Collective Learning in Ongoing Political Conflicts

Abstract

Analyses of political conflict typically concentrate on various structural dimensions and neglect the underlying dynamic processes. In this research, we depart from that practice and take the collective learning process inherent in large-scale and ongoing social conflict as our focus. A mathematical model and its theoretical rationale are developed and evaluated. The model, derived from well-grounded behavioral propositions, assumes that the trend in success/failure rates in this type of conflict is accounted for by the relative learning rates of the parties in conflict. In tests involving several sets of data on conflicts between government and opposition groups in Latin America from 1946 through 1975, the model quite accurately describes the distribution of outcomes.