For lack of space we will not attempt to show just how Quausal Exclusion 2 might be derived from PEX; the interested reader may consult the argument of Macdonald and Macdonald (2006, 544-545), which uses PEX to justify a principle very similar to Quausal Exclusion 2. Rather, our main point about Quausal Exclusion 2 is that, however plausible or implausible PEX may be, and whatever difficulties might be involved in using PEX to justify Quausal Exclusion 2, if Quausal Exclusion 2 is the principle that will be used in the new supervenience argument, then there is a straightforward reply.The reply has been around in the mental causation literature, in various forms, for a few years.32 It is simply this: Quausal Exclusion 2 cannot be used to argue that the causal efficacy of mental properties is “excluded” by that of their supervenience bases, because such an application of Quausal Exclusion 2 would require mental properties and their supervenience bases to be independent. By the very definition of supervenience, there is a necessary connection between mental properties and their supervenience bases, so mental properties are not independent of their supervenience bases.
The “quausal” formulation of the mental causation problem aims to highlight the point that mental causation “ultimately involves the causal efficacy of mental properties” (Kim 1998, 37), and Kim’s challenge to the token physicalist is to explain how a mental event, even if token-identical with a physical event, can cause what it does in virtue of the mental properties it instantiates (where these are distinct from the physical properties it instantiates), and why the causal efficacy of the mental properties is not preempted by the efficacy of the physical properties. However, to merely complain that token physicalism, as such, lacks the resources to account for the efficacy of mental properties33 is not to the point: the relevant issue is whether token physicalism can coherently be supplemented with additional “substantive premises” (e.g., about the physical realisation of higher-level properties, or their implementation in physical mechanisms, etc.)34 so as to give mental properties a causal role—a role which token physicalism as such does not preclude.
9. Is there a problem about token identity?
Why has the fact that Kim’s supervenience argument begs the question against nonreductive physicalist positions that assume token identity not received wide attention? There are two fairly common beliefs that might be responsible for this. One is that “the problem of how events should be individuated” “plays no essential role in questions about epiphenomenalism” (Shapiro and Sober forthcoming, n. 3)—and this is just what Kim appears to think. The other is that there is something wrong with token physicalism, and that nonreductive physicalism ought to be formulated in terms of a Kim-style fine-grained conception of events. Our response to the first belief is implicit in what we said in the previous two sections: the supervenience arguments Kim actually gives us depend essentially on an assumption he makes about the identity criteria of events, and it is questionable whether Kim can construct a convincing alternative to the actual supervenience arguments without making that assumption. Clearly, then, the way we individuate events does play an essential role at least in the questions about epiphenomenalism raised by Kim. Our response to the second belief is that there quite simply are no good arguments against token physicalism; or, less tendentiously and more to the point, that there are no good arguments against token physicalism that are not equally good arguments against type physicalism.35 (This is unsurprising, since type physicalism entails token physicalism; so if the latter is shown to be false, so is the former.) The arguments against token physicalism are notoriously controversial,36 and at any rate, they are of no help to a friend of type physicalism like Kim since, as just remarked, if any of these arguments defeat token physicalism, they defeat type physicalism.37
The reasons that have led some philosophers to prefer “token realisationism” to token identity physicalism are far from persuasive, and/or are largely motivated by a prior commitment to a fine-grained conception of events. Boyd (1980), for example, claims that his realisationist version of nonreductive physicalism enables him to cope with Kripke’s arguments against identity theories. But he himself shows, in the same paper, how identity theories are quite able to resist Kripke’s arguments; so what makes realisationism preferable to token identity? And Papineau (1993, 24) explicitly acknowledges that his reason for preferring realisationism is that, qua terms of causal relations, events are best viewed as fact-like entities, or as structured, fine-grained events à la Kim. If so, how we individuate events again does, pace Kim, play a role in the mental causation debate.