Sogdians were famous as traders along the Silk Road before the Islamization of Central Asia (see SOGDIAN TRADE at The earliest Sogdian record of their activities are the so-called Ancient Letters (s.v.) of the early 4th century discovered some 90 km west of Dunhuang (q.v. at However, close investigations of the Chinese sources lead one to assume that the Sogdians came to China for trading during the Later Han Dynasty, when the name of Sogdiana was first recorded in China (see Shiratori, pp. 94-100). The 5th-8th centuries saw the heyday of Sogdian merchants’ activities; especially during some hundred years before the An Lushan’s rebellion (755-63 C.E.), when Tang controlled Central Asia, a great many Sogdians were encountered in northern China.

In those days, when foreigners came to China, they were given special surnames mostly in accordance with their origins or homelands (TABLE 1). For example, those originating from India bore a surname Zhu (1) [numbers in parentheses refer to Table 1 entries], which is a short form of Tianzhu (2), the ancient Chinese designation of India. The number and the popularity of Sogdians in China was apparently the reason why they were given several different surnames, representing each oasis state from which they originated. Thus, those from Samarqand bore the surname Kang (3), Bukhara An (4), Kabudhan (north of Samarqand) Cao (5), Kushaniyya He (6), Tashkent (older Chach) Shi (7), Kesh Shi (8), and Maymurgh (modern Panjikant?) Mi (9) (Pulleyblank, p. 320). For example, a Sogdian (hu) boy who accompanied the Chinese pilgrim Xuanzang (? – 664 C.E.) at the beginning of his travels was named Shi Pantuo (10), whose surname indicates that he originated from Tashkent. The boy’s given name Pantuo (< *bʻuân dʻâ) is a transcription of a Sogdian name vandak (βntk) “slave,” which is attested as a personal name among the Upper Indus inscriptions (Sims-Williams, 1992, p. 47).

All these surnames were collectively referred to as Zhaowu (jiu) xing (11) “(nine) surnames of Zhaowu,” because Sogdians were believed by the Chinese to have originated from Zhaowu, a town in Gansu; the number nine in this case seems to mean “numerous.” Among the seven surnames, Kang and An are the oldest and were attested already in the Han time. However, in the older days Kang represented Kangju (12), which denotes a nomad state once ruling the area including Sogdiana, while An is short for Anxi (13), a transcription of Aršak “Arsaces,” i.e., the Parthian state, dating back to the Han period. An Shigao (14), who came to China in the 2nd century and was alleged to be a Parthian crown prince, is the earliest example of An as a surname. Later, when Kangju and Anxi no longer existed, the names came to be applied to Samarqand and Bukhara respectively. The others began to be attested later in the 6th century. Mi (< Middle Chinese *miei), Shi (8) (< *ṣi), and Shi (7) (< *źiäk) are characters representing parts of their original names: Maymurg, Kesh, and Chach. However, the reason why Cao and He were selected is not known. Apart from the seven most common surnames, there were a few others, but only Bi (15) has been identified (with Paykent). It is also to be noted that the other six names, but not Mi, were also borne by non-Sogdians (see de la Vaissière, pp. 124-27). On the other hand, due to the intermarriage between Turkish peoples, one sometimes finds Sogdian names among those who bore the surname Zhe (16), which is generally assumed to be a surname of a Turkic tribe, e.g., (Zhe) Hutianpantuo (17) (Li and Wang, p. 113) = Xwatenvandak (xwtʾyn-βntk) "queen's slave."

Names of Sogdians transcribed in Chinese characters are attested both in the Chinese historical records handed down to this day and in the documents or inscriptions unearthed in modem times. However, since Sogdians were foreigners and did not play a main role in historical events, the number of their names recorded in the chronicles and similar historical texts is restricted. Although more and more funeral inscriptions of Sogdian-Chinese people have been discovered in recent years, most of their given names are totally sinicized. Thus, the bulk of Sogdian onomastica transcribed in Chinese characters are attested in Chinese texts discovered in Dunhuang and East Turkestan. In particular, recent excavations of the graveyards of Astana and Karakhoja in Turfan have brought to light many Chinese documents of the 6th-8th centuries, in which numerous Sogdians, both sedentary and itinerant, are mentioned. Of a total of ca. 10,500 persons named in the documents, more than 850 bear Zhaowu surnames; those with Kang number 352, Cao 179, An 118, Shi (< *ṣi) 115, He 59, Shi 38, Mi 12. About one third or one quarter of them bear names that sound like Sogdian. Among the Dunhuang texts there has survived one census document dating back to ca. 750 C.E. for a Sogdian settlement named Conghuaxiang. The register was extensively analyzed by O. Ikeda, who also attempted to reconstruct some of the names in their original forms (Ikeda, 1965); as a Sogdianist, D. Weber followed him in reconstructing others. According to Ikeda's study, out of 236 personal names found in the register, 214 people bear Zhaowu surnames, and 100 personsʼ given names are in Sogdian or at least do not sound like Chinese.

Names of the rulers of Sogdian oasis states consitute an independent group. Each time they sent tribute to the Chinese court, their names were recorded in the Chinese chronicles. Their names, either without surnames or with a surname Zhaowu, were collected by O. I. Smirnova and compared with legends of coins found in Sogdiana and with those in Islamic sources. For example, the three Samarqand kings of the 8th century are as follows: Tu hun (18) (*tʻuət xuən) = Tarkhun (trxwn); Wu lei qie (19) (*˙uo lək g’ia) = Urak (ʾwrʾk); Duo he (20) (*tuət γât) = Turghar (twrγʾr; Smimova, pp. 423-31).

When one tries to find out the original forms of the names phonetically transcribed in Chinese, one first reconstructs the Middle Chinese pronunciation of each character. Then, considering the subsequent phonological change and possible range of correspondence between Chinese sounds and Sogdian phonemes, one searches for a candidate among those Sogdian names so far known; however, one sometimes tries to reconstruct possible but so far not attested forms purely from the transcription. A case in point is An Lushan (21) (? – 757 C.E.), who is perhaps the most famous Sogdian in the history. Son of a Sogdian father and a Turkish mother, he became a favorite of the emperor Xuanzong (r. 712-56 C.E.) but later rebelled against him and attempted to be an emperor himself. His surname suggests his father’s Bukharan origin, and his given name Lushan (< *luk ṣan) was recognized by W. B. Henning as transcribing rwxšn- [roxšn-] “light, bright” (apud Pulleyblank, p. 333, n. 1). Though rwxšn- has not been attested as a personal name in Sogdian texts, Henning compared it with cognate onomastica in other Iranian languages, such as Rōxanē, the princess of Bactria who became the wife of Alexander the Great. His name happened to be the first to be identified with the corresponding Sogdian form.

In some cases popular elements found in Sogdian names help identify the original forms of the Chinese names. Such are βntk “slave,” prn “glory, fortune,” and yʾn “boon,” the three most frequent second members of Sogdian onomastica, and -c, the most common suffix constituting the hypocoristics or short names. Some examples are as follows: Cao Naningpan (22) (*nâ nieng pʻuân; Li and Wang, p. 329) = Nanaifarn (nny-pm); An Fuyemenyan (23) (*bʻi̯u i̯a muən i̯än; Li and Wang, p. 183) = Avyāman-yān (ʾβyʾmn-yʾn); An Fuyemenzhe (24) (*bʻi̯u i̯a muən tśi̯a; Li and Wang, p. 183) = Avyāmanch (ʾβyʾmnc). For the form ending with -βntk, see examples cited above and below. However, it is to be noted that, in view of great differences existing between phonological and syllabic structures of Chinese and Sogdian, uncertainties always surround the proposed identifications, and there remain numerous transcriptions left unreconstructed.

Among the theophoric names one often finds the Zoroastrian deities. Since the names of deities also served as calendar terms, their names may perhaps represent the day or month when they were born. The above-mentioned examples contain nn(y) “Nana(i),” the most popular goddess in Sogdiana, and ʾβyʾmn “Avyāman,” a Sogdian counterpart of Wahman (Sims-Williams, 1992, p. 40). Ahuramazda is referred to as “creator” in Dishebo (25) (*dʻi śi̯a puât; Yoshida, p. 38) = δš(c)yʾpt “protected by the creator.” The Sogdian form of Tištrya is encountered in Zhishifan (26) (*tie śi pʻi̯wen; Li and Wang, p. 312) = tyš-prn “Tishtrya’s glory.” Mithra is nicknamed as “god” by Sogdians (Sims-Williams, 1991 [1992]), and several names containing the element βγ- “god” are found in the materials, e.g., Baobi (27) (*bʻuâk bʻji; Li and Wang, p. 417) = βγyβyrt “obtained from the god.” Another popular goddess is the Moon, as in Mopan (28) (*mâk pʻuân; Li and Wang, p. 129) = mʾx-prn “the Moon’s glory.” Zymtyc, the 11th month of the Sogdian calendar, is named after a god Zymt, who is mentioned in the name Shewupantuo (29) (*dźi̯a mi̯uət bʻuân dʻâ; Ikeda, 1965, p. 64) = zymt-βntk. In the light of its Bactrian counterpart drēmatigano, Sims-Williams (2000, p. 190) proves that the god originated from the Greek earth goddess Demeter. A Middle Persian loanword for “Tuesday” is found in the name Wenhan (30) (*˙˙uən xân, cf. British Library MS S. 542, line 74; Ikeda, 1979, p. 537) = wnxʾn.

Some Sogdians were converted to Buddhism and bear names beginning with buti- “the Buddha,” e.g.—one of the rare female personal names in the material—Fuzhitai (31) (*bʻi̯tu tie dʻai; Li and Wang, p. 180) = pwty-δʾyh “the Buddha’s slave [fem.].” It is interesting to note that these Buddha names are all attested only after the latter half of the 7th century. The fact seems to hint at the period when Sogdians were converted to Buddhism (Yoshida, 1998, pp. 40-41). Names which can be attributed to Christianity or Manicheism are very few. The one such name known to me is Yousuoyan (32) (*ji̯ǰu ṣi̯wo i̯än; Li and Wang, p. 128) = yšw-yʾn “Jesus’ favor.” Hypocoristics are also found: Yena (33) (*i̯a nâ; Li and Wang, p. 332) = yʾnʾkk (derived from yʾn), Funa (34) (*pʻi̯uət nâ; Li and Wang, p. 312) = prnʾkk (derived from prn). Adjectives derived from place names serve as personal names in the following: Boxi (35) (*puât si̯ək; Li and Wang, p. 333) = pʾrsyk “Persian,” Yintujianing (36) (*i̯en tʻuo ka nieng; Ikeda, 1965, p. 65) = ʾyntwkʾnʾk “Indian.”

Ordinary Sogdian words, adjectives, nouns, and compounds are also employed as personal names: Yexi (37) (*i̯a xjei; Ikeda, 1965, p. 65)= yʾxy “brave,” Zhejie (38) (*tśi̯a ki̯et; Ikeda, 1965, p. 65) = cʾkr “warrior,” Fudanyan (39) (*pʻi̯uət tâm i̯än; Ikeda, p. 63) = ʾprtmyʾn “first boon” (a name given to a first born). As a result of intermarriage between Sogdians and Turkish people one sometimes finds Turkish titles as personal names, e.g., Yijin (40) (*i̯et ki̯ən; Ikeda, 1965, p. 65) = Turk. irkin (a title borne by tribal chiefs; Clauson, p. 225).

As Ikeda’s study clearly shows, after settlement in China Sogdians became sinicized and came to be given Chinese names. No case is known where sons are given Sogdian names by fathers with Chinese names. Recently, tombs of a Sogdian family were excavated in Guyuan, and the epitaphs attest names of the members of the families for seven generations. Their geneological table shows that the first four generations continued to give Sogdian names to their sons (Luo, p. 216). Some other names are assumed to be translated from Sogdian originals. According to Henning (apud Pulleyblank, p. 337, n. 3), Shennu (41) lit. “godʼs slave” can be a translation of βγy-βntk. But, it is always possible that such names are genuine Chinese.



G. Clauson, An Etymological Dictionary of Pre-Thirteenth-Century Turkish, Oxford, 1972.

O. Ikeda, “Hasseiki chūyō ni okeru Tonkō no Sogudojin shūraku” (A Sogdian colony at Tunhuang in the mid-eighth century; with English summary), Yūrashia bunkakenkyū (Hokkaidō University)1, 1965, pp. 49-92.

Idem, Chūgoku kodai sekichō kenkyū: gaikan, rokubun (English title: Ancient Chinese household registers and related documents: a historical study: general introduction and collected documents; with English summary), Tokyo, 1979.

B. Karlgren, Grammata Serica Recensa, Stockholm, 1957.

Li Fang and Wang Su, Tulufan chu tu wen shu ren ming di ming suo yin (Index of personal names and place names attested in the Chinese documents unearthed in Turfan), Beijing, 1996.

Luo Feng, Guyuan nan jiao Sui Tang mudi (Graveyard of Sui and Tang dynasties in the south suburbs of Guyuan;with English summary), Beijing, 1996.

E. G. Pulleyblank, “A Sogdian Colony in Inner Mongolia,” T’oung Pao 41, 1952, pp. 317-56.

K. Shiratori, “A Study on Su-tʻê [42] or Sogdiana,” Memoirs of the Research Department of the Toyo Bunko 2, 1928, pp. 8 1-145.

N. Sims-Williams, “Mithra the Baga,” in P. Bernard and F. Grenet, eds., Histoire et cultes de l’Asie centrale pré-islamique, Paris, 1991 [1992], pp. 177-86.

Idem, Sogdian and Other Iranian Inscriptions of the Upper Indus II, London, 1992. Idem, Bactrian Documents I, Oxford, 2000.

O. I. Smirnova, Svodnyĭ katalog sogdiĭskikh monet (A summary catalogue of Sogdian coins), Moscow, 1981.

E. de Ia Vaissière, Histoire des marchands sogdiens, Paris, 2002.

D. Weber, “Zur sogdischen Personennamengebung,” Indogermanische Forschungen 77, 1972, pp. 192-208.

Y. Yoshida, “Sino-Iranica” (in Japanese), Bulletin of the Society for Western and Southern Asiatic Studies (Kyoto University),no. 48, 1998, pp. 33-51.

(Y. Yoshida)

Originally Published: August 15, 2006

Last Updated: August 15, 2006