May 31st at 18:00, in F1.15
Circumstantialism is the thesis according to which the content of sentence (i.e., its proposition) is its truth-conditions, that is the set of circumstances where it is true. The main circumstantialist account in the theory of content is the one for which the content of a sentence is the set of possible worlds where it holds. This induces a natural semantics for knowledge where epistemic states are modelled as sets of possible worlds and an agent is said to know a sentence p if his or her epistemic states is a subset of the content of p. It is not difficult to see that this account has the following undesirable feature: if a sentence p is necessary (i.e., true at all possible worlds) then it is known. This problem is called 'logical omniscience'. It has been said that it follows from the theory of content induced by the possible world semantics. There are two possible reactions to that: reject circumstantialism or try to amend it. I am interested in the latter. The idea is to keep circumstantialism while rejecting the thesis that the relevant circumstances are possible worlds. In this talk, I distinguish between several kinds of logical omniscience and investigate how different sorts of circumstantialist accounts can avoid them. In particular, I study (i) the impossible worlds account according to which the relevant circumstances are worlds but but not necessarily possible and (ii) the truthmaker account according to which the relevant circumstances are parts of worlds. A problem is that avoiding logical omniscience seems to come at the cost of even minimal rationality. Therefore, I propose my own semantics for knowledge, based on Fine's truthmaker semantics, that manage to avoid (some kind of) logical omniscience while maintaining moderate rationality.