In his 1993 book Inductive Inference and Its Natural Ground (MIT Press, 1993) Kornblith argues that inductive knowledge is possible by virtue of a fit between our innate psychological capacities and the causal structure of the world.
This claim is defended in his book Knowledge and its Place in Nature (Oxford University Press, 2002) where Kornblith argues that knowledge, as it is being studied in cognitive ethology, is a sufficiently robust and inductively valuable category to qualify as a natural kind.
Kornblith's case for knowledge as a natural kind provides the basis for his critique of philosophy as conceptual analysis.
At the same time, Kornblith's work in semantics indicates that his claims about conceptual analysis can, in fact, be motivated independently of his view on knowledge as a natural kind.
In particular, he has argued that semantic externalism provides the correct semantics not only for natural kinds but also for artifactual kinds — a claim that, if true, would lend plausibility to the idea that empirical investigation provides a promising philosophical method, quite independently of whether or not a majority of the objects of philosophical inquiry turn out to be natural kinds.