Humianism, so far as that is necessary to our argument, may best be dealt with in the modernized version of Mill. But before doing so, it will be instructive to trace the very similar process of criticism by which the realistic elements were eliminated from the original theory of Kant, and we shall see how their elimination leads to similar sceptical results.
It is important to observe that Kant's starting-point is a hypothetical dualism in many respects similar to that of Locke. Our knowledge refers to things which are other than our knowledge and may be said, in that sense, to lie beyond it. This further reference (which we have some reason to believe essential to the very nature of knowledge) Kant certainly starts with; and whatever results his theory leads him to as regards the kind of knowledge we have of things, he never loses hold of what he calls the thing-in-itself as that which alone gives meaning to the cognitive effort. Our knowledge of things may be imperfect and colored by the infusion of subjective elements, but if there were no 'things-in-themselves,' the whole process of knowledge would be a completely unmotived excursion into the void. Hence, as Kant puts it in the Preface to the Second Edition of the Critique, with his whole system explicitly in view, "while we surrender the power of cognizing, we still reserve the power of thinking objects as things-in-themselves. For otherwise we should require to affirm the existence of an appearance without anything that appears which would be absurd." In other words, our cognitions may be Erscheinungen, merely phenomenal, but as phenomena – as cognitions – they imply real objects, of which they are the cognitions. It is, of course, the peculiarity of the Kantian scheme, that our knowledge is so organized as to defeat its own purpose and cut us off from a knowledge of things as they really are. So far as our knowledge of it is concerned, the thing-in-itself shrinks, therefore, for Kant into a mere unknown somewhat; but in that capacity it remains as the necessary presupposition of the knowing process.
