But how vastly different must have been the case of primitive man! The infant, as a child of to-day, finds a grammar already made to its use, and one which it is bound to learn with the first learning of denotative names. But the infant, as an adult in primeval time, was under the necessity of slowly elaborating his grammar together with his denotative names; and this, as we have previously seen, he only could do by the aid of gesture and grimace. Therefore, while the acquisition of names and forms of speech by infantile man must have been thus in chief part dependent on gesture and grimace, the acquisition by the infantile child is now not only independent of gesture and grimace, but actively inimical to both. The already-constructed grammar of speech is the evolutionary substitute of gesture, from which it originally arose; and, hence, so soon as a child of to-day begins to speak, gesture-signs begin at once to be starved out by grammatical forms. But in the history of the race gesture-signs were the nursing-mothers of grammatical forms; and the more that their progeny grew, the greater must have been the variety of functions which the parents were called upon to perform. In other words, during the infancy of our race the growth of articulate language must not only have depended, but also reacted upon that of gesture-signs—increasing their number, their intricacy, and their refinement, up to the time when grammatical forms were sufficiently far evolved to admit of the gesture-signs becoming gradually dispensed with. Then, of course, Saturn-like, gesticulation was devoured by its own offspring; the relations between signs appealing to the eye and to the ear became gradually reversed; and, as is now the case with every growing child, the language of formal utterance sapped the life of its more informal progenitor. We are now in a position to consider the exact psychological relation of sentence-words to denotative and receptually connotative words. It will be remembered that I have everywhere spoken of sentence-words as representing an even more primitive order of ideation than denotative words, and, _a fortiori_, than receptually connotative words. On the other hand, in earlier parts of this treatise I showed that both the last-mentioned kinds of words occur in children when they first begin to speak, and may even be traced so low down in the psychological scale as the talking birds. This apparent ambiguity, therefore, now requires to be cleared up. Can anything, it may be reasonably asked, in the shape of spoken language be more primitive than the very first words which are spoken by a child, or even by a parrot? But, if not, how can I agree with those philologists who conclude that there is an even still more primitive stage of conceptual evolution to be recognized in sentence-words?