Krishnamurti Bhadriraju. The Dravidian Languages

Подождите немного. Документ загружается.

7.10

Negation in

ﬁnite

and non-

ﬁnite

verbs

351

The above data show that in South Dravidian I the negative sufﬁx occurring in ﬁnite

verb inﬂection had at least ﬁve allomorphs ¯ay/ay∼ ¯a/a∼Ø. The verbal adjective and ad-

verb have presumably incorporated an erstwhile past allomorph

∗

t∼

∗

tt with neutralized

tense meaning. In -¯ay-/-ay- the -y element is not a glide created in sandhi (contrary to

Subrahmanyam’s suggestion in 1971: 348, fn.), because its occurrence is not predictable

in terms of the preceding and following segments, vowels or consonants. To account for

¯a/a and also the y element, we need to set up for Proto-South Dravidian

∗

aH (or more

legitimately

∗

¯aH ) as the negative marker.

7.10.2 South Dravidian II

The South-Central Dravidian data are examined below.

Telugu (Old and Modern), e.g. cepp- ‘to tell’: 1sg cepp-a-nu ‘I do not tell’. There is

no change before other personal sufﬁxes; cepp-a-ni adj ‘that ...not told’ (all tenses),

cepp-aka adv ‘without telling’. In Modern Telugu another auxiliary u

d-an (inf of u

‘to be’) is added to the negative adverb in -aka, e.g. ceppa-ak-u

d-¯a ‘without telling’.

It is to be explored if -k-in-aka- is a reﬂex of an old laryngeal. The negative abstract

noun is formed by adding -a-mi (<

∗

-aH-may), cepp-a-mi ‘not telling’.

In Gondi the negative marker is

∗

-w- before personal sufﬁxes beginning with a front

vowel, and -Ø- before a rounded vowel -o (-wo < -wa), e.g. tin- ‘to eat’:1sg tin(n)-Ø-

on (<

∗

tin-w-on <

∗

tin-w-an, with loss of -w- before a labial vowel), ‘I do not eat’,

1pl tin(n)-Ø-om, 2sg tin-w-i/-in, 2pl tin-w-i

t/-ir,3msgtin(n)-Ø-ol/-or/-o

r/-on/-on

3m pl tin(n)-Ø-or/-o

r, 3n-m sg tin(n)-Ø-o/-oye, 3n-m pl tin(n)-Ø-o

/-oku. The negative

adverb, e.g. s¯u

r-v¯ak ‘without seeing’, and the negative conditional, e.g. veh-v-¯ek¯e ‘if

one does/will not tell’ have -v- in Gondi dialects as the negative marker.

da negative marker in verb inﬂection is /ʔ/. In the Gu

ri dialect of Ko

da /v/ occurs

instead, which should have been the primary form of negation, e.g. 1sg ki-

e ‘I will not

do’ (Gu

ri dialect ki-v-e etc.), 1pl (excl) ki-

-ep, 1pl (incl) ki

t, 2sg ki-

i, 2pl ki-

-ider,

3m sg ki-

-en,3mplki-

-er, 3n-m sg ki-

-ed, 3n-m pl ki-

-u. In all non-ﬁnite forms

involving the negative, negation is indicated by -

-, e.g. ki-

-i (non-past adj), ki-

i-t-i

(past adj), ki-

-e

da (neg adv) ‘without doing’.

Kui has almost the same type of conjugation as Ko

da with -

- occurring as a negative

marker in non-past, e.g. tin- ‘to eat’: 1sg tin-

enu ‘I do not eat’, 1pl (excl) tin-

-amu, 1pl

(incl) tin-

-asu, 2sg tin-

-ai, 2pl tin-

-eru,3msgtin-

-enju,3mpltin-

-eru, 3n-m sg

tin-

-e, 3n-m pl tin-

-o. In the negative past (which will be discussed later), the marker is

a-, e.g. 1sg tin-

a-t-enu (stem-neg-past suff-pers) ‘I did not eat’, tin-

a-n-i (non-past

neg adj), tin-

a-t-i (past neg adj). The negative adverb is formed by adding -arnge/-araa,

e.g. s¯u

r- ‘to see’: s¯u

r-arange/-araa ‘not having seen’.

Kuvi has parallel constructions, except that -a- is replaced by -o- owing to the inﬂuence

of Oriya, e.g. p¯ay- ‘to beat’: 1sg p¯ay-

o-Ø, 1pl (excl) p¯ay-

o-mi, 1pl (incl) p¯ay-

o-hi

352 The verb

‘you do not beat’, 2sg p¯ay-

o-di, 2pl p¯ay-

o-deri, 3m sg p¯ay-

o-si,3mplp¯ay-

o-ri,

3n-m sg p¯ay-

-e, 3n-m pl p¯ay-

-u. The negative past is like p¯ay-

a-t-i ‘you did not beat’,

p¯ay-

a (non-past neg adj), p¯ay-

a-t-i (past neg adj) and p¯ay

a-naha ‘not having beaten’

(neg adv), and the negative verbal noun, p¯ay-

a-tayi ‘not beating’.

Pengo non-past negative inﬂection has -u-(-vu-) as the negative marker, but when

followed by an unrounded vowel, -v- surfaces, e.g. hu

r- ‘to see’: 1sg hu

r-u-

∗

vu-

) ‘I do not see’, 1pl (excl) hu

r-u-p, 1pl (incl) hu

r-u-s, 2sg hu

r-u-y, 2pl hu

r-u-der,3m

sg hu

r-u-n, 3m pl hu

r-u-r, 3n-m sg hu

r-u-t, 3n-m pl hu

r-u-

. After vowel-ending roots

the second person singular is realized as -vi-, e.g. o- ‘to take’: o-v-i ‘you will not take’.

The past negative has forms like hu

r-va-t-a

’I did not see,’ and the negative adjective

r-v-i (non-past), hu

r-vi-t-i (past). The negative adverb has a complex -va-da

added

to the stem, ta-va-da

‘without bringing’.

For South-Central Dravidian (other than Telugu), we need to reconstruct

∗

-wa-asthe

basic negative morpheme.

In Ko

da–Kui–Kuvi

∗

-w- became a glottal stop -

-, and the

following vowel has vanished in Ko

da and Kui, but is preserved in different shapes in

Gondi, Kuvi and Pengo. It appears that we can reconstruct

∗

-Ha- (a part of

∗

aH-aH) as

the negative morpheme for Proto-South-Central Dravidian with

∗

H developing into

∗

-w

in prevocalic position.

7.10.3 Central Dravidian

The data from Central Dravidian can be explained in terms of PD

∗

-aH- as the negative

marker. The reﬂex of

∗

-H - is lost except in Parji where -ay- is derived from an older -aH-.

Kol. s¯ı-e-n ‘I do not give’, s¯ı-e-t-an ‘I did not give’. It has several negative non-ﬁnite

verbs by adding the inﬂected forms of t¯ot- in negation, e.g. present durative participle

and the negative of t¯ot-, e.g. v¯at¯oten ‘I am not coming/will not come’; the past durative

negative is formed by present participle + past negative forms of t¯ot-, e.g. anu

vessa

v¯a tottin

‘fever

would not have come

to me

’; the perfective negative is formed by

adding the negative forms of t¯oten to the perfective participle, ¯an adn tin-t tottantiri

‘although I had not eaten’; the future negative is formed by adding the negative forms

of t¯ot- to the future participle, e.g. varak t¯oten ‘I will not come’; the negative participle

is formed by adding sel-(< ?

∗

< cil- ‘to be not’) to the verb stem, tin-sel ‘not having

eaten’ (DVM: 366–7); the negative adjective is formed by simply adding the negative

marker -e to the stem, tin-e ‘not eating/eaten’ (neg adj in all tenses).

Pa. c¯u

r-a-n-a ‘I do not see’ (the ﬁnal -a- is idiosyncratically a copy of the neg.-a-

repeated after the personal sufﬁx), c¯u

r-ay-Ø-a (3n-m sg) ‘it does not see’ (note -ay-

instead of -a-), c¯u

r-a (non-past neg adj), c¯u

r-aka (adv) ‘without seeing’, Oll. s¯u

r-a-n ‘I

will not see’, s¯u

r-a ‘that ...not see’ (all tenses), s¯u

r-a kerin (adv) ‘without seeing’.

The Ollari past negative is formed by adding the past-tense forms of the archaic v

erb

∗

l- to the negative participle, e.g. man- ‘to be’: man-(a) u

ton ‘I did not stay’; the past

7.10

Negation in

ﬁnite

and non-

ﬁnite

verbs

353

negative durative is formed by adding the past-tense form of man- ‘to be’ with a main verb

in the non-past negative, e.g. ¯ıl- to fall’: ¯ıl-a-n-ima-

t-on ‘I was not falling’. The negative

participle is formed by adding the negative marker to the verb stem, e.g. pun- ‘to know’:

pun-a ‘not knowing’. The present, perfective and future are formed periphrastically by

adding the negative inﬂected forms of man- ‘to be’ in past and non-past to the participles

of the present, past or future, e.g. s¯u

r-i manan ‘I have not seen’, s¯u

r-i man-u

ton ‘I had

not seen’.

7.10.4 North Dravidian

Among the North Dravidian languages Ku

rux and Malto have no negative inﬂectional

sufﬁx. Brahui has a/Ø in negative verbs, preceded by a tense-marking consonant. Br.

tix-p-a-r ‘I do not place’, tix-t-ava

t ‘I did not place’ (past stem followed by the negative

present form of the substantive verb affa

t → ava

t ‘I am not’).

7.10.5 Summary

Subrahmanyam

(1971:

§4.22–4) reconstr

ucts

∗

-¯a- for Proto-Dravidian. He says it was

preserved only in the third neuter in South Dravidian and in the other persons it became

a zero; its shortened form -a- occurred in Telugu, Tu

lu, Central Dravidian and Brahui.

He does not explain the widely distributed allomorphs -¯ay-/-ay- and also how -¯a- could

be totally lost in most of the inﬂectional forms. For South-Central Dravidian, he recon-

structs

∗

-v¯a- (because some Gondi dialects

have a long vowel) and says, ‘At present it is

difﬁcult to explain the additional v in the sufﬁx’ (1971: 387). Alfred Master (1947: 146)

cites a 1935 article of Jules Bloch (BSL no.107, p.35), in which Bloch suggested that

the negative sufﬁx -a- ‘was preceded prehistorically by an ill-determined consonantal

element, possibly guttural, laryngeal or glottal’. It is not clear why Bloch suggested this

‘element’ preceded rather than followed

∗

-a-. He must have thought so because of the

developments in South-Central Dravidian and Brahui. He guessed correctly that Kui

‘glottal stop should be the remnant of an old consonantal articulation’ (Bloch 1954: 67).

Taking the total scenario into account, we can reconstruct

∗

-aHa- (or

∗

-aHaH- > -¯ay-/

-ay-) for Early Proto-Dravidian or Pre-Dravidian which developed into a long grade -¯a-

(by contraction), a short grade

∗

-a- by loss of

∗

H and the zero grade by the loss of

this vowel before personal sufﬁxes beginning with vowels. In some of the languages

(Malay¯a

lam, Kota, Tu

lu and Parji) the morpheme-ﬁnal

∗

H was further softened into

a semivowel -y/-w. These developments explain the entire data of South Dravidian,

lu and Central Dravidian, Telugu of South-Central Dravidian and Brahui of North

Dravidian. Telugu, being geographically close to the South Dravidian literary languages,

had followed the southern pattern by selecting -a- in negative inﬂection.

How can we relate the Proto-South-Central Dravidian reconstruction

∗

-wa-toPD

∗

-aHa(H )-? Three possibilities are suggested: (a) the South-Central Dravidian

∗

-wa-was

354 The verb

an independent innovation not related to PD

∗

-aHa(H )-; (b) Early PD

∗

-aHa(H )- split

into

∗

-¯aH- in Proto-South Dravidian and to

∗

-Ha(H )- in Proto-South-Central Dravidian,

as an innovation that separated these two branches, by a process of contraction and trun-

cation (reducing the morpheme size with loss of -a- in unaccented positions). Then, the

other subgroups (Central Dravidian and North Dravidian) require minimally a proto-

form

∗

-aH- which must have developed on the same lines as SD -aH- >

∗

-aØ- > -Ø[V-;

∗

aHa(H )- in all subgroups except Proto-South-

Central Dravidian, either in its full form as

∗

-aHa(H )- or its contracted form as

∗

-¯aH-,

because both these could lead to

∗

-¯ay-/

∗

-ay- > -¯a-/-a- > -Ø-. In that case, only South-

Central Dravidian innovated

∗

-Ha(H )- by a rule of truncation;

∗

-H - developed to

∗

-w-

originally in

intervocalic or prevocalic position,

later generalized to all

positions; a

further change of

∗

-wa- > -ʔa- occurred in Ko

da–Kui–Kuvi as a shared isogloss.

da–Kui further reduced it to -

- in non-past paradigms, but the trace of the vowel is

preserved in the past negative, e.g. Ko

da ki-

-en ‘he does/will not do’, ki-

e-t-an ‘he

did not do’, ki-

i-t-i ‘you (sg) did not do’. Alternative (a) is ruled out because of the

widespread phonetic, semantic and grammatical similarity of the relevant allomorphs

in different subgroups. Alternative (b) is less likely, because the developments of the

reconstructed morpheme

have to be replicated in different

subgroups independently.

Therefore, alternative (c) seems closer to the truth. I believe that the process of con-

traction

∗

-aHa(H )- >

∗

-¯aH- is not as atypical a sound change as truncation

∗

-aHa(H )-

∗

-Ha(H )-.

Therefore, only Proto-South-Central

Dravidian could have inno

vated the

latter type of change. In any case, the involvement of a laryngeal

∗

H in negative v

erb

inﬂection, as already guessed by Jules Bloch over sixty years ago, would account for the

data better than any of the earlier explanations.

7.10.6 Negation by verbs

∗

al- and

∗

cil- etc.

A peculiarity of the Dravidian languages is the use of a basic verb meaning ‘not to be’,

∗

cil-, which is found in all subgroups. It occurs both as the main verb and as an auxiliary.

In South Dravidian I it became il- (through earlier hil- and sil-). Another verb al- ‘to be

not so-and-so’ occurs mainly in Old Tamil and some other languages of South Dravidian

I with approximately the same meaning.

7.10.6.1 South Dravidian I

In Old Tamil the inﬂected forms of al- ‘to be not’ are added to verb stems to express

negation in non-past; the inﬂected forms of il- ‘to be not’ (<

∗

cil-) are added to the past

participle or the past participial noun to express past negative. In later Ca˙nkam texts there

are also usages of adding the al- forms to bare (untensed) verbal stems, e.g. cel- ‘to go’:

cell-al-am ‘we will not go’, ko

l- ‘to receive’: ko

l-al-ir ‘you (pl) will not receive’, nill-

al-a

n ‘he will not stand’. These forms appear as though they ﬁt into the pattern: stem +

tense + (g)np; in that case, we need to treat -al- as negative-cum-tense marker like -a-,

7.10

Negation in

ﬁnite

and non-

ﬁnite

verbs

355

but this argument is not sustainable, since the ﬁnite forms of al also occur after other

ﬁnite verbs in Old Tamil texts, e.g. kalank-i

n-¯en

all-ø-¯en

[I was disturbed

, not

]‘Iwas

not disturbed’, cel-v-¯em

all-Ø-¯em

‘ we will

not

’. Steever (1988: 42–4) proposes

that forms like cell-al-am have resulted from the telescoping of two ﬁnite verbs, i.e. cel-

v-¯em all-¯em. Something of the kind has happened in South-Central Dravidian but we are

not sure if this is also replicated in South Dravidian I. Note that the inﬂected al- is also

added to pronominalized nouns, e.g. pe

t-ir-¯em

all-Ø-¯em

‘we are not

women

’. The

inﬂected forms of il-, in addition to past participles, are also used with pronominalized

nouns, e.g. va-nt(u)il-¯ar ‘they (h) did not come’, a

ri-nt(u) il-ir ‘you did not know’.

Verbal nouns are also followed by inﬂected il- forms, e.g. a

ri-nt-at(u) il-¯en (lit. ‘knowing

(past)-not-I’) ‘I did not know’. In Modern T

amil

illai ‘it is not’ is added to the

ﬁnitive

or the noun formed on the past participle to express past negative, e.g. var- ‘to come’:

var-a (v) illai, va-nt-atu illai ‘(one) did not come’. Negation in the non-past is rendered

by adding illai to the nominalized verbs in non-past, e.g. varu-ki

r-atu illai [come-dur

sufﬁx-3neu sg not] ‘(one) does not come’, varu-v-atu illai [come-non-past-3neu sg not]

‘(one) will not come’.

In Malay¯a

lam illa ‘it is not’ is added to the non-ﬁnite verbs (past and

present participles

and the inﬁnitive) to express

negation in different tenses,

e.g.

cey-tu illa ‘(one) did not

do’, ceyy-unn(u) illa ‘(one) is not doing’, ceyy-uka/ceyy-(a) illa ‘(one) will not do’.

The Ko

dagu past negative is formed by adding ¨ıle to the past stem of a verb, e.g.

ba-nd-¨ıle ‘(one) did not come’. A past perfect negative was innovated by adding ille to a

stem with double past marking, e.g. ba-nd-it(i) ille ‘(one) has not come’. The non-past

negative is formed by adding -¨ıle to the non-past stem, e.g. ba-pp-¨ıle ‘(one) does/will

not come’, tin- ‘to eat’: tim-b-¨ıle ‘(one) does/will not eat’.

In Kota il¯a is added to the past stem to form negative past, e.g. avn

kekn

c-il¯a

‘he

did not do

the work

’.

In Kanna

da the past and non-past negatives are formed by adding illa to the inﬁnitive

or the nominal of a verb, e.g. n¯o

du- ‘to see’: n¯o

d-al illa ‘one did not see’, n¯o

du-v-ad(u)

illa ‘one does/will not see’.

lu deviates from the other South Dravidian

I languages. Negative suf

ﬁxes -ji/-ri

(different social dialects) are added to tensed stems in distant past, immediate past and

present–future to express negation; the personal sufﬁxes follow in forming negative ﬁnite

verbs, e.g. kal- ‘to learn’: immediate past: kal-t¨ı-j-i (1sg) ‘I did not study’, distant past:

kal-t¨ı-d¨ı-j-i ‘I did not study’, present–future: kal-pu-j-i ‘I am not studying/will not study’.

The personal sufﬁxes are: 1sg/pl -i/-a, 2sg/pl -a/-ar¨ı,3msg-e,3fsg-al¨ı,3hpl-er¨ı, 3neu

sg/pl -i/-a. The past conditional form was innovated by combining the past and non-past

markers, e.g. kal-t¨ıd¨ı-v-ay-e (root-past

-past

-non-past-neg-3m sg) ‘he would not have

learnt’, ba-tt¨ıd¨ı-v-ay-e ‘he would not have come’. Here, -ay- is the negative marker.

Summary: in all languages (except Tu

lu), negation with tense is expressed by an

inherited grammatical template: {tensed stem or tensed nominal + illai}. Here illai is

356 The verb

to be interpreted as a uniform marker of negation meaning ‘not’, although technically it

is neuter singular of the irregular verb, il- <

∗

cil- ‘not to be’.

7.10.6.2 South Dravidian II

In Telugu the negative verb root l¯e- ‘to be not’ (<

∗

il-a- <

∗

cil-a) occurs both as the

main verb and as auxiliary. As a main verb, the personal sufﬁxes are added to the root

with Ø tense marking, 1sg l¯e-nu ‘I am/was not’, 1pl l¯e-mu, 2sg l¯e-wu, 2pl l¯e-ru,3msg

l¯e

du, 3hum pl l¯e-ru, 3n-m sg l¯e-du, 3n-h pl l¯e-wu. The paradigm of l¯e- occurring after

an inﬁnitive has a modal meaning (capabilitative), e.g. OTe. c¯eyu-/ Mdn Te. c¯es- ‘to

do’: c¯ey-a(n) l¯e-nu ‘I cannot do’ etc. (see section 7.15.2(9)viib). The past negative in

Middle and Modern Telugu is formed by adding the 3neu sg form of l¯e- (i.e. l¯e-du) to the

inﬁnitive, e.g. c¯ey-a(n) l¯e-du ‘(one) did not do’ corresponding to Ta. var-av-illai which

has the same structure. In Old Telugu the negative past was formed by two serial verbs

(see section 7.13). The present negative in Modern Telugu is formed by adding l¯e-du to

an action nominal in -a

tam/-a

dam, e.g. c¯ey-a

tam/-a

dam l¯edu (lit. doing-not-it) ‘one is

not doing’ (in all numbers and genders).

Some Gondi dialects use a structure similar to Telugu to form the past negative, e.g.

¯u

d- ‘to see’: ¯u

d-(a) ill-¯a-na ‘I did not see’ (Koya dialect: DVM: 356).

Kon

da has sile ‘it is not’ occurring as the main verb and not as an auxiliary. There are

also other non-past ﬁnite verbs with (g)np sufﬁxes, 3m sg sil-en ‘he is not’, 3m pl sil-er,

3neu sg sil-ed/-e, 3neu pl sil-u.

Kui has periphrastic constructions with the inﬂected negative verb si

d- ‘to be not’

added to present and perfective participles, e.g. t¯ak- ‘to walk’, t¯ak-ai ‘walking’ (present

participle): t¯ak-ai si

d-enu ‘I am not walking’ (present negative), t¯ak-ai si

d-at-enu [lit.

walking-not-past suff.-I] ‘I was not walking’ (imperfect negative), t¯ak-a si

d-enu [lit.

having walked-not-I] ‘I have not walked’ (perfective negative), t¯ak-asi

d-at-enu ‘I had

not walked’ (pluperfect negative). Notice

that these distinctions are brought about by

using the present and past participles of the main verb and the non-past and past inﬂection

of the negative verb,

∗

sil- ‘not to be’. Kuvi also follows the same mode.

In Kuvi hil-(<

∗

sil-) ‘to be not’ is used both as a main verb and as an auxiliary in

the

formation of non-past negatives, e.g.

p¯ay-‘to

beat

’: n¯anu

¯ay-i

hil-

-o [lit. I-having

beaten not-I] ‘I am not beating’ (present negative), n¯anu p¯ay-i hil

e [lit. I-having

beaten not-past-I] ‘I was not beating’. Also the perfect and pluperfect negatives are made

by adding the non-past and past hil- to the perfective participle as in Kui, e.g. n¯anu p¯ay-a

hil-

o/hil-

a-t-

e ‘I have/had not beaten’.

7.10.6.3 Central and North Dravidian

The derivatives of

∗

cil are used with negative inﬂection as the main verb. This will make

the meaning of

∗

cil- ‘to be’. This is true of North Dravidian also. It appears that the

use of inﬂected

∗

cil as an auxiliary is an innovation in Proto-South Dravidian, another

7.11

Other simple

ﬁnite

verbs

357

feature that binds South Dravidian I and South Dravidian II as a major branch of Proto-

Dravidian.

7.11 Other simple ﬁnite verbs (afﬁrmative and negative)

7.11.1 Imperative singular and plural

All Dravidian languages have ﬁnite verb forms in the imperative mood. There is no

particular marker for

the imperative. The verb consists

of the stem

+ the second person

singular and plural. Some languages showevidence of the imperative stem being identical

with the stem in non-past ﬁnite verbs. In most of the languages the imperative singular

is -Ø-, i.e. the singular form is identical with the bare stem without a person–number

marker. One set of the second-person sufﬁxes in the imperative mood is different from

the one found in tensed ﬁnite verbs. The ne

gative imperative is formed

by putting the

marker of negation in the tense slot, i.e. between the stem and the person marker.

7.11.1.1 South Dravidian I

1. Tamil: in Old Tamil, there are several modes of forming the imperative singular,

i.e. -Ø/-¯ay/-m ∼ -mm (generally followed by a clitic -¯o/-¯e)/-ti/-mati, e.g. (a) k¯e

l-Ø ‘hear,

ask!’, p¯a

tu ‘sing!’; (b) cell-¯ay ‘go!’, (c) u

n-m‘eat!’, u

lu-m-¯o ‘think!’, cey-mm-¯e ‘do!’,

(d) teri-ti ‘know!’, (e) karai-mati ‘call!’. The forms under (c) are occasionally used in

the hortative plural also, e.g. vaik-am

va-mm-¯o

‘let us stay

and come

’. The imperative

second plural is formed by adding to the stem -mi

n or -¯ır, e.g. c¯er-min ‘reach!’, cell-¯ır

‘go!’. In Modern Tamil the second singular (informal) is close to the verb base and

not identical with it in all cases, e.g. ceyy-i ‘do!’, pa

ti ‘read!’, p¯o ‘go!’; -um and in

plural -u

n-ka

l are used in the imperative second plural or honoriﬁc singular, e.g. v¯ar-um

‘come!’, ko

t-u

nka

l (coll. ko

t-ung¯o) ‘give!’

The negative imperative or prohibitive is formed by adding the negative sufﬁx -¯a-to

the verb stem followed by 2sg -ti; e.g. nill-¯a-ti ‘do not stand!’, p¯ok-¯a-ti ‘do not go!’;

the 2pl -mi

n/-ir are added to the negative particle -al- which is sufﬁxed to the verb

stem, e.g. koll-a

n-min ‘don’t kill!’ (l → n/

−−−

m), nill-al-ir ‘don’t stand!’ Another type of

prohibitive second person plural is formed with the personal sufﬁx -¯ım added to negative

[nominal] in -¯atu, e.g. cell-¯at-¯ım ‘do not go!’ ModernTamil has a corresponding usage,

e.g. coll-¯a-ti-

nka

l ‘do not tell!’

2. Malay¯a

lam: the second singular is -Ø/-¯ay in Old Malay¯a

lam; the second plural is

formed by adding -vin/-ppin/-min (-ppin after strong verbs and -min after nasal ﬁnal

stems), e.g. 2sg cey,plcey-vin ‘do!’, sg na

ta/na

takku ‘walk!’, pl na

ta-pp-in ‘walk!’,

k¯a

n-min ‘see!’ Modern Malay¯a

lam adds -u (formal) and -

(informal) as imperative

signs both in the singular and the plural. A polite imperative is formed by adding the

inﬁnitive markers -ka/-kka in ﬁnite forms, p¯a

tu-ka ‘sing!’, pa

ra-kka ‘tell!’ The aorist

-¯u (< -um) form added to non-past stems is used in the imperative, e.g. ko

l-v-¯u ‘take!’,

k¯e

l-p-¯u ‘hear!’ The prohibitive is formed by adding -ka in the singular, and -vin in the

358 The verb

plural, to the negative stem formed by the sufﬁx -¯ay-, e.g. ceyy-¯ay-ka ‘(one) should not

do’, nill-¯ay-vin ‘do not stand!’

3. Kota: the imperative second singular is -Ø and the second plural -m, e.g. tin-, tin-m

‘eat! (sg, pl)’. The prohibitive singular is formed by adding to the verb stem -¯ad-¯ı,pl

-¯ad-¯ım, v¯ar-¯ad-¯ı, v¯ar-¯ad-¯ım ‘don’t come! (sg/pl)’.

4. Toda: the imperative singular is -Ø, plural -

s, e.g. nil, nil-s ‘stand! (sg pl)’. The

prohibitive is formed by adding -o

ti to the past stem, pod-o

ti ‘don’t come! (sg/pl)’.

5. Ko

dagu: the imperative singular is -Ø and the plural -ri (<-iri, 2 pl sufﬁx), e.g.

b¯a ‘come! (sg)’, b¯a-ri ‘come! (pl)’. The prohibitive singular and plural are formed by

adding the negative marker and the personal sufﬁxes to the verb stem, i.e. -a-te (sg),

-a-ti (pl), tar-a-te ‘do not give! (sg)’, tar-a-ti ‘do not give! (pl)’.

6. Kanna

da: the second singular is -Ø and the plural -i/-im/-ir/-(i)ri, e.g. m¯a

du ‘do!

(sg)’, m¯a

d-i/-im/-(i)ri ‘do! (pl)’. The prohibitive is formed

by adding the negative par

ticle

-al to the stem with a Ø sufﬁx for person in the singular; -im is added in the plural, e.g. n¯o

al ‘do not see!(sg)’, n¯o

d-al-im ‘do not see! (pl)’. In Modern Kanna

da, the prohibitive is

expressed by a periphrastic construction, n¯o

da b¯e

da ‘you need not see’ (inﬁnitive of

n¯o

du + the negative imperative of b¯e

du ‘is needed’; further see non-ﬁnite verbs, below).

7. Tu

lu: the second-person-singular sufﬁx is -Ø/-la and the plural -i/-le, e.g. u

nu/u

n-la

‘eat! (sg)’, jekk¨ı/jek-la ‘wash! (sg)’, kalp-i/kalpu-le ‘learn! (pl)’, ba-le ‘come! (pl)’. The

prohibitive is formed by adding -a

d- to the verbal base followed by 2sg -a and 2pl -e.

Summary: except for T

lu, the remaining South

Dravidian I languages have the

singu-

lar formed by -Ø or -¯ay, the second-singular sufﬁx of ﬁnite verbs. The plural is formed by

a variant of -

∗

w-in (non-past marker plus the plural sufﬁx

∗

-im/-ir). Malay¯a

lam preserves

evidence of the non-past in -vin/-min-/-pin/-ppin. In Proto-Dravidian there is neutra-

lization of ﬁnal -n/-m (

∗

maram/maran ‘tree’) which explains how -in can be related to

the second-plural sufﬁx -im;-ir is the alternative sufﬁx. The replacement of -¯ım by -¯ır

(2pl) as a personal sufﬁx must have been a Proto-South Dravidian phenomenon, but the

imperatives have relics of the replaced morph in some of the Pre-Tamil descendants (see

Old Tamil, Malay¯a

lam, Kota); Toda has -

s in the second plural, corresponding to Tamil

-l, presumably a reﬂex of the negative particle -al which has acquired a number meaning

in Toda. The Tu

lu -l-V represents the negati

ve particle -

al- followed by the

personal

sufﬁxes which are shortened to a point where they do not bear resemblance to those of

any other South Dravidian I language.

7.11.1.2 South Dravidian II

8. Telugu: in Old Telugu the second-person-singular sufﬁxes are -Ø/-mu/-mmu (the

last after short-vowel-ending

monosyllabic roots) and the second plural -

du after stems

which end in non-morphemic -u. Alternatively, -u can be analysed as part of these

sufﬁxes; e.g. cepp-u(mu) ‘tell! (sg)’, cepp-u

du/-u

du ‘tell! (pl)’. The imperatives can

7.11

Other simple

ﬁnite

verbs

359

be followed by address clitics -¯a/-¯ı. In Modern Telugu, the singular is -u and plural

-a

di. When the singular form is followed by a vowel (generally of the complementizer

-an- ‘to say’), u is replaced by -am. In other words the underlying -um surfaces in some

phonological contexts even in Modern Telugu. The simple prohibitive is formed by the

negative marker -ak- between the verb stem and the personal sufﬁx, cepp-ak-umu ‘don’t

tell! (sg)’, cepp-ak-u

du/-˜u

du ‘don’t tell! (pl)’; in Modern Telugu cepp-ak-u, cepp-ak-

di, respectively. Imperatives in monosyllabic bases are ra-mmu ‘come! (sg)’, ra-

‘come! (pl)’: Mdn Te. r¯a (sg), ra-

di (pl). The corresponding negative imperatives are

r¯a-k-umu, r¯a-k-u

du: Mdn Te. r¯a-k-u, r¯a-k-a

di.

9. Gondi: in the singular the imperative is -m- after stems ending in a long vowel

(C)

V-, -Ø- after stems ending in -n, and -¯a- elsewhere; the plural sufﬁx is -

t, e.g. s¯ı-

‘to give’: s¯ı-m, s¯ı-m-

t ‘give!’ (sg/pl)’, tin- ‘to eat’: tin- (sg), tin-Ø-

t (pl), var-¯a: vara-¯a-

‘come! (sg/pl)’. In the Koya dialect the imperative -mu/-mmu occurs in the singular for

stems of the (C)

V(C) type; the plural sufﬁx -

t is added in the plural. For other stems the

sg is -a or -u, e.g. u

d- ‘to plough’: 2sg u

d-mu, 2pl u

d-m¯u-

tu; tinn- ‘to eat’: 2sg tinn-u,

2pl tinn-¯u-

tu. The prohibitive is formed

by adding -

m¯a in 2sg, -m¯a-

t in 2pl, e.g. v¯ay-‘to

come’: v¯ay-m¯a ‘don’t come! (sg)’, v¯ay-m¯a-

t ‘don’t come ! (pl)’.

10. Ko

da: the imperative second singular is marked by -

a and the second plural by

-du with its sandhi variants (-tu/-

ru/-

du phonologically conditioned): v¯a-/ra- ‘to come’:

a (sg), ra-du (pl), nil- ‘to stand’: nil

a (sg), nin-ru (pl), veR- ‘to tell’: veR-

a (sg),

veR-tu (pl), u

n- ‘to drink’: u

a (sg), u

du (pl).

Irregular verbs

ta- ‘to

bring

’ and s¯ı-

‘to give’ have suppletive variants, si-da ‘give me’(object 1sg), si-da-

t (object 1pl); si

‘give to the 3sg’: si-du ‘give to the 3pl’.

11. Kui: The second-singular sufﬁx is -mu (-amu/-umu) and second plural is -du and

its sandhi variants (-

du,-nju) and -a

tu, e.g. tin- ‘to eat’: tin-umu (sg), tin-ju (<

∗

tin-ru),

n- ‘to drink’: u

n-umu (sg), u

du (pl), k¯o- ‘to reap’: k¯o-mu (sg), k¯o-du (pl), l¯ak- ‘to

sacriﬁce’: l¯ak-amu (sg), l¯ak-a

tu (pl). The prohibitive is formed by adding to the verb

stem the negative sufﬁx -

a- before the second singular -Ø and second plural -

tu, e.g.

k¯o-

a-Ø ‘do not reap’ (sg), k¯o -

tu (pl).

12. Kuvi: the second-singular marker is -amu/-mu and the second plural -adu/-du, h¯ı-

‘to give’, h¯ı-mu/h¯ıy-a-mu (sg), h¯ı-du/h¯ıy-a-du, ven- ‘to listen’: ven-a-mu (sg), ven-ju

(pl). The prohibitive has three slots: stem +

a (neg marker) + 2sg -ni, 2pl -du, e.g.

h¯ı-

a-ni ‘don’t give! (sg)’, h¯ı-

a-du (pl).

13. Pengo: the sufﬁxes are 2sg -a/-am,pl-a

t, e.g. hiy- ‘to give’ : hiy-a (sg), hiy-a

t (pl),

ru- ‘to plough’:

ruv-a (sg),

ruv-a

t (pl). Certain clitics -de/-

re are added in the singular,

e.g. kiy-am-de ‘do!’ (sg). The prohibitive singular is -ma and the plural -ma-

t, just as in

da.

Summary: for South Dravidian II, we can reconstruct

∗

-Vm for sg

∗

-dV for plural.

While the singular is found in all members of the subgroup (except Ko

da in which -m

360 The verb

occurs in the prohibitive only), the plural is found in Telugu (the sufﬁx -V

du presupposes

an older

∗

-ndu),Ko

da, Kui and Kuvi. The other set of sufﬁxes,

∗

-wa (in which -w-

becomes a glottal stop in Ko

da, Kui and Kuvi) in the singular and -V

t in the plural, is

apparently different in origin, found in Gondi (including the Koya dialect), Ko

da (in

the prohibitive), Kui and Pengo. Note the absence of a glottal stop and the presence of a

glide in the corresponding position. Kui has both the sets of sufﬁxes. The Telugu sufﬁx

cannot be derived from -um + -

tu (DVM: 495), because a voiceless -

t goes back to -

and would have remained a voiceless -

t in Telugu as -(u)m

tu and not -(u)

du. Note that

Gondi has a second-plural sufﬁx in -V

t in the personal pronouns and in the verbs (also

in Ko

da ﬁrst-person inclusive -a

t). The -V

t is more widespread in the imperative verbs

only and not simple ﬁnite verbs in the indicative

mood.

7.11.1.3 Central Dravidian

14. Kolami:

the imperative suf

ﬁxes are

singular -

Ø, plural

r/-ur. In the

prohibitive the

negative in -ne- is followed by -m in the singular and -r/-

d in the plural, e.g. tin- ‘to eat’:

tin (sg), tin-ur (pl). The prohibitive: tin-ne-m (sg) ‘do not eat!’, tin-ne-

d/-r (pl). Naiki

has the same prohibitive structure as Kolami. Here, the personal sufﬁx -m compares well

with imperative singular

∗

-Vm of South Dravidian II.

15. Parji, Ollari: the imperative singular is -Ø and the plural -ur. The prohibitive has

base + (e)m + person -en (sg), -or (pl). An epenthetic vowel -e occurs before-m, e.g. c¯u

‘to see’: c¯u

r (sg), c¯u

r-ur (pl); the prohibitive c¯u

r-em-en ‘do not see!’ (sg), c¯u

r-em-or

‘do not see!’ (pl). Ollari imperative has -V

t in the sg and -Vr in the pl, nagu-p- ‘to

make one laugh’: nagup-u

t (sg), nagup-ur (pl). The prohibitive is mark

ed by -

men, e.g.

s¯u

r-men ‘do not see!’ The imperative plural is similar to the one in other ﬁnite verbs.

The singular in the prohibitive has -m, derivable from PD

∗

-Vm.

7.11.1.4 North Dravidian

16. Ku

rux: the imperative sufﬁx is -¯a in the singular and plural added to the verb stem.

When the addressee is a woman or non-human, it is replaced by -ai. When women speak

among themselves, it is -¯e, e.g. es- ‘to break’: n¯ın/n¯ım es

¯a ‘you (sg/pl) break’, f neu

ai; women among themselves, es

¯e. In polite imperative -k¯e is used instead. Here,

-k- appears to be the non-past marker with a new connotation, e.g. bar

¯a ‘come’, bar-k¯e

‘come, if you please’. So also -k¯o/-ko

e are used as ‘a kind of mild imperative’, the

former when addressing boys and the latter in addressing girls. Hahn compares these

forms with Mundari polite imperative -ko-, e.g. senkome ‘please go’ (Hahn 1911: 54–5).

The prohibitive is expressed by preﬁxing amb¯a (f ambai, amb¯e) to the verb, apparently

a non-Dravidian construction, ¯as ambas bardas nekk¯a ‘he shall not come’.

17. Malto: the imperative markers are -a (non-future) and -ke (future) both in the

singular and plural, e.g. amb- ‘to leave’: amb-a, amb-ke. The prohibitive is formed by