Hume, however, had noticed that the principle of causation, the judgment that every event has a cause, is both a priori and synthetic. It is not, he held, derived from experience; rather it is a principle which guides our investigation of experience. It is not got from analysis of the notion of causation, nor is it simply concerned with the agreement or disagreement of our ideas. It asserts the necessary connection of two perfectly separate existing things. Hume himself, as we saw, tried to explain away these uncomfortable facts. He was too wedded to his belief that all knowledge was derived from passively received impressions to face them rightly. Kant, coming to the problem with different prepossessions, with the belief that most knowledge was analytic, was impressed with Hume’s proof that the principle of causation could not be derived from analysis. The very basis of all science of nature, then, contradicted the belief that knowledge was analytical. Kant was also, with Hume, convinced that the principle of causation was not derived from experience, for he saw that experience assumed it. At the same time, he was not prepared, like Hume, to explain it away. Further, he saw that the problem raised by the principle of causation was a wide one. For other judgments, he held, are both synthetic and a priori, among them mathematical judgments. As we shall see afterwards, Kant proved the impossibility of arriving at knowledge of God or the soul by mere analysis of concepts. The judgments of metaphysics, about God or the soul, are also synthetic. But the validity of the judgments of metaphysics is under dispute. If we examine the synthetic a priori judgments of mathematics and of science whose validity is certain, we may then discover whether such judgments in metaphysics can or can not have similar certainty. We may thus see that the problem of the possibility of synthetic a priori judgments is a restatement in logical terms of the problem of the relation between the a priori sciences and metaphysics.