These are the sounds which are distinguished from one another by the most marked differences, which our organs most readily utter, and which are most universally found in human speech: all others are of later origin, having grown out of these in the course of the phonetic changes which words necessarily undergo, as they pass from one generation's keeping to another's. Our race has learned, as we may truly express it, by long ages of practice, of both mouth and ear, what the child now learns, by imitation and instruction, in a few months or years: namely, to add to its first easy utterances others more nicely differentiated, and produced by a greater effort of the organs. In like manner, starting from the mere rudiments of expression in radical monosyllables, the tribes of our family have acquired, through centuries and thousands of years of effort, the distinction and designation of innumerable shades of meaning, the recognition and representation of a rich variety of relations, in the later wealth of their inflective tongues—resources which, being once won, the child learns to wield dexterously even before he is full grown. It will be our next task to review the steps by which our language advanced out of its primitive monosyllabic stage, by which it acquired the character of inflective speech. To follow out the whole process in detail would be to construct in full the comparative grammar and history of the Indo-European dialects—a task vastly too great for us to grapple with here; we can only direct our attention to some of the principal and characteristic features of the development.
The first beginning of polysyllabism seems to have been made by compounding together roots of the two classes already described, pronominal and verbal. Thus were produced true forms, in which the indeterminate radical idea received a definite significance and application. The addition, for example, to the verbal root vak, 'speaking,' of pronominal elements mi, si, ti (these are the earliest historically traceable forms of the endings: they were probably yet earlier ma, sa, ta), in which ideas of the nearer and remoter relation, of the first, second, and third persons, were already distinguished, produced combinations vakmi, vaksi,
