Berger: Theory and ethnography in the modern anthropology of India

This work is licensed under the Creative Commons | © Peter Berger. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 (Online)

Theory and ethnography in the modern anthropology of India

Peter Berger, University of Groningen

Over the last sixty-five years, since the country’s independence, trained anthropologists have conducted extensive ethnographic fieldwork in India. In this time span, anthropological discourses about Indian society have developed their own specificity, while at the same time the anthropology of India has also had a profound impact on the discipline as a whole. This paper provides a critical overview of the general theoretical perspectives that have been employed by these anthropologists and that have been developed on the basis of their ethnographic experiences. In allusion to Ortner (1995), this paper is a plea for “ethnographic approval” in devising theoretical perspectives. It is argued that anthropological theorizing loses its heuristic value (i.e., the ability to help investigate, understand, analyze and compare the particular sociocultural life worlds of humans) when it abandons the dialogue with ethnographic reality.

Keywords: India, theory, ethnography, ethnosociology, subaltern studies, postcolonial studies, ethnographic refusal

If “[e]thnography…has never been mere description” (Nader 2011: 211, original emphasis), anthropological theory has never been mere abstract thinking. Ethnography and theory are predicated upon each other and both are aspects of a single endeavor. Since doing ethnography usually implies making an effort to understand and describe with reference to the complexities of a certain place, generalizations and theories emerge from the particulars of a locality. In a next step, a certain perspective or a generalization of a specific phenomenon is transferred to other regions or other parts of the globe: Trobriand reciprocity is compared to gift giving in India or African lineage theory is tested in highland Papua New Guinea. Such direct transplantations of theoretical perspectives do not always work, but such a comparative perspective helped refine our conceptual tools in a close dialogue with ethnographic data.

Mainly since its independence in 1947—thus for sixty-five years now—trained anthropologists have conducted ethnographic research in all corners of India, though anthropological attention has not been distributed evenly (see Berger and Heidemann, forthcoming). In this contribution I am not concerned with the specific themes these anthropologists have been concerned with (e.g., marriage), nor mainly with the analytical concepts they employed and considered (e.g., sanskritization), but rather with the development of their theoretical outlooks. Thus, it is constructive to ask: Which general perspectives about how to analyze and interpret Indian culture and society have been employed and specifically generated on the basis of doing ethnography in India? Theoretical approaches such as structural functionalism have certainly been imported, but new outlooks have also been devised in relation to the ethnographic experiences in various parts of the country. Through the decades, anthropological discourses about Indian society have developed their own specificity, while at the same time the anthropology of India has also had an impact on the discipline as a whole. As is the case for our field in general, the anthropology of India and its theoretical development are not strictly cumulative. As such, the more or less chronological account that follows does not mean to suggest a linear unfolding of theoretical insights. Let me also emphasize that I am not concerned in any significant way with different national traditions: the contributions of either American, Indian, French, or British anthropologists. The multidirectional traffic of influences, the internal diversity of national traditions and the trajectories of individual biographies often make such clear distinctions problematic.

The method I have adopted in critically outlining the theoretical strands is to focus on what I consider to be key publications. I consider some scholars to be representative of a certain way of theorizing about Indian society and I refrain from listing all of those who I consider to work in a similar way. Several approaches are treated in some detail because I do not only want to mention theoretical labels; I also want to illustrate arguments from within a certain theoretical perspective. Of course, such an attempt cannot be encyclopedic or pretend to be comprehensive or neutral. Rather, I present my reading of the issue and my attempt here should be understood as an invitation to discuss the theoretical particularities of the anthropology of India and the place ethnographic fieldwork has in this generalizing endeavor.

There is no cardinal way to conduct ethnographic fieldwork, since it depends on numerous factors such as topic, personality, region, political context and, perhaps not so frequently acknowledged, coincidence. However, in my view, doing ethnography—or in any case being concerned with ethnographic reality, sociocultural content and form, even if from a distance in time or space—is the very heart of our discipline, and is certainly not restricted to hilltop villages and tropical islands as has been amply demonstrated by now. I argue that my synopsis of the theoretical developments shows that as soon as anthropological theories let go of ethnographic descriptions dealing with the complexities and paradoxes of social life, they become a self-referential mind game, perhaps a kind of social or cultural philosophy. As such, they lose their capacity to help us better understand how ordinary people in some locality make sense of the world they live in and how we can compare their ideas, practices, and experiences with those of others. This does not imply that abstraction, theorizing, or generalizations are impossible; they must, however, remain in dialogue with ethnography. As will be seen, sometimes this detachment from the ethnographic ground may happen gradually, perhaps even to some extent inadvertently, and at other times ethnography is dismissed on purpose. Both scenarios I consider as variants of what Ortner (1995) calls “ethnographic refusal.”1

My discussion begins with the era of “village studies,” from the 1950s onwards, and then moves on to the “transactionalist” and “structuralist” perspectives of the 1960s. From the 1970s onwards, an approach called “ethnosociology” was promoted by American anthropologists. The 1980s, then, saw what has been labeled a “historical turn” and I will discuss three variants of historical approaches.

Village and civilization

Village India: Studies in the little community (Marriott 1955a) was a landmark book in the study of Indian society and culture. It exemplifies a general shift in anthropology away from “deserts, jungles, and arctic wastes” (Geertz 2010: 221) and toward the study of complex societies. The anthropology of India contributed significantly to this shift. Before Independence, few professional Western anthropologists chose India as their field of research and many of those who did, studied tribal societies, as this seemed appropriate then (Rivers, von Fürer-Haimendorf, Mandelbaum). As Bernard S. Cohn stated, “A. A. Aiyappan [who studied in London with Malinowski] was the main exception to the generalization that professional anthropologists studied tribal peoples” (Cohn 1968: 23). After World War II, ethnographic research in India received a new impetus—as did other regions of strategic economic and political interest—and several American universities in particular initiated well-funded research programmes to study Indian “civilization” from a grass-roots empirical perspective. However, Milton Singer still complained in 1965 that “it is a regrettable but undeniable fact that the social anthropology of South Asia…is thought of by many social anthropologists as peripheral to the mainstream of social anthropology” (Singer 1968: vii).

Village India—like its twin India’s villages (Srinivas 1955a)—presented the first results of this new ethnographic endeavour. Many of the contributions were the reports of young anthropologists—Indian, British and American—who had just returned from the field. The framework of the book as a whole, however, was conceived by the established anthropologists Robert Redfield and Milton Singer, both of the University of Chicago, who edited a whole series of broad social-scientific research projects called Comparative studies of cultures and civilizations. While the young McKim Marriott edited the Village India volume, Redfield and Singer wrote a foreword that in effect served as an introduction. For them, the challenge of studying a complex society was mainly one of method and their primary interest was therefore “the effort to understand how to seek understanding of any great civilization and its enormously complex changes through anthropological study of villages” (Redfield and Singer 1955: ix). With this problem in mind, Redfield developed, with some input from Singer, his influential idea of two interacting levels—a local or little tradition and an encompassing or great tradition—which was widely used but also criticized early on (see Dumont and Pocock 1957). I will return to Redfield and Singer’s ideas below.

The different contributors to the Village India volume did not all follow Redfield’s framework, nor did they share a single idea of the village as a unit of study. In fact, the volume displays a variety of approaches, and many of the then young scholars became founding fathers and mothers of different theoretical strands in the anthropology of India, exerting great influence on future generations of anthropologists. I will mention only those who may be taken to represent the various strands in the anthropology of India that continue, albeit often in a transformed way, up to the present.

M. N. Srinivas is one such founding figure of the anthropology of India. Here, I take Srinivas as a representative of the structural-functionalist camp in British social anthropology. As Evans-Pritchard’s “first doctoral student and R-B’s [Radcliffe-Brown’s] last” (Srinivas 1997: 11), he witnessed a decisive transformation in anthropology at Oxford: Evans-Pritchard broke away from the idea of anthropology as a positivist science, adopting a more hermeneutic and historically embedded approach. Nevertheless, Srinivas’s contribution in Village India is a classic example of an account of “social structure.” He perceives Rampura, the pseudonym for his village of study, as “a well defined structural entity” (1955b: 1) and investigates how local castes are, on the one hand, separated by rules of commensality and, on the other hand, interdependent through occupational specialization and patron-client relationships. The dominant caste fulfils, in his view, a unifying function for the village as a whole.

Although academically trained during the “high-water mark period of structural-functionalism” (Lee and Sacks 1993: 182) in Cambridge, Kathleen Gough’s contribution already challenged its limitations. In contrast to Srinivas, Gough did not see “her” village as a “well-defined structural entity.” She observed the severe social consequences of drastic economic changes: decreasing coherence at the village level and increasing rivalry between castes, especially the contestation of Brahmanical dominance in Kumbapettai, a village in the Tanjore District of Tamil Nadu. Not only did Gough state that the village would soon cease to be a meaningful unit of study in such circumstances, she also exemplified a broadening of anthropological focus to include government policies of the new state and party politics (Gough 1955: 51n).

Gough exemplifies another important strand in the anthropology of India. Like André Béteille, Gerald D. Berreman, Joan Mencher, and Frederick G. Bailey, she is mainly concerned with questions of power and dominance. All these scholars share a concern with political economy, class relations, and structures of hegemony and they pay attention to the influence and consequences of colonialism. However, it was only in the 1980s and 1990s that the key focus of Indian anthropology was turned to the significance of colonialism for Indian society and its more general theoretical implications.

Bernard S. Cohn, who pioneered such a perspective and can be regarded as the precursor of subaltern and postcolonial studies, had also participated in the Chicago seminar that led to the publication of Village India. Cohn received his PhD in anthropology from Cornell a year before the publication of Village India, but since Cohn had to do military service it was Marriott who actually composed the text based on Cohn’s oral and preliminary written reports (Cohn 1955: 53n2). In this early article Cohn is concerned with attempts at social mobility made by the Chamar, a major “untouchable” caste of Uttar Pradesh. At this stage of his development, though, his perspective was similar to that of Gough. He used a multifaceted approach that included political status, economy, kinship, and religion to account for the changing situation of the Chamar in a particular village. Like Gough, he stressed that complex processes of change cannot be grasped by focusing on the village alone, and pointed to contradictions in the way the Chamar sought equality within the modern context of the nation while “also trying to borrow and revive for themselves elements of a culture that the higher castes are shedding” (ibid.: 76).

Of the authors in Village India, McKim Marriott takes the theoretical perspective of Redfield and Singer most seriously, but he also tries to develop it further. He opens his chapter with two questions: a) “can such a village be…comprehended…as a whole in itself?” and, b) “can understanding of one such village contribute to the understanding of the greater culture and society?” He goes on to observe that “[t]he two questions are inversely related: if we say ‘Yes’ to the first question, then we must say ‘No’ to the second” (Marriott 1955b: 171). Two characteristics of his long academic life are thus already evident in this early contribution: firstly, he is an anthropologist with theoretical ambitions; secondly, although only tentatively here, he has a formalistic bias and a mathematical passion. He proceeds to reflect on the relationship between these two questions—and hence the relationship between village and civilization—and dismisses several analytic perspectives before settling on Redfield and Singer’s view of “primary civilization”:

This concept of a primary civilizational type and process is one of the most inviting of available models for conceptualizing Kishan Garhi’s [the village of his research] relations with its universe. A primary or ‘indigenous’ civilization is one which grows out of its own folk culture…. The ‘great tradition’ which is characteristically developed by…a primary civilization is a carrying-forward of cultural materials, norms, and values that were already contained in local little traditions…. An indigenous great tradition remains in constant communication with its own little traditions through a sacred literature, a class of literati, a sacred geography, and the rites and ceremonies associated with each of these. (Marriott 1955b: 181)

Marriott exemplifies this process of “constant communication” with reference to three fields: land administration (taking account of the impact of colonialism), caste organization, and religion. Moreover, he refines Redfield and Singer’s analysis by specifying an upward movement, through which aspects of the little tradition are generalized (“universalization”) and a “downward devolution of great-traditional elements and their integration with little-traditional elements” (“parochialization”) (ibid.: 197, 200). Coming back to his initial two questions, he concludes that although Indian villages have probably never been isolates, the study of little communities as sub-systems within more encompassing systems can contribute to the understanding of the processes of universalization and parochialization (ibid.: 191, 218).

Despite the overarching framework provided by Singer and Redfield, Village India shows that the anthropology of India was heterogeneous already at an early stage. The various contributors demonstrate an awareness of the complexity of Indian society and culture and follow diverse analytical perspectives. This diversity would only increase in the decades to come.

Studying transactions—approaching history

A decade after the publication of Village India Chicago was again the location of a general conference followed by a large publication, this time with 20 contributions, mainly by American anthropologists, bearing the title Structure and change in Indian society, published in 1968. Taken as a whole, the volume looks backwards rather than forwards and to some extent involved “flogging away at very dead horses,” as one reviewer commented (Bailey 1969a: 501). Like its predecessor of the previous decade, the topic of caste took center stage: four of the six sections dealt with it. Structure and change can be said to forecast the end of the village study era. Several contributions provide new perspectives, and in this respect I want to look at two of the authors who have already been mentioned: Marriott and Cohn.

In his widely recognized and quoted contribution “Caste ranking and food transactions: A matrix analysis,” Marriott’s approach is quite different from that in Village India (nor was this to be his last metamorphosis). Frederick G. Bailey, in his essay reviewing Structure and change, awarded Marriott’s contribution the victor ludorum (1969a: 500). This is not surprising since Marriott analyzed caste ranking along the theoretical lines Bailey himself had developed. As already mentioned, Bailey—a student of Max Gluckman in Manchester—was mainly interested in the political aspects of social life. Still a prolific writer, he is a staunch critic of the “idealist” approach proposed by Dumont, to be discussed below. The year Bailey’s review was published also saw the appearance of his book Stratagems and spoils: A social anthropology of politics, in which he fully develops his transactional and interactionalist vision of society (Bailey 1969b). According to this perspective human action in any society is similar because, all cultural particularities aside, social actors strive to achieve their (culturally defined) aims. Actors thus try to maximize gain and minimize loss, and they compete with others and make choices. “Society” is regarded as a result of this process. As in the case of Frederick Barth, “social structure” in Bailey’s vision is processual and actor focused.

In much the same vein, Marriott sees intercaste transactions in Kishan Garhi “as a kind of tournament among 24 teams [castes] which make up this village society” (1968: 154). The actors’ goal in this tournament is to gain “dominance over others through feeding them or securing dependence on others by being fed by them” (ibid.: 169). In this perspective, rank is the result of transaction, the latter being the “master conception” of the villagers (ibid.: 145). Significantly, the different values of the materials transacted (e.g., food cooked in clarified butter or in water, leftovers) are irrelevant from the transactional view, since giving results in a high and receiving in a low rank no matter what the substances transacted are: all transactions are symbolically equivalent (ibid.: 170). Consequently, the values of purity and pollution are perceived by Marriott to be “expressions of achievement” of the goals mentioned above (ibid.: 169). This behaviouristic and actor-focused approach is diametrically opposed to Dumont’s theory, as will be seen, but also in many ways quite different from Marriott’s own later views.

Marriott’s methodological procedure also deserves mention. All empirical transactions are entered in a two-dimensional matrix, which eventually comes to represent the abstract differences in net scores between the different castes. The correctness of the matrix is “tested” by comparing its results with the opinions villagers hold on rank. The matrix proves to be largely congruent with the empirical findings. In contrast to his earlier contribution in Village India, which reflected Redfield’s and Singer’s preference for a “middle-run” historical perspective (Marriott 1955b: 186), Marriott’s work at this stage was thus purely synchronic. In his future work the historical dimension similarly received little or no attention, as was also the case with Dumont.

For Cohn, whose BA included a major in history (Dirks 2005), the historicity of culture increasingly gained importance. In Structure and change—which Cohn coedited together with Milton Singer—his contribution (an overview of the study of Indian society and culture from Megasthenes onwards), foreshadowed much of his own later development and that of historical approaches in general. Cohn traced the views and impact of missionaries, Orientalists, and British administrators, the latter providing the “official” view of caste. Especially from the second half of the nineteenth century onwards, administrators and ethnographers focused on caste as a “thing,” something to be counted and listed, the properties of which were to be described. Moreover, they concentrated on caste because they believed that understanding caste was to understand people and hence to understand India. Most importantly, Cohn ventured the idea that in treating caste as they did and freezing a lived reality into a fixed schedule, colonial administrators were instrumental in creating the institutions they thought they were investigating. Cohn himself proposed this argument, which was later to be expressed in a much bolder way by Nicholas B. Dirks, tentatively:

The question remains, however, of what was the cause and what was effect. Did the notion of social precedence on a provincial basis, the enshrining of the categorical level of the caste system as against any real social grouping known in the earlier part of nineteenth century, in effect create that level? (Cohn 1968: 18)

Again anticipating the arguments of later postcolonial studies, Cohn argues that anthropological work in the decade after World War II had been influenced by assumptions about the nature of Indian society that had developed over the previous two hundred years. In combination with the dominant functionalist framework of the discipline at the time, these assumptions had led to the focus on isolated villages as well as to the centrality of caste and the perception of caste as a “thing.” Furthermore, “change” was investigated against the background of an earlier state that was assumed to have been “genuine.” The publication of Village India and India’s villages in 1955, Cohn writes, caused a “shock of recognition that great correction was needed” in terms of methods, interests, assumptions, and theories (ibid.: 25).

While Cohn apparently thought that Structure and change provided such new approaches (which was true in his and Marriott’s cases), as a whole the volume rather testified to the approaching end of a period of the anthropology of India. In his review of the book, the well-known historian Burton Stein, for example, diagnosed that the anthropology of India had entered a “critical phase” (Stein 1969: 453). Stein noted two trends. First, that American anthropologists used “culture-bound analytical categories,” thus elevating indigenous categories, often derived from classical textual sources such as varna,2 to the status of theoretical tools (ibid.: 452). In Stein’s view, this hampered generalization, comparison, and theoretical development, all of which he saw as lacking in the volume. Second, because anthropologists were increasingly concerned with social change, he frequently observed them “acting as historian[s],” engaging in “some kind of archival anthropology” or “some kind of documentary research” (ibid.: 453). Although he blamed his own discipline for having neglected this field, he did not appear to be in favour of anthropologists encroaching in a seemingly amateurish way upon his own academic domain instead of doing what they had been trained for: ethnographic fieldwork.

Stein’s identification of these two trends is significant because, ironically and perhaps to his dismay, they did indeed prove to presage two important theoretical developments in the anthropology of India during the 1970s and 1980s: ethnosociology fully engaging “culture-bound” concepts and ethnohistory developing and establishing “archival anthropology.” Yet, generalization, comparison, and theorizing did not at all suffer because of these approaches. Beside ethnohistory and ethnosociology, also a third strand was at the time putting forward general theoretical claims, namely Louis Dumont’s structuralism. Dumont had been outlining his approach to Indian society already from the mid-1950s onwards but his opus magnum on India had not fully been received in the English speaking world at the time Stein wrote his review; and in the whole volume of Structure and change only Marriott—a rigorous and early critic of Dumont—mentions Homo hierarchicus.

American and French versions of “the whole”

Louis Dumont’s Homo hierarchicus was one of two books that ultimately signalled the end of the village studies era, the other being Mandelbaum’s Society in India. Both books were only possible because of the village studies carried out in the 1950s and 1960s and, if we take the first English translation of Homo hierarchicus (French orig. 1966), they both appeared in 1970. However, the similarities end there. While Mandelbaum’s two volumes are encyclopaedic, inductive, descriptive, and easy to read, Dumont’s book is analytic, deductive, theoretical, and at times difficult to digest.

Society in India is written within a functionalist framework and outlines in great, survey-like detail the different social systems and sub-systems of Indian society, such as family, jati, and village, and the different “roles” actors have in these systems, the latter being understood in behavioural terms, as interactions between different individuals and groups. Accordingly, caste ranking is perceived as an extreme form of “stratification,” and “pervasive inequality” is indicated as a key feature of the caste system (Mandelbaum 1970: 6).

For Dumont, by contrast, caste is not an observable reality in the first place but a “state of mind (Dumont 1980: 34, original emphasis). This means that caste cannot be explained merely as a particular form of social structure or a particular type of social behaviour but primarily in terms of ideas and values. Like Durkheim’s “collective representations,” such ideas and values are basic categories of thought that are social in nature. Moreover, adopting Lévi-Strauss’ structuralism, Dumont stresses the relational properties of such ideas and values, which are integrated into the general cognitive systems he calls ideology. He therefore does not speak of “inequality,” as Mandelbaum does, but of “hierarchy” as a structuring principle, which he claims to have detected in classical Vedic texts dealing with the fourfold societal model of the varna. Religious status as expressed in the opposition of pure/impure is for Dumont the key value of Indian society, and it is represented by the Brahman priest in the varna model. Within the ideology, this value does not merely stand in opposition to its antithesis—power, represented by the kshatriya varna or the king—rather it encompasses the latter. Religion, the pure, and the Brahman thus represent society as a whole. While, according to Dumont, on the ideological level the religious is thus always superior to power, on the empirical level the reverse may be the case: the king being—in terms of power—superior to the materially dependent Brahman priest (see Dumont 1980, esp. introduction, chs. 2 and 3).

The superior encompassing value of purity and the clear distinction between religious status (Brahman) and power (the king) are the main conclusions Dumont draws from his analysis of the varna model. Having postulated this ideological structure as basic for understanding the caste system, he confronts his theory with ethnographic findings relating to marriage, commensality, and local authority. Not only does Dumont argue that all these social fields and relationships can be explained as manifestations of ideological structure, he also claims that hierarchy, as defined by him, is a general feature of systems of ideas. As such, he claims to have added another dimension to Lévi-Strauss’ model of binary opposition.

The composition of Homo hierarchicus suggests that the argument is deductive in nature—a general theoretical hypothesis being confronted with empirical data—which might also lead to the assumption that theory comes first and ethnography, the empirical, second. There is much in Dumont’s writings that supports such a view, for instance the way in which he delegates empirical aspects to the “residual level.” However, Homo hierarchicus was the end product of three to four years of ethnographic research in South and North India and a consequence of an intensive engagement with the ethnographic literature of his time. As the successor of Srinivas at Oxford in the early 1950s, Dumont became a close associate of Evans-Pritchard and David Pocock, and closely aligned with the British empirical tradition of anthropology in general. In fact, he credited Evans-Pritchard with an achievement in “ethnographic theory” for immediately discovering the conceptual principle of Nuer political organization from his ethnographic material (Dumont 1975: 335). Finally, being deeply influenced by his teacher, Marcel Mauss, Dumont held the microscopic aspects of ethnography in high esteem. Therefore, I would argue that the organization of Homo hierarchicus obliterates the fact that Dumont developed his argument in close interaction with ethnography. The theory of hierarchy is not only contained in the varna model but already in the “structural definition” of the Aiyanar temple, on which Dumont published in 1953 (Dumont 1970).

Not only are the outlooks, aims, and methods of Society in India and Homo hierarchicus very different, but so are their impacts on the anthropology of India. While the former has basically fallen into oblivion, the latter could be regarded as the single most important work on Indian society of the twentieth century. Dumont’s theory of hierarchy and his view of Indian society provided the ground for a great deal of debate in the 1970s, as, for example, two symposia on Homo hierarchicus in leading journals testify.3 Not only were Dumont’s daring arguments discussed and many weak points in his theory exposed, his contribution also served as a foil for new theoretical developments. An indicator of the continuing relevance of Dumont’s work—not only of Homo hierarchicus, but also beyond the anthropology of India—is the ongoing flow of publications dealing with his theory of value, which has been put into dialogue with many new ethnographic contexts (see Barnes et al. 1985; Parkin 2002; Khare 2006; Strenski 2008; Rio and Smedal 2009; Berger et al. 2010).

However, other general trends in the discipline of anthropology made themselves felt in India as well: Geertz’s symbolic anthropology, Lévi-Strauss’ structuralism, Asad and Said’s criticism of colonialism and Orientalism, feminist anthropology and neo-Marxist theories, to name just the main currents. As more anthropologists came to conduct fieldwork in India the themes also diversified. Caste and social structure became less dominant topics, while religion began to receive more attention (see Fuller and Spencer 1990).

India from within

The new wind: Changing identities in South Asia (David 1977a) can be taken to represent the main developments of the 1970s. Like its predecessors in the 1950s and 1960s the volume was based on a conference—in fact, three meetings in 1973—in Chicago, and although British and Indian anthropologists participated, the result is clearly a Chicago product. Furthermore, as previously, it was a young anthropologist who took on the responsibility of editing the volume: Kenneth David, who had received his PhD from Chicago only a year before the conference, under the supervision of the by now established McKim Marriott. However, while Village India and Structure and change hardly provided anything in terms of an introduction and basically presented a collection of articles representing the status quo of the (American) anthropology of India, David wrote 135 (!) pages of introduction and epilogue—partly stimulating, at times repetitive and tiring—and also included the conference discussions chaired by von Fürer-Haimendorf in the volume.

The title of the book, The new wind: Changing identities in South Asia, reflects the changing situation in South Asian societies and suggests innovative approaches to the study of the region. In retrospect, it can be argued that this claim was fulfilled, although the terms that figure in the two section headings which provide the frame for the seventeen contributions—“the moving” and “the standing”—appear to be a mere reformulation of the well-known “structure and change” dichotomy. The subtitle of the book also features a new term—“identity”—that derives from psychoanalytic theory and which has figured prominently in anthropology since the 1970s, arguably in American anthropology in particular. The term tends to be ambiguous and it was the British anthropologist Edmund Leach who, towards the end of the conference, asked what the term actually meant (David 1977b: 443). David tries to provide an answer to this question in his long epilogue, and it can generally be said that “identity” refers to the relational position of a person, group, or category in a social environment. In contrast to social structure, identity implies a less rigid, ascribed status but stresses dynamic aspects of social processes; thus “actor,” “arena,” and “strategy” are all terms that can be found in the pages of The new wind. A dominant theme in relation to identity and to Dumont’s arguments concerns ideology (see David 1977b: 508).

In many chapters Dumont’s theory is criticized, applied, modified, and combined. Owen Lynch deals with the theory in general and comes to the conclusion “that Dumont has offered us neither a productive paradigm for sociological research nor a particularly enlightening interpretation of India” (Lynch 1977: 262). Others, however, do not dismiss Dumont’s argument as a whole but rather criticize specific aspects of it. Mahapatra (1977), for example, shows that the clear distinction between the king and the religious sphere cannot be maintained when looking at the Jagganath cult in Orissa, where the God is a ruler and the king is a deity. Steve Barnett’s innovative contribution is an example of a combination of approaches: Dumont’s, Marriott’s (see below), and neo-Marxist approaches are most evident. Starting from Dumont’s concept of “substantialization” of castes, Barnett describes how—due to modern socioeconomic factors—a caste in South India became an independent “ethnic-like” unit (Barnett 1977: 402). The indigenous notion of blood-purity, which crucially defined relational caste and kinship contexts, was transformed in the process, becoming merely a matter of substance or being, while moral and normative aspects (“code”) relating to lifestyle became unstressed. This opened up new avenues for individuals to connect and identify within competing fields such as class, political parties, or Tamil nationalism. Barnett’s article is among those that pursue the book’s theme of changing identities in the most consistent manner.

“Ethnosociology” provides the new theoretical wind in the volume. We are dealing here with another of Marriott’s changing identities, although continuities are certainly discernable. Together with his Chicago colleague, the historian Ronald Inden, Marriott provided a short programmatic contribution to The new wind in which he introduced the new approach of ethnosociology. This piece should be read together with two other publications of his from the mid-1970s, the article “Caste systems,” also coauthored with Inden (Marriott and Inden 1974), and his important—and in many ways more convincing—article “Hindu transactions: diversity without dualism” (Marriott 1976), which provides a more detailed view of his new theoretical approach, fully indulges in matrix analysis, and makes his transactionalist position more explicit.4

While Marriott’s aim in Structure and change was to “understand a local caste hierarchy in village terms” (1968: 133), he was now at the level of “South Asian thought” in general (e.g., Marriott and Inden 1977: 236). While the concepts of “transfer” and “transaction” were still relevant to his framework—especially in the form of “flow”—the behaviourist emphasis had given way to, or was combined with, a stress on cognitive aspects of Hindu culture. While this can be read as reflecting the influence of Dumont’s emphasis on systems of ideas, Marriott conceived his theory in decided opposition to Dumont. The latter, Marriott argued, claimed to be concerned with indigenous ideology, but approached this ideology from a Western dualistic framework: “individualism” versus “holism” and “status” versus “power.” Instead, Marriott claimed, “South Asian thought” is characteristically and thoroughly nondualistic or monistic. Applying the ideas of the cultural anthropologist David Schneider to the Indian context, Marriott argued that “substance” (the natural, material) does not oppose the “code” (the moral, normative) but, from an emic viewpoint, all aspects of reality are natural and moral at the same time. Food transactions, for example, not only imply the transfer of items that nourish the body in a material way, but also a flow of a coded substance that has particular moral qualities (guna) and powers (shakti) that alter the bodies and persons of the actors involved. In perceiving their universe “as a structured flow of coded substance” (ibid.: 236), South Asians conceptualize entities of all kinds as basically porous, malleable and subject to constant reconstitution or change. Hence, like Dumont but for different reasons, Marriott argues that there are no individuals in South Asia but “dividuals” (ibid.: 232), temporary composites in ongoing processes of substance flows. This notion also found its way into general anthropological debates on the concept of the “person” (e.g., Strathern 1988. 13).

Certain methodological aspects of this stage of ethnosociology deserve mention. First, although Marriott has left the “little community” behind to investigate general cognitive systems, he does not (yet) do away with ethnography. Local narratives recorded by the ethnographer can provide “cognitive statements” (Marriott and Inden 1977: 229) as do the Sanskrit texts in which Marriott forages for key indigenous categories. Hence, again like Dumont, Marriott brings into interaction classical texts and ethnography and, as such, Indology and anthropology. In the second place, however, underlying this procedure is the assumption of a fundamental continuity of Indian culture over millennia, which Marriott makes quite explicit. The cognitive mode of nonduality, he writes:

is continuous from the ancient texts of the Vedas, Brahmanas, and Upanishads…through the classical books of moral and medical sciences…and on into twentieth-century explanations of their behaviour by living peoples. (1977: 229)

Finally, and this time unlike Dumont, what is most characteristic of ethnosociology is a kind of monistic perception of social science. Marriott conflates the indigenous system he is researching and his own research. In finding conceptual nonduality he says he is merely restating Indian sources (ibid.: 229), Indian culture is science and Marriott’s academic insights are Indian culture, or so it seems. Accordingly, he speaks of “indigenous ethnosociology” or the “indigenous scientific view” (ibid.: 232n). Below, I will briefly come back to the final twist of Marriott’s theory towards the end of the 1980s, at a time when his former companion Ronald Inden—more in tune with the trend of the time—had a rather critical and historical view of an imagined India.

While Marriott developed and refined his approach to ethnosociology, and increasingly became more of a theoretician than an ethnographer, his students applied his ideas to the field but shed much of his dogmatism (his transfigured behaviourism and mathematical fascination) in doing so. As such, they contributed significantly to the development of the anthropology of India. Susan Wadley, for one, made a careful re-study of Karimpur—the place where William and Charlotte Wiser had conducted what was perhaps the first village study in the 1920s—and exemplified the potential of ethnosociological analysis. In her contribution to The new wind she presented her arguments about the indigenous category of shakti or power. Dumont, she argued, disregarded the indigenous notions of power and wrongly constructed an opposition of “purity” and “power.” Power was not, in Wadley’s view, merely the political domain, but was thoroughly religious, as was purity. Moreover, like purity, shakti was an idea that structures relationships between humans and gods, and between humans. Purity and power are crucial factors in understanding what it means to be a Hindu and to act as a Hindu, and they are related to other key notions of Hindu ideology. To act in the right way (according to dharma), for example by worshipping the gods, provides access to the shakti of gods and impacts on one’s individual fate (karma) (Wadley 1977: 154). Much like Raheja, Dirks, Burghart, and Marglin in the 1980s, Wadley contributes to widening Dumont’s narrow focus on purity as the single value, hierarchy as the single type of relationship, and power as nonreligious.

History, or rather historiography, was a dominant influence throughout the 1980s and 1990s. Each of the approaches discussed so far had had its own relationship to a diachronic perspective. Redfield, Singer, and the young Marriott took notice of historical developments regarding the constant interaction between the two levels of great and little tradition. The Marriott of Structure and change provided a purely synchronic analysis of food transactions, while Marriott and Inden implied a kind of frozen history in their contribution to The new wind with texts from many different periods all equally representing “cognitive statements” of South Asian thought. Dumont was mainly concerned with structural changes in value systems and adopted a broad Maussian evolutionary perspective. In many instances, even if change and “new identities” were discussed, the new seemed—as Cohn had pointed out already in 1968—to be oddly opposed to the genuinely old, as phrases such as “post-traditional” suggest (David 1977b: 444; see Cohn 1987: 27; Spencer 1997: 2n). However, the point in the new developments—which involved literary criticism and history as much as anthropology—was not so much whether to view cultural and social patterns in a historical perspective or whether to locate change on the economic-political or ideological level, but rather to scrutinize the production of history itself, especially colonial history, and to lay bare its entanglement with relationships of power. Ethnohistory and the Subaltern Studies Group were two strands within this overall development that merged with postcolonial studies in the late 1980s and early 1990s (see Prakash 1994; Mathur 2000; Novetzke and Patton 2008).

Historiography from below

At the beginning of the 1980s, the historian Ranajit Guha initiated a scholarly programme of considerable scope and impact known as the Subaltern Studies Group.5 In 1988 a widely recognized collection of essays from the first five volumes of the Subaltern Studies Group was published (characteristically with a foreword by Edward Said). Three contributions by Guha to the first two volumes of the Subaltern Studies Group were presented in the first section (Guha 1988a, b, c). These three texts exemplify the sincerity and open-mindedness of Guha’s ideas, which form the basis of a long and diverse project involving many scholars, among them Bernard Cohn. The first of these texts is the preface to the first volume of the Subaltern Studies Group; the second is the manifesto of the group summarizing in sixteen points the cornerstones of their agenda; and the third, by now a classic text, presents a detailed and sharp analysis of the historiography of peasant rebellions unmasking it as “prose of counter-insurgency” (the title of the essay).

Through his interest in peasant insurgencies in colonial India, Guha became aware of the elitist bias of most historical accounts. The “refusal to acknowledge the insurgent as the subject of his own history” (Guha 1988c: 82) was for him the driving motive to challenge the historiography of India in a radical way, to disclose the often implicit strategies of representation, and to devise alternative methodologies to restore the agency of the subaltern subject. “Subaltern” was a term he borrowed from the Italian philosopher and Marxist Antonio Gramsci, and refers to “the general attribute of subordination in South Asian society whether this is expressed in terms of class, caste, age, gender, and office or in any other way” (Guha 1988a: 35). Although Guha himself was at the time mainly concerned with peasant insurgencies and Indian historiography, the scope of the project as a whole was much broader, as he makes very clear in his preface: “As such there is nothing in the material and spiritual aspects of that [the subaltern] condition, past or present, which does not interest us” (ibid.). I have italicized the word “spiritual” in order to highlight the fact that Guha here assigns a crucial role to culture and religion and thus deviates not only from classical Marxist analyses but also from much postcolonial work that was yet to come. Yet, Sherry B. Ortner remains unconvinced by this statement and argues that what Guha calls “peasant religiosity” is “a kind of diffuse consciousness that is never further explored as a set of ideas, practices, and feelings built into the religious universe the peasant inhabits” (1995: 181). As such she argues that Guha is also guilty of “thinning culture” and “ethnographic refusal” (ibid. 180n).

In “The Prose of Counter-Insurgency,” originally published in 1983, Guha dissects and reorders—some would say he “deconstructs” (see Spivak 1988)—the historiography of peasant insurgencies. He uses the techniques of literary criticism, semiotics, and structural linguistics—notably the theories of Roland Barthes, Emile Benveniste, and Roman Jacobson—to identify textual strategies, meanings and, most importantly, structural omissions in the texts produced on the rebellions. He distinguishes three discourses or genres—primary, secondary, and tertiary—that follow one another chronologically and are thus increasingly further removed from the events they describe. In many ways, the three types of historiography are different and even partly contradict each other in their aims. Yet they all share one failure, and this is Guha’s main point, which is that their descriptions and explanations of the peasant rebellions rely on external factors, and the people who revolt remain an epiphenomenon.

The primary discourse includes texts with an immediate relationship to the rebellions, for example, letters British officials wrote in the face of a revolt on their doorstep. In such texts, events are described and action is called for from the colonial regime, with which the writers fully identify. Thus, this discourse remains completely within the reason of the colonial state.

Texts of the secondary discourse are further removed from the events and have a “duplex character” (Guha 1988c: 51). They do not have the direct administrative purpose of the primary discourse and purport to be “history.” However, despite their potential “aura of impartiality,” the texts—some liberal, others racist—are at the intersection of a system of power (colonialism) and a system of knowledge (historiography); they “interpret the world in order to master it” (ibid.: 70). Guha shows in his analysis how the authors confirm the colonial mindset and policy. Causes of the rebellions are identified that relate to the indigenous situation, such as the exploitation of tribals by Hindu “money-lenders” and the inability of the colonial government to deal with this problem. Within a narrative of colonial continuity, it is then concluded that the administration has to improve in these respects in order to perpetuate the Raj. Thus, something can be learned. Guha summarizes, and partly quotes, a text from the first half of the nineteenth century:

To know the cause of a phenomenon is already a step taken in the direction of controlling it. To investigate thereby understand the cause of rural disturbances is an aid to measures “deemed expedient to prevent a recurrence of similar disorders.” (Guha 1988c: 74, original emphasis)

This connection of colonialist knowledge with power is a key concern of ethnohistory, postcolonial studies, and the Subaltern Studies Group.

The writers of the secondary discourse systematically leave out the perspective of the insurgents. In their interpretations of events (“terrible”), the peasant rebel view (“fine” and “legitimate”) only exists as an unstated other; a “revolt against zamindari” figures in the texts as “defying the authority of the State” (ibid.: 59). In this way, Guha speaks of a prose of counter-insurgency. Only by reading the account against the grain can the subaltern voice be recovered.

Even though texts belonging to the tertiary discourse—which includes post-Independence historiography—are emphatically in favour of the insurgents, they appropriate the latter, Guha argues, because of an inability to overcome the authors’ own leftist ideological preferences. In this “History-of-the-Freedom-Struggle genre” the insurgent’s consciousness of the past is substituted with the historian’s consciousness of the present (ibid.: 77). Hence, they too fail to account for the subaltern view. One domain where this failure of the tertiary discourse becomes most evident is in understanding the role of religiosity in the rebellions, labelled as “fanatical superstition” (ibid.: 79) in the earlier genres. Because of their own Marxist and secularist inclinations, the writers of the tertiary discourse can only perceive religion as propaganda exerted by a few chiefs to motivate the dull masses. For Guha,

[r]eligiosity was, by all accounts, central to the hool [Santal rebellion of 1855]. The notion of power which inspired it, was made up of such ideas and expressed in such words and acts as were explicitly religious in character. It was not that power was a content wrapped up in a form external to it called religion…. [I]t is not possible to speak of insurgency in this case except as a religious consciousness. (Guha 1988c:78)

Guha here again emphasizes the role of religion for understanding subaltern agency. However, I would support Ortner’s criticism mentioned above because Guha effectively dedicates only very little space to Santal religion in this long article (though material would have been available) and does not say anything explicit about the form, content, or practice of it. Nonetheless I consider it significant that Guha so strongly underscores the role of religion—and as such of culture, even though both are not the same of course—in those programmatic essays that set the agenda for the subaltern approach. Even though he fails to implement his own conviction in appropriate—or “thick,” in Ortner’s terms—ways, his statements clearly communicate the necessity of paying close attention to these phenomena within the subaltern perspective. This seems to be a distinctive strand of Guha’s subaltern studies when compared with later postcolonial developments to be discussed below. For Guha, then, power is not a noncultural anonymous force but a phenomenon permeated by culture and religion. It could be assumed that this applies as much to colonial power as to subaltern resistance.

Historical anthropology: not just “another new speciality”

Another key publication of the 1980s—which had a lot in common with Guha’s approach and included an introduction by him—was Bernard Cohn’s An anthropologist among the historians and other essays (Cohn 1987). Articles covering roughly thirty years of Cohn’s work are collected in this book, two of which have already been mentioned above. I will deal here with some of those written in the first half of the 1980s. The first three essays of the book are methodological in nature and explore the possibilities of combining anthropology and history “not just as another new speciality, not just as the means by which more hyphenated histories and anthropologies may be generated, but as the means by which an epistemology and subject matter common to the two disciplines might be reasserted” (1987: 42). In other words, Cohn does not want to add “ethno-history” to the list of sub-disciplines, but argues that the way anthropology looks at its subject matter changes fundamentally once the historicity of culture is fully recognized.

The basic assumption of Cohn’s programme is thus that culture is always “mediated by history” (ibid.: 58). Cohn hereby deliberately seeks to break away from the before/after fallacy that I described above and to which [according to Guha (1987: xx-xxi)] the early Cohn was also prone to succumb. To state that India was “traditional” before the colonization process began is to misconceive the nature of culture, which is always in the making. Related to the before/after dichotomy is the equally problematic notion of “impact.” Colonial power surely had an impact on Indian society, but as in the case above, the term obscures the interaction that was taking place. While “impact” implies a one-way process and a passive reception by the people of India, Cohn claims that colonizers and colonized have to be investigated as actors “united in one analytic field” (Cohn 1987: 44). The production of a “colonial sociology,” which Cohn outlines with reference to Fiji, can serve as an example (ibid.: 58–63). The model of the traditional society constructed by Europeans rested on general European nineteenth-century assumptions about society and culture, and to construct such a model in the first place was their initiative. However, Fijians acted as informants, translators, and “experts” on local culture and hence played a crucial part in the construction process of that model.

What is also at stake here is the process Cohn calls “objectification” (ibid.: 228–31). In the process of interaction with the colonial regime, of interpreting and representing their culture to outsiders, a (smaller or larger) part of the indigenous population is looking at its own culture in a different way, perceiving it as something to think about. One such “analytic field” where colonizers and colonized interacted and that “provided an arena for Indians to ask questions about themselves” (ibid.: 230) was the census operations that started in the nineteenth century. The most profound effect this operation had was on the half a million enumerators, those literate locals who conducted the census and whose duties included listing, naming, and classifying caste and religion. By the 1930s the effect had diffused through much of the population. Cohn demonstrates this with the case of a community in Punjab that wanted to be recorded as “Rajput” in the census of 1911. It succeeded partly and this had immediate effect on its access to Government facilities. Obviously this group, like many others, was already conscious of the relevance and potential such classificatory schemes had in the lives of its members (ibid.: 249). Cohn’s research on the census exemplifies the methodology of anthropology he envisions:

Historical anthropology then will be the delineation of cultures, the location of these in historical time through the study of events which affect and transform structures, and the explanation of the consequences of these transformations. (Cohn 1987: 73)

One striking characteristic of Cohn’s approach is the strong plea for culture and the central place culture—in the sense of Geertz and the “collective representations” of Durkheim—takes in his methodology. This is despite the fact that he acknowledges the political context in which anthropology as a discipline is embedded, for example, when he writes in “History and anthropology: the state of play” (orig. 1980): “in 1942 Americans had to confront ‘the others’…and the concept of culture was drafted to help the war effort” (ibid.: 26). This essay starts in a playful way, outlining the features of anthropologyland and historyland and the endeavours of Philias Fillagap and Lucy Lacuna therein (ibid.: 21). When it comes to culture, however, the tone is quite different. It is worth quoting Cohn at some length:

We cannot get around the study of the cultural order by dismissing it as attitudes, folk beliefs, mere ideological formations, or false consciousness, or as merely the window dressing for the ‘practical realities’ of social life, because the cultural order is the very basis of the institutional order. The true mystification as far as I am concerned is all those theories which try to reduce culture to the epiphenomenal or dependent. There can be no practical realities without the symbolic coding of them as practical; the theory that the social is created out of action—the day-to-day decisions of myriads of people—truly obfuscates the nature of the social. People cannot act as maximizers…without the pre-existence of meaning in cultural terms. (Cohn 1987: 40n, original emphasis)

Although Cohn criticizes a particular brand of history in the section from which this quote is taken, I think his criticism is much more far-reaching in that in one stroke it dismisses vulgar Marxism (as Guha does), transactionalism (such as Marriott in 1968 or Bailey), utilitarianism, and, in anticipation, much of postcolonial and postmodern anthropology that was yet to come.

When Cohn argues that the processes of cultural construction should be studied in cultural terms this also means that the “units of study”—such as “power, authority, exchange”—“should be cultural and culturally derived” (ibid.: 47), that is, understood in a particular cultural context. There is no value or culture-free space, and Cohn recognizes the analytic tools of anthropologists and historians being as much subject to a cultural coding—their ideas of “authenticity” and “chronology” for instance (ibid.: 50n)—as the kinship system of an Indian peasant. Also Marriott acknowledges the “cultural coding” of his theoretical tools. In his case he claims this coding to be “Hindu.”

Marriott in another dimension

In 1989 Marriott added another dimension to his theory of ethnosociology. Five years earlier T. N. Madan was still hesitant to summarize Marriott’s ideas because, as he ironically comments, “Marriott’s position is still evolving—I dare not say it is ‘fluid’” (Madan 1994: 97). However, with the special issues of Contributions on ethnosociology in 1989, which was published as India through Hindu categories a year later (Marriott 1990a), one can say with some confidence that Marriott’s theory had reached a solid state of aggregation. Marriott never summarized his main theoretical ideas in a monograph, and the long introductory chapter “Constructing an Indian ethnosociology” (Marriott 1990b), presents the most detailed account of his ideas that is available. I will deal with this chapter only very briefly, because while it presents the reader with a lot of technicalities and new terms, the added value of all this in relationship to his articles from the 1970s, is not directly evident. Why mention it here at all then? Because of its ambition and the debate that it triggered and, even more relevant to my argument, it shows what happens to theories when they abolish the genuine dialogue with ethnographic material.

Marriott had first endeavoured to understand village structure, then South Asian thought. Now he wants to “expand the world repertory of social sciences” (ibid.: 1). Going against the grain of the postmodern trend of the time, with its aversion to any far-reaching “grand narratives,” Marriott turned for methodological inspiration to Talcott Parsons, a towering figure of mid-twentieth century American sociology, seeking to construct “an alternative general theoretical system for the social sciences of a non-Western civilization, using that civilization’s own categories” (ibid.: 5). As he had in the earlier phase of his intellectual development, he based this aim on the assumption that to understand Indian culture one has to get rid of the categories and thought patterns from one’s own society. What he attempted is to unearth a kind of relational grammar of Hindu thought that would help understand all kinds of its many textual and empirical manifestations.

Briefly, Marriott proceeds by selecting from a range of Indian philosophical texts indigenous classificatory schemes that include three or more categories: the five elements, three humours, three strands, and four human aims. The number three is important here because this leads him to the three-dimensional geometric representation of a cube, the different terms being placed on its different sides. Arguing that the categories within these schemes are largely homologous, that is, they overlap in semantic content, Marriott merges the schemes and provides a new name for each attribute that captures its general meaning (mixing, unmarking, unmatching, grossening, and consciousness). The different cubes can thus be merged into one cube. Moreover, due to the specific nature of Hindu thought, particular cognitive features are translated into mathematical terms, each time an inversion—though not absolute—of Western thinking [i.e., nonreflective, nonsymmetrical, and nontransitive (ibid.: 17)]. For example, as Marriott had already stated in his earlier work, in India people are conceived of not as individuals but as relational “dividuals” and hence nonreflexive. “[A]nything that partakes of ‘substance,’” Marriott states (ibid.: 16), hence everything that exists, can be represented by the dimensions of the cube, which can be turned around and reflected upon from different perspectives. Thus, the cube is supposed to be as dynamic as Hindu thought itself.

There is no need to criticize this version of Marriott’s theory in detail here, as this has been done thoroughly and eloquently by others.6 It is obvious already in my brief summary that the way Marriott settles on the number three in selecting relevant cultural categories in order to develop his cube is completely arbitrary. Moreover, even though many anthropologists find aspects of his ethnosociology inspiring, most doubt that it is feasible to squeeze fluid “Hindu Thought” into a cube that appears to be quite rigid and criticize his undue claim to have founded a new science.

Perhaps more telling than the criticism by his colleagues in other parts of the world is the reaction of the students of the Chicago environment who are also among the other contributors to India through Hindu categories. Besides the general focus on indigenous categories, a common feature of their work is the repudiation of Dumont’s theory.7 Gloria G. Raheja and Nicholas B. Dirks, for example, both reject Dumont’s notion of power (and kingship) as completely separate from the religious sphere and they share a focus on gift giving from the perspective of a politico-religious center. Very few of Marriott’s students seriously consider the cube or related theoretical attempts as exemplified by his latest development. Rather, they put some of his general earlier ideas to test in relation to their ethnographic work, or they turn to a completely different direction.

Raheja’s contribution to India through Hindu categories focuses on the dominant Gujar caste in a North Indian village and the caste members’ central role as givers of religious gifts known as dan. Like Hocart, and in contrast to Dumont, she argues that sovereignty has political and religious aspects, the Gujar having the crucial function of distributing evil and inauspiciousness through their dan prestations. Moreover, although hierarchy is evident in inter-caste relationships, Raheja rejects Dumont’s claim that there is one ideological principle and one type of relationship. Instead, she claims that differently valued relationships come to the fore depending on the context (1990: 81). The “centrality” of the Gujar in the structure of prestations when they distribute inauspiciousness to “others” is one example of this. However, in the different contexts defined locally as transactions between “one’s own people,” relationships between the very same castes are conceptualized as sharing and “mutuality” (ibid.: 91). Thus, as is a common feature of the ethnographic work inspired by ethnosociology, Raheja pays close attention not only to observable processes of exchange but also to linguistic aspects of social action. Like the work of Wadley discussed earlier, her ethnography displays the possibilities of Marriott’s approach while avoiding many of its pitfalls.

Dirks’ contribution clearly departs from Marriott’s theoretical assumptions, methodological outlook, and academic aims. The only sense in which he describes his analysis as ethnosociological is in the repudiation of dichotomies such as “religion” versus “power” (Dirks 1990: 60). His three criticisms he voices of Marriott’s theory can be read as an inverted agenda of his own style of ethnohistory. First, he asks on what ground ethnosociology can assume an “epistemological privilege” in representing indigenous thought? Second, he criticizes the fact that historical questions are mostly excluded and that, third, the relationship between knowledge and power is not dealt with. If power is an issue—as in the case of Wadley discussed above—this is merely done within “a restricted form of cultural analysis” (ibid.: 75). Apparently, when it comes to history, knowledge, and power, Dirks has more on his mind, which is why his work will be discussed in the next, penultimate section.

Anthropology and culture as problems

Dirks received his PhD in 1981 from the Department of History of the University of Chicago. His work makes it evident once again that scholarship on Indian history and culture is part of a much larger academic context, beyond India and anthropology. Just as Village India was written against the backdrop of American functionalist social theory and Dumont’s writing was inspired by structuralism, Dirks’ work on the impact of colonialism on Indian culture (notably caste) was fundamentally influenced by poststructural and postmodern perspectives, as well as so-called reflexive anthropology. The key terms in this debate, “representation,” “voice,” and “power,” all in relation to an ambivalently perceived notion of “culture,” inform his arguments. The key publications discussed here are the article in Marriott’s India through Hindu categories (mentioned above), which is a summary of the argument in Dirks’ book The hollow crown (first published in 1987). Dirks’ further development can be seen in the 1992 volume Colonialism and culture8 he edited with contributions on Africa, South America, and Asia, and in his book Castes of mind: colonialism and the making of modern India (2001).

Dirks’ work, in a way rather similar to Dumont’s Homo hierarchicus, is an invitation to misreading. His arguments are as general as those of Dumont (or Marriott) and as provocative. Moreover, his writing style and terminology may lead to misinterpretation. This concerns not some minor point in his argument but its very crux: the extent of the influence of colonial power. Despite his use of words such as “invention,” “creation,” “production,” and “the making” to frame the effect of British rule, he certainly does not argue that the British invented anything ex nihilo, but rather—less dramatically—that they transformed what was already there.

The two examples of such transformed traditions that Dirks sees as most prominent and related are kingship and caste. In precolonial South India—as in other parts of the country—local kings were part of a dynamic network of major and minor kings struggling for dominance in warfare as much as in worship and the support of temples. Kingship was an institution involving the political and religious domains. Moreover, the social hierarchy was predicated not so much on the ideology of the pure and the impure as on the institution of the king. Because ruling was about people and not about territory in the first place, the main socioreligious practice that kept this social structure going and changing was a system of gift giving. The king bestowed honours, privileges and especially tax-free land on all sorts of institutions and people, such as his militia, affines, castes, priests, and village heads, thereby securing their loyalty and support as well as his own position. Not only was caste as much political as it was religious, values—“key discursive components” in Dirks’ words (1990: 61)—unnoticed by Dumont, such as honour, also lay at the basis of this political system. In receiving gifts the king’s subjects also participated in his socioreligious sovereignty, which was graded and relational, and diffused through the whole system, as were the rights to land (ibid.:67).

The British, Dirks argues, did not understand how this system worked and what it was all about. In effect, they froze kingship by taking away the political and dynamic aspect and turning it into a theatre state, a “hollow crown.” Significantly, Dirks emphasizes that the colonial project was not one great homogenous design of planning and insight but rather a matter of unintended consequences, and colonialism was often “successful in spite of itself” (1992b: 203). In their settlement measures focusing on clear procedures and the extraction of revenue, the British transformed the whole system of land ownership and turned the little king into a landlord. With the kings thus deprived of any real power and the land system transformed, the distribution of royal gifts lost its function and feasibility. If they wanted to engage in political struggles the kings had to enter a completely new field—that of British courts with alien rules and new categories. If the little kings wanted to donate land, their right to do this had to be argued for in this new legal context (ibid.: 201). Thus, while old elites continued to have important roles, the institution of kingship was fundamentally transformed under British rule.

So it was with caste, Dirks’ main theme in many publications. Through the colonial encounter, Dirks argues in Castes of mind, caste came to be regarded as synonymous with “traditional India” and the key representation of its fundamental difference from the West. While the institution was certainly present in precolonial times, it too was transfigured by British rule. Through the complex interplay of missionaries, administrators, and Orientalist scholarship, caste represented India as fundamentally religious and nonpolitical; it became reified into a “system” that was consistent, uniform, and all-encompassing. Significantly, colonial policy changed after the Great Rebellion of 1857, and colonial administrators started systematically to collect empirical information in order to prevent such situations in the future, being well aware that to know means to rule. With the census the “spirit of caste attained its apotheosis” (Dirks 2001: 48). Moreover, the census implied the “ethnographic state:”

To put the matter in bold relief, after 1857, anthropology supplanted history as the principal colonial modality of knowledge and rule. By the late nineteenth century…the colonial state in India can be characterized as the ethnographic state…. The ethnographic state was driven by the belief that India could be ruled using anthropological knowledge to understand and control its subjects. (Dirks 2001: 43n)

After extensive criticism of Dumont and Marriott the chapter concludes: “The state of ethnography—British, French, and American—turns out to be a direct descendent of the ethnographic state” (ibid.: 60). The questions this statement raises relate not so much to the approaches of Dumont and Marriott but rather, more generally, to how Dirks—and postcolonial approaches generally—assesses anthropology, ethnography, and culture.

In 1992 Dirks published Colonialism and culture and in its introduction he gives some clues to how the relationship between these terms is conceptualized. The concept of culture figures in, at least, two senses. First, “the culture in the title of this volume is…the congeries of belief, value, assumption, and habitus identified by anthropology” (1992a: 22). This notion of culture, which is mentioned earlier in the introduction (ibid.: 5), is similar to Cohn’s idea outlined above. The second meaning of culture is that of “a regime in which power achieves its ultimate apotheosis” or “a site of intervention, dislocation, and struggle” (ibid.: 5, 10n). In this perspective, which owes a lot to the literary critic Edward Said and his famous book Orientalism (first published in 1978), culture always implies the creation of differences between “our culture” and “their culture,” or “the Other,” that derive from and legitimate the hegemony of the West (1992a: 9, 2001: 313). It is the latter view of culture as a “project of control” (1992a: 15) that ultimately dominates the theoretical perspective of postcolonial studies outlined in the introduction (and also as it is evident in Castes of mind): “many of us [the contributors of Colonialism and culture] now believe that colonialism is what culture is all about” (ibid.: 11).

What, in my view, is left open is the relationship between these two notions of culture and the relationship between “power” and the first definition of culture as, put simply, a symbolic system. This becomes more problematic because the term “power” is itself left unspecified. Dirks probably assumes that culture as a symbolic system is always permeated by power. The reverse does not necessarily seem to be the case, that is, Dirks at times does not seem to see power as connected to culture as a symbolic order, for example when he states in his final section of Castes of mind, entitled “Toward a postcolonial historiography,” that he has “viewed culture principally as an effect of power” (2001: 313). This is in line with his earlier criticism of ethnosociology as only dealing with power within “a restricted form of cultural analysis” (1990: 75). The issue cannot be settled here, but if this were correct, Dirks would clearly deviate at this point from the historical anthropology envisioned by Bernard Cohn.

Postcolonial theory not only criticizes the complicity of culture with power but generally “targets academic disciplines.” This is the reason why “disciplines devoted to representing the Other…have been less enthusiastic, if not hostile” to this approach, as the historian Gyan Prakash wrote in the volume edited by Dirks (Prakash 1992: 376). The latter stated more generally, that “Western scholarship has consistently been part of the problem rather than the solution” (1992: 9). In this view, anthropology is not only morally dubious because of its entanglement with colonialism right up to the present, and its ignorance of that very fact, but in addition its aims (discovering “authentic” culture) are questionable, its ethnographic method (spending a year in a village and then generalizing the findings) doubtful, and its epistemological assumptions (ignoring the “knower’s involvement in the object of knowledge”) naïve (Prakash 1992: 262n; Dirks 2001:54. 79).

Thus, Dirks is one among several postcolonial scholars who abandoned ethnographic fieldwork because of its alleged indissoluble link with the colonial past and since it constructs radical alterity and implies domination. But it was also abandoned because in understanding “culture” the ethnographic project is suspect, as culture is in itself a “crime” (Dirks 1999: 158n). Again, I need not engage in a detailed criticism here. 9 It is, however, significant to point out the difference between Marriott’s and Dirks’ kinds of “ethnographic refusal.” Marriott was interested in culture and also in ethnography. He obviously stimulated his students to engage in thorough fieldwork with excellent results. However, he himself was carried away by his abstract theorizing and forced ethnographic material into his analytical categories, rather than being further theoretically inspired by it. But he did not forsake ethnography as such. Dirks deliberately abandoned both culture and ethnography. He offered anthropology a dead end and advised us to do his kind of history instead of ethnographic fieldwork.

But is it possible to have the cake and eat it too? Can we think about postcolonial theory and ethnography together, one being inspired by the other? Daniel Münster (2007), for one, provides us with a convincing example: his “postcolonial ethnography”10 of a contemporary South Indian village tries to identify postcoloniality in a particular locality and he argues that this enables us to describe Indian modernity and its postcolonial condition from a nonprivileged position. He thus combines the method of village studies with postcolonial theory (and even refrains from a complete repudiation of Dumont). Münster describes the social life in the village he studied as “postcolonial tradition” since it is predicated upon a complex interplay between the colonial heritage and a self-conscious retention of traditional ideas and practices concerning caste and religion.

Significant ordinariness

At the same time that Colonialism and culture declared the importance of historicization and deconstruction, other voices—certainly not to be misinterpreted as ahistorical—reminded us of the significance of synchronic empirical research, generalization, and comparison. Chris Fuller’s book The camphor flame (first published in 1992), certainly counts as a key publication of the modern anthropology of India. It presents the first general account of Hindu religion and society since Dumont and Mandelbaum, drawing on half a century of ethnography in India. It also exemplifies another strand in the anthropology of India, in which Jonathan Parry is the other major figure. Here British social (and cultural) anthropology took the lead without suggesting new theoretical perspectives. Fuller and Parry’s anthropology is strongly empirical and, although critical, generally sympathetic to the French tradition: they attend to structures without being structuralists. Fuller does not shy away from using terms such as “authentic” or “ethnographic evidence.” Moreover, he argues that,

insights [are] to be gained from a mainly synchronic, ethnographically based analysis. The ethnographic record clearly shows that there are enduring structures within Hindu religion and Indian society, at both the institutional and ideological or symbolic levels, and the objective of this book is to reveal and analyze some of the most important of them. (1992: 6)

A few years later Fuller qualified his stress on “enduring structures,” stating that he had “underestimated the shift away from hierarchical values that is occurring today” (1996b: 11n8). Yet he also reaffirmed that within the sphere of religion these values continue to be relevant.

Fuller, sometimes in collaboration with other colleagues, has edited a number of significant publications on current themes: the transformations of caste in the contemporary public and political environment, the modern Indian state and aspects of globalization in India (Fuller 1996a; Fuller and Bénéï 2001; Assayag and Fuller 2005a). In all these volumes, as in The camphor flame, Fuller and his coauthors bring home a crucial point, namely, that the original contribution of anthropology to a social science of India lies in the close and empirical investigation of the “ordinary” and the “everyday.” In studying the experiences, ideas and practices of “ordinary men and women” (Assayag and Fuller 2005b: 4), anthropologists can approach abstract topics such as the state, globalization, or nationalism, as lived and concrete realities. Moreover, in stressing the value of ethnography Fuller also calls attention to the fact that, if read critically, older ethnographies, although perhaps framed in out-dated theoretical perspectives, are a valuable source of information. Quite obviously this view of ethnography as the necessary and reliable basis of anthropological generalizations is worlds apart from Dirks’ indictment of the same practice.


If, as Dumont and Pocock claimed in 1957, India is one (however debatable this hypothesis was and still is), this certainly cannot be said for the anthropology dealing with it. I have tried to provide a critical overview of the main theoretical currents in the anthropology of India (post-Independence) and their development, though this sketch necessarily simplifies and selects from a much more complex discussion. Most anthropologists—not only the mavericks among them—do not follow any single theoretical paradigm or methodology but pick and choose whatever helps them make sense of their research problems. Furthermore, national anthropological traditions have been internationalized and their contours are less sharp than they used to be (see, for example, Spencer 2000).

Thus, one must perhaps speak of different theoretical currents or perspectives rather than strictly distinct strands or academic traditions. One such perspective that continues to influence the anthropology of India is Dumont’s brand of structuralism. Many still consider his stress on the study of values to be important and apply his ideas to new sets of ethnographic problems and themes, although today few anthropologists would perhaps oppose individualism to holism in the way he did.

A second, truly innovative, theoretical current that has perhaps abated to a greater extent is ethnosociology. Marriott inspired many students who did not care so much about his matrices and cubes but closely listened to what their informants had to say and wrote outstanding ethnographic monographs. Nowadays, their identity as “ethnosociologists” is less marked. Marriott’s late contribution demonstrates the lifelessness and artificiality of theory when the genuine dialogue with ethnography is abandoned.

British anthropologists dealing with India were usually sympathetic to, if critical of, Dumont but much less convinced by Marriott’s theory. Fuller and Spencer’s (1990) review of the 1980s could serve as an example of this. Figures such as Richard Burghart (an American trained at the School of Oriental and African Studies in London) and David Pocock may also demonstrate a typical eclecticism: while studying values they stress ambivalence within ideologies and highlight history and agency. It seems that many British anthropologists share an emphasis on ethnography, and an unease with allegedly universal dichotomies and over-consistent representations of cultural and social phenomena (see Fuller and Spencer 2008; Parry and Simpson 2010). Perhaps one reason that British anthropology did not put forward a general theoretical orientation of its own—after the structural functionalism developed in relation to ethnographic contexts other than India—is the widespread skepticism towards grand theories.

Something similar could probably be said about Indian anthropology, in the sense that anthropologists there follow diverse theoretical orientations and, in my view, no general anthropological theory has been developed there. However, even if L. P. Vidyarthy had been correct in asserting that “Indian anthropology…had been born and brought up under the dominant influence of British anthropology” (1977: 70f.), Patricia Uberoi, Nandini Sundar, and Satish Deshpande’s (2007) recent contribution shows that Indian anthropology can now look back on a vibrant tradition in its own right. But, as I already pointed out, characterizing national traditions is not my aim here.

Finally, while, as I have shown, the different historical approaches are heterogeneous in origin and outlook—oriented towards Indology, the Subaltern Studies Group, or ethnohistory—the postcolonial perspective with its focus on colonialism, knowledge, and power is the most dominant among them, partly assimilating the other currents and exerting a lasting influence on the study of Indian society and culture. The relationship between history and ethnography remains ambivalent in these approaches. Although at least the historical approaches of Guha and Cohn were designed to combine history with ethnography, it appears that many scholars who adopt a historical perspective soon let go of ethnography altogether. Dirks’ version of postcolonialism went a step further and explicitly abandoned the study of culture as a project and ethnographic fieldwork as a method. His “ethnographic refusal” (Ortner 1995) was deliberate and radical, while Marriott just gradually lost touch with concrete ethnographic contexts. Yet scholars like Daniel Münster (2007) also show that “postcolonial ethnography” need not be a contradiction in terms.

Although general theoretical positions are thus more or less intimately connected with particular research questions, this critical overview has been concerned with major theoretical developments in the anthropology of India and not primarily with the subjects in which different anthropologists have been interested. Had this been otherwise, much more literature would have had to be mentioned. Pioneering volumes have opened up new highly relevant areas of research [such as Veena Das’ (1990) volume on violence or David Ludden’s (2005) volume on Hindu nationalism] without suggesting new general theoretical perspectives for looking at Indian realities.11

This outline has shown that theoretical perspectives involved in the anthropology of India have been as diverse as the social and cultural phenomena being studied. No single approach can claim preeminence, nor—despite shifts in urgent issues and popular trends—can particular themes or methodologies claim hegemony over others. However, in my view, to amend an expression of Marshall Sahlins (1985: 149), a theory needs to be “burdened with the world,” that is, with ethnography and the microscopic dimensions ethnographies usually describe. As long as a scholar working from a theoretical perspective is able to engage in an ongoing dialogue with ethnographic material and can generate new analytic impulses on the basis of this correspondence that will in turn facilitate the understanding of social processes and patterns, it will be alive and well; disconnected from ethnography, a theory soon loses its heuristic value for anthropology. Since “ethnography is never impossible” (Ortner 1995: 188), the prospects of further theoretical development in the anthropology of India could be promising.


I am grateful to Helene Basu, Tony Carter, Arnout Couperus, Chris Gregory, Frank Heidemann, Carolyn Brown Heinz, Georg Pfeffer, and Maxine Weisgrau who read and commented upon different versions of this text. I also thank John Griffiths for his corrections, Sean Dowdy for his careful copyediting, and two anonymous reviewers for their feedback. Of course, the responsibility for this contribution remains with the author.


Allen, Nicholas J. 2000. “The field and the desk: Choices and linkages.” In Anthropologists in a wider world, edited by Paul Dresch, Wendy James, and David Parkin, 243–57. New York and Oxford: Berghahn Books.

Assayag, Jackie, and Christopher John Fuller, eds. 2005a. Globalizing India: Perspectives from below. London: Anthem Press.

———. 2005b. “Introduction.” In Globalizing India: Perspectives from below, edited by Jackie Assayag and Christopher John Fuller, 1–14. London: Anthem Press.

Bailey, F. G. 1969a. “Structure and change in Indian society: A review article.” Pacific Affairs 42 (4): 494–502.

———. 1969b. Stratagems and spoils: A social anthropology of politics. Oxford: Blackwell.

Barnes, Robert Harrison, Daniel de Coppet, and Robert Parkin, eds. 1985. Contexts and levels: Anthropological essays on hierarchy. Occasional Papers No. 4. Oxford: Journal of the Anthropological Society of Oxford.

Barnett, Steve. 1977. “Identity choice and caste ideology in contemporary South India.” In The new wind: Changing identities in South Asia, edited by Kenneth David, 393–414. The Hague: Mouton.

Berger, Peter, Roland Hardenberg, Ellen Kattner, and Michael Prager, eds. 2010. The anthropology of values: Essays in honour of Georg Pfeffer. New Delhi: Pearson/Longman.

Berger, Peter, and Frank Heidemann. forthcoming . The modern anthropology of India: Ethnography, themes, and theory. London: Routledge.

Breckenridge, Carol A., and Peter van der Veer, eds. 1993. Orientalism and the postcolonial predicament: Perspectives on South Asia. Philadelphia: The University of Pennsylvania Press.

Clark-Deces, Isabelle, ed. 2011. A companion to the anthropology of India. Oxford: Wiley-Blackwell.

Cohn, Bernard S. 1955. “The changing status of a depressed caste.” In Village India: Studies in the little community, edited by McKim Marriott, 53–77. Chicago: The University of Chicago Press.

———. 1968. “Notes on the history of the study of Indian society and culture.” In Structure and change in Indian society, edited by Milton Singer and Bernard S. Cohn, 3–28. Chicago: Aldine Publishing Company.

———. 1987. An anthropologist among the historians and other essays. New Delhi: Oxford University Press.

Das, Veena, ed. 1990. Mirrors of violence: Communities, riots, and survivors in South Asia. New Delhi: Oxford University Press.

David, Kenneth, ed. 1977a. The new wind: Changing identities in South Asia. The Hague: Mouton.

———. 1977b. “Epilogue.” In The new wind: Changing identities in South Asia, edited by Kenneth David, 443–515. The Hague: Mouton.

Dirks, Nicholas B. 1990. “The original caste: Power, history, and hierarchy in South Asia.” In India through Hindu categories, edited by McKim Marriott, 59–77. New Delhi: Sage.

———. 1992a. “Introduction: Colonialism and culture.” In Colonialism and culture, edited by Nicholas B. Dirks, 1–25. Ann Arbor: The University of Michigan Press.

———. 1992b. “From little king to landlord: Colonial discourse and colonial rule.” In Colonialism and culture, edited by Nicholas B. Dirks, 175–208. Ann Arbor: The University of Michigan Press.

———. 1999. “The crimes of colonialism: Anthropology and the textualization of India.” In Colonial subjects: essays on the practical history of anthropology, edited by Peter Pels and Oscar Salemink, 153–79. Ann Arbor: The University of Michigan Press.

———. 2001. Castes of mind: Colonialism and the making of modern India. Princeton: Princeton University Press.

———. 2005. “Bernard S. Cohn (1928–2003).” American Anthropologist 107 (4): 751–53.

Dresch, Paul, and Wendy James. 2000. “Introduction: Fieldwork and the passage of time.” In Anthropologists in a wider world, edited by Paul Dresch, Wendy James, and David Parkin, 1–25. New York and Oxford: Berghahn Books.

Dumont, Louis. 1970. “A Structural Definition of a Folk Deity of Tamil Nad: Aiyanar, the Lord.” In Religion, Politics, and History in India, 20–32. Paris, The Hague: Mouton.

———. 1975. “Preface to the French edition of E. E. Evans-Pritchard’s The Nuer.” In Studies in social anthropology: Essays in memory of E. E. Evans-Pritchard by his former Oxford colleagues, edited by John H. M. Beattie and R. Gordfrey Lienhardt, 328–43. Oxford: Clarendon Press.

———. 1980. Homo hierarchicus: The caste system and its implications. Chicago: The University of Chicago Press.

Dumont, Louis, and David F. Pocock. 1957. “Village studies.” Contributions to Indian Sociology 1: 23–41.

Fuller, Christopher John, ed. 1996a. Caste today. New Delhi: Oxford University Press.

———. 1996b. Introduction: Caste today. In Caste today, edited by Christopher John Fuller, 1–31. New Delhi: Oxford University Press.

———. 1992. The camphor flame: Popular Hinduism and society in India. Princeton: Princeton University Press.

Fuller, Christopher John, and Véronique Bénéi’, eds. 2001. The everyday state and society in modern India. London: Hurst & Company.

Fuller, Christopher John, and Jonathan Spencer. 1990. “South Asian anthropology in the 1980s.” South Asia Research 10 (2): 85–105.

———. 2008. “Introduction: History and theory in the interpretation of South Asian society.” In The conditions of listening: Essays on religion, history, and politics in South Asia, edited by Christopher John Fuller and Jonathan Spencer, 1–32. New Delhi: Oxford University Press.

Geertz, Clifford. 2010. Life among the anthros and other essays, edited by Fred Inglis. Princeton: Princeton University Press.

Gough, Kathleen. 1955. “The social structure of a Tanjore village.” In Village India: Studies in the little community, edited by McKim Marriott, 36–52. Chicago: The University of Chicago Press.

Guha, Ranajit. 1987. “Introduction.” In An anthropologist among the historians and other essays, by Bernard S. Cohn, vii–xxvi. New Delhi: Oxford University Press.

———. 1988a. “Preface.” In Selected subaltern studies, edited by Ranajit Guha and Gayatri Chakravorty Spivak, 35–36. Oxford: Oxford University Press.

———. 1988b. “On some aspects of the historiography of colonial India.” In Selected subaltern studies, edited by Ranajit Guha and Gayatri Chakravorty Spivak, 37–44. Oxford: Oxford University Press.

———. 1988c. “The prose of counter-insurgency.” In Selected subaltern studies, edited by Ranajit Guha and Gayatri Chakravorty Spivak, 45–86. Oxford: Oxford University Press.

Inden, Ronald B., and Ralph W. Nicholas. 1977. Kinship in Bengali culture. Chicago: The University of Chicago Press.

Khare, R. S., ed. 2006. Caste, hierarchy, and individualism: Indian critiques of Louis Dumont’s contributions. New Delhi: Oxford University Press.

Lee, Richard, and Karen Brodkin Sacks. 1993. “Anthropology, imperialism and resistance: The work of Kathleen Gough.” Anthropologica 35 (2): 181–93.

Lewis, Herbert S. 1998. “The misrepresentation of anthropology and its consequences.” American Anthropologist 100 (3): 716–31.

———. 2004. “Imagining anthropology’s history.” Reviews in Anthropology 33: 243–61.

———. 2007. “The influence of Edward Said and Orientalism on anthropology, or: Can the anthropologist speak?” Israel Affairs 13 (4): 774–85.

Ludden, David, ed. 2005. Making India Hindu: Religion, community, and the politics of democracy in India. New Delhi: Oxford University Press.

Lynch, Owen M. 1977. “Method and theory in the sociology of Louis Dumont: A reply.” In The new wind: Changing identities in South Asia, edited by Kenneth David, 239–62. The Hague: Mouton.

Madan, T. N. 1994. Pathways: Approaches to the study of society in India. New Delhi: Oxford University Press.

Mahapatra, L. K. 1977. “Gods, kings, and the caste system in India.” In The new wind: Changing identities in South Asia, edited by Kenneth David, 159–78. The Hague: Mouton.

Mandelbaum, David G. 1970. Society in India. Berkeley: The University of California Press.

Marriott, McKim. 1955a. Village India: Studies in the little community. Chicago: The University of Chicago Press.

———. 1955b. “Little communities in an indigenous civilization.” In Village India: Studies in the little community, edited by McKim Marriott, 171–222. Chicago: The University of Chicago Press.

———. 1968. “Caste ranking and food transactions: A matrix analysis.” In Structure and change in Indian society, edited by Milton Singer and Bernard S. Cohn, 133–71. Chicago: Aldine.

———. 1976. “Hindu transactions: Diversity without dualism.” In Transactions and meaning: Directions in the anthropology of exchange and symbolic behavior, edited by Bruce Kapferer, 109–42. Philadelphia: Institute for the Study of Human Issues.

———. 1990a. India through Hindu categories. New Delhi: Sage.

———. 1990b. “Constructing an Indian ethnosociology.” In India through Hindu categories, edited by McKim Marriott, 1–39. New Delhi: Sage.

Marriott, McKim, and Ronald B. Inden. 1974. “Caste systems.” Encyclopaedia Britannica 15th edition, Vol. 3: 982–91.

———. 1977. “Toward an ethnosociology of South Asian caste systems.” In The new wind: Changing identities in South Asia, edited by Kenneth David, 227–38. The Hague: Mouton.

Mathur, Saloni. 2000. “History and anthropology in South Asia: Rethinking the archive.” Annual Review of Anthropology 29: 89–106.

Münster, Daniel 2007. Postkoloniale Traditionen: Eine Ethnografie über Dorf, Kaste und Ritual in Südindien. Bielefeld: Transkript.

Nader, Laura. 2011. “Ethnography as theory.” Hau: Journal of ethnographic theory 1 (1): 211–19.

Novetzke, Christian Lee, and Laurie Patton. 2008. “Subaltern.” In Studying Hinduism: Key concepts and methods, edited by Sushil Mittal and Gene Thursby, 378–99. London: Routledge.

O’Hanlon, Rosalind 1988. “Recovering the subject: Subaltern Studies and histories of resistance in colonial South Asia.” Modern Asian Studies 22: 189–224.

Ortner, Sherry B. 1995. “Resistance and the problem of ethnographic refusal.” Comparative studies in society and history 37 (1): 173–93.

Parkin, Robert. 2002. Louis Dumont and hierarchical opposition. New York and Oxford: Berghahn Books.

Parry, Jonathan P., and Edward Simpson. 2010. “David Pocock’s Contributions and the legacy of Leavis.” Contributions to Indian Sociology 44 (3): 331–59.

Pfeffer, Georg 2006. “‘Anthropologies of mind’: Against the ethnocentric removal of tribal meaning.” Unpublished Manuscript presented at the 19th European Conference on Modern South Asian Studies, Leiden.

Pels, Peter, and Oscar Salemink, eds. 1999. Colonial subjects: Essays on the practical history of anthropology. Ann Arbor: The University of Michigan Press.

Prakash, Gyan. 1992. “Writing post-orientalist histories in the third world: Indian historiography is good to think.” In Colonialism and culture, edited by Nicholas B. Dirks, 353–88. Ann Arbor: The University of Michigan Press.

———. 1994. “Subaltern studies as postcolonial criticism.” The American Historical Review 99 (5): 1475–90.

Raheja, Gloria Goodwin. 1990. “Centrality, mutuality, and hierarchy: Shifting aspects of inter-caste relationships in North India.” In India through Hindu categories, edited by McKim Marriott, 79–101. New Delhi: Sage.

Redfield, Robert, and Milton Singer. 1955. “Foreword.” In Village India: Studies in the little community, edited by McKim Marriott, vii–xvi. Chicago: The University of Chicago Press.

Rio, Knut M., and Olaf H. Smedal, eds. 2009. Hierarchy: Persistence and transformations in social formations. New York and Oxford: Berghahn Books.

Sahlins, Marshall D. 1985. Islands of history. Chicago: The University of Chicago Press.

Singer, Milton. 1968. “Preface.” In Structure and change in Indian society, edited by Milton Singer and Bernard S. Cohn, vii–xii. Chicago: Aldine.

Sivaramakrishnan, Kalyanakrishnan 1995. “Situating the subaltern: History and anthropology in the subaltern studies project.” Journal of Historical Sociology 8 (4): 395–429.

Spencer, Jonathan. 1997. “Post-colonialism and the political imagination.” The Journal of the Royal Anthropological Institute 3 (1): pp. 1–19.

———. 2000. “British social anthropology: A retrospective.” Annual Review of Anthropology 29: 1–24.

Spivak, Gayatri Chakravorty. 1988. “Subaltern studies: Deconstructing historiography.” In Selected subaltern studies, edited by Ranajit Guha and Gayatri Chakravorty Spivak, 3–32. Oxford: Oxford University Press.

Srinivas, M. N., ed. 1955a. India’s villages. Bombay: Asia Publishing House.

———. 1955b. “The social system of a Mysore village.” In Village India: Studies in the little community, edited by McKim Marriott, 1–35. Chicago: The University of Chicago Press.

———. 1997. “Practicing social anthropology in India.” Annual Review of Anthropology 26: pp. 1–24.

Stein, Burton. 1969. “Review of structure and change in Indian society.” Philosophy East and West 19 (4): 451–54.

Strathern, Marilyn. 1988. The gender of the gift: Problems with women and problems with society in Melanesia. Berkeley: The University of California Press.

Strenski, Ivan 2008. Dumont on religion: Difference, comparison, transgression. London: Equinox.

Uberoi, Patricia, Nandini Sundar, and Satish Deshpande, eds. 2007. Anthropology in the East: Founders of Indian sociology and anthropology. Calcutta, London, and New York: Seagull Books.

Vidyarthi, Lalita Prasad 1977. “The Rise of Social Anthropology in India (17741972).” In The new wind: Changing identities in South Asia, edited by Kenneth David, 61–84. The Hague: Mouton.

Wadley, Susan. 1977. “Power in Hindu ideology and practice.” In The new wind: Changing identities in South Asia, edited by Kenneth David, 133–57. The Hague: Mouton.

Théorie et ethnographie dans l’anthropologie contemporaine de l’Inde

Au cours des soixante-cinq dernières années, depuis l’indépendance du pays, des recherches de terrain approfondies ont été menées en Inde par des anthropologues de formation. Pendant cette période, le discours anthropologique portant sur la société indienne a développé sa propre spécificité, parallèlement à l’impact profond que l’anthropologie de l’Inde eut sur la discipline même. Cet article propose une présentation critique des perspectives théoriques développées par ces anthropologues à partir de leurs expériences de terrain. En référence à Ortner (1995), cet article défend l’usage d’un système « d’approbation ethnographique » pour le développement de perspectives théoriques. La théorisation anthropologique perdrait ainsi sa valeur heuristique (i.e., sa capacité à investiguer, comprendre, analyser et comparer les mondes socioculturels humains) quand elle cesse de dialoguer avec la réalité ethnographique.

Peter Berger teaches anthropology and Indian religions at the University of Groningen, the Netherlands. His ethnographic research in highland Orissa focuses on religion, food, ritual, social structure, and cultural change. He is the author of Füttern, Speisen und Verschlingen (2007) and coeditor of Fieldwork: Social realities in anthropological perspective (2009), The anthropology of values (2010), and The modern anthropology of India (forthcoming).


1. My aim is not to fetishize fieldwork here, although I consider it to be the most important method anthropologists are devoted to and much of what follows is about the relationship between ethnographic fieldwork and theory. I agree with Nick Allen (2000) that the discipline needs the Malinowski-type as much as the Mauss-type of researcher. Various sources and methods contribute to the general anthropological endeavor. The case of textual sources is particularly obvious in India as the combination of ethnography with textual analysis was at the heart of Dumont and Pocock’s Contributions to Indian Sociology. Moreover, there are other highly significant material, historical sources, and other places beside the field (such as the desk and the museum). “Ethnographic refusal,” in my view, thus not only refers to the refusal to do fieldwork but more generally to the refusal to seriously engage with the specific life worlds of humans, with their ideas, institutions and practices (what I describe as “ethnographic reality” at several points in this paper), or their “socio-cultural forms” (Allen 2000: 247). This is perhaps similar to Pocock’s claim that anthropology should be an “empirical philosophy” (Pocock quoted in Dresch and James 2000: 4).

2. I do not use diacritics in this paper.

3. See Contributions to Indian Sociology Vol. 5 (NS) and Journal of Asian Studies Vol. 35.

4. Ronald Inden played an important part in developing the ethnosociological approach. Besides the texts written together with Marriott he analyzed “Bengali kinship” in ethnosociological terms with the anthropologist Ralph W. Nicholas (Inden and Nicholas 1977). While thus acknowledging his contribution I will speak mainly of Marriott here.

5. See Spivak (1988), O’Hanlon (1988), and Sivaramakrishnan (1995) for critical reviews of the Subaltern Studies Group.

6. For critical reviews see Fuller and Spencer (1990), Madan (1994, ch. 5), and Contributions to Indian Sociology Vols. 24 (2), 25 (2), 26 (1).

7. Notwithstanding the fact that Marriott’s more recent work shares certain features with Dumont’s perspective (though differing in many other respects), for example, the idea that India is fundamentally religious and that it is principally different from Western culture.

8. This volume is one of many dealing with the topic from the 1990s. Other examples are Breckenridge and van der Veer (1993), Pels and Salemnik (1999).

9. See Lewis (1998, 2004, 2007) and, with special reference to India, Münster (2007) and Pfeffer (2006).

10. All translations from German are mine.

11. For an overview of contemporary fields of anthropological research in India, see the excellent volume edited by Isabelle Clark-Deces (2011).