A Quarrel in The Language Family
A Quarrel in The Language Family
A Quarrel in The Language Family
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=cup.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].
Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to Modern
Asian Studies.
http://www.jstor.org
Modern Asian Studies, 27, 4 (I993), pp. 76I-804. Printed in Great Britain.
SiidasienInstitut,UniversitdtHeidelberg
... there is a well-known form of speech in the south of the Panjab called
'Jangali', from it being spoken in the 'Jungle', or unirrigated country border-
ing on Bikaner. But 'Jangali', also means 'boorish' and local inquiries failed
to find a single person who admitted that he spoke that language. 'O yes, we
knowJangali very well,-you will find it a little further on,-not here.' You
go a little further on and get the same reply, and pursue your will-o'-the-wisp
till he lands you in the Rajputana desert, where there is no one to speak any
language at all.
I came upon this passage in Grierson's Linguistic Surveyof India (I927:
I: i: I9) after having spent a year in the provincial town ofJanakpur,
documenting the Maithili language of northern Bihar and
southeastern Nepal. Many local people encouraged and assisted me in
my research, but all told me in good faith that I had come to the
wrong place. I should have gone twenty miles to the southeast, where
the 'authentic' language is spoken. It seems that I had not been alone
in having been urged by informants and well-wishers to go somewhere
else: either in pursuit of languages that do not exist or being redirected
down the road to where the language is really spoken. Unfortunately
visa problems prevented me from taking up the advice of friends, yet a
cursory reading of the literature on regional and social dialectology
would have been enough to turn anyone into a skeptic about what one
might have been gained from such a journey. Subjective dialect
boundaries do not often register on maps of isoglosses, and the objec-
tive methods of linguists usually reveal local perceptions of speech
behaviour to be based on stereotypes.
I am gratefulto SimonWeightman,David Parkin,BurtonSteinand MahadevApte
fortheircommentson an earlierdraftof this essayand to His Majesty'sGovernment
for permission to carry out research in the Nepalese Tarai in I984-1985. My deepest
debt, of course, is to Brikhesh Chandra and Sivendra Lal Karn who taught me their
mother tongue and to Dhireswar Jha whose affection for the language and literature
of Mithila became contagious.
oo26-749X/93/$5.oo + .oo ? '993 Cambridge University Press
761
762 RICHARD BURGHART
India for its narratives and songs, and on the western frontier of
Maithili, whose advocates claim for it the honour of being the sweetest
language in the world. The people of Sarlahi-so a Bhojpuri speaker
from Champaran district (India) informed me-speak Bhojpuri
words with Maithili 'tuning'. They use the lexicon of one language and
the phonemics of another. That's khicari.
The fact that Maithils detect a rice-and-pulse pattern along their
linguistic frontier implies that within and beyond the frontier there is
either rice or pulse, but not both. That is to say, within Mithila there
is a language which is trans-local enough for it to be considered a
language. But what is this language and what is its name? Native
speakers distinguish between two kinds of Maithili: chaste Maithili
(suddh mdithili) and rustic Maithili (dehati mdithili). The distinction is
somewhat evaluative, as well as possessory, for chaste Maithili
speakers sometimes claim more strongly that they alone speak
Maithili; what the others speak is not rustic Maithili but simply rustic
language (dehati bhasa). Here speech participates in personal refine-
ment and civilized values. Maithili is the preserve of those who
cultivate their minds, not their fields. In the former category one finds
Maithils from the so-called 'big' castes: the Maithil Brahmans who
are guardians of Sanskritic knowledge, the Bhumihar Brahmans who
are the traditional landlords of Mithila and the Kayastha scribes who
kept the accounts and revenue records for the Bhumihars. In the
latter category are the so-called 'little' castes who comprise the touch-
able servant and artisan castes, such as the Cowherders (the most
populous caste of Mithila), Blacksmiths, Watercarriers and Barbers;
and the untouchable servant and artisan castes, such as the Oilpres-
sers, Palanquin bearers, Cobblers, Washermen and Sweepers.
Chaste and rustic Maithili are mutually intelligible speech forms
that differ lexically and phonetically. With regard to the lexicon
chaste Maithili is enriched with recent Sanskrit loanwords (tatsama)
while sharing with rustic Maithili much the same native tadbhava
vocabulary that derives historically from Sanskrit but which entered
Maithili through Prakrit, Apabhramsa and the various proto-forms of
modern Maithili. For example, in chaste Maithili one might use either
the tatsama word 'pracin' for ancient or the tadbhava 'purdn'. Rustic
speakers, however, would only use 'purdn'.Research on caste dialects
in Dravidian south India has shown greater lexical borrowing among
the high castes (Bright I960), but considerable research would have
to be completed before the same argument could be advanced for
Maithili. It may be that rustic speakers merely borrow from a dif-
REPRESENTATIONS OF SPEECH IN MITHILA 767
ferent source. For example, speakers of chaste Maithili use the tad-
bhava 'ber' for 'time', but rustic speakers tend to use 'bdji' from the
Persian. Still other chaste and rustic words may have come from a
common tadbhava source, evolving differently in their pronunciation:
e.g. the English 'well' is 'inar' in chaste Maithili but 'in.da'in rustic. In
a somewhat similar vein English loanwords in chaste Maithili are
rendered phonetically into the Maithili sound system. Rustic speakers
also take up English words, but they are more likely to adapt the
source word to local pronunciation. A case in point is the English
word 'cement', which is 'siment' in chaste Maithili but 'simata' in
rustic. Finally, literary words, such as the verb 'to be' derived from
the root 'thik', form part of spoken chaste Maithili, but are not charac-
teristic of rustic speech, which is perforce colloquial.
Bearing in mind that agency in the Hindu universe is expressed by
manual passivity and self-restraint, 'big caste' speakers of chaste
Maithili often aim for a curtailed speech (alp bhasa), leaving rustic
speakers to express through their vociferousness the necessity of their
domination. Alternatively chaste speakers sweeten their speech or
infuse it with sentiment (rasaeb) in order to animate conversation in a
pleasing manner. These various stylistic aims produce a somewhat
higher pitched speech, articulated from the front of the mouth which
is light in manner and low in amplitude. Chaste Maithili speakers also
exhibit a tendency to end words in a consonant, the effect of which is
to give a stop-and-go pattern to the rhythm of their speech. By con-
trast, rustic speakers produce a more lilting sound by ending words in
a long intra-syllabic vowel: e.g. the word 'purdn',mentioned above, is
often pronounced 'purnd'or lengthened to 'purnkd'by rustic speakers.
This vocalic contrast is accentuated by metathesis among the
speakers of chaste Maithili. For example, in uttering the word water
'pani', chaste speakers pronounce the short intra-syllabic 'i' before the
consonant as 'pain'; rustic speakers lengthen the intra-syllabic vowel
and pronounce it after the consonant as 'pdna'. Bearing in mind the
recourse to Sanskrit loanwords, chaste Maithili speakers articulate
the cacuminal 'n' which is found only in Sanskrit loanwords. Rustic
speakers, however, render it as a dental. Similarly the unvoiced alveo-
palatal fricative 's', which in early Maithili may not have been dis-
tinguished from the dental 's', is often articulated by chaste speakers
who are conscious of its pronunciation in Sanskrit loanwords (S. Jha
1958: 187-8).
With regard to the grammar, adjectival and verb terminations often
differ in chaste and rustic Maithili, but the variation does not lie in
768 RICHARD BURGHART
country. As for the rest, they have a 'working language' that sets up
the possibility of mutual intelligibility as 'it' goes, and they go, from
here to there across centres of dominance.
which he refers to the view of Bihar officials that Maithili was just a
gaonwdri boli, that is, 'rustic speech'.
If Maithili were merely the rustic speech of villagers then Grierson
thought it would have to be the dialect of some language, be it Hindi
or Bengali, that served as its standard. Alternatively Maithili could be
a language, distinct from Hindi or Bengali, with its own subsidiary
dialects. For Grierson the synchronic distinction between language
and dialect hinged on two criteria: first, a language serves as a
standard for which dialects appear as variant forms; and second, the
mutual intelligibility of dialects and the mutual unintelligibility of
languages. Despite their conciseness, the application of these criteria
to particular cases is often problematic, and especially so in the case of
Indo-Aryan vernaculars.
With regard to the first criterion, considerable ambiguity clouds the
term 'standard'. Does the word designate the normative form of
speech that is prescribed in schoolbooks, but highly restricted in use?
Or does it designate the normal form of the language as it is dis-
tributed throughout a speech community, the existence of which
social and geographical dialectologists call into question? Grierson
was not unaware of this ambiguity as he set out to establish the
standard of, or possibly for, Maithili. His method was informed by his
classical and orientalist training as much as by his commitment to
language studies as philology (rather than as linguistics, a term which
had not yet gained its modern currency). He represented languages as
paradigms of person and number or of cases according to which verbs
are conjugated and nouns declined. Such a grammatical represen-
tation informed both the European study of Latin and Greek as well
as the prescriptive English grammars used in schools. The standard
was found by having native informants translate Maithili from Hindi
and Sanskrit equivalents in terms of a Latin-derived 'word and
paradigm' representation of language (Grierson 1881: I).
I printed paradigms of all the forms in Hindi and Sanskrit Grammar and
circulated them as widely as possible amongst the Pandits, Village School
Masters and educated Native Gentlemen of Northern Mithila, with direc-
tions to give the exact translation of each of these forms in their own native
language.
I was enabled in this way, to collect some fifty most useful books of forms,
supplied by representatives of all classes of society, from the village guru,who
knew little more than the herd-boys he taught, to the most learned Pandits of
Mithila. I am glad to say that the utmost interest was taken in my design, for
the people are proud of their language and were pleased at the idea of its
being made a polite one, by obtaining the honour of print.
REPRESENTATIONS OF SPEECH IN MITHILA 775
Despite the disparaging reference to the village guru, all of Grierson's
informants were literate, and by implication members of the so-called
big castes. Moreover, they resided in the Madhubani subdivision of
Darbhanga district and the adjoining portion of Bhagalpur-that is to
say the panc kosi region-where Maithili is spoken by Brahmans in its
greatest purity and where they have 'retarded its corruption' (i909:
xi). In short, the Maithil pandit's 'chaste Maithili' became the
European philologist's 'standard Maithili'. By implication, the
Maithili spoken outside the paic kosi region became variant forms,
which were classified as 'dialects'.
In one respect these two language values-the 'chaste' and the
'standard'-are commensurable. Ethnographic accounts of
'Sanskritization' describe the persuasive imitative pull that Brahman-
ical customs, privileges and manners have exerted upon lower castes;
hence the pandit's chaste Maithili is as authoritative as the European
philologist's standard Maithili. Yet the hierarchical ordering of caste
society deprives chaste Maithili of the prescriptive import that the
'standard' language form has in societies with a commitment to
egalitarian values. Chaste Maithili does not impose itself as the nor-
mative form of rustic Maithili (or at least it did not until universal
primary education came to Bihar). Moreover, a certain type of speech
comes 'natural' to a certain type of speaker such that the adoption of
the speech styles of persons from higher or lower stations in life is
unnatural. In medieval drama different ranks of characters were
scripted in different languages: Brahmans and the king in Sanskrit;
women and persons of middle rank in Sauraseni; and comic charac-
ters in Magadhi (Bloch 1965: i9). If one searches for prescriptive
import in spoken or written Maithili then the comparison would be
between the impact of Sanskrit grammar on chaste Maithili and of
Latin grammar on nineteenth-century prescriptive English (the split-
ting of infinitives or the mixing of cases in 'It is me' are 'bad usage'
because they violate the rules of Latin grammar, not the conventions
of spoken English).
As for the second criterion of a language, namely its mutual unintel-
ligibility with other languages, linguistic research in the 1950S and
I96os has demonstrated how problematic such a criterion is. Intelli-
gibility is often a matter of degrees and contexts, so that one is still left
with the problem of determining how unintelligible communication
must be before two dialects may be thought of as separate languages
(Haugen I967). Moreover, perceptions of intelligibility-and there-
fore readiness to make oneself intelligible-are matters of local ethnic
776 RICHARD BURGHART
attitude (Wollf 1959). It is not clear how Grierson tested for mutual
unintelligibility, except to note that comparison was made with
reference to the neighbouring languages of Hindi and Bengali.
Because there was no external standard for Maithili and in view of
its mutual unintelligibility with Bengali or Hindi, Grierson concluded
that Maithili was an independent language (I88I: 2).
For Maithili is a language and not a dialect. It is the custom to look upon it
as an uncouth dialect of untaught villagers, but it is in reality the native
language of more than seven and a quarter millions of people [1880 figures],
of whom, as will be borne out by every official having experience of North
Bihar, at least five millions can neither speak nor understand either Hindi or
Urdu without the greatest difficulty. It differs from both Hindi and Bangali,
both in Vocabulary and Grammar, and is as much a distinct language from
either of them as Marathi or Uriya. It is a country with its own traditions, it
own poets and its own pride in everything belonging to itself.
22-4).
But although we entered on our work with these ideas [the Sanskrit origin of
the nine principal Indian vernacular languages, all the rest being dialects of
Hindi] we were ultimately constrained to relinquish them. First, one
language was found to differ widely from Hindee in point of termination,
then another, and in so great a degree, that the idea of their being dialects of
Hindee seemed scarcely tenable. Yet while they were found to possess
terminations for the nouns and verbs distinct from the Hindee, they were
found as complete as the Hindee itself; and we at length perceived that we
might, with as much propriety term them dialects of the Mahratta or the
Bengalee language, as of the Hindee. In fact, we have ascertained that there
are more than twenty languages, composed, it is true, of nearly the same
words and all equally related to the common parent, the Sungskrit, but each
possessing a distinct set of terminations, and therefore, having equal claims
to the title of distinct cognate languages. Among these we number the
Juypore, the Bruj, the Oodupore, the Bikaner, the Mooltanee, the Marawar,
the Maguda (or South Bahar), the Sindh, the Mythil [Maithili], ... etc.,
languages, the very names of which have scarcely reached Europe, but which
have been recognized as distinct languages by the natives of India since time
immemorial.
posed the three dialects to have derived from the same language. A
name did not exist for this congery of dialects, but Grierson (1883: 2)
made good the absence by calling them dialects of the Bihari
language. The Bihari language was spoken throughout much of Bihar
province and hence served-like Marathi, Panjabi, Bengali and other
languages named after places--to designate the language of the
region. In i909, when the revised edition of The Maithili Language of
North Bihar (I88i) was published, the new title reflected Maithili's
diminished status in the family of Indo-Aryan languages: The Maithili
Dialectof theBihariLanguageas spokenin NorthBihar. In brief, Maithili,
and its neighbouring 'dialects', stemmed from the common Bihari
language source which, in turn, possessed common features with
Bengali, Assamese and Oriya. Bihari became the 'parent' of Maithili
and the 'sister' of Bengali, Assamese and Oriya in the family tree of
eastern Indo-Aryan languages. All four languages, in turn, sprang
from Magadha Apabhramsa.
The philological discovery of the Bihari language defies in some
measure common sense. Grierson offered the comparison of Marathi,
Panjabi and Bengali as territorially defined languages, yet the com-
parison is not entirely apt. True that Bengali is spoken in dialect, yet
the speakers of these various dialects recognize in some sense their
common membership in a Bengali speech community. No such
awareness exists in Bihar of there being a Bihari language, or of there
ever having been a Bihari language. It seems preposterous that
Maithili, Bhojpuri and Magahi speakers could all speak variants of a
language, of the existence of which they remained unaware.
To be charitable, though, the invention of the Bihari language has
some plausibility. Grierson was a philologist, applying the compara-
tive method to formal features of speech (sound, grammar, words) so
that dialects and languages might be found whose relatedness stem-
med from their supposed common origin. At the time of Grierson's
initial research in the I88os comparative philology had been put on a
'scientific basis' by theJunggrammatiker, based in Leipzig, who form-
ulated phonological laws of language change such that lapsed parent
languages could be reconstructed from an analysis of their descendent
forms. Such reconstructed proto-forms, called hypotheticals, were
pure philological inventions, and were indicated as such by an
asterisk. The fascination for Grierson, as much as for other philo-
logists, of Indian languages was their proto-relation with European
languages and the availability of documents and inscriptions extend-
ing over several millennia which enabled the reconstruction of the
REPRESENTATIONS OF SPEECH IN MITHILA 779
Indo-Aryan language family with relatively rare recourse to starred
items. Nonetheless hypothetical forms were part of the stock-in-trade
of the philologist. Grierson's invention of the Bihari language was not
a departure from philological convention.
Given the, by now, commonplace criticisms of the synchronic
language/dialect distinction and the family-tree model of language
change (see Lyons i981), there is no need to take Grierson to task for
being a 'man of his time'. Yet it is curious, his shift in frame from
subdialect to dialect to language. The dialect/subdialect relation was
synchronic and hinged on the distinction between standard and
variant forms; the language/dialect relation was diachronic and
genetic. He implicitly treated Maithili as a language in its synchronic
context, for it was the standard form of its variant subdialects; and he
explicitly treated Maithili as a dialect in its diachronic relations with
so-called Bihari. By constructing a genealogy in which a plethora of
regional dialects stemmed from four languages (Bengali, Oriya,
Assamese and Bihari) that emerged from one Ursprache (Magadha
Apabhramsa), Grierson solved his problem of synchronic hetero-
geneity by imagining a prehistory of homogeneity.
Yet one might expect the socio-linguistic conditions that made the
synchronic language/dialect distinction unworkable in the late
nineteenth century to have also made it unworkable in the prior past.
Indeed, a purview of the philological literature reveals the absence of
any consensus whatsoever on the genealogy of Maithili (reviewed in
Yadav 1981). Kellogg (1876) and Hoernle ( 880) considered Maithili
to be part of eastern Hindi. Grierson separated the three 'Bihari
dialects' from Eastern Hindi and put them with the eastern Prakrit of
Bengali, Oriya and Assamese. Chatterji (1926) granted
'independence' to the three dialects, making Bengali, Assamese and
Oriya the eastern Magadhi; Maithili and Magahi the central
Magadhi and Bhojpuri, Nagpuri and Sedani the western Magadhi.
Misra (1976: 2 ) similarly granted Maithili 'independence', putting it
in with Bengali, Assamese and Oriya, relegating Bhojpuri into the
eastern Hindi family and sending Magahi into oblivion. The common
feature of these fictions is that language families tend to achieve some
measure of identity when they are spoken at the former imperial
centres of the Ganges basin. The so-called Bihari languages, being on
the periphery of Bengal and the Doab, are shifted back and forth
between the languages of two other places as counters in the
reconstruction of the past.
78o RICHARD BURGHART
knowledge. The passage below, drawn from the 1872 Census (p. 152),
reads like three pictures from the exhibition that never took place:
Bihari, therefore, was deemed to be 'Hindi in the wide sense' (p. 314),
a phrase that would reappear in later censuses. The number of Bihari
speakers-comprising the dialects of Maithili, Bhojpuri and
Magahi-was worked out according to the formula Grierson had used
to recover the speech community from its exclusion in the 1891 Cen-
sus. The Maithili formula was: all the people from Darbhanga and
Bhagalpur districts, 6/7ths of the population of Muzaffarpur, 1/2 that
of Monghyr, 2/3rds the population of Purnea and 4/5ths of the Hindi
speakers of the Santhal Pergunnas. According to this calculation in
19oI there were 10,387,897 Maithili speakers, up from 9,207,131 as
from Grierson's calculations from the I891 Census (p. 320).
The administrative convenience of recording Bihari as Hindi suited
the government's emerging policy on language. By the turn of the
century the Government of India had begun considering ways by
which the regional government might be made more effective. It was
clear that the Bengal Presidency was too large, and yet the initial
solution to the problem--the crude division of the Presidency in 1905
into an eastern and western province-angered Bengali speakers who
saw their country being cut in two. Resistance was strong enough to
make the Government rethink the basis of effective administration. A
common language was seen to be a precondition of communication,
and socio-cultural homogeneity was thought to promote the growth of
local loyalties. To this effect plans were made to constitute four
culturally homogenous states: Bengal, Assam, Bihar and Orissa, each
of which possessed its own dominant language. Hindi (Hindustani,
REPRESENTATIONS OF SPEECH IN MITHILA 785
Urdu) was the language of Bihar-both of the people and the govern-
ment-and this difference alone legitimated its separation from
Bengal. At the Delhi Darbar on 12 December 1911 it was announced
that in the following year the western and eastern provinces of Bengal
would be reunited into one administrative unit, but that Assam would
be detached from the Presidency as a Chief Commissioner's Province
and that a separate province of Bihar and Orissa would be created.
Orissa was to be administered from Patna until judged ready for its
own government (Government of Bihar, I954: I2).
It is to the highest degree desirable to give the Hindi-speaking people, now
included within the Province of Bengal, a separate administration. These
people, have hitherto been unequally yoked with the Bengalis, and have
never, therefore, had a fair opportunity for development. The cry of Bihar for
the Biharis has frequently been raised in connection with the conferment of
appointments, an excessive strong belief has grown up among Biharis that
Bihar will never develop until it is dissociated from Bengal.
Under the circumstances it would have been difficult for the O'Malley
Census of 191I to maintain that Hindi was not spoken in Bihar and at
any rate 'it was realized that it would be hopeless to expect the people
themselves to return their languages with any philological exactitude'
(p. 382). A person who speaks Bihari does not call his language
Bihari, but Hindi.
By 1921 the classification scheme of the Linguistic Surveyof India was
adopted for Bihar and Orissa. All Indo-Aryan vernaculars were rel-
egated either to the Outer or the Inner Sub Branch, with the excep-
tion, again, of Bihari which is not brought into the schedule. The
Census Commissioner could not, however, pretend--as O'Malley had
done-that the people of Bihar do 'not recognize such names as
Magahi, Bhojpuri and Maithili as designations for different dialects of
that language' ( 91I: 382). During the first few decades of the twen-
tieth century numerous organizations for the promotion of Maithil
culture emerged in Calcutta, Benaras and the towns of Mithila, of
which the chief of them was the Darbhanga-based Maithila
Mahasabha, formed in I9Io on the instigation of the Benaras-based
Maithili journal Mithila Moda (Misra I976: 223-8). Thus the reason
why Maithili could not be returned in the 1921 Census now lay with
the unsophistication of the enumerators as philologists, not the people
(p. 2 II):
Considerable interest was displayed at this census in the Maithili dialect,
and letters were received from the Maithili Mahasabha of Darbhanga and
the Shree Dharmaamrita Vaishini Sabha of Bhagalpur suggesting that
786 RICHARD BURGHART
In North Bihar 98.7 per cent of the people were said to speak Hindi
where the 'linguistic distribution is as flat and devoid of variety as is
the landscape-facts which are connected as effect to cause, for
languages naturally spread more rapidly where there is no physical
hindrance to the free movement of people who speak them (p. 2 o).' It
is clear from the above statement, which reveals more about the
doctrine of free trade than about language distribution, that a philo-
logist did not write the 1921 Census report.
The I921 Census is of interest, for it confirms that the census had
now become a state institution. The I901 Census in Bengal was
preoccupied with problems of validity. Regardless of Risley's racialist
concept of caste and Grierson's invention of Bihari, there was an
attempt by Gait to base the census categories on the perceived socio-
cultural reality of India. By 1921 the value had shifted to reliability.
There may be a language or dialect called Maithili but there is no
sense in recording it because if the census categories are altered one
can no longer compare the returns of the present census with previous
ones to monitor and measure population movements. It is impossible
to say whether Khotta is Hindi or Bengali but since it was treated as
Hindi in 191I, better to do so in 1921 (p. 209). Moreover, weight is
given to precedence, as in Grierson's formula for recovering Maithili
speakers (1901: 320; I 9 I 1: 388; 192 : 219). Expressions from previous
censuses are now repeated as stock phrases: e.g. Maithili is 'Hindi in
its wide sense'. Finally the census loses its character as an occasional
text. Each report incorporates the previous reports in a complex form
of authorship generated by a concern for legal precedence rather than
scientific validity.
The I951 Census is of interest as it was the first census of an
independent India. The ruling Congress Party found itself, however,
in a linguistic dilemma. Although the Indian National Congress had
committed itself in 1921 to the idea of linguistic provinces in the belief
that this was the only way to reach the mass of people in such a
culturally diverse land, by the time the Congress Party had assumed
power in I949 the need for a common language to strengthen national
unity was also recognized. The language of the former imperial rulers
was an inappropriate symbol of national unity and Hindi was the
most obvious candidate for elevation to that role. Thus there was little
REPRESENTATIONS OF SPEECH IN MITHILA 787
reason for the state to recognize Maithili, for such recognition would
exacerbate claims for regional autonomy and diminish the officially
uncontroversial area where the 'main language' and 'national
language' were one and the same. Yet the non-recognition of regional
languages by Congress would betray the trust of those people who had
fought to free the country of imperial rule (see Government of India,
I956).
The 195i Census schedule cleverly sidestepped the horns of this
language dilemma. If Hindi was to function as the national language
it would have to be the subsidiary language in those regions where it
was not the main language. Hence the innovation of the 1931 census,
namely the enumeration of both main and subsidiary languages, was
retained in the schedule, but for the first time Maithili was listed as a
Bihari dialect, the returns of which would be provisionally
enumerated. All 'Urdu and Bihari and Hindi dialects', however, were
ultimately recorded with the Hindi language in the table.
Respondents could not offer Maithili and Hindi as main and subsidi-
ary languages, for they were the same language. The I95i Census
returns marked the nadir of the Maithili language movement. Coming
so soon after independence, the feeling for national unity was high as
was the commitment of local leaders to sacrifice their regional identity
for the sake of that greater unity. In northern Bihar only 97,685
persons returned Maithili as their mother-tongue.
The years following independence were marked, however, by an
increasing disenchantment with Delhi so that in the 1950S the aims of
Maithili advocates gained greater purchase on public opinion. Hav-
ing failed in their attempt to have Maithili listed as a regional
language in Schedule VIII of the Indian Constitution, they looked
forward to the I96I Census as a forum for demonstrating the justice of
their aims. The Maithili v. Hindi issue, together with the Urdu v.
Hindi issue, made the language section the most controversial part of
the schedule, even creating in some areas a 'law and order' problem
(p. 459). Maithili advocates promoted an unsuccessful attempt to
have the residents of Darbhanga and Bhagalpur districts entered ipso
facto as Maithili speakers (based on Grierson's formula), and
numerous rumours circulated within the districts that enumerators
had been instructed to enter Maithili as Hindi. The result of all this
'consciousness-raising' was an increase of some 5000 per cent over the
I95 figure for the number of Maithili speakers. The total of 4,982,6 5
was, however, still less than 50 per cent of what Grierson had cal-
culated for I90 .
From 1872 in which 'Hindee is the language of upper India' to I96I
788 RICHARD BURGHART
and beyond in which 'the figures for Urdu and Bihari and Hindi
dialects have been shown with Hindi in the Table' the knowledge of
the census commissioners remained in force. Discontinuities appear
only in the changing reasons for maintaining that policy. Either the
people are too unsophisticated to return Maithili; or the enumerators
are too unsophisticated. Either the language does not exist; or it
exists, but to record it would upset the projections based on previous
censuses. In 1911 Maithili is Hindi in its wide sense because it is not
Bengali; from I95I Maithili is Hindi in its wide sense because it is
important that the national language be the main language in as
many regions as possible. The continuity in knowledge/policy during
the early and late colonial periods, both before and after
independence, demonstrates how successfully the interests of the
modern state and of its institutionalized census resisted changes in
government and forms of government as much as the 'will of (half) the
people'. For the state Maithili was and is Hindi 'in its wide sense'.
Regardless of how one uses the terms bhasa and boli as labels, the
context in which they are used invariably describes a relation between
an authoritative language and one caught up in local circumstances.
The authority variously derives from its universality, divinity, orig-
inality, purity or the dignity invested in it by the qualities of its speech
community (human and celestial gods). In short, one returns to the
essential/local contrast between chaste and rustic Maithili described
above in the first section.
The preoccupations of the essential/local structure of language
identity touch on Grierson's diachronic reconstructions but not on his
synchronic classification of eastern Prakrit languages. The mutual
unintelligibility of languages and the mutual intelligibility of dialects
is not an issue for Maithil pandit-grammarians in the assertion of the
linguistic autonomy of Maithili. Nor do they discuss the 'standard'
form of the language. Indeed, they do not even mention the existence
of rustic Maithili, for chaste Maithili is essentially Maithili. Only
Maithil linguists (not pandits), writing in English (S. Jha 1958),
render the chaste form as the 'standard'; and only European scholars
(Grierson I909; myself) betray their alien sense of comprehensiveness
by mentioning rustic speech in their descriptions of the language.
What is at issue is the autonomy of the language; derived genetically
from an authoritative language of the past which ultimately stems
from Sanskrit, or derived from the autonomy of the country in which
the language (sthaniya bhasa) is spoken. In the former case what is
critical is the essence (the 'seed' as G. Jha put it) of the present
language in a previous authoritative one, not the question of one being
the 'standard' of the other. In the latter case what is critical is the
uniqueness of its speech forms as artefacts of the country, not its
degree of intelligibility with neighbouring forms of speech.
A language that possesses its own authoritative seed in the past and
unique, local features in the present is autonomous (svatantra; or
'independent' as stated by Misra I976: 20-2). Diachronically auto-
nomy is established by the purity of one's pedigree, traced back to
Sanskrit through Apabhramsa and Prakrit languages. Dependence
upon Sanskrit, as a universal language, can be acknowledged, but not
the dependence of one modern language upon another; or one
REPRESENTATIONS OF SPEECH IN MITHILA 79I
certain if the form of the language taught to them will be the same
form used by the outside examiner. ... ' Moreover, Maithili gram-
mars written by Maithil pandits rely upon Indo-Aryan grammatical
categories (four gender categories, three or four voice categories);
whereas Maithili grammars written by Maithil linguists rely in some
measure on Latin-derived or descriptive linguistic categories. Indeed,
Yadav, an American-trained Maithil linguist from Nepal, refers to the
grammars of pandits as '. .. Sanskrit grammars, masquerading as
Maithili grammars' (1981: 73). Meanwhile attempts to 'modernize'
the language by simplifying its complex conjugational rules and by
bringing the use ofhonorifics more in line with values of an egalitarian
society remain idiosyncratic.
In sum, without Maithili becoming a medium of communication by
the modern state, it is unlikely that a 'standard' form will emerge.
And such a state of affairs is, in turn, unlikely. Bihar counts within its
frontiers not only Maithili but also Bhojpuri, Magahi and the various
Munda languages (Santali, Ho, etc). Maithili is neither the majority
tongue nor the administrative language; it has state certification, but
lacks public currency.
At the national level Maithili advocates have been even less suc-
cessful. In 1965 the All-India Sahitya Akademi recognized Maithili as
a language, which entitled Maithil writers to receive central funds for
the promotion of Maithili literature, but the national censuses con-
tinue to treat Maithili as a dialect of Hindi. Meanwhile, on the
Nepalese side of the border, the government rid its census returns, if
not its territory, of all trace of Hindi by designating Maithili, Bhojpuri
and Avadhi as autonomous languages, not dialects of Hindi. This
administrative action mentally banished from Nepalese territory the
national language of the dominant regional power, and at the same
time turned Maithili into the second largest language group of Nepal
(Census of Nepal, 1971; see also Gaige 1975). Yet there was no gain
for Maithils. Nepali remains the sole state language and, since 1962,
the sole medium of instruction (Gaige 1975). Although political
leaders from various parties in the general election of 1991
campaigned for a reversal of the language policy of the previous
pancayat regime, Maithili still lacks public currency and its future
remains uncertain.
What works against the aims of Maithili advocates, however, is not
only the absence of public currency, but also the way in which the
essential/local contrast subverts the very notion of linguistic auto-
nomy that it constructs. The subversion is particularly evident with
REPRESENTATIONS OF SPEECH IN MITHILA 795
rhetoric, etc.). Despite little caste people having their own aesthetic
values, they make no political claims on the basis of language
representations. It is chaste Maithili, not rustic, that organizes the
frontiers of speech. Moreover, the resistance against the dominant
models of the LinguisticSurveyof Indiaand that of the Censusof India,in
which Maithili has only the status of a dialect, is mounted by those
persons with an alternative language representation, that is, the mem-
bers of the 'big castes' who predicate their local dominance on speech.
In this case, at least, the census schedule and colonial scholarship was
not decisive in 'objectifying' society (see Cohn 1987), for linguistic
self-awareness as a basis of regional and social identity is as old as the
names 'Maithili', 'rustic language', etc.; if anything, the census
became the focus of a contest on how speech was to be socially
constituted.
The fact that language figures in the dominance of the 'big castes',
but not in the tactics of resistance by the 'little castes' suggests only
that language representations have little bearing on how lower castes
express agency in culture. Rather the speakers of the 'going language'
(cdlubhads)act and defend themselves on other terms. The stereotype
of society in northern Bihar as being 'traditional', 'cast-ridden' and
'backward' does not fit squarely with the lower castes having been at
the forefrontof the freedom movement or the fact that Naxalbari is on
the border between the Maithili- and Bengali-speaking areas. When a
Maithili-speaking rickashaw driver from Sitamarhi District informed
me that the Bihar state police is a khagatibhag,or 'paper tiger', one
gets an idea of what other sorts of literature are read by lantern light
in Mithila. Needless to add, the 'Red Army of Muzaffarpur' is not
fighting against the stigma of Maithili being a dialect. These remarks
help explain not only the political failure of a broad-based Maithili
cultural movement (see Brass I974) but also the fact that in the 196I
census less than half the Maithili-speakers returned Maithili as their
mother-tongue. The lower castes have masterful narrative and
rhetorical skills, but they do not elaborate the autonomy of rustic
speech as language.
Grierson's invention of the Bihari language remains a singular
discovery that encountered almost universal rejection: from census
commissioners, from his Maithil pandit informants and from some of
his philological colleagues. Indeed, in view of Grierson's previously
cited remarks on the imaginary Jangali language, the search for
'Bihari' by the local administrators who prepared the I9OI census
schedule takes on some irony. But apart from Grierson's reconstruc-
REPRESENTATIONS OF SPEECH IN MITHILA 799
tion of Maithili's pedigree, his research in Mithila was otherwise
ambiguous in cultural derivation. He took paic kos' Maithili to be the
standard form of speech; and he meant by standard the correct form,
not the prevalent one. He understood Maithili to have not a boun-
dary, but a frontier that might best be represented by the shading of
colours. Given the inadequacy of the synchronic language/dialect
distinction, he was forced to treat Maithili as the language of a place
(sthdniyabhdsa)whose existence could best be described with reference
to its genesis. Although Grierson and pandit-grammarians differed in
their metaphysics, and therefore in what makes history, nonetheless
the philologist's stammbaum model of language change could have been
readily reworded into the terms of the pandit's 'seed' model. Finally,
the conjuring of the Bihari language stemmed not from the arrogant
use of an alien method, but from his recognition of the inadequacy of
the European language/dialect distinction. In short, Grierson's philo-
logical investigations were not in any simple sense of the term an alien
'construction'. He was enough in the hands of local Brahmans that
the difference between the knowledges of investigator and investigated
became ambiguous.
These few observations, plus the fact that Grierson's scheme was
resisted by the census commissioners, suggest that scholarly practices
ought to be distinguished from the institutionalization of scholarly
knowledge by the state and from the effects of that institutionaliza-
tion. In a recent article Cohn (I985) described the work of an earlier
generation of philologists who during the Company period used the
comparative method in order to construct the genealogical history of
Indian vernaculars. Although it is clearly the case, as Cohn (pp. 292-
5) shows, that philologists, such as Sir William Jones, worked from
their own expectations about legal traditions and that they learned
Sanskrit in order to reduce the Company's dependence upon Indian
interpretors of Hindu and Muslim law, it does not follow that the
comparative method of philologists was in itself a purely alien method
or that by means of this method they could give to Indians their
history (p. 326). What Jones learnt from his Sanskrit teacher in
Nadya remains unclear, and Emeneau (I955) in his homage to Indian
linguistics detects the influence of Panini on the emergence of com-
parative philology in Europe. At the level of philological practices
there could be, as was the case with Grierson in Mithila, ambiguity in
the scholar's understanding of others.
With regard to the institutionalization of knowledge by the state the
language representation that counts most and contrasts most with the
800 RICHARD BURGHART
References