Under what circumstances do people agree that a kind-referring generic sentence (e.g., "Swans are beautiful") is true? We hypothesised that theory-based considerations are sufficient, independently of prevalence/frequency information, to lead to acceptance of a generic statement. To provide evidence for this general point, we focused on demonstrating the impact of a specific theory-based, essentialist expectation--that the physical features characteristic of a biological kind emerge as a natural product of development--on participants' reasoning about generics. Across three studies, adult participants (N=99) confirmed our hypothesis, preferring to map generic sentences (e.g., "Dontrets have long tails") onto novel categories for which the key feature (e.g., long tails) was absent in all the young but present in all the adults rather than onto novel categories for which the key feature was at least as prevalent but present in some of the young and in some of the adults. Control conditions using "some"- and "most"-quantified sentences demonstrated that this mapping is specific to generic meaning. These results suggest that generic meaning does not reduce to quantification and is sensitive to theory-based expectations. (Contains 1 footnote and 5 figures.)