What I said in Paris, more or less
Tue 08 Jul 2014 02:23 PM
The talk I gave in France a few months ago was mostly a paper, but there were parts I hadn't written out. I had bullet points where I was saying things I've shown in other papers or in my book, because I could talk those through just fine without scripting them precisely.
I finally got around to filling in those lacunae, and today I posted a draft.
I am sure there are places where I should say more in order to be clear, but I am not sure where they are. If I elaborated on every point where I could say more, it would recapitulate most of my prior work. So, if you take a look, I'd be glad to hear which parts seem to go by too quickly.
Abstract: When we ask what natural kinds are, there are two different things we might have in mind. The first, which I'll call the taxonomy question, is what distinguishes a category which is a natural kind from an arbitrary class. The second, which I'll call the ontology question, is what manner of stuff there is that realizes the category. When causal accounts of natural kinds are assessed without clearly distinguishing these two questions, they fare poorly. The reason is that causal structure only provides an answer to the ontology question, it does so for many but not all natural kinds, and even where it applies it provides some importantly different kinds of answers. This confusion occurs when philosophers take John Stuart Mill's Kinds to be predecessors of our natural kinds, because it ignores Mill's equal commitment to what he calls natural groups as the right categories for scientific taxonomy. It occurs, too, when philosophers take homeostatic property cluster (HPC) accounts as a candidate definition for `natural kind', because being an HPC is neither necessary nor sufficient for being a natural kind. So we should think of causal accounts as just partial answers to the ontology question. I argue that most philosophers have systematically failed to distinguish these questions but that making doing so would offer several advantages.