Perhaps it was because a Kim student was in the room, but the SMU mind group was hard pressed to come up with objections to the first half of Kim’s recent “Physicalism or Something Near Enough.” It is written in Kim’s usual (unusually) elegant style, and while it retreads much familiar ground, one doesn’t regret the tour. In particular his discussion of causal drainage helps to round out one’s understanding of the debate, and his chapter on substance dualism is one of those pieces which should find its way into standard texts. Despite all this, I think we all felt a need for further explanation on at least one point: the status of non-species specific mental properties.
Famously, Kim thinks that the multiple realizability of a property impugns its status as a natural kind rather than guaranteeing its independent respectability as a property. Multiply realizable properties cannot be reduced—strictly speaking–but why should we wish to reduce such causally heterogeneous properties? Instead we should reduce the property instances. For example, Martians and humans might both have pain, but for Martians silicon-based states play the pain role and for humans carbon-based states play the pain role. We should not say, though, that there is a property pain that is irreducible and that we both share. Instead, we should identify Martian pains with the silicon states that play the appropriate roles, and we should identify human pains with the carbon states that play the appropriate roles. So there is the question—what about pains “in general”? We’ve reduced Martian pains and Human pains, but what about pain?
Here is where it seems Kim has changed his view. In Mind in a Physical World, it seems his answer was that pain in general is merely a concept under which all of its particular realizations fall. This is a sort of irrealist view about pains in general. In the new book, however, Kim seems to shy away from that answer. He still holds that “Neural bases may differ for different instances of pain, but individual pains must nonetheless reduce to their respective neural/physical realizers.” (Kim, 2005, 25) But the change seems to come next:
…the problem of mental causation has a solution for all pain instances. But what of the causal efficacy of pain itself? What should we say about the causal powers of pain as a natural kind? The answer is that as a kind pain will be causally heterogeneous… Pain itself will lack the kind of causal/nomological unity we expect of true natural kinds, kinds in terms of which scientific theorizing is conducted. …On this reductive account, pain will not be causally impotent or epiphenomenal; it is only that pain is causally heterogeneous. (Kim, 26)
So here it seems that Kim wants pain in general to be a property, only a causally heterogeneous one. This realism about pains in general seems a genuine change for Kim. One worries, however, that this concession is not innocuous, for two reasons:
1. We now have two distinct properties that might explain why I say “ouch!”, Howell-pains and pains. Why is there no competition here? If there is no competition here, why think there is competition between functional properties and their realizers?
2. Does pain really play a causal role? Call “schmains” states that consist of either an instance of Howell pains or an instance of being The Great White Whale. Do schmains ever make me say “ouch”? Are there such things? (Do my C-fibre firing and Moby Dick really have anything in common?) My intuition says that there are no schmains, and that even if they were, they could not be said to cause my exclamations. If one says this about schmains, it seems one should say this about pains—unless pains have some further sort of “real unity” that schmains lack. But to say this, I take it, is to become a nonreductivist.
It could be that I am misreading Kim here, but if I’m not, the distance between him and a nonreductivist seems to be shrinking a bit.
Hi Robert,
I don’t think we disagree, although a difference of opinion could arise at a deeper, more detailed level. According to me, there is no causal competition between M and its realizer P, on a given occasion of M’s instantiantion, because the M-instance is identical — or “token” identical–with the P-instance. There is nothing M on this occasion over and above the instantiation of P. Ultimately, as you note, I would embrace the position that M, which has been functionalized, is not a “genuine” property, and that it is better treated as a concept that groups properties (at the realizer level) according to their causal characteristics. Some people, like Ned Block (I believe), would insist that M is a genuine property, a property that gets realized by P1, P2, …, at this world, by Q1, Q2, … at world w1, by R1, R2, … at world w2, and so on, all depending on what laws and hence what causal relations are operative at those worlds. On this sort of approach, there would be a genuine causal competion between a given instance of M and the P-instance that realizes it, and, as you point out, the causal status of the M-instance is going to be undermined by the P-instance. To think that M is a genuine property is ontologically extravagant, and it doesn’t do any good anyway since M turns out to be epiphenomenal. That, at any rate, is one train of thoughts, and a natural one, too.
Of course, most of us think of pain as essentially phenomenal and qualitative. If that is the case (and I believe it to be so), then pain is not functionalizable, and talk of realizers does not apply to pain. That is to say, according to the way I frame the discussion, realizer talk makes sense only for functional properties/concepts/predicates. So the causal role of pain doesn’t get preempted by its realizers since it has no realizers (strictly speaking). But that is not the only way a event/state’s causal status can be preempted. Pain presumably has neural correlates (whether there is one or more than one correlate doesn’t make a difference here), and its neural correlates may very well threaten the causal claims of pain. Unless, that is, pain is identity reduced to neural states; that is, for some neural state N, pain = N. Then pain would have all the causal powers of N and enter into just those causal relations in which N enters. The main problem, of course, is whether we are entitled to identities like “pain = N”. In my view, such identities are incredible–even more so than the idea that pain is a functional state/concept.
Left by Jaegwon Kim on October 17th, 2006