The difference is this: Descartes and Locke had accounted for our knowledge of things by maintaining that things act upon us, and make an impression or sensation--that their action, so to speak, begets ideas. This is a very ancient doctrine as well as a very modern one; it is the doctrine that most men find reasonable even before they devote themselves to the study of philosophy. The totality of such impressions received from the external world, they are accustomed to regard as our _experience_ of external things; and they are inclined to think that any knowledge of external things not founded upon experience can hardly deserve the name of knowledge. Now, Hume, when he cast doubt upon the existence of external things, did not, as I have said above, divest himself of the suggestions of the word "impression." He insists strenuously that all our knowledge is founded upon experience; and he holds that no experience can give us knowledge that is necessary and universal. We know things as they are revealed to us in our experience; but who can guarantee that we may not have new experiences of a quite different kind, and which flatly contradict the notions which we have so far attained of what is possible and impossible, true and untrue. It is here that Kant takes issue with Hume. A survey of our knowledge makes clear, he thinks, that we are in the possession of a great deal of information that is not of the unsatisfactory kind that, according to Hume, all our knowledge of things must be. There, for example, are all the truths of mathematics. When we enunciate a truth regarding the relations of the lines and angles of a triangle, we are not merely unfolding in the predicate of our proposition what was implicitly contained in the subject. There are propositions that do no more than this; they are _analytical_, _i.e._ they merely analyze the subject. Thus, when we say: Man is a rational animal, we may merely be defining the word "man"--unpacking it, so to speak. But a _synthetic_ judgment is one in which the predicate is not contained in the subject; it adds to one's information. The mathematical truths are of this character. So also is the truth that everything that happens must have a cause. Do we connect things with one another in this way merely because we have had _experience_ that they are thus connected? Is it because they are _given_ to us connected in this way? That cannot be the case, Kant argues, for what is taken up as mere experienced act cannot be known as universally and necessarily true. We perceive that these things _must_ be so connected. How shall we explain this necessity?