For lack of a better idea, I propose to use the uniform prior on it.
If we consider only universally quantified statements: Conjunctive statements in each world are redundant as the world itself is a conjunction anyway. If each world contains only disjunctive statements, shouldn’t the worlds that assign truth to longer statements be more likely, as it is easier to satisfy disjunctive statements with more terms?
Um, I don’t see value in excluding statements that have e.g. a forall-exists-forall-exists sequence of quantifiers at the outer level?
But yeah, there may well be some systematic bias, and I have no particular reason to believe that easier satisfiability of long disjunctive statements could not be the cause of such a bias… unfortunately, my insight isn’t good enough to offer anything but guesses. As I said in the post, I’m really not sure whether this is a good choice of prior or not; the main point of having it is to have something concrete to think about, possibly as a stand-in for a better prior over the same set of “worlds”.
If we consider only universally quantified statements: Conjunctive statements in each world are redundant as the world itself is a conjunction anyway. If each world contains only disjunctive statements, shouldn’t the worlds that assign truth to longer statements be more likely, as it is easier to satisfy disjunctive statements with more terms?
Um, I don’t see value in excluding statements that have e.g. a forall-exists-forall-exists sequence of quantifiers at the outer level?
But yeah, there may well be some systematic bias, and I have no particular reason to believe that easier satisfiability of long disjunctive statements could not be the cause of such a bias… unfortunately, my insight isn’t good enough to offer anything but guesses. As I said in the post, I’m really not sure whether this is a good choice of prior or not; the main point of having it is to have something concrete to think about, possibly as a stand-in for a better prior over the same set of “worlds”.
I just wanted to show it seems possible to do better and that was easier to do considering only a subset of statements.
Ah, ok.