Research Group
in Analytic Philosophy

Abstract

Generics in natural language allow us to formulate generalizations that can be variously described as non-strict, as tolerating exceptions, or as holding all things equal or typically. Classic examples are "ravens are black" and "tigers have stripes". Semantically, generics are puzzling because they seem to sometimes have very strict truth-conditions and very weak truth-conditions at other times. "Ravens are black" seems to require that most ravens are black, but "lions have manes" is true if (some of) the males do, regardless of what is true of females. Semantic theories for generics seek to account for this variability by invoking a mixture of linguistic and non-linguistic resources, where the latter might include a kind of probabilistic judgment, primitive inductive procedures, or a proprietary notion of normality.

In this paper, I discuss how current psychological work on categorization and generic judgment bears on the various semantic hypotheses, using a theory of generics in terms of natural kinds I've developed elsewhere as a case study. I'll argue that, with some qualifications, semantics and psychology are independent of each other.