RESUMO
The question of how cooperation evolves and is maintained among nonkin is central to the biological, social, and behavioral sciences. Previous research has focused on explaining how cooperation in social dilemmas can be maintained by direct and indirect reciprocity among the participants of the social dilemma. However, in complex human societies, both modern and ancient, cooperation is frequently maintained by means of specialized third-party enforcement. We provide an evolutionary-game-theoretic model that explains how specialized third-party enforcement of cooperation (specialized reciprocity) can emerge. A population consists of producers and enforcers. First, producers engage in a joint undertaking represented by a prisoner's dilemma. They are paired randomly and receive no information about their partner's history, which precludes direct and indirect reciprocity. Then, enforcers tax producers and may punish their clients. Finally, the enforcers are randomly paired and may try to grab resources from each other. In order to sustain producer cooperation, enforcers must punish defecting producers, but punishing is costly to enforcers. We show that the threat of potential intraenforcer conflict can incentivize enforcers to engage in costly punishment of producers, provided they are sufficiently informed to maintain a reputation system. That is, the "guards" are guarded by the guards themselves. We demonstrate the key mechanisms analytically and corroborate our results with numerical simulations.
Assuntos
Comportamento Cooperativo , Modelos Psicológicos , Humanos , Punição , Evolução Biológica , Registros , Teoria dos JogosRESUMO
Assortative mechanisms can overcome tragedies of the commons that otherwise result in dilemma situations. Assortativity criteria include various forms of kin selection, greenbeard genes, and reciprocal behaviors, usually presuming an exogenously fixed matching mechanism. Here, we endogenize the matching process with the aim of investigating how assortativity itself, jointly with cooperation, is driven by evolution. Our main finding is that full-or-null assortativities turn out to be long-run stable in most cases, independent of the relative speeds of both processes. The exact incentive structure of the underlying social dilemma matters crucially. The resulting social loss is evaluated for general classes of dilemma games, thus quantifying to what extent the tragedy of the commons may be endogenously overcome.