The Xiongnu lost the steppe of Mongolia in the first and second centuries AD, and the Göktürks took it in 552, when Bumin overthrew the Rouran.[1] Between the two peoples stand the Xianbei and the Rouran.[2] The Chinese histories nevertheless compare the Göktürks with the Xiongnu again and again. The Zhou Shu makes the Ashina, the ruling house of the Göktürks, a branch of the Xiongnu.[3] Some modern writers have taken the likenesses as signs of descent. They have taken the Chinese statement of descent as a reason to place the Xiongnu in the line of Turkic states.[4] Others attribute the likenesses to institutions that the peoples of the steppe held in common. They attribute the statement of descent to the way the Chinese histories write of the north.[5]

The likenesses

The comparison begins with the Chinese compilers. The Zhou Shu says that the title Illig Khan, 伊利可汗, which Bumin took in 552, was like the chanyu of old, 猶古之單于也. It says that the title khatun for his wife was like the yanzhi of old, 猶古之閼氏也.[1] Of the Göktürk way of life, the Zhou Shu says that the Göktürks despised the old, honoured the strong and were without shame or rites, like the Xiongnu of old, 猶古之匈奴也.[1] The Sui Shu closes its account of their customs with the remark that they were broadly those of the Xiongnu, 大抵與匈奴同俗.[2] The Bei Shi, which draws on both, has all three remarks. The Tongdian opens with the statement that Göktürk custom was that of the ancient Xiongnu, and then names what differed.[6]

All of these are Chinese judgements. The Göktürk inscriptions neither compare the Göktürks with the Xiongnu nor name them.[7] The Sui Shu ends its chapter with a survey of the northern peoples in which each age has its own. The Xunyu belong to the Five Emperors, the Xianyun to the Three Dynasties, and the Xiongnu to the two Han. The Wuhuan and the Xianbei come after them, and under the Later Wei and the Zhou come the Rouran and the Tujue.[2] According to the text, the chiefs of these peoples followed one another and took turns as lords, 相繼互為君長.[2] The survey sets the peoples in a sequence of masters of the frontier. It does not make them one stock.

Land and court

The Shiji places the court of the chanyu opposite Dai and Yunzhong, two commanderies of the Chinese frontier. The kings of the left controlled the eastern territories as far as Shanggu, while the kings of the right controlled the west as far as Shangjun. Each ruler had his own pastures and moved in search of water and grazing.[8] The Shiji names no mountain and no river for the court. After the Han campaigns of 119 BC the court lay north of the Gobi, and by 105 BC further to the north-west. The excavated settlements of the empire lie in the Khanui valley north of the Khangai. The princely graves of Noin Ula lie in northern Mongolia.[9][10]

The Zhou Shu has the khan of the Göktürks always on the mountain it writes Dujin, 都斤山. The inscriptions call the Ötüken forest the place from which the realm is held.[11] Modern accounts identify Dujin with Ötüken, the mountain country of the upper Orkhon and the Khangai. The accounts that name Ötüken as the Göktürk seat name the Xiongnu among the powers that kept their seat there before them. This shared location therefore rests on modern interpretation. The Shiji’s own description of the chanyu’s court does not establish it.

Titles and the mandate of Heaven

The Han Shu gives the full style of the Xiongnu ruler as Chengli Gutu Chanyu, 撐犁孤塗單于. It glosses chengli as Heaven, gutu as son, and chanyu as a word for vastness, since the ruler is vast like Heaven, 言其象天單于然也.[12] In the letter of 176 BC Bagatur calls himself the great chanyu of the Xiongnu, set up by Heaven, 天所立匈奴大單于.[8] The court of Laoshang adopted the style of the great chanyu born of Heaven and Earth and ordained by the sun and moon, 天地所生日月所置.[8] The chanyu’s consort was the yanzhi, 閼氏.[12] It cannot be decided whether the claim of a heavenly mandate was borrowed from China or belongs to the common traditions of Inner Asia.[13] Chengli, however, is not the Chinese word for Heaven.[5]

The Göktürk khans took a title of their own. Khan and khagan are not Xiongnu words. The word chanyu went out of use with the Xiongnu. The claim the title makes is the same. The Kül Tegin inscription opens Bilge Khan’s address with teŋri teg teŋride bolmış türk bilge kagan, “Bilge Khan of the Göktürks, like Heaven and born in Heaven”. The Bilge Khan stone opens with teŋri teg teŋri yaratmış, “made by Heaven”.[14] Bilge Khan rules “because Heaven ordained it and because I had Kut”, teŋri yarlıkaduk üçün, özim kutum bar üçün.[15] The Chinese went on calling Göktürk rulers chanyu in some contexts. In 650 the Tang named their protectorate over the surrendered Göktürks the Shanyu protectorate.[16] This is a Chinese usage. It does not show how the Göktürks understood themselves. Taşağıl adds that the name was chosen after the state of the Southern Xiongnu, because the Göktürk princes sat in that country.[17] At the court debate of 630 Wen Yanbo proposed that the surrendered Göktürks be settled as Emperor Jianwu of the Han had settled the submitting Xiongnu on the Wuyuan frontier, with their tribes kept whole.[18] Taşağıl reports that Taizong later gave the vassal state of Simo the structure of the Xiongnu empire, with a Worthy Prince of the left and of the right.[19] A remark that the Jiu Tang Shu gives to Taizong speaks of exterminating the Xiongnu and means the Göktürks. The Tang court used the old name for the people on its northern frontier.[20] Scholars argue from the language of the Xiongnu whether gutu in the Xiongnu style is the same word as Kut. The question belongs to the claims of descent.

Wings and offices

The Shiji divides the offices of the Xiongnu into left and right. The Left and Right Worthy Kings came first, and the heir was regularly the Left Worthy King, tuqi. Below them stood twenty-four chiefs of ten thousand horse. The great ministers held their offices by inheritance.[8] The elders sat on the left at the chanyu’s seat and faced north.[8] The ten highest kings, the Four Horns and the Six Horns, were sons and brothers of the chanyu. They were to become chanyu in their order.[21] This pairing of left and right may have been a territorial order of two halves, not Chinese court placement.[22]

The Göktürk realm had two halves as well. The Sui Shu says that Taspar set Ishbara over the east. He made Börü Khan, a son of his brother, ruler in the west.[2] The Kül Tegin inscription says that Ilterish set the Tölis and the Tardush in order. He gave them a yabgu and a shad.[7] The inscriptions place the Tölis “in front” and the Tardush “behind”, which are the east and the west.[23] A shad commanded a wing and was a prince of the Ashina. Bilge Khan was set over the Tardush at fourteen.[14] The Zhou Shu lists twenty-eight grades of officials, the yabgu first and the shad second. It says that all were hereditary.[1] The nearest parallel to the Xiongnu pairing of offices is the division of the Göktürk realm into an east and a west.[22]

The two orders differ in names and in number. Sons and brothers of the chanyu held the Xiongnu offices of the left and the right. Among the great ministers, three clans held them: the Huyan, the Lan and the Xubu.[8] Ashina princes held the Göktürk wings, and the grades of rank ran to twenty-eight.[1]

Rites of the year

The Shiji gives the Xiongnu three gatherings a year. In the first month the chiefs met in a small assembly at the court of the chanyu and made offering. In the fifth month they met in the great assembly at Longcheng. There they sacrificed to their ancestors, to Heaven and Earth and to the ghosts and spirits. In autumn, when the horses were fat, they met at Dailin and counted people and animals.[8] The Hou Han Shu adds that the Southern Xiongnu sacrificed to the spirit of Heaven on the wu day of the first, the fifth and the ninth month.[21] The chanyu came out of his camp each morning and bowed to the rising sun. Each evening he bowed to the moon.[8]

The Zhou Shu names two occasions for the Göktürks. Every year the khan led the nobles to the cave of the forebears, 先窟. In the middle of the fifth month they gathered at the Taren water, 他人水, and sacrificed to the spirit of Heaven.[1] The Sui Shu says that in the fifth month they killed many sheep and horses to sacrifice to Heaven. Of the western Göktürks it says that on the eighth day of the fifth month they gathered to sacrifice to the god. Every year they sent a high minister to the cave in which their forebears had lived.[2] The khan’s tent opened to the east, out of reverence for the place where the sun rises.[1]

Both calendars hold a sacrifice in the fifth month, and both join Heaven with the forebears. The first-month gathering and the autumn count of the Xiongnu have no Göktürk counterpart in the sources.[3] The sources do not say whether the Göktürks took the fifth-month rite over from the Xiongnu or both followed the rhythm of the steppe year.

Marriage and succession

The Shiji gives the Xiongnu a marriage rule. When a father dies, the son marries the stepmother. When brothers die, the survivors take their wives. The Shiji writes this as 父死，妻其後母；兄弟死，皆取其妻妻之.[8] The Han Shu shows it at work. When Fuzhulei succeeded his father Huhanye, he took his father’s consort Wang Zhaojun to wife.[12] The Zhou Shu gives the Göktürks the same rule and widens it. Sons, younger brothers and nephews married the stepmothers, the wives of uncles and the wives of elder brothers, 妻其後母、世叔母及嫂.[1] The Sui princess Yicheng was married to Yami Khan. After him she was married to his sons Shibi, Chula and Ilig in turn.[24]

Succession in the two houses combined descent from father to son with succession from brother to brother. Among the Xiongnu the lineal rule held before the civil strife of the first century BC. After it, succession passed from elder to younger brother.[13] Among the Göktürks the Töre put brother before son. Where the rule ran out, between the sons of brothers, the standing of a prince’s mother and the backing of powerful men decided, as in the crisis of 581.[2]

Burial

The Shiji says that the Xiongnu buried their dead in an inner and an outer coffin with gold, silver, clothing and furs. They raised no mound, planted no trees and wore no mourning dress. The ruler’s favourite servants and concubines followed him in death, hundreds or thousands by the Shiji’s count and tens or a hundred by the Han Shu’s.[25] Excavated Xiongnu graves hold the bones of sacrificed animals.[26]

The Zhou Shu describes another sequence for the Göktürks. The body lay in the tent, and the sons and kin killed sheep and horses before it. They rode round the tent seven times and cut their faces, and wept blood and tears, seven times over. The mourners burned the body with the horse the dead man had ridden and the things he had used. They kept the ashes until the plants turned yellow if he had died in spring or summer, or came into leaf if he had died in autumn or winter.[1] At the grave they set up a stone for each man the dead man had killed, and hung the heads of the sacrificed sheep and horses on a post.[1] The Sui Shu adds a house built at the grave and painted with the dead man’s likeness and the battles he had fought.[2]

The two descriptions agree that animals were killed for the dead. They differ in the grave. The Xiongnu had no mound or trees, and the Göktürks had stones and a post. In the Altai the burial of a man with his horse under a stone mound is an old tradition. The square stone enclosures that mark the Göktürks appear there in the fifth century as something new. No excavated site fully matches the Chinese description of the Göktürk funeral.[27] The sources record following a lord in death for the Göktürks only in the two men who died with Ilig Khan in 634. Neither Tang history calls it a custom.[18][28]

War

The Shiji says that the Xiongnu watched the stars and the moon before an undertaking. They attacked when the moon was full and withdrew when it waned.[8] The Sui Shu says of the Göktürks that they waited until the moon was almost full and then went out raiding, 候月將滿，輒為寇抄.[2] The Shiji has the Xiongnu child ride a sheep and shoot birds and mice. The Sui Shu calls the Göktürks good at riding and shooting.[29] The list of arms in the Zhou Shu includes the whistling arrow, 鳴鏑. In the Shiji the founder Bagatur uses this arrow in drilling his riders.[30]

The Shiji says of the Xiongnu that each man fought for his own gain. He advanced when there was profit and withdrew when there was none, and he felt no shame in running. The captor kept his plunder, and a cup of wine was given for each head.[8] The Sui Shu says of the Göktürks that they valued death by arms and were ashamed to die of illness, 重兵死而耻病終.[2] At this point the Chinese words for the two peoples are not the same. Each side kept a count. The Xiongnu counted people and animals at the autumn assembly.[8] The Göktürk khan raised troops and levied livestock with notched wood. An arrow with a golden head, sealed in wax, gave the notched wood its authority.[1]

Relations with China

The Shiji sets out the settlement after Baideng. It had four terms: a woman of the imperial house as the chanyu’s consort; fixed yearly deliveries of silk, floss, wine, rice and food; open frontier markets; and an oath that made the two rulers brothers.[8] The heqin was renewed at each change of ruler.[8] The Zhou Shu says that once the Zhou court was at peace with the Göktürks, it gave them a hundred thousand lengths of silk, floss, brocade and coloured silk every year. The Qi court emptied its stores to pay them as well. Taspar said he would lack for nothing so long as his two sons in the south were dutiful.[1]

In both records the silk is also the danger. In the Shiji, a Han defector, Zhonghang Yue, tells the chanyu that Chinese goods were no more than two parts in ten of what the Xiongnu used. A taste for them, he says, would put the Xiongnu in Han hands.[8] The Bilge Khan inscription records the other side of the same arrangement. The Chinese give gold, silver and silk in great quantity without being asked. Their sweet words and soft silks draw peoples near, to die. The khan tells the Göktürks to stay in the Ötüken forest and send their caravans from there.[7] Jagchid and Symons set Zhonghang Yue’s warning to the chanyu beside Tonyukuk’s warning to Bilge Khan. Both say that a people few in number lives by moving over the steppe, and that the cities and customs of China would swallow it.[31] They do not claim descent from the likeness.

Assessing the likenesses

A likeness in title, rite or custom does not prove descent. Peoples of the steppe lived by herds and war in the same country and met the same Chinese neighbour. They could arrive at the same offices and sacrifices without one coming from the other. The Chinese records describe them in phrases older than either. Some of the likenesses recur outside the Xiongnu and the Göktürks. The Zhou Shu gives the marriage rule of the Göktürks to the Tuyuhun, a people of Xianbei descent, and says that their custom was the same.[1] The Wei Shu says of the Gaoche that their language was much like that of the Xiongnu with small differences.[32] If that holds, and if the Gaoche spoke Turkic, then the Xiongnu spoke a language close to Turkic.[33] Pulleyblank counted the Gaoche among the Turkic-speaking peoples. He did not take the Wei Shu statement seriously.[34] Kyzlasov quotes the Wei Shu for the Yueban, the part of the northern Xiongnu that stayed behind. Their customs and language were the same as those of the Gaoche, whom he glosses as Turkic-speaking tribes.[35]

The likenesses are real. For the title and its claim, the two halves of the realm, the fifth-month sacrifice and the levirate, the sources of the two peoples say the same thing in the same terms. They show that the Göktürks stood in a long tradition of steppe rule. They do not say through whom the tradition came.

Writers on both sides of the language question allow a continuity of that kind. Pulleyblank, who held that the Xiongnu language was not Turkic, wrote that the Göktürks were the heirs to the pattern of political dominance on the steppe that had begun with the Xiongnu. He saw it in the word tängri and in some titles.[36] In 1962 he derived the Turkic title tarqan from shanyu. He remarked that the supreme title of the Xiongnu had sunk to a rank of high officer among the Turks.[37] Jean-Paul Roux counts the customs as signs of continuity: watching the moon before a campaign, sacrifices at the turn of the year, the marriage of a widow into the dead man’s family, and burials of servants and wives with the dead.[38] He notes also that the religion of the Xiongnu resembles that of the Scythians as much as the later Turkic and Mongolic religions. The likeness therefore says nothing of language.[39]

Positions on kinship

Kinship of the Göktürks with the Xiongnu has been argued on five grounds: a Chinese statement of descent, a Chinese account of the Ashina’s flight in 439, a wolf legend, the language of the Xiongnu with the title of their ruler, and the genetic evidence.

The statement of the Zhou Shu

The Zhou Shu opens its chapter on the Göktürks with a sentence that makes them a branch of the Xiongnu: 突厥者，蓋匈奴之別種，姓阿史那氏, “the Göktürks were, it seems, a separate branch of the Xiongnu, of the surname Ashina”.[1] The particle 蓋 introduces a presumption, and 別種 names a separate kind or stock. The Zhou Shu was compiled in 636. The sentence names no ground for the statement.[1] The Bei Shi gives the sentence with a home “on the right bank of the Western Sea” added. The Tongdian joins it to the Pingliang sentence of the Sui Shu.[6] Both works rest on the Zhou Shu and the Sui Shu. The Xiongnu clause therefore has one source, repeated by two later works. The Xin Tangshu of the eleventh century gives the sentence in its own words. It says that the Ashina of the Tujue were presumably the northern division of the old Xiongnu, 蓋古匈奴北部也, and that the title khagan was like the chanyu. It adds that the wife was called katun.[18] It uses the same particle and the same comparison as the Zhou Shu. It therefore adds no second witness.

The Chinese histories give the same label to other peoples. The Sui Shu says in the same chapter that the Tiele were descendants of the Xiongnu, 鐵勒之先，匈奴之苗裔也, and the Tongdian calls the Jihu a separate branch of the Xiongnu of the Jin period, 晉時匈奴別種.[40] The Jin Shu counts the Jie a branch of the Xiongnu.[41] A label that fits so many peoples leaves open what it meant in each case: descent, rule by a Xiongnu house, a shared way of life, or the Chinese habit of placing a northern people by the best-known of its predecessors. Leonid Kyzlasov says that some of the fifteen Tiele tribes descended from the Xiongnu, and that all of them spoke Turkic languages although their origins and culture differed from those of the other Turkic-speaking peoples.[42] Luc Kwanten gives the Chinese tradition that the Uyghurs were distant descendants of the Xiongnu and that the Tolos confederation of the Tiele was their immediate ancestor.[43] Denis Sinor regards the label as a convention. In his words, Chinese sources are wont to attribute Xiongnu origin to any people of the vast group of northern or western barbarians. He compares the habit with that of Greek writers, who called any nomad people on the steppe Scythian.[44] Kwanten notes a case in which the label served a purpose. Later pedigrees that traced the Rouran to the Xiongnu were of use to the Northern Wei, who wanted to show that the Rouran were a subject people.[45]

Interpretations of the statement

There are three interpretations. The first takes the statement as one of descent, held loosely. In this interpretation the Ashina lived within the Xiongnu confederation as one small group among many. It admits that the sentence does not fit what the sources say of the Ashina’s arrival in Gansu from the west. It allows that the Ashina may have moved back and forth over the centuries.[46] Ahmet Taşağıl takes the Chinese accounts to agree that the Göktürks came from the Xiongnu, either as another line of them or as a people living north of them.[47] Talat Tekin states it more strongly. In his rendering the Chinese sources of the sixth century say that the Göktürks were one of the principal tribes of the old Huns.[48] The Zhou Shu itself says only that they were presumably a separate branch.[1] Luc Kwanten also gives the Chinese tradition that the Turks were a special branch of the Xiongnu, with a state of their own north of the Xiongnu empire. He adds that the factual evidence about the origins of the Turks is very limited.[49]

The second sees the sentence as a Chinese classification of the kind the other peoples received. Such a classification says that the Ashina belonged to the world of the Xiongnu. It does not say that they came from the Xiongnu confederation. It rests on the sequence of the Sui Shu’s survey and on the label’s spread to the Tiele, the Jihu and the Jie. The Chinese account of Xiongnu origins itself conflates several things under one name. These are the peoples of the north who fought the Zhou, the nomad as an ethnographic type, and the people who called themselves Xiongnu.[5] Edwin Pulleyblank put it more bluntly in 1990. He held that the connection with the Xiongnu probably rests on nothing more than the likeness of their role on the Chinese frontier under the Sui and the Tang to that of the Xiongnu under the Han.[36] Sinor adds that the second account of the Zhou Shu, which places the Göktürks in the land of Suo north of the Xiongnu, shows that they were not counted in the Xiongnu confederation.[44]

The third puts the making of the Ashina on the Chinese frontier of Gansu and Eastern Turkestan, among the peoples the Chinese call hu. It leaves the Xiongnu at most as one layer. Peter B. Golden takes the Göktürk union to be a composite of tribes of diverse origins under a ruling clan of possibly, or very probably, alien origin.[50] Owen Lattimore saw the Rouran and the Göktürks as new nomad formations drawn from the margin of China, and not as the Xiongnu continued. In his account the Göktürks were mixed barbarians of the Pingliang region.[51]

Pingliang and the flight of 439

The Sui Shu does not call the Ashina Xiongnu. It says that the ancestors of the Göktürks were mixed hu of Pingliang, 平涼雜胡, of the surname Ashina. When Emperor Taiwu of the Wei destroyed the Juqu, the Ashina fled with five hundred families to the Rouran. They lived for generations on the Golden Mountain, working iron.[2] The Wei Shu dates the fall of the Juqu to 439.[52] The Juqu ruled the Northern Liang in Gansu from 397 and are counted a house of the Xiongnu.[53][54]

The Wei Shu names the Juqu leader Mengxun, who took the titles of governor of Liangzhou and duke of Zhangye, as a hu of the Lushui, 盧水胡沮渠蒙遜. The Hou Han Shu already lists “Qiang and Hu of the Lushui” among the horsemen of the Hexi frontier in its chapter on the Dou family.[55][56] Mixed hu of Pingliang and hu of the Lushui are both frontier peoples. A place names them, not a descent.

Accounts of where the Ashina stood before the flight differ. One reconstruction places them within the Northern Liang state, from which they fled to Gaochang about 439 and on to the southern Altai about 460.[57] A second connects Pingliang with the Helian, the Xiongnu house that founded the Xia.[53] A third understands the wolf legend as the memory of a Wei campaign that laid waste a village of Turkic semi-nomads at Pingliang.[58] In the first two reconstructions the Ashina stood among Xiongnu houses of the fifth century without being Xiongnu by descent. That is one of the senses the Zhou Shu’s sentence could bear. Ahmet Taşağıl identifies the “various hu of Pingliang” as Xiongnu.[59] He calls the link between the Ashina and the Juqu leader Mengxun only a rumour. The histories he names do not mention it, and he gives the southern slope of the Altai as the first certain home of the Göktürks.[60] He regards the name Juqu as the Chinese writing of a Turkic word, a title given to the great among the Xiongnu nobles.[59] He holds that the Zhao, the Xia and the Northern Liang carried on a Turkic culture in north China.[61] Luc Kwanten holds that the dynasties of supposed Xiongnu origin were almost completely sinicized and no longer belong to the history of the steppe.[62] The flight and Pingliang have their own articles.

The wolf legend and the land of Suo

The Zhou Shu tells the wolf legend after its first sentence. A neighbouring state destroys the clan. One boy of about ten survives. A she-wolf feeds him and bears his sons in a cave north of Gaochang, and one of the ten sons is Ashina.[1] The Xiongnu stand in the first sentence only, and the story names no Xiongnu.[1] A second account in the same chapter derives the Göktürks from the state of Suo, 索國, north of the Xiongnu. The chapter closes the two accounts by saying that, however they differ, the Göktürks are in the end of wolf stock, 然終狼種也.[1] The sentence is the compiler’s.

The wolf is told of other peoples whom the sources set beside the Xiongnu. The Shiji has Zhang Qian report that the Xiongnu killed the father of the Wusun king Kunmo. The child was left in the wilds, where a raven brought him meat and a wolf suckled him. The chanyu, thinking it a marvel, took him in and raised him.[63] The Wei Shu says of the Gaoche that a chanyu of the Xiongnu set his two daughters on a high platform to give them to Heaven. An old wolf came, and the younger daughter went down and became the wolf’s wife and the mother of the Gaoche.[32] In the Gaoche story the Xiongnu are the parents of the wolf’s wife. In the Göktürk story no Xiongnu parent appears.

The shared motif has been used in three ways. One study derives the Ashina legend from the Wusun, the neighbours of the Ashina in Gansu. In that case the legend belongs to the Gansu frontier and not to the Xiongnu country.[64] Another sees in the likeness of the motifs a sign of an Iranian component in the making of the Göktürks.[65] The commentators who annotate the Suo account make Suo the old homeland of the Xianbei and conclude that the Göktürks descend from them. In the same passage they warn that such genealogies are not to be believed at once.[66] Lev Gumilev regarded Xianbei, a Mongolic language, as the language of the Göktürk union at its start. He derived the name Ashina from the Mongolic word for wolf.[67] Three descents of the Göktürks thus stand in one chapter and its commentary: from the Xiongnu, from Suo and from the Xianbei. Ahmet Taşağıl puts Suo at the upper Yenisei. He places it neither with the Tabgach nor with the Xianbei, and so leaves the second line of descent outside the Xiongnu country.[68]

The Turkish literature draws the wolf story into the argument for descent. Bahaeddin Ögel understands the wolf legends of the Wusun, the Gaoche and the Göktürks as one line that begins inside the Xiongnu realm.[69] He holds that the wolf belongs to the Turks and not to the Mongols. In his view the Turkic family is built on the father’s line and the she-wolf is the great mother.[70] Jean-Paul Roux calls it the greatest tale of the Turko-Mongol world, with the Xiongnu “in the background”. He holds that the relief of Bugut shows that the Göktürks did not take it up only after they founded their empire.[71]

The Göktürk state itself used the wolf. The Zhou Shu reports a golden wolf’s head on the standards and a guard called fuli, 附離, which it glosses as wolf. The Bugut stone of about 584 bears a relief of a wolf with a man beneath her.[1][72] None of this ties the legend to the Xiongnu. The wolf has its own article.

Eastern Turkestan and the name Ashina

Sergei G. Klyashtorny regards the Zhou Shu legend as a Turkic one. In it the she-wolf’s sons mix with the local population of Eastern Turkestan, whom the Chinese call hu and who spoke Iranian and Tocharian languages.[73] In his account the Ashina were moved in the fifth century to the foothills of the Mongolian Altai. There they headed the tribal union that took the name Türk.[73] He derives the name Ashina from the Khotan-Saka word for “blue”, ässeina, and finds the same meaning in the kök of kök türk.[74] The Old Turkic inscriptions never write the name Ashina.[73]

Golden asks whether the Ashina acquired their way of life and their metallurgical skill through an association with Xiongnu statelets. He does not answer.[75] Both positions put the formative place of the Ashina in Gansu and Eastern Turkestan, and neither makes the Xiongnu confederation their origin. The names of the founders and their descendants, Bumin, Istemi, Mukan and Taspar, are not obviously Turkic. Several are plausibly Iranian.[76]

The language of the Xiongnu

The Xiongnu left no text. Scholars argue their language from a couplet of four words in the Jin Shu and some fifty glossed terms. Four answers have been given. The prevailing one makes the Xiongnu, or their ruling stratum, speakers of Turkic. Its modern advocates include René Grousset, who held the chanyu to be a Turkic title and the khagan a Mongolian one; Annemarie von Gabain, who argued that Turkic and Mongol history should be understood as a continuation of the Xiongnu; Franz Altheim; Jean-Paul Roux, on linguistic and cultural grounds; and Ahmet Taşağıl, who calls the Xiongnu the first Turkic state in history.[77] Alexander Savelyev and Choongwon Jeong find the greater part of the Xiongnu population likely to have spoken Late Proto-Turkic, with an Eastern Iranian group assimilated in time.[78] Anna Dybo found in the fifty-six most transparent glosses an Eastern Iranian stratum, early and almost confined to titles and dairy words, and a Turkic stratum that runs through all periods and fields of meaning.[79]

Other writers hold the Turkic view in less settled forms. Luc Kwanten regards the language of the Xiongnu as Altaic, and the choice between proto-Turkic and proto-Mongol as undecided.[80] Denis Sinor holds that the Göktürks are the first great steppe empire whose language is known with certainty. For the Xiongnu and the Rouran, in his words, any attribution of a language remains highly speculative.[81] Leonid Kyzlasov calls the language of the Xiongnu unknown. He regards the Rouran language as one of the same family, since the ancient sources call the Rouran a separate branch of the Xiongnu.[82] Pulleyblank held in 1990 that the Xiongnu words in Chinese transcription cannot be interpreted as Turkic or as any other Altaic language.[36]

The second answer makes the language Yeniseian, an idea that goes back to Lajos Ligeti and Edwin Pulleyblank.[83] Pulleyblank judged it very unlikely that Xiongnu was Altaic.[84] The newest form of this answer, by Svenja Bonmann and Simon Fries, makes the language an early form of Arin and extends it to the Huns of Europe.[85] A third answer, argued by Harold Bailey, makes the language Eastern Iranian. A fourth, with Gerhard Doerfer, holds that the material permits no classification.[83] The language of the Xiongnu has its own article.

Language bears on the Göktürks in one direction only. If the ruling stratum of the Xiongnu spoke Turkic, the Xiongnu stand in the line of Turkic states and the likenesses of title and rite have a linguistic ground.[4] If it spoke Yeniseian or an Iranian language, the likenesses need another explanation. In either case the language of the Xiongnu does not by itself say where the Ashina came from.

The title Chengli Gutu Chanyu

One argument for a Turkic Xiongnu is a derivation of the ruler’s title. It identifies chengli with Tengri, gutu with Kut and chanyu with an original tanyu, on the analogy of the first shift. It thus reconstructs Teŋri Kut Tanyu, “the ruler endowed by Heaven with fortune”.[86] It sets this beside the Göktürk teŋri teg teŋride bolmış. It notes that Kut and tanyu do not appear among the Mongols. It concludes that the Xiongnu ruling elite spoke Turkic and had a Turkic identity.[87] It adds that the Chinese word chanyu fell out of use with the Xiongnu and returned for the Göktürks. It adds that a commander of Bagatur’s western wing bore the title Turki.[88] The same argument rejects the Han Shu’s gloss of gutu as “son” as a Chinese rendering of a steppe word.[89]

Other explanations of the same words stand beside it. Scholars have explained the glossed words ku-t’u and ch’eng-li as Kettic, as Iranian and as Turkic. Ch’eng-li has been regarded as a probable borrowing from Turkic. By the same reasoning, the majority of the Xiongnu words cannot be explained from Turkic at all.[90] Golden calls the linguistic data at best ambiguous. He grants that the Altaic peoples come to the fore in the Xiongnu era and are by its end the predominant power of Inner Asia.[90] He finds the name of the royal clan, Luandi, very un-Altaic.[84] The data record the derivation as an inference and not as a finding. Jean-Paul Roux renders the title as tengri kutu shanyu and finds the idea of Kut in it. He doubts, though, that the Han Shu is right to gloss gutu as “son”.[91] Paul Pelliot found gutu again in the title of a Rouran khan. He regarded the Göktürk titles that end in -fa as a Rouran legacy of Mongolic form.[92]

The names Duman and Bagatur

The Chinese wrote the names of the first two Xiongnu rulers 頭曼 and 冒頓. Some restore them as Duman, “fog”, and Bagatur, “hero”. They regard both as Turkic words.[93] Golden writes the name of Bagatur with a question mark.[94] Others decline to reconstruct either name. They regard 頭曼 and 冒頓 as personal names that cannot be recovered. The Turkish forms Mete and Teoman do not derive from the graphs. Mete descends from the French transcription Mei-Tei of 1756 and its Turkish translation in the 1920s. Neither Mete nor Teoman comes from the sources.[95] The article on Bagatur sets out how each form was reached.

Claims of continuity in institutions

Writers who hold the Xiongnu to be Turks add further likenesses to those the Chinese give. Bahaeddin Ögel holds that the Göktürks took their roots, and even their practice of founding a state, from the Xiongnu, and that the two had the same administration in Turkestan.[96] He holds that the Göktürks saw themselves as the continuation of the great Xiongnu state, as the many Xiongnu titles in their state show.[97] He regards the Xiongnu refusal to hold walled towns, in the counsel of Wei Lü and in Zhizhi’s failed defence, as the rule that Tonyukuk later urged on the Göktürks.[98] He sets the rule of succession among the Xiongnu, the title hatun and the place names Yanran and Huyen beside their appearance in the Göktürk and Uyghur periods.[99] He also sets the posthumous name of the chanyu beside the Turkic custom of not speaking the name of a dead great man and giving him another, called kong.[100] Finally he finds the Han Shu account of the Xiongnu after the frost of 71 BC close to what the Göktürk inscriptions say of the people scattering and perishing.[101]

İbrahim Kafesoğlu holds that the Xiongnu state took its form from the steppe and not from China. He names a central command, an army ordered by tens, wings on the left and the right, and an assembly of the ruling men.[102] He regards the Tuge, the leading Xiongnu tribe of the third and fourth centuries, as the people whose name stands behind that of the Turks.[103] László Rásonyi asks the same of the Tuge and of the Tuoba clan Tuku, and leaves the question open.[104] Kafesoğlu holds it very probable that the Ashina came from the family of Modu. He bases that on the Chinese statement of descent, the wolf legend and the story of the survivor of the Ashina after the fall of the Northern Liang in 439.[105] Faruk Sümer counts the Chinese derivation of the Göktürks, the Uyghurs and the Kirghiz from the Xiongnu among the main proofs.[106] The Chinese texts derive the Göktürks and the Tiele from the Xiongnu. For the Kirghiz the Chinese tradition names another ancestor. Turkic Atlas treats that under Li Ling.

Sümer also sets the left and right wings and the twenty-four commanders of the Xiongnu beside the two branches and twenty-four tribes of the Oghuz. He does not accept that Oghuz Khan was Modu, because the Oghuz epic was written some fifteen hundred years later and knowledge of the twenty-four tribes goes back no further than the eleventh century.[107] Zeki Velidi Togan accepts the identification, which de Guignes and Bichurin had proposed. He names marvellous arrows and horses, a division into four and twenty-four, a son and successor Gün Han, and the killing of the father on the hunt.[108]

The objections come from writers who interpret the same material differently. Anatoly Khazanov calls the view that the Xiongnu order underlies the Oghuz segments arguable, and he holds that the history of the Inner Asian empires was no single line of development. Continuity existed in single periods, he writes, but it was often broken and sometimes ran backwards.[109] He agrees that the political order of the Xiongnu resembles that of later nomad societies, the Mongols above all.[110] Luc Kwanten holds that the Xiongnu state anticipated the one that the Turkic tribes formed centuries later, and that the shanyu was the predecessor of the qaghan. He adds that it cannot be determined, for lack of information, whether the Xiongnu had a lasting influence on the steppe.[111]

Genetics

A genome contains no language, and the genetic studies of the two peoples ask different questions. Three published Xiongnu genomes from Mongolia are genetically diverse. One is largely eastern Eurasian, and two are clearly mixed between western and eastern sources.[112] The authors consider the mixture consistent with a largely Turkic-speaking population with an Iranian-related component. They find the Yeniseian hypothesis unsupported by the genetic evidence.[113] A study of eighteen individuals from two cemeteries on the western frontier of the empire found that the genetic diversity within those communities was as great as that of the empire as a whole.[114]

A study of 2023 sequenced the remains of Empress Ashina, a daughter of Mukan and a granddaughter of Bumin. It found 97.7 per cent north-east Asian and 2.3 per cent west-Eurasian ancestry.[115] The study reports that she was genetically closest to the Tungusic and Mongolic pastoralists of the steppe after the Iron Age, the Rouran, the Xianbei and the Khitan among them. It also reports that she differed from other ancient Göktürks.[115] That finding cuts both ways in the debate on the Xiongnu. Neither side has yet made much of it. The figures of the two studies come from different models and are not measures of kinship against one another.

National narratives

Every state that now claims the steppe has a version of the Xiongnu and the Göktürks, and each is a position. The Turkish History Thesis of the 1930s and the popular historiography that followed count every steppe empire from the Xiongnu onward as a Turkic state.[64] A Turkish national-historical view, set out by İbrahim Kafesoğlu, treats the Xiongnu realm as the first of a single line of Turkic states and as the model of the later ones.[116] The Mongolian narrative presents the Xiongnu, the Xianbei, the Rouran and the Göktürks as the successive states of Mongolia. The Chinese narrative treats the Tujue as one of the northern minorities of a multi-ethnic Chinese history and their descent from the Xiongnu as a fact of that history. The debate on Turkic origins covers these positions, and the debate on the Huns the question of the Xiongnu in Europe.

Limits of the evidence

The sources give three things. The first is a Chinese statement of 636 that the Göktürks were a branch of the Xiongnu. The second is a Chinese account that puts the Ashina in the mixed population of the Gansu frontier when the Wei destroyed the Juqu in 439. The third is a body of likenesses in title, rite, rank and custom.[3] They contain no Göktürk statement of descent. None of the Chinese texts names the Xiongnu house or lineage from which the Ashina came.[117] Between the Xiongnu in Mongolia and Bumin lie four centuries and more and two other powers, the Xianbei and the Rouran. No Xiongnu site west of the Sayan and the Altai, where the Ashina lived under the Rouran, can be dated after the first century AD.

The data record a Turkic-speaking Xiongnu elite as an inference from the title, the rulers’ names and the heartland of the confederation. They record the kinship of the Ashina with that elite as a Chinese statement and a position. The article on the Xiongnu and the ethnogenesis debate treat the same material from the side of the Xiongnu. What remains is a likeness that the sources attest, a descent that one Chinese source asserts, and an interval of four centuries that no source fills.

References

  1. ^ a b c d e f g h i j k l m n o p q r s t Zhou Shu (Book of Zhou), juan 50
  2. ^ a b c d e f g h i j k Sui Shu (Book of Sui), juan 84
  3. ^ a b c Zhou Shu (Book of Zhou), juan 50; Sui Shu (Book of Sui), juan 84
  4. ^ a b Yavuz 2025, p. 53
  5. ^ a b c Cosmo 2011
  6. ^ a b Bei Shi (History of the Northern Dynasties), juan 99; Tongdian, juan 197
  7. ^ a b c Kül Tegin inscription
  8. ^ a b c d e f g h i j k l m n o Shiji, juan 110
  9. ^ Houle 2011
  10. ^ Werner 1956, p. 47,51
  11. ^ Zhou Shu (Book of Zhou), juan 50; Kül Tegin inscription
  12. ^ a b c Han Shu, juan 94
  13. ^ a b Golden 1992, p. 65
  14. ^ a b Kül Tegin inscription; Bilge Khan inscription
  15. ^ Kül Tegin inscription; Tekin 1968, p. 348, 392–393
  16. ^ Jiu Tangshu (Old Book of Tang), juan 194
  17. ^ Taşağıl 1999, p. 2, 59
  18. ^ a b c Xin Tangshu (New Book of Tang), juan 215
  19. ^ Taşağıl 1999, p. 29
  20. ^ Jagchid 1989, p. 70
  21. ^ a b Hou Han Shu, juan 89
  22. ^ a b Cosmo 2011, p. 46–47
  23. ^ Thomsen 1896, p. 147
  24. ^ Sui Shu (Book of Sui), juan 84; Jiu Tangshu (Old Book of Tang), juan 194
  25. ^ Shiji, juan 110; Han Shu, juan 94
  26. ^ Erkoç 2026, p. 241
  27. ^ Konstantinov 2018
  28. ^ Jiu Tangshu (Old Book of Tang), juan 194; Xin Tangshu (New Book of Tang), juan 215
  29. ^ Shiji, juan 110; Sui Shu (Book of Sui), juan 84
  30. ^ Zhou Shu (Book of Zhou), juan 50; Shiji, juan 110
  31. ^ Jagchid 1989, p. 25
  32. ^ a b Wei Shu, juan 103
  33. ^ Wei Shu, juan 103; Pulleyblank 1990, p. 25–26
  34. ^ Pulleyblank 1990, p. 23, 25–26
  35. ^ Litvinsky 1996, p. 315
  36. ^ a b c Pulleyblank 1990, p. 21
  37. ^ Pulleyblank 1962, p. 256–257
  38. ^ Roux 1994, p. 108, 200, 228–229
  39. ^ Roux 1994, p. 13
  40. ^ Sui Shu (Book of Sui), juan 84; Tongdian, juan 197
  41. ^ Savelyev 2020, p. 5
  42. ^ Litvinsky 1996, p. 319–320
  43. ^ Kwanten 1979, p. 50
  44. ^ a b Litvinsky 1996, p. 323
  45. ^ Kwanten 1979, p. 18
  46. ^ Yavuz 2025, p. 55
  47. ^ Taşağıl 1995, p. 9
  48. ^ Tekin o. J., p. 9
  49. ^ Kwanten 1979, p. 29–32
  50. ^ Golden 1992, p. 121, 126
  51. ^ Lattimore 1951, p. 526, 528
  52. ^ Wei Shu, juan 4
  53. ^ a b Golden 1992, p. 120
  54. ^ 内田吟風 1935, p. 575
  55. ^ Wei Shu, juan 2
  56. ^ Hou Han Shu, juan 23
  57. ^ Golden 1992, p. 121
  58. ^ Yavuz 2025, p. 125
  59. ^ a b Taşağıl 1995, p. 12
  60. ^ Taşağıl 1995, p. 12–13
  61. ^ Taşağıl 1995, p. 13
  62. ^ Kwanten 1979, p. 14
  63. ^ Shiji, juan 123
  64. ^ a b Yavuz 2025
  65. ^ Golden 1992, p. 119
  66. ^ Mau-tsai 1958, Turkish edition
  67. ^ Golden 1992, p. 126–127
  68. ^ Taşağıl 1995, p. 10
  69. ^ Shiji, juan 123; Wei Shu, juan 103; Zhou Shu (Book of Zhou), juan 50; Ögel 1993, p. 13–20
  70. ^ Ögel 1995, p. 111–115, 536
  71. ^ Roux 1994, p. 157–158
  72. ^ Golden 1992, p. 151
  73. ^ a b c Klyashtorny 1994, p. 445
  74. ^ Klyashtorny 1994, p. 446–447
  75. ^ Golden 1992, p. 123
  76. ^ de la Vaissière 2005a, p. 206
  77. ^ Yavuz 2025, p. 52, 54
  78. ^ Savelyev 2020, p. 12
  79. ^ Dybo 2007, p. 114–115
  80. ^ Kwanten 1979, p. 12
  81. ^ Litvinsky 1996, p. 322
  82. ^ Litvinsky 1996, p. 317
  83. ^ a b Savelyev 2020, p. 4
  84. ^ a b Golden 1992, p. 58
  85. ^ Bonmann 2025, p. 29–30
  86. ^ Yavuz 2025, p. 49, 51
  87. ^ Yavuz 2025, p. 51–52
  88. ^ Yavuz 2025, p. 52
  89. ^ Yavuz 2025, p. 50
  90. ^ a b Golden 1992, p. 59
  91. ^ Roux 1994, p. 91, 133–134
  92. ^ Pelliot 1929, p. 228
  93. ^ Yavuz 2025, p. 52, 74
  94. ^ Golden 1992, p. 50
  95. ^ Yavuz 2025, p. 73–74
  96. ^ Ögel 2015, p. 216, 380
  97. ^ Ögel 1993, p. 288
  98. ^ Ögel 2015, p. 114, 160–161, 163
  99. ^ Ögel 2015, p. 103, 105, 106, 372
  100. ^ Ögel 2015, p. 184, 209, 219
  101. ^ Ögel 2015, p. 122
  102. ^ Kafesoğlu 1997, p. 62–63
  103. ^ Kafesoğlu 1997, p. 60, 69
  104. ^ Rásonyi 1993, p. 20
  105. ^ Kafesoğlu 1997, p. 97–99
  106. ^ Sümer 1972, p. 2
  107. ^ Sümer 1972, p. 2–3
  108. ^ Togan 1981, p. 40, 419
  109. ^ Khazanov 1994, p. 179, 257
  110. ^ Khazanov 1994, p. 178
  111. ^ Kwanten 1979, p. 14–15
  112. ^ Savelyev 2020, p. 6
  113. ^ Savelyev 2020, p. 7, 12
  114. ^ al. 2023
  115. ^ a b al. 2023, abstract
  116. ^ Kafesoğlu 1997, p. 61, 63, 253–257
  117. ^ Kül Tegin inscription; Zhou Shu (Book of Zhou), juan 50; Sui Shu (Book of Sui), juan 84

Sources

Literature

  • Emre-E. Yavuz, The Gokturk Empire: 2nd Edition (2025)
  • Peter B. Golden, An Introduction to the History of the Turkic Peoples (1992)
  • Alexander Savelyev; Choongwon Jeong, Early Nomads of the Eastern Steppe and Their Tentative Connections in the West (2020)
  • Anna V. Dybo (А. В. Дыбо), Лингвистические контакты ранних тюрков. Лексический фонд: пратюркский период (2007)
  • Nicola Di Cosmo, Ethnogenesis, Coevolution and Political Morphology of the Earliest Steppe Empire: The Xiongnu Question Revisited (2011)
  • S. G. Klyashtorny, The Royal Clan of the Turks and the Problem of Early Turkic-Iranian Contacts (1994)
  • Yang et al., Ancient genome of Empress Ashina (2023)
  • Juhyeon Lee et al., Genetic Population Structure of the Xiongnu Empire at Imperial and Local Scales (2023)
  • Svenja Bonmann; Simon Fries, Linguistic Evidence Suggests that Xiong-nu and Huns Spoke the Same Paleo-Siberian Language (2025)
  • Étienne de la Vaissière, Sogdian Traders: A History (2005)
  • Vilhelm Thomsen, Inscriptions de l'Orkhon déchiffrées (1896)
  • Talât Tekin, A Grammar of Orkhon Turkic (1968)
  • Liu Mau-tsai, Çin Kaynaklarına Göre Doğu Türkleri (1958, Turkish edition)
  • Jean-Luc Houle; Lee G. Broderick, Settlement Patterns and Domestic Economy of the Xiongnu in Khanui Valley, Mongolia (2011)
  • Joachim Werner, Beiträge zur Archäologie des Attila-Reiches (1956)
  • Hayrettin İhsan Erkoç, Asya Hunlarında Dinî İnançlar (Religious Beliefs of the Xiongnu) (2026)
  • Nikita Konstantinov; Vasilii Soenov; Synaru Trifanova; Svetlana Svyatko, History and culture of the early Türkic period. A review of archaeological monuments in the Russian Altai from the 4th–6th century AD (2018)
  • Uchida Ginpū 内田吟風, 五胡亂及び北魏に於ける匈奴 (1935)
  • İbrahim Kafesoğlu, Türk Millî Kültürü (1997)
  • László Rásonyi, Tarihte Türklük (1993)
  • Jean-Paul Roux, Türklerin ve Moğolların Eski Dini (1994)
  • Bahaeddin Ögel, Büyük Hun İmparatorluğu Tarihi II (2015)
  • Bahaeddin Ögel, Türk Mitolojisi (Kaynakları ve açıklamaları ile destanlar) I (1993)
  • Bahaeddin Ögel, Türk Mitolojisi (Kaynakları ve açıklamaları ile destanlar) II (1995)
  • Zeki Velidi Togan, Umumi Türk Tarihine Giriş I (1981)
  • Faruk Sümer, Oğuzlar (Türkmenler). Tarihleri, boy teşkilâtı, destanları (1972)
  • Luc Kwanten, Imperial Nomads. A History of Central Asia, 500–1500 (1979)
  • B. A. Litvinsky; Zhang Guang-da; R. Shabani Samghabadi (ed.), History of Civilizations of Central Asia, Vol. III (1996)
  • Edwin G. Pulleyblank, The 'High Carts': A Turkish-Speaking People Before the Türks (1990)
  • Edwin G. Pulleyblank, The Consonantal System of Old Chinese, Part II (1962)
  • Ahmet Taşağıl, Gök-Türkler, Bd. I (1995)
  • Ahmet Taşağıl, Gök-Türkler, Bd. II (1999)
  • Talât Tekin, Orhon Yazıtları. Kül Tigin, Bilge Kağan, Tunyukuk (o. J.)
  • Owen Lattimore, Inner Asian Frontiers of China (1951)
  • Anatoly M. Khazanov, Nomads and the Outside World (1994)
  • Sechin Jagchid; Van Jay Symons, Peace, War, and Trade Along the Great Wall. Nomadic-Chinese Interaction through Two Millennia (1989)
  • Paul Pelliot, Neuf notes sur des questions d'Asie Centrale (1929)

Related

What links here7

States
Xiongnu
Places
Noin Ula
Peoples
History of the Turkic peoples · Tiele
Customs & Pastimes
Culture of the Turkic Peoples
Controversies
Archaeogenetics of the Steppe Empires · The Turkic ethnogenesis debate

Cite this article

Yavuz, Emre-E. “The Göktürks and the Xiongnu.” Turkic Atlas, revised 8 November 2026. https://turkicatlas.org/historiography/gokturks-and-xiongnu/

Last revised: 8 November 2026