Reference By Constitution?
I guess I should to clarify my argument. The position I want to argue against consists of the following two claims:
1) "pain" denotes a physical entity, say CFF.
2) For no P that only contains physical terms is "PI am in pain" a priori.
"" is the material conditional. I've chosen the pain example only for brevity: if you think it matters, feel free to replace "pain" by something like "the phenomenal quality of my current red-experience".
I make two further assumptions:
3) For some physical P, "P
4) If "P
(2) and (4) together imply
5) For no physical P is "P
Together with (3) and (1), this implies
6) For no physical P is "P
But, I claim, given (1), (6) is extremely implausible. On any semantics I can think of, sufficient physical information always entails whether a given string, individuated physically, denotes a given physical property. Semantic properties aren't inscrutable. Or rather, in so far as they are inscrutable, they are indeterminate. There is no such thing as inscrutable determinate reference.
Three further remarks on the argument.
First, I realize that a type-B materialist could question (4). But if the rest of the argument goes through, she is still committed to the claim that for some physical P, "P I am in the state denoted by 'pain'" is a priori. And that already sounds very much like type-A materialism.
Second, more needs to be said about "pain" and "denotes": Is (1) false should there turn out to be a language in which the physical string "pain" denotes the moon? No. Presumably, "denotes" means "denotes-in-English", or "denotes-in-my-idiolect". The argument goes through either way. One could also index "pain" by a language, and interpret (1) as saying that "pain"-in-English denotes CFF. For reasons that will become obvious this is not how I want (1) to be read.
Third, the argument also works against certain versions of token-identity theory that are strictly weaker than type-B materialism: simply replace "pain" throughout by something like "token 217 of 'pain'".
David Chalmers suggests that in response, the type-B materialist could argue that "the phenomenal state stands in a relation of constitution to the representation of it". I don't think this works, but I'm not sure I fully understand the constitution relation. Here is Chalmers' explanation in The Content and Epistemology of Phenomenal Belief (p.235 of the dead tree version):
[Sometimes] the content of a phenomenal concept [...] is partly constituted by an underlying phenomenal quality, in that the content will mirror the quality (picking out instances of the quality in all epistemic possibilities), and in that across a wide range of nearby conceptually possible cases in which the underlying quality is varied while background [i.e. other] properties are held constant, the content will covary to mirror the quality.
I guess this is meant to define the constitution relation. But notice the relata: Constitution relates a phenomenal quality on the one hand to the content of a phenomenal concept on the other. It doesn't relate qualities to concepts themselves, let alone to representations of concepts.
This is why constitution doesn't provide an explanation of how (6) above could be true. The question there is what could make it the case that a certain name, individuated non-semantically, denotes the physical entity it in fact denotes. My own answer appeals to facts about use, causal connections, etc. of which we can in principle know a priori that if these facts obtain, then the name has that denotation. But this contradicts (6). Answering that the content of the name is partly constituted by its denotation only shifts the question. I then ask what makes it the case that the name has this particular content.
Perhaps one could say that at least for words in the language of thought, the word itself, as a physical string, is partly constituted by its denotation. One could even hold that words for phenomenal qualities in LOT just are their referents. But then the constitution is arguably derivable from physical information. Moreover, and more importantly, this doesn't answer my question. My question is about "pain", or "the phenomenal quality of my current red-experience", not about words in the language of thought.