It will be appropriate to notice here in a few words Hume's discussion of causality. According to Hume, the concepts of cause and effect have their origin solely in custom. We observe repeatedly that one event follows another and become accustomed to think of them as causally connected, so that we expect the second to occur as soon as we have observed the first. This theory, however, springs from a totally mistaken view of the causal relation. Suppose for several days running I observe the same person whenever I step out of the door of my house, I shall gradually form the habit of expecting the temporal sequence of the two events. But, it will never occur to me to think that there is any causal connection between my own appearance and that of the other person at the same spot. I shall call in aid essentially other items of the world-content in order to explain the coincidence of these events. In short, we determine the causal nexus of two events, not according to their temporal sequence, but according to the essential character of the items of the world-content which we call, respectively, cause and effect. From this purely formal activity of our thinking in the construction of the scientific picture of the world, it follows that the content of every cognition cannot be fixed a priori in advance of observation (in which thinking comes to grips with the Given), but must be derived completely and exhaustively from observation. In this sense, all our cognitions are empirical. Nor is it possible to see how it could be otherwise. For, Kant's judgments a priori are at bottom, not cognitions, but postulates. On Kant's principles, all we can ever say is only this, that if a thing is to become the object of possible experience, it must conform to these laws. They are, therefore, rules which the subject prescribes to all objects. But, we should rather expect cognitions of the Given to have their source, not in the constitution of the subject, but in that of the object. Thinking makes no a priori affirmations about the Given. But it creates the forms, on the basis of which the conformity of phenomena to law becomes manifest a posteriori. From our point of view, it is impossible to determine anything a priori about the degree of certainty belonging to a judgment which embodies knowledge thus gained. For, certainty, too, derives from nothing other than the Given. Perhaps it will be objected that observation never establishes anything except that a certain nexus of phenomena actually occurs, but not that it must occur, and will always occur, in like conditions. But, this suggestion, too, is in error. For any nexus which I apprehend between elements in the world-picture is, on our principles, nothing but what is grounded in these elements themselves. It is not imported into these elements by thinking, but belongs to them essentially, and must, therefore, necessarily exist whenever they themselves exist.