Page:Bergson - Matter and Memory (1911).djvu/125

 certain service which the second memory can render to the first is to bring before it images of what preceded or followed situations similar to the present situation, so as to guide its choice: in this consists the association of ideas. There is no other case in which the memory which recalls is sure to obey the memory which repeats. Everywhere else, we prefer to construct a mechanism which allows us to sketch the image again, at need, because we are well aware that we cannot count upon its reappearance. These are the two extreme forms of memory in their pure state.

Now we may say at once that it is because philosophers have concerned themselves only with the intermediate and, so to speak, impure forms that they have misunderstood the true nature of memory. Instead of dissociating the two elements, memory-image and movement, in order to discover subsequently by what series of operations they come, having each abandoned some part of its original purity to fuse one with the other, they are apt to consider only the mixed phenomenon which results from their coalescence. This phenomenon, being mixed, presents on the one side the aspect of a motor habit, and on the other that of an image more or less consciously localized. But they will have it that the phenomenon is a simple one. So they must assume that the cerebral mechanism, whether of the brain or of the medulla oblongata or of the cord, which