Previous Next Up Comments
SEHR, volume 5, issue 1: Contested Polities
Updated 27 February 1996

ethnographic politics and the discourse of origins

Gauri Viswanathan

Writing on the contested histories of Ayodhya and Somnath, Peter Van der Veer refers to the inimical relations between Hindus and Muslims as "one of the most important master-narratives of colonial orientalism in India."[1] Van der Veer argues that the manner in which Hindu-Muslim relations were construed in British historiography was crucial to legitimating the validity of British rule as one of enlightened disinterestedness. It is also clear, however, that the "facts" marshalled by British commentators are central to the description of Hindu-Muslim relations. The destruction of Hindu temples by Muslim rulers and the forcible conversion of Hindus at sword-point represent one type of fact that underwrites a master-narrative of violent change, as recorded, for example, in James Mill's work, The History of British India (1817). In opposition is another kind of fact, also documented in British historiography in works such as Thomas W. Arnold's The Preaching of Islam (1896): the patronage of Hindu shrines by Muslim saints and Muslim tomb-worship by Hindus; the sharing of titles and names, as well as certain social practices and customs, by both Hindus and Muslims, and so forth. If both "facts" are equally accurate descriptions of Hindu-Muslim relations, then tolerance and intolerance can be defined as the respective absence or presence of violence and forcible change from outside.

But by the last quarter of the nineteenth century, when the first British census reports were commissioned to establish the causes of Muslim expansion in India, the official sociology of India no longer depicted religious change through outside intervention as the central issue: Islamic conversion was represented as having less to do with either the coercive or the charismatic character of Islam than with economic necessity or social ostracism from Hinduism. What models of tolerance or intolerance are then suggested by this reading of Muslim expansion, in which conversion to Islam is the result neither of a gradual mystical insight that incorporates aspects of Hindu worship nor a violent rupture of existing beliefs, but rather of the exclusionary character of caste Hinduism? Indeed, in this description the agency of intolerant action would seem to have shifted from Islam to Hinduism, which, by casting out its members, enables Islam to offer the possibilities for social betterment to these excluded groups. But the meanings of exclusion and incorporation are as volatile as that of tolerance and intolerance, for the conversion to Islam precipitated by lower-caste status in Hinduism also gives rise to movements of reconversion to Hinduism. These attempt to reverse the principle of exclusion and challenge the appeal of rival religious systems by re-absorbing those who had earlier been cast off or had not been fully assimilated. Yet, belying their incorporative philosophy and reformist tendencies, reconversion movements such as the Arya Samaj exhibit a morbid defensiveness that finds expression in group solidarity and an enforced collective identity. Indeed, the point of the reformist discourse of these movements is that, in redefining the boundaries of the Hindu community, reconversion is incorporated into the discourse of Hindu nationalism.

Posing similar questions about incorporation and exclusion (albeit in a different context) in a provocative essay on Tamil art, Vidya Dehejia points to the appropriation of Vaishnavite features in Shaivite art as possible evidence of sectarian tension.[2] Dehejia contests the assumption that wherever there is rivalry or contention between two religious communities, one should expect to find not appropriation or borrowing of features of the rival religious system, but its total destruction, and that the establishment of a separate identity normally requires negation of the competing system. Dehejia's argument raises fundamental questions about whether tolerance can be assumed to be equivalent to syncretism, and intolerance to absolutism and exclusivity.[3] How, for example, do we respond to reconversion movements which attempt to include, in a broadly reformist way, groups that had been formerly excluded? Does the gesture of reclamation and incorporation efface the earlier one of marginalization? Suppose the groups being courted do not want to return to the fold -- are they denying their true origins and willfully affiliating themselves with a community with which they are united not by belief but by social circumstance?

With this discussion of the complex meanings of appropriation and exclusion in mind, my intent in this essay is to locate the point in British discourse where the "facts" of forcible and violent change, as presented in works like Mill's History of British India (1817), and the "facts" of peaceful assimilation to Islam, as presented in Arnold's The Preaching of Islam (1896), are no longer clearly demarcated as self-evident examples of Islamic intolerance in the first instance and tolerance in the second -- where, in short, incorporation and exclusion resist being unproblematically located in ideas of religious syncretism and religious absolutism, respectively.

The point in British discourse at which narratives of tolerance and intolerance acquire a shifting center of reference -- at times it is Islam and at other times Hinduism -- is determined, as I suggested above, by the writing of India's official sociology in the third quarter of the nineteenth century, when the first census reports, settlement reports, and district gazetteers were commissioned. The enumeration of India's populations marks the period when the boundaries of religious communities are redrawn in relation to empirically derived explanations about the expansion of the Muslim population in nineteenth-century India. While Muslim groups are identified as separate from the Hindu community and therefore also a separate political entity, a large majority of Muslims are also recognized for the first time in British discourse, as Peter Hardy notes, as originally having been Hindus who had converted for reasons other than direct force or spiritual illumination.[4] For the first time, the discourse of nationalism is processed through a discourse of origins. For instance, questions confronting British administrators included deciding how the numbers of Muslims in India were to be categorized -- as descendants of those who originally came from Arab lands and were subsequently indigenized (i.e., "hereditary" Muslims), or as descendants of native Hindus who had converted. Were Hindus who had converted to Islam to be considered less Muslim (i.e., more Hindu) than other Muslims of Arab-descent groups?

I want to approach these questions of origins through that institutional instrument most directly focused on the determination of identity -- census-taking -- and, in particular, census-taking as an established feature of British colonial administration. The census reports on India issued between 1872 and 1901 made the first systematic attempt to categorize the religious identities of Indian peoples (including converts) according to criteria of racial origin, customs, and laws. In the course of such categorization, various oppositions were constructed out of the material of enumeration -- oppositions such as foreign/indigenous, national/local, pure/hybrid, lineal descent (or hereditary)/convert. In assessing the strength of the Muslim population in India between 1872 and 1901, the census threw the bulk of its weight on the side of the second term in each of these oppositions to draw a picture of the Indian Muslim, not as an autonomous "other" but as a version of the Hindu at an earlier historical moment before the advent of Arab, Afghan, and Turkish groups -- and before possibly forcible conversion of Hindus to Islam.[5] The "contrived" assimilation of Muslim Indians to Hindu India is not simply a nineteenth-century Indian nationalist strategy of fighting colonial oppression, as it is portrayed by recent Hindu revivalists, but a feature of late nineteenth-century British discourse itself. Indeed, the British assertion of the local origins of Indian Muslims challenged the separatist impulse among Muslims as a claim that was belied by the "facts" accumulated by British ethnographic data, census reports, and commissioned surveys -- facts which placed the Muslims closer in racial features, behavior, habits, and customs to other native inhabitants of India, including Hindus. The "Muslim" represented in the British census reports is marked by an ambivalent identity -- neither truly Muslim nor truly Hindu, riven by social class differences that, in their turn, displace the possibilities of a unity of religious belief or identity. The critical issue in the historiography of Hindu-Muslim relations is not so much that British policy conceived of Hindus and Muslims as separate communities, but that the theory of common origins -- from which other social and religious identities were willingly or forcibly adopted -- produced a crippling situation that disallowed either total unity of Hindus and Muslims or total division between them.

I should point out that I am not reading these reports to suggest a continuity between British colonial discourse and the rhetoric of modern communalism, or to argue that the roots of contemporary communal problems lie in late nineteenth-century techniques of information gathering. Rather, what I would like to argue is that acts of classification, such as the census, establish the categories of knowledge about the racial and religious composition of the people it enumerates that enter the domain of memory for the colonized -- a fluid and shadowy realm of meaning that acquires a suggestive power and resonance in the construction of future relationships between India's ethnic and religious groups. The shift from elite to mass politics in Indian nationalism gave a new importance to the masses of Muslim converts who were denied an origin outside India. As a descriptive catalog of India's ethnic composition, the British census establishes fixities of racial and religious categories, even as it insinuates the possibilities of overlapping and common origins rather than real historical difference. The function of the census to introduce categories of difference and then deny them must be seen to have a complex effect on the structure of perceptions in Hindu-Muslim relations, if not on those relations themselves. My interest lies in examining the mediating role of British ethnography in the production of a field of remembered identities, both Hindu and Muslim, that feeds into the discourses of religious nationalism.

census-takers and the origins of muslims

The first systematic assessment of India's Muslim population was made in the British census reports of 1872. H. Beverly, the superintendent of the census, made the potentially explosive assertion that the large presence of Muslims in Bengal was due not so much to the introduction of foreign blood into the country but to the conversion of the former inhabitants, for whom a rigid system of caste discipline made Hinduism intolerable.[6] Many Bengali Muslims took exception to this conclusion, and Khondkar Fazli Rabi wrote The Origins of the Musalmans of Bengal (1895) to prove that the truth was indeed quite the opposite: he was at pains to point out, for example, that many leading Muslim families could trace their origins to foreign roots -- families such as the Saiads, who refrained from intermarriage with families of more "dubious" ancestry. Piqued by what he took to be Beverly's social condescension, Rabi wrote,

It can safely, and without any fear of contradiction, be asserted that the ancestors of the present Musalmans of this country were certainly those Musalmans who came here from foreign parts during the rule of the former sovereigns, and that the present generation of Musalmans are the offspring of that dominant race who remained masters of the land for 562 years.[7]

Other Muslim historians, however, were less extreme in their claims, and though committed to the theory of the foreign origin of Indian Muslims, they reluctantly admitted that local converts dominated the total. At the same time the figures quoted were generally conservative. Abu A. Ghaznavi, who was asked by the British to respond to Fazli Rabi's claim, calculated that roughly twenty percent of the Muslims living in Bengal were lineal descendants of foreign settlers, fifty percent had a mixture of foreign blood, and the remaining thirty percent, he claimed, were probably descended from Hindus and other converts.[8]

The 1901 Census, however, dismissed these figures as too disproportionate and placed the percentage of converts from Hinduism much higher. The idea of the original "Hindu-ness" of Muslim inhabitants extended to the argument that the early Muslim invaders in Bengal were not even Arabs but Pathans. Yet the fact recorded in the census is that the Muslims who called themselves "Shekh" outnumbered those who professed to be Pathans in a ratio of fifty to one, and furthermore, that many of these "Shekhs" had only recently begun to claim this name and were formerly known as Ashraf in south Bengal and as Nasya in north Bengal.[9] Two different commentaries are thus juxtaposed in a contained narrative of conflicting memories: the descriptive record of Muslim self-definitions as Arab-descended is framed by a commentary that negates those self-perceptions and posits an alternative explanation of Muslim origins in the fractured space of Hindu communities.

Explanation became even more racialized through the ethnographic contributions of Herbert Risley, who was brought into the census-taking operations at a crucial stage of description. The ethnographic scale of measurement, or "Cephalic index," that he devised conclusively "proved" the Hindu origins of Indian Muslims, despite the latter's claims to foreign ancestry that their names and titles presumably asserted. By taking measurements of the proportion of the breadth of the head to its length, as well as of the breadth of the nose to its length, Risley placed Muslims closer in racial features to the lower castes of Chandals and Pods than to Semitic peoples.[10] Here is a clear instance of how the discourse of class, blending indistinguishably with the discourse of race, appropriated the category of religion as uniting both discourses; it became possible to state that "although the followers of the Koran form the largest proportion of the inhabitants [of Rangpur district], there is little reason to suppose that many of them are intruders. They seem in general, from their countenances, to be descendants of the original inhabitants."[11] The split between "original" Muslims, defined as those who comprised the higher classes, and local Muslim converts from Hinduism, who were consistently identified with the lower classes, did two things: first, it accentuated differences not so much between Hindus and Muslims but between Muslims and Muslims on the point of foreign or native descent, with Muslims converted from Hinduism being regarded more ambiguously as Muslim and more relationally placed vis-a-vis Hindus; secondly, the dichotomy of foreign versus locally descended Muslims replaced a unity of Muslim identity -- which the profession of Islam presumably implied -- with categories of differences based on social class. Both factors figure importantly in the reconversion movements led by Hindu groups as early as the nineteenth century, and which continue to function today in certain regions of India (especially in those areas where mass conversions have taken place, such as in Meenakshipuram).

The reconversion movements (which often include rituals of purification, or shuddhi) are a relatively unique phenomenon in that they seek to reverse the total excommunication from Hinduism that apostasy to any other religion generally demanded. Furthermore, reconversion is premised on the activation of remembered identities long since lost or abandoned. The readmission into Hinduism of converts to Islam required, often as a test, that they display types of behavior that no Muslim would ever be identified with, such as eating pork. (Many Muslims who had converted from Hinduism had still not adopted the taboo against pork-eating. For such converts, it was thus possible to exhibit those behaviors that made them acceptable to caste Hindus[12].) Most important, the emphasis in reconversion rituals on practices, habits, and usages as markers of religious identity bears a strong resemblance to a similar emphasis in the British census reports that gave greater weight to customs and practices over the self-declarations of religious identity as a means of classifying religious groups in India. If apostates could be reclaimed back into the fold despite their earlier rejection of Hinduism in favor of another religion, such reclamation was made possible by a political discourse of religious identity whereby a Hindu remained a Hindu by virtue of retaining certain social customs. What the census report does, in other words, is establish a set of scientifically derived representations that enables the Hindu community to claim Muslims among its own by virtue of criteria drawn from racial categories suggesting cultural continuity.

In its preoccupation with the question of Muslim origins, the census revealed its own bias toward downplaying the foreign element in the composition of Indian Muslims, only one sixth of whom were placed as Arab- or Pathan-descended Muslims. The rest were listed as local converts from Hinduism who still preserved habits and usages from the religion they had supposedly repudiated. The census report consistently accentuates the "Hindu-ness" of Muslim converts in proportion to minimizing the self-definitions of those whom it sought to enumerate, the census-taker often assuming the prerogative of listing them under the group to which he thought they belonged, even though extensive inquiry was adopted as a means of eliciting more detailed information from Muslims and other religious groups on where they placed themselves. In almost every category -- age, religion, caste, marital status, and so forth -- questions generated a bewildering range of responses that often led the census-taker to make the determination himself. An infuriating inexactitude of response emerges as one of the central frustrations of census-taking (for age, many inhabitants of Indian villages were known to respond with a blithe bis-chalis" [twenty or forty]), and such vagueness encouraged the census-takers to transfer the authority for self-classification from the subject to themselves. On the matter of religious classification, the census-takers had clear instructions that they were to accept each person's statement about their religious affiliation, no matter how vague or imprecise it might be, but in practice this rule was systematically overlooked. Instead, the census-takers took it upon themselves to decide whether an individual was Hindu or Muslim, which in practical terms often meant determining whether a Muslim could trace his roots to foreign ancestors or whether he was descended from local converts from Hinduism. Often such determinations were based on a combination of the customs, usages, and practices followed by the individual and his racial features, rather than personal declaration of religious identity or religious belief, the latter being routinely effaced in the final classification. (This was also true, incidentally, with regard to Christian converts who were often judged as Hindus, not as Christians, in cases they filed in the British Indian courts for restitution of rights they had forfeited under Hindu law, the basis for the decision being the degree to which their behavior, habits, and manners conformed to those of Hindus, such as preserving the joint family system.)[13]

Classification was made even more problematic in the case of Muslims who appeared to follow Hindu customs to some extent and had half-Hindu names, yet called themselves by an upper-class Muslim title. Some were formerly high-caste Hindus who, on conversion to Islam, were allowed to assume upper-class Muslim titles such as "Shekh" even though they continued to adhere in part to Hindu customs and, in a few rare cases, even to intermarry with those Muslims who were of foreign descent. On the contrary, the lower castes, who were often converts, had to be content with the title "Nau-Muslim," or "New Muslim." It was only in the case of converts who came from functional groups that Hindu names and titles were still retained, such as Kali Shekh, Kalachand Shekh, etc. As a Muslim convert of low social position rose in station, he was likely to assume more high-sounding designations that combined both Hindu and Muslim names. For instance, almost in a crude sort of parody, the gradual upgrading of a low-caste convert like Meher Chand is seen in the progressive combination of names and titles that he acquired through conversion to Islam: Meher Ullah, Meheruddin, Meheruddin Muhammad, Munshi Muhammed Meheruddin, Munshi Muhammed Meheruddin Ahmad, and finally Maulavi Munshi Muhammed Meheruddin Ahmad.

Perhaps the most damaging assessment made by the census report, at least in terms of the repercussions that it had on future constructions of Muslims as "outsiders," was that while the majority of Indian Muslims were identified as local converts or descendants of converts from Hinduism, the conclusion established by the British census-takers was that Indian Muslims saw themselves as "other"-defined, their point of reference for personal identity lying outside India in a quasi pan-Islamic unity:

All Mohammedans look on Arabic as their sacred language and they interlard their conversation with any Persian or Arabic words they can pick up from their Mullahs or from their religious books. The grammar remains Bengali and it is only some of the vocables which are changed. The better educated converts often deliberately abandon their native language. The Garpeda Bhunjas of Balasore furnish an illustration of this. They are descended from a Brahman and the females are still so far imbued with Hindu prejudices that they abstain from beef. But they have completely given up the use of Oriya and now speak Hindustani even in the family circle.[14]

In this accent on the practice of "difference" by those who were indeed drawn away from Hinduism at one point in their history, the British census gave Hindu nationalists of a later generation a language of foreignness and otherness to describe Muslims who had no proper claims to a unique foreign identity, and who yet were said to have made such claims in a gesture of denial of their (in many cases) Hindu origins.

In other words, the British census produced a complex construction of Muslim religious and ethnic identity playing on both the assertion and the denial of difference. The very function of the census was to show, through enumeration, that the assertion of "difference" -- the idea of Muslims as outsiders -- was propagated by Indian Muslims themselves. Once this was established as a specifically Muslim claim, that declaration of difference was promptly denied by the categories adopted in British census-taking, which sought to demonstrate that the bulk of the Muslim population came from local converts. While Indian Muslims were anxious to link their ancestry to Arab roots, British commentators seemed bent on proving their mixed heritage -- that the majority of Muslims in India at the time of the first major census in 1872 were indeed converts and not descendants of Arab settlers and conquerors.

romance or coercion: ethnographic plots

One of the most fascinating sections of the 1901 census is an appendix that lists individual cases of conversion in various districts of east and north Bengal. In this listing, the cause of the vast majority of conversions is established as neither proselytism nor doctrinal conviction, but romance. The elaborate narrative sketch that follows each case reads like a romantic novel in its stress on Hindus converting to Islam primarily to marry men or women with whom they had fallen in love.[15] As in any romance novel worth its name, each story is carefully annotated with the names of principal characters, central episodes, conflict, and climactic resolution.

The comparison with the romantic novel stops at the point of narrative structure, however. Though romance is presented as the main motive for conversion, the play of human desires and feelings has no place here. The enumerated instances of mixed marriages, in which marital union between Hindus and Muslims is achieved only with the conversion of one partner to the religion of the other, become examples of exile, excommunication, and existential isolation. The potential function of interracial love and marriage to offer a model of cultural syncretism -- as a counterpoint to the homogeneity of a politically determined religious culture -- is less emphatic in the British account than the irreversible loss of community, which results from romantic attachments that impel individuals to convert in a final act of desperation. Though twenty-one of the forty reported cases list romance as the motive for conversion, the inner details of each case suggest that caste is the main player, not love: it is not the impulse of love that drives Hindus to embrace the religion of their Muslim spouses, but the fact that they lose membership in their former community as a result of their romantic attachments. In other words, the real cause for conversion still continues to be a condition that is built into Hinduism -- its ability to render caste members as outcastes through mere association with non-Hindus, mainly by having romantic intrigues with them, but also by sharing food with them or coming under their care during illness (the two other primary causes of conversion to Islam listed in the appendix). Only six of the forty cases listed suggest that doctrinal inclination -- the inherent appeal of Islam -- had anything to do with conversion. These instances add up to show the extent to which a vast number of Hindus became Muslims not because they chose to or in order to attain the objects of their desire, or even for reasons of practical expediency, but because the door had been permanently shut on them by Hinduism.

The presentation of case histories is itself marked by an unusual reflexity and self-consciousness. One of its most conspicuous features is the scrupulous annotation of the religious and racial authorship of each section of the returns as Hindu, Muslim, or British. It is to be expected that the sections written by Muslim informants would stress that Hindus converted to Islam voluntarily and as a result of the deep impression made by Koranic teachings, communicated not just by preachers but by enthusiastic lovers as well. It is also to be expected that the descriptions written by Hindus would minimize the role of individual conviction and attribute conversion to force; by extension they would emphasize seduction rather than love. The persistent divisions between Hindus and Muslims on the interpretation of conversion -- Muslims claiming that the bulk of Islamic conversions were voluntary and Hindus claiming that they were forced -- are reproduced in the interpretations of love and marriage between members of the two communities. Whereas the Muslim authors stress the conventionally romantic aspects of Hindu-Muslim liaisons and the attractions of Islamic faith as intrinsic to romantic love, the Hindu authors of the report's various sections dismiss love as a motive existing independently of proselytizing zeal and signifying emotional and spiritual needs.

The sections attributed to British authorship articulate a position that appears to work in a space between the two positions on romance and conversion taken by Muslim and Hindu authors, representing the two extremes of volition and coercion. But the British position turns strategically on a stance of measured uncertainty and ambivalence. Is "falling in love" the result of free will or manipulated desire? That central question is raised but never answered, because the event of "falling in love" is shifted from the category of effect (i.e., the result of conviction, sexual passion, emotional or spiritual needs, desire to establish autonomy in human relationships, etc.) to that of cause (i.e., of excommunication, civil death, and eventually conversion). The focus again returns to the caste features of Hinduism and to the notion of an undivided community before the disruptions wrought by mixed marriages and the threats to a stable religious identity that such marriages pose. This movement parallels the shift in emphasis, so apparent from the very mode of the census-taking operations, from the individual to the community, the latter increasingly subsuming the subjectivity of converts and reducing their actions to the helpless reactions of those who have been shut out from a secure and proper place in Hindu society. The conversion of Hindus to Islam in the context of romance is represented in these reports as a result of excommunication from Hinduism, not as willed change or the exercise of unfettered choice whose unfortunate outcome happens to be exile and excommunication.

origins and the right to self-governance

What are we to make of the pattern discernible in the census reports that suggests a keen British interest in proving the Hindu origins of the Muslim population of India? In considering several hypotheses, we may find it profitable to examine Peter Hardy's provocative argument that deciding whether Muslims were either foreign settlers or local converts was vital to resolving the British debate about whether to confer the right of self-governance on Indians. If, as Hardy speculates, Muslims in British India were descendants of foreign settlers with a culture foreign to India, the British could claim justification for not treating the population they ruled as a united people capable of sustaining self-governing institutions of a kind that required, for their successful functioning, a modicum of shared moral values. I have shown, however, that the census reports seemed not to favor the theory that Muslims were foreigners but rather that they were converts from Hinduism. This suggests that a different type of reasoning may be behind the conclusions put forth by the census reports. Hardy does indeed consider the possibility that the presence of a large number of voluntary converts among Muslims could have suggested to the British political establishment an inherent instability about the Hindu community and innate fissures within its membership, giving grounds for the British suspicion that India could not be dealt with as a homogeneous political community.

I would like to advance two slightly different arguments, however. First, the emphasis on class differences between so-called hereditary Muslims and Muslim converts is part of a well-documented tendency in British commentaries to explain Indian society in terms of the caste system. Hindus converting to Islam presumably repudiated not merely a religion or world-view but caste itself. The British, however, interpreted the converts' imperfect assimilation into Muslim society and the pursuit of titles and rank by low-caste converts not as a genuine desire for upward social mobility, as more recent Muslim historians suggest,[16] but as evidence of the continuing influence of Hindu social ideas and the perpetuation of an invisible caste system even in the new Islamic order.[17] Indeed, the dislike of educated Muslims for the theory that most of the local converts in east and north Bengal were from the lower castes is doggedly read by the British as reflecting a persistent caste mentality. That is to say, if Indian Muslims wanted to be recognized as descendants of foreign settlers, they must have been motivated largely by a desire to conceal their low-class Hindu origins. Edward Gait, the census commissioner of the 1901 report, even remarked:

The Moghals are converts, just as much as are the Chandals [a low-caste tribe of Bengal]. It is only a question of time and place. The Christian religion prides itself as much on converts from one race as on those from another, and except for the influence of Hindu ideas, it is not clear why Muslims should not do so too.[18]

We can infer from this sort of analysis that the British determination to prove the local origins of Indian Muslims assumes that any future political scheme for India would have to consider -- even to the point of reproducing to some extent -- the systems of social stratification by which Indian society had come to be defined in colonial discourse.

The second argument I would make, however, might seem to contradict the first. At the same time that the census reports showed a tacit recognition (and even acceptance) of social stratification, there was also an eagerness to show the forces of change that had been set in motion in "traditional" Hindu society, to the point that outcastes like the Kochs of Bengal were converting to Islam because they had a "disposition to change."[19] That the conversions revealed the existence of a volatile and dynamic society, constantly in flux, appeared to confirm some of the positive consequences of outside intervention, be it by the Mughals or the British. Thus the almost inordinate British preoccupation with proving that the large majority of Muslims in Bengal were originally Hindus ran parallel to the British zeal in demonstrating the validity of catalyzing India into social and cultural change.

But the effect of this dual, contradictory move was to further complicate the ambivalent identity of Indian Muslims, who are represented as having both rejected (either voluntarily or involuntarily) the religion they once professed (even several generations removed) and retained aspects of it in their social orientation. There are, therefore, two kinds of often conflicting memories that inhabit the Hindu past: one, the memory of having once been an undivided community that had been violently torn asunder by foreign invasions, depredations, and cultural violence, of which forcible conversion is the most radical and divisive; and two, the memory of betrayal, repudiation, and willful reaffiliation to another community that the Muslim self-definition as "foreign-descended" appeared to suggest to Hindus. In both cases the Indian Muslim could not readily be identified as either outsider or insider. The sense of betrayal is further accentuated by the mythologies surrounding forcible conversion which were often propagated -- or so claimed the British census reports -- by Muslims themselves. One typical story recorded in the census report tells of a time when the Muslim population in Bengal was still scattered and it was customary for each Muslim dweller to hang an earthen-pot [badana] from his thatched roof as a sign of his religious affiliation. The census report recounts a story about a learned maulvi who, after a few years' absence, went to a Hindu village to visit a disciple dwelling there. Unable to locate the latter's earthen-pot, he was told on inquiry that his Muslim disciple had renounced Islam and joined a tribal group. The maulvi on his return to the city reported this incident to the nawab, who in a fit of rage ordered his troops to surround the village and compel every person there to become Muslim. As part of Muslim folklore, this extravagant story is narrated at one level as an example of Muslim assimilationist zeal and dogmatic pride. But when retold in the context of the census report, it has the effect of mythologizing the increase of the Muslim population in Bengal and removing history and ideology from the construction of a hybrid Indian identity.[20]

In increasingly alarming ways, Hindu revivalists have sought to reinscribe that history and ideology through the reconfigurative instrument of memory and through rituals like reconversion, which in many respects functions as the handmaiden of memory. The whipped-up hysteria surrounding forcible conversion as one of Muslim India's most bitter legacies is not just expressive of Hindu antagonism to Muslims and to the history of violent rule in the past that the Muslim presence connotes. The hysteria is also part of a well-developed, concerted effort to remind Muslims of their original identity as Hindus, inasmuch as it is, at the same time, a sinister reminder to them that Muslim claims to "difference" and "otherness" are falsely founded and therefore untenable. (The BJP's insistence on eliminating separate personal laws for Muslims, concerning such things as marriage, divorce, and inheritance rights, and on developing a common civil code by which Hindus and Muslims would be governed alike derives, I think, from this reclamation of Muslims as Hindus.) Within this extreme logic, if there is anything worse than marauding Arab and Afghan invaders plundering Hindu temples and destroying Hindu religious life and culture, it is the fact that those who were once Hindus and subsequently converted (even if only because they occupied a low or outcaste position in Hindu society) now dare to deny their "true" heritage and make claims to a separate religious (and also political) identity.

What makes the intolerance of Hindus invisible, especially to Hindus themselves, is a rhetorical strategy that can be seen in the British census reports and which Hindu nationalists have subsequently adapted for their own purposes: contrasting the fluid, mercurial status of Muslims -- they are either foreigners or converts, but never presented as having a direct, unmediated relationship to India -- with the fixed, essentialized status of Hindus as the original, real inhabitants. Though the census introduces the category of "Animists" to suggest a pre-Aryan presence in India, the incorporation of the culture of animists to that of the early Aryans is presented as a process outside conversion and religious expansion. While Aryan incorporation of animist features in such things as Aryan stone-images is unchallenged as an example of syncretic adaptation, a similar process of incorporation-as-assimilation (for example, Hindu stone pillars used as steps in mosques) is considered a defilement of Hindu culture by Muslim conquerors. The fact that the former is considered an example of tolerance and the latter of intolerance has a great deal to do with the distinctions drawn between an originary culture and a culture described as derivative and foreign. While, as I suggested above, the census account of the conversion of lower-caste Hindus to Islam would seem to have shifted the agency of intolerance from Islam to Hinduism, the simultaneous representation of Hinduism as the "original" religion of India removes Hinduism from a history of expansion and religious conversion as active as that of, say, Islam. To conceptualize Hindu-Muslim relations as a relationship of native to convert (as the British census does) or native to foreigner is to introduce notions of incorporation and exclusion that become ideologically charged in the struggle to affirm origins. Sadly, history vanishes, leaving only distorted memory in its place.


The BJP/VHP forces claim that the destruction of the Babri Masjid on December 6, 1992, has now made the temple-mosque controversy a non-issue. Nothing could be further from the truth. But that is not because the demolition of the mosque had been preceded by viable alternative ways of resolving the crisis. If there was a strange commingling of modernist and mythic elements in the political symbolism of the Ram movement, its very oddness would have seemed to make it open to swift dismantling. But the reaction against Hindu revivalism has not been particularly successful in enabling a higher level of discourse to develop, for the counterreply either asserts an equally positivist claim (i.e., the Babri Masjid was built on an empty plot of land, Islamic settlements pre-date Hindu presence in Ayodhya, Islam forbids the construction of a mosque over a "pagan" temple, and so forth), or it takes the form of postmodern skepticism against all truth claims (i.e., there was no temple, there was no Ram, there was no Mughal "invasion" or forcible conversions or iconoclasm, for the history of Ayodhya, like all given histories, is subject to doubt and can never be known).

In either case religious belief, Hindu or Islamic, remains unplaced and unaccounted for. No matter what the evidence might be for or against the existence of a Ram temple before the Babri Masjid came to be built, the weight of the evidence has not seemed to affect the authoritativeness with which belief in Ram is accepted by Hindu devotees. Letters to the editor in various national dailies during the period of the Ayodhya crisis suggest that the immediate priority of building the temple had receded into the background ("Ram lives in the heart, not in temples" is the line one encounters most often), but not so the firmness and solidity of belief in Ram. One writer told the editor of the Madras-based The Hindu, "I am prepared to accept the declaration of historians that Ram probably never existed, but that will not stop my believing in him nonetheless."[21] If this can be taken to be a typical response, then the very debate over Ayodhya -- the historicity of Ram, the presence of a mosque on the site of a Hindu temple, and the instances of iconoclasm that accompanied Islamic conversions -- is narrowly concentrated on the verification of facts that in reality have little or nothing to do with the actual problem. That problem is how modern Indian secularism can accommodate and absorb the reality of religion and the power of religious conviction experienced by believers, while at the same time protect the rights of those who believe differently.

Ashis Nandy's contention that Indian secularism has exhausted itself and failed to offer a potent alternative to the rising tide of violence in Indian politics and religion has been construed by some of his critics as a reactionary, anti-secularist argument. What I take Nandy to mean, however, is that Indian secularism has taken the form of the very thing it opposes in principle -- religious intolerance -- and allowed for further divisions between religious ideology and everyday practices of religious belief. If, as Nandy contends, one of the trends in recent South Asian history is the splitting of religion into faith and ideology -- faith defined as a way of life which is "operationally plural and nonmonolithic"[22] and ideology as organized religion which is identifiable with a set body of texts -- the modern Indian state has chosen to define its secular character more in reaction against religious ideology than in relation to religious belief. As an example of how this tendency translates into policy, the regulation of excesses of religious ideology that might threaten the national interest has become an acquired function of the modern state, which is authorized to act as the ultimate arbiter for religious disputes. The reality of individual belief cannot be dealt with as an autonomous reality by the machinery of state, because the secular nation recognizes only the social component of religion -- its hierarchic structures and organizational features. Hence, the state cannot engage with the individual; it can respond only to the material and symbolic orderings of religion as a social institution.

Are the possibilities of religious communication thus foreclosed in the modern secular state, especially if part of the imported baggage of the state is an ingrained skepticism toward personal conviction? The urgent challenge to Hindu revivalism made by secular historians reveals the latter's own difficulties in dealing with religion as a heterogeneous belief system irreducible to mere ideology. If nationalism can be defined as the total set of representational practices that establish the grounds of nationality,[23] then terms like "cultural nationalism" or "religious nationalism" already assume a seamless unity of aspirations, goals, and agendas, a selection and filtering that irons out the contradictions embedded in the processual construction of national identity from the fragments of religious, racial, cultural, and other forms of self-identification. Peter Van der Veer cautions us against this totalizing approach and urges that "we should take religious discourse and practice as constitutive of changing social identities, rather than treating them as ideological smoke screens that hide the real clash of material interests and social classes."[24] However forcefully allegories of the nation, constituting the history of modern secularism, might draw attention to the teleology of its own formation (and by this I refer specifically to the triumphalist rhetoric of rights and citizenship on the model of liberal principle), the narratives produced in the crucial space of negotiation between national and religious identity yield the most visible light on the strains and stresses in community self-identification, especially when community or individual self-perceptions conflict with the definitions accorded them by the nation-state.

Perhaps, as David Krieger suggests in a recent essay,[25] if ideology and faith as polarized terms are replaced by a notion of "cultural metanarratives" at work in non-monolithic, pluralistic societies, it would be easier to conceptualize -- and revitalize -- possibilities for the attainment of a pragmatics of discourse where the meaning rather than the validity of truth claims is foregrounded -- in other words, a discourse that presses to the very limits the contestations of religious ideology, to the point where it can be broken down to illuminate areas of personal belief. In Krieger's conceptualization of the problem of communication, every form of knowing can be construed at a level of discourse higher than argumentation, or what he calls the level of a discourse of limits. Such a discourse accepts the possibility of unknowability and preserves an agonistic concept of "truth" in a pluralistic context, "where discontinuity upon the level of limit-discourse is an inescapable fact."[26] An instance of such discontinuity is the shading of ideology into faith, which is effected by what Krieger calls a "methodological conversion." Krieger draws heavily on religious conversion as a metaphor for a theory of knowledge to suggest the conceptual means by which the gaps between different cultural metanarratives might be bridged. The cognitive rejection of one narrative through the conative acceptance of another is a conceptual analogue to the displacement of religious ideology by faith. "Such a conception is necessary," Krieger writes, "to deal with the problem of how global thinking -- the general validity of knowledge, the universality of norms and a more than merely local solidarity with fellow humans and with nature -- is possible in a radically pluralistic world and a postmodern context."[27]

Strictly at the level of argumentation, an impasse can be overcome only when the ideological premises of the parties to a dispute happen to be similar. If they are not, even the same sets of facts, including those which are entirely non-controversial, can yield totally different conclusions, as is all too apparent in the Ayodhya debate where Hindu and Muslim activists derived completely contrary conclusions from virtually the same evidence. That these facts have a different meaning within the different paradigms involved in the dispute -- mythological and historical -- is further complicated by the absence of a metalanguage to negotiate the conflicting paradigms. The only form of negotiation that is possible is one that entails a transition, or a conversion, from one paradigm to another. Indeed, the process of transition between worldviews emerges, in contexts of pluralism, as the only credible form of negotiation. If discourse beyond the level of argumentation is to materialize, it cannot be grounded in a unitary worldview or religion, but rather in the ability to move between worldviews.

If a higher level of discourse is to be made possible, certain pragmatic conditions for communication would necessarily have to be in place. For one thing, the clash of metanarratives cannot be resolved in terms of the pursuit of knowledge, which is how it had been approached in the Ayodhya debate right up until the time of the mosque's destruction. "True" knowledge is conferred an authority that is belied by the resulting intransigence of both parties in the dispute. The search for knowledge is unavoidably bound up with a struggle for social and political power. For Muslim believers and secular Indians alike, the general fear of losing power to Hindu revivalists had certainly raised the stakes for "proving" the nonexistence of a Hindu temple at the mosque site, to the point that unraveling the truth about Ayodhya had become tantamount to a struggle for hegemonic dominance.

At the end of a provocative and learned essay on perceptions of Islamic conversions by vastly different groups in South Asia -- medieval historians (both Hindu and Muslim), European commentators, and modern scholars -- Peter Hardy speculates:

whether the hypotheses of modern commentators and scholars are themselves essays in conversion, albeit not wholly conscious or deliberate ones: the conversion of agents of the East India Company or of the Crown to particular conceptions of their interests and their duties in India; the conversion of South Asian Muslims to particular conceptions of their future relationships with each other and with non-Muslims; or, to look into an area of inquiry not here entered, namely that of Hindus' interpretations of conversion to Islam, the conversion of Hindus to particular conceptions of their future relationships with Muslims.[28]

The power of conversion as an epistemological concept is that it reclaims religious belief from the realm of intuitive (non-rational) action to the realm of conscious knowing and relational activity. What I hope will emerge in future discussions, even if tentatively, is an examination of that unexplored area pointed to by Hardy -- not just Hindu interpretations of Islamic conversion, but more importantly, the reorientation (or conversion) of Hindus to ways of relating with the Muslim community in India. Conceived in these relational terms, conversion is defined not as a renunciation of an aspect of oneself (as it is in the personal or confessional narrative form), but as an intersubjective, transitional, and transactional mode of negotiation between two otherwise irreconcilable worldviews.[29]

Previous Next Up Comments


1. Peter Van der Veer, "Ayodhya and Somnath: Eternal Shrines, Contested Histories," Social Research 59.1 (Spring 1992): 96.

2. Vidya Dehejia, "Shaivite and Vaishnavite Art: Pointers to Sectarian Tensions?" Unpubl. paper.

3. On this point, Peter Van Der Veer's anthropological fieldwork in Surat offers illuminating insights: he argues, for instance, that the discourse of tolerance and communal harmony is related to the eclipse of the themes of Hindu participation and the influence of Hinduism from the debate about Sufi ritual. See Van der Veer, Religious Nationalism: Hindus and Muslims in India (Berkeley: U of California P, 1994) 33-43.

4. Peter Hardy, "Modern European and Muslim Explanations of Conversion to Islam in South Asia: A Preliminary Survey of the Literature," Conversion to Islam, ed. Nehemia Levtzion (New York: Holmes, 1979).

5. The question of what produced changes in the strength of any religion was settled by reference to three causes: the reproductive power of a religion's adherents, migration, and conversion. By the 1890s Muslims had grown twice as rapidly as Hindus, and the census asks the question: "How far is this due to the conversion of Hindus and how far to the greater fecundity of Muslims?" (E. Gait, The Lower Provinces of Bengal and Their Feudatories, Census of India 1901, 6.1, Report [Calcutta, 1902] 156. Henceforth abbr. Census of India, 1901.)

6. H. Beverley, Report of the Census of Bengal, 1872 (Calcutta, 1872), pars. 348-354. E.g.: "The real explanation of the immense preponderance of the Musalman religious element in this portion of the delta is to be found in the conversion to Islam of the numerous low castes which occupied it. . . . If further proof were wanted of the position that the Musalmans of the Bengal delta owe their origin to conversion rather than to the introduction of foreign blood, it seems to be afforded in the close resemblance between them and their fellow-countrymen who were still from the low castes of Hindus. That both are originally of the same race seems sufficiently clear, not merely from their possessing identically the same physique, but from the similarity of the manners and customs which characterise them."

7. Khondkar Fazli Rabi, The Origins of the Musalmans of Bengal (1895; Dacca: Soc. for Pakistan Studies, 1970) 43. See also Rafiuddin Ahmed, The Bengal Muslims 1871-1906: A Quest for Identity (Delhi: Oxford UP, 1981) for a complex exploration of the construction of Muslim identity. Ahmed argues that a dominant feature of the nineteenth-century campaigns of Islamization in Bengal was the attempted rejection of virtually all that was Bengali in the life of a Muslim as something "incompatible with the ideas and principles of Islam"(106).

8. Census of India 1901 166.

9. Census of India 1901 166. Rafiuddin Ahmed maintains that the Muslim community's claims to family names and alien origins, by way of removing the stigma of their local descent, "were helped by certain government measures like census classification"(Ahmed 184). While it is true that the census did elicit the names by which Muslims called themselves, this should not be taken to mean that it accepted the foreign origins that those names connoted. On the contrary, it often contested their authenticity, incredulously dismissing, for instance, the number of self-proclaimed "Shekhs" as being more than twenty times the estimated population of "Arabia" at that time.

10. H. H. Risley, The Tribes and Castes of Bengal, 2 vols. (Calcutta: Bengal Secretariat P, 1891) 1: xxii-xxxvii.

11. Census of India 1901 167. Emphasis added.

12. The ritual of shuddhi contributed greatly to the increase of Hindu-Muslim antagonism. For many Muslims, the infamous pork-test of the Shuddhi Sabha was taken as the ultimate insult to their religious adherence. But there were many communities and individuals who manifested dual types of behavior, and they were targeted as ripe candidates for shuddhi. For instance, the religious status of the Malkanas, in the western part of what was then called the United Provinces, was a confused one. Their culture showed the influence of Islam, even to the point of using Muslim functionaries in some of their ceremonies. At the same time they retained many Hindu practices. However, in the census they tended to declare themselves Muslims. Several unsuccessful attempts to reconvert them had been made between 1907 and 1910, but as J.T.F. Jordens points out, "the decisive break-through came in 1922 when the Hindu Rajputs in their Kshatriya Upkarini Sabha passed a resolution in support of receiving the Malkanas, and permitting them to be reunited with the Rajput Hindu brotherhood after purification"(158). J.T.F. Jordens, "Reconversion to Hinduism, the Shuddhi of the Arya Samaj," Religion in South Asia, ed. Geoffrey Oddie (Delhi: Manohar, 1982).

13. This argument is elaborated elsewhere in my "Coping with (Civil) Death: The Christian Convert's Rights of Passage in Colonial India," After Colonialism: Imperial Histories and Postcolonial Displacements, ed. Gyan Prakash (Princeton: Princeton UP, 1994).

14. Census of India 1901 166.

15. App. II, "Extracts from District Reports regarding Causes of Conversion to Muhammadism," Census of India, 1901x-.xix.

16. Ahmed 184.

17. Cf. Census of India, 1901, which describes Hinduism as "not so much a form of religious belief as a social organization. . . A man's faith does not greatly matter so long as he recognizes the supremacy of the Brahmans and observes the restrictions of the Hindu caste system"(152). Bernard S. Cohn has powerfully shown how the British system of objectification through census-taking hinged on caste and religion as crucial sociological keys to understanding Indian society and Indian people. Cohn maintains that "ideas about caste -- its origins and functions -- played much the same role in shaping policy in the latter half of the nineteenth century that ideas about the village community and the nature of property played in the first half of the nineteenth century"(243). In the hands of an ethnographer like Herbert Risley, who wielded anthropometric instruments as if they were weapons of war, the caste system fed into theories of racial purity and social hierarchy. See Bernard Cohn, "The Census, Social Structure and Objectification in South Asia," An Anthropologist among the Historians and Other Essays (Delhi: Oxford UP, 1987).

18. Census of India, 1901 172.

19. Census of India, 1901 166.

20. While S.A.A. Rizvi states that Muslim commentators usually give an "altogether exaggerated account of proselytisation," claiming pride in Islam for winning scores of Hindu followers (17), Peter Hardy suggests a more ambivalent reading. Hardy proposes that while it is true that there was a certain amount of exaggerated self-glorification among recorders of Muslim history, Muslim historians were less interested in showing how Islam expanded through force and chance than through the missionary zeal of sufis and pirs. See S.A.A. Rizvi, "Islamic Proselytisation, Seventh to Sixteenth Centuries," Religion in South Asia, and Hardy, "Modern European."

21. "Ayodhya Temple," letter, The Hindu, 10 Dec. 1990.

22. Ashis Nandy, "The Politics of Secularism and the Recovery of Religious Tolerance," Mirrors of Violence, ed. Veena Das (Delhi: Oxford UP, 1990).

23. Simon During, for instance, provides a useful working definition of nationalism as "the battery of discursive and representational practices which define, legitimate, or valorize a specific nation-state or individuals as members of a nation-state." See Simon During, "Literature -- Nationalism's Other? The Case for Revision," Nation and Narration, ed. Homi Bhabha (London: Routledge, 1990) 138.

24. Van der Veer, Religious Nationalism ix.

25. David J. Krieger, "Conversion: On the Possibility of Global Thinking in an Age of Particularism," Journal of the American Academy of Religion 58.2 (1990): 223-243. See also Alan M. Olson, "Postmodernity and Faith," Journal of the American Academy of Religion 58.1 (1990): 37-53. See also Michael C. Banner, The Justification of Science and the Rationality of Religious Belief (Oxford: Oxford UP, 1990) for an illuminating analysis of the problematic opposition between the "rationality" of science and the "irrationality" of religious belief and the circularity of paradigm conflicts to which this gives rise.

26. Krieger 227.

27. Krieger 223.

28. Hardy 99.

29. In their challenge to conceptions of conversion as forcible and radical change, recent advances in the scholarship stress the relational features of conversion, though not always critically or with a view to examining the grounds of relationality. An historiography based on notions of transition rather than change is interested in recovering a pragmatics of intersubjective communication. The work of scholars like Robin Horton, Susan Bayley, Derryck Shreuder, and Geoff Oddie has aimed to supplant the contestational features of conversion with a version that stresses its adaptive tendencies. For instance, in her work on Muslims and Christians in South Indian society, Saints, Goddesses and Kings (Cambridge: Cambridge UP, 1989), Susan Bayley challenges as misleading the view that conversion is a radical movement from one religion to another, resulting in the total repudiation of one for the other. Rather, she emphasizes the fluidity of the original religion which allowed for the "conversion" of its individuals to other religions. Bayley, echoing Robin Horton, contends that conversion is not as transgressive or disruptive of the norms of a society as generally maintained. Applying Horton's theories of African conversion to the Indian context, Bayley argues that conversion is not simply a shift of individual conviction or communal affiliation: if Indians have embraced another religion, it is less so because they believed in some egalitarian message that the new religion had to offer, but because they invested in an alien religious being new divine power associated with the old. See Robin Horton, "African Conversion," Africa: Journal of the International African Institute 41.2 (1971): 85-108; and Deryck Shreuder and Geoffrey Oddie, "What is 'Conversion'? History, Christianity and Religious Change in Colonial Africa and South Asia," Journal of Religious History 15.4 (1989): 496-518.