In the following I look at two fronts on which you might battle Chalmers’ arguments against physicalism about consciousness. The first is from Polgar’s recent paper (discussed recently by Richard Brown) in which Polgar briefly critique’s Chalmers’ implicit theory of reduction. The second is the more common strategy of attacking him for assuming he can conceive of zombies in the first place. I argue that the second strategy is better, though they are not mutually exclusive.
A weaker front: Supervenience and reduction
In his paper Are Sensations Still Brain Processes? Polgar says the following about Chalmers’ zombie argument against physicalism:
[T]he a posteriori identity theorist quite reasonably rejects the problematic notion of “logical” conditions and the positivist demand for a special kind of “reduction” of which there are no non-trivial examples and that has long since been rejected by philosophers of science. The burden is on the objector to show independently that the identity theorist, or any physicalist, should require such reductions. And no such positive argument has been given.
Recall Chalmers’ argument: if physicalism is true, fixing the values of the physical facts logically fixes the value of the facts about consciousness. This is equivalent to the supervenience claim that different consciousness-facts entail different physical-facts. Note when I say ‘physical facts’ I mean facts about things in physics, brains, stuff your run-of-the-mill naturalist likes to use as a supervenience base.
It seems, then, that the inchoate view of reduction Polgar refers to is the ‘logical fixation’ of one set of facts by another. In practice, Chalmers says we need to determine if physical-facts fix consciousness-facts by using a conceivability test. Can we conceive of C-facts shifting while the P-facts stay the same? If so, then physicalism fails.
I’m not sure what is positivisic about Chalmers’ view, or what in his view has long been rejected by philosophers of science. Intuitively, he seems right. When we fix the chemical facts we have fixed the higher-level facts about this cup of water (e.g., is it in liquid or solid phase). When the facts about molecules in this gas are fixed, the facts about its temperature, pressure, etc. come along ‘for free’ so to speak. This seems reasonable, basically a restatement of the claim that consciousness supervenes on the physical.
Based on such considerations, Polgar’s claims seem a weak front on which to attack Chalmers’ argument, especially when communicating with people that don’t know any philosophy. Perhaps a philosopher in the house can clarify or defend what Polgar was saying.
More promising: tempering the conceivability arguments
Another argument against Chalmers accepts his argument strategy, but attacks the high credence he gives his intuitions about how C-facts and P-facts logically relate to one another. How do facts about brains relate to facts about conscious experiences? Our understanding of both sets of facts is so undeveloped that Chalmers’ confidence seems premature.
By analogy, many people don’t understand how facts about energy relate to facts about mass, they can’t conceive of any possible logical route from one to the other. Most people’s understanding of energy is about as clear as our present characterization of phenomenal facts, so this seems an apt analogy. While facts about energy don’t supervene on facts about mass, that shouldn’t change the conceptual point.
The analogy people usually bring up against Chalmers is vitalism. Vitalists couldn’t conceive of how physico-chemical facts related to certain biological facts, and used this to infer that the physico-chemical picture of life was incomplete. The vitalist Driesch stated his argument strategy quite nicely when he claimed (in ‘Science and Philosophy of the Organism’ (1908), p105):
[S]omething new and elemental must always be introduced whenever what is known of other elemental facts is proved to be unable to explain the facts in a new field of investigation.
Driesch’s argument for vitalism was an application of that general inference rule. For instance, he argues (ibid, 142):
No kind of causality based upon the constellations of single physical and chemical acts can account for organic individual development; this development is not to be explained by any hypothesis about configuration of physical and chemical agents. Therefore there must be something else which is to be regarded as the sufficient reason of individual form-production…
This is a good example of a conceivability argument hitting the rocks.
Chalmers would likely argue that the analogy fails because in Dreisch’s argument the target facts were “easy” facts about development, not “hard” facts about consciousness. This would be to miss the point that we need to exercise extreme caution when consulting our conceptual intuitions about what follows from what. That we now believe one group of facts is easy to reach from the other is clearly a contingent fact.What guarantee can Chalmers offer that he isn’t falling victim to a similar failure of imagination, lured by his contingent limited understanding of facts about brains and facts about consciousness?
I don’t really buy Polgar’s argument, but am willing to be set straight. The more obvious and central problem with Chalmers’ arguments is his willingness to take too seriously how he presently conceives of the relationship between two targets that themselves are not particularly well defined. Indeed, we could argue that our present understanding of the two targets is even less well delineated than Driesch’s two targets were when he was writing.
… but attacks the high credence he gives his intuitions.
Quite. Am I the only featherless biped still walking, who was taught in his youth that intuitions are of no value whatsoever and are an embarrassment to appeal to in sophisticated circles?
I will accept any proposal arguendo, but give it no weight whatsoever just because it is accompanied by stamping feet or an especially earnest tone of voice.
on the first point, i think that what polger is really opposing isn’t so much the demand for reduction as the demand for a priori entailment from low-level truths to high-level truths. like other type-b materialists, he thinks that the fixing relation itself is essentially a posteriori.
on the second point, obviously i’ve written a lot about what it would take for further empirical information or further understanding to close the epistemic gap. one relevant place is the material on type-c materialism in “consciousness and its place in nature”
[insert stamping feet and earnest tone of voice here]
Thanks David: if someone disagrees with this a priori entailment business, can’t they still buy the more general argument strategy of refuting supervenience to refute physicalism? That is, the strategy as I described it? There are many ways to reject a supervenience claim, and one of them is empirical after all (I can easily refute the claim, on empirical grounds, that facts about my dog supervene on facts about the pencil in my drawer).
Also, I’m not sure why the type B materialist would mind such a logical entailment as long as the original broad-strokes identity was discovered a posteriori.
I’ll have to read the Block/Stalnacker and your response. I’m embarassed to say I haven’t. I guess that will clear up my confusion.
I myself haven’t heard anyone talk about supervenience in quite a while, but maybe I just don’t hang out in the right circles…I myself find that it is too weak of a notion to do much good partially just because of the kinds of issues that Eric brings up….the main issue here is one of a priori entailment or not and this mirrors the issue of conceivability of zombies or not. If there is an a priori entailment then zombies are not conceivable, if there is no a priori entailment then zombies are conceivable….so you can see why it would be hard for a type-b person to admit that there *could* be an a priori entailment since that would mean that zombies are not conceivable (but the type-b position is pretty much defined as the position that they are conceivable but not (metaphysically) possible….My own position, as you may know, is the just the one you describe. I think that the mind/brain identities will be discovered by us a posteriori even though in principle they are knowable a priori. But I claim that zombies *seem* conceivable to us just because we are not in a position to actually make the a priori deductions, once we are we will see that they never were really conceivable…
Since people are already foot-stomping I figure I would join in…though in a sternly optimistic tone of voce
No you’re not the only one. But you may be the only one who didn’t subsequently realize that you can’t take the tiniest logical step without some kind of intuition to back it up.
Last night I asked:
can’t they still buy the more general argument strategy of refuting supervenience to refute physicalism?
In the morning, the answer is obvious. Sure, that is the whole strategy. But refuting supervenience empirically won’t work for Chalmers, at least for consciousness. It would work if there were evidence for the existence of consciousness in the absence of brain
activity, for instance.
The problem is that
because of his principle of structural coherence between conscious
states and “easy” processes, this empirical strategy will not work for
Chalmers. Hence, the argument must poke through a priori, and this
seems a reasonable point of contention for Polgar etc.. I too am suspicious of such conceptual analysis being a good guide to such complex scientific problems (largely for reasons laid out in the second half of my post, so perhaps there is some synergy there after all).
I guess that positivistic in the sense that they are promoting
conceptual analysis as a useful endeavor, believe in the importance of
analytic philosophy, and other antiquated ideas
Another option I guess would be to say that a priori and a posteriori aren’t mutually exclusive, as people assume. A kind of type Q approach?
Me and my formalist friends, and rather a lot of our physicalist associates.
I have a question regarding the analogy between Chalmers’ and Driesch’s arguments. In the case of vitalism, the claim was that “organic individual development” could not be reduced to interaction between physical processes. But we now know that the motivation for this claim is based in a misunderstanding of physical processes. The reductand (I hope I’m using this term correctly?), life, was known to be a structural/functional process, namely that of organic development. What was inconceivable was that particular kinds of structural/functional processes, namely those of physics and chemistry, could underlie the macroscopic structural/functional processes of life. But both were admittedly structural/functional properties.
In the case of consciousness, however, there still seems to be an argument that consciousness might *not* be a structural or functional property. If our intuitions about possibility could be brought to bear on whether or not consciousness is a structural or functional property in such a way as to vindicate these intuitions, would this be sufficient to disrupt the analogy between dualism and vitalism?
I realize that there are a number of issues with the above approach. First, there are those who claim that consciousness is a posteriori identical with some physical and therefore structural/functional property(ies). Second, one of the primary means of arguing the conclusion I described is to argue for an entailment from conceivability to possibility. But this argument has a number of difficulties and has been resisted in many ways. (These resistances, in my opinion, seem to have been effective inasmuch as they have succeeded in shifting the focus from a conceivability-possibility (i.e. straight metaphysical) argument for dualism to an anti-reductive explanation (i.e. epistemologically mediated) argument for dualism.) But it still seems to me that the argument that consciousness be fundamentally non-structural/functional has not been entirely exhausted. It seems that one would have to find some way of meeting concerns about the possibility of a posteriori identity with a structural/functional property and explaining a tenable link between conceivability and possibility that doesn’t rely on entailment. Not trivial tasks, but not impossible, either (so it seems to me).
If anybody has any comments or resources that might answer my questions, please do share them. Thanks.
David: I tried to get at that in the original post when I wrote:
Also, I discussed the hard problem intuition, that consciousness cannot be explained in terms of causal/functional/physical/biological processes, at my post The Meta-Hard Problem. It’s a bit heavy-handed (I wrote it when I was annoyed), but I still agree with the main point, and especially how I reformulated things in terms of the ‘Hard Problem Template’ in the comments.