The Dynamics of Anarchy

A computational model of interaction in anarchy is presented and used to predict the outcome of anarchy. Hobbes’ and Locke’s classic a priori analyses of the State of Nature are compared, including the reasons for their divergent conclusions. Several game-theoretic models of anarchy are examined that employ Hobbes’ realistic assumption that typically in anarchy some moderates most desire mutual cooperation, while other dominators most desire to exploit others’ cooperation. A priori type-based game-theoretic models yield inconsistent conclusions and rest upon unrealistic assumptions. A dynamical Variable Anticipation Threshold type-based model is explored, where individuals in anarchy modify their behavior as they learn from repeated interactions. Under quite general conditions, a population converges to Hobbes’ war of all against all even if only a small percentage of its members are dominators. This analysis gives a dynamical vindication of Hobbes’ conclusion that for a sizable population in anarchy, war is indeed inevitable.