Acta Orientalia Academiae Scientiarum Hung.

Volume 60 (2), 161 – 194 (2007)

DOI: 10.1556/AOrient.60.2007.2.2

IRANO-TURCICA:
THE KHAZAR SACRAL KINGSHIP REVISITED*
PETER B. GOLDEN

Department of History, Rutgers University

Conklin Hall, 175 University Avenue, Newark, NJ 07102, USA
e-mail: pgolden@andromeda.rutgers.edu

The ruling house and core tribes of the Khazar empire did not share the same tribal or, in many
instances, ethnic origins as those of the Qağanate’s diverse subject population. The Khazar rulers
were heirs of the Türk qağanal charisma. Although aspects of sacral rule and dual kingship can be
seen in the Türk and other Inner Asian nomad-based empires, it was only in Khazaria that the Qağan
became a sacralised, tabuised figure. This transformation occurred in the 9th century and may
reflect the influence of the Ors, the Khwârazmian-Iranian guard of the Qağan and the chief minister
drawn from their ranks.
Key words: Khazars, sacral kingship, royal charisma, Ors, Khwârazmian guard corps.

As was typical of a number of polities in mediaeval Eurasia, the Khazar ruling

qağanal house and its core tribes did not share the same tribal or, in many instances,
ethnic origins as those of the Qağanate’s diverse subject population. Charismatic
ruling houses in much of Eurasia provided a set of royal traditions, origin myths and
ideologies of heavenly mandated rule which could, in the proper circumstances, serve
as the framework to create a “people”. This was true of the Latino-Celto-Germanic
world of the early Middle Ages (Geary 2002, pp. 74–78) (or at least that is argued by
the followers of the Traditionskern school in Gillett 2002) and it also appears to be
true of the steppelands to the East where we find “charismatic” ruling clans and their
immediate clanal or tribal entourages which become the Traditionsträger and foun-
ders of polities. Here too, as György Györffy notes, rulers frequently “constituted a
supra-national organisation”. In “nomadic empires, rulers often took wives from other
countries and surrounded themselves with escorts of foreign bodyguards, scribes and

*
An earlier and much briefer version of this paper was given at “Pre-Modern Russia and its
World. A Symposium Honoring the Work of Thomas S. Noonan”, University of Minnesota, Nov.
1 – 2, 2002 and was published as Golden (2006).
0001-6446 / $ 20.00 © 2007 Akadémiai Kiadó, Budapest
162 P. B. GOLDEN

artisans” (Györffy 1994, p. 87). The Khazar Qağans, surrounded by their wives and
concubines coming from the subject population, and Khwârazmian guard corps (the
Ors) (Ibn Faḍlân1939, Arabic text, p. 44, Germ. trans., p. 100; al-Mas‘ûdî 1966, I, pp.
213–215; see also Golden 1990, pp. 39–46 and Golden 2001, pp. 153–170), nicely
illustrate this phenomenon. However, contemporaries were struck by many unusual
elements of Khazar socio-political structure. Al-Mas‘ûdî comments, for example, that
“among the kings of the East in this region, there is no one who has hired soldiers
(junûd murtazîqa) except for the king of the Khazars” (al-Mas‘ûdî 1966, I, p. 214).
The Khazar ruler was one of the great, larger than life, figures of his age.
According to the Fârsnâma of Ibn al-Balxî (12th century, but reflecting older pre-
Islamic Iranian traditions), the Sâsânid Shah, Xusraw Anûširvân (531–578) kept
three golden thrones on the right and left sides of his throne, reserved for the kings of
China, Byzantium and the Khazars, monarchs on his level, should they come to visit
him (Ibn al-Balkhî 1921, p. 97; see also Christensen 1971, pp. 411–412). This was
an old notion among the ancient empires of Eurasia: the community of heavenly-
mandated rulers. The Chinese sources make reference to the four sons of Heaven: the
crowned heads of China, India, Rome, and the Yuezhi (Konow 1929, pp. 162–163).
The latter represented the “nomadic northlands”.

The Khazar Sacral Kingship

Among the most striking institutions of the Khazar Qağanate was a form of dual
kingship with a fully articulated sacral ruler – although it might be noted that sacral
kingship in one form or another (i.e. rulers claiming divine origins or a heavenly man-
date) was the monarchic norm across Eurasia for much of recorded history (Al-Az-
meh 2004, p. 10). Nonetheless, the Khazar Qağanate had some unique features. Our
most explicit descriptions of the Khazar sacral Qağanate are found in the Muslim
sources dating to the late 9th–11th centuries, but largely based on slightly earlier
accounts as well as contemporary information.1
Al-Ya‘qûbî (writing ca. 891, d. 897), in his Ta’rîx (“History”) has a section on
the kingdoms of the North, descended from the sons of Japheth and of “Tâġarmâ.”

1
The earliest notices come from a group of sources that are derived from the works of al-
Jaihânî, a vezir of probable Khwârazmian origin in the service of the Sâmânids in the early 10th
century. His works, in particular the Kitâb Masâlik al-Mamâlik (ca. 900), have not survived but are
reflected in the geographical and historical writings of his contemporary, Ibn Rusta (d. 913), in the
work of the anonymous author of the Ḥudûd al-‘Âlam (987), and in the histories of Gardîzî (mid-
11th century), al-Bakrî (d. 1094), and al-Marwazî (d. sometime after 1120). The latter has a rather
truncated account that does not touch on the theme of interest to us. Al-Jaihânî’s scholarship and
patronage extended to Abû Zaid Balxî, the founder of the “classical school” of Muslim geographers
and others from whom a number of valuable notices on our theme stem. As Khwârazm, an oasis
city-state, was deeply involved in trade with the steppe peoples, much of the information gathered
by al-Jaihânî must have come directly from Khwârazmian merchants, see Kračkovskij (1955 –
1960, pp. 219ff.); Zaxoder (1962, I, pp. 46 – 49); Kmoskó (1997, pp. 49 – 54); and the very thorough
study of Göckenjan – Zimonyi (2001).
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 163

He notes a series of peoples: “The Burjân, Dailam, Babr, Ṭaylasân, Jilân, Fîlân, Alans
(al-Lân), the Khazars, Dûdâniyya and Arman. The Khazars conquered the people of
the country of Armenia. Over them there is a king called Xâqân and he has a deputy
called *Yazır Bulaš”. (The mss. have: all in variants without dots, most
probably for Yazır Bulaš (‫ ) ﻳﺰﻳﺮ ﺑﻼش‬or ‫ ﻳﻮﻻش‬Yulaš etc. (Al-Ya‘qûbî 1883, I, pp.
203–204).)2
The dating of the Khazar conquest of Armenia (more probably a major raid) is
uncertain. Al-Ya‘qûbî goes on to note that the areas conquered by the Khazars were
“Fourth Armenia”, which was taken by the Sâsânid Shah Qubâd\Kawâd (488–531).
He then describes other Persian conquests in Daghistan and concludes with the Byz-
antine defeat of the Persians and their appointment of a king for the region. Dunlop
(1954, pp. 20–22) was prepared to date the Khazar conquest and the existence of the
dual kingship to the era of Kawâd/Qubâd, but then noted the difficulties in positing a
mature Khazar Qağanate with a “deputy” in an era even before the foundation of the
Türk Empire in 552. Dunlop was also prepared to see the Khazars “on the scene” by
the time of Qubâd and Anûširwân (531–579), but emerging as a Qağanate only after
the decline of the Western Türks. He suspected that al-Ya‘qûbî’s source was Hišâm
al-Kalbî (d. 204/819) who received his information from his father (d. 146/763), thus
placing the data to the mid-8th century. This may be so, but it seems equally possible
that al-Ya‘qûbî updated the information he took from Hišâm al-Kalbî or the latter
himself added this notice about rule among the Khazars by a Qağan and “his deputy”
bringing his information into line with what was known in his time. Moreover,
“deputy” does not necessarily imply a co-ruler. The notice, then, could well refer to
any one of a number of Khazar raids in Transcaucasia in the 8th century with ad-
ditional information on Khazar governance updated to either al-Ya‘qûbî’s or his
source’s era (9th century). It is even more likely that this “conquest” of Armenia
sometime during the era of Qubâd was carried out by some other “Hunnic” peoples
anachronistically now called Khazars by our source who adds the information about
a deputy ruler, a Khazar institution becoming more familiar to Arab readers of his-
torical-geographical literature.
Ibn Rusta (d. 913): “They [the Khazars, PBG] have a king who is called îšâ3,
the greater king, however, is the Khazar xâqân. (But,) he does not have the obedience
(ṭâ‘a) of the Khazars except in name. The (full) extent of the (management of the)
affairs of state are upon the îšâ since, as concerns leadership and the armies, he is in
a position in which he does not heed anyone who is above him. Their greater king is
an adherent of Judaism as is also the îšâ and those from among the leaders and great
ones who sympathise with his inclinations (yamîlu maylahu). The rest of them profess

2
His History goes up to 873. For mss. forms, see Golden (1980, I, pp. 217 – 218).
3
Correctly: Išâd; on this form, see Golden (2005, pp. 212–213). Išâd is a Khazar form of
Ixšad, a high title of Iranian (most probably of Soġdian or Khwârazmian origin, cf. Soġd. ’ġšyd,
ixšâd usually Arabised as Îxšîd/Îxšîd (Ibn Xurdâdhbih, malik Farġâna ‫’[ اﺧﺸﻴﺪ‬xšîd] or ‫ اﺧﺸﺎذ‬Îxšâd)
and related to Türk Šad. These stem from Old Pers. Xšâyaθrîya/xšâita, Avest. xšâêta > Middle
Pers. šâh or cf. also Saka Šao. See also Bombaci (1974), but how xšâyaθiya produced Türk šad
remains unexplained. Clauson (1972, p. 866; 1975, p. 45).
Acta Orient. Hung. 60, 2007
164 P. B. GOLDEN

a faith similar to that of the Turks.” (Ibn Rusta 1892, p. 139.)4 Further on, he notes:
“Their king, the Išâ imposes on the people of power and affluence among them (the
provision) of horsemen according to the amount of their properties (qad waẓẓafa ma-
likuhum Išâ ‘alâ ahli quwwati wa’l-yasâri minhum fursânan ‘alâ qadri amwâlihim)
and the extent (ittisâ‘) of their circumstances (aḥwalihim) with regard to income
(ma‘âš). They raid the Bajânâk (Pechenegs) every year [or: throughout the year].
This Išâ takes upon himself the conduct of the expedition (al-xurûj5) and goes forth
on his razzias with his soldiers. They have a gracious outward appearance. When
they go forth in any direction, they go forth completely armed, in full gear (miḥallât),
(with) flags (a‘lâm), spears (ṭirrâdât) and sturdy coats of mail (jawâšin muḥkama).
His horsemen number 10,000 riders, of whom (there are) those who are bound to
them by wages (mimman huwa murtabiṭun ujriya ‘alaihim6) and among them (there
are those) whom he has imposed ([qad waẓẓafa] as an obligation) on the wealthy.
When they go forth in any direction, something like a sun-shaped disk (šamsa7) is
prepared after the fashion of a tambourine/drum (‘alâ ṣan‘at ad-duff) and a horseman
carries it in his hands, going in front of him.8 He goes and his soldiers follow him,
catching sight of the light of this sun-disk. When they take war booty, they gather

4
See also Dunlop (1954, pp. 104 – 105) for a slightly different rendering; German trans. in
Göckenjan – Zimonyi (2001, pp. 52 – 53).
5
This is the widely shared view of this sentence, e.g. G. Wiet’s translation (1955, p. 157):
“Cet Isha commande en person ces expeditions”; that of Lewicki (1956 – 1988, II/2, pp. 28/29): “Sam
Κâ zajmuje się organizacią wyprawy i udaję się na swe ekspedycje wojenne wraz ze swymi wojow-
nikami.” Göckenjan – Zimonyi (2001, p. 54, n. 21), however, noting that the mss. and Gardîzî have
xarâj, emended here by de Goeje to xurûj, remark that “diese sonst plausible Emendation wird je-
doch von G nicht bestätigt”. Hence, they render the passage as “Dieser Κâ zieht in eigener Person
die Steurn ein und rückt bei seinen Streifzügen mit seinen Heeren aus”.
6
Göckenjan – Zimonyi (2001, p. 54) “mit solchen, denen er einen festen Sold gezahlt hat,
und denen, die er unter den Reichen ausgehoben hat”. Turk. trans. p. 292: İşâ, muhârebeye defterde
mukayyed ve erzakı hükûmet tarafından verilen on bin süvâri ile çıkardı [mimman huwa murta-
biṭun ajrâ ‘alaihim / mimmen huve murtabitun ecrâ alayhim] : kendilerin vazife verilmek sûretiyle
askerliğe bağlanmış olanlarda or askerliğe bağlı olup kendilerine askerlik alıştırılmış.
7
Dozy (1881, I, p. 786): šamsa “ornament rond, petite boule en forme de soleil”; Lewicki
(1956 – 1988, II/2, pp. 64 – 65), referencing Xvol’son (1869, p. 689), also notes the report from the
8th-century Armenian historian Łewond about the victory of Zaid ben Amru el-Xaraši (Sa‘îd b.
‘Amr al-Ḥarašî over the Khazars in 731 (noted also by al-Ya‘qûbî and al-Baladurî) in which he
says that the Arabs captured a copper emblem/standard with an image which the Arabs have kept
up to this day. Xvol’son identified the šamsa with this, suggesting that it was like the standards
used by various Turkic tribes. Lewicki adds “the existence of such an object and its use as a
military standard can be viewed as (as symbol) of the cult of the moon or of the sun which is noted
among the Turkic beliefs of the Khazars (można tłumaczyć kultem księżyca czy słońca wystę-
pującym w tureckich wierzeniach Chazarów). According to the sources, the leader of the Huns [he
has the Xiongnu in mind here, PBG] came out from his camp every morning to bow before the
rising sun and in the evening bowed to the moon. The Huns [Xiongnu, PBG] also had the custom of
going out on campaign at the time of a full moon (see Bičurin 1950, I, p. 50). The brightly flashing
disk (Owa błyszcząca tarcza) of the Khazar leader could, then, symbolise the full moon, which is
propitious for military expeditions.”
8
Göckenjan – Zimonyi (2001, p. 54): “so wird ihm ein schirmähnliches Symbol vorgetra-
gen, das wie eine Hantrommel aussieht”.
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 165

this booty, all of it, in his army camp (mu‘askar). Then, the Išâ selects what he likes
and takes it for himself and he releases the rest of the booty to them to be divided
among them.” (Ibn Rusta 1892, pp. 139–140.)9
The Ḥudûd al-‘Âlam (anonymous author, 987): “…the king, who is called
Ṭarxân Xâqân and is one of the descendants of Ansâ (?). He lives with all his troops
in the western half of the town which possesses a wall. In the other half live Muslims
and idol-worshippers. This king has in this town seven governors belonging to seven
different creeds. At any hour when a more important litigation arises, they ask the king
for instructions, or inform him of the decision (taken) on that litigation.” (Ḥudûd /Su-
tûdah 1340/1962, p. 193; Ḥudûd /Minorsky 1970, pp. 161–162.)
Ânsâ [‫ ]ﺁﻧﺴﺎ‬may be a corruption of Ashina [‫]ﺁﺷﻴﻨﺎ‬, the royal house of the Türks
– although this remains a matter of contention.
Gardîzî (mid-11th century): “They have a king, his name/title (nâm) is *Κâd.10
There is a great king. They call the great king Xazar Xâqân. He has the title and that
is all (wa bas). As for the central (control) of all the affairs of state and the entourage
(ḥašam), this is incumbent on the *Κâd. There is no one greater than the Κâd. Their
greater chief is a Jew and the Κâd is also a Jew. And whoever is inclined to (these)
two, from among the commanders and great ones, is likewise. The rest profess a
religion that resembles the religion of the Ġuz Turks.” (Gardîzî/Ḥabîbî, p. 580)11
With reference to the army, he has more details than Ibn Rusta: “Every year they go
on campaign (B: bi-ḥarb, Ḥ: bi-ġazw on raids) in the land of the Pečenegs. From
there they bring back goods (mâl, perhaps “livestock”) and captives (barda). This
Κâd takes (carries away, mi-sitânad) his own taxes (xarâj-i xûd ) and dispenses (tafri-
qa kunad) [them] to the soldiers. It happens that sometimes (Ḥ: gâh) they also make
war on the Ġuzz (Oğuz) and Burdâs [ġuzz wa burdâs or ġazw-i burdâs “war with the
Burdâs]. They have banners, spears (B: ṭirrâdhâ, Ḥ: ṭarâwahâ “pennants”) and sturdy
coats of mail and good armor. When the king of the Khazars sits upon (bar-našînad
‫ ) ﺑﺮ ﻧﺸﻴﻨﺪ‬his horse to go forth, 10,000 cavalrymen rise out with him. Of them, some
receive salaries (puštagânî-xwâr “stipend,” B: pîškânî-xwâr, Ḥ: bîstagânî “monthly
pension, wages, pay of an army”)12 and some are from the military following (B:
waḍî‘at, Ḥ: waḍ‘iyat) of the powerful (taken as tribute to the king) who go out with
the king with their own arms and military apparatus (bi-sâz u âlat-i xwîš). When they
lead out the army and go somewhere, they leave at home numerous troops (laškar-i
anbûh) in order to guard (az bahr-i nigâh dâštan) (their) families and goods (buna).

19
See also Dunlop (1954, pp. 104 – 105) for a slightly different rendering; German trans. in
Göckenjan – Zimonyi (2001, pp. 52 – 53).
10
In the most recent edition, Gardîzî/Ḥabîbî (1984, p. 580) this is mistakenly given as ‫اﻟﺸﺎد‬
(’lšâd) for ‫’( اﻳﺸﺎد‬yšâd). The latter reading is clear from the two extant mss., see Golden (1980, I, pp.
206 – 207).
11
See also Pers. text and Russ. trans. in Gardîzî / Barthold (1963 – 1977): Pers. p. 36, Russ.
p. 57; Gardîzî / Martinez (1982, pp.152 – 153); Göckenjan – Zimonyi (2001, p. 166).
12
Göckenjan – Zimonyi (2001, p. 167): “Einige von diesen sind Söldner, wahrend einige
(andere) Gefolgsleute der Grossen sind, die mit dem König in eigener Rüstung ausziehen,” here
following Martinez, 154 “Some of these are salaried [troops], while some (others) are from the
retainers and clients of the wealthy (men az jehat o ṣanî‘at-e tavângarân).”
Acta Orient. Hung. 60, 2007
166 P. B. GOLDEN

They have advance guards (ṭalâ‘i) that (always) go before (in front of) the army.
They (always) carry in front of the king candles (šam‘ahâ13) and torches (nafâṭhâ)
made of wax (Ḥ: p. 582) so that by its light (ba-rowšanâ’î-yi ân) he goes with the
army. When they take war booty, they collect it all (B: hama gird kunand) in the army
camp. First their commander removes (bar-dârad) whatever of the war booty he
wants. The rest he divides among the soldiers (Gardîzî/Ḥabîbî 1984, pp. 581–582).14
Al-Bakrî (1094): “Their king does not have their obedience (ṭâ‘a) except in
name. The direction of (state) affairs is in the hands of the ‫( إﻳﺮان ﺷﺎﻩ‬Îrân šâh, obvi-
ously a garbling of Κâd, PBG). He is the one who commands their armies and he
possesses their obedience…They bring their most obscure affairs to the attention of
their great one who is called Xâqân Xazar. He is greater in power among them than
the king.” (Al-Bakrî 1992, pp. 447–448; Göckenjan – Zimonyi 2001, pp. 226, 227.)
Another cluster of information stems from the school of al-Balxî (850–934)
(Kračkovskij 1955–1960, I, pp. 194ff.; Zaxoder 1962, I, pp. 49–51; Kmoskó 1997,
pp. 73–74), preserved in the accounts of al-Iṣṭaxrî (mid-10th century) and Ibn Ḥaw-
qal (wrote ca. 977) who closely follows him. These notices are much more detailed.
Al-Iṣṭaxrî: “The king lives in the western part of [the capital]. The king is
called in their tongue beg (‫ )ﺑﻚ‬and he is also called *yilig (text: ‫ ﺑﺎك‬bâk, recte: ‫ ﻳﻠﻚ‬ylk
= yilig: yelig “prince”15) … As concerns their governance and affairs of the kingdom
among them, their supreme leader (‘aẓîmuhum) is called “Xâqân Xazar”. He is greater
than the king of the Khazars except that the king of the Khazars is the one who
installs him (in office). When they want to install this “xâqân”, they bring him to him
[the king] and choke him with a silk (cord) until his breath is nearly cut off. They say
to him ‘for how long a period do you wish (to have) the kingship?’ And he says so
many and so many years. If he dies before then (there is no problem); he is not killed
unless he reaches that year (that he named). The Qağanate is not permitted (to any)
among them except for the members of a house of notables” [ahl bait ma‘rûfîn]. He
does not have the power to command or forbid anything. Nonetheless, he is held in
esteem and they prostrate themselves before him when they go in to him. No one
comes near him except for a group that goes (to him) such as the king and those of
his stratum (in society). The king does not go in to him, except for a (special) event
(ḥâdita). When he goes in to him, he rolls in the dust, prostrates himself and stands at
a distance until he is permitted to draw near. When great difficulties (ḥazb aẓîm)
befall them, the “Xâqân” is brought out into the open by him. None of the Turks and
those of the different groups of unbelievers who are near them, sees him without de-
parting. They do not fight with him out of respect for him. When he dies and is buried,
no one passes his tomb without dismounting and prostrating himself before him. He
does not mount (again) until his tomb is lost from sight. Their obedience to their king

13
The confusion with šamsa.
14
See also Pers. text and Russ. trans. in Gardîzî / Barthold (1963 – 1977), Pers. p. 36, Russ.
p. 57; Gardîzî / Martinez (1982, p. 154); Göckenjan – Zimonyi (2001, pp. 167 – 168).
15
Cf. the Old Hungarian *yeleġ (’Ie/lex), see discussion in Golden (1975, pp. 37 – 43). This
is from Türk éllig “(junior) king, prince” (cf. Turk. éllig, élig “having a realm, king, ruler, master”
< él “realm”, Clauson (1972, pp. 141 – 142); Erdal (2004, p. 51)).
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 167

is to such a degree that when it is required that one of them, perhaps, is to be killed
and he is from among their great ones and the king does not want to kill him in a
public fashion (ẓâhiran), he orders him to kill himself. He departs for his home and
kills himself. The Qağanate is within a group of relatives (qaum). They do not have
royal power or wealth. When the leadership falls to one of them, they award it to him
without paying attention to his state.” (Al-Iṣṭaxrî 1927, pp. 220, 224; see also Dunlop
1954, pp. 91, 97–98.)
Al-Iṣṭaxrî goes on to note that the Qağan and only the Qağan had a golden
throne and that his domiciles, whether his tent when he went forth or his home in
Atıl/İtil, were always higher than that of the king (al-Iṣṭaxrî 1927, pp. 224–225;
Dunlop 1954, p. 98). He also notes the presence of seven religious judges, two each
for the Jews, Muslims and Christians and one for the pagans (al-Iṣṭaxrî 1927, p. 221;
Dunlop 1954, p. 93). Ibn Ḥawqal’s account is basically the same (Ibn Ḥawqal 1992,
pp. 330, 334–335).
With Ibn Faḍlân we have a report from a man who journeyed to Volga Bulğaria
as part of a caliphal embassy to that uneasy vassal of the Khazar Qağans in 921–922.
His lengthy account appears to be based on communications he received from local
informants about Khazar customs and governance and contains unique information.
Ibn Faḍlân: “As concerns the king of the Khazars, his name is Xâqân. He does
not appear (in public) except for one time every four months for a promenade (muta-
nazzahan16) [apart from the masses]. He is called the Great Xâqân and his deputy
(xalîfatuhu) is called the Xâqân Bah [Bäh].17 He is the one who commands the armies
and governs. He directs the affairs of state and manages it. He makes public appear-
ances and goes on military campaigns. The kings who neighbor him submit to him.
Every day, he enters into the presence of the greater Xâqân, humbly, showing humil-
ity and devout tranquility (sakîna). He only enters his presence barefoot and has
firewood in his hand. Having greeted him, he ignites the firewood in his hands.
When the fuel is all consumed, he sits with the king, on his throne, on his right side.
He has a deputy who is called *Kündü Xâqân [text ‫ آﻨﺪر ﺧﺎﻗﺎن‬recte: ‫ آﻨﺪو ﺧﺎﻗﺎن‬18].
This man also has a deputy, a man called jâwšîġr [‫]ﺟﺎوﺷﻴﻐﺮ‬.19 The customary
practice of the greater king is that he does not sit (in meetings) with the people; he

16
The verb tanazzaha “to be far from something, to go for a walk, to promenade” < nazuha
“to be far from, be untouched, unblemished, steer clear of…” (see Wehr 1994, p. 1125) was under-
stood in the latter sense by Togan in his rendering of the passage, see Ibn Faḍlân (1939, p. 98)
“entfernt (von den Massen)”. It rather denotes the idea of a ceremonial progression by a figure who
remains remote from the mass of people.
17
This is the Qağan Beg. Ibn Faḍlân’s ‫ ﺑﻪ‬bäh = Khazar beh < bex (cf. the πέχ of Constan-
tine Porphyrogenitus, see below) < beġ < beg. The beġ form may have entered Hungarian, beġ >
beü, böü > bő “full, rich”, see TESz (I, pp. 356 – 357) and Golden (2005, pp. 209 – 210).
18
This is the title kündü, found among the Hungarians as kende, on this title see Ligeti
(1986, pp. 49, 254, 368, 482, 484), the title of the Hungarian sacral king. The mss. all have kundur,
see Golden (1980, I, pp. 200 – 202; 2005, pp. 213 – 214).
19
The reading of this title is unclear. Klyashtorny (1997, pp. 22 – 23), suggests: ‫ﺟﺎوﺷﻨﻐﺮ‬: jav
+ šunğar = čav (Old Turk. čavlı “falcon, hunting bird”) and and šungqar “falcon, gerfalcon,” cf.
Qaraxanid Čavlı Beg “head of the royal falcon hunting”, a high dignitary in the Qaraxanid state.
Acta Orient. Hung. 60, 2007
168 P. B. GOLDEN

does not speak with them and no one enters his presence except for those whom we
have mentioned. Sovereignty in the exercise of governmental power and the imple-
mentation of punishments and the direction of the kingdom rests with his deputy, the
Xâqân Bäh. It is the customary practice of the greater king that when he dies a great
abode is built for him. In it are twenty cells20 and a tomb is dug in each of the cells.
The stones are broken until they become like pulverised kohl (antimony) and are
spread out over it and lime is thrown over that. Under the abode is a river. It is a
great river which runs (under it). They direct the river over that tomb and they say
that it does not allow the Devil to draw near nor does it allow a man or a maggot or
insects (to come near). When he is buried, they cut the necks of those who buried
him so that no one will know where his tomb is among those twenty cells. They call
his tomb “Paradise”. They say he has entered Paradise. They spread brocade sewn
with gold over all the cells. It is the customary practice of the king of the Khazars to
have twenty-five wives. Each wife is the daughter of the kings whom he has cap-
tured, taking them either voluntarily or under duress…When this king goes forth on
horseback, the rest of the army rides out on horseback with him. They maintain a dis-
tance of one mile between him and the riders. None of his subjects sees him without
dropping to the ground to prostrate himself in his presence and does not raise his
head until he has passed by. The length of their kingship is forty years. If he goes
past that by one day, his subjects and entourage kill him. They say “his mind has be-
come diminished and his good sense has become confused” (Ibn Faḍlân 1939, Arabic,
pp. 43–44; German pp. 98–101; see also Dunlop 1954, pp. 111–113).
Ibn Faḍlân adds further that he has the power of life and death over his troops
(Ibn Faḍlân 1939, Arabic, p. 44; German p. 101).
Al-Mas‘ûdî, writing in the 930s provides further details. After a discussion of
the ‫[ اﻻﺳﻴﺔ‬al-Orsiyya = Ors], the Khwârazmian guard/comitatus of the Khazar king,
he notes that “king of the Khazars has complete trust in them in his wars”.21 They
have special “arrangements” (šurût) with the Khazar rulers, one of which is the pro-
vision that “the vezirate be from among them. The vezir, in our time, is one of them.
He is Aḥmad b. Kûya (or Kûba)” (al-Mas‘ûdî I, p. 213). Al-Mas‘ûdî then adds: “Our
report was not about the king of the Khazars, (rather), we mean here the Xâqân. That
is because in the Khazar kingdom there is a Xâqân. It is his customary practice that
he is in the hands (i.e. under the authority, PBG) of a king, other than him, (living) in
his abode [dâr] (that of the king, PBG). And the Xâqân is inside a castle. He is not
allowed to ride forth nor to appear before the people of distinction (al-xâṣṣa) nor the
common folk; nor is he allowed to go out from his residence. With him (there in the
king’s house, pbg) is his personal household (ḥaramuhu). He does not give orders,
nor does he proscribe nor manage anything of the affairs of the kingdom in any way.
The royal authority of the Khazars is not in order [lâ yastaqîmu] for their king unless
the Xâqân is with him in the capital of his kingdom, with him in his castle. When the
land (arḍ) of the Khazars suffers from drought or a misfortune befalls their country

20
bait “house, domicile” etc., here “cell”, see Dozy (1881, I, p. 131) bait “alvéole”.
21
Lit. “upon them is the dependance (mu‘awwal) of the King of the Khazars in his wars”.
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 169

or they face a war against other nations or some other matter comes upon them, the
common folk and the people of distinction rush to the king of the Khazars and say to
him ‘we have seen a bad omen (taṭayyaranâ) in this Xâqân and in his reign, we fore-
see (no good) from him. Kill him or hand him over to us so that we may kill him.’
Sometimes, he hands him over to them and they kill him. Sometimes, he takes (the
task) of killing him upon himself. Sometimes, he shows him mercy, defends him (say-
ing) he is free of offense [jurm], he justifies him [istaḥaqqahu] (saying) there is no
crime [danb] which he has committed. This is the customary practice of the Khazars
at this time. I have no knowledge whether this was the case in olden times or if it is
an innovation. The dignity (manṣib) of their Qağanal office belongs to the members
of a family [ahl bait] from among their notables. I am of the opinion that the king-
ship was among them in olden times, but God knows best” (al-Mas‘ûdî 1966–1979,
I, pp. 214–215).
There is a brief reference to the Khazar Qağanal office in the Hebrew letter of
an anonymous Khazar Jew (the so-called Schechter Text, from the Cairo Geniza) dat-
ing to the latter half of the 10th century (960s) on the very eve of the fall of Khazaria:
“The men of the land appointed over them one of the sages as judge (‫ שפט‬šôpêṭ).
They call him in the language of the Qazars, ‫( כגן‬kagan); for this reason the name
given to the judges who arose after him has been kagan until this day…” (see Golb –
Pritsak 1982, pp. 110–113).22
Here the Qağanal office is set within a Judaised context, in the aftermath of
the conversion23 (or “return” as the author puts it) of the Khazars to Judaism. Address-
ing a Jewish audience, our source transforms the Qağan into a judicial figure, per-
haps harkening back to the “judges” 24 of pre-monarchic Israel, a biblical reference
that might be more acceptable to Jewish norms and would be understood by his
learned readers (Petrukhin 2004, p. 272). The Response of King Joseph mentions ha
sar ha gadol “the great Ruler/Prince” whose permission was needed by the “king” to
bring about the conversion to Judaism.25 The Khazar King Joseph (it is unclear if he
is the Qağan or the Qağan-Beg), in his reponse, ca. 950–960, to a letter from Ḥasdai
b. Šaprûṭ, the Jewish courtier of the Spanish Umayyads, recounts the Khazar conver-
sion tale. According to it, Bulan, the lesser king (he is referred to as ‫ מלך‬melek
throughout), after an angelic visitation in a dream, asks the angel to appear to the
Great Ruler (‫ השר הגדול‬ha-sar ha-gadol) who then “gathered all his chiefs (‫)שריו‬26
and servitors and his whole people”. He then holds a religious debate between

22
The transcription of this title kgn or possible kgn clearly show that the author of the docu-
ment was reflecting either local, i.e. Khazarian (Qağan or Xağan), Byzantine (χαγάνος) or even
possibly Slavic (коган) usage and not Arabic (Xâqân).
23
Al-Mas‘ûdî (1966 – 1979, p. 212) dates it to the reign of Hârûn ar-Rašîd (787 – 809).
24
See D. Shapira (1998 – 1999, p. 236): the “judges” of pre-monarchical Israel who were
subsequently “overshadowed” by the kings. See also Kovalev (2005, p. 233).
25
Kokovcov (1932, Heb. p. 21/Russ. p. 76); see also Novosel’cev (1990, p. 136); Zucker-
man (1955, p. 252). Heb. sar has a variety of meanings, “prince, chief, ruler, minister”. Clearly, the
sense here is that of a superior ruler.
26
Here, sar is most probably referring to clan or tribal chiefs.
Acta Orient. Hung. 60, 2007
170 P. B. GOLDEN

Judaism, Islam and Christianity, from which Judaism emerges the victor.27 This
“Great Ruler” is undoubtedly the Qağan, who, at this stage (early 8th century) appears
to be playing a role in affairs (perhaps largely ceremonial), but it is the lesser “king,”
Bulan, who takes the initiative. It is the “Great Ruler” who convenes the other “chiefs”
and the whole people and then approves this new religious orientation. The conver-
sion narrative, as one would expect, presents this as entirely voluntary. Whether there
were elements of coercion involved, taken as axiomatic in a number of modern dis-
cussions of the conversion, is simply a matter of speculation and conjecture. Our
sources, including the Muslim accounts, give no hint of this.
We may summarise the principal features of the Khazar sacral kingship thus:
It is associated with the office of the Qağanate itself which is limited to a special,
charismatic clan. The Qağan reigns but does not rule. The day to day governance of
the state is in the care of another person (presumably representing another line). He is
termed the “king” (Arab. malik, Heb. melek) in our sources and bears the title
Išad /Beg [Beh]/Yilig [Yeliġ] and officiates at the investiture of the Qağan. By the
early 10th century (Ibn Faḍlân’s account), the Qağan and Qağan-Beg are assisted by
two other officials. The Qağan appears to be a descendant of the Ašina (see below),
the ruling clan of the Türk empire. Only members of this clan may become Qağans.
The Khazar rulers provide judges for each of the confessional groups of the state
(a point noted by most of our sources, the Ḥudûd, al-Iṣṭaxrî, Ibn Ḥawqal, al-Mas‘ûdî),
a clear indication of their “universal” rulership. The Qağan is first and foremost the
“law king.” He is superior to the “king” and has the ultimate power of life and death,
but is not to be physically associated with the actual shedding of blood. The Qağan is
enthroned in accordance with the Ašina tradition of ritual strangulation which high-
lights his shamanic powers. The Qağan does not appear to the people or even to the
royal entourage except on a few occasions (four times a year-changing of the sea-
sons?). When he does appear all must prostrate themselves before him. Even the king
when he does go in to have an audience with him, must undergo purification rites.
Human sacrifices are conducted at his burial. His tomb, an elaborate mausoleum, is
holy ground perpetually cleansed by water and requiring passers by to fall prostrate
before it. The Qağans are beyond wealth. The subject rulers all have to send daughters
to the harem of the Qağan. The king must have the Qağan in his residence in order to
legitimate his rule. The Qağan is a heavenly mandated intermediary between the
divine and his state and hence is a talisman for the good fortune of the state. His term
of office has temporal limits (forty years – or some catastrophe), after which his spiri-
tual power is considered diminished. He no longer has qut.28 The life and the death of

27
Kokovcov (1932, Heb. pp. 21ff., 28ff./Russ. pp. 75ff., 93ff.). See also Novosel’cev (1990,
p. 136); Zuckerman (1955, p. 252).
28
In the political context, Turkic qut denoted “heavenly good fortune, charisma” it was, in
essence, the heavenly mandate to rule. See Clauson (1972, p. 594). In general, the term has a wide
range of meanings in Turkic, “soul, life-force” (in this sense akin, perhaps, to Ancient Egyptian no-
tions of ka), spirit, fortune, good fortune, blessing, grace,” etc. In Mongol of the era of Činggis Xan
it was rendered by suu jali: suu (Lessing 1982, p. 740: “distinction, superior(ity), genius”), jali
(Lessing 1982, p. 1031: “flame, spirit”). The original meaning of the latter was “burning coal, fire.”
This element of fire and brightness was added to the notion of suu (the “genius” bestowed by
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 171

the sacral king have cosmic significance.29 The light-reflecting, tambourine or drum-
shaped, sun-disk-like object (šamsa), noted by Ibn Rusta, which is paraded before the
Išâ(d) when he sets forth on campaign may point to a solar cult as well. Solar disks
are well known as representations of the divinely bestowed power of the ruler (xwa-
rəna, farr etc.) in the Iranian world (see Soudavar 2003).30 If this solar disk does
indeed reflect such an emblem and is not merely a signaling device for indicating
where the Išâ(d) and hence army center is, the question may well be asked: why does
the Išâ(d), a non-sacral figure, have this emblem? There are several possibilities. Ibn
Rusta may have confused or conflated a practice actually associated with the Khazar
Qağan during his infrequent processionals from his chambers. Or, the disk-like solar
emblem was borne before the Išâ(d) as a symbol of the divinely mandated power
given to the Qağan in whose name the Išâ(d) commanded the army. If this is so, it
would point to a solar cult associated with the Khazar sacral ruler as well.
Within the steppe world, the Khazar Qağan was a figure of awe and reverence.
How did the Qağanal institution develop in Khazaria? When did it acquire these par-
ticular characteristics? Are there parallels elsewhere in the steppe world for this
particular line of evolution of political power? To understand this we must first say
something – briefly – about the origins of the Khazar state and people. Although
there is general agreement on broad issues – as usual the devil is in the details.31

Khazar Origins

The nomadic and semi-nomadic elements that subsequently constituted the subject
tribes of the Khazar union derived from a succession of Turkic tribal groupings and
confederations that came to the Volga–Caspian–Pontic steppe, the future Khazar
core territories, beginning in the 4th century CE. These migrations were the result of
movements out of the earliest Turkic homeland in Mongolia and Southern Siberia
that were associated with the fall of the Asian Hunnic/Xiongu state (mid-2nd century

————
Heaven) to capture the full sense of the Turkic qut. Skrynnikova (1992, pp. 71 – 85), who has made
a detailed study of these concepts translates suu jali as “Heavenly charisma”. This association with
“light” and “flame” should, in turn be connected with the Old Iranian concepts of xwarəna which
has the same notions of royal glory personified by light and flame, see Gnoli (1999), where, among
other things, xwarəna is defined as “magical force or power of luminous and fiery nature”. See also
Choksy (1988, pp. 36 – 37).
29
See also discussion in Czeglédy (1966, p. 14); Ludwig (1982, p. 133).
30
According to Rašîd al-Dîn (1982, I/1, p. 389), the Mongols had the custom of saying,
when they saw the ruler, “I have seen the golden face of the Emperor” [rûy-yi zarîn-i pâdišâh
dîdîm], clearly a reference to the sun-like brilliance that shone forth from him (see Skrynnikova
1992, pp. 81 – 82, who notes this fiery, glowing light as a sacral substance found in the head of a
person, the ruler). See also below.
31
The literature is too extensive to be discussed in detail here. On Khazar history, see: Dun-
lop (1954); Artamonov (1962); Golden (1980); Ludwig (1982); Novosel’cev (1990). The various
theories are surveyed in Romašov (2000 – 2001, pp. 258ff.). Omeljan Pritsak has dealt with the sub-
ject in a number of publications, most recently: Pritsak (1990, esp. pp. 3 – 5).
Acta Orient. Hung. 60, 2007
172 P. B. GOLDEN

BCE–mid-2nd century CE) and successive nomadic polities in Inner Asia.32 By the
latter half of the 5th century, tribes speaking Oğuric, a form of Turkic that differs
from Common Turkic, were present in the Ponto-Caspian steppes and had established
contact with Byzantium. Oğuric and Common Turkic split certainly before 400 CE
and perhaps as early as 500 BCE (Janhunen 1996, pp. 188–189, 218). In 552, the
Türks, led by the Ašina clan, a tribal union of complex origins initially including
Turkic, Iranian33 and subsequently very probably other ethnic elements (Proto-Mon-
golic or Para-Mongolic, Uralic and Palaeo-Siberian), toppled the Rouran/Asian Avar
state that had been the hegemonic power over the nomads of Inner Asia. They then
advanced westward pursuing the Avar remnants as well as tacking on to their now
burgeoning union other steppe nomads and the Soġdian city-states that dominated the
trans-Eurasian trade routes. In 557, İštemi, also known by his titles Yabğu Qağan and
Sin or Sir Jabğu Qağan,34 the younger brother and Co-Qağan of Bumın, the founder
of the state, allied with Iran, crushed the Hephthalite state35 and then, unhappy with
the Sasanids, made contact with Constantinople by 562 or 568 (Theophanes 1883, I,
p. 245; Menandros 1985, pp. 111ff.; Chavannes 1903, pp. 233–242).36 The Türk Qa-
ğanate was, grosso modo, a dual kingship, but one very different than the Qağanate
in Atıl/İtil.
Buffeted by a major defeat at the hands of the Sâsânids in 589 and by internal
discord, the Western Türks appear to have recovered sufficiently to provide crucial
military assistance to Constantinople in the years 625–628, enabling the Byzantines
to defeat Iran in the exhausting Byzantine–Sâsânid war of 603–638. Although the

32
For an overview, see Golden (1992, chaps II – V); Artamonov (1962, chaps 2 – 5).
33
Chinese and Soġdian sources have: Ashina, *Turkit-Ašinas (Bugut inscription, see Mori-
yasu – Ochir (1999, p. 123), the most recent reading, not universally recognised). The name Ašina is
most probably of Eastern Iranian or Tokharian origin (cf. Khotanese Saka âṣṣeina – âššsena “blue”,
Pers. axšaêna “dark-colored” Tokharian A âśna “blue, dark”) and corresponds to the Türk usage
Kök Türk, see Klyashtorny (1994, pp. 445 – 447). See also Šervašidze (1989, pp. 79 – 80). Ibn Xur-
dâdhbih (1889, p. 40) in his listing of the “kings of the Turk,” notes ‫[ ﺷﺎﺑﺔ‬Šâba, var. lect. ‫ َوﺳَﺎﻧﻪ‬wasâ-
na or “wa Sâna”]. Al-Mas‘ûdî (1966, I, p. 155) has ‫[ ﺷﺎﺑﻪ‬Šâba, with var. lect. ‫( ﺷﺎﻧﻪ‬Šâna), ‫( ﺷﺎﻳﻪ‬Šâya)
who, it is implied, was one of the “kings of the Turks,” or more specifically of the Qarluqs. This
form could be a corruption of ‫( اﺷﻨﻪ‬Ašina).
34
İštemi appears in Middle Iranian (preserved also in later Muslim historians), Armenian,
and Byzantine accounts under the title Sin or Sir Jabğu: ‫[ ﺳﻨﺠﺒﻮ ﺧﺎﻗﺎن‬Sinjibû Xâqân], see the Middle
Pers. Šahrestânîhâ î Êrânšahr, Middle Pers. Text, p. 13 (#9): Yabbu Xâgân, Sinjêbîk Xâgân
(Sinjepuk / Sinjepîg). He is also Sir Jabğu: Σιλζίβουλος, Ζιέβηλ (perhaps to be emended to Ζιέβηχ as
suggested by Pritsak) < Śri Yabğu Qağan (see Dobrovits 2004, pp. 112 –113). This title, in the form
Yabğu Qağan (cf. the Georgian Jibğa and the Armenian Jebu Xak’an, see Pritsak 1996, pp. 234 –
236; Golden 1980, I, pp. 187 – 190; Dobrovits 2004, pp. 112 – 113) was borne also by some of his
successors in the West as well. See also al-Mas‘ûdî (1966, I, p. 307). The form jabğu may reflect
the pronunciation of the local Oğuric-speaking tribes, the Common Turkic form was yabğu, a term
that was borrowed by Turkic from the Kušans and of probable Iranian origins. See Pulleyblank
(1966, pp. 27 – 28); Litvinskij (1968, pp. 21 – 22); Chin. xihou: *hiǝp-hu or *siep-hu.
35
The Hephthalites had evolved out of a mix of “Hunnic” and other, perhaps Proto-Mon-
golic groupings, similar to the War-Huns that had constituted the core of the Asian Avar state. See
Golden (1992, pp. 69 – 72, 76 – 83).
36
On the Soġdians, see Vaissière (2002).
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 173

ethnonym “Khazar” surfaces in some accounts of these campaigns which were cen-
tered on the Caucasus,37 as well as in connection with earlier wars of Iran and the
peoples of Transcaucasia against the “northern peoples,”38 these earlier references are
almost certainly anachronistic and the same may be true with regard to those of the
early 7th century.39

Khazar Ethno-linguistic Affiliations

The most widely accepted notion is that the Khazar union derived from a mix of
Oğuric tribes, the Sabirs40 and the Türks. The ruling strata, in this reconstruction,
stemmed from the Western Türk state and their role, in many respects, was similar to
that of the later Činggisids in the Golden Horde, i.e. a relatively small, outside ruling
house and its entourage which was linguistically distinct from its subject population.
In Khazaria this meant the numerically larger masses of Oğuric Turkic and Alano-As
tribes. The theory linking the Khazars (Qazar) with the Uyğur or more broadly Tiele
Qasar tribe whose name has been derived from Caesar, a Wanderwort remains, in
my view, still problematic (Németh 1930, p. 204; 1991, pp. 162–163; Golden 1980,
I, p. 53; Czeglédy 1983, pp. 104–106).41

37
See the occasionally (chronologically) confusing account of Movsês Dasxuranc’I (1962,
pp. 81 – 88) and the K’art’lis C’xovreba (1955, I, pp. 223 – 225, the chronicle of Juanšer, pp. 374 –
375, the account of Sumbat Davit’is dze). For the latter, see also the English translation by Robert
Thomson of the Georgian text and Old Armenian translation of this chronicle (Thomson 1996, pp.
233 – 234, account of Juanšer).
38
See the earliest Georgian notices in K’art’lis C’xovreba (1955, I, pp. 11 – 12, 19, 27, 59,
63 etc.); see also Thomson (1996, pp. 13 – 15, 20) (the Armenian adaptation makes mention of both
Türks and Khazars, pp. 26, 39, 70, 75 etc.).
39
See discussion in Melikset-Bek (1960, pp. 113 – 118). With two exceptions, the Arme-
nian historians of the 4th– 7th centuries do not mention the Khazars. The two exceptions are Mov-
sês Xorenac’i, an author of uncertain dating (the 5th century is claimed by the source itself, but other
evidence indicates, at the very least, a later hand of the 7th, 8th or perhaps even 9th century, see
Moses Khorenats’i (1978, pp. 210 – 211). On the question of authorship and dating, see Thomson
(1996, pp. 58 – 59) (“predates end of 8th century”); Toumanoff (1963, p. 18); Rapp (1997, pp. 55 –
69), and the historiographically problematic Armenian Geography attributed to Movsês Xorenac’i,
but most probably the work of Ananias Širakec’i (probably 7th – 8th century): Ananias of Širak
(1966, I, pp. 7– 15). The latter makes reference to the defeat of the Onoğur – Bulğar tribal union by
the Khazars and the flight of Asparuk (Ananias of Širak 1966, I, p. 48). This occurred sometime
before 679 when Asparuk (in Byzantine sources ’Ασπαρούχ) led his part of Qubrat’s horde into the
Balkans, laying the foundation of the Balkan Bulgarian state (Theophanes, p. 357; Nikephoros, pp.
86 – 91). The precise dating of this contest is unclear. Presumably, as Romašov has argued, it came
after the Arab – Khazar war of 652 – 653, perhaps sometime around 668, see Romašov (2000– 2001,
pp. 272 – 279).
40
Sabir ethno-linguistic affiliations are not clear. The few names and titles preserved in our
sources can be viewed as Turkic (Golden 1980, I, pp. 256– 258) or perhaps even Proto-Mongolic
(Pritsak 1996, pp. 228 –230).
41
Czeglédy does posit an actual Qasar tribe which, in his view, came to the Ponto-Caspian
steppes ca. 463 along with the Oğuro-Bulğaric migrations. The Qasars are noted in several Uyğur
runic inscriptions (Šine Usu, Tes, Terxin, see Klyashtornyi 1983, pp. 356 – 366; 1985, pp. 137– 156),
Acta Orient. Hung. 60, 2007
174 P. B. GOLDEN

The Emergence of Khazaria

The Western Türk Qağanate, frequently embroiled in internecine strife, succumbed

to the Tang in 659, who had earlier (630) subjugated the Eastern Qağans. The Khazar
state, then, emerged as an offshoot and successor state of the Western Türk Qağanate
in its westernmost zone (Volga–North-Caucasus–Pontic steppes) sometime in the
period between 630 and ca. 650,42 a time of great instability just prior to the collapse
of the Western Türk state. The ethnonym Qazar may have begun as a socio-political
term, related to qazaq “freebooter, one who lives freely”43 the name taken (or given
to) Chinggisid-led groupings that broke away from Abu’l-Xair’s Özbek confedera-
tion in the 15th century. The ethnonym Bulğar “rebel, one who breaks away” had
very much the same meaning (see Németh 1991, pp. 130–131; Kljaštornyj – Savinov
2005, p. 63).
In any event, by the 650s–670s, we have firm evidence for the existence of
the Khazar Qağanate in control of the Western Eurasian steppelands. They had begun
a protracted war with the Arabs for dominion in the North Caucasus and had
subjugated the Onoğur–Bulğar tribal union located in the Kuban and eastern Pontic
steppe zone, driving some of them out of the steppe (Romašov 2000–2001, pp. 310–
318). The Khazar ruling house was most probably of Ašina origin. If “Ânsâ” noted in
the Ḥudûd (Ṭarxân Xâqân… one of the descendants of Ânsâ) is a garbling of Κâd, as
has been argued, we would not expect the title of Qağan to be associated with him
here – unless one conjectures that the Κâd had moved up and was now using the
Qağanal title for which there is no evidence. If that were case, we should expect Ibn
Faḍlân’s Qağan Bäh and not Κâd. The Ḥudûd is a pastiche of notices and part of the
Jaihânî (Ibn Rusta 1892; Gardîzî – Barthold 1963–1977; Gardîzî – Habîbî 1984; Gar-
dîzî – Martinez 1982) tradition. Nonetheless, it does contain unique information and
this possible reference to the Ašina connections of the Khazar royal house cannot be
automatically dismissed.44 The investiture ceremony of the Khazar Qağans, virtually

————
although the historical context is far from clear. Róna-Tas (1982, pp. 349 – 379; 1983, pp. 126 – 133;
1999, pp. 228 – 229) has attempted to make the linkage linguistically with Qasar. However, the name
Qasar appears simultaneously with “Khazar” as an anthroponym among the North Caucasian Huns,
a subject people of the Khazars, whose ruling house may have been of Khazar origin. In the steppe
empires, members of the ruling dynasty were often placed as the heads of subject tribal groupings.
The evidence for a presumed shift s > z (Qasar > Qazar) in Khazar has yet to be adduced.
42
This is the most widely accepted view, see, among others, Ludwig (1982, p. 134).
43
See the comments of Kljaštornyj – Savinov (2005, p. 58), who argue that this was not an
unknown form of developing politonyms. In addition to Qazaq, they note Rouran (= Middle Chin.
*nönör which denoted some kind of group of steppe freebooters led by a chieftain who had broken
away from his clan and tribe). This is the equivalent of the Činggisid era nökör.
44
Ḥudûd / Sutûdah (1340 / 1962, p. 193); Ḥudûd / Minorsky (1970, pp. 161 – 162), cf. the
“king”, Ṭarxân Qağan, is noted by the Ḥudûd al-‘Âlam (see above) as “one of the descendants of
Ansâ”. This name, ‫ ﺁﻧﺴﺎ‬in the manuscript, may be read as ‫*( ﺁﺷﻨﺎ‬šnâ) or perhaps ‫*( ﺁﺷﻴﻨﺎ‬Ašînâ), the
name of the Türk ruling house. It should be noted, however, that this form may also be a garbling
of ‫اﻳﺸﺎ‬, ‫ اﻳﺸﺎد‬etc. noted in the Ibn Rusta – Gardîzî tradition, a viewpoint expressed by Minorsky and
accepted by Zuckerman (2002, pp. 552 – 554). See mss. forms in Golden (1980, I, pp. 206 – 207,
219 – 220).
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 175

identical to that of the Ašina Türks and not noted elsewhere in the Turkic world
points strongly towards the Ašina origins of the Khazar ruling house. Moreover, in
the steppe, one did not adopt the Qağanal title lightly or arbitrarily, one acquired it
by conquest or descent – in this case probably from a Western Türk Ašina who fled
the ongoing strife in that fractious realm. The Khazar rulers, as has been suggested
by Artamonov (1962) and Gumilëv (1967), may have descended from one of the
rival Western Türk Ašina contestants for power. This was, perhaps, Yipi shigui, who
was driven from power (or killed) ca. 651. Either he or one of his heirs may have fled
westward and established himself in what became Khazaria. The history of this period
is very murky.45 The Chinese notices on the Khazars, often note them as Tujue [EMC
dwət kuat, LMC tɦut kyat (Pulleyblank 1991, pp. 311, 168)], Ke-sa [EMC khat sat,
LMC kha’ sat or Tujue He-sa [EMC vat sat, LMC xɦat sat] = Türk Qazar, Türk Xazar
(Chavannes 1903, pp. 145, 170; Pulleyblank 1991, pp. 173, 271, 123).46 Tibetan us-
age, Dru-gu Ge-sar (Róna-Tas 1999, p. 229) follows this. Tujue is only used with
peoples who were part of the Inner Türk world, not as a generic “Turkic”. Needless
to say, there is a considerable variety of viewpoints on the origin of the Khazar qağa-
nal line and state.47

45
Yipi shigui (EMC ʔit bji ʐiajk gwih, LMC ʔit pɦji ʂɦiajk kyj’) or Yipi shegui (EMC ʔit
bji ʐia gwih, LMC ʔit pɦji ʂɦia` kyj’, see Pulleyblank 1991, pp. 367, 236, 279, 283, 115). See
h

Chavannes (1903, pp. 3 – 4, 32 – 33, 58, 59, 265, 266); Artamonov (1962, pp. 170 – 171); Gumilëv
(1967, pp. 218 – 220, 233 – 234, 237 – 238).
46
Chavannes (1903, p. 164) also notes the “Tujue Yantuo”, one of the inner core of tribes
or tribal unions associated with the Türk realm. Chinese cannot represent -z-, hence the forms -s-
(Ke-sa, He-sa) can stand for Qazar or Qasar, Xazar or Xasar.
47
Thus, Novosel’cev (1990) suggests that an already independent Khazar state existed by
the early 7th century. Its ruler assumed the Qağanal title sometime in the period ca. 630 –650. This
state was built on an ethnic base of peoples associated with the Hunnic state, various Turkic peoples,
Ugrians and the older Iranian population of the western steppes. The Khazars were among the Turkic
tribes that arrived in the region, came under the authority of the Türk Qağans, but by the early 7th
century, although still nominally subjects of the Türks, were actually independent. Their ruler bore
the title of Jabğu Qağan. The Qağanal title is attested by the 640s– 650s, but was assumed as early
as 630. Again, the Khazar ruler is associated with the ruling Türk-Ašina (in this instance, Tong Yab-
ğu, a son / grandson (?) of İštemi) who was assassinated in 630. The Khazar ruler while retaining the
older title of Yabğu now also took the Qağanal title, see Novosel’cev (1990, pp. 80– 91). The date
of 630 is also accepted by Romašov (2000 – 2001, p. 302). Pritsak also derives the Khazar ruling
house from the Ašina. He suggests that they fled, after the fall of the Western Türk state, to the land
of the Akatir / Akatzir (*Aq Qazar) union and hence the ethnonym Khazar. Here, they were joined
by the Varâz / Warač / Barč, another Western Türk clan. The Barč now functioned as subordinate
rulers, the Beg and Ixšêd noted in the Muslim sources. The peoples over whom they ruled here
were predominantly speakers of Hunno-Bulğaric which became the lingua franca of the Khazar
state (see Pritsak 1978, p. 261; 1990, pp. 3 – 5). In a recent study Pritsak (1996, p. 236) identified
the name Čepetux with Jabğu Qağan, suggesting that the name is actually *Jabğu Tux (with the shift
Tong > Toġ > Tux in Bulğaric, the youngest son of T’ung Jabğu of the Western Türks and nephew
of Xieli, the Great Qağan of the Eastern (senior) Türk state. This Čepetux/ *Jabğu Tux, who also had
the title Šad, came west to aid Herakleios and he is the progenitor of the Khazar Ašina house, (actu-
ally “Čenastan Čepetux” = “Čepetux of China”). The passage is found in Sebeos (1992, I, pp. 50 –
51), where this same figure is noted as Čembux, the commander of the “great Khak’an, king of the
regions of the North”, (p. 54) “they (the Armenians, pbg), rebelled (against the Persians, pbg) and
Acta Orient. Hung. 60, 2007
176 P. B. GOLDEN

Khazar affiliations within the Turkic or larger Altaic world have yet to be fully
unraveled. Presuming that the royal house was of Western Türk Ašina origin, it, un-
doubtedly, brought with it a solid core of Western Türk followers. The archaeological
sites identified with the Khazar state show a wide variety of ethnic and racial types.48
The fifty odd titles, toponyms and personal names transcribed in our sources reflect-
ing the Khazar tongue are overwhelmingly Turkic, but few of them have specific lin-
guistic markers that would enable us to determine what form of Turkic was domi-
nant.49 It is very likely that several Turkic languages were used – and even written in
variants of the Turkic runic script – on the territory of Khazaria. There are scattered
finds of fragments of runic texts. These still need much work. Oğuric Turkic, spoken
by many of the subject tribes, doubtless, was one of the linguae francae of the state.
It, or some other Archaic Turkic language related, perhaps distantly, to Oğuric, is
clearly seen in a number of Khazar titles and names of Turkic origin, but a number of
sound changes have occurred in these terms which became the source of considerable
confusion to outside observers – some of whom claimed that Khazar shared nothing
in common with any other language of humanity (see Golden 2005, pp. 206–207).
Alano-As was also widely spoken. Eastern Common Turkic, the language of the royal
house and its core tribes, in all likelihood remained the language of the ruling elite in
the same way that Mongol continued to be used by the rulers of the Golden Horde,50
alongside of the Qıpčaq Turkic speech spoken by the bulk of the Turkic tribesmen that
constituted the military force of this part of the Činggisid empire. Similarly, Oğuric, like
Qıpčaq Turkic in the Jočid realm, functioned as one of the languages of government.
However one reconstructs these crucial years it is clear that the Qağanal insti-
tution did not develop locally, but was most probably associated with the Ašina, the
Türk ruling house, the only legitimating source. It was brought in from the outside
and was strengthened by conflict with the Arabs and the Bulğar union for dominion
over the western Eurasian steppes.
We do not have a great deal of information about the functions and workings
of the Qağanal institution among the Western Türks, but what we do find in the
Chinese and Arabic sources is a Qağan who is always portrayed as an active player
in affairs: a political and military leader (Movsês Dasxuranc’i 1962, pp. 87–88).51
————
submitted to the great Khak’an, king of the regions of the north, under the Chinese Chepetukh. They
went from the east to the west across the regions of the north to join the army of that Chepetukh at
the command of their king the Khak’an. Passing through the Pass of Chor with many troops, they
went to assist the king of the Greeks”. This notice seems to imply that a major Türk army was sent
from the core Türk lands, not the periphery near the Caucasus. Howard-Johnston, in the commentary
(Sebeos 1999, II, pp. 185 – 187), places this notice in its proper context and correctly identifies Čem-
bux/ Čepetux with “Jebu Khak’an” et al. The Armenian forms or deformations of this name reflect
their transmission from Pahlavi to Armenian. This identification is well known, see Dunlop (1954,
p. 32).
48
See Pletnëva (1999) and also see Mixeev (1985) who focuses on the Bulğaric, Alanic and
Ugric population of the Don region during the Khazar period.
49
These words have been studied by Golden (1980, I, pp.112 – 255).
50
See comments: Grigor’ev (1981, pp. 81 – 89).
51
During the early 7th century war against the Sâsânids, the actual direction of the Western
Türk army was entrusted to the Šad, a very high title in the Türk administration just below that of
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 177

Among the Türks there was a sacral dimension to the Qağanate, at least in the descrip-
tions of the institution in the Chinese sources. The Türk investiture ceremony with its
shamanic elements and ritual strangulation of the new Qağan is clearly mirrored in
that of the Khazars.52 These and other legends associated with the Ašina point to sha-
manic or shaman-like functions associated with the Türk Qağan. The sunward turning
in felt (a symbol of rebirth) probably points to a solar cult. The nine turns reflect the
ruler’s movement through the nine zones of the heavens. The ritual strangulation of
the Qağan brings him to the other, spirit world and hence his ability to tell how long
he will rule.53 This investiture rite is not mentioned as practiced by other Turkic or
Inner Asian peoples outside of the Türks and Khazars. The ideological formulas of
the Orxon inscriptions with their emphasis on the heavenly origins of the Qağans and
their possession of qut “heavenly good fortune”, the sign of having the mandate of
heaven, all underscore this aspect of Türk beliefs.54 The ruler’s blood was sacred and
could not be shed, hence the use of the silk scarf in the ritual strangulation. The later
Ottoman use of the bow cord to dispatch royal claimants to the throne is probably a
faint and distorted reflection of the tradition of strangling royal rivals, but it was
never practiced in Ottoman investitures.55
In this respect, Türk notions of kingship were very similar to other monarchies
that rested on divine right ideologies of rule. Türk Qağans had sacral perhaps even
sacerdotal qualities, but they were not sacralised (Márton 1997, p.78). Khazar Qa-
ğans, however, were. As we have seen, the Khazar Qağan had become a tabuised
personage, ritually isolated, a talisman for the good fortune of the state, responsible
for order in the cosmos. As a consequence of his sacral status, he was not an active
ruler. He was the “law-king” pure and unsullied by the blood-letting that was part of
real governance. The task of dealing with the daily affairs of state, was left to the
“war-king.” This form of kingship is well known in the anthropological literature
(Hocart 1970, pp. 161–170, 176–179). Thus, the Khazar dual kingship was funda-
mentally distinct from the bipartite division of governance that we see in the Türk
empire with its heavenly ordained Qağans of the East and West. In Khazaria, there
was no division of rule into eastern and western spheres. At the end of his term, de-
termined by the shamanic trance/ritual strangulation noted above, the Khazar Qağan
————
Qağan and usually given to a close relative of the latter, in this case the nephew of the Qağan, “the
king of the north”. On Šad, see above.
52
Liu Mau-tsai (1958, I, p. 8), the report in the Zhou-shu: “When a new ruler is chosen, the
high officials carry him out from his nearest surroundings in a felt covering and turn him, then, in
the direction of the sun nine times. With each turn, all of his subjects bow to him. After (the turn-
ing) and the bowing, they help the prince onto a horse and let him ride. Then, they strangle him
with a silk scarf to the extent that he barely remains alive. Then, they loosen the neck binding and
ask him hurriedly: ‘how many years will you be our Qağan?’ Since the Qağan is in a dazed state, he
cannot say clearly the length of time. They then determine from the words which he blurted out in a
muddled state the length of his term of office.”
53
From this, Waida concluded that sacral kingship among the Türks originated in the ruler
as shaman, see Waida (1976, pp. 180 – 188).
54
See in brief: Roux (1959, pp. 231 – 241).
55
Choksy (1988, p. 41) (inviolability of Sasanid royal blood); Roux (1984, p. 51); Köprülü
(1940, pp. 15 – 23; 1944, pp. 1– 9); Akman (1997).
Acta Orient. Hung. 60, 2007
178 P. B. GOLDEN

was sacrificed – without shedding his blood. He was associated with a sun cult (Czeg-
lédy 1966, pp. 14–26; 1974, pp. 11–17; Márton 1997, p. 74).

Origins of the Sacral Qağanate: Dating – When Did

the Khazar Qağanate Acquire These Unusual Characteristics?

None of these features is evident in our sources dealing with the Khazars up to the
9th century (Ludwig 1982, p. 165). Up to then he is portrayed as a figure actively in-
volved in the military affairs of state – although armies are most often sent forth under
the command of a general, usually a close relative – at least that is the impression left
in our sources.56 In the attempt by the deposed Byzantine Emperor, Justinian II, to
regain his throne in the early 8th century, he married the sister of the Khazar Qağan.
The latter is shown, in the Byzantine sources, as clearly fulfilling all the functions of
an active ruler. There is no mention of a “deputy” or any other intermediary.57
Our notices on the sacral Qağans all come from later Muslim sources writing
in the 10th century, but frequently using materials from the 9th century and are
hinted at in the Khazar Hebrew documents. Al-Ya‘qûbî, who completed his History
in 278/891 (covering the period up to 873) simply points to the existence of a
“deputy” without any comment as to whether the Qağan was an active figure. The
earliest possible indication of a sacral status stems from two post-Khazar sources, al-
Tannûxî (d. 994) and at- al-Ṭarṭûšî (d. 1126). They report that one of the vezirs of the
‘Abbâsid Caliph al-Ma’mûn (813–833), Faḍl b. Sahl (d. 818), heard from a Khazar
ambassador that during a time of severe heat and misfortune in Khazaria, people had
come to the king for help. The king was unable to provide it. His sister, however, the
“Xâtûn” (Xatun/Qatun “empress, queen”) gave them a stirring speech, calling on
them to appeal to God. This quieted them. They then went from the door of the “infe-
rior king” (‘an bâbi’l- maliki’l-adnâ) to that of the “superior king”. The account is not
without problems. Moreover, it can also be read as “they went from the lower door of
the king to the higher door of the king” (al-Tannûxî 1939, p. 178; al-Ṭarṭûšî 1888–
1989, p. 152).58 The important role played by the Qatun was not unusual in the pre-
Islamic steppe world and lends some credibility to the account. Somewhat more cer-
tain is the notice in Constantine Porphyrogenitus on the building of the Khazar fort
Sarkel (ca. 840–841), on the left bank of Don River, for which the Qağan and the Beg
(`O g¦r cag£noj ™ke‹noj kaˆ Ð płc Χαζαρίας) requested Constantinople’s help
(Const. Porph. 1967, pp. 182–183; Ludwig 1982, pp. 167–168).59 Clearly, the Qağan-

56
At one point, with the death of a Qağan, P’arsbit’ (Barsbeg) his mother takes over the
direction of the state and military affairs (Łewond 1982, pp. 69, 107, 125 – 126; Novosel’cev 1990,
p. 139).
57
See discussion in Dunlop (1954, pp. 171ff.); Head (1972, pp. 102ff.); Golden (1980,
I, pp. 60 – 61, 182– 184).
58
See also comments of Togan in Ibn Faḍlân (1939, pp. 263 – 264); and Ludwig (1982, pp.
166 – 167).
59
On the dating of Sarkel, see Zuckerman (1997, pp. 213 – 214).
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 179

Beg had by this time become an important figure in shaping policy and was an offi-
cial “face” that the Khazars presented to the outside world.60 Ibn Khurdâdhbih, in re-
counting the nearly contemporaneous dispatch (ca. 843) of Sallâm the Interpreter by
the ‘Abbâsid Caliph al-Wâthiq (842–847) to the “land of Gog and Magog” (i.e. the
North Caucasus, Khazaria and beyond), reports that the Caliph wrote to the governor
of “Arminiyya” who then wrote on the embassy’s behalf to the “Lord of the Throne”
(roughly the lands of the Daghistanian Avar people of today), who wrote to the king of
the Alans, who, in turn, wrote to the Filânshâh (in the North Caucasus) who “wrote
to Ṭarxân, the king (malik) of the Khazars for us” (Ibn Xurdâdhbih 1889, pp. 162–
163).61 If this “king” – and it should be remembered that this was the term so often
used in the Islamic sources for the Qağan-Beg is, indeed, the Qağan-Beg, it would
indicate that the latter was conducting the affairs of state by the early 840s at the
least. Kovalev, on this basis, has concluded that the Qağan-Beg, already an active
participant in a dual kingship ca. 800, took over the actual governance of the state
sometime between 838 and 843. This Qağan-Beg was Bulan/Sabriel, who played so
prominent a role in the Khazar conversion narrative (Kovalev 2005, pp. 231–232).
Of course, it is entirely possible that the Qağan-Beg was already handling the affairs
of state before this. In the diplomatic correspondence of the Qağan and Beg noted by
Constantine Porphyrogenitus, it would be only natural that the Qağan would, at least,
be nominally a part of the process. The notice is important because it shows us that
the Qağan-Beg was fully on the stage by this time.
Recently, Constantine Zuckerman has argued that the development of sacral
kingship among the Khazars arose out of a series of defeats that he believes the
Khazars sustained at the hands of the Hungarian tribal union which had migrated into
the Pontic steppes. These defeats, which would have cost the Khazars territory west
of the Don River, are the presumed reason for the building of Sarkel and thus should
be placed before 840–841, i.e. in the 830s. This is not a new idea. It was put forward
by Marquart more than a century ago and by Németh in 1930. Marquart also empha-
sised Ibn Rusta’s notice that “in the past”, the Khazars “built trenches around them-
selves in fear of the Majġariyya and other peoples who bordered on their country”
(Marquart 1903, pp. 27–28; Németh 1930, pp. 153–155; 1991, p. 218).62 The passage
itself, unique to Ibn Rusta, may be a later interpolation.63 Zuckerman argues that the
defeats suffered by the Khazars at the hands of the Hungarians (undocumented in our

60
See also discussion in Kovalev (2005, pp. 230 – 231).
61
The sending of the embassy is also mentioned by Ibn Rusta (1892, p. 149) and al-Muqad-
dasî (1906, p. 362). The latter also notes “Ṭarxân malik al-Xazar”. Another embassy, led by Mu-
ḥammad b. Mûsâ al-Khwârazmî, the famous mathematician and geographer, of presumed Kwaraz-
mian origins and hence familiar with these lands, preceded that of Sallâm. See also Dunlop (1954,
pp. 190 – 191).
62
The latter argues that this could not have been against the Pečenegs; Golden (1980, I,
67ff.); Zuckerman (2002, pp. 521, 525ff.); Ibn Rusta (1892, p. 143). Kovalev (2005, pp. 235 – 236)
also views the successful Hungarian occupation of lands west of the Don as one of the factors
accounting for the diminution of the powers of the Qağan and the rise of the Beg.
63
See discussion in Göckenjan – Zimonyi (2001, p. 74, n. 102). This important comment is
not found in the other accounts based on the Jaihânî tradition.
Acta Orient. Hung. 60, 2007
180 P. B. GOLDEN

sources, but certainly a possibility) were sufficient to weaken the power of the
Qağanal house. This paved the way for the rise to power of the Šad (Išad) who also,
he argues, brought about the conversion to Judaism sometime after 861 (Zuckerman
2002, pp. 528–531).64
The dismantling of the western, Pontic zone of Khazaria is conjecture. Ac-
cording to Rus’ tradition, they were still collecting tribute from the Dnepr Slavs in
the 850s.65 In fact, Ibn Rusta (or his later interpolator) only tells us that the Khazars
were building fortifications, beefing up their defenses sometime in the past. It is very
likely that Ibn Rusta or the later interpolator has Sarkel in mind,66 but the Hungarian
tribal union was not the only threat as Ibn Rusta himself notes (“and other peoples”).
Byzantine sources indicate that the Pečenegs, pressured westward by the Oğuz and
others, were also a cause for concern. They were not yet a regular presence in the
Pontic steppes, but given their attacks on the Hungarian union by the end of the 9th
century, Pečeneg raids cannot be excluded as one of the unsettling elements in the
region. Their subsequent impact was traumatic, causing the loss of Khazar-controlled
regions and the displacement of tribes.67 The Pečeneg invasion was far more unset-
tling than whatever earlier problems the Khazars may have had with the Hungarians.
Some would date the beginning of the decline of Khazaria to the Pečeneg invasion
(Romašov 2004, p. 222). Nonetheless, we have no evidence that the Hungarians or
Pečenegs were existential threats to Khazaria that would produce, in essence, a
restructuring of the modes of governance. In the case of the Pečenegs, this would
have occurred, in any event, long after the conversion to Judaism.
The nature of the Khazar–Hungarian relationship remains the subject of
considerable debate and speculation. Constantine Porphyrogenitus’s famous chapter
38 of the De Administrando Imperio which provides some information on these early
movements of the Hungarians, obtained first-hand, apparently, from Árpádids who
were part of a Hungarian embassy to Constantinople in his time (mid-10th century)
as well as from several other accounts, some of earlier provenience, has also been
open to a variety of interpretations. Constantine’s Árpádid guests, as some Hungarian
historians have suggested, may have presented him with a tendentious, highly parti-
san representation of events reflecting only what the victors in internal Hungarian

64
In the process, Zuckerman, among other things, offers a revision of early Hungarian his-
tory, in particular the dating of the Hungarian migration into the steppe zone. This is a topic that
has long been a battleground with rival perspectives and a very detailed literature in Hungarian that
Zuckerman has ignored. He bases his dating on the work of the late Gyula Kristó (1996) a very fine
historian representing one of a number of competing viewpoints.
65
PSRL, II, c. 12 – although the double-edged swords that the Poljane offered them were a
source, according to the Rus’ account, of some concern on the part of the “Khazar elders”.
66
As I have argued previously, see Golden (1980, I, pp. 67 – 77).
67
Romašov (2004, pp. 219 – 222) places the Pečeneg crossing of the Volga in 889, after
considerable pressure from tribes to the east. They ravaged the Khazar Don region, subjugating the
Alanic and Bulğar tribes there and destroying the Khazar defense network. They attacked the Cri-
mea and perhaps at the same time moved against the Hungarians in Etelköz, expelling them from
there ca. 895.
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 181

struggles wanted the learned emperor to hear.68 The dating of the migration of the
Hungarian tribal union from their trans-Volga homeland to the Pontic steppes – one
of the many minefields of Hungarian “proto-history” remains equally contentious
and far from resolved: for example Kristó places it ca. 830, Németh puts it between
830 and 850, others have placed considerably earlier (Kristó 1996, p. 106; Németh
1930, p. 218; see also discussion in Golden 1992, pp. 260–261).69 The De Administ-
rando Imperio notes that the Hungarians “of old” lived “next to Khazaria in the place
called Λεβεδία (Lebedia/Levedia)”, named after their first vojevoda (βοέβοδος, the
latter a Slavic term, perhaps coming via an interpreter or already borrowed into 10th
century Hungarian, cf. Hung. vajda: TESz III, p. 1070). Lebedias (Λεβεδίας), how-
ever is a Hungarian name: < Hung. *Lëvedi (Kristó 1996, pp. 107–108,112, 115).70
He was a primus inter pares among their seven tribes (or clans). Here, according to
Constantine, they were allies of the Khazars for “three years”, bound to them also by
marital ties and here the Magyar-led Hungarian tribes, in Kristó’s view, were formed
into a union (Constantine Porphyrogenitus 1967, pp. 170–171; Kristó 1996, p. 117).
The location of this land, east or west of the Don has also produced an extensive lit-
erature.71 The length of the Khazaro–Hungarian alliance, like so many other aspects
of Pre-Pannonian Hungarian history, remains equally disputed, ranging from Constan-
tine Porphyrogenitus’s “three years” to three hundred years (Dvornik et al. 1962,
p. 148). Regardless of its duration, the period of Khazar dominion over the Hungarian
union was profound and left traces in the organisation of the Hungarian union. The
Hungarians may have adopted a non-qağanal version of the sacral kingship – although
other explanations can be offered. The Hungarian “sacral” ruler, took the Khazar title
Kündü (> kündä > Kende in Hungarian).72 The Hungarian sojourn in the Pontic
steppes was relatively brief. By the late 880s, under pressure from the Pečenegs, they
migrated westward to Etelköz (lit. a mesopotamia, i.e. “the [land] between the
rivers”), located, according to Németh, between the Dnepr and Seret (Németh 1991,
p. 218). By the late 9th century, as noted above, the Pečenegs drove them into Panno-
nia.73 Given the Khazar domination and political and cultural influence after 830 or
thereabouts, the traumatic impact that Zuckerman posits as touching off the crisis
that reduced the position of the Khazar Qağan to that of a largely ceremonial figure,
does not seem very likely. As was typical of nomadic confederations, the newly

68
See the recent discussion in Kristó (1996, pp. 97ff.).
69
Some Hungarian scholars place the migration as early as the 5th century, others set it at
various times in the 8th century. One school of thought posits two migrations to Pannonia, the first
of which took place in the late 7th century and the second at the end of the 9th century.
70
“Lebedia” / Levedia could refer either to Lëved’s core zone or to the whole of the territory
held by the Hungarian tribes under his leadership. The name itself is of Uralic origin, from the root
le-/ lesz- “to be”, see Németh (1991, p. 222).
71
See summary in Kristó (1996, pp. 108 – 111), who opts for the area west of the Don.
Németh (1991, p. 218) locates it between the lower Don and the lower Danube.
72
Ligeti (1986, pp. 253 – 254); and also Kristó (1996, pp. 132 – 136), who places the period
of Khazar domination to ca. 840 – 861.
73
Constantine Porphyrogenitus (1967, pp. 166 – 167), writing ca. 948 – 952, places this some
fifty-five years prior to the time of his writing.
Acta Orient. Hung. 60, 2007
182 P. B. GOLDEN

absorbed were placed on the borders. The Hungarians in the Pontic steppes may have
served as just that, border troops for the Khazars or support forces for the right
wing.74 Indeed, it has recently been suggested by Árpád Berta that (H)ungar- derives
not from Onoğur but from Turko-Khazar onğar “right wing” (Berta 1992, pp. 7–11).
In the account in the Old Slavic Žitije Konstantina/Vita Constantini of the em-
bassy of Constantine/Cyrill to Khazaria in 861, the Qağan appears to be in charge of
the reins of government. The Beg, as such, is not mentioned. Perhaps, he is the “first
councillor” (pr’vyj s’vetnik) who is noted in some of the events. Zuckerman, in par-
ticular, gives special weight to this document, while dismissing others (Zuckerman
2002, pp. 525ff). The text, however, has many “obscurities” (Vlasto 1970, pp. 34–
35) and this presentation of the Qağan may be a later interpolation (Pritsak 1988, p.
298). Constantine’s mission to Khazaria clearly aimed at undoing the Judaisation of
the Khazar ruling elite or at the least securing the rights of the Christian community
there (Vlasto 1970, pp. 34–35).75 Why was this a problem at this stage? The answer
seems self-evident. Khazar Judaisation, already in progress since the early 9th cen-
tury at the upper strata of society, was apparently spreading. According to the text,
Constantine, who may have already had some familiarity with Semitic languages,
acquired Hebrew in the Crimea as a preparation for the expected religious debate in
the Khazar capital. The Vita Constantini portrays Constantine as brilliantly trouncing
his Jewish opponents and so impressing the Qağan that the latter sings the future
saint’s praises in a letter to the Byzantine Emperor, allows any who so desire to con-
vert to Christianity and indeed promises to do so himself. The reality is that Constan-
tine returned home with several hundred “Greeks” who had been held prisoner in
Khazaria (certainly a sign of worsening Khazar–Byzantine relations) and nothing
else.76 Perhaps some amelioration of Khazar–Byzantine relations was achieved, but
the major prize, the conversion of the Qağan, had completely eluded this very tal-
ented diplomat. In this regard, the mission was a failure, a fact barely hidden in this
hagiographical text. This may be viewed as further evidence that Judaism was al-
ready firmly implanted in the Khazar court by that time (Vlasto 1970, p. 35; Róna-Tas
1990, p. 233). The ambiguity about or distortion of the role of Judaism at the Khazar
court found in the Vita Constantini are not surprising. Other Byzantine sources avoid
all mention of Khazar Judaism. This was hardly accidental. The Byzantines who had
an often uneasy entente with the Khazars and had even arranged a marital alliance
between the Byzantine and Khazar ruling houses, were clearly deeply stung by the
Khazar conversion to Judaism (see Shepard 1998, pp. 9–34). Constantine himself, in
the De Administrando Imperio warns against marriages with the northern barbarians
and specifically cites the Khazar marriage (732) of Constantine Copronymous, son of

74
Kristó (1996, pp. 130, 138): an interpretation of Σάβαρτοι ¥σφαλοι of the DAI (pp. 170–
171), an earlier name of the Hungarians, as savard denoting “the one behind the right wing”.
75
Vlasto suggests that it was also to deal with a joint Khazar – Byzantine response to no-
madic raiding. This, a perennial problem, may well have been one of the topics discussed, but we
have no documentation of it.
76
See text in Lehr-Spławiński (1959); see also discussion in Bernštejn (1984, pp. 58 – 76)
which also touches on other references to Constantine’s activities in these critical years.
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 183

Leo the Isaurian, as something that “attached great shame to the empire of the Ro-
mans” (Constantine Porphyrogenitus 1967, pp. 68–73). A recent study of the De
Administrando Imperio concludes that from the Emperor-Historian’s perspective,
Khazaria was at the top of the “hierarchy of hostility, followed by the Rus and the
Magyars…” (Howard-Johnston 2000, p. 314). Indeed, much of the opening chapters
of the De Aministrando Imperio is devoted to strengthening or rather better control-
ling the Pečenegs, Constantinople’s most important ally (followed by the Alans) and
controlling the Khazars (see the comments of Howard-Johnston 2000, pp. 312–313).
Clearly much had changed since the era of marital alliances. Finally, before leaving
the Vita Constantini, we may note that even a sacralised ruler or one increasingly be-
coming so, was still the “head of state” and his ceremonial presence at a high level
diplomatic meeting or at least an audience with him would have probably comported
with the diplomatic usages of the age.
Attempts have also been made to connect the transformation of the Qağan into
a talismanic, sacral king with the conversion of the Khazar elite (and eventually other
strata of society) to Judaism and with the Qabar revolt, which, according to some,
was a reaction to the conversion.77 Both of these events, not necessarily connected,
have produced an extensive and often highly speculative literature which need not
detain us here. What we can say, with some assurance, is that according to al-Mas‘û-
dî, a reliable source who remarks that he composed a work on this event, the conver-
sion of the Khazar elite to Judaism took place during the reign of Hârûn al-Rašîd
(786–809). This is now confirmed by Khazar coinage of 223/837–8: the “Mûsâ
coins” – Khazar imitation Arab dirhams with the inscription Mûsâ rasûl Allâh
“Moses is the Messenger of God”, a re-shaping of the customary Muslim formula
with Moses replacing Muḥammad and hence a clear signaling of the Khazars pre-
senting an official Judaic profile to the world. The conversion, a top – down process,
undoubtedly took place in stages. The coins of 223/837–8, as Roman Kovalev sug-
gests, most probably commemorated a concluding stage of the process, perhaps the
conversion of the Qağan himself (al-Mas‘ûdî 1966–1979, I, p. 212).78 Our informa-
tion on the Qabar revolt on which so many theories regarding the conversion and sub-
sequent Khazar institutional governance are constructed, comes only from Constantine
Porphyrogenitus. He writes that the Qabars, consisting of three clans, were of Khazar
origin (©pÕ tÁj tîn Caz£rwn gene©j), they revolted against the Khazar government,
were defeated and fled. They “came and settled with the Turks (i.e. the Hungarian
union) in the land of the Pechenegs.” Here, they were incorporated into the Hungar-

77
See the remarks of Ludwig (1982, pp. 168 – 175). Ludwig suggests that the transforma-
tion of the Qağan into a sacral figure took place towards the end of the 9th – early 10th century, the
end of a process that began in the 9th century.
78
On on the Khazar imitation Arab dirhams, see Kovalev (2004, pp. 97 – 129; 2005, esp.
pp. 226ff.). The coins, dated well before Zuckerman’s proposed dating of the conversion (after 861)
are strong evidence in favor of al-Mas‘ûdî’s notice. Kovalev (2005, pp. 233 – 234) posits the Beg’s
conversion early in the 9th century (in keeping with al-Mas‘ûdî’s notice) and that of the Qağan
“sometime in the 830s”. This, he argues, led to the minting of the special Khazar coins with their
strongly articulated national sense (arḍ al-Xazar “land of the Khazars” inscribed on them) and overt
Judaism as reflected in the “Moses coins”. This, then, would mark the conversion of the Qağan.
Acta Orient. Hung. 60, 2007
184 P. B. GOLDEN

ian tribal union, learned Hungarian and taught the Hungarians their tongue. They re-
mained bilingual into Constantine’s time. Constantine does not tell us when the re-
volt occurred. Presumably, it was before the Hungarian union left for Pannonia
(895). The Qabars are already noted as part of Hungarian forces raiding Central
Europe in 881 (Annales, p. 742). The “land of the Pechenegs,” most probably refers
to the Pontic steppes, formerly under Hungarian control and in Constantine’s day
now in possession of the Pečenegs. In other words, the Qabar revolt took place
mostly probably sometime after 800 and certainly before 881. The Qabar revolt has
been connected with what were undoubtedly competing religious orientations that
had developed at the Khazar court (Judaic, Christian and Muslim). The Soviet ar-
chaeologist Artamonov attempted to connect the sacralisation of the Khazar Qağa-
nate specifically to the conversion to Judaism. The weakening of Qağanal power is
viewed as beginning with the 737 defeat administered to the Qağan by the Umayyad
Marwân, a defeat which forced the Qağan to convert to Islam – at least temporarily –
although some members of the dynasty are later (see al-Iṣṭaxrî 1927, p. 224; Dunlop
1954, pp. 97–98) depicted as Muslims and as such barred from succeeding to the Qa-
ğanate which had become limited only to those professing Judaism. Artamonov argues
that the conversion to Judaism, although useful as a “demonstration of not only the
independence, but the parity of the Khazar Qağanate with the Byzantine Empire and
Arabian Caliphate”, was, in the long run, fatal for a dynasty of “foreign” origin.
Viewing Judaism as a “national religion”, he declared it unable to unify the “multi-
tribal population” of Khazaria. Judaism, he avers, “underscored the class contradic-
tions and separated the government that professed it from the people”. On the basis
of this and the remark by the Khazar ruler, Joseph, in his letter to Ḥasdai b. Šaprûṭ,
that one of his ancestors, Obadiyah, a successor of Bulan (who brought the Khazars
or at least their ruling groups to Judaism) “renewed the faith, …built synagogues and
schools of learning and gathered the wise men of Israel” (Kokovcov 1932, Heb. text,
pp. 21–24, 28–31; Russ. trans. pp. 75–80, 92–97). Artamonov leapt to the conclu-
sion that Obadiyah had seized power and forced Judaism on the Qağan. This had
produced a civil war in the early 800s which led to the departure of those opposed to
the pro-Judaic orientation (the Qabars) and a coup d’etat which removed the Qağan
from the actual reins of government, i.e. his sacralisation which effectively stripped
him of power. This view became axiomatic in much of Soviet scholarship and was
accepted uncritically elsewhere as well (Artamonov 1962, pp. 278–280, 324ff.).79

79
Some aspects of this thesis have been followed by Pletnëva (1986, p. 62); Pritsak (1978,
pp. 278 –280) and others. Novosel’cev (1990, pp. 137 – 142) offers a slightly different version. In
his view, the Išad / Beg / Yilig took the lead in the conversion to Judaism. Following Artamonov, he
views the Qağanal institution as weakening during the lengthy period of wars with the Arabs and
after the defeat of 737 (when the Umayyad Marwân briefly captured the Qağan and forced him to
convert to Islam – the Qağan apostacised once the Arab armies withdrew). The Išad gradually
pushed out the Qağan as the effective ruler and then took the title of Beg. By the first third of the
9th century, there was indeed dual power in Khazaria and by the 10th century, the Qağan had lost
all effective power which was now concentrated in the hands of the Beg. This power struggle he
connects with the triumph of Judaism, led by the Beg, among a section of the Khazar aristocracy.
A critique of the Artamonov – Pletnëva thesis is given by Golden (1983, pp. 144ff.). The whole thesis
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 185

It is certainly not impossible that conversion to a new religion provoked some

internal discord. This is hardly unknown in world history, but there is nothing to con-
nect the Qabar revolt with religious issues. Moreover, a Qabar archaeological site in
Yugoslavia (Čelarevo) contains Judaic graffiti, indicating that, at the very least, ele-
ments of the Qabars were Judaic (Erdeli 1983). There are other, more nuanced vari-
ants on this theme. Róna-Tas, for example, while noting that the “leading stratum” of
the Qabars may have been connected with Judaism, nonetheless sees the joining of
the Qabars to the Hungarian union as “not wholly independent from the conversion
of the leading Khazar strata to [the] Jewish faith.” It may have been “indirectly…
provoked by the conversion to the Jewish faith,” a move which he suggests may have
“interfered with imperial and political interests.” He further notes, however, that the
Judaised Khazar elite lived peacefully alongside Muslims, Christians and pagans
who also inhabited the state, a fact well-attested in our sources. Khazar beliefs,
monotheistic and otherwise, were “complex and syncretistic.” (Róna-Tas 1999, pp.
348–349). I think Róna-Tas’s view is closer to the mark. The conversion to Judaism
had negative consequences in terms of Khazar–Byzantine relations. These ultimately
soured and by the mid-10th century had become overtly hostile. There may well have
been a pro-Byzantine faction at the Khazar court and they, in all likelihood, would
have opposed this move. However, we have no evidence of Christianity (whose im-
portance in Khazaria is often neglected, see Róna-Tas 1999, p. 233) among the Qabars
and are equally ignorant of Khazaria’s palace politics. Even when the Khazars pre-
sented a Judaic profile to the world, we must remember that the chief vezir of the
Qağanate was Muslim in the 10th century and Muslims played a key role in the com-
mercial life of the state. Khazar religious tolerance, noted by a number of sources,
was not only typical of the pre-Islamic steppe world, but also made good domestic
political sense. We should also bear in mind that this was still the steppe world and
religious syncretism – as in much of Central Eurasia – was commonplace. Religious
wars, until the coming of Islam in a significant way, seem very unlikely in such an
environment. Finally, returning to our theme, why would conversion to Judaism – or
Christianity or Islam for that matter – have resulted in the sacralisation of the Qağan?
Judaism, and the other Abrahamic faiths, would hardly have been in support of a
sacral kingship that recognised the Qağan as a semi-divine figure.80 The ruler as
God’s anointed comports with the Judeo-Christian tradition, even ’ισοαπόστολος – if
you are a Byzantine emperor, but this is not what our Khazar ruler is. Róna-Tas has
noted that sacral kingship is “typical” of “Tengrism” (Róna-Tas 1996, p. 127; 1999,
p. 148). In Khazaria, the Tengri religion (Tengri = supreme celestial deity in the
Altaic world, see Roux 1956, pp. 49–82; 1958, pp. 32–66; 1984, pp. 122–124), with
its emphasis on the qut (heavenly good fortune) of the ruler, was a tradition that had

————
rests on conjecture, underpinned by Artamonov’s negative view of Judaism – a barely disguised
Antisemitism – which he saw as undermining the Khazar state (Artamonov 1962, pp. 457– 458).
On the Antisemitic theme in Soviet Khazar historiography, see Shnirelman (2002).
80
Petruxin (2001, pp. 74 – 75) notes this as one of the “paradoxes” of the Khazar sacral
Qağanate.
Acta Orient. Hung. 60, 2007
186 P. B. GOLDEN

evolved over time and by the 10th century had attained a particularly “advanced”
status. Róna-Tas further suggests that the longer a dynasty was associated with qut,
the more removed it became from the actual, day to day governance of the state. This
feature of the institution was not limited to the steppe world (Róna-Tas 1996, pp.
127–129, 269; 1999, pp. 148–151, 342–343). Indeed, the Merovingian–Carolingian
model in the West comes to mind. This is an interesting and not implausible idea, but
outside of the Khazar milieu, we have no other instances of this in the Inner Asian
steppe world. Moreover, one would like to be able to point to concrete circumstances
that brought about this transformation. We shall return to this point. But first some-
thing should be said about the complexity of the titles born by the lesser “king.” Do
Išâd, Yelig/Yilig and Beg all denote the same institution of the “active” ruler? As
noted previously, Išâd stems from Soġdian (see Bombaci 1974, see also Golden
1980, I, pp. 206–208)81 ixšēδ or more probably from the form ‫[ اﺧﺸﺎذ‬ʼxšâδ noted by
al-Ṭabarî] (al-Ṭabarî 1967– 1969, IV, pp. 473–476) or perhaps from Khwârazmian,
another East Iranian tongue. Soġdians were the great agents of cultural exchange
along the Silk Road. The Khwârazmians were the main link between the Silk Road,
the Islamic world and the western Eurasian steppe and forest peoples. In addition,
Khwârazmians were an important component of the Khazar court and formed the
personal guard (Ors) of the Qağans in Atıl/İtil. The Türk title Šad, which is related to
our Khazar term, as reflected in the Old Türk and Uyğur texts, as well as Chinese ac-
counts, was invariably borne by members of the ruling house. It is very likely, thus,
that the Khazar Išad was also a member of the ruling house, as would have been
consonant with Türk or Old Turkic tradition, and hence there can be no talk of a
change of dynasty. We do not find the title Šad in use beyond the Türk and its suc-
cessor state, the Uyğurs. It is not found among the later Qarluqs, Oğuz – although it
does appear in the origin tale of the Kimek as a personal name (Gardîzî/Ḥabîbî 1984,
pp. 549–550; see also discussion in Golden 1992, pp. 202–203). However, titles are
commonly used as anthroponyms in the Turkic world (Németh 1991, p. 277). The Išad
notices may mark an earlier stage of Khazar organisational history, one that was
typical of the Türk Empire as well. The Išad may not have been the actual ruler, but
as in the Türk realm simply a relative of the Qağan who led the army into combat or
carried out other important activities entrusted only to close family members. The
Türk officer who led the Türk allies of Heraclius against the Sâsânids bore the title
Šad, as we know from Movsês Dasxuranc’i and Kirakos of Gandza (Movsês Dasxu-
ranc’i 1962, p. 88; see also discussion in Golden 1980, I, pp. 206ff.). Šad was later
conflated with the Beg. Similarly, the same may be implied in the other title Yelig
(‫ )ﻳﻠﻚ‬noted in our sources (yilig/yélig [yéllig?] “(junior) king, prince” – although this
title is more neutral in this regard. Beg was, most certainly, a lower title, one reserved
for clan chieftains. The Khazar Beg or Qağan Beg probably points to this office now
being held by a non-Ashina and one, as we have seen, who is functioning as the ac-
tual ruler of the state.

81
The Zhou-shu (see Liu Mau-tsai 1958, I, p. 8) lists the titles Ye-hu (yabğu), Shê (Šad)
and T’ê-lê (Tegin) as the highest ranks.
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 187

Within the nomadic Turkic world, we have few instances of a Khazar-style sac-
ral kingship. The Kende–Gyula dual kingship of the pre-Christian Hungarian union,
if it is indeed akin to it, was most probably taken from the Khazars (see Kristó 1996,
pp. 171–172).82 Kende or *kündä is clearly of Khazar origin (kündü, a term that is
not without problems nonetheless as it is unattested in Turkic, but found in Mongolic
and Tunguso-Manchu).83 Cf. also Alanic karkundâj. The etymology of gyula/ǰula
remains problematic (Róna-Tas 1999, pp. 150, 347). In the anonymous 12th century
Risâlat al-Aqâlim, there is a notice on the Toquz Oğuz (this is in reference to the
post-840 Uyğurs) rulers that describes a similarly tabuised monarch who reigns but
does not rule.84 The notice is late, of uncertain authenticity and not confirmed by any
other source. The Chinese sources make reference to the “Yi ke-han” of the Türks,
the “great families that live at home and do not rule.” These were below the yabğus
(Liu Mau-tsai 1958, II, pp. 498–499.).85 This term may reflect Turkic *eb Qaghan
“house Qaghan”, perhaps a reference to a tabuised, ritually isolated member of the
charismatic ruling house. But, this is speculation. The Chinese accounts do not make
much of these home-bound rois fainéants – whatever they may have been doing, or
not doing. The fact remains, that it is hard to find anything like the Khazar sacral ruler
elsewhere in the Turko-Mongolian steppe world. The Pre-Christian Hungarians con-
tinued, perhaps, something akin to it for a while in their new Pannonian homeland,
but this may be nothing more than the retention of an important steppe tradition asso-
ciated with the Khazars and with the ruling house. When the Türgeš largely supplanted
the Ašina as the dominant element among the On Oq (see Beckwith 1993; Chavannes
1903),86 the Ašina remained, nonetheless, active – albeit occasionally proxies of other
powers (e.g. China, Tibet). They were not tabooised and set aside. The closest parallel
we have to the Khazar sacral ruler is found in the Iranian tradition. Here, the ruler
was considered to be of divine origin or at the least divinely chosen and hence his
person was holy, inviolable, and unapproachable. He possessed divine good fortune
(xwârəna), often depicted as a “nimbus of fire surrounding his head” (Widengren
1959, p. 245).87 The king was rarely seen by the public and was largely invisible
even to court. He was often veiled. The royal countenance was of such power (deriv-
ing from his kinship with the sun and moon), that direct sight of his visage could
cause one to be consumed by fire. Hence, those who approached the Šâh, covered
their faces and said misûzam “I am burning up”. One prostrated oneself before the
Šâh. The ruler was a cosmic figure, the Sâsânian “lord of the seven climes” and “lord

82
See also discussion, with somewhat differing view in Róna-Tas (1996, pp. 269 – 272;
1999, pp. 343 – 347).
83
See Golden (2005, pp. 213 – 214), where I have revised my earlier views, cf. Golden
(1980, I, pp. 200 – 204). Clearly Kündü [‫ آﻨﺪر‬recte: ‫ آﻨﺪو‬kndw] is also to be connected with
*kündäčik, transcribed as ‫ آﻨﺪاﺟﻴﻖ‬kndâjîq.
84
See text in al-Jâḥiẓ (1967, Arabic text, p. 35, Turk. trans. p. 33); see also Ibn Faḍlân
(1939, pp. 263, 268); Dunlop (1954, p. 39).
85
Liu Mau-tsai suggests the reading *äb for Chin. i (yi); Taşagıl (1995, p. 114).
86
A full study of the Türgeš remains a desideratum of early Turkic studies.
87
See Choksy (1988, pp. 35, 37): “the king was at all times considered a mortal and never a
god incarnate”, rather, his kingship was “god-given and sacred”.
Acta Orient. Hung. 60, 2007
188 P. B. GOLDEN

of the age” (Widengren 1959, pp. 245–250; Choksy 1988, pp. 42, 44). The Šâh did
not physically take part in warfare. The Šâh also had priestly functions and was the
final word in religious matters. At his death and burial, human sacrifices were some-
times performed and a mausoleum was erected. Sound a bit familiar?
Although Sâsânid influences could have made their way into what eventually
became Khazaria, in particular after the close encounters the Western Türks had with
Iran during the Byzantine– Iranian war of the early 7th century, it is doubtful that this
was the immediate source or model for the Khazar sacral kingship.88 Why the time
lapse? Iranian influences also came directly from the Iranian steppe nomads such as
the Alano-As tribes that were still present in the Ponto-Caspian steppelands. Alans
constituted the prevailing ethnic element in the Majackoe archaeological complex in
the northwestern part of Khazaria (Vinnikov – Pletnëva 1998, pp. 209–210; Pletnëva
1999, p. 224). But, there is nothing in the Alanic tradition to point to sacral kingship.
I think there are two factors at work here. One is that noted by Róna-Tas: it is
possible that as the Khazar ruling house – and other dynasties elsewhere – but not in
the steppe, gained in longevity, their holy status was emphasised and they retreated
further into ritual status. But, we have no evidence for this having occurred with the
Türks, the progenitors of the Khazars, nor with the long-lived Qarakhanid Qağans
(late 10th–early 13th centuries) who may also have been of Türk-Ašina origin – al-
though Islam would have been a mitigating factor here. Another, perhaps more im-
portant, immediate factor in Khazaria, may have been the Khwârazmian guard, the
Ors and their vezirs. The Ors are the Auruša ('Aorso„ > Oset. Ors/Urs) who were
earlier associated with the region of Khwarazm (Christensen 1971; Golden 1990).
The Khwârazmšâhs were considered descendants of the mythical, hero and demigod,
Siyâvuš (al-Birûnî 1923, p. 35/Russ. p. 47; Tolstov 1948, pp. 74–75, 83). None of the
local rulers in neighbouring Soghdia, bore the title šâh (cf. the Ixšîdh of Farghâna, the
Afšîn of Usrushâna, the tudun of Shâsh, the ṭarxûn of Samarqand, the buxâr-xudât of
Bukhârâ). Shâhs are found in Sharwân, Kabul and Tirmidh (al-Birûnî 1923, pp. 100–
102/Russ. pp. 111–112). Al-Birûnî remarks that Qutayba b. Muslim,when he con-
quered Khwârazm for the second time, made Askajamuk, a descendant of the local
royal house (Afriġids) the Shâh and adds “the rule of this country left the hands of
the descendants of Khusraw, but the title of Shâh remained with them since it was
hereditary. The governance (of Khwârazm) was sometimes in the hands of this family
and at other times in the hands of others until the title of ruler and the dignity of Shâh
left them…” (al-Birûnî 1923, pp. 35–36/Russ. p. 47). This occurred in 995 with the
death of the last of the Afriġids. Thus, it would appear that, at the least, a ceremonial
if not completely sacralised ruler, may have already been present in Khwârazm since
the first half of the 8th century with the Islamic conquest. There is much we still do
not know about the Afriġid Khwârazmshâh state. The kings may have continued to
play some kind of active role. Thus, al-Birûnî mentions fġbryx¸ the first day of the
month Axšrwyry in which the “kings of Khwârazm” rode out because it was the be-
ginning of the cold season and they now wintered outside of their summer abodes,

88
On these encounters, see Noonan (1982, pp. 269 – 302).
Acta Orient. Hung. 60, 2007
 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 189

“driving away the Oğuz Turks from their borders and protecting the frontiers of their
realm from them” (al-Birûnî 1923, p. 236/Russ. p. 257). If this is, indeed, a reference
to the Oğuz, then it must refer, at the earliest, to the last quarter of the 8th century
when the Oğuz first arrived in the region. Or, it may simply be a reference to a cere-
monial function of the ruler with the actual military activity being conducted by
others. We simply do not know. What is clear, however, is that the Ors came from
this old Iranian Khwârazmšâh tradition. Judging from al-Birûnî’s writings, even with
the Islamisation of the region, old traditions expressed in holidays etc. were hardly
forgotten. Indeed, in some instances, they lasted into relatively recent times (Tolstov
1948, p. 87). The Ors were resident in Khazaria from the relatively early days of the
kingdom. They had emigrated there after drought and epidemic had forced them out
of their lands in Khwârazm sometime “after the appearance of Islam.” They were
granted the privilege of not having to participate in wars on their coreligionists when
the Khazars were fighting Muslims (al-Mas‘ûdî 1966– 1979, I, p. 213). Clearly, as
hereditary ministers, the source of manpower for the standing army that surrounded
the rulers and most probably the largest group of Muslims in the capital (in which
Muslims, in all likelihood, constituted the majority), they would have been able to
influence court practices – even with notions of kingship that were decidedly un-
Islamic. Many of the Türk practices, beliefs about royal authority, court titles and
ranks show Iranian influences (cf. qut and xwârəna, yabğu, šad etc.) – stemming
from a variety of sources. Finally, as studies elsewhere have suggested, sacral king-
ships develop in societies where there has been considerable “societal overlap,” where
one group has imposed itself on one or more now subordinate groups (Czeglédy
1966, p. 15). The Khazar state with its layerings of peoples provides a good illustra-
tion of that. A more exalted kingship, especially in an aging state, provided the neces-
sary social glue, particularly for the Turkic steppe element, to hold an otherwise dis-
parate society of nomads, semi-nomads, agriculturalists, and trapping – hunting forest
populations under control.

Bibliography

Akman, M. (1997): Osmanlı Devletinde Kardeş Katli. Istanbul.

al-Azmeh, A. (2004): Monotheistic Kingship. In: Al-Azmeh – Bak (2004), pp. 9 – 29.
al-Azmeh, A. – Bak, J. (eds) (2004): Monotheistic Kingships. The Medieval Variants. Budapest.
Ananias of Širak (1992): The Geography of Ananias of Širak. Ed., trans. R. H. Hewsen. Wiesbaden.
Annales iuvavenses: Annales ex annalibus iuvavensibus antiques excerpti. Ed. H. Bresslau (Monu-
menta Germaniae Historica, scriptorum tomi XXX, pars II). Lipsiae 1834.
Artamonov, M. I. (1962): Istorija Xazar. Leningrad.
al-Bakrî (1992): Kitâb al-Masâlik wa’l-Mamâlik. Ed. A. P. Van Leeuwen and A. Ferre. Qarṭâj,
Tunis, 2 vols.
al-Birûnî (1923): Atâr al-Baqiyya / Chronologie orientalischer Völker. Ed. C. E. Sachau. Leipzig,
reprint: Baghdad (no date).
Beckwith, Christopher I. (1993): The Tibetan Empire in Central Asia: A History of the Struggle for
Great Power among Tibetans, Turks, Arabs, and Chinese during the Early Middle Ages.
Princeton.

Acta Orient. Hung. 60, 2007

190 P. B. GOLDEN

Bernštejn, S. B. (1984): Konstantin-filosof i Mefodij. Moskva.

Berta, Á. (1992): Die chasarische Benennung der Ungarn. In: Fragner, Ch. – Schwarz, K. (eds):
Festgabe an Joseph Matuz. Osmanistik – Turkologie – Diplomatie. Berlin, pp. 7 – 11.
Bičurin, N. J. (1950): Sobranie svedenij o narodax obitavšix v Srednej Azii v drevnie vrmena.
Moskva – Leningrad, 2 vols.
Bombaci, A. (1974): On the Ancient Turkish title šaδ. In: Gururâjamañjarika. Studi in Onore di
Giuseppe Tucci. Napoli, vol. I, pp. 176 – 193.
Chavannes, E. (1903): Documents sur les T’ou-kiue (Turcs) Occidentaux. St. Petersbourg– Paris
(reprint: Taipei 1969).
Choksy, J. (1988): Sacral Kingship in Sasanian Iran. Bulletin of the Asia Institute N.S. 2, pp. 35 –
52.
Christensen, A. (1971): L’Iran sous les Sassanides. Copenhagen (2nd ed. 1994, reprint: Osna-
brück).
Clauson, Sir Gerard (1972): An Etymological Dictionary of Pre-Thirteenth Century Turkish. Oxford.
Clauson, Sir Gerard (1975): The Foreign Elements in Early Turkish. In: Ligeti, L. (ed.): Researches
in Altaic Languages. Papers Read at the 14th Meeting of the Permanent International Al-
taistic Conference held in Szeged, August, 1971. (Bibliotheca Orentalis Hungarica, 20) Bu-
dapest, pp. 43 – 49.
Const. Porph.: Constantine Porphyrogenitus: De Administrando Imperio. Ed. Moravcsik, Gy., trans.
Jenkins, R. J. H. = Corpus Fontium Historiae Byzantinae. Washington, D.C. 1967.
Curta, F. (ed.) (2005): East Central Europe and Eastern Europe in the Early Middle Ages. Ann
Arbor.
Czeglédy, K. (1966): Das sakrale Königtum bei den Steppenvölkern. Numen 13, pp. 14 – 26.
Czeglédy, K. (1974): A szakrális királyság a steppei népeknél. Magyar Nyelv 70, pp. 11 – 17.
Czeglédy, K. (1983): From East to West: The Age of Nomadic Migrations in Eurasia. Trans. P. B.
Golden. AEMA 3, pp. 25 –125.
Dobrovits, M. (2004): A nyugati türkök első uralkodójáról. Antik Tanulmányok 48, pp. 111 – 114.
Dozy, R. (1881): Supplément aux dictionnaires arabes. Leiden. 2 vols.
Dunlop, D. M. (1954): The History of the Jewish Khazars. Princeton.
Dvornik, F. et al. (1962): Constantine Porphyrogenitus. De Administrando Imperio. Volume II.
Commentary. London.
Erdal, M. (2004): A Grammar of Old Turkic. Leiden – Boston, Brill. (Handbook of Oriental Studies.
Section Eight: Central Asia. Ed. Denis Sinor, Nicola Di Cosmo. Vol. 3)
Erdeli (Erdélyi), I. (1983): Kabary (Kavary) v Karpatskom basseine. Sovetskaja Arxeologija 4,
pp. 174 – 181.
Gardîzî / Barthold: Bartol’d, V. V.: Izvlečenie iz sočinenija Gardizi, Zajn al-axbâr. In: Bartol’d,
V. V.: Sočinenija 8. Moskva 1963 – 1977, pp. 23 – 62.
Gardîzî / Ḥabîbî: Târîx-i Gardîzî. Ed. Ḥabîbî, ‘A. Tehran 1363/1984.
Gardîzî / Martinez: Martinez, A. P.: Gardîzî’s Two Chapters on the Turks. AEMA 2 (1982),
pp. 109 – 217.
Geary, P. (2002): The Myth of Nations. The Medieval Origins of Europe. Princeton.
Gillett, A. (ed.) (2002): On Barbarian Identity. Critical Approaches to Ethnicity in the Early
Middle Ages. (Studies in the Early Middle Ages, 4) Turnhout, Belgium, Brepols.
Gnoli, G. (1999): Far(ah). In: Encyclopaedia Iranica. New York, Biblioteca Persica Press, IX,
pp. 312 – 319.
Göckenjan, H. – Zimonyi, I. (2001): Orientalische Berichte über die Völker Osteuropas und Zentral-
asiens im Mittelalter. Die Ğayhânî-Tradition. Wiesbaden.
Golb, N. – Pritsak, O. (1982): Khazarian Hebrew Documents of the Tenth Century. Ithaca.

Acta Orient. Hung. 60, 2007

 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 191

Golden, P. B. (1975): The Q’azar-Hungarian Title/Personal Name ‫ ﻳﻠﻚ‬/ ’Ιέλεχ. AEMA 1, pp. 37 – 43.
Golden, P. B. (1980): Khazar Studies. Budapest. 2 vols.
Golden, P. B. (1983): Khazaria and Judaism. AEMA 3, pp. 127 – 156.
Golden, P. B. (1990): Cumanica III: Urusoba. In: Sinor, D. (ed.): Aspects of Altaic Civilization III.
Bloomington, pp. 39 – 46.
Golden, P. B. (1992): An Introduction to the History of the Turkic Peoples. Wiesbaden.
Golden, P. B. (2001): Some Notes on the Comitatus in Medieval Eurasia with Special Reference to
the Khazars. Russian History/Histoire Russe 28 /1 – 4, pp. 153 – 170.
Golden, P. B. (2005): Khazarica: Notes on Some Khazar Terms. Turkic Languages 9 / 2, pp. 205 –
222.
Golden, P. B. (2006): Sacral Kingship Among the Khazars. In: Tracy, J. – Reyerson, K. – Stavrou,
Th. (eds): Pre-Modern Russia and Its World. (Schriften zur Geistesgeschichte des östlichen
Europas). Wiesbaden, Harrassowitz Verlag, pp. 79 – 102.
Grigor’ev, A. P. (1981): Oficial’nyj jazyk Zolotoj Ordy XIII – XIV vv. In: Tjurkologičeskij Sbornik
1977. Moskva, pp. 81 – 89.
Gumilëv, L. N. (1967): Drevnie tjurki. Moskva.
Györffy, Gy. (1994): Dual Kingship and the Seven Chieftains of the Hungarians in the Era of the
Conquest and the Raids. AOH 47, pp. 87– 104.
Head, C. (1972): Justinian II of Byzantium. Madison, Wisconsin.
Hocart, J. (1970): Kings and Councillors. Cairo, 1936; reprint: 1970 Chicago.
Howard-Johnston, J. (2000): The De Administrando Imperio: A re-examination. In: Kazanski, M.
et al. (eds): Les centres proto-urbains russes entre Scandinavie, Byzance et Orient. Paris,
pp. 301 – 336.
Ḥudûd /Sutûdah (1340/1962): Ḥudûd al-‘Âlam. Ed. M. Sutûdah. Tehran.
Ḥudûd /Minorsky (1970): Minorsky, V.: Ḥudûd al-‘Âlam. “The Regions of the World”. London,
1937; rev. ed.: 1970.
Ibn al-Balkhî (1921): Fârsnâma. Ed. Le Strange, G. – Nicholson, R. A. London.
Ibn Faḍlân (1939): Togan, A. Z.: Ibn Faḍlân’s Reisebericht. (Abhandlungen für die Kunde des
Morgenlandes, XXIV/3) Leipzig.
Ibn Ḥawqal (1992): Kitâb Ṣûrat al-Arḍ. Beirut.
Ibn Rusta (1892): Kitâb al-A‘lâq an-Nafîsa. Ed. De Goeje, M. J. Leiden.
Ibn Xurdâdhbih (1889): Kitâb al-Masâlik wa’l-Mamâlik. Ed. De Goeje, M. J. Leiden.
al-Isṭaxrî (1927): Kitâb Masâlik al-Mamâlik. Ed. De Goeje, M. J. 2nd ed.: Leiden.
al-Jâḥiẓ (1967): el-Câḥiẓ: Hilâfet Ordusunun Menkıbeleri ve Türkler’in Fazîletleri. Ed. Turk. trans.
Şeşen, R. Ankara.
Janhunen, J. (1996): Manchuria. An Ethnic History. Helsinki.
K’art’lis C’xovreba (1955): K’art’lis C’xovreba I. Ed. Qauxč’išvili, S. T’bilisi. – See also Thomsen
(1996).
Klyashtornyi, S. G. (1983): The Terkhin Inscription. AOH 36, pp. 356 – 366.
Klyashtornyi, S. G. (1985): The Tes Inscription. AOH 39, pp. 137 – 156.
Klyashtorny, S. G. (1997): About One Khazar Title in Ibn Faḍlân. Manuscripta Orientalia 3/3,
pp. 22 – 23.
Klyashtorny, S. G. (1994): The Royal Clan of the Turks and the Problem of Early Turkic – Iranian
Contacts. AOH 47, pp. 445 – 447.
Kljaštornyj, S. G. – Savinov, D. S. (2005): Stepnye imperii drevnej Evrazii. Sankt Peterburg:
Filologičeskij fakul’tet Sankt-Peterburgskogo gosudarstvennogo universiteta.
Kmoskó, M. (1997): Mohamedán írók a steppe népeiről. I / 1. Budapest.
Kokovcov, P. K. (1932): Evrejsko-xazarskaja perepiska v X veke. Leningrad.

Acta Orient. Hung. 60, 2007

192 P. B. GOLDEN

Konow, S. (ed.) (1929): Kharoshṭhî Inscriptions with the Exception of those of Aśoka. (Corpus In-
scriptionum Indicarum, vol. II, part 1) Calcutta.
Kovalev, R. K. (2004): What does Historical Numismatics Suggest About the Monetary History of
Khazaria in the Ninth Century? Question Revisited. AEMA 13, pp. 97 – 129.
Kovalev, R. K. (2005): Creating Khazar Identity Through Coins: The Special Issue Dirhams of
837/8. In: Curta (2005), pp. 220– 251.
Köprülü, M. F. (1940): La proibizione de versare il sangue nell’esecuzione d’un membro dell
dinastio presso i Turchi et i Mongoli. Annali del R. Istituto Cuppericre Orientale di Napoli,
n.s. I, pp. 15 – 23. – Cf. Köprülü (1944).
Köprülü, M. F. (1944): Türk ve Moğol sülâlesinde hânedan âzasının idâmında kan dökme
memnuiyeti. Türk Hukuk Tarihi Dergisi 1, pp. 1 – 9. – Cf. Köprülü (1940).
Kračkovskij, I. Ju. (1955 – 1960): Arabskaja geografičeskaja literatura. In: Kračkovskij, I. Ju.: Iz-
brannye sočinenija. IV. Moskva.
Kristó, Gy. (1996): Hungarian History in the Ninth Century. Trans. Novák, Gy., Kelly, E.. Szeged.
Lehr-Spławiński, T. (1959): Żywoty Konstantyna i Metodego (Obszerne). Poznań.
Lessing, F. et al. (1982): Mongolian – English Dictionary. Bloomington, Ind.
Lewicki, T. (1956 – 1988): Źródła arabskie do dziejów Słowiańszczyzny. Wrocław– Kraków– War-
szawa. 4 vols.
Łewond: History of Lewond, The Eminent Vardapet of the Armenians. Trans., commentary, Z. Ar-
zoumanian. Philadelphia, 1982.
Ligeti, L. (1986): A magyar nyelv török kapcsolatai a honfoglalás előtt és az Árpád-korban. Buda-
pest.
Litvinskij, B. A. (1968): Kangjujsko-sarmatskij farn (k istoriko-kul’turnym svjazjam plemën južnoj
Rossii i Srednej Azii. Dušanbe.
Liu Mau-tsai (1958): Die chinesischen Nachrichten zur Geschichte der Ost-Türken (T’u-küe). Wies-
baden. 2 vols.
Ludwig, D. (1982): Struktur und Gesellschaft des Chazaren-Reiches im Licht der Schriftlichen
Quellen. Münster.
Marquart, J. (1903): Osteuropäische und ostasiatische Streifzüge. Leipzig.
Márton, A. (1997): Szakralitás és hatalom a türköknél. AETAS 2 – 3, p. 78.
al-Mas‘ûdî (1966 – 1979): Murûj ad-Dahab wa Ma‘âdin al-Jawhar. Ed. Pellat, Ch. Beirut, 7 vols.
Melikset-Bek, L. M. (1960): Xazary po drevnearmjanskim istočnikam. In: Issledovanija po istorii
kul’tury narodov Vostoka. Sbornik v čest’ akademika I. A. Orbeli. Moskva – Leningrad.
Menandros: Menander Protector: The History of Menander the Guardsman. Ed., trans. Blockley,
R. C. Liverpool, 1985.
Mixeev, V. K. (1985): Podon’e v sostave xazarskogo kaganata. Xar’kov.
Moriyasu, T. – Ochir, A. (eds) (1999): Provisional Report of Researches on Historical Sites and
Inscriptions in Mongolia from 1996 to 1998. Osaka.
Moses Khorenats’i (1978): History of the Armenians. Trans. Thomsen, R. W. Cambridge, Mass.
Movsês Dasxuranc’i (1962): The History of the Caucasian Albanians. Trans. Dowsett, C. London.
al-Muqaddasî (1906): Aḥsan at-Taqâsîm fî Ma‘rifat al-Aqâlîm. Ed. de Goeje, M. J. Leiden, 2nd ed.
Németh, Gy. (1930): A honfoglaló magyarság kialakulása. Budapest, revised 2nd ed.: Budapest
1991.
Nikephoros: Nikephoros Patriarch of Constantinople: Short History. Ed., trans. Mango, C. (Corpus
Fontium Historiae Byzantinae, XIII) Washington, D.C. 1990.
Noonan, Th. S. (1982): Russia, the Near East, and the Steppe in the Early Medieval Period: An
Examination of the Sasanian and Byzantine Finds from the Kama – Urals Area. AEMA 2 pp.
269 – 302.

Acta Orient. Hung. 60, 2007

 IRANO-TURCICA: THE KHAZAR SACRAL KINGSHIP REVISITED 193

Noonan, Th. S. (1998 – 1999): The Khazar–Byzantine World of the Crimea in the Early Middle
Ages: The Religious Dimension. AEMA 10, pp. 207 – 230.
Novosel’cev, A. P. (1990): Xazarskoe gosudarstvo i ego rol’ v istorii Vostočnoj Evropy i Kavkaza.
Moskva.
Petruxin, V. Ja. (2001): K voprosu o sakral’nom statuse xazarskogo kagana: tradicija i real’nost’.
Slavjane i ix Sosedi 10, pp.73 – 78.
Petrukhin, V. Ja. (2004): A Note on the Sacred Status of the Khazarian Khagan. In: Al-Azmeh–Bak
(2004), pp. 269 – 275.
Pletnëva, S.A. (1986): Xazary. Moskva, 2nd ed.
Pletnëva, S. A. (1999): Očerki xazarsksoj arxeologii. Moskva – Ierusalim.
Pritsak, O. (1978): The Khazar Kingdom’s Conversion to Judaism. Harvard Ukrainian Studies 2,
pp. 261 – 281.
Pritsak, O. (1988): Turkological Remarks on Constantine’s Khazarian Mission in the Vita Constan-
tini. In: Farrugia, E. G. et al. (eds): Christianity Among the Slavs: The Heritage of Saints
Cyril and Methodius. Acts of the International Congress Held on the Eleventh Centenary of
the Death of St. Methodius, Rome, October 8 – 11, 1985. Rome.
Pritsak, O. (1990): The Pre-Ashkenazic Jews of Eastern Europe in Relation to the Khazars, the
Rus’ and the Lithuanians. In: Aster, H. – Potichnij, P. J. (eds): Ukrainian – Jewish Relations
in Historical Perspective. 2nd ed. Edmonton, pp. 3 – 21.
Pritsak, O. (1996): The Turcophone Peoples in the Area of the Caucasus from the Sixth to the Elev-
enth Century. In: Il Caucaso: Cerniera fra Culture dal Mediterraneo alla Persia (secoli
IV – XI). Settimane di Studio del Centro Italiano di Studi sull’Alto Medioevo, 43. Spoleto,
pp. 223 – 245.
PSRL: Polnoe sobranie russkix letopisej.
Pulleyblank, E. (1966): Chinese and Indo-Europeans. JRAS 1966, pp. 9 –39.
Pulleyblank, E. (1991): Lexicon of Reconstructed Pronunciation in Early Middle Chinese, Late
Middle Chinese, and Early Mandarin. Vancouver.
Rapp, S. H. Jr. (1997): Imagining History at the Crossroads: Persia, Byzantium, and the Architects
of the Written Georgian Past. PhD dissertation, University of Michigan.
Rašîd al-Dîn: Jâmi‘al-Tawârîx. Ed. Romaskevič, A. A. – Xetagurov, A. A. – Ali-zade, A. A. 2nd ed.
Moskva 1982, I/1.
Romašov, S. A. (2000 – 2001, 2004): Istoričeskaia geografija Xazarskogo kaganata (V–XIII vv.), I–
II. AEMA 11, pp. 219 – 338; IV–V. AEMA 13, pp. 185 – 264.
Róna-Tas, A. (1982): A kazár népnévről. NyK 84, pp. 349 – 379.
Róna-Tas, A. (1983): Újabb adatok a kazár népnév történetéhez. NyK 85, pp. 126 – 133.
Róna-Tas, A. (1996): A honfoglaló magyar nép. Budapest.
Róna-Tas, A. (1999): Hungarians and Europe in the Early Middle Ages. Budapest. (Revised Eng.
trans. of Róna-Tas 1996)
Roux, J. P. (1956, 1958): Tängri, essai sur le Ciel-Dieu des peuples altaïques. Revue de l’Histoire
des religions 149, pp. 49 –82; 154, pp. 32 – 66.
Roux, J. P. (1959): L’origine céleste de la souveraineté dans les inscriptions paléo-turques de Mon-
golie et de Sibérie. In: The Sacral Kingship, pp. 231 – 241.
Roux, J. P. (1984): La religion des Turcs et des Mongols. Paris.
Sacral Kingship (1959): The Sacral Kingship – La regalità sacra. Leiden.
Sebeos: The Armenian History attributed to Sebeos. Trans. Thomsen, R. W., historical commentary
by J. Howard-Johnston and T. Greenwood. Liverpool 1999. 2 vols.
Shapira, D. (1998 – 1999): Two Names of the First Khayar Jewish Beg. AEMA 10, pp. 231 – 241.

Acta Orient. Hung. 60, 2007

194 P. B. GOLDEN

Šahrestânîhâ î Êrânšahr: Šahrestânîhâ î Êrânšahr. A Middle Persian Text in Late Antique Geogra-
phy, Epic, and History. Ed. Daryaee, T. Costa Mesa, CA. 2002.
Šervašidze, I. N. (1989): Eščë raz ob étimologii imën Ašina i Attila-Avitoxol. Sovetskaja Tjurkolo-
gija No. 39, pp. 79 – 80.
Shepard, J. (1998): The Khazars’ Formal Adoption of Judaism and Byzantium’s Northern Policy.
Oxford Slavonic Papers 31, pp. 9 – 34.
Shnirelman, V. (2002): The Myth of the Khazars and Intellectual Antisemitism in Russia, 1970s –
1990s. Jerusalem.
Skrynnikova, T. D. (1992): Predstavleniia mongolov o sakral’nosti pravitelja. In: Kljaštornyj,
S. G. – Petrosjan, Ju. A. (eds): Tjurkskie i mongol’skie pis’mennye pamjatniki. Moskva,
pp. 71 – 85.
Sinor, D. (ed.) (1990): Aspects of Altaic Civilization. III. Bloomington.
Soudavar, A. (2003): The Aura of Kings. Legitimacy and Divine Sanction in Iranian Kingship.
Costa Mesa, CA.
al-Ṭabarî (1967 – 1969): Ta’rîx al-Ṭabarî. Ta’rîx al-Rasul wa’l-Mulûk. Ed. Abu’l-Faḍl Ibrâhîm.
Cairo, 10 vols.
al-Tannûxî: Kitâb al-Mustajâd. Ed. Pauly, L. Stuttgart 1939.
al-Ṭarṭûšî: Sirâj al-Mulûk. Cairo 1306/1888 – 1889.
Taşagıl, A. (1995): Gök Türkler. Ankara.
TESz: Benkő, L. (ed.): A magyar nyelv történeti-etimológiai szótára. Budapest 1967 – 1976. 3 vols.
Theophanes: Chronographia. Ed. de Boor, C. Leipzig 1883; reprint: Hildesheim 1971.
Thomson, R. W. (ed., trans.) (1996): Rewriting Caucasian History. The Medieval Armenian Adap-
tation of the Georgian Chronicles. The Original Georgian Chronicles and the Armenian
Adaptation. Oxford.
Tolstov, S. P. (1948): Po sledam drevnexorezmijskoj civilizacii. Moskva – Leningrad.
Toumanoff, C. (1963): Studies in Christian Caucasian History. Washington, D.C.
Vaissière, E. de la (2002): Histoire des marchands sogdiens. Paris.
Vinnikov, A. Z.– Pletnëva, S. A. (1998): Na severnyx rubežax xazarskogo kaganata. Majackoe
poselenie. Voronež.
Vlasto, A. P. (1970): The Entry of the Slavs into Christendom. Cambridge.
Waida, M. (1976): Notes on Sacred Kingship in Central Asia. Numen 23 / 3, pp. 179 – 190.
Wehr (1994)): The Hans Wehr Dictionary of Modern Arabic. Ed. Cowan, J. M. 4th ed. Ithaca, N.Y.
Widengren, G. (1959): The Sacral Kingship of Iran. In: The Sacral Kingship (1959), pp. 242 – 257.
Wiet, G. (1955): Les atours précieux. Cairo.
Xvol’son, D. (1869): Izvestija o xozarax, burtasax, bolgarax, mad’jarax, slavjanax i russax Abu-Ali
Axmeda ben Omar ibn Dasta. St. Petersburg.
Al-Ya‘qûbî: Ta’rîx. Ed. Houtsma, M. Th.. Leiden 1883. 2 vols.
Zaxoder, B. N. (1962, 1967): Kaspijskij svod svedenij o Vostočnoj Evrope. Moskva. 2 vols.
Zuckerman, C. (1955): On the Date of the Khazars’ Conversion to Judaism and the Chronology of
the Rus’ Oleg and Igor. Revue des Études Byzantines 53, pp. 237 – 270.
Zuckerman, C. (1997): Two Notes on the History of the Thema of Cherson. Byzantine and Modern
Greek Studies 21, pp. 210 – 212.
Zuckerman, C. [K. Cukerman] (2002): O proisxoždenii dvoevlastija u xazar i obstojatel’stvakh ix
obraščenija v iudaizm. Materialy po arxelogii, istorii i étnografii Tavrii 9, pp. 521 – 534.

Acta Orient. Hung. 60, 2007