The Tang legacy on the Silk Road during the Uighur era: urbanisation in the eastern Tianshan region during the ninth to thirteenth centuries

Abstract

Tang expansion in the early seventh century brought about a series of changes to the eastern Tianshan region, including the incorporation of the region into the imperial postal relay and defence system. Important structures, including cities for the soldiers and other immigrants from Tang territory, along with fortresses and relay posts, were established along the major routes in the region, especially on the northern slopes of the Tianshan range. However, the era after the decline of the Tang is not as well known, due to a lack of contemporary sources. This article, based on a comprehensive analysis of documentary and unearthed materials, discusses a previously unacknowledged process of urbanisation in the region during the Uighur era. Uighur immigrants, originally nomads on the Mongolian steppe, occupied not only the cities, but also the garrisons and other infrastructure established by the Tang. As a result, urban settlements were established at sites that had previously served military purposes. Clusters of new cities emerged in the region, especially on the northern slopes of the Tianshan, which had long been part of the nomadic cultural zone. The sedentary and mercantile culture of the Uighurs played an important role in this process, serving as an impetus for economic prosperity along the eastern section of the Silk Road between the Tang and Mongol–Yuan eras.

Introduction

The Tang era, along with the later Mongol–Yuan era, is generally considered the heyday of the Silk Road, in part thanks to the relative abundance of contemporary literary sources on Central Asia. However, due to a relative lack of literary sources, the era after the decline of the Tang is much less well known. As the Tibetans encroached on Tang territory in the Gansu Corridor after the An Lushan Rebellion (755–63), the direct connections to Anxi 安西 and Beiting 北庭 (in what is now Xinjiang) were cut off. Thereafter, information on Central Asia in Chinese sources is fragmentary and scarce, leaving the following history to a great extent in darkness. After a series of battles with the Tibetans in the Tianshan region, the Uighur Khaganate finally took control of the former Tang territories of Beiting and part of Anxi at the beginning of the ninth century.¹ The collapse of the Uighur Khaganate in 840 triggered a mass migration of Uighurs from the Mongolian steppe and their subsequent settlement in the eastern Tianshan region.

The Uighur occupation of the major cities in the region, such as Xizhou 西州 (Uighur (Uig.) Qočo < Chinese (Chin.) Gaochang) and Beiting (Uig. Bešbalïq), is recorded in various historical sources. The decipherment of unearthed manuscripts has resulted in the attestation of additional Old Uighur toponyms, which have been identified as phonetic transcriptions of the earlier Chinese names for cities in the Turfan Basin. This article, however, discusses the fact that the Uighurs not only settled in former Tang cities, but also founded a number of new settlements in the region. Indeed, the Uighur era witnessed an unprecedented process of rapid urbanisation in the eastern Tianshan region. The urbanisation on the northern slopes of the Tianshan is particularly worth noting, since nomadic culture had long been dominant in the area, even when it was Tang territory. Many of the new urban settlements were established on the foundation of walled sites that had primarily served military purposes under the Tang. In contrast to the major oasis cities that existed before the Tang era, the military sites that evolved into cities during the Uighur era are a unique legacy left by the Tang.

The transition of the eastern Tianshan region during the Tang era

Before the seventh century, envoys and travellers from states in northern China normally travelled west using the routes along the northern and southern rims of the Tarim Basin. In contrast, due to its population of nomadic tribes and historical domination by steppe empires, the route along the northern slopes of the Tianshan was not recorded as a regular east–west route until the early seventh century. Pei Ju 裴矩 (d. 627), the Sui official in charge of trade with Sogdian merchants in Zhangye, listed it for the first time as a major route connecting Dunhuang and the Western Regions, running through the territory of the Turkic-speaking nomadic Tiele 鐵勒 tribes, as well as the court of their overlord at the time, the Western Türk Khaganate.² Recorded here for the first time in a Chinese dynastic history as one of the major routes, it was by that time dominated by a nomadic steppe empire.

Following the defeat of the Eastern Türk Khaganate on the Mongolian steppe in 630, the Tang began marching westwards, incorporating Yiwu (Hami), Gaochang (Turfan), and Bešbalïq (modern Jimsar County, Changji Prefecture, Xinjiang) into its territory. After suppressing the rebellion of Ashina Helu 阿史那賀魯 (the last ruler of the Western Türks) in 657, the Tang extended its direct military control over the entire Tarim Basin and the northern slopes of the Tianshan, where the Western Türks had previously dominated. Along with relocating the Anxi protectorate to Kuča and installing the four garrisons deep in the Tarim Basin,³ the Tang ‘opened up routes and installed postal stations’ for travellers in this region.⁴ In 692, when the Tang finally recaptured the four garrisons from the Tibetan empire, they left an unprecedented 30,000 soldiers there,⁵ ushering in an era during which the Tang occupied the Western Regions with massive standing armies. In 702, they also stationed a standing army in Beiting, establishing another protectorate specifically to strengthen the military control of the northern route along the Tianshan.⁶ With the influx of settlers from Tang territory, infrastructures for defence, transportation, and daily life developed. The route along the northern slopes of the Tianshan, once under the control of nomadic powers (as noted by Pei Ju several decades before), was gradually transformed into an area intended for sedentary Chinese travellers. To guard and support this route, a system of cities, garrisons, fortresses, and postal stations was established around Beiting, where Kehan Futu cheng 可汗浮圖城 ‘the stūpa city for the Khagan’, a headquarter for the Türks, was located.⁷ In a note under the entry for Beiting, the ‘Treatise on Geography’ in the Xin Tangshu records not only the cities (counties) within the territory of this protectorate, but also the large and medium-sized garrisons—jun 軍 and shouzhuo 守捉⁸—along the route:

60 li to the west of the west adjunct city of Beiting is located the medium garrison of Shabo 沙鉢. Further (west) is located the medium garrison of Pingluo 馮洛. 80 li further west is located the medium garrison of Yele 耶勒. 80 li further west is located the medium garrison of Juliu 俱六. 100 li further, one reaches Luntai 輪台 county. 150 li further is located the medium garrison of Zhangbao cheng 張堡城. Further on, one crosses the Li yi de jian 里移德建 river and 70 li further is located the medium garrison of Wuzai 烏宰. Further on, one crosses the Baiyang 白楊 river and 70 li further is located the large garrison city of Qingzhen 清鎮. Further on, one crosses the Yeye 葉葉 river and 70 li further is located the medium garrison of Yehe 葉河. Further on, one crosses the Heishui 黑 水 river, and 70 li further is located the medium garrison of Heishui. 70 li further is located the medium garrison of Donglin 東林, and 70 li further the medium garrison of Xilin 西林. Passing the Huangcao bo 黃草泊 lake, a large desert, and a small Gobi desert, one crosses the Shiqi 石漆 river, goes over the Cheling 車嶺 ridge, and reaches Gongyue 弓月 city.⁹

Additionally, under the entry for Yizhou, the Xin Tangshu notes two crucial locations along the westward route to Beiting: the medium garrison of Dushan 獨山 and Pulei 蒲類 County.¹⁰ The installation of a series of garrisons along the route reflects the fact that the region was located at the intersection between the nomadic population and the settled Tang immigrants.

The Uighur occupation of the Tang sites

As Uighur refugees fled to the eastern Tianshan region after the downfall of their Khaganate in 840, they took over the major cities there, including Xizhou (Qočo), Beiting (Bešbalïq), and Yanqi (Solmï), resulting in further prosperity in the following era.¹¹ Thanks to the decipherment of manuscripts dating to the Uighur era, it is clear that smaller cities of the Tang in the Turfan Basin were also taken over by the Uighurs. As already noticed by Takao Moriyasu, Old Uighur and Chinese manuscripts dating prior to the eleventh century sometimes describe the whole Turfan region as ‘the 22 cities’.¹² Since the number of cities in that region was exactly 22 at the end of the Qu Dynasty of Gaochang (501–640, the last local dynasty to rule in Turfan) and during the early-Tang era that followed,¹³ it seems that the Uighurs took over all the cities in the Turfan Basin. An increasing number of Old Uighur toponyms have been identified as phonetic transcriptions of the Chinese names of these cities during Tang or earlier times.¹⁴ Evidence for the overall Uighur occupation of Tang cities in Turfan can also be found in the Chinese text ‘Memo on the Merit of Building a Stupa’, dating to the early period of the West Uighur Kingdom, in which the governor of the entire Turfan region was referred to as being ‘in charge of the affairs of the four garrisons [Chin. fu 府, in the Garrison Millita system (Chin. Fubing zhi 府兵製)] and the five counties of Xizhou’ [in line (l.) 4].¹⁵ The numbers and names of the military and administrative units are identical to those during Tang times.

The rise of the Uighur chief Pugu Jun 僕固俊 from Beiting in 866 is recorded in various Chinese sources; it must have been an influential event at a time when the Chinese dynasty did not have direct connection with the Western Regions. Scholars generally accept this as the beginning of the West Uighur Kingdom, centred in Qočo and Bešbalïq (For the geographical setting of the kingdom, see Figure 1.)¹⁶ Although Moriyasu has already thoroughly surveyed and compared the different records,¹⁷ it is still worth discussing that contained in the Zizhi tongjian, which cites the 866 report by Zhang Yichao 張議潮, the governor of the Guiyi jun 歸義軍.¹⁸ His report lists Xizhou, Beiting, Luntai, and Qingzhen as examples of cities taken by Pugu Jun.¹⁹ It is not surprising that the pre-eminent cities of Xizhou and Beiting are mentioned in this report, but the references to Luntai and Qingzhen are of particular interest. Luntai—one of the three counties under the Beiting Prefecture during Tang times—was also the site where the Jingsai jun 靜塞軍 standing army was stationed.²⁰ Qingzhen, as mentioned above, was a large garrison city of the Tang, where a standing army should have been stationed.²¹ Together with Beiting, the centre of the protectorate, where a third standing army (Hanhai jun 瀚海軍) was stationed, Luntai and Qingzhen were the most important military sites under the Beiting protectorate during Tang times. The record in the Zizhi tongjian for the year 866, almost 80 years after the region had passed from Tang to Uighur control, suggests that Luntai and Qingzhen had ongoing significant roles in the region. It is convincing evidence that, in addition to the major cities, the Uighurs also took over the counties and even military sites of the Tang in the eastern Tianshan region (see Figure 2).

Figure 1. The geographical setting of the West Uighur Kingdom. Source: drawing by the author on a Google Earth map.

Figure 2. Uighur cities in Turfan Basin and on the northern slopes of Tianshan. Source: drawing by the author on a Google Earth map.

The same information can also be gleaned from the newly edited Old Uighur ‘annals’.²² Composed in the Mongol era, they record historical events that took place in the early days of the West Uighur Kingdom.²³ Sections R and S record the migration of a certain nomadic people, ‘the Six Tatar’, from the realm of the Khitans (‘Qïtay’) to that of the West Uighurs during the reign of their first ruler (i.e. in the early years of the kingdom).²⁴ This event can be connected with the pressure that the Khitans exerted on nomadic tribes during their early expansion from their home base on the south-eastern edge of the Mongolian steppe. During the years of Guangqi 光啟 (885–88), the Khitan king Qinde 欽德 took advantage of the chaos in the Tang when the defence along the northern frontier was empty, conquering many tribes, such as the Tatars (Chin. Dada 達靼), the Xi 奚, and the Shiwei 室韋.²⁵ The immigrant ‘Six Tatar’ nomads were initially settled ‘down (from) Bay Taγ as far as Qum Sängir’,²⁶ namely the belt that lies along the eastern rim of Dzungaria,²⁷ the gateway from the Mongolian steppe to the eastern Tianshan region. Later, ‘they flushed into the thriving cities of the Uighur khan, taking them as their cities and settling down’.²⁸ These cities would be located within the territory of the West Uighurs, in the eastern Tianshan region. In the following section W, the annals record: ‘As soon as they left their fine homes where they used to live, they came (to this region, taking it) as their homes for a long (stay), went and settled down in the lower (region) of Yangï Balïq.’²⁹ Given the reference to Yangï Balïq—a city located on the northern slopes of the Tianshan—we can further locate the aforementioned cities in the same region. Since there were only three counties in this region during Tang times, it is likely that many of the ‘thriving cities’, including Yangï Balïq (‘new city’), were either newly founded or had previously been Tang garrisons. A record by Qiu Chuji 丘處機 and his disciples more than three centuries thereafter confirms the latter. In the early thirteenth century, when the West Uighurs had recently submitted to the Mongol empire, the Taoist master (who was on a journey to Central Asia at the invitation of Chinggis Khan) visited their capital city, Bešbalïq. The local Uighurs introduced the city as Beiting from Tang times and revealed that ‘many of the frontier cities from Tang times still exist’.³⁰

From garrison sites to urban settlements

This section discusses the transformation of Tang-era garrison sites into urban settlements, or even major cities, under the Uighurs—something not previously noted by scholars. This transformation may be an even better indicator of the development in this region under the Uighurs than the prosperity of major cities such as Bešbalïq and Qočo.

From Chiting garrison to Čïqtïn city

There was already a civilian settlement in the oasis of Čiqtim during the Uighur era, where the modern town of Qiketai (七克台 < modern Uig. Čiqtim) is located. In Uighur manuscripts, this settlement is referred to as Čïqtïn (> modern Uig. Čiqtim)—a phonetic transcription of Chin. Chiting 赤亭, the name of the garrison set up on the same site in Tang times.³¹ Moriyasu has carried out a thorough survey of Uighur manuscripts unearthed from Čiqtim, illustrating many aspects of life in this civilian settlement during the thirteenth and fourteenth centuries.³² However, his assumption that a town already existed in Čiqtim before the Uighurs arrived³³ needs to be revised here.

Before the Tang occupation, the city of Baile (白艻), located in the oasis of Pičan (modern Pičan County, Turfan Prefecture), was considered the eastern frontier of the kingdoms in the Turfan Basin; no settlements are recorded at that time further east in the oasis of Čiqtim. In manuscripts unearthed from Turfan dating to the period of the Sixteen Kingdoms (304–439), the expression ‘guarding Baile’ (Chin. shou Baile 守白艻) is often attested, reflecting its position on the frontier.³⁴ During the Qu Dynasty of Gaochang, the city was also referred to as Dong zhen cheng 東鎮城, ‘the garrison city in the east’.³⁵ According to the biography of Xuanzang 玄奘, after a six-day journey through the desert, the famous monk made his first stop within the territory of Gaochang at Baile city, which was described as being located ‘on the frontier of Gaochang’.³⁶

After the Tang expansion into the Western Regions, a military outpost was set up to take advantage of the oasis of Čiqtim (located to the east of Pičan oasis). The earliest attestation of Chiting is found in the Chinese manuscript 67TAM78: 38, unearthed from the tombs in Astana, Turfan, which contains an administrative order issued from Puchang 蒲昌 County (modern Pičan = Baile under the Qu Dynasty of Gaochang) to the beacon tower at Chiting (Chin. Chiting feng 赤亭烽) for the delivery of provisions to the ten or so soldiers in the small garrison at Chiting (Chin. Chiting zhen 赤亭鎮).³⁷ Judging from the dates on other manuscripts unearthed from the same tomb, this can be dated to several years after 640,³⁸ the year in which the Tang captured Gaochang. The troops guarding Chiting numbered only ten or so at the time; their provisions would have been supplied from the adjacent oasis of Pičan, the previous frontier of the Turfan Basin. Other manuscripts reveal that postal stations for travellers (Chin. Guan 館) and for relay horses (Chin. fang 坊) were also set up there at around the same time.³⁹

Another document, reconstructed from several fragments unearthed from Astana, is a report from the small garrison of Chiting on two accidental deaths of long-distance relay horses in 705.⁴⁰ The meat of the dead horses in both cases was discarded in the wild, since there was ‘no one to sell to in the desert’.⁴¹ Although one of the incidents happened 35 li east of the garrison site, the other happened within the military colony (Chin. ying nei 營內 ‘in the camp’, l. 5), suggesting that the area of the oasis at that time was quite limited and there was no civilian settlement nearby, only desert.

By approximately the early 720s, the number of garrison soldiers in Chiting had reached 42, as revealed in manuscript 72TAM226: 51, unearthed in Astana.⁴² This manuscript also indicates that they had to cultivate a certain area of land to sustain themselves.⁴³ During the time of Tang colonisation, the area of the oasis was enlarged and thus it was able to support a larger population. The ‘Treatise on Geography’ in the Xin Tangshu refers to Chiting as a shouzhuo (i.e. a medium-sized garrison),⁴⁴ suggesting a growth in size of the garrison during the Tang period. Replacing the oasis of Pičan, the oasis of Čiqtim became the eastern gateway to Xizhou, namely the Turfan Basin.⁴⁵

The first attestation of a civilian settlement in the oasis of Čiqtim is the record of a temple from 982, when the Song envoy Wang Yande 王延德 entered the territory of Qočo and was received by Uighur officials.⁴⁶ As Moriyasu has already pointed out, the settlement located at the former Tang garrison of Chiting was referred to as a city or town (Uig. balïq) in Uighur manuscripts.⁴⁷ Although he has assumed that this urban settlement was much smaller than the major cities,⁴⁸ various sources actually indicate that the oasis was significantly developed. An Old Uighur provision order dating to the Mongol era lists Čïqtïn and the adjacent major city Pučang (< Chin. Puchang of the Tang time; > modern Pičan) as places to be taxed,⁴⁹ suggesting that the city of Čïqtïn was almost as important as Pučang, which had been among the five counties in the Turfan Basin during Tang times, when it functioned as the superior administrative unit over Chiting garrison. The development of the oasis of Čiqtim during the Uighur era can be further exemplified by the rise of another city in Mongol times, namely Töküz, which was depicted on the recently published Menggu shanshui ditu 蒙古山水地圖 (Landscape Map of the Mongols ⁵⁰) and transcribed in Chinese as Tuogusi 脫谷思 (see below). The oasis of Čiqtim, initially settled by Tang soldiers as a military colony, was thus a prosperous area with two cities within it by Mongol times.

From Dushan garrison to Dushan city

During the Tang era, the eastern gateway to Beiting was Dushan 獨山 (‘(at) the lone mountain’) garrison, in a relationship that was similar to that of Chiting to Xizhou. The ‘Treatise on Geography’ in the Xin Tangshu outlines the official route from Yizhou to Beiting, marking the following two stops: Dushan garrison and Pulei County.⁵¹ It is now well accepted that the site of Dushan garrison is the ruined city of Youku (Chin. Youku gucheng 油庫古城) in modern Mulei 木壘 (Mulei County, Changji Prefecture), which lies circa 90 kilometres away from the site of Beiting.⁵² The most recent official survey dates the founding of the city to Tang times,⁵³ though no archaeological excavations have yet been carried out.

The name of Dushan is attested in several different forms in later sources. According to his itinerary, after visiting the Mongol ruler Möngke Khan at Karakorum, Het‘um I, the king of Little Armenia (the Armenian Kingdom of Cilicia), travelled by a place named Berbalex, located before Bešbalïq on his itinerary.⁵⁴ James Russel Hamilton has associated this toponym with Mulei.⁵⁵ Dai Liangzuo 戴良佐 has proposed a Turkic etymon *bir-balïq (‘lone city’) and has correctly identified it with the toponym Bai balie 白拔烈 in the Liaoshi,⁵⁶ another transcription for the same etymon. Moreover, he has identified both the Dushan garrison of the Tang and the Dushan city of Mongol times as the Chinese name for the city of *bir-balïq.⁵⁷ It is now generally accepted that the Dushan garrison of Tang times continued to exist down to the Mongol era.⁵⁸ By the thirteenth century, however, the city must have evolved from a garrison site to an urban settlement. As the Yuanshi records, when Chinggis Khan travelled past the city site during his campaign to the west, he was informed that the ‘civilians’ (Chin. min 民) of the city had migrated to other places due to a famine that had occurred years earlier.⁵⁹

From Pingluo garrison to *Barlïγ city

Pingluo 憑洛 (or 馮洛)—a medium-sized garrison of the Tang located to the west of Beiting—was also transformed into a city after the Uighur occupation. According to Paul Pelliot, the earliest attestation of this site from the Uighur era is perhaps the bihä:rakä (l. 21) in the list of cities preserved in the so-called Staël-Holstein scroll, dated to 925,⁶⁰ although his identification is not generally accepted.⁶¹ The 12th chapter of the Persian source Hudūd al-Ālam (compiled in 982/83) describes the geography of the Toghuzghuz (commonly understood as the West Uighurs).⁶² The toponym bārlugh is listed among the five villages behind the Bogda mountain, including the winter capital Bešbalïq (Panjkath in the text).⁶³ The form can probably be traced back to a Turkic etymon *Barlïγ (< *Barslïγ, ‘(place) with tigers’), which could either be the older place name before the Tang occupation (later transcribed in Chin. as Pingluo) or an Old Uighur adjustment of the transcription of Chinese Pingluo due to folk etymology. In the latter case, Pingluo, pronounced as /pɦiəŋ lak/ in Late Middle Chinese,⁶⁴ might possibly have been pronounced as /pɦi lak/ in the north-western dialect after the 10th century;⁶⁵ it could thus be equated with bārlugh or its etymon *Barlïγ. Given its location and proximity to Bešbalïq, bārlugh can be identified with Pingluo from Tang times. Since the capital city Bešbalïq was also referred to as a ‘village’ in Hudūd al-Ālam, bārlugh was likely an urban settlement, just like Bešbalïq.

More evidence for the existence of an urban settlement can be found in Tarīkh-i Jahān-gushā (History of the World-Conqueror), the historical account of the Mongol empire composed by the Persian author Juvayni in the 1360s. In the chapter on Körgüz (a Uighur official in eastern Persia under the Mongols), Juvayni writes: ‘His birthplace was a small village four parasangs from Bešbalïq called Barligh in the western part of the Uighur country on the route followed by travellers through that region.’⁶⁶ The Persian unit parasang indicates the distance that travellers could travel in one hour; its exact value varied at different times. Considering data from different sources,⁶⁷ the distance from Bešbalïq to Barligh can be estimated to be 18–25 kilometres. As mentioned above, the distance from Beiting to Pingluo recorded in the Xin Tangshu is more than 60 li, namely 26.5 kilometres, roughly matching the distance from Bešbalïq to Barligh.

From Juliu garrison to Kunlü city

Juliu 俱六—another medium-sized garrison on the route between Beiting and Luntai during the Tang era—can also be located as an urban settlement during the Uighur times. Hamilton has identified K‘ullug on the itinerary of Het‘um I as the same toponym, proposing a Turkic etymon *köllüg.⁶⁸ On the historical world map Kangnido ⁶⁹ (compiled in 1402), one finds the toponym Yinliu 因六 along the route from Bešbalïq to Emil—the route that followed the northern slopes of the Tianshan. Chen Dezhi has correctly pointed out that the Chinese written form is a mistake for 固六 (Gu liu), therefore identifying the toponym with K‘ullug and Juliu.⁷⁰

I would like to add here another form of the same name, dating to the early thirteenth century. The biography of an Uighur elite, Xiban 昔班, in the Yuanshi records that his father, Que li bie wo chi 闕里別斡赤, when being awarded for his service during the campaign to the west with Chinggis Khan, requested to be made the darugači (governor) of the city of Kunlü 坤閭 in his own country, namely the West Uighur Kingdom.⁷¹ Kunlü has previously been accepted by several scholars as a transcription of the modern toponym Korla (Korla city, Bayingol Prefecture, Xinjiang).⁷² However, the fact that Korla is not attested in any other sources before the Qing era makes this identification very unlikely. Furthermore, the EM pronouciation of Kunlü can be reconstructed as /k^hun ly/ with rounded vowels in both syllables, contradicting the unrounded vowel in the second syllable of Korla. Kunlü can thus be identified with *köllüg. Since the Turkic denominal nominal suffix +lXg/lXγ was usually transcribed in Chinese with a loss of the final g/γ in Mongol times,⁷³ there is no significant difficulty in identifying the Chinese transcription Kunlü with *köllüg. Thus, the medium-sized garrison of Juliu in Tang times had developed into an urban settlement of Kunlü by the time of Chinggis Khan.

Urbanisation along the northern slopes of the Tianshan during the Uighur era

The defence system in the Beiting protectorate during the seventh and eighth centuries covered the routes that led to Beiting from Barsköl and Hami in the east, from the eastern rim of Dzungaria in the north, and from the Ili valley in the west, where the majority of the population were nomadic Turkic tribes subject to the Tang.⁷⁴ Only the three cities (counties) of Beiting, Luntai, and Pulei had sedentary civilian populations in addition to the standing armies. With the influx of Uighurs into the region, a significant process of urbanisation took place along the northern slopes of the Tianshan (see Table 1). Mahmud Kashgari's description of the West Uighur Kingdom in the 1170s reveals the surprising outcome of this process. The entry ‘Uighur’ in the Dīwān Luγāt at-Turk (Compendium of the Turkic Dialects) lists five major cities in the kingdom: Solmï, Qočo, Čambalïq, Bešbalïq, and Yangï Balïq.⁷⁵ For the first time, the eastern Tianshan region witnessed the major cities on the northern slopes (which had been dominated by nomads even during the Tang era) outnumbering the major cities in the oases along the southern slopes. Unlike Bešbalïq—a major city already established before the Tang era—both Čambalïq and Yangï Balïq were very likely established during the West Uighur Kingdom on the foundation of former Tang garrison city sites. It is generally accepted that Čambalïq evolved from the medium-sized garrison of Zhangbao⁸¹ and Yangï Balïq from the large garrison of Qingzhen.⁸²

Table 1. The Tang legacy on the northern slopes of the Tianshan in Uighur times

Apart from these three cities, the city of Qutaba likely also rose as a major city on the northern slopes of the Tianshan after Kashgari's time; its name is still retained to the present day as Qutubi (> Chin. Hutubi 呼圖壁; modern Hutubi County, Changji Prefecture). It was listed in the official records of the Yuan, along with the aforementioned three cities, as one of the four major cities on the northern slopes of the Tianshan in the realm of the Uighurs (Chin. Weiwuer di 畏兀兒地, = Persian Uighuristan).⁸³ Old Uighur manuscript U5265, unearthed in Turfan, is a private loan contract for a donkey to be used on a two-way journey to Qutaba⁸⁴ in which the city is mentioned as a destination for Uighur merchants. In the early thirteenth century, as recorded in Qiu Chuji's travel account, the region west of Čambalïq was populated mainly by Muslims⁸⁵ and Qutaba was the first stop on the way from Čambalïq to the west. The rise of Qutaba possibly resulted from its location as a gateway of exchange between the Buddhist Uighurs and the Muslims. As Hamilton has pointed out, the medium-sized garrison of Wuzai from Tang times should be located in Qutaba (modern Qutubi),⁸⁶ but the name Wuzai seems to have no connection with Qutaba/Qutubi, which is very likely derived from Arabic Qutbah. This, along with a lack of archaeological data from the oasis of Qutubi, makes it unclear whether the city was established on the foundation of the Tang-era garrison or as a new settlement. Nonetheless, the emergance of Qutaba is a good example of urbanisation on the northern slopes of the Tianshan after the twelfth century.

The growth of cities in the Turfan Basin

Unlike the northern slopes of the Tianshan, the oases along the southern slopes (i.e. the northern rim of the Tarim Basin) had a long history of sedentary culture in an urban setting. As mentioned above, major cities in this region, especially in the Turfan Basin, continued to enjoy prosperity after the Uighurs arrived. Rather than just adapting to urban life, the Uighurs actually promoted urbanisation in the region. According to the Yuan-era Tongzhi tiaoge 通制條格, there were at least 24 Uighur cities in the Turfan Basin by the year 1321,⁸⁷ thus outnumbering the 22 cities they had taken over from the Tang. This section will identify new cities that emerged in the Turfan Basin during the Uighur era, demonstrating the Uighur contribution to urbanisation beyond the legacy of the Tang.

The emergence of Töküz (Tuogusi) city

The Tang-era establishment of the military colony of Chiting in the present-day oasis of Čiqtim and its subsequent transformation into an urban settlement under the Uighurs has been discussed above. By the Mongol era, another city had emerged in the same oasis, reflecting further development in the area. In the eastern part of the Turfan region, the Menggu shanshui ditu depicts a square city with the name Tuogusi 脫谷思, not attested in any other Chinese sources.⁸⁸ Lin Meicun has proposed the Turkic etymon *toquz, simply based on the similarity in pronunciation.⁸⁹ Depicted on the map to the lower left of Bizhan 比站 (Uig. Pičan), its actual location should be to the north-east, as can be proven by the location of Lanzhen 懶真 (Uig. Lämčin) relative to Lucheng 魯城 (Uig. Lükčüng) on the same part of the map; it too is depicted to the lower left, although it actually lies to the north-east. One can therefore situate the city of Tuogusi in the oasis of Čiqtim—the only oasis located to the north-east of Pičan. Kenzheakhmet has identified it with a modern toponym Tekusi 特庫斯 located along the route from Čiqtim to Pičan,⁹⁰ which fits Tuogusi in both pronunciation and location.

This settlement can also be identified in two sources from Yuan times, thus dating its emergence to the second half of the thirteenth century. An Old Uighur loan contract obtained by the German Turfan expedition bears a toponym Töküz, which matches Tuogusi (EM. t^hᴐ ku sz) not only in pronunciation, but also in location. Its original shelf mark *T II 3 Čiqtim 3 indicates that the excavation site was the ruined city of Čiqtim. The contract records the rental of ‘a piece of land for the cultivation of crops in Töküz’ (l. 3, töküztäki tarïγ tarïmaqča yer) for ‘twelve liang paper money’ (l. 4, on iki stïr čo),⁹¹ suggesting that Töküz was very likely located not far from Čïqtïn city—that is, within the oasis of Čiqtim. The reference to the paper money čo (< Chin. chao 钞) narrows the date of the contract to a period between 1260, when Kublai Khan initially issued paper money, and 1304/05, when Uighuristan was incorporated into the Chagatay Khanate.⁹²

In the sixth month of the fourth year of Zhiyuan 至元 (1267), Yelü Xiliang 耶律希亮 returned to Yuan territory from Kuča via the oases of Turfan and Hami, as is well known from his biography in the Yuanshi. ⁹³ However, the text on his shendao bei 神道碑 (‘the stele erected on the path to his tomb’), identified as the basis for his biography,⁹⁴ contains a more detailed itinerary, including two stops between Turfan and Hami at Liuzhong 柳中 (> Uig. Lükčüng~Lükčün) and Jian hou zi 鑯堠子,⁹⁵ neither of which is included in his biography. Of particular interest is the toponym Jian hou zi midway between Lükčün and Hami, which has not been attested in any other sources. As Cen Zhongmian has correctly pointed out, Jian hou zi must have been located in the oasis of Pičan or the oasis of Čiqtim.⁹⁶ If we consider the character jian 鑯 ‘iron product’ as a variant of tie 鐵 ‘iron’, close in both meaning and orthography, this toponym can be identified with Töküz (> Tuogusi), located to the east of Pičan. *Tie hou zi 鑯堠子, in EM. /t^hjε xəw tsẓ/, roughly fits the pronunciation of Töküz and Tuogusi (EM. t^hᴐ ku sz), except that the first vowel is not rounded. If this identification is valid, the rise of Töküz as a major settlement along the route between Turfan and Hami can be dated prior to the year 1267.

The emergence of Sirkip city

There are no records of urban settlements along the route connecting the oases of Lükčün and Lämčin through the modern Sirkip Aγïz (‘Sirkip valley’) in any sources from the Tang era or before. The form tsīräkyepä retained in the Staël-Holstein scroll of 925 (l. 18) has been identified as the Khotanese transcription of Sirkip,⁹⁷ the earliest attestation of this toponym. The collocation with kaṃtha ‘town’ indicates that an urban settlement probably had emerged by 925, after the Uighur occupation of Turfan. The toponym has been attested in the form of tsirkip ⁹⁸ and sirkäp ⁹⁹ in Old Uighur manuscripts; according to Kitsudo Koichi, it can be traced to the name of a Buddhist temple Qiji si 七級寺 located at the very site in Tang times.¹⁰⁰

The existence of a city in Sirkip can be further proven in the Menggu shanshui ditu. A square city with the name Xi er qi 洗兒乞 is depicted between Lucheng and Lanzhen, indicating an actual location between the oases of Lükčün and Lämčin.¹⁰¹ Although Lin Meicun has proposed a Turkic etymon sarïčï,¹⁰² Xi er qi, in EM /si ṛ k^hi/, should rather be understood as the Chinese transcription of Sirkip,¹⁰³ which fits both the location and the pronunciation.

The emergence of Yangxi city

*Yangxi is another city that emerged during Uighur times midway between the oases of Lükčün and Qočo, where no urban settlement was recorded previously. In the Menggu shanshui ditu, it is recorded as Yanghei 羊黑 in Chinese and depicted as a square city.¹⁰⁴ A more accurate description is provided in the Xiyu tudi renwu lüe 西域土地人物略—an outline of lands and peoples of the Western Regions included in the provincial gazetteer for Shaanxi (Chin. Shanxi tongzhi 陝西通志) that was compiled in 1542. In it, Yanghei is recorded as being located to the north of Lükčün and 50 li to the east of Qočo.¹⁰⁵ Although the name has not been attested in its Old Uighur form thus far, we can relate it to the modern toponym Yangxi (> Chin. Yanghai 洋海; Yangxi Village, Pičan County, Turfan Prefecture).

Thus, in addition to cities that were either inherited from the Tang or established on the foundation of Tang-era military settlements, several new cities emerged during the Uighur era. It is noteworthy that most of these cities were located not within the densely populated oases that had a long history of urban culture before the Uighurs arrived, but rather along the major routes between these old cities, suggesting an expansion of oases in the Turfan Basin and a process of urbanisation during the Uighur era.

The Uighur tendency towards urbanisation

The general acceptance of the sedentary legacy of the Tang may well be attributed to the Uighur fondness for urban life that was already evident in the steppe period of their history, according to literary sources and recent archaeological data. Excavations of their steppe capital Ordu Balïq (Karabalgasun) confirm Tamīm ibn Baḥr's description of urban life there.¹⁰⁶ The site of Bay Balïq, built under the command of the second Uighur khagan,¹⁰⁷ has been identified as a complex of three cities according to archaeological surveys.¹⁰⁸ In addition to Uighur cities still in use during the subsequent Liao Dynasty,¹⁰⁹ the famous Mongol city of Karakorum may, according to recent archaeological data, have been built upon a walled site dating back to Uighur times.¹¹⁰ An increasing number of city sites with no record in literary sources have also been dated to the Uighur era.¹¹¹ The total number of cities built under the Uighurs may reach 40,¹¹² outnumbering any other nomadic steppe empire in premodern times.¹¹³ What made the Uighurs so special in this regard compared with their nomadic forerunners on the steppe? One may naturally relate it to their deep involvement, under Sogdian influence, in trade along the Silk Road—in particular the famous ‘horse and silk trade’ with Tang China after the Uighurs helped to pacify the An Lushan Rebellion.¹¹⁴ As the Uighur ruling class continuously obtained large volumes of silk and other high-value objects from the Tang, they perhaps naturally felt the need to build walled palaces and cities to demonstrate their status and to house their property. Similar opinion has been suggested as early as in the Chinese chronicle Zizhi tongjian: ‘As (Bögü Khagan) was meritorious to the Tang, they rewarded him amply. Thereafter, Dengli Khagan (i.e., Bögü Khagan) began to be arrogant, and built palaces to reside.’¹¹⁵ However, this alone seems insufficient to explain the unique fondness of the Uighurs for urban life—a fondness that differentiates them from their forerunners who likewise benefitted from the Silk Road trade.¹¹⁶ We must take into consideration a striking characteristic of the Uighur Khaganate, namely that a very large sedentary population formed in the core of their steppe territory.

In addition to his description of the sedentary life in and around the Uighur capital city, the Abbasid envoy Tamīm ibn Baḥr also revealed that villages and cultivated lands (i.e. a considerable sedentary population) could be found within 20 days’ journey of the capital.¹¹⁷ Whereas Zoroastrians coexisted outside the city, Manichaeans prevailed in and around it, according to his report. During the reign of Bögü Khagan (759–70), the Uighurs converted to Manichaeism under the influence of the Sogdians, becoming the first and only nomadic group in history to embrace this ideology as their state religion. After Bögü Khagan's conversion by the four Manichaean monks whom he brought from Luoyang, a Mahistag, who ranked as the third class in the hierarchy of the eastern branch of the Manichaean Church, led monks and nuns into the country and propagated the Manichaean teaching there,¹¹⁸ resulting in the establishment of a Manichaean monastic order in the central steppe. A recently published Old Uighur manuscript reveals that Bögü Khagan even invited three Možaks (Manichaean apostles), along with 60 senior priests, to the ‘realm of the Orkhun’ (el orxun, i.e. the centre of the steppe); they brought with them 200 scripture books (nom) to preach.¹¹⁹ As an increasing number of Manichaean monks came to settle, preach, and perform rituals in the steppe, monasteries, churches, and other sedentary infrastructures needed to be built.¹²⁰

In addition to a sedentary monastic lifestyle, the Manichaean doctrine also required a purely vegetarian diet. As noted by the Chinese chronicles, this precept was strictly practised by the Manichaean monks in the Uighur empire, who were obliged to drink water, eat spicy vegetables, and abstain from milk products.¹²¹ Since the Manichaean monks were not involved in productive activities, a certain number of the sedentary agricultural population was no doubt required to provide the monks and secular practitioners of Manichaeism with agricultural products that the traditional pastoralist economy did not produce. Moreover, as ‘the (Uighur) khagan usually consulted them on state affairs’, the Manichaean monks enjoyed high political privilege in the empire, so the actual number of sedentary population who were dependent on the Manichaean monasteries may have been much larger than just the farmers who provided their food.¹²² This lifestyle must have influenced the nomadic Uighurs, at least to some extent. The Karabalgasun Inscription, the official monument of the Uighur empire, records that ‘since they accepted the Teaching of Light (i.e., the Manichaeism), their barbarous practices full of bloodshed changed and their state became a country of vegetarians; the country where cattle were slaughtered was transformed into a place where good deeds were encouraged’.¹²³ No wonder Tamīm ibn Baḥr reported that the people in the villages who had cultivated lands were ‘Turks’.¹²⁴

Along with the Sogdian Manichaean influence, the Tang Chinese influence should also be considered as a major factor that promoted the Uighur tendency towards urbanisation. Unlike the Türks who proceeded them, the Uighurs succeeded in maintaining a generally peaceful and friendly relationship with the Tang. The An Lushan Rebellion severely undermined Tang rule in China, compelling them to turn to the Uighurs for military support in their fight against the rebels and later the Tibetans. This not only provided the Uighurs with a huge amount of wealth from trade and diplomacy, as mentioned above, but also promoted the exchange of personnel between the Tang and the Uighur empire, resulting in a strong cultural influence from the former to the latter. Uighur cities and other walled structures from the imperial time have been proven to be the result of significant Tang Chinese influence.¹²⁵ This can be attributed, in the first place, to the frequent introduction of sedentary Tang Chinese population to the steppe as a result of the long-term peaceful Tang–Uighur relationship (e.g. the diplomatic marriages that took place over several generations). At least three Uighur cities are recorded in Chinese sources as residences for the khatun (‘queen’) or Gongzhu (‘princess’).¹²⁶

Moreover, the Tang cultural influence on the lifestyle of native Uighur elites was crucial in the process of urbanisation. In his flight to the northern border of Tang in 841, the Uighur khagan even requested the Tang court to lend him and his Tang princess the city of Zhenwu jun 振武軍 to reside in,¹²⁷ indicating that the Uighur ruling class was already very used to urban life by the end of the empire. The Chinese sources suggest that this influence might have at least partly resulted from the large number of Uighur and Sogdian elites who visited or resided in Tang territory. As early as the year 779, the Tang court issued an edict that required the Uighurs and other foreigners to wear their own dress rather than Tang dress.¹²⁸ Only the Zizhi tongjian records the background:

Previously, the Uighurs who stayed in the capital usually numbered about 1000, and the Sogdians who wore Uighur dress and lived together with them multiplied the number. The city governor provided them with slaughtered or live cattle. They accumulated assets, established mansions, and obtained all the lucrative goods from the market …. Some wore Tang Chinese dress, seducing (local Tang women) to marry them. It is therefore prohibited.¹²⁹

Due to the frequent exchange of personnel between the two empires, large numbers of Uighur and Sogdian elites stayed in the Tang capitals and major cities after diplomatic and trade missions, accustoming themselves to the urban lifestyle of Tang elites and building their own mansions in Tang cities. One can see how thoroughly they merged into the Tang urban lifestyle as, when they wore Tang Chinese dress, they appeared to be Tang nobles. When these Uighur and Sogdian elites returned to the steppe, they likely brought their urban lifestyles back with them, along with a large amount of silk and other Tang luxury goods that they were now used to.

Concluding remarks

The incorporation of the Western Regions into Tang territory had profound impacts. Infrastructure serving the military colonies was set up along the route on the northern slopes of the Tianshan, making the previously nomadic region habitable for a sedentary population. After the Tang retreat, the Uighurs finally defeated the Tibetans in the 790s and began to take control of the region, occupying major cities, towns, as well as garrisons. A strong inclination towards urban life may already have formed among Uighur elites in the second half of the eighth century, prompted by multiple factors, including Sogdian Manichaean and Tang urban life influences. During the massive influx of Uighur and other steppe peoples into the eastern Tianshan region in 840, the rich legacy of infrastructure left by the Tang encouraged this inclination towards a sedentary life in the new homeland and facilitated Uighur settlement in urban environments. More than just occupying Tang urban infrastructure, the Uighurs contributed to an unprecedented rapid process of urbanisation in the succeeding centuries, especially on the northern slopes of the Tianshan. Along with traditional major cities such as Qočo and Bešbalïq, a number of new cities emerged under the Uighurs, either on the previously established Tang garrison sites or as new urban settlements. The mercantile culture of the Uighurs, heavily influenced by the Sogdians, developed significantly after they migrated to the eastern Tianshan region, at the crossroads of the Silk Road.¹³⁰ Subsequent rapid urbanisation should also be regarded as an outcome of economic prosperity in this region. By the Mongol times, some 400 years after they inherited the Tang legacy, the cities in the eastern Tianshan region had already taken deep root in their own tradition and reshaped their memory. According to Juvayni in the thirteenth century, the capital city of Bešbalïq was recorded, in the books of the Uighurs, as having been built by themselves when they migrated there.¹³¹