APR JUN AUG 12 2004 2006 2007 56 captures 28 Apr 99 - 29 Apr 16

Close Help

Peter Marton

1. Zombies are popular nowadays in philosophical circles. Their advocates, the zombists insist that we can imagine or conceive creatures, like just you and me, sharing all the physical-chemical-biological structure with us, differing only in that they lack conscious experience. You and I know what it is to feel this or that, zombies do not. Absolute darkness is in their inside.

The zombist argues that if we can conceive that zombies are possible, then materialism fails, because it claims that there is nothing over and above the totality of basic physical facts. By this claim materialism is interpreted as a supervenience thesis: Higher level facts supervene on the fundamental ones. If this is so, the mere possibility of zombies presents a counter-example against this supervenience thesis. The general structure of the zombist's argument is, as follows:

(1) Zombies are possible.

(2) If zombies are possible then materialism fails.

(C0) Therefore, materialism fails.

It's that simple. Clearly, the materialist is embarrassed by this argument. It is not that he cannot answer to the zombist’s challenge. He has even more than one objection to it. Still, the zombist has the upper hand in this game: He can offer an easily comprehensible and appealing idea (and we all like simple but powerful ideas) and the only thing the materialist can do is—as it seems—to hover on the details.

Nevertheless, to begin the materialist's game we need to clarify some of the details (more details, more clarification in the next section). Let us begin with the materialist's thesis. The materialist holds, according, at least, to the zombist, the following principle on conscious phenomena:

(MT) All facts of (phenomenal) consciousness logically supervene on

The thesis, ascribed to him by the zombist, might give the first line of defense to the materialist. He might wonder why he should consider all the possible worlds.¹ He may argue that we can gain important insights even if materialism is restricted to the scope of nomological necessity. The zombist, however, must ascribe such a high standard to his opponent. If the scope of the materialist’s supervenience thesis is narrower, then it is possible that there is some zombie world which is, however, outside of the materialist’s interest.² The same can be said—and later we will exploit this fact—about the modal system within which the zombist argues: He needs S5 for the success of his argument. Imagine that axiom-schema s5 does not hold. According to the Kripke-style semantics, there might be then some possible world which is not accessible from the actual one. It is not impossible then that materialism is true in the actual world and all the worlds accessible from it, nonetheless, some of the worlds, not available from the actual world, are in fact zombie worlds. Summarily, if either S5 or logical supervenience is denied, then MT is not inconsistent with the existence of a zombie world. In this case the zombist should argue more substantially that his zombie world is within the materialist’s interest. Thus, to be fair to the zombist, his claim of logical supervenience will be granted here (though, not without reservation).

The zombist's position, on the other hand, can be described, as follows:

The zombist’s reason for assenting to Z is the Conceivability principle, as follows:

(CT) Conceivability is sufficient for logical possibility.

This premise is, not surprisingly, the materialists' second favorite target. There are well-known theoretical considerations and counter-examples against this presumption.³ Still, we will grant this point too. Our only focus will be that what are the zombist’s criteria for conceivability.

To summarize, we will accept, and exploit in fact, the zombist's premises, and doing so we will not follow the materialist's usual tracks in refuting the zombist. We do not aim to change the zombist's game, we want to win it.

2. According to the zombist, the fact that we can conceive zombies supports his claim that consciousness does not supervene on the physical. We might ask, however: What supports what? If I can imagine zombies, then I certainly have reason to refuse the supervenience thesis. But why and how can we conceive that zombies might exist? One of the zombist's reasons is that it is simply obvious. There is nothing explicitly contradictory in the idea that zombies might exist. Why? The answer is, I think, that we can conceive this idea just because we have doubts about the truth of MT. If I (have reason to) assent to this supervenience thesis, then I certainly will have difficulties imagining zombies. The zombist's position is circular at this point. The materialist might argue in the following way: "If you deny the supervenience thesis, then you should admit even the possibility of zombies. But if everything is the same, as in the actual world, how could our atom-for-atom replicas differ from us only in the (non-) existence of consciousness. It is just inconceivable."

I have to admit that this reasoning seems to me just as appealing as the zombist's. Unfortunately, the zombist is less impressed by this argument. As Chalmers writes about the zombie hypothesis:

These words give us some standard for conceivability, which can be expressed by the Principle of no explicit contradiction, as follows:

(PNC) A hypothesis is conceivable if it contains no explicit or explicable contradiction.⁴

The zombist claims furthermore that the burden of proof is on the antagonist of the hypothesis. If you doubt the zombie's possible existence, then prove its inconsistency, otherwise you (and everyone else) can conceive it.

Chalmers' words suggest a second important aspect of conceivability: If p and q are two rival hypotheses (i.e. p is true if and only if q is false), and it is not proved that either p or q contains (explicit or explicable) contradiction, then q's conceivability (and so its logical possibility) should not undermine p's conceivability, and vice versa.

3. The charge of methodological circularity did not impress or embarrass the zombist. So let us look for something stronger. From the previous sections we can now reconstruct the details of the zombie argument. We can postulate the following three premises which precede the zombist’s first premise ("zombies are possible"):

(1a) The zombie hypothesis contains no explicit contradiction.

(presumption, burden of proof on the materialist).

(1c) Conceivable theories are possible. (CT).

(1) Therefore, zombies are possible. (1a-1c)

Now let me recall our former second premise:

(2a) If zombies are possible then MT is false.

(C1) Therefore, MT is false.

This is the point where the zombist stops. We should not, however. As we know, MT is a supervenience thesis, so if there is a counterexample against MT, then MT is not only false, it is necessarily false—at least in S5, on which the zombist relies (as was noted above). So we can add now:

(3) If MT is false then MT is necessarily false.⁵

(C2) Therefore, MT is necessarily false.

Now we can "recycle" the zombist’s principles. It’s tautological that

(4) If MT is necessarily false, then it is impossible.

Using (the contraposition of) the zombist's Conceivability principle we can establish that

(5) If MT is impossible then it is inconceivable.

Using his Principle of no explicit contradiction we can add that

(6) If MT is inconceivable, it contains explicit or explicable contradiction.

(C3) Therefore, MT is inconceivable and contains explicit contradiction.

This conclusion, however, does not seem appealing. Materialism, even if it would turn out to be false, seems to be (at least weakly) conceivable. But this is not the point. Rather, we might point out that for concluding (C3) the burden of proof is now on the zombist.

Can the zombist accept this conclusion and can he therefore prove the charge of inconceivability and inconsistency against MT? What options, indeed, does the zombist have? As I see it, he has three options:

(i) The zombist might point out that in the case of MT conceivability is not the adequate criterion for its possibility. Doing so, the zombist maintains that there is some apparent difference between MT and his own thesis, Z. Most probably, he would argue that MT is a necessary proposition and so we cannot conceive both this proposition and its negation without mis-representing at least one of them.⁶ Nevertheless, we should not accept his reasoning. First, by this claim the zombist would tacitly admit that conceivability is not as obvious a standard for possibility as he first claimed. Furthermore, let me ask: What can we conceive, the existence or the possibility of zombies? As most zombists admit, zombies are, at least, nomologically impossible (e.g. Chalmers 37). It means, to stress the point, that zombies are impossible, and so inconceivable in the actual world. Then the answer seems to be that what we can conceive is nothing more than the possibility of zombies.⁷ But this claim is also necessary, at least in S5.⁸ So we can conclude that the zombist has no grounds to ask for a different treatment for MT than for Z.

(ii) Following a different strategy, the zombist—facing our challenge—might insist that MT is in fact inconceivable because it contains explicit or explicable contradiction. As now the burden of proof is on him, we can ask: Where exactly does the contradiction lie? There are two different kinds of answer the zombist might offer. He can either refer, or avoid referring, to the zombie hypothesis itself. Let us see both alternatives:

(iia) The zombist might claim that his argument proves conclusively that MT is inconceivable. He might say, as I imagine, that his argument provides an appropriate counter-example against MT. And then he can ask: what else should be required to prove MT’s failure? However, if the standard of conceivability is the lack of explicit contradiction, this answer is clearly mistaken. The zombie argument demonstrates that if the argument is valid with true premises, then MT is impossible. But to show that it contains explicit contradiction, the zombist must demonstrate that for some proposition, p, MT entails both p and not-p. Or, alternatively, he should prove that MT entails all propositions. Indeed, the materialist can defend MT by demonstrating that some proposition is not implied by it. A good candidate for this demonstration is Z itself. We know that MT entails that ‘zombies are not possible’. If the zombist were right about MT’s inconsistency, then it would also entail that ‘zombies are possible'. But it does not sound convincing. Indeed, the zombist grounds his entire strategy on the fact that the possibility of zombies is relevant about the truth or falsity of MT. Alternatively, if MT were to imply that ‘zombies are possible’, then the zombist could not use this premise against MT. Consequently, he needs an independent and direct proof that MT is inconsistent.

(iib) The zombist, on the other hand, might argue that there are independent reasons to believe ("to make plausible") that MT contains explicit contradiction. If this is so, then he does not need, however, the zombie hypothesis. He then would have a straightforward proof against MT, on which the zombie hypothesis is just parasitic. In this case the zombie argument is incorrigibly insufficient, since it depends on the independent reasoning that MT is contradictory.

There is a further, less obvious reason to refute the idea that MT contains explicit (or explicable) contradiction. The materialist and the zombist agree on some supervenience theses, which have the same logical form as MT has. They agree that biological, chemical, macrophysical facts supervene on the fundamental (micro-) physical facts. If this is so, then it is hard to imagine finding some argument which shows that the supervenience of conscious states imply explicit contradiction, while other higher-level states do not. It is hard to imagine, at least for me, that there is anything in the mere concept of conscious states which could lead to such a difference in the supervenience theses. The point, let me stress, is that it is possible, of course, that biological facts do supervene on the physical, while phenomenal do not, but it is questionable whether the difference is just conceptual.

It is worth noticing that this consideration, along with some of the previous remarks, suggests that—contrary to the zombist’s claim—MT is not a logical supervenience thesis. My point is that MT might be false without being inconsistent and so the scope of necessity in MT must be narrower than the logical one.

(iii) As a third alternative, the zombist might admit, in lack of conclusive proof about the contrary, that MT is free of explicit, or explicable contradiction and, therefore, is conceivable. This, however, is not a viable option for him. If MT is conceivable, then we might ask the zombist to add this statement to his premises. But doing so, his argument becomes overtly inconsistent. Furthermore, if the zombist admits that MT is conceivable, we can go back on the whole chain of his premises. Keeping in mind that one man's Modus Ponens is another's Modus Tollens (Dretske, 129) we can reverse his argument step by step, and at the end we can conclude that the zombie hypothesis is impossible, and so inconceivable. But this is weird, of course. The best result this route can offer is the conditionalization of the argument in both directions:

(C4) The zombie hypothesis is conceivable if and only if MT is

However, this conclusion was, let me remind ourselves, ruled out by Chalmers (at least in one direction) as preconceived, since it contradicts to his standard that the conceivability of rival hypotheses cannot depend on each other.

5. The zombist's attempt to prove MT's falsity fails, as we saw. His argument is either inconsistent or incorrigibly insufficient. But why? The zombist might say, and he says indeed, that each step in his argument seems sound. The source of his failure is that he lowered, on the one hand, the standard of what counts as a real possibility, and he raised, on the other, the requirement (by demanding logical necessity) for materialism.⁹

It’s time to consider what happens if the zombist weakens on his demand that MT must be a logical supervenience thesis. He might argue that the real issue is the following subtle distinction: the zombist claims that zombies are nomologically but not physically impossible. The scope of nomological necessity (in contrast with the physical necessity) is narrowed down to those worlds which share with the actual world not only the fundamental (micro-physical) laws, but also the higher-level, and bridge-laws. Then the debate between the materialist and the zombist is over whether or not there is any such world which is physically but not nomologically possible. The zombist’s position is, that since the zombie world is such a world, the materialist fails again. Notice, however, that the materialist can use the same arguments as before: He can show, once again, that if zombies are (physically) possible, then materialism not only fails, but it is also (physically) impossible and so inconceivable.¹⁰ Consequently, the zombist loses once again.

The zombist, let me put it in this way, tries to make us to believe that his hypothesis, Z, is at a different (quasi-empirical) level, than MT. It is not so, however; the zombist's hypothesis, aimed as a direct refutation of materialism, claims, or should claim, the same ontological/metaphysical position as materialism has. Claiming that there is something over and above the physical facts is as strong a position as claiming that there are no such facts. Alternatively, our conclusion, C4, makes explicit that the zombist is wrong when he pushes the burden of proof to the materialist. He is wrong, since his hypothesis has a logical equivalent, whose burden of proof, according to his own standard, is on himself.

The zombist fails to prove that materialism is untenable. His argument, however, has its own virtues, as I see. First, it supports a skeptical position about the materialist's claim.¹¹ The zombist's hypothesis cannot disprove MT, but it is certainly sufficient to raise reasonable doubts about it. Raising doubts, as we know, requires much less support, and even an unproved hypothesis can do that job. The materialist, being unable to prove that the zombie hypothesis is impossible, has to admit that his position falls short of absolute certainty.¹²

Secondly, the zombie argument has relevant diagnostic value: the answer of different philosophical approaches to the zombie argument can help us to see the consequences of these approaches. On the basis of this and other similar hypotheses we can decide which theories we find appealing and which go against our intuitions.¹³

1) See e.g. Locke, 99.

2) In the last section of this paper, it will be shown that even if the zombist weakens this condition, it could not help him.

3) Actually, there are two questions here, both raised by opponents of the zombists: (i) Is the zombie hypothesis conceivable at all? (see e.g. Dennett, 1995), and (ii) Furthermore, does this conceivable hypothesis imply its own possibility? The second question, of course, is a particular case of a more general philosophical problem about conceivability (see e.g. Van Cleve, Yablo).

4) These two principles (PNC and CT) can be easily connected to Van Cleve’s distinction of weak and strong conceivability (Van Cleve, 37). Roughly, the former principle corresponds to the weak, the latter to the strong conceivability.

5)Let me present here the details of this step. According to the zombist, MT is a logically necessary claim, therefore there is some proposition, p, which fits into the following definition:

(3a) MT=_dfNecp.

p can be read as "all fundamental facts are physical," or something like that, but its exact content, as we will see, is not important. Let us suppose now, with the zombist that

(3b) ~MT (supposition)

(3c) ~Necp (by 3a)

(3d) ~Necp® Nec~Necp (s5)

(3e) Nec~Necp (MP, 3d and 3c)

(3f) Nec~MT (by 3a)

(3) Therefore, ~MT® Nec~MT (3b-3e).

6)The problem is originally due to Kneale, 79-80. For the zombist’s answer, see Chalmers, 67.

7) For an analogous argument see Van Cleve’s Descartes-interpretation, 39.

8)Doesn’t this answer imply that the zombist can claim only that ‘it is possible that zombies are possible’? The answer is in the negative, because this proposition collapses into ‘zombies are possible’, at least in S5.

9) This claim can be supported by Yablo’s analysis of conceivability. The significance of this approach is that it allows a third alternative attitude toward propositions: if p is neither conceivable, nor inconceivable, then it is undecidable. Now we can easily conclude that the zombist falls short of offering a world that in fact verifies p (which is Yablo’s condition of conceivability), so his situation qualifies as undecidable rather than conceivable. This final conclusion, let me emphasize, is neither a presumption on our behalf, nor a claim we aimed to prove in some direct way. No, this is how we can interpret our conclusion within the framework of a theory of conceivability.

10) Notice, however, that the zombist’s reliance on CT is even shakier than it was before.

11) The zombie hypothesis is often characterized as a variation of the ‘other minds’ problem (e.g. Kirk, 43; Moody, 196). Let me emphasize that the skeptical position I suggest here does not belong to this famous problem but goes directly against materialism.

12)We should notice, however, that this skeptical position is tenable only if either MT is a logical supervenience thesis or the nomological possibility of zombies is not excluded.

13) I am deeply thankful to Jaegwon Kim, Terry Horgan, Jamie Phillips, and Dan Weiskopf, for their valuable comments and suggestions.

Chalmers, David J. (1996): The Conscious Mind. In Search of a Fundamental Theory.

New York: Oxford University Press.

Dennett, Daniel C. (1995): "The Unimagined Preposterousness of Zombies." Journal of Consciousness Studies, 2, 322-6.

Dretske, Fred (1995): Naturalizing the Mind. Cambridge, Mass.: MIT Press.

Kirk, Robert (1974): "Sentience and Behaviour." Mind, 83, 43-60.

Kneale, William (1949): Probability and Induction. Oxford: Clarendon Press.

Locke Don ((1976): "Zombies, Schizophrenics, and Purely Physical Objects." Mind, 85,

7-9.

Moody, Todd C. (1994): "Conversations with Zombies." Journal of Consciousness

Studies, 1, 196-200.

Van Cleve, James (1983): "Conceivability and the Cartesian Argument for Dualism,"

Pacific Philosophical Quarterly, 64, 35-45.

Yablo, Stephen (1993): "Is Conceivability a Guide to Possibility?" Philosophy and Phenomenological Research, 53, 1-42.