The Sogdian “Ancient Letters”

In 1907 Aurel Stein found in the rubbish of a ruined watchtower (his site T.XII.a) on the Han frontier wall that guarded the road from Dunhuang to Loulan a group of letters on paper written in Sogdian, the Iranian language of the region around Samarkand: five nearly complete letters (six by V. A. Livshits’ count) and fragments of others. Each was folded several times, with the names of the sender and the addressee on the outside; most were tied with string, and one, wrapped in silk and enclosed in an envelope of coarse cloth, was addressed to Samarkand. At least two were written at Dunhuang and one at Guzang (Wuwei). They seem to be the contents of a mailbag lost or abandoned on its way west, and they are now kept in the British Library.

The longest, Letter II, was sent by Nanai-vandak, an agent stationed somewhere in Gansu, perhaps at Jincheng (Lanzhou), to his partners in Samarkand. He reports a severe famine, fighting between the Chinese and the Huns (Sogdian xwn), who “yesterday were the emperor’s subjects”, the flight of the emperor and the burning of his palace: “Luoyang is no more, Ye is no more!” The Indians and Sogdians in the capital, he writes, had all died of hunger. He mentions a hundred “freemen” from Samarkand living in one Chinese town and gives instructions about his money and a consignment of musk; the letters also mention silver, linen and pepper. Letters I and III were dictated by Miwnay, a woman abandoned at Dunhuang by her husband Nanai-dhat: living on the charity of a priest, she told him that she would rather be the wife of a dog or a pig than his.

The events described in Letter II allowed W. B. Henning in 1948 to date the letters to 312–313: the sack of Ye in 307 and the capture of Luoyang in 311 by the armies of the Xiongnu-led Han-Zhao state, when the Jin emperor Huai was taken prisoner. His dating is generally accepted, and Nicholas Sims-Williams places Letter II in about June or July 313; other proposals — J. Harmatta’s link with an earlier sack of Luoyang in 190 and a dating of the paper to no earlier than the 6th century (A. Fujieda) — are not accepted. The letters are the oldest substantial Sogdian documents and a key source on the early Sogdian colonies in China, and their use of xwn for the Xiongnu is an important argument in the long debate on whether the Xiongnu and the European Huns bore the same name.