Page:Bergson - Matter and Memory (1911).djvu/181

 unexpressed: which amounts to saying that you abandon more and more the hypothesis that each verbal image goes up and fetches down its corresponding idea. In truth, there is here only a question of degree: every language, whether elaborated or crude, leaves many more things to be understood than it is able to express. Essentially discontinuous, since it proceeds by juxtaposing words, speech can only indicate by a few guide-posts placed here and there the chief stages in the movement of thought. That is why I can indeed understand your speech if I start from a thought analogous to your own, and follow its windings by the aid of verbal images which are so many sign-posts that show me the way from time to time. But I shall never be able to understand it if I start from the verbal images themselves, because between two consecutive verbal images there is a gulf which no amount of concrete representations can ever fill. For images can never be anything but things, and thought is a movement.

It is vain, therefore, to treat memory-images and ideas as ready-made things, and then assign to them an abiding place in problematical centres. Nor is it of any avail to disguise the hypothesis under the cover of a language borrowed from anatomy and physiology; it is nothing but the association theory of mind; it has nothing in its favour but the constant tendency of discursive