Page:Bergson - Matter and Memory (1911).djvu/140

 perception by a motor tendency to sketch its diagram. Whence we may conclude that such is indeed the primordial condition of recognition.

But we must pass now from automatic recognition, which is mainly achieved through movements, to that which requires the regular intervention of memory-images. The first is recognition by inattention; the second, as we shall see, is attentive recognition.

This form also begins by movements. But, whereas, in automatic recognition, our movements prolong our perception in order to draw from it useful effects and thus take us away from the object perceived, here, on the contrary, they bring us back to the object, to dwell upon its outlines. Thus is explained the preponderant, and no longer merely accessory, part taken here by memory-images. For if we suppose that the movements forego their practical end, and that motor activity, instead of continuing perception by useful reactions, turns back to mark out its more striking features, then the images which are analogous to the present perception,—images of which these movements have already sketched out, so to speak, the form,—will come regularly, and no longer accidentally, to flow into this mould, though they may have to give up much of their detail in order to get in more easily.

III.—Gradual passage of recollections into move-