Playing Fair

Norms requiring individuals to treat their partners fairly can evolve even in populations that lack well-defined identity groups. The emergence of fairness equilibria in the bargaining problem and in Augmented Stag Hunt is analyzed with inductive and evolutionary learning models applied to populations that are not subdivided into preexisting groups. Inductive learning models applied to the bargaining problem yield distributions of equilibrium solutions centered around the egalitarian solution that corresponds to a norm of equal division of benefits. Inductive and evolutionary learning models applied to the Augmented Stag Hunt yield distributions of equilibrium solutions where each side contributes to a commonly desired good that are supported by costly punishments for noncontributors. These results support the strong reciprocity hypothesis in the social sciences without employing the controversial idea of group selection.