Phenomenal Concepts, This and That
This morning I have been revising (once more...) my reply to Chalmers' "Phenomenal Concepts and the Explanatory Gap". In this paper, Chalmers poses an important objection against the phenomenal concept strategy, and in my reply I try to defend the strategy.
The phenomenal concept strategy tries to respond to conceivability arguments in this way. Remember that CA say that a sentence such as (1) 'if the physical facts are such and such, then Luce is in pain', if a posteriori, has to be false at some possible world. The PC strategy replies by saying that the aposteriority of such sentence can be explained merely in psychological terms, independently of the modal status of the sentence. That is, the sentence is a posteriori because of the nature of the concepts involved, not because it is contingent.
I think this is a very promising strategy. The problem comes when we try to give an account of the difference between physical and phenomenal concepts, so that it follows that sentences such as (1) are a posteriori.
In my paper, I was trying to explain Hill & McLaughlin account (in their (1999) paper, in PPR), but it is not easy to explain it clearly and briefly!
One thing they seem to say is that phenomenal concepts are self-presenting (that is, we can apply them just by virtue of being in the state they refer to), whereas physical concepts are not so. This seems right, but why should this explain that phenomenal concepts are not a priori connected to physical concepts? It seems that more needs to be said in order to infer the a priori disconnection between physical and phenomenal concepts from the fact that the norms we use to justifiably apply physical and phenomenal concepts are different. I'm not sure how to motivate this move.
Another thing they say is that the mechanisms we use to fix the reference of physical and phenomenal concepts are psychologically different. This could explain why concepts such as 'C-fibre firing' and 'pain' are a posteriori related. Still, what are those reference-fixing mechanisms, and why do they have the consequence that the corresponding concepts are not a priori related? H&M say that we fix the reference of 'pain' by means of a recognitional disposition, and the reference of 'C-fibre firing' by means of a physical-theoretical description. This seems right. Is this enough to explain why those two concepts are not a priori connected? This seems more promising to me: there could be an explanation hidden there, somewhere! This is something I'll attempt to flesh out in my thesis.
However, I prefer to take (what I take to be) Loar's line: being a priori connected is a psychological property, which either holds or does not hold between two concepts, and it is just a brute psychological fact that they are so (un)connected. I guess that "a brute psychological fact" is a psychological fact that cannot be explained in psychological terms, but rather in terms of a lower-level (neurobiology, or whatever).
So the crucial claim of this version of the strategy might be the following: it is possible (and plausible) that our cognitive systems are such that physical and phenomenal concepts are not a priori connected (this psychological fact might just be a consequence of structural facts about the realization of our cognitive systems). Then, this explains why sentences such as (1) above a posteriori. This explanation is compatible with (1) being necessary. So (1) being a posteriori does not entail that (1) is false at some possible world. So the conceivability argument fails.
I think this is very promising line of reply. Notice that it is not committed to (1) being necessarily true. It just says that there are plausible claims about our concepts that, if true, would entail that (1) being a posteriori has nothing to do with it being contingent. Therefore, inference from a posteriority to contingency is not warranted, unless one shows that those plausible claims about our concepts are wrong.