Page:Bergson - Matter and Memory (1911).djvu/305

 nothing on which to base the parallelism of their phenomena. It is therefore constrained to account for this parallelism by a habit which makes the actual perceptions of sight, for instance, suggest to us potential sensations of touch. If the impressions of two different senses resemble each other no more than the words of two languages, we shall seek in vain to deduce the data of the one from the data of the other. They have no common element; and consequently, there is nothing common between extensity, which is always tactile, and the data of the senses other than that of touch, which must then be supposed to be in no way extended. But neither can atomistic realism, which locates movements in space and sensations in consciousness, discover anything in common between the modifications or phenomena of extensity and the sensations which correspond to them. Sensations are supposed to issue from the modifications as a kind of phosphorescence, or, again, to translate into the language of the soul the manifestations of matter; but in neither case do they reflect, we are told, the image of their causes. No doubt they may all be traced to a common origin, which is movement in space; but, just because they develop outside of space, they must forego, qua sensations, the kinship which binds their causes together. In breaking with space they break also their connexion with each other; they