Ziebel Qaghan identified

ZIEBEL QAGHAN IDENTIFIED by Étienne de La Vaissière Several works have dealt with Heraclius’ Turkic allies during his war with Khosrau II in 627–8, but questions as to the exact identity of these allies are still pending.1 In order to answer them, we need to present in some detail the politics of the two Turkic Empires during this period. Recent epigraphic discoveries have thrown new light on the genealogy of the Turkic imperial Ashinas dynasty.2 The genealogies drawn up by Chavannes and Liu Mau-Ts’ai require a crucial correction, which has a deep impact on our perception of the sovereigns’ legitimacy in the context of a sacral kingship.3 Thirty years after the victory of the Turks over their Rouran masters, the civil war of 581–3 between the sons of the first qaghans split the Turkic Empire.4 To the East, in Mongolia, the scions of the ephemeral second qaghan Qara, son of Bumïn, managed to grasp the throne, while the scions of Muqan, another son of Bumïn and the actual builder of the empire, fled to the West, to the northern slopes of the Tianshan. The competition between the two dynasties with equal claims to the supreme title of great qaghan lasted for thirty years, up to 610, a fact that remained unnoticed until the recent elucidation of the Mongolküre inscription.5 In 610, Muqan’s heirs were expelled from the West by a third branch of the family, tracing its genealogy back to Bumïn’s brother Ishtemi, which retained from the beginning great power in the West. To the Ishtemid Shegui, who drove out his Muqanid rivals,6 succeeded on the throne, some time before 619, his brother Tong.7 1. Earlier studies are analyzed in A. Bombaci, Qui était Jebu Xak‘an ?, Turcica 2, 1970, pp. 7–24. For more recent studies, see M. Dobrovits, The nomadic ally of Heraclius, Chronica 3, 2003, pp. 3–8. 2. See mainly É. de La Vaissière, Oncles et frères : les qaghans Ashinas et le vocabulaire turc de la parenté, Turcica 42, 2010, pp. 267–77, and also Id., Maurice et le qaghan : à propos de la disgression de Théophylacte Simocatta sur les Turcs, REB 68, 2010, pp. 219–24. 3. É. Chavannes, Documents sur les Tou-Kiue (Turcs) Occidentaux, Paris 1903, pp. 2–4; Liu Mau-Tsai, Die chinesischen Nachrichten zur Geschichte der Ost-Türken (T’u-küe), Wiesbaden 1958, genealogy at the end of vol. 2. See the appendix to this article for an updated genealogical tree. 4. Chavannes, Documents (quoted n. 3), pp. 13–4, 48–51, 219–21 5. La Vaissière, Maurice et le qaghan (quoted n. 2), p. 221. 6. Chavannes, Documents (quoted n. 3), pp. 17–8. 7. Ibid., p. 24. Constructing the seventh century, ed. by C. Zuckerman (Travaux et mémoires 17), Paris 2013, pp. 741–748. 742 ÉTIENNE DE LA VAISSIÈRE The eastern branch of the family was not faring well either. Crushed between the Sui and the expanding proto-Uyghur confederation in the North, the qaghans had to take refuge on the frontier of China and became the Sui’s puppets.8 Only after a civil war had split the Sui Empire could the eastern qaghans regain their position as the main independent nomadic power of Eastern Asia, especially under Xieli (620–30).9 The relationship between Tong and Xieli was very tense during the 620’s. By giving the hand of a Tang princess to Tong, Chinese diplomacy strove to ally the Western Empire to China in order to get rid of Xieli. This is especially true of the years 626–8, just before Tong’s death in 628. As the Chinese annals put it: “Xieli qaghan, unhappy that the Chinese Empire has allied itself by marriage with Tong, several times sent soldiers to raid his country.”10 However, this hostility remained confined to raids and diplomatic threats, without ever breaking out into a full-scale war. The two empires then had little in common, except for a shared past going back some forty years, a period during which seven qaghans ascended the throne in the East and six in the West. Xieli’s Empire was in every respect Eastern Asian: he had no interest in Central Asia and never tried to conquer it. His network of subject tribes ranged from Mongolia and the Ordos east to the Pacific and Korea. From its centre on the Chinese frontier, which Xieli never left, he could pillage Northern China much more conveniently than from Mongolia. This pillage was the main activity of Xieli’s Turks and the source of their wealth. Meanwhile, Tong’s Empire had shifted to the West, as he moved his capital from the northern slopes of the Central Tianshan to Semirech’e, 700 km further west. Its wealth came from the pastures of the Central Asian plateaux and from the taxes paid by the sedentary oases of the region. Although Tong vied with Xieli for the control of the northern fringes of the Turkic world, on the eastern slopes of the Altai, their actual contact zone was limited. Tong expanded his empire towards the south-west, towards Bactria, Persia and the Caucasus.11 Byzantine and Caucasian sources describe Heraclius’ ally Ziebel as the brother of the king of the North, second only to him in the empire (we will discuss the Yabghu qaghan title below).12 This king can only be Tong or Xieli. As regard the former, the difficulty of linking these data with those of the Chinese sources is that none of Tong’s brothers could be viewed as second to him in the empire, and none had a name that could resemble Ziebel.13 If he is the latter, various hypotheses have been proposed, including an attempt to identify the king of the North as Xieli, and Ziebel as Tong Yabghu qaghan. However, this purely ad hoc hypothesis is contradicted by what we have just learned about the 8. Liu Mau-Tsai, Die chinesischen Nachrichten (quoted n. 3), pp. 58–65. 9. Ibid., pp. 134 sq. 10. Jiu Tangshu 194.5182, and Chavannes, Documents (quoted n. 3), p. 53. 11. Compare Chavannes, Documents (quoted n. 3), pp. 24–5 and 52–3 with Liu Mau-Tsai, Die chinesischen Nachrichten (quoted n. 3), pp. 134 sq. 12. Theophanes Confessor: The Chronicle of Theophanes Confessor : Byzantine and Near Eastern history, AD 284–813, transl. with introd. and comment. by C. Mango and R. Scott with the assistance of G. Greatrex, Oxford 1997, p. 447. Movsēs Dasxuranc‘i: The History of the Caucasian Albanians by Movsēs Dasxuranc‘i, transl. by C. Dowsett, London 1961, p. 83. 13. Ziebel was for instance identified with Mohe shad, brother of Tong, known for acting as an envoy to China (Chavannes, Documents [quoted n. 3], p. 55). See Bombaci, Qui était Jebu Xak‘an (quoted n. 1), p. 11.  ZIEBEL QAGHAN IDENTIFIED 743 actual relationship between the two empires: no source provides the slightest hint of any dependency of the Western Turkic Empire on the Eastern one, while we do have abundant proof of their mutual independence. To imagine Tong taking orders from Xieli is in total contradiction to what we know of both qaghans, in particular regarding the far West. Seen from the West, the supreme head of the Turks in 627–8, named “king of the North”, can only be the Yabghu Qaghan Tong. As for Ziebel, an uncle of Tong, Sipi , fits exactly into his western depictions. Sipi, whom Tong put in charge of part of the Turkic tribes,14 carried the singular title “little qaghan” ( ), which corresponded precisely to Ziebel’s title of viceroy (sons and brothers of Tong had different titles). As for his name, its transcription in Chinese, Sipi, was then pronounced in Early Middle Chinese ʐih-bjil, where ʐ is the voiced retroflex fricative, quite similar to Russian ж, while the initial Z- in Ziebel could transcribe a z, ž, š or ǰ.15 Thus, its Byzantine and Chinese versions match. This identification, so obvious, has never been proposed since Sipi, Tong’s uncle, could not be a brother of Tong, Ziebel. However, who gave this information to the Byzantines and the Chinese if not the Turks themselves? A Turk told a Byzantine scribe that Ziebel was the brother of the king of the North. Similarly a Turk told a Chinese scribe that Sipi was Tong’s uncle. The key to the problem is the word they used: äci. In the Turkic vocabulary of kinship, designating with one word several members of the family across generations, äci describes males older than ego in paternal line except for the father and grandfather.16 Thus it is the standard word for both paternal uncle and for elder brother. By contrast, the Chinese vocabulary, like ours, more detailed and precise, reserves a word for each position in the family in each generation. The Chinese chose systematically to translate äci by paternal uncle, sometimes rightly, sometimes wrongly. There are at least four cases of false uncles in the Chinese sources, for whom external evidence, such as inscriptions left by qaghans, ascertains the true genealogical links, proving inaccurate rendering by the Chinese of the ambiguous Turkic word äci. Ziebel is one of these cases.17 He was the äci and viceroy of the king of the North, just as Sipi was the äci and Little qaghan of Tong. For once, western sources are more accurate in their rendering—brother—since Byzantine and Caucasian informants actually met the Turkic elites in the Caucasus, while the Chinese were far removed from the events. Not only the name, title and status fit together, but the chronology as well: during the first operations of Ziebel in the Caucasus, in 625–7, he is consistently under the supreme command of the “king of the North.” For 628–9, however, there is no mention of a supreme king. Moreover, Ziebel is in deep trouble from 629 on, even reported dead by 14. Xin Tangshu 215.6057, and Chavannes, Documents (quoted n. 3), p. 54. The text is: . 15. G. Moravcsik, Byzantinoturcica, Berlin 1958, vol. II p. 33. The second character, pi, is used for names of various other qaghans, for instance in the transcription of the first syllable of Bilgä, . 16. K. Grønbech, The Turkish system of kinship, in Studia Orientalia Ioanni Pedersen Septuagenario A.D. VII id. nov. anno MCMLIII a collegis, discipulis, amicis Dicata, Hauniae 1953, pp. 124–9, and Sh. Baștuĝ, Kök Türük kinship terminology: an Omaha model, Central Asiatic journal, 37, 1–2, 1993, pp. 1–20. 17. For four examples of false uncles in Chinese sources on the Ashinas dynasty, see La Vaissière, Oncles et frères (quoted n. 2). 744 ÉTIENNE DE LA VAISSIÈRE one source. These elements find explanation in the biography of Sipi: he rebelled against Tong and killed him at some time in the second year of Zhenguan (February 628– January 629).18 But in 629, Sipi was himself defeated by Tong’s son and had to take refuge in the Altay, where he in turn was killed in 630.19 We know from numismatic data that the Augusta was sent to Ziebel in spring or summer 629, that is precisely the short period during which Sipi was still the uncontested ruler of the empire, and she was called back when news of his defeat arrived, after August 629, that is when Sipi was contested by Tong’s son. Ziebel and Sipi’s careers are exactly parallel, save for one point: Nikephoros reports him dead in 629, thus allowing Heraclius’ cancelation of his daughter’s marriage.20 In fall 629, Sipi was only a fugitive, expelled from the centre of the qaghanate yet, from the Byzantine point of view, he was as good as dead for an imperial marriage. In any case, nobody could verify his whereabouts and the accuracy of the reasons put forward by Heraclius for cancelling the marriage. If, however, Ziebel held only the second rank in the empire, how could he obtain the hand of an Augusta, an unprecedented honor? This marriage can be considered from two points of view: from Byzantine and from Turkic politics. As for the latter, it is an established fact that not only qaghans but also imperial princes sought the hands of foreign imperial princesses. Shegui, the elder brother of Tong and Sipi, before he expelled the scions of Muqan from the Western Turkic Empire and became qaghan, had benefited from a Chinese imperial marriage. Thus, while only a pretender to the throne from a lateral branch, he obtained the necessary stature to get rid of the last heir of Muqan. Enhancing a pretender’s prestige by means of marriage was a usual way for foreign powers to influence Turkic politics. Ziebel/Sipi was not only promised an Augusta, he also sought the hand of a Chinese princess, hoping to repeat what his elder brother Shegui had successfully achieved twenty years earlier with a Sui princess. And he was equally successful in destroying the power of his rival. From a Byzantine point of view, a marriage between an Augusta and somebody who was not yet supreme qaghan might look less obvious. However, the later proposal to marry the same Augusta to the general of the Arabic forces in Egypt provides a parallel: the patriarch of Egypt would certainly not have dared to propose something so unusual if a precedent was not close at hand.21 The Byzantine and Armenian historians describe her projected marriage to a Turkic ruler without pointing to a problem of inequality of status. From Turkic point of view, this was only business as usual, reflecting the autonomy of the small qaghans in their part of the empire and the actual frailty of Turkic unity. Besides, Heraclius had no other choice.22 18. On the chronology of Tong, see É. de La Vaissière, Note sur la chronologie du voyage de Xuanzang, Journal asiatique 298, 1, 2010, pp. 164–6. 19. Xin Tangshu 215.6057, and Chavannes, Documents (quoted n. 3), p. 54. 20. Nikephoros : Nikephoros Patriarch of Constantinople, Short History, ed., transl. and comment. by C. Mango (CFHB 13), Washington 1990, p. 67. 21. Ibid., pp. 72–5. 22. C. Zuckerman, Au sujet de la petite augusta sur les monnaies d’Héraclius, RN 152, 1997, pp. 473–8, here p. 478.  ZIEBEL QAGHAN IDENTIFIED 745 faut-il The only remaining problem is the title Yabghu qaghan (Jebu Xak‘an) given to the un signe leader of the Turks besieging Tiflis in the History of the Caucasian Albanians.23 The diacritique Byzantine sources do not mention it, saying only that he was second in dignity to the associé au J? qaghan alone. But Yabghu qaghan was the official title of the head of the Western Turkic Empire according to the Chinese sources, including an eye-witness like Xuanzang, and could not belong to anybody else. This contradiction between the Armenian and the Byzantine sources, on which much has been written, can best be resolved if we pay due attention to the nature of our Armenian source. For part of its depiction of the events, the History of the Caucasian Albanians goes back to the testimony of the Albanian catholicos Viroy who met Ziebel’s son in Albania during the winter 628–9.24 Viroy was not in Albania during the siege of Tiflis and all his knowledge of the Turkic power comes from discussions with the Turks during a period in which Ziebel had already deprived Tong of his throne and of his life. In other words, Viroy has never heard of any other Yabghu qaghan than Ziebel, who obviously took this title. I suggest that this title, present in Viroy’s narrative for the years 628 and 629, was projected backward on 627 by one of the compilers, for simplicity and cohesion’s sake. This might have been a member of Viroy’s entourage (Akopyan), the anonymous compiler of 682 (Howard-Johnston), or the early-8th-century compiler (Zuckerman).25 In any case, when the “Viroy narrative” using this name was integrated with events prior to the return of the catholicos to Albania, the resulting text made use of Viroy’s more precise terminology. The Byzantine chroniclers, by contrast, derived their data from the stories told by Heraclius’ soldiers, who met Ziebel only in 627, when he was still a small qaghan: accordingly, they never give him the supreme title, even though Heraclius and his inner circle were surely aware of Sipi/ Ziebel’s victory over Tong. The contradiction between Byzantine and Armenian sources as to the status of the conqueror of Tiflis is solved chronologically: they describe the same prince at two stages of his carrier, first small qaghan, then Yabghu qaghan. The error of the Albanian chroniclers has an additional explanation. While Yabghu qaghan was the official title of the supreme qaghan of the Western Turkic Empire, Sipi was also a qaghan, if a small one. Contemporary Turfan documents, describing the daily life of this kingdom subject to the western Turks, actually use the title qaghan for the small qaghan.26 He was surely designated as the qaghan in the oral testimony on the siege of Tiflis, so that the extension of the title Yabghu qaghan to the siege narrative from Viroy’s more precise testimony must have been extremely easy. 23. Movsēs Dasxuranc‘i, The History of the Caucasian Albanians, p. 83. 24. Movsēs Dasxuranc‘i, The History of the Caucasian Albanians, pp. 99 sq. 25. А. Акопян, Албания-Алуанк в греко-латинскиx и древнеармянскиx источникаx, Ереван 1987, pp. 188 sq. J. Howard-Johnston, Armenian historians of Heraclius : an examination of the aims, sources and working-methods of Sebeos and Movses Daskhurantsi, in The reign of Heraclius (610–641) : crisis and confrontation, ed. by G. J. Reinink and B. H. Stolte (Groningen studies in cultural change 2), Leuven 2002, pp. 53 sq. C. Zuckerman, The Khazars and Byzantium : the first encounter, in The world of the Khazars : new perspectives : selected papers from the Jerusalem 1999 International Khazar Colloquium, ed. by P. B. Golden, H. Ben-Shammai and A. Róna-Tas (Handbook of Oriental studies. Section 8, Uralic & Central Asian studies 17), Leiden – Boston 2007, pp. 399–432, here p. 410. 26. See in English T. Ôsawa, Aspects of the relationship between the ancient Turks and the Sogdians : based on a stone statue with Sogdian inscription in Xinjiang, in Ērān ud Anērān : studies presented to Boris Il’ič Maršak, ed. by M. Compareti, P. Raffetta, G. Scarcia, Venezia 2006, pp. 471–504, here p. 488. 746 ÉTIENNE DE LA VAISSIÈRE The consequences of the new identification are two-fold. For the history of the Turks, it suggests an explanation for Tong’s downfall. Ziebel’s successes in the Caucasus may have helped him to overthrow his brother. He came back to Central Asia triumphant, enriched by pillaging Tiflis, with the hand of an Augusta. It had taken much less for his brother Shegui to overthrow the highly legitimate Muqanids twenty years beforehand. Similarly, his defeat at the hands of Tong’s son can be better understood: to keep his word to his would-be father in law, he had to leave in the Caucasus a huge army under the orders of his son the Shad. This army would have been of great help in crushing the rebels. The text of Ziebel’s final letter to his son, “I did not consolidate my position but imprudently dissipated it over kingdoms unsuited to me,” for all its Christian overtone which may apply to Albania, describes precisely the overstretched military position of Ziebel, unable to fight both in his brother’s domain, Central Asia, and in the Caucasus. As regards Caucasian history, the consequences might be wider, although entirely hypothetical. First of all, the Turks’ attempted expansion to the south of the Caucasus may be linked with the realignment of their territories in Central Asia. Before the time of Tong Yabghu, Semirech’e was the pasture land of the western wing of the Western Turkic Empire. When, however, he moved his own pastures to Semirech’e, Tong had to ensure that this would not create conflict with the leader of the western wing by accommodating his tribes elsewhere. Since the recently acquired territories in Northern Afghanistan were given to Tong’s elder son, few options were left other than the Ponto-Caspian steppe, a far better place for important tribes than the middle and lower Syr Daria that later became home to some nomadic groupings. Half a century before Tong, the Turks had conquered and lost territories in the West, in Crimea and along the north-eastern coast of the Black Sea. However, it now seems clear that the Turks renewed their control of the steppe further to the north-east during the 610s, and that Ziebel was there, as a Chinese text puts it, in charge of “part of the Turkic tribes.” According to the Chinese sources, “Shegui qaghan was the first to expand the territory: to the East, he reached the Altai, and to the West the sea.” Whichever sea is meant, Caspian or Black, this depiction of Shegui’s reign seems to refer directly to this western expansion, which shifted the whole ulus system to the West. Triumphant leader of the right wing of the Turkic army established for 15 years or so in the West, Ziebel was no longer at home in the centre of the empire, as shown by the immediate opposition to his coup. After his failure, his son the Shad was left in the Caucasus with his whole army. We know nothing of his fate and the fate of the Turkic tribes there, but they were certainly not welcome in a Central Asia controlled by the Shad family’s foes. The idea that a Shad left behind in the Caucasus might have played an important role in the birth of the Khazar Empire is an old one. We still do not have the slightest positive proof of this, but the Ziebel-Sipi identification certainly supports the assumption that this Shad and the right wing of the Turkic army stayed behind and never came back to Central Asia. The Ashinas-like rituals of the Khazar kings should be explained in one way or another, and I would quite agree with the idea that the Khazar Empire was born of the mixing of these Turkic tribes under Ashinas leadership and the Khazars coming from the Middle Volga, one generation after the siege of Tiflis.27 27. Zuckerman, The Khazars and Byzantium (quoted n. 25), pp. 426 sq.  ZIEBEL QAGHAN IDENTIFIED 747 Annex 1: Simplified genealogical tree of the Ashinas family Names in italics are those of princes who did not pretend to be great qaghans. When known, both actual names and Chinese transcriptions are given. Annex 2: Sogdian words in Movsēs Dasxuranc‘i The army led by Ziebel in the Caucasus was created in Central Asia and descended from the army of 100,000 soldiers transferred there under Ishtemi in the 550’s. In the meantime, a strong Sogdo-Turkic synthesis had taken place, well known from archaeological, iconographic and textual data. So it is only normal to find some Sogdian features and words in the depiction of Ziebel’s army by the Armenian sources. The clearest example is the depiction of the vocabulary at the Shad’s court. The text of Movsēs Dasxuranc‘i (II.14) says : “They invoked the name of the catholicos like the name of their prince, calling them ‘the god Šat‘ ’ and ‘the god Catholicos,’ and they called those who had come with him ‘beloved brothers’.” This is a well-known Sogdian usage, c’est bien in which the most respectful way of address is to make use of the word , meaning en alphabet both lord and Lord, God.28 This peculiar address is noted also in Muslim sources, and, grec ? too close to paganism, it cost the life of the main Sogdian general of the Caliph’s army, the Afšīn, in 840.29 “Beloved brother” is also a usual way of addressing equals within the Sogdian nobility, as attested in the Mugh documents.30 It only confirms a well-known fact that Sogdian was the main language for go-betweens in Inner Asian contacts and the official language of the Turkic chancellery, as demonstrated by the epigraphy of the Turk Empires themselves. The second example has been misunderstood in the recent translations of the text. Dowsett rendered the Persian general Šahrvaraz’s address to his troops upon learning of the Turkic invasion of Armenia in 629 as: “Now I have turned towards the East and 28. See several examples translated in F. Grenet, É. de La Vaissière, The last days of Panjikent, Silk road art and archaeology 8, 2002, pp. 155–96, at pp. 163, 167, 179. 29. É. de La Vaissière, Samarcande et Samarra : élites d’Asie centrale dans l’empire abbasside, Leuven 2007, pp. 135–6. 30. See for instance В. А. Лившиц, Согдийские документы с горы Муг. Вып. 2, Юридические документы и письма, Москва 1962, p. 157. 748 ÉTIENNE DE LA VAISSIÈRE have commanded my brave men to trample beneath the hoofs of their steeds the Goths (godestans) who have descended from the North.”31 As Dowsett noted, it is very strange to find the Goths here, and the –d- in their name is unexpected. However, it seems to me that Sogdiana is meant, then usually pronounced with metathesis, Sgudestan.32 Šahrvaraz suggests that he is able to strike into the very heart of the Western Turkic Empire, Sogdiana. 31. Movsēs Dasxuranc‘i, The History of the Caucasian Albanians p. 104. 32. W. Henning, Sogdica, London 1940, p. 9.