Wednesday, August 15, 2007

Zombies and anti-zombies (by guest blogger Keith Frankish)

I'm coming to the end of my stint as guest blogger. It's been fun, and I'd like to thank Eric and everyone who has commented. I thought I'd finish with another post about consciousness.

Every schoolboy knows how the zombie argument goes. Zombies -- physical duplicates of us that lack consciousness –- are clearly conceivable. If a scenario is clearly conceivable, then it is metaphysically possible (the conceivability–possibility, or CP, principle). So zombies are metaphysically possible, and therefore physicalism is false. (Physicalism is the view that consciousness supervenes metaphysically on the physical and thus that there is no world where the physical correlates of consciousness are instantiated without consciousness.) I suspect that the first premise here is false -- that zombies are not conceivable, at least in the rigorous way required by the argument. (For a persuasive statement of the case for this view, see Allin Cottrell's paper, 'Sniffing the Camembert'.) But even if that's wrong, I still don't think the argument works. Like many people, I'm suspicious of the CP principle. And one way to highlight the problem is to note that physicalists can also invoke the principle to argue for their position. Here's how it goes.

Consider anti-zombies. These are beings that are physical duplicates of humans, and that have no non-physical properties, but which are nonetheless conscious. They inhabit an anti-zombie world, which is a physical duplicate of ours, but where no non-physical properties are instantiated. (Physicalists think that we are anti-zombies, of course.) Then we can run an anti-zombie argument for physicalism, as follows. Anti-zombies are conceivable and therefore, by the CP principle, metaphysically possible. And if anti-zombies are metaphysically possible, then physicalism is true. The last step may seem a big one, but it should be uncontroversial. In the anti-zombie world consciousness is physical, so the microphysical features of that world are metaphysically sufficient for consciousness, and any world with the same microphysical features will have the same distribution of phenomenal properties. But, by definition, our world has the same microphysical features as the anti-zombie world. Hence the microphysical features of our world are metaphysically sufficient for the existence of consciousness, which is to say that physicalism is true.

The argument has been anticipated by various writers -- notably Peter Marton -- but I've but tried give it a definitive statement in a recent paper (available here for those with a Blackwell Synergy subscription). As I stress in the paper, the only response available to defenders of the zombie argument is to deny the first premise, that anti-zombies are conceivable.

The point can be made independently by considering the unique world that is a physical duplicate of ours and where no further, non-physical properties are instantiated. This should be a zombie world, if any is. But it's also the only candidate for an anti-zombie world. Thus, the possibility of zombies is incompatible with that of anti-zombies. And if conceivability entails possibility, then the conceivability of zombies is incompatible with that of anti-zombies. So defenders of the zombie argument must deny that anti-zombies are conceivable.

Now of course physicalism is the view that we are anti-zombies, so if anti-zombies aren't conceivable then physicalism isn't conceivable either. In short, if you want to endorse the zombie argument, then you have to maintain that physicalism is inconceivable.

That's all from me. So long and thanks for all the fish.


Eric Schwitzgebel said...

Thanks for your delightful posts, Keith!

I'd comment, but I'm not sure "metaphysics" isn't just a branch of psychology!

Anonymous said...

Hi Keith,

I've really enjoyed your stint as a guest blogger here. I didn’t realize it was a limited-run engagement!

Regarding your argument, it’s intriguing, and as a reductive physicalist, I hope it works. However, I have some worries. I haven’t read your paper, so the following remarks are based only on your post.

Consider a version of zombism that doesn’t seem touched by your argument. I don’t believe it, and I don’t know anyone who does, but it is worth considering.

On this version of zombism, consciousness is massively multiply realized. In some worlds—Berkeleyan ghost worlds—there are no physical properties, only consciousnesses. In those worlds, my mental doppleganger is realized by networks of non-physical ectoplasmic psychons, or spook-juice, or something. In other worlds—Cartesian fractured worlds—there exist both minds and bodies, and like the ghost worlds, conscious stuff is realized by non-physical spook-juice. In other worlds—anti-zombic worlds—consciousness is realized by solely physical stuff. In other worlds—zombic worlds--there’s all the same physical stuff as in the anti-zombic worlds, but absolutely no consciousness, perhaps because of different laws that operate there, or something.

So, I guess this imagined zombist would deny your statement that “any world with the same microphysical features will have the same distribution of phenomenal properties.”

I suppose you’ll need some reason why such a move wouldn’t be available to the zombist. What is it?

Anonymous said...

Hi Pete

Thanks for the kind words. I've enjoyed blogging, and I may look for a way of continuing with it. The Splintered Mind and Brain Hammer are great inspirations.

To your question: I'm not sure I see the problem. Is it ok if I make some quick points of clarification (drawing on the developed version of the argument in my paper), and then you come back tell me if they remove your worry?

First, when I say that the anti-zombie (AZ) world is a physical duplicate of ours, I mean that it's a duplicate in all microphysical features – including both laws and distribution of properties. Call this set of microphysical features P.

Second, the crucial step of the argument goes like this: in the AZ world consciousness supervenes metaphysically on P -- in virtue of token identities, say. Thus if there is an AZ world then P is metaphysically sufficient for consciousness. But P is instantiated in our world. Hence we have a physical form of consciousness. (It's true that this doesn't rule out the possibility we have a non-physical form *as well*, but that's little comfort for the dualist.)

Third, the anti-zombie argument aims to show only that the microphysical features of our world are metaphysically sufficient for consciousness, and thus that consciousness is physical in our world. It doesn't aim to show that those features are metaphysically necessary for consciousness, and so doesn't rule out functionalist versions of physicalism, according to which consciousness is in fact microphysically realized but could have been realized in other ways. So the argument is compatible with the existence of Berkeleyan ghost worlds and Cartesian worlds (though the physical set-up of the latter would have to be different from ours, if consciousness was non-physical there.)

Does that clear things up, or are you still dubious?

Best, K

Anonymous said...

I agree that Zombies aren't conceivable in a way that confers justification on the claim that they're metaphysically possible. Van Inwagen is right: unless your conceiving is something as simple as that your chair can exist two inches to the left, it can't prima facie justify a possibility claim.

One might try to run a reductio on this sort of skepticism regarding conceivability: wouldn't that imply that we lack all sorts of humdrum modal knowledge that we actually have? No, it wouldn't. It only implies that such modal knowledge is grounded *in our conceivings*. For my part, it just seems obvious that such modal knowledge is grounded in our knowledge of how the world works. I have good track-record evidence that chairs can be moved, that cars can be painted different colors, etc. Unfortunately for the zombie argument, the modal knowledge we get in this way doesn't extend to claims remote from ordinary experience, such as knowledge of the possibility of zombies.

Anonymous said...

Thanks, Keith. That does help a bit. However, let me press the issue with what might amount to a slightly different line of objection.

I can imagine someone asserting as conjointly possible the following:

1. Whether P suffices for consciousness is relative to a world.

2. There could be two distinct worlds, both of which are identical with respect to P and with respect to the set of mental properties that are instantiated (M), but differ in virtue of what suffices to instantiate the mental properties. For instance, one world is AZ world (where P suffices for M) and the other is Cartesia (where P doesn't suffice for M).

3. We don't know whether our own world is AZ world or Cartesia, but we do know that our world is identical to both worlds with respect to both P and M.

Now, if 3 is true, I don't see how you can establish that P suffices for M in our world. And 3 (pretty much) follows from 2 which in turn follows (pretty much) from 1, so I suppose you'll want to deny 1 or something pretty close to it.

I hope I'm making more sense this time around, but perhaps I'm making less!

Anonymous said...

I think a lot of dualists -- for example, Chalmers -- would take issue with your conception of supervenience. Many people, myself included, take supervenience to be a modal concept. A-properties supervene x-ly on B-properties iff, in any two x-ly possible worlds W and W` such that both have the same B-properties, both have the same A-properties. Under this conception it does not make sense to speak of a metaphysically possible world in which mental properties metaphysically supervene on the physical because doing so is the same as asserting that mental properties supervene metaphysically on the physical. Therefore, when you claim that the anti-zombie world is conceivable your opponents are likely to argue that the conceivability of that world only indicates the conceivability (and, therefore, possibility) of a world relative to which the mental metaphysically supervenes on the physical. Such a world would be one where mental terms happen to have been attached to physical things.

Of course, this line of thought rests on my own limited understanding of supervenience. I'm just a new grad student in philosophy -- maybe you could tell me a little more about supervenience?

Anonymous said...

Pete –
Thanks; that helps. As you suspect, I want to deny (1). The notion of sufficiency invoked in the argument is metaphysical, not merely natural. And of course, to say that P metaphysically suffices for consciousness is to say that any world that instantiates P also instantiates consciousness. We get the claim that P metaphysically suffices for consciousness by conceiving of a world whose existence entails the truth of the claim, and then applying the CP principle. See my reply to James below for more details on this.

Anonymous said...

James –

You make a very good point. I share your conception of supervenience. Does it follow that it’s illegitimate to speak of a possible world where the mental supervenes metaphysically on the physical? (I’m not sure if your objection is that it simply doesn’t make sense or that it would be questioning-begging, but either way, you think I shouldn’t do it.) Perhaps that's right. But I don't need to speak that way, and if I have done so, I shouldn't have. What the argument requires us to do is to conceive of a world where consciousness is physical (and where the physical set-up is the same as in ours). And I don’t think of *physicalism* as a modal thesis; it’s the thesis (roughly) that mental properties are not distinct from physical ones. We can imagine a physicalist world by (for example) imagining one where there are token identities between the mental and the physical. But though physicalism is not a modal thesis, it does *entail* one. If physicalism holds in w, then any world with the same physical features as w (laws+properties) will have the same distribution of mental properties. So it goes like this: we conceive of a world w which is physically identical to ours but where physicalism holds; we apply the CP principle and derive the conclusion that such a world exists. If such a world exists, it follows that the physical features of that world (and hence of ours) are metaphysically sufficient for consciousness. Hey presto!

Incidentally, it is possible to run a metamodal version of the AZ argument, which by-passes the step just mentioned. It goes like this. Let P be the actual physical facts and Q the phenomenal ones. Then it is conceivable that P -> Q is necessary. If it is conceivable that P -> Q is necessary, then (by the CP thesis) it is possible that P -> Q is necessary. If it is possible that P -> Q is necessary, then P -> Q is necessary (S5 principle MLp -> Lp). If P -> Q is necessary, then consciousness is physical. Hence consciousness is physical. Peter Marton's version of the argument was along those lines. I prefer mine, for reasons I explain in my paper.

Anonymous said...


Thanks for your comment. I'm tempted to agree with you – though I think we do have some robust modal intuitions that extend beyond the naturally possible. But this is something I haven't thought enough about. For present purposes, of course, I'm accepting the CP principle in order to blow up the zombie argument from within.

Anonymous said...

Zombies only seem to be conceivable (in any sense strong enough to support
the relevant arguments) when the implications of the concept are insufficiently
explored. In fact, the concept of zombie (as it functions in anti-physicalist
or anti-functionalist arguments) contains a concealed contradiction. It follows
from assumptions required by the arguments that
zombies will inevitably claim to
be conscious, but there is no coherent way to construe the truth value of this
claim. Attempts to construe it either as true (even by the zombie's own special
lights), as false, or as neither true nor false, all inevitably lead to contradictions
and incoherencies. I spelled this out in my article "Zombie Killer" (Thomas,
1998). The arguments have never really been answered.

Presumably nothing like this argument will apply to anti-zombies because
ex hypothesis they are conscious, so any claims they make to be conscious
are straightforwardly true. Nevertheless, the anti-zombie argument looks very
question begging. After all, the metaphysical possibility of physicalism, which
the anti-zombie argument needs to assume, is precisely what anti-physicalists
(at least the zombie loving ones) deny. It is not just contingently false. They
can say: "Actually, anti-zombies are not conceivable because, as
we have proved (via zombies), physicalism is not metaphysically possible." Luckily,
however, we do not need anti-zombies to refute zombie based anti-physicalist
arguments. (The way you move from "anti-zombies are possible" to "physicalism
is true" also looks very suspect to me, but I am too tired to spot the flaw at
the moment.)

The broader moral, I think, is that philosophers need to be a lot more circumspect
about making conceivability claims (at least if they are going to adhere to the
C-P principle). Just because the inherent contradictions in a concept (such as
zombie) do not immediately strike you, does not mean that they may not be there.
After all any reductio ab absurdum proof reveals a contradiction, an
impossibility, but it normally takes quite a lot of work, thought, and even imagination,
to actually understand such a proof. The notion of a highest prime seems at first
glance to be quite conceivable, but Euclid proved, by an ingenious reductio
ab absurdum
it is impossible that there should be a highest prime. In the face of such examples
(and math and logic are full of them), it seems to me that you have only two
choices: either keep the C-P principle but admit that conceivability, in the
strong sense the principle requires, is not something that can be reliably grasped
just by intuition, or else embrace a very weak sense of conceivability (a sense
in which I can conceive of round squares, or that two plus two might equal five,
and in some sense I clearly can concieve these things, I can think of them, as
I can think of a highest prime, or a zombie) and give up the principle

Thomas, Nigel J.T. (1998) Zombie Killer. In S.R. Hameroff, A.W. Kaszniak, & A.C.
Scott (eds.) Toward a Science of Consciousness II: The Second Tucson Discussions
and Debates
(pp. 171-177). MIT Press, Cambridge, MA.

Anonymous said...

Hi Nigel

Thanks for your comments. Just a quick response. I'm sympathetic to what you say about the inconceivability of zombies and the need for circumspection in making conceivability claims. I also liked the argument in your paper very much.

On the anti-zombie argument, I don't see why you say that the argument needs to assume the metaphysical possibility of physicalism. Where exactly is the assumption made? Certainly not in the first premise, which is simply the claim that a creature with certain properties is conceivable. Nor, so far as I can see, do either of the other premises make the assumption; indeed I think both should be acceptable to zombists.

As for the zombist response you suggest ("Actually, anti-zombies are not conceivable because, as we have proved (via zombies), physicalism is not metaphysically possible"), this is question-begging, for reasons I discuss in the full version of paper, and the anti-zombist can respond by running a parallel objection to the zombie argument ("Actually, zombies are not conceivable because, as we have proved (via ant-zombies), dualism is not metaphysically possible"). The objections just trade conceivability intuitions, and the result is a stand-off.

Finally, if do you spot a flaw in the move from the possibility of anti-zombies to the conclusion, do please let me know. I'd be very interested to hear what it is!

All the best, Keith

Unknown said...

I gave an argument similar to the anti-zombie argument in a paper presented at Tucson VII - Toward a Science of Consciousness 2006. One of my conclusions was, as you say, that a zombie-file must hold that physicalism is inconceivable. I had a conversation with Chalmers about it, and he confirmed that that's precisely his view.

Anonymous said...

I like this blog is fantastic, is really good written. Congratulation. Do you want to see something more? Read it...: Costa Rica is a country with a extremely sense of freedom. The landscapes are for much the most green in whole center america.The chances of investement are way to high, the average of Americans, European and people of the entire planet who is buying here is up in the sky !!!
Great investment opportunity in Costa Rica: condos, costa rica real estate, costa rica property. Visit us for more info at: