The making of a riot narrative


back to issue

FOR a society where sectarian tensions are endemic and group violence routine, India has produced surprisingly few, if any, comprehensive works on what Donald Horowitz calls the ‘deadly ethnic riot’ (which we refer to as communal violence). Academic studies of the phenomenon suffer on two counts: one, a lack of scholarly ambition; two, the limitations of interpretive frameworks.

At a level of abstraction, riot studies in India resolve into two broad categories. On the one hand are ‘causal’ accounts of violence and, on the other, ‘hermeneutic’ narratives that seek to understand the meaning and symbolism of violence as a form of collective social action. (Incidentally, this has parallels in the philosophy of social science, particularly in the debate about cause vs. reason.)

In the first camp, which by and large dominates the Indian riot scene, there is a preponderance of what was once loftily called ‘political economy’. The inspiration here is basically ‘Marxist’ (read, some form of economism) – often vulgarly so – but, depending on the sophistication of the scholar, there is room for more, including the autonomy of the political. A good example of this is Asghar Ali Engineer, easily the most prolific of the riot ‘specialists’. The other approach is concerned with violence as social text. Or, questions of meaning, agency and symbolism.

The methodological impulse here draws heavily upon Durkheimian sociology, but there is a fair sprinkling of social psychology and, in the odd case, psychoanalysis as well. From the occasional writings of sociologist Veena Das on communal violence, especially in the context of the 1984 Sikh pogrom, to the meditations of Ashis Nandy on nationalist hyper- masculinity or the analytical excavations of Sudhir Kakar – the list represents a heady interdisciplinary mix.

While the latter approach has yielded many interesting insights into the nature of communal prejudice and collective identities, it does not raise, much less answer, the basic ‘why’ questions about violence. To find an answer to these, one can do no better than look at the causal accounts constructed by scholar-activists such as Engineer.

In what follows, I offer a schematic reading of Engineer, pointing out the limitations of the causal approach he and others like him follow. The intent is to use Engineer as a stalking horse for David Horowitz’s encyclopedic work on the morphology and dynamics of ethnic group violence.

According to Engineer, the roots of sectarian violence in India lie, broadly speaking, in the conjunction of two systemic forces: uneven capitalist development and competitive politics. Even assuming that this is so, it begs several questions, starting with some very basic ones concerning space, time and history. Why, for instance, some cities and not others fall to repeated episodes of collective violence. (After all, the forces Engineer invokes ‘apply’ equally to the whole of India, if not the entire subcontinent). What explains the timing of a riot? Why does the quantum and brutality of violence vary from one riot to the next? If riots are a species of purposive-rational action, then what explains the irrational brutal and expressive forms that they invariably take? (In Horowitz’s terms, why do riots represent ‘a bizarre fusion of coherence and frenzy’?)



Engineer has no real answers to any of these questions. His preferred method is to call upon a number of ‘contingent’ factors – from ‘precipitating’ events to rumours to the collusive role of the state machinery – to ‘explain’ particular episodes of violence. Admittedly, Engineer devotes a great deal of attention to mapping out the macro political context of violence but this is more in the nature of a deux ex machina. The effort is marred by an apparent arbitrariness and subjective license. From Shah Bano to Advani’s rath yatra to the Ramjanmabhoomi movement – to take a random sample – a small subset of ‘political’ events is chosen from a larger set to underline the ‘communalisation’ of the larger environment.

The point is not that this macro narrative is not relevant to understanding the micro communal politics of a mofussil town. Rather that the relevance has to be empirically demonstrated and precisely delineated. Invoking a national political narrative is, heuristically, a non-starter precisely because of its scope. It is too general. Among the first principles of the sociological method, there is one which says: What (potentially) explains everything, actually explains nothing.



A second problem arises from Engineer’s proto-Marxist belief that communal riots are nothing but a species of class conspiracy which is aimed at thwarting working class unity. This places Engineer in the category of those who look at collective violence as instrumental action, with the qualification that the instrumentality of the violence is meant to uphold not the interests of those who carry it out but of political elites who profit from it.

This is the old Marxian bogey of ‘false consciousness’. The question why the poor or the working class so easily succumb to a conspiracy which is not in their material interest is never raised. To get around the problem, there is sometimes a facile acceptance of the popular-political myth that communal violence is the handi-work of organised gangs of thugs hired by politicians or other vested interests. But this does not take care of the problem. It simply shifts it to a different level.

If violence is carried out by paid goons, then how does it win the approval of the aggressor community? Engineer denies the possibility, maintaining the fiction that communal violence is in the nature of a moral aberration which has no sanction in larger society. Ironically, there is sometimes a tacit acceptance of this approval in his categorization of communalism as a political ideology. The villains of the piece in such cases are communal organisations whose objective it is to spread, through indoctrination, the message of communal hatred so as to prepare ordinary Hindus (and Muslims) for the ‘cause’.



The discerning will notice that common to both perspectives – whether causal or hermeneutic – is what political scientist Partha Chatterjee once called the assumption of Indian exceptionalism. Since communalism is supposedly peculiar to India – a legacy of the Raj – the investigation of its causes, forms and meanings cannot be located in a comparative, cross-national framework. A communal conflagration in Khurja or Moradabad cannot be helpfully compared with ethnic strife elsewhere in the world. This is because communal violence is not seen as a species of collective violence in general.

Speaking very broadly, the strength of this approach lies or, more accurately, ought to lie in a closer attention to the historical processes of how identities like ‘Hindu’ and ‘Muslim’ are formed, but Indian scholarship often takes these as a datum. This might partly have to do with a lack of ‘history’ in many of these studies. Communalism or the politicisation of religious communities after all is a process which played itself out, first, in a colonial context in which the distinctiveness of the colonial state (as opposed to a generic modern state) was a central driver. Historians such as Sandra Freitag, David Gilmartin and, in a more oblique way, Partha Chatterjee, Gyan Pandey and others have brought out the peculiar non-universalist nature of the public sphere in colonised societies and how that impacted on the formation of pan-Indian religious identities such as Hindu and Muslim.

But this historical understanding of community formation is largely absent in the here and now investigations of this or that riot. That is perhaps inevitable, particulary because many of the riot studies have been undertaken (by peace and civil rights activists) specifically with a view to pressurizing the Indian state to intervene in an ongoing episode of violence. Or to attend to the process of rehabilitation and justice in the aftermath. Unfortunately though, the same charge applies just as forcefully to less proximate riot studies.



This essay began with the claim that riot studies in India suffer from lack of ambition. What I mean by this is that these studies fail to raise some of the most obvious questions about group violence. Why did the incident take place where it did and when it did? Why not at any other place or time? Why is it that only some cities and not others, with a comparable mix of Hindu and Muslim populations – both economically and sociologically – are prone to the outbreak of violence? What is the exact nature of state involvement in riots? Is state connivance a crucial reason in the outbreak of violence or is it only a subsidiary cause?

What is the role of trigger or precipitating events in a riot? Do they affect the nature and course of sub-sequent violence? Was Godhra incidental to the subsequent pogrom that took place in Gujarat or was it a necessary condition? Would the scale or brutality of ‘reprisal’ killings have been different had Godhra not involved the loss of so many lives?

What is the role of rumour in a riot? Are riots spontaneous or are they organised? Who are the actual participants in a riot situation? Why does an aggressor community justify the aggression and laud the rioters as heroes of the community?

Donald L Horowitz’s study – The Deadly Ethnic Riot – takes on board all of the above questions. And then more. Cutting across disciplinary boundaries, Horowitz embarks on a comparative global study of the structure and dynamics of violent group discord. In the Indian context, the most important methodological departure made by Horowitz is in the very title of the book. Rather than focus on any one kind of identity conflict, he is interested in the whole lot of them. For Horowitz, an ethnic riot is simply an instance of group violence, an ‘intense, sudden, though not necessarily wholly unplanned, lethal attack by civilian members of one ethnic group on civilian members of another ethnic group, the victims being chosen because of their group membership.’



The only kind of group violence that is excluded from his purview is that which is explicitly political in nature, one usually directed against the state. The other violence that he excludes is targeted political violence, as for instance in terrorist killings. This definitional abstraction is obviously premised on the assumption that the specificity of the ascriptive element in a riot situation is not explanatory. Whether the groups involved in violence define themselves in terms of their religious, linguistic or racial identity, there is for the purposes of understanding a substantial common core.

Ethnic conflicts of course presuppose some conception of ethnicity. Horowitz distinguishes between ‘hard’ and ‘soft’ conceptions of ethnicity. ‘The first sees ethnic affiliations as being made of stone, while the second sees them as made of putty.’ For the hard theorist, ethnic groups are ‘ascriptive, firmly bounded entities’ which ‘persist over time’ and are ‘liable to conflict behaviour based on passion (even to the exclusion of calculation).’



For the soft theorist on the other hand, ethnic groups are entities whose boundaries are ‘problematic and malleable’. Group behaviour here is strategic rather than affective, that is to say, based on a calculation – often an individual calculation – of interests. A corollary to the hard and soft distinction is the question about ‘the proper locus of riot investigation’; the choice being between ‘macro societal’ and ‘micro strategic’.

Horowitz’s own preference is summed up in one line: ‘Cleavage drives culture, more than culture drives cleavage.’ This, he is quick to clarify, denies neither ‘the social construction of ethnicity’ nor the role of ‘cultural and political elites’ who harness group ‘passions for their own interests’. But it does deny that the elites have a wide latitude in fomenting ethnic conflict and violence. ‘The constraints of the field in which group interactions occur limit what elites can do and what interests they can pursue.’

If this sounds like common sense dressed up as academic specialism, then Horowitz is not entirely to blame. The lack of analytical acuity and the preponderance of well-meaning ‘political correctness’ in riot studies has often led to naïve assumptions about who is ultimately to ‘blame’ for the violence. But what are the constraints that the elites face? According to Horowitz, these constraints stem from the ‘cognitive’ basis of ethnic affinities and disparities. And with this, he enters a terrain which is entirely foreign to Indian riot scholars, namely, behavioural social psychology.

Without putting too fine a point on it, Horowitz’s argument is that sociality and group affiliation constitute universal human condition, ‘antedat[ing] globalisation, the modern state, the industrial revolution, even the printing press.’ So there is no getting away from ethnicity, indeed of the ‘thick, Gemeinschaft variety.’ The upside is that this very condition also means that elites cannot pull the strings as they wish, their options being limited by ‘their followers’ definition of the situation and what they would be willing to fight over.’

On balance then Horowitz stands with the primordialists who emphasise the ‘thick, compelling character of group membership’ but he wants them to appreciate better the sensitivity of ethnic boundaries and conflict to ‘changing contexts’. In any event, even identities that are so intense and deeply-held as to seem almost primordial to an external observer can quite easily be ‘recently constructed’.



Moving from ethnicity to ethnic violence, Horowitz calls into question what he calls the ‘ontology of the riot’: Is a particular riot a single event or is it a series of distinct events? ‘If, on February 1, members of group A kill three members of group B and then, on (sic) February 15 to 18, kill 30 members of Group B, is this one riot or two?’ Suppose that in the second instance, members of Group B killed members of Group A, ‘were the riots of February 1 a warm-up, a precondition, or an integral part of the events of two weeks later?’

If this is not enough of a problem, then one can add the further dimension of the different shifts and stages in a single episode of violence. What’s more, you have to assign a place within this narrative to nonviolent events? If this seems like pointless academic hair-splitting, then it is very far from Horowitz’s intent. He is especially concerned about precision in marking out the ontology of riots because ‘an excess of connectedness’ and ‘too little individuation’ can destroy ‘the concept of an event’ and will invariably shift attention away from ‘immediate causes to remote causes.’ It will also de-emphasise agency ‘in terms of who did what to whom?’ In other words, specifying the spatio-temporal contours of a riot has an immediate and direct bearing on the causal narrative. Was Godhra the trigger, or was it merely a link in a long chain?



There is a larger question/point in all this: Horowitz quite simply is not satisfied with any taken for granted assumptions about a conflict situation. His attempt is to construct an exhaustive analytical and conceptual taxonomy which renders every known and little-known facet of ethnic conflict open to inquiry. Thus he raises and, where possible, answers questions about under-investigated areas like the temporal rhythm of conflict, the nature and function of rumours, the meaning of expressive or amok vs. non-amok violence, the puzzle about targeting or victim selection, the characteristics associated with target groups, the organizers and participants, the flashpoints that trigger violence, the broader social environment that sanctions violent conflict, the location, diffusion, function and meaning of violence and, finally, a fascinating typology of violent emotions – arousal, rage, outrage and wrath.

In each instance, whether or not one agrees with him, there is a breadth and depth of scholarship which is, without exaggeration, stunning and possibly unparalleled. In the process, Horowitz systematically debunks a number of popular myths that have marred riot studies in India. Among them, those that see a communal riot as sudden, unstructured frenzy, an aberrational act of deviancy, an escalated personal quarrel, a cynically organised plot by manipulative leaders, an attempt to redress specific grievances, a manifestation of straight-forward hatred of otherness or difference.

On the positive side, Horowitz identifies ‘the concatenation of four underlying variables’ that in his view best explains the deadly ethnic riot. These are: ‘A hostile relationship between two ethnic groups’, ‘events that engage the emotions of one of the collectivities’, ‘a keenly felt sense of justification for killing’, and, not least, ‘an assessment of the reduced risks for violence that facilitates disinhibition.’ None of the four, taken in isolation, sounds particularly acute or compelling, but in this Horowitz does not perhaps do enough justice to the complexity of his own investigation. The real insights of the Horowitz study lie in the empirical and analytical richness of detail.



This is not to say that there are no shortcomings to Horowitz’s cross-cultural approach. In his emphasis on the social psychological origins of ethnicity, for instance, he runs the obvious risk of underplaying the complex social, political, economic and historical processes by which ethnic identities are articulated and formed. In that sense, there is an inevitable element of reification in his riot narrative.

To wit, a Hindu-Muslim riot in Bijnore in the 1990s belongs in the same category as an inter-racial conflict in mid-1940s America. Space and time are by themselves not relevant causal categories in understanding riots. Then again, most of the empirical material that he draws upon is already premised on the assumption that a particular ethnic description quite easily ‘fits’ a particular episode of violence.

Unfortunately, Horowitz does not always read his ‘source material’ too critically. What’s more, when he is not dealing with larger theoretical questions relating to violence, his narrative tends to lapse into treating communities as fully-formed entities who act as more or less coherent actors. The antipathy between communities can wax and wane but the lines of division between the communities are themselves fairly distinct and inviolate.

But lest this should create the impression that Horowitz is insensitive to or cannot account for complex, cross-cutting and imperfect motivations at play in a riot, then that is an overstatement. Horowitz is at pains to distinguish between the ‘ethnic’ dimension of a riot from others, the destruction of state property for instance. Or the fact that an episode of violence, lasting over several days, can sometimes be a coalescing of more than one kind of conflict. In this regard, he makes a particular mention of the Gujarat violence of the early eighties where what began as caste violence soon took on a communal colour.

But to fault Horowitz on specific matters of detail or one or another aspect of his meta-theory is to miss the wood for the trees. You can argue with this or that part of his argument, but it is impossible not to be overwhelmed by its sheer scope and vision. What it yields is a fascinating analytical schema in which every small and seemingly inconsequential, even inexplicable, aspect of a riot is carefully dissected, categorized and systematically accounted for.



The final question: Is there a practical pay-off at the end of Horowitz’s breathtaking effort? The answer is: yes and no. In a fascinating excursus in the concluding chapter, Horowitz raises the question about the decline, if not disappearance, of the deadly ethnic riot in the West. On the negative side, he discounts obvious candidates such as material prosperity or democracy as causes for the decline, focusing instead on what he called ‘important attitudinal changes’, particularly a variety of state-sponsored proposals to counter ‘prejudice’ and reduce discrimination. This, he argues, stems from a post-War political consensus in the West that nationalism, or extreme manifestations of ethnic sentiment, is to be discouraged and discredited. In other words, the declining legitimacy of ethnic antipathy eliminated support for the deadly ethnic riot in the West.



Because of this steadfast commitment from the state, he argues, the police in the West cannot be counted on for indifference in the event of group violence, the public authorities cannot be counted on for impunity, and the targets themselves cannot be counted on for passivity. Whether or not this accurately corresponds to the political history of western democracies in the last fifty years, it does point to important normative concerns. Stopping a deadly communal riot is not just a law and order problem – although it is also that to a substantial degree – but it is also looking at the systemic prejudice and discrimination that make minorities easy targets of riots.

The irony of course that post-9/11, racial prejudice, particularly prejudice against Arabs, Muslims and Islam is making a dramatic comeback in the West, aided and abetted by the millenarian rhetoric of leaders like Bush Jr. But that’s a story which takes nothing away from Horowitz’s majestic effort.

If nothing else, The Deadly Ethnic Riot sets a definitive analytical benchmark that we in India can only ignore at our own peril and theoretical impoverishment.


* The Deadly Ethnic Riot by Donald L. Horowitz. Oxford University Press, New Delhi, 2002.