Now if qualia have a one to one correspondence or a one to many correspondence with the internal states of a conscious system (such that a single qualia can be present with a number of different internal states) then it is obvious that we can have knowledge about our own qualia. Simply knowing that such a correspondence exists is enough to be able to observe our own actions and use them to justify the claim that we are in a particular qualitative state. For example, we could justify the claim that we are in pain by appealing to the fact that we verbally claim we are in pain. Because such a correspondence exists there is only one qualia that corresponds to the verbal behavior of asserting that we are in pain (ignoring for the moment the fact that particular verbal behaviors can correspond to a number of different internal states) we can conclude that we are actually in pain, and not mistakenly thinking that we are in pain when we are really experiencing some other qualia. Of course this all turns on there being a reliable correspondence between particular qualia and verbal behavior, but that is a correspondence that we have good reason to believe exists.*
But if there is a many to one correspondence between qualia and internal states then things become more complicated, and I maintain that in such a situation we cannot actually have knowledge about our own qualia. Let us suppose that there is a conscious system with internal state I1 followed by internal state I2. And let us suppose that two different qualia, q11 and q21, can be associated with I1. Now the simple case is when they converge, meaning that only a single qualia q2 can be associated with I2. Obviously then experiencing q2 tells you nothing about whether you previously experienced q11 or q21, and so in cases of such convergence qualia with a many to one correspondence with internal states are unknowable.**
Now lets turn to the harder case, where the qualia don’t converge, and q11 and q21 are followed by q12 or q22 respectively. Now things become trickier. Since the systems never differ in internal state obviously they won’t diverge in behavior, which means that we could never express the knowledge that we were in q11 instead of q21, but perhaps that is because qualia are inexpressible. Maybe we just don’t have the right language for talking about qualia, and never will as a matter of principle. However knowledge requires more than just the true belief that we experienced q11, it requires that such a belief be justified. How could we justify it? Even if we are currently experiencing q12 that doesn’t justify concluding that we experienced q11, unless we can justify the law that q12 only follows q11. But we could never justify such a law. To justify it we would have to know which particular qualia we experienced and then motivate the law as an inference from our observations. But this presupposes that we can know when we are experiencing various qualia, which was the very thing we set out to demonstrate that we could do, and so justifying such knowledge is impossible. Thus since we can’t have knowledge of our own qualia both when they do and don’t converge in the case of a many to one correspondence between qualia and internal states we can’t have knowledge of such qualia at all.
Of course there is a small exception to this, which is if all the qualia that can correspond to a particular internal state have something in common (assuming qualia are complex rather than simple). In that case we could know that we had experienced whatever it is that they all have in common. But this just illustrates the point, the thing that they all have in common has a one to one correspondence with internal states and thus is knowable, although which qualia it was part of (of if there was anything more to it) is unknowable. In any case this shows that if qualia are something that we are supposed to have privileged access to, in the sense that we have knowledge about them that other people don’t, then qualia must have a one to one correspondence with internal states.
* One reason to believe that such a correspondence exists is if you think that experiences only exist when feedback exists within a system. That is to say that the state of the system at one moment has an effect on the way the system processes information subsequently. A convenient way to conceptualize systems in which such feedback exists is like a tree, with the processing starting at the trunk and making its way to the branches, each of which represents a different subsystem / information processing ability. Feedback occurs when some of the branches send information back to the trunk. Now experiences, and hence qualia, only occur when such feedback occurs. Which means that all qualia go through the trunk and thus make their way to each subsystem. And one such subsystem is the language center. And so under such a model it is reasonable to suppose that different qualia will result in different linguistic behavior.
** Here some may question the premise that a second internal state is required in order to have knowledge about the qualia associated with the first. Why can’t we just know directly, while experiencing the first internal state, what qualia are associated with it? Well we can’t because knowledge is a propositional attitude, to the effect that something is the case. Thus a second internal state is required to embody the belief that we experienced a certain qualia when we were in the first state. Of course that state also has its own qualia, but to know we experienced the qualia associated with that state a third state is required, and so on. Now normally we don’t experience such states, which implies that we don’t “know” our own qualia most of the time (though we do experience them, this is not a higher-order claim), but that isn’t two surprising since in that same sense I don’t “know” what 185 – 57 is. Although I have the disposition to be able to acquire the relevant knowledge in both cases I must acquire the correct belief before I can be actually said to have this knowledge.