The data cited above raise the question of whether "simple Bengali" was a rather misleading banner under which to smuggle other ideologically-motivated efforts to shape Bengali speech and writing. The several goals of the EBLC are not equally transparent, but their contribution to what an outside party might call "simplification" is often less clear than their connection to communalistic nationalism. To return to the EBLC's third goal in the light of other writings by committee members and their contemporaries suggests that the "idiom . . .in common use" which the committee had in mind were Perso-Arabic. While Perso-Arabic words were and are in common use, Pakistan had strong religio-nationalistic motivations for displacing Sanskrit with Islamicate lexemes. Bengali-speaking Pakistanis needed to re-imagine themselves and the center of their collective identity; in this re-imagining Calcutta would no longer be acceptable as the cultural center, if it had been at all in recent decades. A. M. Ahmed's call for a Pak Bangla (Pure Bengali) code which would reflect Muslim usage, particularly that of Dhaka and Mymensingh, seems a close parallel to the EBLC's campaign.14 If Pakistan's motivations are plain, Dil's justification for paying no theoretical attention to the communalist dimensions of diglossia and state manipulation are not.
It is significant that the Committee summarized its policy as "simplifying," incorporating into its packaging an egalitarian rationale. In terms of its short label, they were pursuing not a "purified" but a s;ahaj or "simple" language. Committee members fulminated against past elites whose motivations for upholding standard linguistic forms had been to block the full sociopolitical participation of the masses (Dil 1986: 456). Yet through a combination of inadequate sociolinguistic research and conceptualization on the one hand and, on the other hand, what seems to be the work of class-based ideological blinders, much of what passed for "simplification" was either a return to the colloquial standard of Calcutta (calit bha\s>a\ ) or communalistic purism. Neither would facilitate mass participation in non-communalistic public discourse.
How class undermined the simple language campaign
HCB carries a prestige not to be sacrificed. Dil tells of an interview she conducted with Munier Chowdhury. Admittedly, Chowdhury himself had not been an advocate of s;ahaj Bangla . Nonetheless, it is interesting that Chowdhury's speech during the interview included elements of s;a\dhu bha\s>a\ . When Dil pointed this out, Chowdhury "did not defend himself or contradict" the observation (Dil 1986: 460). He explained that he and others of his class were prone to use the supposedly written "pedantic" standard on "/attanta s;ist>a paribes ;/ meaning 'very strictly formal occasions', [or] for discussing /duraha bis>ay, kat>hin pras;anga / meaning 'difficult subjects'" (Dil 1986: 460f). Together, such settings, genres, participant structures, and act sequences of form and content form a whole in the imagination of such speakers which Hymes says should be linked in "ethnographies of communication" (1972). Foucault and others stress the ways access to such discourses is restricted. But rather than explore the social and linguistic practices which cooperate to restrict access and reinforce diglossic boundaries, the 1986 papers leave all too much theoretical work undone.
Dil's empirical investigation of the recorded speech of highly literate speakers confirms the kind of functional complementarity of usage seen in the paradigm cases of diglossia. The fact that she found that certain speakers in "very strictly formal occasions" (Chowdhury, quoted by Dil 1986: 461) actually speaking s;a\dhu bha\s>a\ may not be what Suhas Chatterjee predicted but it is well in line with diglossic complementarity. Chatterjee's 1986 paper-- which claimed s;a\dhu bha\s>a\ was only written and read, never spoken-- does not make use of specific texts or transcripts of recorded speech in the way that Dil's does. The longest transcript of a naturally-occurring speech event in this whole literature is Dil's representation of the words of one participant in a televised debate (Dil 1986: 457). Nashid Kamal's speech, which Dil claims through observation of other debates is typical, includes not only s;a\dhu bha\s>a\ but also English lexemes, although the verb morphology is consistently calit bha\s>a\ . That is, Kamal's lack the "long stem" suffix /i/. Evidently including some s;a\dhu and English nouns in formal speech achieves some enhancement in prestige without the risk they would have incurred by mixing long (s;a\dhu ) and short (calit ) verb morphology.15
Diglossia is bound up with analytic and folk perspectives on the social practice of writing. Diglossic situations crucially entail attitudes toward written and spoken language. To illustrate this, we return to Chowdhury's colleagues who advocated simple Bengali. Why did they call for publications in "Simple Bengali" but, for the most part,16 fail to write such works themselves? There was so little response following the EBLC's specifications that no accepted model of simple Bengali was ever realized. Dil acknowledges insightfully that one reason for this failure was that "'Shahaj Bangla' was a recommendation for a Government appointed Committee and had not originated from the people [emphasis mine; see discussion below] themselves." She also points to the complexity of the EBLC's specific recommendations, touching so many aspects of language that the members themselves "did not (could not) follow their own recommendations" (Dil 1986: 456). Yet it seems remarkable in an article on diglossia to overlook the pull of prestige. Even for the cause of populism being advocated by their spokesmen, the educated elite of Bangladesh hesitated to abandon certain s;a\dhu features in their publications.17
Dil concludes her article with some fascinating observations. I have added line numbers for ease of later reference.
Rather than blaming these failures of analysis or of policy implementation on individual "hypocrisy", it is more illuminating to link three levels of social activity seen here-- a) perceptions of usage, b) actual usage of policy-makers and others, and c) policies advocated-- with a fourth (d), class interest. A tense dialectic unites these four. Simply stated, class-based rationalization-- i.e. a more-or-less unconscious ideology of language-- colored the policies of the EBLC. Like all ideologies of language, that of the Committee reflected their relation to power. As Woolard points out in her review of the notion of ideology in the framework of linguistic ideologies, theorists do not agree on how conscious and explicit "ideologies" tend to be (Woolard 1992: 237f). Indeed, we find variation in the degree of explicitness of the ideologies informing discourse on Bengali linguistic variation and language planning. Siraj, whose purism was described earlier, did not object to religiously-motivated vocabulary but to the inclusion in speech of forms which he regards as archaically literary. What is remarkable here is his reflexive awareness of usage, including his own usage. Siraj explicit form of elitism may be compared with that of Chowdhury, which is partly masked by his appeal to setting and content as determiners of register use. An even stronger contrast can be drawn between Siraj's explicit elitism and the disingenuousness of the EBLC-- describing Islamization as a "simplification" of Bengali. In on-the-record comments (Siraj's remarks to me cannot be considered such), there is a dispreference for using class as an explanatory factor.18 Just how problematic the ideological foundation of research and policy becomes is illustrated in failures to adequately characterize the on-the-ground situation in Bengali. Typically, neither one's own nor others' language usage/habits are conscious-- or at least accurately and explicitly articulated-- even in elite discourse on diglossia.
Several of the examples of lexical borrowing discussed above can be treated together in a final comment on the maintenance of diglossia and H forms in particular. It is not so surprising that some speakers and writers persist in importing words from prestigious codes into what might otherwise seem to be a "colloquial" register. Examples presented above include my general reference to genres filled with either Perso-Arabic or Sanskrit lexemes as well as specific cases-- Dil's record of Chowdhury's speech peppered with s;a\dhu words and the Dhaka University speeches containing English as well as s;a\dhu borrowings. Modernist assumptions notwithstanding, instrumentally efficient communication is not the only goal driving the choice of linguistic forms. Here the situation in East Pakistan/Bangladesh can be compared with that in Indonesia. Errington (to appear) distinguishes between two linguistic ideologies discernible in Indonesian discourse, the instrumentalist or "standardist" ideology of modernism and the "exemplarist" ideology evincing a prenationalist era of Indonesia polity. While both appear to influence public Indonesian discourse, and the "exemplary" is encompassed by the "standard" ideology of language, the distinction is illumining for our discussion of Bengali diglossia. Exemplarist language-- i.e. speech filled with words which recall an earlier era when kings and gods walked the earth-- may not serve any instrumentalist purpose, may lack clarity and fail to make reference, and it may not serve efficient communication. What it does, however, is indicate the speaker's own status; hearerÕs instincts may grant such speech a semi-divine status. The educated classes of Indonesia-- and Bangladesh-- may have one foot firmly planted in modernity, but the exploitation of earlier linguistic ideologies does nothing to threaten their status. In fact it strengthens it by merging old with new forms of power.
Conclusion
A survey of the literature on Bengali diglossia leads me to make a methodological plea. While further investigation is called for, future publications must include empirical data beyond responses to self-report surveys. These data should preferably entail whole transcripts rather than brief excerpts, but minimally, some sort of transcript of tape-recorded discourse. Moreover, this recorded discourse should include at least some speech events which would have occurred without the presence of the analyst. As William Labov's writings so eloquently demonstrate, if it is the nature of natural speech we seek to know, formal interviews are not likely to uncover it.
What are the facts about Bengali linguistic diversity today? Despite calls for a universally-used simple Bengali, I have no doubt that even more adequately grounded studies would find that diglossia is as prevalent today as it was when the EBLC began its work. Thus, a form of Bengali not so different from the H which the EBLC advocated, one which indexed membership in East Pakistan's relatively tiny urban, literate elite, probably continues to this day to receive recognition by most Bangladeshi speakers as the H or prestige form of the language. We have too few empirical studies of speech-in-use and evaluations of language expressed in everyday discourse. Further, an adequate view of East-West and communal linguistic diversity in contemporary Bengal is nowhere to be had.
The studies needed to give us an adequate picture of speech and writing in Bangladesh and W. Bengal must ask significant questions and contribute to building a big picture. Past discourses on linguistic diversity-- historiographic, sociolinguistic, and language planning writings-- have not constructed a model which accounts for the interrelationships of language behaviors and attitudes with class, ideology, and religion. Perhaps reasonable fears of communal violence have contributed to the failure of scholars to adequately theorize the Bengali language situation. Of equal importance are the influences of narrow disciplinary specialization evident in the three discourses on Bengali. Yet the subtlest and most important source of mystification is the social position of the authors themselves.
Bengali society was dissected communally in 1948. The formation of Pakistan gave salience to the ideological affirmation of pan-Muslim solidarity, but then the communal dividing line became less important as most Hindus left East Pakistan. Still, a vertical form of differentiation remained-- social stratification. As Carol Prindle argues in her ethnography of Chittagonian businessmen, Islamic ideology's denial of any ontological reality to social hierarchy makes problematic any direct recognition of class. She argues that Chittagonians have balanced two discourses, dha\rmik vs. sa\ma\jik , the former denying ontological hierarchy and the latter acknowledging a ranking of persons on the basis of social value. A similar duality of discourse is seen in the EBLC's ambivalence towards s;ahaj Bangla. The Committee's official pronouncements simultaneously (a) discursively imagine the nation (or the Eastern province) in terms which erased vertical distinctions yet (b) obscure and thus perpetuate those distinctions through the pretense that religious purification of the language made it-- and political participation-- universal, accessible.
Clifford Geertz (1968: 62) describes how religion was transformed into ideology in the history of the Muslims of Indonesia and Morocco. This transformation exemplifies rationalization in Weber's sense. Similar transformations surely can be seen during India's colonial history-- and in the history of language, including elite written language about language. Just as religion under the impact of colonialism and modernism underwent a rationalizing compartmentalization, a narrowing from pervasive experience to creed, most analyses of the Bengali language have removed it from its links with Bengali life. Treatments of the Bengali language situation from respective disciplinary vantage points have severed it from its multiplex links with Bengali life, with micro and macrosociological phenomena. A better mutual reading of each other's work by historians and sociolinguists would help reverse academic atomization of the Bengali world. But increased reflexivity in academic writing is required if we are to replace mystification with an adequate model of Bengali linguistic variation.
I have argued that, quite apart from any putatively objective recording of the strategic use of markers of communal ethnolinguistic identity, attitudes toward H and L variants partake of the ideological. One sense in which this is true is that governmental and academic consciousness regarding the gap is distorted/distorting. Compartmentalization is one distorting factor. Historians and sociolinguists have made their respective contributions to what is potentially a whole picture, respectively documenting the play of world-civilizational forces in Bengal and contemporary correlations between speech behavior, associated prestige value, and the social status of speakers. One could attribute sociolinguists' failure to address macrosociological and historic factors to the same failure of sociolinguistics per se, especially in its quantitative incarnations, although that failure is being addressed (e.g. Woolard 1985). But the academic compartmentalization works hand in hand with the subtler pull of class-based ideologies, always political at some level, even as they touch language issues.
Thus, the theoretical failings of these sociolinguists are not merely the blinders of an academic discipline but signs of a mystification resembling that to which the jingoistic advocates of Pak Bangla in the Pakistan era were prey. Increased reflexivity-- something akin to off-record comments cited here but translated into academic writing in particular-- is required if we are to replace mystification with an adequate model of Bengali linguistic variation.19
Anderson, Benedict. 1991 (1983). Imagined communities (rev. ed.). London and New York: Verso.
Bakhtin, M. M. 1981. The dialogic imagination. M. Holquist, ed. Austin: University of Texas Press.
Chatterjee, Suhas. 1986. Diglossia in Bengali. In Bh. Krishnamurti, C. Masica, and A. Sinha (eds.), South Asian languages: Structure, convergence and diglossia . Delhi: Motilal Banarsidass, pp. 294- 302.
Chatterji, Suniti Kumar. 1934. The origin and development of the Bengali language (Volume I). London: George Allen & Unwin.
Chowdhury, Munier. 1960. The language problem in East Pakistan. In Charles A. Ferguson and John J. Gumperz, eds., Linguistic diversity in South Asia: Studies in regional, social, and functional variation.
Dil, Afia. 1986. Diglossia in Bangla: A study of shifts in the verbal repertoire of the educated classes in Dhaka, Bangladesh. In Joshua Fishman et. al (eds.), The Fergusonian impact (Vol. 2, Sociolinguistics and the sociology of language ), pp. 451-465.
Dimock, Edward C. Jr.. 1967. Muslim Vaisnava Poets of Bengal. In Language and Areas: Studies Presented to George V. Bobrinksoy , 28-36. Chicago: University of Chicago Press.
Errington, Joseph. To appear. Indonesian('s) authority. In P. Kroskrity, ed., Language ideologies. Santa Fe, NM: SAR. Geertz, Clifford. 1968. Islam observed: Religious development in Morocco and Indonesia. Chicago: University of Chicago.
Halhed, Nathaniel B. 1778/1969. A grammar of the Bengal language. (English Linguistics 1500- 1800, R. C. Alston, ed., No. 138).
Haq, Md. Enamul. 1957. Muslim Bengali literature . Karachi: Pakistan Publications.
Hill, Jane. 1985. The grammar of consciousness and the consciousness of grammar. American Ethnologist 12: 725- 37.
Hymes, Dell. 1972. Models of the interaction of language and social life. In J. Gumperz and D. Hymes (eds.) Directions in Sociolinguistics: The ethnography of communication., pp. 35-71. New York: Holt. Reprinted: Basil Blackwell.
Kopf, David. 1969. British orientalism and the Bengal Renaissance: The dynamics of Indian modernization 1773-1835. Berkeley and Los Angeles: University of California.
Kroskrity, Paul V. 1982. Language contact and linguistic diffusion: the Arizona Tewa speech community. In F. Barkin et al. (eds.), Bilingualism and language contact , 51-72. New York: Columbia Teachers College.
Kroskrity, Paul. 1993. Language, history, and identity: ethnolinguistic studies of the Arizona Tewa . Tucson: University of Arizona Press.
Mannan, Qazi Abdul. 1966. The emergence and development of dobha\s>i literature in Bengal (up to 1855 A.D.). Dacca: Department of Bengali and Sanskrit, University of Dacca.
Oberoi, Harjot. 1994. The construction of religious boundaries: Culture, identity, and diversity in the Sikh tradition. Chicago: University of Chicago/ Oxford: Oxford University.
Prindle, Carol. 1988. Occupation and orthopraxy in Bengali Muslim rank. In Ewing, Katherine P. (ed.), Shari'at and ambiguity in South Asian Islam (Comparative Studies on Muslim Societies, Barbara D. Metcalf, Gen. Ed.), 259-287. Berkeley and Los Angeles: University of California.
Silverstein, Michael. 1981. The limits of awareness. Working Papers in Sociolinguistics 84. Austin: Southwest Educational Development Laboratory.
Silverstein, Michael. To appear. Whorfianism and the linguistic imagination of nationality. In P. Kroskrity, ed., Language ideologies. Santa Fe, NM: SAR.
Singh, Udaya Narayana. 1986. Diglossia in Bangladesh and language planning problems. In Joshua Fishman et. al (eds.), The Fergusonian impact (Vol. 2, Sociolinguistics and the sociology of language ), pp. 431-449.
Umar, Badaruddin. 1970. purbo ba\m>la\r bha\s>a\ a\ndolan o ta\tka\li\n ra\jni\ti (East Bengal's Language Movement and the politics of that era).
Woolard, Kathryn. 1985. Language variation and cultural hegemony: Toward an integration of sociolinguistic and social theory. American Ethnologist 12: 738-48.
Woolard, Kathryn. 1992. Language ideology: Issues and approaches. In Kroskrity, Paul, Bambi Schieffelin, and Kathryn Woolard, eds. Language ideologies. Special issue of Pragmatics (2).
1 This paper profited from discussions at the Bengal Studies Conference, particularly comments by Professor Anisuzzaman of Dhaka University. Naturally this author bears sole responsibility for the perspectives which remain after suggestions have all been processed!
2 According to Mannah (1966: 69f), Bha\rat Candra Ra\y's poem, Annada\man³gal , refers to this styles as bha\s>a\ ya\bani misa\l (lit., "language foreigner mixed," i.e. an Islamicate-mixed idiom).
3 It is not at all clear that the orientalist-trained pandits wrote uniformly in a Sanskritized Bengali from the earliest days of the 19th century. In fact, Dil (1986: 453) describes the great contrast between the Sanskritized style of Mrtyunjoy and the heavily Perso-Arabized lexicon of Ramram Basu.
4 This explanation of the divergence of communal forms of Bengali is foreshadowed in Haq and Chatterji. Haq claims that "Hindu-British hostility to Muslims resulted in a recrudescence of Muslim consciousness of their separate identity and the Muslims especially of lower Bengal, filled their language with more and more Arabic, Persian and Urdu words." As for Chatterji, despite his denigration of Musalmani (see fn. 7), by calling this variant of Bengali, "the Maulavi's reply to the Pam>d>it's sa\dubha\s>a\., he implicitly "blames" Ft. Williams College rather than the Muslims.
5 "'Dobhashi' was an inferior poetic medium to standard Bengali (Haq 1957: 175, cf. 192).
6 Those four examples were High and Low German and Greek, Haitian Creole/French, and classical/colloquial Arabic.
7 This denial is odd, and does not characterize other sociolinguistic analyses of Bengali. For instance, Munier Chowdhury did highlight phonological differences in his depiction of the language situation of East Pakistan in 1960, but he presented those differences in terms of the urban-rural dichotomy and standard vs. regional variants.
8 It is noteworthy here that even such novels as Surya Dighal Bari, whose pages effectively and artistically represent rural speech, reserve for the narrator a HCB (Hi Calit Bengali).
9 If we take Halhed as a reliable witness to Bengali "diglossia" (though he refers to "Hindustani" but that is his term for the form of Indo-Aryan spoken in Bengal, evidently), we might believe that Perso-Arabic borrowings outdid the prestige of Sanskritic borrowings in the 18th century. Halhed wrote, "[A]t present those persons are thought to speak this compound idiom with the most elegance, who mix with pure Indian verbs the greatest number of Persian and Arabic nouns.' (1778: xi).
10 Sanskrit and Hinduism are as inextricably linked in the subcontinent of the twentieth century as are Arabic and Islam. To Pak nationalists, this Sanskritic influence in Bengali became a symbol of its pollution. Conversely, S.K. Chatterji and other twentieth-century Hindu scholars of Bengali have expressed their disgust at Perso-Arabized Bengali. "The literature in [Musalmani] Bengali has no merit and some of the deathless tales of [Persia and Islam] ... have been ruined by the hack versifiers of Calcutta and Chittagong in rendering them in this jargon" (1934: 211f).
11 Both Halhed (1778: xi) and the EBLC (cited by Dil 1986: 454) mention the unintelligibility of heavily Sanskritized registers to the Bengali masses. I myself have seen ###
12 The written discourse she cites is an indirect citation. The piece in question, a\ma\der bha\s>a\r rup, appeared in translation in Munier Chowdhury's 1970 Bangla gadyariti (Dhaka: Central Board for Development of Bengali).
13 For a critical discussion of Benedict Anderson's use of "the trope of 'we'ness" in the imagination of community, see Silverstein (to appear).