Previous work has shown that preschoolers—in comparison to older children and adults—tend to view categories as homogeneous, generalizing properties of individuals broadly to all category members (e.g., this dax has wings, so all daxes do). Here, we explore whether the testimony used to describe category individuals as well as children’s prior knowledge of categories attenuates their homogeneity expectations. Using a novel induction task, 4 to 7-year-olds were asked to predict the distribution of properties among members of familiar/unfamiliar animal categories based on a single exemplar. Exemplars were introduced as “special” to half of participants. Preliminary findings (N = 71) suggest that prior knowledge may contribute to beliefs about category homogeneity: responses for familiar animals varied appropriately given the real-world prevalence of each property whereas children overestimated the property’s prevalence for unfamiliar animals. The complete dataset will speak to how language choice in testimony shifts children’s beliefs about homogeneity.