selecting those which he shall intimate, by such signs as he can make intelligible: there is no other way in which we can make a definition or description, whether for our own use or for that of anybody else. If, for example, a dog is the subject of our effort, we compare our conception of him with those of other sensible objects, and note its specific differences—as his animality, shape, size, disposition, voice. This is so essentially a human procedure that we cannot conceive of the first makers of language as following any other. Then, in finding a designation, it would be impossible to include and body forth together the sum of observed qualities: in the first instance, not less than in all after time, some one among them would necessarily be made the ground of appellation. The sign produced would naturally vary with the instrumentality used to produce it, and the sense to which it was addressed: in the instance which we have supposed, if the means of communication were writing, it would probably be the outline figure of a dog; if gesture, an imitation of some characteristic visible act, like biting, or wagging the tail; if the voice, not less evidently an imitation of the audible act of barking: the dog's primal designation would be bow-wow, or something equivalent to it. But in this designation would be directly intimated the act; the actor would be suggested by implication merely: bow-wow, as name for 'dog,' would literally mean 'the animal that bow-wows.' So in the case of a word like splash, used to imitate and call up before the mind the fall of a stone into water—the collision of the stone and the water would be the immediate suggestion; but a natural act of association might make the sign mean the stone, or the water, or the act of throwing, or the fall. One sign would turn more readily to the designation of a property or action, another to that of a concrete thing, an actor, according to the nature of each, and the exigencies of practical use as regarded it; but both would be inherently a kind of indifferent middle, capable of conversion to either purpose: and, in the poverty of expression and indistinctness of analysis belonging to the primitive stage of linguistic growth, would doubtless bear various offices at
