Quinn Truckenbrod

University at Buffalo

This is a very interesting paper which presents serious challenges to Chalmers. I will focus my comments on three main points. The first is that the title of the paper suggests a more substantial thesis than it ends up with as its conclusion. Phillips begins by telling us that positive and negative possibility cannot save the conceivability-possibility link but later backs off that position and argues only that negative conceivability fails to account for possibility while granting that positive conceivability may do it. Second, I will try to give an account of a scenario that might verify the existence of zombies by invoking the problem of other minds. And lastly, I would just like to suggest that Chalmers might be perfectly happy with the argument that Materialism is in the strongest sense inconceivable.

Let me restate what I take to be the main thesis. The mathematical and metaphysical examples show that negative conceivability is insufficient for possibility, and a zombie world is at best negatively conceivable. In order for Chalmers' anti-materialist argument to work a zombie world must be positively conceivable, but since none is, the argument fails and Materialism is still a viable theory. Now the title of the paper suggested to me that Phillips would be arguing for the stronger, more generalized, thesis that even positive conceivability is insufficient for possibility. I've been trying to extend his arguments to make the stronger claim, but haven't yet convinced myself it can be done. I think he has done a good job to show that both GCH and ~GCH are ideally negatively conceivable, but it is impossible for both to be true. Chalmers considers what he calls the NEGPOS principle, which says that ideal negative conceivability entails positive conceivability, although he admits he does not have a satisfactory argument establishing its truth. If NEGPOS does hold and both GCH and ~GCH are ideally negatively conceivable, then the CP link must be broken.

In his discussion of the difference between positive and negative conceivability Phillips is right to say that the hypothesis, S, cannot be included in the description of the scenario, because then the entailment would be trivial. But there is a sense in which all entailments, all valid arguments, are tautologous. If a given scenario verifies some statement S, then a complete description of that scenario must encode all the information encoded by S. S itself cannot be in D but all the information in S must be in D, otherwise there would be no entailment. It is up to the observer to cull that information from the description. Godel and Cohen have shown that GCH and ~GCH are consistent, but we have been unable to determine whether either is fully encoded in the Zermelo-Frankel axioms. It seems to me there are two sources for this deficit in our knowledge which prevents the move from negative to positive conceivability. Either we just aren't clever enough to see the proper connections which would enable us to pull that information out, or we don't have access to all the facts, perhaps the axiom set is not a complete description and we need to add another axiom. If the description is rich enough, an ideal reasoner should, by virtue of being ideal, be clever enough to see the necessary connections to proceed from negative to positive conceivability. If the problem is rather with the description, we can then ask if this is a human limitation or a limitation in principle. If it is just a human limitation, then the ideal reasoner should be able to locate and identify the missing axioms to formulate a compete description and make the move from negative to positive. However, a third possibility is that even an ideal reasoner would be unable to see all the facts. This seems to distinguish positive conceivability, but then the quip "God knows which is true" is called in to question. If even God cannot know which is true, does it make sense to say one or the other is true? We are left with the possibility that GCH is indeterminate. The question then becomes, indeterminate with respect to a given axiom set, or indeterminate in fact? Does it make sense to say the hypothesis must be determinate in fact? George Lakoff has recently argued that mathematical concepts, specifically conceptions of infinite numbers, are metaphorical human creations and there are no facts of the matter out there in the world. While useful, infinite numbers violate all kinds of mathematical intuitions, such as the intuition that there are exactly half as many even integers as positive integers, so the fact that contradictory hypotheses about infinite numbers are both consistent leads me to believe the problem lies with the hypothesis itself. So even if NEGPOS holds, there might still be room for Chalmers to wiggle out and maintain the CP thesis. If GCH is indeterminate then it is not necessary, and therefore both it and its negation can be both conceivable and possible.

I will move now to the argument that Chalmers' zombie world is not positively conceivable. Phillips argues that the putative zombie scenario, Z, does not verify the existence of zombies but instead verifies the existence of conscious people. Well, if we take seriously the subjectivity of consciousness, then, I would argue, Z only verifies the existence of one conscious person, namely me. For all I know, the rest of you are zombies. Principles of similarity may provide evidence that you are conscious like me, but I can only verify that I myself am conscious. Now when I die, I will leave behind a world of zombies. This only provides for prima facie positive conceivability, but if consciousness is truly subjective, then I don't see why this cannot be extended to the ideal.

Finally, I would ask in which sense of conceivability Phillips sees Chalmers' argument implying that Materialism is inconceivable, and why he thinks this works against Chalmers. It seems to me that if Materialism is not conceivable in any possibility implying sense, this can only serve to bolster Chalmers' argument that Materialism is false.