The cup from which I am drinking water now is yellow, and I know that it is. Why does my belief that the cup is yellow count as knowledge? Presumably, the answer must involve some reference to my current perceptual experience: I see the cup, and I see that it is yellow. What is it for me to see that the cup is yellow? The obvious answer would seem to be that it is for me to stand in a certain relation-namely, the relation expressed by the verb 'to see'-to a proposition, namely, the proposition that the cup is yellow: Perception, that is to say, is a kind of propositional attitude, like belief, though it is also different from belief in many ways. So my seeing that the cup is yellow, being a kind of propositional attitude I take towards that proposition, can count as my reason for believing that the cup is yellow. I've thus got a reason for that belief, and it's a good one. No sort of inference from a prior judgement about my experience is necessary. Rather, I need only import the content of my perceptual state into cognition to believe it.If one could accept that much, then-though there would no doubt be many problems left to discuss-it would make the question how perceptual experience justifies perceptual beliefs significantly more tractable. But the problem, as I see it, is that accepting that much threatens to impose high costs. If the 'importation' model of perceptual justification is to be extended to all perceptually justified beliefs, then every concept that figures in a belief that is perceptually justifiable for a given subject must also be able to figure in the content of her perceptual experience: Only what the subject can conceptualize can play any justificatory role for her. It does not, of course, follow that there is nothing in the content of one's current perceptual experience that one cannot now conceptualize. But even if there is, it can play no role in thought: It cannot so much as figure in thought. A fortiori, non-conceptual elements of experience, if such there be, can play no role in one's decisions about what to believe or, more generally, one might conjecture, in any rational process. Elements of one's current perceptual experience that cannot be conceptualized would be of no significance for the thinking subject. Or, as Kant put it, in a famous passage: