0% found this document useful (0 votes)
90 views136 pages

The Mahābhārata in The Tribal and Folk Traditions of India - Edited by K - S - Singh - Shimla, New Delhi, 1993 - Indian Institute of Advanced Study - 9788185952178 - Anna's Archive

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
90 views136 pages

The Mahābhārata in The Tribal and Folk Traditions of India - Edited by K - S - Singh - Shimla, New Delhi, 1993 - Indian Institute of Advanced Study - 9788185952178 - Anna's Archive

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 136

CS)

Li
— S
Ss
So
eo =
SZ
es
(oT
CD)@) <r
4 >?

This is an important book in Areal Typology of
India as it establishes that typologycally and
genetically diverse languages of the country
including the modern and tribal languages today
share their linguistic structures which are
manifestations of the shared underlying
semantic structures. Drawing examples from a
large number of Indian languages, the author
has been successful in showing that each
language, in a typical language contact
situation, preserves both individual as well as
group identity markers. Basing her arguments
on some significant but hitherto ignored
grammatical aspects, she has tried to prove that
it is mostly those semantic constructs that
pertain to perceptive and sensory abilities of a
human being that materialize in shared
linguistic structures.

This book is a challenge to historical linguists


as the homogenized signifier- signifiants
relationship that holds among diverse and
distinct languages of India, an indicator of
strong and stable multilingual community,
obscures the temporal dimensions of a language
or languages of a ‘family’. .

ISBN 81-85952-17-5 Rs. 100


Digitized by the Internet Archive
in 2022 with funding from
Kahle/Austin Foundation

https://archive.org/details/mahabharataintri0000unse
SEMANTIC UNIVERSALS IN
INDIAN LANGUAGES

deae
ie coe a |
pen oe ar

He
SEMANTIC UNIVERSALS IN
INDIAN LANGUAGES

ANVITA ABBI

INDIAN INSTITUTE OF ADVANCED STUDY


RASHTRAPATI NIVAS, SHIMLA 171005
First Published 1994

© Indian Institute of Advanced Study 1994

All rights reserved. No part of this publication,


may be reproduced in any form, or by any means,
without written permission of the publisher.

Published by the Deputy Secretary for

INDIAN INSTITUTE OF ADVANCED STUDY


Rashtrapati Nivas, Shimla 171005

Lasertypeset by AJ Software Publishing Co. Pvt. Ltd.


and printed at Elegant Printers, Mayapuri, New Delhi.
FOR
PREM SINGH AND V. PRAKASAM
oe
oa . ? are
a
Foreword

Dr. Anvita Abbi came to the Institute as a Fellow for a short period in the
summer of 1990 to work on ‘Semantic Unity in Linguistic Diversity’.
Besides giving lively seminar during her term as a Fellow, she prepared
the first draft of her monograph before leaving the Institute and prepared
the final version subsequently at Jawaharlal Nehru University, New
Delhi. The book presents a strong agrument for a ‘long and stable multi-
lingualism’ in India which has given rise to ‘the “Indianness” of our
languages’, manifested in linguistic structuration and variety. The
similarities of structures arise from the fact that ‘all Indian languages
share acommon semantic base’. Her approach to the problem, therefore,
is refreshing as well as instructive.

Rashtrapati Nivas J.S. Grewal


1 March, 1993 Director
Preface

Language differences are a topic of perennial interest in our country, but


it is not generally understood that most language differences are either
superficial or created by Indian polity. It is true that we speak some odd
number of 106 languages belonging to four distinct genetic families yet,
- along and stable multilingualism of the Indian continent has given rise
to the ‘Indianness’ of our languages. I have tried to define and identify this
common and shared ‘Indianness’ that is easily manifested in the linguistic
structurations of variety of Indian languages. What emerges, when one
tries to analyse these linguistic structures is the fact that all Indian
languages share a common semantic base which is responsible for the
similar structures that we witness today.
I, however, have restricted to only a few of such aspects such as
expressives, echo formation, word reduplication, compound verbs and
non agentive subjects. I am sure many more aspects of the various
languages of India would reflect the underlying semantic unity and would
earnestly hope that many more researchers would take up this worth-
while task. As it has taken almost three years for the present research to
come in print form, readers are encouraged to consult my work on
compound verbs published between the period of 1991-1993 also.
Iam thankful to Professor J.S. Grewal, former Director of the Institute
for inviting me as a Fellow during summer of 1990 to work on this little
monograph. His encouragement and the support of his staff had been very
constructive. Special thanks are due to Dr. Uma Rao, Dr. Raman and Dr.
Jaidev for making my stay at Shimla very pleasant. I am grateful to
Professors V. Prakasam and L.M. Khuubchandani for their comments and
criticisms on the earlier draft of the manuscript.

November, 1993 ANVITA ABBI


oa Artur, su re
Morte sismini

Seino grivtian a
“ina 2Dqu vhs bhi routes: nec Yee Mel sqrt IT
| vtBonicanet Soenworrs, ooleo ato senate fraclatHandHA ahaa

er Iu irs I susan owales gee in ucla 96 |


Breeder ye ; ¥ io

HRA ATI
Contents

Foreword

Preface

Introduction

Areal Typology of Expressives, Echo-Formations


and Word Reduplication

Explicator Compound Verbs and Indian World View

Identity Crisis of Dative Subjects and Experience


Nominals in Indian Languages

Conclusions

Select Bibliography 105


List of Abbreviations

auxiliary
causative
Compound Verb
Complete Word Reduplication
declaration
dubitative
emphatic
Echo Word
feminine
Genitive
Intransitive
locative
masculine
Nominative
nominal
Negative
participle
passive
pre/post position
possessive
progressive
past
reciprocal
Subject
CHAPTER I

Introduction

It has been known that every word, and more generally every
linguistic sign, is an entity with two sides. Every linguistic sign is a
unity of sound and meaning, or in other words, of signifier and
signified. The two components are intimately related, e.g. the Hindi
word caay ‘tea’ is pronounced in such a way that there are definite
sound sequences c followed by a long vowel aa which itself is
followed by a semivowel y. This combination of the three sounds is
the structuration of the word for ‘tea’ in Hindi, which is the signifier
of the word. And the signifier signifies a definite meaning here, for
example it refers to a “drink which is hot and prepared by tea leaves’.
If someone says caay, this signifier evokes in us the corresponding
signified, i.e. the hot drink made of some special leaves.
The 18th century and the 19th century linguists had been pre-
occupied with finding such sound-meaning correspondences across
languages which they called COGNATES. The methodology adopted
came to be known as Comparative Methods which ruled out
borrowings and chance similarity among languages. The basic
assumption on which the comparative method was based happened
to be a little unnatural, i.e. the proto language (the language or a
linguistic form said to be the ancestor of attested languages or forms)
is a homogeneous entity and is devoid of borrowings and
interferences. This implied that if we compared cognates of several
languages we could not only reconstruct the parent language from
which the ‘daughter’ languages must have descended but could also
establish group of languages belonging to the same family of
languages. In other words, relationship like mother-daughter, and
sister-sister languages were established on the basis of
sound-meaning correspondences across languages. We must note
THE LANGUAGES AND
DIALECTS OF
SOUTH ASIA

Language Families and Branches


INDO-EUROPE AN DARDIC BRANCH
FAMILY INDO-ARYAN BRANCH
INDO-IRANIAN
BRANCH
] DRAVIDIAN FAMILY
AUSTRO-ASI
aaATIC MUNDA BRANCH
SINO-TIBETAN NJ TIBETO-BURMESE
FAMILY NS BRANCH
Languages
Languages are shown
thus: (HIND
Dialects
Dialect goup shown thus:
WESTERN HINDI
Literary dialects are shown
thus: Avadhi
IIIT aSoURUIEpUY ‘DMDIDS‘asugQ asajauyuas

jJesIH aSOURUIBPUY ‘ajpg ‘apay “‘lomne‘IUDOg ‘24a

YON 189 ‘DpoYyv] Iqvluvg ‘uDyny ‘IYpulg ‘lunuyspyYodan


‘21a
YINOS SOM ‘y1Yyg ‘uoaviny 1uDyisvioy
‘94a
yinos ‘1YyIDADF ‘IuDyuoy ‘asajoyuig UDIAIPjap
‘DJ2
seq ‘asawupssy‘DsuagDAUQ‘21a
[PINpue ‘np4)-1pulH
‘yD ‘undloyg 1ypomy
‘212

YON « ‘xniny ‘ol]DW inyvig


= jenued
~— ‘n8njay ‘Iany-iny‘IpuoH ‘NuvjoyIp
ynos + ‘mvt ‘wnjpkojvpw‘vpouuny
‘vpol DION

epunyWUON
« ISOyYY
OILVISVOUL
ei SNV YON =— ‘yoyjupg “Uppuny
OR ‘212
epunyy jenusd
—= ‘Diuvyy ‘suone 1]DYON
YyNoS — ‘vlog ‘{DIDD WNLOD

ueyoqrL ‘Ysipog ‘YsluDMaN Yys14svIg


‘24a
of
esowng
Introduction

asoulingyoueiq— + ‘OjOJ “AOH] DISY-DISH


3

"BISV YINOS Jo salwey adensuey *] “By


4 Semantic Universals in Indian Languages

that all this genealogical relationship was established on the premise


that the parent language was a uniform whole and the availability of
cognates across languages was only due to genealogical affiliation.
This methodology gave us classification such as Indo European
family, Dravidian family, Austroasiatic family and Sino-Tibetan
family of languages in the Indian sub continent (Fig.1).
Indo European family is represented by Indo Aryan languages
such as Hindi, Punjabi, Marathi, Konkanki, and Kashmiri; Dravi-
dian family is represented by Tamil, Telugu, Kannada, Malayalam,
and Kurux; Austroasiatic family is represented by Munda languages
such as Mundari, Kharia, Bhumij, Ho, as well as by the only non
Munda language Khasi (spoken in Meghalaya hills) and
Sino-Tibetan family is represented by Tibeto Burman languages such
as Meithei, Kabui, Paite, and Gangte. There is a fifth family of
languages represented in India known as Andamanese (Zide &
Pandya 1968) spoken in Andaman islands of which very little is
known linguistically. Unlike other language families Andamanese
does not form a contiguous chain of linguistic community with the
rest of the country. For the sake of discussion I will concentrate only
on the first four language families. A rough geographical distribution
of these languages are as follows: Indo Aryan languages are spoken
in the Northern, Western, some parts of East, and in Central parts of
India. Dravidian languages are mainly spoken in the South India.
‘Kurux, Malto and Brahui are the only Dravidian languages spoken
in North-Central and North West (Baluchistan) respectively. Munda
occupies the Central India and Bihar region while Tibeto Burman
languages are isolated from the mainstream and spoken in the North
East of India (Map 1).
As against the backdrop of grouping languages on the basis of
genealogical affiliation, the early 19th century linguists (F. von
Schlegel 1808; C.W. von Humboldt 1822) grouped languages on the
basis of shared structural features which go in the building of various
types of signifiers. Special mention should be made of Schmidt (1926)
who distinguished between sprachfamilien (resulting from genetic
classification) and sprachenkriese (typological groupings) based on
some set/sets of features. Linguists divided and classified languages
on the basis of the phonological patterning, on the basis of word
Introduction 5

formation processes and morpho-phonological structuration, etc.


Thus came the terms like Analogical languages and Transpositive
languages (Diderot, D. and J. Le R.D’ Alembert 1772) or
Agglutinative languages and Inflectional languages (Hum-
boldt 1822, 1836), or Organic and Inorganic languages (A.W. von
Schlegel 1818), or Isolating, Agglutinative and Inflecting \janguages
(Humboldt 1822), etc. This kind of classification was based on the
processes that one employed to form words. For instance, inflecting
language like Hindi has a word Jarke ‘boys’ where separation of the
root morpheme from plural masculine suffix which is— e is difficult
to make as the root mortheme /Jarak has gone through a change
by dropping the vowel a and (thus created consonant. sequence)
while in Kurux, a Dravdian tribal language spoken in Bihar
munda+s is ‘boy’ (sg) and munda+r ‘boys’ (pl) is a case of
agglutinative language where affixes can be very neatly detached
from the base/root morpheme.
Of late, linguists have expanded their field of investigation for
typologizing the languages and have gone beyond the domain of
word boundary. Two kinds of investigations have been popular and
important: (i) the word order pattern; i.e. the way any language
decides the sequencing of subject (S), object (O) and verb (V); and (ii)
various syntactic structurations that a language or languages use for
communicative purpose. Language typology today, thus stands for
giving us information whether a language x is of SOV type (like Hindi,
Tamil) or SVO type (like Khasi, Kashmiri & English) or VSO type such
as Austronesian, Sudanic, Salish, etc. It is concerned with questions
like how Subject of a sentence is marked or how causatives sentences
are formed; whether there are lexical causatives or syntactic
causatives (the former ones are found in Indian languages while the
latter ones are respresented in English and other Indo European
languages of Europe), or how the concepts of reflexivity and reci-
procalness are indicated, etc. In fact, languages can be grouped and
classified on the basis of any grammatical aspect. Looking from this
point of view Indian languages have been typlogized as follows:
Morphologically, Indo Aryan languages are inflecting languages
while Dravidian, Munda (Austroasiatic) and Tibeto Burman (Sino-
Tibetan) languages are of agglutinative type (where segmentality and
6 Semantic Universals in Indian Languages

invariance of morpheme is maintained). Tibeto-Burman languages


like Manipuri and Gangte are agglutinative as well as incorporating
type (those languages which take a number of lexical morphemes
(words) and combine them to form a new word). For instance, in
Manipuri cap ‘to cry’; la ‘come’; |cung ‘to turn with a_ steady
motion’; na ‘too much’ and the combination of all these independent
lexical morphemes /a-cung-na-cap means ‘to cry here too much’. In
Kharia which is a Munda language spoken in Bihar and Chota
Nagpur area, agglutination is of very high degree, e.g. jib ‘touch (V)
ji-ni-b ‘a touch’ (N) where nominal infix n+vowel (after vowel
copying rule) changes verb into nouns (more examples: jung ‘to ask’
ju-nu-g ‘question’). Consider Kharia example for a_ typical
agglutinative type of languages:

Kharia

kol-ob-no?-dom-dhab-na-la?-ki-kiyar
rec. caus. eat pass. quickly progr. pst dub.
‘you both were being, fed by each other quickly’ (VM:p. 35-7)

Such structures are reduced and simplified in Modern Kharia due to


language contact situation provided by Indo Aryan languages (see
Abbi 1991 b).
Most of the languages of India maintain SOV word order though
they all did not have the uniformity to begin with. Munda languages
today are SOV type but were derived from the system of SvO type.
Present Khasi and Kashmiri are the two languages which still have
SVO ordering. Tibeto Burman languages are tonal languages while
the languages of the other three language families (barring lexical
oppositions created by the absence or presence of tones in Punjabi)
are not tonal languages. For example Tangkul Naga has rising, level,
and falling tones to distinguish words; e.g. kahung ‘red’; kahung
‘rotten’ and kahung ‘chicken’.
At the phonological level, while [A languages are rich in aspirates
(sounds like bh, dh, gh, or ph th, kh,) Dravidian languages and Tibeto
Burman either have none or have very few, or are of recent borro
wings (this is specially true of voiced aspirates). Munda languages
and Tibeto Burman languages have (glottal stop ?) such as in Kharia
Introduction 7

go? ‘carry’; which Dravidian and Indo Aryan lack it. Nasalization
such as in Hindi hé‘are’ dges not exist in Dravidian, Tibeto Burman,
and Khasi (Austroasiatic). Conversely, retroflex lateral / does not
exist in Tibeto Burman languages. Nor does it exist in Eastern IA
languages. Consult map 2 which suggests North-South divide on the
basis of a single feature, i.e. nasalization. It also suggests East-West
divide on the basis of retroflex lateral.

Special Phonological Features of Indian Languages


Typologically similar languages give us a neat pattern for cross
linguistic comparison, however, we must be aware of those features
also which make the languages of India so special and distinct. The
characteristic features of the sounds, described in the following pages
do not help typologizing Indian languages but draw our attention to
the unique distinctive character of these languages. The speculation
of the author is that these peculiar sounds will either be assimilated to
the nearest sound they resemble or will get extinct in near future
because these are unique, isolated sounds which are characteristic
features of the respective languages. The areal pressure might su
bsume them in the next 100 years. The only factor that can save them
from extinction is ‘language exclusiveness’ or ‘language
fundamentalism’ of its speakers.

Pharyngeal (+) and Glottal Stop (?)

A pharyngeal sound is produced by lowering and retracting the


root of the tongue against the pharynx wall. The phonetic symbol for
this sound is (¢) an inverse question mark without a dot. The sound is
very common in Assamese. and its dialects. The very word assam has
pharyngeal (4) in the beginning of the second syllable, i.e. asam.
As mentioned earlier, languages of the Tibeto Burman group,
Munda and Khasi of Austro Asiatic branch, as well as North Dra
vidian languages make heavy use of the ? glottal stop, (written as a
question mark without a dot underneath) the sound produced by the
constriction of the glottis. Glottal stop in these languages are
phonemically distinctive, i. e. capable of distinguishing meaning of a
word. Thus in Savara (a Munda language) spoken in Central India
bebe ‘to surround’ but b?eb?e ‘satisfactory’; /a ‘to spill’ (tr.) /?a: ‘to
8 Semantic Universals in Indian Languages

spill’ (intr); /u ‘to nourish’ and /?u: ‘to be alive’, in Kurux (North
Dravidian language) ke?na ‘to die’; ho?na ‘to take’, bona ‘hit’;
ci?na ‘to give’, etc.
Khasi (Autroasiatic, Monkhmer branch) has glottal stops in
words like /9? 19? ‘clouds’, daukha? ‘fish’ etc.

Vowels t and n
High, Central unrounded vowel [t] and mean mid back rounded
vowel [n] are prevalant in Dravidian languages like Tamil, and
Autroasiatic languages like Khasi Consider:

Tamil — kndt ‘give’ mottt ‘bud’


kudi ‘join’ —s pnt ‘lie’
vidt ‘leave’ —s_kal ‘kill’
vi-dt ‘house’ pn ¢i = ‘bind’
Khasi dnd ‘liver
p?nt ‘rotten’
shibon ‘many’

Murmur vowels [a]


Gujarati has murmur vowels in words like par ‘maintain’ as
opposed to par ‘gratefulness’ or maro ‘mine’ as opposed to maro
‘beat’ (imperative).

Lamino Dentals

Khasi has word final lamino dentals in the production of which


blade of the tongue touches the lower teeth (Abbi 1978, 1989). Since
the obstruents are unreleased word finally it is difficult to distinguish
between the voiced and voiceless laminodentals. Consider: toid ‘to
flow’; biet ‘mad, dull’; ksaid ‘water fall’, etc.

Implosives

Stops which are produced by using glottalic ingressive air stream


mechanism (for details of processes involved in production of
implosives see Ladefoged [1975]) are abundant in Sindhi (Indo
Aryan language spoken in India and in Pakistan [in Sindh]) and
Introduction 9

Multani (also an IA language spoken in Multan in Pakistan).


Implosive is marked by a curled hook on the top of the consonant
letter.
Consider:

Multani: — billa ‘cat’ tidu ‘small frog’


billa ‘badge’ didhu ‘one and a half
dhidu ‘stomach’

Sindhi budzho ‘curse’ gi-ta ‘song’


Budzho ‘whole’ gi:ta ‘gulp’
da:ba ‘a coin’
da:ba ‘lung?’

Unreleased Stops
Though occur only in certain specific phonological environments
such as in word final position yet quite prevalent in Tibeto Burman
languages, as well as in Khasi, Munda, and Konkani (IA language).
Mundari has ‘kuij’ ‘to dance’; Kharia has goej’ ‘to dance’; Konkani
uses it in rap’ ‘to live’, thak’ ‘to lie’, udak’ ‘water’ etc.; Meitei:
lek’-‘lick’ tak’ ‘teach’ sak’ ‘sing’; Khasi : ksaid’ ‘water fall’; mrad’
‘animal’, etc.

Rolled r
Khasi and Konkani, though genealogically, typologically, and
geographically very distant from each other share this feature among
themselves. Khasi r is however more rolled than Konkani r.
Consider:

Konkani
bairo ‘deap’
bare ‘good’
gharat ‘at home’
rastyan ‘on the road’
10 Semantic Universals in Indian Languages

Khasi
mrad’ ‘animal’
rospa ‘rich’
rben ‘be thick’
shtnrang ‘male’

There are other phonological features which can help classifying


languages in bigger groups. Features which typologize the languages
of this subcontinent into various subgroups have been suggested
earlier (Ramanujan and Masica 1969). The map 2 given earlier is a
good indication of dividing the languages of this country into two big
halves. One can either visualize North-South divide on the basis of
nasalization feature alone or West-East divide on the basis of
retroflex lateral /. In fact, larger the number of features under
consideration, finer the divisions are. It is perhaps better to restrict
the groups of features to typologize languages.
We have thus grouping of languages based either on the genetic
relationship or on typology. India represents, as stated earlier, five
families of languages each diverse in its genetic and typological
makeup.
What I am going to show here that labels like genetic affiliation
and ‘real cognates’ are hard to establish as genealogically diverse
languages of a contiguous area, if exist for a long time, interact with
each other so much so that they share more features among them-
selves than they share with their sister languages spoken elsewhere
distanced by geographical barriers. ‘Languages in contact’ of an area
are marked by high rate of bilingualism and one finds a good number
of ‘borrowed’ lexicon which resemble lexion of the donor language
both in sound and meaning. A long and stable contact of this kind
leads not only in the sound and meaning resemblances but eventually
resemblances in grammatical structuration and meaning
correspondences. Let us call them STRUCTURAL COGNATES. For
instance, the grammatical structuration of Hindi reduplicated word
baithe baithe ‘while sitting’ (sitting sitting) is shared by all Indian
languages though the word is translatable in different sound
sequences (so different that it might be unintelligible for Hindi
speakers). What is important is the fact that all languages share the
Introduction 11

same semanteme of ‘continuity’ manifested in this reduplicated word


‘to sit’. The intimate relationship that I tried to establish between
sound and meaning earlier as in caay is similarly being established
between the linguistic structuration, i.e. verbal reduplication in
adverbial position and the meaning, i.e. the contnuity aspect. baith is
a verb in Hindi but after duplication is used as an adverb such as in vo
baithe baithe so gaya ‘he went off to sleep in a sitting position (or
because he was sitting for a long time)’. Such structurul cognates are
the real indicators of areal contact between languages. The languages
in contact, however, do maintain their grammatical individualities to
some extent, so as the member of a particular language community
identifies himself with it.
An area comprising of common linguisitic features shared by
various genealogically unrelated languages is known as Linguistic
Area, e. g. South-Asia, South East Asia, Meso-America areas, etc.
We can extend the definition of linguistic area to incorporate also
those languages which were originally typologically diverse but later
shift to homogenity due to contact situation. This implies that if a
linguistic feature with its corresponding semanteme is found in
various typologically and genetically diverse languages of a
contiguous area then that particular feature is an areal universal. A
bundle of such areal universals will mark a strong semantic unity
among its diverse languages. Besides, wide existence and use of such
features in various languages will confuse the comparative
methodologist, as he has no way of isolating ‘structural cognates’
from what he calls ‘real cognates’.
Areal features originate as ‘interference’ in languages in initial
stages! of bilingualism. In speech it is like a sand carried by a stream.
It is easily identifiable as ‘foreign’ or ‘borrowed’. In language (i.e. the
system) it is the sedimented sand deposited on the bottom of the lake,
which has become part of the constitution of the lake. The
‘interference’ phenonemon of the initial stages of bilingualism
becomes established and habitualized in the language of the subse
quent speakers who might or might not be bilinguals. The feature
gets established as if it belonged to the language. Hence our earlier
example of baithe baithe type when found in Tamil, Marathi, Kharia
or in Naga, it appears as indigineous to the systems of each of the
YOrVN S3SSO190S!
(H9NOY)
——— xa1s0uLgy
YO) (QaLovu¥lay
IVOldv SdOLS

«seers
“==: IWSYN 13M0A S3N3NOHd

UO/GNY

WS8VN
SIVUALVT

XITIOULIY
dew
Z
<In0g
: “yy uefnuewey
pue ulOD Bose (6961)
Introduction 13

languages though originally the structure developed out of the


convergence of the linguistic features in a typical language contact
situation.
The process of convergence implies a simulteneous process of
divergence of the languages too. When language A becomes like
language B in areal contact situation, it also starts deviating from
other genealogically related languages of its stock. This is inevitable
in the process of language change and of emergence of linguistic area.
India has always been in a constant dynamicity of bi/multilingualism
for as long as 3000 years. The contact has generated several struc
tural cognates, i. e. linguistic structures which are common at the
structural and content levels across various languages. I will consider
three such structural cognates or areal features in the foregoing
discussion: (i) Expressives and Reduplication; (ii) Explicator com-
pound verbs (ECV); and (iii) Dative subject constructions. The
phenomena of reduplication and explicator compound verbs will be
dealt with in the next two chapters and a detailed investigation into
experiential constructions and other constructions which have
oblique marked subjects (commonly termed as ‘Dative Subjects’)
will be taken in the chapter after the next one.

NOTE
''U. Weinreich (1968).
CHAPTER II

Areal Typology of Expressives,


Echo-Formations and Word Reduplication

There are linguistic structures in Indian languages which occur as


repeated occurrences such as Hindi ghar ghar ‘house house’, i.e.
‘every house’. Technically, these structures are known as
reduplicated or simply reduplication. Reduplication, by definition
stands for repetition of all or a part of a lexical item (word) carrying a
semantic modification. Thus Hindi word gharghar ‘house house’
is derived by repeating the word ghar completely and da-dal ‘in the
habit of striking’ in Santhali (a Munda language spoken in parts of
Bihar, Bengal, and Orissa) is derived by partially repeating the word
dal ‘to strike’. In Santhali one can also say dal dal ‘strike intensively’.
Reduplication thus can be partial or complete.
There is a kind of reduplicated structure prevalent in all the lang
uages of the world including Indian which is derived by repeating a
syllable which constitutes a word. An isolated syllable does not carry
a meaning of its own, it acquires the meaningful status only after it is
repeated. The entire repeated construction works as an unit
morpheme as well as an unit word. These are called EXPRESSIVES
(Abbi 1985), e.g. kinship terms such as Hindi caca ‘uncle’ or dada
“grand father’ or onomatopoeic words like khat khat ‘knocking
sound’. Expressives are a kind of Mophological Reduplication while
Structures like in ghar ghar ‘house house’ we saw earlier, where
the syllable which undergoes reduplication is a complete word itself,
is an example of Lexical Reduplication. Both morphological and
lexical reduplication are areal phenomena which do not find their
absence even as exceptions in any single language of the
subcontinent.
Areal Typology of Expressives 15

The question we can ask is if all the languages of the world employ
expressives then what is so great about these structures as far as we
are concerned? The answer to the question is seen in the typical
socio-cultural states of Indian community. The most significant
aspect of Indian expressives is that they indicate five senses of per-
ceptions of.Indian speakers. Secondly, the expressives behave and
function like any regular word and thus forma part of the lexicons of
Indian languages. Regular verbal paradigms are formed by adding
verbal affixes to the expressive word; e.g. Hindi taptap ‘dripping
sound’ can be formed into a regular verbal paradigm as fap
tapata (m) tap tap-ati (f) ‘drips’ or tap tapa- hat ‘dripping’
etc. The nature and abundance of use of expressives in Jndian
languages can be visualized by the examples given below.

Expressives:
1. Acoustic Noises
(a) Animal Noises (for monkey’s chattering)
Methei : u?u?;
Hindi : khaav khaawv
(b) Noises of Natural Phenomena (for rain pattering)
Hindi ; tap tap
Meithei : tap tap
Mizo : klek kpek
(c) Noises made by Humans (laughing sound)
Hindi 2 khi khi
Meithei F khe khe
Mizo ; hek hek
(d) Noises made by miscellaneous inanimate objects
(jingling of anklets)
Methei 3 tchrin tchrin
Mizo : tchek tchek
Hindi : chun chun

2. Sense of sight (shining-twinkling)


Hindi : cam cam
Nepali : bharbharaaun
Mizo : SEP SEP
16 Semantic Universals in Indian Languages

3. Sense of touch (sticky)


Nepali : cyat cyat
Methei ; tsanaap tsanaap

4. Sense of smell
Tamil:
samay ai aarail irundu gama gamanu
vusanai varadu
kitchen in from exp. good smell coming
‘A nice aroma is coming out of the kitchen’

5. Sense of taste
Hindi:
cirpira : spicy hot’
cat pata ; ‘tangy’

Another feature which makes the South Asian languages distinct


from other languages is the use of expressives as manner adverbs.
This use gives Indian speaker the linguistic facility in bringing out
very minute and subtle distinctions in a performance of an action
which might go either. unnoticed or unexpressed in non Indian
culture. To substantiate the claim I have chosen Khasi (an
Austroasiatic language spoken in Meghalaya hills) examples. It
should be noted that learning these languages implies learning of
these expressives as a sentence without a difinite manner adverb is
considered incomplete. The knowledge of expressives, thus,
constitute an integral part of language competence.

Expressives : Khasi

iaid’ ‘to walk’


iaid’ bak bak ‘to walk/go hurriedly and unheedingly’
laid’ dondon . ‘toddle like a child’
laid’ weng weng “swaying”
iaid’ toin toin ‘walking without hesitation’
jaid’ tuin tuin ‘walking with hesitation’
iaid’ ter ter ‘walking in an order/in a line’
iaid’ dot dot/dut dut ‘(of an old person) weak and unsteady
walk’
Areal Typology of Expressives 17

iaid’ dar dar/der der ‘walking quickly and with light


steps’
iaid’ kniah kniah ‘walking mincingly’
laid’ /eng leng ‘walking carelessly’
laid’ Jong long ‘walking blindly’
laid’ tay tay ‘to waddlle’
kem ‘to catch’
kem thrait’ thrait’ ‘to catch something adeptly’ (of a snake)
leh ‘to act’, ‘to do’
leh thrait thrait ‘to do or to act unhesitatingly’
mareh ‘to run’
mareh thrait thrait ‘to run without losing hope or vigour’
beh ‘to pursue’
beh thrait thrait ‘to pursue with an aim of catching
somebody’
(kren) blaapblaap ‘(to speak) jabberingly’
(li?) blajblaj ‘(to be white) bloodlessly’
(jem) nupnup ‘(the way a bear’s fur is) soft and hairy

(khi) kdikkdik ‘(the way the soul throbs) longingly


(le?) bianbian ‘(to do something) carefully’
(?iej) thonthon ‘(to love) intensely’

Interestingly, expressives are not restricted to the perceptual


words and manner of actions. There are many in various languages
which lay bare a large number of feelings, situations, attributes,
disorder, disturbance, confusion, untidiness, etc. Nepali excels in
expressing various types of confusion and disorderliness by expres
sives. At least 15 different expressives to represent disorderliness
have been noticed by the author (for detail, see Abbi 1991 a).
There is an interesting folk puzzle in Telugu which exposes the
general contention that the knowledge of expressives and their
semantic correlates are considered pre-requisite to man’s intelli-
gence. Once a king used an expressive puttukku jara jara dubukku me:
as a puzzle to be solved by one of his court poets. The king
apparantly saw something while taking a round of his kingdom. The
poet, having pondered over it for some time, solves the mystry of the
18 Semantic Universals in Indian Languages

expressives in the following way:

There was a small hut with a thatched roof. A bottlegourd creeper


had spread all over the roof of the hut. The creeper bore many
bottle gourds. One day one of the fruits got disconnected from the
twig (producing the noise putukku) and quickly rolled down
(producing the noise jara jara). There was a goat resting under-
neath the edge of the roof. The bottle gourd fell on the goat
(producing the sound dubukku). When the gourd fell on the goat,
the goat cried ‘me :’
(Peri Bhaskararao 1977)

Many languages of the Munda family form words by expressives.


For instance, in Kharia ra? ra? ‘flower’ dung dung ‘eel’; mas mas
‘relax’ etc. In other words, reduplication of a syllable to constitute a
single lexical category manifest many semantic units which share
more or less similar ‘semantic fields’ accross the languages of South
Asia. M.B. Emeneau (1969) in his essay ‘Onomatopoetics in Indian
linguistic Area’ showed that there were at least 39 sound-meaning
correspondences (cognate type of structures) of Onomatopoetic
across Dravidian and Indo Aryan languages. He terms them Areal
Etymologies. He treats these forms as structural traits of Indian
linguistic area, offering a possibility of treating Onomatopoetics as
subcorpus of linguistic material that will yield to comparative treat
ment. Emeneau compares data from Kota (a non literary South
Dravidian language) and forms from Dravidian Etymological
Dictionary (DED) (1960) and Dravidian Etymological Dictionary-
Supplement (1968) (DEDS) for Dravidian, and Comparative
Dictionary of Indo Aryan Languages for indo Aryan (CDIAL). In
general, he classifies the ononmatopoetic forms into two basic types.
1. Systematic class.
2. Non systematic class.
Onomatopoetic forms of the systematic class are described in
terms of a basic CVC, with either no derivative suffix or one of a few
derivative suffixes: -k,-r,- ak, -a:r, -a:l, -um. These are of two major
types.
1. Non-reduplicated.
2. Reduplicated.
Areal Typology of Expressives 19

I. Non-reduplicated
1. CVC with no derivative suffix:

(bird makes noise while flying up from ground)


tan in -, (to cool off)
2. With derivative suffixes:
CVC + -k/-r/-ak/-a:1/-a:i/
caykn ‘to be sprained’
garrn ‘with noise of thunder’
kulak in- ‘body feels cold’
dabakn ‘with noise of falling crashingly’
gabakn ‘with noise of jumping into water’
cata:r ‘noise like whip crack
daba:r/l ‘noise of body falling with crash’

II. Reduplicated
1. Identical reduplication with or without derivative suffix.
(a) With no derivative suffix:
cor cor ‘walking on dry leaves’
cur cur ‘while roasting meat’
dam dam ‘drum beating’
cur curn ‘boiling hot’
pat patn ‘tears falling in abundance’

(b) With derivative suffix:


civk civk in ‘birds twitter’

2. Reduplication with change of vowel derivative suffix/ change of


initial consonant/ change of vowel + initial consonant.

car cur in ‘noise of snake’s motion’


dap dop ‘repeated beating’
doba.r daba.r- ‘crashing sound’
ve.k me.k in ‘to make sport of someone’
tat pot-n ‘startling suddenly from sleep’
20 Semantic Universals in indian Languages

Other Dravidian languages, Toda and Kodagu show less number


of onomatopoetic words. For Kolami, no forms are recorded, how
ever, material from the literary languages is recorded in DED.
Taking the constructions in Kota as basic, Emeneau identifies such
constructions in comparative terms. The construction as a direct
quotation, followed by an obvious form of the quotative verb
*en/"in/"an is seen in Tamil, Toda, Kannada, Telugu, Konda, Pengo,
,and Kui.
o-enal o-enru
The suffix -n found in forms which are used as expanders of the
verb (adverbs) has as cognate terms in Malayalam (potukkanu
‘suddenly’, Kannada tottene-‘suddenly’), Kodagu, Tulu, Telugu,
Parji, Gondi, Kui, Kurukh, Tamil, Malayalam, and Kannada have
verbs formed directly from onomatopoetic forms.
Tamil - kilukilu ‘to rattle’
Malayalam- kilu kilukka ‘to rattle’
There are traces of this in Tulu and Telugu. Kurukh and Malto
have numerous formations of similar nature which are not, however,
related morphologically to the South Dravidian. Both non-
reduplicated and reduplicated formations are probably found in all
the languages.
Emeneau says that a tentative analysis of Indo Aryan material
(based on Sir Ralph Turner’s Comparative Dictonory of Indo Aryan
Languages) yields interesting preliminary results.

CDIAL 5342
Punjabi jham jham ‘glittering’
Hindi jham jham ‘glittering’
Punjabi jhamak-na ‘to glitter’
Hindi jhamak-na ‘to glitter’
CDIAL 6092
Buddhist Hybrid Sanskrit thara thar-aya ‘to tremble’
Prakrit thara thar-edi ‘feel giddy’
Bangla thara thar ana- ‘to tremble’
Marathi thartharne ‘tremble’
Areal Typology of Expressives 21

Morphological characteristics of the onomatopoetics can be


summarized as:
1. Reduplicated and non reduplicated stems.
2. Reduplications-identical and non identical.
3. Verbal derivations from onomatopoetic stems.

Historical treatment of such forms generally rested on the


assumption that they can be newly created at any time and that the
phonological rules of a language do not generate such formations.
This implies that no genetic origins could be traced for these forms
and that in different languages, similar or even identical forms may at
any time arise independently. If we accept this view of Hoffmann
(1952), then we have to accept that such forms are evolved in
individual languages and onomatopoeia is not an areal structural
feature of Indian linguistic area.
Emeneau argues against this case. The argument that the phono
logical structure of each language does not dictate the forms of its
onomatopoetics is true only within limits. Though the special pho-
nological features of the onomatopoetics are deviant from those of
the language itself, they are reducible to rules and these rules show
some consistency with the total set of rules of the language. So there
is no validity in a statement that similar or even identical forms at any
time arise in different languages independently. In languages of
greatly differing phonological structures, there will probably be
found few such coincidences of form. If several contiguous related
languages, with phonological structures much alike, have both an
onomatopoetic system and many onomatopoetics in common, gene
tic inheritance, or diffusion, or both, can be assumed. Independent
development will be highly improbable. Likewise, if several
contiguous languages of different families have an onomatopoetic
system and a fairly large number of onomatopoetics offer
sound-meaning correspondences, then diffusion will be the probable
explanation. The more divergence the two phonological systems, the
more probable this explanation. The two major families Drividian
and Indo Aryan have very similar systems and also have within those
systems a sufficient number of areal etymologies to make the
explanation of diffusion most plausible.
22 Semantic Universals in Indian Languages

An investigation into the modern languages of South Asia, also


show ‘structural cognates’ in the.area of onomatopoetics across
languages. Interestingly enough, a few expressives denoting various
kinds of noises (true onomatopoetics) across languages show
sound-meaning correspondences too. This can be taken as a marker
of heavy borro wing or complete diffusion of the structure under
consideration. Consult Table 1 for explanation.
Other than expressing acoustic noises, words of expressives types
can be formed from those words from which they draw the semantic
content. Consider the following examples from Drividian, Indo-
Aryan and Tibeto Burman languages:

Word meaning Tamil Hindi Mizo

‘reddening with anger’ civa civa-in tamtam ana sen tawn tawn
‘eyes filled with water’ kangal dab daba mithui parawl
; kulamayina
Reddening with anger:
Tamil civa civa-in/civa ‘red’
Hindi tamtam ana ‘to reddin with anger’

CDIAL 5779
Hindi tamra ‘copper coloured’, ‘dark red’.
Pali tamba ‘red’, ‘copper’
Prakrit tamba red (adjective and noun)’
Dameli tramba ‘red’
Sinhalese tamhba ‘reddish’
Hence there is a possibility that the Hindi onomatopoetic
tamtam-ana had the origin in the word *tamra that means copper or
red.
Eyes filled with water:
In Tamil there is no onomatopoetics as such for this. It is expressed
in the sentence kangal kulamayina ‘eyes became pool’. The word for
‘pool’, however becomes onomatopoeic in other languages.
Hindi dab dab a-
aTaVL
1:
saatssaidxy
ul ueipuy sa8en3ury]

epunyy sasensury] ueipiurig sasensury]

yetpueg
= uepunyw O19qI],
OzI
ueuling
eUlaS eSeN

eT punog
jo Suimoy ypes| ypes ins jn3 eles eyed [ex Tey [ey [ey rey rey
JOA
o¢, punos
JO MOJO ACY ARY rey Jey ey ey rey Jey key Avy
Ae
ye
ey

€ Surjoouy 0} 0} YO) 40} eI ed FoI YO} Fey yeyy yney yney YOY YOY
, s39aq BUIZZNq uns eueunsuns un3 mo mo uns uns uns uns TA 1A Surq Suiq
"G sy Suryoeso - . ied jed ivy jez jeo jeo yeu yeyi S8und 8und
‘9 punos
JO SpiIg pp 9999 WPT oY pp xDD detyo detyo s1y9~—s ayo
ft 1995 S][2q ureyd wureyo weyd wey [eo [eo ureyo ureyo
=| ureyd wey deyua deruy Iny Iny
°g BSulsuts
Jo ay) [[9q
| UIPH UIptj
= SUL SuLI uep uep Sur} Sur} ur} ue} ITY Jtyo Suey) Suey)
6 193em Sutddup doj doi dei dej dei dej di dij dei dei dey dey yod teyeyy
‘OI punos
JO uinip wre} we} 8urpis 3urpis
| wepwep sunyjes wep wep uinp uinp Suap Suap
Bunyies
24 Semantic Universals in Indian Languages

CDIAL 5530
*dabbara ‘mud’, ‘pool’, ‘vessel’
Bangla dabar ‘water vessel’
Hindi dabar ‘low, hollow ground where
water collects’
Old Marwari dabara ‘pool’
Marathi dabura ‘hole dug for water’

These pieces of evidence show that the presence of this type of


expressive is also an areal feature. These words are based on the
lexical items in the respective languages. This proves that though the
individual onomatopoetics are phonetically different from language
to language the system is an areal feature. The signifiant is the same
and shared across languages. The signifier is different only
superficially as it shares the system and pattern of structuration
among contiguous languages. The absence of a reduplicated
expressive in Tamil for ‘eyes filled with water’ and the fact that the
language uses the word ‘pool’ for this signification which it shares
with other I.A. languages that employ the word ‘pool’ in a
reduplicated expressive form proves the point that languages of this
subcontinent draw their signifiers from the common and shared
semantic fields.
Emeneau (1980: 279-80) considers some of the areal etymologies
which are indicators of language contact situation that must have
prevailed around 500 B.c. He draws his data from Dravidian and
Indo Aryan languages to prove his point:

DED 2428

Tamil taka-tak-enal ‘dazzling’


Kota dag dag-in ‘burns brightly’
Tulu = daga daga ‘brightly’
Telugu daga daga ‘glitter’

CDIAL 6704

*dhagg- ‘throb’, ‘glitter’


Prakrit dhagadhagai ‘flares’
Hindi dhagdhagana ‘to glitter’
Areal Typology of Expressives 25

Marathi dhagdhagné ‘to glitter’


Gujarati dhag dhagru ‘to burn fiercely’

CDIAL 5522

“dag ‘tremble’
Prakrit dagamagai ‘trembles’
Punjabi dagmag ‘trembling’
dagdagauna ‘to tremble’
Nepali dagmag ‘to tremble’
Assamese dagmag “‘unsteadiness’, ‘sparkle’
dagmagaiba ‘to glitter’
Bangla dagmag ‘trembling’
dagdagiya ‘glowing’
Oriya dagadaga ‘briskly’, ‘hastily’
Hindi dagdagana ‘burn brightly’
dagmagana ‘to tremble’

In Santhali dak means ‘heat’. This is used as a noun. The word is


also used denotatively to mean vehemence and fierceness.
We can examine another such onomatopoetic term that means
‘doing something quickly’
Tamil catput-in of cat-in
Hindi cat cat
Santhali chat pat
In Santhali, the expressive has taken full load of lexical item in the
language.

Santhali cat pat (adj) ‘active, brisk high spirited’

With minor phonological variations, related meanings are


expressed.
chat pati ‘brisk’, ‘energetic’
chat pat ‘hard working’
chat patia ‘brisk (energetic children)’
‘quick (in mischief).
26 Semantic Universals in Indian Languages

It is more striking to note that the word is also used as a noun and
stands for a lucky plant.
chat pat (N) ‘A small plant’

Evolvolus hirsutus. Wild. When women take


anything to the market to sell, some put this
plant in the rice, etc. or in their hair-top, in the
belief that by doing so they will quickly sell their
products. Men also sometimes put it in their hair
when taking goods or timber away to sell.

Could the semantic content of ‘quickness’ originate from the name


of the plant, which helped selling things quickly? It could be any
body’s guess.
There is another areal etymology of chi chi which is used to denote
anything dirty or filthy. It expresses an emotion, an aversion towards
something. The word is used in Hindi, Tamil, Telugu, Santhali,
Mundari, Sema Naga and other languages. The non reduplicated
from ehi]chi or forms with lengthening of the vowel chi: chi: are
found incall these languages.
Santhali derives verbs and adverbs from this expressive. It is also
used with varying suffixes to give various subtle meanings related to
‘disgust’
Consider:

chi ‘intensive of disgust, disapproval, scorn’


(often reduplicated)
chi ‘used for stronger disgust’
chi ‘spurn’, ‘scorn’, ‘despise’
chi-ketae bae calaka ‘she rejected it with contempt’
chi-a-tin ‘shame on’ (tin-poss-nom-suffix)
chi-atam ‘shame at’
chi-atae ‘shame on yow’
chia- ‘disgust’
chia chia (intensive) of loathing and disgust’
chiak ‘despicable’, ‘contemptable’

Looking at these structures we cannot rule out the possibility of


Munda being the donor of the constructions and the other languages
Areal Typology of Expressives 27

of the area the borrower.


Selvam (1988) cites Telugu expressives which describe feelings like
‘bodily sensations of various kinds, from tactile ones to pain, dizzi
ness and even gut reactions such as impatience, anxiety,
embarassment, loneliness etc.’ (ibid: 47). These feeling expressives
exist in Indo Aryan as well:

Hindi Telugu Gloss


kir kir-i gara gara ‘feeling of a dust particle
in the eye’.
jhar jhar bota bota ‘flowing down of a liquid from
a whole (blood from wound) or
(tears from eyes)’.
tam tam-ana dhuma ‘to be very angry’; ‘seem to be
dhuma very hot’.
phak tes ‘to become pale in the face’.
sunn j(h)um ‘feeling numbness in limbs;
feeling giddy etc’.

To summerize, Indian Expressives manifest feelings, sensations,


perceptions pertaining to five senses, manner of an action and dis-
orderliness. All these relate to the ability of PERCEPTION of the
speaker in a cultural specific environment. The ‘world view’ and its
manifestation in expressive morphology is a typical feature of Indian
linguistic area.
Lexical Reduplication or word reduplication with specific types of
meaning is another important areal feature which has originated due
to language contact. Lexical reduplication can be partial such as
those found in Echo formation, and in Compounding; or it can be
complete. A typical South Asian linguistic feature is Echo Formation
or Echo word; represented words like Hindi khana vana ‘food’ etc. or
Punjabi roti shoti ‘food etc’; Bengali ca-ta tea etc. These are examples
of echo formations where the base word is followed (in some cases
preceded, see Teluge examples given on the next page) by an echo
word. The echo word is formed by replacing the initial syllable of the
word in question by another syllable (this syllable varies from
language to language) keeping intact the canonical shape of the
28 Semantic Universals in Indian Languages

word. Most often it copies the initial vowel of the base word; hence
Hindi sag vag ‘vegetables’ etc. has-a-which is copied in the echo word
also. Dravidian languages, however do not follow the vowel copying
rule but copies the length of the vowel of the initial syllable of the
word, e.g. Telugu cinna ‘small’ and cinna ginna ‘small etc’. but
cu:ttam ‘to see’ cu:ttamgi:ttam ‘to seeetc.’ Dravidian replaced
syllable is gi or ki. Please refer to examples given below. Echo word
echoes the sense of the base word and provides a semantic
extension.

ECHO FORMATIONS

Nominals
Tamil puli ‘tiger’ puli gili ‘tiger, etc.’
Telugu puvu ‘flower’ puvu givu ‘flower, etc.’
Hindi phul ‘flower’ _phul vul ‘flower, etc.’
Adjectives
Tamil pacca ‘green’ _pacca gicca ‘green and the like’
Telugu cinna ‘small’ cinna ginna ‘small, etc.’
Hindi mota ‘fat’ mota-vota ‘fat and plump’

Pronominals
Tamil avan ‘he’ avan givan ‘he’ etc.’
Telugu atanu ‘he atanu gitanu ‘he, etc.’
Hindi mé ‘7 me vé ‘I, etc.

Verbals
Tamil vandu ‘come’ _ vandu gindu ‘come, etc.’
Telugu vintam ‘hear’ _vintam gintam ‘hear, etc.’
Hindi jana ‘go’ jana vana ‘go, etc.’

Aadverbials
Telugu tondarga ‘fast’ tondarga gindarga ‘fast, etc.’
Hindi jaldi ‘fast’ jaldi valdi ‘fast, etc.’
While IA languages, more or less conform to the structure expli
cated above for forming echo words, Dravidian languages have
additional strategies to form such words. For instance, in Telugu,
Areal Typology of Expressives 29

‘extremities’ might be represented by echo formation when echo


word precedes the word in question. The first syllable of the word in
question is repeated, and syllable -tta or -rra is added to this. This
newly derived echo word is then made to precede the word in
question. Consider:

patta pagalu < pazalu ‘broad day time’


batta bayalu < bayalu ‘very open’
motta modalu < modalu ‘right at the beginning’
kotta kosa <_ kosa “extreme end’
urra (u:) ta iets 1) “swinging too much’
virra vi:gu < vigu ‘to be extremely: proud’

For further detail see M. Sastri (1969) and Selvam (1988). Other than
gi- Telugu uses go-, sa (:)-, so-, su- ti-, be- as replacer syllables of the
echo words. It may also use lexically empty putra, natra, gatra added
to the word in question to denote generality.
Nouns are the most echoed grammatical category. Verbs, adjec-
tives, pronouns and personal pronouns are echoed but the frequency
is much less than found in the nominal category.

Tamil ate otacce kitacce pootaate


that break EW NEG (imp)
‘Don’t break (or do such a thing)!’
kuticca kiticca kaal otaccitaate
jump EW leg NEG break (imp)
‘Don’t break your leg by jumping (or such activity).
Malayalam atine potticca kittica kalayaruto
that break EW do NEG (imp)
Don’t break (or do such a thing) that.
The conjunctive participle forms could also be
echoced:
Hindi jaldi naha- vaha-kar aao
soon bathe-EW-CP come (imp)
Lit: ‘Come soon after having bath (and such thing)’
‘After getting ready come soon’.
30 Semantic Universals in Indian Languages

It has generally been accepted by linguists that echo word has no


individual occurrence and meaning of its own. Abbi (1985) does not
consider it lexically empty because it adds meaning of generality and
nonspecificness to the meaning in question; however, she agrees that
this meaning is acquired only after the Echo word is attached to the
word under consideration. Recently my attention has been drawn to
those morphological compounds which are like echo formations in
meaning but not in phonological structurations. That is, no copying
rule or reduplication is used. Instead of using an empty morpheme,
one uses meaningful word from the lexicon to form an echo
formation. Two independent words from the lexicon are drawn to
form a compound where the second word delexicalises itself from its
inherent meaning but serves as an echo word manifesting the
connotation of ‘etc’. ‘generality’ and ‘such things’.
Consider:
Tamil
tuni mani ‘clothes etc’ < tuni ‘cloth’, and mani ‘ball’;
mulla killa ‘thorns etc.’ < mulla ‘thorn’, killa ‘pinch’
The occurrence of killa above could also be due to the real Echo
word in the language which allows ki-replacer syllable. Whether
killa here is Echo word or an independent word meaning ‘pinch’ is
not known. Whatever the origin, the compound here is of echo
formation type. Malayalam examples given below are intriguing as
the two lexically independent words have nothing in common but
still collocate with each other and represent generality, the typical
meaning manifested in echo formations.

Malayalam
paamba ceemba (snake+ vegetable [tuber]) snakes and the like’.
vendi kindi (cost + bronze vessel [used in rituals]), ‘vessels etc.’
vali kili (way+ small cloth purse) ‘way or such thing’

The last two examples, like in case of Tamil, could also be the
instances of replacer syllable of echo word ki- and not of
independent lexical units. If they are independent lexical units, then
Areal Typology of Expressives 31

the question is, should they be considered echo words on the basis of
their semantic import? If yes, we must include collocationally
restricted and lexically selective expressions like the following which
are frozen expressions in the language:

Malayalam:
aana ceena (elephant + vegetable (tuber) ‘unimportant
matters and the like’
tengnga mangnga (coconut + mango) ‘nothing and such’
Hindi:
gajar muli (carrot + reddish) ‘insignificant’
Why the word for ‘vegetable’ is used for denoting generality and the
concept of ‘unimportant’ is anybody’s guess.
Steever (1987) opines that: ‘Echo compounds have two facets of
meaning. First, like other reduplicated compounds, they have
distributive meaning so that the compound conveys the idea, entities,
or actions of which echoed word refers to a random example from
general range. Secondly, echo compounds conventionally carry a
pejorative nuance to the effect that the speaker neither likes nor cares
enough about the entity or action to specify it any further
(744-45).’
The relevant question before us is: should we define echo
formations from the semantic point of view or from the point of view
of phonological structuration? If the basic criteria is semantic then
we must also include collacationally restrictive compounds that we
have just examined in Malayalam and Tamil. However, if we
consider the criterion of phonological structuration, 1.e. the principle
of replacer syllable followed by the canonical shape of the word
under consideration then we have to take into account only those
formations in which one of the words is partially repeated and
semantically non occurring on its own in the language. Linguists
have generally restricited themselves to the latter types of formations
while defining echo formation. We, in this book, are concerned with
the ‘structural cognates’ —the identity relation that exists between
words that have common meaning-structure relation across
languages.
32 Semantic Universals in Indian Languages

We have come to the threshold of deciding whether two collocated


words in which one is completly delexicalized but adds the new
meaning of ‘generality’ following the pattern of other echo word
(which are lexically empty to begin with) are instances of echo
formation or not. Since we are concerned with structural cognates as
defined earlier and above we cannot include items like Malayalam
aana ceena ‘unimportant matters and the like’ under instances of
echo formations. These should be considered under compounds
where either of the two lexical units have the potentiality of losing its
basic meaning. It should be brought to the notice of the readers that
explicator compound verbs (treated later in the book) offer similar
morphosemantic structures where words functioning as explicators
lose their original meaning.
Other than semantic echoing, echo words reflect also
‘FUNCTIONAL ECHOING’. When a word in question is derived into its
echo formation, the construction thus derived echoes the functions of
the items manifested in the base word, For instance, Hindi kursi-vursi
‘chair etc.’ really means ‘chair or any such item which is used for
sitting’. Thus the echo formation kursi-vursi creates a set of items that
is used for sitting purposes. Another good example is caay-vaay ‘tea
and such things’ and pyaalaa vyaalaa ‘cup and such things’. While
the former is an instance of semantic echoing denoting items like tea
and other eatables that may go along with it, the latter is of
functional echoing where any container of a cup type is referred to.
Consider a few examples from Hindi of functional echoing which
basically create sets of entities serving similar functions:
pen ven ‘pen and any such thing (used for
writing)’.
kitab vitab ‘book and any such thing (used for
reading)’.
palang valang ‘bed or any such thing (used for lying
down)’
kagaz vagaz ‘paper or any such thing (used for
writing).’
The demarcation line between the semantic and functional echo |
ing is not very clear cut. The language use and pragmatic factors help
Areal Typology of Expressives 33

us in deciding which is what. For instance, consider the following


two sentence from Hindi, where in the former pen-ven echoes
semantic contents of the base word pen ‘pen’ while in the latter,
pen-ven echoes the functional use of the nominal item ‘pen’.
1. vo pen-ven lene baazaar goyaa hai
he pen-EW get bazzar go-past
‘he has gone to the bazzar to get some stationery’.
2. bhai, ghar mé koi pen-ven hai kya
‘brother, house LOC any pen-EW exists Wh?
‘Is there any pen or any such thing (with which
I can write) in the house ?

To conclude, it is clear that echo formations in Indian languages


share their structural configuration as well as the associated semantic
parameters across language families and language typologies. These
structures manifest generality, casualness, nonspecificity, extremi
ties, and group identity. Echo formations, in nominal category, are
also used to create sets of entities with similar functions. Now let us
consider other kinds of lexical reduplication.
Lexical reduplication could be also of DISCONTINUOUS type when
a syllable is inserted in between the two reduplicated words such as
Hindi kam-se -kam ‘minimum’ from kam ‘a little’; Khasi
toh-sa-toh ‘crafty’ <toh ‘to steal’; Punjabi vad- 16: -vad
‘maximum’ <vad ‘increase’. Dicontinuous lexical reduplication
may manifest concept like ‘non specificness’ also. Consider Hindi
(IA) and Khasi (Austro Asiatic) examples:

Hindi Khasi
kabhi -na-kabhi sano-pa-sano
‘sometime or the other’ ‘sometime’

kahi -na- kahi sano-re-sano


‘somewhere or the other’ ‘somewhere or other’

koi-na-koi uno-re-uno
‘someone or the other’ ‘someone’

Jatt -na -jau


‘may be I go, may be I don’t’
34 Semantic Universals in Indian Languages

aye-na-aye
‘he might come or might not’

These structures indicate attitudes of the speaker. They are used as


hedging strategies which are important discourse level linguistic
units.
Then there is Semantic Reduplication used in forming those
compounds where the two words entering into compounds are either
synonyms like Hindi-Urdu dhan-daulat ‘wealth’ (dhan ‘wealth’
(Skt); daulat ‘wealth’ (Persian), or polor opposites like uthnaa
baithnaa ‘frequenting (uthnaa ‘to rise and baithna ‘to sit’) or related
words such as in mol-tol ‘bargain’ (mol-value’ and tol ‘measure’).
Such compounds are part of language universals and donot interest
us specifically for the investigation under consideration.
The most fascinating and contact induced linguistic structure
among class of reduplication is of COMPLETE WORD REDUPLI
CATION (CWR) which can either be of class changing type or of class
maintaining type. The former is represented when the reduplicated
word under consideration belongs to different grammatical class
membership of its non reduplicated counterpart. For instance, Hindi
verb baith ‘sit’ can be used as reduplicated adverb as in baithe baithe
‘while sitting’. We have changed the verbal membership of baith into
adverbial by reduplication. This is one of the very productive devices
to form words across grammatical classes. Munda languages are very
effecient in creating new words by this process. In fact,the
multilingual situation that exists in these communities make the
speakers adapt and adopt new words and associated concepts by
duplicating the existing ones. For instance, consider Kharia
examples:

Kharia

CWR (Class Changing)


goej’ ‘to die’ (v) goej’ goej’ ‘dead like’ (Adj)
bor- ‘to ask’ (v) bor-bor-lebu ‘beggar’ (N)
no ‘to eat’? (v) no-no ‘right hand (used for eating)’ (N)
Jung ‘to ask’ (v) —_jung-jung-dan ‘engagement’ (N) .
ao ‘to live’ (v) ao ao-kar ‘inhabitant’ (N)
yo?‘to see’ (v) yo? yo?-kar ‘onlooker’ (N)
Areal Typology of Expressives 35

Reduplicated verbal adverbs of our baithe baithe types are uni-


versally used by all South Asian languages to indicate aspects like
simultaneity, continuity, iteration, sequentiality and non precipita-
tiveness. This is one feature which marks the South Asian languages
distinct from the neighbouring languages. It gives the languages of
the subcontinent a group identity. The structural cognates shared by
all Indian languages are very strong and can be confirmed by a small
sample of Hindi examples given below:
Hindi
1. barish hote hote rah gai
rain happen happen stayed went
‘It was about to rain but did not’ (Non Precipitative)
2. main use citthi likh likh ker thok gel
I him letter write write-CP tire went.
‘T got tired of writing him letter (again and again). (Iteration)
3. vo khate khate bola
he eat eat spoke
“He spoke while he was eating’ (Simultaneity)
All these sentences are translatable in other languages of India. I
have specifically chosen the aspect of ‘continuity’ to demonstrate the
presence of reduplicated verbal adverbs across four language
families of India. Consider:
Indo Aryan
Hindi vo baithe baithe thok gaya
Assamese __ tai bahi bahi bhagori gol.
Bangla she boshe boshe haple uthe che
Dogari o bei bei ge utt geya
Gujarati te betho betho thoki gelo
Konkani to: baeson baeson thokon gelo
shmiri su thok b’ih b’ih
Marathi te basun basun thokle
Maithili o balsel balsel thalk gel
Oriya se khali bas/basI thokI goelani
Punjabi o baithe baithe thok geya |
Sindhi ho wahende wahande thakji pyo
Sadari u: baethe baethe thalk gelok
36 Semantic Universals in Indian Languages

Tibeto Burman
Kabui komai la?na dundune
Meitei mahak fam-na fam-na cok th-ram-y
Paite gma tutulai opukta

Munda
Birjia huni duru duru thokayena
Kharia hokker doko doko thakke gotki
Mundari enhore dubdub logaIn
Dravidian
Kannada ovonu ku:tu ku:tu sustada

Kodagu ava arittaar itta cotepoci


Malayalam corkon irinat fa irinatt a vesomiccu
Kurukh as bonte bonte xar dlyes keras
Tamil utkarandt utkarandt ove nokku eluttu
vitta do
Telugu atanu ku:rconi kurconi olasi po:ya: du
The Map no. 3 will give us a perspective as to how wide spread
these structural cognates are among Indian languages and how very
distinct from the neighbouring languages of South East Asia. The
term ‘Reduplicated structures in general’ used in Map 3 refers to all
other kinds of CwR of noun, verb, adjective and adverbial classes
which are again in abundance in Indian languages (Refer to Table 2).
Reduplicated nouns and adjectives manifest in concepts like
distributiveness, emphasis, exclusiveness and attennuation, while
reduplicated adverbs indicate emphasis and intensification.
Most of the Indian languages do not reduplicate finite verbs.
However, a few selected languages of the Austro Asiatic branch and
some of the Tibeto Burman languages such as Taizang do reduplicate
the finite verbs. Most of the languages of the subcontinent, specially
Indo Aryan and Dravidian do not share this feature. Consider the
following sentence from Taizang, a Tibeto Burman language spoken
on the border of Burma. There are approximately 20,000 speakers on
the Indian side. Tiazang has not yet assimilated areal features of
Indian languages and is a proof of how geo-political barriers block
the areal diffusion.
NOILNGIY
JO LSIG GALVOIIdNGSY SAYNLON
NI YLS
OGNI LSVSHLN
VISVOS

; es
VWane

(
VISAVIV

O31V9IIdNIWe8Y3A
G3y @Yy3A0V

G3LV9ITdNaSy SSYNLONY
NI LS 1V493N39

¢ dey
Adon AIOA, SUTYTeM SULATEM, oul OU, skoq sXoq,
ieuu3 ieuus = jedured j (sured
Spay, SY
weluoy wefuoy ey ey npuapou npuepeu
OM}, OM} 31q 31q, ASPJ SRY, 3 oY 94, .Poolq poo;g,
npues npued eAtiad eAtiad e3oA e3A ueae ueAe Ww yer weq}e1 [Ruel
pol pod, BUIOS, nok nof,
eBIWIO} BIWIO}
Say, OY
niy43 njy3 Tel [el ajof aj0f
sq 31q, Aevou IvoU, OM OM, asnoy asnoy,
esueg
OM}, OM}
ojnp oinp 019q 010q ayoey syoey elwe vIUIe a10ys a10y43
yonu, <yonuw .Ppo ppo, suryyem suryyem, nok nod,
1yyxd tYyyAo epjeo epyeo ny 1 ny
IL skoq shoq,
epel epel
Oya au, JSPJ SJ,
iqefung
‘6

3Uul 3 opunu opunut


904}
9914},
31} 31} e}iID eIRI0 ays Wey
pal pol, Zuryyem Suryyem, nok nof,
[el [el a}ed a1]€9 win} Ty wn}
op OAM},
OM}

sq 31q, MOJS MOIS, Ii shoq 4oq,


eleq vieq oyp s1typ 3UI 3UIS ayde] oye] npi-IpulH

sodensuery
Op
sIsyMUeN?)

S v € (4

I
L
soaloolpy SQIOAPY sunouolg sunON
ee ie sea a a De eR Se eS
ie EE
sadenduvy] ueisy ynog ul uoHBddnpay P10
¢ ATaVL
asnia], ur ur ue TuUea Is:nd IS:Nd 1pa
Ipaa exo Tjeyo
ssnoy, oSnoy wry, TY das, das 104, Joy dUO, UO
eppnu eppnw esIopuo} ella eselio
duny, dun, eBSIOpUuo} pos, .Pa1
ISBJ, ASBS
weledeeyp]
eA LI€A a];oUT Q[[aUr eqye[eA ere eve
SMOJ, SMOI Apjos, Anjos eqe[ea yey, Jey
joy9) 104 3} 9UdIOY QUdIDy ITey, ATey
SYSOIQ, SYIOTQ Zuldaom, Ssuldoem
yyniny ued wed BXIO eXI0 O1exnul OJexnwW
Teueu
SUIUIOW SUIUIOUL, pols, <PEHS YORI, PRIQ
jyey,

BAO exYO
9U0,
,2U0

op
Jey9ju09ju0

sunys, Suns poo3, .poos


fing fing pe Ipe worrYyp woTYyp 1040p Jofop suojoqsuoj
npnuoqgnpnuw
souep, oy, oy ATMOS, A[MO]S 3UOT, BuO] presje, au0,Aq <2u0

eueyy
Gyesuury) <Preye
J0q 10q (nqeJ) rt
1e330q, yeym Jeym suoye Zuoje 1INq "IN
fey

{nex

ZuIZuIS, SUIZUIs sure,


[ews
|[euws,

Suryyem
:20IN0g
14QV
S861

SunuM Jlas Ino


,2UIOS aWWOS 8unuUM, .puNnoI punol, BZurlyjem Suryyem, 0} nod, S[JId sjlis,
yex 1ex TY AIzZ HIz wnj-e-win| PAS WAS) ueu {o] ueU jew jouw nu
OM} OM}, 3UI}Vd BuIA, ,poos poo3, .MO[S MOJS, JIesAu 0} J, shoq sKoq,
PALaPALt Ty ou ou yeyd-v-jeyd om Ce) 13] {3] 12y [es [es-od a3uey
(SdATISUD})
08 03, Aqieou, Ajaverpourur, (Aou},
9-119] uef uel ejUeul e}UeUr Ty WY
,2u0 9UO, no 4nd, Jey ej, Apyomb, (gjasanoA) .nod, 21009,
Tomn od romn wod wiod Tey] ley] eq eq tyd ry tyd suou jy Suou Iseuy
€ATAVL
sq], [Bary
|, [edIsoOjodA SdNSlI9}9VIeYD
JO YjNOS uvIsy uoneddnpsy

sad&] Jva180j0ydiopy
sadQqng

jeotsojoyudiopy gy SOAIssaldx

[eoIxa] oysqg «———-suoneuloy


1
pio) (uonesdnpey jenseg
& OYOA-UON SUOTBULIO.
«—|

1
1

snonuuoosiq

0102:
42 Semantic Universals in Indian Languages

Taizang

1. keika papai nangna na sai sai-m


I I go go you you work do do-(decl.)
‘(Since) you are continuously doing your work,
I am leaving !
2. nangna lai si sial
you you book read read (prog.)
‘You are reading your book (continuously)’.
3. kei curacanpur-a ka pa pai
I Churachandpur-Loc [ go go
‘I go to Churachandpur very often (again and again)’.

In short, the morpho-semantic and the morpho- syntactic proper


ties of the reduplicated structures of Indian languages force these
structures to be recognized as strong areal universals (For detail on
this issue see Abbi 1987 and 1991a).
Geographically seen, north central India seems to be the focal
point ofthe distribution of all types of reduplication. This is the area
which is represented by a large number of Indo Aryan languages,
almost all of the Munda languages, and some Dravidian (North
branch) languages. Munda group of languages not only share the
areal character of the reduplication in general, but also use the
structure for varying syntactic functions such as causatives,
reflexives, reciprocalness, passives, etc.
The areal typology of reduplication with its structural cognate
relationships across Indian languages can be schematized
in Table 3.
The semantic constructs are given in dark area which map with
their morphological structuration given on their left under ‘morpho
logical subtypes’. Only the partially reduplicated words have been
given a name, i.e. ‘Echo Formations’. The extreme left column
indicates the basic broad divisions among Indian reduplication from
the point of view of morphophonemics of the languages. As said
earlier, morphological reduplication consists of meaningless
iterative syllables which make up a word Lexical reduplication, on
the other hand consists of two iterated words (lexemes) and function
as a single lexical item.
Areal Typology of Expressives 43

NOTES

'The entries listed in this and following pages are from Turner’s Comparative
Dictionary of Indo Aryan Languages.
2Dravidian Etymological Dictionary ed. by T. Burrow and M.B. Emeneaul,
1960
>Dravidian Etymological Dictionary- Supplement, ed. by T. Burrow and M. B.
Emeneau, 1968.
CHAPTER III

Explicator Compound Verbs and


Indian World View

I would, now, like to exemplify linguistic structuration which signi-


flies common semantic field from which various Indian languages
draw their structural cognates. It is a well known fact (Masica 1976)
that Indian languages make use of the sequence of two verbs V1+ V2
as finite verbs. The main meaning of the verb is contained in V1 while
V2 is lexically emptied, 1.e. grammaticalized. V2 in most cases, serves
as a modifying or explicating element of v1. Thus the term Explicator
Compound Verbs (ECV). The V1 is generally known as the Polar verb.
It is the V2 which takes the burden of representing tense, mood, and
aspect markings. It also receives the agreement markings if the
language so demands. Thus, in the following Hindi sentence the
explicator ja- ‘GO’ is a fully inflected form for temporal, spatial and
agreement information. (Explicators are in capital letters).
1. vo jaldi se baith gai
she hurriedly Instr. sit Go - past. 3 fem sg.
‘she sat hurriedly’
2. vo jaldi se baith jayega
he hurriedly Instr. sit Go - future, 3 mas. sg.
‘He will sit hurriedly’
3. ve uth kor cal diye
they rise-CP walk GIVE - past. 3 mas pl.
‘they left suddenly’
In Hindi cal-dena walk (V1) + give (V2) the meaning ofV1 cal is
the main meaning, the second verb dena ‘give’ indicates ‘suddenness’.
There is a very selective closed set of verbs which can function as
vectors and generally come in pair of opposites; e.g. COME-GO;
Explicator Compound Words 45

TAKE-GIVE; RISE-FALL; KEEP-THROW; SIT-STAND etc. The number of


vector types varies from 6 (Meithe) to 13 (Marathi) across languages.
In addition to being /exically restricted, explicators are also lexically
selective, i.e. ‘a given V2 combines only with such V1 as one
compatible with it or... as have the same semantic potential for it.’
(Hacker 1958, Katenina 1957). This is one reason why explicators
should not be considered aspectual auxiliaries or modals.
Illustrations given below of Hindi V1 mar ‘to hit’ with various
explicators give us the range of possible meanings of the compound
verbs. Some of the expressions are now frozen expres sions and are
used idiomatically.

Verbal Compounds: (Vi+¥2) Hindi


(hit +...) v2 Gloss
1. mar dena 1 GIVE to beat, strike, hit
2. mar dena 2 GIVE to kill
3. mar dalna PUT to kill outright, to slay, to murder
4. mar lena 1 TAKE to kill oneself
5. mar lena 2 TAKE to win (a game etc.)
6. mar lena 3 TAKE to seize, take (a fortress etc.)
7. mar lena 4 TAKE to earn smartly and cleverly;
to cash in, to earn the profits
8. mar lena 5 TAKE to make or save money
(in a business)
9. mar lena 6 TAKE to embazzle, to convert
to one’s own use
10. mar lena 7 TAKE to beat someone and to take
pleasure in doing so
11. mar rokhnal KEEP to lay dead
12. mar rokhna2 KEEP to withhold (something valuable)
wrongfully
13. mar baithna 1 SIT to beat someone unwillingly
14. mar baithna 2 SIT to refuse to return a thing
15. mar lana 1 BRING to obtain by theft or plunder
16. mar lana 2 BRING to trick or swindle a person out
of a thing
46 Semantic Universals in Indian Languages

17. mar giranal FALL to knock down, to lay down


18. mar girana2 FALL to kill (animals/birds)
19. mar bhogana RUN(tr.) to put to flight, to rout
20. maar hotana REMOVE _ to beat back, to drive away

Semantic Typology
Although, there has been much debate and little agreement among
writers on the most-general function of the class of explicators, there
is no doubt that individually in the various South Asian languages
where they occur, explicators indicate similar ‘types’ of meanings. At
the highest level these meanings may be classified into aspectual,
adverbial, and attitudinal types. Each of these types can further be
seen in terms of several sub-types: attitudinal meaning for instance,
may mean the marking of humility, respect, contempt, anger or
surprise that the speaker or narrator feels towards the actor or action
concerned. Consult (Table 5). Aspectual meaning is manifested in
perfictvity, completion or action being seen as a whole (Table 4).
Adverbial subtype may indicate manner (Table 6b) benefactive and
other types (Table 6a).
Consider some examples of ECV in Indian languages (explicators are
translated in capital letters).

Explicator Compound Verbs


I Aspectual
Malayalam (4) vriddhan maRicca pooyi
old man die GO
“The old man died’

Telugu (5) atenu mithai mottom tini vesaidu.


he sweets all eat THROW
‘He ate up all the sweets’

Hindi (6) burha hathi mar gaya


old elephant die GO (pst)
‘the old elephant died’

Kashmiri (7) vagiv jo:ra ni: ge:a kheth


mats two TOOK cow having eaten
‘the cow ate up the two mats.’
Explicator Compound Words 47

IT Adverbial
Malayalam (8) raju sinima kandoa karahné pooyi
Raju film see-pst-prt cry GO
Hindi (9) raju sinema dekh-kar ro para
Raju film, see-CP cry FALL
‘Seeing the film Raju cried (could not control)

Meithei (10) thang-no thin dat’ lei


knife-Inst. pierce DISJOIN-past
‘(He) killed him with a knife’ (mercilessly).

III Attitudinal

Malayalam (11) nookkikolu! aval ippol oru kevita eluti tallum


watch ! she now a poem write PUSH
‘Just watch her write a poem now ! (Contempt)
Hindi (12) ye main kya kar baithi
this I what do SIT
‘Oh what have I done ! (Regret)

The actual semantic articulation of what is basically one notion


may not always be identical in the case of two languages. One of the
important aspectual categories manifested in South Asian
explicators is of PERFECTIV!TY. However, when the notion of
perfectivity is used by varying language speakers it is variously seen
to mean (i) event or action seen as a whole; (11) action thoroughly or
exhaustively done; (111) completely done action; (iv) total
achievement in different languages. Clearly, al! these belong to the
SAME semantic field and it is this we take as significant. The variance
in the semantic constructs is due to different ‘world views’.

Functional Characteristics

Two important points emerge on examination of explicator


systems in individual languages. Firstly, it is clear that explicators
can be multi-functional. In other words, an explicator in a language
can have more that one function and this is true of most if not all
explicators in different languages. Illustration of this is provided by
Hindi-Urdu explicators /e ‘take’ and de ‘give’ that show meanings of
48 Semantic Universals in Indian Languages

TABLE4
Aspectual

Language Perfective/Action drawn to last point/Action


seen as a complete whole
INDO-ARYAN

Hindi jaa ‘go’, aa ‘come’, le ‘take’, de ‘give’


Punjabi jaa ‘go’ aa ‘come’, lai ‘take’, de ‘give’
Kashmiri hyun ‘take’, yun ‘come’, (shunun ‘wear’
Bangla jaa ‘go’, aash ‘come’, ne ‘take’, de ‘give’
Marathi za ‘go’, ve ‘come’, kad ‘draw
DRAVIDIAN

Tamil vitu ‘leave, poo ‘go’, itu ‘put’


Malayalam ita ‘put’, poo ‘go’, kaala ‘throw’
Telugu poo ‘go’, pett ‘keep’, wees ‘throw’
Kannada hoogu ‘go’, bitu ‘release’
Kurukh kaal ‘go’, ci? ‘give’
AUSTRO-ASIATIC

Santhali NA
Kharia godna ‘pluck’, cena ‘go’
Gta? we ‘go’, bi? ‘give’
TIBETO-BURMAN

Meithei thok ‘exit’


Kabui tang ‘do, finish’, /au ‘put?
ATAVL
¢
eurpnynry
—?0_€020606060O60O _——
asensury] AUN iduiaju05 yadsay astiding
ye 19189y
IO 198uy
10 \sn8siq
ssoupa}oadxaun Ayeinsuay
Jo 10 uonesodse

NVAUYV-OCNI AI OTe, yireqals,


| ts id
iqefung req als,
WIWIYsey unysjes
08,

Ariqessapuy
yesuag ysoq als,
TUIeILJy Seq Als,
NVICIAVAG
[rue], [OY ,UTeJUOD, DY Je}, [nse 0RIs, niood and, rejo2 198, ASO]
ood 08, 170 ystiad,

joy
ood

jnue
eeey

Bjeey
nsnjay
038,

uooinny
ais, Mee 3Ip,

.20e18,
MOI),
mori,
edyeyn?

epeuury

ureyuod,
Aonsap,

NSooy
,08,

ysnd-ejj:27

wieyedepeyy
nyeey and,
{nan ¢19 PM18, | reece
jeey 08,
NVWaNg-OLAGIL
TayMIW IG DAI8, UIs 28ueLe, yoy? axa,
10 doo, Iq (2at8,
mqey ey lOD go
—_—_—_—_—
TaVL eo-A

[BIqIZApY (JouuBUI-UON])

osensuery uonoerjousg /2{qIsIaAeq]


su0gpue oneydwq -edionuy WOAOIUTW2AQ
yes Joyj0 qNoyyIM 303 J2A0 oUyoq A10} uonoe uonoe
Apoulol quMm oUOp
ul
sourApe

NVAUV-OGNI
IpulH 3 O4P1, ap als, poop and, 4304 deay, 9) 24k) ap Als,
iqefung 10] OTP}, ap Ald, jins Mo1y}, 10] 241, ap PAIB,

LImWysey undy .o¥e1, ununys} unysjv3


08, ununys] .Jeam, ununys] eam,
Jeom,
vsueg au 24e1, ap QAI8, y3D4 .daoy,
Tyee a8 24k, ap OAS, yi and, aayi daay,

NVICIAVUG

[rmey joy ule\U0dNpoy DAIB


= 00d .08, nj1 and, joy .UIeIUOD,104 daoy,

weyekeley npoy dAI8, 00d ,03, DjDDY MOI}, 104 <daoy,


n3snpoL uoy aye,
= iad aay, 2un deoxy,

epeuury jo¥ .UTe\UOd,


NPOY OAIS, nit (aq,
qyniny 219 nls, I90X P91,
OILVISV-OULSNVY

eueyy vupos ond,

NVWuUNd-OLASLL
yew 19% A138, oIs/uls wnys dooy,
0}, ,Adoo
Jo
0}, 9q .Ul
ATAVL
(9)9
[BIqI9Apy (sauuB]A])

eee Ee
a
o8ensueq /ueppng INO UOM]OA ajeroqtfaqou0q-~—YIM= ou0q Ajisea
= suoq /YUSTOTA, Ajaatsuajuy
ydniqe Aynouytp /Ajjenseo /2AIsIoap AJaANSNeYyX
A[ssaporeo oNseip auop
NWAUV-OGNI
IpulH yin Osu, wd [lej, jpop and, 4049 aavay
= joop and, j00p and,
aod le}, ap OAIS,
iqefung
ppoyo sAea],jins MOIYI,
y LIMyse uodd ey,— unysjp3
08, undy Oki, ununys} ununys}vam

yjo
Jeom,

asl,
40d

jayd
ej,

esueg
MOI},
MOI},

SOQ
jaydjayd

IS,
MOIU},

tyyerey|
nja

[rue
njjoj

24k,
880},

~—s
Njood

ysnd,

00d00d
03,08,

NVICIAVUd
DIDDY
DIDDY
vjjv}

J n3nja 1104 71y,

MOI},
.ysnd,

MOIY},
saan and, Sainod MOI},
DYDDPUN

O4DUL,
Says saam

We|eARTEP
AGMB
Moy},and,
epeuuey nsooy08, NP1G DARI], NPI PARI], nyooy and, nyovy and,
NPIG AR9],
yyniny (19 Pal, (19 Pals,

(19

«
OAIB,
yoasg, 290x
OILVISV-OULSNV
yes {708 nid,
poy dis,

poy
dis,
NWWUNg-OLAGIL
PyVoW [!s/urs ney JO .yeo1q, 1DY]
0}, Adoo MOd8, <urolsip,
Inqey wog Aeys,

eee See—k—
54 Semantic Universals in Indian Languages

‘self-benefaction’ and ‘other-benefaction’ respectively in some cases


and that of ‘perfectivity’ in others.

(13) ae ek kamiz silva lo


one shirt get tailored TAKE-imp.
‘Get a shirt made’ (for self)
(13) b. shyaam ne saara dudh pii liya :
shyam-erg. all milk drink TAKE-pst.
‘Shyam drank up all the milk’ (completive)
(14) a. ek kamiz silva do
one shirt get tailored GIVE-imp
‘Get a shirt made’ (for someone)
(14) b. radha ne baat keh dii
Radha (erg.) matter say GIVE-pst.
‘Radha revealed all’ (completive)

Another illustration of this is seen in the Austro-Asiatic Language


Kharia where the explicator godna ‘pluck’ shows the meanings of
‘perfectivity’ as well as ‘intensity’:

(15) a. ravan raja goej-god-ki


Ravan king die PLUCK-pst. intr.
‘Ravan the king died’ (perfective>

(15) b. da?-te ho-te dul-gore-m


water-acc that-in pour-PLUCK-imp., tr., r sg.
‘you throw away the water there’
(intensive/exhortative)

The second point noted is that explicators just as much as other


grammatical elements or lexial items can be homonymous. What this
means is that a language may use more than one explicator to mark
the same meaning. To illustrate, Kannada explicators hoogu ‘go’ and
bidu ‘release’ both mark lack of volition.
(16) moguvige pada doalli hunnu aaki hoogide
child Gen. foot-in wound happen GO
“Wounds developed in the child’s foot’
Explicator Compound Words 55

(17) moguvige pads dalli hunnu aaki bittide


child Gen. foot-in wound happen LEAVE
“Wounds developed in the child’s foot’.
To conclude we may say that explicators in South Asian languages
are drawn from similar lexical sets and are used with main verbs in
the respective languages in privative contrast? with simple verbs, to
indicate similar types or range of meanings.
Inspite of such sharing at lexico-semantic level, one can identify
typical Indo Aryan explicators and typical Dravidian explicators.
For instance explicators COME and SIT are typical Indo Aryan while
those of the latter are HOLD/CONTAIN and possibly THROW.
Conversely, one can identify a typical explicator meaning of a
specific language family. The case is in point is ‘action
introvert/overt’ (Table 6 a) for Indo Aryan languages. This semantic
feature is conspicuous in the area by its absence in the other three
language families. Another instance of a semantic sub-type whose
marking is familv specific, is the representation of attitudes such as
humility, contempt and respect in Dravidian languages alone
(Table 6 a).
At the level of actual manifestations such as sound sequences or a
length of a word, or a number of explicators and their individual
meanings etc. it is true that languages show differences but they are
strikingly similar from the point of view of the semantic parameters
involved. What ought to be of permanant interest is not the fact that,
Malayalam has explicator tala for intensively done action and
Kashmiri has so:s for carelessly done action, but the fact that the
languages of different stocks and of different typologies belonging to
one linguistic area use a similar linguistic device to represent/signify
the signifiant/kernel semanteme. It is this semantic sameness or unity
that is the underlying principle behind the ‘Indian-ness’ of the South
Asian languages. Further, the semantic similarity among South
Asian languages is what lends credibility to the notion of a ‘Semantic
Area’, possibly one of the factors contributing to what is commonly
known as Indian culture or ethos.
56 Semantic Universals in Indian Languages

NOTES

' I am thankful to Professor Halmut Nespital for supplying me this list.


? For detail see Abbi and Gopalakrishanan ‘Semantics of Explicator Compound
Verbs in South Asian Languages’ in /ndia as a Linguistic Area Revisited edited by
Anvita Abbi; a special issue of Language Sciences (forthcoming). I am indeed
thankful to Dr. Devi Gopalakrishanan for helping me sort out the semantic
classification that these explicators fall into.
3 By privative contrast we mean that sentences with ECV can be substituted by
sentences without ECV without any violation of inherent meaning of the sentence. The
substitution of this type will lose the discourse content as well as several pragmatically
oriented features contained in ECV. This is evident by the two following sentences from
Hindi.
a. hori ne p2kaure kha liye
Hari erg. pakoras eat TAKE-pst.
‘Hari ate up the pekoras’ (undesirable)
b. hari ne sare pokaure khae
Hari ate the pakoras’
CHAPTER IV

Identity Crisis of Dative Subjects and


Experiencer Nominals in Indian Languages

Manis affected by his actions and performances. He can both initiate


and be initiated into an action. He can both be affective to or affected
by his or someone’s actions. He can be an ‘actor’ as well as an
‘experiencer’ of an action. The two qualities are inherent to any
normal human being. There are innumerable number of activities,
happenings and processes which an individual member ‘undergoes’
or ‘experiences’ which are not directly related to his ‘doings’. He is a
passive participant in these experiences. For instance, the sentence ‘I
am happy’ may convey the subject ‘I’ to be in a situation of
‘happiness’ without any direct activity undertaken by ‘I’. ‘I’ is ina
state of happiness because of the factors outside the controlling
powers of the ‘I’. ‘I’ is simply a passive experiencer of a particular
state. In a sentence ‘I washed my face’ ‘I’ is not a passive experiencer
but an active member who performs an action of ‘washing’ at will. ‘I’
here is a ‘doer’ of an ‘action’. In English the experiencer or affected ‘T’
and actor or doer ‘I’ are marked identically (Zero marking). Not all
languages of the world do so. Languages of the world can be
classified into two lots. Those which make a distinction between the
first type (affected) and the latter type (actor) and those which do not
make such distinctions linguistically. The languages of the former
type employ some specific linguistic encoding for the nominal entity
‘undergoing’ an experience or for being in a ‘state’ with reference to
an object and a distinct one for the nominal entity engaged in an
action. Languages of the South Asian subcontinent do make a
distinction between ‘subjects’ who act, do, or perform an action from
‘subjects’ who undergo, experience, have, become or any such
58 Semantic Universals in Indian Languages

phenomenon which is ‘out of control’ of the subject nominal. We


would, for reference sake, call the former as ‘performative’ and the
latter as ‘non performative activities’.
How do the languages of South Asia mark this distinction? Nor
mally the languages use dative/accusative marking on nominal of the
non performative though genitive and instrumental markings are
also used. Masica (1976) calls experiences of the non performative
type as ‘subjective experiences’ which typically include liking and
disliking, state of health, or sickness, happiness and unhappiness,
dreaming, feeling, remembering, thinking, embarrassment, pity,
doubt, pain, thrust, hunger, sleepiness, anger, urgency and knowing.
(1976: 160).
He further points out an interesting paradox that these languages
describe such ‘subjective experience’ (which are a kind of inter
nalized states/experiences) from an, ‘external point of view, that is, by
putting the experiencing subject in an oblique case (most commonly
the dative) and either making the experience itself the grammatical
subject or, less commonly, using an impersonal (and generally
deleted) grammatical subject’ (1976: 160).
The oblique marked subject nominals have been popularly termed
as ‘dative subjects’ in South Asian linguistic literature. Oblique
marked subjects however is a better term because not all languages
mark these nominals by ‘dative’. As mentioned earlier, genitive,
accusative, and in certain rare cases, instrumental markings are also
used. We, in this study, will make use of this term ‘oblique marked
subjects - (obl. m. Subj.) to incorporate the ‘dative subject’ as well as
those subjects which are marked by ‘accusative’, ‘instrumentive’, or
‘genitive’. The examples given below are drawn from Dravidian,
Indo Aryan, Munda, and Tibeto-Burman families of languages.'
Oblique subjects are marked for a variety of constructions. The
illustrations given in Fig. 1 and 2 have been attempted to cover
almost all types of constructions marked so. As we shall see soon that
barring some of the Tibeto Burman languages all other languages do
use obl.m. Subj. for all the types of constructions. Experiencer con
structions are the ones which are most marked by oblique subjects.
The nominal element which serves as a subject in all these
constructions either take dative or any other oblique marking in
Identity Crisis of Dative Subjects 59

South Asian languages.


I will first consider non experiential types of constructions and
then experiential constructions to discuss the issue. The chapter is
divided into four parts. The first part is about those language which
use oblique or any other kind of indirect marking on the so called
‘subjects’ of the non experiential types of constructions. The second
part is concerned with defining and identifying the experiential
constructions on the basis of cross linguistic data drawn from all
language families. Various types of experiential constructions are
taken into account here. The third part of the chapter is concerned
with the linguistic encoding that these languages employ to mark the
experiencer nominals. The fourth part of the chapter is an important
part as it traverses through the theoretical issues like what is subject,
what is accusative, what is dative, etc. The issue of identifying
experiential NP with the ‘subject’ of the sentence is discussed both at
syntactic level incorporating such as verb agreement, conjunction
reduction, anaphoric control, position of the nominal, topicality and
focus, as well as at semantic level involving issues like ‘locational
lypothesis’, “point of view hypothesis’ and ‘passive hypothesis’. The
judgement of ‘subjectivity’ of experiencer NP is reached only after
testing the experiencer NP at these levels.
For the sake of initiating the discussion I will be taking Hindi
examples as base. Since the identification and specification of
semantic notions and roles have universal application and relevance,
taking one specific language as a base to explain the phenomenon of
experiencer nouns, or experiential verbs, or non experiential con
structions neither over-generalizes nor under-estimates the
phenomenon of Obl. m. Subj.

I Non Experiential Constructions


Benefactive
Benefactive verbs such as ‘be’, ‘become’ ‘have ‘own’ in Indian
languages take oblique marked subjects. There can be two types of
benefactive verbs, a State Benefactive verb and a Process Benefactive
verb. A state benefactive verb (SB) specifies that a benefactor is in a
certain state or condition with respect to a given object. It is
accompanied by a benefactive noun which specifies the possessor of
60 Semantic Universals in Indian Languages

the object, and a patient noun which specifies the object possessed.
Let us consider the ‘possessive constructions’.

Possessive

A nominal element can be in possession of alienable and inalienable


entities. With respect to these possessions he is in a certain ‘state’
which he has reached into with or without his will. Many of the South
Asian languages mark the ‘possessor NP’ distinctly from the ‘non
possessor’ or ‘agnetive NP’.
Consider:

Inalienable
Meithei ram-gi makut-ama lay
ram-Gen. hand one be
‘Ram has a hand’ (SKC)
Marathi tya-la: don hat a:het
he - Dat hands be
‘He has two hands’
(Lit: To him two hands are) (RP)
Kannada avani-ge eradu kannive
he-Dat two eyes
‘he has two eyes’ (SNS)
Hindi mere do hath hain
I-Gen. two hands be
‘I have two hands’

Alienable -
Kharia hokra?-te moloy kunru aij-ki
he-Dat/Acc five children be-
‘He has five children’ (VM)
Meithei ram-gi famli-ama loy
Ram-Gen. family one be
‘Ram has a family’ (SKC)
Malayalam kuttikka dha:ra:lam panam unta
child-Dat _ plenty money have-pres.
‘the child has plenty of money’ (K.P.M)
Identity Crisis of Dative Subjects 61

Marathi 8. tala: khu:p kapde a:het


he-Dat many clothes be
‘he has many clothes’
(Lit: Many clothes are to him) (RP)
H. Kannada 9. __yen-ge jamin ille
I-Dat. land not (be)
‘I do not have land’ (HU)
Kashmiri 10. — tem-is ehu setha pas (RB)
he-Dat be much money
‘He has lots of money’
Bhojpuri 11. hamra: ego laika: ba:
I-Gen: one son be
‘I have a son’ (MKV)
Nahali 12. bidi mancho-ke_ ir lana tha
one man-to two sons were
‘A man had two sons’. (FBJK)

As we can obsérve all possessive constructions are translatable by


‘have’ constructions in English. Indian languages make use of dative
or genetive marking for inalienable as well as alienable possessions.
The Tibeto Burman branch of languages treat ‘family’ and its
members as inalienable possessions. In all the sentences the possessor
NP which is in initial position is in a state of possessing the entitites,
hence, literally, the ‘possessed NP is to the possessor NP’.
In addition to possessor NPs the languages of South Asia mark
benefactors NPs of a process benefactive (PB) verbs with dative or
accusative. Consider:
Malayalam (13) baalana pustakam kitti/ labhiccu
boy Dat book Nom receive-past/gain
past
‘The boy received/gained the book’
(14) baalana katto wannu
boy-Dat __letter-Nom come past
‘The boy got a letter’
(Lit: A letter came to the boy)
62 Semantic Universals in Indian Languages

Hindi (15) tum-he kya mila


you-Dat what got-past
‘What did you get?’
Panjabi (16) on-nu sukki roti milli:
he-Dat dry bread receive
‘He received (only) dry bread’
The other kind of process benefactive constructions marked by
Dative Subjects in most of the languages is that with ‘progeny’ verbs.
The verbs for ‘to be born to’, gave birth to’ etc. take Dative NPs.

Malayalam (17) meerikka aankutti janiccu


Mary-Dat. boy-child Nom be born-past
‘Mary had a baby boy’
(Lit: To Mary a baby boy was born)
Hindi (18) neera ko/ke do bete hue
Neera Dat/Gen. two sons were
“Two sons were born to Neera’
The process benefactive verb, thus specifies that a benefactor
undergoes a change of state or condition with respect to a given
object.

Attributive

A nominal element can have certain inherent as well as exterior


qualities which further define the element. If an animate noun is ina
‘state’ of attribution, intrinsic or extrinsic, then that noun is marked
by either dative or locative in Indian languages.
Consider:
Maithili (19) okra: sa:has cahi
he-Dat courage be
‘He has courage’ (MKM)
Marathi (20) — tya:la: himmat —a:he
he-Dat courage is
‘He has courage’
Identity Crisis of Dative Subjects 63

H. Kannada (21) awani-ge _ tilawalike ille (HU)


he-Dat sense not (be)
“He has no sense’
Nahali (2)\eebhate etere-ke aka pati do etere kainiy.
then him-Dat sense come and he said
“Then sense came to him and he said:

Some Indian languages do capture the inchoativeness with regard to


animate nouns. For instance the following sentence from Malayalam
puts the affected nominal in the dative case with the conjunctive verb
beautiful + keep (Adj + V). Such constructions may be considered
of transitory state: Indo Aryan languages, however, do not allow
obl-m: Subj-with transitory state verbs.
(23) kuttika bhangngi weccu
child-Dat beautiful put/keep-past
‘The child become beautiful’
‘(the child is in a_ state of becoming
beautiful) (KPM)

Attribution can be acquired by inanimate objects also. In such a case


the inanimate NP takes the Dative/Genitive and it is in a state of an
‘acquired quality’.
Marathi 24. ghara-ta _—_don da:ra _azhet
house-Dat two doors be
The house has two doors
(Lit. To the house two doors are)
25. = zha:da:-la__ phula a:het
tree-Dat flowers are
‘The tree has flowers’

Malayalam 26. cetikka kompo mulaccu


plant-Dat branch sprout past
‘The plant sprouted branches’

(Lit: To the plant, branches sprouted)


64 Semantic Universals in Indian Languages

Most of the other languages use locative or genitive for such


inamimate attribution:
Hindi 27. per mein naye potte aye hain
tree-loc new leaves come be
“The tree has new leaves’
Panjabi 28. ais ped ca nove
nove phull aye ne
This __tree-loc new
new _ flowers come
past
New flowers have blossomed on this tree.
Hindi 29.~ kursi ki car ta:ngé hoti hain
chairGen fourlegs being are
‘A chair has four legs’.
Semantically, Malayalam and Marathi dative serve the same
function as Hindi and Panjabi locatives and Genitives do.

Out of Control
So far we had been considering the constructions where a nominal
element is in a certain ‘state’, where he does not perform an action
but an action is performed on him or say, the ‘effects’ of an action are
directed to him. Now we would take into account those cons-
tructions where an action is being performed by a nominal element
on which he has no control whatsoever. The constructions are
passive syntactically and semantically. The NP under consideration
takes typically an instrumental marking. Consider:
Hindi 30. neha-se botoal tuut gal
Neha-Instr bottle break-go past.
Bottle was broken by Neha (inadvertently)
31. papa-se gacri chuut gail
papa-Inst. train leave go-past
Papa missed the train.
(Lit: The train was missed by papa)
The Incapabilitative Passives involving ‘intransitive’ verbs will also
come under ‘out of control’ constructions as these also take the NP in
Identity Crisis of Dative Subjects 65

oblique case. Consider:

Hindi 32. ham-se cola nghin jaataa


We-Inst. walk Neg go
“We cannot walk’
(Lit: Walking cannot be done by us)
These NPs which are marked by dative, locative, genitive or instru
mental, as suggested earlier, can be referred to as oblique marked
NPs. The oblique marked NPs in non experiential constructions can
thus occur in (1) Attributive phrases, (2) Benefactive phrases and, (3)
in ‘Out of Control’ phrases. We just examined that attributive
phrases could be of (+ animate) or (—animate) nominals.
The benefactive phrases could be those involving SB verbs or non
SB verbs i.e. PB verbs. The former included possessive constructions
of alienable and inalienable possessed nominals. The Process
Benefactive verb specifies that a benefactor undergoes a change of
state or condition with respect to a given object. It is accompanied by
a benefactive noun which specifies the one who undergoes gain or
loss, and a patient noun which specifies the object which is gained or
lost. The constructions also include verbs of ‘progeny’. The ‘Out of
Control’ phrases include typical passive constructions and those
which indicate that an action may take place without any instigation
or initiation of the doer of the action. In fact, the event becomes out
of the control of the actor (if we can still call him the ‘actor’. He is
more like a patient or ‘affected’ nominal). The following illustration
gives a bird’s eye view of the typology of the Non Experiential
Constructions with oblique marked NPs.
66 Semantic Universals in Indian Languages

Non Experiential

Attributive Out of Control

Possessive Progenetic

[+ Alienable]

[+state]v [—state]v {—action]v

[+state]v [+Transitory state]v

[+ animate]

Fig. 2. Non Experiential Constractions


Identity Crisis of Dative Subjects 67

Now the question is should we or should we not consider obl. m. NPs


in these constructions as subjects. We would like to wait for the
answer till we look into the experiential constructions in great
detail.

II. Experiential Constructions


In this part of the essay we will be concerned with (1) the semantic
nature of the so called experiential constructions; (2) the specific
linguistic encoding that is employed by the languages of South Asia
to represent experiential constructions and finally, (3) the areal
spread of such encoding devices. We will begin with considering
sentences like 33-38 from Hindi :

Hindi (33) mujhe dukh he


I-DAT grief is
(33a) meé dukhi hu
I-NOM sad be
‘I am sad’
(34) mé bimar hi
I-NOM sick be+ 1 sg.
‘T am sick’
(35) mujhe khansi/ zukam_ /bukhar dard he
1+ DAT cough cold fever pain be
‘Thave cough/ cold/ _fever/ pain’

(36) bacce ko khelna pasand he


child-DAT _ play-inf like be
‘Child likes to play’
(37) priya ko kitab cahiye
priya-DAT book want
‘Priya wants a book‘
(38) papako _ sigret pine ki adat he
papa-Dat cigarette drink-gen. habit be
‘Papa is in the habit of smoking/Papa smokes’
In all these sentences the nominals ‘I’, ‘bocce’, ‘Priya’, and
‘Papa’ are neither the instigators of the action nor those who did
something. Rather, the animate being symbolized by these nominals
68 Semantic Universals in Indian Languages

are mentally disposed in some way. Grammatically, sentences (33)


and (34) are instances of predicative adjectives signifying the state of
the nominals; sentences (35) (36) and (38) make use of conjunct verbs
(noun+ verb ‘to be’) and sentence (37) has the verb ‘want’. All these
verbs are experiential verbs which signify a state. Each verb is
accompanied by a typical animate being who is in the state of
experiencing which we would call experiencer. Verbs illustrated in
(33)-(38) are State Experiential Verbs (SE). An SE verb specifies that
an experiencer is in a certain state of condition with/without respect
to a given object. It is accompanied by an Experiencer noun which
specifies the one who is in the psychological state of sensation,
emotion, cognition. It may also be accompanied by a patient noun
which specifies the stimulus for or the content of the experience
(given in bold italics). The following sentence (39) provides for a
patient noun which acts as a stimulus for the experience of the
experiencer noun. Or the patient noun itself could be content of the
experience (40-42).

Hindi (39) mujhe comeliki sugondh ai


I-DAT jasmine Gen. fragrance come past
‘I could smell the fragrance of jasmine’
Bhojpuri(40) = hamra sardi ho gel ba
I-DAT/GEN cold be go _ be
‘I have cold’
Tibetan (41) nga chamba/ la /tsawa du
I-NOM_ cold/cough/ fever be
‘I have cold/cough/fever’
Marathi(42) mala bhukh lag |i
I+ DAT hunger strike past
(42a) mi bhukh le li
I-NOM hunger past part.
‘I am hungry’.
Experiential constructions may also involve verbs which are Process
Experiential Verbs(PE). A Process Experiential Verb specifies that an
experiencer undergoes a change of state with respect to a given
object. It is accompanied by an experiencer noun which specifies the
one who undergoes the change of psychological state, and a patient
Identity Crisis of Dative Subjects 69

noun (given in bold) which specifies the stimulus for or the content of
the experience. Consider the following examples (43) and (44).
Hindi (43) mujhe dhol ki avaz sunai de rahi he
1+DAT drum GEN sound hear give-prog
‘I can hear the sound of the drum’
Hindi (44) mujhe thakanho roahi_ he
I+ DAT exhaustion be prog
‘I am feeling tired/exhausted’.

(44a) main thok raha hu


I-NOM tire prog. be
‘I am feeling tired’.
(The experiencer nouns are in italics, the patient nouns which could
either be stimulus or the content of the experience are in bold letters.)
It is the conjunct verbs such as sunai de ‘be heard’, thakan ho ‘to get
tired’, bukh lag ‘to get hungry’ — which are semantically specified
as Process Experientials and State Experientials.
State Experiential Verbs and Process Experiential Verbs are the
only true experiential verbs. A basic Action Experiential Verb
cannot exist in any language since the notion of experiential verb
itself precludes the possibility of agency or instigation. Our obser
vation across South Asian languages, however, show that one can get
action/action process experiential verbs which are derived by the
application of a resultative unit to a typical action or action process
verb. For instance, Hindi baith - ‘to sit’ is a typical action process.
verb which involves an animate being. The enactment of the verb
vaith can be followed by a prolonged state which is the direct
outcome of the completion of the verb baith and the agent of the
action baith itself could be the experiencer of the resultant state.
Morphologically, this could be derived by a resultative unit applied
to action/action process verbs to give us constructions of -e hue type.
Thus, if a typical action process verb is converted into a state
incohative and if an animate being finds itself in this state then we can
expect to have an experiential construction involving the experiencer
noun which happened to be the agent of the action too. Considser the
sentences (45), (46), and (47).
70 Semantic Universals in Indian Languages

Marathi(45) mala Ithe bosun don tas zhale


1+DAT here sit+ past two hours over/happened
Hindi (46) mujhe yeha baithe hue do ghonte ho gaye
I+DAT here sit + past part. two hours be were
Munda (47) _aiya ke ente du:ra meyd ghonta howai na
I+DAT here sitting one hour be + past
(45, 46, 47) ‘I have been sitting here for two/one hour(s)’

Marathi, Hindi and Mundari examples indicate that ‘I’ had been
in the state of experiencing the trauma of sitting for two hours. “To
sit’ is an activity which has its agent coreferent to the experiencer of
the state of ‘sitting’. By all méans, (45), (46), and (47) are as much
experiential constructions as (33) - (44) have been. We can thus define
a Stative Action Process Experiential (SAPE) Verb as that which
_ specifies that an experiencer is in a certain state or condition with
respect to an action undertaken by himself. In this respect it is always
reflexive. Almost all Indo-Aryan languages have SAPE
constructions.
To review, either we can have State Experiential verbs, or Process
Experiential verbs or Stative Action Process verbs. The only
difference between the first and the last one is that SE verbs are basic
verbs while the SAPE verbs become stative after being derived so, by
resultative unit. In both cases, the verbs describe the experiential
state of the nouns under consideration. Whether a verb is State
Experiential or Process Experiential would be decided by the
semantic nature of the verb itself. Thus Hindi bhukh ho ‘be hungry’ is
state experiential while bhukh lag ‘get hungry’ is process
experiential verb. Interestingly, both types can take dative
experiencer nominals.
An Experiential Construction is then a construction which has an
experiencer verb accompanied by one experiencer noun and one or
zero patient noun. These constructions express our experience of the
phenomenon of the ‘real world’ as well as those of the ‘inner world’
of our consciousness.
Identity Crisis of Dative Subjects 71

Ill
Semantic Typology
Let us consider the various types of experiential constructions
available in South Asian languages. My attempt has been to have as
clear cut division as possible but some overlapping seems to be
inevitable. The divisions between the various types have been made
solely on the semantico-cognitive parameters. Some of the
parameters, however, have already been discussed by Klaiman
(1986) earlier.
1. Sensory and Mental Experiences
2. Physical and Biological Experiences
3. Emotional Experiences
4. Compulsive/Obligattonal Experiences
5. Temporal Situational Experiences
The list is neither exhaustive nor very accurate, because what may be
considered an emotional experience may also be considered a type of
sensory and mental experience, or what may be considered a
temporal situational experience may be part of emotional or physical
experience. Nonetheless, I am going to adhere to these divisions to
account for the data given here. It is a well known fact that minute
hairline distinctions between sensory experiences are difficult to
make even for psychologists. Linguists can do only a partial justice
based on the analyses of linguistic structuration.

1. Sensory and Mental Experiences :


Verbs indicationg internalized desire, want and need such as verbs
for ‘hunger’, ‘thirst’, ‘love’, ‘like’, ‘to be sleepy’ etc. come under this
category. Consider sentences (37), (44), and (48) - (53).

Munda (48) aiya-ke pita poseind mena


I+DAT bread _ like be
‘T like bread’
Mizo (49) avil otam
he hungry
Kannada (50) avanige hasIvagide
he + DAT hunger-strike
72 Semantic Universals in Indian Languages

Khasi (S1) u thongon


he-NOM hungry

Kashmiri (52) suris cha boch _ lejmets


child-DAT aux hunger got
Newari (53) ji-ta pitya-ta
I-DAT hungry aux
(48, 49, 50, 51, 52, 53) *X is hungry’

Under this category we would also like to consider ‘belief’, and


‘claim’ constructions. The experiencer nominal is in genitive form.
Consider:
Hindi (54) nira ka vicar /irada aj parhne ka nahin hai
Nira Gen. thinking/intention today to read neg aux.
‘Nira does not think of/intend to read today.’
(55) pradhanmantrika davahai ki ve
The Prime minister GenclaimAux that he
kimat6 mé komi layé ge
prices in decrease’ will bring
“The prime minister claims that he will bring down the
prices’.

II. Physical and Biological Experiences


Verbs indicating physical ailments and biological experiences such as
cold, cough, wound, pain, heart attack, ‘to get cold’ ‘be hot’, etc.
Consider constructions such as (34), (35), (40), (41), (44), (56), (57),
and (59)
Munda (56) aiya ke bubumena
I+DAT fever be
‘I have fever’
(57) oiyad ke a:du_ lagao tora
I+DAT urinate strike be
‘I feel urinating’
Kannada (58) avanige tumba_ kalleaygide
he + DAT much - ill is
‘He is very ill’
_ Identity Crisis of Dative Subjects 73

Newari (59) ji-ta casu (li) le


I-DAT itch become
‘I am itching’.

Ill. Emotional Experiences


Feeling verbs, the stimulus for which is outside the experiencer such
as verbs for ‘worry’, ‘expectations’, ‘happiness’, ‘sorrow’, ‘know’, ‘to
feel shy’, ‘to remember’, ‘to be angry’, ‘to tire’, ‘to get an idea’, ‘hear’,
‘see’, etc. Sentences like (33), (60)-(68):

Kannada (60) nanage nidde _ barutide


I1+DAT sleep _ getting

Marwari (61) mhane ni:d a:ve


I acc/DAT sleep comes.
‘I am sleepy’

Kannada (62) nanago nacike agutt ode


I+ DAT shyness come be
‘I feel shy’

Khasi (63) nga snou soemthya


I+ NOM feel sleepy
Meithei (64) ai tumnige
I+NOM - sleep be
(63, 64) ‘I am sleepy’
Bangla (67) or lojja hocce
she+DAT shy be
‘She is shy’

Nahali (68) etere khijia


he+NOM got angry
‘He got angry’

IV. Compulsive/Obligational Experiences


Verbs which indicate obligation or compulsion on the part of the
prospective doer or forced doer of the action. Such anticipatory agents
by virtue of being in obligational or compulsive situations can be
identified as experiencers. Such experiencers are in a state of
74 Semantic Universals in Indian Languages

obligation or compulsive experience. Constructions such as (69)-79)


exemplify such experiences.
Hindi (69) tumko ana hi hoga
you DAT come emph. be-fut.
“You will have to come’
Khasi (70) nga dei bon lei sha_ klas
INOM have to go _ to _ class
‘I have to go to the class’
Mizo (71) khombi _ thla kha khangaethla ngeingei toov
certainly song that I+ NOM listen must fut.
‘That is certain that I must listen to that song’.

Hindi (72) ram-ko mithai kha:ni pori/hogi


Ram DAT sweets eat (f) had to/will have to
Ram had to/will have to eat the sweet
Munda (73) amake aiya ora hiju ketona
you + DAT I house come have to
“You have to come to my house’.

Nepali (74) ma-lai ghar janu-cha


I-DAT home go aux
‘I have to go home’
Sinhala (75) ma-ta yana una
I-DAT go _ forced
‘I was forced to go’
Malayalam (76) (a) eni-kku pokonam
I-DAT go have to
‘I have to go’
(b) nyaan pokonem
‘I-NOM should go’
Kannada (77a) nana ge hoga beka gi bantu
I-DAT go compelled
‘I was compelled to go’
(77b) nanu hoga beku.
‘I-NOM want to/must/should go’
Identity Crisis of Dative Subjects 75

Kashmiri (78) temis pazi gasun van’


he-DAT should go now
‘He should go now’
Marathi (79) (a) ram-ni te kam koraylac hawe
Ram-ERG that work do must
(b) ram-la te kam karayla lagel
Ram-DAT that work to do will have to

V. Temporal Situational Experiences :


Constructions that involve either stative action process experiential
verbs or any other verb which puts its patient in a temporally situated
experiences such as ‘be in the habit of. Such verbs are state
experiencer verbs which signify the mental state of an experiencer
with respect to the temporal aspect of the action he is involved in.
Consider sentences such as (38), (45), (47), (80), and (81).
Tamil (80) nan Inge irardmoni nerma:ha:irikkiren
I have two hours past stay -be-cont.
Malayalam (81) nan ivre rondmoni kur irikkina
I here two hours stay
(80,81) ‘I have been sitting here for last two hours’

These five types of experiential constructions are basically


sub-types of the following basic types of experiential verbs.

1. State Experiential Verbs,


2. Process Experiential Verbs,
3. Stative Action Process Experiential Verbs.

Each of the experiential constructions can be formed with either


state or process experiential verb depending upon the basic nature of
the finite verb of the sentence. Only the fifth type, i.e. the Temporal
Situational one can be generated by the Stative Action Process
Experiential Verbs.

IV
The Linguistic Encoding
Having identified the experiential verbs and experiential con-
structions semantically, let us now consider what linguistic encoding
76 Semantic Universals in Indian Languages

does the experiencer nominal assume. The general patterning of


linguistic encoding of the experiencer nouns falls under:
(a) Languages that have direct forms,
(b) Languages that have indirect/oblique forms,
(c) Languages that have both direct and oblique forms which
alternate freely in all constructions and in all situations,
(d) Languages that offer both direct and oblique forms but
alternate them in very restricted occurrences.
Mizo, Meithei, Tibetan, all from Tibeto Burman group, fall under
the category (a). The experiencer nouns do not take any special
marking in any of the experiential constructions discussed above.
Khasi shares this feature with the Tibeto-Burman group.
There are some languages which have only oblique marking (b) for
experiencer nouns. In other words, not all experiencer constructions
in all the languages induce dative/oblique marking. Languages under
category (c) are hard to find as no such language which offers
direct-oblique alternation allows this variation in all types of
experiential constructions.
Most of the South Asian languages fall under the category (d).
Languages under this category offer parallel constructions of the
type (33) and (33a) where one construction has an experiencer noun
with oblique case (generally Dative, uncommon is Genitive), while
the other construction does not have oblique case marking with
experiencer nouns. Also (42), (42a), (44), (44a), (82-82a) are worth
considering :
Hindi (82) vo bahor ane se sharma rahi thi
she out come-inf-pp shy(v) prog. was(fem.)
(82a) use bahar ane mé shorm a rohi thi
she +DAT out coming pp shy-come (CV) prog. was
‘She was feeling shy of coming out’.

The English translation of sentences 82 and 82a is the same. The


question is are these variants of a single semantic construct or
represent two distinct semantic constructs (which unfortunately
cannot be translated in English)?
In a study basically based on Bengali, Klaiman (1986) claims that
‘Bengali speakers select the dative subject counterpart when an
Identity Crisis of Dative Subjects 77

activity is to be spoken of a NONVOLITIONAL, while the direct subject


counterpart is selected when the activity is to be spoken of as
VOLITIONAL’ (1986 :182).
The discussion that is going to follow now will be devoted to this
claim. Klaiman’s argument that direct/oblique counterpart construc
tions are decied by VOLITIONALITY parameter does not hold for
many Indian languages including Bengali. In constructions such as
(33-33a), (42-42a), (44-44a) and (82-82a) both direct experiencers
and dative experiencers are involved in NONVOLITIONAL activity. In
fact, Hindi sentences (82-82a) have the same stimulus bahar ana ‘to
come out’ and the activity involved, i.e. sharmana ‘to feel shy’ is
non-volitional in both direct experiencer construction (82) and
dative experiencer construction (82a). We must remember that cases
where such direct/oblique parallelism is not available and
constructions have only direct case for experiencer nouns (such as in
languages like Khasi, Mizo, Meithei and Tibetan) nonvolitional acts
are anyhow represented by non-dative experiencer nouns.
Further, experiencer constructions of Stative-Action-Process
discussed above take dative though the activity involved is volitional.
Consider again sentences (45), (46), (47), (80), and (81) from Hindi,
Marathi, Munda, Tamil and Malayalam. A closer and deeper
analysis of these constructions indicate that though the activity
involved might be voluntary, the resultattve state that the
experiencer is in might not necessarily be voluntarily acquired. In
such a case these types can be claimed to be ‘non-volitional’. One of
the arguments that Klaiman proposes (1987: 183) is that with
adverbs of volitional only direct counterparts can co-occur : dative
counterparts are excluded. I do not find such a situation to exist in
Hindi and in many other IA languages. Consider sentence (83) which
is an exact translation of her ungrammatical Bengali sentence.
Adverb of volitional dekh kar does not render dative experiencer in
ungrammatical bag.
Hindi (83) amriki filmé dekh ker bacco ko
American films see CP. boys-DAT
anek golot dharnaé ho jati hé
many wrong conceptions become go aux
78 Semantic Universals in Indian Languages

‘Boys get a lot of misconceptions from watching


American films’.
Lastly, Bengali examples such as (84-84a), (85-85a) and many others
like these indicate nonvolitionality on the part of the experiencer in
spite of the fact that counterpart (-a) sentences are in direct case.

Bengali (84) amar lojja holo


I+GEN shyness become
(84a) ami lojja pelam/lojjito holam
I+NOM shyness got/shy became
‘T felt shy’
(85) amar kshudha palya che/khide peyi che
I+GEN hunger get be/hunger get be

(85a) ami kshudharto


I+NOM hungry-am
‘I am hungry’
What perhaps one can say that all dative marked experiential
constructions are non volitional but all non oblique or direct case
constructions are not volitional. The difference thus cannot be
ascribed only to the parameter volitional-nonvolitional. One basic
difference between the direct form and its parallel oblique form
seems to be that of FOCUS. In Dative experiential constructions the
focus is on the patient (grammatical subject) nominal, while in direct
form construction the focus is on the experiencer nominal.
Compulsive/obligational constructions always have dative
experiencer nouns in IA, and Munda. Dravidian languages and
Marathi offer variation in this respect keeping ‘should’/‘must’
constructions distinct from ‘have to’ ones. The non dative marking is
with the obligational constructions while the ‘have to’ types which
are of compulsive nature take dative markings. Tibeto Burman langu
ages and Khasi, it appears, do not use any kind of oblique marking
for any of the experiential constructions. Consider sentences
(86-86a) in Mizo and (70) in Khasi.

Mizo (86) ka tho


I-got up
Identity Crisis of Dative Subjects 79

(86a) ka tho angae


I get up have to

Marathi, it appears, to be substratumized by Dravidian in this


respect (see 79a and 79b).
To sum up:
(1) Not all experiential constructions are marked obliquely in all
Indian languages.
(2) Languages which do allow oblique marking with experiencer
nominals may also allow parallel direct forms with experiencers.
(3) Some experiential constructions with oblique nominal marking
have no parallel direct forms. This is true mainly for experiential type
II and type IV discussed above.
(4) The distinction between direct and oblique forms of exper-
iencers in languages which allow parallels is not based upon
volitional factor but on the parameter of FOCUS. In the former, the
focus is on the experiencer nominal while in the latter, it is on the
patient noun.
(5) Most of the Indo-Aryan languages take conjuct verbs with
oblique marked experiencers.
(6) In oblique marked construction verb does not agree with the
experiencer but with the patient noun or with the right most noun in
the sentence. In non oblique marked construction verb agrees with
the experiencer noun. This statement implies that verbs agree with
the focussed nominal in the sentence.
(7) Typologically Indo-Aryan, Munda and Dravidian languages
more or less share the patterning of direct and oblique experiencer
noun marking. However, in type V, for stative action process experi-
ential verbs, Dravidian languages do not make use of the oblique
(Dative) forms. Consider (80) and (81) again. However, habit
marking sentences are marked by oblique case even in Dravidian.

Consider:
Kannada (87) appanige sigretu sedua sbhyasavida
father + DAT cigarette smoking habit is
‘Father is in the habit of smoking’.
80 Semantic Universals in Indian Languages

V
Dative or Accusative or any other - ative ?
Now we come to the crucial question whether nominals of non
experiential and experiential constructions are ‘subject’ or not.
Should one call these nominals ‘dative subjects’ at all? Are these
nominals subjects? The answer to these questions are important but
not easy as obl.m. nominal constructions in Indian languages expose
a new semantic analysis of verbs as well as raise issues in the realm of
syntax as to of identifying the category ‘subject’. We have already
analysed the semantic nature of the experiential and non-experiential
verbs (Benefactive, Attributive and Out of Control). We must now
look into the constructions from the point of view of seeking answer
to the question of identity of the oblique marked nominals. For this,
we would like to consider the phonomenon of ‘subject’ both at
semantic and syntactic axies. We will begin with the argument
whether it is justifiable to term these constructions as Dative
at all?
The nominal under question in many Indian languages is encoded by
accusative and in others there is no overt difference between Accusa-
tive and Dative markings. How do we decide what is which? The
problem is not a new one as the question of the distinctness between
the two cases had disturbed Otto Jespersen to remark, ‘It is, however,
impossible to keep these two things apart, at any rate in the best
known languages’ ‘(1968: 185). He opines that word order is often
sufficient to indicate the case in a sentence, ‘a case is a purely
grammatical (syntactic) category and not a notional one in the true
sense of the word’ (1968: 185). However when it comes to distinguish
between the Dative and the Accusative his own suggestion of word
order hierarchy crumbles down as he puts it in the following
statement:
... It is impossible to recognize positional dative in ‘I gave it him’
for we have the inverse order ... if we were to speak of separate
datives and accusatives in English, I for one do not know where in
the list the dative goes out and the accusative comes in, and I find
no guidance in those grammars that speak of these two cases ... we
Identity Crisis of Dative Subjects 81

may therefore safely assert that there is no separate dative and no


separate accusative in modern English (ibid : 174).

What Jespersen said of Modern English is true for Modern Indian


languages too. Hindi-ko, Punjabi -nu, Telugu -ku, Kannada -ge,
Munda -ke encodings can serve both as accusative or dative case
markings. Dativity can also not be decided by the criterion of
what can be promoted to the subject position. In sentences like
‘she was awarded a gold medal’ or ‘he was refused admittance’,
etc. she and he are promoted datives. But in parallel sentences
such as ‘a gold medal was awarded to her’ and ‘admittance was
refused to him’ the accusatives (italicized) are promoted to the
subject position. Jespersen further suggested that Dative should not
be equated with indirect object as the frequency of indirect object as
being dative is very low in any given text.
Other than the syntactic criteria, at the sementic level ‘the mean-
ings of the accusative and dative cannot be kept strictly distinct, is
shown also by the fact that the same verb may in some languages take
sometimes One case and sometimes the other. The object after Old
English onfan “take, receive” is now in the accusative, now in the
dative, and now in the genitive’ (ibid. 179). This case is just like what
we have seen with experiential constructions where Hindi uses
dative/ accusative and Bengali uses genitive. Jespersen in his famous
grammar criticises Professor Sonnenschein’s view that ‘a dative is
found only when the verb has also another object (which then is said
to be in the accusative)’ (ibid. 179).
Jespersen was aware of languages which marked distinctly the
datives and accusatives for he commented, ‘in languages with sepa-
rate forms for the accusative and the dative, the person is put in the
dative and the thing in the accusative, the former is called the
indirect, and the latter the direct object. But sometimes we find the
dative when there is only one object, and in some cases, both objects
are in accusative—which shows that the difference between the
dative and the accusative is not a notional one, but purely syntactic,
dependent in each language on idiomatic rules’ (ibid. 162). He
further asserts that ‘it is customary to speak of two classes of cases,
grammatical cases (nom, acc ... etc.) and concrete, chiefly local cases
(locative, ablative, sociative, instrumental, etc.)’. Wndt, in much the
82 Semantic Universals in Indian Languages

same sense, distinguishes between cases of inner determination and


cases of outer determination and Deutrchbein between ‘Kasus des
begriffichen denkens’ and ‘Kasus der anchuung’ (ibid. 185).
The confusion between dative and accusative is further compoun-
ded in Indian languages where we face two pronged problems.
1. Many languages have identical encoding in dative and
accusative.
2. Not all languages use only dative or accusative markings on the
nominals under consideration. Genitive, and instrumental are also
used.
The most appropriate solution seems to be to identify one
‘common’ semantic feature underlying all these constructions. We
have just seen that expriential and non-experiential constructions in
various Indian languages, signify the ‘patient’ nature of the oblique
marked nominals. In the experiential constructions all the nominals
are ‘experiencers’ who undergo an experience, while in non
experiential constructions the nominals under consideration are
passive patient of the various states or processes. This passivity is a
dominating factor in both experiential and non-experiential
constructions. None of the nominals in the subject position are agent
of an action or agent of a resultant situation (this can be also true of
Temporal Situational Constructions). The varying states and
processes be of experiential type, or of benefactive type, or of
attributive type; or the actions of the involuntary, inadvertent
types—all put their affected nominals in some kind of ‘patient’ or say
‘affected’ category. The nominals under discussion are definitely
non-Agentives. “Non agentive’ thus would refer to both experiencer
and non experiencer nominals which are not involved in any
‘action-oriented activity’.
The question now is whether these non agentive nominals can be
considered as subjects ? Or should they be considered objects, or
locatives (as some linguists would prefer to call) ? This calls for a
detailed analysis for what is subject ? What are the criteria to
determine subjectivity ? Are they syntactic, or semantic, or both? Are
there pragmatic or discourse level criteria too ? These are tough
questions nevertheless we would make an attempt to answer all these
questions in the following sections. I would divide the discussion in
Identity Crisis of Dative Subjects 83

three parts (a) to define what is subject or object or predicate; (b)


syntactic criteria to decide on subjectivity. Under this I would like to
discuss the issues of verb agreement, anaphoric control, conjunction
reduction, positional situation, and topicality and; (c) the semantic
and pragmatic criteria which may be concerned with issues like
‘locational hypothesis’. ‘point of view’ and ‘passivity hypothesis’.
What is Subject ?
The definition of subject has been discussed, analysed and dissected
most in the linguistic literature. There are as many as 123 definitions
of the subject ‘subject; starting from ancient grammarians to the
most modern. We possibly cannot discuss each and every definition,
however, we will make an attempt to investigate into two sets of
notional opposites which are relevant for the constructions under
consideration. That is, the distinction between Subject and Predicate
and Logical and Psychological Subject. Jespersen (1928) discusses
mainly three kinds of views proposed then about the subject and the
predicate of a sentence : (i) Subject is old information and predicate
is new information (this is a commonly shared view of the modern
semanticists), (ii) the role of the predicate is to specify or determine
what was at the outset indefinite and indeterminate. That the subject
is thus a determinandum which only by means of the predicate
becomes a determinantum; (iii) Subject is what you talk about and
the predicate is what is said about this subject. Con sidering these
three views it could be proposed that non agentive constructions
have oblique marked nominals as subjects and non oblique marked
nominals as predicates since the former could be considered old
information and the latter as new information, However, in ‘Out of
Control’ sentences and ‘Biological and Physical experiences’
sentences it might be difficult to ascertain which is which. If we take
discourse as a target of analysis then problem is somewhat solved as
the confusion between what is old and what is new will evaporate.
The second point also draws our attention to the fact that oblique
marked nominals are subjects because they serve as determinantum
in the said constructions. If we consider the third view it again points
to the non oblique marked nominals in the predicate position and
oblique marked nominals in the subject position because we ‘talk
84 Semantic Universals in Indian Languages

about’ the latter and the said statement is about the oblique marked
nominal.
The best indicator of the state of cofusion and indeterminacy that
prevail in defining subject is seen in the quotes from a noted
psychologist Stout who, in a famous passage (AP 2. 212 ff) starts
defining it but later takes us at a point which admittedly very far from
the conventional grammarian’s conception of subject and
predicate:

The predicate of a sentence is the determination of what was


previously indeterminate. The subject is the previous qualification
of the general topic to which the new qualification is attached. The
subject is that product of previous thinking which forms the
immediate basis and starting point of further development. The
further development is the predicate. Sentences are in the process
of thinking that steps are in the process of walking. The foot which
is moved forward in order to occupy new ground corresponds to
the predicate ... All answers to questions are, as such predicates,
and all predicates may be regarded as answers to possible
questions. If the statement ‘I am hungry’ bea reply to the question,
‘Who is hungry ?’ then ‘T’ is the predicate. If it be an answer to the
question, ‘Is there anything amiss with you?’ Then ‘hungry’ is the
predicate. If the question is ‘Are you hungry?’ then ‘am’ is the
predicate. Every fresh step in a train of thought may be regarded as
an answer to a question. The subject is, so to speak, the formation
of the question; the predication is the answer. (Jespersen
[Secondary Source] : 1968 : 146).

Stout’s views lead us nowhere. As far as his ‘further development’


theory is concerned we understand that sentences are built in steps
and additions to each step is an ‘add on’. Sentence (1) for instance,
makut-ama lay ‘hand-one-be’ is further development of ram-gi
“‘Ram- poss’. The problem would come only if we reverse the order of
words in these languages. Indian languages most often than not
allow OvS order for emphasis purpose. For example, sentence 14 in
Hindi can be rendered as kya mila tumhe ‘what did you get’. Insucha
case the further development theory would establish kya (obj) mila
(V) as subject and tumhe as predicate. In other words, the theory is
Identity Crisis of Dative Subjects 85

based on positional elements in a sentence. Any change in the word


order will change the emphasis and hence the notion of ‘subject’ and
‘predicate’, Stout’s second suggestion of question-answer strategy
also does not work satisfactorily as it heavily relies on the supraseg
mental markers of emphasis and the formation of the question
pattern in the language. In normal circumstances each sentence is
neither an answer to a probable question nor can it be rendered as an
answer to a real question. Also this would allow the same nominal in
the same sentence sometimes as subject and sometimes as predicate
depending upon to what question this is an answer to.
The other set which has been discussed in great detail in the
literature is the dichotomy between /ogical subject and psychological
subject. The former being decided more on the basis of coindexing of
verb morphology and the latter on the basis of ordering of words. It
has been assumed that what makes speaker think is placed first and is
‘psychological subject’ and what he thinks about is placed next is
‘psychological predicate’. Jespersen quotes Paul (Gr.3.12) in whose
views psychological subject is the idea or group of ideas that is first
presented in the mind of the speakers, and the psychological
predicate as what is joined to it. But he neutralizes this definition
when he adds that even if the subject-idea is the first in the mind of the
speaker, it is sometimes placed later because of pragmatic reasons. In
his former work (p. 283: source: Jespersen) he says that the
psychological predicate was the most mportant element, that which
it is the aim of the sentence to communicate and which therefore
carries the strongest tone. Jespersen also quotes Hoffding who
identified logical predicate with grammatical subject or an adjective
belonging to it. For him, logical predicate was equal to psychological
predicate. Bloomfield (1933) calls logical predicate as ‘the
emotionally dominant element’.
There were many generalized views about the notion of ‘subject’
also. For instace, some linguists considered that any primary word
(in case of English) in a sentence is subject. In such a case ‘Peter gives
a book to Paul has three subjects, i.e. ‘Peter’, ‘Paul’ and ‘book’.
Many grammarians use the term ‘logical subject’ for the demoted
nominal in a passive sentence. In such a case the ‘Out of control’
_ sentences which do not involve any promotion or demotion of NP will
86 Semantic Universals in Indian Languages

be left unclassified. Moreover, what do we do with those languages


(Meithei, Tibetan) which do not have grammatical passive
constructions; or what do we do with those constructions which do
not allow passive transformations such as experiential or benefactive
constructions that we have just seen in section I-IV. The entire debate
on the issue as well as confusing and varying views on the notion
‘subject’ by various linguists prompted Jespersen to say: ‘...one
should give up the notion like “logical’’, ‘“‘psychological’’, “formal”
etc. and retain subject/object notions in the context of grammatical
subject/object only’ (1968 : 150). He suggests three ways of identi
fying this grammatical subject:
(1) Grammatical subject is always primary.
(2) Subject occupies the sentence initial position (this is not true in
all cases).
(3) If the two are connected by is are equally indefinite in form, it
depends on the extension of each which is the subject, e.g. A cat
(Subj) is a mammal.
We shall soon see that it is not even easy to define what is gram
matical subject because in ergative languages such as Meitei (Tibeto
Burman) and in many languages of Indo Aryan stock grammatical
subject is decided by verb agreement morphology which is not
uniformally subject or object coindexing. For instance, in Hindi
progressive tense allows agreement with first nominal (‘grammatical
subject’) while perfective allows agreement with the object noun
(‘grammatical subject’).

Consider.
(88) ram roti kha raha hai
Ram (m) bread (f) eat mas. sg-prog-aux
Ram is eating the bread. (verb agrees with Ram)
(89) ram-ne roti khaaii
Ram-Erg. bread (f) eat-f.sg. perf.
Ram ate the bread. (verb agrees with ‘bread’)
Moreover, languages like Malayalam and Bangla and many others
do not show verb agreement. In that case how do we decide on the
“grammatical subject’ ?
Identity Crisis of Dative Subjects 87

On the basis of the issues discussed so far we can suggest that


oblique marked nominals in Indian languages may be considered
subject (without any further prefixed modifier like grammatical,
logical, psychological ... etc.) because (1) they serve as old
information; (2) they are determinadum and indefinite without the
following predicate; (3) they are being ‘talked about’ in the
construction; (4) they precede the further development of the
activities in a sentence to a large extent (if we disregard the change in
word order for stylistic reason) and, (5) they do work as
psychological subject primarily not because they occupy the sentence
initial position but because the nominal in question finds its
extension in the rest of the sentence. It seems we have solved the
problem ! Not yet. So long we do not judge or evaluate the oblique
marked nominals on the grounds suggested by various syntacticians
we cannot safely identify these nominals as subjects. In other words,
we must discuss the notion of ‘subject’ at the syntactic level and then
decide whether oblique marked nominals can be claimed to be
‘subjects’ at all.

The Syntactic Criteria


Prototypically subject is supposed to have a nominative case. The
problem arises when we take upon ourselves the task of defining
what I termed earlier in the paper as ‘non performative’
constructions. These constructions involve experiential and non
experiential nominals in some kind of passive state. We have to
decide whether this experiencer or patient which most of the time
takes non-nominative marking behaves like subject with nominative
marking or not. Out of the several tests suggested by various linguists
such as Perlmuttur and Postal (1983), Charles N. Li (1976); Comrie
and Keenan (1977), Givon (1976) to identify subject the most
popular and widely tested is the one proposed by N. Chomsky in his
Government and Binding (1981) theory. ‘Subject’ is treated as an
entity. involving syntactic prominence. It is the most prominent
element related to ‘superiority‘ of various kinds—such as agreement
controller, c-commander of the arguments, having structurally
prominent S. Structure position. It is the most prominent element in
the relationship of binding, PRO construal, and in constraining
88 Semantic Universals in Indian Languages

movement. This expanded notion of subject implies the ‘specified


subject condition’ and the ‘tensed-S condition’ which together seem
to constrain the NP movement operations in languages.”
Let us take the argument of ‘verb agreement’ first since this has
been discussed widely in literature. Before we really plunge into the
discussion we must make note of the fact that there is no uniform
pattern of verb agreement in Indian languages. Firstly, Indo Aryan
languages are not fully ergative. Hindi is ergative in perfective only.
While Marathi shows ergativity in ITI p. singular form only. Konkani
shows it in all pronouns. Secondly, Maithili, Bhojpuri, Bengali and
other Magadhan dialects are not syntactically ergative at all.
Thirdly, Dravidian languages are not ergative. Fourthly, the
ergativity of Munda and Tibeto Burman languages is also partial. In
this environment of lack of total agreement with the subject nominal
(which may be in a nominative case) one cannot afford to base the
entire argumentation of the identity of subject on verb agreement.
Nonetheless, as far as ergative languages (whether partial or
complete) are concerned, any postpositional marking on the subject
position nominal (the first nominal in a sentence in a regular word
order paradigm) blocks the verb agreement.
Consider : (Postpositional markings are in bold).
Hindi (90) ma ne kapre dhoye
mother ERG. clothes (m. pl.) work-past-m pl.
“Mother washed the clothes.’
(91) ma ko buxar a goya
mother DAT/ACC fever (m. sg.) come-went (m-sg).
‘Mother had faver’
(92) ma > kapre silne logi
mother clothes (m. pl.) stitch started f. sg.
‘the mother started to stitch clothes’.

The first and the last sentence have transitive verbs where verb is co
indexed with the second (last) nominal in the former and with the first
nominal in the latter because verbs cannot agree with the case
marked subject nominals. This is also true of experiential
constructions like (91) where verb does not agree with the first
Identity Crisis of Dative Subjects 89

nominal but with the last or second nominal buxar ‘fever’. There is no
controversy over the subjectivity of ma in (90) and (92) [inspite of the
lack of verbal agreement with md in (90). The problem is only with
the nominal ma in (91) which is followed by Dat/Acc case marking.
If agreement is the only clue I find no reason why md in (90) can be
considered ‘subject’ but not in (91). Both the constructions do not
allow cross referencing with subject nominals. In fact, in languages
like Hindi and Punjabi in postpositional sentences* verb agrees with
the last nominal in the sentence. If the verb is of conjunct type
(Adj/Noun + verb) then the verb agrees with the nominal
constituent of the conjunct verb. For instance in sentence 15 from
Punjabi (given earlier) the verb mil ‘to receive’ is coindexed with
object noun roti ‘bread’ because the benefactive nominal is marked
by dative case post positional marking. Similarly in Marathi sentence
7 (also given earlier) verb ‘be’ agrees with masculine plural kapre
‘clothes’ because the first nominal of the sentence which is a
benefactor is in oblique case. See also sentences of Stative
Experiential types (45-47) where verbs agree with the temporal nouns
rather than oblique marked animate nouns. See also Process
Experiential verbs (43, 44). In fact, even in compulsive/obligational
constructions verbs agree with the direct object (72). In case of
intransitive verbs of ‘locational’ type verb is in third person singular
which is a neutral form.
The experiential constructions which allow parallel sentences in
direct and oblique marked nominals (such as 33 : 33a, 42 : 42a, 44:
44a, 82 : 82a) have different focussed nominals. We saw in the earlier
section that verb always agreed with the focussed nominals whether it
is the experience or the patient.
Thus, to summarize, verb agreement is not a very reliable criterion
to determine subjectivity because : (1) Not all Indian languages have
verb coindexing features of subject, i.e. verb does not show agree
ment for gender, number and person. (2) Those languages which do
allow such coindexing allow so only partially, i.e. in one of the tenses
or in few of the pronominally marked sentences. (3) In postpositional
sentences verbs agree with the object noun or with the last nominal of
the sentence.* This statement holds true for languages specified in (2)
above. (4) The existence of explicator compound verb in a sentence
90 Semantic Universals in Indian Languages

disturbs the normal verb agreement pattern.°


Other than the verb agreement there could be other criteria to
decide about the ‘subjectivity’ of oblique marked nominals of
experiential and non experiential types such as Anaphoric Control,
Conjunction Reduction, Position in a Sentence and Topicality. Let
us first consider Anaphoric Control.

Anaphoric Control
We can consider three important devices of anaphoric control. They
are : (a) Reflexive, (b) Conjunctive participle (CP) and (c) Redupli
cated adverbs — all refer to the subject of the sentence. Consider
“sentences (93) (Reflexive), (94) (CP), (95) (Reduplicated adverbs) in
which the anaphoric strategies are in bold.
Hindi (93) mujhe apna becca yad aya
I-DAT self child memory came
‘I remembered my own child’
(93a) vo bar bar apne desh ko yad kor rahi thi
she + NOM again again self country Acc. remmember
do prog past.
‘She was missing her country again and again’
(94) use khana bonta dekh kar zoro se bhukh logne
logi
he+ DAT food cook see CP much hunger began
‘Seeing the food being cooked, he started feeling
hungry’
(94a) vo khana bonta dekh kar vohi beth goi
she + NOM food prepare see CP there sit go past
‘Seeing the food being prepared she sat right there’
(95) mujhe yo bethe bethe shorm a rohi(f) he
I+ DAT like this sitting sitting shyness come prog.
‘I am feeling embarrassd sitting like this’
(95a) vo yi bethe bethe shorma raha tha (m)
he + NOM like this sitting sitting shy prog. past
“He was feeling embarrased sitting like this’
The counterpart (a) sentences make use of direct forms and thus
verbs agree with experiencer nouns in gender and number.
Identity Crisis of Dative Subjects 91

Anaphoric control strategies like apna, -kar and reduplicated ad-


verb in these constructions cross refer to the experiencer nouns.
There is no dispute regarding the identification of subject in such
constructions. In (93),*(94) and (95), however, which are non-(a)
counterpart sentences, the use of dative blocks the verb agreement.
In spite of such blocking, reflexive, conjunctive participle and
reduplicated adverbs used in these sentences co-refer to the dative
experiencer nominals respectivaly. On the basis of such anaphoric
control, the oblique marked nominals (both experiential and non
experiential types) should be considered subjects.
Conjunction Reduction
It is considered that in a conjoined structure (nomitative type)
repeated subject-NP can be deleted in the second clause as it has its
antecedent in the first clause which controls the verb agreement. For
instance, in simple nominative construction from Hindi,
(96) ram ne ti vi dekha @ aur so goya
Ram Erg. T.V. (m) and sleep went (m)
Ram watched the TV and (then) slept.
The verb so-gaya ‘slept’ corefers to the deleted equi NP ram
which in turn has the antecedent in the first clause and is the subject
of the sentence. The Dat NPs of experiential type behave the same
way.
Consider :
Hindi (97) mujh-e nind ai aur @ so gaya
I(m) DAT sleep came(f) and sleep went (m)
‘I felt sleepy and went off to sleep.’
(98) hari ko dhol ki avaz sunai detehi @ gav-ki yad a gal.
Hari-DAT drum Gen sound hear (vi) give-emp.
village-Gen memory (f) come-go-past (f)
‘As soon as Hari heard the sounds of the drum (he) was
reminded of the village.’
In (97) experiencer NP ‘I’ is in the dative case which corefers to non
oblique marked ‘I’ of the conjoined clause. The second ‘I’ (not
overtly present) though distinctly marked (or not marked in this
case), in the underlying structure need not be repeated again hence
92 Semantic Universals in Indian Languages

deleted but is coindexed in the finite verb (m. sg.). Similarly, in the
next sentence (98) ‘Hari’ is in dative and so is the equi NP which is
deleted later. As suggested earlier since postpositions block
agreement, the finite verb no longer agrees with the experiencer NP
‘Hari’ nor it agrees with deleted dative NP ‘Hari’ but with the last
nominal of the sentence ‘yad (f) ‘memory’. Agreement with the
deleted NP in (97) and nonagreement with the first nominal in (98)
both prove that dative subjects or experiential nominals have subject
properties.

Position in a Sentence
The fourth criterion to identity subject nominal is by its position in a
sentence. The agentive subjects normally assume the sentence initial
position. Dative subjects or oblique marked nominals also assume
the same position as is clear from all the 98 sentences from different
languages. Any change in the position is brought only for pragmatic
reasons specially for emphasizing. This brings in the question of
‘topicality’:

Are ‘Topic’ the ‘Subjects’ ?


There is a tendency to select agentive nominals as topics as well as the
subject of the sentence. Topic is the thing one ‘talks about’ and the
extension on the topic is the information which is in focus.
Languages with coindexed verb show agreement with the focussed
nominals in transitive perfective constructions. According to Comrie
(1961) there is a natural correlation between agent and topic; and if
subject is prototypically agentive then agent and topic should
coincide. In oblique marked subject constructions the focus is on the
second nominal which might be an object or object nominal
constituent of the conjunct verb (e.g. buxar (N) hona (Vv) ‘fever is’),
while topic is the first nominal and takes the oblique case. Thus in ma
‘mother’ in (90) as well as in (91) (which is an experiential
construction) is the topic and hence the subject. The argument of
‘topicality’ is independent of the argument of verb agreement. In
languages where verb is not coindexed with nominal features the first
nominal is still the ‘subject’ and ‘topic’ and the second element is the
‘object’ and ‘focus’. Linguists like Verma (1988) feel that ‘topic seems
Identity Crisis of Dative Subjects 93

to be a more basic concept than subject, designed to identify,


definitize, or ‘mark’ the NP that a sentence or proposition is about’. Li
and Thompson (1976) point out that it is probably the topic rather
than the subject that controls the deletion of coreferential
constituents. This is clear in topic prominent languages like Hindi
and other Indian languages.
Considering the last three arguments it seems one can draw
following deductions :

Topic > NP initial position > Subject


(controller in Vi and nonperfective)
Focus > NP non-initial position > Object
(controller of the finite verb in the perfective)

Identity with Passives


As in passive constructions, patient nominal has subject as well as
topic properties and occupies the sentence initial position so does the
non-agentive nominal with oblique marking in the case of experien
tial constructions and non-experiential constructions in non
performative types. We observed that a typical oblique marked
nominal occupies the sentence initial position, gets prominence, and
is the ‘topic’ of the construction.
Another similarity with passive constructions is that verb is
coindexed with that patient noun which is free from any adpositional
markings. Hence, in English, a verb, in passive sentence, agrees with
the first nominal (which is free from ‘by’ PP) and in constructions
under consideration, with the last nominal (which is also free from
any PP markings).
As discussed earlier, oblique marked constructions are truly
passive in semantic import. The oblique marked nominal which is
either experiential or non experiential is never involved in any action
oriented activity. The nominals under consideration are patients of
various states and processes. They are purely non agentive in
nature.
Are Oblique Marked Nominals the Locationals ?
Lately, some attempts have been made to consider experiential
and other constructions of non performative types from different
94 Semantic Universals in Indian Languages

‘point of views’. One such idea is to consider all oblique marked


nominals in experiential and benefactive constructions as /ocational
nominals (Pandharipande, 1988) implying :
(1) Datives do not have subject properties, and (2) datives are
really ‘locational’. The identity of datives (or oblique marked
nominals as we proposed to call them) as ‘subject’ has been well
established by now as one saw in sections considered earlier.
As far as the issue of their being ‘locational’ is concerned there
seems to be some confusion in the argument. ‘Locational’ like
‘experiential’ is a semantic case. The markings on each case can be
either nominative accusative, or any other oblique form. The pro
bability .of locational, or experiential, or dative to assume
“‘subjecthood’ is the same. The bone of contention is not whether the
nominals of non performative are experiential or agent or any other
case but whether they are subject or not. Saying that all datives are
locational is basically shifting from one case frame to another or
changing from one set of arguments to another set of arguments. Ifa
language allows an oblique form which is homophonous to several
case markings (i.e. postpositions) such as Pandharipande has cited
for Marathi /a: (which can work for ‘possession’, ‘location’, ‘dative’,
‘accusative’ etc.) then calling /a: by one unique case marking is not
justified. In many languages ergative and instrumental are
homophonous (Comrie 1989). ‘Cases’ are semantic relations holding
between the nominal and verbal elements of a sentence. Their
linguistic encoding is another matter. Her argument that since the
homophonous? suffix -/a: in Marathi which marks the experiencer
NPs typically marks the non-subject (mostly locational Ns) NPs
elsewhere in the language and hence should not be considered subject
is not well founded. Firstly, she considers transitive sentences to
prove her point. The kind of constructions specially the true
‘experiential constructions’ that we are discussing are like
intransitive constructions which have no object nominals.
Comparing these with transitive sentences with object nouns to
prove that the subject nominals of experiencer construction is not
subject is fallacious. Secondly, her definition of experiencer is rather
vague as she takes into account non experiential constructions as
well that we have seen in section I.
Identity Crisis of Dative Subjects 95

Point of View
There have been arguments that the difference between the two types
of parallel constructions such as in 33 and 33a where in one expe-
riencer NP is encoded with oblique marking and in the other it is not is
that in the latter the situation is described from the point of view of
an observer or the speaker while in the former the speaker is
reporting a situation which may not be perceivable to an outsider
(Joshi 1988). The two of them are distinguished by the ‘point of view’
factor on external-internal perspective factors. This argument is too
simple to be accepted. Let us consider sentences 34 and 35 again. In
both the constructions ‘IT’ is in a state of biological/physical ailing
state. Both the sentences are of internal perspective. Where should
we draw the line and how? Similarly one can take almost all the
parallel cases of oblique marked subjects and non oblique marked
subjects analysed in sections I, II for discussion but we can not prove
beyond doubt what is external and what is internal perspective
factor. In fact, all experiential constructions and to some extent all
benefactive and attributive constructions have internal perspective
factors as far the experiencer nominal or benefactor nominal or the
nominal in the state of attribution is concerned. Even if the causing
factor of the experience is outside the domain of the experiencer, the
experiencer perceives the ‘effect’ of the factor and thus is an ‘affected’
party or is an ‘internalized experiencer’. As we have seen, the
difference between 33 and 33a or 34 and 35 is not of ‘volition’ nor is it
of ‘point of view’ but is of FOCUS. The basic difference between the
two parallel constructions emerges due to non focussed element
encoding the oblique marking.

The Solution :
We already know that there is one or the other kind of oblique
marking (mostly dative and accusative) wich follows the subject
nominal of the experiential constructions. The verbs in such
constructions are mainly of conjunct verb type which should be seen
as one whole lexical unit as semantically they are states and processes
rather than object and verb combinations. We can never equate
‘transitive verbs’ that obligatorily take objects with ‘conjunct verbs’
which have either an adjective or a nominal as an integral
96 Semantic Universals in Indian Languages

constituent. If we take the basic experiential construction as


constitued of the following configuration then the constructions will
not be misunderstood. The NP in the beginning of the sentence
assumes the ‘subjecthood’ and the second nominal governs the verb
agreement because the first nominal is no longer available (because
of oblique marking) for agreement control. Temporal situational
constructions and sensory and mental experiential constructions
which have the verbs (such as ‘like’, ‘love’, ‘hate’ etc.) which take
object nouns do not disturb the analysis. Let us consider a Hindi
example to clarify our situation :
Ea

S Te, este + \P

i 32
N < (Adj.) Complex: Verb

Exp. N obl. .m. N+Adj/Nom V, (V2)

(Sensory and Mental Experiences)


Hindi (99) mujh-e shila acchi__logti hai
I-DAT Shila good (f) strike (f) aux
‘T like — shila’
The Object noun ‘Shila’ governs the adjective component of the
conjunct verb in agreement (acchi - fem-sg.) and is preceded by
which is marked for dative case. The object of ‘like’ (accha lagna)
is Shila and the subject of the verb ‘like’ (accha lagna) is ‘TY’. The
oblique marking on ‘I’ in Hindi in this particular case and similar
coding in Indian languages should not deter us from identifying it as
‘subject’. In fact, wherever the conjunct verb is of Adj+¥ type there
will obligatorily be a preceding object noun with which the adjective
and verb will agree in number and gender.
The V,(V2) configuration stands for the possibility of a combi
nation of conjunct and explicator compound verbs, such as in:
Hindi (100) mé dhokha kha gaya
I dececeit (m) eat went (m)
I was cheated (verb agrees with ‘I’ )
Identity Crisis of Dative Subjects 97

(101) —_us-e soza mil go/


he-DAT punishment (f) receipt went
“He was punished’ (verb agrees with
‘punishment’ )
The conjunct verbs are dhokha kha- and soza mil-. The two
constituents of the verbs are Nom+verb. Conjunct verbs, in turn are
followed by explicators such as gaya (100) and ga/ (101). The
entire verb phrase is a complex verb constituted of conjunct verb (V,)
and compound verb (V; + V2). The nominal vé in (100) and us-e in
(101) are both experiential nominals as well as subjects of the
experience.
The agreement is controlled by the subject noun mé in the first case
(100) and by the object noun saza in the latter (101). Verb is a
complex phenomenon in Indian languages both from the syntactic
and semantic point of view.
To summarize, one would propose that:
(1) Syntactically, experiential and non experiential constructions
typically take some kind of oblique marked, sentence initial nominal.
This nominal can be assigned the role of ‘subject’ because it has
‘subject’ like properties of (a) blocking of verb agreement in
postpositional sentences; (b) anaphoric control, (c) conjunction
reduction, and (d) sentence initial position. The constructions under
consideration are like passive constructions. This is the reason why
further passivization of these constructions is not possible.
(2) Semantically, the constructions are ‘involuntary’ in nature and
the affected party is an “experiencer’ or ‘benefactor’ or ‘recipient’ or
just ‘exist’ in a state. In other words, the constructions are Non
Agentive in nature, and indicate a kind of passivity on the part of the
non agent nominal. This passivity is encoded linguistically in most of
the languages by dative or any other oblique marking and in the rest
by the absence of marking. This proves that linguistic encoding is not
the clue to decide upon the ‘passive’ nature of these constructions. It
is the semantic constructs of the verb which decide the ‘non
agentiveness’ or ‘passivity’ of the nominal under question. We have
discussed such verbs in great detail.
(3) Identification of ‘subject’ in such constructions is dependent
upon what is the ‘topic’ and what is the ‘focus’ in the construction.
anbijqo poxrey S[BUIWON
ce
eedeMs
uoN [enuotedxg [enuotadxq

saToRJounO
sg Jo jONUOD sanNqLUny a1S SS3001g 9ANeIS UOT}OYV SS9
|
|
|
31S SSI001g -Iqedeout
§=ayejg_ saeoyouy />es SS9901g je1odwiay Jeuon
Saatssed
= Ae]

conceal

SAISSOSSOgules SSO]
H
datsyndurod Alosuas peotskug [euono
pue pue pue
jeuonesi|qo [equa [eo1sojorg
aqeuarye
©= g[qeualeur

“BLT
‘¢ UON aanuasy anbijqQ payer S[EUIWION
Identity Crisis of Dative Subjects 99

Topic generally assumes the ‘subject’ position and ‘focus’ controls


the agreement in a verb. Examples where subject nominals are
unmarked for any case (either because the language does not allow
any marking or because the language allows parallel constructions
with and without oblique marked subjects) topic and focus coincide
into the subject nominal and verb agrees with the subject nominal not
because of its topicality but because of its being the ‘focus’.
Considering arguments given in (1), (2) and (3) above, I would
propose that these constructions should be called ‘Non-Agnetive
Subject Constructions’.
Passivity in Indian languages, thus, is not restricted to syntactic
passives. It can be manifested in such constructions that we have just
analysed, where there is no agent. What exists in place of agent is an
‘affected’ nominal which typically occupies the sentence initial
position and which may or may not be coindexed in the verb. The
choice of it being coindexed in the verb or itself being coded by some
kind of oblique marking lies in the fact what is the topic and what is
the focus in the sentence; or the language may not allow any choice
because of it idiosyncratic characteristics. In such cases, semantic
criteria are the only criteria that help us decide the subject property
of the nominals under consideration.

NOTES

" Some of the example are cited from the various handouts distributed at the 17th
South Asia Conference held at Madison, Wisconsin in Nov. 1988. The parenthesis
against each sentence referstc the author’s apbreviated names. These are :
DM David Magier
KP Krishna Pradhan
KPM K.P. and Tara Mohanan
MKM Mithilesh K. Mishra
MKV Mainindra K. Verma
RB Rakesh Bhatt
RP Rajeshwari Pandharipande
SLC Shobha L. Chelliah
SNS S.N. Sridhar
HU Helen Ullrich
The papers presented in the said conference have been published by the Stanford
Linguistic Association. See Verma and Mohanan 1990. Nahali examples are from
Kuiper’s Nahali: A Comparative Study, and Kharia sentences are from Veena
Malhotra’s Ph.D. dissertation ‘The Structure of Kharia’.
100 Semantic Universals in Indian Languages

2. M.K. Verma, ‘Topic and Subject’ in Encyclopaedia of Linguistic Terminology (ed.


V. Prakasam).
3. Sentences with any postpositional marking on the very first nominal of a sentence
of regular sov type. \
4 Exception is Nepali (M.K. Verma 1976) where subject does not lose its control
over verb even in perfective. However, it loses its control over experiential and
obligational constructions.
>. See Abbi Agreement in Hindi (forthcoming).
© The author opines that the nature of ahomophonous linguistic unit has generated
more problems and confusion than help as linguists are tempted to club together two
or three distinct and discrete semantic units into one class on the basis of identical
linguistic encoding. This is the problem when a linguist tries to make a journey from
surface to deep rather than from deep to surface.
CHAPTER V

Conclusions

It has been demonstrated that typlogically and genetically diverse


languages, majority and minority languages, literary and tribal
languages of India today share linguistic structures which are
mainfestations of shared underlying semantic structures. Typo-
logically divergent languages may become typologically uniform
languages if they are spoken contiguously in an area, have stable and
long contact among themselves, are marked by bi/mutilingual speech
communities, and are not tampered with artificial insemination (i.e.
Standardization). The discipline devoted to the study of such
phenomena is known as Areal Typology. Areally typologized
languages are indicators of sharing of semantic, cognitive and
socio-cultural aspects of a region. The speakers of such areas inspite
of using varying and different languages, share common ethos and
common socio-cultural ambience.
Linguists generally maintain that contact interference is a
precursor to minority languages loss (Haugen 1989). The process of
‘loan to loss’ cannot be generalized but popularly known to pass
through the stages of acculteration and ‘gratuitous’ loan word
replacing vacabulary items already present in the language. The
Indian situation, however, does not totally fall into this paradigm.
Minority languages’ such as Kharia (Munda), Kurux (Dravidian),
Konkani (Indo Aryan) have heavily borrowed from the dominating
languages which are areally contiguous to them. When languages are
in contagious situation and open to gratuituous borrowing then the
initial borrowing stage gets stabilized and the structures thus
borrowed take part in the journey of language development just as
the non borrowed indigeneous structures do. The fact that inspite
102 Semantic Universals in Indian Languages

the pressures otherwise the minority languages have been maintained


in themultilingual India and the phenomenon of language shift is not
very prevalent are the two key factors which have stopped the
language attrition of these minority speeches. They have been
converged as far as lingustic structuration is concerned, with the
~ neighbouring dominating languages. The convergence, however, has
not been discouraging to ‘linguistic innovations’, which these
languages do experience from time to time.
What I am sugessting is that the process of language change
through language convergence as well as the process of language
development through new-and-newer innovations can continue
simultanously. The parallel movement of two processes force a
language to converge with the other language and at the same time
maintain its identity for social solidarity purpose.
Language change, specially in the context of the minority and
lesser known languages, may lead to Janguage contraction, i.e.
reduction in the number of speakers of language x and the reduction
of domains of use of languages x. Perhaps, a long period of state of
contraction ultimately leads to language obsolescence. But this
would be possible only if non lingustic factors such as adverse
socio-economic situations, lack of educational facilities, rise in inter
group, inter-caste and inter-regional marriages, discriminatory
policies of the government, etc. coexist with linguistic factors
responsible for language decay and death.
The tribal languages of India, today have already entered the
domain of language contraction and one can speculate the time
factor involved in their obsolescence. However, there is one redeem
ing factor. That is, there has been upsurge of ‘language loyalists’ and
‘language purists’ (language awareness in Manipuri by discarding
the present script for the old one, or in Kurux speech region where
movement is to demand Kurux as a medium of instruction in primary
schools, etc.) who might help maintaining the language and save it
from attrition though at a heavy price of blocking the language from
going through ‘change’ due to contact and convergence. This is
because those who have worked on tribal languages in India must
have felt that loyalists motivated to maintain the use of languages are
often also highly conservative monitors of the form of that language.
Conclusions 103

As Woolard (1989: 357) rightly puts ‘the will for maintenance and for
purity do seem to have the same roots’.
Other than the drive by ‘languages purists’ the linguistic and the
social solidarity of the speech community also help preserve some
characteristic features of the language. The typical examples are in
the realm of phonetics and phonology. The lamino-dentals in Khasi,
the glottal stop in Munda and Kurux, the implosives in Sindhi, the
glottal fricatives in Assamese, the vioced apirates in Indo Aryan are
cases in point. These features are individual language identity
markers. The areal features, on the other hand, are area identity
markers. Each language in a typical language contact situation, thus
preserves both individual as well as group identity. The group
identity helps it-to be distinguished from the identity of the
languages of the other neighbouring areas.
Our investigation into the semantic structures of expressives, echo
formations, word reduplication, explicators, and non agentive
subjects reveal a very significant aspect about the Indian languages
and its users. It is mostly those semantic constructs which pertain to
perceptive and sensory abilities of a human being that materialize in
structural cognates. These abilities are predetermined by specific
socio-cultural environment of the speakers of the region. We can ask
ourselves that why don’t we find structural resemblances without
parallal resemblances in meaning? Or why don’t we find varying or
different linguistic structurations for same or similar semantemes?
Why five senses of prceptions manifest themself in expressives ? Why
passivity or out of control situation is encoded by obligue marking
on the subject nominal? Or why inadvertent action is manifested in
the use of an explicator? Surely, these are neither chance resemb
lances nor genetic inheritences. These are language contact induced
phenomena which, having crossed the barriers of history and
geography, have sustained in the various speech communities over a
long period of time.
What is obscure till date is the process of this diffusion and
convergence. But what is transparent is the fact that in a language
contact situation many of the languages change drastically. Drastic
enough to surprise any historical linguist. To him they might appear
as ‘sister’ languages, not being able to segregate historical affinity
104 Semantic Universals in Indian Languages

from areal affinity. Transparency is also reflected in the typlogy of


these languages which undergoes a change and becomes like that of
the adjacent languages irrespective of the genetic or typological
leanings of the latter.
The homogenized signifier-signifiant relationship that holds bet-
ween diverse languages of India is a unique feature that reflects a
nation with strong and stable multilingual community. Perhaps, it is
in the interest of all of us not to have it disturbed by artificial forces
like monolingualism and reduction in the domains of language
use.

NOTE

'Eric Allardts (1984 : 195-205) rejects the idea that restricts the use of the word
‘minority’ to a fixed percentage or a proportion that a linguistic minority cannot
exceed. The decisive factor in making them linguistic minorities is not the size but
rather the social organization and their ‘place in society’. A language can be
subordinate to a dominant language to be claimed to be a minority language.
Select Bibliography

Abbi, Anvita. 1980. Semantic Grammar of Hindi. Bahri Publications, New


Delhi.
——. 1984. Conjunctive Participles in Hindi-Urdu. IJDL. vol 13, no.2,
(252-63).
——. 1985. ‘Reduplicative Structures: A Phenomenon of the South Asian
Linguistic Area’ in For Gordon H. Fairbanks (eds. Vaneeta Z. Acson and
Richard L. Leed). Oceanic Linguistics Special Publication no. 20.
University of Hawaii Press, Honolulu.
——. (ed) 1986. Studies in Bilingudism. New Delhi. Bahri Publications.
——. 1988. ‘Experiential Constructions, and the ‘Subjecthood of
Experiencer NPs in South Asian Languages’. Paper presented at 17th
South Asian Conference, Madison Wisconsin. Proceedings publised by
Standfsdord University Press, 1990.
——. 1990. ‘Reduplication in Tibeto- Burman Languages of South Asia’, in
South Asian Studies, vol. 28 : 2, Sept. 1990.
. 1991a. Reduplication in South Asian Languages. An Areal, Typological
and Historical Study. Allied Publishers, New Delhi.
. 1991b. ‘Language Contraction, Language Shrink, and Language
Conflation : A case study of Kharia’, paper presented at the seminar on
‘Continuity and Change in Tribal Community’, Jan. 14-19, 1990, New
Delhi.
Abbi, Anvita and Mithilesh K. Mishra. 1987. ‘Aspectual Elements of
Simultaneity and Iteration in Indian Languages :A Case for an Areal
Universal’. Studies in the Linguistic Sciences. no. 1, Spring 1987
(1-14).
Acharya, K.P. Rekha Sharma, Sam Mohan Lal and K.S. Rajyashree. 1987.
Pidgins and Creoles as Languages of Wider Communication. C.1.1.L.
Mysore.
Allardts, Erik. 1984. ‘What constitutes a Language Minority?’ Journal of
Multilingual and Multicultural Development vol 5. no. 3, 4 (195-205).
Apte, M.L. 1968.Reduplication, Echo Formation, and Onamatopoeia in
Marathi. Deccan Callege Poona.
106 Select Bibliography

Aronoff, Mark (ed.). 1984. Language, Sound and Structure, M.I.T. Press
Cambridge.
Asher, R.E. 1985. Tamil. Croom Helm Descrptive Series. Routledge.
London.
Bach, Kent. 1987. ‘Thought and Reference, Clarendon Press. Oxford.
Bhatia, Tej. 1988. ‘The Notion ‘Subject’ in Punjabi and Lahanda’. Paper
presented at 17th South Asia Conference, Madison.
Bhatt, Rakesh. 1988. ‘Dative Subjects in Kashmiri Paper presented at 17th
South Asia Conference, Madison.
Bhattacharya, S. 1975. Studies in Comparative Munda Linguistics. 1.1.A.S.
Simla.
Bloch, Jules. 1965 (Eng. Ed.) Indo Aryan: From the Vedas to Modern Times.
Libraire d’ Amerique et d’ orient Adrien Maisonnenue Paris.
Botha, Rudolf, P. 1988. Form and Meaning in word formation. A Study of
Afrikaans Reduplication. Cambridge University Press. New York.
Briggs. Charles L. 1986 (reprinted 1989), Learning How to Ask. A
Sociolinguistic Appraisal of the Role of the Interview in Social Science
Research. Cambridge University Press. New York.
Burrow, T. and Emeneau, M.B. 1960. A Dravidian Etymological Dictionary.
Clarendon Press Oxford. 1968 Supplement.
Butterworth, Brian, B. Comrie and Osten Dahl (eds). 1984. Explanations for
Language Universals Mouton Publishers New York.
Caldwell, Robert. 1856. A Comparative Grammar of the Dravidian or South
Indian Family of Languages. Harrison London.
Campbell, L.T. Kaufman and T.C. Smith-Stark. 1986 ‘Meso-America as a
Linguistic Area’. Language. vol. 62, no. 3 (530-70).
Chatterji, S.K. 1942. Indo- Aryan and Hindi, Gujarat Vernacular Society.
Calcutta (reprinted in 1960 & 1969).
Chelliah, Shobhna L. 1988. ‘Experiencer Subject in Manipuri’. Paper
presented at 17th South Asia Conference, Madison.
Chomsky, N. and H. Lasnik. 1977. ‘Filters and Control’. Linguistic Inquiry.
vol. 8 (425-504).
Chomsky, N. 1981. Lectures on Government and Binding. Amesterdam.
Foris.
Chomsky, N. 1988. Language and Problem of Knowledge the M.I.T. Press,
Messachuttes.
Cole, P. & J.M. Sodock. 1977. Syntax and Semantics : Grammatical
Relations. vol. 8. Academic Press. New York.
Comrie, B. 1977. ‘In Defense of Spontaneous Demotion : the Impersonal
Passive (47-58) in Syntax & Semantics. vol. 8.
——. 1989. Language Universals and Linguistic Typology. Syntax and
Morphology. Basil Blackwell, Oxford.
Cruse, D.A. 1986. Reprinted 1989. Lexical Semantics. Cambridge Textbook
in Linguistics
Select Bibliography 107

Davison, Alice. 1985. ‘Experiencers and Patients as Subjects in Hindi-Urdu’,


in Participant Roles : South Asia and Adjacent Area (eds. Zid et al.).
Dorian Nancy C. 1989. Investigating Obsolescence : Studies in Language
Contraction and Death. Cambridge University Press.
Emeneau, M.B. & T. Burrow. 1962. Dravidian Borrowings from Indo Aryan.
University of California Publication in Linguistics vol. 26. University of
California Press. Berkeley, Los Angeles.
Emeneau, M.B. 1938. ‘Echo Words in Toda’. New Indian Antiquary. vol. 1.
(109-17).
—. 1956. ‘India as a Linguistic Area’. Language 32. 3-16. (reprinted in
Emeneau 1980 : 105-25).
——. 1969. ‘Onomatopoetics in the Indian Linguistic Area’, Language 45 :
2, pt. 1 247-99.
. 1965. India and Linguistic Areas. India and Historical Grammar. (Dept.
of Linguistics, Publication 5), 25-75. Annamalai Nagar, Annamalai
University (reprinted in Emeneau 1980 : 126-66).
——. 1974 ‘The Indian Linguistic Area Revisited’ in Contact and
Convergence in South Asian Languages edited by Franklin C. Southworth
and Mahadev L. Apte. Sp. Publication of International Journal of
Dravidian Linguistics Publication (Trivandrum).
—. 1980. Language and Linguistic Area. Essays by Murray B. Emeneau,
Selected and introduced by Anwar S. Dil. Stanford University Press
Stanford.
Gair, J.W. 1970. Colloquial Sinhalese Clause Structure. Mouton. The
Hague.
Gal, Susan. 1989. ‘Lexical Innovation and Loss : The Use and Value of
Restricted Hungarian’. In Investigating Obsolescence ed. Nancy C.
Dorian.
Giles, Howard and R.N. St. Clair. 1979. Language and Social Psychology.
Basil Blackwell, Oxford.
Giles. H. (ed) 1971. Language, Ethnicity and Intergroup Relation. European
Monographs in Social Psychology no. 13. Academic Press.
Givon, T.1976. ‘Topic, Pronoun and Grammatical Agreement’ in Subject
and Topic (ed. C.N. Li), Academic Press, New York. pp. 149-88.
Gopala |Krishanan, Devi 1983. ‘Verb Sequences in Malayalam’ Ph.D. diss.
J.N.U. New Delhi.
Greenberg, J.H. (ed). 1963, 1966 2nd ed. Universals of Language. The M.I.T.
Press, Massachuttes.
—. 1971. Language, Culture and Communication. Stanford University
Press.
. 1974. Language Typology : A Historical and Analytic Overview
Mouton. The Hogue.
108 Select Bibliography

——. 1978. Charles Ferguson, and Edith A. Moravesik. 1978. Universals of


Human Language. Method and Theory. 1 Stanford University Press,
Reviewed by H. Seiler, Language, vol. 56, no. 4, 1980.
Greenberg, J.H. Charles Ferguson, and Edith A. Moravesik. 1978
Universals of Human Language. Phonology. vot. IJ. Stanford University
Press, Reviewed by Hector Raul Javkin. Language, vol. 56, no. 4.
1980.
Greenberg, J.H., Charles A. Ferguson and Edith A. Moravesik. 1978.
Universals of Human Language. Word Structure-III. Reviewed by
Bernard Comrie, Language, vol. 56. no. 4 (1980).
Greenberg, J.H., Charles A. Ferguson, and Edith A. Moravesik, 1978.
Universals of Human Language. Syntax vol-IV. Reviewed by W.P.
Lehmann Language. vol. 56. no. 4 (1980).
Grierson, G.A.1903-28. Linguistic Survey of India. 11 vols. Reprinted.
Motilal Banarsidass, New Delhi, 1987.
Gvozdanovic, Jadranka. 1985. Language System and its Change on Theory
and Testability. Mouton, the Hague.
Hacker,Paul. 1961. ‘On the Problem of a Method for Treating the
Compound and Conjunct Verbs in Hindi. In Bulletin of the School of
Oriental and African Studies. 24 : 484-516.
Hahn, Revd. Ferd. Grammar of the Kurux Language. Mittal Publications,
Undated. The work was done in 1908.
Haugen, E.1953. The Norwegian Language in America : A Study in Bilingual
Behaviour (rpt. 1969). University of Pennsylvania Press, Philadelphia.
Hill, Jane H. 1989. “The Social Functions of relativization in obsolescent and
non-obsolescent languages’. In Investigating Obsolescence. ed. Nancy C.
Dorian.
Hoffman, Karl. 1952. ‘Wiederholende Onomatopetika in Altindischen,
Indogermanische Forschungen. 60. 254-64
Holland, Dorothy & Naomi Quinn. 1987. Cultural Models in Language and
Thought. Cambridge University Press.
Hook, Peter. 1974. The Compound Verb in Hindi. AnnArbor. The University
of Michigan Publication Ann Arbor Michigan.
Hook, Peter Edwin 1982. ‘South Asia as a Semantic Area. Forms, Meanings
and their Connections’. South Asian Review. vol. 6 (30-41).
Humboldt, Wilhelm Von. 1988. On Language. The Diversity of Human
Language-Structure and its Influence on the Mental Development of
Mankind. Cambridge University Press.
Hutchins, W.J. 1971. The Generation of Syntactic Structure from a Semantic
Base. North Holland Ling. Series. North Holland Publishing Co.
Jakobson, Roman. 1978. Six Lectures on Sound and Meaning. The Harvester
Press.
Jespersen, Otto. 1924. The Philosphy of Grammar. George Allen & Unwin
Ltd., London (rpt. 1968).
Select Bibliography 109

Joshi, Smita. 1988. ‘Two Semantic Properties of Marathi Dative Subject’.


Paper presented at 17th South Asian Conference, Madison.
Kachru, Yamuna. 1980. Aspects of Hindi Syntax. Manohar Publishers
Delhi.
Katenina, T.E.1957. “Kratkii ocherk grammatiki iazyka khindi’. Appendix
to Russko-khindi slovar (ed. A. Beskrovny), Gosudarstvennoe Izadatel’
stvo Inostrannykh i Natsional’ nykh Slovarei, Moscow, 1279-376.
Keenan, E.L. 1976.‘Towards a Universal Definition of ‘Subject’. Subject
and Topic (ed Charles N. Li). Academic Press. New York, 303-33.
Keenan, Edward,L. and Bernard Comrie. 1977. ‘Noun Phrase Accessibility
and Universal Grammar. Linguistic Inguiry, 8 : 66-99.
Keenan, Edward L. 1987. Universal Grammar : 10 Essays. Croom Helm.
Khubchandani, L.M. 1983. Plural Languages, Plura! Cultures:
Communication, Identity, and Sociopolitical Change in Contemporary
India, East West Centre Honolulu.
Klaiman, M.H. 1986. ‘Semantic Parameters and the South Asian Linguistic
Area’. South Asian Languages : Structures Convergence and Diglossia.
(eds. Bh. Krishnamurti, C.P. Masica, and A.K. Sinha) Motilal
Benarsidass, Delhi.
Krishnamurti, Bh., C.P. Masica and A.K. Sinha (eds) 1986. South Asian
Languages: Structure, Convergence and Diglossia. Delhi. Motilal
Benarsidass. Delhi.
Kuiper, F.B.J. 1962. Nahali. A Comparative Study. N.V.
Noord-Hallandsche Uitgevers Maatschappij.
Kuiper, F.B.J. 1966. ‘The Sources of the Nahali Vocabulary’ in Studies in
Comparative Austro-Asiatic Studies (ed. by Norman H. Zide). Mouton,
London. The Hague. Paris.
Kuiper, F.B.J. 1967-68. ‘The Genesis of a Linguistic Area’ in Indo-Iranian
Journal (The Hague) vol. 10, pp. 81-102. Reprinted 1974. in Contact and
Convergence in South Asian Languages. (eds. Franklin C. Southworth
and Mahadev L. Apte). Special Pub. of International Journal of
Dravidian Linguistics Publication, Trivandrum.
Ladefoged,P. 1971. Preliminaries to Linguistic Phonetics, Chicago
University Press, Chicago.
Lehiste, Ilse. 1988. Lectures on Language Contact, The MIT Press.
Li, CharlesN. (ed). 1976. Subject and Topic. Academic Press, New York.
Li, C.N.and S.A. Thompson. 1976, ‘Subject and Topic : A New Typology of
Languages’ Subject and Topic (ed. C.N. Li), Academic Press, New York,
457-89.
Liebermann, Philip, 1975.On the Origins of Language : An Introduction to the
Evolution of Human Speech. 1975. McMillan Press.
Lindholm, J.M. 1976. ‘Nested Caste Relations and the Subject in Tamil’
NSAAL (ed. Verma), University of Wisconsin Publication no. 2
(152-82).
110 Select Bibliography

Magier, David. 1988. ‘Dative/Accusative Subjects in Marwari’. Paper


Presented at 17th South Asia Conference, Madison.
Malhotra, Veena. 1982. ‘The Structure of Kharia : A Study of Linguistic
typology and Language change’. Ph.D. Diss. J.N.U.
Mallinson, Graham and Bary J. Blake. 1981. Language Typology. Cross
Linguistic Studies in Syntax. North Holland Publishing Co.,
Amesterdam.
Masica, C.P. 1976. Defining a Linguistic Area: South Asia. The University of
Chicago Press, Chicago.
Masica, C.P. 1988. ‘The Significance of Varied Case Marking in
Obligational Constructions’. Paper presented at 17th South Asia
Conference, Modison.
Mandelbaum, D. 1963. Selected Writings of Edward Sapir. University of
California Press.
Martinich, A.P. (ed). 1985. The Philosophy of Language. Oxford University
Press.
Mohanan, K.P. 1984 ‘Functional and Araphoric Control’. Linguistic
Inquiry. vol. 14.
Mohanan, K.P. & Tara Mohanan. 1988. ‘Dative Subjects in Malayalam:
Semantic Information in Syntax’. Paper presented at 17th South Asia
Conference, Madison.
Moravesik, E.A. 1978. ‘Reduplicative Constructions’ in Universals of
Human Language. (ed). Greenberg J. vol. 3, Stanford University Press,
Stanford.
Mishra, M.K. 1988. ‘Dative/Experiencer Subjects in Maithili’. Paper
presented at 17th South Asia Conference, Madison.
Muhlhausler, Peter. 1986. Pidgin & Creole Linguistics.
Munizel, M.C. and Lyle Campbell. 1989. ‘The Structural Consequences of
Language Death’. In Investigating Obsolessence (ed. Nancy Dorian).
McAlpin, D.W. 1976. Dative Subject in Malayalam. NSAAL(ed. Verma),
University of Wisconsin, Pub no. 2 (183-94).
Pandharipande, R. 1988. ‘Experiencer (Dative) NPs in in Marathi’. Paper
presented at 17th South Asia Conference, Madison.
Pandit, P.B. 1972. India as a Sociolinguistic Area. University of Poona.
Paul, H. 1909. Prinzipien der Sprachgeschichte 7te Aufl. Halle. (Secondary
Source : O. Jespersen).
1916. Deutsche Grammatik. Halle. (Secondary Source: O.
Jespersen).
Perlmutter, D.M. 1983 Studies in Relational Grammar, vol. 1, Chicago
University Press, Chicago.
Peri Bhaskararao,1977. Reduplication and Onomatopoeia in Telugu. Deccan
College, Poona.
Peterson, Philp L. 1985. ‘Causation, Agency, and Natural Actions’, CLS
Procedings, vol. 21, no. 2, 204-27.
Select Bibliography 111

Plank, Frans (ed). 1980. Objects : Towards a Theory of Grammatical


Relations. Academic Press London.
Plank, Frans (ed) 1985. Relational Typology. Trends in Linguistic Studies,
Monographs 28, Mouton de Gruyter.
Portman, M.V. 1887. A Manual of the Andamanese Languages, W.H. Allen,
London.
Pradhan, Krishna, 1988. ‘Subject & Experiencers in Newari’, Paper
presented at 17th South Asia Conference. Madison.
Prakasam, V. 1985. ‘Case Relations and Realizations’. The Linguistic
Spectrum (ed By V. Prakasam). IILS Patiala.
Prakasam, V. and Anvita Abbi, 1980. Semantic Theories and Language
Teaching. Allied Publishers, New Delhi.
Pray, Bruce, 1976. ‘From Passive to Ergative in Indo Aryan. NSAAL (ed.
Verma) University of Wisconsin, Pub. no.2 (195-211).
Rabel, Lili. 1968. ‘Redundant Expressions in Khasi’. Studies in Indian
Linguistics. (ed Bh. Krishnamurti), Poona.
Rabel- Heyman L. 1976. ‘Sound Symbolism and Khasi Adverbs’. Southeast
Asian Linguistic Studies. vol 2, Pacific Linguistics, Series C, No 42.
Canberra.
Ramanujan, A.K. and Colin P Masica. 1969. ‘Toward a Phonological
Typology of the Indian Linguistic Area’ in Current Trends in Linguistics
(ed. T. Sebeok). vol. 5 (543-577).
Ramat, Paolo. 1987. Linguistic Typology. Mouton de Gruyter.
Report of the Official Language Commission of Gujarat Government.
1960.
Riccardo, Stella and Maris Bortoni, 1985. The Urbanization of Rural Dialict
Speakers: A Sociolinguistic Study in Brazil. Cambridge University
Press.
Robins, R.H. 1969. ‘The History of Language Classification’. Current
Trends in Linguistics (ed. T. Sebecok), vol. 11.
Romaine, Suzanne 1989. Bilingualism. Basil Blackwell, London.
Sastri, K.M.1969. Historical Grammar of Telugu : With Special Reference
to Old Telugu. c. 2000 B.c.-A.D.1000, Andhra Pradesh. Sri Venketeswara
University.
Schieffelin, B.B. & Elinor Ochs. 1986. Language Socialization Across
Cultures. Cambridge University Press. New York.
Sapir, E. 1963. Selected Writings of Edward Sapir. Language, Culture and
Personality. University of California Press.
Sebeok Thomos A. (ed) 1969. Current Trends in Linguistics. Linguistics in
South Asia. vol 5. Mouton. The Hague.
Schiffman, Harold E. & M.C. Shapiro. 1981. Language and Society in South
Asia. Motilal Banarsidass. Delhi.
Selvam. 1988. ‘Expressives in Telugu’ M.Phil. Diss. J.N.U. New Delhi.
112 Select Bibliography

Shopen, Timothy (ed). Language Typology and Syntactic Description. Clause


Structure—vol I, Complex Constructions—vol II, Grammatical
Categories and Lexicon—vol III.
Sridhar, S.N. 1976. ‘The Notion of “Subject” in Kannada in NSALL (ed.
Verma), University of Wisconsin, Pub. no. 2 (212-39).
Sridhar, S.N. 1988. ‘The Semantics of the Dative Subject Construction in
Kannada’. Paper presented at 17th South Asia Conference Madison.
Steever, Sanford B. 1987. ‘Tamil and Dravidian Languages’ in The World's
Major Languages (ed. B. Comrie) Croom Helm, London.
Stout, A.P. and G.F. Stout. 1902. Analytic Psychology. London. (Secondary
Source : O. Jespersen)
Thomason, S.G. and T. Kaufman. 1988. Language Contact, Creolization
and Genetic Linguistics. University of California Press, California.
Reviewed by Jay H. Jasanoff. Language, vol. 65. no. 3 (1989).
Tomlin, Russell. S 1986. Basic Word Order: Functional Principles. Croom
Helm. London.
Trudgill, Peter. 1906. Dialects in Contact. Basil Blackwell. London.
Turner, R.L. 1966. A Comparative Dictionary of the Indo Aryan Languages.
Oxford University Press. London.
Ullmann, Stephen. Descriptive Semantics and Linguistic Typology.
Ulrich, H. 1988. “Subjective Experiencers in H. Kannada’, Paper presented
at 17th South Asia Conference Madison.
Vatter, Harold J. 1969. ‘Sound Symbolism in Language’ in Language
Behavior and Communication : An Introduction. Peacock Publication.
Verma. M.K. (ed). 1976. The Notion of Subject in South Asian Languages.
South Asian Studies. University of Wisconsin, Madison, Pub. no. 2.
Verhaar, John W.M. 1985. ‘On Iconicity and Hierarchy’ in Studies in
Language 9.1. 21-76.
Verma, M.K. 1976. ‘The Notion of Subject and the Data from, Nepali’ in
NSSAL (ed. Verma University of Wisconsin Publication, Madison
Verma, M.K. 1988. ‘Topic and Subject’, in Encyclopaedia of Linguistic
Terminology, (forthcoming), Allied.
Verma, M.K.and K.P. Mohanan (eds.) 1990. Experiencer Subjects in South
Asian Languages. The Stanford Linguistic Association, CSLI,
Stanford.
Vestergaard, Torben and Kim Schroder. 1985. The Language of Advertising.
Basil Blackwell, London.
Weinreich, Uriel, 1968. Languages in Contact. Mouton. The Hague.
Woolard, Kathryn, A 1989. ‘Language Convergence and Language Death as
Social Processes’. In Investigating Obsolescence (eds. Nancy C.
Dorian).
Wundt, S. 1900. Die Sprache. Leipzig
Zide, N.H. (ed). 1966. Studies in Comparative Austro-Asiatic Linguistics.
Mouton The Hague.
Select Bibliography 113

Zide, N.H. 1976. ‘A Note on Gta? Echo Forms’ in Austro-Asiatic Studies (ed
by Nzide Jenner, L.C. Thompson and Stanley Starosta), Pt. 11. Oceanic
Linguistic, Special Publication No. 13. University of Hawaii Press,
Hawaii.
Zide, N.1987 (1984). Robert Colebrook and the first linguistic notice of the
Andamanese Languages. South Asian Review. 8 : 5. (95-100).
Zide, N. and Vishvajit Pandya. 1989. ‘A Bibliographical Introduction to
Andamanese Linguistics’. Journal of the American Oriental Society. 109.4
639-51
= Baber. 6 PeAvereahs pl 2 Duties : i) ee ; msSA Atae ol
A ee
aA ianett. wn cei iT i enarqiattant Niaee a te ~

Ae o's gle hencial er i, *) yoru inv aie), Batt ste) diet ae


a
te) ch copy!aw re, it gual; eesti
te pdietrt
pai jas! = id A One! wee BHP Gs ee a
mest Ny sanivil eerie, of te he aN: st bic

Mie law
. ee

f syn ; : 7. | ae | Giant’
rR Eire 2 Prins rene ace
Hl Sy Flee Atala af Me, wi ote Cling ei ;
ul Wu 2G | Wegget Te » a
i Heertonny: te. ieAe

il Bales eal vw Tite ine ema +


ines aty it Within i yagi , halinon
yi j y i a. 3 ae i) . éarndin
y Jere ?
1 a af ian 1 of pet “ay i ou jared “ean
\ <a 5 es a a
wes oe SSnieu = Vin OL 7 owe .

ind =. ds Fang es tty Lamesa @ Pia OaaEe,


w 1 $~a« ra = wee
a ¥
iis # aN ia atte ploazia iS

* 2 A ‘oh

a
=,
, itsa
“ee
7
+
i Ay A
we
he
Anvita Abbi is Associate Professor of Linguis-
tics at Jawaharlal Nehru University, New Delhi.
Educated at the University of Delhi, Cornell
University, U.S.A., she has widely travelled and
given extensive lectures at the universities in
Europe and America.

An Areal Typologist, she has extensively


worked on tribal languages and lesser known
languages of India. Language areas she has
worked in include Indo Aryan, Dravidian,
Austroasiatic and Tibeto Burman. Her recent
publications include: Semantic Grammar of
Hindi, A Study in Reduplication (1980), Studies
in Bilingualism (1986), Semantic Theories and
Language Teaching ( with V. Prakasam)
(1986), and Reduplication in South Asian
Languages - An Areal, Typological and
Historical Study (1991). She is currently
engaged in the research on Explicator Com-
pound Verbs in South Asian languages. She has
recently edited a special volume of “Language
Sciences’ on India as a Linguistic Area
Revisisted. (Tokyo-Japan).

She has been a Fellow at the Indian Institute of


Advanced Study, Shimla, and a Visiting Fellow
at Osmania University, Hyderabad.

You might also like