Re: "What I have for dinner depends on what

[Pycyn@aol.com]

What there is on the semantics of indirect questions in the linguistic (or, 
better, logic) end of things is an appendix to the stuff on (direct) 
questions.  The lead names here are Belnap and Harrah and the lead theory is 
that a question is the set of its proper answers (full sentences, by and 
large, since those are easier for logic to deal with).  Some purists 
(Montagovians, by and large) hold out for the set of true answers, but that 
does not work as well for indirect questions.   To be sure, in the case of 
"know," we want to partition the set of answers into the true ones, all of 
which the knower knows, and the false ones, none of which he even believes 
(though he may also not even believe their denials in the case of unknown 
potential, but not actual, party goers).  On the other hand,  with, say, 
"wonder" the whole set is involved apparently and the issue just what the 
partition is.  
By the way, what is the restriction on preds that can take an indirect 
question? We can know or wonder also sorts of them, but we can't believe one 
(but then how about "You won't believe who I saw yesterday"?) or think or 
claim or ..., all of which take regular indirect discourse.
pc