Human life is riddled with norms, many though not all of which are costly for individuals to adopt. Similarly, human ecological adaptation relies on costly-behaviors that often generate non-rivalrous and non-exclusionary benefits for group-members. Yet, in a dynamic world, innovations, environmental change, and information-revelation mean that what norms are beneficial for a group to adopt will inevitably change over time. However, multiple game-theoretic models studying the various mechanisms stabilizing normative behaviors have demonstrated that the stability of a norm does not depend on the benefits it confers. In turn, explanations of normative change have either relied on group-selective mechanisms to explain the presence of adaptive norms or have failed to identify conditions under which normative change occurs. We study normative change by means of costly-punishment and conflict resolution. We identify social differentiation in goals and punishment capacity as a key condition permitting normative change. While normative change that results from such social differentiation need not be group beneficial it will be beneficial to some subset of agents in the population. We additionally discuss how the intra-societal forces of normative conflict that we study might interact with group-selective forces and in turn determine the dynamics and outcomes of group-selection.

Understanding how humans successfully stabilize public good contributions is a major ongoing question in the social and behavioral sciences. The use of targeted sanctions against defecting strategies is an important solution to this problem. However, ethnographic and behavioral evidence suggest that punishment is sometimes not used against defectors to stabilize cooperation. Sanctions instead are either light and insufficient to coerce cooperation or take the form of verbal repudiations, urging defectors to reform their behavior. Should defectors not reform, they are then ostracized from groups. We construct a cultural evolutionary game-theoretic model to study the evolution of ostracizing strategies in public goods games. We demonstrate that simple ostracizing strategies are unlikely to be evolutionarily viable and can neither encourage the evolution of contrite-defectors, who respond to punishment with cooperation, nor can invade recalcitrant-defecting populations, which ignore punishment. Motivated by the ethnographic literature, we then consider a hybrid sanctioning-ostracizing strategy that lightly-sanctions defectors before ostracizing repeat defectors. Such a strategy demonstrates clear advantages over simple sanctioning strategies. It can afford to impose light-sanctions when common because these sanctions are irrelevant when coercing future cooperation from defectors. More so, when recalcitrant defecting strategies have some possibility of arising in a population, sanctioning-ostracizing strategies dominate pure sanctioning ones, stabilizing cooperation with greater efficiency. Finally, our model makes psychological predictions concerning the reasoning processes that defectors will go through when defectors are coerced to cooperate by the threat of ostracism as opposed to sanctioning.