istic property of geometrical space, even though it be incompatible with reality, even though it does not even admit of being conceived as realized."
The concept of repetition unites the two antithetical concepts of permanence and change. Hence, one may direct one's attention to the identical elements in the consecutive phenomena, or, on the contrary, attend to differences and modifications. The identity of phenomena is only an hypothetical concept. One may postulate the identity of two phenomena, and admit that the same phenomenon may occur twice; this we shall call 'integral repetition.' So far as stress is laid upon the diversity of phenomena, which cannot be absolute, but rather implies a certain unity, one may form an antinomical concept, suppose a combination of identical and non-identical elements, and thus form the notion of a phenomenon repeated with alteration, i.e., with progressive modification of the phenomenon in the course of repetition. We shall designate this concept and the principle involved as the concept and the principle of 'altering repetition.' The object of the present paper is to show the relations which unite the concepts of repetition, defined as above, with the concept of time, and to make it appear that, whichever point of view one adopts, one is led to conceive of time under one of two essentially different aspects. (1) 'Integral repetition.' The science of the 'phenomenon-in-itself,' i.e., of external nature, depends for the most part on this principle: phenomena admit of repetition without alteration. While necessarily an assumption to start with, this principle finds itself justified by the steady growth of the physical sciences. As regards the concept of time which corresponds, it is evident that a phenomenon capable of integral repetition bears to duration only the relation of content to that which contains. Time is inert with regard to it, is a void and homogeneous medium. That which distinguishes time as medium from space is only the fact that it has but a single dimension. This concept of time is no more an illusion than the concept of integral repetition. They are both necessary conditions of the possibility of the science of 'phenomena-in-themselves.' (2) 'Altering repetition.' We have hitherto spoken of a single mode of existence, of objective existence, or existence-in-itself. But there are phenomena which exist not only in themselves,
