Dreaming, Belief, and Emotion (by guest blogger Jonathan Ichikawa)
Colin McGinn thinks that we believe what we dream. In his book Mindsight, McGinn raises lots of questions about dreaming and imagination, but only briefly considers the suggestion that we "no more believe our dreams than we believe our daytime reveries." That's my view. He rejects it on the grounds that without invoking belief, we cannot explain our emotional involvement with dreams. Here's a quote:
The sure test that dreams are suffused with belief is their ability to generate emotions that are conditional on belief, such as fear and elation—with which dreams are full. (p. 112)
The problem for the imagination model, then, can be stated thus:
(1) When I dream that p, I experience fear, elation, and other emotions of a certain type.
(2) Emotions like fear and elation, arising from an attitude that p, can only arise from a belief that p. Therefore,
(3) When I dream that p, I believe that p.
This argument strikes me as pretty bad. This parallels a puzzle in the philosophy of literature involving emotional responses to fictions. Fictions arouse emotions in us without causing belief; we seem to be happy that p, even though we do not believe that p. This is what philosophers of literature call the "Paradox of Fiction," and it takes the form of an apparently-inconsistent triad:
(1') When I read in a fiction that p, I experience fear, elation, and other emotions of a certain type.
(2) Emotions like fear and elation, arising from an attitude that p, can only arise from a belief that p. Therefore,
(3') When I read in a fiction that p, I do not believe that p.
It is generally accepted that (3') is true, so philosophers of fiction agree that it is either (1') or (2) that has to go. We may, with Kendall Walton, deny that we really experience fear and elation, but rather experience different, similar states, which he calls quasi-fear and quasi-elation. Or, we may say with Derek Matravers and others that belief isn't really necessary for fear; imagination can also play the role that belief often plays in fear, and likewise for the other emotions. I prefer the latter option, but I think it's obvious that one of these has to be right.
If we take the latter option to solve our puzzle about fiction, then we have also directly avoided the problem for the imagination model by denying the shared premise (2). If, on the other hand, we insist that these emotions include a cognitive element, denying instead that fictions really generate these emotions, then we may very well ask whether dreams really generate them either. It will be, perhaps, more plausible if we qualify our denial of emotional involvement in dreams thus: dreams don't involve emotions, except in the way that fictions do.