Wolfgang Schwarz

This is a good point. I'd also note that the move from the possibility of zombies to the necessity (or even possibility) of dualism relies on the notion that if physicalism is false, we need something nonphysical to account for consciousness. But unless we construe the presumably nonphysical something-we-know-not-what as essentially conscious-making, zombies are every bit as "clearly and distinctly" conceivable (Chalmers' criterial phrase for metaphysical possibility) even on the dualist's picture.

Consider an arbitrary person, call him Al. On the assumptions of dualism, Al comprises or instantiates certain natural properties N and certain nonnatural properties O. The dualist postulates that N and O are necessary and sufficient conditions for consciousness. Yet we can "clearly and distinctly" conceive of zombie-Al, who is a duplicate of Al in respect of both N and O, but who is nonetheless phenomenally void.

This means that the zombie argument defeats dualism in just the same way it defeats physicalism. In fact, it would defeat any ontology that isn't stipulated a priori as containing ingredients essential to consciousness.

yes, also a good point: merely adding ectoplasma doesn't seem to help at all. So the lesson from the anti-materialist arguments -- assuming they work -- can't merely be that there is more stuff in the world than just physical stuff. It must be something like: there are fundamental facts about the stuff in the world that can only be captured in phenomenal vocabulary.

Hi Wo. Re your point in the last paragraph of the post, the argument that no amount of structural/dynamic truths a priori entail phenomenal truths is roughly that the only plausible route to such an entailment is a functional analysis of phenomenal concepts, and there's no such analysis. This isn't really intended as an argument against analytic functionalists, but as an argument against a priori-entailment materialists who aren't analytic functionalists. (In the taxonomy of "Consciousness and its Place in Nature", it's an argument against type-C materialists rather than type-A materialists.) The case against analytic functionalism turns on more basic intuitions: conceivability of functions without experience, knowledge gaps, explanatory gaps.

Re the penultimate paragraph: for me the conceivability of P&~Q is as strong as the conceivability of the possibility of P&~Q. Re the issue you mention, one can get around this by making Q be a truth about someone else's consciousness. Then e.g. I just need to conceive that things are physically like so (and that's all) and it's not the case that you are conscious. I think this is pretty straightforward -- e.g. it's the sort of thing we conceive all the time in the context of the problem of other minds. Claims about the conceivability of "nobody was ever conscious" aren't really required for the argument against materialism. (Though I do think it's conceivable that no-one was ever conscious. There's a big difference between "No experience could convince me of P" and "P is not conceivable".)

Re Q's point: It's not obvious that zombie argument works for arbitrary base properties O. Obviously it doesn't work when O includes fundamental phenomenal properties. It's also not out of the question that there are protophenomenal properties and concepts thereof such that there's an a priori relation between these concepts and phenomenal concepts. The zombie argument suggests that this won't work for a purely structural/dynamic base, but on the face of it these considerations don't generalize to an arbitrary base.

David is right that the zombie argument wouldn't work with respect to "fundamental phenomenal properties," since (as I said) such properties are by stipulation essentially conscious-making. But since it's the zombie argument itself that lends color to the intuition that we need to appeal to fundamental phenomenal properties in the first place, I think this stipulation begs the question.

It's a simple enough exercise, e.g., to imagine a minimal physical duplicate of our universe that is "gravitationally void." But of course the implication of an ontologically separable domain of "fundamental gravitational properties" here strikes us as artificial because (1) we realize that gravity plays a *dynamic* role in the motion of bodies, and (2) we have the linking intuitive concepts to see at least roughly how that's so. Because it's so less clear to us how the structure of physical reality plays out *dynamically* in the generation of consciousness, however, the potentially illusory quality of an ontologically separable domain of fundamental phenomenal properties isn't as transparent.