A "Zombie" is a special term of art in the philosophy of mind. A philosophical zombie is a hypothetical creature that is stipulated to be micro-physically identical to a normal human being, but completely lacking in conscious experience(1). Zombies are, again by stipulation, behaviourally indistinguishable from normal human beings. But to use a phrase from Tomas Nagel(2), there is nothing it is like to be a zombie. A zombie will talk about how it feels, about qualia, and about conscious experiences just like a normal human being. But there is nothing mental there. The talk is just the micro-physically generated behavior resulting from whatever input stimuli are applied.
In contemporary philosophy of mind, "physicalism" is the doctrine that all that exists, is just the physical -- the purview of the physical sciences(3). Minds and conscious experience are based on, and can be explained by, physical properties. Physicalism maintains that mental phenomena -- the "how it feels", the qualia, the propositional attitudes, and all those conscious experiences -- are the products of physical properties of physical existents. (Physicalism is also called "materialism". But "physicalism" is often preferred because of it relation to "physical sciences", and because the physical sciences also deal with "immaterial" stuff -- concrete things like forces and fields, and abstract things like numbers and phase-spaces.)
Zombies are key elements of the "Argument from Conceivability" against the physicalist conception of mind and conscious experience(4). A compressed version of the argument can be stated as --
P1 Zombies are conceivable. (A creature physically identical to a normal human being, but lacking in conscious experience, is conceivable.)
P2 If zombies are conceivable, then zombies are possible. (Conceivability entails possibility.)
P3 If zombies are possible, then physicalism is false. (If zombies are possible, then a normal human being with conscious experience must be more than just the physical.)
C Therefore, physicalism is false.
Although the argument as stated is logically valid, as an attack against the notion of physicalism, the argument suffers because both the first and second premises are quite debatable. Physicalists will maintain that the argument is unsound because the two are false. Anti-physicalists will, of course, deny this. The problem with the first premise is establishing just exactly what is meant by something being "conceivable". The problem with the second premise is establishing whether or not conceivability really does imply possibility. And just what sort of possibility needs to be established to maintain the conclusion of the argument.
David Chalmers has developed the most detailed analysis of the possible meanings of both conceivability and possibility involved in this argument(5). He details eight different notions of conceivability -- prima facie versus ideal; positive versus negative; primary versus secondary. (For the purposes of this essay, however, I will limit my consideration to just the first pair -- prima facie versus ideal. Arguments based on the other two pairs are unnecessary if the first pair does not carry the load assigned.)
Prima facie conceivability is conceivability "at first glance", or "without a great deal of thought". Ideal conceivability is conceivability with full considerations of all logical entailments. Chalmers uses Goldbach's conjecture(6) to highlight the difference. While Goldbach's conjecture and its negation are both prima facie conceivable, only one is ideally conceivable because Goldbach's conjecture is either true or false. So if it turns out to be true, conceiving it as false would involve a logical contradiction, and would thus not be ideally conceivable. It would only be prima facie conceived as false.
This separation of prima facie conceivability from ideal conceivability presents a fatal challenge to the Argument from Conceivability. Consider the zombie argument as set up this way --
P0 Either physicalism is true or physicalism is false. (I am not aware of any arguments that suggest that the Principle of Bivalence should be abandoned when considering the Argument from Conceivability)
P1 If physicalism is true, then zombies are not ideally conceivable. (It would involve a logical contradiction.)
P2 If zombies are not conceivable, then zombies are not possible.
P3 If zombies are not possible, then physicalism is true.
C Physicalism is true.
P1* If physicalism is false, then zombies are ideally conceivable. (It would not involve a logical contradiction.)
P2* If zombies are conceivable, then zombies are possible.
P3* If zombies are possible, then physicalism is false.
C* Physicalism is false.
By making clear that the notion of "conceivability" that is being employed is "ideal conceivability", it makes it obvious that the Zombie argument in particular, and the Argument from Conceivability in general, simply begs the question. Zombies are considered to be ideally conceivable only if one has already adopted (tacitly or explicitly) the premise that physicalism is false. If one has already pre-committed to a physicalist notion of mentality, then zombies are simply not ideally conceivable.
But now let's deal with the second premise -- the relation between conceivability and possibility. One can (prima facie) conceive of a flying horse -- Pegasus. But one cannot "really" conceive of a flying horse because (at least in this actual possible world) the physics of aerodynamics, and the bio-physics of muscle power dictate that anything that remotely looks like a horse could not power wings large enough to enable it to fly. A winged horse would be either too heavy to fly, or too light boned to gallop across the plains. In other possible worlds, however, where various forms of magic (and perhaps Harry Potter) or different physical laws reign, flying horses (like hippogriffs and thestrals, and even perhaps flying pigs) may be possible. But this gets us into the various notions of "possible" being employed.
David Chalmers identifies two different notions of "possible" that he applies to his discussion of zombies -- primary possibility and secondary possibility. Both of these notions he defines in terms of possible worlds. (S is primarily possible when S is true at a world considered as actual, and S is secondarily possible when S is true at a world considered as counterfactual.(5)) But whatever is true of a possible world is true only by stipulation -- possible worlds not being empirically accessible. So one is free to stipulate whatever S one happens to like, and then simply interpret the rest of the possible world accordingly (maintaining logical consistency by changing the rules as necessary). According to his definitions, magic and flying horses (and the rest of the denizens of either or both of Harry Potter's world or Peirs Anthony's Xanth world) are both primarily and secondarily possible. There are possible worlds that can be described where such things exist. So Chalmers' analysis of "possibility" is not as useful as it might be. One can easily define a possible world where Cartesian Substance Dualism (for example) is true, and zombies are therefore possible (both primarily and secondarily).
A more informative analysis of possibility is the difference between logical, metaphysical, and nomological possibility. A conception is considered logically possible (Kripke calls this "epistemic possibility"(7)) if there is no logical contradiction within it. As shown above, given a presumption that physicalism is true, zombies are not logically possible. In the absence of that presumption, zombies are logically possible. There is no logical contradiction in conceiving of a possible world consistent with (say) Cartesian Substance Dualism.
When one factors in Kripke's a posteriori necessities(8), one reduces the scope of logically possible worlds to those that are metaphysically possible. The classic example is "water is H2O". This is (according to Kripke) metaphysically necessary, but not logically necessary. Hence a possible world containing a Twin Earth(9) where water is not H2O is logically possible, but not metaphysically possible. Some physicalists would argue that because it is true, we will discover that there is an a posteriori necessary relationship (supervenience perhaps?) between physical states (properties, processes, events) and mental states (properties, processes, events). If true, zombies are metaphysically impossible. But one is not going to argue that zombies are metaphysically impossible unless one has already accepted physicalism, rendering zombies not ideally conceivable. And equivalently, one is not going to argue that zombies are metaphysically possible unless one has already accepted anti-physicalism, rendering zombies ideally conceivable.
The next level of possibility is nomological possibility. Most philosophers since David Hume have held that there could have been different natural laws than the ones that actually obtain - that the laws of nature are metaphysically contingent. It is not logically or metaphysically impossible, for example, for Pegasus to exist. It would just require different laws of nature governing lift, drag, weight, muscle strength, and so forth. But given that the laws of nature are what they are, flying horses are not nomologically possible. Physicalists will argue that given the actual laws of nature in this world, zombies are likewise not nomologically possible. Anti-physicalists will, of course, deny this. Anti-physicalists will argue that since the mental is not reducible to the physical, the mental is beyond the reach of the currently known laws of nature.
So whether or not conceivability implies possibility depends intimately on which notion of conceivability and possibility one is employing. And which notions of conceivability and possibility one entertains will depend on one's prior commitment to physicalism or anti-physicalism. David Chalmers, for example, goes to great lengths to demonstrate that there is at least one combination of his 8 notions of conceivability and 2 notions of possibility where the conceivability of zombies does imply the possibility of zombies. But then, he needs to find this result, because he is an anti-physicalist of the property-dualism sort. He needs to demonstrate that the conceivability of a zombie "proves" that physicalism is false. But despite his efforts, he is still begging the question. It does not matter whether one can conceive of a possible world where zombies are possible (primarily or secondarily). All that matters is that in this actual world, with the laws of nature that actually obtain, zombies are not nomologically possible if physicalism is true. They are only nomologically possible if physicalism is false. In other words, regardless of the efforts of various philosophers on the topic, the essay's title argument begs the question. Zombies are only conceivable if you are an anti-physicalist to start with.
In the words of Christopher Hill and Brian McLaughin:(10)
"We reject the argument as invalid: we accept ["In our world, there are conscious experiences."] but deny that ["One can imagine all the physical facts holding without the facts about consciousness holding."] follows. It is false that if one can in principle conceive that P, then it is logically possible that P; and, in particular, the corresponding conditional of the conceivability-possibility argument is false. We hold that it is an a posteriori truth that types of sensory states are identical with types of physical states. Given psychophysical identities, it is an a posteriori fact that any physical duplicate of our world is exactly like ours in respect of positive facts about sensory states."
Notes & References
(1) Kirk, Robert; "Zombies" in The Stanford Encyclopedia of Philosophy (Summer 2012 Edition), Edward N. Zalta (ed.), URL=<http://plato.stanford.edu/archives/sum2012/entries/zombies/>.
(2) Nagel, Thomas; "What it is Like to be a Bat" in The Philosophical Review, Vol. 83, No. 4 (Oct., 1974), pp. 435-450. URL=<http://www.jstor.org/stable/2183914>.
(3) Stoljar, Daniel; "Physicalism" in The Stanford Encyclopedia of Philosophy (Fall 2009 Edition), Edward N. Zalta (ed.), URL=<http://plato.stanford.edu/archives/fall2009/entries/physicalism/>.
(4) Kirk, Robert; The Inconceivability of Zombies" in Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, Vol. 139, No. 1 (May, 2008), pp. 73-89. url=<http://www.jstor.org/stable/40208892>.
(5) Chalmers, David; The Character of Consciousness (Philosophy of Mind), Oxford University Press, New York, New York, 2010. ISBN 978-0-195-31111-2.
"Does Conceivability Entail Possibility" in Conceivability and Possibility, T.Gendler & J.Hawthorne Eds.
(6) "Every even integer greater than 2 can be expressed as the sum of two primes."
(7) Soames, Scott; Kripke on Epistemic and Metaphysical Possibility: Two Routes to the Necessary Aposteriori, Cambridge Books Online, 2011. ISBN 978-0-511-78062-2. URL=<http://ebooks.cambridge.org/chapter.jsf?bid=CBO9780511780622&cid=CBO9780511780622A010>.
(8) Kripke, Saul Aaron; Naming and Necessity. Harvard University Press. Cambridge, Massachusetts. 1980. ISBN 0-674-59845-8.
(9) Putnam, Hillary; "Meaning and Reference" in Journal of Philosophy, Vol 70, No 19 (1973), Pgs 699-711. URL=<http://www.jstor.org/stable/2025079>.
(10) Hill, Christopher S. & McLaiughlin, Brian P.; "There are Fewer Things in Reality than Dreamt of In Chalmers's Philosophy" in Philosophy and Phenomenological Research, Vol 59,No 2 (Jun 1999), Pgs 445-454. URL=<http://www.jstor.org/stable/2653682>.
Gendler, Tamar Szabo & Hawthorne, John (Eds); Conceivability and Possibility, Oxford University Press, New York, New York, 2002. ISBN 978-0-198-25090-6. Kindle Edition.
Kirk, Robet; Zombies and Conscioousness, Oxford university Press, New York, New York, 2005. ISBN 978-0-199-22980-2. Kindle Edition.