Most of these concepts are what is called empirical. We get at them by observing likenesses and differences in what we perceive, and observing which are significant and important, and which are what we call accidental. Science, in its discovery of laws, is only carrying further this process which is implied in all simple judgments. By observing likenesses and differences, their uniformities and variations, and discovering those which are a key to the rest, we improve our concepts, and thereby have more knowledge of what we call natural laws, and can more and more anticipate experience. With these empirical concepts and their development Kant is not concerned. But there are certain concepts of which Hume had observed that they are not obtained in the ordinary way from an examination of the contents of experience. The two with which he chiefly concerned himself were substance and cause. These concepts seems to play an especially important part in the ordering and arranging of the concepts of experience. For the work of science, in moving from a simple observation of likenesses and differences to a knowledge of empirical laws, depends upon certain assumptions or principles, like the principle of causation or the principle of the conservation of energy. These principles imply concepts not derived, like the others, from generalisation from experience; they are the synthetic a priori judgments which, as we have seen, constituted a special problem for Kant. Kant is first concerned to ask where these a priori concepts come from, and how many of them there are. This inquiry he calls the metaphysical deduction of the categories. Having answered that question, he then goes on to ask by what right we assume these principles in our dealing with experience. This, the most important and difficult section of the _Critique_, he calls the transcendental deduction of the categories.