tried to compare it now with Semitic and now with Aryan. In this respect it resembles the Finnic idioms, where agglutination has so disguised itself under the mask of inflection as to tempt a scholar like Weske to suggest their inclusion within the Indo-European family. In fact, any distinction that can be drawn between the Finnic and the Aryan verb is a purely artificial one; the forms in both have originated in agglutination, and become what they are through the influence of phonetic decay. So far as form is concerned, there is little difference between the Ostiak madâdm, madân, madâ; madau, madár, madâda, and the Sanskrit bhavâmi, bhavasi, bhavati; abhavam, abhavas, abhavat. In the declension, too, the postpositions have in many instances ceased to be independent or even semi-independent words; indeed, the marks of certain of the cases (the genitive -n(a), the abessive -ta, the adessive -, &c.) are throughout the Turanian or Ural-Altaic world mere symbols, whose origin has been long forgotten. But for all that the Finnic idioms remain agglutinative, the Aryan languages inflectional. The Aryan languages started with flection, and made their agglutinated compounds conform to the prevailing analogy; the Finnic idioms owe the appearance of flection which they possess to the wear and tear of time. In the one case analogy, in the other case phonetic decay has worked the change. The two groups of tongues have met, as it were, in the same spot, after starting from opposite quarters; and the fact need not surprise us any more than the common resemblance in many points presented by English and Chinese. After all, languages, however unallied, have all originated under

similar circumstances from men of similar mould; they are but varying species of one and the same genus. Hence that gradual passage from one form of speech to another, described in a former chapter, and that sporadic participation of one form of speech in the characteristics of another. We may discover the principle of flection in the agglutinative Dravidian of western India, where the Tulu dialect forms the frequentative malpevé, and the causative mălpăvé from the active malpuvé, “I do,” or in the Bâ-ntu of southern Africa, where the final vowel of the noun has a passive meaning if it is -, an active or causative one if -i, a neutral one if -a,' while in Mpongwe mi kámba is "I speak," mi kâmba, "I do not speak." In the Finnic languages we can actually trace a change of signification in a root accompanying a change of vowel, and so be reminded of our own distinction between incense and incense, torment and torment. Thus karyan is “to ring" and "to lighten ;" kar-yun and kir-yun, “to cry," but kir-on, "to curse;" kah-isen, koh-isen, kuh-isen, “to hit" or "stamp" käh-isen, köh-isen, "to roar;" keh-isen, kih-isen, "to boil." 2 What is this but the Semitic mode of indicating a change of signification by a change of vowel ? The difference between the two is that the one utilizes the variation of vowel for lexical, the other for grammatical, purposes; it is the only difference, but, for determining the morphological position of a language, it is a most important one.

1 Bleek: <<

guages," p. 138.

Comparative Grammar of the South African Lan

2 Donner in the "Z. D. M. G.," xxvii. 4 (1873).

CHAPTER VIII.

THE AGGLUTINATIVE, INCORPORATING, POLYSYNTHE

TIC, AND ISOLATING LANGUAGES.

"L'idée de l'infériorité des nations touraniennes, de leur inaptitude à l'art et à la civilisation, est un vieux préjugé qui a fait son temps, et qui ne doit guères son origine qu'aux affirmations vaniteuses, et surtout intéressées des nations germaniques.” — FR. LENORMANT.

PUTTING aside the polysynthetic dialects of America, the majority of the languages of the world belong to the agglutinative class. But just as the inflectional families of speech differ one from another, so also do the agglutinative; indeed, there is a greater difference between the rude and unformed Bushman and the polished Finnic, with its semblance of flection, or the Dravidian of Western India, with its power of modifying the sense by internal vowel-change, than there is between any two groups of inflectional speech. Agglutination, too, may be of more than one kind. The agglutinated adjuncts may be either prefixed, as in Kafir, or affixed, as in Ural-Altaic; or, again, they may be almost wholly dispensed with, as in Malayo-Polynesian. The root may be modified in sound during the process of agglutination, or may remain fixed and unchangeable, whatever incrustations may attach themselves to it. A verbal stem may exist apart from a

nominal stem, or, as in Polynesian, a verb may not have emerged into existence at all. The root may influence the suffixes, producing that law of vowel harmony which assimilates the vowel of the suffix to the vowel of the root, or suffix and root may resemble two atoms in close contact which each keep their own unalterable character.

The important part played in history and civilization by the races who speak the various dialects of the UralAltaic or Turanian family makes a brief review of the leading languages of this family as necessary as a review of the Aryan or Semitic families of speech. From the eastern shores of Siberia to Scandinavia and western Russia extends a group of tongues which can all be traced back to a common mother speech. The Finns and Lapps of the North, the Esths and Ugric tribes of Russia, the Magyárs of Hungary, the Osmanlis of Turkey, the Tatars, the Samoieds, the Mongols, the Mantchus, and the Tunguses all share the fragments of a common patrimony. Possibly Japanese may have hereafter to be added to the list; for the present, however, it must remain isolated and unclassified. The oldest monuments of Turanian speech have been of late revealed to us by the cuneiform monuments of Babylonia; the wild hilltribes of Media and Susiania, the citizens of the ancient empire of Elam, and the primitive population of Chaldea itself all spoke cognate languages, which, it would seem, must be assigned to the Ural-Altaic group. Already the same intellectual power which to-day distinguishes the Finn or the Magyár had begun to show itself; and the Accadians of primæval Babylonia were the inventors of

the cuneiform system of writing, the builders of the great cities of the country, the first students of mathematics and astronomy, and, in short, the originators of the culture and civilization which was handed on to the Semites, by whom they were afterwards conquered and dispossessed. Contemporaneous records prove that Western Asia possessed its China in Turanian Accad at least five thousand years ago; and that the "wisdom of the Chaldeans," stored up in their imperishable libraries of clay, was no imaginary dream of a later age, but a startling and solid fact.

Of course it does not follow that the communities which now speak the allied dialects of the Turanian family all belong to the same race. The Lapps, in fact, though now using a Finnic idiom, are not related to the Finns in blood, and it is more than doubtful whether we can class the Mongols physiologically with the TurkishTatars or the Ugro-Finns. It is even possible that the Mongolian dialects themselves were originally distinct from those of the Turanian group, and owe their present inclusion in the group to their common agglutinative character, and to a long and close contact with the Turkish-Tatar languages, which have made them approximate so nearly to the latter as to compel us to classify them together. However this may be, the whole Turanian family is bound together by its structure, its grammar, its stock of roots, and its law of vocalic harmony. It may be divided into five branches, the FinnoUgric, the Turko-Tatar, the Samoyedic, the Mongolian, and the Tungusian, the first two representing the culti