Xiongnu

The word "Xiōngnú" means "fierce slave". They were identified by the Han Chinese as invaders from the north who rode on horseback. The pronunciation of 匈奴 as Xiōngnú [ɕjʊ́ŋnǔ] is the modern Mandarin Chinese pronunciation, from the Mandarin dialect spoken now in Beijing, which came into existence only in the second millennium CE. The Old Chinese pronunciation has been reconstructed as *xiuoŋ-na or as *qhoŋna according to Wang Li and Zhengzhang Shangfang respectively, the two most widely recognized Old Chinese reconstructions in China. Sinologist Axel Schuessler (2014) reconstructs the pronunciations of 匈奴 as *hoŋ-nâ in Late Old Chinese (c. 318 BC) and as *hɨoŋ-nɑ in Eastern Han Chinese; citing other Chinese transcriptions wherein the velar nasal medial -ŋ-, after a short vowel, seemingly played the role of a general nasal – sometimes equivalent to n or m –, Schuessler proposes that 匈奴 Xiongnu < *hɨoŋ-nɑ < *hoŋ-nâ might be a Chinese rendition, Han or even pre-Han, of foreign *Hŏna or *Hŭna, which Schuessler compares to Huns and Sanskrit Hūṇā. However, the same medial -ŋ- prompts Christopher P. Atwood (2015) to reconstruct *Xoŋai, which he derives from the Ongi River (Mongolian: Онги гол) in Mongolia and suggests that it was originally a dynastic name rather than an ethnic name.

The territories associated with the Xiongnu in central/east Mongolia were previously inhabited by the Slab Grave Culture (Ancient Northeast Asian origin), which persisted until the 3rd century BC. Genetic research indicates that the Slab Grave people were the primary ancestors of the Xiongnu, and that the Xiongnu formed through substantial and complex mixture with West Eurasians.

During the Western Zhou (1045–771 BC), there were numerous conflicts with nomadic tribes from the north and the northwest, variously known as the Xianyun, Guifang, or various "Rong" tribes, such as the Xirong, Shanrong or Quanrong. These tribes are recorded as harassing Zhou territory, but at the time the Zhou were expanding northwards, encroaching on their traditional lands, especially into the Wei River valley. Archaeologically, the Zhou expanded to the north and the northwest at the expense of the Siwa culture. The Quanrong put an end to the Western Zhou in 771 BC, sacking the Zhou capital of Haojing and killing the last Western Zhou king You. Thereafter the task of dealing with the northern tribes was left to their vassal, the Qin state.

To the west, the Pazyryk culture (6th – 3rd century BC) immediately preceded the formation of the Xiongnus. A Scythian culture, it was identified by excavated artifacts and mummified humans, such as the Siberian Ice Maiden, found in the Siberian permafrost, in the Altai Mountains, Kazakhstan and nearby Mongolia. To the south, the Ordos culture had developed in the Ordos Loop (modern Inner Mongolia, China) during the Bronze and early Iron Age from the 6th to 2nd centuries BC. Of unknown ethno-linguistic origin, it is thought to represent the easternmost extension of Indo-European-speakers. The Yuezhi were displaced by the Xiongnu expansion in the 2nd century BC, and had to migrate to Central and Southern Asia.

Western Han historian Sima Qian composed an early yet detailed exposition on the Xiongnu in one liezhuan (arrayed account) of his Records of the Grand Historian (c. 100 BC), wherein the Xiongnu were alleged to be descendants of a certain Chunwei, who in turn descended from the "lineage of Lord Xia", a.k.a. Yu the Great. Even so, Sima Qian also drew a distinct line between the settled Huaxia people (Han) to the pastoral nomads (Xiongnu), characterizing them as two polar groups in the sense of a civilization versus an uncivilized society: the Hua–Yi distinction. Sima Qian also mentioned Xiongnu's early appearance north of Wild Goose Gate and Dai commanderies before 265 BC, just before the Zhao-Xiongnu War; however, sinologist Edwin Pulleyblank (1994) contends that pre-241-BC references to the Xiongnu are anachronistic substitutions for the Hu people instead. Sometimes the Xiongnu were distinguished from other nomadic peoples; namely, the Hu people; yet on other occasions, Chinese sources often just classified the Xiongnu as a Hu people, which was a blanket term for nomadic people. Even Sima Qian was inconsistent: in the chapter "Hereditary House of Zhao", he considered the Donghu to be the Hu proper, yet elsewhere he considered Xiongnu to be also Hu.

Ancient China often came in contact with the Xianyun and the Xirong nomadic peoples. In later Chinese historiography, some groups of these peoples were believed to be the possible progenitors of the Xiongnu people. These nomadic people often had repeated military confrontations with the Shang and especially the Zhou, who often conquered and enslaved the nomads in an expansion drift. During the Warring States period, the armies from the Qin, Zhao and Yan states were encroaching and conquering various nomadic territories that were inhabited by the Xiongnu and other Hu peoples.

Pulleyblank argued that the Xiongnu were part of a Xirong group called Yiqu, who had lived in Shaanbei and had been influenced by China for centuries, before they were driven out by the Qin dynasty. Qin's campaign against the Xiongnu expanded Qin's territory at the expense of the Xiongnu. After the unification of Qin dynasty, Xiongnu was a threat to the northern border of Qin. They were likely to attack the Qin dynasty when they suffered natural disasters.

The first known Xiongnu leader was Touman, who reigned between 220 and 209 BC. In 215 BC, Chinese Emperor Qin Shi Huang sent General Meng Tian on a military campaign against the Xiongnu. Meng Tian defeated the Xiongnu and expelled them from the Ordos Loop, forcing Touman and the Xiongnu to flee north into the Mongolian Plateau. In 210 BC, Meng Tian died, and in 209 BC, Touman's son Modu became the Xiongnu chanyu.

In order to protect the Xiongnu from the threat of the Qin dynasty, Modu Chanyu united the Xiongnu into a powerful confederation. This transformed the Xiongnu into a more formidable polity, able to form larger armies and exercise improved strategic coordination. The Qin dynasty fell in 207 BC, and was replaced by the Western Han dynasty in 202 BC after a period of internal conflict. This period of Chinese instability was a time of prosperity for the Xiongnu, who adopted many Han agriculture techniques such as slaves for heavy labor and lived in Han-style homes.

After forging internal unity, Modu Chanyu expanded the Xiongnu empire in all directions. To the north he conquered a number of nomadic peoples, including the Dingling of southern Siberia. He crushed the power of the Donghu people of eastern Mongolia and Manchuria as well as the Yuezhi in the Hexi Corridor of Gansu, where his son, Jizhu, made a skull cup out of the Yuezhi king. Modu also retook the original homeland of Xiongnu on the Yellow River, which had previously been taken by the Qin general Meng Tian. Under Modu's leadership, the Xiongnu became powerful enough to threaten the Han dynasty.

In 200 BC, Modu besieged the first Han dynasty emperor Gaozu (Gao-Di) with his 320,000-strong army at Peteng Fortress in Baideng (present-day Datong, Shanxi). After Gaozu (Gao-Di) agreed to all Modu's terms, such as ceding the northern provinces to the Xiongnu and paying annual taxes, he was allowed to leave the siege. Although Gaozu was able to return to his capital Chang'an (present-day Xi'an), Modu occasionally threatened the Han's northern frontier and finally in 198 BC, a peace treaty was settled.

Xiongnu in their expansion drove their western neighbour Yuezhi from the Hexi Corridor in year 176 BC, killing the Yuezhi king and asserting their presence in the Western Regions.

By the time of Modu's death in 174 BC, the Xiongnu were recognized as the most prominent of the nomads bordering the Han empire According to the Book of Han, later quoted in Duan Chengshi's ninth-century Miscellaneous Morsels from Youyang:

Also, according to the Han shu, Wang Wu (王烏) and others were sent as envoys to pay a visit to the Xiongnu. According to the customs of the Xiongnu, if the Han envoys did not remove their tallies of authority, and if they did not allow their faces to be tattooed, they could not gain entrance into the yurts. Wang Wu and his company removed their tallies, submitted to tattoo, and thus gained entry. The Shanyu looked upon them very highly.

The ruler of the Xiongnu was called the chanyu. Under him were the tuqi kings. The Tuqi King of the Left was normally the heir presumptive. Below him in the hierarchy were more officials in pairs of left and right: the guli, the army commanders, the great governors, the danghu and the gudu. Beneath them were the commanders of detachments of one thousand, of one hundred, and of ten men. This nation of nomads, was organized like an army.

After Modu, later leaders formed a dualistic system of political organisation with the left and right branches of the Xiongnu divided on a regional basis. The chanyu or shanyu, a ruler equivalent to the Emperor of China, exercised direct authority over the central territory. Longcheng (around the Khangai Mountains, Otuken) (Chinese: 龍城; Mongolian: Luut; lit. "Dragon City") became the annual meeting place and served as the Xiongnu capital. The ruins of Longcheng were found south of Ulziit District, Arkhangai Province, in 2017.

North of Shanxi with the Tuqi King of the Left was holding the area north of Beijing and the Tuqi King of the Right was holding the Ordos Loop area as far as Gansu.

In the winter of 200 BC, following a Xiongnu siege of Taiyuan, Emperor Gaozu of Han personally led a military campaign against Modu Chanyu. At the Battle of Baideng, he was ambushed, reputedly by Xiongnu cavalry. The Emperor was cut off from supplies and reinforcements for seven days, only narrowly escaping capture.

The Han dynasty sent commoner women falsely labeled as "princesses" and members of the Han imperial family to the Xiongnu multiple times when they were practicing Heqin (Chinese: 和親; lit. 'harmonious kinship') marriage alliances with the Xiongnu in order to avoid sending the Emperor's daughters. The Han sent these "princesses" to marry Xiongnu leaders in their efforts to stop the border raids. Along with arranged marriages, the Han sent gifts to bribe the Xiongnu to stop attacking. After the defeat at Pingcheng in 200 BC, the Han emperor abandoned a military solution to the Xiongnu threat. Instead, in 198 BC, the courtier Liu Jing was dispatched for negotiations. The peace settlement eventually reached between the parties included a Han princess given in marriage to the chanyu; periodic gifts to the Xiongnu of silk, distilled beverages and rice; equal status between the states; and a boundary wall as a mutual border.

This first treaty set the pattern for relations between the Han and the Xiongnu for sixty years. Up to 135 BC, the treaty was renewed nine times, each time with an increase in the "gifts" to the Xiongnu Empire. In 192 BC, Modun even asked for the hand of Emperor Gaozu of Han widow Empress Lü Zhi. His son and successor, the energetic Jiyu, known as Laoshang Chanyu, continued his father's expansionist policies. Laoshang succeeded in negotiating terms with Emperor Wen for the maintenance of a large scale government sponsored market system.

While the Xiongnu benefited handsomely, from the Chinese perspective marriage treaties were costly, very humiliating and ineffective. Laoshang Chanyu showed that he did not take the peace treaty seriously. On one occasion his scouts penetrated to a point near Chang'an. In 166 BC he personally led 140,000 cavalry to invade Anding, reaching as far as the imperial retreat at Yong. In 158 BC, his successor sent 30,000 cavalry to attack Shangdang and another 30,000 to Yunzhong.

The Xiongnu also practiced marriage alliances with Han dynasty officers and officials who defected to their side by marrying off sisters and daughters of the chanyu to Han Chinese who joined the Xiongnu and Xiongnu in Han service. The daughter of Laoshang Chanyu (and older sister of Junchen Chanyu and Yizhixie Chanyu) was married to the Xiongnu General Zhao Xin, the Marquis of Xi who was serving the Han dynasty. The daughter of Qiedihou Chanyu was married to the Han Chinese General Li Ling after he surrendered and defected. Another Han Chinese General who defected to the Xiongnu was Li Guangli, a general in the War of the Heavenly Horses, who also married a daughter of the Hulugu Chanyu. The Han Chinese diplomat Su Wu married a Xiongnu woman given by Li Ling when he was arrested and taken captive. Han Chinese explorer Zhang Qian married a Xiongnu woman and had a child with her when he was taken captive by the Xiongnu.

The khagans of the Yenisei Kyrgyz Khaganate claimed descent from the Chinese general Li Ling, grandson of the Han dynasty general Li Guang. Li Ling was captured by the Xiongnu and defected in the first century BC. And since the Tang royal Li family also claimed descent from Li Guang, the Kirghiz Khagan was therefore recognized as a member of the Tang Imperial family. This relationship soothed the relationship when Kyrgyz khagan Are (阿熱) invaded Uyghur Khaganate and executed Qasar Qaghan.

The Xiongnu confederation was unusually long-lived for a steppe empire. The purpose of raiding the Central Plain was not simply for goods, but to force the Central Plain polity to pay regular tribute. The power of the Xiongnu ruler was based on his control of Han tribute which he used to reward his supporters. The Han and Xiongnu empires rose at the same time because the Xiongnu state depended on Han tribute. A major Xiongnu weakness was the custom of lateral succession. If a dead ruler's son was not old enough to take command, power passed to the late ruler's brother. This worked in the first generation but could lead to civil war in the second generation. The first time this happened, in 60 BC, the weaker party adopted what Barfield calls the 'inner frontier strategy.' They moved south and submitted to the dominant Central Plain regime and then used the resources obtained from their overlord to defeat the Northern Xiongnu and re-establish the empire. The second time this happened, around 47 AD, the strategy failed. The southern ruler was unable to defeat the northern ruler and the Xiongnu remained divided.

Historians disagree on whether the Xiongnu were of Iranian, Turkic, Mongolic or Yeniseian origin. However, some fragments from Chinese chronicles may indicate that the Xiongnu were of Turkic rather than Mongolian origin. Fragments relating to the Six Kingdoms period state that the Xiongnu spoke a language different from that of the Xianbei, who were of Mongolian or Paramongolian origin. However, the origin of the Xiongnu remains a subject of speculation.

The Xiongnu empire is widely thought to have been multiethnic. There are several theories on the ethnolinguistic identity of the Xiongnu, though there is no consensus among scholars as to what language was spoken by the Xiongnu elite.

The Xiongnu-Hun hypothesis was originally proposed by the 18th-century French historian Joseph de Guignes, who noticed that ancient Chinese scholars had referred to members of tribes which were associated with the Xiongnu by names which were similar to the name "Hun", albeit with varying Chinese characters. Étienne de la Vaissière has shown that, in the Sogdian script used in the so-called "Sogdian Ancient Letters", both the Xiongnu and the Huns were referred to as the γwn (xwn), which indicates that the two names were synonymous. Although the theory that the Xiongnu were the precursors of the Huns as they were later known in Europe is now accepted by many scholars, it has yet to become a consensus view. The identification with the Huns may either be incorrect or it may be an oversimplification (as would appear to be the case with a proto-Mongol people, the Rouran, who have sometimes been linked to the Avars of Central Europe).

Most scholars agree that the Xiongnu elite may have been initially of Sogdian origin, while later switching to a Turkic language. Harold Walter Bailey proposed an Iranian origin of the Xiongnu, recognizing all of the earliest Xiongnu names of the 2nd century BC as being of the Iranian type. Central Asian scholar Christopher I. Beckwith notes that the Xiongnu name could be a cognate of Scythian, Saka and Sogdia, corresponding to a name for Eastern Iranian Scythians. According to Beckwith the Xiongnu could have contained a leading Iranian component when they started out, but more likely they had earlier been subjects of an Iranian people and learned the Iranian nomadic model from them.

In the 1994 UNESCO-published History of Civilizations of Central Asia, its editor János Harmatta claims that the royal tribes and kings of the Xiongnu bore Iranian names, that all Xiongnu words noted by the Chinese can be explained from a Scythian language, and that it is therefore clear that the majority of Xiongnu tribes spoke an Eastern Iranian language.

According to a study by Alexander Savelyev and Choongwon Jeong, published in 2020 in the journal Evolutionary Human Sciences, "The predominant part of the Xiongnu population is likely to have spoken Turkic". However, important cultural, technological and political elements may have been transmitted by Eastern Iranian-speaking Steppe nomads: "Arguably, these Iranian-speaking groups were assimilated over time by the predominant Turkic-speaking part of the Xiongnu population".

Lajos Ligeti was the first to suggest that the Xiongnu spoke a Yeniseian language. In the early 1960s Edwin Pulleyblank was the first to expand upon this idea with credible evidence. The Yeniseian theory proposes that the Jie, a western Xiongnu people, spoke a Yeniseian language. Hyun Jin Kim notes that the 7th century AD Chinese conpendium, Jin Shu, contains a transliterated song of Jie origin, which appears to be Yeniseian. This song has led researchers Pulleyblank and Vovin to argue for a Yeniseian Jie dominant minority, that ruled over the other Xiongnu ethnicities, such as Iranian and Turkic people. Kim has stated that the dominant Xiongnu language was likely Turkic or Yeniseian, but has cautioned that the Xiongnu were definitely a multi-ethnic society.

Pulleybank and D. N. Keightley asserted that the Xiongnu titles "were originally Paleo-Siberian words but were later borrowed by the Turkic and Mongolic peoples". Titles such as tarqan, tegin and kaghan were also inherited from the Xiongnu language and are possibly of Yeniseian origin. For example, the Xiongnu word for "heaven" is theorized to come from Proto-Yeniseian *tɨŋVr.

Vocabulary from Xiongnu inscriptions sometimes appears to have Yeniseian cognates which were used by Vovin to support his theory that the Xiongnu has a large Yeniseian component, examples of proposed cognates include words such as Xiongnu kʷala 'son' and Ket qalek 'younger son', Xiongnu sakdak 'boot' and Ket sagdi 'boot', Xiongnu gʷawa "prince" and Ket gij "prince", Xiongnu "attij" 'wife' and proto-Yeniseian "alrit", Ket "alit" and Xiongnu dar "north" compared to Yugh tɨr "north". Pulleyblank also argued that because Xiongnu words appear to have clusters with r and l, in the beginning of the word it is unlikely to be of Turkic origin, and instead believed that most vocabulary we have mostly resemble Yeniseian languages.

Alexander Vovin also wrote, that some names of horses in the Xiongnu language appear to be Turkic words with Yeniseian prefixes.

An analysis by Savelyev and Jeong (2020) has cast doubt on the Yeniseian theory. If assuming that the ancient Yeniseians were represented by modern Ket people, who are more genetically similar to Samoyedic speakers, the Xiongnu do not display a genetic affinity for Yeniseian peoples. A review by Wilson (2023) argues that the presence of Yeniseian-speakers among the multi-ethnic Xiongnu should not be rejected, and that "Yeniseian-speaking peoples must have played a more prominent (than heretofore recognized) role in the history of Eurasia during the first millennium of the Common Era".

Bonmann and Fries (2025) argued that the Xiongnu and succeeding Huns were of multi-ethnic origin, but had, at least a partial, Paleo-Siberian, specifically Yeniseian ethnic core, corresponding to the early Arin language.

According to a study by Alexander Savelyev and Choongwon Jeong, published in 2020 in the journal Evolutionary Human Sciences, "The predominant part of the Xiongnu population is likely to have spoken Turkic". However, genetic studies found a mixture of haplogroups from western and eastern Eurasian origins that suggested large genetic diversity, and possibly multiple origins of Xiongnu elites. The Turkic-related component may be brought by eastern Eurasian genetic substratum.

Other proponents of a Turkic language theory include E.H. Parker, Jean-Pierre Abel-Rémusat, Julius Klaproth, Gustaf John Ramstedt, Annemarie von Gabain, and Charles Hucker. André Wink states that the Xiongnu probably spoke an early form of Turkic; even if Xiongnu were not "Turks" nor Turkic-speaking, they were in close contact with Turkic-speakers very early on. Craig Benjamin sees the Xiongnu as either proto-Turks or proto-Mongols who possibly spoke a language related to the Dingling.

Chinese sources link several Turkic peoples to the Xiongnu:

According to the Book of Zhou, History of the Northern Dynasties, Tongdian, New Book of Tang, the Göktürks and the ruling Ashina clan was a component of the Xiongnu confederation,

However, the Ashina-surnamed Göktürks were also stated to be they were "mixed barbarians" (雜胡; záhú) who fled from Pingliang (now in modern Gansu province, China). or from an obscure Suo state (索國), north of the Xiongnu.

Uyghur Khagans claimed descent from the Xiongnu (according to Chinese history Weishu, the founder of the Uyghur Khaganate was descended from a Xiongnu ruler).

The Book of Wei states that the Yueban descended from remnants of the Northern Xiongnu chanyu's tribe and that Yueban's language and customs resembled Gaoche (高車), another name of the Tiele.

The Book of Jin lists 19 southern Xiongnu tribes who entered the northern Chinese borders, the 14th being the Alat (Ch. 賀賴 Helai ~ 賀蘭 Helan ~ 曷剌 Hela); Alat being glossed "piebald horse" (Ch. 駁馬 ~ 駮馬 Boma) in Old Turkic.

Bonmann & Fries (2025) criticize Turkic theories, believing that the Turkic identification of the Xiongnu and their language is impossible. They argue that the hypothesis of a Turkic identity for the Xiongnu is "rather unlikely from the outset", since there is "neither direct nor indirect evidence" supporting the claim of any sort of Turkic presence in Inner Asia between the 3rd century BC and the 2nd century AD, let alone amongst the Xiongnu. They also note that such arguments rely primarily on written sources dating after the Xiongnu period, whereas the earliest secure evidence for Turkic languages appears much later. They conclude that the available evidence does not substantiate a Turkic linguistic presence among the Xiongnu during the period of their empire.

Chinese sources also ascribe Xiongnu origins to the Para-Mongolic-speaking Kumo Xi and Khitans.

Mongolian and other scholars have suggested that the Xiongnu spoke a language related to the Mongolic languages. Mongolian archaeologists proposed that the Slab Grave Culture people were the ancestors of the Xiongnu, and some scholars have suggested that the Xiongnu may have been the ancestors of the Mongols. Nikita Bichurin considered Xiongnu and Xianbei to be two subgroups (or dynasties) of but one same ethnicity.

According to the Book of Song, the Rourans, which the Book of Wei identified as offspring of Proto-Mongolic Donghu people, possessed the alternative name(s) 大檀 Dàtán "Tatar" and/or 檀檀 Tántán "Tartar" and according to the Book of Liang, "they also constituted a separate branch of the Xiongnu". The Old Book of Tang mentioned twenty Shiwei tribes, which other Chinese sources (the Book of Sui and the New Book of Tang) associated with the Khitans, another people who in turn descended from the Xianbei and were also associated with the Xiongnu. While the Xianbei, Khitans, and Shiwei are generally believed to be predominantly Mongolic- and Para-Mongolic-speaking, yet Xianbei were stated to descend from the Donghu, whom Sima Qian distinguished from the Xiongnu (notwithstanding Sima Qian's inconsistency). Additionally, Chinese chroniclers routinely ascribed Xiongnu origins to various nomadic groups: for examples, Xiongnu ancestry was ascribed to Para-Mongolic-speaking Kumo Xi, as well as the Turkic-speaking Göktürks and Tiele;

Genghis Khan refers to the time of Modu Chanyu as "the remote times of our chanyu" in his letter to Daoist Qiu Chuji. Moreover, the sun and moon symbol of Xiongnu discovered by archaeologists is similar to Mongolian Soyombo symbol.

Since the early 19th century, a number of Western scholars have proposed a connection between various language families or subfamilies and the language or languages of the Xiongnu. Albert Terrien de Lacouperie considered them to be multi-component groups. Many scholars believe the Xiongnu confederation was a mixture of different ethno-linguistic groups, and that their main language (as represented in the Chinese sources) and its relationships have not yet been satisfactorily determined. Kim rejects "old racial theories or even ethnic affiliations" in favour of the "historical reality of these extensive, multiethnic, polyglot steppe empires".

Chinese sources link the Tiele people and Ashina to the Xiongnu, not all Turkic peoples. According to the Book of Zhou and the History of the Northern Dynasties, the Ashina clan was a component of the Xiongnu confederation, but this connection is disputed, and according to the Book of Sui and the Tongdian, they were "mixed nomads" (traditional Chinese: 雜胡; simplified Chinese: 杂胡; pinyin: zá hú) from Pingliang. The Ashina and Tiele may have been separate ethnic groups who mixed with the Xiongnu. Indeed, Chinese sources link many nomadic peoples (hu; see Wu Hu) on their northern borders to the Xiongnu, just as Greco-Roman historiographers called Avars and Huns "Scythians".

The original geographic location of the Xiongnu is disputed among steppe archaeologists. Since the 1960s, the geographic origin of the Xiongnu has attempted to be traced through an analysis of Early Iron Age burial constructions. No region has been proven to have mortuary practices that clearly match those of the Xiongnu.

In the 1920s, Pyotr Kozlov oversaw the excavation of royal tombs at the Noin-Ula burial site in northern Mongolia, dated to around the first century AD. Other Xiongnu sites have been unearthed in Inner Mongolia, such as the Ordos culture. Sinologist Otto Maenchen-Helfen has said that depictions of the Xiongnu of Transbaikalia and the Ordos commonly show individuals with West Eurasian features. Iaroslav Lebedynsky said that West Eurasian depictions in the Ordos region should be attributed to a "Scythian affinity".

Portraits found in the Noin-Ula excavations demonstrate other cultural evidence and influences, showing that Chinese and Xiongnu art influenced each other mutually. Some of these embroidered portraits in the Noin-Ula kurgans also depict the Xiongnu with long braided hair with wide ribbons, which is seen to be identical with the Ashina clan hair-style. Well-preserved bodies in Xiongnu and pre-Xiongnu tombs in Mongolia and southern Siberia show both East Asian and West Eurasian features.

Analysis of cranial remains from some sites attributed to the Xiongnu have revealed that they had dolichocephalic skulls with East Asian craniometrical features, setting them apart from neighboring populations in present-day Mongolia. Russian and Chinese anthropological and craniofacial studies show that the Xiongnu were physically very heterogenous, with six different population clusters showing different degrees of West Eurasian and East Asian physical traits.

Presently, there exist four fully excavated and well documented cemeteries: Ivolga, Dyrestui, Burkhan Tolgoi, and Daodunzi. Additionally thousands of tombs have been recorded in Transbaikalia and Mongolia.

The archaeologists at a Xiongnu cemetery in Arkhangai Province said the following: "There is no clear indication of the ethnicity of this tomb occupant, but in a similar brick-chambered tomb of the late Eastern Han period at the same cemetery, archaeologists discovered a bronze seal with the official title that the Han government bestowed upon the leader of the Xiongnu. The excavators suggested that these brick chamber tombs all belong to the Xiongnu (Qinghai 1993)."Classifications of these burial sites make distinction between two prevailing type of burials: "(1) monumental ramped terrace tombs which are often flanked by smaller "satellite" burials and (2) 'circular' or 'ring' burials." Some scholars consider this a division between "elite" graves and "commoner" graves. Other scholars, find this division too simplistic and not evocative of a true distinction because it shows "ignorance of the nature of the mortuary investments and typically luxuriant burial assemblages [and does not account for] the discovery of other lesser interments that do not qualify as either of these types."

A 2003 study found that 89% of Xiongnu maternal lineages are of East Asian origin, while 11% were of West Eurasian origin. However, a 2016 study found that 37.5% of Xiongnu maternal lineages were West Eurasian, in a central Mongolian sample.

According to Rogers & Kaestle (2022), these studies make clear that the Xiongnu population is extremely similar to the preceding Slab Grave population, which had a similar frequency of Eastern and Western maternal haplogroups, supporting a hypothesis of continuity from the Slab Grave period to the Xiongnu. They wrote that the bulk of the genetics research indicates that roughly 27% of Xiongnu maternal haplogroups were of West Eurasian origin, while the rest were East Asian.

Some examples of maternal haplogroups observed in Xiongnu specimens include D4b2b4, N9a2a, G3a3, D4a6 and D4b2b2b. and U2e1.

According to Rogers & Kaestle (2022), roughly 47% of Xiongnu period remains belonged to paternal haplogroups associated with modern West Eurasians, while the rest (53%) belonged to East Asian haplogroups. They observed that this contrasts strongly with the preceding Slab Grave period, which was dominated by East Asian patrilineages. They suggest that this may reflect an aggressive expansion of people with West Eurasian paternal haplogroups, or perhaps the practice of marriage alliances or cultural networks favoring people with Western patrilines.

Some examples of paternal haplogroups in Xiongnu specimens include C2, Q1b, R1, R1b, O3a and O3a3b2, , J2a, J1a and E1b1b1a.

According to Lee & Kuang, the main paternal lineages of Xiongnu Elite remains in the Egiin Gol valley belonged to the paternal haplogroups N1c1, Q-M242, and C-M217. Xiongnu remains from Barkol belonged exclusively to haplogroup Q-M242. They argue that the haplogroups Q, N and C2 likely formed the major paternal haplogroups of the Xiongnu tribes, while R1a was the most common paternal haplogroup (44.5%) among neighbouring nomads from the Altai Mountains, who were probably incorporated into the Xiongnu confederation and may be associated with the Jie people. Alternatively, if the Jie, who formed a separate branch of the Xiongnu that founded the Later Zhao Dynasty (319–351 AD), were indeed a Yeniseian-speaking people, they may have been carriers of haplogroup Q and resembled modern-day Kets of central Siberia. Elite Xiongnu samples from Duurlig Nars belonged to R1a1 and to C-M217, while elite Xiongnu remains from Tamir Ulaan Khoshuu beloged to R1a-Z95, R1a-Z2125, Q1a-M120, J1 and N1a1. According to Juhyeon Lee sample DA39, a likely chanyu of the Xiongnu empire from Gol Mod 2 site in central-north Mongolia was assigned to haplogroup R1 and nobility from Shombuuzyn Belchir, samples SBB009, SBB010 and SBB001 belonged to haplogroup C-F1756. Högström et al. (2024) later reassigned DA39 to haplogroup R-Y56311.

A study published in the American Journal of Physical Anthropology in October 2006 detected significant genetic continuity between the examined individuals at Egyin Gol and modern Mongolians.

A genetic study published in Nature in May 2018 examined the remains of five Xiongnu. The study concluded that Xiongnu confederation was genetically heterogeneous, and Xiongnu individuals belonging to two distinct groups, one being of primarily East Asian origin (associated with the earlier Slab-grave culture) and the other presenting considerable admixture levels with West Eurasian (possibly from Central Saka) sources. The evidence suggested that the Huns probably emerged through minor male-driven geneflow into the Saka through westward migrations of the Xiongnu.

A study published in November 2020 examined 60 early and late Xiongnu individuals from across Mongolia. The study found that the Xiongnu resulted from the admixture of three different clusters from the Mongolian region. The two early genetic clusters are "early Xiongnu_west" from the Altai Mountains (formed at 92% by the hybrid Eurasian Chandman ancestry, and 8% BMAC ancestry), and "early Xiongnu_rest" from the Mongolian Plateau (individuals with primarily Ulaanzuukh-Slab Grave ancestry, or mixed with "early Xiongnu_west"). The later third cluster named "late Xiongnu" has even higher heterogenity, with the continued combination of Chandman and Ulaanzuukh-Slab Grave ancestry, and additional geneflow from Sarmatian and Han Chinese sources. Their uniparental haplogroup assignments also showed heterogenetic influence on their ethnogenesis as well as their connection with Huns. In contrast, the later Mongols had a much higher eastern Eurasian ancestry as a whole, similar to that of modern-day Mongolic-speaking populations.

A Xiongnu remain (GD1-4) analysed in a 2024 study was found to be entirely derived from Ancient Northeast Asians without any West Eurasian-associated ancestry. The sample clustered closely with a Göktürk remain (GD1-1) from the later Turkic period.

Although the Xiongnu were ethnically heterogeneous as a whole, it appears that variability was highly related to social status. Genetic heterogeneity was highest among retainers of low status, as identified by their smaller and peripheral tombs. These retainers mainly displayed ancestry related to the Chandman/Uyuk culture (characterized by a hybrid Eurasian gene pool combining the genetic profile of the Sintashta culture and Baikal hunter-gatherers, otherwise known as Baikal EBA), or various combinations of Chandman/Uyuk and Ancient Northeast Asian Ulaanzuukh/Slab Grave profiles.

On the contrary, high status Xiongnu individuals tended to have less genetic diversity, and their ancestry was essentially derived from East Eurasian-related Ulaanzuukh/Slab Grave culture groups, or alternatively from the Xianbei, suggesting multiple sources for their Eastern ancestry. High East Eurasian ancestry was more common among high status female samples, while low status male samples tended to be more diverse and West Eurasian shifted. DA39, a likely chanyu identified by his prestigious tomb, was shown to have had similar ancestry as a high status female in the "western frontiers", deriving about 39.3% Slab Grave (or Ancient Northeast Asian) genetic ancestry, 51.9% Han (or Yellow River farmers) ancestry, with the rest (8.8%) being Saka (Chandman) ancestry.