Abstract Human behavior and collective actions are strongly affected by social institutions. A question of great theoretical and practical importance is how successful social institutions get established and spread across groups and societies. Here, using institutionalized punishment in small-scale societies as an example, we contrast two prominent mechanisms - selective imitation and self-interested design - with respect to their ability to converge to cooperative social institutions. While selective imitation has received a great deal of attention in studies of social and cultural evolution, the theoretical toolbox for studying self-interested design is limited. Recently Perry, Shrestha, Vose, and Gavrilets (2018) expanded this toolbox by introducing a novel approach, which they called foresight, generalizing standard myopic best response for the case of individuals with a bounded ability to anticipate actions of their group-mates and care about future payoffs. Here we apply this approach to two general types of collective action – “us vs. nature” and “us vs. them” games. We consider groups composed by a number of regular members producing collective good and a leader monitoring and punishing free-riders. Our results show that foresight increases leaders' willingness to punish free-riders. This, in turn, leads to increased production and the emergence of an effective institution for collective action. We also observed that largely similar outcomes can be achieved by selective imitation, as argued earlier. Selective imitation by leaders (i.e. cultural group selection) outperforms self-interested design if leaders strongly discount the future. Foresight and selective imitation can interact synergistically leading to a faster convergence to an equilibrium. Our approach is applicable to many other types of social institutions and collective action.

中文翻译：

---

通过选择性模仿和自利设计发展集体行动机构

摘要 人类行为和集体行动受到社会制度的强烈影响。一个具有重要理论和实践重要性的问题是成功的社会制度如何在群体和社会中建立和传播。在这里，以小规模社会中的制度化惩罚为例，我们对比了两种突出的机制——选择性模仿和自利设计——关于它们向合作社会机构融合的能力。虽然选择性模仿在社会和文化进化的研究中受到了极大的关注，但用于研究自利设计的理论工具箱是有限的。最近，Perry、Shrestha、Vose 和 Gavrilets（2018 年）通过引入一种他们称之为远见的新方法扩展了这个工具箱，概括标准的近视最佳反应，适用于具有有限能力来预测其团队成员的行为并关心未来收益的个人。在这里，我们将这种方法应用于两种一般类型的集体行动——“我们与自然”和“我们与他们”的游戏。我们考虑由一些生产集体利益的普通成员和一个监督和惩罚搭便车的领导者组成的团体。我们的研究结果表明，远见增加了领导者惩罚搭便车者的意愿。这反过来又会导致生产增加和集体行动的有效机构的出现。我们还观察到，如前所述，选择性模仿可以实现大致相似的结果。领导者的选择性模仿（即 如果领导者强烈地低估未来，则文化群体选择）胜过利己主义的设计。远见和选择性模仿可以协同相互作用，导致更快地收敛到平衡。我们的方法适用于许多其他类型的社会机构和集体行动。