I've just been reading this old post from the Language Log on the Sapir-Whorf Hypothesis, and it reminded me of a question I wanted to think through a while back and never got around to. The question is this: does Burge-style anti-individualism underwrite a principled philosophical defense of the Sapir-Whorf Hypothesis? Here is why you might think that it does.
According to Burge, mental states have externalist individuation conditions, not just with respect to the local environment, but also with respect to the linguistic community in which one is situated. We individuate beliefs (and other propositional attitudes) via their content, and we individuate content in accordance with constraints on the publicity of meaning. Suppose, for example, that there are two concepts associated rhuemetoid ailments. Arthritis is the concept of a rheumetoid ailment restricted to the joints; tharthritis, by contrast, is the concept of a rhuemetoid ailment not so restricted. The English word "arthritis" expresses the concept of arthritis. Now consider an individual, Oscar, who tells his doctor that he believes that his arthritis is spreading to his thigh. What Oscar says, namely, "I believe that my arthritis is spreading to my thigh," is false since by fiat arthritis can't occur in the thigh.
But what about Oscar's belief itself? There are two possibilities. (1) Oscar has the false belief that his arthritis is spreading to his thigh; (2) Oscar has the (possibly) true belief that his tharthritis is spreading to his thigh. Burge argues convincingly that the correct answer is (1). Oscar's concept is the concept of arthritis, and he (as it were) mistakes that concept for the concept of tharthritis.
Now consider a situation in which everything in Oscar's linguistic history is held fixed but, entirely outside the realm of his experience, the medical community has fixed "arthritis" to mean tharthritis. Nothing internal to Oscar's psychology has changed, only external facts about the language community. In this case, the proposition Oscar expresses when he utters "I believe that my arthritis is spreading to my thigh" is the (possibly) true proposition that Oscar believes his tharthritis is spreading to his thigh. Moreover, his belief is no longer a belief about arthritis, but about tharthritis! But clearly the only thing that has changed between the two cases is which linguistic community Oscar is embedded in. Thus, our mental states (i.e., what thoughts we have) are individuated in part by our linguistic community.
As the case has been described, we have held everything fixed except the external linguistic environment. This allows Burge to locate the shift in belief in the shift in the linguistic environment. But once the conclusion has been established, it would seem to hold more or less across the board: under normal circumstances the contents of our thought is dependent upon the linguistic community in which we happen to be located.
Now here is the million dollar question: do the semantics of at least some linguistic communities differ significantly enough from one another that members of the two communities would typically exhibit Burge-like divergences in their thoughts? If the answer here is "yes", then we seem to have the makings of a relatively robust defense of the Sapir-Whorf Hypothesis (or something like it). Put generally (and putting to one side further anti-realist considerations), the Sapir-Whorf Hypothesis is the thesis that culture, via its influence on language, substantially affects our thoughts about the world.
The answer to this question is, at least in part, an empirical matter. But I do think that there are some general considerations in philosophical semantics that might well lend support to the S-P Hypothesis. In particular,there are good philosophical reasons for thinking that many (perhaps most) words have various types of categorial information associated with them and that this categorial information might vary culture-wise. One such bit of information is whether or not the meaning of the word is functional or (micro)structural. For example, most philosophers of language are now convinced that the concept expressed by the word "water" is microstructural, in the sense that nothing could be water (i.e., fall under that concept) unless it is composed of H2O molecules. But it is difficult to believe that this is the only reasonble way of handling the semantics of a word of the "water" category. [By this, I just mean a word which is at least roughly translatable as by the English word "water".] In particular, there seems to be no principled barrier to having a "water" word which is categorially functional/descriptivist rather than microstructural. Call the resulting concept macro-water (Korman). Now, if the anti-individualist thesis is correct, then whether your thoughts are water thoughts or macro-water thoughts will depend (in part) on facts about your language community. And it isn't unreasonable to think that different cultures might well differ quite generally in their propensity to take one or the other (or, if you are fond of two-dimensionalism, both) of these tracks. The former we might broadly characterize as atomistic cultures; the latter, as functionalistic ones. [Brian Weatherson had a link to a paper that argued something like this, but alas I couldn't find the reference.]
Does this really give us anything like a substantial S-W Hypothesis? Well, consider the concept expressed by a functionally oriented community who conceives of time as cyclical or nonlinear or whatever!