Excerpts

From work in progress. I have taken Scott Soames' Understanding Truth to be a paradigm example of a certain view of propositions that posits them in order to perform a variety of functions such as being the primary bearers of semantic truth and establishing an "objective" synonymy relation between sentence utterances.

It might be argued that since it is prima facie tempting to believe that someone is talking about some sort of abstract entity when one speaks of “what is stated” on an occasion of utterance that there are actually such entities and that this is all we need to begin at any rate explaining how they will fit into our taxonomy of linguistic objects. The problem, even more that one of “individuation”--as Quine might put it--is identification. It is taken as clear, by this view, how to metaphorically point out the particular proposition one wants to discuss. This is usually found in the supposedly unproblematic statements like, for a sentence called ‘p’

(EP1) The sentence p expresses the proposition that there is a cat on the mat.

What proposition is supposed to be pointed out by the italicized phrase? I should note that italicizing the beginning of the that-clause as

(EP2) The sentence p expresses the proposition that there is a cat on the mat

won’t get us anywhere since the italicized sentence ‘there is a cat on the mat’ is at issue: what, pray tell, is its content? Likewise a problem with

(EP3) … that *s*

where ‘s’ stands for the name of a sentence [and '*' replaces the notation of corners introduced by Quine]. This move might avoid pointing out a particular proposition. But that is question-begging since the point is that we get nowhere without being able to pick out particular propositions for discussion. What is the problem?

The difficulty is that we use language to attempt to express propositions about propositions, if the view that there are propositions and they are expressed by sentence utterances holds. [I have since come to seriously question the line just described.] Meaning, as Quinian stimulus meaning or some other similar notion, has been done away with (at least by Soames) in favor of the notion of a function that maps sentences in all their potential utterances to the propositions uniquely expressed by those utterances. (EP1) is an illustration of the notion that indeterminacy undermines the idea that we can identify the propositions expressed by utterances. If we suppose that the italicized sentence expresses a proposition, it will be the case that only a sentence in my idiolect or one that I can already translate into my idiolect will be one for which I can identify the proposition. The only way to point out the proposition will be to use a sentence in an idiolect at a time, and of course no other idiolect will contain that sentence unless identical to the original utterance. Other specifications of the proposition will be allowed only under translation. But of course there can be no translation except relative to a scheme. So the proposition supposedly expressed by a sentence can only be identified by the speaker of the utterance at the time of utterance (and for as long as she keeps her idiolect); from within no other idiolect will there be any correct answer as to which proposition an utterance expressed. This is contradictory to Soames’ assertion that a synonymy relation holds between two sentences expressing the same proposition, unless the only thing synonymy amounts to is that sameness of proposition. This may be rebutted as an (merely) epistemic problem. When I use my sentence in my idiolect in a context, there is one and only one proposition that it expresses. That no one else, including myself later, can pick it out is a problem of identification but not one that indicates there is nothing to be found. However if this is all synonymy amounts to, then identity of behavioral dispositions gets us the same relation between sentences without mucking about with abstract entities. Metaphysical parsimony would indicate dropping propositions from this role.

A further complication is that sentences with the same meanings may have utterances in different languages. The sentence of which ‘Der Schnee ist Weiss’ in German is an instance has the same meaning function as the sentence of which ‘Snow is white’ in English is an instance. So the two sentences types in context mentioned in the previous sentence are two inscriptions of one sentence. The obvious next step is to become Davidsonians and deny that there is more than one conceptual scheme, for every natural “language” potentially contains ways of expressing exactly the same things as every other. But I mentioned at the outset that sentences for Soames appeared to be individuated syntactically. It is not supposed to be loose talk that sentences (1) and (2) from the first section are different sentences that mean the same thing—that they are synonymous utterances and not part of the same sentence. Yet if propositions are not themselves syntactically differentiated, something else must be at work behind the scenes; the candidate that comes to mind is a historical story about the origin of the sentence’s syntactico-lexical form. [In Kaplanesque fashion.]

All of this by the way of establishing the ground on which the dispute is to be resolved. The ultimate point, it's turning out, is just that I will show propositions to be useless for establishing a synonymy relation between sentences. The basic line seems to be that such a more presupposes such a relation and is therefore question-begging.