuzzy because of the Udihe features present in the Khadi dialect. On the other hand, the interaction of the Tumnin and Khungari dialects with Nanai and Ulcha has resulted in the borrowing of lexical items from Nanaic, especially terms related to the local environment, such as the names of the different kinds of fish in the Amur. Phonological criteria allow words with an initial or medial p or medial r to be identified as Nanaic loanwords, e.g. paatila- ‘to hit’ ← Ulcha paatɪla- (= Nanai paacɪla-), aapʊ/n ‘hat’ ← Ulcha aapʊ/n (= Nanai aapo/n), araki ‘liquor’ ← Nanai-​ Ulcha arakɪ (cf. Udihe, a’i, Neghidal ayakii), mʊri/n ‘horse’ ← Ulcha mʊrɪ/n (= Nanai morɪ/n, cf. Udihe mui). An initial x for Proto-Tungusic *x also signals Nanaic origin. e.g. xosikta ‘star’ ← Nanai xosɪkta (= Ulcha xosɪta, cf. Udihe wahikta). Some items are present in doublets, with one variant representing original heritage, while the other is a Nanaic borrowing, e.g. iggi (= Udihe igi < *irgi < *xürgü) ~ xuggu ‘tail’ ← Nanai xurgu (= Ulcha xuju < *xurgu < *xürgü). Items shared by Oroch with Ulcha and possibly borrowed from the latter include bai&bai ‘difficult, dangerous’ (= Ulcha bai&bai), gaƞgayʊ ‘little finger’ (= Ulcha gaƞgakʊ). There are a few cases in which Oroch uses a word different from the other Tungusic languages. For instance, the negative existential noun in Oroch is maaki ~ maimaaki (cf. Udihe anci, Ewenki aacin, Nanai aba, Ulcha kewe/n, Manchu akū = aqu). Also, the colour term ceekke/n ‘white’ is idiosyncratic (cf. Udihe caligi, Ewenki bagdarin), though Oroch also has caagja/n ‘white’, perhaps borrowed from Nanai-Ulcha caagja/n. A change of meaning is observed in e.g. sagdi ‘big’ (also in Udihe) < ‘old’, diggan- ‘to speak, to say’ (also in Udihe) < ‘to make a sound’, uli ‘water, river’ (also in Udihe) < ‘river’. Some Oroch words have a cognate only in Uilta, e.g. gaam ~ gaamdi ‘all’ (= Uilta geem, cf. Nanai xem), laki ‘near’ (= Uilta laxa, cf. Nanai laƞ), which may suggest a contact between the two languages. In this context, the fact that Oroch and Uilta share the word *ulaa-r ‘domestic reindeer’ is also of relevance. Oroch has also a large number of Manchu and earlier Jurchenic loanwords, most of which were, however, transmitted by Nanai-Ulcha, or also by Udihe. The same is true of the Chinese elements in Oroch. Naturally, there are also many loanwords from Russian,

324 Shinjiro Kazama

some of which were borrowed before the Soviet period, e.g. xiliaba ‘bread’ ← Russian xleb, kami/n ‘iron stove’ ← Russian kamín. The Russian items occasionally show varying stages of adaptation, e.g. skola ~ iskolo ‘school’ ← Russian shkóla. Like the other Tungusic languages of the Amur region, Oroch has been actively used in oral literature. The principal genres of folklore are nimaapʊ ‘folktales’, teeluƞu ~ teelumu ‘legends’, and soxori ‘foreign stories’. REFERENCES AND FURTHER READING Abramova, G. S. [Г. С. Абрамова] (1994) Русско-орочский разговорник [Russian-Oroch phrasebook], Хабаровск [Khabarovsk]: Комитет образования администрации Хабаровского края. Abramova, G. S. [Г. С. Абрамова] (2002) Картинный словарь орочского языка [Picture dictionary of the Oroch language], Санкт-Петербург [St. Petersburg]: “Дрофа”. Alonso de la Fuente, José Andrés (2014) ‘Oroch uktä ‘meat’, the Tungusic “heteroclisis” and analogical morphophonology’, Cahiers de Linguistique Asie Orientale 43 (2): 69–90, Leiden. Alonso de la Fuente, José Andrés (2017) ‘An Oroch word-list lost and rediscovered: A critical edition of Tronson’s 1859 pseudo-Nivkh vocabulary’, Bulletin of the School of Oriental and African Studies 80 (1): 97–117. Arsen’ev, V. K. [В. К. Арсеньев] (2008) Русско-орочский словарь: Материалы по языку и традиционной культуре орочей и удэгейцев [Russian-Oroch dictionary: Materials on the language and traditional culture of the Oroch and Udihe], edited by A. Kh. Girfanova [А. Kh. Гирфанова] & N. L. Sukhachev [Н. Л. Сухачев], Санкт-​ Петербург [St. Petersburg]: Санкт-Петербургский государственный университет. Avrorin, V. A. (1979) [В. А. Аврорин] ‘Родственные связи орочского языка с другими тунгусо-маньчжурскими языками’, in: История и диалектология языков Сибири [History and dialectology of Siberian languages], 3–22, Новосибирск [Novosibirsk]: ИИФФ CO Н СССР. Avrorin, V. A. [В. А. Аврорин] & E. P. Lebedeva [Е. П. Лебедева] (1966) Орочские сказки и мифы [Oroch tales and myths], Новосибирск [Novosibirsk]: “Наука”. Avrorin, V. A. [В. А. Аврорин] & E. P. Lebedeva [Е. П. Лебедева] (1968) ‘Орочский язык’ [The Oroch language], in: Монгольские, тунгусо-маньчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 191–209, Ленинград [Leningrad]: “Наука”. Avrorin, V. A. [В. А. Аврорин] & E. P. Lebedeva [Е. П. Лебедева] (1978) Орочские тексты и словарь [Oroch texts and dictionary], Ленинград [Leningrad]: “Наука”. Avrorin, V. A. [В. А. Аврорин] & B. V. Boldyrev [Б. В. [Grammar of the Oroch language], Новосибирск [Novosibirsk]: ИФ СО РАН. Baranova, V. V. [В. В. Баранова] & K. A. Maslinskii [К. А. Маслинский] (2014) ‘Есть ли жизнь после смерти? По следам экспедиции к орочам 2001’ [Is there life after death: In the wake of the expedition to the Oroch people in 2001], in: В. Ф. Выдрин [V. F. Vydrin] & Н. В. Кузнецова [N. V. Kuznecova] (eds.), От Бикина до Бамбалюмы, из варяг в греки: Экспедиционные этюды в честь Елены Всеволодовны Перeхвальской [Field-inspired essays in honour of Elena V. Perekhvalskaya], 5–11, Санкт-Петербург [St. Petersburg]: “Нестор-История”.

Oroch 325

Bereznickii, S. V. [С. В. Березницкий] (1999) Мифология и верования орочей [Mythology and beliefs of the Oroch], Фольклор народов Маньчжурии [Folklore of the peoples of Manchuria] 2, Санкт-Петербург [St. Petersburg]: Центр “Петербургское востоковедение”. Cincius, V. I. [В. И. Цинциус] (1949) ‘Oчерк морфологии орочского языка’ [A sketch of Oroch morphology], Ученые записки ЛГУ 98, Серия востоковедческих наук 1: 119–160, Ленинград [Leningrad]. Girfanova, A. Kh. [А. X. Гирфанова] (2007) Словарь орочско-русский и русскоорочский [Oroch-Russian and Russian-Oroch dictionary], Санкт-Петербург [St. Petersburg]: “Дрофа”. Kazama, Shinjirō 風間伸次郎 (1996)『オロチ語基礎資料』[Basic linguistic materials on Oroch], 鳥取 [Tottori]: 鳥取大学教育学部 =『ツングース言語文化論集』 [Publications on Tungus Languages and Cultures] 7, http://сое.аа.tufs.ac.jp/​tungus/​ home.​html. Kazama, Shinjirō 風間伸次郎 (2009)「オロチ語とウデヘ語の異同について」[On the differences and similarities between Oroch and Udihe],『語学研究所論集』14: 1–12, 東京 [Tokyo]: 東京外国語大学. Ko, Seongyeon (2012) ‘Oroch vowel harmony revisited’,『언어학』Linguistic Association of Korea Journal 63: 31–​53, 서울 [Seoul]. Lar’kin, V. G. [В. Г. Ларькин] (1964) Орочи (историко-этнографический очерк с середины XIX в. до наших дней) [The Oroch: a historical-ethnographical sketch from the 19th century to our days], Москва [Moscow]: “Наука”. Lebedeva, E. P. [Е. П. Лебедева] (1997) ‘Орочский язык’ [The Oroch language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 215–226, Москва [Moscow]: “Индрик”. Leontovich, Sergei [Сeргѣй Леонтовичъ] (1896) Краткiй русско-ороченскiй словарь съ грамматической замѣткой [Concise Russian-Orochen dictionary with a grammatical note], Записки Общества изученiя Амурскаго края 5 (2), Владивостокъ [Vladivostok]: Типографiя Н. В. Ремезова. Protodiakonov, P[rokopii] [П. Протодьяконов] (1888) Краткiй русско-ороченскiй словарь [Concise Russian-Orochen dictionary], Казань [Kazan]: Изданiе Православнаго миссiонерскаго общества. Schmidt, P[eter] (1928) ‘The language of the Oroches’, Acta Universitatis Latviensis 17: 17–62, Riga. Tronson, John M. (1859) ‘Some observations on the Ghilack or Ghailack race: their manners, customs, and the regions they inhabit’, Transactions of the China Branch of the Royal Asiatic Society 6: 107–118. Turaev, V. A. [В. А. Тураев] (2001) (ed.) & al. История и культура орочей: Историко-​ этнографические очерки [History and culture of the Oroch: Historical and ethnographical sketches], Санкт-Петербург [St. Petersburg]: “Наука”.

CHAPTER 13

UDIHE Elena Perekhvalskaya and Juha Janhunen

Udihe (udihe) is a Tungusic language of the Orochic group spoken in the southern part of the Sikhote Alin Mountains, between the Ussuri and Middle Amur basins to the west and the coast of the Sea of Japan to the east. The name Udihe (also Udie, Ude, Udehe, Udege, Udeghe, Udykhe) was introduced into linguistic literature in the early 20th century in Russia (Russian MASC udègéec : PL udègéicy : FEM udègéika) to replace the older term Oroch ~ Orochën (pejoritively in Russian also rachón : PL rachóny), which had been used generically for both Oroch and Udihe. Another name used historically in reference to the Udihe speakers is Taz or Tazy (Russian taz : PL tázy, from Chinese 韃子 dazi), which today officially refers to a minor (sub)ethnic group with mixed Chinese and Udihe ancestry and with a Mandarin-based language. The ethnonym Udihe in various shapes is known from Korean and Chinese sources referring to the mediaeval and later populations living in the eastern border zone between Korea and Manchuria, and although it may have had a more diffuse reference it is likely to have comprised the linguistic ancestors of the modern Udihe. The same populations used to be known to the Manchu by the term weji aiman ‘forest people’, and it is possible, though not fully confirmed, that the element udiin udi-he is a cognate of Manchu weji ‘forest’ (< *wedi, possibly from *ure-di < *xöre-dii). The Udihe ethnic territory, which until 1858–1860 was part of the Manchu Empire, is today divided between the Khabarovsk and Primorsk Krais of the Russian Far East. In recent decades, the officially registered ethnic population of the Udihe has declined from 1902 (1989) to 1496 (2010). The population is more or less equally divided between the two krais, with 620 individuals listed as resident in Khabarovsk Krai and 793 in Primorsk Krai in 2010, as well as 83 elsewhere. Traditionally, the Udihe used to be seminomads, moving within a limited territory usually along a particular river and its tributaries, thereby forming territorial groups which typically consisted of several clans. The principal territorial groups occupied, from south to north, the basins of the rivers Iman, Bikin, Samarga, Khor, Anyui, and Khungari (Gur). Additionally, there was a group living in the basins of the rivers Kur and Urmi to the west of the Lower Amur, as well as a group living at the coast of the Sea of Japan. Most of these groups were identified by names derived by the element -ƞ-kA from the corresponding hydronyms: imaƞka on the Iman, bikiƞka on the Bikin, samargiƞka on the Samarga, xuƞka on the Khor, uniƞka on the Anyui, xuƞgake on the Khungari, as well as namuƞka (a term also used for the Oroch) on the coast (from namu ‘sea’). In the 1930s, the Udihe were forcibly made sedentary. Every territorial group was settled in one or several permanent settlements, comprising Sanchikheza on the Iman, Olon, Syaïn, and Mitakheza on the Bikin, Agzu on the Samarga, Gvasyugi on the Khor, Bira on the Anyui, Kun on the Khungari, and Kukan on the Kur-Urmi. In 1934, the Jewish Autonomous Oblast’ was created on the Middle Amur, comprising also the Urmi basin, and soon afterwards the Udihe village of Kukan became a place of exile for political DOI: 10.4324/9781315728391-13

Udihe 327

prisoners. In the same period, after the construction of the railway from Khabarovsk to Sovetskaya Gavan’, the Udihe village of Kun became a railway station. As a result, the Udihe ended up constituting an insignificant part of the population in many of their “national” villages. In the 1960s and 1970s, in the course of the “consolidation-of-villages” campaign, several small Udihe villages were liquidated, including Sanchikheza on the Iman, Syaïn and Mitakheza on the Bikin, and Bira on the Anyui. The Bikin Udihe were resettled in the new village of Krasnyi Yar, while the Iman Udihe were transferred into neighbouring Russian villages. At present, the formerly compact Udihe ethnic territory, divided only by narrow watersheds, is split into a few disconnected settlements. The proportion of ethnic Udihe is largest at Agzu (80 % of the local population), Gvasyugi (65 %), and Krasnyi Yar (55 %). As a language, Udihe is closely related to Oroch, with which it was formerly lumped into a single “Oroch” language. Compared with Oroch, Udihe is, however, much more innovative, especially as far as its phonology is concerned, making mutual intelligibility rather limited. There are also dialectal differences within the Udihe language, corresponding to the territorial groups, but these differences are mainly connected with phonetics and lexicon and do not interfere with mutual intelligibility. From the linguistic point of view, there are significant similarities between the Iman and Bikin dialects, on the one hand, and between the Khor and Anyui dialects, on the other. This allows the Udihe dialects to be divided into a northern (Khor-Anyui) and a southern (Iman-Bikin) dialect cluster. In this framework, the Samarga dialect occupies an intermediate position. The information on the Kur-Urmi, Khungari, and coastal dialects, all of which are today extinct, is insufficient to define their position, though it seems that the Kur-Urmi dialect was strongly influenced by Ewenki and may have been transitional towards Kilen. Historically, the southern groups of the Udihe have been known by the name Kyakala (< *kïra+kala ‘border clan’), a name that is also the source of the exonyms used for the Udihe by the Nanai-Ulcha (kıakar) and Ghilyak (kéqal). Part of the Kyakala entered the composition of the Manchu, among whom their descendants are still recognized as a subethnic entity (Chinese 恰卡拉 qiakala). At the same time as the Udihe ethnic population has been declining, the number of people fluent in the Udihe language has been declining even more rapidly. According to the official census data there were 462 speakers in 1989, 227 in 2002, and 103 in 2010. However, these figures are overestimated. In reality, in 2002 only 155 people named Udihe as their “native language” in the places of traditional residence, while 72 people (31 %) of the alleged Udihe speakers lived outside the traditional territory and are unlikely to have spoken the ethnic language. By a more reliable estimate (A. Kh. Girfanova) the number of Udihe speakers in 2000 did not exceed 50. The drop has been particularly sharp in Primorsk Krai, where even the official census gave the number of native speakers as only 16 in 2010. This is connected with the unfavorable social conditions in the ethnic settlements, where especially men have a very short life expectancy. The situation is slightly better in Khabarovsk Krai, where the official number of 55 speakers in 2010 may have been close to correct, assuming that it includes also semi-speakers and individuals with a passive knowledge of the language. Some of the last speakers live today in the village of Gvasyugi. DATA AND SOURCES The first fixation of Udihe vocabulary items belongs to Richard Otto Maack, who in 1855 carried out an expedition to the Amur region on the assignment of the Russian

328 Elena Perekhvalskaya and Juha Janhunen

Geographical Society. Along with the Latin names of local animals and plants, he gives their names in Chinese, Nanai and Manchu, but in a few cases also in Udihe (Maack 1859: 112–113, 133–134, 138). In the subsequent decades, several researchers, including Nadarov (1887), Margaritov (1888), Protodiakonov (1888), and Leontovich (1898) did not distinguish Oroch and Udihe, referring to both by the name “Oroch” or “Orochën”. S. N. Braïlovskii (1901), who was the first to use the name “Udihe” (along with “Tazy”), also insisted on the cultural and linguistic unity of all “Oroch-Udihe”, but linguistically he separated the southern Udihe dialects from northern Udihe and Oroch. A somewhat similar point of view was held by Peter Schmidt (1928), and even much later, the ethnologist V. G. Lar’kin (1959: 5) considered Oroch and Udihe to be dialects of a single language, with Udihe further divided into several subdialects. Lexical material on Udihe and Oroch was also collected in the context of ethnographic documentation by V. K. Arsen’ev (posthumously 2008) and Stanisław (S. F.) Poniatowski (posthumously 1966, as discussed by Simonov 1990). The separate status of Udihe was first fully recognized by E. R. Schneider, who in the early 1930s created a literary norm and orthography for Udihe based on the Khor dialect and using the Unified Northern Alphabet. Apart from several school textbooks he compiled a dictionary and a grammatical sketch of the Khor dialect (1936), as well as a description of the Anyui dialect (1937). The literary use of Udihe was subsequently discontinued until a new Cyrillic-based orthography, also reflecting the Khor dialect, was created in the 1980s by M. D. Simonov and V. T. Kyalundzyuga. This orthography was used in a collection of Udihe folklore (Simonov & Kyalundzyuga & Khasanova 1998), in a three-volume Udihe dictionary (Simonov & Kyalundzyuga 1998), and in a number of school textbooks. However, the attempts at creating a unified orthography have not been successful, and both linguists and Udihe speakers continue using also other systems of notation. Following grammatical summaries by O. P. Sunik (1968, 1997) and K. H. Menges (1968), the first modern work on Udihe, containing an ethnographic introduction, a grammatical description, texts, and a dictionary, was published by I. V. Kormushin (1998) on the basis of fieldwork carried out in 1974–1984. The most fundamental description of Udihe, based on the Bikin dialect and containing an extensive grammar, sample texts, and thematic word lists, was published by Irina Nikolaeva and Maria Tolskaya (2001). A concise description of Udihe grammar was published by Albina (A. Kh.) Girfanova (2002), who also compiled a dictionary containing all major lexicographic material published earlier on Udihe (Girfanova 2001). Apart from Simonov & al. (1998), significant collections of texts have been published by Nikolaeva, Perekhvalskaya, and Tolskaya (2002, 2003), as well as, in Japan, by Kazama Shinjirō (2004ab, 2006–2010). Udihe language material was also collected by Tsumagari Toshirō, who published a collection of sample sentences (1997, 1998), a folktale (2011), and an extensive autobiographical narrative authored by Aleksandr Kanchuga (2002, 2003–2006). A grammatical topic is discussed by Kazama (2008), who has also authored a more comprehensive grammar with diachronic insights (2022). Particularly many specific studies have been devoted to the phonetic and phonological properties of Udihe, features that are relevant from the taxonomic point of view, though they also have practical implications for orthographical and notational solutions. A pioneer in the experimental analysis of the phonetics of Udihe (and of several other “northern” languages) was L. R. Zinder (1948), followed by Uzbek Baichura (1978). Other relevant studies, especially of the vowel system of Udihe, include Radchenko (1988),

Udihe 329

Simonov (1988), Janhunen (1999), Nikolaeva (2000), and Perekhvalskaya (2010). Another area of special studies is connected with the ethnohistorical position of Udihe speakers. The background and connections of the term “Udihe” are analysed by Hiu Lie (1978), while the historical and current position of the Kyakala is discussed in Janhunen & al. (1999) and Hölzl (2018). The current state of the linguistically Sinicized (later Russianized) group of Taz is summarized in Belikov & Perekhvalskaya (2002), as well as in Kazama & Podmaskin (2002). On the history of research of the Udihe language, see Girfanova (2003). In addition to the above-listed published materials, the present chapter is based on field data collected by one of the authors (Elena Perekhval’skaya) in 1984–2018. SEGMENTAL STRUCTURE The peculiarity of Udihe within the Tungusic family is largely connected with its phonetics and phonology, especially with the existence of vowel oppositions not attested in the other Tungusic languages, including Oroch. The northern dialect has four distinct series of vowels, which may phonetically be characterized as short, long, pharyngealized, and glottalized or laryngealized. The southern dialects have merged the pharyngealized series with the long vowels, and have, therefore, only short, long, and glottalized vowels. Pharyngealization (breathy voice) and glottalization (creaky voice) are features connected with phonation type, while the opposition between short and long vowels, which is also present in most other Tungusic languages, involves the suprasegmental feature of duration. However, it has been suggested that the four-way opposition of short, long, pharyngealized, and glottalized vowels could also be analysed in terms of tonal distinctions. This would mean that Udihe is a tonal language of the same type as Mandarin Chinese, which also has four (dialectally three or five) tones, and which is known to have been in contact with Udihe, especially in connection with the formation of the subethnic entity of the Taz. As in Mandarin, the vowels in Udihe are produced with a combination of phonational and durational features. Irrespective of how the four series of vowels in Udihe are analysed synchronically, their background is well known. Thus, the short vowels represent original short and long vowels, inherited from Proto-Tungusic, meaning that Udihe has, in general, lost original vowel length, which is still partially, though not systematically, preserved in Oroch. For this reason, the synchronic long vowels in Udihe are typically results of secondary contraction, caused by the loss of an intervocalic consonant, which in most cases was *g or *r. In a somewhat similar way, the pharyngealized and glottalized vowels are results of contraction, with the difference that they retain a trace of a segmentally lost intervocalic *s in the case of pharyngealization or *k in the case of glottalization. Considering this diachronic information, the four series of vowels are here transcribed as V (short) vs. VV (long) vs. VhV (pharyngealized) vs. V’V (glottalized), as in ƞala ‘hand’ (< *ƞaala), daama ‘(small of the) back’ (< *darama), ahanta ‘woman’ (< *asa-nta), ba’a- ‘to find’ : PRS-3SG ba’a-ni (< *baka-ra-ni). Several other notational solutions have been proposed by different authors; however, it should be noted that the graphic sequences VV vs. VhV vs. V’V do not necessarily imply a phonemic analysis, for they allow both monosegmental and sequential interpretations. An argument in favour of analysing the elements of pharyngealization (h) and glottalization (’) as separate segments is that they can be surrounded by different vowel qualities, as in ahikta ‘spruce’ (< *xasï-kta), uhende- ~ wehende- ‘to throw’ (< *useendee-), xa’i ‘liver’ (< *xaakï/n < *paa-ku/n), na’u ‘hen’ (< *nako).

330 Elena Perekhvalskaya and Juha Janhunen

As far as vowel qualities are concerned, the system seems to be rather uniform crossdialectally, although opinions differ concerning the details. For both northern (Khor-Anyui) and southern (Bikin) Udihe, a system of seven to eight short vowels has been postulated, comprising, maximally, two rounded back vowels u o [u o], two rounded front vowels ü ö [y ø], two unrounded back or central vowels a e [a ə], and two unrounded front vowels i ä [i æ] (Table 13.1). TABLE 13.1 UDIHE SHORT VOWELS u

e

ü

i

o

a

ö

ä

The rounded front vowels ö ü are diachronically secondary and synchronically marginal, and especially ü may even be absent as a distinct segment in the northern dialects. The short unrounded low front vowel ä is also relatively rare, and some idiolects of the northern dialect cluster are, in fact, reported to have the simple five-vowel system comprising only the segments a e i o u. Altogether, both the synchrony and the diachrony of the vowels in Udihe are still poorly understood, and the information given in the various descriptions of the language should not be taken at face value. The system of long (double) vowels (Table 13.2) is basically analogous to that of the short (single) vowels, except that it also includes an opposition between a low ää [eæ ~ æ:] and a mid-high ie [ie ~ e:], both of which can be of a contractive origin, as in kää ‘edge, border’ (< *kïra), siekte ‘thread’ (< *sire-kte), though ää is also the regular reflex of the original diphthongoid sequence *ïa, as in bää ‘moon, month’ (< *bïaga). The long rounded mid-high front vowel öö is often pronounced as diphthongoid [yø], as in möö [myø] ‘neck’ (< *mie < *miire < *müire), while the long rounded high front vowel üü is normally realized as [yi], as in düi ‘cradle’ (< *duri < *döri), though in some items it alternates dialectally with the sequence iu, as in (Bikin) küü ‘honey’ ~ (Khor) kiu ‘honey’. TABLE 13.2 UDIHE LONG VOWELS uu oo

 

üü

ee

öö

aa

ii ie ää

Morphological segmentation speaks strongly in favour of analysing the long vowels as sequences of two segments, cf. e.g. dili ‘head’ : PX1SG dili-i (< *dïlï-bï). The same is true of sequences of two different vowels, as in kusige ‘knife’ : PX1SG kusige-i : PX1PL kusige-u, in which the segments -i and -u function as separate morphemes. Strictly speaking, the system of long vowels merges with that of vowel sequences, and any distinction between the two is arbitrary. Commonly occurring sequences of two different vowel segments contain the high vowels i and u as either the first or the second segment and include (with i as the second segment) ai ei oi ui, (with u as the second segment) au äu eu ou öu, (with i as the first segment) io iu, as well as (with u as the first segment) ua ue. The system of pharyngealized vowels may also be understood as comprising both monophthongoid and diphthongoid sequences accompanied by the feature (or interrupted by the segment) of pharyngealization. The long vowels aa ee ii oo uu, corresponding to the five basic vowel qualities, have the monophthongoid pharyngealized counterparts

Udihe 331

aha ehe ihi oho uhu, while the four additional long vowels ää ie öö (= üö) üü (= üi) have, at least phonetically speaking, the diphthongoid pharyngealized counterparts ehä ihe ühö ühi. Most of the other diphthongoid sequences are also attested with pharyngealization. Again, morphological segmentation suggests that the pharyngealized vowels are, at least at the lexical level, composed of two independent vowel segments, separated by a pharyngeal element, as in umi- ‘to drink’ : PRS 2SG umi-hi : 2PL umi-hu. The situation is slightly different with the glottalized vowels, for the development of intervocalic *k to a glottal element seems to have been confined to the position after the non-high back vowels *a *o (including *aa *oo), yielding originally only the two glottalized monophthongs a’a o’o, as in ga’akta ‘cranberry’ (< *gaka-kta), ko’olo ‘mitten/s’ (< *kokolo < *koka-l.sa), but secondarily also ä’ä, as in bä’äsa ‘river’ (< *bïra-ka-caa/n, dialectally also > bääsa > bäsa). A glottalized e’e is reported to occur at morpheme boundaries, as in etete- ‘to work’ : PRT etete-’e. In glottalized diphthongoid sequences the first component is normally either a or o and the second component either i or u, as in ga’i ‘crow’ (< *gaakï), so’u ‘scoop’ (< *soka-bon). Due to secondary developments, the sequence e’i is also attested, as in ina’i ~ ine’i ‘dog’ (< *ƞïndakï/n) vs. sula’i ‘fox’ (< *sulakïï). The consonant system of Udihe is in line with the general Tungusic pattern. Not regarding the elements of pharyngealization (h) and glottalization (’), which, depending on the analysis, could also be classified as separate consonant phonemes, Udihe has a set of 19 consonants (Table 13.3), comprising, according to the manner of articulation, four nasals (m n ñ ƞ), four weak and four strong stops (b d j g vs. p t c k), three fricatives (f s x), two glides (w y), and two liquids (l r), and divided, according to the place of articulation, into labials (m b p f w), dentals (n d t s l r), palatals (ñ j c y), and velars (ƞ g k x). All these segments, except the vibrant r, are attested word-initially, but in other positions their distribution is regulated by a number of restrictions. The original *r being regularly lost (> Ø) in all positions, the synchronic r [r] is secondary and rare, and may be classified as a marginal phoneme, present only in a few recent loanwords, as in merge/n ‘hero’ (from Nanai), as well as in ideophones, as in mur&mur ‘toothless’. In a few borrowed items r alternates with l, as in oro ~ olo ‘reindeer’ (← oro/n). TABLE 13.3 UDIHE CONSONANTS m

n

ñ

ƞ

b

d

j

g

p

t

c

k

f

s

w

x y

l r

Intervocalic labial and velar stops show a tendency of spirantization, here considered to be non-phonemic for /g/ [ɣ] but phonemic for *b > /w/ [w], /p/ [ɸ ~ f], and (though only sporadically) *k > /x/ [x]. A distinctive /f/ [f] is also attested in word-initial position in recent loanwords, as in fuetiƞe ‘airplane’ (← Chinese feiji 飛機). Dialectal variation is present in the pronunciation of the palatal stops c j, which in the Khor dialect are generally realized as hushing (palato-alveolar or alveolo-palatal) affricates, but which

332 Elena Perekhvalskaya and Juha Janhunen

in the Bikin dialect have the hissing (dental) values [ts dz] when occurring before the phonetic velar vowels a e o u. The dentals n s, but not d t, tend to be palatalized before the phonetic palatal vowels ä i ö ü, meaning that the distinction between n [n] and ñ [ɲ] is phonemic only before a e o u. In the initial sequences ni and ƞi the nasal is often dropped in the Iman dialect, sporadically also in the Bikin dialect, as in ƞica’a ~ ica’a ‘little’, niƞka ~ iƞka ‘Chinese’. It may be noted that although several of the original Proto-Tungusic consonants in intervocalic position have undergone significant changes in Udihe, the corresponding segments have been restored due to secondary developments. Thus, the original intervocalic sibilant *s was debuccalized to *x [x ~ h], which is synchronically represented as the element of pharyngealization (h), but the niche of intervocalic *s was subsequently taken by *c, which was deaffricated to s [s], as in ise- ‘to see’ (< *ice-), while there is also a new medial x [x ~ h], distinct from the element of pharyngealization and representing original intervocalic *k or *ks, as in oloxi ‘squirrel’ (< *xölö-kii), buxe ‘cartilage’ (< *bö-kse), and appearing also in loanwords, as in bexe ‘ink’ (from Manchu). PHONOTACTICS AND MORPHOPHONOLOGY The basic syllable in Udihe is of the type (C)V(C). Independent word roots can consist of one, two, or three syllables. In word-final position, only open syllables are permitted, while medially clusters of two consonants can occur. The lexical root can, however, also end in an unstable nasal, which can be restored before suffixes, though many original nasal stems have been restructured as vowel stems, and the use of the nasal stem is often optional, e.g. xoto/n ‘town’ : NOM xoto : DAT xoton-du ~ xoto-du. The principal types of consonant clusters include the following: two weak stops, e.g. gegbi ‘name’; two strong stops, e.g. nakta ‘boar’; nasal + weak stop, e.g. oƞgo ‘cave’; nasal + strong stop, e.g. santu ‘fist’; two nasals, e.g. aƞma ‘mouth’. The nasal preceding a stop is typically homorganic with the latter, but in loanwords heterorganic combinations are also attested, e.g. yaƞja ‘pattern’ (← Chinese yangzi 樣子). In recent borrowings, the liquids l r can also occur in clusters containing a stop as the second component, e.g. alta ~ arta ‘tinplate’ (← Nanai alta/n). Geminates are rare and secondary, but at least kk is attested in regular vocabulary. Deviations from these phonotactic rules are often encountered in ideophones and onomatopoeia, as in gaataƞ&gaataƞ [flapping wings], kägdaf&kägdaf [river became covered with ice], kokkoleu [something moving very fast], punäll ~ puñarr [smoke rising above the house]. Rhythmically, each vowel segment counts as a single mora. Thus, a syllable with one short vowel (V) has one mora, while syllables containing complex vowels with length (VV), pharyngealization (VhV), or glottalization (V’V), have two morae. When used in isolation, monosyllabic (monomoraic) roots containing a short vowel tend to be lengthened, e.g. ja > jaa ‘kin’ (< *ja < *jaa), cf. jaa ‘ten’ (= jaa/n < *ju(w)a/n). This may have etymological reasons, as all original roots of this type, with the exclusion of pronominals and auxiliaries, contain a long vowel. The secondary (or etymological) length may also appear under other conditions, as before a high vowel of the following syllable. As a result, monosyllabic roots can show morphophonological alternation between short and long vowels, as in wa- ‘to kill’ (< *waa-) : FUT-3SG wa-jaƞa-ni : CV.CONN waa-mi. Note, however, that the long vowel observed at the surface could theoretically also be of a contractive origin, cf. e.g. ata waa ‘he will not kill’, in which waa = CONNEG *wa.a (< *wa-ra < *waa-ra).

Udihe 333

The moraic structure of words has consequences for word-level prosody. According to a traditional formulation (Cincius), a “bisyllabic” word in Udihe has two “stressed” syllables: the “dynamic stress” on the initial syllable, and the “musical stress” on the second syllable. In this context, “syllable” is to be understood as a mora, while “musical stress” refers to higher pitch. All words are divided into bimoraic sequences, each of which may be considered to constitute a rhythmic “foot” of its own. When a bimoraic word is expanded by suffixes, the pitch (in the following marked by the acute accent) moves to the last mora in trimoraic sequences, while in quadrimoraic sequences it is present on both the second and the fourth mora, e.g. tadá ‘arrow’ : INSTR tada-jí : DIR tadá-tigí. This prosodic structuring of the word is not phonologically distinctive, but it may be seen as a supplementary means marking the internal coherence of words and word forms. Vowel harmony in Udihe is rather seriously disrupted because of the changes in the vowel paradigm. Even so, both roots and suffixes preserve a basic opposition between the harmonically active vowels a e o and the harmonically inactive (neutral) vowels u i. Suffixes with a non-high vowel (*A) have normally three harmonic variants with a e o, whose distribution depends on the vowel of the preceding syllable, as in mafa ‘bear’ : LOC mafa-la, meime ‘harpoon’ : LOC meime-le, jolo ‘stone’ : LOC jolo-lo. The secondary vowels ä and ö behave like a and o, respectively, as in jaƞgä ‘chief’ : ACC jaƞgä-wa, tokö ‘cloud’ : ACC tokö-wo, while the secondary vowel ü behaves like u and is neutral. Within roots, the neutral vowels u i can be followed by a distinctive a or e, as in kuliga : ACC kuliga-wa ‘snake’ vs. kusige ‘knife’ : ACC kusige-we, but when located in the final syllable of the stem u and i require e in the suffix, as in ataxi ‘spider’ : ACC ataxi-we. A stem-final a is often represented as e before a suffixal -i, as in anda ‘friend’ : PX1SG ande-i (< *anda-bï). Violations of vowel harmony are also observed in Chinese and Russian loanwords, as in bouje ‘steamed pastry’ (← Chinese baozi 包子), ceradaka ‘attic’ (← Russian cherdák). WORD FORMATION Udihe retains the basic agglutinative typology of Proto-Tungusic, but agglutination is combined with elements of fusion especially in the verbal paradigm. The root always occupies the extreme left position in a word; it is followed by derivational and inflectional suffixes, which form a chain that can comprise up to four or five suffixes in the case of verb forms, as in etete- ‘to work’ : etete-ne-kce-he-hu ‘you would go there in order to work’ (work-AND-PROSP-PST-2PL), uli- ‘to sew’ : uli-gi-li-weƞ-ki-ti ‘they made (somebody) start sewing again’ (sew-REV-INCH-CAUS-PST-3PL). Just how much fusion Udihe has depends on how the complex vowels are analysed: in a sequential analysis they can be divided into segments belonging to different morphemes, meaning that any “fusion” in them takes place only at the phonetic level, while a monosegmental analysis would require the postulation of rather complicated morphophonological processes. Obviously, for the description of the morphological system, the sequential analysis is to be preferred. Most inflectable stems end in a vowel, with no formal difference between nominals and verb(al)s, though both parts of speech also have relicts of consonant stems, which are more prominent in verbal morphology. Both nouns and verbs have specific sets of inflectional and derivational suffixes. Derivation allows nouns to be converted to verbs (denominal verbs) and vice versa (deverbal nouns), but it is also used for semantic modification or specification without conversion (denominal nouns and deverbal verbs).

334 Elena Perekhvalskaya and Juha Janhunen

(1) Denominal nouns. This group comprises evaluative suffixes for diminutives, as well as a number of more specific suffixes with a varying degree of productivity. •

•

• •

•

• •

•

•

Diminutives are formed by the productive suffix -jiga’a, which may be added to both animate and inanimate nouns, e.g. ba’ata ‘boy’ : DIM ba’ata-jiga’a, bei ‘bow’ : DIM bei-jiga’a, jug-di ‘house’ : jug-jiga’a. A non-productive diminutive suffix is -sA’A, present mainly in kinship terms, e.g. mama ‘old woman’ : mama-sa’a ‘wife’, mafa ‘old man’ : mafa-sa’a ‘husband’, (*nexun :) nexu-se’e ‘younger sibling’, jooƞku ‘poor’ : jooƞku-so’o ‘poor man’, yoxo ‘pot’ : yoxoso’o ‘small pot’. The suffix -ƞkA derives nouns denoting inhabitants from names of places, e.g. biki/n [Bikin] : biki-ƞke ‘inhabitant of the Bikin valley’, guasi [Gvasyugi] : guasi-ƞka ‘inhabitant of the Gvasyugi valley’, solo’o ‘upstream’ : solo’o-ƞko ‘inhabitant of the upper reaches of a river’, jugdi ‘house’ : jugdi-ƞke ‘household member’, kolxojo ‘collective farm (kolkhoz)’ : kolxojo-ƞko ‘collective farmer’. The suffix -ƞA indicates deceased persons, e.g. uila’a ‘co-wife’ : uila’a-ƞa ‘late co-wife’, but it also has a pejorative function, e.g. uja [proper name]: PEJ uja-ƞa. The destinative suffix -ƞati denotes material prepared for producing something, e.g. uƞta ‘boot/s’ : DEST uƞta-ƞati ‘leather for sewing boots’, jugdi ‘house’ : DEST jugdi-ƞati ‘materials for a house’. It is often combined with the possessive suffixes, as in ei sexi bii tege-ƞati-i ‘this cloth is for my dress’ (this cloth 1SG dress-DEST-PX1SG). It is also attested in two lexicalized examples denoting persons: mafa ‘husband’ : DEST mafa-ƞati ‘bridegroom’, mama’asa ‘wife’ : mama’asa-ƞati ‘bride’. The suffix -lA, combined with the possessive suffix PX3SG -ni, derives relative nouns of measure from adjectival nominals, e.g. egdi ‘numerous, many’ : egdi-le-ni ‘multitude’, sagdi ‘big’ : sagdi-la-ni ‘size’, suƞta ‘deep’ : suƞta-la-ni ‘depth’. The suffix -(ƞa)’i denotes thickets of plants, e.g. amigda ‘poplar’ : amigda-’i ‘poplar thicket’, kuƞka ‘cedar’ : kuƞka-ƞa’i ‘cedar thicket’. The suffix -mulA, which represents the suffixalized form of the postpositionally used particle mule ‘together’, denotes pairs of individuals in a symmetrical relationship with each other, e.g. gagda ‘(an)other’ : gagda-mula ‘married couple’, tie ‘pair’ : tie-mule ‘pairs’; xa’a ‘relative’ : xa’a-mula ‘relatives’. Obscured suffixes with restricted productivity, but retaining a degree of semantic coherence, include: -ktA for various types of mass nouns, e.g. (fruit and plants) ahi.kta ‘spruce’, kusi.kte ‘Manchurian nut’, xai.kta ‘grass (sp.)’, (types of soil) oño.kto ‘sand’, lipa.kta ‘mud’, (small animals and insects) gawa.kta ‘butterflies’, ƞama.kta ‘mosquitoes’, sili.kte ‘worms’, (hair and feather) iƞa. kta ‘wool’, ofo.kto ‘feather’, gua.kta ‘beard’, sami.kta ‘eyelash’, (other) wa. kta ‘tree bark’; -mji for individuals, e.g. boƞgo ‘first, foremost’ : boƞgo-mji ‘leader’, sagdi ‘old’ : sagdi-mji ‘the eldest person; -ntA for people and groups of people, e.g. aha.nta ‘woman’, nii.nta ‘man, male’, bäo.nta [mythological forest dwellers], sagdi ‘old’ : sagdi-nta ‘old people’; -xi for animals and objects, e.g. olo-xi ‘squirrel’, ee-xi ‘frog’, tibje-xi ‘lynx’, sewe-xi ‘idol’, mule-xi ‘bucket’. Suffixes with obscured semantics include: -gA, e.g. kuli.ga ‘snake’, xalu.ga ‘hammer’, begdi ‘leg’ : begdi-ge ‘leg of table’, talu ‘birch’ : talu-ga ‘birch bark’, (*te- :) te-ge ‘clothes’ (cf. te-ti- ‘to put on clothes’); -pti ~ -kti, e.g. uña ‘finger’ : uña-pti

Udihe 335

~ uña-kti ‘ring’, too ‘fire’ : too-pti ‘fireplace’; -(C)ugu ~ -(C)ugA, e.g. cige ‘urine’ : cige-m.uge ‘bladder’, moo ‘tree, wood’ : moo-m.ugu ‘stack of firewood’, saƞña ‘smoke’ : saƞña-b.ugu ‘flue for smoke’, yaa ‘eye/s’ : yaa-gd.ugu ‘glasses’. (2) Deverbal nouns. This group comprises a number of semi-productive suffixes deriving full nouns with varying semantics and mostly with lexicalized meanings. •

•

• • •

The most frequent suffix is -ƞku, which has a wide range of meanings: (instrument of action) akpu- ‘to sweep’ : akpu-ƞku ‘broom’, jawa- ‘to take’ : jawa-ƞku ‘handle’, punde- ‘to blow’ : punde-ƞku ‘camera’, (result or object of action) nima-si- ‘to tell’ : nima-ƞku ‘tale’, nagbu- ‘to guess’ : nagbu-ƞku ‘riddle’, (place of action) au- ‘to wash’ : au-ƞku ‘washing place’, dau- ‘to cross’ : dau-ƞku ‘bridge’, (actor nouns, usually with a negative connotation) diga- ‘to eat’ : digaƞku ‘glutton’, joo- ‘to need’ : joo-ƞku ‘poor person’. The suffix -kci, often used in combination with the stative suffix -si-, denotes the place of action, e.g. STAT dalau-si- ‘to feed (animals)’ : dalau-si-kci ‘feeding place’, STAT diga-si- ‘to eat’ : diga-si-kci ‘dining room’, STAT ƞua-si- ‘to sleep’ : ƞua-si-kci ‘sleeping room’. The suffix -pti(-lA) ~ -kti(-lA) denotes the result or object of action, e.g. ulaa‘to soak’ : ulaa-pti ‘soaked fish’, cam-na- ‘to break’ : cam-pti-le ‘splinter’. The suffix -lA denotes people who are experts in a certain activity, e.g. aki- ‘to harpoon’ : aki-la ‘lucky harpooner’, wa- : waa- ‘to kill’ : waa-la ‘lucky hunter’, yexe- ‘to sing’ : yexe-le ‘good singer’. Actor nouns are formed by the suffix -mni, e.g. bele- ‘to help’ : bele-mni ‘helper’, STAT tatu-si- ‘to teach’ : tatu-mni ‘teacher’.

(3) Denominal verbs. Conversion between nominals and verb(al)s can, though only rarely, take place without derivational suffixes, e.g. yexe ‘song’ : yexe- ‘to sing’. Many of these cases are diachronically secondary, e.g. bude- ‘to die’ : bude ‘death’ (< AOR *bu-de < *bö-de). In most cases, verbalizing suffixes are used. •

• • •

The basic verbalizer -lA- (< *-lAA-), attested in all Tungusic languages, has in Udihe the aspectual connotation of singulative action, as opposed to semelfactive action expressed by the suffix -ndA-, and stative action expressed by the suffix -si-, e.g. kusige ‘knife’ : SING kusige-le- ‘to cut off with a knife’ : SEM kusige-nde- ‘to cut (a little) with a knife’ : STAT kusige-si- ‘to use a knife (in general)’, cf. also anda ‘friend’ : SING anda-la- ‘to make friends’ : STAT anda-si- ‘to be friends’. The verbalizer -dA- (< *-dAA-) denotes an instrumental relationship, e.g. gieu ‘oar’ : gieu-da- ‘to row with oars’, sa’i ‘salt’ : sa’i-da- ‘to salt’, silo ‘skewer’ : silo-do- ‘to roast fish on a spit’. The suffix -mA- forms captatives, e.g. olondo ‘ginseng’ : olondo-mo- ‘to collect ginseng’, sugjää ‘fish’ : sugjää-ma- ‘to fish’. It can also have an instrumental meaning, e.g. ikte ‘teeth’ : ikte-me- ‘to bite’, dukta ‘ski’ : dukta-ma- ‘to ski’. Translative verbs from nouns and adjectival nominals are derived by the suffix -nA-, e.g. amba ‘evil spirit’ : amba-na- ‘to become an evil spirit’, aƞnaxi ‘strange’ : aƞnaxi-na- ‘to become strange’, baya ‘rich’ : baya-na- ‘to become rich’, daampi ‘shabby’ : daampi-ne- ‘to become shabby’, lusa ‘Russian’ : lusa-na- ‘to become Russian ~ Russified’, sagdi ‘old’ : sagdi-nta ‘old people’ : sagdi-nta-na- ‘to become old (of people)’. This suffix is also present in verbs

336 Elena Perekhvalskaya and Juha Janhunen

•

referring to natural phenomena, e.g. imaa ‘snow’ : imaa-na- ‘to snow’, suu ‘sun’ : suu-ne- ‘to shine (of the sun)’. Other verbalizers with more diffuse functions include -di-, e.g. saƞa ‘hole’ : saƞa-di- ‘to mend clothes’, kumu ‘flea’ : kumu-di- ‘look to fleas’; -li-, e.g. sau ‘torch’ : sau-li- ‘to light a torch’, wee ‘mountain’ : wee-li- ‘to cross a mountain’; and -ƞi-, e.g. gegbi ‘name’ : gegbi-ƞi- ‘to name’, kawa ‘hut’ : kawa-ƞi- ‘to make a hut’, lala ‘gruel’ : lala-ƞi- ‘to cook gruel’. The suffix -li- is also attested with adjectival nominals, often with an essive or translative meaning, e.g. egdeƞge ‘surprising’ : egdeƞge-li- ‘to behave in an eccentric manner’, gili ‘cold’ : gili-li‘to get cold’, guläla ‘crazy’ : guläla-li- ‘to go crazy’.

(4) Deverbal verbs. Deverbal verbs in Udihe, as in the other Tungusic languages, are a complex group of derivational forms somewhat diffusely connected with the categories of voice, modality, aspect, and Aktionsart. The productivity and level of grammaticalization of these forms varies, and their taxonomy is open to various interpretations. However, because of their connection with grammatical functions most of them are best dealt with in the context of verbal morphology. The following are a couple of form groups that remain outside of the mainstream categories of deverbal verbs. •

•

Andatives (also known as “directives”) are formed by the productive suffix -nA-, e.g. wakca- ‘to hunt’ : AND wakca-na- ‘to go to hunt’, egbesi- ‘to swim’ : AND ebgesi-ne- ‘to go to swim’, olokto- ‘cook’ : AND olokto-no- ‘to go to cook’. Intensives are formed by the semi-productive suffix -mA-, e.g. tukä- ‘to run’ : INTENS tukä-ma- ‘to run quickly’, xukti- ‘to jump’ : INTENS xukti-me- ‘to jump quickly away’. With transitive stems, this suffix implies the presence of multiple objects, e.g. saula- ‘to light (a lamp)’ : INTENS saula-ma- ‘to light (many lamps)’, diga- ‘to eat’ : INTENS diga-ma- ‘to eat a lot of everything’.

NUMBER AND CASE Udihe has lost the primary plural markers *-l and *-r, and it also lacks any productive traces of the complex marker *-sA-l, as still attested in Oroch. The suffix -ntA, sagdi ‘old’ : sagdi-nta ‘old people’, may in some cases be understood as a marker of a collective plural. Two other markers of collective plurality are -jigA and -getu, used optionally in reference to humans or, more rarely, animals. The marker -jigA ~ -jiga (apparently borrowed from Chinese jige 幾個 ‘several’) is used with nouns, e.g. exe ‘elder sister’ : PL exe-jiga ~ exe-jige, tukca ‘hare’ : PL tukca-jiga, maƞmu ‘Nanai’ : PL maƞmu-jige. It can also indicate associative plurality, e.g. kolko [proper name] : PL kolko-jige ‘Kolko and his friends’. The marker -getu (from the noun *getu ‘people’, as attested in Oroch) is mainly used with participles, as well as with some adjectival nominals and pronouns, e.g. bagdi‘to live’ : PTCP.PRS bagdi-i ‘living’ : PL bagdi-i-getu ‘living beings, people’, bude- ‘to die’ : PTCP.PRS bude-i ‘dying’ : PL bude-i-getu ‘dying people’, wakca- ‘to hunt’ : PTCP. PRS wakca-i ‘hunting’ : PL wakca-i-getu ‘hunting people, hunters’, xoƞto ‘other, alien’ : PL xoƞto-getu ‘strangers’, DIST uti ‘that’ : PL uti-getu ‘those (people)’. There is also a possessive plural in -na/n-, used with animate nouns in combination with the possessive suffixes, e.g. a’i/n ‘elder brother’ : PL-PX1SG a’i-na-mi ‘my elder brothers’, nakta site-ne-ni ‘wild piglets’ (boar child-PL-PX3SG).

Udihe 337

With non-specific objects, singular is used also with people. Plural is not used in the presence of quantifiers, e.g. jube ugda ‘two boats’ (two boat). On finite predicates, third-person plural forms are used only in agreement with plural animate subjects. The system of nominal cases comprises eight suffixally marked forms: accusative, designative, dative, locative, prolative, ablative, directive, and instrumental (Table 13.4). There are two stem types: vowel stems and nasal stems. The nasal stems end in the unstable nasal */n, which in Udihe is lost in absolute position, but which appears in inflection and conditions allomorphic variation of certain case markers. The former stems ending in *g and *l (> *g), as still present in Oroch, have been restructured in various ways, e.g. jug- : jug.di ‘house’ (< *juu/g-), yaa < *yäha ‘eye/s’ (< *ïa-sa-l).

TABLE 13.4 UDIHE CASE MARKERS  

V

/n

ACC

-wA

-mA

DES

-nA/n-

DAT

-DU

n-dU

ABL

-digi

n-digi

LOC

-lA

n-dulA

PROL

-li

n-dUli

DIR

-tigi

n-tigi

INSTR

-ji

n-ji

Stem types: V = vowel stems, /n = nasal stems. Segmental alternations: A = a e o (harmonic variation), U = u ~ i (contextual variation), D = d ~ n (contextual variation).

•

•

The unmarked nominative is used as the quotative form of a noun; it typically codes the subject argument of a transitive or intransitive verb; it is used in copular constructions and in the function of a nominal predicate. In possessive and genitive constructions, it marks the modifier (possessor) in combination with the possessive form of the headword, e.g. ba’ata ƞala-ji-ni ‘by the boy’s hand/s’ (boy hand-INSTRPX3SG), similarly in spatial constructions, e.g. jugdi doo-ni-ni ‘inside the house’ (house inside-DAT-PX3SG). The nominative is required by the verbs o.do- ~ edeand o.si- ‘to become’, e.g. alagdiga’a belie osi-gi-he-ni ‘she became a beautiful girl’ (beautiful girl become-REV-PST-3SG). The accusative, marked by the suffix -wA for vowel stems and -mA for nasal stems, prototypically codes the direct definite or specific object (patient, goal, stimulus, theme, etc.). An indefinite or non-specific noun in object position is unmarked. The accusative is used with speech verbs denoting the information transmitted, e.g. bude-i-getu keye-we-ni teluƞusi-jeƞe-i ‘I will tell words of dying people’ (die-PTCP. PRS-PL word-ACC-PX3SG tell-FUT-1SG). It denotes the beneficiary in constructions with the verb belesi- ‘to help’, e.g. nii ajiga-wa belesi-i-ni ‘who will help the girl?’ (who girl-ACC help-PRS-3SG). The accusative is also used with directional verbs to denote the route of movement, e.g. bä’äsa-wa solo’o-[h]o-ti ‘they went up the river’ (river-ACC go.upstream-PST-3PL).

338 Elena Perekhvalskaya and Juha Janhunen

•

•

The designative case, marked by the suffix -nA/n and always combined with the possessive suffixes, indicates a new object or a new status resulting from the action, e.g. bii sina-na-mi oo-mi ‘I make myself a backpack’ (1SG backpack-DES-PX1SG make-1SG), bii sii mafa-na-hi ba’a-jaƞa-i ‘I will find you a husband’ (1SG 2SG husband-DES-PX2SG find-FUT-1SG), befule-du solo’o-du jogjo-no-ti cem&cem oo-ho-ni ‘to the nuthatch and the goldcrest he made sharp beaks’ (nuthatch-DAT goldcrest-DAT beak-DES-PX3PL sharp make-PST-3SG). Most often, the designative is used with inalienable nouns (kinship terms, names of personal objects). When used with alienable nouns, the mark of alienability is added before the case marker, e.g. mene&mene be’äsa-ƞi-na-fei jawa-ha-ti ‘each one [of the Udihe clans] took a river’ (REFL&REFL river-AL-DES-RX.PL take-PST-3PL). The basic dative marker is -du, but the variant -di is facultatively used, especially in the Khor dialect, before or after a neighbouring i, e.g. eniƞe ‘mother’ : DAT-PX1SG eniƞe-di-i. With certain spatials, the combination DAT-PX3SG -du-ni is replaced by -ni-ni, e.g. doo ‘inside’ : dektuge doo-ni-ni ‘inside the box’ (box inside-DATPX3SG). The dative has both functional and spatial uses. In functional use it codes (1) the recipient (“indirect object”), e.g. exi-ni amba-du diga-laga-ni bu-ho-ni ‘the elder sister gave the food to the forest spirit’ (elder.sister-PX3SG forest.spirit-DAT eat-CV.PURP-PX3SG give-PST-3SG); (2) the agent in passive constructions, e.g. bii aga’a-di-i uu-ja-hi ‘you will be killed by my elder brother’ (1SG elder.brother-DATPX1SG be.killed-OPT-2SG); as well as (3) the causee in causative constructions, e.g. uti nada lusa-jiga-du sugala-wa teti-weƞ-ki-ni ‘she made those seven Russians put on skis’ (that seven Russian-PL-DAT ski-ACC put.on-CAUS-PST-3SG). In spatial use the dative denotes (4) the location where an action takes place, e.g. udi[h]e maƞmu omo baa-du wakca-[h]a-ti ‘the Udihe and Nanai used to hunt in the same place’ (Udihe Nanai one place-DAT hunt-PST-3PL); or the time when an action takes place, e.g. ali=da ekin-du ‘in a certain moment’ (when=PTCL time-DAT), geäwan ‘dawn’ : DAT geäwan-du ‘at dawn’. The spatial uses give rise to a wide range of functional extensions, including, most importantly, the possessive construction, e.g. mama-du o’i bie ‘grandmother has poultry’ (grandmother-DAT poultry be.PRS). The dative form of participles functions as a quasiconverb with a circumstantial, temporal, or causal meaning, as in [1]:

[1] a’i-wa tene ba’a-han-di-fei liquor-acc for.the.first.time find-ptcp.prf-dat-rx.pl ‘When they found spirits for the first time, umi-mei sokto-ho-ti drink-cv.conn become.drunk-pst-3pl they drank it and got drunk.’ •

The ablative is marked by the suffix -digi (< *-dugi) for both stem types. Exceptionally, the ablative of jugdi ‘house, home’ is jug-digi, suggesting either the original consonant stem jug- (< *juu/g) or a haplological shortening of *jugdi-digi, cf. DAT jugdi-du. Functionally, the ablative denotes the starting point of a movement or the source of an action, e.g. a[h]i-ƞa’i-digi aima die-li-he-ni ‘a hazel hen flew out of the spruce thicket’ (spruce-DX-ABL hazel.hen fly-INCH-PST-3SG), tokö-digi tigdei-ni ‘it is raining from the cloud’ (cloud-ABL rain-PRS-3SG), abuga-digi-i okto-wo

Udihe 339

•

•

•

gaju-gi-ye ‘bring me medicine from your father’ (father-ABL-RX medicine-ACC bring-REV-IMP). Temporal and metaphoric uses are also common, e.g. uta-digi caixi ‘from now on’ (this-ABL further), mafa-digi seke-he-ti bie ‘they descend from a bear’ (bear-ABL generate-PST-3PL be.PRS). In the comparative construction, the ablative marks the standard of comparison, e.g. ogböö käƞa-digi sagdi ‘an elk is bigger than a deer’ (elk deer-ABL big). The locative is marked by the suffix -lA for vowel stems and -dulA ~ -dilA for nasal stems. The locative marks the achieved reference point and is typically used with the verbs ƞene- ‘to go’, eme- ‘to come’ : REV eme-gi- ‘to come back’, ihi- : REV ihi-gi- ‘to reach’, nagda- ‘to hit’, e.g. tee auƞga-la i[h]iƞ-ka-u ‘there we reached the hunting shelter’ (there hunting.shelter-LOC reach-PRT-1PL.EXCL), keämikta-la-ni nagda-ha-ni ‘he hit the temple (of a boar)’ (temple-LOC-PX3SG hit-PST-3SG). Compared with the dative, with which it is often interchangeable, the locative expresses a more diversified relation with the reference point, cf. e.g. mene aka-la-i ‘on(to) her back’ (REFL back-LOC-RX), uli ‘river’ : LOC uli-le ‘at the river’ ~ ‘out of the river’, often implying a physical contact with the reference point, e.g. dili-le-ni jawa-’a=de ‘(he) grabbed his (another person’s) head’ (head-LOC-PX3SG take-PRT=PTCL). The locative also indicates the source of information, e.g. yaa-la saa-ja-hi ‘you will know by the eye’ (eye-LOC know-OPT-2SG). More rarely, the locative indicates time, e.g. jaa see-le-i ‘when I was ten’ (ten age-LOC-PX1SG). A lexicalized example is tima­ dula ‘in the morning’ (possibly from LOC *tima-r-dula), cf. timana ‘tomorrow’. The prolative parallels the locative and is marked by the suffix -li (for vowel stems) and -duli ~ -dili for nasal stems. It codes the location through, along, or across which the movement takes place, e.g. igee tege-li akku-si-i-ni ‘the gadfly stings through the clothes’ (gadfly clothes-PROL bite-STAT-PRS-3SG), xoƞto xokto-li ƞeni-i-ni ‘(he) is returning by another road’ (other road-PROL go-PRS-3SG), biki deen-duli-ni weli-he-ni ‘(she) crossed the mountain across the upper reaches of the river Bikin’ (Bikin upper.reaches-PROL-PX3SG cross.mountain-PST-3SG). The prolative is also used in reference to a guideline according to which a movement is realized, e.g. topto topto-i-li-ni ƞen[e]-he-ni ‘(he) went back in the direction of the owl’s hooting’ (screech.owl hoot-PTCP.PRS-PROL-PX3SG go-PST-3SG), mafa xokto-li-ni ƞeneh e-ni ‘(he) went by the bear’s footprints’ (bear road-PROL-PX3SG go-PST-3SG). The directive, marked by the suffix -tigi, indicates the reference point in the direction of which an action takes place (with no contact with the reference point). It is used with verbs of motion and verbs of directed activity, such as buisi- ‘to sneak (up to somebody)’, tuunde- ‘to spit (at somebody)’, gakpa- ‘to shoot (at something)’, e.g. anana bi-sin-tigi-fei ƞeni-li-he-ti ‘they set off in the direction of the place where they lived before’ (before be-PTCP.PRF-DIR-RX.PL go-INCH-PST-3PL). The directive is also required by the verbs guailana- ‘to turn (into something)’ and mafala- ‘to marry (of women)’, e.g. bua bui-tigi-ni guailana-’a-si ‘having turned into a taiga beast . . . ’ (taiga beast-DIR-PX3SG turn-PRF-CONJ), jabdala’a yele-tigi nii mafala-ja ‘who will marry a snake lizard?’ (snake lizard-DIR who marry-OPT). With speech verbs and some verbs of directed perception and emotional activity the directive marks the addressee of the speech, the object of the directed perception, or the causator of the emotion, e.g. ñajakta-hi bi-si=de ñajakta-tigi gele-muse-i ‘if you had a brother, I would have asked him’ (younger.brother-PX2SG be-PST=PTCL youngerbrother-DIR ask-DESID-1SG), xuñaji-ni tagda-ha-ni mafa-tigi-ni ‘his sister was angry with her husband’ (sister-PX3SG be.angry-PST-3SG husband-DIR-PX3SG).

340 Elena Perekhvalskaya and Juha Janhunen

•

The instrumental, marked by the suffix -ji, indicates the means (instrument, tool, material, etc.) of action, e.g. mafa sä’ä-ji-ni uƞtu-de-he-ni ‘she drummed with a bear’s shoulder blade’ (bear shoulder.blade-INSTR-PX3SG drum-VBLZ-PST-3SG), moo pa’i-wa-ni ƞala-ji jawa-ha-ni ‘(he) took a piece of wood with his hand’ (wood pieceACC-PX3SG hand-INSTR take-PST-3SG), naata-ji-ni xula-na-mi woo-gi-he-ni ‘(he) made a blanket from skin’ (skin-INSTR-PX3SG blanket-DES-RX make-REVPST-3SG), mui-ji xukti-gi-si-weƞ-ki-ni ‘(he) made (her) ride a horse’ (horse-INSTR run-REV-STAT-CAUS-PST-3SG), nii keye-ji-ni soƞo-li-he-ni ‘(he) began to cry with a human voice’ (human voice-INSTR-PX3SG cry-INCH-PST-3SG). The instrumental is also used in comitative and reciprocal constructions. In the comitative function it is often, though not always, accompanied by the postposition geye ‘together’, bi=de sin-ji geye ƞene-je-mi ‘let me go with you’ (1SG=PTCL 2SG-INSTR together go-OPT-1SG), ogbö[ö] eni-ji-i buga-du oƞko-i-ti ‘elks graze with their females in the clearing’ (elk female-INSTR-RX graze-PRS-3PL). The instrumental is required by some verbs of emotion and attitude, including ƞele- ‘to be afraid’, xelene- ‘to treat’, susa- ‘to avoid’, wa’ali- ‘to fight (over something)’, e.g. ina’i jeu-ji-ni wa’alii-ti ‘the dogs are fighting over the food’ (dog food-INSTR-PX3SG fight-PRS-3PL).

ADJECTIVES Adjectives, or adjectival nominals, retain in Udihe their basic nominal characteristics with some derivational and syntactic properties different from regular nouns. They are typically used as adnominal modifiers, but they can also function as independent head nouns, in which role they can take case suffixes and be modified by a noun, e.g. аyawa=da gehe-we=de wa-ha-mi ‘I (seem to have) killed both bad (meat) and good (meat)’ (good-ACC=PTCL bad-ACC=PTCL kill-PST-1SG), niƞka sexi-ma tege ‘Chinese cotton dress’ (Chinese fabric-DX clothes). However, unlike regular nouns, adjectives can be modified by adverbs, e.g. teu bali ‘completely blind’ (all blind), asi alagdiga’a ‘very beautiful’ (very beautiful). Some adjectives may have a second argument, e.g. oño-lo paki a[h]anta ‘a woman skilful in embroidery’ (embroidery-DX skilful woman). Adjectival nominals in the instrumental case form function as modal adverbs, e.g. aya ‘good’ : INST aya-ji ‘well’. Qualitative adjectives may form comparative constructions, e.g. yaa ina’i-digi sagdi ‘a cow is bigger than a dog’ (cow dog-ABL big). The superlative function is expressed by the adverb co ‘most’, e.g. muyi teu-ji-ni co sagdi ‘horse is the biggest of all’ (horse allINSTR-PX3SG most big). A few adjectives, usually referring to people, can form superlatives with the suffix -dimA, e.g. a’in-dima ‘eldest’, maƞga-dima ‘strongest’. Intensity of the quality can also be expressed by the suffix -ƞku, e.g. aya ‘good’ : aya-ƞku ‘better, best’, igdi ‘loud’ : igdi-ƞku ‘very loud’, gehe ‘bad’ : gehe-ƞku ‘very bad’, eƞme ‘wide’ : eƞme-ƞku ‘very wide’, as in keye-ti igdi-ƞku eƞme-ƞku bi-si-ni ‘their voices were loud and sonorous’ (voice-PX3PL loud-DX wide-DX be-PST-3SG). The class of adjectival nominals comprises both basic roots and complex stems derived from nominal and verbal roots. •

The most important category of derived adjectives is formed by the proprietives, marked by the suffix -xi, e.g. xuñaji ‘sister’ : PROPR xuñaji-xi ‘having a sister’. Many proprietive forms have a lexicalized meaning, e.g. unugu ‘ilness’ : PROPR unugu-xi ‘ill’, meye ‘mind’ : meye-xi ‘clever’. One of the functions of proprietives

Udihe 341

•

•

•

•

is to express possession, as in bu site-xi bi-u ‘we have children’ (1PL.EXCL childPROPR be-1PL.EXCL), ila ñeƞuhe mama’asa-xi o-si-ti ‘the three ogres got married’ (three ogre wife-PROPR become-PST-3PL). The possessed object may be modified by a noun in the instrumental case, as in a’i-ni wokto-ji tege-xi bi-si-ni ‘his elder brother had a shirt of grass’ (elder.brother-PX3SG grass-INSTR clothes-PROPR be-PST-3SG). When negated, the proprietive forms are replaced by the privative construction, which involves the “partitive” form in -lA of the noun in combination with the privative noun anci, as in jää-ƞi-le anci bi-[h]i-ni ‘he has no money’ (money-AL-PART PRIV be-PRS-3SG). The Udihe “partitive” suffix -lA represents diachronically a variant of the designative case marker -nA, which in Oroch still has the allomorphs -yA ~ -lA ~ -nA depending on the stem type. Material from which an object is made is expressed by the suffix -mA, e.g. aisi ‘gold’ : aisi-ma ‘golden’, wokto ‘grass’ : wokto-mo ‘made of grass’, talu (moo-ni) ‘birch (tree-PX3SG)’ : talu-ma ‘made of birch bark’. This suffix is also used to denote a class or a subclass to which the modified noun belongs, e.g. eexi-me mama’asa ‘frog wife’ (frog-DX wife), sewe-me anda ‘a spirit friend’ (spirit-DX friend), cuguse-me o’o ‘duck’ (duck-DX fowl). Reference to gender and ethnic context can also be expressed in this way, e.g. ba’ata-ma site ‘boy’ (boy-DX child), guasa’a-ma ina’i ‘bitch’ (bitch-DX dog), lusa-ma jugdi ‘a Russian house’ (Russian-DX house). The suffix -ligi forms adjectives referring to external features, e.g. bombo ‘ball’ : bombo-ligi ‘spherical’, kico ‘top’ : kico-ligi ‘(with a) sharp (top)’, kede ‘stripe’ : kede-ligi ‘striped’. This suffix is used in many colour terms, e.g. paa-ligi ‘black’, caa-ligi ‘white’, ñöö-ligi ‘green, blue’. The same roots may also take the suffix -lA ~ -li, denoting a variant of the colour, e.g. paa-la ‘intense black’, xoo-lo ‘yellowish’, ñöö-lo ‘dark-green’. The suffix -nA’use points at an unusual (often negative) feature of the noun modified, e.g. ikte ‘tooth’ : ikte-ne’use ‘with large teeth’, lawa ‘failure’ : lawa-na’use ‘failing (in hunting, etc.)’. This suffix can also form adjectives from verbs, e.g. oƞmo- ‘forget’ : oƞmo-no’use ‘forgetful’, soƞo- ‘to weep’ : soƞo-no’use ‘crybaby’. The suffix -fi ~ -pti forms adjectives from temporal nouns and adverbs, e.g. anana ‘before, in the past’ : ana-fi ~ ana-pti ‘former, ancient’, eineƞi ‘today’ : eine-fi ‘today’s’, neki ‘spring’ : neki-fi ‘vernal’.

NUMERALS The numerals for the basic digits represent regular heritage from Proto-Tungusic: 1 omo, 2 juu ~ jube, 3 ila, 4 dii, 5 tuƞa, 6 ñuƞu, 7 nada, 8 jakpu, 9 ye(y)i, 10 jaa. Of the decades, only 20 we(y)i (< *xorï/n) is expressed by a separate lexeme, while the others are formed by multiplication with the element 10 -jAA, e.g. 30 ila-jaa (3 × 10), 40 (dii-jee), 50 tuƞajaa, etc. The intermediate numerals are expressed additively, e.g. 11 jaa omo, 12 jaa juu ~ jaa jube, 13 jaa ila, 14 jaa dii, 15 jaa tuƞa, etc. For the powers of ten, 100 (omo) taƞgu, 1000 (omo) miƞga, are used. Other numerals are formed by a combination of additive and multiplicative constructions, e.g. 352 ila taƞgu tuƞa-jaa juu, 888 jakpu taƞgu jakpu-jaa jakpu, 10,000 jaa miƞga, 100,000 omo taƞgu miƞga. Most of the numeral roots, including also 5 tuƞa (< *tuƞa), but excluding 2 juu, 100 taƞgu, and 1000 miƞga, are nasal stems, e.g. 3 ila = ila/n : ACC ila-ma : DAT ilan-du : ABL ilan-digi : LOC ilan-dile : PROL ilan-dili : DIR ilan-tigi : INSTR ilan-ji. The form 2 jube is historically a petrified accusative ju-be, which retains an indirect trace of the

342 Elena Perekhvalskaya and Juha Janhunen

earlier consonant stem in */g < *-l (< *-r), i.e. *jul-be (< *jöör-be). Synchronically, this form can be used both as a nominative and as an accusative, though the pleonastic form ACC jube-me is also attested, the other case forms being DAT juu-du : ABL juu-digi : LOC juu-dile : PROL juu-dili : DIR juu-tigi : INSTR juu-ji. It may be noted that the numeral 9 ye(y)i (< *yegi/n) in Udihe represents a different root variant as compared with Oroch xuyu/n (< *xöyöö/n), the Oroch form being shared by Nanaic and Ewen, while the Udihe form is identical with the rest of Ewenic. This difference between Udihe and Oroch is, however, secondary, as the Oroch form is a recent borrowing from Nanaic. Ordinal numerals are formed by the suffixes -Ati ~ -nti, e.g. omo-iti ‘first’, ile-äti—ilenti ‘third’, dii-eti ~ dii-nti ‘fourth’, tuƞ-äti ~ tuƞa-nti ‘fifth’, etc., a lexicalized exception being gagda ‘second’. Multiplicatives are formed by the suffixes -lie ~ -lää, e.g. juu-lie ~ juu-lää ‘twice’, ila-lie ~ ila-lää ‘three times’, dii-lie ~ dii-lää ‘four times’, etc., with emne’e ‘once’ as an exception. Other derivatives from the numeral stems include the limitatives (‘only’) in -tA’A ~ -sA’A, e.g. omo-so’o ~ emu-se’e ‘only one’, jube-te’e ‘only two’, ila-ta’a ‘only three’, etc., and the augmentatives (‘as many as’) in -ndimali, e.g. jube-ndimali ‘as many as two’, ila-ndimali ‘as many as three’, tuƞa-ndimali ‘as many as five’. The function of distributive numerals is expressed by the reduplication of the corresponding cardinal numerals, e.g. omo&omo ‘one each’, juu&juu ~ jube&jube ‘two each’, ila&ila ‘three each’, etc. PRONOUNS The basic personal pronouns in Udihe are fully identical with those in Oroch, with separate oblique stems used before the case markers, and with additional stem variants present in the accusative and possessive forms (Table 13.5). In the Khor dialect the full accusative forms with ACC -we can be replaced by the shorter variants 1SG mine : 2SG sine : 1PL mune : 2PL sune, similar to the pronominal partitives attested in Ewenic. The basic first person plural pronoun has an exclusive reference. The corresponding inclusive pronoun is minti : OBL minti- (vs. Oroch biti : biti-). For the third person, but mainly with an obviative function, the stem nua/n- is used, almost always expanded with the possessive suffixes and yielding 3SG nua-ni : ACC nua-me-ni : POSS nua-ƞi(-ni) : OBL nuan-CXni vs. 3PL nua-ti : ACC nua-me-ti : POSS nua-ti-ƞi : OBL nuan-CX-ti. Alternatively, the demonstrative uti ~ uta ‘this’ or the regular noun beye ‘body, person’ : PX3SG beye-ni : PX3PL beye-ti can be used for the third person. The stems nua- and beye- are used only in reference to humans. TABLE 13.5 UDIHE PERSONAL PRONOUNS   SG

1 2

PL

1 2

EXCL

NOM

OBL

ACC

POSS

bii

min-

min.e-we

mini-ƞi

sii

sin-

sin.e-we

sini-ƞi

buu

mun-

mun.e-we

muñu-ƞu

suu

sun-

sun.e-we

suñu-ƞu

The possessive forms are used in predicative position, e.g. uti laƞi nii-ƞi—mini-ƞi ‘whose is this trap?—mine’ (this trap who-POSS—1SG-POSS), ei ugda nua-ti-ƞi ‘this boat is theirs’ (this boat 3P-PX3PL-POSS). They are normally not used adnominally,

Udihe 343

though such usage is occasionally attested in the modern language, perhaps due to language attrition, e.g. sini-ƞi baa ‘your land’ (2SG-POSS land). The reflexive pronoun is me/n-, which in object position takes the singular and plural reflexive suffixes with no case marking: SG me-mi ~ me-mui : PL me-fi ~ me-fei. In the marked case forms the case marker precedes the reflexive suffixes: DAT SG men-di-i : PL men-di-fi, ABL men-digi-i : PL men-digi-fi, LOC SG men-dile-i : PL men-dile-fi, PROL SG men-dili-i : PL men-dili-fi, DIR SG men-tigi-i : PL men-tigi-fi, INSTR men-ji-i : PL men-ji-fi. Exceptionally, the designative forms are SG men-dine-mi : PL men-dine-fi, containing the complex suffix -di-nA/n- (= DAT-DES). The oblique forms are used in adverbal position, e.g. buu ñöö ñaa-wa-ni men-dine-fi waa-mu ‘we caught the sable skins for ourselves’ (1PL.​EXCL sable skin-ACC-PX3SG REFL-DES-RX.PL kill-1PL.EXCL). The reflexive pronoun also has the possessive forms SG meni-ƞi-i : PL meni-ƞi-fi. In adnominal position the form mene ‘one’s own’ is used, e.g. mene kekese-i mene do’opo ooƞ-ki-ni ‘she made her servant look like herself’ (REFL servant-RX REFL likeness make-PST-3SG). The emphatic function is expressed by the construction mene beye-ji (REFL person-INSTR), e.g. lusa mene beye-ji naa saƞa-ni cul ƞene-he-ni ‘the Russian went into that hole in the ground by himself’ (Russian REFL person-INSTR ground hole-PX3SG DESCR go-PST-3SG). The reduplicated form mene&mene has a distributive function (‘each one’), e.g. mene&mene xokto-ji-fi ƞene-kte-gi-[h]e-ti ‘each one of them went along his own road’ (REFL&​REFL road-INSTR-RX.PL go-DISTR-REV-PST3PL), while the corresponding form marked for reflexivity me-fi&me-fi ~ me-fei&me-fei (= REFL-RX.PL) functions as a reciprocal pronoun, e.g. nua-ti me-fei&me-fei teti-gi-i-ti ‘they dress each other’ (3P-PX3PL REFL-RX.PL&REFL.​RX.PL dress-REV-PRS-3PL). The basic demonstrative pronouns are PROX eyi ~ ei ‘this’ vs. DIST teyi ~ tei ~ ti ‘that’. For the distal function, the expanded form u.teyi ~ u.tei ~ u.ti is commonly used. In the case forms, other stem variants also occur: PROX a- ~ o- vs. DIST ta-, yielding the paradigms ACC ei-we ~ a-wa vs. ta-wa : DAT o-du vs. (u)ta-du : ABL o-digi vs. (u)tadigi : LOC o-lo vs. (u)ta-la : PROL o-li vs. (u)ta-li : INSTR o-ji vs. (u)ta-ji. The directive forms are irregular also as far as the case marker is concerned: eu-xi vs. (u)tau-xi. It is difficult to make a clearcut distinction between demonstrative pronouns and pronominal spatials, as the majority of the case forms are used as adverbs. The demonstrative pronouns lack the designative case. The plural forms are PROX ei-getu ‘these’ vs. DIST uti-getu ‘those’. The demonstratives have two main functions: deictic and anaphoric. In both functions they can occur as adnominal modifiers, e.g. ei bua-du ‘at this place’ (this place-DAT), uti lusa ‘that Russian’ (that Russian), but also as independent head nouns with case marking, e.g. ei-we=tene waa-jaƞa-fi ‘we will kill this (animal)’ (this-ACC=FOC kill-FUT-1PL. INCL), sii a-wa eji xule=gde ‘do not open it!’ (2SG that-ACC PROHIB open=FOC). The distal forms uti ~ uta : PL uti-getu are often used in the function of a third-person personal pronoun, e.g. bua-na-fei gala-kta-mi xuli-si-i-ti uti-getu ‘they travel looking for a land for themselves’ (land-DES-RX.PL seek-DISTR-CV.CONN travel-STAT-PRS-3PL that-PL). The demonstratives also occur in a number of stable forms and expressions, e.g. e=bede ‘like this’ vs. ute=bede ‘like that’, ei=tene ‘now’, ei julie-ni ‘up to now, so far’ (this frontPX3SG), o-digi cai-xi ‘from now on’ (this-ABL far-DIR). The basic interrogative pronouns are ye’u ~ ye’e ~ yi ‘what?’ and nii ‘who?’. The root ye- ~ ya- (< *ïa-) : ye’u ~ ye’e- (< *ïa-ko/n) has a more or less regular nominal paradigm: ACC ye(’e)-we ~ ya-wa : DES ye’e-ne/n- : DAT ye’e-du : ABL ye’e-digi : LOC ye’e-le : PROL ye’e-li, except that the directive form is of the spatial type ye’u-xi ~ yau-xi, cf. also

344 Elena Perekhvalskaya and Juha Janhunen

the derivative ya-ma ‘made of what?’. Examples: ye-we digana-i-ti jugdi doo-lo-ni ‘what are they saying inside the house?’ (what-ACC say-PRS-3PL house inside-LOC-PX3SG), ye-digi eme-he-hi ‘where have you come from?’ (what-ABL come-PST-2SG), suu ye’u-xi ƞene-jeƞe-hu ‘where will go?’ (2PL what-DIR go-FUT-2PL), sii dili-we-hi nii igde-he-ni? ‘who combed your hair?’ (2SG head-ACC-2SG who comb-PST-3SG). Most of the forms based on ye’u have semantic parallels based on the apparently etymologically separate root ii- (< *ii-), which yields the interrogative adverbs DAT ii-du ‘where? why?’ : ABL ii-digi ‘from where?’ : LOC ii-le ‘whither?’ : PROL ii-li ‘which way?’ : DIR ii-tigi ‘in which direction?’ : INSTR ii-ji ‘how?’, though the roots ye- ~ ya- and ii-, which are still separate in Oroch, may have become confused in Udihe. Additional interrogatives include adi ‘how many?’, ali ‘when?’ : DX ali-fi ‘how old?’, ono ‘how?’ : DX ono-fi ‘what sort of?’. When reduplicated, the interrogative pronouns have the universal meaning ‘all’, ‘all kinds of’, e.g. ye’u&ye’u sexi-ni, ye’u&ye’u tege-ni bi-si-ni ‘there were all kinds of fabrics, all kinds of clothes’ (what&what fabric-PX3SG, what&what clothes-PX3SG be-PST-3SG), ya-ma&ya-ma xaju-i jawa-i ‘they took all their belongings’ (what-DX&what-DX belonging-RX take-PRS). The cliticized form =ye’u is used as the marker of the general question, e.g. sii emu-se’e bi-hi=ye’u ‘do you live alone?’ (2SG one-DX be-PRS=INTERR). When repeated, =ye’u codes the alternative question (‘either—or’), e.g. ogbö[ö] xokto-ni bi-se[‘e]=ye’u, mafa xokto-ni bi-se[‘e]=ye’u ‘were these the tracks of an elk or the tracks of a bear?’ (elk road-PX3SG be-PRT=​INTERR, bear road-PX3SG be-PRT=INTERR). The plain interrogative root functions also as a verb: ya- ~ ye- ~ ii- ‘to do what’ (< ïa- ~ *ii-), e.g. ono ya-ja-fi ‘what shall we do?’ = ‘what will happen to us?’ (how do.what-OPT-1PL.INCL), which yields the lexicalized forms CV.CONN ya-mi ~ ye-mi ~ ii-mi ‘why?’ and AND-CV.CONN ya-na-mi ‘what for?’, e.g. ye-mi sii dili-i kapta-[h]i ‘why do you wrap up your head?’ (do.what-CV.CONN 2SG head-RX wrap-2SG), ya-na-mi eme-he-hi? [do.what-AND-CV.CONN come-PST-2SG] ‘what for have you come?’. Indefinite pronouns are formed from the interrogatives by adding the enclitic particles =kA(i) ~ =ke, =dA, =dAkA, =dAlA, or =gu, which may be followed by the adverb dehem ~ dem ‘no matter’, e.g. ye’u=ke ‘something’, yauxi=de ‘towards some place’, iimi=ke dehem ‘no matter for what reason’, nii=de ‘somebody’, ono=do ‘somehow’, etc. Examples: julie-le ye’u=ke uli-le patina-i-ni ‘something is flapping ahead of us in the water’ (front-LOC what=​INDEF water-LOC flap-PRS-3SG), ye-le=de uni-si-i-ni ‘it hurts somewhere’ (what-LOC=​INDEF hurt-STAT-PRS-3SG), nii=dele site-ni me-mi waa-ha-ni ‘someone’s son has killed himself’ (who=​INDEF child-PX3SG REFL-RX killPST-3SG), ono=gu ono=gu agda-gi-jeƞe-hi ‘you will somehow return (me) the favour’ (how=INDEF how=INDEF thank-REV-FUT-2SG). In combination with negation, the indefinite pronouns have a connegative function, e.g. nii=de e-si-ti alausi ‘nobody had taught (them)’ (who=INDEF NEG-PST-3PL teach), sama ali=da e-hi eme-gi ‘the shaman will never come back’ (shaman when=INDEF NEG-PRS come-REV). In some cases, indefinite pronouns can have a negative connotation with an affirmative verb, e.g. ye’uxi=de ƞene-kce-mi diƞ&diƞ bi-si-ni ‘going in whatever direction was impossible’ = ‘there was nowhere to go’ (what-DIR=INDEF go-PROSP-CV.CONN DESCR be-PST-3SG). They are also used in interrogative sentences, e.g. ono=ko ya-ja meisi-he-ni yegdige’e ‘what would she do, the hero thought’ (how= INDEF do.what-OPT think-PST-3SG hero). A special indefinite item is aƞi-, which substitutes a forgotten word of any content class, usually a noun or a verb. It acquires all grammatical markers of the word forgotten,

Udihe 345

e.g. aŋi-la ba’a-[h]a-mi bii ‘I found this however you call it’ (INDEF-LOC find-PST-1SG 2SG). PERSON MARKING Like Oroch, Udihe has almost totally lost the original distinction between possessive suffixes (PX) and predicative personal endings (VX). Even so, it is possible to distinguish between up to five different sets of person markers, including two of the possessive type and three specific to verbal forms depending on their morphological category. The actual differences concern the first person singular and plural forms in both the possessive and predicative sets, as conditioned by both the stem type and other factors. In addition, the third person forms remain unmarked in two of the predicative sets. There are also reflexive suffixes (RX), separate for singular and plural reference (Table 13.6). TABLE 13.6 UDIHE PERSON MARKERS     SG

PX  /n

I

1

-i

-mi

-mi

2

-hi

-hi

3

-ni

-ni

RX PL

VX  

V

1

-i

-mi

EXCL

-u

-mu

INCL

II

III -i

-Ø

-u

-fi

-fi

2

- hu

-hu

3

-ti

-ti

RX

-fi ~ -fei

-ti -Ø

Stem types: V = vowel stems, /n = nasal stems. I–III = categories of verbal forms taking different sets of person markers.

Finitely used verbal forms can take possessive suffixes of either the vowel-stem or the nasal-stem type, depending on the underlying form, but also the three predicative sets of person markers (I–III), each of which shows slight differences with regard to the others. In many cases it is difficult to determine which suffixes represent direct heritage from Proto-Tungusic, and which are results of analogical levelling. This is also because some of the verbal paradigms are of a mixed origin. Most importantly, although the aorist as a separate finite category has been lost in Udihe, the paradigm that has been identified as the “present indicative” contains both forms unmarked for tense (in the first and second persons) and forms based on the imperfective participle in -i- < *-RII (in the third person). The markers 1SG -mi : 2SG -hi in this paradigm are likely to represent regular reflexes of Proto-Tungusic *-m-bI : *-n-sI, containing the aorist marker *-n-, as also present in Ewenic. However, the analogical influence of the corresponding possessive suffixes cannot be ruled out. The plural forms in 1PL EXCL -u (< *-bO) : INCL -fi (< *-pU) : 2PL -hu (< *-sO) must, in any case, involve various types of analogical developments.

346 Elena Perekhvalskaya and Juha Janhunen

The third-person singular possessive suffix -ni is used as the head-marking expression of a possessive relationship in constructions involving two nouns, e.g. ba’ata ina’i-ni ‘the boy’s dog’ (boy dog-PX3SG). Apart from actual possession, this construction can indicate several other kinds of connection, cf. e.g. moo pa’i-ni ‘piece of wood’ (wood piece-PX3SG), ogböö imoo-ni ‘fat of an elk’ (elk fat-PX3SG), uli kää-ni ‘the bank of a river’ (river bank-PX3SG). Spatial nouns used as postpositions are organized in the same way, e.g. jugdi doo-ni ‘the space inside the house’ (house inside-PX3SG) : jugdi doo-digi-ni ‘from inside the house’ (house inside-ABL-PX3SG), jugdi doo-li-ni ‘through the inner space of the house’ (house inside-PROL-PX3SG). The same construction is also encountered in sequences in which a noun categorizes another in terms of species, gender, ethnicity, or other property, e.g. kuƞka moo-ni ‘cedar tree’ (cedar tree-PX3SG), ñau amina-ni ‘rooster’ (hen male-PX3SG), lusa a[h]anta-ni ‘Russian woman’ (Russian woman-PX3SG), suaƞka see-ni ‘Suanka kin’ (Suanka family), biki uli-ni ‘the Bikin river’ (Bikin river-PX3SG). Another function of the third-person singular possessive suffix -ni is to express definiteness. In this usage this suffix does not mark the head of a possessive construction. It may express both situational and contextual definiteness, e.g. ei moo-wo-ni kusige-ji eji tinda ‘do not cut these trees with a knife!’ (this tree-ACC-PX3SG knife-INSTR PROHIB cut), ƞää-la-ni=de juu nii=de ƞua-ha-ni ‘and two people slept on the bank’ (bank-LOC-PX3SG=PTCL two person=PTCL sleep-PST-3SG). Alienable possession is expressed by the suffix -ƞi-, which follows any derivational suffixes or plural marker and precedes the case markers and possessive suffixes (N-DXNX-AL-CX-PX/RX), e.g. giuse-jige-ƞi-we-ni ‘his roe deer’ (roe-PL-AL-ACC-PX3SG). Items classified as inalienable typically comprise body parts, relatives, domestic animals, clothes, and personal belongings, cf. e.g. bii site-i ‘my child’ (1SG child-PX1SG) vs. bii bä’äsa-ƞi-i ‘my river’ (1SG river-AL-PX1SG). OTHER WORD CLASSES Lexical items falling outside of mainstream nominals and verb(al)s include spatials, adverbs, postpositions, as well as some marginal types of expressive vocabulary, such as ideophones and interjections. Strictly speaking, many adverbs, most postpositions, as well as all spatials, are nouns, though their inflectional paradigm may involve deviations from that of regular nouns, or they are used only in a limited number of relict forms. Some of the principal spatials (spatial nouns) include amää- ‘back part’ > ‘behind, after’, julie- ‘front part’ > ‘in front, before’, bua- ‘space, environment, nature’ > ‘outside’, doo- ‘inside’, dasa’a- ‘vicinity’ > ‘near, close’, goo- ‘remote(ness)’ > ‘far away’, kää ‘side’ > ‘beside’, xegie- ‘underside’ > ‘below, down’, ui- ~ wei- ‘upside’ > ‘up’, as well as ejie- ‘downstream’, so’olie- ‘upstream’, bagää- ‘the other side (of the river)’ > ‘across the river’. Many of these can take the suffixalized spatial element -jA(-) ‘side’, yielding forms like amää-ja(-) ‘back side’, bagää-ja(-) ‘the other side’, bua-ja(-) ‘outside’, cf. also ta.u-je(-) ‘that side, opposite side’ (from DIST ta- ‘that’). The case paradigm of spatials follows mainly that of regular nouns, but some spatials take exceptionally the marker -fA in the accusative, -xi in the directive, and -ƞAji in the ablative, e.g. julie- ‘front’ : ACC julie-fe : DAT julie-du : ABL julie-ƞeji : LOC julie-le : PROL julie-li : DIR julie-xi : INSTR julie-ji. The complex ablative marker -ƞAji is also attested in the pronominal form ye’e-je-ƞeji ‘from where?’ (what-side-ABL). Spatials are normally used with possessive suffixes, e.g. хokto doo-ni-ni ‘on the road’ (road

Udihe 347

inside-DAT-PX3SG), namu ta’aƞki-le-ni ‘to the middle of the sea’ (sea middle-LOCPX3SG), bii amää-la-i ‘after me’ (1SG back.part- LOC-PX1SG). Non-spatial nominal postpositions include guu/n- ‘absence’, e.g. nexuse guundi-ni ‘in the brother’s absence’ (brother absence-DAT-PX3SG) and oli/n- *‘place’ > ‘instead’, e.g. ami-ni olin-di-ni ‘instead of the father’ (father-PX3SG place-DATPX3SG). There are also a few synchronically invariable postpositions, including amää.ta ‘after’ (based on the spatial amää- ‘back part’), cai.xi ‘further, next’ (containing the directive marker -xi), meƞde ‘with’ (for inanimates), as well as mule and geye ‘with’ (for animates). The case forms of spatials are used in the sentence as local or temporal adverbs. Many other adverbially used words are also nouns or nominal forms. The basic forms of nouns denoting parts of the day, succession of days, or seasons, are used as temporal adverbs with no case marking, e.g. sikie ‘(in the) evening’, tue ‘(in) winter’, tineƞi ‘yesterday’ (< DIST ta + ineƞi ‘day’). Other underived adverbs include egdi ‘many, much’, ñaƞga ‘a little’ : ñaƞga&ñaƞga ‘just a little’. Adverbs of manner are often instrumental forms of adjectival nominals, e.g. maƞga ‘strong’ : INSTR maƞga-ji ‘strongly’, kata ‘strong’ : INSTR kata-ji ‘strongly’ > ‘very’, or converbal forms of verbs, e.g. kesem < kese-mi ‘hardly, with difficulty’ (suffer-CV.CONN). Some adverbs of manner are derived from nominals and verbs with the help of special suffixes, including -lie, e.g. monto ‘circle’ : monto-lie ‘around’, and -ptigi, e.g. kumte- ‘to overturn’ : kumte-ptigi ‘upside down’, wani ‘wide’ : wani-ptigi ‘widely’, ogdo ‘side’ : ogdo-ptigi ‘from side to side’. Reduplication is also used, e.g. teu&teu ~ tu-tu ‘silently, quietly’, yaadam&​yaadam ‘randomly’. Other types of adverbs include (degree) kamdu ‘totally’, ketu ‘very, very much’, kie&kie ‘entirely’, lääsi ‘very, very much’, te[h]u ‘all, completely’; (modality) saina ‘probably’, bije ‘maybe’ (= OPT bi-je from bi- ‘to be’), (addition) ñaa ‘once more, again’, xai ~ xaisi ‘also’. Ideophones and onomatopoeia are items outside of regular lexicon. They often contain non-canonical phonotactic structures and reduplications, and they involve a high degree of irregular formal and semantic variation, cf. e.g. ata-ta-ta [laughter], bokcom&​bokcom [wrinkled], keät&keät [just in time], kiil ~ kiir [sound of a file], kitoom&kitoom [waddling], kotooro&kotooro [the sound of a red deer walking in water], ludu&ludu [a lot of idle chatter], pce&pce [sound of a harpoon being pulled out], xaƞna&xaƞna [completely dry]. Interjections have similar properties and are often pronounced with a specific intonation, e.g. ene ~ eneñeñe [it hurts!], ge↑ [starting a new topic], ñaxatuƞaxa [exclamation of discontent, vexation], ssüü↓ [sigh of tiredness after hard work]. VERBAL MORPHOLOGY As in Oroch, the loss of the aorist has led to a disruption of the original system of verbal stem types. Unlike Oroch, Udihe has also lost the connegative as a clearcut form. However, traces of the original stem types are still preserved in a number of morphological differences, which are especially noticeable in the formation of the past tense paradigm, but in some cases also of the present tense forms, which incorporate parts of the former aorist paradigm. This makes it possible to trace the evolution of the four historically distinct stem types: (1) vowel stems, (2) consonant stems, (3) verbs with the aorist stem in *-dA-, and (4) verbs with the aorist stem in *-sI-. Consonant stems may be further divided into (2a) non-nasal and (2b) nasal stems. In addition, vowel stems have developed a secondary difference between (1a) low-vowel stems, ending in the vowels a ä e o, and (1b) high-vowel stems, ending in the vowels u i, which show a difference in the formation of the finite preterite paradigm.

348 Elena Perekhvalskaya and Juha Janhunen

(1) Vowel stems form the present tense paradigm by adding the person markers (VX type I) directly to the plain stem, except in the third person singular and plural, which are based on the imperfective participle in -i. The plain stem also functions as a connegative, suggesting that the original aorist and connegative marker *-rA- has been fully lost either by an irregular phonetic development or by morphological analogy; a regular phonetic development would only have deleted the *r with a long vowel or vowel sequence as a result. The finite past tense paradigm, which is based on the perfective participle combined with person markers (PX for nasal stems), has the marker -hA/n- > Bikin -A/n-, which represents an irregular reflex of earlier *-xA/n (< *-ksA/n), as attested in Oroch and Nanaic. In the finite preterite paradigm, all vowel stems took originally the marker *-kA-, which, due to a regular phonetic split has yielded -’A- for (1a) low-vowel stems and -gA- for (1b) high-vowel stems, combined with person markers (VX type III). Sample paradigms: ƞene- ‘to go’ (low-vowel stem) : PRS 1SG ƞene-mi : 2SG ƞene-hi : 3SG ƞene-i-ni : 1PL EXCL ƞene-u : INCL ƞene-fi : 2PL ƞene-hu : 3PL ƞene-i-ti; umi- ‘to drink’ (high-vowel stem) : PST 1SG umi-he-mi : 2SG umi-he-hi : 3SG umih e-ni : 1PL EXCL umi-he-mu : INCL umi-he-fi : 2PL umi-he-hu : 3PL umi-he-ti; yexe- ‘to sing’ (low-vowel stem) : PRT 1SG yexe-’e-i : 2SG yexe-’e-hi : 3SG yexe-’e : 1PL EXCL yexe-’e-u : INCL yexe-’e-ti : 2PL yexe-’e-hu : 3PL yexe-’e; au- ‘to wash’ (high-vowel stem) : PRT 1SG au-ge-i : 2SG au-ge-hi : 3SG au-ge : 1PL EXCL au-ge-u : INCL au-ge-ti : 2PL au-ge-hu : 3PL au-ge. (2) Consonant stems form the present tense paradigm by a stem which actually contains the original aorist marker *-rA-, represented as -DA- (= -dA- ~ -tA-) after (2a) non-nasal stems and -A- (< */n-dA-) after (2b) nasal stems. From the synchronic point of view the stems exhibiting -DA- or -A- behave like vowel stems, taking person markers directly to the stem, except in the third person forms, which incorporate the participle marker -i. It may be noted that the presence of the participle marker (*-rII) after what is actually an aorist marker (*-rA-) is historically pleonastic, for which reason the complexes SG3 -i-ni : PL3 -i-ti have in some descriptions of Udihe been analysed as indivisible personal markers, i.e. SG3 -ini : PL3 -iti. The finite past tense paradigm and the corresponding perfective participle of consonant stems have the marker (*)-ki/n- (unrelated to the marker -hA/n- of vowel stems), while the finite preterite paradigm has the marker -kA- (etymologically identical with the markers -’A- and -gA- of vowel stems). In the modern language, consonant stems tend to be transformed to vowel stems based on the original aorist stem. Sample paradigms: nagda- ‘to hit (the aim)’ (non-nasal consonant stem) : PRS SG1 nagda-mi : SG2 nagda-hi : SG3 nagda-i-ni, etc., PST SG1 nak-ki-mi, etc., PRT nak-ka-i, etc.; digan- ‘to speak’ (nasal stem) : PRS 1SG digana-mi : 2SG digana-hi : 3SG digana-i-ni, etc., PST 3SG digaƞ-ki-ni ~ digana-ha-ni, etc., PRT 3SG digaƞ-ka ~ digana-’a, etc. (3) Vowel stems originally taking the aorist marker -dA- remain only as a relict class and tend to be transformed to regular vowel stems, e.g. AOR *ga-da- > gada- ‘to take’, though the past tense paradigm of some verbs can still reflect the original pattern, e.g. bude- ‘to die’ : PST 3SG buk-ki-ni ~ bude-he-ni, o- ‘to become’ : PRS odo- : PST o-si- (< *o-cï-). (4) The aorist stem in *-sI- > -hi- is regularly retained by the two auxiliaries COP/EXIST bi- ‘to be’ and NEG e- (negation verb), which form their present tense paradigms from the stems bi-hi- and e-hi-, respectively, though some of the personal forms are

Udihe 349

simplified due to haplology or vowel contraction, cf. e.g. bi- ‘to be’ : PRS 1SG bi-hi-mi : 2SG bi-hi-hi > bi-hi : 3SG bi-hi-ni : 1PL EXCL bi-hi-u > bi-hu : INCL bi-hi-fi : 2 PL bi-hi-hu > bi-hu : 3PL bi-hi-ti. The past tense is formed by the marker -si- (< *-ci-), e.g. NEG e- : PST 1SG e-si-mi : 2SG e-si-hi : 3SG e-si-ni, etc., while the preterite paradigm is marked by -se’e- (< *-cA-kA-), e.g. PRT 1SG bi-se’e-i : 2SG bi-se’e-hi : 3SG bi-se’e, etc. Note also PTCP.PRS bi.i ~ bi.e. VOICE AND ASPECT The criterion for dividing verbal suffixes into derivational and inflectional is their behaviour in the negative construction: only inflectional suffixes are attached to the negation verb. The principal grammatical categories of the verb expressed by derivational means are voice and aspect. The category of voice involves valency and argument-structure-changing operations and comprises the productive passive, causative, medial (decausative), and reciprocal forms, as well as the non-productive cooperative form. •

• •

•

•

The passive voice is formed by the suffix -u-, with the agent in the dative case, e.g. sina-i teti-gi-kce-mi diƞe-u-he-ni sina-di-i ‘while trying to put on his rucksack, he was crushed down by it’ (rucksack-RX put.on-REV-PROSP-CV.CONN press-PASSPST-SG3 rucksack-DAT-RX), dukte-u-je-hu min-du ‘you will be beaten by me’ (beatPASS-OPT-2PL 1SG-DAT). The passive has often a necessitative connotation. Causatives are marked by the suffix -wA/n- ~ -wAnA-, e.g. edigin-di-i eji ise-wene cu’ai-we ‘do not show the bird cherries to your grandfather!’ (grandfather-DAT-RX PROHIB see-CAUS bird.cherry-ACC). Medials are marked by the primary suffix -p-, which in the present tense incorporates the aorist marker *-tA- (< *-rA-) and in the past tense a metathesized reflex of the perfective participle marker *-ki/n, yielding synchronicallyl PRS -p-tA- : PST -kpi/n-, e.g. eniƞe-ni moo-ƞi-ni tehu mala-kpi-ni ‘mother’s firewood is all used up’ (mother-PX3SG wood-AL-PX3SG all spend-MED.PST-3SG). Reciprocals (mutual action) are formed by the suffix -mAsi- (< *-mAAcI-), e.g. utemi seƞgite-we oo-li-he-ti, gaa-masi-he-ti ‘therefore, they became in-laws and took (wives) from each other’ (then in.law-ACC become-INCH-PST-3PL take-RECIPR-PST-3PL). Cooperatives (joint action) are formed by the suffix -gdi- (< *-ldU-), present only in a few stems with lexicalized meanings, e.g. ba’a- ‘to find’ : COOP ba’a-gdi- ‘to meet’, diga/n- ‘to speak’ : COOP diga-gdi- ‘to conspire’, jawa- ‘to take’ : COOP jawa-gdi- ‘to copulate’.

Aspectual forms in Udihe express differences of linear aspectuality and comprise the categories here identified as prospectives, inchoatives, duratives, semelfactives, singulatives, multidirectionals, distributives, and reversives. Prospectives mark the stage of preparation for an action, inchoatives the starting point, and duratives an on-going action. Semelfactives and singulatives describe the action as having no duration. Multidirectionals (diversatives) express an action taking place in different directions, while distributives refer to an action performed by all members of a group one after the other. Reversives (reversive-repetitives or refactives) imply that a previous action is performed again. •

Prospectives are formed by the suffix -kcA-, e.g. ili-gi-kce-he-ni kanda’a mafa tääk&​ tääk=da lagbaƞ-ki-ni ‘the old man tried to get up but he stuck tightly’ (stand-REVPROSP-PST-3SG old man DESCR=FOC stick-PST-3SG).

350 Elena Perekhvalskaya and Juha Janhunen

• • • • • • •

Inchoatives are formed by the suffix -li- (< *-lU-), e.g. neƞu-ni moo-we gada-li-he-ni ‘the younger sister started collecting firewood’ (younger.sibling-PX3SG wood-ACC take-INCH-PST-3SG). Duratives are formed by the suffix -si- (< *-cI-), e.g. maƞga goo-ji baatuga-si-ga-u ‘we went hunting for a very long time’ (strong distant-INSTR hunt-DUR-PRT-1PL.EXCL). Semelfactives are formed by the suffix -ndA-, e.g. ƞala-ji-i eke bagba-nda-ha-ni saƞa cul-de ‘he gently poked his hand right into the hole’ (hand-INSTR-RX gently poke-SEM-PST-3SG hole straight=FOC). Singulatives are formed by the suffix -lA-, e.g. selemege tiƞme-le-he-ni, lupam=da ‘Selemege fell (on the knife)’ (Selemege fall-SING-PST-3SG DESCR=FOC). Multidirectionals are formed by the suffix -wAsi- (< *-wAcI-), e.g. wui-xi eteƞi-he-ni: yegdige’e mama’asa-ni die-wesi-ni ‘he looked up: the hero’s wife was flying back and forth’ (upside-DIR look-PST-3SG hero wife-PX3SG fly-MULTID-3SG). Distributives are formed by the suffix -ktA-, e.g. dogbo-gi-he-ni ƞua-gi-kte-li-he-ti ‘the night came and they started falling asleep one by one’ (night.come-REV-PST3SG sleep-REV-DISTR-INCH-PST-3PL). Reversives are formed by the suffix -gi- (~ -ƞi-) (< *-rgU-), e.g. tege-i=le ayakta-wa-ni=de teti-gi-he-ni ‘she put on her best clothes (once again)’ (clothes-RX=PTCL good.thing-ACC-PX3SG=​PTCL put.on-REV-PST-3SG), cf. a’i-mi tege-we-ni teti-he-ni ‘he put on his elder brother’s clothes (for the first time)’ (elder.brother-RX clothes-ACC-PX3SG put.on-PST-3SG). The reversive suffix is also used with a number of nomina-verba denoting chronological cycles, e.g. dogbo(-) ‘night’ : REV dogbo-gi- ‘to come (of night)’, neki(-) ‘spring’ : REV neki-gi- ‘to come (of spring)’.

PARTICIPLES Udihe retains the three basic participles present also in Oroch: the imperfective participle marked by the suffix -i (< *-rII) for all stem types, the perfective participle marked by the suffix -hA/n (< *-xA/n) for vowel stems and (*)-ki/n for consonant stems, and the futuritive participle marked by the suffix-jAƞA (< *-jA-ƞA) for all stem types. Exceptionally, however, the two auxiliaries COP/EXIST bi- and NEG e- take the marker -si/n (< *-ci/n) in the perfective participle. The imperfective and futuritive participles have passive (impersonal) counterparts, containing the passive marker -u- and yielding the complex suffixes -u-yi and -u-jAƞA, respectively, (Table 13.7).

TABLE 13.7 UDIHE PARTICIPLES  

 

V

/n

C

ACT

IMPRF

-i

-A.i

-DA.i

-ki/n

-Ki/n

PASS

PRF

- A/n

FUT

-jAƞA

IMPRF

-u-yi

FUT

-u-jAƞA

h

Stem types: V = vowel stems, /n = nasal stems, C = non-nasal consonant stems. Segmental alternations: A = a e o, D = d t, K = k p.

Udihe 351

The three basic participles are also used in finite temporal functions, corresponding to the present, past, and future tenses. In addition, there are three deverbal nominal forms outside of the tense system which functionally resemble impersonal participles, but with aspectual connotations. These forms comprise the resultative in -sA, the completive in -ktu, and the designative in -kci. Sample paradigms: etete- ‘to work’ (low-vowel stem) : PTCP IMPRF etete-i ‘working’ : PRF etete-he/n ‘(one) who has worked’ : FUT etete-jeƞe ‘(one) who will work’; gu/n- ‘to say’ (nasal stem) : PTCP IMPRF gune-i ‘saying’ : PRF guƞ-ki/n ‘(one) who has said’ : FUT gun-jeƞe ‘(one) who will say’; uli- ‘to sew’ (high-vowel stem) : PTCP PASS IMPRF uli-u-yi ‘that is being sewn’ : FUT uli-u-jeƞe ‘that will be sewn’ : PTCP.RES uli-se ‘that has been sewn’ : PTCP.COMPL uli-ktu ‘that has already been sewn’ : PTCP.DES uli-kci ‘that is meant to be sewn’. The non-finite functions of the participles include their use as adnominal modifiers, independent head nouns, and sentential arguments. As adnominal modifiers, participles take no additional marking and precede their nominal headword, e.g. yaƞse-i nii bukta yaa-ni bi-hi-ni ‘the steering man had bad eyes’ (steer-PTCP.IMPRF person broken eyePX3SG be-PRS-3SG), taluna-i nii-we gakpa-ha-ti ‘they shot a man who was spearing (fish)’ (spear-PTCP.​IMPRF person-ACC shoot-PST-3PL). In independent use, participles can function both as actor nouns, in which role they can take plural marking, e.g. ei bua-du eke-jeƞe-getu gogdi&gogdi ƞua-ha-ti ‘those who were going to stay in this place were sleeping tightly together’ (this place-DAT stay-PTCP.FUT-PL DESCR sleepPST-3PL), and as action nouns (“nominalizations”), in which role they can take marking for case and person, e.g. ba’ata ami-ni ƞene-jeƞe-ji-ni ƞele-i-ni ‘the boy is afraid that his father would go away’ (boy father-PX3SG go-PTCP.FUT-INSTR-PX3SG be.afraidPRS-3SG). As sentential arguments with marking for case and person, participles function as quasiconverbs, e.g. ali eme-hen-ji ƞua-mie=de bi-hi-hi ‘when did you come as you are already asleep?’ (when come-PTCP.PRF-INSTR sleep-CV.CONN=PTCL be-PRS2SG). The instrumental form of the perfective participle can also function as an adverbial argument, e.g. mine-we jeä-han-ji gaji-ya ‘bring me in secretly’ (1SG-ACC steal-PTCP. PRF-INSTR bring-IMP). The resultative, completive, and designative forms are taxonomically difficult to place in the system of verbal morphology. As they lack an overt passive marker, their passive function must be diachronically secondary. The completive and designative, in particular, resemble deverbal nominal derivatives, though they are not attested in adnominal position and they seem to lack a nominal paradigm, unless lexicalized as full nouns. The completive forms are based on telic verbs and denote a result of a previous action, e.g. tuu kapta-ktu dili-ni ‘his head is all in bandages’ (all bind-PTCP.COMPL head-PX3SG), a’ana-ti campi-ktu bi-si-ni ‘their boat was broken’ (boat-PX3PL break-PASS.RES be-PST-3SG). The designative forms denote an action for which an object is designated, e.g. komnata gegbi-ni: au-si-kci, tatu-si-kci, diga-si-kci, ƞua-si-kci ‘the names of the rooms are: washroom, class room, eating room, and sleeping room’ (room name-PX3SG: wash-STATPTCP.DES, learn-STAT-PTCP.DES, eat-STAT-PTCP.DES, sleep-STAT-PTCP.DES). The resultative marker -sA, which has parallels in Oroch and Nanaic, is a reflex of the perfective participle in *-cAA, as attested in Ewenic in both participial and finite functions with no passive connotation. This may be the reason why this form can facultatively be combined with an accusative object, e.g. ule jii-se ‘the meat is cut’ (meat cut-PTCP.

352 Elena Perekhvalskaya and Juha Janhunen

RES) ~ ule-we jii-se (meat-ACC cut-PTPC.RES). Unlike the completive and designative forms, the resultative form can also, though apparently rarely, have an agent in the dative case, e.g. ei oloxi-we ba’ata-du jawa-sa ‘this squirrel was caught by the boy’ (this squirrel-ACC boy-DAT take-PTCP.RES). As the resultative is always attested in the function of a finite predicate, it could perhaps synchronically also be classified as a finite form. TENSE AND MOOD In finite use, the three basic participles function as the three tense forms of what may be identified as the unmarked or indicative mood. The present tense is formally exceptional, in that it incorporates the participial marker only in the third person form, while the first and second person forms are synchronically based on the plain stem, though diachronically they incorporate traces of the original aorist paradigm. The preterite (also called the “perfect”) is not attested in participial functions and is only used as a finite form. Note that all the four finitely used forms take different sets of person markers: the past and future tenses use possessive suffixes of either the vowel-stem type (for the futuritive participle) or the nasal-stem type (for the perfective participle), while the present tense and preterite paradigms have in some respects idiosyncratic sets (VX type I for the present tense and VX type III for the preterite), though the differences concern only a few forms. The functional difference between the participial past tense and the finite preterite involves a number of temporal, aspectual, modal, and evidential factors. The preterite codes, (1) perfective meaning, e.g. bii uya-ge-i ‘I (have) missed’ (1SG miss-PRT-SG1), (2) former experience, e.g. sii ali=da a’i-wa umi-ge-[h]i ‘have you ever drunk liquor?’ (2SG when=​INDEF liquor-ACC drink-PRT-2SG), (3) an avertive or unsuccessful attempt, e.g. nua-ni ketige tiƞmele-’e ‘he almost fell’ (3P-PX3SG almost fall-PRT), (4) admirativity or “hot news”, e.g. lääsi patina-li-ge dawa ‘the chum salmon flapped intensively’ (intensively flap-INCH-PRT chum.salmon). Most importantly, the preterite expresses (5) direct knowledge witnessed by the speaker, while the past tense based on the perfective participle transfers reportative evidentiality and is therefore widely used in narratives. Cf. e.g. [2], in which the verb gun- ‘to say, to tell’ appears in the past tense form in the narrative part and in the preterite form in the direct quotation. [2] [sine-we bii mafa-i [cinda’a sile-we-ni umi-ne-ye] sg2-acc 1sg husband-px1sg bird soup-acc-px3sg drink-and-imp ‘“Come and eat a bird soup! my husband guƞ-ke] guƞ-ki-ni say-prt say-pst-3sg told you,” she said.’ Apart from the indicative, there are a few other modal categories expressed by forms with no temporal oppositions. These comprise the following: •

Imperative. The imperative, directed at the second person, is formed by the suffix -yA : PL -yA-hu, e.g. buu- ‘to give’ : IMP SG bu-ye ‘give!’ : 2PL bu-ye-hu. The imperative can be optionally accompanied by the “hortative” clitic =jA, e.g. bagdi-i-le-ni ba’a-han-tigi ƞen[e]-ye=je ‘go back there where you have found a family!’ (livePTCP.PRS-LOC-PX3SG find-PTCP.PRF-DIR go-IMP=HORT).

Udihe 353

•

•

•

•

•

•

Optative. The form here termed “optative” (but which has also been called “subjunctive”) is based on the original progressive marker *-jA-, which also functions as a future tense marker (the “first future tense”) in Ewenic. Since this is originally a derivational form, it has a full personal paradigm, but with a specific set of person markers (VX type II), with no marking in the third person singular and plural. The optative expresses a wish directed at the subject person, normally with a future tense reference, e.g. dili-we-hi kumedi-je-mi ‘let me examine your head for lice’ (headACC-PX2SG search.lice-OPT-1SG). A lexicalized optative form is bi-je ‘maybe’ (be-OPT). Permissive. The permissive is marked by the suffix -tA- combined with person markers of the same type as used in the optative (VX type II), but with the obligatory addition of the final “hortative” particle =jA, e.g. gun- ‘to say, to tell’ : PERM 1SG gun-te-m[i]=je ‘I’d rather tell’ : 2SG gun-te-hi=je ‘you’d rather tell’ : 3SG gunte=je ‘he’d rather tell’, etc. The permissive expresses, very much like the optative, a wish directed at the subject person, and it can function as a soft imperative, e.g. min-tigi timanaƞi jima-ta-hi-ja ‘come to visit me tomorrow!’ (1SG-DIR tomorrow visit-PERM-2SG=HORT). Desideratives are formed by the derivational suffix -mu- : -mu[h]i- (< *mO-sI-), which diachronically contains the aorist marker *-sI-, e.g. umi- ‘to drink’ : umi-mui‘to want to drink’, ƞene- ‘to go’ : DESID ƞene-mui- ‘to want to go’. The same suffix, in the nominalized aorist form -mu[h]i (< *-mO-sI) is used to derive adjectival meanings, e.g. je/p- ‘to eat’ : je-mui ‘hungry’, kogo- ‘to be thirsty’ : kogo-mui ‘thirsty’. A related suffix is -mu.gdi, which has been termed the “desiderative noun”, but which actually lacks a nominal paradigm and is normally used in the predicative position, e.g. ƞua- ‘to sleep’ : ƞua-mugdi ‘one who wants to sleep’, as in nua-ni ƞua-mugdi ‘he wants to sleep’ (3P-PX3SG sleep-DX.DESID). Subjunctive. The form here termed “subjunctive” (but which has also been called “conditional”) is formed by the suffix -musA- (< *-mO-cAA), combined with person markers of the same type as used in the preterite (VX type III), e.g. buu- ‘to give’ : SUBJ 1SG bu-muse-i ‘I would give’. This form is typically used in a main clause in combination with the finitely used PTCP.PRF = PST bi-si of bi- ‘to be’ in the subordinate clause, e.g. bii sun-ji geye ƞene-muse-i, ugda bi-si ‘I would have come with you, if there had been a boat’ (1SG 2PL-INSTR together go-SUBJ-1SG boat be-PST). A special tense-related construction, outside of regular morphology, is the “expressive past”, coded by the clitic =k ~ =t, attached to present or past stems. This construction, which also violates the rules of regular phonotactics, expresses a sudden and intensive action or a series of actions in the past, e.g. yegdige’e=le neme-i kop wende-me=t ‘the hero immediately threw away the cover’ (hero=TOP cover-RX DESCR throw-INTENS=EXPR), tee-gi-[h]e-ni, tee-gi-me=t, gida-i jawa-gi-[h]e=t ‘he jumped up and seized his spear’ (sit-REV-PST-3SG sit-REV-INTENS=​EXPR spear-RX take-REV-PST=EXPR). This form is often accompanied by the verb oo‘to do’, e.g. olokto-ho-ti, jepte=t oo-ho-ti ‘they cooked (food) and ate’ (cook-PST-3PL, eat=EXPR do-PST-3PL), jugdi-tigi ƞen[e]-[h]e=k oo-ho-ni ‘she went immediately home’ (house-DIR go-PST=EXPR do-PST-3SG). Another exceptional feature connected with the finite use of certain temporal and modal forms is the so-called “plural” or pluritative in -du-. This element, which should not be mistaken for a person marker, is attested in reference to a plural subject

354 Elena Perekhvalskaya and Juha Janhunen

before the markers of preterite tense, as well as of the optative, permissive, and subjunctive moods, e.g. umi- ‘to drink’ : PLUR-PRT umi-du-ge ‘they have drunk’, etete- ‘to work’ : PLUR-OPT etete-du-je ‘let them work!’, wakca- ‘to hunt’ : PLURSUBJ wakca-du-muse ‘they would hunt’. The position before the temporal or modal marker suggests that -du- is originally a derivational suffix implying multiple actors. However, in the Bikin dialect, for low-vowel stems, -du is reported to be added after the preterite marker, which may also be lost, e.g. yexe- ‘to sing’ : PRT yexe-’e ‘(he) sang’ : PRT-PLUR yexe-’e-du ~ yexe-du ‘they sang’. CONVERBS Compared with Oroch, the Udihe system of converbs is more innovative, a circumstance that illustrates the instability and rapid evolution of converbal morphology. The system comprises four forms with parallels elsewhere in Tungusic: the conjunct connective and simultaneous converbs with no synchronic person marking, and the ambivalent contemporal and purposive converbs with either possessive or reflexive marking, as well as four complex forms specific to Udihe and functioning as conjunctive and conditional converbs, with both conjunct and disjunct reference (Table 13.8). TABLE 13.8 UDIHE CONVERB MARKERS  

 

PL

CONN

-mi

-mei ~ -mu

PX

SIMULT

-nA

CTEMP

-ƞAhA/n- ~ -ƞie/n-

+

+

+

+

PURP

-lAgA-

+

+

+

+

CONJ

PRS-si-

+

DS

+

+ +

+

-li.si-li

SS +

PRT-si COND

RX

-lie

+ +

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e o.

•

The connective converb (which in Udihe studies has also been called “gerund” or “infinitive”) is marked by the suffix -mi, with cognates in all Tungusic languages. It also has the plural form in -mu ~ -mei (< *-mArI), though the latter is rarely used in the modern language. This converb can be used in the position of a subject or object, e.g. damisi-mi gehe ‘smoking is bad’ (smoke-CV.CONN bad), moolo-mi wadi-he-ti ‘they finished chopping’ (chop-CV.CONN finish-PST-3PL). It may also form an adverbial clause with the subject coreferential with the subject of the matrix clause, normally expressing simultaneous or anterior action, e.g. ñuu-mi tiƞme-le-he-ni ‘coming out, he fell down’ (come.out-CV.CONN fallSING-PST-3SG), uti tada-ji waa-mei bi-li-he-ti ‘they began to live hunting with those arrows’ (that arrow-INSTR kill-CV.CONN.PL be-INCH-PST-3PL). When repeated, it points to an action that has lasted for some time, e.g. tee-mi tee-mi loƞko-i culi ñuxaƞ-ki-ni ‘(she) was sitting for some time and then suddenly

Udihe 355

•

•

•

•

•

dived into the cauldron’ (sit-CV.CONN sit-CV-CONN cauldron-RX suddenly dive-PST-3SG). The simultaneous converb in -nA, although absent in Oroch, is apparently inherited from Proto-Tungusic and is probably related to the Ewenic converb of simultaneous action in (*)-nA, though in Udihe it can also refer to an action which shortly precedes the action of the main clause. This form is attested only with a few stems and is normally accompanied by a verb of movement in the main clause, e.g. sugese xebu-ne ƞene-he-ni ‘taking an axe she went out’ (axe carry-CV.SIMULT go-PST-3SG). The contemporal (“imperfective past”) converb is marked by the suffix-ƞAhA/n- ~ (Bikin) -ƞie/n- (< *-ƞAsA-). This form can take both possessive marking (in disjunct use) and reflexive marking (in conjunct use) and denotes an action which took place in the past—often in the remote past—simultaneously with another action, e.g. anana bii sike bi-ƞehe-i sugalada-[h]a-mi bi-se’e ‘long ago, when I was young, I used to go skiing’ (long.ago 1SG young be-CV.CTEMP-RX ski-PST-1SG be-PRF), bii bei-xi ƞene-ƞie-i nua-ni yexe-i bi-se’e ‘when I was going out he was singing’ (1SG outside-DIR go-CV.CTEMP-PX1SG 3P-PX3SG sing-PTCP.PRS be-PRT). The purposive converb (“supine”) is marked by the suffix -lAgA-, which can also be combined with both possessive marking (in disjunct use) and reflexive marking (in conjunct use). Like its Oroch cognate -(A)lA(k)A-, this form denotes the purpose or target of an action, e.g. yoxo-we olokto-ma-ya diga-laga-fei ‘put the kettle on (fire) so that we can eat!’ (kettle-ACC cook-INTENS-IMP eat-CV.PURP-RX.PL), sii teluƞusi-ye nua-ti sa-laga-ti ‘tell them, so that they would know!’ (2SG tell-IMP 3P-PX3PL know-CV.PURP-PX3PL). The form here termed “conjunctive” (also called “perfective”) converb has a specific structure, not typical of converbs, in that it involves a marker -si(-) (< *-ci) that is attached to finitely used tense forms. For disjunct use, with possessive suffixes, -siis attached to the finite present tense form, based on the present participle, while for conjunct use, with no person marking, it follows the finite preterite form of the verb. In both constructions, the resulting complex form most often expresses temporal anteriority, e.g. (disjunct) ute-bede nixe-i-si-ni eme-gi-he-ni ‘after he had done like that, (she) came back’ (that-like do-PRS-CONJ-PX3SG come-REV-PST-3SG), (conjunct) bii sagdi ede-’e-si wakca-jaƞa-i ‘after I have become big, I will hunt’ (1SG big become-PRT-CONJ hunt-FUT-1SG). Occasionally, the implication may also be one of simultaneity (‘when’) or causality (‘because’). The complex structure of these forms leaves the actual status of the element -si(-) open. Since it is not attached directly to the verbal stem, it cannot be a converb marker in its own right. The fact that it can take person marking suggests, however, that it may have a verbal origin and could perhaps be a trace of a periphrastic construction. The conditional converb also has separate forms for conjunct and disjunct use. The plain suffix is -li ~ -lii = -li-i, used for singular conjunct reference, with -lie as the corresponding plural form. These forms apparently contain traces of reflexive marking for singular and plural, respectively. In disjunct use, with possessive suffixes, the suffix has the form -li.si-, with the final .si- identical with the element present in the conjunctive converbs. The conditional converb expresses general conditionality (‘if’), which may also involve temporality (‘when’), e.g. (conjunct) bii xoton-digi ƞene-li-i sin-du kompeta-wa gada-jaƞa-i ‘if I go to the city, I will buy you some candy’ (1SG city-DIR go-CV.COND-RX 2SG-DAT candy-ACC take-FUT-1SG), (disjunct) bii xuñaji-i gehe bi-lisi-n[i] eme-wene-te-hi=je ‘if my sister behaves badly

356 Elena Perekhvalskaya and Juha Janhunen

you should make her come (back home)’ (1SG sister-PX1SG bad be-CV.CONDPX3SG come-CAUS-PERM-2SG=HORT). COMPLEX PREDICATES Complex predicates involving analytical constructions are formed with the two principal auxiliaries COP/EXIST bi- ‘to be’ and NEG e- (negation verb). The constructions with bi- yield several types of grammaticalized temporal, aspectual, and modal meanings: • •

•

•

•

The combination of the finite preterite form PRT bi-se’e with the perfective participle of the preceding verb expresses the “pluperfect”, e.g. bii anana o-ho-mi bi-se’e ‘I (had) made it a long time ago’ (1SG long.ago do-PTCP.PRF-1SG be-PRT). The combination of the finitely used participle PTCP.PRF = PST bi-si with a preceding participial predicate expresses counterfactual conditionality (irrealis mood), aya bi-muse bii xunaji-i yexe-i bi-si ‘it would have been good if my sister had sung’ (good be-SUBJ 1SG sister sing-PTCP.PRS be-PST). A similar construction is also possible with a nominal predicate, e.g. eke-xi bi-si bii eme-muse-i ‘if I had had time, I would have come’ (time-PROPR be-PST 1SG come-SUBJ-1SG). Habitual action (habitive aspect) is expressed by the combination of the imperfective participle of the main verb with the various forms of bi-, e.g. anana ami-ƞa-hi ebede kepte-i bi-se’e ‘in the past, your late father always used to lie like this’ (long.ago father-DX-PX2SG thus lie-PTCP.PRS be-PRT). Progressive action (aspect) is expressed by the combination of the connective converb of the main verb with the various forms of bi-, e.g. ajiga-jiga soƞo-mi bi-si-ti ‘the girls were crying’ (girl-PL cry-CV.CONN be-PST-3PL), bii sine-we galakta-gi-mi nada-ma aƞa-ni bi-hi-mi ‘I have been looking for you for seven years’ (1SG 2SG-ACC search-REV-CV.CONN seven-ACC year-PX3SG be-PRS-1SG). A necessitative function is expressed by combining bi- with the futuritive participle of the preceding verb, e.g. etete-jeƞe bi-si-u ‘we had to work’ (work-PTCP.FUT be-PST-1PL.EXCL), also impersonally with the passive marker in the first verb, e.g. uli-u-jeƞe bi-si-ni ‘it had to be sewn’ (sew-PASS-PTCP.FUT be-PST-3SG).

The negation verb e- retains its original verbal character and forms a regular verbal paradigm, with the active participles PRS e-hi- : PRF e-si/n- : FUT e-jeƞe-, the passive participles PASS-PRS e-u-yi : PASS-FUT e-u-jeƞe, the finite preterite PRT e-se’e, the modal forms OPT e-je- : SUBJ e-muse-, and the converbs CTEMP e-ƞehe/n- : PURP e-lege- : PRS-CONJ e-i-si- : COND e-si : e-lisi-. Morphologically irregular and/or idiosyncratic are the modal forms PROHIB e-ji : DESID a-ta-, the converb CV.CONN e-hi-mi, and the participle PTCP.REM e-ptile (the participle of “remote past”, a form otherwise not attested in Orochic, but with parallels in Ewenic). After all these forms, the semantic main verb (content verb), including any derivational elements it may contain, occurs in the basic form, or, in the case of consonant stems, in a vowel-stem form, which synchronically functions as an unmarked connegative, e.g. e-i-ni tigde ‘it is not raining’ (NEG-PRS-3SG rain), eji soƞo, e-jeƞe-u waa ‘don’t be afraid, we will not kill you’ (PROHIB cry, NEG-FUT-1PL.EXCL kill), mafa määwa-ni ahanta-du eji-hu je-wene ‘do not let women eat a bear’s heart!’ (bear heart-PX3SG woman-DAT PROHIB-2PL eat-CAUS), a-ta-hi nodo-pto ‘let you not get lost!’ (NEG-DESID-2SG lose-MED), e-li buu mine-we gehe eji gune ‘if you do not give (it), do not say that I am bad!’ (NEG-CV.COND give

Udihe 357

1SG-ACC bad PROHIB say), emugde-i e-lege-ni uni ‘(this is) to prevent stomach ache’ (stomach-RX NEG-CV.PURP-PX3SG ache). The negation of existence is coded by the invariable privative noun anci ‘there is not’, which is combined with the copula bi-, and which is also used in the privative construction, e.g. a[h]anta anci bi-i-si-ni goo mama’asa-la anci bi-[h]i-ni ‘when there is no woman, he lives a long time without a wife’ (woman PRIV be-PRS-CONJ-PX3SG distant woman-PART PRIV be-PRS-3SG). SYNTAX Udihe shows the main features typical of a left-branching SOV language: nominal phrases, postpositional phrases, and non-finite clauses are head-final, the modifier precedes the head noun, and dependent clauses are embedded before the matrix verb, cf. e.g. a[h]anta ba’ata-ma site-we ba’a-[h]a-ni ‘the woman gave birth to a boy’ (woman male-DX childACC find-PST-3SG). In the matrix clause the word order is largely motivated by the information structure. The inverse word order OVS is most often observed when the verb is in a focus position, e.g. ei=tene mafa-la anci osi-gi-ho-mi bii ‘now I am left without a husband’ (this=PTCL husband-PART PRIV become-REV-PST-1SG 1SG), uma’a-i ise-ne-gi-[h]a-ni yegdige’e ‘the hero went to check his fishing hooks’ (fishing.hook-RX see-AND-REV-PST-3SG hero). Verb finality is violated usually under the influence of Russian. However, the direct object may be postposed after the verb if an element other than the direct object stands in focus. In the following a few additional issues concerning sentence structure are discussed in more detail. •

•

•

•

Possessive groups are always head-final, with possession expressed suffixally in the head noun, e.g. ba’ata jugdi-ni ‘the boy’s house’ (boy house-PX3SG). The modifier may also comprise two nominals, e.g. ba’ata tie ko’olo-wo-i jawa-gi-he-ni ‘the boy took back his pair of mittens’ (boy pair mitten-ACC-RX take-REV-PST-3SG). In the nominal phrase, there is normally no agreement between the modifier and the head noun. An exception is made by sequences of a numeral and a noun, in which, in object position, the accusative marker is often added to both the numeral and the noun, e.g. ila-ma sina-wa oo-ho-ti ‘they made three backpacks’ (threeACC backpack-ACC make-PST-3PL). In sequences with other quantifiers, the quantifier tends to be placed after the noun, with both taking accusative marking, e.g. sugjä-wa egdi-we wa-ha-ni ‘he caught a lot of fish’ (fish-ACC many-ACC kill-PST-3SG). Personal pronouns in subject position are regularly omitted (“pro-drop”), cf. e.g. selemege-tigi wuhe-nde-he-ni ‘(he) threw (it) to Selemege’ (Seleme-DIR throwSEM-PST-3SG)—ui-nu=pehi=de tewe-ƞi-he-ni ‘(Selemege) caught (it) in the air’ (top-DAT=FOC=​EMPH catch-REV-PST-3SG). The predicates of dependent clauses are expressed by non-finite verbal forms, that is, converbs and case-marked participles. Usually, these forms indicate whether their subject is the same with (conjunct) or different from (disjunct) the subject of the main clause. This indication can be either inherent in the form (as in some converbs) or it is expressed by the distinction between possessive (for disjunct use) and reflexive (for conjunct use) suffixes, cf. e.g. (conjunct) neƞu-ni mene mala-jaƞa-i bodo-si-mi olokto-ho-ni ‘the younger brother cooked as much as he intended to eat himself’ (younger.sibling-PX3SG REFL spend-PTCP.FUT-RX think-STAT-CV.

358 Elena Perekhvalskaya and Juha Janhunen

•

•

•

•

•

CONN cook-PST-3SG) vs. (disjunct) bii käƞa buni-i-we-ni jakta-mi ‘I imitate the deer roaring’ (1SG deer roar-PTCP.PRS-ACC-PX3SG imitate-1SG). Polar questions are marked by the interrogative pronoun ye’u ‘what’, cliticized to =ye’u, e.g. eyi jali doo-ni-ni bii bi-hi-we-i e-[h]u ise=ye’u ‘didn’t you see that I live in this barn?’ (this barn inside-DAT-PX3SG 1SG be-PTCP.PRS-ACC-PX1SG NEG-2PL see=​INTERR). Another general question marker is =nu (= Oroch =nUU), e.g. ceje digana-hi=nu ‘are you telling the truth?’ (truth tell-2SG=INTERR). When repeated it codes, like =ye’u, alternative questions, e.g. bu-jeƞe-hi=nu, e-jeƞe-hi=nu ‘will you give (it) or not?’ (give-FUT-2SG=INTERR NEG FUT-2SG=INTERR). Quotation is expressed by the quotative particle gum(u), based on the verb gun- : gune- ‘to say’. This particle also has uses related to epistemic modality, cf. e.g. sii ami-hi ye-le bie gumu ‘and where, then, lives your father?’ (2SG father-PX2SG what-LOC be.PRS QUOT), nagda-ha-ti gumu ehi-mi waa ‘they hit him, it seems, but did not kill’ (hit-PST-3PL QUOT NEG-CV.CONN kill). A related particle is (=)gune(i), but in the modern language it functions as a mirative marker, usually connected with sensory perception, jugdi doo-lo-ni yeƞpeñeñeñe o’o gune ‘inside the house there was a hum—birds!’ (house inside-LOC-PX3SG DESCR fowl MIR), tutu bi-si-ni too=de anci gune ‘it was quiet (in the house), even no (sound of) fire!’ (quiet be-PST-3SG fire=FOC PRIV MIR). Another mirative particle is (=)bubu ~ (=)bebu, often used in combination with (=)gune(i), e.g. tinekti nii ilakta-ƞi-ha-ni=bebu ‘yesterday’s man reappeared, indeed!’ (yesterday person appearREV-PST-3SG=​MIR), e-e, ga’i daktä-ni bagdi-he-ni gunei bebu ‘oh, the wings of the crow have grown, indeed!’ (INTERJ crow wing-PX3SG grow-PST-3SG MIR MIR). Other discourse particles include (=)yaja ‘of course’ = ya-ja (do.what-OPT), e.g. nii xaña-la anci bude-i=yaja ‘a person without a soul dies, of course’ (person soulPART PRIV die-PRS=of.course), and (=)gieni ‘so, consequently’, e.g. mene xegdeli-he-mi=le, säägi-jeƞe-i gieni ‘I myself tore it off, so I will also fix it’ (REFL tear. off-INCH-PST-1SG=FOC join-FUT-1SG consequently). Conjunction between syntactically equal nouns is expressed by the additive clitic =dA ‘also, even’, often repeated after each member of the group, e.g. baita-la maƞaxi=da bi-si-ni, agdaxi=da bi-si-ni, xaisi aƞnaxi=da bi-si-ni ‘there were arbitrators, mediators and members of strange clans in those trials’ (trial-LOC arbitrator=ADD be-PST-3SG mediator=ADD be-PST-3SG also stranger=ADD be-PST-3SG). The corresponding disjunctive clitic is =(e)s, repeated as =(e)s . . . =(e)s ‘either . . . or’, ‘neither . . . nor’, e.g. jali doo-ni-ni mama’asa-n​=es exi-n=es bi-hi-ti ‘in the barn there is either his wife or his sister’ (barn inside-DAT-PX3SG wife-PX3SG=DISJ sister-PX3SG=DISJ be-PRS-3PL). The clitic =dA is also used as a general emphatic marker of topic or focus, e.g. [topic] sama sewe-ni=de xei-i-ni ‘as for the shaman’s helper spirit, it wanders (everywhere)’ (shaman helper.spirit-PX3SG=TOP wander-PRS-3SG), [focus] omo dogbo-ni kahi=da je-kce-he-ni ‘he would have eaten all the night’ (one night-PX3SG whole=FOC eat-PROSP-PST-3SG). It may also be attached to an extraposed element, e.g. nii egdi bi-si-ti anana=da ‘there were many people, in old times’ (person many be-PST-3PL long.ago=FOC). A new topic is also marked by the clitic =lA, e.g. wali=le juke=de degde-he-ni, ƞen[e]-he-ni ‘the raven rose with difficulty and flew away’ (raven=TOP hardly=FOC rise-PST-3SG go-PST-3SG). A subject introducing a new topic with the particle =lA may take the final position in the sentence, e.g.

Udihe 359

•

ihi käämbu-we-ni tagdi-nde-he-ni uja’a=la ‘Udza pulled out a crooked larch tree’ (larch.tree crooked-ACC-PX3SG pull-SEM-PST-PX3SG Udza=TOP). Emphasis, contrast, definiteness, and several other diffuse discourse features are expressed by a wide selection of additional particles, many of which have lexical, rather than fully grammaticalized functions. These include =bede [similative], e.g. a’i-mula, neƞu-mule=bede bi-si-ti ‘they were like elder and younger brothers’ (elder.brother=SIM younger.​brother=SIM be-PST-3PL); =gdA(-lA) [emphatic], e.g. ono=gdo eme-he-mu=de e-u saa ‘and how we arrived here, we do not know’ (how=EMPH come-PST-1PL.​EXCL​=FOC NEG-1PL.​EXCL know); =gu ‘any, some’ [indefiniteness on pronouns], e.g. ga’i odo’o bi-si-ni, ye’u=gu ye’u=gu diga-mi tuu bi-si-ni ‘crow grandfather lived like that, ate something’ (crow grandfather be-PST3SG what=INDEF what=INDEF eat-CV.CONN so be-PST-3SG); =kA [emphasis on pronouns], e.g. bi-ke e-jeƞe-i menje ‘I will not do housekeeping’ (1SG-EMPH NEG-FUT-1SG do.housekeeping); =mA’i ~ =meli [limitative], e.g. omo nehä=me’i osi-gi-ho-ni ‘only the skin remained’ (one skin=LIM become-REV-PST-3SG); =sA’i [terminative], e.g. uta-digi namu-la=sa’i dataƞ-ki-ti ‘then they went down the river up to the sea’ (this-ABL sea-LOC=TERM move.to.mouth-PST-3PL); =sA’AnA ~ =sAnA [restrictive], e.g. nii-ni anci bi-si-ni, omo keku=se’ene bi-si-ni ‘there was no one, only a cuckoo’ (person-PX3SG PRIV be-PST-3SG one cuckoo=RESTR be-PST-3SG); =tAnA [contrastive], e.g. amuliƞka amuli-tigi ƞene-he-ti, kancuga=tene kañumi be’äsa-tigi-ni ƞene-he-ti ‘the Amulinka clan went to the Amuli river, while the Kanchuga clan went to the Kanyumi river’ (Amulinka Amuli-DIR go-PST3PL Kanchuga=CONTR Kanyumi river-DIR-PX3SG go-PST-3PL).

LEXICON AND LANGUAGE CONTACTS The basic lexicon of Udihe is of a predominantly Tungusic origin, and the neighbouring Tungusic languages are also the principal sources of loanwords. Jurchen-Manchu has been a major source of cultural vocabulary, but many items have probably been transmitted in the Amur basin via Nanai, and it is difficult to identify any possible direct borrowings in the lexical corpus. Via Manchu, a number of Mongolic, Para-Mongolic, and Turkic items have also entered Udihe. More recently, Chinese and Russian have been sources of direct lexical influence. Examples of Chinese loanwords are tuduje ‘potato’ (← tudouzi 土豆子), xuece ‘train’ (← huoche 火車), and (Bikin) sunje ‘grandchild’ (← sunzi 孫子). Older Russian loanwords include, for instance, lafka ‘shop’ (← lávka), takana ‘drinking glass’ (← stakán), doktori ‘physician’ (← dóktor). Of potential importance are the words with no cognates elsewhere in Tungusic and with no convincing external etymology. These comprise basic vocabulary, topographical and zoographical terms, as well as cultural items, and include, for instance, a’ana ‘boat’, a’asa ‘bay’, da’a ‘cotton wool’, dühi ‘brain’, gehe ‘bad’, gobo’o ‘fly’, jahi ‘wild boar’, ta’aƞki ‘middle’, tehu ‘all’, etc. A widespread Eurasian cultural item is sa’i (< *caki) ‘salt’, but it is different from Oroch daksʊ/n ‘salt’ and its cognates in several other Tungusic languages, which all go back to Mongolic *dabu-su/n. REFERENCES AND FURTHER READING Arsen’ev, V. K. [В. К. Арсеньев] (2008) Русско-орочский словарь: Материалы по языку и традиционной культуре орочей и удэгейцев [Russian-Oroch dictionary:

360 Elena Perekhvalskaya and Juha Janhunen

Materials on the language and traditional culture of the Oroch and Udihe], edited by A. Kh. Girfanova [А. Kh. Гирфанова] & N. L. Sukhachev [Н. Л. Сухачев], Санкт-​ Петербург [St. Petersburg]: Санкт-Петербургский государственный университет. Baichura, Uzbek (U. Sh.) (1978–1979) ‘Word intonation in Udehe according to instrumental-phonetic data’, Orientalia Suecana 27–28: 108–117, Uppsala. Belikov V. I. [В. И. Беликов] & E. V. Perekhvalskaya [Е. В. Перехвальская] (2002) ‘Тазов язык’ [The Taz language], in: V. P. Neroznak [В. П. Нерознак] (ed.), Языки народов России: Красная книга, Энциклопедический словарь-справочник [Red book of the languages of the peoples of Russia: An encyclopaedic dictionary], 170– 174, Москва [Moscow]: “Аcademia”. Braïlovskii S. N. [С. Н. Браиловскiй] (1901) ‘Тазы или удихэ: Опытъ историкоэтнографическаго изслѣдованiя [The Tazy or Udihe: A historical and ethnographic sketch]’, Живая старина 2: 129–216, 3–4: 323–455. С.-Петербургъ [St. Petersburg]. Girfanova A[lbina] Kh. [А. Х. Гирфанова] (2001) Словарь удэгейского языка [A dictionary of the Udihe language], Санкт-Петербург [Sankt-Petersburg]: Nauka. Girfanova, Albina (2002) Udege, Languages of the World/Materials 255, Munich: LINCOM Europa. Girfanova A[lbina] Kh. [А. Х. Гирфанова] (2003) ‘Из истории изучения удэгейского языка’ [On the history of Udihe language studies], in: Juha Janhunen & Asko Parpola (eds.), Remota Relata: Essays on the History of Oriental Studies in Honour of Harry Halén, Studia Orientalia 97: 13–23, Helsinki. Hölzl, Andreas (2018) ‘Udi, Udihe, and the language(s) of the Kyakala’, International Journal of Diachronic Linguistics and Linguistic Reconstruction 15: 111–146. Janhunen, Juha (1999) ‘Tonogenesis in Northern Asia: Udeghe as a tone language’, Journal de la Société Finno-Ougrienne 88: 31–38, Helsinki. Janhunen, Juha & Fu Yuguang & Guo Shuyun (1999) ‘The Kyakala in China: History and present situation’, Journal de la Société Finno-Ougrienne 88: 248–252, Helsinki. Kanchuga, Aleksandr カンチュガ・アレクサンドル　(2002) Tsumagari Toshirō 津曲 敏郎(ed.),『ウデヘ語自伝テキスト』[An Udehe Autobiographical Text with a Russian Translation],『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 17 = ELPR Publications Series A2–019, Suita. Kanchuga, Aleksandr カンチュガ・アレクサンドル　(2003–2006), Tsumagari Toshirō 津曲敏郎 (ed.),『ウデヘ語読本自伝』,『ウデヘ語自伝テキスト 2–4』[Autobiographical story, in the Udehe language with Russian translation], [1]『少年時 代』[Boyhood], 2『青年時代』[Youth], 3『学生時代』[Student years], 4『妻ファ ヤの思い出』[Memories of my wife Faya],『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 22 [2003], 28 [2005], 29 [2006], 33 [2006], 札幌 [Sapporo]: 北海道大学大学院文学研究科. Kazama, Shinjirō 風間伸次郎 (2004a)『ウデヘ語テキスト (A)』[Udihe Texts (A)], 『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 24/A, 東京 [Tokyo]: 東京外国語大学アジア・アフリカ言語文化研究所. Kazama, Shinjirō 風間伸次郎 (2004b)『ウデヘ語テキスト (B)』[Udihe Texts (B)], 『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 24/B, 札幌 [Sapporo]: 北海道大学. Kazama, Shinjirō 風間伸次郎 (2006–2010) 『ウデヘ語テキスト 2–6, 付：文法 概説』[Udihe texts 2–6, with a grammatical synopsis],『ツングース言語文化 論集』[Publications on Tungus Languages and Cultures] 31 [2006], 35 [2007], 42

Udihe 361

[2008], 44 [2009], 47 [2010], [2, 4–6] 東京 [Tokyo]: 東京外国語大学アジア・アフ リカ言語文化研究所, [3] 札幌 [Sapporo]: 北海道大学. Kazama, Shinjirō (2008) ‘The ‘plural’ markers in Udihe.’ In: Kurebito Tokusu (ed.), Ambiguity of Morphological and Syntactic Analyses, 229–246, Tokyo: Tokyo University of Foreign Studies, Research Institute for Languages and Cultures of Asia and Africa (ILCAA). Kazama, Shinjirō (2022) ‘A grammar of Udihe’,『思言 東京外国語大学記述言語学 論集』 Shigen: Tokyo University of Foreign Studies Descriptive Linguistic Papers 18: 3–42, Tokyo: Tokyo University of Foreign Studies. Kormushin I. V. [И. В. Кормушин] (1998) Удыхейский (удэгейский) язык: Материалы по этнографии, очерк фонетики и грамматики, тексты и переводы, словарь [The Udihe language: Materials on ethnography, a sketch of phonetics and grammar, texts and translations, dictionary], Москва [Мoscow]: “Наука”. Lar’kin V. G. [В. Г. Ларькин] (1959) Удэгейцы: Историко-этнографический очерк с середины XIX в. до наших дней [The Udihe people: A historical and ethnographic sketch from the middle of the XIX century to the present day], Владивосток [Vladivostok]. Leontovich, Sergei [Сeргѣй Леонтовичъ] (1896) Краткiй русско-ороченскiй словарь съ грамматической замѣткой [A concise Russian-Orochen dictionary with a grammatical note], Записки Общества изученiя Амурскаго края 5 (2), Владивостокъ [Vladivostok]: Типографiя Н. В. Ремезова. Lie, Hiu (1978) ‘Über die Benennungen der Udihe: Ein Beitrag zur historischen und ethnographischen Untersuchung über den Volkstamm Udihe’, in: Gerhard Doerfer & Michael Weiers (eds.), Beiträge zur nordasiatischen Kulturgeschichte, Tungusica 1: 179–201, Wiesbaden: Otto Harrassowitz. Maack, R[ichard Otto] [Р. К. Маак] (1859) Путешествiе на Амуръ [Travel on the Amur], совершенное по распоряженiю Сибирскaго Отдѣла Русскaго Географическaго Общества, въ 1855 году, Санктпетербургъ [St. Petersburg]: Bъ типографiи Карла Вульфа. Margaritov V. P. [В. П. Маргаритовъ] (1888) Объ орочахъ Императорской гавани [On the Oroch of Imperatorskaya Gavan’]. Санкт-Петербургъ [St. Petersburg]: О[бщест] во изученiя Амур[скаго] края въ г. Владивостокѣ. Menges, K. H. (1968) ‘Die Sprache der Udihe’, in: Altaistik: Tungusologie, Handbuch der Orientalistik I.V.3: 129–171, Leiden: E. J. Brill. Nadarov I. P. [И. П. Надаровъ] (1887) Сѣверно-Уссурiйскiй край (съ картою) [The northern Ussuri region (with a map)], Записки Императорскaго Русскaго Географическaго Общества по общей географiи 17 (1), Санктпетербургъ [St. Petersburg]: Военная типографiя. Nikolaeva, Irina (2000) ‘The vocalic system of Udihe’, Eurasian Studies Yearbook 72: 113–142, Berlin & London: Eurolingua. Nikolaeva, Irina & Elena Perekhvalskaya & Maria Tolskaya (2002) Udeghe (Udihe) folk tales, Tunguso-sibirica 10, Wiesbaden: Otto Harrassowitz. Nikolaeva, Irina & Elena Perekhvalskaya & Maria Tolskaya, with preface by Toshiro Tsumagari (2003) Udeghe (Udihe) texts, Endangered Languages of the Pacific Rim (ELPR) Publications A2–025. Osaka: Osaka Gakuin University, Faculty of Informatics. Nikolaeva, Irina & Maria Tolskaya (2001) A Grammar of Udihe, Mouton Grammar Library 22, Berlin: Mouton de Gruyter.

362 Elena Perekhvalskaya and Juha Janhunen

Perekhvalskaya, Elena [Е. В. Перехвальская] (2010) ‘Аспирированные и глоттализованные гласные в диалектах удэгейского языка’ [Aspirated and glottalized vowels in Udihe dialects], Вопросы языкознания (6): 60–87, Москва [Moscow]. Poniatowski, Stanisław (1966) Dziennik wyprawy (do kraju goldów i oroczonów w 1914 roku) [Diary of an expedition to the Gold and Orochon in 1914], Wstępem zaopatrzyła Barbara Heidenreich, Wrocław: LUD. Protodiakonov, P[rokopii] [П. Протодьяконов] (1888) Краткiй русско-ороченский словарь [Concise Russian-Orochen dictionary], Казань [Kazan]: Изданiе Православнаго миссiонерскаго общества. Radchenko G. L. [Г. Л. Радченко] (1988) ‘Спорные вопросы фонетики удэгейского языка’ [Controversial issues of Udihe phonetics], in: M. D. Simonov (ed.) (1988), 35–43. Schmidt, P[eter] (1928) ‘The language of the Oroches’, Acta Universitatis Latviensis 17: 17–62, Riga. Schneider, E. R [Е. Р. Шнейдер] (1936) Краткий удэйско-русский словарь, с приложением грамматического очерка [A concise Udihe-Russian dictionary, with a grammatical sketch]. Москва [Moscow] & Ленинград [Leningrad]: Государственное учебно-педагогическое издательство. Schneider, E. R. [Е. Р. Шнейдер] (1937) Материалы по языку анюйских удэ (фонетика, морфология, лексика) [Materials on the language of the Anyui Udihe: Phonetics, morphology, lexicon]. Труды научно-исследовательской ассоциации Института Народов Севера 11. Москва [Moscow] & Ленинград [Leningrad]. Simonov M. D. [М. Д. Симонов] (1988) ‘Фонологическая система удэгейского языка’ [The phonological system of the Udihe language], in: M. D. Simonov (ed.) (1988), 44–88. Simonov M. D. [М. Д. Симонов] (1990) ‘Удэгейская лексика в записях польского этнографа С. Понятовского’ [Udihe words in the fieldnotes of the Polish ethnographer S. Poniatowski]’ in: Alfred F. Majewicz (ed.), The Collected Works of Bronisław Piłsudski, vol. 2: 80–85. Mouton de Gruyter. Simonov, M. D. [М. Д. Симонов] (ed.) (1988), Историко-типологические исследования по тунгусо-маньчжурским языкам [Historical and typological studies of the Tungusic languages], Новосибирск [Novosibirsk]: “Наука”. Simonov M. D. [М. Д. Симонов] & V. T. Kyalundzyuga [В. Т. Кялундзюга] (1998) Словарь удэгейского языка: Удэгейско-русско-удэгейский / An Udeghe Language Dictionary: Khor River Region Dialect, vols. 1–3. Preprint, published by Alfred F. Majewicz. International Institute of Ethnolinguistic and Oriental Studies, Monograph series 15/1–3. Ste̹ szew: IIEOS. Simonov M. D. [М. Д. Симонов] & V. T. Kyalundzyuga [В. Т. Кялундзюга] & M. M. Khasanova [М. М. Хасанова] (1998) Фольклор удэгейцев: ниманку, тэнунгу, эхэ [Folklore of the Udihe: nimanku, tènungu, èkhè], Памятники фольклора народов Сибири и Дальнего Востока 18. Новосибирск [Novosibirsk]: “Наука”. Sunik, O. P. [О. П. Суник] (1968) ‘Удэгейский язык’ [The Udihe language], in: Монгольские, тунгусо-маньчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 210–232, Ленинград [Leningrad]: “Наука”. Sunik, O. P. [О. П. Суник] (1997) ‘Удэгейский язык’ [The Udihe language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 236–248, Москва [Moscow]: “Индрик”.

Udihe 363

Tsumagari, Toshirō 津曲敏郎 (1997)「ウデヘ語文例」[Sample sentences in Udihe], 『言語センター広報』5: 89–91, 小樽 [Otaru]: 小樽商科大学言語センター. Tsumagari, Toshirō 津曲敏郎 (1998)「ウデヘ語文例補遺」[Additional sample sentences in Udihe], 『言語センター広報』6: 107–110, 小樽 [Otaru]: 小樽商科大学 言語センター. Tsumagari, Toshirō (2011) ‘An Udihe folktale text: “Solomo and Tausima”’,『北方人 文研究』Journal of the Center for Northern Humanities 4: 75–93, 札幌 [Sapporo]. Zinder, L. R [Л. Р. Зиндер] (1948) ‘Экспериментальное изучениe фонетики северных языков [Experimental study of the phonetics of northern languages]’, Известия Академии Наук СССР, Отделение литературы и языка [Bulletin of the Academy of Sciences of the USSR, Section of literature and language], 7 (6): 579–581, Москва [Moscow].

CHAPTER 14

NA NA I Shinjiro Kazama

Nanai (Nanay) is the language of the largest Amur Tungusic ethnic group, historically living in the Middle and Lower Amur basin, in the region of the confluence of the rivers Amur, Sungari, and Ussuri, on both sides of the modern Sino-Russian border. The ethnonym Nanai (Russian MASC nanáec : PL nanáicy : FEM nanáika) is based on the endonym naanaɪ ~ naanɪ < naa+naɪ ‘land person’ = ‘local people’, also used by the Ulcha and Oroch. In older Russian and international literature, the Nanai used to be known by the exonym Gold (Russian gol’d : PL gól’dy), a term of unknown etymology but attested in Neghidal as goldɪx and in Oroch as gogdi (< *goldï) in reference to the Nanai. Another exonym of unknown origin, used by the early Russian settlers for the Nanai in the 17th century was “Natki” (nátki). In Chinese the Nanai, traditionally classified under the generic term yupi dazi 魚皮韃子 ‘Fishskin Tatars’, are officially known as Hezhe 赫哲 or Hezhen 赫真, a name based on the Tungusic root *xeje ‘lower course (of a river)’ : xeje/n ‘inhabitant of the lower course’, used by the Neghidal in the form xejen in reference to the Ulcha and by some Nanai groups in the form xeje-ni or xeje naɪ-nɪ in reference to those Nanai (and Ulcha) who live lower down the river. As a borrowing, this name is also attested in Manchu in the form xejen (hejen). In both Russia and China the Nanai are recognized as an official ethnic minority. In Russia, the Nanai ethnic territories are administered in the composition of Khabarovsk Krai as well as, more marginally, the Maritime Province (Primor’e). The Nanai are the titular ethnic group of the Nanai Raion, located on the Amur just north of the city of Khabarovsk. In China, the Nanai (Hezhe zu 赫哲族) inhabit the region between the mouths of the Sungari and Ussuri, where they have been assigned three township-level titular areas within the districts of Jiamusi and Shuangyashan, Heilongjiang Province. The number of the Nanai in Russia has been reported at c. 11,700 (2010) and in China at c. 5,400 (2010). While the population has been stable, or even growing in China, the number of fluent speakers has gone down to c. 200 on the Russian side and only a few individuals, at most, in China. In this situation, the language is severely endangered, but there is information that some generational transmission may still be taking place in places like the ethnic village of Dada in the Nanai Raion, with a Nanai-dominated local population of some 400. Nanai is divided into several dialects and subdialects. The main division is between three dialect groups, traditionally known by the somewhat misleading names as the “Upper”, “Middle”, and “Lower Amur” dialects, all of which are, technically speaking, spoken on the Lower Amur, as well as on the lower courses of the Sungari and Ussuri. The “Upper Amur” dialect, today extinct, which has also been called the “Ussuri dialect”, was spoken in the Lower Ussuri basin, including the Bikin basin on the Russian side. The “Middle Amur” dialect is spoken in the actual Lower Amur basin and is divided into DOI: 10.4324/9781315728391-14

Nanai 365

three varieties, spoken in the ethnic villages of Sikachi-Alyan, Naikhin, and Dzhuèn, respectively. The “Lower Amur” dialect is spoken at the villages of Bolon’ and Èkon’, in the region of the lakes Bolon’ and Khummi, as well as in the basin of the river Gorin, including the village of Kondon. The Gorin Nanai are historically known as the Samar or Samaghir, an entity which comprises also Neghidal speakers in the northern part of the territory, adjacent to the upper course of the Amgun’. In addition to the actual Nanai dialects, the Nanai-Hezhen complex comprises two important “mixed” varieties, technically known as Kili and Kilen, which may also be classified as separate languages. Kili was spoken in the basin of the rivers Kur and Urmi on the Russian side and is also known as the Kur-Urmi “dialect” of Nanai, while Kilen is spoken on the Chinese side of the Lower Sungari and Amur and is in Chinese sources identified as the “Kilen” (qile’en 奇勒恩) “dialect” of “Hezhen”. The terms Kili and Kilen, also attested as clan names (and surnames) among the Nanai, are based on the tribal name kile ~ kilen, used as an endonym by several groups of Ewenki and as an exonym for the Ewenki by the speakers of the other Tungusic languages. This reflects the fact that, as languages, Kili and Kilen are admixtures of Ewenki-type lexicon with Nanai-type grammar. This means that they are grammatically very close to the “other” dialects of Nanai, but their basic lexicon is virtually identical with that of Ewenki, with several diagnostic phonetic and semantic developments of the Ewenki type. It has, however, also been noted that Kili and Kilen are not identical with each other, for while the lexical component of Kili is unambiguously of the Ewenki type, the Kilen lexicon contains some features that point to an early source of an Orochic type, without, however, containing later Oroch-Udihe developments. Attempts to write Nanai in the Cyrillic script were made already in the late 19th century. A primer in the “Upper Amur” dialect, also in Cyrillic, was published in 1928 by N. A. Lipskaya-Val’rond. Later, the Naikhin dialect was adopted as the basis of the literary language. Between 1931 and 1937, several school textbooks were published in the Unified Northern Alphabet, replaced after the war by a new Cyrillic orthography. Today, Nanai continues to be taught at a few schools, but few children know it as a native language. There is also an online course designed in 2011 by A. S. and L. T. Kile and M. N. Zavalishina (available at “Portal Khabarovsk”). In addition, the Nanai language has occasionally been used in works of belles-lettres, newspapers, and theatre pieces. DATA AND SOURCES The first information concerning the Nanai language was provided by the early Russian explorers on the Amur, including A. S. Orlov (1869). Samples of Nanai lexicon were first collected and published by Aleksandr Protodiakonov (1869–1870), as analysed by I. I. Zakharov (1876). Larger vocabularies were made available by Prokopii Protodiakonov (1901) and Wilhelm Grube (1900). Other early works, containing both ethnographic and linguistic information are those by I. A. Lopatin (1922) and Owen Lattimore (1932), of which the latter is focused on the Nanai of the Lower Sungari basin. In the Soviet Union, the study of the Nanai language was initiated by A. N. Ulitkin (1933ab) and T. I. Petrova (1933). The latter was the author of the first school textbooks in the Latinized orthography, but she also published the first actual dictionary (1935) and grammar (1941), continuing her work after the war with lexical studies (1948, 1954, 1957) and another dictionary (1960). Another early school teacher among the Nanai was A. P. Putinceva, who later published a series of methodical tools (1955, 1956, 1957) as

366 Shinjiro Kazama

well as important data on the Gorin (Samar) dialect (1954ab). At about the same time, the question concerning the linguistic and ethnic identity of the Samar/Samaghir was taken up by Vilmos Diószegi (1953), who correctly noticed the dual Nanai-Neghidal affiliations of this group. The major authority on Nanai in the Soviet Union came to be V. A. Avrorin, who published a large number of papers on a variety of issues concerning Nanai grammar (1948, 1953, 1956), as well as phonetics and phonology (1957, 1958ab). His work culminated in a massive Nanai grammar (1959–1961), followed by a grammatical sketch (1968), a volume with syntactic studies (1981), and a collection of folkloric texts (1986). Later Russian works on Nanai include those by A. V. Smolyak (1970) on anthroponymy, G. L. Radchenko (1985, 1986) on phonetics, as well as, most recently, Sofia Oskolskaya (S. A. Oskol’skaya) on morphological functions and other issues (2014, 2015, 2016, 2020ab), also in collaboration with N. M. Stoinova (2019). Particularly valuable documentation work has been carried out by several native Nanai scholars. Most importantly, S. N. Onenko published a series of bilingual Russian-Nanai and Nanai-Russian dictionaries (1958, 1980, 1982, 1986) and related lexical studies (1955, 1958ab). He worked also on issues pertaining to phonetics (1975), orthography (1981a, 1983), morphology (1977, 1981b), and bilingualism (1972). N. B. Kile was mainly active in the field of lexical and onomastic studies (1973, 1976, 1977ab, 1981, 1986). He published also a thematic dictionary (1994) and a collection of folklore (1996). A more recent thematic dictionary has been published by L. Zh. Zaksor (2003), a specialist on Nanai native language teaching. In Japan, the study of the Nanai language has long been in the focus of Kazama Shinjirō, who, on the basis of his own field work, has published, in close collaboration with native speakers, two collections of texts (1991–1993) and 13 volumes of folkloric materials (1995–2012). Most of these publications, which contain texts with an interlinear translation into Japanese, are accompanied by audio material on disc. Kazama has also published papers dealing with various lexical and grammatical topics, including agreement (1994), kinship terms (1995), rhetorical questions (2005/2007), participles (2006b, 2010a), pragmatics (2008), aspect (2010b), modality (2011ab), voice (2013a), transitivity (2014), complex sentences (2015b), and information structure (2016, 2019). Dialectological work on Nanai in the Soviet Union was initiated by O. P. Sunik, who focused on the special position of Kili (1948, 1958a), in which connection he also examined the varieties spoken by the Nanai in China (1958b). This work stimulated Gerhard Doerfer to author a series of studies on Kili (1973, 1975, 1984) and its relevance to general linguistics as a universally rare case of “language mixing”. At the same time, dialectological work was continued in the Soviet Union by L. I. Sem, who published a monographic study of the Ussuri (Bikin) dialect (1976). She also authored a grammatical sketch of Nanai in a collective volume (1997). In China, Hezhen and Kilen were first described by An Jun (1986), followed by another description of Kilen (in English) by Zhang & Zhang & Dai (1989). Li Linjing (2014, 2020) has worked on the verbal morphology of Kilen. Text samples from several dialects of Nanai spoken on the Russian side, with interlinear translation into Japanese, have been made available by Kazama (2006, 2012–2021, 2015a), who was also the first to identify the taxonomic difference between Kili and Kilen (1996/1998). The only general source on Nanai in English is the grammatical description by Ko Dongho & Yurn Gyudong (2011), accompanied by the text collection by Kim Juwon & al. (2015). Earlier works in Western languages include the summary by K. H. Menges (1968), cf. also Menges (1964), and the phonological note by Janhunen (1986).

NANAI 367

The present chapter is based on the Naikhin dialect, with occasional reference to the specific features of other dialects. SEGMENTAL STRUCTURE Nanai has a symmetrical system of 6 vowel qualities (Table 14.1) comprising two rounded back vowels u o, two unrounded front vowels i ɪ, and two unrounded central vowels e [ə] a. TABLE 14.1 NANAI VOWELS u o

i e

ɪ

a

The same distinctions are also attested as long vowels (VV): uu oo ii ɪɪ ee aa, which can appear in all syllables. Harmonically, the vowels represent two sets, which may be identified by the conventional terms “hard” and “soft”, respectively. The opposition between the two sets involves both height and tongue root position, in that the “soft” qualities u e i [u ə i] are pronounced higher and with a relatively more advanced tongue root, while the “hard” qualities o a ɪ [o a ɪ] are pronounced lower and with a relatively more retracted tongue root. The vowel o contains also the reflexes of earlier *ʊ (< *u), still preserved in Ulcha, e.g. Nanai ogda ‘boat’ vs. Ulcha ʊgda (< *ugda). Nanai still preserves the distinction between *ɪ (*ï) and *i (< *i), but the distinction is phonetically small and is best observed in the initial syllable, where ɪ can come close to the quality of [e], especially if the following syllable has a, as in ɪnda [enda] ‘dog’, while in non-initial syllables both i and ɪ tend to be pronounced as [i]. The distinction is nevertheless phonemic, for roots containing ɪ in the first two syllables contrast phonetically with roots containing i, and the difference affects the choice of suffixal vowels, e.g. a vs. e, as in sɪsɪ [sɪsɪ] ‘larchtree’ : ACC sɪsɪ-wa vs. ini [iɲi] ‘day’ : ACC ini-we, ɪsɪ- ‘to reach’ : PTCP.PRF ɪsɪ-xan vs. ii- ‘to enter’ : PTCP. PRF ii-xen. In final position, i and ɪ are often devoiced and can be dropped. Also, all vowel qualities can phonetically merge with a final n, yielding nasalized vowel qualities, e.g. sɪan [ɕiã] ‘ear’. In addition to the long monophthongs the vowels can form diphthongoid sequences. In the latter, following the rules of vowel harmony, the vowels e a can be combined with i ɪ u o, yielding the combinations ie ɪa ue oa and ei aɪ eu ao, but the vowels i ɪ u o can also be combined with each other, yielding the combinations iu ɪo ui oɪ. In practice, especially the combinations ei aɪ eu ao are often pronounced as the corresponding monophthongs ii ɪɪ uu oo, particularly over morpheme boundaries, e.g. japa- ‘to take’ : PTCP.IMPRF japa-ɪ > japɪɪ [dzapi:], ene- ‘to go’ : PTCP.IMPRF ene-i > enii [əɲi:], ta‘to do’ : PASS-PRCP.IMPRF ta-o-rɪ > toorɪ [to:ri]. The fact that both long vowels and vowel sequences can be formed at morpheme boundaries suggests that neither the long monophthongs nor the diphthongoid sequences are independent phonemic entities. In the consonants, Nanai retains the typical Tungusic system of 18 segments (Table 14.2), identical with that of, for instance, Siberian Ewenki, as well as Ulcha.

368 Shinjiro Kazama TABLE 14.2 NANAI CONSONANTS m

n

ñ

ƞ

b p

d

j

g

t

c

k

s w

x y

l r

The distinction between the weak and strong obstruents b d j g vs. p t c k is, at least in the dialects spoken the Russian side, based mainly on voice (in the weak set). It may be noted that Nanai preserves Proto-Tungusic *p as a stop, while x represents the original Proto-Tungusic velar fricative *x. However, in a few words, a dialectal or idiolectal variation between k and x in initial position is observed, e.g. kupi- ~ xupi- ‘to play’. Also, in the “Upper Amur” dialects, apparently due to an areal development shared with Manchu and Udihe, word-initial and postvocalic p has been recorded with the fricative value f [ɸ], e.g. Bikin mafa ‘grandfather’ > ‘bear’ = Naikhin mapa. The palatal stops j c are realized as alveolo-palatal affricates [dʑ tɕ] before the vowels i ɪ, but as dental affricates [dz ts] before other vowels, a pronunciation also observed in the Khor dialect of Udihe. The velar nasal is preserved as a distinctive segment in initial position, e.g. ƞaala ‘hand’ < *ƞaala, but in several words it has been lost. In at least some cases the loss must be recent, since it has not taken place in Ulcha and Uilta, e.g. Nanai ui ‘who?’ < *ƞui > Ulcha & Uilta ƞui, Nanai ene- ‘to go’ < *ƞene- > Ulcha & Uilta ƞene-. Some items show an alternation between ƞ and w, e.g. ƞonɪmɪ ~ wanɪmɪ ‘long’, possibly because of Udihe influence. In words containing “hard” vowels, the velars k g are pronounced with the uvular qualities [q ɢ]. The velar fricative x, on the other hand, has values ranging from palatal [ç] to velar [x] to laryngeal [h] depending on the vowel environment. Historically, *x has been palatalized to s before *i and *ï in Nanai, as also in Ulcha and Uilta, e.g. Nanai siƞekte ‘bird-cherry’ = Ulcha siñekte < *xinƞekte > Ewenki inƞekte, Nanai sɪmana ‘snowfall’ = Ulcha sɪmana < *xïma-na > Ewenki imana. Similarly, the dentals *t *d have been palatalized to c j before *i and *ï, as also in Manchu and Uilta; however, this development is not shared by Ulcha, meaning that it took place independently in both Nanai and Uilta, e.g. Nanai cɪmɪɪ ‘morning’ vs. Ulcha tɪmaɪ < *tɪmakɪ > Uilta cimai, Nanai jɪlɪ ‘head’ vs. Ulcha dɪlɪ < *dɪlɪ > Uilta jili. An analogous palatalization of n [n] to [ɲ] is also observed in Nanai and Uilta, but not in Ulcha; however, in this case the phenomenon remains phonetic, as no opposition between n and ñ is possible before the vowels i ɪ, cf. e.g. Nanai eni [əɲi] ‘mother’ : PX3SG eni-ni [əɲiɲi] = Uilta eni-ni [əɲiɲi] vs. Ulcha eni-ni [ənini]. Consonant clusters are generally well preserved in Nanai, including, for instance, clusters of two stops. e.g. pokto ‘road’ < *pokta > Ewenki xokto, tugde ‘rain’ < *tügde > Ewenki tigde, and clusters of nasal+stop, e.g. ɪnda ‘dog’ < *ƞïnda > Ewenki ƞina-kin, xemburen ‘sharp’ < *xembür- > Ewenki emir. The labial+velar sequences pk bg mƞ are, however, preserved only in Bikin, but metathesized in Naikhin, e.g. Naikhin jakpon ‘eight’ < *japkon > Bikin jafkon = Ewenki japkun, Naikhin sogbo ‘fish skin’ < *sobgo > Bikin sobgo = Neghidal sobgu, Naikhin aƞma ‘mouth’ < *amƞa > Bikin amga = Ewenki

Nanai 369

amƞa. In Bikin, a metathesis in the opposite direction is attested over morpheme boundaries, e.g. aag ‘elder brother’ : PX1SG *aag-bɪ > aabgɪ. In Gorin, regressive assimilation is observed in these cases. The sequences *ns *ls are represented as nt lt in Nanai, e.g. nanta ‘skin’ < *nan-sa > Ewen nanra, polta ‘blanket’ < *pul-sa > Ewen xulra. The sequence *ms seems to have remained intact, e.g. xumsə- ‘to overturn, to tilt’ < *xömse> Oroch umse-. An important feature of Nanai, shared with other Tungusic languages in the Amur region, is the tendency to reduce the load of the vibrant (*)r. In Nanai, this segment is lost in intervocalic position at the boundary between the second and third syllables, e.g. ice- ‘to see’ : PTCP.IMPRF ice-i < *ice-ri/i, but not between the first and second syllables, e.g. goro ‘distant’ < *gora, nor after long vowel elements, e.g. ao- ‘to sleep’ : PTCP. IMPRF ao-rɪɪ. The clusters rk rg, as present in Ewenic, are represented as ik ig in Nanai, except in the Gorin dialect, where the representation is tk dg, e.g. Naikhin uike vs. Gorin utke (Ewenki urke), Naikhin xuige vs. Gorin xudge ‘heavy’ (Ewenki urge). Note also the correspondences in Naikhin aɪƞanɪ ‘year’ vs. Gorin anƞanɪ vs. Kilen arƞa (Ewenki anƞanii). A trend shared with Udihe is the glottalization of intervocalic *k, which in Nanai has led to complete loss, e.g. baa- ‘to find’ < *baka- > Udihe ba’a-. PHONOTACTICS AND MORPHOPHONOLOGY The general syllabic structure of Nanai may be schematized as (C)V(V)(C). Monosyllabic free forms of the type (C)V are always pronounced long, though lexically they may be understood as containing a single vowel, e.g. 1SG mii = √mi. In other structural types the distinction between short and long vowels is functionally significant and may be illustrated by many minimal pairs, e.g. pikte ‘child’ vs. piikte ‘nettle,’ un- ‘to say’ vs. uun- ‘to melt,’ osɪ- ‘to become’ vs. oosɪ- ‘to clear up’. The analysis of long vowels as sequences is supported by the fact that they behave quantitatively in the same way as vowel sequences: for instance, they condition the preservation of a suffix-initial *r, as in ii- ‘to enter’ : PTCP.IMPRF ii-rii, ao- ‘to sleep’ : PTCP.IMPRF ao-rɪɪ. Some authorities (notably Avrorin) have maintained that several vowel sequences can occur in two phonetically and phonologically distinctive types, termed “descending” and “ascending”. While this seems to be true of at least some varieties of the “Middle Amur” dialects, including Naikhin, the distinction should not be mistaken to involve a true tonal or accentual opposition. More likely, the “ascending” sequences should be analysed as containing an intervocalic glide y w, e.g. (“ascending” vs. “descending”) nayɪ ‘harpoo shaft’ vs. naɪ ‘man, person’, sawo ‘whitefish’ vs. pao ‘gun’, xuwel ‘across’ vs. juer ‘two’. The sequences ie ɪa iu ɪo, as in e.g. jɪa ‘friend’, pɪo ‘ticket’, are supposed to be always of the “ascending” type, but in the lack of an opposition, there is apparently no need to postulate an intervocalic glide for them. Apart from vowel harmony of the “tongue root” (vertical) type, which divides the vowels into the “soft” (higher) and “hard” (lower) sets, i.e. u e i vs. o a ɪ, there are no combinatory restrictions concerning the vowels. In connection with the merger of the quality *ʊ (< *u) with o, the conditions for labial harmony were lost. As a result, traces of labial harmony remain in lexical roots, but the phenomenon is not active in suffixes. In this respect, Nanai differs from Ulcha and Uilta, e.g. Nanai oƞbo- ‘to forget’ : PST-3SG oƞbo-xa-nɪ vs. Ulcha oƞbo-xo-nɪ = Uilta oƞbo-xo-ni. As far as consonants are concerned, morphophonological alternations are mainly manifested as assimilations by voice or nasality. Thus, a stem-final voiced obstruent remains

370 Shinjiro Kazama

voiced before a suffix-initial voiced consonant, but is devoiced before an unvoiced consonant and nasalized before a nasal, e.g. aag- ‘elder brother’ : PX 1SG aag-bɪ : 2SG aak-cɪ : 3SG aaƞ-nɪ. As in other Tungusic languages, syllable-final obstruents could also be analysed as unmarked “archiphonemic” entities, e.g. aaG-bɪ : aaG-cɪ. A lexical factor is involved in the alternation of a final l with g (= G) in a few nouns, e.g. puril ‘children’ (historically a plural form in -l) : PX 1SG purig-bi : 2SG purik-ci : 3SG puriƞ-ni. Several phenomena are connected with the behaviour of the unmarked nasal n. When occurring as a final segment of a lexical root, it can be lost before suffixes either for lexical or phonotactic reasons. Phonotactic reasons require the deletion of n before the suffix-initial segments n m s l. As a result, the clusters *nn *nm *ns *nl are absent in the language. Before suffixes beginning with an obstruent a final n assimilates to the place of articulation of the obstruent, e.g. xaosan ‘paper’ : ACC xaosam-ba, jɪlgan ‘voice’ : PROPR jɪlgaƞ-ko. In these cases, the assimilation could, of course, again be analysed as automatic and involving an unmarked “archiphonemic” nasal (= N). Nanai prosody has not been studied in detail, but it is probably correct to say that the primary expiratory stress falls on the initial syllable of the word, though there may be a raised pitch on a later syllable, especially on the penultimate “mora”, as in several other Tungusic languages. However this may be, prosody is non-distinctive in Nanai. WORD FORMATION Inflection and derivation in Nanai depend entirely on suffixes. Examples of reduplication are almost absent, except in onomatopoeia and a few other cases, e.g. goɪ ‘different’ : REDUPL goɪ&goɪ ‘various’. There are almost no compounds, or words derived from compounds. As in other Tungusic languages, there is a basic morphological distinction between nominals and verb(al)s, with the class of nominals comprising regular nouns, spatials (locational nouns), adjectives, numerals, and pronouns. Derivational morphology follows this division and allows derivational forms to be classified into the following main types: (1) Denominal nouns: This group comprises, above all, the productive category of alienable possession, but also diminutives and a number of other derivational forms. •

•

Alienable possession is marked by the suffix -ƞgO-, which always follows other derivational suffixes and precedes case markers and possessive suffixes, e.g. naɪ jɪlɪ-nɪ ‘a person’s (own) head’ (person head-PX3SG) vs. naɪ jɪlɪ-ƞgo-nɪ ‘a head that a person has (such as the head of an animal)’ (person head-ALPX3SG). Adjectives used as nouns are normally marked for alienability when combined with possessive suffixes, e.g. turge-ƞgu-ji-e-ni moco-go-xaarɪ ‘return quickly afterwards!’ (quick-​AL-INSTR-OCX-PX3SG return-REVIMP.FUT.2SG). As may be seen, alienable possession is a category that is positioned between derivation and inflection. Its definition as a derivational category is supported by the fact that it forms a secondary stem that takes the actual inflectional markers for case and possession. Diminutives are formed by the suffix -kAAn, e.g. joo/g ‘house’ : DIM joo-kaan ‘little house’. When attached to numeral bases, the diminutive suffix conveys a limitative meaning, e.g. juer ‘two’ : LIM juer-keen ‘only two’. Unlike the Ewenic languages, Nanai does not have augmentative nouns, though augmentatives can be formed from adjectives.

Nanai 371

•

•

Other productive suffixes deriving nouns from nominals are -ƞkAn [inhabitants of a certain place], e.g. ɪxon ‘village’ : ɪxo-ƞkan ‘villager’, boa ‘land, place’ : bue boa-ƞkAn ‘a person from our place’ (1PL land-DX); -ksA [skin of an animal], e.g mapa ‘bear’ : mapa-ksa ‘bear skin’; -lAA [property nouns from adjectives in combination with a possessive suffix], e.g. daayɪ ‘big’ : DX-PX3SG daayɪ-laa-nɪ ‘size’, gogda ‘high’ : DX-PX3SG gogda-​laa-nɪ ‘height’. A non-productive suffix, used in a number of lexicalized items is -ptOn [coverings], e.g. nasal ‘eyes’ : nasa-pton ‘spectacles’. As in other Tungusic languages kinship terms often show exceptional stem alternations, with separate forms of address and reference. The address forms or “vocatives” typically end in the long vowel -AA, e.g. amɪn- ‘father’ : VOC am-aa, enin ‘mother’ : VOC eñ-ee, aag- ‘elder brother’ : VOC ag-aa, eike‘elder sister’ : VOC eg-ee.

(2) Deverbal nouns: The principal class of deverbal nouns is formed by the participles, which, however, belong to the verbal paradigm. Other deverbal nouns, with no verbal properties, may be classified into action nouns, actor nouns, and instrumental nouns: • • •

Action nouns are formed by the suffix -n, e.g. bi- ‘to be, to live’ : bi-n ‘being, existence, life’, jobo- ‘to work’ : jobo-n ‘work’. Actor nouns are formed by the suffix -mjI, e.g. aloosɪ- ‘to teach’ : aloosɪ-mjɪ ‘teacher’, bota- ‘to fish’ : bota-mjɪ ‘fisherman’. Instrumental nouns are formed by the suffix -kO, e.g. xado- ‘to mow (hay)’ : xado-ko ‘scythe’, japalɪ- ‘to grasp’ : japalɪ-ko ‘handle’.

(3) Denominal verbs: Some nominal stems can also be used in the function of verbal stems without any derivational suffix, meaning that they are nomina-verba, which are ambivalent with regard to the part of speech, e.g. tugde ‘rain’ : tugde- ‘to rain’, xedun ‘wind’ : xedun- ‘to blow (of wind)’, jarɪ ‘song’ : jarɪ- ‘to sing’, nɪƞmaan ‘folktale’ : nɪƞmaan- ‘to tell a tale’. Nomina-verba often yield constructions in which the predicate and the subject or the object are lexical and etymological cognates, e.g. tugde tugde-i-ni ‘it rains’ (rain rain-PRS-3SG), jarɪ-wa jarɪ- ‘to sing a song’ (song-ACC sing). In other cases, the conversion of nominals to verbs takes place by suffixes: •

•

•

The principal verbalizing suffix is -lA-, which is used in a variety of functions, e.g. [instrument] gɪda ‘spear’ : gɪda-la- ‘to stick, to thrust (a spear)’; [object] aapon ‘hat’ : aapo-la- ‘to put on a hat’; [result] coƞdo- ‘pit’ : coƞdo-la- ‘to make a pit’; [purpose] mue ‘water’ : mue-le- ‘to fetch water’; [role] cowoo ‘thief’ : cowoo-la- ‘to steal’; [social union] asɪ ‘wife’ : asɪ-la- ‘to take a wife, to marry (of men)’, eji ‘husband’ : eji-le- ‘to take a husband, to marry (of women)’. Translative verbs are derived from adjectives by the suffix -nA-, e.g. daayɪ ‘big’ : daayɪ-na- ‘to become big’. In some items, this suffix is also attached to nouns, e.g. juke ‘ice’ : juke-ne- ‘to be covered by ice’, jolo ‘stone’ : jolo-na- ‘to become petrified’, abaa [negative existential noun] : aba-na- ‘to disappear’. Sensive verbs are derived from adjectives by the synchronically non-productive suffix -sI-, e.g. uleen ‘good’ : ulee-si- ‘to regard something as good’ > ‘to like’, orkɪn ‘bad’ : orkɪ-sɪ- ‘to regard something as bad’ > ‘to dislike’.

372 SHINJIRO KAZAMA

(4) Deverbal verbs: Most types of deverbal verbs еxpress the grammaticalized categories of voice, aspect, and mood and are better discussed in connection with verbal conjugation. The following is a selection of a few other derivational types: • •

•

Andatives are formed by the suffix -(I-)ndA-, e.g. ice- ‘to see’ : AND ice-nde‘to go see’, tee- ‘to sit’ : tee-nde- ‘to go to sit’. In the “Lower Amur” dialects, this suffix has the form -ƞdA-, as also in Ulcha. Pretentives are formed by the suffix -kAAcI-. These forms express playful action or pretense, e.g. aksa- ‘to be offended’ : PRET aksa-kaacɪ- ‘to pretend to be offended’, susuen- ‘to ski, to sled’ : PRET susue-keeci- ‘to have fun skiing or sledding’, meu‘to perform a shamanic rite’ : PRET meu-keeci- ‘to play the role of a shaman’. Transitives are formed from intransitives by the suffix -bO- ~ -wO-, which can also form causatives from transitive verbs. Typically, the intransitive bases in these cases are motion verbs, e.g. ii- ‘to enter’ : CAUS ii-wu- ‘to put/bring in’, nie- ‘to go out’ : CAUS nie-wu- ‘to put/take out’, agbɪn- ‘to appear’ : CAUS agbɪm-bo- ‘to take out’, tuu- ‘to fall’ : CAUS Naikhin tug-bu- ~ Bikin tubgu‘to drop’, too- ‘to go up from the riverside’ : CAUS too-bo- ‘to take up from the riverside’, eu- ‘to go down to the river’ : CAUS ee-wu- ‘to take down to the river’, xaa- ‘to dock’ : CAUS xaa-bo- ‘to bring something onshore from a boat’. The transitivized verbs of motion are used with a regular object in the accusative case, cf. e.g. (intransitive) tei naɪ jook-cɪ-yɪ ii-gu-xe-ni ‘he went into his (own) house’ (that person house-DIR-RX enter-REV-PST-3SG) vs. (transitive) tei naɪ moo-wa jook-cɪ-yɪ ii-wu-gu-xe-ni ‘he brought the firewood into his house’ (that person wood-ACC house-DIR-RX enter-CAUS-REV-PST-3SG).

NUMBER AND CASE In the nominal template, the derivational suffixes, with the marker of alienable possession in the last position, are followed by the markers for number, case, and possession (N-DX-AL- NX-CX-PX). The plural marker is -sAl, taken by both animate and inanimate nouns. Most commonly, this suffix is used with nouns denoting humans, e.g. naɪ ‘man, person’ : PL naɪ-sal ‘people’, but it also occurs with other types of nouns, e.g. xulu ‘squirrel’ : PL xulu-sel, moo ‘tree’ : PL moo-sal. Exceptionally, the suffix can have the form -sIl, e.g. jɪa ‘friend, fellow’ : PL jɪa-sɪl. Some nouns have an irregular stem form used before the plural marker: ekte ‘woman’ : PL ek-sel, gurun ‘(mass of) people’ : PL gur-sel, elci ‘slave’ : PL. elciu-sel, pujin ‘heroine’ : PL pujiu-sel. The primary plural marker *-l is attested only in relicts, e.g. puril ‘children’ (= PL puri-l), nasal ‘eye/s’ (= PL nasa-l), but also in verbal conjugation as the third person plural personal ending, while the marker *-r (for nouns ending in *-n in the singular basic form) is preserved in several clan names and ethnonyms, e.g. samar (= PL sama-r ‘shamans’), kiler ‘Ewenki’ (= PL kile-r). Nouns modified by numerals and other quantifiers can be marked for plural, e.g. jue ek-sel ‘two women’ (two woman-PL), egji gur-sel ‘many people’ (much people-PL), though this concord is not obligatory. The plural suffix can also be attached to adjectives used without a following head noun, as in [1]: [1] eike-ni elder.sister-px3sg

tama-xan gather.ptcp.prf

gaakta-ƞgo-nɪ berry-al-px3sg

daayɪ-sal big-pl

ulee-sel good-pl

Nanai 373

bi-ci-ni be-pst-3sg ‘The berries gathered by the elder sister were big and good ones.’ Associative plurality is marked by the suffix -nAA, e.g. amaa- ‘father’ : PL ASSOC amaa-naa ‘father and others’. A small group of lexicalized collective nouns is formed by the suffix -jOAn, e.g. nuuci-keen-juen ‘children’ (small-DIM-COLL). The case system in Nanai comprises six primary suffixally marked cases, which may be identified by their conventional names as the accusative, designative, dative, locative, directive, and instrumental (Table 14.3). TABLE 14.3 NANAI CASE MARKERS  

V

C

ACC

-wA

-bA

DES

-gO-

DAT

-dO

LOC

-lA

DIR

-cI

INSTR

-jI

-dO-lA

Stem types: V = vowel stems, C = consonant stems. Segmental alternations: A = a e, O = o u, I = ɪ i.

As in the other languages of the Nanaic group, the system is less diversified than those attested in the Ewenic languages. Also, the suffix allomorphy of the case markers is limited to the impact of vowel harmony, as well as a couple of rather minor differences between vowel stems and consonant stems. Most consonant stems end in the nasal n, but there are also stems ending in the lateral l, e.g. gɪol ‘oar’, xapol ‘broom’, the vibrant r, e.g. juer ‘two’, yeƞgur ‘wolf’, as well as the weak velar stop g (positionally alternating with k or ƞ), e.g. aag- ‘elder brother’. At least some of the stems ending in l and r are petrified plural forms. A special type of allomorphy is exhibited by the limited group of stems in which the final l of the basic form alternates with g in the oblique forms, e.g. derel ‘face’ : OBL dereg- (positionally also derek-, dereƞ-). An exceptional item is joo = joo/g- ‘house’, which in the oblique forms has the alternant stems joo- and joog- (positionally jook-, jooƞ-), e.g. ACC joo-wa ~ joog-ba : LOC joo-la ~ joog-dola : DIR joo-cɪ ~ jook-cɪ. When combined with the possessive suffixes of the second and third persons, the case markers ending in the vowels O I are followed by an additional vowel segment -A- = -a~ -e-. This vowel segment has no obvious meaning of its own, and it could be analysed as a semantically void epenthetic element (E). Alternatively, it could be understood as part of the allomorphy of either the case markers or the possessive suffixes. Here, the analysis will be adopted that it is recognized as a separate element that functions as an “oblique coaffix” (OCX) between the case marker and the possessive suffix, e.g. jakpa ‘side’ : DAT jakpa-do : DAT-OCX-PX3SG jakpa-do-a-nɪ ‘at the side of’. No such element is overtly present after the case markers ending in the vowel A, e.g. LOC-PX3SG jakpa-la-nɪ ‘to the side of’. The “oblique coaffix” is a specific feature of the “Middle Amur” dialects of Nanai, and it is absent in the other dialects, as well as in Ulcha.

374 Shinjiro Kazama

From the functional point of view, the cases may be divided into the categories of grammatical, local, and modal cases. The grammatical cases (proper) include the accusative and designative, as well as the unmarked nominative. The local cases are the dative, locative, and directive, while the instrumental functions as a modal case, though ultimately, all cases also have some grammatical functions. •

The unmarked nominative is the case of the subject. It also functions as a nominal predicate in combination with the copular verbs bi- ‘to be’ and osɪ- ‘to become’, e.g. neu-ni esi daayɪ ekte osɪ-xa-nɪ ‘now her younger sister has become a big woman’ (younger.​sibling-​PX3SG now big woman become-PST-3SG). In causative sentences, the complement remains in the nominative, e.g. sim-bie jaƞgɪan osɪ-gowaan-dam-bɪ ‘I will make you a director’ (2SG-ACC director become-REV-CAUSAOR-1SG). Since there is no adnominal genitive case, the nominative, combined with the possessive form of the head noun, also expresses the possessor e.g. tei naɪ gerbu-ni ‘that persons’s name’ (that person name-PX3SG). • The accusative in -wA (after vowels) ~ -yA (after i ɪ) ~ -bA (after consonants) ~ -pA (in some spatials and pronouns) marks the direct object of transitive verbs. Occasionally, the object may remain unmarked, but this does not imply indefiniteness, as, for instance, in Mongolic and possibly in Solon. Some verbs like ɪsɪ- ‘to reach, to arrive’ can be used both transitively with the object in the accusative and intransitively with the goal in the locative or directive cases. In emotional expressions adjectives can take what looks like the accusative marker without any object function, e.g. xeree uleem-be xuse pikte bi-ci-ni ‘there was a very good boy’ (INTERJ good-ACC male child be-PST-3SG). • The designative in -gO-, which corresponds to the partitive (partitive-designative) in the Ewenic languages, is used only in combination with possessive suffixes. This case indicates that something is designated for the referee of the possessive suffix. Syntactically, the noun in the designative form functions as the object of a transitive verb, but it can also co-occur with an accusative object in a single sentence, e.g. pikte-gu-yi xuse-we baa-xa-nɪ ‘s/he got a boy as his/her child’ (child-DES-RX male-ACC find-PST-3SG). The verbs requiring the designative case are largely confined to what may be called “effective verbs”, such as baa- ‘to find, to get’, gele- ‘to want’, aƞgo- ‘to make’. The objects of these verbs appear as the result of the action. In some cases, however, the designative case can also be used with an intransitive predicate, such as bi- ‘to be’, e.g. bue jɪa-go-po bi-ruu ‘stay (here) in the capacity of our friend!’ (1PL friend-DES-PX1PL be-IMP.2SG). • The dative in -dO indicates the location of an action or existence, e.g. joog-do naɪ arkɪ-wa omɪ-ɪ-cɪ ‘the people drink liquor in the house’ (house-DAT person liquor-ACC drink-PRS-3SG), tei jaka takto-do bi-i-ni ‘that one is in the shed’ (that thing shed-DAT be-PRS-3SG). It can also refer to the location where an object will be after an action is completed, for which reason verbs like nee- ‘to put’, loo- ‘to hang’, tee- ‘to sit down’, doo- ‘to perch’, and xaa- ‘to dock (a boat)’ require the dative. With some verbs it refers to an exact point within an entity, e.g. uike-du-e-ni japa-raa ii-xe-ni ‘s/he grabbed the door and entered’ (door-​DAT-OCX-PX3SG grab-CV.ANT enter-PST-3SG). With animate nominals the dative expresses recipient or possessor, e.g. xemtu gurun-du ‘to/for all people’ (all people-DAT), nuuci ekte-keen-sel-du akoan-sal bi-ci-n ‘small girls had puppets’ (small woman-DIM-​PL-DAT puppet-PL

Nanai 375

be-PST-3PL). With participles, the dative forms quasiconverbal constructions with a temporal function, as in [2]: [2] ene-i-du-e-ni go-ptcp.prs-dat-ocx-px3sg ‘When he went, he saw the boy.’ •

ice-xen see-pst

tei that

naonjokaam-ba boy-acc

Compared with the dative, the locative in -(dO)-lA signals a more restricted location. It can, for instance, refer to an exact point within an entity, as in [3]:

[3] tei naɪ mapa-wa mɪawan-dola-nɪ that person bear-acc heart-loc-px3sg ‘He shot the bear in the heart’.

garpa-xa-nɪ shoot-pst-3sg

Since Nanai has no prolative case, the locative also has a prolative funtion, e.g. uike-le ii-xe-ni ‘s/he entered through the door’ (door-LOC enter-PST-3SG), xukte-le-ni seekse xeye-xe-ni ‘blood flew through/from his teeth’ (tooth-LOC-PX3SG blood flowPST-3SG). In some cases the locative may be understood as indicating a locational limit, e.g. mue peiƞen-dule-ni bi-ci-ni ‘the water reached up to his/her knees’ (water kneeLOC-PX3SG be-PST-3SG). •

•

The directive in -cI indicates the direction of motion. It is typically used with verbs like ii- ‘to enter’, or naƞgala- ‘to throw’. It also expresses the addressee of verbs of verbal activity and emotional feelings, such as un- ‘to say’, medesi- ‘to ask a question’, gele- ‘to ask’, aksa- ‘to be annoyed’, cuursi- ‘to dislike’, and ñoxoraan- ‘to bow’. The instrumental in -jI indicates instruments and materials, e.g. xurme-ji gɪdala-xa-nɪ ‘s/he pierced (her finger) with the needle’ (needle-INSTR spear-VBLZPST-3SG). It also has a comitative meaning, often used in combination with the postposition gese ‘together’, e.g. tei naɪ-jɪ gese ji-cim-bi ‘I came with him’ (that person-INSTR together come-PST-1SG). Importantly, the instrumental has also several uses that would normally be expressed by an ablative case. For instance, it indicates a starting point in time, as in [4]:

[4] mii baljɪ-xan-jɪ-yɪ ei meen joog-do-yɪ baljɪ-ɪ 1sg live-ptcp.prf-instr-rx this self house-dat-rx live-prs ‘I have been living in this house of mine since I was born.’ Some emotional verbs, such as ɪlaajɪ- ‘to be ashamed’, ƞeele- ‘to be afraid’, and ayaktala- ‘to get angry’, also require the instrumental case, e.g. xaɪ-jɪ ƞeele-xe-si ‘what are you afraid of?’ (what-INSTR be.afraid-PST-2SG), eji min-ji ayaktala-ra ‘do not get angry at me!’ (PROHIB 1SG-INSTR get.angry-CONNEG). Prevention (anti-​purpose) is expressed by the instrumental case in combination with a negative participle, as in [5]: [5] ui=dee baa-rasɪ-jɪ-a-nɪ paajɪ-kaan nee-uri who=ptcl find-ptcp.neg-instr-ocx-px3sg separate-dim put-impers.prs ‘They have to be separated so that nobody finds them.’

376 Shinjiro Kazama

In addition to the six basic cases listed above there are several other forms that have case-like functions: •

•

• •

Many descriptions of Nanai mention also an actual separative case, marked by the suffix -jIAjI. The separative indicates physical separation and is in this function synonymous with the instrumental case, e.g. jayɪ-jɪ ~ jayɪ-jɪajɪ xuluƞ-ki-ni ‘s/he came out from the boat’ (boat-INSTR ~ boat-SEP come.out-PST-3SG), mue-ci ogda-jɪ-a-cɪ ~ ogda-jɪajɪ-a-cɪ okalɪƞ-kɪ-nɪ ‘s/he fell into the water from the boat’ (water-DIR boat-INSTR-OCX-PX3PL ~ boat-SEP-OCX-PX3PL fall-PST-3SG). The initial part -jIA- of the separative marker may be identified with the secondary spatial element -jIA(-) ‘side’, as in bagɪa : bagɪa-jɪa ‘opposite shore’. However, there is also a diachronic connection with the separative/elative markers in several other Tungusic languages, e.g. Ewenki *-gIIjI. A cognate of the Ewenic ablative in *-DUU-kI is also preserved in Nanai in the form -dOI, which, however, functions only as the marker of a “referential” (comparative) case indicating the reference point of comparison, e.g. tei ɪnda tei pikte-dui nuuci ‘the dog is smaller than the child’ (that dog that child-ABL small). An actual ablative case in -dOkI with a wider range of functions is present in the Gorin dialect. A kind of comitative case is marked by -mOlIA ‘together with’ (cf. Oroch -mUnA), e.g. pikte-mulie baljɪ-xa-nɪ ‘s/he lived with her/his child’ (child-COM live-PST-3SG). Nanai has also the suffix -ƞgI, which forms possessive nouns, e.g. ɪnda ‘dog’ : POSS ɪnda-ƞgɪ ‘a thing for dogs, the dog’s one’, loca ‘Russia’ : POSS loca-ƞgɪ ‘a Russian thing, a thing made in Russia’. This suffix has the scope of a phrase, e.g. goɪ naɪ-ƞgɪ ‘another man’s one’ (other person-POSS).

Spatials (locational nouns) show in Nanai characteristics similar to those observed in other Tungusic languages. In the sentence they are used as local/temporal adverbs, as well as in the role of postpositional head words of constructions in which they define the local or temporal position of an adnominal noun. As head words they can also take possessive marking, e.g. tei jaka takto doo-la-nɪ bi-i-ni ‘that one is inside the shed’ (that thing shed inside-LOC-PX3SG be-PRS-3SG). The spatials have a defective and in several respects irregular paradigm which comprises the locative in -lA-, the spatial directive in -sI (< *-sI), and the exceptional accusative in -pA, used in a prolative meaning, e.g. juli.e- ‘front side’ : LOC-PX3SG juli.e-le-ni ‘before’ : DIR jule-si ‘forewards’ : ACC juli-pe ‘ahead’, xamɪ.a- ‘back side’ : LOC-PX3SG xamɪ.a-la-nɪ ‘behind’ : DIR xama-sɪ ‘backwards’ : ACC xamɪ-pa ‘behind’, uye- ‘upper side’ : LOC-PX3SG uye-le-ni ‘above’ : DIR ui-si ‘upwards’ : ACC ui-pe ‘over’, pegi.e- ‘underside’ : LOC-PX3SG pegi.e-le-ni ‘underneath’ : DIR pei​-si ‘downward’ : ACC pegi-​pe ‘under’, xaƞgɪ.a- ‘other side’ : LOC-PX3SG xaƞgɪ.a-la-nɪ ‘on the other side’ : DIR xaƞgɪ.a-sɪ ‘in another direction’ : ACC xaƞgɪ-pa ‘along another place’. While not all spatials have all these case forms, some have an instrumental form in -jI, used in a separative function, e.g. INSTR xamɪ.a-jɪ ‘from behind’. Many spatials are connected with orientation in the natural environment. For the Nanai, locations and directions are often determined by their relative position to rivers, e.g. solɪa- ‘upper course’ : LOC-PX3SG solɪ.a-la-nɪ : DIR solɪ : ACC solɪ-pa ‘up the river’, xeji.e- ‘lower course’ : LOC-PX3SG xeji.e-le-ni : DIR xeyi : ACC xeji-pe ‘down the river’, duy.e- ‘mountain side’ : LOC-PX3SG duye-le-ni : DIR dui-si : ACC dui-pe ‘away from the river’, way.a- ‘river side’ : LOC-PX3SG waya-la-nɪ : DIR waɪ-sɪ : ACC

Nanai 377

waɪ-pa ‘towards the river’, talgɪ.a- ‘middle of the river’ : LOC talgɪ-la ‘far from the shore in the river’ : ACC talga-wa ~ talgɪ-pa ‘along the middle of the river’. Besides the actual spatials, there are other locational nouns, e.g. tokon ‘center’ and aldan ‘space in between’, which have a regular nominal declension with no special spatial case forms. These may be identified as “secondary spatials”. The suffixal spatial -jIA ‘side’ is also a secondary spatial; it can be added to nouns with a locational or orientational meaning, e.g. perxi ‘west’ : perxi-jie ‘the western side’, but also to some actual spatials or their derivatives, e.g. doo ‘inside’ : doo-kɪa-jɪa ‘inner side’. ADJECTIVES As in the other Tungusic languages, adjectives in Nanai are morphologically nominals, and when used as independent head words they behave like nouns. In adnominal position, they do not agree with the head noun in number or case. They have, however, derivational and morphosyntactic properties that distinguish them from other nominals. The two most important derivational types are the proprietive adjectives, expressing possession, and the degree adjectives, expressing modifications of the strength of the adjectival meaning. Several other derivational forms are more marginal and less productive. •

•

•

Proprietive adjectives are formed by the suffix -kO, e.g. asɪ ‘wife’ : PROPR asɪ-ko ‘having a wife’ > ‘married (man)’. When used in predicative position, proprietive adjectives form a possessive construction, e.g. bue asɪ-ko-po ‘we have wives’ = ‘we are married men’ (1PL wife-​PROPR-1PL), mii=dee neu-ku-yi ‘I too have a younger sister’ (1SG=PTCL younger.​sibling-PROPR-1SG), tei saman uƞcuxuƞ-ku ‘that shaman has a drum’ (that shaman drum-PROPR). When the possessed object has an adjectival or numeral modifier, the latter takes the instrumental suffix, e.g. tei ekte ƞonɪmɪ-jɪ nukte-ku ‘she has long hair’ (that woman long-INSTR hair-​PROPR), ei joo ɪlan-jɪ paawa-ko ‘this house has three windows’ (this house three-​INSTR window-PROPR). This implies that the possession is permanent. Semantically, the possessive construction with a proprietive adjective equals an existential construction with the dative case, e.g. ei joog-do ɪlaan paawa bi-i-ni ‘there are three windows in this house’ (this house-DAT three window be-PRS-3SG). If the possession or existence is temporary, only the latter construction can be used. For the negation of possession or existence (the privative or caritive function) an analytic construction with the privative noun anaa ‘absence, absent’ > ‘without’ is employed, e.g. pikte anaa naɪ ‘a man without a child’ = ‘a childless man’ (child PRIV person). Degree adjectives are formed by the suffixes -lAA [moderation], e.g. egji ‘much’ : MOD egji-lee ‘rather much’; -jImA [contrastive-emphatic], e.g. nuuci ‘small’ : CONTR nuuci-jime ‘(the) smaller (one)’; -OOkAAn [augmentative], e.g. gogda ‘tall’ : AUGM gogd-​ookaan ‘very tall’, -kAAn ~ -kOO-kAAn [diminutive], e.g. otokoo ‘few’ : DIM otokoo-kaan ‘very few’, nuuci ‘small’ : DIM nuuci-kuukeen ‘very small’. Other derivational suffixes forming adjectives include -mA [material, from noun bases], e.g. moo ‘tree, wood’ : moo-ma ‘wooden’; -sI [possessing a property], e.g. amta ‘taste’ : amta-sɪ ‘tasty’, gici ‘cold (water)’ : gici-si id.; -OlI [no synchronic meaning, from adjectival bases], e.g. pakcɪ ‘dark’ : pakcɪ-olɪ id.; -pcI [no synchronic meaning, from adjectival bases] goro-pcɪ ‘old (thing)’, jala-pcɪ ‘ancient (event)’.

378 Shinjiro Kazama

NUMERALS The basic numerals are, for the digits: 1 emun, 2 juer, 3 ɪlan, 4 duyin, 5 toɪƞga, 6 ñuƞgun, 7 nadan, 8 jakpon, 9 xuyun, 10 joan, and for the lower decades: 20 xorɪn, 30 gocɪn, 40 dexi, 50 sosaɪ ~ sosɪɪ. The items for the decades 20–40 are early borrowings from Mongolic, possibly transmitted by Jurchen-Manchu. The numerals for the higher decades are formed by the suffix -IƞgO: 60 ñuƞu-iƞgu, 70 nada-ɪƞgo ~ nad-ɪƞgo, 80 jakpo-ɪngo, 90 xuyu-iƞgu. For the powers of ten, 100 em taƞgo (one hundred), from Manchu, and 1000 em mɪƞgan (one thousand), from Mongolic via Manchu, are used. Numerals are nominals used both as independent head words and as adnominal modifiers. In the latter function, the first two numerals can also have the special short forms 1 em, 2 jue. As independent head words, numerals take the case suffixes of regular nominal declension. The accusative form is used in an approximative function, e.g. joan ‘ten’ : ACC joam-ba ‘about ten’. Ordinals are formed by the suffix -A-cIA ~ -IA-cIA ~ -yA-cIA: jue-ye ~ jue-ye-cie ‘second’, ɪl-ɪa ~ ɪl-ɪacɪa ‘third’, du-ye ~ du-ye-cie ‘fourth’, toɪƞg-acɪa ~ toɪƞg-ɪacɪa ‘fifth’, ñuƞgu-cie ~ ñuƞg-iecie ‘sixth’, nada-cɪa ~ naj-ɪacɪa ‘seventh’, jakpo-cɪa ~ jakp-ɪacɪa ‘eighth’, xuyu-cie ~ xuy-ecie ‘ninth’, joa-ya ~ joa-yacɪa ‘tenth’. For the ordinal ‘first’, the suppletive form boƞgo ‘first, foremost’ is used. Collective numerals are formed by the suffix -tOƞA, e.g. jue-tuƞe ‘the two, both’, ɪlan-toƞa ‘all (the) three’, duyin-tuƞe ‘all (the) four’, etc. For the counting of days, Nanai preserves the Common Tungusic suffix -ltA, e.g. ɪla-lta ‘(for) three days’, also on the pronominal stem xado ‘how many’ : xado-lta ‘(for) how many days?’. The numeral 1 em- can take the contrastive suffix -jImA, em-jime ~ em-jiƞe ‘(the) one of them’. Another derivative is emu-ckeen ‘alone’. In general, especially when compared with the Ewenic languages, the derivational morphology specific to numerals is rather limited in Nanai. Multiplicatives, for instance, are formed analytically from the cardinal and ordinal numerals by the noun modan ‘time’, e.g. juer modan ‘twice’, juecie modan ‘for the second time’. PRONOUNS The pronominal system in Nanai comprises the conventional classes of personal, reflexive, demonstrative, and interrogative pronouns. The personal pronouns (Table 14.4) are mainly in line with the other Tungusic languages. However, like the other Nanaic languages, Nanai lacks the distinction between exclusive and inclusive pronouns in the first person plural. TABLE 14.4 NANAI PERSONAL PRONOUNS  

 

NOM

OBL

ACC

POSS

SG

1

mii

min-

mim-bie

mi-ƞgi

2

sii

sin-

sim-bie

si-ƞgi

1

bue

bun-

bum-bie

bue-ƞgi

2

sue

sun-

sum-bie

sue-ƞgi

PL

Nanai is unique among the Tungusic languages in that the basic form of the first person singular pronoun has an initial m instead of b. At the same time, the oblique

Nanai 379

stem of the first person plural pronoun has an initial b against m in the other Tungusic languages. Since these features are absent even in the other members of the Nanaic group, they must be due to relatively recent morphological generalizations in Nanai. In fact, the Bikin dialect follows the mainstream Tungusic representation and has 1SG bii : OBL min- : ACC mim-be and 1PL buu : OBL muun- : ACC muum-be. In Sikachi-Alyan the first person plural pronoun is mue : OBL mun-, in analogy to the corresponding singular forms. The possessive forms 1SG mi-ƞgi ‘mine’ : 2SG si-ƞgi ‘thine’ : 1PL bue-ƞgi ‘ours’ : 2PL sue-ƞgi ‘yours’ are used as possessive pronouns, which, like other denominal forms in -ƞgI are not true case forms but derivatives with a regular nominal morphology. An actual adnominal genitive with a Proto-Tungusic background is, however, present in the Bikin dialect, which has 1SG mini ‘my’ : 2SG sini ‘thy’ : 1PL muuni ‘our’ : 2PL suuni ‘your’. The oblique case forms of the personal pronouns take the regular case markers, added to the oblique stems. The personal pronouns also have the “ablative” forms 1SG min-dui : 2SG sin-dui : 1PL mun-dui : 2PL sun-dui, which are used to indicate the reference point of comparisons (‘than me’, etc.). There is also a kind of “equative” form in -ieci, 1SG min-ieci : 2SG sin-ieci : 1PL bun-ieci : 2PL sun-ieci (‘like me’, etc.). In previous research it has been claimed that Nanai also has the third person pronoun ñoa- : OBL ñoan-, a cognate of the third person pronoun (*nuƞan ~ *nugan) in several other Tungusic languages. Like its cognates, this pronoun follows the inflectional pattern of taking the third person singular and plural possessive suffixes after the case markers: SG NOM PX3SG ñoa-nɪ : ACC-PX3SG ñoam-ba-nɪ : DAT-OCX-PX3SG ñoan-do-a-nɪ : LOC-PX3SG ñoan-dola-nɪ : DIR-OCX-PX3SG ñoan-cɪ-a-nɪ : INSTR-OCX-PX3SG ñoan-jɪ-a-nɪ, PL NOM ñoan-cɪ : ACC-PX3PL ñoam-ba-cɪ : DAT-OCX-PX3PL ñoando-a-cɪ : LOC-PX3PL ñoan-​dola-​cɪ : DIR-OCX-PX3PL ñoan-cɪ-a-cɪ : INSTR-OCXPX3PL ñoan-jɪ-a-cɪ. However, this is synchronically not a regular third person pronoun but, rather, an obviative third person (“fourth person”) pronoun, used in reference to a newly introduced (“non-salient”) third person in contrast to a contextually dominant (“salient”) third person. Reference to the third person in Nanai is normally made by phrases like tei naɪ ‘that person’ = ‘he’ or tei ekte ‘that woman’ = ‘she’. The corresponding indefinite phrases are em naɪ ‘a man, a person’, em ekte ‘a woman’. There is also an indefinite pronoun poa-nɪ ‘someone’. The reflexive pronoun has the stem meen- ‘(one)self’. The form meen is used adnominally as a reflexive possessive pronoun in the meaning ‘(one’s) own’, e.g. meen aag-bɪ ‘one’s own elder brother’ (REFL.POSS elder-brother-RX). The other forms take the singular or plural reflexive markers after the case suffixes, except in the object form, which has no accusative case marker: SG RX mee-pi : DAT-RX meen-du-yi : LOC-RX meen-dule-yi : DIR-RX meen-ci-yi : INSTR-RX meen-ji-yi, PL RX.PL meepeeri : DAT-RX.PL meen-du-eri : LOC-RX.PL meen-dule-eri : DIR-RX.PL meen-cieri : INSTR-​R X.PL meen-ji-eri. The possessive forms are SG POSS-RX mee-ƞgi-yi : PL POSS-RX.PL mee-ƞgi-eri. In addition, there is the independent emphatic form mene ‘(by) oneself’, e.g. xaɪ-wa=daa xem mene mute-i-ni ‘s/he can do everything by her/himself’ (what-ACC=PTCL all REFL be.able-PRS-3SG). The form mene yields the reduplicated construction men&men ‘everybody by her/himself’, as well as fixed phrases like mene beye-ji ‘by oneself’ (REFL body-INSTR), mene murun ‘as you like’ (REFL thought).

380 Shinjiro Kazama

The demonstrative pronouns are ei ‘this’ (proximal) vs. tei, often pronounced [ti:], ‘that’ (distal). In the oblique forms, which are mainly used in adverbial functions, the stems ñe- (proximal) vs. ca- (distal) are also attested. The case paradigm of the demonstratives shows several exceptional features: ACC ei-we vs. ca-wa : DAT ei-du ~ ñe-du ‘here’ vs. ta-do ~ ca-do ‘there’ : LOC ei-le ~ ñe-le ‘here’ vs. taya-la ‘there’ : DIR eu-si ‘hither’ vs. tao-sɪ ‘thither’ : INSTR eyee-ji ‘from here’ vs. tayaa-jɪ ‘from there’ : SEP eyee-jieji ‘from here’ vs. tayaa-jɪa-jɪ ‘from there’, DER ewe-ƞki ‘from here, from now on’ vs. tawa-ƞkɪ ‘from there, since then’. The two demonstratives are often used in combinations like eye taya ‘here and there’ and DIR eu-si tao-sɪ ‘this way and that’. Another distal form is tui ‘so, that way’, which is often used in combination with the verb ta- ‘to do’, yielding conjunctional constructions like tui ta-raa > totaraa ‘then’ (so do-CV. ANT), tui ta-pɪ=o ‘after a while’ (so do-CV.​CSEC=​PTCL), tui ta-pɪ=mat ‘finally’ (so do-CV.CSEC=PTCL). The basic interrogative pronouns are xaɪ ‘what?’ and ui ‘who?’. The pronoun xaɪ has several related stem forms, including xa-, xaya-, xaa-, xao-, xoo-. The resulting set of question words includes the more or less regular case forms ACC xaɪ-wa ‘what?’ : DES-RX xaɪ-go-yɪ ‘for what?’ : DAT xaɪ-do ‘where?’ : INSTR xaya-jɪ ‘from where? how?’ : DIR xao-sɪ ‘whither?’, as well as the exceptional forms xaa-lɪ ‘when?’, xoo-nɪ ‘how?’, xa-macaa ‘what kind of?’, xa-do ‘how many?’. The stem xaɪ- ~ xaa- is also used as a pro-verb in the meaning ‘to do what?’. Some of the verbal forms are attested as fixed expressions: CV.CONN xaɪ-mɪ ‘why?’, AND-CV.CONN xaa-nda-mɪ ‘in order to do what?’. In both the nominal and the verbal function the stem xaɪ(-) can be used as a dummy element, with the function of giving the speaker time to think of the next word, as in [6]: [6] ca-do eike-ni xaɪ-sal-ba xaɪ-xa-nɪ what-pl-acc do.what-pst-3sg that-dat elder.sister-px3sg ‘At that time her elder sister did those-whatever, well-what-did-she-do, xajon-sal-bɪ loo-kta-go-xa-nɪ tool-pl-rx.sg hang-distr-rev-pst-3sg she hung up the tools.’ The interrogative pronouns are often used in rhetorical questions, e.g. ca-wa xorɪ-orɪ xaɪ maƞga ‘why is it difficult to save him/her? (implying: of course it is easy)’ (thatACC save-PTCP.IMPERS.IMPRF what hard). To express a feeling of astonishment or surprise the pronoun xaɪ can be expanded by the emphatic particle =dAA, e.g. erdeƞge xaɪ=daa ~ xaɪ=daa erdeƞge ‘how strange it is!’ (what=PTCL strange). Another lexicalized form based on xaɪ is xaɪ-loo ‘very’, e.g. xaɪloo guckuli arcokaan bi-ci-ni (very beautiful girl be-PST-3SG), tei xaɪloo uleen naɪ=goanɪ ‘that is a wonderful man’ (that very good person=PTCL). PERSON MARKING Nanai preserves the distinction between possessive suffixes (PX), reflexive suffixes (RX), and verbal predicative endings (VX), but the system has been somewhat reduced (Table 14.5). As in the personal pronouns, there is no separate inclusive form in the first person plural.

Nanai 381 TABLE 14.5 NANAI PERSON MARKERS     SG

PL

PX 

VX 

V

C

AOR

SUBJ

1

-yI

-bI

-bI

-yI

2

-sI

-cI

-sI

3

-nI

RX

-yI

1

-pO

-pO

2

-sO

-sO

3

-cI

-l

RX

-ArI

-Ø -bI

-bArI

Stem types: V = vowel stems, C = consonant stems. Segmental alternations: A = a e, O = o u, I = ɪ i.

The possessive suffixes are used on nouns to express possessor, but also on predicatively used nominals and nominalized verbs to express the subject. The predicative personal endings are used on a limited set of finite forms, including, in particular, the Common Tungusic aorist and a number of related aorist-based derivational forms. Several verbal forms, including the subjunctive in -mcA- are conjugated with a mixed set of personal markers. One feature characteristic of “true” finite forms is, however, that they do not take a person marker in the third person singular, while in the third person plural they take the original plural marker *-l. Altogether, the system is in many respects rather different from that attested in, for instance, the Ewenic languages. A peculiarity of the possessive declension is that the first person singular oblique case forms take an additional final extension element -wA ~ yA after the possessive suffix, e.g. mii jɪlɪ-do-yɪ-wa ‘on my head’ (1SG head-DAT-PX1SG-EXT). The status and diachronic origin of this element is unclear, though it has been suggested (Avrorin) that it might be identical with the “oblique coaffix” (OCX) -A- used in certain case forms between the case marker and the possessive suffixes. This suggestion is, however, questionable in view of the different position of the two elements in the string of morphemes. Synchronically, the sequence PX1SG-EXT could also be analysed simply as the oblique form of the first person possessive suffix, i.e. PX1SG.OBL. This form is also attested in Uilta, but not in the “Lower Amur” dialects of Nanai or Ulcha. Its function is obvious, for it allows the first person possessive forms in -yi (< *-bI) to be distinguished from the otherwise homonymic singular reflexive forms in -yi (< *-bII), cf. e.g. tei naɪ sogdata-ƞgo-yɪ gaajo-xa-nɪ ‘he brought his (own) fish’ (that person fish-AL-RX bringPST-3SG) vs. tei naɪ sogdata-ƞgo-yɪ-wa gaajo-xa-nɪ ‘he brought my fish’ (that person fish-AL-PX1SG-EXT bring-PST-3SG). OTHER WORD CLASSES Most adverbs in Nanai are of a nominal origin. Adjectival nominals can occur as modal adverbs without any marking or, alternatively, in the instrumental case form, e.g. turgen ~ turgen-ji tutu-i-ni ‘s/he runs fast’ (rapid ~ rapid-INSTR run-PRS-3SG). Adjectives can also be used as degree adverbs, e.g. maƞga ‘hard’ > ‘very’, goro ‘far’ > ‘much’, amban

382 Shinjiro Kazama

~ ambaan ‘big’ > ‘somewhat, rather’, e.g. maƞga bayan mapa ‘a very rich grandfather’ (hard rich grandfather), amaa goro orkɪn ‘father is much worse’ (father far bad), ambaan egji-we japa-xa-nɪ ‘s/he got rather many (of them)’ (big many-ACC get-PST-3SG). In many cases it is difficult to determine whether an item is basically an adverb or an adjective. For instance, xeme can mean both ‘silent’ and ‘silently’, e.g. xeme bi-ruu ‘be silent(ly)’ (silent be-IMP.2SG), cf. also the form AL-INSTR-​OCX-​PX3SG xeme-ƞgu-jie-ni ‘silently’. There is, however, a formal difference, in that adverbs cannot be used in adnominal position without a copula. For this reason, for instance, the item baɪ ‘merely, simply, in vain, for free’ (← Chinese bai 白) is best classified as an adverb, cf. e.g. baɪ jobo-xa-mbɪ ‘I worked for nothing’ (simply work-PST-1SG), baɪ bi-i naɪ ‘a simple man’ (simply be-PTCP.PRS man). Items clearly lexicalized as degree adverbs include bajɪ ‘more’, e.g. bajɪ ayaktala-xa-nɪ ‘s/he got more and more angry’ (more get.angry-PST-3SG), cuu ‘most’ (superlative), e.g. cuu daayɪ ‘biggest’ (most big), and teƞ ‘just, only, most’, e.g. teƞ em naɪ ‘only one man’ (just one person), teƞ uleen jaka ‘the best thing’ (just good thing). There are also several adverbial quantifiers, e.g. xem ‘totally, all’ and asoo ‘(not) particularly’ : asoo=​daa ‘(not) at all’, e.g. asoo=daa sɪa-rasɪ-nɪ ‘s/he does not eat at all’ (at.all=PTCL eat-NEG.​PRS-​ 3SG). Temporal adverbs include esi ‘now’, nee ‘immediately’, guci ‘again’, and tul&tul ‘all the time’, but many temporal adverbs are simply nominals in adverbial use, e.g. boƞgo ‘first’ > ‘at first’, ini ‘day’ : REDUPL ini&ini ‘every day’. There is also a group of three aspectual adverbs, comprising ketelee ‘almost’, sɪlaan ‘barely’, xaolɪa ‘somehow’, which are always used in combination with the past tense form of the predicate, e.g. ketelee waa-xa-nɪ ‘s/he almost killed (it)’ (almost kill-PST-3SG), sɪlaan ɪsɪ-go-xa-nɪ ‘s/he barely reached (there)’ (barely reach-PST-3SG), naambɪca-mɪ xaolɪa cɪxala-waaƞ-kɪ-nɪ ‘s/he persuaded (him/her) and somehow s/he made her/him agree’ (persuade-CV.CONN anyway agree-CAUS-PST-3SG). Likelihood is expressed by saɪnaa ~ xaɪnaa ‘probably’, e.g. saɪnaa bui-ki-ni bi-jeree ‘s/he probably died’ (probably die-PST-3SG be-FUT). Another group of auxiliary words that are often of a nominal origin is formed by the postpositions (as understood here). Like the spatials, especially the secondary spatials, many postpositions are simply postpositionally used nouns that can take markers for case and possession. The nominal bases of postpositions are often synchronically transparent, as in taon ‘count’ : PX3SG tao-nɪ ‘every’, e.g. ini taon-do-a-nɪ ‘every day’ (day count-DAT-OCX-​PX3SG), (*)oron ‘place’ : LOC-PX3SG oron-dola-nɪ ‘instead of (somebody)’, e.g. amɪm-bɪ oron-dola-​nɪ amɪ-ƞgo-yɪ gele-xem-bi ‘I looked for somebody to take the place of my father’ (father-PX1SG place-LOC-PX3SG father-AL-RX. SG search-​PST-​1SG). In other cases, the nominal base is synchronically obscured, as in baaroa-nɪ ‘towards’, e.g. arcokaan baaroa-nɪ ji-ci-ni ‘s/he came to the girl’ (girl towards-PX3SG come-PST-3SG). There are, however, also morphologically invariable postpositions, some of which can govern specific nominal cases, suggesting that they are of a verbal origin. Examples are: pule- ‘to be left’ : NOM + CV.CONN pule-mi ‘more than’, e.g. joan pulemi bi-ci-ni ‘there were more than ten’ (ten more.than be-PST-3SG), aca- ‘to meet’ : ACC + aca-pcɪ ‘against’, e.g. baɪgoam-ba acapcɪ ene-uri ‘we should go against the enemy’ (enemy-ACC against go-IMPERS.PRS), dabalɪ- ‘to pass’ : ACC + dabal ‘over’, e.g. porom-bɪ dabal naƞgala-​xa-​nɪ ‘s/he threw (it) over her/his head’ (top.of.head-RX over throw-PST-3SG), keendeli- ‘to turn’ : ACC + keendeli ‘around’, e.g. nakam-ba keendeli akoam-ba deri-gu-xe-ni ‘she spread the dolls around the heated floor’ (heated.floorACC around doll-ACC spread-REV-PST-3SG), siuri- ‘to pass’ : ACC + siur ‘(passing)

Nanai 383

by’, e.g. ñoam-ba-nɪ siur ene-xe-ni ‘s/he went by him/her’ (3P.OBV-ACC-PX3SG by go-PST-3SG), ACC + suƞguree ‘through’, e.g. irgem-be suƞguree solɪ ene-psiƞ-ki-ni ‘s/he started going upstream through the village’ (village-ACC through upstream go-MOM-PST-3SG), DIR + sɪak ‘until’, e.g. irgen due-ci-e-ni sɪak ene-xe-ni ‘s/he went to the end of the village’ (village end-DIR-OCX-PX3SG until go-PST-3SG), INSTR + kamor ‘together with’, e.g. pikte-ji-e-ni kamor puƞne-gu-u-su ‘kick her out together with her child!’ (child-INSTR-OCX-PX3SG together.with kick.​out-REV-IMP-2PL). Among other invariables there is a heterogeneous group of interjections, including gee ‘well, now, come on!’, maa ‘here, take it!’, pede ‘cheer up!’, bala ‘hurry up!’, enene [moan from pain], xeree [exclamation of surprise]. Lexicalized curses include: PX3SG saktɪɪ-​nɪ ‘you fool!’, deregbece ‘damn you!’ < dereg-be acapcɪ (face-ACC against), eni kaato-nɪ ‘fuck you!’ (mother female.genitals-PX3SG). VERBAL MORPHOLOGY The Nanai verbal forms may be divided into the four conventional classes of finite indicative forms, finite imperative forms, participles, and converbs. Preceding the markers of these basic classes the verbal template can contain derivational suffixes marking the categories of voice, aspect, and Aktionsart. The system of verbal stem types in Nanai shows both simplifications and secondary complications as compared with Proto-Tungusic. The original conjugational classes, dividing the verbs into vowel stems, consonant stems, nasal stems, as well as those taking the elements -dA- or -sI- in the aorist, are preserved, but only in traces, while a new distinction has arisen between short-vowel stems and long-vowel stems. As in other Tungusic languages, the differences between the stem types are particularly relevant to the aorist stem formation, which affects also the form of aorist-related participles and converbs. However, in Nanai there are also other morphological differences involved. Altogether, the verbs in Nanai may be classified into the following stem types: (1a) Stems ending in a short vowel (always bisyllabic or longer), e.g. japa- ‘to take’, ice- ‘to see’, soƞgo- ‘to cry’, tutu- ‘to run’, ɪsɪ- ‘to reach’, pukci- ‘to gallop’. In the imperfective participle (based on the aorist) these verbs take the marker -I, which can regressively assimilate the stem-final vowels a e, e.g. ice-i [itɕi:]. (1b) Stems ending in a long vowel, e.g. (monosyllabic) baa- ‘to find’, ñaa- ‘to decay’, paa- ‘to evaporate’, saa- ‘to know’, xaa- ‘to dock’, waa- ‘to kill’, nee- ‘to put’, tee- ‘to sit down’, oo- ‘to get on’, doo- ‘to perch’, poo- ‘to saw’, too- ‘to go up from the riverside’, buu- ‘to give’, tuu- ‘to fall’, ii- ‘to enter’, sii- ‘to pass’, (bisyllabic) boyaa- ‘to break’ TR, kɪƞdaa- ‘to break (long objects)’ ITR, mañaa- ‘to run out’, talaa- ‘to cook sliced raw fish’, puyuu- ‘to cook’. This type also comprises stems ending in a diphthongoid vowel sequence, e.g. (monosyllabic) cɪa- ‘to overturn’, gɪa- ‘to shave with a knife’, pɪa- ‘to whet’, sɪa- ‘to eat’, nie- ‘to go out’, xue- ‘to tan’, xaɪ- ‘to do what?’, ao- ‘to sleep’, dao- ‘to cross’, eu- ‘to go down to the river’, meu‘to perform a shamanic rite’, (bisyllabic) xalɪa- ‘to hook dogs to the dog-​sleigh’, loptoa- ‘to untie’. In the imperfective participle these verbs take the marker -rII, e.g. waa- : waa-rɪɪ. Incidentally, the imperfective participle and related forms help to verify the distinction between sequences with or without an intervocalic glide y, that is, monosyllabic V+I vs. bisyllabic V+yI; thus, for instance, the verb soyɪ- ‘to scold’ : PTCP.IMPRF soyɪ-ɪ is a bisyllabic short-vowel stem.

384 Shinjiro Kazama

(2a) Stems ending in non-nasal consonants. There are very few verbs belonging to this type, and the choice of final consonants is limited to two: the lateral l and the labial stop p. The former group comprises the verbs oral- ‘to echo’, kaal- ‘to stalk’, maal(~ maalbɪa-) ‘to die out’, masal- ‘to squirt’, xool- ‘to sound’, xuul- ‘to scatter, to spill’, and a few others, while the latter comprises the verbs guup- ‘to fade, to go out’, jalop- ‘to fill’, and jep- ‘to eat’ (in the Naikhin dialect largely replaced by sɪa‘to eat’). In the imperfective participle these verbs take the suffix -dII (after l) ~ -cII (after p), e.g. xool- : xool-dɪɪ, guup- : guup-cii. Another feature of the consonant stems is that they take the suffix -kIn in the perfective participle, e.g. xool- : xool-kɪn, while all vowel stems take the etymologically different suffix -xAn, e.g. saa- : saaxan. The sequence p-k undergoes, however, metathesis, e.g. jalop- : (*)jalop-kɪn > jalok-pɪn. (2b) Stems ending in the nasal n, e.g. un- ‘to say’, taon- ‘to count’, toan- ‘to stretch’, agbɪn- ‘to appear’, gɪran- ‘to step’, jɪlgan- ‘to raise voice’, teyin- ‘to rest’, etc. This is a rather large group, which also comprises the causative stems in -wAAn-. In the imperfective participle these verbs take the marker -dII, e.g. un- : un-dii. Like the other consonant stems, these verbs also take the suffix -kIn in the perfective participle, e.g. un- : uƞ-kin. (3) Stems originally belonging to the limited category of “change-of-state verbs”, whose aorist stem was formed by the element *-dA-. In Nanai these verbs survive only in rudiments, the two examples normally quoted being ga- ‘to take’ and ji- ‘to come’. As a trace of their original stem type, these verbs form the imperfective participle, like the consonant stems of types (2a and 2b) by the marker -dII, i.e. ga-dɪɪ, ji-dii. In the perfective participle they take the suffix -cIn, i.e. ga-cin, ji-cin. There are, however, also two other verbs that diachronically belong to this same type, and which still synchronically share similar properties, though with irregularities. These are o- ‘to become’ and bu- ‘to die’. The verb o- follows basically the same pattern as the ga- and ji-, in that it has the expected participial forms IMPFR o-dɪɪ and PRF o-cɪn. However, it also has the alternative stem o-sɪ-, which is conjugated like a regular bisyllabic vowel stem. The verb bu-, on the other hand, shows several irregular features. Although it has the imperfective participle form bu-dii, it has the participial forms PTCP.IMPERS.IMPRF bur-buri : PTCP.PRF bui-kin. Because of these features it has also been classified with the consonant stems of type (2a). (4) Stems originally taking the element *-sI- in the aorist. Nanai preserves only two items of this type: the copula bi- and the negation verb e-. The copula bi- has, however, lost the aorist in -sI- and forms the imperfective participle as bi-i [bi:], while the perfective participle follows the model of type (3), yielding bi-cin. The aorist form of the negation verb, which originally preceded the invariable connegative form of the main verb, has been suffixalized and now follows the main verb. Even so, the suffixalized negation verb retains its original stem type and appears in the shape -A-sI-, e.g. saa-r-a-sɪ-nɪ ‘s/he does not know’ (know-​CONNEG-NEG-PRS3SG). The corresponding past tense form is marked by -A-cI-, e.g. saa-r-a-cɪ-nɪ ‘s/ he did not know’ (know-CONNEG-NEG-PST-3SG). In these forms, it is difficult to determine where the morpheme boundary between the connegative suffix and the suffixalized negation verb synchronically goes, but it is likely that the vowel element -A- in the suffix is a direct trace of the stem of the negation verb (*-e-). It is also somewhat unclear whether the negation form of present tense reference is based on the original finite aorist form of the negation verb (*e-si-) or on the

Nanai 385

corresponding imperfective participle (*e-si-i). Synchronically, the two forms seem to have merged in Nanai, cf. e.g. PTCP.IMPRF.NEG saa-r-a-sɪ ‘(one) who does not know’ (< *saa-ra+e-si-i). An ambiguity is also indicated by the fact that in predicative use, the suffixalized negation verb can in some forms take two alternative types of person marking (PX and VX). From the point of view of suffix allomorphy, the types (1a) and (1b), as well as (3), which end in a vowel in the basic stem form, may be collectively identified as “vowel stems”, while the types (2a) and (2b), which end in a consonant, may be identified as “consonant stems”. This difference is also relevant for a number of derivational suffixes. The verb bu- : bur- : bui- ‘to die’ is in most cases treated as a consonant stem. VOICE AND ASPECT Voice and aspect are marked by derivational suffixes on the verbal base. There are three voices: the causative, reciprocal, and medial. •

The causative is marked by the suffix -wAAn- (after vowel stems) ~ -bO-wAAn(after consonant stems). The basic function is the causative, e.g. elci-we ene-weeƞki-ni ‘s/he made the slave go’ (slave-ACC go-CAUS-PST-3SG), but a passive connotation is also present, e.g. eji saa-waan-da ‘don’t let (it) be known’ (PROHIB know-CAUS-CONNEG). If the verb is transitive, the causee appears in the accusative case, and the sentence may contain two accusatives, e.g. ñepulte-keem-be xaolɪa mim-bie waa-waan-doo ‘please let me hunt for a small boar’ (boar-DIMACC somehow 1SG-ACC kill-CAUS-IMP.2SG). The causative can also form transitives from intransitive bases, e.g. tei mue-we xeye-weeƞ-ki-ni ‘s/he poured the water’ (that water-ACC flow-CAUS-PST-3SG). In some contexts, it can indicate an unintentional action, e.g. moo-wa ta-mɪ nasal-cɪ-yɪ toɪko-waaƞ-kɪm-bɪ ‘when I was making firewood, I accidentally hit my eye’ (wood-ACC do-CV.CONN eyeDIR-RX hit-CAUS-​PST-1SG). The causative can also be used in order to keep the subject of a subordinate clause the same as in the main clause, as in [7]:

[7] tee siun ui-si too-waan-daa there sun above-dir rise-caus-cv.ant ‘Then, after the sun had risen, baarajɪ-go-raa ene-i=goanɪ get.ready-rev-cv.ant go-prs=ptcl he got ready and started to go.’ In the above type of construction, the causative can also function as an adversative passive, e.g. amɪm-bɪ bur-buween-dee ñoa-nɪ emuckeen baljɪ-ɪ-nɪ ‘after his father died he is living alone’ (father-RX die-CAUS-CV.ANT 3P.OBV-PX3SG alone live-PRS-3SG). •

The reciprocal is marked by the suffix -mAAcI-, e.g. sorɪ- ‘to fight’ : RECIPR sorɪ-maacɪ- ‘to fight each other’, ojo- ‘to kiss’ : ojo-maacɪ- ‘to kiss each other’. However, some verbs in this form have lexicalized meanings, e.g. gele- ‘to want’ : RECIPR gele-meeci- ‘to look for’, ice- ‘to see’ : RECIPR ice-meeci- ‘to look for something carefully’.

386 Shinjiro Kazama

•

• • •

• • • •

•

The medial is marked by the suffix -p-tA-, e.g. xuede- ‘to lose’ : xuede-pte- ‘to be lost, to disappear’. Diachronically, the suffix -p-tA- consists of the original medial (medio-passive) marker -p- and the aorist marker -tA- (< *-rA-), as is still obvious from the morphological behaviour of the verbs of the category: the imperfective participle is based on the aorist stem in -p-tA-, e.g. xuede-pte-i, while the perfective participle is based on the simple stem in -p-, e.g. xuede-k-pin < *xuede-p-kin. Thus, these verbs are morphologically ambiguous and have properties of both vowel stems and consonant stems. The derivational markers for aspect include those for inchoative, momentative, durative, stative, iterative, distributive, habitive, and reversive action. The inchoative in -lO- marks the beginning of an action, e.g. soƞgo-lo-xa-nɪ ‘s/he started to cry’ (cry-INCH-PST-3SG). The momentative in -psIn- refers to the sudden start of an action. e.g. seekse xeyepsiƞ-ki-ni ‘blood starting streaming’ (blood flow-MOM-PST-3SG). The durative in -AcI- ~ -cI- indicates prolonged action, e.g. degde- ‘to fly’ : DUR degde-eci-, puyuu- ‘to cook’ : DUR puyuu-ci-, ñoxoraan- ‘to bow’ : DUR ñoxoraa-cɪ-. A durative meaning can also be expressed by repeating the verb form, e.g. ene-i=dee ene-i ‘(s/he) goes and goes’ (go-PRS=PTCL go-PRS). The stative in -sI- indicates static action, e.g. tee- ‘to sit’ : STAT tee-si- ‘to be sitting’, ɪlɪ- ‘to stand’ : ɪlɪ-sɪ- ‘to be standing’. In some verbs, the suffix -sI- has been petrified as a part of the stem, e.g. enu-si- ‘to be ill’, cf. enuu ‘illness’. The iterative in -nAsI- indicates an action repeated many times or over a long period of time, e.g. arakɪ-wa omɪ-nasɪ-ɪ-nɪ ‘s/he drinks liquor many times ~ for a long time’ (liquor-ACC drink-ITER-PRS-3SG). The distributive in -ktA- indicates an action repeated separately several times, e.g. boa-la xajom-bɪ nee-kte-gu-xe-ni ‘s/he put her various tools outside’ (land-LOC tool-RX put-DISTR-REV-PST-3SG). The habitive in -lsI- indicates habitual action, e.g. xuede- ‘to lose’ : HAB xuedelsi- ‘to lose all the time’. In moo-la- ‘to fetch firewood’ : HAB mool-sɪ- and muele‘to fetch water’ : HAB muel-si- the suffix is haplologically shortened to -sI-, e.g. moo-wa mool-sɪ-ɪ, mue-we muel-si-i ‘(s/he) is always fetching firewood and water’ (wood-ACC fetch.wood-​H AB-PRS water-ACC fetch.water-HAB-PRS). The reversive (repetitive-reversive) in -gO- (< *-rgU-) indicates an action performed repeatedly or in a reverse direction, e.g. japa- ‘to take’ : guci japa-goxa-nɪ ‘s/he took it back again’ (again take-REV-PST-3SG), degde- ‘to fly’ : xama-sɪ degde-gu-xe-ni ‘(it) flew back’ (back-​DIR fly-REV-PST-3SG). Some verbs have irregular forms, e.g. ji- ‘to come’ : REV ji-ju- ‘to come back’, baa- ‘to find’ : REV bao-go- ‘to find again’, nee- ‘to put’ : REV nee-ku- ‘to put back’, while others have a lexicalized meaning, e.g. buu- ‘to give’ : REV buu-gu- ‘to return’, saa- ‘to know’ : REV saa-go- ‘to regain consciousness’. The suffix -gO- is also taken by the verbally used terms for the times of the day and seasons, e.g. sikse- ‘(to become) evening’ : sikse-gu-xen ‘it became evening (again)’ (evening-REV-PST), joa ‘(to become) summer’ : joa-go-xan ‘it became summer (again)’ (summer-​R EV-PST).

PARTICIPLES Nanai uses actively two basic participial forms, here termed the imperfective and perfective participles. Diachronically, the imperfective participle represents the Common Tungusic extension (*-rA-I ~ *-dA-I ~ *-sA-I) of the aorist, well known also from the

Nanai 387

Ewenic languages, while the perfective participle is a feature absent in the Ewenic languages and is manifested by three apparently suppletive suffixes (-xAn, -kIn, -cIn), distributed according to the stem types. The participles have personal (active) and impersonal (passive) forms, with the latter containing the reflexive-passive suffix -bO(w)- ~ -pO(w)- ~ -O- ~ -w-, as well as negative forms, based on the combination of the connegative form of the verb with the suffixalized negative verb -A- and its participial forms (Table 14.6). TABLE 14.6 NANAI PARTICIPLES  

 

V

VV

C

PERS

IMPRF

-I

-rII

-JII

PRF

-xAn

NEG-IMPRF

-A-sI

-r.A-sI

-D.A-sI

NEG-PRF

-A-cIn

-r.A-cIn

-D.A-cIn

IMPRF

-O-rI

-BO-rI

PRF

-O-xAn

-BO-xAn

NEG-IMPRF

-w-A-sI

-BOw-A-sI

NEG-PRF

-w-A-cI

-BOw-A-cI

IMPERS

-kIn

Stem types: V = short-vowel stems, VV = long-vowel stems, C = consonant stems. Segmental alternations: A = a e, O = o u, I = ɪ i, B = b p, D = d t, J = d c.

The verbs ga- ‘to take’, ji- ‘to come’, o- ‘to become’, bu- ‘to die’, which originally had the element -dA- in the aorist, as well as the copula bi-, which originally had the element -sI- in the aorist, have in many respects idiosyncratic participial forms: PTCP.IMPRF ga-dɪɪ, ji-dii, o-dɪɪ, bu-dii, bi-i; PTCP.PRF ga-cɪn, ji-cin, o-cɪn, bui-kin, bi-cin; NEGPTCP.IMPRF ga-d-a-sɪ, ji-d-e-si, o-d-a-sɪ, bu-d-e-si, bi-e-si; NEG-PTCP.PRF ga-da-cɪ, ji-d-e-ci, o-d-a-cɪ, bu-d-e-ci, bi-e-ci; PASS-PTCP.IMPRF ga-o-rɪ, ji-u-ri, o-po-rɪ, bur-bu-ri, bi-u-ri; PASS-​PTCP.PRF ga-o-xan, ji-u-xen, o-po-xan, bur-bu-xen, bi-u-​xen; PASS-NEG-PTCP.IMPRF ga-w-a-sɪ, ji-w-e-si, o-pow-a-sɪ, bur-buw-e-si, bi-w-e-si; PASS-NEG-​PTCP-PRF ga-w-a-cɪ, ji-w-e-ci, o-pow-a-cɪ, bur-buw-e-ci, bi-w-e-ci. The participles have three functions: as attributive modifiers, as sentential arguments, and as finite predicates. As finite predicates they function as present and past tense forms, respectively (glossed PRS and PST). In this role they normally take person markers of the possessive type, but person marking can also be absent, e.g. ene- ‘to go’ : mii ene-i ‘I go’ (1SG go-PRS) : sii ene-i ‘you go’ (2SG go-PRS) : tei naɪ ene-i ‘he goes’ (that person go-PRS). The paradigms with person markers may be illustrated as follows: ice- ‘to see’ (type 1a, short-vowel stems) : PTCP.IMPRF ice-i : PRS 1SG ice-i-yi : 2SG ice-i-si : 3SG ice-i-ni : 1PL ice-i-pu : 2PL ice-i-su : 3PL ice-i-ci. un- ‘to say’ (type 2b, nasal stems) : PTCP.IMPRF un-dii : PRS 1SG un-dii-yi : 2SG un-dii-si : 3 SG un-dii-ni : 1PL un-dii-pu : 2PL un-dii-su : 3PL un-dii-ci. sɪa- ‘to eat’ (type 1b, long-vowel stems) : PTCP.PRF sɪa-xan : PST 1SG sɪa-xam-bɪ : 2SG sɪa-xa-sɪ : 3SG sɪa-xa-nɪ : 1PL sɪa-xa-po : 2PL sɪa-xa-so : 3PL sɪa-xa-cɪ. un- ‘to say’ (type 2b, nasal stems) : PTCP.PRF uƞ-kin 1SG uƞ-kim-bi : 2SG uƞ-ki-si : 3SG uƞ-ki-ni : 1PL uƞ-ki-pu : 2PL uƞ-ki-su : 3PL uƞ-ki-ci.

388 Shinjiro Kazama

Variation is present in the third person plural of the negative paradigm (NEG-PRS), which can take either the possessive suffix -cI or the plural marker -l, e.g. omɪ-a-sɪ-cɪ ~ omɪ-a-sɪ-l. This may be connected with the dual origin of the negative paradigm from both the simple aorist and the participial form of the negation verb. In the other persons, markers of the possessive set are used: omɪ- ‘to drink’ (type 1a, short-vowel stems) : NEG-PTCP.IMPRF omɪ-a-sɪ : NEG-PRS 1SG omɪ-a-sɪm-bɪ ~ omɪ-a-sɪ-yɪ : 2SG omɪ-a-sɪ-sɪ : 3SG omɪ-a-sɪ-nɪ : 1PL omɪ-asɪ-po : 2PL omɪ-a-sɪ-so : 3PL omɪ-a-sɪ-cɪ ~ omɪ-a-sɪ-l. The negative participle has also an emphatic form in =dAA-sI, incorporating the additional element =dAA-, which is obviously identical with the emphatic particle =dAA, as used also on nominals. The preceding verb is in the connegative form, e.g. pikte-yi waaxam-ba-nɪ em-be=dee naƞgala-a=daa-sɪ ‘(s/he) did not throw away anything that her/ his son had caught’ (child-RX kill-PTCP.PRF-ACC-​PX3SG one-ACC=PTCL throw-​ CONNEG=​EMPH-PTCP.IMPRF.NEG). When a participle is used as a modifier, the head noun may be the subject (conjunct use), but also the object or an oblique argument (disjunct use). In the latter case, the participle is marked for possession, e.g. (subject) ji-cin naɪ ‘the man who came’ (come-​ PTCP.PRF person), (object) ekte naɪ japa-ɪ-nɪ jaka ‘the things that women use’ (woman person take-PTCP.IMPRF-PX3SG thing), (place) tui gusereendu-xem-beri beun-due-ni ‘at the place where they were talking’ (so talk-PTCP.PRF-RX.PL place-DAT-OCXPX3SG), (other) naɪ meu-rii(-ni) sɪasɪn ‘the sound of a man perfoming a shamanic rite’ (person perform.​shamanic.rite-PTCP.IMPRF-PX3SG sound). When a participle is used as a sentential argument, it takes the regular nominal markers for case and possession, e.g. sɪrɪ-xan-jɪ-yɪ agbɪƞ-go-ɪ-nɪ ‘she came out from where she had hidden’ (hide-PTCP. PRF-INSTR-RX appear-REV-PRS-3SG). The impersonal participle has a broad meaning and may express: (1) proper action (‘should’), e.g. xoonɪ ta-orɪ ‘what should we do?’ (how do-IMPERS.PRS), tei naɪ-ya uyun bi-weem-buwesi ‘we should not let that man live’ (that man-ACC alive be-CAUS-IMPERS. NEG.PRS); (2) possibility (‘can’), e.g. pikte-yi bao-go-orɪ=m ecie murci-e ‘I did not think I could find my child again’ (child-RX find-REV-PTCP. IMPERS.PRS=PTCL NEG.​PST think-CONNEG), bue xoonɪ saa-orɪ ‘how can we know?’ (1PL how know-​IMPERS.​​​PRS), xaosɪ=daa ene-wesi ‘no one can go anywhere’ (where=PTCL go-​IMPERS.NEG.PRS); (3) passive meaning, e.g. aag-bɪ xorɪ-go-​orɪ-wa gele-mi soŋgo-ɪ-nɪ ‘she cried wanting her brother to be raised from the dead’ (elder. brother-RX revive-REV-​PTCP.​IMPERS.IMPRF-ACC want-CV.​CONN cry-PRS-​3SG); (4) usage or purpose, e.g. naɪ-cɪ pulsi-uri tetue-sel ‘clothes for visiting someone’ (person-DIR go.around-PTCP.​IMPERS.IMPRF clothes-​PL). Because of its functional flexibility, the impersonal imperfective participle is conventionally used as the dictionary form of Nanai verbs. It can also take the third person possessive marker -nI, though the exact conditions when this happens remain still to be clarified. As a marginal feature, Nanai also has the synchronically isolated passive resultative participle in -cA, as also attested in Oroch. As in Oroch, this form is in Nanai used mainly as a sentential argument, e.g. asɪ-nɪ nukte-yi pacɪ-ca-wa-nɪ caalɪ-xa-nɪ ‘his wife cut off one of her braids’ (wife-PX3SG hair-RX plait-PTCP.RES-ACC-PX3SG cut-PST-3SG). More rarely, it is attested as an adnominal modifier, e.g. ulpi-ce xukue-keen ‘a small sewn bag’ (sew-​PTCP. RES bag-DIM). Altogether, -cA looks synchronically more like

Nanai 389

a derivational suffix, which, moreover, is relatively seldom used. A somewhat similar suffix is -ktO, which produces adjectival nominals with a completive, resultative, or anticausative meaning, e.g. nixeli- ‘to open’ : nixeli-ktu ‘open’. As adjectives, these forms can also be used in the role of predicates, e.g. paawa nixeli-ktu ‘the window is open’ (window open). FINITE INDICATIVE FORMS The participle-based finite forms are doubled by an analogous set of monofunctional finite forms, which, unlike the participles, cannot take case markers or be used in adnominal position. Like their participial analogues, the finite forms distinguish between present tense and past tense reference. Diachronically, the form with the present tense reference represents the Common Tungusic aorist (AOR), while the form with the past tense reference has been identified (Benzing) as the “preterite” (PRT). Additionally, there is a future tense, also with a Common Tungusic background. As in other Tungusic languages, the formation of the aorist stem involves morphological complications and is connected with the same verb types as regulate the differences between the participial allomorphs (of which those of the imperfective participle are directly based on the aorist). Thus, the original aorist marker *-rA- appears as -A- after short-vowel stems, e.g. jobo- ‘to work’ : AOR jobo-a-, as -rA- after long-vowel stems, e.g. ao- ‘to sleep’ : AOR ao-ra-, and as -dA- ~ -tA- after consonant stems, e.g. un- ‘to say’ : AOR un-de-. The variant -rA(-) can, however, also be used as an alternative to -A(-) in the third person singular and plural forms. The verbs ga- ‘to take’, ji- ‘to come’, o- ‘to become’, and bu- ‘to die’ preserve their original aorist stem in -dA-, yielding the modern forms ga-da-, ji-de-, o-da-, and bu-de-, respectively. In accordance with its historical origin the aorist takes the person markers of the predicative set (VX), which differ from the possessive set (PX) in the first and second persons singular, as well as in the third person singular and plural, e.g.: omɪ- ‘to drink’ (type 1a, short-vowel stems) : AOR omɪ-a- : 1SG omɪ-am-bɪ : 2SG omɪa-cɪ : 3SG omɪ-a ~ omɪ-ra : 1PL omɪ-a-po : 2PL omɪ-a-so : 3PL omɪ-a-l ~ omɪ-ra-l. bu- ‘to die’ (type 3, synchronically irregular) : AOR bu-de- : 1SG bu-dem-bi : 2SG bu-de-ci : 3SG bu-de : 1PL bu-de-pu : 2PL bu-de-su : 3PL bu-de-l. bi- ‘to be’ (type 4, synchronically irregular) : AOR bi-e- : 1SG bi-em-bi : 2SG bi-e-ci : 3SG bi-re : 1PL bi-e-pu : 2PL bi-e-su : 3PL bi-re-l. In difference from the Ewenic languages, which employ the personally unmarked aorist form for the third person plural, Nanai uses this form for the third person singular, while the third person plural incorporates the plural marker (*)-l, which is preserved in the nominals only in petrified relicts. The plain aorist has also the function of the connegative, which is synchronically present before the suffixalized negation verb, but which is also used independently after the negative forms NEG e-m ‘not’ : PROHIB e-ji ‘do not!’, e.g. eji jobo-ra ‘do not work!’ (PROHIB work-​CONNEG). In the second person plural, the connegative takes, rather unexpectedly, the corresponding person marker, e.g. eji jobo-a-so ‘do not work!’ (PROHIB work-CONNEG-2PL). A similar strategy is used with the form NEG.PST e-cie ‘did not’, which expresses past tense negation in combination with the personal forms of the connegative. In this particular usage, person marking takes place by markers of

390 Shinjiro Kazama

the possessive type, e.g. tei naɪ ecie saa-ra-nɪ ‘he did not know’ (that person NEG.PST know-​CONNEG-3SG). The plain aorist without person marking is also used in certain fixed expressions and constructions, as in questions of the type mii xaɪ saa-ra ‘how can I know it?’ (1SG what know-AOR). A special form is AOR un-de from the verb un- ‘to say’, which has been grammaticalized as a quotative marker used after other finite predicates in the aorist form, e.g. tei naɪ ene-e un-de ‘he goes, they say’ (that person go-AOR say-AOR). This quotative marker is especially common in folktales. The finite preterite is marked by the suffix -kA- for all regular vowel and consonant stems, as well as for bu- ‘to die’ : PRT bui-ke(-), while the verbs ga- ‘to take’, ji- ‘to come’, o- ‘to become’, and bi- ‘to be’ take the suffix -cA-. The latter suffix is apparently identical with the Common Tungusic perfective participle marker *-cAA, as also present in Nanai in the marginally attested (quasi-)participial form in -cA. The preterite is also otherwise closer to participles in that it takes person marking of the possessive type in the first and second persons singular, though in the third person it follows the aorist pattern, e.g.: omɪ- ‘to drink’ (type 1a, short-vowel stems) : PRT omɪ-ka- : 1SG omɪ-ka-yɪ : 2SG omɪka-sɪ : 3SG omɪ-ka : 1PL omɪ-ka-po : 2PL omɪ-ka-so : 3PL omɪ-ka-l. ga- ‘to take’ (type 3) : PRT ga-ca- : 1SG ga-ca-yɪ : 2SG ga-ca-sɪ : 3SG ga-ca : 1PL ga-ca-po : 2PL ga-ca-so : 3PL ga-ca-l. At the same time as participles have become the dominant forms of finite predicates, the actual finite forms have received secondary modal and evidential connotations. In general, the aorist and preterite, which have also been called the “assertive mood” (Avrorin), tend to indicate a stronger statement than a participial form, cf. e.g. sii un-dii-si ‘you are saying’ (2SG say-PRS-​2SG) vs. sii un-de-ci ‘you are saying, aren’t you?’ (2SG say-AOR-2SG), mii uƞ-kim-bi ‘I said’ (1SG say-PST-1SG) vs. mii uƞ-ke-yi ‘I did say’ (1SG say-PRT-1SG). The third person forms normally imply the presence of visual evidence and perhaps surprise (mirativity), e.g. ei-du ao-ra=tanɪɪ ‘oh, s/he is sleeping here!’ (thisDAT sleep-​AOR=PTCL), tei naɪ ca-wa sɪa-ka ‘he ate it, I saw!’ (that person that-ACC eat-PRT). By contrast, the participles express only the factual circumstances, e.g. tei naɪ esi ca-wa sia-rɪɪ-nɪ ‘he is eating it now’ (that person now that-ACC eat-PRS-3SG). The future tense is formed by the Common Tungusic suffix -jA-, originally a marker of progressive action. In Nanai, -jA- remains morphologically a derivational element and the forms containing it are conjugated according to the aorist model: omɪ- ‘to drink’ (type 1a, short-vowel stems) : FUT omɪ-jaa/n- : 1SG omɪ-jaam-bɪ : 2SG omɪ-jaa-cɪ : 3SG omɪ-jaraa : 1PL omɪ-jaa-po : 2PL omɪ-jaa-so : 3PL omɪ-jaraa-l. ga- ‘to take’ (type 3) : FUT ga-jaa/n- : 1SG ga-jaam-bɪ : 2SG ga-jaa-cɪ : 3SG ga-jaraa : 1PL ga-jaa-po : 2PL ga-jaa-so : 3PL ga-jaraa-l. Both the participial present tense and the aorist can also be used with a future reference. The actual future which has also been called the “assumptive mood” (Avrorin), conveys, however, a notion of uncertainty, though the functional difference is very small, cf. e.g. cɪmana mii xoton-ci ene-i-yi ~ ene-em-bi ~ ene-jeem-bi ‘tomorrow I will go to the city’ (tomorrow 1SG city-DIR go-PRS-1SG ~ AOR-1SG ~ FUT.AOR-1SG), in which the variant predicate forms indicate statement, assertion, and assumption, respectively.

Nanai 391

IMPERATIVES Nanai has a full paradigm of personally differentiated imperative forms (Table 14.7). The forms of the first and third persons are marked by the Common Tungusic suffix -gI, to which the regular person markers of the possessive set are added in the third person singular and plural, while the first person singular form retains the original idiosyncatic element -tA. In practice, the first person forms are rarely used in the modern language (with the exception of the Lower Amur dialects, where they are still common today). The second person forms are based on the aorist and are differentiated according to the verb type. The verbs ga- ‘to take’, ji- ‘to come’, o- ‘to become’, bu- ‘to be’, and bi- ‘to be’ have the second person imperative forms ga-doo(-), ji-duu(-), o-doo(-), bu-duu(-), bi-ruu(-), though variants with -s-, e.g. bi-suu(-), have also been attested. An idiosyncratic form is jep- ‘to eat’ (not preserved in active use in the Naikhin dialect) : IMP.2SG jep-uu (= Manchu jef-u). TABLE 14.7 NANAI IMPERATIVE FORMS  

 

V

SG

1

-gI-tA

2

-(rO)O

3

-gI-

1

-gI-tO

2

-O-

3

-gI-

PL

VV

C

-rOO

-dOO

PX

-nI -rOO-

-dOO-

-sO -cI

Stem types: V = short-vowel stems, VV = long-vowel stems, C = consonant stems. Segmental alternations: A = a e, O = o u, I = ɪ i.

The imperative forms are used to express command and related modal contents, e.g. mii=dee jarɪ-gɪta ‘let me also sing!’ (1SG=PTCL sing-IMP.1SG), uleen-ji jobo-o-so ‘work well!’ (good-INSTR work-IMP-2PL), tei ekte ene-gi-ni ‘let her go!’ (that woman go-IMP-​3SG). The third person imperative forms are, however, often replaced by second person imperatives from causatives, e.g. pikte-yi xupi-gi-ni ~ xupi-ween-duu ‘let my child play!’ (child-PX1SG play-IMP-3SG ~ play-CAUS-IMP.2SG). The same construction can also be used with a first person reference, e.g. aya-jɪ baljɪ-waan-doo ‘let me/us live well!’ (good-​I NSTR live-CAUS-IMP.2SG). Prohibition is expressed by the prohibitive particle eji (= NEG.IMP.2P e-ji). There are also other ways to convey the notion of command. In the second person, the suffix SG -xAArI : PL -xAAr-sO is used to mark the so-called “future imperative”, e.g. xola- ‘to read’ : IMP.FUT 2SG xola-xaarɪ : 2PL xola-xaar-so ‘read it (later)!’ In the first person, the purposive converb in -(pO)-gO(-) with possessive or reflexive marking can be used in the function of a cohortative, e.g. gese osɪ-go-raa baljɪ-go-arɪ ‘let us live together!’ (together become-REV-CV.ANT live-CV.PURP-RX.PL). OTHER MODAL FORMS Apart from the imperatives, there are several other grammaticalized forms with a modal content. Among these, the subjunctive and optative, with Common Tungusic parallels,

392 Shinjiro Kazama

are “true” moods, in that they have only a single personal paradigm, but not other forms. Other forms with a modal content, like the intentives and necessitatives involve deverbal verbs with a full verbal paradigm. •

The subjunctive mood is marked by the suffix -mcA-, which takes a mixed set of person markers, e.g. omɪ- ‘to drink’ : SUBJ 1SG omɪ-mca-ɪ : 2SG omɪ-mca-sɪ : 3SG omɪ-mca : 1PL omɪ-mca-po : 2PL omɪ-mca-so : 3PL omɪ-mca-l, often followed by the particle =ma. The subjunctive has a counterfactual meaning, as in [8], but it can also express a euphemistic proposal, as in [9]:

[8] naɪ beye-ji-eri baljɪ-xa-so osɪnɪ, if person body-instr-rx.pl live-ptcp.prf-px2pl ‘If you had been born in the form of a normal human being, xoonɪ=a amɪ-sɪ agda-mca=ma how=ptcl father-px2sg be.pleased-subj=ptcl your father would have been very pleased.’ [9] cɪxala-ɪ-sɪ agree-ptcp.imprf-2sg ‘If you agree,

osɪnɪ if

tei mii pikte-ji-yi-ye mamala-mca-sɪ=ma that 1sg child-instr-px1sg-ext marry-subj-2sg=ptcl how about marrying my daughter?’ •

•

•

•

The optative mood is marked by the suffix -ƞA-, followed by a mixed set of person markers of the subjunctive type and the particle =tAnII, e.g. jobo- ‘to work’ : OPT 1SG jobo-ƞa-yɪ=tanɪɪ : 2SG jobo-ƞa-sɪ=tanɪɪ : 3SG jobo-ƞa=tanɪɪ : 1PL jobo-ƞapo=tanɪɪ : 2PL jobo-ƞa-so=tanɪɪ : 3PL 3SG jobo-ƞa-l=tanɪɪ. The optative expresses a wish or consent and has a future tense reference, e.g. mii ei daƞsa-wa xola-ƞa-yɪ=tanɪɪ ‘I (am) will(ing to) read this book’ (1SG this book-ACC read-OPT-1SG=PTCL). In the modern language the optative has, however, virtually disappeared, leaving only the second person singular form in occasional use, e.g. jiju-pi sɪa-ƞa-sɪ=tanɪɪ ‘when you return, please eat!’ (return-CV.CSEC eat-OPT-2SG=PTCL). Intentives are formed by the suffix -(k)IcA-, e.g. ɪlɪ- ‘to stand up’ : INTENT ɪlɪɪca- ‘to intend to stand up’. This suffix requires an animate subject, e.g. tei gaakɪ degde-ice-i-ni ‘that crow is about to fly away’ (that crow fly-INTENT-PRS-3SG). Some verbs in this form yield a lexicalized meaning, e.g. ice- ‘to see’ : INTENT ice-ice- ‘to seek’. Necessitatives are formed by the suffix -gIlA, e.g. ene- ‘to go’ : NEC ene-gile- ‘to have to go’. This form is often used in rhetorical questions, e.g. nɪƞmaan pikte-ni xaɪ goɪda-mɪ ure-gile-i ‘why should a child in a folktale take a long time to grow up?’ (folktale child-PX3SG what last.long-CV.CONN grow-​NEC-3SG). Desideratives are formed by the suffix -mO-, which in the participial forms takes the marker -sI(-), e.g. jep- ‘to eat’ : DESID je-mu-(si) ‘to be hungry’, omɪ- ‘to drink’ : omɪ-mo-(sɪ) ‘to be thirsty’. In predicative position, these verbs are used impersonally, and, like adjectives expressing emotions or feelings, they are combined with the dative form of the nominal, e.g. min-du cieci-mu ‘I want to urinate’ (1SG-DAT

Nanai 393

urinate-DESID), cf. e.g. min-du peku ‘I feel hot’ (1SG-DAT hot). They can also be combined with a nominative subject, in which case the latter is interpreted as the causer of the feeling of a first-person experiencer, e.g. tei naɪ iñe-mu ‘he makes me laugh’ (that person laugh-DESID). CONVERBS Compared with several other Tungusic languages, including Ulcha and Uilta, Nanai has a somewhat limited system of converbal forms. Disregarding quasiconverbal forms, which perform functions identical with the fully grammaticalized converbs, but which synchronically involve transparent case forms of the participles, there are only six converbal forms in Nanai, here termed the connective, consecutive, conditional, purposive, terminative, and anterior converbs (Table 14.8). The connective and terminative converbs have a Common Tungusic background, while the consecutive converb is specific to Southern Tungusic. The anterior converb has aorist-based parallels in most Tungusic languages, but the formal details vary. The conditional converb is an Amur Tungusic feature, while the purposive converb is apparently a specifically Nanai innovation. By their morphosyntactic behaviour the converbs can be divided into the conjunct (same-subject) and disjunct (different-subject) types. Some converbs are, however, ambivalent in this respect. The conjunct converbs are coreferential with the subject of the main clause and do not take person marking, though the connective and conditional converbs, which diachronically include reflexive marking, do have plural forms. The disjunct converbs take the possessive suffixes for person marking. TABLE 14.8 NANAI CONVERB MARKERS  

 

PL

CONN

-mI

-mArI

PX

RX

SS +

CSEC

-pI

-pArI

+

CTXT

-OcIA-

+

PURP

-(pO)gO-

+

TERM

-dAlA

+

ANT

-(dA-)rAA

+

DS (+) +

+

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e, O = o u, I = ɪ i, R = r d.

•

The connective converb in -mI : PL -mArI (~ -mAArI) can express simultaneous action, e.g. soƞgo-mɪ ɪlɪ-sɪ-ɪ-nɪ ‘s/he stands crying’ (cry-CV.CONN stand-STATPRS-3SG), beici-mi ene-lu-xe-ni ‘s/he started to go hunting’ (hunt-CV.CONN go-​ INCH-PST-3SG)’, or a condition, e.g. bum-bie mana-mɪ sum-bie waa-nda-jaraa ‘if (s/he) destroys us, s/he will come and kill you’ (1PL-ACC destroy-CV.CONN 2PL-ACC kill-AND-​FUT). It is also used with auxiliary verbs as the form of the semantic main verb, as well as with the adjective aya ‘good’ in predicative position in a common phrase expressing permission, e.g. ii-mi aya ‘may I come in?’ (enter-CV.CONN good). This construction may imply an obligatory action, e.g.

394 Shinjiro Kazama

•

•

•

•

•

sargancoo bu-dii-we-ni xorɪ-mɪ aya=goanɪ ‘the beauty is almost dying, we must save her life’ (beauty die-PTCP.PRS-ACC-PX3SG save-CV.CONN good=PTCL). The connective converb is negated by a construction in which negation is marked both by a preverbal particle and a postverbal suffix, e.g. em saa-ra-mɪ ji-cim-bi ‘I came without knowing (it)’ (NEG know-AOR-CV.CONN come-PST-1SG). Here, the main verb contains superficially only the marker of the connective converb after the aorist marker, but in reality, at least diachronically, the converb marker is a trace of the suffixalized form *e-mi ‘not doing’, which is also the source of the particle em (= e-m), meaning that the diachronic derivation is saa-ra-mɪ < *saa-ra+e-mi. The consecutive converb in -pI : PL -pArI (~ -pAArI) expresses temporal or conditional subordination, e.g. tei ekte ca-wa ice-pi enim-bi jooƞgo-ɪ-nɪ ‘whenever she sees it, she remembers her mother’ (that woman that-ACC see-CV.CSEC mother-RX remember-PRS-​3SG). It can also imply a sequence of actions, e.g. tei jobom-ba xojɪ-pɪ gese caɪ-ya omɪ-nda-​go-o ‘when you/we finish that work, come and drink tea together!’ (that work-DX-ACC finish-CV.CSEC together tea-ACC drinkAND-REV-IMP.2SG), cf. also the idiom bekteen bi-pi=e ‘after a while’ (while be-CV.CONN=EXT). Exceptionally, if there is a close relationship between the subject of the subordinate clause and that of the main clause, the consecutive converb can also be used in a disjunct construction. The contextual converb in -OcIA-, in the “Lower Amur” dialects -OcA-, as also in Ulcha, functions as the disjunct counterpart of the consecutive converb and expresses non-​coreferential temporal or conditional subordination, e.g. cɪmana tugde-ucie-ni mii ene-e​sim-bi ‘if it rains tomorrow, I will not go’ (tomorrow rain-CV.CNTXPX3SG 1SG go-​NEG.PRS-1SG). Almost the same meaning can also be expressed analytically by the conjunction osɪnɪ ‘if, when’, e.g. cɪmana tugde-xen osɪnɪ mii ene-esim-bi ‘if it rains tomorrow, I will not go’ (tomorrow rain-PTCP.PRF if 1SG go-NEG.​PRS-1SG). The conjunctional construction with osɪnɪ can also be used coreferentially, e.g. tei ekte ca-wa ice-i-ni osɪnɪ enim-bi jooƞgo-ɪ-nɪ ‘whenever she sees it, she remembers her mother’ (that woman that-ACC see-PTPC.​IMPRFPX3SG if mother-RX remember-PRS-​3SG). This construction has a strong modal force directed to the hearer, and it is particularly often used to express a counterfactual condition in conversational contexts, in which the counterfactual conditional is expressed only by the counterfactual conjunction osɪnɪ. The purposive converb in -gO- (for vowels stems) ~ (-pO)gO- (for other stem types) can be used in both corefential and non-coreferential constructions. It expresses purpose (‘in order to’). In coreferential usage it takes the reflexive suffixes, e.g. xao-sɪ ene-gu-yi ta-ɪ-sɪ ‘where do you intend to go?’ (where-DIR go-CV.PURP-RX do-PRS-2SG). In non-coreferential usage it takes the possessive suffixes preceded by the “oblique coaffix” (OCX), e.g. neu-yi ji-pugu-e-ni sɪa-go-arɪ cek bargɪ-xa-nɪ ‘s/he prepared their meals in order to welcome his/her younger sibling’s arrival’ (younger.sibling-RX come-CV.PURP-OCX-PX3SG eat-CV.PURP-​RX.​PL ready prepare-PST-3SG). The terminative converb in -dAlA expresses a temporal limitation, e.g. yada-dala soƞgo-xa-nɪ ‘s/he cried until s/he got tired’ (get.tired-CV.TERM cry-PST-3SG), dolbo ine-dele ‘until the day dawns’ (night dawn-CV.TERM). A lexicalized item is ele-dele ‘(eat) to the full’ (eat.enough-CV.TERM). The anterior converb is based on the aorist and has the forms -rAA ~ -dAA depending on the stem type of the verb. The verbs ga- ‘to take’, ji- ‘to come’, o- ‘to become’,

Nanai 395

•

and bu- ‘to die’ take the complex suffix -dA-rAA: ga-da-raa, ji-de-ree, o-da-raa, bu-de-ree, while the verb bi- ‘to be’ yields bi-ree. The anterior converb expresses anterior action, e.g. japa-raa sɪa-xa-nɪ ‘s/he took (it) and ate (it)’ (take-CV.ANT eat-PST-PX3SG). A relatively strongly grammaticalized quasiconverbal form that may be mentioned in this context is built by the combination of the instrumental case form of the imperfective participle with the possessive suffixes and the adverb gese ‘together’. This combination expresses an action that is immediately followed by another action (‘as soon as’), e.g. japa-ɪ-jɪ-yɪ gese naƞgala-xa-nɪ ‘no sooner had s/he taken (it) than s/ he threw (it) away’ (take-​PTCP.IMPRF-INSTR-RX together throw-PST-3SG). The same meaning can also be expressed by the imperfective participle in combination with the complex form noljɪn-jɪ-a-nɪ(=tol) ‘at the moment when’ (moment-INSTR-​ OCX-​PX3SG=PTCL), e.g. nie-rii noljɪn-jɪ-a-nɪ=tol ca-wa baƞsala-xa-nɪ ‘s/he kicked it the moment s/he went out’ (go.out-​PTCP.IMPRF moment-INSTR-OCXPX3SG=PTCL that-ACC kick-PST-3SG).

SYNTAX Nanai is a verb-final language (SOV) in which the modifier precedes the modified. There is no agreement in number or case. Exceptionally, a numeral or another quantifier can follow a noun in object position and both can take the accusative marker, e.g. sɪa-orɪ jaka-wa egji-we aƞgo-xa-nɪ ‘s/he cooked a lot of food’ (eat-PTCP.IMPERS.IMPRF thing-ACC much-ACC make-​PST-3SG). This should, however, not be viewed as agreement, but, rather, as a signal of the ambiguous status of quantifiers, as is also evident from the seemingly inverted order of the constituents. Agreement between a constituent with the subject is indicated by the reflexive suffixes, cf. e.g. tei naɪ sogdata-ƞgo-yɪ sɪa-xa-nɪ ‘he ate his (own) fish’ (that person fish-AL-RX eat-PST-3SG) vs. tei naɪ sogdata-ƞgo-a-nɪ sɪa-xa-nɪ ‘he ate his (= another person’s) fish’ (that person fish-AL-OCX-PX3SG eat-PST-3SG). Coordination is expressed by juxtaposition, e.g. eike neu baljɪ-xa-cɪ ‘there lived a girl and her younger brother’ (elder.sister younger.sibling live-PST-3PL). Clausal subordination takes place with the help of the converbs and quasiconverbal constructions, in which the temporal context is expressed by the finite verb of the main clause, e.g. a-psɪƞgo-ɪ-do-yɪ xem oosɪ-go-xa-nɪ ‘when s/he went to bed, s/he cleaned up everything’ (sleepMOM-REV-PTCP.IMPRF-DAT-RX all clean-REV-PST-3SG), mii jiju-i julie-le-ni puƞne-gu-usu ‘before I come back, drive them out!’ (1SG come.back-PTCP.IMPRF before-LOC-PX3SG drive.out-REV-IMP-2PL). The subject of both finite and non-finite clauses is in the unmarked nominative; only in quotational clauses is the accusative case possible, e.g. mii sim-bie tui ta-jaraa=m ecie murci-e-yi ‘I did not think that you would do like that’ (1SG 2SG-ACC so do-FUT=PTCL NEG.PST think-CONNEG-1SG). There are no relative clauses, but correlative pairs of a dependent clause and a head clause can be linked by the repetitive use of interrogative pronouns, e.g. ui jobo-asɪ osɪnɪ ui sɪa-rasɪ ‘he who does not work shall not eat’ (who work-NEG.PRS if who eat-NEG.PRS). The following is a discussion of a selection of a few other syntactic features and patterns: •

Topic marking is not obligatory, but the topic can be marked by the enclitic particles =tAnII and =gola, e.g. ca-wa=tanii tei ogda doo-cɪ-a-nɪ nee-xe-ni ‘as for that one, s/he put it in the boat’ (that-ACC=TOP that boat inside-DIR-OCX-PX3SG put-PST-3SG). Another element used as a topic marker is the conjunction osɪnɪ ‘if,

396 Shinjiro Kazama

•

•

•

when’, e.g. tei=tenii nukte-ni osɪnɪ sadaa sadaa bi-i ‘as for her, her hair is stiff and dry’ (that=TOP hair-PX3SG TOP DESCR DESCR be-PRS). The conjunction osɪnɪ = PX3SG o-sɪ-nɪ is based on the verb o- : o-sɪ- ‘to become’, and its use as a topic marker, which has parallels in both Tungusic and Mongolic languages, derives directly from its conditional function. Equative sentences with an affirmative present tense reference do not normally require a copula, but a first or second person subject is indicated by person marking of the possessive type on the predicatively used nominal, e.g. mii uyum-bi ‘I am alive’ (1SG alive-1SG) : sii uyun-si ‘you are alive’ (2SG alive-2SG), bue ɪlan ekte-pu ‘we are three women’ (1PL three woman-1PL). For first and second person subjects, the auxiliary ta- ‘to do’ can also be used as a copula, but only for affirmative present tense reference, e.g. mii sii amɪ-sɪ ta-am-bɪ ‘I am your father’ (1SG 2SG father-PX2SG do-AOR-1SG), sii=kee erdeƞ-ku ekte ta-a-cɪ ‘you are a woman with special powers (= you are a shaman)’ (2SG=PTCL method-PROPR woman do-AOR2SG). If negation or past tense have to be expressed, the copula bi- ‘to be’ is used, e.g. mii neu-yi=tenii [. . .] orkɪn naɪ bi-esi ‘my younger sibling is not a bad person’ (1SG younger.sibling-PX1SG=TOP bad person be-NEG.PRS), teje=dee bi-ci-ni ‘it was true’ (true=PTCL be-PST-3SG). The basic copulas bi- ‘to be’ and osɪ- ‘to become’ can also be combined with verbal forms to express complex temporal and modal contents, e.g. elee ɪsɪ-ɪ-nɪ bi-ci-ni ‘s/he was about to arrive’ (soon reach-PTCP.IMPRF-PX3SG be-PST-3SG), aosɪ-nɪ jiju-xe-ni bi-ci-ni ‘his/her maternal uncle had (already) come back’ (maternal.unclePX3SG come.back-​PST-​3SG be-PST-3SG), piuseenji-ji-e-ni=tul tuu-rii-ci bi-jeree ‘they had probably fallen because of the storm’ (storm-INSTR-OCX-PX3SG=PTCL fall-PTCP.PRS-3PL be-FUT), ñoan-cɪ-a-nɪ ene-i-ni osɪ-xa-nɪ ‘(it was decided that) she will have to go to them’ (3P.OBV-DIR-​OCX-PX3SG go-PTCP.IMPRF-PX3SG become-PST-3SG). An approximative function (‘almost, about’) is expressed by the suffix -mOlAA, e.g. elee tuu-mulee bi-ci-ni ‘s/he almost fell’ ~ ‘s/he was about to fall’ (soon fall-APPR be-PST-3SG). The auxiliary ta- ‘to do’ also forms a periphrastic construction with participles, e.g. mue-we muel-si-i moo-wa ii-wu-i ta-xa-nɪ ‘s/he did things like fetching water and gathering firewood’ (water-ACC fetch.water-​ HAB-PTCP.IMPRF.PRS wood-ACC enter-CAUS-PTCP.IMPRF do-PST-3SG). Other auxiliaries form aspectual and modal constructions with the connective converb of a semantic main verb. Examples: cɪɪla- ‘to do something in vain’, e.g. simbie xalacɪ-mɪ cɪɪla-xam-bɪ ‘I waited for you in vain’ (2SG-ACC wait-CV.CONN fail-PST-1SG); deruu- ‘to start’, e.g. sɪa-mɪ deruu-xe-ni ‘s/he started to eat’ (eat-CV. CONN start-PST-3SG); korpɪ- ‘to have time to do’, e.g. mii ɪraso-mɪ korpɪ-asɪm-bɪ ‘I do not have time to carry (her)’ (1SG carry-CV.CONN have.time-NEG.PRS-1SG); mute- ‘to be able to’, e.g. mii tui egji-we sɪa-mɪ mute-esim-bi ‘I cannot eat so much’ (1SG so much-ACC eat-CV.CONN be.able-NEG.PRS-​1SG); otolɪ- ‘to know how to’, e.g. paorɪ-mɪ otolɪ-asɪ ‘s/he cannot swim’ (swim-CV.CONN know.how-NEG.PRS); tepciu- ‘to start’, e.g. aƞgo-mɪ tepciu-xe-ni ‘s/he started to make (it)’ (make-CV. CONN start-PST-3SG). The adjective aya ‘good’ > ‘suitable’ and the lexicalized participle aca-asɪ ‘unsuitable’ > ‘not good’ (from aca- ‘to meet’) function in the same way, e.g. aag-bɪ gele-nde-gu-mi aya ‘it is best that I go to search for my elder brother’ (elder.brother-PX1SG search-AND-REV-CV.CONN good), ene-mi=dee aca-asɪ moco-go-mɪ​=daa aca-asɪ ‘it is not good to go (further), nor to return’ (goCV.CONN=PTCL suit-NEG.PRS return-REV-​CV.CONN suit-NEG.PRS).

Nanai 397

•

•

•

•

Negation is morphologically and morphosyntactically split into several structural patterns. The negation of the participles and finite indicative present-tense predicates takes place by the suffixalized negation verb -(A)-si- (< AOR *+e-si-), while the imperatives are negated by the independent preverbal petrified forms e-m and e-ji, of which the former is also used to negate the connective converb, which additionally takes the suffixalized negation marker -m (< CV.CONN *+e-mi). Finite past tense predicates are negated by the preverbal past tense negator e-cie. All these forms require the use of the connegative form of the main verb, which in combination with ecie can also take person marking of the possessive type, e.g. mii ecie saa-ra-yɪ ‘I did not know’ (1SG NEG.PST know-​CONNEG-​1SG), tei naɪ ecie saa-ra-nɪ ‘he did not know’ (that person NEG.PST know-​CONNEG-​3SG). Another petrified form of the negation verb is e-bur (< PASS-PTCP.IMPRF *e-bu-ri), which is synchronically used as a discursive negator in the meaning ‘of course not’, e.g. ebur ene-jeem-bi ‘of course I will not go’ (NEG.DISC go-FUT-1SG). Privation is expressed by the negative existential nouns aba/a and anaa ‘absence, absent’. Both can be used as sentential arguments with marking for case and person, e.g. ñoa-nɪ aba-do-a-nɪ egji naɪ ji-ci-ni ‘during his absence many people came’ (3P.OBV-PX3SG NEG.EXIST-​DAT-OCX-PX3SG much person come-PST​-3SG), sii anaa-do-a-sɪ neku xoonɪ baljɪ-ɪ-nɪ ‘if you were not here, how would our child live?’ (2SG PRIV-DAT-OCX-PX2SG younger.sibling how live-PRS-3SG). More typically, aba/a is used as a finite predicate, e.g. min-du xaɪ=daa abaa ‘I have nothing’ (1SG what=PTCL NEG.EXIST). It is also used as a response in the pair (affirmative) ii/n ‘yes’ : (negative) abaa ‘no’. With person marking and in combination with the connective converb of a main verb it offers an alternative construction for past tense negation, e.g. mii xola-mɪ aba-yɪ ‘I did not read’ (1SG read-CV.CONN NEG.EXISTPX1SG). This construction has been considered as the past tense of the “assertive mood” (Avrorin). Non-finite adnominal and adverbal privation is normally expressed by anaa, e.g. asɪ anaa naɪ ‘an unmarried man’ (wife PRIV person), jɪlgan=daa anaa osɪ-xa-nɪ ‘s/he was struck silent’ (voice=PTCL PRIV become-PST-3SG). Interrogation in polar questions, as well as in negative echo questions, is expressed by the enclitic particle =nOO, e.g. jook-cɪ-yɪ ɪsɪ-go-orɪ=noo abaa=noo ‘shall I reach my home or not?’ (house-DIR-RX reach-REV-IMPERS.PRS=INTERR NEG.EXIST=INTERR). When attached to interrogative pronouns, this particle expresses indefiniteness, e.g. dere oya-la-ni xaɪ=noo bi-i-ni ‘there is something on the table’ (table above-LOC-PX3SG what=INTERR be-PRS-3SG), ui=nuu uike-we dukte-i-ni ‘somebody is knocking on the door’ (who=INTERR door-ACC knock-PRS-3SG). Alternative questions can also be marked by the particles =os and =o, e.g. ji-ci-ni=os abaa-nɪ=os ‘I wonder if s/he has come or not’ (come-PST3SG=ALT NEG.EXIST=ALT), bur-mi=o baljɪ-mɪ=o ui daaljɪ-ko-so ‘whether I die or live is no business of yours’ (die-CV.CONN=ALT live-CV.CONN=ALT who relation-PROPR-PX2PL). Emphasis is expressed by several enclitic particles, most importantly by =dAA, which is often used in a cumulative or coordinative function (‘also, and’), e.g. seepe sɪƞakta-nɪ=daa bi-i gormaxon=​daa bi-i xaɪ=daa xem bi-i ‘there are sable furs, and there are rabbits, and there is everything’ (sable fur-PX3SG=PTCL be-PRS rabbit=PTCL be-PRS what=PTCL all be-PRS). When attached to interrogative pronouns and combined with negation, =dAA has a connegative function, e.g. xaɪ=daa abaa ‘there is nothing’ (what=PTCL NEG.EXIST). This particle is also used on the extraposed

398 Shinjiro Kazama

•

element when the verb-final word order is exceptionally broken, e.g. esi=tenii enexen, pikte-yi kuu-ree=dee ‘now she left, after having nursed her child’ (now=PTCL go-PST child-RX nurse-CV.ANT=PTCL). Other enclitic particles have a variety of discursive functions. They include: =bAkI ‘finally’, e.g. ca-do=bakɪ bu-dii-ni ‘finally he dies then’ (that-DAT=PTCL diePRS-3SG); =goanɪ [assertive], e.g. sokta-jɪ pulsi-uri osɪnɪ jaa-nɪ=goanɪ ‘it is easy to go skiing’ (ski-​INSTR go.around-IMPERS.PRS TOP easy-PX3SG=PTCL); =kAA [concessive] naɪ=kaa naɪ xuigu-ni ɪnda ‘he may be a man but he has a dog’s tail’ (person=PTCL person tail-PX3SG dog), also used with negation to express a request of consent, e.g. abaa un-deci-si=kee ‘you said there is nothing, didn’t you?’ (NEG.EXIST say-NEG.PST-2SG=PTCL); =kAAcI [similative], e.g. xaalɪ=daa gese bi-cin=keeci ‘(she behaved) as if she had lived together (with him) for a long time’ (when=PTCL together be-PST=PTCL); =kOcI [interrogative focus], e.g. ei daƞsa-wa=kocɪ ga-cɪ-sɪ ‘is this the book that you bought?’ (this book-ACC=PTCL take-PST-​2SG); =m : PL =mAl [quotative], e.g. ca-do nee-mi aya bi-jeree=m un-dii-ni ‘you can put it there, s/he said’ (that-DAT put-CV.CONN good be-FUT=QUOT say-PRS-3SG); =mAt ~ -mAA ‘finally’, e.g. ca-do=mat ji-ci-ni ‘then s/ he finally came’ (that-DAT=PTCL come-​PST-3SG); =maña : PX3SG mañaa-nɪ : PL =mañaa-l ‘nothing but’, e.g. pokto-nɪ seekse=​maña bi-ci-ni ‘his footprints were nothing but (= full of) blood’ (footprint-PX3SG blood=​PTCL be-PST-3SG); =rAgdA : PX3SG =rAgdA-nI : PL =rAgdA-l ‘only’, e.g. mapa-caam-​ba=ragda gaajo-ɪ-nɪ ‘s/he brought only the old man (but not the old woman)’ (grandfather-​ DIM-ACC=PTCL bring-PRS-3SG); =tOl [focus], e.g. cɪmanaa=tol jiju-jeem-bi ‘it is tomorrow that I will be back’ (tomorrow=PTCL come.back-FUT-1SG); =xei [focus], sii=xei ga-cɪ-sɪ ‘is it you who bought (it)?’ (2SG=PTCL take-PST-2SG).

LEXICON AND LANGUAGE CONTACTS Apart from the Common Tungusic inherited lexicon, Nanai shares regional vocabulary with the other Tungusic languages at several different levels: with Ulcha and Uilta within the Nanaic group, with Oroch and Udihe in the Amur Tungusic context, and with Jurchen-​Manchu at the larger Southern Tungusic level. For some items of basic vocabulary Nanai shows idiosyncratic developments. For instance, the original word for ‘to eat’ (*)jep-, as attested in the other Nanaic languages, as well as in Ewenic and Manchu, has been replaced in mainstream Nanai by sɪa-, originally ‘to bite’. Also, some items show exceptional phonetic developments, occasionally shared with Ulcha. For instance, Nanai and Ulcha have an initial nasal in nasa-l ‘eyes’ against *ïasa-l in the other Tungusic languages. Loan words have entered Nanai from all the surrounding languages. (Jurchen-) Manchu has been the principal source of secondary cultural vocabulary, including some numerals, e.g. taƞgo ‘hundred’ ← Manchu tanggu. Dialectally, basic vocabulary can also show Manchu features: for instance, the “Upper Amur” dialects have borrowed muke ‘water’ ← Manchu muke vs. Naikhin mue (< Proto-Tungusic *möö). Mongolian vocabulary has probably mainly been received via Manchu, e.g. daosʊn ‘salt’ ← Manchu dabsun ← Mongolic *dabu-su/n, but some items may have entered directly, e.g. damaxɪ ‘tobacco’ ← Mongolic tamaxi vs. Manchu dambagu. Some Mongolic items show Daghur features, e.g sarpɪ ‘chopsticks’, cf. Daghur (*)sarpa (< *sabka) vs. Upper Amur sapkɪ ~ saf kɪ.

Nanai 399

Modern cultural vocabulary is borrowed from Russian and Chinese, depending on the political context. Chinese loanwords have, however, also entered the dialects today spoken on the Russian side, e.g. duduuse ‘potatoes’ ← Chinese tudouzi 土豆子, ganjɪn ‘clean’ ← Chinese ganjing 干净, sɪanto-la- ‘to punch’ ← Chinese quantou 拳頭 ‘fist’, though in many cases they may have been transmitted by Manchu. There are also Chinese lookalikes in the kinship lexicon, suggesting social contacts in the past, e.g. papa ‘mother’s elder brother’ ~ Chinese baba 爸爸 ‘father’, dada ‘mother’s elder sister’ ~ Chinese taitai 太太 ‘aunt, lady’, gugu ‘father’s younger sister’ ~ Chinese gugu 姑姑 ‘aunt’. The natural and cultural context of the Nanai speakers is intimately reflected by the Nanai lexicon. There is, for instance, a rich vocabulary related to fishing, including regional names for varieties of fish, fishing instruments, boats and boat parts, etc. Orientation is based on the riverine environment, especially the Amur, and relative positions are determined by spatials with regard to the river. Taboo affects the use of vocabulary items in connection with totemism, bear cult, and other factors. Some idiomatic expressions may also be due to taboo, e.g. beye-du osɪ- ‘to get pregnant’ (body-DAT become). There is a rich vocabulary of descriptive words of various types, e.g. kirke ‘very old (man)’, walɪɪ walɪɪ ‘very tired’, caƞsop ‘(to rise) on one’s feet’, duƞ duƞ ‘(bind) tightly’, kukcelbelbel ‘flickering’, xaoncok ene-xe-ni ‘s/he fainted’ (unconscious go-PST-3SG). It has to be added that the Nanai speakers also have a rich oral literature, divided into the genres of nɪƞmaan ‘folktales’, teeluƞgu ‘legends’, and sɪoxor ‘foreign stories’. One of the poetic devices used in folklore is alliteration, as in [10]: [10]

beyum-be beici-i too-wa toɪkaan-dɪɪ, beast-acc hunt-prs elk-acc hit-prs ‘He hunts for wild beasts, he hits elks, gasa-wa garalɪ-ɪ ta-a waterfowl-acc beat-prs do-aor he beats waterfowl, they say.’

un-de say-aor

REFERENCES AND FURTHER READING An, Jun 安俊 (1986) 赫哲語簡志 [A sketch of the Hezhe language],『中國少數民族語 言簡志叢書』[Languages of the minority nationalities of China], 北京 [Peking]: 民 族出版社. Avrorin, V. A. [В. А. Аврорин] (1948) ‘О категориях времени и вида в нанайском языке’ [On the categories of tense and aspect in the Nanai language], Язык и мышление 11: 29–55, Москва [Moscow] & Ленинград [Leningrad]. Avrorin, V. A. [В. А. Аврорин] (1953) ‘Об ошибках в освещении некоторых вопросов грамматического строя нанайского языка и его истории’ [On mistakes made in the description of some issues concerning the grammatical structure and history of the Nanai language], Доклады и сообщения Института языкознания АН СССР 5: 107–130, Москва [Moscow] & Ленинград [Leningrad]. Avrorin, V. A. [В. А. Аврорин] (1956) ‘Предикативно-притяжательные формы в нанайском и других тунгусо-маньчжурских языках’ [Predicative-possessive forms in Nanai and other Tungusic languages], Вопросы языкознания (3): 93–100, Mосква [Moscow].

400 Shinjiro Kazama

Avrorin, V. A. [В. А. Аврорин] (1957) Основные правила произношения и правописания нанайского языка [Principal rules of the pronunciation and orthography of the Nanai language], Ленинград [Leningrad]: Учпедгиз. Avrorin, V. A. [В. А. Аврорин] (1958a) ‘Основные закономерности сочетания согласных в нанайском языке [The principal regularities of consonant combinations in the Nanai language], in: Филология и история монгольских народов, 177– 189, Москва [Moscow]: Издательство восточной литературы. Avrorin, V. A. [В. А. Аврорин] (1958b) ‘Сингармонизм гласных в нанайском языке’ [Vowel harmony in the Nanai language], Доклады и сообщения Институтa языкознания АН СССР 11: 137–150, Mосква [Moscow]. Avrorin, V. A. [В. А. Аврорин] (1959–1961) Грамматика нанайского языка [Grammar of the Nanai language], vols. 1–2, Москва [Moscow] & Ленинград [Leningrad]: Издательство Академии Наук СССР. Avrorin, V. A. [В. А. Аврорин] (1968) ‘Нанайский язык’ [The Nanai language], in: Монгольские, тунгусо-маньчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 129–148, Ленинград [Leningrad]: “Наука”. Avrorin, V. A. [В. А. Аврорин] (1981) Синтаксические исследования по нанайскому языку [Syntactic studies on the Nanai language], Ленинград [Leningrad]: “Наука”. Avrorin, V. A. [В. А. Аврорин] (1986) Материалы по нанайскому языку и фольклору [Materials on the Nanai language and folklore], Ленинград [Leningrad]: “Наука”. Diószegi, V[ilmos] (1953) ‘Le problème de l’ethnogenèse des samaghirs’, Acta Orientalia Academiae Scientiarum Hungaricae 3 (1/2): 33–44, Budapest. Doerfer, Gerhard (1973) ‘Das Kur-Urmiische und seine verwandten’, Zentralasiatische Studien 7: 567–599, Bonn. Doerfer, Gerhard (1975) ‘Ist Kur-Urmiisch ein nanaischer Dialekt?’, Ural-Altaische Jahrbücher 47: 51–63, Wiesbaden. Doerfer, Gerhard (1984) ‘Die Körperteilbezeichnungen des Kili (Ein Beitrag zur Frage der dialektologischen Klassifikation)’, Nyelvtudományi Közlemények 86: 238–246, Budapest. Grube, Wilhelm (1900) Goldisch-Deutsches Wörterverzeichniss mit vergleichender Berücksichtigung der übrigen tungusischen Dialekte, Reisen und Forschungen im Amur-Lande in den Jahren 1854–1856: Linguistische Ergebnisse 2, St. Petersburg: Kaiserliche Akademie der Wissenschaften. Janhunen, Juha (1985) ‘On the diphthongs in Nanay’,『北方文化研究』 Bulletin of the Institute for the Study of North Eurasian Cultures 17: 103–115, Sapporo. Kazama, Shinjirō 風間伸次郎 (1991–1993)「ナーナイ語テキスト」[Nanai texts] =『ツ ングース言語文化論集』 [Publications on Tungus Languages and Cultures] [1/1991] #1: 61–134, 札幌 [Sapporo]: 北海道大学; [2/1993] #4, 小樽 [Otaru]: 小樽商科大学言 語センター. Kazama, Shinjirō 風間伸次郎 (1994)「ナーナイ語の「一致」について」[On the so-called agreement in Nanay],『北大言語学研究報告』5, 札幌 [Sapporo]: 北海道大 学文学部言語学研究士室. Kazama, Shinjirō 風間伸次郎 (1995)「ナーナイ語の親族名称について」[Nanay kinship terminology],『北海道立北方民族博物館研究紀要』Bulletin of the Hokkaido Museum of Northern Peoples 4: 43–64, 網走 [Abashiri]. Kazama, Shinjirō 風間伸次郎 (1995–2012)『ナーナイの民話と伝説』 [Nanay folk tales and legends] [1/1995] 小樽 [Otaru]: 小樽商科大学言語センター; [2/1996] 鳥取 [Tottori]: 鳥取大学教育学部; [3/1997, 5/2000, 12/2010, 13/2012] 東京 [Tokyo]: 東京外

Nanai 401

国語大学; [4/1998, 8/2005, 9/2006] 千葉 [Chiba]: 千葉大学文学部; [6/2001, 7/2002] 吹田 [Suita]: 大阪学院大学情報学部; [10/2007, 11/2008] 札幌 [Sapporo]: 北海道大 学大学院文学研究科 =『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 5, 8, 10, 12, 14, 15, 18, 27, 32, 36, 40, 48, 53. Kazama, Shinjirō 風間伸次郎 (1996/1998)「ヘジェン語の系統的位置について」,『言語 研究』Journal of the Linguistic Society of Japan 109: 117–139, 京都 [Kyoto] = ‘Genetical position of Hezhen’,『環北太平洋の言語』Languages of the North Pacific Rim 4: 115–131, 札幌 [Sapporo]: 北海道大学. Kazama, Shinjirō 風間伸次郎 (2005/2007)「ナーナイ語の疑問詞による反語表現につ いて」, in:『環北太平洋の言語』Languages of the North Pacific Rim 12: 129–163, 札幌 [Sapporo]: 北海道大学 = ‘Rhetorical questions with interrogative markers in Nanai’, in: Kawaguchi Yuji & al. (eds.), Linguistic Infomatics and Corpus Linguistics, 305–321, Tokyo: Tokyo University of Foreign Studies. Kazama, Shinjirō 風間伸次郎 (2006a)「ナーナイ語コンドン方言テキスト」[Texts of the Kondon dialect of Nanai],『環北太平洋の言語』Languages of the North Pacific Rim 13: 83–126, 札幌 [Sapporo]: 北海道大学. Kazama, Shinjirō 風間伸次郎 (2006b)「ナーナイ語の形動詞について」[On the participles in Nanai],『言語情報学研究報告』11: 95–108, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2008)「ナーナイ語における語用論的研究 : ―短いテ キストにおけるケーススタディ―」[Pragmatic study of Nanai language: A case study of a short text],『言語の研究 : ユーラシア諸言語からの視座』16: 159–168, 東京 [Tokyo]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 (2010a)「ナーナイ語の非人称形動詞について」 [Impersonal participles in Nanai],『アジアとアフリカの言語と言語学』4: 133–148, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2010b)「ナーナイ語のアスペクト」[Aspect in Nanai],『 語学研究所論集』15: 181–191, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2011a)「ナーナイ語の複文について : 条件表現の使い 分けを中心として」[The complex sentence in Nanai: with a special focus on conditionals],『北方言語研究』1: 115–138, 札幌 [Sapporo]. Kazama, Shinjirō 風間伸次郎 (2011b)「ナーナイ語のモダリティ」[Modality in Nanai],『 語学研究所論集』16: 57–74, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2012–2021)『ナーナイ語諸方言の研究』[Research on Nanai dialects] =『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 54, 57, 68, 東京 [Tokyo]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 (2013)「《データ： 「補遺」》ナーナイ語」[Voice in Nanai]『語学研究所論集』18: 426–435, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2014)「《データ： 「他動性」》ナーナイ語」[Transitivity in Nanai],『語学研究所論集』19: 323–330, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2015a)『ナーナイの文化と生活1』 [Nanai culture and living 1],『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 60, 東京 [Tokyo]: 東京外国語大学. Kazama, Shinjirō 風間伸次郎 (2015b)「《データ：「（連用修飾的）複文」》ナ ーナイ語」 [Complex sentences in Nanai] 語学研究所論集』20: 205–214, 東京 [Tokyo]. Kazama, Shinjirō 風間伸次郎 (2016)「《データ： 「情報構造と名詞述語文」》ナーナイ 語」[Information structure and nominal predicates in Nanai],『語学研究所論集』21: 259–268, 東京 [Tokyo].

402 Shinjiro Kazama

Kazama, Shinjirō 風間伸次郎 (2019)「《データ：「情報表示の諸要素」》ナー ナイ語」 [Information structure in Nanai],『語学研究所論集』22: 25–45, 東京 [Tokyo]. Kile, A. S. [А. С. Киле] (1994) Нанайско-русский тематический словарь [NanaiRussian thematic dictionary], Хабаровск [Khabarovsk]: Хабаровский государс твенный педагогический университет. Kile, N. B. [Н. В. Киле] (1973) Образные слова нанайского языка [Ideophones of the Nanai language], Ленинград [Leningrad]: “Наука”. Kile, N. B. [Н. В. Киле] (1976) ‘Лексика связанная с религиозными представлениями нанaйцев’ [Lexicon connected with the religious concepts of the Nanai], in: Природа и человек в религиозных представлениях народов Сибири и Севера [Nature and man in the religious concepts of the peoples of Siberia and the North], 189–202, Ленинград [Leningrad]: “Наука”. Kile, N. B. [Н. В. Киле] (1977a) ‘Нанайские топонимы района озера Болонь’ [Nanai toponyms around Lake Bolon’], in: Филология народов Дальнего Востока (ономастика) [Onomastics of the peoples of the Far East], 59–63, Владивосток [Vladivostok]: ДВНЦ АН СССР. Kile, N. B. [Н. В. Киле] (1977b) ‘Антропонимы нанайцев’ [Nanai anthroponymy], in: Филология народов Дальнего Востока (ономастика) [Onomastics of the peoples of the Far East], Владивосток [Vladivostok]: ДВНЦ АН СССР. Kile, N. B. [Н. В. Киле] (1981) ‘Фольклорная лексика в нанайском языке [Nanai folkloric lexicon], in: Этнография и фольклор народов Дальнего Востока [Ethnography and folklore of the peoples of the Soviet Far East], 132–136, Владивосток [Vladivostok]: ДВНЦ АН СССР. Kile, N. B. [Н. В. Киле] (1986) ‘Словообразовательные типы образных слов нанайского языка и их грамматические значения [Derivational types and grammatical meanings of Nanai ideophones], in: Актуальные вопросы языков народностей Севера [Current issues of the languages of the peoples of the North], 111–116, Якутск [Yakutsk]: Якутский филиал СО АН СССР. Kile, N. B. [Н. В. Киле] & al. (1996) Нанайский фольклор: Нингман, сиохор, тэлунгу [Nanai folklore: Ningman, siokhor, telungu], Памятники фольклора народов Сибири и Дальнего Востока 11, Новосибирск [Novosibirsk]: “Наука”. Kim Juwon & Ko Dongho & Antonina Kile & Choi Moonjeong (2015) The Life and Rituals of the Nanai People, Altaic Languages Series 08, Seoul: Seoul National University Press. Ko, Dongho & Yurn Gyudong (2011) A Description of Najkhin Nanai, Altaic Languages Series 07, Seoul: Seoul National University Press. Lattimore, Owen (1932) The Gold Tribe, “Fish-skin Tatars”, of the Lower Sungari, Memoirs of the American Anthropological Association 40, Menasha, WI. Li, Linjing 李林静 (2014)「ホジェン語の動詞屈折形式とその統語機能」[Hezhen verbal inflectional forms and their syntactic functions],『北方言語研究』4: 111–126, 札 幌 [Sapporo]. Li, Linjing 李林静 (2020)「ヘジェン語における二つのアスペクト形式 -mi bi- と -re bi- について」[Two aspect forms -mi bi- and -re bi- in Hezhen],『北方言語研究』10: 117–134, 札幌 [Sapporo]. Lopatin I. A. [И. А. Лопатин] (1922) Гольды амурские, уссурийские и сунгарийские [The Gold of the Amur, Ussuri and Sungari], Владивосток [Vladivostok]: Типографiя Управл. Внутр. Дѣлъ.

Nanai 403

Menges, K. H. (1964). ‘Zur Verbal-morphologie des Nanaj’, Ural-Altaische Jahrbücher 35: 397–402, Göttingen. Menges, K. H. (1968). ‘Die Sprache der Nanai (Goldi)’, in: Altaistik: Tungusologie, Handbuch der Orientalistik I.V.3: 171–246, Leiden: E. J. Brill. Onenko, S. N. [С. Н. Оненко] (1955) ‘Фонетические нормы словарных заимствований из русского языка’ [Phonetic norms of lexical borrowings from Russian], В помощь учителю школ Крайнего Севера 6: 100–110, Ленинград [Leningrad]. Onenko, S. N. [С. Н. Оненко] (1958a) ‘Характеристика некоторых областей лексики современного нанайского языка’ [Characteristics of selected areas of modern Nanai lexicon], Просвещение на Советском Крайнем Севере 8: 257–280, Ленинград [Leningrad]. Onenko, S. N. [С. Н. Оненко] (1958b) ‘О частицах нанайского языка’ [On particles in modern Nanai], Доклады и сообщения Институтa языкознания АН СССР 11: 95–106, Mосква [Moscow]. Onenko, S. N. [С. Н. Оненко] (1958c) Русско-нанайский словарь [Russian-Nanai dictionary], Ленинград [Leningrad]: Учепедгиз. Onenko, S. N. [С. Н. Оненко] (1972) ‘Характеристика двуязычия у нанайцев’ [Bilingualism among the Nanai], Проблемы двуязычия и многоязычия [Problems of biand multilingualism], 208–213, Москва [Moscow]: “Наука”. Onenko, S. N. [С. Н. Оненко] (1975) ‘Заднеяычно-велярные согласные нанайского языка’ [Back velar consonants in the Nanai language], Известия СО АН СССР, Серия общественных наук 2: 135–139, Новосибирск [Novosibirsk]. Onenko, S. N. [С. Н. Оненко] (1977) ‘Значение суффикса -го/-гу в современном нанайском языке’ [On the meaning of the suffix -go/-gu in modern Nanai], in: Исследования по языкам народов Сибири, 62–73, Новосибирск [Novosibirsk]: Институт истории, филологии и философии СО АН СССР. Onenko, S. N. [С. Н. Оненко] (1980) Нанайско-русский словарь [Nanai-Russian dictionary], Москва [Moscow]: “Русский язык”. Onenko, S. N. [С. Н. Оненко] (1981a) ‘Coздание и развитие нанайской письменности’ [Crestion and development of the literary use of Nanai], in: Письменность народов Сибири: История и перспективы [Writing and literacy of the peoples of Siberia: History and perspectives], 96–101, Новосибирск [Novosibirsk]: Институт истории, филологии и философии СО АН СССР. Onenko, S. N. [С. Н. Оненко] (1981b) ‘Реликтовые формы продольного и отложительного падежей в нанайском языке’ [Relict forms of the prolative and elative cases in Nanai], in: Языки и фольклор народов Севера [Languages and folklore of the peoples of the North], 33–37, [Novosibirsk]: Институт истории, филологии и философии СО АН СССР. Onenko, S. N. [С. Н. Оненко] (1982) Cловарь нанайскo-pусский, pусско-нанайский [Nanai-Russian-Nanai dictionary], Ленинград [Leningrad]: “Просвещение”. 2nd edition (1989). Onenko, S. N. [С. Н. Оненко] (1983) Усовершенствование нанайской орфографии [Improving the Nanai orthography], Preprint, Новосибирск [Novosibirsk]. Onenko, S. N. [С. Н. Оненко] (1986) Русско-нанайский словарь [Russian-Nanai dictionary], Москва [Moscow]: “Русский язык”. Orlov, A. S. [А. С. Орловъ] (1869) Замѣтка о языкѣ гольдовъ и ходзеновъ’ [A note on the language of the Gold and Hezhen], Иркутскiя епархиальныя вѣдомости, Прибавленiя 13: 165–166, Иркутскъ [Irkutsk].

404 Shinjiro Kazama

Oskol’skaya, S. A. [С. А. Оскольская] (2014) ‘Морфологические свойства аспектуальных показателей в нанaйском языке’ [Morphological features of Nanai aspect suffixes], in: В. Ф. Выдрин [V. F. Vydrin] & Н. В. Кузнецова [N. V. Kuznecova] (eds.), От Бикина до Бамбалюмы, из варяг в греки: Экспедиционные этюды в честь Елены Всеволодовны Перeхвальской [Field-inspired essays in honour of Elena V. Perekhvalskaya], 43–54, Санкт-Петербург [St. Petersburg]: “Нестор-История”. Oskol’skaya, S. A. [С. А. Оскольская] (2015) ‘Показатель “косвенных” падежей в нанайском языке’ [A marker of “oblique” cases in Nanai], Acta Linguistica Petropolitana 11 (2): 379–397, Санкт-Петербург [St. Petersburg]. Oskol’skaya, S. A. [С. А. Оскольская] (2016) ‘К вопросу об аспектуальной системе нанайского языка’ [On the question concerning the aspectual system of the Nanai language], Вопросы языкознания (1): 76–93, Москва [Moscow]. Oskol’skaya, S. A. [С. А. Оскольская] (2020a) ‘Каритив в нанайском языке’ [Caritive in Nanai], Томский журнал лингвистических и антропологических исследований 27: 32–43, Томск [Tomsk]. Oskolskaya, Sofia (2020b) ‘Nanai and the Southern Tungusic languages’, in: Martine Robbeets & Alexander Savelyev (eds.), The Oxford Guide to the Transeurasian Languages, 305–320, Oxford Guides to the World’s Languages, Oxford: Oxford University Press. Oskol’skaya, S. A. [С. А. Оскольская] & Н. М. Стойнова [N. M. Stoinova] (2019) ‘Взаимодействие отрицания с акциональностью в нанайском языке’ [Interaction of negation with actionality in Nanai], in: Д. В. Герaсимов [D. V. Gerasimov] & С. Ю. Дмитренко [S. Yu. Dmitrenko] & Н. М. Заика [N. M. Zaïka], Сборник статей к 85-летию В. С. Храковского [Festschrift for V. S. Khrakovskii], 389–415, Москва [Moscow]: Издательский дом ЯСК. Petrova, T. I. [Т. И. Петрова] (1933) Грамматические таблицы нанайского языка [Grammatical tables of the Nanai language], Стеклограф. Ленинград [Leningrad]: Институт народов Севера. Petrova, T. I. [Т. И. Петрова] (1935) Краткий нанайско-русский словарь [Concise Nanai-Russian dictionary], Ленинград [Leningrad]: Учпедгиз. Petrova, T. I. [Т. И. Петрова] (1941) Очерк грамматики нанайского яыка [Sketch of Nanai grammar], Ленинград [Leningrad]: Учпедгиз. Petrova, T. I. [Т. И. Петрова] (1948) ‘Oбразные слова в нанайском языке’ [Ideophones in the Nanai language], Известия АН СССР 7 (6): 527–537, Mocква [Moscow]. Petrova, T. I. [Т. И. Петрова] (1954) ‘Oбразные слова служащие для передачи световых и цветовых впечатлений в нанайском языке’ [Ideophones conveying impressions of light and colour in Nanai], Ученые записки ЛГПИ 101: 111–130, Ленинград [Leningrad]. Petrova, T. I. [Т. И. Петрова] (1957) ‘Удвоенные и парные сочетания в нанайском языке’ [Doubled and paired combinations in Nanai], В помощь учителю школ Крайнего Севера 7: 134–150, Ленинград [Leningrad]. Petrova, T. I. [Т. И. Петрова] (1960) Нанайско-русский словарь [Nanai-Russian dictionary], Ленинград [Leningrad]: Учпедгиз. Protodiakonov, A[leksandr] [Aлександръ Протодiаконовъ] (1869–1870) ‘Краткiй русскогольдскiй словарь’ [Concise Russian-Gold dictionary], Иркутскiя епархиальныя вѣдомости, Прибавленiя 32, 40–41, 44–45 (1869), 17, 19, 22, 24–25, 29, 47, 50 (1870). Иркутскъ [Irkutsk].

Nanai 405

Protodiakonov, P[rokopii] [Прокопiй Протодьяконовъ] (1901) ‘Гольдско-русскiй словарь’ [Gold-Russian dictionary], Извѣстiя Восточнаго Института 2 (3): 300– 350, Владивостокъ [Vladivostok]. Putinceva, A. P. [А. П. Путинцева] (1954a) Морфология говора горинских нанай [Morphology of the dialect of the Gorin Nanai], Диссертация на соискание ученой степени кандидата наук [PhD thesis], ЛГПИ им. А. И. Герцена. Putinceva, A. P. [А. П. Путинцева] (1954b) ‘Числительные в говоре горинских нанай’ [Numerals in the dialect of the Gorin Nanai], Ученые записки ЛГПИ 101: 131–159, Ленинград [Leningrad]. Putinceva, A. P. [А. П. Путинцева] (1955) ‘К изучению имени существительного в нанайской школе’ [Studying the noun in the Nanai school], В помощь учителю школ Крайнего Севера 5: 53–68, Ленинград [Leningrad]. Putinceva, A. P. [А. П. Путинцева] (1956) ‘К изучению имени прилагательного в нанайской школе’ [Studying the adjective in the Nanai school], В помощь учителю школ Крайнего Севера 6: 122–143, Ленинград [Leningrad]. Putinceva, A. P. [А. П. Путинцева] (1955) ‘К изучению имени числительного в нанайской школе [Studying the numeral in the Nanai school], В помощь учителю школ Крайнего Севера 7: 70–86, Ленинград [Leningrad]. Radchenko, G. L. [Г. Л. Радченко] (1985) ‘Функциональная нагрузка ларингала в нанайских словоформах [The functional load of the laryngeal consonant in Nanai word forms], in: Лексика в тунгусо-маньчжурских языках Сибири [Lexicon of the Tungusic languages of Siberia], 97–105, Новосибирск [Novosibirsk]: Институт истории, филологии и философии СО АН СССР. Radchenko, G. L. [Г. Л. Радченко] (1986) ‘К вопросу о долгих гласных нанайского языка’ [On long vowels in Nanai], in: Фонетика языков Сибири и сопредельных регионов [Phonetics of the languages of Siberian and adjacent regions], 54–57, Новосибирск [Novosibirsk]: Институт истории, филологии и философии СО АН СССР. Sem, L. I. [Л. И. Сем] (1976) Очерки диалектов нанайского языка: Бикинский (уссурийский) диалект, Ленинград [Leningrad]: “Наука”. Sem, L. I. [Л. И. Сем] (1997) ‘Нанайский язык’ [The Nanai language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 173– 188, Москва [Moscow]: “Индрик”. Smolyak, A. V. [А. В. Смоляк] (1970) ‘Личные имена нанайцев’ [Personal names of the Nanai], in: Личные имена в прошлом, настоящем, будущем: Проблемы антропонимики [Personal names in the past, today, and in the future: Problems of anthroponymy], 173–177, Москва [Moscow]: “Наука”. Sunik, O. P. [О. П. Суник] (1948) ‘О языке нанайцев на р. Куре’ [On the language of the Nanai on the river Kur], Известия АН СССР, Отделение литературы и языка, 7 (6): 537–551, Москва [Moscow]. Sunik, O. P. [О. П. Суник] (1958a) Кур-урмийский диалект: Исследования и материалы по нанайскому языку [The Kur-Urmi dialect: Studies and materials on the Nanai language], Ленинград [Leningrad]: Учпедгиз. Sunik, O. P. [О. П. Суник] (1958b) ‘О языке зарубежных нанайцев’ [On the language of the transborder Nanai], Доклады и сообщения Институтa языкознания АН СССР 2: 168–184, Москва [Moscow].

406 Shinjiro Kazama

Ulitkin, A. N. [А. Н. Улиткин] (1933a) ‘Гольдский (нанайский) язык в свете нового учения о языке’ [The Gold (Nanai) language in the light of the new doctrine of linguistics], Язык и мышление 1: 111–118, Москва [Moscow]. Ulitkin, A. N. [А. Н. Улиткин] (1933b) ‘Материалы по гольдскому (нанайскому) языку [Materials on the Gold (Nanai) language], Язык и мышление 1: 119–147, Москва [Moscow]. Zakharov, I. I. [И. И. Захаровъ] (1876) ‘О матерiалахъ для изученiя гольдскаго языка, доставленных отцомъ Ал. Протодiаконовымъ’ [On the materials for the study of the Gold language presented by Father Al. Protodiakonov], Извѣстiя Императорскаго Русскаго географическаго общества 6: 579–581, С.-Петербургъ [St. Petersburg]. Zaksor, L. Zh. [Л. Ж. Заксор] (2003) Нанайский язык в таблицах: Нанайско-русский тематический словарь [The Nanai language in tables: A Nanai-Russian thematic dictionary], Санкт-Петербург [St. Petersburg: “Дрофа”. Zhang, Yanchang [張彥昌] & Zhang Xi [張晰] & Dai Shuyan [戴淑艳] (1989), The Hezhen Language [赫哲语], 長春 [Changchun]: 吉林大學出版社 [Jilin University Press].

CHAPTER 15

ULCHA Shinjiro Kazama

Ulcha is a language of the Nanaic group, closely related to Uilta, and spoken in the Lower Amur basin immediately north of Nanai and south of the Palaeosiberian Ghilyak (Nivkh) language. The Ulcha ethnic territory, today administered as the Ulcha Raion (Ul’chskii raion) of Khabarovsk Krai, is located along a section of the Amur where the river flows relatively close to Nevelskoy Strait, the narrowest point of the Strait of Tartary, with which it is connected by a route of local rivers, lakes, and overland passages. Traditionally, the Ulcha speakers live in about a dozen villages, comprising, from south to north: Kalinovka, Sofiisk, Mariinskoe, Bulava (the center of the raion), Auri, Mongol, Dudi, Savinsk, Koima, Soloncy, Kol’chëm, Bogorodskoe, Ukhta, and Nizhnyaya Gavan’. In the north, as in the village of Tyr, Ulcha speakers live mixed with the Ghilyak and Neghidal. Most villages with an Ulcha population are today dominated by ethnic Russians. The ethnonym Ulcha (Russian ul’cha ~ MASC ul’ch : PL úl’chi : FEM ul’chánka), earlier also Olcha, is not an original endonym of the Ulcha, though it has reportedly been used as an exonym for the Ulcha by the Neghidal, from whom it apparently entered Russian usage, perhaps also because of a confusion with the Uilta. The actual endonym of the Ulcha speakers is naañɪ = naa-ñɪɪ ‘local people’, as also used by the Nanai and Oroch. In early Russian and European sources the Ulcha were known by the name “Mangun” (Russian PL mangúny), based on Ulcha maƞgʊ-ñɪɪ ‘Amur people’. Another name of historical importance is “Santan”, Japanese santan 山丹, used in connection with the Santan trade network, which linked the Japanese Islands and Sakhalin across the Tartar Strait with Manchuria and China. Ulcha speakers played a dominant role in this trade, which, in fact, may have been a factor behind the formation of the Ulcha as an ethnic group. Culturally, the Ulcha speakers are very similar to the Ghilyak, with whom they also share intimate social ties. This may mean that the Ulcha are essentially descendants to earlier Ghilyak speakers who only recently shifted over to a Tungusic language of the Nanaic type. A recent linguistic expansion is also suggested by the fact that Ulcha as a language is very homogeneous with no clear distinctions between dialects. In connection with the Santan trade, the Ulcha absorbed also some Sakhalin Ainu elements, as is still evident from the Ulcha clan name Kuisali ‘Ainu’. According to the most recent official census the population of the Ulcha was 2,913 (2010). In recent decades, the population has been rather stable, and until the end of the Soviet period the Ulcha seem to have retained their ethnic language relatively well. In the Post-Soviet period there has been a rapid decline, however, leaving only some dozens of fluent speakers today. The language has no official status, and it is not taught at school, leaving Russian as the principal written medium for all Ulcha speakers. There are, however, signs of increasing ethnic activism among the Ulcha, and interest in the ethnic language has also grown. This has resulted in some publications, including an Ulcha primer, in which the language is written in a preliminary Cyrillic-based orthography. DOI: 10.4324/9781315728391-15

408 Shinjiro Kazama

TAXONOMIC STATUS In view of its rather transparent similarity with Nanai, and perhaps also because of the self-identification of the Ulcha speakers as “Nanai” (naañɪ), Ulcha was initially classified as a mere “dialect” of Nanai. Indeed, as a member of the Nanaic group of Tungusic, Ulcha shares many diagnostic features with Nanai, such as the general preservation of initial *p and *x and medial *mb and *nd, as well as the representation of *ü as u. A closer look at the data shows, however, that Ulcha differs from Nanai in many respects. Some of these differences involve archaic features that have been lost in Nanai, while others are due to innovations that have taken place in Ulcha. Among the Nanaic languages, Ulcha takes the middle position between Nanai and Uilta both linguistically and geographically. From the genetic point of view, however, Ulcha is nearer to Uilta, for many of the innovations observed in Ulcha are also present in Uilta. This means that Nanai was separated from the common ancestor of Ulcha and Uilta before the latter were differentiated from each other. The common features of Ulcha and Uilta are present at all levels of linguistic structure and substance: phonetic, inflectional, derivational, and lexical. Phonetic innovations distinguishing Ulcha and Uilta from Nanai include the following: • *x- > Ø before u, e.g. Ulcha & Uilta uisi- ‘to sound’ vs. Nanai xuisi-; Ulcha unce ‘elbow’ = Uilta uite vs. Nanai xuice; Ulcha & Uilta weede- ‘to miss’ vs. Nanai xuede-. • *r > l in clusters before m b t, e.g. Ulcha galmakta ‘mosquito’ = Uilta nalmakta vs. Nanai garmakta; Ulcha & Uilta gelbu ‘name’ vs. Nanai gerbu; Ulcha bʊltaɪ ‘completely’ = Uilta bultai vs. Nanai bortaɪ. • *k *g > Ulcha c j = Uilta c j (before i) > t d (before other vowels) in clusters after l and ƞ > Uilta n, e.g. Ulcha tolcɪn ‘dream’ = Uilta tolci vs. Nanai tolkɪn; Ulcha bɪlja ‘throat’ > Uilta bilda vs. Nanai bɪlga; Ulcha moƞjɪ-cʊ- ‘to make skins soft by kneading’ = Uilta monji- vs. Nanai moƞgɪ-. • *rk *rg *rƞ > Ulcha c j ñ > Uilta t d n, e.g. Ulcha uce ‘door’ > Uilta ute vs. Nanai uike < *örke; Ulcha bajʊn ‘enemy’ > Uilta baduu vs. Nanai baɪgoan < *bargu-; Ulcha paña ‘palm’ > Uilta pana vs. Nanai paɪƞa < *parƞa. • *l > r before g k in secondary borrowings from Manchu transmitted by Nanai, e.g. Ulcha ɪrga ‘decoration’ = Uilta irga vs. Nanai ɪlga ← Manchu ilga ~ ilha (ilha); Ulcha & Uilta erke ‘silent, slow’ vs. Nanai elke ← Manchu *elge ~ elhe (elhe). • *ks > sk sporadically, e.g. Ulcha & Uilta weske ‘sleeve’ vs. Nanai xuekse; Ulcha & Uilta peske- ‘to be surprised, to wonder’ vs. Nanai pekse-; note also Oroch pekse-, Neghidal peske-, attested in Ghilyak as feske-. Features of inflectional and derivational morphology that link Ulcha with Uilta, as opposed to Nanai, include, the following: • Nominal morphology: the prolative in *-kI > Ulcha -kI = Uilta -kki, absent in Nanai; the contrastive-emphatic suffix of adjectives in *-dUmA > Ulcha -dUmA = Uilta -duma vs. Nanai -jImA; the frequent use of the adjectival suffix -UlI, often corresponding to -sI in Nanai, e.g. Ulcha amta-ʊlɪ ‘tasty’ = Uilta apta-uli vs. Nanai amta-sɪ, Ulcha got-ʊlɪ ‘bitter’ = Uilta goccu-uli ‘salty’ vs. Nanai gocɪ-sɪ ‘bitter’, Ulcha ñama-ʊlɪ ‘warm’ = Uilta nama-uli vs. Nanai ñama ~ ñama-sɪ, Ulcha xuje ~ xuje-uli ‘heavy’ = Uilta xede-uli vs. Nanai xuige, etc. • Verbal morphology: the finite future in *-rIlA/n- > Ulcha -rIlA/n- = Uilta -rilA/n- vs. Nanai -jA/A-; the future imperative in *-sAArI- > Ulcha -sArI ~ -sAlI = Uilta -ssAAri ~ -ttAAri vs. Nanai -xAArI; the subjunctive in -lAxA/n- vs. Nanai -mcA (also attested in Ulcha); the purposive converb in Ulcha -bdA- = Uilta -buddoo- vs. Nanai -gO-.

Ulcha 409



•

•

In the case of lexical differences it is not always possible to determine which side, Nanai or Ulcha-Uilta, is more innovative than the other. Often, both sides simply preserve a different inherited item. In many cases, the differences are confined to irregular derivational or phonetic details, while in others totally different lexemes are used. Derivational differences and other irregular developments are involved in, for instance, the following items: Ulcha wasa/n ‘heron’ = Uilta wasa vs. Nanai waacaka; Ulcha & Uilta nu-lji- ‘to move’ vs. Nanai nu-kte-; Ulcha & Uilta baasala- ‘to kick’ vs. Nanai baƞsala-; Ulcha & Uilta akpan- ‘to lie down’ < *ap-kan- vs. Nanai ap-sɪn-. Different lexemes are used in, for instance, the following cases: Ulcha aldʊ/n ‘news’ = Uilta aldu vs. Nanai mede (from Mongolic via Manchu); Ulcha makʊrɪ ‘dried fish’ = Uilta mauri vs. Nanai ogosan; Ulcha wagbaƞgɪ ‘gloves’ = Uilta wambakka ~ mambakka vs. Nanai kacama; Ulcha & Uilta duuse ‘tiger’ vs. Nanai puren ambanɪ ‘the big one of the forest’; Ulcha cʊpal ‘everything’ = Uilta cuppal ~ cipal vs. Nanai xem (also in Uilta geem); Ulcha jep- ‘to eat’ = Uilta deptu- vs. Nanai sɪa- (dialectally also jep-); Ulcha sɪɪ- ‘to bite’ = Uilta see- vs. Nanai sekpen-; Ulcha & Uilta somɪ- ‘to shut (the door)’ vs. Nanai dasɪ-; Ulcha & Uilta nimeri- ~ ñimeri- ‘to visit’ vs. Nanai jɪɪma-; Ulcha pan-sɪ- ‘to ask a question’ = Uilta panu- vs. Nanai medesi- (from mede ‘news’); Ulcha yaya- ‘to sing’ = Uilta yaaya- vs. Nanai jarɪ-.

DATA AND SOURCES Peter Schmidt was probably the first to publish data of Ulcha, collected by himself and his collaborators in the field in 1908. His publication (1923) consists of a lexicon of about 2,000 words and some texts. Slightly before him, Ulcha lexical items had been collected by Bronisław Piłsudski, but his materials, comprising about 1,500 words, were published only decades later by Alfred and Elżbieta Majewicz (1985). The first overall description of Ulcha was prepared by T. I. Petrova, who still treated Ulcha as a “dialect” of Nanai. Her publication (1936) contains a grammar with comparative notes, a dictionary of about 3,000 words, and 10 texts with translation. After this, no new material became available until O. P. Sunik started publishing his field materials collected in 1966–1976. His first grammatical sketch (1968) was followed by a more extensive monograph (1985), containing a grammatical description, a dictionary of about 7,000 words, and 19 texts with translation. He also compiled a practical dictionary with about 4,000 words (1987), as well as, later, yet another grammatical sketch (1997). V. A. Avrorin, who had worked on Ulcha already in 1948, published a collection of folklore (1981), comprising seven texts with translation. Today, the Russian tradition of Ulcha studies is continued by N. M. Stoinova (e.g. 2018). In Japan, the study of Ulcha was first taken up by Ikegami Jirō (1961, 1969, 1999), who made available unique materials based on the documents recorded by early Japanese explorers of the Edo period. More recently, Kazama Shinjirō has carried out field work among the last speakers of Ulcha and published several collections of folklore texts with translation, grammatical notes, and audiodiscs (1996–2010). He has also published articles on the diachronic position and areal context of Ulcha (1998, 2004, 2008, 2009, 2016). The first primer for Ulcha children was prepared by K. F. Dechuli and L. I. Sem (1992). Colloquial conversation texts have been published by S. V. Angina (1993), while

410 Shinjiro Kazama

Val’dyu (2005) is a bilingual dictionary for practical use. Another publication intended for the members of the Ulcha ethnic community is Kazama & Ogihara & Rosugbu (2007). The native Ulcha scholar A. O. Rosugbu has published on local toponymy (1977), which, however, remains a largely unstudied field up to the present day. SEGMENTAL STRUCTURE Early descriptions of the language postulated for Ulcha a system of six vowels of the same type as in Nanai. In this system there would be an opposition between two rounded back vowels u o, two unrounded front vowels i ɪ, and two unrounded central vowels a e [ə]. As in Nanai, the phonetic distinction between i and ɪ is small and can disappear altogether in non-initial syllables in favour of the value [i]. Even so, in the initial syllable the vowel ɪ is pronounced with a lowered quality and can approach the value of [e], especially when the following syllable has a, e.g. ɪxan [ɪxɑ̃] ~ [exɑ̃] ‘cow’. The distinctivity of ɪ is confirmed by items like ɪsɪ- [ɪɕi-] ‘to reach’, in which no contextual factor is present. However, later studies have also revealed the presence of ʊ, as in ʊgda [ʊgda] ‘boat’. The distinctive status of ʊ (ʊʊ) is suggested by, at least, the item jʊʊ [dʑʊ:] ‘house’, though the vowel is clearly marginal in the modern language and is probably merging with o and/or u. There are also traces of a distinction between *o and *ö > ɵ [o/ɵ], as is suggested by variants like [xulu] ~ [xolo] ‘squirrel’ for *xölö. Synchronically, the vowel ɵ is very rare and attested only in combination with labial consonants, but it seems to be distinctive in, at least, mɵɵ ‘water’ (< *möö) vs. moo ‘tree’ (< *moo), where it represents Proto-Tungusic *ö. It is also observed in the words bɵyɵ/n ‘wild animal’ (< *beyön) and ɵmɵ/n ‘one’ (< *emön). Obviously, ɵ, like ʊ, is today only marginally present in the system, but its few remaining occurrences suggest that it must have been a regular phoneme at some earlier stages of the language, as is also confirmed by its consistently distinctive status in Uilta. This means that we may with some reservations postulate a system of eight vowels for Ulcha (Table 15.1). TABLE 15.1 ULCHA VOWELS u

 

ʊ

i ɪ

ɵ

e

o

a

This system, which represents a verticalized version of the Proto-Tungusic vowel paradigm, is essentially identical with the systems attested in Ewen, Orochen, and Solon. It may be assumed that Nanai also once had a similar system, though it subsequently lost the distinctions *ʊ vs. *o and *u vs. *ɵ. An analogous tendency of reducing the number of distinctions from eight to six is observed in Neghidal. In view of the somewhat marginal status of the vowels ʊ and ɵ, it seems that Ulcha has also been on the way towards a sixvowel system, as attested in the neighbouring Ghilyak language. As in Nanai, the vowels in Ulcha can form both monophthongoid and diphthongoid sequences (VV). The diphthongoid sequences -aɪ -oɪ -ʊɪ and -ei -ui in non-initial syllables tend to be pronounced as monophtongoid ɪɪ ~ ii [i:], which can also be shortened to [i], e.g. japa- ‘to take’ : PTCP.IMPRF japa-ɪ [dʑapi:] ~ [dʑapi], ƞene- ‘to go’ : PTCP.IMPRF ƞene-i [ŋəni:] ~ [ŋəni], soƞgo- ‘to cry’ : PTCP. IMPRF soƞgo-ɪ [soŋgi:] ~ [soŋgi]. This

Ulcha 411

tendency is more general than in Nanai, where only -aɪ and -ei undergo monophthongization. Original short ɪ and i in final position are often devoiced and can be dropped. In the consonants, Ulcha retains the Proto-Tungusic system of 18 distinctive segments (Table 15.2). The system is identical with that of Nanai. TABLE 15.2 ULCHA CONSONANTS m

n

ñ

ƞ

b

d

j

g

t

c

k

p

s w

x y

l r

The palatal stops c j are pronounced as alveolo-palatal affricates [dʑ tɕ] before all vowels. The sibilant s has the alveolo-palatal value [ɕ] before ɪ i, for some speakers (in the southern part of the Ulcha territory) also before e u. However, unlike both Nanai and Uilta, Ulcha retains the sequences ti tɪ and di dɪ distinct from ci cɪ and ji jɪ, respectively, cf. e.g. dɪlɪ ‘head’ < *dïlï > Nanai jɪlɪ. Also, Ulcha retains a distinctive palatal ñ, as opposed to the dental n, before the vowels ɪ i, cf. e.g. ñɪɪ ‘person’ (< *ñarï) vs. nɪƞma/n ‘tale’ (< *nïƞma-), a unique feature among the Tungusic languages. Phonetic palatalization can also be caused by the vowel sequence ɪa (< *ïa), which, like its counterpart in Ewenic, is normally pronounced as a long monophthongoid [e:], e.g. bɪa [bje:] ‘moon’ (< *bïaga). At the level of allophony, the velar stops k g have the uvular variants [q] and [ɢ] before the original velar vowels a o ʊ. The nasal n in word-final position is realized as the nasalization of the preceding vowel, but the nasalization can also be absent, leaving no trace of n (here written as /n), e.g. sɪa/n ‘ear’ [ɕeẽ] ~ [ɕe:]. The lateral l in word-final position is always palatalized to [ʎ]. In several words, this palatalization has developed into an actual vowel segment, yielding a variation between l and the sequences lɪ li, e.g. juel ~ jueli ‘two’, xasal ~ xasalɪ ‘wing’, ɪsal ~ ɪsalɪ ‘eye’, seul ~ seuli ‘oar’. PHONOTACTICS AND MORPHOPHONOLOGY The phonotactic properties of Ulcha are very similar to those of Nanai. Thus, the general syllabic structure is (C)V(V)(C). The distinction of short (single) and long (double) vowels can be illustrated by minimal pairs like pikte ‘child’ vs piikte ‘nettle’, osɪ- ‘to become’ vs oosɪ- ‘to clear up’, but the functional load of the distinction is probably smaller than in Nanai, and long vowels can be lost in non-initial syllables. Phonotactically, the long monophthongs are equal to diphthongoid sequences, cf. e.g. ii- ‘to enter’ : PTCP.IMPRF ii-rii, aʊ- ‘to sleep’ : PTCP.IMPRF aʊ-rɪɪ vs. ice- ‘to see’ : PTCP.IMPRF ice-i. Monosyllabic free forms with no final consonant are always pronounced long, e.g. 1SG bii. The principal factor causing morphophonological alternations is vowel harmony, which opposes the “hard” (originally back) vowels ʊ o a ɪ to the “soft” (originally front) vowels u ɵ e i and requires the vowels in suffixes to harmonize with those in the root. Ulcha also preserves more traces of labial harmony, which requires the use of the rounded vowels o ɵ instead of a e if the preceding syllable has o ɵ. Labial harmony is, however, not fully productive in Ulcha, and there are many exceptions of it.

412 Shinjiro Kazama

For diachronic reasons there are several restrictions concerning the occurrence of consonants and their combinations in clusters. For instance, the only nasal that can occur in word-final position is n, which, therefore, can also be viewed as an unmarked archiphonemic nasal consonant. A stem-final /n is dropped before suffixes beginning with the consonants n m l s. Otherwise, n is regressively assimilated to the place of articulation of the following consonant, e.g. xaʊsa/n ‘paper’ : ACC xaʊsam-ba. Regressive assimilation also affects a stem-final g of nominal stems (here written /g), which is absent in absolute position, but which appears as g ~ k ~ ƞ before a suffix-initial consonant, e.g. dere/g = dere ‘face’ : PX 1SG dereg-bi : 2SG derek-si : 3SG dereƞ-ni. As in Nanai, the primary stress in Ulcha may be described as falling on the first syllable, which also controls the harmonic status of the word. At the same time there is, however, a high pitch on the penultimate “mora” of the word. WORD FORMATION The morphological patterns of Ulcha are basically similar to those in Nanai and Uilta. There is a basic distinction between nominals and verb(al)s, which divide the categories of derivation into the four standard classes of denominal and deverbal nouns and verbs. (1) Denominal nouns comprise, above all, the productive category of alienability, as well as a number of more marginal and often lexicalized derivational types: •

•

•

•

Alienability is marked by the suffix -ƞgU- which is placed immediately before the markers of case and possession, e.g. sʊgdata ‘fish’ : AL-PX3SG sʊgdata-ƞgʊ-nɪ ‘his/her fish’. Adjectives used as independent head nouns also normally take this suffix, ulee/n ‘good’ : AL-INSTR ulee-ƞgu-ji ‘well’. Alienability is typically expressed on nouns denoting foods, plants, wild animals, fish, birds, heavenly bodies, and natural phenomena. Spatials, kinship terms, and integral parts of entities are never marked for alienability. Diminutives are marked by the suffix -kAA/n, e.g. xagdʊ/n ‘house’ : DIM xagdʊ-kaa/n, loxo/n ‘sword’ : DIM-ACC loxo-koom-ba. When added to pronominal stems, this suffix has a limitative function, e.g. tɪɪ-kaan-jɪ bi-i ‘we lived by only that one’ (that-DIM-INSTR be-PRS). It is also present in some lexicalized items, e.g. eñi/n ‘mother’ : DIM eñe-keen ‘mother-in-​law’ (cf. Nanai eƞpe id.), edi ‘husband’ : DIM ede-keen ‘father-in-law’ (cf. Nanai aƞpa id.). There seem to be no augmentative nouns in Ulcha. The suffix -ncA/n refers to people living at a certain place, e.g. gasa/n ‘village’ : gasa-nca/n ‘villager’, aʊrɪ-nca/n ‘a person from the village Aori’. A very productive suffix denoting people is -ñII ~ -ñI (< -ñAI < *+ñarï ‘man, person), as in naa-ñɪ ‘Ulcha’. When attached to the imperfective participle, it forms actor nouns, e.g. bʊta- ‘to fish’ : PTCP.IMPRF bʊta-ɪ > bʊtɪɪ : bʊtɪɪ-ñɪɪ ‘fisherman’. It can also be added to numeral stems, e.g. buu jue-​ñii-​kee/n ‘we are only two’ (1PL two-person-DIM). As in Nanai, several kinship terms have special “vocative” forms: amɪ/n ‘father’ : VOC amaa, daamɪn- ‘grandfather’ : VOC daama, eñi/n ‘mother’ : VOC eñee, aag- ‘elder brother’ : VOC agaa, eige- ‘elder sister : VOC egee. For some kinship terms, the basic form is also used as a vocative, e.g. NOM = VOC daada ‘grandmother’. Terms for deceased relatives are marked by the suffix -ƞAsA, e.g. daada-ƞasa ‘late grandmother’. This suffix, known from the Ewenic

Ulcha 413

languages and Uilta, is not attested in Nanai. Ulcha has a kinship-term system based on a patrilineal society and levirate marriage. (2) Deverbal nouns express action, actor, or instrument. Unlike participles, these forms have no verbal functions and often involve lexicalized meanings. • •

•

Action nouns are formed by the suffix -/n, e.g. baldɪ- ‘to live, to be born’ : baldɪ-/n ‘life’, deƞsi- ‘to work’ : deƞsi-/n ‘work’. Actor nouns are formed by the suffix -mdI, e.g. belecu- ‘to help’ : belecu-mdi ‘helper’, saa- ‘to know’ : CAUS saa-wan- : saa-wa-mdɪ ‘messenger’. Occasionally, this suffix can also be attached to nominals, e.g. sagdɪ ‘old’ : sagdɪ-mdɪ ‘oldsters’. Another actor noun suffix, implying good skills, is -lA, e.g. bʊta- ‘to fish’ : bʊta-la ‘good fisherman’. Instrumental nouns with lexicalized meanings are formed by the suffix -kU, e.g. ana- ‘to push’ : ana-kʊ ‘key’, ice- ‘to see’ : ice-ku ‘binoculars’.

(3) Denominal verbs: Like Nanai, Ulcha retains a few Common Tungusic ambivalent nomina-verba. e.g. xedu/n ‘wind’ : xedun- ‘to blow (of wind)’. Otherwise, verbs are derived from nouns by suffixes. •

•

The main verbalizers are -lA- and -dA-, which often yield lexicalized meanings, e.g. jolo ‘stone’ : jolo-lo- ‘to throw’, asɪ ‘wife’ : asɪ-la- ‘to marry (of men)’, edi ‘husband’ : edi-le- ‘to marry (of women)’, naa-ñɪ ‘Ulcha’ : naa-ñɪ-da- ‘to speak Ulcha’. Other denominal verbs include the captatives in -mAsI-, e.g. gasa ‘duck’ : gasamasɪ- ‘to hunt for ducks’, the possessives in -cI-, e.g. aapʊ/n ‘hat’ : aapʊ-cɪ- ‘to have a hat’, and the translatives (of nouns denoting seasons) in -jU-, e.g. bolo ‘autumn’ : bolo-jʊ- ‘to become autumn’, jʊwa ‘summer’ : jʊwa-jʊ- ‘to become summer’.

(4) Deverbal verbs: Apart from the derivatives expressing voice, aspect, or mood, deverbal verbs comprise the following types, as also known from Nanai and Uilta: •

• •

Andatives are formed by the suffix -ƞdA- (= Uilta -ƞdA-, Nanai -nda-), e.g. waa- ‘to kill’ : AND waa-ƞda- ‘to go to kill’. This suffix can be expanded by the element -sU-, which adds a habitive meaning, e.g. beici-ƞde-su-i-ni ‘he always goes hunting’ (hunt-AND-​HAB-PRS-3SG). Pretentives are formed by the suffix -kAAcI-, e.g. gele- ‘to search’ : PRET gelekeeci- ‘to be looking for, to pretend to be searching’, paka- ‘to play with a ball’ : PRET paka-kaacɪ- ‘to play casually with a ball’. Transitives are formed from intransitives by the suffix -bU- ~ -wU-, which may also be understood as a causativizer. All actual examples are based on motion verbs, e.g. ii- ‘to enter’ : CAUS ii-wu- ‘to put/bring in’, nie- ‘to go out’ : CAUS nie-wu- ‘to put/take out’, agbʊn- ‘to appear’ : CAUS agbʊm-bʊ- ‘to take out’, tuku- ‘to fall’ : CAUS tug-bu- ‘to drop’, too- ‘to go away from the riverside’ : CAUS toog-bʊ- ‘to take away from the riverside’, eu- ‘to go down to the river’ : CAUS ee-wu- ‘to take down to the river’, xaa- ‘to dock’ : CAUS xaag-bʊ- ‘to bring something onshore from a boat’.

NUMBER AND CASE Most nominal stems end in a vowel or the nasal /n. More rarely, the lateral l is also attested as a stem-final consonant. In addition, there are a few stems ending in a “hidden”

414 Shinjiro Kazama

/g, which is absent in the basic form but manifested as g k ƞ before suffixes; this group comprises the following items: jʊʊ/g ‘house’, dere/g ‘face’, peruu/g ‘trousers’, and aag‘elder brother’ (always used with a possessive suffix). Plural is marked by the suffix -sAl ~ -sAlI, used mainly on nouns denoting animate beings, e.g. mʊrɪ/n ‘horse’ : PL mʊrɪ-sal/ɪ. Irregular plural forms are present in eekte ‘woman’ : PL eke-sel ~ ek-sel, guru/n ‘people’ : PL gur-sel, jɪa ‘fellow’ : PL jɪa-sɪl. A few items preserve the plural suffix -l, e.g. puru-l ‘children’, nasa-l ‘eye/s’ (< *ïasa-l). Associative plurality is marked by the suffix -nAA, e.g. am-naa ‘fathers, father and others’, mapa-naa ‘grandfathers, grandfather and others’. The Ulcha case paradigm, with seven suffixally marked cases, is essentially identical to that of Nanai, except that it contains a separate prolative case (Table 15.3). TABLE 15.3 ULCHA CASE MARKERS  

V

C

ACC

-wA

-bA

DES

-jU-

DAT

-dU

LOC

-lA

PROL

-kI

DIR

-tI

INSTR

-jI

-dU-lA

Stem types: V = vowel stems, C = consonant stems. Segmental alternations: A = a e, U = ʊ u, I = ɪ i.

• •

•

•

The unmarked nominative is used as the form of the subject, nominal predicate, and the “translative noun” before the verb o- ‘to become’, e.g. tɪɪ pikte ambaan daayɪ o-cɪ-nɪ ‘that child has become rather big’ (that child rather big become-PST-3SG). The accusative in -wA ~ -bA marks the direct object, e.g. tɪɪ gasam-ba ɪsʊ-xa-tɪ ‘they reached that village’ (that village-ACC reach-PST-3PL), dere oya-la-nɪ jepuwu-we tɪs nee-xen ‘s/he put plenty of food on the table’ (table top-LOC-PX3SG food-ACC very put-PST). As in Nanai, accusative marking can occasionally be absent, but this does not imply indefiniteness. Unlike Nanai, Ulcha does not use the accusative in an approximative function on numerals, nor in an emotional function on adjectives. The designative in -jU- (cf. Uilta -du, Nanai -gO-) is almost always used with “effective verbs”, such as baka- ~ baa- ‘to find, to get’, gele- ‘to want’, anjʊ- ‘to make’, japa- ‘to take’, etc., e.g. bii asɪ-jʊ-yɪ baa-xan, sii edi-ju-yi baa-xan ‘I found a wife for myself, and you found a husband for yourself’ (1SG wife-DES-RX find-PST, 2SG husband-DES-RX find-PST). It can also co-occur with the accusative, e.g. neuwe-si asɪ-jʊ-yɪ gele-ƞde-xem-​bi=e ‘I have come to ask you for your sister to be my wife’ (younger.sibling-ACC-PX2SG wife-DES-RX want-AND-PST-1SG=PTCL). The dative in -dU indicates primarily the location of an action or existence, or also the location where an object will be after an action is completed. In this usage, it is typically combined with verbs like bi- ‘to be’, nee- ‘to put’, loko- ‘to hang’, tee‘to sit’, etc., e.g. ɵm biruu-du bi-ci-ni ñɪɪ-sal ‘in a village there lived people’ (one village-DAT be-PST-3SG person-PL), pikte-yi moo-dʊ loko-​raa emsu-i bi-cin ‘she rocked her child hanging in a cradle from a tree’ (child-RX tree-DAT hang-CV.ANT rock.baby.in.cradle-PTCP.PRS be-PST). It can also have a temporal reference, e.g. mɵɵ dalan-dʊ-nɪ puli-kte-xe-ni ‘he walked around at the time of the floods’ (water

Ulcha 415

•

•

•

flood-DAT-3SG go.around-DISTR-PST-3SG). With verbs of “giving” the dative marks the recipient (“indirect object”), e.g. otɪɪtaka otɪɪtaka gasan ñɪɪ-dʊ-nɪ buu-ktexen ‘little by little s/he gave (them) to the people of the village’ (little little village person-​DAT-PX3SG give-DISTR-PST). The locative in -lA ~ -dU-lA- indicates a restricted or definite location, or also a locational limit, e.g. mɵɵ min peñe-dule bi-ci-ni ‘the water reached up to my knees’ (water 1SG.GEN knee-LOC be-PST-3SG). It is often used in reference to a part of an entity (e.g. a body part), e.g. ƞaala-la-nɪ japa-xa-nɪ ‘s/he took (him/her) by the hand’ (hand-LOC-3SG take-PST-3SG), nukte-le-ni gapakala-xa-nɪ ‘s/he grabbed (him/her) by the hair’ (hair-​LOC-​3SG grab-PST-3SG). The prolative in -kI (with a cognate in Uilta but not in Nanai) indicates movement through or along, e.g. uce-ki ii-mi=dee mute-esi-ni=guni ‘s/he wanted to enter through the door, but s/he could not’ (door-PROL enter-CV.CONN=PTCL be.ableNEG.PRS-3SG=​PTCL), mɵɵ-ki ƞene-i-ni ‘s/he goes along the water’ (water-PROL go-PST-3SG). The directive in -tI indicates the direction or target of motion. In this function it is typically used with motion verbs like di- ‘to come’, ii- ‘to enter’ : ii-wu- ‘to put/bring in’, too- ‘to go away from the riverside’ : too-bʊ- ‘to take away from the riverside’. It is also used to mark the addressee of verbs of verbal activity and emotional feelings, including wen- ‘to say’, jaralɪ- ‘to talk’, paƞsɪ- ‘to ask a question’, gele- ‘to ask’, aksa- ‘to be annoyed’, tagdansɪ- ‘to get angry’. The directive function can also be expressed by the enclitic spatial =bAA/n ‘in the direction of’, as in [1]:

[1] gee tɪmana daʊ-wʊ interj tomorrow cross-ptcp.impers.prs ‘Well, tomorrow we shall cross the river tɪɪ xotom=baa-nɪ tɪɪ that town=towards-px3sg that towards that town, to his town.’

•

ñɪɪ person

xoton-tɪ-nɪ town-dir-px3sg

The Ulcha enclitic directive marker is used more frequently than its Nanai cognate OCR-PX3SG baaro-a-nɪ. It has basically all the same functions as the original directive case, which it seems to be replacing, but its secondary origin is obvious from the fact that it can deviate from the rules of vowel harmony, cf. e.g. bii sin=baa-nɪ di-cim-bi ‘I came to you’ (1SG 2SG.GEN=towards come-PST-1SG). The instrumental in -jI indicates in its basic function instruments and materials, e.g. sʊlta-jɪ di-dii ‘s/he is coming on skis’ (ski-INSTR come-PRS). It has, however, also the function of a separative case, e.g. meen xagdʊn-jɪ ƞene-xe-ni ‘s/he came from her own house’ (REFL house-INSTR go-PST-​3SG), which is why it is used to indicate the standard of comparison, e.g. kaljamɪ tɪɪ xagdʊ-jɪ=daa gʊgda ‘(the mythical monster) Kaljami is even higher than that house’ (Kaljami that house-INSTR=PTCL high). For the same reason, the instrumental is required by a number of emotional verbs, like ƞeele- ‘to be afraid of’, e.g. min-ji eji ƞeele-re ‘do not be afraid of me’ (1SG-INSTR PROHIB be.afraid-CONNEG). Moreover, the instrumental functions as a comitative, as well, e.g. tɪɪ eekte-ji asɪ-la-xa-nɪ ‘he married her’ (that woman-INSTR wife-VBLZ-PST-​3SG). When attached to a participle, the instrumental forms a

416 Shinjiro Kazama

•

quasiconverbal construction indicating a temporal starting point, e.g. jekpi-ni xodaxan-ji-yi akpan-ci-ni ‘s/he finished eating and lay down’ (eat.PST-3SG finish-PTCP. PRF-INSTR-RX lie-PST-3SG). Ulcha also has the possessive marker -ƞgI, which marks possessive forms of nominals, e.g. ɪƞda ‘dog’ : POSS ɪƞda-ƞgɪ ‘dog’s (own)’. As in Nanai, these are not true case forms, but derivationally formed nominal bases that can themselves take case marking when used as independent head nouns.

Spatials (locational nouns) typically have a defective paradigm which comprises the regular locative and prolative cases, as well as the special spatial directive in -sI, e.g. juli‘front side’ : LOC-PX3SG juli-le-ni : DIR jule-si : PROL-PX3SG juli-ki-ni, xamɪ- ‘back side’ : LOC-​PX3SG xamɪ-la-nɪ : DIR xama-sɪ : PROL-PX3SG xamɪ-kɪ-nɪ, ui- ‘upper side’ : LOC-​PX3SG ui-le-ni : DIR ui-si : PROL-PX3SG uye-ki-ni, peji- ‘underside’ : LOC-PX3SG peji-le-ni : DIR pei-si : PROL-PX3SG peji-ki-ni, xaƞgɪ- ‘other side’ : LOC xaƞgɪ-la : DIR xaƞgɪ-sɪ. Spatials connected with orientation according to rivers have also the special accusative in -pA, e.g. solɪ- ‘upper course’ : LOC-PX3SG solɪ-la-nɪ : DIR solɪ : ACC solɪ-pa : PROL-PX3SG solɪ-kɪ-nɪ, xeji- ‘lower course’ : LOC-PX3SG xeji-le-ni : DIR xei : ACC xeji-pe : PROL-PX3SG xei-ki-ni, dui- ‘mountain side’ : LOC-PX3SG dui-le-ni : DIR dui-si : ACC dui.r-pe : PROL-PX3SG dui-ki-ni, waɪ- ‘river side’ : LOCPX3SG waɪ-la-nɪ : DIR waɪ-sɪ : ACC waɪ.rpa : PROL-PX3SG waɪ-kɪ-nɪ. Some spatials have a mixed paradigm with both regular and special case forms, e.g. bajɪ- ‘opposite side’ : LOC-PX3SG bajɪ-la-nɪ : ACC bajɪ-pa : DIR bajɪ-tɪ. Secondary spatials can be formed by the suffix -jIA ~ -jI ‘side’, e.g. toko/n ‘center’ : tokon-jɪa ‘middle part’, namʊ ‘sea’ : nam-jɪ ‘sea side’. The prolative forms uye-ki-ni ‘up’ and peji-ki-ni ‘down’ are also used in comparative constructions in the meanings ‘taller, superior’ vs. ‘smaller, inferior’, as in [2]: [2] kaljamɪ Kaljami

tɪs very

beye-ni body-px3sg

daayɪ, big

nekte moo uye-ki-ni low tree up-prol-px3sg (but) taller than low trees.’

bi-i be-prs

gʊgda high

moo tree

peji-ki-ni bi-i under-prol- be-prs px3sg ‘(The mythical monster) Kaljami is very big, he is smaller than high trees

ADJECTIVES Adjectives are distinguished from other nominals by a number of morphosyntactic properties, analogous to those observed in Nanai and Uilta: •

Proprietive adjectives are formed by the suffix -cU (cf. Nanai -kO), which typically implies permanent inalienable possession, e.g. bii tɪs kusun-cu ‘I am very strong’ (1SG very strength-PROPR), bii mene pikte-cu o-cɪm-bɪ ‘I got a child of my own’ (1SG REFL child-​PROPR become-PST-1SG), xetue-ni tɪs ɪrga-cʊ ‘the collar (of her dress) was all decorated’ (collar-PX3SG very decoration-PROPR). This suffix functions over the noun phrase, e.g. tɪɪ ñɪɪ daayɪ dere-cu ‘he has a big face’ (that person big face-PROPR). In such cases, the modifier in the noun phrase normally takes the

Ulcha 417

•

•

instrumental suffix, e.g. meƞgu-ji xosɪn-cʊ aɪsɪn-jɪ xete-cu bi-i merge ‘a hero with a silver skirt and golden coat’ (silver-INSTR skirt-PROPR gold-INSTR coat-PROPR be-PTCP.IMPRF hero), ɪlan-jɪ pati-cʊ ɪlan-jɪ dɪlɪ-cʊ puimul agbʊn-cɪn ‘a monster with three tails and three heads appeared’ (three-INSTR tail-PROPR three-INSTR head-PROPR monster appear-PST). Proprietive adjectives can be used both adnominally, e.g. sɪñakta-cʊ caʊlʊ/n ‘hairy chick’ (hair-PROPR chick), and adverbially, e.g. aapʊn-cʊ ii-ju-xen ‘s/he entered with her/his hat on’ (hat-PROPR enter-REV-PST). For the corresponding negative meaning, the privative noun anaa is used, e.g. tɪɪ ñɪɪ asɪ anaa bi-i-ni ‘he lives unmarried’ (that person wife PRIV be-PRS-3SG). Degree adjectives are formed by the suffixes -dUmA [contrastive-emphatic], e.g. ñawjʊa/n ‘young’ : ñawjʊan-dʊma ‘(the) younger (one)’, -AkU [augmentative], e.g. daayɪ ‘big’ : daayɪ-akʊ ‘very big’, -kAA/n ~ -kUU-kA/n [diminutive], e.g. ñuuci ‘small’ : ñuuci-​kuuke/n ‘very small’. Normally, emphasis is expressed by analytic means. A special emotional suffix taken by adjectives in combination with the interjection anayɪɪ ‘oh!’ is -lAnI, e.g. anayɪɪ, gʊgda-lanɪ ‘oh, how tall s/he is!’ (INTERJ high-EMPH). Another derivational suffix forming adjectives is -mA [material, from noun bases], e.g. sele ‘iron’ : sele-me ‘(of) iron’, sugbu-me tetu ‘garment made of fish skin’ (fish. skin-DX garment). Also, many adjectives in Ulcha incorporate the suffix -UlI, which can appear on both nominal and verbal bases, e.g. peku ‘heat, hot’ : peku-uli ‘hot’, aama-sɪ- ‘to be sleepy’ : aama-ʊlɪ ‘sleepy’.

NUMERALS The basic numerals are, for the digits: 1 ɵmɵ/n, 2 juel/i, 3 ɪla/n, 4 duyi/n, 5 tʊnja, 6 ñuƞgu/n, 7 nada/n, 8 jakpʊ/n, 9 xuyu/n, 10 jʊa/n, and for the lower digits: 20 xorɪ/n, 30 gʊtɪ/n, 40 dexi ~ deyi, 50 sʊsaɪ. With the exception of regular phonetic correspondences, these are identical with their cognates in Nanai. The numerals for the higher decades are formed by the suffix -njU (cf. Nanai -I-ƞgO): 60 ñuƞgu-nju, 70 nada-njʊ, 80 jakpʊ-njʊ, 90 xuyu-nju. For the lower powers of ten, 100 taƞgʊ and 1000 mɪƞga/n are used. Ordinals are formed by the suffix -I, yielding, in the first decade: ɵmɵ-i ‘first’, jue-i ‘second’, ɪlɪ-ɪ ‘third’, duyi-i ‘fourth’, tʊnjɪ-ɪ ‘fifth’, ñuƞgu-i ‘sixth’, nada-ɪ ‘seventh’, jakpʊ-ɪ ‘eighth’, xuyu-i ‘nineth’, jʊa-ɪ ‘tenth’. For collective numerals the suffix -tUƞAsA/r is used, e.g. jue-tuƞese/r ‘both of the two’, tʊnja-tʊƞasa/r ‘all five’; this suffix is normally combined with the markers for the instrumental case and reflexive possession, e.g. naa-tɪ jue-tuƞeser-ji-eri di-ci-ti ‘they both came’ (3P.OBV-PX3PL two-COLL-INSTR-RX.PL come-​PST-3PL). Days are counted by the Common Tungusic suffix -ltA. e.g. ɪla-lta ‘(for) three days’, also in the pronominal form xasʊ-lta ‘(for) how many days?’ ~ ‘(for) some days’. Multiplicatives are formed analytically by the noun mʊda/n ‘time/s’. The numeral 1 ɵmɵ/n has the short form ɵm, used as an equivalent of an indefinite article, e.g. ɵm ñɪɪ ‘a man’, ɵm eekte ‘a woman’. This numeral also has the emphatic-​contrastive derivative ɵmɵn-dume ‘one of them’. PRONOUNS The types and forms of pronouns in Ulcha are closely similar to those in Nanai. In the personal pronouns, Ulcha, like Nanai and Uilta, lacks a distinction between inclusive and exclusive pronouns in the first person plural. In general, however, the forms of the personal

418 Shinjiro Kazama

pronouns (Table 15.4) are more in line with Common Tungusic and lack the specific innovations that have taken place in Nanai. A distinctive feature of Ulcha, as compared with Nanai, is that the oblique stems 1SG min : 2SG sin : 1PL mun : 2PL sun are used in the function of adnominal genitives. It may be noted that the final nasal in these forms remains an invariable segment that does not assimilate to the place of articulation of the following consonant, suggesting that it is latently followed by the vowel i, i.e., min/i : sin/i : mun/i : sun/i, as is confirmed by the similar forms in Bikin Nanai and (the southern dialect of) Uilta. The possessive forms are 1SG mi-ƞgi : 1PL buu-ƞgi : S2G si-ƞgi : 2PL suu-ƞgi. TABLE 15.4 ULCHA PERSONAL PRONOUNS  

 

NOM

OBL

ACC

GEN

POSS

SG

1

bii

min-

mim-be

min

mi-ƞgi

2

sii

sin-

sim-be

sin

si-ƞgi

1

buu

mun-

mum-be

mun

buu-ƞgi

2

suu

sun-

sum-be

sun

suu-ƞgi

PL

For third-person reference, phrases like tɪɪ ñɪɪ ‘the man’ = ‘he’ and tɪɪ eekte ‘the woman’ = ‘she’ are used. The so-called third-person pronoun has the shape naa- : OBL naan-, which always requires the presence of a possessive suffix: SG NOM PX3SG naan/ɪ : ACC-PX3SG naam-ba-n/ɪ : DAT-PX3SG naan-dʊ-nɪ : LOC-PX3SG naan-dʊla-n/ɪ : DIR-​PX3SG naan-tɪ-n/ɪ : INSTR-PX3SG naan-jɪ-n/ɪ, PL NOM PX3PL naa-tɪ : ACCPX3PL naam-ba-tɪ : DAT-PX3PL naan-dʊ-tɪ : LOC-PX3PL naan-​dʊla-​tɪ : DIR-PX3PL naan-tɪ-tɪ : INSTR-PX3PL naan-jɪ-tɪ. As in Nanai and Uilta (as well as in Udihe), this pronoun is, however, used in the obviative function. In addition, Ulcha has the indefinite pronoun paa-n/ɪ ‘someone’. The reflexive pronoun is mee- : OBL meen- ‘(one)self’, which in inflection requires the presence of the reflective markers: SG mee-pi (unmarked object form) : DAT-RX meen​ -du-​yi : LOC-RX meen-dule-yi : DIR-RX meen-ti-yi : INSTR-RX meen-ji-yi : PL-ACC-​ RX.PL mee-pe-eri : DAT-RX.PL meen-du-eri : LOC-RX.PL meen-dule-eri : DIR-RX.PL meen-ti-eri : INSTR-RX.PL meen-ji-eri. In the emphatic (adverbial) function the form mene is used, e.g. bii mene saa-rɪɪ ‘I know (it) myself’ (1SG REFL know-PRS), mene anjʊ-xa-tɪ ‘they made (it) by themselves’ (REFL make-PST-3PL). In the adnominal possessive function the forms mene and meen can both be used, e.g. meen ~ mene xagdʊm-bɪ ‘one’s own house’. The reduplicated form mene&mene means ‘everybody’. The demonstrative pronouns in Ulcha are ei : OBL ei- ~ ye- ~ e-we- ‘this’ (proximal) vs. tɪɪ : OBL tɪ- ~ ta-ya- ~ ta-wa- ‘that’ (distal), which yield the following sets of inflected forms: ACC ei-we vs. tɪ-wa ~ tʊ-wa : DAT ei-du ~ ye-du vs. tɪ-dʊ : LOC ye-le vs. tɪ-la : DIR ewe-si ~ eu-si vs. tawa-sɪ : PROL ei-ki ~ ye-ki vs. tɪ-kɪ : INSTR ei-ji ~ ye-ji vs. tɪ-jɪ : SEP ewe-nci vs. taya-jɪ ~ tawa-ncɪ : PL ei-sel/i vs. tɪ-sal/ɪ. The demonstratives are often used in combinations, e.g. tawa-sɪ ewe-si ‘that way and this’. The form tɪɪ has also the adverbial meaning ‘so’, and in combination with the verb ta- ‘to do’ it forms lexicalized constructions like tɪɪ ta-raa ~ tɪtaraa ~ tʊtaraa ‘then’ (so do-CV.ANT). The basic interrogative pronouns are xaɪ : OBL xaɪ- ~ xa- ~ xaa- ~ xa-wa- ~ xoo‘what?’ and ƞui ‘who?’. The former yields a series of partially lexicalized case forms, including ACC xaɪ-wa ‘what?’ : DES-RX xaɪ-jʊ-yɪ ‘for what?’ : DAT xaɪ-dʊ ‘where?’ : LOC xaɪ-la ‘where?’ : DIR xaɪ-tɪ ~ xaɪ=baa-nɪ ~ xawa-sɪ ‘whither?’ : PROL xaɪ-kɪ ‘which

Ulcha 419

way?’ : INSTR xaɪ-jɪ ‘from where? how?’, as well as the special forms xaalɪ ‘when?’, xoo-nɪ ‘how?’, xamataa/n ‘what kind of?’, xasʊ/n ‘how many?’. The stem xaɪ- is also used as a pro-verb in the meaning ‘to do what?’, yielding forms like CV.CONN xaɪ-mɪ ‘why?’, AND-CV.CONN xaɪ-ƞda-mɪ ‘in order to do what?’. As in Nanai, the stem xaɪ : xaɪ- can be used as a dummy element, e.g. tɪɪ=meec xaɪ-xan, jololo-xon ‘(s/he) did, well, she threw (it) like that’ (so=PTCL do.what-PST throw-PST). The pronoun xaɪ and related forms are also often used in rhetorical questions, e.g. ei aldam-ba pul-puri xaɪ maƞga ‘why should it be so difficult to go there and back?’ (this interspace-ACC go.around-PTCP.IMPERS. IMPRF what difficult). In combination with the emphatic particle =dAA, xaɪ can express admiration or surprise, e.g. xaɪ=daa pʊkʊlɪ ‘how warm it is!’ (what=PTCL warm). PERSON MARKING The Ulcha system of possessive (PX) and reflexive (RX) markers and predicative personal endings (VX) differs only minimally from that of Nanai (Table 15.5). The biggest difference is that Ulcha has generalized a nasal initial to the first person plural predicative ending -mU, as opposed to the corresponding possessive suffix which has variation between -pU and -mU. TABLE 15.5 ULCHA PERSON MARKERS    SG

PL

PX 

VX 

V

C

AOR

SUBJ

1

-yI

-bI

-bI

-yI

2

-sI

-tI

-sI

3

-n/I

-Ø

RX

-yI

1

-pU ~ -mU

-mU

2

-sU

-sU

3

-tI

RX

-ArI

-bI

-Ø ~ -l

-pU -l

-bArI

Stem types: V = vowel stems, C = consonant stems. Segmental alternations: A = a e, O = ʊ u, I = ɪ i.

The predicative personal endings are used in the aorist and related finite forms, while the possessive suffixes are used on nominals to express possession, as well as on finitely-used nouns and nominalized verbs to express the subject. Some finitely used verbal categories, such as the subjunctive in -mcA-, use a mixed set of person markers. A few kinship terms have exceptional stems when combined with the possessive suffixes: amɪ/n ‘father’ : PX1SG am-bɪ, eñi/n ‘mother’ : PX1SG eƞ-bi. Some kinship terms are used only with possessive and reflexive suffixes, e.g. aag- ‘elder brother’ : PX 1SG aag-bɪ : 3SG aaƞ-nɪ : DIR-RX aak-cɪ-yɪ. OTHER WORD CLASSES When used adverbially, adjectival nominals normally take the instrumental suffix, e.g. uleen-ji bi-meri ocʊ-xa-tɪ=tanɪɪ ‘they began to live well’ (good-INSTR be-CV.CONN.PL become-​PST-​3PL=PTCL). Some adjectives are also used as degree adverbs, e.g. amban

420 Shinjiro Kazama

~ EMPH ambaan ‘big’ > ‘somewhat, rather’, e.g. ambaan malxʊ ɪƞda ‘rather many dogs’ (big many dog). Other items used as degree adverbs include tɪs ‘very’ (< ‘full, fully’), e.g. ñɪɪ tɪs malxʊ ‘(there were) very many people’ (person very many), otɪɪ ‘very’ (< ‘little’), e.g. otɪɪ ñuuci ‘very small’ (very small), and jiƞjiƞ (< ‘great, terrible’), used postpositionally, e.g. tɪɪ largɪ ñɪɪ jiƞjiƞ ‘he was such an extremely wonderful man’ (that wonderful person extremely), ket ‘much’ : ket=dee ‘too much’, e.g. ket=dee ñuuci ‘too small’. For the comparative and superlative functions badɪ ~ baadɪ ‘(even) more’ and cuu ‘most’ are used, e.g. badɪ ñuuci ‘smaller’ : cuu ñuuci ‘smallest’. Temporal adverbs include esi ‘now’, nee ‘immediately’, sup ~ sup&sup ‘always’, balanaa ‘earlier’, ooronɪ ‘recently’, xaɪsɪ ‘still, again’, guci ‘again’, and elee ‘soon, already’, e.g. elee xoda-xan ‘(it) has already finished’. An aspectual and modal content is expressed by juke ‘barely’ and xaʊlɪ ‘barely, finally’, as in [3]: [3] jue-i ineƞ-du kɪɪƞa-xan, jɪljan juke=dee, two-ord day-dat speak.with.weak.voice-pst voice hardly=ptcp ‘On the second day he spoke weakly, with a hardly audible voice, ɪlɪ-ɪ ineƞ-du xaʊlɪ tuncu-lu-xen three-ord day-dat finally move-inch-pst and on the third day he finally started moving a little.’ Apart from the spatials, which are nominal words, there are few actual postpositions. Some of them, especially items of a verbal origin, require a specific case of the preceding noun. For instance, suƞgurem ‘through’ and acaptɪ ‘facing’ (from aca- ‘to meet, to face’) require the accusative case, e.g. naam-ba-tɪ acaptɪ tee-si-ni ‘s/he is sitting face to face with them’ (3P-ACC-PX3PL against sit-PRS-​3SG), Most adverbs and postpositions in Ulcha have a transparent etymological connection with nominal or verbal roots. The class of truly invariable lexemes comprises mainly a few interjections, e.g. gee ‘well, now, come on’, ma ‘here (take it!)’, bala ‘hurry up!’, enene [moan with pain], anayɪɪ [admiration or surprise]. VERBAL MORPHOLOGY Verbal stems in Ulcha may be classified according to their differences with regard to aorist formation. The factors regulating these differences are both lexical (different aorist markers) and phonological (different stem-final segments): (1a) Stems ending in a short vowel, e.g. japa- ‘to take’, ice- ‘to see’, soƞgo- ‘to cry’, tutu- ‘to run’, dooljɪ- ‘to hear’. In the imperfective participle (based on the aorist) these verbs take the marker -I (< *-rII), which tends to assimilate the stem-final vowel, yielding a long II, which, moreover, is often shortened, e.g. japa-ɪ > japɪɪ [dʑapi:] ~ japɪ [dʑapi], ice-i > icii [itɕi:] ~ ici [itɕi], soƞgo-ɪ > soƞgɪɪ [soŋgi:] ~ soƞgɪ [soŋgi], tutu-i > tutii [tuti:] ~ tuti [tuti]. Sometimes, regressive assimilation may also occur, especially before the suffix of the impersonal participle, e.g. ɪsɪ- ‘to reach’ : PTCP.IMPERS.IMPRF ɪsɪ-wʊ > ɪsʊ-wʊ. As a result, the stem-​final vowel is inherently unstable, which may lead to variation throughout the paradigm, especially between the representations I ~ U ~ Ø, e.g. PTCP.PRF ɪsɪ-xa/n ~ ɪsʊ-xa/n ~ ɪs-xa/n.

Ulcha 421

(1b) Stems ending in a long vowel or a vowel sequence, e.g. (long vowel) saa- ‘to know’, waa- ‘to kill’, nee- ‘to put’, tee- ‘to sit down’, too- ‘to go away from the riverside’, ii- ‘to enter’, sii- ‘to pass’, (vowel sequence) xaɪ- ‘to do what?’, aʊ- ‘to sleep’, daʊ- ‘to cross’, eu- ‘to go down to the river’, ñie- ‘to go out’. In the imperfective participle these verbs take the marker -rII, normally shortened to -rI, e.g. saa-rɪɪ > saa-rɪ, eu-rii > eu-ri. The verbs baka- ‘to find’ and tuku- ‘to fall’ preserve the medial k in Ulcha and belong to group (1a), while their cognates in Nanai, baa- and tuu-, are long-vowel stems. (2a) Stems ending in the obstruent p. There are very few verbs of this type, the principal examples being jep- ‘to eat’ and guup- ‘to extinguish’. In the imperfective participle these verbs take the marker -tII (< *-rII), normally shortened to -tI, e.g. jep-tii > jep-ti. In the perfective participle these verbs take the marker *-kI/n, which, when combined with the stem-final p, yields the metathetical cluster kp, e.g. jekpin < *jep-kin. (2b) Stems ending in the sonorant consonants n l. While there are many stems ending in the nasal n, e.g. wen- ‘to say’, including also the causatives in -wAAn-, the only examples of stems ending in l in active use are kaal- ‘to stalk’ and maal- ‘to destroy, to be annihilated by an epidemic’. In the imperfective participle these verbs take the marker -dII (< *-rII), normally shortened to -dI, e.g. wen-dii > wen-di, kaal-dɪɪ > kaal-dɪ, while in the perfective participle they take the marker -cI/n, e.g. wen-ci/n, kaal-cɪ/n. It may be noted that the marker -cI/n is in these cases a regular reflex of earlier *-kIn (in the clusters *lk *ƞk), as also attested in Nanai, and does not derive from original *-cI/n (as in group 3). (3) Stems of the category of “change-of-state” verbs, comprising ga- ‘to take’, di- ‘to come’, bu- ‘to die’, o- ‘to become’. In the imperfective participle these verbs take the marker -dII (< *-dII), normally shorted to -dI, i.e. ga-dɪ/ɪ, di-di/i, bu-di/i, o-dɪ/ɪ, while in the perfective participle they take the marker -cI/n, i.e. ga-cɪ/n, di-ci/n, bu-ci/n, o-cɪ/n. Like their cognates in Nanai, the verbs bu- ‘to die’ and o- ‘to become’ have additional idiosyncracies, in that they also have the exceptional stems bul- (< *bur-) and o-cʊ- ~ o-sɪ- ~ o-po- ~ o-pʊ-, e.g. PASS.PTCP.PRS bul-wu, o-sɪ-wʊ ~ o-po-wʊ ~ o-pʊ-wʊ. (4) Stems originally taking -sI- in the aorist. As an obscured trace of this stem type Ulcha preserves several verbs whose stem can be enlarged by the element -sI-, e.g. pul- ~ pul-si- ‘to walk around, to come and go’, xer- ~ xer-si- ‘to call’, en- ~ en-si‘to be sick’, etc. Although the element -sI- in these cases is historically a conjugational marker of this particular group of “stative” verbs, it is often synchronically better analysed as a derivational suffix (stative aspect) or, simply, as a part of the stem. In fact, some of these verbs, notably pulsi- have already been reanalysed as regular short-vowel stems (type 1a). This means that original forms like 3SG PRS *pul-si-ni : PST *pul-ci-ni, as still preserved in Uilta, have in Ulcha been replaced by 3SG PRS pulsi-i-ni : PST pulsi-xe-ni, as also in Nanai. Of the other verbs of this group, the copula bi- has been restructured to an irregular vowel stem, with PTCP.IMPRF bi-i : PTCP.PRF bi-ci/n as the participial forms. As in Nanai, the negation verb *e- : AOR *e-si- preserves the aorist marker -si-, though only in suffixalized position, e.g. saa-r-a-sɪm-bɪ ‘I do not know’ (know-​CONNEG-NEG. PRS-1SG), but unlike in Nanai the corresponding past tense forms are no longer used in Ulcha.

422 Shinjiro Kazama

VOICE AND ASPECT Voice and aspect are expressed derivationally. The voices are the causative, reciprocal, and medial, with forms and functions closely similar to those in Nanai. •

•

•

• • •

•

The causative is formed by the suffix -wAn- (after vowel stems) ~ -bU-wAn- (after consonant stems). The exceptional stems (types 3 and 4) have the forms CAUS ga-wan- ‘to take’, di-wen- ‘to come’, bul-buwen- ‘to die’, opo-wan- ~ osɪ-wan- ~ ocʊ-wan- ‘to become’, and bi-wen- ‘to be’. The causee stands in the object form (ACC or RX), e.g. pikte-yi ƞene-wen-ci-ni ‘s/he made/let her/his child go’ (child-RX go-CAUS-PST-3SG), naam-ba-nɪ eji saa-wan-da ‘do not let her/him know (it)!’ (3P. OBV-ACC-PX3SG PROHIB know-​CAUS-CONNEG). If the verb is a transitive one, the sentence can contain two objects, e.g. tʊ-wa tɪɪ pikte-we olbɪm-bʊwan-dɪ-nɪ ‘s/he made that child bring it’ (that-ACC that child-ACC bring-CAUS-PRS-3SG). The causative can also be used as a device to avoid the change of subject between a subordinate clause and a main clause, e.g. andaxa di-wen-deree bii turgen=dee uñureju-xem-bi ‘a guest came and I quickly cooked for him’ (guest come-CAUS-​ CV.ANT 1SG quick=PTCL cook-PST-1SG). In such cases, the causative can also function as an adversative passive, e.g. am-bɪ bul-buwen-deree tɪɪ ñɪɪ emke bi-i-ni ‘after his father died he is living alone’ (father-RX die-CAUS-​CV.ANT that man alone be-PRS-3SG). The reciprocal is formed by the suffix -mAcI-, e.g. tata-macɪ-xa-tɪ ‘they pulled each other’ (pull-RECIPR-PST-3PL), tɪxala-macɪ-xa-tɪ ‘they agreed with each other’ (agree-​RECIPR-PST-3PL), ñɪɪ-jɪ sorʊ-macɪ-mɪ=daa kewe/n ‘I have never fought with anyone’ (person-​INSTR fight-RECIPR-CV.CONN=PTCL NEG.EXIST). The medial is formed by the suffix -p- ~ -p-tA-. Like its cognate in Nanai, this form has an ambiguous paradigm: the perfective participle in *-p-kI/n > -k-pI/n is based on the simple consonant stem in -p-, while the imperfective participle in -p-tII is based on the vowel stem in -p-tA-, which incorporates the aorist marker -tA- (< *-rA-), e.g. meen pikte-yi cup weedek.pin ‘s/he lost her/his own child’ (REFL child-RX entirely miss.PTCP.PRF), xagdʊn too tɪɪ ice-pti-i-ni elee ‘there appeared a house already’ (house far so see-MED-PRS-3SG already)’. It may be noted that Ulcha, like Nanai and Uilta, does not have an actual passive, which is why passive-like meanings are expressed by other means, including the causative and medial forms, as well as the impersonal participle. The aspectual categories are also formally and functionally very similar to those in Nanai and Uilta and comprise the following: The inchoatives in -lU-, e.g. guci ƞene-lu-xen ‘(we) started going again’ (again go-INCH-​ PST), tuncu-lu-xe-ni ‘(it) started moving’ (move-INCH-PST-3SG). The momentatives in -psIn-, e.g. tɪɪ ñaʊjaka muru-psin-ju-xe-ni ‘that young man sank in deep thought again’ (that young.man think-MOM-REV-PST-3SG). The duratives in -pAcI- ~ -cI-, e.g. ɵm gasa-kan ui-ki-ni xerel-​peci-i-ni ‘a waterfowl is circulating in the sky’ (one waterfowl-DIM above-PROL-PX3SG move.aroundDUR-​PRS-​3SG), dolbo ineƞ yaaya-cɪ-marɪ saman-dɪɪ-tɪ ‘they sang and shamanized night and day’ (night day sing-DUR-CV.CONN.PL shamanize-PRS-3PL). The iteratives in -nAsI-, e.g. tɪɪ jolo-wo ewru-nesi-i-ti ‘everyone tried to lift that (big) stone’ (that stone lift-ITER-PRS-3PL).

Ulcha 423

• • •

The distributives in -ktA-, e.g. buu-kte-mi buu-kte-mi ñɪɪ-sal-tɪ cʊpal ʊmɪ-waan-cɪ-tɪ ‘s/he poured and poured and let everyone drink’ (give-DISTR-CV.CONN giveDISTR-​CV.​CONN person-PL-DIR all drink-CAUS-PST-3PL). The habitives in -lsI- ~ -sI- , e.g. moo-wa mool-sɪ-ɪ, mɵɵ-we mɵɵl-si-i ‘(s/he) is always fetching firewood and water’ (wood-ACC fetch.wood-​HAB-PRS water-ACC fetch.​ water-​HAB-PRS). The reversives (repetitive-reversives) in -jU- (cf. Uilta -du, Nanai -gO-), e.g. meem=​bee-ni daʊ-jʊ-x ‘s/he crossed (the river) towards her own place’ (REFL=towards-PX3SG cross-REV-PST). Some verbs have irregular forms and/or lexicalized meanings, e.g. di- ‘to come’ : REV ji-ju- ‘to come back’, buu- ‘to give’ : REV buu-ju- ‘to return’, saa- ‘to know’ : REV saa-jʊ- ‘to regain consciousness’, too- ‘to go up from the riverside’ : REV too-kʊ- ‘to return up from the riverside’. The verbs referring to the times of the day and seasons can also take this suffix, e.g. sikse- ‘(to become) evening’ : sikse-ju-xen ‘it became evening (again)’, tue- ‘(to become) winter’ : tue-ju-xen ‘it became winter (as usual)’.

PARTICIPLES The Ulcha system of participles (Table 15.6) retains the basic distinctions between the imperfective vs. perfective, personal (active) vs. impersonal (passive), and affirmative vs. negative types. However, the impersonal and negative participles have only the imperfective form. TABLE 15.6 ULCHA PARTICIPLES  

 

V

VV

C

PERS

IMPRF

-I/I

-rI/I

-DI/I

PRF

-xA/n

NEG

-A-sI

IMPRF

-wU

-BUwU

NEG

-w-A-sI

-BUw-A-sI

IMPERS

-cI/n -r.A-sI

p *-kI/n

-D.A-sI

Stem types: V = short-vowel stems, VV = long-vowel stems, C = consonant stems. Segmental alternations: A = a e, U = ʊ u, I = ɪ i, B = b p, D = d t.

Several older descriptions of Ulcha (Petrova, Sunik) mention also the impersonal perfective participle in -wU-xA/n ~ -BUwU-xA/n, but it seems to have been lost in the modern language and is not attested in any actual text material. These descriptions also quote the impersonal imperfective participle in the long form -wU-rI ~ -BUwu-rI, but the element -rI has also been lost in the modern language. The exceptional verbs ga- ‘to take’, di- ‘to come’, o- ‘to become’, bu- ‘to die’, and bi- ‘to be’ form the participles as follows: PTCP.IMPRF ga-dɪ/ɪ, di-di/i, o-dɪ/ɪ, bu-di/i, bi-i; PTCP. PRF ga-cɪ/n, di-ci/n, o-cɪ/n, bu-ci/n, bi-ci/n; NEG-PTCP ga-d-a-sɪ, di-d-e-si, o-d-a-sɪ, bu-de-si, bi-e-si > bi-i-si, PTCP.IMPERS.IMPRF ga-wʊ, di-wu, o-sɪ-wʊ ~ o-po-wʊ, bul-bu, biwu, PTCP.IMPERS.NEG ga-w-a-sɪ, di-w-e-si, o-pow-a-sɪ, bul-buw-e-si, bi-w-e-si. The participles are used in the function of attributive modifiers, sentential arguments, and finite predicates. In the finite function, they express the present (PRS) and past (PST)

424 Shinjiro Kazama

tenses and can be used without person marking, e.g. ƞene- ‘to go’ : bii ƞene-i ‘I go’ (1SG go-PRS) : sii ƞene-i ‘you go’ (2SG go-PRS) : tɪɪ ñɪɪ ƞene-i ‘he goes’ (that person go-PRS). More often, however, they take person markers of the possessive set, yielding personal paradigms of the following types: ice- ‘to see’ (type 1a, short-vowel stems) : PTCP.IMPRF ice-i [itɕi] : PRS 1SG ice-i-yi [itɕi:] : 2SG ice-i-si : 3SG ice-i-ni : 1PL ice-i-pu : 2PL ice-i-su : 3PL ice-i-ti. wen- ‘to say’ (type 2b, nasal stems) : PTCP.IMPRF wen-di/i : PRS 1SG wen-di-yi : 2SG wen-di-si : 3 SG wen-di-ni : 1PL wen-di-pu : 2PL wen-di-su : 3PL wen-di-ci. buu- ‘to give’ (type 1b, long-vowel stems) : PTCP.PRF buu-xe/n : PST 1SG buuxem-bi : 2SG buu-xe-sɪ : 3SG buu-xe-ni : 1PL buu-xe-pu : 2PL buu-xe-su : 3PL buu-xe-ti. wen- ‘to say’ (type 2b, nasal stems) : PTCP.PRF wen-ci/n 1SG wen-cim-bi : 2SG wen-​ ci-si : 3SG wen-ci-ni : 1PL wen-ci-pu : 2PL wen-ci-su : 3PL wen-ci-ti. The negative participle is based on the combination of the connegative form (aorist stem) of the main verb and the suffixalized imperfective participle of the negation verb. The finite forms of this participle may actually involve a merger with the aorist paradigm of the negation verb, which explains why the person marking in these forms represents a mixture of possessive suffixes and predicative personal endings: saa- ‘to know’ (type 1a, short-vowel stems) : NEG-PTCP.IMPRF saa-r-a-sɪ : NEG.PRS 1SG saa-r-a-sɪm-bɪ : 2SG saa-r-a-sɪ-tɪ ~ saa-r-a-sɪ-sɪ : 3SG saa-r-a-sɪ-nɪ : 1PL saar-a-​sɪ-pʊ : 2PL saa-r-a-sɪ-sʊ : 3PL saa-r-a-sɪ-tɪ. In attributive use, the participles remain unmarked when referring to the subject (conjunct use), e.g. di-cin ñɪɪ ‘the man who came’ (come-PTCP.PRF person), but when referring to other constituents (disjunct use) they take possessive marking, e.g. min gaajʊ-​xam-bɪ sʊgdata-wa xaɪ-dʊ nee-xe-si ‘where did you put the fish that I brought?’ (1SG.GEN bring-PTCP.PRFPX1SG fish-ACC what-DAT put-PST-2SG). When used as sentential arguments, participles take regular nominal marking for case and person, e.g. am-bɪ aldacʊ-​xam-ba-nɪ bii aldacʊ-ɪ=ta ‘I will tell you what my father told me’ (father-PX1SG tell-PTCP.​PRF-​ACC-PX3SG 1SG tell-PRS=PTCL). The impersonal (passive) participles have often a modal connotation of permission or possibility, e.g. tɪ-wa naañɪ-da-mɪ jaɪ wem-buwu ‘we may call that (a small boat) jaɪ in Ulcha’ (that-ACC Ulcha-VBLZ-CV.CONN jaɪ say-IMPERS.PRS), xoon/ɪ ta-wʊ ‘what should be done?’ (how do-IMPERS.PRS), xoon=daa ñie-wesi ‘it is impossible to get out (of there)’ (how=PTCP go.out-IMPERS.NEG.PRS). The imperfective impersonal participle is also conventionally used as the dictionary form of verbs. FINITE INDICATIVE FORMS The finite tense forms based on the participles are complemented by the aorist, which represents the original monofunctional set of finite forms. The formation of the aorist stem depends on the stem type of the verb. The aorist marker *-rA- appears as -A- ~ -rA- after short vowel-stems, e.g. ana- ‘to push’ : AOR ana-a- (> ana-) ~ ana-ra(-), as -rA- after long-vowel stems, saa- ‘to know’ : AOR saa-ra-, as -dA- after sonorant stems, e.g. wen- ‘to say’ : AOR wen-de-, and as -tA- after obstruent stems, e.g. jep- ‘to

Ulcha 425

eat’ : AOR jep-te-. The exceptional verbs ga- ‘to take’, di- ‘to come’, o- ‘to become’, bu- ‘to die’, and bi- ‘to be’ form the aorist as follows: ga-da-, di-de-, o-da-, bu-de-, bi-e-. In the personal paradigm, the aorist uses person markers of the predicative type (VX). In the third person plural it takes the nominal plural marker (*)-l, which may, however, also be absent: ʊmɪ- ‘to drink’ (type 1a, short-vowel stems) : AOR ʊmɪ-a- : 1SG ʊmɪ-am-bɪ : 2SG ʊmɪ-a-tɪ : 3SG ʊmɪ-a ~ ʊmɪ-ra : 1PL ʊmɪ-a-mʊ : 2PL ʊmɪ-a-sʊ : 3PL ʊmɪ-a/-l ~ ʊmɪ-ra/-l. The plain aorist stem is also used as the connegative form after the prohibitive and “nondumitive” negators eji ‘do not!’ and eciel/i > eceel/i ‘not yet’, both of which are petrified forms of the negation verb *e-, e.g. eji ice-re ‘do not look (at it)!’ (PROHIB see-CONNEG), eji wen-de ‘do not say (it)!’ (PROHIB say-CONNEG), eji di-de ~ eji di-dere ‘do not come!’ (PROHIB come-CONNEG). The “nondumitive” negator is a cognate of the Nanai past tense negator ecie, which in Ulcha has received the connotation of ‘not yet’ (“nondum”), e.g. naa-n eceel saa-ra ‘s/he does not know (it) yet’ (3P.OBVPX3SG NONDUM know-​CONNEG). As in Nanai, the connegative can in such cases take person marking of the possessive type, e.g. eceel saa-ra-yɪ ‘I do not know (it) yet’ (NONDUM know-AOR-1SG). Since the function of a regular finite present tense is synchronically filled by the finitely used imperfective participle, the aorist has received additional modal and evidential connotations, as also in Nanai. The basic connotation of the aorist seems to be speaker involvement, which is why the third-person forms are used only when the speaker has personally witnessed the event. An implication of surprise (mirativity) may also be present, often enhanced by the enclitic particle =tAnII, e.g. mee-pi saa-ra=tanɪɪ ‘s/he knows it by her/himself’ (REFL-RX know-AOR=PTCL). An important feature of Ulcha is that it lacks the finite past tense or “preterite” (Nanai -kA-) corresponding to the aorist. For this reason, the function of the finite past tense is filled only by the finitely used perfective participle. Ulcha also lacks the future tense markers attested in other Tungusic languages, including both the progressive marker *-jA- (cf. Nanai -jAA/n-) and the future participle in *-jA-ƞAA. The progressive marker is preserved only in a few lexicalized relicts, e.g. ice- ‘to see’ : ice-je- ‘to watch’. For the finite future tense function, Ulcha uses the innovative complex suffix -RIlA/n- (= -IlA/n- ~ -rIlA/n- ~ -dIlA/n- ~ -tIlA/n-), shared with Uilta but not attested in any other Tungusic language. This suffix shows the same initial alternations as the imperfective participle in -RI/I and may be connected with the latter, though the details remain to be elucidated. It is conjugated according to the aorist pattern, e.g. buu- ‘to give’ : FUT-1SG buu-rilem-bi, wen- ‘to say’ : FUT-2SG wen-dile-ti. Exceptionally, however, the third person forms always contain the particle =mA, e.g. jep- ‘to eat’ : FUT 3SG jep-tile=me : 3PL jep-tile-l=me. In the modern language, the third-person singular form in -tIlA=mA tends to replace all other personal forms of the paradigm. IMPERATIVES As in Nanai and Uilta, the second person singular imperative in Ulcha is based on the aorist and marked by the suffix -U ~ -rU ~ -U-rU depending on the stem type. The corresponding

426 Shinjiro Kazama

plural form takes the person marker -sU, preceded by the consonant k, which is unique to Ulcha, e.g. ana- ‘to push’ : IMP 2SG ana-ʊ ~ ana-rʊ : 2PL ana-ʊ-k-sʊ, buu- ‘to give’ : IMP 2SG buu-ru : 2PL buu-ru-k-su, wen- ‘to say’ : IMP 2SG wen-u ~ wen-u-ru : 2PL wenu-k-su, jep- ‘to eat’ : IMP 2SG jep-u ~ jep-u-ru : 2PL jep-u-k-su. The exceptional verbs ga- ‘to take’, di- ‘to come’, o- ‘to become’, bu- ‘to die’, and bi- ‘to be’ have the imperatives 2SG ga-dʊ : ga-dʊ-k-sʊ, di-su : di-su-k-su, osɪ-rʊ : o-dʊ-k-sʊ, bu-du : bu-du-k-su, bi-su : bi-su-k-su. The total paradigm of the imperative (Table 15.7) comprises also the first and third person forms, based on the element -jI- (< *-gI-) and marked by person markers of the possessive type except in the idiosyncratic first person singular form in -jI-tA (< *-gItA). In the modern language, however, the first person forms are no longer actively used. TABLE 15.7 ULCHA IMPERATIVE FORMS  

 

V

SG

1

-jI-tA

2

-(r)U

3

-jI-

1

-jI-

2

-U-k-

3

-jI-

PL

VV

C

-rU

-U(-rU)

PX

-nI -pU -rU-k-

-U-k-

-sU -tI

Stem types: V = short-vowel stems, VV = long-vowel stems, C = consonant stems. Segmental alternations: A = a e, U = ʊ u, I = ɪ i.

Negative imperatives for the second person are formed with the help of the prohibitive negator eji, followed by the connegative form of the main verb, to which the complex person marker -ksU is added in the plural, e.g. tawa-sɪ eji pul-te-ksu ‘do not go there!’ (there-DIR PROHIB go.around-CONNEG-2PL). For the second person there is also the so-called “future imperative” in SG -sA(A)rI : PL -sA(A)r-tU ~ -sA(A)r-sU (cf. Uilta -ssAAri ~ -ttAAri, Nanai -xAArI), e.g. buu- ‘to give’ : IMP.FUT 2SG buu-seeri : 2PL buu-seer-tu. A confusion with the consecutive converb seems to be present in jep- ‘to eat’ : IMP.FUT 2SG jep-seeri ~ jek-peeri : 2PL jep-seer-tu ~ jek-peer-tu (< *-p-k-). OTHER MODAL FORMS Other modal contents are expressed by the subjunctive and optative, which have only a single personal paradigm, as well as by the derivationally formed intentives and desideratives. •

The subjunctive mood is marked by the Common Tungusic suffix -mcA-, which takes a mixed set of person markers, e.g. ana- ‘to push’ : SUBJ 1SG ana-mca-yɪ : 2SG ana-​ mca-​sɪ : 3SG ana-mca : 1PL ana-mca-pʊ : 2PL ana-mca-sʊ : 3Pl ana-mca-l. However, in Ulcha this suffix is normally used only in the dependent clause, while the predicate of the main clause is marked by the complex suffix -(r)IlA-xA/n-, containing the finite future marker in combination with the finitely used perfective participle, as in [4]:

Ulcha 427

[4] min ana bi-mce-yi 1sg.gen neg.exist be-subj-1sg ‘If I had not been there, amba ikte-du-ni bi-ile-xe-si devil tooth-dat-px3sg be-fut-pst-2sg you would have been in the devil’s teeth’. This construction is shared by Uilta, but not by Nanai. While it is still used by some speakers today, there is a tendency in the modern language to replace the personal forms in the dependent clause by the exceptional third-person form bi-mce-n (be-SUBJ-3SG), which, then, functions like a conjunction in the meaning ‘if’, e.g. bii saa-rɪɪ bimcen, belecu-ile-xem-bi ‘if I had known, I would have helped’ (1SG know-PTCP.IMPRF if help-FUT-PST-1SG). There are also examples of -mcA- in the main clause, as in Nanai, often combined with the particle =mA, e.g. naa-tɪ mum-be ice-xe-ti bimcen, waa-mcal=ma ‘if they had seen us, they would have killed us’ (3P.OBV-PX3PL 1PL-ACC seePST-3PL if kill-​SUBJ-​​PL=PTCL). Another particle used in the main clause is =guni, e.g. min am-bɪ tɪ-dʊ bi-mce-n belecu-ile-xe-ni=guni ‘if my father had been there, he would have helped (him)’ (1SG.GEN father-PX1SG that-DAT be-SUBJ-3SG help-FUT-PST-​ 3SG=​PTCL). •

The optative mood is marked by the suffix -ƞA-, before which consonant stems show some special developments. The personal forms are of the subjunctive type, followed by the particle =tAnII, e.g. ana- ‘to push’ : OPT-1SG=PTCL ana-ƞa-yɪ=tanɪɪ ‘let me push!’, wen- ‘to say’ : OPT-2SG=PTCL we-ƞe-s(i)=tenii ‘may you say!’, jep- ‘to eat’ : OPT.3SG=​PTCL jem-ƞe=tenii ‘may s/he eat!’, soƞgo- ‘to cry’ : OPT-1PL=​PTCL soƞgo-ƞo-p(ʊ)=tanɪɪ ‘let us cry!’, bi- ‘to be’ : OPT-2PL=PTCL bi-ƞe-s(u)=tenii ‘may you be!’, buu- ‘to give’ : OPT-3PL=​PTCL buu-ƞe(-l)=tenii ‘may they give!’. The first person forms of both the singular and the plural are no longer actively used in the modern language. The optative functions as a soft admonition with a future tense reference, as in [5]:

[5] meen uleesi-i baa-tɪ-yɪ ƞene-mi refl like-ptcp.imprf place-dir-rx go-cv.conn ‘Go to a place of your own preference and tɪ-dʊ kooka-ƞa-sɪ=tanɪɪ that-dat stay-opt-2sg=ptcl stay there!’ •

•

Intentives are formed by the suffix -(d)ɪcA (cf. Uilta -(k)itA ~ -jitA, Nanai -(k)ɪcA-), e.g. tɪɪ eekte tɪ-wa nee japa-ɪca-xan ‘she intended to take it immediately’ (that woman that-ACC immediately take-​INTENT-PST), tɪ-dʊ xagdʊm-ba anjʊ-ɪca-mɪ joombʊɪ-tɪ ‘at that time the idea occurred to them that they would build a house’ (that-DAT house-ACC make-​INTENT-​CV.CONN come.to.think-PRS-3PL). Desideratives, which normally indicate a physical desire or need, are formed by the suffix -m(A)sI-, e.g. jep- ‘to eat’ : DESID je-msi- ‘to be hungry’, aa- ‘to sleep’ : DESID

428 Shinjiro Kazama

aa-​masɪ- ‘to be sleepy’, cf. also kooja-msɪ- ‘be thirsty’ (with no synchronic base). The element -sI- in the suffix is connected with the marker of stative verbs (type 4). CONVERBS Ulcha has some seven fully grammaticalized converbal forms (Table 15.8), here termed the connective, successive, consecutive, conditional, purposive, terminative, and anterior converbs. In addition, there are a number of quasiconverbal constructions, based on the case forms of participles. As in other Tungusic languages, the converbs may be divided into conjunct (same-subject) and disjunct (different-subject) types, with some converbs being ambivalent. The disjunct converbs take person marking of the possessive type. Some of the conjunct converbs include reflexive markers and can have plural forms. TABLE 15.8 ULCHA CONVERB MARKERS  

 

PL

PX

RX

SS

CONN

-mI

-mArI

+

SUCC

-m-dII

-mArI-dII

+

CSEC

-pI

-pArI

COND

-wUcA-

+

PURP

-bdA-

+

TERM

-dAlA

+

ANT

-(RA-)rAA

+

DS

+ + +

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e, U = ʊ u, I = ɪ i, R = r d t.

•

•

•

The connective converb has originally the forms -mI : PL -mArI, but in the modern language the plural form is receding and can be replaced by the invariant singular form. This converb expresses simultaneous action, e.g. ƞene-mi lʊca gasam-ba-nɪ baka-xa-tɪ ‘when they were going, they found a Russian village’ (go-CV.CONN Russian village-ACC-PX3SG find-PST-3PL), or a succession of events, e.g. bii ɪlalta bi-mi eu-si daʊ-jʊ-ɪ ‘after staying (there) for three days I will cross (the river) to this side again’ (1SG three-day be-CV.CONN here-DIR cross-REV-PRS). It is also used in combination with auxiliary verbs and the adjective aya ‘good’, e.g. ii-mi aya ‘may I come in?’ (enter-CV.CONN good), ñɪɪ gasam-ba-nɪ xecu-mi aya ‘we should drop in other people’s village’ (person village-ACC-3SG drop.in-CV.CONN good). The successive converb in -m-dII : PL -mArI-dII is based on the connective converb, to which the originally enclitic element -dII has been added. This converb, which has a cognate in Uilta and Oroch, but not in Nanai, indicates a prolonged or repeated action after which another action occurs, e.g. tɪɪ baldʊ-mdɪɪ wen-ci/n ‘after living like this (for a long time) s/he (once) said’ (so live-CV.SUCC say-PST). A lexically fixed example of this form is ta-​mdii ‘after doing’ > ‘then’ (do-CV.SUCC), e.g. ta-mdɪɪ xaʊncak ƞene-xe-ni ‘then s/he fainted’ (do-CV.SUCC unconscious go-PST-3SG). The consecutive converb also originally has the forms -pI : PL -pArI, but the plural form is no longer used in the modern language. This converb expresses temporal or conditional subordination, e.g. aʊ-pɪ dolbo sene-jʊ-xe-ni ‘s/he slept, and then woke

Ulcha 429

•

•

•

•

•

up in the night’ (sleep-CV.CSEC night wake.up-REV-PST-3SG), tuku-xen, bekte bi-pi, tee-ju-xen ‘s/he fell down, and after a long time, got up again’ (fall-PST much be-CV.CSEC get.up-REV-PST), tɪɪ biruu-ti ƞene-pi bak-ɪla-tɪ ‘they will find (it), if/ when they go that village’ (that village-DIR go-CV.CSEC find-FUT-3PL). The conditional converb in -wUcA- is always used in a disjunct function. Like the conjunct consecutive converb, this form indicates temporal or conditional subordination, e.g. edi-ni too-wʊca-nɪ xeeri-ƞde-xe-ni ‘when her husband came up to the mountainside, she went to call him’ (husband-PX3SG go.up-CV.COND-PX3SG call-AND-PST-3SG), sin xaa-​wʊca-​sɪ tɪɪ mapa wen-dile ‘when you dock your boat, grandfather will tell you’ (2SG.GEN dock-CV.COND-PX2SG then grandfather sayFUT). As in Nanai, condition can also be expressed with the help of the petrified form o-sɪ-nɪ ‘if’ (become-AOR-3SG), e.g. baa uleen osɪnɪ ƞene-wu ‘if the weather is nice, we should go’ (sky good if go-IMPERS.PRS), ye-du baldʊ-wʊ osɪnɪ largɪ uleen-ji baldʊ-wʊ ‘if we live here, we will live very well’ (this-DAT live-IMPERS. PRS if good good-INSTR live-IMPERS.PRS). The purposive converb in -bdA- is ambivalent and can take both possessive and reflexive marking, e.g. ice-bde-ni xeeri-ƞde-ween-ci/n ‘s/he made him go and call so that he could see (it)’ (see-CV.PURP-PX3SG call-AND-CAUS-PST), je-bd-i bakaasɪ ‘s/he cannot find anything to eat’ (eat-PURP-RX find-NEG.PRS). The terminative converb in -dAlA is mainly attested in fixed expressions like ele-dele ‘(eat) to the full’ (eat.enough-CV.TERM), sagdan-​dala ‘till old age’ (grow.old-CV. TERM), e.g. sagdan-dala jepu-we-ci-mi bi-le=me ‘(they) will feed us until we grow old’ (grow.old-​CV.TERM eat-CAUS-DUR-CV.CONN be-FUT=PTCL). The anterior converb is marked by the suffix -(RA-)rAA depending on the stem type. The exceptional verbs ga- ‘to take’, di- ‘to come’, o- ‘to become’, bu- ‘to die’, and bi- ‘to be’ have the forms ga-daraa, di-deree, o-daraa, bu-deree, bi-ree. This converb can occur in sequences, e.g. eu-reree ƞene-ree gapakala-raa too-bʊ-jʊ-xa-nɪ ‘s/he went down to the river, took (it), and came up to the mountainside again’ (go. down-CV.ANT go-CV.ANT seize-CV.ANT go.up-CAUS-REV-PST-3SG). This converb has also the negative form eƞ-dee, which requires the connegative form of the main verb, e.g. eƞ-dee aʊ-ra ƞene-xe-ni ‘s/he left without sleeping’ (NEG-CV.ANT sleep-CONNEG go-PST-3SG), erkee ƞene-wu, eƞ-dee yada-a ta-bd-ɪ ‘we should go slowly, so as not to get tired’ (slow go-IMPERS.PRS NEG-CV.​ANT get.tired-CONNEG do-CV.PURP-RX). A similar negative meaning can also be expressed by the combination of the connective converb and the instrumental case form of the negative existential noun kewe/n, e.g. aʊ-mɪ=daa kewen-ji ƞene-xe-ni ‘s/he left without sleeping’ (sleep-CV.CONN=PTCL NEG.EXIST-INSTR go-PST-3SG). A synchronically transparent quasiconverbal construction with an immediative meaning is formed by combining the instrumental case form of the perfective participle with the postposition gese ‘together’, e.g. je-mi xʊda-xan-jɪ-yɪ gese aaƞ-nɪ ƞenexe-ni ‘no sooner had his brother finished eating than he went away’ (eat-CV.CONN finish-PTCP.PRF-​INSTR-​RX together elder.brother-​PX3SG go-PST-3SG). The analogous construction in Nanai is based on the imperfective participle.

SYNTAX The syntactic properties of Ulcha are very similar to those of Nanai and Uilta. For instance, modifiers normally precede their headword with no agreement in number or case, but

430 Shinjiro Kazama

quantifiers can follow a case-marked noun and take the same case marking, suggesting that they have an ambivalent status, e.g. gee, mamaca, ulse-we-ni egdi-we jep-uru ‘hey, grandmother, eat much meat’ (INTERJ grandmother meat-ACC-PX3SG much-ACC eatIMP.2SG). Quantifiers, including numerals, can also function as independent headwords, e.g. tʊtaraa juel-be aʊnja-xan ‘then s/he stayed two nights’ (then two-ACC stay-PST). In subordinated constructions, agreement with the subject is shown by the reflexive suffixes, e.g. min-ti di-dii-du-yi epem-be ga-saarɪ ‘when you come to me, please buy a loaf of bread!’ (1SG-DIR come-PTCP.PRS-DAT-RX bread-ACC take-IMP.FUT). If the subject of the subordinate clause is different from that of the main clause, the first and second person subjects are expressed by the adnominal genitive forms of the corresponding pronouns in combination with the possessive suffixes, e.g. min dabda-xam-bɪ osɪnɪ min pikte-we-yi sii japa-ɪ ‘if I lose, (you) will take my child’ (1SG.GEN lose-PST-1SG if 1SG.GEN child-ACC-​PX1SG 2SG take-​PRS), bii sin ƞene-xe-si xamɪ-la-nɪ akpandɪla=ma ‘after you leave I will go to bed’ (1SG 2SG.GEN leave-PTCP.PRF-PX2SG back. side-LOC-PX3SG lie-FUT=PTCL). A third person possessive suffix in a sentence with a third person suffix refers to a separate (“fourth”) person, cf. e.g. tɪɪ ñɪɪ meen xagdʊn-dʊ-yɪ sʊgdata-ƞgʊ-wa-nɪ jekpi-ni ‘he ate his (= another person’s) fish in his (own) house’ (that person REFL house-DAT-RX fish-AL-​ACC-PX3SG eat.PST-3SG) vs. tɪɪ ñɪɪ meen xagdʊn-dʊ-yɪ sʊgdata-ƞgʊ-yɪ jekpi-ni ‘he ate his (own) fish in his (own) house’ (that person house-DAT-RX fish-AL-RX eat.PST-3SG). Correlation between two clauses can also be expressed by interrogative pronouns, e.g. ƞui deƞsi-esi tɪɪ ñɪɪ jep-tesi ‘he who does not work shall not eat’ (who work-NEG.PRS that person eat-NEG.PRS). In the following, a few other features of Ulcha syntax will be discussed: •

•

•

Basic unmarked equative sentences do not normally require a copula, but in folkloric texts the copula ta- ‘to do’ is occasionally used for the first and second persons, e.g. bii sin neu-si ta-a-m(bɪ)=tanɪɪ ‘I am your younger brother’ (1SG 2SG.GEN younger.sibling do-​AOR-​1SG=​PTCL), buu baa ejen ta-a-pʊ ‘we are the lord of heaven’ (1PL sky master do-​AOR-​1PL). For past tense reference and negation, the copula bi- ‘to be’, is used, e.g. tɪɪ mapa daayɪ mapa bi-ci-n=guni ‘that grandfather was a great grandfather’ (that grandfather big grandfather be-PST-3SG=PTCP), ye-du bi-esi-ni ‘(it) is not here’ (thisDAT be-NEG.PRS-​3SG). The copulas bi- ‘to be’ and o- ‘to become’ also function as auxiliaries in complex tense forms, e.g. meen gasa-nca-sal-ba cʊpal jepe-wen-ci-ni bi-cin ‘he had been feeding all the people of his village’ (REFL village-​inhabitant-PL-ACC all eat-CAUS-PST-3SG be-PST), tʊksa-wa=daa beice-i o-cɪ-n/ɪ ‘he began to hunt for rabbits as well’ (rabbit-ACC=PTCL hunt-PTCP.PRS become-PST-3SG). Other auxiliaries, used to express aspectual and modal contents in combination with the connective converb of the main verb, include korpɪ- ‘to have time to do’, e.g. ñɪɪ tɪs bu-cin, malxʊ, ƞui korpɪ-ɪ-nɪ xumu-mi ‘so many people died, who will have time to bury them?’ (person very die-PST many who have.time-PRS-3SG bury-CV. CONN), mute- ‘to be able to’, e.g. xoonɪ=daa japa-mɪ mute-esi ‘there is no way to catch (it)’ (how=PTCL catch.CV.CONN be.able-NEG.PRS), xawa-sɪ=daa ƞene-mi mute-wesi ‘it is impossible to go anywhere’ (where-DIR=PTCL go-CV-CONN beable-IMPERS.NEG.PRS). Negation in Ulcha is split, as in Nanai, between different areas of morphology and syntax. The negation of finite predicates in the present tense takes place by the negative forms of the verb (based on the suffixalized imperfective participles and/

Ulcha 431

•

•

•

or aorist forms of the negation verb), while for the negation of commands and past tense forms the specific negators PROHIB eji and NEG.PST eciel (‘not yet’) are used in combination with the connegative form of the main verb. Privation is expressed by the nominals anaa and kewe/n (cf. kewke ~ keweke in the “Lower Amur” dialect of Nanai), of which the former functions as a privative marker (‘not having, without’), while the latter is a negative existential noun (‘absence, not being there’), cf. e.g. bii pikte anaa-yɪ ‘I do not have a child’ (1SG child PRIV1SG), min-du pikte kewe/n ~ kewe-ni ‘I do not have a child’ (1SG-DAT child NEG. EXIST ~ -PX3SG). In combination with a preceding connective converb of a verb, kewe/n is also used as an alternative to express past tense negation, e.g. tisee bii tɪ-wa saa-mɪ kewem-bi ‘yesterday I did not know about it’ (yesterday that-ACC know-CV.CONN NEG.EXIST-PX1SG). This construction is unique to Ulcha and may be due to Ghilyak influence. Another (quasi-)negator used in combination with the connective converb is baɪbaɪ ‘impossible, useless’ (← Chinese baibai 白白), e.g. xoon=daa ta-mɪ=daa baɪbaɪ ‘there is no way to do it’ (how=PTCL do-CV. CONN=PTCL impossible)’. Interrogation in polar questions is expressed by the interrogative particle =nUU, e.g. di-dii-si=nuu ‘are you coming?’ (come-PRS-2SG=INTERR). When attached to interrogative pronouns, =nUU expresses indefiniteness, e.g. tɪɪ namʊ=nʊʊ xaɪ=nʊʊ kɪra-dʊ-nɪ eke-sel bi-i-ti ‘there were women at the seashore or somewhere there’ (so sea=INTERR what=​INTERR woman-PL be-​PRS-3PL), while in alternative questions it conveys a notion of uncertainty or wonder, e.g. tɪɪ-sal tɪ-wa waa-xa-tɪ=nʊʊ, xoon ta-xa-tɪ=nʊʊ ‘I wonder if they killed him, or what did they do?’ (that-PL thatACC kill-PST-3PL=INTERR how do-PST-​3PL=INTERR). Occasionally, =nUU is attested in a context that could be interpreted as a subordinate question, e.g. xoon ƞenexe=nuu, ƞui=dee saa-rasɪ ‘nobody knows how they went there’ (how go-PST=INTERR who=PTCL know-NEG.PRS). Corrogation in content (non-polar) questions is expressed by the corrogative particle =kAA, e.g. xaɪ-mɪ aʊ-rɪɪ-sɪ=kaa ‘why are you sleeping?’ (do.what-CV.CONN sleepPRS-​2SG=​CORR), xaɪ-wa mʊdala-ɪ-sɪ=kaa ‘what are you frowning at?’ (whatACC frown-PRS-​2SG=CORR), xoon t(a)-ɪla-sɪ=kaa ‘how will you do (it)?’ (how do-FUT-2SG=CORR). This particle is also used on negated predicates to express cross-questioning or request of confirmation, e.g. tɪ-dʊ bi-esi=kee ‘it is there, isn’t it?’ (that-DAT be-NEG.PRS=CORR), largɪ naa-nɪ, tʊ-wa ɪs-xa-tɪ bi-esi=kee ‘they reached a good place, didn’t they?’ (good place-PX3SG, that-ACC reach-PST-3PL be-NEG.PRS=CORR). Probably the same element is also used as a focus particle, e.g. bii=kee tɪmana too-rɪɪ-yɪ ‘I will go hunting in the mountains tomorrow’ (1SG=PTCL tomorrow go.up-PRS-1SG). Emphasis is expressed by the enclitic particle =dAA, which has a variety of functions, including cumulative or coordinative (‘also, and’), e.g. tuci=dee bi-i ɪƞda=daa bi-i ‘there are both a sled and dogs’ (sled=PTCL be-PRS dog=PTCL be-PRS), tɪɪ ekte=dee xaɪs tɪɪ=meeci tɪɪ ta-xa-nɪ ‘she also did in the same way’ (that woman=PTCL again so=PTCL so do-PST-​3SG). This particle is also used in the connegative function on interrogative pronouns in combination with negation, e.g. xaɪ=daa kewe/n ‘there was nothing’ (what=PTCL NEG. ​EXIST), ƞui-ti=dee xaɪ-wa=daa jaralɪ-asɪ ‘s/he did not say anything to anybody’ (who-​DIR=PTCL what-ACC=PTCL speakNEG.PRS). In clause-final position, =dAA can express a concessive relationship (‘although’), as in [6]:

432 Shinjiro Kazama

[6] naañɪ gasan baka-xan=daa Ulcha village find-pst=ptcl ‘Although they found an Ulcha village, tɪ-la=daa xecu-mi=dee that-loc=ptcl drop.in-cv.conn=ptcl they did not drop in there, either.’ •

kewen neg.exist

Other enclitic particles include =gdAl/I [contrast], e.g. gee tɪ-wa=gdal doolda-raa di-ci-ti ‘now, they heard it and came here’ (INTERJ that-ACC=PTCL hear-CV.ANT come-​PST-​3PL); =gUnI [emphasis of predicate], e.g. tɪ-wa xaʊlɪ japa-jʊ-xa, yebe= guni ‘s/he finally caught (it) and was relieved’ (that-ACC finally catch-REV-PST useful=PTCL); =lA [topic], especially on esi=le ‘now’, e.g. esi=le xawa-sɪ=daa eji ƞene-re ‘now do not go anywhere!’ (now=PTCL where-DIR=PTCL PROHIB go-CONNEG); =m ~ =m=dA [quotative], e.g. wen-dii-ni baɪ wen-cim-bi=m=de ‘s/ he said: I did nothing but say it’ (say-PRS-3SG only say-PST-1SG=QUOT=PTCL); =mAl/I ‘nothing but’, e.g. beye=mel daʊ-xan=daa aya=gʊnɪ ‘it is fine if you just take yourself across the river’ (body=PTCL cross-PST=PTCL good=PTCL), sʊgdata-jɪ=mal balda-ɪ ‘s/he lives on fish only’ (fish-INSTR=PTCL live-PRS), tɪ-dʊ simata=malɪ ‘there was nothing but snow’ (that-DAT snow=PTCL); =meet ~ =meeci [similative], e.g. tɪɪ=meet ta-xa-tɪ ‘they did (it) like that’ (that=PTCL do-PST-3PL), tɪɪ ñɪɪ min=meet bi-esi-ni ‘he is not like me’ (that person 1SG.GEN=PTCL be-NEG. PRS-3SG), degde-i-ni=meeci ƞene-i-ni ‘s/he goes as if s/he was flying’ (fly-PTCP. IMPRF=PTCL go-PRS-3SG), tɪɪ pikte xerekte-ni cʊpal sele=meeci bi-i-ni ‘that child’s skin is like made of iron’ (that child skin-​PX3SG all iron=PTCL be-PRS3SG); =mek ‘at last’, e.g. tɪ-dʊ=mek saa-jʊ-xa-nɪ ‘at that time s/he finally recovered her/his consciousness’ (that-DAT=PTCL know-REV-PST-3SG); =tA [emphasis of intention of the first person], e.g. bii daʊ-rɪɪ=ta ‘I do cross the river’ (1SG crossPRS=PTCL); =tAmA [focus], e.g. buu=teme ɪla-lta bi-mi eu-si too-kʊ-ɪ ‘we will come back here to the mountainside in three days’ (1PL=PTCL three-day be-CV. CONN here-DIR go.up-REV-PRS); =tAnII [topic], e.g. daʊ-mɪ=tanɪɪ aya=gʊnɪ ‘it is all right to cross the river (there)’ (cross-CV.CONN=PTCL good=PTCL), [emphasis of predicate] ñɪɪ peji-ki-ni bi-uxen osɪnɪ tɪɪ tuk-ile-si=tenii ‘if you are inferior to common people, you will remain lying’ (person under-PROL-PX3SG be-IMPERS. PST if so fall-FUT-2SG=PTCL).

LEXICON AND LANGUAGE CONTACTS Although lexical differences within the Nanaic group often link Ulcha with its closest relative Uilta, rather than with Nanai, there are also many lexical items shared by Ulcha and Nanai, but not by Uilta, e.g. Ulcha uƞtuxu/n ‘shaman’s drum’ = Nanai uƞcuxen vs. Uilta daali, Ulcha ñaʊja/n ‘young’ = Nanai naonjoan vs. Uilta purige, Ulcha di- ‘to come’ = Nanai ji- vs. Uilta sinda-, Ulcha nee- ‘to put’ = Nanai nee- vs. Uilta ekse-, etc. In many of these cases, Ulcha and Nanai retain Common Tungusic heritage, while Uilta shows an innovation. In some cases, however, Ulcha itself shows a language-specific innovation: for instance, the cognates of Ulcha xagdʊ/n ‘house’ in the other Tungusic languages,

Ulcha 433

including both Nanai and Uilta, mean ‘bear’s den’. A typical Ulcha word is the negative existential kewe/n, though it is also attested in “Lower Amur” Nanai in the form kew(e)ke. Because of its geographical location, Ulcha has few, if any, direct borrowings from Mongolian, Manchu, or Chinese, for most foreign items borrowed from the south have been transmitted by Nanai, unless they were transmitted already to Proto-Nanaic, cf. e.g. Ulcha daʊsʊ/n = Nanai daosun ‘salt’ ← Manchu dabsun ← Mongolic *dabu-su/n. In some cases, however, Ulcha shows an idiosyncratic form, different from both Nanai and the possible source languages, e.g. Ulcha daƞpɪ ‘tobacco’ vs. Nanai damaxɪ ‘tobacco’ ← Mongolic tamaxi vs. Manchu dambagu. The major non-Tungusic contact partner of Ulcha, disregarding the very recent impact of Russian, has been Ghilyak or, more exactly, Amur Nivkh. Considering that Ghilyak is the probable substrate language of Ulcha, surprisingly few Ghilyak items have so far been identified in Ulcha. Examples can nevertheless be found in the cultural lexicon, e.g. ma.kʊrɪ ‘dried fish’ ← Ghilyak ma, mosɪ/n [fish-skin jelly with berries] ← Ghilyak mos. A non-cultural item, shared also with Neghidal is Ulcha malxʊ/n ‘many, much’ = Neghidal malxʊn ← Ghilyak malʁo-n- vs. Nanai egji, Uilta bara/n. In a larger areal context, Ulcha and Ghilyak share many vocabulary items, reflecting more ancient language contacts in both directions and involving also the other languages of the region. Ulcha has a rich vocabulary of descriptive words, many of which have an exceptional structure, including reduplication and phonotactic peculiarities, e.g. cum&cum [sharp, acute], lap [manner of grasping something], pos [penetrating or drilling through something], kɪcatak [manner of rising or springing to one’s feet quickly], etc. The oral literature of Ulcha speakers is basically very similar to that of the Nanai and comprises the genres of nɪƞmaa/n ‘folktales’, teeluƞgu ‘legends’, and sɪaxʊrɪ ‘foreign stories’. Alliteration is used as a common stylistic device, e.g. xecuwes xewendu, biwes bɪrandʊ, odoncʊra oonjodʊnɪ, xasta xamɪlanɪ, waƞta wayaalani, sisi sirgeduni ɵm xagdʊ baldaxanɪ . . . The traditional folktales in Ulcha begin with the phrase daa xajɪlatɪ ~ daa xaɪjilatɪ, apparently containing the interrogative stem xaɪ(-) ‘(to do) what?’. REFERENCES AND FURTHER READING Angina, S. V. [С. В. Ангинa] (1993) Русско-ульчский разговорник [Russian-Ulcha phrasebook], Богородское [Bogorodskoe]: Издательский отдел при Упримформпечати админиcтрации Хабаровского края. Avrorin, V. A. [В. А. Аврорин] (1981) ‘Ульчские тексты: Материалы для грамматических и этнографических исследований’ [Ulcha texts: Material for grammatical and ethnographical studies], in: Морфология имени в сибирских языках [Nominal morphology in Siberian languages], 129–164, Новосибирск [Novosibirsk]: СО АН СССР. Dechuli, K. F. [К. Ф. Дечули] & L. I. Sem [Л. И. Сем] (1992) Букварь для 1 класса ульчских школ, Санкт-Петербург [St. Petersburg]: “Просвещение”. Ikegami, Jirō (1961) ‘On the Santan vocabularies of the eighteenth and nineteenth centuries’, Ural-Altaische Jahrbücher 33 (1–2): 73–77, Wiesbaden. Republished in: Jirō Ikegami, Researches on the Tungus Languages, 309–317, Tokyo: Kyuko Shoin (2001). Ikegami, Jirō (1967)「サンタン言葉集」 [A collection of Santan words],『北方文化研 究』Bulletin of the Institute for the Study of North Eurasian Cultures 2: 27–87, Sapporo. Ikegami, Jirō (1999)「サンタンことば拾遺」[Gleanings for a collection of Santan words], 『満族史研究通信』8: 1–9, 東京 [Tokyo]: 東洋文庫.

434 Shinjiro Kazama

Kazama, Shinjirō 風間伸次郎 (1996–2010)『ウルチャ口承文芸原文集』[Ulcha oral literature] [1/1996] 鳥取 [Tottori]: 鳥取大学教育学部; [2/2002] 吹田 [Suita]: 大 阪学院大学情報学部; [3/2006] 札幌 [Sapporo]: 北海道大学大学院文学研究科; [4/2008, 5/2010] 東京 [Tokyo]: 東京外国語大学 =『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 9, 20, 30, 43, 49. Kazama, Shinjirō 風間伸次郎 (1998)『ツングース諸語におけるウルチャ語の位置 について』[On the genetic position of Ulcha among the Tungusic languages],『少数 民族言語調査報告』[Fieldwork Reports on Minority Languages], 東京 [Tokyo]: 東 京大学文学部. Kazama, Shinjirō 風間伸次郎 (2004)「ツングース諸語におけるIII群の形成につい て」 [On the formation of group III of the Tungusic languages],『環北太平洋の言 語』 Languages of the North Pacific Rim 11: 91–114, 札幌 [Sapporo]: 北海道大学. Kazama, Shinjirō (2008) ‘The diachronic development of the Group III of Tungusic languages’, Linguistic Typology of the North 1: 103–123, Fuchu: Tokyo University of Foreign Studies. Kazama, Shinjirō 風間伸次郎 (2009)「ニブフ語と近隣諸言語との類型的異同・ 言語接触について」[Topics in the typology and language contact of Nivkh and its neighbouring languages], in: 津曲敏郎 [Tsumagari Toshirō] (ed.),『サハリンの言語 世界』[The linguistic world of Sakhalin], 127–144, 札幌 [Sapporo]: 北海道大学大 学院文学研究科北方研究教育センター. Kazama, Shinjirō & Shinko Ogihara & I. P. Rosugbu (2007) Булавинское наследие— Мифы, легенды и сказки ульчей Булавы [The heritage of Bulava—Myths, legends, and tales of the Ulcha of Bulava], 千葉 [Chiba]: 千葉大学. Majewicz, Alfred F. & Elżbieta Majewicz (1985) ‘Bronisław Piłsudski’s Olchan-Polish glossary, with English equivalents’, Lingua Posnaniensis 27: 71–96, Poznań. Petrova, T. I. [Т. И. Петрова] (1936) Ульчский диалект нанайского языка [The Ulcha dialect of the Nanai language], Moсква [Moscow] & Ленинград [Leningrad]: Государственное учебно-педагoгическое издательство. Rosugbu, A. O. [А. О. Росугбу] (1977) ‘О ложной этимологии топонимов Ульчского района’ [On some false etymologies of the toponyms of the Ulcha raion], Филология народов Дальнего Востока (ономастика) [Philology of the peoples of the Far East (onomastics)], 64–66, Владивосток [Vladivostok]. Schmidt, P[eter] (1923) ‘The language of the Olchas’, Acta Universitatis Latviensis 8: 229–288, Riga. Stoinova, N. M. [Н. М. Стойнова] (2018) ‘Глагольное отрицание в ульчском языке: полевые данные’ [Verbal negation in Ulcha: field data], Сборник научных статей по материалам шестой конференции-школы “Проблемы языка: взгляд молодых ученых” [Problems of language: a view from young scholars], 721–734, Moсква [Moscow]: “Канцлер”. Sunik, O. P. [О. П. Суник] (1968) ‘Ульчcкий язык’ [The Ulcha language], in: Монгольские, тунгусо-маньчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 149–171, Ленинград [Leningrad]: “Наука”. Sunik, O. P. [О. П. Суник] (1985) Ульчcкий язык: Исследования и материалы [The Ulcha language: Studies and materials], Ленинград [Leningrad]: “Наука”. Sunik, O. P. [О. П. Суник] (1987) Словарь ульчско-русский и русско-ульчский [Ulcha-​ Russian and Russian-Ulcha dictionary], Ленинград [Leningrad]: “Просвещение”.

Ulcha 435

Sunik, O. P. [О. П. Суник] (1997) ‘Ульчcкий язык’ [The Ulcha language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 248–260, Москва [Moscow]: “Индрик”. Val’dyu, G. G. [Г. Г. Вальдю] (2005) Словарь ульчско-русский и русско-ульчский [Ulcha-Russian and Russian-Ulcha dictionary], Богородское [Bogorodskoe]: Амурский маяк.

CHAPTER 16

UILTA Toshiro Tsumagari and Yoshiko Yamada

Uilta is a language of the Nanaic group spoken on the island of Sakhalin by a very small number of people. Taxonomically, Uilta is closely related to Ulcha, though it is separated from the latter by several idiosyncratic innovations, which must have taken place after its translocation from the Lower Amur basin to Sakhalin some centuries ago. On Sakhalin, the Uilta speakers occupy a belt in the central part of the island, where their historical neighbours are the Ghilyak or, more exactly, the Sakhalin Nivkh (Nighvng) in the north and the Sakhalin Ainu (Enchiw) in the south. Sakhalin (in Japanese known as Karafuto), was until the mid 19th century a part of the Qing empire of the Manchu, but in the subsequent decades it came to form a frontier between Russian and Japanese territorial interests. In the period 1905–1945 the island was officially divided between a Russian (northern) and a Japanese (southern) part, until it, after World War II, came in its entirety under the jurisdiction of Russia (the Soviet Union). During the decades of political division, the Uilta speakers were correspondingly divided into a northern (Russian) and a southern (Japanese) group. After the war, a large part of the Uilta on the Japanese side were evacuated to the Japanese island of Hokkaido, where some speakers of the language survived until the 2000s. The ethnonym Uilta is used as an endonym by a part of the Uilta speakers and is today considered to be the politically correct appellation of the ethnic group, used also in Russian (úl’ta or úil’ta ~ uíl’ta) and Japanese (uiruta ~ wiruta). Formally, Uilta, phonemically u(y)ilta or wilta, is a regular cognate of the ethnonym Ulcha (*ulca), an item of unknown origin but often thought to derive from the Uilta word ulaa ‘reindeer’, a popular etymology stimulated by the fact that the Uilta are traditionally engaged in small-scale reindeer herding. Another name of the Uilta is Orok, also used in Russian (órok : PL óroki) and Japanese (orokko), which likewise is popularly thought to be connected with the Tungusic word oron ‘reindeer’ (an item actually absent in Uilta). Among the Uilta there is today some confusion and division of opinions as to which name they should prefer, and some Uilta actually identify themselves as Orochon (Russian MASC orochón : FEM orochónka : PL orochóny), a name which should not be confused with the appellations used for the Oroch and Orochen in the technical sense. Historically, the neighbouring ethnic groups on Sakhalin have identified the ancestors of the Uilta as “Uryangkhai” (*uraƞka-ta), which has yielded the exonyms used for them by the Sakhalin Ainu (urakata) and Sakhalin Ghilyak (orƞař). Well into the 20th century the Uilta used to live a nomadic life very similar to that of the Siberian Ewenki and Ewen, with a subsistence economy based on hunting, fishing, and reindeer herding, and with the conical tent as the principal form of dwelling. After World War II all Uilta have lived in fixed settlements. Reindeer herding is still maintained on a limited scale among a group of Uilta today living around the village of Val, close DOI: 10.4324/9781315728391-16

Uilta 437

to the mouth of the river Val on northeastern Sakhalin. This group represents mainly the descendants of those Uilta who lived on the Russian side during the political division of the island. Although the Uilta on the Japanese side were largely evacuated to Hokkaido, some individuals remained, leaving descendants who today live in the basin of the river Poronai, close to the city of Poronaisk on southern Sakhalin. This geographical and historical division corresponds also to a dialectal division of Uilta into a northern and a southern variety. The Uilta have always been a small ethnic group with a population of no more than a few hundred individuals at most. According to the 1989 census, the total population of the Uilta was c. 200, of which 44.7 per cent (89 persons) spoke Uilta as their mother tongue. Recent data give numbers as high as 300–400 individuals, but the number of native speakers has suffered a steady decrease down to between 25 (according to Ozoliņa 2001) and only 16 (according to E. A. Bibikova, p.c. 2007). Today, only a couple of fully fluent speakers survive, with Russian having become the only language of almost all ethnic Uilta. Historically, there has also been some bilingualism between Uilta and Ewenki, which has been present on Sakhalin since the mid 19th century, and there are still a few bilingual semispeakers with a fragmentary knowledge of both languages. As it is, Uilta is today a critically endangered language. Some efforts have been made to revitalize the language and to produce teaching materials, but the results still remain to be seen. DATA AND SOURCES The first noteworthy record of the Uilta language was made by a Japanese explorer in the mid-19th century, supposedly Matsuura Takeshirō, who wrote down about 350 Uilta words in Japanese syllabic transcription (Ikegami 1971, 1993b). A more substantial documentation of the language, including some 2000 words with grammatical remarks and short texts, was made by Bronisław Piłsudski in the early 20th century (Majewicz 1985ab, 2011, Ikegami 1985, Tsumagari 1985b, 1987, 2014). During the Japanese rule of southern Sakhalin, some Japanese linguists also investigated the Uilta language. Among them, Nakanome Akira (1917) published a grammar with a glossary of some 1000 words, later followed by a German translation (1928). Magata Hisaharu started working on Uilta in 1928 and collected a substantial corpus of vocabulary, which was, however, published only much later (Magata 1981). Another early scholar was Kawamura Hideya, who collected lexical material and folklore (Ikegami ed. 1983). After World War II, strictly linguistic research on Uilta was taken up by Ikegami Jirō, who worked with the Uilta speakers who had been evacuated from Sakhalin to Hokkaido. Over the years, Ikegami published a large number of papers, dealing with different aspects of the Uilta language, including phonology (Ikegami 1953, 1955), morphology (1956, 1959, 1968, 1973, 1993a, 2001b), kinship terms (1970), and language contacts (1990). He also collected and published texts (1984, 2002, 2007) and compiled a dictionary (1997), which, with some 4500 entries, is the most reliable lexical source on Uilta. The field data collected by Ikegami represent the southern dialect of Uilta, but information from Soviet and Russian sources, as well as, later, contacts with speakers of the northern dialect allowed him to make conclusions about the differences between the southern and northern dialects (1994a). Work with speakers of the northern dialect also resulted in his activity on the issue of language revitalization and the development of a written language for Uilta. Following a series of preliminary papers on orthographical issues (1994b, 1996, 1998), this work finally resulted in the first ever Uilta primer,

438 Toshiro Tsumagari and Yoshiko Yamada

published in collaboration with native Uilta speakers, and with the financial support of a local enterprise (Ikegami & Bibikova & al. 2008). Ikegami’s work stimulated the descriptive analysis of Uilta morphophonology by Hayata Teruhiro (1979). Also, following Ikegami’s framework, Tsumagari Toshirō took up the study of Uilta. After a couple of papers dealing with the lexicon and accentology (Tsumagari 1980, 1983), he authored an outline of Uilta grammar (1985a, revised in 2009a, which is the basis of the present chapter). He also published on language contacts (2009b) and syntax (2010). More recently, Uilta has become the research focus of Yamada Yoshiko, who has worked with speakers of the northern dialect and published several papers on grammatical issues (2007, 2008, 2010, 2017a) and dialectology (2009, 2017b). She has also collected and published various types of texts with grammatical analysis (2011abc, 2012, 2014, 2015a, 2016a), as well as collections of sample sentences (2015b, 2016b, 2017c). An important text published under her editorship, and in collaboration with Tsumagari, is the long epic (niƞmaa) of Siiƞuuni (Yamada ed. 2014), originally recorded by Ikegami from his main consultant, Ms. Satō Chiyo (c. 1910–1985), also known by her Uilta name Napka, a speaker of the southern dialect. In collaboration with E. A. Bibikova, Yamada has also published the Uilta translation of The Little Prince of Antoine de Saint-Exupéry (Yamada ed. 2016). In Russia, the first descriptions of Uilta (Orok) were authored by T. I. Petrova, who published both a book-length monograph and a concise sketch of Uilta (Orok) grammar (Petrova 1967, 1968). Another sketch grammar appeared under the authorship of A. I. Novikova and L. I. Sem (1997). As a contemporary specialist, L. V. Ozoliņa has published a grammar (2013), which remains the most comprehensive to date, as well as a bilingual dictionary of some 12,000 words (2001), the largest source available today on Uilta lexicon. Another dictionary, with some 5000 words from Uilta to Russian and 4000 words from Russian to Uilta, was published by Ozoliņa in collaboration with the native speaker I. Ya. Fedyaeva (2003). L. I. Missonova (2013) has extensively studied the lexicon of Uilta from a historical and ethnographical perspective. Other Russian scholars who have worked on Uilta include A. M. Pevnov, who has published important papers on the taxonomic position and linguistic contacts of Uilta (Pevnov 2009, 2014, 2016, 2017, 2020), and Valentin Gusev, who has contributed a paper on the multidimensional phenomenon of figurae etymologicae in Uilta (Gusev 2016). Outside of Russia and Japan, Uilta historical phonology has been studied by Alonso de la Fuente (2013), while syntax is the main object of Patryk Czerwinski (2019). Ethnonymic issues have been discussed by Janhunen (2014). The present sketch owes much to the pioneering study of the southern dialect of Uilta by Ikegami. Actually, some explanations and interpretations in this chapter are no more than a brief digest of his analysis. In particular, most sections concerning the phonology and morphology follow very closely his framework. Additional data, mostly on syntax, reflect the results of later fieldwork carried out by Tsumagari. Additional information on the northern dialect derives from Yamada’s fieldwork on Sakhalin. SEGMENTAL STRUCTURE Although closely related to Ulcha, Uilta has a vowel system that is somewhat reduced and shows a transition towards a type reminiscent of Ewenki. Qualitatively, the system comprises seven vowels: the three rounded back vowels u ɵ o, the two unrounded central vowels e a, and the two unrounded front vowels i ie (Table 16.1). All vowel qualities have both single or short (V) and double of long (VV) varieties.

Uilta 439 TABLE 16.1 UILTA VOWELS u

i

ɵ

e

o

a

ie

An important feature of Uilta is the presence of the distinction between the two nonhigh rounded vowels ɵ o, which represent the original pair *ö vs. *o. Phonetically, ɵ [o ~ ɵ] is pronounced higher and somewhat more centralized than o [ɔ]. The distinctive status of ɵ is obvious in materials deriving from the southern dialect (as first established by Ikegami), but in the modern northern dialect the opposition between o and ɵ is lost at least in suffixes in favour of an indistinctive o, reducing the number of vowel qualities in the system to six. Even so, some speakers recognize the opposition within roots, e.g. poro ‘grouse’ vs. pɵrɵ ‘thumb’. The vowel e is pronounced, as in most other Tungusic languages, with the centralized value [ə], which can occasionally also stand for original *ö especially after labial consonants, e.g. southern dialect (Ikegami) bɵyɵ ‘bear’ = northern dialect (Ozoliņa) boyo ~ buyu ~ beye. The vowel ie is pronounced with the unrounded front quality [e ~ ɛ] and is mainly attested as the long variety [e: ~ ɛ:], which represents the earlier diphthongoid sequence *ïa. However, in a few words, a short iĕ is also observed, mostly in the initial syllable, making this quality at least potentially distinctive also as a single vowel. In the consonants, Ulcha retains the Proto-Tungusic system of 18 segments, with the four nasals m n ñ ƞ, the four weak stops b d j g, the four strong stops p t c k, the two voiceless fricatives s x, the two glides w y, and the two liquids l r (Table 16.2). TABLE 16.2 UILTA CONSONANTS m

n

ñ

ƞ

b p

d

j

g

t

c

k

s w

x y

l r

The difference between the weak and strong stops is mainly based on voice (in the weak set). In intervocalic position, the weak velar stop g is fricativized to [ɣ]. The palatal stops j c are pronounced as affricates with a sibilant release, i.e. [dʒ tʃ] ~ [dʑ tɕ]. The fricative s [s] is also a sibilant and is normally palatalized to [ʃ ~ ɕ] before the vowels i ie, in the southern dialect also before u e. The nasal n is palatalized to [ɲ] before the vowels i ie, which means that its opposition with regard to ñ is neutralized in this position, allowing the sequences to be written ni nie, phonetically [ɲi] and [ɲe:], respectively. Because of the historical palatalization of *t *d to c j before the vowels i ie, the segments t d are normally not attested in this position. The lateral l and the vibrant r, which may also be pronounced as a flap, are often devoiced when followed by a voiceless consonant.

440 Toshiro Tsumagari and Yoshiko Yamada

PHONOTACTICS AND MORPHOPHONOLOGY The syllable structure in Uilta may be generally represented as (C)V(V)(C). Apart from double or long vowels, there are also diphthongoid sequences of two non-identical vowels, in which, however, the second segment is always one of the two high vowels i or u, with the additional exclusion of the combination i+u. This leaves the sequences ai ei oi ɵi ui and au eu ou ɵu, with the reservation that the sequences ɵi ɵu can occur only in those varieties that preserve the vowel ɵ as a distinctive segment. In the sequences ei eu the segment e is normally pronounced with the fronted value [e]. Any consonant may occur in syllable-initial or syllable-final position. All consonants are also attested word-initially, but the vibrant r is marginal and present only in the northern dialect in a few recent loanwords from Russian. Except for onomatopoetic words, only the consonants m n l can be present in word-final position, with the additional restriction that the final nasals m n are only attested in monosyllabic words. There is, however, a large group of stems ending lexically in the primary unstable nasal (*)/n, which is lost in prepausal position and before certain suffixes, but which appears as an “archiphonemic” homorganic nasal segment before other suffixes, e.g. nada ‘seven’ : ACC nadam-ba. Clusters (CC) can only occur in medial position. In clusters consisting of a velar followed by a labial the sequence is the opposite in the northern dialect, e.g. (southern vs. northern) dakpa vs. dapka ‘border’, sugbu vs. subgu ‘fish skin’, aƞma vs. amƞa ~ aƞma ‘mouth’. A syllable can be divided further into rhythmic entities or morae, which comprise a primary mora (= the initial consonant, if present, and the following vowel segment) and up to two secondary morae (= the second component of a double vowel or diphthongoid sequence and the syllable-final consonant, if present). Any word consists of at least two morae, and monosyllabic words always contain a vowel sequence, which means that they can only be of the types (C)VV or (C)VVC. The division of the syllable into morae is relevant for the location of the pitch accent, which, although a non-phonemic feature, is conspicuous in the pronunciation of a word and may function as a boundary signal. The highest pitch is placed on the penultimate mora. This may be identified as the accent peak of the word. If, however, the penultimate mora is a secondary mora, the pitch is moved backwards to the preceding primary mora. In words with two moras, the accent peak (in the following marked by the acute accent) falls on the first mora, e.g. nári ‘man’, móo ‘tree’. In words with three moras, the accent peak falls on the penultimate mora or, if the latter is a secondary mora, on the preceding primary mora, e.g. patála ‘girl’, iláa ‘three’, isál ‘eye’, ƞáala ‘hand’, áƞma ‘mouth’, gíĕul ‘oar’. In longer words, the higher pitch starts with the second mora and is maintained up to the accent peak, e.g. unígéri ‘star’, sinásúu ‘knife’, xuséńníe ‘man’. In such cases, the start of the accent peak can also be located on a secondary mora, e.g. xaúsáli ‘paper’, puútáa ‘bag’, including a sonorant consonant, e.g. xaḿdátta ‘leaf’. Secondary moras do not, however, maintain the high pitch after the accent peak, e.g. éekte ‘woman’, isáaptu ‘eyeglasses’, buwáata ‘island’. Apart from the pitch accent, the current speakers of the northern dialect tend to stress long vowels (VV) in various positions. This may be due to the influence of Russian phonetics. Also, when a syllable with a single o is located next to a syllable with a long vowel, there is a tendency to pronounce it as [ɑ] (as if following the rules of Russian akan’e), e.g. oksoo ‘reindeer sled’ > [ɑkˈsɔː], sietosici ‘they rustle’ > [ˈɕɛːtɑɕitɕi]. The vowels are harmonically arranged into two classes, corresponding to the original front vs. back series, subsequently rotated to an opposition based on tongue height (higher vs. lower), which may also have involved tongue root position (advanced vs.

Uilta 441

retracted) at some stage in the past. However, the vowels u i ie are harmonically neutral, leaving only the pairs a vs. e and o vs. ɵ as harmonically active (Table 16.3). TABLE 16.3 UILTA VOWEL HARMONY HIGH

e

ɵ

LOW

a

o

u

i

ie

Basically, each word can only comprise vowels of the higher (e ɵ) or the lower (a o) set in combination with the neutral vowels. As a trace of labial harmony, the vowels a e u cannot be followed by the short (single) vowels o ɵ, though there is no corresponding restriction with regard to the long vowels oo ɵɵ. Due to vowel harmony, most suffixes have either two (O = o ɵ) or four (A = a e o ɵ) harmonic variants depending on the preceding stem vocalism. In the northern dialect, however, the vowel ɵ has merged with o, resulting in a three-way alternation (A = a e o). Although Uilta morphology is based on agglutinative suffixation, there is a special morphophonological phenomenon that makes Uilta unique among the Tungusic languages and also less intelligible to speakers of even closely related idioms like Ulcha and Nanai. This phenomenon involves the loss of suffix-initial consonants, followed by the contraction of the syllables and the compensatory gemination of the preceding consonant. The stems that participate in the gemination process typically end lexically in a sequence of the type -VCV#, while the suffix-initial consonants that are lost in intervocalic position include *b *r *k, e.g. ute ‘door’ : ACC *ute-be > uttee, ƞene- ‘to go’ : IMP.2SG *ƞene-ru > ƞenneu. The same phenomenon can also be active in suffixes, e.g. buda ‘rice’ : AL-RX.PL *buda-ƞu-bari > buda-ƞƞoori. Diachronically, the processes of contraction and gemination in such cases involve a language-specific innovation that must have happened only after the separation of Uilta from Ulcha. However, the effects of this innovation may also be seen as a synchronic system of ordered processes, which obscure the surface manifestation of certain stems and suffixes and add a fusional character to Uilta morphology. Some speakers, especially of the modern northern dialect, show a tendency of avoiding the fusional forms by replacing them with agglutinative forms when possible. In most cases, the agglutinative forms are secondary innovations based on analogy. A tool for this analogy is offered by the unstable nasal (*)/n, which can be added also to stems that do not contain it etymologically, e.g. seduxi ‘berry’ : ACC *seduxi-be > seduxxee ~ seduxim-be (in the southern dialect also siduxxɵɵ ~ siduxum-be), peete ‘seal’ : ACC *peete-be > peettee ~ peetem-be (in the southern dialect peettee), gilee ‘Ghilyak’ : ACC *gileke-be > gilekkee ~ gileem-be (in the southern dialect gilekkee). WORD FORMATION The two basic morphological classes of inflectable words are nominals and verb(al)s. The class of nominals comprises regular nouns, spatials, adjectives, numerals, and pronouns. The derivation of nominals and verbs from nominal and verbal roots follows patterns similar to the other languages of the Nanaic group. The following is a selection of some of the derivational suffixes attested in Uilta (data adapted from Ozoliņa).

442 Toshiro Tsumagari and Yoshiko Yamada

(1) Denominal nouns: -dA/n [seasonal places of living], e.g. duwa/n ‘summer’ : duwada/n ‘summer pasture’; -ktA [frequent but synchronically non-productive suffix for collective nouns with individualizable units], e.g. pulme-kte ‘wild rose’, xosi-kta ‘fingernail/s’, siru-kte ‘ant/s’; -nnie (< *+ñarï ‘man, person’) [inhabitants], naa ‘land’ : naa-nnie-ni ‘local person’, namu ‘sea’ : PX3SG namu-nnie-ni ‘seaside inhabitant’; -ptu/n [coverings], e.g. kotoo/n ‘finger’ : kotoo-ptu/n ‘glove/s’; -skA (< *-ksA) [hide/ skin of animals], e.g. peete ‘seal’ : peete-ske ‘seal skin’; -tA/n (< *-cA/n) [young of animals], e.g. siro/n ‘wild reindeer’ : siro-ta/n ‘calf of wild reindeer’. Secondary spatials are formed by the suffix -jie ‘side’, e.g. namu ‘sea’ : namu-jie ‘seaside’; suu- ‘southern Sakhalin’ < ‘south’ (< *suu/n ‘sun’) : suu-jie id., cf. also PX3SG suunnie-ni ‘inhabitant of southern Sakhalin’. Adjectival nominals are derived from nouns by the suffixes -li [possession of a characteristic] kusu/n ‘power’ : kusu-li ‘strong’; -lu [proprietives], e.g. asi ‘wife’ : asi-lu ‘having a wife’ > ‘married (man)’; -mA [material], e.g. girapsa ‘bone’ : girapsa-ma ‘made of bone’; -ƞAci [similatives], e.g. gasa ‘bird’ : gasa-ƞaci ‘bird-like’; -unA [possession of a characteristic], e.g. nirukte ‘hair’ : nirukte-une ‘hairy’. Adjectival nominals derived from adjectival roots include the forms in -kA ~ -ukA [moderatives], e.g. ƞonimi ‘long’ : ƞonimi-ka ‘rather long’, nekte ‘low’ : nekte-uke ‘quite low’; -li ~ -uli [observable characteristic], e.g. maƞga ‘hard, strong’ : maƞga-uli id.; -pci [temporal association], e.g. sagji ‘old’ : sagji-pci ‘of old age’. (2) Deverbal nouns: -kku [instrument, place, process, or result of action], e.g. karga- ‘to watch’ : karga-kku ‘binoculars’, bi- ‘to be, to live’ : bi-kku ‘place of living’, uile- ‘to work’ : uile-kku ‘work’, nanda- ‘to borrow’ : nanda-kku ‘debtor’; -mji [actor nouns], e.g. (southern) ɵlɵ- ~ (northern) ule- ‘to cook’ : (southern) ɵlɵ-mji ~ (northern) ulemji ‘cook’, beleci- ‘to help’ : beleci-mji ‘helper’; -ƞku ~ -ƞki [instrument or result of action], e.g. nama- ‘to saddle’ : nama-ƞku ~ nama-ƞki ‘saddle girth’; -nAu [place of action] xumu- ‘to bury’ : xumu-neu- ‘place to bury’; -ptu [object or result of action], e.g. ciree- ‘to press’ : ciree-ptu ‘pillow’, ƞokki- ‘to smell’ : ƞokki-ptu ‘smell’. Adjectival nominals are derived from verbs by the suffixes -gbA [characteristic feature], e.g. daxuri- ‘to undertake’ : daxuri-gda ‘obedient’; -ktu [resulting state] e.g. dausula- ‘to salt’ : dausula-ktu ‘salted’; -psie [characteristic feature], e.g. agda- ‘to rejoice’ : agda-psie ‘joyful’. (3) Denominal verbs: -lA- [multifunctional default verbalizer], e.g. moo ‘tree, wood’ : moo-lo- ‘to fetch firewood’, asi ‘wife’ : asi-la- ‘to marry (of men)’; -dA- ~ -dAi[multifunctional secondary verbalizer], e.g. kalta ‘split’ : kalta-da- ‘to split’, duwa ‘summer’ : duwa-da- ‘to spend the summer’, luca ‘Russian’ : luca-dai- ‘to speak (in) Russian’; -mA- [captatives], e.g. tuksa ‘hare’ : tuksa-ma- ‘to hunt for hares’; -mAn[instrumentals], e.g. ikte ‘tooth’ : ikte-men- ‘to bite’; -su- [multifunctional], e.g. (southern) dɵɵ ~ (northern) duu- ‘two’ : (southern) dɵɵ-su- ~ (northern) duu-su- ‘to double’. Suffixes deriving verbs from adjectival nominals include -lA- [factitives], e.g. gugda ‘high’ : gugda-la- ‘to make higher’; -dA- [essives], e.g. kuta ‘greedy’ : kuta-da- ‘to be greedy’. (4) Deverbal verbs: As in other Tungusic languages, deverbal verbs in Uilta express several grammaticalized meanings of voice, mood, aspect, and Aktionsart. •

Voice is expressed by the suffixes -(B)OOn- ~ -(B)un- [causatives and passives], e.g. dep- ‘to eat’ : CAUS de-pɵɵn- ‘to feed’, ƞene- ‘to go’ : CAUS ƞenne-un‘to send’; -un- [causatives from andatives], e.g. gele- ‘to search’ : AND-​CAUS

Uilta 443

• •

•

gele-nde-un- ‘to send (somebody) searching’; -ptu- [medials], e.g. dolji- ‘to hear’ : MED dolji-ptu- ‘to be heard’; -mAci [reciprocals], e.g. ataali- ‘to meet’ : RECIPR ataali-maci- ‘to meet each other’, caici- ‘to drink tea’ : RECIPR caici-maci- ‘to drink tea together’. Modal distinctions are expressed by the suffixes -KitA- [intentives], e.g. ite- ‘to see, to watch’ : INTENT ite-gite- ‘to intend to watch’, -musi- [desideratives], e.g. umi- ‘to drink’ : DESID umi-musi- ‘to want to drink’. Distinctions of aspect and Aktionsart are expressed by the suffixes -ci- ~ -cci[duratives], e.g. ite- ‘to see’ : DUR ite-ci- ‘to keep seeing’; -du- [reversives], e.g. tuksa- ‘to run’ : REV tuksa-du- ‘to run again’; -lu- [inchoatives], e.g. gitu- ‘to walk’ : INCH gitu-lu- ‘to start walking’; -nA- [iteratives], e.g. gele- ‘to request’ : ITER gele-ne- ‘to request all the time’, garpa- ‘to shoot’ : ITER garpa-na- ‘to shoot many arrows’; -tA- [completives], e.g. gaa- ‘to take’ : COMPL gaa-ta- id. Andatives are formed by the suffixes -ndA- ~ -ni-, e.g. waa- ‘to hunt’ : AND waa-nda- ‘to go hunting’, gaa- ‘to take’ : AND gaa-ni- ‘to go to take’.

NUMBER AND CASE Nominal plurality is marked by the suffixes -l or -sAl, which are mainly attached to stems meaning humans or animals, but also, though more rarely, inanimate things, e.g. nari ‘man, person’ : PL nari-l ~ nari-sal ‘men, people’, ulaa ‘reindeer’ : PL ulaa-l, buwaata ‘island’ : PL buwaata-l. The plural forms of a few nouns are based on irregular stem variants, e.g. putte ‘child’ : PL puri-l, ƞinda ‘dog’ : PL ƞinji-l. Several kinship terms also have exceptional plural forms, e.g. ami/n ‘father’ : PL ami-l ~ ami-nil ~ ami-til ‘fathers’, eni/n ‘mother’ : PL eni-l ~ eni-nil ~ eni-til ‘mothers’. In general, the use of plural forms implies emphasis, and the unmarked form can also be used in reference to plural entities. This is, in particular, the case when the noun is preceded by a numeral or other quantifier, e.g. ilaa nari ‘three men’ (three main), bara putte ‘many children’ (many child). Apart from the unmarked nominative, the Uilta case paradigm has nine suffixally marked cases: accusative, designative, dative, ablative, locative, prolative, directive (allative), instrumental, and comitative (Table 16.4). The designative and prolative occur only in combination with possessive marking. TABLE 16.4 UILTA CASE MARKERS  

CV

VV

C

/n

ACC

-A

-wA

-bA

m-bA

DES

-ddOO-

-dOO-

n-dOO-

DAT

-du

n-du

ABL

-duu

n-duu

LOC

-lA

-du-lA

n-du-lA

PROL

-kkie-

-kie-

k-kie-

DIR

-tAi

INSTR

-ji

COM

-ndOO

t-tAi n-ji -dOO

n-dOO

Stem types: CV = single-vowel stems, VV = double-vowel stems, C = consonant stems, /n = nasal stems. Segmental alternations: A = a e o ɵ, O = o ɵ.

444 Toshiro Tsumagari and Yoshiko Yamada

Some case markers have different allomorphs depending on whether the nominal stem ends in a single vowel (CV#), a double vowel (VV#), the unstable nasal (/n), or another consonant (l g). The nasal is absent in the unmarked nominative of polysyllabic stems, while in the other cases it is represented as a nasal homorganic with a following weak stop. In the prolative and directive cases it adapts to the following strong stop. Sample paradigms: ute ‘door’ (stem ending in CV#) : ACC uttee : DES ute-dɵɵ- : DAT ute-du : ABL ute-duu : LOC ute-le : PROL ute-kkie- : DIR ute-tei : INSTR ute-ji : COM ute-ndɵɵ. bie [place in a dwelling on both sides of the entrance] (stem ending in VV#) : ACC bie-wa : DES bie-ddoo- : DAT bie-du : ABL bie-duu : LOC bie-la : PROL bie-kkie- : DIR bie-tai : INSTR bie-ji : COM bie-ndɵɵ. tuƞe : tuƞen- ‘chest’ (stem ending in the nasal /n) : ACC tuƞem-be : DAT tuƞen-du : ABL tuƞen-duu : LOC tuƞen-dule : PROL tuƞek-kie- : DIR tuƞet-tei : INSTR tuƞen-ji : COM tuƞen-dɵɵ. In the accusative, stems involving a single medial consonant before a short final vowel (VCV#) participate in the processes of contraction and compensatory gemination, as in ute : ACC uttee (< *ute-be). Among the stems ending in a double vowel, exceptions are formed by dɵɵ ~ duu ‘two’ and ulaa ‘domestic reindeer’, which have the accusative forms dɵɵ-be ~ duu-be and ulaa-ba, respectively (instead of the expected forms in -wA). This is historically connected with the fact that these stems originally ended in the consonant *-r, a well-known fact for the numeral ‘two’ (*jöör), but obviously also true of the item for ‘domestic reindeer’ (*ulagar, as first observed by Pevnov). Outside of the actual case paradigm, Uilta preserves also the Common Tungusic suffix -ƞi ~ -ni, which marks the possessive form of nominals, used in predicative position and as an independent head word with a regular nominal declension, e.g. ƞinda ‘dog’ : GEN ƞinda-ƞi ~ ƞinda-ni ‘of the dog, dog’s (own)’. NUMERALS The Uilta cardinal numerals of the first decade are: 1 gieda ~ gida, 2 (southern) dɵɵ ~ (northern) duu, 3 ilaa/n, 4 jiin, 5 tunda, 6 nuƞu/n, 7 nada/n, 8 (southern) jakpu/n ~ (northern) japku/n, 9 xuyu/n, 10 joon. With the exception of the items for 1 gieda, 2 dɵɵ ~ duu, and 5 tunda, these are all nasal stems, but in absolute position the final nasal is preserved only in the monosyllabic items for 4 jiin and 10 joon, while in the others it appears in the oblique forms. Most of the numerals are regular cognates of the corresponding Ulcha items, but the item for 1 gieda is a Uilta innovation based on Common Tungusic *gïa ‘other’, cf. Nanai gïagda = Ulcha giegda/n ‘single, unmarried’. The original numeral *emön ‘one’ is preserved only marginally in the shape (northern) umu/n, as well as in derivatives like (southern) ɵmɵɵkɵ ‘alone’, ɵmɵttɵ ‘equal’. The numeral 2 dɵɵ ~ duu differs from Ulcha juel/i, in that its accusative form preserves an indirect trace of the original final *r, as still present in Nanai juer. The form of the numeral 4 jiin < *diin is also specific to Uilta, but it apparently nevertheless represents a direct reflex of Nanaic *duin < *duyin, as in Ulcha and Nanai, rather than a secondary borrowing from Ewenic *digin. Of the higher numerals, Uilta preserves the Nanaic items for 20 xori/n, 100 (gieda) taƞgu, 1000 (gieda) miƞga/n, and 10,000 (gieda) tume/n. A peculiar item for higher numerals is daxu ‘hundred thousand’ ~ ‘million’. The other decades are formed by the

Uilta 445

suffix -dOO (< *-jU < *-gU), i.e. 30 ilaan-doo, 40 jiin-dɵɵ, 50 tunda-doo, 60 nuƞun-dɵɵ, 70 nadan-doo, 80 jakpun-doo ~ japkun-doo, 90 xuyun-dɵɵ. For the intermediate numerals additive combinations are used, e.g. 11 joon gieda, 12 joon dɵɵ ~ joon duu, etc. Derivational forms of numerals include the multiplicatives in -ltA (‘times’): 3 ilaa-lta, 4 jii-lte, 5 tunda-lta, 6 nuƞu-lte, 7 nada-lta, 8 jakpu-lta ~ japku-lta, 9 xuyu-lte, 10 joo-lto. The first two numerals have the special forms 1 gieda-ra ‘once’ and 2 dɵɵ-rɵ ~ duu-re ‘twice’, though the regular forms gieda-lta, especially in 11 joon gieda-lta ‘eleven times’, and duu-lte ‘twice’ are also attested. The forms in -ltA can also be used for counting days, e.g. nada-lta ‘seven days’ > ‘a week’, and occasionally they can serve as ordinals, as well. Actual ordinals, e.g. (northern) gied-inniĕ ~ gieda-lliĕni ‘first’, (southern) dɵɵ-yyie-ni ~ (northern) duu-yeni ~ duu-lliĕni ‘second’, are today used mainly in reference to months after pregnancy or years after birth. Distributive numerals are formed by the suffix -tAl : INSTR -tAl-ji, e.g. gieda-tal(-ji) ‘one by one’, tunda-tal(-ji) ‘five each’. PRONOUNS The pronominal system of Uilta is in most respects similar to that of Ulcha. The personal pronouns (Table 16.5) have the same basic (nominative) forms and oblique stems as in Ulcha. However, the accusative forms have the extended forms in -bie > -biĕ in the singular (as in Nanai) and (southern) -bie-pe ~ (northern) -bɵɵ-pe in the plural (unlike Ulcha and Nanai), i.e., 1SG mim-biĕ : 2SG sim-biĕ : 1PL (southern) mum-bie-pe ~ (northern) mum-bɵɵ-pe : 2PL (southern) sum-bie-pe ~ (northern) sum-bɵɵ-pe. The function of adnominal genitives is filled by the extended oblique stems 1SG mini : 2SG sini : 1PL munu : 2PL sunu, while in predicative position and as independent possessive nouns the possessive forms 1SG mini-ƞi : 2SG sini-ƞi : 1PL munu-ƞi : 2PL sunu-ƞi are used. The adnominal genitives are also used as substitutes for a nominative subject in subordinate clauses. Like the other Nanaic languages, Uilta has no distinction between the exclusive and inclusive functions in the first person plural. There is also no actual pronoun of the third person, as the reflex of the Common Tungusic third person pronoun, in Uilta noo- : PX3SG noo-ni ‘s/he’: PX3PL noo-ci ‘they’ (with the case markers placed before the possessive suffixes), functions as an obviative pronoun and refers to the third person only in contrast with another third person actor. Usually, the third person is referred to by a demonstrative pronoun with or without a suitable noun. TABLE 16.5 UILTA PERSONAL PRONOUNS  

 

NOM

OBL

ACC

GEN

POSS

SG

1

bii

min-

mim-biĕ

mini

mini-ƞi

2

sii

sin-

sim-biĕ

sini

sini-ƞi

1

buu

mun-

mum-bie-pe

munu

munu-ƞi

2

suu

sun-

sum-bie-pe

sunu

sunu-ƞi

PL

The reflexive pronoun has the stem mee- : meen- ‘(one)self’, which in the oblique forms always takes the reflexive markers after the case suffix, e.g. LOC-RX.SG meendul-i : LOC-RX.PL meen-dul-eri. In object position, however, the plain forms SG mee-pi : PL mee-peri, unmarked for case, are used. There is also the adnominal form meene ~ meen ‘(one’s) own’, after which the head noun takes reflexive marking. The same form

446 Toshiro Tsumagari and Yoshiko Yamada

meene is used in the emphatic (adverbial) function ‘(by) oneself’, e.g. meene dep-ci-wi ‘I eat (it) myself’ (REFL eat-PRS-1SG). The demonstrative pronouns are eri (~ eye) : OBL eri- ~ ye- ~ yee ‘this’ (proximal) vs. tari : OBL tari- ~ ca- ~ caa ‘that’ (distal), with the plural forms eri-l ~ eri-sel ~ ye-sel ‘these’ vs. tari-l ~ tari-sal ~ caa-l ~ ca-sal ‘those’. The singular forms eri and tari may be used both independently and attributively; in the latter case they take the forms yee vs. caa before nouns in oblique cases, e.g. tari nari ‘that man’ > ‘he’ : ACC caa narri-e ‘him’. The plural forms are employed only independently with the meaning ‘these people’ vs. ‘those people’ > ‘they’. For more distant referents, the emphatic forms too-tori and (even more distant) taa-tari are used. The basic interrogative pronouns are ƞui : OBL ƞui- ‘who?’ and xai : OBL xai- ‘what?’ : DAT xai-du ‘where?’. Related forms include xaawu : ACC xaakkoo ‘which?’, xasu : ACC xasum-ba ‘how many? how much?’, xamaciga ~ xamacie ‘what kind of?’, all of which can be used adnominally, e.g. xaawu nari=ga ‘which person?’ (which man=CORROG). The root xai- can also be used as a verb in the meaning ‘to do what?’, yielding the lexicalized form CV.CONN xai-mi ‘why?’. Other lexicalized question words, functioning as adverbs, include xaali ‘when?’ and xooni ‘how?’. PERSON MARKING The distinction between possessive suffixes (PX) and verbal predicative endings (VX) is less prominent in Uilta than in the other Nanaic languages. This is mainly because the role of the aorist paradigm is reduced and somewhat obscured. Even so, there are a few differences between the possessive and predicative sets (Table 16.6), as a few aorist-related forms require person markers different from the possessive set. There is also a set of reflexive suffixes (RX) for singular and plural reference. The possessive and reflexive suffixes have allomorphic variation conditioned by the same stem types as with the case suffixes, i.e. single vowel stems (CV#), double vowel stems (VV#), consonant stems (l g), and nasal stems (/n). The first person singular possessive suffix -bi ~ -wi is identical in shape with the singular reflexive suffix. In sentences, however, they are distinguishable because the reflexive form does not stand for the subject of a sentence, and the possessive suffix is extended to -wwie in the oblique forms, e.g. PX1SG ute-bi (subject form) : ACC-PX1SG. EXT utte-​e-wwie : LOC-PX1SG.EXT ute-le-​wwie. The extended form has a parallel in Nanai, but not in Ulcha. A similar extension is observed in the corresponding plural suffix, e.g. PX1PL ute-pu : ACC-PX1PL.EXT utte-e-ppɵɵ : LOC-PX1PL.EXT ute-le-ppɵɵ, TABLE 16.6 UILTA PERSON MARKERS      SG

PX   

VX

CV

VV

C

/n

AOR

1

-bi

-wi

-bi

m-bi

-mi

2

-si

-ci

c-ci

3

-ni

RX

-bi

-ni -wi

-bi

m-bi

-Ø

Uilta 447 PL

1

-pu

2

-su

3

-ci

RX

-bAri

p-pu -cu

c-cu c-ci

-wAri

-bAri

(-l)

m-bAri

Stem types: CV = single vowel stems, VV = double vowel stems, C = consonant stems, /n = nasal stems. Segmental alternations: A = a e o ɵ.

but not in the other persons, cf. e.g. PX2SG ute-si : ACC-PX2SG utte-e-si : LOC-PX2SG ute-le-si. Another difference between the possessive and reflexive declensions is that the reflexive markers take the forms SG -i : PL -A.ri before case suffixes, at the same time as the consonant of the case suffix undergoes gemination, except after nasal stems and in the plural locative. In this connection, the singular designative, dative, and instrumental suffixes are merged to the form—(j)j-, while in the plural the designative and dative suffixes are merged to the form -(d)d.A-. Also, the ablative and directive suffixes take the long (etymologically full) forms -dukk.A- and -tAkk.ie-, respectively. It has to be noted, however, that in the modern northern dialect many of the morphophonological complications have been simplified in the direction of a more agglutinative structure. Sample paradigms: ute ‘door’ (stem ending in CV#) : RX ute-bi : DES & DAT & INSTR ute-jj-i : ABL : ute-dukk-i : LOC ute-ll-i : PROL ute-kk-i : DIR ute-tekk-i : RX.PL utte-eri : DES & DAT ute-ddɵ-ɵri : ABL ute-dukkɵ-ɵri : LOC ute-le-ri : PROL ute-kki-eri : DIR ute-​ tekki-​eri : INSTR ute-jji-eri. bie [place in a dwelling on both sides of the entrance] (stem ending in VV#) : RX bie-wi : DES & DAT & INSTR bie-jj-i : ABL : bie-dukk-i : LOC bie-ll-i : PROL bie-kk-i : DIR bie-takk-i : RX.PL bie-wari : DES & DAT bie-ddo-ori : ABL bie-dukko-ori : LOC bie-la-ri : PROL bie-kki-eri : DIR bie-takkie-ri : INSTR bie-jji-eri. tuƞe : tuƞen- ‘chest’ (stem ending in the nasal /n) : RX tuƞem-bi : DES & DAT & INSTR tuƞen-j-i : ABL : tuƞen-dukk-i : LOC tuƞen-du-ll-i : PROL tuƞek-k-i : DIR tuƞe.ttekk-i : RX.PL tuƞem-beri : DES & DAT tuƞen-dɵ-ɵri : ABL tuƞen-dukkɵ-ɵri : LOC tuƞen-​du-le-ri : PROL tuƞek-kkie-ri: DIR tuƞet-tekkie-ri : INSTR tuƞen-ji-eri. As an exception conditioned by its diachrony, the stem ulaa ‘domestic reindeer’ takes the reflexive markers with initial b, i.e. RX ulaa-bi : RX.PL ulaa-bari. This exception is preserved also in the modern northern dialect. The possessive and reflexive suffixes can be combined with the marker of alienability, which is -ƞu-, e.g. ulise ‘meat, flesh’ : PX1SG ulise-bi ‘my flesh’ vs. AL-PX1SG uliseƞu-bi ‘my meat’, jili ‘head’ : PX1SG jili-bi ‘my head’ : AL-PX1SG jili-ƞu-bi ‘the head (of a game animal) that belongs to me’. The forms containing the alienability suffix are inflected like single-vowel stems (CV#), e.g. buda ‘rice’ : AL-RX.PL buda-ƞƞo-ori, patala ‘girl’ : AL-INSTR-PX1SG patala-ƞu-ji-wwie. The presence or absence of the alienability suffix is determined by culturally conditioned conceptual factors, cf. e.g. ƞinda ‘dog’ : PX1SG ƞinda-bi (unalienable domestic animal) vs. bɵyɵ/n ‘bear’ : AL-PX1SG bɵyɵ-ƞu-bi ‘my bear’ (alienable wild animal), cikte ‘louse’ : PX1SG cikte-bi ‘my louse’

448 Toshiro Tsumagari and Yoshiko Yamada

vs. sura ‘flea’ : AL-PX1SG sura-ƞu-bi ‘my flea’, kitaa/n ‘needle’ : PX1SG kitaam-bi vs. kupe/n ‘thread’ : AL-PX1SG kupe-ƞu-bi ‘my thread’. VERBAL MORPHOLOGY Uilta preserves the distinction between vowel-stem and consonant-stem verbs. The vowel stems may be further divided into single-vowel stems (CVC-) and double-vowel stems (VV-), while the consonant stems (C-) normally end in the nasal n, but also in the segments l g p. The distinction between the stem types is visible in how the original aorist marker *-rA and related forms are represented. Since the simple finite aorist paradigm is not used, the plain aorist marker is synchronically best observed in the connegative function, as used with the negation verb e-, e.g. (single-vowel stem) ƞene‘to go’ : CONNEG ƞenne-e, (double-vowel stem, southern) bɵɵ- ‘to give’ ~ (northern) buu- : CONNEG (southern) bɵɵ-re ~ (northern) buu-re, (nasal stem) un- ‘to say’ : CONNEG un-de, (other consonant stem) xɵɵl- ‘to spill’ : CONNEG xɵɵl-dɵ. The verb dep- ‘to eat’ : CONNEG dep-te has also the vowel-stem alternant deptu- : CONNEG deptu(-re). In addition, there are verbs that take the element -si in the aorist stem and, hence, in the connegative form, e.g. andu- ‘to make’ : CONNEG andu-si. The negation verb e- also has the aorist stem e-si-. Finally, there is a small group of exceptional verbs with cognates in the other Nanaic languages: ga- ‘to take’ : CONNEG ga-da, o- ‘to become’ : CONNEG o-do ~ o-si, bul- ‘to die’ : CONNEG bu-de, bi- ‘to be’ : CONNEG bi-e [be:]. Verbal forms in Uilta can be divided into the standard categories of participles, indicative finite forms, imperatives, and converbs. Of these, the participles are multifunctional forms that can be used as adnominal modifiers before a head noun, as independent head nouns with or without case marking, and as finite predicates with person markers. The other verbal categories are monofunctional, being restricted to either finite (imperatives, indicative finite forms) or non-finite (converbs) functions. The role of the indicative finite forms is rather reduced, since the function of the basic unmarked tense forms has been taken over by the finitely used participles, as also in Ulcha and Nanai. At the same time, the original finite forms have developed additional evidential functions. PARTICIPLES The two participles in active use in Uilta are the imperfective participle in .i-e ~ -ri ~ -ji (< *-rII) ~ -si (< *-sii) depending on the stem type, and the perfective participle in -xA/n (for vowel stems) or -ci/n (for the other stem types), e.g. ƞene- ‘to go’ : PTCP.IMPRF ƞenn.i-e : PTCP.PRF ƞene-xe/n, baa- ‘to find’ : PTCP.IMPRF baa-ri : PTCP.PRF baa-xa/n, andu‘to make’ : PTCP.IMPRF andu-si : PTCP.PRF andu-ci/n, un- ‘to say’ : PTCP.IMPRF un-ji : PTCP.PRF uc-ci/n (with c-c < *n-c), buyal- ‘to break’ : PTCP.IMPRF buyal-ji : PTCP. PRF buyal-ci/n. The verbs ga- ‘to take’, o- ‘to become’, bul- ‘to die’, bi- ‘to be’, and the negation verb e- have the participle forms ga-ji : ga-cci, o-si : o-cci, bu-ji : bu-cci, bi-i [bi:] : bi-cci, and e-si : e-cci, respectively. In the role of finite predicates, the participles function as tense markers for the present (PRS) and past (PST) tenses, respectively. Person is marked by the suffixes of the possessive type, except that the first person singular suffix -bi yields in combination with the final nasal /n of the past tense markers the complex suffix -mbi. In these cases, the location of the morpheme boundary is synchronically somewhat ambiguous, e.g.

Uilta 449

ƞene- ‘to go’ : PST.1SG ƞene-xem-bi ~ ƞene-xe-mbi. It may be noted that the possessive forms of all other persons of the perfective participle (in all functions) lack the final nasal. In the northern dialect, a secondary future participle in -l-i- (< *-lU-rII) has developed from the imperfective participle of the inchoative forms in -lu-. This future participle can also be used finitely, yielding a secondary finite future tense (FUT), e.g. bii sit-tei tamali-wi ‘I will reward you’ (1SG 2SG-DIR reward-FUT-1SG). FINITE INDICATIVE FORMS The finite indicative set involves three temporally differentiated forms used only in the finite function and referring to present, past, or future action, respectively, with the additional feature that the present and past forms have a connotation of evidentiality (witnessed by the speaker). This is similar to Nanai and Ulcha, where the original finite forms have likewise been marginalized to an evidential function, while the unmarked reference to tense is expressed by the finitely used participles. In Uilta, the situation is complicated by the circumstance that the present and past forms are morphologically complex, in that they involve the additional element +kkA ~ +gA ~ +wA ~ +A, which follows the person markers. The finite form with present tense reference is marked by -A- ~ -rA- ~ -dA- ~ -si-, depending on the stem type. This is the Common Tungusic aorist, which, consequently, survives in Uilta in finite use only in the complex present-tense evidential form, which may be identified as the “extended aorist” (AOR-EXT). The finite form with past tense reference is marked by -tA-, which apparently is a reflex of the Common Tungusic past-perfective marker *-cAA with marginally attested cognates also in Nanai and Ulcha. In Uilta, the resulting complex past-tense evidential form may be identified as the finite “extended past” (PST-EXT) form. The finite future tense represents the synchronically still transparent combination of the imperfective participle with the general verbalizer -lA-, as also attested in Ulcha. Like the extended aorist and past forms, the complex future form takes person markers of the aorist type. The actual difference with regard to the possessive person markers is synchronically only that the first person singular has the marker -mi (and not -bi ~ -wi or -mbi), while the third person forms have no explicit person markers, with the plural indicated by the nominal plural marker -l. Sample paradigms: dep- ‘to eat’ (consonant stem): AOR-1SG-EXT dep-te-mi-ge : AOR-2SG-EXT dep-tesi-ge : AOR-EXT dep-te-kke : AOR-1PL-EXT dep-te-pu-we : AOR-2PL-EXT depte-su-we : AOR-EXT-PL dep-te-kke-l. ƞene- ‘to go’ (single-vowel stem): PST.EVID-1SG-EXT ƞene-te-mi-ge : PST.EVID2SG-​EXT ƞene-te-​si-ge : PST.EVID-EXT ƞene-te-e : PST.EVID-1PL-EXT ƞene-tepu-we : PST.EVID-2PL-EXT ƞene-te-​su-we : PST.EVID-​EXT-PL ƞene-te-e-l. tee- ‘to sit down’ (a double-vowel stem with an optional aorist stem in -si-): FUT-1SG tee-​sile-mi : FUT-2SG tee-sile-si : FUT.3SG tee-sil(l)e(-e) : FUT-1PL tee-sile-pu : FUT-​2PL tee-sile-su : FUT.3PL tee-sil(l)e(-e)-l. Another finite form, present in the northern dialect, is the habitive in -bukki ~ -wukki, e.g. tacci nari agda-mi bi-wukki ‘a diligent man always lives happily’ (diligent man be.happy-CV.CONN be-HAB). This form takes the possessive set of person markers in the first and second person, but has only number marking in the third person.

450 Toshiro Tsumagari and Yoshiko Yamada

IMPERATIVES The imperatives involve a closed set of finitely used modal forms. The basic second person singular imperative is marked by the aorist-based suffix -u ~ -ru ~ -su ~ -du, depending on the stem type, e.g. ƞene- ‘to go’ : IMP.2SG ƞenne-u, baa- ‘to find’: IMP.2SG baa-ru, andu- ‘to make’ : IMP.2SG andu-su, buyal- ‘to break’ : IMP.2SG buyal-du. The verbs ga- ‘to take’, o- ‘to become’, bul- ‘to die’, and bi- ‘to be’ have the imperative forms ga-du, o-su, bu-du, bu-u [bu:], while the negation verb yields the idiosyncratic prohibitive form e-jjie (< *e-ji-). For explicit plural reference, the person marker 2PL -su can be added. A polite request is marked by -yA : PL -yA-l-tu on all verb stems, e.g. ƞene-ye ‘please go!’ For the other persons no systematic paradigm of imperative forms exists. In the first person singular, the imperfective participle (= present tense) with the particle =tA is used, e.g. ƞenn.i-e=te ‘I will go, let me go!’, while the first person plural is marked by the complex ending -ƞA-p=tA (< *-ƞA-pu=tA), e.g. ƞene-ƞe-p=te ‘let us go!’. An invitational meaning can also be expressed by the second person plural form of the imperfective participle, e.g. uiltadai-ri-su ‘let us speak Uilta!’ (the title of the first Uilta primer). In the northern dialect, a third person imperative (permissive) reference is expressed by the combined use of the particle bijini ‘maybe’ (from the verb bi- ‘to be’) and a verb form in PTCP.IMPRF-loo, e.g. bijini nooni sinj.i-e/loo ‘let him/her come!’ (maybe 3SG come-PERM). CONVERBS The Uilta system of converbs shows some differences as compared with the other Nanaic languages (Table 16.7). Most importantly, it contains the contemporal converb in -ƞAssie(‘when in the past’), a reflex of Proto-Tungusic *-ƞA-si, not attested in Ulcha or Nanai. The other converbs are the connective converb in -mi (‘at the same time as’), the consecutive converb in -pie (‘if, when, after’), the contextual converb in -GAcci (‘when, after’), the terminative converb in -dAlAA (‘until’), the conditional converb in -KutA- (‘if, after’), and the aorist converb in -RA(y)i- (‘if, when, after’).

TABLE 16.7 UILTA CONVERB MARKERS  

 

PL

CONN

-mi

-mAri

PX

EXT

SS +

DS

CSEC

-pie

-pissAA

+

CTXT

-GAcci

-GAccieri

+

TERM

-dAlAA

CTEMP

-ƞAssie-

+

1P

COND

-KutA-

+

1–3P

+

AOR

-RA-(y)i-

+

1P

+

+

+

+

+

Functional types: SS = same-subject (conjunct), DS = different-subject (disjunct). Segmental alternations: A = a e o ɵ, G = g k w, K = k w Ø, RA = aorist marker depending on the stem type. Extended person markers: 1P = first person, 1–3P all persons.

Uilta 451

The connective, consecutive, and contextual converbs are always used in same-subject (conjunct) constructions. Their basic forms contain etymologically the reflexive suffix -i, and they have the corresponding plural forms with the plural reflexive suffix -Ari, except that the plural form of the consecutive converb has the idiosyncratic suffix -pissAA. The terminative converb is ambivalent (conjunct/disjunct), while the other converbs, which take person marking, are used in different-subject (disjunct) constructions. In these cases, the person markers are based on the possessive set, but they appear in extended forms with a long vowel element in the end, similar to the first person singular and plural extended forms as used in the oblique cases of the nominal possessive declension. Moreover, the contemporal and aorist converbs take the extended markers only in the first person: 1SG -wwie : 1PL -ppOO, as attested in the nominal declension, while the conditional converb takes extended markers also in the other persons: 2SG -ssie : 3SG -nnie : 2PL -ssOO : 3PL -ccie. The contemporal converb can also be used conjunctively without person marking. The markers of the contextual, conditional, and aorist converbs show considerable variation of various types. In the aorist converb, this variation is connected with the morphological and morphophonological rules of aorist formation, while in the contextual and conditional converbs only morphophonology is involved, e.g. ƞene- ‘to go’ : CV.AOR ƞenne-eyi- : CV.CTXT ƞene-gecci : CV.COND ƞene-ute-, baa- ‘to find’ : CV.AOR baarai- : CV.CTXT baa-gacci : CV.COND baa-wuta-, andu- ‘to make’ : CV.AOR andu-sii- : CV.CTXT andu-wacci : CV.COND andu-uta-, buyal- ‘to break’ : CV.CTXT buyal-kacci : CV.AOR buyal-dai- : CV.COND buyal-kuta-. The verbs ga- ‘to take’, o- ‘to become’, bul‘to die’, bi- ‘to be’, and the negation verb e- have the forms CV.AOR ga-dai-, o-sii-, bu-dei-, bie-, e-sii-, CV.CTXT ga-gacci, o-wocci, bul-kecci, bi-gecci, e-gecci, and CV.COND ga-uta-, o-uta-, bul-kute-, bi-wute-, e-ute-, respectively. PHRASE STRUCTURE Uilta has a headfinal phrase structure, with both adnominal modifiers (adjectival nominals, demonstrative pronouns) and adverbal arguments (object, adverbials) preceding their nominal or verbal head words. In the nominal phrase there is normally no agreement for number, case, or person between an adjectival modifier and the nominal head word, and the morphological marking is taken only by the head noun, e.g. aya ulaa ‘good reindeer’ : aya ulaa-l (good reindeer-PL) : aya ulaa-ba (good reindeer-ACC) : aya ulaa-bi (good reindeer-PX1SG). The same is true of adnominally used demonstrative pronouns, except that they have special oblique stems, e.g. eri nari ‘this man’ : yee narri-e (this.OBL man-ACC), tari nari-l ‘those men’ (that man-PL) : caa nari-l-ba (that.OBL man-PL-ACC). Occasional agreement in number has, however, been reported from the northern dialect (Petrova), e.g. daayi-l duku-l ‘large houses’ (large-PL house-PL) = southern daayi duxu-l (large house-PL). An adjectival nominal can also be used as a noun, in which case it can function as the head noun of a phrase with a nominal modifier, e.g. ulaa aya-ni ‘the good (best of the) reindeer’ (reindeer good-PX3SG). Numerals show some deviations from the agreement patterns, suggesting that they may in some cases not be simple modifiers but, rather, parallel head nouns. A nominal following a numeral is normally unmarked but can also take plural marking, e.g. ilaa nari ~ ilaa nari-l ‘three men’ (three man-PL). Case agreement, mostly only for the accusative, is also occasionally observed between a numeral and a following nominal, e.g. ilaa narri-e (three man-ACC) ~ ilaam-ba narri-e (three-ACC man-ACC).

452 Toshiro Tsumagari and Yoshiko Yamada

Possession is expressed by the possessive suffixes, but the phrase can facultatively also contain the first and second person pronominal genitives or the adnominal form of the reflexive pronoun, e.g. (mini) ƞinda-bi ‘my dog’ (1SG.GEN dog-PX1SG) : (mini) ƞinda-a-​wwie (1SG.GEN dog-ACC-PX1SG), (sini) suƞdatta-ƞu-si ‘your fish’ (2SG.GEN fish-​AL-PX2SG) : (sini) suƞdatta-ƞƞo-o-si (2SG.GEN fish-AL-ACC-PX2SG), (meene) ƞaala-bi ‘one’s own hand’ (REFL.ADN hand-RX). With a nominal possessor the possessive relationship is indicated by the third person possessive suffix attached to the head noun, cf. e.g. eekte putte-ni ‘the woman’s child’ (woman child-PX3SG) vs. eekte putte ‘daughter, girl’ (woman child). The third person possessive suffix is also freely used with reference to non-human possessors, e.g. gasa omo-ni ‘bird’s nest’ (bird nest-PX3SG), xulda oo-du-ni ‘at the corner of the box’ (box corner-DAT-PX3SG). A participle followed by a head noun functions as the equivalent of a relative clause. If the head noun is the subject of the relative clause (conjunct use), no person marking is required, but in other cases (oblique or disjunct use) the subject is marked by the possessive suffixes, cf. e.g. suƞdatta-a waa-xa nari ‘the man who caught fish’ (fish-ACC kill-PTCP.PRF man) vs. tari nari waa-xa-ni suƞdatta ‘the fish which he caught’ (that man kill-PTCP. ​PRF-PX1SG fish). Some nominal arguments other than the subject and object can also be relativized without any indication of their grammatical role in the relative clause, e.g. suƞdatta-a mii-xem-bi kucige ‘the knife with which I cut fish’ (fish-ACC cut-PTCP.​PRF-​PX1SG knife), suƞdatta-a ekse-xe-si ostooli ‘the table on which you put the fish’ (fish-ACC put-PTCP.​PRF-PX2SG table). SENTENCE TYPES In the basic sentence pattern, the subject and object precede the predicate in this order (SOV). Deviations from this order are, however, often found, both for functional reasons and because of Russian interference. In the following a few other sentence types and related phenomena are discussed in some more detail: •

•

In equative sentences with a nominal (including adjectival) predicate and with a present tense reference, no copula is needed, but in the past tense the past tense form of the copula bi- ‘to be’ : PST bi-cci is used. The nominal predicate can optionally take person markers of the first and second persons. In this connection, the first person singular marker is -mbi also after nouns that do not contain an etymological final nasal, e.g. bii uilta(-mbi) ‘I am a Uilta’ (1SG Uilta-1SG) : bii uilta bi-cci-mbi ‘I was a Uilta’ (1SG Uilta be-PST-1SG); sii purige(-si) ‘you are young’ (2SG young-2SG) : sii purige bi-cci-si ‘you were young’ (2SG young be-PST-​2SG); tari ulaa ‘that is a reindeer’ (that reindeer) : tari ulaa bi-cci(-ni) ‘that was a reindeer’ (that reindeer be-PST-3SG); ye-du namauli ‘it is warm here’ (this-DAT warm) : ye-du namauli bi-cci ‘it was warm here’ (this-DAT warm be-PST). Existential sentences always require the copula bi-, e.g. ostooli oyo-du-ni bicixe bi-i-ni ‘there is a book on the table’ (table top-DAT-PX3SG book be-PRS-​3SG), tari nari dug-j-i bi-i-ni ‘he is at home’ (that man home-DAT-RX be-PRS-3SG). This pattern can also be used as a possessive construction, e.g. min-du ilaa ulaa bi-i-ci ~ bi-i-ni ‘I have three reindeer’ (1SG-​DAT three reindeer be-PRS-3PL ~ be-PRS3SG), but possession can also be expressed by an equative sentence with the predicate in the proprietive form in -lu. In this construction, an accompanying numeral takes the instrumental marker, e.g. bii ilaan-ji ulaa-lu-bi ‘I have three reindeer’ (1SG

Uilta 453

•

three-INSTR reindeer-PROPR-1SG). When the possession is alienable, the alienability marker -ƞu- is added, e.g. tari nari gumasikka-ƞu-lu ‘he has money’ (that man money-AL-PROPR). The form in -ƞu-lu can also be used attributively, e.g. gumasikka-ƞu-lu nari ‘a man with money’ (money-AL-PROPR man). Negation is expressed by the negation verb e- : AOR = PRS e-si- : PST e-cci- in combination with the connegative form of the main verb, e.g. tari nari e-si-ni ƞenne-e ‘he does not go’ (that man NEG-PRS-3SG go-CONNEG) : tari nari e-cci-ni ƞenne-e ‘he did not go’ (that man NEG-PST-3SG go-CONNEG). The negation of equative sentences requires the presence of the copula bi- ‘to be’, e.g. bii uilta e-si-wi bi-e ‘I am not a Uilta’ (1SG Uilta NEG-PRS-1SG be-CONNEG), sii purige e-cci-si bi-e ‘you were not young’ (2SG young NEG-PST-2SG be-CONNEG), tari munu ƞinda-pu e-si bi-e ‘that is not our dog’ (that 1PL.GEN dog-PX1PL NEG-PRS be-CONNEG). Prohibition is expressed by the imperative form of the negative verb e-jjie, e.g. ejjie ƞenne-e ‘don’t go!’ (PROHIB go-CONNEG). Uilta, like Ulcha, has also the “nondumitive” negator e-ccieli ‘not yet’, with which the main verb takes person marking of the aorist type, e.g. bii eccieli ƞenne-e-mi ‘I do not go yet’ (1SG NONDUM go-CONNEG-1SG). For the negation of existential sentences, the privative noun anaa ‘without’ can be employed, e.g. tamaciga nari anaa ~ e-si-ni bi-e ‘there is no such man’ (such man PRIV ~ NEG-PRS-3SG be-CONNEG). In combination with the privative noun the proprietive suffix is replaced by the “partitive” suffix -la, e.g. sii gumasikka-​ƞu-la anaa-si ‘you have no money’ (2SG money-AL-PART PRIV-2SG). In the past tense, the privative noun is combined with the copula bi-cci, as in [1]:

[1] goropci-du suƞdatta bi-ccin=dee buda anaa old.time-dat fish be-pst-ptcl rice priv ‘In olden times there was fish, but there was no rice.’ •

•

bi-cci be-pst

Interrogation in polar questions is indicated by the interrogative clitic =i ~ =yi, usually accompanied by a rising intonation, e.g. eri aya ulaa=yi ‘is this a good reindeer?’ (this good reindeer=INTERR), sii e-cci-si tukkɵ-ɵ=yi ‘didn’t you fall?’ (2SG NEG-PST-2SG fall-​CONNEG=​INTERR). In the answer, the particle ii ‘yes’ can confirm both an affirmative and a negative question, e.g. sii ƞenni-e-si=i ‘do you go?’ (2SG go-PRS-2SG-​INTERR)—ii, ƞenni-e-wi ‘yes, I (do) go’ (AFF go-PRS1SG) vs. ii, e-si-wi ƞenne-e ‘no, I do not go’ (AFF NEG-PRS-1SG go-CONNEG). Alternatively, a negative answer can also be introduced by the privative noun anaa, e.g. anaa, e-si-wi ƞenne-e ‘no, I do not go’ (PRIV NEG-PRS-1SG go-CONNEG). For non-polar questions the non-obligatory corrogative clitic =ga ~ =ka can be used, e.g. eri xai=ga ‘what is this?’ (this what=CORR), ƞui sinda-xa-ni(=ga) ‘who came?’ (who come-PST-3SG=CORR), sii xai-wa ite-xe-si(=ge) ‘what did you see?’ (2SG what-ACC see-PST=CORR). In the comparative construction the standard of comparison takes the instrumental case marker, e.g. eri moo caa moo-ji gugda ‘this tree is taller than that tree’ (this tree that.OBL tree-INSTR high), eri ƞinda caa ƞinda-ji kusal-ji tuksi-e-ni ‘this dog runs faster than that dog’ (this dog that.OBL dog-INSTR fast-INSTR run-PRS-3SG). For the positive and superlative degrees, the items gese ‘together’ > ‘equally’, SIM =ƞAci ‘similar(ly)’, and tenee ‘most’ are used, e.g. eri moo caa moo-du gese gugda ‘this tree is as tall as that tree’ (this tree that.OBL tree-DAT together high), eri moo caa moo=ƞoci gugda e-si bi-e ‘this tree is not so tall as that tree’ (this tree that.OBL

454 Toshiro Tsumagari and Yoshiko Yamada

•

•

tree=SIM high NEG-PRS be-CONNEG), eri moo tenee gugda ‘this tree is (the) tallest’ (this tree most high). Passive and causative are expressed by adding the passive-causative suffix -(B)OOn~ -(B)un- (with a wide range of allomorphy) to the verbal stem. As a rule, the agent takes the dative case, while the causee stands in the accusative, e.g. bii caa nari-du paacc-ooc-ci-mbi ‘I was hit by him’ (1SG that.OBL man-DAT hit-PASS-PST-1SG), bii caa nari-du ulaa-bi waa-wooc-ci-mbi ‘my reindeer was killed by him’ (1SG that. OBL man-DAT reindeer-RX kill-PASS-PST-1SG), bii caa narri-e ƞenne-uc-ci-mbi ‘I let/made him go’ (1SG that.OBL man-ACC go-CAUS-PST-1SG). If there are two accusative arguments in a causative sentence, the causee usually precedes the patient, except in semantically obvious cases, e.g. bii caa narri-e caa eekte-e itteuc-ci-mbi ‘I let him see her’ (1SG that.OBL man-ACC that.OBL woman-ACC seeCAUS-PST-1SG), yee mɵɵ-wɵ caa narri-e um-moon-u ‘let him drink this water!’ (this.OBL water-ACC that.OBL man-ACC drink-CAUS-IMP.2SG). Wish and intention are expressed on the verb by the suffixes -mu- : -mu-si- [desiderative] and -kitA- ~ (y)itA- : -kicc- ~ (y)icc- [intentive], respectively, e.g. bii ca-la ƞenemu-si-wi ‘I wish/want to go there’ (1SG that-LOC go-DESID-PRS-1SG), bii e-si-wi deptu-mu-si ‘I don’t want to eat’ (1SG NEG-PRS-1SG eat-DESID-CONNEG); bii ca-la ƞene-icci-e-wi ‘I intend to go there’ (1SG that-LOC go-INTENT-PRS-1SG), also in combination with the causative suffix: bii caa narri-e de-pɵɵk-kicci-e-wi ‘I intend to let him eat’ (1SG that.OBL man-ACC eat-CAUS-INTENT-PRS-1SG).

COMPLEX SENTENCES Complex sentences are built with the help of subordinate verbal forms, which include both the actual converbs and various quasiconverbal constructions (case forms of participles used as independent head nouns). •

• •

Conjunct converbs: patala ine-mi sinda-xa-ni ‘a girl came smiling’ (girl smile-CV.​ CONN come-PST-3SG), bii duk-takk-i ƞene-gecci akpac-ci-mbi ‘I went home and lay down’ (1SG home-DIR-RX go-CV.CTXT lie.down-PST-1SG), putte bi-ƞessie ye-le sinda-xa-mbi ‘I came here when I was a child’ (child be-CV.CTEMP this-LOC come-PST-1SG), sini sinda-dalaa bii xalacc-i-wi ‘I will wait until you come’ (2SG. GEN come-CV.TERM 1SG wait-PRS-1SG), tari nari sinda-pie ye-we icci-e-ni ‘he will see this when he comes’ (that man come-CV.CSEC this-ACC see-PRS-3SG). Disjunct converbs: tari nari sinda-uta-nni-e bii ƞenni-e-wi ‘I will go once he comes’ (that man come-CV.COND-3SG-EXT 1SG go-PRS-1SG), tari nari sinda-ayi-ni bii ƞenni-e-wi ‘I will go if he comes’ (that man come-CV.AOR-3SG 1SG go-PRS-1SG). Quasiconverbal constructions: tari nari buc-cin-du-ni bii soƞo-xo-mbi ‘when he died, I cried’ (that man die-PTCP.PRF-DAT-PX3SG 1SG cry-PST-1SG), ƞeele sinda-xan-du-ni ulaa tuta-xa-ci ‘the reindeer ran away because there came a wolf’ (wolf come-​PTCP.​PRF-​DAT-​PX3SG reindeer run.away-PST-3PL). Participles in the accusative case form are used as equivalents of complement clauses, e.g. bii sii ~ sini sinda-xam-ba-si saa-ri-wi ‘I know that you came’ (1SG 2SG ~ 2SG.GEN come-PTCP.PRF-ACC-PX2SG know-PRS-1SG), bii tari nari e-si-we-ni sinda-a saa-ri-wi ‘I know that he will not come’ (1SG that man NEG-PTCP.​IMPRFACC-PX3SG come-CONNEG know-PRS-1SG). Quotations are, however, inserted

Uilta 455

•

in the main clause in finite form with no special marker of subordination, e.g. tari nari [mɵɵ-wɵ bɵɵ-ru] uc-ci-ni ‘he told (me) to give (him) water’ (that man water-ACC give-IMP.2SG say-PST-3SG), bii [tari nari sinji-e-ni] mɵrɵcci-wi ‘I think that he will come’ (1SG that man come-PTCP. IMPRF-PX3SG think-PRS-1SG). Concessive clauses are formed by the clitic =ddAA ‘(even) though’, as in [2]. This clitic can also be added to nominal predicates, as in [3]:

[2] tari nari sinda-xan=daa bii e-si-wi that man come-pst=conc 1sg neg-prs-1sg ‘Even though he has come, I will not go.’

ƞenne-e go-conneg

[3] tari eekte aya eekte=ddee bii e-si-wi anana-si that woman good woman=conc 1sg neg-prs-1sg like-conneg ‘Even though she is a beautiful woman, I do not like her.’ LEXICON AND LANGUAGE CONTACTS In its geographical context on Sakhalin, Uilta has for several centuries been in contact with Sakhalin Ghilyak (Nighvng) in the north and Sakhalin Ainu (Enchiw) in the south. As a result of extensive trading activities between Japan, Sakhalin, and Manchuria, Uilta has historically interacted also with the nearby Tungusic languages on the continent, including Manchu, as well as with Yakut. In most of these contact relationships, Uilta has been the receiving part. The impact of Japanese and Russian started growing in the 19th century and resulted during the period 1905–1945 in different patterns of bilingualism for the southern (Japanese) and northern (Russian) groups of Uilta. The contacts with Ghilyak have left many loanwords of Ghilyak origin in Uilta, including even such items of basic vocabulary as ƞoyokko ‘egg’ ← Ghilyak ƞoyq (*ƞoyoko), unigeri ‘star’ ← Ghilyak uñɣř (*unigeri). Other items are connected with local technology, e.g. kalumuri ‘board’ ← Ghilyak qalmř (*kalumuri), xala/n ‘dog collar’ ← Ghilyak halƞ (*halan). In these and several other items Uilta preserves an otherwise undocumented early shape of the Ghilyak originals. Some Ghilyak words present in Uilta are also attested in Ulcha, e.g. musi/n [fish skin jelly with berries] = Ulcha mosɪ/n ← Ghilyak mos (*mosi). There is also a large corpus of regional vocabulary shared by Uilta and Ghilyak for which the ultimate origin and direction of borrowing remain difficult to determine. Ainu loanwords in Uilta are mainly connected with marine products and related technology, e.g. nuna [kind of sea urchin] ← Ainu nona, sarukki ~ sirukki ‘herring’ ← Ainu (Taraika) herokki, tasuri ‘hatchet’ ← Ainu tasiro, sairi [wooden bar for hanging and drying fish] < *sakiri ← Ainu sakir. The Ainu have also acted as mediators between the Uilta and the Japanese, as suggested by the ethnonym sisa/n ‘Japanese’ ← Ainu sisam. Many Japanese words have entered Uilta via Ainu, e.g. occii : ACC occikki-e [Japanese tray] < *occiki ← Ainu otcike ← Japanese *wosiki, sittoo : ACC sittokko-o < *sittoko ‘barrel’ ← Ainu sintoko ← Japanese *sintoko, tuŋa ‘hoe’ ← Ainu tonka ← Japanese tō-guwa 唐鍬. Such items obviously represent Japanese objects transmitted to the Uilta by the Ainu.

456 Toshiro Tsumagari and Yoshiko Yamada

There are also words that may have borrowed from Japanese to Uilta without the intermediation of Ainu, e.g. icaruu : ACC icarukko-o kashi), kamusai : ACC kamusakki-e ‘sake [Japanese rice wine]’ < *kamusaki ← Japanese *kamu-sake ‘sake for/of gods’, masaari ‘ax’ < *masakari ← Japanese masa(-)kari, satu ‘sugar’ ← Japanese satō, temme ‘barge’ ← Japanese temma (tenma). Some of these borrowings must be relatively old, dating back to at least the 19th century, since they are based on Japanese originals with obsolete meanings and/or archaic (premodern) shapes. The Manchu lexical items present in Uilta represent objects and concepts transmitted by trade. Typically these items are also attested in the other Tungusic languages of the Amur region, including Nanaic, Orochic, and Neghidal. Some of them may be relatively old borrowings, e.g. cauxa ‘soldier’ ← Manchu *cauxa (cooha), sumala/n ‘bag’ ← Manchu sumala (sumala), while others have probably been secondarily transmitted to Uilta by the neighbouring languages, e.g. gecuweri [Manchurian formal costume bought from the Ulcha as treasure] ← Manchu gecuxeri (gecuheri), tekkere [musical instrument of the Ghilyak type] = Ghilyak t‘əƞrəƞ (*t‘eƞgere) ← Manchu teƞgeri (tenggeri) [three-stringed lute]. Some of the Manchu items derive from Chinese, e.g. moicca ~ miĕucca ‘gun’ ← Manchu *myaucan (miyoocan) ← Chinese niaoqiang 鳥鎗 ‘shotgun’. Borrowings from Yakut and Ewenki (which has been present on Sakhalin since the mid 19th century) are mainly connected with reindeer herding, e.g. ari ‘butter’ (which the Uilta used to make from reindeer milk) ← Ewenki arii ← Yakur arï ï, juu.r-egde ‘two-headed reindeer’ ← Ewenki juur-egde, cooraa [large bell attached to the neck of a male reindeer pulling a sledge] ← Yakut cuoraan, naarta ~ naaritta [tall reindeer sledge introduced by the Yakut] ← Yakut naarta ← Russian nárta. There are also a couple of words connected with card games: yeiregde ‘number nine (in playing cards)’ ← Ewenki yegi-regde, toora ‘clubs (in playing cards)’ ← Yakut tuora. Russian borrowings cover a wide range of fields, e.g. (religion) baacikka ‘Orthodox priest’ ← Russian bát’ka, satana ‘devil’ ← Russian sataná, (guns) coompira ‘cleaning rod for guns’ ← Russian shómpol, sɵmmɵɵki ‘breechblock’ ← Russian zamók, (monetary economy) gumaska ~ gumasikka ‘money’ ← bumázhka ‘paper’, gumaanikka ~ gumaarinikka ‘wallet’ ← Russian bumázhnik, (foodstuff) lipieske ‘bread’ ← Russian lepëshka ‘flat bread’, saakari ‘cube sugar’ ← Russian sákhar, (household items) puuntu ‘balance’ ← Russian funt, ostooli ‘(dining) table’ ← Russian stol, kookkuna ‘kitchen’ ← Russian kúkhnya, (clothing) ulbaaxu ‘shirt’ ← Russian rubákha, (card games) cielba ‘hearts (in playing cards)’ ← Russian chérva, tuusi ‘ace’ ← Russian tuz, (other) warinaakka ‘fugitive prisoner’ ← Russian varnák. In view of the semantic and formal adaptation, many of these loanwords must be relatively old, dating back to the 19th century. A large number of modern Russian words have been borrowed later in the Soviet and Post-Soviet periods. Apart from lexical borrowings, Uilta may have adopted some grammatical elements and features from the neighbouring languages. It has been proposed that, for instance, the habitive marker -bukki ~ -wukki may have been borrowed from the Ewenki suffix for habitual action -w(u)kii. Similarly, the polite imperative marker -yA could be a borrowing from Ghilyak, where -ya is the regular marker of the second person singular imperative. The background of most other morphological and morphophonological properties of Uilta, which distinguish it from its closest relatives, remains unknown.

Uilta 457

REFERENCES AND FURTHER READING Alonso de la Fuente, José Andrés (2013) ‘Vowel epenthesis in Orok ulisä ‘meat’’,『北 方人文研究』Journal of the Center for Northern Humanities 6: 121–127, 札幌 [Sapporo]. Czerwinski, Patryk (2019) ‘Complex predicates in Uilta (Orok)’, in: Éva A. Csató & Lars Johanson & Birsel Karakoç (eds.), Ambiguous Word Sequences in Transeurasian Languages and Beyond, Turcologica 120: 99–115, Wiesbaden: Harrassowitz Verlag. Gusev, Valentin (2016) ‘Figura etymologica in Uilta’,『北方人文研究』Journal of the Center for Northern Humanities 9: 59–74, 札幌 [Sapporo]. Hayata, Teruhiro (1979) ‘Orok noun morphophonology’,『アジア・アフリカ文法研 究』Asian and African Languages and Linguistics 8: 129–162, 東京 [Tokyo]. Ikegami, Jirō 池上二良 (1953)「ツングース語オロッコ方言の母音音素ɵについ て」[On the vowel phoneme ɵ in the Orokko dialect of Tungus],『言語研究』Journal of the Linguistic Society of Japan 22/23: 75–78, 東京 [Tokyo]. Republished in Ikegami (2001a). Ikegami, Jirō 池上二良 (1955)「トゥングース語」[The Tungusic languages], in: 市河 三喜 [Ichikawa Sanki] & 服部四郎 [Hattori Shirō] (eds.)『世界言語概説』（下）, 441–488, 東京 [Tokyo]: 研究社. Ikegami, Jirō (1956) ‘The substantive inflection of Orok’,『言語研究』Journal of the Linguistic Society of Japan 30: 77–96, 東京 [Tokyo]. Republished in Ikegami (2001a). Ikegami, Jirō (1959) ‘The verb inflection of Orok’,『国語研究』9: 34–79, 東京 [Tokyo]. Republished in Ikegami (2001a). Ikegami, Jirō (1968) ‘The Orok third person pronoun nooni’, Ural-Altaische Jahrbücher 40: 82–84. Republished in Ikegami (2001a). Ikegami, Jirō (1970) ‘Orok kinship terminology’,『北方文化研究』Bulletin of the Institute for the Study of North Eurasian Cultures 4: 133–156, 札幌 [Sapporo]. Republished in Ikegami (2001a). Ikegami, Jirō 池上二良 (1971)「19世紀なかごろのオロッコ語集」[A vocabulary of mid-19th century Orok],『北方文化研究』Bulletin of the Institute for the Study of North Eurasian Cultures 5: 79–184, 札幌 [Sapporo]. Ikegami, Jirō (1973) ‘Orok verb-stem-formative suffixes’,『北方文化研究』Bulletin of the Institute for the Study of North Eurasian Cultures 7: 1–17, 札幌 [Sapporo]. Republished in Ikegami (2001a). Ikegami, Jirō 池上二良 (ed.) (1983)『川村秀弥採録「カラフト諸民族の言語と民 俗」』[Hideya Kawamura’s materials on the languages and folklore of the peoples of Sakhalin] =『ウイルタ民俗文化財緊急調査報告書』[Reports on urgent research concerning Uilta culture and folklore] 5, 札幌 [Sapporo]: 北海道教育委員会. Ikegami, Jirō 池上二良 (1984)『ウイルタ口頭文芸原文集』[Uilta oral literature: A collection of texts] =『ウイルタ民俗文化財緊急調査報告書』[Reports on urgent research concerning Uilta culture and folklore] 6, 札幌 [Sapporo]: 北海道教育委員会. Ikegami, Jirō 池上二良 (1990)「日本語・北の言語間の単語借用」[Lexical borrowings between Japanese and neighbouring languages in the north],『北海道方言研究 会会報』30: 2–11, 札幌 [Sapporo]. Republished in Ikegami (2004). Ikegami, Jirō (1985) ‘B. Pilsudski in Uilta and Olcha studies’, in: Proceedings of the International Symposium on B. Pilsudski’s Phonographic Records and the Ainu Culture, 168–172, Sapporo: Hokkaido University.

458 Toshiro Tsumagari and Yoshiko Yamada

Ikegami, Jirō 池上二良 (1993a)「ウイルタ語代名詞とその格変化」[Uilta pronouns and their inflections], in: 『地域・情報・文化：札幌大学女子短期大学部創立25周 年記念論文集』 [Volume commemorating the 25th anniversary of Sapporo Women’s Junior College], 363–371, 札幌 [Sapporo]: 響文社. Republished in Ikegami (2001a). Ikegami, Jirō (1993b) ‘A brief history of the study of the Uilta language’, in: 村崎 恭子 [Kyōko Murasaki] (ed.)『サハリンの少数民族』[Ethnic minorities of Sakhalin], 65–71, 横浜 [Yokohama]: 横浜国立大学. Republished in a revised version in Ikegami (2001a). Ikegami, Jirō 池上二良 (1994a)「ウイルタ語の南方言と北方言の相違点」[Differences between the southern and northern dialects of Uilta], 『北海道立北方民族博 物館研究紀要』 Bulletin of the Hokkaido Museum of Northern Peoples 3: 9–38, 網 走 [Abashiri]. Republished in Ikegami (2001a). Ikegami, Jirō [Дзиро Икэгами] (1994b) ‘Проект письменности уйльтинского языка’, Acta Slavica Iaponica 12: 253–258, 札幌 [Sapporo]. Republished in Ikegami (2001a). Ikegami, Jirō [Дзиро Икэгами] (1996) ‘Письменная практика на уйльтинском языке: дополнение к проекту письменности уйльтинского языка’. Acta Slavica Iaponica 14: 120–123, 札幌 [Sapporo]. Republished in Ikegami (2001a). Ikegami, Jirō 池上二良 (1997)『ウイルタ語辞典』[A dictionary of the Uilta language spoken on Sakhalin], 札幌 [Sapporo]: 北海道大学図書刊行会. Ikegami, Jirō [Дзиро Икэгами] (1998) ‘Письменная практика на уйльтинском языке: продолжение’ Acta Slavica Iaponica 16: 181–183, 札幌 [Sapporo]. Republished in Ikegami (2001a). Ikegami, Jirō 池上二良 (2001a)『ツングース語研究』 [Researches on the Tungus Language], 東京 [Tokyo]: 汲古書院. Ikegami, Jirō 池上二良 (2001b)「ウイルタ語動詞活用大要」[A concise outline of the verb inflection of Uilta],『環北太平洋の言語』Languages of the North Pacific Rim 7: 157–166, 吹田 [Suita]: 大阪学院大学 [Osaka Gakuin University]. Ikegami, Jirō 池上二良 (2002)『増訂ウイルタ口頭文芸原文集』[Uilta oral literature: A collection of texts, revised and enlarged edition, with an audio CD], 吹田 [Suita]: 大阪学院大学. Ikegami, Jirō 池上二良 (2004)『北方言語叢考』[Essays on languages of Northeast Asia], 札幌 [Sapporo]: 北海道大学図書刊行会. Ikegami, Jirō 池上二良 (2007) Сказания и легенды народа уйльта, translated by E. A. Bibikova [Е. А. Бибикова] and edited by Toshirō Tsumagari [津曲敏郎],『ツン グース言語文化論集』[Publications on Tungus Languages and Cultures] 38, 札幌 [Sapporo]: 北海道大学大学院文学研究科. Ikegami, Jirō [Дзиро Икэгами] & E. A. Bibikova [Е. А. Бибикова] & L. R. Kitazima [Л. Р. Китазима] & S. Minato [С. Минато] & T. P. Roon [Т. П. Роон] & I. Ia. Fedjaeva [И. Я. Федяева] (2008) Уилтадаирису: Говорим по-уйльтински [A Uilta primer], Южно-Сахалинск [Yuzhno-Sakhalinsk]: Сахалинское книжное издательство. Janhunen, Juha (2014) ‘On the ethnonyms Orok and Uryangkhai’, Studia Etymologica Cracoviensia 19: 71–81, Krakόw. Magata, Hisaharu 澗潟久治 (1981)『ウイルタ語辞典』[A dictionary of the Uilta language], 網走 [Abashiri]: 網走市北方民俗文化保存協会. Majewicz, Alfred F. (1985a) ‘Vicissitudes of B. Pilsudski’s lexicological collections’, in: Proceedings of the International Symposium on B. Pilsudski’s Phonographic Records and the Ainu Culture, 173–183, Sapporo: Hokkaido University.

Uilta 459

Majewicz, Alfred F. (1985b) Materials for the Study of the Orok (Uilta) Language and Folklore, I: Фонетические и грамматические замечания по языку ороков, орокские тексты [Phonetic and grammatical remarks on the Orok language and Orok texts], Poznań: Adam Mickiewicz University. Majewicz, Alfred F. (ed.) (2011) The Collected Works of Bronisław Piłsudski 4: Materials for the Study of Tungusic Languages and Folklore. Berlin: De Gruyter Mouton. Missonova, L. I. [Л. И. Миссонова] (2013) Лексика уйльта как историко-​ этнографический источник [Uilta lexicon as a source on history and ethnography], Москва [Moscow]: “Наука”. Nakanome, Akira 中目覚 (1917) 『オロッコ文典』 [A grammar of the Orokko language]. 東京 [Tokyo]: 三省堂. Nakanome, Akira (1928) Grammatik der Orokko-Sprache, aus dem Japanischen übersetzt von Dr. W. Othmer, Osaka: Tōyō Gakkai. Novikova, A. I. [А. И. Новикова] & L. I. Sem [Л. И. Сем] (1997) ‘Орокский язык’ [The Orok language], in: Монгольские языки, тунгусо-маньчжурские языки, японский язык, корейский язык [Mongolic, Tungusic, Japanese, Korean], Языки Мира [Languages of the World] [3]: 201–215, Москва [Moscow]: “Индрик”. Ozoliņa, L. V. [Л. В. Озолиня] (2001) Орокско-русский словарь [Orok-Russian dictionary]. Новосибирск [Novosibirsk]: Издательство СО РАН. Ozoliņa, L. V. [Л. В. Озолиня] (2013) Грамматика орокского языка [A grammar of Orok (Uilta)]. Новосибирск [Novosibirsk]: Академическое издательство “Гео”. Ozoliņa, L. V. [Л. В. Озолиня] & I. Ya. Fedyaeva [И. Я. Федяева] (2003) Орокскорусский и русско-орокский словарь [Orok-Russian and Russian-Orok dictionary]. Южно-​Сахалинск [Yuzhno-Sakhalinsk]: Сахалинское книжное издательство. Petrova, T. I. [Т. И. Петрова] (1967) Язык ороков (ульта) [The language of the Orok (Ul’ta)]. Ленинград [Leningrad]: “Наука”. Petrova, T. I. [Т. И. Петрова] (1968) ‘Орокский язык’ [The Orok language], in: Монгольские, тунгусо-маньчжурские и палеоазиатские языки, Языки народов СССР [Languages of the Soviet Union] 5: 172–190, Ленинград [Leningrad]: “Наука”. Pevnov, Aleksandr [M.] (2009) ‘On some features of Nivkh and Uilta (in connection with prospects of Russian-Japanese collaboration)’, in: 津曲敏郎 [Tsumagari Toshirō] (ed.),『サハリンの言語世界』[The linguistic world of Sakhalin], 113–125, 札幌 [Sapporo]: 北海道大学大学院文学研究科北方研究教育センター. Pevnov, A. M. [А. М. Певнов] (2014) ‘Текст на орокском (уильта) языке с комментариями’ [An Orok (Uilta) text with comments], in: В. Ф. Выдрин [V. F. Vydrin] & Н. В. Кузнецова [N. V. Kuznecova] (eds.), От Бикина до Бамбалюмы, из варяг в греки: Экспедиционные этюды в честь Елены Всеволодовны Перехвальской [Field-inspired essays in honour of Elena V. Perekhvalskaya], 55–66, Санкт-​Петербург [St. Petersburg]: “Нестор-История”. Pevnov, Aleksandr [M.] (2016) ‘On the specific features of Orok as compared with the other Tungusic languages’, in: Ekaterina Gruzdeva & Juha Janhunen (eds.), Crosslinguistics and Linguistic Crossings in Northeast Asia, Studia Orientalia 117: 47–63, Helsinki. Pevnov, Aleksandr M. (2017) ‘On the origin of Uilta (Orok) nōni ‘he or she’’,『北方人 文研究』Journal of the Center for Northern Humanities 10: 71–78. 札幌 [Sapporo]. Pevnov, Aleksandr M. (2020) ‘Historical contacts of the Uilta (Orok), with special attention to elements of unknown origin’, in: Anna Król & Wioletta Laskowska-Smoczyńska (eds.), Bronisław Piłsudski w stulecie śmierci. W stronę niepodległej: materiały

460 Toshiro Tsumagari and Yoshiko Yamada

pokonferencyjne/Bronisław Piłsudski on the centennial of his death. Towards an independent homeland: proceedings of the conference, 208–217, Kraków: Manggha Museum of Japanese Art and Technology. Tsumagari, Toshirō 津曲敏郎 (1980)「『ツングース・満州諸語比較辞典』のウイル タ語単語の検討」[Remarks on Uilta words in the Comparative Dictionary of the Tungusic Languages], 『ウイルタ族言語文化調査研究報告』Research Reports on the Language and Culture of the Uiltas (Oroks) 1: 11–25, 札幌 [Sapporo]: 北海道大学文学部. Tsumagari, Toshirō 津曲敏郎 (1983)「ウイルタ語のアクセント」[Accent in Uilta],『アジア・アフリカ文法研究』Asian and African Languages and Linguistics 12: 75–84, 東京 [Tokyo]. Tsumagari, Toshirō (1985a) ‘Grammatical outline of Uilta’,『アジア・アフリカ文法 研究』 Asian and African Languages and Linguistics 14: 1–15, 東京 [Tokyo]. Tsumagari, Toshirō (1985b) ‘On B. Pilsudski’s Orok vocabulary’, Proceedings of the International Symposium on B. Pilsudski’s Phonographic Records and the Ainu Culture, 184–189, Sapporo: Hokkaido University. Tsumagari, Toshirō 津曲敏郎 (1987)「B. ピウスツキのオロッコ語文法記述につい て」[Remarks on B. Pilsudski’s Orok grammatical sketch], in: Katō Kyūzō 加藤九祚 & Kotani Yoshinobu 小谷凱宣 (eds.),『ピウスツキ資料と北方少数民族文化の研 究』[Piłsudski’s Materials on Northern Peoples and Cultures] = 国立民族学博物館 研究報告別冊5 [Bulletin of the National Museum of Ethnology, Special Issue No. 5]: 283–294, 吹田 [Suita]: 国立民族学博物館. Tsumagari, Toshirō (2009a) ‘Grammatical outline of Uilta (revised)’, Journal of the Graduate School of Letters 4: 1–21, [Sapporo]: Hokkaido University. Tsumagari, Toshirō 津曲敏郎 (2009b)「サハリンの言語世界：単語借用から見 る」[The linguistic world of Sakhalin, with a view on lexical borrowing], in: 津曲敏郎 [Tsumagari Toshirō] (ed.),『サハリンの言語世界』 [The linguistic world of Sakhalin], 1–10, 札幌 [Sapporo]: 北海道大学大学院文学研究科北方研究教育センター. Tsumagari, Toshirō 津曲敏郎 (2010)「ウイルタ語文例」[A collection of short sentences of Uilta],『環北太平洋の言語』Languages of the North Pacific Rim 15: 159– 177, 富山 [Toyama]. Tsumagari, Toshirō 津曲敏郎 (2014) ‘Remarks on the Uilta folktale text collected by B. Pilsudski’, 『北方人文研究』 Journal of the Center for Northern Humanities 7: 83–94, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2007)「ウイルタ語後置文の機能論的分析」[A functional analysis of postposing in Uilta sentences],『環北太平洋の言語』Languages of the North Pacific Rim 14: 87–102, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2008)「ウイルタ語口頭文芸の伝聞形式：サハリンに おける言語接触の可能性」[Hearsay forms in Uilta oral literature: a possibility of language contact in Sakhalin]『北海道民族学』Hokkaido Journal of Ethnology 4: 63–71, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2009)「ウイルタ語北方言調査の課題と展望」[Topics and prospects for the study of Uilta dialects], in: 津曲敏郎 [Tsumagari Toshirō] (ed.),『サハリンの言語世界』[The linguistic world of Sakhalin], 11–26, 札幌 [Sapporo]: 北海道大学大学院文学研究科北方研究教育センター. Yamada, Yoshiko 山田祥子 (2010) 「ウイルタ語北方言にみられる動詞語尾 -li に ついて」[On the verb ending -li in the northern dialect of Uilta],『環北太平洋の言 語』Languages of the North Pacific Rim 15: 85–100, 富山 [Toyama].

Uilta 461

Yamada, Yoshiko 山田祥子 (2011a)「ウイルタ語北方言テキスト：思い出話2 編」[Uilta texts in the northern dialect: two stories on recollections], 『北海道民族 学』 Hokkaido Journal of Ethnology 7: 48–59, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2011b)「ウイルタ語北方言テキスト：suluktaの作 り方」[Uilta texts in the northern dialect: How to cook sulukta], 『北方言語研 究』 Northern Language Studies 1: 217–228, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2011c)「ウイルタの振楽器「ヨードプ」：復元工程 についてのウイルタ語北方言テキストを中心に」[The musical instrument joodopu of the Uilta: A text in the northern dialect],『北方人文研究』Journal of the Center for Northern Humanities 4: 24–59. 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2012)「ウイルタ語北方言テキスト：ありがとう、池 上先生」[Uilta texts in Northern dialect: Thank you, Ikegami-sensei],『北方人文研 究』Journal of the Center for Northern Humanities 5: 159–172. 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2014)「ウイルタ語北方言テキスト：アザラシ肉に関 する体験談」[Uilta texts in the northern dialect: A story about to seal meat], 『北海 道民族学』 [Hokkaido Journal of Ethnology] 10: 104–113, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2015a)「ウイルタ語北方言テキスト：人喰いお化 けの話」 [Uilta texts in the northern dialect: A story about an ogre],『北方言語研 究』Northern Language Studies 5: 261–280, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2015b)「ウイルタ語調査報告：北部方言の文例 (1)」[A report on the Uilta language: Sample sentences in the northern dialect (1)] 『 北海道立北方民族博物館研究紀要』 Bulletin of the Hokkaido Museum of Northern Peoples 24: 39–58, 網走 [Abashiri]. Yamada, Yoshiko 山田祥子 (2016a)「ギシクタウダ（マリーヤ・ミヘエワ）の生 涯 : ウイルタ語北方言テキスト」[The life of Gisiktauda (Maria Mikheeva): An Uilta text in the northern dialect], 『北方言語研究』 Northern Language Studies 6: 179–201, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2016b)「ウイルタ語調査報告：北部方言の文例 (2)」[A report on the Uilta language: Sample sentences in the northern dialect (2)], 『北海道 立北方民族博物館研究紀要』 Bulletin of the Hokkaido Museum of Northern Peoples 25: 45–66, 網走 [Abashiri]. Yamada, Yoshiko 山田祥子 (2017a)「ウイルタ語北方言の音韻的・形態的特徴： 南方言との相違点を中心に」[Some phonological and morphological features of the northern dialect of Uilta: with a view on differences as compared with the Southern dialect], 『北方人文研究』 Journal of the Center for Northern Humanities 10: 51–70, 札幌 [Sapporo]. Yamada, Yoshiko 山田祥子 (2017b) ‘Language contact among Sakhalin indigenous peoples and some dialectal features of the Uilta language’, in: The Proceedings of the 31st International Abashiri Symposium, 45–52, 網走 [Abashiri]: 一般財団法人北方文化 振興協会. Yamada, Yoshiko 山田祥子 (2017c)「ウイルタ語調査報告：北部方言の文例 (3)」[A report on the Uilta language: Sample sentences in the northern dialect (3)],『北海道立 北方民族博物館研究紀要』Bulletin of the Hokkaido Museum of Northern Peoples 26: 45–66, 網走 [Abashiri]. Yamada, Yoshiko 山田祥子 (ed.) (2014)『ウイルタ長編英雄物語：シーグーニ物語 テキスト』[The Uilta epic story about Siiƞuuni], recited by Satō Chiyo [佐藤チヨ], recorded and transcribed by Ikegami Jirō [池上二良], Russian translation by E. A.

462 Toshiro Tsumagari and Yoshiko Yamada

Bibikova [Е. А. Бибикова], supervised by Tsumagari Toshirō [津曲敏郎], 札幌 [Sapporo]: 北海道大学大学院文学研究科 =『ツングース言語文化論集』[Publications on Tungus Languages and Cultures] 58. Yamada, Yoshiko [Ëсико Ямада] (ed.) (2016) Нучӣкэ принси: Маленький принц на уйльтинском языке [“The Little Prince” in the Uilta language], translated from Russian by E. A. Bibikova [Е. А. Бибикова] and supervised by Tsumagari Toshirō [津曲 敏郎], 札幌 [Sapporo]: 北海道大学大学院文学研究科=『ツングース言語文化論 集』[Publications on Tungus Languages and Cultures] 61.

CHAPTER 17

SPOKEN MANCHU Veronika Zikmundová and Gao Wa

By Spoken Manchu is understood the complex of Manchu oral varieties that have been used since the formation of the Manchu as an ethnopolitical entity in the late 16th century. As speakers of a Jurchenic language, the Manchu continued the linguistic tradition of the mediaeval Jurchen, but the replacement of Written Jurchen by Written Manchu as a new normative literary language after 1599 created a growing difference between the written and spoken forms of the language. Although the written norm has constantly influenced the spoken language, the dispersal of the Manchu during the Qing dynasty led to the diversification of the spoken language into a number of regional and local varieties. At the same time, it is possible that the spoken forms of Manchu have preserved some of the dialectal variation that existed already before the Qing period. During the early part of the Qing dynasty, Manchu was widely spoken as a first language by the Manchu bannermen in different parts of the Qing empire, which gradually expanded from Manchuria to cover also Inner Mongolia (1636), China (1644), Outer Mongolia (1691), Tibet (1720), and Jungaria (1757). The geographical dispersal of the ethnic Manchu led, however, to the gradual loss of the ethnic language, leaving only Written Manchu in active use until the end of the Qing rule (1911). While in the 18th century there may still have been a million native Manchu speakers, as well as a considerable number of people from other Manchurian ethnic groups speaking Manchu as a second language, the number of speakers in the late 19th century was already down to less than 100,000. The decline continued during the Republic of China (1912–1945), Manchukuo (1931–1945), and the People’s Republic, leaving only a few dozen individuals more or less fluent in Spoken Manchu today. This is in stark contrast with the official “minority nationality” of the Manchu (Man zu 滿族), who number today over 10 million (2010). Although Manchu bannermen were dispersed to many different parts of the empire, it was ultimately in the periphery where the language survived longest. Apart from the Jungarian diaspora group of the Sibe, officially classified as representing an ethnic group separate from the Manchu, the Manchu language survived in several garrison towns located in central and northern Manchuria, including Qiqihaer (Tsitsikar), Nenjiang (Mergen), Ning’an (Ningguta), and Aihui (Aigun), all in Heilongjiang Province. The last Manchu speakers have been living in a few villages around these former garrison towns, notably Sanjiazi in Fuyu County and Yibuqi in Tailai County of Qiqihaer City, as well as Dawujiazi, Xiamachang, and Sijitun of Aihui District, in the vicinity of Heihe City on the Amur. All these Manchu garrison communities originated from resettlements during the Qing dynasty. The speakers were families of Manchu bannermen of varying origins who were sent to stay at military outposts, mainly outside of the original Jurchen speaking area. Their language was a variety of standardized Manchu which was spread through government schools in Manchu garrisons. Spoken in relative isolation, the language DOI: 10.4324/9781315728391-17

464 Veronika Zikmundová and Gao Wa

remained sufficiently free of the impact of Chinese and was at the beginning of the 20th century considered more “proper” than Peking Manchu, which at that time was already heavily Sinicized. Until recently, Manchu has been retained best at Sanjiazi (Manchu Ilan Boo ‘three households’), a relatively compact settlement, where Manchu was still used in everyday communication by c. 300 individuals in the 1960s. Later, the interest of researchers also seems to have contributed to the somewhat better preservation of Manchu at Sanjiazi. Even so, the number of fluent speakers at Sanjiazi is today down to just a few, with perhaps some more semi-speakers. According to information from 2017 no speakers use Manchu as a means of communication, but several individuals are capable of carrying a simple conversation, translating sentences from Chinese, or repeating short narratives they have memorized for the purpose of interviews. The situation is probably similar at Yibuqi, where still in the 1980s all elderly Manchu people were fluent in the language. There are, however, no current data available for this village. In the Aihui region Manchu was also retained relatively well until the 1960s, but has declined rapidly, leaving only a small number of speakers in the early 2000s. As of 2019, only one fluent speaker was believed to remain in the area. While Spoken Manchu has been declining, there have been significant efforts to revive the Manchu language on different levels since the 1990s. Both Written and Spoken Manchu have been officially taught in several primary schools in Manchu ethnic areas, and, even more effectively, on many private language courses. Besides this organized effort, a small number of ethnic Manchu of various ages and origins have studied the language independently with Sanjiazi speakers and have acquired a relatively good command of Spoken Manchu. At the same time, efforts have been made to identify and locate all recently used varieties of Spoken Manchu, most of which are extinct today. DIALECTAL DIVISION The Manchu varieties spoken at Sanjiazi, Yibuqi, and Aihui are all fully mutually intelligible and also, to a somewhat lesser degree, mutually intelligible with Sibe. Compared to Sibe, the contemporary Manchurian varieties are, however, less well preserved, having notably limited vocabulary, as well as simplified morphology and sentence structure. Another variety, Lalin Manchu, also known as the Jing (京) or ‘Capital’ Manchu, used to be spoken by descendants of Manchu transferred in 1744 from Peking to the basin of the river Lalin close to Harbin with the aim of opening virgin lands and training military skills which had declined during the life in the imperial capital. Thus, the Lalin dialect was originally close to both Written Manchu and Peking colloquial Manchu. Lalin speakers were bilingual in Peking Mandarin and intentionally preserved their dialectal features, as well as their life-style and religious differences with regard to the older local Jurchenic speakers. In Chinese specialist literature, the varieties spoken by the Manchu in the garrisons at and around Qiqihaer and Aihui, including the variety spoken by the Lalin Manchu, are collectively known as “Bannermen Manchu” (Qiren Manyu 旗人滿語). Another two idioms, known as Alchuka and Bala, have been considered to be Manchu “dialects” by some authors, while others have referred to them as separate “languages”. The Alchuka variety was spoken by the old local Jurchenic population in central Manchuria between Ningguta (Ning’an) and Beduna (modern Songyuan), with the center at Acheng (Alcuka), the former “Upper Capital” of the Jurchen of the Jin dynasty. The traditional occupation of Alchuka speakers was hunting and fishing. Bala was the Manchu label for another local community

Spoken Manchu 465

of Jurchen hunters and fishermen who, instead of entering the banner system, stayed in the densely wooded mountains of the Zhangguangcai range in eastern central Manchuria. Towards the end of the Qing dynasty they were relocated and settled mainly among the speakers of Alchuka Manchu, who gradually absorbed and assimilated their last traces. Both Alchuka and Bala possess features in phonology and lexicon that distinguish them from “Bannermen Manchu”. To some extent these differences may reflect problems in the sources, which may not be fully reliable for these already extinct varieties. At the same time it seems that some of their specific features may indeed reflect traces of dialectal differences that existed in Jurchen before the Qing period. The better known Alchuka variety had less than 20 speakers in the 1960s, the last of whom survived until the 1980s. Another apparently extinct variety may have been spoken by the so-called “Chinese Kyakala” (Qiakela 恰克拉), who represent descendants of an Udihe group (Kyakala) on the left side of the Ussuri, and who still survive as a subethnic entity in the composition of the Manchu. Ultimately, all varieties of Spoken Manchu have to a varying extent been influenced by “Standard Manchu”, which was used in the Manchu army, and which, in turn, was influenced by Written Manchu. There also used to be local interaction between Alchuka and Lalin Manchu, with some influences transferred in both directions. At the level of details, the Sanjiazi, Yibuqi, and Aihui varieties, although mutually intelligible, have also differences. The Yibuqi variety, in particular, shows several features in the phonology and lexicon that are shared with the neighbouring Khorchin Mongolian. The present chapter is based on the Sanjiazi variety. Other varieties will also be touched upon as far as possible and relevant to the discussion. DATA AND SOURCES The rise and decline of the Manchu language as a spoken medium during the Qing period and later is still an unstudied issue in the details, though general summaries and case studies of the historical background have been published by, for instance, Robert H. G. Lee (1970), Ch’en Chieh-hsien (1976), Li Shu (1986), and Guo Meng-xiu & Yin Tie-chao (2008). Daniel Kane (1997) discusses also the more recent process of language revitalization among the Manchu. The existence of Manchu speakers in Manchuria was mentioned several times during the first half of the 20th century. According to Martin Gimm (1981) about ten settlements with an estimated total of several thousand Manchu speakers had been reported by various authors before the Cultural Revolution. Of “Bannermen Manchu”, the first variety to be studied was probably Aihui Manchu. In his work on the social organization of the Manchu Shirokogoroff (1924) quotes examples of Aihui Manchu. The Japanese researchers Yasui Katsumi (1943), and Kobori Iwao (1949) published short reports, and the probably first linguistic study of Aihui Manchu was published by Kōno Rokurō (1944). In the 1960s several linguists in PR China carried out fieldwork in Manchuria. Materials from this fieldwork were mostly published after the Cultural Revolution, the most comprehensive treatments being those by Wang Qingfeng (1984) and Ji Yonghai et al. (1989) for Aihui Manchu, Enhebatu (1995) for Sanjiazi Manchu, and Zhao Jie (1989) for Yibuqi Manchu. Wang Qingfeng (1986) and Zhao Jie (1987) also wrote on Manchu phonology, while Zhao Jinchun (1986) studied a morphological issue. Later, Wang Qingfeng (2005) published a more comprehensive description of Spoken Manchu, based on materials collected earlier from Aihui and Sanjiazi and containing a grammatical overview, comparisons with written Manchu, a vocabulary, and sample texts. While all of these works are in Chinese, Kim Juwon et al. published a volume in English (2008), containing

466 Veronika Zikmundová and Gao Wa

the outcome of fieldwork carried out at Sanjiazi during 2005–2006. In addition to a sketch of phonology and morphology, the volume contains a substantial Manchu-English and English-Manchu vocabulary, as well as a corpus of some 700 sentences with English and Chinese translations. At present, audiovisual material on Manchu from speakers at Sanjiazi and Aihui is also available on Youtube and on Chinese social media. Compared to the relatively well documented and researched varieties of Sanjiazi, Yibuqi, and Aihui, much less material and very few studies are available for the Bala, Alchuka, and Lalin varieties. Probably all extant data on the Bala and Alchuka varieties were collected by the Kyakala scholar Mu Yejun, as published in a series of papers in the late 1980s (Mu 1985, 1986a, 1987b, 1988ab). Based on these data, Ikegami Jirō (1993) and Andreas Hölzl (2017, 2021) have discussed the dialectal position of these two varieties and their general relationship to the Jurchenic languages. Mu Yejun (1986b, 1987a) was also the first to write about the Lalin variety. Slightly different material of Lalin Manchu is presented by Aisin Gioro Yingsheng (1987–2004), partly based on his own knowledge of Jing Manchu, which he acquired in his childhood. Mu Yejun’s materials on Bala, Alchuka, and Lalin Manchu later served as the base for a monographic publication by Zhao Aping and Chaoke (2000), who do not, however, supply any new linguistic material. Their volume also contains a short sketch of the phonology, morphology, and lexicon of Sanjiazi Manchu. The most obscure variety of Manchu remains Kyakala, on which, however, Andreas and Yadi Hölzl (2017, 2018, 2019) have recently presented some new insights. The most comprehensive survey of the historical phonology of all known Manchu varieties is contained in the dissertation of Andrew Joseph (2018). As for actual “Capital Manchu” as used in Peking, it was to some extent still spoken during the first decades of the 20th century but apparently became extinct before any systematic research had been done on it. The only available material on Peking Manchu are the Manchu words used by the Mandarin speaking Manchu in Peking, as collected by Aisin Gioro Yingsheng (1993) and others. There have also been studies, like that of Zhao Jie (1995), on the phonetic and grammatical influence of Manchu on Peking Mandarin. Ultimately, Peking Mandarin, as a whole, may be viewed as Chinese spoken by the Manchu, as has been pointed out by Okada Hidehiro (1992) and Keyou Liu (2021). SEGMENTAL PHONEMES Different opinions exist concerning the Sanjiazi Manchu vowel system. Some researchers have postulated up to eight vowel phonemes (Zhao & Chaoke), while others (Kim et al.) list only five. In the accessible material only the original five Manchu vowels a e i o u can be identified unambiguously (Table 17.1).

TABLE 17.1 SPOKEN MANCHU VOWELS u o

i e a

The vowels of the triangle u a i are pronounced with the standard values [u a i], while e has the mid central value [ə] and o is often realized as a diphthong [uo] or [ou]. The vowels a e o u, which are phonetically velar or central, have also the palatal allophones

Spoken Manchu 467

[ε e ø y], respectively, which most often occur in the vicinity of i and alternate with palatal diphthongs of varying types, e.g. dulin ‘half’ (Written Manchu dulin) = [tylin] ~ [tuilin] ~ [tielin]. Vowel length is not phonologically distinctive, with the exception of several cases of secondary long (double) vowels caused by apparently sporadic intervocalic consonant elision, cf. e.g. taa- (tafa-) ‘to climb’ vs. ta- (tuwa-) ‘to look’. An initial i can be accompanied by a prothetic glide [j], as in Mandarin. Although apparently non-distinctive, this glide can in some cases represent an etymological segment, e.g. iƞ [jiŋ] ‘bird-cherry’ (yengge). Sanjiazi Manchu has also some seven diphthongs or vowel sequences, which can be divided into the closing (ai ei), opening (ia ie io ua), or ambivalent (ui) types, e.g. aisin ‘gold’ (aisin), eixen ‘husband’ (eigen), kia ‘draught’ (hiya), tie- ‘to sit’ (te-), gio ‘roe deer’ (gio), juan ‘ten’ (juwan), tuix ‘cloud’ (tugi). The high vowels i u in the diphthongs could possibly also be analysed as manifestations of the glides y w, or some of the diphthongs could be analysed as containing an intervocalic glide (as proposed by Kim et al.). However, the presence of vowel sequences with two identical components favours the analysis of the diphthongs also as sequences of two independent vowel segments. The other varieties of “Bannermen Manchu”, as well as Sibe, show some deviations from these patterns. For instance, the counterpart of Sanjiazi o is in the other varieties often represented as u, e.g. Aihui udun ‘wind’= Sibe udun vs. Sanjiazi odun (edun), Aihui yu- ‘to go’ = Sanjiazi yo- (yo-). The data from Lalin Manchu, which in many respects preserves characteristics reminiscent of Written Manchu at all levels, show a mixture of retentions and innovations. While the syllable structure and vowel qualities are often preserved intact in Lalin, there are also peculiar developments, perhaps results of the influence of Peking Mandarin, such as the triphthong iue, e.g. giueraƞgi ‘bone’ vs. Sanjiazi giliaƞ (giranggi) (Zhao & Chaoke). Such data should not be taken at face value, as they may be due to a confusion between phonetic realizations and phonemic segmentation. The consonant system has 18 phonemes, divided into the nasals m n ƞ, the weak stops b d j g, the strong stops p t c k, the fricatives f s š x, the glides w y, and the liquid l (Table 17.2). In terms of the place of articulation, the system comprises the labials m b p f w, the dentals (alveolars) n d t s l, the palatals j c š y, and the velars ƞ g k x. Among the labials, the fricative f is pronounced as a dentilabial (labiodental) [f], as in Mandarin, while the glide w is basically bilabial but can also be realized as a dentilabial [v] in initial position. A difference with regard to Sibe is that there are no distinctive postvelar (uvular) consonants, though the velars have postvelar allophones in the vicinity of the vowels a o. Another important feature is that there is only one liquid phoneme, here written l, which is normally realized as a lateral [l], but which can allophonically also be realized as a flap [ɾ] in intervocalic position and as an approximant [ɹ] in syllable-final position.

TABLE 17.2 SPOKEN MANCHU CONSONANTS [ɲ]

ƞ

d

j

g

t

c

k

š

x

m

n

b p f

s

w

y l

468 Veronika Zikmundová and Gao Wa

As in the other Tungusic languages of the Manchurian sphere, the distinction between the two sets of stops is based primarily on aspiration (in the strong set), but in medial position also on voice (in the weak set). The velar stop k can in intervocalic position be spirantized to [x χ], e.g. jake [tʂaχə] ‘thing’ (jaka), without, however, losing its distinctivity against x, which, like the other fricatives f s š is voiced in sonorant environment, e.g. indaxo [jindaʁo] ‘dog’ (indahun), jafe- [tʂavə-] ‘to catch’ (jafa-), tusa [thuza] ‘help’ (tusa), kušo [khuʐo] ‘knife’ (huwesi). In the palatal column, the stops j c are realized as affricates with a sibilant release, varying, like the palatal sibilant š, between the palato-alveolar or retroflex (before the vowels a e o u) and alveolo-palatal (before i) values. The nasal n has the palatal allophone [ɲ] in the position before i (including the sequences ia ie io). In the Aihui variety initial s is alternatively realized as [ts], which is otherwise one of the typical features of Alchuka and Bala, e.g. Alchuka [tsama] ‘shaman’ vs. Sanjiazi samen (saman). A prominent feature of Alchuka and Bala, distinguishing them from both Sanjiazi and Written Manchu, is the apparent preservation of *d *t in several words before the vowel *i, e.g. di- ‘to come’ vs. Sanjiazi ji- (ji-), tixe ~ tixo ‘rooster’ vs. Sanjiazi coko (coko). This feature is occasionally also observed before other vowels, e.g. dusiba ‘firefly’ (juciba). Another possible archaism is the presence of *m before *i in a few items where other forms of Manchu have n, e.g. Bala misixa ‘little fish’ (nisiha). In some cases, Alchuka and Bala would seem to preserve a word-initial p against the expected f, e.g. Alchuka piniege = Bala pinierge ‘hair’ vs. Sanjiazi fenixe (funiehe). There are also items that would suggest a dual representation of the original word-initial velar nasal *ƞ, which in Manchu normally yields g, but which in Bala can also be lost, e.g. ele- ‘to be afraid’ vs. Sanjiazi gele- (gele-) < Proto-Tungusic *ƞeele-. In other items, Alchuka and Bala would seem to preserve the original ƞ, e.g. Bala ƞala = Alchuka ƞal ‘hand’ = Sanjiazi gale (gala) < *ƞaala. The most intriguing feature of Alchuka is the presence of an initial k in a number of words which in other forms of Manchu have no initial consonant, e.g. Alchuka karpuƞ ‘form’ (arbun). In some words, this initial k would seem to correspond to Proto-Tungusic *x, e.g. kai ‘what?’ vs. Sanjiazi ai (ai) < *xaï, while in others it corresponds to original zero, e.g. ker ‘this’ vs. Sanjiazi ele (ere) < *eri. Most curiously, however, it also appears in a few rather recent Mongolian loanwords, e.g. katali ‘like’ (adali) ← Mongolian adali, kerdem ‘virtue’ (erdemu) ← Mongolian erdem (ultimately from Turkic). The inevitable conclusion is that it is a secondary segment, added sporadically to some items. The possibility of transcriptional errors in the relatively unreliable corpus cannot be ruled out, either. PHONOTACTICS AND MORPHOPHONOLOGY The maximal syllable structure in Spoken Manchu is CVC. There are no initial or final clusters, and the typical coda consonants are the sonorants m n ƞ l. The dental nasal n in final position represents the original Proto-Tungusic primary *n (*/n), which is morphologically unstable and can be absent before suffixes. In pronunciation it can be realized as the nasalization of the preceding vowel, but in the position before other nasals and stops it assumes the place of articulation of the latter. In many etymons, an original final *n has simply been lost, e.g. nima ‘goat’ (niman). By contrast, the velar nasal ƞ in final position represents the original sequence *ƞi < *ƞgi (< *ƞI), e.g. iniƞ ‘day’ (inenggi). The labial nasal m in final position also implies the loss of a final vowel, as in the suffix CV.CONN -m, e.g. tate- ‘to pull’ : CV.CONN tate-m (tatame).

Spoken Manchu 469

The loss of final vowels is connected with the general process of vowel reduction, which can affect all vowels in non-initial syllables, yielding the uniform neutralized value e [ə], e.g. amele ‘behind’ (amala), gisele- ‘to speak’ (gisure-). Occasionally reduced vowels lose their syllabicity, e.g. acenaxo [aʂənaxo] ‘wrong’ (acanarakū), which may be seen as a precondition for their ultimate loss. The reduction and loss of vowels in Spoken Manchu is, however, not a fully regular process, being conditioned by a complex combination of phonological, lexical, and individual factors, many of which are not well understood. Some words show a morphophonological alternation between forms with and without a final vowel, e.g. emu ‘one’ (emu) : em dulin ‘one half’, etc. As a special case, final i and u in suffixes are often not reduced, but, instead, extended with an offglide, e.g. je- ‘to eat’ : PRS je-mi [tʂəmie] : EXT-PTCP.AOR-NEG je-ce-l-ku [tʂətʂəlkuo]. Combinatory changes that have affected the qualitative status of vowels in Sanjiazi include palatalization, labialization, and delabialization, all of which have taken place with some regularity. Palatalization (palatal umlaut) involves the regressive palatalizing effect of i on the quality of the vowel of the preceding syllable, a phenomenon also attested in the adjacent varieties of Mongolian. The palatalized vowels often alternate with diphthongs, e.g. tali- [thεli-] ~ [thaili-] ‘to cultivate’ (tari-), but basically palatalization remains a phonetic, rather than a phonemic, process. Labialization, by contrast, is a phonemic phenomenon, which involves the rounding of unrounded vowels under the regressive or progressive impact of rounded vowels in adjacent syllables, e.g. ose [ouzə] ‘net’ (asu), otu- [otu-] ~ [outu-] ‘to wear’ (etu-), muko ‘water’ (muke), owude ‘here’ (ubade). Delabialization concerns the loss of roundedness of u after labial consonants, e.g. medan ‘time’ (mudan). Another vowel change, though less regular, is palatal breaking, e.g. niumkuƞ ~ niuƞkuƞ ‘sickness’ (nimeku), biagan ~ biaƞgem ‘wild’ (bigan). Compared with Written Manchu, Spoken Manchu is characterized by various types of simplification. For instance, the Written Manchu “diphthong” ua (uwa) is often represented as a monophthong, e.g. taki- ‘to guard’ (tuwakia-), while the corresponding “triphthong” uai (uwai) is represented as a diphthong, e.g. xaiti- ‘to bind’ (hūwaita-). Medial consonants tend to be weakened and can yield unexpected results, e.g. sixden ~ siƞe ‘between’ (siden). Assimilation within clusters is frequent, e.g. alš [aɹʃ] ‘how?’ (absi), inna- ‘to reach’ < *isna- (isina-). The dropping of entire syllables occurs especially at morpheme boundaries, e.g. toogde [tho:ʁdə] ~ taode [thaodə] ‘in the village’ (tokso de). Much of the random variation in Spoken Manchu must be connected with the ongoing deterioration of the language skills of the last speakers. There is also some interaction between the spoken language and the written norm. Although the speakers deny knowledge of the Manchu script, they often distinguish between two manners of pronunciation, one which is rather close to the written language, and another, less careful, which involves extensive consonant weakening and elision, often causing merger of two syllables into one, e.g. DIR bo-wal ~ bo-wul > bool ‘to(wards) the house’ (boo baru). There is no productive vowel harmony in Spoken Manchu, and the distribution of the vowels in syllables is unpredictable, except for the tendency to reduce all vowels in non-initial syllables, merging them with e. Some suffixes retain, however, sporadic traces of earlier harmonic variants. The coherence of the phonological word is also to some extent indicated by a phonologically non-distinctive prosodic prominence or stress, which in polysyllabic words normally falls on the second syllable. The stress is accompanied by a high pitch, which extends from the second syllable to the end of word, with the initial syllable being pronounced with a lower pitch.

470 Veronika Zikmundová and Gao Wa

WORD FORMATION Spoken Manchu retains the morphological difference between nominals and verb(al)s, but a certain confusion is suggested by some examples in which nominals would seem to take suffixes that are basically used on verbs, e.g. saken ‘good, beautiful’ (saikan) : CV.CONN sake-me ‘well, properly’, kia ‘dry, drought’ (hiya) : PTCP.PRF-INFIN kia-xeniƞe ‘(the way how) dry (it was)’. In such cases, it may be assumed that the nominal root is first verbalized to make it capable of taking verbal suffixes. The difference between the principal parts of speech is also observed in derivation. It is, however, difficult to establish the current degree of productivity of Sanjiazi derivational suffixes, as derivation is not widely used, and many derived forms may be considered to have been lexicalized. Even so, there are a few productive and semi-productive suffixes of both the denominal and the deverbal type. The suffix -niƞe, identical with the secondary possessive marker -ni-ƞe of nominal declension, produces, with a rather vague concretizing function, nouns from other nouns, e.g. alwen ‘(public) service’ (alban) : DER alwen-niƞe ‘subordinate’. It can also be attached to participles, transferring them to full nouns (infinitives), e.g. taci- ‘to learn’ (taci-) : PTCP.PRF.INFIN taci-xe-niƞe ‘the things (one) has learned’. Another denominal suffix with a diffuse adjectivalizing function is -s-xun, e.g. yamji ‘evening’ (yamji) : yamji-sxun ‘late, (of the) evening’ (yamjishūn). For deverbal nominals the suffix -ko ~ -ku is attested, e.g. otu- ‘to wear (clothes)’ (etu-) : otu-ku ‘clothes’ (etuku). The suffix -la ~ -le- functions as the principal verbalizer of nominals, e.g. tusa ‘help’ (tusa) : VBLZ tusa-le- ‘to help’ (tusala-). It also makes verbs from Chinese loanwords, e.g. mafan-le- ‘to cause trouble’ ← Chinese mafan 麻煩 ‘trouble’. Another verbalizing suffix is -ni-, which derives translative verbs from nominal bases ending in the nasal n, e.g. san ‘good’ (sain) : TRANSL san-ni- ‘to become good, to improve’. Among deverbal verbs, the andatives (“translocatives”) in -na- and venitives (“cislocatives”) in -n-ji-, which have been lost in Sibe, are productive in Sanjiazi, e.g. AND in-na- ‘to arrive (there)’ (isina-) vs. VEN in-ji- ‘to arrive (here)’ (isinji-), based on the secondarily abstracted root in- (isin-). The causative-passive suffix -bu- is also frequent, e.g. ili- ‘to stand up’ (ili-) : CAUS ili-bu- ‘to stop (a cart)’ (ilibu-). Another voice suffix is -na- for reciprocals, attested mainly in lexicalized examples like tal-na- ‘to fight (each other)’ (tantanu-). Some verbs show in certain forms irregular extensions with no synchronic function, e.g. je- ‘to eat’ (je-) : EXT je-ce- id. (jete-). Words can be joined into compounds and fixed phrases. In these cases, the first member of the sequence is often truncated, e.g. manju+niama > man-niama ‘Manchu person’ (manju niyalma), bitke sewe > bit-sewe ‘teacher’ (bithe sefu). NUMBER AND CASE Nouns are currently rarely marked for number in Sanjiazi Manchu. There is only one plural suffix -sa ~ -se, which occurs only in several lexicalized forms, such as niam-se ‘people’ (niyalma-sa), ju-se ‘children’ (ju-se). Plurality is generally expressed by analytical or lexical means, such as numerals, other quantifiers, or suppletive forms, e.g. gulun ~ niam-gulun ‘people’ (niyalma gurun). In addition to the unmarked nominative, Sanjiazi Manchu has four primary suffixally marked cases: the accusative, genitive, dative (-locative), and ablative (-instrumental), as well as three secondary cases: the possessive, directive and the comitative. The case markers are identical for stems ending in a vowel (V) or a consonant (C), but original nasal stems (n) take a separate allormoph in the accusative (Table 17.3).

Spoken Manchu 471 TABLE 17.3 SPOKEN MANCHU CASE MARKERS

• •

•

•

•

•

VC

n

ACC

-we

m-be ~ -me

GEN

-i

POSS

-ni-ƞ/e

DAT

-d/e

ABL

-deli ~ -dele- ~ -dile ~ dili

DIR

-wal/e ~ -wol/e

COM

-maƞ/e

The unmarked nominative is basically the case of the subject, adnominal attribute, and the nominal predicate, but it can also replace the accusative as the case of the object, especially an indefinite object. The accusative marks the direct definite object, e.g. sawe-be otu-xe ‘(I) put on (my) shoes’ (shoe-ACC wear-PST), though its use is not obligatory. In some constructions it may be understood as indicating the source, e.g. mini soku-be sawe ale ‘make shoes of my skin!’ (or: ‘make my skin to shoes!’) (1SG.ADN skin-ACC shoe make.IMP). The genitive marker -i is not frequent. It mostly marks the nominal attribute in fixed expressions in storytelling, e.g. yilxa-i bila ‘the Flower River’ (flower-GEN river). In regular genitival functions, including adnominal usage, it is replaced by the complex possessive form in -ni-ƞ/e, e.g. sewu-niƞ bo ‘the teacher’s house’ (teacher-POSS house), yilxa-niƞe siliƞ ‘flower dew’ ~ ‘(the) dew of (the) flowers’ (flower-POSS dew). The dative marks static location, e.g. tale-de emke gio sawe-xe ‘on the plain (I) spotted one roe deer’ (plain-DAT one.EMPH roe.deer see-PST), direction, e.g. še-de weile-na-xe ‘(I) went to the town for work’ (town-DAT work-AND-PST), temporal specification, e.g. ice sunja-de ‘on the fifth day after the new moon’ (new five-DAT), and indirect object, e.g. bi in-de bu-le-ko ‘I will not give (it) to him’ (1SG 3SG-DAT give-PTCP.AOR-NEG). Occasionally, it can also indicate source, e.g. nialma-de san ba-xe ‘(he) received help from a man’ (person-DAT good receive-PST), or means, e.g. so ai alexe-de nimaxe tanna-mi ‘by what method do you catch fish?’ (2PL what method-DAT fish catch-PRS). The ablative indicates the point of departure, e.g. ailin-deli waili-nji-xe ‘(they) came back from the mountains’ (mountain-ABL return-VEN-PST), the point of reference or comparison, e.g. owu-dili nadin ba golo ‘seven miles from here’ (here-ABL seven place far), the route of movement, e.g. ai juxun-dile ji-xe-niƞe ‘by which way did you come?’ (what way-ABL come-​PTCP.PRF-INFIN), or the starting point in time, e.g. ajigin-dele ‘from childhood’ (young-ABL). In some uses, it marks the instrumental function, e.g. gale-deli fete-xe ‘(he) dug it with (his own) hands’ (hand-ABL dig-PST). The directive in -wal/e ~ -wol/e is based on the postposition bali ‘towards’ (baru) and indicates the direction of movement and the addressee of verbal communication. The suffix variant -wol/e appears after stems containing o, while otherwise the variant -wal/e is used, e.g. bo-wole yawe-xe ‘(he) left for home’ (house-DIR leavePST), xexe-wale gisel-mi ‘(he) says to (his) wife’ (wife-DIR say-PRS). There is also another directive in Spoken Manchu, marked by the suffix -ci, as present in the

472 Veronika Zikmundová and Gao Wa

•

regular case paradigm in Sibe. Based on the suffixalized noun *ici ‘direction’ (ici), this form is in Sanjiazi Manchu attested only in a few pronominal adverbs, notably e-ci ‘hither’ (< *e+ici) and ye-ci ‘whither?’ (ya ici). The comitative in -maƞ/e, which has a cognate in Sibe -maqe, is based on the postposition maƞƞa ~ maƞƞe ‘together with’, but is probably better understood as a case form. As its cognate in Sibe, it is itself often combined with the postposition embade (< eme ‘one’ + DAT ba-de ‘place’), e.g. xexe-maƞ embade ‘together with (one’s) wife’.

As compared with Sanjiazi, the Alchuka and Bala varieties have in general a richer morphology. For instance, Alchuka has four plural markers: -tsa ~ -tse (< *-sA), -ta ~ -te (< *-tA), -si, and -ri. The case paradigm of Alchuka and Bala differs from Sanjiazi mainly by the larger number of allomorphs of the case markers. Depending on whether the stem ends in a vowel, a non-nasal consonant, or a nasal, the accusative marker appears as -we ~ -be ~ -me, the genitive marker as -i ~ -gi ~ -ni, and the dative marker as -de ~ -te ~ -ne. For the ablative function, the primary suffix -ti ~ -ci ~ -ji is used. NUMERALS The Sanjiazi numerals are etymologically identical with those of Written Manchu. The basic numerals of the first decade are: 1 eme ~ em, 2 juo ~ ju, 3 ilan, 4 duin, 5 sunja, 6 niƞun ~ niuƞun, 7 naden, 8 jakun, 9 uyin, 10 juan. The numeral 1 eme has the special emphatic form em-ke. The teens are formed by combining 10 juan with the basic numerals, with occasional assimilations at the boundary: 11 jom-mo, 12 jon-jo, etc. An exception is 15 tokon [toxon] (tofohon). The decades are expressed by separate lexical roots for the lower decades (from 20 to 50) or by combining the basic numerals with a trace of the numeral 10 juan for the higher decades (from 50 to 90), yielding the series 20 olin, 30 gosin, 40 dexe ~ dixe, 50 susai ~ sunja-je, 60 niƞun-jo, 70 naden-jo, 80 jakun-jo, 90 uyin-jo. For the powers of ten the numerals 100 (em) taƞ, 1000 em) miƞa, and 10,000 (em) tumen are used. The suffix for ordinal numerals is -ci, which in the modern language seems to be used only to identify siblings, e.g. ila-ci gexe ‘third elder sister’ (three-ORD elder. sister). For counting, predominantly with digits except 1, combinations of the numeral roots with the general counter -wali (fali) ‘piece’ are used, often added to a truncated stem, e.g. ji-wali ‘two (pieces)’, ila-wali ~ ye-wali ‘three (pieces)’, nad-wali ‘seven (pieces)’. The number of ‘times’ is expressed by the noun medan (mudan); note em+medan > em-dan ‘once’. For the concept of ‘(one) half’ the word (em) dulin ‘middle, center’ > ‘half’ is used. The Alchuka variety deviates from Sanjiazi on several points. The ordinal suffix has the shape -ti (< *-tI/n), multiplicatives are formed by the suffix -meri, and for counting without a head noun -ke (apparently from Chinese ge 個) is used. Most importantly, Alchuka has preserved the whole set of the Jurchenic numerals for the teens 11–19. They have been recorded in two variants, one of which represents the more original full shape, as known from Jurchen, while the other is a reduced colloquial shape that was used in actual speech: 11 enso > ensi, 12 tirgon > tio, 13 koxo > koo, 14 tuxu > tuu, 15 tofxo > too, 16 niulxu > niuu, 17 toxoƞ > toƞ, 18 niokon > nion, 19 kuniku > uniu (here rendered in a transcription modified from Mu Yejun; note that initial in these data stands for both /t/ and /d/).

Spoken Manchu 473

PRONOUNS Sanjiazi Manchu preserves the full set of Manchu personal pronouns, including, importantly, the third person items, which have been lost in Sibe. The basic pronouns are 1SG bi : 2SG si : 3SG i ~ in : 1PL.EXCL bo : 2PL so : 3PL ce, with the oblique stems 1SG min- : 2SG sin- : 3SG in- : 1PL.EXCL mon- : 3PL son- : 3PL cen- (Table 17.4). The main difference with regard to Written Manchu is the vocalism of the plural pronouns 1PL. EXCL bo (be) : 2PL so (suwe), a feature also shared by Sibe. TABLE 17.4 SPOKEN MANCHU PERSONAL PRONOUNS   SG

PL

NOM

OBL

ACC

GEN

POSS

1

bi

min-

mim-be

min-i

mini-ƞ/e

2

si

sin-

sim-be

sin-i

sini-ƞ/e

3

i/n

in-

im-be

in-i

1

EXCL

bo

mon-

mom-be

mon-i

moni-ƞ/e

2

so

son-

som-be

son-i

soni-ƞ/e

3

ce

cen-

cem-be

The personal pronouns are inflected according to the regular nominal pattern, e.g. 1SG-ABL min-deli, 3SG-COM im-maƞe, though complete paradigms are unavailable. The first person plural inclusive pronoun is mese (muse), which may be formally understood as a plural form in -se. Postnominal enclitic possessive markers are not attested in the available Sanjiazi Manchu material, though they are present in Sibe. To indicate possession, the adnominal possessive forms, which may also be analysed as genitives in -i, are used, e.g. mini ama ‘my father’, soni toxso ‘your village’. The system also comprises the possessive forms in -ƞ/e, which are used in predicative position, e.g. ele bitke mini-ƞe ‘this book is mine’ (this book 1SG-POSS), but which are also attested adnominally, e.g. mini-ƞe indaxu ‘my dog’ (1SG-POSS dog). In a construction with two nouns, possession is expressed only by word order, with the head noun remaining unmarked even when no possessor marking is present on the modifier, e.g. gio betke ‘the roe deer’s leg’ (roe.deer leg). The demonstrative pronouns are ele ‘this’ : PL e-se (proximal) vs. tele ‘that’ : PL te-se (distal), also used in the extended forms ele-m ‘this one’ vs. tele-m ‘that one’ (with -m < +eme ‘one’). The distal demonstrative tele : PL tese is often used to replace the actual third person pronouns, especially in the plural. In the possessive construction, instead of the genitive form tese-i, the basic form tese is used, e.g. te-se bo ‘their house’ (that-PL house). Lexicalized derivatives of the demonstrative pronouns include e-sexe ‘in this manner, to this extent’ vs. te-sexe ‘in that manner, to that extent’ and o-wude ‘here’ (ubade < *e+ba-de ‘at this place’) vs. to-wude ‘there’ (tubade < *te+ba-de ‘at that place’). The function of a reflexive pronoun is filled by the noun beyi ‘body’ (beye), e.g. beyi jielin gene-xe ‘(he) went for the sake of himself’ (REFL because.of go-PST). It is also used emphatically, e.g. beyi fenci-xe ‘he was left all alone’ (REFL remain-PST). The basic interrogative pronouns are ai ‘what?’, we ‘who?’, and ya ‘which?’. Other interrogative forms and expressions are ai-tiƞ ‘when?’ (atanggi), ai-se(xe) ‘to what extent?’ : ai-sexe lawdo ‘how much?’, ya-wude ‘where?’ (yabade < *ya+ba-de ‘at what

474 Veronika Zikmundová and Gao Wa

place?’), udu ~ ude ‘how many?’ (udu), ane ‘why?’. The pro-verb ai-ke- ‘to do what?’ is mainly used as a filler word and an echo expression, e.g. buda aike-m ale-m ‘she makes food and so on’ (food do.what-CV.CONN make-CV.CONN). A lexicalized pro-verbal form is adeleme = CV.CONN adele-me ‘how? why?’ (adarame). VERBAL FORMS The system of verbal inflection in Sanjiazi Manchu is highly reduced as compared with Written Manchu, not to mention the other Tungusic languages. Even the other forms of Spoken Manchu, including not only Sibe, but also Alchuka and Bala, show a somewhat more diversified verbal morphology. However, the formal and functional differences between the standard categories of imperatives, finite indicative forms, participles, and converbs can still be distinguished also in Sanjiazi Manchu. In addition, there is a new secondary category, which will here be termed “infinitives”. •

•

•

The imperatives in Sanjiazi distinguish person, but not number. The plain verbal stem is used as the basic second person command form. Some monosyllabic stems have irregular imperative forms, of which the following are attested: ji- ‘to come’ : IMP jo (jio), je- ‘to eat’ : IMP jewe ~ jewu (jefu), bai- ‘to seek’ : IMP baise (baisu), bi- ‘to be’ : IMP bile (*bire vs. bisu). In the case of bi- ‘to be’, the plain form bi is also attested, e.g. si owude bi ‘be here!’ (2SG here be.IMP). For the “voluntative” function of the first person the suffix -ki is used, e.g. sim-me cumaxe siƞe ali-ki ‘let me wait for you until tomorrow!’ (2SG-ACC tomorrow until wait-VOL), while the “permissive” function of the third person is expressed by the suffix -kene, e.g. xaxeji yowu-kene ‘may the boy go!’ (boy go-PERM). The suffix -kene = -ke-ne contains the probably only trace preserved in Sanjiazi Manchu of the otherwise lost third person possessive suffix *-nI. The system of converbs is also very restricted, comprising only two forms: the connective converb in -me ~ -m and the conditional converb in -ci. Rarely, the consecutive converb in -wei (-fi) is also attested in lexicalized expressions, e.g. yamji-wei ‘after the night came’ (become.night-CV.CSEC). The consecutive converb is better attested in Alchuka, where it has the form -pi ~ -wi, while the conditional converb has the form -ti ~ -ci ~ -ji. Alchuka also has the terminative converb in -tala ~ -tele ~ -tolo. The participles include two forms: the aorist (imperfective) participle in -le ~ -l and the perfective participle in -xe ~ -xa. After stems with the vowel o, the variants -lo and -xo are used, respectively. Several monosyllabic stems require a synchronically unmotivated extension before the aorist participle marker, e.g. je- ‘to eat’ : EXT-PTCP.AOR je-ce-le (jetere), ji- ‘to come’ : EXT-PTCP.AOR ji-si-le (*jisire vs. jidere), bi- ‘to be’ : EXT-PTCP.AOR bi-se-le ~ bi-si-le (bisire). Also, some monosyllabic stems require the variant -ke of the perfective participle suffix, e.g. je- ‘to eat’ : PTCP.PRF je-ke.

The aorist and perfective participles are both polyfunctional forms, but with a slightly different distribution of functions. Both are used as attributes, e.g. je-ce-le jaxe ‘things to eat, food’ (eat-EXT-PTCP.AOR thing), dule-xe ani ‘last year’ (pass-PTCP. PRF year). Both can also take case endings and function as independent heads of subordinate clauses. However, only the perfective participle is used as a finite predicate, in

Spoken Manchu 475

which position it indicates past tense, e.g. elwixe suji-xe ‘the sable ran away’ (sable run. away-PST). Participles play also an important role in the formation of negated forms of verbs. Both participles are negated by the negative existential noun aku ~ ako ‘absence, absent’, which is normally suffixalized to the participle in the form -ku ~ -ko, phonetically also [-xu] ~ [-xo], e.g. sa- ‘to know’ : PTCP.AOR-NEG sa-le-ku ‘not knowing, (one who) does not know’, bu- ‘to give’ : PTCP.PRF-NEG bu-xe-ku ‘having not given, (one who) has not given’. With the negation marker, the aorist participle can also be used as a finite predicate, e.g. axe axe-le-ku ‘it is not raining’ (rain rain-PTCP.AOR-NEG). Converbs are negated by adding to the negative participles the converbal forms of the auxiliary verb o- ‘to become’, in which process both the participle marker and the auxiliary stem can disappear, yielding new negative converb markers like -ku=ci, e.g. inna-ku=ci = inna-Øku-Ø=ci ‘if (one) does not reach’ (reach-PTCP.AOR-NEG-AUX=CV.COND). • •

Both participles can take the suffix -ƞe ~ -ni-ƞe, transforming them to full action nouns or “infinitives”, which can function as nominal constituents in the sentence, e.g. suji- ‘to run’ : PTCP.AOR-INFIN suji-le-niƞe ‘(the act of) running’. There are no primary finite indicative forms in Spoken Manchu, for all the forms used in the function of finite predicates are based on either the two participles or the connective converb, forming an integrated system that also comprises the infinitives (Table 17.5). A similar system, though a morphologically slightly more elaborate one, is attested in Sibe. TABLE 17.5 SPOKEN MANCHU VERBAL FORMS  

PTCP

NEG

PRS

CV

FIN

-m/e

-m-i ~ -m/e ~ -mie

INFIN

AOR

-le

-le-ku

-le(-ni)-ƞe

PRF/PST

-xe

-xe-ku

-xe(-ni)-ƞe

While the participles serve for the expression of the affirmative finite past tense as well as the negative present and past tenses, the affirmative present tense is expressed by the finitized connective converb with the suffix -mi = -m-i (-mbi), containing the auxiliary element -i (< *+bi ‘to be’). This suffix can be both strengthened to -mie and reduced to -me ~ -m, in which case it formally merges with the actual converb suffix -m/e. The choice between the suffix variants is based on stylistic and/or phonotactic motivations. The temporal range of the present (non-past) tense comprises both present and future references, e.g. ani-deli geme tieli-m/e ‘we till (it) every year’ (year-ABL all cultivate-PRS), bi cumaxe nimaxe buta-na-mi ‘tomorrow I will go fishing’ (1SG tomorrow fish hunt-AND-PRS). COMPLEX PREDICATES Several temporal, aspectual, modal, and evidential functions which cannot be expressed by the limited paradigm of the finitely used monoverbal forms can be expressed by complex predicates, which are typically formed with the help of the auxiliary verbs bi- ‘to

476 Veronika Zikmundová and Gao Wa

be’ : PRS bi and o- ‘to become’. The semantic main verb in these constructions stands in either a participial or a converbal form. •

•

• •

•

Although the combination of the connective converb with the auxiliary bi- has been grammaticalized into the synchronically indivisible finite present tense marker -mi (-mbi), the combination -me+bi- (-me bi) survives synchronically and expresses durativity, e.g. bi gele amge-me bi ‘I was still sleeping’ (1SG still sleep-CV.CONN be). The combination of the perfective participle with bi yields a perfect “tense”, implying a completed action whose consequences continue in the present, e.g. in tiatia xaiti-xe bi ‘he is tied up very tightly’ (3SG tightly bind-PTCP.PRF be.PRS). The corresponding pluperfect is formed by using the past tense form bi-xe of the auxiliary, e.g. ame juse ju niama o-xo bi-xe ‘they had become only two persons: father and son’ (father child two person become-PTCP.​PRF be-PST). The form bi-xe as a part of complex predicates may also add a mirative modality, e.g. si amba nimaxe ba-xe bi-xe ‘oh, you have caught a really big fish!’ (2SG big fish find-PTCP.PRF be-PST). The auxiliary verb o- indicates a change of state, e.g. si bo-wole yo-me mute-xe o-xo ‘(the time has come that) you will be able to go home’ (2SG house-DIR go-CV. CONN be.able-PTCP.PRF become-PST). Other complex predicates consisting of two or more verbs include combinations of full verbs, constructions with semantically bleached verbs, and reduplicated expressions. In the chaining of verbs, not only a converb, but also a participle may be followed by another verb, e.g. indaxo daxe-bu-xe suji-xe-niƞe ‘he was running, taking the dog with him’ (dog follow-CAUS-PTCP.PRF run-PTCP.PRF-INFIN). Combinations of full verbs mostly express manner or purpose of an action, e.g. sake-me ulu ‘feed (it) well!’ (do.well-CV.CONN feed-IMP), ale-na-me gene-xe ‘(he) went to tell’ (tell-AND-CV.CONN go-PST). Semantically bleached verbs are used to aspectually modify the meaning of the main verb, which is mostly in a converbal form, e.g. je-me waji-xe ‘(he) finished eating’ (eat-CV.CONN finish-PST). Reduplication is employed to add certain temporal or aspectual meanings, e.g. goina-me goina-xe ‘(he) thought for a while’ (think-CV.CONN think-PST). A special type of complex predicate is present in the prohibitive construction, which involves the combination of the prohibitive particle eme (ume) ‘do not!’ with the aorist participle of the main verb. The aorist participle fills in this construction the function of a connegative form, e.g. si eme gele-le ‘don’t be afraid!’ (2SG PROHIB be.afraid-CONNEG). The prohibitive particle, though synchronically indivisible, is still transparently identifiable as the connective converb of the original negation verb (negative auxiliary) *e-, which is otherwise lost in Sanjiazi Manchu.

SENTENCE TYPES In spite of being rather heavily influenced by Chinese, Sanjiazi Manchu retains its basic syntactic orientation as a verb-final (SOV) language. The normal order of constituents in the sentence is: temporal determination—subject—spatial determination—indirect object—direct object—predicate. In a noun phrase modifiers mostly precede the head noun. Temporal and spatial determinations are often expressed with petrified case forms of spatial nouns, e.g. amele ‘behind, after’ (amala), nuƞude ‘above’ (ninggude), used postpositionally after nominals and nominalized verbs. Other features of Sanjiazi Manchu syntax include the following (with some of the examples adapted from Kim et al.):

Spoken Manchu 477

•

•

•

•

Topic marking is not obligatory but is nevertheless often present. Topic markers include the particles niƞe and ne (the latter possibly influenced by Chinese ne 呢), e.g. eixe niƞe mo sace-mi ‘the husband was cutting wood’ (man TOP wood cut-PRS), ele eme bo ne utxa ila aƞala niame bi ‘in this house there were three people’ (this one house TOP thus three mouth person be.PRS). The converbal form CV.COND o-ci ‘if it becomes’ can also be used in a function reminiscent of a topic marker, e.g. tele oci we ji-xe ‘then, who came?’ (that become-CV.COND who come-PST). In existential sentences and in the possessive construction the auxiliary bi- ‘to be’ : PRS bi : PST bi-xe is used, e.g. tele xaxaji bi ‘he has a son’ (that boy be.PRS), bi bo-de bi-xe ‘I was at home’ (1SG house-DAT be-PST). In the past tense, the form bi-xe is also used as a copula in equative sentences, but in the present tense, the Chinese copula še (shi是) is used, e.g. tel še sewe ‘he is a teacher’. In such examples, the Chinese copula could also be interpreted as a subject marker. Negation is expressed by the negative forms of verbs, e.g. bi sim-be je-ce-le-ko ‘I will not eat you’ (1SG 2SG-ACC eat-EXT-PTCP.AOR-NEG), or, in existential sentences, by the independent negative existential noun aku ~ ako (akū), e.g. em niam geme aku ‘there was not a single person’ (one person all NEG.EXIST). The negative copula is wake (waka), e.g. bi sajwe wake ‘I am not a student’ (1SG student NEG.COP). Interrogation in polar questions is expressed by the sentence-final interrogative particle no, e.g. so ani-deli gem sog tali-m no ‘do you grow vegetables every year?’ (2PL year-ABL all vegetables cultivate-PRS INTERR). Other frequently used particles include ino and eƞ [affirmation] and ba (Chinese ba 吧) [probability].

COMPLEX SENTENCES The typology of Sanjiazi complex sentences is remarkably influenced by Mandarin. Although there are structures shared with Written Manchu or Sibe, there are also, and perhaps more frequently, structures of the Mandarin type, such as subordinate clauses attached to the main clause by periphrastic or paratactic means. • •

•

The equivalents of relative clauses are expressed by participles, e.g. nimiƞ sekte-le jugun-dili waili-nji-xe ‘(they) came back by a road which was covered by snow’ (snow cover-PTCP.AOR road-ABL return-VEN-PST). Object clauses may end in a participle in the accusative case, e.g. sim-me yawule-be cia-ku ‘I do not want you to leave’ (2SG-ACC leave-PTCP.AOR-ACC willing-NEG)—note that the predicate in this clause is a nominal word (cihakū = ciha+aqu ‘without willingness’), although it takes an argument in the accusative. Object clauses may, however, also be embedded without case marking, e.g. xaaj yeci yawe-xe bi sa-le-ku ‘I do not know where the boy went’ (boy whither leave-PTCP. PRF 1SG know-PTCP.AOR-NEG). Temporal and conditional clauses are connected to the main clause with converbs and quasiconverbs. Quasiconverbs are mostly used in temporal subordinate clauses, the two most common constructions being based on the dative and genitive forms of the perfective participle, e.g. bo-wole waili-nji-xe-de tini jakun eilin ‘when I came home, it was only eight o’clock’ (house-DIR return-VEN-PTCP. PRF-DAT only eight time), suju-xa-i ele indaxo jowe-me bici-xe ‘having been running (for a long time) the dog died of exhaustion’ (run-PTCP.PRF-GEN this dog

478 Veronika Zikmundová and Gao Wa

suffer-CV.CONN die-PST). Converbs are used to connect both temporal and conditional subordinate clauses, e.g. jugun dulin inna-me eme iwagun sawe-xe ‘when he reached half-way, he saw a ghost’ (road middle reach-CV.CONN one ghost see-PST), bi bici-ci eme ji-le ‘if I die, don’t come!’ (1SG die-CV.COND PROHIB come-CONNEG). The converbal form waji-me (finish-CV.CONN) following the verb of a subordinate clause is often employed to specify that the action took place before that of the main verb, e.g. donji-me waji-me inji-me ‘having heard (it) (he) smiled’ (hear-CV.CONN finish-CV.CONN smile-PST). The immediate consequence of an action is expressed with a participle followed by the emphatic particle (=)li, as in [1]: [1] ji-xe come-ptcp.prf ‘As soon as (I) came,

li ptcl

elwixe ula&ula sila-me yawe-me sable one.by.one continue-cv.conn leave-cv.conn (I saw that) the sables were passing one after another’. •

•

bi-xe be-pst

In a periphrastic construction, the converbs are replaced by sequences of a participle and a full noun in a proper case form, e.g. yawe-le eilin-de bi sin-de ale-xe ‘I told you (at the time) when I was leaving’ (leave-PTCP.AOR time-DAT 1SG 2SG-DAT tell-PST). The “infinitive” forms in -niƞe are used as equivalents of the Mandarin constructions with the “particle” de (得) in its attributive-marking and nominalizing functions. A participle “infinitivized” with -niƞe usually ends an embedded clause with or without an overt subject, which anticipates an expression of the manner or extent of the main action, e.g. sin-i molin suji-le-niƞe sain xodun ‘your horse runs pretty quickly’ ~ ‘the running of your horse is pretty quick’ (2SG-GEN horse run-PTCP. AOR-INFIN good quick), kia-xe-niƞe ba na gume xuaji-xe ‘it was (so) dry that the soil all cracked’ (be.dry-PTCP.PRF-INFIN place land all crack-PST). As in the corresponding Mandarin construction, an “infinitivized” perfective participle can also be used as a finite predicate, denoting an already known action to which new information is added, e.g. bi alga aku bu-xe-niƞe ‘(the fact that) I gave (it to him) (was because) I had no other way’ (1SG method NEG.EXIST give-PTCP.PRF-INFIN). Further, in a construction with emdan ‘once’, a participle followed by -niƞe marks an action immediately preceding another, e.g. emdan ta-xe-niƞe ele xaaj aku ‘as they looked, the boy was not there’ (once look-PTCP.PRF-INFIN this boy NEG.EXIST). A subordinate clause ending with an adjective followed by -niƞe usually denotes the characteristics of some nominal constituent of the main sentence, thus being close to a relative clause, as in [2]:

[2] fennixe gem siaƞen-niƞe hair all white-infin ‘Having completely white hair, eme sakda niama in julgi-de one old person 3sg front-dat an old man appeared in front of him.’

cici-xe appear-pst

Spoken Manchu 479

•

Quotation is not embedded and is introduced by converbal forms of quotation verbs such as gisele- ‘to say’, fenji- ‘to ask’, goina- ‘to think’, as in [3]:

[3] xaxeji-de fenji-me, boy-dat ask-pst ‘(He) asked the boy: si in-de ai jaxe 2sg 3sg-dat what What did you give him?’ •

bu-xe thing

give-pst

In narration, paratactic sequences of simple finite clauses are used along with chains of converbal phrases, as in [4]:

[4] bi miocen tuki-xe muxali-we 1sg rifle lift-pst round-acc ‘I lifted the rifle, loaded the bullet,

gida-xe press-pst

gio-we aike-xe eme miocen sinda-xe roe.deer-acc do.what-pst one rifle set-pst the roe deer, well, I triggered the rifle once, eme miocen utxa gio-we tuƞ-na-xe one rifle thus roe-deer-acc fall-and-pst that is, (with) one rifle (shot) I killed the roe deer.’ Another way of expressing a quick sequence of events is to use the plain verbal stem, which, then, functions as a zero-marked converb, e.g. suji, molin senda-me waji . . . eme tale-de in-na-xe ‘(he) ran, set the horse free . . . and came to a plain’ (run horse set-CV. CONN finish . . . one plain-DAT reach-AND-PST). LEXICON AND LANGUAGE CONTACTS Spoken Manchu has for centuries evolved under the dual areal impact of Mongolian and Chinese. The Mongolian impact is particularly evident in some phonetic developments, especially the phenomena of palatal breaking and palatal umlaut. There is also a large number of lexical borrowings from Mongolian, but they are generally shared by Written Manchu, making it difficult to specify any direct influences restricted only to Spoken Manchu. The situation is somewhat similar with Chinese, but the widespread bilingualism in Chinese during the last few generations has increased the number of Chinese loanwords much beyond what is attested in the written language. The most conspicuous Chinese feature in Spoken Manchu is the merger of the liquids *r and *l, which must be relatively recent, since it is absent in Sibe. In fact, judging by recordings made in the late 20th century, the distinction was still retained by some of the last speakers of Manchu in the Aihui region. Apart from this detail, Chinese influence is observed in many other aspects of pronunciation, especially in the vowels, as well as in the syntax. Other contact languages of Spoken Manchu have been Udihe-Kyakala, Nanai-Hezhen, and Solon, as well as Daghur, but in these cases, Manchu has been in the role of the dominant partner. Within Spoken Manchu, the Alchuka, Bala, and Lalin varieties used to be geographically close and are likely to have influenced each other. Some specific Alchuka

480 Veronika Zikmundová and Gao Wa

features are shared by Bala, but it is difficult to determine, whether they are based on common heritage, shared innovations, or borrowing. Alchuka and Bala also share some lexical items that are absent in Written Manchu and Sanjiazi, and that seem to derive from older forms of Jurchenic, cf. e.g. Alchuka ansin = Bala antsugun ‘gold’ < Jurchen ancun vs. Written Manchu aisin = Sanjiazi aisin ‘gold’ (but note also Manchu ancun ‘earring’). Another old item preserved by Alchuka and Bala is dili ‘head’, the Common Tungusic counterpart of Written Manchu uju ‘head’ = Sanjiazi ujo (but note also Manchu jili ‘the base of the horn on deer’). Mongolic items present in Alchuka and Bala, but absent in Written Manchu and Sanjiazi, include nikuru ‘friend’ ← Mongolic *nekür (from which also Manchu neku ‘female friend’) and terin ‘head’ ← Mongolic *teriün. The impact of Written Manchu and “Standard Manchu” on the varieties of Spoken Manchu is rather difficult to estimate. Although the school system spread the knowledge of a uniform standard language, not all speakers were literate. Alchuka and Bala, in particular, remained long outside of the Manchu written culture. They did, however, receive indirect influences of the standard language via contacts with the Lalin variety. The last speakers of the Bala variety are reported (by Mu Yejun) to have used two registers of speech, one of which was an “elegant style” (yayu 雅語), close to the standard language, while the other was a casual style (suyu 素語), which had more dialectal features. In addition, Bala is reported to have had a set of “local words” (tuyu 土語) absent in Written Manchu. Unfortunately, the Alchuka and Bala varieties were never documented in a sufficiently reliable and detailed way to allow them to be fully incorporated into the context of Jurchen-Manchu dialectology and comparative Tungusic studies. REFERENCES AND FURTHER READING Aisin Gioro, Yingsheng 愛新覺羅瀛生 (1987–2004)「談談滿語的景語」[Discussing the Jing dialect of Manchu] 1–15,『滿語研究』[1–2] 1986 (2): 2–15, [3] 1989 (1): 4–20, [4] 1990 (2): 22–37, [5] 1991 (2): 3–16, [6] 1992 (2): 1–17, [7] 1994 (1): 15–36, [8–1] 1995 (1): 13–20, [8–2] 1996 (1): 1–12, [9] 1997 (1): 3–14, [10] 1998 (1): 1–10, [11] 1999 (1): 5–19, [12] 2000 (1): 1–17, [13] 2002 (1): 3–8, [14] 2003 (1): 5–14, [15] 2004 (1): 6–11, 哈爾濱 [Harbin]. Aisin Gioro, Yingsheng 愛新覺羅瀛生 (1994)『北京土語中的滿語』[Manchu vocabulary in the Peking dialect of Chinese], 北京 [Peking]: 燕山出版社. Ch’en, Chieh-hsien (1976) ‘The decline of the Manchu language in China during the Ch’ing period (1644–1911), in: Walther Heissig (ed.), Berichte und Vorträge der XVII. Permanent International Altaistic Conference (Bonn/Bad Honnef 1974), 137–154, Wiesbaden: Otto Harrassowitz. Enhebatu 恩和巴圖 (1995)『滿語口語研究』[Study of Spoken Manchu], 呼和浩特 [Huhehaote]: 内蒙古大學出版社. Gimm, Martin (1981), ‘Zur Mandschuristik in der Volksrepublik China 1980’, T’oung Pao, Second Series 67 (3/5): 269–287. Guo, Meng-xiu & Yin Tie-chao (2008) ‘Changing process and causes of the Manchu language in Heihe District’, Altai Hakpo『알타이학보』18: 11–30, Seoul. Hölzl, Andreas (2017) ‘New evidence on Para-Mongolic numerals’, Journal de la Société Finno-Ougrienne 96: 97–113, Helsinki. Hölzl, Andreas (2018) ‘Udi, Udihe, and the language(s) of the Kyakala’, International Journal of Diachronic Linguistics and Linguistic Reconstruction 15: 111–146.

Spoken Manchu 481

Hölzl, Andreas (2021) ‘The only known text from Bala, an extinct Tungusic language’, Studia Orientalia Electronica 9: 173–191, Helsinki. Hölzl, Andreas & Hölzl Yadi (2019) ‘A wedding ceremony of the Kyakala in China: Language and ritual’, International Journal of Diachronic Linguistics and Linguistic Reconstruction 16: 87–144. Ikegami Jirō 池上二良 (1993)「満州語方言研究における穆曄駿氏採集資料につ いて」[Mu Ye-jun’s materials viewed [in the context of] the study of Manchu dialects],『言語文化接触に関する研究』5: 1–24, 東京 [Tokyo]. Republished in Ikegami (1999). Ikegami, Jirō 池上二良 (1999)『満洲語研究』[Researches on the Manchu Language], 東京 [Tokyo]: 汲古書院. Ji, Yonghai 季永海 & Zhao Zhizhong 趙志忠 & Bai Liyuan 白立元 (1989)『現代滿語 八百句』, 北京, [Peking:] 北京中央民族學院出版社. Joseph, Andrew Jonathan (2018) The Historical Phonology of Manchu Dialects, PhD dissertation, Cornell University. Kane, Daniel (1997) ‘Language death and language revivalism: The case of Manchu’, Central Asiatic Journal 41 (2): 231–249. Kim, Juwon & Ko Dongho & D. O. Chaoke & Han Youfeng & Piao Lianyu & B. V. Boldyrev (2008) Materials of Spoken Manchu, Altaic Languages Series 01, Seoul: Seoul National University Press. Kobori, Iwao 小堀巌　(1949)「璦琿附近の滿洲族の言語について」[On the language of the Manchu in the vicinity of Aigun],『民族學研究』The Japanese Journal of Ethnology 14: 59–64, 東京 [Tokyo]. Kōno, Rokurō 河野六郎 (1944)「滿洲國黑河地方に於る滿洲語の一特色」[One characteristic of the Manchu language of the Heihe region of Manchukuo], 『學叢』 3: 190–215, 京城 [Keijō]. Republished in Kōno (1975–1980). Kōno, Rokurō 河野六郎 (1975–1980)『河野六郎著作集』(Collected works), vols. 1–4, 東京 [Tokyo]: 平凡社. Lee, Robert H. G. (1970) Manchurian Frontier in Ch’ing History, Harvard East Asian Series 43, Cambridge, Massachusetts: Harvard University Press. Li, Shu 李書 (1986)「三家子陶氏家族史料」[Material on the history of the Tao clan of Sanjiazi],『滿語研究』1986 (2): 115–119, 哈爾濱 [Harbin]. Liu, Keyou (2021) L’influence du mandchou dans le pékinois moderne: une étude comparative sur le contact linguistique entre le mandchou et le pékinois moderne, PhD thesis, INALCO, Paris. Mu, Yejun 穆曄駿 (1985)「阿勒楚喀滿語語音簡論」[Short discussion about Alchuka Manchu phonemes],『滿語研究』1985 (1): 5–15, 哈爾濱 [Harbin]. Mu, Yejun 穆曄駿 (1986a)「阿勒楚喀滿語的數詞与格助詞」[Numerals and case markers in Alchuka Manchu],『滿語研究』1986 (1): 2–16, 哈爾濱 [Harbin]. Mu, Yejun 穆曄駿 (1986b)「拉林滿語語音概論」[A survey of Lalin Manchu phonetics],『滿語研究』1986 (2): 2–30, 哈爾濱 [Harbin]. Mu, Yejun 穆曄駿 (1987a)「十二詞頭拉林口語讀法解」[Pronunciation of the twelve heads of the Manchu alphabet in the Lalin spoken language],『滿語研究』1987 (1): 16–50, 哈爾濱 [Harbin]. Mu, Yejun 穆曄駿 (1987b)「巴拉語」[The Bala language],『滿語研究』1987 (2): 2–31, 哈爾濱 [Harbin]. Mu, Yejun 穆曄駿 (1988a)「論巴拉語的語音変化」[Discussing the phonetic changes in the Bala language],『滿語研究』1988 (1): 1–26, 哈爾濱 [Harbin].

482 Veronika Zikmundová and Gao Wa

Mu, Yejun 穆曄駿 (1988b)「阿勒楚喀語元音発声的音発特点」[Specific features of phonetic change in Alchuka Manchu vowels], 『滿語研究』1988 (2): 1–24, 哈爾濱 [Harbin]. Okada, Hidehiro (1992) ‘Mandarin, a language of the Manchus: how Altaic?’, Aetas Manjurica 3: 165–187, Wiesbaden: Otto Harrassowitz. Shirokogoroff, S. M. (1924) Social Organization of the Manchus: A study of the Manchu Clan Organization. Shanghai: Royal Asiatic Society. Wang, Qingfeng 王慶丰 (1984)「愛輝滿語概況」[A survey of the Manchu language in Aihui], 『民族語文』(5): 55–66, 北京 [Peking]. Wang, Qingfeng 王慶丰 (1986)「試論滿語的元音o, u, ū [A preliminary discussion of the Manchu vowels o, u, ū],『滿語研究』1986 (1): 2–16, 哈爾濱 [Harbin]. Wang, Qingfeng 王慶丰 (2005)『滿語研究』[Study of the Manchu language], 北京 [Peking]: 民族出版社. Yasui, Katsumi 安井克巳 (1943)「璦琿滿洲語」[The Manchu language of Aigun],『 音声學協会会報』 The Bulletin of the Phonetic Association of Japan 74 (5): 20–22, 東京 [Tokyo]. Zhao, Aping 趙阿平 & Chaoke 朝克 (2000)『黑龍江現代滿語研究』[A study of modern Heilongjiang Manchu], 哈爾濱 [Harbin]: 黑龍江教育出版社. Zhao, Jie 趙杰 (1987)「泰來滿語音位解析」[Pholonogical analysis of Tailai Manchu],『滿語研究』1987 (1): 51–66, 哈爾濱 [Harbin]. Zhao, Jie 趙杰 (1989)『現代滿語研究』[Study of modern Manchu], 北京 (Peking): 民 族出版社. Zhao, Jie 趙杰 (1995)「清初滿語京語重音前移及其對京腔漢語的影响」[The front shift of accent in Early Qing Peking Spoken Manchu and its influence on the Peking dialect of Chinese],『滿語研究』 1995 (1): 21–30, 哈爾濱 [Harbin]. Zhao, Jinchun 趙金純 (1986)「初探三家子滿語中動詞 “時” 的表示法」[Preliminary study of the expression of verb “tense” in Sanjiazi Manchu],『滿語研究 1986 (1): 49–55, 哈爾濱 [Harbin].

CHAPTER 18

SIBE Veronika Zikmundová

Sibe (Sibo, Xibo, Xibe) is a Southern Tungusic language of the Jurchenic group, spoken in Jungaria, in the north-west of the modern Xinjiang (Sinkiang) Uighur Autonomous Region of PR China, in the vicinity of the city of Ghulja (Yining). The Sibe speakers, who use the self-appellation siwe (sive), are officially classified in China as a separate minority nationality (Chinese Xibo zu 錫伯族, Russian SG MASC sibínec : PL sibíncy), with representatives also in the Manchurian provinces of Liaoning (Mukden), Jilin (Girin), and Heilongjiang. The Sibe in Xinjiang are descendants to Manchu garrison soldiers sent from Manchuria to Jungaria in 1764. Their history goes back to the Jurchen-speaking Sibe tribe, which in the 16th century lived in central and southern Manchuria as vassals of the Khorchin tribe of the Mongols, who at that time controlled over a large part of Manchuria. In the 17th century the Manchu emperor Kangxi bought out the Sibe, whom he regarded as a tribe related to the Manchu, from the Khorchins. Later the Sibe were integrated in the banner system of the Manchu army, and as such they were dispersed, moved, and resettled several times. The modern Sibe in Xinjiang represent a trace of one such forced migration. The Sibe in Xinjiang number currently around 30,000, a considerable part of whom are more or less fluent speakers of the Sibe language. In terms of speaker numbers, Sibe may today be the “largest” Tungusic language, though with little generational transmission it is probably less viable than Solon. In addition, some 150,000 ethnic Sibe still live in Manchuria, but they have lost their ethnic language in favour of Chinese several generations ago. In Xinjiang, the majority of Sibe speakers live in the Chabchal (Qapqal) Sibe Autonomous County (Chabuchaer Xibo zixhixian 查不查爾錫伯自治縣) in eight villages named niur ‘banners’ (niru), a reminiscence of their past as Manchu soldiers. Smaller enclaves in Ice Gasen ‘New Village’ or Khorgas (Huocheng), Nilka, Gongliu and Tarbagatai (Tacheng) originated in connection with resettlements in the 19th century. The Manchu writing system has been in use among the Sibe continually since its introduction (1599). A standardized pronunciation, some half-way between the written and the spoken form, and known as bitke gisun ‘literary language’, has traditionally been used for loud reading of written texts and recitation by heart. In the 1960s several changes were made in the orthography. The written language subsequently recorded in the reformed writing system is known as Written Sibe, and it differs from “Standard” Written Manchu also by being closer to the spoken language in terms of grammar and idiomatics. Written Sibe was widely taught in schools before the Cultural Revolution. During the Cultural Revolution (1966–1976), however, teaching of the written language was interrupted, and although later resumed, it has been less effective in the past decades. In connection with recent political developments, all teaching of Sibe was stopped in 2017. Even so, books, a journal, and a newspaper are still being published in Written Sibe, and the Chabchal TV has until recently been broadcasting news read in Sibe. Attempts have been made DOI: 10.4324/9781315728391-18

484 VERONIKA ZIKMUNDOVÁ

repeatedly, but insofar unsuccessfully, to codify a Latin transcription for spoken Sibe. A language committee whose task it is to enrich the Sibe vocabulary with terms for contemporary phenomena works continually in Chabchal. Some of the neologisms introduced by this committee have been used in the written language, but virtually none of them have become property of the spoken language. Taxonomically, Sibe is probably best described as a dialect of Spoken Manchu. It is mutually intelligible with the Sanjiazi dialect of Manchu, as still spoken by a few individuals in Manchuria. It differs from several other documented Manchu dialects by features which may be ascribed to the influence of Khorchin Mongol (on the lexical level), as well as to the impact of Kazakh and the Oirat Mongol dialects (on the phonetic level). Inside the Sibe speaking community, minor differences in phonetics and lexicon exist between Chabchal and the other enclaves, but these are not an obstacle to full mutual intelligibility. The varieties of Nilka, Gongliu and Tarbagatai have been more influenced by Kazakh, which used to be the majority language in these areas until the 1960s. Elderly speakers in these enclaves are bilingual in Sibe and Kazakh, while middle-aged and younger speakers usually speak only Mandarin, which became the majority language after the 1960s. Some bilingualism exists also between Sibe and Uighur, especially in Chabchal, where even some Uighur residents speak Sibe. There is also some differentiation between the “banners”, into which the Sibe used to be organized. Within Chabchal, the speech of the 5th banner slightly differs from the rest by being phonetically more conservative. An oral tradition holds that the members of this banner were originally speakers of a different language who were forcibly taught normative Manchu, which caused their speech being closer to the “literary language”. The speech of the 1st and 3rd banners, located at the greatest distance from the administrative center, is considered by the Sibe to be particularly “pure”, representing a kind of “standard” variety. In general, the language is best retained by elderly and middle-aged individuals in Chabchal. There are significant differences among individual speakers, which are mainly conditioned by the level of literacy in Written Sibe and the level of knowledge of Chinese. Literacy in the Sibe script is common among the older generation, while the degree of proficiency in Chinese increases in the younger generation. Most speakers today are bilingual in Chinese, code-switching is widespread, and Chinese loanwords form a significant part of the vocabulary. In the speech of the younger generation, the Chinese impact increases to a level close to a “mixed language”. The greater part of young Sibe have only a limited or no competency at all in the spoken language. DATA AND SOURCES The first substantial amount of Sibe language material, a collection of folk tales, was gathered by F. V. Muromski at the beginning of the 20th century, as published and analysed much later by Stanisław Kałużyński (1977, 1987). This material, which is of great interest for the study of Sibe culture, is probably an example of the “literary language”, since it involves a pronunciation used for reading or recitation of written texts and does not reflect the contemporary colloquial language. The first works dealing with oral Sibe therefore seem to be the phonological description by Hattori Shirō and Yamamoto Kengo (1956) and the “Classified Dictionary of Spoken Manchu” by Yamamoto (1969), followed by the “Sketch of Sibe morphology” by Jerry Norman (1974). Other early works on Sibe published outside of China include the historiographical paper by L. M. Gorelova (1988) and the small dictionary of Sibe-specific Manchu words by Giovanni Stary (1990).

SIBE 485

In China, the first linguistic studies of spoken Sibe appeared in the 1980s with the publication of the descriptions by Li Shulan (1985, 1986), partly in collaboration with Zhang Qian (1984) and Wang Qingfeng (1984), which include both grammatical and lexical material. A phonological sketch was authored by Wang Xiaohong and Guo Meilan (1985). Additional materials on the spoken language, with an English translation, were published by Jin Ning (1991, 1993). More recently, annotated texts have been published in Japan by Jia Xun (2014) and Kogura Norikazu (2018a). Kogura (2018b) has also authored an extensive monograph on Sibe modality, while Kubo Tomoyuki (1997, 2005, 2008) has studied issues of phonology and semantics, based on his own field research in Chapchal. The first textbook of genuine spoken Sibe was published by Kubo in collaboration with Kogura and Zhuang Sheng (2011). A grammatical description of Sibe in Chinese was published by the Korean linguist Jang Taeho (2008), followed by two syntactic papers in collaboration with Thomas E. Payne (Jang & Payne 2012, Payne & Jang 2012). Most recently, Jang (2020) has authored another general language sketch of Sibe. The most detailed description of Sibe in English to date, with a focus on morphology, is Zikmundová (2013). A considerable amount of language material in Written Sibe has been published in Urumchi (Ürümqi) mainly since the 1980s. They contain editions of both oral and written Sibe texts, literature of scholarly character, primary school textbooks, translations from Chinese and, importantly, original poetry and prose written by modern Sibe authors. A bilingual journal and a newspaper have been published until recently on a regular basis. However, after some three decades of publishing activity, new Sibe books have become rare due to the lack of young writers (and readers). At present, however, websites with texts, video and audio recordings in spoken Sibe are appearing. SEGMENTAL STRUCTURE The phoneme system of Sibe is somewhat difficult to delimit because of the variation between speakers in the degree of integration of Chinese elements. Also, especially the vowel system can be analysed in several alternative ways. However, a minimal system of vowels contains six segments (Table 18.1), comprising the two rounded back vowels u o, the two unrounded front vowels i ä, and the two central vowels e a. The higher central vowel e has the diffuse centralized value [ə], as in most other Tungusic languages, while the lower unrounded front vowel ä, which represents an innovation in comparison with Classical Manchu, is pronounced with the value [ɛ] of the lower mid level. Phonetically, the number of different vowel segments in Sibe is considerably larger, and most descriptions postulate at least one additional vowel phoneme, the high rounded front vowel ü [y], sometimes also the corresponding mid-high vowel ö [ø], but in these cases a sequential analysis is probably more correct, as in diof [døf] ~ duäf [dyəf] ‘fox’ (Written Manchu dobi). All vowels have lowered and velarized allophones in the vicinity of uvular consonants; this is particularly conspicuous in the case of u, which in uvular environment has the lowered and possibly pharyngealized value [ʊ]. TABLE 18.1 SIBE VOWELS u o

i e a

ä

486 VERONIKA ZIKMUNDOVÁ

Vowel length is not distinctive, but vowels can be combined into diphthongoid sequences, which are of either a closing or an opening type. The closing sequences comprise ai äi ei oi ao [ai ɛi ei oi ao], while the opening sequences are ia ie io ua uä [ia ie io ua yɛ], with ui iu [ui iu] being ambivalent. However, unlike the vowel sequences in several other Tungusic languages, the Sibe “diphthongoids” are rhythmically equal to single vowels, which is why it would probably be phonologically more correct to analyse them as sequences of vowels and glides, e.g. /ay äy ey oy uy aw/ and /ya ye yo wa wä/. Phonetically, these sequences are rather unstable, and they typically alternate with monophthongs, the most frequent variations being [io ~ ø ~ iɛ], [ai ~ æ ~ ɛ], and [iu ~ y]. The Sibe consonant system (Table 18.2) has some 21–24 phonemes, differentiated according to five places of articulation: labial, dental, palatal, velar, and postvelar or uvular, with the uvular set being a specifically Manchu feature. By the manner of articulation, the consonants are divided into the nasals n m ƞ, the weak stops b d j g ɢ, the strong stops p t c k q, the fricatives f s x h [χ], the glides w y, and the liquids l r. In terms of phonetic values, the strong stops p t k q in word-initial position are aspirated, while the weak stops b d g ɢ, are unvoiced and unaspirated. The palatal stops c j are realized as affricates with values varying from palato-alveolar [tʃ dӡ] or retroflex [tʂ dʐ] before non-front vowels to alveolo-palatal [tɕ dʑ] before front vowels. The labials f w are realized as dentilabial (labiodental) [f v], but the segment w is probably nevertheless best classified as a glide together with the palatal glide y. The strong labial stop p is diachronically secondary and still relatively marginal, occurring only in loanwords and onomatopoetic items. The speech of individuals fluent in Chinese may contain several other marginal consonant phonemes. In the position before i, the sibilant s and the nasal n are palatalized to [ɕ] and [ɲ], respectively. Some speakers may also have a palatal š [ʃ ~ ɕ] as a marginal phoneme, but for most speakers the distinction between s and š is not phonemic. It would also be possible to postulate a distinctive palatal nasal ñ in cases like niur = ñur [ɲur] ~ [ɲyr] < *niru (niru). In word-final position n is pronounced as a “goal toward which the tongue moves, but which it rarely (or perhaps never) reaches” (Norman). The velar nasal ƞ [ŋ] has the uvular allophone [ɴ] before the vowels a o u. The lateral l in syllable-final position is pronounced as an alveolar approximant [ɹ], e.g. ilha [iɹʁa]. As compared with Written Sibe and Written Manchu, spoken Sibe exhibits certain regular deviations connected with recent innovations and dialectal properties. For instance, there are examples of raising and palatalization of vowels, e.g. mirin ‘shoulder’ < *meiren (meiren), fiaƞqalen ‘low’ < *faƞqalen (fangkala). In bisyllabic words, the combinations e—u and u—e are normally assimilated to u—u, e.g. udun ‘wind’ < *edun (edun), muku ‘water’ < *muke (muke). In the initial syllable, the vowel i shows a tendency of breaking, TABLE 18.2 SIBE CONSONANTS m

n

[ɲ]

ƞ

[ɴ]

b

d

j

g

ɢ

p

t

c

k

q

f

s

[ɕ]

x

h

w

y l r

SIBE 487

at the same time as the vowel of the second syllable, if open, is reduced and can be lost, e.g. niur ‘arrow’ < *niru (niru). In some items, the initial y has been replaced by n [ɲ], e.g. nionhun ~ yonhun < *inahun (indahūn), niumesun < *umiesun (umiyesun). There are also examples of labial breaking, e.g. wilxe ‘sleeve’ < *ulxi (ulhi). In the original clusters gd and gj, the initial component g has developed variously to either o or i, e.g. aojun ‘thunder’ < *agjen (akjan), äide- ‘to believe’ < *akda- (akda-). Among other changes, the elision of r before consonants may be mentioned, e.g. saɢenji ‘girl’ < *sarɢanji (sargan jui). MORPHOPHONOLOGY AND PHONOTACTICS In spoken Sibe there is a strong tendency towards vowel reduction and loss, cluster formation and subsequent changes within clusters, as well as towards voicing assimilation and intervocalic sonorization of consonants. The overall phonetic shape of spoken Sibe with its palatal vowels and frequent consonant clusters appears, in comparison with the Sanjiazi Manchu dialect, to be influenced by neighboring Kazakh and Oirat. The reduction, fortition and lenition processes are partly ongoing changes which result in the parallel co-existence of two or more forms. In this situation, it is impossible to describe the syllable structure of Sibe in an unambiguous way at a single uniform synchronic level. There is a general tendency towards spirantisation of plosives and neutralization of voicing opposition in non-initial positions. The labials *b *w *f in non-initial positions, except after m n, have merged into w [v], which is voiced intervocalically, e.g. awele—‘to hunt’ < *abala- (abala-), and before voiced (weak) consonants, e.g. jawde- ‘to manage in time’ < *jabda- (jabdu-), but devoiced before voiceless (strong) consonants or a pause, e.g. awhe [afχ] ‘leaf’ < *awaha (abdaha), uwe [uf] ‘fate’ < *ube (ubu). The velar and uvular stops k q are usually spirantized in preconsonantal position, e.g. oqte [oχt] ‘medicine’ < *oqto (okto), while in original prevocalic position they retain their plosive prounciation, e.g. aneqe [anəq] ‘key’ < *anaka (anakū). In non-initial positions the velars g and x mostly merge into a uniform velar fricative which is voiced intervocalically, e.g. gege [kəγə] ‘elder sister’ < *gege (gege), yigə [jiγə] ‘goose’ < *yixe (niyehe), but devoiced before voiceless (strong) consonants or a pause, e.g. PL gex-te [kəxtə] ‘elder sisters’. The uvulars ɢ and h [χ] behave in an analogous manner, e.g. aɢa [aʁa] ‘rain’ < *aga (aga), arɢe [arχ] ‘method’ < *arɢa (arga), boh [poχ] ‘deer’ < *buhu (buhū). Of two consecutive fricatives the second one is subject to fortition and realized as a stop, e.g. wesixun [vəɕixun] > weiskun [veiʃkun] (wesihun). The sibilant s is voiced to [z ʑ] in intervocalic position, e.g. gisun [kizun] ‘language’ (gisun). The affricates c j merge to a spirantized c [ʃ ~ ɕ] before strong obstruents, e.g. tacqu [thaʃqu] ‘school’ < *taciqu (tacikū), bacqan [paʃqan] ‘a little more’ < *bajiqan (bajikan), as well as word-finally, e.g. IMP ɢac [qaʃ] ‘bring!’< *gaju (gaju). Vowel loss, in particular, has several phonetic and phonotactic consequences. In word-final position, most unstressed vowels are reduced and lose their syllabicity, which may lead to their complete loss. The rounded vowels o and u can also leave a trace of labialization on the preceding consonant, e.g. niuƞku [ɲuƞkw] ‘illness’. In trisyllabic words, the vowel of an open second syllable is also frequently lost (Mittelsilbenschwund), especially before velars and uvulars, e.g. [qoɕiχʊŋ] > ɢosqun [qosqʊn] ‘bitter’ (gosihon). Clusters of an affricate and a plosive are gradually simplified by the spirantization of the first element, e.g. dijiku [tidʑikw] > dick [tiʃk] ‘fuel’ (deijiku). Due to these developments, the maximal syllable structure in Sibe in CVCC, e.g. betk ‘foot’ (bethe). There are no

488 VERONIKA ZIKMUNDOVÁ

initial clusters, but in some cases, as if to prevent the formation of an initial cluster, a prothetic vowel is added (unless we are dealing with metathesis), e.g. sita- [iɕita-] ‘to be late’ < *sita- (sita-). The occurrence of the consonants is also subject to several additional positional restrictions. Thus, for instance, the labial glide w [v] occurs only before the unrounded vowels a e i. The palatal glide often alternates with n [ɲ] before palatal vowels, e.g. yineƞ [jinəŋ] ~ nineƞ [ɲinəŋ] ‘day’ < *ineƞe (inenggi). The segments r and ƞ do not occur word-initially. A stem-final or word-final n assumes the place of articulation of a following obstruent and is realized as [m ~ n ~ ɲ ~ ŋ]. This can happen both before suffixes and (in sandhi) at word boundaries. When in such cases the cluster mb comes to a word-final position because of the loss of a following vowel, the segment b is pronounced as an unreleased stop, e.g. yiqan ‘Chinese’ < *niqan (nikan) : ACC yiqam-be > yiqamb [jiqamb˺]. At the prosodic level, stress is non-distinctive, with the main stress normally falling on the first syllable, as is also suggested by the widespread reduction of vowels in non-initial syllables. However, the final vowel of some words has resisted vowel reduction, in which case it is pronounced as stressed and relatively long, e.g. afqa [afqa:] ‘sky’ (abka). The reasons underlying this behaviour are unclear, but the situation would seem to suggest an earlier prosodic contrast of either the accentual (*afqá) or durational (*afqaa) type, even though vowel duration is otherwise not distinctive. Vowels in monosyllabic words with open syllables are also pronounced as phonetically long, e.g. fa [fa:] ‘window’ (fa), we [ve:] ‘who?’ (we), 2SG si [ɕi:] (si), bo [po:] ‘house’ (boo). As in other varieties of Manchu, vowel harmony is not productive in Sibe, and suffixes have only a single harmonically invariant form, the only exception being the semiproductive diminutive (moderative) suffix -KAn, which appears in the variants -qan ~ -qon ~ ken, e.g. lawdu-qan ‘more’, yono-qon ‘funny’, ert-ken ‘early’. However, the distribution of vowels is to some extent governed by combination rules, which favour certain combinations. Following are some examples of frequent combinations: (a—a) amban ‘official’ (amban), (a—e) jalen ‘joint’ (jalan), (a—u) jaqun ‘eight’ (jakūn), (a—i) alin ‘mountain’ (alin), (o—e) ower ‘nose’ (oforo), (o—u) orun ‘position’ (oron), (o—i) orin ‘twenty’ (orin), (u—a) fuqa ‘circle’ (fuka), (u—i) ulin ‘property’ (ulin), (u—u) turun ‘flag’ (turun), (e—e) ewen ‘bread’ (efen), (i—a) bira ‘river’ (bira), (i—e) ice [itʃə:] ‘new’ (ice), (ä—i) ämil ‘male fowl’ (amila). WORD FORMATION Spoken Sibe distinguishes nominal and verbal derivation and inflection. Nominal words (nouns, pronouns, adjectives, numerals, spatials) follow in general one inflectional pattern, with suffixal markers for number and case, with the exception of spatial nouns which have a special marker for the directive case. Most of the derivational suffixes forming nominals in Written Manchu are fully lexicalized in Sibe. The few suffixes that remain partially productive include -Kan, -ƞe and ​‑liƞe. The suffix -Kan, which originally forms diminutives, indicates a moderately increased intensity of a quality and is often combined with the particle =di of Chinese origin (← di 地 的), e.g. ajig ‘small’ (ajige) : DIM aji-ken=di ‘little, very small’. The suffix -ƞe forms adjectival nominals from nouns, e.g. morin ‘horse’ (morin) : mori-ƞe ‘equestrian’ (moringga). This same suffix, or perhaps a homonymous element, has also the grammatical function of producing noun forms, or infinitives, from participles. The suffix -liƞe forms similatives, e.g. daiwe ‘medical doctor’ (daifu) : SIM daiwe-liƞe ‘like

SIBE 489

a doctor’. When added to a predicatively used participle it implies probability, e.g. ji- ‘to come’ (ji-) : PTCP.PRF-SIM ji-xe-liƞe ‘it looks like (he) came, (he) seems to have come’. In verbal derivation two suffixes are fully productive: the denominal suffix -le- and the deverbal suffix -we-. The suffix -le- functions as a general verbalizer of nominals, e.g. safqe ‘chopsticks’ (sabka) : safqe-le- ‘to eat with chopsticks’, and is widely used on Chinese bases, e.g. jao-le ‘to take a photograph’ (← Chinese zhao 照). The suffix -we- functions as a passive-causative marker, e.g. tici- ‘to come out’ (tuci-) : CAUS tici-we- ‘to take/let out’ (tucibu-), tande- ‘to beat’ (tanta-) : PASS tande-we- ‘to be beaten’ (tantabu-). An important innovation of spoken Sibe is the inchoative suffix -maci-, probably based on the combination of the connective converb in -me with the verb aici- ‘to move’ (aci-), e.g. gisere- ‘to speak’ (gisure-) : INCH-PST gisere-maci-hei ‘started speaking’. An important difference of the Sibe vocabulary as compared to Written Manchu is the tendency towards using analytic means rather than derivational morphology. For example, the Manchu directional deverbal suffixes -na- ‘to go to do something’ (andatives) and -nji- ‘to come to do something’ (venitives) are not productive in Sibe, which uses complex verbal expressions instead, cf. e.g. ale-me gene- ‘to go to tell’ (tell-CV.CONN go-) vs. Written Manchu AND ala-na- id. Some types of expressions, mainly of the iconic type, are formed by reduplication. Reduplicated adverbial roots, sometimes complemented by the verbs se- ‘to say’ (quotative) or are- ‘to make’, are widely employed in reference to various sensoric perceptions and emotions e.g. cir&cir are- ‘to be prickly, to feel anxiety’. In Written Sibe there are a number of synthetically formed neologisms coined and proposed by the language committee, e.g. sukdu+jen ‘car’ from sukdun ‘steam’ + sejen ‘cart’ (a calque of Chinese qiche 汽車), deye+tun ‘airplane’ from deye- ‘to fly’ + sonitun ‘machine’ (a calque of Chinese feiji 飛機). These are, however, not widely used in the spoken language, which, apart from employing Chinese loanwords, prefers to form new expressions by semantic change, e.g. sejen ‘car, bicycle’ < ‘cart’, or by analytic means. For new nominal meanings, phrases consisting of a participle and a generic nominal are typically used, e.g. jaqe je-re ba ‘restaurant’ (thing eat-PTCP.AOR place). NUMBER AND CASE For historical reasons, all Sibe nominal stems end either in a vowel (V) or in the primary (unstable) nasal n. This difference causes suffix allomorphy in the accusative case marker. The loss of final vowels has produced a large number of secondary consonant stems (C), but morphologically they still behave like vowel stems. The category of number in spoken Sibe is expressed by two derivational suffixes: -se and -te. The suffix -se is attached to nominals denoting humans, and it implies a specified group of people, e.g. gucu ‘friend’ : PL gucu-se. Before this suffix, a stem-final n is normally dropped, e.g. hawen ‘official’ : PL hawe-se [χafsə]. The suffix -te has the additional limitation of being restricted to kinship terms, e.g. gege ‘elder sister’ : PL gege-te ~ gege-se. The suffix ​‑se is also used in reference to a group of people associated with someone (associative plural), e.g. iktan-se ‘Iktan and his friends’, and as polite reference or address to someone, e.g. gege-se te ‘sister, take a seat!’ (elder.sister-PL sit.IMP). Both plural markers are optional. Multiplicity can also be expressed by lexical means, such as the noun gurun ‘state, people’, e.g. xexe ‘woman’ : xexe gurun ‘women’. The Sibe case system has six marked forms: the accusative, genitive, dative, ablative, directive, and instrumental (Table 18.3).

490 VERONIKA ZIKMUNDOVÁ TABLE 18.3 SIBE CASE MARKERS VC

n

ACC

-w/e

m-b/e

GEN

-i

DAT

-d/e

ABL

-deri

DIR

-ci

INSTR

-maqe

The final vowel can be dropped in the accusative and dative markers, resulting in the phonetic forms ACC [-f] (after vowels) ~ [-b˺] (after nasals), DAT [-d ~ -t] (after all stems). A reduced final vowel of the nominal stem is normally restored before case suffixes, usually as e [ə], but after a rounded stem vowel also as u [u]. Sample paradigms: jaqe ‘thing’ (unrounded vowel stem) : ACC jaqe-w/e : GEN jaqe-i : DAT jaqe-d/e : ABL jaqe-deri : DIR jaqe-ci : INSTR jaqe-maqe. fuƞku ‘handkerchief’ (rounded vowel stem) : ACC fuƞku-w/e : GEN fuƞku-i : DAT fuƞku-d/e : ABL fuƞku-deri : DIR fuƞku-ci : INSTR fuƞku-maqe. erin ‘time’ (nasal stem) : ACC erim-b/e : GEN erin-i : DAT erin-d/e : ABL erin-deri : DIR erin-ci : INSTR erim-maqe. The unmarked nominative functions mainly as the form of the subject, while the object and attribute are often unmarked too, unless they stand in the accusative and genitive, respectively. The genitive is altogether rare, used predominantly with pronouns and certain nouns such as nane ‘person’ : GEN nane-i to mark a possessive attribute. The genitive is also used to mark the subject of relative clauses. The accusative is mostly used to mark a definite direct object, but it also has a prolative function. In addition, it can mark the causee in causative constructions. The dative (dative-locative) marks location at or movement to something or somebody. In the possessive construction it marks the possessor, and in causative constructions it can also mark the causee. The directive (lativedirective) expresses direction towards something or somebody. The ablative expresses the source of action or movement. It is also used in the comparative construction to mark the point of comparison. The instrumental (comitative-instrumental) expresses the instrument of action (‘with the help of’) or the company of joint action (‘together with’). In the latter function it is often used in combination with the postposition embade (< eme ‘one’ + DAT ba-de ‘place’), e.g. sim-maqe embade ‘together with you’ (2SG-INSTR together). Apart from the regular nominal paradigm there is the special directive marker -si, which is used only with spatial nominals, e.g. ame- ‘back, north’ : DIR ame-si ‘backwards, to the north’, nuƞu- ‘top, up’ : DIR nuƞu-si ‘upwards’. Spatials also have other exceptional forms, though most of them have fallen out of use in spoken Sibe. A suffix still used in the oral language is -rxi > -rx, which produces adnominal forms from spatial stems, e.g. jule- ‘front, before’ : jule-rx > julx ‘the one in the front, southern’. NUMERALS The Sibe system of numerals is more or less identical with that attested in other varieties of Manchu. The basic numerals for 1–10 are: 1 em/e ~ DIM em-ken, 2 ju, 3 yilan, 4

SIBE 491

duyin, 5 sunja, 6 niƞun, 7 naden ~ nadin, 8 jaqun, 9 uyin, 10 juan. The form em/e is used attributively, while emken is used for the independently standing numeral. The items for the decades are variants of the Written Manchu ones: 20 orin, 30 ɢusin, 40 dixe, 50 susai, 60 in-je, 70 nadin-je, 80 jaqun-je, 90 uin-je, 100 taƞe. For the higher powers of ten the Mongolic loanwords 1000 miƞan and 10,000 tumen are used. The final nasal n in 10 juan, 20 orin and 30 ɢusin is dropped before digits beginning with a vowel, e.g. 11 juan+emu > jua-emu [ʤomu], 13 juan+(y)ilan > jua-ilan [ʤuailan]. Also, before the noun se ‘year of age’, all numerals drop the final n, e.g. yilan+se > yila-se [jilazə] ‘three years old’, juan+jaqun+se > juan-jaqu-se [ʤuanʤaqʊsə] ‘eighteen years old’. All native numerals for numbers higher than 3 are, however, often replaced by Chinese borrowings. Exceptional items for teens are preserved in 15 tofohon (tofohon) and the expressions omson bia ‘eleventh lunar month’ (omšon biya), jurhun bia ‘twelfth lunar month’ (jorgon biya). The ordinal numerals are formed by the suffix -c/i, before which the final nasal n is dropped, e.g. yila-ci ‘third’, duyi-ci ‘fourth’, sunja-ci ‘fifth’, niƞu-ci ‘sixth’, nade-ci ‘seventh’, jaqu-ci ‘eighth’, uyi-ci ‘nineth’, jua-ci ‘tenth’. Exceptional suppletive forms are ujui ‘first’ and jai ~ jaci ‘second’. Multiplicatives, are formed by the cliticized noun =medan (< *mudan ‘curve’), e.g. em=medan [əmdan] ‘once’, yilan=medan [jilamdan] ‘three times’. PRONOUNS The personal pronouns (Table 18.4) preserve the general pattern known from both Manchu and other Tungusic languages. The basic forms 1SG bi : 2SG si : 1PL.EXCL bo : 2PL so have the expected oblique stems 1SG min- : 2SG sin- : 1PL.EXCL mon- : 2PL son-, with the plural pronouns 1PL.EXCL bo : 2PL so showing a vocalism shared also by other forms of Spoken Manchu. The case declension follows the regular nominal paradigm, with the case markers added to the oblique stems. The first person plural inclusive pronoun is mese (muse), with a regular nominal declension, e.g. ACC mese-we : GEN mese-i, etc. The role of personal pronouns for the third person is filled by the demonstratives, often in combination with the nouns nane ‘person’ or jaqe ‘thing’. While Sibe lacks the original suffixal person markers of both the possessive (PX) and the predicative (VX) type, the adnominal genitival forms of the basic personal pronouns are used enclitically as secondary possessive markers (=PX) after nouns, e.g. bo=sini ‘your home’ (home=PX2SG). The postnominal enclitic forms differ from the corresponding prenominal forms phonetically only by being stressed more prominently on the final vowel, yielding the pronunciations [miˈɲi] : [ɕiˈɲi] : [moˈɲi] : [soˈɲi]. These forms are, however, marginal, for possession is normally expressed prenominally, e.g. min-i bo ‘my home’ (1SG-GEN home). By contrast, possession in the third person (both singular and plural) is regularly expressed by the enclitic possessive marker =ni, which represents TABLE 18.4 SIBE PERSONAL PRONOUNS   SG PL

NOM

OBL

ACC

GEN

=PX

1

bi

min-

mim-be

min-i

=mini

2

si

sin-

sim-be

sin-i

=sini

bo

mon-

mom-be

mon-i

=moni

so

son-

som-be

son-i

=soni

1 2

3

EXCL

=ni

492 VERONIKA ZIKMUNDOVÁ

a trace of the original third person singular pronoun *i : OBL *in- (i : OBL in- GEN in-i) otherwise lost in spoken Sibe. In the possessive construction involving two nouns the modifier is normally unmarked, while the head noun bears possessive marking, e.g. haheji gewe=ni ‘the boy’s name’ (boy name=PX3P). The demonstratives are ere ‘this’ (proximal) vs. tere ‘that’ (distal), which have a regular nominal case paradigm. The corresponding plural forms e-se : te-se are mainly used in reference to persons and function synchronically as true personal pronouns. The local case forms of the demonstratives can be used as independent adverbs with a local or temporal meaning, e.g. ABL e-deri ‘from now on’, DIR er-ci ter-ci ‘here and there’. There are also a number of special forms, e.g. e-wa ‘this place, here’ (uba) vs. te-wa ‘that place, there’ (tuba), er.a-liƞe ‘like this’ vs. ter.a-(li)ƞe ‘like that’. Emphatic demonstratives are formed by the prefix m(e)-, a borrowing from Khorchin Mongolian: mer = m-er ‘this very’ vs. meter = me-ter ‘that very’ : me-ter.aƞe ‘exactly so’. The function of a reflexive pronoun is filled by the noun beye ‘body’ > ‘self’, often with reduplication, e.g. beye beye-maqe gisere-mi ‘(s/he) is talking to her/himself’ (REFL REFL-INSTR speak-PRS), also in emphatic use, e.g. bi beye gene-xei ‘I went (there) myself’ (1SG REFL go-PST). The interrogative pronouns and composite interrogative expressions are ai ‘what? what kind of? : ai-tiƞe ‘when?’ : ai-liaƞe ‘like what?’ : awsi ‘how?’, we ‘who?’, ya ‘which?’ : yamken ‘which one?’ (< ya+emken ‘which one?’) : ya-wa ~ äwi ‘what place? where?’ (yaba), udu ‘how much/many?’. All the interrogative pronouns can be inflected according to the nominal paradigm. Examples of frequently used forms include, e.g. DAT ai-de ‘for what?’ : ABL ai-tiƞe-deri ‘from when?’, DAT yawa-de ~ ye-de ~ äwi-de ‘where?’ : DIR ye-ci ~ yawa-ci ‘whither?’, also with the third person possessive marker: ACC yamkem-be=ni ‘which one of them?’, ACC udu-we=ni ‘how many of them?’. VERBAL FORMS Spoken Sibe verbal morphology is comparatively simple, featuring a basic opposition of past vs. non-past (aorist/present-future) in finite forms and perfective vs. imperfective in non-finite forms. All verbal stems end lexically in a vowel, which can be lost or restored depending on the form. When restored, this vowel normally appears as e [ə], but after a rounded stem vowel also as u [u], following the same pattern as in nominal stems. Verbal forms include the standard categories of imperatives, finite indicative forms, participles, and converbs, complemented by the additional category of infinitives. With the exception of the imperatives, all these categories are formally linked with each other, with, in particular, the participles playing a central role in the system. •

The imperatives form a closed system of finite forms, which, however, is very much reduced as compared with most other Tungusic languages. The basic imperative, used in commands directed at the second person (singular and plural) is unmarked (-Ø), e.g. bu- ‘to give’ : IMP bu ‘give!’. A few historically irregular verbs have, however, special stem forms in the imperative: ji- ‘to come’ : IMP ju, je- ‘to eat’ : IMP jewe, bi- ‘to be’ : IMP bise, o- ‘to become’ : IMP ose, giä- ‘to take’ : IMP giäse, biä- ‘to look for’ : IMP biäse. The command can be softened by the “benedictive” suffixes -me and -ki, e.g. qarme- ‘to protect’ : BEN qarme-me ‘please protect!’, je‘to eat’ : BEN je-ki ‘please eat!’. The suffix -ki ~ -ci also yields the “voluntative” of the first person (singular and plural) and, in combination with the third person

SIBE 493

•

•

possessive marker =ni, the “permissive” in ‑ki=ni of the third person (singular and plural), e.g. ucule- ‘to sing’ : VOL ucule-ki ‘let me/us sing!’, bahe- ‘to find’ : PERM bahe-ki=ni ‘may s/he/they find!’. The converbs also form an essentially closed system, comprising four or five forms, including the connective converb in -me, the two consecutive converbs in -fe and -maqe, the conditional converb in -ci, and the terminative converb in -cini. The terminative converb is, however, marginal and is mainly used in idiomatic expressions, e.g. gelewe-cini ‘in a frightening manner’ (frighten-CV.TERM). Also, the consecutive converb in -fe is rare and seems to be borrowed from the written language, being mostly used by literate speakers as an alternative to -maqe. The conditional converb functions only as a head of subordinate clauses, expressing a condition or temporal subordination, often in combination with the enclitic particle =da, e.g. gene-ci=da [kənəʃta] ‘(even) if (one) goes’ (go-CV.COND=PTCL). The consecutive converb in -maqe is used in complex constructions, e.g. feksi-maqe ji-xei ‘came running’ (run-CV.CSEC come-PST), as well as in the role of the predicate of subordinate clauses, e.g. cekse ji-maqe nane aqu ‘when I came yesterday, nobody was (here)’ (yesterday come-CV.CSEC person NEG.EXIST). The connective converb indicates a closer connection with the main verb, often implying a parallel action. However, in many cases the connective and consecutive converbs are interchangeable, e.g. dosi-me ~ dosi-maqe ji-xei ‘[somebody] came in’ (enter-CV.CSEC ~ enter.CV.CSEC come-PST). It may be noted that the marker of the consecutive converb -maqe is identical with the instrumental case marker -maqe, suggesting a secondary transition from nominal to verbal inflection. The participles include three forms: the aorist (imperfective) participle in -re, the perfective participle in -Ke = -xe ~ -he ~ -ke ~ -qe, and the progressive participle in -mahe. The stem-final vowel is normally restored in the aorist participle, but lost in the perfective participle, while in the progressive participle both forms occur, e.g. ale- ‘to tell’ : PTCP.AOR ale-re [aler] : PTCP.PRF ale-he [aɹχ] : PTPC.PROGR ale-mahe [aləmaχ ~ aɹmaχ]. The verbs bi- ‘to be’, o- ‘to become’, je- ‘to eat’, and ji- ‘to come’ have the irregular aorist participle forms bisi-re, ojo-re, je-re ~ jete-re, and ji-re ~ jide-re, respectively. The distribution of the suffix variants -xe ~ -he of the perfective participle depends on the historical vocalism (palatal vs. velar) of the stem, e.g. gene- ‘to go’ : PTCP.PRF gene-xe [gənx ~ gənγə] vs. tate- ‘to pull’ : PTCP. PRF tate-he [thatχ ~ thatχə]. When the last consonant of the stem is a velar or uvular fricative, the resulting geminate is simplified, e.g. tuxe- ‘to fall’ : PTCP.PRF tuxe-xe > tuxe [thuxə], bahe- ‘to hit’ : PTCP.PRF bahe-he > bahe [paχə]. The plosive variants -ke ~ -qe are mainly required by the rule of phonotactics of the spoken language, e.g. taci- ‘to study’ : PTCP.PRF taci-he > tac-qe [thaʃq], though there are also a few diachronically conditioned cases, e.g. je- ‘to eat’: PTCP.PRF je-ke.

There is also a set of negative participles, formed by adding the suffix -qu (< NEG. EXIST aqu), whose uvular pronunciation causes the marker of the perfective participle to assume the uvular form -Qa- = -ha- ~ -qa-, e.g. gene- ‘to go’ : PTCP.PRF gene-he : PTCP.​ PRF-​NEG gene-ha-qu. The negative counterpart of the aorist participle has synchronically no participle marker, e.g. PTCP.AOR gene-re : NEG gene-qu. Participles mainly function as adnominal attributes, but also as finite predicates. In the latter function, the final vowel of the participle markers is retained, e.g. nane ji-xe [nan dʑiγə] ‘a person (= someone) came’ (person come-PTCP.PRF) vs. ji-xe nane [dʑix nan]

494 VERONIKA ZIKMUNDOVÁ

‘the person who came’ (come-PTCP.PRF person). The aorist participle rarely functions as a finite predicate, but with dative case marking it is frequently used in quasiconverbal constructions with a temporal content, e.g. yawe-re-de sawe-hei ‘(I) saw (it) upon leaving’ (leave-PTCP.​AOR-DAT see-PST). •

•

The participles can take the suffix -ƞe, which functions as a “nominalizer” in the sense that it forms nouns from nominalized verbs, that is, from the participial forms. The resulting forms AOR -re-ƞe : PRF -Ke-ƞe : PROGR -mahe-ƞe may be identified as “verbal nouns” or “infinitives”. In the sentence they normally take the role of the subject, e.g. yawe-re-ƞe=ni hudun ‘her/his walking is quick’ ~ ‘s/he walks quickly’ (walk-PTCP.AOR-INF=PX3P quick), but occasionally also that of the object, e.g. sin-i yawe-re-ƞe-we sawe-hei ‘(I) saw you walking’ (2SG-GEN walk-PTCP.AORINF-ACC see-PST). The infinitive corresponding to the perfective participle is used as a finite predicate with modal (evidential) semantics (as studied in detail by Kogura), e.g. jaqe je-ke-ƞe=na ‘did you eat?’ (thing eat-PTCP.PRF-INF=​INTERR) : je-ke-ƞe ‘yes, I did’ (eat-PTCP.PRF-INF). The finite indicative forms are monofunctional finite forms which express the present (non-past) and past tenses, as well as the progressive aspect. From the formal point of view, they are members of an integrated system which also comprises participial and converbal forms, to which the auxiliary element -i (< *+bi ‘to be’) has been added (Table 18.5). Thus, the finite present tense marker in -mi [-m ~ -mie] (-mbi) is based on the connective converb marker -me, while the past tense and progressive aspect markers -Kei and -ma-he-i > -mäi are based on the perfective and progressive participles, respectively. In practice, these forms double the participles in finite use, but unlike the latter, they have no other functions. For negation, the corresponding negative participles are used.

Following are samples of the conjugational paradigm of verbs. The verbs inflected here are yawe- (unrounded vowel stem) ‘to go, to leave, to walk’, ute- ~ utu- (rounded vowel stem) ‘to wear, to put on’, and o- (irregular stem) ‘to become’. Imperatives: IMP yawe [jaf], ute [utə] ~ utu [utu], oso [os]; VOL yawe-ki [jafki], ute-ki [utki], o-ki [oki]; PERM yawe-kini [jafkiɲi], ute-kini [utkiɲi], o-kini [okiɲi]. Converbs: CV.CONN yawe-me [javm], utu-me [utum], ome [omə]; CV.CSEC1 yawe-fe [jaffə], utu-fe [utufə], o-fe [of]; CV.CSEC2 yawe-maqe [yavmaq], utu-maqe [utumaq], o-maqe [omaq]; CV.COND yawe-ci [jaftɕi], ute-ci [utətɕi], o-ci [otɕi]; CV.TERM yawe-cini [yaftɕiɲi], ute-cini [utətɕiɲi], o-cini [otɕiɲi].

TABLE 18.5 SIBE VERBAL FORMS  

PTCP

NEG

PRS

CV

FIN

-me

-m-i

INFIN

AOR

-re

-Ø-qu

PRF/PST

-Ke

-Qa-qu

-Ke-i

-Ke-ƞe

PROGR

-ma-he

-ma-ha-qu

-ma-he-i > -mäi

-ma-he-ƞe

-re-ƞe

SIBE 495

Participles: PTCP.AOR yawe-re [javər], utu-re [utur], ojo-re [odʑor]; PTCP.PRF yawe-he [jafχ], ute-xe [utx], o-he [oχ ~ oʁʊo]; PTCP.PROGR yawe-mahe [yavmaχ], utu-mahe [utumaχ], o-mahe [omaχ]. Finite indicative forms: PRS yawe-mi [javəm ~ javəmie], utu-mi [utum ~ utumie], o-mi [om ~ omie]; PST yawe-hei [javʁəi], ute-xei [utxui ~ utkəi], o-hei [oʁʊi]; PROGR yawe-mahei [javmaʁəi] ~ yawe-mäi [javmɛi], utu-mahei [utumaχəi] ~ utu-mäi [utumɛi], o-mahei [omaχəi] ~ omäi [omɛi]. COMPLEX PREDICATES In addition to the synthetic forms of verbal conjugation, there are several verbal constructions that express grammaticalized functions by analytical means. These constructions typically involve the use of auxiliary verbs, especially bi- ‘to be, to exist’ and o- ‘to become, to have the identity of, to be(come) possible’, but also others. In combination with regular verbs, these auxiliaries form complex predicates with temporal, aspectual, modal, and evidential contents additional to those expressed by the monoverbal forms. •

•

•

The past tense form bi-xei ‘was’ is used in combination with nominal and verbal predicates to express various evidential and mirative meanings, e.g. ewa sahurun bi-xei ‘(I realize that) it is cold here (and I am surprised by it)’ (this.place cold be-PST), bi taqe-me bi-xei ‘(I am surprised that) I actually know him’ (1SG know-CV.CONN be-PST), si gisere-mahe bi-xei ‘(I remind you that) you were actually speaking (about it)’ (2SG speak-PTCP.​PROGR be-PST). In combination with the perfective participle of the main verb, the form bi-xei is also used to express perfectivity in the past (pluperfect), e.g. tere erin-de yawe-he bi-xei ‘at that time s/he was already gone’ (that time-DAT leave-PTCP.PRF be-PST). Predicates with the auxiliary verb o- ‘to become’ express various aspectual meanings such as prospectivity, e.g. aɢa da-me o-hei ‘it is going to rain’ (rain fall-CV. CONN become-PST), as well as modal meanings of possibility and necessity, e.g. gene-me o-mi ‘it is possible to go’ (go-CV.CONN become-PRS), gene-qu=da o-hei ‘it is possible not to go’ = ‘it is not necessary to go’ (go-NEG=PTCL become-PST). In combination with the negative participles the auxiliary o- forms the negative equivalents of converbs, e.g. gene-qu o-ci oju-qu ‘it is impossible not to go’ = ‘it is necessary to go’ (go-NEG become-CV.COND become-NEG). In these constructions, the auxiliary verb tends to merge with the preceding negative participle, yielding complex synthetic forms like yawe-qu-me [javʊqʊmə] < yawe-qu+o-me ‘not leaving’ (leave-NEG+become-CV.CONN), yawe-qu-ci [javʊqʊʧi] < yawe-qu+o-ci ‘in case of not leaving’ (leave-NEG+become-CV.COND). Other multiverbal predicates can be classified into two groups in accordance with their semantics: symmetrical, consisting of full verbs, and asymmetrical, involving combinations of a full verb with an auxiliary which partly retains its lexical meaning. Sequences of full verbs are used to describe closely connected actions, e.g. dosi-me tici-xei ‘went in and out, dropped in’ (go.in-CV.CONN come.out-PST). Some of these constructions are lexicalized, e.g. gene-me ji- ‘to maintain contacts’ (go-CV. CONN come). This pattern is also used to express the purpose of an action, with the verb denoting the purpose preceding the verb denoting the action, e.g. ta-me gene‘to go to see (someone), to visit’ (see-CV.CONN go).

496 VERONIKA ZIKMUNDOVÁ

In the asymmetrical type of construction the auxiliaries narrow the semantics of the main verb in different ways. The auxiliaries yila- ‘to stand’ and dude- ‘to lie’ add a progressive meaning, e.g. yawe-me yila- ‘to keep going’ (go-CV.CONN stand), giseremaqe dude- ‘to keep talking’ (speak-CV.CSEC lie). The verb yawe- ‘to go’ can convey the notion of habitual aspect, as in weile-me yawe-hei ‘(s/he) used to work’ (work-CV. CONN go.PST). The verbs bu- ‘to give’ and giä- ‘to take’ form benefactives, e.g. are-me bu- ‘to write for (someone)’ (write-CV.CONN give), eje-me giä- ‘to note, to take notes’ (note-CV.CONN take). The verb ji- ‘to come’ implies direction towards the speaker and can also be used temporally, e.g. baitele-maqe ji-xei ‘(we) have been using (it) since a long time’ (use-CV.CSEC come-PST). •

As a trace of the original negation verb (negative auxiliary) *e- Sibe preserves the forms eme = NEG-CV.CONN *e-me ‘not’ (ume) and endi ‘not yet’ (unde). Both are used in combination with the aorist participle form of the main verb, which in these cases retains its function as a connegative. Synchronically, the negator eme functions as a prohibitive particle placed before the verb, eme yawe-re ‘don’t go!’ (PROHIB go-PTCP.AOR = go-CONNEG), and it can also be used with the permissive form, e.g. eme je-kini ‘let him not eat!’ (PROHIB eat-PERM). The negator endi has the function of a “non-duminitive” particle and is placed after the connegative form of the verb, e.g. yawe-re endi ‘(s/he) has not left yet’ (leave-PTCP.AOR = leave-CONNEG NONDUM).

SYNTAX Sibe is an SOV language. The normal constituent order is subject—time—place— indirect object—direct object—predicate. In the noun phrase most modifiers (with the exception of the enclitic possessive forms of the personal pronouns and some quantifiers such as gume ‘every’) precede the head noun. There is no agreement, and the marking on the head belongs to the whole noun phrase, e.g. tere eme aji saɢenji-deri ‘from that one little girl’ (that one little girl-ABL). Spatial, temporal and causal relations are expressed by postpositionally used nouns, including both the actual spatials, e.g. nuƞu-de ‘above’ (top-DAT), and a few other lexicalized items, e.g. turgun-de ‘because’ (reason-DAT). Following are a few other syntactic properties characteristic of spoken Sibe: •

Topic marking has become almost obligatory in the colloquial language. Normally, the topic is placed first in the sentence and is marked by one of several topic markers, which for nominal constituents include the converbs o-me (become-CV.CONN) and o-ci (become.CV.​COND) and the possessive marker =ni, e.g. tere-we=ni bi ɢaji-hei ‘this was brought by me’ (this-ACC=TOP 1SG bring-PST). Another topic marker is the emphatic particle =da, which is particularly often used with converbs and participles. Topic markers can also be used together, e.g. bo=ni ome=da . . . niuƞku dase-mi ‘as for us, we cure diseases’ (1PL=TOP TOP=TOP . . . disease curePRS). Spoken Sibe has also some other features of a topic-prominent language. For instance, constructions with “double subjects” are frequent, e.g. ere jaqe betke ambu ‘he has big feet’ (this thing foot big). Also, constructions without a clear subject/ object/agent marking occur more or less idiomatically, most often as analogies to

SIBE 497

•

•

•

Chinese constructions, e.g. ame jasigan ji-xei ‘Father sent me a letter’ ~ ‘I received a letter from Father’ (father letter come-PST), tere jaqe baite tici-xei ‘he has caused some problems’ ~ ‘he has got into problems’ (that thing problem come.out-PST). Confirmation or denial of the content of a statement is expressed by the sentence particles menjaƞe ‘indeed, (it is) so’ vs. waqe ‘no, (it is) not so’ (waka), e.g. bi siwe nane menjaƞe ‘I am indeed a Sibe’ (1SG Sibe person AFF) vs. bi siwe nane waqe ‘I am not a Sibe’ (1SG Sibe person NEG). The particle waqe forms the equative counterpart to the negative existential aqu (akū), cf. e.g. ere nane waqe ‘this is not a human being’ (this person NEG) vs. ewa-de nane aqu ‘there is nobody here’ (this.place-DAT person NEG.EXIST). Other sentence particles include dere, ere, uru, bare, which express modality ranging from probability to mild emphasis, e.g. bi ere nane-we taqa-qu uru ‘but I don’t know this person’ (1SG this person-ACC know-NEG PTCL), si cimare ji-mi bare ‘you will come tomorrow, won’t you?’ (2SG tomorrow come-PRS PTCL). Interrogation in polar questions is expressed by the question particle =na. Before this particle, the finite present and past tense markers appear in a short form without the final vowel, e.g. gene-mi=na > gene-m=na [kənəmna] ‘will (you) go?’ (go-PRS=INTERR). All question types retain the normal word order. Subordinate clauses are marked by converbs or participles, which often take a subject in the genitive case. Equivalents of relative clauses are expressed by participles which modify their head noun, e.g. sin-i gisere-xe nane ji-xei ‘the person you were talking about came’ (2SG-GEN speak-PTCP.PRF person comePST). In noun clauses, when the participle has no head noun, the corresponding infinitive is used as an independent head noun, e.g. sin-i gisere-xeƞe ya emken ‘which is the one you spoke about?’ (2SG-GEN speak-INF.PRF what one). In object position, the infinitive stands in the accusative case, e.g. ai yihan acereƞe-we sa-qu ‘he did not know on which bull to load (it)’ (which bull load-INF. AOR-ACC know-NEG). Alternatively, the participle can be “nominalized” by the generic noun baite ‘matter’, e.g. sin-i yawe-re baite-we bi sa-qu ‘I do not know (the fact) that you are leaving’ (2SG-GEN leave-PTCP.AOR matter-ACC 1SG know-NEG).

Embedded temporal clauses are attached to the main clause mainly by converbs and quasiconverbs (case-marked participles). The converbs (proper) are normally used in conjunct (same-subject) constructions, e.g. yawe-maqe ɢuida-hei ‘he left long ago’ (leave-CV.​CSEC take.time-PST), while in disjunct (different-subject) constructions quasiconverbs, normally with dative marking, are used, e.g. sin-i yawe-re-de sim-be bene-ki ‘when you are leaving, I will see you off’ (2SG-GEN leave-PTCP.AOR-DAT 2SG-ACC see.off-VOL), daci sawe-he-de teraƞe gisere-xei ‘when I saw him earlier, he told me like that’ (earlier see-PTCP.PRF-DAT like.that speak-PST). The perfective participle can also imply subordination without case marking (zero instrumental, equivalent to the written form -ha-i), in which case it can be combined with the particle =da, a construction which emphasizes durativity, e.g. yawe-he gene-xe=da eme bo-de isine-xei ‘having walked (for some time) he came to a house’ (walk-PST go-PST=PTCL one house-DAT arrive-PST). In a somewhat similar construction, the infinitives can also be used, e.g. mim-be sawereƞe=da feksi-maci-hei ‘when s/he saw me s/he started running’ (1SG-ACC see-INF. AOR=PTCL run-INCH-PST).

498 VERONIKA ZIKMUNDOVÁ

•

The quotative function is expressed by the converbal form CV.CONN se-me of the verb se- ‘to say’, e.g. eni ji-xe=na se-me fienji-xei ‘he asked if Mother had come’ (mother come-PTCP.PRF=INTERR say-CV.CONN ask-PST).

LEXICON AND LANGUAGE CONTACTS Sibe shares most of its vocabulary, including older Mongolian and Chinese loanwords, with other varieties of Manchu. Among the few exceptions is nane ‘person’, which, at least superficially, is more similar to forms like Nanai naɪ than to mainstream Manchu ñalma (niyalma) > Sanjiazi Manchu niama [ɲama]. For Written Sibe, modern technical vocabulary is being coined following the Manchu derivation and compounding patterns, and considering Chinese semantic models, e.g. fehi.tun ‘computer’ from fehi ‘brain’ + sonitun ‘machine’. In spoken Sibe, Chinese loanwords are used without significant phonetical adaptation, e.g. [diennao] ‘computer’ ← diannao 電腦 ‘electric brain’. Of all known varieties of Manchu, Sibe is the one most profoundly influenced by Mongolic. The possible sources of this influence are Khorchin, Daghur, and Oirat. Mongolic loanwords are phonologically fully adjusted in Sibe. Most Sibe-specific lexical borrowings pertain to Buddhism and shamanic practices, e.g. douce [toutʃ], cf. Khorchin duuc [tu:tɕ] [a type of shamanic healer], but other lexemes are also attested, such as kurwu [khurvo] ‘bridge’, cf. Daghur xuurwu < *köörgö. Some grammatical markers, such as the third person possessive marker =ni, have direct analogues in Mongolic and may have been influenced by them. A borrowing is, in any case, the emphatic prefix m(e)- of demonstrative pronouns, based on the Mongolian deictic pronoun (*)mön ‘this/that very’, which is also used prefixally in Khorchin, e.g. Khorchin mi-tii ‘exactly like that’. After the migration of the Sibe to Jungaria, several loanwords from Russian, Uighur and, more locally, Kazakh have entered the language. Russian loanwords were both acquired through direct contact since the late 19th century and transmitted via Uighur. They include words denoting previously unknown concepts, e.g. padruga ‘girlfriend’ ← Russian podrúga, famdor ‘tomato’ ← Russian pomidór, masina ‘sewing machine’ ← Russian mashína, garman ‘accordion’ ← Russian garmón’. Russian loanwords are more or less adapted to Sibe, whereas Uighur loanwords are, as a rule, used without adaptation. Uighur words such as yataq ‘dormitory’, parang ‘talk, chat’ or bazar ‘market’ occur in the speech of most Sibe. In the multilingual communities of Ghulja and Urumchi many more Uighur words and phrases are used, e.g. caxcaq ‘joke’, cataq boldi ‘there is a problem’ cataq yoq ‘no problem’. There are also examples of language “mixing”, as in the pun yoq-deri yaxši ‘better than nothing’, based on Uighur yoq-tin yaxši (NEG.EXIST-ABL good), where the Uighur ablative marker ​‑tin has been replaced by its Sibe counterpart deri. REFERENCES AND FURTHER READING Gorelova, L. M. (1988) ‘The Sibe dialect of Manchu language as interpreted by Russian and foreign linguists’, in: M. S. Andronov & Bhakti P. Mallik (eds.), Linguistics—A Soviet Approach, 300–317, Indian Journal of Linguistics, Calcutta. Hattori, Shirō 服部四郎 & Yamamoto Kengo 山本謙吾 (1956)「満州語口語の音韻体 系と構造」[The phonological system and structure of Spoken Manchu],『言語研究 30: 1–29, 京都 [Kyoto].

SIBE 499

Jang, Taeho 张泰锅 (2008)『錫伯語語法研究』[A study of Sibe grammar], 昆明 [Kunming]: 雲南民族出版社. Jang, Taeho (2020) ‘Xibe and the Manchuric languages’, in: Martine Robbeets & Alexander Savelyev (eds.), The Oxford Guide to the Transeurasian Languages, 269–287, Oxford Guides to the World’s Languages, Oxford: Oxford University Press. Jang, Taeho & Thomas E. Payne (2012) ‘Dependency and clause combining in Xibe’, in: Lindsay J. Whaley & Andrej L. Malchukov (eds.), Recent Advances in Tungusic Linguistics, 229–256, Wiesbaden: Harrassowitz Verlag. Jia, Xun 郟勛　(2014)「シベ語テキスト—手延べ麺とシベ風パンの作り方—」 [Sibe texts: How to cook Tatere beda and Falexe eweN],『北方言語研究』4: 199– 212, 札幌 [Sapporo]. Jin, Ning (1991) ‘The Sibe-Manchu legend of Blackening the Face’, Aetas Manjurica 2: 147–173, Wiesbaden. Jin, Ning (1993) Sibe-English Conversations, with a foreword by Giovanni Stary, Wiesbaden: Harrassowitz Verlag. Kałużyński, Stanisław (1977) Die Sprache des mandschurischen Stammes Sibe aus der Gegend von Kuldscha, Warszawa: Państwowe wydawnictwo naukowe. Kałużyński, Stanisław (1987) ‘Характеристика сибинского языка’ [Characteristics of the Sibe language], in: Проблемы языков Азии и Африки [Problems of Asian and African languages], Warszawa, 195–210. Kogura, Norikazu 児倉徳和 (2018a)「シベ語」[Sibe language], in: Li Linjing 李林静 & Yamakoshi Yasuhiro 山越康裕 & Kogura Norikazu 児倉徳和 (eds.),『中国北方危 機言語のドキュメンテーション』, 東京 [Tokyo]: 三元社, 55–101. Kogura, Norikazu 児倉徳和 (2018b)『シベ語のモダリティの研究』[Studies of modality in Sibe language], 東京 [Tokyo]: 東京外国語大学アジア・アフリカ言語 文化研究所. Kubo, Tomoyuki (1997) ‘“Come” and “Go” in Sive Manchu’, Saksaha—Review of Manchu Studies 2: 19–24. Kubo, Tomoyuki (2005) ‘“Inclusive/Exclusive” Distinctions in Sive (Spoken Manchu) and Khalkha Mongolian’,『九州大学言語学論集』25–26: 255–270, 福岡 [Fukuoka]. Kubo, Tomoyuki (2008) ‘A sketch of Sibe phonology’,『語学研究フォーラム』16: 127–142, 東京 [Tokyo]: 大東文化大学. Kubo, Tomoyuki 久保智之 & Kogura Norikazu 児倉徳和 & Zhuang Sheng 庄声 (2011) 『シベ語の基礎』[Basic Sibe], 東京 [Tokyo]: 東京外国語大学アジア・アフリカ 言語文化研究所. Li, Shulan 李樹蘭 (1985)「錫伯語的狀詞」[Imitatives in Sibe],『民族語文』5: 12–25, 北京 [Peking]. Li, Shulan 李樹蘭 (1986)「錫伯語的話語材料」[Materials of oral Sibe],『民族語 文』6, 76–79, 北京 [Peking]. Li, Shulan 李樹蘭 & Zhong Qian 仲謙 (1986),『錫伯語簡志』[A sketch of the Sibe language],『中國少數民族語言簡志叢書』[Languages of the minority nationalities of China], 北京 [Peking]: 民族出版社. Li, Shulan 李樹蘭 & Zhong Qian 仲謙 & Wang Qingfeng 王慶豐 (1984),『錫伯語口語 研究』[Study of colloquial Sibe], 北京 [Peking]: 中央民族出版社. Norman, Jerry (1974), ‘A Sketch of Sibe Morphology’, Central Asiatic Journal 18: 159– 174, Wiesbaden.

500 VERONIKA ZIKMUNDOVÁ

Payne, Thomas E. & Taeho Jang (2012) ‘Topic marking and the construction of narrative in Xibe’, in: Pirkko Suihkonen & Bernard Comrie & Valery Solovyev (eds.), Argument Structure and Grammatical Relations: A Crosslinguistic Typology, 151–175, Studies in Language Companion Series 126, Amsterdam: John Benjamins. Stary, Giovanni (1990) Taschenwörterbuch Sibemandschurisch-Deutsch, Wiesbaden: Otto Harrasowitz. Wang, Xiaohong 王小虹 & Guo Meilan 郭美蘭 (1985「錫伯語口語音位系統」[Phonological structure of spoken Sibe],『満語研究』1: 48–53, 哈爾濱 [Harbin]. Yamamoto, Kengo 山本謙吾 (1969),『満州語口語基礎語彙集』A Classified Dictionary of Spoken Manchu with Manchu, English and Japanese indexes, 東京 [Tokyo]: 東 京外国語大学アジア・アフリカ言語文化研究所. Zikmundová, Veronika (2013). Spoken Sibe: Morphology of the Inflected Parts of Speech. Prague: Charles University, Karolinum Press.

CHAPTER 19

SOCIOLINGUISTIC ASPECTS OF TUNGUSIC Nadezhda Mamontova

Tungusic languages are currently spoken in two political states: Russia (the Russian Federation) and (the People’s Republic of) China. Tungusic languages spoken only in Russia include Neghidal, Ewen, Ulcha, Uilta, Oroch, and Udihe, while languages spoken only in China include Orochen, Solon, Manchu, and Sibe. Languages with speakers on both sides of the border include Ewenki and Nanai. Strictly speaking, however, Nanai is recently extinct on the Chinese side, while Ewenki in China survives in only two marginal communities, the Khamnigan and the “Manchurian Reindeer Tungus”, both of which ultimately represent 19th to 20th century migrant groups from the Russian side. In addition, there are two very minor groups with “mixed” languages, the Kili (Kur-Urmi Nanai) in Russia and the Kilen (Hezhen) in China, of which only the latter has some speakers left today. Nowadays all Tungusic languages are endangered. The two most resilient languages are Solon and Sibe, both of which are spoken on the Chinese side, but their status is also becoming increasingly precarious. The other Tungusic languages spoken in China, that is, Orochen and (Spoken) Manchu, are on the verge of extinction, leaving only Khamnigan Ewenki with some prospects of survival. The present discussion will focus on the current sociolinguistic status of the Tungusic languages in Russia: (Siberian) Ewenki, Neghidal, Ewen, Oroch Udihe, Nanai, Ulcha, and Uilta. The speakers of all these languages have the official status of “Small Indigenous Peoples of the North, Siberia, and the Far East” (korennye malochislennye narody Severa, Sibiri i Dal’nego Vostoka). Historically, all of the Tungusic-speaking populations in Russia have been involved in traditional activities of hunting, fishing, and gathering. In addition, reindeer breeding has been practised by the Ewenki, Neghidal, Ewen, and Uilta. Fishing has been of particular importance to the groups living in the Amur basin, that is, the Nanai and Ulcha, as well as the Neghidal. Even today, a significant part of Tungusic speakers are dependent on traditional types of subsistence economy. The reindeer-breeding groups have well into the 20th century been nomadic, while the groups living in the Amur basin have been living in more fixed settlements, often with separate summer and winter camps. Under the impact of the Soviet rule and postwar resettlement policies, most Tungusic speakers are today concentrated into a number of relatively large rural villages with a typically multiethnic population and with Russian as the dominant interethnic language. This has inevitably affected the status of the ethnic languages with large-scale endangerment as the result. DEMOGRAPHIC BACKGROUND Today, the situation for virtually all Tungusic-speaking groups in Russia is that fluent speakers can be found only among the oldest generation, while the children of the DOI: 10.4324/9781315728391-19

502 Nadezhda Mamontova

indigenous communities have, with few exceptions, at most a limited ability to understand simple phrases and words. The process of language loss is also evident from the official census data, which record the size of the ethnic population and the proportion of individuals recognizing affiliation with the ethnic language (Table 19.1). Although the census data in Russia reflect mainly the attitude of the minority groups towards their ethnic languages, rather than their actual linguistic competence, it is clear that there has been a systematic decline of native language command in the last several decades for all Tungusic languages. The decline started during the Soviet period and has only accelerated later. Typically, while the ethnic populations have in most cases remained stable, or even grown (the left column for each census year in the table), the percentage of individuals affiliated with the ethnic language has systematically diminished (the right column) for all the ethnic groups concerned. For instance, between 1989 and 2010, the size of the Ewenki ethnic group increased from 29,901 to 37,843 individuals, but the number of individuals who regarded Ewenki as their mother tongue decreased from 30.5 per cent to 11.5 percent, or in absolute terms from c. 9100 to c. 4300. In the same period, the percentage of ethnic language speakers fell from 44 to 22 per cent for the Ewen, from 44 to 7 for Nanai, from 31 to 5 for Ulcha, and from 24 to 5.5 for Udihe. For Oroch and Uilta the percentage dropped to below 5 percent, reaching only 2 per cent for the Oroch, but even this figure is exaggerated, for, as we know, Oroch was already extinct in 2010. TABLE 19.1 DECLINE OF THE ETHNIC LANGUAGES     Ewenki

1959  

1970 %

 

1979 %

1989

2002

2010

 

%

 

%

 

%

 

%

24583

56

25051

51

27294

42.5

29901

30.5

35527

19.5

37843

11.5

Ewen

9023

81.5

11819

55.5

12452

57

17055

44

19071

33

22383

22

Nanai

7919

87

9911

69

10357

56

11883

44

12160

26.5

12003

7

Ulcha

2049

85

2410

61

2494

37

3173

31

2913

13

2765

5

Udihe

1395

74

1396

54.5

1431

27.5

1902

24

1657

9

1496

5.5

Oroch

779

68.5

1037

47

1040

33

883

18

686

4

596

0.5

Neghidal

350

495

52

477

43

587

27

567

7

513

4

179

42

346

4

295

3.5

Uilta

The situation reflects the generational effect of linguistic assimilation, as young people normally do not know their ethnic languages even passively and regard Russian as their mother tongue. This phenomenon is especially obvious in regions with a high level of interethnic marriages. Russian is, however, not the only language to which Tungusic speakers have been shifting. In Yakutia young Ewenki and Ewen individuals often acquire Yakut as their new non-Tungusic vernacular. Until recently, a similar shift from Ewenki to Buryat was going on among the Ewenki speakers in Buryatia. In many cases it would be mistaken to talk about complete ethnic assimilation, for the linguistically assimilated individuals still tend to consider themselves as members of the various Tungusic groups.

Sociolinguistic aspects of Tungusic 503

LEGISLATIONAL STATUS To some extent, the current situation of the individual Tungusic languages is also conditioned by their legal status. Among the federal subjects of Russia, there are only a few regions where the indigenous languages are protected under regional legislation: •

•

•

•

In the Republic of Sakha (Yakutia), Ewenki and Ewen are recognized as local official languages. According to the “Law on the Languages in the Republic of Sakha (Yakutia)” (N 1170-XII, October 16, 1992), these languages can be used in all domains in localities with indigenous population along with Russian and Yakut, the two official languages of the republic. The republic also has a separate law protecting and promoting mass media in the minority languages: “On State Support of Mass Media in the Languages of Indigenous Peoples in the Republic of Sakha (Yakutia)” (N 858-IV, November 10, 2011). In Krasnoyarsk Krai, the “Law on the Native Languages of the Indigenous Peoples of the North, Siberia and the Far East Living in Krasnoyarsk Krai” (N 9–3816, November 5, 2015) provides for a right to use indigenous languages in public domains and in media, as well as in road signs and names of places and buildings. Among the Tungusic languages this law concerns Ewenki, while the other languages protected by it are (Samoyedic:) Nenets, Enets, Nganasan, Selkup, and (Yeniseic:) Ket, as well as (Turkic:) Dolgan and Chulym. The law further specifies that the local minority languages can be used as official languages in places of traditional residence of the indigenous groups and applied as means of tuition and/or taught as a subject in kindergartens and schools. Special legal protection is also granted to the languages of the indigenous populations of Sakhalin, where the “Law on the Languages of Indigenous Peoples Living in the Territory of Sakhalin Oblast’” (N 91, October 16, 2007) is applied to three Tungusic languages: Uilta, Ewenki, and Nanai, as well as to the Palaeosiberian (Amuric) Ghilyak (Nivkh) language. It may be specially noted that Nanai also enjoys the rights and privileges stipulated by this law, although the Nanai community on Sakhalin is not autochthonous to the island, but represents the result of resettlement from the continent in 1947. In other regions, indigenous languages are only mentioned in laws concerning the protection of the indigenous populations’ way of life, as well as in various programmes and strategies that aim at making it possible for the indigenous communities to use their ethnic languages in cultural activities and education. Laws granting special protection of the local indigenous languages have been passed in Magadan Oblast’ (N 2013, March 30, 2016), as well as on Chukotka (N 65, October 23, 2017) and Kamchatka (N 72, March 30, 2017). Some of the new regional initiatives are based on the “Strategy of Sustainable Development of the Small Indigenous Peoples of the North, Siberia and the Far East”, approved by the Russian government in 2009. A state-supported regional programme “On the Development of the Indigenous Peoples of the North, Siberia and the Far East Living in Khabarovsk Territory” was adopted in 2011 and is since then annually amended.

DATA AND SOURCES Language endangerment has become an increasingly important topic in general linguistics only since the 1980s. The first handbook of the endangered languages of Russia,

504 Nadezhda Mamontova

including all the Tungusic languages spoken in Russia, was published under the editorship of V. P. Neroznak (1994). The sociolinguistic situation of the “Small Peoples of the North, Siberia, and the Far East” during the last decade of Soviet rule was summarized by Janhunen (1990), followed by a more extensive discussion of the topic by Vakhtin (2001), cf. also Janhunen (2005) for a general overview of language endangerment among Tungusic speakers in Russia and China. A detailed study of the use of ethnic languages in education among the “northern peoples” is Arefiev (2014). The current status of Ewenki in Siberia and the Russian Far East has been the object of several specific studies, including Bulatova (1997), Sirina (2002, 2012), Turaev (2004, 2006), Kazakevich (2007, 2011ab), Mamontova (2014), Morozova (2014), Procukovich (2015), as well as Funk & al. (2012), while Ewen is discussed in Tarasova (1998, Ulturgasheva (2012), Kirillova (2015), and Tarabukina (2015), and Neghidal in Aralova & Pakendorf (2018). Gerasimova (2002) and Osipova (2013) focus on Nanai and Ulcha. Regional trends are discussed by Elaeva (2012) and Dyrxeeva (2014) for Buryatia, Zamorshhikova & al. (2010) for Yakutia, and Mamontova (2016) for Sakhalin. The following is a summary of this information with a view on selected regional and functional parameters. EWENKI AS A SUPRAREGIONAL LANGUAGE Ewenki differs from the other Tungusic languages by its wide geographical distribution, which covers historically almost all of North Asia. Although the total population of ethnic Ewenki in Russia is only c. 38,000 people (2010), the nomadic way of life has allowed the Ewenki to occupy historically a huge territory, extending from the eastern tributaries of the Ob and the Yenisei basin in the west to the Amur basin, the Sea of Okhotsk and the island of Sakhalin in the east. The principal areas of Ewenki traditional residence include the oblast’s (provinces) of Amur (1481 individuals in 2010), Irkutsk (1272), Sakhalin (209), Tomsk (95), and Tyumen’ (87), the krais (territories) of Krasnoyarsk (4372), Khabarovsk (4101), Zabaikal’sk (1387), and Primorsk (130), as well as the republics of Sakha (Yakutia) (21,008) and Buryatia (2974). In view of this wide geographical dispersal, it is obvious that the situation of language maintenance varies from one region to the other. •

Most of the western Ewenki groups live in Krasnoyarsk Krai and the Tyumen’ and Tomsk Oblast’s. Historically, a comparatively satisfactory sociolinguistic situation has prevailed in the Ewenki Raion (formerly Autonomous Okrug), or “Ewenkia”, of Krasnoyarsk Krai, but by the present day the majority of families have lost the language transmission between generations. There are, however, local differences: in the Tungussko-Chunsk group of settlements the language shift has resulted in a rapid loss of the ethnic language even among the middle generation, while in the Ilimpiya and Baikit groups of settlements there are still quite many middle-aged fluent speakers and even a few young people with a good command of the Ewenki language. In the settlement of Sovrechka of Turukhansk Raion, the situation is also relatively good. According to a survey conducted in 2006, the entire adult population of Sovrechka, including some non-Ewenki residents, could understand Ewenki and used it in simple communicative situations. All middle-aged Ewenki were fluent speakers, and even children preserved some knowledge of their ethnic language. By contrast, in the settlement of Sym of the neighbouring Yeniseisk Raion there were only three fluent speakers and a few more individuals, including one ethnic Russian,

Sociolinguistic aspects of Tungusic 505

•

•

•

who could understand Ewenki. A similar situation was documented for the Sym Ewenki of Tomsk Oblast’, where the language of communication in the surveyed settlements was Russian, and the ethnic language was not used in education or in media. Only two residents of Belyi Yar were fluent in Ewenki. On Taimyr, in the northernmost part of Krasnoyarsk Krai, Ewenki is still spoken in several villages, with some 50 speakers altogether (c. 40 in Khantaiskoe, 5 in Potapovo, and a few elderly individuals in Dudinka). The Ewenki language is also declining in Yakutia, which in absolute terms has the largest ethnic Ewenki population. According to the 2010 census, only 5.6 % of the Ewenki in Yakutia possess some knowledge of the native language. However, as in Krasnoyarsk Krai, the situation varies considerably from one settlement to another. In the settlement of Iengra of Neryungri Raion about 78 % of the local population has been reported to be fluent in both Ewenki and Russian, while the figure for the settlement of Tanya is only 5 %. Yet, field data collected in 2018 in Iengra shows that the situation regarding Ewenki language maintenance was not much different from other regions where Ewenki reside. In Kutana of Aldan Raion, only 2 % were able to communicate in Ewenki. Due to a high level of mixed marriages, Yakut served as a primary language of communication in most localities not only in public domains, but also at home. In some settlements Yakut is currently replacing both Ewenki and Russian in almost all domains. According to the census data, 81 % of Ewenki in Yakutia know Yakut and 91 % know Russian. Even so, most ethnic Ewenki keep their Ewenki identity even while being far more fluent in Yakut and/or Russian. The situation seems to be different in Buryatia, where trilingualism in Ewenki, Buryat, and Russian was still common in the 1990s. According to the 2010 census data, however, only 6 % of the ethnic Ewenki in Buryatia know Ewenki, while only 4 % know Buryat, despite the fact that Buryat has been obligatory in 80 % of the local schools. The situation is similar in the neighbouring Zabaikal’sk Krai, where only 5 % of the ethnic Ewenki know Ewenki, while 7 % know Buryat. In Irkutsk Oblast’, where a rapid language shift to Russian has been going on during and after the Soviet period, some 10 % of the ethnic Ewenki are reported to know Ewenki. Russian is the main medium of communication in all settlements with Ewenki population, and only some individuals of the older generation can fluently communicate in the Ewenki language. The Russian Far East has long been perceived as a region with a high level of language maintenance among the Ewenki. The number of speakers is, however, rapidly declining. According to recent data, there were only about 65 individuals, or 11 % of the ethnic population, fluent in Ewenki in Amur Oblast’. For example, in the village of Bomnak of Zeya Raion there were 43 speakers, or 22 % of the ethnic population, while in the village of Ivanovskoe of Selemdzha Raion the number of speakers was 15, corresponding to only 4 % of the local ethnic population. In Mazanovo Raion, which comprises a few nomadic camps located along the river Nora and its tributaries, 7 individuals, or 21 % of the ethnic Ewenki, have a very good command of Ewenki. On the whole, the language is better preserved in the Tynda and Zeya Raions and worse in Selemdzha Raion.

The current state of the Ewenki language in Khabarovsk Krai is largely unknown. A survey conducted in the 1990s revealed that 8 % of the residents in Tuguro-Chumikansk Raion, where the majority of the ethnic Ewenki are concentrated, spoke Ewenki fluently. The language was better preserved in the settlement of Torom located on the western

506 Nadezhda Mamontova

coast of the Sea of Okhotsk. Additionally, some ethnic Ewenki in the Tuguro-Chumikansk and Ayano-Maisk Raions were bilingual in Ewenki and Yakut. According to the 2010 census, some 7 % of ethnic Ewenki in Khabarovsk Krai still know the ethnic language. Subsequent surveys, conducted among the Ewenki in the Tuguro-Chumikansk and Ayan-Maya Raions in 2017, 2018 and 2021, have showed that the situation of the ethnic language is stable. The native language is best preserved in the settlements of Torom, Algazeya, Tugur (Tuguro-Chumikansk Raion) and Nelkan (Ayan-Maya Raion). It is unclear how many native speakers of Ewenki are left on Sakhalin, where the ethnic Ewenki live mainly in the villages of Viakhtu and Trombaus on the western coast of the island. However, already in the 1990s the situation was reported to be far from satisfactory. Research conducted in 2016 showed that there were no more than five speakers left at Viakhtu. There are also some Uilta or mixed Ewenki-Uilta individuals in the village of Val on the eastern coast who have some command of Ewenki. EWEN IN NORTHEASTERN SIBERIA The Ewen are dispersed widely over all parts of northeastern Siberia, including Yakutia (15,071 people in 2010), Khabarovsk Krai (1128), Magadan Oblast’ (2635), Kamchatka (1872), and Chukotka (1392). •

•

According to the 2010 census, 22 % of the ethnic Ewen living in Yakutia know the Ewen language, but with considerable regional differences. For instance, the westernmost dialect spoken in the village of Sebyan-Kyuël’ (Yakut Sebeen Küöl) of the district (ulus) of Kobiai (Kebeeyi) is still reported to be retained well by the majority of the ethnic Ewen in a situation of trilingualism in Ewen, Yakut, and Russian. Although Yakut is spoken more frequently in this locality, Ewen is also acquired by children and used by young people. By contrast, in the districts of Verkhoyansk and Oimyakon, as well as in the titular “Ewen national” districts of Moma (Muoma) and Eveno-Bytantai, the ethnic Ewen prefer Yakut as a lingua franca, while in the districts of Abyi, Lower and Upper Kolyma, Ust’-Yana, and Bulun the Russian language dominates in all communication. Ewen is also relatively well retained on Chukotka (Chukotskii avtonomnyi okrug) with some 14 % of the Ewen ethnic population speaking the language. An even higher level of language maintenance has been reported from Kamchatka (Kamchatskii krai), where some 27 % of the ethnic population know Ewen. These figures include a small number of children and young people still speaking Ewen as a first language. The situation is, however, considerably worse among the small group of Ewen living in the basin of the river Bystraya on central Kamchatka, where the language is highly endangered and is not being transmitted to children any more, with virtually all speakers being above the age of 40. In Magadan Oblast’ the language is virtually lost, being spoken by only some 2 % of the ethnic population. In Khabarovsk Krai, the language is retained by some 19 % of the ethnic population and is still being transmitted to children and learnt at school in some locations, notably at the village of Arka north of Okhotsk.

THE AMUR-SAKHALIN REGION The region extending from the Sungari, Ussuri, and Lower Amur basins to the island of Sakhalin forms a coherent geographical complex where the subsistence of the ethnic

Sociolinguistic aspects of Tungusic 507

populations is highly dependent on the resources of the rivers and the adjoining coastal waters. Altogether eight Tungusic languages are—or have until recently been—spoken in this region: Neghidal, Oroch, Udihe, Nanai, Ulcha, Uilta, Kili, and Kilen. Taxonomically, these languages represent four groups: Ewenic (Neghidal), Orochic (Oroch and Udihe), Nanaic (Nanai, Ulcha, Uilta), and “mixed” (Kili and Kilen). Kilen may still have some speakers on the Chinese side of the border, while the other languages are—or have been—spoken on the Russian side in the administrative context of Khabarovsk Krai and the Maritime Province (Primorskii krai or Primor’e). All the languages in the region are seriously endangered or moribund, and some are recently extinct. •

•

•

In the Nanaic group, Nanai is the most dispersed and relatively best preserved ethnic language. Its speakers live in the Nanai, Amursk, Komsomolsk, Solnechnyi, Khabarovsk, and Ulcha rural districts (raions) of Khabarovsk Krai, as well as in the Pozharsk and Olginsk districts of the Maritime Province. The Nanai on Sakhalin, resettled from the continent in 1947, live in the Poronaisk district, adjacent to the town of Poronaisk. According to a study carried out in 2002, the largest number of fluent speakers were found in the titular Nanai district. All of them were, however, above 40 years of age. The 2010 census indicates that the number of Nanai speakers was 714, or 6.5 % of the total Nanai population of Khabarovsk Krai. More recent information tells us, however, that interest in the Nanai language is currently growing in the village of Dada, the center of the Nanai district, and there are at least several semi-speakers among the younger generations, as well. In the Maritime Province and on Sakhalin, the Nanai language is no longer used as a means of communication. Kili, which used to be officially classified as a Nanai dialect has also recently become extinct. Ulcha and Uilta have declined even more rapidly than Nanai. According to a survey conducted in 2013 only 12 % of ethnic Ulcha residents of the titular Ulcha district in Khabarovsk Krai understood the ethnic language. The language functions today more as a symbol of ethnic identity, and it is not acquired by children. Nevertheless, there are still fluent speakers, though only among the oldest generation, and recent ethnic activism has raised the status of the language. The situation is similar for Uilta, but with the important difference that the size of the ethnic group and the absolute number of speakers are much smaller. Estimates made in the current millennium give the number of Uilta speakers variously from 16 to 30, but in 2017 only four speakers were identified. Most of the last speakers and semi-speakers represent the northern dialect, spoken at the settlement of Val, where a few Uilta individuals have also some command of Ewenki. There may also be some semi-speakers of the southern dialect left in Poronaisk Raion, though most of the Uilta ethnic population from South Sakhalin (Karafuto) was evacuated to Japan after WW2. In the Orochic group, Oroch is recently extinct, with no active speakers left. A survey made in 2001 revealed a few individuals aged 80 or more who had spoken the language in their childhood, but even they had switched to Russian and almost lost their competence in the ethnic language. Udihe is in a slightly more favourable situation thanks to the larger overall size of the ethnic population, but its speakers are likewise restricted to the oldest generation. In this context, it may be noted that the Oroch and Udihe are the only officially recognized Tungusic-speaking groups whose ethnic population has diminished during the last few decades. In the period from 1989 to 2020, the officially registered Oroch population (which itself is difficult to estimate because of confusion with the Orok and Orochen) decreased some 33 % (from 883 to 596), while for the Udihe the drop was about 21 % (from 1902 to 1495). In 2002,

508 Nadezhda Mamontova

•

the number of Udihe speakers was estimated at 100, of whom some 50 had a fluent command. It seems that Udihe is nowadays preserved mostly as a symbolic marker of ethnicity, but as such it is not used as a spoken language. Even so, in 2010 close to 5 % of the ethnic Udihe still claimed Udihe as their “native language”. Neghidal used to be spoken in two varieties, Lower (Russian nizovskii) and Upper (verkhovskii) Neghidal. There are no speakers left today of Lower Neghidal, which was until recently spoken in the villages of Tyr and Beloglinka on the Lower Amur. Among the speakers of Upper Neghidal, three individuals with a fluent knowledge and a few more people who could understand the language passively were registered in 2008 in the village of Vladimirovka in the Amgun’ basin.

TUNGUSIC LANGUAGES IN EDUCATION Considering the fact that almost all Tungusic languages are not spoken at home any more, the children can only study their ethnic languages at schools, where learning begins from zero level. Many native speakers of the parent and grandparent generations prefer to speak only Russian at home because they do not think there could be any sphere in which the children would need a knowledge of their ethnic language. Therefore, the role of school education in language transmission is potentially very important. However, the current practices of ethnic language teaching in education involve many problems. As it is, no single Tungusic language is used as a means of tuition at any level of education. In some kindergartens the indigenous languages may be present in the upbringing process as an “ethnocultural component”, which, however, does not involve their use in a communicative function. At schools, these languages are taught either as a subject or as an optional class. The time assigned to ethnic language teaching varies from 1 to 2 hours per week, which is obviously far from enough to develop any active communicative skills. Outside the school the children normally do not have any sphere where they could practise and develop further the knowledge acquired at school. There are also considerable differences between rural and urban environments, as well as between the individual Tungusic languages (Table 19.2, based on information for 2015/2016 available on the website of the Russian Ministry of Education). TABLE 19.2 ETHNIC LANGUAGES AT SCHOOL Language

Schools

Pupils per grade 1–4

urban

rural

urban

5–9 rural

urban

10–11 rural

urban

Total

rural

urban

rural

Ewenki

2

29

10

382

23

525

38

88

71

995

Ewen

1

19

0

485

50

547

75

127

125

1159

Nanai

1

16

3

458

0

473

0

34

3

965

Ulcha

0

1

0

0

0

0

0

0

0

0

Udihe

0

0

0

0

0

0

0

0

0

0

Oroch

nd

n.d

Neghidal

0

0

0

0

0

0

0

0

0

0

Uilta

0

0

0

0

0

0

0

0

0

0

Sociolinguistic aspects of Tungusic 509

As may be seen, most schools where the ethnic languages are taught are located in rural environments, a situation which, of course, is a direct consequence of the fact that the ethnic populations live mainly in rural localities. Also, the Tungusic languages included in the official system of school education comprise only Ewenki, Ewen, and Nanai. The teaching of these languages at school actually comprises the majority of all children in the ethnic groups concerned. In 2013, for example, 70 % of ethnic Nanai children in Khabarovsk Krai studied Nanai at school. Ewenki was studied by 68 % of ethnic Ewenki children in Ewenki Raion of Krasnoyarsk Krai, 88 % in Amur Oblast’, and almost 100 % in Zabaikal’sk Krai. Among the latter, 14 % studied Ewenki as subject and 86 % as an optional class. Recently, there has been a trend to move ethnic language teaching to optional classes. As a result, over the past 10–15 years the proportion of pupils studying their ethnic languages as a subject has significantly diminished, with a drop of 16.7 % for Ewenki, 3.5 % for Ewen, and 6.4 to 9.5 % for Nanai. The total number of children attending optional language classes has also dramatically decreased, especially in the case of Ewen, with a drop of 22 %. There have also been attempts of offering optional classes also in Udihe and Neghidal, but with the general decline of these languages this option has been discontinued. These developments indicate the worsening situation of ethnic language teaching. While reflecting a change in the attitude towards ethnic languages, they are also conditioned by the unavailability of competent teachers in many remote settlements. In public opinion, the ethnic languages are increasingly often regarded as a mere attribute facilitating the process of cultural acquisition, rather than as a functional means of communication. In the following, an approximate survey of the present-day situation is given region by region. Note that the situation is liable to rapid changes, and the official information does not always correspond to the reality. Even the presence of ethnic languages in the official school curricula does not guarantee that actual classes in them are offered. •

•

•

In Khabarovsk Krai, according to information from the local Ministry of Education, four Tungusic languages (Nanai, Ulcha, Ewenki, Ewen) are taught as a subject in 21 schools (999 pupils) and as optional classes in 11 schools (237 pupils). The total number of pupils learning ethnic languages has decreased over the past few years, though for Nanai a slight increase has been observed in 14 schools. Nanai is the only ethnic language for which a full programme from the 1st to the 11th grade is offered. Ewen and Ewenki are both taught in grades 2 to 9, the former in one school in Okhotsk Raion and the latter in three schools in the Tuguro-Chumikansk, Upper Bureya, and Ayan-Maya Raions. Ulcha is taught in Ulcha Raion as a subject in one elementary school and as an optional class in three schools in grades 1 to 4 (altogether 10 pupils). Neghidal and Udihe used to be offered as optional classes in two schools. Oroch has not been taught at all. In the Maritime Province, in the late 1990s, Udihe was studied as a subject in the settlements of Krasnyi Yar and Agzu. Subsequently, the language classes became optional, and in 2013 they were dismissed from the school schedule due to the lack of financial support. Local activists attempted to introduce free Udihe courses for adults, but this initiative did not result in any significant change. Nanai and Ewenki have not been taught in any settlement in the Maritime Province. On Sakhalin, some positive changes may be mentioned with regard to the Uilta language. The language remained unwritten until the perestroika period, and it was

510 Nadezhda Mamontova

•

•

never taught at school. In the 1990s, thanks to the efforts of the Japanese linguist Ikegami Jirō, a written form was developed for Uilta, based on the southern dialect. A few years ago Uilta lessons finally became available in the local boarding school located on the island of Yuzhnyi of Poronaisk Raion. Today about 10–15 children learn the language there in the elementary school. However, in the settlement of Val, where the majority of the remaining Uilta speakers live, the language has never been taught. Lessons of Ewenki have, however, been recently introduced in the settlement of Viakhtu. In Amur Oblast’, Ewenki is an obligatory class for all pupils of the first grade. Starting from the second grade it is taught as an optional class for Ewenki children. There are also two nomad schools for children of reindeer herders who spend most of their time in the taiga. This project, called “Ewenki-French nomad school”, was initiated by the anthropologist Alexandra Lavrillier in collaboration with local Ewenki activists in the early 2000s. One of the aims of the project is to preserve the language by keeping the children in the nomadic environment with their Ewenki-speaking parents. Unfortunately, evidence from other regions suggests that this approach may not be as effective as one would wish. Research conducted in a few reindeer brigades at Surinda (Ewenki Raion) and Sovrechka (Turukhansk Raion) in Krasnoyarsk Krai proves that even in the taiga the children are switching to Russian. In families where both parents have a perfect mastery of the Ewenki language and are involved in nomadic activities, communication with children nevertheless often takes place in Russian. The children may acquire herding terminology, which they use as a professional jargon, but they do not necessarily learn to speak the Ewenki language. Therefore, children learn the ethnic language only if their parents know it and speak it with them, irrespective of whether they live in a settlement or in a brigade. Similar nomad preschools, schools and summer camps functioning in reindeer brigades are also operating in Yakutia, where the very first nomad school was established as early as 1991. In the period between 2003 and 2012, the authorities of the Republic of Sakha passed a few important laws aimed at developing nomad education, including the laws “On the ancestral and tribal nomad communities (obshhina) of the indigenous peoples” (2003), “On nomad schools in the Republic of Sakha (Yakutia)” (2008), and “On state support of the indigenous peoples of the Republic of Sakha (Yakutia) leading a nomadic way of life” (2012). As a part of the UNESCO project “Education for All”, the Centre for Development of Nomad Educational Institutions (established in 2006) has implemented a programme on “Socialization of nomadic children with their parents”. Nowadays there are already 11 nomad schools, including 4 for the Ewenki and 5 for the Ewen, providing teaching for all grades. The nomad-oriented education includes family-based teaching, allowing parents and other community members to work as tutors. The graduates of nomad schools may continue their studies in institutions of professional and higher education. However, the language of instruction in all of these schools is either Russian or Yakut. In rural settlements, the Ewenki and Ewen languages are taught as subjects or optional classes in 13 and 22 schools, respectively. Although in 12 schools the children may attend Ewen and Ewenki classes from the 1st to the 11th grades, the ethnic language does not serve as a means of tuition in any of them. Yakut is the most widely used language of communication and the preferable tool of teaching among all indigenous minorities in the republic. A research conducted in 2007–2008 revealed that 18 to 35 per cent of Ewenki and Ewen individuals wanted their children to attend schools

Sociolinguistic aspects of Tungusic 511

•

•

•

with Yakut as the language of instruction and with Russian as a subject. The percentage of those who would prefer a Yakut school with the indigenous languages as a subject varied from 4 to 29 depending on the district. Neighbouring to Yakutia, the Ewen language is taught on the elementary level in 2 schools in Bilibinsk Raion of Chukotka (about 100 pupils) and in 3 schools in the Penzhinsk and Bystrinsk Raions of Kamchatka (about 80 pupils). In the settlement of Gizhiga of the Severo-Èvensk Raion of Magadan Oblast’, there are Ewen lessons from the 1st to the 11th grades (less than 20 pupils). As an optional class Ewen is also taught in the elementary school of the settlement of Takhtoyamsk in the Ol’sk Raion. In Bauntov Raion of the Republic Buryatia, the Ewenki language is taught two hours per week from the 2nd to the 9th grades. In 2012, the number of children attending Ewenki classes was about 150. There are also free language courses for adults organised by the local centres for Ewenki culture. In Zabaikal’sk Krai, Ewenki is taught as an optional class in 6 elementary schools. In Irkutsk Oblast’ only a minimal amount of ethnic language teaching is offered. As of 2013/2014, the Ewenki language was taught there two hours per week as an optional class in the secondary school of Yerbogachen and one hour from the 3rd to the 9th grades in the settlement of Vershina Tutury. There were no classes in the boarding schools in the Kachugsk and Kazachinsko-Lensk Raions. Over the past decade, the total number of pupils studying the ethnic language has never exceeded 50. The leading position of Ewenki language teaching is held by the Ewenki Raion, or Ewenkia, of Krasnoyarsk Krai, where Ewenki is taught in the majority of schools from the 2nd to the 9th grades and up to the 11th grade at the village of Baikit. Altogether, the language is taught as a subject in 15 out of 22 schools. As of 2013, the number of Ewenki pupils studying the ethnic language in Ewenkia was 670, or 67.6 % of the total number of Ewenki children. However, Ewenki is not used as a means of teaching in any school in Ewenkia. Within Krasnoyarsk Krai, Ewenki is also taught in the elementary school of the settlement of Khantaiskoe Ozero on Taimyr. Importantly, there are also three training centers involved in the upgrading of the Ewenki language teachers’ qualifications, where 9 to 13 per cent of the rural school teachers are being retrained on an annual basis.

The following is a summary of ethnic language teaching by the region as of 2013 (Table 19.3, data from Arefiev 2014 and personal observations). TABLE 19.3 LANGUAGE TEACHING BY THE REGION Region   Khabarovsk Krai

Language  

Schools subject

Pupils

optional

subject

Teachers

optional

 

Nanai

14

6

737

102

19

Ulcha

1

3

5

93

4

Ewenki

1

2

30

24

3

Ewen

1

0

99

0

1

Udihe

0

1

0

20

1

Neghidal

0

1

0

3

1

Oroch

0

0

0

0

0 (Continued)

512 Nadezhda Mamontova (Continued) Region   Maritime Province Sakhalin

Language  

Schools subject

Pupils

optional

subject

Teachers

optional

 

Udihe

0

0

0

0

0

Nanai

0

0

0

0

0

Uilta Ewenki Nanai

1

0

14

0

1

n.d.

n.d.

n.d.

n.d.

1

0

0

0

0

1

Amur Oblast’

Ewenki

2

2

102

110

7

Yakutia

Ewen

22

0

888

0

26

Ewenki

13

0

875

20

19

Chukotka

Ewen

2

0

98

0

2

Kamchatka

Ewen

2

1

76

14

4

Magadan Oblast’

Ewen

1

1

19

6

2

Buryatia

Ewenki

5

0

247

0

6

Zabaikal’sk Krai

Ewenki

2

6

31

187

6

Irkutsk Krai

Ewenki

1

3

3

44

3

Ewenkia

Ewenki

9

11

671

835

15

Taimyr

Ewenki

1

2

4

21

3

Turukhansk Raion

Ewenki

0

0

0

0

n.d.

Tomsk Oblast’

Ewenki

0

0

0

0

n.d.

TUNGUSIC LANGUAGES IN THE PUBLIC SPHERE Official literary languages using the Latin-based Unified Northern Alphabet were created in the early Soviet period for Ewenki, Ewen, Nanai, and Udihe, all of which were discontinued in the late 1930s. New Cyrillic-based orthographies were subsequently created only for Ewenki, Ewen, and Nanai. All of these languages have been used over the years for writing and publishing a small number of original and translated belletristic works, newspapers, as well as school textbooks. After the Soviet period new Cyrillic-based orthographies have also been created for Neghidal, Oroch, Udihe, Ulcha, and Uilta, though all of these have had very limited use, confined mainly to language tools and textbooks. In the published media the Tungusic languages are usually present in the form of supplements to leading regional newspapers. For example, the newspaper Èvenkiiskaya zhizn’ ‘Ewenki Life’ has been published in Ewenkia weekly since 1933. Each issue contains an insert in the Ewenki language. Materials in Ewenki are also sporadically published by a few regional newspapers in Yakutia, Buryatia, and Zabaikal’sk Krai. Articles in Ewen may be found in such newspapers as Ilken (Yakutia) and Aborigen Kamchatki (Kamchatka). The majority of regions with Tungusic speaking inhabitants have neither television nor radio programmes in Tungusic or other minority languages. However, in Buryatia the stations Radio Buryatiya and Ulgur occasionally broadcast in Ewenki. On Chukotka, the regional radio broadcasts in Ewen.

Sociolinguistic aspects of Tungusic 513

Book publishing activity in minority languages is traditionally comparatively developed in Russia. Even in recent years, works by several prominent indigenous writers and poets, as well as fairytales and other folklore materials have been published. Additionally, some biblical texts and the “Declaration of the United Nations” have been translated into a few Tungusic languages (Ewenki, Ewen, Nanai, Uilta). However, a demand still remains for modern literature and learning tools in the ethnic languages, first and foremost modern textbooks. For example, although Ewenki is better provided in this respect than the other Tungusic languages, most Ewenki textbooks and manuals are outdated, with the exception of perhaps only a single textbook, Ewediit tatkaljaraw by F. M. Lekhanova & et al. (2012), which is designed for those who do not speak the language. For some languages only ABC-books and primers for the elementary level have been published. It is, however, worth to mention that the situation is changing with the introduction of modern social media. The number of websites and electronic programs using indigenous languages is steadily growing. One of the positive examples is the promotion of the Ewenki language in cyberspace. Some Ewenki leaders and activists from Yakutia have already designed the first Ewenki digital teaching aid and a Russian-Ewenki dictionary that can be downloaded and installed in devices using the Android operating system. A few years ago the Russian social network VKontakte (vk.com) was translated by the members of a group of Ewenki activists into the Ewenki language. This stimulated some fresh interest in the language, and the users started to create pictures and memes in Ewenki, as well as posters facilitating the acquisition of Ewenki grammar and the most commonly used words. In Vkontakte there are also groups for those who wish to study Nanai and Ewen. FUTURE PROSPECTS Among all the Tungusic languages spoken in Russia, only Ewenki, Ewen, and Nanai may be regarded as relatively “well” preserved, since children in a few settlements still occasionally acquire them as their mother tongues. However, even these languages are seriously endangered. Ulcha and Udihe have apparently already ceased to be transmitted to children, while Neghidal and Uilta are on the verge of extinction, and Oroch is recently extinct. In spite of this, almost all Tungusic languages, with the exception of Oroch, are still being taught, or have until recently been taught at schools. Again, the situation is better for Ewenki, Ewen, and Nanai, especially in those communities and places of traditional residence, where these languages are still used as means of communication. The status of the ethnic languages is weakened by the increasingly commonly applied optionality of the ethnic language classes. Udihe and Neghidal have been present in education only as optional classes, and their teaching has apparently been discontinued altogether. A major problem is that the Tungusic languages are nowhere used as means of tuition in either schools or kindergartens. In some settlements the teaching of the ethnic languages extends only to the third grade, meaning that the children have no possibility to continue language learning on more advanced levels. Moreover, there is a discrepancy between the system of language teaching and the actual language skills of the children. In Soviet times, the majority of textbooks were created for children who had spoken the ethnic language from their childhood, while today almost all children have little or no knowledge of their ethnic language when they enter school. Today all indigenous communities still lack study material that would employ modern methods of teaching the ethnic languages as foreign ones. This problem is also relevant for those adult native speakers

514 Nadezhda Mamontova

and other individuals who wish to improve or revive their language skills. Free language courses are arranged irregularly and held mainly in administrative centers. The only sphere where all Tungusic languages are still used is in cultural and folklore activities. But languages that are only used in stage performances tend to turn to elements of folklore themselves, which only increases the marginalization of their communicative role. It still remains to be seen whether modern methods of active language revitalization can be successfully applied to the Tungusic languages. Much depends on whether a new generation of young speakers can be raised with the efforts of the ethnic communities themselves. Some encouraging developments may, however, be observed on the Internet, where, in particular, Ewenki, Ewen, and Nanai are today occasionally used by young and middle-aged people as a true means of communication. These languages are also more actively being utilized in posters, memos, and other visual material. REFERENCES AND FURTHER READING Aralova, Natalia & Brigitte Pakendorf (2018) ‘The endangered state of Negidal’, Language Documentation and Conservation 12: 1–14. Arefiev, A. L. [А. Л. Арефьев] (2014) Языки коренных малочисленных народов Севева, Сибири и Дальнего Востока в системе образования: История и современность [Languages of the indigenous minorities of the North, Siberia and the Far East in the educational system: Past and present], Москва [Moscow]: Центр социального прогнозирования и маркетинга. Bulatova, N. Ya. [Н. Я. Булатова] (1997) ‘Эвенкийский язык и его региональные варианты в социолингвистическом аспекте’ [The Ewenki language and its regional varieties from a sociolinguistic perspective], in: Малочисленные народы Севера, Сибири и Дальнего Востока: Проблемы сохранения и развития языков [The Small Peoples of the North: Problems of language preservation and development], 42–53, Санкт-Петербург [St. Petersburg]: ИЛИ РАН. Dyrkheeva, G. A. [Г. А. Дырхеева] (2014) ‘Языковые приоритеты республики Бурятия: система образования’ [Linguistic priorities in the Republic of Buryatia: The system of education], in: Языковая политика и языковые конфликты в современном мире [Language policies and language conflicts in the modern world], Moсква [Moscow]: ИЯ РАН, 101–107. Elaeva, I. È. [И. Э. Елаева] (2012) ‘Этничность, образовательная политика и языковая ситуация в республике Бурятия’ [Ethnicity, educational policies and the linguistic situation in the Republic of Buryatia], Власть 2012 (6): 145–155, Moсква [Moscow]: ИС РАН. Funk, Dmitry & Nadezhda Mamontova & Kirill Shakhovtsov & Aleksandra Teryokhina (2012) ‘On the mission for gathering information in view of preparation of research and recommendations to support endangered languages (at the example of the Evenki language), in Alexey Kozhemyakov & Sergey Sokolovskiy (eds.), European Charter for Regional or Minority Languages in the Russian Federation: Research, Reports, Recommendations Made in the Framework of the Joint Project “Minorities in Russia: Developing Languages, Culture, Media and Civil Society”, 2009–2012, Strasbourg: Council of Europe. Gerasimova, A. N. [А. Н. Герасимова] (2002) ‘Нанайский и ульчский языки в России: сравнительная характеристика социолингвистической ситуации’ [The Nanai and

Sociolinguistic aspects of Tungusic 515

Ulcha languages in Russia: A comparative survey of the sociolinguistic situation], Языки коренных народов Сибири 12: 246–257, Новосибирск [Novosibirsk]. Janhunen, Juha (1990) ‘Ethnic death and survival in the Soviet North’, Journal de la Société Finno-Ougrienne 83: 111–122, Helsinki. Janhunen, Juha (2005) ‘Tungusic: an endangered language family in Northeast Asia’, in: David Bradley (ed.) Language Endangerment in the Sinosphere, International Journal of the Sociology of Language 173: 37–54, Berlin. Kazakevich, O. A. [О. А. Казакевич] (2007) ‘Два поселка, два языка, две языковые ситуации’ [Two villages, two languages, two linguistic situations], Полевые исследования студентов РГГУ: этнология, фольклористика, лингвистика, религиоведение 2: 324–330, Moсква [Moscow]. Kazakevich, O. A. [О. А. Казакевич] (2011a) ‘На руинах автохтонных языков Томской области’ [On the ruins of the autochthonous languages of Tomsk Oblast’], in: Л. Д. Раднаева [L. D. Radnaevа] (ed.), Проблемы изучения и сохранения языков и культур народов России [Problems of the study and preservation of the languages and cultures of the peoples of Russia], 67–73, Санкт-Петербург [St. Petersburg]: Филологический факультет СПбГУ. Kazakevich, O. A. [О. А. Казакевич] & al. (2011b) ‘Таймырская (эвенкийско-энецконганасанско-долганская) экспедиция 2011 г.’ [The 2011 Taimyr (Ewenki-Enets-Nganasan-Dolgan) expedition], field report online: www.lcl.srcc.msu.ru/​grants_​ pages/expedition_2011.htm. Kirillova, A. I. (2015) ‘Scientific assessments of Bystrinski Evens socio-cultural development in 1990–2000s’, Asian Social Science 11 (6): 105–110. Lekhanova, F. M. [Ф. М. Леханова] & T. F. Aleksandrova [Т. Ф. Александрова] & K. I. Makarova [К. И. Макарова] & A. I. Maksimova [А. И. Максимова] & V. P. Marfusalova [В. П. Марфусалова] (2012) Эвэдыт таткалдярав: Начинаем изучать эвенкийский язык [We start learning Ewenki], Санкт-Петербург [St. Petersburg]: Просвещение. Mamontova, Nadezhda (2014) ‘What language do real Evenki speak? Discussions surrounding the nomad preschool’, Anthropology and Archeology of Eurasia 52 (4): 37–75, Taylor & Francis Online. Mamontova, Nadezhda (2016) ‘The sociolinguistic landscape of the island of Sakhalin’, in: Ekaterina Gruzdeva & Juha Janhunen (eds.), Crosslinguistics and Linguistic Crossings in Northeast Asia, Studia Orientalia 117: 227–247, Helsinki. Morozova, O. N. [O. H. Морозова] (2014) ‘Уровень владения эвенкийским языком в Зейском, Селемджинском и Мазановском районах Амурской области’ [The level of competence in the Ewenki language in the Zeya, Selemdzha and Mazanovо raions of Amur Oblast’], in: Обучение иностранному языку на современном этапе студентов высших и средних общеобразовательных учреждений [Current state of foreign language еducation of middle and high school-level students], 207–214, Благовещенск [Blagoveshchensk]: Амурский ГУ. Neroznak, V. P. [В. П. Нерознак] (ed.) (1994) Красная книга языков народов России: Энциклопедический словарь-справочник [The Red Book of the languages of the peoples of Russia: The encyclopaedia], Москва [Moscow]: “Аcademia”. Osipova, M. V. [М. В. Осипова,] (2013) ‘К проблеме сохранения родного языка и культуры ульчей—коренного малочисленного народа Хабаровского края’ [On the preservation of the language and culture of the Ulcha—the indigenous people of

516 Nadezhda Mamontova

Khabarovsk Krai], Фундаментальные и прикладные исследования: проблемы и результаты 2013 (3): 48–52, Новосибирск [Novosibirsk]. Procukovich, E. A. [E. A. Процукович] (2015) ‘Уровень владения эвенкийским языком в местах компактного проживания эвенков Амурской области’ [The level of competence in the Ewenki language in localities with a compact Ewenki population in Amur Oblast’], Теоретическая и прикладная лингвистика 1: 85–93, Благовещенск [Blagoveshchensk]. Sirina, A. A. [А. А. Сирина] (2002) Катангские эвенки в XX веке: расселение, организация среды жизнедеятельности [The Katanga Ewenki in the 20th century: distribution and organization of the social environment], 2nd revised edition, Иркутск [Irkutsk]: “Оттиск”. Sirina, A. A. [А. А. Сирина] (2012) ‘Иркутская область’ [Irkutsk Oblast’], in: Н. И. Новикова [N. I. Novikova] & Д. А. Функ [D. A. Funk] (eds.), Север и северяне: Современное положение коренных малочисленных народов Севера, Сибири и Дальнего Востока, 121–136, Москва [Moscow]: ИЭА РАН. Tarabukina, U. P. (2015) ‘The development of native languages in educational institutions in the conditions of the dispersed indigenous small-numbered peoples in the North, Siberia and the Russian Far East, in the Republic of Sakha (Yakutia) (on the example of the Even language)’, Journal of Siberian Federal University, Humanities & Social Sciences 6 (8) 2015: 1111–1120, Krasnoyarsk. Tarasova, Marina (1998) ‘Even language in the early stages of education’, in: Erich Kasten (ed.) Bicultural Education in the North: Ways of Preserving and Enhancing Indigenous Peoples’ Languages and Traditional Knowledge, 187–190, Münster: Waxmann Verlag. Turaev, V. A. [В. А. Тураев] (2004) ‘Охотские эвенки в ХХ веке: от этнокультурной эволюции к социальной деградации’ [The Okhotsk Ewen in the 20th century: from ethnocultural evolution to social degradation], in: Этнос и культура в условиях общественных трансформаций [Ethnicity and culture under conditions of social transformation], 3–36, Владивосток [Vladivostok]: “Дальнаука”. Turaev, V. A. [В. А. Тураев] (2006) ‘Ассимилятивные процессы у дальневосточных эвенков’ [Assimilational processes among the Ewenki in the Far East], Вестник ДВО РАН 2006 (3): 104–121, Владивосток [Vladivostok]. Ulturgasheva, Olga (2012) Narrating the Future in Siberia: Childhood, Adolescence and Autobiography among the Eveny, New York: Berghahn. Vakhtin, N. B. [Н. Б. Вахтин] (2001) Языки народов Севера в XX веке: Очерки языкового сдвига [The languages of the peoples of the North: Studies of the language shift], Санкт-Петербург [St. Petersburg]: “Дмитрий Буланин”. Zamorshhikova, L. S. [Заморщикова, Л. С.] & V. M. Toburokova [В.М. Тобурокова] & K. I. Fedorova [К. И. Федорова] & N. E. Zakharova [Н. Е. Захарова] (2010) ‘Языковая ситуация и региональные особенности русского языка в республике Саха (Якутия)’ [The linguistic situation and regional features of the Russian language in Yakutia], in: Arto Mustajoki & Ekaterina Protasova & Nikolay Vakhtin (eds.) Sociolinguistic Approaches to Non-Standard Russian, Slavica Helsingiensia 40: 94–105, Helsinki.

CHAPTER 20

TUNGUSIC IN TIME AND SPACE Juha Janhunen

The placing of the Tungusic language family in a temporal and spatial context is a complex task that requires proper consideration of both linguistic and extralinguistic facts. On the linguistic side, the basic underlying assumption is that all synchronically and historically attested varieties of Tungusic are ultimately descendants of an originally relatively uniform protolanguage which underwent diversification, apparently in several successive stages, resulting in the gradual crystallization of the individual branches, sub-branches, languages, and dialects. A relative chronology for this diversification can be established by considering the features uniting and separating the individual varieties with regard to each other. While some of the diversifying developments that took place in each given variety were linked to internal factors, an at least equally important role was played by external factors, conditioned by contacts with neighbouring non-Tungusic languages. If only the linguistic data were considered, neither internal processes nor external contacts would offer a firm basis for making conclusions about the underlying temporal and spatial setting. For an understanding of the latter, the linguistic data have to be linked to the information provided by extralinguistic realities such as cultural and physical anthropology, human genetics, oral literature, history, and archaeology. The relevance of these fields varies, however, and it is particularly hazardous to search for direct connections between languages, archaeological cultures, and human populations. This is simply because languages, cultures, and genes do not evolve or move in tandem. While, for instance, the method of linguistic palaeontology, using the shared corpus of cultural lexicon as a point of reference, provides a rough tool for assessing the cultural stage of the population that once spoke the common protolanguage, it does not allow the speakers of the protolanguage to be immediately identified with any particular prehistorical culture. Still less is it likely that genes would preserve evidence of linguistic movements or connections. The fact is that the principal mechanism of linguistic expansion is not migration and population replacement, nor large-scale population mixing, but, rather, language shift with relatively little gene flow. It may, nevertheless, be taken for certain that Proto-Tungusic was an actual natural language, spoken in a limited region by a coherent population at a fixed historical or prehistorical time level. Therefore, it is a relevant task to search for the original location and chronological framework of this protolanguage. For this the best basis is offered by the linguistic and geographical context of the modern and historical Tungusic languages, that is, the physical features of the regions where they are or were spoken, as well as the languages that are spoken, or are known to have been spoken, in their neighbourhood. Looking at the distribution of the Tungusic languages on the map, we can see that the DOI: 10.4324/9781315728391-20

518 JUHA JANHUNEN

majority of them are spoken, or used to be spoken, in the sphere of continental Manchuria, comprising, above all, the Sungari-Amur basin, but including also the Liao basin, and extending to the northern limits of the Korean Peninsula in the south and the coast of the Sea of Japan in the east. The language families that surround Tungusic in this region include Mongolic in the west, Koreanic in the south, and Amuric in the north. It is reasonable to look for the Tungusic homeland against this very linguistic and geographical context. A rough dating for Proto-Tungusic can also be obtained from an assessment of the size of the family-internal comparative corpus, as compared with other language families that have a more reliably established chronology. Preliminary glottochronological calculations, though not necessarily carried out with proper etymological discretion, have typically given for Proto-Tungusic depths between 1500 and 2000 years. Perhaps not accidentally, this time range, corresponding to the conventional concept of Iron Age, is compatible with what is known of the extralinguistic developments in the presumed homeland region. It should be noted that Manchuria is covered, though only marginally, by the early historical sources of both China (since the 1st millennium BZ) and Korea (since the middle of the 1st millennium AZ). Although the information supplied by these sources is vague as far as linguistic and ethnic identities are concerned, the general protohistorical political setting of the region is likely to contain a key to the understanding of the underlying linguistic situation both with regard to Tungusic and the other languages in the region. The protohistorical polities that have either direct or indirect relevance to the history of Tungusic include the so-called “Three Kingdoms” of Korea and southern Manchuria, that is, Goguryeo, Baekje, and Silla, as well as their northern predecessor Buyeo in central Manchuria. While the conclusion concerning the former existence and subsequent diversification of Proto-Tungusic is based on factual evidence deduced from the comparative material by the methods of diachronic linguistics, it has been considerably more difficult to find generally acceptable factual evidence in support of any deeper-level affinities between Tungusic and other language families. Proposals concerning the possibility of such affinities have been made in several different frameworks, on different chronological levels, and under various names, including “Altaic”, “Transeurasian”, “Ural-Altaic”, “Nostratic”, and “Eurasiatic”, but the common trait for all of them is that they are based on a teleological interpretation of the data and are carried out without a sufficiently sophisticated methodology. This means that the information available so far on time levels beyond Proto-Tungusic is confined to what is provided by the method of internal reconstruction, as well as by the analysis of the traces of external contacts preceding the level of the protolanguage. DATA AND SOURCES The question concerning the prehistory and origin of the Tungusic languages has most often been discussed in the context of the so-called Altaic Hypothesis, first formulated as a coherent doctrine by G. J. Ramstedt (1952–1965), followed by Nicholas Poppe (1960, 1965, 1975). Expanded versions of this hypothesis have been propagated, among others, by Roy Andrew Miller (1971) and S. A. Starostin (1975), as well as, under the name “Transeurasian”, by Martine Robbeets (2005, cf. also Robbeets & Savelyev, eds. 2020). An attempt to prove the Altaic Hypothesis by a massive lexical corpus was made by Starostin & al. (2003), while visions concerning the even more macroscopic frameworks of

TUNGUSIC IN TIME AND SPACE 519

Nostratic and Eurasiatic have been supported by Aharon Dolgopolsky (1998) and Joseph F. Greenberg (2000–2002). The serious problems inherent in all these proposals have been pointed out by several scholars, as summarized by Janhunen (2023). The prospects of the binary comparison of Tungusic with other “Altaic” languages have also been studied, with varying conclusions, by Lee Ki-moon (1958) and Kim Dongso (1980) for Korean, Tsumagari Toshirō (2003) for Japanese, as well as Janhunen (1996b) for Mongolic. For the understanding of the areal position of Tungusic in Eurasia, the contributions of Robert Austerlitz (1970, 1980) provide an important background. The actual contacts of Tungusic with other languages and language families have been widely discussed in various contexts, as summarized by Janhunen (2013). A problem in this connection is that some of the contact relationships have a long history, extending from prehistorical times up to the present day, and it is not always easy to distinguish early contact features, dating back to Pre-Proto-Tungusic time levels, from later ones, involving the individual Tungusic branches and languages. Moreover, some parallels shared by Tungusic with other language families are of such a generic and diffuse nature as to have served as a basis for speculations concerning possible long-range affiliations. This is particularly true of the comparisons made between Tungusic and Uralic, as discussed by Denis Sinor (1975) and István Futaky (1988), though traces of actual contacts involving the individual branches of Uralic, especially Samoyedic and Khantic, have also been studied, as by Futaky (1975, 1983), Joki (1976, 1977), Helimski (1985), Anikin & Helimski (2007), and Gusev (2020). In a general framework, the contacts between Samoyedic and Tungusic have been discussed by Janhunen (1977). The most important contact partner of Tungusic has always been Mongolic. The fact that there are Mongolic loanwords in Tungusic has also been recognized by the proponents of the Altaic Hypothesis, as by Poppe (1966). The most careful analysis of this topic was carried out by Gerhard Doerfer (1984, 1985), who listed all the lexical parallels shared by Mongolic and at least two branches of Tungusic, cf. also the uncompleted list of Shherbak (posthumously 2020). However, there are still many open questions especially concerning the early contacts between Mongolic and Jurchen-Manchu. It is generally acknowledged that Manchu is the Tungusic language that has been under the most intensive Mongolic influence, as was noted already by G. D. Sanzheev (1930) and later elaborated in more detail by William Rozycki (1994). On the Mongolic side, this influence has also involved Para-Mongolic idioms, as discussed by Ligeti (1960), Okada (1962), Kuz’menkov (1988), and Janhunen (2015). Apart from the lexicon, the contact has involved the borrowing of grammatical elements, on which cf. Cincius (1977), as well as, especially in the more recent history of Manchu, interaction in the fields of syntax and morphosyntax, as analysed by Kyōko Maezono (1992, 2007ab). The interaction between Tungusic and Mongolic continues today at a local level between Solon and Daghur, as well as between Khamnigan Ewenki and Khamnigan Mongol, two pairs of languages whose relationship involves not only areal contact, but also, in the first place, ethnic bilingualism, in which virtually all speakers of Solon are bilingual in Daghur, while all speakers of Khamnigan Ewenki are bilingual in Khamnigan Mongol. A somewhat similar situation exists also between several groups of Ewen and Ewenki speakers with regard to Yakut, though, in general, contacts between Tungusic and Turkic are historically marginal and secondary, as also in the case of Sibe, which incorporates some recent areal influences from the neighbouring Central Asian Turkic languages. The same is true of the contacts with Russian, which, apart from becoming the dominant language of most Tungusic speakers in Siberia, has also received loanwords

520 JUHA JANHUNEN

from Tungusic, especially Ewenki and Ewen, as discussed by Anikin (1990). Chinese has, at least potentially, had much older contacts with Tungusic, especially with Jurchen-Manchu, but most of the lexical borrowings from Chinese to Tungusic, as listed by Peter Schmidt for Manchu (1932–1933) and Gianni Coati (1990) for Amur Tungusic, are nevertheless of a relatively recent origin. Besides Mongolic, another ancient contact partner of Tungusic has been Amuric, today represented by the modern varieties of Ghilyak (Nivkh and Nighvng), spoken on the Lower Amur and on Sakhalin. The interaction between Tungusic and Amuric seems to have involved both structural and lexical aspects, as discussed by Kreinovich (1955), Panfilov (1973), Pevnov (1992), Janhunen (2016), Gruzdeva & Janhunen (2020), and Knapen (2021). By comparison, contacts with the other “Palaeo-Asiatic” language families, including Yeniseic, Kolymic, Kamchukotic, and Kurilic, have been more marginal. Typically, these contacts have involved relatively late bilateral relations of one or several non-Tungusic languages with one particular Tungusic language: Ket and Yugh (Yeniseic) with Ewenki in the northwest, Forest and Tundra Yukaghir (Kolymic) and Chukchee-Koryak (Kamchukotic) with Ewen in the northeast, and Sakhalin Ainu (Kurilic) with Uilta in the southeast. For a discussion of relevant issues in the light of selected lexical case studies, see, for instance, Ikegami (1980, 1994), Tsumagari (2010), Georg (2008), and Khabtagaeva (2019). Information concerning the general ethnohistorical situation in Manchuria and adjacent regions is available in Janhunen (1996) and Shimunek (2017). The linguistic identity of the “Three Kingdoms” of Korea has been a source of wide differences of opinion. An important circumstance in this connection is that Japonic was also involved, though the conventional assumption of Japonic, or Para-Japonic, as the language of Goguryeo and even of Buyeo, as defended by Chris Beckwith (2004) and Andrew Shimunek (2021), has been challenged on good grounds by Toh Soo Hee (2005, cf. also Janhunen 2005). On the other hand, Koreanic, which is known to have been the language of Silla, has by some scholars, as by Alexander Vovin (2006ab, 2007, 2013), been assumed to have started its expansion from central Manchuria. Even so, it is a generally accepted fact that southern Manchuria was historically divided between Mongolic, or Para-Mongolic, in the west, and Tungusic, or Jurchenic, in the east, with the basic division line being formed by the Liao basin, as summarized in Janhunen (2008, 2012a). Toponymic evidence concerning the early presence of Amuric in the Sungari basin and the late arrival of Tungusic to the Yenisei basin has been presented by Janhunen (2012b, 2022). A curious hypothesis, based on the so-called Buyla or Nagyszentmiklós inscription, suggesting an early expansion of Tungusic across the Eurasian steppe to the west in the context of the Avar khaganate, was presented by Eugene Helimski (2000, 2003), but the idea has been correctly refuted by Alonso de la Fuente (2015). EXTERNAL RELATIONSHIPS According to the Altaic Hypothesis, Tungusic is one of the three language families belonging to the so-called “Micro-Altaic” or “Core Altaic” group, comprising also Mongolic and Turkic. In the expanded version of the hypothesis, known as “Macro-Altaic”, Koreanic and Japonic are added, with the underlying assumption that all these language families descend from a common “Proto-Altaic” ancestor. It is, indeed, true that the “Altaic” languages share a number of properties that allow them to be viewed as a coherent complex. These properties are, however, of a structural character and comprise also

TUNGUSIC IN TIME AND SPACE 521

the Uralic languages, which means that the proper context of Tungusic is not “Altaic”, but “Ural-Altaic”. Opinions are divided as to whether structural similarities are sufficient to justify the postulation of shared ancestry, but the majority of scholars today seem to favour the conclusion that both Altaic and Ural-Altaic are to be seen as areal phenomena, deriving from prolonged contacts, which have mainly taken place along a chain of bilateral relations between Uralic and Turkic, Turkic and Mongolic, Mongolic and Tungusic, Tungusic and Koreanic, and Koreanic and Japonic. Direct contacts between other pairs of families, such as Uralic and Mongolic, Uralic and Tungusic, Turkic and Tungusic, and Tungusic and Japonic, have been more marginal and are typically connected with recent interaction between individual languages. An important fault line within the Altaic complex separates Koreanic and Japonic from Turkic, Mongolic, and Tungusic. Not only are Koreanic and Japonic in some respects less “Altaic” typologically, but, in spite of attempts to show the contrary, they also lack any significant number of material similarities with the three Micro-Altaic families, or with each other. In the context of the Altaic Hypothesis, this is “explained” in terms of a family tree which divides “Proto-Altaic” into two primary branches, the one leading to Koreanic and Japonic, and the other to Turkic, Mongolic, and Tungusic. Because of the diffuse relation between these “branches”, Macro-Altaic is sometimes characterized as a “macro-family” or “phylum”. By contrast, Micro-Altaic involves a relatively large corpus of shared lexicon. Conventionally, this corpus is taken as evidence of a genealogical connection between the families concerned, especially as some of the shared items show regular correspondences reminiscent of sound laws. However, the critics of the Altaic Hypothesis have demonstrated the invalidity of this conclusion. For one thing, the shared lexicon does not comprise items of basic vocabulary, with the exception of a few loanwords. Also, the lexical items are not shared evenly between Turkic, Mongolic, and Tungusic, but in pairs comprising Turkic and Mongolic, on the one hand, and Mongolic and Tungusic, on the other, leaving no corpus of items shared by Turkic and Tungusic only. This means that any items shared by all the three families are results of a flow of lexical influence from Turkic to Mongolic to Tungusic. Finally, the regular correspondences, although an undeniable reality, are better explained in an areal framework. A crucial role in the areal contacts between Turkic and Mongolic was played by Bulgharic, or the Bulghar Turkic branch of Macro-Turkic, which yielded the first significant layer of Turkic items to Mongolic. To take just one example, the phenomenon of the so-called rhotacism-lambdacism, which involves a correspondence between the sibilants *z [z] and *š [ʃ] with the liquids *r and *l, respectively, was originally based on an innovation which took place in the Bulgharic branch of Macro-Turkic and changed the original sibilants *s and *š to the liquids *r and *l under certain positional conditions. The rhotacist-lambdacist representation is still observed in Chuvash, the only surviving descendant of Bulgharic, while the rest of Turkic, also known as Micro-Turkic or Common Turkic, shows *z and *š. Lexical items that involve these correspondences between Turkic and Mongolic are borrowings from Pre-Proto-Bulgharic to Pre-Proto-Mongolic, and some of these items were further transmitted from various stages of Mongolic to the individual Tungusic branches and languages. In most cases we are dealing with typical cultural vocabulary, as in Common Turkic *öküz ‘ox’ Pre-Proto-Bulgharic *pökür (>> Chuvash văxxăr) → Pre-Proto-Mongolic (with irregular and variable vowels) *püker > Proto-Mongolic *xüker (> Modern Mongolian uxer) ~ *xökör → Ewenic *xökör (> Ewenki xukur > ukur).

522 JUHA JANHUNEN

It may be concluded that the lexical comparisons between Tungusic and the other “Altaic” languages have been carried out without proper consideration of the established methods of historical and comparative linguistics. This is nowhere as evident as in the fruitless attempts that have been made to relate Tungusic lexical items to Koreanic or Japonic “cognates” in the context of Macro-Altaic comparisons. However, it is surprising that even areal comparisons of Tungusic lexical items with Koreanic and/or Japonic have yielded only minimal results, suggesting that the contact was never particularly intensive, although the seeming lack of material parallels does not rule out the possibility of structural interaction. As it is, there are only a few apparently early continental loanwords in Japonic with cognates in Tungusic, but most of them are also present in Mongolic, and it is difficult to identify their ultimate source, as in the case of Japanese (*)kutu ‘shoe’ = Mongolic *gutu.l id. = Manchu gūlha id. and Japanese (*)siru ‘soup, juice’ = Mongolic *silö/n id. = Tungusic *sile/n id. A possible candidate for a lexical parallel shared by only Japonic and Tungusic is the numeral for ‘seven’, Japonic (*)nana = Tungusic (*)nada/n, but an isolated example like this is open to alternative explanations, including chance similarity. Traces of early lexical interaction between Tungusic and Korean, if they exist, await to be identified. The only language family that would still seem to offer prospects for genealogical comparisons with Tungusic is, not surprisingly, Mongolic. Although most of the similarities between the two families are results of areal interaction deriving from their geographical adjacency, there are also a few items of basic vocabulary, as well as some grammatical elements, including auxiliaries and pronouns, that might imply a common origin. At the same time, it has to be noted that the bulk of their basic vocabulary, including terms for body parts and bodily functions, items for natural environment, as well as numerals, are different, suggesting that their relationship, if real, must be very remote. Moreover, some of their shared features have a wider distribution in Eurasia, which means that their value for comparative purposes is seriously reduced. Even so, it is instructive to review some of the relevant parallels and determine what their potential value is. •

•

The perhaps most striking parallel is observed in the class suffixes for homogeneous (ucountable) masses and individualizable (countable) entities. The Proto-Tungusic suffixes for these functions are *-sA and *-tA, respectively, often expanded to *-GsA and *-G-tA, as in *see-k.se ‘blood’, *xüü-k.te ‘tooth’. For the same functions, Pre-Proto-Mongolic used the suffixes *-s- and *-d-, respectively, as in *cï-s-u/n ‘blood’, *si-d-ü/n ‘tooth’. It is true that on the Mongolic side the suffixes are synchronically less transparent and reliably recoverable only by internal reconstruction, while in Tungusic they are more widely used and can even retain limited productiveness in some contexts. Even so, the parallelism is non-trivial and would be difficult to explain by borrowing. Among independent lexemes with grammatical functions, the most obvious parallel is present in the singular personal pronouns, which are in Tungusic SG1 *bi : OBL *min-, SG2 *si : OBL *sin-, SG3 *i : OBL *in- and in Mongolic SG1 *bi : OBL *min-, SG2 *ci : OBL cin-, SG3 *i- : OBL *in-. The second person forms are probably secondary in both languages, for in Mongolic *ci < *ti, as is confirmed by the corresponding plural pronoun 2PL *ta, while in Tungusic, also, the original form is likely to have been *ti, which is the shape presupposed by the inclusive forms 1PL.INCL *bi+*ti : *mi/n+*ti ~ *bö+*ti, as well as by the corresponding inclusive person markers. However, the relevance of the personal pronouns is diminished by

TUNGUSIC IN TIME AND SPACE 523

•

•

•

•

the fact that closely similar pronominal roots are also attested in several other language families of Eurasia, including Turkic, Uralic, Kolymic, Kamchukotic, and even Indo-European—a circumstance that has nourished the hypotheses concerning an ancient Nostratic or Eurasiatic language family or phylum. It is obvious that this similarity has been conditioned by factors other than common ancestry. On the other hand, it may be argued that the parallelism between Tungusic and Mongolic comprises also the oblique forms, but even these are not unique, for similar oblique forms, extended by what has been called the “pronominal n”, are also attested in, at least, Turkic and Uralic. Another case of pronominal parallelism is present in the demonstrative roots, which are in Tungusic PROX *e- vs. DIST *ta- and in Mongolic PROX *e- vs. DIST *te-. Again, similar roots are present in several other language families of Eurasia, including, in particular, Uralic. However, Tungusic and Mongolic also seem to share a pronominal extension element with the vibrant *-r-, as attested in Tungusic PROC *e-ri vs. DIST *ta-rï and Mongolic DIST *te-r-e. Although it is unlikely that these pronouns would involve direct borrowing, it has to be noted that the Manchu form of the distal pronoun tere (tere) is either borrowed from or secondarily influenced by its Mongolic counterpart. A less trivial parallelism is present in the reflexive pronouns. The independent forms of the reflexive pronoun in Tungusic are SG *mee/n : PL *mee-r, but the corresponding reflexive suffixes RX.SG *-bII < *-bA-(y)I : RX.PL *-bA-r-I suggest that the original forms of the independent pronouns were SG *bee/n : PL *bee-r. These forms are closely parallelled by the Mongolic reflexive suffix *-xAn < *-pAn (Written Mongol bav ~ ijav ~ -qhav) and reflexive pronoun *öxer < *ö-pe-r (Written Mongol vuibar, suggesting the reading †öber). Although the segmentation of the Mongolic pronoun is uncertain, and although it does not synchronically imply plurality, the comparison with Tungusic might suggest that the element *-r in *ö-pe-r represents the same plural marker as is present in the Tungusic plural forms. Supposing that this is so, plural formation of the type SG */n : PL *-r would be a feature common to Tungusic and Mongolic. Tentatively, other potential plural forms in *-r could also be identified in Mongolic, cf. e.g. Mongolic LOC emü/n-e (> *ömü/n-e) ‘in front of’ : NOM *emü/n < *ebü/n : PL *ebü-r (> *öbür) ‘front’ (< ? ‘breast/s’). The case systems of Tungusic and Mongolic show very few, if any, primary material similarities. The only potential parallelism would seem to be present in the directive suffixes for spatials, which are *-sI in Tungusic, as in *xama-sï ‘backwards’, and *-g-sI in Mongolic, as in *ïna(-xa)-g-sï (> *naa-si) ‘hither’. An adverbial suffix like this could, in principle, be borrowed, but its restricted use in a small number of lexicalized spatials, including pronominal spatials, would rather not favour the assumption of borrowing. An important auxiliary morpheme that would seem to be shared by Tungusic and Mongolic is the negation verb *e-. In Tungusic, this functions morphologically as a regular verb which in the aorist has the stem *e-si-, and which is combined with the invariable connegative form (also identical with the aorist stem) of the semantic main verb, as in Ewenki e-si-m saa-ra ‘I do not know’ (< *e-si-n-bi saa-ra). In Mongolic, the independent form *ese appears as one of the two standard negation particles, used preverbally to negate fully inflected verbs in finite position. However, *ese itself can also be inflected in echo constructions of the type *ire-kü ese-kü (> Modern Mongolian ir-ex es-ex) ‘to come or not to come’ (with both verbs in the form

524 JUHA JANHUNEN

•

•

•

of the futuritive participle). Moreover, it seems that the Mongolic negation particle is also attested in suffixal position in a few verbs denoting privation, as in *öl ‘food’ : *öl-ö.s- < *öl-es- ‘to be without food’ > ‘to be hungry’. The verbal properties of Mongolic *ese would favour the assumption that it is not a borrowing, but a true cognate of Tungusic *e-si-. However, it has to be noted that, once again, there is a material parallel elsewhere in Eurasia, in that Uralic also has a negation verb with the form *e-, combined with an invariable connegative form of the main verb. Nevertheless, the element of stem extension *-sV- might represent shared heritage of Tungusic and Mongolic, even though it has lost any synchronic function in Mongolic. Another auxiliary morpheme showing a material parallel between Tungusic and Mongolic is the copula-existential, Tungusic *bi- : AOR *bi-si- ‘to be’ vs. Mongolic *bü- id. The Mongolic item is attested in a number of relict forms, such as CONF bü-lüxe (> Modern Mongolian bilee) : PTCP.FUT *bü-küi, including also the petrified copular particle *bü-i (> Modern Mongolian bii, Written Mongol bui). A problem is that, in this case, the element of stem extension *-sV- is not present in Mongolian. Also, copula-existential verbs with the shape *bV- or *pV- are attested in other language families of Eurasia, including Amuric and even Indo-European (English be). The fact is that the tendency to express simple meanings with simple forms may also produce accidental lookalikes. There are also several spatial roots that show material similarities between Tungusic and Mongolic. The most interesting case is offered by the root *deg- ‘upper side’, which in Tungusic is used as a verb in the meaning ‘to rise, to fly up’ (from which *deg-ii ‘bird’), but which in Mongolic appears in inflected forms of the lexicalized spatial noun *dege- > *dexe-, as in LOC *dexe.r-e ‘above, upon’ : DIR *dexe-g.si ‘upwards’. Similarity is also present between Tungusic *doo ‘inside’ (the source of the dative case marker *-dOO) and Mongolic *do.ta- ~ *do.ca- id. : LOC *do.ta-r-a ~*do.ca-x-a, suggesting a primary root *do- ‘inside’, as well as between Tungusic *du.lï/n ‘middle’ and Mongolic *du.lï ~ *dü.li id., LOC *du.m.d-a ‘(in the) middle’, suggesting a primary root *du- ‘middle’. A pronominal spatial is involved in the parallel between Tungusic *caa- ‘further away; the other side’ and Mongolic *ca- : DX-LOC *ca-xa-n-a ‘behind, beyond, over there’. Among non-grammatical basic vocabulary items, two relatively convincing candidates for true cognates are Tungusic *moo ‘tree, wood, forest’ ~ Mongolic *mo-du/n id. (with the class marker *-d- for countables) and Tungusic *umï- ‘to drink’ ~ Mongolic *um- id. in *um-da-xa/n < *um-da-ga/n ‘drink’ (containing what looks like the passive suffix *-dA- and the nominalizer *-xA/n < *-gA/n, i.e., ‘something to be drunk’). Altogether there are around 50 lexical parallels that would seem to date back to the protolanguages of both families, and for which there is no formal reason to assume borrowing, nor any criterion to determine the direction of borrowing. However, most of these items are culturally or geographically specific and therefore likely to involve areal contact. Incidentally, it has been assumed that items relating to fishing or the maritime environment in Mongolic might be of a Tungusic origin. This is, however, difficult to verify, as both language families seem to have evolved in the vicinity of rivers and the seaside, and there is no reason to project the well-known steppe culture of the historical Mongols upon their linguistic ancestors, who did not live in the steppe belt. Therefore, items like Tungusic türe-k.se ‘fish eggs’ ~ Mongolic *türi-s.ü/n id. (which also share the class markers *-k-sA = *-s- for uncountables) remain ambiguous: they can involve borrowing in either direction, but they could

TUNGUSIC IN TIME AND SPACE 525

also be taken as evidence of common heritage. The possibility that this item can involve borrowing is suggested by the fact that it is also attested in Samoyedic with the Proto-Samoyedic shape *türä-mä ~ *tirä-mä ‘fish eggs’, though the ultimate source and route of borrowing are unknown. Obviously, the question concerning the possibility of a genealogical connection between Tungusic and Mongolic must remain open for the time being. If a reality, the underlying language family, which has also been called “Khinganic”, would have to be very deep, perhaps deriving from the context of the archaeologically established Neolithic Hongshan (紅山) culture in the Liao basin (early 5th to early 3rd millennium BZ). In this connection, however, it has to be recalled that, irrespective of whether they share a common genealogical background or not, Tungusic and Mongolic have been interacting since ancient times with both lexical and grammatical borrowing as a result, and many elements conventionally considered to be inherited from a common protolanguage are actually borrowings. It should not be expected that true cognates show perfect matches in both form and function. For instance, the Tungusic verbal voice suffix expressing reciprocal or cooperative action *-ldU- (> Ewenki -ldi- ~ -ldii-, Udihe -gdi-, Nanai -ltO-) is identical with Mongolic *-ldU- and must be a relatively old borrowing from the latter. Modern examples of intensive contact or ethnic bilingualism between Tungusic and Mongolic, as between Khamnigan Mongol and Khamnigan Ewenki, provide many similar examples of suffixal borrowing, though the synchronic aspects of the issue would certainly deserve to be studied in more detail. AREAL POSITION The general areal position of Tungusic is best defined by its rather systematic adherence to what is known as the “Altaic typological sphere”. This means that it shares most of its typological properties with its “Altaic” or “Ural-Altaic” neighbours, including, especially, Mongolic, but also Turkic and Uralic, as well as, to a somewhat lesser extent, Koreanic and Japonic. In Eurasia, Altaic typology may originally be viewed as a phenomenon of continental Northeast Asia, as opposed to Sinitic typology, dominant in Southeast Asia, and the so-called Pacific Rim typology—or typologies—represented by the genetically and typologically diversified “Palaeo-Asiatic” languages spoken mainly in maritime Northeast Asia and including, in the immediate neighbourhood of Tungusic, the Amuric (Ghilyak), Kurilic (Ainu), and Kolymic (Yukaghir) families. There has been interaction between all these typologies, leading to the spread of Altaic features to languages with a basically non-Altaic typology, and to the loss of Altaic features in favour of non-Altaic ones in languages with a basically Altaic typology, two opposite processes known as “Altaicization” and “de-Altaicization”, respectively. In the Manchurian context, the syllable structure of Udihe, which involves contrasts reminiscent of tones, may possibly be seen as a Sinitic feature, while Ghilyak is a language that has adopted several Altaic features from its Tungusic neighbours. Typological features that link Tungusic with its “Altaic” neighbours are evident in all areas of grammatical structure, including phonology, morphophonology, morphology, morphosyntax, and syntax. Some of the features shared with Mongolic could, in principle, also be explained as common heritage in the context of “Khinganic”, but most of them, if not all, are more likely to be due to secondary areal interaction. For instance, the segmental paradigms of both vowels and consonants in Proto-Tungusic and Proto-Mongolic

526 JUHA JANHUNEN

were virtually identical, but a closely similar paradigm was also present in Pre-Proto-Turkic, as well as, at least as far as the vowels are concerned, in Pre-Proto-Koreanic. The phenomenon of vowel harmony is an obvious Northeast Asian areal feature, shared by all “Ural-Altaic” languages except Japonic, but also by Amuric and Kamchukotic. The dual manifestation of this feature as a palatal-velar harmony (PVH) in the west and as a tongue-root harmony (TRH) in the east likewise has an areal basis, but it involves additionally a diachronic aspect, in that tongue-root harmony has been expanding from east to west, ultimately replacing palatal-velar harmony in most of Tungusic and Mongolic, though not in Turkic or Uralic. There are other areal phenomena, especially in historical phonology, which have affected the individual Tungusic branches and languages. A particularly widespread sound change in Northeast Asia is the development of the strong labial stop *p [p ph] to the labial fricative *f [ɸ f] and further to the velar or laryngeal fricative or spirant *x [x h], with a complete loss as the final result. In Tungusic, the development *p > *x in initial position is shared by Ewenic and Orochic and must have taken place in the ancestral form of Northern Tungusic, while the parallel development *p > f is present in Manchu and dialectally in Nanai (as a secondary phonetic trend also in Udihe). The same development is, however, attested, at different time levels, in Pre-Proto-Turkic and Pre-Proto-Mongolic, as well as in Japanese and two Samoyedic languages (Mator, Nganasan). Although the succession [ph > ɸ > x > h > Ø] is phonetically trivial and attested in many languages having no connection with each other, it is likely that this change in Tungusic was triggered by areal factors. It seems that this was an early innovation spreading in the west-toeast direction from Turkic to Mongolic to Tungusic. To the east of this innovation there is a coherent area, comprising Koreanic and Amuric, as well as most of Nanaic, where *p in initial position is preserved intact, though in medial intervocalic position spirantization has taken place also in these languages. A very similar, but later areal development covering several languages in large parts of Siberia is the desibilization or “debuccalization” of the sibilant *s [s] to (*)x [x h]. This development is attested, with varying positional restrictions and dialectal exceptions in Ewenki, Ewen, Orochen, Oroch, and Udihe, but normally not in Neghidal, Solon, Nanaic or Manchu, although there is the possibility that synchronic x in Nanaic and Manchu can represent the original cluster *ks. Elsewhere in Siberia the analogous development *s > x [x h] is present as a relatively recent feature in Yakut (Turkic) and Buryat (Mongolic), as well as in a few Samoyedic languages (Forest Nenets, Selkup dialectal), suggesting that it was conditioned by substratal and adstratal factors whose ultimate origin remains difficult to specify. Some of the languages that do not show this development, including Neghidal, Solon, and Manchu, but also Orochen, have tended to fricativize *k to (*)x [x h], a feature also shared by several Mongolic languages, including Buryat. It may be noted that most Buryat dialects keep the reflexes of *k and *s distinct, in that the former yields a velar and the latter a laryngeal segment, i.e. /x/ vs. /h/. However, no Tungusic variety is known to have more than one velar to laryngeal fricative, which, then, can represent any one or two of the original segments *p *s *k, or also, in Nanaic, the original primary *x of Proto-Tungusic. One of the most interesting issues of the areal typology of Tungusic concerns its position with regard to Amuric, represented synchronically by the modern varieties of Ghilyak (Nivkh and Nighvng). Ghilyak is in many respects very different from Tungusic. Some of the differences seem to be due to relatively recent idiosyncratic developments in Ghilyak itself, but others are more fundamental. For instance, the direct object, which

TUNGUSIC IN TIME AND SPACE 527

in Tungusic is marked by the object cases (accusative and partitive-designative), is in Ghilyak expressed by the incorporation or “synthetization” of the object nominal into the verb, accompanied by morphophonological changes in the initial segment of the verbal root, a unique feature in the Manchurian context. Also, words with adjectival meanings (property words), which in Tungusic are nominals, are in Ghilyak verbals (static verbs), an inherently “non-Altaic” feature which, incidentally, is shared by Koreanic, suggesting that there may be an ancient “non-Altaic” areal link between Amuric and Koreanic. On the material side, Amuric shares a considerable number of lexical items with Tungusic, transmitted as loanwords in both directions. Thus, some items have a confirmed Tungusic etymology, as in Ghilyak ñ-řaƞq ‘(one) hundred’ (< *ñi+taƞq ‘one hundred’) ← Jurchenic taƞgu (Manchu tanggū, possibly transmitted to Ghilyak via Amur Tungusic) < Proto-Tungusic *taƞu ‘count’ : *taƞu- ‘to count’ (> Ewenki taƞi- id.), while others are more likely to derive from Amuric, e.g. Tungusic *sama(a)n ‘shaman’ (with no internal etymology) ← Amuric *saman (> Ghilyak camƞ ‘shaman’ = ‘eagle’, suggesting a semantic transition from ‘eagle’ to ‘shaman’). In some cases, Ghilyak seems to preserve the original root of a word, while Tungusic has borrowed only derivatives, as in Ghilyak la ‘sea, the Amur river’ > ‘wind from the sea, wind from the east’ (attested only in Amuric) : DX *la-mo ‘water basin’ (not attested synchronically in Amuric) → Tungusic *laamo ‘sea’ (> Ewenki laamu, Manchu namu id.) : DX-DX *la-mo-s ‘wind from the east’ (> Ghilyak lams id.) → Tungusic *laamos id. (> Ewenki laamus id.). However, there are also items for which there are no internal criteria to determine the direction of borrowing. On the Tungusic side, many of these items are specific to Manchu and Amur Tungusic, and it is possible that they represent traces of ancient cultural vocabulary of the Manchurian region, cf. e.g. Manchu aisin (aisi/n, also in Amur Tungusic) ‘gold’ = Amuric *aysin id. (> Ghilyak ays/ƞ), Manchu ulgiyan (ulgya/n, also in Amur Tungusic) ‘pig’ = Amuric *ulgan id. (> Ghilyak olʁoƞ). For items connected with maritime environment, especially terms for marine mammals, an Amuric origin is often assumed and can in some cases be verified, as in Ewenki kieƞa-s ‘dolphin skin’ : Oroch-Udihe käƞa ‘small whale, dolphin’ ← Amuric *kéƞa id. (< *kï+ƞa > Nivkh kéƞ), containing Amuric (*)+ƞa ‘animal’. PROTOHISTORICAL SETTING The political situation in protohistorical Northeast Asia started evolving with the coexistence and interaction of the two “tribal confederations” Xiongnu (匈奴) and Xianbei (鮮 卑), known from Chinese historical sources as approximate contemporaries of the Han (漢) dynasty of China (BZ 206–220 AZ). The Xiongnu were based in Mongolia, while the Xianbei were based in southwestern Manchuria, a regional difference that may have correlated with differences in ecological adaptation and social traditions, though both confederations were internally diversified and engaged in several fields, including agriculture, cattle breeding, warfare. and commerce. Because of their large size and diffuse nature, both are also likely to have comprised speakers of several languages. Even so, it is reasonable to assume that confederations of this type had one dominant language. Since no unambiguous information is preserved of these languages, they are best identified by circumstantial evidence, which would seem to link them with the first major wave of Turkic loanwords in Mongolic, that is, the period of interaction between Pre-ProtoBulgharic and Pre-Proto-Mongolic. If this is so, the dominant language of the Xiongnu would have been Pre-Proto-Bulgharic, while the Xianbei were dominated by speakers of Pre-Proto-Mongolic.

528 JUHA JANHUNEN

The identification of the Xiongnu and Xianbei with the Turkic and Mongolic language families, respectively, is supported by subsequent developments in Mongolia and Manchuria, during which Mongolia was ruled by the Kök Türk (546–744) and Uighur (744– 840) khaganates, dominated by speakers of Common Turkic, a lineage related to, though not identical with, Bulgharic, while southwestern Manchuria became the location of the Northern Wei (北魏 Bei Wei) of the Para-Mongolic-speaking Tabghach (386–534), and, later, of the Liao (遼) empire of the likewise Para-Mongolic-speaking Khitan (907–1125). This allows the historical eastern border of Mongolic to be placed, roughly, in the Liao basin. Considering that the Liao basin was also the western border of the Jin (金) empire of the Jurchen (1115–1234), who were confirmed Tungusic speakers and contemporaries of the Khitan, it is reasonable to assume that the local predecessors of the Jin empire in southeastern Manchuria, including Goguryeo (高句麗 Gaogouli, BZ 37–668 AD) and Barhae (渤海 Bohai, 698–926), had also been dominated by Tungusic speakers. The assumption that Tungusic was the dominant language of Goguryeo is congruent with the linguistic evidence which points to a prolonged contact between Mongolic and Tungusic, as well as, more specifically, between Para-Mongolic and Jurchenic. Although in popular historiography it is often claimed that the “Three Kingdoms” of the Korean Peninsula and southern Manchuria were all Korean-speaking, there is no linguistic evidence in support of this claim. On the contrary, the fact that there are no known traces of early contacts between Korean and either Tungusic or Mongolic suggests that Koreanic was originally spoken rather far away from continental Manchuria, that is, in the region which later (in the 4th century AZ) became the core territory of Silla (新羅 Xinluo) in southeastern Korea, from where Silla ultimately unified the southern part of the Korean Peninsula and initiated the expansion of the Koreanic lineage to the territories of the neighbouring polities (in the 7th century AZ). The first target of Koreanic expansion was Baekje (百濟 Baiji) in southwestern Korea, which for both historical and philological reasons may be assumed to have been dominated by Japonic or Para-Japonic speakers. The northward push of the Korean language then continued for several centuries, reaching the modern borders of Korea only during the Joseon (朝鮮 Chaoxian) dynasty (1392–1910). It is, consequently, obvious that each of the “Three Kingdoms” on the Korean Peninsula had a dominant language of its own: Tungusic in Goguryeo, Japonic in Baekje, and Koreanic in Silla. Another language family in the region must have been Amuric, whose linguistic interaction with Tungusic clearly dates back to times preceding the Proto-Tungusic stage. The modern position of Ghilyak on the Lower Amur and on Sakhalin, and the recent expansion history of the modern varieties (Nivkh and Nighvng), suggest that the Amuric language family was gradually pushed northwards along the Sungari-Amur basin from a more southerly location, apparently under the expansive pressure of Tungusic. Some local ethnicities, notably the Ulcha and Neghidal, incorporate descendants of former Ghilyak speakers, suggesting a recent language shift. Possible evidence for the geographical relocation of Amuric is also provided by the river names Sungari and Ussuri, which would seem to be of Amuric origin (containing the Amuric element -ri ‘river’, also attested in the modern territory of the Ghilyak speakers and probably connected with the Nivkh word eri ‘river’). All of this allows the origins of the Amuric language family to be connected with some likelihood with the protohistorical state of Buyeo (夫餘 Fuyu, BZ 2nd c. to 4th c. AZ), which was the first political formation in Manchuria and a regional predecessor of Goguryeo. The interaction between Buyeo and Goguryeo also explains the early lexical connections between Tungusic and Amuric. Later, the contact has continued up to premodern times.

TUNGUSIC IN TIME AND SPACE 529

Compared with the local actors formed by Buyeo and the “Three Kingdoms”, as well as their successor states, the historical role of China (proper) used to be more marginal in Manchuria until premodern times. Even so, Chinese was present in the region as a written language since the first millennium BZ, and it must have been used also for oral communication in political and economic contacts with the Chinese sphere. The state of Yan (燕) in the northeast of what is conventionally known as the “China” of the Warring States (戰國 Zhanguo) period (BZ 475–221) certainly also comprised Chinese speakers among its population, even though both the masses and the local elite were more probably composed of speakers of Pre-Proto-Mongolic, predecessors of the later rulers and dominant ethnic groups of the northern Wei and the Liao. Most importantly, however, Chinese must have been present in southern Manchuria and northern Korea in the context of the so-called “military commanderies” of the Han period of China, of which the best-documented is Lelang (樂浪 Nangnang, BZ 108–313 AD). It is so far difficult to say whether any linguistic traces remain from this time of early contacts between Tungusic and Chinese, but possible candidates are, at least, several items of the type Manchu fafun (fafun) ‘law’ < *pap.u/n ← Chinese *pap > Modern Mandarin fǎ (法) id., Manchu lamun (lamun) ‘blue’ < *lam.u/n ← Chinese *lam > Modern Mandarin lán (藍). These items are formally difficult to date and might also represent later borrowings from Middle Chinese to Jurchen, or also elements initially transmitted to Para-Mongolic and only secondarily to Jurchenic. STAGES OF EXPANSION There are, then, few alternatives to placing the homeland of the Tungusic language family in the region comprising northern Korea and southeastern Manchuria, a region where the early Tungusic speakers established a succession of protohistorical and historical political states. While still located in the homeland, Pre-Proto-Tungusic interacted with Pre-Proto-Mongolic to the west and Pre-Proto-Amuric to the north. Contacts with Pre-ProtoJaponic were more limited, while contacts with Pre-Proto-Koreanic were probably absent altogether, although the mutual adjacency of Koreanic, Japonic, Tungusic, and Mongolic must have contributed to the regional consolidation of the “Altaic typological sphere”, to which Turkic to the west of this region also adhered. At the same time, the Amuric family also received its share of Altaic features, though it always retained its basic “non-Altaic” typological orientation. There must have been other ethnicities, languages, and possibly language families in the region. Chinese and Korean historical sources provide the names of some of them, starting with the diffuse Yilou (挹婁) and Sushen (肅愼) of the first millennium BZ and ending with the somewhat later Ye (濊 穢 薉 Hui), Maek (貊 貉 Mo), Okjeo (沃沮 Woju), and Malgal (靺鞨 Mohe) in northeastern Korea and adjacent regions. Some of these ethnonymic entities may have comprised speakers of Tungusic and/or Amuric, or also Para-Tungusic and/or Para-Amuric, representing lineages different from the surviving ones. However, in the absence of factual evidence, any conclusions about their languages remain hypothetical. The political developments connected with the pulsation of China and the unification of Korea, as well as the rivalry between the local tribal and political entities, led to the pressure of northward expansion. This may have been triggered by the expansion of Koreanic on the Korean Peninsula, or the incorporation of Buyeo into the political sphere of Goguryeo, or simply by the growing prestige of both Buyeo and Goguryeo, which led to the gradual absorption of the local populations to the north, leading them to adopting

530 JUHA JANHUNEN

first an Amuric and later a Tungusic language. The subsequent diversification of Tungusic would seem to have taken place in several successive stages: •

•

•

•

•

At the first stage, the original community that spoke Proto-Tungusic was split into a southern and northern group. The northern group, which came to constitute the linguistic basis of Northern Tungusic, started moving northwards along the Sungari-Amur basin, not so much as a coherent population in migration, as in the form of a speech community which moved on the map by acquiring new speakers on its northern margins and perhaps losing old speakers on its southern margins. At the second stage, internal developments led to the linguistic differentiation of the southern and northern groups, with changes such as *ü > *i and *p > *x taking place in the language of the northern group against *ü > *u and *p ≡ *p in the language of the southern group. At this stage, the northern group must have occupied the lower course of the Sungari and, possibly the junction of the Sungari and the Amur, pushing ahead of themselves the linguistic ancestors of Amuric speakers. At the third stage, the southern group, which constitutes the linguistic basis of Southern Tungusic, initiated a new wave of northward expansion, leading to the separation of Nanaic from Jurchenic, and leaving only Jurchenic in close contact with Para-Mongolic. Moving along the Sungari-Amur basin, the Nanaic branch created a split between Ewenic (to the west) and Orochic (to the east). Later interaction between Nanaic and Orochic led to the formation of the areal union of Amur Tungusic, in the context of which Orochic acquired some secondary southern features, while Ewenic moved upwards along the Middle Amur and ended up in a secondary homeland in the Zeya basin. At the fourth stage, Ewenic started its expansion from the Middle Amur and Zeya to all over Siberia, using rivers as the main routes for moving in both the south-tonorth and the east-to-west direction, and reaching ultimately the Yenisei basin in the west, the Arctic Ocean in the north, and the Pacific coast in the east. All over this macroscopic territory, Ewenic absorbed the speakers of former languages, of which no information remains, unless it is assumed that Kolymic (Yukaghir) is one of the language families that were displaced to the northeast by the expansive push of Ewenic. It may be noted that there are very few traces of early contacts between Tungusic and Kolymic, the most convincing candidate being Tungusic *nime- ‘to visit’ : *nime(-) ‘neighbour’ (> Ewenki nime-, also in Orochic and Nanaic) = Kolymic *nime- ‘dwelling’ (> Tundra Yukaghir nime-ƞ, Forest Yukaghir numo, but note also Chukotic nem&nem [nəmnəm] ‘dwelling, settlement’). Another northern item, with a possible Chukotic connection, is Amur Tungusic *surka ‘(walrus) tusk’ ← Ewenki surka ~ xurka id. ← Ewen xurka ~ urka id. ← Chukchee rerka [rərka] ‘walrus’. Remaining basically in the original homeland, Jurchenic was the only branch of Tungusic that always retained a link with state building in the proper sense. The ultimate result of this was the consolidation of the Manchu as an ethnic and military entity and the conquest of the neighbouring regions, including China, Mongolia, Tibet, and East Turkestan. Although this development brought the Manchu language all over the Qing empire, the demographic odds were against an actual linguistic expansion with the exception of diaspora groups, of which the Sibe are the only surviving example.

TUNGUSIC IN TIME AND SPACE 531

As far as datings are concerned, there is no reason to assume that the breakup of the Tungusic protolanguage took place earlier than the Goguryeo period, which gives it a depth of approximately 2000 years. Soon after this, the successive waves of northward expansion led to the crystallization of the initial forms of the four branches of Tungusic. The Ewenic expansion from the Middle Amur region to Siberia started only a millennium later and was probably triggered by the rise of the historical Mongols in western Manchuria and the subsequent expansion of Mongolic to Mongolia, where it replaced the earlier Turkic idioms. Historical and toponymic evidence confirms that the Ewenic expansion has continued up to premodern times, with Ewenki reaching the Yenisei basin no earlier than the 15th–16th centuries, that is, shortly before the arrival of the first Russians in the region, while the presence of Ewenki on Sakhalin and Ewen on Kamchatka dates back to as late as the 19th century. Even so, the Tungusic expansion, as a whole, has been completed in a relatively short time, especially considering the distances that have been covered. In this connection, it is useful to recall once again that the linguistic expansion of Tungusic has had little to do with migrations or the replacement of former populations by Tungusic speakers. Rather, in each given locality, the former population has adopted a Tungusic language from a small number of Tungusic speakers who, for reasons of social, cultural and/or political prestige have been able to assimilate their neighbours. This is why the prospects of identifying traces of the Tungusic linguistic expansion in the genetic composition of the modern Tungusic speakers are limited. Studies made so far only confirm that the overall gene pool of Tungusic speakers has some common traits on the large Northeast Asian level, and more so on the regional Manchurian level, but this gene pool is also shared by the neighbouring non-Tungusic speakers, especially the Mongols. At the same time, genetic studies reveal a considerable local variation, which correlates with differences in the contact histories of the individual Tungusic languages. For instance, the Ulcha speakers, who are known to incorporate the former local population whose language was Ghilyak, show a genetic transition towards the modern Ghilyak speakers. Also, the Ulcha incorporate a genetic component traceable back to the Sakhalin Ainu, who are known to have contacted over several centuries with the ethnic groups of the Lower Amur. Altogether, the contribution of modern human genetics to our knowledge of the diachrony of the Tungusic languages has so far been minimal, and much of the information supplied by genetic studies agrees with what was known before by the methods of classical human anthropology. SIC TRANSIT GLORIA As a language family, Tungusic offers an interesting case of “rise and fall”. In the context of Northeast Asia, Tungusic is one of the three expansive “Altaic” families, the others being Mongolic and Turkic. While the expansion of the latter two was directed mainly towards the Eurasian steppe belt, Tungusic expanded primarily towards the north and occupied ultimately the entire eastern half of the Eurasian taiga zone, making it the most widespread language family in Northeast Asia. All of this happened in the course of the last couple of millennia, within about the same time frame as the Turkic and Mongolic expansions. As a language family of the boreal taiga zone, Tungusic is parallelled by Uralic, with which it has been in marginal contact across the Yenisei basin. Compared with Tungusic, however, the history of the expansion and internal diversification of Uralic is much older.

532 JUHA JANHUNEN

As if its wide geographical distribution was not enough, Tungusic has also been involved in state building in its Manchurian home region, becoming, as it seems, the dominant language of the kingdoms of Goguryeo and Barhae, as well as of the Jurchen and Manchu empires. The Manchu, in particular, were second only to the historical Mongols to create a political unity comprising the entire “Greater China”, though it should be emphasised that the base of their empire was not “China”, but Manchuria. During the Qing dynasty (1616–1911), Manchu was the official language of the entire empire, which made it a tool of administration and culture over a population of a few hundred million people, even though the principal means of communication in most parts of the empire was Chinese. Against this background it is surprising that the Tungusic language family today is on the verge of extinction. Not only has Manchu lost its official status, but it has also disappeared as a spoken language in Manchuria, being remembered by only a handful of elderly semi-speakers in a couple of rural locations, though, importantly, it is still being spoken by part of the ethnic Sibe in East Turkestan. Of the other Tungusic idioms spoken within the territory of P.R. China, only Solon, as well as, possibly, Khamnigan Ewenki, survive as spoken languages still transmitted to children, while Orochen has lost generational transmission and is rapidly disappearing. The varieties of Nanai until recently spoken on the Chinese side are extinct, leaving only Kilen with a small number of residual speakers. On the Russian side, apart from Arman, which became extinct some decades ago, Neghidal, Oroch, and Kili have lost their last native speakers in the years after 2000, with Udihe also being very close to extinction. Of the Nanaic languages, only Nanai (proper) is possibly still being transmitted to children in a few families, while Ulcha has apparently lost transmission, and Uilta has only a couple of fluent speakers. Ewenki and Ewen, the two Ewenic languages with a relatively large ethnic background community, still count some thousands of speakers, but their transmission to children has virtually ceased with the exception of a few locations, especially in Yakutia. However this may be, the conclusion is that all Tungusic languages in Russia are either extinct, moribund, or critically endangered. In summary, the future of the Tungusic language family looks bleak. The total number of Tungusic speakers in the world today is hardly more than 30,000 people, while the number of people transmitting the language to the next generation is probably in the hundreds, rather than thousands. The single Tungusic language which for the time being would seem to have the best chances of survival is, somewhat surprisingly, Solon, whose main asset seems to have been the traditional multilingualism of its speakers. Even so, monocultural and monolingual solutions dictated by the dominant society and the globalized world based on nation states and the “international order” are a serious threat that can probably not be stopped before it is too late. In this situation, it is potentially important that the Tungusic languages are relatively well documented, which would allow the data available on them to be used for purposes of language revitalization should such enterprises be initiated on a serious scale. REFERENCES AND FURTHER READING Alonso de la Fuente, José Andrés (2015) ‘Tungusic historical linguistics and the Buyla (a.k.a. Nagyszentmiklós) inscription’, Studia Etymologica Cracoviensia 20: 17–46, Kraków.

TUNGUSIC IN TIME AND SPACE 533

Anikin, A. E. [А. Е. Аникин] (1990) Тунгусо-маньчжурские заимствования в русских говорах Сибири [Tungusic loanwords in the Russian dialects of Siberia], Новосибирск [Novosibirsk]: “Наука”. Anikin, A. E. [A. E. Аникин] & E. A. Helimski [Е. А. Хелимский] (2007) Самодийскотунгусо-маньчжурские лексические связи [Lexical connections between Samoyedic and Tungusic], Москва [Moscow]: Языки славянской культуры. Austerlitz, Robert (1970) ‘Agglutination in Northern Eurasia in perspective’, in: Roman Jakobson & Shigeo Kawamoto (eds), Studies in General and Oriental Linguistics Presented to Shirô Hattori on the Occasion of his Sixtieth Birthday, 1–5. Tokyo: TEC Corporation for Language and Educational Research. Austerlitz, Robert (1980) ‘Typology and universals on a Eurasian East-West continuum’, in: Gunter Brettschneider & Christian Lehmann (eds), Wege zur Universalienforschung: Beiträge zum 60. Geburtstag von Hansjakob Seiler, 235–244. Tübingen. Beckwith, Christopher I. (2004) Koguryo, the Language of Japan’s Continental Relatives, Brill’s Japanese Studes Library 21, Leiden: Brill. Cincius, V. I. [В. И. Цинциус] (1977) ‘Монгольские аффиксы в маньчжурском языке’ [Mongolian affixes in the Manchu language], in: Gyula Décsy & Christo D. Dimov-Bogoev (eds), Eurasia Nostratica: Festschrift für Karl Heinrich Menges, Ural-Altaische Jahrbücher 48: 42–48. Wiesbaden: Otto Harrassowitz. Coati, Gianni (1992) ‘I prestiti cinesi nelle lingue tunguse dell’ Amur’ [The Chinese loanwords in the Amur Tungusic languages], Journal de la Société Finno-Ougrienne 84: 65–87, Helsinki. Doerfer, Gerhard (1984) ‘Prolegomena zu einer Untersuchung der dem Tungusischen und Mongolischen gemeinsamen Wörter’, Journal de la Société Finno-Ougrienne 79: 65–85, Helsinki. Doerfer, Gerhard (1985) Mongolo-Tungusica, Tungusica, Band 3, Wiesbaden: Otto Harrassowitz. Dolgopolsky, Aharon (1998) with an introduction by Colin Renfrew, The Nostratic Macrofamily and Linguistic Palaeontology, Papers in the Prehistory of Languages [1], Cambridge: The McDonald Institute for Archaeological Research. Futaky, István (1975) Tungusische Lehnwörter des Ostjakischen, Veröffentlichungen der Societas Uralo-Altaica 10, Wiesbaden: Otto Harrassowitz. Futaky, István (1983) ‘Zur Frage der nganasanisch-tungusischen Sprachkontakte’, Urálisztikai Tanulmányok 1: 155–162, Budapest. Futaky, István (1988) ‘Uralisch und Tungusisch’, in: Denis Sinor (ed.), The Uralic Languages: Description, History and Foreign Influences, Handbuch der Orientalistik VIII.1: 781–791, Leiden: E. J. Brill. Georg, Stefan (2008) ‘Yeniseic languages and the Siberian linguistic area’, in: Alexander Lubotsky & Jos Schaeken & Jeroen Wiedenhof (eds.), Evidence and Counter-Evidence: Essays in Honour of Frederik Kortlandt, Studies in Slavic and General Linguistics 33, vol. 2: 151–168. Amsterdam: Rodopi. Greenberg, Joseph F. (2000–2002) Indo-European and Its Closest Relatives: The Eurasiatic Language Family, vols. 1–2, Stanford, California: Stanford University Press. Gruzdeva, Ekaterina & Juha A. Janhunen (2020) ‘Notes on the typological prehistory of Ghilyak’, International Journal of Eurasian Linguistics 2 (1): 1–28, Leiden. Gusev, V. Yu. (2020) ‘Новые эвенкийско-самодийские этимологии’ [New Ewenki-Samoyedic etymologies], Томский журнал лингвистики и антропологии [Tomsk Journal of Linguistics and Anthropology] 2020 (2): 9–17.

534 JUHA JANHUNEN

Helimski, Eugene [E. A. Хелимский] (1985) ‘Самодийско-тунгусские лексические связи и их этноисторические импликации’ [The lexical relations between Samoyedic and Tungusic and their ethnohistorical implications], in: Урало-алтаистика: Археология, этнография, язык [Ural-Altaic studies: Archaeology, ethnography, language], 206–213, Новосибирск [Novosibirsk]: “Наука”. Helimski, Eugene [Е. А. Хелимский] (2000) ‘Языки аваров: тунгусо-маньчжурский аспект’ [The language of the Avars: the Tungusic aspect], in: Studia in Honorem Stanislai Stachowski Dicata, Folia Orientalia 36: 135–148, Kraków. Helimski, Eugene [Е. А. Хелимский] (2003) Тунгусо-маньчжурский компо[н]ент в аварском каганате и славянская этимология [The Tungusic component in the Avar khaganate and Slavic etymology], Hamburg: Institut für Finnougristik/Uralistik. Ikegami Jirō 池上二良 (1980)「アイヌ語のイナウの語の由来に関する小考 : ウイ ルタ語のillauの語原にふれて」[On the origin of the Ainu word inau: in reference to the Uilta word illau],『民族学研究』44 (4): 392–40, 東京 [Tokyo]. Ikegami Jirō 池上二良 (1994)「アイヌ語の大陸的要素」[Continental elements in the Ainu language],『アイヌの集い : 知里真志保を継ぐ』, 159–169, 札幌 [Sapporo]: 北海道出版企画センター. Janhunen, Juha (1977) ‘Samoyed-Altaic contacts: present state of research’, Mémoires de la Société Finno-Ougrienne 158: 123–129, Altaica: Proceedings of the 19th Annual Meeting of the Permanent International Altaistic Conference, Helsinki: Suomalais-Ugrilainen Seura. Janhunen, Juha (1996a) Manchuria: An Ethnic History; Mémoires de la Société Finno-Ougrienne 222, Helsinki: The Finno-Ugrian Society. Janhunen, Juha (1996b) ‘Prolegomena to a comparative analysis of Mongolic and Tungusic’, in: Giovanni Stary (ed.), Proceedings of the 38th Permanent International Altaistic Conference (Kawasaki 1995), 209–218, Wiesbaden: Harrassowitz Verlag. Janhunen, Juha (2005) ‘The lost languages of Koguryo’, Journal of Inner and East Asian Studies 2 (2): 65–86, Seoul. Janhunen, Juha (2008) ‘Mongolic as an expansive language family’, in: Kurebito Tokusu (ed.), Past and Present Dynamics: The Great Mongolian State, 127–137, Tokyo: Tokyo University of Foreign Studies. Janhunen, Juha (2012a) ‘The expansion of Tungusic as an ethnic and linguistic process’, in: Andrej L. Malchukov & Lindsay J. Whaley (eds.), Recent Advances in Tungusic Linguistics, Turcologica 89: 5–16, Wiesbaden: Harrassowitz Verlag. Janhunen Juha (2012b) ‘Etymological and ethnohistorical aspects of the Yenisei’, Studia Etymologica Cracoviensia 17: 67–87, Kraków. Janhunen, Juha (2013) ‘The Tungusic languages: A history of contacts’, in: Kim Juwon & Ko Dongho (eds.), Current Trends in Altaic Linguistics: A Festschrift for Professor Emeritus Seong Baeg-in on His 80th Birthday, 16–60, Seoul: Altaic Society of Korea. Janhunen, Juha (2015) ‘Observations on the Para-Mongolic elements in Jurchenic’, Acta Linguistica Petropolitana 11: 575–592, St. Petersburg. Janhunen, Juha (2016) ‘Reconstructio interna linguae ghiliacorum’, in: Ekaterina Gruzdeva & Juha Janhunen (eds.), Crosslinguistics and Linguistic Crossings in Northeast Asia, Studia Orientalia 117: 3–27, Helsinki. Janhunen, Juha (2022) ‘Amuric hydronyms in Manchuria and the Puyŏ connection of Ghilyak’, International Journal of Eurasian Linguistics 4 (1): 23–40, Leiden. Janhunen, Juha (2023) ‘The unity and diversity of Altaic’, Annual Review of Linguistics 9: 135–154. Online at linguistics.annualreviews.org.

TUNGUSIC IN TIME AND SPACE 535

Joki, Aulis J. (1976) ‘Some Samoyed-Tungusic word comparisons’, in: Walther Heissig & John R. Krueger & Felix Oinas & Edmond Schütz (eds.), Tractata Altaica: Denis Sinor sexagenario optime de rebus altaicis merito dedicata, 321–323, Wiesbaden: Otto Harrassowitz. Joki, Aulis J. (1977) ‘Die Tungusen und ihre Kontakte mit anderen Völkern’, Studia Orientalia 47, Commentationes in honorem Pentti Aalto, 109–118, Helsinki: Finnish Oriental Society. Khabtagaeva, Bayarma (2019) Language Contact in Siberia: Turkic, Mongolic, and, Tungusic Loanwords in Yeniseian, The Languages of Asia Series 19, Leiden: Brill. Knapen, Martijn G. T. M. (2021) ‘The oldest layer of Amuric-Tungusic lexical contacts’, MA Thesis in Linguistic Diversity and Digital Humanities, University of Helsinki, https://helda.helsinki.fi/handle/10138/337472. Kreinovich E. A. [Е. А. Крейнович] (1955) ‘Гиляцко-тунгусо-маньчжурские языковые параллели’ [Linguistic parallels between Ghilyak and Tungusic], Доклады и сообщения Института языкознания АН СССР 8: 133–167, Москва [Moscow]. Kuz’menkov, E. A. [Е. А. Кузьменков] (1988) ‘Монгольские элементы в маньчжурском и диалектная база старомонгольской письменности’ [Mongolic elements in Manchu and the dialectal base of Written Mongol], Вопросы языкознания (4): 117–122, Москва [Moscow]. Lee, Ki-moon (1958) ‘A comparative study of Manchu and Korean’, Ural-Altaische Jahrbücher 30: 104–120, Göttingen. Ligeti, Louis (1960) ‘Les anciens élements mongols dans le mandchou’. Acta Orientalia Academiae Scientiarum Hungaricae 10: 231–248, Budapest. Maezono, Kyōko (1992) Kasus-Entsprechungen des Mongolischen und Mandschu: Anhand von Sagang Secen’s Erdeni-yin Tobci und seiner mandschurischen Übersetzung, Tungusica, Band 4, Wiesbaden: Otto Harrassowitz. Maezono, Kyōko (2007a), Intransitiv-, Transitiv-, Kausativ- und Passivverben im Mandschu und Mongolischen, Tunguso-Sibirica 18, Wiesbaden: Harrassowitz Verlag. Maezono, Kyōko (2007b) Verbbildungs-Suffixe im Mandschu und Mongolischen, Tunguso-Sibirica 19, Wiesbaden: Harrassowitz Verlag. Miller, Roy Andrew (1971) Japanese and the Other Altaic Languages, Chicago: University of Chicago Press. Okada, Hidehiro (1962) ‘Color-names in Manchu’, in: Nicholas Poppe (ed.), American Studies in Altaic Linguistics, Indiana University Uralic and Altaic Series 8: 225–227, Bloomington & The Hague. Panfilov, V. Z. [В. З. Панфилов] (1973) ‘Нивхско-алтайские языковые связи’ [Linguistic connections between Nivkh and Altaic], Вопросы языкознания (5): 3–12, Москва [Moscow]. Pevnov, A. M. [А. М. Певнов] (1992) ‘Нивхский и тунгусо-маньчжурские языки: проблемы контактов’ [Nivkh and the Tungusic languages: problems of contacts], in: Б. О. Пилсудский—исследователь народов Сахалина [B. O. Piłsudski—a researcher of the peoples of Sakhalin], Материалы международной конференции 2: 25–29, ЮжноСахалинск [Yuzhno-Sakhalinsk]: Сахалинский Областной краеведческий музей. Poppe, Nicholas (1966) ‘On some ancient Mongolian loanwords in Tungus’, Central Asiatic Journal 11: 187–198, Wiesbaden. Poppe, Nicholas (1960) Vergleichende Grammatik der altaischen Sprachen, Teil I: Vergleichende Lautlehre, Porta Linguarum Orientalium, Neue Serie 4, Wiesbaden: Otto Harrassowitz.

536 JUHA JANHUNEN

Poppe, Nicholas (1965) Introduction to Altaic Linguistics, Ural-Altaische Bibliothek XIV, Wiesbaden: Otto Harrassowitz. Poppe, Nicholas (1975) ‘Altaic linguistics: An overview’, Sciences of Language 6: 130– 186, Tokyo Institute for Advanced Studies of Language. Ramstedt, G. J. (1952–1966) Einführung in die altaische Sprachwissenschaft, vols. 1–3, Pentti Aalto (ed.), Mémoires de la Société Finno-Ougrienne 104: 1–3, Helsinki: Suomalais-Ugrilainen Seura. Robbeets, Martine (2005) Is Japanese Related to Korean, Tungusic, Mongolic and Turkic? Turcologica 64, Wiesbaden: Harrassowitz Verlag. Robbeets, Martine & Alexander Savelyev (eds.) (2020) The Oxford Guide to the Transeurasian Languages. Oxford Guides to the World’s Languages. Oxford: Oxford University Press. Rozycki, William (1994) Mongol Elements in Manchu, Indiana University Uralic and Altaic Series 157, Bloomington. Sanzheev, G. D. [Г. Д. Санжеев] (1930) ‘Маньчжуро-монгольские языковые параллели’ [Linguistic parallels between Manchu and Mongolian], Изветия АН СССР: Отделение гуманитарных наук, 8–9: 601–627, 673–70, Москва [Moscow]. Schmidt, Peter (1932–1933) ‘Chinesische Elemente in Mandschu’, Asia Major 7: 573– 628, 8: 233–276, Leipzig. Shherbak, A. M. [A. M. Щербак] (2020) Очерки для этимологического словаря тунгусо-маньчжурских языков: Заимствованный фонд: тюркизмы, монголизмы [Essays for an etymological dictionary of Tungusic: Turkic and Mongolic elements], Санкт-Петербург [St. Petersburg]: Институт лингвистических исследований РАН. Shimunek, Andrew (2017) Languages of Ancient Southern Mongolia and North China: A Historical-Comparative Study of the Serbi or Xianbei Branch of the Serbi-Mongolic Language Family, with an Analysis of Northeastern Frontier Chinese and Old Tibetan Phonology, Tunguso-Sibirica 40, Wiesbaden: Harrassowitz Verlag. Shimunek, Andrew (2021) ‘Loanwords from the Puyo-Koguryoic language of early Korea and Manchuria in Jurchen-Manchu’, Altai Hakpo『알타이학보』31: 65–84, Seoul. Sinor, Denis (1975) ‘Uralo-Tungus lexical correspondences’, in: Louis Ligeti (ed.), Researches in Altaic Linguistics, Budapest: Akadémiai Kiadó, 245–265. Starostin, S. A. [С. А. Старостин] (1991) Алтайская проблема и происхождение японского языка [The Altaic problem and the origin of the Japanese language]. Москва [Moscow]: “Наука”. Starostin, Sergei & Anna Dybo & Oleg Mudrak (2003), with assistance of Ilya Gruntov and Vladimir Glumov, Etymological Dictionary of the Altaic Languages, vols. 1–3, Handbuch der Orientalistik VIII.8, Leiden: Brill. Toh, Soo Hee (2005) ‘About early Paekche language mistaken as being Koguryŏ language’, Journal of Inner and East Asian Studies 2 (2): 13–31, Seoul. Tsumagari Toshiro 津曲 敏郎 (2003)「ツングース語と上代日本語の文法上の類 似」[Some grammatical similarities between Tungusic and Old Japanese], in: Alexander Vovin & Osada Toshiki (eds.),『日本語系統論の現在』[Perspectives on the origins of the Japanese language] =『国際日本文化研究センター共同研究報 告』[Nichibunken Japanese Studies Series] 31: 237–247, 京都 [Kyoto]. Tsumagari Toshiro (2010) ‘Long journey of walrus: A linguistic survey’,『北方人文研 究』Journal of the Center for Northern Humanities 3: 45–57, 札幌 [Sapporo].

TUNGUSIC IN TIME AND SPACE 537

Vovin, Alexander (2006a) ‘Why Manchu and Jurchen look so non-Tungusic?’, in: Alessandra Pozzi & Juha Antero Janhunen & Michael Weiers (eds.), Tumen jalafun jecen akû: Manchu Studies in Honour of Giovanni Stary, Tunguso-Sibirica 20: 255–266, Wiesbaden: Harrassowitz. Vovin, Alexander (2006b) ‘Are Koguryŏ and Paekche different languages or dialects of Old Korean?’ Journal of Inner and East Asian Studies 2 (2): 107–142. Vovin, Alexander (2007) ‘Korean Loanwords in Jurchen and Manchu’, Altai Hakpo『알 타이학보』17: 73–84, Seoul. Vovin, Alexander (2013) ‘From Koguryŏ to T’amna: Slowly riding to the south with speakers of Proto-Korean’, Korean Linguistics 15 (2): 222–240, Amsterdam.

INDEX accent see prosodic features action nouns 14, 351, 371, 413, 475 actor nouns 14, 49, 113, 301, 335, 351, 371, 412–413, 442 adjectives 12, 111, 151, 175, 188–189, 192–193, 196, 203, 211–212, 240, 266–267, 272–273, 282, 288–289, 303, 306–307, 340–341, 370–372, 374, 377, 388–389, 392, 408, 412, 414, 416–417, 419, 441, 488 adstratal influences 526 agent 60, 114, 157, 191, 201, 228, 243, 270, 285, 338, 349, 352, 454, 496 agglutination 111, 266, 333, 533 agreement 70, 131, 174, 192, 195, 199, 201, 226, 261, 266, 272, 274–276, 278, 280, 282–284, 286, 321, 337, 357, 366, 395, 429–430, 451, 496 Aldan Ewenki 33, 143, 505 allomorphy 52, 196, 197, 246, 248, 316, 373, 385, 454, 489 allophony 43, 108, 110, 143, 148, 150, 209, 238, 264, 411, 466–468, 485, 486 Altaic 5, 6, 7, 10, 14, 36, 518–519, 520–531; see also Ural–Altaic Altaic Hypothesis 518–521 Amur xxviii, 1–2, 8–9, 11–12, 19, 23, 44, 63, 76–77, 139–143, 145, 176–177, 179, 184, 229, 234–236, 241, 257–258, 294–295, 306, 311, 323–324, 326–327, 359, 364–365, 368–369, 372–373, 381, 391, 393–394, 398–400, 407, 431, 433, 436, 456, 463, 501, 504–506, 508–510, 518, 520, 527–528, 530–531 analytic constructions 172, 213, 254, 377 animacy 213 anthroponymy 366 archaisms 141, 468 areal features 11, 525–528 Arman 4–5, 8–9, 29, 143, 261, 264, 273, 287–288, 532 articles 195, 245, 417 aspect see temporal–aspectual systems aspiration 11, 43, 108, 147, 186, 209, 468 auxiliary verbs 230, 393, 428, 475–476, 495 Avar khaganate 520 Baikal region 139, 142–144, 177 banner system 465, 483 Barguzin 143–145 basic vocabulary 4, 10, 132, 177, 207, 257, 288, 297, 359, 398, 455, 521–522, 524

bilingualism 10, 177, 204, 366, 403, 437, 455, 479, 484, 519, 525 Borzya Khamnigan Ewenki 9–10, 64, 141–144, 148–149, 159, 161, 193, 204, 209 breaking see vowel breaking Buddhism 2, 229, 498 Bulgharic see Bulghar Turkic Bulghar Turkic 133, 521, 527–528; see also Proto–Bulgharic Buryat 139, 143, 145, 177, 204, 502, 505, 526 causative constructions 114, 198, 228, 303, 338, 490 Central Eurasia 2 Central Tungusic 8 chaining see clause chaining Chapogir 31, 144 China xviii, 1, 6, 19, 23, 76, 106, 140–141, 178, 185, 206–207, 364, 366, 407, 463, 465, 483–485, 501, 504, 518, 527, 529–530, 532 China, People’s Republic of 1, 463, 501 Chinese xx–xxi, xxiv, 3–4, 6, 19–22, 26, 29, 31–32, 60, 76–81, 83–88, 90–92, 95–96, 98–99, 104–106, 108–110, 112, 120, 132–133, 140–141, 178, 184–185, 187, 204, 206–207, 209–211, 229–230, 237, 299, 312, 323, 326–328, 331–333, 336, 340, 359, 364–365, 382, 399, 431, 433, 456, 464–466, 470, 472, 476–477, 479, 483–486, 488–489, 491, 497–498, 501, 507, 520, 527, 529, 532 Chinese script xxi, 3, 76, 79–80, 83–84, 86 Chinese sources 19–21, 76, 224, 326, 365; see also dynastic histories Christianity 2 Chukchee 31 Chulym Turkic 503 Chuvash 521 Classical Manchu 103, 485 classifiers see numeral classifiers class suffixes 522 clause chaining 71, 287 cliticization see enclitic particles Common Tungusic 141–142, 198, 298, 301, 305, 308, 310, 313, 378, 381, 386, 389–391, 393, 398, 413, 417–418, 426, 432, 444–445, 449, 480 Common Turkic 521, 528 comparison 55, 154, 215, 227, 271–272, 304, 339–340, 376, 382, 415, 420, 453, 471, 490 complementizer 203

INDEX 539 compounding 115, 241, 266, 273, 498 conjunctional particles 131, 286, 380, 394 conjunctions 13, 176, 203, 257 connective converb 68–69, 97, 111, 125–126, 128, 169–170, 201–202, 213, 223–225, 227–228, 239, 255–256, 276, 281–282, 287, 319, 322, 354, 356, 393–394, 396–397, 428–431, 450–451, 474–476, 489, 493–494 connective vowels xxi, 52, 149, 154, 156–157, 164, 167, 195–196, 200, 239, 248 connectors 176, 224, 286 connegative forms 13–15, 66, 71, 98, 130, 132, 163, 168, 175, 197, 220, 228, 243, 247, 249, 251, 267, 274, 277–278, 312–313, 315, 318, 322, 344, 347–348, 356, 384, 387–389, 397, 424–426, 429, 431, 448, 453, 476, 496, 523–524 conservativeness 9–10, 141, 204, 264, 484; see also archaisms contraction see vowel contraction copulas 199, 396, 430 corrogative forms 14, 71, 322, 431, 453 counters see numeral classifiers Cultural Revolution 185, 465, 483 cultural vocabulary 72, 230, 288, 359, 398–399, 521, 527 Cyrillic script xxi, xxiv, 4–5, 23, 105, 145–146, 148, 178, 260, 264, 328, 365, 407, 512

European languages 170 evidentiality 67, 352, 449 Ewen xvii–xviii, 2–5, 7–12, 26, 28–29, 31–32, 37–40, 42, 47, 49–53, 56–58, 64, 69–72, 140–141, 143–144, 146, 149, 175, 177, 184–185, 190, 208, 213, 223, 236, 238, 242, 260–293, 297, 299, 307, 342, 369, 410, 436, 488, 501–504, 506, 508–514, 519–520, 526, 530–532 Ewenic 8–9, 10, 12, 35, 38, 40–44, 46, 48–49, 51, 53–54, 57–64, 66–71, 132, 139–143, 146 148, 155–156, 167, 173, 176–178, 184, 187, 190, 193–200, 202, 206, 209, 211, 213–215, 217–220, 222–223, 234, 237–240, 242, 244–246, 249, 255–257, 260, 265, 274, 277, 281, 288, 295–296, 300, 307–308, 342, 345, 351, 355–356, 369–370, 373–374, 376, 378, 387, 389, 398, 411–412, 444, 507, 521, 526, 530–532 Ewenki xxvii–xxviii, 2–5, 7–12, 26, 28–31, 33, 38–40, 42, 44–72, 184, 187, 190, 206, 234, 238, 242–243, 261–265, 270–271, 274, 277, 280, 283–285, 288, 294–297, 307, 323, 327, 365, 367–369, 372, 376, 436–438, 456, 501–514, 519–521, 523, 525–527, 530–532; see also Khamnigan Ewenki; Siberian Ewenki Ewenki dialectology 142–145, 150, 157, 165, 177, 185, 187, 261, 288 exclusive forms see inclusive/exclusive forms

Daghur 10, 31, 76, 103, 140, 177–178, 204, 206–209, 211, 217–218, 229–230, 398, 479, 498, 519 debuccalization see desibilization definiteness 53, 195, 346, 359 descriptive words 399, 433 desibilization 143, 526 diacritics 103 dialectology 145, 208, 438, 480 dictionary form 68, 170, 388, 424 digraphs 106, 146 diphthongization 261 discourse structure 14, 283, 288, 358–359 dissimilation 72 Dolgan 143, 178, 503 Dutch 26 dynastic histories 19

feminine forms see gender marking French 2, 26, 31, 33, 104–105, 510 fricativization see spirantization fusion 333

Eastern Turkestan 1, 19, 103, 140, 530, 532 elision see vowel elision embedded constructions 125, 357, 477–479, 497 emphatic particles 131, 322, 380, 388, 419, 478, 496 enclitic particles 197, 257, 266, 274, 323, 344, 395, 397–398, 432 endangerment see linguistic assimilation English 2, 6, 105–106, 235–236, 466, 485 Equestrian Tungus 139 ethnic identity 366, 507 ethnonymy 3, 76, 78, 93, 103, 139, 140, 184, 206, 260, 294, 326, 364, 407, 436, 455 etymology 36, 257, 359, 364, 436, 527 Europe 26, 106

gemination 211, 299, 441, 444, 447 gender marking 115, 341, 346 generic plural 3, 52, 190 German 2, 4–6, 26, 29, 105, 235, 262, 437 glottalization 329, 331–332, 369 government 282 grammatical cases 53, 155, 191, 374 grammaticalization 53, 61, 69, 336 graphemes 106 Greater China 532 Hailar basin xxviii, 140 Hailar Dagur 217 Hailar District 206, 229 Hailar Solon 206–208, 217–218, 221, 223, 229 Heihe region 463 Heilongjiang 103, 184–185, 206, 364, 463, 483 hiatus 88–89, 92–93 Horse Tungus 139–140 Hua–Yi Yiyu 78 Hulun Buir 140, 184, 206, 229 Ili region 140 inclusive/exclusive forms 12, 58, 60, 62, 121, 161, 164–167, 172, 176, 194, 196, 200, 217, 222–223, 246, 248, 253, 273–274, 279, 308, 342, 378, 380, 417, 445, 473, 491, 522 Indo–European 523–524

540 INDEX initial clusters 110, 239, 488 Inner Mongolia 140, 184–185, 206, 463 inscriptions 21, 77–78, 86, 92–98 interjections 226, 276, 311–312, 346–347, 383, 420 internal reconstruction 36, 45, 47, 72, 518, 522 International Phonetic Alphabet xx interrogation 14, 71, 176, 322, 397, 431, 453, 477, 497 intonation see prosodic features Irkutsk Province 139, 504–505, 511–512 Islam 2 isoglosses 72, 184 Itelmen 31 Japan 6, 19, 23, 104, 106, 326, 328, 366, 409, 438, 455, 485, 507, 518 Japanese xx–xxi, 6, 8, 19, 22–24, 32, 85–86, 105, 132, 204, 207, 230, 236, 262, 285, 366, 407, 409, 436–437, 455–456, 465, 510, 519, 522, 526 Japonic 10, 85, 520–522, 525–526, 528–529 Jilin 86, 483 Jin dynasty 19, 76–77, 86, 464 Jungaria xxviii, 2, 8, 206, 229, 463, 483–484, 498 Jurchen xviii, xxi, 2, 8–9, 19–23, 36, 54, 56, 76–103, 107, 120, 130, 133, 140, 216, 229, 256, 359, 378, 398, 463–465, 472, 480, 483, 519–520, 528–529, 532 Jurchenic 8–10, 12–13, 35, 38, 40, 42–44, 46–52, 54–60, 62, 64, 67–68, 70, 87, 99, 103, 132, 163, 177, 319, 323, 463–464, 466, 472, 480, 483, 527–528, 530 Jurchen script 20, 77–80, 82–87, 96 Kangxi 20, 483 Kazakh 230, 484, 487, 498 Khamnigan see Khamnigan Ewenki; Khamnigan Mongol Khamnigan Ewenki 4, 9–10, 48, 56, 64, 140–146, 148–150, 158–161, 165–169, 177–178, 184, 193, 204, 206, 209–210, 501, 519, 525, 532 Khamnigan Mongol 10, 56, 140, 143, 145, 147, 177–178, 180, 206, 519, 525 Khitan see Khitan scripts Khitan scripts 76–77, 80–81, 84, 86–87, 92, 95, 98–99, 132–133, 528 Khorchin 229, 465, 484, 492, 498 Kilen 4, 6, 8–10, 12, 35, 40, 296–297, 308, 313, 315, 327, 365–366, 369, 501, 507, 532 Kili 4–6, 8–10, 12, 16, 35, 140, 365–366, 400, 501, 507, 532 Korea 7, 19, 21–23, 76, 80, 89, 104, 106, 326, 518, 520, 528–529 Korean xx–xxi, 21–23, 32, 76, 116, 118, 132, 207, 326, 518–519, 522, 528–529; see also Middle Korean; Old Korean Koryak 31 labial breaking 487 labial harmony 11, 43, 94, 111, 149–150, 187, 210–211, 239, 299, 369, 411, 441

labialization 469, 487 laryngeals xx, 11, 43–44, 142–143, 148, 238, 263, 368, 405, 526 Late Han Chinese 19–20 Latin 3, 23, 260, 263, 328, 484, 512 Lena xxviii, 139, 142, 177 letters see orthography lexical borrowing see loanwords lexicalization 49, 111–112, 114, 154 lexical taboo 76, 399 Liao dynasty 20, 76, 85–86, 99, 518, 520, 525, 528–529 Liaoning 483 ligatures 106 lingua francas 140, 506 linguistic assimilation 178, 501–504 linguistic interaction see areal features loanwords xxiv, 44, 90, 99, 109–110, 133, 148, 159, 176–178, 187, 193, 204, 209–211, 216, 229–230, 238, 257, 289, 299, 323, 331–333, 359, 398–399, 440, 455–456, 468, 470, 479, 484, 486, 489, 491, 498, 519, 521–522, 527 local cases 53–55, 156, 274, 276, 374 Macro-Altaic 520–522 Manchu xviii, 2–12, 14, 19–24, 26, 28–33, 36–44, 46–60, 62–72, 76, 78–80, 84, 87–99, 129–131, 140–141, 143, 151–152, 173, 178, 184, 189, 204, 206–210, 216–217, 229–230, 245, 257–258, 307, 323, 326–328, 332, 359, 364, 368, 378, 391, 396–401, 456, 501, 519–520, 522–523, 526–527, 529–530, 532; see also Spoken Manchu; Written Manchu Manchu alphabet see Manchu script Manchu empire 1–2, 103, 326 Manchuria 1, 8, 19, 76, 85, 133, 140, 142, 144, 178, 184, 229, 326, 407, 455, 463–465, 483–484, 518, 520, 527–529, 531–532 Manchurian Reindeer Tungus 141, 184, 501 Manchuric 8, 26; see also Jurchenic Manchu script xx, xxiv–xxvi, 84, 107, 117, 469 Mandarin see Chinese manuscripts xix marginal phonemes 264 Mergel 140 Mergen 463 metathesis 40–41, 46, 229, 296, 312, 369, 384, 488 Middle Korean 20, 85, 132 Middle Mongol 99, 133 Ming dynasty 76 mirativity 66, 390, 425 missionaries 32 “mixed” languages 4, 8, 10, 35, 501 modal cases 53–54, 155, 157, 374 modal particles 321–322 Modern Mongolian 133, 521, 523–524 Mongol see Mongolian Mongolia 1, 8, 140, 463, 527–528, 530–531 Mongolian 1, 10, 21–22, 56, 84, 88, 104, 159, 204, 206–208, 217, 220, 227, 229–230, 279,

INDEX 541 433, 456, 468–469, 479, 492, 498, 521, 523–524 Mongol script 77, 87, 103, 106, 108, 117 Mongols, historical 95, 483, 524, 531–532 monophthongization 411 morphophonology xviii, 109, 117, 148, 166, 187, 210, 238, 264, 299, 332, 369, 411, 438, 440, 451, 468, 487, 525 Nanai xxvii, 3–10, 13, 37–40, 42–46, 48–60, 62–63, 65, 67–71, 92–93, 97, 235–237, 256–258, 269, 294–297, 299, 301, 304–307, 314, 319–320, 323, 327–328, 331–332, 336, 338, 359, 364–419, 421–422, 425–433, 441, 444–446, 448–450, 479, 498, 501–504, 507–509, 511–514, 525–526, 532 Nanaic 8–10, 12, 35–36, 38, 40–46, 48–51, 53–54, 56–60, 62, 64–70, 87, 130, 132, 177, 295–296, 300, 307, 312–313, 315, 319, 323, 342, 348, 351, 373, 378–379, 398, 407–408, 432–433, 436, 441, 444–446, 448, 450, 456, 507, 526, 530, 532 nasalization 46, 150, 265, 295, 411, 468 nativization 178 negation 13–15, 45, 53, 64, 66, 71, 97–98, 129, 163, 168, 175, 197, 220, 228, 247, 249, 251, 253, 257, 266, 274, 277–278, 280, 282, 284, 312–313, 315–316, 318, 321–322, 344, 348, 356–357, 377, 384–385, 388–389, 391, 394, 396–398, 421, 425, 430–431, 448, 450, 453, 475–477, 494, 496, 523–524 Neghidal xxvii, 3, 5–6, 8–9, 37–40, 42–43, 49, 52–53, 56–58, 63, 67–70, 72, 140–144, 146, 149, 151, 155, 177, 184, 186–187, 190, 193, 199, 209, 213, 223, 234–259, 273, 296, 307, 323, 364–366, 368, 407–408, 410, 433, 456, 501–502, 504, 507–509, 511–513, 526, 528, 532 Nenjiang County 463 Nenjiang River 140, 206 neologisms 115, 484, 489 Nerchinsk 144 Nerchinsk Tungus 30, 33 nominalization 49, 66, 69, 151, 167, 170–171 nomina-verba 47, 149, 211, 301, 371, 413 Nonni xxviii, 140, 206–208, 217, 229 Northeast Asia xviii, 1, 36, 176, 525–527, 531 Northern Tungusic 8–10, 15, 26, 35, 38–40, 42, 44, 46, 48–49, 51–57, 60, 62–63, 65, 69, 257, 260, 526, 530 Northern Wei 528–529 Nostratic 518–519, 523 numeral classifiers 120, 159, 194 object marking 12, 215 Oirat 37, 484, 487, 498 Olcha see Ulcha Old Japanese 85–86 Old Korean 20, 132 onomatopoeia 109, 332, 347, 370 Onon 176 Oroch xvii, 3–6, 8–9, 11, 13, 37–40, 42, 44, 46, 49–54, 56–61, 65, 68–69, 71–72, 184, 256,

294–329, 336–337, 341–342, 344–345, 347–348, 350–351, 354–355, 358–359, 364–365, 369, 376, 388, 398, 407–408, 428, 436, 501–502, 507–509, 511–513, 526–527, 532 Orochen xxvii, 3–4, 6–10, 37–39, 42, 48–49, 52–53, 56–58, 66, 69–70, 72, 140–143, 146, 149, 155, 178, 184–205, 206, 208–209, 211–216, 219–222, 225, 227, 236, 238, 294–295, 297, 307, 326, 328, 410, 436, 501, 507, 526, 532 orthography xx–xxi, xxiv–xxv, 3, 21, 46, 79, 106–109, 145–146, 148, 150, 178, 260, 263–264, 328, 365–366, 407, 483 Outer Mongolia 463 overdifferentiation 106 palatal breaking 469, 479 palatal harmony 265 palatalization 93, 368, 411, 439, 469, 486 palatalness 110 palatal umlaut 469, 479 Para-Mongolic 56, 76, 86, 96, 98–99, 120, 133, 216–217, 359, 519–520, 528–530 passive constructions 114, 243, 338 Peking 32, 76, 464, 466–467 periphrastic constructions 128, 221, 266, 272, 277, 280, 283, 355, 396, 477–478 pharyngealization 37, 263, 329–332 pharyngeals 261, 264, 331 Pinyin system xx, 3, 141 pitch see prosodic features poetry 485 populations xviii, 1–2, 140–142, 229, 326, 501–503, 507, 509, 517, 529, 531 possessive constructions 48, 153, 158, 161, 215, 246, 269, 272–274, 282, 305, 388, 413, 416, 470, 491–493, 496, 498 postpositional constructions 131, 174, 196, 226, 244, 257, 266, 272–273, 276, 282–283, 305, 311–312, 340, 346–347, 357, 375–376, 382, 420, 429, 471–472, 490 Post-Proto-Mongolic 288 Post-Proto-Tungusic 38–39, 42 prefixation 193 prepositions, lack of 131 Pre-Proto-Amuric 529 Pre-Proto-Bulgharic 133, 521, 527 Pre-Proto-Ewenic 148, 167 Pre-Proto-Japonic 529 Pre-Proto-Jurchenic 56 Pre-Proto-Koreanic 526, 529 Pre-Proto-Mongolic 56, 99, 120, 133, 193, 521–522, 526–527, 529 Pre-Proto-Tungusic 52, 58, 60, 519, 529 Pre-Proto-Turkic 521, 526 privative construction 14, 175, 306, 322, 341, 357, 377, 417, 431, 453 progressive assimilation 150, 239, 265, 295–296 progressive construction 51, 68, 96, 128–129, 150, 153, 168–169, 171–172, 198, 200, 213, 220, 241, 252, 254, 267, 280, 314, 353, 356, 390, 425, 469, 493–494, 496 progressive vowel harmony 11, 47, 148

542 INDEX prohibition see negation prosodic features 47, 148, 188, 211, 239, 254, 261, 266, 300, 333, 347, 370, 412, 440, 453, 469, 488 prothesis 92 Proto-Altaic 520–521 Proto-Ewenic 142–143, 146, 148–149, 167, 176–177, 222, 265 Proto-Mongolic 133, 176, 521, 525 Proto-Nanaic 433 Proto-Samoyedic 525 Proto-Tungusic 35–75, 88, 92–93, 95, 121–122, 132–133, 142–143, 146–147, 149, 151, 155, 158, 160, 163, 167, 185–187, 208, 212, 236–238, 257–258, 262, 264, 273, 277, 323, 329, 332–333, 341, 345, 355, 368, 379, 383, 398, 410–411, 439, 450, 468, 517–519, 522, 525–530 Proto-Turkic 177 Qianlong 20–21 Qing dynasty 19–20, 103–104, 463, 465, 532 quantifiers 247, 337, 357, 372, 382, 395, 430, 470, 496; see also numeral classifiers quasiconverbs 13–14, 68, 70–71, 171, 251, 257, 286–287, 351, 477, 497 quotative constructions 228, 337, 358, 390, 398, 432, 489, 498 reduplication 123, 193, 225, 266, 342, 347, 370, 433, 476, 489, 492 relativization 199, 226–227, 286 restructuring 13, 38, 42, 61, 65–66, 261 rhotacism 211, 229 rhotacism of Daghur 211 rhotacism-lambdacism of Bulghar Turkic 521 Romanization xx, xxiv, 3, 79, 103, 107–108 Roman script xx rotation see vowel rotation Ruanruan 19 Russia 1–2, 5–6, 23, 26–27, 31, 104, 106, 139, 142, 144, 224, 235, 326, 364, 376, 436, 438, 501–504, 513–514, 532 Russian xx–xxi, 1–6, 13, 26, 29, 31, 33, 104–106, 109, 139–142, 144–146, 159, 178, 184–185, 204, 207, 229–230, 234–236, 258, 260, 262–264, 288–289, 294, 297, 303, 324, 326–327, 333, 335, 338, 341, 343, 346, 357, 359, 364–366, 368, 376, 399, 407, 409, 428, 433, 436–438, 440, 442, 452, 455–456, 483, 498, 501–502, 504–508, 510, 513, 519, 532 Russian empire 235 Russian Far East 2–3, 6, 139–140, 234, 294, 326 504–505 Russian Federation 1, 139, 260, 501 Russians 2, 76, 139, 143, 178, 338, 408, 531 Samoyedic 2, 139, 143, 176, 178, 288, 503, 519, 525–526 sandhi 240, 488 Sanskrit 132 schwa vowels xx

scripts see writing systems Selenga Ewenki 30 sentence types 13, 227, 283, 452, 476 shamanism 2, 144 Sibe xvii, 1, 6–9, 12, 14, 48, 66, 71, 103, 230, 463–464, 467, 470, 472–475, 477–479, 483–501, 519, 530, 532 Siberia 1–3, 8, 139, 142–144, 177–178, 229, 260, 501, 503–504, 506, 514, 519, 526, 530–531 Siberian Ewenki xxvii, 4, 9–10, 12, 37, 52, 64, 139–187, 190–191, 193–194, 198–200, 203–204, 206, 209–211, 213, 220, 222–224, 226–228, 234, 237–238, 242, 245–246, 248, 257, 274, 367, 436, 501 singulative 335 Sinitic languages 525 Sinkiang see Xinjiang Sino-Russian border 3, 364 Solon xxvii–xviii, 3–4, 6, 8–12, 14, 37–40, 42, 48–49, 52–53, 55–58, 66, 69, 71–72, 76, 140–143, 146, 148–149, 155, 177–178, 184–185 187, 190, 193, 198–199, 204, 206–233, 236, 299, 307, 374, 410, 479, 483, 501, 519, 526, 532 Southern Ewenki 29 Southern Tungusic 8–9, 35, 38, 41–42, 46, 50, 52, 54, 56–57, 65, 273, 393, 398, 483, 530 Soviet period 3–5, 26, 139–140, 144–146, 178, 234, 365–366, 407, 436–437, 456, 502, 504–505, 512–513 specificness 53 spirantization 148, 331, 487, 526 Spoken Manchu 7, 29, 33, 48, 66, 103, 107–108, 117, 463–482, 484, 491, 501 standardization 103 stem types 94, 123–125, 157, 164–165, 190–191, 195, 200, 214, 219–220, 222, 242, 248, 268, 275, 302, 310, 312, 315, 337–338, 345, 347, 350, 373, 381, 383, 387, 391, 394, 414, 419, 423, 426, 443, 446–448 stress see prosodic features subjunctive 131, 173, 201, 213, 221, 253, 275, 279, 318, 353–354, 381, 391–392, 408, 419, 426–427 subordination 202–203, 223, 225, 256, 319–320, 394–395, 428–429, 455, 493, 497 substratal influences 142, 176, 178, 526 superlative see comparison supine 69, 170, 173, 200, 202, 222–223, 253, 256, 275, 279, 281, 355 syllable structure of 3, 299, 440, 467–468, 487, 525 symbolic status 1, 508 syncretism 55 Tabghach 528 taboo words see lexical taboo Tangut 80, 84 Tarbagatai 483–484 taxonomy 7–10, 35–36, 48, 140–141, 298, 336 temporal–aspectual systems 13–14, 51, 66–68, 70, 72, 96–97, 123, 128, 144, 153, 168–170, 172,

INDEX 543 186, 188, 199–201, 208, 211–212, 219–222, 227, 241, 249–252, 254, 266, 275–280, 283, 301, 312–316, 318, 321, 336, 345, 347–349, 351–356, 366, 372, 382–386, 387, 389–390, 392, 396–397, 399, 413, 421–422, 424–425, 427, 430–431, 442–443, 448–450, 452–453, 475–477, 494–496, 497, 526 tense see temporal–aspectual systems terminology xviii, xxi, 14–15, 156, 176, 198, 256, 288, 510 Tibet 1, 463, 530 Tibetan 21, 132 tones xx, 11, 329, 525 topic marking 395, 477, 496, 500 Transbaikalia 1, 140, 144–145 transcription xx–xxi, xxiv, 5, 17, 20–22, 24, 77–78, 87–92, 96, 98, 100, 103, 105, 137, 145, 207, 437, 472, 484 Transeurasian 6, 10, 518 transitivity 208, 366 transliteration xx, xxiv, 79, 103; see also Romanization Tungus 2–4, 6, 30 Tungus Ewenke 206 Tungusic, definition of xviii, 1–7 Türk 528 Turkestan see Eastern Turkestan Turkic xviii, 2, 7, 10, 19, 37, 93, 133, 143, 176–178, 193, 230, 283–284, 287, 359, 468, 503, 519–521, 523, 525–529, 531 Turkish 106 Turks, historical see Türk typology xxv, 7, 13, 106, 118, 173, 257, 262, 333, 477, 525–526 Udihe xx, xxvii, 3–4, 7–9, 11, 38–40, 42–44, 46, 49–51, 53, 55–59, 63–64, 67–68, 71–72, 294–298, 301, 306–308, 310, 312–313, 317, 319–320, 323–324, 326–363, 365, 368–369, 398, 418, 465, 479, 501–502, 507–509, 511–513, 525–527, 532 Uighur, ancient 21, 528 Uighur, modern 483–484, 498 Uilta xvii, 3–6, 8–9, 23, 36–40, 42, 44, 49–4, 56–59, 62, 67–72, 294–295, 299, 307, 311, 319, 323, 368–369, 381, 393, 398, 407–418, 421–423, 425–429, 432–433, 436–462, 501–503, 506–510, 513, 520, 532 Ulcha xvii, 3–6, 8–9, 23, 37–40, 42, 44, 46, 48–60, 62, 66, 68–72, 92, 234–235, 256–258, 294–295, 299, 307, 311, 319, 323, 327, 364, 367–369, 372–373, 381, 393–394, 398,

407–436, 438–439, 441, 444–446, 448–450, 453, 455–456, 501–502, 504, 507–509, 511–513, 528, 531–532 umlaut see palatal umlaut Ural-Altaic 10–11, 70, 148, 155, 173, 282, 286, 518, 521, 525–526 Uralic 1, 10, 519, 521, 523–526, 531 Urulyungui 143 valency 266–268, 349 verbalization 13, 151, 163, 275 vocabularies 4, 20, 77–79, 87, 297, 365, 433 voice 11, 13, 43, 50, 55, 72, 87, 89, 111, 114, 147, 153, 186, 188, 198, 208–209, 212–213, 219–220, 228, 241, 250, 268, 276, 278, 296, 298, 301, 306, 313–315, 329, 336, 349–350, 366, 368–369, 372, 385–386, 413, 422–423, 439, 442, 468, 470, 525 voicing 147, 487 vowel breaking 469, 479, 486–487 vowel contraction 40, 115, 329, 349, 441, 444 vowel elision 11, 263, 265, 467, 469, 487 vowel harmony 7, 11–12, 36, 41–43, 47, 53–54, 63, 72, 92, 110–111, 117, 120, 124, 129, 146, 148–150, 185, 187, 197, 210, 214, 219, 236, 239–240, 262, 264–265, 276, 333, 367, 369, 373, 411, 415, 441, 469, 488, 526 vowel reduction 263, 469, 487–488 vowel rotation 38–39, 41, 149 word order 13, 70, 78, 131, 173–174, 203, 227, 257, 282–284, 357, 398, 473, 497 writing systems xx–xxi, xxiv–xxv, 19–20, 23–24, 76–87, 89, 93, 96, 103, 106–108, 117, 207, 234, 260, 365, 469, 483–484 Written Manchu xx, 5, 48, 103–138, 463–465, 467–469, 472–474, 477, 479–480, 483, 485–486, 488–489, 491 Written Mongol xx, 106, 523–524 Xianbei 19, 86, 527–528 Xinjiang 2, 206, 483 Xiongnu 19, 86, 527–528 Yakut xxviii, 10, 37, 72, 141, 143, 145, 159, 176–178, 184, 204, 206, 258, 260–261, 263, 288–289, 455–456, 502–503, 505–506, 510–511, 519, 526 Yeniseic 178, 503, 520 Yuan dynasty 20, 77 zero derivation 47, 211, 240