Transcendental idealism is a lot like normal idealism with some subtle differences (differences that make a world of difference when we are considering taking it seriously). Transcendental idealists generally don’t deny that a world external to us exists, or that it is objective. They do, however, claim that this world is constructed or constituted by our perceptions of it. According to the transcendental idealist there is no such thing as part of the world that cannot be perceived because our perceptions to some extent define the world. (Of course even within transcendental idealism there are variations, for example does part of the world exist when no one is perceiving it?) We can contrast this position with transcendental realism, which holds that there is some independently existing world, but that we can’t get at it except through appearances.
In my opinion both transcendental idealism and transcendental realism are flawed, but today I am going to deal only with transcendental idealism. Transcendental idealism is certainly preferable to plain old idealism; regular idealism would have us abandon the notion of objectivity, and perhaps slide into solipsism, which is not a problem with transcendental idealism. Transcendental idealism does, however, share a problem with regular idealism, dealing with error and disagreement.
If it is true that the world is constituted or defined by our perceptions of it then how are we to recover the idea of a perception being in error? Let me address specifically a Husserlian kind of transcendental idealism, where objects in the world are said to exist as a harmonious synthesis of perceptions (that means that the perceptions about the object are in some kind of agreement and that they present themselves as perceptions of the same object). What are we to say when our perceptions of an object are in disagreement? (For example, disagreement over time, where an object looks one way close up, but differently at a distance; or disagreement between modalities, where an object looks a certain way, but our sense of touch presents it as another.) Are we to say that no object is being constituted? Or that there are different objects at different times or for different modalities? We could say such things, just as the idealist is free to say that there is no such thing as objectivity, but to do so seems to be to throw out something we wish to save. We want to be able to say that the world is a certain way, and that we can perceive it either accurately or inaccurately, but without allowing for the existence of a world outside of perception transcendental idealism blocks us from describing cases of error in this way. And even the notion of objectivity may be threatened by this problem, when two people disagree about the contents of the world because of their different perceptions of it. How can the transcendental idealist decide who is right, since truth is generally not thought of as a democratic process?
I do not raise such issues because transcendental idealism in general is a powerful force in philosophy at the moment; of course there are some who work with transcendental idealism still, but it does not dominate contemporary philosophy as it once did. However, the philosophy of mind still does grapple with a kind of limited transcendental idealism. Specifically it is claimed that there is nothing more to conscious experience than what appears as part of conscious experience. And unlike normal transcendental idealism this seems to be a rational claim, after all being aware of some fact does seem to be a requirement if it is to be conscious. And to adhere to such a position is not to say that every mental act and structure must be available to us, there could be a large supporting, unconscious, framework that is responsible for consciousness. Additionally, when dealing with consciousness the inability to deal with error may not be a problem at all, since it does seem rational to think that people cannot be in error about their own consciousness, at least not at the moment it is occurring.
And I do not, in principle, have a problem with these claims. However, sometimes we are tempted to take the method of transcendental idealism too far with respect to the mind. It is going too far to say that because it appears to us that there is some feature of consciousness that there must be such a feature. This is to confuse the appearances with the underlying reality, because while transcendental idealism may be a good approach when dealing with consciousness by itself it falls apart when we are considering the mind as a whole. And when we are wondering about some feature of consciousness we are wondering “is there really such a feature in the mind?”, that is, does the reality match up with appearances? And the answer may very well be “no”. It may be that all there is to some conscious feature is the appearance of that feature. For example, it appears to us as if our blind spot is filled in. Is the visual data available to consciousness filled in, or do we just think that it is? This is a question that can’t be answered by appealing to appearances. Sure it appears to be filled in, but that appearance could have a number of causes, not all of which correspond to some kind of filling in of the visual field. Ultimately when we seek to explain consciousness we want to explain and understand the underlying reality, and deal with the appearances only as appearances.
To really make the distinction clear we need to pin down exactly how something can be an appearance to consciousness, which would have the effect of highlighting how the appearances in consciousness, that we can’t be mistaken about, can differ from reality, which we can be mistaken about. I propose that we do so via an “as if” definition. More specifically I am proposing that saying something appears (in consciousness) to be a certain way is to assert that subsequent mental states are constructed as if that appearance was the case. Let me give a concrete example. Going back to our blind spot we can say that there is an appearance of the blind spot being filled in because consciousness proceeds as if it was filled in. We make judgments as if the spot was filled in, we talk about our experience as if the spot was filled in, we orient ourselves to the world as if the spot was filled in, we think about the contents of our visual field as if the spot was filled in, ect. And of course there are many ways in which this could come about, one of which being that the visual data really is filled in. But, as this description of what it means for something to appear a certain way in consciousness points out, we can’t jump from our knowledge of appearances to knowledge about the way things really are. A jump people seem ready to make in the case of qualia, when they assume that because it appears to us as if there are qualia then there must really be qualia, as they appear to us.