knowledge is no longer so radically, so definitely relative as certain philosophers say, if we can establish that it bears on a reality of inverse order, a reality which we always express in mathematical laws, that is to say in relations that imply comparisons, but which lends itself to this work only because it is weighted with spatiality and consequently with geometry. Be that as it may, it is the confusion of two kinds of order that lies behind the relativism of the moderns, as it lay behind the dogmatism of the ancients. We have said enough to mark the origin of this confusion. It is due to the fact that the "vital" order, which is essentially creation, is manifested to us less in its essence than in some of its accidents, those which _imitate_ the physical and geometrical order; like it, they present to us repetitions that make generalization possible, and in that we have all that interests us. There is no doubt that life as a whole is an evolution, that is, an unceasing transformation. But life can progress only by means of the living, which are its depositaries. Innumerable living beings, almost alike, have to repeat each other in space and in time for the novelty they are working out to grow and mature. It is like a book that advances towards a new edition by going through thousands of reprints with thousands of copies. There is, however, this difference between the two cases, that the successive impressions are identical, as well as the simultaneous copies of the same impression, whereas representatives of one and the same species are never entirely the same, either in different points of space or at different moments of time. Heredity does not only transmit characters; it transmits also the impetus in virtue of which the characters are modified, and this impetus is vitality itself. That is why we say that the repetition which serves as the base of our generalizations is essential in the physical order, accidental in the vital order. The physical order is "automatic;" the vital order is, I will not say voluntary, but analogous to the order "willed." Now, as soon as we have clearly distinguished between the order that is "willed" and the order that is "automatic," the ambiguity that underlies the idea of _disorder_ is dissipated, and, with it, one of the principal difficulties of the problem of knowledge.