Home | Philosophy

Fading and Dancing Qualia - Moving and Shaking Arguments.
By: Brian Crabb

In a by now well-known paper entitled Absent Qualia, Fading Qualia, Dancing Qualia [1] David Chalmers claims to present a persuasive case for what he calls ‘Organizational Invariance’ - the thesis that a being’s functional profile, suitably characterised, is sufficient to fix its state of conscious experience. This fixed relation between functional profile and conscious state is, he intuits, binding at the natural or nomological level. So while we can happily continue to cherish the intuition that functional and experiential states are logically, conceptually, even metaphysically independent, we should be persuaded that in the actual world, at least, it seems very likely that the two are inextricably correlated. In support of this claim, Chalmers develops two thought experiments, which I shall now examine in turn.

1. The Fading Qualia scenario.
This argument takes the form of a reductio ad absurdum, although with an explicit disclaimer attached. It is intended, says Chalmers, as a persuasive rather than logically water-tight demonstration of his thesis, staking no claim to strict, logically sound status. This is fine in itself; after all, we are all appealing to intuitions and degrees of subjective plausibility in this area of enquiry, and it would be unreasonable to demand anything more rigorous of Chalmers. What I hope to show, however, is that even as a device of persuasion his argument rests on intuitions that we simply do not have.

The first premise of the argument concerns the possibility of a zombic ‘functional isomorph’ of ourselves; a being that resembles us in all relevant functional respects, but which nevertheless lacks qualia, or even consciousness per se. In line with Chalmers’ own terminology, and since the supposed creature might be constructed entirely from silicon-based materials, I refer to it as ‘Robot’. We find that we can accept the possibility of Robot as it stands, perhaps, and so Chalmers wishes to explore some likely consequences of what we can accept in order to establish that we should not. His argument thus consists of a reductio ad absurdum on this initial assumption; the possibility of Robot. In what follows, then, the functionally and experientially normal human Joe is to be the subject of Chalmers’ thought experiment. For convenience, I will volunteer as Chalmers’ adversary here, sometimes referring to myself when I wish to refer to defenders of the Robot possibility.

The first part of Chalmers’ reductio involves the preservation of Joe’s original functional organisation, and hence his reports about his own qualia (Chalmers also wants to infer the preservation of Joe’s original beliefs, but as we shall be arguing, there might be a problem here). We can imagine replacing a small section of Joe’s brain with a silicon implant, which has been so-configured as to leave his functional profile intact. Hence, irrespective of what we might want to say about his qualia, he continues to report that his qualia are exactly the same as before. Next, another, then a larger piece of hardware, is introduced, and finally the whole brain, or the entire person, is composed of silicon. And throughout the process it seems highly plausible to Chalmers that such a transformation, with the original functional organisation intact, should be possible. In Chalmers’ words:

“As long as the chip [or any subsequent and larger silicon chunk] has the right input/output function, the replacement will make no difference to the functional organisation of the system” (p. 314)

In short, then, what Chalmers has explicitly described is the steady transition from a normal human being to a silicon-based functional simulacrum. His strong intuition here is that functional profile at the behavioural level can be preserved. Thus:

“We can imagine that throughout, the internal system is connected to a body, is sensitive to bodily inputs, and produces motor movements in an appropriate way, via transducers and effectors. Each system in the sequence will be functionally isomorphic to me at a fine enough grain to share my behavioural dispositions. But while the system at one end of the spectrum is me, the system at the other end is essentially a copy of Robot.” (p. 314)

And if functional profile at the behavioural level can be preserved, it follows that the resultant robot or semi-robot will be making the usual reports about its experiences. It will not even report (sincerely) that its qualia are fading out.

The second part of the reductio looks at the implications of our assumption that a functional simulacrum of ourselves might not be conscious. If Robot were possible, it should also be possible to transform a human being gradually into a mindless robot by progressively replacing its brain matter with a silicon substitute. Why? Well, if mindless robots that function just like us are possible, then at the very least Joe, who is the recipient of the silicon substitutions, must surely experience a transformation in his qualia. Suppose, to take the extreme case, that silicon-based organisms are just incapable of sustaining consciousness, and that this is what justifies our belief in Robot. It then seems quite plausible to infer that as Joe’s brain is gradually replaced with a silicon substitute his consciousness will begin to degrade. Nor do I have any problem accepting Chalmers’ verdict that for such a being-in-transformation qualia must fade out gradually, if at all, rather than suddenly blink out on a whim. Since, ex hypothesi, Joe is heading for oblivion, there will surely be a point where he has taken on so much silicon that his experiences finally cease, but the erosion through that point is still likely to be gradual.

The unacceptable consequence of all this, for Chalmers, is now upon us. A functionally ‘human’ being, partially-composed of silicon, will continue to report its qualia as being as bright and vivid as ever, even though they are fading out. And this, Chalmers urges, is just wildly implausible. How could a being which is still half-human, still conscious, be so sadly out-of-touch with its own immediate experiences? Or, to turn the question into an assertion, he might say that of all the possible silicon adaptations which (we might all agree) would in fact preserve Joe’s functional state, there are none which could be accompanied by fading qualia. And we know this because it would be utterly implausible for a conscious being to be so far out of touch with its own changing experiences. Intuitively, there just couldn’t be a case in which Joe continued to report that his qualia remained intact even though they were seriously fading, because it is very unlikely that such experiential reports could become so mistaken.

This, then, is how Chalmers’ reductio is supposed to work. In the first place, we are asked to accept that:

1. For as long as Joe remains conscious he will remain disposed to respond as if to the original, vivid qualia. He will not respond as if to a fading of his qualia. In particular, he will not even report that his qualia are fading.

This, coupled with the intuition that:

2. Joe would be very unlikely to respond as in 1. unless his qualia were in fact constant.

leads to the desired conclusion that:

3. It is very likely that Joe’s qualia will remain constant.

Now we can see immediately that premise 2 requires support. Obviously, the fidelity of one’s experiential reports can be expected only if they are sincere; they must accurately reflect Joe’s beliefs. So premise one must be refined to provide that support. What Chalmers clearly needs to add to 1. is that:

1’. Joe’s experiential reports are accurate expressions of his experiential beliefs.

Immediately, the argument works. Firstly, it is surely obvious that with the appropriate technological jiggling the silicon implants can be contrived to preserve Joe’s original reports. Next, those reports are true indicators of his experiential beliefs. But then, since it is extremely unlikely that Joe would continue to believe that his experiences remain unchanged unless they did, it follows that they do. The fading qualia that the assumption of Robot suggested are just not going to happen, and the Principle of Organizational Invariance prevails.

I think the argument breaks down where Chalmers infers from this impossibility that the original robot (permanently devoid of qualia) could not exist either. For if there are special difficulties in coming up with a conscious Joe who can be systematically wrong about his own experiences, those special difficulties exist precisely because Joe is conscious; to the extent that, for whatever reason, a conscious creature must be restricted to true beliefs about its own qualia. The full-blown zombie, or Robot, does not suffer from those difficulties, simply because it is not conscious, and therefore has no qualia. Turning this point around, then, the possibility of Robot does not, as Chalmers’ reductio requires, in itself entail the possibility of a conscious Joe who becomes seriously wrong about his own fading qualia. So the question I want to ask is why Chalmers thinks I am committed to such an errant Joe.

Preserving Joe’s original functionality, for Chalmers, involves merely securing the correct input and output values with the silicon implant in place, and he seems to assume that as a simple matter of common sense this must always be possible. Indeed, for the purpose of his argument, it must always be possible. For only in that case, and on the further plausible assumption that Joe’s reports will remain accurate, could he infer that the posited fading qualia could not occur. But if, as he suggests, it is extremely unlikely that Joe’s qualia reports will become seriously inaccurate, then the onus is on Chalmers to explain how Joe’s functional organisation would be preserved in the face of his fading qualia. For unless it can be preserved in these circumstances he has no argument. Remember: the reductio takes as its initial assumption that Robot is a possibility. That is supposed to be my mistake. It is then supposed to follow that (a) Joe’s qualia will fade, while (b) his qualia beliefs and reports can nevertheless be held constant. Each of these is supposed to be an independently plausible premise. (a) is supposed to follow from the absent-qualia assumption, and (b) is supposed to be plausible per se. Thus, the absent-qualia assumption, combined with common sense, must be shown to lead to an absurd set of consequences. But unless (b) really does stand as a matter of common sense, so that for me there must be a case where the two obtain at once, Chalmers argument simply falls apart.

The problem is that Chalmers has not explained why I should be committed to (b) irrespective of (a). Joe’s fading qualia, which we can all accept as a consequence of the absent-qualia possibility, are posited on the assumption that a silicon brain cannot be conscious. Given that assumption, it then remains to be shown how Joe’s original beliefs (functional organisation) could be sustained anyway. But Chalmers wants it both ways. Even after (a) has been accepted, he must say that I am committed to (b) being possible anyway, irrespective of (a), in order to set up the argument. But I can only know that (b) is possible irrespective of (a) if I already know that Joe’s qualia have no effect on his functional profile. Otherwise, I would have to allow that his faded qualia might restrict the functional profiles available to him, and thence allow that (b) might be false. And since there is no independent justification for my commitment to (b), no reason why I should be committed to (b) once I have conceded (a), the reductio cannot even get started. Either it will always be possible to secure (b), irrespective of Joe’s qualia, simply by arranging for the correct input and output values to preserve functionality, or it will be possible only when Joe’s qualia permit. To the extent, if any, that fading qualia would force a change in functionality for Chalmers, they would do so for me too. He cannot have it both ways.

To summarise: Chalmers' argument depends on my being committed to (b), that a silicon implant can always be configured with the right input/output function to preserve Joe’s functionality. That is how I am supposed to fall into his trap. I acknowledge commitment to (b) and then Chalmers points out that I have just reaped the absurd consequence of my original position. I have not. If fading qualia really do present a problem for sustained functionality, in virtue of the strong tendency of conscious beings to have reliable beliefs about their own experiences, then they present that difficulty equally for both Chalmers and me. Hence, given my assumption that Joe’s qualia fade, I would never be forced unconditionally to accept (b) in the first place, and the ‘absurdity’ of my position would never arise.

Finally, it might be possible to make something out of the following. Even though Chalmers has failed to dismiss the possibility of Joe’s qualia fading out as the silicon takes over, his reliability thesis might still stand. Thus, it might still be conceded that Joe’s fading qualia must be reflected in his changing qualia reports. So doesn’t that show, anyway, that by the time Joe’s consciousness disappears his qualia reports will disappear too?
It seems that his qualia and his functional organisation are connected after all, at least to some extent, and hence that the mindless Robot must be an impossibility. But as I have already pointed out, what applies to a conscious Joe need not apply to the permanently unconscious Robot. Specifically, any strong link between qualia and qualia beliefs in a conscious subject will evaporate once consciousness is lost. We need only remind ourselves that it is precisely because Robot, being unconscious, is not so implausible per se that Chalmers constructed the story of conscious Joe. In fact, we should note that even a semi-conscious Joe with permanently feeble qualia remains possible, since Chalmers’ argument depends crucially on the intuition that Joe’s fading qualia would be noticed. He is not in a position to extend that claim to the case of a permanently semi-conscious Joe, for whom there would be no changes to notice.

2. Dancing Qualia
The situation envisaged under this heading is thought to be even more compelling than the first. Since the first argument collapsed irrespective of how compelling the envisaged absurdity might have been, however, because the derivation of that absurdity was itself logically flawed, we might find ourselves hoping for more rigour, rather than a higher level of compulsion, when considering the second. The envisaged situation this time involves two functionally isomorphic but constitutionally dissimilar beings – one carbon-based and the other silicon-based – who have quite different colour sensations when looking at one and the same object. The cast on this occasion includes the original carbon unit, his silicon-based functional isomorph, and Bill, who is receiving the progressive replacement of carbon- with silicon-based brain material. So Carbon experiences red when looking at a tomato, Silicon experiences blue, and Bill undergoes a gradual experiential shift from one to the other. Just to make things a little more vivid, Bill is fitted with a switch, which enables him to flip between the silicon- and carbon-based brain matter at will.

Here again, then, the initial consequence we are asked to accept, this time of allowing that Silicon’s inverted spectrum is possible, is that Bill is going to experience a change in his qualia. As in the first argument, it seems pointless to dispute this. For all I know it might be correct; as Bill receives the silicon implant and it is switched on his red qualia turn blue, and as it is switched off again they turn red. Of course, he is going to notice this. So we can at least assume for the sake of Chalmers’ argument that we would have to accept this description of Bill’s dancing qualia as a consequence of the initial inverted-qualia possibility. That is the compelling part.

Now Chalmers’ intended reductio once again depends on my being committed to two incompatible consequences, this time of the inverted-qualia possibility. In order to carry it through he needs to show firstly that I am committed to (a) Bill’s qualia changing from red to blue and back again, but then also that I am committed to (b) Bill failing to notice or respond to this change. (a), we are conceding, seems a natural consequence of the inverted-qualia assumption, since if silicon brains experience blue, where carbon brains experience red, there must surely be an experiential shift for a being migrating from one constitution to the other. But again there is (b). Clearly, the possibility of Bill holding onto his original functional organisation – his qualia beliefs and reports included – does not follow as a consequence of the inverted-spectrum possibility per se. The fact that carbon units experience red qualia, where silicon units experience blue qualia, and the two units are functionally isomorphic, tells us nothing in itself about the functional possibilities for Bill. For Bill’s experiences undergo a change, and it is his ability to notice this change that the argument relies on.

I think it is easier this time around to spot the error in Chalmers’ reasoning. Once again he is simply assuming, and expecting us to assume with him, as a matter of common sense, that, even if the inverted qualia hypothesis is correct, it must always be possible to preserve Carbon’s functional organisation, regardless of what is going on with his qualia. Having fallen into that trap, we are then promptly tripped up by the reductio – an appeal to the absurdity of where we have ended up. Again, Chalmers is having it both ways. Once I have put my money on the inverted-qualia possibility, it seems I am bound to accept that Bill’s qualia would change noticeably when he switched from carbon to silicon and back again. This much we can agree on. But given this much, we then have to ask whether it would be possible for Bill to preserve his original functional organisation – his qualia beliefs and reports – even so; even though his qualia are changing from red to blue. And, as in the case of fading qualia, it seems not. For if Bill’s qualia change noticeably the change will be noticed; his beliefs will follow suit. Hence, there will be no absurdity in our resultant position. As Chalmers insists, Bill will probably exemplify the principle of organisational invariance to a good approximation, as he notices and responds to his changing qualia. The commonsense assumption that Bill’s functionality could be held constant is just wrong. But again, that has no bearing on the possibility of two numerically distinct functional isomorphs having permanently dissimilar qualia.

In each case, then, Chalmers’ argument depends on the intuition that (b) it will always be possible to maintain Joe’s, or Bill’s, functionality, simply by arranging for the appropriate inputs and outputs to be in place. This, combined with the expectation that qualia reports will tend to be fairly accurate, would rule out (a); the possibility of fading or dancing qualia. But since there is no independent support for (b), and no support for (b) to be derived even from the absent-qualia or inverted-qualia hypothesis, the commonsense response would be simply to abandon (b) in deference to the obviously correct (a). The qualia-reports of a conscious being will tend to follow any changes in its qualia, and there would be no commitment to supposing otherwise. But this tells us nothing at all about the relationship between qualia and qualia-reports either in an unconscious Robot or in experientially stable conscious beings.

3. Further Note
From my defence of the absent-qualia and inverted qualia hypotheses against Chalmers’ arguments here, it would be tempting to infer that I actually subscribe to them. This would not be correct. My sole purpose has been to point out and clarify a logical non-sequitur in Chalmers’ argument. My claim is simply that his reductio does not work. It is perfectly consistent to accept the absent-qualia possibility and at the same time allow that a conscious being undergoing Chalmers’ silicon substitution will experience and respond reliably to his fading qualia. The two cases are logically unconnected, in virtue of the fact that only in one of them is there any conscious monitoring of qualia-changes at all. If that conscious monitoring does indeed effect restrictions on how Joe reports and responds to his fading or inverting qualia, even then there is no reason to suppose that any such restrictions are imposed on our entirely unconscious functional simulacrum. There is no prima facie absurdity in assuming the latter possibility, and Chalmers’ argument concerning conscious or semi-conscious, self-monitoring, Joes does nothing to change that. My position with regard to the inverted-qualia possibility is exactly parallel.

On the other hand, nor do I reject the absent-qualia or inverted-qualia possibility out of hand. I do not, in particular, find Daniel C. Dennett’s rejection [2] of absent qualia at all convincing. In essence, his position seems to be that a functional simulacrum of ourselves would, by dint of sharing all of our functional profiles, share all of our beliefs - even about our conscious experiences. But if, like Chalmers and unlike Dennett, we start out with the premise that our experiences are not fully describable in functional terms, Dennett’s rejection of absent qualia will miss the mark. While I do not attempt to argue the point here, for both Chalmers, I take it, and myself, beliefs about experience have qualitative content which no functional description can capture. Hence, on that view, there is no reason to suppose that a zombie which exactly emulates my own experience-related functional profile has any conscious experiences at all. By the same token, nor do I have any prima facie cause for concern that my own experiences might not be qualitatively at least as rich as I believe them to be. My point here, however, is just that the possibility of zombies has not been settled either way.

[1] David J. Chalmers, “Absent Qualia, Fading Qualia, Dancing Qualia,” in Thomas Metzinger (ed.), Conscious Experience (Schoeningh-Verlag, Paderborn, 1995), pp. 309-28.

[2] See, for example, “The Unimagined Preposterousness of Zombies: Commentary on Moody, Flanagan, and Polger.” (In Brainchildren: Essays on Designing Minds. Cambridge, Mass.: MIT Press, A Bradford Book, 1998. pp171-177 ).

Brian G. Crabb,
University of Liverpool,
February 2008

Article Source: http://journal.ilovephilosophy.com

Please Rate this Article

 

Not yet Rated

Click the XML Icon Above to Receive Philosophy Articles Via RSS!
Additional Articles From - Home | Philosophy
Sign Up as an Author.
Submit Articles
Member Login
Top Authors
Most Popular Articles
Submission Guidelines
Ezine Notifications
Article RSS Feeds
New Stuff
Forum
Link to Us
Contact Us

Powered by Article Dashboard