I am a man who can recognize an unnamed town by its skeletal shape on a map. . . . So I knew their place before I crashed among them, knew when Alexander had traversed it in an earlier age for this cause or that greed. I knew the customs of nomads besotted by silk or wells. When I was lost among them, unsure of where I was, all I needed was the name of a small ridge, a local custom, a cell of this historical animal, and the map of the world would slide into place.
—Michael Ondaatje, The English Patient
It is the State which first presents subject-matter that is not only adapted to the prose of History, but involves the production of such history in the very progress of its own being.
—G.W.F. Hegel, The Philosophy of History
The archive, that primary site of state monumentality, is the very institution that canonizes, crystallizes, and classifies the knowledge required by the state even as it makes this knowledge available to subsequent generations in the cultural form of a neutral repository of the past. Colonial governmentality was not merely dependent on knowledge, it was also embedded in the forms of knowledge that provided the basis for the principal practices of the colonial state. Colonial conquest made possible (even as it was made possible by) the marking of new territories with the dimensions and coordinates of colonial interest. Resources were converted into colonial commodities through a conquest based as much on knowledge as on military success. But the conquest of knowledge ran far deeper than the mere conquests of armies, for it was in the inscription of colonized spaces that colonial power was translated into useable forms of knowledge. Colonial conquest was about the production of an archive of (and for) rule. This was not an archive that was imagined as the basis for a national history, for it was only designed to reap the rewards and to tell tales of imperial interest.
Early imperial interest in India was principally concerned with commodities for trade, though the limited resources of northern Europe meant that England could only trade bullion, secured from the legacies of Spanish imperium in the new world, for the textiles, spices, and other riches of the subcontinent. Trade required not only exchange but also translation, as tantalizing in its promise of wealth as it seemed tainted by its dependence on mediation. Imperial interest in controlling the terms of trade led inexorably to imperial expansion of the theater of trade itself, from forts to factories, and then from the establishment of port cities to the negotiation of rights to collect revenue in territories surrounding those ports.1 Even as Clive and Hastings made themselves rich nabobs by engaging in “private” trade through their “public” positions, the East India Company became increasingly restless, seeking a monopoly not only in relation to other imperial interests (within Britain and across Europe) but also to the private interests of its own servants. The Company might have professed that it was no more than a trading company, but by 1757 its enmity with the French, its alliances with some local rulers and its wars against others, and its interest in expanding trade made it look very like a full-scale imperial presence. By the end of the century, with the consolidation of the spoils of Plassey, the taming of Awadh and Arcot, the relentless wars against the southern chiefs (and even the failed efforts against the Marathas), and the final defeat of Tipu Sultan, the trading company had turned into a colonial state.
With the Regulating Act of 1784, the colonial state that was once a trading company was reined in by Parliament and brought under the direction of the “Board of Control.” And it had been charged with the establishment of legitimate procedures for the delineation of revenue and property rights, on the basis of which it was to support itself not from private trade but rather from the collection of revenue. Trade was not unimportant—and the Company was ever in search of new commodities and resources it could collect on favorable terms and export for massive profit—but it became an activity of an enterprise that increasingly took on a life of its own. As the Company became like a state—albeit one at the service of and controlled by political interests in England—it developed a bureaucratic form and rationality that gave to land revenue a salience that even trade no longer had. The land that had been conquered by the military and political victories of eighteenth-century statecraft had now to be surveyed, mapped, and organized in relation to a new form of imperial interest. This interest was coded historically, for not only did the state need to assume the claims of previous (or for that matter present) rulers to own either the land itself or the right to its produce, it also needed to legitimate its assumption of these claims through historical precedent and bureaucratic management. And the state needed to fill its coffers without fomenting either revolution or massive chaos, and without occasioning the outrage of new Burkes in distant England who might conceive the greedy appropriations of the state as a new form of private corruption.
The turn of the century, then, was a time not only of major transformation but also of frenetic contradiction. Mackenzie’s historical interests were rapidly overtaken by those of a new generation of colonial bureaucrats. As it turned out, Thomas Munro, the contemporary and good friend of Mackenzie who designed the “ryotwari” settlement of Madras presidency, became far more important for the production of a colonial archive than Mackenzie and his vast collection.2 This new archive buried with astonishing rapidity the useless detritus of the age of conquest, leaving eighteenth-century history (and its historians) behind in several related senses.
Unsurprisingly, what survives at the center of the new colonial archive are not the myriad records and manuscripts collected by Mackenzie but rather the land records that became so fundamental to the debates over land tenure and settlement in the initial years of British rule.3 These documents—used so extensively by historians of agrarian relations—turn out to be far more than assessments of different land parcels and their potential (or actual) productivity. They were, rather, interventions in the way the colonial state worked to constitute land relations as the basis of the state’s ultimate right of ownership and, more generally, delineated relationships between state and society.
The early colonial state was chiefly an agrarian state that used various representations of “oriental despotism” to justify its legitimacy and fortify its claims to ultimate power through the bureaucratic regulation of landed property. Building on arguments between those who held that the East India Company was inheriting the king’s right of ownership over all property, and those who used a Ricardian theory of rent to claim for the Company the right to set revenue rates and collect taxes as fundamental to the custodial project of the state, the British gradually established a state bureaucracy that focused primarily on land revenue. Decisions about whether the bureaucracy should accord proprietary rights to landlords (zamindars), village brotherhoods, or principal cultivators (ryots) became critical interventions in the relationship between state and society, at the same time that these decisions both produced and were produced by a variety of different histories of India that were important parts of early colonial rhetorics of rule.
When I first waded through settlement land records, I did so to determine the nature of agrarian relations in different parts of southern India, as well as to assess arguments made by different administrators about the nature of the precolonial village community. The arguments were complex and robustly documented, and always assumed that historical forms were necessary predicates for colonial policies. Intellectual histories of some of the key players of the period by historians such as Ranajit Guha, Eric Stokes, and Burton Stein have revealed how integral historical argument was to political ambition and European experience. Cornwallis was influenced by the physiocrats and driven by his ambition to recreate in India the authority and position of the landed gentry in Britain (already under major assault and in considerable defensiveness, given the events of the revolution in France); Munro was captured by Burkean rhetorics of paternalistic responsibility and a Scottish sense of the folk heroism of the yeoman cultivator; Elphinstone was enamored by Bentham but still in favor of local systems of administration based on the experience of those most grounded in Indian affairs. The discrete land records—with their plenitude of facticity—reflected these genealogies and policies, and are the result of the documentation project of the early colonial state around matters of land and revenue. Indeed, from the swashbuckling time of Clive and Hastings to the events of the Great Rebellion of 1857, conquest was the pretext not only for political control but also for revenue collection. As surely as the mercantilist logic of imperial trade yielded to the bureaucratic imperatives of the imperial state, land became what the colonial state was all about.
Edmund Burke had not only called Hastings to account during the years he held impeachment hearings, he had also questioned the very foundations of the British imperial presence in India. Part of the problem was that the British presence was not in fact imperial; it was based instead on the activities and provenance of a trading company that had managed to secure a monopoly but nothing like a charter that would justify imperial expansion across the Indian subcontinent. When Warren Hastings became governor-general of India in 1774, the position of the Company was still precarious. By making alliances with rival factions in succession disputes, choosing battles to fight in which there were high probabilities of success, and extorting massive sums of money from dependant allies, Hastings was able to secure extraordinary territorial gains. His methods were not always scrupulous; he financed his war efforts by the same combination of plunder and tax farming used by the Marathas and Mughal emissaries of the time. It is small wonder that Hastings’s brilliant expansionary efforts became vulnerable to parliamentary review, which under siege by Burke’s rhetorical attacks quickly seized upon both the irregularity of Hasting’s successful politics and the fact that he conducted them on behalf of a trading company that had no authorization for even the most tentative imperial ambition. But there was far too much profit from the lucrative textile trade, and far too many connections between Company directors and political interests, to suspend the India operations altogether. The Pitt Act of 1784 established a Board of Control in London that was supposed to oversee the actions of the governor-general, and gave voice to concerns in Parliament that the Company either clean up its business or cede its private monopoly. And when Lord Cornwallis—who followed up his disgrace in North America with an impressive military victory against Tipu Sultan in 1792—established the Bengal Permanent Settlement in 1793, he did so in concert with a major reform in which Company servants were paid much higher wages in return for further discouragement against engaging in private trade.
The Permanent Settlement was the brainchild of Philip Francis, a lifelong foe of Warren Hastings. Hastings was by this time in disgrace, however much his actions in India had made it possible in the end for Francis, Shore, and Cornwallis to propose the comprehensive land settlements represented by the 1793 reform. Indeed, Hastings had skillfully balanced the Mughals, along with the nizams of Awadh and Hyderabad, against the Marathas and the rising power of the Mysoreans. But although he had set the stage for the Permanent Settlement by bringing Bengal under nominal British control during his rule, and had prepared the way for Britain’s dominance on the subcontinent by the last decade of the eighteenth century, the Permanent Settlement was an attempt to erase Hastings’s legacy in more ways than one. As formulated initially by Francis and implemented by Cornwallis, it was meant to regularize Company revenues through a steady tax rather than by extortion, to normalize administration by setting high public standards for the service of Company officers, and to create a loyal elite based on landed property rather than military alliance, by restoring the putatively traditional landholders to their rightful position. The erasure of Britain’s history of Indian conquest began in the embarrassment around the aggressive expansion of Hastings, but soon took a different form under the leadership of Cornwallis, who was intent on reproducing the landed gentry of England, in a dramatic enunciation of imperial policy that seemed a denial of the entrepreneurial origins of Indian empire even as it sought to stabilize a new kind of Indian elite. But Cornwallis was also expressing his frustration about the task of knowing, surveilling, and administering India. The Permanent Settlement was in part the result of British inability to get a handle on the actual levels of production in agrarian tracts as well as on the best way to adjudicate competing claims for local-level proprietorship, tenurial right, and revenue relief. Given this level of uncertainty, it is perhaps no surprise that the settlement in fact led to many estate auctions and a high turnover in the rural elite, at least in the short term.4 And although Francis had justified the settlement by his theory that control over profits would motivate Indian landlords to improve their properties in systematic ways, once inflation and increased agrarian productivity conspired to increase profit, most landlords became absentee, using their estates to support other pursuits in Calcutta rather than reinvesting money and energy in agrarian management.
The Permanent Settlement was bold and astonishingly simple. It entailed setting a single amount of tax for each estate that would remain the same in perpetuity, with each contracting zamindar free to alienate proprietary title and, by implication, the fixed revenue obligation. Property was introduced along with a fixed revenue demand, though the right of property was vested at a level that sometimes mimicked a feudal structure of old, and at the very least encouraged the maintenance of large estates with tenurial relations and agrarian management under the centralized control of estate managers. Francis and Cornwallis assumed that many of the assets of the estates were hidden from view, and that in any case the application of enlightened managerial attention would bring new profits to light in short order, and so the revenue demand was set at high levels. After all, once fixed it could not be changed. But the colonial interest in fixing agrarian structures and property relations misfired in serious respects. Not only were many of the contracting zamindars hardly representative of the landed aristocracy Cornwallis had hoped to identify and encourage but a great many zamindars defaulted on their revenue payments, sending virtually half of the estates to the auction block within the first twenty-five years of the settlement. In Bengal, the landed gentry turned out to be as “open,” and as much part of a larger history of social transformation, as Lawrence Stone has argued was the case for the landed gentry in seventeenth- and eighteenth-century England.5
Despite its detractors, the Permanent Settlement was such a compelling idea, and so alleviated British anxiety about having first to determine and then to manage the microlevel details of agrarian production, that it seemed to many the only way to look after, and collect the necessary funds from, the areas over which the Company assumed direct control.6 But as useful as the settlement was for the transition from trading company to company state, it began to incur criticism from Company officers who worked in localities where they steadily took on more direct control over agrarian affairs.7 Among these British officers was Thomas Munro, who along with his initial supervisor Alexander Read had already begun to experiment with making direct settlements with cultivators as early as the mid-1790s. Nevertheless, the idea of the permanent settlement was still compelling, particularly in those areas where powerful chiefs—palaiyakarars as well as zamindars—made the threat of potential resistance as worrying as the difficulties that might accompany the acquisition of reliable information on agrarian matters within the countryside. Accordingly, in 1802 approximately one-third of Madras presidency was, in fact, settled under Cornwallis rules, establishing “loyal” as well as particularly powerful chiefs as the Madras version of Cornwallis gentry. But the Madras settlement tended to be in areas that were not nearly as productive, or central to the agrarian economy, as had been the case in Bengal, and was frequently contracted with local chiefs who had been loyal supporters of Company rule. Indeed, the zamindari settlement was effected in areas that had managed substantially to escape a steady process of encroachment and agrarian conversion under the combined forces of the Tanjore Marathas, the Mughal emissaries in Arcot and Hyderabad, and the especially aggressive bureaucratic policies of Haidar Ali and Tipu Sultan. The agriculturally rich riverine and deltaic regions of the Tamil and Andhra plain were kept for further revenue experimentation, and soon succumbed to the changing character of imperial rule, first enabled by the military victories over Tipu Sultan and the French, and then as a part of the general transformation of the Company state into a bureaucratic enterprise of a very different sort than that imagined even by Burke in his most imaginative renderings of the future of the British imperial presence in India.
Thomas Munro played a particularly critical role in this shift of British opinion and administration. Credited as he is with the invention of the ryotwari system, which meant “direct” revenue settlements with the cultivators, he in fact embodied in his Indian career some of the primary transformations in British rule from the late eighteenth to the early nineteenth centuries. He went to India in 1780 as a military man and spent his first four years in warfare against Haidar Ali of Mysore, not at first with any success. It was in part this early experience that made him so wary of the threat to British interests represented by Haidar Ali, and later his son Tipu Sultan. It certainly set a theme for his Indian career, for when he became a Company administrator, entrusted with the task of settling the Baramahal, Kanara, and then the Ceded Districts, he found himself often following in Tipu Sultan’s footsteps. Indeed, in these assignments he often carried on Tipu Sultan’s efforts to suppress the power of the palaiyakarars and to introduce in its stead regularized bureaucratic procedures designed for civil administration and revenue collection. In Baramahal he experimented with annual and individual settlements and, despite his expressed concern for the local cultivator, he set the revenue rates at an exorbitant level. In Kanara, where joint village arrangements had been even more transformed by Tipu Sultan’s interventions, he developed an historical theory to justify his reliance on the idea of individual private property. And yet he also recommended that provisional settlements be made for at least five years to assure the Company that they would set their assessments on the basis of an adequate understanding of agricultural assets. In the meanwhile, he recommended lower assessments than elsewhere, no doubt in part reflecting the residual power of local interests, chiefs, and palaiyakarars prominent among them. Curiously, at the very moment that he took full responsibility for a revenue settlement of his own, he stood on considerably shakier ground than he had in Baramahal, where the palaiyakarars had already been significantly displaced and village institutions were in a stronger condition. Successful nevertheless, he was sent to the much vaster Ceded Districts in October of 1800 with the task of settling a region that posed even greater difficulties to the Madras government, which was financially in a bad way after the military campaigns of the previous decade and politically stressed by the rapidly escalating administrative responsibilities that were part of its new charge.8
Munro managed to pursue an aggressive policy towards the many active palaiyakarars who remained in positions of local power. He no sooner reduced many of them to pensioners than he managed to coopt them by the provision of inam benefits for those who promised loyalty.9 Munro was frequently attacked by those who felt that he was too harsh on local chiefs—a symptom in part of his military rather than civilian background—but it was his very success in establishing local order and generating considerable revenue that convinced even his detractors and set the stage for his successful advancement of the ryotwari cause. Once again he flew in the face of considerable local evidence that suggested the institutional integrity of the village as a revenue unit, and selected dominant cultivators for proprietary title, fashioning a new system based on the isolation of elite farmers for a steadily escalating revenue demand.10 Although he justified his policies with historical argument and precedent, he developed a method of revenue collection and local administration that was successful in large part because of a growing Utilitarian sentiment in Madras and London, as well as his evident success in raising cash and restoring order. Munro returned to England in 1807, to find a wife and advance his political fortunes. After an uncertain wait for a few years, he was vindicated when the Fifth Report of 1812, written to advise Parliament about the state of administration in India as part of the Charter Renewal debate of 1813, chose the ryotwari scheme as the revenue method of highest regard.11 Munro returned to India in 1814 to oversee the revision of judicial administration in Madras, and went on to play a major role in the military victory against the Marathas in 1818, ultimately rising to the position of governor of Madras in 1819. Although his ascendency has been generally understood in terms of the rise of a certain kind of administrative paternalism and the canonical ryotwari system of revenue collection, I would emphasize instead the extent to which Munro played a major role in the establishment of the Company as a bureaucratic state system, which during these years developed extraordinary confidence in its capacity to know, to regulate, and to profit from local social forms. And although Munro justified his new approach through an egalitarian rhetoric that was posed against the oligarchic regime of Cornwallis in years past, the ryotwari system aided both the radical extension of Company power into the Indian countryside and the introduction of new forms of capitalist social and economic relations.
In order for Munro’s ryotwari system to work, the Company had to collect massive amounts of local information, and in the process it depended on local accountants and assistants to a greater extent than ever before. The zamindari system had been an outgrowth of an older system of tributary relations, even as it reflected a provisional compromise with the landed structures already in place. The ryotwari system—however much it was justified by resort to historical precedent—represented an unprecedented level of state governmentality, an explicit effort to engineer a new kind of political and economic elite and to micromanage a local-level agrarian economy. It involved not just the identification of individual cultivators and the adjudication of different claims to local landholding but also the periodic assessment of individual lands, rated according to soil classification, irrigational resources, and other factors that affected production such as climate, topography, and so on. Munro’s successful advocacy, and implementation, of this system was part of a larger transformation in the authoritative production of information, an imperial archive that was expanding rapidly in the early years of the nineteenth century. It was an archive that had as its principal raison d’être the collection of revenue. Accordingly, Munro’s judicial reforms, intended as they were to unite the roles of the magistrate and the collector, organized local government within the guiding framework of a revenue state, and conferred judicial authority on those charged with the administration of the revenue system.
Munro’s style of local government was both utilitarian and paternalistic, and set the stage as well for Elphinstone’s pragmatic appropriation of Peshwai precedent to fashion a colonial state in western India after the defeat of the Marathas in 1818. In the colonial setting, bureaucratic state forms developed in order to enlarge the coffers of the state, balancing exploitation against potential rebellion by the careful manipulation of local interests and newly empowered elites. The state forms were more intrusive and interventionist than their rhetoric made them out to be; they were also more closely wedded to colonial forms of domination than could have been possible in a metropolitan politics that witnessed at this very time new political movements and the beginning of serious pressure for increasing democratization, in the wake of the French Revolution. Even as the forms of state governmentality were established through the mechanisms of a revenue bureaucracy, Munro’s paternalism was substituted for sovereignty and was designed to justify the transformation of the power of the colonial state.
This, then, was the emerging context in which the British state sought to know more about the social structure as well as the productive resources of the Indian countryside. This area was expanding in extent for the British by leaps and bounds, given the defeat first of Tipu Sultan and then of the Marathas, as well as the progressive isolation of the nawab of Arcot and later the nizam of Hyderabad. Officials such as Mackenzie and Buchanan had developed different forms of the survey in Mysore, where princely rule meant a fixed tribute rather than a direct revenue system and potential resources would only have tangential value. But the new generation of government servants under Munro and men like him held revenue matters as the major activity of the state and the principal purpose of collecting information. Land records of the time are punctuated by periodic revenue accounts that systematize new levels of detail about matters related to agricultural production. Whereas Buchanan had been chided for being too prolix in his descriptive memoir of Mysore, new surveys were increasingly statistical, constructed for ready comparison and quick accountability. It is instructive that by the end of the first decade of the nineteenth century, when Buchanan was commissioned to undertake a new survey of the districts of Bengal and Bihar, he was himself engaged in a very different kind of survey, with extensive tables and classificatory rubrics used precisely to allow a statistical comparison across time and space for immediate administrative consultation.12 Mackenzie’s scrupulously collected and translated manuscripts concerning the histories of villages, localities, temples, kingdoms, and chiefly families might have been seen as of great (at least potential) value in the climate of the zamindari settlement and in the context of an early colonial state that was still principally concerned with identifying the proper points of mediation. But despite Munro’s considerable personal support of Mackenzie’s project—based in large part on his own extensive historical interest in the past of local social, political, and economic institutions—the system ushered in by Munro made this kind of information seem increasingly antiquarian. In retrospect, it should come as no surprise that Mackenzie’s collection fell through the cracks of the congealing sense of a colonial archive.
But what did caste have to do with this new informational regime or, to put the matter more directly, how did notions of caste change in relationship to it? Very little, indeed. When Buchanan was commissioned to undertake his survey of Mysore in 1800, caste as such was not even mentioned. Instead, he was asked, among many other things, to take into account that “The different sects and tribes of which the body of the people is composed will merit your observance; you will likewise note whatever may appear to us worthy of remark in their laws and customs, and state with as much accuracy as may be in your power the nature of their common usages in matters of personal traffic at their markets their weights and measures the exchange of money and the currency amongst the lower order of people and such matters in respect to their police as may seem to you to have immediate or particular tendency towards the protection, security, and comfort of the lower orders of the people.”13 Buchanan’s journal makes many mentions of Brahmans, in large part because Brahmans were contacted on a regular basis by him for information regarding the histories and customs of particular places. As was the case for Mackenzie as well, Brahmans were widely accepted as the local literati, and even when they were not the specific objects of inquiry, they were referred to by others as those who could provide reliable accounts of matters ranging from local history to sociology, from revenue systems to the existence of local texts and records. Buchanan was aware that the Brahmans had prejudices of their own, as for example in a village of Canara, where the Brahmans disparaged the knowledge of the local accountant, a non-Brahman: “These Bahudundas the Vaidika Brahmans hold in great contempt; but as the office of Shanaboga has in numerous instances continued for many generations in the same family, I am inclined to think that from this source much historical information might be procured.”14 Buchanan noted further that the Brahmans in question had contempt even for other Brahmans. But he nevertheless took for granted the role of Brahmans as an intellectual elite, if not always given to the spiritual otherworldliness that characterized them in some contemporaneous literature.
In fact, Buchanan reserved some of his sparing ethnographic comments for Brahmans themselves, struck no doubt by the contradictions attending Brahmanic status. In one case he noted that despite the obsessive endogamy of a particular “nation” of Brahmans and the prohibitions on widow remarriage, the custom of Brahman men was to take a woman who had been dedicated to the temple goddess (as a devadasi) as a “concubine.”15 Buchanan used the term “caste” only for the four varnas, and used terms such as “sect,” “tribe,” or “nation” for subcastes of a more specific sort. Occasionally he relayed some information about the customs of groups other than Brahmans, though with little regularity, and virtually no anticipation of any kind of colonial ethnographic interest.
In some cases, Buchanan came across statistical statements of the population, known as “kaneh shumareh,” sometimes as “caneeshamari.” The lists were broken into caste or subcaste groups, counted by number and gender.16 As Buchanan noted in the case of a list he procured from southern Kanara, “The different castes are detailed in the usual confused manner, with which they are spoken of by the native officers of revenue.”17 In a population of slightly less than 400,000, divided among almost 80,000 households, there were 122 caste groups listed; they were headed by Brahmans but including such miscellaneous caste groups as Kankanies (bankers and traders), Rajputs, Muslims (exclusive of Moplays), Bhats (genealogists and poets), Marathas, Parsis, Garwadys (snake catchers), Reddis (farmers), Jogies (religious mendicants), Julais (weavers), Jettys (wrestlers), and Dhobis (washermen). The list is miscellaneous, mixing recognizeable “caste” names with occupations, religions, ethnicities, and other social categories. As Buchanan indicated here, and as is clear from records elsewhere, these population tables were produced by revenue officials (sometimes Brahmans, sometimes not) who maintained village accounts. Although they consistently break the population down into caste groupings, they do so in ways that suggest the haphazard character of caste as a marker of identity. In Mackenzie’s Mysore survey, which includes a more regular compilation of caneeshamari tables than can be found in Buchanan’s descriptive memoir, some of the documents attempt to group the long list into caste varna groups (Brahmans and other twice-born, Sudras, and others), but these attempts seem half-hearted and hardly helpful, when they classify the bulk of the population listed as “Soodra.”
Once again, however, the Mysore survey gives little evidence of a colonial interest in caste per se, or of the prominence of caste as a social category that demanded attention except insofar as certain Brahmans were clear about declaiming their own distinct, and distinguished, position in society. As Mackenzie noted about the survey, he was successful in collecting materials and documents of different kinds, which concerned such matters as “the genious and manners of the People, their several systems of Government and Religion, and of the predominant causes that influence their sentiments and opinions to this day . . . the tenures of lands, other origins and varieties of assessment of rents and revenues, and the condition of the peoples and the privileges of the different classes.”18 But if Mackenzie collected, and wrote, almost nothing about “caste” aside from the caneeshamaris that were part of his cartographic surveys, it was both because of the relative unimportance of caste in many of the areas of British intervention and interest and because of the much greater salience of other institutions; these extended from the character of property and revenue relations to issues of local governance and political legitimacy. The relative ignorance among the British of caste matters, even concerning Brahmans, was reflected when Francis Ellis, a government servant in Madras and an eminent philologist and scholar, noted with impatience that James Mill, in his large history of British India, had failed to understand what was meant when Brahmans were designated as priests. In fact, as he observed, the lowest category of Brahmans were those who actually served as priests in temples (though those who served in sacerdotal capacities for domestic rituals of Brahman families were of a different category).19 When H. H. Wilson edited a new edition of Mill’s history, he had many occasions to correct fundamental mistakes, as well as to chide Mill for his ignorant contempt of Indian institutions. But even Wilson had little in the way of “ethnographic” knowledge of caste matters. The continued prominence of Dubois’s antiquated text was the result of the combined effects of ethnographic inexperience and the fundamental disinterest in what only later became the basis for a colonial sociology of knowledge. For most of the early nineteenth century, Orientalist recodings of the increasingly reified dharma texts of Manu seemed perfectly sufficient to summarize the social forms of the Hindus. What in retrospect is most extraordinary about Wilson’s edition of Mill, his substantial disagreements notwithstanding, is that such a volume was possible at all. The Anglicists and the Orientalists might have differed in the ways they viewed India’s past and such Indian institutions as caste, but the two both groups shared fundamental assumptions about contemporary Indian society, and neither accumulated much in the way of empirical information about it.
What I have suggested in this chapter, therefore, is that questions of land tenure were seen as far more fundamental to the colonial understanding of rural social structure than empirical knowledge about caste. At most, caste was important in relation to debates over historical forms of land tenure. Occasionally, learned Orientalists made major contributions to this debate, as for example in the substantial treatise by Francis Ellis on mirasi rights in southern India. Ellis used a wide variety of historical sources to argue that the Vellalar settlers of Tondaimandalam, the area around Madras, had true property rights in the land, and were not tenants at the will of the king in the mode assumed by those holding the Oriental Despotism thesis.20 He further argued that they all held land jointly, rotating their use of particular fields annually among the village shareholders. Significantly, his argument was solely textualist, and he used no ethnographic/administrative knowledge to support the claim. Munro argued back in part on the basis of his long experience of life in Indian villages rather than mere textual knowledge about them; he was hostile to this argument, and indeed to any notion that would suggest corporate brotherhoods among or within dominant caste groups. Indeed, Munro argued that although the village community was of primordial importance in Indian civilization, landholders always held their land directly from the state, except during moments of extreme disruption, such as when Tipu Sultan appropriated local land rights. Munro was moved most by the needs of his own ryotwari proposals. When he argued that the Indian state had been oppressive, and had engaged in excessive and capricious taxation that discouraged security and improvement, he was justifying his own revenue settlement; he was justifying, as well, his vision of a partnership between the colonial state and the dominant peasants, mediated only through a Utilitarian commitment to property. Munro was simply uninterested in caste; it served neither to explain some old system of community landholding nor to justify skepticism about the future consequences of colonial innovations predicated on ideas of individual property in land.
In retrospect, Louis Dumont argued vociferously against early colonial writings on the importance of the village community because of his sense that these writings completely ignored caste.21 Dumont, for whom community could only be understood in relation to caste and ideas of hierarchy, was deeply concerned that contemporary anthropologists might take too seriously early colonial sources that depicted community in ways dependent on political, economic, and social modes of production and distribution. He was doubtless correct to caution scholars about the romantic character of unhistorical versions of village republics. He was unaware, however, of the extent to which his own interest in caste was dependent upon the emergence of equally romantic epistemological predispositions and an accompanying body of knowledge that was produced only during the late nineteenth century. Early colonial notions of Oriental Despotism and autonomous villages were both wrongheaded and tied to a colonial interest in organizing relations of production around new state systems and regimes; they do reflect, however, the extent to which caste relations were either subordinate to or perhaps embedded in both economic relations within localities and the political organization of relations between agrarian locales and different state units and structures.
When caste was singled out for empirical notice, it often appeared in awkward ways. As I have mentioned above, it occasionally appeared in lists called caneeshamaris, another example of which can be seen in the 1823 census of Tirunelveli. The Tirunelveli census provides a population breakdown by caste groups, crudely divided into eight broad categories, in rank order as follows: Brahmans, Ksatriyas, Sudras connected to the religious establishment, Sudras of different denominations, “Muhamedans,” “Christians of different castes,” inferior Sudras, and the low castes (or untouchables). These macro categories reflect an uncertain marriage between the varna scale of the Manu Dharma Sastras and contemporaneous revenue classifications. These classifications differentiated the broad and largely meaningless group of Sudras into those who were supported by inam benefices and those who had direct revenue relationships with the state; these latter were alongside “Muhamedans” and “Christians of different castes” in the general caste scale (that is, groups placed above only the inferior Sudras and the untouchables). But the categories also make clear the limits of the varna scheme for the empirical reality of social differentiation in the far south. Ksatriyas could barely be found (and the classification of a group such as Goswamy in that group was a major stretch); Vaisyas simply did not exist. Although Brahmans were a discrete, if much subdivided, community, the appelation Sudra had meaning only in the sense of demarcating difference vis-à-vis Brahmans on the one hand and untouchables on the other. The actual subdivisions among Sudras reflect a highly differentiated social landscape, which distinguished “caste” groups such as Vellalars and Mudaliyars, occupational groups such as weavers and potters—among many other kinds of groups—all with territorial affiliations that appear every bit as significant as any other aspect of social difference. The list is not only miscellaneous, it is also unranked within categories, no doubt because of the impossibility of securing any consensus about rank outside of small localities where exchanges both established and reflected at least some of the character of local hierarchy.22
There is one fragment in the papers of Colin Mackenzie that sits out of place, a trace of a single ethnographic experiment that one of Mackenzie’s assistants during the years of the Mysore survey must have attempted in order to understand the nature of local-level rank. On a single sheet of parchment, a curious list of caste groups was positioned on a grid, marked in such a way as to document the relationship of social rank to food exchange. Those marked N–1 were said to be allowed to eat in the houses of those under which column they are placed, and those with a blank are said to be prohibited from eating in the houses of those under which column they are placed. Anticipating by a century and a half the ethnographic proposals of McKim Marriott, the document demonstrated that within a village setting the principle and practice of food exchange—by which is meant simply whether or not one group will eat food prepared, or offered, by another group—was correlated with social rank, in this case stretching from Lingayats to “Vellauls” to Chetties, Reddys, and Coravurs.23 What is also interesting (and in agreement with the early interactional proposals of Marriott), however, was the assumption that varna- (or any macro-) level classifications were irrelevant. The “caste” groups were highly localized and heterogeneous, including as they did families of accountants (or Kurnams), and local priests (Pundaram), as well as miscellaneous occupational groups. As in all the other documentation from the early-nineteenth-century, there was still nothing formal, or systematic, in the organization of caste categories.
Despite rare exceptions, most early-nineteenth-century catalogues of caste betray both the uncertain character of caste and the extent to which knowledge about caste was fashioned specifically for colonial interests. This can be clearly seen in one book on caste that was written by another of Mackenzie’s assistants, C. V. Ramaswamy, Boria and Lutchmia’s younger brother. In 1847, Ramaswamy “compiled” a book entitled “A Digest of the different castes of the southern division of southern India, with descriptions of their habits, customs, etc.”24 The work was specifically “dedicated to the British public of India, by their most obedient and humble servant.” Ramaswamy began the book by noting that “According to Hindoo mythology, there were originally created by the God almighty four castes of Hindus who were placed in the land of Bharata Khanda, or India,” namely, Brahmans, Ksatriyas, Vaisyas, and Sudras. But, having provided a short synopsis of their “dharmic” duties, he went on to say that it would be most useful for the British to know the actual caste groups with whom they would come into contact while in India. “Trusting that those individuals both European and Native, who may honor my humble undertaking with their patronage, particularly the former, for whom the work is principally intended, may receive that gratification and instruction which it is my anxious desire to impart,” Ramaswamy went on to provide portraits, both actual drawings and short literary sketches, of the relevant caste groups. The first such category is “Butler,” the second “Dubash” (meaning translator), the third “Cook.”
The caste groups of Ramaswamy’s book—including such categories as grass cutter and waterman, wine cooler and hookah man, horsekeeper and dog boy, dancing girl and Brahman votary, agriculturalist and accountant—reveal an inventory of service castes that put official British households and their lavish princely scale at the center of the Indian social world. A few nonessential groups were thrown in, for example the town priest and the almanac Brahman, perhaps to provide pleasing subjects for the picturesque drawings that adorn each caste type. But the book is an extraordinary artifact of its time, an archaeological finding that cannot be seen in retrospect as merely idiosyncratic. Ramaswamy was no isolated crank but rather a product of one of the most sophisticated efforts to generate local social information about peninsular India in the early nineteenth century. Although Ramaswamy’s Borgesian work sits astride other more “familiar” works, from the Abbé Dubois’s ramblings about Brahmanic ritual practices to the statistical lists of maps and early census documents, it sits in such a way as to disturb any effort to reclaim the early colonial period as a time when caste had the same significance that it came to have only a short time later. Caste was not a unitary signifier; there were no uniform understandings about what might be meant by a treatise on caste—or an inclusive list that might compile more than either a high textual view of some abstract order of things, on the one hand, or the actual distribution of highly differentiated social groups in some locality, or specified context such as the British official household, on the other. Caste was as variable as the Indian social world in the early, still tentative grip of colonial knowledge; it was far from being the comprehensive means for specification of the social order or for the interpretation of the cultural cartography of the subcontinent. Caste, as we have come to know it, did not yet exist.
In this context, the varna classificatory scale appeared useful, indeed necessary, for early colonial commentators on Indian society. Since so little was known about caste relations, it must have made eminent sense to fill in the blanks of ethnographic knowledge by the regular rote references to the Manu Dharma Sastra, as we observed so often in the accounts narrated in Chapter 2. Although the varna view of caste was not of great significance in most explanatory accounts, it served a variety of purposes and came to occupy a canonic—if only marginally relevant—place in early colonial concerns about Indian society. Thus it was that the accounts of Orientalists and Anglicists alike on the subject of caste seemed very similar, even if the emphases and judgments were as different as their respective views on the merits of Indian civilization. And thus we saw such peculiar sediments as the fragile overlay of varna categories on the territorial, revenue, and microsocial classifications of rural Tirunelveli in the 1823 census referred to above. It was not that the varna system was invented at this time but rather that its utility for explaining Indian social relations as a whole became naturalized. As we shall see, even when varna categories were most threatened by the explosion of empirical knowledge in the late nineteenth century, varna came both to signify all caste relations and to explain them in some ultimate sense.
The early imperial archive, the body of knowledge that was cross-referenced by the convictions of colonial common sense and the exigencies of quotidian state practice and concern, reflects the heavily sedimented remains of British interests in the extraction of revenue from land and agrarian regimes. The colonial state established itself in India first as the source of authority for mercantilist trading monopolies and then as an entity that had to negotiate a double legitimation crisis—in both England and India—in the late eighteenth century. Built on greed and conquest, the colonial state had now to justify itself to the interrogation of Edmund Burke, who set public standards for empire that subjected history to the scrutiny of metropolitan moral codes. Burke brought articles of impeachment against Hastings for what everyone knew was business as usual. In Burke’s obsessional litany of Hastings’s excess, what was embarrassing was neither Hastings’s greed nor his methods so much as his manifest success in making the horrors and the pleasures of empire realizable. In the wake of Burke’s attack, a colonial bureaucracy was established to monitor the greed with which all Britons went to India from the late eighteenth century on. Burke shifted the balance of power to the state rather than the mercantile elite, and it was under his scrutiny that the colonial state was born. Colonial rapacity could not be curtailed either by Hastings’s recall or the India Act of 1784, however; it could only be bureaucratized through the high-minded rhetoric of the land settlements detailed in this chapter. British rule represented its interest in securing steady revenue through a language of improvement predicated on the rule of property and the benevolent intent of a new “postdespotic” state.
What Burke neglected was the other side of the legitimation crisis. All the talk of improvement notwithstanding, no Indian public was recruited by (or to) the contradictory logic of colonial sovereignty. Colonial governmentality consisted of bureaucracy without sovereignty, or rather a form of sovereignty abstracted from even the most minimal conceits of political representation. Sovereignty was to perform itself through the extension of landed property and state security. As a result, sovereignty was yoked to bureaucracy most conspicuously in the documentation project of the early colonial state, which created an archive that bears the traces of colonial failure even as it pressed down upon the lives of ordinary Indians with a rapacity that could not have been exacted through direct military or even political means without a vastly larger apparatus of colonial rule. Thus colonial knowledge was far more powerful than the colonial state ever was (a very different balance of power from that which existed in metropolitan forms of governmentality); and thus the colonial documentation project encoded British anxiety that rule was always dependent on knowledge, even as it performed that rule through the gathering and application of knowledge. As I have argued throughout the first two parts of this book, the idea of caste was not yet fundamental to either colonial knowledge or rule. But things would change, especially after the Great Rebellion of 1857 and the assumption of direct Crown rule in 1858. In the historical unfolding of colonial governmentality, caste would emerge as of fundamental importance for the colonial struggle to know and to rule India. From the crisis around state security, as well as from the direct challenge to state sovereignty, came the anthropological idea that caste could be seen as the colonized form of civil society that would both substitute for and explain away the problem of political sovereignty. Ethnography became the primary colonial modality of representation, linking politics and epistemology in a tight embrace for the last century of British colonial rule in India.