Forbes takes a critical look at the "contact hypothesis" - the assumption commonly held by social scientists that increased contact between different ethnic groups gives each group more accurate information about the other and thus reduces friction. By distinguishing aggregate from individual relations, Forbes suggests a way out of the perplexities induced by current social science literature on prejudice and discrimination. Scientific research suggests that increased contact between culturally distinct groups in some cases gives rise to more intense conflict. Yet individuals who get to know each other better generally like each other better. Can these apparently conflicting generalizations both be true? asks Forbes. They are, he argues, and he takes contemporary social science to task for failing to show how and why this is possible. The author clarifies the weaknesses of contact theory, develops an alternative "linguistic model" of ethnic conflict, and concludes with penetrating reflections on the politics and methodology of the social sciences today.