Sunday, September 21, 2008

An Objection from Stalnaker, and a Reply

I'm not sure if this will work. Let me know what you think.

The linguistic picture of content (LP) can be characterized as the conjunction of the following two theses:

(T1) The structure of an attitude ascription ^a v’s that P^ mirrors the syntactic structure of a sentence used to ascribe the belief that P to a.

(T2) The structure of an object of belief, P, mirrors the syntactic structure of those elements of sentences that are used to designate P.

If (LP) is true, then it seems to follow that propositional attitude ascriptions describe relations holding between agents and determinate, sentence like objects. The inference would basically look like this:


(i) If (LP) is true, then [the structure of an attitude ascription ^a v’s that P^ mirrors the syntactic structure of a sentence used to ascribe the belief that P to a] and [the structure of an object of belief, P, mirrors the syntactic structure of those elements of sentences that are used to designate P].

(ii) If [the structure of an attitude ascription ^a v’s that P^ mirrors the syntactic structure of a sentence used to ascribe the belief that P to a] and [the structure of an object of belief, P, mirrors the syntactic structure of those elements of sentences that are used to designate P], then an attitude ascription of the form ^a v’s that P^ expresses a relation between a and a determinate, sentence like object o.

(iii) So, if (LP) is true, then an attitude ascription of the form ^a v’s that P^ expresses a relation between a and a determinate, sentence like object o.

On 64/65, Stalnaker describes a case of propositional attitude ascription. If his description accurate, we have a counterexample to the principle expressed in the consequent of (iii), which would entail the falsity of (LP) (and hence the falsity of either (T1), (T2), or both, assuming that (LP) is meant to hold in all cases of attitude ascription). Here’s a version of the argument he runs.

Consider the following two true propositional attitude ascriptions:

(A1) Dan believes that Godel’s first incompleteness theorem is true.

(A2) Baby logic Bob believes Godel’s first incompleteness theorem is true.


(1) (A1) and (A2) are true attitude ascriptions.

(2) If (1), then, if an attitude ascription of the form ^a v’s that P^ expresses a relation between a and a determinate, sentence like object o, then there exists a (unique) determinate, sentence-like object o such that Dan believes o and Bob believes o.

(3) So, if an attitude ascription of the form ^a v’s that P^ expresses a relation between a and a determinate, sentence like object o, then there exists a (unique) determinate, sentence-like object o such that Dan believes o and Bob believes o. (1,2)

But consider what Dan and Bob are able to infer on the basis of the beliefs attributed to them in (A1) and (A2), respectively. Presumably, Dan is able to infer all kinds of stuff, while Bob is able to infer next to nothing. We infer that while Dan fully understands the content of his assertion, Bob does not.


(4) Dan fully understands the content of his assertion.

(5) Baby logic Bob does not fully understand the content of his assertion.

(6) [(4) & (5)]

(7) If (6), then it is not the case that there exists a (unique) determinate, sentence-like object o such that Dan believes o and Baby logic Bob believes o.

(8) But if that’s right, then it is not the case that an attitude ascription of the form ^a v’s that P^ expresses a relation between a and a determinate, sentence like object.

(9) And if that’s right, then (LP) is false.

(10) So, (LP) is false. (6-9, iii)

I think there is at least one way out of this problem for the structured propositionalist. They can block the argument in premise (7), by denying the consequent of that conditional, and say the same thing in this case that they say when faced with problems posed by the substitution of co-designative proper names in propositional attitude contexts. One way to do this is to analyze an attitude ascription of the form ^a v’s that P^ as expressing a two place relation (v) that holds between a and P iff a three place relation holds between a, P, and some mode of presentation of P. If we construe a mode of presentation as a set of propositions that one is able to infer on the basis of believing P, we can see how it is that Dan and Bob are in a position to infer different things on the basis of standing in a two place relation of belief to P. They can do this because each of them stands in a two place relation of belief to P in virtue of being one of the relata of a three place relation holding between themselves, P, and some mode of presentation of P, where the mode of presentation of P in each of their respective cases is not the same.

No comments: