It has been generally accepted that you can’t know what a sensation is like without experiencing it yourself. No experience of pain, no understanding of pain; no seeing of red, no knowing what seeing of red is. Possessing the concept requires instantiating what it is a concept of. This is not true of states designated physical: you can know what measles is without having it, or how a bat’s wing is constructed without having one, or what C-fibers are without having any in your brain. So sensations differ from physical states in this epistemic respect: sensations require having them to know them, but physical states can be known whether you have them or not. But why is this? What is it that makes it the case that there is this epistemic asymmetry? Why is it that I can know what C-fibers are without having any C-fibers but I can’t know what pain is without being in pain (at some time)?
The answer is crushingly obvious: I can see C-fibers in someone’s brain (perhaps using a microscope), so forming the concept; but I can’t see pain in someone’s brain (with any microscope), so forming the concept. Lacking sensory perception of pain, I have to fall back on acquaintance with instances of it in my own person (I can’t be aware of pain in another person in this way). I have to get the concepts from somewhere, and perception (mainly vision) allows me to get the concept of C-fibers by looking inside brains (or things similar to brains), while the concept of pain cannot be acquired in that way, but must be based on introspection.  I can only introspect myself, but I can perceive many things other than myself. Thus it appears that we can assert the following counterfactual: If we could see sensations, we would know what they are without ourselves needing to have them. Suppose I couldn’t see C-fibers or anything like them, or sense them in any other way: then I wouldn’t know what C-fibers are—they would be like an exotic species I’ve never set eyes on (or heard about from someone who had). They would be an epistemic blank to me. That is our position with respect to sensations: we have never seen bat experiences, for example. Nor have we ever seen our own sensations, but here we have direct acquaintance with them in acts of introspection—as we don’t for bat sensations. But if we could see sensations–as they intrinsically are, as we see physical things–then we would not need to fall back on introspection. The trouble is that this counterfactual looks like it has an impossible antecedent: we necessarily can’t see sensations. This necessity claim is very strong, and even if true applies only to human beings. Is it true that no possible being could conceivably perceive sensations? It is not as if sensations lack causal powers or spatiotemporal coordinates—why should they not be in principle perceivable? And yet it is hard for us to conceive of what perceiving them might be like. That might just be a limitation on the human imagination, born of our particular cognitive faculties; maybe there are logically possible sensory systems that can respond to sensations as ours respond to streaks of lightning or cold air. But even if this is not possible the counterfactual still holds, only now with an impossible antecedent: if (per impossibile) we could see sensations, then we would know what they are in that way. So this is still the explanation of why we can form concepts of physical states, but not of sensations, without having them: the reason for the asymmetry is that one is perceptible and the other is not.
Consider a character, Billy, suspended in a tank under conditions of sensory deprivation: suppose Billy has many bodily sensations but no perception of outer objects; in particular, he feels pain but has never seen a brain or anything like one. He knows very well what pain is, since he feels it every day, but he has no idea what C-fibers are, having never seen any, or anything like them. He is an expert phenomenologist but a physical ignoramus. One day he is liberated from his tank and given outer senses: a brain is placed before him with C-fibers prominently displayed and he gazes at it for a good long time. Now he knows what C-fibers are, not having known this before, even though he knew very well what pain is: he has learned something new. Therefore C-fiber firing is not reducible to (identical with) pain: it is something over and above pain. We have thus proved that there is more to the world than sensations: it would not be plausible to maintain that C-fibers are nothing more than pain—or else Billy would have known about C-fibers just by knowing about pain.  Whether or not this argument reaches its conclusion, it illustrates the dependence of certain concepts on perception: these concepts cannot be derived from introspection alone. The reason for Billy’s earlier ignorance about the nature of C-fibers is that he lacked perception of them; and the same can be said for our ignorance of (say) the sensations of bats—we lack perception of them. We only know the nature of our own sensations because we have another route to such knowledge, i.e. introspection. If we lacked introspection, we would have no concept of our own sensations either.
What if we could know both things both ways? Suppose we could know sensations from our own case andknow them via perception: then we wouldn’t have any problems of conceptual limitation vis-à-vis other types of experience. We could know what it’s like to be a bat without having bat-type experiences. This would open up our knowledge of mind considerably; maybe it would enable us to solve the mind-body problem, by locating the psychological in the realm of perceptible things.  It would certainly involve a conceptual transformation. Similarly, suppose that we could know physical states by perception and know them by introspection: then a lack of perception of them wouldn’t bar us from arriving at an accurate conception of what they are (unlike Billy). One might have a sensation as of one’s C-fibers firing (not the same as a sensation of pain); and this too might contribute to solving the mind-body problem, by locating the brain in the field of consciousness. As it is, the duality of perception and introspection underscores the duality of physical and mental states, but that duality might reflect our epistemic predicament more than any underlying ontology. A perceptual concept for sensations would render sensations more objectively comprehensible, while an introspective concept for brain states would render them more subjectively comprehensible (someone who lacks the C-fiber sensation can’t know what it’s like to have that sensation). As it is, however, we are stuck with a sharp duality of understanding: one thing we understand perceptually, the other introspectively. We are thus saddled with an unbridgeable subjective-objective divide. The point I have wanted to make here is diagnostic: the reason for the divide is that we can’t see sensations. That may be a contingent truth or it may be a necessary truth, but it is why our concepts for sensations are as they are; if we conceived them perceptually, we would have no trouble extending our psychological understanding beyond our own case. Bat minds would be transparent to us—as transparent as their bodies and brains. I strongly suspect that the limitation here is contingent, though human beings would have to change dramatically (unrecognizably) in order to become mind perceivers: I think there must be beings in possible worlds that can see (etc.) sensations. You just have to hook up a sensitive surface to the intricacies of sensations in such a way that the corresponding percepts reflect fine distinctions in the sensations perceived. In any case, that is the ground of the difference: we can see brains but we can’t see minds. Just think how much simpler your intellectual life would be if you could see both. The “world-knot” might unravel before your eyes, literally.
 For one thing, seeing sensations would allow us to draw pictures of them, as we can draw pictures of C-fibers. We could then set the pictures side by side and compare them. As it is, we can draw no picture of pain to compare with our picture of C-fibers. What would that do to the mind-body problem?