Agents and Conflict: Civil War Resolution As A Complex Adaptive System

Michael Findley, Political Science, UIUC

Civil wars pose one of the most frequent threats to peace in the post-WWII era. Parties opposing peace have successfully subverted negotiated agreements in contexts as diverse as Rwanda, Northern Ireland, and Bosnia. In Rwanda, a Hutu faction subverted the 1993 Arusha Accords and led the genocide of approximately 800,000 Rwandans. This project addresses (1) the conditions under which such groups "spoil" peace processes, and (2) the emergent effects of this behavior. Existing models of conflict resolution are primarily game-theoretic and assume away important actor and conflict characteristics. Empirical work demonstrates, nonetheless, that conflict resolution is both complex and uncertain. As such, agents are likely to be heterogeneous, incompletely informed, and thus, boundedly rational. Moreover, they are capable of learning from history and adapting their behavior - all hallmarks of a complex adaptive system. Accordingly, I employ an agent-based model to capture these characteristics, and the results of the model suggest novel explanations of civil war resolution.

Audio Slides