Negation patterns in Bengali

HANNE-RUTH THOMPSON
School of Oriental and African Studies

I. Introduction
Negation is a challenging chapter in Bengali grammar which has so far
attracted remarkably little attention. Bengali negation is complex: there are no
negative pronouns and adverbs but there are negative verbs, and negation
can be linked to tense. In this article I offer an overview of Bengali negation
patterns as they exist in modern Standard Bengali, without going into great
detail on etymological morphology or, for that matter, a complex debate on
logical aspects or truth conditions.
Of all the negational features in Bengali, the status of ni as a tensed nega-
tive has attracted the most attention, but findings are far from conclusive and
some very basic questions have remained unasked. A relatively recent essay
by Gillian Ramchand ‘Two types of negation in Bengali’ (unpublished, 2001)
gives a semantic analysis of the two forms na and ni and offers some interest-
ing new perspectives.
Suniti Kumar Chatterji’s tentative claim that ni is a variation of nai, and
his uncertain stance on the role of the incomplete verb ach- in the formation
of perfect tense forms has been generally accepted but not examined. Chatterji
deals with negative verbs in one short paragraph in the ‘Additions and correc-
tions’ to his 1,300-page opus The Origin and Development of the Bengali
Language, and makes no meaningful distinctions between the different forms.
Anderson, in his useful, if dated, Manual of the Bengali Language (Cambridge,
1920) devotes all of six lines to the negative in a paragraph entitled ‘The
negative verb’, which I quote here in full:
Negation is expressed by putting na after the finite verb and before the non-
finite verb, with one remarkable exception. It is not usual to say korlam na.
The proper negative of the preterite is kori nai. This applies to all preterites
except chilam. It is permissible to say chilam na. (p. 17).
Much of this is controversial for modern Bengali and I hope to show,
among other things, how peacefully korlam na and korini (the equivalent of
Anderson’s kori nai) can co-exist.
One thing that is immediately apparent in Bengali is that there are no
negative adverbs or pronouns. How do Bengalis say ‘never’? They don’t—and
this simple fact alone constitutes a significant difference between negation
in Bengali and negation in English. Whereas in English we can say ‘Nobody
came’, in Bengali we can only say keu ase ni, which literally means ‘anyone
did not come’. The negative element is moved from word-level to sentence-
level and it is logical to start with the tentative assumption that in Bengali all
negation happens on the syntactic level. We do not encounter the ambiguity
of double negatives such as ‘I won’t have none of that’, ‘I didn’t say nothing’
because negative adverbs and pronouns like nothing, nobody, none, nowhere
or never do not exist.
There are double negatives of a different kind, however, involving lexical
negation. This is a fairly simple morphological feature consisting of negative
affixes and does not really concern us here. But for the sake of a well-rounded
picture, these affixes should at least be mentioned. Most Bengali prefixes have
two allophones, one for preceding vowels and one for preceding consonants.
Milne (1913) gives a complex listing of Bengali affixes and their very varied
Bulletin of SOAS, 69, 2 (2006), 243–265. © School of Oriental and African Studies. Printed in the
United Kingdom.
244 HANNE-RUTH THOMPSON

semantic impact. Of those that can be classed as having a negating quality o-

(on) is the most common. Bengali morphology is dealt with in detail in many
grammar text books, so here are just a few examples. Nouns (including verbal
nouns) and adjectives take the same prefixes; suffixes are word-class specific.
What I have given as the ‘nearest root word’ is in most cases not the basic
form, but it is sufficient to show the building blocks (see Table 1).
Below are some sentences containing more than one negative element.
These are mostly combinations of a negatively prefixed noun or adjective
with a sentence negation, but occasionally (sentence ii) two negatively prefixed
elements come together. The rhetoric element in these sentences is equivalent
to English structures such as ‘it is not impossible...’, ‘it is not unheard of...’.
All these sentences are from Otol joler dike by Prophullo Ray.
(i) nijer odhikarer sima kotodum r poryonto Dtana, setD a tar ojana noy1
He was not unaware of how far he could stretch the limits of his
privileges.
(ii) tãr modhye emon ek probol byktitvo royeche, ya omanyo kora
osombhob
Apart from this she had such a strong personality that it was
impossible to ignore it.

Table 1.
Nearest root word ! Affixed negative form
prefix o (before ksD omota power oksD om powerless
consonants)
sombhob possible osombhob impossible
moron D death omor immortal
pum rn
Do full opum rn
Do incomplete
sahajyo help osohai helpless
gon D ona counting, ogon D yo uncountable,
reckoning innumerable
prefix on onto end, termination ononto endless,
(before vowels) infinite
upoyukto suitable onupoyukto unsuitable
odhikar right onodhikar interference
prefix nir olos lazy nirolos unceasing,
untiring
asa a hope nirasa a despair
ohonkar
q pride niroho nkari
q unassuming,
humble
suffix hin (adj.) polok wink, twinkling polokhin unblinking
uddesa yo purpose uddesa ohin purposeless,
aimless
ortho meaning orthohin meaningless
svartho self-interest svarthohin unselfish

1
The transliteration used in this paper differs from the Standard Sanskritic transliteration in the
following areas: (1) The Bengali inherent vowel is given as o, not as a. I consider this a correction
of a longstanding problem. The inherent vowel in Bengali is pronounced o or om but never a. (2)
and yophola are all given as y. (3) Bisorgo is given as h—this
D corresponds to the Library
of Congress use. (4) No distinction is made between donto t ( ) and khonD do D t ( )—both are given
as t.
 NEGATION PATTERNS IN BENGALI 245
(iii) yodi onekdin kachakachi thaka yeto or ke pheranor osombhob hoto na
If he had managed to stay close, it would not have been impossible to
bring her back.
(iv) garD i calate calate pronD obesa onoboroto kotha bole yacchilen Dthiki, tobe
ei ghor raghuritD a nichok uddesaohin bhromonD noy
While he was driving, Pranobesh kept on talking incessantly, but this
travelling around was not just an aimless journey.
The numbered example sentences in this article are taken from the Dhaka
literary journal kali or kolom (Ink and Pen) Vol 7, August 2004, and from a
range of contemporary novels. They are cited as primary literature. My term
‘real’ or ‘live’ examples excludes all sentences from grammars, courses or text
books which serve to demonstrate particular structures.

II. A comparative detour

(1) Negative particles and speech acts
Before we look at the actual structures, let us take a quick comparative detour.
In English, apart from these negative adverbs and pronouns, we have
the sentence negator ‘not’ and the lexical unit ‘no’, which can function as an
adjective, as in ‘no bread’ and also forms the negative response to a question.
When you look up the word ‘no’ in the dictionary, it is interesting to see that
we are moved from the lexical to the speech act level. ‘No’ as a negative
response has no word class assigned to it but comes under the category ‘formu-
laic expression’, ‘interjection’ or ‘speech convention’, as of course does the
word ‘yes’.
In Bengali the word na is both the sentence negator and the negative
answer to questions. Linguistic terminology is unsatisfactory here. Formulaic
expressions are phrases such as excuse me, thank you, pardon, goodbye and
hello in some definitions, other sources list phrases like by the way, as a matter
of fact, if you will, more or less—interjections are words like ah, well, eh, hmm
and oh. Should the words ‘yes’ and ‘no’ be considered equivalent to these?
It seems to me that what the above expressions have in common is that
they are non-essential parts of speech, whereas the words ‘yes’ and ‘no’ can
convey crucial information. The New Collins Dictionary classes the word ‘yes’
as ‘sentence substitute’—perhaps this is more appropriate, though it is still
rather vague. I do, however, agree that in order to classify words like ‘yes’ and
‘no’ satisfactorily we have to step outside the lexical word class enclosure and
recognize their illocutionary force.
Both ‘no’ in English and na in Bengali can stand as statements on their
own. There are other ways of answering a question in the negative. In English
we have phrases like not at all, far from it, in Bengali mor Dtei na ‘not at
all’, emotive expressions like dumr, which literally means distance but is used
to strongly reject a suggestion, and words with negative prefixes such as
osombhob ‘impossible’.
For the purpose of clarity I will make a distinction between what I will call
the illocutionary na, which is used as a negative answer to questions, and an
adverbial na in connection with verb forms.

(2) negative potential

The fact that Bengali has no negative adverbs and pronouns seems at first sight
to constitute a lack. In English we can make subtle distinctions, as between
246 HANNE-RUTH THOMPSON

‘I’ve said nothing’ and ‘I haven’t said anything’, the first version of which is
more definite. Bengali does not have this option. Nor does it employ emphatic
negatives like not ever, never ever, none at all. Indefinite adverbs kor thaor
‘somewhere’, kokhono ‘ever’, and pronouns keu ‘someone’, karom ‘someone’s’,
kichu ‘some’, kor no ‘any’, etc., are used in conjunction with a sentence-final
negation. Some simple examples with the English negatives and the Bengali
indefinites underlined are given in Table 2.
From Table 2 it may seem that Bengali is more restricted than English
in expressing negative concepts, but a closer analysis of Bengali patterns will
show that through devices like flexible word order, negative verbs and non-
finite structures a great range of expressions is available. But what about more
complex sentences?
The sentence-final negative can serve to negate more than one indefinite
adverb/pronoun, and the English translation of the sentence shows that, in
spite of its greater lexical range, English does not, in standard speech, allow
more than one negative element per sentence either, whether this be an adverb
or a sentence negator. Instead the first negative adverb/pronoun causes all
subsequent adverb/pronouns to appear in their indefinite form, as shown here:
keu kokhono ei kotha janbe na No-one will ever know about this.
or ra kokhono kauke kichu bolbe na They will never tell anybody anything.

III. Negation patterns in Bengali

As Anderson (1920) rightly says, the negative particle na is added after the
verb. I will, for the moment, consider ni to be a tensed version of na and
discuss the relationship between the two forms further on (see under III, 4).
But we still have two other sentential negative elements whose uses need to be
looked at.
Purely on the surface structure we have three basic syntactic patterns:
• sentences with the negative element no-;
• sentences with the negative element nei;
• sentences where na (or ni) is added to the verb.
Both no- and nei can be considered negative verbs, which means that they
occur in sentences where no other verb form is present and can therefore be
assumed to contain the verbal element. Bengali linguists, including Pabitra
Sarkar, list no- and nei indiscriminately together as forms of negation. As
far as I know, no one has attempted a proper differentiation and comparison
of the two verbs. In order to make clear the syntactic nature of this study the
verb nara ‘be unable to’ should be briefly mentioned. This verb is analogous
to the verb para ‘be able to’ and is no longer in common use. In spite of its

Table 2.
English Bengali gloss
Nobody knows. keu jane na anyone does not know.
Nothing needs to be said. kichu bolar dorkar nei something of the saying
need is not
They will never lose this. or ra setD a kokhono they ever will not lose this
harabe na
I have no doubts. amar kor nor sondeho nei my any doubt is not
 NEGATION PATTERNS IN BENGALI 247
negative meaning it is syntactically a regular verb and has no unusual features
which concern the negation structures of Bengali.

(1) Sentences with the negative element no-

This is a negative verbal element; it is not used in connection with other verb
forms and is the negative of the zero verb. The zero verb in Bengali has often
been interpreted as the absence or omission of a verb form in the present tense.
I disagree with this reading and have discussed it in detail in the chapter on
verbs of being in my PhD thesis. The zero verb is the true copula in equational
sentences. Neither hoor ya with its semantic properties of ‘becoming’ nor ach-
with its existential features can fulfil this role. The confusion which has existed
around these structures for many years is perpetuated by Bengali scholars
when they classify the noi, noor , noy, non forms as ‘the negative na with copula
verb hoor ya (Dakshi 1995: 137). This classification is wrong on both counts:
the noi, noor , noy, non forms are not the negative forms of hoi, hoor , hoy, hon,
and hoor ya is used as a copula only in a very few specific cases. One of them is
the use of third person present continuous hocche and simple past holo as a
procopula, e.g. aj holo budhbar ‘today is Wednesday’. These forms are never
negated. A detailed discussion of these can be found in my thesis. Another use
is the definition of human relationships and this is the only case where the
forms of no- can be interpreted as a direct negative to forms of hoor ya, as for
instance in:
se tor mar ki hoy, literally: ‘What is he to you?’
to which the answer would be:
se amar bhai hoy ‘He is my brother’.
The negative of this is se amar bhai noy ‘He is not my brother’, but there is a
subtle shift in focus here. se amar bhai hoy makes a statement about a relation-
ship, se amar bhai noy makes a statement about a person. The difference is,
however, slight, and I have found no occurrences of se amar bhai hoy na.
I have no argument with Chatterji who says that etymologically the form
noy is a contraction of na hoy. This is on the morphological level and sounds
probable but in modern usage and from a syntactic point of view noy and hoy
na are quite distinct.
hoomya like other verbs, negates with na
ekhane gom hoy Wheat grows here.
! ekhane gom hoy na Wheat doesn’t grow here.
kalke brDsDti hobe There will be rain tomorrow.
s
! kalke brDsDti hobe na There will be no rain tomorrow.
s
tar osukh hoyeche He has been ill.
! tar osukh hoy ni He has not been ill.
This is the only correct way of negating these sentences.
The verb no- has the present-tense conjugated forms only: noi (for first
person), noor (for second person familiar), noy (for third person familiar), non
(for honorific forms, second and third person). There are no non-finite verb
forms, no imperatives and no other tenses. The stem no- as well as the verbal
noun, which is given as noha in the Samsad Dictionary, are assumed dictionary
creations with no existence in real language. The old forms nohi for noi and the
equivalent forms for second and third person are now obsolete. There is a con-
ditional participle noile which is used in alternation with na hole ‘if this is not
so’, but its semantic features are not in line with the conjugated forms and I
248 HANNE-RUTH THOMPSON

would consider it a separate lexical entity. I will come back to this. Forms of
no- appear in simple sentences such as:
(1) boitD a amar noy The book is not mine.
(2) aj budhbar noy Today is not Wednesday.
(3) gachtD a besa i bor
D o noy The tree is not very big.
(4) baba nobi non Father is not a prophet.
All these examples in their positive form would have the zero verb. Past tenses
would be with chilo na, the negative past tense of ach-, the future tense is most
likely with a form of hoor ya ‘be, become’, the conditional also with hoor ya. This
touches on the complex issue of overlap among the Bengali verbs of being, and
requires separate treatment.
But let us look at some real examples, moving from simple to more
complex structures (all from kali or kolom):
(5) amar tor nisa cito noi. But we are not sure.
(6) ei ki yothesD Dto noy? Is this not enough?
(7) se poth amar noy. This is not my way.
(8) amar kaje ami mor Dtei trpto noi. I am not at all satisfied with my work.
(9) kebol mukher kotha noy. s (They) are not just empty words.
(10) lekhok bor dh hoy praibhetD bisa vobidyaloygulor somporke khub besa i
obohito non.
The writer is perhaps not very familiar with private universities.
(11) etD a komitD ir siddhanto, amar ekar noy.
This is the committee’s decision, not just mine alone.
(12) ekhan theke koto dum r apnar bar D i? besa i noy.
How far is your home from here? Not far.
It has to be kept in mind that the function of a copula is to link (or equate) two
elements, the subject and its complement, though the subject can sometimes be
omitted, as for instance in sentence (9). Even with more complex structures it
is possible to reduce them to their essentials in order to see the logic. The basic
positive structure is [Subject] equals [Complement], with the subject either
a noun or a pronoun, the complement a noun, pronoun or an adjective. In
sentence (12) we have an interesting example of a locative context, which is
normally associated with ach- and therefore nei, but if we look more closely we
see the equational element: dum r besa i noy ‘The distance is not great’. As we shall
see in section IV there are a number of areas where the uses of noy and nei can
indeed come very close, and careful thought or reliable intuition is needed to
distinguish between the two.
In addition to this verbal role, noy has a modal sentence function express-
ing appropriateness or acceptability. This appears, though not frequently, in
connection with certain non-finite verb forms. Two instances are as follows:
(i) noy following a conditional participle: ajke na gele noy, which
translates as ‘It won’t do if (we) don’t go today’.
(ii) noy following a verbal noun in the genitive: etD a hoor yar noy, ‘This is
not supposed to happen’.
Both of these structures contain a modal element and go beyond mere nega-
tion. An overview of these and other modal structures involving non-finite
verb forms is given at the end of section (2) but I have not yet found sufficient
examples to come to any firm conclusions about their internal composition.
 NEGATION PATTERNS IN BENGALI 249
(2) Sentences with nei
This negative verbal form is unchangeable and is the negative of ach- ‘to be,
to exist, to be present’. On the surface this is very simple. ach- itself is an
incomplete verb with only present-tense achi (first person), acho (second per-
son, familiar), ache (third person, ordinary), achen (second and third person,
honorific), and the equivalent simple past-tense forms chilam, chile, chilo,
chilen, but no non-finite verb forms, no imperative and no other tenses. For
these, the verb thaka ‘stay, remain’ is the substitute.
As learners of Bengali, our first acquaintance with the verb ach is in sen-
tences such as amar gar D i ache ‘I have a car’ and its equivalent negative amar
gar
D i nei ‘I don’t have a car’. To translate ach- with ‘have’ is an approximation.
There is no overtly possessive element in ach-; a more accurate translation
of amar gar D i ache is ‘my car exists’ nei is also the negative for the locative use
of ach- where the verb forms are conjugated. Thus we have:
ami ekhane achi I am here. ami ekhane nei I am not here.
tumi ekhane acho You are here. tumi ekhane nei You are not here.
se ekhane ache He (ord) is here. se ekhane nei He (ord) is not here.
tini ekhane achen He (hon) is here. tini ekhane nei He (hon) is not here.
nei is therefore the only verb in Bengali which treats me, you, ordinary and
honorific, in the same way, and this does, of course, have a compelling logic
behind it: absence does not distinguish gods, ghosts or specks of dust. This
fairly straightforward semantic concept, expressed by nei, can be divided into
two types of absence: absence from a particular place (specific absence (a))
and non-existence (general absence (b and c)). For the convenience of English
speakers, sentences denoting general absence can again be divided into those
which can be translated with ‘have’ (b) and those best translated with phrases
like ‘there is no’ (c).
The presence of a logical subject in the genitive usually indicates that a
translation with ‘have’ is appropriate. The following example sentences are
divided into groups which show these differences clearly.
(a) examples of specific absence, locative sentences:
(13) uni ajke basay nei. He is not at home today.
(14) keu tor ekhane nei. There is no-one here.
(15) uni emon svostite nei. He is not in such great comfort.
(16) nosaDto korar motor poryapto somoy tar hate nei.
There is not a lot of time to waste on his hands. (a clumsy
translation!)
(17) amader desa er praibhetD bisa vobidyaloygulor te bamd la bhasD ar kor nor sthan
nei.
In the private universities of this country there is no room for the
Bengali language.
(b) examples of general absence with a genitive ‘logical’ subject, ‘have
sentences’:
(18) mayer apon rokter keu nei
lit: of mother of own blood someone is not
Mother has no blood-relatives.
(19) amar ar kichu bolar nei
lit: of me something more of saying is not
I have nothing more to say.
250 HANNE-RUTH THOMPSON

(20) er ninda korar bhasD ar amar jana nei

lit: of this language of condemnation my knowing is not
I have no language with which to express (sufficient) condemnation
of this.
(21) tãr songe
q amar bisa esD dvimot nei
lit: with him of me particular disagreement is not
I have no particular argument with him.
(c) examples of general absence without a possessive element, ‘there is not
sentences’
(22) ar kor nor poth nei?
Is there no other way?
(23) amake na cene emon lor k nei
lit: me does not know such a person is not
There is no one who doesn’t know me.
(24) er sotyota osvikar korar upay nei
lit: of this truth of the denying way is not
There is no way of denying the truth in this.
(25) robindronather sarbik bhabnar phosol i ye e melay
lit: the harvest of all Rabindranath’s thinking that in this fair
rum payonD korar cesD Dt a hoyechilo tate sondeho nei
of the shaping there was the effort in this doubt is not.
There is no doubt that an effort was made to incorporate all of
Rabindranath’s thinking into the character of the fair.
My arrangement and interpretation of these sentences is meant to show that,
whatever the degree of complexity or abstraction, we have in all of them the
underlying simple concept of absence. Particularly in the category of general
absence, there are a number of abstract nouns which are commonly used. In
sentence (24) we have upay ‘way’, in sentence (25) sondeho ‘doubt’. Other such
nouns are dorkar ‘need’, kotha ‘word’, proyor jon ‘necessity’, somoy ‘time’, jayga
‘place’, karon
D ‘reason’, sombhobona ‘possibility’, subidha ‘convenience’. The
phrase osubidha nei ‘there is no inconvenience’ is as common as the English
equivalent stock phrase ‘no problem’. Syntactically nei almost always follows a
noun. In sentences (19) and (20) we have two verbal nouns; amar ‘my’ (or any
other possessive pronoun) jana nei ‘I have no knowing’ is a common phrase
which often, but not always, implies information irretrievably lost or unavail-
able and the phrase offers a convenient way out of a personal and active
statement ami jani na ‘I don’t know’. In sentence (19) we have a genitive verbal
noun bolar ‘of the saying’ attached to kichu ‘something’. This is also a very
common structure and is, from an illocutionary point of view, equivalent
to shifting the attention away from the speaker. If we compare these two
sentences
(i) active ami ar kichu korbo na I won’t do anything more.
(ii) with neiamar korar ar kichu nei There is nothing more for me to do.
we see immediately that the structure with nei is not only impersonal but
implies an abdication of responsibility. In fact, taking a step away from syntax
to the communication level, the word nei in itself has a somewhat fatalistic air
and is often accompanied by an open-handed waving gesture to indicate that
there is nothing to be done.
Like noy, nei also has a modal role in connection with non-finite verb
forms, this time with the infinitive. In ekhane thakte nei ‘it is not appropriate
 NEGATION PATTERNS IN BENGALI 251
to stay here’, nei adds the element of suitability, permission or acceptability. I
have not yet encountered many live examples of this structure, but it belongs
to a pattern of a particular type. Some common verbs in connection with cer-
tain non-finite verb forms take on a modal meaning, for instance etD a hoor yar
noy ‘this should not occur/happen’, etD a korte hobe ‘this has to be done’, ekhane
thakte ache ‘it is permissible to stay here’, etD a kora yay ‘it is possible to do this’.
The verbs involved in these structures are predominantly verbs of being and
the verbs yaor ya ‘go’ and cola ‘move’.
Table 3 shows a preliminary overview of these structures.

Table 3.
Non-finite verb form + 3rd person verb form modal meaning
pos neg
cond. part. gele hoy hoy na sufficiency
cond. part. gele cole cole na sufficiency
inf. yete hoy hoy na obligation
inf. yete ache nei permission
cond. part. gele – noy appropriateness
vbN (gen.) yaor yar – noy compulsion
vbN yaor ya yay yay na possibility

(3) The properties of ni

ni has a different status from nei and the forms of no-. It is not a negative
verb but a negative particle used together with simple present-tense verb forms,
and is usually classified as the negative of the present perfect and of the past
perfect. This classification is based on the non-appearance of the expected
forms *korechi na (present perfect) and *korechilam na (past perfect), not a
very reliable base for interpretation. What we do know is that, in addition to
the simple present negative ami kori na, we have the structure ami korini and
that there is clearly a difference of tense between these two forms in that ami
korini has a past tense meaning. This is a vague but safe place to start.
Rafiqul Islam in his 2001 condensed grammar Bangla Byakaran Samikha
points to a fundamental difference between na and ni. na is an independent
particle, ni is not. In written language ni is often attached to the verb form, na
is written separately. But this is merely a convention and has no bearing on the
internal connection between these forms. A Bangladeshi variation of ni is nai,
and this is frequently separated from the verb form in writing. From what we
have seen earlier, there is no doubt that na has the monopoly on being the
illocutionary negative particle, but how does na differ from ni in its adverbial
use? In addition to its negative function ni has a tense element built into it and
is therefore more closely linked to the present-tense verb form—it has a doubly
changing effect on the verb form. Rafiqul Islam makes much of the fact that
ni cannot stand on its own and this is a valid observation. ni always follows
a simple present verb form directly, no adverbial can be inserted between the
verb and ni, and ni never occurs on its own in answer to questions.
It is disconcerting to see that even something as basic as the actual verb
forms used is a matter of disagreement. Qazi Din Muhammad calmly states
that ‘the use of the particle ni with the simple present is found to function
as the corresponding negative implications of the perfective present and
252 HANNE-RUTH THOMPSON

progressive and perfective past in all aspects and persons and grades’ (1963:
33). What he is saying is: ami khai ni is the negative not only of the present
perfect ami kheyechi and past perfect ami kheyechilam but also of the past
continuous ami khacchilam. Smith (1997) makes no mention of the past con-
tinuous in this context but allows the occasional use of korechilam na (the past
perfect) (p. 23). Our starting position is therefore far from clear. In order to
refute Muhammad’s claim, all we need to do is to find some examples of past
continuous uses with na and this is easily done.
(26) bhiruta katD iye utD hte parchilen na onuradha
Onuradha was unable to overcome her fear.
(27) osommantD a ekebare bhulte parchilen na
He was unable to forget the slight.
These sentences suggest that the ni negation does not apply to the past
continuous but they are taken from the same novel and both involve the verb
para ‘be able to’ so we need to look a bit further.
(28) ami kichu bolchilam na I wasn’t saying anything.
(29) pron
D obesa ki uttor deben bhebe pacchilen na
She could not imagine what Pranobesh would reply.
(30) sthir thakte para sombhob hocchilo na
It was not possible to remain quiet.
All of these are instances of past continuous verb forms negated with na, and
since this confirms the general consensus on which tenses are affected by the
ni negation, we can safely assume that Muhammad’s claim is wrong. Smith’s
statement that the past perfect negative is occasionally formed with na is
more difficult to refute—the fact that I have not found any examples for this
is insufficient proof. But since he is talking about an occasional occurrence
rather than a grammatical rule, we can leave this with a question mark.
It is much more important and interesting to try and establish the semantic
properties of these ni structures. Designating the present and past perfect
as their area of application is a default solution and does not yet tell us
very much. In dealing with real language we cannot speculate about what a
sentence would have been without the negative element. We can only go by
what we find, and it is reassuring to find sentences like:
(31) sabu ar tar premik eksathe thakte sa uru koreche ki koreni khobor peye
lor ktD ir stri gãye hajir holo
The man’s wife turned up in the village with news about whether
Sabu and her lover had started living together or not.
where the positive koreche and negative koreni forms appear in direct
juxtaposition.
An indication of past perfect contexts is given in these examples:
(32) kichu ektD a bolte cesD Dta korechilen tini kintu golar svor phor Dteni
He tried to say something, but his voice failed him. (lit: his voice
didn’t emerge)
(33) tarpor Dthik ki hoyechilo tãke janayni se
Exactly what happened afterwards, he didn’t tell him.
These are fairly firm indicators of a tense-context, but from a functional point
of view we are on safer ground referring to the structure in question as the ni
past tense rather than as the present and past perfect negative.
 NEGATION PATTERNS IN BENGALI 253
In order to get a firmer grip on these structures, let us have a look at past-
tense negative verb forms in general. In sentences (29) and (30) the past con-
tinuous negative forms have a particular purpose and are the logical forms to
use. Bengali tense use often seems arbitrary to us but it has its own semantic
logic. Both these verb forms occur in narrative, simple past contexts. In sen-
tence (29) the past continuous tense expresses the ongoing frustrating worry
about what the husband’s reaction might be. In sentence (30) the continuous
verb form, in spite of the stative properties of sombhob hoor ya ‘be possible’ adds
the itchy restlessness which the author wants to convey. No other tense would
have had the same effect. These kinds of examples are not frequent but when
they do occur, they are used for specific purposes. In neither case is the ni past
tense a likely alternative. Apart from the past continuous we have the simple
past ami khelam na and the past habitual ami khetam na. In order to establish
the role of the ni past tense, we need to take all these forms into account.
Anderson is wrong when he says that the simple past korlam na is unusual.
There are plenty of examples like (34) with or without time adverbials:
(34) se din ar keu elo na
No-one else came that day.
Gillian Ramchand in her article ‘Two types of negation in Bengali’ sets out
to ascertain the semantic range of the ni past tense through the analysis of
linguistic features and with complete disregard to traditional categorization.
We need this kind of purely linguistic focus because it offers an objective
analysis of contemporary language. If we want to classify the ni form appro-
priately, we need something more tangible than somewhat arbitrary traditional
assumptions.
Ramchand’s article offers a new approach and I was fascinated by her
attempt to prove that the present perfect is semantically incompatible with the
na negative. I was thinking along the same lines and would have liked to
come up with a clear semantic distinction between sentences like ami khelam na
(simple past with na) and ami khai ni (ni past tense). One fairly clear semantic
component is that of volition. While ami khai ni is the simple reporting of a
fact, ami khelam na contains a deliberate element ‘I didn’t eat because I didn’t
want to eat’. Ramchand sees the distinction in the aspective features involved
in the two tenses, the simple past and the present perfect. Briefly stated, she
says that a perfective aspect cannot be made incomplete. A comparable struc-
ture in English is a sentence containing ‘already’ which represents a perfective/
completive aspect. The only way to negate an already completed action is to
abandon the completedness and revert to an unfinished state of events by
replacing already with yet, e.g. ‘I have already seen him’ becomes ‘I have
not seen him yet’. In the move from the completed ami kheyechi ‘I have
eaten’ to a negation of this event we have to change the completive verb form.
The reasoning behind this can be shown even more clearly if we look at the
common Bengali compound verb structure of a perfective participle plus the
verb phela ‘throw’. In many ways, this structure is the epitome of perfectivity
or completedness. ami bhat kheyechi means ‘I have eaten’ or ‘I have eaten the
rice’. ami bhat kheye phelechi means ‘I have eaten the rice up’. This sentence
cannot be negated. In order to express the negated concept, the verb phela,
with its completive aspect, needs to be removed and the simple verb form
restored ami bhat khai ni or ami bhat asesD kori ni ‘I have not finished the rice’
i.e. ‘there is still some left’.
Ramchand’s analysis is thorough and careful and adds a new perspective
to the understanding of these forms. She lacks, however, a realistic view of
254 HANNE-RUTH THOMPSON

Bengali tenses. This is perhaps simplifying her arguments more than is admis-
sible but it seems to me that her understanding of the present perfect is based
on this aspective feature of completion, a telic reading, whereas the simple past
has to her an inherently atelic character. These issues need further consider-
ation, but they take us far beyond our present topic to a semantic analysis of
Bengali tense use. Add to this the distributional factor of these tenses: the main
use of the simple past is in narrative text while the ni past tense is the more
common tense in spoken language.
In order to gain a clearer understanding of the ni past tense, I have carried
out a random study of narrative texts (mainly fiction) on the initial assumption
that the ni past tense does not mix well with simple past tense storytelling.
From normal everyday spoken use of the form, it is easy to get the impression
that it has a ‘not yet’ feature built into it. If we compare the simple past ami
bajare gelam na and the ni past tense ami bajare yai ni both sentences can
be translated as ‘I didn’t go to the market’. In order to clarify the difference
between these two sentences, we can add time adverbials and have
simple past (i) ami gotokal bajare gelam na
I didn’t go to the market yesterday.
ni past tense (ii) ami ekhonor bajare yai ni
I haven’t yet gone to the market.
This is the kind of pattern we would like to find and which corresponds well
with English tense use, i.e. the semantic difference between simple past and
present perfect. It is also the pattern that Ramchand identifies. Using time
adverbials to test semantic features is a valid and useful procedure and, indeed,
in a great number of cases the distribution of time adverbials seems to confirm
our desired pattern. Let us look at some examples:
(35) ekhane emon kichu ekhonor ghotD eni
Nothing like that has happened here yet.
(36) apnake tor bola hoy ni
You have/had not been told.
(37) keno etodin tor mar songe
q yor gayor g rakhte parini, setD a tor mar or jana noy
You know why I was unable to stay in touch with you for so long.
(38) ami onek khor̃ j korechi kintu kopi pai ni
I searched a lot but have not found any copies.
(39) se kokhonor biyete raji hoy ni
She has never agreed to marriage.
Even in the absence of overt time adverbials these sentences have a perfective
character which indicates a future change, e.g. in (36) it can be assumed that
the addressee will now be told, in (37) that contact has now been restored, in
(39) that the lady in question is now changing her mind about getting married.
But is the use of the ni past tense restricted to such contexts? What about its
use in simple past narrative contexts? Given the fact that in such contexts we
find frequent leaps to the simple present or the past perfect and therefore a
general instability of tense use, there is no indication at all that the ni past tense
does not fit into these contexts. Here are three poignant examples:
(40) pron
D obesa yokhon nije theke boleni tokhon e niye kor utuhol prokasa
koratD a obhaboniyo
When Pranobesh didn’t say anything himself, it would have been
unthinkable to reveal his curiosity.
 NEGATION PATTERNS IN BENGALI 255

(41) tar bor n tulu bombete phor n korechilo. tader khobor nebar por laintD a
pron D obesa ke dite bollo. omla tokhon take bolen pron D obesa bombete nei.
kor thay gechen setD a tulu jante cayni
His sister Tulu phoned Bombay. After hearing their news she asked
to speak to Pranobesh. Amla told her that Pranobesh was not in
Bombay. Tulu didn’t ask where he had gone.
(42) strir songe
q mithyacarer jonyo kharap lagchilo pron D obesa er. kintu e
char D a ki korar ache? moner gor pon alor chaya-ghera omd sa e ãtipãti kore
khum jeor protaron D ar samanyo obhiyor gtD ukuor paor ya gelo na. na omlake
tini Dthokanni
Pranobesh felt bad about lying to his wife. But what else could he
do? Even diligently searching the darkest recesses of his mind he
could find no charge of deceit. No, he had not cheated Amla.
In (40) we have in one neat sentence a convincing indication that the ni
past tense can occur with a fixed point time adverbial (yokhon) and refer to a
past event. In (41) we have a simple series of events with the usual mixture of
tenses. We have two negatives. The first one pron D obesa bombete nei is reported
speech. In Bengali this is usually conveyed directly, so the simple present in this
sentence is expected. The second one is in the ni past tense. Out of the many
passages I examined I chose this relatively simple one to show that the ni past
tense is used for completed actions (or non-actions) in the past without any
indication of our ‘not-yet’ feature. Example (42) is even more convincing. The
last sentence na omlake tini Dthokanni serves as a summing-up of a past tense
train of thought and actions with no connection to the present tense. A more
thorough investigation will have to be done on these past-tense negatives, but
this should be in connection with tense use in general.
In addition to these fairly straightforward findings, it is becoming increas-
ingly clear to me that there is a style issue involved here. The more colloquial
the style, the more ni past tenses appear. In novels with a more formal style we
often have a change to the simple present or the present continuous for nega-
tive statements. It is almost impossible to draw any firm conclusions on this,
but we can be sure that the ni past tense is a fully participating member of past
tense narrative verbal variations.

IV. Syntactic realities

(1) Overlap of nei and noy
Learners of Bengali have a certain amount of difficulty in distinguishing
between the uses of nei and noy. We can explain to them that nei is the negative
of ach-, noy the negative of the zero verb; we can say that sentences with nei
involve absence, sentences with noy the non-equation of two elements, but the
structures can indeed become very close. If we look at example sentence (3) se
poth amar noy ‘this is not my way’ we can replace noy with nei and translate
the sentence as ‘I don’t have that option’. We may want to change the word
order to amar se poth nei to see that this is the same structure as amar gar Di
nei ‘I don’t have a car’. But word order is fairly flexible in Bengali and the
underlying structure is not always easy to detect. In Milne’s Practical Bengali
Grammar we find the sentence:
bina nimontronD e sekhane yaor ya amar iccha noy
I don’t wish to go there without an invitation.
256 HANNE-RUTH THOMPSON

Milne’s book has many errors in it and we are tempted to regard this as one
of them because we expect amar iccha nei ‘I have no wish’. But a nei structure
would clearly require the verbal noun yaor ya to be in the genitive: amar sekhane
yaor yar iccha nei ‘of me the wish of going is not’. With a nominative verbal
noun an equational reading is entirely possible ‘to go there my wish is not’.
The meanings are to all intents and purposes the same, but the example shows
not only that these kinds of sentences need very close reading, but also that the
fluctuation between nominative and genitive verbal nouns, which is sometimes
dismissed as sloppiness, in fact has a much more solid syntactic basis. Syntac-
tically, in structures with nei the immediately preceding noun is the sentence
subject, while everything else serves to modify this subject, e.g. amar sekhane
yaor yar dorkar nei ‘of me the need of going is not’, dorkar ‘need’ is the subject
and amar sekhane yaor yar ‘of my going there’ qualifies this need. In structures
with noy we are dealing with two nominative elements, the subject and the
complement in an equational relationship.
This relationship status can sometimes be obscured by omission of the
subject. We had this in sentences (9) and (12) and we have it here:
(43) ekhon somoy protibad korar protibad sa udhu noy protiror dh korar
Now is the time to protest, not only to protest, but to prevent.
The basic sentence is somoy protibad korar noy ‘the time is not to protest’. This
is a more uncompromising statement than ekhon protibad korar somoy nei
‘now there is no time to protest’ would be, and it shows that the abstract nouns
mentioned above are by no means automatically linked with nei. In some
sentences the copulative structure is plain to see:
(44) manusD sa udhu premer karon
D e noy
Human beings don’t just exist for love.
But in sentences like (45), again from Milne, we need to identify the
nominative subject taha ‘that’ in order to grasp fully the structure.
(45) se ye amake ki birokto kore tuliyechilo taha ar bolibar kotha noy
I can’t adequately express how greatly he annoyed me.
The second part of the sentence can be glossed as
taha ar bolibar kotha noy
that more of the saying word is not
If we leave out the non-essentials ar ‘more’ and bolibar ‘of the saying’, and
look at
taha kotha noy
that word is not
we see that this is exactly parallel to, for instance,
ta mithya noy
that lie is not
‘That is not a lie’.
This detailed demonstration of a structure which is intuitively accessible
to native speakers has a significance beyond the learning needs of foreigners
because it offers a real insight into the way syntactic structures reflect semantic
realities. This level of understanding opens up the logic of Bengali grammatical
structures in a new way and alerts us to these intriguing case variations.
 NEGATION PATTERNS IN BENGALI 257
Here are three sentences from Abdul Rouf Choudhury’s novel Pordeshe
Porbasi, which show the same pattern.
(46) tobe eksD etre somoy tor bor
D o kotha noy
But time is not the real issue here.
The equational structure is easy to see here. In both the following sentences,
however, a combination of kotha with nei would be more usual, and we note
that the use of noy causes the sentence subjects (niculor k (47) and osubidhe (48))
are in the nominative, where nei would have required the genitive. The possible
nei versions of the sentences are given in brackets:
(47) imd lyan
D dD e tor krDsoker mor to niculor k thakar kotha noy
s Dsoker mor to niculor ker thakar kotha nei)
D dD e tor kr
(imd lyan
But in England s such lowly people as farmers are not supposed to
exist.
(48) apnar kor nor osubidhe hoor yar kotha noy
(apnar kor nor osubidhar hoor yar kotha nei)
You are not supposed to have any problems.
These examples show the complex internal structuring of Bengali sentences.
Unfortunately, the matter does not end there. Whilst it is true that sentences
with noy need a nominative subject, they do not always have a nominative
complement. We have seen this in simple structures like
boytD i amar noy
the book mine is not
‘The book is not mine’.
But we also find it with verbal nouns
(49) e kolkatar gonq gay ki ek Dtan ache taha bhor lbar noy
There is some attraction in the Ganges in Kolkata which is
unforgettable.
For our purposes, we need only look at the basic structure
ta bhor lbar noy
that of the forgetting is not
There is no sign of any hidden, modal meaning in this. If we had come across
this sentence in isolation, we might have taken it as an instance of ellipsis,
short for ta bhor lbar jinis noy ‘this is not a thing to be forgotten’ but alongside
the simple boitD a amar noy ‘the book is not mine’ we can accept it as a complete
sentence and at the same time take note of the wide semantic range of the
genitive. As I said at the end of section (1), I do not yet have enough examples
of this kind for a proper analysis and, taking sentence (49) into account, it is
clear that the modal character of these structures is by no means certain. Nor
are the structures themselves clearly defined.
Both the infinitive with nei and the genitive verbal noun with noy express
a prohibition, though whether these imply a strict ‘must not’ or a less severe
recommendation ‘not to’ has not yet been examined conclusively.
We can have:
(50) por na mach dhorte nei
fingerling to catch is not
‘To catch fingerlings is not allowed’.
258 HANNE-RUTH THOMPSON

or
(51) por na mach dhorbar noy
fingerling of the catching is not
‘Fingerlings are not to be caught’.
In sentence (52) we have, with the same structure, an element of necessity
rather than prohibition.
(52) ta amader bole debar noy
this to us of the saying is not
We don’t need to be told this. Or, indeed
It is not up to us to say this.
The modal element is there in these sentences but it seems to have a wide
semantic range. The syntactic scaffolding alone does not tell us enough.

(2) Positioning of na
This section can be subdivided into the following occurrences of na:
(i) na following a finite verb form;
(ii) na preceding a finite verb form for emphasis;
(iii) na preceding a finite verb form in subordinate clauses;
(iv) na preceding a non-finite verb form.
(i) The negative particle na follows the verb and, since verbs usually come at
the end of sentences, na is generally the last word in a sentence. Rafiqul Islam
interprets the fact that ni is usually attached to the verb form as a particularly
close connection. The connection between verb form and na is, however,
almost as close. I have found only one instance of the particle tor inserted
between the two.
(53) aror dujon yatri peye gele tor make niye yete kor nor osubidha hobe na
kintu pacchi tor na.
If two other passengers can be found, there won’t be a problem
taking you. However, I’m not finding any.
tor is an adversative particle which strengthens the effect of kintu ‘but’. It could
just as easily have been placed after na. Examples like this—and there are only
very few—show that the verb form and na are not entirely inseparable but that
they have a strong adherence to one another.
There are two regular patterns for na preceding the verb and they are
syntactically linked to one another: subordinate clauses and non-finite verb
forms.
(ii) But first we have an emphatic use of na:
(54) na ache bamd la bhasD a, na ache bamd la bibhag
There is no Bengali language and there is no Bengali department.
(55) na yabo bombete, na yabo kor thaor , ekhanei thakbo tor mar songe
q
I’m not going to Bombay, I’m not going anywhere. I’ll stay here with
you.
These examples are relatively rare but all the more striking. In (54) we have not
only a reversed order of verb and negative particle but also the combination of
ache with na. Ferguson has another similar example:
 NEGATION PATTERNS IN BENGALI 259
(56) sekhane na ache ghor-dor, na ache kichu
There are no houses there, there is nothing.
It seems from these few examples that this word order is linked to an
anaphoric sentence arrangement, a string of parallel structures. This use of na
before a finite verb form can be considered a rhetoric device for emphasis. One
other instance of preceding na is found in (57) with a third person imperative:
(57) amar chele ache. ar keu asa roy na dik, se debei
My son is there. Let no one else give shelter—he will give it.
It seems likely to me that this is the normal position of a negation with a third
person imperative rather than an emphatic device, but I have not seen enough
examples to corroborate this. For the moment, I will classify it here.
(iii) The system of subordinating clauses and their relationship to the main
clause is quite different in Bengali from in English. Although my research in
this area is still incomplete, I can make some preliminary comments. Every
subordinating conjunction has a corresponding conjunction in the main clause.
In fact, the existence of a correlative in the main clause can be taken as a
definitional feature for subordination in Bengali. Conjunctions without a cor-
relative are coordinating. We have here a very interesting overlap of word
classes which has not been recognized as such. The familiar concept of relatives
and co-relatives is recognized for relative clauses of the type ye-se ‘he-who’,
ya-ta ‘that which’ and by extension for conjunctions like yemon-temon as-as. It
also applies to pairs like yodi-tahole ‘if’, yodior -tobuor ‘although’, yokhon-tokhon
‘when’. Instead of having a clear distinction between pronouns and conjunc-
tions, we have the phenomenon of correlation, to which pronouns as well as
conjunctions are subject.
But how does this affect negation patterns? The particle na can precede
a finite verb form in subordinate clauses, but not in main clauses, except for
emphatic use as shown above. With some conjunctions this change in word
order is almost obligatory, e.g. in sentences with yodi ‘if’.
tumi yodi na yaor if you don’t go
uni yodi ese na thaken if he hasn’t arrived
br DsDti yodi na hoto if it didn’t rain
s
With other conjunctions the positioning of na seems to be linked to aspectual
features such as progressiveness:
(58) yotoksD onD na rumi asche ekhanei opeksD a korben
As long as Rumi didn’t turn up he would wait for her here.
In the same novel, a few paragraphs later, the same concept is given with
yokhon—na appears after the verb.
(59) yokhon rumi elo na tokhon sor ja keyatolay phire elen
When Rumi didn’t turn up he went straight back to Keyatala.
The difference between yotoksD onD and yokhon is their extension in time. While
yotoksD onD refers to an extended period ‘as long as, for as many moments as’,
yokhon gives us a point in time ‘when, at that moment’. In another context we
find:
(60) kintu yotodin na prthibi dhvomd so hocche, totodin tor manusD er onner
s
somd sthan korte hobe
260 HANNE-RUTH THOMPSON

But as long as the world is not on the point of extinction, provisions

have to be made for man to eat.
We can therefore provisionally state that with durative conjunctions the nega-
tive particle is drawn into the phrase, appears before the verb form and often,
but not always, forms a close semantic unit with the conjunction, implying ‘as
long as not’, whereas with positional (point in time) conjunctions, na comes
after the verb form.
In an essay on the difficulties of unifying the language, very poignantly,
there is:
(61) yotoksD onD niroksD orta dum r na hobe ebomd desa er sorbosadharonD ontoto
ekhonkar obosthay or ek sa o joner modhye sa otokora DsatD jon sotyii
as iksD alabh na korbe totodin bamd labhasD ar or unnoti duDhsadhyo
As long as illiteracy is not abolished and while at least at the moment
60% of the population don’t benefit from education, an improve-
ment in the Bangla language is hard to accomplish.
What does the author mean by improving the Bangla language? There does
not seem to be much wrong with it and it is clear that the negative particle
plays its own subtle role in the semantic interplay of sentence parts and in
retaining the internal balance which is so typical for this language. In the cases
where na is placed before the verb it loses its exposed sentence-final power but
it gains a stress which is never present when na follows the verb.
(iv) na preceding a non-finite verb form: I started with the hypothesis that in
Bengali all negation is on the sentence level, that na at the end of a sentence
negates the whole sentence. But this somewhat rigid regime is loosened up
in connection with non-finite verb forms. Non-finite verb forms play a crucial
role in Bengali syntax. In fact, they are so prominent that where two or more
actions or events are related and reported together, a non-finite verb form is
almost always the first port of call. Subordinate clauses with finite verb forms
are usually only the back-up option. Let us have a brief look at what non-finite
verb forms can do in places where English has subordinate clauses.
English subordinating conjunctions can be grouped into the following
semantic areas:
• Time (after, before, when, while, since, until)
The perfective participle can deal with anything related to sequences of
events (when, after, before). The present participle exists for the purpose of
expressing simultaneity (while). For ‘since’ and ‘until’ the verbal noun with
the postpositions theke ‘from’ and poryonto ‘until’ are the obvious choices.
• Cause and effect (because, since, now that, as, in order that, so)
The locative verbal noun can express cause, so does the genitive verbal
noun with the postposition jonyo ‘for’ (because, since) or the postposition
pore ‘after’ (now that). Anything relating to purpose or effect can be
expressed through an infinitive (in order that/in order to).
• Opposition (although, even though, whereas, while)
The adversative particle or added to the conditional participle or to the
perfective participle can cover most of these meanings. Verbal nouns
with postpositions such as sottveor ‘despite’ or poriborte ‘instead of’ have a
similar function.
• Condition (if, unless, whether or not, even if, in case)
Again, the conditional participle is the main candidate for expressing these
concepts. Verbal nouns with the postposition pache ‘unless’ offer another
option.
 NEGATION PATTERNS IN BENGALI 261
This very roughly indicates the wide range of non-finite uses in Bengali and it
is no wonder that negation patterns are also affected by it. By positioning the
negative particle na before non-finite verb forms we gain dramatically the pos-
sibility of internal negation. I will restrict myself to two examples for each verb
form. There is often a considerable difference between the Bengali sentence
structure and an idiomatic English translation, so an approximate gloss is also
given.
—Verbal noun
(62) tai byapartD a ar narD acarDa na korai bhalo
so this matter more stirring not doing good
‘So it would be good not to stir up this matter any further’.
(63) tini na asbar poryonto kor nor kaj hobe na
he until not coming any work will not be
‘Nothing will get done until he arrives’.
The nominal character of verbal nouns has the effect that a preceding na
becomes almost equivalent to a prefix so that we think of non-arrival, non-
occurrence. But this melding together of negative and verb form is not
restricted to verbal nouns.
—Conditional participle
Glosses for the conditional participle are given as ‘if doing, going, saying’,
etc. This is not very idiomatic in English but it preserves the person-neutral
character of the form.
(64) kokhonor kokhonor jor r na korle cole na
sometimes force if not doing does not go
‘A little bit of force now and then is necessary’.
(65) sokal na hole kichui kora yabe na
morning if not becoming something the doing will not go
‘Nothing can be done before morning’.
In (65) we have the very common form na hole ‘if not becoming’, which is used
by itself as a conjunction meaning ‘if not, otherwise’. Here again one feels that
the negative and the conditional participle meld into one semantic unit.
(66) na paor ya gele ek ek kore
not finding if going one by one
onyo haramjadagulor r bar
Di yabo
the houses of the other scoundrels I will go
If she is not found I will go to every single house of the other
scoundrels.
In sentence (66) it is difficult to decide whether the negative particle belongs to
the verbal noun or to the conditional participle. The two non-finites form such
a close unit that the negative embraces both forms and a change to paor ya na
gele ‘the finding, if not going’, would, in my opinion, make no difference.
The occurrence of na hole in sentence (65) and dealing with conditionals
brings us back to the peculiar role of noile ‘if not, otherwise’, which was men-
tioned earlier on. As I said, I feel that this form has moved away from its origi-
nal, verbal character and has become an isolated lexical unit. There are many
cases where verb forms take on the role of another word class. We have diye
‘having given’, niye ‘having taken’, kore ‘having done’, theke ‘having stayed’—
all of these can all but lose their verbal character when used as postpositions
262 HANNE-RUTH THOMPSON

or adverbs, but they can at any time revert to being full verb forms. The form
noile has lost its environment. There are cases where na hole and noile are
interchangeable, but in (65), for instance, noile could not be used because it is
no longer capable of fulfilling its conditional function within a syntactic struc-
ture. It is not possible to say *sokal noile nor even *ta noile. Let us look at the
following three sentences to establish the uses of noile.
—noile
(67) prothome khor̃ j kore dekhben ekhon histD ri klas hocche kina. yodi hoy
opeksD a korben. noile ei DdipartD men
D Dter chatrochatri beyara yake pan
take onuror dh korben...
First he will find out if there is a history class going on now. If so he
will wait. If not, he will ask the students or staff of this department
...
In this sentence noile appears in direct juxtaposition to yodi hoy ‘if it is’ and
has its full conditional character.
(68) rate ghumer modhye setD a bujhiya theke geche. noile sokal utD htei take
mone porD be keno
This understanding must have come to him during his sleep in the
night. Otherwise why would it be on his mind as soon as he woke up
in the morning?
The syntactic relationships are getting blurred here. There is no longer any
sense of equating a subject with a complement—instead noile calls the whole
previous statement into question. This is by far the most common use of noile,
at the beginning of a sentence with no syntactic ties. Sentence (69) makes this
even clearer:
(69) tãr esomoy kolkatay asa tader poksD e bhaloi hoyeche, noile ei mase
bombe chutD te hoto
His coming to Kolkata at this time was convenient for them.
Otherwise they would have had to shoot off to Bombay.
We have a perfect equational relationship in the first sentence, which can be
rendered with a zero verb and negated with noy:
tãr kolkatay asa tader poksD e bhalo (noy)
His coming to Kolkata is (not) convenient for them.
The second sentence, starting with noile is, however, not concerned with this
equation, only with the consequences of his whereabouts. noile has lost its
verbal character and become a coordinating conjunction meaning ‘otherwise’.
I take this as sufficient grounds to claim (see p. 248) that noile is no longer,
either syntactically or semantically, used as the conditional participle of the
verb no-.
—Perfective participle
(70) kor nor uttor na diye tini takiye thaken
any answer not having given he having stared stays
He kept staring without answering.
(71) e kotha svikar na kore upay nei
this word not having admitted way is not
There is no denying this.
 NEGATION PATTERNS IN BENGALI 263
The perfective participle is, of course, the main candidate for sentence-internal
negation, simply because it is the most common non-finite form. Negated
perfective participles can, as in (70), often be translated with ‘without’ when we
are looking at a sequence of events, but we need to have a closer look at the
catchment area of the sentence-final negative particle. I was surprised to find
the following interpretation in Clinton Seely’s Intermediate Bengali (p. 306).
He interprets the sentence
sekhane giye tar sonq ge dekha korlam na
there having gone with him seeing I did not as
*I didn’t go there and I didn’t meet her,
but accepts a separation of the two verb forms in:
sekhane giyeor tar sonq ge dekha korlam na
there although having gone with him seeing I did not as
I went but didn’t meet her.
I can only assume that this is a straightforward mistake. The concessive or in
the second sentence adds a semantic nuance but it cannot change the truth
conditions of the sentence. In both sentences the act of going is not subject to
the negation. The interpretation of these sentences may have been influenced
by mixed compound verb structures such as niye yaor ya, bhule yaor ya, rekhe
yaor ya, whose internal set-up is very different to that of two sequential but
separate verbs. In
ami take niye yaini
I him having taken I did not go
we cannot accurately guess at the real events without more information. In
most cases the sentences will mean: ‘I went but I didn’t take him with me’,
which means that the sentence-final negation affects the first, but not the
second of the two verbs. The more usual way of stressing the fact that the visit
happened without him would be to say:
ami take rekhe giyechi
I him having left I went
but the structure with niye yaor ya is perfectly acceptable. The reason for this
somewhat unorthodox state of affairs is quite simply the fact that these com-
pounds are in a constant process of semantic bleaching and the finite verb
yaor ya ‘go’ can at times be an almost meaningless attachment to the main verb
neor ya, as it is in the compound bhule yaor ya ‘forget’, where in most cases
no ‘going’ takes place at all. But the situation is quite different where two
consecutive actions or events are described. Since every perfective participle is
capable of having its own negative particle, there is no reason to assume that
a sentence-final negation extends to preceding verbal actions and I have seen
no evidence of it.
—Infinitive
(72) tader sonq ge sob somoy jhogr D a na korte para yay na
with them always not to quarrel be able does not go
It is impossible not to quarrel with them all the time.
(73) se din theke onuradhar sonq ge kotha na bolte arombho korechilen
from that day with Anuradha not to speak he began
From that day onwards he started not speaking to Anuradha.
264 HANNE-RUTH THOMPSON

In both of these sentences we have conjunct verbs, a noun-verb combination

which is negated. This is no coincidence. Combinations of a simple verb infini-
tive and a following verb like para ‘be able to’ caor ya ‘want’, yaor ya ‘go’, deor ya
‘allow’ are much less likely to have a negated infinitive, e.g. ? ami take na yete
dilam ‘I allowed him not to go’. This is due to the semantic awkwardness of
such structures more than to a syntactic consideration, and Bengali is more
likely to reformulate the concept in a positive sentence, e.g. ami take thakte
dilam ‘I allowed him to stay’. An infinitive followed by another verb often
functions as the object of this verb, the two verb forms represent one concept
and this is why internal negation is rare but syntactically quite possible,
particularly in negative sentences:
se take bhat na khete deyni.
She didn’t allow him not to eat.

V. Conclusion
Linguistic phenomena like negation do not occur in isolation. Perhaps this
is one reason why Bengali linguists have been somewhat negligent of this
feature—it simply does not occur to them to separate negatives from the rest
of the language. Every occurrence of a negative rubs shoulders with features
like tense, syntactic hierarchy, the status of a verb form, case issues and word
order; and at every step of the analysis one has to avoid getting sucked into the
uncertain quagmire of other grammatical issues. Separating out the grammati-
cal feature of negation, as I have done here, can be seen as an artificial exer-
cise, an isolation of only one piece of the puzzle, but a close observation of one
area gives us not only a focus but a firmer starting point for these other issues
and a much better understanding of which questions need to be asked. As a
result of this analysis of negation patterns, we can identify the following areas
for further research:
• case use;
• tense use;
• syntactic subordination;
• the overlap of verbs of being in Bengali.
In the following areas some progress has been achieved. We have seen:
• the uses of nei and no- and a clearer dividing line between the two uses;
• the semantic range and tense use of the negative particle ni;
• a classification of noile as a lexical conjunction;
• the flexibility gained by internal negation;
• a new way of looking at subordination in Bengali.

REFERENCES

Primary literature—Journals
Bangla Journal Vol 4, 1–2, Toronto, April–August 2002
Vol 5, 1, Toronto, December 2004
Kali o kolom (Ink and Pen), Vol. 7, Dhaka, August 2004
Desh, Kolkata, February 2005

Fiction
Ahmed, Humayun. 1991. Ekjon mayaboti. Dhaka: Noroj Kitabistan.
Bandopadhyay, Atin. 1987. Ekti joler rekha. Kolkata: Pratibhash.
Choudhuri, Abdul Rouf. 2003. Pordeshe porbasi, Inglyander dinguli. Dhaka: Pathok Shomabesh
Books.
 NEGATION PATTERNS IN BENGALI 265
Debi, Moytreyi. 1986. Bidhi o bidhata. Kolkata: Praima Publications.
Debsen, Nobonita. 1996. Sita theke shuru, Kolkata: Anondo Publishers.
Haq, Hassan Azizul. 1973. Jibon ghoshe agun. Dhaka: Khan Brothers.
Haydar, Rashid. 1988. Dhaka: Mabuhai Aksar.
Hossein, Selina. 1987. Nirontor ghontadhvoni. Dhaka: Ahmed Publishing House.
Majumdar, Samaresh. 1998. Daybondhon. Kolkata: Anondo Publishers.
Ray, Praphullo. 1997. Otol joler dike. Kolkata: Dey’s Publishing.
——. 1981. Akaser nice manush. Kolkata: Dey’s Publishing.

Non-fiction
Poscimbanga Bangla Academy. 1993. Prosongo banglabhasha (essays).
Ara, Begum Jahan. 2005. Bangla bhasha prosongo. Dhaka: Hemi Books.
Murshid, Ghulam. 1992. Kalantore bangla godyo. Kolkata: Ananda Publishers.

Secondary literature
Anderson, James Drummond (the elder). 1920. A Manual of the Bengali Language. Cambridge
(reprint, 1962).
Bangla Academy. 1993. Prosongo Banglabhasha (essays). Kolkata: Poscimbanga Bangla Academy.
Chatterji, Suniti Kumar. 1926. The Origin and Development of the Bengali Language. Kolkata:
Calcutta University Press.
Dakshi, Alibha. 1995. Learning Bengali, A Self-Tutor on a Phonetic Basis. Kolkata: Subarnarelcha.
Islam, Rafiqul. 2001. Bangla byakoron somiksa. Dhaka: Globe Library.
Milne, W. S. 1913. A Practical Bengali Grammar. Kolkata
Muhammad, Qazi Din. 1963. Some Syntactic Structures of Bengali in PL (Pakistani Linguistics).
——. 1969. Bangla Kryiapod. Dhaka: Bangla Academy Potrika.
Ramchand, Gillian. 2001. ‘Two types of negation in Bengali’ (unpublished).
Sarkar, Pabitra. 1998. Poket Bangla Byakoron. Kolkata: Ajkal Publishers.
Seely, Clinton B. 2002. Intermediate Bengali. Munich: Lincom Europa.
Smith, W.L. 1997. Bengali Reference Grammar. Stockholm: Association of Oriental Studies.
Thompson, Hanne-Ruth (PhD) ‘Toward a definitive grammar of Bengali: a practical study and
critique of research or selected grammatical structures’, SOAS, 2003.