Imagination and Theoretical Inference

I spent the weekend at the Virginia Tech Graduate Philosophy Conference, where I presented my paper on dreaming and imagination. It went over well, I think -- it seemed to get people interested. It also got *me* interested in it again. Here's a question I've been focusing on for the past couple of days, largely thanks to insightful questions by Colin Klein and Jason Decker: how is it that imaginings can lead to beliefs? They clearly can and do -- I imagine one figure rotating and moving to the position of another and form the belief that they are congruent, or I imagine a fistfight between Al Sharpton and Al Gore and form the belief that Al Sharpton would win such a fight, or I imagine my apartment burning down and how I would react to it, and form the belief that I don't have a good enough evacuation plan. How does this work? I'd like to be able to tell a story about this on the model of theoretical inference, so the first step for me will be to figure out how plain old reasoning works -- I believe that p and I believe that p implies q, and I somehow manage to form the belief that q. I'd like to read up on how that works. Any suggestions? I know John Broome talked some about that here at Brown last year... pointers to published papers, online resources, etc. would be very welcome.