When the melting pot boils over

Violence is part of human experience, and many of its forms have been naturalised to the point where we hardly notice their existence. But the quotidian violence that is experienced in the family, the workplace, the pub, or school, bounded by rules, sanctions and expectations, is of a different order from the murderous attacks by one group on another in which the very humanity of the victim is sought to be denied and where the state is often powerless to protect or deliver justice, even where it is not itself directly complicit. During these episodes exemplary violence is inflicted on the bodies of the victims in ways that have become distressingly familiar from newspaper and other media accounts.

South Asia has been no stranger to such events, whether at the time of the partition riots in 1947, or the more recent riots that occurred in India at the time of the demolition of the Babri mosque in 1992, or the prolonged civil war in Sri Lanka, one of whose defining moments was the anti-Tamil pogrom of July 1983. Such violence quickly becomes mythologised by all concerned, and these mythologies often legitimate and "explain" further conflict, so that there is great urgency to the task of exploring and understanding what has happened. The four books under review, written from different disciplines and by authors from a range of backgrounds, go some considerable way to achieving this.

Paul Brass points out in the introduction to his edited collection of case studies that most scholars seek to place riots within a master narrative. Some of these narratives may in effect be universal, telling a story in which at times of economic, social or political disruption the unruly mob is always prone to spontaneous violence, either out of sheer bloodlust or in the hope of loot. Others are more specific, especially those that emerge out of orientalising accounts of Asian or African societies divided by tribal or communal (in this context usually religious) divisions which somehow naturally produce violent confrontations when left to themselves. The state's role is to preserve order by a brave willingness to use force whenever necessary so as to avoid greater bloodshed later on. Among recent writers a new line of argument has developed which emphasises the element of manipulation. The puppet masters can be either ambitious local politicians, or cynical businessmen anxious to destroy a rival or to obtain control, for example, of valuable land for property development. These approaches, taken separately, seem inadequate when confronted with the actuality of violence. The empirical evidence presented in each of the books under review shows, for example, how rarely riots start spontaneously. "Riot captains" or specialists in violence are encountered in most of the case studies and commonly used metaphors of tinder and sparks are superficial and misleading. Yet the crowd has to be taken seriously, and issues of agency and responsibility addressed.

One particularly instructive episode, treated at length by both Virginia van Dyke in the Brass volume and Stanley Tambiah, and referred to by Sudhir Kakar, is the anti-Sikh violence in Delhi which followed immediately after the assassination of Indira Gandhi on October 31 1984. For three or four days there were uncontrolled attacks on Sikhs which left as many as 2,000 dead. In ways that are familiar from other situations, deliberate efforts were made to dehumanise the members of the community, although women were spared.

It is clear that a prime role was played in the direction of the riots by local leaders of the Congress party (the individuals concerned have often been named, but more than a decade later none of the leading figures has been convicted). The victims of this cruelty were mostly poor day labourers, living on the margins of the city, and with little social or political connection to the principal actors in the Punjab crisis that had developed in the previous few years. The men who did the actual killing appear to have been, in part at least, local villagers for whom Delhi's rapid expansion had been disruptive in both economic and psychic ways. Even if they were paid and organised to carry out murder and mayhem, they cannot simply be regarded as hired labourers. An adequate account of a riot, in good Weberian fashion, thus needs to link together different domains of causation.

Kakar's work is a good starting point for such accounts. It focuses on a major communal riot in the large southern city of Hyderabad in 1990, and in particular on the experience of one Hindu and one Muslim locality. Using his skills and sensibility as a Freudian psychoanalyst, Kakar has interviewed a range of "warriors", in particular the pahalwans or wrestlers who provide a form of traditional local leadership while making a living from "land business", and "victims'', in order to probe the way riots change people's perceptions of themselves and their neighbours. Individuality is under pressure at the present time, Kakar argues, and people often seek refuge in group membership. Purely instrumentalist accounts of communal identity fail, in his view, to give sufficient weight to individuals' anxieties and concerns.

These anxieties of course are based on fantasies about the supposed qualities of the "other", and Kakar notes how they are passed down from generation to generation by reflecting on how he as a child came to perceive Muslims at the time of the 1947 partition massacres. This leads him to an extended analysis of speeches made by leading figures on each side of the communal divide. Sadhvi Rithambara, who played a major role in inciting communal hatred during the Ayodhya confrontation, appears to play on sexual anxieties among Hindu males, while Ubedullah Khan Azmi attempts to "trigger and stoke a persecutory anxiety in his audience", that can be resolved only by closer and closer adherence to the tenets of Islamic practice.

Kakar's approach only goes so far, however: it is not intended to provide a complete analysis of riots and their aetiology. Tambiah, who has already written on "ethnic fratricide" in his native Sri Lanka, offers the largest and in some ways most ambitious of the other books under review. He takes a series of case studies: the 1984 riots in Delhi, the post-Ayodhya riots in India, successive riots in Karachi in the 1980s, riots in Sri Lanka in 1983, and, by way of historical depth and comparison, the 1915 riots in colonial Ceylon involving primarily Muslims and Sinhalese. The studies are scrupulously drawn and could be used by researchers looking for comparative material.

Inevitably, though, Tambiah has to rely on secondary material, and the specialist will find relatively little that is new or revealing. The second part of the book moves into a more general discussion in which he wrestles with the question of how to understand collective action at times of riot. In search of leads, he goes back to the early work on crowd psychology, especially that of Gustave Le Bon, who first talked of the "collective mind", and to later writing, especially that of Elias Canetti in the 1980s, who avoids Le Bon's reactionary tone.

Tambiah's own awareness of the political and economic context of the riots he has studied, however, soon leads him to discard this as a useful approach. The more important part of the discussion concerns the way collectivities, which have usually only taken shape in the colonial period, come to see themselves as holding entitlements to power and resources; and here historians, political scientists and sociologists all have something to say.

As the 1984 Delhi riots indicate, it is the role of the state in all its aspects that has to become the special focus of attention. The police, as Brass describes, are often grossly biased in favour of one community. They take on and operationalise discourses which have emerged in the dialectic between state and society and which often stigmatise one particular group as the enemy of national unity. Individual politicians may exploit communal hostility for a range of purposes. The legacy of the colonial state, with its insistence on enumeration and classification as means of control, persists in many of the administrative practices of contemporary regimes. This is now combined with forms of economic development which are premised on national markets being integrated into the global economy in ways that can be hostile to the survival of identities that appear to challenge the basic national formula.

These themes are picked up in E. Valentine Daniel's work. Cast in a rather different mould from the others, although equally passionately concerned to elucidate the way violence is constructed, Daniel focuses on one group, the so-called "estate" Tamils who were brought to Sri Lanka to work on the tea plantations in the 19th century. Drawing extensively and illuminatingly on the philosophical insights of Heidegger and C. S. Peirce, he writes of the ways "history" can be used to create and destroy communities, and of the contrasts between epistemic and ontological discursive practices in this regard. The growing dominance of the former practice, for example the privileging of language as an essential marker of identity, makes the position of all Tamils in Sri Lanka increasingly difficult, especially for estate Tamils. They have been victims of pogroms, as in 1983, or of torture, at the hands of the state security forces, of Indian troops, and of rival Tamil militant groups. Some have been forced, or have decided, to emigrate to the United Kingdom despite increasing immigration restrictions, or to Canada and elsewhere.

What strikes Daniel is the way many Tamils have rejected the effort to construct a nationalised past, or at least one that makes exclusive claims on territory. Deprived of the opportunity to link past and future through the present, Tamils may be forced to choose between a search for a new nationality (which may in any case not be available to them) or to look for a "thoroughly new constellation of habits" of living in the world, in which a sense of national identity is replaced by something radically different.

Daniel sees the case of the Tamils as one of many in the present world where, to quote W. E. Connolly, refugees disturb "established priorities of identity/difference through which social relations are organised". One consequence, though it is on a much smaller scale than in their homeland, is prejudice and violence in their new places of settlement.

David Taylor is senior lecturer in politics, School of Oriental and African Studies, London.