caused the development of the roots themselves into varying phonetic forms; and these have then been applied, at first to support, and afterwards to replace, the primitive means of grammatical expression: an internal flection has come in upon and supplanted the original aggregation. All Indo-European forms are originally of the kind here first illustrated, mere agglutinations of independent elements, whereof a part are reduced to a subordinate value and formal significance; but they tend, in a marked degree, to pass over into the other kind, indicating formal relations by internal change in the root or theme, instead of by external additions alone.
This tendency is generally regarded as constituting the highest characteristic of the Indo-European dialects, as making them properly inflective; and languages possessing in this sense an inflective character are reckoned to stand at the head of all the forms of human speech. Some, however, are inclined to claim a more original and fundamental importance for the process of internal change in the history of the tongues of our family, to regard a capacity of significant variation of vowel as inherent in their roots, and bearing a regular and conspicuous part in even the earliest steps of their development. The evidence upon which this claim is founded I cannot but regard as altogether insufficient to sustain it. Wherever, in the most ancient as well as the more modern processes of word-formation and inflection, we find internal changes of the root, they are, I am persuaded, of secondary growth, inorganic; they are called out ultimately by phonetic causes, not originated for the purpose of marking variation of meaning, though sometimes seized and applied to that purpose. To prove the element of internal flection one of prime value in the growth of Indo-European language, it would be necessary to show that the variation of vowel had a distinctly assignable office in the primitive production of words; that it regularly distinguished from one another certain parts of speech, certain classes of derivatives, certain forms of declension or conjugation; that it formed guiding analogies, which could be and actually were imitated continuously in the further processes of word-making. But this is far from being the case; on the con-
