Étienne de la Vaissière

École pratique des hautes
études Sciences historiques
et philologiques, Paris


he Sogdians were the inhabitants of fertile valleys surrounded by deserts, the most important of which was the Zeravshan valley, in today’s Uzbekistan and Tajikistan. This Iranian-speaking people had a fifteen-centuries-long historical identity between the sixth century BCE and the tenth century CE when it vanished in the Muslim, Persian-speaking world. Although the Sogdians constructed such famous towns as Samarkand and Bukhara, they are quite unknown. Only specialists on the Silk Road know that they were among the main go-betweens of the exchanges in the steppe, in Central Asia, and in China during the first millennium CE, and especially between the fifth and the eighth centuries CE. During this period, the “inland silk road” and the “Sogdian trading network” are almost synonymous. The contemporary Sogdian, Chinese, Arabic, Byzantine, and Armenian sources describe the Sogdians as the great traders of Inner Asia. They managed to sell their products - musk, slaves, silverware, silk and many other goods - to all the surrounding peoples. A Greek text describes their trading embassies to Byzantium, some caravaneers’ graffiti prove that they were in India, Turkish vocabulary is a testimony to their cultural and economic power in the Turkish steppe...

But their main market was always China. The Chinese branch of their network is by far the best known, and in China the number of new discoveries on the Sogdians is quickly growing.

When did the first Sogdian arrive in China to trade? Various answers have been given to this quite simple question, but one of the most popular, which can still be found in many recent books or articles, makes Alexander the Great the villain of the story. Due to the disaster which befell Sogdiana during his campaigns there in 329-328 BCE, the Sogdians would have been forced to emigrate as far east as China. The creation of the Sogdian trading network between Samarkand and China thus was a by-product of the Greek conquest of the Achaemenid Empire. Needless to say, there is nothing in the available sources to support such an idea, which survived only because of its classical flavor and its familiar ring to the historian’s ears. In fact the Chinese, Yuezhi, Bactrians, Indians and Sogdians who created the historical Silk Road did not need Greek help. Trade is yet another item that should be removed (after irrigation, town-planning and state-formation) from the long list of supposed Greek influences in the history of Central Asia. The list of real influences is already full enough with coinage, iconography and the alphabet (in Bactria)!

It has long been known that Chinese diplomacy towards the nomads in the second century BCE was instrumental in creating in Central Asia and further west, in Parthia, a market for Han products, especially silk. Chinese embassies traveled with thousands of bolts of silk but at very irregular intervals. The merchants in northwestern India and eastern Iran were quick to appreciate the potential for this exchange and followed the steps of the Chinese ambassadors back to China. As Du Qin, a Chinese statesman, put it in 25 BCE, “There are no members of the royal family or noblemen among those who bring the gifts. The latter are all merchants and men of low origins. They wish to exchange their goods and conduct trade, under the pretext of presenting gifts” [Han shu, Hulsewé 1979, p. 109]. The Sogdians were doing exactly the same thing at the same time, and the first testimonies, in 29 and 11 BCE, on a Sogdian in China might also be found in the Han shu: “If in view of these considerations, we ask why [Kangju] sends his sons to attend [at the Han court], [we find] that, desiring to trade, they use a pretence couched in fine verbiage” [Han shu, Hulsewé 1979, p. 128]. Kangju, a nomadic state, the center of which was in what is now the southern part of the Tashkent oasis, included Sogdiana during the first century BCE. This desire to trade was a Sogdian one.

It is difficult to understand the next step in the establishment of Sogdian communities in China. It seems that some of the ambassadors and their families settled in China, especially in Gansu. Some late genealogies of Sogdian families in China seem at least to imply such a reconstruction. We know on a firm textual basis that as early as 227 CE, in Liangzhou (Gansu), when a conquering army was approaching from the South, “The various kings in Liangzhou dispatched twenty men including Zhi Fu and Kang Zhi, the ennobled leaders of the Yuezhi and Kangju Hu, to receive the military commander, and when the large army advanced north they competed to be the first to receive us” [Sanguo zhi, 4, p. 895]. The Hu from Kangju are the Sogdians, while the Yuezhi are the traders from Bactria and Gandhara, the Kushan Empire created by the Yuezhi tribes. The leaders of the biggest trading communities in Gansu were sent to the invading army, and the Sogdians were already on a par with the greatest merchants of Antiquity, the Kushan ones.

The Sogdian Sabao and ambassador
An Jia negotiating with a Qaghan

The next step in the history of the Sogdians in China is provided by the “Sogdian Ancient Letters.” These letters were left in the ruins of a Han watchtower, 90 kilometers west of Dunhuang in 313. Sir Aurel Stein discovered them there in 1907. They were sent by some Sogdian traders from Gansu to the West. One of them was sent from Gansu to Samarkand and described the political upheavals in Northern China. The Xiongnu, who were then sacking the main towns there, were called for the first time in an Indo-European language by the name which would be theirs in Europe a century later: the Xwn, Huns. The letter described also the ruin of the Sogdian trading network in these towns: “The last emperor, so they say, fled from Luoyang because of the famine and fire was set to his palace and to the city, and the palace was burnt and the city [destroyed]. Luoyang is no more, Ye is no more! [...] And, sirs, if I were to write to you about how China has fared, it would be beyond grief: there is no profit for you to gain from it [...] [in] Luoyang... the Indians and the Sogdians there had all died of starvation” [Sims-Williams, 2001, p. 49]. But the same text describes a Sogdian network well established in Gansu which was still there a century later. A Chinese text explains how “Merchants of that country [Sogdiana] used to come in great number to the district Liang [the present Wuwei in Gansu] to trade. When Guzang [i.e. Wuwei] was conquered [by the Wei in 439] all of them were captured. In the beginning of the reign of Gaozong [452-465] the king [of Sogdiana] sent embassies to ask for their ransom” [Wei shu, Enoki 1955, p. 44].

The fifth and sixth centuries were certainly the high days of Sogdian emigration to China. After the disruption of the Inner China network in the fourth century, a new network of Sogdian communities was created then. Many Sui and Tang texts or funerary epitaphs of Sogdian families describe how the great grandfather came to China during the Wei as Sabao, i.e. chief caravaneer. These families established themselves first in Gansu, the next generation moved into the main Chinese towns, and some Sogdians managed to reach the court. For instance, the biography of An Tugen in the Bei shi (chap. 92 p. 3047) describes how An Tugen’s great grandfather came from Anxi (western Sogdiana) to the Wei and established himself in Jiuquan (the western end of Gansu). Later on, An Tugen rose from the position of merchant to Grand Minister of the Northern Qi in the middle of the sixth century.

New discoveries from Guyuan in the Chinese province of Ningxia (Southern Ordos) provide a very good example of Sogdian families in China who did not achieve such high distinction. Six graves of one Sogdian family have been excavated there [Luo Feng, 1996 and 2001]. According to its name, the family should have originated from the Sogdian town of Kesh (Shahr-i Sabz, in Uzbekistan), and the texts of the funerary epitaphs describe indeed how the family migrated from the western countries. The archaeological content of these looted tombs confirms these western links, as some Byzantine and Sasanian coins, a seal stone inscribed in Pahlavi, and a Zoroastrian symbol were found there. The great grandfather, Miaoni, and the grandfather, Boboni, “served their country in the capacity of Sabao.” The father, Renchou, “idled away his life, accomplishing nothing in his official career.” A member of the fourth generation, Shi Shewu (d. 610), was the great man of the family, and through him the family became integrated into Chinese society. He was a military officer of the Sui, and his grave and funerary epitaph are Chinese. His elder son, Shi Hedan (d. 669), was translator in the Imperial Secretariat of the Tang. Another son, Shi Daoluo (d. 658) was a soldier. A grandson, Shi Tiebang (d. 666), was in charge of an army horse-breeding farm near Guyuan. Shi Daode (d. 678), from another branch of the Shi clan, and his uncle, Shi Suoyan, who is buried in same graveyard, were also members of the military and officialdom. We know also by name many other members of the family. Some of these names are simply transcriptions of Sogdian names: Shewu is the honorific personal name, but the public name was Pantuo. Shewu was pronounced Jia-mut, and Pantuo banda; together they provide Jimatvande, a well-known Sogdian name, “servant of Demeter”, which was only divided in two halves for the need of the interpretatio sinica. The first names of the next generation sound Chinese, except in the elder branch: Shi Hedan and his son Shi Huluo have first names that seem to be transcriptions. After so many generations in China, some of the members of the families still married in the Sogdian milieu, among them Shi Hedan, great-great-grandson of Shi Miaoni, who married a Kang (the Chinese surname of the natives of Samarkand), and Shi Suoyan, who married an An (from Bukhara).

The main (western) wall of the Afrasiab painting in Samarkand:
Chinese, Korean, and Iranian ambassadors.

These data are fascinating, because with the Shi family we can follow the destiny of a specific Sogdian family in China and see how it became integrated first in the Sogdian milieu and then how it became sinicized in the Chinese administration. So far we lack other examples, due mainly to the novelty of the interest towards the Sogdian families in Chinese archaeology. Some other Sogdian graveyards have been found but badly published [See de la Vaissière and Trombert 2004, forthcoming, for other examples from textual sources].

The ancestors of Shi Shewu bore a title, Sabao, which is very interesting for the history of the Sogdians in China. It proves that the Sogdian communities in China were deeply rooted in the caravan trade. This title is a transcription of the Sogdian word sartapao, itself a Sogdian transcription of the Indian sârthavâha, chief-caravaneer, through a Bactrian intermediary [Sims-Williams, 1996]. In India the sârthavâha was not only the chief-caravaneer but also the head of the traders’ guild. In China the title was imported first as early as the second century CE under the form sabo directly from India, for example in Buddhist texts which described the Buddha as a chief-caravaneer, and then a second time from the Sogdian form with the precise administrative meaning “head of the local Sogdian community.” The “river of the sabao” in Chinese geography is the Zeravshan, which flows in Samarkand and Bukhara. So the heads of the Sogdian communities in China bore the titles of “chief caravaneer,” and it is on this ambiguity that Shi Shewu played when he wrote that his ancestors “served their country in the capacity of Sabao.” For a Chinese ear, they were officials, while for a Sogdian one, they were simply heads of caravan! We can see in the textual and epigraphical sources many such Sabao installed in China. Most of the main towns of Northern China had in the sixth and seventh centuries their Sogdian community headed by a Sabao, who received a mandarinal rank in the offical hierarchy, at least from the Northern Qi to the Tang. These nouveaux riches had some wealthy funerary beds carved for them, where they displayed both their Sogdian culture and their integration into Chinese society, in a way the iconographic counterpart of the epitaphs of the Shi family. These funerary beds were an old Chinese tradition well suited for Zoroastrian purposes because it isolated the body from earth and water. Some of them have been known for a long time; for example, the Anyang (Ye) one, which shows a Sabao in his garden receiving the members of his community. Others have been found recently and have found their way into Chinese or foreign museums.

One of the most interesting was discovered two years ago near Taiyuan by a team of the Shanxi Archaeological Institute. The tomb of Yu Hong, who died in 593 at age 58, contained a funerary bed in the shape of a Chinese house, adorned by 53 carved panels of marble, originally painted and gilded. Yu Hong had traveled extensively, acting as an ambassador to the Ruanruan, in Persia and Bactria or Gandhara (Yuezhi), and to the Tuyuhun tribes near lake Qinghai. Then he served the Northern Qi and Zhou, and the Sui. He became Sabao in 580 and then nominal governor of a town. We know this because the funerary epitaphs of Yu Hong and his wife were discovered in the tomb. And the iconography fits very well the geography of the texts: we see on the panels Yu Hong hunting with nomads on horses, but also hunting on an Indian elephant or banqueting with his wife. Zoroastrian symbols are clearly displayed: two priests half-bird, half-human wearing the traditional padam (a piece of cloth in front of the mouth) and Mithra and his sacrificial horse facing each other on each side of the entry [Marshak 2002, and Riboud 2003].

The trade links with Central Asia provided the communities from the fifth to the eighth centuries with waves of new immigrants. One specific example of these links is from a discovery on the main stage of the route, Turfan, in Xinjiang. Many Chinese documents were used there to cut paper clothing for the dead who were buried in the Astana cemetery. Among them is a list of taxes paid on caravan trade in the Gaochang kingdom (Turfan) in the 620s. The text is not complete but gives a fairly good idea of the identity of the main traders in Turfan: out of 35 commercial operations in this text, 29 involved a Sogdian trader. In 13 instances both the seller and the buyer were Sogdians.

In Inner China, the seventh century saw an evolution in the official position of the Sogdian communities. It seems that the Tang transformed quite independent and autonomous Sogdian communities loosely integrated in the mandarinal hierarchy into more controlled “submitted counties” without Sogdian hierarchy. The Sabaos disappeared from the epigraphical and textual sources after the middle of the seventh century. But this period, up to the middle of the eighth century, was certainly the climax of Iranian influence on Chinese civilization. If the communities were suppressed, the families and individuals who were before inclined to stay within the Sogdian communities now were integrated more thoroughly into Chinese society. We can see people with typical Sogdian surnames, such as Kang, getting involved in all the fields of Tang social life.

Obviously many of them were merchants: around the main markets of the capitals, Chang’an and Luoyang, Sogdian temples, Sogdian taverns, and Sogdian shops flourished. They sold to the Tang elite the Western goods that were then à la mode [Schafer 1963]. Many young nobles or drunken poets celebrated the charms of the Sogdian girls, and the most famous of them, Li Bo, wrote:

That Western hour with features like a flower
She stands by the wine-warmer, and laughs with the breath of spring Laughs with the breath of spring,
Dances in a dress of gauze!
Will you be going some where, milord,
now, before you are drunk?
[transl. Schafer 1963, p. 21]

A segment of the procession depicted on the south wall of the
palace at Afrasiab

It was not just merchants, but soldiers, monks, and high or low officials who were of Sogdian descent. We have seen that the sons and grandsons of Shi Shewu served in the army. To quote another example, the New History of the Tang describes the family of An Chongzhang, Minister of War from 767 to 777. His ancestors were sabao in Wuwei (Gansu) during three generations. At the fourth generation, a member of the family An Xinggui became “Wude-era meritous servant” and from then on the family belonged to the administration: the sixth-generation An Zhongjing was military Vice-commissary-in-chief of Hexi, and An Chong-zhang was the leading member of the seventh generation.

Research on such materials is just beginning. Many funerary epitaphs of the Tang period should be used to understand the rapid pace of the sinicization of the Sogdians that took place under the Tang. It is already clear though that the rebellion of An Lushan is a major stage in this process. An Lushan was the main military governor of northeastern China on the frontier with Korea and the Kitans. His father was a Sogdian installed in the Turk Empire and his mother was Turk; his first name is a straight transcription of the Sogdian Rokhshan, “luminous” (the same as Roxane, Alexander’s wife). He established himself as a young boy in northeastern China, acted as a translator there in the markets, became a soldier and climbed from the rank and file to the top of the army. His rebellion in 755 nearly destroyed the Tang dynasty and put an end to one of China’s Golden Ages. The rebellion was quelled only in 763 with the help of the Uighur nomads.

The rebellion has been described as if it was only a military coup by autonomous and very powerful armies [Pulleyblank, 1955]. The Sogdian identity of the rebels has never been investigated as such. Yet many texts described it as a Sogdian rebellion and described how many Sogdian traders supported An Lushan. Furthermore, some new discoveries prove that this idea was not due to a xenophobic bias in the Chinese descriptions of the revolt but in a way was something claimed by the rebels themselves: Shi Seming, the second successor of An Lushan, himself a Sogdian, put the Sogdian royal title of Jamuk (Jewel, transcribed Zhaowu in Chinese) on par with Huangdi on his recently discovered ceremonial jades. The troops of the rebels bore the Sogdian name of Zhejie, a fair transcription of the Sogdian Châkar “professional soldier” [de la Vaissière 2004, forthcoming].

The Sogdian milieu was torn apart by the rebellion, many Sogdians in China siding with the Tang. But from then on the Sogdians in China began to conceal their foreign origins. Maybe the clearest example is An Chongzhang, the Minister of War. In 756 he asked for the authorization to change his family name, “being ashamed to bear the same name” as An Lushan. He became Li Baoyu and the exchange was retroactive: his ancestors’ family name changed also [Forte 1995, pp. 24-7]. There are many other examples of this kind of social hiding. To the degree that we can follow the destiny of some Sogdian families in Northern China up to the ninth century, we see that their great days are already by then gone. The pace of sinicization grew faster for security reasons, while the international trade with its new waves of immigrants was totally disrupted in the second half of the eighth century and reborn only on a very low level, if at all, during the ninth century. The Persian traders, arriving by sea in the main harbors of Southern China, became the main traders of the age. This was the end of one millennium of a Sogdian presence in China.

About the Author

Étienne de la Vaissière is Assistant Professor at the Ecole pratique des hautes études, in Paris, where he teaches Medieval History of Central Asia. His dissertation on the Sogdian trade, a comprehensive analysis of the trading network from its creation to its end based on all the available sources, archaeological and textual, has been published (see below; it can be ordered on www.deboccard.com). He is currently working on Central Asian medieval sociology, on the Chinese as well as the Iranian side.


Detail from mural of procession on south wall of the
palace at Afrasiab
(photographed in Afrasaib Museum)

From the Editor | The Archaeology of Sogdiana | Returning to Varakhsha |Sogdians in China: A Short History and Some New Discoveries | The Pre-Islamic Civilization of the Sogdians (seventh century BCE to eighth century CE): A Bibliographic Essay (studies since 1986) | Bamiyan: Professor Tarzi's Survey and Excavation Archaeological Mission, 2003| 'Knowing the Road That Leads You Home': Family, Genealogy, and Migration in Post-Socialist Kazakhstan¹ | Among the Kazakhs of Xinjiang | Announcements