Chris Clarke will be talking at CamPoS about causal-explanatory kinds.
Race, gender, moral value, money, kinship, psychiatric disorders, species, emotions, beliefs. Which of these are real kinds and which are merely conventional? Most philosophers endorse a Causal Explanatory View of kinds: positing real kinds does causal-explanatory work, whereas positing conventional kinds does not. And they add that the Causal Explanatory View entails that inductive inferences are reliable when performed over real kinds, but not over conventional kinds. In this respect, real kinds are projectable.
I argue against both claims. Firstly, the Causal Explanatory View doesn't yield a distinction between real kinds and conventional kinds that is worthy of the name. This is because it validates many clearly gerrymandered kinds as real, I show. Secondly, the Causal Explanatory View doesn't entail that inferences over real kinds are reliable, to any degree. Projectability and causal-explanatory power are much less tightly connected that most have supposed. I spell out what is required for such inferences to be genuinely reliable, and thus what it takes for a kind to be projectable. I then urge that, if one must draw a distinction between real kinds and conventional kinds, one should draw it between projectable and non-projectable kinds.
The talk is taking place on Wednesday 4 March at 1pm in the Department of History and Philosophy of Science, Seminar Room 2.