Friday, December 09, 2005

Liar, Liar

A few days ago Rad Geek guestblogged a fascinating post on insolubilia at Philosophy, etc. that I wanted to say something briefly about. I'll have to return to the issue when I have the time, but I am much in favor of the medieval approach to insolubilia. Take a typical paradox:

(L) This statement (L) is false.

The modern reading regards this as problematic on the face of it: if (L) is true, (L) is false, and if it is false it is true, so we cannot assign a truth-value to (L) -- that is, it cannot be true or false. The medievals had a completely different approach. The medievals tend to see insolubilia as sophistical inferences. On their view, the real paradox lies not in sentences like (L) but in what we are assuming that we can infer from (L)'s being true or false. They are certainly right about that. A modern reader of (L) might hold, of course, that the sort of inferences that generate the paradox are necessary; but even if you hold this view, the medievals are to be congratulated on not taking it for granted. On most medieval views, (L) would be false; and inferences from its truth-value would all be illegitimate. (On one possible view this would also be true of the truth-value of "This statement is true", which yields no paradox; i.e. the restriction is not necessarily ad hoc, but done on principle.) This sort of discussion gets very complicated very quickly; I'll have to look at it more closely at a future time.