Wolfgang Schwarz

Type-B Materialist Semantics (Again)

Brandt Van der Gaast points out that Michael McDermott proposes something like the semantics I sketched on behalf of type-B materialism in his "The Narrow Semantics of Proper Names" (Mind 1988). That's true. But I think McDermott is almost silent on the matter crucial to type-B materialism, and there is no acceptable way to fill the silence without spoiling type-B materialism.

McDermott defends a kind of global descriptivism modulo experiential ('input') concepts and 'output' concepts. (The latter are presumably concepts in terms of which decisions to act are represented, but he doesn't say much about them.) The idea is that all other terms get their meaning in Ramsey-Lewis manner from their place in total theory. The crucial question for type-B materialism then is how the input and output concepts get their meaning. Not by their place in theory, says McDermott. We identify sensations "directly", not via descriptions, he says.

Fine. Assume first my total theory is represented by certain physical tokens in my brain -- sentences in my language of thought. According to McDermott, "the only non-logical terms in [LOT] are names for kinds of in- and output" (p.236). So here is one of these names. Let's say it denotes a kind of red sensation. But why? What makes it the case that it denotes a red sensation rather than a blue sensation, or Gottlob Frege? Not because of theory, okay. But that can't be the end of the story. The only answer I can see is that the name stands in some physically describable relation to red sensations: probably a causal relation, perhaps even identity. But this won't help type-B materialism. For the type-B materialist claims that no amount of physical information can tell you what the name denotes.

Could he argue that even though the name gets its reference by means of a certain physical relation, this fact is itself not entailed by any physical information? That sounds just as incredible. Once we know that the name is systematically correlated with physical event (and candidate sensation) A but not with B, it doesn't make much sense to wonder whether it might nevertheless denote B rather than A.

As McDermott notes, it is more plausible to avoid any talk about LOT and instead construe the content of someone's beliefs as whatever makes best sense of his behaviour by the light of ideal psychology. McDermott also claims that belief ascriptions in ideal psychology must be framed in a special, "narrow" vocabulary whose only non-logical components are, again, names for sensations and outputs. But we can't just introduce new names, say "Fred", "Paul", "Jones", etc. and hope that they somehow manage to attach to specific in- and outputs. Instead, ideal psychology has to use descriptions to pick out the in- and outputs, like "the sensation typically caused by visual confrontation with red objects". At this point the game is lost for type-B materialism. McDermott adds that such a characterisation is not part of the ascribed content: the content is somehow "de re" about the characterised sensation. But that doesn't save type-B materialism.

(By the way, even though I agree with McDermott that ideal psychology should be "narrow", I don't agree that therefore belief ascriptions must only contain narrow vocabulary. The content of a belief system is not the ascribed sentences. It's not sentences at all. At best it is what these sentences express: something like a distribution of credence over a set of centered worlds. Narrow vocabulary is not the only way to specify such a distribution. With some care, it can also be done in English.)

Compare Quine's holism, which closely resembles McDermott's: the primary unit of significance is total theory; the content of a theory is (determined by) the observation sentences that confirm or disconfirm it. How do the observation sentences get their content, the content in virtue of which they can confirm and disconfirm? Not via total theory. Quine says they "wear their meanings on their sleeves" (Word and Object, §10, p.42). This sounds like they somehow got their meaning "directly", by the kind of magic required for type-B materialism. But what Quine means is just that their content can be identified with their stimulus meaning. That is, observation sentences have their content in virtue of being suitably correlated with particular stimuli (in all relevant speakers). No hope for type-B materialism here either.