Abstract

Realist accounts of natural kinds rely on an account of causation where the relata of causal relations are real and discrete. These views about natural kinds entail very different accounts of causation. In particular, the necessity of the causal relation given the instantiation of the properties of natural kinds is more robust in the fundamental sciences (e.g. physics and chemistry) than it is in the life sciences (e.g. biology and the medical sciences). In this paper, I wish to argue that there is a difference in kind between the putative natural kinds of the fundamental sciences and those of the life sciences, such that a uniform account of causation cannot capture both. The upshot is that we must either reject the claim that the kinds of the life sciences are genuine natural kinds, or accept that there are different kinds of causal relations involving the relata of natural kinds. I accept the latter. I reject the objection that the true causal relations that relate macro-level kinds are to be found by “going down a level” to causal relation at the fundamental kind, because the relevant causal mechanisms are not at the fundamental level. Since, autonomous mechanistic accounts of causal relations at the macro-level can be provided (e.g. in Biology and medicine), I argue that realism about the natural kinds of the life sciences is justified. I address the problem of negative causation as a counterexample to the positive account of causation that is entailed by realism about natural kinds in the life sciences. I argue that an acceptance of realist accounts of two different kinds of natural kind makes a uniform analysis of causation look unpromising. (277 words)