/grabbing _Dennett and his Critics_ off the book shelf...
Not much of a fan of his. Nice to have people floating in the deep end, but I'll continue to wade for a while longer.
My current theory (which has as much research behind it as one coming from someone who does not do neurological studies all day can) is this: sleep is the integration of short-term memory into long-term memory. Dreams are your body's reaction to this, hallucinogenic due to various -cholines, and isolated from most motor circuits by the depression of -amines. cf. J. Allan Hobson and Antonio Damasio for that latter chemical bit; first part's mine--developed after reading _GEB_ in high school. This fits well with my own personal experience, at least: when I run short on sleep for a period of time, my dreams begin to have fewer long-term elements; then, when I can catch up on sleep, I find my dreams including elements from further and further back until I reach the point at which I began to seriously lose sleep. Don't know about y'all but I have *no* difficulty "interpreting" my dreams--the elements are readily mapped to my recent experience.
And to boxley, perhaps you're basing dream-interpretation on extreme emotions because your conscious mind naturally puts more focus on them? Dreams have just as many boring or quirky elements as they do anxious or epiphanic ones--don't ignore the boring bits when you develop your theory. :)