Coarse Social Reasoning and Prosocial Choice

People often represent others through coarse social categories rather than as distinct individuals. We study how the granularity of these representations affects prosocial behavior. A social representation partitions a set of rich, person-specific “inferred stories,” and stories pooled in the same category receive a common other-regarding social weight. When the representation becomes finer, distinctions that were previously compressed become behaviorally available. We show that the resulting social weights form a mean-preserving spread of those generated by the coarser representation. If prosocial action responds convexly to social weights—as in a threshold choice problem—finer representations can increase expected prosocial action even though average social valuation is unchanged. The framework links individuation in social cognition to models of categorical and coarse reasoning in economics. Literary fiction provides a motivating application because it repeatedly exposes readers to the motives, histories, ambiguities, and contradictions that make individuals difficult to reduce to generic types.