PART ONE
CONCEPTS

2
Conceiving Caste

Capitalism and Caste

The nineteenth and twentieth centuries witnessed a sea change in India's commercial and political relations with the rest of the world—a change that, in Wallerstein's (1974, 1980) terms, "incorporated" and "peripheralized" India in a British-dominated world system. To say this, however, is not to accept auxiliary allegations about the noncapitalist qualities of India's precolonial political economy or other allegations about the locus of agency and the role of Indian class interests in the colonial transformation. All such claims seem to me to deserve a thorough rethinking. Yet, although much of what I have to say in the present book bears on these topics, I will not take the opportunity to comment on them. Instead, I shall focus on an additional claim, rooted in the classic sociological theories of Marx, Weber, and Simmel: namely, the claim that any capitalist transformation entails (as cause, consequence, or both) a form of commodity fetishism and objectification of labor that brings about the dissolution of social groups and the creation of alienated individuals.

The exploitative essentialism and evolutionism inherent in these and most comparisons between modern Western society and every other society remain dominant in the social sciences and have recently received considerable critical attention (Chandravarkar 1985; Fabian 1983; Fuller 1989; Inden 1986, 1991; Said 1978). This book participates in that criticism by taking issue with the underlying assumptions behind orientalist interpretations of Indian political economy—whether classic or contemporary, Marxist or a product of modernization theory. Toward this end, the present chapter considers the implications of Nakarattar caste structure and role in

Indian society both for standard theoretical treatments of caste and for generalizations about caste drawn from case studies of other castes.

The Standard View of Caste

"Standard" anthropological views of caste were well established by the 1950s and 1960s. By and large, they were shaped by studies of dominant agrarian castes. Some studies focused their attention on peasant cultivating castes, others on warrior or robber castes, and still others on nondominant, landless laboring castes. Almost no studies focused on mercantile castes. From the standard but agrarian perspective, castes were held to be commensal, endogamous groups, ritually ranked with other castes in consideration of local standards of relative purity. They were also held to be economically specialized with respect to their members, who were, therefore, interdependent with members from other, differently specialized castes (Dumont 1980 [1970]; Leach 1960; Marriott 1976; Srinivas 1962). Indigenous conceptions about varna (the religious ideal of the caste order) were seen, in Geertz's (1973) terms, to act as models of and for interaction between members of different castes within villages (Dumont 1980 [1970]; Marriott and Inden 1974; Rocher 1975). Concomitant conceptions about kinship were seen to regulate interaction between members of the same caste who lived in different villages (Dumont 1980 [1970]; Mayer 1960). All of these characteristics were supposed to be maintained by the corporate organization of castes within traditional, politically defined regions referred to as chiefdoms or '"little kingdoms"—at least in South India (Dumont 1980 [1970]; Miller 1954; Srinivas 1962). Each corporate caste was governed by the operation of its own caste panchayat: a decision-making body composed of elder caste members whose purpose was to establish and sanction castewide standards of behavior.

Within this standard view there were disagreements about the relative importance of economic interdependence versus competition, of politico-economic power versus religious ideology, of attributional versus interactional systems of ritual ranking, and of other important considerations in the understanding of Indian society. Despite these disagreements, there was broad agreement that castes played a fundamental role in the lives of Indian people. For most social scientists, caste was the characteristic unit of Indian social organization: Indian society was caste society.^[1]

With this general tenet, I have no disagreement. Nevertheless, it is a central theme of the present study that recent studies of mercantile castes including the Nakarattars require a radical revision of the standard concept of caste. No such proposal occurs in a vacuum. On the contrary, a general dissatisfaction with the standard view has given rise to various calls

for its revision during the last few decades. The remainder of this chapter reviews an important segment of this revisionary work, setting the stage for the substantive analyses of Nakarattar caste organization to follow.

Caste as Symbolic Capital in Political Ethnicization

Recent historical studies of caste groups carry out a radical reconceptualization of the institution of caste and, by implication, raise serious questions about the adequacy of the standard social science view. I refer, in particular, to analyses that fall under such rubrics as the "substantialization," "ethnicization," or "racialization" of caste: the alleged twentieth-century transformation of caste from one kind of social formation into another. Research on this phenomenon raises questions, both about the kind of social formation caste is taken to have been prior to its transformation and about the kind of social formation it is said to be in the process of becoming. In both cases, the most frequent answers reflect a distinctive analytic bias.

Historical treatments of the changing nature of caste appear most prominently in connection with more broadly defined historiographic concerns about the growth of Indian self-rule prior to independence (Gallager, Johnson, and Seal 1973). Within this context, it is not surprising that historians have focused their attention on the twentieth-century emergence and role of caste associations: sodalities that seem to integrate aspects of caste organization with aspects of organization in Western-style political parties. The growth of these sodalities marks an apparent watershed in the way members of subcastes sharing a common name interact with one another, and this change has led influential historians of South India (including Baker 1976; Stein 1980; and Washbrook 1975, 1976, 1982) to raise important doubts about the putative corporate status of castes, the collective action of castes, and sometimes, it almost seems, the very relevance of caste groups within South India, at least for any level of territorial organization larger than a village.^[2]

As an alternative, these radical historians argue that, prior to the late nineteenth century, South Indian castes constituted unintegrated clusters of localized, endogamous subcastes that had little in common except name. This view of caste is one that follows in the wake of an anthropology that emphasized village studies without giving adequate attention to extravillage and regional social phenomena. Because of the absence of regional caste studies, and because of their predisposition to look for the impact of caste in the Indian independence movement, some radical historians have based their findings on investigations of the twentieth-century development of caste associations formed for political purposes (e.g., Hardgrave 1969; Rudolph and Rudolph 1967).^[3]

If I understand the arguments of these historians correctly, South Indian caste identity has gradually been shaped into a vague umbrella concept that covers a range of potential practical and moral linkages, but that entails no specific rights or obligations among people who share a common caste name. Rather, caste identity defines a category of people who are candidates for alliance in virtually any common cause. They do not exist as permanent corporate groups, but they can be mobilized in temporary political factions (Boissevain 1968) or quasi groups (Mayer 1966) on an occasional basis by powerful politicians pursuing individual ends. Caste identities, in this view, constitute a form of what Bourdieu (1977) called symbolic capital , available for investment by politicians with the skill to manipulate its meanings.^[4]

The most elaborated version of this position is represented in a series of papers and books by David Washbrook (1975, 1976, 1982, 1984). Developing a sophisticated comparative model of ethnicity and social stratification, Washbrook carefully orients himself within the luxuriant literature about Indian castes. For example, with respect to sociological and historical theories about the "modernity of tradition" (Rudolph and Rudolph 1967), Washbrook sees postconquest and colonial caste associations as the results of novel manipulations of traditional symbols. At the same time, Washbrook is too good a historian to posit a timeless tradition. Rather, along with scholars such as Cohn (1960, 1983, 1987) Carroll (1978), and Dirks (1987), he theorizes that "traditional" Indian symbols and ideas about caste may themselves have been recent inventions, born out of interaction between European conquerors and indigenous elites. In this view, the apparent nineteenth- and twentieth-century modernization of caste represents a continuation of seventeenth- and eighteenth-century processes.

Washbrook is also careful to place his viewpoint in relation to more culturally based theories of caste. Thus, despite a sustained critical attack against the theories of Dumont (1980), he shares a number of Dumont's interpretations of Indian society, especially as elaborated by Stephen Barnett (1973). Like Dumont and Barnett, Washbrook finds that precolonial caste identity fits more or less closely with what I have characterized as the standard model of caste society. In particular, he holds that precolonial caste membership applied only disjointedly to distinct social identities within localized caste hierarchies. The social obligations of membership were wholly specified by the local situation and were not generalized to caste members belonging to regionally disbursed subcastes sharing nothing other than a common name. By contrast, the "substantialized," "ethnicized," or (in Washbrook's terms) "racialized" systems of caste identity in colonial and independent India have reversed this emphasis. That is to

say, subcastes have come to be identified on the basis of kinship within a commonly named caste cluster; clusters as wholes are ranked with respect to legal and ritual prerogatives that are sanctioned by pan-regional, governmental authorities; and finally, interactions between members of the same local subcaste and between members of different subcastes sharing a common caste name are free from all caste-defined regulations, except those preserving endogamy.^[5]

Washbrook marshals a number of arguments that complement those of Dumont and Barnett. These, however, are incidental to his primary interest in explaining "the rise of ethnic and racial sentiment" as "a characteristic response to certain kinds of economic and political instability" (1982: 153). In his view, the evolution of a caste cluster into an ethnic group within a racially stratified system reflects a complex political process. Elite cluster members create and "sell" a myth of cluster solidarity to the emerging colonial government, position themselves as brokers of political influence for members of their caste clusters, and employ these newly created patronage resources to build a base of political support.^[6] As the multicaste members of caste clusters cooperate to support an elite member, they create temporary political factions whose memberships change with the political fortunes of their leaders. Such cluster-based factions exemplify regional caste units that never existed in pre-colonial India. Their members disregard local variation and conceive of themselves in substantialized, ethnicized, and racialized terms that are equally novel and that, in fact, resemble Western-style ethnic or racial groupings. In Washbrook's apt restatement of Srinivas, "the djinn (of caste) was released from the bottle [of little kingdoms].... All hell broke loose as society began to form itself into 'castes' which, so far from accepting the hierarchy of the varna scheme, began furiously to contest their own places within it" (1988: 26–27).^[7]

"Where"s the Rest of Me?" The Case of the Komatis

My own research persuades me that the caste-as-symbolic-capital interpretation of Indian political society has considerable merit, but that it also has a markedly distortive Western accent. In the first place, it restricts its focus to the way that Western -style political activities are structured by, and structure, ideas of caste. In the second place, it dichotomizes social organization into Western-style corporate groups and Western-style political factions in a way that is quite insensitive to a variety of South Indian , caste-based social forms that fit neither category. In other words, the doubly accented, Western interpretation of caste falls to consider the likely possibility that non-Western, caste-based social forms might occu-

py South India's social landscape. A closer look at some of the details in Washbrook's conception of caste as a politically mobilized ethnic group will illustrate my concerns.

To begin with, Washbrook's elaborated discussion significantly qualifies the basic ethnicization thesis by applying it primarily to case studies of various mercantile and service castes. According to Washbrook (1975: 164) these castes exhibited important organizational differences from the castes of their agrarian brothers, differences that made factional organization at the provincial level an attractive and feasible political strategy that was unworkable for agriculturalists.

Washbrook illustrates his model of the ethnicization of merchant castes by analyzing the historical transformation of the Komati Chettiars, a merchant-banking caste whose members dominated South Indian commerce in their role as European/Indian brokers until the mid-nineteenth century (Suntharalingam 1974). The Komatis constituted a highly segmented caste cluster split by religious-sectarian, occupational, territorial, and affinal boundaries (Thurston 1909 III:306–348; Washbrook 1975: 152–159). Washbrook cites some meager evidence for inter-regional cooperation, such as use of a common trade language (Thurston 1909 III:308) and a common cross-cutting system of exogamous clans which, in principle, could have allowed inter-subcaste marriage exchanges (312–314). But the only hard evidence that suggests inter-subcaste cooperation within the cluster as a whole is the presence of a few elite families operating throughout the Madras Presidency from the sixteenth to the nineteenth century.

It would clearly be wrong to argue from the absence of information about Komati organization that the Komatis therefore lacked pan-regional organization, especially without any systematic effort to establish what kinds of cooperation may have existed. Yet, having asserted that Komatis were commercially organized throughout South India, Washbrook proceeds to reject every possible social mechanism for making such organization work and, in so doing, contradicts his thesis that mercantile castes were distinctively preadapted to twentieth-century ethnicization. According to Washbrook's model, Komati marriage was restricted to localized subcastes; collective worship of the caste goddess at a shrine in the Komati homeland, north of Madras, had no economic, political, or other discernible social implications; and, finally, the only Komatis who truly conducted commerce on a pan-regional scale were elites, who, according to Washbrook, had separated themselves from specialized commercial interaction with their fellow Komatis in order to pursue alternative and somehow incompatible commercial opportuni-

ties offered by the developing British East India Company and colonial economies.

These propositions, as I have said, run contrary to Washbrook's important insight about the preadaptive qualities of mercantile caste organization for twentieth-century caste associations. Yet, because issues outside the politicized organization of castes are secondary to his major arguments, Washbrook does not explore the possibilities for nonpolitical forms of subcaste cooperation or the possibilities for elite-nonelite interaction. As a consequence, it seems to me that it is premature to judge the case of the Komatis. Although Washbrook claims that they and other mercantile and service castes successfully form caste associations because of pre-twentieth-century organizational characteristics unique to mercantile and service castes, in the end, he never really identifies the relevant preadaptive organizational features and therefore implicitly suggests that they may not have been distinctive after all.^[8] The present study agrees with Washbrook's original suggestion that mercantile caste organization was distinctive. And although I am unable to offer any further information about the Komatis, my study of the Nakarattars, the merchant-banking caste who succeeded the Komatis as dominant players in the Indo-European credit market, explores precisely those properties of caste organization that were characteristic of a mercantile specialization. There remains, however, one more lesson about caste that I wish to draw from Washbrook and from one of his critics.

Segmentary Structure and Structural Variation

The criticism I have just made of Washbrook's treatment of caste-based political formations in the twentieth century bears a superficial similarity to criticism that Michael Roberts (1982) has also aimed at Washbrook. According to Roberts, Washbrook is guilty of a "sociological naivete" that imparts a "misplaced concreteness" to the study of caste (1982: 194). Roberts means by this that Washbrook ignores the social implications of "cultural typifications" about caste (1982: 194, 202–204). In particular, Roberts argues that Washbrook fails to consider the fluid, segmentary organization manipulated by a caste's membership (1982: 197).

I agree with Roberts that Washbrook does not give adequate consideration of non-Western forms of pan-regional social organization. I disagree with Roberts's claim that Washbrook is unaware of the "cultural"—in contrast to the sociological—nature of typifications about caste. On the contrary, the central point of Washbrook's interpretation is precisely the flexible political use of cultural typifications. Where Washbrook goes wrong, or so it seems to me, is not in any failure to appreciate either the

cultural nature of caste as cultural typification or the fluidity of caste-based political structures, but in limiting attention narrowly to a single use of typification: the construction of Western-style political coalitions. In Washbrook's restricted view, symbols of caste identity simply fail to operate in the organization of pan-regional social groups, except in temporary nineteenth- and twentieth-century political coalitions. In other words, although both Roberts and I agree that Washbrook does not adequately appreciate the segmentary nature of caste formations, my criticism of Washbrook is that he ignores important social variables, not that he ignores cultural variables.

From my view, caste identity has been and continues to be an important (though certainly not the exclusive) form of symbolic capital invested in the segmentary production of a variety of South Indian social formations. To recognize the segmentary nature of caste-based formations, however, marks the beginning of investigation—not, as Roberts seems to suggest, the end. Different castes face different opportunities, constraints, and threats in different regions and at different times. They invest their caste identities accordingly, producing no uniform segmentary structure but instead varieties of segmentary structures, structured by—and structuring in turn—the actions of their members.

The "Corporacy" of Castes: Kayasths and Nakarattars Compared

Interpreting caste as a form of symbolic capital raises a number of definitional questions about the nature of corporate and quasi-corporate groups in which the capital of caste is invested.^[9] To this end, it is helpful to consider Karen Leonard's (1978) important study of the Kayasths of Hyderabad. Less radical in its position than those of Baker, Stein, or Washbrook, Leonard's study argues that although castes may not constitute corporate wholes in themselves, their membership may nevertheless be organized in a variety of supervillage and even superregional corporate units. In particular, Leonard indicates that actively cooperating caste members most commonly form corporate units at the level of families, lineage segments, marriage networks, or restricted bilateral kindreds, all of them operating within multiple urban contexts. She points out that, in any given historical period, a caste is characterized by a typical constellation of such component kin groups, each of which conducts distinctive economic and political activities. These constellations of kin groups vary according to the conditions of the wider politico-economic environment. She describes (1978: 15) the following progression of constellations in the case of the Kayasths:

In the eighteenth century military aristocracy, an individual and his immediate family were the important social unit. In the nineteenth century, kin groups formed, and these small groups of closely related families controlled hereditary positions within the Muglai administration. Due to the educational and administrative changes of the late nineteenth century, some Kayasths organized on the basis of common sub-caste membership; in the twentieth century, they have organized on the basis of common caste membership.

Notice that, in at least one important respect, Leonard's view of caste supports those of Baker, Stein, and Washbrook. Excepting only a brief period of political organization at the beginning of the twentieth century, Leonard claims that not even Kayasths have operated in a corporate fashion on a caste level (i.e., by organizing the unrestricted kindreds comprising the Kayasth caste cluster). She differs from the three South Indian historians in presenting a much richer picture of multilocale, intracaste organization. Nevertheless, not even her enriched model of caste organization alters the unified picture of internal caste politics presented by all of the historians under review. In particular, no constellation of kin groups seems determined by any decision-making body acting in the interests of the whole. Nor does there exist any such body having powers or rights to act with regard to any other matter of collective concern. The caste cluster as a whole is simply a culturally distinguished set of individuals and groups—a kindred of recognition in contrast to a kindred of cooperation (cf. Mayer 1966). It is a cultural category designating an aggregate of people with the potential for participating in useful but less inclusive forms of social organization.

It should be pointed out (Leonard's conclusions notwithstanding) that even if our analysis is confined to political organization, the absence of a centralized, decision-making body operating within either the kindred of cooperation or the kindred of recognition is not a sufficient condition for withholding recognition of a subcaste's or a caste cluster's corporate status. This is especially the case when subunits of a caste (such as the Kayasths) exhibit traits of segmentary organization similar to those that have been well analyzed in anthropological studies of African political systems. Without further study of what may prove a wide variety of political mechanisms, the absence of effective castewide panchayats should not be taken as conclusive.

Nevertheless, assuming that Leonard's evidence would have reflected Kayasth corporate organization if it had existed, her study offers us an important contrast with the Nakarattars of colonial south India. To begin with, Nakarattar identity was unexceptionably based on jural principles of

kinship and was regulated by the operation of temple-clan committees (see Chapter 9). Secondly, Nakarattars as a body made a variety of economic, ritual, and political claims upon each other in virtue of their common caste identity. Moreover, they traditionally and often successfully claimed special rights in the administration and execution of government in those states in which they played a commercial role. Finally, Nakarattars exercised a variety of collective rights over their ancestral homeland of Chettinad. It is true that the Nakarattars were politically acephalous—that they were not organized by any single institution such as kingship or a caste panchayat . But this did not prevent the caste from making collective decisions about a wide variety of issues and disputes through a system of institutions that embodied both a nonhierarchical, crosscutting form of segmentary organization and also considerable centralization of power vested in the hands of a relatively small body of elite Nakarattars (adathis ).

All this, however, may be besides the point, for Nakarattar caste organization is best viewed as primarily economic in orientation rather than political. For individual Nakarattars, for their families, and for the caste group as a whole, it was commerce, not politics, that dominated almost every consideration and every action. Commercial considerations affected institutions that regulated practices ranging from philanthropy and politics to marriage and adoption. Conversely, the net effect of institutional adaptation to the needs of business was to render the caste especially efficient in the task of capital accumulation and distribution. In the case of the Nakarattars—particularly in their nineteenth- and twentieth-century incarnation as merchant-bankers—there is little room for doubt concerning the putative corporate status and collective action of the jati across village, regional, and even international boundaries.

In sum, although I endorse many recent historiographic interpretations of caste in its modern and flexible embodiment as a form of political coalition, I disagree that ethnicized and substantialized caste factions are the only forms that caste groups take outside the level of the village. At the very least, for Nakarattars and other large-scale merchant-banking castes (and I suspect for many landholding and military castes as well), such relatively recent interpretations of caste, like the more standard anthropological views they superceded, ignore a variety of corporate institutional arrangements between caste members. This conclusion in no way contradicts political analyses of castes as constituting—among other things—potential factions occasionally mobilized by politicians. It does indicate that factional mobilization represents only part of the organization and functioning of caste in Indian society; in particular, an exclusive focus on

factional politics ignores the economic resources that castes provide their members and the ways in which these resources are organized. By contrast, if economic variables are taken into account, I suggest that castes will typically emerge as complex, multilayered, multifunctional corporate kin groups with enduring identities, a variety of rights over property, and crucial economic roles, often within large regions. This is precisely the conclusion that can be drawn from recent historical analyses of agrarian castes, as will be addressed in the following section.

Ecologically Qualified Views of Caste

The standard view of caste society deals spectacularly poorly with the on-the-ground diversity of India's caste systems. No "little kingdom" in India has ever been the same as any other. Some have been governed by authentic kings. Others have been governed by assemblies representing important clans of a dominant caste. Some little kingdoms incorporate as many as fifty castes and include a substantial "untouchable" population. Others comprise only a single caste and structure their social relationships primarily in terms of kinship. The importance of ritual purity as an ordering principle for social relations varies widely; and so does its concrete expression in articulated prescriptions and sanctions on behavior. All of these variables, and many more as well, threaten to render the standard view of caste meaningless, unless some order can be found that links the cultural capital embodied in the structure of historically specific castes to the opportunities for investing this capital available in specific historic environments.

Two recent works on South Indian rural history make significant progress in such an inquiry: David Ludden's Peasant History in South India (1985), and Christopher Baker's An Indian Rural Economy (1984). Ludden's and Baker's descriptions of the agrarian castes of South India arise in the context of an inquiry into agricultural production and land revenue, and they are only indirectly concerned with the conceptualization of caste. Nevertheless, I have found their treatment of caste enormously helpful in my own work. Using statistical information culled from early revenue reports, censuses, and government reports on land, climate, soils, agricultural techniques, rural credit, and innumerable other topics, Ludden and Baker construct a socioecology of caste. Baker's regional division of Tamil Nadu into valleys, plains, and Kongunad is inspired by Ludden's regional subdivision of Tirunelveli District into wet-farming, dry-farming, and mixed-farming zones. Both map the distribution of dominant agrarian castes onto their respective rural landscapes and suggest a link between modes of production and modes of political domina-

tion: namely, that kinship-based "tribal" formations control resources in the dry-farming plains, and kingship-based "ritual state" formations control resources in the wet-farming valleys. To put this another way, they link differences in the segmentary structures of different castes to the goals of caste members as they try to exploit their natural and social environments.

Ludden and Baker are, perhaps, somewhat insensitive to variations in caste segmentation that social anthropologists have found significant throughout the world and that seem operative in the Tamil countryside. They contrast the structure of royal honors and service in the wet-farming valleys with the segmentary lineage structure of the dry-farming plains, but they fail to consider the kinship structure of the valleys and the kingship structure of the plains—an omission all the more remarkable since the best available studies of Tamil kingship concern precisely the little kingdoms of the plains and their integration with kinship organization (Dirks 1988; Dumont 1957a, 1957b; Price 1979). Nevertheless, the lesson that I draw from their studies concerns, again, adaptive variability in the social organization of different castes. My analysis of the Nakarattars attempts to complement their findings by characterizing a mercantile caste that is territorially dominant in its residential homeland, but which has a resource base that lies neither in agrarian production nor in military control.

Mercantile Caste Variation: Some South Indian Dimensions

To say that a particular caste functions as a regional or pan-regional commercial organization is not to claim that all commercially organized castes function in the same way. Different castes specialize in different forms of trade, characterized by location, geographic scope, and types of commodities traded. The geographic distribution, internal organization, standards for behavior, and resource opportunities of the castes vary accordingly. In other words, however much Indian businessmen might exhibit class resemblances, they differ from one another with respect to the resources available to them through their castes, the roles they play in their castes, and the roles their castes play in Indian society. Put concretely, Nakarattars did business differently than Komatis, who did it differently than Nadars, who did it differently than did the multiple species of South Indian Chettis.

The exact nature of these differences is still very much at issue. The two best-known studies directed at the historical roles of mercantile castes throughout India—Hardgrave's (1969) study of Nadars, and Timberg's

(1978) study of Marwaris—do not provide the kind of information needed to carry out a systematic comparison of the forms of mercantile organization in question. Hardgrave focuses on politics and cultural identity; Timberg, on the historic and geographic diaspora of the Marwaris. Neither provides a detailed analysis of social organization. Under the circumstances, results from my own investigation can only suggest avenues for future comparative research. Nevertheless, even such preliminary observations indicate important dimensions of variation between the differently specialized commercial groups located, in this case, on South India's historical, commercial landscape.

On one hand, for example, Nakarattars were unlike many of their brethren Chettis, who conducted small-scale wholesale and retail trade in groceries, edible oils, jewelry, and other commodities. Such "petty" Chetti castes tended to operate their businesses within a small radius around their native towns or villages.^[10] By contrast, Nakarattars conducted their business activities well beyond their residential homeland. In this respect they resembled some other pan-regional mercantile castes—including, for example, both the merchant-weaving Kaikkolars (or Sekuntar Mudaliars) and the grain-trading Nadars.

On the other hand, Nakarattars differed from both of these large-scale mercantile castes in a variety of other respects. To begin with, Nakarattars carried out their commercial activities over a much larger territory than that covered by even the Kaikkolars or Nadars. Both of these were confined by and large to the Tamil-speaking regions of South India. Nakarattars, however, conducted business as far north as Calcutta, and the bulk of their dealings actually took place in British Southeast Asia. In addition, unlike the merchant-artisan Kaikkolars or the grain-trading Nadars, Nakarattars acted primarily as bankers, extending financial credit, discounting bills of exchange, and receiving deposits—in short, performing all the functions of a modern bank. In other words, unlike the other two pan-regional mercantile groups, Nakarattars kept themselves relatively free from the requirements of commodity production or processing, especially the necessity imposed by activities of production to tie up resources in relatively fixed capital investments. Nakarattars much preferred to keep their resources liquid and thus concentrated their activities primarily in financial operations.^[11]

A Closer Look at Merchant-Artisans: The Case of the Kaikkolars

It was not feasible for me to collect detailed information about concomitant variation in the social organization of most of the differently special-

ized commercial castes. Fortunately, the Kaikkolar caste, as the subject of an ethnohistorical study by Mattison Mines (1984), represents an unusual exception to the general lack of scholarly attention given the organization of commercial castes. Accordingly, it becomes possible to contrast Kaikkolar organization with that of the Nakarattars. In his study of Kaikkolars, Mines is primarily concerned with the spatially segmented organization of their "seventy-two-natu " system extending throughout Tamil Nadu, and the contrast between this form of organization and that of agrarian castes.^[12] Within the seventy-two-natu system, interaction between Kaikkolar caste members tended to be territorially segmented into natus: small local regions where sites of descent group formation and marriage alliance coincided with the sites of business activity. In fact, notwithstanding the pan-regional scope of commerce represented by the Kaikkolar caste as a whole, individual traders operated within their local territories much as did the small-scale Chettis, forming what amounted to microcastes, sub-jatis , or even segmented, endogamous jatis within the Kaikkolar caste cluster. They extended the total range of their caste's operation by maintaining a network of ties between different local segments: in pre-twentieth-century days through their ancient seventy-two-natu system, and more recently through the operation of a modern caste association (Mines 1984). Thus it was the caste as a whole that conducted business beyond the range of the disbursed and discrete residential homelands of its individual members. Caste members as individuals, however, tended to merge residential and business domains.^[13]

As clearly demonstrated by Mines, the difference between Kaikkolar and agrarian castes is based on their differing economic specializations. Kaikkolars produced textile commodities that were easily transportable and that, in fact, have been the object of international trade for a thousand years or more. By contrast, agrarian castes specialized in production for local consumption. Only recently have they had the means for transporting their bulky and perishable produce. For agriculturalists, there was never any question of combining production with long-distance commerce. For the Kaikkolars, without long-distance commerce there would have been no justification for production. The pan-regional Kaikkolar natu system established standards of quality for textile goods, provided its members with market intelligence, adjudicated conflicts over production and market monopolies, and interceded with local governments on behalf of its members. There was no comparable corporate organization that tied together agricultural castes across spatially dispersed residential communities.

Mines views the division of Indian society into agriculturalists and merchant-artisans as breaking down in the twentieth century under the

impact both of technological innovations that facilitate transport of agrarian produce and of an altered political system that discourages caste organization in favor of organization by industrial sector. But, he argues, the kind of dual organization exemplified by agriculturalists and Kaikkolars represented the norm for all of South Indian society for at least a thousand years prior to the 1920s.

The Merchant-Banker Nakarattars Contrasted

On the whole, I believe Mines is premature in constructing a simple dual model of agriculturalists and merchant-artisans for the entire range of South Indian society. In particular, such a model ignores important variations in the conduct of South Indian commerce. My study of the Nakarattars, for example, identifies aspects of their social organization that directly contradict the central features of Mines's mercantile model. I emphasize, however, that my findings with regard to the Nakarattars do not have implications for Mines's analysis of the Kaikkolars, but only for his generalizations to all mercantile organization. Notwithstanding any disagreement over the scope of Mines's conclusions, I remain in total agreement with his identification of the topic of commercial organization as a crucial and neglected topic in Indian studies. His study represents an important contribution to the field and provides useful comparative material for understanding variation between mercantile castes.

For example, there are important organizational differences associated with differences in the economic specializations of Kaikkolars and Nakarattars. In contrast to Kaikkolars, Nakarattars conducted their business activities in "business stations" that could lie hundreds of miles beyond their Chettinad homeland. Why the two groups should differ in this regard is not clear. It seems reasonable, however, to suppose that there was a functional link between occupation and organization. The production and processing of trade commodities such as textiles was conducted more efficiently with the participation of the domestic group (the wives and children) of merchant-producers than it could have been without the group's participation. At the same time, Kaikkolar concern for ritual purity of the endogamous group operated to restrict the distance that the combined domestic and producing unit could travel. By contrast, Nakarattar merchant-bankers traveled long distances to strange lands. But their banking operations could operate just as efficiently with paid male employees as with labor supplied by the merchant's family (see Chapter 6). Concomitantly, Nakarattars satisfied their ritual concerns with purity by leaving the domestic group behind in the circumscribed homeland of Chettinad.

Whatever the truth of these speculations, the presence of pan-regional commercial groups, organized in the ways I have ascribed to, respectively, merchant-producing castes and merchant-banking castes, extends back in history to a period when guilds rather than castes were the primary institutions for long-distance trade. A thousand years ago, guilds termed manigrammams exhibited a split between their residential homelands and their circuits of trade that is similar to the more recent organizational pattern of the Nakarattar caste. In the same period, other guildlike groups, functioning as municipal councils for market towns, exhibited the territorial segmentation of present-day merchant-producing castes such as Kaikkolars or Nadars.^[14]

In other words, the different organizational patterns exemplified by Nakarattars and Kaikkolars cannot be interpreted as modern adaptations to South Indian involvement in the colonial economic system. Rather, my proposal is that longstanding cultural concerns with status purity combined with the exigencies of business operations and the political influence of local and regional rulers to produce two separate organizational structures. Moreover, these two distinct structures transcend the institutional differences between castes and guilds.

Other differences in the internal organization of mercantile castes cannot be traced back in history to the same extent, but they are no less significant. For example, not even within their homeland of Chettinad did the Nakarattar jati develop a territorially segmented system of microcastes like the one developed by the Kaikkolars. There did exist what might be recognized as incipient microcastes in the division of Chettinad into sub-territories called vattakais (Chapter 8). But membership in temple-clans created links of association that cut across tendencies toward vattakai endogamy and precluded segmentation of the caste into localized kin groups beyond the level of village-bound lineage segments. (Chapters 9 and 10). Similarly, commercial activities in Nakarattar business stations reflected the crosscutting divisions of their kinship and territorial organization in Chettinad. That is, Nakarattars from the same village, vattakai , or temple-clan might conduct business in different business stations. Conversely, Nakarattars operating in the same business station might belong to entirely separate villages, vattakais , or temple-clans.

Again, it seems reasonable to suppose that Nakarattar social organization represented an adaptation to the exigencies of their specialization in banking activities. In particular, the crosscutting nature of different Nakarattar kin groups may have operated to maximize the flow of information and credit extension throughout the entire caste, without any impediment from a noncrosscutting, segmentary system like the Kaikko-

lar system, which fostered a corresponding division in the production of specialized textiles.

All of these suggested associations between business specialization and social organization are highly speculative and require systematic comparative study before they can be accepted with certainty. Moreover, the issue becomes more problematic the farther one looks back in time. For example, only two to three hundred years ago, Nakarattars specialized in a geographically localized salt trade rather than in colonial-era banking operations. But this change in occupation may not have involved adaptations to the kinds of considerations suggested above. In particular, as far as we can tell, Nakarattars were never involved in the production of salt, but only facilitated its transport between producers and consumers. Furthermore, it seems as though the social organization which underpinned the Nakarattars' colonial banking operations was already in place. Yet both of these conclusions are based on indigenous records (in the form of palm-leaf manuscripts) maintained at a Nakarattar site of pilgrimage and trade (cf. Chapter 7). And these records deal only briefly and indirectly with the topic of trade and social organization.

Symbolic Capital and the Cultural Ecology of Caste

What implications do these various reflections about diverse castes have for anthropological and historical conceptions about caste in general? Perhaps the ultimate implication is that although India is a caste society, requiring (even today) analysis in terms of its constituent castes, there is no single thing properly identified as a caste. There are individual castes formed by unique histories of investments and transactions in symbolic capital. There are local configurations joining several different castes, similarly formed. Each caste and each configuration bears some family resemblance to every other caste and configuration. But as Srinivas (1952) showed us long ago, the similarity in the relationships between castes and their local configurations covers an enormous range of diversity: from the elaborate, multicaste villages of rice-growing river valleys to the single-caste villages of "tribal" groups in the hills. More recently, as studies by Dumont, Barnett, the Rudolphs, and Washbrook have shown, castes have become ethnic groups subject to manipulation by regional politicians and mass media. Within this diversity, some castes and configurations are undoubtably more similar, in important respects, to social groupings and configurations outside India than they are to local caste systems several steps removed in the Indian chain of resemblances.

Yet, if there is no single thing that can properly be called 'caste' in India, there are, perhaps, factors of mind and society that influence the

social and cultural ecology of different castes, including factors that consist of opportunities for economic investment in a capitalist world. The existence of such factors is what I try to suggest in this book by looking at an Indian caste both like and unlike other castes from which we have tried to abstract a unitary definition in the past. The following chapter, however, takes up another conceptual difficulty no less problematic than that provided by Western definitions of caste—one that, in fact, complicates and is complicated by those definitions.

3
The Study of Commerce in Indian Society

Giants on Our Shoulders

A major difficulty standing in the way of adequate historical understandings of Indian commerce is that we labor under the burden of past misunderstandings.^[1] For example, despite contemporaneous and near-contemporaneous accounts to the contrary (Buchanan 1870; Crawfurd 1971[1837]), most studies of India's precolonial economy have, until recently, simply assumed an agrarian and noncommercial system. Historians who were concerned with economic issues at all directed their efforts to a debate about India's economic development (or underdevelopment) during the nineteenth and twentieth centuries. On the one hand, nationalist historians of India argued that colonialism stifled Indian development through unfair taxation and trade regulations. On the other hand, proponents of Western modernization theories argued that colonialism fostered development by providing a peaceful environment and by promulgating the growth of efficient transportation, communication, education, and government. But both sides shared the standard social science models of precolonial Indian economy developed by Marx and Weber. These classic models divide India into autonomous peasant villages and oriental despots who siphon off agricultural surplus. They characterize India as lacking a commercial life worthy of study, except insofar as she provides an example of a "premodern" society incompatible with commerce. They deny the existence of commerce and hence of institutional involvement in commercial activities.^[2]

More recent scholars place India's nineteenth-century economy into perspective by examining seventeenth- and eighteenth-century docu-

ments connected with mercantile and "protocapitalist" systems of the period (Bayly 1983; Das Gupta 1967, 1970; Habib 1969, 1980; Leonard 1979; Mendels 1972; Pearson 1976; Perlin 1983; Subrahmanyam 1990). The result is a developing revisionist view of India's precolonial economy. It envisions a significant manufacturing power, producing perhaps a third of the world's manufactured goods—principally textiles (Washbrook 1984). In this view, the colonial transformation had two major effects. It destroyed India's textile and other manufacturing industries, and it subverted her "traditional," noncommercial systems of agricultural production by forcing agriculturalists to produce exportable crops at the expense of food crops and subsistence agriculture (Habib and Raychaudhuri 1982; Neale 1957; Thorner 1960). In other words, some scholars now see colonialism as having destroyed India's precolonial industrial capabilities and created a colonial economy whose narrow basis in agricultural production was, ironically, to become the model for mythical reconstructions of her precolonial past.

Despite considerable progress in demythologizing the concept of the traditional Indian economy, most of the revisionist work directs its attention toward aggregate measures—of trade volume, terms of trade, or quantity of money. (Deyell 1970; Habib 1982; Hasan 1969; Moosvi 1980; Prakash 1976; Prakash and Krishnamurty 1970). Excepting only some of the most recent studies, such as those by Baker (1984), Bayly (1983), and Subrahmanyam (1990), almost no one addresses the specific institutions that were agents of Indian commercial activity in the colonial or precolonial periods.^[3] For the most part, our understanding of India's precolonial social formations retains many traditional assumptions about India's noncommercial character. This conservative perspective is particularly noticeable in work influenced by formulations of neo-Marxist and substantivist economic anthropologists. By and large, proponents of these schools of historical interpretation arrive at their conclusions deductively, through evolutionary or dialectical theories about the rise of capitalism. From their perspectives, only evolutionarily advanced societies whose economic activities are organized by contract law, commodities markets, stock markets, central banking systems, and other institutions characteristic of Western commerce are capable of conducting commercial activities. Conversely, the absence of Western institutions is seen as the hallmark of a precapitalist society and as logically equivalent to—or at least providing prima facie evidence for—incompatibility with activities oriented around private property, market production, capital investment, and credit extension for long-distance trade.

None of these revisionist perspectives diverge from the basic premise that India lacked indigenous commercial systems. Excepting the subversive intrusion of the capitalist system into commodity production, no other remotely commercial activity figures in neo-Marxist and substantivist accounts of the colonial period. Any precolonial evidence of commercial activity is interpreted by a Procrustean set of categories whose chief characteristic is that they are the opposite of commerce. If goods are produced, it is for purposes of "subsistence" rather than exchange. If goods are exchanged, it is a transaction of "prestations" rather than a buying and selling of commodities. If commodities are bought, it is for purposes of "consumption" rather than investment. If commodities are valued, they are said to have a "use value" rather than an exchange value. If profits from an exchange are saved, the savings constitute a "primitive hoard" rather than an accumulation or reserve of capital.

Neo-Marxist and substantivist analyses of Indian economy share an additional property besides their doctrinal similarity. They are remarkably ahistorical. Pitting their analyses of contemporary or colonial India against an idealized model of "traditional" India, proponents ignore evidence of extensive commercial activity extending back to the third or fourth century B.C. (Maloney 1970; Thapar 1966). Perhaps this is only to be expected. Only recently has there been any progress in addressing the powerful and complex non-Western commercial apparatus that underlay the Indian economy. Yet, despite growing recognition that India has long maintained itself as a formidable commercial society, we still understand very little about the people who engaged in commerce, the institutional structures by which they controlled credit and money, the ways they used these structures for investment, and the values that underlay these uses.

Three Stereotypes of the Indian Moneylender

The powerful influence of nineteenth-century sociological theory is most clearly seen in three widely occurring stereotypes about Indian money-lenders and their relationships with agricultural producers. Perhaps the most prominent of these stereotypes is illustrated by the wonderfully evocative descriptions in R. K. Narayan's The Financial Expert (1952), a novel about a twentieth-century moneylender who guides his actions more with an eye to the goddess Lakshmi than with any consideration of economic rationality. Margayya, the moneylender of Malgudi, begins his business life by assisting peasants in dealing with the town's Central Cooperative Land Mortgage Bank. He collects some fees for this service. More importantly, he actively creates money for his clients. If a peasant requires a loan for marriage expenses, Margayya persuades a better-off

peasant to borrow money from the bank and loan it to the first peasant. The better-off peasant pays the interest he owes the bank from the slightly higher interest paid him by the first peasant. Margayya receives a fee for his assistance in the transaction on the occasion of the initial bank loan, which is made at no risk to himself. Money is plentiful. Only the interest is repaid, never the principal. Everybody is happy.

Irrationality enters the picture when Margayya steps beyond the role of broker and begins to loan money that he himself has borrowed or has received on deposit. By loaning money on outrageously large margins, Margayya plants the seeds of his own downfall. For he has neither the reserves of the cooperative bank nor any guarantee of government intervention in a crisis. As the book reaches its climax, Margayya's clients start a run on his money-lending operation, and he is forced into insolvency. Mystical, befuddled, and irrational, Margayya never comprehends the potential for disaster until it is too late.

A second and even more prevalent stereotype portrays Indian money-lenders as all too rational: coldly preying upon their cultivator clients, luring them further and further into debt, and finally sucking them dry of surplus, savings, property, and liberty. A classic ethnographic depiction of the usurious moneylender may be found in Darling's (1947) account of Punjabi peasantry (for a literary illustration of the stereotype, see Raja Rao's Gandhian indictment of village moneylenders in his 1939 novel Kanthapura ).^[4] In this stereotype, it is peasant cultivators who take over the burden of irrationality, while moneylenders emerge as rational but immoral.

Some of the excesses in these stereotypes have been addressed in an article by Michie (1978), who explores the blend of rationality and morality that actually characterizes interaction between moneylenders and peasant farmers. Unfortunately, even such a welcome corrective essay perpetuates one further stereotype: the image of the moneylender as an independent, strictly small-scale entrepreneur whose business activities are confined to credit transactions with his client agriculturalists. For example, even though Michie mentions the ability of moneylenders to manipulate links between villages and wider market networks (1978: 50), he pays no attention to the kinds of organization these links presuppose. His focus is tightly confined to the moneylender/cultivator relationship, defined by an exchange of loans and repayments. His primary concern is the organization of production. Missing from Michie's account is any description of market networks and moneylender/moneylender relationships, defined by exchanges of deposits, letters of credit, bills of exchange, and other financial instruments. In other words, he does not address the organiza-

tion of finance and trade. By default, Michie assigns moneylenders to small-scale operations limited by the assets they can generate from their cultivator clients.

Legal Stereotypes and Historical Myopia

It is this final stereotype about the scale of Indian finance that invariably colors the writing of colonial administrators, of economists during the colonial period, and of historians today. In general—to the extent that Indian credit operations are recognized at all—Indian moneylenders are taken as unreliable and irrational, or rational to the point of usurious immorality, but in any case as strictly small scale in the size of their assets and the scope of their credit extension activities. For example, it is easy to find reports devoting large amounts of space to efforts to distinguish between petty moneylenders (representing the Indian stereotype) and large-scale bankers (operating in the fashion of Western banks).

Reflecting this distinction, the legal history of the period is replete with judicial efforts to define indigenous financial instruments such as the hundi (a kind of bill of exchange used by moneylenders but not by "true bankers"). Ultimately, the courts concluded that such instruments lacked explicit statements stipulating conditions for certain kinds of transactions between multiple trading partners. Accordingly, their negotiability could not be appealed to a court of law (Krishnan 1959; Weersooria 1973). The implication of such a finding is that instruments such as hundis , which lack legal standing, could not possibly function effectively outside of a specific local community's ability to apply customary sanctions; therefore, hundis must be ineffective instruments for any kind of large-scale or long-distance trade.

Such a conclusion might be appropriate for a jurist or administrator who looks only to the courts for sanctions on contracts or authoritative judgment of disputes. On the other hand, it is certainly inappropriate for any person dealing with the day-to-day operation of an Indian commercial enterprise. The difficulty is that it simply ignores customary sanctions on hundi transactions that are rigorously enforced by multilocale, multiregional, and even multinational communities of businessmen. Indeed, the considerable negotiability established by hundis is a testament to the adequacy of these customary sanctions (see discussion in Chapter 5). When jurists' failure to appreciate these important financial instruments is placed in the context of stereotypic views about Indian bankers as merely clever (and sometimes irrational or usurious) moneylenders, it is clear that British and British-trained jurists never really comprehended the systematic operation of Indian financial institutions.

Stereotypes of the Indian moneylender even affect scholars who explicitly recognize indigenous systems of trade and banking during the colonial period. A. K. Bagchi (1972), for example, cites the monopolistic access of British entrepreneurs to "organized" banking systems as one of their important advantages in preempting Indian investment in industrial opportunities well into the twentieth century. Apparently, Bagchi has in mind characterizations of the Indian banking system in which Madras Presidency and British exchange banks, which made credit available to the British, are compared with the developing Indian joint-stock banks of the early twentieth century, which made credit available to Indians. An example is provided by Vakil and Muranjan (1927: 532), whose analysis of capital within the "organized" banking system for the period from 1913 to 1917 indicates a ratio of British credit to Indian credit of approximately five to two.^[5]

Buried in Bagchi's own footnotes is evidence that he unconsciously discounts (perhaps "devalues" is the better term) indigenous credit markets—presumably the "unorganized" banking sector. Compare the following two passages quoted by Bagchi from official government reports about British investment credit and the Nakarattars or Nattukottai Chettiars, the focus of the present book. The passages were published in 1901 and 1930, on either side of the period characterized by Vakil and Muranjan; nevertheless, they illustrate the point.

It was established in 1901 by Sir Edward Law [the member in charge of finance in the Viceroy's Council] that [British] banking capital available in India for trade purposes was less than £10 million [approximately Rs. 80 million], after making allowances for the share of the capital of the exchange banks which was held outside India; the amount required was estimated as £12 million [Rs. 96 million].^[6]

By contrast:

According to [Chettiar evidence in the Report of the Madras Provincial Banking Enquiry Committee], the capital of the Chettiars had increased from Rs. 100 million in 1896 to Rs. 800 million in 1930, and the capital employed by them (including borrowed capital) at home and in Madras came to Rs. 750 million, which is equal to what Furnivall claimed to be the capital employed by them in Burma alone.^[7]

Notice that the Chettiars' own estimate of Chettiar capital in 1896—not to mention Furnivall's (1956) less conservative estimate originally made during the same period—are both far in excess of the amount that the colonial government claimed was available for British trade in 1901; an amount which, according to Bagchi, represented superior British access to

investment capital. And this estimate does not even begin to take into account the capital controlled by non-Chettiar banking castes such as Marwaris, Parsis, or Baniyas. Yet none of this capital enters into calculations about the credit available to Indians (or Europeans) in India's organized banking sector. Thus, even modern economic historians such as Bagchi continue to accept Western colonial views that India lacked an institutional system capable of providing the large-scale finance necessary for industrial investment.^[8]

Lacking in all such views of Indian credit resources is any appreciation for the complex network of financial debts, opportunities, and possibilities that indigenous moneylenders and bankers could activate outside of Western-style banks through relationships of kinship and caste or through common participation with potential investors and lenders in a variety of religious and secular institutions. It is scarcely surprising that the scale and scope of Indian financial operations have been denied, when the very mechanisms for their transaction have gone unrecognized.

What are we to make of colonial administrators who regularly denied the existence of large-scale Indian commerce while interacting with its institutions on a daily basis? How are we to account for the strange myopia of British administrators and the historians who study them? On one hand, I suspect that verbal denial constituted a device by British banking interests for attaining special treatment from the emerging government-regulated banking systems in South and Southeast Asia (e.g., Buchanan 1870; Crawfurd 1971[1837]). On the other hand, as we have seen, anticolonial historians may give too much credit to claims made by colonial administrators in pronouncements on commercial policy while failing to consider the massive capital controlled and invested by Indian financiers. But these provocative assertions are themes for another essay. The present book attempts to modify the existing stereotypes (whatever their basis) by examining aspects of a large-scale system for credit provision—a non-Western banking system—operated by a South Indian caste during the colonial period.

Village Studies of Indian Commerce

One reason for the longevity of nineteenth-century biases about Indian society is that there have been remarkably few studies of the operation of even contemporary Indian commercial and other economic systems. Anthropologically, the focus has tended to fall on operations of the so-called jajmani system, a putatively India-wide redistributive system that operates within strictly local village contexts. According to standard models of jajmani , members of occupationally specialized castes exchange

their services for shares of agricultural produce controlled by members of the dominant landed caste in a village.^[9]

But there is growing reason to believe that the entire anthropological orientation toward and emphasis on jajmani models of village economic life merely reflect an untenable presumption about the noncommercial autonomy of precapitalist peasant villages. Recent systematic and detailed studies of village economic transactions have begun to show that the jajmani does not represent a pan-Indian phenomenon and that where jajmani or jajmani -like systems do operate, they make up only part of the total village economy (Commander 1983; Fuller 1989; Good 1982, 1991). Perhaps the most frustrating aspect of anthropological emphasis on jajmani is the impression it gives of India's economy as consisting of nothing else. Against any such misapprehension, it must be remembered that throughout Indian history, 40 percent to 60 percent, and under extreme circumstances perhaps as much as 80 percent, of village produce has left the village (Habib 1969; MBPEC 1930 1:35–85; Nicholson 1895; Rajayyan 1964–65; Robert 1983; Thorner 1960).

A large part of this exported village surplus takes the form of taxes in kind or money levied by various governmental institutions. Although not studied by anthropologists, such expropriation of village surplus has received attention from historians attempting to understand the financial basis of the great Indian empires, Mughal and British (Chandra 1966; Chaudhuri 1971; Habib 1969; Ludden 1985; Marshall 1976; Stokes 1978). Other historians, strongly influenced by anthropological concerns with indigenous ethos and world view, have attempted to integrate values believed to underlie village jajmani transactions with values believed to underlie the interactions between regional kings and local peasants (Dirks 1988; Greenough 1983; Price 1979). This last set of important studies identifies hierarchical values of "royal generosity" and "indulgence" as structuring the behavior of two sets of roles, as Greenough puts it: "the first set comprising powerful, resource controlling 'parents,' the second set comprising helpless needy 'children.' The roles in each set are equivalent in the sense that those in the first (kings, gods, masters) are all destined providers of subsistence ... while those in the second (devotees, subjects, dependents) are all persons requiring nurture" (1983: 841).

Studies such as these break down the artificial boundaries that existed between village geography and Indian economy. But they still ignore a major sector of the Indian economy not governed by parent/child values of hierarchical indulgence and generosity. What they miss are the relatively egalitarian values governing commercial relationships. Role sets of persons engaged in a commercial transaction are not split into two asymmetrical

categories of "subsistence providers" and "nurture requirers." Instead, they involve elaborate chains of lenders and debtors who are often mutually lenders and debtors to each other. For example, on one occasion, an agriculturalist may borrow money to purchase land, seed, or fertilizer. On another occasion, he may deposit excess profits from his harvest with a banker for investment on his account.^[10] Moreover, a landholder or trader who borrows money from a moneylender may loan out that money himself. And the original moneylender may easily borrow up to 90 percent of the monies he loans out (Robert 1983). In such relationships, values taken to underlie behavior cannot be understood as any simple transformation of generosity and indulgence from holders of parental roles to holders of child roles. Rather, the operative values seem to entail nonhierarchical relationships of trust and trustworthiness, often created and symbolized by religious gifting on the part of those engaged in commerce.

In other words, anthropologists and historians have paid little attention to an entire range of significant economic activities occurring beyond the level of the village. This observation applies to trade at a variety of periodic markets and especially to money-lending and banking activities involved in sophisticated indigenous systems for providing credit to farmers, traders, and governments. Without institutions of credit extension, India could not have maintained either its notorious tax levies or its extensive system of medium- and long-distance trade in agricultural and nonagricultural commodities.

Beyond the Village Moneylender

Another explanation for the continuing influence of stereotypes about Indian moneylenders may lie simply in the failure to consider implications of financial transactions beyond the level of the village. This omission, in turn, may be due to a marked bifurcation between money-lending peasants in village contexts and merchant-banking firms in urban contexts.^[11] This is not to say that social scientists have totally ignored nonvillage India, but only to indicate that nonvillage studies have failed to have any impact on prevailing views about the organization of Indian finance.

To the extent that scholars have addressed Indian financial organization beyond the village, their studies seem to fall into two categories: (1) studies of bazaar economy, and (2) studies of major commercial centers, sometimes referred to as "burgher cities" (Bayly 1978, 1983), and of Indian "burghers," sometimes called "portfolio capitalists" (Subrahmanyam 1990). Analyses of Indian bazaars look explicitly at sources of capital for petty shopkeepers, retail traders, and moneylenders operating within regional markets.^[12] In contrast to studies of village-based credit

markets, these bazaar studies explicitly raise questions about the scale of moneylender resources and the kind and degree of financial cooperation between moneylenders acting as insurers for and investors in one another. But the change in venue does not alter the analytic bias. By and large, the focus of these studies is confined to small-scale activities of moneylenders oriented toward what Fox (1969: 302)—drawing a parallel with putatively risk-aversive peasants—calls "subsistence trade." In keeping with prevailing stereotypes, moneylenders and shopkeepers of the bazaar are portrayed as irrational, as requiring only small amounts of capital, as lacking permanence, as conservative or even stagnant, and so forth.^[13] Bazaar studies perform an important service in broadening our understanding beyond agrarian production in the village. But they scarcely scratch the surface of Indian commerce. The role, operation, and even existence of large-scale merchant-bankers remains, with few exceptions, unknown or unconsidered.

The major exception to this general indictment consists of studies focused on coastal entrepôts and port cities extending from Saurastra through Gujarat, down the Malabar Coast and up the Coromandel to Golconda.^[14] In addition, a relatively small amount of attention has been paid to the role of merchants in inland cities such as Allahabad or Benares (Bayly 1978, 1983). Subrahmanyam's (1990) work on late precolonial South India and Baker's (1984) work on "modern" South India (1984) provide two additional examples of a focus that extends to inland commerce. With the exception of the last three writers, however, the predominant concern of these studies remains the working of empires—or, more recently, the working of local or regional polities. Their strength lies in their focus on politics; their weakness, in their lack of attention to commerce. In effect, they constitute analyses of political power in cities which happen to have a marked mercantile orientation. But the implications of mercantile organization and values as opposed to royal, professional, or administrative organization and values are not seriously examined.^[15]

Indian Burghers and Portfolio Capitalists

The recent studies by Bayly and Subrahmanyam require additional comment, for, as I have already indicated, they represent a major turning point to the general trends I have just described, and at first glance their findings seem to stand in radical contradiction to the conclusions presented in the present study.^[16] I begin with Bayly, who explicitly joins the issue and, in this respect, provides the best opportunity to scout out potential differences and agreements in our views. In several places Bayly presents detailed historical data and interpretations about North Indian commercial

towns and cities that correct largely unfounded, Weberian stereotypes about Indian commerce (Gadgil 1959; Lamb 1959; Sjoberg 1970). In particular, he is concerned with the organization of commerce in so-called burgher cities, such as Allahabad or Benares, which exhibit long histories of financial, commercial, and industrial activity and an elite, multicaste commercial community. Among other topics, Bayly addresses Weberian ideas concerning the relationship between caste and commerce within these burgher cities. In a concluding section of one paper on the topic, he points out that

caste at the level of geographically extended kin groups had an important role in the organization of trading diasporas; at the level of the commensal jati group it was relevant to the social and economic organization of artisan groups; at the level of varna it had some implications for general mercantile status. Nevertheless, it is difficult to see how caste in any sense could have been the prime parameter of mercantile organization in complex cities. Forms of arbitration, market control, brokerage, neighborhood communities, and above all conceptions of mercantile honor and credit breached caste boundaries, however construed, and imposed wider solidarities on merchant people. (1978: 192)

In this passage, Bayly's primary argument is directed against the tendency to reduce all merchant organization to caste organization. Unfortunately, the argument is put so strongly that a casual reader might conclude that Bayly sees little or no commercial role for caste at all. The polemical issue of whether caste is a prime parameter of mercantile organization overshadows potentially interesting questions about its role as a nonprime but still significant parameter, not to mention questions about the systematic interaction of various parameters of mercantile organization. The trick is to avoid an all-or-nothing view of the significance of caste.^[17]

The positive contributions of Bayly's research are too rich for detailed comment in the present context. Along with work by a small number of fellow historians (e.g., Habib 1960, 1973; Leonard 1979; Pearson 1976; Perlin 1983; Subrahmanyam 1990), it represents an important correction to views limiting Indian financial operations to those of small-scale moneylenders. In particular Bayly identifies an upper stratum of powerful merchant-bankers who maintain interregional trade in various commodities and credit notes and who provide important treasury and remittance facilities for regional and imperial authorities. The point I wish to make about Bayly's studies, however, is that they are open to serious misinterpretation. Bayly's focus on complex organizational networks that crosscut ties of caste or neighborhood, combined with his historically specific

analyses of urban commercial groups who elevate ties of class over ties of caste, are designed (among other things) to counter the classic view that castes are uniformly specialized with respect to occupation. It does not follow from this observation, however, that no castes were ever specialized. In particular, it does not follow that no caste ever specialized in commercial activities. Moreover, any such conclusion flouts the spirit of Bayly's work, which emphasizes the diversity of principles operating in the organization of Indian commerce, including principles of caste (Bayly, personal communication).

It may indeed be the case that the commercial elites of Allahabad and Benares gave no special precedence to relationships of caste. But this was not the case for all commercial magnates, especially those belonging to the Marwari caste (Timberg 1978) or the Kaikkolar caste (Mines, 1984). Nor was it the case for the Nakarattars, whose caste organization constituted a corporate financial institution in Indian society. Nevertheless, given the potential for misunderstanding, it is worth emphasizing that my approach to these issues is resolutely anti-Weberian: it no more forecloses the possibility for caste-neutral studies of elite businessmen than caste-neutral studies foreclose studies of the commercial role of caste. On the contrary, both kinds of study are needed. Chapter 7 outlines one way of beginning such a larger project.

The Historiographic Gap

Before presenting my analysis of the collective character of the commercial practice and social organization of the Nakarattars, it is worth commenting about a peculiar historiographic blind spot concerning their important role in colonial Asian society. By 1870 (and, perhaps, for some time before this) the Nakarattars had no contemporary South Indian competitors within the region comprising South India and British Southeast Asia. They were the premier merchant-banking caste of the region. Moreover, individual Nakarattars were among the first Tamil businessmen to divert their assets from banking and trade to capital-intensive industry. Nakarattars also played major roles in providing financial support and management for temples and charitable institutions wherever they did business. Finally, prominent Nakarattars wielded considerable political influence in a variety of local arenas throughout the Madras Presidency and Southeast Asia and, from 1920 onward, held important political offices in the city of Madras and the provincial government.

Despite the important role the Nakarattars played in Indian society, despite the availability of relevant information, even despite the publication of a handful of books and papers about Nakarattars, most accounts of

Indian commerce scarcely acknowledge them. Part of this omission follows in the wake of Weberian pronouncements about the putative incompatibility of Hinduism and capitalism, noted above. Another part may be attributed to too narrow a focus on village economy, or to stereotypic views of Indian moneylenders, which have also been discussed. But part of the problem arises from two further sources of historiographic bias. On one hand, scholars who focus explicitly on the activities of merchants in the seventeenth and eighteenth centuries tend to focus only on Indian-European connections in the organization of the textile export trade, especially as these connections were established in Madras and other port cities. As a consequence, they tend to look only at specific groups who were prominent in these port cities, such as Telugu Chettis and Brahmans or Tamil Vellalas. They do not explore the inland "country trade" of internal South India during the sixteenth and seventeenth centuries.^[18]

Nor do they address the growth of new economic actors in commercial activities during the late eighteenth and early nineteenth centuries.^[19] Rather, the dominant historiographic emphasis on the rise of Indian nationalism in the nineteenth century raises another difficulty. In post-1850s Madras, mercantile political power seems to have given way to growing influence by administrative and professional elites. And scholars whose research focuses on political developments in the period after 1850 tend to give mercantile and commercial elites only passing attention (Appadurai 1981; Lewandowski 1976; Suntharalingam 1974). In fact, the literature barely touches on them except in their capacity to bankroll politicians and ideologues (both the anti-Brahmans and those in Congress). Moreover, even here little attention is given to the source of the funds expended by the bankrollers.

Apart from such disciplinary biases, the whole issue seems to be conditioned by early European interactions with and efforts to master the Indian economic system. From the seventeenth to the nineteenth century, much of the European trade depended on financing by Indian capital under the control of large-scale merchants (Appadurai 1974; Arasaratnam 1980; Basu 1982; Brennig 1977; Chaudhuri 1978; Furber 1951; Lewandowski 1976; Subrahmanyam 1990). Not surprisingly, the relationship was always strained on both sides.^[20]

From the European point of view, Indian brokers held entirely too much power and, moreover, often employed it in competition with the Europeans themselves. By 1680, European merchants were already attempting to alter the indigenous system by insisting on dealing with groups of merchants operating like their own joint stock companies rather than with individuals with privileged claims on their European patrons

and monopolistic power over native producers (Brennig 1977: 338–340; see also Arasaratnam 1979; Subrahmanyam 1990). But these efforts to circumvent the Indian mercantile elite were by and large unsuccessful in that even the joint stock companies continued to be dominated by a small number of highly powerful "chief merchants" (Arasaratnam 1980).

This is not to say that efforts to circumvent the power of chief merchants were totally ineffective. From the mid-eighteenth century onward, an increasingly large European private trade and a growing English bureaucracy increased the number of Englishmen desiring contact with a variety of Indian producers, traders, and local authorities. This in turn led to increased opportunities for Indian middlemen, until virtually every major officer of the East India Company and every private European trader had his own personal middleman (Arasaratnam 1980; Basu 1982). Such middlemen came to be known as dubashes . According to some historians, the appearance of these dubashes represents a significant institutional change from the seventeenth-century situation, in which "chief merchants" served these functions. Individual dubashes are said not to have had the power of chief merchants. But this argument fails to take account of differences in the wealth and influence of dubashes situated in various places in the colonial power structure. Although the number of eighteenth-century dubashes grew dramatically relative to the number of seventeenth-century chief merchants, there remained a small number of individuals who continued in or acquired the position of dubash for leaders in the European community. These "chief dubashes " maintained a preeminence in Indo-European trade comparable to that of chief merchants in the seventeenth century.

In other words, the overall function and position in society of important merchants and financiers did not change from the seventeenth to the nineteenth century. There was a growth in the absolute number of Indian middleman-merchants. And all such individuals, whatever their station in colonial society, were designated by the title dubash . But there continued to be a small circle of elite merchants who exercised a dominant role in the organization and financing of Indo-European trade.^[21]

If there was no important change in the power and institutional role of preeminent Indian merchants, there was, nevertheless, a significant impact on the ethnic make-up of merchants holding positions of power and there were also important changes in the organization of ethnicity itself. As with the early position of chief merchant, competition for dubash positions was at least initially factionalized between groups from right-hand castes (primarily Telugu Chettis) and left-hand castes (Telugu Brahmans and Tamil Vellalars). Over the course of the eighteenth century, individu-

als belonging to left-hand groups gradually gained positions of power in a hierarchy of dubashes that corresponded to the hierarchy of European traders whom they served (Appadurai 1974; Basu 1982; Lewandowski 1976). Thus, while the position of chief merchant/dubash continued to play an important role, the period was witness to a change of ethnic guard in the personnel who filled these positions.

The process, however, seems to have altered the nature of ethnic identity or at least the nature of social ties that connected members belonging to the same caste. Increasingly, during the seventeenth and eighteenth centuries, individual merchants who rose to positions of chief merchants or dubashes were recruited either from castes that lacked a tradition of mercantile organization (e.g., Brahmans or Vellalas) or else from mercantile castes whose organizational structure failed to adjust to the lure of European commerce. In both cases, the chief merchants and dubashes of the Coromandel seemed to operate with increasing independence from special ties to their caste groups. Multicaste left-hand and right-hand factions gradually lost their ability to mobilize for coordinated action. Historians are unable to discern any form of caste organization in the nineteenth century—even within a single mercantile caste group such as the Telugu Komati Chettis—in which caste members coordinated their actions to promote a common commercial goal. Of the earlier forms of multifunctional caste cooperation (Krishna Rao 1964), only shared ritual observances at caste shrines and prohibitions enforcing caste endogamy remain. And the relationship between these ties of caste and the conduct of business were apparently so attenuated as to play no role of any kind.^[22]

The Historiographic Rejection of Caste in Commerce

Sanjay Subrahmanyam (1990) denies any caste basis for commercial activity, not only in the eighteenth and nineteenth centuries, but in the sixteenth and seventeenth centuries as well. Writing less than a decade after Bayly, Subrahmanyam presumably finds the evidence for multicaste and intercaste commercial activity so strong, and Weber's orientalism so patent, that he sees no need even to criticize the thesis of incompatibility between Hinduism and capitalism.^[23] Like Bayly, he focuses on prominent merchants, among them the chief merchants and dubashes of the Coromandel. Unlike Bayly, he refers to them as "portfolio capitalists" rather than "burghers," perhaps to emphasize that their commercial activity was not confined to a single city but extended up and down the Coromandel, into the Tamil and Telugu kingdoms of inland South India and outward to Southeast Asia. In other important respects, however, Subrahmanyam depicts his portfolio capitalists in the same way as Bayly depicts his North

Indian burghers. Specifically, he describes them as more or less individualistic entrepreneurs who occupied "the middle ground between mercantile capitalism and political capitalism" (1990: 298), gaining and using political positions to further commercial interests, harnessing commercial operations to gain political influence, and in general operating at a scale unimaginable according to the orientalist conception of India as the land of jajmani and village moneylenders. In Subrahmanyam's words, they formed

a group of persons who, from the second half of the sixteenth century, emerge gradually into prominence, eventually assuming formidable proportions in the first half of the seventeenth century. These persons, whom we have termed "portfolio capitalists," occupied a shadowy middle ground between state and producing economy, and combined a role in the fiscal structure with participation in inland trade, currency dealing, movements of bills of exchange, and even seaborne trade on a quite considerable scale. (1990: 355)

Although Subrahmanyam does not deign to address Weberian stereotypes, he is careful to raise three questions about more recent historiographic interpretations. In the first instance, as indicated in the preceding quotation, he calls into question the assumption (and, at times, the explicit postulate) that the phenomenon of chief merchants' acting as portfolio capitalists in the seventeenth and eighteenth centuries represents an innovation in reaction to commercial demand from the newly arrived European companies. In contrast, Subrahmanyam forcefully argues a position with which the present study is entirely sympathetic: that the portfolio capitalists who mediated between Indian textile producers and European companies performed commercial roles that extend back before the arrival of the European companies to at least the intra-Asian trade of the sixteenth and seventeenth centuries. Subrahmanyam's second characterization of portfolio capitalists is also one with which I am entirely sympathetic: that there existed no sharp separation of roles between merchants and "a 'militarily oriented elite' whose culturally sanctioned activities were land activities."^[24]

It is only with a third feature of Subrahmanyam's picture of portfolio capitalists that I have reservations. Citing Washbrook's (1975, 1976, 1982) study of the nineteenth-century rise of mercantile caste associations (see Chapter 2, above), Subrahmanyam argues against close identification of any functionally defined role with any ethnographic category, by which he means caste: "We cannot believe axiomatically that southern Indian society in the sixteenth and seventeenth centuries comprised of a set of well-defined and watertight compartments, in one of which it is possible to site

the mercantile communities of the region" (1990: 340). Subrahmanyam thus differs from Washbrook in denying that the precolonial organization of mercantile castes may have possessed properties that were "preadaptive" for the late nineteenth-century formation of political associations. Yet, as with Washbrook's treatment of nineteenth-century Komati merchants, it is important to emphasize that Subrahmanyam provides no information about intracaste cooperation between portfolio capitalists and that this omission leaves the question of mercantile caste organization entirely open. Moreover, it is important to separate the two conceptions of caste and commerce that Subrahmanyam attacks. It is one thing to argue that merchants engaged in war and politics, while warriors and politicians engaged in commerce. It is another thing to argue that absence of an exclusive occupational calling for some individuals rules out the possibility of any occupational specialization in the kin and caste groups to which they belong.

In this context, it is worth noting that Subrahmanyam's brief treatment of a dynastic merchant family of Balija Chettis is tantalizingly suggestive that ties of kin and caste did play an important but unexplored role in South Indian commerce. It may have been the case, as Subrahmanyam suggests, that this extended family operated with no special ties to other, less prominent members of their caste. But this was surely not always the case. And, as I shall argue for the Nakarattars, equally prominent merchant-bankers maintained commercially crucial ties to castemates that are invisible in the standard historical records.

Subrahmanyam's understanding of pre-Western portfolio capitalists is important because it deprives Weberians—and, indeed, all modernization theorists—of the argument that only involvement with Western financial institutions and rational, bureaucratic forms of government can trigger the forces of capitalism. Like more Eurocentric historians, however, Subrahmanyam still finds confirmation that Indian businessmen participated in capitalism—albeit an Indian capitalism—to the exclusion of caste and kinship. Yet, as indicated earlier in this chapter, it is precisely at the point where the unfolding commercial history of South India seems to provide an irrefutable case in support of these views that historians have shifted their attention to other issues. Had they continued with the economic concerns that motivated their studies of periods before the middle of the nineteenth century, the verdict would surely be less secure. For it is just at this time that the Nakarattars emerge on the scene in a major way. And it is precisely the qualities of their caste organization that enable them to take advantage of the changing colonial economy and become the chief merchant-bankers of South India and Southeast Asia.