make their speech also analytical. But if they were not themselves analytical, they could not have done so.
Now it is very certain, as certain as anything can well be, that all the non-Aryan languages of India are still in the agglutinative stage. If, then, they exercised any influence on the structure of the Aryan speech, such influence would tend to make that also agglutinative; in other words, the Aryans would have had to go backwards, and try and find out what were the agglutinated particles from which their own inflectional terminations had arisen; and having found them, would have been led to use them, no longer conveniently incorporated into their words, but disintegrated and separate. Thus, a vulgar Aryan who said homi, for "I am," would have had to re-construct out of his inner consciousness the older form bhavâmi, and, not content with that, to further resolve bhavâmi into its component elements of bhû and âmi, and henceforth to use these two words adjacent to each other, but unincorporated. This we see at once is out of the question, and absurd. What our vulgar Aryan really did was in course of time to drop the final i of homi, and to nasalize the m, at the same time imparting a broader and more rustic pronunciation to the vowel, thus producing हौं haun. He also changed hosi, "thou art," and hodi, "he is," both into hoï; and thus having got two words of similar sound, he had to use the pronouns tû and vah to distinguish them: which is precisely the opposite process to that which imitation of the Dravidians or Kols would have led him to follow, and precisely similar to that which his brother the vulgar Roman followed when he changed sum into sono, and sunt also into sono; so that, getting two sonos, he had no means of distinguishing between them except by constantly prefixing the pronouns io, "I," and eglino, "they"; and just that which the Englishman followed when he changed ga, gœth, and gath, all three into go, and then had always to prefix I, he, we, ye, they, to make his meaning clear.
