The importance of being coherent: Category coherence, cross-classification, and reasoning

Category-based inference is crucial for using past experiences to make sense of new ones. One challenge to inference of this kind is that most entities in the world belong to multiple categories (e.g., a jogger, a professor, and a vegetarian). We tested the hypothesis that the degree of coherence of a category—the degree to which category features go together in light of prior knowledge—influences the extent to which one category will be used over another in property inference. The first two experiments demonstrate that when multiple social categories are available, high coherence categories are selected and used as the basis of inference more often than less coherent ones. The second two experiments provide evidence that ease of category-based explanation of properties is a viable account for coherence differences. We conclude that degree of coherence meaningfully applies to natural social categories, and is an important influence on category use in reasoning.