Content uploaded by Mark W Post
Author content
All content in this area was uploaded by Mark W Post on Oct 21, 2014
Content may be subject to copyright.
Foreword i
North East Indian Linguistics
Volume 3
Edited by
Gwendolyn Hyslop • Stephen Morey • Mark W. Post
Delhi • Bangalore • Mumbai • Kolkata • Chennai • Hyderabad
ii North East Indian Linguistics Volume 3
Published by
Cambridge University Press India Pvt. Ltd.
under the imprint of Foundation Books
Cambridge House, 4381/4 Ansari Road, Daryaganj, New Delhi 110002
Cambridge University Press India Pvt. Ltd.
C-22, C-Block, Brigade M.M., K.R. Road, Jayanagar, Bangalore 560 070
Plot No. 80, Service Industries, Shirvane, Sector-1, Nerul,
Navi Mumbai 400 706
10 Raja Subodh Mullick Square, 2nd Floor, Kolkata 700 013
21/1 (New No. 49), 1st Floor, Model School Road, Thousand Lights,
Chennai 600 006
House No. 3-5-874/6/4, (Near Apollo Hospital), Hyderguda,
Hyderabad 500 029
Agarwal Pride, ‘A’ Wing, 1308 Kasba Peth, Near Surya Hospital,
Pune 411 011
© Cambridge University Press India Pvt. Ltd.
First Published 2011
ISBN 978-81-7596-???-?
All rights reserved. No reproduction of any part may take place without the
written permission of Cambridge University Press India Pvt. Ltd., subject to
statutory exception and to the provision of relevant collective licensing
agreements.
Cambridge University Press India Pvt. Ltd. has no responsibility for the
persistence or accuracy of URLs for external or third-party internet websites
referred to in this book, and does not guarantee that any content on such
websites is, or will remain, accurate or appropriate.
Typeset at Sanchauli Image Composers, New Delhi.
Published by Manas Saikia for Cambridge University Press India Pvt. Ltd.
and printed at …
Foreword iii
Contents
About the Contributors v
Foreword ix
A Note from the Editors xvii
The View from Manipur
1. Person-Marking Prefixes of Purum
Surmangol Sharma and Gopendro Singh 1
2. The Evolution and Recent Development of
the Meitei Mayek Script
Harimohon Thounaojam Singh 24
The Sal Group
3. Three Meanings of “Language” and “Dialect” in
North East India
Robbins Burling 35
4. An Initial Reconstruction of the Proto-Bodo-Garo
Noun Phrase
Daniel Wood 46
5. Nocte and Jingphaw: Morphological Correspondences
Scott DeLancey 61
iv North East Indian Linguistics Volume 3
6. Tangsa Agreement Markers
Stephen Morey 76
Tibeto-Burman Nominalization
7. Nominalization and Related Phenomena in Marma
Huziwara Keisuke 105
8. Functions of Nominalization in Karbi
Linda Konnerth 120
Tani
9. Topographical Deixis and the Tani Languages of
North East India
Mark W. Post 137
10. Morphosyntactic Variation in the Pagro
and Sa:jaŋ
Dialects of the Mising Community
Jugendra Pegu 155
Eastern Indo-Aryan
11. The Referring Systems and the Determinative Elements of
Noun Phrases in Assamese
Jagat Kalita 173
12. Copular Sentences in Asamiya
Runima Chowdhary 198
13. Case Marking in Hajong
Virginia Crowell Phillips 224
Austroasiatic
14. Derivational Morphology in Pnar
Hemanga Dutta 241
Foreword v
About the Contributors
Editors
Gwendolyn HYSLOP (glow@uoregon.edu) is a PhD candidate at the
University of Oregon, USA. She is a specialist in the East Bodish
languages of Bhutan and Arunachal Pradesh and has also authored and
coauthored several publications on other Tibeto-Burman languages. She
has been a Director of the Aienla Project, a 501(c)3 non-profit organization
dedication to the preservation of languages and cultures of North East
India, since 2004.
Stephen M
OREY (s.morey@latrobe.edu.au) is a research fellow at the
Research Centre for Linguistic Typology, La Trobe University. He is the
author of two books on tribal languages in Assam, from both Tai-Kadai
and Tibeto-Burman families. He is the co-chair of the North East Indian
Linguistics Society and has also written on the Aboriginal Languages of
Victoria, Australia.
Mark W. P
OST (mark.post@jcu.edu.au) is currently Postdoctoral Research
Fellow in Anthropological Linguistics at The Cairns Institute of James
Cook University in Cairns, Australia. He is a specialist in the Tani
languages of Arunachal Pradesh, and has also published in the typology
and diachrony of other South East Asian languages. He has been Secretary
of the North East Indian Linguistics Society since its founding in
2006.
vi North East Indian Linguistics Volume 3
Authors
Robbins BURLING (rburling@umich.edu) first visited North East India in
1954–56 when he studied the kinship organization of the Garos. He has
written on the ethnology and linguistics of North East India and adjacent
Burma and Bangladesh, especially on the Garos. He is Professor Emeritus
of Anthropology and Linguistics at the University of Michigan, USA.
Runima C
HOWDHARY (runimachowdhary@yahoo.com) is Associate
Professor in Linguistics at Gauhati University in Assam, India. She has
authored a monograph on Assamese verbs and a number of research
articles on various aspects of Assamese linguistics. She is a life member
of the Linguistic Society of Assam and the Linguistic Society of India.
Scott D
ELANCEY (delancey@uoregon.edu) is Professor of Linguistics at
the University of Oregon, where he has taught since 1982. His research
concentrates on Tibeto-Burman languages and languages of western North
America, and on functional syntax, particularly in the areas of alignment,
case theory, and grammaticalization.
Hemanga D
UTTA (hemangadutta1@gmail.com) is a research scholar at
Jawaharlal Nehru University, New Delhi. He has been awarded the S.
M. Katre award in 2006, as well as a short term PhD fellowship by
Utrecht University in 2009.
HUZIWARA Keisuke (kejcxan@gmail.com) is Research Fellow of linguistics
at Kyoto University, Kyoto, Japan. He has been studying Tibeto-Burman
languages in the Chittagong Hill Tracts, Bangladesh, since 1999 and
completed his PhD thesis on Cak language in 2008.
Jagat Chandra K
ALITA (jagat_kalita@sify.com) is Assistant Professor of
Assamese at Abhayapuri College in Abhayapuri, Assam, India.
Linda K
ONNERTH (lkonnert@uoregon.edu) is a PhD student of Linguistics
at the University of Oregon in Eugene, OR, USA. She is currently
working on a descriptive grammar of Karbi.
Arup Kumar N
ATH (arupjnu@gmail.com) is a PhD scholar of the Center
for Linguistics at Jawaharlal Nehru University, New Delhi. At present
Foreword vii
he is working on Tiwa and Deori, two Tibeto-Burman languages of
Assam, and is also involved in the multilingual online descriptive
dictionary project of Assamese and other North Eastern languages (http://
www.xobdo.org/).
Jugendra P
EGU (jugen@tezu.ernet.in) is a Lecturer at the Department of
English at Jonai Girls’ College, Assam, and is currently pursuing his
PhD in Linguistics on Mising language at the Department of English and
Foreign Languages, Tezpur University, Assam. He is also engaged in
the North East Language Development Programme (NELD) funded by
the Central Institute of Indian Languages (CIIL), Mysore, Karnataka,
India.
Virginia Crowell P
HILLIPS (ginny_phillips@sall.com) works with SIL
International as a linguist and literacy specialist. She was born and raised
in Punjab, India by American parents and now facilitates language
development for Eastern Indo-Aryan languages in West Bengal, Bihar
and North East India. She has an MA in Linguistics from the University
of North Dakota.
H. Surmangol SHARMA (surmangol@yahoo.co.in) is Reader in Linguistics
at Deccan College Postgraduate and Research Institute, Pune, India. He
is Editor of the Manipuri-English Learner’s Dictionary.
Chungkham Yashawanta S
INGH (chungkhamyash@gmail.com) is
Professor of Linguistics, specializing in Tibeto-Burman Linguistics at
Manipur University, Manipur, India. He is author of two books: Manipuri
Grammar and Tarao Grammar. He was a Commonwealth Fellow and
Fulbright Fellow.
Nameirakpam Gopendro S
INGH (gopendron@yahoo.com) has a PhD in
Linguistics and is presently working as Research Assistant at the
Language Cell, Directorate of Education (School), Government of
Manipur, India. His current research is focused on Purum grammar, as
well as compilation of a Manipuri-English dictionary.
Thounaojam Harimohon S
INGH (harimohon@yahoo.co.in) has a PhD in
Linguistics and has worked as a Lecturer for seven years. Currently, he
is Research Assistant in the Language Cell of the Directorate of Education
About the Contributors vii
viii North East Indian Linguistics Volume 3
(School), Government of Manipur, India. He has conducted research on
several endangered languages of Manipur such as Lamkang, Purum,
Thadou, Paite, etc., as well as some articles on Meetei. He is currently
co-compiling a Manipuri to English Dictionary.
Foreword ix
Foreword
Chungkham Yashawanta Singh
Manipur University
1. Linguistic diversity in India
It is a fact that India is one of the most linguistically diverse countries
in the world, with seven language families; namely: Indo-Aryan (of
Indo-European), Dravidian, Austroasiatic, Tibeto-Burman (of Sino-
Tibetan), Tai (of Tai-Kadai), Andamanese and Great Andamanese which
is newly added by Abbi (2006). The North East can be described as a
mini-India because there are so many different communities having
various cultures and speaking various mother tongues. For example, from
the Dravidian language family, particularly Tamil but also Malayalam
from Kerala are spoken by inhabitants in Moreh in Manipur, a small
Indo-Myanmar border trade town 109 km from Imphal. They were driven
out from Mandalay in 1960 by the Burmese Army. The North East is a
garden of various languages and dialects with more than one hundred
various mother tongues, especially Tibeto-Burman (TB) languages – most
of which are neither fully documented nor scientifically described.
Manipur has officially recognised 33 tribal languages. Besides these,
Manipuri (also called Meitei) itself has six dialects, namely, Sekmai,
Khurkhul, Andro, Phayeng, Koutruk and Kwatha. All these tribal
languages are not yet fully and scientifically described; of these some
are endangered, namely, Tarao, Monshang, Moyon, Koireng, etc.
Documentation of these languages is, of course, felt to be very necessary;
loss of a language goes with the loss of beliefs, culture, ecology and
medicinal herbal knowledge. In Manipur there was a language called
x North East Indian Linguistics Volume 3
Chakpa, spoken in Andro and Phayeng; now it was extinct. Documentation
was not properly done.
2. Linguistic situation
There are some extremely complex linguistic situations in the North East.
One is in the Churachandpur district of Manipur where the young Hmars
understand Paite, Mizo, Vaiphei, and Thadou while the old folks
understand three more speech forms, namely Simte, Zou and Gangte.
These speech forms are mutually intelligible to some extent and share a
great number of common features in phonology, morphology and syntax.
It remains to be demonstrated which of these should be treated as separate
languages, and which as dialects of the same language. The situation is
complicated by other facts. Whereas the Hmars of Churachandpur
understand Paite, the Thadous of Sajik Tampak in Chandel district, know
Paite only a little.
Opposite to this is Ukhrul district where there are more than 200
villages, each with a distinct speech form. They are mutually unintelligible.
When they talk to each other (e.g. between Hundung and Toloy villages)
they automatically use either the Ukhrul dialect, the standard variety of
the headquarters of the district, or Manipuri as their lingua franca
(Singh 1995).
Unlike the linguistic situations of the above two districts, there is a
peculiar linguistic situation that may be found in the small border town
of Moreh, on the Indo-Myanmar border. There, three language families
co-exist, viz. Indo-Aryan (Hindi and Punjabi), Dravidian (Malayam and
Tamil) and TB languages. Many ethnic groups speaking different speech
forms are concentrated in this small town. They are Hmar, Thadou, Paite,
Gangte, Tiddim-Chin, Tezang, Mizo and Meitei. These various groups
use Manipuri as their lingua-franca.
The influence of the Manipuri language is great. Myanmarese traders
also speak Manipuri and the various ethnic groups inhabiting this small
town have borrowed many day-to-day words both from Manipuri as well
as Burmese. Since the town is a business centre, interaction – especially
between the Manipuri and Myanmarese – is via a trade language. It is
conceivable that a new pidgin may emerge ultimately from the three
predominant languages of the area – Hindi, Manipuri and Burmese
(Singh 1995). It is also noticed that there are many Manipuri-Burmese
Foreword xi
bilinguals along the Myanmar border, as well many Tamil-Burmese
bilinguals.
It is also a fact that in every district of Manipur, communities live
together and are in constant contact with each other. From such contact
there may arise new speech forms, as is the case in, for example, the
border areas of Brazil, Uruguay and Argentina. On the Iberian peninsula,
as is well known, there is a geographical dialect continuum where dialects
of Catalan, Spanish and Portuguese merge gradually into one another,
and where the number of languages recognized as being spoken depends
on the autonomous, standard varieties that have succeeded in raising
themselves over the dialect continuum (see Kurath 1972). Similar
dynamics may be found in our region.
The opening of Department of Manipuri Language and Literature in
Jawaharlal Nehru Post Graduate Centre, Imphal in 1972, has made a
turning point in the scientific study of language in Manipur. Some
unpublished PhD theses have come up, viz., M. S. Ningomba’s Maring
Grammar (1976), P. C. Thoudam’s A Grammatical Sketch of Meiteilon
(1980), P. Madhubala’s Manipuri Grammar (1979), Ch. Y. Singh’s Some
Aspects of Manipuri Syntax (1984), Nilamani Singh’s An Outline of
Meiteilon Interrogative (1985), U. Chetan’s Structural Analysis of
Manipuri Language (1976), and Nonigopal Singh’s A Meitei Grammar
of Roots and Affixes (1987). Scientific studies on various aspects of a
variety of languages of Manipur have grown in significance since 1986,
when the Department of Linguistics was established in Manipur
University. Studies on various aspects and on various languages have
started, namely, on the Manipuri language and others, for instance, Aimol,
Kom, Hmar, Tangkhul, Lamkang, Chothe, Zou, Rongmei, Chiru, Sizang,
Sukte, etc. These studies are on the descriptive front.
The second and third fronts of linguistic research are on Sociolinguistics
and Lexicography. Many MPhil. dissertations and many PhD theses on
various aspects are made; MA field reports on various tribal languages
have been completed since 1986.
Studies on TB languages of Manipur by the Department are
remarkable. Some of the projects worth mentioning are as follows.
Dr N. Pramodini is undergoing “Development of NE Languages:
Manipuri, Nurturing NE Languages on Digital Medium”; Dr S. Imoba’s
“Dictionary Project on Manipuri to English (2002–03)”; “Dictionary
Project on Manipuri-Hindi-English (2006–09)” and another project on
“Survey of Surnames in Manipuri” (ongoing). Prof. Madhubala is
xii North East Indian Linguistics Volume 3
investigating a project on “Manipuri Verbal Lexicon” (2006–09). Prof.
Yashawanta Singh had completed two projects: “Studies of endangered
tribal languages of Manipur” and “Studies of tribal languages of
Manipur”, which has now stepped into the Northeast WordNet
Consortium. Significant works have come up from many scholars, namely
Thoudam’s Remedial Manipuri Grammar (1988), Ningomba’s Meitei
Lonmit (1981), D. N. S. Bhatt and Ningomba’s Manipuri Grammar
(1997); Ch. Y. Singh’s Manipuri Grammar (2000) and Tarao Grammar
(2002); N.S. Singh’s A Grammar of Paite (2006); Madhubala’s Manipuri
Phonology (2002), Imoba’s Manipuri to English Dictionary, (2004);
Surmangol Sharma’s Learner’s Manipuri-English Dictionary (2006),
Hidam Dolen’s Structure of Manipuri Meaning (2004), and H. Singh’s
Chothe Grammar (2008). Manipuri has also been studied at the University
of Northern Texas by Shobhana Chelliah, producing A Grammar of
Meithei (Chelliah 1997). Chelliah and Harimohon Thounaojam Singh
then wrote Lamkang Grammar which was published in Linguistics of
Tibeto-Burman Area (Chelliah and Singh 2007).
As yet, no joint venture with any foreign university has been made.
However, the Central Institute of Indian languages, Mysore (CIIL) has
taken a magnificient step in the study of the TB languages of the Northeast
including the languages of Manipur. CIIL has taken up a grand project/
programme called North-East Language Development (NELD) since
2003. Under this programme some languages have started to be described,
for instance, Kom, Thangal, Chothe, Aimol, Hmar, Sukte, Mao, and
Tangkhul, in Manipur; other languages of the Northeast under this
programme are Adi, Nocte, Ao, Dimasa, Tiwar, Mishing, Riang, Karbi,
Pnar, and Galo, etc.
Other Indian scholars working on languages of the North East include
U. V. Joseph, whose mammoth grammar of Rabha has now been
published (Joseph 2007), Shobha Satyanath, who has worked on
Bishnupuriya, and A. K. Mishra.
The contacts built between scholars from Manipur University and
those from outside India have been strengthened by the NEILS
conferences and volumes such as this, giving Manipuri linguists a great
opportunity to share their knowledge with overseas scholars and interact
with them. Languages of the North East, such as Tai Phake and Aiton,
Boro, Garo, Meitei, Mongsen Ao, Deuri, Karbi, Khasi, Deuri, Kok Borok,
Tai Ahom, Singpho, Galo and Atong have been the source of substantial
language descriptions by scholars like Banchob, Burling, Chelliah, Coupe,
Foreword xiii
Grüssner, Jacquesson, Morey, Post and van Breugel, from Thailand,
USA, Australia, Germany, France and the Netherlands. In addition,
scholars like Jackson Sun from Taiwan and Alfons Weidert from
Germany added very much to our knowledge of historical and comparative
linguistics in the North East.
3. Points of relationship
The TB languages of the Northeast have some common lexical items,
especially in body parts, numerals, and heavenly bodies. For example,
the words mi[k]/[t] ‘eye’, sam ‘hair’, lu (kok) ‘head’, panga/nga ‘five’,
etc. are common amongst the TB languages of North East India. Not
only this, the occurrence of a velar nasal [ŋ] is found in all the languages
of Manipur as well as in the TB languages of North East India, and also
in Burmese and Lai (Hakha and Falam) for instance, ngaa ‘fish’.
Morphologically there are usually only two major categories, namely,
noun and verb; suffixes are more prevalent than prefixes. Syntactically,
functional types of sentences are formed by the addition of the respective
markers to the verb.
Manipuri has some peculiar linguistic features, such as semantic
reduplication; for instance, paamba kei ‘tiger’ (paamba ‘tiger’ is the
word for tiger found at the time of the Khuman principality (14 century)
and kei is ‘tiger’ in the speech of the Ningthouja clan of the Meitei.
Another feature is of polite forms, or honorifics: for example, a polite
form of the second person pronoun nang ‘you’ is asom/adom; and, polite
sentences can be formed by addition of -bi/-pi to the verb as well as by
use of a lexeme canbiduna/ thoujanbiduna ‘kindly’. Expressions can be
made even more polite by use of an archaic alternant such as cep-pi-ro
(instead of the more common tum-bi-ro ‘please sleep’ (cep = tum
‘sleep’).
4. Conclusion
Many things remain to be done; the scope of study of TB languages of
North East India is very vast. A scientific and systematic study of the
TB languages of North East India may lead to new challenges to our
current linguistic theories. The study of the more than 100 TB languages
of this region may weaken or may strengthen Greenberg’s Universals in
xiv North East Indian Linguistics Volume 3
the field of phonology, morphology and syntax, for example (Singh 2000).
Detailed research can be conducted on the following points:
a) the descriptive study of TB languages,
b) sociolinguistic study of each ethnic group,
c) the historical and comparative study of grammars, and
d) re-examination of sub-grouping of TB languages.
This book and the papers herein take modest but important steps in
these directions. We all look forward to more.
References
Abbi, A. (2006). Endangered Languages of the Andaman Islands. München,
Lincom Europa.
Bhatt, D. N. S. and M. S. Ningomba. (1997). Manipuri Grammar. München,
Lincom Europa.
Chelliah, S. (1997). A Grammar of Meithei (Mouton Grammar Library 17).
Berlin, Mouton de Gruyter.
Chelliah, S. and H. Th. Singh. (2007). “The Lamkang language: Grammatical
sketch, texts and lexicon.” Linguistics of the Tibeto-Burman Area 30(1):1–
212.
Chetan, U. (1976). Structural Analysis of Manipuri Language. PhD Thesis,
Manipur University.
Dolen, H. (2004). Structure of Manipuri (Meiteiron) Meaning. Imphal,
MALADES International.
Greenberg, J. H. (1963). “Some Universals of Grammar with Particular Reference
to the Order of Meaningful Elements.” In Greenberg, Ed. Universals of
Language. Cambridge, MA, MIT Press: 73–113.
Imoba, S. (2004). Manipuri to English Dictionary. Imphal, Manipur, S. Ibetombi
Devi.
Joseph, U. V. (2007). Rabha. Leiden, Brill.
Kurath, H. (1972). Studies in Area Linguistics. Bloomington, Indiana University
Press.
Madhubala, P. (1979). Manipuri Grammar. PhD Thesis, Manipur University.
_____. (2002). Manipuri Phonology. Imphal, Manipur, Potsangbam Bhuban
Singh.
Ningomba, M. S. (1976). Maring Grammar. PhD Thesis, Manipur University.
_____. (1981). Meiteilonmit [Meitei Language]. Imphal, Manipur, Imphal Board
of Education.
Sharma, S. (2006). Learner’s Manipuri-English Dictionary. Imphal, Manipur,
Sangam Book House.
Foreword xv
Singh, B. H. (2008). Chothe Grammar. New Delhi, Akansha.
Singh, C. Y. (2002). Tarao Grammar. New Delhi, Akansha.
_____. (2001). Manipuri Grammar. New Delhi, Rajesh.
_____. (2000). “The Tibeto-Burman languages of North Eastern India.” In
Manipur University Research Journal of Humanities and Social Sciences.
Vol. V, No. 1. Imphal, Manipur University.
_____. (1995). “The linguistic situation in Manipur.” Lingusitics of the Tibeto-
Burman Area 18.1: 129–34.
_____. (1984). Some Aspects of Manipuri Syntax. PhD Thesis, Manipur
University.
Singh, Nilamani. (1985). An Outline of Meiteilon Interrogative. PhD Thesis,
Manipur University.
Singh, Nonigopal. (1987). A Meitei Grammar of Roots and Affixes. PhD Thesis,
Manipur University.
Singh, N. S. (2006). Grammar of Paite. New Delhi, Mittal.
Thoudam, P. C. (1980). A Grammatical Sketch of Meiteilon. PhD Thesis,
Manipur University.
_____. (1988). Remedial Manipuri Grammar. Imphal, Land.
xvi North East Indian Linguistics Volume 3
xvi blank
Foreword xvii
A Note from the Editors
The papers collected for this volume were initially presented at the Third
International Conference of the North East Indian Linguistics Society
(NEILS), held on January 18–22, 2008 at the Don Bosco Institute in
Guwahati, Assam. The conference was ably hosted by the Department
of Linguistics of Gauhati University, in collaboration with scholars from
the Research Centre for Linguistic Typology of La Trobe University
(Melbourne, AU) and from the Department of Linguistics of the
University of Oregon (USA). As the third such volume in the North East
Indian Linguistics series, this collection offers plentiful evidence that
North East India is not only one of the richest and most diverse cultural-
linguistic areas in all Asia, but is quickly becoming one of the richest
and most diverse research areas in the field of descriptive linguistics as
well.
We are especially pleased to note that the present volume both
continues the existing trend of diversity – papers are included here from
well-known scholars based in Australia, in several states of the USA, in
Japan, and throughout North East India – and extends this by introducing
several committed young researchers who are now making their first
enduring marks on the field. All papers appearing here, as in previous
volumes of the series, were anonymously peer-reviewed by leading
international specialists in the paper subfields, underwent a rigorous
revision process in close consultation with the editors, and were finally
subjected to approval by the editorial staff at Cambridge University Press
India Pvt. Ltd. Our aim, here as in previous volumes, has been to bridge
the longstanding gaps between local and international researchers and
readerships, and to produce a volume of the highest quality obtainable
while reflecting the current state of research in our field.
xviii North East Indian Linguistics Volume 3
An important focus of the third NEILS conference was on the
languages of Manipur. One of the tiniest of Indian states, Manipur is
also among the richest and yet least well-understood cultural-linguistic
areas in Asia. Home to an untold number of mainly Kuki-Chin languages
– most of them practically unknown to the outside linguistic world –
Manipur is also home to a dedicated and active network of indigenous
linguists working on such fascinating and little-known languages as
Aimol, Anal, Inpui, Koireng, Kom, Chothe, Moyon, Monshang, Sukte,
and of course, Meithei (Manipuri) the language spoken by a majority of
the people of Manipur. We are very pleased to be able to bring some of
the work from this group of scholars to the attention of a wider
international audience.
Three papers from this group were finished quite early and were thus
included in the second NEILS volume (Aimol 2010; Betholia 2010; Singh
2010). The present volume begins with a foreword by Yashawanta Singh,
senior most and one of the most active of Manipuri linguists, and is
followed by a special section dedicated to the Languages of Manipur.
Surmangol Sharma and Gopendro Singh open this section with a paper
on pronominal proclitics in the little-known Kuki-Chin language Purum,
also comparing them with similar forms in Manipuri. Thounaojam
Harimohon, another indigenous linguist of Manipur, then presents a rare
and fascinating insight into the evolution and development of Meitei
script, a Brahmic-type script whose story provides a glimpse into aspects
of the historical diffusion of South Asian orthographies across South
East Asia.
Branching away from Manipur, the next section focuses on the “Sal”
group of languages, a hypothesized genetic meso-grouping within Tibeto-
Burman of Bodo-Garo, “Konyak Naga” and Jinghpaw first proposed by
Robbins Burling (1983). Burling himself first sets the scene with a
background discussion of three meanings of the terms “language” and
“dialect” in North East India. Speaking from decades of experience
of interacting with linguists and non-linguists in the North East who often
use these terms with different goals and meanings, Burling helpfully
provides an analysis of each usage in an attempt to increase mutual
understanding. Dan Wood follows with a preliminary reconstruction of
the Bodo-Garo noun phrase, one of several recent advances in
comparative-historical Bodo-Garo studies emerging from the University
of Oregon, others of which will be included in the next NEILS volume.
The next paper by Scott DeLancey, also of the University of Oregon,
Foreword xix
focuses on one corner of the “triangle” of the Sal hypothesis. Arguing
that morphological correspondences among Nocte and Jinghpaw in the
tense-agreement complex provide strong evidence of a relationship at
the Proto-Sal level, DeLancey speculates that the absence of such
structures in Proto-Bodo-Garo may owe to partial creolization of the
latter (a hypothesis also anticipated by Burling 2007). In the same
geographical, genetic and typological area, Stephen Morey rounds off
the Sal group papers with a discussion of a very complex system of
portmanteau agreement markers in Tangsa varieties, some of which here
receive their first ever descriptive treatment.
Turning to the perennially fascinating topic of nominalization and
nominalization-based constructions in Tibeto-Burman, two papers add
important new genetic and typological dimensions to the discussion.
While most past studies have focused either on relatively isolating
languages like Lahu (Matisoff 1972) or strongly agglutinating languages
of the Eastern Himalaya like Athpare and Belhare (Bickel 1999), Keisuke
Huziwara here presents a historically well-backgrounded discussion of
nominalization in Marma, an almost completely undescribed dialect of
Arakanese spoken in the Chittagong Hill Tracts. Linda Konnerth follows
with an analysis of nominalization in Karbi (Mikir), a relatively more
accessible but still severely undescribed Tibeto-Burman language of
Assam.
Two papers present new studies of aspects of the Tani languages of
Arunachal Pradesh, Tibet and Assam. In the context of a broader study
of the environmental shaping of language in Tibeto-Burman, Mark W.
Post presents a treatment of the interaction between different terrain types
and the semantic evolution of “topographical deixis” (up/down/same-level
directionals) in the Tani languages, focusing especially on Apatani, Galo
and Mising. In the first ever comparative study of Mising dialects,
Jugendra Pegu then presents a paper on the behaviour of the ubiquitous
Tani article-cum-copula -ə. On the basis of his study, Pegu helpfully
suggests a broad division of Mising dialects into geminating (more Galo-
like) and non-geminating (more Adi-like) morphophonological types. It
is likely that Pegu’s findings will eventually prove to be of real
significance for our understanding of the areal and genetic shaping of
Mising, as well as, perhaps, the Tani languages more generally.
The final two sections turn to two of the remaining three genetic
groups represented in North East India, namely Eastern Indo-Aryan and
Austroasiatic. In addition to the expected papers on Standard Assamese,
A Note from the Editors xix
xx North East Indian Linguistics Volume 3
in which Jagat Kalita presents a study of NP referential marking and
modification and Runima Chowdhary works through a detailed analysis
of copula constructions, Virginia Phillips reminds us of the underlying
(and little-studied) diversity of Eastern Indo-Aryan languages in her study
of case marking in Hajong. In the tradition of Satyanath and Laskar
(2008), Phillips attempts to unravel diverse layers of Tibeto-Burman and
Indo-Aryan origin in a language which is obviously a product of intense
and enduring population contacts. Finally, Hemanga Dutta adds
significantly to the existing literature on Austroasiatic languages of North
East India – which has been overwhelmingly dominated by studies of
Standard Khasi – in his study of derivational morphology in Pnar, a
little-known and highly endangered Austroasiatic language of the Jaintia
Hills.
Editorial duties for the present volume have been equally shared
among the alphabetically-listed editors, who engaged in long-term
collaboration among a large group of geographically-dispersed people
– most of whom, to make matters more difficult, are almost constantly
on the move! Paper drafts often went through four or five revisions,
shuttling between authors who might be in Imphal, Tezpur, or Ann Arbor,
reviewers who might be in Delhi, Paris or Canberra, and editors who
might be in Pasighat, Bangkok, or Thimphu on any given day of the
week. It goes without saying that the patience and hard work of many
people – together with the twin modern-day godsends of email and mobile
telephones – were required to bring it off successfully, and we thank
them most sincerely. As always, our greatest thanks go to the staff and
students of the Department of Linguistics at Gauhati University, both for
hosting the 3
rd
NEILS conference and for doing so much to provide a
local and international nexus for the study of North East Indian languages.
Thanks also go to Father Thomas of the Don Bosco Institute for once
again extending use of the beautiful Don Bosco conference facilities –
replete with expansive views of the mighty Brahmaputra! – to Gauhati
University’s ex-Vice Chancellor Prof. Amarjyoti Chowdhury for enabling
financial and logistical support for the conference, and to the staff of the
Research Centre for Linguistic Typology of La Trobe University – in
particular Ms Siew-Peng Condon – for logistical support including
periodical maintenance of the NEILS website (http://www.latrobe.edu.
au/rclt/Neils/Neils.htm). We also thank the editorial staff of Cambridge
University Press India Pvt. Ltd., both for continuing to support the NEILS
endeavour to build and maintain international bridges of quality research
Foreword xxi
and publications in linguistics and for, shall we say, keeping us on our
stylistic toes. One of the distinct pleasures of writing and publishing in
English in India is that one often encounters so-called “non-native
speakers” of English who are both quite prepared to instruct an American
or an Australian on the rules of “proper” English phrasing and entirely
equal to the task! We also thank the dozens of contributors and anonymous
peer-reviewers who worked so hard to make this volume as good as it
could be.
But finally, we would like to take this opportunity to thank the dozens
and possibly hundreds of native-speaker consultants who provided the
database for every paper in this volume, and who continue to maintain
that database by speaking their languages and teaching them to their
children. Everyone who comes to the North East, it seems to us, ends
up wanting to stay, and this is clearly just as much due to the fascinating
opportunities for research and study that exist here as it is due to the
wonderful human-cultural environment in which one is privileged to be
able to reside for a time. So, people of the North East, we thank you,
and to you we dedicate this volume.
Gwendolyn Hyslop
Eugene, Oregon, USA
Stephen Morey Melbourne, Australia
Mark W. Post New Delhi, India
References
Aimol, C. D. (2010). “A comparative study of Kom and Aimol numerals.” In S.
Morey and M. W. Post, Eds., North East Indian Linguistics Vol. 2. New
Delhi, Cambridge University Press India: 155–65.
Bickel, B. (1999). “Nominalization and focus constructions in some Kiranti
languages.” In Y. P. Yadava and W. G. Glover, Eds., Topics in Nepalese
Linguistics. Kathmandu, Royal Nepal Academy: 271–96.
Burling, R. (1983). “The Sal languages.” Linguistics of the Tibeto-Burman Area
7(2): 1–32.
______ (2007). “The lingua franca cycle: Implications for language shift,
language change, and language classification.” Anthropological Linguistics
49(3–4): 207–36.
Betholia, C. (2010). “Universal quantifiers in Manipuri.” In S. Morey and M. W.
Post, Eds., North East Indian Linguistics Vol. 2. New Delhi, Cambridge
University Press India: 210–21.
A Note from the Editors xxi
xxii North East Indian Linguistics Volume 3
Matisoff, J. A. (1972). “Lahu nominalization, relativization and genitivization.”
In J. Kimball, Ed. Syntax and Semantics, Vol. 1. New York, Seminar Press:
237–57.
Satyanath, S. and N. B. Laskar (2008). “Lexicon in a contact language: The case
of Bishnupriya.” In S. Morey and M. W. Post, Eds., North East Indian
Linguistics. New Delhi, Cambridge University Press India: 75–92.
Singh, H. B. (2010). “Negation in Chothe grammar.” In S. Morey and M. W.
Post, Eds., North East Indian Linguistics Vol. 2. New Delhi, Cambridge
University Press: 221–31.
Person-Marking Prefixes of Purum 1
The View from Manipur
2 North East Indian Linguistics Volume 3
2 blank
Person-Marking Prefixes of Purum 3
1
Person-Marking Prefixes of Purum
H. Surmangol Sharma
Deccan College, Pune
N. Gopendro Singh
Language Cell, Directorate of Education (S), Govt. of Manipur
1. Introduction
Purum is a language belonging to the Kuki-Chin group of the Tibeto-
Burman family (Grierson 2005 [1904]). Benedict includes the language
under Kuki-Naga (Kukish) (Benedict 1972). The word Pu rum means
‘hide from tiger’ (Shakespear 1912: 150). Purum is recognized as one
of the 33 languages spoken in Manipur by the State Government of
Manipur, and the language is considered to be highly endangered.
According to the Census of India (2001), the number of Purum speakers
is 503. The teaching of the language has not yet been introduced even
at the primary level of school education. There are two Purum villages,
situated in the Senapati district of Manipur. The names of these villages
are Purum Khullen and Purum Likli; the former is about 18 kms and the
latter 27 kms north of Imphal (the capital of Manipur). Most Purum
speakers are able to speak Manipuri due to the fact that they are in contact
with Manipuri speakers through education, administration, business
transactions, and so on.
There has been confusion as to whether two groups, the Purum and
the Chote, are the same, as we can see from the title of the proceedings
of a symposium held in 1985, “Proceedings of the Symposium on Purum
(Chote) Revisited”. In those proceedings, Hiyang et al. point out that the
4 North East Indian Linguistics Volume 3
name Chote, which has later come to be spelled Chothe, seems to be
confused with Purum because the Chothe people settled in some places
called Purum (Hiyang et al. 1985: 25–26).
However, we consider that Chothe and Purum are distinct languages.
The distinction between Chothe and Purum is shown by the following
facts: first, the verb in Chothe is not inflected with person-marking
prefixes, but just the opposite is the case for Purum, with verb structure
being inflected with person-marking prefixes (see G. Singh 2003; B.
Singh 2008). Second, the core vocabulary in these two languages is
different. Some examples are provided here, where the symbols in the
parentheses stand for language names: ‘Ch’ for Chothe and ‘P’ for Purum,
and the symbol ʰ stands for aspiration: ni (Ch) kini (P) ‘sun’; kərosompə
(Ch) sortʰə (P) ‘moon’; nisuknə (Ch) niso (P) ‘east’; nitək (Ch) təytək
(P) ‘west’; əŋəwpə (Ch) əkəbəŋ (P) ‘white’, əhoŋ (Ch) əkəvəm (P) ‘black’,
and so on.
2. An overview of person-marking prefixes
Personal prefixes are often reduced forms of independent pronouns in
Chin languages (Matisoff 2003: 89) and this may be true for Purum too.
The prefixes in Purum are added to both free nominal words and bound
nominal roots. Here we consider bound nominal roots to be forms of
nouns that do not occur as independent words, for instance, kin terms
such as pa- ‘father’, nu- ‘mother’, etc. in (1a)–(1b). Note that person-
marking prefixes when added to nouns function as possessive markers.
However, these prefixes when occurring on verbs serve as subject-object
agreement markers in a transitive sentence, as shown in (4). The
interesting patterns and combinatory complexity of different person-
marking prefixes for subject and object on a transitive verb will be
discussed in detail in §3, referring to examples (15)–(17) of §3.1 and
(18)–(21) of §3.2.
There are ten person-marking prefixes in Purum and they are given
in Table 1. First, these prefixes are broadly classified into subject/
possessive prefixes and object prefixes. The subject/possessive prefixes
are then categorized into singular and plural ones. There is a distinction
between exclusive and inclusive in the first person subject/possessive
plural prefix, as kin- ‘1.PL.EXCL’ and ni- ‘1.PL.INCL’. The singular person-
marking prefixes maintain a simple and uniform pattern. It seems that
the free pronouns are reduced to the first syllable, which is a single mora,
Person-Marking Prefixes of Purum 5
to form the subject/possessive prefix. On the contrary, each plural subject/
possessive prefix follows a different pattern. Note that object prefixes
exhibit an entirely different pattern from that of subject/possessive
prefixes in that they show no singular-plural distinction. Although the
object agreement on the verb is marked by prefix that does not discern
number, the corresponding free object pronoun could be singular or
plural.
The forms of first person object prefix and second person object prefix
are found to be homophonous, suggesting that it may be a single prefix,
i.e. a non-third person object prefix. Finally, the third person object prefix
is null, for which we use the symbol Ø in the paper. An example of the
SUBJ and OBJ prefixes occurring together on a transitive is found in (18)
below. In §6, we will discuss the Purum negative construction where we
find that verbs are not inflected with the person-marking prefixes
furnished in Table 1, but that agreement between the verb and its
argument(s) is maintained through various negative suffixes.
2.1. Vowel harmony
Purum has six vowels as given in Table 2. There are two vowels each
in front, central and back positions. Non-low front vowels are /i/ and /e/,
non-high central vowels are /
ə/ and /a/, and non-low back vowels are /o/
and /u/.
The vowels of the first person singular prefix and second person
singular prefix undergo regressive vowel harmony with the vowel of the
stem. The vowel in the prefix is mid central /ə/ if the first vowel of the
stem is one of the non-high central vowels, i.e. /ə/ or /a/. Likewise, the
vowel in the prefix is high front /i/ if the stem contains /i/ and mid-high
Table 1 Purum person-marking suffixes
SUBJ/POSS prefixes OBJ
Singular Plural
prefixes
Free Prefixes Free Pronouns Prefixes
Pronouns
EXCL INCL EXCL INCL
1P
kəy kə- kəyni əyni kin- ni- nə-
2P
nəŋ nə- nəŋni nin- nə-
3P
əma ə- ənmani ən- Ø
6 North East Indian Linguistics Volume 3
vowel /e/ in the prefix if the stem contains /e/. This is nicely shown in
examples (1) and (2). The behaviour of the back vowels is the same as
that of the front vowels. Table 3 below presents a pattern of regressive
vowel harmony and this applies both to the subject and object prefixes,
as we can see from (18), where we find kini-; here, the combination of
the 1st person singular subject prefix and the object prefix are realised
with the vowel /i/ in harmony with the vowel of the stem minreŋ
‘love’.
Table 2 Purum vowels
Front Central Back
High i u
Mid-High e o
Mid
ə
Low a
Table 3 Regressive vowel harmony
Vowels of Prefixes Type of vowels
(1
SG and 2SG) present in the stem
ə ə, a
i i
e e
o o
u u
2.2. Person-marking prefixes on nouns
Purum person-marking prefixes function as possessive markers when
they occur on nouns which are inalienable, such as kin terms and body-
parts. The names of a few domestic animals, which are traditionally
considered very essential, can take person-marking prefixes. Consider
examples (1a)–(1b) showing person-marking prefixes with kin terms,
(2a)–(2c) with body-parts and (3a)–(3b) with domestic animals. Person-
marking prefixes do not occur on alienable nouns as evidenced from
Person-Marking Prefixes of Purum 7
examples (3c)–(3d). The elicited data we use here are taken from G.
Singh (2003).
(1a)
kəpa (1b) kinpi
kə-pa kin-pi
1SG-father 1PL-grandmother
‘my father’ ‘our grandmother’
(2a) nokor (2b) ninlubur
nə-kor nin-lubur
2SG-ear 2PL-head
‘your ear’ ‘your heads’
(2c) əkut
ə-kut
3SG-hand
‘his/her hand’
(3a) kovok (3b) kisirat
kə-vok kə-sirat
1SG-pig 1SG-cow
‘my pig’ ‘my cow’
(3c) *nunulur (3d) *nuskul
nə-nulur nə-skul
2SG-snake 2SG-school
‘my snake’ ‘your school’
2.3. First person singular prefix
The first person singular prefix is realized as
kə- ~ ki- ~ ke- ~ ku- ~ ko-.
The alternation is triggered by the presence of a specific vowel in the
stem as discussed in §2.1. These are glossed as ‘1SG’ in the language
data. These are exemplified in (4), where they are now functioning as
subject prefixes, though the form of the prefix is the same as the
possessive with nouns, as already shown in (1a)–(1b) and (3a). The third
person object prefix is marked as Ø ‘3OBJ’ in (4) and will be discussed
further in §3.1.
(4)
kəy bu kəbək.
kəy bu kə-Ø-bək
I rice 1SG-3OBJ-eat
‘I eat rice.’
8 North East Indian Linguistics Volume 3
2.4. First person plural prefix
The “inclusive” and “exclusive” distinction of first person plural pronouns
and their corresponding prefixes refers to first person dual or plural
pronouns. A pronoun is said to be “inclusive” where the addressee as
well as the speaker is included while an “exclusive” pronoun does not
include the person being addressed (Crystal 1999: 158). This distinction
is seen to be applicable for Purum. When the first person plural subject
pronoun includes the listener(s), the verb takes ni- ‘1
PL.INCL’ when the
listener(s) are not considered as a part of the subject argument, the verb
takes kin- ‘1PL.EXCL’. The same prefixes indicate possessor pronouns
when occurring on nouns. This plural inclusive/exclusive distinction is
still maintained when a noun is prefixed with ni- or kin-. Examples
(5a)–(5b) are with kin terms and (6a)–(6d) are with verbs.
(5a) kinpa (5b) nipa
kin-pa ni-pa
1
PL.EXCL-father 1PL.INCL-father
‘our father’ (exclusive) ‘our father’ (inclusive)
(6a) kəyni bu kinbək.
kəyni bu kin-Ø-bək
we.EXCL rice 1PL.EXCL-3OBJ-eat
‘We eat rice.’ (exclusive)
(6b) əyni bu nibək.
əyni bu ni-Ø-bək
we.INCL rice 1PL.INCL-3OBJ-eat
‘We eat rice.’ (inclusive)
(6c) *əyni niŋ nəŋ niminreŋ.
əyni niŋ nəŋ ni-minreŋ
we.INCL AGT you 1PL.INCL-love
‘I and you love you.’
(6d) kəyni iŋkhet le iŋkhət kinkinrieŋ.
k
əyni iŋkhət le iŋkhət kin-kin-rieŋ
we.EXCL one ASSO one 1PL.EXCL-1PL.EXCL-love
‘We love each other.’
Example (6c) turns out to be ungrammatical because the first person
plural inclusive subject co-occurs with a second person in the object
position. Note that in (6d), the verb is inflected with kin- ‘1
PL.EXCL’ twice
Person-Marking Prefixes of Purum 9
and the sentence has a different verb stem, i.e. rieŋ rather than minreŋ
‘love’.
2.5. Second person singular prefix
The second person singular prefix
nə- has the allomorphs nu- ~ ni- ~
ne- ~ no-. The condition of allomorphy for nə- is found to be the same
as presented in Table 3. In (7a)–(7b), nə- is added to the kin terms, and
in (8a)–(8d) it is added to the verbs. NP arguments in a Purum sentence
are not always distinguished through case marking. The subject and the
object are assigned with case markers when the speaker is referring to a
specific rather than general entity, or when it is necessary to distinguish
the subject from the object or vice-versa. Note that in (8a)–(8c) the
subjects of these sentences do not co-occur with niŋ ‘AGT’ for the simple
reason that the personal pronoun nəŋ ‘you’ can easily be recognized as
the actor of the verb rather than the other NPs bu ‘rice’ in (8a) and tuy
‘water’ in (8c) and in (8b) where the verb is an intransitive motion verb
it easily selects its actor as nəŋ ‘you’; hence, the subject does not require
niŋ ‘AGT’, although it is also possible to include it.
Example (8d) suggests that there are two NPs of equal status, either
of which could be the subject or the object in the sentence; therefore,
the subject and object NPs are identified by niŋ ‘AGT’ and hi ‘ACC’
respectively. It is also possible that only the subject NP is marked with
niŋ ‘AGT’ while the object NP (though it is of equal status with the subject
NP) is left without case marking as in (8e). We might assert at this level
of analysis that the subject NP tends to be case marked rather than the
object NP when distinguishing the grammatical function of one from the
other. In (8f) the nature of the verb sop ‘wash’ determines what NPs are
its subject and object but still the subject is marked with niŋ to establish
the actor’s mode of engagement in the particular action as it might be
performed at a certain point of time, or pointing out a particular actor.
We assume here that the agentive suffix could also be relative to a certain
point of time, with the subject being seen to be engaged in the particular
action on a temporary/limited basis, rather than a habitual or long-term
basis. Thus, the particular action in which the subject engages might not
yet be denoted as a sort of profession or occupation for the actor, nor
yet achieving a situation where the name of the particular action could
be used as an “attribute” to the actor. In general, we would expect that
habitual sentences would not have agent marking. We will not, however,
10 North East Indian Linguistics Volume 3
present a detailed analysis of the Purum case system for reasons of its
exceeding the scope of this paper.
(7a)
nəpa (7b) nunu
nə-pa nə-nu
2SG-father 2SG-mother
‘your father’ ‘your mother’
(8a) nəŋ bu nəbək. (8b) nəŋ skul nese.
nəŋ bu nə-Ø-bək nəŋ skul nə-se
you rice 2SG-3OBJ-eat you school 2SG-go
‘You eat rice.’ ‘You go to school.’
(8c) nəŋ tuy niin.
nəŋ tuy nə-Ø-in
you water 2SG-3 OBJ-drink
‘You drink water.’
(8d) nəŋ niŋ əma hi numunom.
nəŋ niŋ əma hi nə-Ø-munom
you AGT he ACC 2SG-3OBJ-like
‘You like him.’
(8e) nəŋ niŋ bil niminreŋ.
nəŋ niŋ bil nə-Ø-minreŋ
you AGT Bill 2SG-3OBJ-love
‘You love Bill.’
(8f) nəŋ niŋ kʰeŋ nosop.
nəŋ niŋ kʰeŋ nə-Ø-sop
you AGT plate 2SG-3OBJ-wash
‘You are washing plates (I know).’
2.6. Second person plural prefix
The prefix nin- ‘2
PL’ can be added to both nouns and verbs. Moreover,
the prefix is not subject to vowel harmony rules, which are applicable
to the second person singular prefix. Vowel harmony doesn’t apply to
first person plural prefix either, nor to third person singular prefix as we
will see in the next section. The reason why vowel harmony applies to
some prefixes and not others is a matter that will require further research.
In (9a)–(9b), nin- is added to kin terms while in (10) it is added to
verbs.
Person-Marking Prefixes of Purum 11
(9a) ninpa (9b) ninnu
nin-pa nin-nu
2PL-father 2PL-mother
‘your (pl.) father’ ‘your (pl.) mother’
(10) nəŋni niŋ əma hi ninminreŋ.
nəŋni niŋ əma hi nin-Ø-minreŋ
you AGT he ACC 2PL-3OBJ-love
‘You all love him.’
2.7. Third person singular prefix
As with other singular person-marking prefixes discussed above,
ə- ‘3SG’
can also be added to both nouns and verbs but this prefix does not exhibit
vowel harmony. In (11a)–(11b), ə- is prefixed to kin terms and in (12)
to the verb.
(11a)
əpa (11b) əpi
ə-pa ə-pi
3SG-father 3SG-grandmother
‘his/her father’ ‘his/her grandmother’
(12) əma bu əbək.
əma bu ə-Ø-bək
(s)he rice 3SG-3OBJ-eat
‘(S)he eats rice.’
2.8. Third person plural prefix
ən- ‘3PL’ can be added to both nouns and verbs. In (13a)–(13b) ən- is
prefixed to kin terms, which enables us to identify whose mother, father,
etc. is being referenced, and in (14) it is seen to be co-occurring with
the verb as a subject agreement of person and number.
(13a)
ənpa (13b) ənnu
ən-pa ən-nu
3PL-father 3PL-mother
‘their father’ ‘their mother’
(14) ənmani bu ənbək.
ənmani bu ən-Ø-bək
they rice 3PL-3OBJ-eat
‘They eat rice.’
12 North East Indian Linguistics Volume 3
3. Agreement features of subject and object on verb
This discussion about the agreement features, especially of number
marking of subject and object personal prefixes on the transitive verb, is
done on the basis of whether the sentence contains a free third person
object pronoun or a free non-third person object pronoun. A sentence
containing a free third person object pronoun is seen to have its verb
distinctly marked with a singular or a plural subject prefix in terms of
the corresponding free subject pronoun’s number. In other words, if the
subject is plural, the verb is inflected with the corresponding personal
plural prefix, and this is also the case if the free subject pronoun is
singular. In addition, a third person object personal prefix is marked on
the verb as Ø ‘3OBJ’ irrespective of whether the corresponding free object
pronoun in the sentence is singular or plural. We can see this in Nos. 2,
3, 4, 6, 7, 10 and 11 in Table 4 where the verbs are seen to be inflected
with singular and plural subject prefixes and the third person object prefix
as Ø, regardless of the number of free object pronoun. For the relevant
examples, see (15)–(17).
Turning back to Table 4, Nos. 1, 5, 8 and 9 show examples containing
a non-third person (first or second person) free object pronoun and a free
subject pronoun (first, second or third person). In this case, a non-third-
person object is marked on the verb by a single prefix, i.e. nə-, which is
glossed in the paper as ‘1OBJ’ or ‘2OBJ’ depending on the specific
corresponding free object pronoun; moreover, the prefix nə- does not
denote the number of the free object pronoun.
Another interesting point we may note here is the change in behaviour
of subject prefix-marking on the verb where the sentence has a free non-
third person object. In that case the verb is marked with a singular subject
prefix, albeit that the corresponding free subject pronoun is plural as in
(18a). In other words, the Purum subject-verb agreement system in the
presence of a non-third person object might be asserted as showing no
agreement of number (feature) between the subject and the verb because
(a) the verb is always marked with singular subject prefix, and (b) though
the verb is marked with singular subject prefix, the corresponding free
subject pronoun could be plural.
3.1. Sentences with a third person object
As evidenced from Nos. 2, 3, 4, 6, 7, 10 & 11 of Table 4 above, the
third person object prefix is represented as zero. Hence, the expected
Person-Marking Prefixes of Purum 13
position of third person object prefix in the verb is marked by the symbol
Ø in the following examples (15)–(17).
(15a)
kəy niŋ əma kiminreŋ.
kəy niŋ əma kə-Ø-minreŋ
I AGT he 1SG-3OBJ-love
‘I love him.’
(15b) kəyni niŋ əma kinminreŋ.
kəyni niŋ əma kin-Ø-minreŋ
we AGT he 1PL.EXCL-3OBJ-love
‘We love him.’
Table 4 Subject and object marking on the verb
No. Subject Object Verb
1 First person (
SG/PL) Second person (SG/PL) kə-nə-V
1
SG-2OBJ-V
2 First person (
SG) Third person (SG/PL) kə-Ø-V
1
SG-3OBJ-V
3 First person (
PL.EXCL) Third person (SG/PL) kin-Ø-V
1
PL.EXCL-3OBJ-V
4 First person (
PL.INCL) Third person (SG/PL) ni-Ø-V
1PL.INCL-3OBJ-V
5 Second person (
SG/PL) First person (SG/PL) nə-nə-V
2
SG-1OBJ-V
6 Second person (
SG) Third person (SG/PL) nə-Ø-V
2
SG-3OBJ-V
7 Second person (
PL) Third person (SG/PL) nin-Ø-V
2
PL-3OBJ-V
8 Third person (
SG/PL) First person (SG/PL) ə-nə-V
3SG-1OBJ-V
9 Third person (
SG/PL) Second person (SG/PL) ə-nə-V
3SG-2OBJ-V
10 Third person (
SG) Third person (SG/PL) ə-Ø-V
3
SG-3OBJ-V
11 Third person (
PL) Third person (SG/PL) ən-Ø-V
3
PL-3OBJ-V
14 North East Indian Linguistics Volume 3
(15c) kəy niŋ ənmani kətʰətye.
kəy niŋ ənmani kə-Ø-tʰət-ye
I AGT they 1SG-3OBJ-kill-PERF
‘I have killed them.’
(16a) nəŋ niŋ əma nətʰətye.
nəŋ niŋ əma nə-Ø-tʰət-ye
you AGT he 2SG-3OBJ-kill-PERF
‘You have killed him.’
(16b) nəŋni niŋ əma nintʰətye.
nəŋni niŋ əma nin-Ø-tʰət-ye
you.PL AGT he 2PL-3OBJ-kill-PERF
‘You all have killed him.’
(16c) nəŋni niŋ ənmani ninminreŋ.
nəŋni niŋ ənmani nin-Ø-minreŋ
you (pl) AGT they 2PL-3OBJ-love
‘You all love them.’
(17a) əma niŋ nurul ətʰətye.
əma niŋ nurul ə-Ø-tʰət-ye
he AGT snake 3SG-3OBJ-kill-PERF
‘He has killed a snake.’
(17b) ənmani niŋ əma əntʰətye.
ənmani niŋ əma ən-Ø-tʰət-ye
they AGT he 3PL-3OBJ-kill-PERF
‘They have killed him.’
3.2. Sentences with non-third person object
This section deals with the type of sentences in which the object is either
first person
SG/PL or second person SG/PL. Examples (18)–(19) present
sentences in which objects are second person, subjects are first person
and third person while in (20)–(21) objects are first person, subjects are
second person and third person.
(18a)
kəyni niŋ nəŋ kiniminreŋ.
kəyni niŋ nəŋ kə-nə-minreŋ
we AGT you 1SG-2OBJ-love
‘We love you.’
(18b) kəy niŋ nəŋni kiniminreŋ.
kəy niŋ nəŋni kə-nə-minreŋ
I AGT you.PL 1SG-2OBJ-love
‘I love you all.’
Person-Marking Prefixes of Purum 15
(19a) ənmani niŋ nəŋ əneren.
ənmani niŋ nəŋ ə-nə-ren
they AGT you 3SG-2OBJ-scold
‘They scold you.’
(19b) əma niŋ nəŋni əniminreŋ.
əma niŋ nəŋni ə-nə-minreŋ
he AGT you.PL 3SG-2OBJ-love
‘He loves you all.’
(20a) nəŋni niŋ kəy niniminreŋ.
nəŋni niŋ kəy nə-nə-minreŋ
you.PL AGT I 2SG-1OBJ-love
‘You all love me.’
(20b) nəŋ niŋ kəyni nənətʰətye.
nəŋ niŋ kəyni nə-nə-tʰət-ye
you AGT we 2SG-1OBJ-kill-PERF
‘(The situation turns out to be as) you have killed us.’
(21a) ənmani niŋ kəy əniminreŋ.
ənmani niŋ kəy ə-nə-minreŋ
they AGT I 3SG-1OBJ-love
‘They love me.’
(21b) əma niŋ əyni
1
əniminreŋ.
əma niŋ əyni ə-nə-minreŋ
he AGT we.INCL 3SG-1OBJ-love
‘He loves us.’
Note that sentences (18)–(21) with non-third-person objects have their
verbs inflected with subject and object person-marking prefixes. Referring
to (18) it is clear that a first person object prefix is marked differently
(as
nə- ‘1OBJ’) than a first person subject prefix (kə- ‘1SG’) on the verb.
Purum prefixes follow a SUBJECT PREFIX – OBJECT PREFIX – VERB ordering
pattern, which is the same as the language’s basic SUBJECT – OBJECT – VERB
word order. A sentence in which the subject is second person and the
object is first person, such as example (20), exhibits a sequence of
homophonous prefixes along with the verb; for instance, nə-nə-V ‘2SG-
1OBJ-V’. In that case, the order of the full pronouns will disambiguate
1
Our consultant Ms Sana Bapuri prefers that an independent first person plural
object be in the inclusive form rather than the exclusive form when the subject
in the sentence is third person singural/plural.
16 North East Indian Linguistics Volume 3
the roles. However, the question of how fixed these orders are must be
left to further research.
Another important point that needs to be explained here is that the
object prefix nə- may refer to first person object or second person object
regardless of the subject. By merely looking at the structure of the verb
we may fail to identify what nə- represents, so the overt presence of a
free object pronoun in the sentence is considered obligatory. Thus, the
correspondence between the prefix nə- and independent (first person/
second person) object can be established without much difficulty.
Consider examples (22a)–(22b) in which the underlined portions – being
considered obligatory – enable us to determine whether nə- represents a
first person object or a second person object.
(22a)
kəy əneren.
kəy ə-nə-ren
I 3SG-1OBJ-scold
‘He/She/They scold(s) me.’
(22b) ənmani niŋ nəŋ əneren
ənmani niŋ nəŋ ə-nə-ren
they AGT you 3SG-2OBJ-scold
‘They scold you.’
4. A comparison with Aimol
At this point, we would like to highlight a few important points of Aimol
person-marking system in Table 5, because Aimol is a closely related
language to Purum
2
. Aimol uses first person subject prefixes such as
kə- ‘1SG’, kən- ‘1PL’ and object prefix nə- ‘1OBJ’. In both the languages,
the first person object prefix does not make singular-plural distinction.
Further, Aimol uses subject person-marking prefixes such as nə- ‘2SG’,
nən- ‘2PL’ and object person-marking suffixes -ce ‘2SG.OBJ’, -ceu ‘2PL’.
OBJ’ while Purum uses only one second person object prefix. The third
person object prefix in Aimol is marked as zero on the verb as in
Purum.
2
Aimol data appearing in this paper are provided by C.D. Aimol (p.c.), and
should be considered preliminary in this context.
Person-Marking Prefixes of Purum 17
5. Ditransitive verbs
In Purum the verb can also be inflected with an indirect object prefix. In
(23) we see a first person subject and second person indirect object both
marked. If the indirect object is a third person, its prefix on the verb is
represented as zero, as in (24).
(23a)
kəy niŋ nəŋ deŋ kolom kənəpek.
kəy niŋ nəŋ deŋ kolom kə-nə-Ø-pek
I AGT you DAT pen 1SG-2OBJ-3OBJ-give
‘I give you a pen.’
(23b) əma niŋ kəy deŋ kolom ənəpek.
əma niŋ kəy deŋ kolom ə-nə-Ø-pek
he AGT I DAT pen 3SG-1OBJ-3OBJ-give
‘He gives me a pen.’
(23c) əma niŋ (nəŋ deŋ) kolom ənəpe kəy niŋ kəjət.
əma niŋ (nəŋ deŋ) kolom ə-nə-Ø-pe
he AGT (you DAT) pen 3SG-2OBJ-3OBJ-give
kəy niŋ kəjət
I AGT know
‘I know, he gave you a pen.’
(24a) jon niŋ əma deŋ jedət əpe.
jon niŋ əma deŋ jedət ə-Ø-pe
John AGT he DAT book 3SG-3OBJ-give
‘John gives him a book.’
Table 5 Aimol subject and object person-marking affixes
Subject/Object Singular Plural Context
1 1P Subject
kə- kən- with 2P/3P object or the verb is
intransitive
2 1P Object
nə- with 2P/3P subject
3 2P Subject
nə- nən- with 1P/3P object or the verb is
intransitive
4 2P Object -ce -ceu with 1P/3P subject
5 3P Subject
ə- ən- with 1P/2P/3P object or the verb is
intransitive
6 3P Object
Ø- with 1P/2P/3P subject
18 North East Indian Linguistics Volume 3
(24b) jon niŋ bil deŋ iŋlis əmincu.
jon niŋ bil deŋ iŋlis ə-Ø-mincu
John AGT Bill DAT English 3SG-3OBJ-teach
‘John teaches English to Bill.’
6. Sentences without person-marking prefixes
This section deals with negative sentences in which a negative suffix is
added to the verb but there is no person-marking prefix on the verb.
Purum uses different negative suffixes that agree with subjects in person
and number as shown in Table 6. It is interesting to note that both the
first person singular negative suffix and the first person plural exclusive
negative suffix end in the velar nasal /ŋ/ and the first person plural
inclusive negative suffix uses a different form as in No. 2 of Table 6.
The third person negative suffixes end in the velar stop /k/. Note that the
negative sentence uses the same object prefixes as found in the assertive
sentence; the non-third person object prefix is nə- and the third person
object prefix is represented as Ø. Consider the examples in (25).
Table 6 Negative suffixes
Subject Negative suffix
1 First person singular -
məŋ/-moŋ
2 First person plural inclusive -
məkmi
3 First person plural exclusive -mu
ŋ
4 Second person singular -
məkce
5 Second person plural -mukcu
6 Third person singular -
mək
(25a) kəy tʰəyhay bəknomməŋ.
kəy tʰəyhay Ø-bək-nom-məŋ
I mango 3OBJ-eat-want-1SG NEG
‘I don’t want to eat mango.’
(25b) əyni tʰəyhay bəknomməkmi.
əyni tʰəyhay Ø-bək-nom-məkmi
we (INCL) mango 3OBJ-eat-want-1PL NEG
‘We don’t want to eat mango.’
Person-Marking Prefixes of Purum 19
(25c) nəŋ daktər iməkce kəy niŋ kəjət.
nəŋ daktər i-məkce kəy niŋ kəjət
you doctor be-2SG.NEG I AGT know
‘I know, you are not a doctor.’
(25d) nəŋni bu bək nomməkcu kəy niŋ kəjət.
nəŋni bu Ø-bək-nom-məkcu kəy niŋ kəjət
you (PL) rice 3OBJ-eat-want-2PL.NEG I AGT know
‘I know, you all don’t want to eat rice.’
(25e) əma yu inmək.
əma yu Ø-in-mək
he wine 3OBJ-drink-3SG.NEG
‘He doesn’t drink wine.’
(25f) ənmani bəjarrə semuŋ.
ənmani bəjar-rə se-muk
they market-loc go-3PL.NEG
‘They didn’t go to the market.’
(25g) kəy niŋ nəŋ niminreŋmoŋ.
kəy niŋ nəŋ nə-minreŋ-moŋ
I AGT you 2OBJ-love-1SG.NEG
‘I don’t love you.’
(25h) əma niŋ kəy nərəcakmək.
əma niŋ kəy nə-rəcak-mək
he AGT I 1OBJ-hit-3SG.NEG
‘He doesn’t hit you.’
7. Other major types of sentences
We are concerned here with interrogative, imperative and reflexive
sentences, in order to check whether the verbs in these sentences are
inflected with person-marking prefixes and if so, in what ways.
7.1. Interrogative sentences
Person-marking prefixes are used in interrogative sentences. Yes/No
questions in Purum are formed by the addition of mi ‘
INT’ to a noun or
a verb. In (26a), the suffix is added to a noun while in (26b) to a verb.
In both cases the person-marking prefix is also present.
20 North East Indian Linguistics Volume 3
(26a) əma hi nəŋ ta nəsəpami?
əma hi nəŋ ta nə-səpa-mi
he ACC you GEN 2SG-son-INT
‘Is he your son?’
(26b) nəŋ yu niinmi?
nəŋ yu ni-in-mi
you wine 2SG-drink-INT
‘Do you drink wine?’
In a Wh-question construction, the verb is inflected with a person-
marking prefix as in (27).
(27)
nəŋ niŋ satrə kʰoŋjətmo iŋlis nimincu?
nəŋ niŋ satrə kʰoŋjətmo iŋlis ni-mincu
you AGT student how many English 2SG-teach
‘How many students do you teach English to?’
7.2. Imperative sentences
In imperative constructions, the command suffix -ro ‘
COMD’ is added to
the verb. The person-marking prefix that corresponds to the subject of
the sentence does not occur on the verb but the prefix corresponding to
the object occurs on the verb as in (28).
(28) kolom n
ə-pe-ro.
kolom nə-pe-ro
pen 1OBJ-give-COMD
‘Give me a pen.’
7.3. Reflexive sentences
In this kind of construction the person-marking prefix corresponding to
the subject of the sentence occurs on the verb; however, that of the object
does not. Note however that the verb also carries a prefix kin- ‘
REFL’. It
is also seen that the object in a reflexive sentence bears the person-
marking prefix which corresponds to the subject concerned. Consider
example (29).
(29)
əyni nibiŋŋa nikinminreŋ.
əyni ni-biŋ-ŋa ni-kin-minreŋ
we (INCL) 1PL.INCL-self-LOC 1PL.INCL-REFL-love
‘We love ourselves.’
Person-Marking Prefixes of Purum 21
8. Conclusion
Purum personal prefixes function as possessive markers when occurring
with inalienable nouns and as agreement prefixes, marking the person
and number of the subject and object on the verb. Alienable nouns do
not occur with person-marking prefixes; so, for example, the construction
*nə-skul ‘your school’ is unacceptable. It is clear from the data reviewed
above that Purum treats third person objects differently as there is no
overt prefix corresponding to this argument type on the verb, whereas
first and second person object prefixes are represented via a single form,
i.e. nə- ‘1OBJ/2OBJ’. When there are both subject and object prefixes
together, the subject is expressed in the singular. Hence, the prefix carries
only the agreement feature of person, and not the agreement feature of
number, in that case.
As the prefixes of first person object and second person object are
homophonous, the free object slot (intended to be filled in by first person/
second person) in the sentence is not allowed to be left empty, especially
when the subject is third person. A ditransitive verb is inflected for the
person prefixes of subject and indirect object, not of direct object. The
verb in a negative sentence is seen without a person-marking prefix due
to the fact that the negative suffix on the verb itself carries the agreement
of person and number with the subject. In this case, Purum might be said
to be employing an alternative agreement system of person and number
between the verb and its argument. That the verb in an imperative
sentence is seen without a subject person-marking prefix is due to the
fact that the subject in such a sentence is always understood as second
person. In reflexive sentences the verb is inflected with a subject prefix
and a reflexive prefix but not with an object prefix. This might be because
of the fact that the reflexive prefix determines that the subject and the
object in the sentence are the same.
At this stage of study, we are still left with some important issues
which require further investigation. For example: (a) what causes a first
person to acquire an object prefix which is different from its subject
prefix, though this is not so for second person? (b) is there any other
strategy (not discussed in the paper) to disambiguate the homophonous
prefix of first person object and second person object?
22 North East Indian Linguistics Volume 3
Abbreviations
1PL First person plural
1SG First person singular
2PL Second person plural
2SG Second person singular
3PL Third person plural
3SG Third person singular
ACC Accusative
AGT Agentive
ASSO Associative
COMD Command
DAT Dative
GEN Genitive
INDIR Indirect
INT Interrogative
INT Negative
NOM Nominalizer
OBJ Object
PL Plural
POSS Possessive
REFL Reflexive
SG Singular
SUBJ Subject
V Verb
References
Benedict, P. K. (1972). Sino-Tibetan: A Conspectus. Cambridge, Cambridge
University Press.
Census of India (2001). Downloaded from http://www.censusindia.in/. Accessed
12/8/2008.
Crystal, D. (1999). A Dictionary of Language. Chicago, University of Chicago
Press.
Grierson, G. A., Ed. (2005 [1904]). Linguistic Survey of India, Volume 3: Tibeto-
Burman Family, Part 3: Kuki-Chin and Burma Groups. New Delhi, Low
Price Publications.
Hiyang (Khiyang), Thambaljao, M. Yuhlung, C. Yuhlung, H. Yuhlung, and A.
Yuhlung. (1985). “The original history of Chothe (Chote) tribe.” In B. K.
Das Gupta, A. Basu, B. Raychaudhari, and P. Gupta, Eds., Proceedings of
Person-Marking Prefixes of Purum 23
the Symposium on Purum (Chote) Revisited. Calcutta, Anthropological
Survey of India.
Matisoff, J. A. (2003). Handbook of Proto-Tibeto-Burman: System and
Philosophy of Sino-Tibetan Reconstruction. Berkeley, University of
California Press.
Shakespear, J. (1983 [1912]). The Lushai Kuki Clans. Delhi, Cultural Publishing
House.
Singh, B. H. (2008). Chothe Grammar. New Delhi, Akansha Publishing House.
Singh, G. N. (2003). A Descriptive Grammar of Purum. PhD Thesis. Imphal,
Manipur University.
24 North East Indian Linguistics Volume 3
2
The Evolution and Recent Development of
the Meetei Mayek Script
Harimohon Thounaojam Singh
Language Cell, Directorate of Education, Government of Manipur
Manipur is a small state in North East India. Many different communities
inhabit the state. The total population of the state is approximately
2,400,000 (Directorate of Census Operations, Manipur 2001). There are
thirty-three different tribal communities which are recognized by the
state government. They are Aimol, Anal, Angami, Chiru, Chothe, Gangte,
Hmar, Kabui, Kacha Naga, Kharam, Koirao, Koireng, Kom, Lamkang,
Mao, Maram, Maring, Monsang, Moyon, Paite, Paomai, Purum, Ralte,
Sema, Simte, Sahte, Tangkhul, Tarao, Thadou, Vaiphei, Zou, any Kuki
and any Mizo (Lushai). They speak different languages, all of which
belong to the Tibeto-Burman family. Among them nine different
languages are recognized as Modern Indian Languages. They are Thadou,
Paite, Tangkhul, Hmar, Rongmei (Kabui), Mao, Zhou, Kom and Vaiphei.
There are also Meetei Muslims, called Pangals, in the state.
In addition to these groups are the Meeteis, who comprise the largest
community in terms of population. The people of Manipur are collectively
known as Manipuri, and this term is also used for the state official
language of Manipur, which is one of the Modern Indian Languages.
This language, which is in fact the first language of the Meetei people,
is also called Meeteirol/Meeteilon. Manipuri/Meeteilon is the lingua
franca of the state. The language has been regarded as one of the most
important regional languages of the nation and has been listed in the
Eighth Schedule of the constitution of India. People speaking the language
The Evolution and Recent Development of the Meetei Mayek Script 25
also live in the neighbouring states of Assam and Tripura and in some
parts of Bangladesh and Burma, now Myanmar.
Manipuri/Meeteilon is a Tibeto-Burman language. It has its own script.
This script is found in the old Manipuri/Meetei books known as puya,
and is known as Meetei Mayek script. It is difficult to establish the exact
period of the script’s origin. The accepted date for the earliest use of the
script is between the 11th and 12th centuries AD (Manihar Singh 1984:
23). A stone inscription found at Khoibu in Tengnoupal district of the
state contains royal edicts of Kiyamba (1467–1508), one of the Meetei
kings. The royal chronicle Cheitharol Kumbaba commenced from this
time (Ibungohal and Khelchandra 2005). In the written records it is stated
that the script was established before the reign of the Meetei king
Khagemba (1592–1652). He popularized the spread of education and the
production of manuscripts using the script. The script continued to be
used to write the Manipuri/Meeteilon language until the late 18 century
AD.
During the reign of the Meetei king Garibaniwaz (also known as
Pamheiba, 1709–48), there was a religious revolution, with large numbers
of people changing to Hinduism in the state. The king embraced Hinduism
in 1717 and many Hindu texts, such as the Ramayana and the
Mahabharata, were translated into Manipuri/Meeteilon in the Meetei
Mayek script. After the Meetei adopted Hinduism, many literary works
written in Meetei script about the pre-Hindu religion as well as other
historical documents were burnt and the Bengali script was imposed on
the Meeteis between 1709 and 1941, thereby replacing Meetei Mayek.
The Bengali script subsequently became the official script of the palace
(known as the konung). Consequently the use of Meetei Mayek for all
genres of writing has declined and the Bengali script has been used by
Meeteis as their own script till now.
There have been many dissenting voices from Meetei scholars about
the use of the Bengali script. The most strident was that of one
Lainingthou Naoriya Phullo. In 1930, he initiated a movement to revive
the Meetei traditions. He developed a new script and named it after
himself, calling it Naoriya Mayek. The voice of revivalism, together with
the new script, soon spread throughout Manipur (for further information
see Sadokpam (2008)).
In 1947, an organization called Meetei Marup was established in
Manipur to propagate the Naoriya Phullo script. After three years, in the
26 North East Indian Linguistics Volume 3
year 1950, serious debates on the script began. Many different opinions
came from among the Meetei scholars. The most controversial aspects
regarded the number of letters in the alphabet, the order and way of
reciting the alphabet, and its origin. In 1958 a state level committee called
Mayek Luptin Committee (MLC) was then formed to conduct a study
of these controversies. A finding of the Committee was that there are
only eighteen letters in the Meetei Mayek script. Though many of the
puyas had been destroyed, some puyas like Wakoklon Thilel Salai
Amailon Pukok and Khununglon Sakok Thilel survived. These were
written in the Meetei script with eighteen consonants. After a thorough
study of the puyas it was confirmed that the genuine Meetei Mayek had
only eighteen letters. So the committee recommended that those eighteen
letters only be used.
The following observations are based on the author’s conversations
with supporters of two differing versions of Meetei Mayek. The main
controversy is whether Meetei Mayek as written with eighteen letters,
or rather another version of Meetei Mayek which is based more closely
on Bengali with thirty-five letters, be used. The two scripts are also
different in the way the letters are organized: the thirty-five-letter alphabet
is organized in Bengali alphabetical order, according to points of
Table 1 The eighteen letters of the traditional Meetei Mayek alphabet,
arranged in the traditional body-part order
a s d
kok ‘head’
səm ‘hair’ lai ‘forehead’
f l k
mit
‘eye’ pa ‘eyelash’ na ‘ear’
j
h g
cil
‘lips’ til ‘saliva’ khəu ‘throat’
q w e
ŋə
u ‘pharynx’ thəu ‘chest’ wai ‘navel’
r t p
ya
ŋ ‘backbone’ huk ‘lower spine’ un ‘skin’
o
i u
i
‘blood’ phəm ‘placenta’ ətiya ‘sky’
The Evolution and Recent Development of the Meetei Mayek Script 27
articulation, while the eighteen letter alphabet is organized according to
parts of the human body, as shown in Table 1.
There are some additional differences. The people who support the
eighteen letter alphabet argue that since the script does not have any
letters for voiced sounds, there were in fact no voiced sounds in archaic
Meetei
1
. So, the script with eighteen letters of the alphabet is the only
original Meetei script. They further argue that the thirty-five-letter
alphabet was manufactured and modeled on the Bengali script after the
coming of Hinduism to Manipur. Additionally, the script includes four
symbols for [s] and like sounds such as [š] and [kš], being Bengali-
derived orthographic contrasts that are not in fact pronounced contrastively
by Meetei speakers. Finally, since the script was based closely on Bengali
it is recited in the same order as Bengali is.
The thirty-five-letter alphabet is exemplified in Grierson (2005 [1904]:
22) and also in Chelliah (1997: 359), where the last letter of Table 1
ətiya is included with the vowels. Those people who support the thirty-
five-letter alphabet argue that the eighteen letter alphabet is not sufficient
for reading all the puyas written in the Meetei Mayek script. In their
view it is therefore necessary to learn the thirty-five-letter alphabet in
order to able to read the puyas. They conclude that theirs is the only
script that can be used to read all the puyas. Consequently they claim
that the thirty-five-letter alphabet is the only viable Meetei script.
Meanwhile supporters of the eighteen letter alphabet claim that the
puyas which cannot be read with knowledge of the eighteen letter alphabet
are not in fact the original puyas. They are the puyas mainly written on
Hindu epics and mythologies after the coming of Hinduism in the
state.
The thirteenth meeting of the Expert Committee of the state
government, held on 2 February 1979, discussed the inclusion of
characters representing nine voiced consonants to the existing eighteen
letters of the alphabet in order to represent additional phonemes in
modern-day Meeteilon/Manipuri. The Committee approved the increase
of eighteen consonant letters to twenty-seven for the convenience of
1
There are voiced sounds in modern spoken Meetei, although as Chelliah
(1997: 18) points out, they are only contrastive with voiceless stops in word
medial position. For a full list of phonemes in the modern spoken language,
including a voiced aspirate series only used in words borrowed from Indo-
Aryan languages, see Chelliah 1997: 17.
28 North East Indian Linguistics Volume 3
modern-day users, and submitted their report to the State Government
(Meetei Mayek Expert Committee 1979). Regarding this report the State
Government recognized the twenty-seven-letter alphabet in the year 1980
vide Govt. letter No. 1/2/78/SS/E of 16/04/80, Gazette No. 33 of
22/04/1980. On the 19 January 1983, the Department of Education
(School), Government of Manipur, prescribed the Meetei Mayek Tamnaba
Mapi Lairik (the Primary Book of Meetei Mayek) as a textbook for
students of Class VI.
Even though the State Government approved the twenty-seven-letter
alphabet, controversy still exists. One of the consequences of this
controversy is the following: Meeteilon was included in the Eighth
Schedule of Indian Constitution on the 20/08/1992. Since the language
became a scheduled language its script was to be included in Indian paper
currency. However, because of the ongoing controversy regarding the
competing alphabet sets it has not yet been included
2
. The State
Government could, in the end, do nothing about it. This controversy has
thus caused a delay in the development of the script as well as the
language.
The present existing twenty-seven letters of the Meetei Mayek
alphabet in order of their teaching are given in Table 2.
Many non-governmental organizations who wish to take up the cause
of Meetei Mayek have tried to develop the script as well as the language
and pursued the State Government to start teaching the script at the
primary school level. Consequently, the State Government recently took
some actions with respect to the development of the script. They released
a letter, No. 16/11/02-SE(S) dated 27/05/2005, which states that Meetei
Mayek should be taught in schools from class (standard) I and II from
the academic session 2006–07 without fail. The Government aims to
then upgrade the script by at least one class level per year. Towards this
end the Government has prepared all necessary textbooks. Many primary
teachers of the state have also been trained for this purpose. As of 2009,
the script is now being taught statewide up to class (standard) V.
In the classification of scripts provided by K. S. Singh and Manoharan
(1993: 26–29), Meetei Mayek is considered to be part of the Tibetan
group of scripts, which ultimately originate from the Gupta Brahmi script.
2
Indian banknotes use both Devanagari and Roman scripts. Additionally, the
value of the banknote is spelled out in words in all the official scripts of India,
but not, as yet, in Meetei Mayek.
The Evolution and Recent Development of the Meetei Mayek Script 29
Meetei Mayek was modified from its source to accommodate the
phonemic distinctions made in Meetei language. Like other Brahmic
scripts, Meetei Mayek is syllabic. The unified Meetei script makes use
of explicit final consonants which have no inherent vowel. Its writing
system may be considered as an alphabetic writing system. The script is
written from left to right
3
.
Among the twenty-seven letters given in Table 2, the eighteen
“original” letters are named using different parts of the human body. The
order in which they are recited also follows the organization of the human
body. These original letters are called iyek ipi. The nine subsequently
added letters are called lom iyek. These letters can occur in word-initial
positions only.
Table 2 The existing twenty-seven letters of Meetei Mayek
a
s d
kok ‘head’
səm ‘hair’ lai ‘forehead’
f
l k
mit
‘eye’ pa ‘eyelash’ na ‘ear’
j
h g
cil
‘lips’ til ‘saliva’ khəu ‘throat’
q
w e
ŋə
u ‘pharynx’ thəu ‘chest’ wai ‘navel’
r
t p
ya
ŋ ‘backbone’ huk ‘lower spine’ un ‘skin’
o
i u
i
‘blood’ phəm ‘placenta’ ətiya ‘sky’
y
z x
gok
jhəm rai
c
v b
ba
jil dil
I
m n
gh
əu dhəu bhəm
3
See also Chelliah (1997: 355) for a comprehensive description of the script.
30 North East Indian Linguistics Volume 3
In syllable-final position, seven consonants and one vowel have a
contracted form which is called lonsum iyek. These are called “unreleased”
in most descriptions of this script. They are given in Table 3.
Kok lonsum stands for syllable-final [k]; lai lonsum for syllable-final
[l]; mit lonsum for syllable-final [m]; pa lonsum for syllable-final [p]; na
lonsum for syllable-final [n]; til lonsum for syllable-final [t]; ŋəu lonsum
for syllable-final [ŋ]; and i lonsum for syllable-final [i]
4
.
The twenty-seven-letter alphabet, which is hereafter termed Unified
Meetei Script, can represent five initial vowels using just three independent
vowel characters. The remaining two vowels and syllable-initial
diphthongs are made up of an initial letter u followed by one of a distinct
Table 3 Syllable-final consonants, the lonsum iyek
A
D F L
ok lonsum lai lonsum mit lonsum pa lonsum
K
H S O
na lonsum til lonsum
ŋəu lonsum i lonsum
4
It is not clear why there needs to be a separate character for syllable-final [i]
or why none of the other possible syllable-final vowels have been given a
distinct syllable-final character. This is something which the Meetei Mayek
Expert Committee probably needs to review.
Table 4 Meitei Mayek final vowel markers
Marker Name Sound
E otn
əp [o]
Y in
əp [i]
P at
əp [a]
Q yetn
əp [e]
R
səunəp [əw]
U un
əp [u]
W čəyn
əp [əy]
T nu
ŋ [ŋ]
The Evolution and Recent Development of the Meetei Mayek Script 31
set of final vowel markers. There are eight final vowel markers – also
including one velar nasal – which are called cheitap iyek. These are listed
in Table 4.
Table 5 Meitei Mayek punctuation
V . `
əpun iyek
‘joining underline’ čəikh
əi ‘fullstop’ lum iyek
‘tone mark
’
Three punctuation marks, including one tone mark, are included in
Unified Meetei Script. These are called khudam iyek; they are listed in
Table 6 Meitei Mayek numerals
0 phul ‘zero’
1 əma ‘one’
2 əni ‘two’
3 əhum ‘three’
4 məri ‘four’
5 məŋa ‘five’
6 təruk ‘six’
7 təret ‘seven’
8 nipan ‘eight’
9 mapən ‘nine’
Table 5. All other punctuation uses standard international punctuation
marks.
The Unified Meetei Script has ten numeral symbols, called cheising
iyek. These are listed in Table 6.
Linguistically speaking, twenty-seven letters are quite enough for
reading and writing in Manipuri/Meeteilon. If one knows the script, one
can make oneself understood clearly. Many Meetei people have been
able to read and write using these twenty-seven letters for more than
twenty years now. However, there has been a recent popular interest in
32 North East Indian Linguistics Volume 3
reinstating the use of the traditional orthographic system. A Unicode font
has also been proposed which will help with computer use of the script
(Everson 2008; Qureshi 2007). If it is completed it will be very helpful
to users of Meetei Mayek.
References
Chelliah, S. (1997). A Grammar of Meithei [Mouton Grammar Library 17].
Berlin, Mouton de Gruyter.
Directorate of Census Operations, Manipur. (2001). Provisional Population Totals,
Paper 2 of 2001. New Delhi, Census of India.
Everson, M. (2008). “Proposal for Encoding the Meetei Mayek Script in the
UCS and associated Unicode Technical Reports”. Downloaded from http://
std.dkuug.dk/ JTC1/sc2/wg2/docs/n3206.pdf. Accessed 2008.
Government of Manipur. (1980). Manipur Gazette, Extraordinary. Published by
Authority, No.33 Imphal, Tuesday, 22/04/1980. (Vaisakha 2, 1902).
Grierson, G. A., Ed. (2005 [1904]). Linguistic Survey of India, Volume 3: Tibeto-
Burman Family, Part 3: Kuki-Chin and Burma Groups. New Delhi, Low
Price Publications.
Ibungohal Singh, L. and Khelchandra Singh, N. (2005). Cheitharol Kumbaba.
Imphal, Manipuri Sahitya Parishad.
Manihar Singh, C. (1984). “The Meeteis: Origin and affinity”. Journal of
Manipur University, 1(1): 9–25.
Meetei Mayek Expert Committee. (1979). Report of the Meetei Mayek Expert
Committee. Imphal, Meetei Mayek Expert Committee.
Qureshi, T. (2007). “Meetei Mayek and Unicode – Making the Script Immortal”.
Downloaded from http://e-pao.net/channel.asp?what=meiteimayek. Accessed
2008.
Sadokpam, D. (2008) “History of Meetei Mayek.” Downloaded from http://
www.docstoc.com/docs/ 5302443/The-History-Of-Meitei-Mayek. Accessed
2008.
Singh, K. S. and S. Manoharan (1993). People of India: Languages and Scripts
[National Series, Anthropological Survey of India Vol. 9]. New Delhi,
Oxford University Press.
Three Meanings of “Language” and “Dialect” in North East India 33
The Sal Group
34 North East Indian Linguistics Volume 3
34 blank
Three Meanings of “Language” and “Dialect” in North East India 35
3
Three Meanings of “Language” and “Dialect”
in North East India
1
Robbins Burling
University of Michigan
1. Introduction
In the course of a number of trips to North East India, I have come to
realize that people there often use the words “language” and “dialect”
to mean something quite different from the meanings that I usually give
them. In some cases, their usage and mine are close enough to let us
understand one another, but occasionally we run into trouble. In fact,
“language” and “dialect,” are used in three different ways in North East
India, each one emphasizing different aspects of communication. I will
describe these varying usages in the hope that we can learn to understand
each other better.
1
Most of what I have written in this article comes from many years of conver-
sations with friends in North East India, and I can no longer select out the
particular people from whom I have learned the most. I do owe a special debt
to Mark W. Post. He and I have had long discussions about the ways that
“language” and “dialect” are used in North East India. It was Mark who first
suggested that I describe these alternatives at NEILS 3 in Guwahati. I had long
been aware of the difference between the linguists’ meaning and the tribal
meaning, but it was Mark who first called my attention to the orthographic
meaning.
36 North East Indian Linguistics Volume 3
2. The linguist’s meaning
First is the linguist’s definition, and as a linguist this is the definition
that I find most useful. Linguists are usually clear that by a “language”
they mean the collection of spoken dialects that are mutually intelligible.
For example, my native speech is a dialect of North American English,
but I can understand people who speak British English and people who
speak Indian English, so all of these ways of speaking count as “dialects”
of a single language. It is the collection of all these dialects that we call
“English.” On the other hand my speech is not mutually intelligible with
Assamese, Angami, French, or Japanese, so these count as “languages”
that are different from the English “language”. When I use my native
speech, I am using a language that is different from Assamese, but I am,
at the same time, using a dialect when compared, for example, to the
English of London. It makes no sense, by this definition, to ask whether
I am speaking a “dialect” or a “language”. That depends entirely on what
my speech is being compared to.
This definition is reasonably clear to linguists and they generally try
to adhere fairly closely to it. Mutual intelligibility is a useful breaking
point because it groups people together into communities whose members
should be able to communicate with reasonable ease. Only with those
speaking other “languages” will they need an interpreter. Still, the matter
is not quite so simple. Mutual intelligibility is not an all or none matter.
How easily intelligible must two forms be to count as dialects of the
same language? Is effortless conversation required? Or is it enough to
get the gist of what someone is driving at? In fact, no sharp line for
mutual intelligibility can be found. I can have difficulty with some forms
of British or Indian English. I can generally make myself understood,
and I can generally figure out what others are saying, but communication
can be much more difficult with the English, Scots, or Indians than with
my neighbours in the United States.
The linguist’s definition faces another problem. It can happen that
dialect A is mutually intelligible with dialect B, and that B is mutually
intelligible with dialect C, even though those who speak dialects A and
C cannot understand one another. In such a case, how many languages
are there? Where do we draw the lines? The linguist’s definition of
“language” and “dialect” simply breaks down in this situation, for there
is no single way to apportion the dialects among languages, and we really
should not try to draw any lines at all.
Three Meanings of “Language” and “Dialect” in North East India 37
In spite of these problems, linguists usually manage to understand
each other reasonably well. We have a rough but useful way of
distinguishing forms of speech that are fairly similar to each other from
others that are less so. Whether or not you and I can understand each
other is important. We can relate to each other in a different way if we
understand each other’s speech than if we do not. In clear cases it is
reasonable to label similar forms “dialects” and less similar forms
“languages”.
3. The tribal meaning
Many people in North East India give a very different meaning to the
two words, a meaning that relates a “language” to the north eastern idea
of a “tribe”. A tribe is seen and talked about, in North East India, as if
is has sharp boundaries. These boundaries are presumed to separate it
unambiguously from all other tribes or ethnic groups. Each tribe is
expected to occupy a territory, usually but not always contiguous, and
its people are presumed to share their history and customs: their traditional
religion, their house style, their clothing. Above all, each tribe is presumed
to have its own “language,” and reciprocally, every “language” has its
“tribe”. Draw a map of the tribes and you will have simultaneously drawn
a map of the languages. Even the language names are generally the same
as the names of the tribes. The Angamis speak Angami; the Mizos speak
Mizo; the Garos speak Garo; and the Assamese speak Assamese.
It sometimes happens, however, that some people who count as
belonging to the same tribe cannot understand each other’s speech. It
can also happen that people belonging to different tribes can understand
one another easily. Many people of the North East whom I have known
will describe the varying forms of speech used by members of the same
tribe as “dialects” even if they are not mutually intelligible. Nothing is
strange about this, because, by the tribal definition, mutual intelligibility
is simply irrelevant to what counts as a language. As long as people
belong to the same tribe, their ways of talking are taken to be dialects
of the same language whether or not the people can understand each
other. If the map of tribes has its boundaries drawn in the same place as
the map of languages, then all the members of a tribe must speak the
same “language,” while people on opposite sides of a tribal border speak
different “languages”. For some people from the North East, that is simply
the definition of what constitutes a “language”. The forms spoken by
38 North East Indian Linguistics Volume 3
sub-tribes are called “dialects” whether or not they are mutually
intelligible.
If everyone who belongs to the same tribe could understand one
another, and if they could understand no one else without learning a
second language, then the tribal definition would correspond to the
linguist’s definition, but this correspondence is often lacking. All the
members of the Tangkhul “tribe” of north eastern Manipur, for example,
are said to speak the Tangkhul language. The Tangkhul “tribe” and the
Tangkhul “language” have the same boundary. They occupy the same
area on a map, just as they occupy the same area on the ground. However,
Tangkhuls have told me that there are at least three mutually unintelligible
varieties of Tangkhul (Mortsensen 2003). These are spoken by “sub-
tribes,” so their speech forms are called “dialects”. By the linguist’s
definition, of course, these “dialects” would be considered separate
“languages,” but Tangkhuls are rarely trained in linguistics, and according
to the “tribal definition” they are certainly “dialects,” simply because all
are spoken by members of the same tribe.
On the other hand, each of the many Kuki groups, most of which are
scattered along the western and southern boundary of Manipur, is always
considered to be a separate “tribe,” and each of these Kuki “tribes” is
always said to have its own distinct “language”. Nevertheless, Kukis
have told me that when Kukis get together they can all understand each
other’s “languages”. Even Mizos can often understand at least some of
the Kuki “languages”. Clearly, when Kukis describe their various ways
of talking as “languages” they are not using the word “language” as
linguists generally use the word. They are using the tribal definition, not
the linguistic definition.
If linguists spoke only with other linguists and North East tribals
spoke only with each other, everyone could use his own definition without
misunderstanding, but when linguists talk with people who use the tribal
definition, there can be serious misunderstandings. For example, the
diagram in Figure 1 is taken from a monograph by David Bradley that
was published by Pacific Linguistics (Bradley 1997: 1–72). Most readers
of Pacific Linguistics are, like Bradley, linguists, and they expect that
words in linguistic journals will be used according to the definitions and
habits of linguists. They also expect that the length of a line up to a fork
on a family tree of languages will be roughly proportional to the linguistic
difference between the languages. Notice, however, what a short branch
leads to Tangkhul. The various mutually unintelligible forms of Tangkhul
Three Meanings of “Language” and “Dialect” in North East India 39
are ignored completely on this diagram, all invisibly clustered together
on the short stem that is labelled “Tangkhul”. Any linguistic reader would
assume that all varieties of Tangkhul are mutually intelligible.
Ao
Sangt
am (N)
Lotha
Yimchungrü
Ntenyi/Meluri
Tangkhul
Maring
Sema
Angami (W)
Chakhesang
Chokri
Khenzhama
Mao
Rengma (W)
Maram
Zeliangrong
Mzieme
Zeme
Liangmai
Puiron
Nruanghmei
Meithei
Rangkhol
Bete
Hallam
Langrong
Hmar
Anal
Kom
Chawte
Mayol
Lamgang
Other Old Kuki
(various)
Lai(zo)(various)
Mizo
Ashö
Khami/Khumi
Mara
Arleng (Karbi, Mikir)
‘Old Kuki’
‘Old Kuki
’
Central Chin
South Chin
Figure 1 Classification by Bradley (1997: 1–72)
40 North East Indian Linguistics Volume 3
The branches leading to the various forms of Kuki are much longer
than the stubby twig that leads to Tangkhul. The fork that separates Kuki
from Mizo is close to the top of the tree. A linguist who knows little of
North East India will mistakenly conclude that the difference between
Mizo and the Kuki languages is much greater than any small difference
among the forms of Tangkhul. Bradley must have taken his information
from writers whose usage reflected the tribal definitions, for by this
definition all forms of Tangkhul count as one “language,” while Mizo,
and all the different forms of Kuki, count as different “languages”.
Bradley, of course, is simply summarizing the available descriptions of
these languages, and he is hardly to be blamed for failing to realize that
the people of North East India use these words so differently than linguists
do. For a linguistic audience, however, the diagram is seriously
misleading.
Consider, also, the case of the Garos (see Figure 2). For the most
part, the boundaries of the Garo “tribe” are clear. Who is a Garo and
who is not a Garo, is rarely in dispute, and all Garos are said to speak
one or another “dialect” of the Garo “language”. These “dialects” include
A’beng, A’we, Matchi, Dual, Ruga, Atong and some others. Most of
these “dialects” are easily mutually intelligible, but Ruga and Atong are
not mutually intelligible with the others. Ruga is now almost, if not
totally lost, spoken at most by three or four of elderly people. Atong is
more viable, though probably shrinking gradually as its speakers shift to
Garo. No Garo ever questions the fact that the Atongs are Garos, and no
one doubts that Atong is a “dialect” of Garo. Most Garos, certainly all
the Atongs themselves, know very well that the Atong way of speaking
is not mutually intelligible with most dialects of Garo, but since it is
shared tribal membership rather than mutual intelligibility that makes
something a “dialect,” mutual intelligibility is irrelevant
2
. A number of
communities that speak one of the Koch “languages” are found along
the western and northern borders of the Garo Hills. The Koch are not
regarded as Garos, so the Koch are never considered to speak a dialect
of Garo. Nevertheless, their “languages” are closer to Ruga and Atong
2
The Garos tell a lovely story about the origin of the word “Atong”. In the
Atong language, “Atong” is a question word meaning “what?” According to
the story, when other Garos spoke, in ordinary Garo, to Atong speakers, the
Atongs, not understanding, would reply Atong? Atong? “What? What?” The
Garos, thinking that they were hearing an explanation of why the others had
not understood, took that to be their sub-tribal name.
Three Meanings of “Language” and “Dialect” in North East India 41
(and also to Rabha) than any of these are to A’we or A’beng or to any
of the other core dialects of Garo. Again it is tribal membership that
matters. Degree of linguistic relatedness is irrelevant.
To a linguist, a still more anomalous case is the status of a people
who live in the westernmost part of the Khasi hills, right along the border
with the Garo Hills. These people are known as “Lyngngams” to the
Khasis but as “Megams” to the Garos. The Garos consider the Megams
to be a subtribe of the Garos and describe their speech as a “dialect” of
Garo. The Khasis, however, consider the Lyngngams to be a subtribe of
the Khasis, and they consider Lyngngam to be a dialect of Khasi. As far
as I know, no one has yet bothered to ask the Lyngngam-Megams what
they consider themselves to be, but on this point, a linguist would have
to agree with the Khasis. The speech of Lyngngam-Megams is clearly
allied linguistically to Khasi rather than to Garo. By the linguist’s
definition, Lyngngam is either a dialect of Khasi or a closely related
language. Like all forms of Khasi, Lyngngam-Megam is not even Tibeto-
Burman, and so it is not related, even distantly, to Garo. Megam-
Lyngngam belongs, instead, to the Mon-Khmer language family. To a
linguist it seems absurd to call the speech of the Lyngngam-Megams a
dialect of Garo, but if the Megam-Lyngngams are Garos, then by the
tribal definition they must certainly speak a dialect of Garo.
Figure 2 Classification by the “Linguist’s” and the “Tribal” definition
Tibeto-Burman
Boro-Garo
Rabha Koch
Ruga Atong
A’we A’beng
Matchi etc.
Megam-
Lyngngam
Jaintia Wa
r etc.
Garo (Tribal Definition)
Garo (Linguist’s
Defonition
Khasi
Mon-Khmer
42 North East Indian Linguistics Volume 3
4. The orthographic meaning
For many people, both in the west and in North East India, it is neither
mutual intelligibility nor tribal membership that defines a “language”.
Instead, it is writing. According to the “orthographic” meaning, the only
speech that counts as a “language” is speech that has a written form.
Unwritten forms of speech are “dialects” and an “unwritten language”
is a contradiction in terms.
Even in the west, those who are not linguists often use “language”
for the collection of “dialects” that are spoken by everyone who uses the
same writing system. For a language like English this corresponds
reasonably closely to the linguist’s definition of mutual intelligibility.
All the people who speak mutually intelligible dialects of English also
use the same written conventions, or at least conventions that differ only
trivially. Reciprocally, all those who use the same English written
conventions can generally understand each other’s spoken languages
without too much difficulty. The same correspondence between writing
and mutual intelligibility is also characteristic of many other European
languages, but there are exceptions. Swedes and Norwegians understand
each other easily. They no more bother to learn each other’s “languages”
than do British and Americans bother to learn each other’s “dialects”.
Because the Swedes and Norwegians use different written conventions,
“Swedish” and Norwegian” are usually described as distinct “languages”.
There are even two different ways of writing Norwegian. Norwegians
have no problem understanding one another’s speech, but they can choose
between two rather different ways of transcribing their speech on paper.
The two written forms are often described as different “languages”.
The “languages” of the Croatians and the Serbs are also mutually
intelligible, but Croatians use the Roman alphabet while Serbs, like the
Russians, use Cyrillic. The difference in writing is enough to make most
people count Serbian and Croatian as separate “languages”. Hindi is
written with the Devanagri alphabet and Urdu with the Arabic alphabet,
so they, too, are spoken of as separate “languages” even though most
forms of spoken Hindi and spoken Urdu are mutually intelligible.
In China, the situation is reversed. The Chinese have several mutually
unintelligible spoken languages − Mandarin, Hakka, Cantonese and others
− but they all use the same written language. The spoken forms of Chinese
are almost always described as “dialects” because their speakers use the
same writing system. In popular usage, they are said to be “dialects” of
Three Meanings of “Language” and “Dialect” in North East India 43
the Chinese “language” even though they are by no means mutually
intelligible. Even linguists can forget their own definitions. We may call
Norwegian and Swedish “languages” and Mandarin and Cantonese
“dialects,” and thereby obscure the reality that spoken Mandarin and
spoken Cantonese are considerably more different from each other than
are the spoken forms of Norwegian and Swedish.
If we are able to count the “languages” of Europe it is only because
we can count the large but still finite number of writing systems that
Europeans use, and because we ignore the very much larger number of
“dialects”. Most “language” maps of Europe (or of the world, for that
matter) are actually maps of the writing systems used, not maps of the
spoken languages. For example, the dialects of Dutch and German spoken
on opposite sides of political border are mutually intelligible, but most
writing on one side is done in Dutch and on the other side in German.
That allows the “language” map to show a sharp boundary at the border,
even though this fails to reflect the way people talk.
The connection between “language” and writing also appears when
“dialect” is used to mean a form of speech that is not written at all.
Americans have sometimes asked me: ‘Did you learn any of the “dialects”
while you were in India?’ By “dialect” they mean a form of speech that
is local, spoken by a relatively small number of people and, crucially,
not written. By this definition, unwritten ways of speaking are “dialects”.
Only when they gain a conventional writing system do they deserve to
be called “languages”. For many people, it seems, written “languages”
are more important than mere spoken “dialects”. Such judgments seem
prejudicial, and even wrong, to most linguists, for they violate the deep
respect that we cultivate for the spoken languages of the people from
whom we learn. Linguists have usually regarded writing as secondary to
the spoken language. After all, speech came first in human history just
as it came first to each of us in childhood. Much of the world, however,
sees writing as the “real” language while speech seems fleeting and
insubstantial.
The orthographic definition of “language” and “dialect” is reflected
in North East India by an intense concern with alphabets and writing.
Unlike Western Europe, where all languages are written with closely
related forms of the Roman alphabet, the people of North East India
sometimes feel that, in order to count as a serious “language,” a form of
speech should have its own unique alphabet. Languages that seem
important, such as English, Hindi, Urdu, and Bengali all have their own,
44 North East Indian Linguistics Volume 3
quite different, alphabets. To an outsider, the differences between the
Bengali and Assamese alphabets seem trivial, but the differences take
on enormous importance to the Assamese because they mark their
“language” as different from Bengali. The focus on writing makes
understandable the ferocity of disputes among speakers of some
north eastern communities over which alphabet to use when they write
their language. It also makes understandable the popularity of stories
about how the ancient alphabets of tribal people were lost − eaten by a
dog or dropped over the side of a boat as the people crossed the
Brahmaputra river. To believe that they once had an alphabet implies an
importance for the language that it would otherwise not have.
Conclusion
The Indian Constitution recognizes certain languages as having official
status. The Eighth Schedule to the Constitution confers official status on
22 languages, four of which are spoken by substantial numbers of people
in the North East: Assamese, Bengali, Meitei and Bodo. The last of these
was added only in 2004, as one provision of the ninety-second amendment
to the Constitution
3
. Official status is supposed to bring official support
for these languages, and a Government Resolution, as adopted by both
Houses of Parliament in 1968, specified that “. . . a programme shall be
prepared . . . for the development of all these [officially recognized]
languages, alongside Hindi so that they grow rapidly in richness and
become effective means of communicating modern knowledge . . .
4
” In
addition to the nationally recognized official languages, the states can
give official status to other language that are used within the state’s
borders. Table 1 lists the languages that are recognized as official in the
seven states of North East India
5
. Constitutional recognition has great
symbolic significance. Its actual pragmatic effect in daily life is not so
clear.
For North East India, the most important consequence of the
orthographic definition is that people sometimes dismiss forms of speech
that are not written as mere “dialects,” but dignify forms that are written
3
Constitution of India: http://indiacode.nic.in/coiweb/welcome.html
4
The Official Language Resolution, 1968: http://www.rajbhasha.gov.in/dol-
resolutioneng.htm
5
Official languages of India: http://en.wikipedia.org/wiki/Official_languages_
of_India
Three Meanings of “Language” and “Dialect” in North East India 45
as “languages”. In India, people who speak a “dialect” may even be in
danger of losing the legal protection that the Constitution gives to
“languages”. Most linguists would like to encourage all forms of speech,
even those that are not written, to be treated as honourable ways to talk,
but we need to be clear about what we and others mean by the words
“language” and “dialect”. When we are talking or writing for other
linguists, we should be able to use the linguist’s definition with confidence.
If we are listening to advocates with a political agenda, we should never
forget that they are almost certainly using the words in some other way
than we usually do. If we want to support the linguistic minorities of
North East India, we should, whenever we have the chance, insist that
the way people talk is important, and that each person’s way of talking
deserves respect, whether or not it is written. In principle, whether
something is called a “language” or a “dialect” should not matter. But
it matters a great deal if, by having their language dismissed as a “dialect,”
people are deprived of their rights to use their own form of speech.
References
Bradley, D. (1997). “Tibeto-Burman languages and classification” In David
Bradley, Ed., Papers in Southeast Asian Linguistics No 14: Tibeto-Burman
Languages of the Himalayas. Canberra, Pacific Linguistics: 1–72.
Mortensen, D. (2003). “Comparative Tanghkul.” Unpublished Manuscript.
Table 1 Languages officially recognized by states
State Official Language Other Officially
Recognized Languages
Arunachal Pradesh English, Assamese None
Assam Assamese Bengali, Boro
Manipur Meitei None
Meghalaya English Khasi, Garo
Mizoram Mizo None
Nagaland English None
Tripura
English, Bengali, Kokborok None
46 North East Indian Linguistics Volume 3
4
An Initial Reconstruction of the Proto-Boro-Garo
Noun Phrase
1
Daniel Wood
University of Oregon
1. Introduction
The Boro-Garo (BG) sub-branch of the Tibeto-Burman (TB) language
family consists of thirteen languages found in North East India, as well
as across the border in Nepal and Bangladesh; these languages include
Boro, Deuri, Dimasa, Tiwa, Riang, Kokborok, Kachari, Garo, Megam,
A’tong, Koch, Rabha and Ruga. Though much work remains to be done
on all the modern languages in the family, this paper attempts an initial
reconstruction of the Proto-Boro-Garo (PBG) noun phrase (NP), as well
as some NP categories.
2. Reconstruction of some grammatical NP elements
The sound change patterns from PBG to the modern languages have been
adopted from Joseph and Burling (2006), and are used as the basis for
the reconstruction. Of the thirteen languages representing the Boro-Garo
family, only Boro, Garo, Deuri, Rabha, Dimasa and Kokborok will be
examined given the availability of language data. In the instances where
1
I would like to thank Scott DeLancey, attendees of the 3
rd
International Con-
ference of the North East Indian Linguistics Society, as well as the reviewers
of this paper, whose input has greatly helped improve it. All mistakes, how
-
ever, remain my own.
An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase 47
data from other BG languages were available, they have been used to
aid in the reconstruction. Phonetic transcriptions, specifically those of
vowels and tones, have been regularized among the languages following
Joseph and Burling’s (2006) phonological reconstruction. High tones are
marked with ´, while low tones are unmarked. A ‘~’ over a vowel
indicates nasalization.
2.1. The PBG pronouns: Personal PNs and demonstratives
The modern Boro-Garo languages are said to have a relatively shallow
time depth, meaning their split from the proto-language occurred more
recently in comparison to other Tibeto-Burman languages (Burling 2006).
As a consequence, we might expect to find a clear correlation between
pronominal forms across the languages investigated. However, it is not
uncommon for TB languages to innovate new pronominal forms, even
within relatively short periods of time (Thurgood 1985). We find some
evidence of this in the modern BG languages, as not all the person/number
distinctions made are consistently found across the family, nor are all
the forms cognate (Table 1).
Table 1 demonstrates the closeness in form for the 1
S, 2S and 1P
pronouns, which easily reconstruct to the original PBG pronominal forms.
From this, we are left to assume that some of the 3S pronominal forms,
either u or bV, have been innovated since the split from PBG, and we
must account for this variation and determine which is the older form.
Table 1 Pronoun forms across Boro-Garo
Garo Boro Deuri Dimasa Kokborok Rabha *PBG
1S aŋ(a) an ã aŋ aŋ aŋ *aŋ
2S naʔ(a) nən nõ níŋ nɯŋ naŋ *naŋ
3S bi(a) bí la bó bɔ u
ua ba o
1P naʔchiŋ zən jou jiŋ cɯŋ ciŋ *jaŋ
anchiŋ jeŋ
2P naʔsoŋ nənsər lou nísí nɔrɔk
3P bisoŋ bísər bau bósí
uamaŋ bíphər lau bon-si
48 North East Indian Linguistics Volume 3
The demonstrative system, though typically a reliable etymological source
for 3S referents cross-linguistically, does not provide strong evidence of
innovation with respect to either form (Table 2).
Table 2 Modern Boro-Garo demonstratives
Garo Boro Deuri Dimasa Kokborok Rabha
‘this’ (h)i-(a) be la ebo i i
‘that’ (h)u-(a) bí ba bo u u
The demonstrative pronouns in Table 2 evidence a similar disparity
in forms across the modern languages as both u and bV forms are found.
One possible etymological source for the BG demonstratives could be
found in the Tibetan demonstrative system. Franke writes:
The demonstrative pronoun is often replaced or strengthened by the
optional article bo, which may be added to any noun, furnished with
an article (sexual determination) or not… khyi-bo ‘that dog’, myi-bo
‘that man’. (Jäschke et al. 1929: 112)
The close similarity between BG bV demonstratives and Tibetan -bo,
which both seem to indicate definiteness of the noun referent, suggests
a similar historical connection. However, there is also evidence elsewhere
in TB for *u ‘3
S’ and *i ‘this’ vs. *u ‘that’, respectively. Benedict (1983)
reconstructs *(h)i ‘this’ and *(h)əw ‘that’ in Western Bodic, which are
clearly cognate to the u and i demonstratives found in the modern BG
system – especially the Garo allomorphs, hia and hua, respectively. Other
evidence supporting a reconstructible *u ‘3S’ can be found in the PTB
transitive verbal suffixal paradigm (Table 3).
Table 3 The PTB transitive verbal suffixal paradigm (from DeLancey 1988)
Undergoer
→
Actor
↓ 1
st
2
nd
3
rd
1
st
(-na/-ŋ) -ŋ
2
nd
-ŋ (-na)
3
rd
-ŋ -na [-u]
An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase 49
It is widely accepted that the -ŋ ‘1S’ and -na ‘2S’ suffixes developed
from the PTB 1S and 2S pronouns *ŋa and *na(ŋ). Following this pattern,
it is a logical step to assume that the 3>3 index has a diachronic
connection to a pronominal *u. Similar cognate forms can be found in
the verbal suffixal paradigms of other modern TB languages (DeLancey
1988).
Thus we have evidence that BG u does have some diachronic
connection to either a deictic demonstrative or 3S pronominal element
as we find it occurring across multiple branches of the Tibeto-Burman
family. Therefore, u must date back to a time preceding the split of PBG
into the modern daughter languages, while the bV forms evidence some
level of innovation within the class. The more recently innovated forms
have not replaced the PBG forms in all of the BG languages, however.
Of particular interest are instances where the two sets of 3S pronouns
occur interchangeably, which can be found in Garo.
Burling (1961: 39) writes that Garo speakers alternate between ua
and bia for the distal demonstrative, commonly relying on the ua form
for written language, whereas bia is used orally. This example could
provide even further evidence that ua represents the older 3S pronoun in
BG.
It is worth noting that neighboring Indo-Aryan languages have similar
demonstrative forms with Indic sources (Table 4; cf. Jain and Cardona
2003).
Table 4 Indo-Aryan demonstratives
Bengali Bhojpuri Assamese
‘this’ ea i i
ehi hai iye
ini eye
‘that’ o u hi
ui hau hiye
uni unhi
heye
hunhi
We cannot automatically interpret this as evidence that the BG forms
are borrowed, given the evidence for their TB provenance. However, it
is certainly possible that the existence of similar forms in these
neighbouring languages has supported their preservation.
50 North East Indian Linguistics Volume 3
2.2. Numeral classifiers
The modern BG languages employ a numeral classifier system (CL)
which was clearly part of the PBG grammar. With so many varied forms
suggesting recent incorporation into the system, we are able to reduce
the PBG classifier system to its most basic distinctions, which
grammatically encode animacy, human-ness, and shape of the referent.
Table 5 Cognate classifiers across BG and reconstructed PBG forms
Garo Boro Rabha Deuri Kokborok Dimasa PBG
Humans sak- sá- sak- dug(u)- khɔrɔk sao *sak
ak- kh
á- ka-
Animals maŋ- má- maŋ- ma mâ *maŋ
1D OBJ diŋ- thó- tɯŋ- tuŋ- *tɯŋ
2D OBJ kiŋ- gán- khó- ha- kaŋ garáŋ *khó
3D OBJ roŋ- tháy- ti- thay- thai- *tai
Residual kan- gó- gu- kay- *gV
category
geʔ- ku-
While Joseph and Burling (2006) provide reconstructed forms for
these elements, they offer no explanation of how the system has changed
since PBG. For example, modern Garo has replaced the PBG three-
dimensional object classifier *tai with a more recently innovated form
roŋ-, which is derived from the noun roŋ ‘stone’. Other languages have
simplified the PBG classifier system such that Boro has lost the residual
category, Rabha the 3D object category, Deuri the animal and 1D object
categories while Dimasa has lost the 1D object and residual categories
(Wood 2008).
2.3. Relator nouns
The modern BG languages also utilize a system of relator nouns (RN)
to indicate the location of a referent in terms of the location of the
ground.
An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase 51
(1) Garo
2
naʔa sa-ni kisaŋ-o
you who-GEN behind-LOC
chadeŋ-a-miŋ
stand-NEUT-PST
‘Who did you stand behind?’ (Burling 2004: 221)
Typologically similar to many Tibeto-Burman languages, (1) demonstrates
how RN constructions in BG function as full NPs (DeLancey 1997). The
two arguments in this sentence are naʔa ‘you’ and genitive-inflected
sani ‘who’, while kisaŋo ‘behind’, derived from the noun kisaŋ ‘buttocks’
(Burling 2004: 230) takes the locative marker -o, and functions as a
modifier of sani. In doing so, this locative-inflected noun effectively
describes the spatial relationship between the two referents.
Other modern BG languages exhibit similar syntactic patterns, in spite
of the fact that examples of relator noun constructions from these
languages are not readily available. Thus, we cannot justifiably describe
the PBG relator noun system in terms of the actual relator nouns used,
though we can comfortably claim, based on the syntactic similarity of
these constructions across BG, that PBG also relied on RN constructions
to frame spatial relationships between figure and ground.
2.4. Case-endings
Using the same methodology applied in the reconstruction of the PBG
classifiers, those case-forms that are consistent across the modern
languages will be assumed to represent the case-endings in PBG. The
discussion of case-endings will only involve those forms exhibiting
variation across the family, thus requiring that we determine which have
been more recently innovated. There are more case distinctions made
across the modern BG languages than are presented in Table 6, but with
such varied forms and unclear etymologies that we must assume that
although their functional categories very well existed in PBG, they do
not necessarily represent the oldest forms.
2
Garo examples herein have been altered from Burling (2004) in order to main-
tain consistency across the languages investigated. Word-parsing is based on
my own analysis of the lexical and grammatical elements and may contrast
with Burling’s (2004) division of the elements, which was often based on syl
-
lable structure, whereas mine are purely based on morpheme boundaries.
52 North East Indian Linguistics Volume 3
2.4.1. Nominative case
The data seem to suggest there could have been an overt nominative case
suffix in PBG, despite the fact that so few of the modern languages have
retained this behaviour. Evidence for the existence of this case-marker
in PBG can be found in Boro, where this suffix consistently surfaces on
S and A arguments. In terms of Garo, a fossilized form of the PBG
nominative is found on monosyllabic pronouns filling S and A grammatical
roles (Burling 2004; Joseph 2006). When other case-inflections are
suffixed onto these pronouns, the fossilized -a is lost. Meanwhile, the
other modern languages have simply lost this particular case-ending.
2.4.2. Accusative
Modern BG demonstrates some variation in form for accusative case-
marking across the family. Some of the modern languages, exemplified
here by Deuri and Kokborok, rely upon a “primary object” pattern in
which the same form is used for dative arguments and for animate or
definite direct objects. Conversely, we find that Garo, Boro, Rabha and
Dimasa have a cognate form for the accusative case, -k(h)V, which is
quite distinct from the dative declension. A similar form,
-(ɔ)k, which
also exhibits preferential object marking is found in neighbouring
Assamese (Masica 1991: 230–39). This suggests that the k- initial forms
in BG could have been borrowed.
Table 6 Case-ending suffixes across Boro-Garo
Garo Boro Deuri Kokborok Dimasa *PBG
NOM -Ø -Ø -Ø -Ø -Ø *-a
-a -a
ACC -ko -khow -khɔ *-kho
-khe
DAT -na -nə -na -nɔ -ne *-na
GEN -ni -ni -o -ni -ni *-ni
ABL -ni -níphray -capi -ni -niphəraŋ *-ni
LOC -o -ha -(h)o -ɔ -ha *-ao
INSTR -cha -zən -coŋ -bay -jaŋ *-caŋ
An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase 53
Investigation of other languages within the greater Boro-Konyak-
Jinghpaw family evidences a different form for the accusative. We find
-hpe in both Jinghpaw (Hanson 1917: 11) and Assam Singpho (Morey
2011), which functions as “primary object marking” in both languages.
The fact that it is not uncommon for other languages within the greater
Boro-Konyak-Jinghpaw to utilize primary object-marking could suggest
that PBG also relied on this pattern of case-marking, in which case the
k- initial morphemes would therefore represent more recent innovations.
However, since we do find similar k- initial morphemes across multiple
branches within BG, it is plausible that this was the PBG accusative
marker. The assumption could then be made that those languages using
primary object-marking have merely simplified the PBG system and rely
upon primary object-marking. The accusative must therefore be
reconstructed as *kho.
2.4.3. Ablative
With the exception of Deuri
-capi, the ablative forms across the modern
BG languages are either identical to or retain a remnant of the genitive
declensions. Boro -níphray and Dimasa -niphəraŋ can be explained by
the fact that -phray/-phəraŋ are both instances of borrowing of the
Assamese postposition para ‘from’. Given the functional and semantic
similarities between the genitive and ablative cases, as well as the ubiquity
of these particular cases sharing the same form cross-linguistically, we
may assume that the genitive and ablative case-markers shared the same
form in PBG.
2.4.4. Locative
The reconstructed PBG *-ao ‘
LOC’ is taken from Joseph and Burling
(2006). An interesting feature of the locative suffix in Boro, Kokborok
and Dimasa is the initial h-, and how similar these elements appear
compared to the Garo deictic demonstratives, (h)i-a ‘this’ and (h)u-a
‘that’, seen above in Table 2 (§2.1). This could imply some historical
connection between the demonstratives and locative suffix.
3. Syntactic description and reconstruction
In the following section, I will reconstruct the order of elements in the
PBG NP.
54 North East Indian Linguistics Volume 3
3.1. The PBG NP: Order of the elements
3.1.1. Possession
Across the five languages, genitive-inflected possessor consistently
precedes the possessed. From this we can assume that the same was true
for PBG.
(2) Garo
chiŋ-ni achak-piʔsa-ba
1P-GEN dog-DIM-also
‘our puppy also’ (Burling 2004: 167; author’s own gloss)
(3) Boro
án-ní bízáb-khow how
1
S-GEN book-ACC send
‘Send my book.’ (Bhattacharya 1977: 149)
(4) Deuri
ã ã-yõ now-o dũ-ĩ
1S 1S-GEN boat-LOC is-NEUT
‘I am in my boat.’ (Jacquesson 2005: 231)
(5) Kokborok
Ram-ni uskul hakcal-o
Ram-GEN school far-LOC
‘Ram’s school is far away.’ (Pai 1976: 58)
(6) Dimasa
níŋ-ni no baraha
2S-GEN house where.is
‘Where is your house?’ (Dundas 1908: 20)
*PBG: POSS N
3.1.2. Demonstratives and the nouns they qualify
As in possession, demonstratives precede the nouns they qualify. This
suggests that the syntactic pattern in Proto-Boro-Garo was the same.
However, from the multiple sources used for this reconstruction, it has
not yet been possible to find an example in which both a
GEN possessor
and a demonstrative pronoun surface.
Given robust typological patterns
around the world, it does seem highly likely that the word order would
be
*DEM POSS N if both these elements were to occur in a clause. This
is, however, pure speculation given the lack of evidence available.
An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase 55
(7) Garo
[ia paŋ-ko] daraŋ-ba
[this tree-ACC] anyone-IND
denʔ-nabe ia -de aŋ-ni-sa
cut-NEG.IMP this-but 1S-GEN-only
‘Don’t anyone cut [this tree]. It is mine.’
(Burling 2004: 217)
(8) Boro
[bé báthrá-khow] bun
[this information-ACC] speak
‘Speak this information.’ (Bhattacharya 1977: 122)
(9) Deuri
[hela iskul-a ] atay-o-ne
[that school-TH] all-GEN-FOC
‘That school is public.’ (Jacquesson 2005: 257)
(10) Kokborok
[i manɯy-rɔk] tub-ɔy thaŋ-di
[this thing-many] take-PROG go-IMP
‘Go taking these things.’ (Pai 1976: 32)
(11) Dimasa
[ebo áncha] hámbi
[that child] good
‘That child is good.’ (Dundas 1908: 5)
*PBG: DEM POSS N
3.1.3. Plural marking
Plural inflection is suffixed directly onto the noun it modifies in the
modern languages. It precedes any case-marking that may be relevant to
the noun’s grammatical role in the sentence consistently, and therefore
this order is reconstructed to PBG.
(12) Garo
mande-daŋ-ni haʔ-ba-soŋ-ba
person-PL-GEN fields-also-village-also
baŋʔ-a-miŋ
much-NEUT-PST
‘The Mandis’ property (fields and villages) were plentiful.’
(Burling 2004: 180)
56 North East Indian Linguistics Volume 3
(13) Boro
má-phr-khow nan-gow
what-PL-ACC need-DES
‘What are needed?’ (Bhattacharya 1977: 123)
(14) Deuri
jimosaya mira-mirasi-laju-wa cagu-ho
Deuri boy-young-PL-TH route-LOC
kirobo-ho jeŋ-laju-na gusa-i
go-LOC obstacle-PL-ACC remove
‘The young Deuris remove the obstacles when they go on the
road.’ (Jacquesson 2005: 280)
(15) Kokborok
bɯ-rɯy-chikla-rɔg-nɔ rɔhɔr-di
3S-girl-young-many-ACC send-IMP
‘Send the young girls.’ (Pai 1976: 54)
(16) Dimasa
subúng-nishi-ni
man-PL-GEN
‘of the men/men’s’ (Dundas 1908: 3)
*PBG: DEM POSS N-PL
3.1.4. Adjectives
Adjectives, viewed strictly in terms of their functioning as modifying
elements within the NP, vary across the languages with respect to their
word order, occurring either before or after the qualified noun. Since this
pattern is found across BG, we may assume that this was also the case
in PBG.
(17) Garo
a. dalʔ-gipa achak
big-NMZ dog
‘big dog’
b. achak dalʔ-gipa
dog big-NMZ
‘big dog’ (Burling 2004: 133)
(18) Deuri
dẽ popõ
tall tree
‘a tall tree’ (Jacquesson 2005: 111)
An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase 57
(19) Kokborok
wag-jəla-kɔsɔm-nɔ
pig-male-black-DAT
‘to the black male pig’ (Pai 1976: 92)
(20) Dimasa
ácha hámba
child good
‘good child’ (Dundas 1908: 5)
*PBG: DEM POSS {ADJ} N-PL {ADJ}
3.1.5. Classifiers and numerals
For the most part, classifiers and numerals follow the noun they modify,
though there are some instances in some languages where they can
precede the nouns.
(21) Garo
a. te
ʔrik roŋ-sa
banana CL-one
‘one banana’ (Burling 2004: 248)
b. kan-sa kodam-ko-in aŋ-na raʔ-ba-bo
CL-one pillow-ACC-FRG 1S-DAT bring-IMP
‘Bring me a pillow.’ (Burling 2004: 185)
(22) Boro
msów-ha áthín thən-brəy don
cow-LOC leg CL-four exist
‘A cow has four legs.’ (Bhattacharya 1977: 142)
(23) Deuri
a. ã-yõ gu-ja gumo dũ-i
1S-DAT CL-one hat is
‘I have one hat.’ (Jacquesson 2005: 262)
b. motosaikl gu-ja ko-ri
motorcycle CL-one come
‘A motorcycle came.’ (Jacquesson 2005: 262)
(24) Kokborok
than-sa cempay bəthay-gənaŋ naŋ-gɔ
CL-one basket fruit-POSS need-PRES
kaŋ-cha japa naŋ-gɔ
CL-one basket need-PRES
‘A basket full of fruits and a japa (basket) are needed.’
(Pai 1976: 130)
58 North East Indian Linguistics Volume 3
(25) Dimasa
botol thai gini
bottle CL two
‘two bottles’ (Dundas 1908: 7)
Examples (22)
–(25) demonstrate how there is some amount of freedom
for the classifier-numeral morphemes to either precede or follow the
qualified noun, at least when the classifier occurs in conjunction with the
number one. While we may assume a similar syntactic patterning in PBG,
further investigation is required to confirm whether the same syntactic
freedom applies to the other numerals.
*PBG: DEM POSS {CL-NUM ADJ} N-PL {ADJ CL-NUM}
3.1.6. Case inflections and relator nouns
Case inflection consistently cliticizes onto the end of the NP, and is
preceded by the RN. This behavior suggests that this too was the syntactic
restriction for these elements in PBG.
*PBG: DEM POSS {CL-NUM ADJ} N-PL {ADJ CL-NUM} RN =CASE
4. Conclusion
Given the available data, we find that the modern languages do not differ
greatly in their syntactic ordering of NP elements nor in the grammatical
forms. This is not terribly surprising, given that this branch of Tibeto-
Burman has a relatively shallow time-depth. Continued efforts to
document and describe the modern BG languages will only further
contribute to the overall thoroughness of these reconstructions.
Abbreviations
1D obj One-dimensional object
2D obj Two-dimensional object
3D obj Three-dimensional object
ABL Ablative
ACC Accusative
ADJ Adjective
CL Classifier
An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase 59
DAT Dative
DEM Demonstrative
DES Desiderative
DIM Diminutive
FEM Feminine
GEN Genitive
IMP Imperative
IND Indicative
INSTR Instrument
LOC Locative
MASC Masculine
N Noun
NEG.IMP Negative Imperative
NEUT Neutral Tense
NMZ Nominalizer
NOM Nominative
NUM Numeral
PL Plural
POSS Possessive
PRES Present
PROG Progressive
PST Past
S Singular
TH Theme
References
Benedict, P. (1983). “This and that in TB/ST.” Linguistics of the Tibeto-Burman
Area. 7(2): 75–97.
Bhattacharya, P. C. (1977). A Descriptive Analysis of the Boro Language.
Gauhati, Gauhati University Department of Publications.
Burling, R. (1961). A Garo Grammar. [Deccan College Monograph Series No.
25.] Poona, Deccan College Postgraduate and Research Institute.
_____. (2004). The Language of the Modhupur Mandi (Garo), Vol. I: Grammar.
New Delhi, Bibliophile South Asia.
_____. (2006). “The Lingua Franca cycle and the map of Tibeto-Burman
languages.” Paper presented at the 39th International Conference on Sino-
Tibetan Languages and Linguistics, Seattle WA, September 16.
DeLancey, S. (1984). “Etymological notes on Tibeto-Burman case particles.”
Linguistics of the Tibeto-Burman Area 8(l): 59–77.
60 North East Indian Linguistics Volume 3
_____. (1988). “On the evolution of the Kham agreement paradigm.” Linguistics
of the Tibeto-Burman Area, 11(2): 51–61.
_____. (1997). “Grammaticalization and the gradience of categories: Relator
nouns and postpositions in Tibetan and Burmese.” In J. L. Bybee, J. Haiman
and S. A. Thompson, Eds., Essays on Language Function and Language
Type Dedicated to Talmy Givón. Amsterdam/Philadelphia, John Benjamins:
51–69.
Dundas, W. C. M. (1908). Outline Grammar and Dictionary of the Kachari
(Dimasa) Language. Shillong, Eastern Bengal and Assam Secretariat Press.
Hanson, Rev. O. (1917). A Grammar of the Kachin or Jinghpaw Language.
Rangoon, American Baptist Mission Press.
Jacquesson, F. (2005). Le Deuri: Langue Tibéto-Birmane d’Assam. Paris,
Peeters.
Jain, D. and G. Cardona. (2003). The Indo-Aryan Languages. London,
Routledge.
Jäschke, H. A., A. H. Francke and W. Simon. (1929). Tibetan Grammar by
Jascke; Addenda by A. H. Francke; Assisted by W. Simon. Berlin, Walter de
Gruyter.
Joseph, U. V. and R. Burling. (2006). The Comparative Phonology of the Boro-
Garo Languages. Mysore, Central Institute of Indian Languages.
Masica, C. P. (1991). The Indo-Aryan Languages. Cambridge, Cambridge
University Press.
Morey, S. (forthcoming 2010). “Nominalization in Numhpuk Singpho.” In F.H.
Yap and J. Wrona, Eds., Nominalization in Asian Languages: Diachronic
and Typological Perspectives. Volume 1: Sino-Tibetan and Indo-Iranian
Languages. [Typological Studies in Language Series.] Amsterdam/
Philadelphia, John Benjamins.
Pai, P. (1976). Kokborok Grammar. [CIIL Grammar Series No. 3]. Mysore,
Central Institute of Indian Languages.
Thurgood, G. (1985). “Pronouns, verb agreement systems and the subgrouping
of Tibeto-Burman.” In G. Thurgood, J. A. Matisoff and D. Bradley, Eds.,
Linguistics of the Sino-Tibetan Area: The State of the Art: Papers Presented
to Paul K. Benedict on his 71
st
Birthday. Canberra, Dept. of Linguistics,
Research School of Pacific Studies, Australian National University: 376–
400.
Wood, D. (2008). An Initial Reconstruction of Proto-Boro-Garo. MA Thesis.
Department of Linguistics, University of Oregon.
Nocte and Jinghpaw: Morphological Correspondences 61
5
Nocte and Jinghpaw: Morphological
Correspondences
1
Scott DeLancey
University of Oregon
1. Introduction
The genetic relationships among the Tibeto-Burman languages of eastern
India and western Burma have always been problematic. While several
linguists, including myself, have made stabs at sorting the problem out
at higher levels, we can expect that serious progress will start with
establishing lower-level groupings, on the order of Burling’s (1983)
hypothesis of a special relationship among Bodo-Garo, the Konyak Naga
languages, and Jinghpaw. This paper discusses data which offer significant
support to Burling’s “Sal” hypothesis – I will present here what I think
is strong comparative evidence for a quite close relationship between
Jinghpaw and Nocte, which nails down one side of the triangular Sal
grouping.
2. The Jinghpaw Tense-Aspect-Agreement Complex
Jinghpaw
2
has a very peculiar verbal system, as yet far from thoroughly
understood. Along with a large set of grammaticalized verbs (Matisoff
1
I am grateful to LaRaw Maran, my first teacher in Tibeto-Burman, who
originally provided the Jinghpaw data in this paper (the evident lacunae and
any other errors in their interpretation are mine), and to the late Dr Alfons
Weidert, who generously shared with me his unpublished notes on Nocte.
2
Examples are in the standard orthography as used in Hanson 1906, 1917,
save that glottal stop, not represented in Hanson’s orthography, has been
62 North East Indian Linguistics Volume 3
1974; DeLancey, Diehl, et al. 1978), there is a system of morphologically
complex final elements, of one or two syllables, which Dai and Diehl
(2003) label “Sentence Final Words”, or SFW’s.
3
Dai and Diehl’s SFW
category actually consists of two distinct categories of SFW, which I
will label prefinal and final. We will see that the prefinal SFW construction
corresponds directly to an almost identical SFW construction in Nocte.
A sentence in Jinghpaw typically ends with some kind of mood
particle. In a declarative sentence the default is the final particle ai, which
otherwise functions as a nominalizer.
4
This ai is the simplest of the final
SFW’s. If there is no prefinal SFW, the final particle may prefix 1
st
or
2
nd
person agreement morphemes, creating disyllabic, bimorphemic final
SFW’s, such as ngng-ai in (1) and nd-ai in (2) etc.
(1) ngai MaNang hpe kayat ya ngng-ai.
I PN OBJ beat give 1-FIN
‘I beat MaNang.’
(2) nang MaNang hpe kayat ya nd-ai.
you PN OBJ beat give 2-FIN
‘You beat MaNang.’
(3) shi MaNang hpe kayat ya ai.
s/he PN OBJ beat give FIN
‘S/he beat MaNang.’
These final SFW’s are quite peculiar. They clearly reflect, somehow, the
pan-TB agreement elements 1
st
*-ŋ and 2
nd
*-n, and the 2
nd
person form
apparently also includes the *te element which Bauman (1975) identifies
as a widespread 2
nd
person element. These inflected final particles are
disyllabic, the first nasal carrying a syllable pulse. Thus the 1
st
person
form is a syllabic 1
st
person /ŋ/ followed by another copy of the same
morpheme prefixed to the final marker to form a second syllable – which
is fairly bizarre even for Tibeto-Burman. (The resemblance between the
indicated in the SFW’s where it is important to my comparative argument.
Jinghpaw has distinctive tone, but tones are not traditionally represented in
the standard orthography.
3
This may not be the most perspicuous term for the phenomenon, but for
lack of a better alternative I will use it in this paper.
4
The connection between nominalizers and final particles in Tibeto-Burman
is notorious; Jinghpaw ai is mentioned in Matisoff’s classic study of Lahu
ve (Matisoff 1972), with which it is probably cognate.
Nocte and Jinghpaw: Morphological Correspondences 63
pronoun ngai and the bimorphemic SFW ngng-ai is coincidental; the
sentence final ai is an old copula, reflecting PTB *way).
The prefinal SFW’s are CVC syllables, or disyllables formed by
prefixing plural mă- or punctual să- to a basic SFW. The CVC base
consists of two morphemes: an initial consonant which indicates tense/
aspect, mood, or other verbal categories, and a rime consisting of a vowel
plus an agreement suffix. I call these “prefinal” SFW’s because they are
followed by a final SFW. The 1
st
and 2
nd
person agreement endings of
the prefinal SFW’s occur in two series, nasal and voiceless (Table 1).
Table 1 Jinghpaw prefinal SFW agreement endings
1
st
-ing / -eng -iʔ/ -eʔ
2
nd
-in / -en -it / -et
The glottal finals in the 1
st
singular forms reflect earlier *k (PTB final
*k > ʔ is a regular Jinghpaw sound change), thus the voiceless stop forms
are the original nasal suffixes subject to some kind of devoicing process.
Since this process has no synchronic phonological motivation in either
language, it must be of some reasonable time depth. The voiceless forms
are followed by the simple final SFW ai, but the nasal forms are followed
by the inflected form with person agreement (4)–(8).
(4) ngai sa n-i
ʔ ai.
I go n-1 FIN
‘I went.’
(5) ngai sa wa s-ing ng-ai.
I go start s-1 1-FIN
‘I am starting to go.’
(6) ngai shanhte hpe ndai jaw n-iʔ ai.
I they OBJ this give n-1 FIN
‘I gave them this.’
(7) nang shanhte hpe ndai jaw n-it ai.
you they OBJ this give n-2 FIN
‘You gave them this.’
(8) shi shanhte hpe ndai jaw n-uʔ ai.
s/he they OBJ this give n-3
rd
FIN
‘S/he gave them this.’
64 North East Indian Linguistics Volume 3
The consonantal morphemes which occur as the initials of prefinal SFW’s
– n-, s- in examples (4)–(8), and s-, d- r-, l- and m- below – express
various deictic, tense/aspect/modal, and other categories (Hanson 1917;
DeLancey, Diehl, et al. 1978; Liu 1984; Dai 1992). They apparently are
the remnants of old auxiliary verbs which have lost the rest of their
original form through phonological reduction. We will see exactly the
same phenomenon in Nocte. Functionally, these elements present a very
complex and confusing picture; for a comprehensive description with
many examples see Dai (1992). Some of them seem to have basically
aspectual values, others appear to focus one or the other argument. The
choice of vowel is sometimes relevant – u is clearly associated with 3
rd
person – but the explanation for alternations between /i/ and /e/ is not
yet clear (9)–(10).
(9) ngai sa n-i
ʔ ai.
I go n-1 FIN
‘I went.’
(10) ngai sa w-eʔ ai.
I go w-3 FIN
‘I went, I did go.’
The n- and w- morphemes in the preceding examples probably reflect
PTB copular roots *na and *way; their precise aspectual value remains
unclear. An r- initial with motion verbs marks the cislocative ‘hither’
category (i.e. motion toward the deictic center, where the speaker and
hearer are), as we will see below; it has further tense/aspect functions
which I don’t entirely understand. There are d- and m- morphemes which
occur only in transitive (d- most typically in ditransitive) clauses, and
they appear to represent, respectively, subject and object focus (DeLancey
1980b). An l- morpheme, which I will discuss below, marks agreement
with the possessor of an argument. The main points of interest about this
paradigm for our present purpose are two facts which have striking
correspondence in Nocte – the morphophonemic alternation in the 1
st
and 2
nd
person forms between nasal and voiceless stop alternants, and
hierarchical agreement which reflects the highest argument on a hierarchy
of person, rather than of grammatical role.
Nocte and Jinghpaw: Morphological Correspondences 65
2.1 Alternation in the 1
st
and 2
nd
person agreement suffixes
The fact about the Jinghpaw system which is the most important for
comparison with Nocte is the alternation between nasal and voiceless
stop variants of the 1
st
and 2
nd
person endings. I cannot yet identify a
clear and consistent function to this alternation, but, in general, where
we can find minimal pairs, the devoiced alternant has a perfective or
punctual sense relative to the nasal (11)–(16).
(11) ngai sa wa s-ing ng-ai.
I go go s-1 1
-FIN
‘I am starting to go.’
(12) ngai sa wa s-eʔ ai.
I go go s-1
FIN
‘I have started to go.’
(13) ngai tsawm wa s-ing ng-ai.
I pretty go ASP-1 1-FIN
‘I am getting pretty.’
(14) ngai tsawm wa s-e
ʔ ai.
I pretty go s-1
FIN
‘I have started getting pretty.’
(15) nang dung n-in d-ai.
you sit n-2 2-FIN
‘You are sitting.’
(16) nang dung n-iʔ ai.
you sit n-2 FIN
‘You have sat down.’
2.2 Hierarchical agreement and External Possession
The second point of interest in the Jinghpaw paradigm is hierarchical
agreement – clauses where agreement indexes a 1
st
or 2
nd
person object
rather than the subject. In a transitive or ditransitive clause agreement is
often with a Speech Act Participant (SAP) in preference to a 3
rd
person
argument, regardless of subject or object status (17)–(18).
(17) ngai MaNaw hpe gumhpraw jaw
I
PN OBJ money give
n-iʔ ai.
n-1
FIN
‘I gave MaNaw money.’
66 North East Indian Linguistics Volume 3
(18) MaNaw ngai hpe gumhpraw jaw
PN I OBJ money give
n-iʔ ai.
n-1 FIN
‘MaNaw gave me money.’
Note that we have 1
st
person agreement in both clauses.
A related phenomenon, not, as far as I know, paralleled in Nocte, is
agreement with a 1
st
or 2
nd
person “external possessor”, the possessor of
the argument which would otherwise govern agreement. This is the
function of the l- morpheme (19)–(20).
(19) na a manang wa grai pyaw
your
GEN friend go very happy
l-it d-ai.
EP-2
2-EP
‘Your friend is very happy.’
(20) shi nye a baw hpe adup ya
He my GEN head OBJ hit give
l-iʔ ai.
EP-1 FIN
‘He hit me on the head.’
Note 2
nd
person agreement in (19), indexing the possessor of the 3
rd
person subject; 1
st
person agreement in (20) is with the possessor of the
object.
3. Nocte
Nocte, under the name Namsangia, was recorded already in the Linguistic
Survey of India as having verb agreement. The fuller paradigms presented
by Das Gupta (1971) made the language a more useful witness in the
reconstruction of verb agreement in Tibeto-Burman (DeLancey 1980a;
1981a), but as far as I know the resemblance between the Nocte and
Jinghpaw verbal systems has not been previously explored. Cited forms
are originally from Das Gupta (1971), but presented here in a
phonologically more complete form from unpublished notes of Weidert
(see also Weidert (1985)), which record glottal stop, which is relevant
to our comparisons with Jinghpaw.
Nocte and Jinghpaw: Morphological Correspondences 67
The Nocte agreement system is shared with Tangsa (Das Gupta 1980,
Morey this volume), although not all varieties of Tangsa show hierarchical
agreement, and there is no question that something much like the system
described here existed in Proto-Konyak.
3.1 SFW’s in Nocte
The verbal system of Nocte is formally very similar to that of Jinghpaw,
in that tense/aspect and other categories are combined with agreement
in a sentence-final syllable quite analogous to the Jinghpaw SFW. Das
Gupta (1971: 15) describes Nocte as marking tense “by adding suffixes
to the root”. These “suffixes”, however, are all syllabic, and Weidert
records them as separate words. In Nocte, the SFW may consist simply
of an agreement morpheme with a supporting vowel, indicating non-past
tense, illustrated here with the verb ka ‘go’ (Table 2).
Table 2: Nocte intransitive agreement paradigm
singular plural
1
st
ka
1
ʌŋ ka
1
ɛ
2
nd
ka ɔʔ ka
1
ʌn
3
rd
ka
1
a
Or they may be combined with one of a set of morphemes which,
like the initial elements in Jinghpaw SFW’s, consist of a single consonant,
and indicate various tense/aspectual and other verbal categories (21)–
(25).
(21)
1
ŋaa ka
1
ʌŋ.
I go 1SG
‘I go.’
(22)
1
ŋaa ka t-ʌk.
I go PAST-1SG
‘I went.’
(23)
1
ŋaa ka
1
r-ʌŋ.
I go CIS-1SG
‘I come.’
68 North East Indian Linguistics Volume 3
(24)
1
ŋaa ka
1
k-ʌŋ.
I go PROG-1SG
‘I am going.’
(25)
1
ŋaa ka m-ʌk.
I go NEG-1SG
‘I don’t go.’
Note the structural correspondence between the final syllables in examples
(21)–(25) and the Jinghpaw SFW’s – each consists of an initial consonant
with the sort of meaning which we typically find associated with
auxiliaries, plus a rime which expresses person agreement. As in
Jinghpaw, we find both nasal and stop versions of the person marker.
Apparently in Nocte the auxiliary element determines which form of
agreement suffix will occur, but the fact that the past tense form requires
the stop suffix and the progressive and cislocative the nasal is consonant
with the distribution of the two forms in Jinghpaw.
Another striking trait which Nocte shares with Jinghpaw is hierarchical
agreement: the verb agrees with a 1
st
or 2
nd
person in preference to a 3
rd
,
regardless of which is subject or object (26)–(29).
(26)
1
ŋaa-
1
mɛ
1
ʔʌ
1
te-
2
nʌŋ vaat
1
ʌŋ.
I-ERG s/he-OBJ beat 1SG
‘I beat him.’
(27)
1
ʔʌ
1
te
-1
mɛ
1
ŋaa-
2
nʌŋ vaat
1
h-ʌŋ.
s/he-ERG I-OBJ beat INV-1SG
‘He beat me.’
(28)
1
nʌŋ
-1
mɛ
1
ʔʌ
1
te-
2
nʌŋ vaat ɔʔ.
you-ERG s/he-OBJ beat 2
‘You beat him.’
(29)
1
ʔʌ
1
te
-1
mɛ
1
nʌŋ-
2
nʌŋ vaat h-ɔʔ.
s/he-ERG you-OBJ beat INV-2
‘He beat you.’
Hierarchical agreement occurs elsewhere in Tibeto-Burman (Delancey
1980; 1981a; 1988; 1989), but it is nevertheless noteworthy to find it in
these two systems which show so many other correspondences. A striking
feature of the Nocte paradigm is explicit marking of the grammatical
category called inverse. The h- morpheme indicates that agreement, rather
than being with the subject, as is typical in Indo-European languages, is
Nocte and Jinghpaw: Morphological Correspondences 69
4. A comparison of the Nocte and Jinghpaw SFW systems
The 1
st
person forms correspond perfectly, in both the nasal and stop
versions. The Nocte 2
nd
singular suffix is anomalous; the 2
nd
plural
strongly resembles the Jinghpaw 2
nd
person nasal form, but since there
is also a well-attested plural -n attested in other TB agreement paradigms,
Jinghpaw 2
nd
-Vn and Nocte 2
nd
pl. -ʌn are not necessarily cognate. The
Jinghpaw form presumably must reflect the old 2
nd
person root, but the
Nocte form could equally well be a reflex of the PTB 2
nd
person plural
suffix *-ni. There is little overlap between the set of initials in Jinghpaw
and in Nocte, and with one exception, what resemblances there are seem
to be coincidental. In spite of the resemblance in form, it seems unlikely
that Jinghpaw d-, which has something to do with focus, is related to
Nocte past tense t-. Jinghpaw m- has no functional similarity at all to
the Nocte negative morpheme, and they could hardly be cognate.
Table 3 Nocte transitive agreement paradigm
OBJECT
→ 1
st
2
nd
3
rd
SUBJECT
↓
1
st
SG.
1
ɛ
1
ʌŋ
t-iʔ
t-ʌk
PL.
1
ɛ
t-i
ʔ
2
nd
SG. h-
1
ʌŋ ɔʔ
1
th-ʌŋ t-ɔʔ
PL. h-ɛʔ
1
ʌn
th-ɛʔ
t-ʌt
3
rd
h-
1
ʌŋ h-ɔʔ
1
a
1
th-ʌŋ th-ɔʔ t-aʔ
with whichever argument is higher on the hierarchy 1
st
/2
nd
> 3
rd
. Inverse
marking is a feature shared by a few other TB languages (DeLancey
1980a; 1980b; 1981a; 1981b; Ebert 1991). Like many other TB languages,
Nocte has special form for a transitive verb with 1
st
person subject and
2
nd
person object. The transitive paradigm of Nocte is given in
Table 3.
70 North East Indian Linguistics Volume 3
One form clearly is cognate, the r- morpheme which in Nocte indicates
cislocative, that is, motion toward the deictic center. It is “only possible
in present tense” according to Weidert, and in Jinghpaw has tense
functions in the realm of future/progressive, but yet also has a clear
cislocative force (DeLancey, Diehl et al. 1978). The formal correspondence
is evident in Table 4, with Jinghpaw sa ‘go’ + r- and Nocte ka(t) ‘go’
+ r- both meaning ‘come’.
Table 4 Jinghpaw and Nocte cislocative paradigms
Jinghpaw Nocte
singular singular plural
1
st
sa r-ing ka r-
1
ʌŋ ka r-
1
i
2
nd
sa r-in ka r-ɔʔ ka r-
1
ʌn
3
rd
sa r-aʔ ka r-
1
a NONPAST
ka r-aʔ
PAST
The Jinghpaw plural forms are distinguished not by distinct agreement
suffixes, as in Nocte, but with the plural
mă- prefixed to the SFW, as in
ex. (30).
(30) nga ni wa mă-r-a
ʔ ai.
cow PL go PL-CIS-3 FIN
‘The cattle return (home).’
Functionally, in both languages the r- morpheme occurs with motion
verbs to indicate motion toward the deictic center (31)–(36).
Jinghpaw
(31) shi sa n-u
ʔ ai.
s/he go n-3 FIN
‘He is going/has gone.’
(32) shi sa r-uʔ ai.
s/he go CIS-3
FIN
‘He is coming/will come (= will arrive).’
(33) shi du n-uʔ ai.
s/he arrive
ASP-3
FIN
‘He has arrived.’
Nocte and Jinghpaw: Morphological Correspondences 71
(34) shi du r-uʔ ai.
s/he arrive CIS-3 FIN
‘He has arrived (here).’
Nocte
(35) ʔʌ
1
tɛ mɛ
1
pe kʌt
1
a.
s/he DEM OBJ take up 3. PRES
‘He takes it.’
(36) ʔʌ
1
tɛ mɛ
1
pe kʌt
1
r-a.
s/he DEM OBJ take up CIS-3. PRES
‘He brings it.’
So we have correspondence in form, meaning, and combinatorial
behaviour between not only Jinghpaw and Nocte r-, but the whole SFW
construction in which they participate.
5. Toward a reconstruction of the Proto-Sal verbal system
Based on the formal, structural, and functional correspondences between
the Jinghpaw and Nocte systems, we can reconstruct something like a
SFW category for the common ancestor of the two languages. Although
both systems clearly have roots in the PTB paradigm (DeLancey 1980a;
1981a; 2008), the SFW phenomenon must be a later development, as it
is found only in Konyak-Jinghpaw and Kuki-Chin (DeLancey 2008).
Thus we can reconstruct it to some point in the common history of
Jinghpaw and Nocte following the separation of some ancestral form of
Sal from PTB, but prior to the separation of Jinghpaw and Konyak.
Without a detailed tree of relationships for the Sal group, we cannot
automatically assume that this common ancestor is Proto-Sal, since it is
hypothetically possible that Jinghpaw and Konyak might constitute a
genetic subunit within Sal, defined by the common innovation of the
SFW system. However, I expect that we are reconstructing here at the
Proto-Sal level, or even beyond – the suffixal agreement paradigms of
Northern Chin (Henderson 1957; Stern 1963; DeLancey 2008) seem to
have the same SFW structure.
This hypothesis, of course, requires some explanation for the lack of
the SFW structure, and indeed of agreement in any form, in Singpho,
Konyak languages such as Chang and Konyak, and, most significantly,
the entire Bodo-Garo branch, and thus presumably Proto-Bodo-Garo. In
72 North East Indian Linguistics Volume 3
a recent paper (DeLancey 2009) I have argued that the unusual
transparency and regularity of Bodo-Garo grammar, which is
reconstructible for Proto-Bodo-Garo, reflects a stage in the development
of PBG when it was a widespread lingua franca throughout the
Brahmaputra Valley (as evidenced by the wide geographical distribution
of the family down to the present), and underwent the kind of extreme
morphological simplification which is characteristic of such languages.
A similar explanation can be advanced for the simplified morphological
structure of Singpho, which has spread into North East India in significant
part by being adopted by communities which originally spoke other
languages (see Leach 1954; DeLancey 2008).
The source structure for the SFW construction must have been an
inflected auxiliary following the uninflected (or, possibly, nominalized)
lexical verb, a construction which occurs throughout the family. In form
a Nocte SFW or Jinghpaw prefinal SFW is precisely analogous to the
last word in the Hayu (Kiranti) sentence in (36).
(36) gu khok la-ŋ.
I walk go-1.SG
‘I’m leaving.’ (Hayu)
Syntactically they are different, in that Hayu la- is a lexical verb, which
can occur independently as the main verb of a sentence, while Nocte-
Jinghpaw r- no longer is. There is, however, ample comparative evidence
for a Tibeto-Burman verb *ra ‘come’, which can thus be identified as
the source for the Nocte-Jinghpaw morpheme (DeLancey 1980a;
1985).
Thus prefinal SFW’s are phonologically reduced inflected auxiliary
verbs. The remaining problems are the Jinghpaw final SFW construction
and the structure of the Nocte agreement syllable. The agreement elements
themselves pose only minor problems. The relation between the 1
st
person
forms in both languages, and 2
nd
person forms in Jinghpaw, and the
corresponding pronouns is evident (Table 5).
The independent pronouns in both languages are transparent
inheritance of the PST pronominal roots *ŋa and *na(ŋ). The
corresponding agreement syllables must represent grammaticalization, at
some chronological level, of the pronominal roots. However, the pronouns
provide no source for the initial vowel in the agreement syllables in
Nocte. Reflexes of them do occur elsewhere in the family in disyllables
Nocte and Jinghpaw: Morphological Correspondences 73
with some sort of syllabic prefix (Bauman 1975), but these are secondary
developments, and the presence in both languages of the unchanged
original pronominal forms shows that nothing of the sort has occurred
in the history of this branch.
The Nocte agreement syllables must have the same origin as the other
SFW’s; the vowel represents an older AUX element, likely a copula. In
any case, whatever the origin of the Nocte ʌŋ/ʌn forms, I suspect that
they represent the source of the strange Jinghpaw final SFW forms, which
then represent the conflation of this vowel-inital SFW with the final
particle ai, something like the development sketched in Table 6.
Table 6 Source of Jingpaw SFW forms
σ σ σ σ σ
σ
| | | | | |
*V
ŋ ai > V$ŋai > ŋ$ŋai
Abbreviations
1 First person
2 Second person
3 Third person
σ Syllable
$ Syllable boundary
ASP Aspect
AUX Auxiliary
CIS Cislocative
DEM Demonstrative
ERG Ergative
Table 5 Correspondence of pronominal and agreement forms
Pronoun Agreement
1
st
JP ŋai iŋ ~ eŋ
N
1
ŋaa
1
ʌŋ
2
nd
JP naŋ in ~ en
N
1
nʌŋ ɔʔ SG
ʌn PL
74 North East Indian Linguistics Volume 3
FIN Sentence-final particle
GEN Genitive
INV Inverse
OBJ Object marker
PL Plural
PN Personal name
PRES Present
SFW Sentence-final word
SG Singular
References
Bauman, J. (1975). Pronouns and Pronominal Morphology in Tibeto-Burman.
PhD Dissertation. Berkeley, University of California.
Burling, R. (1983). “The Sal languages.” Linguistics of the Tibeto-Burman Area
7(2): 1–32.
Dai, Q. (1992). Jingpo Yu Yufa. [Grammar of the Jingpo Language.] Beijing,
Zhongyang minzu xueyuan chubanshe. [in Chinese]
Dai, Q. and L. Diehl. (2003). “Jinghpo.” In G. Thurgood and R. J. LaPolla, Eds.,
The Sino-Tibetan Languages. London/New York, Routledge: 401–8.
Das Gupta, K. (1971). An Introduction to the Nocte Language. Shillong, North-
East Frontier Agency.
_____. 1980. The Tangsa Language (a Synopsis). Shillong, North-East Frontier
Agency.
DeLancey, S. (1980a). Deictic Categories in the Tibeto-Burman Verb. PhD
Thesis. Bloomington, Indiana University.
_____. (1980b). “Parameters of empathy.” Journal of Linguistic Research 1(3):
40–49.
_____. (1981a). “The category of direction in Tibeto-Burman.” Linguistics of
the Tibeto-Burman Area 6(1): 83–102.
_____. (1981b). “An interpretation of split ergativity and related patterns.”
Language 57(3): 626–57.
_____. (1985). “The analysis-synthesis-lexis cycle in Tibeto-Burman: A case
study in motivated change.” In J. Haiman, Ed., Iconicity in Syntax.
Amsterdam/Philadelphia, John Benjamins: 367–89.
_____. (1988). “On the evolution of the Kham agreement paradigm.” Linguistics
of the Tibeto-Burman Area 11(2): 51–61.
_____. (1989). “Verb agreement in Proto-Tibeto-Burman.” Bulletin of the School
of Oriental and African Studies 52: 315–33.
Nocte and Jinghpaw: Morphological Correspondences 75
_____. (2008). “Towards a history of verb agreement in Tibeto-Burman.” Paper
presented at the 14
th
Himalayan Languages Symposium, Gothenburg,
Sweden, August 21–23.
_____. (2009). “On the origins of Bodo-Garo.” Paper presented at the 4
th
International Conference of the North East Indian Linguistics Society,
Shillong, Meghalaya, India, January 15–17.
DeLancey, S., L. Diehl and L. Maran. (1978). “A localistic account of aspect in
Jinghpaw.” University of Michigan Papers in Linguistics 2(4): 49–88.
Ebert, K. (1991). “Inverse and pseudoinverse prefixes in Kiranti languages:
Evidence from Belhare, Athpare and Dungmali.” Linguistics of the Tibeto-
Burman Area 14(1): 73–92.
Hanson, O. (1906). Kachin Dictionary. Rangoon, Baptist Board of
Publications.
_____. (1917). A Hand-book of the Kachin or Jinghpaw Language. Rangoon,
American Baptist Mission Press.
Henderson, E. J. A. (1957). “Colloquial Chin as a pronominalized language.”
Bulletin of the School of Oriental and African Studies 20: 323–7.
Leach, E. R. (1954). Political Systems of Highland Burma: A Study of Kachin
Social Structure. London, Bell and Sons.
Liu, L. (1984). Jingpozu Yuyan Jianzhi (Jingpoyu) [Brief description of the
language of the Jingpo people (Jingpo language)]. Beijing, Nationalities
Press. [in Chinese]
Matisoff, J. (1972). “Lahu nominalization, relativization, and genitivization.” In
J. Kimball, Ed., Syntax and Semantics I. New York, Seminar Press: 237–57.
_____. (1974). “Verb concatenation in Kachin.” Linguistics of the Tibeto-
Burman Area 1(1): 186–207.
Morey, S. (this volume). “Tangsa Agreement Markers.” In G. Hyslop, M. Post
and S. Morey, Eds., North East Indian Linguistics Volume 3. New Delhi,
Cambridge University Press: 76–101.
Stern, T. (1963). “A provisional sketch of Sizang (Siyin) Chin.” Asia Major 10:
222–78.
Weidert, A. (1985). “Paradigmatic typology and its application to verb agreement
analysis.” In U. Pieper and G. Stickel, Eds., Studia Linguistica Diachronica
et Synchronica: Werner Winter, sexagenario anno MCMLXXXIII. Berlin/
New York, Mouton de Gruyter.
_____. (undated). Unpublished notes on Nocte.
76 North East Indian Linguistics Volume 3
6
Tangsa Agreement Markers
1
Stephen Morey
Research Centre for Linguistic Typology, La Trobe University
1. Introduction
In all of the various linguistic varieties that are grouped under the heading
Tangsa, verb complexes are marked by portmanteau morphemes that
convey information about person, and tense/aspect or polarity. Consider
(1), an example from Cholim Tangsa. Here the reciprocal verb răbom
‘say’ is marked by the 2
nd
person plural past marker, king. As we will
see below in §2 (Table 4), -ing is the marker of 2
nd
person plural and
k- is present on all 1
st
and 2
nd
person past forms.
1
This paper has been produced with the assistance of a fellowship from the
DoBeS program (Dokumentation der Bedrochter Sprachen), funded by the
Volkswagen Foundation. The Cholim, Lochhang and Yongkuk was collected
between January 2007 and January 2009 and is in the process of being archived
at DoBeS (www.mpi.nl/DoBeS), where recordings of the texts analysed here
will soon be available. The complete text corpus is available at the Tai and
Tibeto Burman Languages of Assam website,
http://purl.oclc.org/assam-tai. I
am very grateful to the Research Centre for Linguistic Typology, La Trobe
University and I thank Randy LaPolla, Mark W. Post, Scott DeLancey and
Alec Coupe for helpful comments. Krishna Bodo of Gauhati University was
of invaluable assistance in the collection of the Cholim word list and in the
analysis of Das Gupta’s data. Thanks also go to Palash Nath and Jyotiprakash
Tamuli of the GU Linguistics Department. My greatest thanks go to my
Tangsa teachers, Lukam Tonglum (Lukam Cholim), Chonja Tonglum and
Nongtang Langching.
Tangsa Agreement Markers 77
1) nyim ni ămoe wa răbom king.
ɲim¹ ni² əmɤ² βa² rə-bom² kiŋ³
2DL two how PRT RECIP-speak PST.2PL
‘You two, what did you say?”
The function of these markers is similar to the ‘sentence final words’ in
Jinghpaw and Nocte as discussed by DeLancey (this volume), though
much simpler than the system described for Jinghpaw (Jinghpo) in China
by Dai (1992), briefly reprised in Dai and Diehl (2003).
We will term these ‘agreement markers’ because they carry person
agreement with at least one of the arguments of the verb. In the Cholim
and Lochhang varieties (
§2 and §3), this agreement is only with the actor,
whereas in Moklum, as reported by Das Gupta (1980), and briefly
discussed in §5, there is hierarchical marking, in other words 1
st
and 2
nd
person undergoers are marked together with the actor in at least some
situations. This difference of structure between Cholim and Lochhang
on one hand and Moklum on the other has implications for the
subgrouping of Tangsa varieties. Although much more data needs to be
collected, particularly on Moklum, it would appear that Moklum should
be treated as belonging to a different subgroup within Tangsa.
These markers are not obligatory. As we will see in §2.7, there are
examples of bare verbs and also clauses containing markers carrying
information about Tense/Aspect that are invariant for person, and are
thus are not agreement markers. These may or may not occur in
combination with the agreement markers.
In this paper we will closely examine agreement in Cholim (§2),
compare that with another Tangsa variety, Lochhang (§3) and then in §4
compare both of these with the varieties discussed by Das Gupta (1980),
particularly Moklum (§5).
1.1. A brief typological overview of Tangsa
These brief comments are based largely on the Cholim variety of Tangsa,
though from our examination of other varieties, we expect them to apply
more generally.
Tangsa words are mostly monosyllabic and tone bearing. In the
varieties examined so far, there are three tones on open syllables. The
tones of Cholim, Lochhang and Yongkuk are given in Table 1. There is
a tonal contrast in stopped syllables in Cholim, but seemingly not in
Lochhang.
78 North East Indian Linguistics Volume 3
The basic prosody of multisyllabic words is iambic, accompanied by
a general downdrift in pitch across longer words.
Clauses are verb final, although sometimes an actor or other argument
may be post-posed. Verb serialisation is permitted, and the agreement
markers that are the topic of this paper have scope of the whole verb
complex. Nominalization is frequently used for subordinate clauses, and
is marked by a prefix a-, also realised as ə- which is also found as general
nominal prefix.
There is no agreed practical orthography in use for Tangsa at this
stage. The draft orthography used in this paper has been discussed with
Cholim speakers in Kharang Kong, with ue standing for [ɯ], oe for [ɤ]
and ă for reduced vowels, ch for [ʨ], chh for [ʨʰ], w for [ß] j for [ʒ] and
q for glottal stop. Vowel length is not distinctive.
1.2. The Tangsa people and their languages
Most but not all Tangsa people regard themselves as Naga, a name given
to many Tibeto-Burman speaking communities in the hill regions between
Burma and India. The Naga groups in the north and east speak languages
of the Bodo-Konyak-Jinghpaw group (also known as Sal), while the
remaining Naga groups are categorised by Burling (2003: 182) in terms
of the Konyak group to their north and the Mizo, Kuki and Chin to their
south. As he says, “between these geographical extremes we find massive
heterogeneity and uncertainty”.
The most recent version of Ethnologue lists 8 languages within
Konyak, of which Tangsa (there termed Naga-Tase), Nocte and probably
Tutsa seem to form a sub-group, the use of agreement markers being
one of its features.
According to Saul (2005: 28), “groups or sub-groups such as the
Muklum, Longchang, and the Havi are loosely gathered under the
umbrella title of Tangsa, a term coined in 1956 and embracing thirty-two
identified sub-groups”. He added that “The Tangsa are divided into two
Table 1 Tangsa tones
1 2 3
Cholim high level, glottal final low falling high falling
Lochhang low, glottal final high falling, glottal
Yongkuk low, glottal final mid-high level high falling
Tangsa Agreement Markers 79
sections depending on their residence in India. The first group of settlers
are known as Tangwa, while the later arrivals are known as Pangwa or
Pangsa”.
According to Morang (2008: 17) Tangsa can be divided into
(1) Tikhak, (2) Lungchang, (3) Moklum (4) Pangwa or Pewai, (5) Junkuk
and (6) Kato. Of these, Tikhak, Lungchang (Longchang), Moklum,
Junkuk (Yongkuk) and Kato are apparently both clan/group names and
also linguistic varieties. Tikhak, Lungchang, Junkuk (Yongkuk) and Kato
are groups that have been in Assam for a long time. There are apparently
no remaining Kato speakers, but Tikhak speakers have reported that the
Yongkuk and Longchang varieties are closest to their own. All three
have similar agreement systems, such as a future marked by post-verbal
particle/postclitic that has initial ch-, as we see from Das Gupta’s data
on Tikhak and Longchang in Table 7.
Our Yongkuk consultants stated that Yongkuk and Longchang were
nearly identical. We tentatively conclude that Tikhak, Longchang and
Yongkuk form a linguistic subgroup.
Morang’s Pangwa, on the other hand, includes a wide variety of
groups, at least some of which have more recently arrived in India from
Burma. Both Cholim and Lochhang are included by Morang within
Pangwa. These two varieties are much less mutually intelligible than
Longchang and Yongkuk, as witnessed by the fact that our main
Lochhang consultant, Nongtang, does not speak Cholim though living
in Cholim village.
Linguistically Moklum is probably the most divergent, given its
different agreement system (see §5).
Table 2 compares linguistic data collected for this project (Cholim,
Lochhang and Yongkuk) with data from Das Gupta (1980), Bandyopadhyay
(1989)
2
(abbreviated as Bandyo-), and that in Weidert (1987)
3
. We have
grouped this table into four sections marked by bold lines, sections that
may represent linguistic subgroupings. The groups that are listed by
Morang as Pangwa are given first, followed by Yongkuk-Longchang-
Tikhak. Third are two varieties discussed by Das Gupta and regarded as
somewhat divergent by some other groups are Havi and Ponthai. Both
2
From the data in Table 2, we suggest that Bandyopadhyay collected a list
similar to Moklum, as reported by Das Gupta.
3
Weidert does not make clear which variety he was talking about, for at one
point he talks about “Tangsa or Moshang” (1979: 51), while elsewhere stating
that the tone system was that of the Jugli dialect (1979: 85).
80 North East Indian Linguistics Volume 3
have the form ingat for ‘one’. Finally the three sources for Moklum are
listed. One feature of Moklum is final -l on ‘fire’, ‘sun’ and ‘good’. This
feature is partially shared by Havi, Mosang and perhaps Morang.
The data on Cholim in §2 and §3 were collected in Kharang Kong
village, Margherita subdivision of Tinsukia District, Assam, shown in
the map in Figure 1. Kharang Kong is a complex of villages, consisting
of a Cholim speaking village set up in the 1950s, an older Longri (Tangsa)
village where Singpho is the spoken language, and an Assamese village.
Just to the north of Kharang Kong is the village of Ninggam (a Singpho
word meaning ‘cliff’) which consists of a Tai Phake village of 22 houses,
an Adivasi (tea-tribe) village and an older Tangsa Lochhang village. In
the latter Lochhang is now used only by older people; Singpho (Bodo-
Konyak-Jinghpaw) is the lingua franca between Tangsa groups, and to
some extent between all of these groups.
The ethnonyms for the various Tangsa groups vary. For example, the
Cholim call themselves [ʨo¹lim³], literally ‘yam hot’, whereas the
Lochhangs call them [tjaŋ²lam³] which means ‘yam boiled’ in Lochhang.
The general name used by other Tangsas, and by Cholims with outsiders,
Table 2 Comparison of Tangsa words
‘one’ ‘hill’ ‘fire’ ‘water’ ‘dog’ ‘sun’ ‘good’
Cholim ße¹si¹ kɯ¹ ßai¹ kham² ko²ko³ raŋ²xai² a-hai³
Lochhang sai¹ kau¹
ße¹ khe³ ku²ku³ ra³si³ a-he²
Longri ashi kan wei (vei) kham koko rangshei ahe
Kimcing ashi ku war kham kuki rashe ahe
Ronrang waci pum wan kam gihe rangmit ahan
Morang ashi kin war (var) kham kuku rosyel (?) axa
Mosang ashi ku/kan var (war) dzu (ju) gui rangshal ahal
Yongkuk asi¹ kaŋ¹ ßɯ¹ kham³ hui¹ raŋ²sa³
Longcang asi kang vi (wi) kham hui rang-sa ahu
Tikhak asi kang vi (wi) kham hoi rang-sa ahu
Havi ingat hakan wal (val) jong hu: rangsa jachal
(kham)
Ponthai ingat kan we tong hui rang-sa jase
Moklum ashe ha-kan wal jung hi rangsal axal
Bandyo- aa
še haaʔ Baal juŋ hí raáŋsaal aaxaal
kaan
Weidert ¹ʌ¹si(ˀ) ¹vʌl ²d̟ʌi ¹hi(ˀ) ³rʌŋ³šal
Tangsa Agreement Markers 81
is Tonglum. The Lochhangs call themselves [lo³ʨʰaŋ³], literally ‘eagle
rhinoceros’ according to Nongtang Lanching, but their ‘general name’
is Langching.
The linguistic relationship between Cholim and Lochhang Tangsa has
been explored in some detail. Table 3 lists the correspondences between
the seven Cholim simple vowels and cognate words in Lochhang. The
correspondences in Table 3 hold for content words, but there are
differences in the correspondences between the agreement markers,
discussed below in §3.
Figure 1 Location of Kharang Kong village (approx 27°22’ N, 95°48’ E)
Table 3
Correspondence of Cholim simple vowels with Lochhang
Cholim Rhyme Example Lochhang Rhyme Example Gloss
/i/
ße¹si¹ /ai/ sai¹ ‘one’
/e/
ne² /i/ ɲi² ‘ear’
/a/ ʨʰ
a¹ /o/ ʨʰo¹ ‘elephant’
/a
ŋ/ raŋ² /a/ ra³ ‘sky’
/u/ men
¹su² /aɯ/ min¹ saɯ³ ‘cattle’
/u/
pu³ /au/ pau² ‘snake’
/o/
mo² /u/ mu² ‘you (Sg)’
/(j)o/
ɲip¹pjo¹ /(j)aŋ/ ɲep pjaŋ³ ‘breast’
/ɯ/ ßɯ¹ /au/ ɣau¹ ‘fowl’
/jɤ/ ʒjɤ¹ /i/ ʤi¹ ‘leg’
82 North East Indian Linguistics Volume 3
2. Agreement in Cholim
The Cholim agreement system is given in Table 4, where V stands for
the position of the verb. As already indicated in §1, these markers indicate
the person of any actor argument, stated or unstated, and also mark
tense/aspect features such as future/irrealis (§2.1), past (§2.4), continuous
(§2.3) as well as negation (§2.5). Attributive and other subordinate clauses
are marked by a combination of the nominalization prefix and the
agreement marker and are discussed in §2.2.
Table 4 Cholim agreement markers
1SG 1PL 2SG 2PL 3
Future/Irrealis me¹ V aŋ³ me¹ V i³ me¹ V u³ me¹ V iŋ³ me V
Attributive
a¹ V aŋ³ a¹ V i³ a¹ V u³ a¹ V iŋ³ a V
Continuous V laŋ³ V li³ V lu³ V liŋ³ V lu
³
Past V kyo³ V ki³ V ku³ V kiŋ³ V tuʔ¹
Negative V maŋ³ V mi³ V mu³ V miŋ³ V muʔ¹
In case of both the future/irrealis and the attributive form, there is a
preverbal element and a post verbal agreement marker, the latter glossed
as AGR. In most cases the continuous, past and negative forms subsume
this agreement marker, so that we can probably analyse the 1
st
person
singular continuous /laŋ³/ as containing a continuous marker /l-/ and an
agreement marker /-aŋ³/. In our glossing we have subsumed agreement
and continuous into a single element, glossing this as CONT.1SG, rather
than CONT.AGR.1SG.
These agreement markers are not obligatory, as we will see in §2.7,
where examples with bare verbs and also with particles that are invariant
for person and number are exemplified. In addition to those, there is an
imperative/hortative system which is marked for a 1
st
person/2
nd
person
distinction, but not for number, which will be exemplified in §2.6.
Returning to Table 4, we see that the 3
rd
person is not always marked
by a final element, as with the 3
rd
person future. This differs from
Moklum, as we will see below in Table 7, where all persons are always
marked.
The only 1
st
singular form without final -aŋ is the past. The form
here is kyo³, which historically probably had final -ŋ, when we consider
Tangsa Agreement Markers 83
both that the cognate form in Lochhang is keŋ³, and that there is a sound
correspondence between the Cholim rhyme /jo/ and the Lochhang rhyme
/jaŋ/ (see Table 3 above).
One of the issues in the discussion of Tangsa agreement markers is
whether they should be regarded, prosodically, as particles, i.e. separate
words, or as clitics that are prosodically dependent on the verb.
Consider (2) where we see a serial verb construction tădyoeq gep nyen
‘able to make him fall and shoot him’, the whole of which is in the
future
4
.
2)
…khălung along lalong ara ămoe re
khəluŋ² a²loŋ³ la²loŋ³ a²ra² əmɤ² re²
eagle PN great.eagle here how SEQ
wa rang kho chue wa me
βa² raŋ² kho² cɯ³ βa² me¹
PRT sky side up PRT FUT
tădyoeq gep nyen ang…
tədjɤʔ² gep² ɲen² aŋ³
[CAUS-fall shoot able AGR.1SG]
‘…How can I manage to make the great eagle in the sky fall and
shoot him? ...’ Cholim Naga Story
5
, told by Lukam Cholim, No
(274)
The wave form for the verbal complex is shown in Figure 2, which
clearly demonstrates that the forms me and ang are prosodically separate
from the rest of the construction. The causative
tădyoeq ‘cause to fall’
is closely bound to the verb gep ‘shoot’, and that structure is also bound
to nyen ‘able’. The schema for the whole verb complex is prosodically
and grammatically [me [[tădyoeq gep] nyen] ang]. The final agreement
marker ang follows a hesitation on the part of the speaker, the presence
of which suggests that it is not a prosodically bound form.
On the other hand, there are cases where we might want to argue for
treating these as prosodically bound. The predicate me ke ang ‘I will go’
is often realised as [me¹ k
e
jaŋ³], where the verb and agreement particle
4
We will use the draft orthography when discussing examples in the text,
rather than the more cumbersome phonemic transcription.
5
This text is numbered SDM12-2008Tascam-055 and appears under that name
on the Tai and Tibeto-Burman Languages of Assam website (
http://purl.oclc.
org/assam-tai).
84 North East Indian Linguistics Volume 3
are prosodically fused and the vowel of the verb is assimilated to the
following element. This fusing has been recorded only in fast speech,
and only when the final segment of the verb is a vowel or vowel plus
glottal stop, and the initial segment of the agreement marker is a vowel.
Because in most cases the agreement markers are prosodically
independent, we will notate these as particles, except for the nominalizer
prefix that is discussed in relation to (4) below. As we shall see, this
analysis of agreement markers as prosodically independent particles does
not hold for Lochhang, as discussed in §3.
As already mentioned in §1, these markers show agreement with the
actor. In Cholim Tangsa, the semantic role of experiencer is not subsumed
under the macro-role of ‘actor’ (see Van Valin and LaPolla 1997 for a
discussion of macro-roles). This is shown in (3), where we have an
experiencer ‘I’, marked by the anti-agentive, and 3
rd
person agreement
on the verb.
3) nye maq akhing ra lu wa.
ɲe¹ maʔ¹ a¹khiŋ² ra² lu³ βa²
1SG A.AG time need CONT.3 PRT
‘I need time.’ Cholim sentence, spoken by Chonja Tonglum
Sentences like this are sometimes said to have ‘experiencer subjects’,
or ‘dative marked subjects’, and this example could be literally translated
as ‘to me there is need of time’. The function of the anti-agentive is to
mark recipients, animate patients and some experiencers as in
(3). All
Figure 2 Wave form for me tădyoeq gep nyen ang
0.1224
0
0.09459
0
Time (s)
3.61304
me
ta-dyoeq
gep
nyen
ang
Tangsa Agreement Markers 85
of these semantic roles are subsumed under the general heading of
undergoer in Cholim.
Clearly the agreement on the verb in this example
is not with the 1
st
person singular. We can conclude from this that in
Cholim agreement is typically occurring with the semantic roles that can
be grouped together as ‘actor’, most often the agent.
We have seen that several verbs can occur in a single verb complex,
as in (2) above, where all are marked within the same structure. In (4),
on the other hand, one of the verbs ke ‘go’ is marked in the future and
the other in nominalized. The phrase ăsai me ke i consists of a matrix
clause within which there is a subordinate, nominalized clause, ăsai, here
translated as ‘to kill, for the purpose of killing’.
4) păngai pu joelyoe le
pəŋai³ pu³ ʒɤ³ljɤ² le²
thus snake if
ă-sai me ke i ngo re…
ə-sai¹ me¹ ke¹ i³ ŋo² re¹
[NOMZ-cut FUT go AGR.1PL] say SEQ
‘Having said, “If he is indeed a snake, we will go to kill him”
...’ Cholim Naga Story, told by Lukam Cholim, No (55)
In this example, the scope of the agreement marker is only over the
verb ke. An alternative way of expressing this would be me sai ke i ‘we
will go and cut’, with both verbs ‘cut’ and ‘go’ included within the scope
of the marker. We will see this again in (5), where the second verb kueq
(which as a full verb means ‘give’) has the function of the beneficiary,
and is thus bleached of its full verbal meaning. This bleached verb alone
cannot carry the agreement, so we cannot re-formulate (5) as *ăgvlo me
raq kueq i.
We will now discuss and exemplify the Cholim agreement system as
shown in Table 4. This will be followed by an exemplification of the
imperative, which marks person in an incomplete paradigm (§2.6) and
follow that with a discussion of those cases where agreement is not
marked (§2.7).
2.1. Future/Irrealis
The future is expressed by a combination of an invariant particle me that
precedes the verb, and an agreement particle following, as in (5), a serial
verb construction shown bracketed:
86 North East Indian Linguistics Volume 3
5) ahai me gălo kueq i…
a¹-hai² me¹ gəlo² kɯʔ¹ i³
NOMZ-good [FUT make GIVE AGR.1PL]
‘It is good, we will make (them) for you.’ Cholim Naga Story,
told by Lukam Cholim, No (78)
In (5), the verb complex is me g
ălo kueq i, literally ‘we will make
give’. The grammatical function of the verb ‘give’ in this example is to
show that the action is for the benefit of the addressee, hence the
translation ‘we will make (them) for you’ (two sons have been asked to
make boats for their father). Although we write kueq and i as separate
words, this combination meets the conditions in which the marker is
sometimes prosodically bound, and in this example they are realised as
a single prosodic unit, with a clear differentiation of the tone of kueq
and i, as [kɯˀ¹i³].
2.2. Attribute
Example (6) shows the nominalizing prefix on the verb in combination
with the agreement markers. This structure is used for adjectives, relative
clauses and other subordinate clauses. We have termed it ‘attribute’. It
is exemplified in the second line of (6), where the bracketed portion a-
khyoe toet ing means ‘(which/when) you are finished planting’. The verb
toet ‘finish’ conveys aspect, while the construction a V ing conveys that
the whole structure is attributive or subordinate.
6) nyim ni
sephue raq kămchhin kămpet
ɲim¹ ni
2
se¹-phɯ¹ raʔ¹ kəm²chin³ kəm²pet²
2PL two child-eld.bro AG gourd.type pumpkin
a-khyoe toet ing
a¹-khjɤ² tɤt¹ iŋ³
[NOMZ-plant finish AGR.2PL]
păra nang la ke kyoeq.
pəra² naŋ² la² ke¹ kjɤʔ²
that at look go IMP.away
‘You two brothers being finished planting the gourds, look away!’
Cholim Naga Story, told by Lukam Cholim, No (38)
The
a- is analysed as a prefix both because it attaches directly to the
main verb (which the agreement particle does not), and because
Tangsa Agreement Markers 87
prosodically it is never recorded as a separate element. This accords with
the generally iambic nature of Cholim multisyllabic words. This form is
also used for relative clauses, as in (7).
7) nyi ni raq a-khyv i
ɲi¹ ni² raʔ¹ a²khjɤ² i³
[1PL.EXCL two AG NOMZ-plant AGR.1PL]
kămchhin
kəm²chin³
gourd.type
‘the gourd which we two planted’ Cholim Naga Story, told by
Lukam Cholim, No (49)
2.3. Continuous
The remaining three structures in Table 4 consist only of a particle
following the verb complex. The first of these is the continuous, used
for present time and habitual action. It is exemplified in (8).
8)
ăre miqwe agyo maq păra re
əre² miʔ¹βe¹ agjo² maʔ¹ pəra² re²
thus male NEG.have A.AG that SEQ
nyim dim ăse khe due lap ling.
ɲim¹ dim² əse¹ khe¹ dɯ² lap² liŋ²
2PL PL child what at [get CONT.2PL]
‘So, there being no men, where are you getting/ do you get your
children from?’ Cholim Naga Story, told by Lukam Cholim, No
(162)
The context of this example is that the younger son of the Naga landed
in a country where there were only females, yet they were bearing
children, so he asked one of the females the question in (8).
The continuous is commonly combined with the final particle wa and
less commonly with the final particle ngai. The latter is demonstrated in
(9).
9)
ăra phang kham păra leq re
əra² phaŋ¹ kham¹ pəra² leʔ² re²
this tree.name post that bring SEQ
88 North East Indian Linguistics Volume 3
chi lu ngai.
ci² lu³ ŋai³
cut CONT.3 PRT
‘And this phang post having been brought, used to be split.’
Building a Cholim house, by Lukam Cholim, No (9)
When this example was translated, Chonja Tonglum said that because
the phang tree was no longer available, ngai would be added, and offered
the minimal pair in (10).
10) chi lu wa ‘just now cutting’
chi lu ngai ‘used to cut, long ago’
In the continuous, the particle ngai is associated with habitual action,
which might, as in (9), have occurred at some time in the past and no
longer occurs. When combined with the continuous, wa refers to present
continuous.
2.4. Past
The past is exemplified in (11). The past almost always co-occurs with
the particle wa
.
11) a
kămchhin kămpet jak phak seq
a² kəm²chin³ kəm²pet² ʒak² phak² seʔ²
HESIT gourd.type pumpkin leaf eat eat
kho ke ki wa.
kho³ ke¹ ki³ βa²
[see go PST.1PL PRT]
‘We went and saw him eating pumpkin and gourd leaves.’
Cholim Naga Story, told by Lukam Cholim, No (44)
We categorise this form as a past rather than perfective because it
always relates to past time. The past is not used in combination with the
negative, so that (12) below is not marked for past though it is in past
time. Nor is the past used when the clause takes the attribute structure
discussed in
§2.2 above, although past time is often implied in examples
like (6).
Tangsa Agreement Markers 89
2.5. Negative
This is exemplified in (12). The negative marker cannot co-occur with
any of the other markers and temporal and aspect marking such as future/
irrealis, past and continuous cannot be marked on a negated clause.
12) a-we le kho ke mi.
a²-βe¹ le² kho³ ke¹ mi³
1SG-father if see go NEG.1PL
‘We didn’t see our father.’ Cholim Naga Story, told by Lukam
Cholim, No (48)
There is an example of the negative in combination with the perfective,
a form that does not mark person, illustrated in (18) below. That
construction was not accepted by all speakers.
2.6. The Imperative/Hortative
The imperative and hortative markers are given in Table 5. As can be
seen, whilst these do distinguish person (the hortative having 1
st
person
reference and the imperative 2
nd
person reference), they differ from the
agreement particles in Table 4 in that they do not distinguish number,
with the imperative taking the same form, whether singular or plural.
The imperative and hortative are both exemplified in (13):
13) cham seq i wang laq.
cam¹ seʔ² i³ βaŋ² laʔ²
rice eat HORT.1 come IMP
‘(You) come, let’s eat rice.’
Table 5 Cholim imperative/hortative markers
Func. (
SG) Form Func. (PL) Form
1SG V + i³ 1PL V + i³
1PL.INCL V + pha³
2SG V + laʔ² 2PL V + laʔ²
2SG.away V + kjɤʔ² 2SG.away V + kjɤʔ²
prohibitive nak² + V + ke¹ prohibitive nak² + V + ke
¹
90 North East Indian Linguistics Volume 3
In this example, imperative laq was addressed to me by Lukam
Cholim, as 2
nd
person singular, but the i hortative was addressed to both
himself and to me. He added that if the whole group of people present,
including his daughters and wife were to be invited to eat, then the correct
form would be cham seq pha.
There is only one example of pha in the texts that we have recorded,
given in (14). In this example pha has a singular referent, in that it is
addressed by two sons to their mother, but the undergoer of the verb
chhung ‘us’ is a 1
st
person. This pha is almost certainly related to the
Moklum ph- forms, which mark 1
st
person undergoers of imperatives,
as shown in Table 10 and Table 11.
14) a-we ming chhung woen pha…
a²-βe¹ miŋ² chuŋ¹ βɤn² pha²
1SG-father name tell COS IMP
‘Tell us our father’s name ...’ Cholim Naga Story, told by Lukam
Cholim, No (27)
Further data is required to make a clear categorisation of the function
of pha. It seems that pha is always inclusive of everyone present, either
as actors or undergoers, and for that reason that we have glossed it as
1
st
person inclusive in Table 5.
There is one special imperative form, kyoeq, which is used when the
action is away from the speaker; consider (15).
15) ă-lue nang ke jaq gue kyoeq.
ə-lɯ¹ naŋ² ke¹ ʒaʔ² gɯ² kjɤʔ²
NOMZ-far at go return COS IMP.away
‘Go far away!’ Sentence spoken by Lukam Cholim
The prohibitive is expressed by nak V ke, as shown in (16
)
16) ngarai
gămung păra nak bom ke.
ŋa¹rai² gəmuŋ² pəra² nak² bom² ke¹
sin matter that PROH speak GO
‘Don’t speak in this sinful way!’ Cholim Naga Story, told by
Lukam Cholim, No (59)
Tangsa Agreement Markers 91
2.7. Clauses without agreement marking
Not all clauses are marked by the agreement markers in Table 4.
Sometimes a verb is unmarked, particularly when expressing present
time, if the context is fully recoverable, as in (17).
17) chong nang ke.
coŋ² nang² ke¹
school at go
‘I am going to school.’
(17) was spoken by a young lady, Jamkha; I was sitting on the porch
and could see her walking somewhere, and she called out this sentence.
It would be possible to say chong nang ke lang wa ‘school at go
CONT.1SG
PRT’ (using the continuous), but in present time the bare form is probably
more frequently used in conversation.
Finally there are two particles that convey information about Tense/
Aspect but which are invariant for person and number. These are noq
‘perfective’ and roq ‘habitual’.
The form noq is used to indicate that something is done. When asked
whether one has eaten or not, if the answer is yes, the form cham seq
noq ‘I have eaten rice’ will be used.
According to Lukam Cholim, noq can be used in combination with
the person-marked forms, at least with negative, as in (18).
18) napphyo kho nang ăpang cham
nap¹phjo² kho² naŋ² ə-paŋ¹ cam¹
morning side at NOMZ-much rice
seq noq mang.
seʔ² noʔ² maŋ³
eat PERF NEG.1SG
‘In the morning I don’t eat so much rice.’ Sentence spoken by
Lukam Cholim
A younger speaker, Chonja Tonglum, specifically rejected the
possibility that this could be said, saying that it should be seq lot mang,
(/lot²/ ‘able’).
A habitual form roq, which does not vary for person, is exemplified
in (19), compared to the continuous in (20).
92 North East Indian Linguistics Volume 3
19) la raq do woen roq.
la² raʔ¹ do² βɤn² roʔ¹
eagle AG lift.up COS HAB
‘The eagles continue to lift them up.’
20) la raq do phak lu păra.
la² raʔ¹ do² phak² lu³ pəra²
eagle AG lift.up eat CONT.3 that
‘The eagles lift up and eat them.’
Cholim Naga Story, told by Lukam Cholim, No (290)–(291)
3. Lochhang
The Lochhang paradigm is somewhat different to that of Cholim, and is
given in Table 6.
Table 6 Lochhang agreement markers
1SG 1PL 2SG 2PL 3
Fut. mә=V=a³ mә=V=ai³ mә=V=au³ mә=V=an³ mә=V
Att. a=V=a³ a=V=ai³ a=V=au³ a=V=an³ a=V
Cont. V=t
әkeng³ V=tә-kai³ V=tә-lau³ V=tә-lan³ V=tә-lә-(wa)
Past V=keng³ V=kai³ V=lau³ V=lan³ V=t
ә-(wa)
Neg. V=mau³ V=mai³ V=mau³ V=man³ V=mok
As can be seen in Table 6, Lochhang agreement markers are treated
as clitics. They are prosodically closely bound to the verb, as shown in
(21), and in Figure 3, the wave form for (21).
21) mu=r
ăq=man dong mă=saq=au.
mu²=rəʔ=man¹ doŋ² mə=səʔ=au³
2SG=AG=also rice FUT=eat=2SG
‘You are also eating/you will also eat.’ Sentence spoken by
Nongtang Langching
The future clitic
mă= in Lochhang has only been recorded as
unstressed proclitic, whose tone cannot be determined. Lochhang is more
strongly iambic than Cholim, and as a consequence when such
grammatical morphemes as the future marker, the agentive or others are
Tangsa Agreement Markers 93
followed by another element in the phonological word, the rhyme
centralizes to schwa and they become a light syllable (see Donegan and
Stampe 1983), Thus raq=man in (21) is realised as [rəman¹] and
mă=saq=au as [məsəˀau³].
The past clitic =tă, which also reduces when followed by another
element, as we see in (22), has been heard in final position with the form
/te³/, carrying full stress and tone. A fully stressed alternate of the future
prefix may thus have existed at earlier stage of the language.
When comparing Figure 3 with Figure 2 above, we can see that there
is no gap between the future clitic mă= and the verb, nor between the
verb and the agreement particle. Both figures occupy nearly the same
time length, and the difference in prosody can clearly be seen. In all of
our recordings of Lochhang made so far, we have observed the same
kind of prosodic structure of agreement markers that is observed here,
and in general one of the more salient differences between Lochhang
and Cholim is the prosodic status of the agreement markers.
0.1284
0
0.1109
0
Time (s)
3.54535
mu-raman
dong
ma=saq=au
Figure 3 Wave form for (21)
The vowel correspondences between Cholim and Lochhang in Table 3
above are regular for content words, but not for agreement markers. For
example, the Cholim 1
st
singular past agreement marker is kyo but the
Lochhang form is keng rather than the expected kang. Similarly the
Cholim 2
nd
singular continuous agreement particle is lu, and the cognate
Lochhang form is /lau³/, instead of the expected /laɯ³/. Furthermore,
the Lochhang form lau is actually marking the past, whereas the cognate
Cholim lu is the continuous.
In the first and second persons, the Lochhang continuous is derived
from the past by means of a particle tə-. A similar system seems to apply
in Kimsing, as reported by Das Gupta (1980) and briefly reprised
94 North East Indian Linguistics Volume 3
underneath Table 7 below. In Cholim, on other hand, the continuous is
not derived from the past, rather it is a different combination of initial
l- and the agreement marker.
These markers are also sometimes found marking words that would
otherwise be treated as nouns, as in (22) and (23). In this case we can
perhaps translate kok as ‘hold a festival’.
22) măchhau man kok =tălă=nai .
mәtɕʰau¹ man¹ kok =tәlә=nai³
now also festival =CONT.3=QN
‘Now is there a festival there?’
23) kau=nă=kaq kok =mok ahaq .
kau¹=nə=kaʔ kok =mok a
H
haʔ¹
hill=at=this festival =NEG.3 maybe
‘Is there not any festival in the hills?’
4. Das Gupta’s survey of Tangsa
Das Gupta (1980) gave examples of sentences in a number of Tangsa
varieties, but not either Cholim or Lochhang. Table 7 is a summary of
the forms of what Das Gupta termed ‘present’, ‘past’ and ‘future’ in 11
Tangsa varieties
6
. From Table 7 we can see that there is considerable
variety in the form of Tangsa agreement markers. In most of the varieties
surveyed by Das Gupta, the ‘future’ is expressed only by an element that
follows the verb, rather than having a preceding element me or mă- as
we see in both Cholim and Lochhang.
One feature common to all the Tangsa agreement markers is that
number is shown, but only for first and second person. In all recorded
cases third person singular and plural receive the same marking. First
person singular is usually -ang or something very similar, reflecting PTB
*ŋa. First person plurals usually have a high front vowel final.
In Kimsing the past and present first person markers are the same.
According to Das Gupta, “ta is inserted for continuous or habituative
sense. mi is prefixed for future action” (1980: 15). He gave examples as
in (24)–(25).
6
This table was produced by Krishna Boro, Department of Linguistics, Gau-
hati University.
Tangsa Agreement Markers 95
Table 7 Tangsa agreement makers from Das Gupta (1980), tabulated by Krishna Bodo
Present Past Future
1
st
person 2
nd
person 3rd 1
st
person 2
nd
person 3rd 1
st
person 2
nd
person 3rd
Sg Pl Sg Pl Sg Sg Pl Sg Pl Sg Sg Pl Sg Pl Sg
Moklum
-ang -i -a -tang -ti -ta -nang -ni -na
Longcang
-rang -rai -rao -r(e)n -ro(h) -tang -tai -tao -t(o)n -ti -cang -cai -cao -c(o)n -ci
Tikhak
-rang -re -ang -te -cang -ce -ci
Kimsing
-kang -kai -lo -kang -kai -ta -ha
Jogli -lak -lai -lu -lit -tha -tak -tai -tu -tit -ta -ang -i
Mosang -tauk -ti -to -tat -ta -shaung -shi
-sho(u)
Hawi -ang -i
-tang -ti -ta
Longri -koi -ki
Morang -ta -ti
Ronrang -tang -ti
-to
Ponthei -tang
-athang -athe
96 North East Indian Linguistics Volume 3
24) ngi cam si-ta-kang ‘I am eating rice.’
(ngi ‘I’, cam ‘rice’ si ‘eat’)
25) ngi cam cung-kang ‘I boiled rice.’
(cung ‘boil’)
For Longri only three example sentences were given by Das Gupta:
nga kakoinge ‘I went’, nyidim kakinge ‘we went’ and eke ka-te ‘they
went’ (1980: 43). In his table of pronouns, Das Gupta gave nyi-dim as
‘we (inclusive)’ and eke as ‘he’. The form of the verb ‘go’ is ka. It
appears that the agreement particles for the 1
st
persons are koi and ki
respectively. The examples he has given appear to parallel Cholim ke
kyo ngai and ke ki ngai.
The Ponthei forms listed under ‘past’ are translated with a perfect in
English (1980: 48): nga wang athang ‘I have come’; ngi wang athe ‘we
have come’.
5. Some details of the system in Moklum, as reported by
Das Gupta
Moklum differs from Cholim and Lochhang, and probably most other
Tangsa groups, in having a system of hierarchical agreement marking
also found in Jinghpaw and Nocte (DeLancey, this volume). As reported
by Das Gupta, and confirmed by a very brief recording made of a
Ngaimong speaker who has lived in the Moklum area since he was a
child, in transitive and ditransitive clauses where there are 1
st
or 2
nd
person objects, additional marking is required. For ditransitives this is
detailed in Table 8, with some of Das Gupta’s examples (glosses added
by Stephen Morey and Krishna Boro).
26) nga nang-ma kolom
7
a-she ko -ta-phang.
1SG 2SG-A.AG pen CLF-one give PST.1SG>2
‘I gave you a pen.’
27) nga pi-ma kolom a-she ko -tang.
1SG 3SG-A.AG pen CLF-one give PST.1SG
‘I gave him a pen.’
7
The word kolom is an Assamese loan.
Tangsa Agreement Markers 97
28) nang-a nga-ma kolom a-she ko -ta-phu
2SG-AG 1SG-A.AG pen CLF-one give PST.2SG>1
‘You gave me a pen.’
The ending -phu in (28) suggests that final -u marks the 2
nd
singular
and that in Moklum the 2
nd
person singular past marker might be tu,
missing from Table 7.
Monotransitives behave in exactly the same way, as we can see in
(29)–(31).
29) nga nang wat -ta-phang.
1SG 2SG beat PST.1SG>2SG
‘I beat you.’
30) nang-a nga wat -ta-phu.
2SG-AG 1SG beat PST.2SG>1SG
‘You beat me.’
31) nang nga hal -ta-phu.
2SG 1SG call PST.2SG>1SG
‘You called me.’
In (30), the 2
nd
person agent is marked by -a, called nominative by
Das Gupta but here glossed as ‘agentive’. Its presence with the verb
‘beat’ suggests a stronger agentivity/transitivity for ‘beat’ than for ‘call’
in (31).
The hierarchical ph- form also appears as a 1
st
person singular
imperative/hortative, as suggested from (32) and (33).
32) nga bang-ri duk -phang.
1SG tree-CASE cut HORT.1SG
‘Let me cut the tree.’
Table 8 Ditransitives in Moklum (after Das Gupta)
Ex. Agent Recipient Theme Verb Marker Value
(26) 1
SG-Ø 2SG-A.AG pen-Ø give -ta-phang 1SG>2SG
(27) 1SG-Ø 3SG-A.AG pen-Ø give -tang 1SG>3
(28) 2SG-AG 1SG-A.AG pen-Ø give -ta-phu 2SG>1SG
98 North East Indian Linguistics Volume 3
33) hitang-a bang-ri duk -i.
1PL.EXL tree-CASE cut HORT.1PL
‘Let us cut the tree.’
The imperatives are given in Table 9, all with ka ‘go’.
Table 9 Imperatives in Moklum (after Das Gupta)
Value Form Examples (Das Gupta 1980: 11)
2SG -u ka-u Go!
2DL -shin ka-shin Go (you two)!
2PL -in ka-in Go (you all)!
There is also hierarchical marking of imperatives when 2
nd
person
acts on 1
st
person, as shown in Table 10, with sin ‘look’.
Table 10 Imperatives with 1
st
person objects in Moklum (after Das Gupta)
Value Form Examples (Das Gupta 1980: 11)
2SG -phu sin-phu Look at me!
2PL -phin sin-phin Look at me (you all)!
Thus, the whole imperative system appears to be as in Table 11.
Table 11 Imperatives/hortatives in Moklum
1
SG V + phang 1PL V + -i
2
SG V + u 2PL V + in
2
SG>1SG V + phu 2PL>1SG V + phin
2DL V + shin
Conclusion
All Tangsa varieties have agreement marking on verbs. They can perhaps
be divided into those that have hierarchical marking, such as Moklum,
and those that do not, such as Cholim and Lochhang and most probably
most other Tangsa varieties.
Tangsa Agreement Markers 99
A second feature that distinguishes the varieties is whether the future/
irrealis is expressed only by a postclitic/post-particle as in Moklum and
Longcang, as reported by Das Gupta (Table 8), or whether, as in
Lochhang and Cholim, there is a proclitic/pre-particle expressing future/
irrealis and a postclitic/post-particle expressing person agreement.
A third feature that distinguishes the varieties is whether the agreement
markers should be considered clitics, phonologically and prosodically
bound, as they clearly are for the recordings we have made in Lochhang,
or whether they should be considered particles, phonologically and
prosodically independent words, as in Cholim.
When compared with the Nocte and Jinghpaw systems discussed by
DeLancey (this volume), we see tantalising similarities, such as the high
front vowel ending for 1
st
plural, and final nasal for 2
nd
plural seen in
both Tangsa and Nocte. One intriguing difference is that whereas Nocte
has a final glottal stop on the 2
nd
singular agreement marker, Cholim
does not; in Cholim final glottal marks some 3
rd
person agreement
markers.
So far, within the Bodo-Konyak-Jinghpaw group within Tibeto-
Burman, such agreement has been reported only for Nocte, Tangsa and
Tutsa (Rekhung 1992), which appear to form a subgroup within Konyak,
and in Jinghpaw, but not in Singpho (although there are some signs of
a remnant agreement system in the Diyun variety of Singpho). A thorough
survey of agreement in all Tangsa varieties is needed and following that
an attempt at positing a proto-Tangsa agreement paradigm, which may
then lead us to be able to say more about the history of this fascinating
phenomenon.
Abbreviations
A.AG Anti-agentive
AG Agentive
AGR Agreement
CAUS Causative
CLF Classifier
CONT Continuous
COS Change-of-state
DL Dual
FUT Future
HAB Habitual
100 North East Indian Linguistics Volume 3
HORT Hortative
IMM Immediate
IMP Imperative
NEG Negative
NOMZ Nominalizer
PERF Perfective
PL Plural
PL.EXCL Plural exclusive
PN Proper name
PRES Present
PROH Prohibitive
PRT Particle
PST Past
QN Question
SEQ Sequential
SG Singular
References
Bandyopadhyay, S. K. (1989). “A Tangsa wordlist.” Linguistics of the Tibeto-
Burman Area 12(2): 79–91.
Burling, R. (2003). “The Tibeto-Burman Languages of Northeastern India.” In
G. Thurgood and R. LaPolla, Eds. The Sino-Tibetan Languages. [Routledge
Language Family Series.] London, Routledge: 169–191.
Dai, Q. X. (1992). Jĭngpōyŭ Yŭfă [Jinghpo Grammar]. Beijing, Zhongyang
minzu xueyuan chubanshe [Central Institute of Nationalities Press]. [In
Chinese and Jinghpo]
Dai, Q. X. and L. Diehl. (2003). “Jinghpo.” In G. Thurgood and R. J. LaPolla,
Eds., The Sino-Tibetan Languages. [Routledge Language Family Series.]
London, Routledge: 401–408.
Das Gupta, K. (1980). The Tangsa Language (a Synopsis). Shillong, The
Philology Section, Research Department, North East Frontier Agency [now
Govt. of Arunachal Pradesh].
DeLancey, S. (this volume) “Nocte and Jinghpaw: Morphological corresponden
ces.” In G. Hyslop, S. Morey and M. Post, Eds., North East Indian Linguistics
Volume 3. New Delhi, Cambridge University Press India: 61–75.
Lewis, M. P., Ed. (2009). Ethnologue: Languages of the World, Sixteenth Edition.
Dallas, TX, SIL International. Online version: http://www.ethnologue.com/.
Matisoff, J. A. (2003). Handbook of Proto-Tibeto-Burman: System and
Philosophy of Sino-Tibetan Reconstruction. Berkeley, University of
California Press.
Tangsa Agreement Markers 101
Morang, H. K. (2008). Tangsas – The Children of Masui Singrapum. Guwahati,
G.C. Nath on behalf of AANK-BAAK.
Rekhung, W. (1992). Tutsa Language Guide. Itanagar, Government of Arunachal
Pradesh Directorate of Research.
Saul, J. (2005). The Naga of Burma. Bangkok, Orchid Press.
van Valin, R. and R. J. LaPolla (1997). Syntax:
Structure, Meaning and Function.
Cambridge, Cambridge University Press.
Weidert, A. (1987). Tibeto-Burman Tonology. Amsterdam/Philadelphia, John
Benjamins.
Nominalization and Related Phenomena in Marma 103
Tibeto-Burman Nominalization
104 North East Indian Linguistics Volume 3
104 blank
Nominalization and Related Phenomena in Marma 105
7
Nominalization and Related Phenomena in Marma
1
Huziwara Keisuke
Kyoto University
1. Introduction
Marma is a dialect of Arakanese, a Southwestern variety of Burmese. It
is spoken mainly in the Chittagong Hill Tracts (henceforth CHT),
Bangladesh, where the population of Marma speaking people is 154,216
(1991 Census
2
). In India, Marma is commonly known as Mogh and the
corresponding population is 30,559 (2001 Census
3
). There are hardly
any linguistic reports on Marma/Mogh, except for Konow (1903), Bernot
(1958; 1966), Huziwara (2003) and Maggard et al. (2007).
1
This paper is a thoroughly revised version of my presentation under the same
title at NEILS 3. The text in Marma attached to the original appendix in this
paper is omitted owing to space constraints. I am indebted to an anonymous
reviewer for his/her detailed comments on the first draft, to Atsuhiko Kato,
Mark W. Post and Sawada Hideo for their comments on the second draft, and
to Mark W. Post for his comments on the third draft. This
study was partly
supported by the Grants-in-Aid for JSPS Fellows from the
Ministry of Educa-
tion, Culture, Sports, Science and Technology of Japan for
the research entitled
‘Indoken to sinaken kara mita tyakkugo no kizyutu to
hikakugengogakuteki
kenkyuu’.
2
It is strange but the Population Census of Bangladesh 2001 no longer lists any
tribal population, so the latest official tribal figure is that of 1991 (BBS 2002:
143)
.
3
This figure is available from a webpage of the Census of India, http://www.
censusindia.in/ (last accessed on 18 July 2008).
106 North East Indian Linguistics Volume 3
Marma has 7 vowels /a, e, i, ɔ, o, u, ə/, 27 syllable-initial consonants
/p, ph [pʰ], b, t, th [tʰ], d, c[ʨ], ch[ʨʰ], j[ʥ], k, kh [kʰ], g, m, hm,
4
n,
hn, ŋ (which also occurs as a syllable final), hŋ,
5
l, hl, r, hr, y, w, θ,
6
ʃ,
h/ and one syllable-final consonant /ʔ/ with four distinctive tones: the
high tone is marked with an acute accent; the rising tone, with a hacek;
7
the stop tone, with a word-final glottal stop; and the low tone is not
marked.
The general grammatical information of Marma is as follows. Lexemes
are classified into nouns, verbs and particles. Adjectives occur as a
subclass of verbs; adverbs, as a subclass of nouns. The most frequent
word order in a sentence is SOV, although arguments may occur before
or after the predicate, depending on context. Marma has agglutinative
morphology, predominantly using suffixes. Temporal/procedural setting
and/or modification is accomplished via combinations of auxiliary verbs
and predicate markers. Clause chaining occurs with the aid of various
kinds of subordinating/converb-forming markers. Marma has a highly
developed classifier system and a variety of “elaborate expression” (in
the sense of Matisoff (1991)). Roughly speaking, then, its characteristics
are quite similar to those encountered in Burmese (Yangon-Mandalay
dialect). In fact, many lexical items in Marma are almost identical to
those in Burmese, although Marma forms exhibit a more conservative
phonological profile, especially among initials.
8
Many grammatical
particles, however, are considerably different from those in Burmese. In
this paper, I will be focusing mainly on one of the particles not found
in Burmese.
4
Almost all voiceless bilabial nasals in written Burmese (henceforth WrB) cor-
respond
to hm in Marma. However, there is one, but very often used, exception
to this: WrB hmaa vs. Marma ma, ‘a locative marker’.
5
Palatal nasals in Burmese regularly correspond to palatalized velar nasals: WrB
ñaa vs. Marma ŋya ‘right’ (with reference to side), WrB hño: vs. Marma hŋyú
‘wither’. Further, note that Burmese
o spoken with a high tone (marked with
a colon) corresponds to ú in Rajasthali Marma, on which this paper is based.
6
Bernot (1958) reports [ð] phonetically, but there is no [ð] in my data on
Marma.
7
The rising tone is accompanied by a creaky vowel, while its counterpart in
Burmese consists of a creaky vowel quality with a high falling tone.
8
See Nishi (1999) for the historical phonology of Burmese.
Nominalization and Related Phenomena in Marma 107
On the basis of my data, most of which are drawn from narrative
texts, I will describe the various uses of the particle =ca in Marma.
9
2. Various usages of =ca
Marma =ca has four functions: nominalization of clauses, formation of
relative clauses denoting non-future events, topicalization of NPs and
sentential nominalization of non-future events.
2.1 Nominalization of non-future clauses
Marma
=ca is used to nominalize a clause. In (1), =ca functions as a
nominalizer since it enables a verb to stand before a postposition.
(1) ... laŋməyá oiʔ=ca=ma ...
husband.and.wife sleep= NOM=LOC
‘…at/in (the place) the couple was sleeping…’
=ca is also used to nominalize a verb in the subject position. In this
case, a topic marker =gá ~ =ká is also employed very often as in (2).
At the same time, it is possible to not use it, as in (3).
(2) ŋa thaŋ=ca=gá...
1.SG think=NOM=TOP
‘What I think is that ...’
(3) ... hlɔʔke [khaiŋtha]
NP
[ŋyɔ=ca]
NP
pre-phǒ.
then body tired=NOM recover -FUT
‘...and then the tired body will recover.’
=ca is occasionally used as a postadnominal marker indicating that
the referent in question is a property of the head noun.
(4) [yáŋ ɔiŋ]
NP
əyaŋ [θuthé tə-
that deep.pond in.past.times rich.man one-
yɔʔ=ca]
NP
hǐŋ-li-re
CL:man=NOM be-CMPL-NF
‘In the past, that deep pond belonged to a rich man.’
9
=ca has a free variant =ja that appears sporadically after vowel or velar nasal.
I will deal with the Rajasthali variety, which is slightly different from
the
Bandarban variety, the main dialect of Marma in CHT, where =ca
is pro-
nounced
as =cɔ.
108 North East Indian Linguistics Volume 3
2.2 Relativization of non-future events
=ca is generally used as a relative marker. It is used when a relative
clause refers to a non-future event. It is possible to relativize various
arguments in the relative clause. But it is not possible to relativize a
genitival argument.
In (5a), a nominative subject in a sentence like (5b) is relativized.
(5) a.
[[yəma rɔʔ=ca]
R
lu]
NP
=gá ...
over.there arrive=REL man =TOP
‘the man that arrived over there ...’
b. [lu] [yəma] rɔʔ-te.
man over.there arrive-NF
‘A man arrived over there.’
In (6a) a direct object in a sentence like (6b) is relativized.
(6) a.
[thú [mráŋ=ca]
R
cəraiʔ]
NP
=ma təná
that see=REL place =LOC little
ná-gaiʔ-me.
take.rest-VPL-FUT
‘Let’s take a rest at that place (that we) see.’
b. [ŋərǒ]
S
[thú cəraiʔ]
O
mráŋ-re.
we that place see-NF
‘We see that place.’
In (7a), an oblique (in this case, locative) argument in a sentence like
(7b) is relativized.
(7) a.
[yáŋ [cúŋ hǐŋ=ca]
R
tɔ́]
NP
=dǒ rɔʔ-te.
that ghost exist=REL forest =ALL arrive-NF
‘(He) arrived the forest where that ghost lived.’
b.
[yáŋ cúŋ]
S
[tɔ́ =ma]
OBL
hǐŋ-re.
that ghost forest =LOC exist-NF
‘That ghost is in the forest.’
As in (8), it is possible to form a nominal complement construction
by
=ca. It is not exactly a relative clause in that none of its arguments
are coreferential with the following head noun, but is similar in the sense
Nominalization and Related Phenomena in Marma 109
that the clause functions as a nominal modifier. In nominal complement
constructions, head nouns tend to be abstract nouns.
(8) [[
cɔiʔ ɔŋ=ca]
R
əpyɔ=ca]
NP
...
war win=REL joy=DEF
‘The joy that (he) felt on winning the war ...’
2.3 Topicalization of NPs
=ca often appears after the subject NP. In this function, =ca is glossed
as =DEF in this paper.
10
The =ca of this function is generally preceded
by animate nouns.
(9) [[
bəlú=ca]
NP
nɔiŋ-phǒ]
NP
mə-pɔ.
ogre=DEF can-FUT NEG-go.up
‘It is not that the ogre can go up.’
More rarely,
=ca is preceded by a limited set of inanimate nouns.
Although the reason for this is not clear, the noun is most often əkuŋ
‘all’.
(10) [
[əkuŋ=ca] =go] mwı́ŋ rǎ-phǒ
all=DEF =OBJ ask must-FUT
‘(We) must ask about everything.’
In the main clause of the relative-correlative construction,
11
=ca
sometimes appears to show that the subject is a definite argument.
10
=ca in this position is labelled as TOP in Dockum (2006). I am often inclined
to do so as well. In fact, the function of this =ca is in many cases that of
topicalization. Nonetheless, there are three reasons for not labeling it as a
TOP here. First, we already have =gá ~ =ká as a topic marker, and =ca
(as a nominalizer) sometimes concurs with =gá ~ =ká. Thus, I do not want
to use this TOP again for another particle (see §3.3 below for the usage of
=gá ~ =ká). Second, the Bangla (Bengali) definite particle =ṭa appears
just in this position as per my Marma teacher’s translation into Bangla.
Third, demonstratives are often the source of nominalizers in Tibeto-Bur-
man languages, and this might be also the case in Marma =ca (see §4.3
below for further discussion).
11
The Marma relative-correlative construction is different from neighbouring
Indo-Aryan and other TB languages in that both a question word (such as
ja ‘what’) and a sentence particle of question (=lé ‘=Q’) are employed in
the subordinate clause, which is perhaps accidentally very similar to the
situation in Kannada, a Dravidian language (Nadkarni 1975).
110 North East Indian Linguistics Volume 3
(11) [yáŋ θu]
NP
, [[ja jəga=ma] [ja]
that man what.COL place=LOC what.COL
louʔ rǎ-phǒ-lé], [yáŋ=ca=go]
work must-FUT-Q.COL that=DEF=OBJ
mə-taiʔ.
NEG-know
‘This man did not know what he must do and where he must work.’
(lit. ‘That man, in what place and what (he) must work, did not know
that=thing.’)
2.4 Sentential nominalization of non-future events
A verb that is nominalized in
=ca sometimes occurs as the head of a
finite/non-embedded clause, although the exact function of such usage
is not entirely clear. It is difficult to explain the difference between
sentence endings with and without =ca.
12
Therefore, we must be content
with some examples.
13
(12) kywaiŋdɔ maiŋ=nǎ cá-bəlɔiʔ=ca.
I anger=COM eat-PERF=NOM
‘It is that I ate (it) completely with anger.’
In this usage,
=ca is followed by other predicate markers.
(13) mráŋ cho
-ge, nɔiŋ=ca-bya.
horse say-COND can=NOM-PERF
‘If (it is) a horse, (it) must be able to do it.’
=ca is also used in an interrogative sentence (14).
(14) de cúŋ
-rǒ ... jəgǎ rɔʔ=ca-lé?
this ghost-PL from.where arrive=NOM-Q
‘From where did these ghosts come ...?’
12
Although it is difficult to explain, =ca of this type appears to be very
similar to the =no=da construction in Japanese, where =no serves as a
kind of nominalizer and has an explanatory force. See Martin (1988:
851–56) for illustrative examples.
13
See Matisoff (1972: 246–47) for similar constructions in Lahu. Note, however,
that in Lahu, it is possible to use this construction irrespective of tense.
Nominalization and Related Phenomena in Marma 111
3. Grammatical devices other than =ca that perform the
functions of =ca
In the previous section, we have seen various ways in which =ca is used.
However, there are certain other grammatical devices that also perform
the same functions as =ca does.
3.1 Nominalization by grammatical devices other than =ca
There is another marker in Marma that serves as a clausal nominalization
marker. The future tense marker
-phǒ is used as a nominalizer of future
clause. It is possible to regard it as a nominalizer because, as shown in
(15) (=(9)), it precedes directly to the negative predicate.
(15) [[
bəlú=ca]
NP
nɔiŋ-phǒ]
NP
mə-pɔ.
ogre=
DEF can-FUT NEG-go.up
‘It is not that the ogre can go up.’
While clausal nominalization is done by =ca
or -phǒ, lexical
nominalization is realized by affixation and reduplication. Both
derivational processes are relatively productive.
(16) a. əlouʔ ‘work(n)’ < ə- ‘prefix’ + louʔ ‘work(vt.)’
b. khrákhrá ‘various(adv/n)’ < khrá ‘divide’ (full
reduplication)
c. ŋəŋě ‘soft(adv./n.)’ < ŋě ‘soft(vi.)’ (partial
reduplication)
3.2 Relativization by grammatical devices other than =ca
When a relative clause refers to a future event, a purposive marker -pho
̌
is employed (17).
(17) [[
rwa+praŋ thwɔʔ-phǒ]
R
láiŋ]
NP
=ma ...
village+outside go.outside-PURP road =LOC
‘at the road that goes outside of the village...’
It is sometimes possible for a verb to modify a noun without a relative
clause marker (18).
14
14
It might be possible that the =ca is just adding definiteness to a similarly zero-
relativized clause. However, zero-relativized clauses are very rare in Marma.
112 North East Indian Linguistics Volume 3
(18) [[θaiŋdaphúŋ=go graŋ-khənɔ́iŋ hlɔiʔ]
R
θaiŋ]
NP
.
small.window=OBJ ONOM-ing open sound
‘(She heard) the sound of the opening of the small window.’
NOTE: -kh
ənɔ́iŋ is a kind of adverbializer that is always used
with onomatopoeia.
NOTE: θaiŋ is originally əθaiŋ. Here, the prefix is dropped.
3.3 Topicalization by grammatical devices other than =ca
There is another topic marker in Marma, namely,
=gá ~ =ká.
15
This
morpheme appears as =ká after a glottal stop, and as =gá otherwise.
The difference between =gá ~ =ká and =ca as topic markers is not yet
clear. However, compared with =ca, =gá ~ =ká can be attached to
various elements in a sentence and it appears to convey a meaning of
contrastive focus: it can follow a noun (19), postposition (20), nominalizer
=ca (21),
16
verb (22)
17
and verbal predicate (23).
(19)
əkuŋ+θu=gá əhlǎpǎ.
all+people=TOP beautiful
‘All the people are beautiful.’
(20) ŋǎ=ma =gá tə-khǔ=lé əkyú
1.OBL=LOC=TOP one-CL:thing=even benefit
mə-pɔ.
NEG-appear
‘With me, there is not even a single benefit.’
15
=gá ~ =ká might be related to Proto-Tibeto-Burman locative/ablative marker
*ka-i (DeLancey 1984: 66) although the direct reflex of *ka-i is the ablative
marker
=gǎ ~ =kǎ whose rising tone is, perhaps, derived from *-i, which bears
a creaky tone as seen in WrB
i ‘a genitive marker’ and is reconstructed as the
general subordinating particle
*-ki (Benedict 1972: 96; Thurgood 1981: 40–
41).
16
This example has a conditional meaning. As Haiman (1978) points out, it is
common for a language to employ the same word as a topic marker and a
conditional marker. In South Asia, for example, the Hindi =to is used in this
manner.
17
It seems that there is no =ca in the sentence-final position in negative sen-
tences.
However, in these cases, the sentence-final =gá ~ =ká forms a kind
of nominal
complement construction instead.
Nominalization and Related Phenomena in Marma 113
(21) lu cho=ca=gá tə-nı̌ŋ tə-raʔ
man say=NOM=TOP one-CL:day one-CL:day
θi rǎ-phǒ.
die must-FUT
‘If you are a man, you must die one day.’ (lit. ‘The so-called
man must die one day.’)
(22) təphě=lé mə-hraʔ=ká.
only.a.few=even NEG-feel.ashamed=TOP
‘It is not shameful even a bit.’
(23) ŋǎ yɔʔphǎθe nɔiŋ-re=gá, mráŋ
1.GEN brother.in.law be.able.to-NF=TOP horse
+rú=əkhɔŋ=go.
+be.mad=even?=OBJ
‘My brother in law can (ride) even the mad horse.’
3.4 Basic predicate markers other than =ca
Basic predicate markers (i.e. verbal particles that are used obligatorily
to denote tense/aspect of a sentence) other than =ca
are summarised as
follows. Owing to space constraints in this paper, examples are referred
to only by their corresponding numbers.
(24) a. future:
-me (6), -phǒ (3, 10, 11); the difference between
-me and -phǒ is not clear.
b. non-future: -re (4, 8, 17, 18, 23) ~ -te (7); -te appears only
after a glottal stop.
c. perfect: =bya (13) ~ =pya; =pya appears only after a glottal
stop.
4. In search of the origin of =ca
In this section, I will investigate both the forms and functions of several
grammatical equivalents for Marma =ca in Burmese and Arakanese.
Then, I will try to elaborate the possibility that Marma =ca is related to
demonstratives in Tibeto-Burman.
18
18
Discussions as well as grammatical forms of Burmese and Arakanese in this
section owe foremost to the insightful comments of an anonymous reviewer.
114 North East Indian Linguistics Volume 3
4.1 Comparison of Marma with Burmese
Considering the various uses of the nominalizer =ca,
sañ
19
in literary
Burmese seems likely to be cognate, because it works not only as a
nominalizer, but also as a relativizer (with an induced creaky tone),
sentence-final predicate marker, and topic marker,
20
as well as a
demonstrative. However, the problem is that its initial and rhyme do not
conform with =ca in Marma.
21
Another relative clause marker so in WrB cannot be cognate with
the Marma =ca because of the same reason.
In WrB, we have the nominalizer taa and the realis marker tay.
However, WrB taa is a fusion of the realis marker tay and the nominalizer
haa (Okell 1969: 22). As for WrB tay, the Marma -te ~ -re is the direct
reflex. In general, WrB t corresponds to Marma t. So, it is not probable
that WrB taa and tay are directly related to Marma =ca. Further, note
that, unlike WrB, the Marma -te ~ -re does not carry an induced creaky
tone and cannot be used as a relative marker.
Another candidate for a Marma =ca cognate in Burmese is, although
not attested as an independent nominalizer, the verb nominalizer suffix
caraa ‘things for’ in WrB (-səra in Arakanese), where s corresponds
to c in WrB and Marma and the schwa reduced from and written as a.
For this form, -cəra ~ -jəra is found in Marma. It might be a cognate
but is different in that it shows obligatory voicing alternation.
The last candidate is WrB cwaa, which is also attested in Old
Burmese (Ohno 2005: 267, 275). It is used as an adverbializer or an
intensifier for preceding verbs. Phonological correspondence with Marma
is fairly good, but the meaning does not necessarily match.
22
19
Nishida (1989: 807–808) suggests that *s was once a past tense marker in
Proto-Tibeto-Burman, which is still reflected in WrB sa
ñ as well as in Jingpho
past/perfect predicate marker sai
.
20
According to an anonymous reviewer, the topical use of sañ is usually attrib-
uted
to influence from Pali through Nissaya (calqued Pali) Burmese and its
function is not exactly to topicalize. See Okell (1967: 103) for examples in
Nissaya Burmese.
21
As a rule, Marma /c/ corresponds to /c/ in WrB, and WrB /añ/ to Marma /e/.
See Huziwara (2003: 262, 269) for details.
22
In WrB, cwaa is used as in pyo-hrwang-cwaa-phrang. ‘happily (lit. happy-
cwaa-by
)’, where cwaa can be regarded as a nominalizer as it precedes the
postposition phrang (Kato Atsuhiko, p.c.). cwaa is also used as a kind of
nominalizer in
idiomatic expressions like mrat-cwaa-bhu-raa: ‘The Noble
Nominalization and Related Phenomena in Marma 115
4.2 Comparison of Marma with Arakanese
The direct reflex of Marma =ca in Arakanese is a nominalizer
sɔ ~ zɔ.
23
It functions as a special head noun and also appears in sentence-final
position (Okell 1995: 28). However, there are two differences: Marma
=ca does not show compulsory voicing alternation and Arakanese sɔ ~
zɔ is not used as a topic marker.
In Arakanese, to mark the subject explicitly, a grammatical particle
re is employed. It is also used to emphasize the object (Okell 1995: 27).
As in the case of WrB sañ, re is also used as a declarative marker, where
it has an allomorph te (Okell 1995: 28–29).
24
Although the form is
completely different, this kind of bifunctionality is reminiscent of Marma
=ca. Here, the difference is only with regard to the position in the
accusative noun phrase: the Arakanese re is attached after the case marker
(Okell 1995: 47 (2.59), 51 (3.25)), while Marma =ca is attached before
it (see examples 10 and 11).
4.3 Possible demonstrative origin of =ca
It is well known that demonstratives may give rise to nominalizers (as
English that). Very similar types of phenomena are found in the Burma/
North East India corridor (Mark W. Post, p.c.): Galo (this volume), Lotha
(Herring 1991: 66), Rawang (LaPolla 2008) and Mizo (Chhangte 1993:
70) are languages in which the phenomenon of demonstratives and/or
articles giving rise to nominalizers and/or relativizing morphemes has
been identified, to name a few.
There are three demonstratives in Marma:
de, yáŋ and thú. Among
them, de is the direct cognate of both WrB demonstrative dii and
Arakanese demonstrative de (Okell 1995: 33). WrB dii has a free variant
sañ, which is, in turn, used as a nominalizer as we have seen in the
previous section.
25
Lord (Buddha; lit. noble-cwaa-buddha)’ (Sawada Hideo, p.c.). See Okell and
Allott (2001: 58
–59) for various uses of cwaa in Burmese.
23
Remember that =ca appears as =cɔ in the Bandarban dialect, which is spoken
by the majority of Marma people living in CHT.
24
According to an anonymous reviewer, the Arakanese re carries an induced
creaky
tone and serves as a relativizer, but Okell (1995: 17) states that the
Arakanese re
does not carry a creaky tone. In my observation, in Marma re
does not serve as a relativizer, and the relativizer =ca
never carries a creaky
tone.
25
See Simpson (2009) for further discussions and insights on possible links
116 North East Indian Linguistics Volume 3
Marma =ca by itself is not used as a demonstrative. However,
although it is not used as a nominalizer, one of the possible cognates for
Marma =ca is the Tamangic demonstrative ca.
26
It is not impossible
that Marma =ca has something to do with demonstratives.
5. Conclusion
In this paper, I briefly described various ways in which the Marma particle
=ca is used. Its principal function is nominalization, which extends to
relativization as well as a sentence-final declarative use. This kind of
versatile particle is common among Tibeto-Burman languages (Matisoff
1972, for instance).
The various functions and forms of the Marma =ca in Burmese (WrB)
and Arakanese are roughly summarised in Table 1.
between demonstratives, ir/realis markers and nominalizers in Burmese. How-
ever,
note that the demonstrative use of WrB sañ is not attested in inscrip-
tional
Burmese, while the topical/nominative use (from 12c onward) and the
sentence-final/nominalizer use (from 14c onward) are attested. Thus it is not
probable for the Burmese sa
ñ to be originally demonstrative (Sawada Hideo
p.c.).
26
See Honda (2007) for various forms of demonstratives in Tamangic languages.
Lotha topic marker
có (Herring 1991) and Mizo demonstrative/relative clause
marker
cû (Chhangte 1993: 70) might be cognate with Marma =ca, too.
Table 1 Marma =ca
and its equivalents in Burmese and Arakanese
Marma Burmese Arakanese
Nominalization =ca taa, sañ
sɔ~zɔ
Relativization =ca tay., sa
ñ., so re
Sentence-final =ca, -re~-te taa, tay, sañ
sɔ~zɔ, re~te
Topicalization =ca, =g
á~=ká kaa:, sañ re
Demonstrative de dii, sañ de
Adverbializer
— cwaa —
What is most striking in the case of Marma, however, is that =ca is
also used as a definite/topic marker, which is pervasive in Tibeto-Burman
languages of NE India, but not so common in elsewhere. The versatile
use of =ca is, perhaps, a result of years of language contact in this
region.
Nominalization and Related Phenomena in Marma 117
Abbreviations
- Morpheme boundary
= Clitic boundary
+ Word boundary in compound noun
1 First person
ALL Allative
CL Numeral classifier
CMPL Completive
COL Correlative marker
COM Comitative
COND Conditional
DEF Definite
FUT Future tense
GEN Genitive
LOC Locative
NEG Negative
NF Non-future tense
NOM Nominalizer
[...]NP Noun phrase
OBJ Objective
OBL Oblique
ONOM Onomatopoeia
PERF Perfective
PL Plural
PURP Purpose
Q Question
[...]R Relative clause
REL Relative clause marker
SG Singular
TOP Topic
VPL Auxiliary verb indicating plurality of subject or action
References
BBS (Bangladesh Bureau of Statistics), Ed. (2002). Statistical Pocketbook of
Bangladesh 2000. Dhaka, Statistic Division, Ministry of Planning,
Government of the People’s Republic of Bangladesh.
118 North East Indian Linguistics Volume 3
Benedict, P. K. (1972). Sino-Tibetan: A Conspectus. Cambridge, Cambridge
University Press.
Bernot, D. (1958). “Rapports phonétiques entre le dialecte marma et le birman
[Phonetic correspondences between the Marma dialect and Burmese (in
French)].” Bulletin de la société de linguistique de Paris 53: 273–294.
_____. (1966). “Êtes-vous fâchée, Belle-Mère?: Conte Marma [Are you angry,
mother-in-law?: A Marma folktale].” In B. Shin, J. Boisselier and A. B.
Griswold, Eds., Papers on Asian History, Religion, Languages, Literature,
Music Folklore, and Anthropology: Essays Offered to G. H. Luce by his
Colleagues and Friends in Honour of his Seventy-Fifth Birthday. Vol. I.
Ascona, Switzerland, Artibus Asiæ Publishers: 59–66. [in French]
Census of India Webpage http://www.censusindia.gov.in (last accessed on 18
July 2008).
Chhangte, L. (1993). Mizo syntax. PhD Thesis. Eugene, OR, University of
Oregon Department of Linguistics.
DeLancey, S. (1984). “Etymological notes on Tibeto-Burman case particles.”
Linguistics of the Tibeto-Burman Area 8(1): 59–77.
Dockum, R. (2006). “Convergence in Khumi and Marma morphosyntax.” Paper
presented at the 39
th
International Conference on Sino-Tibetan Languages
and Linguistics, 17 September, University of Washington, Seattle, USA.
Haiman, J. (1978). “Conditionals are topics.” Language 54(3): 564–589.
Herring, S. C. (1991). “Nominalization, relativization and attribution in Lotha,
Angami, and Burmese.” Linguistics of the Tibeto-Burman Area 14(1): 55–
72.
Honda, I. (2007). “A comparative and historical study of demonstratives and
plural markers in Tamangic languages.” In R. Bielmeier and F. Haller, Eds.,
Linguistics of the Himalayas and Beyond. Berlin/New York, Mouton de
Gruyter: 97–118.
Huziwara, K. (2003). “Marumago no onsei ni kansuru koosatu [A phonetic
analysis of Marma (in Japanese)].” Kyoto University Linguistic Research 22:
237–300.
Konow, S. (1903). “Notes on the Maghı̄ dialect of the Chittagong Hill Tracts.”
Zeitschrift der Deutschen Morgenländischen Gesellschaft 57: 1–12.
LaPolla, R. J. (2008). “Nominalization in Rawang.” Linguistics of the Tibeto-
Burman Area 31(2): 45–65.
Maggard, L., S. Ahmad and M. Sangma (2007). The Marma and Rakhine
Communities of Bangladesh: A Sociolinguistic Survey. Dhaka, SIL
Bangladesh.
Martin, S. E. (1988). A Reference Grammar of Japanese. Rutland/ Tokyo, C. E.
Tuttle.
Nominalization and Related Phenomena in Marma 119
Matisoff, J. A. (1972). “Lahu nominalization, relativization, and genitivization.”
In J. P. Kimball, Ed. Syntax and Semantics. Vol. I. New York/San Francisco/
London, Seminar Press: 237–57.
_____. (1991). “Endangered Languages of Mainland Southeast Asia.” In R. H.
Robins and E. M. Uhlenbeck, Eds., Endangered Languages. Oxford, Berg:
189–228.
Nadkarni, M. V. (1975). “Bilingualism and syntactic change in Konkani.”
Language 51(3): 672–83.
Nishi, Y. (1999). Four Papers on Burmese: Toward the History of Burmese (The
Myanmar Language). Tokyo, Institute for the Study of Languages and
Cultures of Asia and Africa, Tokyo University of Foreign Studies.
Nishida, T. (1989). “Tibetto-birumagoha [Tibeto-Burman languages].” In T.
Kamei, R. Kono and E. Chino, Eds., 言語学大辞典 Gengogaku Daijiten
[The Sanseido Encyclopaedia of Linguistics]. Vol. II. Tokyo, Sanseido: 791–
822. [in Japanese]
Ohno, T. (2005). “The structure of Pagan period Burmese.” In J. Watkins, Ed.,
Studies in Burmese Linguistics. Canberra, Pacific Linguistics: 241–305.
Okell, J. (1967). “Nissaya Burmese: A case of systematic adaptation to a foreign
grammar and syntax.” Journal of the Burma Research Society 50(1): 95–
123.
_____. (1969). A Reference Grammar of Colloquial Burmese. London, Oxford
University Press.
_____. (1995). “Three Burmese dialects.” In D. Bradley, Ed. Studies in Burmese
Languages. Canberra, Pacific Linguistics: 1–138.
Okell, J. and A. Allott. (2001). Burmese/Myanmar Dictionary of Grammatical
Forms. Richmond, Surrey, Curzon Press.
Post, M. (this volume). “Nominalization and nominalization-based constructions
in Galo.” In F. H. Yap and J. Wrona, Eds., Nominalization in Asian
Languages: Diachronic and Typological Perspectives. Volume 1: Sino-
Tibetan and Iranian Languages. [Typological Studies in Language Series.]
Amsterdam/Philadelphia, John Benjamins.
Simpson, A. (2009) “The Grammaticalization of clausal nominalizers in
Burmese.” In M. J. López-Couso and E. Seoane, Eds., Rethinking
Grammaticalization: New Perspectives. Amsterdam/Philadelphia, John
Benjamins: 265–288.
Thurgood, G. (1981). Notes on the Origins of Burmese Creaky Tone. Tokyo,
Institute for the Study of Languages and Cultures of Asia and Africa.
120 North East Indian Linguistics Volume 3
8
Functions of Nominalization in Karbi
1
Linda Konnerth
University of Oregon
1. Introduction
This paper is a comparative approach to functions of nominalization in
Karbi (formerly called Mikir/Arleng), a Tibeto-Burman (TB) language
from Assam in central North East India (NEI). Nominalization and its
various functions is a topic of considerable current interest in TB studies,
and one with both typological and historical implications. There may be
several nominalization constructions with deep roots in TB, including
nominalization with the kV- prefix, as found in Karbi, and nominalization
with the post-verbal particle -pa, which is widely attested in the family.
There is still disagreement as to where Karbi fits into the family. It is
often classified as particularly close to Meithei (DeLancey 1987: 800),
but other scholars prefer to place it in a branch by itself (Burling 2003:
187). Both approaches acknowledge resemblances between Karbi and
the different groups of “Naga
2
” and Kuki languages, even though those
are spoken in far eastern NEI. The status of these resemblances for the
genetic classification has yet to be determined. While this paper is
1
I would like to thank Karl-Heinz Grüßner, author of the Karbi grammar
(Grüßner 1978), who provided me with additional documents and analyses
from his fieldwork on Karbi. I am also grateful to my advisor Scott DeLancey
for very helpful clarifications on this topic.
2
“Naga” (referring to a group of languages) is put in scare quotes following
Burling (2003: 172), who points out that these languages do not constitute a
linguistic subgroup.
Functions of Nominalization in Karbi 121
primarily descriptive rather than comparative, I have included data from
two other TB languages of NEI: Angami Naga (Tenyidie), where we see
the same velar prefix as in Karbi, and Meithei (Manipuri), putatively
closely related to Karbi, but using the -pa nominalization construction
instead of the velar prefix.
The outline of this paper is as follows: §2 is devoted to derivational
nominalization, including the derivation of nouns that denote activities,
states, or events, and the derivation of agent nouns. Next, nominalization
in adjectival attribution is discussed in §3, i.e. nominalization of (a subset
of) verbs to derive nominal modifiers. The following four sections present
data of syntactic nominalization, where nominalization yields full NPs
figuring on the clausal level. The clausal counterpart to adjectival
attribution, relative clauses, is discussed first, being followed by the two
other types of subordination, complement clauses and adverbial clauses.
Finally, evidence from the peculiar case of nominalized main clauses is
presented. The data in this paper come from published sources: Grüßner
(1978) for Karbi, Chelliah (1997) for Meithei, and mostly Herring (1991)
and Giridhar (1980) for Angami Naga. Data in this article are represented
as they are in the publications from which they were drawn.
The primary purpose of this paper is to present a descriptive overview,
and limited space does not allow for a detailed discussion of any one of
the functions of nominalization. In particular, the sections on nominalization
in complement clauses and adverbial clauses only sketch out the agenda
of what needs to be done in order to determine the role of nominalization
in these types of subordination.
2. Derivational nominalization
The kV-
3
prefix in Karbi is productive in deriving nouns from verbs. In
(1), dúk ‘(to) suffer’ is nominalized by ke- yielding kedúk ‘poverty’,
which is a noun that can then be determined by the preceding
demonstrative. Furthermore, we see an a- prefix, which is glossed as the
attributive, being a marker that generally indicates a relationship between
the noun and some other modifier to this noun in Karbi. This a- prefix
is very frequent, showing up on the head noun in adjective constructions
as well as on the head noun of relative clauses as we will see in the
course of this article.
3
There are two allophones of this prefix, ke- and ka-, but since the trigger(s)
for the alternation remains unclear in Grüßner (1978: 95), I will refer to this
form as kV- throughout the paper.
122 North East Indian Linguistics Volume 3
(1) là-bàng-sō a-ke-dúk
4
DEM-CLF:G-DIM ATTR-NMLZ-suffer
‘this poverty’ (Grüßner 1978: 96)
(2) is another instance of event nominalization, figuring here inside a
sentence with a nominal predicate, where the copula is missing (note
that Karbi also has an overt copula used in other constructions).
(2) ka-cho+b
éy-kè
5
páp.
NMLZ-lie-EMPH sin
‘Lying is a sin.’ (Grüßner 1978: 96)
The data in (3) and (4) are examples of agentive nominalization. This
construction includes the indefinite pronominal classifier
bàng, here with
the attributive a-, to signal the resulting NP as an agent noun. The verb
may be nominalized by kV- as in (3), but does not have to be, as in
(4).
(3) lō ke-thán a-bàng
book NMLZ-teach ATTR-CLF:G
‘somebody who teaches/teacher’ (Grüßner 1978: 78)
(4) cho-nghū a-bàng
MID-steal ATTR-CLF:G
‘somebody who steals/thief’ (Grüßner 1978: 78)
3. Nominalization in Adjectival Attribution
In Karbi as well as in Meithei and Angami Naga, there are no adjectival
roots in the lexicon. Instead, adjectives are derived from (a subset of)
verbs by means of nominalization.
Grüßner (1978) describes Karbi as having two different adjective
constructions (ACs): definite and indefinite ACs. Functionally, the
difference is that in definite ACs, the emphasis is on the adjective, which
4
Glosses of the Karbi examples (all from Grüßner 1978) are mostly my own,
since Grüßner himself did not consistently gloss the examples but only trans
-
lated and referred to single words or morphemes.
5
I use ‘+’ as opposed to ‘-’ between what used to be a morpheme boundary,
but has become more of a lexicalized unit (specifically, cho is an otherwise
productive middle marker, but
béy cannot stand by itself to mean ‘lie’).
Functions of Nominalization in Karbi 123
is interpreted as the “center of the construction” (Grüßner 1978: 123)
– the adjective is marked and interpreted as the head (semantically and
syntactically). In indefinite ACs, on the other hand, the emphasis is
rather on the noun. Table 1 summarizes these observations.
The data in (5) and (6) show definite ACs. The adjectival verb is
obligatorily prefixed by the attributive marker a- and by the nominalizer
kV-. Since the a- prefix otherwise goes on head nouns or possessed
elements, thus indicating that it is being modified, this is a structural sign
of head-marking. According to Grüßner, èr ‘be.red’ and bī ‘be.small’
are emphasized
6
in (5) and (6). Also note that the nouns may or may not
be marked attributive
7
.
(5) cha
́m a-ke-èr
colour ATTR-NMLZ-be.red
‘red colour’ (Grüßner 1978: 124)
(6) a-plāng a-ke-bī-nèy ēn-sí
ATTR-bread ATTR-NMLZ-be.small-SPLT take-EMPH
‘She took the smallest bread.’ (Grüßner 1978: 123)
In (7) and (8), we see instances of the indefinite AC. The adjectival
verb may be nominalized as in (7), but does not have to be as in (8) – in
any case, it cannot be marked attributive by a-. And, again, note that
the nouns may or may not be marked attributive, just as in the definite
AC.
Table 1 Definite and indefinite ACs (Grüßner 1978: 123–4)
Formal Difference Functional Difference
Definite Both a-
‘ATTR’ and kV- ‘NMLZ’ Adjectival verb is the head
AC obligatory on adjectival verb
Indefinite
No a- ‘ATTR’ and only optional Modified noun is the head
AC
kV- ‘NMLZ’ on adjectival verb
6
Unfortunately, (contextualized) data to determine whether ‘emphasized’ here
means (or could mean) something like ‘contrastive focus’ are not available.
7
Grüßner does not note a functional difference here – the use of the attributive,
thus, might provide a subtle pragmatic distinction.
124 North East Indian Linguistics Volume 3
(7) pé ke-lòk
cloth NMLZ-be.white
‘white cloth’ (Grüßner 1978: 124)
(8) ingnàr-kè a-só dīng+lèn dō.
elephant-TOP ATTR-tooth be.long exist
‘The elephant has long teeth.’ (Grüßner 1978: 124)
What will be interesting for the following section on relative clauses
is that both ACs have a fixed order, where the noun precedes the adjectival
verb. The same order in adjectival attribution can be found in Meithei
and Angami Naga.
Example (9) is from Meithei. The stative verb sa ‘be.hot’ is
nominalized, and, further, carries the prefix ə-, which is most likely to
be cognate to Karbi a-. This ə-, however, has a specialized function in
Meithei, only appearing on adjectives. Since -pə is the general nominalizer
in Meithei, occurring in all kinds of different nominalization constructions,
we could argue that ə- may be called an adjectivizer (given that it has
indeed no other function). Following this, it would make sense to posit
a separate structural category of adjectives. Of course, this should not
obscure the obvious morphological hierarchical pattern of these adjectives
in Meithei that identifies adjectives derived from nouns derived from
verbs.
(9) sadhəbiget əsabə páyrgə
sadhəbiget ə-sa-pə páy-lgə
certificate ATTR-be.hot-NMLZ hold-AFTER
‘…after securing a false certificate’ (Chelliah 1997: 165)
In Angami Naga, the nominalizer
kə̀- can prefix either onto the verb
or onto what is glossed as ‘verbal marker’ by Herring (1991), apparently
something like an auxiliary verb. As the presence of a ‘verbal marker’
is a characteristic of verbs, the nominalization mēziē kə̀tá in (11) appears
less nominal, and, in that, adjectival attribution in Angami Naga resembles
the indefinite AC in Karbi.
(10) kí [kə̀-sa] ù …
house NMLZ-be.new the
‘the new house…’ (Herring 1991: 57)
Functions of Nominalization in Karbi 125
(11) tèpf [mēziē kə̀-tá] ù
dog be.tired NMLZ-VM the
‘the tired dog…’ (Herring 1991: 58)
4. Nominalization in Relative Clauses
Relative clauses (RCs) are functionally similar to adjectives in that both
serve to modify a head noun. There is also a general structural similarity
in languages like Karbi, Meithei, and Angami Naga, where ‘adjectives’
are nominalized verbs, and, likewise, RCs are formed with nominalized
verbs. In all three languages, the two constructions – adjectival attribution
and RCs – are kept structurally distinct by means of constituent
ordering
8
.
Table 2 Constituent Order within Adjectival Attribution vs. RCs
Position 1 Position 2
Adjectival Attribution Noun [
HEAD] Modifier [DEP]
Relative Clause Modifier [
DEP] Noun [HEAD]
The Karbi examples (12) and (13), where the modifying RC precedes
the head noun, can be contrasted with (5–8) above, where the modifying
derived adjectives follow the head noun.
(12) thē
ke-chō a-monít
fruit NMLZ-eat ATTR-man
‘the man who is eating the fruit’ (Grüßner 1978: 96)
(13) kám ke-klém-ráp thèk-thē a-tūm
work NMLZ-do-together know-NEG ATTR-people
‘those who cannot work together’ (Grüßner 1978: 96)
In Karbi, the RC constituent order is the same as in possessive
constructions such as (14).
8
Note, however, that Meithei has prenominal and postnominal adjectives: “Post-
nominal adjectives are simply descriptive but prenominal adjectives identify
the modified noun as belonging to a class of similar things” (Chelliah 1997:
165). This seems to be a functional difference between a non-restrictive and
restrictive reading, similar to the difference between (pronominal) RCs and
(postnominal) ‘adjectives’ in Angami Naga (see the following discussion).
126 North East Indian Linguistics Volume 3
(14) rechó a-hēm
king ATTR-house
‘the house of the king’ (Grüßner 1978: 52)
Unlike Karbi, which has two different adjective constructions but
only one RC type, Angami Naga employs the same construction for
adjectival attribution and RCs, and only distinguishes between the two
syntactically: (15) and (16) below structurally contrast with (10) and (11)
only by the ordering of head noun and modifier.
(15) [ki
́-nù kə̀-bá] tèpf ù …
house-LOC NMLZ-VM dog the
‘the dog that is in the house …’ (Herring 1991: 58)
(16) [mhà lē kə̀-bá] thēmie̠ ù …
thing think NMLZ-VM man the
‘the man who thinks…’ (Herring 1991: 57)
Since the same verbs qualify for occurrence in the ‘adjective’ slot
and the RC verb slot in Angami Naga, syntactic minimal pairs are
possible, which trigger a functional difference between a rather “inherent
or internal” interpretation of the attributed state for the adjectival syntax,
and a prototypically “circumstantial or external” interpretation for the
RC syntax (Herring 1991: 58)
9
.
In the case of Meithei, it is not only reversed constituent order that
distinguishes RCs from adjectival attribution, but – as shown above –
there actually seems to be a morphologically marked structural category
of adjectives as derived by the ə- prefix. The Meithei RC construction
in (17) is to be contrasted with that in (9).
(17) ŋəsi čtpə nupadudi
[…]
ŋəsi čt-pə nu-pa-tu-ti
today go-NMLZ person-MASC-DDET-DLMT
‘the boy who is going today […]’ (Chelliah 1997: 157)
9
For example, (11) as it is will be interpreted as ‘a dog which is always tired,’
if the modifying part
[mēziē kə̀-tá] ‘be.tired NMLZ-VM’ came before the head
noun (yielding an RC), the whole NP would be interpreted as ‘a dog which
just happens to be tired (because it was just running around).’
Functions of Nominalization in Karbi 127
5. Nominalization in complement clauses
The following introduction to complement clauses (CCs) in Karbi,
Meithei, and Angami Naga will present evidence of nominalized CCs in
all three languages. A more detailed look at Karbi reveals some of the
variation in CC types beyond nominalization.
First, nominalization may be optional with the same complement-
taking verb such as mò ‘hesitate’, as (18) and (19) show. Note that -jí
appears to be obligatory, a marker of irrealis or future.
(18) k-arjū-jí mo
̀-ló.
NMLZ-ask-IRR hesitate-PAST
‘He hesitated to ask.’ (Grüßner 1978: 125)
(19) thán dàm-jí mò-ló.
tell go-IRR hesitate-PAST
‘He hesitated to tell (start telling).’ (Grüßner 1978: 89)
Further, a nominalized CC is also found with pa+ngch
èng ‘start’ as
in (20), and with thèk-lóng ‘see-discover’ as in (21). In both examples,
an overt S argument
10
is expressed within the CC, and, like (18) and
(19), both complement verbs suffix -jí ‘IRR.’
(20) arwè ke-ja
́ng-jí pa+ngchèng-ló.
rain NMLZ-fall-IRR start-PAST
‘It started raining.’ (Grüßner 1978: 125)
(21) a-pō ke-wàng-jí thèk-lóng-sí.
ATTR-father NMLZ-come-IRR see-discover-EMPH
‘She saw that her father was coming.’ (Grüßner 1978: 180)
Another type of CC construction in Karbi is the juxtaposition of main
verb and complement verb without nominalization at all. In (22), ingt
úng
‘wish; want’ appears to form a syntactic union with chō ‘eat’ in this
same-subject desiderative construction: the classifier -dùk ‘soft’, referring
to the fruits, suffixes onto ingtúng, and thus the whole complex predicate,
and not onto the semantically connected verb chō.
(22) bo
́kurí a-thē chō ingtúng-dùk má.
Bokuri ATTR-fruit eat wish-CLF:soft INTR
‘Do you want to eat Bokuri fruits?’ (Grüßner 1978: 89)
10
The term “S argument” refers to the single argument in an intransitive clause.
128 North East Indian Linguistics Volume 3
An even more integrated CC construction is given in (23), where the
‘complement-taking verb’ has become just a suffix on the ‘complement
verb’, which is the light verb klém ‘do’ here.
(23) a-k
ám klém-langláng-bōm-ló.
ATTR-work do-try-PROG-PAST
‘They were trying to do the work.’ (Grüßner 1978: 122)
So in Karbi, CCs vary between being nominalized – and, thus, being
embedded as a core argument into a superordinate main clause – and
having come to be the main clause themselves, with the notional
complement verb being a complex predicate.
Examples (24) from Meithei and (25) from Angami Naga show
nominalized CCs in both languages, parallel to Karbi (18; 20–21)
above.
(24)
nəŋ čtpə dərkar oy.
nəŋ čt-pə dərkar oy
you go-NMLZ duty is
‘It is necessary for you to go.’ (Chelliah 1997: 170)
(25) [thèvo̠-tshə tsə̠ pə̀krā kə̀-tsē] kə̀mìciè.
pig-flesh eat a.lot NMLZ-VM danger
bá.
have
‘[Eating too much pork] is dangerous.’ (Herring 1991: 58)
6. Nominalization in Adverbial Subordination
Nominalization is also used for adverbial clauses in all three languages.
Table 3 summarizes which semantic adverbial clause type is nominalized
(‘+NMLZ’), which is non-nominalized (‘-NMLZ’), and for which the
language has a nominalized and a non-nominalized construction (‘+/-
NMLZ’).
Even though Table 3 is only a very rough and preliminary sketch of
where nominalization figures in adverbial subordination in the three
languages, some important questions can be raised:
Is consistent nominalization in all types of adverbial subordination
(as in Meithei) the older pattern, and did Karbi and Angami develop
finite constructions more recently? Is there a semantic motivation, or
Functions of Nominalization in Karbi 129
actually a semantically motivated hierarchy of pairings of adverbial
subordination type and syntactic construction (i.e., nominalized or non-
nominalized)? What is the semanto-pragmatic difference between a
nominalized and a non-nominalized construction for the same
subordination type?
This last question asks about the functional difference between two
constructions like (26) and (27) below. The latter is nominalized, the
former is not, but both are purposive clauses. The subordinator in (26)
is the suffix -jí+sí, whereas in (27) it is the postposition a-phān that
otherwise follows (NPs containing) lexical nouns as in (28).
(26) lāng náng ch+inglú-jí+sí wàng-ló.
water PART
11
bathe-in.order.to come-PAST
‘They came in order to take a bath in the water.’ (Grüßner 1978:
102)
(27) àn ke-chō-jí a-phān.
rice NMLZ-eat-FUT ATTR-GOAL
‘in order to eat rice’ (Grüßner 1978: 80)
Table 3 Nominalization patterns in adverbial subordination
Adverbial semantics Karbi Angami Meithei
while +/-NMLZ +/-NMLZ + NMLZ (+ASSOC)
as soon as
- NMLZ - NMLZ
before
+ NMLZ
after
+/-NMLZ + NMLZ + NMLZ (+LOC)
as long as
- NMLZ + NMLZ
until
- NMLZ - NMLZ
if
+/-NMLZ - NMLZ + NMLZ (+DLMT)
even though
- NMLZ + NMLZ (+’also’)
because
+ NMLZ + NMLZ + NMLZ (+ABL/INST)
in order to
+/-NMLZ + NMLZ + NMLZ (+GEN)
11
The function of náng is not clear and has yet to be determined. Grüßner (1978:
133) preliminarily describes it as an ethical dative (derived from
nàng
‘2.
SG’).
130 North East Indian Linguistics Volume 3
(28) chiklō jōn kép a-phān
month CLF ten ATTR-GOAL
‘for ten months’ (Grüßner 1978: 80)
Here, an internal reconstruction of the subordinating suffix
-jí+sí would be very interesting. The first part is very likely to be the
suffix -jí ‘IRR’ (which has already shown up in the discussion of CCs).
And, actually, the second part, the suffix -sí, can also be found on lexical
nouns as in (29), where it is described as an emphatic suffix by Grüßner
(1978: 62) conveying contrastive stress and/or translating into English
‘only’ or ‘just’.
(29)
nàng òk-sí chō-pò.
2.SG.POSS meat-EMPH eat-FUT
‘They will eat only your meat.’ (Grüßner 1978: 62)
Thus – if -s
í actually is a nominal suffix – the seemingly non-
nominalized construction in (26) might actually be just as well
nominalized in a different type of nominalization construction.
7. Nominalization in Main Clauses
The last function of nominalization to be addressed here is main clause
marking. In Karbi, kV- on verbs in main clauses expresses the progressive
aspect according to Grüßner (1978: 95).
(30) nè ka-chi+r
ú-ló.
1.SG NMLZ-cry-PAST
‘I was crying.’ (Grüßner 1978: 95)
(31) konát-sí nàng-lì ke-dàm-jí.
where-DIR 2.SG-HON NMLZ-go-INTENT
‘Where are you going?’ (Grüßner 1978: 95)
For Angami Naga, neither Herring (1991), nor Hutton (1921), Giridhar
(1980), or Giridhar (1991) report such a function of the
kə̀- prefix.
However, if we look at (32), we see that there seems to be a progressive
marker -wē in Angami Naga. This is interesting, because there is a
nominalizer or copula *way reconstructed for Proto-Tibeto-Burman (see
Matisoff 2003: 209). Therefore, nominalized main verbs may actually
exist in Angami Naga, and perhaps employ the same function of
Functions of Nominalization in Karbi 131
progressive marking as in Karbi – even though it is not the putatively
cognate nominalizer kV- that carries out this function.
(32) ā pu-wē.
I speak-?? < *way
‘NMLZ/COP’?
‘I speak; I am/was speaking.’ (Hutton 1921: 306)
Chelliah (1997: 297) notes that the suffix sequence -pəni (the
nominalizer -pə plus the copular suffix -ni) can occur on main verbs.
This is structurally similar to Karbi and Angami (with the difference of
the preserved overt copula), but functionally, main verb nominalization
in Meithei undertook a different grammaticalization path; it has become
an evidential marker.
[…] where a nominalized clause is made into a copulative sentence, the
speaker indicates knowledge about the truth of a proposition […] based
on sensory evidence […], predictable habitual activity[…] or a gener-
ally known fact […]. (Chelliah 1997: 297)
An example of this nominalization evidential construction is shown
in (33).
(33) má ŋayhák tumməkpəni.
má ŋay-hák tum-lək-pə-ni
he wait-here sleep-DIST-NMLZ-COP
‘(I could swear that) up to now he was here sleeping.’ (Chelliah
1997: 297)
Conclusion
Table 4 summarizes the patterns of nominalization across the three
languages and across four of the six constructions discussed above.
Excluded from the table is the section on derivational nominalization,
where data only from Karbi were provided, as well as the last section
on main clause nominalization, as the progressive marking in Karbi and
possibly Angami cannot be readily compared to the evidential marking
in Meithei.
We see that only Meithei consistently uses nominalization for all
functions. In both Karbi and Angami, there are non-nominalized adverbial
clause types, and, likewise, some Karbi complement clause types are not
nominalized. An interesting line for future research would be to study
132 North East Indian Linguistics Volume 3
the potential pragmatic or sociolinguistic motivations for choosing one
construction over another.
Regarding the modifier constructions, all three languages distinguish
between one-word modifiers (‘adjectives’) and clausal modifiers (‘relative
clauses’) by means of word order, as ‘adjectives’ consistently follow
their head noun, but ‘relative clauses’ precede it. Whereas this difference
in constituent order is the only distinction Angami makes, both Karbi
and Meithei also have separate morphological means of marking this
distinction.
Inventories of nominalization constructions like this (still preliminary)
one for Karbi, Meithei, and Angami Naga provide evidence for typological
characteristics of nominalization in TB languages. Future research on
languages of the family will clarify what constructions are recurrent, and
which ones are rather idiosyncratic. For example, the existence of two
adjectival constructions, as we find it in Karbi, seems to also exist in
Dimasa (a Bodo language from NEI), in an apparently parallel fashion,
both in form and in function (see Jacquesson 2008: 36). Furthermore,
the pattern of distinguishing between adjectival and RC modification
constructions by means of constituent ordering appears to be recurrent,
and has been analysed by DeLancey (1994) based on Tibetan. Arriving
at a better understanding of the role of nominalization in complement
and adverbial clauses might be a more distant goal of future research
considering the time-consuming nature of going through the different
semantic types. Still, each construction will tell us more about the
(development of) functions of nominalization in TB, and so will knowing
Table 4 Summary
Construction type Karbi Meithei Angami Naga
Modifier
Adjectival (1) +NMLZ plus a- +NMLZ +NMLZ
attribute (2) +/-NMLZ plus ə- [word order]
[word order] [word order]
Relative +NMLZ plus a- +NMLZ +NMLZ
clause [word order] [word order] [word order]
Core argument Complement (1) +
NMLZ +NMLZ +NMLZ
clause (2) –NMLZ (biclausal)
(3)
–NMLZ (unified)
Oblique Adverbial (1) +
NMLZ +NMLZ (1) +NMLZ
argument clause (2) –
NMLZ (2) –NMLZ
(3) +/-
NMLZ (3) +/-NMLZ
Functions of Nominalization in Karbi 133
about recurrent patterns and distributions of those functions across the
branches of the family.
Abbreviations
: Additional function/sense (in portmanteau-morphemes
- Morpheme boundary
+ (Former) morpheme boundary between lexicalized
combinations
. Used in gloss if more words are needed for translation
1 First person
2 Second person
ABL Ablative
AFTER After V-ing
ASSOC Associative
ATTR Attributive
CLF Classifier
COP Copula
DDET Distance Determiner
DEM Demonstrative
DEP Dependent
DIM Diminutive
DIR Directional
DIST Distal
DLMT Delimitative
EMPH Emphasis
EX Exclusive
FUT Future
G Generic
GEN Genitive
GOAL Goal and primary object marker
HON Honorific
INST Instrumental
INTENT Intentional
INTR Interrogative
IRR Irrealis
LOC Locative
MASC Masculine
MID Middle voice
134 North East Indian Linguistics Volume 3
NEG Negation
NMLZ Nominalizer
PART Particle
PAST Past Tense
POSS Possessive
PROG Progressive
SG Singular
SPLT Superlative
TOP Topic
VM Verbal Marker
References
Burling, R. (2003). “The Tibeto-Burman languages of Northeastern India.” In
Thurgood, G. & R. J. LaPolla, Eds., The Sino-Tibetan Languages. London,
Routledge: 169–91.
Chelliah, S. (1997). A Grammar of Meithei. Berlin/New York, Mouton de
Gruyter.
DeLancey, S. (1987). “Sino-Tibetan languages.” In Comrie, B., Ed., The World’s
Major Languages. London, Croom Helm: 797–810.
_____. (1994). “Grammaticalization and linguistic theory.” Proceedings of the
1993 Mid-America Linguistics Conference and Conference on Siouan/
Caddoan languages. Boulder, Dept. of Linguistics, University of Colorado:
1–22.
Giridhar, P. P. (1980). Angami Grammar. Mysore, Central Institute of Indian
Languages.
_____. (1991). “On the word in Angami Naga.” Linguistics of the Tibeto-Burman
Area 14(1): 1–54.
Grüßner, K.-H. (1978). Arleng Alam, die Sprache der Mikir: Grammatik und
Texte. (Beiträge zur Südasienforschung, Bd. 39). Wiesbaden, Franz Steiner
Verlag.
Herring, S. C. (1991). “Nominalization, relativization, and attribution in Lotha,
Angami, and Burmese.” Linguistics of the Tibeto-Burman Area 14(1): 55–
72.
Hutton, J. H. (1921). The Angami Nagas. With Some Notes on Neighbouring
Tribes. London, MacMillan & Co.
Jacquesson, F. (2008). A Dimasa Grammar. Unpublished manuscript. Available
at: http://brahmaputra.vjf.cnrs.fr/bdd/IMG/pdf/Dimasa_Grammar-2.pdf.
Matisoff, J. A. (2003). Handbook of Proto-Tibeto-Burman – System and
Philosophy of Sino-Tibetan Reconstruction. Berkeley/Los Angeles,
University of California Press.
Topographical Deixis and the Tani Languages of North East India 135
Tani
136 North East Indian Linguistics Volume 3
136 blank
Topographical Deixis and the Tani Languages of North East India 137
9
Topographical Deixis and the Tani Languages of
North East India
1
Mark W. Post
The Cairns Institute, James Cook University
1. Introduction
Many if not most Tibeto-Burman languages exhibit a means of referencing
the location or trajectory of a referent in terms of one of three
topographically-oriented planes: upward, downward and on the same or
an unknown level (as a shifting reference point). A brief example from
Lare Galo (Tibeto-Burman > Tani; Arunachal Pradesh) illustrating use
of an upward distal demonstrative is first given in (1).
(1)
tə̂ə hɨɨrêk tə̂ə dà, japóm əcìn rədâk
tə̀ hɨɨrèk tə̀ da japóm ə=cìn r-dàk
DST.UP banyan DST.UP CNTR demon TOP=ADD exist-COS
1
I gratefully acknowledge the assistance of my primary Galo consultants for
this paper `Igoo Rwbaa,
`Ilww Rwbaa, and `Miilww `Xodu, my Mising
consultants Avn Dole:, Polobu Dole:
and Jugendra Pe:gu: and my Apatani
consultants Habung Taki and Nending Jorang. I am also indebted to Stuart
Blackburn and Pascal Bouchery for helping me to better understand the situation
in Apatani. This paper was written while on fieldwork in Arunachal Pradesh,
for which a grant was provided by the Research Centre for Linguistic Typology
at La Trobe University.
138 North East Indian Linguistics Volume 3
bentù.
bèn=tu
EVID=AURV
‘As for that banyan up there, there certainly also now seems to
be an evil spirit living (in it).’
In this paper, I will introduce the term topographical deixis to describe
this sort of system as it is found in the Tani languages, which constitute
a medium-sized branch of Tibeto-Burman languages spoken primarily
in Arunachal Pradesh and Assam, North East India, as well as in small
numbers in Tibet (Sun 1993). Although a survey of topographical deixis
and related systems in Tibeto-Burman has been initiated and remains
ongoing (Post in preparation), in this paper I will mainly restrict
discussion to an explication of the system as it operates in those Tani
languages which exhibit it. The paper is thus primarily descriptive in
intent, and may also, I hope, be used as a benchmark reference for the
study of similar and possibly related systems in other North East Indian
languages.
The remainder of the paper has the following structure: §2 discusses
the environmental context of the Tani languages. In §3, we outline the
basic characteristics of a system of topographical deixis, also discussing
its hypothetical origins and semantic extensions. §4 presents a detailed
overview of the system of topographical deixis found in Lare Galo, based
on the recent description of Post (2007), and §5 discusses the status of
topographical deixis in other Tani languages.
2. The environmental context of the Tani languages
The majority of modern-day Tani languages are spoken in and around
central Arunachal Pradesh state, North East India. The locus of highest
intra-Tani linguistic diversity is found in north-central Arunachal Pradesh,
which may well be the approximate location of the earliest reconstructible
Tani “homeland” (Post forthcoming 2010). Located roughly between 27
and 29 degrees latitude, central Arunachal Pradesh sits squarely within
the Eastern Himalaya, where altitudes range between highs of around
5,000–6,000 metres above sea level, at and beyond the Tibet border, all
the way down to 200–300 metres at the border with Assam and the
Brahmaputra floodplain.
Topographical Deixis and the Tani Languages of North East India 139
Mountain topography shapes almost every aspect of the traditional
Tani lifestyle. Most Tani-speaking villages and towns in central Arunachal
Pradesh are located at altitudes of between 500–2,000 metres, usually
on sloping hillsides near to, but not within, river valleys (Figure 1).
Traditional agriculture is centred on the practice of jhum, or shifting,
hillside cultivation. Jhum fields are often located above or adjacent to
villages, on mountain slopes which would be too steep to settle, and
which are often barely traversable to a novice. Terraced wet rice fields
were traditionally not widely cultivated by most Tani tribespeople,
however efforts to boost productivity have led to their introduction in
several areas.
There are, however, at least two important exceptions to the above
description:
First, within Arunachal Pradesh, the Apatani are the only Tani group
who have traditionally lived almost exclusively within a plateau area.
Legendary for their high agricultural productivity, Apatani devote nearly
every square inch of land available in their plateau to intensive wet rice
cultivation-cum-pisciculture, i.e., paddy-based fish farming (von Fürer-
Figure 1 rɨlúu (Galo) village, West Siang district, Arunachal Pradesh, with
moodi
̀i-rɨk (jhum fields) in background and terraced isì-rɨk (wet fields) in
foreground
140 North East Indian Linguistics Volume 3
Haimendorf 1980). Densely-arranged villages ring the large plateau,
while surrounding hills are primarily employed as plantation grounds,
where forest resources such as bamboo and firewood are obtained;
hillsides are not generally cultivated by Apatani (Figure 2).
A second important exception is that of the largest single population
of Tani tribespeople, the Mising, who are found almost exclusively within
the Brahmaputra floodplain of Upper Assam. Topographically quite unlike
most parts of central Arunachal Pradesh, the Upper Assam plains lack
hills of almost any size, permit only wet rice agriculture (supplemented
by kitchen gardens and small forest plantations intermixed with villages),
and undergo extensive flooding sometimes for several months in a year
(Figure 3).
In addition to those Tani language speakers currently found in North
East India, small numbers are also found in Tibet. Very little current
information of any reliability is available regarding their distribution or
environmental context, although a few remarks may be found in Sun
(1993).
3. Topographical deixis
Deixis may be understood quite generally as a set of “ways in which
languages encode…features of the context of utterance” (Levinson
Figure 2 Ziro, Apatani plateau; Lower Subansiri district, Arunachal Pradesh
Topographical Deixis and the Tani Languages of North East India 141
1983: 54). As Senft (1997a: 2) writes, “when we communicate, we
communicate in a certain context, and this context shapes our
utterances”.
In a system of topographical deixis, referents’ locations or trajectories
are identified in spatial terms, as being upward of, downward of or on
the same (or an unknown) level as a shifting “deictic” centre (often,
though not always, the place of speaking). Similar or identical systems
are quite common in Tibeto-Burman, and have been fully or partially
described from various perspectives as “environmental space deixis”
(Bickel 2001), “altitudinal case markers” (Ebert 2003) and “vertical case”
(Noonan 2006). Outside the Tibeto-Burman area, languages which exhibit
some form of topographical deixis (whether as the primary or secondary
basis for some system of spatial reference) are common in Papua New
Guinea and Oceania (Senft 1997b; 2004), and seem to occur in some
form in Uto-Aztecan and in Dyirbal (see discussion in Dixon 2003).
3.1. Environmental source
The ultimate source of topographical deixis, as well as its prototypical
domain of reference, would appear to be human interaction with a
Figure 3 View of Upper Assam from the West Siang district border, with
Mising and Bodo-Kachari villages in distance
142 North East Indian Linguistics Volume 3
topographically varied environment such as that roughly schematized in
Figure 4 (to be compared with Figure 1).
3.2. Environmental extensions
From the topographical source illustrated in Figure 4, a number of
environmental extensions are also observed. The first is riverine
orientation: in this extension, upward is transferred to upriver, downward
to downriver, and same/unknown level is effectively extended to mean
not along a river course (Figure 5).
Upward
Same/unknown
level
Same/unknown
level
Downward
Figure 4 Environmental source of topographical deixis
Not along
river course
Uprive
r
No
t along
river course
Downriver
Figure 5 Environmental extension 1: Riverine orientation
The second extension is compass orientation. In this extension,
upward is transferred to northward, downward to southward, and same/
unknown level to east/west/unknown direction (effectively, neither north
nor south; Figure 6).
Topographical Deixis and the Tani Languages of North East India 143
The reasons why these particular extensions are possible seem
straightforward: both reflect what are virtually necessary interpretations
of the central Arunachali environment (§1). Just as it is the case that
altitude decreases in this area as one moves southward – but remains,
overall, more or less the same as one moves from east to west – it is
also (necessarily, therefore) the case that all Arunachali rivers ultimately
flow from north to south; hence the syncretism. For two telling exceptions
to this generalization for the Tani languages, however, see §5.
4. Topographical deixis in Galo
In Galo, topographical deixis is found in two primary linguistic domains,
with a variety of sub-domains. Within the predicate domain, topographical
deixis is encoded by verb roots and by directional
predicate derivations
(§4.1). Within the noun phrase domain, topographical deixis is encoded
by demonstratives and demonstrative postpositions (§4.2)
2
.
Neither North nor South
South
Neither North nor South
North
N
E
S
W
Figure 6 Environmental extensions 2: Compass orientation
2
Descriptions of topographical deixis in other Tibeto-Burman languages are
often not sufficiently detailed to give a clear sense of how widely this particular
set of domains is distributed; however, it is striking to find that Belhare – an
Eastern Kiranti language of Nepal which is not usually thought to be particularly
close to Tani within Tibeto-Burman – seems to share exactly this fundamental
set, albeit with some further complexities (a vowel alternation encodes an
additional dimension of deictic displacement, which is not found in Tani)
(Bickel 2001). This perhaps unexpectedly close correspondence – which also
involves at least some cognate formatives – could in principle support a theory
that topographical deixis reconstructs quite deeply within Tibeto-Burman. This
possibility is currently under investigation.
144 North East Indian Linguistics Volume 3
4.1. Predicate domain
The set of Galo verb roots encoding topographical deixis are presented
in Table 1. An example of use is in (2). Note that one of the three verb
roots is polysemous with two other deictic domains (motion towards a
reference point and motion into an enclosed space); note also that
although the English translations are necessarily somewhat elaborate,
these are among the most frequent and ordinary verbs used for expressing
motion to a goal in Galo.
Table 1 Galo verb roots encoding topographical deixis
Function Form Gloss
MOTION UPWARD/UPRIVER/NORTHWARD càa- ‘ascend’
MOTION DOWNWARD/DOWNRIVER/ SOUTHWARD ìi- ‘descend’
MOTION ON THE SAME OR UNKNOWN LEVEL/ áa- ‘come; enter; move
NOT ALONG RIVER COURSE/EAST/WEST laterally’
Table 2 Galo directionals encoding topographical deixis
Function Form Gloss
DIRECTION UPWARD/UPRIVER/NORTHWARD -càa ‘GOAL:UP’
DIRECTION DOWNWARD/DOWNRIVER/SOUTHWARD -lòo ‘GOAL:DOWN’
DIRECTION ON THE SAME TOPOGRAPHICAL LEVEL/
NOT ALONG RIVER COURSE/EAST/WEST OR -áa ‘GOAL:SLEV’
UNKNOWN TRAJECTORY
(2) caarûu làp mər ə̀m...
càa-rûu-lapə̀ mə-r=ə̀m
ascend-DEF-CTZR:PURP think-IRR=ACC.TSUB
caalâa r.
càa-làa-r
ascend
-ABIL-IRR
‘If they definitely want to move in
3
, they’ll be able to.’
The set of directional predicate derivations (“directionals” for short)
encoding topographical deixis in Galo are presented in Table 2. The
3
‘Move in’ here refers to ascending the ladder of a newly-built house for the
first time.
Topographical Deixis and the Tani Languages of North East India 145
function of a directional is to modify a predicate by specifying the
direction or trajectory of an activity verb (such as a verb of throwing)
or transferred location of a state (for example, when the location in which
a state is said to obtain becomes different from where it may be said to
have previously obtained). Directionals in at least some Tani languages
are also applicative, in that they additionally function to add an obligatory
(usually locative-marked) goal argument to the clause. In Table 2, note
also that two out of three forms are cognate with the directional verb
roots in Table 1; this would suggest that directionals have an ultimate
source in motion verbs, but that since the sharing of forms is not complete,
the relationship may no longer be active
4
.
An example of use is in (3). Note here that the obligatory GOAL
argument hogò ‘(to) here’ is licensed by the (applicative) directional -lòo
‘GOAL:DOWN’; in absence of -lòo, hogò could only be understood as an
optionally-mentioned location at which ‘fleeing’ simply happens to take
place.
(3) hodu
̂m-hor hòg
[hodùm-hor]
S
[hogò]
GOAL
barking.deer-boar SPRX.LOC
keŋŋám loodûu kutú!
[kéK-ŋám-lòo-dùu-kú]
PRED
=tu
flee-EXH-GOAL:DOWN-CMPL=AURV
‘The game animals are all fleeing down here, don’t you see
(because there’s a tiger roaming about up in the mountain)!’
4.1. Noun phrase domain
Within the noun phrase, Galo distal demonstratives encode topographical
deixis (Table 3); note that they do not seem to be cognate with forms
found in the predicate domain
5
. As in most Tibeto-Burman languages,
Galo demonstratives function both as pronominals and as referential
modifiers. An example of pronominal use is in (4) (also cf. (1) above).
4
See also Post (2010) for a general discussion of the etymologies of Tani
predicate derivations, including directionals.
5
à ‘DST.SLEV’ and -áa ‘GOAL:SLEV’ may at first look promising, however the
Proto-Tani forms *va
ŋ ‘come’ and *a(a) ‘distal demonstrative’ (Sun 1993)
suggest otherwise.
146 North East Indian Linguistics Volume 3
(4) bə̂ə jəkkə̂ə làa?
bə̀ jə̀-kə̀=ə làa
DST.DOWN who-GEN=COP.IPFV CQ
‘Whose is that (thing down there)?’
In addition to distal demonstratives, a very wide range of relational
markers employ topographical deixis in Galo (Table 4). Quite clearly
etymologically complex, they also usually include a distal demonstrative
as one of the incorporated formatives. However, since their phonological
forms involve vowel harmony processes which are not generalizable
beyond a subset of any given paradigm (and which therefore cannot be
predicted by a synchronic rule), they are treated as morphologically non-
compositional in terms of modern Galo.
Relational markers employing topographical deixis encode a variety
of relational functions, including locative/allative, ablative/partitive, and
genitive/semblative. Due to their ability to either precede the noun, follow
the noun, or occur on both sides of a noun – like Galo demonstratives
– together with their ability to mark relational functions – like Galo
postpositions – they are assigned the hybrid label demonstrative
postposition
6
. A set of manner demonstratives seems to share a similar
origin and structure and is also included here; however, the manner
demonstratives cannot precede a head noun, unlike all simple
demonstratives and demonstrative postpositions. Examples of use are in
(5)–(7). In (7), note the parallel use of a ‘same level’ demonstrative
postposition and ‘same level’ motion verb.
Table 3 Galo distal demonstratives encoding topographical deixis
Function Form Gloss
LOCATION UPWARD/UPRIVER/NORTHWARD tə̀ ‘DST.UP’
LOCATION DOWNWARD/DOWNRIVER/ SOUTHWARD bə̀ ‘DST.DOWN’
LOCATION ON THE SAME TOPOGRAPHICAL LEVEL/
NOT ALONG RIVER COURSE/EAST/ WEST OR à ‘DST.SLEV’
UNKNOWN TRAJECTORY
6
The statistically most common position is post-head; it is likely that ability to
occur pre-head (and to
“bracket” the head simultaneously) owes to earlier
grammaticalization of a topic-comment construction in which a pronominal
head is marked for relationality; schematically, something like:
“(as for) up
there, the mountain up there…” → “that there mountain up there…”.
Topographical Deixis and the Tani Languages of North East India 147
(5) ŋôk tòk ˀêz
i
go.
ŋó-kə̀ tokə̀
ˀezə̀=go
1.SG-GEN DST.ABL.UP clothing=IND
lâaz
i
takèe!
làa-zí-tà(a)=kée
take-BEN-MOT=HORT.POL
‘Go get my shawl from up there!’
(6) ˀastrée lijáa bəmbə̀ kaí rə̀m, buppɨ̂ɨ.
ˀastreelijáa bəmbə̀ kaí-r=ə̀m buppɨ̂ɨ
PLACE DST.MAN.DOWN big-IRR=TSUB all
min rələ̀m paarûu r.
min
ərl=ə̀m pàa-rûu-r
mineral(<Eng)=ACC get-CERT-IRR
‘If Australia down there is so big, certainly you’ll find all (man-
ner of) minerals.’
(7) âl nám âl ŋó aakáa r.
alo
̀ nam alò ŋó áa-káa-r
DST.LOC.SLEV house DST.LOC.SLEV 1.SG come-TENT-IRR
‘I’ll go check that house over there.’
5. Topographical deixis in other Tani languages
The basic outlines of the system of topographical deixis described for
Lare Galo in §4 seems to be shared by the great majority of Western and
Eastern Tani languages, almost always employing clearly cognate
formatives; this would suggest that a system not very much different
from that discussed above must be reconstructible to the Proto-Tani stage
(and indeed, as suggested in fn 2, quite probably well beyond). Table 5
presents the sets of topographical-deictic demonstratives found in all
Tani languages for which any data are currently available to me.
Table 4 Galo demonstrative postpositions exhibiting topographical deixis
Function Upward Downward Same/unknown level
DST.LOC/ALL tolò bolò alò
DST.ABL/PART to(lo)kə̀ bo(lo)kə̀ a(lo)kə̀
DST.SEMB/GEN təkə̀ bəkə̀ akə̀
DST.MAN təmbə̀ bəmbə̀ ambə̀
148 North East Indian Linguistics Volume 3
There are, however, two interesting exceptions to the general trend
here: Apatani and (some dialects of) Mising.
Table 5 Distal demonstratives exhibiting topographical deixis in a sample
set of Tani languages (phonetic values are approximate in most cases)
Language Up Down Slev
Proto-Tani
7
*tà *bà *à
Bori (Megu 1988) taa
bə (?) aa
Galo (Post 2007)
tə̀ bə̀ à
Bokar (Sun 1994)
təə bəə aa
Bengni (Sun 1994)
tɯɯ bɯɯ ɯɯ
Tagin (Keepor Mara, p.c.)
tə bə a
Aashing (Megu 2003)
tə bə a
Ramo (Badu 2004a)
8
tolo bolo alo
Hills Miri (Simon 1976)
tə bə ə
Karko (Megu 1993)
tə- bə- ə-
Paadam (Tayeng 1983)
tə bə ə
Mising (Post field notes)
tə bə ə
Pasi (Yankee Modi, p.c.)
tə bə ə
Minyong (Post field notes)
tə̀ bə̀ ə̀
Nishi (Post field notes) te be
e
Tangam (Badu 2004b) te- me- e-
Milang (Post field notes) ʨə̀- mo
̀- jò-
Apatani (Simon 1972; Abraham 1985) – –
–
7
Although the two consonant initials are clear enough, the vowels are simply
placeholders; there is too much variation in the data – possibly in some cases
reflective of paradigm-interaction and/or irregular sound changes – to advance
anything with confidence at present. Vowel lengths in the Sun (1994) Bokar
and Bengni data appear to reflect regular operation of a Western Tani Bimor
-
aic phonological word constraint, which is generally not found in suffixed
forms.
8
The cited source lists the locative demonstrative postpositions, but no simple
demonstratives. It is, however, more than likely that a more complete description
would uncover corresponding simple demonstratives in this language.
Topographical Deixis and the Tani Languages of North East India 149
In Apatani (Tani > Western), topographical deixis appears to be well-
represented in the predicate domain. In the descriptions of Abraham
(1985; 1987) and Simon (1972) (dialect(s) not identified), we find
directionals -ca ‘move up to perform the act’ -yi ‘move down to perform
the act’ and -bo ‘move on level ground to perform the act’ (Abraham
1987: 123). Among motion verbs, Post (field notes) has attested four
verbs of seeming relevance in Apatani, càa- ‘ascend; move north’, tòʔ-
‘descend; move south’, ín- ‘go; move on the same level’ and bòo- ‘cross
(a river); move east or west’. All of the listed forms have solid cognates
with similar meanings in other Tani languages. However, when we turn
to the noun phrase grammar, we find no sign of topographical deixis in
Apatani at all. Demonstrative deixis in Apatani instead seems to
distinguish only proximity, and exhibits speaker-proximate, addressee-
proximate and speech-act-participant (SAP)-distal forms
9
.
The second exception is found in Mising (Tani > Eastern). Although
in Mising, a similar set of forms to that described in §4 is exhibited, in
some (not all) Mising dialects, a comprehensive shift in compass
orientation has been observed. In the subdialect of Pagro Mising which
I studied in the North Lakhimpur area, very near to the north bank of
the Brahmaputra River in Upper Assam, the upward forms also indicate
eastward (not northward), downward forms indicate westward (not
southward) and same/unknown level forms indicate north/southward (not
east/westward). A sample set is given in Table 6; examples illustrating
9
Although one may automatically suspect descriptive incompleteness here, these
observations have been confirmed by the folklorist Stuart Blackburn and the
social anthropologist Pascal Bouchery (in personal communication), each of
whom possesses an enviably rich knowledge of Apatani language and culture,
having independently conducted extensive field research in the area over several
years.
SAP-sensitivity is found in other Western Tani languages (it is rare in or
possibly absent from Eastern Tani), however it generally co-exists with a
system of topographical deixis, as is the case in, for example, Galo (see Ab
-
raham (1985: 31–35) for Apatani and Post (2007: 341–355) for Galo). Topo
-
graphical deictic demonstratives are also missing from some descriptions of
Tibet-area Tani languages such as Bengni and Bokar (Sun, Zhang et al. 1980;
Ou-Yang 1985), however since they were identified in the same languages by
Sun (1994), it does seem that descriptive incompleteness would apply in this
case. That said, if at least some dialects of any Tibet-area Tani languages could
be demonstrated to lack topographical deictic demonstratives, this fact would
both complicate and potentially enrich the present overview considerably.
150 North East Indian Linguistics Volume 3
use of the Pagro Mising set of topographical-deictic directionals are in
(8)–(10).
(8)
ədə-m gə-aa-lɨk-tok!
SDST-ACC throw/carry-SAME.LEVEL-APPL:INTO-IPTV
‘Throw it over/along a north/south axis!’
(9) ədə-m gə-saa-lɨk-tok!
SDST-ACC throw/carry-UPWARD-APPL:INTO-IPTV
‘Throw it upward/to the east!’
(10) ədə-m gə-jii-lɨk-tok!
SDST-ACC throw/carry-DOWNWARD-APPL:INTO-IPTV
‘Throw it downward/to the west!’
What accounts for these differences? In the case of North Lakhimpur
Pagro Mising, the explanation would appear straightforward. As shown
in Figure 2, there are virtually no hills of any size in the Pagro area,
quite in contrast with the almost uniformly mountainous terrain of central
Arunachal Pradesh (cf. Figure 1). Instead, the overwhelmingly dominant
environmental feature is the Brahmaputra river, which flows from east
to west. Accordingly, it would appear that the cognitive anchor of the
system (on which compass orientation is predicated) has shifted from
topography to riverine orientation. To the extent that Tani oral histories
– which uniformly point to a historical Mising migration from the hills
of Arunachal down to the plains of Assam – have any validity, this would
appear to represent a clear instance of linguistic adaptation to shifting
environmental circumstances
10
.
Table 6 Sample set of terms employing topographical deixis in Pagro Mising
Upward/East Downward/West Same-level/
North or South
Motion verbs
saa- tok- aa-
Directionals -saa -jii -aa
LOC/ALL Dem. Pos. tolo bolo olo
10
Jugendra Pe:gu: (in personal communication) reports that this shift has not
taken place in all Mising dialects, some of which continue to resemble the
Galo system described in
§4. I have not yet had the opportunity to study any
non-shifting Mising dialects myself.
Topographical Deixis and the Tani Languages of North East India 151
The absence of topographical deixis in the Apatani noun phrase
grammar may have a similar explanation. Apatani – unique among the
Tani languages – is spoken almost exclusively within a plateau area,
where wet-rice-cultivation-cum-pisciculture is well-known to have been
practiced over a wide and flat expanse of land since well before modern
times (von Fürer-Haimendorf 1980). It may be that in the Apatani area,
there was neither a topographical factor nor a riverine factor of sufficient
saliency to anchor some aspects of the system, whereby the inherited set
of topographical deictic forms may have partially disappeared from the
language. Why only in the noun phrase, and not the verb phrase? The
only truthful answer is that we don’t know. However, if one is to
speculate, it may reflect a possible fact that systems describing upward
and downward motion are more easily anchored by human social and
environmental interaction than are systems employed for upward and
downward reference. Note that the former occurs in (or is based on) a
large, open class of lexemes (memorized in a large list, just like English
‘ascend’ and ‘descend’) whereas the latter occurs in a small, closed class
of functional words, which both arise and coalesce due to, and are
subsequently maintained primarily by virtue of, discourse-frequency
(Bybee and Hopper 2001). In the mountainous central Arunachali
environment, awareness of the relative topographical situation of referents
is virtually inescapable, whereas in the Apatani plateau, opportunities to
describe contrasting heights of referents would have been relatively much
rarer. Note that in Mising, by contrast, speaker awareness of the relative
riverine orientation of referents would have been relatively robust. This
evidently led to the formal conservation of topographical deixis even as
its semantic contents were adapted to reflect novel environmental
realities.
Conclusion
Topographical deixis has been described in this paper as a unified
cognitive system underlying the grammars of most Tani languages, which
is probably reconstructible to Proto-Tani, and indeed beyond. It is
represented in four linguistic domains: verb roots, directionals, distal
demonstratives and demonstrative postpositions. While the principal (and
probably in some sense “original”) basis for the system is viewed as
topographical, extensions are also observed in the areas of compass and
riverine orientation. When a salient topographical anchor to the system
152 North East Indian Linguistics Volume 3
is no longer available, a salient riverine anchor to the system may “take
over.” When no anchors of sufficient salience are available, the system
may partly erode or perhaps obsolesce entirely. Accordingly, topographical
deixis would appear to provide good evidence of the shaping of diverse
linguistic structures by means of human interaction with diverse geo-
topographical environments; however, it probably does not make a good
genetic sub-grouping criterion. Further research will address these and
similar topics in a broader Tibeto-Burman context.
Abbreviations
ABIL Ability
ABL Ablative
ACC Accusative
ADD Additive
ALL Allative
APPL Applicative
AURV Auto-Revelative
BEN Benefactive
CERT Certainty
CMPL Completive
CNTR Contrastive
COP Copula
COS Change of State
CQ Content Question
CTZR Complementizer
DAT Dative
DEF Definitive
DST Distal
EVID Evidential
EXH Exhaustive
GEN Genitive
HORT Hortative
IND Individuative
IPFV Imperfective
IPTV Imperative
IRR Irrealis
LOC Locative
MAN Manner
Topographical Deixis and the Tani Languages of North East India 153
MOT Motion
PART Partative
POL Polite
PRED Predicate
PURP Purpose
S Intransitive subject
SDST Speech-act-distal
SEMB Semblative
SG Singular
SLEV Same-level
SPRX Speaker-Proximate
TENT Tentative
TOP Topic
TSUB Temporal subordinator
References
Abraham, P. T. (1985). Apatani Grammar. Mysore, Central Institute of Indian
Languages.
_____. (1987). Apatani-English-Hindi Dictionary. Mysore, Central Institute of
Indian Languages.
Badu, T. (2004a). Ramo Language Guide. Itanagar, Director of Research,
Government of Arunachal Pradesh.
_____. (2004b). Tangam Language Guide. Itanagar, Director of Research,
Government of Arunachal Pradesh.
Bickel, B. (2001). “Deictic transposition and referential practice in Belhare.”
Journal of Linguistic Anthropology 10(2): 224–47.
Bybee, J. L. and P. Hopper (2001). Frequency and the Emergence of Linguistic
Structure. Amsterdam/Philadelphia, John Benjamins.
Dixon, R. M. W. (2003). “Demonstratives: A cross-linguistic typology.” Studies
in Language 27(1): 61–112.
Ebert, K. (2003). “Kiranti languages: An overview.” In G. Thurgood and R. J.
LaPolla, Eds., The Sino-Tibetan Languages. London, Routledge: 505–17.
Levinson, S. C. (1983). Pragmatics. Cambridge, Cambridge University Press.
Megu, A. (1988). Bori Phrase Book. Itanagar, Directorate of Research,
Government of Arunachal Pradesh.
_____. (1993). The Karkos and Their Language. Itanagar, Directorate of
Research, Goverment of Arunachal Pradesh.
_____. (2003). Aashing Language Guide. Itanagar, Directorate of Research,
Government of Arunachal Pradesh.
Noonan, M. (2006). “Contact-induced change in the Himalayas: the case of the
Tamangic languages.” Paper presented at the International Colloquium on
154 North East Indian Linguistics Volume 3
Language Contact and Contact Languages, University of Hamburg, July 6–8,
2006.
Ou-Yang, J.-Y. (1985).
珞巴族语言简志
（
崩尼-博嘎尔语
)[Luobazu Yuyan
Jianzhi (Bengni Boga'eryu)/Brief Description of the Luoba Nationality
Language (The Bengni-Bokar Language)]. Beijing, 民族出版社 [Minzu
Chubanshe/Nationalities Press].
Post, M. (2007). A Grammar of Galo. PhD Dissertation. Melbourne, La Trobe
University Research Centre for Linguistic Typology.
_____. (2010). “Predicate derivations in the Tani languages: Root, suffix, both or
neither?” In S. Morey and M. Post, Eds., North East Indian Linguistics
Volume 2. New Delhi, Cambridge University Press India: 175–97.
_____. (in preparation). “Topographical deixis in Tibeto-Burman.” Cairns,
James Cook University.
_____. (forthcoming 2010). “The language, culture, environment and origins of
Proto-Tani speakers: What is knowable, and what is not (yet).” In S.
Blackburn and T. Huber, Eds., Origins and Migrations among Tibeto-
Burman Speakers of the Extended Eastern Himalaya [Provisional title].
(Tribal Cultures in the Eastern Himalaya.) Leiden, Brill.
Senft, G. (1997a). “Introduction.” In G. Senft, Ed., Referring to Space: Studies
in Austronesian and Papuan Languages. (Oxford Studies in Anthropological
Linguistics.) Oxford, Oxford University Press: 1–38.
_____. Ed., (1997b). Referring to Space: Studies in Austronesian and Papuan
Languages. (Oxford Studies in Anthropological Linguistics.) Oxford, Oxford
University Press.
_____. Ed., (2004). Deixis and Demonstratives in Oceanic Languages. Canberra,
Pacific Linguistics.
Simon, I. M. (1972). An Introduction to Apatani. Gangtok, Sikkim, Government
of India Press.
_____. (1976). Hill Miri Language Guide. Shillong, Government of Arunachal
Pradesh.
Sun, H.-K., J.-C. Zhang, et al. (1980).
门巴，珞巴，登人的语言
[Menba,
Luoba, Dengren de Yuyan/The Language of the Menba, Luoba and Dengren].
Beijing, 中国社会科学出版社 [Zhongguo Shehui Kexue Chubanshe/
Chinese Social Sciences Press].
Sun, T.S. J. (1993). A Historical-Comparative Study of the Tani (Mirish) Branch
of Tibeto-Burman. PhD Dissertation. Berkeley, University of California
Department of Linguistics.
_____. (1994). “Linguistic characteristics of the Tani (Mirish) branch of Tibeto-
Burman.” Bulletin of the Institute of History and Philology, Academica
Sinica 65(1): 175–220.
Tayeng, A. (1983). A Phrase Book in Padam. Shillong, Director of Research
(Philology), Government of Arunachal Pradesh.
von Fürer-Haimendorf, C. (1980). A Himalayan Tribe: From Cattle to Cash.
Berkeley/Los Angeles, University of California Press.
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 155
10
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ
Dialects of the Mising Community
Jugendra Pegu
Tezpur University
1. Introduction
Mising is listed in the Indian constitutional order of Scheduled Tribes,
1950, as Miri
1
. When and how the name Miri came to represent the
Misings remains a matter of conjecture
2
. In fact, “never a tribe known
as Miri existed before”, claims Dr N. C. Pegu. In the words of Prasad
(1991: 4), the word Miri is derived from ami ‘man’ + iri ‘virtue’ = miri
‘virtuous man’. Today, the Mising speakers consider the name Miri to
be derogatory and in fact refer to themselves as Mising. Hence, they are
now known by this name for all purposes. Sun, in his (1993) study of
Tani languages, states that the term Abor-Miri-Dafl a, composed of names
of the three major tribal groups speaking these languages, must be
discarded, since the outdated exonyms it is based on are now resented
by native speakers. These terms are no longer accepted or widely used
1
Mising refers to both the people and their language.
2
Prasad (1991: 4) states that the priests of the Adi (Hillmen) were called Miri
and when a group of Mising people came down to the Assam Valley, the plains
people, at the fi rst contact, came to know that the newcomers follow the cult
of Miri, i.e. worship of natural forces. Thus, the term Miri came to refer to the
whole tribe. On the other hand, Doley (2004: Preface) opines that the term
Miri is said to be derived from the Mising word miri = miruo, which signifi es
a priest of the Mising. According to him the Miri acted as the authority in
Mising villages, accomplishing marvelous acts and was considered to possess
superhuman powers.
156 North East Indian Linguistics Volume 3
in North-East India, having been replaced by the names Adi, Mising, and
Nyishi, respectively.
The Mising people are believed to have once been hill dwellers, and
to have lived in the hilly tracts of modern-day Arunachal Pradesh. The
precise date and circumstances of entry of the Misings to the Brahmaputra
Valley is not known. There are many controversies regarding the time
of their settlement in Assam as well as in the adjoining plains of
Arunachal Pradesh. N. C. Pegu and Dimbeswar Doley have estimated
the Mising migration to have taken place around the time of the Ahom
invasion in the early 13 century AD (Doley 2008: 48), although this has
not yet been independently confi rmed.
1.1. The language
Mising belongs to the Eastern Tani sub-group of the Tibeto-Burman
language family
3
. The language of this community was recognized for
educational purposes by the Government of Assam on 30 October, 1985.
Mising is categorized under the non-scheduled languages of Tribal and
Border Languages of India. Since its recognition, Mising has been using
a Roman-based script. At fi rst, native scholars of this community, under
the banner of misiŋ agom kə:baŋ (Mising Literary Society), experimented
with an Assamese-based orthography. However, it was later discontinued
due to problems in representing some Mising sounds. For instance, there
are no characters in Assamese script which can easily represent the Mising
high central vowel /ɨ/ and the mid central vowel /ə/ (both long and short).
Hence, the agom kə:baŋ opted for modifi ed forms of the Roman letters
í and é or v and c to represent these vowels, respectively. In this paper,
which is primarily intended for an audience of linguists, I have decided
to make use of the International Phonetic Alphabet (IPA).
Mising has twenty-nine phonemes, of which fourteen are vowels and
fi fteen are consonants. Tables 1 and 2 show the twenty-nine phonemes
based on Taid (2005).
3
Sun (2003: 456) refers to Tani as a compact cluster of Tibeto-Burman lan-
guages situated at the eastern end of the Himalayas, in an area skirted on four
sides by Tibet, Assam, Bhutan, and Burma. The Tani languages fall into two
divisions: Eastern Tani consisting of Mising plus many of the Adi dialects,
and Western Tani consisting of dialects spoken by the Nyishi(ng)-Bengni,
Tagin, Galo, and the Hills Miri tribes, among others. Tani languages are esti-
mated by Sun to be spoken by about 600,000 indigenous people of Arunachal
Pradesh and Upper Assam, although this fi gure should probably be revised
signifi cantly upward.
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 157
1.2. Mising dialects
Eight Mising dialects are traditionally recognized: pagro, dəlu, sa:jaŋ,
ojan, mo:jiŋ, dambuk, somua, and samuguria. Grierson (2005 [1909]:
584) mentioned only four dialects, viz. sa:jaŋ, ojan, dambuk, and somua
in the Linguistic Survey of India. It may be surmised that as the Misings
migrated from the hilly tracts of Arunachal Pradesh to different parts of
the plains of Assam, some dialects like pagro, dəlu, and mo:jiŋ, may
not have come to the notice of the survey team. Some later studies have
identifi ed the eight Mising dialects pagro, dəlu, sa:jaŋ, ojan, mo:jiŋ,
dambuk, somua, and samuguria (Prasad 2001: 4)
4
. On the other hand,
Taid (1995: viii) gives a list of six dialects, viz. dəlu, pagro, ojan, sa:jaŋ,
Table 1: Consonants in Mising (+/- V = voicing)
Place → Bilabial Dental Alveolar Palatal Velar
Manner ↓ → -V +V -V +V -V +V -V +V -V +V
Stop p b t d k g
Fricative s z
Nasal m n ɲ ŋ
Lateral l
Tap r
Continuant j
Table 2 Vowels in Mising
Front Central Back
Short Long Short Long Short Long
High i i: ɨ ɨ: u u:
Mid e e: ə ə: o o:
Low a a:
4
samuguria can no longer be described as a dialect of the Mising language
per se, inasmuch as it has almost completely merged with Assamese. Assamese
infl uence is found at all levels, with the speakers retaining only some Mising
fi lial terms and phonological features. However, the culture and traditions
of Samuguria speakers remains largely the same as those of other Mising
people.
158 North East Indian Linguistics Volume 3
mo:jiŋ, and dambuk. However, Taid has also mentioned somua as a
dialect which he believes to be spoken in some parts of Arunachal
Pradesh. Very little reliable information concerning the somua dialect is
currently available.
1.3. Population and demography
The Indian census report of 2001 reports the existence of 587,310 Mising
speakers, constituting a full 17.8% of the total tribal population of Assam.
However, a recent unoffi cial count jointly conducted by the takam misiŋ
porin kə:baŋ (All Mising Students Union), misiŋ mimag kə:baŋ (Mising
Revolutionary Organization), and takam misiŋ mimə kə:baŋ (All Mising
Women Organization) has put the population of Mising at 1,257,596
living in 1,724 villages in Assam (Doley 2008: 52).
Native speakers of Mising are primarily concentrated in the upper
half of Assam, spreading over the districts of Dhemaji, North Lakhimpur,
and Sonitpur on the North Bank of the Brahmaputra river and Tinsukia,
Dibrugarh, Sivasagar, Jorhat, and Golaghat on the South Bank. The
largest concentrations of Mising speakers are in Jonai and Majuli Sub-
Divisions in the districts of Dhemaji and Jorhat respectively. However,
a sizeable number of Misings are also settled in the East Siang District
of Arunachal Pradesh.
1.4. Goals and organization of the paper
This paper is divided into four sections. Some background information
on the language and its dialects is given in §1. The main topic of the
paper is the syntactic functions of the marker -ə in the pagro and sa:jaŋ
dialects of Mising (PD and SD hereafter); these are discussed in §2. In
§3, morphosyntactic variation among the two dialects mentioned is
discussed, followed by a discussion of morphophonemic changes brought
about by suffixation of -ə to a root word in §4. A conclusion then
summarizes the presentation.
2. Syntactic functions of -ə in pagro and sa:jaŋ dialects
In PD and SD, -ə has two different syntactic functions. It has the function
of a copula as well as that of a generic/defi nite non-specifi c marker.
These two functions are discussed in the following two sub-sections.
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 159
2.1. -ə as copula
In PD and SD, the marker -ə operates as a copula when it is suffi xed to
a sentence-fi nal nominal or adjective. Consider examples (1a)–(2b).
(1a) zon-bɨ ŋo-k kaі-ə.
John-3.
SG 1.SG-GEN elder brother-COP
‘John is my elder brother.’ (PD)
(1b) zon-bɨ ŋo-k sasi-ə.
John-3.SG 1.SG-GEN elder brother-COP
‘John is my elder brother.’ (SD)
(2a) mɨmbɨr-də kaŋkan-ə.
young.girl-DEF beautiful-COP
‘The young girl is beautiful.’ (PD)
(2b) mіjum-də kampu-ə.
young.girl-DEF beautiful-COP
‘The young girl is beautiful.’ (SD)
As shown in (1a) and (1b) above, we see that -ə links the subject
zon-bɨ ‘John-3.
SG’ to kaі ‘elder brother’ and sa:si ‘elder brother’ in PD
and SD respectively. Here, we see that the marker -ə links the subject
to another noun, or rather, one NP to another NP. Likewise, in (2a) and
(2b), the subject mɨmbɨr-də ‘young girl-DEF’ or mіjum-də ‘young girl-
DEF’ is linked to an adjective kaŋkan ‘beautiful’ or kampu ‘beautiful’.
Thus, going by the traditional defi nition of a copula as a linking element,
one may say that the suffi xation of -ə to a sentence-fi nal nominal or
adjective amounts to a copula function in PD and SD.
2.2. -ə as a generic marker
In §2.1 we have argued for the function of -ə as a copula. Apart from
this, -ə occurs as a generic marker when it attaches to a subject NP.
Consider examples (3a)–(5b).
(3a) іkі:-ko ra-la-duŋ.
dog-
INDEF bark-NF-EXIST
‘A dog is barking.’(PD)
(3b) əkі:-ko ra-la-duŋ.
dog-INDEF bark-NF-EXIST
‘A dog is barking.’(SD)
160 North East Indian Linguistics Volume 3
(4a) іki:-də ra-dak.
dog-DEF bark-PRES
‘The dog barks.’(PD)
(4b) əkі:-də ra-dak.
dog-DEF bark-PRES
‘The dog barks.’(SD)
(5a) іkі:-ə ra-dak.
dog-GENR bark-PRES
‘Dogs bark.’(PD)
(5b) əki:-ə ra-dak.
dog-GENR bark-PRES
‘Dogs bark.’(SD)
The marker -ko in (3a) and (3b) is almost certainly derived from the
numeral ako ‘one’. But here it is not occurring as an enumerative
modifi er, but rather as a referential marker. When an NP is marked by
-ko, the reading is indefi nite in both PD and SD, giving the sense of ‘any
given dog’. When an NP is marked in -də and -ə, the understanding is
defi nite; -də gives the reading of a defi nite specifi c noun phrase and -ə
as that of a generic/non-specifi c noun phrase. For instance, while in (4a)
and (4b) the reading is that of an individual dog, in (5a) and (5b), the
reading is that of all dogs as a natural class.
However, -ə may have different functions in different languages
belonging to the Tani group. For instance, it has a somewhat different
role in Galo
5
. Consider the following examples from the Lare dialect:
(6a) bɨ̂ɨ tic r əənà.
[bɨ̀ɨ]
CS
[tic r]
CC
[əə]
COP
=na
3.SG teacher COP.IPFV=DECL
‘He or she is a teacher (currently).’ (Post 2007: 423)
6
5
Galo is spoken in Arunachal Pradesh. It is a “transitional” (Eastern-convergent)
language of the Western Tani group (Post 2007).
6
[Transcription of a long vowel əə here is faithful to the source cited.
However, it may be that the source is incorrect in this respect, and that the
Galo form, like its Mising cognate, may be underlyingly short. Surface length-
ening which has been observed in Galo seems likely to be the result of (pho-
nological) word minimality rules. – Ed. Note]
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 161
(6b) hobbə́...jadɨ̂ɨ lokòm ohóo ogg rəllà...
[hob =əə]
CS
[jadɨ̀ lo=kòm ohóo-ók-gər -là(a)
mithun=TOP how.much LOC=ADD rope-tie.up-ACNC-NF
rɨ̂ɨl amáa namm .
rɨ́ɨ-là(a) á-máa-nam]
CC
[=əə]
COP
tie.up.animal-NF keep-NEG-NZR.RLS =COP.IPFV
‘The mithun…is the only animal that never ever gets tied up.’
(Post 2007: 425)
(6c) dooɲ caakú namm …
dooɲí=əə càa-kú-nam=əə
sun=TOP ascend-CMPL-NZR:RLS=TOP
attór kaakú.
attór-káa-kú
be.fi rm/hard-PF-CMPL
‘The sun having risen…(the earth) fi rmed up.’
(Post 2007: 407)
In both (6a) and (6b) we notice that, like in Mising, Galo makes use
of -ə as a sentence-fi nal copula. However, in (6b) and (6c), -ə occurs
both with a subject NP and with a backgrounding clausal nominalization;
it is analyzed by Post (2007: 423) as a topic marker, entailing both
defi nite specifi c and non-specifi c functions, among others. Thus, it also
includes the function of the defi nite marker -də in Mising (for which
there is no corresponding form in Galo). Accordingly, we may say that
-ə has closely similar patterning in Mising and Galo, but has a wider
range of uses in Galo when marking an NP.
3. Morpho-syntactic variation in pagro and sa:jaŋ dialects
Morphosyntactic variation among PD and SD is observed in the functions
described above. In PD the form of a word is frequently altered when
-ə is suffi xed to it. The form of a word in SD, on the other hand, does
not get affected in any way with the suffi xation of -ə to the root word.
Along these lines, it is possible to divide dialects of Mising into two
groups: geminate and non-geminate. Dialects which fall under the
geminate group are pagro, dǝlu, ojan, and dambuk. The non-geminate
group includes sa:jaŋ, mo:jiŋ, and somua
7
. In this section, we focus on
7
Due to extensive Assamese infl uence the Samuguria data are insuffi cient in
this area to assign Samuguria to either of these two groups, so it is left aside.
162 North East Indian Linguistics Volume 3
variations brought about by suffi xation of -ə to nominal stems. Consider
the following examples:
(7a) bɨ zon-kə ommə.
bɨ zon-kə omə-ə
3.
SG John-GEN daughter-COP
‘She is John’s daughter.’(PD)
(7b) bɨ zon-kə omə-ə.
bɨ zon-kə omə-ə
3.SG John-GEN daughter-COP
‘She is John’s daughter.’(SD)
(8a) tabbə zon-mə pa-to.
tabɨ-ə zon-mə pa-to
snake-GENR John-ACC bite-PERF
‘Some snake has bitten John.’(PD)
(8b) tabɨ-ə zon-mə pa-to.
tabɨ-ə zon-mə pa-to
snake-GENR John-ACC bite-PERF
‘Some snake has bitten John.’(SD)
It is evident from the examples given above that the root words omə
‘daughter’ and tabɨ ‘snake’ in both the dialects are same, but the
combination of -ə with the root words bring about morphophonemic
changes in PD and not in SD. The alternation is further exemplifi ed in
Table 3.
Among the PD forms shown, when -ə is suffi xed to a root word,
gemination of the medial consonant and deletion of the word-fi nal vowel
are observed. But we see that the suffi xation of -ə to an SD root word
does not bring about any change to the form of the word. Notably, the
Table 3 Geminate and non-geminate forms of PD and SD: Root words
along with their syllabic representations
Root + -ə Structure Gloss PD SD
tabɨ-ə
CV.CV-V ‘snake-GENR’ tabbə tabɨ-ə
(
CVC.CV) (CV.CV.-V)
omə-ə
V.CV-V ‘daughter-COP’ ommə omə-ə
(
VC.CV) (V.CV.-V)
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 163
alternation is relevant only to forms ending in vowels; forms ending in
consonants are not affected (9a–10b).
(9a) ədə bɨ-k kujab-ə/gimur-ə/apoŋ-ə/matsik-ə.
that 3.
SG-GEN spade-COP/cotton-COP/liquor-COP/knife-COP
‘That is his spade/cotton/liquor/knife.’(PD)
(9b) ədə bɨ-k kujab-ə/gimur-ə/apoŋ-ə/matsik-ə.
that 3.SG-GEN spade-COP/cotton-COP/liquor-COP/knife-COP
‘That is his spade/cotton/liquor/knife.’(SD)
(10a) ukum-ə/tapat-ə/oŋer-ə/tamik-ə
house-GENR/leech-GENR/prickly.ash-GENR/mosquito-GENR
jumra-sə ka-ma.
forest-LOC exist-NEG
‘There is no house/leech/prickly ash/mosquito in this
forest.’(PD)
(10b) ǝkum-ə/tapat-ə/oŋer-ə/tamik-ə
house-GENR/leech-GENR/prickly.ash-GENR/mosquito-GENR
jumra-sə ka-ma.
forest-LOC exist-NEG
‘There is no house/leech/prickly ash/mosquito in this
forest.’(SD)
It is evident from the above examples that a word ending with a
consonant phoneme, whatever it may be, is not affected when suffi xed
by -ə in either PD or SD.
4. Morphophonemic variations
In §3, we saw that when -ə is suffi xed to SD root words ending with
either a vowel or a consonant, no change takes place, whereas some
morphophonemic changes occur in PD root words ending with a
vowel.
However, not all the available vowels trigger the change in PD. For
instance, words ending with a long high front vowel /i:/ or long mid front
vowel /e:/ do not undergo any change. For instance; iki: + ə → iki:-ə
‘dog.COP’ and take: + ə → take:-ə ‘ginger.COP’. A few more examples
of words ending in i: and e: which are common to both PD and SD and
do not trigger any morphophonemic changes are ke:di: ‘mango’, ki:ni:
164 North East Indian Linguistics Volume 3
‘navel’, be:siri: ‘young female monkey’, ope: ‘family’, sibe: ‘monkey’,
take: ‘ginger’, me:re: ‘charcoal’, and gule: ‘pot’.
In PD and SD alike, most types of long vowel occur in non-word-fi nal
(word-initial and word-medial) positions only. For instance: gə:r ‘ask
someone to move’, a:buŋ ‘river’, ga:m ‘chieftain or head of a village’,
u:saŋ ‘lift, especially clothes’, tu:li ‘pot’. Some examples in SD are ə:
zo ‘old woman’, sa:si ‘elder brother’, a:nə ‘river’, tɨ:k ‘louse’, and jo:jo
‘grandmother’. In PD and SD, only i: and e: are found word-fi nally
among long vowels.
In the following sections, then, we will be primarily concerned with
changes that occur in PD root words ending in short vowels when
combined with -ə. In PD, a set of morphophonological rules apply when
-ə is added to nominal stems and adjectives ending in various types of
short vowels. The vowels triggering the changes form natural classes:
mid polar vowels e and o, central vowels ɨ, ə, and a, and high polar
vowels i and u each trigger different processes when combined with -ə.
The phonological operations involve vowel coalescence, vowel deletion,
and gemination, which are determined by the type of vowel a word ends
with.
4.1. Coalescence: words ending in high polar vowels i and u
In general, coalescence is a phonological process whereby two adjoining
or contiguous segments converge or fuse into one element such that the
new segment is phonologically distinct from the input segments. Crystal
(1980: 65) defi nes coalescence as “the coming together of linguistic units
which were originally distinguishable. Allophones of a phoneme may
coalesce, as many different phonemes and different morphemes.”
In PD, coalescence occurs when -ə is attached to root words ending
in high polar vowels i and u. Not all words ending with these two vowels
allow coalescence; it is determined by the syllabic structure of a word.
The fusing of the vowels triggers word-medial gemination which is again
allowed or restricted by the presence or absence of a heavy initial syllable.
Given below are few examples with words ending in i with various
syllabic structures:
(11a) ui + -ə → uijə ‘spirit.
COP’
(11b) asi + -ə → asse ‘water.COP’
(11c) pəsi + -ə → pəsse ‘needle.COP’
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 165
(11d) ampi + -ə → ampe ‘sticky rice.COP’
(11e) rugzi + -ə → rugze ‘a kind of fern.COP’
(11f) do:ɲi + -ə → do:ɲe ‘sun.COP’
(11g) baboi + -ə → baboijə ‘father’s brother.COP’
Examples (11a–11g) above exhibit words ending in i with various
possible syllable types (VV, V.CV, CV.CV, VC.CV, CVC.CV, CVV.
CV. and CV.CVV) in PD. Here, we have both monosyllabic and
disyllabic words. We notice the presence of light and heavy initial and
fi nal syllables in the given disyllabic words. Words with light initial and
fi nal syllables (VC.V and CV.CV, as in (11b–11c)) both allow coalescence
and produce word-medial gemination when -ə is attached. On the other
hand, a disyllabic word with heavy initial syllable and fi nal light syllable
(VV, CV.CV, CVC.CV, and CVV.CV) allows coalescence but not word-
medial gemination. The fusing of i and -ə usually gives the resultant
vowel e, and this can be represented as V
1
(high, front) + V
2
→ V
3
(mid,
front), or i + -ə → e. It is worth noting here that a monosyllabic word
in PD which is in diphthongal form (11a) or a disyllabic word in which
the word-fi nal syllable is a diphthong (11g) neither coalesces nor exhibits
word-medial gemination; rather, we notice the insertion of the palatal
continuant j in between the root-fi nal vowel and the suffi x -ə.
In words ending in u include the structures VV, CVV, CV.CV, VC.CV
and CVV.CV. Consider the following examples:
(12a) ou + -ə → ouwə ‘mother.
COP’
(12b) bau + -ə → bauwə ‘father.COP’
(12c) anu + -ə → anno ‘new.COP’
(12d) pəpu + -ə → pəppo ‘owl.COP’
(12e) pe:gu + -ə → pe:go ‘a Mising surname.COP’
As above, words which have light-light syllable structure (V.CV and
CV.CV) as in (12c) and (12d), allow coalescence and as well as
gemination when -ə is fused with word-fi nal u. On the other hand, words
with heavy-light syllable structure allow coalescence but do not trigger
word-medial gemination (see (12e)). Here, the fusing of u and -ə results
with o as the output vowel and it can be represented as V
1
(high, back)
+ V
2
→ V
3
(mid, back) or u + -ə → o.
Monosyllabic words with VV and CVV structure (12a-b) neither
coalesce nor exhibit gemination in the word-medial position. However,
166 North East Indian Linguistics Volume 3
we notice the insertion of /w/ in between the root-fi nal vowel and the
suffi x -ə.
The discussion above establishes the presence of vowel coalescence
when -ə is attached to words ending in i and u, except when i and u are
second members of diphthongs as in (11a), (11g) and (12a–12b).
4.2 Deletion: words ending in central vowels
ɨ
,
ə
, and a
In PD, the rule of vowel deletion applies instead of vowel coalescence
when -ə follows words ending in ɨ, ə, and a. In this rule, the word-fi nal
vowel is deleted while -ə is preserved (13a–13e).
(13a) apɨ + -ə → appə ‘egg.
COP’
(13b) pəkɨ + -ə → pəkkə ‘dove.COP’
(13c) omə + -ə → ommə ‘daughter.COP’
(13d) karə + -ə → karrə ‘raised bamboo platform.COP’
(13e) sira + -ə → sirrə ‘wild boar.COP’
As shown in (13a–13e), gemination is triggered at the word-medial
consonant in words with light-light syllable structures V.CV and CV.CV.
Further, we notice that in (13a–13b) and (13e), deletion of the word-fi nal
vowel ɨ and a is clear, whereas in (13c–13d) one cannot easily assume
which one of the vowels gets deleted as V
1
(ə) and V
2
(-ə) are identical.
However, from our observation of the other central vowels (ɨ and a), we
may suppose that it is the word-fi nal vowel (V
1
) which gets deleted, and
not the suffi x vowel (V
2
). This can be re-written as V
1
(central) + V
2
→
V
2
.
Words which end in short central vowels with a heavy initial syllable
allow vowel deletion but do not exhibit word-medial gemination (14a–
14c).
(14a) ka:kɨ + -ə → ka:kə ‘mother’s brother.
COP’
(14b) e:nə + -ə → e:nə ‘female pig.COP’
(14c) ki:ra + -ə → ki:rǝ ‘trivet (tripod).COP’
From the above discussion it is evident that both vowel deletion and
word-medial gemination occur with a word ending in central vowels ɨ,
ə, and a with light-light syllable structures (V.CV and CV.CV) when
suffi xed in -ə. On the other hand, the presence of a heavy initial syllable
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 167
Table 4 Morphophonemic changes in PD root words (ending in short vowels) with various syllabic structures
V
1
Syllabic structure V
2
Coalescence V
3
Deletion WMG
i
VV (diphthong) Does not occur Not triggered
V.CV Occurs Triggered
CV.CV Occurs e ----------- Triggered
CVC.CV Occurs Not triggered
CVV.CV (IS ending in a long vowel) -ə Does not occur Not triggered
e
V.CV Triggered
CV.CV Does not occur e V
2
deleted Triggered
CVC.CV Not triggered
ɨ
V.CV Triggered
CV.CV Does not occur ə V
1
deleted Triggered
CVV.CV (IS ending in a long vowel) Not triggered
ə
V.CV Triggered
VV.CV (IS ending in a long vowel) Does not occur ə V
1 deleted
Not triggered
CV.CV Triggered
v
C.CV Not triggered
a
CV.CV Triggered
CVV.CV (IS ending in a long vowel) Does not occur ə V
1
deleted Not triggered
CVC.CV Not triggered
u
V.CV Triggered
CV.CV Occurs o ------------- Triggered
CVV.CV (IS ending in a long vowel) Not triggered
o
V.CV Triggered
CV.CV Does not occur o V
2
deleted Triggered
CVV.CV (IS ending in a long vowel) Not triggered
Note: WMG = word-medial gemination, IS = initial syllable
168 North East Indian Linguistics Volume 3
CVV.CV as in (14a–14c) triggers vowel deletion but restricts word-
medial gemination in the same conditions.
4.3. Deletion: words ending in mid polar vowels e and o
In §4.2 we observed that word-fi nal central vowels ɨ, ə, and a are deleted
when followed by -ə. However, in words ending with mid, polar vowels
e and o, we fi nd deletion of the suffi xal vowel -ə (V
2
), not the word-fi nal
vowel (V
1
). A few examples are given in (15a–15f).
(15a) eŋe + -
ə
→ eŋŋe ‘a kind of arum.COP’
(15b) take + -
ə
→ takke ‘crab.COP’
(15c) bomze + -
ə
→ bomze ‘a big village.COP’
(15d) əpo + -
ə
→ əppo ‘winnowing basket.COP’
(15e) mibo + -
ə
→ mibbo ‘guest.COP’
(15f) po:lo + -
ə
→ po:lo ‘moon.COP’
As shown, the combination of e and o (V
1
) with -ə (V
2
) results in
the deletion of the latter; this can be rewritten as V
1
(mid, polar) + V
2
→
V
1.
Here, words with V.CV, CV.CV structure ending in both the
vowels also trigger word-medial gemination. However, words with fi nal
e and o with heavy initial syllables, as discussed earlier, exhibit vowel
deletion but restrict word-medial gemination.
Conclusion
In this study of two Mising dialects, namely pagro and sa:jaŋ as spoken
in the districts of Dhemaji and Tinsukia (both in Upper Assam), we found
that the marker -ə in both the dialects have two different syntactic
functions: (1) as a copula when suffi xed to sentence-fi nal nominal or
adjective; (2) as a generic or defi nite non-specifi c marker when suffi xed
to a subject NP. The attachment of -ə to PD and SD root words exhibit
differences: PD may be described as a geminate variety and SD as a
non-geminate variety.
The suffi xation of -ə to an SD root word triggers no morphophonemic
changes, regardless of syllable structure or the nature of the word-fi nal
vowel. In PD, words ending in consonants and long vowels i: and e:
(i.e., words with fi nal heavy syllables) are not affected either, whereas
various morpho-phonemic changes occur in words ending with a short
fi nal vowel (i.e. words with light fi nal syllables).
Morpho-Syntactic Variation in the Pagro and Sa:jaŋ Dialects ... 169
Suffi xation of -ə to PD words ending in short high polar vowels,
central vowels, and mid polar vowels variously triggers vowel coalescence,
vowel deletion, and word-medial gemination. Coalescence occurs only
in words ending with high polar vowels, whereas deletion occurs in words
ending with central and mid vowels. In the case of central vowels, it is
the word-fi nal vowel which is deleted, whereas in the case of mid, polar
vowels, it is the suffi x which is deleted. Word-medial gemination occurs
in words with a light-light syllable structure ending in any short vowel,
but it is blocked in words with a heavy initial or fi nal syllable. Finally,
we note that diphthongs ending in high polar vowels i and u exhibit none
of the above-mentioned changes when suffi xed in -ə, but rather exhibit
insertion of a homorganic liquid j or w. Table 4 summarizes the observed
set of changes.
Abbreviations
1 First person
2 Second person
3 Third person
+ Morpheme boundary
→ Written as
ACC Accusative
COP Copula
DEF Defi nite
EXIST Existential
GEN Genitive
GENR Generic
INDEF Indefi nite
LOC Locative
NEG Negative
NF Non-fi nite
PD Pagro dialect
PERF Perfective
PRES Present
SD Sa:jaŋ dialect
SG Singular
V
1
Vowel that ends a root word
V
2
Vowel suffi xed to root word as copula or generic marker
V
3
Output vowel
170 North East Indian Linguistics Volume 3
References
Crystal, D. (1980). A First Dictionary of Linguistics and Phonetics. Boulder,
CO, West View.
Doley, J. (2004). The Pronouncing Mising-English-Assamese Dictionary.
Nalbari, Purba Bharati Publication.
Doley, R. (2008). Inter-Lingual Code-Switching as a Socio-Linguistic
Phenomenon amongst the Misings of the Brahmaputra Valley: A Study. PhD
Thesis. Gauhati, Indian Institute of Technology.
Grierson, G. A., Ed. (2005 [1909]). Linguistic Survey of India, Volume 3: Tibeto-
Burman Family, Part 1: General Introduction, Tibetan Dialects, Himalayan
Dialects, & North Assam Groups. New Delhi, Low Price Publications.
Post, M. (2007). A Grammar of Galo. PhD Thesis. Research Centre for Linguistic
Typology. Melbourne, La Trobe University.
Prasad, B. R. (1991). G. D. P. Sastry and P. T. Abraham, Eds., Mising Grammar.
Mysore, Central Institute of Indian Languages.
Sun, T.-S. J, (1993). A Historical-Comparative Study of the Tani (Mirish) Branch
in Tibeto-Burman. PhD Thesis. Department of Linguistics. Berkeley,
University of California.
_____. (2003). “Tani languages.” In G. Thurgood and R. J. LaPolla, Eds., The
Sino-Tibetan Languages. [Routledge Language Family Series.] London/
New York, Routledge: 456–66.
Taid, T. (1995). A Dictionary of Mising Language. Jorhat, takam misiŋ poji:rnə
kə:baŋ [All Mising Teachers’ Association].
The Referring Systems and the Determinative Elements ... 171
Eastern Indo-Aryan
172 North East Indian Linguistics Volume 3
172 blank
The Referring Systems and the Determinative Elements ... 173
11
The Referring Systems and the Determinative
Elements of Noun Phrases in Assamese
Jagat Chandra Kalita
Abhayapuri College
1. Introduction
Along with number, gender and case, nouns in Assamese and all other
Indo-European languages have another important grammatical category,
the discussion of which was not found in traditional grammars of olden
days of east or west. Nouns, especially class nouns, need to be determined
by some devices when they are used in sentences to realize any argument
– subject, object or complement. The noun man in English or manuh in
Assamese, for example, can refer to all the human beings in one use and
in another use the reference of the same noun can be restricted to only
one person. The reference of the noun man in former use is called generic
and in later the reference is said to be of specific. The specific reference
of a class noun can be again definite or indefinite.
One of the reasons behind the absence of discussion on the
determinative elements of nouns in traditional grammar may be the
absence of a definite article in the older Germanic languages in the west
(Hawkins 1978: 13) and any such element in Old and Middle Indo-Aryan
Languages in the east (Kakati 1972: 278) to show contrast between
definiteness and indefiniteness of NPs. In modern linguistics, however,
this feature has been studied by scholars from many different backgrounds
and perspectives. Traditional grammarians (Christopherson 1939;
Jespersen 1949), philologists (Heinrichs 1954; Hodler 1954), structuralists
174 North East Indian Linguistics Volume 3
(Yotsukura 1970), transformational syntacticians (Smith 1964; Higham
1972), and philosophers and logicians (Russell 1905; Strawson 1950)
have all contributed to the currently available data and descriptive
insights. Nowadays, in most grammars of prominent languages,
discussions on determinative elements of NPs are commonly found.
In this context, Assamese and some other New Indo-Aryan languages
such as Bangla and Oriya are somewhat peculiar in terms of the
determination of noun reference. In these languages, certain morphemes
are added to nouns, pronouns, numerals and/or adjectives which function
as the determinative element of the concerned nouns. Different scholars
have presented valuable discussions on these morphemes and their
functions from different points of view. Chatterji (1926) tried to work
out the roles and origins of these morphemes in Bengali, terming them
“enclitic definitives” or “numeratives”. He also mentioned that these
morphemes are “commonly described as articles”, which suggests that
other scholars had been discussing this feature of these languages before
him. Following Chatterji, Kakati (1941) gave a detailed discussion on
these morphemes in Assamese along the same lines, but mentioned that
these morphemes “have the value of the [English] definite article the”.
Medhi (1978) later attempted to trace the origin of these morphemes –
which he called “particles” – in Assamese, firmly advocating that the
origin of the system in Assamese is Bodo (Tibeto-Burman). G. C.
Goswami, in various works (e.g. Goswami 1968; 1982; 1990) as well as
in a joint work with J. Tamuli (Goswami and Tamuli 2003), dealt with
this feature synchronically and somewhat more objectively. In Goswami’s
analysis, the so-called Assamese “plural markers” -bʊr, -bilak, -hɔt and
-mɔkʰa – which were excluded from discussions of noun determination
by his predecessors – were rightly included in the discussion. He called
these morphemes “definitives”, and divided them into two classes:
singular definitives and plural definitives. At the same time, Goswami
pointed out other roles which are played by these morphemes, such as
indicating various types of number, gender, status, shape, and size, etc.,
according to the different classes of nouns in addition to their reference.
Dutta Barua has given a comparative description of this feature in
Assamese, Bangla and Oriya in his comparative studies of Assamese vs.
Oriya and Bangla (Barua 1978; 1994). Later, Dasgupta (2003: 379)
discussed these morphemes in terms of referential definiteness,
indefiniteness and antidefiniteness phenomena of the Bangla language.
The Referring Systems and the Determinative Elements ... 175
Finally, Kalita took up the issue of these and other determinative elements
of NPs in Assamese in his (2004) PhD Thesis.
The contrast of generic and specific reference in Assamese will be
first exemplified in (1)–(2)
1
.
(1) bag
ʰ-e goru kʰa-i.
tigers-NOM cow eat-3
‘A tiger eats a cow.’
(2a) bagʰ-tʊ-e ɛ-ta goru kʰa-l-e.
tiger-DTV-NOM one-DTV cow eat-PAST-3
‘The tiger has eaten a cow.’
(2b) bagʰ-tʊ-e goru-tʊ kʰa-l-e.
tiger-DTV-NOM cow-DTV eat-PAST-3
‘The tiger has eaten the cow.’
The NPs bag
ʰ and goru in subject and object slots in sentence (1) do
not refer to any specific tiger or specific cow and the reference, therefore,
is said to be generic (Quirk et al. 1985). In sentences (2a) and (2b) on
the other hand, the NPs have the same head nouns as in (1), but this time
refer to a specific tiger and a specific cow respectively; that is, their
reference is specific. They differ in that the specific reference of the NP
bagʰtʊe in (2a) is definite, whereas the NP ɛ-ta goru has indefinite
reference. In (2b) both NPs are of definite reference. It should be noted
here that a numeral cannot be used in sentences like (2a) without the
particle -ta. The absence of -ta, however, doesn’t change the meaning
of the sentence, but simply renders it unacceptable.
2. Determiners and determinatives
The term “determiner”, in some traditions, covers both the determinative
words of an NP and dependent words of a verb phrase, i.e. the auxiliaries
(Matthews 1981). In this study, however, the term “determiner” will be
used to refer only to the former. “The kind of reference a particular noun
phrase depends on its DETERMINATIVE element, i.e. the item, which
determines it” (Quirk et al. 1985: 253). In English, “this function is
typically realized by a set of closed class items, or DETERMINERS, which
occur before the noun acting as head of the NP (or before its pre
1
The data in this paper is drawn from Kalita (2004). Most of the examples were
provided by me, as a native speaker of Assamese.
176 North East Indian Linguistics Volume 3
modifiers)” (Quirk et al. 1985: 253). Assamese also has this same system
of noun reference i.e. referring by means of determiners. In addition to
that, Assamese has one more system for the same purpose which is
morphological and which is not found in English. Both of these two
systems can be illustrated with the example (2a) above. In the case of
ɛta goru, the second NP of the sentence ɛ-ta is used as a determiner
much like a in English. The English translation of bagʰ-tʊ-e, the first
NP of the sentence, is given as the tiger (nominative), i.e. in place of
determiner the in English Assamese uses a morpheme like -tʊ which is
affixed to the noun. So, we get two systems of determination of noun
reference in Assamese – one is by using a determiner like ɛ-ta before
(or after in few cases) the noun acting as head of the NP and another is
by affixing a morpheme (free or bound) like -tʊ with the noun acting as
head of the NP or any dependent words of that noun.
In this system, morphemes which are added to nouns, or to certain
other elements within the NP, as determinatives (DTV), have some
similarities with certain elements of some other languages which have
been generally termed “classifier” (CL). Some important similarities
between these two types of determination system are that both are related
to the quantification and measurement of nouns, and that their selection
depends upon some semantic properties of nouns as illustrated in
Table 1
2
.
Table 1 Comparison of Assamese determinatives and Boro classifiers
Assamese determinatives Boro classifiers
a) ɛ-zɔn manuh mansi sa-se
one-
DTV man man CL-one
‘a man’ ‘a man’
b)
ɛ-ta goru mosow ma-se
one-
DTV cow cow CL-one
‘a cow’ ‘a cow’
c) du-kʰɔn kitap
bizab gaŋ-nøi
two-
DTV book book CL-two
‘two books’
‘two books’
d)
pas-ta koni bidoi boi-ba
five-DTV egg egg CL-five
‘five eggs’ ‘five eggs’
2
The source of Boro data is Bhattacharya (1977).
The Referring Systems and the Determinative Elements ... 177
Even some such morphemes are common to both Boro and Assamese
(Table 2); this would suggest that the two systems may at least partly
share a common origin.
In view of these and certain other minor similarities, Assamese
determinatives are also called classifiers in some literature. But if we
compare the role of these two elements in sentences we notice that they
are actually very different. The role of the Assamese elements is not to
classify the nouns but to determine the reference of the noun i.e. whether
the item or object referred by the noun is generic or specific, if specific
then definite or indefinite, whether one or more than one, if more than
one or uncountable what is the quantity, whether animate or inanimate,
if animate whether human or non human, whether masculine or feminine,
if human whether inferior, familiar or honorific, if inanimate whether
flat, round, long or oblong? In short, in Assamese the kind of reference
a particular noun phrase has mainly depends on its determinative that is
attached to either the head noun or any of its dependent words, provided
the head noun is not a proper noun. As Goswami observes:
the classifiers no longer classify the Assamese nominals into distinct
classes; nor is there any evidence that they ever did in the earlier periods
of the language. They, however, convey some sense and give some idea
about the object of which they become the adjunct, whether it is male
or female, respectful or general, big or small, round or flat or oblong, in
bunches or otherwise, etc. All classifiers in the language may be called
definitives when they are I(mmediate) C(onstituent)s with a preceding
noun or pronoun. (Goswami 1968: 111)
Table 2 Comparison of Assamese determinatives and Boro classifiers
Meaning Assamese determinative Boro classifier
‘bundle of long things’ -mut
ʰa/-mutʰi mutʰa-
‘unit of six paisa’ -ɔna/-ana ana-
‘thousand’ -hezar/-hazar hazar-
‘pair’ -z
ʊra/-zuri zøra-
‘bundle or collection’ -mɔkʰa mɔkʰa-
‘kilogram’
-ħer ser-
‘40 kilogram unit’ -mʊn mʊn-
178 North East Indian Linguistics Volume 3
A short list of the determinatives found in Assamese is given in
Table 3, together with illustrative examples of their use.
Furthermore, it is not clear that the terms “enclitic definitives” or
“numeratives”, used by Chatterji and Kakati, or “definitives”, used by
Goswami and Tamuli, can be unproblematically applied to this system.
This is because, when such elements are added to a numeral and used
together as determiner to a noun, the reference of that noun then becomes
indefinite; i.e. such elements can be used to make a noun phrase both
definite specific and indefinite specific. And when such an element is not
in the NP then the head of that NP is generic in reference (cf. (15a)–
(15b)). Therefore, it would seen that the term DETERMINATIVE (DTV) for
this class of elements is preferable, as determining the reference of nouns
is their main role in a sentence in Assamese. To refer to both determiners
and determinatives I shall use the common name DETERMINATIVE ELEMENT.
Though it was mentioned above that the kind of reference a particular
NP has depends on its determinative elements, NPs which are headed
Table 3 Determinatives in Assamese
Determinative Example
Determinative Example
-t
ʊ kɔlɔm-tʊ -kʰɔn kitap-kʰɔn
pen-DTV book-DTV
‘the pen’ ‘the book’
-zɔn lora-zɔn -zoni soali-zoni
boy-DTV girl-DTV
‘the boy’ ‘the girl’
-tʰʊpa aŋur-tʰʊpa -akʰi kɔl-akʰi
grapes-DTV banana-DTV
‘the bunch of grapes’ ‘the bunch of bananas’
-bɔtɔl pani-bɔtɔl -kahi bʰat-kahi
water-DTV rice-DTV
‘the water in the glass’ ‘the rice on the plate’
-duta - tamʊl-duta -keizɔn manuh-keizɔn
betel nut-DTV man-DTV
‘the two betel nuts’ ‘the men’
-dɔzɔn koni-dɔzɔn -mʊn dʰan-mʊn
egg-DTV paddy-DTV
‘the one dozen eggs’ ‘the 40 kg. of paddy’
The Referring Systems and the Determinative Elements ... 179
by proper nouns are generally excluded from this phenomena, as proper
nouns generally need no determinative element. A proper noun can take
a determinative element only when it is in high focus. In the following
passage from Surjya Kr Bhuyan’s printed lecture (Bhuyan 1997) ħɔraigʰat
ase ħɔraigʰat nai, the name of Guwahati is mentioned five times; but
only once is a determinative -kʰɔn affixed to it, namely when it is under
focus. (The king, the speaker, there is highly concerned for the important
place Guwahati)
Guwahati tetia mʊgɔlɔr tɔlɔt, sɔrgɔdeue patrɔ-mɔntrik kole, “mʊk gu-
wahati-kʰɔn loi dehɔk.” patrɔ-mɔntrie kole, “guwahati loi dim, kintu
etia tar pɔr hʊa nai…” ɔħomiyai mʊgɔlɔk kati-mari guwahatir pɔra
kʰedile. guwahati kamrup punɔr ɔħɔmɔr hol.
“At that time when Guwahati was under the Mughals, the King (sɔrgɔdeu)
said to his ministers “you recapture Guwahati for me.” The ministers
said “we must recapture Guwahati for you, but this is not the right
time…” The Assam side has driven out the Mughals from Guwahati by
killing or injuring them. Guwahati i.e. Kamrup has been regained by
Assamese.”
On the other hand, if and when the reference of a noun is not in any
way important in a particular context, even one single or a few item(s)
or substance(s) can be referred without using any determinative element;
in this case, the noun is said to be generically referring, as in (3)–(4).
(3) darʊga-i pistɔl ulia-i guli kor-ibɔ kʰuz-isil.
officer-NOM pistol bring out-PART shoot-PART try-DPAST.3
‘Bringing out pistol the Police officer was ready to shoot.’
(Agarwala 1981)
(4) ɛnete naw-ɔr boitʰa-r pɔra hat er-i
then boat-GEN oar-GEN from hand leave-PART
mʊgɔl-bor-e sapori mar-ibɔloi dʰor-il-e.
Mughal-DTV-NOM clap beat- PART start-PAST-3
‘Then the Mughals leaving oars of their boats started to clap.’
(Bhuyan 1997)
The NPs pist
ɔl in (3) and naw < naʊ in (4) are used in generic
reference because their specificity is not important in the sentences. This
type of sentence, however, has a chance of being ambiguous. Hearer(s)
may have to ask which pistol the police officer brought out, or which
180 North East Indian Linguistics Volume 3
boats are referred to here. In the sentence immediately following (3),
from a work of literature, the author used the noun pistɔl in definite
reference as in (5).
(5)
mɔi tʰap mar-i pistɔl-tʊ lo-i….
I snatch-PART pistol-DTV take-PART
‘Snatching his pistol I…’
Even the use of a DTV with pistɔl in (3) and naw in (4) does not make
the sentences unacceptable; i.e., they can be used in specific reference
also. Then the sentences will be as in (3a) and (4a).
(3a) dar
ʊga-i pistɔltʊ uliai guli kor-ibɔ kʰuz-isil.
(4a) ɛnete naʊ-bilak-ɔr boitʰa-r pɔra hat er-i mʊgɔl-bor-e sapori
mar-ibɔloi dʰor-il-e.
We will now discuss the processes of determination of NP reference
and the determinative elements in Assamese in more detail.
3. Different classes of determiners
In Assamese there is no permanent class of determiner words such as
the articles the and a/an in English. In Assamese, lexemes of different
word classes (nouns, pronouns and also adjectives), in different forms,
can realise the function of determiner in an NP. One important point
about the use of some of the words in determiner functions is that a DTV
is obligatorily added to such NPs, either with the determiner, as in (6),
or with its head noun, as in (6a). Due to the absence of a DTV, (6b) is
unacceptable. There is, however, no stacking of determiners in
Assamese.
(6) ħei-t
ʊ kɔlɔm an-a.
that-DTV pen bring-2.FAM
‘Bring that pen.’
(6a) ħei kɔlɔm-tʊ an-a.
that pen -DTV bring-2.FAM
(6b) *ħei kɔlɔm an-a.
The Referring Systems and the Determinative Elements ... 181
Both (6) and (6a) are unacceptable without the DTV -tʊ in the NP.
This, however, is not the case with all types of determiners. In fact,
demonstratives and other some pronouns of Assamese are not exactly
the same as in English. In Assamese, most pronouns are in fact bound
forms and are therefore called pronominal roots. The precise character
of their boundness is, however, somewhat peculiar. Demonstrative
pronouns cannot be used as pronouns without a co-occurring DTV. On
the other hand, when they are used in determiner functions, the DTV can
be added either to the demonstrative itself or to its head.
Another important point to make here is that the choice of the DTVs
used with determiners is made according to the class of the head noun
and has nothing to do with the class of the determiner.
The following subsections §3.1–§3.6 discuss the classes of words or
lexemes which are used in determiner functions in Assamese.
3.1 Pronouns
Demonstrative, interrogative, indefinite and relative classes of pronouns
are found in determiner functions. Table 4 displays the determiner
function of Demonstrative, interrogative and relative classes of
pronouns:
In the determiner use of demonstrative pronouns, the addition of a
DTV is obligatory in some cases, as shown in (7)–(7a), but in some cases
it is optional, as in (8)–(8a).
(7)
ei-zɔn manuh-ɔk mɔi zan-ʊ.
this-DTV man-ACC I know-1
‘I know this man.’
(7a) *ei manuh-ɔk mɔi zan-ʊ.
(8) ei-bidʰ gɔs ɔħɔm-ɔt pʊ-wa ne-za-i.
this-DTV tree Assam-loc find-part NEG-be-3
‘This species of trees is not found in Assam.’
(8a) ei gɔs ɔħɔm-ɔt pʊ-wa ne-za-i.
Despite their grammatically “optional” status, there is a strong
tendency of DTVs occurring in such phrases, such that the usage ei-bidʰ
gɔs (8) would be more frequent than ei gɔs (8a). Interrogative pronouns
are generally used as determiners without any DTV being added to them.
The only exception is the personal interrogative kʊn, which is capable
of taking all the DTVs in determiner as well as in pronoun functions. There
182 North East Indian Linguistics Volume 3
is, however, an important difference between NPs with an interrogative
determiner and NPs with a demonstrative determiner. In the latter case,
the DTV can be affixed either to the determiner or to the head; in the
former case, the DTV can be combined with the determiner only (9)–
(11).
(9)
kʊn-gɔraki manuh tʊma-r deuta-ra?
which-DTV man you-GEN father-N2FAM ?
‘Which man (out of these) is your father?’
(9a) *kʊn manuh-gɔraki
Table 4 Determiner function of three different classes of pronouns
Pronoun classes Pronouns Determiner function
Demonstrative ei- ‘this’ ei-tʊ kɔlɔm mʊ-k nɛ-lag-e.
this-
DTV pen 1-ACC NEG-want-3
‘I don’t want this pen.’
ħei- ‘that’ ħei-tʊ kɔlɔm an-a.
that-
DTV pen bring-2.FAM
‘Bring that pen.’
ħou- ‘that’ ħou-dal ħap-e bʰekuli dʰor-is-e.
that-
DTV snake-NOM frog catch-IMPFV-3
‘That snake has caught a frog.’
i- ‘it’ i-gɔraki manuh mʊ-r kɔkaideu.
this-
DTV man 1-GEN elder brother
‘This man is my elder brother.’
ħi- ‘that’ ħi-gɔraki manuh mʊ-r kʰuradeu.
that-
DTV man 1-GEN uncle
‘That person is my uncle.’
Relative zi ‘who/ zi bostu p
ʊ-wa an-ib-a.
which’ which things find-2.FAM bring-FUT-2.FAM
‘Bring whatever things you get.’
Interrogative
kʊn ‘who’ kʊn manuh-e mʊ-k bisar-isil?
who man-
NOM 1-ACC search- DPAST.3
‘Who is the man that wants to meet me?’
ki ‘what’ tumi ki bostu an-is-a?
you what things bring-
PERF-2.FAM
‘What things have you brought?’
The Referring Systems and the Determinative Elements ... 183
(10) kʊn-kʰɔn kitap tʊma-k lag-e?
which-DTV book you-ACC want-3?
‘Which book (out of these) do you want?’
(10a) *kʊn kitap-kʰɔn
(11) kʊn-dal gɔs kat-im?
which-DTV tree cut-FUT.1?
‘Which tree (out of these) shall I cut?’
(11a) *kʊn gɔs-dal
The NPs above, the references of which are determined by
kʊn + DTV,
are not purely definite in reference but they are less indefinite than an
NP with an interrogative determiner without DTV, as in (12)–(14).
(12) kʊn manuh-e tʊma-k sakori ni-di-e?
which man-
NOM you-ACC job NEG-give-3?
‘Who denies you a job?’
(13) ki kitap tʊma-r priyɔ?
what book you-GEN favourite?
‘What kind of book do you like?’
(14) ki gɔs ataitkoi ʊkʰɔ.
what tree most tall
‘What (species of) tree is tallest?’
Relative pronouns are used as determiners with or without the
DTVs,
as in (15a) and (15b).
(15a) zi manuh b
ɛmari ħi ɔħukʰi.
which man ill he unhappy
‘People who are sick are unhappy.’
(15b) zi-zɔn manuh bɛmari ħi ɔħukʰi.
which-DTV man ill he unhappy
‘The man who is sick is unhappy.’
Most of the indefinite pronouns are used as determiners, some of them
with
DTVs and some in bare forms. NPs with these determiners in bare
forms are indefinite in reference, as in (16) and those with DTVs are
definite, as in (17).
184 North East Indian Linguistics Volume 3
(16) an manuh-ɔk kɔtʰa-tʊ nɔ-kɔ-b-a.
other man-ACC matter-DTV NEG-tell-FUT-2.FAM
‘Don’t tell others about the matter.’
(17) an-zɔn manuh-ɔk kɔtʰa-tʊ nɔ-kɔ-b-a.
other-DTV man-ACC matter-DTV NEG-tell-FUT-2.FAM
‘Don’t tell the other person about the matter.’
3.2 Cardinal numerals
Cardinal numerals (
CN) are generally regarded as a subclass of nouns,
but they are more frequent in determiner functions than in head function
of NPs. In head function they are inflected for functional and other
categories as in (18) and (19). However, in determiner function only DTVs
can be added to them, as in (20) and (21).
(18) sari-e tini-e mil-i ħat hɔi.
four-
NOM three-NOM combine-PART seven be-3
‘Four plus three makes seven.’
(19) pas-tʊ bʰalkoi likʰ-a.
five-DTV properly write-2.FAM
‘Write the five properly.’
(20) mʊ-k du-kʰɔn kitap lag-e.
1-ACC two-DTV book need-3
‘I want two books’
(21) tini-zɔn manuh-e ei kam-tʊ kor-ibɔ par-e.
three-DTV man-NOM this work-DTV do-PART can-3
‘Three (male) persons can do this work.’
In Assamese determinatives not only determine definiteness status or
the generic vs specific reference of a noun, but also determine the number,
gender, status, etc. of the same, as was mentioned in §2. In this way,
-zɔn in (21) determines that the object referred-to by the noun manuh is
three in number, animate in nature, human as a class, masculine in gender
and familiar in status. Therefore, the reference of the noun manuh is not
generic but specific; i.e. tinizɔn manuh means three male persons of
familiar status. Accordingly, it would seem preferable to view such forms
as determinatives.
The choice of DTV is not dependent on the numeral which it is
agglutinated to. Rather, its selection depends upon the class of the noun
The Referring Systems and the Determinative Elements ... 185
the number or amount of which is referred by the numeral or
quantifier.
This class of determiners has some special features. Firstly, the DTV
-ta combines with this class and with other quantifiers but not to other
determiners or nouns; e.g., du-ta ‘two of something’ and kuri-ta ‘twenty
of something’ but not *ei-ta, *kʊn-ta, *manuh-ta, etc. Secondly, the
same CN + DTV forms like du-zɔn, tini-zɔn, etc., which are used as
determiners in an NP – making them indefinite in reference, as in (22a)
– can again be used as DTV; i.e. they can be affixed to nouns, etc. making
the host NPs definite in reference, as in (22b).
(22a)
du-zɔn manuh ah-ib-ɔ.
two-DTV person come-FUT-3
‘Two persons will come.’
(22b) manuh-du-zɔn ah-ib-ɔ.
person-DTV come-FUT-3
‘The two persons will come.’
It is interesting to notice that when affixed to nouns, etc. as
DTV, lower
prominent CNs up to 99 (niranɔbboi) are generally used only after taking
another DTV with them; i.e. manuh-du-zɔn ‘the two men’ and koni pas-ta
‘the five eggs’ are acceptable but *manuh-dui and *koni pas are
unacceptable. However, CNs after 99 are acceptable as DTV only without
another DTV i.e. tɔka-ħɔ ‘the one hundred rupees’, manuh-hazar ‘the
one thousand people’ are acceptable and *tɔka-ħɔ-ta, *manuh-hazar-zɔn
are unacceptable. More interesting is that in DTV use the CN ɛk (one)
obligatorily drops and a DTV alone can mean ‘definite one’ of the noun
referred (e.g. manuh-zɔn ‘the man’, koni-tʊ ‘the egg’, kitap-kʰɔn ‘the
book’ etc.). Finally, though CNs in determiner use are normally used with
DTVs, in a storytelling style the DTV can be dropped, as in (23) and
(24).
(23)
ɛk rɔza as-il.
one king be- PAST
‘There was a king.’
(24) rɔza-r ɛ-hezar hati as-il.
king-GEN one-thousand elephant be- PAST
‘The king had one thousand elephants.’
186 North East Indian Linguistics Volume 3
Nevertheless, the use of CNs as determiner without DTV is very rare
and should be regarded as exceptional.
3.3 Ordinal numerals
In Assamese there are only about three common ordinals, pohila ‘first’,
dʊsʊra ‘second’ and sotʰa ‘fourth’. They are used only as modifiers to
nominal heads, as in pohila bɔhag ‘first day of the month of Bahag’
(U. N. Goswami 1997). Sanskrit ordinals are used in both modifier and
determiner functions, as in (25) and (26).
(25) ei-kʰɔn mʊ-r prɔtʰɔm kitap.
this-DTV 1-GEN first book
‘This is my first book.’
(26) ditijɔ lora-zɔn-ɔk mɔi sini nɛ-pa-ʊ.
second boy-DTV-ACC 1 recognise NEG-get-1
‘I don’t know the second boy.’
However, most Assamese people do not know the Sanskrit ordinals
to a complete extent. They may know from prathama ‘first’ to about
pa
ñcadasha ‘fifteenth’, but only a relatively few people who use them
in writing or similar functions may learn them to a greater extent.
When ordinal numerals are used as determiners one DTV is added
either to it or to its head, as in (27)–(27a) and (28)–(28a).
(27) pr
ɔtʰɔm manuh-gɔraki mʊ-r deuta.
first man-DTV 1-GEN father
‘The first man is my father.’
(27a) prɔtʰɔm-gɔraki manuh mʊ-r deuta.
first-DTV man 1-GEN father
‘The first man is my father.’
(28) mʊ-r ditijɔ lɔra-zɔn daktɔr.
1-GEN second boy-DTV doctor
‘My second son is a doctor.’
(28a) mʊ-r ditijɔ-zɔn lɔra daktɔr.
1-GEN second-DTV boy doctor
‘My second son is a doctor.’
The Referring Systems and the Determinative Elements ... 187
3.4 General ordinals
The form and function of general ordinals used as determiner are the
same as ordinal numerals (29)–(29a).
(29) nat
ɔk-kʰɔn-ɔr ħeħ drissɔ-tʊ bɔr korun.
play-DTV-GEN last scene-DTV very tragic
‘The last scene of the drama is very tragic.’
(29a) natɔk-kʰɔn-ɔr ħeħ-tʊ drissɔ bɔr korun.
play-DTV-GEN last-DTV scene very tragic
‘The last scene of the drama is very tragic.’
3.5 Quantifiers
Quantifiers are generally regarded as adjectives; however, in Assamese
they function mostly as pronouns and determiners. Though quantifiers
can themselves depict the quantity of the substance referred to, some of
them can further take an approximative
DTV when in determiner functions,
as shown in Table 5.
Table 5 Use of DTV with quantifiers
bohut manuh bohut-bilak manuh
many
man many-DTV man
‘many men’
ħɔkɔlʊ kobi ħɔkɔlʊ-bilak kobi
all poet all-
DTV poet
‘all the poets’
ɔ
tɔ tɔka money ɔtɔ-bʊr
so much
tɔka so much-DTV money
‘
so much money’
3.6 Genitive forms
The main function of genitive (or -
r/-ɔr) forms of nouns, pronouns and
adjectives is to determine the reference of an NP. The important point
about possessive phrases in which the possessor noun is in genitive form,
is that although possessive phrases are by their nature specific and
definite, yet in some cases the addition of a DTV is almost obligatory, as
shown in examples (28)–(28a) and (29)–(29a). Because DTVs in Assamese
188 North East Indian Linguistics Volume 3
determine not only the specificity and definiteness of a noun but also
other some factors like the number or quantity, gender, status, size, etc.
(cf. § 2) according to different classes, in order for determination of these
other factors of a noun in a possessive phrase to be realized the use of
a DTV is obligatory.
4. Determinatives
The morphemes that are used to determine the reference of NPs in
Assamese are indefinite in number, heterogeneous in nature and quite
difficult to analyse in many cases. Some aspects of them have been
exemplified above and some are discussed below briefly in terms of three
characteristics: morphological, syntactic and semantic.
4.1 Morphological aspects of determinatives
Some
DTVs are bound morphemes in Assamese, e.g. -tʊ in mas-tʊ ‘the
fish’ and -kʰɔn in kitap-kʰɔn ‘the book’. However, some are free
morphemes (nouns), e.g. gilas ‘glass’ as in gakʰir-gilas ‘the full glass
of milk’ and gʱɔr ‘house’ as in manuh-gʱɔr ‘the family’ (manuh = ‘man’).
Still others are semi-bound morphemes, e.g. akʰi ‘bunch’ as in kɔl-akʰi
‘the bunch of bananas’.
Consequently, it is difficult to decide whether the addition of these
morphemes to a given word in an NP is a process of affixation or
compounding. There is, however, no doubt that each of the constructions
from kitap-kʰɔn to gakʰir-gilas is single word, both phonologically and
grammatically.
In case of some pronoun bases, the addition of a DTV to them seems
to be a derivational process. By affixing -hɔt to ħi (third person singular
inferior) and tɔi (second person singular inferior) we get ħihɔt (third
person plural inferior) and tɔhɔt (second person plural inferior), forms
which have the status of a single grammatical word. However, addition
of the same DTV -hɔt or any other DTVs to noun bases cannot be regarded
as derivation, because lora-hɔt ‘the boys’, deuta-hɔt ‘father and the
others’, kitap-kʰɔn ‘the book’, gakʰir-gilas ‘the full glass of milk’, etc.
are not distinct lexemes but rather grammatical bases or stems to which
case markers may be added.
The Referring Systems and the Determinative Elements ... 189
4.2 Syntactic aspects
Though words carrying determinatives fulfil other criteria of word
recognition, they lack internal stability. That is, some
DTVs have positional
mobility within the NP. They can be added to any of the three constituents
of an NP, determiner, modifier and head. So, for example, (30) can be
like (30a) and (30b); cf. also (27), (27a), (28), 28a), (29), and (29a).
(30) ei-zɔn burha manuh ram-ɔr deuta-k.
this-DTV old man Ram-GEN father-N3
‘This old man is Ram’s father.’
(30a) ei burha-zɔn ram-ɔr deuta-k.
this old-DTV Ram-GEN father-N3
‘This old man is Ram’s father.’
(30b) ei burha manuh-zɔn ram-ɔr deuta-k.
this old man- DTV Ram-GEN father-N3
‘This old man is Ram’s father.’
This is, however, not a general characteristic. For example, when the
determiner is a numeral, as in (21), -zɔn cannot be added to the head or
modifier of the NP, harming the internal stability of the word tini-zɔn
(numeral + -zɔn). That is, (21) cannot be rewritten as (21a).
(21a) *tini manuh-z
ɔn-e ei kam-tʊ kor-ibɔ par-e.
Words with the internal composition NUMERAL+DTV can, in some cases,
be used as determiner to a noun and also as determinative to the same
noun. For example, duzɔn manuh ‘two men’ and manuhduzɔn ‘the two
men’ are both acceptable (cf. (22a)–(22b)). Their referential meanings
are, however, different. The former is indefinite but the latter is
definite.
A few DTVs are fully mutually exclusive, some are partly so and some
others are not at all mutually exclusive. -bʊr is fully replaceable with
-bilak (cf. (4a)), -zɔn can be replaced by -tʊ in some uses (cf. (28)) but
not in some others (cf. (27)). On the other hand, -dal cannot be replaced
by -kʰɔn. The selection of alternative DTVs depends on many factors in
some cases. For example, in the case of human nouns, selection depends
upon the impression of the speaker about the person(s) referred to. In
example (4) above, the Mughals are referred to definitely with a
190 North East Indian Linguistics Volume 3
DTV -bʊr in which a neglecting attitude is expressed. With a more
respectful attitude, the same referent can be referred by -bilak. But with
full respecting attitude they can be referred by -ħɔkɔl. So also, to refer
a man definitely, as in (30), the speaker may use any one of the forms
manuh-tʊ, manuh-zɔn, or manuh-gɔraki according to his attitude to the
man referred-to. Even the non-personal DTV -dal can be affixed to manuh
with a highly neglectful or even angry mind-set.
With only two exceptions, all DTVs can be affixed to nouns, pronouns,
adjectives and numerals. -tʊ cannot be affixed to numerals when they
are used as quantifiers. -ta can be affixed only to numerals in quantifier
function, and not to any other words (cf. §3.2).
4.3 Semantic aspects
When used as a
DTV, the lexical meanings of free morphemes are
backgrounded and each takes on grammatical or functional meanings.
Gilas ‘glass’ in panigilas, for example, doesn’t refer to a glass but rather
indicates definite reference to a quantity of water kept in a glass.
Therefore the English translation of panigilas cannot be ‘the glass in
which the water is contained’ but is rather ‘the quantity of water that
would be kept in a glass’. When agglutinated, DTVs generally mark any
NP either as specific, definite or as specific, indefinite, as in (2a–2b). On
the other hand, absence of these morphemes or any other determiner
marks an NP as generic, as in (1). Another common grammatical meaning
of DTVs is that they mark the quantity or amount of the substance referred-
to or at least help a quantifier to do the same. For example, manuh-tʊ
means one definite person, manuh-duzɔn means two definite persons,
manuh-bʊr means some definite persons, and so on.
In case of countable substances, the numerals serially up to 100 and
also the numerals for thousand, lakhs (hundred-thousands), etc. are used
if and when the exact number is important or possible to give. Otherwise
one approximative plural marker (e.g. bʊr, bilak, mɔkʰa, etc.) is used;
for example, manuh-bʊr or manuh-bilak. In case of uncountable
substances, either approximative plural markers or unit words (e.g. ħer,
pʊwa, gɔnda, pʊn, mitar, litar, kezi, kuintɔl, etc.) or the name of the
container of the substance (as an amount signifier) is used. E.g., gakʰir-
pʊwa ‘the ¼ litre of milk’, tamʊl-gɔnda ‘one unit of four betel nuts’,
tel-litar ‘the one litre of oil’, gakʰir-gilas ‘the full glass of milk’, etc.
It should be noted that when the exact number of a countable substance
is almost impossible to give or unimportant in the context, then they are
The Referring Systems and the Determinative Elements ... 191
referred to like uncountable substances. For example, saul-bɔsta ‘the
full bag of rice’, manuh-gari ‘the full car of people’, kɔmɔla-pasi ‘the
full basket of oranges’, etc. Besides these two common meanings,
different classes of DTVs may mark other different features like gender
(in case of human and animals), status (in case of humans)
3
and size,
etc. (in case of nonliving things) of NPs. The word manuhzɔn in (30)–
(31), for example, is definite in reference, single in number, male in sex
and familiar in status. kitap-kʰɔn in (32) is definite in reference, single
in number and flat in size (cf. §2).
(31) manuh-z
ɔn bɔr dʰunija.
man-DTV very handsome
‘The man is very handsome.’
(32) kitap-kʰɔn bɔr muljɔban.
book- DTV very valuable
‘The book is extremely valuable.’
By using different
DTVs even different positions of a substance can
be indicated. For example, a rope in open position is referred to as rosi-
dal. If it is kept as a bundle with a knot then it is referred to as rosi-
muthi. And if the same is kept as a bundle without a knot then it is
referred to as rosi-kʊsa. If on the other hand the rope is short then it is
referred to as rosi-tukura.
Though the reference of an NP is mostly dependent on these
morphemes, they have no such meaning within them. In fact, the reference
of an NP depends, generally, on the presence or absence of these
morphemes (cf. (1), (2a–2b), (9), (10), (15a–15b), (16), (17)), or in some
cases when they are present, then on the position of these morphemes
within the noun phrase (cf. (22a–22b)).
It was mentioned earlier that some of the DTVs are mutually exclusive
but some are not (§4.2). In use, they show more idiosyncrasies than
regularities. Some can be added to certain words but not to others; e.g.,
-zɔn, -gɔraki can be combined with human animate nouns not with non-
human animate nouns ((cf. (2a)–(2b)). However, this doesn’t mean that
3
For example, the translation of the English NP ‘the man’ may be of three
different forms according to the social status of the man referred to: manuh-t
ʊ,
manuh-z
ɔn, or manuh-gɔraki, the last term being the highest in status. This
difference of status is also maintained in verbal endings in case of 2
nd
persons.
192 North East Indian Linguistics Volume 3
they are used according to the class of the nouns, because -zoni can be
combined with both human and nonhuman classes (e.g. sʊwali-zoni ‘the
girl’, goru-zoni ‘the cow’) and -tʊ, -bʊr, and -bilak can be combined
with both living and nonliving classes of nouns (e.g. manuh-tʊ ‘the man’,
am-tʊ ‘the mango’). -bʊr, -bilak, -kʰini, etc., can be combined with both
countable and uncountable nouns (cf. Table 5), but -zɔn, -kʰɔn, -dal,
etc., can be combined with countable nouns only (cf. Table 3) and -tʊpa,
-solu, etc., can be combined with uncountable nouns only. Some can
show gender contrast in the case of human nouns (-zɔn, -zoni), some
cannot (-tʊ, -gɔraki). -dal is used with nouns referring long substances
(bah-dal, pensil-dal) but with similar substances (such as kɔlɔm ‘pen’)
-tʊ is used. These and other examples given above show that DTVs are
multipurpose in showing contrast among NPs in a lot of different features
and it is not possible to present them in a grammar as a paradigm as it
is possible, for example, with case markers in Assamese.
5. The history of the system
The determination of NP reference by affixing DTVs to its head noun or
pronoun or to any of its dependent words was not found in Old or Middle
Indo-Aryan languages (Kakati 1941; Medhi 1978). Many linguists
therefore talk about the influence of Sino-Tibetan languages, especially
the influence of the geographically adjacent Boro (Kachari) and Garo
languages (Medhi 1978: 435) as the possible reason behind the evolution
of DTVs in Assamese and other some New Indo-Aryan (NIA) languages,
seeing the similarities between numeral classifiers of Boro, Garo and
other classifier languages and the DTVs of Assamese, Bangla, etc., about
which we mentioned with some examples earlier. We have also shown
there that these two features of classifier languages and these New Indo-
Aryan languages are quite different in function, due to which we have
called them determinatives in Assamese instead of classifiers. On the
other hand, DTVs of these languages are notably different in morphology
from numeral classifiers. Numeral classifiers can be added only to
numerals; however, besides numerals DTVs can be added to nouns,
pronouns and even to adjectives, as we have exemplified in (30)–(30b).
Above all, in some classifier languages like Boro there are, though small
in number, elements which are more similar both morphologically and
syntactically with the DTVs of Assamese and other NIA languages.
Examples (33a)–(34b) from Assamese and Boro show this.
The Referring Systems and the Determinative Elements ... 193
(33a) mɔi ɛ-ta mas pa-l-ʊ.
1 one-DTV fish get-PERF-1
‘I have got a fish.’ (Assamese)
(33b) aŋ na ma-se mʚn-bai.
1 fish CL-one get-PERF
‘I have got a fish.’ (Boro)
(34a) mɔi mas-tʊ pa-l-ʊ.
1 fish-DET get-PERF-1
‘I have got the fish.’ (Assamese)
(34b) aŋ na-kʰow mʚn-bai.
1 fish-DTV.ACC get-PERF
‘I have got the fish.’ (Boro)
The morpheme -
kʰow of Boro in sentence (34b) is almost the same
as Assamese DTV -tʊ; indeed, the Boro morphemes are likely to have
been borrowed from Assamese, Bangla, etc., and therefore could be called
a DTV, though there are some minor differences between Assamese and
Boro DTVs. For these reasons, we suggest that evidence of the influence
of Boro “classifiers” in the evolution of the system of referring nouns
by affixing DTVs in Assamese, Bangla, etc., is very thin.
Some other linguists think that there could be an element of
Austroasiatic influence in the evolution of this feature of Assamese due
to some similarities among these two systems (Kakati 1972: 281).
However, these claims remain unproven and need further research.
The view adopted here is different from that of the above. It rather
seems more likely that the evolution of the system of noun reference was
within the Indo-European family itself, and did not result from language
contact. If English can produce the definite article ‘the’ to refer to a noun
definitely from an unstressed demonstrative (Hawkins 1978), then why
is the evolution of such a system for the same purpose to be considered
improbable in the case of Assamese? To compensate for the loss of
grammatical gender, number, etc., of the highly synthetic Old Indo-Aryan
language from which Assamese evolved through the Middle Indo-Aryan
stage, and as a result of Assamese becoming relatively more analytical,
this process of affixing DTVs to NPs may have evolved in this language
as a means of functional compensation. It is worth noting that all
morphemes that are used as DTVs in Assamese were originally free
morphemes, namely lexemes, at an earlier stage of the language, some
of which having lost their word status later on. This process of using
194 North East Indian Linguistics Volume 3
nouns as referential modifiers begins in the NIA stage only, and that is
why the origin of these morphemes can be ascertained relatively easily.
No doubt, there are some morphemes (nouns) used as DTVs which have
been borrowed from other languages (as is the case in general with the
lexicon of Assamese). But it is also seemingly true that none of them
came to Assamese as DTVs. Rather, they came as nouns, and after being
fully nativized they started to be used as DTVs with other nouns.
Perhaps the process in Assamese (and in other some NIA languages)
began with words of OIA origin, which were compounded with some
nouns giving them a weak sense of definiteness. jana, gana, brinda,
khanda/khandaka, kunda, patra, etc., are some such nouns, the changing
forms of which are frequently used as DTVs in present-day Assamese. It
seems therefore to be that the process of using new nouns as DTVs of
other nouns is still in force. As for example, the English origin words
gilas < ‘glass’, plet < ‘plate’, kap < ‘cup’, which have probably been
inducted into Assamese not before 120 years ago, are frequently used
DTVs in present days, as in panigilas, bundiaplet ‘the full plate of
sweetmeat’, sah-kap ‘the full cup of tea’, etc.
The view of Emeneau (1956) relating to this phenomenon in North-
Eastern Indo-Aryan languages, quoted in B. K. Kakati (1972), is worth
reproducing here.
He holds that the phenomenon of the use of the enclitics “is of limited
areal range, but appears in all three families”, viz, Indo-Aryan, Dravid-
ian and Munda, “having spread from Indo-Aryan, though it is not an
Indo-European phenomenon…My reconstruction…is that the construction
(so far as India is concerned) is originally Indo-Aryan. It spread thence
to the other languages as a total construction consisting of numerical +
classifier, and then was elaborated in some of the languages with native
material, the native numerals, native morphemes as additional classifiers,
etc.” (Emeneau (1956), quoted in Kakati (1972: 10, 14))
Abbreviations
ACC Accusative
CL Classifier
CN Cardinal numeral
DTV Determinative
DPAST Distant past tense
FAM Familiar
The Referring Systems and the Determinative Elements ... 195
FUT Future
GEN Genitive
IMPFV Imperfective
NEG Negative
NOM Nominative
NP Noun Phrase
PART Participle
PERF Perfective
PAST Past tense
1 First person
2 Second person
3 Third person
N2FAM Nominal second person familiar
N3 Nominal third person
References
Agarwala, J. P. (1981). Jyoti Prasad Rasanavali. Gauhati, Asom Prakashan
Parishad.
Bhattacharya, P. C. (1977). A Descriptive Analysis of the Boro Language.
Guwahati, Gauhati University Press.
Bhuyan, S. K. (1997). “Saraighat ase saraighat nai.” In M. Neog, Ed., Snatakar
Kathabandha. Guwahati, Gauhati University Press: 227–37.
Chatterjee, S. K. (1970). [1926]. Origin And Development of Bengali Language.
London, George Allen & Unwin.
Christopherson, P. (1939). The Articles: A Study of Their Theory and Use in
English. Copenhagen, Munksgaard.
Dasgupta, P. (2003). “Bangla.” In G. Cardona and D. Jain, Eds., The Indo-Aryan
Languages. London, Routledge: 386–428.
Dutta Baruah, P. N. (1978). A Comparative Analysis of the Morphological Aspect
of Assamese and Oriya. PhD Thesis. Guwahati, Gauhati University.
_____. (1994). A Comparative Analysis of the Syntactic Aspects of Assamese
and Bengali. PhD Thesis. Mysore, University of Mysore.
Emeneau, M. B. (1956). “India as a linguistic area.” Language 32: 1–16.
Goswami, G. C. (1968). Classifiers and Quantifiers in Assamese. In B.
Krishnamurti, Ed., Studies in Indian Linguistics. Poona, Centre for Advanced
Study in Linguistics, Deccan College: 110–17.
_____. (1982). Structure of Assamese. Guwahati, Gauhati University Press.
_____. (1990). Asomiya Byakaranar Moulik Bisar. Guwahati, Bina Library.
Goswami, G. C and J. Tamuli. (2003). “Asamiya.” In G. Cardona and D. Jain,
Eds., The Indo-Aryan Languages. London, Routledge: 429–84.
196 North East Indian Linguistics Volume 3
Goswami, U. N. (1997). Asomiya Bhasar Byakaran. Guwahati, Mani-Manik
Prakash.
Hawkins, J. A. (1978). Definiteness and Indefiniteness. London, Groom Helm.
Heinrichs, H. (1954). Studien zum bestimmten Artikel in den germanischen
Sprachen. Giessen, Schmitz.
Higham, J. (1972). “Relational ‘the’.” Edinburgh, University of Edinburgh.
Unpublished manuscript.
Hodler, W. (1954). Grundzüge einer germanischen Artikellehre. Heidelberg,
Winter.
Jespersen, O. (1949). A Modern English Grammar on Historical Principles, Vol.
VII. Copenhagen, Munksgaard.
Kakati, B. K. (1972 [1941]). Assamese: Its Formation and Development,
Guwahati, Lawyers Book Stall.
Kalita, J. C. (2004). Nouns and Nominalisations in Assamese: A Microlinguistic
Study. PhD Thesis. Gauhati University Department of Linguistics.
Matthews, P. H. (1981). Syntax. Cambridge, Cambridge University Press.
Medhi, K. (1978). Asomiya Byakaran Aru Bhasatatwa. Guwahati, Asam Prakasan
Parishad.
Quirk, R., S. Greenboum, G. Leech, and J. Svartvik. (1985). A Comprehensive
Grammar of the English Language. London, Longman.
Russell, B. (1905). “On denoting.” Mind 14: 479–93.
Smith, C. S. (1964). “Determiners and relative clauses in a generative grammar
of English.” Language 40: 37-52.
Strawson, P. F. (1950). “On referring.” Mind 59: 320–44.
Yotsukura, S. (1970). The Articles in English: A Structural Analysis of Usage.
The Hague, Mouton.
Copular Sentences in Asamiya 197
12
Copular Sentences in Asamiya
1
Runima Chowdhary
Gauhati University
1. Introduction
Copular constructions have since long drawn the attention of scholars
engaged in explorative studies of a particular language or across languages
because of the intriguing features that set copular constructions apart
from other types of constructions in language(s). The term “copula” (COP),
as a constituent of a copular construction, has been used in typological
studies to refer to any morpheme (affix, particle or verb) that links or
“couples” a subject with a copula complement in a “family” of
constructions, collectively often referred to as “predicate nominal
constructions” (Payne 1997: 111–14). In contrast to the dimension of
verb complementation distinguished in terms of the number of objects
present, a copular construction, characterized by a copula, contains an
obligatory copula complement (CC)/predicative complement (PC),
predicated of the copula subject (CS) (Dixon 2004; Huddleston & Pullum
2002). Cross-linguistically, “the most basic copular construction” is used
to encode the meanings of classification (inclusion/group membership)
or identification (specification/equation
2
of two participants “normally
1
Asamiya, pronounced [ɔħɔmija] in the vernacular medium, stands for the As-
samese language (as well as the people of Assam).
2
Proper inclusion is when “a specific entity is asserted to be among the class
of items specified in the nominal predicate”, while equative clauses “assert
that a particular entity is identical to the entity specified in the predicate
nominal” (Payne 1997: 114).
198 North East Indian Linguistics Volume 3
encoded as noun phrases,” others being “less basic
3
” (Curnow 2000: 2).
Typological studies show that besides that of inclusion and equation,
copular sentences are used to express other assertions like that of location,
attribution, possession or existence of a referent (Payne 1997; Declerck
1988; Pustet 2003). The distinctively different categories of overt copulas
that have been attested across languages are mostly verbal, pronominal
and particle (Stassen 1997: 85, cited in Pustet 2003: 45; Payne 1997:
118–119).
The present study is a descriptive account of copular sentences in
Asamiya (Assamese). The work explores the different grammatical
strategies of copular constructions that the language employs for encoding
semantically similar yet functionally different as well as “pragmatically
marked” structures
4
. Keeping in mind the constraints of space, the paper
focuses on explicating the forms and functions of the mono-clausal
copular sentences only. Additionally, the study attempts to unravel the
influence of pragmatics in determining the choice of one grammatical
strategy as against another.
The structure of the paper is as follows. §2 discusses the types of
copular sentences used to encode various semantic assertions. §3
examines the distinctive morphosyntactic properties of each type. §4
describes the functions and uses of each type of copular sentences
distinguishing one from the other as well as from its analogous non-
copular counterpart. §5 summarizes the findings.
2. Types of copular sentences
Asamyia copular sentences broadly fall into two types: primary, a
structurally more elementary type and quasi, a non-basic or structurally
less elementary type. The primary type has two subtypes – a non-overt
one with a zero copula, labeled zero copular sentence, and an overt one,
with an inflected verbal copula, labeled primary overt copular sentence.
3
“Less basic” is the one of the two constructions essentially used for encoding
the same information, if it adds additional meaning, which is missing from the
first (i.e., the basic one) (Curnow 2000: 2).
4
Clauses with a focused constituent, not evident in other clauses, are “prag-
matically marked” (Payne 1997: 261–68). A contrastive focus in a clause can
be marked by different devices, viz. differences in intonation or constituent
order, use of formatives (particles or affixes), clefting and so on (1997: 268–
82).
Copular Sentences in Asamiya 199
The two overt inflected verbal copulas characterizing the primary
overt copular sentence are: as- ‘be’ and hɔ- ‘become; be,’ the former of
which is homophonous with the lexical verb as- ‘possess; have; own’
or ‘exist’ as well as the progressive auxiliary and the latter, with the
lexical verb hɔ- ‘happen; occur’ and the passive auxiliary in the language
5
(discussed in §3.1). Unlike the lexical verbal counterparts with semantic
contents, the copulas are “semantically empty” in the sense of Payne
(1997: 115) in that they cannot themselves express the intended meanings
of the predications unless accompanied by some other lexical items. The
quasi copular sentence, on the other hand, is marked by the presence of
a kind of fossilized verbal particle hɔl or hoise ‘BE,’ labeled quasi copula
(
COP
Q
) in this study.
The various elementary assertions expressed by primary copular
sentences are as in the following.
(a) attribution of some inherent properties or characteristics to some
entity
(b) establishing equation between two entities
(c) specifying a possessional relationship with some possessed entity
(d) identifying some specified location of some entity
(e) expression of some physical or mental state experienced by some
entity
(f) statement on comitativeness/accompaniment or benefactiveness of
some entity
(g) specifying some specified physical condition of some entity
Table 1 Types of copular sentences
Primary Overt type
non-overt type
Zero copular sentence Primary overt Quasi copular
copular sentence sentence
The quasi copular sentence is characterized by an overt non-inflected
verbal copula. These three subtypes of copular sentences are schematically
represented below.
5
A similar phenomenon is reported by Pustet (2003: 70) relating to Basque,
which has two copulas, one homophonous with the lexical verb ukan
‘have’
and the other, with izan ‘exist’, both being also used as auxiliaries in certain
morphosyntactic contexts.
200 North East Indian Linguistics Volume 3
Both the primary copular sentences are used to share most of the
elementary assertions with each other, but each encodes some specific
additional semantic overtones, not possible in the case of the other.
However, as opposed to the primary types, the quasi copular sentence is
used to express an identifying relation between two referents, one of
which is highlighted or focused upon by the use of a fossilized verb.
3. Morphosyntactic properties of copular sentences
The basic morphosyntactic parameters that are taken into consideration
for distinguishing the various strategies of copular constructions in
Asamiya are: 1) inflectional paradigms of copulas, 2) constituent order,
3) constituent structure and case marking of constituents.
3.1. Inflectional paradigms of copulas
A verbal paradigm in the language may contain the imperfective suffix
-(
i)s and/or past tense suffix -(i)l and a person suffix from a set of five,
inclusive of the tripartite second person markers reflecting the honorific
hierarchy, viz. -ʊ ‘1’, -e ‘2
1
’, -a ‘2
2
’, -ɔ ‘2
3
’ and -e ‘3’. A verb in the
language is prototypically cross-referenced with the subject in terms of
person.
The copula hɔ-, like other verbs in the language, can be potentially
inflected for all the five paradigms of the fused categories of aspect
and/or tense and person, viz. present indefinite, present imperfective,
simple past, past imperfective and future. By contrast, as-, like its
homophonous lexical counterpart, is restrictively inflected for only two
paradigms: present indefinite and simple past. The following Tables
2a–2b illustrates the inflectional paradigms of the two copulas representing
the sequential order of verbal base + (aspect) + (tense) + person.
One of the striking features of the primary copulas, which distinguishes
them from their homophonous lexical verbal counterparts used in non-
copular sentences, is morphological. While the copulas can be inflected
for all person categories, the lexical verbs as- ‘exist’ and hɔ- ‘happen;
occur’ are invariant in form in terms of the category of person.
The data as in (1) and (2) exemplify primary copular sentences with
the copula as- exhibiting cross-referencing with the CS in terms of
person.
Copular Sentences in Asamiya 201
(1) [
NP
mɔe]
CS
[
ADVP
g̤̤ɔr-ɔt]
CC
[
V
as-ʊ]
COP
1.SG home-LOC be -1
‘I am at home.’
(2) [
NP
tumi]
CS
[
AP
ħɔsakoi bɔr duhħahoħi]
CC
2
2
SG really very audacious
[
V
as-il-a]
COP
be-PST-2
2
‘You were indeed very audacious.’
Compare the homophonous intransitive verb as-
‘exist’ in an
existential sentence which exclusively co-occurs with a 3
rd
person
subject.
(3) [
NP
b̤ɔgɔban]
S
[
V
as-e]
INT
God exist-3
lit. ‘God exists,’ i.e. ‘There is a God.’
Table 2 Finite inflectional paradigms of primary copulas
(a)
hɔ-
Person PRS.IND PRS.IPFV PST PST.IPFV FUT
categories
1 ho-ʊ ho-is-ʊ ho-l
-ʊ ho-is-il-ʊ ho-m
2
1
hɔ-e ho-is-e ho-l ho-is-il-(e) ho-b-o
2
2
hʊ-a ho-is-a ho-l-a ho-is-il-a ho-b-a
2
3
hɔ-ɔ ho-is-ɔ ho-l-i ho-is-il-i ho-b-i
3 hɔ-e ho-is-e ho-l ho-is-il-(e) ho-b-o
(b) as-
Person PRS.IND PRS.IPFV PST PST.IPFV FUT
categories
1 as-ʊ as-il-ʊ
2
1
as-e as-il-(e)
2
2
as-a as-il-a
2
3
as-ɔ as-il-i
3 as-e as-il-(e)
202 North East Indian Linguistics Volume 3
Another homophonous lexical verb as- ‘possess; have; own’ used in
co-occurrence with a possessor subject, obligatorily assumes a default
3
rd
person form, irrespective of the person category of the subject, as
exemplified below.
(4) [
NP
mʊ-r]
S
[
NP
du-kʰɔn gari]
DO
[
V
as-il]
TRN
1.SG-GEN two-CL car possess-PST.3
‘I had two cars.’
As can be seen, the lexical verb as- ‘possess; have; own’ inherently
presupposes two arguments: a possessor subject and a possessed object
complement. The resultant sentence is, hence, a transitive one.
In another non-copular rendition, as- functions as the progressive
auxiliary. It is worth noting here that unlike the other lexical verbs with
the same form, as- as the progressive auxiliary is the only non-copular
form which exhibits cross-referencing with respect to person, as shown
in (5) in the following.
(5) [
NP
ami]
S
[
V
porh-i as-il-ʊ]
TRN
1.PL study-NF AUX-PST-1
‘We were studying.’
The following exemplifies copular sentences characterized by the
copula
hɔ-
6
which exhibits agreement with the given CS.
(6) [
NP
tumi]
CS
[
AP
bohut ħɔkɔt]
CC
[
V
ho-l-a]
COP
2
2.
SG
very fat be-PST-2
2
‘You have put on a lot of weight.’
6
It is worth noting here that in addition to its non-copular use as a lexical
or auxiliary verb, hɔ- as an invariant particle hɔe may be used as a question
tag in colligation with the question marker -ne in tag questions in the
language (a)–(b).
(a) [
NP
ħi]
S
[
V
n-ah-il-(e)]
INT
[
V
hɔe-ne]
TAG
3
3
. M.SG NEG-come-PST-3 PART-Q
‘He didn’t come, did he?’
(b) [
NP
tumi]
S
[
V
za-b-a]
INT
[
V
nɔ-hɔe-ne]
TAG
2
2
SG go-FUT-2
2
NEG-PART-Q
‘You will go, won’t you?’
Copular Sentences in Asamiya 203
(7) [
NP
lora-tʊ]
CS
[
NP
kʰeluwoi]
CC
[
V
ho-b-o]
COP
boy-CL player be-FUT-3
‘The boy will be a player.’
In contrast, a non-copular sentence with the intransitive lexical verb
hɔ- ‘happen; occur’, invariably in the 3
rd
person form, obligatorily
requires a 3
rd
person subject as in (8)
7
.
(8) [
NP
ɛ-ta daŋɔr g̤̤ɔtɔna]
S
[
V
ho-l]
INT
one-CL big event happen-PST.3
‘A significant event took place.’
That cross-referencing exhibited by the copula
hɔ- is lacking in non-
copular verbal counterparts is further evidenced in the following.
(9) [
PP
ama-r (dara)]
S
[
NP
sitʰi-bʊr]
DO
1.PL-OBL
8
(by) letter-CL
[
V
di-a ho-l]
TRN
deliver-NF AUX-PST.3
‘The letters were delivered by us.’
The sentence as in (9) is a passive sentence with hɔ- as the passive
auxiliary canonically in default 3
rd
person form, with its subject being
invariably relegated to the oblique complement of the postposition (either
overt or covert).
7
Another non-copular use of hɔ- includes its role as an integral verbal
constituent of a conjunct verb in conjunction with a nominal or adjectival
constituent.
[
NP
kagɔz]
S
[
V
ħeħ ho-is-e]
INT
paper end become-
IPFV-3
‘The stock of papers has run out.’
8
The gloss OBL(ique) is used here to distinguish the homophonous case-marker
of the oblique complement in a POSTP in the language from that of the
GEN(itive) case marker of the possessor complement in a NP as in (4) on the
basis of their differences with respect to syntactic distribution and semantic
function.
204 North East Indian Linguistics Volume 3
The two quasi copulas hol and hoise as in (10) are non-inflected
fossilized verbal forms: the former, homophonous with the 3
rd
person
past tense and the latter, with the 3
rd
person present indefinite form
respectively (see Table 2a).
(10) [
NP
ħukh-ɔr utsɔ]
CS
happiness-GEN source
[
V
hoise/hol]
COP
Q
[
NP
boisitrɔ]
CC
BE variety
‘The source of happiness is variety.’
To sum up the basic findings at this juncture, unlike the homophonous
verbal counterparts, the primary overt copulas exhibit cross-referencing
with the given subjects that set them apart from the former used in non-
copular constructions. In contrast, the quasi copulas are invariant and
morphologically indivisible in form in terms of tense, aspect and person
categories.
3.1. Constituent orders
Like other Indic counterparts, Asamiya also canonically follows SOV
order. However, any possible alternative orders are permissible as
thematic variants of the same proposition, with the grammatical relations
between core constituents being identifiable by the rich case system of
the language.
The
constituent order of primary copular sentences also follows the
canonical SOV order of the language, with the CS invariably occurring
in the sentence-initial position, followed by the CC and the verbal copula
(COP) in sentence-final position as in (11).
(11) [
NP
tekʰet]
CS
[
NP
ħikkʰɔk]
CC
[
V
as-il]
COP
3
1
.M.SG teacher be-PST.3
‘He was a teacher.’
The zero copular sentence, on the other hand, is formed by
juxtaposition of CS and CC in a sequential order, with the position of
the verb being empty as in (12).
(12) [
NP
lora-tʊ]
CS
[
AP
ħahoħi]
CC
boy-CL brave
‘The boy is brave.’
Copular Sentences in Asamiya 205
In a striking contrast to that of the primary overt type, the quasi-
copular sentence is characterized by the non-canonical SVO order. The
CS and the CC of this type are conjoined by the fossilized verbal copula
(COP
Q
) hoise or hol ‘BE,’ used to focus or highlight some element as in
(13) where the focused constituent is canonically the post-verbal NP,
i.e., the CC.
(13)
[
NP
ɔħɔm-ɔr prɔd̤an ħɔmɔsja]
CS
Assam-GEN primary problem
[
V
hoise/hol]
COP
Q
[
NP
banpani]
CC
BE floods
‘The primary problem of Assam is floods.’
Unlike the non-copular counterparts, the copular sentences are
resistant to inversion of constituent order, as substantiated by the
following examples
.
The sentence (14), resultant from inversion of the CS and CC of the
primary overt copular sentence as in (11), is an ill-formed one.
(14) *[
NP
ħikʰɔk]
CC
[
NP
tekʰet]
CS
[
V
as-il]
COP
teacher 3
1
.SG be-PST.3
On the other hand, the inversion of the CS and the CC of a zero
copular sentence as in (12) renders a phrasal interpretation instead of a
sentential one, as in (15).
(15) [
NP
[
AP
ħahoħi] [
NP
lora-tʊ]]
brave boy-CL
‘the brave boy’
However, a quasi copular sentence differs from the primary overt
copular sentences in that the positions of the CS and the CC can be
inverted without affecting the basic proposition.
(16) [
NP
banpani]
CC
[
V
hoise/hol]
COP
Q
floods BE
[
NP
ɔħɔm-ɔr prɔd̤an ħɔmɔsja]
CS
Assam-GEN primary problem
‘Floods are the primary problem of Assam.’
206 North East Indian Linguistics Volume 3
To recapitulate, the primary copular sentence of both the subtypes
follows the canonical SO(V) order of the language. By contrast, the quasi
copular sentence exhibits SVO order. While the inversion of CS and CC
in a primary copular sentence may affect its interpretation or render an
ill-formed sentence, it does not affect the basic proposition of a quasi
copular sentence.
3.2. Constituent structure and case marking of constituents
The constituent structure of CS, i.e., the subject of copular sentence,
either of the overt or non-overt type, is no different from that of the other
types of sentences in that any NP, be it a pronominal, proper noun,
common noun, possessive construction or a nominalized clause, can be
used as the subject of a copular sentence.
It is worth noting that case marking of constituents in copular
sentences is highly sensitive to the given semantic roles. Like the subject
of an intransitive sentence
9
, a CS in Asamiya in the semantic role of a
theme is morphologically unmarked for case as exhibited by the copular
sentences encountered so far as in (1), (2), (6), (7), (10), (11) and (12)
above.
But a CS is marked for genitive case when it functions in the semantic
role of an experiencer. The motivating factor influencing the genitive
case marking of the CS is the presence of a nominal CC encoding some
physical/mental states of a referent like oħukʰ ‘illness’ (as in (17)),
anɔndɔ ‘happiness’, dukʰ ‘sadness’, kɔstɔ ‘hardship’ and so on.
Table 3 Constituent orders of copular sentences
Zero copular Primary overt copular Quasi copular
sentence sentence sentence
[NP]
CS
+[XP]
CC
[NP]
CS
+[XP]
CC
+[V]
COP
[NP]
CS
+[V]
COP
Q
+[XP]
CC
Note that (16) represents the same proposition as that of (13), but
with the inverted CS and CC.
Table 3 in the following summarizes the constituent orders of the
three subtypes of copular sentences.
9
In terms of Huddleson and Pullum (2002: 77–78) copular sentences are a type
of intransitives, designated “complex-intransitives” as opposed to “ordinary
intransitives.”
Copular Sentences in Asamiya 207
(17) [
NP
lora-tʊ-r]
CS
[
NP
oħukʰ]
CC
boy-CL-GEN illness
‘The boy is ill.’
This accounts for the ill-formedness of (18) characterized by an
adjectival CC, which does not collocate with an experiencer CS.
(18) *[
NP
lora-tʊ-r]
CS
[
AP
oħukʰia]
CC
boy-CL-GEN ill
The CC slot of the primary copular sentence may be filled by any
phrasal category, viz., NP, AP, ADVP or POSTP. An NP functioning as
the CC is, by default, morphologically unmarked as in (19)
10
.
(19) [
NP
deuta]
CS
[
NP
ei onustʰan-ɔr ħɔb̤apoti]
CC
father this organization-GEN president
‘(My) father is the president of this organization.’
A nominal CC is marked for genitive case in the possessor role. The
nominal functioning as the CC of this type preferably has to be a human
referent as in (20).
(20) [
NP
kitap-kʰɔn]
CS
[
NP
mʊ-r]
CC
book-CL 1 SG-GEN
‘The book is mine.’
A nominal category functioning in the semantic role of a location in
an ADVP as in (21) or a relator noun (Delancey 1997)
11
like ʊpɔr ‘top’,
functioning as the head of a POSTP as in (22), is marked locative case
as exemplified in the following.
10
It is important to note here that intonation plays a vital role in resolving
ambiguities in a juxtaposed CS and CC, in case both the slots are filled by
NPs. The CS, prototypically marked by a rising intonation is distinguished
from the CC, marked by a falling intonation.
11
The relator noun refers to a class of postpositions like top, or front as in
on top of, or in front of in English, developed from nouns through gram-
maticalization (Delancey 1997: 58).
208 North East Indian Linguistics Volume 3
However, one of the striking features of a quasi copular sentence is
that only a NP, unmarked for case, can occur as its CC as in
(10), (13).
To summarize this section, the CS of all the types of copular sentences,
barring the one with an experiencer role in the genitive case, is unmarked
for case. While the CC slot in the primary copular type can be filled by
any phrasal category, marked for case in accordance with the given
semantic role, the CC position in the quasi copular sentence can be filled
exclusively by a non-case-marked NP.
4. Functions and uses of copular sentences
The use of different subtypes of copular sentences may signal different
temporal interpretations and semantic connotations depending on context.
Following the lines of assumptions in Akmajian et al. (2004: 388), the
notion of “context” has been assumed in this study in an extended sense
that encompasses not only “the previous and anticipated utterances in
Table 4 Case marking of CS and CC in primary copular sentences
Syntactic Function Syntactic Semantic Case-marking Examples
category role
Copula subject
NP
Theme Unmarked 1, 2, 5, 6, 7, 10
(CS) Experiencer Genitive 17
NP
Theme Unmarked 7, 10, 17, 19
Copula Possessor Genitive 20
complement (CC)
ADVP
Location Locative
1, 21
POSTP 22
(21) [
NP
tɛʊ-r g̤̤ɔr]
CS
[
ADVP
zʊrhat-ɔt]
CC
3
2
.SG-GEN home Jorhat-LOC
lit. ‘His home is in Jorhat.’ i.e., ‘He hails from Jorhat.’
(22) [
NP
kitap-kʰɔn]
CS
[
PP
mɛz-ɔr ʊpɔr-ɔt]
CC
[
V
as-e]
COP
book-CL table-OBL top-LOC be-3
‘The book is on the table.’
A schematic representation of the case marking patterns of the CS
and CC in primary copular sentences is shown in Table 4 below.
Copular Sentences in Asamiya 209
the discourse or conversation,” but also extends to “the immediate
physical and social environment as well as general knowledge.”
The following description exemplifies the functions and uses of the
different types of copular sentences. The various contextual factors
influencing the interpretation of each and the selection of a type as against
another are also dealt with in the following.
4.1. Primary copular sentences
The basic function of a zero copular sentence is to express a simple
present/past habitual interpretation exemplified earlier in (12), (17),
(19)–(21). The sentence as in (19) is repeated here as (23) for descriptive
convenience.
(23) [
NP
deuta]
CS
[
NP
ei onustʰan-ɔr ħɔb̤apoti]
CC
father this institute-GEN president
‘(My) father is the president of this institute.’
In contrast to the inherent present reading of a zero copular sentence,
one of the principal functions of the past paradigm of the primary overt
copula as- is to mark the temporal interpretation of past as in (24).
(24) [
NP
deuta]
CS
[
NP
ei onustʰan-ɔr ħɔb̤apoti]
CC
father this organization-GEN president
[
V
as-il]
COP
be-PST.3
‘My father was the president of this organization.’
However, a past reference can also be rendered by a zero copular
sentence provided that the given CC is a NP. The two options in that
case are as follow: (a) use of an appropriate prenominal modifier like
prakt
ɔn ‘former’ as in (25) or (b) use of an adverbial modifier like tetia
‘then’ or agɔtɛ ‘earlier’ for denoting a past reading in a narrative as
in (26).
(25) [
NP
deuta]
CS
[
NP
ei onustʰan-ɔr praktɔn
father this institute-GEN former
ħɔb̤apoti]
CC
president
‘My father was the former president of this institute.’
210 North East Indian Linguistics Volume 3
(26) [
NP
silɔŋ]
CS
[
ADVP
tetia]
A
[
NP
ɔħɔm-ɔr razd̤ani]
CC
Shillong then Assam-GEN capital
‘Shillong was the capital of Assam at that time.’
The use of an inflected overt copula as-, besides rendering a temporal
distinction, may add an extra semantic overtone, not present in its non-
overt counterpart. Note the semantic difference between the zero copular
sentence, as in (27), and its overt counterpart, as in (28).
(27) [
NP
upohar-tʊ]
CS
[
PP
tʊma-r karɔnɛ]
CC
gift-CL 2
2
SG-OBL for
‘The gift is for you.’
(28) [
NP
upohar-tʊ]
CS
[
PP
tʊma-r karɔnɛ]
CC
gift-CL 2
2
SG-OBL for
[
V
as-il]
COP
be-PST.3
‘The gift was meant for you, (not to be given away to anybody
else).’
A zero copular sentence with an ADVP or AP as the CC as in (29)
may convey a habitual state of the referent or may have a present perfect
interpretation of the same, depending on context.
(29) [
NP
kʰiriki-kʰɔn]
CS
[
AP
kʰʊla]
CC
window-CL open
(i) ‘The window is open.’
(ii) ‘The window has been open.’
(29) can have the interpretation (i), when it is used as a positive response
to a closed question: ‘Is the window open?’ The interpretation (ii)
expresses an anticipated state of the referent. In this particular instance
it may be used in a context when the speaker has seen the window being
left open earlier and anticipates it to remain open till now (i.e., at the
time of speaking) on the basis of his/her general knowledge/previous
experience.
Unlike a zero copular sentence, an overt copular counterpart with
as- can be used to imply an assertion of the physical condition of the
referent with present or past reference, as warranted by context. For
Copular Sentences in Asamiya 211
example, the sentence (30) may be used while describing the present
state of a house.
(30) [
NP
kʰiriki-kʰɔn]
CS
[
AP
kʰʊla]
CC
[
V
as-e]
COP
window-CL open be-3
‘The window is kept/left open.’
However, in a specific context, (30) may also be used as a reminder to
someone to close the open window.
Cross-linguistic typological studies show that there is a similarity
among locational, existential and possessional sentences (Clark 1978;
Schachter 1985). In Asamiya the similarity lies in the conspicuous
presence of the homophonous verb as- in all these three constructions;
but it is important to note that all uses of as- are not copular in terms of
the elementary assertions as shown in §2 (a–g) (For argumentation on
logical distinctions among the three constructions see Lyons 1971:
388–395).
Sentences elicited in response to a query regarding some spatial
location of a given referent for past or present interpretation obligatorily
induce the use of the overt copula as- as encountered earlier in (22),
repeated here as (31).
(31) [
NP
kitap-kʰɔn]
CS
[
PP
mɛz-ɔr ʊpɔr-ɔt]
CC
[
V
as-e]
COP
book- CL table- OBL top-LOC be-3
‘The book is on the table.’
But the sentence becomes ill-formed without an overt copula as
in (32).
(32) *[
NP
kitap-kʰɔn]
CS
[
PP
mɛz-ɔr ʊpɔr-ɔt]
CC
book- CL table-OBL top-LOC
Compare (31) with an analogous counterpart as in (33) characterized
with the lexical verb as- ‘exist’.
(33) [
PP
mɛz-ɔr ʊpɔr-ɔt]
A
[
NP
ɛ-kʰɔn kitap]
S
table-OBL top-LOC one-CL book
[
V
as-e]
INT
exist-3
‘There is a book on the table.’
212 North East Indian Linguistics Volume 3
As evidenced, the CS in the copular sentence as in (31) is in the canonical
sentence-initial position, followed by the locative adverbial CC mɛz-ɔr
ʊpɔr-ɔt ‘on the table’, whereas it is the locative phrase which occurs in
the sentence-initial position in the existential sentence as in (33),
indicating a difference with respect to ‘Perspectival center’ (Partee and
Borschev 2007)
12
. Pragmatically, what is given or presupposed
information in a locative copular sentence, is new or asserted information
in an existential sentence and vice versa. That existential sentences are
constructions of a kind devised to play a different role from that of the
copular constructions in the language is reinforced by the following
formulaic sample of a prototypical sentence used in a folktale/fairytale
for introducing a protagonist as new information, realized invariantly by
an indefinite NP.
(34) [
ADVP
ɛ-ħɔmɔj-ɔt]
A
[
ADVP
ɛ-kʰɔn deħ-ɔt]
A
one-time-LOC one-CL country-LOC
[
NP
ɛ-zɔn rɔza]
S
[
V
as-il]
INT
one-CL king exist-PST.3
‘Once upon a time there was a king in a country.’
The following exemplifies the juxtaposition of a CS with an adjectival
enumerator du-ta ‘two-CL’, composed of a cardinal numeral suffixed with
a classifier, functioning as the CC.
(35) [
NP
tɛʊ-r lora sʊwali]
CS
[
AP
du-ta]
CC
3
2
.SG-GEN boy girl two-CL
lit. !‘His children are two’, i.e., ‘The number of his children is
two.’
The zero copular sentence as in (35) cannot have an overt counterpart
with as- without adding a contrastive or specific interpretation. For
12
Following the lines of assumptions in Partee and Borschev (2007: 9), the
distinctions between the two can be accounted for in terms of the relative dif
-
ference assumed to be in the “speaker’s perspective structure,” which presup
-
poses a choice for verbs that take both a
LOCation and a THING as arguments,
viz.,
V(LOC,THING) (ibid.). When the THING is chosen as the Perspectival center,
“its existence is presupposed, and the sentence speaks of its
LOCation and
potentially about other properties or states or actions in the situation;” when
the
LOCation is chosen, it “speaks about what THINGs there are or not in that
situation and potentially about what is happening in the situation.”
Copular Sentences in Asamiya 213
example, the use of the present indefinite form here will infer a proposition
that he indeed has two children, but they are worthless or of not worth
mentioning.
However, the zero copular sentence as in (35) can have an analogous
non-copular counterpart with a lexical verb as- ‘possess’ as in (36).
(36) [
NP
tɛʊ-r]
S
[
NP
du-ta lora sʊwali]
DO
[
V
as-e]
TRN
3
2
.SG-GEN two-CL boy girl possess-3
‘He has two children.’
Like any prototypical adjective, the enumerator du
-ta ‘two-CL’ which
functions as the CC in the copular sentence (35), functions as the modifier
to the nominal head lora sʊwali ‘children’ in the non-copular counterpart
(36) (Dixon 2004). The syntactic difference between the two is supported
by the intonational pauses separating the respective constituents. In terms
of context of discourse, the former may be used as a response to the
question: ‘Has he any family/children?’, whereas, the latter is elicited as
a response to the query: ‘How many children has he?’
An adverbial phrase in the CC position juxtaposed with a CS may be
used to connote a locational interpretation in the present as in the
following.
(37) [
NP
tɛʊ-r g̤̤ɔr]
CS
[
ADVP
zʊrhat-ɔt]
CC
3
2
SG-GEN home Jorhat-LOC
lit. ‘His home is in Jorhat.’ i.e., ‘He hails from Jorhat.’
(38) [
NP
ħɔb̤a-kʰɔn]
CS
[
ADVP
sari-bɔza-t]
CC
meeting-CL four-sound-LOC
‘The meeting is at four o’clock.’
While the past form of copula
as- is obligatory for past reference as
in (24), the present indefinite form, as warranted by context, may be used
mostly to yield contrastive interpretations. Compare (39) and (40) with
(37) and (38) respectively.
(39) [
NP
tɛʊ-r g̤̤ɔr]
CS
[
ADVP
zʊrhat-ɔt]
CC
[
V
as-e]
COP
3
2
SG.-GEN home Jorhat-LOC be-3
lit. ‘His home is in Jorhat (not in whatever place the hearer
presumes it to be).’
214 North East Indian Linguistics Volume 3
(40) [
NP
ħɔb̤a-kʰɔn]
CS
[
ADVP
sari-bɔza-t]
CC
[
V
as-e]
COP
meeting-CL four-sound-LOC be-3
‘The meeting is scheduled for four o’clock (not now).’
The presence of an overt copula as- in colligation with a deverbal
ADVP as the CC used for encoding an assertion of physical state is
mandatory in primary copular sentences as shown below.
(41) [
NP
gɔs puli-bʊr]
CS
[
ADVP
mor-i]
CC
[
V
as-il]
COP
sapling- CL die-ADVL be-PST.3
lit. ‘The saplings were in a decaying state.’
i.e., ‘The saplings have withered.’
(42) [
NP
kapʊr-bʊr]
CS
[
ADVP
ħuka-i]
CC
[
V
as-e]
COP
13
clothes-CL dry-ADVL be-3
lit. ‘The clothes are in a dry condition.’
i.e., ‘The clothes have dried up’
The copula
hɔ-, as noted earlier, is primarily used for non-present
interpretations, i.e. future as in (43) and simple past as in (44) by the use
of future and past tense respectively (see Table 2a).
(43)
[
NP
mɔɛ]
CS
[
NP
ħikkʰɔk]
CC
[
V
ho-m]
COP
1.SG teacher be-FUT.1
‘I will be a teacher.’
(44) [
NP
ħɔb̤a-kʰɔn]
CS
[
ADVP
sari-bɔza-t]
CC
[
V
ho-l]
COP
meeting-CL four-sound- LOC be-PST.3
‘The meeting was (held) at four o’clock.’
The three other forms of
hɔ- are basically used to encode aspectual
interpretations. The present indefinite form may be used to render a
13
Note the use of the identical nonfinite causative verb ħuka-i- ‘make dry’ in
colligation with the progressive auxiliary as-.
[
NP
ħihɔt-e]
S
[
NP
kapʊr-bʊr]
DO
[
ADVP
zui-t]
A
[
V
ħuka-i as-e]
TRN
3.PL-NOM clothes-CL fire-LOC make.dry-NF AUX-3
‘They are drying the clothes in front of the fire.’
Here the causative verb in the nonfinite form
ħuka-i- ‘make dry’, often found
in the speech of many, is the contracted version of the full form
ħukuwa-i.
Copular Sentences in Asamiya 215
habitual interpretation for which the presence of an overt adverbial
modifier like praje ‘often’ as in (45) is mandatory
14
.
(45) [
NP
lora-tʊ-r]
CS
[
ADVP
praje]
A
[
NP
oħukʰ]
CC
boy-CL-GEN often illness
[
V
hɔ-e]
COP
be-3
‘The boy often falls sick.’
The following exemplifies the use of the two other forms of hɔ- for
expressing other aspectual distinctions like present perfect as in (46) and
past habitual as in (47).
(46) [
NP
ta-r]
CS
[
NP
oħukʰ]
CC
[
V
ho-is-e]
COP
3
3
.SG-GEN illness be-IPFV-3
‘He has been sick.’
(47) [
NP
ħɔb̤a-kʰɔn]
CS
[
ADVP
sari-bɔza-t]
CC
[
V
ho-is-il]
COP
meeting-CL four-sound-LOC be-IPFV-PST.3
‘The meeting used to be held at four o’clock.’
As evidenced from the data above, the zero copular sentence is
basically used to encode present reference-simple present, present habitual
or present perfect, depending on context. However, the primary overt
copular sentence, mostly used for past or future reference, may also be
used to add additional semantic overtones or imply aspectual interpretations
as warranted by pragmatic factors. Consequently, they are in
complementary distribution with one another in that one cannot occur in
a context in which the other is found; nor can it be used as a paraphrase
of the other.
4.1. Quasi copular sentences
The quasi copular sentence is the most marked variety of copular
sentences in the language in that it is a structurally less elementary and
14
Counterexamples to this norm have been encountered in two situations –
(a) when it is used to imply reassertion/reaffirmation to what is predicated of
the referent, and (b) when it is taken as a semantic equivalent to the English
overt copula ‘be’ by a younger generation of speakers, exhibiting a case of
structural transference from one language to another.
216 North East Indian Linguistics Volume 3
less frequently used construction. Semantically, it is more or less akin
to cleft constructions in English, where some element is highlighted or
brought into focus
15
. As noted earlier (§2 and §3), the sentence of this
type is characterized by a fossilized verbal copula hol or hoise, neutral
with respect to the morphosyntactic properties of tense, aspect and person.
Further, unlike the canonical falling pitch on a verb, this variety of copula
is marked by a rising pitch. It is ‘unverblike’ not only because it is inert
to any temporal or person distinctions, but also is characterized by
constraints regarding negation, interrogation, subordination, coordination,
relativization or other similar constructions (Dasgupta 1983; Dasgupta
2006)
16
. However, these issues have not been addressed here as it would
take us far beyond the scope of the paper.
A quasi copular sentence is primarily used to express a relationship
between an identified and an identifier, whereby the quasi copula hol or
hoise, characterized by a rising pitch, is inserted between the two to
highlight/focus the identified, i.e. the CC, with the CS being the identifier
as in (48) and (49)
17
.
(48)
[
NP
ɔħɔm-ɔr prɔd̤an ħɔmɔsja]
CS
Assam-GEN primary problem
[
V
hoise/hol]
COP
Q
[
NP
banpani]
CC
BE floods
‘The primary problem of Assam is floods.’
15
I am grateful to Scott DeLancey for his suggestions and comments on issues
relating to this topic.
16
Dasgupta (1983) discusses at length these “unverblike” properties of this type
of verb with reference to Bangla. Dasgupta (2006) draws the attention of
scholars of historical and comparative Indo-Aryan studies towards the impor
-
tance of a diachronic approach to this type of construction, not attested in MIA
languages, but found particularly in the contemporary eastern languages of
India, viz., Bangla, Asamiya and Oriya.
17
A preliminary survey of the syntactic structural patterns of the early modern
Asamiya carried out by the author by a comparative study of the compiled
volumes of the two pioneering monthly magazines – Orunodoi 1846–54 (Neog
2003) and Jonaki 1889–98 (Saikia 2001) – shows that while the earlier work
has been conspicuous by the absence of this construction, the later one has
been replete with instances of such constructions. It may be assumed that this
kind of sentence evolved only after 1854, although no recorded evidence was
perceptible until the inception of Jonaki, which marked the beginning of the
modern period of Asamiya literature
Copular Sentences in Asamiya 217
(49) [
NP
pʰɔlmul-or bib̤innɔ upadan-ħomuh]
CS
fruit-GEN various constituent-CL
[
V
hoise/hol]
COP
Q
[
NP
ħɔrkɔra kɛlsiam
BE
sugar calcium
kʰadjɔpran kʰɔniz lɔbɔn itjadi]
CC
vitamin mineral salt etc.
‘The various constituents of a fruit are sugar, calcium, vitamin,
mineral salt, etc.’
As observed earlier, the inversion of the nominal constituents of a
sentence as in (48) (cf. (16)) does not alter the basic proposition of the
sentence as the focused element remains the same. One piece of evidence
in support of this claim is that both the sentences can be elicited as
response to a question: ‘What is the primary problem of Assam
?’, but
not ‘What are floods?’ Additional support for the claim comes from the
fact that the focused constituent can also be marked by the emphasizer
-(e)i for intensity in its inversed version as in (50), and reinforced by
the unacceptability of (51) as its paraphrase.
(50) [
NP
banpani-ei]
CS
[
V
hoise/hol]
COP
Q
[
NP
ɔħɔm-ɔr
floods-EMP BE Assam-GEN
prɔd̤an ħɔmɔsja]
CC
primary problem
‘It is floods which are the primary problem of Assam.’
(51) [
NP
ɔħɔm-ɔr prɔd̤an ħɔmɔsja-i]
CS
[
V
hoise/hol]
COP
Q
Assam-GEN primary problem-EMP BE
[
NP
banpani]
CC
floods
‘Out of many problems of Assam, the floods are an acute one.’
However, sentences of this type need to be contrasted with the marked
stylistic variant of the primary overt copular sentence with the copula
occurring in between the CS and the CC as in (52), where the reordering
of the constituents from the canonical SOV to SVO is used as a device
to render a literary and/or formal style to an otherwise colloquial form
as in (53).
218 North East Indian Linguistics Volume 3
(52) [
ADVP
tetia]
A
[
NP
akaħ]
CS
[
V
as-il]
COP
then sky be-PST.3
[
NP
ħompurnɔ nila]
CC
complete blue
‘Entirely blue the sky was during that time.’
(53)
ADVP
tetia]
A
[
NP
akaħ]
CS
then sky
[
NP
ħompurnɔ nila]
CC
[
V
as-il]
COP
complete blue be-PST.3
‘The sky was entirely blue during that time.’
A quasi copular sentence may also have a secondary presentative
function in colloquial speech, when the speaker uses it while introducing
someone/something to the addressee. The verbal form preferably used
in this function is hoise rather than hol as in (54) and (55). In such a
sentence the CS is a 3
rd
person gender-neutral proximal pronoun like ɛʊ
or ekʰet ‘s/he’ or an indefinite proximal pronoun like eija ‘this’ or other
variants of ei- ‘it,’ suffixed with an appropriate classifier like -tʊ ,-kʰɔn,
-bʊr, etc.
(54) [
NP
eija]
CS
[
V
hoise]
COP
Q
[
NP
mʊ-r ħɔhɔkɔmi dɔktɔr
this BE 1.SG-GEN colleague doctor
ħɔikia]
CC
Saikia
‘This is my colleague Dr Saikia.’
(55) [
NP
ei-tʊ]
CS
[
V
hoise]
COP
Q
this BE
[
NP
ama-r poitrik b̤eti]
CC
1.PL-GEN paternal foundation
‘This is our ancestral place.’
To recapitulate, the quasi copular sentence is inferred to be the most
pragmatically marked construction (see footnote 4) which employs
various devices like the use of a particle (i.e., the invariant copula), a
different intonation, and word order to highlight some focused element
in its primary function. Additionally, a sentence of this kind may also
be used to present or introduce someone/something to the addressee in
colloquial speech in its secondary function.
Copular Sentences in Asamiya 219
Conclusion
The study reveals that there is a non-correspondence among the three
subtypes of copular sentences in terms of their morphosyntactic
characteristics and functional scope. The zero copular sentence, devoid
of an overt copula in comparison to the overt copular sentences, is found
to be structurally less complex. However, the two overt copular sentences
are set apart from one another with respect to the form of the verbal
copula, constituent order and the use of phrasal category as the CC. The
quasi copular sentence is identified to be the most marked variety on the
basis of its structural uniqueness and delimited manifestation or use.
It has been observed that there is a significant correlation between
the morphosyntactic structure of a subtype of copular sentence and its
semantic function. A zero copular sentence inherently encodes a present
interpretation – simple present, present habitual or present perfect, as
warranted by context. By contrast, the primary overt copular sentence
with an inflected verbal copula is used basically to encode a non-present
temporal interpretation, i.e., past or future, while the quasi copular
sentence with an invariant verbal copula, inert in terms of temporal
interpretations and neutral with respect to cross-referencing, is simply
used to highlight or give focal prominence to some element. It is thus
inferred that the choice of a type as opposed to the other is governed by
pragmatic/semantic considerations. Consequently, the three subtypes of
copular sentences are in complementary distribution with one another in
that they cannot be used as optional variants. This clearly accounts for
why deletion or insertion of a copula in a subtype without any structural
alteration either yields an ill-formed sentence or changes the semantic
overtone of a sentence. However, in this context the role of the
grammatical feature of compatibility of lexical categories with copulas
also needs to be taken into consideration.
One of the striking features that has been unraveled from the study
is that apart from the apparent syntactic and semantic distinctions, the
copular sentences are distinguished from the analogous counterparts
characterized by homophonous verbs in the language in terms of the
coreferencing ability with the subject which is lacking in the latter.
220 North East Indian Linguistics Volume 3
Abbreviations
1 1
st
person
2
1
2
nd
person honorific
2
2
2
nd
person familiar
2
3
2
nd
person inferior
3 3
rd
person
3
1
3
rd
person honorific
3
2
3
rd
person familiar
A Adjunct
ADVL Adverbializer
ADV Adverb phrase
AP Adjective phrase
AUX Auxiliary
CC Copula complement
CL Classifier
COP Copula verb
COP
Q
Quasi copula verb
CS Copula subject
DO Direct object
EMP Emphasizer
F Feminine
FUT Future
GEN Genitive case
IND Indefinite
INT Intransitive verb
IPFV Imperfective
LOC Locative case
M Masculine
NEG Negative morpheme
NF Nonfinite
NOM Nominative case
NP Noun phrase
O Object nominal
OBL Oblique case
PART Particle
PL Plural
POSTP Postpositional phrase
PRS Present
Copular Sentences in Asamiya 221
PST Past
Q Question marker
S Non-copula subject
SG Singular
TAG Question tag
TRN Transitive verb
V Verb
XP Any phrasal variable
[ ] Marker of constituent boundary
* Marker of ill-formed sentence
! Marker of unacceptable sentence
References
Akmajian, A., R. Demers, A. Farmer and R. Harnish (2004). Linguistics: An Intro-
duction to Language and Communication. 5
th
Edn. New Delhi, Prentice-Hall
of India Private Ltd.
Clark, E. V. (1978). “Locationals: Existential, locative and possessive constructions
in universals of human language.” In J. Greenberg, Ed.,
Syntax. Stanford, CA,
Stanford University Press: 85–126.
Curnow, T. J. (2000). “Towards a cross-linguistic typology of copular constructions.”
Proceedings of the 1999 Conference of the Australian Linguistic Society.
http://www.linguistics.uwa.edu.au/research/als99/proceedings. Accessed
4/1/2010
Dasgupta, P. (1983). “Bangla equatives, complementizers, final focii and roots.”
Linguistic Analysis 11(2): 103–37.
_____. (2006). “The positive polarity copula in Bangla: An item of unexplored
origin.” Indian Linguistics 67(1–4): 61–64.
Declerck, R. (1988). Studies on Copular Sentences
, Clefts and Pseudoclefts. Dor-
drecht, Foris.
Delancey, S. (1997). “Grammaticalization and the gradience of categories: Relator
nouns and postpositions in Tibetan and Burmese”. In J. Bybee, J. Haiman and
S. A. Thompson, Eds., Essays on Language Function and Language Type
Dedicated to T. Giv
ón. Amsterdam/Philadelphia, John Benjamins: 51–69.
Dixon, R. M. W. (2004). “Adjective classes in typological perspective.” In R. M.
W. Dixon & A. Y. Aikhenvald, Eds., Adjective Classes: A Cross-linguistic
Typology. Cambridge, Cambridge University Press: 1–45.
Huddleston, R. and G. K. Pullum. (2002). The Cambridge Grammar of the English
Language. Cambridge, Cambridge University Press.
Lyons, J. (1971). Introduction to Theoretical Linguistics. Cambridge, Cambridge
University Press.
222 North East Indian Linguistics Volume 3
Neog, M. (compiled and edited) (2003). Orunodoi. 2
nd
Edn. Guwahati, Asom Pra-
kasan Parisad.
Partee, B. and V. Borchev. (2007). “Existential sentences,
BE and the genitive nega-
tion in Russian.” In I. Comorovski and K. von Heusinger, Eds., Existence:
Semantics and Syntax, Dordrecht, Kluwer/Springer: 147–190.
Payne T.E. (1997). Describing Morphosyntax. Cambridge, Cambridge University
Press.
Pustet, R. (2003). Copulas: Universalization in the Categorization of the Lexicon.
New Delhi, Oxford University Press.
Saikia, N. (Collected, compiled and edited) (2001). Jonaki. Guwahati, Asom Sahi
-
tya Sabha, G. L. Media Services Ltd. Press.
Schachter, P. (1985). “Parts-of-speech systems.” In T. Shopen, Ed., Language Typol
-
ogy and Syntactic Description
. Volume I: Clause Structure. Cambridge, Cam-
bridge University Press: 3–61.
Stassen, L. (1997). Intransitive Predication. Oxford, Clarendon.
Case Marking in Hajong 223
13
Case Marking in Hajong
1
Virginia Crowell Phillips
Summer Institute of Linguistics
1. Introduction
The Hajong group is one of the scheduled tribes of North East India.
They live mainly in Assam and Meghalaya in India and in the Mymensingh
District of Bangladesh. Hajong is classified in the Ethnologue as Indo-
European > Indo-Iranian > Indo-Aryan > Eastern zone > Bengali-
Assamese (Gordon 2005). Although Hajong is classified as Indo-Aryan
(IA), several Hajong case markers are not cognate with those of the large
neighbouring IA languages, Bangla and Asamiya. This raises the question
of the origin of these case markers. Some authors have speculated that
the Hajong language has a Tibeto-Burman (TB) sub-stratum. The Hajong
people are ethnically and culturally closer to the surrounding Tibeto-
Burman people groups such as Garo and Koch than to the Bengali
population
2
. The cultural and linguistic similarities could be due either
to a common origin or to a mutual influence between the groups who
have been living in proximity for generations.
1
I would like to thank the many speakers of the Hajong language who have
given the data used for this paper, and who have helped to transcribe and
translate that data. I especially thank Abhijit Barman and Mamata Hajong for
their hours of help. I also referenced Guts (2007) and Kinny & Zeliang (2005)
for information on Hajong.
2
For example, the Hajong traditional women’s dress (pathin) is identical in
pattern and the way it is worn to the Koch traditional dress.
224 North East Indian Linguistics Volume 3
In this paper, I will describe the system of case marking in Hajong
– that is, the morphemes which specify the syntactic function of the noun
phrase. I will describe the syntactic functions associated with each case
form and I will compare the Hajong case forms to their equivalents in
four geographically proximate languages. From the IA family, Hajong
will be compared to Standard Colloquial Bangla (Dasgupta 2003) and
Standard Colloquial Asamiya (Goswami and Tamuli 2003) (both
classified as Indo-European > Indo-Iranian > Indo-Aryan > Eastern zone
> Bengali-Assamese). From the TB family, I will compare Hajong to
Garo (Burling 2003) and the Wanang dialect of Koch (Kondakov and
Kondakova 2007) (both classified as Sino-Tibetan > Tibeto-Burman >
Jingpho-Konyak-Bodo > Konyak-Bodo-Garo, > Bodo-Garo). Garo is the
language of wider communication in the Garo Hills of Meghalaya where
a majority of the Hajong population is currently located. Koch is
culturally similar to Hajong and also located in the Garo Hills of
Meghalaya. Where the case forms seem unique to Hajong, I will explore
their possible origins. In doing this, I not only provide data on a hitherto
practically undocumented variety of Indo-Aryan, but explore the
interaction between the IA and TB language families.
2. The Hajong Case Marking System
In Hajong, the formatives
3
which specify the syntactic function of a noun
phrase occur as postpositions either immediately following the head noun
(or its classifier) or with a case marker intervening after the head noun.
The dative, genitive, locative, allative, ablative and instrumental markers
immediately follow the head noun or its classifier as shown in Table 1.
Based on both diachronic and synchronic properties, Masica (1991:
231ff.) identifies three layers of formatives with case-like functions in
New Indo-Aryan (NIA) languages. Layer I consists of inflectional affixes
inherited from OIA/MIA; these affixes are characterized by declensional
differences and singular/plural differences; they attach directly to the
base with morpho-phonemic adjustments. This layer is essentially missing
in Asamiya and other Eastern IA languages including Hajong. Layer II
3
Formatives are the markers of inflectional information. They are different from
words since they cannot govern or be governed by other words, cannot require
or undergo agreement, and cannot head phrases (Bickel and Nichols 2007:
172–3).
Case Marking in Hajong 225
may be attached to the base indirectly, may be mediated by a Layer I
element and/or is invariant for all nouns and the same for both numbers.
Layer II elements may be either agglutinative suffixes or analytic particles
and sometimes even clitics. Layer III is mediated by a Layer II element;
it lacks morphophonemic variants and often has a transparent connection
with an independent word; it is semantically more specific than a Layer
II element. All of the Hajong case formatives listed in Table 1 above fall
into Masica’s category of Layer II affixes since they are invariant for all
nouns and the same for both numbers. These case formatives are not
phonologically dependent on the noun. They can occur directly after a
noun or after its classifier. In this paper, I will usually separate the case
formatives from the preceding noun words. The exception is when the
case formative follows a pronoun in a form which cannot stand alone.
Here, the case formative is hyphenated.
Table 1 Case markers which immediately follow the noun
Hajong Gloss Translation Case
buri-rɯ
4
old.woman-DEF ‘the old woman’ unmarked
buri-r
ɯ
5
ge old.woman-DEF DAT ‘to the old woman’ dative
buri la old.woman
GEN ‘of the old woman’ genitive
buri ni old.woman
LOC ‘to/at the old woman’ locative
buri b
ʰaʲ old.woman LOC ‘to the old woman’ allative
buri t̪ʰiki old.woman
ABL ‘from the old woman’ ablative
buri diɯ old.woman INST ‘through/by the help instrumental
of the old woman
’
4
The Hajong definite marker (or generic classifier) -rɯ has allomorphs -ra,
-da, -la, and -ʒa. It should not be confused with the nominative marker of
Bangla,
-ra, which is etymologically related to the genitive (Toulmin 2006:
155). The Hajong
-rɯ is equivalent to the classifier -ʈa of Bangla, Asamiya
and related languages. A similar shift of the [
ʈ] to [r] for this classifier is at-
tested in the Rohinga dialect of Chittagonian (Indo-European > Indo-Iranian
> Indo-Aryan > Eastern zone > Bengali-Assamese) (Lloyd-Williams, per
-
sonal communication).
5
The definite marker is obligatory before the DAT case marker but optional
before the other case markers.
226 North East Indian Linguistics Volume 3
The markers for specific location such as ‘on top of’ or ‘underneath’
can occur with or without the genitive intervening. The genitive marker
comes with a change of meaning, as shown in examples (1) and (2).
(1) tibil up
ʰur ni
table top LOC
‘on (the surface of) the table’
(2) tibil la upʰur ni
table
GEN top LOC
‘over the table’
These formatives are semantically more specific than the general
locative case markers
ni and bʰaʲ and are always followed by either ni
or bʰaʲ.
The scope of this paper is limited to the Layer II case markers, i.e.
formatives which come immediately after a head noun or its classifier.
This includes the six formatives listed in Table 1, i.e. dative, genitive,
locative, allative, ablative, and instrumental. Table 3 lists the forms for
Hajong and compares them to the equivalent forms in Standard Colloquial
Asamiya, Standard Colloquial Bangla, Wanang Koch and Garo
6
.
In the following sections, I will look first at the syntactic function of
each of these markers and then their form and etymology.
3. Accusative/Dative
3.1. Function
The accusative/dative case marker,
ge or gon, marks the object of a
transitive clause and the recipient or goal of a di-transitive verb. It is
Table 2 Case markers which follow the genitive
Hajong Gloss Translation Case
buri la bede old.woman GEN BEN ‘for the old woman’ benefactive
buri la t͡ʃ
ɯjɯ old.woman GEN COM ‘than the old woman’ comparative
Hajong also has Layer III elements which specify the function of a
noun phrase. The benefactive and comparative markers are mediated by
the Layer II genitive marker, as shown in Table 2.
6
Each of these languages has nominative-accusative alignment.
Case Marking in Hajong 227
common in NIA languages to use one marker for both of these functions
(Masica 1991: 365). The two forms, ge and gon, are synonymous and
in free variation although a given speaker will choose one or the other.
Examples (3) and (4) illustrate that, subject to animacy conditions
elaborated on below, both ge and gon can be used for the accusative
case, i.e. the object of a transitive clause.
(3) ud
ɯ bandor-ra gon ni-ɯ ahi-ba lagi-se.
that monkey-DEF ACC take-ing come-INF start-PERF
‘He is bringing a monkey and coming.’
(4) u-kuinɯ-rɯ ge ni-ɯ…
that-bride-DEF ACC take-ing
‘Taking that bride…’
Table 3 Cross-linguistic case marker comparisons
Hajong Standard Standard Wanang Garo
Colloquial Colloquial Koch
Asamiya Bangla
Nominative
7
∅ ∅, -e ∅, -ra (PL, ∅, -a
animate)
Accusative ge
, gon -[ɒ]k -ke, -[e]re
-ko
(animate) -go-re
8
Dative ge, gon -[ɒ]k (animate) -ke, -[e]re -na -na
Genitive la -[
ɒ]r -[e]r -go
9
-ni -ni
Locative 1 -[
ɒ]t -[ɒ]t -e, -te -o
(inanimate)
Locative 2 ni
Allative
bʰaʲ -[ɒ]lɔi -waʲ -chi, -ona
Ablative
t̪ʰiki,
t̪ʰokon,
t̪ʰaki -GEN pɒra -t̪ʰeke -oni
Instrumental di
ɯ, de -ere, -re -dia -chi
7
Zero-marking for nominative is common in languages of North East India.
8
Found in the eastern dialect (Dasgupta 2003: 365)
9
Dasgupta (2003).
228 North East Indian Linguistics Volume 3
Example (5) illustrates that ge is used for marking the dative case,
i.e. the recipient of a di-transitive verb.
(5) adʒi o-ge t͡ʃɯl di-ba na-lag-e.
today 3SG-DAT rice give-INF not-need-IPFV
‘No need to give him rice today.’
In Asamiya the accusative marker -
k is used exclusively for animate
objects (Dasgupta 1993: 92). Rajbanshi and related lects use the dative
formative only if the head noun is either human or both animate and
discourse prominent (Toulmin 2006: 152). Likewise, Hajong uses the
dative/accusative formative for humans or animate nouns which are
discourse prominent. In example (5) above, the direct object t͡ʃɯl ‘rice’
does not take ge/gon since it is inanimate. Since the ge/gon formative
must follow a discourse prominent noun, it must follow a noun marked
by either the definite -rɯ or the plural -glɯ. Masica (1991: 365) notes
that when an accusative suffix is limited to animates, its function is less
syntactic than pragmatic. The marker is stressing the patienthood of
human nouns which is a marked status (humans are normally agents).
LaPolla (1992) calls this type of marking “anti-ergative.” He points out
that ergative marking marks an argument that is an agent, but “anti-
ergative” marking marks an argument that is not an agent. Usually this
is an animate object that might otherwise be interpreted as an agent.
Bossong calls this selective marking of objects Differential Object
Marking (DOM) (1991).
This accusative/dative formative can also be used on the non-finite
verbs of complement clauses although it is not obligatory. When it does
occur, it is usually because the marked verb is removed from its normal
position directly before the main verb. Again, ge and gon are
interchangeable in this function as shown in examples (6) and (7).
(6) poka gusti-rɯ ge kamra-ba ge na-de.
insect guest-
DEF DAT bite-INF DAT not-give
‘[He] didn’t allow the mosquitoes to bite the guest.’
(7) oi o-la guru-rɯ
3SG 3SG-GEN ox-DEF
mo-la bagan kʰa-ba gon di-le.
1.SG-GEN garden eat-INF DAT give-IMM.PST
‘He has allowed his cow to eat my garden.’
Case Marking in Hajong 229
In example (6), the non-finite verb phrase poka gustirɯ ge kamraba
‘mosquito bite the guest’ is a complement of the main verb nade ‘not
allow.’ This whole phrase takes the object marker ge which comes after
the non-finite verb kamraba ‘to bite.’ Within the complement clause, the
object of the verb kamra ‘bite’ is gustirɯ ‘guest’ which then also takes
the object marker ge. Since both the non-finite verb and its object are
marked with ge, it looks like they are agreeing with each other. However,
example (8) shows that a non-finite verb and an object of a different verb
phrase can both be marked with ge.
(8) to-ge k
ʰa-ba ge de-ba
2.SG-DAT eat-INF DAT give-INF
kisui nɯi pa-i
anything neg get-IPFV
‘[He] was not able to give you anything to eat.’
In example (8), the pronoun to ‘2.
SG’ is marked with ge although it
is the object of deba ‘to give’ which is not marked with ge. In example
(9), although the object of the verb, mo ‘1.SG’, is marked with ge, the
verb is left unmarked.
(9) mo-ge basa-ba pa-bo
1.
SG-DAT save-INF able-IRR
‘[You] will be able to save me.’
Based on the evidence of examples (8) and (9), ge does not have the
function of an agreement marker.
3.2. Form
There are two variants of the accusative/dative formative in Hajong, ge
and gon. The first, ge, is cognate with the Bangla -ke. The cognate words
for ‘pond’, Hajong pagar and Bangla pukur, are another example where
Hajong has a voiced consonant and the Bangla cognate has a voiceless
consonant. However, there is no cognate for the accusative/dative variant
gon in standard Asamiya, Bangla, or Garo
10
. Table 4 compares the
accusative/dative formatives in Hajong, Asamiya, Bangla and Garo.
10
Data for Wanang Koch are unavailable.
230 North East Indian Linguistics Volume 3
One possible cognate for the Hajong gon is the Sylheti
11
classifier for
animate nouns which is -gu in the singular and -guin in the plural (Lloyd-
Williams, personal communication). The morpheme -in is the most
common plural in Sylheti. Although these classifiers in Sylheti are not
restricted to the accusative or dative positions, they are used exclusively
for animate nouns as the Hajong accusative/dative is used only for
animate nouns. The Hajong gon could also be related to the archaic
Bangla plural morpheme gon.
It is interesting to note that two other words in Hajong have an
optional -on ending. The comitative loge is in free variation with logon
and the ablative
t̪ʰiki is in free variation with t̪ʰokon.
4. Genitive
4.1. Function
The genitive formative la is applied to the same range of uses as the
Asamiya and Bangla genitive morphemes. It is used to denote material
possession, as in example (10), before words showing the specific location
of an object, as in example (11), and in expressions of time, as in example
(12).
(10)
dʒoto dinɯ sib la gʰor ni t̪ʰaki-bo.
as.many day Shiv GEN house LOC stay-IRR
‘As long as I stay in Shiv’s house…’
(11) to-la upʰur bʰaʲ utʰi-ɯ dʒa-ba lagi-bo.
2.SG-GEN up LOC climb-ing go-INF need-IRR
(monkey says to tiger) ‘I will have to climb on top of you.’
(12) pak aha la pore t͡ʃini de-i.
boil come
GEN after sugar give-IPFV
‘After it has come to a boil, add sugar.’
Table 4 Cross-linguistic comparison of the accusative/dative formative
Hajong Asamiya Bangla Garo
Accusative ge, gon -k, -ɒk -ke, -[e]re -ko
Dative ge, gon -k, -ɒk -ke, -[e]re -na
Case Marking in Hajong 231
In this paper, three hypotheses for the etymology of this formative
are presented. One hypothesis is that la is a cognate of the Bangla and
Asamiya genitive -r. Chatterji (1926: 755) traces the etymology of the
Bangla genitive -r from the OIA kera and kara which had a variant kela.
Although Chatterji does not comment on the etymology of Hajong la, it
is possible that it traces its descent from this kela.
In the Linguistic Survey of India, Grierson (1903–28) gives some
Hajong (Haijong) data from Mymensingh and Sylhet Districts in which
the genitive is listed variously as la, lak and lag. The presence of the
final velar consonants raises the question of whether the genitive is
derived from the Hajong (also Hindi, Bangla, Asamiya) lag meaning
‘attach.’ Presumably lag ‘attach’ is the source of the Hajong comitative
loge, logon. It is easy to imagine the semantic transition from the verb
‘attach’ to the genitive marking formative since an object which is
‘attached’ to you or ‘with’ you is yours – the genitive. It is not difficult
to explain the subsequent drop of the final velar consonants since these
are often unreleased in Hajong and difficult to hear.
A third hypothesis is related to the theory that the Hajong language
has Sino-Tibetan origins. If it was originally a Tibeto-Burman language
that was relexified by Bangla, some hints of its origin may show through.
Several TB languages have a genitive [lə] or [la] morpheme such as
Gamale Kham
13
(Watters 2003: 689), Manange
14
(Hildebrandt 2004),
4.2. Form
The morpheme la is not transparently cognate with the genitive case
markers in Asamiya, Bangla, Koch or Garo
12
, as shown in Table 5.
Table 5 Comparison of genitive case markers
Hajong Asamiya Bangla Koch Garo
Genitive la -r, -ɒr -r, -er -ni -ni
11
Indo-European > Indo-Iranian > Indo-Aryan > Eastern zone > Bengali-Assa-
mese.
12
All of these languages have dependent case marking for the genitive.
13
Kham, Gamale: Sino-Tibetan > Tibeto-Burman > Himalayish > Mahakiranti
> Kham-Magar-Chepang-Sunwari > Kham (Gordon 2005).
14
Manangba: Sino-Tibetan > Tibeto-Burman > Himalayish > Tibeto-Kanauri >
Tibetic > Tamangic (Gordon 2005).
232 North East Indian Linguistics Volume 3
and Tamang
15
(Mazaudon 2003). Interestingly, Gamale Kham also has
a [ni] ablative (cf. the Hajong locative ni). Kurtöp
16
(Hyslop in prep.)
and Lepcha
17
(Plaisier 2007) also have [n] based ablatives [ning] ~ [ni]
and [nun] ~ [nu], respectively. These languages are spoken in the hills
to the north of the Hajong population. Various authors, such as Biren
Hajong (Hajong 2002), have hypothesized that the Hajong people
originally migrated down from Tibet.
5. Locatives
5.1. Functions
There are four formatives used to mark locational function in Hajong:
-(
ɒ)t, ni, bʰaʲ, and t̪ʰiki. The first is the t based locative, cognate with
Asamiya and Bangla and most likely a borrowing from those languages.
Sometimes it is used in a context where it would be replaced by ni or
bʰaʲ if the speakers wanted to dissociate themselves from speakers of
Bangla or Asamiya. At other times it is idiomatic and cannot be replaced,
as in example (13).
(13) nam-ra mon-ot pahri-le ela
name-
DEF mind-LOC forget-IMM.PST now
‘I have forgotten the name right now.’
Of the remaining three locative markers, ni has the broadest scope as
it can be used as both a general locative and an allative case marker. It
can also be used metaphorically for location in time.
bʰaʲ is limited to
allative case and t̪ʰiki marks ablative case.
5.1.1. ni and bʰaʲ
The locative formatives ni and bʰaʲ are used in free variation to express
allative case, as shown in examples (14) and (15).
15
Sino-Tibetan > Tibeto-Burman > Himalayish > Tibeto-Kanauri > Tibetic >
Tamangic (Gordon 2005).
16
Kurtokha: Sino-Tibetan > Tibeto-Burman > Himalayish > Tibeto-Kanauri >
Tibetic > Tibetan > Eastern (Gordon 2005), but see Hyslop (forthcoming) for
more information regarding the classification of Kurtöp.
17
Lepcha: Sino-Tibetan, Tibeto-Burman, Himalayish, Tibeto-Kanauri, Lepcha
(Gordon 2005)
Case Marking in Hajong 233
(14) gʰor ni ahi-le.
house LOC come-IMM.PST
‘(He) came home.’
(15) gʰor bʰaʲ ahi-le.
house LOC come-IMM.PST
‘(He) came home.’
However, ni shows precise location as in example (16) where it means
‘in’. bʰaʲ cannot be used in this way as shown in example (17).
(16) bugn
ɯ ni pani gorom de-i.
pot LOC water hot give-IPFV
‘Heat water in the pot.’
(17) *bugnɯ bʰaʲ pani gorom de-i.
pot LOC water hot give-IPFV
Only
ni can be used metaphorically for location in time, as illustrated
by the data in (18).
(18) te b
ʰijɯn ni bʰat-tat kʰɯ-jɯ…
then morning LOC rice eat-ing
‘Then in the morning, after eating rice…’
5.1.1.
t̪ʰiki, t̪ʰokon, t̪ʰake
The ablative markers
t̪ʰiki, t̪ʰokon, t̪ʰake are cognate with the Bangla
t̪ʰeke and are used in free variation in the same contexts as the Bangla
ablative. They can be used directly following a place name or pronoun
as in examples (19), (20) and (21).
(19) mo-la d
ʒoŋgol t̪ʰiki kene kʰuri-ra ne-i?
1.SG-GEN forest ABL why? firewood-DEF take-IPFV
‘Why are you taking firewood from my jungle?’
(20) golpara t̪ʰokon e-bʰaʲ ahi-se.
Goalpara ABL this-LOC come-PERF
‘From Goalpara we came here.’
(21) idɯ moi ei-t̪ʰokon bʰaga-i bʰala.
this 1.SG this-ABL flee-IPFV good
‘It’s better for me to flee from here.’
234 North East Indian Linguistics Volume 3
They can also be used after bʰaʲ
18
or ni as in examples (22) and
(23).
(22) pas
bʰaʲ t̪ʰiki o-gon kɯibɯ
back side ABL 3SG-DAT someone
dʰuri-bɯ nekʰan lagi-se.
grasp-INF like attach-PERF
‘He felt like someone was touching him from behind.’
(23) o-ge moi kun ni
t̪ʰokon di-bo?
3SG-DAT 1.SG where LOC from give-IRR
‘From where shall I give to him?’
Finally, they can be used metaphorically for time as in (24).
(24) kunb
ʰola t̪ʰiki taifoid hu-se?
when? ABL typhoid be-PERF
‘From when have you had typhoid?’
5.2. Form
5.2.1. ni
The etymology of the locative morpheme ni is not easily identified.
Grierson (1903–28), writing a hundred years ago, gives mi as the locative
marker for Hajong (Haijong) of Sylhet district. mi is reminiscent of the
Hindi mẽ and Maithili me, both meaning ‘in’. Neither ni or mi has a
cognate in Asamiya or Bangla. Neither do they have transparent cognates
in the surrounding TB languages although -ni shows up as a genitive
affix in Garo and Koch. DeLancey comments that there often seems to
be a relationship between genitive and locative cases in TB languages;
however, tracing the development from one to the other is difficult:
It is not unusual to find homophony between genitive and locative, abla-
tive or ergative case. While there is some evidence for the conceptual
relationship between possession and location, the question of the dia-
chronic development of genitive from locative case (or vice versa?) is
an open one, and we cannot for the present assume a historical direc-
tionality here. (DeLancey 1984: 66)
18
In this instance, bʰaʲ behaves more like a lexeme than a formative and is
therefore glossed differently. Its possible status as a lexeme is discussed in
section 5.2.2.
Case Marking in Hajong 235
5.2.2. bʰaʲ
Likewise, the source of the locative bʰaʲ is not easily identified as it is
not transparently cognate with anything in Asamiya or Bangla. There is
evidence for a locative marker waʲ in Wanang Koch (Kondakov and
Kondakova 2007), although more research on Koch is needed to verify
this. This could point either to Hajong being historically related to the
Koch language, or to borrowing between the languages whose speakers
have been living in close proximity for generations. It would be interesting
to see if there are other cognates in these two languages with the same
[bʰ]-[w] alternation.
In addition to its function as a formative, bʰaʲ also looks like a noun
at times, as in example (25).
(25) to-la lok-gl
ɯ ni-se gor bʰa ʲ
2.SG-GEN friend-PL take-PERF bottom part
tui an-se agal bʰaʲ.
2.SG bring-PERF top part
‘Your friends took the bottom part; you brought the top part.’
Formatives are often grammaticalized from nouns. In both Assamese
and Hajong there is a noun
bʰag ‘part.’ It would be interesting to see if
there is a phonological pattern of a velar in Assamese becoming
palatalized in Hajong.
5.2.3.
t̪ʰiki, t̪ʰokon, t̪ʰake
The ablative markers
t̪ʰiki and t̪ʰake are clearly cognate with the Bangla
t̪ʰeke and therefore not surprising in Hajong. However, the alternative
form t̪ʰokon is unexpected. These forms are in free variation. The source
of -on in t̪ʰokon is unknown. It is possible that the first [o] vowel is
lowered to match the vowel of the second syllable as there is a pattern
of vowel harmony in Hajong. As mentioned in section 3.2 the -on ending
is also seen as an alternative on both the comitative loge and the
accusative/dative ge.
6. Instrumental
The instrumental marker diɯ is cognate with the Asamiya and Bangla
instrumental markers. It is often shortened and pronounced [de] in
Hajong. It is used as in the following examples.
236 North East Indian Linguistics Volume 3
(26) iŋkɯke aŋgul diɯ gutɯ-i…
thus finger INST poke-IPFV
‘Poking with his finger like this…’
(27) baŋla bʰasa diɯ ko-i sɛn pukʰi.
Bangla language INST say-IPFV sen bird
‘In Bangla it is called sen bird.’
Conclusion
The case markers of Hajong are interesting in that there is a mix of some
obvious cognates with Indo-Aryan languages, some possible cognates
with Tibeto-Burman languages and some cases where finding a cognate
is a stretch. Table 6 highlights the possible cognates.
Although the dative/accusative and the ablative marker have partial
cognates in Bangla, the [on] endings remain unexplained. The genitive
la in Hajong does not have an obvious cognate in any surrounding
language although it may be related to the -[e]r of Bangla. The locative
ni may be cognate with the Garo and Koch genitive but the relationship
between them is not easily identified. The locative bʰaʲ seems to have a
cognate in Koch waʲ. The instrumental is clearly cognate with Bangla.
This survey of the case markers of Hajong provides some data and
description of a relatively undocumented variety of IA. It also explores
the interactions between the IA and TB language families in North East
India. This paper raises many possibilities and questions regarding the
origin of Hajong case markers. Although there are some possible cognates
with TB languages, more evidence is needed to assert that Hajong has
Table 6 Possible cognates for Hajong case markers
Hajong IA TB
(Bangla, Asamiya) (Koch, Garo)
DAT/ACC ge, gon -ke
GEN la possibly -[e]r
LOC ni -ni (GEN)
LOC bʰaʲ -waʲ
ABL t̪ʰiki, t̪ʰake, t̪ʰokon -t̪ʰeke
INST diɯ -dia
Case Marking in Hajong 237
a Tibeto-Burman sub-stratum. At the same time, the differences between
Hajong and the surrounding IA languages are too many to ignore the
suggestion that the language has some origin or influence from outside
the IA family.
Abbreviations
1 First person
2 Second person
ABL Ablative
ACC Accusative
BEN Benefactive
COM Comparative
DAT Dative
DEF Definite
GEN Genitive
IMM.PST Immediate past
INF Infinitive
INST Instrumental
IPFV Imperfective
IRR Irrealis
LOC Locative
PERF Perfective
PL Plural
SG Singular
References
Bickel, B. and J. Nichols. (2007). “Inflectional morphology.” In T. Shopen, Ed.,
Language Typology and Syntactic Description. Volume 3: Grammatical
Categories and the Lexicon. 2
nd
Edn. Cambridge, Cambridge University
Press: 169–240.
Bossong, G. (1991). “Differential object marking in Romance and beyond.” In
D. Kibbee and D. Wanner, Eds., New Analyses in Romance Linguistics.
Amsterdam/Philadelphia, John Benjamins: 143–70.
Burling, R. (2003). “Garo.” In G. Thurgood and R. J. LaPolla, Eds., The Sino-
Tibetan Languages. London/New York, Routledge: 387–400.
Chatterji, S. K. (1926). The Origin and Development of the Bengali Language.
Reprinted (1985 [1926]). New Delhi, Rupa.
Dasgupta, B. B. (1993). Assamese Self-Taught. Calcutta, Dasgupta Prakashan.
238 North East Indian Linguistics Volume 3
Dasgupta, P. (2003). “Bangla.” In G. Cardona and D. Jain, Eds., The Indo-Aryan
Languages. London/New York, Routledge: 351–90.
DeLancey, S. (1984). “Etymological notes on Tibeto-Burman case particles.” Lin-
guistics of the Tibeto-Burman Area 8(1): 59–77.
Gordon, R. G., Jr., Ed. (2005). Ethnologue: Languages of the World. 15
th
Edn. Dal-
las, SIL International. Online version: http://www.ethnologue.com/. Acessed
11/12/2008.
Goswami, G. C. and J. Tamuli. (2003). “Asamiya.” In G. Cardona and D. Jain, Eds.,
The Indo-Aryan Languages. London/New York, Routledge: 391–443.
Grierson, G. A. (1903–28). Linguistic Survey of India. Repr. Delhi 1967. Calcutta,
Motilal Banarsidass.
Guts, Y. (2007). Phonological Description of the Hajong Language. MA Thesis.
Amsterdam, Vrije Universiteit.
Hajong, B. (2002). The Hajongs and Their Struggle. Assam, Janata Press.
Hildebrandt, K. A. (2004). “A grammar and dictionary of the Manange language.”
In C. Genetti, Ed., Tibeto-Burman Languages of Nepal: Manange and Sherpa.
Canberra, Pacific Linguistics: 2–189.
Hyslop, G. (forthcoming). Kurtöp and the classification of the languages of Bhutan.
Proceedings from the 45
th
Meeting of the Chicago Linguistic Society. Chicago,
University of Chicago Press.
_____. (in preparation). A grammar of Kurt
öp. PhD Thesis. Department of Linguis-
tics. Eugene, University of Oregon.
Kinny, E. and I. Zeliang. (2005). A Sociolinguistic survey among the Hajong of
India. Unpublished manuscript.
Kondakov, A. and A. Kondakova. (2007). Koch dialects of Meghalaya and Assam:
A Sociolinguistic survey. Unpublished manuscript.
LaPolla, R. J. (1992). “Anti-ergative marking in Tibeto-Burman.” Linguistics of the
Tibeto-Burman Area 15(1): 1–9.
Masica, C. P. (1991). The Indo-Aryan Languages. Cambridge, Cambridge Univer
-
sity Press.
Mazaudon, M. (2003). “Tamang.” In G. Thurgood and R. J. LaPolla, Eds., The
Sino-Tibetan Languages. London/New York, Routledge: 291–314.
Plaisier, H. (2007). A Grammar of Lepcha. [Brill’s Tibetan Studies Library, Lan
-
guages of the Greater Himalayan Region 5.] Leiden, Brill.
Toulmin, M. W. S. (2006). Reconstructing Linguistic History in a Dialect Contin
-
uum: The Kamta, Rajbanshi, and Northern Deshi Bangla Subgroup of Indo-
Aryan. PhD Thesis. Canberra, Australian National University.
Watters, D. E. (2003). “Kham.” In G. Thurgood and R. J. LaPolla, Eds., The Sino-
Tibetan Languages. London/New York, Routledge: 683–704.
Derivational Morphology and Compounding in Pnar 239
Austroasiatic
240 North East Indian Linguistics Volume 3
240 blank
Derivational Morphology and Compounding in Pnar 241
14
Derivational Morphology and Compounding in Pnar
Hemanga Dutta
Jawaharlal Nehru University
1. Introduction
This paper presents a description of the derivational morphology and the
process of compounding in Pnar, an Austro-Asiatic language of the Mon-
Khmer branch spoken in the Jaintia hills of Meghalaya. Derivational
processes in Pnar are prefixing in nature and result with the formation
of new words along with a change in lexical category. The notable
derivational prefixes are noŋ-, i- and ɟiŋ-.
An attempt has been made in §2 of this paper to provide a brief
linguistic description of Pnar. §3 discusses the nature of the derivational
prefixes used in this language to derive new words. The next section of
this paper is devoted to the causatives in Pnar which are morphologically
realized. In §5, I discuss the poly-functional nature of the prefix wa- in
Pnar as it is assigned the role of adjectivalizer, complementizer and
conjunction although this prefix as a derivational morpheme is attached
exclusively with adjectives. The next section analyses reduplication in
Pnar as a derivational process. This language exhibits the instances of
complete word reduplication which can be interpreted under the purview
of derivational morphology. The process of compounding plays a vital
role in the word formation process in Pnar which is discussed in detail
with adequate examples in §7 of this paper.
242 North East Indian Linguistics Volume 3
As far as the methodology of this paper is concerned, it is based on
primary data collected from three consultants
1
. All the consultants are
from Jowai, the region in Jaintia hills where Pnar is mainly spoken. Since
the consultants were highly educated it was quite convenient for me to
use English as medium of interaction through out the period of collecting
the primary data. The fact that they were equally competent in Khasi
proved to be beneficial for my interpretation of the data. The data of
Khasi incorporated in this paper are collected from the same consultants
and on the basis of these limited data an attempt has been made in this
paper to make a comparative analysis between Pnar and Khasi.
2. A brief description of Pnar
Pnar, which literally means “dwellers of the upper hills”, belongs to the
Mon-Khmer sub branch of Austro-Asiatic family of languages. Pnar has
always been considered a dialect of Khasi and it is not a medium of
instruction in school curriculum. As told by our consultants, Pnars
understand Khasi though the Khasi speakers hardly comprehend Pnar.
Pnar has a distinctive linguistic structure of its own.
Every noun in this language is morphologically marked for gender.
Thus, every noun and pronoun in this language must be either preceded
by u (3.M.SG) or by ka (3.F.SG) which denotes gender. Reduplication is
a very productive word formation process in Pnar. It is used in pronouns,
the formation of adverbials, expressives, etc. Additionally, compounding
appears to be an important process of word formation in Pnar. In the
elicitation of the Pnar data we have found the instances of endocentric,
exocentric, copulative and synthetic compounds. Pnar is SVO in structure
where modifiers follow the modified and the quantifiers precede.
1
The collection of linguistic data in an unknown, unexplored language undoubt-
edly involves lots of challenges. However, in this regard I must owe my thanks
to these consultants who have been patient listeners to my innumerable queries
throughout the survey with their warm and active response. The consultants
were Felicita Pakma (age: 28 yrs, gender: female), Virginia Hemory Pakma
(age: 23 yrs, gender: female) and Ronald (age: 23 yrs, gender: male). All of
them were from the Jowai region of Meghalaya where Pnar is mainly spoken
as mother tongue. They were quite competent in English and Khasi in addition
to their native tongue, Pnar.
Derivational Morphology and Compounding in Pnar 243
3. Derivational morphology in Pnar
Both inflection and derivation in Pnar are prefixing in nature, like its
sister language Khasi. Derivation does not only result with the formation
of new words along with a change in their lexical categories but with a
change in the grammatical structure. In this section I discuss the
morphemes i-, noŋ-, and ɟiŋ- which function as nominalizers in Pnar,
as well as briefly discuss the morpheme por-, which can be considered
a class-maintaining prefix. The addition of the three prefixes i-, noŋ-,
and ɟiŋ- results in the change of grammatical category of the word from
verb to noun while por- has been found in one instance, prefixed to a
noun to form another noun.
The nominalizer i- is a derivational prefix that changes a verb into a
noun, as shown by the data in (1)–(3). In (1) we see a simple sentence
with the finite verb bam ‘eat’ and two arguments. In (2) however, bam
has been nominalized with the i- prefix and is now an argument of the
verb e ‘give’. The data in (3) also show an example of bam ‘eat’,
nominalized by the prefix i-.
(1) u ram bam sapen u.
3.
M.SG Ram eat mango M.SG
‘Ram eats a mango.’
(2) e i-bam ja u kulle.
give NOM-eat DAT 3.M.SG horse
‘Give the horse the food.’
(3) ja ini i-bam wom u bam smat
ACC this NOM-eat NEG DVM eat fast
da u k
h
innaʔ.
AGENT DVM child
‘The food cannot be eaten by the child.’
no
ŋ- is an agentive nominalizer. Prefixation of noŋ- to a verb root
derives a nominal indicating the agent of the action denoted by the verb.
Example (4) illustrates a derivation of the nominal ‘teacher’ by prefixation
of noŋ to the verb sikai ‘teach’. This derivational prefix noŋ plays the
same grammatical role of nominalizer as evident from the data in (5)–(8)
where the nominals ‘rescuer’, ‘singer’, ‘maid servant’, and ‘dancer’ are
derived via prefixation of noŋ- to the verbs ɟarop ‘rescue’, rwai ‘sing’,
kre ‘work’, chat ‘dance’, respectively.
244 North East Indian Linguistics Volume 3
(4) ka noŋ-sikai da p
h
aʔ ko u
3.F.SG NOM-teach PST CAUS AGRS 3.M.SG
k
h
innaʔ u k
h
re kot.
child 3.M.SG study book
‘The lady teacher made the child study.’
(5) sindet sindet ki noŋ-jarop pin man ja
Suddenly PL NOM-rescue CAUS AGRS ACC
u wa u latim u.
3.M.SG AZR DVM escape AGRS
‘Suddenly the rescuers make him escape.’
(6) toʔ u noŋ-rwai?
QP 3.M.SG NOM-sing
‘Is he a singer?’
(7) ka noŋ-kre
3.F.SG NOM-work
‘the maid servant’
(8) ka noŋ-c
h
at
3.F.SG NOM-dance
‘the dancer’
Khasi has the same nominalizer no
ŋ- which changes a verb to a noun,
as shown in example (9). It shows the derivation of nominal ‘beater’ by
prefixation of noŋ- to the verb soʔ ‘beat’.
(9) u
dei u noŋ-soʔ loʔ.
3.M.SG NOM 3.M.SG NOM-beat wife
‘He is a wife beater.’ (KH)
ɟiŋ- is a derivational prefix used in Pnar to derive nouns from verbs.
However the difference between ɟiŋ- and noŋ- lies in the fact that the
former is mostly attached with the inanimate objects unlike noŋ-, which
is attached with animate entities. In example (10) we get the derivation
of nominal ‘writing’ by prefixation of ɟiŋ- to the verb t
h
oʔ ‘write’.
(10) du neib
ɦ
aʔ ka sinrun wom je de
just because 3.F.SG cold NEG ABIL AGRS
u pin dep ja ɟiŋ-t
h
oʔ.
DVM CAUS finish ACC NOM-write
‘Because of cold writing cannot be done.’
Derivational Morphology and Compounding in Pnar 245
por- may also be considered a derivational prefix in Pnar which results
in the creation of new word, although it does not change the grammatical
category. In other words, it is a class maintaining derivational prefix.
From the interpretation of the data it has been found that this prefix is
used to denote a sense of time frame or duration as illustrated in the
following example (11). In the example (11) por- is prefixed to the word
k
h
innaʔ ‘child’ to derive the word ‘childhood’. A preliminary analysis
may be that this prefix serves to create a more abstract noun from a
concrete one, in the sense that ‘childhood’ is more abstract than ‘child’.
However, further research is needed to fully support this notion.
(11) im toʔ u em i ja por-
NEG BE DVM have AGRS ACC duration-
k
h
innaʔ wa snauk men hac
h
wa u
child AZR feel happy prerequisite DVM
t
h
oʔ i ja kand
write AGRS ACC any
‘Not to have had a happy childhood is a prerequisite for writing
novels.’
Table 1 Nominalization of verbs with Khasi nominalizer ɟiŋ-
Verb Noun
njo ‘to count’ ka ɟiŋ-njo ‘counting’
kunal ‘save’ ka
ɟiŋ-kunal ‘saving’
pulei ‘to study’ ka
ɟiŋ-kunal ‘studies’
ajum ‘to irrigate’ ka ɟiŋ-ajum ‘irrigation’
Table 2 Nominalization of adjectives with Khasi nominalizer
ɟiŋ-
Adjective Noun
ɟilliʔ ‘smooth’ ka ɟiŋ-ɟilliʔ ‘smoothness’
salak ‘clever’ ka
ɟiŋ-salak ‘cleverness’
duk
‘poor’ ka ɟiŋ-duk ‘poverty’
donokor ‘polite’ ka
ɟiŋ-donokor ‘politeness’
In Khasi, this prefix is used not only to derive nouns from verbs like
Pnar but also nouns from adjectives as illustrated by the data in
Table 1–Table 2.
246 North East Indian Linguistics Volume 3
4. Causatives in Pnar
Causatives, which have been postulated as the “linguistic instantiation
of the conceptual notion of causation” (Payne 1997: 175–176), can be
interpreted within the rubric of derivational morphology from the
perspective of grammatical function changing rules. Bybee (1985: 98)
claims that in verbal morphology prime candidates for derivational
expressions are GF or grammatical function changing rules. These rules
refer to the processes that alter the number and nature of NP arguments.
A causative predicate always involves more than one argument than the
caused predicate. So the addition of the causative predicate enhances the
argument structure of the verb, such as the intransitive will become
transitive and transitive will become ditransitive. This interpretation can
be applied to the Pnar morphs pin- and tin- which are morphologically
expressed as shown by the data from (12)–(17).
(12) pin
-k
h
oid ja ka juŋ
CAUS-be.clean ACC 3.F.SG house
‘Clean the house.’
(13) pin-c
h
i
CAUS-light
‘enlighten’
(14) pin-man
CAUS-to.be
‘to make’
(15) pin-k
h
ian
CAUS-be.small
‘to make small’
The causative morpheme tin- is used in Pnar to create the notion of
violence as shown in the data (16)–(17).
(16) o tin-
jap stars u.
3.M.SG CAUS-die chicken AGRS
‘He killed the chicken.’
(17) ŋa wa tin-piaʔ ja u kloŋ.
1.SG AZR CAUS-be.fragile ACC 3.M.SG bottle
‘I broke the bottle.’
Derivational Morphology and Compounding in Pnar 247
From the analysis of the data it has been found that the two causative
morphs in Pnar are morphologically realized. They are treated as
derivational prefixes in the sense that the addition of these causative
morphs changes the number and nature of the NP arguments. As far the
distribution of these causative morph is concerned, they vary from speaker
to speaker. However as reported by my informants the causative morph
tin- exclusively convey a sense of violence.
5. The polyfunctional nature of wa- in Pnar
In the analysis of the Pnar data, the derivational prefix wa- can be treated
either as an adjectivalizer or as a complementizer. In Pnar almost all the
adjectives are preceded by the morpheme wa- although there are a few
exceptions. In Pnar, adjectives are post nominal.
wa- can be prefixed to either a verb or a noun. In either case, the end
result is an adjectival. This is evident from the data in (18)–(19)
2
, where
the adjectives ‘broken’ and ‘fevered’ are derived via prefixation of wa- to
the verb piaʔ ‘break’ and the noun k
h
oiri ‘fever’, respectively.
(18) u kloŋ wa-piaʔ
3.M.SG bottle AZR-break
‘broken bottle’
(19) u k
h
innaʔ u wa-k
h
oiri da jap
3.M.SG child 3.M.SG AZR-fever PST die
u
3.M.SG
‘The fevered boy died.’
wa- also functions as relative pronoun. In Pnar the head noun precedes
the relative clause. In the example (20) the head noun ka k
h
innaʔ precedes
the RP wa- and in (21) it is preceded by tje tje ‘who all’.
2
In the examples (18)–(19) the distribution of u is noticeable. It can occur not
only before the noun as shown in (18) but also with noun and adjective as
shown in (19). In Pnar, all nouns, pronouns and adjectives are preceded by
either
u (3.M.SG) or ka (3FSG), which denote gender. All lexical words are
specified in terms of gender. Why it does not appear before the adjective in
(18) is unknown.
248 North East Indian Linguistics Volume 3
(20) ka k
h
innaʔ ka wa-c
h
oŋ
3.F.SG child 3.F.SG RP-sit
‘the girl who is sitting; the sitting girl’
(21) tje tje wa-da wan c
h
a ka
who who RP-PERF come to 3.F.SG
khawai?
party
‘Who all came to the party?’
Sometimes wa- is used in conjunction and coordination. It is evident
from the data as shown in (22) in which wa-
is functioning as a
conjunction marker between Jack and Jill.
(22) u ɟak wa-ka ɟil
3.M.SG Jack CONJ-F.SG Jill
daŋ bam ki jo sapen.
PROG eat AGRS ACC mango
‘Jack and Jill are eating a mango.’
Like Pnar wa-, Khasi does have ba- which is assigned the functions
of adjectivalizer and relative pronoun as illustrated by the data in (23)
and (24).
(23) ka k
h
anna ba-runar
3.F.SG young .girl AZR-naughty
‘a naughty girl’
(24) ka miau ka ba-ŋa la joji
3.F.SG cat AGRS RP-I PERF see
ka t
h
nam
3.F.SG hungry
‘The cat which I saw was hungry.’
Thus it is clear from the analysis that the prefix wa- in Pnar is assigned
several functions ranging from adjectivalizer, complementizer to
conjunction and coordination. But its role as derivational prefix is
exclusively used with adjectives. In Pnar, almost all adjectives are post
nominal and preceded by wa-.
Derivational Morphology and Compounding in Pnar 249
6. Reduplication in Pnar as a process of derivation
In Pnar, the complete reduplication of a word may be considered
derivation of a new word due to the change in grammatical category.
This view can be put forward taking into consideration the arguments
given by Anderson (1992) and Aronoff (1976). According to them the
process of derivation can be treated as a morphological operation, and
both inflectional and derivational morphology are characterized by
reduplication. In addition to attaching phonologically specified affixes to
a stem, derivation is often marked by the full or partial reduplication of
a part of the stem attached to it. In Pnar the complete word reduplication
may be considered derivation as it either results with the formation of a
new word or with a change in the grammatical category as illustrated by
the data in (25) and (26). In (25) we see the adverb ‘now’ changing to
denote the manner adverb ‘hurriedly’.
(25) kre
ʔ u katni katni.
work AGRS now now
‘He does things hurriedly.’
Likewise, in (26), the adverb ‘fast’ is reduplicated to produce a manner
adverb meaning ‘rapidly’.
(26) bam smat
smat.
eat fast fast
‘Eat rapidly.’
In Pnar, the adverbial compounds are formed by reduplicated words
as illustrated by the data in (27)–(30).
(27)
c
h
eiwon c
h
eiwon
where where
‘nowhere’
(28) minnu minnu
when when
‘never’
(29) katni katni
now now
‘rapidly’
250 North East Indian Linguistics Volume 3
(30) sindet sindet
fast fast
‘suddenly’
From the above examples it is quite clear that reduplication is a
productive word formation process in Pnar. Most of the adverbials and
expressives in this language are formed by complete reduplication.
7. Compounding in Pnar
Compounding, one of the processes of derivation, appears to be an
interesting phenomenon in Pnar. In the analysis of the Pnar data we have
endocentric compounds, exocentric compounds, copulative compounds
and synthetic compounds.
An endocentric compound is one where the meaning is determinable
from one of its members. Semantically it indicates a sub-grouping within
the class of entities that the head denotes, as illustrated by the data in
(31)–(36).
(31)
k
h
innaʔ skur
child school
‘student’
(32) kso chor
ring ear
‘earring’
(33) t
h
lu um
hole water
‘well’
(34) luti kaɟat
road leg
‘foot path’
(35) sŋi u blai
day M.SG God
‘Sunday’
(36) sŋi sait ɟain
day wash cloths
‘Saturday’
Derivational Morphology and Compounding in Pnar 251
Unlike endocentric compounds, the meaning of an exocentric
compound is opaque. It is difficult to work out what an exocentric
compound means from the sum of the meaning of its constituents. It is
impossible to define them in terms of “x is a kind of y” as shown in the
following examples.
(37)
t
h
pai din
ashes fire
‘hearth’
(38) sintu den
flower tree
‘orchid’
(39) min step
duration morning
‘tomorrow’
Copulative compounds are so called because they have two words
which are coupled or conjoined. Although from syntactic perspective
they can be headed, from semantic point of view, the coupled elements
are of equal status, with neither element being regarded as the head that
dominates the entire word as illustrated by the data in (40) and (41).
(40) klom sier
talk foul
‘gossip’
(41) sula
muɟa
shirt socks
‘garments’
Like Khasi, Pnar also has a particular compound in which the head
of the derived word comprises of verb and agentive nominalizer as shown
in (42).
(42) u no
ŋ-lai kurim
3.M.SG NOM-go husband
‘bridegroom’
Pnar is very rich in the process of compounding. In the analysis of
the data we have got the instances of endocentric, exocentric and
252 North East Indian Linguistics Volume 3
copulative compounds. In Pnar the endocentric compounds are left
headed.
Conclusion
In conclusion it can be said that derivation in Pnar is carried out by
multiple means, bearing close affinity with Khasi. In Pnar the prefixes
i- and noŋ- are instances of nominalizers because the addition of these
affixes results in the derivate which differs semantically from its base,
along with a change in category. This study also presented some evidence
of a class-maintaining derivational prefix por-. Our data suggests this
prefix makes abstract nouns from concrete ones, but more work is needed
to prove this preliminary hypothesis. The role of the derivation prefixes
wa- is poly-functional as evident from the examples discussed in §5 of
this paper. Causation is morphologically realized in Pnar by the prefixes
pin- and tin-. Pnar exhibits instances of complete reduplication which
can be treated under the rubric of derivational morphology, as the process
of complete reduplication results either with the formation of new word
or with a change in grammatical category. Even adverbial compounds
are formed by reduplicated words as shown in the data (25)–(30) of this
paper. Compounding, as discussed in §7, appears to be an interesting
word formation process where we have the instances of endocentric,
exocentric, copulative and synthetic compounds. In Pnar the endocentric
compounds are left headed. While this study offers some data and anlyses
of Pnar, an otherwise little-studied Mon-Khmer language of Meghalaya,
a constant endeavour is nevertheless needed on the part of the linguists
to explore the unique and interesting characteristics of this least studied
language of Meghalaya.
Abbreviations
1 First person
3 Third person
ABIL Abilitative
ACC Accusative
AZR Adjectivalizer
AGRS Subject Agreement
CAUS Causative
CONJ Conjunction
Derivational Morphology and Compounding in Pnar 253
DAT Dative
DVM Distributive verb marker
F Feminine
M Masculine
NEG Negative
NOM Nominalizer
PERF Perfective Aspect
PST Past
PL Plural
PROG Progressive
RP Relative Pronoun
SG Singular
KH Khasi
References
Anderson, S. R. (1992). Amorphous Morphology. Cambridge, Cambridge
University Press.
Aronoff, M. (1976). Word Formation in Generative Grammar. Cambridge, MA,
MIT Press.
Bybee, J. (1985). Morphology: A Study of the Relations between Meaning and
Form. Amsterdam/Philadelphia, John Benjamins.
Payne, T. E. (1997). Describing Morphosyntax: A Guide for Field Linguists.
Cambridge, Cambridge University Press.