A Judaeo-Persian Letter from Dandan-Uiliq.

The British Library’s Asian and African studies blog, which I’ve posted about more than once (e.g.), has An eighth century Judaeo-Persian letter from Dandan-Uiliq (originally posted there by Ursula Sims-Williams 19 June 2020):

[…] The document was provisionally dated to the end of the eighth century when the site was abandoned, and this dating was confirmed by an analysis of the paper by Professor J. Wiesner (Margoliouth, pp. 742-3) which found that the structure was indistinguishable from the paper of Chinese documents found at Dandan Uiliq, dating from between 781 and 790.

The letter proved to be written in Judaeo-Persian, i.e. Persian written in Hebrew script. However since the beginning and end of each line was missing, there was only a limited amount of contextual information to be deduced (for an edition and translation see Utas, 1968 below). Mention of sheep trading and cloth indicates the document’s commercial nature and a reference to the author having written “more than 20 letters[1]” attests perhaps to a thriving trade. There is also an intriguing request for a harp required for instructing a girl how to play (see Yoshida, pp. 389-90 for a possible explanation of this).

In 2004, however, an almost intact leaf (BH1-19) of a similar document was acquired by the National Library of China. Published in 2008 (Zhang Zhan and Shi Guang), it appears to be the initial page of possibly the same letter and gives a more detailed historical context by referring to the defeat of the Tibetans at Kashgar which happened around 790.

The letter (translated by Zhang Zhan in Hansen, pp. 381-2) is from a Persian speaking Jew of Khotan to the ‘lord master’ Nisi Chilag, Abu Sahak and others on the subject of sheep trading. It lists bribes to officials, arranged no doubt in order of sociological importance and headed by a local ruler (dihgān) who can perhaps be identified with the King of Khotan or someone of equal status (Yoshida, p. 392). The gifts include a vase, scent, silk cloth, raw silk, sugar and other items which are not yet fully understood. Perhaps the most important information was the news from Kashgar that “They killed and captured all the Tibetans”. The writer himself contributed “a sum worth 100 strings of coins, or 100,000 coins” for the war effort.

As demonstrated by the montage above, the two documents are almost certainly part of the same letter with the National Library fragment forming the opening page and the British Library fragment a subsequent folio. From a morphological, palaeographical, and content-wise point of view we can be fairly certain that both were written by the same Judaeo-Persian trader. The author is identified in the second letter as ‘Sogdian,’ and despite being written in Persian, Yutaka Yoshida has convincingly argued on the basis of various sogdianisms in the letter itself that he was most likely a Persian speaking Sogdian Jew (Yoshida, pp. 390-92).

Taking both parts together the Dandan-Uiliq letter is probably the oldest surviving document of substance to be written in early New Persian, marking the first phase of the Persian language after the Islamic conquest. As such it provides important evidence for the development of the Persian language in addition to documenting the history of eighth-century Khotan.

(Via A Comprehensive Edition of Tocharian Manuscripts’ Facebook post.) And a happy Hanukkah to those of my readers who celebrate it!


  1. I am happy to see that that they specify that what they mean by “Judaeo-Persian” is merely “Persian written in Hebrew script.”

  2. I have been taught that Aramaic and its script had a long history of support and use in early Persian empires, and that the familiar Hebrew script was derived from Aramaic models.

    Is it by evolved differences in Hebrew script, or by the topics in the letter itself, that we know that the writer was Jewish?

    All those Nestorians wandering around central Asia, did they use Greek or Aramaic or don’t we know?

  3. David Marjanović says

    AFAIK, they used Aramaic, but in the Syriac script.

  4. ə de vivre says

    By the time the Church of the East, or the Nestorians, appear, the language they use is usually called Syriac rather than Aramaic. In fact, Syriac is often defined as the Eastern Aramaic dialect from Edessa that was adopted as the vehicle for Christianity by the Aramaic-speaking world. In the intervening time between the Achaemenids and the rise of Christianity, the Syriac script has become pretty far removed from the Aramaic script that gave rise to the Hebrew one.

    The Church of the East used their own flavour of the Syriac script, which you can see on Wikipedia if you’ve got the fonts, but the differences between the different styles of Syriac writing were less categorical than contemporary taxonomies make it seem.

    That said, while Syriac was the liturgical language, it wasn’t necessarily the vernacular of Eastern missionaries. Syriac was also adapted to write Soghdian, where it existed alongside the “native” Soghdian script that was itself derived from the earlier Imperial Aramaic writing that had spread east with the Achaemenids.

  5. By Jewish languages they [sc. the editors of Journal of Jewish Languages] mean ‘languages spoken by Jews’.

    (Boy, a sentence in italics with two NPs in roman, used in different senses.)

  6. The Hebrew script is clearly distinct from the other Aramaic-derived scripts that would have been in use at the time, and Aurel Stein instantly recognized the letters as being in a cursive version of this script.

    The Pahlavi Psalter from the 6th or 7th century found in nearby Turfan records a Middle Persian text from as early as the 4th century, a translation from Syriac, in a script that has come to be called Psalter Pahlavi. You can see it here. This and other versions of Aramaic-derived scripts used for writing Middle Iranian languages, collectively known as Pahlavi scripts, look very different from the Hebrew script (and sometimes from each other).

    Nothing in the translation of the letter seems to identify the writer clearly as Jewish, so the inference seems to be based solely on the choice of script.

  7. Indeed, the letter forms are more characteristic of Hebrew than any other Aramaic-derived abjad that I can remember seeing. (The next closest is maybe Palmyrene.) If you only compare it to scripts that are known to have been in use at the time the letter was written, I don’t think any other script is even close.

  8. I’m only familiar with the Palmyrene alphabet in the form of inscriptions, as on the monuments you can see in the Louvre and other places, so the similarities with the Hebrew alphabet written in ink with its distinctive serifs in letters such as the aleph were not as obvious to me. Apparently there are also fragments of the Palmyrene alphabet written on perishable materials, but I haven’t seen them. In my search for examples, however, I’m finding some Palmyrene inscriptions that almost look like they have rudimentary serifs and look more like Hebrew than the ones I remember.

Speak Your Mind