KHAZARS, a national group of general Turkic type, independent and sovereign in Eastern Europe between the seventh and tenth centuries C.E. During part of this time the leading Khazars professed Judaism. The name is frequently pronounced with an a-vowel, as in the Greek Χάξαροι and Arabic Khazar (Ḥazar), but there are traces of a different pronunciation in Hebrew (Kuzari, pl. Kuzarim), Greek (Χότξιροι), and Chinese (Kʿo-sa). The name has been explained as having derived from Turkish qazmak ("to wander," "nomadize (?)"), or from quz ("side of mountain exposed to the north"). The latter etymology would account for the o/u-vowel in some forms of the name, for which no satisfactory explanation has been given.
The Origin of the Khazars
The Khazars, of Turkic stock, originally nomadic, reached the Volga-Caucasus region from farther east at some time not easily determinable. They may have belonged to the empire of the Huns (fifth century C.E.) as the Akatzirs, mentioned by Priscus. This name is said to be equivalent to Aq-Khazar, i.e., White Khazars, as opposed to the Qara-Khazar or Black Khazars mentioned by al-Iṣṭakhrī (see below). The Khazars probably belonged to the West Turkish Empire (from 552 C.E.), and they may have marched with Sinjibū (Istämi), the first khāqān of the West Turks, against the Sassanid (Persian) fortress of Ṣul or Darband.
In the time of Procopius (sixth century) the region immediately north of the Caucasus was held by the Sabirs, who are referred to by Jordanes as one of the two great branches of the Huns (Getica, ed. Mommsen, 63). Masʿūdī (tenth century C.E.) says that the Khazars are called in Turkish, Sabīr (Tanbīh, ed. Cairo, 1938, 72).
In 627 (Theophanes, Chronographia, ed. De Boor, 1 (1883), 315) "the Turks from the East whom they call Khazars" under their chief, Ziebel, passed the Caspian Gates (Darband) and joined Heraclius at the siege of Tiflis. In view of what is known of a dual kingship among the Khazars (see below), it would be natural to assume that Ziebel, described by Theophanes as "second in rank to the khāqān," was the subordinate Khazar king or beg. However, there are grounds for thinking that Ziebel stands for yabgu, a Turkish title – in the parallel Armenian account (Moses of Kalankatuk, trans. Dowsett, 87) he is called Jebu Khāqān – and that he is T'ung-ye-hu, Ye-hu Khagan of the Chinese sources, i.e., T'ung Yabgu, Yabgu Khāqān, the paramount ruler of the West Turks, who is represented as second in rank to "the King of the North, the lord of the whole world," i.e., the supreme khāqān of the Turks. In the narratives of Theophanes and Moses of Kalankatuk respectively, the Khazars are also called Turks and Huns. From 681 C.E., we hear much in the latter author of the Huns of Varach ʿ an (Warathān), north of Darband, who evidently formed part of a Khazar confederation or empire. Their prince Alp Ilutver was often in attendance on the Khazar khāqān and was converted to Christianity by an Albanian bishop.
It will be seen that the question of the precise racial affinities of the Khazars is not readily solved (see also below). There appears to be insufficient evidence to warrant the conclusion of K. Czeglédy that the Khazars were of Sabīr origin and distinct from the Caucasian Huns and West Turks ("Bemerkungen zur Geschichte der Chazaren," Acta Orientalia… Hungariae, 13 (1961), 245), since it is not known how far these ethnic names mean the same thing.
Consolidation of the Khazar State
According to Theophanes (ibid., 358), the ruler of the Bulgars in the region of the Kuban River (West Caucasus) died c. 650 C.E., leaving five sons of whom only the eldest remained in his inheritance, while the others moved further west, as far as the Danube. On this, the Khazars, described as a "great nation … from the interior of Berzilia in the First Sarmatia," emerged and took possession of the territory as far as the Black Sea. The change of position was completed by 679, when one of the brothers crossed the Danube and conquered present-day Bulgaria. Earlier than this, in 576 C.E., a West Turkish force had been present at the siege of Bosporus (Kerch) in the Crimea (Menander Protector, ed. Bonn, 404), but hitherto there is no mention of the Khazars as such so far to the west. The advance of the Khazars to the Black Sea and Crimea area appears to be mentioned also in the Reply of Joseph (see below, Khazar Correspondence), where a great Khazar victory over the W-n-nt-r is referred to. A people north of the Khazars called W-n-nd-r is mentioned in the Ḥudūd al-ʿĀlam (Regions of the World, trans. by V. Minorsky (1937), 162). Both names are best explained as corresponding to Onogundur, an old name in Greek sources for the Bulgars. The advent of the Khazars on the Black Sea was clearly of great consequence for the future, for they now came within the sphere of Greek political and cultural influence. By 700 C.E. or earlier there were Khazar officials in Bosporus and Phanagoria. Henceforth the Crimea, as well as the Volga and the Caucasus, came to be specially associated with the Khazars, and a further way westward was opened for them toward both Kiev and the Slav lands via the Dnieper (see below).
Arabs and Khazars had already been in conflict on the line of the Caucasus (first Arab-Khazar war, 642–52 C.E.). *Bāb al-Abwāb at the eastern end of the range was occupied by the Arabs in 22 A.H. (643). In the same year the caliph Omar sent instructions to advance northward. Though the Arabs attacked *Balanjar repeatedly, they were unable to take it. The defeat and death of the Arab general at Balanjar in 32 A.H. (653) practically marks the end of the war and the close of the first phase of Arab-Khazar relations. According to Mus ʿ ūdī, the Khazar capital was at this time moved from *Samandar to *Atil, but he says elsewhere that Balanjar was the former capital.
Further Relations with Byzantium and the Arabs
After the exile of Justinian II to the Crimea in 695, the Khazars on several occasions played an important, even determining, part in Byzantine politics. Toward 704 the khāqān helped the emperor at a crucial moment and gave him his sister Theodora in marriage. Justinian returned to Constantinople to reign a second time. His successor Bardanes (711–13) was likewise indebted to the khāqān. In 732 the emperor Leo the Isaurian married his son, the future Constantine V, to a Khazar princess called in the sources Irene. The child of this marriage was Leo IV, the Khazar (775–80). It is to be understood that Irene and Theodora above are baptismal, i.e., not Khazar, names.
The second Arab-Khazar war began in 722 or earlier, and ended in 737 with the defeat of the Khazars by Marwān b. Muhammad (later Marwān II). The Khazar khāqān is said at this time to have professed Islam. If so, we hear no more about it. Later the khāqān was a Jew, as we know from the Arabic geographers Ibn Rustah (c. 290/903), Iṣṭakhrī (c. 320/932), Ibn Ḥauqal (367/977), etc., and it is implied in the Reply of Joseph that the beginnings of Khazar Judaism dated as far back as 112/730, when the Khazars defeated the Arabs south of the Caucasus, and from the spoils consecrated a tabernacle on the Mosaic model. The conversion of the leading Khazars to Judaism perhaps took place toward 740 C.E. (see below). It seems at all events certain that the Khazars successfully resisted the Arabs for several decades, and that they were reduced only with difficulty and at a time when the internal situation of the caliphate prevented the Arabs from exploiting their victory: Marwān was called away to become the last *Umayyad Caliph (744) and to struggle against ever-growing opposition, until
his death in 750 at the hands of *Abbasid soldiers in Egypt. The dynastic crisis probably saved Khazaria. At the same time the situation had wider implications, for if Marwān had been able to hold the Khazar territory permanently, the history of Eastern Europe might have been very different.
The Khazar Double Kingship
This was a phenomenon found among other Turkic peoples, e.g., the Qara-Khanids, and not unknown elsewhere; compare the double kingship at Sparta in antiquity, and the shogun and mikado of medieval Japan. How far back the institution goes among the Khazars cannot be exactly determined. Ya ʿ qūbī (ninth century) speaks of the Khazar khāqān and his representative (khalīfa) apparently in the sixth century (Historiae, ed. by M.T. Houtsma, 1 (1883), 203; cf. above for Ziebel Jebu Khāqān in 627). Arab accounts, in Ṭabarī, Ibn al-Athīr, etc., of the Arab-Khazar wars (see above) afford no precise evidence of the dual kingship, yet the Arab geographers regularly mention it. The account of al-Iṣṭakhrī, written c. 320/932, is as follows (Viae regnorum, ed. by M.J. De Goeje (1927), 223ff.): "As to their politics and system of government, their chief is called khāqān of the Khazars. He is greater than the king of the Khazars [elsewhere called by al-Iṣṭakhrī the bak or bāk, i.e., beg], except that the king of the Khazars appoints him. When they wish to appoint this khāqān, they bring him and throttle him with a piece of silk, till, when his breath is nearly cut off, they say to him, 'How long do you wish to reign?' and he says, 'So and-so many years.' If he dies short of them, well and good. If not, he is killed when he reaches that year. The khaqanate is valid among them only in a house of notables. He possesses no right of command nor of veto but he is honored, and people prostrate themselves when they enter his presence.…. The khaqanate is in a group of notables who possess neither sovereignty nor riches. When the chief place comes to one of them, they appoint him, and do not consider his condition. I have been informed by a reliable person that he had seen a young man selling bread in one of the sūqs. People said that when their khāqān died, there was none more deserving of the khaqanate than he, except that he was a Muslim, and the khaqanate is not conferred on any but a Jew."
A remarkable parallel to the inauguration ceremony described by Iṣṭakhrī is found in a Chinese source on the Turks in the sixth century C.E., the Chou Shu (trans. by Liu Mau-Tsai, Die chinesischen Nachrichten zur Geschichte der Ost-Tuerken, 1 (1958), 8). Recently the theory of A. Alföldi that the double kingship among nomadic peoples corresponds to leadership of the two wings of the horde ("Türklerde çift krallik," Ikinci Türk Tarih Kongresi, Istanbul, 1943, 507–19) has won wide acceptance, but does not apply particularly well to the Khazars. Masʿūdī had already suspected that the Khazar khāqān represented a dynasty which had been superseded (Murūj al-Dhahab, ed. by B. de Maynard and P. de Courteille, 2 (1878), 13). K. Czeglédy (op. cit.) has suggested that the khāqān was the representative at the Khazar capital, Atil,
Date of the Khazar Conversion to Judaism
This has already been referred to above (see *Būlān and below Khazar Correspondence). The date c. 740 C.E. is suggested by converging considerations, namely, the circumstances of the reported conversion to Islam in 737 and the dating given by *Judah Halevi in the Kūzari (Cosri). The absence of distinct references to the Judaism of the Khazars in the biographies of St. Abo of Tiflis, who was in Khazaria c. 780 C.E. and of Constantine (Cyril), who was there c. 860, should not be pressed as proof that the conversion to Judaism took place only later (cf. also M.I. Artamonov, Istoriya Khazar, 332–3). Mas ʿ ūdī states positively that the king of the Khazars became a Jew in the caliphate of Hārūn al-Rashīd (786–809 C.E.). This may well refer to the reformation c. 800 under *Obadiah of which the Reply of Joseph speaks. S.P. Tolstov has sought to explain the Khazar conversion to Judaism as a result of the conquest of Khwārizm (*Khorezm) by the Arab general Muslim ibn Qutayba in 712.
The Khazar Empire
The extent of the territory ruled by the Khazars has been variously estimated. Thus B.A. Ribakov ("K voprosu o roli khazarskogo kaganata v istorii Rusi," Sovetskaya Arkheologiya, 18 (1953), 128–50) makes Khazaria a small territory on the lower courses of the Volga and Don, to include Sarkil (see below) and the Khazar capital (assigning separate localities to Atil, Khamlīj, and al-Bayḍā', usually taken to be the same place). This is based principally on the data in the world map of Idrīsī, which offers a somewhat misleading picture of Khazaria (see K. Miller, Mappae Arabicae, 1 (1926), Heft 2). On the other hand, S.P. Tolstov envisages a Khazaria united with Khwārizm under one ruler to form a single state, a view for which the evidence is slight.
It must be allowed, however, that at one time Khazar rule extended westward a long way beyond the Crimea-Caucasus-Volga region which for the Greek and Arabic sources is Khazaria. The Russian Primary Chronicle ((1953), 58–59; Chronicle of Nestor, Povest vremennykh let) reports that at an unspecified date the Polians south of the Middle Dnieper paid tribute to the Khazars of a sword per hearth, and that in 859 C.E. the Polians, Severians, and Viatichians paid them a white squirrel skin per hearth (trans. Cross and Sherbowitz-Wetzor, 58, 59). Later these payments in kind ceased to be made, being evidently replaced by money payments; e.g., the Radimichians paid the Khazars a shilling or dirham apiece until 885 C.E., according to the Chronicle (61), and the Viatichians until 964, the same per plowshare (ibid., 84). All these peoples were exposed to attack by any strong forces coming up the valleys of the Don and Donets from the Khazar territory. Kiev itself was occupied by the Khazars for some period before 862, but presumably was not built by or for them (ibid., 60, cf. 54), unlike Sarkel or *Sarkil on the Don, which on the application of the khāqān and beg to Emperor Theophilus was constructed by Byzantine workmen in 833 C.E. All of these territories were to be taken from the Khazars, some already in the ninth century, by the advancing Russians.
East of the Volga, in the direction of Khwārizm, the situation is obscure. Al-Iṣṭakhrī tells of caravans passing between Khwārizm and Khazaria, mentioning specifically Slav, Khazar, and Turkish slaves and all kinds of furs among the principal merchandise of Khwārizm. On the other hand, he says that Khwārizm has the nomad Turks (Ghuzz) on its northern and western frontier, not the Khazars. According to Tolstov, a "royal road" led from Khorezm to the Volga, traces of which may be seen from the air, and he finds in it an indication of the emergence of a great Khorezmian-Khazar state in the tenth and beginning of the 11th century (cf. above).
The Extent of Khazar Judaism
While the Khazars were generally known to their neighbors as Jews (cf. notably the narrative of Ibn Faḍlān), they seem to have had little or no contact with the central Jewish organization in Iraq, and they tend to be mentioned less by Rabbanite than by Karaite authors. This is not to say that the Khazars were Karaites, a view which has not lacked defenders, at least since the time of A. *Firkovich. Yet such contemporary or nearly contemporary documents as we possess offer no evidence of the Karaism of the Khazars. On the other hand, it would seem that the lack of interest in the Khazars on the part of the Jewish authorities, as reflected in the literary works at our disposal, was due at least partly to their imperfect adherence to Judaism. This is illustrated notably in their retention of a number of pagan (shamanist) customs, dating back to their Turkic past, which are duly noted by the Arab geographers.
We may here consider the position of H. Baratz that in the oldest Russian writings of a legal character there are Hebrew, mostly biblical-talmudic, elements, and that these go back to Khazar times. Thus the fact that early Russian codes, including the Zakon sudni liudem ("Law for the Judging of the People"), contain traces of Mosaic and talmudic legislation,
The Downfall of Khazaria
The Reply of Joseph mentions that the Khazars guarded the mouth of the Volga before 961 C.E. and prevented the Russians from reaching the Caspian. On several occasions, notably c. 913 and again in 943, the Russians made raids down the Volga, passing through Atil. Later, apparently in 965, Khazaria was the object of a great Russian attack, which was aimed at the Khazar capital and reached as far as Samandar, as we know from Ibn Ḥawqal. From this disaster the Khazars appear to have recovered only partially. Again at this time (cf. above) we hear of a Khazar khāqān adopting Islam. His motive is said to have been to secure the help of the people of Khwārizm (Miskawayh, ed. Amedroz, II, 209; Ibn al-Athīr, VIII, 196).
After 965 the Khazars are still mentioned occasionally, but scarcely for long as an independent people. We cannot use the Cairo Genizah document published by J. Mann, concerning a messianic movement supposedly in Khazaria in the time of al-Afḍal, the great Fatimid vizier who ruled 1094–1121 (REJ, 71 (1920), 89–93; 89 (1930), 257–8), as proof of continued Khazar existence until this time, since it has been shown that the movement in question took place in Kurdistan (see S.D. Goitein, "Obadyah, a Norman Proselyte," in JJS, 4 (1953), 74ff.). Furthermore, Oleg, the same who, according to the Russian Chronicle, established himself in Tmutorokan in 1083, is called in a seal of the 11th–12th century "archon of all Khazaria" (N. Bǎnescu in Bulletin of the Romanian Academy, Hist. Sect. 22 (1941), cited by A.V. Soloviev, For Roman Jakobson (1956), 478). Whatever is precisely indicated here by "Khazaria" – e.g., the Khazar country in the Crimea – such a claim could not have been made prior to 965. We must therefore see the Khazar state as having subsisted until the second half of the tenth century, or the 11th century at the latest. By the 12th century the Qipchaqs or Cumans (identified also with the Polovtsi) appeared in the steppes once ruled by the Khazars. At the time of the Mongol invasions in the 13th century, it was they, not the Khazars, who were in possession.
The Khazar Correspondence
This name is usually given to what appears as an interchange of letters in Hebrew between *Ḥisdai ibn Shaprut, a well-known personality of Muslim Spain in the tenth century, and *Joseph, king of the Khazars. M.I. Artamonov (Istoriya Khazar, 12) includes the Cambridge Document as well as the Letter of Ḥisdai and the Reply of Joseph in the Khazar Correspondence, but this would seem to be contrary to general usage. The Reply is available in a Long Version and a Short Version (LV and SV). The Correspondence involves serious critical difficulties, and its authenticity has been much debated.
The Letter of Ḥisdai begins with a piyyut containing an acrostic which gives his own name and that of Menaḥem b. Saruq, the latter presumably acting as Ḥisdai's secretary and being the author of the piyyut. The prose part, after compliments, refers to the geographical situation of al-Andalus and Khazaria and describes the natural wealth of al-Andalus and Ḥisdai's own position there. It seems that his interest has been aroused by his having heard repeatedly that the Khazars are Jews. The Letter mentions attempts made by Ḥisdai to get in touch with the Khazar king. He was finally successful through the instrumentality of two Jews, Mar Saul and Mar Joseph, who accompanied an embassy which arrived at Cordoba from the "king of the G-b-līm, who are the Ṣaqlab" (see below). The Letter of Ḥisdai was conveyed to the East by their means, i.e., overland, and eventually was put into the hands of the Khazar king, according to the Reply, by a certain Jacob or (LV) Isaac b. Eliezer, a Central European Jew. The tone of the Letter of Ḥisdai is mostly one of enquiry, and it invites an answer to questions which range over a variety of topics: Is there a Jewish kingdom anywhere on earth? How did the Jews come to Khazaria? In what way did the conversion of the Khazars take place? Where does the king live? To what tribe does he belong? What is his method of procession to his place of worship? Does war abrogate the Sabbath? Has the Khazar king any information about the possible end of the world? Ḥisdai mentions that ʿ Abd al-Raḥmān III al-Nāṣir is the reigning king of al-Andalus. This gives 961 as the terminus ad quem for the Letter, with 953–55 as a possible terminus a quo, for in those years Cordoba was visited by John of Gorz, as envoy of the German emperor Otto I, who may be the "king of the G-b-līm, who are the Ṣaqlab" already referred to.
The Reply of Joseph begins by referring to the principal contents of the Letter and recapitulates a number of its questions. It then relates the early history of the Khazars, and proceeds to deal at length with the conversion to Judaism under Būlān. The conversion is initiated by a dream of Būlān, which he communicates to a certain general among them (LV), apparently the beg. From the spoils of a Khazar attack on Ardabil, south of the Caucasus, for which we have the synchronism 730 in the Arabic sources, a tabernacle on the biblical model is set up. A religious debate between representatives of Judaism, Christianity, and Islam is held, after which Būlān and the principal Khazars accept the religion of Israel. Under a later king, Obadiah, there was a reform of religion. Synagogues and schools were built, and the Khazars became familiar with Torah, Mishnah, Talmud, and the liturgy, i.e., rabbinic Judaism was introduced. Joseph then traces his descent from Obadiah and gives a description of his country and capital. He refers
The correspondence has been available since the appearance of the work Kol Mevasser of Isaac Akrish in or after 1577, and more generally since the two letters were published by the younger *Buxtorf in his edition of the book Cosri (Kūzārī) of Judah Halevi in 1660. It is not known what manuscript source was used by Isaac Akrish; Buxtorf depended on Kol Mevasser. The only known manuscript of the Correspondence as a whole, containing the Letter of Ḥisdai and the Reply of Joseph (SV), is in the library of Christ Church, Oxford. This manuscript is very similar to the printed text, which, it has been suggested, is a transcript. There appear to be no special grounds for this opinion, though the manuscript, which is undated, has no claims to great antiquity. Nothing is sure about its provenance, but it is thought to have belonged originally to the celebrated Dr. Fell (1625–1686).
A longer version of the Reply of Joseph was published by A. *Harkavy in 1874, from a manuscript of the Second Firkovich Collection in the Leningrad Public Library. The Long Version bears no indication of any alterations or additions, and is supposed to date from the 13th century. Harkavy, in spite of his very critical attitude to Firkovich, regarded it as the undoubted original of the Short Version.
It appears impossible to suppose that the Khazar Correspondence is a fabrication of the 16th century in view of a reference to it, with the citation of part of the Reply of Joseph, agreeing in general with the Long Version, in the Sefer ha-Ittim of Judah b. Barzillai al-Bargeloni, dated between 1090 and 1105, and a similar reference in the Sefer ha-Kabbalah of Abraham *Ibn Daud in the 12th century. It cannot be admitted that these works were interpolated in the 16th century or later, to support the authenticity. Nor does it appear at all plausible that the letters forming the Khazar Correspondence were forgeries of the tenth century, composed with a view to informing the Jews about the Khazars. It is demonstrable that the literary style of the Letter of Ḥisdai differs from that of the Reply of Joseph in a marked manner. The classical Hebrew construction of vav conversive with the imperfect to express the past tense is freely used in the Letter of Ḥisdai, actually 48 times as against 14 times when the past tense is rendered by simple vav with the perfect. In the Reply (LV), on the other hand, vav conversive with the imperfect occurs not more than once or twice, while the past is expressed by the perfect and simple vav nearly 100 times. Further, in the Short Version of the Reply the vav conversive with the imperfect to express the past, instead of simple vav with the perfect, occurs in a number of passages where the wording is different from the Long Version. There is a new proportion of vav conversive with the imperfect to simple vav with the perfect: 37 to 50. It may therefore be affirmed that there is a separate authorship for the Letter and the Reply, and assumed that the Long Version of the Reply, or something very like it, has been worked over by a third hand to produce the Short Version. There are grounds for thinking that the Reply originally was written in a non-Arabic-speaking environment. Most people would agree with Kokovtsov's cautious statement that as basis for both versions there is the same original text, in general better preserved in the Long Version. B.A. Ribakov supposed that an authentic letter of King Joseph was worked over in Tmutorokan toward the end of the 11th century ("about 1083"), which resulted in the Long Version, and that some time afterward the text of the Long Version was modified by Jews of Barcelona to produce the Short Version of the Reply.
D.M. Dunlop, History of the Jewish Khazars (1954, p. b. 1967), includes extensive bibliography; idem, in: Roth, Dark Ages, ch. 8, and index; M.I. Artamonov, Istoriya Khazar (1962), especially valuable for the archaeology; V. Minorsky, in: Oriens, 11 (1958), 122–45 (review of Dunlop's History…); G. Moravcsik, Byzantinoturcica, 2 (1958), 334–6 (refers to Greek sources); A. Zajączkowski, in: Acta Orientalia Hungaricae, 12 (1961), 299–307 (regards the Karaites as successors of the Khazars); Szyszman, in: Revue de l'Histoire des Religions, 152 (1957), 174–221 (an original short treatment from the Karaite standpoint); A.N. Poliak, Kazariyyah (Heb., 19513); A. Yarmolinsky, in: Bulletin of the New York Public Library, 42 (1938), 695–710; 63 (1959), 237–41 (bibliographies); B.D. Weinryb, in: Studies in Bibliography and Booklore, 6 (1963), 111–29 (updates Yarmolinsky's bibliographies); B.A. Ribakov, in: Sovetskaya Arkheologiya, 18 (1953), 128–50.