The keystone of Kant's philosophy, sometimes called critical philosophy, is contained in his Critique of Pure Reason (1781), in which he examined the bases of human knowledge and created an individual epistemology. Like earlier philosophers, Kant differentiated modes of thinking into analytic and synthetic propositions. An analytic proposition is one in which the predicate is contained in the subject, as in the statement "birds have wings". The truth of this type of proposition is evident, because to state the reverse would be to make the proposition self-contradictory. Such propositions are called analytic because truth is discovered by the analysis of the subject itself, in this case, of the concept "bird". Synthetic propositions, on the other hand, are those that cannot be arrived at by pure analysis, as in the statement "that bird is blue". All the common propositions that result from experience of the world are synthetic.

Propositions, according to Kant, can also be divided into two other types: empirical, or a posteriori, and a priori. Empirical propositions can be known only through sense perception, but a priori propositions can be known without the use of such perception. The difference between these two types of proposition may be illustrated by the empirical "that bird is blue" and the a priori "two plus two makes four". Before Kant, it was assumed that only the a priori judgements were analytic ones. Kant's thesis in the Critique, however, is that it is possible to make synthetic a priori judgements.

In describing how this type of judgement is possible, Kant regarded the objects of the world as fundamentally unknowable; from the point of view of reason, they serve merely as the raw material from which sensations are formed. Objects in themselves are unknowable, and to think of them as situated in time and space is to apply the ideas of time and space, which are in fact part of the mind, to raw perceptions in order to organize them and make them intelligible.

Kant stated that in addition to the ideas of time and space, which he called "pure forms of intuition", the mind necessarily contains a number of fundamental concepts, which he called "categories". He divided the categories into four groups: those concerning quantity, which are unity, plurality, and totality; those concerning quality, which are reality, negation, and limitation; those concerning relation, which are substance-and-accident, cause-and-effect, and reciprocity; and those concerning modality, which are possibility, existence, and necessity.

We apply the pure forms of intuition and our concepts to our "intuitions", or raw perceptions, in order to make sense of them and to make a posteriori judgements about the world. This is how people gain most of their knowledge. However, we can also reflect on the fact that we must apply these forms of intuition and categories in order for us to have any intelligible experience at all, because these forms of intuition and categories are intrinsic to the self-conscious mind. In the part of the Critique called the "transcendental deduction", Kant attempts to prove this for the categories. It follows that any intelligible experience will be organized through the forms of intuition and categories, and this fact can be expressed in a series of synthetic a priori judgements that apply to the world of experience, the only world we can know. For example, the fact that we have to organize all our experience through the category of "cause and effect" can be expressed in the synthetic a priori judgement that "every event has a cause". This judgement is a priori because we can know it simply by reflecting on the fact that the category of cause and effect is essential to intelligible experience as such. We do not have to check whether in our experience every effect does indeed have a cause. However, it is synthetic because the concept "event" does not contain the concept of "caused" in the way that the concept of "bird" contains the concept of "winged". Such synthetic a priori judgements form the fundamental principles of science.

Therefore, for Kant the forms of intuition and the categories belong to the mind, and they are applied by the mind to our raw perceptions in order to gain knowledge. Accordingly, our conceptual knowledge can only be of the world as it is "for us", that is, as we experience it. Kant calls this view "transcendental idealism". The opposite view would be that time, space, and the categories are inherent in the structure of the world as it is "in itself", that is, as it is independently of its being experienced by people. It would then be possible to gain knowledge that goes beyond experience by applying these categories to the world as it is in itself. Kant regards this opposite view as the source of most of the errors of philosophy. In the section of the Critique called the "transcendental dialectic" he attempts to show that it leads to a series of pairs of contradictory propositions or "antinomies", both of which can be proved true. For example, it can be shown that people are both free and determined. The only way to resolve these contradictions is to adopt the standpoint of transcendental idealism, and to distinguish clearly between things as they are for us---or phenomena, of which we can have knowledge---and things as they are in themselves---as well as noumena, such as the human agent, which can be conceived of through reason but which cannot be known. Thus, the human being as a phenomenon is determined, but as a noumenon is free.