South Asia includes the modern republics of India, Pakistan, Bangladesh and Sri Lanka — all formerly under British colonial rule — and the kingdom of Nepal. Although the foundations of modern South Asian anthropology lie in the systematic ethnographic reportage begun during the colonial period, the work of post-independence fieldworkers has had the greatest impact in mainstream anthropology, not least through the reaction to Louis Dumont’s celebrated attempt at theoretical and empirical synthesis, Homo Hierarch-icus (1980). The first wave of post-independence work was village-based and concentrated on questions of caste and kinship. More recent work has focused on religion, particularly Hinduism and Buddhism, gender, history and the politics of collective (or ‘communal’) identities.
Colonial society and ethnography
From the beginning, colonial officials recognized the need to understand the customs and institutions of their subject peoples, and considerable effort was therefore expended on collecting information about social structure, land systems, political and military organization, religion and law. On the whole, officials assumed that they were investigating indigenous traditions, but in reality they were participating in an inventive reconfiguration of them that played a formative role in the institutional consolidation of ‘traditional’ South Asian society under British rule. The consolidating process was complex and should not be caricatured as if the British single-mindedly imagined a ‘traditional’ India or Ceylon that they then set out to construct. Nevertheless, between the mid-eighteenth and mid-twentieth centuries, the overall direction of change is clear. Partly by design and partly by default, partly with the cooperation of their native subjects and partly in the face of their resistance, British colonialism — especially in the Indian heartland — brought into being a social order that was ‘traditional’ in the sense that it conformed more closely to the Orientalist image of an unchanging, hierarchical, religiously minded, mainly village-based society than anything that had actually existed in the precolonial era. In the Orientalist vision of ‘traditional’ society, virtually impotent kings reigned over a mass of ostensibly self-sufficient ‘village republics’, although the latter were also internally stratified by a hierarchical, hereditary caste system. This system was regulated by complex rules of ritual pollution and purity, presided over by the highest caste, the Brahmans, and anciently legitimated by the amorphous religion of Hinduism. So powerful was caste that in most of South Asia non-Hindus, as well as Hindus themselves, were under its sway.
In establishing caste as the foundation of society in India — and as crucially important even on its periphery — colonial ethnography played a critical role. Systematic ethnography effectively began in 1871, when the first national censuses of India and Ceylon were carried out. The Indian census, like all subsequent ones until 1931, included questions about caste membership. Through the decennial censuses, ‘scientific’ data about caste were amassed, which would supply the core material for the ethnographic sections of the official gazetteers, the ‘Castes and Tribes’ compendia, and many influential general treatises on caste written by census commissioners. The same materials were also important for the classic study of Indian religion by Max Weber (1967 [1917]). The censuses and the official literature dependent on them are still an invaluable record for modern anthropologists. Yet those works were also a crucial component in the imperial discourse that consolidated ‘traditional’ Indian society, as well as a principal source of the anthropological image of that society as one founded on a rigid, hierarchical caste system, from which only the marginal tribal populations were seemingly free.
The ‘village studies’ era and Dumont’s theory of caste
Most literature dependent on the censuses was determinedly empiricist, although the perspective of evolutionism was generally reproduced. Even publications from the end of the colonial period were largely uninfluenced by the rise of functionalism during the interwar years. All this changed when a new wave of British, French and American ethnographers arrived in South Asia after 1945 to conduct intensive fieldwork, principally in villages. The first anthropologist to publish a modern, functionalist ethnography, however, was Srinivas (1952), who completed in Oxford his analysis of the South Indian Coorgs before returning home to carry out a new village study.
From their reading of colonial ethnography, the post-independence ethnographers concluded that caste was the central institution of Indian society, but they also wanted to know how caste actually functioned within a local community. By the mid-1950s, the ‘village studies’ era had begun (Marriott 1969 [1955]), and for the next twenty years a stream of publications appeared, mostly about Indian villages, although similar work was also done in villages in Sri Lanka, Pakistan and Nepal. Village studies monographs normally contained either a rounded picture of village society (e.g. Lewis 1958), or a more detailed analysis of the local caste system, often in relation to kinship (e.g. Mayer 1960). In either case, however, the caste system emerged as central to village social structure, so that village studies were simultaneously local caste studies. Even in studies of villages outside India, the influence of Indian ethnography was so strong that caste tended to occupy a disproportionately important place.
Village ethnographies, partly owing to their functionalist orientation, usually depicted ‘traditional’, caste-based villages as if they were static, bounded communities only very recently affected by significant change. Not until the 1980s was it generally recognized that such villages were themselves the historical product of the consolidation of ‘traditional’ society during the colonial period. Nevertheless, despite their unhistorical character, village studies formed the bedrock of modern South Asian anthropology and provided the ethnographic basis for significant theoretical progress, especially in relation to caste and hierarchy.
Indisputably, the most important single theoretical work of the village studies era was Dumont’s Homo Hierarchicus (1980). In very simple terms, Dumont argued that caste was the unifying institution of Indian society, and that it was founded on a religious principle of hierarchy, defined by both the opposition between the pure and impure, and the absolute separation of religious status from politico-economic power. Hence the caste system, as reported in village ethnographies, was the concrete manifestation of an ancient Hindu system of values. India was an unhistorical, holistic, hierarchical society: the antithesis of historical, individualist, egalitarian society in the modern West. By insisting that India as a whole was the true object of a structuralist sociology of India, Dumont transcended the empiricist functional-ism of the village studies that provided his subject matter of caste.
The legacy of Dumont’s theory
For the anthropology of South Asia during the 1970s and 1980s, Dumont defined the main terms of debate, although his theory was more often criticized than endorsed (Raheja 1988). Post-Dumontian anthropology is very diverse, but five principal bodies of scholarship may be roughly distinguished.
First, there is work (dating back to the 1960s) that focuses on the political economy of South Asia, notably on the agrarian class structure and rural production systems (Alavi and Harriss 1989). Class — rather than caste — is treated as primary, particularly by Marxist anthropologists and sociologists, and neo-Marxist scholarship has increasingly situated itself outside mainstream anthropology. Neo-Weberian analyses focused on the relationship between caste and class, as exemplified by fBeteille’s work (1991), remain more centrally within anthropology and represent a powerful critique of Dumont’s overdrawn contrast between India and the modern West.
Second, there is the ‘ethnosociological’ school developed by Marriott and others at the University of Chicago since the 1970s (Marriott 1990), which explicitly criticizes Dumont’s structuralism and draws its inspiration from modern American cultural anthropology. The central axiom of ethnosociology is that in South Asian culture moral ‘code’ and bodily ‘substance’ are indivisible, and are constantly subject to transactions in ‘coded substance’. Consequently – and contrary to Dumont’s theory — South Asian society is not founded on rigid hierarchy, but on an array of transactions that generates a perpetual fluidity in the constitution of authority, rank and even of persons themselves.
Third, since the mid-1970s, the anthropology of religion has expanded greatly. Popular Hinduism and Buddhism have been studied most intensively, and much of this research also draws heavily on textual scholarship. The anthropology of Hinduism in particular (Fuller 2004 [1992]) has rejected Dumont’s very narrow definition of ‘religion’, which notably excluded any concern with deities, rituals and the king’s religious role. Thus the anthropology of religion in South Asia challenges the sociological reductionism pervading Dumontian theory, and insists that religion must be understood in its own distinctive terms.
Fourth, especially since the late 1970s, the anthropology of Sri Lanka, and more recently of Nepal, Pakistan and Bangladesh, has progressively escaped from the Indocentric distortions imposed by the Dumontian legacy. Thus, for example, in research among Buddhists and Muslims throughout South Asia, emphasis is increasingly placed on the autonomous — and distinctively non-Hindu or non-Indian — character of their social and cultural systems.
Fifth, partly owing to the influence of modern historiography, South Asian ethnohistory has developed since the early 1980s and has convincingly undermined Dumont’s premise that ‘India has no history’ (Cohn 1987). Furthermore, historical writing — together with feminism and other critical scholarship — has encouraged growing interest in women, low-status groups and other subordinate or subaltern sections of society, whose dissenting voices were marginalized by most earlier anthropologists.
Contemporary change and communal politics
From the mid-1980s onwards, South Asia was subject to unprecedently rapid socioeconomic change, although the most dramatic eruptions were in the sphere of communal politics. Hindu— Muslim and Hindu—Sikh violence worsened in India, Sri Lanka was torn by a civil war between Sinhalas and Tamils, and Pakistan witnessed ethnic rioting. These rapid changes, especially the bloody exacerbation of communalism, convinced many anthropologists of South Asia that they had to address these issues directly (Das 1990), and many earlier preoccupations — notably those made central by Dumont — came to look increasingly irrelevant. The anthropology of the region began to free itself from its post-Dumontian legacy, not because Dumont’s theoretical achievement was discounted, but because South Asia itself had changed so much.
