By the second half of the nineteenth century, the colonial state in India was about to undergo several major transformations. Whereas the revenue and authority that accrued from the relationship between land and the state were fundamental to the formation of the early colonial state, the general agrarian revolts that followed hard on the heels of the 1857 “Mutiny” and the steadily increasing economic investment in imperial power (propelled in particular by the joint stock arrangement of the railways and other infrastructural projects) made it clear that things had to change. Land tax was still an important source of revenue through the century, as was much of the trade that had been fundamental to the mercantile origins of empire. It became clear, nevertheless, that the extractive colonial state was facing other challenges that required a new kind of imperium; accordingly, imperial ambition, and anxiety, moved to different levels and concerns. The steady absorption of new lands through the aggressive policies of Lord Dalhousie—that in the taking of Awadh in 1856 had led directly to the Great Rebellion—was brought abruptly to a halt, and policies of indirect rule were mobilized to accommodate, and ultimately appropriate, the incomplete project of colonial conquest. At the same time, the rebellion made it clear that some communities in India could be counted as loyal, whereas others became doomed to perpetual suspicion. These latter groups were to be replaced in the armed forces by the “martial” races; Macaulay’s hyperbole was translated into state policy. In the new rhetorical economy of colonial rule, political loyalty replaced landed status, and the form of knowledge and argument that seemed most appropriate to assess matters of loyalty rather than revenue was, of course, knowledge of peoples and cultures. To put the matter in bold relief, after 1857, anthropology supplanted history as the principal colonial modality of knowledge and rule. By the late nineteenth century, as I will go on to show, the colonial state in India can be characterized as the ethnographic state.
In the wake of the Great Rebellion, and as British rule became increasingly secure, caste began to spin a career of its own. Whereas in the early part of the century India’s feudal past and then its village communities seemed far more important than the caste system, the colonial ethnographic curiosity that flowered, especially from 1870 on, took caste as the primary object of social classification and understanding. The relative silence about caste matters in early official writings as well as in collections of local texts such as Mackenzie’s gave way to new kinds of compendia—from miscellaneous collections and volumes, official manuals and gazetteers, to the census—in which caste figured as the most important subject and classificatory schema for the organization of India’s social world. With the memory of the “Mutiny” still lively, concerns about revenue gave way to a preoccupation with social order and the maintenance of rule. To keep India, the British felt the need to know India far better than they had, and now the knowledge had to be about the society of India, not just its political economy. Thus colonial ethnology took the place that had once been held by colonial history. The recognition on the part of Bentinck and Elphinstone that certain forms of knowledge were inaccessible to the British was replaced by the obsessional conceits of high imperial ambition (and anxiety). The ethnographic state was driven by the belief that India could be ruled using anthropological knowledge to understand and control its subjects, and to represent and legitimate its own mission. By the late nineteenth century, ethnological knowledge became privileged more than any other form of imperial understanding, much as Kipling suggests when he made Colonel Creighton both head of the Ethnological Survey and a master spy in the Great Game in his classic colonial novel, Kim.
The more the British believed they could know India—with that peculiar colonial intransitivity that made it possible for them to think the more they knew the less the native could know them in turn—the more, of course, they doubted their knowledge. The confidence of imperial empiricism was always accompanied by a sense that facts spin out of control. As Colonel Creighton, Kipling’s head of the Ethnological Survey, says, “The more one knows about natives the less can one say what they will or won’t do.”1 Indeed, the more H. H. Risley, census commissioner and superintendent of the Ethnographic Survey, went on to refine caste categories to allow the enumeration of the entire population of India by caste, the more it seemed that caste categories were overlapping, unstable, and contested. But such difficulties never made Risley, or the British more generally, doubt their fundamental methods, let alone their ultimate capacity to know. The problem was that knowledge of India was always put in terms of alterity, with unknowability a natural implication of the language of difference. Between difference and the perpetual possibility of deception, knowledge was invariably uncertain. The empiricist response was always to know more, even as the British could never acknowledge the deep uncertainty about the possibility of real knowledge about subjects increasingly cast in terms of incommensurability. The flip side to imperial empiricism was the sense of ultimate inscrutability. And the story that follows charts the murky waters created as a result.2
Throughout the nineteenth century, the collection of material about castes and tribes and their customs, and the specification of what kinds of customs, kinship behaviors, ritual forms, and so on, were appropriate and necessary for ethnographic description, became increasingly formalized and canonic. Gradually the institutional provenance of caste expanded, affecting the recruitment of soldiers into the army (particularly after the Great Rebellion), the implementation of legal codes that made the provisions of the law applicable on caste lines, the criminalization of entire caste groups for local policing purposes, the curtailment of the freedom of the land market when excessive amounts of land were thought to be sold by “agricultural” to “merchant” groups, and the assessment of the political implications of different colonial policies in the area of local administration in caste terms—to mention only a few examples.3 In the years after the rebellion, the detailed compilation of empirical material on British India escalated dramatically, first in occasional manuals of local districts, such as J. H. Nelson’s The Madura Country,4 then in the gazetteers and statistical surveys that proliferated as the ethnographic state gained momentum. Early manuals were faulted for being prolix and insufficiently statistical. This concern was expressed throughout the century, for example in response to Buchanan’s account of his journeys through Mysore and Kerala in the years just after the 1799 defeat of Tipu Sultan. When some years later Buchanan was commissioned to engage in a similar survey of the Bengal districts, he was instructed to reduce the narrative sections of the account, and to collect more statistics, to make his findings both more readily digestible and comparable. Similarly, when Nelson submitted his manual of the Madura country to the Madras government, he received a lukewarm response: “The board [of Revenue] are of opinion that much greater brevity is desirable, a larger proportion of facts as compared with arguments and theories, and more copious references to the sources whence further details on the various points treated of may be obtained.”5 The board noted in general terms that although the preparation of district-level manuals would be necessary to provide the background for the preparation of a gazetteer for Madras presidency, they were concerned about the “deficiency in statistics, and the prolixity of details” in a work such as Nelson’s. Nelson’s volume was declared “useless to the general public.”6 This raised special questions for the Madras government because Nelson’s work had been commissioned to serve as an example for other district manuals.
The government was concerned to find some method that could produce useful and uniform knowledge for all of India. “His Excellency in Council declared his conviction that immediate steps should be taken to ensure uniformity, and, to that end, appointed Mr. W. W. Hunter, LL.D., of the Civil Service, to visit the Local Governments, and ascertain what had been done in each, with a view to his drawing up a comprehensive scheme, prescribing the principles for the compilation of the Provincial Gazetteers and for their ultimate consolidation into one work.”7 In 1869, Hunter was accordingly appointed director general of statistics to the government of India, and over the years he produced and supervised a series of gazetteers that sought to systematize official colonial knowledge about India. Each manual and gazetteer had an ethnological chapter, in which the local castes and tribes were listed and described, with more detail reserved for certain caste and tribe groups specific to the area, under the heading of “manners and customs.” Marriage systems and kinship patterns, funeral rituals, adherence to Brahmanical priesthood and principle, clothing, and the geographical distribution of different groups made up the bulk of the descriptive material. By the time of Maclean’s general manual of Madras presidency, published in 1892, the subject matter of ethnology had become standardized: “An ethnological account of an Indian people must consist of not less than five separate subjects; their race or descent, their language, their caste, their religion or sect, and their traditional habits and customs.” Maclean went on to be more specific: “Of these subjects the first [the issue of race] is the most difficult to examine because it is the most involved, and the second [language] is the easiest because it is the most capable of definition and the most accessible. The other three, caste, religion and customs, are little more than matters of observation; but on the other hand they are very imperfect elements in anthropological inquiry, caste probably taking precedence in India among the three.”8 By this time, the ethnological section had become more important than the historical prologue, both preceeding it and exceeding it greatly in length.
One of the first general compilations of material on caste after the rebellion was assembled by the Reverend M. A. Sherring, who, starting in 1872, published an influential three-volume work entitled Hindu Tribes and Castes. The work aims to be the first text to “give in English a consecutive and detailed account of the castes of India,” and finds it strange that no one had hitherto attempted to do so.9 It is encyclopedic in coverage, starting with Brahmans then moving to Ksatriyas and on “down” the varna scale. But unlike earlier colonial works that relied on textually derived varna categories as a general guide about Indian society, Sherring used these categories as the frame for his attempt to marry textual and empirical knowledge. The footnotes refer to district manuals, James Tod’s three-volume work on Rajasthan, early settlement reports, and Wilson’s glossary, among many other sources. Manu’s text is hardly abandoned; it helps to orient the chapters on Brahmans and the generation of myriad actual castes, and provides a dharmic explanation for multiplicity because of intermarriage. But this becomes in part a pretext for a strictly biological theory of origins for caste groups that are seen now to hold to endogamous principles without breach: “it would be . . . correct to regard the numerous Indian tribes and castes as so many distinct integers, complete in themselves, independent and unassociated . . . the honourable condition of marriage between separate castes, and to a large extent between branches of the same caste, is absolutely prohibited.”10 The tone of Sherring’s work is more empirical than textual, and provides a bridge between early Orientalist forms of knowledge and a new kind of empirical quest. The text seems a perfect prelude for the census, the first all-India version of which was conducted the year of its publication.
Sherring’s text betrayed not only a new empiricist tone but also the author’s Christian affiliation.11 It is striking that a certain kind of Christian position could now accommodate official views of caste, rather than simply being at war with them. By this time Sherring had already accepted the impossibility of mass conversion across India, and even he wrote with approval of Hindu social reformers who had clearly refused to abandon their own religious convictions. But his condemnation of caste, and of Brahmans in particular, carries on a long tradition of missionary and evangelical writing. Sherring followed Max Müller in asserting that “caste as now existing was totally unknown to the Hindu race on first entering into India.” He then explained the rise of caste distinction in terms reminiscent of the most frustrated of Christian missionaries. First, caste explains the total religiosity of the Hindu (“the Hindu, from the outset . . . has been engrossed by his religion, which has been at once a magnet to draw him and a pole star to direct him).” Second, caste has made the Hindu servile, and it is “this credulous and servile condition of the Hindu mind [that] has afforded a golden opportunity to the wily Brahman, thirsting for rule and for the exercise of his superior gifts. And third, it is the “wily Brahman” who is especially at fault. The Brahman is not only wily, he is “arrogant and proud,” “selfish,” “tyrannical,” “intractable,” and “ambitious.” Sherring maintained that the rise of caste after the Vedic period was most certainly the result of a Brahman conspiracy. “Caste, therefore, owes its origin to the Brahman. It is his invention.”12
In inventing caste, the Brahman secured his claim to his very being: “The Brahman could not now exist, and could not have existed at all, . . . without having caste as the objective form in which his ideas were realized.” Sherring used a theory of divide and rule, one that, like his theory of invention, would in fact seem far more appropriate for his own tribe than that of the colonized Brahman. He noted that the Brahman was “pleased that all the castes were animated by the spirit of themselves—pleased at the prospect of their own authority and majesty with every increment added to the castes—and pleased above all at the thought that their own order was at the head of the entire system, and exercised command over all its ramifications.” And although caste did not exist at the beginnings of Aryan life in India, the Brahman was able to engineer it in the “childhood of the Hindus.”13 In his concluding essay on the “Prospects of Hindu Caste,” he began his opening series of paragraphs with the following assertions: “Caste is sworn enemy to human happiness”; “Caste is opposed to intellectual freedom”; “Caste sets its face sternly against progress”; “Caste makes no compromises”; “The ties of caste are stronger than those of religion”; and “Caste is intensely selfish.”14 Sherring held, however, that the greatest force of change in India was not missionizing but the actions of certain “caste-emancipated Bengalees.” Although he was ambivalent about the social reform movement in Bengal (unsure as he was about the adoption of Western customs across the board—he observed that in “our judgment, it is far better for natives of India to adhere to their own customs than to adopt those of foreigners”), he believed that a great change for the better has commenced with the the first steps of “some of the foremost thinkers and actors in Bengalee society.” And although he was uncharacteristically optimistic about indigenous transformations in India, he encouraged the formation of serious anti-Brahman sentiment: “let them suddenly awake to the thought that they are as well educated, as able, as intelligent as the Brahmans . . . and exert a much deeper and a far better influence over Hindu Society at large; and they can, if they be so inclined, destroy Brahmanism, root and branch—can utterly annihilate it.”15 Although Sherring’s text reiterated earlier missionary condemnation of Brahmans, it accepted the salient reality of racial and cultural difference, which for the missionary always meant the general impossibility of conversion.
Such explicit editorializing did not feature in the census, which from 1872 took over the authoritative function of producing empirical information on caste from the writings of missionaries and officials alike. By the time of the census, caste had become especially important both because of strongly held official views and because of the ways caste was confronted by a variety of Indian reformers, activists, and intellectuals as emblematic of Indian society. The spirit of caste attained its apotheosis with the census. A vehicle for the consolidation of imperial ideology, the census became the means for the collection of empirical knowledge the likes of which could not have been imagined by previous commentators. And yet even the extraordinary compilations of the census were ultimately seen as inadequate to the task of colonial ethnology. The inauguration of the ethnographic survey of India, announced in the first issue of Man in 1901, made clear that the census was necessary for colonial knowledge, but hardly sufficient for colonial rule:
It is unnecessary to dwell at length upon the obvious advantages to many branches of the administration in this country of an accurate and well-arranged record of the customs and the domestic and social relations of the various castes and tribes. The entire framework of native life in India is made up of groups of this kind, and the status and conduct of individuals are largely determined by the rules of the group to which they belong. For the purposes of legislation, of judicial procedure, of famine relief, of sanitation and dealings with epidemic disease, and of almost every form of executive action, an ethnographic survey of India, and a record of the customs of the people is as necessary an incident of good administration as a cadastral survey of the land and a record of the rights of its tenants. The census provides the necessary statistics; it remains to bring out and interpret the facts which lie behind the statistics.
Thus at the start of the twentieth century the political centrality of caste was formally announced. Caste was the site for detailing a record of the customs of the people, the locus of all important information about Indian society. This information, which the colonial state felt increasingly compelled to collect, organize, and disseminate, would thus become available for a wide variety of governmental initiatives and activities, and would relate to “almost every form of executive action.”
If the ethnographic survey announced the preeminence of caste for colonial sociology, it was the decennial census that played the most important institutional role not only in providing the “facts” but also in installing caste as the fundamental unit of India’s social structure. There was general agreement among most of the administrators of the census, which began on an all-India basis in 1871/2, that caste should be the basic category used to organize the population counts. But there was far less agreement about what caste really was. Various commissioners, for instance, debated whether a caste with fewer than 100,000 persons should be included, or how to organize the “vague and indefinite” entries that in 1891 exceeded 2,300,000 names. There were also debates about whether, and if so how, to list the castes on the basis of “social precedence.” When Risley adopted a procedure to establish precedence in the 1901 census, caste became politicized all over again. Caste associations sprung up to contest their assigned position in the official hierarchy, holding meetings, writing petitions, and organizing protests. By 1931 some caste groups were distributing handbills to their fellow caste members to tell them how to answer questions about their religious and sectarian affiliations, as also their race, language, and caste status. After 1931, the British could no longer ignore the political fallout of the census, and abandoned the use of caste for census counting altogether.
The rise of caste as the single most important trope for Indian society, and the complicity of Indian anthropology in the project of colonial state formation, are documented in a great many texts, perhaps nowhere more fully than in H. H. Risley’s classic work, The People of India.16 Risley, who was the census commissioner of India for the 1901 census (the regulations of which greatly influenced the 1911 census, as well), had earlier produced the multi-volume work The Tribes and Castes of Bengal, published in 1891.17 The People of India resulted directly from Risley’s work as census commissioner, and is an expanded version of the commissioner’s report on the 1901 census (written with the assistance of E. A. Gait) that, among other things, summarized his views on the origin and classification of the Indian races based on his historical speculations and his anthropometric research. Risley was criticized by contemporary as well as subsequent writers for overemphasizing the racial basis of caste and stressing anthropometry. William Crooke, author of many ethnographic works and perhaps his most important critic, argued against Risley with particular vehemence, suggesting that occupational criteria provided much more comprehensive and accurate indices for understanding caste as a system than did race.18 The anthropometric researches of subsequent scholars steadily eroded the confidence of the anthropological establishment that racial types in India were anywhere near as pure or clear as Risley had assumed. But Risley’s general views of caste as a social system and force in India were little challenged. Risley seemed to speak for many in both colonial and academic establishments when he wrote that caste “forms the cement that holds together the myriad units of Indian society. . . . Were its cohesive power withdrawn or its essential ties relaxed, it is difficult to form any idea of the probable consequences. Such a change would be more than a revolution; it would resemble the withdrawal of some elemental force like gravitation or molecular attraction. Order would vanish and chaos would supervene.”19 At the dawn of the twentieth century, it would be difficult to put the case much more strongly than that.
In the proposal for the ethnographic survey of India, it was observed that anthropometry was a science that would yield particularly good results in India precisely because of a caste system that organized social relations through the principle of absolute endogamy: “Marriage takes place only within a limited circle; the disturbing element of crossing is to a great extent excluded; and the differences of physical type, which measurement is intended to establish, are more marked and more persistent than anywhere else in the world.”20 Thus the government justified its project, and its choice of Risley, for a survey that was specifically directed “to collect the physical measurements of selected castes and tribes.” Risley’s advocacy of anthropometry and his theories about the relation of race and caste were clearly fundamental to the definition of the ethnographic project in turn-of-the-century colonial India. The scientific claim about caste reflects Risley’s assumption that he could actually test in India the various theories about race and the human species that had been merely proposed on speculative grounds in Europe. At the same time, these claims concealed the continuity between the assumption that castes were biologically discrete and his belief that in cultural as well as biological terms castes in India were like individuals in the West.
The ethnographic survey resulted in a series of volumes organized around the encyclopedic delineation of the customs, manners, and measurements of the castes and tribes of the different regions of India, and although not all the surveyors shared Risley’s anthropological views entirely, the volumes nevertheless reflect Risley’s general sense of what the survey should entail. Each entry includes such salient ethnographic facts as caste origin stories, occupational profiles, descriptions of kinship structure, marriage and funerary rituals, manner of dress and decoration, as well as assorted stories, observations, and accounts about each group. The texts were obviously designed as easy reference works for colonial administrators, for the police as well as revenue agents, district magistrates, and army recruiters. It was clear you could know a man by his caste.
Although colonial ethnographers rarely addressed directly the political implications of their scientific projects, Risley did precisely that in his The People of India, where he confronted the question of nationalism. In one of the two new chapters written for the 1909 publication of the book, Risley assessed the role caste might play in the future of India’s political development. And he quoted with approval the words of Sir Henry Cotton, who surmised that “The problem of the future is not to destroy caste, but to modify it, to preserve its distinctive conceptions, and to gradually place them upon a social instead of a supernatural basis.”21 Here Cotton, and Risley, advocated precisely what I have suggested colonialism in India set out to do: to reconstitute caste as a necessary complement to social order and governmental authority, and to formulate it as a new kind of civil society for the colonial state.
In Risley’s view, caste has an ambivalent status. It is both a religious institution and a social or civil one. It is anarchic, yet encourages the development of monarchy. It is particularistic, though the necessary and inevitable basis for any unity in the Indian context. On the one hand Risley noted, basing his conclusions largely on the lectures of Sir John Seeley, that “The facts are beyond dispute, and they point to the inevitable conclusion that national sentiment in India can derive no encouragement from the study of Indian history.” On the other hand, Risley wrote that “the caste system itself, with its singularly perfect communal organization, is a machinery admirably fitted for the diffusion of new ideas; that castes may in course of time group themselves into classes representing the different strata of society; and that India may thus attain, by the agency of these indigenous corporations, the results which have been arrived at elsewhere through the fusion of individual types.” These contradictions are interestingly resolved in (and by) the colonial situation. And here we confront the colonial mind in its most liberal guise. For Risley wrote that “The factors of nationality in India are two—the common use of the English language for certain purposes and the common employment of Indians in English administration.”22
Risley thus held out a kind of limited but realistic hope for national development in India, measured by his sense that caste ideas and institutions would stand in the way, though optimistic that a steady (and English) pragmatism on the part of Indian leaders could sow the seeds of a new mentality. But of course his liberalism was entirely complicit in the general project of British colonialism, since it supported the idea that caste was simultaneously a barrier to national development and an inevitable reality for Indian society in the foreseeable future. Furthermore, he suggested that caste, as he had interpreted it, could be made into a virtue out of its necessity. It could accommodate and shape a gradually developing class society, perhaps even softening its potential conflicts and antagonisms, and provide a model (in its idealized varna version) for the articulation of an all-embracing ideology that might work at a general level to confound and even counteract the fissiparous tendencies of caste as a specific social institution. Caste in this sense was the key to the great transition from feudalism to capitalism/democracy—except that in the colonial situation that transition could never fully be made; the teleology of self-rule was here, as always, couched in a future that had no temporal reality.
And so once again caste expressed the fundamental nature of India, and opposed itself to the possibility of history, past or future. At the same time that Risley’s work united official and academic knowledge, it also revealed the extent to which basic understandings of caste, and of India more generally, were tied to colonial assumptions about the absence of politics and the overpowering and yet divisive force of caste as a social principle. And it will become clear that this view of both caste and India has continued to the present day to be pervasive in Western academic views of India, long after the end of empire.
G.F.W. Hegel was not the first to suggest that the absence of history in India was a consequence of the natural force of the caste system, but he was among the most influential. Curiously, Hegel viewed India as advanced beyond China, on his civilizational scale, in part because of the break between state and society produced by caste. “The different castes are indeed, fixed; but in view of the religious doctrine that established them, they wear the aspect of natural distinctions.”23 For Hegel this meant that in India the whole of society was not absorbed into the despotism of the ruler, as was the case in China. However, the admirable inauguration of a separation between the religious and the secular is distorted by the fact that caste is ultimately a religious principle. Caste resists not just despotism but also any meaningful exchange between social and political developments. For Hegel, as a result, caste fails to establish a relationship with history, and India remains plunged in a dreamlike state that necessitates its subjection to Europe.24
If Hegel thus confirmed the British colonial sense that India had no history, and that the entire force of the caste system was to hinder political development, he also naturalized the establishment of a sociological view of India that went well beyond the specific claims of colonial representation. Marx’s writings on India, both in his journalistic coverage of the Great Rebellion and in his development of a theory of the mode of Asiatic production, also drew on colonial sociology (though more with respect to the village community and an associated view of despotism than with respect to caste); this was true even when Marx was critical of British imperial rule. Max Weber also ratified colonial views in his comparative sociology, which led him to write long treatises on India and China to accompany his extraordinary work on economy, society, and values in the West. Weber not only saw caste as the fundamental institution of Hinduism, organized in relation to and by the Brahman, he accepted the categories (and even the lists of social precedence) of the 1901 census as the empirical basis for his own attempt to correlate caste to religious orders, status groups, kinship categories, and economic units (such as guilds).
Classical social theory both reiterated the conventional wisdom of colonial ethnography and influenced later colonial ethnographers, one of the last of whom was J. H. Hutton, who returned to Britain after his stint as census commissioner in 1931 to assume a professorship in anthropology at Cambridge.25 Although he strayed well away from both the racial and textual views of Risley, having quarreled with Risley’s legacy concerning the use of caste in the census when he directed it in 1931, he still drew his principal material not just from colonial sources and preoccupations but from the census itself. Perhaps most importantly, however, he served as a direct bridge between colonial officialdom and academic certification, directly calling upon his colonial experience as the basis for his anthropological expertise. In fact, academic anthropology, at least in Britain, had been preoccupied not with the peasant societies represented by caste but rather with the idea of primitive society and with the islands and hill regions that were often on the borders of imperial rule. Hutton, who also had the acuity to draw upon a growing interest in French anthropology concerning the nature of the caste system (viz. Émile Senart and Célestin Bouglé), worked to change all that, and to train a new generation of anthropologists in Britain who sought to unite the study of primitive and peasant.
The effects of World War II and the decolonization it brought in its wake had an even greater impact on the world of academic knowledge. Perhaps most important, the United States academy took charge in the years after the war, fortified by massive investment from government and private foundations in an interest to understand, and control, the processes of modernization that colonialism and decolonization had unleashed in the third world. A new empiricist social science felt itself freed from the shackles of a colonial past, and combined positivist method with a vaguely developmentalist agenda. Modernization was viewed as a natural process rather than one linked to political projects and interests in ways that might still bear the traces of imperial ideologies. And yet it is now clear how postwar America assumed a new imperial role, displacing earlier imperial regimes both by its own political position and by its genealogical relationship to these regimes through theories that used cultural theory to excuse colonial history as much as to disguise American hegemony.
Village India, a volume of essays edited by the young American anthropologist McKim Marriott, was the herald of a new anthropology of South Asia.26 A product of the new American interest in “underdeveloped” societies (driven, and funded, in part by the imperatives of the cold war), the turn of anthropological interest to peasant societies was dramatically linked to modernization theories that operated as the ideological charter of a new American claim for postimperial domination.27 Spawned in large part under the influence of Robert Redfield, whose concern was to locate anthropology around the study of little communities in complex societies and in relation to civilizational processes, Village India made a strong case for the fundamental relationship between caste as a civilizational idea and the village. Beyond the particular theoretical imperatives of the project, the volume exemplified a more general trend of the 1950s, when American anthropologists joined British and Indian social scientists in villages across the subcontinent to chart the social organization of the primary unit of India, namely, the village itself. Some American social scientists sought to see caste as one particular instance of social stratification in a comparative context that linked racial discrimination in the United States with caste prejudice in India. Others sought to document in rich detail the way actual relations of intermarriage and interdining determined each local manifestation of caste. But as language study in the United States developed to a new level, and area studies initiatives linked the social sciences with the humanities, caste came increasingly to be seen as a marker of something unique in Indian civilization, a sign at once of general tradition and specific alterity. During the 1960s, caste became central once again to the academic study of India. And the impetus for this came not only from internal developments within British, American, and Indian anthropology but also from the monumental intervention of the French anthropologist Louis Dumont, whose major work on caste, Homo Hierarchicus, was published in 1966.28
When I first went to graduate school in the early 1970s I was told that I was joining a group of scholars who were on the verge of a major breakthrough in the understanding of caste.29 Those were heady days in the history of American area studies, when the accumulation of language learning and field experience made it possible for some American academics to believe they could discover authentic Indian categories of social and cultural knowledge.30 Using theoretical prescriptions from Talcott Parsons and David Schneider, two centuries of Indological knowledge, and the accumulation of research experience based on intensive studies by anthropologists during the twenty-five years since Indian independence, social scientists such as McKim Marriott and Ronald Inden at Chicago developed a comprehensive theory of caste society that claimed to reveal the conceptual underpinnings of caste relations, based on the monistic indissolubility of natural substance and moral code. Like James Mill one hundred and fifty years before, they announced their views in the new edition of the Encyclopedia Britannica.31 The appeal of the theory was in part its apparent sensitivity to Indian conceptual categories and in part the accompanying claim to be working against both British colonial understandings and Western social science. In retrospect, the theory and the claim were both seriously flawed; not only was the approach to theory ahistorical, it was based on an idealist notion of knowledge that refused any contamination by the politics of knowledge. It was not so opposed to Western social science, or colonial understandings, after all.32 In recent years, the theoretical impact of postcolonial studies has made all such knowledge claims deeply problematic. And leaving aside the epistemological disrepute into which this neo-Orientalist project of idealist representation would necessarily fall in the wake of Edward Said’s devastating critique, the implication of this kind of area studies in the constitution of a timeless Hindu India available for use by the forces of right-wing Hindu fundamentalism has subsequently sent shudders down many a scholarly spine. When Inden and Marriott wrote in 1976 that they had “found that the presumptions evident in Vedic thought do provide a good basis for formulating Indian unity and for understanding much of South Asian history and ethnography consistently,”33 they wrote in an idiom that has since been taken over by the Hindu right.
The Chicago school of ethnosociology had a certain impact through the many graduate students who imbibed it throughout the 1970s, but it suffered from the abstruse objective of converting social science to a methodological monism, with the ultimate objective (at times mystical, and not a little colonial) of making “the knowers—somewhat like those South Asian objects that they would make known.”34 In any case, it never produced a systematic work of social theory, and as a result failed to overturn the work of its chief interlocutor and critic, Louis Dumont. Dumont, along with his collaborator David Pocock, had announced his initial proposals not in the Encyclopedia Britannica but in a journal he founded that soon became the leading journal of South Asian anthropology, Contributions to Indian Sociology. Like Inden and Marriott later on, Dumont was interested in generating a holistic theory of the caste system that took into account belief as well as action, though it soon became clear how different were the perspectives of French and American social science. But well before most American anthropologists began searching for the clues to Indian society in ancient texts, Dumont proposed, in the inaugural volume of Contributions, that “the first condition for a sound development of a Sociology of India is found in the establishment of a proper relation between it and classical Indology.” Dumont went on to clarify the sociological implications of this methodological assertion: “By putting ourselves in the school of Indology, we learn in the first place never to forget that India is one. The very existence, and influence, of the traditional higher, Sanskritic, civilisation demonstrates without question the unity of India.”35 In the first major theoretical and synthetic work on caste since those of J. H. Hutton and G. S. Ghurye, Dumont provided a theoretically rigorous and all-encompassing theory of the caste system that based its argument on the idea that India was one, across both time and space. The proposals of Contributions anticipated the publication of Homo Hierarchicus in 1966, an ambitious book that was hailed as a major work of theory and insight by anthropologists and Indologists alike, and was immediately installed as the benchmark for debates on Indian society and culture for years to come.36
Edmund Leach, a prominent British social anthropologist who was critical of Dumont’s work, nevertheless wrote that “it is probably the most important work ever published,” on the subject of caste.37 Other important anthropologists also took the book very seriously, for the most part praising its originality and elegance. Stanley Tambiah wrote that Homo Hierarchicus was “a profound contribution to Indian studies . . . perhaps the time was ripe for a man of vision to attempt a telescopic view of the society as a whole.”38 Ravindra Khare wrote that Homo Hierarchicus was “a brilliant piece of intellectual accomplishment . . . that by all odds secures a pioneering place in Indian caste studies with deserved amour propre,” in the context of a long and at times critical review essay.39 Khare noted, as did others, that Dumont was writing primarily for a French audience. Gerald Berreman dismissed Dumont’s work by characterizing it as based on the Brahman’s point of view, but this important critique was curiously taken by most on both sides of the Atlantic at the time as reductive.
One of the most engaged, if critical, early reviews of Dumont’s work came from McKim Marriott, who nonetheless saw the book as replacing the “older standard works by Ghurye and Hutton.”40 Marriott faulted Dumont for his use of data and his choice of texts, but was most disturbed by the intellectualism of the work, the ways in which “ideology” seemed abstracted from the countless interactions that produce the effective ranks and orders that make up hierarchy in Indian society. Although he had not yet fully formulated a monistic theory of Indian society by this time, Marriott was already distressed by the dualism of the account, which made ideology seem arbitrary and removed from the possibility of either empirical test or effective action in the world (other than the world of French theory). As Marriott developed his new view of caste in collaboration with the cultural historian Ronald Inden, he took as his primary target the work of Dumont, even though to some outside observers the two views (one deriving from French structuralism, the other from American culturalism, but both seeking certification from Sanskritic texts and key Hindu ideals), shared in a great deal more than they differed. From within, however, the debate was bitter. And Dumont, whose position was indeed unabashedly intellectualist, if committed in its own way both to ideological and empirical analyses, was perhaps correct in his critique of Marriott’s later position. In a new preface to Homo Hierarchicus, he wrote that Marriott combined his early empirical “transactionalism” with his later “analysis of culture, or of symbols and meanings, in the manner of David Schneider, thanks to a monist metaphysic that permits them to coincide like matter and idea. . . . He does this unstintingly, and in Marriott and certain of his followers one finds a syncretism of disparate notions, taken out of context, which goes far beyond the known feats of Hinduist popularization.”41 He did not realize, of course, how a similar characterization, differing only in emphasis, could be applied to his own effort to devise a sociology for all of India (ancient, medieval, and modern).
Dumont’s work thus indexed not only a particular trajectory within postwar French social theory but also a far more general anthropological fascination with “caste society” that worked to reinstall caste once again as the major symbol for Indian society. This symbol was a powerful reminder of how that society is organized by religious (read Hindu) rather than secular values. Despite the abstractness of Dumont’s account, what is astounding in retrospect is how very seriously it was taken—in France, Britain, North America, and India—and how, as a European work of social theory, it overshadowed many other issues relating to India in disciplines as diverse as anthropology, sociology, history, political science, and Indology for many years, at a time when there were arguably more important issues to contemplate. Perhaps most astounding of all is the fact that at the very time that postindependent India was struggling with the resilient legacies of colonial history in the context of the question of secularism, a proposal that religion encompassed politics in India would take such pride of place in international scholarship. Although Dumont’s intent was not specifically colonial, and although he attended to modern changes in India in a variety of important essays, he resurrected colonial categories and arguments at a time when the West mistook their overdeter-mined reality for an explanation of the East’s failure, and colonialism was either forgotten or consigned to Raj nostalgia. India’s postcolonial struggle to reinvent the nation and the state, and to find a basis for civil society that would neither be overtaken by conflict between religious and ethnic loyalties nor defined by religious principles, was hardly helped by the rebirth of colonial Orientalism in contemporary Western social science.
Dumont argued that the political and economic domains of social life in India had always been encompassed by the religious domain, articulated in terms of an opposition between purity and pollution. For Dumont, as for Weber and others before him, the Brahman represents the religious principle, inasmuch as the Brahman represents the highest form of purity attainable by Hindus. The king, although important, only represents the profane political world, and is accordingly inferior to and encompassed by the Brahman. The overarching value accorded to the religious domain is the central feature of the ideology of caste—which Dumont characterized with the single word “hierarchy.” Dumont argues that the sociological significance of hierarchy has been systematically missed by modern writers obsessed with the ideology of equality. Caste is fundamentally religious; and hierarchy is about the valorization of society over the individual. Dumont reasoned, in ways not entirely dissimilar from later formulations by Marriott and Inden, that varna and jati were not opposed but rather that the principles of varna underlay the actual organization and articulation of hierarchical relations between and among jatis. But what was distinctive for Dumont was his sense that even if the classical Brahman and Ksatriya were not present in a distinct social system, the values associated with them—status and power, respectively—would always be there, and were always held as foundationally separate and resolutely hierarchalized. And, as he put it, “this was not enough: for pure hierarchy to develop without hindrance it was also necessary that power should be absolutely inferior to status. These are the two conditions that we find fulfilled early on, in the relationship between Brahman and Kshatriya.”42 Whereas Marriott and Inden encoded the significance of the Brahman in a monistic theory of “biopower” that sought to avoid what they took to be an exogenous dual categorization of status (the domain of the ideal) and power (the domain of the real), Dumont insisted on the separation and hierarchical encompassment of the two categories as generative of the caste system itself.
Dumont thus identified the political and economic aspects of caste as relatively secondary and isolated. In assessing recent changes in the caste system, he noted that the British colonial government’s policy of “not meddling in the domain of religion and the traditional social order, while introducing the minimum of reforms and novelties on the politico-economic plane,” significantly reduced the extent of change and conflict under colonial rule.43 Not only did this accept the self-representations of colonial policy, it also suggested that the British were right to make the distinctions they did between religion and society (distinctions that were always vexed, and that even according to colonial policy makers never worked quite right). In fact, the British introduced these distinctions for the convenience of colonial rule and, given the force of colonial presence and power, they took root in serious ways during two hundred years of rule. Dumont thus not only relied, perforce, on colonial sources, he also reproduced colonial ideology even when he assumed he was being most critical of Western social and political theory. The peculiar complicity of Dumont’s sociology in colonial sociology was clear in his original call for a “sociology of India.” His insistence on the unity of India—an insistence he claimed had nothing to do with the somewhat spurious political uses of the idea in the independence movement—not only worked to seal India’s borders from the rest of the world but to mark India as unique in particular because of the “existence of castes from one end of the country to the other, and nowhere else.”44 Dumont used an argument about history to deny India history, much as Hegel did before: “If history is the movement by which a society reveals itself as what it is, there are, in a sense, as many qualitatively different histories as there are societies, and India, precisely because she is indifferent to history, has carefully laid it down in the form of her society, her culture, her religion.”45 Thus Dumont not only exempted the colonial state from any role in the constitution of modern India, he exempted India from history altogether. With this view, it is hardly surprising that in his theories the state—traditional or modern—would be of so little importance.
In Dumont’s view, caste not only subordinated the political, it also reduced the individual to a position of relative unimportance. The individual only has ideological significance when placed outside society, or to put it in Dumont’s terms, as “the individual-outside-the-world.”46 This is the individual as the renouncer, the sannyasi who must leave both society and the mundane world to attain transcendental truth. Dumont’s position was stated more forcefully by Jan Heesterman, a Dutch Indologist who with Dumont played an important role in defining the discourse of Indian sociology: “Here we touch the inner springs of Indian civilization. Its heart is not with society and its integrative pressures. It devalorizes society and disregards power. The ideal is not hierarchical interdependence but the individual break with society. The ultimate value is release from the world. And this cannot be realized in a hierarchical way, but only by the abrupt break of renunciation. . . . Above the Indian world, rejecting and at the same time informing it, the renouncer stands out as the exemplar of ultimate value and authority.”47 The individual as renouncer thus occupies a critical position in what Heesterman called the “The Inner Conflict of Tradition,” in a transcendental critique of the possibility of politics, economics, or history in the Indian world.
The prominence of Indologists and Indology in anthropological discourses on India has not only elided the monumental role of Islam in the history of the subcontinent, it has also worked to secure a specialized scholasticism for India. For much of the academic anthropology of India, Indology replaced history, and has been used to dehistoricize both India and anthropological practice in India. Not only has Islam been erased and the state been ignored as a potent force in the constitution and transformation of Indian society but the colonial history of India has been rendered entirely insignificant, as we have just seen in Dumont’s peculiar sense of caste’s compatibility with empire.48 I find this compatibility unsettling, and as must be clear by now, see other, not unrelated, similarities between the view that the precolonial state was weak, the assertion that traditional society was organized by social and religious rather than political principles, and the conferral on caste of the exemplary status of a traditional form that has resisted the development of modern state and social structures. I become even more unsettled when I read words such as those written by Jan Heesterman, in the introduction to his book: “The modern state . . . wants to bring the ideal of universal order from its ultramundane haven down to earth. The inner conflict then becomes explosively schismatic, as eventually became clear in the drama of the Partition.”49
And so the endless rehearsals of the essential difference between East and West, between the recent histories of India and Europe. In India caste, so colonial sociology had it, always resisted political intrusion, and was already a kind of civil society in that it regulated and represented the private domain, such as it was. But caste could not be more different from modern Western society, for it neither permitted the development of voluntarist or politically responsible social institutions nor did it work to collaborate with the modern state. Indeed, caste actively resisted the modern state even more than it did the old, for the modern state opposed rather than supported dharma. The catch here was that under colonialism the modern state was not a viable option, since its development depended in large part on the conquest and exploitation of lands where premodern states fell to the technological, military, and economic power of the ascendant West. But colonialism was predicated on more than simple economic exploitation, and its effects were as various as they are difficult to untangle, even now, from the presumed weight of tradition on colonized societies.
Colonialism in India produced new forms of society that have been taken to be traditional; caste itself as we now know it is not a residual survival of ancient India but a specifically colonial form of (that is, substitute for) civil society that both justified and maintained an Orientalist vision. This was a vision of an India in which religion transcended politics, society resisted change, and the state awaited its virgin birth in the late colonial era. Thus caste has become the modernist apparition of India’s traditional self. Under colonial rule, caste—now defined by the dharmic idea of varna, disembodied from its former political contexts, and available as the principal object of colonial knowledge—could take on a new and different form. In this dissociated form it was appropriated, and reconstructed, by British rule. And even after decolonization, academic preoccupations have continued to be fascinated with the same chimeric forms that so preoccupied the British, even as they have mistaken the effects of British rule for the traditional predicates of it.50 What anthropology and Indology together have done most successfully in the postcolonial context has been to assert the precolonial authority of a specifically colonial form of power and representation, not only disguising the history of colonialism and the essentially contingent and political character of caste but also reproducing what might be the most extraordinary legacy of colonial rule in the contemporary social life of caste and Hinduism in India today. The state of ethnography—British, French, and American—turns out to be a direct descendent of the ethnographic state.