CHINESE TURKESTAN i. Geographical Overview

The eastern portion of the Central Asian land mass (see central asia i. geography), between 70° and 100° E and 25° and 45° N, encompasses Chinese Turkestan, now Sinkiang (Xin-jiang) Uighur Autonomous Region of the People’s Republic of China, with the Tarim basin and the high plateaus and mountains surrounding it (capital Urumchi [Wu-lu-mu-chi]); Tibet (capital Lhasa [La-sa]); the eastern portions of the Kazakh Soviet Socialist Republic between the Tien Shan and Altai mountains; and the Gobi (Shamo) desert, including parts of the Mongolian People’s Republic and Inner Mongolian Autonomous Region of China.

The area is characterized by spectacular peaks and plateaus at high elevations, interspersed with generally arid lowlands. The steppes of Kazakhstan and Sinkiang, continuing into eastern Mongolia and northwestern Manchuria, form the easternmost portion of the desert belt that extends across the Persian plateau and Arabia as far as western North Africa. In contrast to Soviet Central Asia, however, in Inner Asia the natural severe aridity of the climate has not been altered by large-scale irrigation projects. The plains can be classified in three zones of elevation: 100-500 m, 1,000-1,500 m, and 3,000-4,500 m. As in Central Asia, those in the lowest zone generally lie on the northern and western flanks of the great mountain ranges, the Alai (Fergana [Farḡāna] valley: 250-400 m), Tien Shans, Dzungarian Alatau, Tarbagatay, and Altai. The intermediate zones, on the other hand, generally lie south of the mountains, particularly the Tarim basin and Mongolia as far as the Greater Khingan range (Ta Hsing-an-ling [Xing-an-ling] Shan) in Manchuria and Inner Mongolia. Nevertheless, the second deepest depression on earth after the Dead Sea (-395 m), the Turfan oasis (-154 m), is found on the southern flank of the Tien Shans. The highest steppe zones include the Pamirs and the Tibetan plateau. Again following the pattern of Central Asia, the plateaus generally rise from east to west: For example, the elevation of the Tsaidam swamp in Tsinghai province on the northeastern perimeter of Tibet is 2,700 m, that of Lhasa 3,600 m, that of Chiang-tzu (Gyangtse) 4,000 m, and that of Ka-erh (Gartok) in western Tibet 4,467 m. Fertile yellow loess covers the plains at every altitude, even beneath the sand of the deserts. Today mechanized agriculture is breaking up huge areas of the loess plains, sometimes even removing the sand cover of the deserts, in order to convert them into arable land. Permanent settlements are generally confined to the northern foothills of the great mountain ranges and adjacent portions of the plains. On the steppes beyond only the nomadic life of the livestock herder is possible.

Precipitation, though minimal in the flatlands, increases with elevation; at 1,000 m it may reach a maximum of 1,000 mm a year, that is, almost as much as in coastal areas. The perpetual snow line generally lies between 3,000 and 3,500 m in peripheral ranges but slightly above 4,000 m in the massive systems farther east. In the central Tien Shans, for example, passes above 4,000 m are generally free of snow during the entire summer. In the Pamirs and most parts of the southernmost chains of the Himalayas the snow line lies as high as 5,500 m. The snow line on the northern slopes is often as much as 500-700 m lower than on the sunny southern slopes. In contrast to the broad expanses of steppe north of 45° N, where heavy snow cover lasts for a relatively long period because of low winter temperatures, on the southern borders of Kazakhstan it rarely exceeds a depth of 20 cm. It is not unusual for the arid zone to extend right up to the perpetual snow line. Taller vegetation is found only in irrigated areas, and forests are confined to higher altitudes along the northern mountain slopes.

The southern mountain barrier and Tibet. Except for the Greater Khingan range, which runs north-south, the crestlines of the mountain systems, like those of western Turkestan, generally follow east-west axes, which seem to fan out from the Pamirs. Extending southeast from the Pamirs the Karakoram range, the Himalayas, and the parallel Trans-Himalayas (the Kailas [Kang-ti-ssu] range) describe a huge arc curving eastward and forming the boundary between the Indian subcontinent and Inner Asia. In the southeastern Chamdo (Khams) region of Tibet, between 90° and 110° E, these giant ranges turn southeast and south, forming the relatively unexplored mountainous region of southwestern China and northern Indochina. In the northwestern Karakoram, in the portion of Jammu and Kashmir controlled by Pakistan, the famous K2 (Mount Godwin Austen), the world’s second highest peak, rises to 8,611 m. In the Himalayan and Trans-Himalayan chains are several very high peaks—Nanga Parbat, also in Pakistani Jammu and Kashmir (8,126 m); Mount Dhaulagiri in Nepal (8,172 m); Mount Everest (Tib. Cho-mo-lung-ma), the highest peak in the world (8,848 m), on the border between Nepal and Tibet; and Kangchenjunga, , the third highest (8,579 m), between Nepal and Sikkim—and a great many at elevations just under 8,000 m.

Valley glaciers clustered around such peaks are the basic and permanent sources of all the rivers that flow through Inner Asia, but, whereas most empty into self-contained lakes with no outlet to the oceans, those rising on the Tibetan plateau include many of the major river systems in southern and southeastern Asia. The Baltoro, Siachen, and Remo glaciers south and southeast of K2, as well as others on the northern slopes of the Himalayas, are particularly important; although most are in retreat, thick moraine cover in their lower reaches deters thawing. The Indus and the Brahmaputra originate in close proximity in a geological fold between the giant ranges of the Karakoram and the Himalayas, the watershed between them marked by Ma-na-sa-lo-wu (Manasarowar), a large salt lake 150 km southeast of Ka-erh (elev. 4,602 m). The Indus (Tib. Seng-ge Kha-bab) flows northwest through a number of deep, narrow gorges in the western Himalayas onto the plains of the Punjab, where the major rivers draining the southern slopes of the mountains in the west flow into it. The Brahmaputra (Tib. Rta-mchog Kha-bab) follows a similarly intricate course around the eastern end of the Himalayas into the huge, tropical Assam valley. The Tibetan plateau itself is an arid steppe, dotted with large salt lakes. The four great river systems of Southeast Asia and China have their sources in the eastern portion. The three southernmost—the Salween, the Mekong, and the Yangtze (Kinsha-kiang)—flow generally south and southeast, separated by high parallel ranges extending into Indochina. The fourth, the Yellow river (Huang Ho), follows the Yangtze closely in its upper course, then bends east and northeast toward Mongolia. Most of the population of Tibet is concentrated in the river valleys along the southern margin of the plateau.

The Kunluns and the Tarim basin. Directly east of the Pamirs and the parallel Mu-ssu-t’a-ko-a-t’e (Muztagh Ata range; highest peak Muztagh Ata: 7,546 m) the Kunlun Shan and its eastern extension, the Nan Shan (Southern mountains), separate the Tibetan plateau from the Tarim basin and the Gobi desert farther east. Wu-luk’o-mu-shih (Ulugh Muztagh: 7,723 m) is the highest peak in the Kunlun range, located at 87.5° E, where the A-erh-chin Shan (Altin Tagh) splits off to the north; many other peaks rise above 6,000 m, particularly in the western portion of the range. Farther east some of the southern chains, for example, the A-ni-ma-ch’ing Shan (Amne Machin Shan; highest peak Amne Machin, 7,164 m) and Pa-yen-k’a-la Shan (Bayan Kara Shan), which separate the upper reaches of the Yellow (Tib. Rma Chu) and Yangtze (Tib. Bri Chu) rivers, extend deep into western China; in their western stretches there are apparently many high peaks that have not yet been surveyed or remain unpublished. North of the A-ni-ma-ch’ing Shan are the southern chains of the Nan Shan system, which itself continues as the Ch’in Ling (Tsinling/Qin-ling) Shan as far as western Honan province in China, forming the watershed between the Wei and Han rivers. The elevations surveyed in the Nan Shans so far do not exceed 5,000 m and are progressively lower farther east.

North of the Kunluns lies the Tarim basin, nine-tenths of its surface consisting of sand desert, the Takla Makan (Ta-ko-pi), with dunes up to 20 m high; its constantly shifting sands are steadily encroaching on the cultivated areas. The Yarkand (Yeh-erh-ch’iang) river, which rises in the eastern Karakoram, flows together with the Kashgar (Kyzylsu, K’a-shih-ka-erh), which has its sources in the Alai and Trans-Alai ranges, to form the Tarim river (T’a-li-mu Ho), the main tributary of which is the A-k’o-su (Aksu), descending from the Tien Shan mountains to the north; all the former southern tributaries, which rise in the Kunluns, disappear into the dry earth before they reach the Tarim. The river flows eastward through the desert to feed the shallow Lop Nor (Lo-pu Po, elev. 730 m), the “wandering lake” discovered by Sven Hedin in 1896 and again in 1934, when it had shifted to the northeast of its former position. Such shifts are frequent and reflect both irregular changes in the volume of the Tarim and the effects of wind on the shallow waters. The major towns in the basin are laid out along two branches of the ancient Silk Road starting from Kashgar (K’a-shih; elev. 1,230 m) at the western end. The northern branch followed the southern slopes of the Tien Shans through Aqsu (A-k’o-su, 1,010 m), Kuçā (K’u-ch’e), Qārāšahr (1,090 m) near the Baghrash Köl (Po-ssu-t’eng Hu, 890 m), and Turfan (T’u-lu-fan, -50 m), crossing the mountains to Barkol (Pa-li-k’un,1,720 m). The southern branch followed the northern slopes of the Kunluns through Yarkand (So-ch’e, 1,200 m), Khotan (Ho-t’ien, 1,410 m), Keriya and Niya (both 1,430 m), and Cherchen (1,280 m). The ancient city of Krorain (Lou-lan) was situated northeat of Lop Nor.

The Tien Shans and the northern plains. The Kok Shaal Tau branches off from the Pamirs to the northeast, linking them with the Tien Shans, which run generally parallel to the Kunluns. Beginning about 70 km east of Tashkent, the western Tien Shans, the highest peaks generally exceeding 3,000 m in elevation and snow-capped most of the year, separate the fertile Fergana valley from the southern Kazakh steppes. The major cities in the valley, Andizhan (Andejān), Namangan, Margelan, Kokand (Ḵoqand), and Leninabad (Ḵojand) along the upper Syr Darya, are in the Uzbek S.S.R. Osh, near ancient Uzgen in the southeastern comer of the valley, is in the Kirgiz S.S.R.

Farther east the Tien Shans comprise a number of parallel chains extending along the northern perimeter of the Tarim basin and separating it from the low plateaus of eastern Kazakhstan and Dzungaria. About 600 km east of Tashkent the northern chains divide into the Kungey and Zailiĭskiĭ (Trans-Ili) Alatau on the north and the Terskey Alatau on the south, flanking the large basin of the Issyk Kul (Issiq Köl, lit. “warm lake,” because it does not freeze), 1,623 m above sea level and 700 m deep. Located about 240 km east-southeast of Issyk Kul is Victory Peak (Pik Pobeda; 7,439 m), the highest in the Tien Shan system; slightly to the north is Khan-Tengri (6,995 m) with the associated Inilçäk glacier, 65 km long and 400 m thick, extending down to 2,883 m above sea level.

The highest elevation in the eastern Tien Shans is Po-ko-to Shan (Mong. Bogdo-ola “venerable mountain”; 5,445 m), about 70 km east of Urumchi. Although the highest peaks of the Tien Shans thus cannot compete with those of the Alai and Hindu Kush, the average elevation of the relatively even crest line is only slightly below 4,000 m. The elevations of the Tien Shans decline considerably at around 95° E; a number of lower parallel ridges, generally known in Chinese as the Pei-t’a Shan (Mong. Baytag Bogdo “Northern mountains”), form a kind of barrier between the northern perimeter of the Tarim basin and the Gobi desert. Nevertheless, between the eastern reaches of the Tien Shans and the Nan Shans the two expanses open into each other, forming part of the continuous desert belt across Asia; temperatures in these arid continental regions can reach as high as 50° C.

Several relatively short parallel ranges lie between the Tien Shans and the Altai range farther north, for example, the Dzhungarian Alatau (Char-ka-erh-a-la-t’ao Shan), which includes one peak rising to 4,158 m at 78° E, where it approaches the P’o-lo-k’o-mu Shan (Borokhoro) chain of the Tien Shans, and the generally much lower Tarbagatay range farther north and east, broken by large valleys serving as natural passageways between the plateaus of Inner Asia and the vast lowlands of northern Kazakhstan and Siberia. Mus Tau, marking the transition from the Tarbagatay range to the Altai at 85° 40’ E, rises to 3,805 m. Urumchi, the capital of Sinkiang, lies on the ancient northern caravan route that passed north of the Tien Shans through Kuldja (I-ning, at present the second largest city of Sinkiang), near the present Soviet border, and Dzharkent (Panfilov in the Kazakh S.S.R.).

Near the western end of the Tien Shans, at about 72° E, the low Karatau ridge splits off to the northwest, defining the western border of the Muyunkum desert (elev. 300-700 m) of the southern Kazakh S.S.R. Two small rivers flowing from the Tien Shans, the Talas and the Chu, vanish in its sands. The Chu forms the boundary between this desert and the Yetisu (lit., “seven rivers,” Rus. Semirech’e) desert farther east, which takes its name from a few small watercourses that cross it from the Dzungarian Alatau and empty into Lake Balkhash, 340 m above sea level and 26 m deep. The surface of the lake is shrinking at a rapid rate and may soon be cut in two. The most important river in this region is the Ili, which drains the long valley between the eastern Kungey Alatau and the P’o-lo-k’o-mu Shans, then flows north across the plains to a marshy delta at the southern end of Lake Balkhash. The population of these desert areas is concentrated along the southern perimeter, particularly in Dzhambul (former Aulie Ata) on the upper Talas river, Frunze (former Pishpek), capital of the Kirgiz S.S.R., on the upper Chu; and Alma-Ata (former Vernyi), capital of the Kazakh S.S.R. at the base of the Kungey Alatau.

To the east of Lake Balkhash lie other shallow lakes, for example, the Sasykkol (Stinky lake) and Alakol (Blue lake) north of the Dzungarian Alatau and the smaller Ebi Nor (Ai-pi Hu) south of it, at the western end of the central Dzungarian desert, all fed by local streams.

CHINESE TURKESTAN ii. In Pre-Islamic Times

In antiquity the Tarim and Dzungar (Zungar, Jungar) basins lay at the crossroads of three main Eurasian routes (see i, above; EIr. II/6, p. 595 fig. 47). These three routes included the Southern Silk Road, which skirted the northern foothills of the Altin Tagh (A-er jin Shan) and Kunlun ranges; the Northern Silk Road along the southern foothills of the Tien Shans; and a northern route passing between the Bogdo-ola (Bo-ko-tuo) range and the Tien Shans, then along the northern edge of the latter, through the Dzungarian Gate between lakes Alakol and Ebi Nor (Ai-bi Hu), and across the Eurasian steppes as far as Europe. In the Tarim basin, which consists mainly of the desert of Takla Makan, a number of small oasis states provided halting places for travelers (Samolin, 1964, pp. 9-18); they included Kamul (Ha-mi), Turfan (Tu-lu-fan), Qarāšahr (near modern Yan-qi), Kucha (Ku-che), and Aqsu (A-ke-su) on the Northern Silk Road and Miran (near modern Ruo-qiang), Niya near modern Endere Langan (An-ti-er Lan-gan), Khotan, and Yarkand (Suo-che) on the Southern Silk Road. These two routes merged at Kashgar (Ka-shi) at the western end of the basin.

The principal textual references to the geography and early inhabitants of Central and Inner Asia are found in the Old Persian inscriptions of Darius I (521-486 b.c.e.) and Xerxes I (486-65), the works of Greek authors beginning with Herodotus in the 6th century b.c.e., and early Chinese histories of the Han (206 b.c.e.-220 c.e.) and preceding, largely legendary dynasties. The Old Persian inscriptions include lists of the territories in the Persian empire; farthest to the northeast were those of the Sogdians, Gandharans, and Sakas (Scythians). The easternmost peoples named by Herodotus, in his description of the Silk Road, were the “Issedonians” (4.25-26), who were probably located in the northern steppes as far east as the Altai mountains (Pauly-Wissowa, IX/2, cols. 2235-46), but he knew little of the land between them and the Scythians farther south and west. Claudius Ptolemy (fl. 127-45 c.e.), drawing on information from traders and other travelers along the silk roads, described the country of the Sakas (6.13; Ronca, pp. 37-39), located east of the Jaxartes (Syr Darya), and the “Scythians beyond the Imaos” (the Pamirs; 16.14; Ronca, pp. 49-51), to the north and east of the Sakas. He mentioned no Saka cities but did refer to “Scythian” Issedon, probably in the vicinity of modern Kashgar. Still farther east lay Serica (6.16; Ronca, pp. 52-58), approximately equivalent to the eastern Tarim and Dzungar basins, with another Issedon, perhaps located near the later site of Kroraina (Lop-Nor; Pulleyblank, 1962, p. 127), and in the extreme northeast Throana, later Dun-huang, which is mentioned in the 4th-century Sogdian Ancient Letters as ’rwʾʾn (the origin of the name is uncertain; Pulleyblank, 1966, p. 21, proposed a Yue-zhi origin, H. W. Bailey, 1982, pp. 20 n. 46, derived it from Ir. *druvāna “stronghold,” and V. Mair, forthcoming, argues for a relationship with Ind. draṃga “frontier post,” itself also ultimately an Iranian loanword). The Chinese sources include a few references to the area that may have been preserved from pre-Han times. King Mu of the Zhou dynasty (traditionally 1122-256 b.c.e.) was supposed to have traveled there in approximately 985-80 b.c.e., but the account of his journey (Mu tian-zi zhuan), the extant text of which probably dates from before 300 b.c.e., is extremely difficult to interpret in the light of information available from other sources (Haloun, pp. 301-2; Watson, 1983, p. 537 and n. 1; Pulleyblank, 1966, p. 19). There are also vague references to western peoples and places in other early texts like Shu-jing (Book of documents, compiled largely in the Han period and perhaps preserving some Zhou elements) and Shan-hai-jing (Book of mountains and seas, composed at an indeterminate date before the beginning of the current era; Loewe, p. 658; Pulleyblank, 1966, p. 21), but they, too, are difficult to correlate with known archeological and historical data. Several reports by travelers who went as far as Syria under the Han dynasty (e.g., Hou Han-shu; Chavannes, 1907) include considerable information about the geography and history of eastern Turkestan. Persian and Arabic authors, on the other hand, had only vague notions of the geography of the area before the Muslim expansion into the region in the 4th/10th century (see, e.g. Ebn Ḵordāḏbeh, pp. 29-31; Ḥodūd al-ʿālam, tr. Minorsky, pp. 83-86, 223-35); the Pahlavi books of the 8th and 9th centuries contain no mention of the area.

During a brief interlude of intensive archeological exploration in the late 19th and early 20th centuries Russian, French, German, Swedish, English, and Japanese scholars discovered traces of a variety of ethnic groups and political configurations that had existed in the Tarim basin during the 1st millennium c.e. (Yaldiz). Subsequently, however, apart from systematic Chinese excavations in the Kucha oasis begun in 1950 (Chao et al., pp. 2-3) and a few isolated finds from salvage excavations elsewhere in Sinkiang (Xin-jiang), little further progress has been made. The Tarim basin as a whole has not yet been the object of sustained archeological investigations comparable to those carried out in Soviet Turkestan since World War II.

From the archeological finds it is now clear that the earliest known inhabitants of the Tarim basin spoke Indo-European languages. The two most prominent groups were Tokharians and Iranians, about whose origins and cultures little is known. The only certainly datable Tokharian inscriptions in Indian Brāhmī script are a group found in the Kucha oasis, written in the mid-7th century c.e. (Chao et al., pp. viii, 61, 128-33), but, if, as T. Burrow concluded (1935), one group of Prakrit documents in Kharoṣṭhī script found at Niya, Endere, and Lou-lan includes linguistic elements borrowed from a Tokharian language, then the presence of these people in the eastern Tarim basin would be attested as early as the 2nd-3rd centuries. The Tokharians spoke a language of the so-called centum (Lat. “hundred”) branch of the Indo-European languages, which is more closely related to Celtic, Germanic, Italic, and Greek than to the Indian and Iranian languages, which belong to the satəm (Av. “hundred”) branch. They were probably not immigrants from far western regions, however; more likely they represented an old, marginal section of the proto-Indo-European community that had been displaced by the expansion of central groups, including the Indo-Iranians. The date at which they arrived in the Tarim basin is still a matter of debate. It has been customary to identify as “Indo-Scythians” or Iranians the Yue-zhi (variant pronunciation Ru-zhi) and the Wu-sun (perhaps pronounced *ʾa[g]sən in Old Sinitic; see Schnessler, pp. 645b, 592b), mentioned in Chinese texts of the 2nd century b.c.e. as inhabitants of the area around Dun-huang (see Pulleyblank, 1970). Other scholars, on the other hand, most recently E. G. Pulleyblank, have identified the Yue-zhi with the Tokharians, presumably displaced eastward by Indo-Iranian migrations at an early date (1966, p. 14; 1983, p. 457-58; cf. Grousset, pp. 22-28; Narain, 1987, pp. 7-15; chinese-iranian relations ii). If this identification is accepted and if, as several scholars have argued, the Yue-zhi were already present on the western borders of China in the late 2nd millennium b.c.e. (Narain,1987, pp. 4-7; cf. Pulleyblank, 1983, p. 457), then the Tokharian migrations must have occurred even earlier.

Iranians in the Tarim basin. Artifacts and artistic motifs commonly found in Sinkiang reveal clear links with known “Scythian” pieces from neighboring areas, and it is clear that Iranians must have been present in the Tarim basin from an early date (for the Zaman-Baba culture, see Past Worlds, p. 149). Among their descendants must be included the peoples still settled in the southwestern Tarim basin and the Pamirs (see vi, below); the Khotanese, who lived along the Southern Silk Road east of Khotan; a group on the Northern Silk Road around present-day Maralbashi (Ba-chu) and nearby Tumshuq (Tu-mu-xiuke) west of Aqsu; and the Sogdians, who, though settled in Central Asia, were active as traders throughout what is now Chinese Turkestan from the early centuries of the present era. Plants (e.g., alfalfa, grapes; Laufer, Sino-Iranica, pp. 208-45; cf. Schafer, pp. 141-45), animals first domesticated on the Persian plateau (goats, sheep; Schafer, pp. 75-76; see boz; cattle; dāmdārǰ), musical instruments (balloon guitar, harp; see, e.g., Hayashi, pp. 98, 128), and other aspects of Iranian culture reached the Tarim basin and China quite early. The elaborate underground irrigation system (kārēz) in the Turfan basin (see, e.g., Le Coq, 1929, pp. 52-53) reflects technology developed in Persia and Armenia in the first half of the 1st millennium b.c.e. (see Gaube, p. 28 nn. 2-3 with refs.; D. Hill, pp. 29, 33-36, and refs. pp. 45-46), and already in the Western Han period the horses of Fergana (Farḡāna) were in great demand in China and the Tarim basin (Samolin, 1964, pp. 22-23 and n. 16). The Sogdians and Khotanese were primarily responsible for dissemination throughout the region of the four great pre-Islamic religions: Buddhism, Mazdaism, Manicheism (see vii, below), and Christianity. Iranian-speaking peoples were also involved in the eastward spread of Islam.

The Sogdians’ principal base of operations was the fertile Zeravshan valley around the capital, Afrasiyab (Afrāsīāb, near Samarkand), which was visited by Chinese ambassadors in the 7th century (depicted on murals of the residential quarters preserved in the Afrasiyab museum; Mair, personal observation). According to the Xin Tang-shu (New Tang history, comp. 1060 c.e.), Sogdians went “wherever profit is to be found” (Pulleyblank, 1952, p. 317). They established colonies at all the main entrepôts along the eastern portion of the silk roads, in the Han capital at Lo-yang, and far into Inner Mongolia (Pulleyblank, 1952). According to the Sogdian Ancient Letters, there was a large and economically active community at Dun-huang in the 4th century c.e., attested by the large number of individuals with the surname Kang (< Kang-qu = Samarkand; Pulleyblank, 1952, p. 320) who lived there (Grenet and Sims-Williams). So prominent were the Sogdians along the Asian trade routes that until the 7th/13th century their language was the virtual lingua franca in the region; it was then replaced by Persian (Pelliot, 1912, p. 105).

The kingdom of Khotan near the western end of the Southern Silk Road was founded, according to legends reported in Tibetan and Chinese sources, during the reign of the Maurya king Aśoka (ca. 272-31 b.c.e.; see Stein, 1907, pp. 156-66; Emmerick, 1967, pp. 15-25; aśoka iv) in India, but no mention is made of Iranians in these stories. Nor is it possible to conclude from several references to Khotan in the annals of the Han dynasty, beginning in about 140 b.c.e., whether or not the inhabitants were Iranians (Chavannes, 1903, passim; Samolin, 1964, passim). The kingdom expanded after the middle of the 1st century c.e., but in about 73 the king was forced to submit to the Chinese, who established a garrison in his territory (Chavannes, 1903, p. 156). After 150 it became fairly independent again, but coins from this period found at the site of Khotan, inscribed in Chinese and Kharoṣṭhī (Cribb), provide no clue to the ethnic identity of the Khotanese. The first concrete evidence of an Iranian presence in the country is found in a document probably of the 3rd century, discovered by M. A. Stein at the site of Endere (facsimile in Stein, 1921, pl. xxxviii; transcription in Boyer and Senart, p. 249; tr. 1940, p. 137; cf. Emmerick, 1979, p. 168 and n. 7). It was written in a local Middle Indian dialect in Kharoṣṭhī script by Khotana maharaya rayatiraya hinajha Vij’ida Siṃha “General Vijida Simha, great king, king of kings of Khotan” in his tenth chuna (< Khot. kṣuṇa) “regnal year.” The Khotanese title hīnāysa (pronounced hīnāza, lit. “army leader”) is also attested in much later indigenous texts. Khotan was an early center of Buddhism and was visited over several centuries by Chinese Buddhist travelers, who reported on the land and its people. The earliest was Fa-xian, who was there in 400 (Beal, I, pp, xxv-xxvii).

In the high Pamirs there was a kingdom around Tashqurghan (Ta-shi-ku-er-gan) mentioned in the Jiu Tang-shu (Old Tang history, 945) as Cong-ling (Chavannes, 1903, pp. 32, 73 n.) and later as He-pan-tuo, Han-tuo, Ke-le-tuo, and Jie-pan-tuo (Chavannes,1903, p. 124). According to the Xin Tang-shu (XX, p. 6234; tr. Chavannes, 1903, pp. 124-25, cf. p. 162), it was situated south of Kashgar (Su-le) and east of Wāḵān (Wakhan, Hu-mi) and Šeḡnān (Shughnan, Shi-ni or Shi-qi-ni). The royal family had come originally from Kashgar, and the people, considered violent, looked like the Khotanese and spoke the same language (note that the modern Iranian dialect spoken in Wāḵān is, in certain respects, more closely related to Khotanese than to any other Iranian language; for a survey, see Skjærvø, 1989, p. 375). The Xin Tang-shu adds that people in Kashgar had “green eyes and tattoos on their bodies” and that their king was from the Pei family, that is, the royal family of Kashgar (cf. Chavannes, 1903, pp. 121, 124). In 435-39, according to the Xin Tang-shu, Cong-ling first established relations with China, under the rule of the Northern Wei dynasty, (386-534; Chavannes, 1903, pp. 124-25). According to the Bei-shi, which includes information gathered by the party with which Song Yun traveled in 519, Ke-pan-tuo, Khotan, Wāḵān, Bukhara, and Gandhara were among the countries subject to the Hephthalites (Chavannes, 1903, pp. 224-25). In 635 an ambassador from Cong-ling appeared at the Tang court, and in 659 the three kingdoms of Kashgar, Zhu-jiu-bo (south of Yarkand), and He-pan-tuo joined together in the conquest of Khotan (Xin Tang-shu; Chavannes, 1903, p. 72). Between 713 and 741 China conquered the entire western Tarim basin and placed it under the protectorate of An-xi (see below). The Korean Buddhist monk Hui-chao (Kor. Hye-ch’o) traveled to Cong-ling in around 726, taking fifteen days to pass through Wāḵān (north of which, he was told, lay Šeḡnān), but he found no population in the area, as King Pei-xing had rebelled and fled to Tibet (Fuchs, p. 455; Yang et al., pp. 55-56).

The Iranians on the Northern Silk Road inhabiting the area around Tumshuq and Maralbashi spoke an Iranian language (Tumshuqese, also referred to as “Tumshuq”) most closely related to Khotanese but in some respects more archaic (Konow; Bailey, 1958, pp. 147-54; Emmerick, 1989). Very little is known about them, however. Some fifteen documents in this language survive, most of them official and private letters, with a few Buddhist Hinayana (Nikāya) texts (among them a fragment of the Araṇemi-jātaka, Bailey, 1968, p. 44) and one Manichean text (W. B. Henning apud Konow, p. 171). They include the names of kings and their regnal dates (Konow), but as these rulers are not mentioned in other sources their reigns cannot be synchronized with what is known of the history of the region. The Tumshuqese must have been in close contact with the Tokharians in the nearby Kucha oasis, with whom they shared a common form of the Brahma alphabet and from whom they borrowed the names for poetic meters (Konow, pp. 172-73; Bailey, 1968, p. 44), probably in the 7th century.

Relations with China. China, like all the major states bordering on what is now Chinese Turkestan, was interested in controlling and exploiting the region from an early date. Although throughout history civil and military units were nominally charged with supervision of the Tarim and Dzungar basins, they seldom exercised real or lasting authority beyond Dun-huang at the eastern tip of the Tarim basin, the strategic garrison town at the western end of the long Gan-su corridor. A significant Chinese agricultural population was settled there, and the oasis was protected by an elaborate system of defensive works, constituting the westernmost extension of a line that was eventually replaced by the Great Wall. It was there that the true boundary between China and the Western Regions (Xi-yu, until recently the Chinese name for Inner Asia), between Han and nomad, lay (Mair, 1989).

The first reliable records of Chinese activity in Inner and Central Asia are connected with the mission of the Chinese emissary Zhang Qian (Chang Ch’ien), which lasted from 138 to126 b.c.e. He had been dispatched by the Han emperor Wu-di (r. 141-87 b.c.e.) to forge an alliance with the Yue-zhi against the nomadic Xiong-nu, who were said to have displaced them from their ancestral lands around Dun-huang. Having followed the Southern Silk Road, he found the Yue-zhi in Sogdiana (between the Oxus and Samarkand), where they had settled after conquering Bactria and dividing it into five kingdoms (Shi-ji; cf. Pulleyblank, 1970, p. 159). One of these kingdoms was Kušān (Gui-shuang), which came in subsequent centuries to have enormous influence in India and Central Asia. Zhang Qian failed to convince the Yue-zhi to help fight the Xiong-nu, but he did succeed in opening Inner Asia to Chinese power and interests. As a result of his explorations, Wu-di sent expeditionary forces into the region, erected a line of limites and watchtowers, and created a network of government offices for dealing with the Western Regions (see, e.g., Loewe, pp. 163-65; Yü, pp. 405-12). The main reasons for looking west were to prevent Asian nomads from pouring into Chinese territory and to ensure the uninterrupted flow of commerce, consisting mainly of the export of silk; there is only scant evidence for what might have been imported in exchange (Watson, 1983, pp. 547-53), though the horses of Fergana (see Pulleyblank, 1966, p. 22 with n. 5; Samolin, 1964, pp. 22-23 with n. 18) were much sought after in China.

When the Chinese began to extend their direct control into the Tarim basin at the end of the 2nd century b.c.e. there were thirty-six independent kingdoms along the trade routes south of the Tien Shans (Hou Han-shu; Chavannes, 1907, p. 153; Pulleyblank, 1966, p. 15); the mountains themselves and the area north of them were inhabited by nomadic tribes, including the Xiong-nu, and the Wu-sun. During the final years of the Former (or Western) Han dynasty (206 b.c.e.-23 c.e.) the thirty-six kingdoms split into fifty-five kingdoms, formed an alliance against China, and submitted to the Xiong-nu (Hou Han-shu; Chavannes,1907, p. 155). The Chinese sources contain no useful information about the ethnic composition of any of these kingdoms, however.

This general political alignment persisted through the Later (or Eastern) Han period (25-220 c.e.), with only rare breaks, most particularly in the years 73-102, when China exercised genuine authority in the Tarim and Dzungar basins. In 73, according to the Hou Han-shu (Chavannes, 1907, p. 156), Emperor Ming-di (29-75) sent his generals to attack the Xiong-nu, and the kings of Khotan and the other states sent their sons to serve the emperor (i.e., as hostages). In the year 77 the extremely able general Ban Chao was established in Khotan (Hou Han-shu; Chavannes, 1907, p. 158). Through a combination of clever strategy and ruthless machinations he managed to install a number of local rulers who were more favorably disposed to the overlordship of the Han than to that of the Xiong-nu or other powers. In 91 he became protector-general, with his seat at Kucha, whence in 94 he attacked Qarāšahr farther east and in 97 even sent a mission as far west as Parthian Mēšān (Mesene; see characene and charax) at the head of the Persian Gulf (Hou Han-shu; Chavannes, 1907, pp. 159, 177-78; cf. Samolin, 1964, pp. 35-44). Immediately after his death in 102, however, Chinese influence in the region began to wane (Bielenstein, p. 269; Yü, p. 413).

By the beginning of the 3rd century China was too weak to exercise control over the Western Regions and for most of the next four centuries was itself (or at least that part of its territory north of the Yangtze [Yang-zi]) ruled by a succession of foreign peoples, Xiong-nu, Tibeto-Burmans, proto-Mongols (e.g., Xian-bei [< *Särbi], Wu-huan [Avars]), and Turks; it was finally reunited only under the short-lived Sui dynasty (581-618). Of these groups the Wu-huan were in control of the Tarim and Dzungar basins in the 3rd-5th centuries (Pulleyblank, 1983, pp. 452-54; Samolin, 1964, pp. 52-54; Grignaschi, pp. 225-31). By the beginning of the Tang period (618-907) various other Turkic groups were also gradually filtering into the Tarim basin from the north. These movements were the first in a gradual process of Turkization of Inner and Central Asia.

Although documentation is scarce, it seems that the Hua, or Hephthalites (Ye-dai-yi-li-tuo; on the identification of the Hephthalites see Enoki and Ghirshman, who concluded they were basically Iranians; and chionites), supplanted the Wu-huan in the Tarim basin. The Hua invaded Khotan in 445 (Chavannes, 1903, p. 224), and between 502 and 556 the country appears to have been subject to them (cf. Stein, 1907, p. 171). Beginning in the 6th or 7th century reports in the Chinese records can be compared with Khotanese documents, which contain names of kings and references to local events (Zhang and Rong, 1987; idem, 1988, in which a 5th- or 6th-century date is proposed for some Khotanese legal documents on wood found at the site of Domoqo, q.v.; see also maps in Zhang and Rong, 1987, p. 86).

Under the Tang the Chinese embarked on a series of conquests of the oasis states surrounding the Tarim basin, though the actual campaigns were conducted mainly by Turkish (Tu-jue) condottieri on their behalf (Grignaschi; Samolin, 1964, pp. 56, 59-60). In 644 the famous Buddhist traveler Xuan-zang described Khotan in some detail (Beal, II, pp. 309-24). Four years later, in 648 or 649, the Chinese placed the country under the An-xi (Guazhou) protectorate and transferred the administrative seat from Qočo (Gao-chang) in the Turfan basin to Kucha; Kucha, Kashgar, Khotan, and Qarāšahr were designated thenceforth as the “four garrisons” (see map in Wechsler, p. 227). The names of the kings of Khotan, as well as some key events during the period between 648 and the Tibetan conquest of the country in 791, have also been recorded in the Tibetan versions of what were apparently official annals of Khotan (J. E. Hill). From the mid-7th to the mid-8th century China sent huge armies into Inner and Central Asia and decimated much of the indigenous population. Expansion to the west was, however, brought to a sudden halt in 751, at the battle of Talas (Ṭarāz), when Tang troops under the Korean general Gao Xian-zhi were defeated by Muslim armies led by Zīād b. Ṣāleḥ (Ebn al-Aṯīr, V, p. 449; Chavannes, 1903, pp. 143, 152-53; Barthold, Turkestan ³, pp. 195-96; Samolin, 1964, pp. 66-67; see also central asia iv. in the islamic period up to the mongols).

The corrupt Tang government was also experiencing difficulties at home. At the end of 755 the general An Lu-shan (703-57; see also an-hsi), a favorite of the Tang emperor, Xuan-zong (685-762), staged a coup d’état. His father had been part of a group of Sogdian origin attached to the eastern Turks in Mongolia and his mother a Turkish aristocrat, and he was said to be fluent in six non-Chinese languages. He weighed more than 400 pounds and was renowned at the Tang court for his ability to dance the “Sogdian whirl” (Hu-xuan wu). Soon after his forces had succeeded in taking the capital he was killed by a eunuch slave in collusion with his own son An Qing-xiu. The restored Tang dynasty remained weak, however, and it can fairly be said that An Lu-shan was directly instrumental in its demise (Pulleyblank, 1955; Samolin, 1964, p. 67).

Toward the end of the 8th century the Tibetans extended their domination over the Tarim basin, capturing Khotan in 791 (Samolin, 1964, pp. 69, 80-81). In one Khotanese document, dated in the sixteenth regnal year of Viśa Kīrrti (ca. 803; Skjærvø, forthcoming), it was noted that the land of Khotan was being “watched over” by the Tāguttas, most probably to be identified as the Tibetans (Bailey, 1968, p. 91). Not much is known about the kingdom during the Tibetan occupation, which lasted until about 842 (see, e.g., Beckwith, pp. 155, 171; Samolin, 1964, pp. 69-70), at which time a combination of internal struggles and pressure from the eastern Turkish Uighurs, who had migrated from their homeland in Mongolia after their capital city on the Orhon was sacked by the Qirghiz in 840 (Samolin, 1964, pp. 68-70; Beckwith, pp. 146-71; Hamilton, 1986, p. xv), brought an end to Tibetan power in the Tarim basin. One group of Uighurs settled along the Northern Silk Road, with its capital at Qočo, remaining independent until 1130, when it submitted to the Qarā Ḵeṭāy (Samolin, 1964, p. 70). A second group settled in Gan-su and founded a small state with its capital at Ganzhou; it became a vassal state of the Tanguts (Tibeto-Burmans), or Xi-xia, in 1028 (Samolin, 1964, pp. 70, 75 n. 15, 83). At approximately the same time that the Uighurs were consolidating their hold on the eastern Tarim basin the Īlek-ḵāns (Qarakhanids), with their capitals at Balāsāḡūn on the Chu river in what is now eastern Kazakhstan and at Talas, were expanding both westward into Samanid territory and eastward into Fergana and the western Tarim basin. After the Samanids took Talas in 280/893 the second Qarakhanid capital was shifted to Kashgar (Samolin, 1964, pp. 78-79).

A large number of Khotanese documents from the 10th century attest continued close relations with the Uighur rulers of Sha-zhou (Dun-huang) between 914 and 1014 (Zhang and Rong, 1984, p. 41). In 938, according to the Chinese histories, diplomatic relations between Khotan and China were reestablished when King Viśa Saṃbhava (r. 912-ca. 966; Skjærvø, forthcoming), whom the Chinese called Li Sheng-tian, was then officially recognized as the legitimate ruler of the “jewel country of Khotan” (Zhang and Rong, 1984, p. 29 n. 18). The embassy dispatched to his court arrived in his twenty-ninth regnal year (ca. 941). Viśa Saṃbhava/Li Sheng-tian married a princess of the Chinese governing family at Sha-zhou, the Cao, and both are identified by inscriptions on wall paintings in the caves at Mo-gao/Dun-huang (Zhang and Rong, 1984, p. 42; See caves of the thousand buddhas). Their son Viśa Śūra led a victorious army against the Muslim Īlek-ḵān (tāžik) of Kashgar and reported on the campaign in a letter to the ruler of Sha-zhou, dated 17 February 970, in which he mentioned the capture of a dancing elephant; both the original letter and a record of it in the Chinese annals have been preserved (Pulleyblank, 1954, pp. 91-92; Bailey, 1968, pp. 58-61). One of Viśa Śūra’s brothers, Cong-de, has been identified with Tcūṃttehi, known to have composed poems in Khotanese (Kumamoto, 1986, pp. 231-32). King Viśa Śūfra was succeeded by Viśa Darma in 978 (Skjærvø, forthcoming), and relations with the Sha-zhou government seem to have been broken off shortly thereafter.

The continued presence of the Uighurs along the borders of Khotan and in Sha-zhou and Gan-zhou is reflected in a series of Khotanese documents, probably dating from 990-93 and connected with the resumption of relations (Hamilton, 1977; Kumamoto, 1982). If the dates are correct, these documents are among the latest that have been preserved from Khotan before the Muslim conquest, which was accomplished by 396/1006 (see iii, below). Buddhist culture and the Khotanese language itself disappeared, and few traces of the early Iranian or Tokharian languages formerly spoken in the Tarim basin remain.

CHINESE TURKESTAN iii. From the Advent of Islam to the Mongols

Chinese influence in the Tarim basin began to wane after the battle of Talas (Ṭarāz) in 134/751, when Tang troops under Gao Xian-zhi were defeated by Muslim armies led by Zīād b. Ṣāleḥ (Ebn al-Aṯīr, V, p. 449; Chavannes, 1903, pp. 143,152-53; Barthold, Turkestan ³, pp. 195-96; see ii, above; central asia iv. in the islamic period up to the mongols), though Islam did not gain a permanent foothold there until much later. On the eve of the conquest by the Muslim Īlek-ḵāns (Qarakhanids, 388-607/998-1212) Buddhism was the official religion of both Khotan (He-tian) on the Southern Silk Road and of the Uighurs who controlled the eastern Tarim basin; Manicheans and Nestorian Christians (see christianity iii. in central asia and chinese turkestan) were also active in the region.

The Turkic-speaking Īlek-ḵāns had two capitals, Kashgar/Ordukänd (Chin. Ka-shi) and Balāsāḡūn/Qūzordu or Qūzolus (Moqaddasī, p. 275; Kāšḡarī, I, p. 148; Abu’l-Fedā, Taqwīm, pp. 493, 500). In 396/1006 Abu’l-Ḥasan Naṣr b. ʿAlī, ruler of Kashgar and Bukhara (388-406/998-1015), and his brother Yūsof Qadr (Qadïr) Khan (r. as great khan 417-24/1026-32), ruler of Khotan (He-tian) on the Southern Silk Road, invaded the territory of Maḥmūd of Ḡazna. Khotan must thus have already been conquered by that time (Barthold, Turkestan ³, p. 273; Samolin, pp. 81-82). As rulers over Iranian and Turkic populations, both sedentary and nomadic, the Īlek-ḵāns initiated the transition from a society characterized by linguistic, ethnic, and religious diversity to a more homogeneous community in which Islam, and eventually Turkic languages, came to predominate. The Iranian-speaking inhabitants of Osh (Ūš) in the southeastern comer of the Fergana (Farḡāna) valley, as well as of Kashgar, Yarkand (Suoche), Khotan, and other oasis cities on the Southern Silk Road, were thus subject to a gradual process of Turkization (Haneda), attested by the names of small settlements, fortresses, mountains, and mountain passes in Fergana and the western Tarim basin; in addition, many cities were known by both Iranian and Turkic names, for example, Ḵotan/Udun (Odon) and Kučā/Küsän (Kāšḡarī, II, pp. 114-15, 308; cf. Pelliot, I, pp. 411-18). In general the indigenous populations ruled by the Īlek-ḵāns either became bilingual or evolved dialects in which the pronunciation reflected the influence of both languages (Kāšḡarī, I, p. 83). Nevertheless, Khotan in particular remained a stronghold of Iranian traditions and culture. Although the inhabitants were regarded by their overlords as “settlers in the lands of the Turks,” they had “both a script and a language of their own” (Kāšḡarī, 1, p. 83). In addition, at this period Persian was becoming increasingly popular as a literary language at the courts of the Īlek-ḵāns. The indigenous commercial ruling classes of the former city-states in the Tarim basin, known in the 5th/11th century as Ṣīn (lit. “China”) or Barḵān (Lower) Ṣīn (Kāšḡarī, I, p. 341), were also able to retain some of their power, gradually forming a coalition with the military leaders of the steppe peoples. The Īlek-ḵāns’ government was based on this aristocracy, whose members were responsible for administering their own appanages (Barthold, 1956, I, p. 94). Despite the fact that the rulers were themselves Muslims, they seem to have exhibited a certain degree of tolerance toward the inhabitants of their domains. For example, according to Gardīzī (ed. Ḥabībī, p. 270), two churches were still functioning in Khotan in about 442/1050 (cf. Dauvillier, p. 287).

Throughout this period the cities on the Northern Silk Road east of Kucha (Ku-che) had remained under the control of the Uighurs at Qočo (Gao-chang) in the Turfan basin. Their domain was a center of Manichean and Buddhist culture, even incorporating some Christian elements (Gabain, p. 20). In the 5th/11th century the Īlek-ḵāns considered Kucha itself a city on the frontier of the Uighurs (Kāšḡarī, I, p. 308). The religious differences between the Uighurs and their Muslim neighbors farther west overshadowed their linguistic affinities, and Ṭāher Marvazī included discussion of the Uighurs in his history of China, rather than of the Turks (pp. 18-20).

In about 529/1135 the Qarā Ḵeṭāy (Liao), a Far Eastern people of indeterminate ethnic origin, advanced north of the Tien Shans as far west as Balāsāḡūn, where they established their capital. The Īlek-ḵāns remained in control of the southern and western Tarim basin until about 536/1141, when they were defeated in battle by the Qarā Ḵeṭāy (Wittfogel and Feng, pp. 621-23; Bosworth, p. 581). The Qarā Ḵeṭāy did not continue the system of administration through appanages, but they did rely on a decentralized system of vassal rulers, including the Uighurs on the Northern Silk Road (Jovayni, ed. Qazvīnī, I, p. 32; tr. Boyle, I, pp. 44-45; cf. Barthold, Turkestan ³, pp. 361-62), from whom they collected tribute (cf. Bosworth, p. 582). Under their rule Manicheism and Buddhism, which had never completely disappeared from the western Tarim basin, experienced a revival, and the number of Christians in the region increased as well. The Nestorian patriarch Elias III (1176-90) founded a metropolitan see in Kashgar (Dauvillier, p. 287; Barthold, 1901, pp. 57ff.), and Marco Polo’s report that there were Christians there in the 7th/13th century is corroborated by contemporary Syriac documents (Pelliot, I, pp. 208-9). Garbled reports of these developments reached Europe and fueled the legends of Prester John, a supposed Christian ruler in Asia who opposed the Muslims (Beckingham).

As Mongol power rose in the northeast, both the Buddhist Uighurs and the urban Muslims of the western Tarim basin turned against the Qarā Ḵeṭāy and welcomed the invasions of the Nayman Mongol Küčlüg (Kūčlok) beginning in 607/1211 (Jovaynī, ed. Qazvīnī, I, pp. 32, 48-49; tr. Boyle. I, pp. 44-45; cf. Barthold, Turkestan ³, pp. 361-63, 368-69). After consolidating his power, however, Küčlüg adopted a policy of persecuting Muslims (Wittfogel and Feng, pp. 651-54) and forcing them to convert to Christianity or Buddhism (Jovaynī, ed. Qazvīnī, I, p. 49; tr. Boyle, I, pp. 65-66; Barthold, Turkestan ³, pp. 356-57, 368). After the Mongol Čengīz Khan conquered the Dzungar and Tarim basins in 615/1218 (q.v.; Jovaynī, ed. Qazvīnī, 1, pp. 49-51; tr. Boyle, I, pp. 66-68) Islam was tolerated once again. From that time onward its influence was uncontested in the western Tarim basin and in fact continued to spread eastward (see iv, below).

It was not until the end of the 8th/14th century, however, that Islam became firmly established on the Northern Silk Road, when the Uighurs at Qočo were conquered by the Muslim ruler of Mongolia (Moḡūlestān) Ḵeżr Ḵojā (Ḵᵛāja; r. ca. 1389-99). The population was still primarily Buddhist when an embassy from the Timurid ruler Šāhroḵ (807-50/1405-47) visited Turfan in 823/1420 (Ḥaydar Doḡlāt, p. 52).

CHINESE TURKESTAN iv. In the Mongol Period

On the eve of the Mongol conquests the eastern oases along the northern perimeter of the Takla Makan desert and the towns in the Dzungarian (< Mong. Zunghar) steppes north of the Tien Shans (see i, above) were inhabited primarily by the Uighur Turks, concentrated in the Turfan basin and around Bešbalıq (Bīš Bālīḡ) north of the Tien Shans between present-day Qi-ta and Urumchi (Spuler). The eastern oases south of the Takla Makan were controlled by the Tangut (Tangūt, Tanokgūt), a mixed group of Tibetan and Turkish ancestry who had founded the Xi-xia dynasty in northwestern China. The western portion of the Tarim basin was inhabited by a mixture of Turkic and Iranian peoples, many of whom were Muslims (see iii, above).

Once Čengīz Khan had consolidated his power and been proclaimed great khan over the Mongols in 1206, he turned his attention to these areas. He had subdued the Tangut by 1209, though he had dispatched two probes earlier to determine the enemy’s strength (Boyle, EI ² II, p. 42). In the same year the Uighur ruler (iduq-qut/īdīqūt), Barchuq (Bārjūq) Art-tegin (ca. 1206-ca. 1230) sent an embassy to the Mongols, and two years later he personally submitted to Čengīz Khan. In 1218 he joined with the Mongol armies under Jebe to overthrow Küčlüg (Kūčlok; see iii, above), who had usurped the throne of the Qarā Ḵeṭāy in the west. Between 616/1219 and 622/1225 he and his troops campaigned with Čengīz Khan against Sultan Moḥammad Ḵᵛārazmšāh in Transoxania and Fergana (Farḡāna; Jovaynī, ed. Qazvīnī, I, pp. 32-35; tr. Boyle, pp. 44-47), and he also assisted the Mongols against the rebellious Tangut in 1226. As a reward Barchuq was given one of Čengīz Khan’s daughters in marriage (Jovaynī, ed. Qazvīnī, I, pp. 33-34; tr. Boyle, pp. 47-48; Rašīd-al-Dīn, Jāmeʿ al-tawārīḵ, Baku, pp. 338-41).

Although Čengīz Khan died in the summer of 1227, his troops shortly afterward won an overwhelming victory over the Tangut, which resulted in their elimination as a distinct people. The Uighurs were thus the principal surviving group in the eastern part of the Tarim basin (Rašīd-al-Dīn, Jāmeʿ al-tawārīḵ, Baku, pp. 324-26). Čengīz Khan had divided his empire among his four sons. The second of them, Čaḡadai (Čaḡatāy; 624-39/1227-41), was granted all the territory that lay between the land of the Uighurs in the east and Bukhara and Samarkand in the west (see chaghatayid dynasty). The third son, Ögödei Qaʾan (Ūktāy Qaʾān; r. 627-39/1229-41), received an area in the Tarbagatay mountains and adjacent steppes and in northern Afghanistan. He also succeeded his father as great khan and in that capacity appointed the merchant Maḥmūd Yalavāč Ḵᵛārazmī as supreme minister (ṣāḥeb al-moʿāẓẓam) over all the settled territory of Central and Inner Asia some time before 636/1239; subsequently Yalavāč was transferred to China, and his son Masʿūd Beg was appointed governor of the Uighur country, the entire Tarim basin, and Transoxania (Jovaynī, ed. Qazvīnī, I, pp. 84, 86; tr. Boyle, I, pp. 107-8, 110-11; Rašīd-al-Dīn, ed. Blochet, pp. 85-86; Barthold, Turkestan ³, p. 396 n. 3).

The Uighurs, as the first people of advanced culture to have submitted voluntarily to the Mongols, also played an important role in the Mongol administration. Their literary skills were particularly in demand, and Uighur script, which was derived from Sogdian script (see viii, below), was adopted for the writing of Mongolian. Uighurs often served as interpreters, translators, and dārūḡa s (resident commissioners; Rachewiltz, pp. 282-92) in Mongol-occupied territories in China and Central Asia. After much intrigue a man named Körgüz (Kūrgūz) was even appointed governor of Khorasan in 633/1236; he served until his death in about 640/1242 (Jovaynī, ed. Qazvīnī, II, pp. 228-42; tr. Boyle, II, pp. 489-525; cf. Barthold, Turkestan ³, pp. 474-75).

In the late 640s/1240s competition between the descendants of Ögödei and of Čengīz Khan’s son Tolui (Tūlī) for control of the throne split the empire and brought conflict to Central Asia, China, and Persia. For example, the Uighur iduq-qut, Barchuq’s son Salindi (Sālendī), supported the house of Ögödei and was executed by Tolui’s son Möngke (Mengü, Mankū; 649-58/1251-60) in 1253 (Jovaynī, ed. Qazvīnī, I, pp. 34-38; tr. Boyle, I, pp. 47-52). After Möngke’s death Salindi’s successor, his brother Ögüṇč (Ūgeṇč), refrained from taking sides in the ensuing struggle (1260-64) between Qubilai (Qūbīlāy) Khan (658-93/1260-94), the eventual victor and founder of the Yuan dynasty of China (1260-1368), and his brother Arïḡ Böke (Arīq Būkā); a third brother, Hülegü (Hūlāgū), Il-khanid of Persia (654-63/ 1255-65), supported Qubilai.

Qaidu (Qāydū), a great-nephew of Ögödei who controlled territory in what is now southern Kazakhstan, had supported Arïḡ Böke. He became a champion for conservative Mongols who sought to preserve the traditional customs and practices of their nomadic past and resented the growing identification between Qubilai Khan and his sedentary Chinese subjects, on one hand, and the Il-khanids and their Persian subjects, on the other. In 668/1270 he supported Masʿūd Beg and the Chaghatayid khan Baraq (Barāq; r. 664-70/1266-71) in their attack on Persia, but Baraq was soundly defeated by the Il-khanid Abaqa (663-80/1265-82) at the battle of Herat on 1 Ḏu’l-ḥejja/22 July. The conflict between Qaidu and the Yuan rulers was not resolved so quickly. In 670-71/1271 Qubilai dispatched one of his sons, Nomuqan (Nūmūḡān), to crush him, but squabbles in the army and the elusiveness of the enemy frustrated his efforts (Tu Qi, 76 p. 8b; Yuan shi, p. 265; Pelliot, II, p. 795). The most crushing blow was Qaidu’s capture of Nomuqan himself in 675/1276; he remained a captive for almost a decade (Yuan-shi, pp. 144, 3082; Rašīd-al-Dīn, tr. Boyle, p. 266). In 678-79/1280 Qubilai dispatched a second expedition, but Qaidu also seized its commander, Qi Gong-zhi, in 683-84/1285 (Yuan-shi, p. 6937). Qubilai continued to encounter enormous difficulties in imposing his authority over the Tarim and Dzungar basins. He was never able to achieve economic and military self-sufficiency for the oases, yet the supply lines necessary to maintain the Yuan armies and assist friendly local inhabitants were long and fragile; furthermore, constant harassment by the elusive nomads intimidated his soldiers. His failure to gain control of the region also prevented him from reasserting his authority over his Ilkhanid relations in Persia.

In 688/1289 Masʿūd Beg died and was buried at Bukhara (iii. after the mongol invasion). His three sons succeeded him in turn, the last ruling from Kashgar. Qubilai’s successors wavered between aggressive attempts to control the Tarim and Dzungar basins and passive acquiescence to domination by the Chaghatayids and Qaidu. Between 697/1298 and 703/1303 his grandson Temür (Tīmūr; 693-706/1294-1307) conducted a series of campaigns, during which Qaidu was killed and the Chaghatayid khan, Baraq’s son Duʾa (Doʾā; ca. 691-706/1291-1306), was wounded. But by 705/1305 the Chaghatayid rulers had reasserted control over the entire Tarim basin and had expelled Qaidu’s son Čäpar (Čāpār) from the area north of the Tien Shans. The Yuan rulers accepted the status quo until 717/1317, when another of their armies temporarily expelled the Chaghatayid khan Esen Buqa (Būqā; 709-17/1309-18). This campaign was the last by the Yuan in Turkestan, however. The court in Beijing decided to focus on defense of the homeland in Mongolia and could thus not spare the necessary troops to occupy the region. One inevitable result was a decline in diplomatic and commercial contacts with Persia. By the 720s/1320s the Yuan had abandoned Turkestan to the Chaghatayids. The Chaghatayid khan Tarmašīrīn (726-34/1326-34) became a Muslim (cf. Barthold, Turkestan ³, p. 54), but the eastern branch of his house continued to resist the spread of Islam.

CHINESE TURKESTAN v. Under the Khojas

The Khojas (Ḵᵛājas, Ḵᵛājagān), descendants of the Naqšbandī Sufi Aḥmad Ḵᵛājagī Kāsānī (d. 949/1543), known as Maḵdūm-e Aʿẓam, ruled over Chinese Turkestan between 1089/1678 and 1173/1759. As sayyeds (claiming descent from the Prophet Moḥammad), they had married into some of the leading families (begs) of the Eastern Chaghatays (see chaghatayid dynasty), including the ruling Chingizid clan, thus consolidating a broad power base in the region of Altïšahr (six cities), consisting of Aqsu (A-ke-su), Uch Turfan, Khotan (Hedian), Kashgar (Ka-shi), Yarkand (Suo-che), and Kucha (Ku-che). Throughout the 11th/17-18th centuries the two main lineages of the Khojas, the Esḥāqīya (Qarataḡlïq, Black Mountain) and the Āfāqīya (Aqtaḡlïq, White Mountain), descendants respectively of the two sons of Maḵdūm-e Aʿẓam (see Table 39), extended their power, ruling over a number of city-states with puppet Chaghatayid princes as nominal rulers (Saguchi, 1963, chaps. 1-2; Fletcher, History, chaps. 3-4). In the late 1080s/1670s the Oirad (Ūyrāt) Mongols, originally from Dzungaria and western Mongolia, invaded Altïšahr; with their support in 1089/1678 Ḵᵛāja Āfāq Hedāyat-Allāh, the great-grandson of Maḵdūm-e Aʿẓam for whom the Āfāqīya lineage is named, seized the Chaghatayid throne in Yarkand (Kāšḡarī, tr. Hartmann, p. 214). His descendants continued to rule Kashgaria until the mid-12th/18th century. Their domain was relatively prosperous, characterized by irrigation agriculture dependent upon the qanāt (underground aqueduct) system originally borrowed from Persia. The Khojas’ power was, however, severely limited by their overlords, the Dzungar khans from what was then the leading Oirad tribe, who required payment of tribute and frequently insisted on hostages to ensure their vassals’ obedience. In 1168/1755 the armies of the Manchu Qing dynasty (1644-1911) invaded Dzungaria, where they found two brothers of the Āfāqīya lineage, Qïlïč Borḥān-al-Dīn and Ḵᵛāja Jahan, being held hostage (Saguchi, 1978, pp. 67-80). Although the Chinese attempted to enlist the support of the brothers, they chose to resist and fled to their own territory in the western Tarim basin. Their independence was short-lived, however, and the Chinese conquest of Kashgaria in 1173/1759 brought the rule of the Khojas to an end (Fletcher, 1982, pp. 170-71). Descendants of Qïlïč Borḥān-al-Dīn (who fled to Badaḵšān in 1173/1759) and his brother led Muslim resistance in the region until the second half of the 13th/19th century.

Table 39. The Khoja LineagesView full image in a new tab

Although an indigenous Muslim and non-Muslim Turkic literature is attested in eastern Turkestan from an early period, the earliest surviving works embodying the historical traditions of the Chaghatayids in the 10th/16th century are in Persian, for example, the Tārīḵ-e rašīdī (952/1546) of Mīrzā Moḥammad-Ḥaydar Doḡlāt (q.v.). Especially during the reign of Ḵᵛāja Āfāq (1089-1106/1679-94) the heterogeneous populations of the Altïšahr began to develop a common identity based on allegiance to Islam, especially the teachings of the Naqšbandī order. One of the most pious Naqšbandīs, Šāh-Maḥmūd b. Mīrzā Fāżel Čorās (d. ca. 1107/1696), left two important works in Persian: a political history intended as a continuation of the Tārīḵ-e rašīdī through the first three quarters of the 11th/17th century (Čorās, 1976) and a hagiography of the Esḥāqīya lineage entitled Anīs al-ṭālebīn (1107/1696). The Khojas encouraged both translation of Persian works into Eastern Turkic and a revival of literary forms original to the latter language. This diverse literary culture continued to thrive long after the dynasty came to an end. For example, the major work on Maḵdūm-e Aʿẓam, Jāmeʿ al-maqāmāt, had been written in 1026/1617 by his grandson Abu’l-Baqāʾ, who belonged to a third lineage that had remained uninvolved in the politics of the Tarim basin. In 1208/1794 Moḥammad-Ṣādeq Yārkandī began a translation of Jāmeʿ al-maqāmāt into eastern Turkic; he completed it, with a supplement on the Khojas of Kashgar, in 1235/1819 under the title Majmūʿat al-moḥaqqeqīn (ms. Orient. Oct. 1680, Staatsbibliothek, Marburg, pp. 15, 21, 154). From this series of works it is possible to assess the importance of histories and tales about Muslim saints, especially those from different parts of the Tarim basin, in forging a sense of common Islamic identity. The Khojas also founded shrines to some of these saints and established pious endowments (waqfs) to support them. The shrines became sites of pilgrimage, at which the inhabitants of different parts of the region were brought closer together. Among the most important pilgrimage centers were the Yaḡdu tombs near Kashgar, where Ḵᵛāja Āfāq (d. 1106/1694) is buried (Kāšḡarī, tr. Hartmann, p. 217 n. 2).

CHINESE TURKESTAN vi. Iranian Groups in Sinkiang since the 1750s

Between the late 17th and 19th centuries many Iranian-speaking peoples from Šeḡnān (Shughnan) and Wāḵān (Wakhan) migrated to the region of the eastern Pamirs around Lake Zorkul (Sarīqūl, also known as Lake Victoria), now in the Tadzhik S.S.R., and mingled with the nomadic groups of Iranian descent already established there. In 1168/1755 the armies of the Manchu Ching (Qing) dynasty (1644-1911) of China destroyed the Zungar khanate, then turned their attention to the oases south of the Tien Shan range (see v, above). They advanced steadily through the Tarim basin, and the Khoja rulers at the western end fled, pursued by the army as far as Badaḵšān in northern Afghanistan. A detachment commanded by General Fu-de was sent to search for them near Lake Zorkul, where the population submitted to the Ching empire without a struggle (Fu et al., chap. 77, fols. 1r-4v, 24v-25v; Kuznetsov). Although the Iranian inhabitants of the region were known to the Chinese simply as “Tajiks” (Ta-ji-ke; Mathews nos. 5978-476-3320), they probably spoke the Pamir languages Wakhi, Shughni, and Sarikoli, rather than Tajiki Persian (see central asia xiii. iranian languages). This seminomadic group is also known as “mountain Tajiks” (tājīkān-e sar-kūhī, Uighur sarqari or sarqiri; Sotūda, p. 563) and is distinct from the sedentary Persian-speaking Tajiks farther north and west. At the beginning of the 12th/18th century the mountain Tajiks were converted to Ismaʿili Shiʿism (Xiao, 1983, p. 9).

The Chinese governed the newly conquered territory through indigenous begs, who collected taxes, maintained order, supervised irrigation, and so on under the direction of Manchu officials. In the eastern Pamirs, where Tashqurghan (Ta-shi-ku-er-gan) was the most important town, seven begs were appointed, one fifth-level ḥākem beg and four sixth-level and two seventh-level begs, who were answerable to a Manchu imperial agent (pan-shi da-chen) stationed at Yarkand (Suo-che; Fu et al.,1970, chap. 30, fols. 21r-v). During the first century of Chinese rule the population of the region gradually increased. According to the report of one local administrator in 1759, the Tajiks had “formerly numbered 500 households, paying 50 taels [a unit of weight] of gold to Yarkand, but now there are only slightly more than 110 households with a population of about 400” (Fu et al.,1970, chap. 77, fols. 24v-25r). At the end of the 18th century mountain Tajiks began to settle in villages among the predominantly Turkish population around Yarkand (So-che), Poskam (Po-si-kan-mu), and Karghalik (Ye-cheng), along the old Southern Silk Road. A census probably taken in the 1850s listed 432 Tajik households in fifteen villages, paying a tax of 27.7 taels of gold and 1,700 jins of saltpeter. The number of sixth-level begs had been reduced to two and that of seventh-level begs increased to five, for a total of eight (Ya-er-jiang-cheng; cf. Hori).

Beginning in the 1820s Chinese rule in the western Tarim basin was seriously challenged by descendants of the Āfāqī Khojas, most notably Ḵᵛāja Jahāngīr in 1826; the khans of Kokand (Ḵokand) in Fergana (Farḡāna), who were descended from the Uzbeks, supported these efforts, in order to obtain a share of the trade with China (Tōru, pp. 405ff.; Kim, pp. 16-39). Ḵᵛāja Jahāngīr and several other family members made repeated incursions and sometimes even succeeded in occupying larger towns like Kashgar and Yarkand. The mountain Tajiks, though Shiʿites, and the neighboring Kirghiz Turks participated in these “holy wars” against the “infidel” (Valikhanov, III, p. 142), lured by the prospect of booty. Nevertheless, when the khanate of Kokand began actively to expand into the Tien Shans and the Pamirs, even attempting to occupy Tashqurghan, the begs and their supporters, whose prerogatives were in jeopardy, opposed their fellow Muslims (Zhong-guo, p. 233; Valikhanov, III, p. 148).

In 1282/1865 an emirate was established in Kashgaria under the control of Yaʿqūb Beg (1282-1304/1865-77), who was originally from Kokand. The mountain Tajiks suffered under his rule, and after 1868 many were forcibly resettled in Kashgar, where they could be more easily controlled and posed no further threat to security on the borders (Hayward, p. 109; Sayramī, p. 231); according to T. D. Forsyth, in 1873 “the hamlets are at present in a wretched looking state, the houses have fallen to ruin.” The number of households in Tagharma near Tashqurghan, for example, had shrunk from fifty to four (Forsyth, pp. 223, 269). According to A. N. Kuropatkin, who visited Kashgaria in 1876, Tashqurghan was at that time administered as an independent district of the emirate (p. 40). Meanwhile, after the forced removal of the Tajiks the area near Lake Zorkul was populated mainly by Kirghiz (Kuropatkin, p. 50).

Yaʿqūb Beg died at the end of May 1877, a year after the Kokand khanate had been incorporated into the expanding Russian empire. Although Ḥākem Khan Türa, one of the Āfāqī Khojas, was proclaimed his successor, severe conflict between him and Yaʿqūb Beg’s son Beg Qolī put an end to the emirate. By the end of the year the Manchus had fully reconquered the Tarim and Dzungar basins as far west as the new Russian border. At the beginning of 1879 Ḥākem Khan invaded Kashgaria with about 2,000 soldiers and succeeded in taking the Tashqurghan area, but he was expelled by the army of Liu Jin-tang, commander of the Manchu troops in the Tarim basin (Yuan, chap. 97, fols. 10v-22v). In response to the pressures of Russian expansionism the Chinese government sought to strengthen its own control of the Tarim basin by incorporating the region as the province of Sinkiang (Xin-jiang) and instituting a number of reforms, including abolition of the beg system and concomitant reorganization of the administrative structure.

In 1895 Lake Zorkol and all the territory north of the Wāḵān corridor as far east as Qoqrāš Qōl, now Povalo-Shveĭkovskiĭ peak, were incorporated into the Russian empire (see boundaries iii. boundaries of afghanistan). In 1902 the remnant of the region that remained under Chinese rule, the area around Tashqurghan, was placed under the Yarkand administrative district, and an official with the title “second-class subprefect” (tong-pan) was dispatched to administer it. In 1911, after the establishment of the Republic of China, the Yarkand district, with a total of twenty-seven villages, was designated Pu-li (known locally as Varšīda; Pakhalina, p. 3) after an ancient city-state in the region (Feng, pp. 93-94; Xiao, 1983, p. 17). The population of this remote border area was largely Iranian, in contrast to the other areas of Sinkiang, where Turks predominated.

The new Chinese government was, however, beset by serious internal conflicts and unable to exercise effective control of Sinkiang; the region was thus dominated by a series of warlords, recognized by the central government but operating virtually independently. Although the beg system was abolished in other parts of the province, it continued in Pu-li until 1926, probably owing to the peculiar ethnic character of this remote mountain region. The Tajik population grew steadily, reaching 6,169 (1,805 households) in Pu-li alone (Sung, chap, 1, fol. 11r). The first warlord of Sinkiang, the autocratic Yang Zeng-xin, was assassinated by political rivals in 1928 (Xin jiang, pp. 76-91; Forbes, pp. 33-37). He was succeeded by Jin Shuren, who showed himself unable to deal with insurrections in Ha-mi (former Qomul) at the eastern end of the basin and with an invasion by Ma Zhong-ying, a Muslim warlord from Kansu province, whose intervention was invited by the rebels (Hedin, 1936; Boorman, II, pp. 463-64). In 1933 he was succeeded by Sheng Shi-cai, a Han Chinese who had begun his career under Chang Kai-shek and who at first pursued a policy of compromise among ideological and ethnic groups (Boorman, III, pp. 120-23); by then the Tajiks, who numbered 8,867, had been officially recognized as a minority (Lattimore, et al., p. 110). In the early 1940s, however, the ruling Kuomintang (Guo-min-dang) party adopted a policy of forced assimilation of minorities and determined opposition to the spread of the Chinese communist party, a policy that Sheng pursued vigorously in Sinkiang and that was continued by his successors after his departure in 1944 (Xinjiang; Forbes, pp. 158-69). These developments resulted in considerable turmoil, particularly north of the Tien Shans. In the western Tarim basin the Tajiks organized a revolutionary force of about 200 men and occupied the Pu-li district in 1945. In 1946 they reached an agreement with the government and disbanded (Xiao, 1983, pp. 41-43), but the Kuomintang nevertheless continued its oppressive policy. The Tajiks thus joined the communist Partisans of the Red Tents, the local communist party (Lattimore et al., p. 139). In December 1949 Sinkiang came under the control of the People’s Republic of China, and in 1955 it was incorporated into the Sinkiang-Uighur Autonomous Region. In 1954 Pu-li district became Tashqurghan Tajik Autonomous District. Nevertheless, at the same time the central government decreed that only Chinese and Uighur could be taught in the schools of the region (Sotūda, p. 564). As a result about two-thirds of the present Tajik population in the area around Tashqurghan speak some Uighur, along with their native languages, Wakhi, Shughni, and Sarikoli, and Tajiki Persian. Those settled in the villages in the oases of the western Tarim basin generally speak only Uighur. Some urban Tajiks attained positions of importance in local government; for example, Zolāl, a Persian-speaking Tajik, was governor of Kashgar during the Chinese Cultural Revolution (1965-68; Sotūda, p. 564).

According to the 1978 census, the total Tajik population in Sinkiang was 22,000, of whom 60 percent were concentrated in Tashqurghan and the remainder scattered among other cities of Kashgaria (Zhong-guo, p. 230). In 1985 there were approximately 26,000 “mountain Tajiks” in China, 68 percent of them living in the Tashkhurqan governorate (Xiao, pp. 2, 8, 9, 80). The remainder were settled in autonomous communities (xiang) in the nearby districts of Yarkand (Zarepshat xiang), Poskam (Buyluk xiang), Guma (Pi-shan; Nurawat xiang), and Akhtu (Bostan xiang). According to Manūčehr Sotūda (p. 563), in 1988 there were only twelve Tajik families left in Kashgar.

Of the fifty-six nationalities in the modern Chinese population only the Tajiks and the Russians, who immigrated in modern times, are of Indo-European stock. The Tajiks in the Pamirs around Tashqurghan are seminomadic. They are monogamous; large families are typical, with three or four generations represented in a single household. Each household is a self-contained productive unit governed by the head of the family, usually the oldest male of the senior generation. The people live mainly from raising livestock, especially sheep and yaks. In this region the tail of a single donba (fat-tailed) sheep may yield 15 kg of fat. In addition, some cold-resistant crops like barley and beans are grown; a few cows are kept for milk, both camels and horses are used for transport, and oxen draw the plows (Sotūda, p. 564). Each year after the spring planting in the valleys the men take their animals up the mountains, returning to the villages once or twice during the summer in order to water and weed the crops and in the autumn for the harvest. Once snow begins to fall in the mountains, usually in late October, the animals are brought down to pasture near the villages, where the herdsmen spend the winter. The houses are usually cubical, with flat roofs, and are constructed of wood and clay. Often the interior is not subdivided into rooms; rather, clay platforms are built around the inner walls for sitting and sleeping. The most characteristic item of Tajik dress is the hat. Men wear tall black-fur hats with rounded tops, and women wear embroidered cotton caps with rounded tops and ribbons at the back.

As Ismaʿili Shiʿites most Tajiks do not perform the five daily prayers or observe Ramażān as their predominantly Sunnite Turkish neighbors do. Sometimes they prefer to pray in a jamʿ-ḵāna (communal hall), rather than in the local mosque (Sotūda, p. 564). Their religious leaders are called īšān (yi-chan; lit “they,” used to show respect). Believers from a particular household often follow one īšān and his successors. Upon a man’s death his property is often divided equally among his sons. If he dies away from home, his body must be brought back for religious burial.

The Tajiks in Sinkiang possess a rich store of folk tales, poetry, music, dances, ceremonials, and crafts. Their cultural heritage is rooted both in the traditions of the Pamirs and in the rich Persian past of Central Asia; it has also been influenced by the neighboring Turkish culture. In the past the literature of the Tajiks was predominantly oral, but since the spread of public education under the People’s Republic of China they have developed a written literature. Because official teaching is in Uighur, however, educated Tajiks generally write in that language and script.

The Tajik peasants around Yarkand, Poskam, and Khaqilik sing traditional Tajik melodies with Uighur words. In the past early marriage and marriage between cousins were common. Wedding ceremonies often last three days and usually includes singing, dancing, and the game of oqlaḵ tartišmak (diao-yang). Other Tajik celebrations include the spring festival Qie-to-qia-te-er (“to clean away dust”), the planting festival Tai-he-mu-zi-wa-si-to, and Zi-wan-er (festival of repairing canals and irrigation channels).

Tajik males greet each other by shaking hands or embracing. A younger woman kisses the palm of an older woman, while the older one kisses her cheeks and eyes. Women of the same generation usually kiss each other on the cheek.

CHINESE TURKESTAN vii. Manicheism in Chinese Turkestan and China

Historical survey. Manicheism was probably introduced into Inner Asia by Sogdian (Hu) merchants, though the process of its diffusion there is entirely obscure. It was certainly well established in Tokharistan and the Tarim basin before the 8th century c.e. (Henning, 1936, pp. 11-12). The earliest references to Manicheism in Chinese sources are of the 8th-9th centuries under the Tang dynasty (618-907), a period noted for active contacts between China and Central Asia (Schafer, pp. 407-9). According to a report in the early 17th-century Min-shu (7.32a) of He Qiao-yuan, however, the religion arrived in China during the reign of the Tang emperor Gao-zong (650-83; Pelliot, 1923, pp. 199, 203). It may be that it was formally introduced at court at that time, for in the same source there are hints that it was patronized by the empress Wu (684-704), who usurped the throne under the dynastic title Zhou (Pelliot, 1923, pp. 199-200, 203-4). Her unpopularity may have helped to stimulate conservative hostility to Manicheism. In 731 the Tang court ordered a summary translation of Manichean beliefs (see below), and in the following year the religion was banned by imperial edict (Tong-dian 44.229c). Nevertheless, it persisted among foreigners in China, mainly Sogdians, until the end of the Tang period. During the chaotic years following the rebellion of An Lu-shan (q.v.; see also ii, above, and chinese-iranian relations i. in pre-islamic times) in 755 the qaghan of the Uighur-Turkish kingdom on the Orkhon river in Mongolia, who led a force against the rebels in 762-63, was converted to Manicheism by Sogdian missionaries. A terse account of his conversion is included in the trilingual (Chinese, Sogdian, and Uighur) inscription of Karabalghasun (Sogdian version ed. Hansen; for the Chinese version see Chavannes and Pelliot, 1913, pp. 177-99). In the Chinese version a [mo]-xi-xi-de (Mathews, nos. 4388, 2426, 2506, 6162; < Mid. Pers. mahistag “elder, presbyter,” a term also found in the Compendium and designating a rank in the Manichean hierarchy; cf. Boyce, Reader, p. 11) is reported to have arrived as a pioneer missionary in the Uighur kingdom (Chavannes and Pelliot, 1913, pp. 192, 197; Pelliot, 1939, pp. 248-50).

Having gained the patronage of the Uighurs, supporters of the Tang dynasty, the Manicheans were able to establish temples in four major Chinese cities: Qing, Yang, Hong, and Yue (Chavannes and Pelliot, 1913, p. 262). They served as bases for missionary activity, and there is evidence that the religion gained converts among the indigenous Chinese (Lieu, 1985a, p. 195). Perhaps for that reason, after the Uighurs had been driven from Mongolia by the Kirghiz in 840 (see ii, above) and were no longer able to play a significant role in Chinese affairs, Manicheism was the first “foreign” religion to be banned, in 843; temples were closed and foreign priests expelled. According to the Min-shu, however, one Manichean priest escaped to Fu-tang (now Fu-qing) in southern China, whence he propagated the religion throughout Fukien province (Pelliot, pp. 206-8).

Meanwhile, the Uighurs established a new kingdom around the Turfan oasis in the eastern Tarim basin, and Manicheism continued to play an important part in Uighur religious life. According to surviving documents (Zieme; Geng), the monasteries there owned considerable landed property, from which they derived substantial rents, and they were also organized to conduct trade (Zieme). The monastic scriptoria seem to have been very active, and the quantity of fragmentary Manichean texts in Iranian and Turkish languages discovered by German archeologists at Qočo (Gao-chang) in the early 20th century belong to this period (Boyce, 1960, pp. ix-xxi). Nevertheless, Manicheism faced increasing Buddhist (See buddhism i) and Nestorian Christian (see christianity iii) competition for the favor of the Uighur rulers (Lieu, 1985a, p. 201; Geng and Klimkeit; cf. Hamilton, 1955, p. 133), and there is little doubt that it was no longer of major importance by the Mongol period (1218-1370), when Islam began to spread through the Tarim basin (see iv, above; cf. Lieu, 1985a, p. 201; Pinks, p. 114).

In southern China it seems that Manicheism, though defined as an illegal religion during the Sung period (960-1279), reappeared in the guise of an esoteric Buddhist or Taoist sect known as the “religion of light” (ming-jiao; Mathews, nos. 4534, 719), in which vegetarianism was especially emphasized. The authorities believed the Fang La rebellion of 1120-22 had been instigated by such illegal religious groups and therefore closed down Manichean meeting places and seized Manichean writings (Lieu, 1985a, pp. 236ff.; Forte), but the sect nevertheless continued to enjoy a considerable following in the region, among both the educated and the common people. The Mongol (Yuan) dynasty (1271-1368) appears to have tolerated Manicheism, perhaps owing to the advice of Marco Polo and his uncle Maffeo, who counseled Qubilai Khan (658-93/1260-94) on foreign religions (Lieu, 1980, pp. 76-79). The only extant Manichean temple in China, near Quan-zhou (Ch’üan-chou, q.v.; see also chinese-iranian relations vii. persian settlements in southeastern china), was built during this period. After the final expulsion of the Mongols by the Ming (1368-1644), however, the sect was subject to severe persecution, from which it never recovered (Pelliot; Lieu, 1980; Bryder, 1988; Lin, 1989).

Chinese Manichean texts. Almost all present knowledge of the doctrinal aspects of Manicheism in the Tarim basin and China is derived from three texts in Chinese. All were found at Tun-huang and were compiled before the end of the 9th century; they are generally known as the Traité, the Hymnscroll, and the Compendium respectively. Although documentary evidence of the Manicheans in southern China is plentiful, especially from the Wen and Fu prefectures, details of doctrine and liturgical practices are extremely scant. One official document connected with the suppression of religious sects at the time of the Fang La rebellion (Forte) contains a list of scriptures seized from the ming-jiao in Wen prefecture; the titles of some are unmistakably linked to the earlier texts recovered from Tun-huang. Despite the Manicheans’ self-protective adoption of Buddhist and especially Taoist elements, the inscriptions from their temple at Quan-zhou are unmistakably Manichean in content, though the building itself exhibits many Buddhist features (Lieu, 1985a, pp. 212-13).

The Traité is, despite its title (Moni jiao cao jing, lit. “fragmentary [Mathews, no. 6689] Manichean scripture”), a long text in an excellent state of preservation, with only a few lines missing at the beginning. It was first fully published with a facsimile by Edouard Chavannes (q.v.) and Paul Pelliot in 1911 and is frequently known as Traité Pelliot. Their transcription (including typographical errors) was reproduced in the Chinese translation of the Buddhist Tripiṭaka (Taishō, no. 2141 B, LIV, pp. 1281a16-1286a29); that text was in turn reproduced with critical notes by Helwig Schmidt-Glintzer (1987b, pp. T. 81-86). A more accurate transcription was published by Chen Yuan in 1923 (pp. 531-44), and a new collation based on a reexamination of the original photographs of the manuscript has now been published by Lin Wu-shu (1987, pp. 217-29), with the photographs.

The Traité contains a number of discourses attributed to Mani. They consist of answers to questions on cosmogony, ethics, and other matters posed by his disciple A-to (perhaps Addā, who is known to have been sent as a missionary to the western part of the Persian empire; Mir. Man. II, p. 69 [360]). The first complete discourse is focused on the aftermath of the rescue of primal man by the powers of light and the subsequent attack by the demon of greed (i.e., the prince of darkness), which led to creation of the universe and the eventual victory of the light (see cosmogony and cosmology iii. manichean). A number of parallels with the Traité have been found in Manichean texts in other languages. For example, themes like the enumeration of nights and days and the symbol of trees are found in western, especially Coptic, Manichean texts (Lieu, 1985a, p. 206). Exact parallels can also be found in Turkish Manichean fragments (Le Coq, pp. 16-21; Taisho, pp. 1282a-1285c). Werner Sundermann (1983) has discussed fragments from approximately twenty-two Parthian manuscripts that contain parallels to almost every part of the Traité. One group of these fragments includes the running head “finished is the sermon of the light-nous.” He concluded that the work was originally composed in Parthian, translated into both Sogdian and Turkish, then from one of those languages into Chinese.

Linguistically the Chinese version is important for the interpretation of a number of Parthian terms of ethical and anthropological significance, which first became intelligible through comparison with their Chinese equivalents. For example, Parthian wdyšnʾsgyft, literally “bad knowing,” is shown by its Chinese equivalent (yu-chi; Mathews, nos. 7624, 1025) to mean “folly,” rather than “sinful knowledge” or “evil intention,” as had previously been thought (Salemann, text 34 p. 11, glossary p. 72; Henning, 1943, pp. 58, 63, text h 1.156).

The Hymnscroll (Moni jiao xia-bu zan, lit. “the lower (second?) section of the Manichean hymns”; text in Lin, 1987, pp. 234-65; Taishō, no. 2140, pp. 1270b21-1279c10; Engl. tr. Tsui; Ger. tr. Schmidt-Glintzer, 1987b, pp. 11-67), now in the Stein collection of Tun-huang manuscripts in the British Library, London, contains about thirty hymns probably translated into Chinese from Parthian. Several are simply phonetic transcriptions of the original Parthian hymns and must have made little sense to the common Chinese reader; in one of them traces of Aramaic words have been detected (Yoshida). The first canto of the Parthian hymn cycle known as Huyadagmān (formerly read Huwīdagmān; Boyce, 1954, pp. 68-77; cf. Nyberg, Manual II, p. 225; MacKenzie; Sundermann, 1990, pp. 9-10) is also included, in Chinese translation. The scroll ends with an appeal for blessing, with the information that it was translated and compiled in Turfan (Schmidt-Glinzer, p. 74).

The Compendium (Moni guang-fu jiao-fa yi-lüe, lit. “outline of the teachings and rules of Mani, Buddha of light”; text in Lin, 1987, pp. 230-33; Taishō, no. 2141 A, pp. 1279c17-1281a11; Eng. tr. of pp. 1279c17-1280c12 in Haloun and Henning; Fr. tr. of pp. 1280c12-1281a11 in Chavannes and Pelliot, 1913, pp. 107-16; Ger. tr. of entire text in Schmidt-Glintzer, 1987b, pp. 69-75). This manuscript has been divided; the main portion is in the British Library, and a large fragment, containing the concluding sections, is in the Bibliothèque Nationale, Paris (Lin, 1988). According to the first paragraph, the work was translated in response to an imperial order issued 16 July 731. The Compendium contains a summary of Manichean doctrines, beginning with an account of Mani’s birth that is clearly modeled on that of the Buddha and has no known parallel in Iranian or western Manichean texts. It also contains a long passage from the Taoist polemic Lao-zi hua-hu jing (Lao-tzu converts the barbarians; see, e.g., Schmidt-Glintzer,1987b, p. 71), in which Mani was depicted as an avatar of Lao-tzu, the traditional founder of Taoism; this “scripture” was a focus of controversy between Taoists and Buddhists in China and is unlikely to have been translated directly from one of the Central Asian language. The association of Mani with Lao-tzu was probably partly instrumental in the survival of Manicheism in China after the Tang period.

CHINESE TURKESTAN viii. Turkish-Iranian Language Contacts

Contacts between the Iranian peoples and the Turks are known to have occurred at least as early as 552 C.E., when the ancient Turks spread out from their northern settlements and established an empire extending from the Greater Khingan mountains on the western border of Manchuria to the Aral Sea (Herrmann, pp. 31, 35) and including territories of the Choresmians (see choresmia) and Sogdians farther west. Subsequently Buddhism and Manicheism became the predominant religions among both Turks and Iranians in what is now Chinese Turkestan. Three general phases in the development of the Old Turkic language can be distinguished in the area of east of 80° E, where the earliest texts have all been found: the very archaic language of the 8th-10th centuries (Old Turkic A, sometimes erroneously called Uighur), that of the 10th-12th centuries (Old Turkic B, Old Uighur), and the language that has been evolving since the Mongol period in the 13th-14th centuries (Old Turkic C, Late Uighur). The texts in Old Turkic A are predominantly pagan and Manichean, though some early Buddhist influences are apparent; those of Old Turkic B and C are predominantly Buddhist, though there are some Manichean survivals. Within this framework, however, the language of the documents does vary somewhat with the contents: Even in the third phase religious texts contained a great many archaisms, whereas private documents—for example, sales contracts—were linguistically more advanced, sometimes more advanced than religious documents that are demonstrably later.

In the Tarim basin Iranians were living in and around Khotan on the Southern Silk Road and in the areas of Maralbashi (Pa-chu) and nearby Tumshuq on the Northern Silk Road (see ii, above); in addition, probably as early as the 1st century Sogdian trading colonies had been established along both routes as far east as China (Laut, p. 4; Bazin, pp. 40-41, 44; cf. Pelliot, 1912; idem, 1916-17; Barthold, 1956, p. 88; Gabain, 1961, pp. 15-16; idem, 1974, pp. 245-46; Pulleyblank; Spuler, p. 130; Ḥodūd al-ʿālam, tr. Minorsky, p. 225; Klyashtornyĭ). There were also Manichean refugees from Persia, who spoke Middle Persian and Parthian. Contacts between the Turks and these Iranian groups left traces in the respective languages spoken before the Muslim conquest of the area, which took place shortly after 1000. It is not always certain from which Iranian languages loanwords came into Old Turkic A and B, either because a specific word was common to several Middle Iranian languages (and even non-Iranian languages like Tokharian) or because it is not directly attested in any of those with which the Turks had contact. Recently, however, J. P. Laut has analyzed the features that distinguish Sogdian and Tokharian loanwords in the oldest level of the basic vocabulary found in Old Turkic Buddhist texts.

Old Turkic and Sogdian. The earliest surviving Turkic texts, of the 8th-10th centuries, are in Old Turkic A. The influence of this language on Sogdian appears to have been largely limited to names and titles (e.g., tu’un, qaγan, tarxan; cf. Doerfer, II, pp. 460-74), found chiefly in Sogdian documents from Mount Mug and Marv (Mary) in the Tadzhik and Turkmen S.S.R.s respectively; these documents are datable to 712-15 (Freĭman; Livshits; Bogolyubov and Smirnova; Gabain, in Camb. Hist. Iran III/1, p. 615). The Turks, on the other hand, received their writing systems (except the runic script, the origin of which is disputed) from the Sogdians, particularly the Old Sogdian and Sogdian cursive (or Uighur) scripts, which, according to W. B. Henning, were developed during the 7th century (“Mitteliranisch,” pp. 55-56; cf. Gabain, 1964, pp. 180-81; idem, 1973, pp. 24-25; Sims-Williams; Laut, pp. 5-6; Bazin, pp. 39-40; cf. Bazin, p. 43). The Turks began by copying Sogdian (and Tokharian) texts (Tezcan, 1978, pp. 286, 315; Gabain, 1974, p. 246). In fact, the oldest surviving inscription dealing with the history of the Turkish emperors, a funerary stele found 10 km west of Bugut, in North Khangai province of the People’s Republic of Mongolia (Klyashtornyĭ and Livshits, 1972; Čağatay and Tezcan; Laut, pp. 3-4; Tezcan, 1978, p. 278; Gabain, in Camb. Hist. Iran III/1, pp. 617, 621; Bazin, p. 43), is written, not in Old Turkic, but in Sogdian language and script. It is datable to 581 and thus confirms the statement in one 6th-century Chinese source that the first script employed by the Turks was Sogdian (Bazin, pp. 39-40; Laut, p. 5), rather than runic. The only Turkic words in the inscription are titles (e.g., γʾγʾn = qaγan “emperor,” tykyn = tegin “prince,” tw’wn = tu’un “commissioner”; see respectively Doerfer, III, pp. 141-79, II, pp. 533-41, III, pp. 207-10; the ending -n is characteristic of Tavghach [Tuo-ba, Wei], a Mongolian language). This inscription is also important for revealing the actual forms of the names of the emperors, formerly known only from their imprecise Chinese transcriptions (e.g., Tʾspʾr = Taspar or Tasbar, Nwʾʾr = Nevar or, according to the Chinese sources, ñebar, with the suffix -bAr, typical not only of the Old Turkic title eltäbär but also of the names of the Juan-Juan emperors).

With the possible exception of a small inscription found near Choyren in Mongolia, which may have been written between 688 and 692, the earliest inscriptions in Old Turkic belong to the first half of the 8th century (Bazin, pp. 43-44). The short and defective inscription datable to 762 on a stele from Sevrey in the southern Gobi desert (Klyashtornyĭ and Livshits, 1971; Tezcan, 1978, p. 280) is in both Old Turkic A (in runic script) and Sogdian. Another, from Karabalghasun on the left bank of the Orkhon, is dated 821 (or 810, cf. Tezcan, 1978, p. 280; or ca. 820. cf. Bazin; or 808-21, cf. Chavannes and Pelliot, 1913, p. 383; Henning, “Mitteliranisch,” p. 56) and includes a description of the conversion of Bögü Qaḡan to Manicheism (see iii, above) in 762-63 (Klyashtornyĭ and Livshits, 1971, p. 14; ); it is written in Sogdian in cursive script (text and tr., Hansen, 1932; rev. text and Jap. tr. in Yoshida; cf. Inscriptions, pp. xxi-xxiv, xxvii-xxxviii, 24-25, tables 44-63), in Old Turkic A in runic script (Chavannes and Pelliot, 1913, pp. 177-99, pp. 229-31; Orkun, I, pp. 229-31), and in Chinese (Chavannes and Pelliot, 1913, pp. 177-99; Orkun, pp. 240-46).

Certain differences between dialects had already appeared in the Turkic of the runic inscriptions, for example, bän, män “I,” barḍĭγĭz, bardïŋïz “you (pl.) went,” and such differences became even more pronounced with the passage of time. The linguistic influence of Iranian, and especially Sogdian, is most apparent in Buddhist and Manichean texts in Old Turkic, for example, a Manichean text edited by Wilhelm Bang and Annemarie von Gabain, which also describes the conversion of Bögü Qaḡan; although it is written in archaic Old Turkic A, the specialized Manichean terminology includes some typical Sogdian loanwords, for instance, dindar “elect,” n(i)γošak “auditor,” γuan “sin,” bačaγ “fasting,” možak “bishop.” Two words from the old Iranian religions, originally adopted by the Manicheans and transmitted through Sogdian Manichean texts, are uštmaq “paradise” (Manichean Sogd. wštmʾx, cf. Av. vahištəm ahūm, lit. “best existence”) and šimnu “demon” (Manichean Sogd. šmnw; cf. Av. Aŋra Mainiiuš, Mid. Pers. Ahriman). In secular contexts there are Sogdian loanwords like boṛč “debt” (< pwṛč), maraz “hireling” (< mrʾz), šük “silent” (< šwk), and känd “town” (< kʾn’). Sogdian names for the planets are listed in one Old Turkic text in runic script (Andreas). A significant number of loanwords are actually not of Sogdian origin, though their forms in Old Turkic show that they were borrowed from Sogdian, for instance, tamu “hell” (< Sogd. tm- < Skt. tamas), nirvan “nirvana” (< Sogd. nyṛβʾn < Skt. nirvāna), and nom “religious law” (< Sogd. nwm < Gk. nómos).

Most of the Sogdian loanwords in Old Turkic disappeared with the Islamic conquest of the western Tarim basin in about 1006, and only a few survive in the Turkic languages spoken in the region today (cf. Mansuroğlu, p. 108; Asmussen, p. 127; Menges, pp. 168-72; Gabain, 1977; idem, in Camb. Hist. Iran III/1, pp. 617-24; Laut, pp. 89-148). In one Old Turkic Manichean manuscript written in archaic language č(ä)r(i)g türk uluš Aṛγu T(a)las “the warlike noble (or “Turkish”) tribe Aṛγu Talas” and altun Aṛγu uluš “the "golden" Aṛγu tribe” are named (Röhrborn, I/3, p. 178). The Aṛγu were also mentioned in the 5th/11th century by Maḥmūd Kāšḡarī (I, pp. 84, 151), who described them as a tribe living between Asfījāb, Talas (Ṭarāz), and Balāsāḡūn and speaking both a special dialect of Turkic and Sogdian, from which he quoted the word maraz (I, p. 312). According to Kāšḡarī’s information, in the Aṛγu dialect of Turkic daγ meant “is not” (II, p. 227), and n occurred instead of y (< Old Turk. ñ [I, p. 84], e.g., qoñ “sheep” > qon). These features appear to have survived in modern Khalach, a group of Turkic languages spoken in western Persian villages lying between Sāva, Qom, and Arāk (Doerfer, 1987) and containing a large number of words of obscure origin, perhaps originally from a dialect of Sogdian (e.g., Khalach bēšmi “thread” < *brēšumī < Sogd. *brēšum “silk”?).

Old Turkic and Khotanese. In the Khotanese documents of the 9th-10th centuries Turks (ttrükä), Uighurs (hvauhū:rä), and Oghuz (ūhūysä) are mentioned, and Turkic proper names, titles, and ethnics appear frequently (Bailey, 1939; idem, 1985, pp. 101-9; Hamilton, 1958; idem, 1977a; idem, 1977b; Hovdhaugen). One manuscript (Bibliothèque Nationale, Paris, Pelliot 2892, ll. 165-84) even contains a Turkic-Khotanese vocabulary of terms for body parts, archery, and horse trappings (Bailey, 1943-46; Hovdhaugen; Clauson, 1973). In contrast to most Old Turkic texts, which were written in runic script, in which not all the vowels were distinguished (with the important exception of the Manichean Turkic texts in Northern Brāhmī script; cf. Gabain, 1954), Turkic words in the Khotanese texts are written in Southern Brāhmī, an Indian alphabet in which minute phonological details are recorded. These texts are thus important in establishing the phonology of Old Turkic, especially the vowel system. For example, Old Turkic bwdwn (runic b1wd1wn1) “people” was earlier interpreted as bu’un, but the Khotanese transcriptions all have ā in the first syllable, which in late Khotanese (9th-10th centuries; Emmerick, pp. 215-46) represented /o/, and ā, au, or ū in the second syllable; these features suggest the pronunciation bo’un or bo’on, with assimilation (Hovdhaugen, p. 182). No Turkic loanwords are found in Khotanese, however, and there are at most only a few Khotanese words in Turkic, which is not surprising, as Turks and Khotanese did not come into direct contact until the late 9th or 10th century (see, e.g., Hamilton, 1986, I, pp. xiv-xvi). Some words previously thought to have been borrowed from Khotanese have now been shown to have other origins, for example, tōn “garment, clothing” (Khot. thauna-; see, e.g., Menges, p. 171), which is probably genuine Turkic (Clauson, 1972, p. 512; Doerfer, IV, p. 450; Tezcan, 1978, p. 321). Two Buddhist terms, ultimately from Sanskrit, were actually borrowed through Tokharian, rather than Khotanese: sažan, sažin “religious discipline” (Toch. śāsaṃ, Khot. śāsanā- [pronounced źāsana] < Skt. śāsana; Gabain, in Camb. Hist. Iran III/1, p. 618; Menges, p. 171; Toch. łššȧ; Laut, p. 138) and tužit “one of the heavens” (Khot. ttuṣita- [pronounced tuẓida] < Skt. tuṣita; Toch. tuṣit, Laut, p. 140).

Old Turkic, Middle Persian, and Parthian. In what is now Chinese Turkestan Middle Persian and Parthian were widely used as ecclesiastical languages, especially among the Sogdians (Henning, “Mitteliranisch,” p. 76; Boyce, 1968, pp. 68, 72; Hansen, 1968, pp. 93-94), and West Iranian religious formulas are found even in Old Turkic texts in runic script (Le Coq). On the other hand, the late Middle Persian Manichean texts (8th-9th centuries) contain many Turkish names and titles; in the so-called Mahr-nāmag (Müller), for example, there are long Turkic phrases like Ay täŋridä qut bulmïš alp bilgä Uyγur qaγan “the valiant, wise Uighur emperor who has found hail (grace) from (or at) the moon god (Jesus)” (cf. the same title in the Karabalghasun inscription; for other texts containing Turkic names and titles see Boyce, Reader, text dw; idem, 1960, nos. 158, 1573, I B 6371; for possible Middle Persian loanwords, e.g., yazd “god,” cf. Menges, p. 170, though some of those suggested may be Sogdian and qamaγ “all” is genuine Turkic).

H. W. Bailey has proposed Iranian etymologies for a number of Turkic words and names (e.g., 1985, pp. 104-9), though they are somewhat speculative. A mixed group of Taraṇči and Kāšḡarlık presently inhabiting parts of western China, the Soviet Union, and Afghanistan have been known as “Uighurs” since the 13th/19th century, though they are not descended from the Uighurs of the eastern Tarim basin (Barthold, “Taraṇči”; idem, “Turks”). The modern Turkic languages spoken by them contain very few Iranian elements antedating the Islamic conquest.