On the origins and emergence of the Qaŋlï Turks

Abstract

The Qaŋlï (Qangli) Turks were a numerous people, active in Eurasia in the twelfth to thirteenth centuries, but their ultimate origins remain a matter of debate. Often considered by modern scholars to be a part of the Kipchaks (Cumans), others have different opinions. One of these links them to cart-riding early medieval Turkic tribes called Tägräks, known in Chinese sources as Tiele 鐵勒, among other forms. This article examines the earliest possible (eighth-century) references to the Qaŋlïs in the historical sources, and points to the potential links between them and various tribes seen among Turko-Mongol groupings of the ninth to tenth centuries mentioned in the Chinese sources, such as the Black Carts (Heichezi 黑車子). Another aspect that this article focuses on is how both historical and mythological texts of the Mongol period show the Qaŋlïs to be a people distinct from the Kipchaks. Ultimately, this study, which is based on both historical sources and modern research, proposes to locate the origins of the Qaŋlï Turks among Tägräk tribes.

Introduction

Qaŋlï (lit. “cart”; also spelled as Qangli and Kanglı)¹ was the name of a medieval Turkic people who had important roles in Eurasia between the twelfth and thirteenth centuries. Most famously known for their significant political and military positions in the late Khwarazmshah Empire, they also briefly served the Anatolian Seljuks and Ayyubids, even capturing Jerusalem in 1244 and triggering the Seventh Crusade of 1248–54 led by King Louis IX of France. Most of the available information regarding the Qaŋlïs begins in the twelfth century; they seem to have inhabited a very large territory, starting from modern Gansu 甘肅 in northwest China, stretching westwards to the Caspian Sea and Ural Mountains. Large groups of Qaŋlïs submitted to the Mongol Empire and eventually dispersed as a separate people, as seen with many other Turko-Mongol groupings of that period, but numerous groups of Qaŋlï origin can still be found among many modern Turkic peoples. Often considered as the Asian branch of the Kipchaks (also known as the Cumans), very little is known about them before the twelfth century and their eventual origins remain a matter of debate.² My opinion is that the Qaŋlïs are possibly connected with the Tägräks, early medieval cart-riding nomadic Turkic tribes that roamed the vast Eurasian steppes.³ This article aims to shed light on the origins and emergence of the Qaŋlï Turks based on sources from earlier historical periods.

Modern scholarly opinions on the origins of the Qaŋlïs

For more than a century, modern scholars have worked to uncover the origins of the Qaŋlï Turks who spread over large parts of Central Asia. They have suggested the following different origins:⁴

• • Kangju 康居 (Kanguj in the works of Russian researchers), which starts to appear in Chinese sources in the second to first centuries bc, located on the banks of Jaxartes to the west of Zhetysu (Semireč’e) and east of the Aral;⁵

• • Kangguo 康國 (Kang Country), also seen in the Chinese sources;

• • Avestan Kañha (Kangha);

• • Mixture of the Kangars (known as the ‘noble tribes of Pechenegs’), Oghuz, Kimäks and Cuman-Kipchaks;

• • Eastern branch of the Kipchaks;

• • Continuation of Gaoche (High Carts) tribes of the fourth to sixth centuries.

Of course, there were some objections to these hypotheses.⁶ Scholars who attempted to identify the origins of Qaŋlïs among the Kangju/Kañha on the banks of Jaxartes viewed the name Qaŋlï more as an Iranic or Tokharian toponym than a Turkic ethnonym. However, this study demonstrates that the earliest groupings carrying the name Qaŋlï appear in mid-eighth century Southern Siberia, not in Turkestan.

Earliest possible mentions of the Qaŋlïs

Dilek and Yorulmaz drew attention to the fact that an affinity and kinship between the Qaŋlïs and Uyghurs is stressed in both various historical sources and the work of some researchers.⁷ Kumekov wrote that in the second half of the ninth century, an Uyghur emissary mentioned a name in the form Qara Qanglï (Kara Kanğlık or Qara Qanγlïq) who dwelt in the Altai region neighbouring the Basmïls, Toquz Oghuz and Qarluqs. According to him, it is somewhat difficult to interpret this name.⁸ Even though Kumekov did not note where he acquired this information from, the source he mentions is the BPN, written inside the manuscript Pelliot tibétain 1283 currently kept in the Bibliothèque nationale de France in Paris. A legendary people comprised of giants is mentioned between lines 50 and 60 of this text, which is the Tibetan translation of an Uyghur intelligence report whose original is generally thought to date from the 750s. According to the document, a mountain range covered with deserts was located to the north of a tribe of the Kyrgyz (Gir tis གིར་ཏིས་) and a people of giants lived to the north of these mountains. In the description of the culture of this people, it is mentioned that the Kyrgyz (Hir kis ཧིར་ཀིས༌) sent a messenger to them. The giants tied up and tortured this messenger, then they asked “Our herdsman of calves and sheep, the one in the clan which is called Ga ra gang lig ག་ར་གང་ལིག, where does he dwell?”, which is in line 57 of the document translated by Venturi. To the north of this people of giants lived the Basmïl (Ba sme བ་སྨེ་) tribe, who established a confederation with the Uyghurs (Hor ཧོར་) and Qarluqs (Gar log གར་ལོག་). In the following parts of the narrative, the destruction of the Türk Qaghanate by this alliance is described.⁹

Pelliot indicated that the equivalent of the name Ga ra gang lig mentioned here could be Qara-Qanglï(γ) (literally “Black Qaŋlï[γ]”), although he also stated that he was not entirely certain of this conclusion.¹⁰ Other scholars who studied this document, such as Clauson, Tezcan, Moriyasu and Venturi, did not give any explanations of the name Ga ra gang lig.¹¹ As Yorulmaz also indicated, this Tibetan document, which I examine, is the first source that mentions the name Qaŋlï. Since the name Black Qaŋlï is mentioned in the document, Yorulmaz stated that groupings carrying the names Yellow Qaŋlï or White Qaŋlï should also have existed. On the basis of the document, according to him, the Black Qaŋlïs must have lived to the southeast of the Altai Mountains, in the regions stretching from Eastern Turkestan to northwest China. The name Black Qaŋlï can be seen among the Kyrgyz Qaŋlïs of the Zhetysu region in much later periods. In addition, the Black, Yellow and Red Qaŋlïs are mentioned in a Kazakh šežire (šajara, genealogy) recorded in the nineteenth century. This genealogy describes that Qaŋlïs on the banks of the Jaxartes are from the Yellow Qaŋlï, while Qaŋlïs in Zhetysu comprise the Yellow and Black Qaŋlïs. Yorulmaz emphasized that using colours in the naming of Qaŋlïs in this manner is related to a practice indicating the geographical or administrative statuses of tribes among Turkic peoples.¹² As is well known, the word Qara in Old Turkic had a variety of meanings, including “black”, “northern”, “strong” and “commoner/subject”.¹³ Here, in the case of the Black Qaŋlïs, it is not easy to determine whether they are “Northern Qaŋlïs” or “Commoner/Subject Qaŋlïs”, but I think the latter is more likely.

Apart from this document in Tibetan, there is a recently discovered Uyghur document in Old Turkic that mentions the Black Qaŋlïs. A Uyghur manuscript fragment 81TB10: 06–3a, giving information about Uyghur conversion to Manichaeism, was discovered at Bezeklik in 1981 and first published by Zieme. Here, Manichaean priests from the West are described as bringing religious scriptures and bolts of silk to the Orkhon Uyghur Qaghanate during the reign of Bögü Qaghan in the 760s. In line 7 of the document, it is mentioned that they journeyed via Tuγurïstan (Eastern Turkestan) and Qara Qaŋlï, and they reached the qaghanate after passing through Ärḍiš (Irtysh). Both Zieme-Wang and Moriyasu identified these Qara Qaŋlïs as the Black Carts (Heiche 黑車/Heichezi 黑車子) mentioned in the Chinese sources I will examine below, a view with which I agree.¹⁴ This geographic designation of the Black Qaŋlïs in the documents points to a location most probably around Dzungaria (Jungaria) and the Altai Mountains.

The earliest possible mention of the name Qaŋlï in Chinese sources might be under the title Horse Marks [Tamghas] of Various Administrative Centers (Zhu Jian Mayin 諸監馬印) found within the section dealing with military matters in THY. Here, there is a list of Turko-Mongol tribes living to the north of China as well as the tamgha signs these tribes branded on their horses. This list includes a tribe with the name Kangheli 康曷利 and it is stated that its tamgha looks like the Chinese character zhai 宅.¹⁵ In order to study this ethnonym, one should also check the Middle Chinese (MC) reconstructions of this name proposed by various Sinologists:

• • Karlgren: γɒng-γât-lji- (Ancient Chinese);

• • Pulleyblank: kʰaŋ-ɣat-liʰ (Early MC) and kʰaŋ-xɦat-liˋ (Late MC);

• • Coblin: *khaŋ and *li for the Sui-Tang period Chang'an 長安 dialect, mid-Tang Chang'an dialect and ninth- to tenth-century pronunciations of kang 康 and li 利; *ɣat for the Old Northwest Chinese pronunciation of he 曷;

• • Baxter and Kroll: khang-hat-lijH (MC);

• • Schuessler: kʰâŋ-ɣât-li^C (MC).¹⁶

Given this name's transcription as kan-hé-li and its MC pronunciation as kang-γat-lji, Zuev maintained that the name's Chinese transcription reflects kangγatlï or, in other words, Kangarlïγ. According to him, this name is one of the oldest variants of the ethnonym Kanglï (Qaŋlï). Matsuda thought that Kangheli is the equivalent of Kängäräs in the Orkhon Inscriptions, but Zuev emphasized that the plural suffix -s in Kängäräs is not seen within Kangheli.¹⁷ Although Zuev's hypothesis looks tempting, the character he 曷, which seems to be rendering Old Turkic *-γar-, poses some problems so the equitation Kangheli 康曷利 = Qaŋlï needs additional evidence, as I will point out again below.

Also according to Zuev, it is possible to see the tribal name Kangheli within the title of a Türk nobleman who lived in the seventh century. The title of this nobleman in question is recorded in JTS's Account of the Türks as Tigin Kangshaoli (Teqin Kangshaoli 特勤康稍利),¹⁸ while in THY's Account of the Northern Türks it appears as Kangshaoli 康鞘利.¹⁹ Although Zuev did not refer to him, this Türk nobleman is also mentioned in other Chinese sources. The person in question is mentioned in these sources as follows:

• • Da Tang Chuangye Qijuzhu 大唐創業起居注: Pillar of the State Kangshaoli (Zhuguo Kangshaoli 柱國康鞘利), Kangshaoli 康鞘利 and Shaoli 鞘利;

• • TD's Account of the Türks: Kangshaoli 康稍利;²⁰

• • Liu Wenjing's 劉文靜 biography in JTS: Commander Kangshaoli (Jiang Kangshaoli 將康稍利);

• • CFYG and XTS's Account of the Türks: Tigin Kangshaoli (Tele Kangshaoli 特勒康稍利);

• • ZZTJ: Pillar of the State Kangshaoli 柱國康鞘利 and Kangshaoli 康鞘利.²¹

Implying that the characters such as shao 鞘 and shao 稍 used in the writing of this title are scribal errors, Zuev changed them with the character he 曷 and thought that the title of the aforementioned Tigin should be Qangarlïγ Tigin. According to him, the Qangarlïγ must have been a tribe subordinate to the Eastern Türk Qaghanate, and when the timespan of Kangshaoli Tigin going to and from China is considered, they were not far from Shibi Qaghan's 始畢可汗 administrative centre.²²

I have some hesitation regarding Zuev's opinion that Kangshaoli 康稍利/康鞘利 is a scribal error and its correct form should be Kangheli 康曷利. If Kangshaoli 康稍利/康鞘利 was mentioned in only a single source or the characters shao 鞘 and shao 稍 looked quite similar to the character he 曷, I might have leaned towards this possibility. However, the fact that this title is written in seven different Chinese sources with the characters shao 鞘 and shao 稍 makes me suspicious that a scribal error exists here. Besides, considering that Kangshaoli carries the title Tigin (Prince) might be pointing to the fact that he belonged to the imperial Türk dynasty from the Ašïna (Ashina 阿史那) rather than a tribe subordinate to the Eastern Türk Qaghanate. As is already known, Tigin is a title generally given by the Türks to members of the dynasty, with only a few exceptions.²³ In the family tree of the Türks he prepared in his book, Taşağıl showed Kangshaoli Tigin among Shibi Qaghan's sons, indicating that he is a member of the dynasty. Togan et al. also accepted this person as a member of the Türk ruling house.²⁴ Except for the problem of identifying Kangshaoli with Qaŋlï, I am inclined to consider Zuev's opinion as a possibility, which accepts Kangheli 康曷利 as one of the earliest known examples of the ethnonym Qaŋlï, though I do not accept this for certain. The reason why I hesitate to entirely acknowledge this possibility is because Kangheli's MC reconstructions, as also shown by Zuev himself, probably reflect *Qaŋγarlïγ. If Kangheli 康曷利 did not have the character he 曷, it would have been much easier to identify this ethnonym with Qaŋlï; however, the existence of this he 曷 causes problems. The reconstruction *Qaŋγarlïγ itself points to the ethnonyms and toponyms Qangar, Kängü and Kängäräs that are to be found in Western Turkestan. However, the early Qaŋlï groupings mentioned in BPN, 81TB10: 06-3a as well as in the Chinese sources regarding the ninth century, which I discuss below, appear to have dwelt in Southern Siberia, the Altai Mountains and Dzungaria, areas located to the northeast of that region. Thus, additional information and more research seem to be necessary to fully confirm the identification of Kangheli 康曷利 (*Qaŋγarlïγ) with Qaŋlï.

“Black Carts”: Heichezi 黑車子

A grouping similar to the name of the Black Qaŋlïs mentioned by BPN and 81TB10: 06-3a emerges exactly a hundred years after the mid-eighth century, the time when events in these documents are described. As far as I could determine, Chinese sources start mentioning a grouping named the Black Carts (Heichezi 黑車子) for the first time in the year 841. The earliest record about them is in HCYPJ. Alongside the Tatars (Dada 逹怛), the Black Carts are mentioned in this work among the tribes (buluo 部落) subordinate to the Uyghurs in around 841.²⁵

According to the Chinese sources, the Uyghurs (Huigu 回鶻) dispersed and migrated to different regions after the destruction of the Uyghur Qaghanate by the Kyrgyz in 840. The Uyghur nobleman Ögä Tigin (Wujie Tele 烏介特勒) became leader of one of these groups and acquired the title Ögä Qaghan (Wujie Kehan 烏介可汗). However, as a result of Ögä Qaghan's raids into China and especially to the You Prefecture (Youzhou 幽州), located in today's Beijing 北京, the Chinese, Desert Turks (Shatuo 沙陀), Qibi 契苾 and other groupings united in 842 to launch a surprise attack on him. After being defeated, Ögä Qaghan fled to the Black Carts (Heichezi 黑車子) and the remainder of his troops surrendered to the Tang Dynasty at You Prefecture. During the diplomatic talks conducted between the Kyrgyz and Tang Dynasty in 843, it was decided that the Kyrgyz had to attack the Black Carts in order for the Kyrgyz Qaghanate to be formally recognized by the Tang Dynasty. As a result, the Kyrgyz and Chinese agreed in 844 to conduct a mutual operation against the Uyghurs and Black Carts. However, the operation could not be carried out in 845, and so in 846 the Chinese gave up this plan. Ögä Qaghan was killed in 846 and succeeded by his brother Enian Tigin (Enian Tele 遏捻特勒). Their numbers seriously depleted, during the years 847–8 this Uyghur group took refuge among the Qai (Xi 奚, Tatabï) and Shirvi (Shiwei 室韋), who were among the Mongolic peoples of Manchuria. After the group was dispersed as a result of Chinese pressure, Enian Tigin and his family fled westwards, while the remaining people were ruled by the Shirvi. Yet, the Kyrgyz attacked the Shirvi, taking the Uyghurs to the north of the Gobi Desert (Qi 磧). Those Uyghurs that managed to flee to the forests and mountains migrated westwards to unite with their kinsmen in Eastern Turkestan.²⁶

In an edict recorded in HCYPJ, Li Deyu 李德裕 wrote that the Black Carts lived in the desert located a thousand li 里 (~500 kilometres) away from the Chinese border. Remarking that Chinese armies could never have reached those regions in the past, Li Deyu stated that this group sought to go to Eastern Turkestan (Anxi 安西) after realizing the Uyghurs’ real intentions. A copy of the same edict exists in the CFYG. From some of the records in Chinese sources that give information about Ögä Qaghan's taking refuge among the Black Carts and his eventual murder, one can also acquire some understanding of the geographical position of this grouping around the year 846. As I mentioned before, it is recorded in Biographies of Zhang Zhongwu in JTS and XTS that this qaghan wanted to flee to Kängü (Kangju) located on the banks of the Jaxartes or, in other words, in Western Turkestan. Also, according to the Account of the Uyghurs in JTS, the qaghan was killed in the Altai Mountains. If the qaghan was indeed killed by Yiyin Chor in the Altais, then he must have left the Black Carts with whom he had previously taken refuge.²⁷ The fact that Ögä Qaghan's people took refuge with the Qai and Shirvi following his death indicate that the Black Carts, with whom the qaghan had previously taken refuge, were living in a region close to those peoples, most probably in the Khingan Mountains.²⁸ Again from another record in JTS, it is understood that the Black Carts controlled the regions between China and Eastern Turkestan in the mid-ninth century. According to the Imperial Annals of Xuanzong 宣宗 in JTS and ZZTJ, a legation sent by the Tang Dynasty to the Uyghurs in the tenth month of Dazhong's 大中 eleventh year (22 October–20 November 857) had to return to China because the Black Carts had blocked their path.²⁹As is known, most of the Uyghurs in this period now resided in the Turfan (Qocho/Qochu, Gaochang 高昌) region, while some of them lived in Gansu. As the Black Carts blocked the path of the legation going from China to Eastern Turkestan, they must have been located around Gansu-Turfan because anyone who wanted to travel to Eastern Turkestan from China absolutely had to cross from Gansu. The records in BPN and 81TB10: 06-3a also indicate the presence of Black Carts in regions close by. However, if one considers the fact that the Uyghur grouping of Ögä Qaghan took refuge with the Qai and Shirvi following his death, it should be accepted that the Black Carts with whom the qaghan had taken refuge were living in the east. Yet, the Black Carts blocking the China-Eastern Turkestan path in 857 must have been dwelling in the west, quite far from there. In this case, the existence of Black Carts groupings living in both the east and the west but carrying the same name can be considered.

The same name Black Carts (Heichezi) can also be seen in Hu Qiao's 胡嶠 work Xianluji 陷虜記, in which he described his service to the Khitans (Qidan 契丹) between 947 and 953. The text of this work, which gives some information about the Khitans, is transmitted in Xin Wudaishi 新五代史 and QDGZ. According to Hu Qiao, the Khitans’ 契丹 neighbours were: the Qai 奚 to their south, the Türks 突厥 and Uyghurs 回紇 to their west and a people named Yujuelü 嫗厥律 to their northwest. To the west of the Yujuelü lived the Kyrgyz (Xiajia 轄戛) whose northern neighbours were the Chanyu Turks 單于突厥, and to their north lived the Black Carts 黑車子. They were good at making tents with carts (chezhang 車帳) (that is to say, tents pulled on carts), their people showed respect to family elders and their lands were poor and barren. According to a rumour, an ancestor of Khitans frequently served the Uyghurs; later he fled to the Black Carts and started learning how to make tents with carts. To the north of the Black Carts lived the Turks with Oxen Hooves (Niuti Tujue 牛蹄突厥), and to their northwest, a people named Wajiezi 韈劫子. To the three sides of these peoples, whom the Khitans and other states feared, were located the Shirvi 室韋 groupings comprising the Shirvi, Yellow-head Shirvi (Huangtou Shiwei 黃頭室韋) and Wild Shirvi (Shou Shiwei 獸室韋). To the north of the Wajiezi is the Country of the Dogs (Gouguo 狗國) where a legendary people with dog heads lived.³⁰ The Black Carts mentioned by Hu Qiao seem to coincide, from a geographical point, with the Qara Qaŋlïγ of BPN and Qara Qaŋlï of 81TB10: 06-3a, that is to say the Black Qaŋlïs.

Yet, in a note placed in the section of ZZTJ where Ögä Qaghan's refuge with the Black Carts is narrated, it has been explained that they are a Shirvi grouping. According to this note, if the New and Old Books (Xinjiushu 新舊書) are examined in detail, it can be seen that the Black Carts are a branch (zhong 種, literally “offspring”) of the Shirvi. When an imperial edict was sent to the Kyrgyz, the Black Carts were a thousand li away from the Chinese border. ZZTJ also transmitted a quotation from the Kaoyi 考異 (Zizhi Tongjian Kaoyi 資治通鑑考異) made from JTS's Account of the Uyghurs. According to this, Ögä Qaghan fled 400 li (~200 kilometres) towards the northeast following his defeat and took refuge with the Hejie Shirvi (Hejie Shiwei 和解室韋),³¹ forged kinship with the Shirvi via marriage and became allies with them. After this information is given in the note in ZZTJ, the record in Zhang Zhongwu's biography in JTS regarding the qaghan's desire to flee to Kängü and his refuge within the Black Carts is quoted. As the note continues, there is a quotation from the Li Deyu Jishenggong Bei 李德裕紀聖功碑. It narrates that Ögä Qaghan intended to secure himself via the Dingling 丁令, desired to flee to Kängü and took refuge with the Black Carts (Heiche 黒車). Before the note ends, it states that the ancient Xiongnu ruler Zhizhi Chanyu 郅支單于 also wanted to take refuge with the Kängü but this was a mistake; thus, Ögä Qaghan is likened to Zhizhi Chanyu.³²

There are records in LS regarding a grouping carrying the name Black Cart Shirvi (Heichezi Shiwei 黑車子室韋). Abaoji 阿保機, founder of the Khitan Empire, attacked the Black Cart Shirvi twice, in 904 and 905. After Abaoji attacked them again in 907, eight tribes 部 of this aforementioned grouping surrendered to him. The Khitan ruler organized another campaign and defeated them in the same year. In the following year, Abaoji again ordered his brother Sala 撒剌 to attack the Wuwan 烏丸 and the Black Cart Shirvi. Towards the end of 909, a Khitan army attacked the Black Cart Shirvi and defeated them once again.³³ According to the records in LS, this grouping paid tribute to the Liao Dynasty ruled by the Khitans in the years 938, 940, 944 and 945.³⁴ The Black Cart Shirvi have been mentioned in the Armies of Vassal Countries (Shuguo Jun 屬國軍) list found in LS's Monograph of the Military (Bingweizhi 兵衞志).³⁵ In a similar manner, the Governorship of the King of the Black Cart Shirvi Country (Heichezi Shiweiguo Wangfu 黑車子室韋國王府) is given among the Officials of Vassal Countries in the Northern Regions (Beimian Shuguo Guan 北面屬國官) in the Monograph of the Hundred Officials (Baiguanzhi 百官志, also referred to as Monograph of Various Ranks and Positions).³⁶

Wang Guowei 王國維 thought that although the original homeland of the Heichezi Shiwei mentioned in LS was located to the north of the Khitans, they must have migrated south towards the region of Yin Mountains (Yinshan 陰山, Čuγay Yïš and Čuγay Quzï of the Türk inscriptions), somewhere in today's Chakhar region, according to Wittfogel and Fêng.³⁷ The latter have also accepted the Heichezi, the grouping who taught the Khitans how to make carts with tents, as a Shirvi tribe, and thus the same grouping as the one mentioned in LS.³⁸ Ahmed Zeki Velidî Togan thought that the Heichezi were the Usuns (Wusun 烏孫) who lived to the east of Lobnor in the regions stretching towards Koko Nor and later became vassals of Chinggis Qan (Genghis Khan). Çandarlıoğlu accepted the Hei-ch'e-tse (Heichezi) as a Mongolic people together with the Tatars. Based on Wang Guowei, Çandarlıoğlu was of the opinion that during their migrations after 840, the Uyghurs brought Tatars and Heichezi with them to the Ganzhou 甘州 region in the south.³⁹ Tsai believed that the Heichezi mentioned in the Chinese sources describing the dispersion of Uyghurs after 840 belonged to Shirvi tribes and he translated the ethnonym Heichezi into Turkish as “Kara Arabacı” (“Black-cart [maker]”). Moreover, this people was also called Black Tatar according to him. As to Taşağıl, he translated the name Heichezi into Turkish as Kara Arabalı Oğulları (Sons of the Black Carts).⁴⁰ Drompp was also of the opinion that the Heichezi of the 840s were from the Shirvi.⁴¹ According to him, the Heichezi should be identified with the Hejie Shirvi.⁴² Although the Black Carts are mentioned in ZZTJ and LS as a Shirvi tribe, no tribe bearing this name can be found in the lists of Shirvi tribes given in TD, JTS, THY and XTS's Accounts of the Shirvi.⁴³ The Black Carts with whom Ögä Qaghan took refuge in the ninth century and who are understood to have been living in the Khingan Mountains might be the Black Carts Shirvi, who fought with the Khitans and became their vassals in the tenth century. However, groupings named the Black Carts are also seen in regions like Southern Siberia, Dzungaria and the Altai Mountains during the eighth to ninth centuries. Besides, during the same period when the Black Carts Shirvi existed, Hu Qiao described the region where the Black Carts lived as in Siberia to the north of the Kyrgyz and mentioned them as a people separate from the Shirvi. What is to be understood from this is that there were most probably two distinct groupings that carried the name Black Carts, in other words Black Qaŋlïs: one between Siberia and Gansu and the other in the Khingan Mountains.

I should point out that these earliest mentions of the Qaŋlïs in Tibetan and Chinese sources are quite a bit earlier than the appearance of the ethnonym Kipchak. I consider this possibility to be evidence that supports suspecting the opinions of scholars linking the Qaŋlïs with the Kipchaks, who believe that the Qaŋlïs first appeared as a Kipchak sub-group and broke away from them in the eleventh to twelfth centuries.

The Qaŋlïs as a distinct people

Most of what we know about the Qaŋlïs concerns their activities during the twelfth and thirteenth centuries, when they became a prominent people in a vast area stretching from the Urals to Gansu. This information is generally found in Islamic, Chinese, Mongol and European sources that do not give much information about how they appeared and where they came from. There are, of course, some exceptions, which label the Qaŋlïs as a part of the Kipchaks. Except for the records about possible early Qaŋlïs I listed above, most of the historical sources start mentioning the Qaŋlïs as a prominent people from the early twelfth century. Prior to that, they are usually not counted as a major people; some lists of prominent Turkic peoples found in tenth- and eleventh-century sources do not mention them. One of these sources – the tenth-century anonymous Persian geography work Ḥudūd al-ʿĀlam (section 12–22) – gives a list of Turkic peoples and their descriptions. This list includes Toγuzγuz (Toquz Oghuz, Uyghurs), Yaγmiyā (Yaghma), Xirχīz (Kyrgyz), Xalluχ (Qarluq), Čigil, Tuχs (Tokhsï/Tukhsï), Kimäk, Γūz (Oghuz), Turkish Bajanāk (Pecheneg), Xifčāχ (Kipchak) and Majγarī (Magyar), but not the Qaŋlï.⁴⁴ S.6551, a tenth-century Chinese text from Dunhuang 敦煌, listed the neighbours (most of them Turkic peoples) of the Turfanese Uyghurs as Qarluqs (Gelu 葛祿), Yaghmas (Yaomo 藥摩), Strange Looking Tatars (Yimao Dadan 異貌達但), Tokhsï/Tukhsï (Duxi 獨西), Tibetans (Tubo 土蕃), Basmïls (Baximi 拔悉密) and Kyrgyz (Xiajiasi 黠戛私), but does not mention the Qaŋlïs.⁴⁵ Similar to Ḥudūd al-ʿĀlam, Maḥmūd al-Kāšγarī provides a list of major Turkic peoples in his eleventh-century work DLT. Stating that the Turks were actually composed of 20 tribes (qabīla ﻘَﺒِﻴﻠَﺔً), Kāšγarī listed these as Bäčänäk (Pecheneg), Qïfčaq (Kipchak), Oγuz (Oghuz), Yämäk, Bašγïrt (Bashkurt), Basmïl, Qay, Yabaqu, Tatar, Qïrqïz (Kyrgyz), Čigil, Toχsï/Tuχsï, Yaγma, Oγraq, Čaruq, Čomul/Čömül, Uyγur (Uyghur), Taŋut (Tangut), Xïtay (Khitan) and Tawγač (Tabghach). As can be seen, the Qaŋlï are not on this list.⁴⁶ All these accounts demonstrate that although the Qaŋlïs were most probably living further northwards in Southern Siberia, Dzungaria and the Altais, they were still mostly outside the range of sedentary sources, perhaps also still rather small in population.

As I mentioned earlier, there are differing opinions both in the historical sources and between scholars on the origins of the Qaŋlïs and to which Turkic people they originally belonged. Among these views, the most widespread is the one that Qaŋlïs were the eastern branch of Kipchaks. The reason for this is that the Qaŋlïs were also called Kipchaks or Cumans in some Islamic and Mongol-period European sources. Scholars who maintain that the Qaŋlïs were of Kipchak origin and later broke away from them based their arguments on an expression in DLT, alongside records in other sources I will mention below.⁴⁷ Kāšγarī explained the word qaŋlï ﻗَﻨﻜْﻠﻰ as “A wagon for carrying loads” and the name Qaŋlï ﻗَﻨﻜْﻠٖﻰ as a “Name of an important man of Qifčāq.”.⁴⁸ As I stated earlier, he also did not list the Qaŋlïs among the major Turkic peoples. Yorulmaz had two explanations for this, with which I also agree: 1) Kāšγarī actually described parts of the Turkic world located rather closer to Islamic lands, instead of describing all of it. Other Turkic tribes dwelling in the East and located far from the Islamic world are also not mentioned in DLT; 2) mention of Qaŋlï as the name of a Kipchak nobleman in DLT may point to a Qaŋlï grouping living under Kipchak rule.⁴⁹ As we shall see below, HDSL counted the Qaŋlïs as a people separate from the Kipchaks.⁵⁰ Numerous other accounts from Mongol-period sources with similar designations will also be discussed below.

Çandarlıoğlu argued that after the fall of the Uyghur Qaghanate in 840, Tiele tribes such as the Qibi and Bughu (Pugu 僕骨/僕固) lived in regions between the Yin Mountains and Qumul (Yiwu 伊吾, Hami 哈密) as crowded groupings, even in the tenth century.⁵¹ It can be observed that Uyghurs lived in the Gansu region and especially in Northern Ganzhou.⁵² Hence it is also known that some of the Qaŋlïs lived around Gansu, Southeastern Altais, Southwestern Mongolia and Gobi in the late twelfth and early thirteenth centuries. How and when they came here is still unknown; at this point, one can do nothing but speculate. They might have been formed from some Toquz Oghuz-Uyghur groupings following the eleventh century; since S.6551 does not mention the Qaŋlïs there in the tenth century, they were either still a small group at that time or they came there later. However, we do know for certain that they were living in these regions by the early thirteenth century, and if the Black Carts were indeed a part of the Qaŋlïs, then it can be considered that the Qaŋlïs were already in these regions during the ninth century, maybe not in great numbers as they have escaped the attention of S.6551 and other tenth- and eleventh-century sources. Some Qaŋlï noblemen from these regions, whose biographies were recorded in Chinese sources, entered into Mongol service in the thirteenth century. In these records, those regions are described as containing the ancestral graves of Qaŋlï nobles; according to Yorulmaz, this is an indication that those territories were the ancestral lands of Qaŋlïs who had lived there for many generations, a view with which I agree.⁵³ In YS, the name of these Qaŋlïs is given as Kangli 康里⁵⁴ while the name of their country is written as Kangliguo 康里國 (Qaŋlï Country)⁵⁵ and Kangguo 康國 (Qaŋ Country).⁵⁶ One of these Qaŋlï noblemen who served the Mongols was Buhumu 不忽木. In his biography in YS, it is explained that the name Kangli 康里 stands for the Country of the High Carts (Gaocheguo 高車國) of the Han period.⁵⁷ Another form of Qaŋlï in YS, mentioned by Pelliot and Hambis, is Hangjin 杭斤 and (Hang)jin〔杭〕斤 seen in the biographies of the famous Mongol general Sübe'edei Ba'atur (Subutai 速不台, Xuebutai 雪不台). Here, it is noted that Sübe'edei fought with numerous peoples, including Qaŋlïs and Kipchaks (Qincha 欽察); these two are counted as separate peoples.⁵⁸ Similarly, the biography of Ismāʿīl (Hesimaili 曷思麥里, a Turkestani Muslim originally from the Qara Khitan Empire but later serving the Mongols) in YS mentions the famous Mongol general Jebe Noyan (Zhebo哲伯) attacking the Kipchaks 欽察 and Qaŋlïs 康里 separately.⁵⁹ The HDSL gives a list of countries (guo 國) conquered by the Mongols, which also includes Qaŋlï (Kangli 抗里); the source notes that this is the name of a Turkic country (Huihuiguo 回回國).⁶⁰ According to another note in the same work, the Kipchaks (Kebishao 克鼻稍) are also a Turkic country 回回國 and they are a race 種 of the Uyghurs 回紇.⁶¹

According to section 198 of SHM, the Merkid and Naimans fought against the Mongols on the banks of River Irtysh, eventually dispersing after being defeated. Those Merkid that managed to cross the river passed through the land of Qaŋlïs (Qanglin) and Kipchaks (Kimča'ud).⁶² In section 262 of SHM, it is recorded that during his Khwarazmian Campaign, Chinggis Qa'an (Chinggis Qan) dispatched his commander Sübe'dei Ba'atur northwards to advance until the lands of 11 countries and peoples. The Qaŋlïs (Qanglin) and Kipchaks (Kibča'ud) are listed separately among these. Hence according to SHM section 270, Sübe'edei Ba'atur obeyed this command and reached those regions where the Qaŋlïs (Qanglin) and Kipchaks (Kibča'ud) also lived. However, as the commander fell into a difficult situation in these lands, Ögödei Qa'an dispatched the princes Batu, Böri, Güyüg, Möngke and others to help him.⁶³ According to SHM section 274, these princes launched a campaign in the region and subjugated the Qaŋlïs (Qanglin), Kipchaks (Kibča'ud) and Bashkurts (Baǰigid).⁶⁴ John of Plano Carpini listed the Qaŋlïs (Cangiti) among the peoples conquered by the Tatars (Mongols), and referred to them as different from the Cumans (Comani, Kipchaks).⁶⁵ In the same manner, Friar C. de Bridia counted the Qaŋlïs (Cangite) as separate from the Cumans (Comani) in the lands conquered by Mongols.⁶⁶ Mentioning the Mongol conquests again in another part of his work, C. de Bridia wrote that after the Qaŋlïs (Kangitae), the Mongols conquered Cumania (Comania).⁶⁷ According to him, among the places conquered by Batu (Bati) was the land of the Qaŋlïs (terra Kangitarum) and Great Cumania (Magna Comania), which were listed in Batu's conquests as separate regions.⁶⁸ While travelling to the Mongols, the delegation led by John passed through Cumania (Comania) and the Qaŋlïs’ (Kangittæ) land (Terra Kangittarum) located to its east. After this, the delegation reached the land of Bisermini (“Muslims”, banks of Jaxartes). This record is another example of the Qaŋlïs counted separately from the Kipchaks in John's work, in which the lands of the Qaŋlïs and Kipchaks are mentioned as different regions.⁶⁹ On the other hand, William of Rubruck wrote that the Cumans (Comani) called Qaŋlï (Cangle) lived in the steppes north of the Caspian Sea before the Tatar (Mongol) conquests, thus identifying the Qaŋlïs with Kipchaks.⁷⁰ Not long after, William wrote that in the past Qaŋlïs (Cangle) had lived in the lands north of the Caspian which he passed while going to the Mongol capital and that they were related to the Cumans (Comani).⁷¹

During the Mongol period, various legends of the Oghuz began to be written down; these included narratives about Oγuz Qaγan (Oghuz Qaghan), mythological ruler of that Turkic people. This ruler is mentioned in the legend's non-Islamic form written with the Uyghur script as Oγuz Qaγan, and in its Islamic versions as Oγuz Xan (Oghuz Khan). According to the legend, Oghuz Qaghan gave names to Turkic peoples or certain characters representing them, according to particular incidents that had taken place. It can be seen in the legends about Oghuz Qaghan that, just as in the records I cited above, the Qaŋlïs are mentioned as a distinct Turkic people. Stories about the appearance of the ethnonym Qaŋlï are also narrated in these legends. In lines 257–88 of the Uyghur-scribed Turkic manuscript of the Legend of Oghuz Qaghan, the discovery of carts (qaŋa)⁷² and emergence of the ethnonym Qaŋaluγ are narrated, according to Danka's translation:

Then one day the grey-furred, grey-maned, wolf did not walk (further), it stopped. Oghuz Kaghan also stopped. Setting up camp, he stopped. It was an uncultivated, flat land. They (always) called it Jurched (ǰürčäd).⁷³ It was a great country and people. Their livestock were many, their gold and silver were plenty, their jewels were many. Here, the Jurched Kaghan and people came against Oghuz Kaghan. A fight started. With arrows and swords, they fought. Oghuz Kaghan attacked, he crushed the Jurched Kaghan and killed him. He cut off his head. He made the Jurched people heed his words. After the fight, Oghuz Kaghan's army, bodyguards and people obtained so many inanimate goods that an insufficiency of beasts of burden (lit. horse, mule, ox) turned out to load (the goods) on and carry them away. There in Oghuz Kaghan's army, there was an intelligent, good, clever man, his name was Barmaklig Josun Bellig (barmaqlïɣ ǰosun bellig). This clever (man) built a cart (qanq˚). Onto top of the cart (qanq˚), the inanimate goods he put, in front of the cart (qanq˚), the animate goods he put. They trailed them away and left. All the bodyguards, and people saw this and were astonished, and they built more carts (qanq˚). These carts (qanq˚), while moving, were making the noise “qanq˚ qanq˚”. Due to that, they were given the name qanq˚. Oghuz Kaghan saw the carts (qanq˚) and laughed. Then he told (him), “Let the living make the lifeless walk with the carts! You having the cart (qanq˚luɣ), a name is necessary for you, let the cart (qanq˚) manifest it!” he said, then he went away.⁷⁴

The emergence of Qaŋlïs and the ethnonym Qaŋlï (Qanqlī ﻘَﻨَْﻘْﻠﻰٖ ) has been narrated in the Islamic version of the legend transmitted by Rašīd al-Dīn in Persian as follows, according to Thackston's translation:

The Qanqli. At the time Oghuz was fighting his father, uncles, brothers, and cousins and raiding the province, these were among the tribes that joined him. While others were pillaging and carrying their plunder on animals, these used their minds to devise carts to carry their plunder. Since a cart is called qanqli in Turkish, they were therefore given this name. All the Qanqlis spring from them.⁷⁵

Relying mostly on Rašīd al-Dīn's work, the seventeenth-century Khivan ruler Abū al-Γāzī Bahadur Xan (Abu'l-Ghazi Bahadur Khan) also narrated the discovery of carts (qanq) and the origins of Qaŋlïs (Qanqlï) in the legend's Chaghatai version:

Every year Oγuz Xan used to fight against the people dwelling in the Mongol lands and win. Finally he captured them all and those that managed to flee and save themselves took refuge with the χan of the Tatars. At those years the Tatar people dwelled close to Jürjät. Jürjät was the name of a great realm. Its cities and villages are plenty. It is located to the north of China. The Indians and Tajiks call there Čīn Mačīn. Oγuz Xan went there and attacked the Tatars. The χan of Tatars came with many soldiers and fought. Oγuz Xan emerged victorious and decimated his soldiers. So many inanimate booties fell to Oγuz Xan's soldiers that mounts were not sufficient enough to load them. There was a fine and clever man. He thought and built the cart. All those that saw it from him built carts, loaded their booty and went back. They called the cart “qanq”. Before that, neither its name nor it existed. Because it made a sound like “qanq qanq” while moving, they called it qanq. To the one that built it, they called him Qanqlï. All the Qanqlï people are his children.⁷⁶

The people bestowed with names by Oghuz Qaghan in the Uyghur script version of the legend are listed as Slav (Saqlap), Kipchak (Qïpčaq), Qarluq (Qaγar-luq), Khalach (Qalač) and Qaŋlï (Qaŋa-luγ).⁷⁷ Rašīd al-Dīn listed the peoples descended from the people given names by Oghuz Khan as Uyghur (Ūyγūr ﺍُﻮﻴْﻐُﻮﺮ ), Qaŋlï (Qanqlī ﻘَﻨَْﻘْﻠﻰٖ ), Kipchak (Qīpčāq ﻘِﭙْﭽٰﺎﻖ ), Qarluq (Qārlūq ﻘٰﺎﺮْﻠُﻮﻖ ) and Khalach (Qalač ﻘَﻠَﭻ). Yet, according to him, the Aghachäri (Āγāčerī ﺁﻏٰﺎﭽِﺮﻯٖ) people were a newly formed grouping and appeared after the Oghuz came to Iran.⁷⁸ Hence, while talking about Turkic groupings, Rašīd al-Dīn expressed that the Kipchaks, Qaŋlïs, Qarluqs and Khalach are different branches of the Oghuz.⁷⁹ Legend has it that the Uyghurs, Kipchaks, Qaŋlïs, Qarluqs, Khalach and Aghachäris descended from Oghuz Khan's brothers and some of his cousins.⁸⁰ According to Abū al-Γāzī, the ils (peoples) descended from the people and groupings given names by Oghuz Khan were similarly Uyghur, Qaŋlï (Qanqlï), Kipchak (Qïpčaq), Qarluq (Qarlïq) and Khalach (Qalač).⁸¹ All these narratives demonstrate to us that between the thirteenth and fourteenth centuries, the Oghuz saw Uyghurs, Qaŋlïs, Kipchaks, Qarluqs and Khalach as the major Turkic peoples (apart from themselves) and counted the Qaŋlïs as a people distinct from the Kipchaks.

Conclusion

As can be seen, nomadic Turkic tribes carrying the name Qaŋlï started to appear in historical sources in the middle of the eighth century as Qara Qaŋlïγ and Qara Qaŋlï (Black Qaŋlï). The regions in which these tribes were seen were Southern Siberia, the Altai Mountains and Dzungaria. Exactly a century later, tribes bearing a very similar name to that ethnonym (in the form of Black Carts) are observed to have lived dispersedly both in Gansu in the west and the Khingans in the east. THY may have counted the Qaŋlïs as one of the Turko-Mongol tribes to the north of China in the beginning of the tenth century, but this is not certain. As this indicates, some nomadic Turkic tribes living in carts became a grouping distinct from the Uyghurs and other Toquz Oghuz tribes during the eighth to tenth centuries. These earliest possible records about the Qaŋlïs also predate the ethnonym Kipchak, hence casting doubt on the hypothesis that the origins of Qaŋlïs lie within the Kipchaks. These early records about the Qaŋlïs also demonstrate that the Qaŋlïs, emerging from Southern Siberia, probably had nothing to do with the Kangju/Kañha/Kängü/Kängäräs located on the banks of Jaxartes. In the middle of the tenth century, Black Carts are again seen in Southern Siberia. However, during this period a grouping named Black Cart Shirvi are also seen to be living to the east among the Mongolic people named Shirvi, who might be the Black Carts of the Khingans from the ninth century. The ethnonym Black Carts is not seen in the Chinese sources listing Shirvi tribes; these might correspond to the Shirvi's Hejie tribe. Thus, this draws to our attention two possibilities: 1) the Black Carts seen among the Mongolic Shirvi could originally have been Turkic and might later have migrated eastwards, eventually joining the Shirvi; 2) instead of being an ethnic name, Black Carts might be a generic name pointing to the nomadic lifestyle of various tribes that built carts and lived in the felt tents erected above these vehicles. The Black Cart Shirvi can be considered as a part of this people who specialized in building carts and eventually multiplied as tribes, or this name might be an alternative name for the Hejie Shirvi. Unfortunately, we do not possess a list of tribes and clans that formed the Black Carts during the eighth to tenth centuries, both in Southern Siberia, Altais and Dzungaria, and among the Shirvi of the Khingans in the east. Thus, it is not so easy to comment on the origins and ethnic structures of these aforementioned tribes. However, I am of the opinion that at least the Black Carts living in Southern Siberia, Altais and Dzungaria were Turkic, not Mongolic, which is also indicated by the fact that they were carrying a Turkic ethnonym. I am also inclined to reach the conclusion that the grouping mentioned in BPN as Qara Qaŋlïγ and in 81TB10: 06-3a as Qara Qaŋlï, described as living in Southern Siberia, were the same people as the Black Carts in Siberia and Gansu. When it comes to the twelfth to thirteenth centuries, dispersed but numerous Turkic tribes carrying the ethnonym Qaŋlï are seen to be living in a wide area from the Aral Sea up to Inner Mongolia. Some of these dwelt around Gansu, Southeastern Altais, Southwestern Mongolia and Gobi at that time. From among these regions, Gansu especially witnessed various Uyghur and Toquz Oghuz settlements during the ninth to tenth centuries as well. Starting from the thirteenth century, most of the historical sources mentioning the Qaŋlïs considered them as a distinct people separate from the major Turkic peoples of that period such as the Oghuz, Uyghurs, Qarluqs, Kipchaks and Khalach. However, some of these sources also emphasize that these peoples had a common origin.

During the Türk Qaghanate period, the vast majority of dispersed Turkic tribes living from Mongolia up to the Pontic steppes were named Tägräk (Tiele), in other words “Wheel”. If we consider this information, we can postulate that the Qaŋlïs of the eighth to thirteenth centuries should be a part of them. Although we lack concrete data, we could guess that, starting from the middle of the eighth century, some Tägräk tribes in Southern Siberia might have started separating from the Uyghurs and Toquz Oghuz, acquiring the ethnonym Qaŋlï. It is also possible that these Southern Siberian tribes were never part of the Toquz Oghuz union in Mongolia, as we know of the existence of some other non-Toquz Oghuz Tägräk tribes in that region. The name Qaŋlï might also be a generic name depicting a nomadic lifestyle conducted on carts, which is indicated by the Black Carts in the Khingans. It is known that throughout history, on some occasions the nomadic peoples of Eurasia were called by more than one, or an alternative, name. Saka for the Scythians, Turkmen for the Oghuz, Cuman for the Kipchaks and Tatar for the Mongols are examples of this phenomenon. In some cases, these alternative names were used for different peoples in the past, of which both Turkmen and Tatar are examples. It is already well known that the ethnonym Türkmän (Turkmen) was initially the name of a Turkic grouping different from the Oghuz, while the Tatars were a people different from the Mongols before the thirteenth century. It is my opinion that we can think that, just as in the case of Scythian/Saka, Oghuz/Turkmen, Kipchak/Cuman and Mongol/Tatar, the ethnonyms Tägräk, Oghuz and Qaŋlï might be alternative names for each other. Although the name Tägräk cannot be seen in the Turkic inscriptions of the Türk and Uyghur Qaghanates periods, Tiele and its different versions in the older periods are recorded in Chinese sources. Starting from the middle of the seventh century, the usage of Tiele starts to gradually decrease in these sources, and the Tägräks living in Eastern Türk territories in particular begin to be mentioned as the Nine Surnames. The Oghuz and Toquz Oghuz seen in Türk and Uyghur inscriptions are the names of Tägräk tribes dwelling in Mongolia, as mentioned in the Turkic inscriptions, while they are equivalents of the Nine Surnames in the Chinese sources. In fact, the usage of the ethnonym Oghuz can be dated further back to older periods. During the time between the collapse of the Hunnic Empire and the foundation of the Avar Qaghanate, Turkic tribes that dominated the Pontic steppes carried the name Oghur, who are thought to be descendants of the Dingling, that is to say, the Tägräks. As can be understood, usage of the ethnonym Tägräk decreased in some periods, and usage of the names Oghuz and Qaŋlï became more widespread. Qaŋlï, a term meaning “Cart”, is a description pointing to the nomadic lifestyle of the Turkic tribes of that period, just as in the case of Tägräk meaning “Wheel”. The semantic connection between “Cart” and “Wheel” is already obvious. Hence, when one looks at this from this respect, it is no coincidence that during the Tabghach period the Chinese called Tägräks “High Carts”. Thus, in my opinion Tägräk, Oghuz and Qaŋlï seem to be the names widely used for the Turkic tribes that generally led a nomadic life in the Eurasian steppes, and these ethnonyms are most probably alternatives for each other. Therefore, I believe it is appropriate to look at the Tägräks for the ultimate origins of the Qaŋlï Turks, rather than at the later Kipchaks or other peoples.

One problem we still face is that we simply do not know how a group of Turkic tribes started distinguishing themselves from the Uyghurs-Toquz Oghuz and other Turkic peoples to become a distinct people in the twelfth to thirteenth centuries. The appearance of the ethnonym Black Qaŋlï/Black Carts in the eighth to ninth centuries indicates that this process might have started in Southern Siberia. The designation “black” also shows that other Qaŋlï groupings probably existed at that time as well, but there are no contemporary sources about them. Since Ḥudūd al-ʿĀlam, S.6551 and Maḥmūd al-Kāšγarī do not mention the Qaŋlïs as a distinct Turkic people, the Qaŋlïs seem to have lived quite far away from sedentary realms before the twelfth century, in a wide region stretching between Southern Siberia, Dzungaria, Inner Mongolia and Gansu. Some of these areas correspond to the lands of Black Qaŋlïs/Black Carts, as we have seen. The lack of sources prevents us from making firm statements, but my hypothesis is that some Tägräk tribes in these regions started an ethnogenesis process during the eighth century and became the Qaŋlï people, who went on to become a major grouping in the early twelfth century.