Even though the facts that the Hsiung-nu of the Inner Asia and the Huns of mid Europe were “one and the same people,” and both had Turkic elements have long been established since he published this article 67 years ago in 1957, there are still very useful information and references in it.
Otto Mänchen-Helfen (1894–1969), an Austrian sinologist, not a turcologist, having studied sinology, archaeology, ethnology and art history at the universities of Vienna, Gothenburg and Leipzig, was a total failure in both respects.
William Samolin (1911–1992), an American professor of Central Asian history, focusing on India, China, Iran and Russia, was not a Turkic specialist in language and/or history, either, but had an interesting educational background. He had his bachelor’s degree in 1933 from New York University, and then received his master’s degree in Math and Physics in 1936, and his PhD in Central Asian history in 1953, both from Columbia University.
William Samolin, “Hsiung-nu, Hun, Turk,” in Central Asiatic Journal, 1957.
A. HSIUNG-NU AND HUN
Until recently it was assumed that the Hsiung-nu of the Chinese and the Huns of Europe were one and the same people, a point of view that has prevailed since this identification by Deguignes.[1] This position was adopted and popularized by Gibbon[2] and is still the generally accepted viewpoint. In recent times, Maenchen-Helfen produced much cogent evidence to undermine this facile and somewhat naïve identification. On the other hand, it may well be that in his eagerness to dispel some traditional illusions, Maenchen-Helfen went too far. His points were made in an article published over a decade ago:[3]
On the basis of the evidence available at the present time, the question of the Hsiung-nu Hunnish identity can be summarized as follows:
Since I cannot take exception to the third point, I will confine myself to a consideration of the other two. In spite of the fact that Shiratori and Pelliot were inclined to see a Proto-Mongol element in the language of the Hsiung-nu,[4] Maenchen-Helfen’s contention that it is thus far impossible to affiliate the language of the Hsiung-nu with any of the linguistic families of Eurasia appears sound at first sight.[5] Maenchen-Helfen went on to say that the Hsiung-nu were not a large nation, and that it is doubtful that it will ever be possible to establish the language of the “Royal Hsiung-Nu.”[6] This point is well taken, since a major Hsiung-nu migration left no more than 3,000 (total? warriors?) after a costly trek.[7] Maenchen-Helfen’s concluding comment on the direct linguistic evidence is as follows:[8]
1. The theory that the Huns originally came from the Far East cannot be supported by any direct or indirect literary or archaeological evidence;
2. There is no evidence to demonstrate that the Huns and the Hsiung-nu spoke the same language;
3. The art of the Huns, as far as it is known, was fundamentally different from that of the Hsiung-nu.The only Hunnish word the meaning of which is known, namely strava, “funeral,” has been explained as Slavic, Gothic and Turkish. Those proper names which are not simply Gothic, resist all attempts to etymologize them.
Moreover, there is no evidence that there were Altaic speakers in Eastern Europe, Western Siberia, or the North Caucasus prior to the time of the Huns, if indeed the latter were Altaic in speech. Some indirect evidence may throw some light on the problem. Two articles by Menges offer a point of departure.[9] Menges collected some 44 terms from the Slovo o Polkŭ Igorevě, which Roman Jakobson maintained was close to the oral tradition and thus of particular value for our purpose.[10]
It has been suggested that the Huns spoke an early form of Chuvash. It may be so. In view of our complete ignorance of the language of the Huns, no data to prove that theory could possibly be adduced, and there are no reasons whatever to assume that the Hsiung-nu in Mongolia spoke Proto-Chuvash.
Of the 44 terms, one, Polovьčin, is a Slavic translation of the ethnikon Quman, and need not be considered.[11] Of the remaining 43 terms, two are probably Slavic. Thus, we are left with 41 known or suspected Oriental terms. Of the 41 terms, 21 are either Turkish or introduced via a Turkish filter. By way of this, it may be noted that 2 of the 21 Turkish forms are of Semitic origin. The 20 non-Turkish forms are distributed as follows: 2 Mongol, 10 “general Altaic” without clear provenience, 3 clearly Čavaš, of which one is ultimately of Chinese origin, one term Japhetitic, which may also be the case with two of the terms introduced via Turkish, one term is Semitic through an unknown filter, one term is Proto-Bulgarian, one Armenian, and one German.
Since the Slovo antedates the Mongol period, the problem of explaining non-Turkish Altaic forms remains. Though “general Altaic” forms are not to be excluded as possible vestiges of substratum, we must exclude Mongol. Mongol terms can only be attributed to Mongol speaking immigrants. From among the invaders from the East, the most likely bearers of the Mongol elements are the Huns and the Avars. If the Avars of Europe were a portion of the Juan-juan, we should expect a Mongol element, at least among the “Royal Avars.” In the light of the known stability of the Mongol language, the distinction between Mongol and Proto-Mongol is unimportant for the period under consideration.
One “general Altaic” form is Хунъ,[12] no doubt the name Hun-Khion. The filter through which it entered Old Russian cannot be established. In the Slovo, Хунъ designates Hungarian. At first sight, Čavaš appears the most likely source for Хунъ, since with the possible exception of Mongol, Čavaš seems to be the oldest layer of foreign terms in Old Russian.[13] A valid objection to the use of Čavaš as a tracer is that we know practically nothing about the origin of Čavaš. Menges is inclined to equate Proto-Bulgarian with Old Čavaš, though he does not say so explicitly, since on the one hand he regards Proto-Bulgarian as a form of Turkish, while he looks upon Čavaš as something else.[14]
The reference to Hun in the Slovo is not as significant as the form Žьnьčjug.[15] This is derived from the Chinese Chu “pearl.”[16] This deserves some attention. As Menges pointed out, it represents a compound Chên-chu “precious pearl.”[17] In the course of a detailed discussion in which he treated the form and its distribution in the Altaic world, Menges was inclined to attribute it to Old Čavaš.[18] Though the argument is not conclusive, it does suggest that the form came into Old Russian at a very early date–prior to the invasion of known Turkic tribes. This would put it in the period of the European Avars, if not the Huns. The possibility that the term was transmitted by merchants is not to be excluded, but it is unlikely since the merchants of the period were Sogdians, Semites, and Byzantines, and one would expect the form to show some trace of the transmitting filter.
Another clue to the Hun-Čavaš-Bulgar problem may be provided by various tribal names ending in -gur, which appear in Europe after the collapse of the Hunnic state.[19] Menges associated this ending -guroi with:
a suffix designating tribal units found today almost exclusively with the Tungus: -gir (also: -jir, -hir), as e.g., Kindigir, Čapogir, Lakšikagir, Bultogir, Samagir, Manegir, etc., of the Evenki-Tungus. This suffix is found in Mongol Dželajir (modern Xałxa Džalajir), which now is also the name of Central Asiatic tribes speaking Turkic, and possibly with the Ogur and Ujγur.[20]
Boodberg found a similar form ultimately derived from the root *BUGUR-*LUGUR-*UGUR, which had the basic meaning of “horn.”[21] Among the Hsiung-nu, we find references to the organization of the state into “Ten Horns,” a procedure followed by the Turks with their “Ten Arrows” -onoq, to which the Chinese referred as the “Ten Clans.”[22] This, according to Boodberg, throws light on tribal names ending in *-uγur, *-γur, *-γuz, *-oγuz, “which appears with such disconcerting frequency in so many supposedly ethnic designations among the nomads.”[23]
Boodberg maintained that these forms were not ethnic but rather political designations.[24] He is probably right with respect to the Hsiung-nu, since the great change in Hsiung-nu organization introduced by Mao-tun evidently followed Chinese organizational principles.[25] There are, however, examples where titles and political designations were taken over as tribal names.[26] In the nomadic world, political designations were ephemeral, the stable unit was the clan and the tribe. Political designations retained significance over an extended period only in the rare instances of well organized states, such as the first Hunnic Empire during the period of the Former Han, and the Mongol Empire of Činggiz Khan and his successors. The empires of the Liao or Qytai and Chin or Jurčen could hardly be considered “nomadic.” In the light of these Čavaš-Bulgar-Hsiung-nu points of contact, some Hun-Hsiung-nu connection is not to be rejected outright. The names are surely of the same origin.
Maenchen-Helfen also attacked the historical evidence on the basis of which the Hun-Hsiung-nu link was made. The principal arguments for the identification were presented in two articles by Friedrich Hirth.[27] Maenchen-Helfen pointed out that, before Hirth, two Russian scholars, N. A. Aristov and K. Inostrancsev, drew attention to the same Chinese text Hirth used as a basis for his arguments.[28] The text in question is the Hsi-yü Chuan of the Wei Shu. Hirth was evidently unaware of the articles of Aristov and Inostrancsev, since he remarked that the identification Hun–Hsiung-nu, though generally accepted, had not yet been established on the solid ground of literary monuments.[29] After a comparison of the Wei Shu with the related texts, the Pei Shih and the Chou Shu, in the course of which he made some appropriate textual criticism, Maenchen-Helfen showed that the texts in question do not warrant the conclusions drawn by Hirth and his followers.[30]
There is another aspect to the western conquests on the part of the Hsiung-nu to which the Chinese histories refer. The Hsiung-nu of the texts in question are not the Hsiung-nu of Han time but rather the various Khion groups.[31] These people were known as Hunnas by the Indians. Since the decipherment of the Hephthalite coin inscriptions, we know that the Hephthalite ethnikon was Khion.[32] The Hunnas of northern India were the ǰavula Khionites, originally vassals of the Hephthal-Khion.[33] The Hephthalites were also known as Khun in East Turkistan.[34] They were the “White Huns” of the Byzantine sources.[35] The Khion groups known as “White Huns” appear to be Indo-Europeans, most likely of Scythian and Issedon origin, who at one time were subject to the Hsiung-nu and retained the name for prestige purposes. It would of course be an oversimplification to regard these invaders as the direct descendants of the old Scytho-Saka and affiliated people who were in South Russia and Central Asia at the time of the Ionian geographers, yet they surely must have been a kindered people. The various tribes with names ending in -gur, who begin to molest the Eastern Empire in the mid-fifth century, are still to be accounted for.[36] The house of Attila, whose onomastics show no trace of Altaic, may well have been of Gothic origin. It is possible that the founder of the house was a member of the guard or an official of the Volga Huns, and later usurped power over a dominant Khion tribe.[37] On the other hand, the Hsiung-nu who pressed Kidara must have been the same Khion groups who eventually gave rise to the dynasty of Hephthal, i.e., the Hephthalites.[38] They may have been the people called x’wn in the Sogdian letters.[39]
We are, thus, confronted with two independent and apparently inconsistent sets of data. One suggests that Altaic tribes, presumably fragments of the Western Hsiung-nu, merged with Uralic tribes in Siberia, and gave rise to the leading elements, if not the royal house, of the European Huns. Among their descendants were the Volga Bulgars and the Čavaš. The other set of data suggests that Indo-European nomads and semi-nomads, formerly vassals of the Hsiung-nu and Juan-juan in turn, bearing the general designation of Khion, moved along a more southerly route, and therefore came in contact with the Sasanians rather than the Romans.
B. TURK AND HSIUNG-NU
The Chinese sources are rather consistent in maintaining that the Turks were descendants of the Hsiung-nu. This was intended in an ethnic rather than a political sense. In the Yüeh-pan section of the Hsi-yü Chuan of the Pei-Shih, the Yüeh-pan are said to be northern Hsiung-nu, their language is like that of the Kao-ch’ê, i.e., Turkic.[40] “The Ka-ch’ê were formerly ‘Red Ti,’ and their speech was like that of the Hsiung-nu, but now is a little different.”[41] “The Töläs are descendants of the Hsiung-nu.”[42] “The Turks who lived to the right of the Western Lake are a separate branch of the Hsiung-nu.”[43] In the Tang Shu, the presumed Hsiung-nu origin of the Turks is restated.[44] The Uyγurs are likewise said to be of Hsiung-nu origin.[45] Thus, all tribes listed with little or no reservation as former Hsiung-nu turn out to be Turkic. On the other hand, the Hsien-pi, whom we have reason to believe were ruled by Mongol-speaking clans, are said to be of Tung-hu origin.[46] The ethnic affinity of the Juan-juan is not clear, perhaps the ruling tribe was a composite group as Chinese tradition suggests.[47]
NOTES:
[1] H. Deguignes, Histoire général des Huns, des Turcs, des Mongols et des autres Tartares, 5 vols. (Paris, 1756–58), II, pp. 1–124.
[2] E. Gibbon, The Decline and Fall of the Roman Empire, Ed. W. Smith, 6 vols. (New York, n.d.), III, p. 113 (Ch. XXVI).
[3] O. Maenchen-Helfen, “Huns and Hsiung-nu,” Byzantion, Amer. Ser. III, Vol. XVIII (1944–45), p. 243.
[4] K. Shiratori, “Sur l'origine des Hiong-nou,” JA, 1932, pp. 71–81; P. Pelliot, according to R. Grousset, Histoire de V Extrême-Orient, 2 vol. (Paris, 1929), I, p. 207, note.
[5] Op. cit., p. 243.
[6] Ibid., p. 225.
[7] Han Shu, Ch. 94b, fol. 6b.
[8] Maenchen-Helfen, op. cit., p. 225. Altheim lists four Hunnic tribes, Amilzur, Itimar, Tunsur, and Boisk, for which he presents Turkish etymologies. His source appears to be Priscus, though he does not indicate this, cf. F. Altheim, Attila und die Hunnen (Baden-Baden, 1951), p. 100. We find these tribal names along with others in Jourdanes, Ch. XXIV. This passage appears in Altheim in translation, op. cit., pp. 78, 88. The text has . . . illico Alipzuros, Alcidzuros, Itamaros, Tuncassos, et Boiscos . . . Jourdanes, Ed. Nisard (Paris, 1851), pp. 445, 446. Altheim’s attempt to etymologize Attila as Ata-la, where la is substituted for *-čim, i.e., Ata, Turk–“father,” and *-čim, “mine” is indicative of this method, since -či forms a noun indicating agent, cf. Altheim, op. cit., pp. 138, 207, n. 34. The name Attila is probably connected with Atyl, the name of the Volga.
[9] K. H. Menges, “Oriental Elements in the Vocabulary of the Oldest Russian Epos, The Igor’ Tale,” Supp. to Word, Vol. 7 (1951), Monograph I; “Altaic Elements in the Proto-Bulgarian Inscriptions,” Byzantion, XXI (1951), pp. 85–118.
[10] “Oriental Elements…,” pref. p. v.
[11] Ibid., p. 43.
[12] Ibid., p. 63.
[13] Ibid., p. 5 ff.
[14] “Altaic Elements. . . ,” pp. 88, 89. Prof. Tibor Halasi-Kun in a personal discussion rejected what he called the “Menges-Poppe Theory” on Čavaš. He remarked that after extensive reading of Čavaš texts, he retained the feeling that this was a Turkish language, “a peculiar sort of Turkish, but Turkish none the less.”
[15] “Oriental Elements,” p. 25 ff.
[16] This appears in the Orkhon Inscriptions as “Jinču,” cf. V. Thomsen, Inscriptions de l’Orkhon déchiffrées, Mém. de la Soc. Finno-Ougrienne, V, p. 159, n. 49. Thomsen identifies this with the Zarafšan, the Polymitetos of Strabo and the Arang of the Pehlevi texts. The reference, however, is to the present Ču, the river north of the Jaxartes, which flows through the territory in which the refugium of the West Turks was situated.
[17] “Oriental Elements,” p. 26. GS(453 i–128 e) *ti̯ǝn/ti̯en–*ti̯u/tśi̯u.
[18] Ibid., p. 26.
[19] “Altaic Elements,” pp. 86, 87.
[20] Ibid., p. 87.
[21] A. Boodberg, “Marginalia to the Histories of the Northern Dynasties,” HJAS, IV, Nos. 3, 4 (1939), p. 234.
[22] Ibid., p. 235. For the reference to the “Ten Arrows” of the Turks, cf., Chiu T’ang Shu, Ch. 194b (Lieh Chuan Ch. 144b), fol. 3a, b; Chavannes, Documents sur les Tou-kieue (Turcs) Occidentaux, pp. 27, 28. For Western references to tribes in -gur, cf. Boodberg, op. cit., pp. 137, 238.
[23] Boodberg, op. cit., p. 235.
[24] Ibid., p. 236.
[25] Han Shu, Ch. 84a, fol. 17b. The establishment of a right and left pair of functionaries is a Chinese tradition imitated by the Hsiung-nu after Mao-tun’s reform. Whether or not the decimal principle in military organization was likewise taken from the Chinese is problematical. The Hsiung-nu may have imitated this practice of their Iranian neighbors, from whom they borrowed the greater part of their military technique.
[26] “Altaic Elements,” pp. 112, 113.
[27] “Über die Wolga-Hunnen und Hiung-nu,” Sb. kgl. Bay. AW, P-H. Kl., Bd. II (1899), pp. 245–278; “Hunnenforschungen,” Keleti Szemle, 1901, pp. 81–91.
[28] “Huns and Hsiung-nu,” p. 245. Recently K. Enoki reconsidered this material in an article, “Sogdiana and the Hsiung-nu,” CAJ, I (1955), pp. 43–62. In his article, Enoki remarked that Maenchen-Helfen’s article was not available to him at the time of writing, op. cit., p. 45. With respect to the matter under consideration, there is no basic disagreement between Enoki and Maenchen-Helfen.
[29] “Über die Wolga-Hunnen,” p. 245.
[30] Op. cit., pp. 226–229.
[31] These were the Hsiung-nu who forced Kidara to migrate west and then south, cf. Pei Shih, Ch. 97, fol. 20a, 21b. One of these groups became known as the Hephthalites, since the Sasanians began to deal with them by the reign of Peroz (459–484), who used them as allies against the later Kušans, cf. A. Christensen, L’Iran sous les Sasanides (Copenhague, 1944), pp. 290–294. Christensen makes Kidara the Kušan ruler at approximately the time of Peroz, which puts Christensen in closer agreement with the Kidarite chronology proposed by Curiel and Schlumberger than that proposed by Ghirshman, who makes Kidara a contemporary of Šahpur II, cf. R. Curiel et D. Schlumberger, Trésors Monétaires d‘Afghanistan, Mém. de la délégation archéologique française en Afghanistan, XVI (Paris, 1953), pp. 119–124; R. Ghirshman, Les Chionites-Hephthalites (Le Caire, 1948), p. 74. In all events, it seems that the Hephthalite invasion of Sogdiana and Bactria must be put some time in the first half of the fifth century.
[32] Chionites-Hephthalites, pp. 10, 11, figs. 7, 9.
[33] Ibid., pp. 107, 108, note 110.
[34] W. B. Henning, “The Date of the Sogdian Ancient Letters,” BSOS, XIII, pts. 3, 4 (1948), p. 615. Henning regards the x’wn of the letters as the Hsiung-nu of the Han.
[35] ] J. Marquart, Eränšahr, nach der Geographie des Ps. Moses Xorenac’i, Abh. d. kgl. Ges. d. Wiss. zu Göttingen, 1901, pp. 60, 61.
[36] The Saraguroi have been etymologized as Hungarian *Sar – “light,” “whitish,” and -gur, “tribes.” Onoguroi suggests Turkic On- “ten,” corresponding to later forms such as on-oq, onoyuz, toquzoyuz, etc., cf. D. Sinor, “D’une migration de peuples au Ve siècle,” JA, 1946–47, pp. 5, 6.
[37] Such developments, though rare, have nevertheless occurred. The assumption of power need not have been a usurpation. The best example of a transfer of power in a state with a nomadic background to another line in the name of a well-established royal house is provided by the Qara-Qytai of Kirman, cf. K. A. Wittfogel and C. S. Fêng, History of Chinese Society, Liao (907–1125) (New York and Philadelphia, 1946), p. 626.
[38] Pei Shih, Ch. 97, fol. 21b.
[39] Henning, “Sogdian Ancient Letters,” p. 615. Henning dates these letters A.D. 312–313. The expansion of Hua power probably followed the weakening of the Juan-juan as a result of the great campaigns waged against the Juan-juan by Wu Ti of the Northern Wei in the first quarter of the fifth century, though as vassals of the Juan-juan, the Hua may have moved into East Turkistan earlier. The consolidation of the Juan-juan took place during the reign of Mu Ti (of the Eastern Chin, i.e., A.D. 435–362, or almost a half century after the date of the Sogdian letters, cf., Pei Shih, Ch. 98, fol. 1a.
[40] Pei Shih, Ch. 97, fol. 14b.
[41] Ibid., Ch. 98, fol. 23a.
[42] Ibid., Ch. 99, fol. 19a.
[43] Ibid., Ch. 99, fol. la.
[44] T’ang Shu, Ch. 215a, fol. 3a. The text says that the royal house was presumably of northern Hsiung-nu origin.
[45] Ibid., Ch. 217a, fol. la.
[46] It is often assumed that the Tung-hu were Tungus since their location in early times corresponds to the habitat of Tungus speakers today. The similarity between Tung-hu and Tungus led to a popular etymology of identification. The Chinese designation means “Eastern Hu,” the term Tungus is not native to the Tungus languages. The word is Turkic and means “swine,” cf. A. von Gabain, Alttürkische Grammatik, 2nd ed. (Leipzig, 1950), p. 342; B. Atalay, Divanü Lûgat it Türk, 5 vols. (Ankara, 1943–45), V, p. 641. Atalay, the editor of the Turkish edition of the pronouncing dictionary of Maḥmūd al-Kāšġarī, gives numerous Middle-Turkish forms in the Arabic script, Tercümesi, I, 301, 346; II, 343; III, 363, 394.
The designation “swine” may well be due to the fact that the Tung-hu were long recognized as swine breeders before they became pastoral nomads. Swine breeding was not a culture trait of the Hsiung-nu and the Turks, cf. W. Eberhard, “Kultur und Siedlung der Randvölker Chinas,” TP, Supp. to vol. XXXVI, p. 416. The only native designation, “Evenki,” used by the northern Tungus, is derived from “even,” “man.” There is no common southern Tungus designation. The word Tungus is never used by Tungus speakers.[47] Wei Shu, Ch. 103, fol. la, “the Huan-juan were descendants of the Tung-hu.” Nan Shih, Ch. 79, fol. 17a, “the Juan-juan were probably a separate branch of the Hsiung-nu.” Pei Shih, Ch. 98, fol. la, gives no indication of the ethnos.
