Learning about Inductive Potential from Generic Statements

Generic statements (e.g., "Climbers drive Subarus") shape what categories people take as meaningful bases for generalization. After hearing a generic, people not only learn about the prevalence of a feature in a category (e.g., how many climbers drive Subarus), but also about the inductive potential of the category (e.g., that climbers share many features). Here, we propose a Bayesian model of how people infer inductive potential from generics. To test our model, we introduced adults (n = 284) to nothing (baseline), or to members of a novel social category accompanied by generic statements (e.g., ``Zarpies sleep in trees'') or specific statements (e.g., ``This Zarpie sleeps in trees''). We then measured inferred inductive potential by eliciting the prevalence of novel features. As predicted, generics increased while specific statements decreased the category's inductive potential, relative to baseline. Our account explains how generics facilitate the cultural transmission of social categories believed to be bases for generalization.