The Xiongnu built the first empire of the eastern steppe. How they did it is the foundational question of the field, because everything that came after them inherited the shape of what they made. The Göktürk khaganate is included in this. The answers on offer divide on a single point. One family of explanations locates the cause on the Chinese frontier, so that the steppe empire is a function of the agrarian one. The other locates the cause in the steppe, and holds that the frontier explains the timing at most. Turkic Atlas sets out both, and then shows what the evidence does to them.

Frontier models

Owen Lattimore’s account assumed a sharp divide between the pastoralists of the north and the farmers of China. The divide was ecological, economic and political. Specialisation on both sides left the nomads dependent on goods they could not produce. From that dependency follows the pattern usually called trade or raid: the forceful appropriation of what the steppe economy lacked.[1]

Lattimore’s own pages put the case somewhat differently. In them China is the active partner. The frontier along the Great Wall took its shape from Chinese growth, which came before the great nomadism of the steppe, and not from nomad pressure. The wall served as much to keep the Chinese in as to keep the nomads out.[2] The steppe, in the same book, is a society that produces its own clothing, food and shelter. Its mobility and its economic independence make it strong in war, and the danger lies in the mixed forms of trade and tribute at its margin.[3] A further view drawn from the book is that the early wars of the Han on the northern frontier were wars against the frontier commanders, who went over to the Xiongnu when pressed, rather than against the Xiongnu themselves.[4] Another view lets the steppe empire grow out of the margins. Groups of semi-nomads from the edges of China, the Altai forests and the oases moved into the open steppe. There, once they had left the compromises of border life behind, they took the form that appears with Duman and Bagatur.[5]

Thomas Barfield built the best-known version on that foundation. In it the steppe polities of Mongolia are shadow empires: secondary formations that arise in answer to Chinese imperial expansion. Their stability depends on extorting wealth from China through pillage, tribute, border trade and the re-export of luxuries. It does not depend on taxing the steppe. When China is centralised and powerful, the nomadic empire is too. When China collapses into anarchy, the unified steppe polity collapses with it.[1] The nomadic world unites, on this view, when it needs enough force to extract goods from a neighbour that has grown both richer and stronger.

Two objections have been pressed. Michael Drompp, reviewing the formation and disintegration of the first and second Göktürk khaganates and the Uyghur empire, found the picture more complex than the model allows and the system of diplomacy and trade multipolar rather than bipolar.[1] Peter Turchin objected that there is no time lag between the unification of China and the rise of a steppe empire, and that steppe states came into existence in several cases when China was weak and divided.[1]

Turchin’s own model keeps the frontier but lengthens the clock. Over several centuries, he argues, polities on both sides of the frontier grew larger and more complex together. As larger Chinese states swallowed the smaller ones, the nomads scaled up in step. On this model the Xiongnu empire is the product of that coevolution, not a reaction to a finished Chinese empire.[1] The weakness of the model is one of evidence. What happened on the Chinese side is documented. The parallel development in the steppe is a postulate, and Nicola Di Cosmo holds it impossible to demonstrate on either textual or archaeological grounds.[1]

Anatoly Khazanov added two points. On his account the unification of China brought about the unification of the steppe nomads under the Xiongnu, though with some delay, and the first walls against the north date from the fourth and third centuries BC.[6] He also sorts nomadic states by what the conquerors made of the conquered. A state of the first type took tribute and left the settled lands as they were. A state of the second type joined nomads and farmers in a single order, and one of the third type rested on the division of labour between herdsmen and peasants. On this reckoning the Xiongnu and the Göktürks belonged to the first type.[7]

Origin of the institutions

The decisive test lies not in the timing but in the apparatus. The Xiongnu had an army organised on decimal lines, a government of paired offices, imperial rituals, a centralised conduct of trade and diplomacy, and a language of heaven-given sovereignty. An older interpretation, that of Kurakichi Shiratori in 1930, derives the title chanyu, the sacrifice to Heaven and the use of bronze images among the Xiongnu from Qin and Han China.[8] If that apparatus is Chinese in origin, the frontier models have their mechanism. If it is not, they lose it. Scholars have offered three elements as evidence of Chinese derivation. Di Cosmo takes each of them and argues that it fails.

The first is the title. The Han Shu gives the ruler’s style as Chengli Gutu Chanyu and glosses it son of Heaven, which invites comparison with the Chinese Tianzi.[9] Chengli, however, is tengri. Scholars reconstructed it as such with little dissent, long before this argument was made. A borrowing from China could not have produced a native word different from the Chinese one. If the non-Chinese word was already in place, there is no reason to think the concept attached to it was not.[1] Chanyu has been compared with a later Altaic dārγā, which belongs to a body of political terms outside the Chinese tradition altogether.[1]

The second is the division of every office into a left and a right. Chinese titles were paired that way from the Zhou down to the Han, with the left taking precedence. The arrangement has been derived from the ruler’s facing south, so that his left hand lay east, toward the rising sun. The Shiji, however, describes the Xiongnu ritual otherwise. When the Chanyu and his men sit down, the senior officials sit on the left and face north. All of them do so, not only those styled of the left.[10] The Xiongnu pairing is therefore not ceremonial placement at all. It is a territorial order, with two branches of government administering two halves of the empire. Its closest parallel is not in China. It is the division of the Göktürk realm into an eastern and a western half.[1]

The third is colour symbolism. The Shiji says that when the Han emperor Gaozu was surrounded, the Xiongnu cavalry was white on its western side, grey-spotted on the east, black on the north and bay on the south. This matches the Chinese system of the Five Elements.[10] The association is certainly there. Whether the Xiongnu took it from China is another matter. Colour-and-direction correlations belong to a shamanic substratum that early Chinese civilisation shared with the rest of inner Asia, and they may reach back to the Bronze Age or earlier.[1]

The rest of the apparatus points to the west. The decimal organisation of the army has its closest and perhaps its only parallel in the Achaemenid empire. The oath sworn over wine mixed with the blood of a sacrificed animal and stirred with a jinglu blade is a Scythian rite. So is the cup of wine given for an enemy’s head taken in battle. Scholars have reconstructed the name of the sacrificial dagger itself as an Iranian word.[1] A comparison of the name with Teleut qiŋiraq, “double-bladed knife”, has been called striking in sense and sound, but the Turkic word is itself joined to Iranian forms in Wakhi, Sogdian and Persian, and the relation between the Turkic and the Iranian forms remains to be elucidated. The Xiongnu form seems likely to have been closer to the Turkic than to the Iranian forms.[11] This agrees with what the vocabulary shows: an Eastern Iranian stratum in the Xiongnu titles and dairy terms, which Turkic Atlas sets out under the language debate.[12]

Di Cosmo concludes that the roots of Xiongnu political culture are not in China. A Chinese origin has still to be demonstrated, and with a Chinese influence excluded the coevolutionary hypothesis cannot be sustained. The evidence points instead to a continuum between the nomadic world north of China and the Scythian cultures of Siberia and central Asia, in forms of political organisation as well as in material culture.[1]

The question of state or chiefdom

A second argument concerns what to call the Xiongnu polity. Nikolay Kradin sets four criteria for an early state: access to managerial position by some selection outside clan and kin; regular taxation to pay officials; a judicial power separate from the political one; and a class of functionaries running an administration for the whole community. By that measure the Xiongnu polity meets the requirements at an embryonic level at best, and Kradin calls it a super-complex chiefdom.[1]

The case on the other side is that the military and civil apparatus stood apart from the kin hierarchies. A centre maintained its commanders and functionaries, and a charismatic ruler headed the centre and conducted the rites of the whole community. The judicial power, on this view, was in fact separate.[1] An empire of that extent must have held a great many middle-ranking posts: commanders of smaller units, local administrators, and superintendents of the post, of supplies and of the court. One cannot simply assume that the three or four royal clans named in the Shiji filled all of them.[1] The argument about lesser kings and regional leaders in the Xiongnu article bears directly on this point.

The two labels are in any case compatible with a third. Whether the polity was an early state or a super-complex chiefdom, it was an empire. It reached far beyond its original territory and its original people, and it held a variety of lands and peoples under an imperial centre.[1]

The archaeological evidence

Both families of model were built when Xiongnu archaeology meant tombs. The survey work of the last decades has changed the evidence under them. A full-coverage survey of 246 square kilometres in the lower Egiin Gol valley recorded 572 sites. Those of the Xiongnu period are the most common.[13] The survey found a settlement hierarchy, long-term occupation and grain production. The Shiji describes this country as a land without walled cities, fixed dwellings or agriculture of any kind.[13] In the Khanui valley the domestic sites yield the four herd animals in the proportions of a working pastoral economy. They also yield iron slag at six of the excavated occupation areas. Smelting was therefore not confined to the workshops of fortified places like Ivolga. It went on at open seasonal camps.[14] The people of Egiin Gol were agropastoralists living in settlements. That alone is enough to dispose of any model that rests on the Xiongnu economy being exclusively one of animal husbandry.[1]

The radiocarbon dates carry a further point, and it bears hardest on every model of sudden imperial birth. In Egiin Gol the people using Xiongnu-style pottery lived in the valley for generations before the Xiongnu appear in the Han histories. They went on living there after the Xiongnu had ceased to be worth recording and other names had risen to prominence in the north-east.[13] The name entered and left the Chinese record faster than the population entered or left the valley.

Bearing on the Göktürks

The Göktürks are the test case on which the frontier models were checked and found wanting. They are also the heirs of the arrangement whose origin is in dispute. The left and right of the Xiongnu became the eastern and western halves of the first khaganate. The decimal army persisted. The heaven-given fortune of the ruler, Kut, is the same claim in the same words.[1] If that inheritance came from China, the Göktürk state is a Chinese derivative at one remove. If it came out of the steppe, the state is what it appears to be in the inscriptions: an order that the Göktürks took over from those who held the Ötüken before them.

There are two other interpretations of the Göktürk case. On the first, the khaganate lasted as long as the court at Chang’an could be made to pay. The first khaganate fell when its last khan, short of revenue from China, taxed his own tribes.[15] On the second, the khans of 682 came to power on the unrest of the common people, who had resisted the Chinese ways that the nobles had taken up.[16] A third holds that the titles ending in -t came to the Göktürks from the Rouran, whom it counts as Mongolic, and that the Göktürks learned statecraft from them.[17]

A national-historical interpretation takes the whole matter the other way. It holds that the steppe produced one steppe culture, Turkic at its core and resting on the horse and on iron, whose bearers were the Turks from the Andronovo period. On it the Xiongnu, the Huns, the Tuoba, the Göktürks and the later steppe empires carried the culture on.[18]

References

  1. ^ a b c d e f g h i j k l m n o p q r Cosmo 2011
  2. ^ Lattimore 1951, p. 240,460–461
  3. ^ Lattimore 1951, p. 487–488
  4. ^ Lattimore 1951, p. 478–480
  5. ^ Lattimore 1951, p. 452–454,523–527
  6. ^ Khazanov 1994, p. 96
  7. ^ Khazanov 1994, p. 231–233,254–256
  8. ^ Kurakichi 1930, p. 70–77
  9. ^ Han Shu, juan 94
  10. ^ a b Shiji, juan 110
  11. ^ Pulleyblank 1962, p. 222
  12. ^ Savelyev 2020
  13. ^ a b c Wright 2009
  14. ^ Houle 2011
  15. ^ Khazanov 1994, p. 256
  16. ^ Barthold 2004, p. 14
  17. ^ Barthold 2004, p. 15
  18. ^ Kafesoğlu 1997, p. 213–226

Sources

Literature

  • Nicola Di Cosmo, Ethnogenesis, Coevolution and Political Morphology of the Earliest Steppe Empire: The Xiongnu Question Revisited (2011)
  • Bryan K. Miller, Xiongnu “Kings” and the Political Order of the Steppe Empire (2014)
  • Joshua Wright; William Honeychurch; Chunag Amartuvshin, The Xiongnu Settlements of Egiin Gol, Mongolia (2009)
  • Jean-Luc Houle; Lee G. Broderick, Settlement Patterns and Domestic Economy of the Xiongnu in Khanui Valley, Mongolia (2011)
  • Alexander Savelyev; Choongwon Jeong, Early Nomads of the Eastern Steppe and Their Tentative Connections in the West (2020)
  • Peter B. Golden, An Introduction to the History of the Turkic Peoples (1992)
  • Emre-E. Yavuz, The Gokturk Empire: 2nd Edition (2025)
  • Owen Lattimore, Inner Asian Frontiers of China (1951)
  • Anatoly M. Khazanov, Nomads and the Outside World (1994)
  • Edwin G. Pulleyblank, The Consonantal System of Old Chinese, Part II (1962)
  • Shiratori Kurakichi, On the Territory of the Hsiung-nu Prince Hsiu-t'u Wang and His Metal Statues for Heaven-worship (1930)
  • İbrahim Kafesoğlu, Türk Millî Kültürü (1997)
  • V. V. Barthold, Orta Asya Türk Tarihi Dersleri (Orta Asya Türk Tarihi Hakkında Dersler) (2004)

Related

What links here6

Persons
Duman (Touman)
Peoples
Eurasian Nomads
Lineages
Luandi
Institutions and Titles
Khan (Khagan)
Controversies
The language of the Xiongnu · The Turkic ethnogenesis debate

Cite this article

Yavuz, Emre-E. “Xiongnu: Making the First Steppe Empire.” Turkic Atlas, revised 8 November 2026. https://turkicatlas.org/historiography/steppe-empire-formation-debate/

Last revised: 8 November 2026