I have just read through the latest draft of fellow Brains contributor Pete Mandik’s paper Transcending Zombies which he has up over at Brain Hammer. In it Pete wants to
develop advice to the reductionist about consciousness in the form of a transcendental argument that depends crucially on the sorts of knowledge claims concerning consciousness that, as crucial elements in the anti-reductionists’ epistemic gap arguments, the anti-reductionist will readily concede.
I have some qualms about the line of argument he offers. The basic argument goes as follows
P1. If I know that I am not a zombie, then phenomenal character is (a certain kind of) conceptualized egocentric content.
P2. I know that I am not a zombie.
P3. Phenomenal character is (a certain kind of) conceptualized egocentric content.
P4. Fixing my physical properties fixes my conceptualized egocentric contents.
C. Fixing my physical properties fixes my phenomenal properties.
Pete claims that the bulk of the work is done by P1. So let’s look at Pet’s argument for that before looking at some other issues. The basic idea is that in order for me to have certainty with respect to the fact that I am not a zombie I need to be able to have a belief that I have qualia and that needs to be true. This, however, isn’t enough. I need to be able to characterize my experience in terms of concepts, represent the qualia as being mine (the representation must be egocentric in Pete’s terms), and further, the conceptualized, egocentric contents must be identical to the qualitative character of my experience.
We must first point out that so far there is nothing here that the anti-reductionist need necessarily object to. They may simply affirm the line of argument while holding that the conceptualized egocentric content is not reducible to any physical state of Pete. This means that a dualist like Chalmers can accept the argument all the way to P3. So it really is P4 that is doing the heavy lifting, so let’s look at the argument for it.
Prima facie, it looks like physical similarity would entail conceptual similarity, that my physical doppelganger is my conceptual doppelganger.
But now we have left the realm of things that are neutrally acceptable between the reductionist and the anti-reductionist! A dualist like Chalmers will insist that in so far as non-physical qualitative properties are (partly) constitutive of my phenomenal concepts, a physical duplicate of me could have a pseudo phenomenal concept but they would not have the full-blown phenomenal concept that I do. Pete is aware of this kind of objection and tries to head it off by saying,
Some, like Chalmers (2003), may object that fixing my physical properties does not fix all of my concepts, since phenomenal character is non-physical and there are some concepts, so-called direct phenomenal concepts, which can be possessed only if one has had or is having a state with phenomenal character. My main complaint against this response is that I don’t think there are such concepts as concepts one can only have if one has had or is having a state with phenomenal character. I address this issue at greater [length] elsewhere (Mandik, 2009b). One brief point to make here, though, is that really direct phenomenal concepts, concepts had only while one is currently having a state with phenomenal character seem not to be concepts at all for their violation of both the reidentifiability criterion and the criterion of endogenous triggering (Prinz 2007 pp.207-208 makes a similar point).
Since Pete gives only one reason for doubting the anti-reductionist claim, let’s look at it. He objects to what he calls ‘really’ direct concepts by pointing out that they fail two plausible constraints on concept possession. But the anti-reductionist can agree with this and insist that a full-blown phenomenal concept is one that is partially constituted by a non-physical qualitative property and that physical duplicates can lack these non-physical properties and so lack full-blown phenomenal concepts. The anti-reductionist can agree that really direct phenomenal concepts are strange and insist that once one has acquired the relevant full-blown phenomenal concept it is available for endogenous triggering and reidentification. Thus the argument fails at what is really the crucial stage. What we need is an argument that phenomenal concepts are such that they can be acquired without having the relevant experience and that is not given (it may be given in the reference Pete cites, but the point is that since it plays a crucial role in teh argument it should be given here. What is given here is insufficient to make the point).
Secondly, it is not clear why the anti-reductionist must hold that full-blown phenomenal concepts work like this. Why couldn’t they simply adapt the typical empiricist claim that full-blown phenomenal concepts are innate? It would then be the case that the we could have full-blown phenomenal concepts without having the relevant experience (though perhaps they have to be triggered or something like that). At the very least more needs to be said to make this crucial step of the argument work.
The necx stage of the argument is to try to establush that fixing Pete’s physical properties fix his egocentric properties. But there is a problem with this kind of argument. Pete argues that a physical duplicate of him will necessarily have distinct egocentric properties (that duplicate will be thinking of itself, not the actual Pete). But since this is the case the argument he gives will not move the anti-reductionist. The anti-reductionist can agree that fixing Pete’s physical properties fix his phenomenal properties, since the phenomenal nomologically supervenes on the physical. What should be at issue is whether or not fixing the physical properties of a physical duplicate is enouogh to fix that duplicate’s egocentric contents, arguing that it works here, with these laws, is to miss the point of the zombie argument. Everyone agrees that given the way things are here in out world fixing physical properties is enough to give you phenomenal properties. The question is whether a physical duplicate has to have their phenomenal properties fixed. It is unclear how talking aout the actual Pete helps with this question. Now Pete, again, acknowledges that a move like thismight be made but says that it gets the same response as the move about phenomenal concepts already discussed. But that move is problematic, and so this move is problematic.
Therefore the advice given is not helpful advice for defeating the zombie intuitions that anti-reductionists have. The anti-reductionist can accept the argument as given by Pete while insisting that it doesn’t defeat intuitions about zombies.
A better strategy is to deploy a style of argument that the anti-reductionist is committed to. This is why the zoombie and shombie arguments are better for defeating zombie inuitions. All of the moves made are identical, but opposite, to those made by the anti-reductionist. It thus becomes very clear that what we have here is two parties merely asserting that something is conceivable (zonbies & ghosts/zoombies & shombies). Since both can’t really be conceivable one of us must be wrong and nothing that we know at this point can settle the issue. Thus issues about physicalism/dualism are empirical matters that cannot be settled by a priori methods.