Supervenience Argument

Our critique of Kim’s exclusion argument aimed to show that the reasons he gives for holding that nonreductive physicalism entails epiphenomenalism are not convincing, resting as they do on a conception of events that rules out from the start a classic and perfectly coherent version of nonreductive physicalism that we believe still merits consideration—a conception of events that is, furthermore, only dubiously consistent with another central premise of his arguments. To have accomplished this aim is not, of course, to have provided a positive account of how mental causation is possible within the confines of nonreductive physicalism thus construed. To provide such an account one would have to explain, minimally, how mental events can have causal efficacy not merely in virtue of their being token-identical with physical events, but also in virtue of the mental properties they instantiate. Although the question of how mental causation is possible is surely a legitimate one, it is not one that we need to address here; it suffices to have shown that there is no compelling argument that Kim (or anyone else, to our knowledge) has provided for the conclusion that a satisfactory account of mental causation cannot be provided within the bounds of nonreductive physicalism.

Still, while the question of how mental causation is possible is legitimate and important, there is something perplexing about the almost exclusive attention that it has received from Kim and others for so many years. There is no prima facie reason to suppose that the question of how mental causation is possible is more special, or more difficult, than the question of how biological, or chemical, or, indeed, physical causation is possible.³⁸ Moreover, Kim’s worry that physical causation may somehow “exclude” mental (or other higher-level) causation presupposes that there is physical causation, and that it is somehow less problematic than mental or other higher-level causation. These presuppositions, however, cannot be taken for granted. In fact, we think this stance gets matters backwards: if the reality of any causation can be taken for granted, we think the reality of causation at higher levels, including the mental, should be, whereas the reality of physical causation should be considered an open question.³⁹

The reason is this. The concept of causation we are interested in is not a philosophers’ invention; it does not get its content from metaphysical principles—Closure and others—of the sort that are discussed in the literature on mental causation, but from its use in science and common-sense understanding. And if we are interested in finding out what really causes what, we should listen to what science in particular has to say about this matter. Now it happens that the concept of cause is found only in the special sciences, and never in fundamental physics. That being so, we should accept the reality of higher-level causation (the kind that the special sciences talk about), including mental causation, and remain noncommittal, until some arguments are provided, on the reality of physical causation. Unless physics changes dramatically and its practitioners begin to talk about causation, these arguments will have to be philosophical arguments. Perhaps the most straightforward argument for physical causation is one that assumes token physicalism and the reality of causation at some higher level, call it L (L could be, but doesn’t have to be, the mental level): all L-causes are events, and all events are physical, so some physical events are causes. But this still leaves open the question of whether physical events cause other events in virtue of being instances of particular physical properties.

These may seem like strange things to say, but consider the fact that physical events are instances of physical properties. We assume that physical properties are those properties that are ascribed by predicates that occur in quantum physics (or in whatever the correct lowest-level theory turns out to be), or by open sentences containing only such predicates. Instances of physical properties are virtually never cited as causes, at least not under their physical descriptions. Thus it takes an argument to show that instances of physical properties can be causes. If we assume that all instances of higher-level properties, some of which are commonly cited as causes both in (higher-level) science and common sense, are identical to instances of physical properties, then it is clear that there is physical causation all around us; but this still leaves all the physical events that are not identical to instances of any higher-level properties unaccounted for (presumably there are many such physical events), as well as the causal powers of the physical properties themselves. Trying to identify the causes and effects, and the causally efficacious properties, in the world as described by physics is not an easy task; it would require us to have a correct and informative analysis of the conditions under which one event can be said to cause another, as well as of the conditions under which a property instantiated in a cause can be said to be causally efficacious, and it’s not obvious that we have such an analysis.⁴⁰

Not only is the reality of physical causation less obvious than the reality of higher-level causation, but arguments similar to those found in the mental causation debate could be used for “excluding” physical causation. The problem could be put in terms of a Malcolm 1965-style explanatory exclusion principle: given that we can have, in principle, a complete covering-law explanation for every physical event in the vocabulary of quantum physics, and that these explanations make no use of the concept of cause, what role is there left for physical causation to play? Physical causation is “excluded” by the initial conditions and the laws of physics, which determine, without positing any metaphysical glue binding one event to another, the occurrence (or objective probability of the occurrence) of every physical event.

Not that we think this is a good argument, but its similarity to some of the arguments used to motivate epiphenomenalist worries about mental events and properties should raise some questions about the seriousness of those worries.