the fusion of earlier polysyllabic forms (like the Saxon lord from hlaford, or the English won't from will not). It is probable that the verbal deliquescence which we observe in later stages of inflected languages, dissolving out the pronouns and particles, tense and mood signs, that are imbedded in their grammatical forms, was equally active in melting down the combinations that made up the primitive words. Even such positively agglutinative structures as the Tartar and North American dialects mark distinctions between the syllables by emphasis, gesture, and tone, which are properly the beginning of disintegration. On the other hand, inflection also is but a finer and closer kind of agglutination, peculiar to certain race-qualities, and belongs probably to a much earlier stage than we are wont to suppose : nor is it necessarily preceded by a monosyllabic formation. Thus the disintegrative and constructive processes, so readily recognized in later stages of linguistic growth, probably go on side by side in very early ones. Man, the centre of polarities, at once destroyer and builder, is true to his functions even in these infantile motions of his thought. Mind and body are but sides of one and the same incessant interchange of waste and repair; the dual movement whence harmony issues, the Yin and Yang of progress.
Formation of "roots" by separation and fusion.
Such processes of separation and fusion, going on together, would obviously result in shortening words - from polysyllabic concretions into simple condensed forms, with breathing spaces, as it were, separation between them. It is probably in this way that many of the monosyllables were produced which we have agreed to call roots.
Doubts as to the usual division of languages.
Is there sufficient ground for dividing languages into a monosyllabic, an agglutinative, and an inflected stage? Do even the Aryan tongues afford the evidences of such a process? Humboldt has indeed
