Page:Bergson - Matter and Memory (1911).djvu/306

 have nothing in common between them, nor with extensity.

Idealism and realism, then, only differ in that the first relegates extensity to tactile perception, of which it becomes the exclusive property, while the second thrusts extensity yet further back, outside of all perception. But the two doctrines are agreed in maintaining the discontinuity of the different orders of sensible qualities, and also the abrupt transition from that which is purely extended to that which is not extended at all. Now the principal difficulties which they both encounter in the theory of perception arise from this common postulate.

For suppose, to begin with, as Berkeley did, that all perception of extensity is to be referred to the sense of touch. We may, indeed, if you will have it so, deny extension to the data of hearing, smell and taste; but we must at least explain the genesis of a visual space that corresponds to tactile space. It is alleged, indeed, that sight ends by becoming symbolic of touch, and that there is nothing more in the visual perception of the order of things in space than a suggestion of tactile perception. But we fail to understand how the visual perception of relief, for instance, a perception which makes upon us an impress sui generis, and indeed indescribable, could ever be one with the mere remembrance of a sensation of touch. The association of a memory with a present perception may complicate