It is by means of the transcendental unity of apperception that all the manifold given in an intuition is united into a conception of the object. On this account it is called objective, and must be distinguished from the subjective unity of consciousness. ... The transcendental unity of apperception is alone objectively valid; the empirical possesses only subjective validity. (Critique, 80-81)OK, Roger, help me out:
It is important to understand this phrase, which contains in embryo much of Kant's philosophy. 'Apperception' is a term taken from Liebnizian metaphysics; it refers to any experience of which the subject is able to say "this is mine". In other words, 'apperception' means 'self-conscious experience'. The unity of apperception consists in the '"I think" which can accompany all my perceptions' [B.131-2], to borrow again Kant's version of Descartes. It consists of my immediate awareness that simultaneous experiences belong to me. I know immediately that this thought, and this perception, are equally mine, in the sense of belonging to the unity of consciousness that defines my point of view. Doubt here is impossible: I could never be in the position that Dickens in Hard Times attributes to Mrs Gradgrind on her deathbed, knowing that there is a pain in the room somewhere but not knowing that it is mine. This apprehension is called 'transcendental' because I could never derive it from experience. I could not argue that, because this pain has such a quality, and this thought such another, they must belong to a single consciousness. If I did that, I could make a mistake; I could be in the absurd position of ascribing to myself some pain, thought or perception that belonged, not to me, but to someone else. So the unity that I apprehend in my point of view is not a conclusion from experience, but a presupposition of experience. Its basis 'transcends' anything that experience could establish ... the transcendental unity of apperception provides the minimal description of our point of view. I can know one thing: that there is unity of consciousness. [Scruton, 43-44]Conceivably the reason why this isn't as well known as the Descartean cornerstone cogito it seeks to reform and reesetablish is because its much harder to grasp. Or conceivably the reason has to do with its wrongness ... a splendid, rather glorious wrongness, to be sure, that I can't be sure if Scruton is slyly acknowledging here -- I mean the way that, to illustrate the notion that my pain is mine and your pain yours and we two could never get confused on that point, he brings in a Dickensian character who contradicts exactly that point. I don't think he is trying to suggest Kant's wrongness, though, because the larger objection remains unmade (in fact, reading his book, I suspect he thinks Kant is right, here). Let not that larger objection remain unmade.
The larger objection, of course, is that human consciousness is not monadic in that crass sense that Leibniz's 'windowless' sometimes implies. This is so in large part because of that universal human quality -- empathy -- that erodes the capacity crisply to distinguish between my-pain-which-is-mine and your-pain-which-is-yours. Or to put it another way: Kant is saying we must have transcendental unity of apperception, since I think my thoughts and you think your thoughts; and that without it we might be in the situation of me thinking your thoughts and you thinking mine, which would be crazy. But this means that I refute Kant's transcendental unity of apperception simply by contemplating it, because by contemplating it, by running it through my mind, I am precisely having Kant's thoughts.