A mathematical model previously used to describe chemical interactions may also be able to predict locations where ethnic violence is likely to break out, according to research recently published in Science by Yaneer Bar-Yam '78, PhD '84, and colleagues at the New England Complex Systems Institute. Bar-Yam, the institute's president, explains that the model predicts violence on the basis of a given area's ethnic composition. The basic conclusion is that "when people are well mixed, they're not likely to engage in violence--and also not when they're well separated, and there are clear boundaries between groups," he says. "It's in the intermediate case, where there are patches of a characteristic size and poorly defined boundaries, that we would expect to see violence."

When the researchers tested the model against reports of violence in India and the former Yugoslavia, they saw strong correlations. Bar-Yam says this shows that it may not be necessary to analyze complex social factors in order to predict trouble. "Violence takes place in many places in the world, and the circumstances of those places are quite different, from cultural and economic points of view," he says. "We wanted to be able to be very clear about identifying locations where violence takes place. But it doesn't mean that the other factors don't play a role."

Xavier de Souza Briggs, an associate professor of sociology at MIT and the director of the Institute's Community Problem-Solving Project, which provides free problem-solving tools for people and institutions, says policies directing population settlement have been controversial in the past but aren't entirely "off the menu" now. He says that by identifying bad population mixes, the model could be invaluable in determining how to house people in temporary settlements, such as those that become necessary in a crisis. He adds, however, that interracial friendships and other ties based on nonethnic aspects of identity, such as profession, can help relieve ethnic tensions. In permanent settlements, he says, it may be easier to use public policy to influence such ties through institution building than to change where people live.