What people learn from punishment: a cognitive model

Authorities, from parents of toddlers to leaders of formal institutions, use punishment to communicate disapproval and enforce social norms. Ideally, from whether and how severely a transgression is punished, targets and other observers learn how to comply with shared social norms. Yet in light of every punitive choice, observers also evaluate the motives and legitimacy of the authority. Here, we show that the effects of punishment can only be understood by considering these inferences simultaneously. Punishment inevitably communicates information both about the transgression and about the authority’s motives, and the tradeoff between these inferences shapes when punishment succeeds and how it fails. We measured human observers’ joint inferences empirically in three pre-registered experiments (N=1254), and developed a rational Bayesian model using an inverse planning framework, that captures and explains these inferences quantitatively and parsimoniously. Individual differences in ideological authoritarianism were correlated with differing prior beliefs about authorities, and subsequent interpretations of punishment; the model predicted that observing punishment by the authority can entrench these differences and lead to further polarization. This work reveals the rational logic behind how people learn from punishment, and a key constraint on the function of punishment in establishing shared social norms.