Page:Bergson - Matter and Memory (1911).djvu/237

 similarity to similar objects, embroidering upon the similarity, as on their common stuff or canvas, the variety of individual differences. And we go on also from the whole to the parts, by a process of decomposition the law of which will appear later, a process which consists in breaking up, for the greater convenience of practical life, the continuity of the real. Association, then, is not the primary fact: dissociation is what we begin with, and the tendency of every memory to gather to itself others must be explained by the natural return of the mind to the undivided unity of perception.

But here we discover the radical vice of associationism. Given a present perception which forms by turns, with different recollections, several associations one after another, there are two ways, as we said, of conceiving the mechanism of this association. We may suppose that the perception remains identical with itself, a true psychical atom which gathers to itself others just as these happen to be passing by. This is the point of view of associationism. But there is also another,—precisely the one which we have indicated in our theory of recognition. We have supposed that our entire personality, with the totality of our recollections, is present, undivided within our actual perception. Then, if this perception evokes in turn different memories, it is not by a mechanical adjunction of more and more numerous