topics: |  linguistics | india |

What is interesting about this book is that none of the chapters on these languages of India and Africa are contributed by any scholar in these regions. The only Indian in the work, Yamuna Kachru, has been at the U. Chicago for many years. Also interesting is the fact that the articles written by europeans cite almost no non-european authors, e.g. the articles on Sanskrit, Arabic, Dravidian, Swahili. Suniti Chatterji (bAnglA) and A. Bamgbose (Yoruba) are the major exceptions.

2 Sanskrit : George Cardona

Sanskrit saMskr^ta ( ‘adorned, purified’) refers to several varieties of Old Indo-Aryan, whose most archaic forms are found in Vedic texts: the Rigveda , Sāmaveda, Atharvaveda, Yajurveda... the early sūtra works can properly be called late Vedic. Also of the late Vedic period is the grammarian pANini (not later than early fourth century BC), author of the aShTAdhyAyI, who distinguishes between the language of sacred texts (chhandas) and a more usual language of communication bhAShA (bhASh, from ‘speak’), tantamount to Classical Sanskrit.

Epic Sanskrit is so called because it is represented principally in the two
epics, Mahābhārata and rAmAyaNa. The date of composition for the core of
early epic is considered to be in the first centuries BC.  It is in the
rAmAyaNa that the term saMskr^ta is encountered probably for the first time
with reference to the language. Classical Sanskrit is the language of major
poetical works, dramas, tales and technical treatises on grammar,
philosophy and ritual.  It was not only used by Kalidasa and his
predecessors but continued in use after Sanskrit had ceased to be a
commonly used mother tongue. 

Sanskrit is used as a lingua franca by paNDitas from different parts of
India, and several thousand people claim it as their mother tongue.

1.2 Diachronic Changes Within Sanskrit

Linguistic changes are discernible in Sanskrit from earliest Vedic down to
the language pANini describes.  The nominative plural masculine in -āsas
(devāsas ‘gods’), which has a counterpart in Iranian, is already less
frequent in the Rigveda than the type in -ās (devās), and continues to lose
ground; in brAhmaNas, -ās is the norm. 

Bibliography

Burrow (1965) is a summary of the prehistory and history of Sanskrit,
including Vedic, with references to Middle Indo-Aryan; somewhat personal
views are given in places, but the work remains valuable. For a good summary
of views on the dialects of Old Indo-Aryan, with discussion of theories
proposed and references, see Emeneau (1966).  The standard reference grammar
is Whitney (1889). Renou (1956) is an insightful summary of the grammar,
vocabulary and style of different stages of Sanskrit, including Vedic, with
text selections and translations. Wackernagel (1896-) is the most thorough
reference grammar of Sanskrit, but remains incomplete: the published volumes
are: I (Lautlehre), reissued with a new ‘Introduction générale’ by L.Renou
and ‘Nachträge’ by A.Debrunner (1957); II, 1 (Einleitung zur Wortlehre,
Nominalkomposition), 2nd ed. with ‘Nachträge’ by A.Debrunner (1957); II, 2
(Die Nominalsuffixe), by A.Debrunner (1954); III (Nominalflexion—Zahlwort—
Pronomen) (1930); there is also a Register zur altindischen Grammatik von J.
Wackernagel und A.Debrunner by R.Hauschild (1964).

References

Burrow, T. 1965. The Sanskrit Language, 2nd ed. (Faber and Faber, London)
Emeneau, M.B. 1966. ‘The Dialects of Old Indo-Aryan’, in H.Birnbaum and
    J. Puhvel (eds.), Ancient Indo-European Dialects (University of
    California Press, Berkeley and Los Angeles), pp. 123–38.
Renou, L. 1956. Histoire de la langue sanskrite (IAC, Paris)
Wackernagel, J. 1896–. Altindische Grammatik (Vandenhoek and Ruprecht, Göttingen)
Whitney, W.D. 1889. Sanskrit Grammar, Including Both the Classical Language
    and the Older Dialects, of Veda and Brahmana, 2nd ed. (Harvard University
    Press, Cambridge, Mass.) 

3 Hindi-Urdu : Yamuna Kachru p.37

Urdu, a language closely related to Hindi, is spoken by twenty-three million
people in India and approximately eight million people in Pakistan as a
mother tongue. It is the official language of Pakistan and the state language
of the state of Jammu and Kashmir in India.

Three stages in the evolution of Indo-Aryan languages

It is difficult to date the beginnings of the New Indo-Aryan languages of
India. Scholars generally agree that the development of Indo-Aryan languages
of India took place in three stages. The Old Indo-Aryan stage is said to
extend from 1500 BC to approximately 600 BC. The Middle Indo-Aryan stage
spans the centuries between 600 BC and AD 1000. The Middle Indo-Aryan stage
is further subdivided into an early Middle Indo-Aryan stage (600–200 BC), a
transitional stage (200 BC–AD 200), a second Middle Indo-Aryan stage (AD
200–600), and a late Middle Indo-Aryan stage (AD 600–1000). The period
between AD 1000–1200/ 1300 is designated the Old New Indo-Aryan stage because
it is at this stage that the changes that began at the Middle Indo-Aryan
stage became established and the New Indo-Aryan languages such as Hindi,
Bengali, Marathi etc. assumed distinct identities. [...]

The name Hindi is not Indian in origin; it is believed to have been used by
the Persians to denote the peoples and languages of India (Verma 1933). Hindi
as a language is said to have emerged from the patois of the market place and
army camps during the period of repeated Islamic invasions and establishment
of Muslim rule in the north of India between the eighth and tenth centuries
AD. The speech of the areas around Delhi, known as khARI bolī, was adopted by
the Afghans, Persians and Turks as a common language of interaction with the
local population. In time, it developed a variety called urdū (from Turkish
ordu ‘camp’).  This variety, naturally, had a preponderance of borrowings
from Arabic and Persian. Consequently, it was also known as rextā ‘mixed
language’. The speech of the indigenous population, though influenced by
Arabic and Persian, remained relatively free from large-scale borrowings from
these foreign languages. In time, as Urdu gained some patronage at Muslim
courts and developed into a literary language, the variety used by the
general population gradually replaced Sanskrit, literary Prakrits and
apabhraMShas as the literary language of the midlands (madhyadeśa). This
latter variety looked to Sanskrit for linguistic borrowings and Sanskrit,
Prakrits and apabhraMShas for literary conventions. It is this variety that
became known as Hindi. Thus, both Hindi and Urdu have their origins in the
khARI bolī speech of Delhi and its environs although they are written in two
different scripts (Urdu in Perso-Arabic and Hindi in Devanāgarī). The two
languages differ in minor ways in their sound system, morphology and
syntax. 

Hindi and Urdu have a common form known as Hindustani which is essentially a
colloquial language (Verma 1933). This was the variety that was adopted by
Mahatma Gandhi and the Indian National Congress as a symbol of national
identity during the struggle for freedom. It, however, never became a
language of literature and high culture (see Bhatia 1987 for an account of
the Hindi-Urdu-Hindustani controversy in the late nineteenth and early
twentieth centuries).

Both Urdu and Hindi have been in use as literary languages since the twelfth
century. The development of prose, however, begins only in the eighteenth
century under the influence of English, which marks the emergence of Hindi
and Urdu as fully-fledged literary languages.

Bibliography

The standard reference grammar for modern standard Hindi is Guru
(1920). Other reference grammars are Sharma (1958) and McGregor (1972), the
latter directed to the needs of learners of Hindi. Kellogg (1875) describes
Hindi and the major dialects of the Hindi area, and contains a good
introduction to Hindi prosody; data are drawn mostly from literary texts of
the period and the work is hence dated. For Urdu, see Bailey (1956). For an
account of the parallel development of Hindi-Urdu see Rai (1984).  Ohala
(1983) is a phonological description of Hindi, while Kachru (1980) describes
syntactic constructions of Hindi in non-technical language. Verma (1933) is a
brief sketch of the history of the Hindi language. Kachru (1981) contains a
supplement on transplanted varieties of Hindi-Urdu and one on transitivity in
Hindi-Urdu. Bhatia (1987) discusses the native and non-native grammatical
tradition.

References
Bailey, T.G. 1956. Teach Yourself Urdu (English Universities Press, London)
Bhatia, T.K. 1987. A History of Hindi (Hindustani) Grammatical Tradition
    (E.J.Brill, Leiden)
Guru, K.P. 1920. Hindi (Kashi Nagri Pracharini Sabha, Banaras)
Kachru, Y. 1980. Aspects of Hindi Grammar (Manohar Publications, New Delhi)
Kachru, Y. 1981. ‘Dimensions of South Asian Linguistics’, Studies in the
    Linguistic Sciences, vol. 11, no. 2
Kellogg, S.H. 1875. A Grammar of the Hindi Language (Routledge and Kegan Paul, London)
McGregor, R.S. 1972. Outline of Hindi Grammar (Oxford University Press, London)
Ohala, M. 1983. Aspects of Hindi Phonology (Motilal Banarsidass, Delhi)
Rai, A. 1984. A House Divided (Oxford University Press, Delhi)
Sharma, A. 1958. A Basic Grammar of Modern Hindi (Government of India,
    Ministry of Education and Scientific Research, Delhi)
Verma, Dh. 1933. Hindi ka itihas (Hindustani Academy, Allahabad)

4 Bengali : M.H.Klaiman

Historical and Genetic Setting

Bengali, together with Assamese and Oriya, belongs to the eastern group
within the Magadhan subfamily of Indo-Aryan. In reconstructing the
development of Indo-Aryan, scholars hypothetically posit a common parent
language from which the modern Magadhan languages are said to have
sprung. The unattested parent of the Magadhan languages is designated as
Eastern or Magadhi APABHRAMSA, and is assigned to Middle Indo-Aryan. Apart
from the eastern languages, other modern representatives of the Magadhan
subfamily are Magahi, Maithili and Bhojpuri.

Within the eastern group of Magadhan languages, the closest relative of
Bengali is Assamese. The two share not only many coincidences of form and
structure, but also have in common one system of written expression, on which
more details will be given later.

Historically, the entire Magadhan group is distinguished from the remaining
Indo- Aryan languages by a sound change involving sibilant
coalescence. Specifically, there occurred in Magadhan a falling together of
three sibilant elements inherited from common Indo-Aryan, dental /s/, palatal
/š/ and retroflex /s./. Among modern Magadhan languages, the coalescence of
these three sounds is manifested in different ways; e.g. the modern Assamese
reflex is the velar fricative /x/, as contrasted with the palatal /š/ of
Modern Bengali.  

The majority of Magadhan languages also show evidence of historical
regression in the articulation of what was a central vowel /ă/ in common
Indo-Aryan; the Modern Bengali reflex is .  Although the Magadhan subfamily
is defined through a commonality of sound shifts separating it from the rest
of Indo-Aryan, the three eastern languages of the subfamily share one
phonological peculiarity distinguishing them from all other modern Indo-Aryan
languages, both Magadhan and non-Magadhan. This feature is due to a
historical coalescence of the long and short variants of the high vowels,
which were distinguished in common Indo-Aryan. As a result, the vowel
inventories of Modern Bengali, Assamese and Oriya show no phonemic
distinction of /ĭ/ and /ī/, /ŭ/ and /ū/. Moreover, Assamese and Bengali are
distinguished from Oriya by the innovation of a high/low distinction in the
mid vowels. Thus Bengali has /æ/ as well as /e/, and /c-inv/ (/O/) as well as
/o/. Bengali differs phonologically from Assamese principally in that the
latter lacks a retroflex consonant series, a fact which distinguishes
Assamese not just from Bengali, but from the majority of modern Indo-Aryan
languages.

Besides various phonological characteristics, there are certain grammatical
features peculiar to Bengali and the other Magadhan languages. The most
noteworthy of these features is the absence of gender, a grammatical category
found in most other modern Indo-Aryan languages. Bengali and its close
relative Assamese also lack number as a verbal category. More will be said on
these topics in the section on morphology, below.

Writing and literature have played no small role in the evolution of Bengali
linguistic identity. A common script was in use throughout eastern India
centuries before the emergence of the separate Magadhan vernaculars. The
Oriya version of this script underwent special development in the medieval
period, while the characters of the Bengali and Assamese scripts coincide
with but a couple of exceptions.

the availability of a written form of expression was essential to the
development of the rich literary traditions associated not just with Bengali,
but also with other Magadhan languages such as Maithili. However, even after
the separation of the modern Magadhan languages from one another, literary
composition in eastern India seems to have reflected a common milieu scarcely
compromised by linguistic boundaries.  Although vernacular literature appears
in eastern India by AD 1200, vernacular writings for several centuries
thereafter tend to be perceived as the common inheritance of the whole
eastern area, more so than as the output of individual languages.

This is clearly evident, for instance, in the case of the celebrated Buddhist
hymns called the caryA-pada, composed in eastern India roughly between AD
1000 and 1200.  Though the language of these hymns is Old Bengali, there are
reference works on Assamese, Oriya and even Maithili that treat the same
hymns as the earliest specimens of each of these languages and their
literatures.  Bengali linguistic identity is not wholly a function

Shift in the course of the Ganga

there was a massive shift in the course of the Ganges River between the
twelfth and sixteenth centuries AD, concomitant to the rise of Islamic rule.
Whereas it had earlier emptied into the Bay of Bengal nearly due south of the
site of present-day Calcutta, the river gradually approached and eventually
became linked with the Padma River system in the territory today called
Bangladesh.  The shift in the Ganges has been one of the greatest influences
upon material history and human geography in eastern India; for, prior to the
completion of the river’s change of course, the inhabitants of the eastern
tracts had been virtually untouched by civilisation and sociocultural
influences from without, whether Islamic or Hindu. Over the past four
centuries, it is the descendants of the same people who have come to make up
the majority of speakers of the Bengali language; so that the basis of their
Bengali identity is not genetic and not religious, but linguistic.

With over 70 million native speakers in India and over 100 million in
Bangladesh, Bengali has perhaps the sixth largest number of native speakers
among the languages of the world, considerably more than such European
languages as Russian, German and French.

Orthography

[derived from Brahmi]

Bengali orthography reads from left to right, and is organised according to
syllabic rather than segmental units.  Accordingly, a special diacritic or
character is employed to represent a single consonant segment in isolation
from any following vowel, or a single vowel in isolation from any preceding
consonant. 

the first consonant character is called kO [/O/ = inv-c]. The designation of
the latter is such, because the corresponding sign in isolation is read not
as a single segment, but as a syllable terminating in /O/, the so-called
‘inherent vowel’.

Vowels:

অ               ক     O         ri ঋ          কৃ 	ri  
আ               কা   a             এ          েক 	e 
hrOsso i ই      িক   i            ঐ          ৈক 	oy
dirgho i ঈ      কী	           ও          কো 	o 
hrOsso u উ      কু    u             ঔ          কৌ 	ow
dirgho u ঊ      কূ    u   -       

Consontants: 

ক k     চ c   ট t.    ত ৎ t   প p  
খ kh    ছ ch  ঠ t.h   থ th    ফ ph 
গ g     জ j   ড d.    দ d     ব b  
ঘ gh    ঝ jh  ড় r.    ধ dh    ভ bh 
ঙ ং ṅ  ঞ ñ   ঢ d.h   ন n     ম m  
              ঢ় r.h
              ণ n. 

Ontostho jO য j
Ontostho c য় y, w
র r
ল l
talobbo sO শ s´
murdhonno sO ষ s.
donto sO স s
হ ঃ h

Special diacritics : cOndrobindu ৺
		     hOsonto ৲

Table 23.2 Segmental Phonemes of Bengali

Consonants
		     Labial  Dental Retroflex  Palatal Velar Post-velar
Obstruents
voiceless:
unaspirated		p	t	t.	c	k
aspirated 		ph 	th 	t.h 	ch 	kh
voiced:
unaspirated		b	d	d.	j	g
aspirated 		bh 	dh 	d.h 	jh 	gh
Nasals 			m 	n 	n.		ṅ
Flaps 				r 	r.
Lateral 			l
Spirants					s		h

Vowels
			Front 		Back
High 			i 		u
High mid 		e 		o
Low mid 		æ 		O (/inv-c/)
Low 				a

Vowel raising

The most interesting Modern Bengali phonological processes involve the vowel
segments to the relative exclusion of the consonants. One process, Vowel
Raising, produces a neutralisation of the high/low distinction in the mid
vowels, generally in unstressed syllables. Given the stress pattern of the
present standard dialect, which will be discussed later, Vowel Raising
generally applies in non-word-initial syllables. Evidence for the process is
found in the following alternations:

	mOl ‘dirt’    -> Omol ‘pure’
	sO ‘hundred’  -> ækso ‘one hundred’
	æk ‘one’      -> Onek ‘many’

A second phonological process affecting vowel height is very significant
because of its relationship to morphophonemic alternations in the Bengali
verbal base. This process may be called Vowel Height Assimilation, since it
involves the assimilation of a non-high vowel (other than /a/) to the nearest
succeeding vowel segment within the phonological word, provided the latter
has the specification [+ high]. Outside the area of verbal morphophonemics,
the evidence for this process principally comes from the neutralisation of
the high/low distinction in the mid vowels before /i/ or /u/ in a following
contiguous syllable. Some alternations which illustrate this process are:

	æk ‘one’ 	ekt.i ‘one’ (plus classifier -t.i)
	lOjja ‘shame’ 	lojjito ‘ashamed’
	nOt.  		‘actor’ not.i ‘actress’
	æk ‘one’ 	ekt.u ‘a little, a bit’
	tObe ‘then’ 	tobu ‘but (then)’

Higher frequency of the retroflex

the retroflex flap /r. / of Bengali has no counterpart in Sanskrit, and
its presence in modern standard Bengali (and in some of its sisters) is due
to a phonological innovation of Middle Indo-Aryan. Furthermore, while the
other retroflex segments of Modern Bengali (/t./, /t.h/, /d. /, /d. h/) have
counterparts in the Old Indo-Aryan sound system, their overall frequency
(phonetic load) in Old Indo-Aryan was low.
Modern Bengali (along with the other Magadhan languages, especially the
eastern Magadhan languages) demonstrates a comparatively high frequency of
retroflex sounds. Some external, i.e. non-Aryan influence on the diachronic
development of the Bengali sound system is suggested.

Further evidence of probable non-Aryan influence in the phonology is to be
found in the peculiar word stress pattern of Modern Bengali. Accent was
phonemic only in very early Old Indo-Aryan, i.e. Vedic. Subsequently,
however, predictable word stress has typified the Indo-Aryan languages; the
characteristic pattern, moreover, has been for the stress to fall so many
morae from the end of the phonological word. Bengali word stress, though, is
exceptional. It is non-phonemic and, in the standard dialect, there is a
strong tendency for it to be associated with word-initial syllables.

3 Morphology

Morphology in Modern Bengali is non-existent for adjectives, minimal for
nouns and very productive for verbs. Loss or reduction of the earlier
Indo-Aryan adjective declensional parameters (gender, case, number) is fairly
typical of the modern Indo- Aryan languages; hence the absence of adjectival
morphology in Modern Bengali is not surprising. Bengali differs from many of
its sisters, however, in lacking certain characteristic nominal
categories. The early Indo-Aryan category of gender persists in most of the
modern languages, with the richest (three-gender) systems still to be found
in some of the western languages, such as Marathi. Early stages of the
Magadhan languages (e.g. Oriya, Assamese and Bengali) also show evidence of a
gender system.

In Modern Bengali, it is only in a few relic alternations (e.g. the earlier
cited pair nOt.  ‘actor’/not.i ‘actress’) that one observes any evidence
today for the system of nominal gender which once existed in the language.

Noun declension: 
			Singular	   Plural
Nominative 		NULL 		   -ra/-era; -gulo
Objective 		-ke 		   -der(ke)/-eder(ke); -guloke
Genitive 		-r/-er 		   -der/-eder; -gulor
Locative-Instrumental	-te/-e or -ete 	   -gulote

parameters not revealed in this table are animacy, definiteness and
determinacy.

Generally, the plural markers are added only to count nouns having animate or
definite referents; otherwise plurality tends to be unmarked. Compare,
e.g. jutogulo dO rkar ‘the (specified) shoes are necessary’ versus juto dO
rkar ‘(unspecified) shoes are necessary’.  Further, among the plurality
markers listed in Table 23.3, -gulo (nominative), -guloke (objective), -gulor
(genitive) and -gulote (locative-instrumental) are applicable to nouns with
both animate and inanimate referents, while the other markers co-occur only
with animate nouns.  Hence: chelera ‘(the) boys’, chelegulo ‘(the) boys’,
jutogulo ‘the shoes’, but *jutora ‘the shoes’.

The usage of the objective singular marker -ke, listed in Table 23.3, tends
to be confined to inanimate noun phrases having definite referents and to
definite or determinate animate noun phrases.  Thus compare kichu (*kichuke)
caichen ‘do you want something?’ with kauke (*kau) caichen ‘do you want
someone?’; but: pulis caichen ‘are you seeking a policeman/ some policemen?’
versus puliske caichen ‘are you seeking the police?’.

the sole parameters for subject–verb agreement in Modern Bengali are person
(three are distinguished) and status. the Bengali verb is marked for three
status categories (despective/ordinary/honorific) in the second person and
two categories (ordinary/honorific) in the third.

The most interesting area of Bengali morphology is the derivation of
inflecting stems from verbal bases. Properly speaking, a formal analysis of
Bengali verbal stem derivation presupposes the statement of various
morphophonological rules. However, for the sake of brevity and clarity, the
phenomena will be outlined below more or less informally.

But before the system of verbal stem derivational marking can be discussed,
two facts must be presented concerning the shapes of Bengali verbal bases,
i.e. the bases to which the stem markers are added.

First, Bengali verbal bases are all either monosyllabic (such as jan- ‘know’)
or disyllabic (such as kamr. a- ‘bite’). The first syllabic in the verbal
base may be called the root vowel. There is a productive process for deriving
disyllabic bases from monosyllabics by the addition of a stem vowel. This
stem vowel is -a- (post-vocalically -oa-) as in jana- ‘inform’; although, for
many speakers, the stem vowel may be -o- if the root vowel (i.e. of the
monosyllabic base) is [+ high]; e.g. jiro-, for some speakers jira-
‘rest’. Derived disyllabics usually serve as the formal causatives of their
monosyllabic counterparts. Compare: jan- ‘know’, jana- ‘inform’; ot.h-
‘rise’, ot.ha- ‘raise’; dækh- ‘see’, dækha- ‘show’.

Table 23.4 Bengali Verbal Inflection

--------------------------------------------------------------------

		1st person 2nd person  2nd p	  3rd p	    Honorific
			   despective  ordinary  ordinary  (2nd, 3rd
							   persons) 

--------------------------------------------------------------------

Present imperative  –        NULL       -o         -uk        -un
Unmarked indicative   -i     -is        -o         -e         -en
and -(c)ch- stems
-b- stems             -o      -i        -e         -e         -en
-t- and -l- stems     -am     -i        -e         -o         -en

--------------------------------------------------------------------

the stem-deriving marker -(c)ch signals continuative aspect and is used,
independent of any other derivational marker, to derive the present
continuous verbal form. The element (c) of the marker -(c)ch- deletes
post-consonantally; compare khacche ‘is eating’ (from kha-) with anche ‘is
bringing’ (from an-). In forming the verbal stem with -(c)ch- the high
alternate base is selected, unless the base is disyllabic or is a
monosyllabic base having the root vowel /a/. Compare the last examples with
ut.hche ‘is rising’ (from ot.h-), ot.hacche ‘is raising’ (from ot.ha-).

In a formal treatment of Bengali morphophonemics, the basic or underlying
form of the stem marker could be given as -i(c)ch-; in this event, one would
posit a rule to delete the element /i/ after Vowel Height Assimilation
applies, except in a very limited class of verbs including ga- ‘sing’, sO-
‘bear’ and ca- ‘want’. [gAiche, saiche, cAiche]
In forming the present continuous forms of these verbs, the element /i/
surfaces, although the element (c) of the stem marker tends to be
deleted. The resulting shapes are, respectively: gaiche ‘is singing’ (gacche
is at best non-standard); soiche (*socche) ‘is bearing’; caiche ‘is wanting’
(cacche does, however, occur as a variant).

The stem-deriving marker -b- (see Table 23.4) signals irrealis aspect and is
used to derive future verbal forms, both indicative and imperative (except
for the imperative of the second person ordinary, which will be treated after
the next paragraph). In Bengali, the future imperative, as well as the
present imperative, may occur in affirmative commands; however, the future
imperative, never the present imperative, occurs in negative commands.
ut.hbo ‘I/we will rise’ (from ot.h-), but ot.habo ‘I/we will raise’ (from
ot.ha-); janbo ‘I/we will know’ (from jan-), debo ‘I/we will give’ (from
de-). Compare, however, dibi ‘you (despective) will give’, where Vowel Height
Assimilation raises the root vowel.

It is possible, again, to posit an underlying /i/ in the irrealis stem
marker’s underlying shape (i.e. -ib-), with deletion of the element /i/
applying except for the small class of verbs noted earlier; thus gaibo
(*gabo) ‘I/we will sing’, soibo (*sobo) ‘I/we will bear’, caibo (*cabo) ‘I/we
will want’.

Vowel height assimilation

 
Second, monosyllabic bases with non-high root vowels have two alternate forms,
respectively called low and high. Examples are:

Alternate base forms: 
		Low    High
	-----------------------
	‘know’	jan-	jen-
	‘see’	dækh-	dekh-
	‘sit’	bOs-	bos-
	‘buy’	ken-	kin-
	‘rise’	ot.h-	ut.h-

When the root vowel is /a/, /e/ is substituted to derive the high alternate
base; for bases with front or back non-high root vowels, the high alternate
base is formed by assimilating the original root vowel to the next higher
vowel in the vowel inventory (Table 23.2). The latter behaviour
suggests an extended application of the Vowel Height Assimilation process
discussed in the preceding section. It is, in fact, feasible to state the
rules of verb stem derivation so that the low/high alternation is
phonologically motivated; i.e. by positing a high vowel (specifically, /i/)
in the underlying shapes of the stem-deriving markers. In some verbal forms
there is concrete evidence for the /i/ element, as will be observed
below. Also, Vowel Height Assimilation must be invoked in any case to account
for the fact that, in the derivation of verbal forms which have zero marking
of the stem (that is, the present imperative and unmarked (present)
indicative), the high alternate base occurs before any inflection containing
a high vowel.  Thus 

	dækh- ‘see’, dækho ‘you (ordinary) see’, 

but 
	dekhi ‘I see’, dekhis ‘you (despective) see’, 
	dekhun (honorific) ‘see!’, etc. 

That there is no high–low alternation in these inflections for disyllabic
bases is consistent with the fact that Vowel Height Assimilation only applies
when a high syllabic occurs in the immediately succeeding syllable. Thus
ot.ha- ‘raise (cause to rise)’, ot.hae ‘he/she raises’, (*ut.hai) ‘I/we
raise’, etc.

Word Order

In some literature on word order types, Bengali has been characterised as a
rigidly verb-final language, wherein nominal modifiers precede their heads;
verbal modifiers follow verbal bases; the verbal complex is placed
sentence-finally; and the subject noun phrase occupies the initial position
in a sentence.  In these respects Bengali is said to contrast with earlier
Indo- Aryan, in which the relative ordering of sentential constituents was
freer, notwithstanding a statistical tendency for verbs to stand at the ends
of their clauses.

It is true that the ordering of sentential elements is more rigid in Modern
Bengali than in Classical Sanskrit. However, the view that Bengali represents
a ‘rigid’ verb-final language does not adequately describe its differences
from earlier Indo-Aryan word order patterning.

Word order in NPs and Sentences

Word order within the Modern Bengali noun phrase is stricter than Sanskrit, 
[in part due to the relative impoverishment of the Modern Bengali case system.]

An adjective or genitive expression is always placed before the noun it
modifies. By contrast, in earlier Indo-Aryan, adjectives showed inflectional
concord with their modified nouns and consequently were freer in their
positioning; more or less the same applied to the positioning of genitive
expressions with respect to nominal heads. 

In Modern Bengali, the mutual ordering of noun phrases within the sentence is
strict as well, much more so than in earlier Indo-Aryan. The subject noun
phrase generally comes first in a Modern Bengali sentence, followed by an
indirect object if one occurs; next comes the direct object if one occurs;
after which an oblique object noun phrase may be positioned.

Bengali case markers are, nonetheless, supplemented by a number of
postpositions, each of which may govern nouns declined in one of two cases,
the objective or genitive.

Verbal complex: Causative markers

    At the Old Indo-Aryan stage exemplified by Classical Sanskrit, 
markers representing certain verbal qualifiers (causal, desiderative,
potential and conditional) could be affixed to verbal bases, as stem-forming
markers and/or as inflectional endings. 

In Modern Bengali, the only verbal qualifier which is regularly affixed to
verbal bases is the causal. (See the discussion of derived disyllabic verbal
bases in Section 3 above.)  The following pair of Bengali sentences
illustrates the formal relationship between noncausative and causative
constructions: 

	chelet.i cit.hit.a por.lo 
	(the-boy the-letter read) 
	‘the boy read the letter’; 

	ma chelet.i-ke diye cit.hit.a pOr.alen 
	(mother to-the-boy by the-letter caused-to-read) 
	‘the mother had the boy read the letter’. 

It will be noted that in the second example the non-causal agent is marked
with the postposition diye ‘by’ placed after its governed noun, which appears
in the objective case. Usually, when the verbal base from which the causative
is formed is transitive, the non-causal agent is marked in just this way. The
objective case alone is used to mark the non-causal agent when the causative
is derived either from an intransitive base, or from any of several
semantically ‘affective’ verbs – transitive verbs expressing actions whose
principal effect accrues to their agents and not their undergoers. Examples
are: ‘eat’, ‘smell’, ‘hear’, ‘see’, ‘read’ (in the sense of ‘study’),
‘understand’ and several others.

Negation

    In Old Indo-Aryan (Classical Sanskrit), the marker of sentential
    negation tended to be placed just before the sentential verb. 

The particle of sentential negation in Bengali is na. In independent clauses
it generally follows the sentential verb; in subjoined clauses (both finite
and non-finite), it precedes. Thus: boslam na (I-sat not) ‘I did not sit’;
jodi tumi na bO so (if you not sit) ‘if you don’t sit’; tumi na bosle (you
not if-sit) ‘if you don’t sit’. Bengali has, it should be mentioned, two
negative verbs. Each of them is a counterpart to one of the verbs ‘to be’;
and in this connection it needs to be stated that Bengali has three verbs ‘to
be’.  These are respectively the predicative hO - ‘become’; the existential
verb ‘exist’, having independent/subjoined clause allomorphs ach-/thak-; and
the equational verb or copula, which is normally NUL but in emphatic contexts
is represented by hO - placed between two arguments (compare, for example,
non-emphatic ini jodu (this-person NUL Jodu) ‘this is Jodu’ versus emphatic ini
hocchen jodu (this-person is Jodu) ‘this (one) is Jodu’).  While the
predicative verb ‘to be’ has no special negative counterpart (it is negated
like any other Bengali verb), the other two verbs ‘to be’ each have a
negative counterpart.  Moreover, for each of these negative verbs, there are
separate allomorphs which occur in independent and subjoined clauses. The
respective independent/subjoined shapes of the negative verbs are existential
nei/na thak- (note that the verb nei is invariant) and equational nO -/na
hO-. It bears mentioning, incidentally, that negative verbs are neither 
characteristic of modern nor of earlier Indo-Aryan. They are, if anything,
reminiscent of negative copulas and other negative verbs in languages of the
Dravidian (South Indian) family, such as Modern Tamil.

Interrogative marker ki

    In Sanskrit, the sentential interrogative particle was often placed at a
    distance from the verbal complex.

The Modern Bengali sentential interrogative particle ki is inherited from an
earlier Indo-Aryan particle of similar function. The sentential interrogative
ki may appear in almost any position in a Bengali sentence other than
absolute initial; however, sentences vary in their presuppositional nuances
according to the placement of this particle, which seems to give the most
neutral reading when placed in the second position (i.e. after the first
sentential constituent). To illustrate, compare: tumi ki ekhane chatro?  (you
interrogative here student) ‘are you a student here?’; tumi ekhane ki chatro?
(you here interrogative student) ‘is it here that you are a student?’; tumi
ekhane chatro (na) ki?  (you here student (negative) interrogative) ‘oh, is
it that you are a student here?’.

Reduplicatives and Classifiers

Bengali has two lexical features of a type foreign to Indo-Aryan. These
features are, however, not atypical of languages of the general South Asian
language area (and are even more typical of South-East Asian languages). One
of these is a class of reduplicative expressives, words such as: kickic
(suggesting grittiness), mit.mit. (suggesting flickering), t.OlmOl
(suggesting an overflowing or fluid state).  There are dozens of such lexemes
in current standard Bengali.  

The other un-Aryan lexical class consists of around a dozen classifier words,
principally numeral classifiers. Examples are: du jon chatro (two
human-classifier student) ‘two students’; tin khana boi (three
flat-thing-classifier book) ‘three books’.

OTHERS: TA, Ti, Tu, To, Tuku, -ek [approximative, e.g. goTA chArek], 

It is probable that the features discussed above were absorbed from other
languages into Bengali after the thirteenth century, as the language came to
be increasingly used east of the traditional sociocultural centre of
Bengal. That centre, located along the former main course of the Ganges (the
present-day Bhagirathi–Hooghley River) in western Bengal, still sets the
standard for spoken and written expression in the language.  Thus standard
Bengali is defined even today as the dialect spoken in Calcutta and its
environs. It is a reasonable hypothesis nevertheless, as suggested above in
Section 1, that descendants of non-Bengali tribals of a few centuries past
now comprise the bulk of Bengali speakers. In other words, the vast majority
of the Bengali linguistic community today represents present or former
inhabitants of the previously uncultivated and culturally unassimilated
tracts of eastern Bengal. Over the past several centuries, these newcomers to
the Bengali-speaking community are the ones responsible for the language’s
having acquired a definite affiliation within the South Asian linguistic
area, above and beyond the predetermined and less interesting fact of its
genetic affiliation in Indo-Aryan.
 

Bibliography

Chatterji (1926) is the classic, and indispensable, treatment of historical
phonology and morphology in Bengali and the other Indo-Aryan languages. A
good bibliographical source is Čižikova and Ferguson (1969). For the relation
between literary and colloquial Bengali, see Dimock (1960).  The absence of a
comprehensive reference grammar of Bengali in English is noticeable. Ray et
al. (1966) is one of the better concise reference grammars. Chatterji (1939)
is a comprehensive grammar in Bengali, while Chatterji (1972) is a concise
but thorough treatment of Bengali grammar following the traditional scheme of
Indian grammars. Two pedagogical works are also useful: Dimock et al. (1965),
a first-year textbook containing very lucid descriptions of the basic
structural categories of the language, and Bender and Riccardi (1978), an
advanced Bengali textbook containing much useful information on Bengali
literature and on the modern literary language. For individual topics, the
following can be recommended: on phonetics-phonology, Chatterji (1921) and
Ferguson and Chowdhury (1960); on the morphology of the verb, Dimock (1957),
Ferguson (1945) and Sarkar (1976); on syntax, Klaiman (1981), which discusses
the syntax and semantics of the indirect subject construction, passives and
the conjunctive participle construction in modern and earlier stages of
Bengali.

References

Bender, E. and T.Riccardi, Jr. 1978. An Advanced Course in Bengali (South
    Asia Regional Studies, University of Pennsylvania, Philadephia)
Chatterji, S.K. 1921. ‘Bengali Phonetics’, in Bulletin of the School of
    Oriental and African Studies, vol. 2, pp. 1–25
Chatterji, S.K. 1926. The Origin and Development of the Bengali Language, 3
    vols. (Allen and Unwin, London)
Chatterji, S.K. 1939. (Calcutta University, Calcutta)
Chatterji, S.K. 1972. Sarala , revised ed. (Bāk-sāhitya, Calcutta)
Čižikova, K.L. and C.A.Ferguson. 1969. ‘Bibliographical Review of Bengali
    Studies’, in T.Sebeok (ed.), Current Trends in Linguistics, vol. 5:
    Linguistics in South Asia (Mouton, The Hague), pp. 85–98
Dimock, E.C., Jr. 1957. ‘Notes on Stem-vowel Alternation in the Bengali
    Verb’, Indian Linguistics, vol. 17, pp. 173–7
Chatterji, S.K. 1960. ‘Literary and Colloquial Bengali in Modern Bengali
    Prose’, International Journal of American Linguistics, vol. 26, no. 3,
    pp. 43–63
Chatterji, S.K. et al. 1965. Introduction to Bengali, part 1 (East-West
    Center, Honolulu; reprinted South Asia Books, Columbia, Mo., 1976)
Ferguson, C.A. 1945. ‘A Chart of the Bengali Verb’, Journal of the American
    Oriental Society, vol. 65, pp. 54–5 

Contents

Preface vii

  Introduction 			BERNARD COMRIE 	1
1 Indo-Aryan Languages 		GEORGE CARDONA 	14
2 Sanskrit 			GEORGE CARDONA 	20
3 Hindi-Urdu 			YAMUNA KACHRU 	37
4 Bengali 			M.H.KLAIMAN 	54
5 Iranian Languages 		J.R.PAYNE 	73
6 Persian 			GERNOT L.WINDFUHR 	81
7 Pashto 			D.N.MACKENZIE 	101
8 Afroasiatic Languages 	ROBERT HETZRON 	116
9 Semitic Languages 		ROBERT HETZRON 	121
10 Arabic 			ALAN S.KAYE 	129
11 Hebrew 			ROBERT HETZRON 	147
12 Hausa and the Chadic Languages 	PAUL NEWMAN 	162
13 Tamil and the Dravidian Languages 	SANFORD B.STEEVER 	178
14 Niger-Kordofanian Languages 	DOUGLAS PULLEYBLANK 	195
15 Yoruba 			DOUGLAS PULLEYBLANK 	203
16 Swahili and the Bantu Languages 	BENJI WALD 	219