IRANIAN SOURCES ON THE KHAZARS
Dan D.Y. Shapira
1.0 In the following, three groups of sources dealing with Khazar his-
tory are surveyed. What is common to these sources is their Iranian prov-
enance. The sources can be divided into three groups according to the
three languages in which they are recorded: Middle Persian (or Pahlavi);
Arabic; New Persian. This division reflects not only the language of the
texts, but also, to some extent, their historical value as sources belonging
to a certain cultural and temporal milieu. Some of the sources in Arabic
are supposed to reflect Pahlavi originals now lost, while those in New
Persian, in almost all cases, go back to Arabic originals (some of them
being reworking of older Pahlavi material). There are sources in other
Iranian languages, such as Sogdian or Saka-Khotanese, but they contain
data on Ancient Turks and have no direct bearing on Khazar studies in
the strict sense.
1.1 The Khazars are still believed by some scholars to have appeared
on the historical stage first in the last decades of the Sasanian dynasty in
Persia, as an ally of Iran’s foes.1 Shortly after Heraclius delivered a mor-
tal blow to Iran, new enemies sprang up from the Arabian desert and
conquered Iran within a few years. Zoroastrian Iran thus ceased to exist,
and a new period in the long history of this country began, a period
which is characterized by an adaptation of Islam to the patterns of Ira-
nian civilization, and by a profound Iranization of Islam as a whole. The
first centuries of Islamic rule in Iran brought about interesting devel-
opments in pre-Islamic Iranian literary traditions, developments which
are generally seen within the framework of the Šuūbīya trends; we still
cannot appreciate with precision the very complicated mechanism of
these cultural changes, but in a schematized way, they can be described
as a double-track development: on the one hand, absorption of Iranian
1
However, one should stress that during this period, Khazars cannot be yet discern-
ible from other Western Turks, the so-called Türküts. Compare now C. Zuckerman’s
contribution in this volume; cf. also my own second contribution here.
292 dan d.y. shapira
traditions into the so-called “Islamic lore”, i.e., into the newly-emerg-
ing civilization united by Arabic language, Arabic script, and, to some
extent, by Islam; on the other hand, there was a parallel development,
namely, reshaping and consolidation of pre-Islamic Iranian traditions
within the community of faithful Zoroastrian believers, which became,
in the course of time, an oppressed minority in their own country.
1.2 It should be stressed that historical and geographical lore pre-
served in the traditional literary language of Sasanian Iran, Middle
Persian, was recorded in the form which came down to us only during
the first Islamic centuries; we may be quite certain that during the first
formative centuries of Islamic rule, Zoroastrians formulated anew their
self-awareness, so the historical and geographical material of our extant
Pahlavi sources, though valuable, should be treated with caution.2
1.3 The same applies also to the “Sasanian” material preserved in Ara-
bic sources; though authors writing in Arabic possessed an impressive
body of information about pre-Islamic Iran, most of which is generally
held to go back to the Sasanian Book of Kings (being thus translated into
Arabic3 in the early Islamic Period), it should be nevertheless remem-
bered that the date, origin and authenticity of such material should be
closely scrutinized in each case separately.
1.4 As to the sources in New Persian, they are much later than those in
Arabic, and as a rule, they rework earlier Arabic material.
2.1 To the first group of our sources, those in Zoroastrian Middle Per-
sian, belong a few interesting pieces of information. This is how Ayādgār
ī Jāmāspīg (12.8–9,15),4 a composition of (partly) geographical charac-
ter, describes Turkestān:
Turkestān wuzurg gyāg ud hamāg sard, wēšag bawēd, u-šān draxt ī bar-
war ud mēwag ī xvarišnīg ud *any ciš nihang. hast az awēšān kē Māh
2
Cf. now D. Shapira, “Was there Geographical Science in Sasanian Iran?,” Acta Ori-
entalia Academiae Scientarum Hungaricae, 54 (2–3), 2001, pp. 319–338; D. Shapira,
“Zoroastrian Sources on Black People,” Arabica, XLIX, 1 (Paris 2002), pp. 117–122.
3
In Arabic, Siyar-al-mulūk.
4
Cf. G. Messina, Libro apocalittico persiano Ayātkār ī Žāmāspīk, Biblica et Orientalia
9, Roma 1939.
