Does it make sense to speak of creatures which behave exactly like conscious beings, but are not actually conscious? Such creatures are called Philosopher’s zombies, and they are very different from Hollywood zombies. Hollywood zombies are noticeably different from normal human beings, they drool, shuffle etc. However, when we use this definition of “Hollywood zombie” for philosophical purposes, we must remember that there are many Hollywood zombies who are not sufficiently behaviorally anomalous to land movie contracts. In this technical philosophical sense, if a zombie’s behavior deviates from the behavior of conscious beings in any way whatsoever, she (it?) is a Hollywood zombie.
I think that philosopher’s zombies could not possibly exist, because they are as self-contradictory as round squares. The following argument is one reason why. The logic of the argument is unassailable, so it will thus stand or fall depending on the truth of the three premises, which I will now defend in turn.
1) Zombies are possible if and only if subjective experiences are epiphenomenal.
Philosopher’s Zombies are possible only if there are mental states, often called qualia, which have no causal impact on the physical world whatsoever. If these qualia have any impact at all on the behavior of the creature that possesses them, then in principle another person could detect that impact, and consequently that creature would be a Hollywood zombie, not a philosopher’s zombie. There can never be such a thing as objective evidence for this kind of purely subjective consciousness, which means it would be epiphenomenal.
2) Subjective experiences are epiphenomenal if and only if we have a direct awareness of them.
Descartes claimed we have direct awareness of our mental states. It supposedly gives us absolute certainty, but granting this kind of certainty to our awareness of mental states gives second class status to our knowledge of everything else (including the mental states of other people). If we don’t have a direct awareness of our subjective states, how else could there be any reason for believing that they exist? We have already stipulated that there must be no evidence for their existence in the external world. Everyone agrees that there is no such thing as what Dan Dennett calls a zagnet i. e. an object which behaves like a magnet but lacks some kind of inner “magnetismo”. The only difference between a zagnet and a zombie is that we supposedly have a direct awareness of these epiphenomenal states which we possess and zombies lack.
3) There is no such thing as direct awareness.
See Sellars’ attack on the Myth of the Given and Quine’s on the second of the two dogmas of empiricism. These arguments are too complex to summarize briefly, but most people who have carefully studied them find them to be decisive. Check them out, if you can find the time. Also, anyone who has studied the physiology of perception knows that there is nothing direct about how we become aware of our sensations. There is lots of complicated processing going on; they don’t just jump into our awareness and say “hi , I’m the color red.”
Therefore, (by hypothetical syllogism and modus tollens) subjective experiences are not epiphenomenal, and zombies are not possible.
For my argument, I only need conditionals (if. . .then) not biconditionals (if and only if). However, I think the standard argument for zombies flips these conditionals around, (which is only possible with a biconditional), adds a modal operator, and creates an argument that looks like this.
A) If we have a direct awareness of our subjective experiences, then they could be epiphenomenal.
B) If our subjective experiences are epiphenomenal, then Zombies are possible.
C) We do have a direct awareness of our subjective experiences
Therefore, (by hypothetical syllogism and modus ponens), Zombies are possible.
The punchline of all this is that the best way to settle the zombie controversy is to refocus on the question of direct awareness. That is the only premise in this argument that both sides disagree on, and the two classic papers by Quine and Sellars, linked above, pretty much finished off the concept of direct awareness, as far as many of us are concerned. I hear it’s coming back, though, and I don’t think it’s a coincidence that many of its supporters also think that zombies are possible. Let’s redirect the zombie problem to the question of whether there is direct awareness. If direct awareness goes, zombies go with it.
Monday, July 28, 2008
Why Zombies Are Impossible (by guest blogger Teed Rockwell)
Posted by Eric Schwitzgebel at 12:25 PM
Labels: self-knowledge, stream of experience, Teed Rockwell
Subscribe to:
Post Comments (Atom)
22 comments:
Thanks for the interesting post, Teed! I'm inclined to think that zombies are conceptually possible in the weak sense that we can imagine them coherently (unlike round squares) -- but not possible in any more robust sense.
Since we don't require evidence for things we posit as conceptual possibilities, I don't think we need to accept the second premise of your argument; we could posit undetectable qualia.
Now how much we can learn about the real nature of mind from such posits and such conceptual possibilities -- that's another question!
I would say that the concept of undetectable qualia is every bit as contradictory as a round square. But I suppose that depends on what one considers an essential property of qualia. I think the burden of proof would be on someone who claims there are undetectable qualia. How could you distinguish undetectable qualia from undetectable neurological states?
Eric,
Wouldn't so-called undetectable subjective experiences be things that not even the people that have them could be aware of? Red qualia that are *mine* but that not even *I* could detect the presence of? If so, it's hard to see what would make these either subjective or experiences.
On pain of turning into full blown absurditites like Teed's round squares qualia/subjective experiences better turn out to be possible objects of the awareness of at least one mind.
BTW, Teed, I think you're right to put the focus on direct awareness as well as to raise Sellarsian and Quinean considerations against it.
For a similar sort of line, specifically focusing on Quiney stuff vs. direct reference, take a look at section 3 of this paper Josh Weisberg and I coauthored:
Type-Q Materialism
I'm concerned about the support for premise (1). Suppose that qualia have causal powers. Does it follow that zombies are impossible? You write:
If these qualia have any impact at all on the behavior of the creature that possesses them, then in principle another person could detect that impact, and consequently that creature would be a Hollywood zombie, not a philosopher’s zombie.
But I don't think that causal impact guarantees detectable impact; after all, counterfactual accounts of causation are false. There seems to me to be no incoherence in the claim that qualia have causal powers, and zombies have other, non-qualia things, who perform the same causal roles.
Eric might have a case for undetecTED qualia being possible, but not UndetectABLE qualia. Qualia that could be detected but haven't yet do seem to me to be possible, although interestingly different from the standard view of qualia. I said something about that
in this paper.
Jonathan writes:
But I don't think that causal impact guarantees detectable impact
Causal impact guarantees IN PRINCIPLE detectable impact. Some thing that exists but is in principle not detectable in any way whatsoever is surely a contradiction. Maybe Kant wouldn't think so, but I would say that his ding an sich was in principle detectable, just not detectable by us. We could after all, detect it through reasoning.
Jonathan also writes:
There seems to me to be no incoherence in the claim that qualia have causal powers, and zombies have other, non-qualia things, who perform the same causal roles.
Yes but what sort of things would those be? They would be exactly like qualia, except that they wouldn't be qualia. That seems like a contradiction to me. The only way they could be non-qualia and yet possess all the causal powers of qualia, would be to lack epiphenomenal properties that could be directly intuited. So, we're back to the issue of direct awareness again.
Some thing that exists but is in principle not detectable in any way whatsoever is surely a contradiction.
This isn't obvious to me. Do you mean this to be a weaker claim than logical positivism?
When I was in philosophy we talked about "the problem of other minds", and we didn't have all this high-falutin "zombie" business.
But this was 15 years ago, before "consciousness" became the cottage industry it is now.
Cf. Chalmers: "it is worth noting that even the possibility of zombies does not obviously entail epiphenomenalism. To see this, note that an interactionist dualist can accept the possibility of zombies, by accepting the possibility of physically identical worlds in which physical causal gaps (those filled in the actual world by mental processes) go unfilled, or are filled by something other than mental processes. The first possibility would have many unexplained physical events, but there is nothing metaphysically impossible about unexplained physical events. Also: a Russellian "panprotopsychist", who holds that consciousness is constituted by the unknown intrinsic categorical bases of microphysical dispositions, can accept the possibility of zombies by accepting the possibility of worlds in which the microphysical dispositions have a different categorical basis, or none at all."
I should have known Chalmers has already said pretty much exactly what I was saying above. Cool.
I think saying you can imagine a zombie can be likened to imagining a round square hidden inside a person.
Its easy to imagine the person and for most purposes it seems you have the mental imagery perfect - except for the round square bit.
Your support of (3) seems to me to cause problems for what you say in support of (2). As you say in support of (3), we don't really have 'direct awareness' of anything; but we clearly do have some kind of awareness of them; so you can have awareness of things of which you do not have direct awareness. So it seems we could have awareness of epiphenomenal experiences without having direct awareness; and, if so, that would answer the question "If we don’t have a direct awareness of our subjective states, how else could there be any reason for believing that they exist?" So if (3) is accepted, the support of (2) you actually give in the post is inadequate; and if (2) is not adequately supported, (3) would not appear to be relevant. What the support of (2) lacks is an argument that nothing else that we call awareness would be able to provide any sort of evidence for the existence of epiphenomenal experiences; as it stands it appears to assume that all awareness would have to be direct awareness, which, combined with (3) would require us to conclude that no one is aware of anything (in which case zombies aren't impossible, because we are all zombies!).
But I don't think that causal impact guarantees detectable impact
I think this is a problem. Say you have a causal process such that we move from a given physical state to a state + qualia which then causes a new given physical state. Then imagine via more random processes that we simply move to that state without the qualia but randomly. How do we detect the difference? The only way would be to calculate how probable something is. But that seems difficult to do, even in theory, let alone practice.
Put an other way, how would you go about verifying that this was happening by qualia rather than by some other process?
Clark, it sounds like you (and Teed) are assuming some kind of very strong connection between verification and metaphysics. Suppose you're right -- there's no way to tell whether someone's behavior is caused by qualia or some other non-qualia things (or nothing at all). So what? The claim of the post is that zombies are impossible, now merely that we'd never be able to recognize them.
I'd suggest that to think that "undetectable qualia" is like "square circle" is to conflate the epistemic and the phenomenal aspects of the term "consciousness". I don't see why these can't come apart in conception. I think they also come apart quite a bit in fact; but regardless there's a word of difference between conception and fact. I draw no metaphysical consequences.
Hi Teed, may I recommend Nagel's _Psychophysical Nexus_, a wonderful essay in which he explains the illusion of contingency between phenomenology and physiology, which underlies conceivability arguments. For more: http://philosophy.fas.nyu.edu/docs/IO/1172/nexus.pdf
Cheers,
Kevin
Eric,
You write: "I'd suggest that to think that "undetectable qualia" is like "square circle" is to conflate the epistemic and the phenomenal aspects of the term "consciousness"."
I don't see why such conflation need be entailed. Suppose someone held that one way in which one could be conscious of or aware of a quale is in this so-called non-epistemic phenomemenal sense of "conscious of" or "aware of". Suppose they also held that a quale had to be at least a potential object of consciousness or awareness: a quale in which one is in no way possibly aware of or conscious of is no quale at all.
Such a position would refuse to countenance undetectable qualia, but as long as the phenomenal "conscious of" can be viewed as a phenomenal kind of detection, then your charge of conflation doesn't stick.
Fair enough, but you'd need an argument for the conceptual entailment of "necessarily available to introspection" (or something like that) from "has phenomenal character". I don't think such an argument is to be had, except an invalid argument that depends on equivocation between epistemic and phenomenal senses of words like "conscious" and "aware".
We're likely on the threshhold of rehashing an old dispute, but I'd say the burden would be on the lover of the epi/phen distinction to supply a non-question begging argument that there is such a thing.
The burden for what purpose?
The original post advertised itself as an argument establishing that zombies are impossible.
One challenge that those of us fond of the epi/phen distinction should take up, I think, is the question of why there seem to be several terms that show this particular kind of ambiguity. It's not like bank/bank. Something deeper is going on, linguistically -- but what?
Post a Comment