One thing that I, and others, have been pressing Devitt on is his treatment of semantic types. As he has admitted in class, he is primarily concerned with token sentences as a way of evaluating the thoughts that those tokens are taken to be representitive of (confirming his commitment to P-semantics). This is what leads him to say that something is a meaning if and only if it plays a semantic role, which means that it can be used to explain bevaior and as a guide to reality.
When it was pointed out that this seems to indicate that sentence types, or for that matter word types, don’t have any meaning Devitt appealed to the distinction between physical types and semantic types. This wasn’t a suprise since he had said the same thing to me after he read Kripke, Devitt, Bach. By physical type Devit means the physical structure that all tokens of ‘Aristotle’ share. The physical type ‘Aristotle’ is ambiguous for him. There is a distinct semantic type for each thing named Aristotle (which is a collection of all the tokens that are causally related to the particular object that they name), though only on physical type (at this point Neale asked if utterances and sentences could belong to the same physical type or not, to which Devitt said ‘I haven’t thought a lot about types!’).
He then offered the following tentative definition of the meaning of types
A property of a physical expression type is a meaning if and only if that property, together with the context, determines a meaning for all of its tokens
The second occurence of meaning here is supposed to be the P-semantic one, the one he defined in Comming. He says ‘ a meaning for all its tokens’ because he thinks that tokens can have more than one meaning. According to this definition, then, for the physical type ‘Aristotle’ to have a meaning is for it to have a property which determines a meaning for all of its tokens.
But this is itself ambiguous. It could mean
1. The property of the physical type determines a single meaning that all of the tokens share (though there may be other meanings that all the tokens share)
2. The property of the physical type determines a meaning for each token that may vary from token to token.
If 1 is the case then it will turn out that every token of ‘Aristotle’ will have the property of picking out Aristotle the philosopher no matter who the token actually refers to. This in effect means that every name token refers to every bearer of the name…wierd!
If 2 is the case then there must be some property that the name type has that acts as a function that when applied to a name token, in a given context, will give us the refferent of that name. What could such a property be? The natural candidate is something like NDT (being the theory that a name N is semantically equivelent to the description ‘the bearer of “N”‘). Who the bearer of “N” is will be determined by the causal relation that some object has to the thought that is being expressed by the sentence in which the token occurs. So it looks like Devitt should endorse frigidity as an L-semantic theory…though I somehow doubt that he would agree. When I brought this up in class, he said ‘perhaps names are a bad example…’