Locke is often cited as a precursor to contemporary natural kind realism. However, careful attention to Locke’s arguments show that he was unequivocally a conventionalist about natural kinds. To the extent that contemporary natural kind realists see themselves as following Locke, they misunderstand what he was trying to do. Locke argues that natural kinds require either dubious metaphysical commitments (e.g., to substantial forms or universals), or a question-begging version of essentialism. Contemporary natural kind realists face a similar dilemma, and should (...) not appeal to Locke in their defense. (shrink)

According to a doctrine that has been popularized by Kripke and Putnam, a natural kind term like 'bird' rigidly designates the kind with the microstructure of sample birds. This microstructure is the essence of birdhood, so our learning what the relevant microstructure is our discovery of the kind's essence. We have discovered that some statement like 'The bird is the taxon with such and such DNA structure' is true. Further, it is commonly added, the discovered microstructure is essential to each (...) particular bird. ;I argue that this general picture is mistaken. I do accept and defend the view that terms like 'bird' are rigid designators. However, the picture at issue has other difficulties. Individual birds are not essentially birds. And the view that biologists have discovered or will discover that the bird kind is identical to some DNA structure or other type of structure seems wrong. We don't discover just what particular characteristic a kind is to be identified with. Rather, the meaning of natural kind terms changes; as science advances, more sophisticated concepts attending kind terms replace older ones. So the Kripke-Putnam account of our discovering the truth of theoretical identity statements seems wrong. ;Fortunately, the account retains some interest despite its falsity, because of its association with the claim that statements like 'The bird is the taxon with such and such DNA structure' are necessary but a posteriori. The above sentence does not express a truth, much less a necessary truth, but the idea that it could have expressed the sort of necessary a posteriori truth it is supposed to express is epistemologically interesting. The very possibility of such necessary a posteriori knowledge deserves some account. I offer an account, arguing that this sort of knowledge is no more impressive from an epistemological standpoint than is knowledge of analytic necessity: indeed, I argue, it is best understood as knowledge of a species of analytic necessity. Independently of this claim, I argue in a concluding discussion that analyticity ought to be accepted by proponents of the necessary theoretical identity statements in question. (shrink)