Analytical game theory is useful for identifying optimal strategies by hyper-rational individuals, but it is difficult to apply to strategic interactions in "everyday life" -- the habits of association that generate unthinking compliance with social norms. In contrast, evolutionary game theory allows the unit of analysis to shift from the individual to the underlying rule. I use computer simulation of Prisoner's Dilemma games to show how rule-based evolutionary models can provide new insights into the self-organization of social order. In the first experiment, rules for in-group altruism flourish when cultural influence is associated with group status rather than individual attainment. In the second experiment, a xenophobic rule suddenly transforms into a rule to trust and cooperate with "outsiders" who resemble neighbors. In both cases, the outcome decisively depends on the structural embeddedness of interaction.

Back to Talks Page