It seems to be a central claim of possible worlds semantics that propositional attitudes have, all of them, the same content. Furthermore, possible worlds semantics seems to assume that perception and belief states can have the same content – provided that perception is also understood as a propositional attitude. I think these claims fail to account for a very simple and intuitive difference between perception and belief. That is, the intuition that external objects are not objects of belief, whereas they clearly are objects of perception.
This intuition is supported by our ordinary accounts of perception and belief. We do not ordinarily think that the house in front of me can be believed or thought of, but we clearly speak of it as if it could be seen and touched. There is nothing at all problematic about the sentence ‘I cannot see that house’, but there is something terribly mistaken about the sentence ‘I cannot believe that house’. Notice how different this is from the sentence ‘I cannot believe that there can be such a house’. This one sounds right. However, it seems mistaken to say ‘I cannot perceive that there can be such a house’.
Thought, on the other hand, seems quite amiable with belief. One can substitute any belief sentence for a thought sentence and nothing weird comes out. Similarly, one cannot substitute any perceptual sentence for a thought sentence without saying something wrong. It is mistaken to say ‘I cannot think that house’, but nothing problematic comes out of saying ‘I cannot think of there being such a house’ – though you would be confessing a pretty limited imagination thereby.
Nonetheless, according to possible worlds semantics, the content of a belief state is a set of possible worlds, or a set of possible individuals, or a set of relevant doxastic alternatives according to Lewis. If so, then the objects of belief and thought are just the same as those of perception. Possible objects and individuals certainly are perceivable (at least the actual ones). But what makes us think that they are also believable, or thinkable? It seems then, that possible world semantics cannot account for this intuitive distinction between objects of perception and objects of thought.
One might try to make it up by arguing for a difference content and object of the attitude. If one’s theory of propositional attitudes takes the ‘content’ and the ‘object’ of a mental state to be metaphysically distinct entities, then nothing should be inferred about content. It might still be, for all that theory claims, that perceptual and intellectual (i.e., perception and thought) content are the same. I don’t know how much can be done here. I am not convinced that there is actually any good distinction between the object and the content of a propositional attitude. It is clear to me, however, that these states do not have the same objects. We perceive objects, but do not believe them. Thus, if one claims that the content of a mental state is a possible object and, hence, that it is an object of perception, like possible worlds semantics does, then one is failing to account for an important distinction between objects of perception and objects of thought.
A controversial way to make this distinction would be, I guess, to say that there is no such thing as belief ‘de re’ if the ‘res’ of the beliefs are the same as those of perception. Perception, on the other hand, seems quite amiable with its being ‘de re’ and, also possibly, ‘de dicto’.