Admin


Blog powered by Typepad

« Wikipedia on consciousness | Main | More online papers »

October 10, 2005

Comments

Damon Woolsey

I can't wait to read this paper, just as soon as I get a break from my studies! But just posting this paper has sparked renewed debate over conceivability claims at The Ponderer's Guild.

Ellis Seagh

Interesting -- a truly exhaustive (and exhausting!) analysis.

It seems to me, though, that there should be a materialism other than what you're referring to as Type A and Type B, and an example of that would be the assertion that the experience of qualia is itself a material phenomenon, required for the functioning of a particular kind of mechanism (i.e., in that kind of conscious mechanism, such an experience would be causal -- both caused and causing). If this were the case, then it would seem to make no more sense to say we can "conceive" of a philosophical zombie than it would to say we can conceive of a kind of zombified car in which every material part functions exactly as it does in the normal car, but in which the rotation of the drive shaft, say, is without causal effect. (Or, perhaps closer, in which the drive shaft is simply missing, but that of course would contradict the assumption that every material part functions the same.)

Peli Grietzer

Ellis- The issue is, as I understand it, that even if we posit a casually efficient "qualia", its qualitative properties do not logically supervene on its casual, functional properties.
I. E, any interactionalist model can have a zombified version, in which the "qualia" has only casual effects and no.. well... qualia.
If our mind is a product of an actoplasmic soul transmitting to the brain, there can still be a zombie world with actoplasmic souls that are phenomenologically empty.

Ellis Seagh

Paul:

But in the case of qualia, qualitative properties are all they have -- so if those qualitative properties are not themselves causally effective, the qualia are not causally effective, contradicting the assumption. Can qualtitative properties by themselves be causally effective? Yes, if we think of those properties as signals, received by a mechanism specialized to receive such signals. Then if we try to imagine an alternative version of such a mechanism that lacks only the qualitative properties of the qualia, we will have lost the information necessary for the operation of the mechanism.

djc

This sounds to me like a version of either what I call type-D dualism or type-F monism (see Consciousness and its Place in Nature). If the qualia are doing causal work that is not done by any underlying mechanism that one can specify without mentioning qualia, then it looks like either the qualia are an addition to physics that plays a fundamental causal role (type-D) or they lie at the roots of physical processes (type-F). If on the other hand one can characterize an underlying physical causal mechanism without mentioning qualia, then one will have a version of type A, type B, or type E.

Ellis Seagh

djc: I think it comes down to whether you want to think of signals or information as being able to "do causal work" in a pure mechanism, or not. I think there are advantages to such a viewpoint, and I don't see that it requires any new physics, but you may want to insist that "information" purely as qualitative difference is undetectable by a mechanism. In which case, however, I'd want to say that such a notion of "information" already has a built-in implication of at least some species of dualism, and hence begs the materialist/dualist/idealist question. (The materialist position, in the latter case, as I see it, would then be that there can be no such thing as "qualia", defined as information purified of all material or causal effect. But I'll admit I've gotten a bit lost with the alphabet labels.)

It's "conceivable" that a zombie car can have every physical (micro and/or macro) component exactly identical to a regular car, but have a drive shaft deprived of any causal effect. Would that work as an argument against causality (call it H)?

More generally, I'd just say that, of the arguments against materialism put forth in "Consciousness and its Place in Nature":
- the Explanatory Argument fails against materialist argements that consciousness, qualia and all, is itself both structural and functional;
- the Conceivability Argument fails because (as we see above) it puts too much weight on the vague notion of "conceivability" to support real-world conclusions;
- and the Knowledge Argument fails because it simply begs the question (of whether there are "truths about consciousness that are not deducible [in principle] from physical truths").

I'll add, though, that I think the last argument has more substance than the first two, because it at least hints at issues re: knowledge and experience, and the peculiar situation of a knower embedded in a materialist world.

mjgeddes

Why restrict the definition of 'causality' only to functional systems? Why not posit several different kinds of causality? One could then concede that qualia don't play any role in *physical* causality (functional systems) but they *do* play a role in the posited non-material kind of causality. For instance one could posit a 'volitional' type of causality defined to be the changes in one's belief system (memetic system) as an effect of choices one's makes i.e Mental agency is the system, 'choice' is the cause and changes to beliefs the 'effect'. This definition of causality is not necessarily reducible to the physical definition of causality i.e the evolution of functional systems

Dogz

I have a question from a physics perspective. I'll use your refined version of the "knowledge argument", since I find that most accessible:

(1) Mary is in a position to know all facts deducible from the narrowly physical facts.
(2) Mary is not in a position to know all the phenomenal facts.
(3) If a phenomenal fact is not deducible from the narrowly physical facts, it is not necessitated by the narrowly physical facts.
___________
(4) Not all phenomenal facts are necessitated by the narrowly physical facts.

I buy the argument, given the premises. But what are the "narrowly physical facts"? As far as I know, the only _narrowly_ physical facts anyone knows are facts about quantum-mechanical wavefunctions. Unfortunately, there are relatively few facts that can be deduced from those narrowly physical facts. For example, not even the location of an object or its coherence (as an object) is deducible from the wavefunction. To get those "higher-level" facts one has to add some kind of "collapse" postulate, none of which are without controversy.

So it seems that phenomenal facts are hardly special: almost none of the "facts" that we take for granted are necessitated by the narrowly physical facts. Given that, shouldn't we look for a solution to the problem of consciousness in the same place as we look for a solution to the problem of wavefunction collapse? Can that be construed as a form of Russellian Monism?

To clarify: there is no physical theory of anything other than the QM wavefuntion that does not involve introducing a conscious observer at some point (at least not a theory that is widely accepted by physicists). So Zombies, which are certainly classical beings and hence much more than the wavefunction, cannot be described by physics without introducing a conscious observer. Thus, zombies are not necessitated by the narrowly physical facts alone.

Isn't this devastating for the whole argument? Or are the philosophers waiting for the physicists to patch the gap between the QM wavefunction and the classical world without needing for a conscious observer? If so, I think they'll be waiting a long time - it has been nearly 100 years and still no such theory.

djc

See my reply to "qm skeptic" in this thread. Certainly, one option is to think that the gap between QP and CP involves consciousness itself: see e.g. the discussion of type-D dualism in the paper mentioned above, as well as the discussion of Everett in Chapter 10 of The Conscious Mind. I don't think anything here does anything to undermine the argument, though. Note that all the formal argument relies on is the conceivability of P&~Q (N.B. it doesn't mention "zombies" per se). If anything, the considerations you mention make the gap between P and Q wider than ever.

Dogz

First, thanks for replying to my questions. This is something I have grappled with for a long time, so I really appreciate it.

However, I guess I didn't express myself very well. I don't think the gap between QP and CP hurts your argument, it just makes it seem somewhat pointless. After all, if QP constitutes the narrow physical facts, and QP does not entail CP (without introducing conscious observers, which of course begs the question), then why argue about whether QP entails Q?

Or put another way, conceivability of QP & ~Q is hardly controversial. QP is just a big wavefunction, no distinguished basis, no coherent objects, just a grey soup. Nothing about QP looks remotely like our conscious experience Q, so conceiving of QP & ~Q is a no-brainer.

But when people think of CP, they more or less think of a classical world that precisely mirrors our conscious world, both in terms of classical "causes" of conscious experience (in CP, there really is a keyboard here in front of me, a lamp on my desk over there, reflecting light off the keyboard into my eyes, etc), and in terms of classical "correlates" of conscious experience (we're not there yet, but it seems uncontroversial that one day we'll have a good understanding of the neural processes associated with conscious experience). Given such a complete "explanation" of Q from CP, conceivability of CP & ~Q is naturally far more controversial (and, BTW, no doubt why anti-materialist positions such as yours are so damn difficult to argue (I find it hard enough to read the arguments, let alone generate them)).

So I am in a different place, which could be summarized like this:

1) The narrow physical facts are QP.

2) QP does not entail Q (no big deal there).

3) QP does not entail CP (huge deal there).

4) CP is intimately connected to Q (we've not found any "anti-zombies" yet - conscious beings with no physical support, and we have very good "explanations" of Q from CP. In light of 3, I think this is a much bigger deal than is usually given credit).

So it seems that the question of how Q arises from QP and how CP arises from QP are probably going to be answered at the same time (my hunch is that CP and Q are two aspects of the same thing, but I don't yet know what that thing is). But the question of how Q arises from CP (which I know you say your argument doesn't require, but it is certainly the stumbling block for most people) is probably nonsensical.

djc

Well, first, it's not uncontroversial that QP doesn't entail CP, because it's not clear what QP is. On some interpretations of quantum mechanics, e.g. a Bohm interpretation and arguably certain collapse interpretations, QP will entail CP. On other interpretations (certainly a pure Schrodinger evolution interpretation), QP will not entail CP. But for precisely that reason, there's good reason to think that this sort of QP is incomplete. What sort of thing might complete it? If it's completed by more physical structure and dynamics, then the argument I gave will apply to the completed P, and there will still be a gap between this and consciousness. So either way, it looks like the argument goes through.

The remaining option (which perhaps is what you have in mind) is that QP needs to be completed by something other than just more physical structure and dynamics. The obvious alternative is that it might be completed by consciousness itself (or protoconsciousness or something in the vicinity). I suggested something along these lines in my discussion of Everett in The Conscious Mind, where I argued that the right theory of consciousness could explain at least why we experience a classical world in an Everett universe. Maybe there are other options in this vicinity too (e.g. the different collapse interpretations discussed under type-D in "Consciousness and its Place in Nature"). But on this option we'll already have a dualism at the basic level, so once again the conclusion of the argument will be correct. And furthermore, far from rendering the argument pointless, I'd say that on this option the general irreducibility of consciousness is doing even more work on this option than on the previous options -- in a way, it's responsible for explaining the emergence (or appearance) of the classical world.

mjgeddes

Here's a comment on your paper posted by someone on another bulletin board. How would you answer this David?

"The primary argument in this paper is a version of the Lob Theorem,
expressed as Modal logic, if x is concievable, then x is possible. He
cleverly constructs the argument such that this extension proves that
materialism is false.

However, this argument contains no descriminatory power, because you
can easily construct the identical argument to prove that dualism is
false, by inverting some of the terms. It would remain just as true,
and just as pointless.

It's no more special than the teleological arguments for the existence
of god by neccesary qualities.

Again, Chalmers depends on semantic logic to prove a quality which is
a prerequisite of the argument he's making. if P&~Q is so, in a world
without phenomenological phenomena, his arguments work exactly the
same as they would in a world with phenomenlogical phenomena. They
don't actually descriminate, which is why he ends up having muck
around with the idea that ~Q can't be a positive truth, but rather a
negative possibility, to prevent the argument from being symmetrical.
He gives no reasons for this, aside from some quibbling about a
conscious being's ability to concieve positively of zombies."

djc

This is just a mistake. As I observe in the paper, in a world where Q is false, one of the premises of the argument is obviously false.

Istvan Aranyosi

Hello Dave

I was thinking about an objection to your argument raised by John Perry, namely that you have to presuppose epiphenomenalism in order to assert the premise that zombies are conceivable. The objection fails, for reasons you offered in your reply to Perry’s book. But I think there is still something that you must assume in order to assert the first premise: the causal closure of the physical, and more exactly the idea that interactionism is excluded. How else could you conceive of zombies, since if interactionism were true, the antecedent neural events would not suffice to determine behavior? But whether the physical is causally closed or not is neither a conceptual issue, nor a truth of physics. So you cannot include this principle in P.

If what I say is correct, then whether zombies are conceivable is not a matter of concepts, but a matter of the truth or falsity of causal closure (and of interactionism). And, a priori, both it and its negation are conceivable.

A similar point appears in a recent paper by Robert Kirk, “Why there couldn't be zombies.” Aristotelian Society, Supplementary Volume 73:1-16, 1999. More exactly, Kirk says at some point that the friend of zombies is not a traditional Cartesian dualist: for such a dualist, if you destroy the qualia of a subject, she will fall like a puppet.

In conclusion, it seems one has to assume the falsity of interactionist dualism in oder to assert the first, conceivability premise. (Interestingly, you share this premise of the falsity of interactionism with David Lewis (in his seminal “An argument for the identity theory”, J. Phil, 1966))

djc

Hi Istvan, actually, as I mention in the paper and in a little more depth in my reply to Perry, I think that interactionism is consistent with the conceivability of zombies. If interactionism is true and the causal closure of the physical is false, then zombies (physically identical to conscious beings in our world, not just initially but also over time) will have causal gaps in their processing -- gaps that would be filled by consciousness in our world, but aren't filled by anything in the zombie world. Causal gaps are strange, but they aren't incoherent or inconceivable.

Eray Ozkural

Marvin Minsky has made a comment about your paper on ai-philosophy list. He has included an interesting section from his forthcoming book that was removed from the draft. I tend to agree with him (as usual), and would be interested in your response.

djc

Minsky's main points seem to be that the notion of conceivability is imprecise and reliable, and that zombie analogs in other domains (e.g. car engines) are conceivable too. Re the first point, for precisification and more on reliability see "Does Conceivability Entail Possibility". Re the second, see The Conscious Mind and "Conceptual Analysis and Reductive Explanation".

I should say that the paper on the 2-D argument is a technical paper that I don't really expect to make much sense to people without a reasonably extensive background in philosophy. Non-philosophers would probably find it more productive to focus on the way these issues are cast in "Facing Up to the Problem of Consciousness", or perhaps "Consciousness and its Place in Nature" for something intermediate.

Dogz
If it's [P] completed by more physical structure and dynamics, then the argument I gave will apply to the completed P, and there will still be a gap between this and consciousness. So either way, it looks like the argument goes through.

Right. But I didn't disagree with you on this. As I said, the gap between QP and CP doesn't hurt your argument. However, I seriously doubt QP is going to be completed by more physical structure and dynamics. Bohmian mechanics is the closest as you point out, but it has its own (pretty severe IMO) problems - no relativistic formulation and hence no Bohmian field theory.

I'd say that on this option [QP completed by something other than physical structure and dynamics] the general irreducibility of consciousness is doing even more work on this option than on the previous options -- in a way, it's responsible for explaining the emergence (or appearance) of the classical world.

That seems to presuppose a certain kind of final theory of QP and Q - one in which Q explans CP. An unlikely option IMO. I suspect in the final theory Q and CP are going to be two aspects (for want of a better expression) of something more fundamental - call it D. So there'll be QP, D, CP and Q. The causality will go from D to CP and D to Q, which will explain why CP looks so much like Q.

That's why I think it is a little pointless to be so focused on arguing that Q is irreducible. People have been arguing that since the dawn of time (or at least since Descartes). The real surprise of the last 100 years is that CP is also apparently irreducible. So we should be focusing our energy on connecting the two things we know are irreducible, not continuing to argue in ever finer details that they are irreducible.

It surprises me that philosophers don't pick up on this more - nature has given us a big hint, why ignore it?

Istvan Aranyosi

Dave: Back to the idea of ineractionism as having to be presupposed for your conceivability premise to be asserted. Indeed, you are right, causal gaps are not incoherent. We can conceive of the zombie world as one in which the causal role that consciousness fulfills in the actual world is occupied by nothing, so there are causal gaps between neural states and behavior. But then I wonder whether you can assert that in that zombie world we have a duplication of the functional properties of the actual world. Actual instantiations of functional properties are, following the Lewis/Armstrong analytic functionalism, involve causal efficacy. So in the causally gappy zombie world you cannot say that those zombies are functional duplicates of us, because their neural properties are not causal role occupiers, they are probably caused by, but do not cause physical happenings. Yet the idea should be that zombies are both physical and functional duplicates of us. Is this right?

djc

Istvan: You're right that if interactionism is true, zombie won't be functional duplicates of us, strictly speaking. But functional duplication isn't required for the argument against materialism. All that argument requires is physical duplication.

Dogz: Well, the "D" option would be a version of what I was calling "protoconsciousness": something more fundamental that underlies consciousness (and maybe underlies some of physics too). As for the issue of what philosophers should be doing, my own view is that while there's a lot of value in getting clear on the foundational issues re irreducibility and the like, it would also be nice if philosophers spent more time on the sort of positive project you mention. I did a little bit along those lines in Chapter 10 of The Conscious Mind. But by and large, philosophers are a conservative lot.

Vadim Vasilyev

Dear David, but what do you mean by "physical duplication" (in your reply to Istvan)? Does such a physical duplicate duplicate physical laws as well? It seems not. Then the usual meaning of "physical" would simply evaporate. I think, another problem with your argumentation is that your idea of ideal conceivability needs clarifying. Is it legitimate at all? Are you sure that such a concept is not a self-contradictory one? If yes, how to prove that?

Robin Herbert

A quick question - if you were your behavioural twin in zombie world - would you conclude that you were not conscious?

(Apologies if this post is a duplicate - I received an error last time I tried to post)

Robin Herbert

Further to my previous comment it appears to me that P&~Q cannot possibly be conceivable according to the definition given of positive primary conceivability, because it would entail, for example, a zombie concluding that he was not conscious.

More generally the phrases "physically identical" and "differs slightly" show that P&~Q is a statement about more than one world and could not be verified by any world in particular.

Negative conceivability also seems problematic - since it consists of a positive claim that no a priori argument exists that could rule out S. How could you know that?

In any case the premise "P&~Q is conceivable" is simply assuming the conclusion that there are truths that are not physical truths.

Kappatoo

Hi Robert,

I hope it's okay if I try to respond:
For my zombie twin to be my behavioral duplicate (assuming that interactionist dualism is wrong), he would have to conclude that he is conscious. He would just be wrong.

Why does P&~Q entail anything about a zombie's conclusions? You duplicate P, thus Zombie-Kappatoo's bodily motions, including his utterances, will have to be identical. Presumably, his judgments will also be identical to mine. It is just that he doesn't have any qualia. I'm afraid I did not get this - why can P&~Q not be conceivable then?

You certainly cannot be sure that P&~Q really is positively conceivable - after all, we do have some cognitive limitations and overlook some hidden inconsistency (easy enough, as we are far from even knowing P). But you could also shift the burden of proof here - if you are an a priori physicalist, you have to assume that there is a hidden inconsistency. But is there good reason that there is and if yes, where?

Your last statement is most likely not true - after all, there are physicalists who accept that P&~Q is conceivable - they just deny that it is possible. I think that'P&~Q is conceivable' should rather be read as an intuitive judgment - similar to the intuition that Jackson's Mary learns something new.

Kappatoo

Hi Robin,

and sorry for calling you 'Robert'. Somehow your family name must have interfered.

Robin Herbert

"Hi Robert,
I hope it's okay if I try to respond:"

Hi, you are very welcome.

"For my zombie twin to be my behavioral duplicate (assuming that interactionist dualism is wrong), he would have to conclude that he is conscious. He would just be wrong.

Why does P&~Q entail anything about a zombie's conclusions? You duplicate P, thus Zombie-Kappatoo's bodily motions, including his utterances, will have to be identical. Presumably, his judgments will also be identical to mine. It is just that he doesn't have any qualia. I'm afraid I did not get this - why can P&~Q not be conceivable then?"

What I said was that it could not be positively primarily conceivable, here is the relevant definition from David's paper with my emphasis:

"David Chalmers wrote: One can likewise define a notion of positive primary conceivability, so that S is positively primarily conceivable when S can imagine a coherent situation that verifies S, where a situation verifies S when, under the hypothesis that the situation actually obtains, *the subject should conclude that S*."

So the one thing that the subject in this case could not possibly conclude is that he is not conscious. So a zombie world could never verify P&~Q.

And also, as I said, Chalmers has explicitly defined P&~Q as a statement about more than one world, so one world could never verify it in any case.

"Your last statement is most likely not true - after all, there are physicalists who accept that P&~Q is conceivable - they just deny that it is possible."

You mean in the sense that they accept that it is conceivable that they are wrong? That would seem to fall into prima facie conceivability, so it does not make my statement wrong. The argument does not work unless P&~Q is primarily ideally conceivable and this is equivalent to saying P does not contain Q.

"I think that'P&~Q is conceivable' should rather be read as an intuitive judgment - similar to the intuition that Jackson's Mary learns something new."

But as an intuitive judgement I don't think it implies possibility - it requires "ideal rational reflection".

"and sorry for calling you 'Robert'. Somehow your family name must have interfered."

No problem, that happens a lot. The name seems to form a sort of optical illusion.

Kappatoo

Hi Robin.

"What I said was that it could not be positively primarily conceivable, here is the relevant definition from David's paper with my emphasis:

"David Chalmers wrote: One can likewise define a notion of positive primary conceivability, so that S is positively primarily conceivable when S can imagine a coherent situation that verifies S, where a situation verifies S when, under the hypothesis that the situation actually obtains, *the subject should conclude that S*."

So the one thing that the subject in this case could not possibly conclude is that he is not conscious. So a zombie world could never verify P&~Q."

(The passage you quote is a bit confusing, because S seems to stand for the statement which is to be conceived and for the conceiving subject?!)
The passage does imply that if my zombie-twin were to conceive of the zombie-scenario and would then hypothesize that this scenario actually obtains, he would conclude that he, inter alia, is not conscious. In this sense, he could verify 'P&~Q’. But the same is true for me. I don’t see why this could be problematic. In fact, I do not really believe that the zombie-scenario is actual and thus I do not believe that I am not conscious. Again, the same is true for my zombie-twin: Even if he conceives ‘P&~Q’, he (falsely) does not believe that he inhabits such a world and thus he still believes that he is conscious.

"Chalmers has explicitly defined P&~Q as a statement about more than one world, so one world could never verify it in any case."

You mean because you have to evaluate more than one world to see whether the physical truths expressed by ‘P’ are really identical to those in our world? That’s true. But on the other hand, if you take ‘P’ just as a description of the physical state of a world, you don’t have to know that P also describes the physical state of our world in order to evaluate ‘P&~Q’.

"You mean in the sense that they accept that it is conceivable that they are wrong? That would seem to fall into prima facie conceivability, so it does not make my statement wrong. The argument does not work unless P&~Q is primarily ideally conceivable and this is equivalent to saying P does not contain Q."

No, I meant physicalists who concede that P&~Q is ideally conceivable but deny that it is metaphysically possible (in 2D-terms, that would probably be ‘primarily ideally conceivable but not secondarily possible’).

To illustrate, here’s one version of the 2D-argument from the paper:
“(1) P&~Q is conceivable
(2) If P&~Q is conceivable, P&~Q is 1-possible
(3) If P&~Q is 1-possible, P&~Q is 2-possible.
(4) If P&~Q is 2-possible, materialism is false.
_______________
(5) Materialism is false.“

Here, the philosophers I have in mind accept (1), (4), and maybe (3) but reject (2). They thus deny the crucial premise that there is a conceivability-possibility link, even under ideal rational reflection. In the paper, this is discussed in sections 6-9.

"[A]s an intuitive judgement I don't think [the conceivability of P&~Q] implies possibility - it requires "ideal rational reflection"."

Yes, you are right, of course. Unfortunately, we are not ideally rational and thus we cannot be sure that P&~Q is really ideally conceivable. So of course you can deny that it is. You could thus reject (1), at least for ideal conceivability (which is obviously needed for (2) to have a chance to be true). Objections of this kind are discussed in section 4 of the paper.

Robin Herbert

Hi,
_______________
"(The passage you quote is a bit confusing, because S seems to stand for the statement which is to be conceived and for the conceiving subject?!)
The passage does imply that if my zombie-twin were to conceive of the zombie-scenario and would then hypothesize that this scenario actually obtains, he would conclude that he, inter alia, is not conscious. In this sense, he could verify 'P&~Q’. "
_______________
But in what circumstances could he do that?
_______________
"But the same is true for me. I don’t see why this could be problematic"
_______________
Could you, in any circumstances, coherently conclude that you were not conscious?
_______________
"In fact, I do not really believe that the zombie-scenario is actual and thus I do not believe that I am not conscious. Again, the same is true for my zombie-twin: Even if he conceives ‘P&~Q’, he (falsely) does not believe that he inhabits such a world and thus he still believes that he is conscious. "
_______________
But are you suggesting that it is possible that you might believe the zombie-scenario is actual and conclude that you were not conscious?
_______________
" Robin wrote"Chalmers has explicitly defined P&~Q as a statement about more than one world, so one world could never verify it in any case."
You mean because you have to evaluate more than one world to see whether the physical truths expressed by ‘P’ are really identical to those in our world? That’s true. But on the other hand, if you take ‘P’ just as a description of the physical state of a world, you don’t have to know that P also describes the physical state of our world in order to evaluate ‘P&~Q’."
______________
I don't think the argument works that way, though. There is no inconsistency in having a different set of microphysical truths that do not include consciousness.
______________
"No, I meant physicalists who concede that P&~Q is ideally conceivable but deny that it is metaphysically possible (in 2D-terms, that would probably be ‘primarily ideally conceivable but not secondarily possible’).
To illustrate, here’s one version of the 2D-argument from the paper:
“(1) P&~Q is conceivable
(2) If P&~Q is conceivable, P&~Q is 1-possible
(3) If P&~Q is 1-possible, P&~Q is 2-possible.
(4) If P&~Q is 2-possible, materialism is false.
_______________
(5) Materialism is false.“
Here, the philosophers I have in mind accept (1), (4), and maybe (3) but reject (2). They thus deny the crucial premise that there is a conceivability-possibility link, even under ideal rational reflection. In the paper, this is discussed in sections 6-9."
________________
Let me put it this way. Suppose Q was "somebody has a nose".

I can conceive of a world where nobody has a nose.

Would the argument still work?

Kappatoo

Hi again.

"Could you, in any circumstances, coherently conclude that you were not conscious?"

I don't know, but why not? After all, there are philosophers who deny that consciousness exists and who thus actually believe they themselves are not conscious. But I assume you would deny that this is possible? You could, e.g., say that they are not ideally rational – then it would still be ideally inconceivable that you are not conscious. Notice that in Chalmers’ sense, even ‘I do not exist’ is conceivable (as ‘I exist’ is arguably not a priori true): To conceive of a world which verifies this, just conceive of an empty world, or whatever. I guess this would be an even clearer example to support your objection against Chalmers’ definition of positive conceivability?! However, when you conceive of a world which is not really actual, you have to suspend your empirical beliefs about the world anyway. I guess one should take it like a premise in a reductio argument: You don’t really have to accept its truth, you only have to hypothetically accept it and see what follows from it. (Apologies if this still doesn’t solve your problem. I think I understood your problem with the definition of positive conceivability now, but I still did not quite get the connection with the zombie’s judgments.)
Anyway, the 2D argument does not require conceiving of you being not conscious – it suffices to conceive of a physical duplicate of yours which is not conscious.

Re the second point: That’s a misunderstanding. I do concede that the argument requires the conceived world to be a physical duplicate of ours. All I wanted to say is that this doesn’t mean that P&~Q cannot be verified by a single world.
Suppose someone gives you a description of the physical state of some world. Call this ‘P*’. Then he asks you if P* is compatible with there being no consciousness. Then you conceive of ‘P*&~Q’ and have thus conceived of a world which verifies P*&~Q. Now assume further that P* happens to be a description of our world. It follows that you have conceived of a world which verifies P&~Q.
Admittedly, as I said, in order to know that this speaks against the truth of materialism, you have to know that P (or P*) is a description of the physical state of our world.

Robin Herbert

Hi,
___________________Kapatoo
"(Robin) Could you, in any circumstances, coherently conclude that you were not conscious?"
I don't know, but why not?
___________________Robin
Because concluding that you were not conscious would itself be a conscious act.
___________________Kapatoo
After all, there are philosophers who deny that consciousness exists and who thus actually believe they themselves are not conscious.
___________________Robin
I don't know who you are referring to but a conscious person could not coherently and truthfully make the claim "I am not conscious".
___________________Kapatoo
However, when you conceive of a world which is not really actual, you have to suspend your empirical beliefs about the world anyway.
___________________
Actually empirical beliefs are not about the world - they are about your own sensations. It is important to keep that distinction, between Empiricism and Materialism.
___________________Kapatoo
I guess one should take it like a premise in a reductio argument: You don’t really have to accept its truth, you only have to hypothetically accept it and see what follows from it.
____________________Robin
But this is not a reductio argument, it is modus ponens- a reductio argument only works if the premise turns out to be false. MP depends upon the premises being true.

So H in a reductio argument is necessarily conceivable and not possible
____________________Kapatoo
(Apologies if this still doesn’t solve your problem. I think I understood your problem with the definition of positive conceivability now, but I still did not quite get the connection with the zombie’s judgments.)
____________________Robin
The connection is quite explicitly and precisely stated by Chalmers himself in the definition I quoted above: "...the subject should conclude that S".

You tell me - who is the subject in a zombie world, if not a zombie?
____________________Kapatoo
Anyway, the 2D argument does not require conceiving of you being not conscious – it suffices to conceive of a physical duplicate of yours which is not conscious.
____________________Robin
According to Chalmer's own definition of "conceive" it depends very much upon a zombie concluding that it is not conscious. It is his own words that I am using.
____________________Kapatoo
Re the second point: That’s a misunderstanding. I do concede that the argument requires the conceived world to be a physical duplicate of ours. All I wanted to say is that this doesn’t mean that P&~Q cannot be verified by a single world.
Suppose someone gives you a description of the physical state of some world. Call this ‘P*’. Then he asks you if P* is compatible with there being no consciousness. Then you conceive of ‘P*&~Q’ and have thus conceived of a world which verifies P*&~Q. Now assume further that P* happens to be a description of our world. It follows that you have conceived of a world which verifies P&~Q.
____________________Robin
Only in the general sense of "conceive", not in the sense Chalmers defines it for the argument.

"A possible world verifies S" is part of Chalmer's definition of "conceivable". So if you say "It is conceivable that a world verifies P&~Q", you are in effect saying "There is a possible world that verifies that there is a world that verifies P&~Q".

And after all the premise does not say "A world which verifies P&~Q is conceivable", it says "P&~Q is conceivable".

Again, Chalmers has explicitly defined P and Q as statements about more than one world so the conceivability (at least the primary positive conceivability) of P&~Q is ruled out by the semantics of the argument.
____________________Kapatoo
Admittedly, as I said, in order to know that this speaks against the truth of materialism, you have to know that P (or P*) is a description of the physical state of our world.
____________________Robin
But P is defined as a description of the physical state of our world. So whatever that state was, P would be it.

But you *would* have to know that Q wasn't part of the description of the physical state of our world.

But I would be interested in your answer to my last question - would the argument work the same if Q was "somebody has a nose"?

Kappatoo

Hi Robin,

I hope it's okay if I only pick some of your points.

I don't see why one has to be conscious to conclude something.
I would certainly deny that empirical beliefs are about my sensations.
Yes, the subject in a zombie world has to be a zombie, but he doesn't have to conclude that he himself is a zombie to conceive of 'P&~Q'.
'P&~Q is conceivable' has to be true in order for the argument to work, that's right. But'P&~Q' does obviously not have to be true - in fact, the argument would have little point if it was.
And re your last point: No, the argument wouldn't work, because 'P&~N' is not conceivable - because e.g. someone who is physically identical to me has to have a nose.

Robin Herbert

Hi Kappatoo,

>>I hope it's okay if I only pick some of your points.
That's fine, but let me first address of these, whether the argument would work when Q="somebody has a nose"
>> And re your last point: No, the argument wouldn't work, because 'P&~N' is not conceivable - because e.g. someone who is physically identical to me has to have a nose.

So are you saying that someone who is physically identical to you does not have to have consciousness?

Kappatoo

Hi Robin,

no, Chalmers does - I was just trying to illustrate his point. I am not sure whether zombies are conceivable.

Robin Herbert

However you neatly illustrated my own point. Chalmers is simply assuming his conclusion.

You suggest that a person without a nose cannot be physically identical to you because you make the assumption "noses are physical".

Chalmers suggests that P&~Q is conceivable because he starts with the assumption "consciousness is not physical".

Whereas I, who have never given any thought to consciousness being anything other than physical cannot conceive of P&~Q because someone without consciousness would not be physically identical to me.

So it all depends upon the assumptions we start with.

In other words, whether or not we can conceive of ~Q is quite irrelevant - we must be able to conceive of P&~Q.

Imagine the argument being run in the early middle ages and Q="somebody has a disease". Back then they might easily have considered P did not contain this Q - that diseases were non-physical things. So P&~Q would have been conceivable then, but not now.

Robin Herbert

Hi Kappatoo,

And let me say that I find the following statement quite puzzling:

"There is no more problem with clearly and distinctly imagining a situation in which there is no consciousness than in imagining a world in which there are no angels, or in imagining a world with one particle and nothing else."

Is Chalmers really claiming that he does not see the difference between imagining the absence of angels and the imagining the inability to imagine?

I sometimes think that for all the time he has spent on this, he has really not given it that much thought.

The comments to this entry are closed.