Sunday, June 25, 2006

Why do semantics?

Why should one do semantics? There is a quote from Davidson (I think I read it in the early chapters of the Heim and Kratzer text) that says roughly, semantics should not tell you anything you don't already know as a speaker of the language. My question is obviously directed at specific kinds of semantics, namely formal semantics in either the Montagovian or Davidsonian paradigm. It applies somewhat to other kinds though. If these theories aren't telling us anything about the meaning of sentences that we didn't already know, what is the reason for pursuing them? There are a few technical kinks to work out in different places. In the Montagovian tradition there are problems with type mismatching and non-intersective adjectives. In the Davidsonian tradition there are problems with context sensitivity.

I suppose that philosophers are interested in semantics because it is supposed to make clearer some philosophical problems. The best example of this would be to help in coming up with a general theory of meaning. By creating semantic theories for different languages, we can see what sorts of properties they share, and so what sorts of properties meaning itself has.

Another shot at an answr is that by getting clear about what the formal meaning of a sentence is, then we have taken a step at getting natural language into a form on which we can do serious logical inquiry. We can start proving theorems and finding out what follows from what. This is an extension of the Fregean-Russellian idea about the logical form of language.

Maybe producing clear, formal truth-conditions for sentences will help us make sense of philosophically loaded words like 'necessity' or 'virtue'. This seems pretty doubtful.

Studying syntax, as linguists do it, gives (or is purported to give) a glimpse of the furniture of the human mind, so even if nothing else comes of syntactic theory, it will have given us some more understanding about the mind. Of course, it looks like other stuff will come of syntactic theory, so it is doing pretty well for itself.


Aidan said...

On certain views of the kind of knowledge we attribute when we attribute knowledge of a semantic theory (if that's our wont), the Davidson quote actually makes a fair bit of sense. Dummett, for example, held that we shouldn't attribute explicit knowledge, because then the process of producing a correct semantic theory for one's language should be easy. (I'm bracketing issues about the conceptual resources and technical competence we might have to attribute to each competent speaker too on this view).

He did think, however, that we could ascribe implicit knowledge of a semantic theory to competent speakers of a language. I don't remember what Davidson had to say on this matter, but if one thinks of semantic knowledge as implicit knowledge, one can reconcile Davidson's remark with regarding the construction of a theory of meaning as a non-trivial and potentially philosophically and linguistically illuminating task.

Shawn said...

That is a helpful. I have a tendency to view philosophical semantics in isolation from other areas. This causes a bit of existential angst, if I may be permitted that phrase. It also makes me want to branch out some. Being able to tie things back to semantics is kind of a nice though. I'm not that familiar with the Davidson-Dummett debate, although it seems like that would be a good thing to look in to.

Brad said...

Certainly Davidson regarded "the construction of a theory of meaning as a non-trivial and potentially philosophically and linguistically illuminating task" (of course, he had a unique conception of 'theory of meaning--- "theory of meaning is not a technical term, but a gesture towards a family of problems..." but that isn't the point) . The thought that this position would need reconciling with Davidson's "remark" stems from the fact that no remark of Davidson was given, but only a rough charecterization of a remark of Davidson.

Presumably, Heim and Kratzer were quoting some early period Davidson (if they were quoting from anything not collected in Inquiries, I will be flabbergasted), perhaps in the spirit of this ditty from T&M:

"The theory reveals nothing new about the conditions under which an individual sentence is true; it does not make those conditions any clearer than the sentence itself does. The work of the theory is in relating the known truth-conditions of each sentence to those aspects ('words') of the sentence that recur in other sentences, and can be assigned identical roles in other sentences. Empirical power in such a theory depends on success in recovering the structure of a very complicated ability--the ability to speak and understand a language. We can tell easily enough when particular pronouncements of the theory comport with our understaning of the language; this is consistent with a feeble insight into the design of the machinery of our linguistic accomplishment."

Now, the question is
"If these theories aren't telling us anything about the meaning of sentences that we didn't already know, what is the reason for pursuing them?"

Davidson's original line of thought on this is summarized at the end of the same paper.

"In this paper I have assumed that the speakers of a language can effectively determine the meaning or meanings of an arbitrary expression ( if it has a meaning), and that it is the central task of a theory of meaning to show how this is possible. I have argued that a charecterization of a truth predicate describes the required kind of structure, and provides a clear and testable criterion of an adequate semantics for natural language."

I don't think Davidson ever really repudiated this view, but an important change did come when he moved from the truth and meaning stage of his career into the radical interpretation stage, and then again into the unified theory state, and then finally into the triangulation stage (this might better be described as a repeated refining of his original position). By the end of his career he was saying things like

"Though radical interpretation cannot, of course, explain or describe how a first language is learned, it is hard to believe it does not throw light on what helps make that accomplishment possible." (Comments on the Carlovy Vary papers, p. 294 of Interpreting Davidson)

Also, I think the Davidson-Dummett debate that your talking about is the A Nice Derangement-Comments on Hacker and Davidson-Social Aspect of Language triad. The two Davidson papers are collected in the last oup collection of his papers. The Dummett is in the Lepore collection from the 80s called Truth and Interpretation.

Brad said...

For what it is worth, I was leafing through the collection Interpretations and Causes this morning and saw the Lepore has an article in there that talks about the issue Aidan discussed (in what respect we can say a speaker has knowledge of a semantic theory). That collection might be good to get out of the library anyways because it has a really good paper by Bilgrami at the end of it.