History of Xinjiang: Central Eurasia Ascendant (9th-16th Centuries)

Eurasian Crossroads: A History of Xinjiang. Editor: James A Millward. New York, NY: Columbia University Press, 2007.

Long-lived, territorially vast empires are convenient for mapmakers and students of history. One pair of dates and a single name can cover a big chunk of time and is easily represented by a broad monochrome space on the map. By contrast, eras of invasions, migrations and political flux are hard to grasp. Space and time are repeatedly divided and subdivided, and there are the names of multiple ephemeral rulers and dynasties to be recalled. Just compare a historical map of the Roman empire at its greatest extent—an easily legible band of solid colour encircling the Mediterranean—with that of the post-classical Europe and the Völkerwanderung, in which a tangle of arrows representing barbarian invasions writhe like serpents across a patchwork quilt.

During the centuries covered in this chapter, nomadic powers in this era enjoyed their greatest advantage over the agrarian states of the Eurasian rimlands. In the east the Uyghur, Tangut, Khitan and Jurched all established large states incorporating agrarian areas; in the west the Qarakhanids, Ghaznavids and Seljuks did the same. Next the Mongols outdid them all with an empire that spanned the continent. Chinggis (Genghis) Khan’s successors instituted an enduring pattern of Central Eurasian aristocracies ruling over pastoralist and agrarian areas from Iran to China. Timur (Tamerlane), himself a Turk, temporarily reprised the western Mongol empire.

For all their might, however, Central Eurasian nomad powers were fractious. Their customary acceptance of either lateral (to brothers, uncles or cousins) or patrilineal succession, depending on who won the political and military contest to inherit the khanship, ensured any number of bloody transitions and political fragmentations. Besides internecine struggles, campaigns of conquest and wars between tribal confederations displaced states and ruling élites with relative rapidity. Even states with a long track record, such as that of the Uyghurs based in the region of modern Turfan, shifted through several complex political relationships with the dominant powers in the neighbourhood.

The events covered in this chapter, then, are complex. The names of rulers, dynasties and places flash by bewilderingly. Perhaps for this reason, many accounts of Xinjiang and Central Eurasian history skip or radically abbreviate coverage of this period. I have provided somewhat more detail, however, because this period exemplifies the principal themes of this book: the region’s geography made the Tarim Basin a prize over which nomadic powers based in Mongolia, Zungharia or Semirech’e struggled, and from which powers in China hoped to interdict them. The shifts in ruling power were more rapid than changes in underlying demography and culture, but nonetheless left their mark. The Tarim population became linguistically and, arguably, genetically more Turko-Mongolian in this period; it also became Islamic. Though the names of states pass by in quick succession—Qarluq, Qarakhanid, Qara Khitay, Mongol, Cha-ghatayid, Moghulistan, Yarkand khanate—by the end of this period some more familiar names have mounted the stage: Uyghur, Kazak, Kirghiz, Uzbek. Though these names do not quite signify what they will come to mean in modern times, their emergence does show that it was in the aftermath of the Mongol empire that the seeds of modern religious and national identities began to be sown.

The era of the Tibetan, Arab, Tang and Turk rivalry in Central Asia bequeathed a complex legacy to Xinjiang and world history. The Tibetan empire left its mark on frescos in Dunhuang and some architectural artefacts in the southern Tarim, including a fortress on the Khotan River at Mazartagh and ruins at Miran. Although the Arabs did not penetrate Xinjiang in the eighth century, Islam would spread westward from Central Asia into Xinjiang over subsequent centuries. Tang military farms and settlements in the Turfan area and close communications with China during the Tang period left an enduring stamp upon local culture and administration in eastern Xinjiang. Besides the imports to China mentioned above, trade and other exchanges between Tang and Central and Western Eurasia also left cultural traces in Central Asia and the West; just one example would be the famous Tang three colour glaze (sancai) of which versions may still be seen in pottery across Eurasia and even as far west as Eastern Europe. Chinese coins continued to circulate extensively in Xinjiang. Although the direct influence of the Tang military garrisons in Zungharia and Transoxiana was short-lived, associations with north Chinese states continued to be a source of prestige even for later Turkic empires. The khans of both the Qarakhanid and Qara Khitay empires (see below) referred to themselves as Chinese emperors—though the terms they used for ‘Chinese’—Tabghach and Khitay—were actually names of Inner Asian conquerors who had established states in north China.

The greatest influence of this period was that of the Türks, politically, genetically and linguistically. The Türk empire and its dissolution led to one of history’s great Völkerwanderung, a movement of Turkic-speaking tribes into Xinjiang and across Central Eurasia. The Tang use of Turkic soldiers likewise sped the westward Turkic migrations. Whereas much of Central Eurasia, including the oases of what is now Xinjiang, had been Indo-European speaking, the seventh century saw the beginnings of a process of linguistic Turkicisation that would ultimately replace Tokharian and Soghdian in the Tarim Basin with Turkic languages and create the modern patchwork of Turkic and Iranic speakers in Transoxiana. Steppe empires after the Kök Türks, even if their élites were not Turkic, were made up largely of Turkic tribes. Much more than was the case with the Xiongnu, the memory of the Türk khaghanate would have a momentum of its own, inspiring subsequent states to emulate it and adopt certain of its symbols of legitimacy. Many peoples and now states across Eurasia have continued to identify themselves as Turks, despite the rise of Islam and of the later Mongol empire, which greatly surpassed the territorial extent of the Türk khaghanate.

The Uyghur Kingdom in Xinjiang

The Uyghurs, the next major conquering power on the Xinjiang scene, originated, like many of their predecessors, in the Mongolian core lands of the Orkhon river valley (see Map 2). The tribes known by the term Uyghur (and other names, including Toqquz Oghuz, ‘the nine tribes’), were former components of the Turk khaghanate, spoke and wrote a language virtually identical to that of the Turks before them, and were at first somatically Mongoloid—that is, they resembled other eastern Altaic peoples with epicanthic fold and sparse facial hair, and not the high-nosed, bearded Iranian peoples who were then still the primary inhabitants of the Tarim, and who remain a major component of Xinjiang’s historical population. The Uyghur patrons depicted on the walls of temple 9 at Bezeklik near Turfan probably give a good idea of what the Uyghur aristocracy in Xinjiang initially looked like, before centuries of intermarriage with local peoples of eastern Iranian stock. (A twelfth-century Chinese envoy would describe the oasis inhabitants of the Uyghur region with ‘curly hair, deep-set eyes, straight and thick brows, many have curly beards’).

The Uyghurs of the Uyghur khaghanate were primarily nomads, though they also built impressive cities, such as their capital, Ordu Baliq (literally,’Royal Camp Town’), and some engaged in agriculture. They were not Muslims, but rather Manichaean. Later, the Uyghur state in eastern Xinjiang tolerated Buddhism and Christianity among its urban population—it opposed Islam. The Uyghur aristocracy of the Uyghur khaghanate who migrated to the Turfan Basin, therefore, while certainly among the ancestors of today’s Uyghur people, are not their only ancestors. They and other Turkic migrants compose some part of modern Uyghurs’ genetic makeup; the Uyghurs from the Uyghur khaghanate were also some cultural distance away from the sedentary agriculturalist Muslim Turki of the Tarim oases in the eighteenth and nineteenth centuries who came by the twentieth century to call themselves ‘Uyghur’.

As noted above, the Uyghurs, component tribes of the Türks, joined the Qarluqs and Basmils in overthrowing the Eastern Türk khaghanate.

The Qarluqs and the Uyghurs then displaced the Basmil khan, and in 744 the Uyghurs drove the Qarluqs west and established their own Uyghur royal house at the head of an empire based in central Mongolia and extending into north-west China, parts of Zungharia and at times as far west as Ferghana. This Uyghur khaghanate provided the military aid that allowed the Tang dynasty to survive the An Lushan rebellion, but exacted a heavy price for it: in the course of fighting An Lushan, the Uyghurs sacked and plundered the Chinese city of Luoyang, even burning the temples in which people sought refuge. Later, for decades they maintained an extortionate trade of horses for Tang silk, exchanging one horse for an average of 38 bolts of silk. An Arab traveller describes the Uyghur kings tent as made of gold and pitched on top of his castle; this monarch, he relates, received annual tribute from China of 500,000 pieces of silk.

The Soghdians were deeply involved in the rise of the Uyghur state. Their commercial networks linked north China, Mongolia and the Xinjiang regions, in all of which the Uyghurs had interests. Moreover, not only did the Soghdians work as middlemen in the silk-horse trade with the Tang, but they provided the Uyghur khaghanate with a cultural and administrative model critical to its transition from tribal power to far-flung empire ruling both nomadic and settled subjects. This cultural package included the Soghdian script (itself ultimately derived from Aramaic), which was adopted for writing the unrelated Uyghur language, along with such loan vocabulary as was necessary to imperial administration over oases and to life at a fixed court. It also included the systematised religion of Manichaeism. In 762 or 763 the Uyghur khaghan Bôgö (Ch. Mouyu) converted to this faith after contact with Soghdians in Luoyang. Many Soghdians served the Uyghurs in administrative capacities.

It is interesting that, given the choices available to them, the Uyghurs chose for their cultural and political infrastructure not the Chinese but the Soghdian model, despite (or perhaps because of?) the Uyghur relationship with the Tang as military ally and supplier of horses. The memory of how the Eastern Türks had been co-opted and incorporated by the Tang was no doubt still fresh; indeed the Turk khaghan Kül Tegin had only thirty years earlier carved into his memorial stone in the Orkhon valley an explicit caution to the Turk people not to be co-opted by the southerners. Beware the ‘sweet words and soft materials’ of the Chinese, he wrote. It was a warning the Uyghurs heeded.

In 840 the steppe capital of the Uyghur state was itself destroyed and the tribes scattered by yet another tribal confederation, the Kirghiz (Qirghiz, Kyrgyz), who moved south from the area of today’s Tuva and after twenty years of conflict ousted the Uyghurs from the Orkhon River valley. Juvaini, the Persian historian of the Mongol empire, described the Uyghur diaspora in vivid terms:

The tribes and peoples of the Uighur, when they listened to the neighing of horses, the screaming of camels, the barking and howling of dogs and beasts of prey, the lowing of cattle, the bleating of sheep, the twittering of birds and the whimpering of children, in all this heard the cry of ‘köch, köchF[‘move, move!’] and would move on from their halting-place. And wherever they halted the cry of’ koch, köch! would reach their ears. Finally they came to the plain where they afterwards built Besh-Baligh, and here that cry was silenced.…

Some Uyghurs fled south to China; others moved to the Gansu/Qinghai border region with Tibet, where they established a kingdom that survived until absorbed by the Tanguts around 1030. Another group of tribes migrated to north-eastern Xinjiang, establishing a state centred on Beshbaliq as their winter capital (where the nomadic tribesmen remained to pasture their herds all year) and Qocho as their summer capital. Qocho’s impressive ruins of houses, palaces and temples may still be seen today (the site of this ancient Uyghur capital near Turfan is also known as Gaochang, Huozhou and Qarakhoja). The domains of the Qocho Uyghur state extended east as far as Hami and west to Kucha.

Enduring from the ninth to the thirteenth century, the Qocho Uyghur state was longer-lived than any previous imperial power in the region. In the 1130s it accepted the Qara Khitay (Western Liao; see below) as suzerain, sending Uyghur royal family members as hostages to the court of this new dynasty. In 1209 it submitted promptly to the rising Mongol empire, thus assuring its own continued local authority until the 1370s, when it was finally destroyed and incorporated by the Mongols. Even as vassals, however, the Uyghurs exercised a strong cultural influence upon Chinggis Khan’s empire. By then a firm amalgam of Inner Eurasian oasis and steppe traditions, the Uyghurs provided the Mongols with a writing system and officials to employ it—just as the Soghdians had earlier done for the Uyghur khaghanate. A Uyghur taught Chinggis Khan’s sons to read and write, in the Uyghur script.

The Uyghur state initially resembled the paradigmatic form we have already noted in Xinjiang’s history: a Turko-Mongolian nomadic power ruling Indo-European oasis agriculturalists indirectly from across the Tianshan. Yet the Uyghurs reached south far enough to maintain an administrative capital (Qocho) in the Turfan Basin. Over time, moreover, the populations and cultures of nomad ruler and oasis ruled blended; religious, political and cultural influences from Soghdia, India and China were incorporated as well. The Qocho Uyghur rulers dropped the title ‘khan’ in favour of idiqut (iduqqut), a term meaning ‘sacred majesty’. This state continued to enjoy good relations with China-based dynasties of the Tang, the Five Dynasties-period northern states, the Song, the Liao and the Jin, with whom they and the ubiquitous Soghdian merchants traded a variety of pastoral, agricultural and mineral products. Frescos at Bezeklik (near Qocho) and Qizil (Kucha area), lands under Uyghur control, provide evidence of Uyghur patronage of Manichaeism, Buddhism and Nestorian Christianity. These same paintings reveal, in their variety of religious imagery, dress and hairstyles (including braided queue, coiled bun, bobbed or loose and long styles) a complex and diverse society. Owing to the warm climate and the availability of run-off from the Tianshan, Qocho’s agriculture flourished, allowing double-crops of grain, as well as the production of a wide variety of fruits and vegetables. Cotton had been grown in the region for centuries, and as this fibre was not yet prevalent in China, cotton cloth was among the exports to the east. Grapes, then as now, were a local specialty, and wine was bartered and collected by the Qocho government as tax.

A description from a Song (Chinese) envoy who travelled to Qocho in 984 gives a taste of the city’s diversity and prosperity:

There is no rain or snow here and it is extremely hot. Each year at the hottest time, the inhabitants dig holes in the ground to live in.… The earth here produces all the five grains except buckwheat. The nobility eat horseflesh, while the rest eat mutton, wild ducks and geese. Their music is largely played on the pipa and harp. They produce sables, fine white cotton cloth, and an embroidered cloth made from flower stamens. By custom they enjoy horseback riding and archery.… They use the [Tang] calendar produced in the seventh year of the Kaiyuan reign (719).…They fashion pipes of silver or brass and channel flowing water to shoot at each other; or they sprinkle water on each other as a game, which they call pressing out the sun’s heat to chase off sickness. They like to take walks, and the strollers always carry a musical instrument with them. There are over fifty Buddhist temples here, the names inscribed over their gates all presented by the Tang court. The temples house copies of the Buddhist scriptures (da zang jing) and the dictionaries Tang yun, Yupian and jingyun. On spring nights the locals pass the time milling about between the temples. There’s an ‘Imperial Writings Tower’ which houses edicts written by the Tang emperor Taizong kept carefully secured. There’s also a Manichaean temple, with Persian monks who keep their own religious law and call the Buddhist scriptures the ‘foreign Way’ …. In this land there are no poor people; anyone short of food is given public aid. People live to an advanced age, generally over one hundred years. No one dies young.

Worth noting in this description, besides what seems a rather pleasant way of life, is the continuing respect for Buddhism and the legacy of Tang-era Chinese rule, even under a nomadic Uyghur élite with its own religion and customs.

A Persian geography from the same period, though confused about some details, gives a similar overall picture of the Uyghur or ‘Toghuzghuz’ (Toqquz Oghuz) kingdom to the west of ‘Chinistan’ (China). The source, the Hudud al-’Alam, describes the Toghuzghuz kings as a warlike people, once rulers of all Turkestan (i.e. roughly the lands of the Türk empire), who lived as nomads on the northern steppes, lands that produce musk, furs and horn. Their capital Qocho, interestingly enough, is here called ‘Chinese town’; it is pleasant in winter but terrifically hot in summer (something any summer tourist to Turfan can attest to). The ‘Five Villages’ (a standard term indicating five principal Uyghur towns) were inhabited by Soghdians, among them Christians, Zoroastrians and heathens—perhaps indicating Buddhists? A variety of Soghdian Christian texts have been discovered in the Turfan Basin.

To the south and east of this Uyghur kingdom the Tibetan empire was crumbling, and by the tenth century the Uyghurs definitively wrested the northern rim of the Tarim Basin and the Tianshan range from their influence. By 938 the city-state of Khotan resumed sending its own diplomatic missions to the Chinese court. However, by the 990s the Tangut (Xixia) state had spread westward from its base in the Ordos oxbow of the Yellow River, and was asserting itself in the Gansu corridor and Qinghai. By the 1030s it had absorbed the Gansu Uyghur state, including perhaps Dunhuang. Although the scanty records of this strongly Buddhist state reveal that Uyghur monks and even princesses were present in the Tangut lands, this ‘Great State of White and High’ obstructed Qocho communication with the China-based courts of Song, Liao and Jin.

The Qarakhanids

To the west of the Qocho Uyghur polity a new confederation emerged in the ninth century from the welter of Qarluq, Yaghma and other Turkic tribes driven west by the rise and fall of the Orkhon Uyghur state. These tribes ultimately formed an empire in Kashgaria, Semirech’e and Tran-soxiana, and are known to modern scholars as the Qarakhanid dynasty.

To understand the rise of the Qarakhanids one must appreciate the politico-religious importance of Turkic royal titles, as well as of the imperial heartland territories, particularly the Ötükän forest mentioned in the Orkhon inscriptions and the Western Turk capital of Balasaghun, on the Chu River near Issyk Kul. Before they overthrew the Turk state the Uyghur and Qarluq rulers had styled themselves yabghu—a term that implies descent from the Turk royal clan (Arshila/A-shi-na). The Turk rulers themselves, on the other hand, were khaghans (khans), a term that became reserved for those controlling these sacred refuges of the Eastern or Western Turks. After establishing their own Orkhon-based imperial rule, the Uyghurs themselves could become khaghans. After the Kirghiz drove them from the sacred Orkhon heartland, however, the Uyghur rulers no longer used the khanal title. Nor did the Kirghiz adopt the term. However, the Qarluqs and their confederated tribes did assume the status of rightful imperial successors of the Kök Turks, and their rulers adopted the title khan (khaghan) by virtue of their occupation of the western sacred refuge on the Chu River. When the royal family split into two branches, one line of khans became the arslan, or lion khans, and the other the bughra, or bull camel khans. In Islamic sources their polity was known as al-khaqaniyya (the Khaghanal house), or al-khaniyya (the khanal [kings]). Interestingly, the Qarakhanids also employed the title ‘Tamghaj (Tabghach) Khan’, meaning’khan of China’.

As during the previous era, then, the region we now know as Xinjiang was from the ninth through the thirteenth centuries once again divided into separate spheres, with the Uyghur state in the north-east ruling Beshbaliq and the northern Tarim cities as far west as Kucha; the Tanguts controlling the south and south-east; and the Qarakhanids ruling the west.

Islamicisation of the Turks

Through the Qarakhanids the western Tarim became closely linked to the Islamic world of Transoxiana. While the Uyghur state provided today’s Uyghurs with their ethnonym, the Qarakhanids were largely responsible for their religion. A quasi-legendary account from the Kashgar area attributes the Islamicisation of the Turks to the following event. A brother of the ruler of the Irano-Islamic Samanid state in Transoxiana is said to have sought asylum with Oghulchaq, a member of the junior, Bughra branch of the Qarakhanids ruling in Kashgar. The Bughra Khan installed this eminent Samanid refugee as Governor of Artush, an important caravan station just outside Kashgar. There the new governor built a mosque. Somewhat later Oghulchaq’s nephew Satuq was inspecting caravan goods in Artush when he noticed the Muslim merchants stopping all business to pray as soon as they heard the muezzin’s call. Satuq, profoundly moved by this sight, began studying the Qur’an himself. (In another version, Satuq while out hunting one day encounters a talking rabbit who turns into a shaykh and induces him to repeat, ‘There is no god but God, and Muhammad is his prophet.’)

After his conversion Satuq fought and defeated his uncle the Bughra Khan, and then, perhaps with Samanid assistance, took Balasaghun and displaced the senior Arslan khans. Satuq Bughra Khan promoted Islam among his subjects. According to Islamic sources, Islamicisation in Xinjiang achieved a breakthrough in 960, five years after Satuq’s death, with the conversion of’200,000 tents of the Turks’. Satuq Bughra Khan remains a revered figure in the Artush and Kashgar area, where his tomb, associated college and shrine, was rebuilt and endowed by Ya’qub Beg (see Chapter 3) and is now again under reconstruction.

This heroic tradition notwithstanding, it is likely that the conversion of the Qarakhanids took place more gradually. A Muslim scholar, Kalimati, in the Qarakhanid court around this time may have played a role; also of likely significance were shaykhs and mystics on the steppes, and the penetration everywhere of Muslim merchants.

The conversion of the Turks to Islam was an event of world-historical significance, for the Qarakhanids went on to destroy the Samanid dynasty (1000) and assume control of firmly Muslim Transoxiana. They would be the first of several Turkic ruling dynasties in the Islamic world. From the perspective of modern Uyghur nationalism, however, the discrete histories of the Qarakhanids and the Uyghur Qocho state present a problem. While the modern Uyghur people take their name from the Uyghur empire and Qocho state (known as Uyghuristan in Islamic sources), it was the Qarakhanids in northern Xinjiang and the western Tarim Basin, including what is today the quintessentially Uyghur city of Kashgar, who adopted Islam.

Many Uyghur and some Han Chinese scholars have embraced an alternative view of the origins of the Qarakhanids and Satuq Bughra Khan. The relationship between the Qarluqs, the Qarakhanid state and its other component tribes is uncertain, but it is known that the Yaghma, one of the Toqquz Oghuz tribes that fled west after the fall of the Uyghur Orkhon state in 840, were in early possession of Kashgar; moreover, Yaghma rulers used the title ‘Bughra Khan’, which would later be employed by the Kashgar Qarakhanids. As noted, the Toqquz Oghuz tribes are closely associated with the Uyghur. Thus it may be argued, as did the Russian orientalist Elias, that ‘the dynasty of Ilak Khans [i.e. Qarakhanids] … were, according to the best authorities, Uighurs’, or at least, as Golden puts it, that the Yaghma were a continuation of the Uyghur royal house. Thus both the Qocho Uyghur state and the Qarakhanid dynasty are ‘Uyghur’ in as much as the ancestry of their ruling houses may be traced to the Orkhon khanate of the Uyghurs. Others, disputing the connections between the Yaghma and both Orkhon Uyghurs and Qarakhanids, argue differently.

In either case, underlying the quibbling about the identities of ninth and tenth century Turkic tribes (a heated academic and touchy political issue in the PRC) is the modern nationalist concern over the identity of the Uyghur minzu in China today and its historical ‘claim’ to the Xinjiang region. However, debate on this question seems to me misdirected. Both Qocho Uyghur and Qarakhanid regimes were established by outside conquest elites who ruled over and intermarried with a local population with its Tokhanrian, Iranian, possibly Indian and, in Qocho, Chinese elements. Besides a genetic legacy, such imperial courts left a political, historical and cultural legacy upon the territories and peoples they ruled. All of these have been bequeathed to later inhabitants of this region and together become part of their history, regardless of what name the conquering tribes went by—or in the case of the Qarakhanids, what name later historians ascribed to them. One cannot construct a neat unilinear narrative of Uyghur history, but that does not mean no narrative is possible or that certain branches must arbitrarily be excluded.

In keeping with this broader conception of the modern Uyghurs’ ancestry is the high place occupied by Qarakhanid figures in the pantheon of Uyghur national and cultural heroes. One of these is Mahmud Kashghari, a well-travelled, well-educated member of the Qarakhanid ruling family who studied in the Halik madrasa (college) in Kashgar. He later moved to the Baghdad court of the Abassid caliph, where in the 1070s he wrote an Arabic-Turkic dictionary that describes and compares the various Turkic dialects in great detail. His Compendium. of the Turkic Languages (Divanu lugat-it-Türk) brought the sophisticated techniques of Arabic lexicography to the study of what the Islamic world had generally despised as the crude language of barbarians who roamed the steppes and mountains between Anatolia and China. Perhaps the fact that the caliphate had recently become a protectorate of the Turkic Seljuk dynasty has something to do with Kashghari’s compiling his dictionary at this time. In any case, Kashghari displays a clearly Turkic sense of self: despite his erudition in Arabic and Persian, he defines Uyghur and Qarakhanid dialects as ‘the most elegant’ and ‘purest’ of the Turkic dialects because they had not mixed with Persian. Besides its linguistic value, the terms, verses and proverbs Kashghari cites in his work are a rich store of literary, historical and cultural information and give a vivid sense of Turkic life in the eleventh century.

Yusuf Khass Hajib is another Uyghur literary hero from the Qarakhanid era, likewise a former student at the Halik madrasa. His Wisdom of Royal Glory (Kutadgu Bilig), completed in Kashgar in 1069, is a guide to statecraft, a mirror for princes, written for the Turkic rulers. The book draws on ideas from the older Islamic-Iranian tradition, as well as the values of Sufism, and sets them in Turkic verse. It consists of a series of dialogues between four allegorically-named characters—a king, a vizier, a sage and an ascetic—who respectively represent the principles of Justice, Fortune, Wisdom and Man’s Last End (Religion). The conversations involving the ascetic (named ‘Wide Awake’), occupying the latter half of the book, turn on a debate then prevalent in the Islamic world over whether it was better to become actively involved in affairs of the community and state, or to withdraw to the mountains and devote oneself to religious contemplation. As we will see below, Sufis in Xinjiang later found it possible to reconcile spiritual and temporal concerns, to serve both God and king, and did so with a vengeance.

Both Mahmud Kashghari and Yusuf Khass Hajib are celebrated as national heroes in Xinjiang PRC today. Kashghari’s tomb has been restored, and a modern shrine built to honour Yusuf—both sites have entered Kashgar tourist itineraries.

The Conversion of Khotan

The one Tarim city-state still independent of either Qarakhanid or Uyghur control at this point was Khotan, a Buddhist kingdom whose inhabitants, like those of early Kashgar and Yark and, spoke the Iranian Saka tongue. Khotan’s indigenous dynasty (all of whose royal names are Indian in origin) governed a fervently Buddhist city-state boasting some 400 temples in the late ninth/early tenth century—four times the number recorded by Xuan Zang around the year 630. Khotan enjoyed close relations with the Buddhist centre at Dunhuang: the Khotanese royal family intermarried with Dunhuang élites, visited and patronised Dunhuang’s Buddhist temple complex, and donated money to have their portraits painted on the walls of the Mogao grottos. Through the tenth century Khotanese royal portraits were painted in association with an increasing number of deities in the caves, suggesting the Khotanese royalty knew they were in trouble.

The trouble, specifically, was the Qarakhanid empire. Satuq’s son, Musa, began to put pressure on Khotan in the mid-900s, and sometime before 1006 Yusuf Qadir Khan of Kashgar besieged and took the city. This conquest of Buddhist Khotan by the Muslim Turks—about which there are many colourful legends—marked another watershed in the Islamicisation and Turkicisation of the Tarim Basin, and an end to local autonomy of this southern Tarim city-state. Around the same time, the Qarakhanids destroyed the Samanid dynasty in Transoxiana (or, as the region was now called in Arabic and Persian sources, Mawarannahr, ‘land beyond the [Amu] river’). The Qarakhanids thus became the first Turkic dynasty to rule in Islamic Central Asia, and they did so with the support of the Islamic ’ulama. (The Qarakhanids went on to struggle with the Ghaznavids and Seljuks, likewise Turkic Central Asian powers.) The complete Qarakhanid conquest of the western Tarim and western Zungharia, together with Transoxiana, inaugurates a period in which powers based in Semirech’e controlled Xinjiang, especially the western Tarim, together with Transoxiana and Ferghana. This unit of territory had of course been sporadically unified under the Western Turks earlier, hence the general name of ’Turkestan’ used by the Islamic sources to encompass Transoxiana and Xinjiang. For the next several centuries the history of Xinjiang would be closely entwined with that of Transoxiana, even while the fractious tendencies of nomadic powers made it rare that one ruler controlled both Samarkand and Kashgar.

The Qara Khitay

The now familiar drama of one set of overlords displacing another played itself out again with the arrival in Xinjiang of another ruling dynasty, the Qara Khitay (‘Black Khitay’). The Khitay or Khitan (Qidan) were Mongolian speakers who ruled in north China as the Liao dynasty from 907 to 1125. (Turkic and Mongolian are major divisions of the Altaic language family; early speakers of each were centred on what is today Outer Mongolia. To say that the Khitan were speakers of “Mongolian’ is an anachronistic convention as at this point, before the rise of Chinggis Khan, the Mongols were but one of several tribes speaking Mongolian). Incidentally, the name of the Khitan ruling clan, via Islamic sources and Russian, gave us ‘Cathay’, the medieval and early-modern-European name for China.

The Khitan Liao state was destroyed (and many of their number absorbed) by the Jurchens, a people from Manchuria who established their own Jin dynasty in its place. But one of the Khitan royal clan, Yelü Dashi, fled in the late 1120s to Kedun, a garrison on the northernmost frontiers of the Liao in Mongolia, where he gathered followers from among the tribesmen and took control of the Liao imperial horse herds. Then, after a failed attack on the Kirghiz, Dashi marched west to Beshbaliq, where he assembled the chiefs of a large group of tribes and gathered more adherents, including the Uyghurs. He pressed on west, establishing a tent-city capital in the town of Imil, just east of Lake Balkash, and then took Balasaghun, Kashgar, Khotan and eventually Samarkand and Transoxiana from the eastern and western Qarakhanid rulers. (On some of its Central Asian campaigns the Qara Khitay used elephants to besiege walled cities.) By 1142 Yelü Dashi and subsequent ‘gurkhans’ (khan of khans) ruled over the former Uyghur and Qarakhanid empires, a territory that extended from the Amu darya (Oxus) river, Balkh (Northern Afghanistan) and Khotan in the west, to the Tanguts (Xixia) in the southeast, and to the lands of the Naimans, a powerful Mongolian tribe, in the north-east—in other words, today’s Xinjiang, Kyrgyzstan, Uzbekistan, Tajikistan and southern Kazakstán. Though not fully subordinated to the Qara Khitay, the Islamic state of Khwarazm on the Aral Sea paid tribute to the gurkhans.

The Qara Kitay poses a challenge to any simple imposition of ethnohistorical identities upon Central Asian polities. Because of the Qara Khitays’ background as rulers of north China for two centuries, they brought with them knowledge of Chinese governing practices and institutions; for example, they used Chinese official titles and minted copper coins bearing their names and reign periods in Chinese characters. Perhaps because of this Chinese imperial background, unlike the Turk and Qarakhanid khanates, the Qara Khitay did not subdivide their empire into eastern and western divisions or dole out appanages to princes—for which one modern Chinese scholar praises the dynasty for avoiding ‘feudal’ institutions. Nevertheless, Yelü Dashi did not institute a centralised Chinese-style empire in Turkestan. To the contrary, Qara Khitay rule over its vassal countries remained indirect to the extreme. Local rulers enjoyed almost complete autonomy and maintained their own armies as long as they paid their taxes and tributes to the gurkhans. Throughout the empire agrarian, pastoral and commercial enterprises continued with no major changes. The Uyghur idiquts remained in power in their twin capitals of Qocho and Beshbaliq, and the Qarakhanid khans still reigned in Kashgar and Khotan. Although the Qara Khitay rulers posted financial officials to the courts of conquered regions, they stationed no troops there, dispatching armies only as needed to quell rebellion. Moreover, although Chinese was used for some administrative and prestige purposes (as it was in Uyghur territory), Persian, Turkic and the native Khitan language and script were also employed; it is doubtful to what extent even the Khitan among the ruling élite actually spoke Chinese, though Yelü Dashi himself was well educated in the language.

How, then, should we identify the Qara Khitay? As with the Qarakhanids and many Central Eurasian steppe empires, the label by which we classify the period and region can be misleading, for it is no more than a handy name for the ruling clan, and by no means fully encompasses the ethnic, linguistic or tribal identities even of the Qara Khitay conquest élite, let alone the people ruled in subject polities in the Tarim Basin or Transoxiana. Yelü Dashi fled China with only 80-200 followers; he gathered 10,000 supporters at Kedun and a similar number from diverse tribes in Zungharia. He subsequently took in some 16,000 Khitan former mercenaries at Balasaghun, and also enlisted local Turkic tribesmen. Ultimately the force he took to Central Asia thus included Khitan-, Turkic-, Mongolian-, Tangut-, Iranian-, Chinese- and maybe Jurchen-speaking soldiers; his administrative retinue included Chinese as well as Uyghurs (a Uyghur judge tutored Dashi’s sons). Dashi ruled, moreover, over Iranian-, Turkic-, and any remaining Tokharian-speaking peoples throughout the settled regions of western and eastern Turkestan. Like the Mongol empire that would follow them, the Qara Khitay were a multiethnic polity rather than a particular national or ethnic group.

Küchlük Takes Over the Qara Khitay State

Juvaini described the Qara Khitay empire as a ‘mighty wall’ protecting the lands of Islam from Chinggis Khan. The Qara Khitay decline forms an important chapter in the story of the rise of the Mongols. By the end of the twelfth century a succession of weak gurkhans permitted Qara Khitay officials to exploit populations under their supervision, engendering unrest throughout Zungharia, the Tarim Basin and Transoxiana. The Qarakhanid khan of Kashgar rose against the gurkhan in the first years of the thirteenth century, and was imprisoned in Imil after his rebellion failed. To the west, ‘Ala ad-Din Muhammad, the new shah of Khwarazm (on the south-western shores of the Aral sea), challenged Qara Khitay suzerainty in Transoxiana by campaigning against the former Qarakhanids still ruling as Qara Khitay vassals in Bukhara and Samarkand. In Qocho the Uyghurs turned on their overbearing Qara Khitay supervisor in 1209, beheaded him and threw his head out a tower window. The idiqut, Barchuq, then quickly allied with Chinggis Khan (the first head of state to do so). Barchuq proved a loyal vassal, fielding troops and even riding out himself on campaign with the Mongols against the Khwarazmshah and the Tanguts. Chinggis gave Barchuq one his daughters to marry and the honorific title of’fifth son.’ Barchuq remained idiqut of Uyghuristan.

Up to this point Chinggis had been busy subduing and unifying the many tribes in the Hobbesian world of twelfth-century Mongolia. In 1208 his growing confederation defeated an alliance of the Naiman and Merkit tribes, forcing the Naiman chief, Küchlük, to flee west into the protection of the Qara Khitay gurkhan. The fugitive Küchlük drew upon his reserves of guile and charisma to convince the beleaguered Qara Khitay gurkhan to allow him to re-assemble a force of Naiman and Merkit tribesmen. While the gurkhan was engaged with the Khwarazm-shah’s forces in Transoxiana, Küchlük gained a strategic footing in Semirech’e, and by around 1211 effectively usurped control of the Qara Khitay state.

The Qara Khitay gurkhans had been hospitable to the various religions of their steppe and oasis empire. Kashgar was a Nestorian Bishopric, and Christianity also flourished around the northern capital on the Chu river, as did Judaism in Samarkand. Policies towards Islam, too, were liberal, though the Muslims in the former Qarakhanid lands viewed the Buddhist Qara Khitay as infidels. However, when Küchlük took command of the Qara Khitay empire he reversed this sensible policy. Himself a Buddhist convert from Christianity, he launched a pogrom against Muslims in his new territory. For three years running (1211-13) he sent troops to raid Kashgar at harvest time, starving the city until it submitted to him. The persecution was particularly acute in Khotan, a city already forcibly converted once before, when it fell to the Qarakhanids two centuries earlier. On conquering this southern Tarim city, Küchlük, in the words of Juvaini,

… compelled the inhabitants to abjure the religion of Mohammed, giving them the choice between two alternatives, either to adopt the Christian or idolatrous creed or to don the garb of the Khitayans. And since it was impossible to go over to another religion, by reason of hard necessity they clad themselves in the dress of the Khitayans.… The muezzin’s call to prayer and the worship of monotheist and believer were broken off; and the schools were closed and destroyed. One day, in Khotan, he drove the great imams out on to the plain and began to discuss religion with them.

One of the imams, ‘Ala ad-Din Muhammad of Khotan (not to be confused with the Khwarazm-shah of the same name), ‘girded the belt of truth about the loins of veracity’ and engaged the blaspheming Naiman chief in this religious debate. The mullah concluded his presentation with a curse (’Dust be in thy mouth, thou enemy of the faith, thou accursed Küchlüg!’), and Küchlük had him crucified on the door of his own college. But justice was at hand, for ‘God Almighty, in order to remove the evilness of Küchlüg, in a short space dispatched the Mongol army against him.’

Indeed, Chinggis Khan harboured a personal animus against Küchlük, and sent his trusted general Jebe (‘weapon’) against him in 1216-18. Jebe took Almaligh (near modern Ghulja) and accepted the peaceful submission of tribes in Zungharia and Semirech’e; Küchlük fled south and died in an attempt to escape over the Pamirs.

The Mongol Imperial Period

After Küchlük’s intolerance, the Xinjiang Muslims greeted Jebe and the Mongols as liberators who restored rights of worship and refrained from looting. (That being said, one noteworthy aspect of the Mongol conquest was that Islam became less prevalent on the steppes of Zungharia, as the Mongols, who were shamanist but also favoured Christianity and Buddhism, took control of Muslim Turkic lands). With the Qara Khi-tay gurkhans and Küchlük out of the way, the Yili, Talas and Chu River valleys,’Kashgaria’ in the south-western Tarim and Uyghuristan in the north-east all joined the empire of’the world-conquering Khan’. The way was open for the Mongols’ westward campaign to Transoxiana, Afghanistan and Persia.

Before his death in 1227 Chinggis Khan divided his vast empire into an ulus (a unit of land and people here equivalent to a khanate) for each of his sons. Jöchi, the eldest, received lands in western central Eurasia ‘as far as the hooves of Mongol horses have tramped’ (including places not yet conquered, where Jöchi’s own son Batu would carve out an empire). The ulus of the second son, Chaghatai (Chaghatay, Jagatai etc)., comprised essentially the territories of the old Qara Khitay empire, centred near Almaligh (Ghulja), and in theory including the band of oases extending from Turfan to Samarkand. The lands of the third son and Chinggis’ designated successor, Ögedei, took in central Siberia and eastern Zungharia, with main pastures on the I mil and Irtysh rivers near modern Tacheng (Chughuchaq, Tarbaghatai) in northern Xinjiang. By custom, as ‘prince of the hearth’, the youngest son Tolui received as his own ulus the Mongol homelands centred on the Orkhon valley, as well as the bulk of Chinggis’ army

This arrangement, though neat in appearance, planted the seeds of centuries of instability, particularly in regard to the territory of the former Qara Khitay empire. It would have been natural for the Chaghatayids, based around Issyk Kul and the Yili Valley, to control the oases of Transoxiana and southern Xinjiang. However, the revenue from these Muslim and Uyghur cities and rights to invest rulers and residents was, in practice at least, the prerogative of the Great Khans further east. Uyghuristan (as the sources now call the Qocho Uyghur kingdom) remained subject to the Mongol Great Khans in Beijing until the start of the fourteenth century, as did Khotan until around 1375.

Another built-in source of tension arose from Ögedei’s inheritance of the superior position of Great Khan, while Tolui occupied the symbolically important Orkhon valley lands—where Ögedei would in fact build the Mongol imperial capital of Karakoram. The Ogedeids and Toluids would thus compete over the Great Khanship; conflict would also arise later between those Mongols who remained in Mongolia and those ruling from China. All of this assured that Xinjiang would be drawn into the politics of the Mongol imperial centre to the east, and that the region would again play a geostrategic role similar to that under the Han/Xiongnu and the Tang/Türk rivalries. The rich cities of Transoxiana to the west, in theory part of the Chaghatayid legacy, would at times be ruled together with the Tarim and Zungharia, and at other times fall to locally based Mongol descendents and Türkic strongmen. Khans based in Xinjiang would repeatedly raid Afghanistan and northern India. Generally, the fissiparousness, opportunistic alliances and fratricidal struggles of the Mongols and Turks assured a long and—to the student of the region’s history—bewildering epoch of military flux throughout Turkestan and neighbouring areas.

Here is a summary of the major political and military events in Xinjiang during the Mongol imperial period. After the death of Chinggis khan (1227), the house of Chaghatai remained closely allied, even subordinate to, the new Great Khan Ögedei (r. 1229-41) and his descendents until 1251. The settled areas of the Tarim Basin, as well as Uyghuristan, were now under administrative arrangement similar to that of the rest of the Mongol empire, with Mongol ‘residents’ (Mo. darugha, Ch. guanren) stationed in cities to assure that Mongol demands were met: these included hosting official visits, dispatching royal hostages to the Mongol court, registering the population, paying taxes and irregular exactions, raising a militia, and maintaining postal relay stations. A Khwarazmian Muslim, Mahmud Yalavach, served as darugha overseeing Transoxiana, the Tarim Basin and, by ögedei’s reign, Uyghuristan as well. His son, Mas’ud Beg, succeeded him. The Persian sources praise these governors for wise administration that established consistent taxation systems and rebuilt cities and economies destroyed by Mongol conquest.

Ögedei was succeeded by his widow as regent, by his son for two years, and then by his son’s widow. Then, after a bitter struggle, a new house seized the Great Khanship: Möngke (r. 1251-9) was the son of Tolui, and from his reign the Toluids would hold the Great Khanship and the emperorship of China until replaced by the Ming Dynasty in 1368. After coming to power in 1251, Möngke viciously pursued his ögedeid and Chaghatayid rivals, killing or exiling adult Chaghatayids and raising their children within his own court. He selected Chaghatai’s widowed daughter-in-law, Orghana, as regent for her infant son Mubarak Shah (the first Muslim Chinggisid). Möngke then struck a deal with Batu, khan of the Golden Horde in what would become the Russian steppe. Together, they would split the steppe along a line running through Semirech’e.

Xinjiang thus came fully under the control of the Great Khan, but only for a few years. After Möngke ‘s death, the Toluids themselves fractured, electing rival khans Arigh Böke (r. 1260-4) in Mongolia, and Khubilai (Qubilai;r. 1260-94) in China. This constellation of power and geography should by now be familiar: Xinjiang became a critical source of grain and tribute for the Mongolia-based Arigh Böke after Khubilai cut off grain shipments from China. Khubilai tried, but failed, to conquer the Tarim Basin and Zungharia, though Uyghurstan remained loyal to him. Arigh Böke sent Chaghatai’s grandson, Alghu, to manage the Chaghatai ulus, and after rallying Chaghatayid princes in Kashgar, Alghu sent armies to retake Transoxiana from the Golden Horde and extend power into Afghanistan and eastern Persia. But Alghu was no tame puppet. He turned on his former sponsor, Arigh Böke, who lost the support of the nomads in Zungharia and was forced to return east and submit to Khubilai. Alghu married the durable Orghana (his late uncle’s wife), but died (1264) before he could enjoy her or his newly reconstituted and consolidated Chaghatayid khanate in Transoxiana, the Tarim and Zungharia. Orghana’s son Mubarak Shah thus became the inhertor of the Chaghatai ulus.

Khubilai, from his capital in Beijing (Dadu), now moved again to control Turkestan, dispatching Mubarak’s cousin, Boraq, with an edict naming him co-ruler of the Chaghatai ulus. Boraq quickly took power in Transoxiana, outstripped Mubarak, and then, true to form, rebelled against Khubilai, defeating the latter’s armies on the Oxus and in the Tarim, and in the bargain plundering Khotan, once Khubilai’s city. Of Xinjiang territories only Uyghuristan now remained subject to the Great Khan in China.

Meanwhile, a new contender from the house of Ögedei had been gathering strength in the north. Qaydu (Khaidu) drew on support from the Golden Horde to the west to challenge Alghu, and after the latter’s death took advantage of Boraq’s wars with Khubilai’s forces to seize control of the Talas river plain. After an uncharacteristically peaceable conference (quriltai) in 1269, Qaydu, Boraq and representatives of the Golden Horde divided Transoxiana between them, all the nomad monarchs agreeing to stay out of the way and allow Mas’ud Beg to administer the towns. Upon Boraq’s death around 1271, Qaydu became de factoruler of the entire Chaghatai khanate for the next thirty years, though since he was himself an Ogedeid he continued to prop up Chaghatayid princes as formal khans. His rule marks a watershed of sorts, for neither the khans of the Golden Horde to the west nor the khan-emperors of Mongol-occupied China (the Yuan dynasty) had influence in Xinjiang or Transoxiana after Qaydu gained power. Thomas Allsen has pointed out that the same fiscal-military dilemma that undermined Han and Tang power in the ‘western regions’ made it impossible for the Yuan to hold the area: whereas nomad powers in Zungharia and Mongolia could raid and tax eastern Xinjiang with ease and profit, and in fact relied on grain and commercial connections available there, a power based in north China faced great continuing expenditures to access and defend the region via the Gansu corridor, costs for which the region’s produce could not compensate.

Uyghurs in the Mongol Period

Throughout the nine decades between Chinggis and Qaydu, Uyghurs and Uyghuristan occupied a unique position in the Mongol empire. Because of their early submission and continued loyalty to Chinggis, the Uyghur idiquts enjoyed a good deal of autonomy; Chinggis did not bestow Uyghuristan as part of an ulus upon any of his sons, and not until Ögedei’s reign were Mongol residents stationed there. Li Zhichang, chronicler of the journey of Daoist master Chang Chun west from China to where Chinggis’was then encamped near the Amu River, noted the same features of Qocho as had earlier descriptions: the heat, the rich fields watered by karez, fruits and ample supplies of grape wine. In Qocho and Beshbaliq the travelling Daoists met Confucians as well as many Buddhist and Mamchean priests, indicating the continued local prominence of these religions.

Uyghuristan was not immune to the effects of Mongol succession struggles, however. In the years following Ögedei’s death, the idiqut Salindi supported Ögedei’s grandson, Shireniün, who ultimately lost out to Möngke. When Möngke successfully seized the Great Khanship in 1251, he tried and executed Salindi and several Uyghur nobles. Among the charges levelled at Salindi was his alleged conspiracy with Oghul Khainiish (regent and mother of Shireniün) to massacre the Muslims of Uyghuristan during Friday prayers. While the background and true motivations underlying this incident are unclear, it does suggest that Islam had made inroads into eastern Xinjiang by the mid-thirteenth century.

As they had for the Qara Khitay, Tangut, Naiman and other neighbouring powers, Qocho Uyghurs served the Mongols as a sort of’steppe intelligentsia’, providing crucial administrative skills to the nomads as they built an empire. The Uyghur Tatar Tongga, formerly a secretary in Naiman service, introduced to Chinggis’ new state the use of official seals in tax recording as well as the Uyghur script (which was adopted for the Mongolian language and remains in use in Inner Mongolia today). Uyghurs tutored Mongol imperial princes and served as darughas and in other official capacities throughout the empire, many reaching high rank. One such figure was Körgüz, a Uyghur, possibly a Christian (his name is equivalent to ‘George’) from a village near Beshbaliq. Juvaini provides us with a sketch of his career and a sense of what was possible for an ambitious and literate Uyghur in the Mongol era.

Körgüz’ father died while he was still young, and although custom dictated that he marry his widowed step-mother, she refused and later prepared to marry someone else. His patrimony at stake, Körgüz went to the idiqut who ordered the step-mother to compensate Körgüz with property. Körgüz then learned the Uyghur script and began teaching. Still not well off, he sold himself into bond; with the money he bought a horse and rode off to the camp of Batu, khan of the Golden Horde. Though he entered Batu’s service as a herdsman, he soon gained recognition for his literacy and intelligence and began to rise in rank, from tutor to a secretary of the imperial viceroy of Khwarazm, and ultimately to Viceroy himself, governing the eastern Persian province of Khurasan. Instrumental to his gaining this position was the patronage of Chinqay, an important emir who may himself have been a Uyghur, but was at any rate Turkic. Körgüz is said to have restored the economy and governed justly; he rebuilt the treasury, reduced corruption in the tax collection system, and established the yam postal service to free people from arbitrary commandeering of their horses as official mounts. Under his tenure local élites regained sufficient confidence to invest in the underground irrigation canals on their estates and to rebuild the bazaar. Nevertheless, Körgüz fell victim to Mongol politics, first becoming embroiled in an intrigue involving a rival of Chinqay, and later, in the 1240s, when he insulted a Chaghatayid emir and showed disrespect to ögedei’s widow. For this his mouth was stuffed with stones and he was executed.

Despite such vulnerability to political vicissitudes, Uyghurs and Uyghuristan weathered the epoch of Mongol rule better than Transoxiana, avoiding the destructive conquests and subsequent struggles among rivals that all but levelled many Central Asian cities. The kingdom’s luck ran out, however, during Qaydu’s reign. In the late thirteenth century the territory from Beshbaliq to Hami became the front line in the conflict between Qaydu and the Yuan Dynasty under Khubilai. Raiding by Qaydu and the Chaghatayid princes allied with him inflicted heavy damage to the hydraulic infrastructure so critical to growing crops in the dry heat of the Turfan depression; insecurity forced the Uyghur ruling house to withdraw first from Beshbaliq to Qocho, and then east to the Gansu corridor, where they and a large number of Uyghur refugees became dependents of the Yuan state. The former capital cities of Uyghuristan, meanwhile, were transformed first into Yuan military outposts, colonised by Mongol and Chinese troops, and then in the fourteenth century occupied by Chaghatayids, who enthroned their own pro-Chaghatayid line of idiquts.

Qaydu died after a failed attack on Karakoram in 1301, and a series of Chaghatayid khans ruled Xinjiang and Transoxiana in their own right thereafter, also venturing into Afghanistan and north-west India and warring with the Mongol Ilkhanate in Persia. However, by the 1340s another geo-historical watershed was crossed when the Chaghatai khanate split in two. One source of this split was religious: the Chaghatayid khanTarmashirin, who plundered northern India in the late 1320s from his base in Transoxiana, was a convert to Islam. In 1333-4 the Buddhist, Christian and shamanist nomads of Issyk Kul and the Yili region rebelled against him and established a new khan. Tarmashirin’s line of Chaghatayids carried on in Transoxiana, where it provided Turkic rulers with figurehead Chinggisid khans. The other branch ruled over the Tarim and Turfan Basins from what now came to be called ‘Moghulistan’, a geographic notion that included Zungharia as well as the steppe lands between the Syr Darya and Lake Balkash. Like their brothers and cousins in Transoxiana, the Moghulistan Chaghatayids, too, became tools of others: the wealthy and powerful Mongol Dughlat clan served as emirs while playing kingmaker with the Chaghatayid royals in what is now Xinjiang.

Moghulistan: The Muslim Chaghatayids

In the late 1340s the Dughlats proclaimed as khan a supposed Chaghatayid prince, Tughluq Temür (r. 1347-63), who is famous both as a convert to Islam and as a conqueror. When still a young man, Tughluq met the shaykh Jamal ad-Din, a ‘member of the sect of Khwajas’, who explained to him the meaning of faith and the ‘duties of a Musulman’. Impressed, Tughluq promised that should he ever become khan, he would accept Islam. Some years later Tughluq was indeed the khan, and though the shaykh had died, his son Arshad ad-Din presented himself by calling the prayers outside Tughluq’s tent early in the morning. Thus abruptly roused from his bed, Tughluq was at first enraged. When he learned who the holy man was, however, Tughluq not only made good on his personal promise, but agreed to make all his princes convert as well, or put them to the sword. All the princes obeyed except Jaras, who would only accept the faith if the shaykh could defeat a wrestler in Jaras’ entourage, a giant famous for having once thrown a two-year-old camel. As is common in such stories of religious contests in Inner Asia, the holy man miraculously prevailed, and the wrestler recited the Islamic creed as soon as he regained consciousness. Thereupon ‘the people raised loud shouts of applause, and on that day 160,000 persons cut off the hair of their heads and became Musulmans. The Khan was circumcised, and the lights of Islam dispelled the shades of Unbelief.’ Though the story is mythologised, the Mongols (or, as they are known in Persian sources, Mughals) of the Chaghatai khanate were largely Islamicised by the mid-fourteenth century, the end of the Mongol imperial period across Eurasia. Outside China and Mongolia, the Mongols were by now all Muslims.

As a conqueror, Tughluq Temür campaigned in Transoxiana and Afghanistan, temporarily restoring the unity of the Chaghatai khanate on either side of the Pamirs. After Tughluq, however, the Chaghatayids again descended into a chaos of puppet khans and concurrent reigns in different parts of what was once the Chaghatai ulus. In the west Tamerlane (Timur-i Lang; r. 1370-1405) built a short-lived empire from India to the Middle East to the Russian steppe as emir under a nominal Chinggisid khan, leaving his own descendents to vie with remaining Chaghatayids for Transoxiana. It was a descendent of both Timnir and Chaghatai, Babur (1483-1530), who founded the dynasty in India which has come to be known as ‘Mughal’. (In fact, Babur and his successors preferred to think of themselves as Timurids, the term ‘Moghul’ indicating rather the Chaghatayid rulers of Moghulistan whom they considered false Muslims and uncouth bumpkins).

Meanwhile, back in Moghulistan, a Dughlat emir, Qamar ad-Din, usurped the khanship following Tughluq Temür’s death in 1363 and did away with most of Tughluq’s offspring. Other Dughlats in turn challenged Qamar’s rule in Moghulistan and Altishahr, as did Tamerlane, whose armies penetrated as far as Karashahr in 1389, driving Qamar into the Altai mountains. Khizr Khwaja, Tughluq Temür’s one surviving son, was then placed on the throne by another Dughlat who had sheltered him until he reached adulthood. In the 1390s Khizr personally embarked upon a holy war (ghazat) against ‘Khitay’, conquering Turfan and Qocho (by then called Qarakhoja) in the core of Uyghuristan, and reportedly achieving the conversion of the Uyghuristan populace. In fact, the disappearance of local Buddhism from Turfan took some time. In the early fifteenth century the rulers of Turfan sent embassies to Beijing led by Buddhist priests. Hafiz Abru, a Timurid envoy passing through Turfan and Qarakhoja en route to Ming China in July of 1420 commented on the richness and quantity of’idol temples’ and the prominent image of Sakyamuni. Only by the 1450s were mosques common enough in the area to feature in travellers’ accounts. Nevertheless, Khizr’s conquest of Uyghuristan does mark the end of Uyghur rule and inauguration of Chaghatayid, or Moghul, control of the Turfan Basin.

Khizr also concluded a truce with Tamerlane, marrying a sister to him, thus giving Tamerlane a Chinggisid connection which allowed him to call himself by the prestigious title ‘son-in-law’ (of the khans). However, this family tie did not prevent further Timurid invasions of the Tarim Basin:Tamerlane’s grandson, Mirza Alexander, attacked Kashgar, Yarkand and Aqsu in 1399-1400; the residents of Aqsu saved their city from sacking only by delivering the city’s rich ‘Chinese’ merchants up to Mirza.

The political and military narrative continues in this vein from the late fourteenth through the seventeenth century, its complexity compounded by the fact that chronologies in the Persian and Chinese sources are hard to reconcile. The fluctuating fortunes of individual rulers and families aside, however, four important developments characterise this Moghulistan period of Xinjiang’s history: the decline and disappearance of the Chaghatayid Moghuls; the rise as new regional powers of the Kazaks, the Kirghiz and the Oirats (Zunghars); the revival of trade linking China with Transoxiana and India; and, to be taken up in the next chapter, the further advance of Islam as both religion and political force in the Xinjiang area.

Chaghatayid Decline and the Formation of New Peoples

The Chaghatayids lost their status as rulers rapidly in Transoxiana, and more gradually in the eastern parts of their former domain. The western Chaghatayids had served as convenient legitimating puppets under the Timurids; however, the Uzbeks (the Shibanid dynasty, 1500-99) who drove the Timurids out of Transoxiana, could boast their own Ching-gisid lineage (through Shiban, a son of Chinggis’ eldest son,Jöchi), and thus had no need for Chaghatai’s descendents. In Xinjiang Chaghatayid khans still reigned and even ruled in some places, but from the late fourteenth through seventeenth centuries the region became subdivided into distinct smaller realms or city-states: in Moghulistan to the north, in Altishahr (‘six cities’, as the sources begin to refer to Kashgar and the Tarim Basin) and in the Turfan area. Occasionally a single ruler succeeded in uniting the whole region, but more often these realms and cities were ruled separately either by competing Chaghatayid descendents, by the Dughlats, or, as will be discussed below, by the Khojas, ruling either through a puppet Chaghatayid or in their own right. As the power of the old Moghul khans declined, they lost control of the nomads. New tribal confederations emerged in the steppes and mountains.

One of these was that of the Kazaks, who lived nomadically in Zungharia, Semirech’e and on the steppes north of Transoxiana, roughly the territory of Kazakstán and northern Xinjiang today. The Kazaks as a tribal and political entity emerged from a succession crisis among the remnants of the Golden Horde (also known as the White Horde), the Mongol imperial khanate that Chinggis had bequeathed to his eldest son, Jöchi. In the fifteenth century two Jöchid descendents contested the khanal succession. In the end Abulkhayr (Abu’l Khayr) Khan killed his rival Baraq Khan and emerged victorious as khan. Abulkhayr’s grandson, Muhammad, went on to conquer Transoxiana and to found what we know as the Uzbek (Ozbeg) state. Meanwhile, Baraq Khan’s sons, Janibek and Giray, fled with 200,000 followers to Semirech’e and Zungharia. By the 1520s their confederation numbered a million tribesmen, was known by the name of Kazak (Qazaq, Kazakh), and had expanded its influence south-west to the Syr Darya. The Kazak tribes by the eighteenth century had formed into three subdivisions, known from west to east as the Lesser, Middle and Greater Hordes—or ‘hundreds’, as the divisions are known in Kazak (Kishijüz, Ortajüz and Ulujüz).

The Kirghiz were not a new people, but many Kirghiz migrated to new lands in western Moghulistan during this period. The Turkic-speaking Kirghiz from the southern Siberian watershed of the Yenesei River had destroyed the Orkhon khaghanate of the Uyghurs in the mid-ninth century. A century later they were themselves forced out of the Orkhon valley by the Khitan, at which point some Kirghiz returned north, and others may have begun migrating to the south-west. Kirghiz participated in Qara Khitay and later Mongol imperial campaigns further west; many joined Qaydu’s armies in the later thirteenth century. By the fifteenth century, the Tianshan, Issyk Kul area and Pamir had become a Kirghiz homeland, which they defended against the expansion of the Oirats (see below). Kirghiz tribesmen were the focus of concerted missionary efforts by Sufi shaykhs in the sixteenth and seventeenth centuries, and by the eighteenth century had become fervent Muslims lending military muscle to various factions in the frequent wars of religion and conquest around Kashgar and the Ferghana Valley They remained largely unassailable in their mountain redoubts, and it was only in the twentieth century that they came under direct control of outside political powers.

A group of non-Chinggisid Mongols, speaking a distinct western Mongolian dialect, also came to power in Zungharia from the fifteenth century. The Oirats (Oyrad, Ölüd, Eleuth; known in Persian and Russian sources as Kalmyk, Qalmuq, Kalmuk etc). vied for control of Mongolia and put pressure on the Ming in the early 1400s; later they were forced west. By the seventeenth century a sub-confederation of Oirat tribes, the Zunghars, would wrest control of Moghulistan from the last Chaghatayids and form the last great steppe empire in world history (see Chapter 3). As for the Chaghatayids themselves, our best source covers only up to the mid-sixteenth century. Though details are scarce, we know that Chaghatayid princes later paid visits to the court of the Qing dynasty, so they survived at least in the capacity of figureheads. Still, after the mid-sixteenth century no Chaghatayids ruled on a regional scale, and even where they commanded individual cities they remained under the sway of the religious charismatics known as Khojas (Khwajas; see Chapter 3).

Chinese-Central Asian Relations and the East-West Caravan Trade

The death of Tamerlane (1405) and of his aspirations for far eastern conquest corresponded to the accession of the Ming Yongle emperor (r. 1403-24), a Chinese ruler open to diplomatic and trade contacts with Central Asia. While exchanging letters and envoys with the new Timurid Sultan in Herat on a basis of de facto equality, Yongle sent a gift of silk in 1406 to Turfan’s eastern Chaghatayid ruler, who reciprocated the following year. In 1408 the Ming court received a Turfani embassy led by Buddhist priests. From this point on, missions to the Chinese court to present ‘tribute’ in return for gifts and trading opportunities became the focus of Turfan’s relations with China for over two centuries. Indeed, the cities of Altishahr, Samarkand and Herat dispatched envoys to and traded with the Ming via Turfan.

These exchanges of goods, through which the Ming received horses, jade and other items from Central Asia in return for gifts, fall within the ‘tribute system’ model familiar to students of Chinese history. It is generally agreed that the Ming court covered the envoy’s expenses, opened border markets for certain powerful neighbours, and knowingly overpaid for their goods in return for the prestige of having ostensibly obsequious visitors come to the court to present ‘tribute’—the Chinese term for these diplomatic present, gong, is usually translated as ‘tribute’, but it would be better to simply call them ‘gifts’ since ‘tribute’ implies a subservient and extractive relationship that did not exist. Though all at the time saw through the charade, official Chinese annals present the official line that emissaries bearing gifts were pledging submission to the Ming emperor. Some writers in the PRC today, in order to enhance China’s historical prestige and claim on Xinjiang, still maintain the fiction that envoys presenting gifts were Chinese vassals, as in the following passage on fifteenth-century relations between Turfan and the Ming:

The fact that the latter Chaghatayid princes of Beshbaliq and Turfan repeatedly sought to present gong to the Ming court shows us, first, that they saw themselves as Ming vassals (fanshu), members of the Chinese nation (zhonghna minzu); and second, that at the same time economically they were mutually inseparable [from China]…

The development of Ming-Central Asian trade relations was punctuated after the 1460s by conflict between the Ming and the Chaghatayids in control of Turfan. In the 1460s and 1470s, Moghulistan (now with a capital at Aqsu) and the Turfan region were reunited by Yunus Khan (r. 1462-81) with assistance from the Timurids. Yunus was Persian-educated and deeply cultured; a visiting cleric from Transoxiana had expected the Moghul chieftain to be ‘a beardless man, with the ways and manners of any other Turk of the desert’, but upon meeting him had been surprised to find ‘a person of elegant deportment, with a full beard and a Tajik face, and such refined speech and manner, as is seldom to be found even in a Tajik’. Yunus was thus more sophisticated and powerful than the Ming perhaps appreciated, for it limited his trade missions and snubbed his requests for such status items as four-clawed dragon robes. The Ming, moreover, attempted to maintain Hami as a military outpost (weisuo), under a line of rulers invested by the Ming court, to counter-balance Turfan’s regional power. As a result, until the mid-sixteenth century Turfan and the Ming faced off over Hami, which was repeatedly conquered and plundered. Nevertheless, trade missions continued more or less uninterrupted. The Ming needed the horses.

Trade became easier still after 1514, when Yunus’ grandson, the Chaghatayid Sa’id Khan (r. 1514-33) put a temporary end to the squabbling among Chaghatayid and Dughlat contenders for power in the west and established himself in ‘Kashgaria’ (Kashgar, Yarkand and Khotan). He reached a settlement with his brother, Mansur (r. 1503-43) who reigned in Moghulistan and Turfan—that is, in the Yili region, Aqsu, Kucha, Karashahr and Turfan. Sa’id’s reign inaugurates what Uyghur and Han Chinese scholars call the Sa’idiyya state or the Yarkand khanate, a distinct dynasty ruling in the south-western Tarim while other Chaghatayids continued to reign elsewhere. The generally amicable arrangement between Sa’id in the west, and Mansur in the north and east, ushered in a rare episode of relative stability, as the main source for this period, the Tarikh-i Rashidi, notes:

From this peace and reconciliation between the two brothers, resulted such security and prosperity for the people that any one might travel alone between Kamul [Hami] or Khitai [China] and the country of Farghana without provision for the journey and without fear of molestation.

The Ming initially squabbled with Mansur as it had with his grandfather, but ultimately gave in and allowed larger and more frequent trade missions. For example, Ming annals list 150 ‘princes’ from Samarkand, Turfan, Mecca and elsewhere visiting the Chinese capital in 1536. By the sixteenth century, then, something like regular commerce and diplomatic communication again linked Central Asia and Ming China, and northern and southern Xinjiang (Moghulistan and Altishahr) enjoyed a measure of political unity.

Although the symbolic aspects of the mis-named ’tribute system’ have been much discussed, it is perhaps more useful to view this arrangement as an attempt to maintain a state monopoly on foreign trade to use as a diplomatic and strategic tool. One side effect of this Ming policy à-vis Central Asia was effectively to afford rulers in Xinjiang a similar monopoly on eastbound trade, which for them proved highly profitable. This much is clear from the records of Bento de Goes, a Jesuit lay-brother who travelled to the Tarim Basin from India in 1603 on a mission to ascertain whether ‘Cathay’—known to Europeans from Marco Polo and other medieval travellers—was the same place as ‘China’—where Jesuits were recently established at court. After travelling from Kabul, Goes laid over in Yarkand for a year waiting for an eastbound caravan to join. During this interval Muhammad Sultan, Chaghatayid ruler of the Kashgar-Yarkand area, put the right to serve as caravan leader up for sale. The highest bidder (who paid 200 bags of musk for the contract) took on other members, who likewise paid to join the caravan under his authority. In the end Goes and his stocks of jade (recently acquired in Khotan) joined a caravan consisting of the statutory seventy-two members. The merchants then proceeded to the borders of the Ming carrying passports from the Yarkand khan charging them as his ‘ambassadors’. After Turfan’s rapprochement with the Ming, then, long-distance trade between Central Asia and the Ming again became routine, if not exactly trouble-free (Goes died on the Chinese border in Suzhou [Gansu], possibly poisoned by fellow travellers to whom he had unwisely lent sums of money). Despite the formal strictures on the trade imposed by the Ming, moreover, there seems to have been an active trade at the border: Matteo Ricci noted that a community of Central Asian Muslims had established itself in Suzhou (Gansu) by the mid-seventeenth century to broker trade with Kashgar and other western cities.

The Silk Road in Early Modern Times

That historical records indicate a revival of long-distance trade relations from the early fifteenth century (and continuing, despite interruptions, through Qing times) is interesting in light of what has become a common assumption in world history circles: that the rise of European maritime trade in the Indian Ocean and South China Sea rang the death knell of Silk Road trade across central Eurasia, marking the beginning of the region’s long decline. This bell apparently tolled for some time, since orbituarists have fixed the patient’s death at various points over two centuries: For the authors of the Cambridge History of Islam, ‘Central Asia was thus isolated from the early sixteenth century … and … led an existence at the margin of world history … The discovery of the searoute to East Asia rendered the Silk Road increasingly superfluous.… ’ To Niels Steensgaard, ‘the destructive effects of the discovery of the sea route to Asia upon the traditional intercontinental trade routes were not felt until after the elapse of an entire century … and at the end of the sixteenth century the transcontinental caravan trade reached dimensions which must presumably be regarded as its historical culmination.’ IsenbikeTogan sees the ‘closure of the silk routes’ as happening in the latter seventeenth century, when rising Zunghar, Qing and Russian power choked off the independent merchant intermediaries.

Much depends on what one means by ‘Silk Road’: whether one considers only goods that went from one end of the continent to the other, or choses to include the Chinese, Central Asian, Russian, Indian or Middle Eastern regional segments of the Eurasian trade webs in one’s conception of ‘Silk Road’; or, indeed, whether one includes ‘trans-ecological’, ‘north-south’ necessity trade between sedentary societies and pastoral nomads as well as ‘trans-civilisational’, ‘east-west’ luxury exchange between the continent’s political and population centres. Silk from China was already considered inferior in Europe to that from the Middle East by the thirteenth century, yet the Mongol period is nonetheless considered a Silk Road highpoint. Trade was, to be sure, disrupted on many occasions by wars and political fragmentation, which made travel unsafe and raised protection costs. Nevertheless, viewed from a long perspective taking in the fifteenth through nineteenth centuries, there is no evidence that competition along sea routes affected the volume of overland trade at the centre of Eurasia, and little indication that this trade entered a secular decline from the sixteenth or seventeenth centuries. Rather, the Ming records of ‘tribute trade’ furnish a good deal of evidence that large caravans plied both desert and steppe routes across Xinjiang between Transoxiana and north China until the late sixteenth century; thereafter, missions were less frequent, but nonetheless occurred. Moreover, as we will see in the next chapter, following the turmoil of the Ming-Qing transition in China in the mid-seventeenth century, Qing, Zunghar and Russian unification and expansion quelled petty warring khanates and boosted economic activity across the region.