There has been a surprising amount of talk about zombies recently around the blogosphere. Here I thought the zombie issue was settled back in the ’90’s; but I suppose that’s what I get for forgetting that there aren’t any solved problems in philosophy, and it is in the nature of zombies to come back from the dead so I suppose I shouldn’t be all that surprised.
At any rate this zombie flare-up was brought about by Richard Chappell’s argument against physicalism. It starts by claiming that there is a possible world where there are creatures exactly like us in every physical way but who lack conscious experience. They scream when stabbed and ohh and ah over Matisse but ‘all is dark inside’; they are merely going through the motions since they do not have any conscious mental experiences. Since this world is conceivable it is metaphysically possible and since this world is metaphysically possible that means that physicalism about consciousness is false. This is, of course, not new with Chappell; this is the argument developed by David Chalmers. Chappell then presses JAckson and Chalmers’ argument that the popular Kripkean response fails to meet the zombie challenge. He argues that the zombie world is conceivable and whatever we decide about the way we want to use our words (i.e. whether we decide to apply our word ‘consciousness’ to the zombies), the metaphysical possibility of that world is enough to refute physicalism.
I have been arguing that Chappell’s argument fails to address the most plausible physicalist response to the zombie argument (and that his defense of modal rationalism itself adopts a version of the Kripkean strategy). This is to deny that the zombie world is actually conceivable. Sure, it seems to Chappell that he is imagining a world where there are physical duplicates of me (or you) and no consciousness but he is really imagining a world that LOOKS a lot like there are physical duplicates of me (or you) which lack consciousness. This is what I have been calling the Kripkean response because it is exactly the strategy that Kripke adopts in Naming and Necessity. It seemed to people that they were imagining a world where water wasn’t H2O (or where Aristotle wasn’t Aristotle) but they are really imagining a world where there is stuff that LOOKS like water does to us which isn’t H2O (or a world where there is a person who LOOKS like Aristotle (or satisfies most or all of the descriptions that Aristotle satisfies in the actual world) who isn’t Aristotle). Why should we think this is really what is going on? There are many reasons:
1. Do Conscious Mental States Cause Behavior?
If the answer to this question is yes then it is obvious that zombies are inconceivable. The zombie world is a world physically just like ours but which lacks conscious experience. Conscious experience act as causes of behavior, so a world that lack them would not be one where things went on as usual. It would be a world that was missing a bunch of causes and so, if just like ours, would have creatures that looked like us but would not act just like us. This is basically the argument that John Perry developed in his “Knowledge, Possibility, and Consciousness”.
2. Do We Have Epistemic Access to Our Conscious Mental Experience?
If the answer to this question is yes then the zombie world is inconceivable. This is the argument developed by Robert Kirk in his book “Zombies and Consciousness” (and the one being talked about over at Overcoming Bias). Here is a quick summary of the argument (from my review of the book for phil. psych.)
The anti-zombie argument can be stated rather easily. According to the ‘zombist’ there can be a creature that is a molecule-for-molecule-duplicate of me and yet lacks phenomenal consciousness. At the same time they want to hold that we have ‘epistemic access’ to our phenomenal consciousness. These two claims are not consistent with each other. To see why, imagine a zombie world that is identical to ours except in respect of phenomenal consciousness. Since that world is just like ours we can assume that it is causally closed under the physical. Now, continues Kirk, it should be possible to add to that world whatever it is that the zombist thinks will transform it into a world that does have phenomenal consciousness. But since whatever we added would have to be nonphysical, since their world is identical to ours (excepting consciousness), and so could not interact causally with the physical world (which is closed under the physical), it follows that we could not know anything about these ‘e-qualia’. Therefore, we could not have ‘epistemic access’ to them.
To make this vivid he offers what he calls the ‘sole-pictures’ argument. Again, consider our zombie world. Let’s add whatever it is that the zombist thinks will transform it into a world like ours. Now let’s imagine that by a “strange shift in the natural laws” of the zombie world the visual processes that in me cause e-qualia instead cause
sequences of constantly changing pictures to appear on the soles of [the zombie twin’s] feet. The changing colored patterns on his soles are isomorphic to those neural process in the same way as my e-qualia are isomorphic…to similar process in my brain. (p. 45)
Is there any reason to think that the zombies will have any access to these sole-pictures? Kirk’s answer is ‘NO!’ If not then zombies are not conceivable. The zombist commits what he calls the ‘jacket fallacy’: They treat qualia as something that can be stripped off a world without changing anything in the way that I can remove a jacket and remain the same. In the second half of the book Kirk tries to construct a theory of perceptual-phenomenal consciousness that does not commit the jacket fallacy.
This argument is convincing to me, and others. Has anyone responded to it?
3. Do qualitative properties (partially) Constitute our Phenomenal Beliefs?
If the answer to this question is yes then zombies are inconceivable. This is the option that Chappell and Chalmers opt for. They argue that qualitative properties are epiphenomenal in that they do not causally interact with the physical states of my brain but since the partly constitute the phenomenal beliefs I have (or in other words, my beliefs about what phenomenal states I am in). My zombie twin will believe that he sees red when I do but whereas I have a belief whose content is partly constituted by the epiphenomenal non-physical property ‘redness’ (or whatever), my zombie twin’s phenomenal beliefs lacks this property entirely and so we have very different beliefs even though we are physically identical. This seems conceivable and so it may seem that this response avoids the kind of physicalist response I am suggesting. But this is too quick.
The reason it is too quick is because of the discovery of pain asymbolia. Pain asymbolia is a rare condition where people report that they are in pain, and can even tell you what kind of pain it is (burning, pinching, stabbing, etc) yet they do not feel it as painful. The distinctive painfulness of the pain is absent. Presumably then these patients have an our world equivalent of the kind of beliefs that Chappell and Chalmers take the zombies to have. But far from acting in the same ways that we do when we have the same beliefs + qualitative painfulness these pain asymbolics laugh and smile at the pains. So the world that Chappell is imagining is not physically just like ours. If it were then the zombies would not behave in all the same ways that we do.
All of these discoveries come from a posteriori investigation. Even on Chappell’s modal rationalism these all turn out to be a posteriori discoveries because they depend on which world is actual. A purely rational being would not a priori whether the answers to 1-3 were ‘yes’ or ‘no’ until looking at the world. This is the Kripkean strategy. Chappell hasn’t argued against it and it doesn’t fall prey to the arguments he does develop.
23 thoughts on “The Inconcievability of Zombies”
Quick clarification: you wrote, “What is new with Chappell is his argument that the popular Kripkean response fails to meet the zombie challenge.”
Alas, I cannot take credit for this argument — it is textbook 2-dimensionalism, also due to Chalmers (along with Frank Jackson, as quoted in my ‘Kripke’ post).
Thanks for giving a short explanation of Krik’s argument, after saying in the other post that my argument was similar to it. In fact, I’m not arguing against inconceivability of the zombie-world. Merely that if we remove “philosophers are aware that they are conscious” as an abstract-level explanation of the behavior of the epiphenomenalist-zombies, there seems no other good explanation for the epiphenomenalist-zombie behavior. I tried to follow few possible explanations in my post here, but while I follow the idea, I just keep stumbling into weird consequences.
Sorry about that Richard! I should have read the post more carefully.
So you think that zombies are concievable? It seems to follow from your argument that they aren’t…
On the substantive issue, I think you’re overemphasizing the importance of “LOOKS”. You did better in your parenthetical remark: “(or satisfies most or all of the descriptions that Aristotle satisfies in the actual world)“. The real lesson from Kripke et al. is that often a term ‘T’ rigidly designates whatever Y actually fills the T-role. We can imagine a world where something different fills the T-role, but that doesn’t mean that T might not have been Y. It just means that the T-role might not have been filled by Y. Replace the rigid designator (‘T’) with its reference-fixing description (‘the T role’, or ‘the T-ish stuff’), and we are now — according to Kripke et al. — perfectly able to determine what’s possible in these terms. There *really is* a world containing watery stuff but no H2O. We’re absolutely right about this. Identity talk is risky, but descriptive accounts of imagined possible worlds are perfectly safe (insofar as they’re a priori coherent), and there’s nothing in Kripke to suggest otherwise.
Consider now the zombie world. We think we are imagining a world containing all our physical stuff but no consciousness. But you worry that ‘consciousness’ may be a rigid designator, like ‘water’, so we are misdescribing the world we’re conceiving of. Okay, that’s easily solved: replace ‘phenomenal consciousness’ with ‘the phenomenal-ish stuff’. Now we can safely say that we are imagining a world containing our physical stuff but no phenomenal-ish stuff. It’s a world physically identical to ours, but different in respect to the fact that there is nothing filling the p-consciousness role. Obviously this suffices to refute physicalism, whether the missing stuff gets to be described as ‘consciousness’ or not.
That’s the positive argument. Now to defend against your three objections:
(1) I answer ‘no’, as you know. But in fact even a positive answer does not entail that zombies are inconceivable — as Perry himself eventually granted in light of Chalmers’ counter-argument.
(2) From your description, it sounds like Kirk is begging the question by assuming a causal theory of knowledge. You assert that because qualia “could not interact causally with the physical world… it follows that we could not know anything about these ‘e-qualia’.” But that doesn’t follow at all, according to many epistemological theories (e.g. Knowledge as Sufficiently Safe Belief). Incidentally, Chalmers offers a brief response to Kirk here.
(3) “Presumably then these patients have an our world equivalent of the kind of beliefs that Chappell and Chalmers take the zombies to have.”
Nope. (Pain asymboliacs still have fully-fleged beliefs involving phenomenal properties, they’re just different phenomenal properties from those that feel unpleasantly painful to us. They still feel like something.) Anyway, pain asymboliacs are not physically identical to us, so I don’t see what your objection here is meant to be.
Well, conceivable given the picture that the epiphenomenalist starts with, I guess.
I mean, wouldn’t most agree that we can’t comprehend/understand how the movement of particles gives rise to specific conscious experiences (and exactly those, and not some others)?
I think that anyone who thinks that, is implicitly agreeing that they can’t see that it is necessary that given any kind of events in the brain, some kind of experience would arise. But, if we can’t see any (metaphysical or logical) necessity there, I don’t think we can argue that it isn’t conceivable.
But, it seems to me that the proposal of epiphenomenalism ends up with consequences which are hard to buy. That is, the issues you mention in the post, the ones I pointed to, and the ones which were mentioned at Overcoming Bias. will commit epiphenomenalist to take more and more weird position (though not metaphysically or logically contradictory)
Oops, “But, if we can’t see any (metaphysical or logical) necessity there, I don’t think we can argue that it isn’t conceivable.” should be: But, if we can’t see any (metaphysical or logical) necessity there, I don’t think we can argue that zombies aren’t conceivable.
Hmmm… You can’t kill zombies, they’re already dead. But, whatever:
On your response to 3 using pain asymbolia, I take it Chalmers and Chappell will deny that “these patients have an our world equivalent of the kind of beliefs that Chappell and Chalmers take the zombies to have. ” On what grounds can they do so? And who is guilty of question-begging first?
They might say that pain asymbolics fail to apply the ordinary functional-role notion of pain–they do not believe that these states cause screaming, crying, etc. That is, they don’t believe they are in pain, despite what they say. If they did, they’d show all the ordinary behavioral effects of pain belief, and act just like us (or at least that’s conceivable–is that all they need here?). So they are not our-world zombies.
In any event, even if my gloss isn’t right, I take it that’s the move in general. Something is messed up in pain asymbolics beyond the lack of phenomenal pain. BUt I could be missing the subtelites of zombophilia and -phobia. Let me know.
PS Aren’t you an ethicist? What’s with all this mind stuff? 😉
“Now we can safely say that we are imagining a world containing our physical stuff but no phenomenal-ish stuff.”
Your safely being able to say that is what we are disputing. You cannot simply assert that is conceivable because it seems to you that it is. I have been arguing that a case can be made that it is the world that is physically just like our without phenominal-ish stuff that is inconceivable. You seem to keep missing that point I never meant to suggest that any of these responses were knock-down. Only that your move above is suspect. We need an argument that the world you say you are conceiving is in fact not being misdescribed. You write as though it is obvious that the semantically neutral version of the zombie world is conceivable, but it isn’t. That debate is wide open, and if it is then your argument just begs the question. Though, as I have said, if there are people who say that the semantically neutral world you describe is really conceivable (not just seemingly-so) then your argument works just fine. But it does nothing to the kind of physicalist I am defending. You need to fight that kind of physicalist with another argument; one that establishes the conceivability (not just seemingly conceivablility) of the semantically neutral zombie world. I haven’t seen any such argument.
“Nope. (Pain asymboliacs still have fully-fleged beliefs involving phenomenal properties, they’re just different phenomenal properties from those that feel unpleasantly painful to us. They still feel like something.) Anyway, pain asymboliacs are not physically identical to us, so I don’t see what your objection here is meant to be.”
I guess I was a bit sloppy making the point I had in mind, so it is no wonder that you don’t see what the objection is supposed to be. Let me try again. The pain asymbolic is someone who says that they are in pain, they feel the pain and recognize what kind it is and say that it hurts, but they are not in pain. They do not find it unpleasent or want it to stop or anysuch thing. So, granted that there is something that it is like for these people, they do not have the pain quale (whatever that turns out to be). They are not physically identical to us, and as we would expect from that they do not behave in the same way as a ‘normal’ person does. I take this as evidence that the pain quale matters. If it were changed it would alter the behavior, shouldn’t we expect it to be even more obvious of a change were it to be absent? When you are trying to imagine a physical duplicate of yourself that believes it in pain and yet lack the painful quale (when you don’t) you are trying to imagine something impossible; we have evidence for this from the fact that when we find people with beliefs that they are in pain and who lack the pain quale they behave differently than we do. I know that though they lack the pain quale they have some other quale in place of it, but the important point is that pain beliefs without the painfulness don’t do anything in a world like ours, so the world you are trying to imagine can’t be physically just like ours.
I’m excited to head out to Tucson tomorrow…what day are you getting in? I did a little research and found a cool blues jam on Friday night…could be funky…
But anyways, I’m a META ethicist…which I argue is properly seen as a project in cognitive science (naturalized metaethics).
Josh, the problem with your suggestion is that these pain asymbolics recognize that the pains should lead to screaming. The get some information that something really horrible is comming and then, when it isn’t there, they think it’s amusing that it should be labeled as so terrible. They are trying to apply the normal functional concept, but find that it what they feel doesn’t pack the promised punch.
By the way, Tanasije,
Ah, so you are one of the people that Richard C. is targeting.
“You write as though it is obvious that the semantically neutral version of the zombie world is conceivable, but it isn’t. That debate is wide open, and if it is then your argument just begs the question.”
My point is that there’s no reason to doubt that we’re at risk of “misdescribing” worlds even in semantically neutral terms. There’s nothing in Kripke or Putnam to suggest such a thing. On the contrary, they all presuppose that we do in fact have a solid grasp of, e.g., twin earth, in descriptive terms. So I don’t think the debate is as ‘wide open’ as you think it is — your position is actually an extraordinarily radical form of modal skepticism.
“When you are trying to imagine a physical duplicate of yourself that believes it in pain and yet lack the painful quale (when you don’t) you are trying to imagine something impossible; we have evidence for this from the fact that when we find people with beliefs that they are in pain and who lack the pain quale they behave differently than we do.”
That’s fallacious reasoning. There are two ways to remove pain qualia: (1) remove the psycho-physical bridging laws that give rise to consciousness, or (2) change the underlying brain state that the bridging laws act upon to give rise to qualia.
Given the actual psycho-physical bridging laws, the absence of pain qualia implies a difference of the second kind, i.e. in cognition / brain state. (For one thing, these people lack the belief that their experience is unpleasant.) That’s entirely consistent with my view, and no reason at all to think that a difference of the first kind (leaving brain + behaviour unchanged) is impossible.
Heh, oops, that should have been ‘no reason to believe that we’re at risk…’
Richard, you said: The pain asymbolic is someone who says that they are in pain, they feel the pain and recognize what kind it is and say that it hurts, but they are not in pain. They do not find it unpleasent or want it to stop or anysuch thing. So, granted that there is something that it is like for these people, they do not have the pain quale (whatever that turns out to be). They are not physically identical to us, and as we would expect from that they do not behave in the same way as a ‘normal’ person does. I take this as evidence that the pain quale matters.
I guess I’m not sure now why you think they lack the pain quale, given that you say they feel pain. This seems better evidence that quale play no functional role. The pain asymbolic lacks the connections between pain qualia and normal function. Block tends to go for this sort of case.
Also, given that Chalmers denies that zombies are nomologically possible, he’s ok with a behavioral difference here, in the same manner as Richard C. says. There’s a nomlogical connection between qualia and functional states on this view, but it is not metaphysically necessary.
But my question about the pain asymbolic is why think they are correct when they say they are in pain, in this world? Or that they know what concept they are applying at all?
Meta-ethics! My bad! (Or is it, my meta-bad?)
To me, the whole discussion about philosophical zombies is being “conceivable” or not is pointless. What is conceivable simply depends on the mind of the person who is ding the conception. All a priori “knowledge” is in fact nothing but the outcome of the logic game we are playing. It says nothing about the state of reality. In that way, it’s true that “Philosophy Sucks”. I don’t mind people playing the philosophy game, but I don’t like it if they think that that game defines reality. So call me a fan of Nietzsche. 🙂
It is up to those who say that philosophical zombies can exist, to produce evidence that they really do or did exist in reality. Perhaps some human being has these qualities? Or, perhaps we are all are “philosophical zombies”? The burden of evidence of something is on those who claim that that thing exists.
“Heh, oops, that should have been ‘no reason to believe that we’re at risk…’”
I liked it better the first way 😉
“your position is actually an extraordinarily radical form of modal skepticism.”
Being from California, I think ‘Radical Skepticism’ is an emotive expression of approval I can almost hear Spicoli (sp?) say “all I need is some cool waves and some tasty babes, radical skepticism!” 🙂 Actually, sounds like a nice band name.
But seriously, Richard, I don’t know why you insist that it is so radical. It just follows from the claim that what is actual determines what is possible. Is it possible that we could be presented with ‘watery stuff’ that isn’t H2O, sure seems reasonable. Is it possible that we could be presented with ‘painish stuff’ that wasn’t a brain state? Sure, that seems reasonable. Does that refute physicalism? Of course not, all that means is that we can seem to be in pain when we are not (just as we can seem to be drinking H2O when we are not), and this is something for which there is evidence.
Can there be a physical duplicate of me that lacks ‘painish stuff’? I say no (or at least, ‘not obviously yes’). That brings us to your last response;
“That’s fallacious reasoning. There are two ways to remove pain qualia: (1) remove the psycho-physical bridging laws that give rise to consciousness, or (2) change the underlying brain state that the bridging laws act upon to give rise to qualia. “
I suppose I was not sufficiently sensitive to the implausibility of your views. It is not obvious that your (1) above is even coherent. Suppose I take a steam driven train. The whistle’s blowing on that train is epiphenomenal. Now suppose that we ‘change the bridge laws that give rise to whistle blowing’. Is it reasonable to expect that the train will go on functioning as normal? No! We should expect that engine would NOT function just as it did before. So, say we take the actual world to be one where these bridging laws hold, and let’s say that God decides he has had enough qualia for now and he (secretly) decides to abolish that bridging law (say he did it last night). Should we expect the world to go on as normal today? Why should we expect this? because you think it seems reasonable? That just seems brazenly rash to me; akin to Kant pronouncing that Aristotle’s logic was complete and that Newtonian physics was right. We don’t have all the facts about the physical world and so it is useless to try to figure out what the world would be like ‘if one law were changed’. The fact that you, a limitedly rational being, does not see (or notice) anything contradictory does not entail that there is nothing contradictory about the claim that you are making. A priori can tell us when certain of our assumptions lead to contradiction, but not that they don’t unless we have all the physical facts and laws set , which we don’t. Science is the only way of knowing about the physical world (i.e. setting the facts that we then feed into a priori reasoning). This claim does not fall to your counter-arguments (for instance, this is an inductive claim that has been thoroughly tested for thousands of years and so is a scientific claim itself. The failure of anyone anywhere to produce a plausible theory of a priori knowledge that is not just plain mysterious (while we have an alternative theory of a posteriori knowledge that is not mysterious) is evidence against modal rationalism and for what you call scientism).
Beoran, I mostly agree with you, but I would quibble about the claim ‘that what is conceivable simply depends on the mind of the person who is doing the conception’. I agree that some things will seem conceivable to some while not to others. But what I disagree with is the claim that these things are REALLY conceivable by an ideally rational being who had access to all the facts. For that being there would be only one set of conceivable things all and only those which don’t entail contradictions. The argument I am making is that we don’t have all the facts and we are not ideally rational and so it is a bit far fetched to think that WE, as we stand, have access to the things which are REALLY conceivable. More like, it seems to me, is that we may be missing something.
Josh, I’ll have to get back to you.
Like in being target group for the epiphenomenalist-marketing campaign? 🙂
Maybe. After all, I will buy good metaphysical a priori over slight weirdness or empiricism every time. But I think this goes beyond slight weirdness.
“It is not obvious that your (1) above is even coherent.”
I thought you were trying to offer a further argument against epiphenomenalism here, rather than simply reiterating your modal skepticism. I hope you can at least see that from the epiphenomenalist’s perspective, pain asymbolia is no objection to their view at all. (Though the steam train analogy makes me wonder whether you’ve misunderstood the view. Who on earth thinks there are “bridge laws that give rise to whistle blowing”? Whistles are part of the causal fabric of the world, not ‘epiphenomena’ in the property dualist’s sense.)
Yes, I understand that the analogy is imperfect, but the point is rather that we have no reason to be so confident that we have a handle on what changing one law in the physical world would really do. I am between classes right now, but I will try to think of a better analogy to make the point.
Your defense of the conceivability of Zombie World by analogy to Twin Earth feels unconvincing. On Twin Earth, while there is no water, there IS something (XYZ) filling the water-role. In Zombie World, the phenomenological-stuff role is filled by nothing. A better analogy would be to an Earth where there is no water, yet people still drink out of empty glasses, build elaborate plumbing systems to move nothing, where the sun may be hidden from view with no physical cause, jellyfish drift through inexplicable vacua in deep valleys, etc etc.
This isn’t quite a priori impossible in the sense of 2 + 2 = 5, but it’s considerably less than intuitively obvious that the scenario as described can be conceived. To anyone who follows the abundant intuitive evidence that states of consciousness can cause physical effects (like R. Chappell making blog posts–I know you’d deny this, but the denial seems rather logically strained) it certainly SEEMS inconceivable, and such a person can easily point out numerous doubts that could be cast on the possibility of the scenario, even if these doubts cannot be made fully rigorous at this time.
So even if it hasn’t been demonstrated that Zombie World is inconceivable, I think you’re just a tad early in claiming that it successfully ‘defeats’ either physicalism or substance dualism.
Actually, having thought about it, I am not sure why it matters that the steam is physical. It is caused by the working of the engine, and we can imagine changing the laws that result in the steam arising. All of the relevant stuff is the same.
>your position is actually an extraordinarily radical form of modal skepticism.
When defending a fairly radical view it’s a bit odd to accuse another person of having a radical view as being a counter to that view. Particularly when I would presume modal skepticism is actually fairly common – common enough to be an almost knee-jerk response to some theories despite that one might wonder why they would hang around debating that sort of topic with people.
I saw Chalmer’s brief response to Kirk and unlike much of the rest it seemed rather weak. I think I can formulate a stronger attack than his interpretation of Kirk (probably Chalmers not fully representing him I guess) and yet Chalmers seems to respond by proposing an impossibility in the appropriate world view. I.e. what sense can one make of a disembodied ‘being’ that can’t detect any part of itself?
[…] really conceiving what you think that you are. Since we do not know if materialism is true or not we do not know if we are really conceiving the zombie world without contradiction or not. And that is the point. Without knowing whether or not materialism is true we cannot know if the […]
[…] was an active participant. Oh, my bad, I guess I really did miss the point of all those posts, like The Inconceivability of Zombies, where I was arguing that the first premise of the zombie argument was false; I guess RC […]
[…] we can *almost* conceive of zombies. This is a point that I have made going back to some of my earliest writings about zombies. If the identity theory is true, or if some kind of biological view about consciousness is true, […]