After having a nice little chat about this with Tobias and Isaiah elsewhere, I feel compelled to start a thread about Kant's Transcendental Unity of Apperception. Here it is, then.

I will give a brief introduction to the basic tool of Kantian philosophy. Those that are familiar with it can skip the introduction without loss.

The latin verb transcendere means "to surpass", to go or be beyond something. Kant distinguishes sharply between the transcendent (that which is beyond something) and the transcendental. The latter is very important in Kantian philosophy: the transcendental is something which lies in the border of something - in Kantian philosophy, basically in the limit of our cognitive capacities.

So if x is transcendent, it lies beyond a fence. If x is immanent (a term that Kant is rarely if ever interested in), then it is on this side of the fence. If x is transcendental, then it is on the fence - perhaps a property of the fence in question. The idea is, thus, actually rather simple, if one can get over the horrid word wink.

Kant is interested in the necessary conditions of knowledge. He is not asking whether knowledge is possible anymore (unlike most of his predecessors), but under what circumstances knowledge would be or is possible in the first place. His main idea is to refute skeptical idealism (that of Hume's) by starting out from a premise they accept: we have knowledge of our own mental states, and trying to prove that this is only possible if we can have knowledge of external objects as well - that is, the necessary condition of subjective experience is experience of objects.

Kenneth R. Westphal put it aptly enough to quote in his book Kant's Transcendental Proof of Realism: Kant is identifying some of our key cognitive capacities by identifying some of our key cognitive incapacities.

He is saying that Kant is trying to find out what are the necessary conditions of our knowledge by finding out what we cannot comprehend. This reflection he calls transcendental reflection, and it simply means that he is reflecting on his own mental borders, trying to find out what the fence is like. Or, as I like to analogize it: we are all in a box according to Kant, and he is trying to find out what the walls of the box look like.

Kant's philosophy is extremely complex, but the basic idea is simple. We experience, and we experience something. Kant calls this something that we experience the thing in itself. He then uses transcendental reflection to show that there are some necessary conditions for this thing in itself to be represented by us. He notes that we cannot imagine something that is not in space and in time. In one of his arguments he uses transcendental reflection in a very enlightening way: take an object. A chair, for instance. Then start ripping it of its properties. You can take a lot of properties away from it, but when you have, you end up with some two things that you cannot strip it of: the space it takes and the time it exists in. Therefore: space and time are necessary conditions for something to be presented in our experience.

But that is not all: there are the Categories. First let me note that Kant divided the cognition into three pieces, of which two are of immediate concern: sensibility that receives intuitions (basically: perceptions) and understanding that arranges them into concepts. Space and time are the forms of sensibility, that is, the coordinates into which every sensible perception must conform to in order for it to be able to present itself to us. These could be compared to the borders of a vase, into which things are dropped into: any object that enters the vase must fit into the limitations the vase itself sets for it.

The Categories of understanding are then the necessary conditions for something to be conceived. This is where our interests lie, at the moment. Those Categories are numerous, but the most important ones of those are, for example, that of causality (the ability to set consecutive experiences into some sort of relation with each other) and that of quantity - the ability to numerically distinguish objects from eact other.

Kant claims that these Categories and the forms of sensibility (space and time, remember?) our cognition imposes on the objects we perceive. That is, they are a filter through which every object of experience must pass through. The thing in itself is then the object considered as independent of these conditions (as it is when it is not experienced), and the appearance or the phenomenon is the object as considered within our cognition, as subsumed under the conditions set for it by us.

If Kant can prove his system, he would then refute skepticism. The initial proof for these Categories of understanding (he thought that the Categories are hierarchically higher, so the forms of sensibility too are subsumed under the Categories) he called the Transcendental Deduction (of the Categories). It is within the steps of this proof, that he himself called "the most difficult task ever undertaken in the service of metaphysics" and that took him 10 years to complete, there is the Transcendental Unity of Apperception, to which we now turn. For it is one of the most important single ideas in the history of philosophy - regardless of how the Transcendental Deduction as a whole fares.

Kant embarks on a mission to find out what it would take for one to experience. The first step of the Transcendental Deduction (and the only step I will concern myself with here) concerns the necessary conditions for someone to even experience something. This is why it is so important regardless of the Transcendental Deduction in general, because it basically sets down the foundations of consciousness in a single philosophical idea.

The Principle of Transcendental Unity of Apperception is then the principle which alone gives rise to consciousness. It is something that is actual to the philosophy of mind even today, and something that has been in various forms of a host of philosophers since Kant - including, to name a few, Hegel, Husserl and Wittgenstein.

But before we can get into the actual principle, we must consider the idea of transcendental subject (something that too is present in the philosophy of many since Kant).

In the introduction I presented Kant's idea of understanding as something that arranges experiences and conceives them. One important thing to understand about understanding is that it is active. The sensibility mainly receives passively, whereas the understanding actively organizes the intuitions received by the sensibility. Kant's idea is simple: each and every perception is always singular, particular. In order for there to be any wholes at all, the mind must organize the singular perceptions into conceivable wholes. From this Kant then concludes (through lengthy additions, as always) that the basic way for a subject to be is to act. The subject is not merely active, but subjectivity IS activity.

The transcendental subject is the necessary condition of any subjective experiences. But the transcendental subject is not some material being, but in fact it is pure activity. What the subject brings to the matter given by the object is the activity through arranging.

It could be seen this way, as well: the necessary condition for any cognition is that there is some object that is cognized, and some subject that does the cognizing. Whereas the thing in itself is pure objectivity, the transcendental subject is pure subjectivity. The transcendental subject is subjectivity stripped of all the influence of objects. And for Kant, the transcendental subject is pure activity without any objects of activity.

It is important to understand that the transcendental subject is a theoretical limit (hence the word "transcendental") of subjectivity, and as such is very different from the empirical subject (i.e. you and me). It is beyond our cognition in the same sense as the thing in itself is - even though they are polar opposites.

But what is this activity of the transcendental subject?

According to Kant, the necessary condition for us to have any (conscious) experiences at all is that we are able to recognize the experiences as ours. That is, for an experience to be an experience for us, we must be able to say "I think" of that experience. (Note that "I think" is pretty much equivalent to Descartes' Cogito, but given a new interpretation in Kantian framework). This ability to reflect upon our experiences and saying "I think" of them he calls apperception. Apperception is a faculty of understanding, a sort of subset of it. It is one way the understanding functions, and it is the first step in the Transcendental Deduction. Of that, no more.

This is very important. Basically Kant here invents something totally new: subjectivity is a process, not a substance. The Cartesian substantial mind is substituted by Kantian processual mind, and in this he is way before his time (after all, it was only after the mid 20th century that even crude functionalist accounts of mind begun to emerge - and Kant's idea is actually far better than those).

But this is not the crux of the matter, yet. Because if we stopped here, we would, according to Kant, have at best some flickering moments of consciousness, where the consciousness was always the different consciousness. That is, even though "I think" would accompany two of experiences, there was no telling whether these "I thinks" were somehow related. So, Kant ponders, what do we still need to have a consciousness (which we clearly have!).

And so, he stumbles upon the synthetic nature of apperception. Not only must we be able to say "I think" of every experience, we must be able to say "I think" in the same consciousness of every experience belonging to that consciousness. That is, we must be able to recognize the different "I thinks" as belonging to the same subject. Therefore, he says that we must have synthetic unity of apperception, that is, the ability to take particular "I thinks" under one unity of subjectivity through the synthetic activity of the understanding. And because this is a necessary condition for consciousness, Kant has produces a new term:

The Transcendental (Synthetic) Unity of Apperception.

This principle (that we must accept the transcendental unity of apperception) is the foundation of (empirical) consciousness for Kant. Simply: the ability to recognize particular experiences as belonging to a single subject: us.

It means a lot, of course, but generally. Well, as said, Kant turns from substantial conception of mind to processual (we are still waiting for some philosophers of mind to enter the 19th century with Kant). The consciousness is a constant process of self-recognition. This has actually received a lot of scientific evidence lately, because it seems that memory plays a crucial role in consciousness and personality: the continuity of self-recognition is the basis of consciousness. This is what is Kant's message here.

What is this activity? What recognizes? The understanding, of course, as said before, but that is not enough. It is the transcendental subject, the pure activity of subjectivity that is the act of recognition. Kant is not, notoriously so, very clear at this point. For this raises the question: what is this pure activity? Shouldn't it be based on something - something that acts? The ways to answer this are many, and Fichte for one said that one must simply either start from the subject or the object, and himself started from the subject.

Moreover, Kant is not really that interested in that question. He assumes consciousness, and only tries to chart the boundaries of consciousness. As he ends up saying that nothing can be said of the transcendental subject beyond its effects that we see as its activity (much like with the thing in itself), then he must, in order to be coherent, leave the question as to the nature of the transcendental subject ultimately unanswered. To him, the question is about something about something transcendent, something beyond the limits of our cognition, and thus, basically, meaningless.

So, there you go. There is no easy way to explain the Transcendental Unity of Apperception. Or, there is: the recognition of experiences as one's own within the same subject. But the problem with such explanations is that they are ultimately rather vague. The problems arise when one tries to plunge deeper into the concepts, to see the system behind it.

One must bear in mind always that the single most important point to understand about Kant is his transcendental method. He is NOT trying to prove that we have consciousness (you can disagree, if you will, but somehow I doubt it will be anything but half-hearted playing of a devil's advocate), nor is he trying to prove that we exist. He is assuming that we have a consciousness, that we exist and experience, and then trying to show that there are some pre-requirements for those. If one can prove that the necessary condition of x is y, and that x, then he can prove that y.

His transcendental method is also the reason behind his Copernican Revolution, i. e. his turn from ontology to epistemology (and metaphysics has never been the same since). Through this transcendental method he was able to recognize the active role our mind has, and to somehow get his mind to grasp something as modern as subjectivity as a continuous process of self-identification. It doesn't much matter whether the rest of his Transcendental Deduction succeeds or not: this is a feat worthy of praise in itself.