Should it still be said, "Do not the idiot and the brute often act as if they were determined by the knowledge of things past? Do they not distinguish persons with whom they have lived; places where they have resided; and perform many other actions, which necessarily imply memory? And does not all this prove that memory proceeds not from the power of reflection?" It must already have been perceived, that I distinguish two kinds of memory, infinitely different in their causes, though somewhat similar in their effects. The one consists in the impressions of our ideas; and the other, which I would rather term reminiscence than memory, is nothing more than the renovation of our sensations, or of the vibrations by which they were occasioned. The former issues from the mind, and is much more perfect in man than the latter; which is produced merely by the renovation of the vibrations of the internal sense, and is the only memory possessed by brutes or idiots. Their preceding sensations are renewed by their present ones; the present, and principal, calls forth the former, and the accessory images; they feel as they have felt, and therefore they act as they have acted; they behold together the present and the past, but without distinguishing or comparing, and consequently without knowing them. As another proof of the existence of memory in animals, I may be told of their dreams. It is certain that brutes, while asleep, have the things represented to them with which they have been occupied while awake. Dogs bark when they are asleep; and though this barking is feeble, yet it is easy to distinguish in it the cry of the chace, accents of rage, sounds of desire, of murmur, &c. It is not to be doubted, then, but that dogs have a lively and active memory, different too from that of which we have now been speaking, since it acts independent of any exterior cause. To clear up this difficulty, it is necessary to examine the nature of dreams, and to inquire whether they proceed from the mind, or depend entirely on our internal material sense. If we could prove that they reside solely in the latter, it would be an answer to the objection, and another demonstration, that in brutes there is neither understanding nor memory.