general similarity, and the verb in Buriat, which differs from other Mongol languages, exhibits a development parallel to Turkish.
The want of resemblance in vocabulary between the three classes of languages is remarkable. The numerals, for instance, in Turkish, Mongol and Finno-Ugric are entirely different, and considerable changes have to be assumed before the identity of words can be proved. A comparison of Turkish words with Mongol equivalents makes it probable that the former are in many instances contractions: thus dagh, mountain, yol, road, correspond to the Mongol dabaga, yabudal and perhaps represent an earlier tavagh and yavol. The best-known Turkish languages, particularly Osmanli, have borrowed an enormous number of Arabic and Persian words which disguise the characters of the native vocabulary and to some extent affect the grammar.
Compared with the Finno-Ugric group, the Turkish languages are remarkably uniform. Indeed, allowing for the lapse of time and the importation of foreign words, it is hardly an exaggeration to say that from the Lena to Constantinople, from the Orkhon inscriptions till now, we have merely one language in different dialects. The native vocabulary and grammar remain substantially the same. The linguistic type is evidently strongly individual and persistent, and its separation from Mongol, &c., is probably very ancient.
Radlov divides the Turkish languages or dialects into four groups, according to their phonetic system, (1) Eastern: Altai, Baraba, Lebed, Tuba, Abakan, Küärik, Soyon, Karagass and Uighur. (2) Western: Kirghiz, Bashkir, Irtysh and Volga dialects. (3) Central Asiatic: Jagatai, Taranji, &c. (4) Southern: Turkmani, Azerbaijani, Krimmi, Anadoli and Osmanli. But this classification does not seem entirely satisfactory. As one passes across Asia from the Yakuts, through Kashgar, Turkestan and Azerbaijan to Constantinople, the pronunciation of the Turkish languages becomes decidedly softer, the suffixes become more intimately united with the words to which they are appended (approaching though not attaining the unity of Finnish inflexions), and the verbal forms grow more numerous and more complicated. Thus in the east we find nin, ni, ga as suffixes for the genitive, accusative and dative, and man for that of the first personal pronoun (e.g. durman, I stand or I am) corresponding to -in, -i, -a and -im in Osmanli, which have clearly assumed the character of inseparable terminations more completely than the older forms. Osmanli possesses more copious verbal forms than the other dialects, some of which (such as the future in -ajak) seem to be recent formations. On the other hand, the dialects of Turkestan use in speaking, though not in writing, forms which indicate a process of composition followed by contraction, more remarkable than any change which has taken place in the west. For instance, wopti, a contraction of bolup irdi, is said to be currently used in Khokand for “has become.” Yakut (which can still be best studied in Böhtlingk's excellent grammar of 1851) is the dialect which is most distinct from the others, but does not appear always to preserve the oldest forms. Thus it has lost the genitive, which is replaced by a pronominal periphrasis (e.g. örüs bas-a, horse head-his, i.e. horse's head), and has verbal forms like bisabin, I cut, bispappin, I do not cut, apparently standing for bisarbin, bispatbin. The negative suffix is pa not ma. The resemblance between the Turkish dialects is increased by the fact that they are nearly all written in a somewhat artificial and standardized form which imperfectly represents the variety existing in conversational speech.
Several alphabets have been employed to write Turkish, (1) Arabic characters are everywhere used by Mahommedan Turks, almost without exception; yet this alphabet is extremely ill suited to represent Turkish sounds. It cannot distinguish the hard and soft vowels, so that oldu, “he was” is written in the same way as öldü, “he died.” In some cases the consonants indicate the character of the vowels which are to be supplied after them, hard consonants being followed by hard vowels and soft by soft. Thus the word spelt with the letters kaf, re, he is pronounced as kara, but that spelt with kef, re, he as kerre. Further the orthography often follows an antiquated pronunciation and the letters have many sounds. Thus the single letter kef can be used to express k, ky, g, gy, y, v, w and n. The result is that pure Turkish words written in Arabic letters are often hardly intelligible even to Turks and it is usual to employ Arabic synonyms as much as possible because there is no doubt as to how they should be read. Osmanli documents are often little more than a string of Arabic words with Turkish terminations.
2. The Uighurs and Eastern Turks used in the middle ages a short alphabet of fourteen letters derived from a Syriac source and probably introduced among them by Nestorian missionaries; similar characters may also have been employed by Manichaeans. The Mongol and Manchu alphabets represent further variations of this writing. Though very like the modern Nestorian, it is in some respects more nearly allied to the Estrangelo and Syro-Palestinian alphabets of the 6th and 7th centuries. The most important document in this alphabet is a MS. preserved at Vienna of the Kudatku Bilik, “The Blessed or Fortunate Knowledge,” a poem composed at Kashgar about 1065. A colophon states that the MS. was written at Herat in 1465, and that it is a copy of one written in 1085. Inscriptions in a similar alphabet have also been found in China.
3. The most interesting forms of Turkish writing are those used on the inscriptions found in Siberia near the Yenisei and Orkhon rivers. For some time it has been known that stones bearing inscriptions as well as roughly carved figures and hunting scenes were to be found on the upper waters of the Yenisei, particularly near its tributary the Abakan in the district of Minusinsk. They are greatly venerated by the Soyotes inhabiting the region. They were first discovered by Messerschmidt in 1722, and some of them were represented in the plates of Strahlenberg's Das nord. und östliche Theil von Europa und Asia (1730). They were generally attributed to Scythians or Chudes. The knowledge of them did not much advance until the researches of Castren (1847) and the Finnish Society of Archaeology, which in 1889 published the text of thirty-two, chiefly from the Uibat, Ulukem, Altynkul and Tes. Even more interesting are the monuments discovered in 1889 and known as the Orkhon or Kosho-Tsaidam inscriptions, as they were found in Mongolia to the south of Lake Baikal, between the river Orkhon and Lake Koshp-Tsaidam. The most important are a mortuary inscription in Turkish and Chinese, bearing a date corresponding to 733, in honour of Kül-tegin, and another recounting the exploits of Bilgä Kagan. A third inscription at Kara-Balgassun probably dates from 800-805. The inscriptions were deciphered and translated by Thomsen and Radlov, and Donner examined the origin of the alphabet. He came to the conclusion that the Yenisei alphabet is rather older than that of the Orkhon inscriptions, and that both are derived from the Aramaic alphabet and most nearly allied to the variety of it used on the coins of the Assacid dynasty. In the 3rd century A.D. a section of the Kirghiz, who subsequently moved northwards, were in West Sogdiana and in touch with the Yüe-Chi, who had been for some time in contact with Persia. The old Turkish characters bear a superficial resemblance to runes; the Yenisei letters have the simplest shapes, those of Kara-Balgassun the most complicated. But they are mostly traceable to Aramaic prototypes and have no connexion with Scandinavia. The vowels are generally omitted, even at the beginning of words, and, as in the modern Turkish method of using the Arabic alphabet, their quality is often indicated by the consonants, many of which have two forms, one used with soft the other with hard vowels. Thus bar and bär are differentiated not by the vowels but by the consonants employed to write them.
4. Turkish-speaking Armenians and Greeks often write it in their own alphabets. Turkish newspapers printed in Armenian characters are published in Constantinople, and Greek characters are similarly employed in several parts of Asia Minor.
Bibliography: (a).—General works on the history and ethnography of the Turks: Deguignes, Histoire des Huns; Vambéry, Das Turkenvolk (Leipzig, 1885), Ursprung der Magyaren (Leipzig, 1882), and several other publications; Radlov, Aus Sibirien (Leipzig, 1884); W. Grigoriev, Zemlewjedjenie K. Rittera Wostotschni ili Kitaiski Turkestan; Neumann, Die Volker des südlichen Russland (Leipzig, 1847). We may add the historians of the Mongols—D'Ohsson, Howorth and others—the numerous journals of travellers amongst Turkish peoples, and several articles in the Russische Revue; Journ. Royal Asiatic Soc.; Revue orientale pour les études Oural-altaiques, and other Oriental periodicals; Skrine and Ross, Heart of Asia (1899); Cahun, Turcs et Mongols (Paris, 1896); E. H. Parker, A Thousand Years of the Tartars (1895), and numerous articles, especially in the Asiatic Quarterly by the same author on Chinese accounts of these tribes; Chavannes, Les Tou-kiue occidentaux (St Petersburg, 1903).
b. For the study of Turkish dialects the subjoined books may be used. (1) Osmanli: the grammars, dictionaries and chrestomathies of Wells (1880), A. Wahrmund (1884) and Redhouse (1890). (2) Uighur: the works of Klaproth; Abel Rémusat, Recherches sur les langues tatares (Paris, 1820); Vambéry, Uigurische Sprachmonumente und das Kudatku Bilik (Innsbruck, 1870), and a newer edition by W. Radlov (St Petersburg, 1900). (3) Jagatai: the dictionary of Pavet de Courteille and Vambéry, Jagataïsche Sprachstudien (Leipzig, 1867). (4) Eastern Turki: Shaw's grammar and vocabulary (Journ. Roy. As. Soc. of Bengal, 1877). (5) Tatar dialects: the grammars of Kasimbeg-Zenker (Leipzig, 1848), Ilminski (Kazan, 1869) and Radlov (Leipzig, 1882); Dictionary of Trojanski (Kazan, 1833); the chrestomathies of Béresine (Kazan, 1857), Terentiev and specially Radlov, Proben der Volksliteratur der türkischen Stämme Süd-Sibiriens (St Petersburg, 1872). (6) Yakuti: Böhtlingk, Die Sprache der Jakuten (St Petersburg, 1851); Radlov, Yakutische Sprache in ihrem Verhältniss zu den Turksprachen (1908). (7) Inscriptions: Société finlandaise d'archéologie, Inscriptions de l'Iénisei and several works by O. Donner, W. Radlov and V. Thomsen—especially Thomsen, Inscriptions de l'Orkhon déchiffrées (Helsingfors, 1896); Donner, Sur l'origine de l’alphabet turc (Helsingfors); Radlov, Die alt-türkische Inschriften der Mongolei (St Petersburg, 1897); Marquardt, Chronologie der alt-türkischen Inschriften (1898). (C. El.)
