The Past Tenses of the Mongolian Verb. Meaning and Use

8/20/2019 The Past Tenses of the Mongolian Verb. Meaning and Use

1/259


2/259

Te Past enses o the Mongolian Verb


3/259

Empirical Approaches toLinguistic heory

Managing Editor

Brian D. JosephTe Ohio State University, USA

Editorial Board

Artemis Alexiadou, University o Stuttgart, Germany

Harald Baayen, University o Alberta, Canada

Pier Marco Bertinetto, Scuola Normale Superiore, Pisa, Italy Kirk Hazen, West Virginia University, Morgantown, USA

Maria Polinsky, Harvard University, Cambridge, USA

VOLUME 1

Te titles published in this series are listed at brill.nl/ealt


4/259

Te Past enses o theMongolian Verb

Meaning and Use

By

Robert I. Binnick

LEIDEN • BOSON2012


5/259

Tis book is printed on acid-ree paper.

Library o Congress Cataloging-in-Publication Data

Binnick, Robert I.

Te past tenses o the Mongolian verb : meaning and use / by Robert I. Binnick. p. cm. — (Empirical approaches to linguistic theory; 1) Includes bibliographical reerences and index. ISBN 978-90-04-21429-3 (alk. paper)1. Mongolian language—Verb. 2. Grammar, Comparative and general—ense. I. itle. II. Series.

PL473.B56 2012 494’.2356—dc23

2011035786

ISSN 2210-6243ISBN 978 90 04 21429 3

Copyright 2012 by Koninklijke Brill NV, Leiden, Te Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing,IDC Publishers, Martinus Nijhoff Publishers and VSP.

All rights reserved. No part o this publication may be reproduced, translated, stored ina retrieval system, or transmitted in any orm or by any means, electronic, mechanical,photocopying, recording or otherwise, without prior written permission rom the publisher.

Authorization to photocopy items or internal or personal use is granted by Koninklijke Brill NVprovided that the appropriate ees are paid directly to Te Copyright Clearance Center,222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.Fees are subject to change.


6/259

For Sodnomdorj Gongorand or all those who have

in one way or anotherthrough the years

encouraged my interestin the Mongolian language,

not least:James Bosson

Lucia HammarJohn KruegerNicholas PoppeWayne Schlepp

serenchunt LegdenYidamjab Meng


7/259


8/259

CONENS

Editorial Foreword ............................................................................ ixPreace ................................................................................................. xiAcknowledgments .............................................................................. xv Conventions and ranscription ...................................................... xviiAbbreviations ..................................................................................... xxi

I. Te Problem o the Mongolian Past enses ........................... 1 1. Te Mongolian Past enses .................................................. 1 1.1. Te Verbal Systems o the Mongolic Languages ...... 1 1.2. Te Problem o the Past enses .................................. 10 2. Semantic Teories .................................................................. 14 2.1. Teories Based on ense and Aspect ......................... 14 2.2. Te Finite Indicative Verbs .......................................... 20 2.3. Te Participles ................................................................ 25 2.4. Metric (Degrees o Remoteness) Teories o the

-jee and -lee enses ........................................................ 33 3. oward A Pragmatic Teory ................................................ 37 3.1. Discourse Functions ...................................................... 37 3.2. Te Evidential ................................................................. 40 3.3. Te Modality o -v ......................................................... 46 3.4. Te Inerential ................................................................ 50 3.5. Chuluu’s Critique .......................................................... 54

II. Use and Interpretation o the Past enses in the SpokenLanguage ....................................................................................... 61 1. Evidential and Inerential ..................................................... 61 1.1. Te Opposition o Evidentiality and Inerentiality ... 61 1.2. Inerential -jee ................................................................ 62 1.3. Evidential -lee ................................................................. 70 1.4. -sen in speech ................................................................. 74 2. Distal and Proximal ............................................................... 79 2.1. Distal and Proximal ....................................................... 79

2.2. Future -lee ....................................................................... 82 2.3. Te Pragmatics o Immediacy ..................................... 88 2.4. Spoken -v and the Past enses in Questions ............ 92


9/259

viii

3. Deictic and Anaphoric ........................................................ 102 3.1. Reerence imes ........................................................... 102

3.2. Denite, Deictic, and Anaphoric enses ................. 105 3.3. An Implicative Hierarchy ........................................... 108

III. Use and Interpretation o the Past enses in the WrittenLanguage ..................................................................................... 113

1. Spoken and Written Language ........................................... 113 1.1. Competing Grammatical Systems ............................. 113 1.2. Te Non-equivalence o the Written enses .......... 116 1.3. Te Language o the Internet and Levels o Usage 122 2. Te Past enses in Writing ................................................ 132 2.1. Written -v ...................................................................... 132 2.2. -sen and -sen baina ...................................................... 138 2.3. -jee and -sen baina ....................................................... 140 2.4. Distal -lee ....................................................................... 145

IV. Te Discourse Functions o the enses ................................. 147 1. Te Functions o the enses in Discourse and ext ...... 147

1.1. Te Functions o Utterances ...................................... 147 1.2. Te Tree Levels o Discourse Coherence ............... 149 2. Te Functions o the Past enses ...................................... 161 2.1. Past enses and emporal Reerence ....................... 161 2.2. Past enses and Grounding ....................................... 171 2.3. Past enses and Te opics o Treads ................... 188 2.4. Te Paragraph ............................................................... 195 3. Te Functions o the Past enses in Various Genres .... 198 3.1. Meaning, Use and Genre ............................................ 198

3.2. Diegetic and Mimetic Genres .................................... 202 3.3. Genre and enses ........................................................ 207 3.4. Past enses in the Various Genres ........................... 209

Remarks in Lieu o a Conclusion ................................................... 215

Appendix ............................................................................................. 221List o Works Cited ........................................................................... 223Index .................................................................................................... 229


10/259

EDIORIAL FOREWORD

Te present volume, Te Past ense of the Mongolian Verb. Meaningand Use, by Robert Binnick, inaugurates a new series by Brill, entitled“Empirical Approaches to Linguistic Teory”. I am proud to be themanaging editor o the series, and am joined in this enterprise by astrong team o editorial board members: Artemis Alexiadou, HaraldBaayen, Pier Marco Bertinetto, Kirk Hazen, and Maria Polinsky.

Te goal o this series is to offer contributions to our understand-ing o language in general—the key desideratum o linguistic theory—through highly empirically based studies. Te series is eclectic as totheory and does not privilege any particular theoretical rameworkover any other. We editors expect that each volume will advance ourknowledge o how human language works through solid theoreticallysophisticated description and through empirical testing o theoreticalconstructs and claims.

Dr. Binnick is particularly well known or decades o work on tense

and on Mongolian, so this study represents a joining o these two areaso his expertise. In this case, Mongolian provides the empirical basis,and the realization and value o temporal reerence constitute the the-oretical constructs that are tested by the Mongolian data.

We envision that the series will consist mainly o monographicresearch studies, but do not rule out the possibility o volumes thatare ocused collections o papers on a common theme.

We look orward to seeing many volumes appear under this imprint

in the years to come.

Brian D. JosephEAL Series Managing EditorColumbus, Ohio USA1 August 2011


11/259


12/259

PREFACE

One o the interesting eatures o the Mongolian language is theexistence o four different past tense orms o the verb. o translate‘came’, or example, one can choose (in the written language basedon Khalkha Mongolian) between ирэв irev , ирлээ irlee, иржээ irjee, and ирсэн irsen. extbooks and reerence grammars have contained various accounts o the differences between these endings, generally

vague, sometimes mutually contradictory, and ultimately inadequatelyinormative concerning this signicant topic. Te question, naturally,is why Mongolian has our different past tense endings, and how theirmeanings and/or uses differ rom one another.

Tis question may be illustrated by a couple o passages rom Erdenebulsan aral , the Mongolian translation o R. L. Stevenson’s novel rea-sure Island. In the novel, when both time and his ormer shipmatesnally catch up with the old pirate “Billy Bones,” the mysterious lodgerat the Admiral Benbow Inn, one o the pirates enters the inn to call

on Bones, only to come running out almost at once to inorm hisellows that Bill üxčixjee—“Bill’s dead!” It turns out that old Bill haslef a treasure map and soon the hero o the tale, Jim Hawkins, ndshimsel a member o the crew o a vessel sent to seek out that treasure.When the crew lands on the treasure island o the title, Jim encountersBen Gunn, marooned there years beore by the cruel Captain Flint.Panicked at the sight o Jim’s ship, Gunn asks him, “Tat’s not Flint’sship, is it?” At which Jim assures him that it isn’t, and, urthermore,

that Flint üxčixsen—“Flint is dead.”But why is it that the pirate declares that Billy Bones üxčixjee, whenJim tells Gunn that Captain Flint üxčixsen? Is the choice o differenttenses simply ortuitous, or merely a matter o style, or does it reectsome real difference in meaning and/or use?

Until the last two decades the grammatical literature was at bestunhelpul, and at worst misleading, where the past tense endings oMongolian are concerned. Binnick (1979) was an early attempt at pos-ing, and pointing towards a solution or, the problem. More than a

decade later, Binnick (1990) termed the differentiation o the tenses“pragmatic,” thereby claiming that the difference between the pasttenses was not , as previously thought, semantic, and did not have to


13/259

xii

do primarily with their literal, context-ree meanings in terms o tenseand aspect, but rather with how they are used in context. In the early

90s, when editorial pressure orced me to provide simple labels orthe endings -jee and -lee in a contribution on Mongolian, I chose, onwhat seems now to have been insufficient evidence (albeit ollowingthe approach o my 1990 article), to term them inferential and eviden-tial respectively.

Unknown to me at that time, a similar suggestion had been pub-lished in the meantime by Svantesson (1991), and in the next ewyears a number o scholars put orward similar analyses o the Mon-golian past tense system, based on essentially the modal oppositiono evidentiality and inerentiality—Wu (1995, 1996), Kullmann andserenpil (1996), Song in 1997 (and 2002), Ujeyediin Chuluu (1998),Nelson et al. (1998), Sanders and Bat-Ireedüi (1999), and serenchuntand Luethy (2000)—but as o the mid-90s, no one but Svantessonand mysel had suggested anything o the sort. (A different, thoughinsightul, approach appears in Dugarova 1991.)1

And while by the beginning o the present century the idea thatMongolian might have a past tense system based at least in part on

an opposition o evidentiality and inerentiality was no longer novel,the proposal was based largely, i not entirely, on native-speaker intu-ition, and moreover was so vague and general as to provide little, iany, guidance to the non-native-speaker wishing to properly use andinterpret the various past tense endings. Nor did it clariy the roles othe so-called “neutral” endings -v and -sen—how they differed romthe “non-neutral” endings -lee and - jee, as well as rom one another—though there are useul, albeit limited, suggestions in a number o theworks mentioned above.

Te intention in the present work is to construct an argument or,and to esh out the details covered by, the labels o “evidential” and“inerential,” and as well to provide an account o the “neutral” pasttenses. I this goal has been ullled it is due principally to the assis-tance o a native speaker, Sodnomdorj Gongor, and to a lesser extent tothe advent o the World Wide Web, which has provided easy access tosamples o a wide range o genres in contemporary Mongolian. I have

1 My less than adequate Russian has unortunately not allowed me to prot romDugarova’s work as much as I might have done.


14/259

xiii

also proted greatly by the insightul work, and the generous responsesto my queries, o another native speaker, serenchunt Legden.

It would have been useul to have known, beore I completed myresearch, o the 1998 article by Nelson et al., which anticipates many omy conclusions.2 Teir article is perhaps the most insightul work onthe Mongolian past tenses to come out o the twentieth century.

Many o the questions raised by the Mongolian past tenses are arrom ully resolved, but hopeully the present work has, at the veryleast, provided a more reliable and useul guide to usage than has hith-erto existed, and laid the oundation or urther investigation into anumber o aspects o this ascinating language.

2 Tere are some signicant differences between our conclusions, however, andsome methodological differences; while there is concern in their brie article or thedistribution o the past tense endings, there is still primary dependence on subjective

judgments on the part o speakers. Moreover, some o their conclusions are quitegeneral, so that on the whole the article strikes one as programmatic. For all that, itconstitutes a milestone in our understanding o the Mongolian tenses.


15/259


16/259

ACKNOWLEDGMENS

I would above all like to thank Sodnomdorj Gongor or letting memake use o his native speaker knowledge o, and intuition or, theKhalkha Mongolian language; or his toleration o my minimal andtruly execrable spoken Mongolian; or his numerous interesting anduseul observations, both spontaneous and in response to my ques-tions; and or his comments on the nal manuscript and corrections

o numerous errors, which I have sometimes silently corrected in themanuscript. Without his assistance, this research would literally havebeen impossible.

I would also like to thank serenchunt Legden or her encourage-ment; or her communications, which have been invaluable and romwhich I have beneted greatly; and or both the ne textbooks sheproduced with Sharon Luethy, and the Web site (http://www.indiana.edu/~celcar/intermediate/mongolinter.html) to which she contrib-uted useul grammatical comments. As regards the past tense endings,

her works are amongst the most insightul published, and I regret thatI only became aware o the Web site (and through it, her textbooks)when I had with much travail independently re-discovered what shehad already written about the spoken language. I would like to thankher, too, or her comments on the nal manuscript, her observationson a number o errors, which I have sometimes silently corrected, andor kindly providing translations or a dozen or so colloquial examplestaken rom the Internet.

I would similarly like to thank Diane Nelson or providing me witha copy o the 1998 Nelson et al. article, which proved, alas, to havebeen a very signicant piece o work, which I much regretted some-how ailing to nd in my earlier bibliographic searches.

Also due thanks are six ormer or present colleagues at the Univer-sity o oronto. From the Scarborough campus: Harald Ohlendor,who checked most o my translations rom the German; Michal Schon-berg, who provided some o the translations rom Dugarova with theassistance o Mr. Anatoly Oleksiyenko (whom I also wish to thank);

Corrine Beauquis, who lent me her intuitions as a native speaker oFrench; and Yoonjung Kang, who arranged or Kenji Oda (whom Ialso grateully thank) to provide me with a précis o the 1993 article by


17/259

xvi

Hashimoto (which, despite its title, is in Japanese, and accompaniedby a wholly inadequate summary in English). From the St. George

campus, I wish to thank Christina Kramer and Wayne Schlepp, theormer or arranging or Jan Schallert (whom I also hereby thank) toprovide the translation o one o the passages rom Dugarova’s book,and the latter or his helpul responses to various queries.

Acknowledgment is also due to the anonymous reviewers whosenumerous suggestions have gone ar to helping to improve the nal version o this book, and to all those at Brill who have worked so hardto transorm the manuscript into this volume, especially to MirjamElbers, its production editor.

Finally, I would like very much to thank my riends and loved onesor understanding, and in the main orgiving, my neglect o them dur-ing the writing o this book, and or their support throughout.

Te research upon which this work is based was partly unded, andlargely made possible, by a grant rom the University o oronto. Mucho the work on this book was done while on a sabbatical research leavein the Winter session o 2008.


18/259

CONVENIONS AND RANSCRIPION

C

Unattributed glosses and comments in italics are those o SodnomdorjGongor, the native speaker o Khalkha Mongolian who gave me thebenet o his native-speaker intuitions. Where glosses both in Englishand some other language such as French or German are given, the

English gloss is mine and the other gloss is that o the source.Unless otherwise noted, all the word analyses and detailed glosses

accompanying the examples (as in example 8 below) are mine. Insome cases it is impossible to provide a unique and precise Englishequivalent and the glosses should accordingly be taken as purely nom-inal. I have separated affixes rom their stems using a dot, so: bai.na,and have arbitrarily assigned the -g- used to separate long vowels tothe verb stem rather than to the suffix (thus baig.aa ‘be-’). Inexamples drawn rom other works, a hyphen indicating a morpheme

boundary (e.g., nom-un) has been replaced with a dot (nom.un).1

8. Önöö öglöö bi zurgaan tsag.t boso.v. this morning I six hour- get up- ‘Tis morning I got up at six o’clock.’ (Street 1963: 122)

In the glosses, the terms and abbreviations ollowing the hyphens(e.g., and in example 8 above) represent grammatical cat-egories. (See the table o abbreviations.) Te -x orm is simply glossedin this work as “innitive/uture verbal noun” but the -sen ormis glossed as when it is a predicate and as (perective verbalnoun or participle) when it is a modier or noun; precedes an auxil-iary verb (as in example 3a below), a copula, postposition, or the ques-tion particle be/ve; or ollows another verbal noun affix (as in examples

1 It should be noted that in the vertical-script language, case (and some other)affixes are generally (though not in all contexts) written as separate words rom theirstems, so that nom.un, or example, actually is written as nom un. A hyphen is conven-tionally used (as in nom-un) to indicate the connection between the two. See Grøn-bech and Krueger (1955: 20), Poppe (1964: 30), and the examples in Kullmann andserenpil (1996: 84–100).


19/259

xviii

37–39). Although somewhat arbitrary, these labels generally reect itsuse in the various types o examples in question. Te non-past (pres-

ent-uture) tense affix simply has been labeled (“present”), as in(3a). a, the plural/polite second person pronoun similar to Frenchvous has uniormly been glossed ‘you (plural)’. In contrast, singular či is glossed simply as ‘you’.

3a. Xugar.san bai.na. Break- be- ‘It’s broken.’ (Sanders and Ireedüi 1999: 191)

Te affix -čix - has not been glossed. It likely has no counterpart oradequate gloss in English, but in any case no completely satisactory,denitive statement o its meaning or use has appeared, publishedaccounts tending to be brie and vague.2

ranslations (in single quotes) placed to the lef o the name, date,and page o the source publication (as in examples 8 and 3a above) arethose o the source publication; translations to the right o (i.e., below)the source name (as in example 16 below), are mine.

16. Dašdorjiin Natsagdorj 1906 on.d tör.jee. (Yatskovskaya 1976: 8) Dashdorjiin Natsagdorj 1906 year- be born- ‘Dashdorjiin Natsagdorj was born in 1906.’

In the present work, affixes containing vowels are represented by theirwritten orm containing e, e.g., -lee.

Examples in the literature are sometimes cited in phonetic or phone-mic transcription. When such examples are used here, the transcrip-tion used in the source (e.g., Ramstedt 1902) is employed, with somemodications, which have been noted. On the Internet, Mongolian

2 Tus Street (1963: 83) simply labels it “perective,” and in a similar vein, Kull-mann and serenpil (1996: 133) call it a marker o “[ully] completed action,” with “aslight perect meaning” and gloss ted duusčixjee as ‘they’ve nished (it)’ (rom duus-‘nish’). Sanders and Bat-Ireedüi (1999: 87), however, say it orms “intensive” verbs(terminology echoing Poppe 1951: 51), indicating that the “action is complete andunexpected,” and contrast bosov ‘got up’ with bosčixov ‘sprang up’.


20/259

xix

is not inrequently written in the Latin alphabet and such exampleshave been noted, and presented as they originally appeared. Examples

written in the old vertical script have been transliterated using a airlystandard transliteration. Where such examples are already transliter-ated in the source rom which they have been taken, that translitera-tion is generally used here.

Mongolian examples written in the Cyrillic alphabet are transliter-ated in this book into the Latin alphabet. Standard schemes o translit-eration have been ollowed, especially those employed by Street (1963),Vietze (1974), Sanders and Bat-Ireedüi (1995, 1999), and Kullmannand serenpil (1996). Te letters or which there is some variancebetween transliteration schemes are shown in able 1 below.3

able 1

Cyrillic Street Vietze(“popular

transliteration”)

Sandersand

Bat-Ireedüi

Kullmannand

serenpil

Tis book

е ye, yѳ je yö ye ye, yöë yo jo yo yo yo

ж j dsch j j jй i j i i iѳ ѳ ö ö o öу ʉ ü ü u üх x ch kh xц c ts ts ts tsч č tsch ch ch čш š sch sh sh šъ ә ” ”ы ɨ y y ii ii

ь ә j ’ (or i) ’э e e e e eю yu, y ʉ ju yu, yü yu yuя ya ja ya ya ya

3 serenchunt points out (p.c., October, 2008) that “Now Proessor George Karaand other scholars use ‘w’ or Mongolian and it has been accepted by manypeople. So ‘w’ is better than ‘v’ . . .: yavaad > yawaad, yavj > yawj.” Te use o the letter is so well established in the grammatical literature, however, that I have chosennot to use here in its place. Similarly, it has become common to transliterateCyrillic as , as in the name (Narmandah) o one o the authors o Nelsonet al. (1998), but this book continues the tradition o using Latin .


21/259

xx

Where Russian is concerned (or example, names and titles in the re-erences), the transliteration employed here is a airly standard one.

Te Cyrillic letter , transliterated as in the case o Mongolian,is given as in that o Russian. Also, in the Russian transliterations, is represented as (and as ). Tere are a ew otherdifferences in the transliterations, reecting differences between theCyrillic alphabets o the two languages, but none should cause anytrouble or the reader.


22/259

ABBREVIAIONS

ABL ablative caseACC accusative (object) caseAGVN agentive verbal nounCAUS causative affixCOM comitative caseCONDC conditional converb

CONC continuative1 converbCOPP copular particle2

DA dative caseEMPHP emphatic particleGEN genitive caseHABVN habitual (requentative, generic) verbal nounIFVN innitive/uture verbal nounIMP imperativeIMPFC imperective converb

IMPFVN imperective verbal nounINFERP inerential particleINSR instrumental caseIPA International Phonetic AlphabetMODC modal converbMODP modal3 particle

1 As with many aspects o the grammar, even though there is general agreement asto the meaning o this orm, there is no standard terminology. Poppe (1951: 89) callsit the Konverbum abtemporale and describes it as meaning “eine Handlung, seit derenEintritt bereits die Haupthandlung ausgeübt worden ist” (“an action, since the begin-ning o which the main action has been perormed”). Vietze (1974: 140) says similarlythat as the predicate o a subordinate clause it can be translated “seit” [‘since’]. Withan auxiliary verb such as bai- ‘be’, he notes, it conveys a “ortdauernden Zustand bzw.eine (immer noch) andauernde Handlung” (“continuous state or a (still) continuousaction”). Similarly, Sanders and Bat-Ireedüi (1999: 105) call it the “continuous” con-

verb. Tey note that with bai- it can be translated ‘keeps on’; hence Kullmann andserenpil (1996: 168) call it the “progressive” converb.

2 Yum, mön, and bii are identied in the present work as copular particles (COPP).Kullmann and serenpil (1996: 337ff.) label them modal particles.

3 Tis is the term used by Sanders and Ireedüi (1999: 83) or the particle č . Tisparticle has a wide range o uses, described by Kullmann and serenpil (1996: 346–8),who call it a “ocus” particle. (C. note 29 on p. 206 on l.)


23/259

xxii

NEG negative affixNOM nominative (subject) case

PASS passive affixPAS past tensePFC perective converbPFVN perective verbal nounPL pluralPRFU present/uture (non-past) tenseQP question particleRP reexive-possessive affixERMC terminal converbVOL voluntative (rst-person imperative)


24/259

CHAPER ONE

HE PROBLEM OF HE MONGOLIAN PAS ENSES

1. M P

1.1. Te Verbal Systems of the Mongolic Languages

Te problem o the past tenses o the members o the Mongolic (or“Mongolian”) language amily is one o the most challenging puzzlesin the study o that amily and has long resisted solution. From a cer-tain point o view, the problem is easily stated. Some members o theamily have three or more different affixes which seem to be markerso the past tense, so that in Khalkha, or example, irev, irjee, irlee, andirsen all can translate, and be translated by, the English past tense verbcame. Te question is how these orms differ in meaning and/or use,assuming that they do differ in some way (or ways).

From another point o view, however, the problem o the Mongo-lian past tenses is not so easily stated, first because o the complexitieso the term “Mongolian,” and secondly because o the complexities othe type o verbal system typical o these languages. Te reader mayfind some background inormation on both these topics helpul inunderstanding the problem o the Mongolian past tenses.

Te issue o what precisely is meant by “Mongolian” is relevanthere because the puzzle o the multiple past tenses ound in Khalkhaextends not only to the majority o spoken Mongolic dialects and all

o the corresponding written languages today, but goes back to the very beginnings o the recorded history o the members o the amily.o discuss the problem purely within the context o Khalkha dialectsand/or the Mongolian language written in the Cyrillic alphabet is todeprive onesel o insights rom other members o the amily and toartificially limit the scope o the inquiry at the outset. o a certainextent there is only a terminological issue involved here, or it is ofeneasier (but not necessarily precise) to reer to aspects o a “Mongolian”grammatical system than to speciy a more specific language or lan-guages. Te problem with this is that the term “Mongolian” has beenused to reer to many different spoken and written languages, that thestatus o various members o the amily as independent languages or


25/259

2

as mere dialects is by no means clear, and the historical and classifica-tory relationships o the members o the Mongolic language amily

remain controversial.At present the Mongolic dialects spoken in Inner Mongolia, Mongo-

lia (the ormer Mongolian People’s Republic, at an earlier time OuterMongolia), the adjacent Buriat Republic in Russia, and nearby partso China and Asiatic Russia, orm the bases or our different writtenlanguages. Spoken Khalkha, the language o the vast majority (some80% or more) o the inhabitants o Mongolia, is the basis or writ-ten Khalkha (also reerred to as “Mongol” or “Modern Mongolian”),which utilizes the Cyrillic alphabet. Te Buriat language, written ina slightly different Cyrillic script, differs rom written Khalkha morethan the corresponding spoken languages differ rom one another,partly due to different spelling conventions and partly to the muchgreater effect o Russian on Buriat than on Khalkha, especially where vocabulary is concerned.

Te various Khalkha and Buriat dialects are airly close to many,perhaps most, o the Mongolic dialects spoken in Inner Mongolia,such as Chakhar. In Inner Mongolia, a somewhat modernized version

o the old vertical script language is used, but is read as i representa-tive o the modern spoken language, much as English-speakers writestraight but read the word as the spoken equivalent, strate (phoneti-cally, something like [strejt]).

Te verbal systems o these various “dialects” and “languages” are orthe most part essentially the same, though the endings are spelled di-erently, pronounced somewhat differently, and consequently appearhere in different transliterations.

It is a different matter where the dialects which have variously been

called Oirat or Kalmuck are concerned. Historically, they ormed thebasis or a written language in a modified version o the Mongolian vertical script. Tis written language still has currency amongst theOirats o China, though the Kalmucks in Russia today use a writtenlanguage with a Cyrillic alphabet. Both the spoken and written lan-guages differ quite a bit rom Khalkha and the dialects o Inner Mon-golia, and clearly orm a distinct language (or languages) rom them.At the same time, there are sufficient similarities in the verbal systemso Khalkha, Buriat, and Inner Mongolian on the one hand, and Oiratand Kalmuck on the other, or the grammars o the two groups olanguages to be mutually enlightening. What Bläsing (1984) has to sayabout the finite indicative verb orms o Kalmuck, or example, bears


26/259

3

on the analysis o the corresponding verb orms o the languages inthe ormer group, though to be sure there are important differences

between the grammars o the two.Finally, there are those Mongolic languages whose verb systems are

sufficiently different rom the preceding as to render their analysisquite distinct rom the systems o the first two groups. Tis groupincludes the Mongolic languages o China outside o Inner Mongo-lian: principally Dagur (Daur) in the ormer Manchuria and in Xin- jiang; and in the province o Gansu, u (Monguor), Bonan (Baoan)and Dongxiang (unghsiang). Te verbal systems o the members othis third group require entirely separate treatment, and are ignoredhere. Monguor and Moghol (a now deunct language o Aghanistan)differ rom Mongol in much, much more than the pronunciation otheir names.

Discussion o the Mongolic languages also raises the issue o theirrelationship to the other members o the Altaic group o languages.Aside rom the Mongolic languages, this group also includes theurkic and ungusic language amilies, the best-known members owhich are, respectively, urkish and Manchu. While ormerly there

was largely a consensus that these three amilies ormed branches oan Altaic super-amily, possibly along with Korean and/or Japanese,many, perhaps even most, scholars today believe the Altaic languagesto be a group o genetically unrelated amilies which have convergedwithin a language union (Sprachbund ).1 For present purposes it reallydoes not matter which is the case. Despite the paucity o cognates andhence systematic sound correspondences linking the three amilies—letalone the three together with Japanese and/or Korean—there are sig-nificant structural similarities common to all five, which, despite real

differences between them, justiy speaking, at least in some regards, oan Altaic type o language.2

Tus while the present work largely restricts itsel to a discussiono, and principally draw its data rom Khalkha and relatively closely

1 Poppe (1965)—ollowing Ramstedt (1952–56), Menges (1975) and most strongly,Miller (1971, 1991, 1996) represent the school o thought believing in a genetic rela-tionship between most or all o the five; this view is criticized by Clauson (1956),Doerer (1963, 1985) and Vovin (2005).

2 Where the verb is concerned, the ormation and use o aspectual, tense, andother markers o verbal categories show remarkable similarities across the Altaiclanguages.


27/259

4

related Mongolic languages, we have occasion as well to reer to atleast one other Altaic language, namely urkish.

Te Mongolic languages, like the other Altaic languages, are note-worthy or their extraordinarily rich and complex verb systems, someacets o which should be pointed out, especially or the reader una-miliar with languages o this type.

Tey have numerous auxiliary verbs and copular particles, whichcan be strung together, along with content verbs, ofen to orm longstrings o verbs, such as those italicized in the examples in (1) below.(Te italicizations in the examples are mine.)

1. a. Šaardlag.iig arai čamai xanga.j Demand- with great difficulty provide-

amji.j bai.na. succeed- be-

‘[Te restaurant] is just barely able to satisy the demand . . .’ (lit. ‘issucceeding satisfying’ ) (Street 1963: 149)

b. er ene tuxai med.sen bai.j taar.na. Tat this about know- be- match-

‘He must have known about this.” (Kullmann and serenpil 1996: 209)

c. . . . bid nar.t yar’.j tusal.j ögö.x.güi yuu? We - speak- help- give-- ‘will you please tell us . . .’(http://www.mongolianoralhistory.org/samples/transcriptions/

R060402.xml; http://www.mongolianoralhistory.org/samples/translations/EN060402.xml)

d. er nada.d exel.j zaxia biči.x.güi bol Tat me- begin- letter write-- i

bi tüün.d zaxia biči.x.güi bai.saar bai.x I that- letter write-- be- be-

bol.no. become-

‘I he does not write me a letter first, I won’t write a letter to him.’(Kullmann and serenpil 1996: 199)

Combinations o auxiliaries or copulas with other verbal orms mark various distinctions o tense and aspect. In (2a) the “continuative converb” (-seer ) combines with the copular verb baix ‘to be’ to orm acontinuative perect (‘have been waiting’). In (2b) the “imperective verbal noun” (-ee) combines with the copula to orm a kind o pro-gressive not unlike the English progressive construction, though i thesentence is positive (without the - güi ‘not’), the “imperective converb”(- j) is used instead o the corresponding verbal noun, as in (2c).


28/259

5

2. a. Bi tüün.iig xülee.seer bai.na. I that- wait- be-

‘I have been waiting or him.’ (Sanders and Bat-Ireedüi 1999: 105) b. Yuu č xiig.ee.güj3 bai.na. What do-- be- ‘(I’m) not doing anything.’ (Sanders and Bat-Ireedüi 1999: 63) c. Yuu xii.j baina ve? What do- be- ‘What are [you] doing ?’ (Sanders and Bat-Ireedüi 1999: 63)

Mongolian verbs today have at least our different verbal nouns orparticiples (in Khalkha, those marked with the affixes -sen, -ee, -x,

and -deg ) and a dozen converbs, differentiated mostly by temporalrelationships; or example, the “perective” or “perect” verbal nounormed with -sen and the corresponding perect(ive) converb in -eed contrast with their “imperective” or “imperect” counterparts in -ee and - j as perective aspect to imperective, or sometimes as past tenseto present.4

Te deverbal, non-finite verb orms entering into these combina-tions have other uses as well. As is typical o Altaic languages, in theMongolic languages there are three different types o deverbals—(1)nominalizations like itgel ‘aith, belie’ rom itgex ‘to believe’ and idee ‘ood’ rom idex ‘to eat’; (2) verbal nouns (essentially participles) likethe “perect” or “perective” suusan ‘(having) sat’ and “imperect” or“imperective” xiigee ‘doing’; and (3) converbs like the “perective” yavaad ‘having gone’ and “imperective” yavj ‘going’.

Verbal nouns can be used just like any other noun, but they alsocan serve as the predicates o sentences, with (3a) or without (3b)accompanying copulas. Both verbal nouns, like the non-past one or

“infinitive” in (3c), and converbs, like the “terminal” or “terminative”converb in (3d), are used as the main verbs o subordinate structureswhich unction syntactically as phrases, but semantically as clauses.Converbs can also occur on their own as adverbs (3e). Converbsare also the etymological sources o conjunctions such as bögööd ‘and’in (3), which is the “perective converb” o an obsolete verb (bü-)

3 I have arbitrarily assigned the -g- used to separate long vowels to the verb stem

rather than the suffix. Te root o xiigeegüi is xii-.4 raditionally perective aspect and imperective aspect are considered to contrastas marking complete action vs. incomplete. Tus the verb in the English sentence Iwent home would be considered to be perective, while the progressive construction,e.g., I was going home, would be imperective. Tis pre-theoretical understanding othe aspects is unsatisactory, but should suffice or present purposes.


29/259

6

meaning ‘be’. Bolon ‘as well as’ (3g), too, has the orm o a converb—the “modal” converb—o a verb, namely bolox ‘to become’. (Te ital-

izations in (3) are mine.)

3. a. Xugar.san bai.na. Break- be- ‘It’s broken.’ (Sanders and Ireedüi 1999: 191) b. er ajil.d.aa yav.san. Tat work-- go- ‘He has gone to work.’ (Street 1963: 207) c. Üün.iig marta.x.aas.aa ömnö xii. Tis- orget--- beore do-

‘Do it beore [you] forget .’ (Altangerel 1998: 33) d. Namar bol.tol ted end ajilla.na. Autumn become- those here work-

‘Tey’ll work here until [it becomes] autumn.’ (adapted rom Sandersand Ireedüi 1999: 106)

e. Namaig bitgii širte.n xar! Me- don’t stare- look- ‘Don’t stare at me!’ (Kullmann and szerenpil 1996: 158) . Minii naiz German.d sur.č baig.aa My riend Germany- study- be-

bögööd uda.x.güi ir.ne. and delay-- come-

‘My riend studies in Germany and will come soon.’ (Kullmann andserenpil 1996: 299)

g. er oros bolon xyatad xel.eer sain Tat Russian and Chinese language- well

yar’.dag. speak-

“He speaks Russian and also Chinese well.” (Kullmann and serepil

1996: 300)

In addition to these complexities, there are those presented by thetense/aspect morphology o the content verbs. Te topic o the pres-ent volume comes rom the act, already noted above, that Altaiclanguages typically have, in addition to their present tense or tenses,more than one past tense. In the case o the Mongolic languages, thismultiplicity o past tenses presents a real problem, as it has never beenmade precisely clear in what manner o meaning and/or o use these

tenses differ, assuming that they do. While every textbook and gram-mar necessarily comments on the various “tense” endings, we shall seethat there has been a wide divergence in opinions, ofen based on little


30/259

7

more than naïve speaker intuitions and sometimes on mistaken com-parisons with the quite different verbal systems o other languages. It

is only in the last couple o decades that linguistic research ocused onthis problem has seriously advanced its solution.

o be specific, members o the Mongolic language amily typicallycontain three different past tense suffixes, as in the vertical-scriptexample (4) below.5 For the sake o convenience, these are reerred tohere, respectively, as the past tenses in -v, -lee, and - jee.

ba4. ta nom ungsi- la

jai you (plural ) book read- ‘you read a book’ (Wu 1995: 94)

Because o vowel harmony, each ending in a typical modern Mongoliclanguage potentially appears in between two and eight different orms,because the vowels in the endings generally adjust themselves to the vowels o the stems as regards rounding and a eature until recentlycalled palatality or backedness. In the old written language in vertical

script, the letters represented in transliteration as g and γ reflect velarsounds which obey vowel harmony, g occurring with “ront” vowels,γ with “back” ones. Te importance o this or present purposes issimply that in written Khalkha, or example, the -lee ending and the-sen ending have our different orms each (table 2):

able 2

Stem -lee orm -sen orm

ava- ‘take’ (back, unround) avlaa avsandeve- ‘wave’ (ront, unround) devlee devsenoro- ‘enter’ (back, round) orloo orsonög- ‘give; do or someone else’ (ront, round) öglöö ögsön

5 Te past tense endings and their uses are, with some noteworthy exceptions, very similar in most Mongolian dialects and languages. For surveys o the orms, seePoppe (1955) and Wu (1996). For example, Bläsing (1984: 38) says, “Sieht man einmal

ganz ab von Partizipialeinheiten (z.B. -сн . . .), so konkurrieren im Kalmückischen undKhalkha-Mongolischen . . . drei Formen au der Zeitstue der Vergangeheit miteinander(klm. -ла, -в, -ж; xlx. -лаа, -ав, -жээ).” ‘Aside rom participial units such as -сн, inKalmuck and Khalkha there compete three orms in the past tense: Kalmuck –ла [-la],-в [-v], -ж [-j]; Khalkha -лаа [-laa], -ав [-av], -жээ [-jee].’ (My translation—)

{ }


31/259

8

Te -jee ending is unusual in not undergoing vowel harmony. Tusavjee ‘took’ but also irjee ‘came’ and ögjee ‘gave’. Both - jee and -lee

also have short, vowelless variants, or example in questions beorethe question particle uu/üü: Bag š xödöö yavj uu? ‘Did the teacher goto the countryside?’ (Kullmann and serepil 1996: 186). As well, theending - jee also has variants in -č ollowing certain consonants, e.g.,zogsč(ee) ‘stopped’.

In written Mongolian in the old vertical script, the three endings calledhere -v , -lee, and - jee, are usually transliterated as, respectively, -ba/-be, -luγa/-lüge (modern -la/-le), and - ǰ uqui/- ǰ üküi/-čuqui/-čüküi (modern- ǰ ai, - ǰ ei, -čai, -čei).6 In written Khalkha in Cyrillic script, the corre-sponding orms are -в (-v ), -лаа/-лээ/-лоо/-лѳѳ (-laa/-lee/-loo/-löö),and -ж(ээ), ч(ээ) (- j(ee)/-č(ee)).7

As noted in the section on Conventions, in the present work, affixescontaining vowels are reerred to by their written orm containing e,e.g., -lee. (It is quite common in the literature, however, to reer to theaffixes using their back-vocalic orms, e.g. -laa.)

In addition to the three finite indicative past tense endings, there aretwo verbal nominal (participial) orms that belong in any discussion

o the past tenses. Te first o these, the verbal noun whose endingin vertical-script Mongolian is transliterated as - gsen/-γsan (= writtenKhalkha -сан/-сэн/-сон/-сѳн, -san/-sen/-son/-sön), usually termed the perfect or perfective (sometimes the past ) verbal noun,8 ofen unctionsas a predicate bearing some sort o past meaning. As a participle itofen bears a relative past (i.e., anterior) meaning (e.g., yavsan ‘gone,having gone’). Te second is the verbal noun ormed with the affix

6 At various times there also occurred the orms -bei/-bai and -lügei/-luγai,-legei/-laγai, -luqa, -luqai (Poppe 1955: 265.; Weiers 1969: 149, 153.). Wu (e.g., 1996:77) cites the modern orms -la/-le. Other orms include - ǰ ükü/- ǰ uqu (Poppe 1955:265.; Weiers 1969: 158.). Wu (e.g., 1996: 73) cites the orms - ǰ ei/- ǰ ai/-čei/-čai or themodern language. As in the Cyrillic script, the consonants j and č vary according tothe final sound o the stem the ending.

7 Depending on the final sounds o the stem the ending -v is added to, there maybe a (harmonizing) linking vowel, as in ол-о-в (ol-o-v ) ‘ound’. C. oro-v ‘entered’. Inthe present work, this linking vowel is assigned in analyses to the stem.

8 Sanders and Bat-Ireedüi (1999: 25) reer to it as the perfective verbal noun, a prac-tice generally ollowed here. Tis is more accurate than “perect,” but the latter termis traditional and hence amiliar, is shorter, and should occasion no conusion as longas the actual meaning and use o the orm are borne in mind.


32/259

9

- ge/-γa (= Khalkha -аа/-ээ/-оо/-ѳѳ, -aa/-ee/-oo/-öö).9 Tis is usuallycalled the imperfect or imperfective (also the present ) verbal noun.10

As with any other verbal nouns in the Mongolic languages, theseparticiples may serve, without any accompanying copular particle orcopular auxiliary verb, as the main predicate o an independent clauseor sentence, as uilsan does in example (5). In such sentences, there islittle or nothing to distinguish these participles rom finite verbs.

5. seren iči.sen.d.ee uil.san seren be ashamed--- cry- ‘When he was ashamed, seren cried.’ (Poppe 1970: 134)

Consequently, ollowing the lead o some grammarians, the perective verbal nominal (participial ) orm is reerred to in the present work asthe past tense in -sen, since there is ample reason to regard the -sensuffix as a ourth past tense marker, alongside the finite tense endings-jee, -lee, and -v . (It is also, however, reerred to as the “perective” par-ticiple, and when, as in example (3a), it is unctioning as a participle,it is labeled rather than .)

3. a. Xugar.san bai.na. Break- be- ‘It’s broken.’ (Sanders and Ireedüi 1999: 191)

Tere is little reason, however, to similarly label the -ee ending a pasttense marker, since it is debatable whether it really unctions as a past-tense ending, as it only does so when negated; compare the sentencesin (6).

6. a. Minii sar.iin temdeg . . . sar ir.ee.güi. My month-genitive mark . . . month come--

‘I haven’t had my period or . . . months.’ (Sanders and Bat-Ireedüi1995: 128)

9 Mongolian regularly inserts g between two long vowels (including diphthongs);hence the -ee ending appears as - gaa, -gee, -goo, or - göö immediately ollowing a long vowel (as in nuugaa ‘hiding, concealing’) or diphthong (as in baigaa ‘being’). As hasalready been noted, here this linking consonant is, however, arbitrarily assigned inanalyses to the stem, as in baig.aa.

10 In Sanders and Bat-Ireedüi (1999: 63), the imperfective verbal noun. As with theperective (see note 8 on p. 8), we generally ollow their practice, while not excludinguse o the term imperfect so long as the actual meaning o the orm is borne in mind.


33/259

10

b. Mongol uls ix xögji.j baig.aa. Mongolia(n) nation greatly develop- be-

‘Die Mongolei entwickelt sich sehr.’ (Vietze 1974: 84)‘Mongolia is being greatly developed.’

1.2. Te Problem of the Past enses

Although there is general (though not universal), agreement that the various past tense orms o Mongolian differ in meaning and/or use—afer all, there must be some reason why a language would have thedifferent past tense orms ound in example (4)—it remains debatable,and, to this day, much debated, how they so differ.11

ba4. ta nom ungsi- la jai you (plural ) book read- ‘you read a book’ (Wu 1995: 94)

Tere has been a wide divergence o opinion regarding almost everyaspect o the meanings and uses o these orms. As Chuluu puts it (in

chapter 4 o his doctoral thesis, Ujeyediin 1998),12

. . . it is clear that any o the our morphemes [-v, -jee, -lee, -sen] can bechosen by the speaker to indicate the past tense and the choice o su-fix is perhaps up to the speaker. Tere is no doubt that they can be allregarded as past tense markers, the question is, however, why Mongolianhas our different morphemes or the same past tense. . . . It has been gen-erally concluded that these morphemes are distinct in certain ways butattempts to establish clear and reliable criteria or distinguishing themare not conclusive.

Until the last twenty years or so almost all scholars accepted some ormo the theory, first articulated by Ramstedt (1902), but having its roots

11 Sanžeev (1973: 92) states that “the differences between [the three past tense]orms, which are almost imperceptible, are still debated by students o Mongolian.”(Oddly enough, he also says, “the three past tenses are used interchangeably.”) It isinteresting, and suggestive, that while Sodov’s Foreign Literature Reader (1967: 60) usestörjee in Šyekspir 1564 ond . . . törjee ‘Shakespeare was born in the year 1564’, Altanger-el’s English-Mongolian dictionary (1998: 40) uses törsön in Šyekspir 1564 ond törsön.

12 His name is given in his thesis as Ujeyediin Chuluu. He has published also underthe name Chaolu Wu (i.e., in the Chinese style, Wu Chaolu). I generally reer to himas Chuluu, but reer to the various works by the name o the author given in eachcase.

{ }


34/259

11

at least as ar back as the grammar o Schmidt (1831), that the distinc-tions between the endings have to do with time.13 I call this account

the “semantic” theory. Few i any scholars other than the ones notedin the preace had, prior to the mid-to-late 1990s, suggested a differenttype o account, one based on modal distinctions, a type o accountcharacterized here as a “pragmatic” theory. In the last decade or so,elements o essentially such a modal account have been presented, orexample, in the grammar by Kullmann and serenpil (1996), worksby Song (1997, 2002), the thesis o Ujeyediin Chuluu (1998), and thetextbook o serenchunt and Luethy (2000) (as well as on a Web sitewith grammatical notes by serenchunt).14

However, there are a number o inadequacies shared by these vari-ous works. None o them presents much (i any) evidence other thansheer native speaker intuition to support their accounts; none pres-ents a sufficiently detailed account to allow the non-native-speaker toproperly use and interpret the various orms;15 and none o them makeclear that, as in the case o French, the tense/aspect systems o spoken

13 Already in Schmidt (1831) something approaching Ramstedt’s theory is present,i only in embryo. Te -ba (-v) past he calls (p. 56) the “Präteritum imperectum” (i.e.,imperect past). Schmidt’s “Perectum” (‘perect’, p. 57) includes both the -luγa (-lee)and - ǰ uqui (- jee) orms, which he glosses using the German present perect:

i. bi γar.iyar bari.luγ-a I hand- take- ‘ich habe es mit der Hand ergriffen’ (‘I have seized it with my hand’)ii. qola γa ǰ ar.ača ire. ǰ üküi ar land- come- ‘er ist vom ernen Landen gekommen’ (‘he has come rom a distant country’)14 As noted in the preace, I only became aware o this Web site (which at the pres-

ent time is available at http://www.iub.edu/~celcar/intermediate/mongolinter.html)and, through it, the serenchunt/Luethy textbooks, at the point at which my researchon the spoken language with a native speaker, Sodnomdorj Gongor, was almost com-plete, and I had just realized that the written language was a distinct problem romthat posed by the spoken language. I had read Kullmann and serenpil’s grammar,but obviously had not taken in what they had to say on this topic, since their accountcame as news to me when I reerred to their grammar while reading serenchunt’stextbook. Whether their grammar (unconsciously) influenced my research, I cannotsay. As I noted in the preace, too, I only became aware o Nelson et al. (1998) afercompleting my research and the drafing o this book.

15 Interestingly enough, the Mongolian phrasebook in the popular Lonely Planet series (Sanders and Bat-Ireedüi 1995) simply omits any mention o the verb endings-lee and - jee in its grammatical sketch, despite the act that these are amongst themost commonly used endings in the language. (Tese authors’ 1999 textbook does,however, discuss these affixes.)


35/259

12

Mongolian and written Mongolian are not identical.16 Also, rom thetheoretician’s perspective, most o these works ail to deal with Mon-

golian tense/aspect in terms o contemporary theories o semanticsand pragmatics.

Accordingly, the present study aims to present a new and compre-hensive account o the meaning and use o the past tense endings inspoken and written Khalkha, which holds true (at least in broad out-line) as well or the closely allied Mongolic languages (such as InnerMongolian, Buriat, and to a large extent even Kalmuck/Oirat), andthereby provides implicit suggestions or new directions in the studyo yet other members o the amily, such as Dagur (Daur) and Mon-guor (u).

Te main points I argue or here are these:

a) Te Khalkha verbal system is undamentally an evidential/inferen-tial system. Tat is, apart rom tense (past vs. non-past), perhapsthe most important distinction marked in the verb is between whatthe speaker personally can vouch or (evidential ) and what thespeaker cannot—what he or she is merely reporting or inerring

(inferential ), or has reshly discovered (mirative).b) Te system also makes a undamental distinction between the dis-tal and the proximal . Tus -lee can be used as a present—or evena uture—time marker, as well as a past time one, so long as thesituation recounted is in some way part o the speech act situation(situation o utterance), that is, part o what is happening when

16 When I finished the first draf o this book, I discovered that Nelson et al. (1998)did state that “Mongol has grammaticalised . . . stylistic eatures related to spoken vs.written discourse” (p. 115) and generally anticipated my conclusions regarding thedistribution o the orms and their different uses in spoken and written language. osum up what they have to say regarding spoken and written use (pp. 117–18):

/-jee/ . . . appear[s] in both spoken and written Mongol. It [is] particularly preva-lent in spoken storytelling, particularly when setting the scene or introducinga new event in the discourse.

/-laa/ . . . occurs in both spoken and written language, especially to convey a senseo immediacy.

/-v/ . . . appears in written, rather than spoken, language. . . .. . . [/-san/] as a finite verb affix is extremely prevalent in spoken language.

In conclusion, two o the affixes, /-jee/ and /-laa/, occur in both spoken and writ-ten language . . . . . . . /-v/ and [/-san/] . . . are largely confined in their distribution towritten and spoken Khalkh Mongol, respectively.

We see below, however, that some o their conclusions are not quite correct, and oth-ers do not tell the ull story.


36/259

13

and where the speaker is uttering the sentence in question.17 Whentalk is o an occurrence separated—and especially when distant—in

time rom the time o utterance, the - jee orm is generally usedinstead.

c) Te past tenses marked by - jee and -lee are essentially deictic andindenite (they simply indicate that some occurrence happenedsome time in the past, without necessarily relating it to any par-ticular contextual time), while those in -sen and -v are anaphoric (relating the occurrence recounted in their clause to a contextualtime) and denite (relating it to a particular time).

d) When, or some reason, neither the evidential, proximal -lee, northe inerential, distal - jee is appropriate to use, a neutral orm isused instead, -sen in the spoken language, -v in the written. At leastsome o the contexts triggering use o these latter tense markers arespecifically those in which definite, anaphoric tenses are appropriate.

e) Within a topical thread (a passage or string o passages sharinga common subject or theme), -lee may indicate a new subject ortheme, or may suggest urther inormation to come, while - jee maymark the conclusion o the thread.

) In texts and extended discourses, past tense use depends on genre.In narration, the tenses are anaphoric; -jee indicates an occurrencepreceding the reerence time (roughly, the current time defined bythe context), -lee one which ollows it, and the neutral -v (-sen)indicates one occurring at the reerence time, which allows its useto oreground narrative material (that is, to mark it as part o themain events o the narrative). In non-narration, the tenses are deic-tic and unction like metric tenses, that is, tenses marking degreeso remoteness18 rom the present time: - jee is a distant past, -lee a

recent past, and -v is neutral in this regard.

17 Nelson et al. (1998: 121) propose that -lee “does not mark past tense per se, butrather signals the relative proximity o the speaker with the situation being related,which may include the past or uture,” rom which they conclude that “[i]n mostcases, -/laa/ is a discontinuous tense that includes past or uture but excludes thepresent,” which they compare (p. 122.) to a discontinuous tense reported by Comrie(1985: 88–89) in an Australian aboriginal language, Burera. (Tey mistakenly datethis to 1976.) However, they later note (p. 125) that “[t]here are some types o predi-cates where a present tense interpretation or /-laa/ is possible, or more precisely,where a state is interpreted as continuing into the present,” so that (p. 126) “[i]n theseexamples, /-laa/ is not a discontinuous tense.” Te upshot would seem to be that, asproposed in the present work, -lee is not a discontinuous tense.

18 Tat is not to say that the tense systems o the Mongolic languages are remote-ness or metric systems o the kind typified by the Bantu languages.


37/259

14

g) Te meanings o the various endings and their unctions in dis-course are mutually determinant. Not every tense can be used in

every context, nor in every context can every temporal relationshipbe expressed. Meaning and use are not independent.

h) Te uses o the tenses in written language differ rom those inspoken language. Formal speech approximates the usages oundin writing, while inormal writing approximates that o speech.Speech and writing alike range over a variety o usages, rom themost colloquial and inormal to the most ormal and standard. Teevolving language o the Internet and non-traditional media exhib-its many o the usages o each o writing and speech, and in someregards, its usage is unique.

2. S

2.1. Teories Based on ense and Aspect

For almost a century and a hal, the correct analysis o the tense/aspectsystem o Mongolian largely eluded even the greatest scholars. Tis

is not at all surprising, given that their accounts o the semantics othe verb were nearly all based on a alse assumption, namely that thegrammatical “tenses” o the Mongolian verb marked by the affixes -jee,-lee, and -v are differentiated primarily by tense (time o the occur-rence or state relative to the present, or some other, time) and aspect(roughly, completeness o the occurrence). While such theories havebeen superseded by more adequate ones, such as the accounts pre-sented in Kullmann and serenpil (1996), Sanders and Ireedüi (1998),19 and serenchunt and Leuthy (2000), their prevalence in older works

still widely consulted today and perpetuation in some recent worksrequires their review, particularly as it serves to provide a backgroundto the development in the last two decades o more satisactory theories.

Te various accounts o the meanings and uses o the orms given inthe older, and much o the recent, literature are ofen vague and some-times contradictory. Scholars do not agree with each other, and some,such as Poppe, even differ with themselves. For example, Poppe’s 1951

19 Teir characterizations o the tenses are not as satisactory as those in the ser-enchunt/Leuthy textbook or the Kullmann/serenpil reerence work, but still mark anadvance on the accounts in the older textbooks.


38/259

15

grammar has it that the - jee ending marks a witnessed or certain pastoccurrence (which is how he characterizes the -lee ending in his 1970

Handbook), then his 1954 grammar calls - jee a pluperect marker,while the Handbook calls it an extended past. (Tis last characteriza-tion also is to be ound in some Russian and Mongolian works, orexample Kas’yanenko 1968 and Jančivdorj and Ragčaa 1967). At vari-ous times Poppe terms the -v ending an indicator o a completed pastoccurrence or a perect (1951), a marker o the recent past (1954), or just the “simple” past (1970).

But it is - jee which is called a recent past by Street (1963)—as is -lee by Ševernina (1958) and by Kas’yanenko (1968). Ševernina character-izes - jee as a distant past (as does Schlepp 1983). Te sense that - jee marks a distant past, even a pluperect (the time o the occurrence isearlier than a given past time), may be due to its marking a regress, aconcept applied by Dugarova (1991: 55), a “superficial” regress beingmarked by the orm -lee, a proound one by - jee. Dugarova writes:20

Формы -лаа и -жээ относят ситуации, происходящие на томэтапе, который достигнут в повествовании, назад, к болееранним этапам, и являются в каком-то смысле “дейктическими”

в пределах хронологии художественного “мира” повествования.Эти глагольные формы выражают подобие предшествования, хотя,согласно принятым в настоящей работе дефинициям, являютсятехнически не темпоральными. Функциями форм -лаа и жээ, такимобразом, являются обозначение “регресса”, если воспользоватьсятермином Э. Кошмидера [1962 387], возвращения “назад”, кпредшествующим событиям, обусловившим ситуацию, наличную кмоменту, до которого продвинулся рассказ. “Регресс” может быть“неглубоким” (возвращение к непосредственно предшествующемуэтому моменту событию) или, напротив, “глубоким” (возвращение

к более отдаленным событиям). “Неглубокий” регресс маркируетсяформой -лаа, глубокий—формой -жээ.

(Te orms -laa and -jee relate situations, which are occurring at thestage which has (already) been achieved in the narration, back to ear-lier stages, and are in some sense ‘deictic’ within the ramework o thechronology o the artistic ‘world’ o the narrative. Tese verbal ormsexpress something similar to precedence, although, in keeping withdefinitions accepted in the present work, are technically not temporal.Tus the orms -laa and -jee serve to denote ‘regression’, i one is to useE. Koschmider’s term [1962: 387], a turning ‘backwards’, to previous

events which have provided conditions or the situation that exists in

20 Extended quotations in translation here are mine, unless otherwise noted.


39/259

16

the present moment, and to which point the story has progressed. Te‘regression’ can be ‘shallow’ (a return to events immediately preceding

this moment) or, on the contrary, it can be ‘deep’ (a return to moreremote events). ‘Shallow’ regression is marked by the orm -laa, and‘deep’ regression by the orm -jee.)21

What Dugarova is suggesting is that these markers orm part othe background , rather than o the narrative foreground . Instead oadvancing narrative time, they may halt it or even send it backwards.An illustrative example rom English is (7a), where the explana-tory second sentence involves an event that precedes in time theevent recounted in the first sentence. Tis contrasts with a narrative

sequence, such as Caesar’s boast (7b), in which each sentence recountsa later event than the one beore. In such examples, all o the sentencesare equally on the main narrative line and orm part o the oregroundo a narrative.

7. a. Te itanic sank. It hit an iceberg. b. I came, I saw, I conquered.

Te present perect tense has sometimes been identified with a recentpast, and the term “present perect” has been applied both to - jee—Street (1963: 122) says it may emphasize “the present result o a pastaction or o a state that existed in the past (and may continue into theuture)”—and to -lee (by Ramstedt, and ollowing him, by Poppe 1951,1955, 1970 and Sanžeev 1964).

In light o this (brie and unsystematic) survey, Chuluu’salready-quoted conclusion (Ujeyediin 1998) that “attempts to establishclear and reliable criteria or distinguishing them are not conclusive”

seems an understatement.Several scholars indeed have suggested that there is no temporal orsemantic distinction between the endings, but their suggestions are onthe whole vague, generally lack adequate supporting evidence, and donot point toward definite alternative analyses.

Ramstedt, in his pioneering (1902) study, already warned against asimple account o the endings as “tenses,” though his warning provedinadequate to prevent subsequent authors rom taking too seriouslythe labels he assigned the endings. Although he (on p. 21) calls the

21 ranslation by Jan Schallert, thanks to Christina Kramer.


40/259

17

endings -ne (the “present-tense”, i.e. the non-past or present-uture),-lee, -jee, and -v “tempora indicativi” (“tenses o the indicative”), he

warns that “kann man auch hier den namen ‘tempora’ nur mit vorsichtgebrauchen” (“one can use the name ‘tempora’ only with care”); apartrom “certain objective temporal distinctions,” “[d]ieser khalkhassis-chen verbalormen bezeichnen nämlich in der that, ausser gewissenobjektiven zeitunterschieden, auch verschiedene zeit- und aktionsarten(action imperecta & actio perecta)” (“these Khalkha verb orms indi-cate in reality besides certain objective temporal distinctions also vari-ous kinds o time and aspect (actio imperecta & actio perecta”).22

A number o scholars have stated something similar to Poppe’sobservation (1954: 163, noted by Bläsing 1984: 37) that “there is notemporal difference between various present (or past) orms, but onlya difference o the point o view o the person speaking and the lat-ter’s subjective attitude.” Similarly, Ševernina (1958: 39) warns that the various orms are ofen used or one another and are not “definite.”23

Even Ramstedt himsel, despite his use o terminology based ontense and aspect, does not ully subscribe to the view that the Mongo-lian “tense” orms are simply markers o tense and aspect, but implic-

itly presents another, non-temporal, perspective. He notes that thepast tenses ofen have uses that he characterizes as modal . He writes(1902: 21), albeit somewhat obscurely, that

In vielen ällen tritt bei den tempora die aktionsart deutlicher als dierelative zeitstue hervor, und wir sollten dann von “modi”, nicht von“tempora” reden. So, wenn die -wɒ-bildung zu hypothetischen ann-hamen gebraucht wird. So auch die -lā-bildung, wenn durch sie etwasals sicher geschehend oder als demonstriert bezeichnet wird. Die-nɒ-bildung dagegen hat eine so allgemeine bedeutung, dass sie eher die

hundlung ganz tempuslos als au eine bestimmte zeit hinweisend aus-sagt. Die gegenseitigen verhältnisse der khalkhasischen tempusormenentsprechen also, trotz ihrer hier gegebenen namen, nicht denjenigender objektiven zeitstuen.

(In many cases aspect comes out more clearly in the tenses than relativetime, and we should then talk o ‘modi’, not o ‘tempora’. Tus when the[-v ]-orm is used or hypothetical suppositions. Tus too the [-lee]-orm,when through it something is indicated as clearly happening or as

22 Te translations here rom Ramstedt are mine. I wish to thank my colleague,Proessor Harald Ohlendor, or his help with the translations rom German given inthe present work, though ultimate responsibility or them remains mine.

23 o use Bläsing’s (1984: 37) term.


41/259

18

demonstrated. Te [-ne]-orm on the contrary has such a general mean-ing, that it expresses the action entirely atemporally (timelessly) rather

than as reerring to a definite time. Te reciprocal relationships o theKhalka tense orms thus do not correspond, despite the names giventhem here, to those o the objective times.)

Ramstedt suggests designating them not “tempora” (tenses), but rather“modi” (moods). But as Bläsing (1984: 37) comments, Ramstedt doesnot provide us with “an elucidation or precise definition” o suchterms as modus and actio. Nor is it entirely clear in the end whetherRamstedt views the endings as modal, aspectual, or temporal, or somecombination thereo.

But notwithstanding Ramstedt’s warnings, most scholars throughthe end o the 1980s took the distinctions between the past tense mark-ers to be semantic ones involving both tense and aspect . Ramstedt’sterminology (given in table 3 below) suggests a theory in which theour24 finite, indicative tenses o Mongolian are semantically definedby the temporal (i.e., tense) opposition o past versus non-past, andthe aspectual distinction o perective (i.e., complete) and imperect(incomplete).25

able 3

Ending Label

-ne Imperfektivisches präsens, Präsens Imperfekti‘imperective present’, ‘present o the imperect’

-v Perfektivisches präteritum, Präteritum perfekti ‘perective preterite’,‘preterite o the perect’

-lee Perfektivisches präsens, Präsens Perfekti‘perective present’, ‘present o the perect’

-jee Imperfektivisches präteritum, Präteritum Imperfekti‘imperective preterite’, ‘preterite o the imperect’

24 Some Mongolian languages show more than one present tense, however. Forexample, the vertical script language has a second present tense ending, -yü/-yu (Poppe1954: 92). For the orms and uses o the present tense endings in the Mongolian lan-guages, see Poppe (1955: 261ff.), Weiers (1969: 131–46), and Wu (1996: 58–68).

25 Although the terms perfect and imperfect are ofen used in this connection in theMongolistic literature, the aspectual distinction in question is actually that o perfec-tive and imperfective.


42/259

19

Tus, ollowing Ramstedt, the -v past, reerred to simply as the pasttense by Hangin (1968: 24) and as the denite past by Bosson (1964: 27),

is called by many scholars the past perfect or past of the perfect ;26 whilethe past in - jee, termed the narrative past by Hangin (1968: 114), is onthe contrary called the past imperfect or past of the imperfect , reflectinga purely—i perhaps only nominal—aspectual contrast between thetwo.27 Te -lee past, on the other hand, is considered by some scholarsnot to be a true past tense at all.28 Instead, they term it a present perfect ,which differs rom the past perect in tense (i.e., time) and rom thepresent imperect in aspect.29 Te contrasting present imperfect is orRamstedt and those ollowing him the orm in -ne;30 this marks thenon-past (present-uture) tense.31 Tese contrasts are summarized intable 4 below.

able 4

tense\aspect perect imperect

past -v -jeenon-past -lee -ne

26 Poppe, ollowing Ramstedt’s präteritum perfekti (1902: 19), has Präteritum perfecti (1951: 80), Praeteritum perfekti (1955: 266), and past of the perfect (1970: 131).Ramstedt also writes o the perfektivisches präteritum (p. 24). In Russian this is per-

fektnyi preterit (Sanžeev 1964: 193), hence Mongolian öngörön tögssön tsag , önggerentegüsegsen caγ, ‘past perect tense’ (as in Beffa and Hayamon 1975: 81).

27 Poppe, again ollowing Ramstedt (1902: 18), who has präteritum imperfekti, writeso the Präteritum imperfecti (1951: 80), Praeteritum imperfekti (1955: 265), and pastof the imperfect (1970: 131). Ramstedt also has imperfektivisches präteritum (p. 24).In Russian, this is imperfektnyj preterit (Sanžeev 1964: 190).

28 Strangely enough, some scholars who reer to it as the present perect list it as oneo the three past tenses. Tus Poppe (1954: 92), who consistently calls it the presentperect, includes it in the “three orms o the past.”

29 Poppe, ollowing Ramstedt’s präsens perfekti (1902: 17), has Präsens Perfecti(1951: 80), Praesens perfekti (1955: 265), and present of the perfect (1970: 130). Ram-stedt writes o the perfektivisches präsens (p. 24). In Russian this is perfektnyj prezens (Sanžeev 1964: 188). Oddly enough, Sanders and Bat-Ireedüi’s textbook (1999: 39)calls this orm the past perect tense.

30 Ramstedt calls this the imperfektivisches präsens (1902: 24) or präsens imperfekti(p. 15), ollowing which come Poppe’s Präsens imperfecti (1951: 79), Praesens Imper-

fecti (1955: 260) and present of the imperfect (1970: 130), and, in Russian imperfektnyj prezens (Sanžeev 1964: 185).

31 Similarly in Mongolian (Jančivdorj and Ragčaa 1967: 113) it is the odoo baireedüi tsagiin dagavar ‘present and uture tense suffix’. Beffa and Hayamon (1975:80) call it the effectif non passé (i.e., non-past ), in Mongolian edüge/odu-a čaγ ziγaqukhelber, odoo tsag zaax xelber ‘orm indicative o present time’.


43/259

20

As has already been suggested, Ramstedt explains these labels poorlyand indeed warns the reader (p. 22) against taking them too seriously:

Die gegenseitige verhältnisse der khalkhassischen tempusormen ent-sprechen also, trotz ihrer hier gegebenen namen, nicht denjenigen derobjektiven zeitstuen. Die bedeutung und verwendung jeder einzelnenorm gründet sich, unabhängig von denen der anderen, nur au ihrehistorische unterlage, und man hat nur mit einer gewissen reiheit desredenden zu rechnen unter den vorhandenen ausdrucksmitteln die ürseine zwecke passendsten zu wählen.

(Te reciprocal relationships o the Khalka tense orms do not corre-spond . . . to those o the objective times, despite the names given to them

here. Te meaning and use o each particular orm is based, independento those o the others, only on its historical base, and one has but toreckon with a certain reedom o the speaker to choose amongst theavailable means o expression that serves his purpose.)

Nonetheless there has been a tendency by later scholars to elaborateon, and to attempt to justiy, Ramstedt’s labels in terms o varioussemantic distinctions.

2.2. Te Finite Indicative Verbs

Te name “past perect” suggests that the past tense in -v “reers topast and perected action” as Hangin (1968: 24) puts it, that is, to anevent completed in the past, as in (8, 9).

8. Önöö öglöö bi zurgaan tsag.t boso.v. this morning I six hour- get up- ‘Tis morning I got up at six o’clock.’ (Street 1963: 122)

9. Či öčigdör nom unši.v uu?

you yesterday book read- ‘Hast du gestern ein Buch gelesen?’ (Vietze 1974: 44) ‘Did you read a book yesterday?’

Te term “past imperect,” in parallel with the label “past perect,”should suggest an event not completed in the past. But in act incom-pletion as such does not usually enter into characterizations o thisorm, which is more properly the “present o the perect,” to translateRamstedt’s term.

Rather than the incompleteness o the action, some writers empha-size prolongation in time, so that “past imperect” means little more


44/259

21

than “extended past.”32 Poppe (1970: 131), or example, says that thepast tense verb ormed with - jee “denotes an action that took place in

the past and lasted or some time,” and he offers the examples num-bered (10, 11) below:33

10. Ürgelj ündes tüü.j. all the time roots collect- ‘all the time he collected [edible] roots’

11. iim sain xüü bai.jee. such good boy be- ‘there was such a good lad’

Beffa and Hayamon (1975: 82) claim that the action may extend intothe present, saying this orm indicates “an action commenced in thepast, and which is prolonged or repeated in the present.” Tey offerthe examples (12, 13):

12. Mongol.iin büx nutg.iin dundaj öndör Mongolia- whole country- middle height

1550 metr a.j. 1550 metres be- ‘l’altitude moyenne de toute la Mongolie est de 1550 mètres’ (‘the mean altitude o all o Mongolia is 1550 metres’)

13. xoër baildagč naadam.d barilda.x two combatants Naadam- wrestle-34

ge.j zodog šuudag.tai ir.jee. say- wrestler’s-jacket wrestler’s pants- come-

‘les deux combattants sont arrivés avec leur costume de lutter aux jeux(au Naadam)’ (‘the two combatants have come with their fighting suitsto the games [to the Naadam]’)

32 Beffa and Hayamon (1975: 82) call this the passé prolongé ou aoriste ‘prolongedpast or aorist’, in Mongolian, öngörön ürgel ǰ ilsen tsag , önggeren ürgülzilegsen caγ ‘pass-ing extended time’. Similarly Kas’yanenko (1968: 20) says this expresses “prolongedpast time” ( prošedšee dlitel’noe vremya).

33 Te translations are Poppe’s.34

Te orm sometimes called the “infinitive” is the so-called uture participle or verbal noun, the “nomen uturi.” Besides an infinitive, it also unctions as a present-uture or non-past orm.


45/259

22

As a matter o act, the - jee past, supposedly imperective in aspect, can serve to mark completed past actions (14–16):

14. xaa oči.j? where go- ‘kuda pošel?’ (Kas’yanenko 1968: 20) ‘where did s/he go?’

15. nar gar.č sun come out- ‘the sun rose’ (Poppe 1970: 131)

16. Dašdorjiin Natsagdorj 1906 on.d tör.jee. (Yatskovskaya 1976: 8)

Dashdorjiin Natsagdorj 1906 year- be born- ‘Dashdorjiin Natsagdorj was born in 1906.’

Although Ramstedt (1902: 16) calls the -lee past the “present o theperect” ( präsens perfekti), he does not justiy this name. His statements(p. 17) about the orm, indeed, are a bit conusing, or though he callsit “the present o the perect,” he observes that it is a kind o preterite(past tense), though it can also serve as a present or a uture:

Diese orm wird angewendet in sätzen, wo eine handlung als sicher abge-schlossen oder ein zustand als sicher erreicht angegeben wird. Sie giebtnicht nur eine bestimmte zeitstue, die gegenwart des vollendeten, son-dern auch die sicherheit des ausgesagten an. Es wird dadurch eine hand-lung au etwas, was ür die richtigkeit der aussage irgendwie eine stützegiebt, miteinbegrien. Der redende kann sich entweder au die persönli-chen erinnerungen des ausgeredeten (‘es war ja’ od. ‘wie bekannt’) oderau die zeitlichen verhältnisse beruen (‘schon, jetzt’). So angewendetentspricht diese orm unseren präteritis (imp., per. und plusqup.).

Es kann aber die rightigkeit der aussage auch aus der äusseren situa-tion im momente des sprechens (‘sieh da’, ‘sich doch’) oder als logischekonsequenz aus dem vorher gesagten hervorgehen (‘es wird ja’, ‘es ver-steht sich’). In solchen ällen müssen wir bei dem übersetzten das prä-sens oder uturum gebrauchen.

(Tis orm is used in sentences, where an action is represented as firmlyconcluded or a state as firmly achieved. It indicates not only a certaintense, the present o the perect, but also the firmness o the statement. Itincludes thereby a reerence to an indication o something that somehowsupports the correctness o the statement. Te speaker can reer either tothe personal recollections o the addressee (‘it was, as you know’ or ‘as

is well known’) or to the temporal relationships (‘already, at present’).Used thus, this orm corresponds to our preterite (imperect, perect,and pluperect).


46/259

23

Te correctness o the statement can also arise however out o theexternal situation at the moment o speech (‘look at that!’, ‘do look!’) or

as a logical consequence o what has been said beore. In such cases wemust use the present or uture in the translation.)

Accordingly he glosses example (17) below thus: “‘(er) stirbt schon’,od. ‘ist ja schon gestorben’,” i.e., “‘he is already dying’ or ‘you know,he’s already dead’.” Poppe (1951: 80) glosses it as ‘ist ja tot’, i.e., ‘isalready dead, don’t you know’.35

17. Üxe.lee. die-36

Others,37 however, have viewed the orm, in contrast to Ramstedt’s“present perect” as an imperfective, i.e., an uncompleted, past, indicat-ing “an action which has been started but is still unfinished.”38 Tusin (18), “that person has already started to come here, but has not yetarrived” (Wu 1995: 95).

18. tere kümün ire.le. that person come- ‘Tat person is coming.’

Chenggeltei (1981: 298, in Wu 1995: 97) sees the orm as indicative oa bounding point, rather than o a period o time:39

(a) the moment when an action is about to start; (b) the moment whenan action just started; (c) the moment when an action is about to finish;(d) the moment when an action is just finished or has already finished.

35 By using ja, Ramstedt and Poppe are trying in their glosses to capture a certainnuance o the -luγa (= -lee) orm, namely a sense that the act is well-known to, orshould be recalled by, the addressee. You know and don’t you know are my attemptsto

Documents

The Past Tenses of the Mongolian Verb. Meaning and Use