Meillassoux and the Realist Sense of Statements

I’m reading Meillassoux’s After Finitude in order to teach it next week in the realism seminar (note to those who saw Harman’s post linking to my site: I am badly behind on updating it, and plan to do so when it’s needed). I am not arguing myself for an anti-realism (in Michael Dummett’s sense) but I’ve always thought the weakest part of this work is the chapter on ancestrality, but that’s like saying that I really enjoyed the new car you gave me for the holidays but I didn’t like the bow you put on it. Basically, Meillassoux offers a not-very-good rendition of phil of science after Popper, and not a very good linguistic theory to back it up. I’m leading to a sincere question, fyi, and I’m only a short way through, so i’ll likely post more:

1) He argues that there is a “real sense” to scientific statements. (I don’t have time to summarize the work for those who haven’t read it. Here’s a summary review I published a year ago.) He then says that, of course, after Popper, no one thinks science is anything but revisable, but that the meaning of its statements are real, if contingent (they can be proved wrong). But this would seem to lead to an infinite deferral of a real reference: you’ll have to wait for it to be proved conclusive, and thus for now we can say “as of now, according to x, y, and z scientific theorems, this is the referent for this statement” … Again, I am not an anti-realist, but it does seem that Meillassoux’s argument here has built-in a number of problems a good phil of science class would drive right through.

2) Ok, so then comes in the problem of the referent .Now, here’s where Dummett can be brought in. (Never thought I would write that.) Realism would be any discourse that claims a “literal” and “real” referent. Ultimately Dummett claims that one can test the truth claims of a conceptual scheme without this referent (and in fact, the more one leans on reference, the more one has to give up on truth–but that’s another matter), but what’s important is that Meillassoux claims there’s no compromise between the correlationist and the archifossil (I would cite the English version page numbers, but I only have the French version in front of me). The ancestral statements, he writes, have a “meaning only on the condition that its literal sense is also its ultimate sense” (p. 35). I’ve marked in my own text a number of places where Meillassoux’s implicit linguistic theory is employed. In any event, he says that the ancestral statements have “only a realist sense.”

The whole relation between reference and meaning is something I’ll leave to the Fregens Fregians readers of early analytic philosophy, but it’s obviously fexed and Dummett owes quite a bit to revisiting Frege’s work. My question, though, is more straightforward: doesn’t Meillassoux’s claim about the ancestral apply to all sorts of discourses?

Let me cite the example that would irk him the most, I think: the Freudian unconscious, just for discussion’s sake. (1) it can never have a sense except as a referent beyond the statement since that in the strictest sense is exactly what it is; (2) it cannot be made the correlation of any perceiver; (3) yet it is said to be “real.” One can thus say there is no “compromise between the correlationist and the archi-drives of the unconscious.”

Of course, then you’ll say to me, but there’s a difference. And I’ll say, well, you were talking about sense and reference, and if that’s the case, all realist discourses are analogous: they take their conceptual schemes and add one crucial statement (just as the correlationist adds one crucial statement to all scientific facts): all this and these statements refer to a non-given, independent reality. (But the stuff on donation and 1st and 2nd meanings is pretty knockdown of Heidegger saying scientific statements are “secondary” to the object’s ready-to-handness.) How are we not left with a circular “but mine comes with set theory, no matter what mathemes Lacan would start using…”? Again, I’m not arguing for an anti-realism, but opening up a question…