Krishnamurti Bhadriraju. The Dravidian Languages

Подождите немного. Документ загружается.

1.4 Names

of languag

es and geog

raphical distribution

Proto-Dravidian

Proto-South Dravidian Proto-Central Dravidian Proto-North Dravidian

Proto-South Dravidian I

(South Dravidian)

Proto-South Dravidian II

(South-Central Dravidian)

Proto-South Dravidian I

Proto-South Dravidian II

Tamil

Malayalam

Kodagu

Kurumba

Toda

Kota

Badaga

Kannada

Koraga

Tulu

Telugu

Gondi

Konda

Kui

Kuvi

Pengo

Manda

Kolami

Naikri

20a

Naiki

20b

Parji

Ollari

Gadaba

Kurux

Malto

Brahui

Proto-Central Dravidian

Broken lines reflect uncertainty as to a language’s position within the group.

Proto-North Dravidian

Irula

Figure 1.2 Family tree of the Dravidian languages

the remaining non-literary languages, some of which are spoken by pre-literate, tribal

populations.

root of mo

zi ‘word’. The root must be

∗

z- and not

∗

z-. In Malay¯a

lam, one can be sure of

malai ‘mountain’ and ¯a

l ‘man’, i.e. ‘mountain dweller’; alternatively, ¯a

z-am ‘depth, ocean’, ‘the

land between the mountains and the ocean’. Tenungu/telungu with l/n-alternation have several

suggested sources derived from Skt. tri

naga ‘three hills’, trili˙nga ‘three lingas of Siva’, besides

deriving them from

∗

ten- ‘south’.

∗

l-/te

n- ‘sesame’ <

∗

l- meaning the place where oilseeds

grow. Kanna

da is plausibly derivable from kar(u)- ‘black’, n¯a

du ‘country’, i.e. ‘the land of black

soil’.

22 Intr

oduction

1. Tamil (true native name Tami

z; other names: Malab¯ari, Dr¯avi

d¯ı, Tamul, Aravam).

Pop. 53,006,368; 3.35 million in Sri Lanka and over 2 million in South Africa, Malaysia,

Singapore, Mauritius, Fiji and Burma. Cave inscriptions (some seventy-six) in Tamil

Br¯ahm¯ı script were found in Madurai and Tirunalv¯eli districts c. second century BC

(Mahadevan 1971: 83–4). The ﬁrst known work, Tolk¯appiyam, is a treatise on gram-

mar and poetics ascribed to the early pre-Christian era, presupposing a large body of

literature before it, available in the form of anthologies. Although the inﬂuence of early

Sanskrit grammars (ﬁfth century BC) is obvious in certain grammatical concepts like

Tamil k¯alam ‘tense, time’ (Sanskrit k¯ala- ‘time, tense’), Tamil peyar ‘name’ for the

noun (Sanskrit n¯aman- ‘name, noun’), Tamil v¯e

rrumai ‘separation, division’ for ‘case’

(Sanskrit vibhakti- ‘case marker’, literally ‘division’), there is much

that is original in

Tolk¯appiyam.

As in the case of Pre-Modern Greek and Arabic, Tamil has ‘diglossia’ (Ferguson 1964),

which means that the standard written and spoken variety of Tamil, called centami



‘beautiful Tamil’, is based on the classical language of an earlier era and not on any of the

contemporary regional dialects. The spoken variety is called ko

tuntami



z ‘crooked/vulgar

Tamil’ and is not used for formal roles in speech and writing. The newspaper language

and the language of political speeches are ‘high’. The ‘low’ variety is used in conver-

sations and lately for movie dialogues. There are both geographical and social dialects

of the ‘low’ variety. Some of the social dialects

are distinguished on the basis

of caste

(Asher 1982: ix–x). The writing system of the ‘high’ variety is used even for the spoken

‘low’ varieties in plays, etc.; this practice naturally leads to many problems of spelling.

The rate of literacy in Tamil Nadu for all ages in 7+ was 63.7 per cent (1991). There

were 1,863 newspapers published in Tamil in 1999, of which 353 were dailies (India

2001).

2. Malay¯a

lam (Malay¯a

zma). Pop. 30,377,166; west-coast dialect of Tamil till about

the ninth century AD; ofﬁcial language of Kerala state. The V¯a

zappa

li inscrip-

tion of Raja´s¯ekhara of the ninth century AD is considered the earliest document

(Gopalakrishnan 1985:

31). The

ﬁrst literary work

R¯amacaritam (c. twelfth centur

and the ﬁrst grammar, L¯ıl¯atilakam (fourteenth century), written in Sanskrit. More than

Kanna

da and Telugu, and unlike Tamil, Malay¯a

lam has borrowed liberally from Sanskrit

not only words but even inﬂected words and phrases. A new style called Ma

niprav¯a

(diamond and coral) was a literary innovation in Malay¯a

lam, representing a harmo-

nious blend of bh¯a

sa ‘native language’ and samsk

rta ‘Sanskrit’. Kerala has the highest

rate of literacy in India, 90.6 per cent (1991). Because of a high rate of literacy, the

publishing and newspaper industry has been ﬂourishing in Malay¯a

lam. The religious

minorities, though comparatively very high, Muslims (21.3 per cent) and Christians

(20.6 per cent), mostly speak Malay¯a

lam; the linguistic minorities constitute only 5.2 per

1.4 Names

of languag

es and geog

raphical distribution

cent. Malay¯a

lam had 1,373 newspapers (dailies 213) in 1999 (India 2001). Malay¯a

lam

does not have diglossia of the Tamil kind.

There has been no good description of Malay¯a

lam dialects. ‘Malay¯a

lam has many

different regional and social dialects ...since the 1930s or earlier, distinguished creative

writers have in narrative passages deliberately used a language that is close to what they

use in normal conversation ...examples provided aim to be as dialectally neutral as

possible, though there may be a slight bias toward the variety used by educated speakers

in central Kerala’ (Asher and Kumari 1997: xxv). This is what we tentatively take as

Modern Standard Malay¯a

lam.

3. Kanna

da (Kanarese, Canarese, Kar

n¯a

taka). Pop. 32,753,676;the ofﬁcial language of

Karnataka state; the ﬁrst inscription is dated 450 AD by Kadamba K¯akutstha Varma from

Halmi

di, Belur Taluq, Mysore district; the ﬁrst literary work Kavir¯ajam¯arga, a treatise

on poetics, belongs to the ninth century: S’abdama

nidarpa

na is the ﬁrst comprehensive

grammar written in Kanna

da (thirteenth century). Modern Standard Kanna

da is based on

the educated speech of southern Karnataka (Mysore–Bangalore) and differs considerably

from the northern (Dharwar) and coastal varieties. There are also caste dialects reported

within each of the regions. The literacy rate of Karnataka was 56 per cent (1991). There

were 1,561 newspapers published in the Kanna

da language in 1999, including 314 dailies

(India 2001).

4. Telugu (Tel˜ugu, Ten˜ugu,

Andhram, Gentoo, Wa

dugu, Warug¯ı). Pop. 66,017,615;

ofﬁcial language of Andhra Pradesh. Telugu place names occur in Prakrit inscriptions

from the second century AD onwards. The ﬁrst Telugu inscription is dated 575 AD

from Erragu

dip¯a

du of the Ka

dapa district by a prince of the C¯o

da dynasty; the ﬁrst

literary work, a poetic translation of a part of the Mah¯abh¯arata, belongs to the eleventh

century AD. The ﬁrst Telugu grammar written in Sanskrit,

Andhra´sabdacint¯ama

ni,is

said to have been composed by the author of the ﬁrst literary work. There are four re-

gional dialects in Telugu, namely (i) northern, nine Telugu-speaking districts of the old

Nizam’s Dominions, called Telang¯a

n¯a, merged with Andhra Pradesh in 1956, (ii) south-

ern, four southern districts called R¯ayalas¯ıma plus two coastal districts of Nellore and

Prakasham, (iii) eastern, three northeast districts, Visakhpatnam, Vijayanagaram and

Srikakulam, adjoining Orissa, called the Kalinga country, and (iv) four central coastal

districts, Guntur, Krishna, East Godavari and

West Godavari. Modern Standard Telugu

(Krishnamurti and Gwynn 1985) is based on the speech and writings of the elite of the

central coastal dialect (Krishnamurti 1998c: 51–108). In the case of social dialects the

main criterion is the level of education. There are differences between educated and un-

educated speech in phonology, morphology and lexicon (Krishnamurti 1998c: 110–20).

Although it is genetically closer to its northern neighbours, as a literary language

Telugu has a great measure of give and take with Kannna

da; Telugu and Kanna

24 Intr

oduction

have a common stage of evolution in their script called the Telugu–Kanna

da script

(seventh to thirteenth century). There were several Saivite poets who wrote both in

Telugu and Kanna

da. The Vijayanagara King Krishnadevaraya patronized both Kanna

and Telugu poetry. Consequently, there are extensive lexical borrowings between Telugu

and Kanna

da both ways. Literacy in Andhra Pradesh was 45.1 per cent (1991). There

were 1,106 newspapers published in Telugu in 1999, of which 151 were dailies (India

2001).

1.4.2 Minor literary and non-literary languages

I. South Dravidian (SD I) 5. Tu

lu (Tuluva). Pop. 1.6 million; Dakshi

na Kanna

district of Karnataka and the Kasaragode Taluk of Kerala on the west coast. Kanna

script is adopted; the Brahmin dialect is heavily inﬂuenced by Kanna

da. All educated

people are bilingual in Kanna

da which is used for formal communication. The widely

used common Tu

lu is the variety of the non-Brahmin castes. There are two major regional

dialects (north and south) and two major social dialects (Brahmin and Common), which

gether give rise to a four-wa

y dialect division, North Brahmin

(NB), North Common

(NC), South Brahmin (SB), and South Common (SC). The ﬁrst Tu

lu inscription was

dated to the ﬁfteenth century. Two epic poems are said to have been composed in the

seventeenth century, but there has been no continuous literary tradition (Bhat 1998:

158–60). There is now a growing modern literature in Tu

lu.

lu seems to share several phonological, grammatical and lexical features with the

members of the Central Dravidian subgroup, namely Kolami-Parji, etc. An elementary

grammar was pub

lished by Brigel (1872) and a

lu–English Dictionar

ybyM

¨anner

(1866). Bhat (1967) analyses the current language. Tu

lu Lexicon (1987–97), a new

dictionary in six volumes, edited by Haridas Bhat and Upadhyaya, has been published

recently by Rashtrakavi Govinda Pai Samshodhan Kendra, Udupi.

6. Ko

dagu (Coorgi, Ko

dag¨ı, Ko

dava). Pop. 93,000 in the Ko

dagu (Coorg) district

of Karnataka bordering on Kerala. Ko

dagus use Kanna

da as their ofﬁcial language and

as the language of education. A grammar by Cole (1867) and ﬁeldnotes by Emeneau

(1935–8) were the earlier sources. Balakrishnan’s two volumes (1976, 1977) deal with

the phonology, grammar and vocabulary of Ko

dagu. Recently a short descriptive sketch

of ‘Ko

dava’ by Karen Ebert (1996) was published in the series entitled Languages of

the World (Materials 104).

7. Iru

la (Irula,

la). Pop. 5,200. Nilgiri hills. Difﬂoth (1968), Zvelebil (1973, 1979,

1982b), Perialwar (1978a,b).

8. Kur

umba (several dialects). Pop. about 5,000. Nilgiri hills. Kapp (1984, 1987),

Zvelebil (1982a, 1988).

9. Toda (Tuda). Pop. 1,600. Western regions of Nilgiri Hills. Emeneau (1957, 1984).

Earlier accounts are not reliable.

1.4 Names

of languag

es and geog

raphical distribution

10. Kota. Pop. 1,400. Mainly craftsmen among the Nilgiri tribes. Emeneau (1944–6).

11. Ba

daga (Badagu, Vadagu). Pop. 125,000. Nilgiri hills. It was considered a dialect

of Kanna

da after the sixteenth century, but Pilot-Raichoor claims an idependent status

for it as a language. Hockings and Pilot-Raichoor (1992).

12. Koraga. Pop. about 1,000. Basket makers in the South Kanara district. A number

of them are bilingual in Tu

lu. It looks more like an off-shoot of Tu

lu at a recent past,

although Bhat suggests genetic closeness with North Dravidian (Bhat 1971: 3).

Kota, Toda, Iru

la and Kur

umba have preserved the three-way distinction of the coronal

stop consonants, namely t (dental),

t (alveolar) and

t (retroﬂex), which was a feature of

Proto-Dravidian. Toda has the largest number of vowels (14) and consonants (37) which

have developed through numerous sound changes and not through borrowing. Ba

dagas,

supposed to speak a dialect of Kanna

a, moved to the Nilgiri hills

in the sixteenth

century. They are the dominant community both in numbers and in the economy of the

area. A great deal has been published on the languages, geography and ethnography of

the Nilgiris during the past two decades (see Hockings 1989, 1997).

II. South Central Dravidian (SD II) 13. Gondi (native name K¯oytor). Pop. 2,395,507

(includes the Koya dialect spoken in Andhra Pradesh); it has many dialects scattered over

four neighbouring states, Madhya Pradesh, Maharashtra, Orissa and Andhra Pradesh.

The main dialect division is between west, north and northwest, on the one hand, and

south and southeast, on the other. Some of these dialects are probably mutually unintelli-

gible, particularly Maria Gondi and Koya in the south and southeast to the speakers of the

other dialects. The earliest writings include Driberg and Harrison (1849), Hislop (1866)

and Williamson (1890). In the twentieth century Lind (1913), Trench (1919, 1921),

Mitchell (1942), Moss (1950), Subrahmanyam (1968b), Tyler (1969) and Natarajan

(1985). The last three descriptive grammars cover different dialects. Several PhD dis-

sertations from Indian universities treat various aspects of Gondi grammar. Burrow

and Bhattacharya (1960) is the main source of comparative vocabulary drawing on the

sources available up to that period.

14. K¯ui (K¯uinga, Kandh, Khond, K¯odu). Pop. 641,662;spoken in Ganjam and Phulbani

districts of Orissa. Census reports confuse Kui and Kuvi, both of which are called Khond

or Kandh. Lingum Letchmajee (1853) wrote the ﬁrst grammar. Winﬁeld’s grammar

and vocabulary (1928, 1929)

are still the main source of information on grammar and

vocabulary. Winﬁeld (1928: 226–9) discusses the history and etymology of the names

of the tribes and languages 14 and 15.

15. K¯uvi (K¯uvinga, Khond, Kondh, K¯odu, K¯oju, S¯amantu, J¯at¯apu). Pop. 246,513.

Spoken in the districts of Ganjam, Kalahandi and Koraput of Orissa, Visakhapatnam

and Srikakulam of Andhra Pradesh. Most published sources are neither comprehensive

nor reliable. Schultze (1911, 1913), Fitzgerald (1913). The most recent is Israel (1979).

26 Intr

oduction

16. Ko

da (Ko

da Dora, K¯ubi, K¯ubi

ŋ). Pop. 17,864; mainly spoken in the hills of the

northeastern districts of Andhra Pradesh, and is linguistically closer to Telugu than Kui

or Kuvi. Krishnamurti (1969a) is the main source.

17. Pengo. Pop. 1,300. It is spoken in the Navrangpur district of Orissa. Burrow and

Bhattacharya (1970) is the only source for Pengo.

18. Man

da. Pop. not known; spoken near Thuamul Rampur of the Navrangapur dis-

trict of Orissa; closely related to Pengo. A grammatical sketch by Burrow (1976) and

the vocabulary available from DEDR (1984) based on the ﬁeldnotes of Burrow and

Bhattacharya.

III. Central Dravidian 19. Kolami (K¯ol¯am¯ı). Pop. 99,281. Adilabad district of Andhra

Pradesh, Yeotmal and Wardha districts of Maharashtra. Main sources Sethumadhava Rao

(1950), Emeneau (1955b). It has borrowings from Telugu from a very early period.

20a. Naik

ri. Pop. 1,500 (1961). Spoken in Andhra Pradesh and Maharashtra in the

vicinity of Kolami. Burrow and Bhattacharya considered this a dialect of Kolami without

any supporting arguments. I am considering Naik

ri and Naiki of Chanda as related

languages. Thomasiah’s (1986) unpublished PhD thesis is the main source, besides the

cognate list in DEDR.

20b. Naiki (Chanda). Pop. (?)

54,000. Chanda district of Madhya

Pradesh.

Bhattachar

’s

article (1961) is the main source.

21. Parji (Poroja, Dhurwa). Pop. 44,001 in the Bastar district of Madhya Pradesh

and the adjacent hills of the Koraput district of Orissa. Burrow and Bhattacharya (1953).

22. Ollari (Hallari). Pop. 9,100. Burrow and Emeneau treat Ollari; and Gadaba as

dialects of the same language, but I have kept them apart as languages. Bhattacharya’s

monograph (1957) is the only source.

23. Gadaba (Kon

dekor Gadaba). Pop. 9,197 in Census 1981; in 1991 ?54,000.

Srikakulam District of Andhra Pradesh and Koraput district of Orissa. There is a

Mundarian Gadaba called Gutob Gadaba. The Census reports do not distinguish the

two although they belong to different families. If we take only the population ﬁgures of

Andhra Pradesh they number less than 10,000. Fieldnotes of Krishnamurti of the 1950s

included in DED (1961) and P. Bhaskararao (1980).

IV. North Dravidian 24. Ku

rux (Kurukh

r¯a¯o

). Pop. 1,426,618. Bhagalpur

and

Chota Nagpur districts of Bihar, Raygarh, Sarguja districts of Madhya Pradesh,

Sundargarh, Sambalpur districts of Orissa. Ku

rux is in contact with both Indo-Aryan and

Munda languages. There is a dialect of Ku

rux, called Dhangar, spoken by 10,000 per-

sons in Nepal. The earliest grammar is by Hahn (1911). The main sources are Grignard’s

grammar (1924a) and dictionary (1924b).

1.5 T

ypological featur

es of Dravidian

25. Malto (R¯ajmah¯al¯ı, M¯aler¯ı). Pop. 108,148. Spoken in the Rajmahal hills in the

Santal Paraganas (northeast) of Bihar. Not geographically adjacent to Ku

rux now.

Droese (1884), Mahapatra (1979).

26. Brahui (Br¯aʔ¯u¯ı, Br¯ah¯u¯ı). Pop. 1.7 million (International Encyclopedia of Ling-

uistics 1991); 1.4 million (estimate by Breton 1994: 204). Spoken in Baluchistan of

Pakistan. Bray (1909, 1934), Emeneau (1962d), Elfenbein (1997, 1998). This is the

farthest removed of the Dravidian languages. Some writers considered Brahui as the

ﬁrst column (?) of speakers to branch off in the third millennium BC. Since it does not

retain any archaic features of Proto-Dravidian, it is likely that the speakers of Brahui

had migrated westward from the mainland where they were together with the speakers

of Ku

rux and Malto. Some sound changes shared by these three languages suggest a

common undivided stage deeper in history. Surrounded by Indic and Iranian languages

for many centuries, Brahui is said to have only 10 per cent of Dravidian words, 20 per

cent Indo-Aryan, 20 per cent from Balochi, 30 per cent from Perso-Arabic and 20 per

cent of unknown origin (Elfenbein 1997: 810).

For a better idea of the geographical location of the Dravidian languages, see map 1.1.

A number of other Dravidian languages are listed in the Encyclopedia of Linguistics

(1991) and most recently by Zvelebil in an article (1997). Most of the names represent

dialects of the main languages listed above. Koraga (Bhat 1971) is almost like Tu

lu in

most respects and is tentatively shown as an off-shoot of Pre-Tu

lu.

1.5 Typological features of the Dravidian languages

1.5.1 Phonology

There are ﬁve short and ﬁve long vowels in Dravidian /ieaou¯ı¯e¯a¯o¯u/; the long vowels are

indicated with a suprasegmental phoneme of length [], marked by a macron here. Only

some of the languages of the Nilgiris and Ko

dagu have phonemic centralized vowels /¨ı¨e/,

which have developed from retracted allophones of the front vowels before retroﬂex con-

sonants. Only Toda and Iru

la have also developed front rounded vowels /¨u¨o/ (see chap-

ters 3 and 4 for descriptive and historical details). The favoured syllable pattern of the

word (free form) in Dravidian is (C)

VCV/(C)VCCV/(C)VCVCV (three morae each).The

other types are infrequent. There is no phonemic stress in any of the Dravidian languages.

Normally stress (loudness) falls on the initial syllable of a word, short or long. A non-

initial long vowel is louder than short vowels. Words can begin with vowels or consonants.

There are seventeen consonantal segments in Proto-Dravidian, six stops, four nasals,

two laterals, one trill, one approximant (frictionless continuant) and three semivowels

including a laryngeal, which patterns with semivowels. Voicing and aspiration are not

phonemic. The three-way distinction, dental-alveolar-retroﬂex /t t

t/ in the stop series, a

separate series of phonemic retroﬂexes with different articulatory effor

z/ (stop,

28 Intr

oduction

nasal, lateral, approximant), absence of voice contrast in the stop series are the typo-

logically important features of the Proto-Dravidian consonantal system. Alveolars and

retroﬂexes occurred only in the medial position in Proto-Dravidian. Voicing was sub-

phonemic to start with, but it became phonemic in most of the languages, due to internal

changes and borrowing from Indo-Aryan. Consonant clusters occur non-initially, mainly

geminates or nasal + stop series. Qualitative changes of segments in radical syllables

are not common. Short vowels in non-root syllables (in unaccented position) tend to be

lost. Among apical consonants alveolar

∗

t, and the retroﬂex approximant

∗

z are the most

marked segments.

1.5.2 Morphology

The Dravidian languages are agglutinating in structure. There are no preﬁxes or inﬁxes.

Grammatical relations are expressed only by sufﬁxation and compounding. A Dravidian

root is (C)V, or (C)VC in which the V can be short or long in the case of verbs. These

are followed by sufﬁxes, which originally denoted tense–voice contrasts; a number of

languages have lost their original meaning and they have become only voice markers in

some, or mere formatives (without any discernible meaning) in others (see chapter 5).

These are NP (nasal + stop) in intransitive or NPP (nasal + stop + stop in transitive)

following (C)V-roots; a V

(iau) is inserted when these follow (C)VC-roots. Many

other sufﬁxes are added to the roots and stems described above in different languages to

form words.

1.5.2.1 Nouns, adjectives

Nominals (including nouns, pronouns, numerals and adverbs of time and place) are all

inﬂected for case. In several classes of nominals the nominative stem is the basic form.

An oblique stem, which occurs before case sufﬁxes, is different from the nominative;

a series of oblique sufﬁxes is added to form it. In most languages, the genitive and the

oblique stems are identical. Case relations are expressed by bound morphs (accusative,

dative, genitive, instr

umental/locative) or by in

ﬂected words, which have become

gram-

maticalized, e.g. instrumental, sociative, ablative, some genitive and locative. This latter

set are known as postpositions. Different Dravidian languages have developed indepen-

dent postpositions meaning ‘by means of, from (being in a place/time), before, after,

above, below, up to, until, through’ etc. Case markers and postpositions are added to the

oblique stems of nouns, in the singular and in the plural.

Gender and number are interrelated categories. Singular (unmarked) and plural

(marked) are the numbers denoting ‘one’ and ‘more than one’. The plural is differen-

tiated originally between human and non-human categories. The categories ±Animate,

±Human, ±Male human underlie gender classiﬁcation, which is mainly based on mean-

ing and not on form. There are two plurals of the ﬁrst-person pronoun, one including

1.5 T

ypological featur

es of Dravidian

the person addressed (inclusive) and the other excluding the person addressed (exclu-

sive). Adverbial nouns have no gender or number; personal pronouns (ﬁrst and second)

are distinguished for number and not gender. Gender and number are relevant only in

the third person. Gender-number-person agreement is expressed by ﬁnite verbs. An in-

ﬂected noun has the structure Stem + number + oblique + case. By adding personal

sufﬁxes to nouns and adjectives, pronominalized nouns (Caldwell’s ‘appelative verbs or

conjugated nouns’) are formed.

Adjectives precede the noun head that they qualify. Adjectives do not agree with

the noun head in gender and number. Adjectives are mainly a syntactic class. Numer-

als and adverbs of time and place can become adjectives by sufﬁxation. There is no

morphological device for comparative and superlative degrees.

1.5.2.2 Verbs, adverbs

The verb in Dravidian is ﬁnite or non-ﬁnite. The ﬁnite verb has the structure Stem

(Root + (transitive) + (causative)) + Tense + Gender–Number–Person (gnp). A stem

can be complex (as

above) or compound. A compound

stem has one or more coverbs

attached to an uninﬂected noun, or an inﬂected (inﬁnitive or perfective) main verb; in

the latter case, the coverb is either an explicator/operator/vector type (lexical) or a modal

auxiliary (grammatical) (section 7.15). Basically there are two tenses in Dravidian, past

and non-past. A ﬁnite verb may carry negation as part of its tense category (mainly

non-past). A non-ﬁnite verb heads subordinate clauses, like the durative, perfective,

conditional, concessive, etc. Certain ﬁnite verbs may not carry agreement in gnp.

Adverbs of time and place function as nominals morphologically. Only manner ad-

verbs derived by adding an inﬂected form of the verb ‘to be’ or iteratives (ga

naga

‘ring of bells’) are true adverbs. Adverbs precede verbs syntactically.

There are clitics (grammaticalized words) which are added to autonomous syntactic

units – words, phrases, clauses – to signal a variety of meanings, tag questions, emphasis,

interrogation, surprise, doubt, etc.

1.5.3 Syntax

The Dravidian languages are of the OV type, head-ﬁnal and left-branching. A simple

sentence consists of a subject and a predicate. The subject argument is generally ex-

pressed by a noun phrase

(NP), but a postpositional or casal phrase with the head

nominal in the dative case can also function as the subject. The latter is called a dative

subject sentence. The predicate has either a verb or a nominal as head. Sentences with

nominal predicates are equative sentences, which lack the copula or the verb ‘to be’ in

most of the languages. A noun phrase (NP) has a nominal head, optionally preceded by

one or more adjectives and/or a relative clause, in a ﬁxed order. A verb phrase (VP) has a

verb as head, optionally preceded by one or more complements, carrying different case

30 Intr

oduction

markers or postpositions, encoding different grammatical relations with the verb. I call

these postpositional phrases (PP = NP + case/postposition). A subordinate clause can

be either verbal (with a non-ﬁnite verb as head) or nominal, i.e. a pronominalized verb

or a relative clause with a noun head.

Interrogative sentences are formed either by the addition of an interrogative particle

(yes–no type) or by using an interrogative word substituted for the questioned noun.

Any head nominal can be questioned whether it is a constituent of the main clause or

subordinate clause.

Nominal and verbal predicates have different negative words to express sentence

negation. A negative word is an inﬂected verb meaning ‘to be’ or ‘to be not’. Non-ﬁnite

verbs, which head subordinate clauses, have afﬁrmative and negative counterparts.

The argument NPs which occur as complements to a verb derive from the semantic

structure of the verb; for instance, an intransitive verb requires only one argument, Agent/

Object/Experiencer/Source. A transitive verb requires an Agent + Object (Patient); a

causative verb, Agent (causer) + Agent (causee) + Instrument + Object. Reﬂexivity

and reciprocity are expressed in the verb by an auxiliary or by reﬂexive pronouns. The

passive voice is rarely used in modern Dravidian languages.

A quotative clause is embedded in the main clause as NP by an inﬂected complemen-

tizer of the verb ‘to say’. The difference between direct and indirect (reported) speech is

sometimes noticed in the change of the subject/object pronoun of the quoted clause. The

quotative particle is used to signal a variety of meanings including ‘like, because’, etc.

1.6 Dravidian studies, past and present

1.6.1 1856–1950

During the nineteenth century, after Caldwell’s work (1856), there were grammars and

bilingual dictionaries of the major Dravidian languages prepared by missionaries and

administrative ofﬁcials. Some of these became standard reference works (see section 1.3

above). Among the minor languages, M¨anner’s dictionary (1886) and Droese’s grammar

and vocabulary lasted as standard writings into the twentieth century. The Linguistic

Survey of India edited by Sir George A. Grierson has brought

out short accounts of the

Dravidian languages (1906: LSI IV) outside southern India, which was not covered by

the Survey.

Traditional scholars of the major languages took for granted the Sanskrit

origin of the Dravidian languages and this fact was presumably responsible for the long

spell of disinterest in pursuing Caldwell’s ground-breaking study for nearly ﬁve decades.

What follows is a summary of three survey articles that I have published (see Krishnamurti

1969b, 1985, 2001b: ch. 21).

The Madras Presidency, the princely states of Mysore, Travancore and Cochin, Coorg and the

Nizam’s Dominions (LSI 1: 1. 25 [1927]).