Toward a Science of Ethnic Conflict?

The Deadly Ethnic Riot. By Donald L. Horowitz. University of California Press, 2001. 605 pp.

Ethnic Conflict and Civic Life: Hindus and Muslims in India. By Ashutosh Varshney. Yale University Press, 2002. 384 pp.

Ethnic conflict is revealed in many kinds of behavior, but it has no more striking and disturbing manifestation than the sudden and indiscriminate slaughter of members of one ethnic group by those of another, triggered by some relatively innocuous precipitating event that dramatizes a latent conflict between them. No one presumably needs to be reminded of the many deaths and injuries attributable to ethnic rioting and the enormous number of refugees it creates. Nor is there any need to give examples of its astonishing brutality. The human costs of ethnic rioting will be familiar to the readers of this journal. What deserves attention here is rather the question of its causes, and in particular of its relation to democracy.

These are the questions addressed, directly and indirectly, in these two books. Rather than dwelling on the destructiveness of ethnic rioting and deploring its injustice, both aim to throw light on its causes, with a view to effective prevention. Exemplifying a coolly scientific attitude, both make major contributions to the extensive social scientific literature on ethnic conflict and violence, but both are also vulnerable to the old objection that science, as now understood, has grave limitations when applied to the study of politics and social life. These limitations are more apparent in the more limited and methodically disciplined investigation reported by Ashutosh Varshney than in the broader and more flexible synthesis of existing knowledge offered by Donald L. Horowitz.

Varshney's impressive but relatively narrow inquiry focuses on whether there is a causal relation between civil society and ethnic violence. Do preexisting networks of civic engagement between potentially antagonistic communities provide an indispensable basis for regulating and managing their differences? Do such linkages constitute the "immune system" that can shield a multiethnic city or region from violence, so that the precipitating events—the "exogenous shocks" or "viruses"—that trigger riots elsewhere will leave it unscathed? Is the absence of such networks a necessary condition of communal violence? These are the main questions addressed in Varshney's investigation of Hindu-Muslim violence in India between 1950 and 1995.

Varshney's thesis is a variation on a familiar theme. Much has been written in the past decade about the importance of social capital for effective democratic government. The greater the social capital—the more numerous the trust-inducing private associational linkages among citizens—this reasoning runs, the better the government. For Varshney, a certain kind of social capital, one which "bridges" or "bonds" across ethnic boundaries via a dense network of voluntary engagements, is the key to ethnic peace.

The evidence for this thesis comes from a study that is at once quantitative and qualitative. A statistical analysis of the occurrence, magnitude, and location of all Hindu-Muslim riots (as reported in the Times of India between 1950 and 1995) shows no clear trend in the overall amount of such violence from 1950 until the mid-1970s, then an upward trend, peaking in 1992, when Hindu militants destroyed the Babri Masjid (or Mosque of Babur) in Ayodhya. The statistics also establish clearly that rioting in India is an urban phenomenon: less than 4 percent of recorded deaths in the 46-year period occurred in rural areas. Moreover, and more importantly, the violence is highly concentrated in a few cities (about half of all the riots took place in eight very large cities) and absent from most others.

What distinguishes the riot-prone cities from the peaceful ones? The qualitative side of Varshney's study is a systematic comparison of three pairs of cities, each pair consisting of a violent and a peaceful city having roughly similar Hindu-Muslim proportions. The main result of these comparisons can be quickly summarized: the peaceful cities are distinguished by more numerous and more important intercommunal civic engagements. Their neighborhoods are remarkably integrated, and so is the business and professional life of their citizens. In the riot-prone cities, by contrast...