Hume's skepticism arises when he asks how we perceive causal connections between events. Reason alone cannot tell us about connections between things in the world, and experience alone cannot infer universal generalizations such as "every event has a cause." Hume concludes that in fact we have no rationally justified knowledge of cause and effect. He suggests instead that our concept of causation is justified only by the habit of seeing certain events follow from certain other events.
Kant agrees that we cannot discover the concept of cause and effect either in experience or by means of reason. However, he does not conclude with Hume that this concept is merely a result of habit or custom.
Rather, he suggests, causation is an a priori concept of the understanding applied to appearances. We can know nothing about things in themselves; we can know only how they appear to us in the form given to them by our faculties of sensibility and understanding. The concept of cause and effect is not to be found in these appearances; rather, it is part of the form given to them by the understanding. Causation is not a "thing" that we can discover, either by means of reason or experience. Causation is a form given to experience that makes it intelligible to us. Hume asked how we can derive pure concepts (such as causation) from experience, and answered that we cannot. Kant agrees: we cannot derive pure concepts from experience; rather, we derive experience from these pure concepts.
Pure concepts of the understanding make experience legible, so to speak, but cannot tell us anything about things in themselves. Because pure concepts, as well as our pure intuitions of space and time, are a priori and therefore necessary, we are tempted to think that they can give us knowledge beyond that which we find in experience. However, our pure concepts and pure intuitions provide only form, and no content. They help us make connections between appearances, and as such, they deal only on the level of appearances. They cannot tell us anything about the things in themselves behind these appearances.
Nature, understood as the totality of all our sensations, is possible—as we saw in the first part—by means of our pure intuitions of space and time. Nature, understood as the totality of experience as understood and connected by laws, is possible—as we saw in this part— by means of our pure concepts of the understanding. We cannot go farther and ask how the faculties that give us our pure intuitions and pure concepts are possible, because it is precisely these faculties that help us make sense of experience. We have no further faculty that would help us understand what is behind these faculties.
Sensations themselves teach us nothing about the connections between them or the laws that govern them: these are all provided by our faculties of sensibility and understanding. These faculties, then, are what make nature itself possible, insofar as nature is our intelligible experience. Whatever laws or universality we find in experience comes not from the sensations themselves but from the form given to them by our faculties. Thus, Kant concludes: "the understanding does not derive its laws (a priori) from, but prescribes them to, nature."
Kant wraps up this part with an example showing how we derive astronomical principles from our pure intuition and concepts rather than from experience itself, and with an appendix discussing his system of categories.
Early modern philosophy, from Descartes to Hume, is roughly divided between rationalists—like Descartes, Spinoza, and Leibniz—and empiricists—like Locke, Berkeley, and Hume. Rationalists place a strong emphasis on pure reason and metaphysics, suggesting that the unaided intellect can discover metaphysical truths about the nature of life and the universe. Empiricists place a strong emphasis on knowledge gained from experience, confining the intellect to reasoning about experience based on what it finds in experience. Kant attempts a synthesis between these two camps by showing that they are both flawed in a fundamental respect. The empiricists are wrong to think of experience as consisting only of sensations, and the rationalists are wrong to think that the intellect can give us any insight into the essence or nature of things in themselves. Kant attacks both empiricists and rationalists in this part of the text.
His main object of criticism in the empiricist camp is Hume, who Kant nonetheless deeply admires—after all, Kant's system was motivated by Hume's skepticism. Hume's mistake, according to Kant, lies in misunderstanding the nature of experience. Like pretty much everyone before Kant, Hume believes that experience is fundamentally a set of simple impressions. Ideas and complex impressions are built by connecting various simple impressions. Experience for Hume is what Kant would call "sensations": it is the simple sense-data of what we see, hear, smell, etc.
Kant criticizes this view by pointing out that experience is more than just simple impressions that we receive as neutral observers. Experience comes in an already organized form. Everything that we can experience happens in space and time. Space and time, Kant argues, are not impressions or sensations: we do not learn about them through experience because they are pure intuitions. Further, everything that happens in space and time is subject to the law of cause and effect. This law is also not an impression or a sensation but a pure concept of the understanding. Pure intuitions and concepts organize experience for us and give it its form. We are not neutral observers of an objective world; we actively shape the world we perceive so as to make it intelligible.
Hume argues that we can have no knowledge of cause and effect because we cannot find causes or effects in the simple impressions that constitute the entirety of our experience. Kant agrees that there are no causes and effects to be found in experience, but suggests that our understanding applies the concept of causation to experience in order to make it intelligible. To modify Bertrand Russell's analogy, we see the world through causation-colored sunglasses.
If we can dismiss Hume's skeptical empiricism by pointing out that we can have a priori knowledge of cause and effect, space and time, we encounter an equally unsavory dogmatism in the rationalist camp. If synthetic a priori knowledge is possible, then we can learn substantial truths about the universe that have a necessity and universality that we cannot find in experience. While experience can only tell us about appearances, about how things seem to us through the imperfect medium of the senses, reason can tell us about things in themselves.
Kant's answer to this temptation is twofold. First, our a priori pure intuitions and pure concepts only help us to make sense of appearances. Kant makes no claim that the laws of cause and effect apply to things in themselves, only that we apply them with necessity to appearances. Second, our pure intuitions and pure concepts do not give us any substantial knowledge. All they do is prescribe the form our experience takes. They are the cup into which experience is poured, so to speak: we cannot drink the cup, but without the cup we wouldn't be able to drink at all.
Kant's pure intuitions and pure concepts reconcile the empiricist and rationalist camps by showing that, on one hand, we can have a priori knowledge relating to experience and on the other hand, this a priori knowledge does not tell us anything about things in themselves. Kant recognizes that we are not passive recipients of our sensory experience. What we perceive is given its form by faculties that are innate. These faculties do not tell us anything about what the world is really like, but they determine the patterns according to which the world appears to us.