Kant thought of knowledge as a process of extending the bounds of perception, of piecing together the fragmentary glimpses we get of the world, stretching them out in spatial and temporal determinations that go beyond what we have actually experienced, connecting and linking up the events which we perceive discontinuously. As science extends, the range of our knowledge widens, but the process of extension never reaches its completion. There are always more facts to be discovered and explained. Science, therefore, can never rest content with its achievements, but must always demand that the investigation of conditions should be pushed further back and on. From this sense of the incompleteness of all actual knowledge, and of all there is that might be but is not known, arises what Kant calls an ideal of reason, a demand that, in all investigation into the conditioned, we should go on till we come to the totality of conditions. This ideal he holds to be serviceable and necessary. It has, however, a natural tendency to pass from an ideal to an idea, and in so doing it gives rise to the contradictions with which the _Dialectic_ is concerned. If all our investigation is governed by the thought that it must go on until it reaches completion, we naturally speculate on the fulfilment of that ideal, and try to form an idea of that totality of conditions, of how we should think the world if we knew it in its completeness. Herein we hypostatize the ideal or make it an idea, and we fall into contradiction; for we cannot really know the whole without knowing all its parts. If we give up the slow and never-completed process of knowing one part after another, and try to jump to the idea of the whole, we reach quite contrary results, as we apply to the conception of the whole one or other of two assumptions implied in our investigation of the parts. Kant sharply distinguishes between the principles of the pure understanding and the ideas of reason. The former are implied in all our knowledge, and the fact that experience is not chaotic confirms them at every moment. The second are ideals which guide knowledge, but are never realised. He calls them ideas of reason, because it is the special task of reason to lay down rules for the proper and complete working of the understanding. This task, he thinks, is exemplified in the logical nature of the syllogism which brings into unity the judgments of the understanding. As he used the forms of judgment as a guiding thread to discover a complete list of categories of the understanding, so he uses the forms of syllogism to discover a complete list of the ideas of reason. In both cases Kant’s reference to logical forms is far-fetched. Actually the list in the _Dialectic_ seems to be influenced by a number of considerations not always consistent.