Page:Bergson - Matter and Memory (1911).djvu/307

 this perception by enriching it with an element already known, but it cannot create a new kind of impress, a new quality of perception: now the visual perception of relief presents an absolutely original character. It may be urged that it is possible to give the illusion of relief with a plane surface. This only proves that a surface, on which the play of light and shadow on an object in relief is more or less well imitated, is enough to remind us of relief; but how could we be reminded of relief if relief had not been, at first, actually perceived? We have already said, but we cannot repeat too often, that our theories of perception are entirely vitiated by the idea that if a certain arrangement produces, at a given moment, the illusion of a certain perception, it must always have been able to produce the perception itself;—as if the very function of memory were not to make the complexity of the effect survive the simplification of the cause! Again, it may be urged that the retina itself is a plane surface, and that if we perceive by sight something that is extended, it can only be the image on the retina. But is it not true, as we have shown at the beginning of this book, that in the visual perception of an object the brain, nerves, retina and the object itself form a connected whole, a continuous process in which the image on the retina is only an episode? By what right, then, do we isolate this image to sum up in it the whole of percep-