Krishnamurti Bhadriraju. The Dravidian Languages

Подождите немного. Документ загружается.

9.2 Simple

sentences

421

section 9.2.5). The latter type is called the dative subject. The Predicate is represented

either by a Verb Phrase (VP) or a NP. Both these may have PPs as complements or

adjuncts. Sentences that have NP predicates are generally equative. The four types are

illustrated below, taking (to the extent possible) one language from each subgroup:

(1) NP + VP

a. Ta. ava

nva-nt-¯an

[he come-past-3m-sg]

‘he (informal) came’

b. Te. w¯a

du wacc-

[he come-past-3m-sg]

‘he (informal) came’

c. Pa. tolen-kul verrar

[brother-pl come-past-3m-pl]

‘the brothers came’

d. Ku

r. tam-bas ke-cc-as

[he-gen-father die-past-3m-sg]

‘his father died’

(2) NP + NP

a. Ta. avar e

n¯aciriyar

[he-hon I-poss teacher-hon]

‘he (polite) is my teacher’

b. Te. w¯aru m¯a guruwu-g¯aru

[he-hon we-gen teacher-hon]

‘he (polite) is my teacher’

c. Kol. inne-tp¯er t¯an-ed

[you-sg-gen name-nom what-nom]

‘what is

your name?

’

d. Malt. ¯ah e

ki baya-h

[he-nom my brother-nom-3m-sg]

‘he (is) my brother’

(3) PP

dat

+ VP

a. Ta. ava

nukku k¯opam va-nt-atu

[he-dat anger come-past-3n-sg]

‘he got angry’

b. Te. w¯a

diki k¯opam wacc-in-di

[he-dat anger come-past-3n-sg]

‘he got angry’

This usage is from Adilabad dialect. The Wardha dialect has a copula anda ‘is’ added to the

NP + NP sentences owing to Marathi inﬂuence (Subrahmanyam 1998: 323).

422 Syntax

c. Kol. an-u

karu va-tt-in

[I-dat hunger come-past-1sg]

‘I am hungry’

d. Br. banda



-as-¯eir¯am¯ar assur

[man-one-dat/acc two son(s) be-past-3pl]

‘a man had two sons’

(4) PP

dat

+ NP

a. Ta. ava

n-ukku oru makan

[he-dat one son]

‘he has a son’

b. Te. w¯a

d-i-ki oka ko

duku

[he-dat one son]

‘he has a son’

c. Pa. ¯on-u

sat-jan cind-ul

[he-dat seven-people-class children]

‘he has seven children’

d. Malt. sardare-k¯ena go

tu g¯aydu

[chief-dat how many c

ow(s)-nom]

‘how many

cows does the chief have?

’

The presence of the four patterns in all four subgroups allows them to be reconstructed

for Proto-Dravidian. The Brahui sentence in (3) should normally belong to (4) but a

copula is used under the

ﬂuence of the surrounding

Indo-Aryan.

There is agreement in gnp between the subject NP and the predicate NP in (2) as in the

case of verbal predications, i.e. NP + VP in (1).

In South Dravidian (SD I and SD II),

for equative sentences, even the ﬁrst- and second-person subjects have agreement in the

predicate NP, not only in number

and person, but also in gender. The verb in NP

+ VP

shows agreement only in number and person and not in gender, when the subject is in

the ﬁrst or second person. These are discussed, in detail, under pronominalized nouns

in section 6.7 and the pattern is reconstructible for Proto-Dravidian. Examples:

(5) a. OTa. y¯am ...¯or ...uyir-am

[we ... one ...breath-1pl]

‘we (are) of one breath’

b. OTa. n¯am n¯a

t-¯om

[we town-adj-1pl]

‘we (are) town-folk’ (n¯a

tu ‘town, country’)

Brahui also has NP predicates, e.g. num¯aˇsahra

t¯ı a

tur¯a/¯o [your village-loc how-many houses-

nom] ‘how many houses (are there) in your village?’ However, verbal predicates are preferred

(Elfenbein 1998: 410–11).

See section 7.5.1 for discussion on the loss of personal sufﬁxes in Middle and Modern Malay¯a

lam.

9.2 Simple

sentences

423

c. MTe. n¯enu manci-w¯a

ni (←w¯a

d-i-ni)

[I good-man-obl-1sg]

‘I am a good man’

d. MTe. n¯enu manci-d¯anni (←d¯an-i-ni)

[I good-she-obl-1sg]

‘I am a good woman’

The 2m sg is (nuwwu) manci-w¯a

d-i-wi ‘you are a good man’, 2f sg (nuwwu) manci-

d¯an-i-wi ‘you are a good woman’, 2pl m¯ıru manic-w¯a

lu ‘you (pl) are good persons’,

3m sg w¯a

du manic-w¯a

du ‘he is a good man’, 3h pl w¯a

lu manci-w¯a

lu ‘they are good

persons’, 3non-m sg ¯ame/adi maici-di ‘she/it is a good one’, 3neu pl awi manci-wi ‘they

(non-hum) are good ones’. Modern Tamil has lost the corresponding constructions.

In equative sentences

with NP predicates there

is no reference to tense and aspect.

Therefore, the implied time reference is either ‘generic’, i.e. a feature habitually ascribed

without reference to a particular time, as in (6), or it refers to the time of the utterance,

as in (7) and (8), e.g.

(6) Te. him¯alay¯alu

anni

parwat¯al-a-kann¯a

c¯al¯a

ettu

[Himalayas all mountains-obl-than much height]

‘Himalayas

(are) much

higher

than

all

the mountains

’ (generic use)

(7) Te. r¯amayya-ku mugguru pillalu

[Ramayya-dat three-classiﬁer children]

‘Ramayya has three children’ (present time)

(8) Te. ¯ame tella-ni telupu

[she white-adj whiteness]

‘she is very fair’ (lit. she is white as whiteness)

When an abstract noun occurs as the predicate NP, gnp agreement with the subject NP

is suspended as in (6) and (8).

9.2.1 Sentence types based on the declarative

With a slight modiﬁcation, we can relate interrogative and imperative sentences to declar-

ative, illustrated in this section. Imperative sentences require the imperative singular

and plural ﬁnite verbs, instead of the tensed ﬁnite verbs of declarative sentences; these

have been discussed in section 7.11.1. There are two kinds of interrogative sentences:

(i) those that require a yes–no answer are formed by adding the interrogative clitic

∗

-¯a (-¯e,-¯o in some languages) to the whole clause or to the constituent that is questioned

(see section 8.4.3), and (ii) those intended to elicit information are framed by replacing

any constituent NP or AdvP by a corresponding question word, e.g.

424 Syntax

(9) a. Ta. kum¯ar

enru-kku varu-v-¯an

[Kumar what-day-dat come-fut-3m-sg]

‘which day

will Kumar come?’

b. Ka. nimma hesaru ¯enu

[you-hon-obl name what]

‘what

is your name?’

c. Te. m¯ı ku enta-mandi

pillalu

[you-dat how-hum-class child-pl]

‘how

many

children do you have?’

d. Ko

da n¯ınu embe

soRi mani

[you-sg where

go-past-2sg be-non-past-2sg]

‘where

have you gone?’

e. Ga. ¯ı kor ¯eyr-ne

[this fowl who-gen]

‘whose

fowl is this?’

f. Malto ¯ek¯a¯al-as bar-c-as

[which man come-past-3m-sg]

‘which man

came?’

g. ¯ek¯a¯al-¯ı

bar-c-a

[which woman come-past-3f-sg]

‘which woman

came?’

Note that the question word occurs in the same position within the clause where the anti-

cipated answer occurs, e.g. for (9c) an answer could be m¯aku mugguru pillau ‘we have

three

children’; enta-mandi ‘how many’ and mu-gguru ‘three (persons)’ occur in the

same position of the respective clauses. Negative sentences occur with a negative ﬁnite

verb, substituting for the afﬁrmative ﬁnite verb in non-past or past (see section 7.10).

9.2.2 Core and peripheral arguments

The ﬁnite verb is an obligatory constituent of a verb-ﬁnal sentence. Since the personal

ending encodes information about the subject, it is possible to use a ﬁnite verb, even

without specifying the subject. A sentence like Te. pani

ant¯a

c¯es-

[work whole

do-past-3m sg] ‘(he) did

the whole

work

’ is well-formed, even without a subject

argument. Similarly, there are no complements that are obligatory, as in English, like

the direct object argument with a transitive verb, e.g. Te. ewaru an-n-¯aru? ‘who said?’

is a well-formed sentence even without the direct object. The criteria used to distin-

guish a complement from an adjunct in terms of core and peripheral arguments are not

adequate for Dravidian (see Matthews 1996 [1981]: 123–45, Dixon and Aikhenvald

2000: 2–4).

9.2 Simple

sentences

425

An intransitive clause has S (= Subject) as the core argument. A transitive clause is

said to have two core arguments: A (subject) and O (object). It is possible also to have

non-normal (non-canonical) case marking of both intransitive and transitive clauses in

Dravidian when the subject is marked for dative and not nominative (see Aikhenvald

et al. 2001: viii). Both intransitive and transitive clauses may have the subject either

in the nominative or in the dative case, e.g. intransitive clause: Te. w¯a

du we

‘he went’ (subject in nominative), Te. w¯a

d-i-ki k¯opam-g¯a undi [he-dat anger-advl be-

3neu-sg] ‘he is angry’(subject in the dative). Transitive clause: w¯a

annam

tin-n-

¯a

[he food-acc eat-past-3m-sg] ‘he

ate

a meal

’ (subject in the nominative and

object in the accusative); Te. w¯a

d-i-ki udy¯ogam dorik-in-di [he-obl-dat job be-found-

past-3neu-sg] ‘he found a job’ (lit. to-him a-job was-found). This can be interpreted as a

transiti

ve sentence with the subject

in the dative and the object

in the nominative. In such

sentences the object NP is also a core argument, since non-mention of the object in the

sentence produces an ungrammatical sentence like

∗

w¯a

d-i-ki dorik-in-di ‘to him (it) was

found’. Similarly, predicates like ‘see, know, like’ have the subject NP in the dative (as

a PP) and the object NP in the nominative; e.g. Te. n¯aku ayskr¯ım i

tam [I-dat ice-cream

likeable] ‘I like ice-cream’, Te. m¯ıku k¯af¯ı k¯aw¯al-

? [you-dat coffee be-wanted-Q] ‘do

you want coffee?’ Dixon and Aikhenvald (2000: 3) set up SE (= extension to core)

as core arguments for such intransitive clauses. A transitive clause has A (subject of

a transitive verb)

and O (object) as core arguments,

sometimes including E (recipient,

beneﬁciary), when the verb is ditransitive (‘give, tell’ etc). Non-canonical marking of

the logical subject by dative and of object by nominative have to be accommodated for

both S and AO clauses (for further examples of dative subject, see section 9.2.5.4 (B)).

9.2.3 Free word order, shift of focus by clefting

Because of rich nominal and verbal morphology expressing grammatical relations (with

case marking in nominals and gnp marking in verbs), the grammatical burden on word

order

is reduced in Dravidian. Therefore, it is possible to change

the unmarked word

order of constituents, keeping the verb in the ﬁnal position, without any erosion to the

semantic structure of the sentence (Krishnamurti and Gwynn 1985: 16–31, Lehmann

1989: 176–80, 368–70)

(10) a. Ta. n¯e

rru

mantiri-avarka

zaint-ai.kku

-p.paric-ai

k.ko

tu-tt¯ar

[yesterday minister-hon-pl child-dat prize-acc give-past-3hon-pl]

‘yesterday

, the minister

gave

the child

a prize

’

The adverb (n¯e

rru

), the subject NP (mantiri-avarka

) and the two PPs (indirect object

zaint-ai-kku

and direct object paric-ai

) can be shifted to any position in the sentence,

keeping the ﬁnite verb (ko

tu-tt-¯ar

) in the sentence-ﬁnal position. Any of the phrases

can be focussed by (i) nominalizing the ﬁnite verb, and (ii) shifting the focussed phrase

426 Syntax

to the ﬁnal (predicate) position; alternatively, if any clitic (emphatic -t¯a

n or -¯e) is added

to any of the constituents, that constituent receives focus without being shifted to the

predicate position or disturbing the nominalized ﬁnite verb. The resultant sentence with

the nominalized verb behaves like an equative sentence.

b. Ta. n¯e

rru

zaint-ai.kku

-p. paric-ai

k.ko

tu-tt-atu

– mantiri.y-

avarka

‘yesterday

, (it was) the minister

(who) gave

the child

a prize

’

c. Ta. n¯e

rru

mantiri.y-avarka

paric-ai

k.ko

tu-tt-atu

– ku

zaint-ai.kku

‘yesterday

, (it was) to the child

(that) the minister

gave

a prize

’

d. Ta. n¯e

rru

mantiri.y-avarka

zaint-ai.kku

k.ko

tu-tt-atu

– paric-ai

‘yesterday

, (it was) a prize

(that) the minister

gave

to the child

’

e. Ta. mantiri.y-avarka

zaint-ai.kku

-p. paric-ai

k.ko

tu-tt-atu

–

n¯e

rru

‘it was yesterday

, that the minister

gave

the child

a prize

’

If any clitic like the emphatic

particle -

t¯a

n is added to any of the constituents, the

nominalized ﬁnite verb need not be shifted from the sentence-ﬁnal position:

f. Ta. n¯e

rru

mantiri.y-avarka

l- t¯an

zaint-ai.kku

- p. paric-ai

tu-tt-

atu

[yesterday minister-hon-pl-emph child-dat prize-acc give-past-3n-sg-

nom]

‘yesterday

, it was the minister

who gave

the child

a prize

’

Malay¯a

lam (Asher and Kumari 1997: 181–2) has a similar usage of clefting for emphasis,

but instead of an emphatic clitic the verb ¯a

‘be-present’ is added to the constituent

focussed. The nominalized ﬁnite verb is not shifted from the sentence-ﬁnal position.

(11) a. Ma. r¯aman innale kr

n-n

pustakam ko

tuttu

[Raman yesterday Krishnan-dat two book(s)-acc give-past]

‘Raman gave Krishnan two books yesterday’

b. Ma. r¯aman-¯a

innale kr

na-nn

pustakam ko

tutt-at

[Raman-it is yesterday Krishnan-to two books give-past-3neusg]

‘it was Raman

that gave two books to Krishnan yesterday’

Similarly, different constituents can be emphasized by adding ¯a

,asininnale-¯a

‘it

was only yesterday ...’, kr

n-n-¯a

‘it was to Krishnan ...’, ra

pustakam-¯a

‘it

was two books that....’

In Kanna

da (Sridhar 1990: 139–40) the ﬁnite verb is nominalized by adding -adu

in the place of the personal sufﬁx and by shifting the focused constituent to the ﬁnal

position as is done in Tamil, e.g.

9.2 Simple

sentences

427

(12) a. Ka. t¯ar¯a

nenne

bar¯o

da-kke

h¯o-d-a

[Tara yesterday Baroda-to go-past-3f-sg]

‘Tara

went

to Baroda

yesterday

’

b. Ka. nenne

bar¯o

da-kke

h¯o-da-ddu

– t¯ar¯a

[yesterday Baroda-dat go-past-3neu-sg Tara]

‘it was Tara that went to Baroda yesterday’

Telugu (Krishnamurti and Gwynn 1985: 31–6), a member of South Dravidian II, also

has such cleft constructions as a means of shift of focus/emphasis within a sentence, e.g.

(13) a. Te. r¯amu

ninna

kamalaku

pustakam

icc-

[Ramu yesterday Kamala-dat book give-past-3m-sg]

‘Ramu

gave

a book

to Kamala

yesterday

’

b. Te. ninna

kamalaku

pustakam

icc-in-di

− r¯amu

[yesterday Kamala-dat book-acc give-past-3neu-sg – Ramu]

‘it was Ramu that gave a book to Kamala yesterday’

c. Te. r¯amu

ninna

pustakam

icc-in-di

– kamalaku

[Ramu yesterday book-acc give-past-3neu-sg – Kamla-dat]

‘it was to Kamala

that Ramu

gave

the book

yesterday

’

Te.

r¯amu

ninna

pustakam

icc-in-di

– kamalaku

k¯adu

[Ramu yesterday book-acc give-past-3neu-sg – Kamala-to be-not-3neu-

sg]

‘it was (not)

to Kamala

that Ramu

gave

the book

yesterday

’

Similarly, the object NP pustakam and the AdvP ninna can also be brought into focus

position. Note that (13b, c) are verbless equative sentences, which can be negated by

adding k¯adu ‘it is not’ at the end of the sentences as in (13d). The position of the

nominalized verb need not be shifted to the penultimate position, if a clitic is added to

any of the constituents as in Tamil, e.g.

e. Te. r¯amu

ninna

kamalak-¯a

pustakam

icc-in-di

‘was it to Kamala

that Ramu

gave

the book

yesterday

?’

Since members of two subgroups SD I and SD II have formation of cleft sentences as

a means of focussing individual constituents in a sentence, this phenomenon could be

taken as an inherited one, which is either lost or not reported from the other subgroups.

The function of nominalizing the ﬁnite verb seems to be to enable any of the arguments to

occupy the focus position without involving agreement. The sentence therefore becomes

equational and not verbal.

An important question is whether free word order in the Dravidian languages is clause-

bounded or not. Mohanan and Mohanan (1994: 161) argue with illustrations from Hindi

428 Syntax

and Malay¯a

lam that ‘dependent’ words of an embedded clause can be interspersed

with ‘dependent’ words in the main clause. They say that scrambling need not be clause-

bounded. Another claim is about the canonical word order for SOV languages as ‘subject

< adjunct < secondary object < object < predicate’ (Mohanan and Mohanan 1994:

170, fn 6ii). While the positions of the subject and predicate are ﬁxed, it is not certain

if the order proposed for the other constituents is canonical in Dravidian. Very often

a time/place adverbial precedes the subject. The so-called unmarked constituent order

and the principles governing topicalization in discourse have to be investigated together

to ﬁnd a solution. In Telugu it is found that direct objects of the verbs in the subordinate

clause and main clause cannot be shifted away from their governing verbs in scrambling,

e.g.

f. Te. ninna amma [pilla-l(a)-ni naww-inc-a

t¯ani-ki] oka katha cepp-in-di

[yesterday mother children-acc laugh-caus-noml-dat one story tell-past-

3fn-sg]

‘yesterday, mother told a story to make the children laugh’



.Te.

∗

amma nawwinca

t¯aniki ninna oka katha pillalni ceppindi

[mother laugh-caus-noml-dat yesterday one story boy-acc tell-past-3fn-

sg]

In (13f



) both the direct objects are shifted away from their respective governing verbs,

nawwinca

t¯aniki ...pillalni, oka katha ...ceppindi, and placing them together in a wrong

order is what makes the sentence ungrammatical. If the direct objects from the main

clause and the subordinate clause remain next to their respective verbs the sentence re-

mains grammatical. This particular phenomenon may indicate a closer cohesion between

transitive verbs and direct objects, which is not true of the other complement phrases

with respect to predicates.

9.2.4 Noun phrase

A noun phrase has a noun as head, optionally preceded by one or more modiﬁers. The

order of such modiﬁers within an adjectival phrase (AdjP) is Determiner + Quantiﬁer

+ Descriptive Adjective + Noun (Head). Even among the descriptive adjectives it is

possible to set up an order on a semantic basis, e.g. value + age + colour + dimension

etc.; alternatively, ‘dimension’ may immediately follow ‘age’ and precede ‘colour’.

There may be as many as six adjectives modifying the head noun, although normally

the number does not exceed one or two. Examples:

(14) a. Ta. anta

m¯unru

periya

‘those

three

big

boxes

’

b. Ma. ¯a

karutta

valiya

pucca

‘that

big

black

cat

’, pa

zaya

valiya

n¯ıla

k¯o

, ‘old

big

blue

coat

’

9.2 Simple

sentences

429

c. Ka. ¯a

era

mays¯uru

mallige

h¯uv-ina

h¯ara-ga

‘those

two

big

Mysore

jasmine

ﬂower-gen

garland pl

’

d. Te. ¯a

manci

kotta

tella

pedda

pustak¯alu

‘those

two

good

new

white

big

books

’

e. Ko

da ru

p¯ulu

‘two

ﬁstfuls

of ﬂowers

’, unri

kota

kilpa

‘one

new

comb

’

Possessive adjectives also occur as modiﬁers as in (14c) above. One or more relative

clauses may also occur as a common type of noun modiﬁer within a NP, e.g.

(15) a. Te. idi

pustakam

‘this

(is) a book

’

b. Te. idi

n¯a

pustakam

‘this

(is) my

book

’

c. Te. idi

n¯enu

r¯asina

pustakam

‘this

(is) a book

that I

wrote

’

d. Te. idi

n¯enu

r¯asina

ta-moda

pustakam

‘this

(is) the very ﬁrst

book

that I

wrote

’

n¯enu r¯as-ina pustakam [I write-past-adj book-nom] ‘the book I wrote’ is a relativiza-

tion of n¯enu pustakam r¯as-

-nu ‘I wrote a book.’ A NP with an embedded relative clause

as a modiﬁer in a nominal predication (15d), in terms of a tree structure, would be:

Pron

adi

AdjP

nenu rasina motta-modati pustakam

A clause can be transformed into a NP by replacing the ﬁnite verb with a verbal noun

(see sections 9.3.2.2, 9.3.3.3). Alternatively, the tense and personal sufﬁxes are replaced

by a nominal sufﬁx, for instance, -a

tam/-a

dam in Telugu (section 9.3.2.5):

(16) a. Te. m¯ıru eppu

du amerik¯a-nunci tirig-i wacc-

-ru?

[you-pl. when America-from

return-ppl come-past-2pl]

‘when did you return from America?’

430 Syntax

b. Te. m¯ıru amerik¯a-nunci tirig-ir¯aw-a

tam eppu

du?

[you-pl America-from return-ppl come-noml when]

‘when (was/would be) your returning from America?’

Although (16b) could be synonymous with (16a), there is no explicit time reference in it;

it can refer to either future time or past time, depending on the discourse context. (16b)

is a nominal predication and an equative sentence, whereas (16a) is a verbal predication

with explicit tense marking. Alternatively, a tense-based nominal may replace a ﬁnite

verb as in (13) above.

9.2.5 Postpositional phrase (PP)

A postpositional phrase consists of NP + case/postpositional marking. A PP occurs as

a complement to the main verb in the predicate phrase. According to some scholars

(Steever 1988, Lehmann 1989: 177) there is no VP with a direct object and the main

verb as immediate constituents in Dravidian, as there is in English. The VP, according

to them, is the main verb, followed by auxiliaries, which carry tense and gnp markers.

In the descriptive grammars of Tamil (Lehmann 1989), Malay¯a

lam (Asher and Kumari

1997), and Kanna

da (Sridhar 1990), NPs followed by postpositions and cases are treated

differently. For instance, they consider a NP in the nominative as the subject, but NPs

inﬂected with non-nominative cases, which function as complements to predicates, are

also treated simply as NPs. Only NPs inﬂected with postpositions are treated as PPs.

Some postpositions are

said to occur after the nominati

ve of the head noun of a NP,

where they are in free variation with case markers, e.g. Ma. t¯ı k o

[ﬁre

take-ppl]

‘with

ﬁre’ is in free variation with t¯ı y-in-¯al, in which -¯al is the instrumental case marker (Asher

and Kumari 1997: 210), e.g.:

(17) a. Ma. ka

lakk¯ar t¯ı ko

til na´sippiccu

[bandit-pl ﬁre-take-ppl hut-acc destroy-past]

‘the bandits destroyed the hut with ﬁre’

b. Ma. ka

lakk¯ar t¯ıi-in-¯al ku

til na´sippiccu

[bandit-pl ﬁre-instr hut-acc destroy-past]

‘the bandits destroyed the hut with ﬁre’

Case markers are sufﬁxes and bound morphemes, whereas postpositions are gram-

maticalized words (nominal or verbal). Some of these are frozen only as postpositions

This claim needs to be examined in the light of sentence (13f, f



) above. In a complex sentence

there seems to be a constraint on the relative distance between a transitive verb and its direct

object in scrambling. The ﬁnite verb governing a PP representing a direct object argument can

be relativized without exception. But it is not true of other arguments within a predicate phrase

(see discussion in section 9.3.2.1 below).