Turks

TURKS. The use of the name “Turks” has never been limited in a clear and definite way from the time of the Byzantine authors to the present day. To the former, as also to the Arabs, it has a collective sense like Scythians or Huns; at the present day we are wont to restrict the name to the Osmanli Turks, though they themselves refuse to be called Turks, having, as they hold, ceased to be such in becoming imbued with Arabo-Persian culture. On the other hand, when we speak of Uigurs and Tatars, we mean tribes who style themselves Turks and really are such. It is only by the aid of historical and linguistical evidence that we can determine the true limits of the Turkish name.

The national Turkish traditions, preserved by the Persian historians Rashid ed-Dīn and Jowaini from Uigurian books which are now lost, point to the region watered by the river Selenga and its affluents, the Orkhon and the Tugila, as the primitive seat of the Turkish people. Rashīd ed-Dīn combines this tradition with that of the Mohammedan descendants of Oghuz, who, in accordance with Moslem traditions, derive the whole Turkish stock from Japhet, the son of Noah, or more accurately from Turk, the son of the former (Yafiz-oglan), and pretend that he pitched his tents in the vicinity of Lake Issyk-kul (in Semiryetchensk). But, though Turkish tribes did wander so far to the west, and even farther, in remote antiquity, it seems pretty certain that the Uigurian tradition has preserved the memory of the true origin of the race, that Turks and Mongols were originally different stems of a single people, and that these two members of the Ural-Altaic ( q.v.) family were more closely related to each other than to any other member of the same family (Finno-Ugrians, Samoyedes, Tungus-Manchus). The evidence for this rests, not on the ethnological system of Rashīd ed-Dīn, though it affords a secondary argument, but on the indubitable affinity of the Mongolian and Turkish languages and the similarity of the ethnological characters of the two races. Here, of course, we do not argue from the Osmanlis, who have lost all their original race-characters and have become “Caucasians” of the best type, but rather, for instance, from the Kirghiz, who are considered as the typical Turks of the present day, and are described by Ujfalvy as being midway between the Mongol and the Caucasian. We must now turn our attention to the wanderings of the Turks and their subsequent fate,—a rather difficult task, owing to the want of accurate information. The only truly historical records are to be found in the Chinese chronicles and encyclopaedias, where, however, the Turkish proper names appear in such distorted forms as to be unrecognizable ; yet , till the 6th century of our era, no other accounts are available.

It is generally admitted that the first Turkish people mentioned by the Chinese are the Hiong-nu, who, wandering to the west, occupied the country south of the Altai Mountains and expelled (about 177 b.c. ) the former occupants of those regions, the Yue-chi, Kan-goi, and Usun (U-ssun),—tribes of unknown nationality, but possibly also Turks. The Hiong-nu were identified by Deguignes with the Huns, this denomination being used in a political or collective sense, and including, besides the Huns proper, the Ephthalites or White Huns, Avars, Bulgars, Magyars, Khazars, and Petchenegs, who are styled by several scholars Hunnic or Scythian peoples,—a term of no scientific value whatever, as the main body of these peoples consisted really of Mongol-Turks or Finno-Ugrians. As, however, separate articles have been devoted to most of these ethnical names, we abstain from further details, as also from discussion of the question of the Turkish origin of the Magyars and the Khazars, though that of the former seems to us as improbable as that of the latter is certain. Be this as it may, the Hiong-nu are, so to speak, protoTurks, and the history of the Turks proper begins with the Tu-kiu, the Chinese equivalent of the word Turk. Originally a division of the Hiong-nu, almost extirpated by wars, but miraculously saved from complete destruction, the Tu-kiu settled south of the Kin-Shan (Altai ?) Mountains, and were miners and iron-smelters in the service of the Juen-Juen (“les Tartares Geou-gen” of Deguignes). About 552 A.D., however, they conquered their former masters and founded a mighty empire under princes who took the title of Ili khan. In these Tu-kiu Deguignes recognized the Turks who entered into friendly relations with Byzantium, and to whom Justin II. sent two ambassadors, —Zemarchus (568) and Valentinus (575). The narratives of these ambassadors are preserved in the fragments of Menander Protector; and (comparing the variations of the corrupt text with the record of Tabari) from him we learn that at the first date the reigning prince was Sinjibulus (Arabic Sinjibu). From the Greek and the Arabo-Persian accounts it seems that Sinjibu put an end to the empire of the Ephthalites or Haitāls in those regions. He shared the conquered country with Khosrau I., the Oxus becoming the frontier between Iran and Turān. The memory of the empire of Sinjibu and of its political strength has been preserved by the Arabic authors Ibn Khordādbeh and Mas'udī, who inform us that the Turkish tribe of the Karluks, settled in the provinces of Ferghana and Shāsh (Tashkend), were of old the mightiest of all the Turks, and that their sovereign, the khakān of khakāns, was obeyed by all the neighbouring princes. To them they reckon the mythical Afrāsiāb and the historical Shāwa. It is uncertain at what epoch the empire of the Karluks came to an end; but the Chinese assert that about 650 they reduced the inhabitants of the Ili and Jaxartes territory, though they were unable to protect them afterwards against the inroads of the Arabs under Kotaiba b. Moslim (706-714). The latter defeated the armies sent to their aid under Kurbogha Noyon, a sister’s son of the Chinese emperor. It is also doubtful if the so-called Afrāsiāb kings or Ilek-khans, who reigned in the 10th century at Kāshgar and Balāsāghun and conquered (999) the dominions of the Samanids in Transoxiana, belonged to the Karluks, as is supposed by Grigorieff and Lerch, or to the Uigurs, as others think.

The name Uigurs is very common during the Mongolian period, and Rashīd ed-Dīn and others use it (by an anachronism) in speaking of remote antiquity, though it is wholly unknown to the Arabic geographers, and, as Vambéry has shown, to the Uigurs themselves,—nay, even impossible in the old Turkish language, in which the form would be Utkur. The name Ugur, Ogur, or Ogor of Byzantine authors is really different; but Grigorieff has recognized the name in the corrupt Arabic form of Tagazgaz, which must be read Toguz-Ugur, the “Nine Ugurs,” to distinguish them from another division of the same tribe, the On-Ugur or “Ten Ugurs.” In the time of Ibn Khordādbeh and Mas'udī these Turks had gained the supremacy amongst their brethren, and had their residence at Kushān, which has been identified with the Kiao-chang of the Chinese. According to their accounts, the Kiao-chang form the southern division of the Hui-khe (Hoei-ke of Deguignes), and were settled before the Christian era south and east of the Tian-Shan up to the Pamir plateau and the Kuen-Lun. The Arabic authors make them adherents of Manichaeism; but, as the original Turkish Shamanism has developed into a dualistic system, this statement may rest on a partial misapprehension. It seems, however, certain that Buddhism reached these Turks on its way towards China, for we know that this religion spread in the 2d century b.c. throughout the adjacent kingdom of Bactria, and was still flourishing when Hwen-T’sang visited (7th century) those regions. Thus we can understand why the old Ural-Altaic religion bears a Sanskrit name. The northern division of the Hui-khe, which remained unknown to the Arabs, wandered from the Selenga region to the sources of the Yenisei, vanquished the Tu-kiu (745), and founded an empire from the Selenga to Lake Balkash, till they were overthrown (841) by the Ha-kas (identified with the Kirghiz). These northern Uigurs are called by the Chinese Kao-che, Chi-le, Di-li, and Te-le. The history of the southern branch is unknown, for the chronological data of Rashīd ed-Dīn and Abu-’l-Ghāzi are contradictory and useless, though their statements that the prince bore the title of Idi-kut and submitted to the Mongols have full historical weight. That the Uigurs rose during the Mongolian epoch to a certain supremacy by higher culture is attested by Rashīd ed-Dīn and Jowaini, who often mention Uigurian books.

The Petchenegs (Gr. ∏ ατξιvακαι , ∏ ατξιvακιται ; Magyar Besenyö ; Lat. Bisseni) were of old, as Constantine Porphyrogenitus tells us, settled about the lower Ural and Volga, but were driven thence (894-899) by the Ghuzz (Ouzoi). A part of them returned afterwards to their ancient abode, but the great majority wandered westward and settled on both sides of the Dnieper, driving the Hungarians before them to the Carpathians. Here they annoyed the neighbouring peoples by their raids, and engaged readily in the Russian expeditions against the Greek empire, till the policy of the Byzantine court incorporated large numbers of them with its own armies, sometimes with fatal result, as was experienced by Romanus Diogenes, when these auxiliaries passed to the camp of his antagonist Alp Arslan. At the period of the first crusade the Christian armies met with them on their march through Servia and Bulgaria; but the Petchenegs are not mentioned after the 13th century. The learning of Orientalists has discovered faint traces of the language once spoken by them in the Turkish dialect of the Bosnians.

Comans or Cumani (Russ. Polowtze, Magyar Palocz and Kun) is a term chiefly used by Europeans for the Turkish tribes that occupied Moldavia and the adjacent regions of south Russia. The origin of the name is uncertain; but it seems to be Turkish, though it rarely occurs in Oriental records. The most probable conjecture regarding the people denoted by it is that they were a mixture of Ghuzz and Petchenegs. Oriental authors know much more of their neighbours to the east, the Kipchaks, a very common name of Turkish clans down to the present day. Sometimes both names are combined: Rubruquis speaks of the Coman Kipchaks. Anna Comnena informs us that the Comans spoke the same dialect as the Petchenegs, a dialect well known to European scholars from the so-called Codex Cumanicus. On the arrival of the Mongols in these regions, the Kipchaks suffered great hardships, and large numbers of them were sold as slaves throughout the Mohammedan world. From them sprang the Bahrite Mameluke sultans of Egypt (1250-1380). The Comans sought refuge amongst the Hungarians and became Christians; but their arrival, causing internal dissensions in Hungary, greatly favoured the advance of the Mongolian arms. The remnants of the Comans, Kipchaks, and other Turkish tribes continued to dwell in southern Russia under Mongolian rule (see Mongols), whilst others became merged with the Hungarians.

The Ghuzz dwelt originally in the far East amongst the Toguz-Ugur, but migrated in the reign of the caliph Al-Mahdí (775-785) to Transoxiana, where they adhered to the cause of the famous Al-Mokanna', not from religious predilection, but to satisfy their love of war and plunder. In the same manner they afterwards served every warlike prince in their neighbourhood, and entered like other Turks into the service of the caliphs. The main body of the life-guard of the 'Abbāsids consisted of Turks, and some individuals rose very soon to high commands. Entrusted with the administration of distant provinces, they founded independent princely houses, such as those of the Tulunids and Ikhshids in Egypt (vol. vii. p. 750) and the Ghaznavids (see Ghazni). In the meantime fresh bands of the Ghuzz poured from the east and the north into Turkestan, the region becoming overstocked with a nomadic population. Some of them sought and found an outlet to the west by occupying the territory of the Petchenegs and joining the Turkish population of southern Russia; but the great majority, seeing the ruined state of the empire of the caliphs, crossed the Oxus and overran the northern and eastern provinces of Persia. How these loose desultory bands were guided to subsequent victories, and moulded with the peoples amongst whom they settled into regular political bodies, has been already narrated under Seljuks ( q.v. ).

Meanwhile they underwent a great change in their outward appearance, habits, &c., as Rashīd ed-Dīn relates, owing to the influence of the air and the water, and, we may add, to frequent intermarriage with the inhabitants of the countries invaded by them. After some generations the change was great enough to strike their Iranian neighbours, who called them Turkmans (Turcomans), a term implying resemblance to Turks. It is therefore quite natural that the modern Osmanlis should have become Caucasians; for, if Rashid ed-Dīn in the 13th century noticed the difference between a Turkman and a genuine Turk, the six centuries which have elapsed since amply suffice to have obliterated all original Ural-Altaic characteristics. The old name Ghuzz, originally, as it seems, the Turkish Oghuz (an eponymous hero of whom Turkish chronicles tell many fables) was wholly superseded by the new name Turkman and by other political names.

During the Seljukian period there arose in Transoxiana the empire of the Khārizm shahs, founded by Mohammed b. Anushtegīn, upon whom the government of Khārizm (Khiva),—which down to 995 had belonged to princes of Iranian descent—was conferred (1097) by the Seljuk prince Barkiyārok. His son Atsiz became independent (1138), but his empire seemed destined to early ruin by the arrival of the Kara-Chitai, who defeated the Seljuk prince Sinjar (1141) and became for a time supreme masters of Turkestan. Nevertheless the Kharizmian dynasty held its place as a tributary sovereignty, and developed great power under the princes Takash and Mohammed his son. The former defeated and killed (1194) the last Seljuk prince of 'Irak, and the latter extended his dominion from the Caspian to the Indus and from the Jaxartes to the coast of 'Omān. His cunning antagonist, the 'Abbasid caliph An-Nāsir, invoked the aid of Jenghiz Khān, who scarcely required this invitation to attack Mohammed. The sack of Bokhara was followed in the same year (1220) by that of the other principal cities of Transoxiana and by the persecution of the unhappy prince, who died in a forgotten island of the Caspian. His son Jelāl-ed-Dīn was driven towards India, but by a change of fortune ascended once more the throne of his fathers, till the new Mongol khān, Ogdai, sent fresh armies against him and forced him to seek refuge in the Kurdish Mountains, where he was murdered (1231). The fate of the Turks of Transoxiana was shared by their brethren in Asia and Europe, and new Mongol-Turkish empires arose, of which an outline has been given under Mongols. As the Mongol rule grew weaker, there arose in Persia and on the frontiers of Asia Minor the Turkman dynasties of the Ak Koyun-lu, the Kara Koyun-lu, the Zu’l-kadria, and the Ramazān Oglu, whose history is closely connected either with that of the rise of the Ottoman power (see below) or with the history of Persia (see vol. xviii. p. 632 sq.).

At the present day the Turkish people occupy a very extensive area, the centre of which lies in Khiva, Bokhara, and Khokand, and which stretches from the lower Lena in Siberia to the Danube and from the Crimea to Kermān and India. Politically they belong to Russia, Turkey, Persia, China, and Afghanistan. In religion the great majority are Mohammedans; a few tribes in Russia are baptized Christians; and some others adhere to the original Shamanism, which has also influenced the religious conceptions of the Christian and Mohammedan Turks. The principal Turkish peoples are the following. (I.) By a popular distinction the Turks of Siberia and Russia, with some colonies in Turkey, are styled Tatars (see Tartars), though the Yakuts of northern Siberia are not usually included in this term. The Yakuts, who are perhaps a mixture of Turkish and Tungus tribes, deviating from the ordinary course of Turkish wanderings, are settled about the lower Lena, and number probably 200,000 (Rittig, 80,000; Lansdell, 210,000). They are nominally Christians. (II.) On the Kirghiz (Kara-Kirghiz and Kazaks) and Kara-Kalpaks see Kirghiz ; but note that the Kipchaks, named there as a separate tribe, really form a subdivision of the Kazīk-Kirghiz, and are perhaps akin to the Kitai-Kipchaks, who are reckoned to the Uzbegs. (III.) Uzbeg is a political, not an ethnological denomination, originating from Uzbeg Khan of the Golden Horde (1312-1340). It was used to distinguish the followers of Shaibāni Khān (16th century) from his antagonists, and became finally the name of the ruling Turks in the khanates as opposed to the Sarts, Tajiks, and such Turks as entered those regions at a later date and are known to be Kirghiz, Kara-Kalpaks, or Taranjis. The Uzbegs are therefore a mixed race of different Turkish tribes. According to Kostenko, they number 201,972 in the Russian provinces of Sir-Daria, Ferghana, Zerafshan, and Amu-Daria, and Vambéry conjectures that there are 1,000,000 more in Bokhara, 700,000 in Khiva, and 200,000 under Afghan supremacy, giving a total number of about 2,000,000. They are agriculturists or inhabit the cities; a few are semi-nomads. (IV.) The eastern Turks on the southern slopes of the Tian-Shan Mountains at Kashgar, Ust-turfan, Ak-su, Sairam, Kutcha, Yarkand, Khotan, &c., are the remnants of the ancient Uigurs; and of the same origin are the Taranjis ( = agriculturists), settled in the Ili valley and elsewhere. The number of the latter is given as about 50,000; that of the former may be estimated from the statements of Forsyth and Kuropatkin at about 1,000,000 for the whole district, the great majority being Turks and the rest Mohammedan Chinese (Sungans). (V.) The Turcomans (properly Turkmans) inhabit the steppe east of the Caspian and south of the Oxus from Astrabad to the Paropamisus. The term is sometimes taken to include their brethren in Persia and Asia Minor, who will be treated separately. The following are the principal tribes:—(1) the Tchaudors and Imrailis, in the northwestern part of the Ust-Urt to the Gulf of Karaboghaz; (2) the Yomuts, extending from Khiva across the Ust-Urt to the Caspian, and along the sea-board to Persia; (3) the Göklen, on Persian territory, between the upper Görgen and Atrek; (4) the Tekkes, the most numerous tribe at the present day, divided into the Akhal Tekkes and the Merv Tekkes, so named after the centres where their greatest numbers are found; (5) the Sakars, on the left bank of the Oxus, to the east of Tcharjui, considered by Vambéry as a division of the Tekkes; (6) the Sariks, at Penjdeh and Yul-utan on the north-western slopes of the Paropamisus; (7) the Salors, one of the oldest Turkman tribes, who suffered greatly from the Tekkes, till they finally migrated (1857) to Zurabād in Persia, and left their former districts to the Tekkes and Sariks; (8) the Ersaris, on the Oxus about Khoja Salih; and (9) the Ali-elis, about Andkhui. Their total number, inclusive of some Turkmans who do not belong to any of these tribes, and are scattered throughout the provinces of Syr-Daria, Amu-Daria, Zerafshan, and Astrakhan (about 16,000), is estimated by Vambéry at about 1,000,000, and by Grodekoff at 1,170,000. The Turkmans are, with few exceptions, nomads, and were formerly the terror of their neighbours, who feared them as the “man-stealing Turks”; but since Merv has been annexed to Russia (1884) they have been compelled to abandon their predatory habits. (VI.) The Turkish nomads scattered throughout Persia are partly the descendants of the Ghuzz tribes that invaded the country at the Seljukian period; others have migrated thither in the following centuries. They are known by the name of liât or Iliyāt (meaning tribes or peoples) and consist of several tribes, having each its own chieftain, the Ilkhānī, appointed by the shah. An accurate list of the names of these tribes does not exist; but the most powerful and most numerous are the following. (1) The Kajars, who dwelt in Transcaucasia down to the time of Abbas the Great, by whom one division of them was compelled to settle at the south-east corner of the Caspian near Astrabad. To this division belongs the present dynasty of Persia. (2) The Afshars or Aushars, a very numerous tribe, in the province of Adarbaijān (Azerbijan). A division is also settled in the mountainous regions of the Anti-taurus; its members are nominally subjects of the Ottoman empire but really independent. (3) The Shekakis and Shah-sewen. The latter is not a tribal, but a political name, meaning those who love the shah, i.e ., partisans of the Safawī dynasty (1499-1736) and the Shi'ite faith. (4) The Kara Koyun-lu, near the town of Khoi, the remnants of the once powerful tribe named above. Besides these, many other names are recorded of tribes wandering in the Transcaucasian regions and in the provinces of Adarbaijān and Mazenderan, but many of them are very uncertain. All these Turks are comprehended under the general denomination of Adarbaijāni Turks; they are nomads or semi-nomads and speak a peculiar Turkish dialect, the Turk Azeri or Adarbaijāni Turkish. Some specimens of it have been published by Chodzko, Bergé, Melgunoff, and Barbier de Meynard. In the southern provinces of Persia are settled the (5) Kashkais, (6) Abul-werdis, (7) Kara - Gözlüs, (8) Bahar-lu, and (9) Inan-lu. To the first named are reckoned by some the Khalaches, an old Turkish tribe which was already settled near Herat before the Seljukian period and has given rise to some Indian dynasties. Vambéry thinks that the total number of Iranian Turks may amount to about two millions, or, if we add the Caucasian Turks under Russian supremacy, three millions. (VII). The Osmanlis, under which term are comprehended all the Turkish subjects of the sultan of Turkey, consist chiefly of the following elements.

(1) Turkmanian tribes and Turks of every description, who poured into Asia Minor after the defeat of Romanus Diogenes (1071); to these we may also reckon the Ottomans proper, though they did not enter the country till after the downfall of the Kharizmian empire. The Mongolian invasion drove the obscure ancestors of this the most illustrious Turkish dynasty to Asia Minor, whence they gradually spread to the province of Khodawendikyār (Bithynia).

(2) Tatars scattered amongst the rest of the population, but forming a large colony in the Dobrudja. In part they occupied their present settlements before the conquest of Constantinople; but others have immigrated into Asia Minor during the last two centuries from the Crimea and Caucasus, since the Russian conquests of those regions. They have fared very badly under Turkish rule, as is attested by Captain Wilson. That tribes of Turkish origin were settled in Europe long before the rise of the Ottoman power is known from the Byzantine authors, who mention a colony of them (about 30,000) as early as the 10th century in the Vardar valley in Macedonia.

(3) The so-called Kizil-bashis or “Red Heads,” a nickname of the Shi'itic Turkish immigrants from Persia, who are found chiefly in the plains from Kara-hissar along Tokat and Amasia to Angora. During the Avars with Persia the Turkish sultans forced them to settle here. They are agriculturists and highly praised by several travellers for their honesty and laborious habits. (4) Turkmenian tribes—Yuruks and Götchebes (words meaning “nomads” and characteristic of their most distinctive quality),—who occupy the mountains in summer and descend into the plains in winter, though some are settled in the plains of Cilicia near Tarsus and Adana, the rest being semi-nomads. Reclus estimates the total number of Turks in Europe at 1,500,000 and 35,000 Tatars. For Asia Minor statistics are wanting; but P. de Tchihatchef, the chief authority for matters relating to this peninsula, thinks that 6,000,000 is a failestimate for the total population, including Greeks, Armenians, Kurds, &c., but excluding the islands. It appears therefore necessary to reduce the already moderate number of Osmanlis given by Vambéry (10,000,000) to about 6,000,000.

Language.

The Turkish, or, as some prefer to say, the Turco-Tatar language, is a member of the Ural-Altaic family (see Ural-Altaic) and comprehends many dialects, which differ considerably in their vocabulary and in a less degree also in their grammar. The study of these dialects has made great advances during the 19th century. Abel Rémusat in 1820 knew only of four, viz., the Uigurian, Jagatai, Tatar, and Osmanli. Béresine in 1848 distinguished nineteen, grouped round three types, viz., (1) Jagatai dialects (Uigur, Coman, Jagatai, Uzbegian, Turkmani, Kazani literary language); (2) Tatar dialects (Kirghizian, Bashkiri, Nogai, Kumi, Karatchai, Kara-Kalpaki, Meshtcheryaki, and Siberian); (3) Turki dialects (Derbendi, Adarbaijāni, Krimmi, Anadoli, and Rumili). Böhtlingk (1851) added the Yakuti, and Shaw (1877) the Eastern Turki. Radloff (1882) subdivided the one Siberian dialect of Béresine into more than a dozen different dialects. On phonetic principles the last-named proposes the following classification, which seems, however, not quite satisfactory: (1) Oriental dialects (Altai, Baraba, Lebed, Tuba, Abakan, Küärik, Soyon, Karagass dialects, and Uigur); (2) Occidental (Kirghizi, Irtish, Bashkir, and Volga dialects, with numerous subdivisions); (5) Central-Asiatic (Taranji, Jagatai, &c.); and (4) southern (Turkmani, Adarbaijāni, Caucasian, Anadoli, Krimmi, and Osmanli). It would be premature to criticise this system till the author publishes the second part of his grammar, which will treat of the real etymological phenomena of the north Turkish dialects.

On the phonetical characteristics of each of these dialects ample information is given in his Phonetik der nördlichen Türk. Sprachen.

These great dialectical varieties are easily accounted for by the want of a common Turkish literary language understood everywhere. The most developed and refined Turkish tongue, that of the Osmanlis, which is very rich in literary monuments, has admitted too many Arabic and Persian words, grammatical forms, and even whole sentences, and has been too much spoiled by the precepts of Persian rhetoric, to produce a popular literature. With the exception of some tales and novels, this literature has remained an exotic production, unintelligible even to the people who are supposed to speak the same language (see Turkey, p. 656 above). The Jagatai and Uzbegian dialects would have answered the purpose better, and present the best type of a (hypothetical) general Turkish language, of which the most prominent features may be here given.

The Arabic alphabet is in general use, though some tribes in Russia make use of Russian and others in Asia Minor of Armenian and Greek characters. But the oldest Turkish alphabet, the Uigurian, is a direct transformation of the Syriac, and has fourteen characters. When and by whom it was invented is uncertain; the Arabic author of the Fihrist does not mention it, and the Uigurian MSS. which we possess date for the most part from the 15th century. It is commonly supposed to be the work of Nestorian missionaries, who may have preached the Gospel amongst the Turks as early as the 6th or 7th century. In the age of Sinjibu the Turks seem to have used the Sogdian characters in their political intercourse with Byzantium; but as a rule they remained illiterate till their conversion to Islam. As the Semitic languages are characterized by the three-radical system and the constancy of the consonants, all Ural-Altaic languages are dominated by the law of vowel harmony and agglutination. We have therefore in Turkish a double range of vowels, commonly eight in number, of which a, i, o, u denote the hard or guttural and ä, ï, ö, ü the soft or palatal vowels, the vowels in every separate word being of the same range. The i only is in most dialects indifferent. The law of agglutination is derived from the same principle, but has regard, not only to the vowels, but also to the consonants and the syllables; it is an abuse of the term if it is taken to mean that in Turkish no real etymology exists, but only an agglutination of themes and roots.

As regards the etymology we observe the absence of gender, of a separate form for the dual, and of the nominative in the nouns. There are commonly five oblique cases—genitive, dative, accusative, commorative, and ablative—though Böhtlingk has shown that in the Yakut dialect, which distinguishes ten cases, the genitive is wanting. The adjective, unless used as substantive, is uninflected both as attribute and as predicate; the comparative is formed by the suffix -rak (-rek ), and takes the compared noun in the ablative; the superlative has no specific form, though a peculiar intensive is formed by prefixing to the adjective (though in writing always as two words) a syllable beginning with the same consonant, and ending in a labial p or m : for instance, kap kara, “intensely black”; kip kizil, “intensely red.” The decimal system has prevailed over an original septimal system. The article does not exist. The relative pronoun has been borrowed from the Persian in many dialects; it is absent in the original Turkish. The theme of the verb is seen in the imperative, from which are derived various participles and gerunds, used either separately or combined with pronominal suffixes. These combinations supply the forms of the simple tenses and moods, though different dialects use different forms of participle and gerund for this purpose. Compound tenses and moods are expressed by means of auxiliary verbs. The theme of the imperative may, by the addition of a simple consonant, vowel, or syllable, be modified into a negative, passive, reflexive, reciprocal, impossible, causative, or doubly causative form, which are conjugated in the same manner as the original form. The causative forms again admit of a passive negative, &c., so that in fact the number of possible verbal forms derived from a single theme has been calculated by Shaw at 29,000. There are no prepositions, only postpositions.

In syntax the order of the words and clauses of a period is almost the inverse of what seems natural to us, the subject and its predicate being placed at the end, while all hypothetical, causal, prohibitive,—in short all subordinate—clauses come first. In the simple style of illiterate peasants, and in popular romances and tales, this method presents no inconvenience as regards easy understanding, but in the artificial, often excessively long periods of an Osmanli stylist, it presents serious difficulties to a European reader.

Bibliography.—(a) General works on the history and ethnography of the Turks: Deguignes, Histoire des Huns', Vanibéry, Das Türkenvolk (Leipsic, 1885), Ursprung der Magyaren (Leipsic, 1882), and several other publications; Radloff, Aus Sibirien (Leipsic, 1884); W. Grigorieff, Zemlewjedjenie K. Rittera Wostotschni ili Kitaiski Turkestan ; Neumann, Die Völker des südlichen Russland (Leipsic, 1847). We may add the historians of the Mongols — D'Ohsson, Howorth, and others—the numerous journals of travellers amongst Turkishpeoples, an<l several articles in the Russische Revue, Journal of the Royal Asiatic Soc., &c. A full bibliography of works relating to Central Asia may be found in V. J. Mejoff, Recueil du Turkestan (St Petersburg, 1878-84), and a useful excerpt at the end of vol. ii. of Lansdell's Russian Central Asia. Other works have already been cited in the course of this article.

(b ) For the study of Turkish dialects the subjoined books may be used. (1) Osmanli : the grammars and dictionaries of Redhouse, Mallouf, Zenker, Barbier de Meynard, &c. (2) Uigur: the works of Klaproth; Abel Rémusat Recherches sur les Langues Tartares (Paris, 1820); and Vambéry, Uïgurische Sprachmonu-mente und das Kudatku Bilik (Innsbruck, 1870). (3) Jagatai : the dictionary of

Pavet de Courteille, and Vambéry, Jagataïsche Sprachstudien (Leipsic, 1867). (4) Eastern Turki : Shaw’s grammar and vocabulary (Jour. Roy. As. Soc. of Bengal, 1877). (5) Tatar dialects : the grammars of Kasimbeg-Zenker (Leipsic, 1848), Ilminski (Kazan, 1869), and Radloff (Leipsic, 1882); Dictionary of Trojanski (Kazan, 1833); the chrestomathies of Béresine (Kazan, 1857), Terentieff, and specialiy Radloff, Proben der Volksliteratur der türkischen Stämme Süd-Sibiriens (St Petersburg, 1872). And (6) Yakuti : Böhtlingk, Die Sprache der Jakuten (St Petersburg, 1851). (Μ. T. H.)