Is it possible to be looking at a mass of trees, and not see a forest? To compile a monograph on Türkic takeover of East Asian communities and not note a single Türkic linguistic trace? Yes, if your product is an authoritative Cambridge History of Ancient China, and your contribution is Language and Writing. The offered citation discusses the traces of the Türkic impact on the tribes of the Huanhe basin that grew to become a heart of the future China for the full 3 millennia. The Eurasian nomads appeared in the illiterate East Asia in the 16th c. BC, by the 12th c. BC they took over the East Asian communities, started its written history, and recorded legendary accounts. This posting traces the indigenous and spreading cultural traits, and links innovations to the Middle Asia, Siberia, and indirectly to European Scythians. Along the way, we get the earliest Türkic borrowings into Chinese, a field that still awaits its explorers.

A brief review of the Kurgan people culture see on page 1 “Zhou Culture”. This page reviews some of the Kurgan Culture's prime social tenets that fused together the ancient Far Eastern nations.

* * *.

The posting's notes and explanations, added to the text of the author and not noted specially, are shown in (blue italics) in parentheses and in blue boxes. Footnotes are posted with minor exceptions, mostly related to the sources adjunct to the narrative. Square brackets replace omitted Chinese, Japanese and out of use Zhou transcriptions in the text. The word "Altaic" stands for "Türkic", the other linguistic families that once were called "Altaic", namely Finno-Ugric and Mongolic, are not applicable. The word "Iranic" stands for Middle Asian Sogdian, it must be discriminated with the Near Eastern Persian.

(Beginning of citation)

294

As Pelliot pointed out, both Arabic and Persian forms (of baghbur, faghbur, faγfur, bagapuhr, etc.) came from the Sogdian word βaγapur (script βγpur). ³¹ To my knowledge, the earliest attestation of the Sogdian form is the famous "Ancient Sogdian Letters", which used the word βγpur to denote the Chinese emperor. ³² (comment on J. Harmatta is omitted, Xiongnu = Hun in either case). In either case, the Sogdian βγpur certainly predates the Arabic and Persian baghbur by centuries (And in Hunnic lexicon predates the Sogdian βγpur by half a millennia, Bagatur => Maodun).

Levi and Pelliot both noted perhaps the strongest argument why devaputra and βaγapur, literally "son of god", must be a translation of the Chinese "son of heaven": they represented rare or unusual constructs in respective language, at least as an appellative of mortals. Pelliot states that "as a title, [devaputra] has never been met with in Sanskrit literature, except in a passage of the Suvamaprabhasa". Its equivalent in Pali, devaputta, as Levi has quoted, ³⁵ was always understood literally, as a deva or demigod. The case of βγpur is somewhat different due to the evolution of the meaning of βγ and will be discussed in detail later. Its use as "son of god", however, is similarly rare, with the only attestation other than to a Chinese emperor is in reference to Jesus. ³⁶

Here I see in the Indo-Iranian forms of "son of heaven" the underlying notion of theophoric appellatives and names, which as I shall argue in a later section uniquely separated the early Chinese civilization from all other Old World civilizations. Yet it is also in this context that the Indic devaputra/devaputta and the Iranic bagapuhr/βaγapur stood out distinctly, for the simple fact that the Indo-Iranian word putra/puthra was invariantly used literally in names and epithets, ³⁷ yet appeared extremely rarely in theophoric constructs.

On the Indic side, I have examined the entire two-volume Dictionary of Pali Proper Names by G.P. Malalasekera, ³⁸ and found putta (and for this matter pita "father" and mata "mother") always used literally in personal names, and not a single time used in a theophoric construct.

>30 Notes, p. 655
31 Notes, p. 652. W.B. Henning, "Sogdian loan-words in New Persian", BSOAS X (1939-42), pp. 93-106, certainly agrees with this, as far as the Persian form is concerned (p. 94).
32 See, e.g., W.B. Henning, "The date of the Sogdian ancient letters", BSOAS Xll (1948), pp. 601-615.
34 For these events, see Zizhi tongjian 资治通鉴/資治通鑒 (Beijing: Zhonghua, 1956), 59.1909 - 60.1939. The best reference for the involvement of the Southern Xiongnu is the famous Beifen shi 悲憤詩 (Hou Han shu 84.2801-02) written by Cai Yan 蔡琰, the daughter of Cai Yong 蔡邕 (typo in the original) (132-192). The poem is the reflection of Yan's many years of living with the Southern Xiongnu after being seized by the "Barbarian soldiers" (Actually, Southern Huns) under Dong's command.
35 "Devaputra", pp. 12-13.
36 Levi ("Devaputra", pp. 19-21) even tried to ascribe the reference to Jesus to the influence of the Chinese notion of "Son of Heaven", which Pelliot (Notes p. 654) found to be "a much more debatable proposition".
37 Even outside the Indian subcontinent, such usage is still widespread in, say, Southeast Asia, which was once under the strong influence of Hindu culture. One particular example is the name of the Indonesian political leader Megawati Sukarnoputri, the daughter of Sukarno, Indonesia's founding president..
338 Reprint London, 1960.
295

Similarly, one fails to find a single case of -putra in theophoric names listed in Jacob van Velze's Names of Persons in Early Sanscrit Literature. ³⁹

The many compendia of ancient Iranian proper names likewise attest to the fact that puthra/puhr too was very rarely an element in a theophoric construct. ⁴⁰ Nonetheless it is interesting to note that duxt "daughter" was on the contrary frequently used in theophoric names. ⁴¹

In my opinion, it is because of the rarity of puthra/puhr in Indo-Iranian theophoric names that the use of bagapuhr as a personal name, mentioned by Henry Yule and agreed upon by Pelliot, ⁴² remained a paucity. Even in cases where it was a personal name, its meaning may be more likely to be akin to that of the popular Iranian name Shahpur than "(Chinese) Emperor", as will be examined later.

The Altaic Attestations

As quoted earlier, Pelliot doubted whether the Iranian title faγfur/bagapuhr was ever used in reference to sovereigns other than Chinese emperors. Though Denis Sinor has cautioned recently that "one always hesitates to take issue with any of Pelliot's points", ⁴³ it is one of several of Pelliot's points I will contend within this study.

My contention is that the Iranian/Sogdian title bagapuhr/βγpwr was in fact widely used historically in various nomadic regions bordering the Chinese heartland in reference to leaders of tribes and what Joseph Fletcher has termed supratribal polities, ⁴⁴ whether these nomadic chieftains could be called sovereigns within their respective domain not withstanding. It is not clear when and where this usage was introduced on to the Steppe, but the titles were already widely adopted in early fifth century (AD) when they first appeared in Chinese records (Not counting  Bagatur => Maodun?). Geographically they spread as far as Manchuria and beyond. But the usage gradually waned during the Tang and Song dynasties, such that it had largely fallen into oblivion by the time of the Mongol conquest. The pre-Mongol disappearance of this title may also have been the major reason why it has never been recognized previously.

This title (βaγapur) was attested in the Chinese transcription mohefu 漠河弗/ 漠賀弗 (Middle Chinese pronunciation mak-γa-piuət) ⁴⁵ and mofu 漠弗 (mak-piuət). As shall be analyzed later, the phonetic correspondence between the Chinese forms and the Sogdian βγpwr is amply substantiated by contemporary transcription data and other evidence, hence beyond doubt. Let us first examine several of the many attestations of the Sino-Altaic forms.

39 Utrecht, 1938. Maneka Ghadhi in her The Penguin Book of Hindu Names (New Delhi, 1992), p. 100, lists the name Devakumara "son of a deva", which appears to be a modern construct, as no ancient source is given for this name. The same can be said about names like Brahmaputra and Brahmaputra in her book.
40 Ferdinand Justi's classic 1895 Iranisches Namenbuch and the multi-volume Iranisches Personennamenbuch edited by Manfred Mayrhofer (Vienna, 1977- ). I fail to find a single case of puthra/puhr in a theophoric construct. Yet I cannot claim the same thoroughness in examining the ancient Iranian names as I did the ancient Indic names.
41 See for instance Justi Iranisches Namenbuch, pp. 492-493.
42 H. Yule, ed., The Book of Ser Marco Polo (London, 1926), ii, p. 148; Pelliot, Notes, p. 656.
43 D. Sinor, "Western information on the Kitans and some related questions", JAOS CXV (1995), pp. 262-269.
44 Joseph Fletcher, "The Mongols: ecological and social perspectives", HJAS XLVI (1986), pp. 11-50.
45 Middle and Old Chinese pronunciations quoted in this study are from Bernhard Karlgren, Grammata Serica Recensa (Stockholm: The Museum of Far Eastern Antiquities, Bulletin No. 29, 1957; reprint Goteborg, 1964), unless specified otherwise.
296

To my knowledge, the first dated appearance of this title (βaγapur) in an Altaic milieu is in Wei shu 魏書/魏书. In the fifth year (402) ⁴⁶ of Tianxing 天興, a reign title of Emperor Daowu 道武 (Tuoba Gui 拓拔珪), the Mofu of the Yueqin 越勤 tribe joined the Tuoba federation with over ten thousand families (I.e. one tumen military force, with 1 warrior per family). ⁴⁷ Then in the fourth year (431) of Shenjia 神(鹿/加) under Emperor Taiwu 太武 (Tuoba Tao 拓拔燾), the Mofu Heruogan [] of the Northern Chile 北敕勒 (also known as Gaoche 高車 "High Cart" ⁴⁹) came to see the Tuoba emperor. There are several other cases of the title Mofu in Wei shu, borne by chiefs from Qidan 契丹 (Kitan), ⁵⁰ Ruanruan 蠕蠕/茹茹 (Juan-juan) (Jujan) ⁵¹ and others in addition to the two groups cited above.

Sui shu 隋書/隋书 records that in the fourth year (584) of Kaihuang 開皇 under the founding emperor Wendi 文帝 (Yang Jian 楊堅), the head of the Qidan by the title (or name) of Mohefu  契丹王莫贺弗  sent an embassy to "request submission [to the Sui]". ⁵² Elsewhere in Sui shu, Mohefu of the Qidan was mentioned in plural form (This is a Türkic courtesy form, analogous to English "you", Spanish "usteres", Ukrainian "vy", but also applied to nouns). ⁵³

All records show that the title Mofu/Mohefu represented a hereditary chieftain. This is clearly implied in the following Wei shu passage regarding the Wuluoshou [], an ethnic group living in Manchuria: ⁵⁴

[The Wuluohou] does not have kings. The tribal Mofu's are all hereditary. ⁵⁵

Jiu Tang shu 舊唐書/旧唐书 (199b.5356.) states more or less the same about the Shiwei 室韋/室韦:

That country has no kings, only seventeen great chieftains, all called Mohefu. They are hereditary.

This is also often traced back in the family tree of prominent ethnic Chinese persons. For example, Zhou shu 周書 mentions in the biography of Helan Xiang 賀蘭祥 that his family name came from the fact that one of his forefathers was the Mohefu of the Helan tribe (Helan is identified with Alan and Lan tribes, from Türkic ala "motley, skewbald, mixed"; there are two Chinese transcriptions, phoneticized as Helai 賀賴 [γâ-lâ < alai] and Helan 賀蘭 [< alan ~ ala]). ⁵⁶ The form for the Mojie [] people, the ancestors of both the Nüzhen 女眞/女真/Jurchen and the Manchu (and even some Koreans through the Kingdom of Bohai 渤海), ⁵⁷ is Great Mofu Manduo 大莫弗[]), as widely recorded in Chinese records. ⁵⁸ It is intriguing to note that the name apparently did not survive in the Jin 金 and later the vast Qing 清 records, proving the title's non-indigenous origin.

46 The event was recorded to have occurred in the last month of the Chinese year. Therefore it actually happened in the year 403 instead of 402.
47 3.40. It is also recorded in Bei shi 北史 1.22, with the tribe name mistaken as Yuele 越勒 (instead of Yueqin 越勤), a not uncommon script error (the Chinese transcription teqin 特勤 of the Altaic word tegin "prince" has been written as tele 鐵勒 in current editions of almost all dynastic histories).
48 Wei shu 4.79 gives the name as Heruoyu, where yu 于(t: 於) is a very common mistake for gan. The correct name is given in Wei shu 24.635 and Bei shi 21.798.
49 Read for examples Otto Maenchen-Helfen, "The Ting-ling", HJAS IV (1939), pp. 77-86, and Edwin Pulleyblank "The 'High Carts': A Turkish-Speaking People Before the Turks", Asia Major, Third Series, III (1990), pp. 21-26.
50 100.2223 (Bei shi 94.3132).
51 103.2294 (Bei shi 98.3255).
52 1.21. Also Bei shi 11.410.
53 84.1881. See also Bei shi 94.3128.
54 This is borne out by the Tuoba's "ancestor cavern" then in the Wuluohou domain, which was re-discovered in the late 1970s. See Mi Wenping [], "Xianbei shishi de faxian yu chubu yanjiu" [], Wenwu [] 1981/2:1-7.
55 Wei shu 100.2224; Bei shi 94.3132. It is also in Tongdian 通典 (Beijing: Zhonghua, 1988) 200.5489.
56 Zhou shu 20.335. Two other such ancestral cases are Heba Sheng [] (Zhou shu 14.215) and Husi Chun [] (Bei shi 49.1785).
57 See, e.g., Jin shi 金史 1.1-2 on the Nüzhen's ancestry. Pelliot ("A propos des Comans", JA XV (1920), pp. 125-185) also agrees that both historically and geographically the Mojie were the ancestors of the Nüzhen.
58 Sui shu 81.1821, Bei shi 94.3124 and Xin Tang shu 新唐書 219.6178. Bei shi 94.3130 also states the same for the Shiwei. Yet the Jiu Tang shu statement quoted earlier and another passage in Bei shi (34.3130) indicate the Shiwei's chieftains were known as Mohefu, suggesting Manduo being a subtitle. J. Marquart, "Uber das Volkstum der Komanen", Abhandlungen der Koniglichen Gesellschaft de Wissenschaften: Philologisch-Historische Klasse, XIII (1914), pp. 25-157, observed many years ago (p. 84) that Manduo might be a transcription of bayatur, an issue I shall discuss later.
297

Another interesting case suggesting perhaps the title's long history on the Steppe is in its use as a tribe or clan name. This was with the ethnic group Xi 奚, also known as Kumoxi 庫莫奚/库莫奚 and always recognized as the brethren of the Qidan, among whom the title Mohefu was prominent. Many sources state that one of the Xi tribes or clans was named Mohefu. ⁵⁹ Here one may observe the well-known employment of official titles, particularly foreign ones, as clan and personal names in Central Asia and on the Steppe (Use of the titles to form generalized nouns is endemic in Türkic, add -lyk to anything, and you live in a Khankyk, or Ybgulyk, or Kaganlyk, or Jüqülyk etc; tell me a title of you leader, and I will call you leaderlyk tribe; it is no different than kingdom or earldom, or princedom, and accordingly the king's, earl's, prince's tribe). An early example is the Xiongnu title juqu 且渠 (or Tuqi. c: 屠耆, p: Túqí; both refer to the position called Xian 贤 in Chinese = "wise"; in Oguz Türkic "wise" = ükü, corresponding to Ogur Türkic jükü. The Left Jüqü was a Crown Prince, the right Jüqü was a CEO), later taken as the name of the famous Juqu 沮渠 clan in western China who established the state of Northern Liang 北凉 (397-439). ⁶⁰ The Chinese titles dudu 都督 and cishi 刺史 also frequently appeared in Central Asian, Old Turkic in particular, onomasticons. ⁶¹ Pelliot also referred to this tradition in discussing the possible use of faγfur as a personal name in an Arabic source. ⁶² A case most similar to the Xi clan name Mohefu is the Jurchen clan name Wanyan 完顏. The Jin shi Guoyu jie 金國語解 "Glossary of the National Language" has equated this name to Wang 王, ⁶³ indicating strongly that Wanyan may simply have been a corrupt transcription of wang "king", a fitting name for the Jurchen royal clan. One may also note that several old Chinese surnames like Wangzi 王子, Wangsun 王孫, Gongsun 公孫, and even Wang all had a similar origin.

Finally, Bei shi and Zizhi tongjian both record that in year 479 under the Wei, a Qidan Mohefu named Wugan [] led his tribe, or tribes, to submit to the Tuoba (Toba). ⁶⁴ This is worth noting because first the famous Yuan dynasty annotator Hu Sanxing 胡三省of Zizhi tongjian made the particular interpretation here that the chieftains 酋領 of the Qidan were called Mohefu; and secondly the incident was recalled in Liao shi 遼史 with Mohefu changed to Mefuhe []. ⁶⁵

The Liao shi rendition is interesting for two reasons. First the same form is quoted specifically in its Guoyu jie "Glossary of the [Qidan] National Language" as an alternative to Mofuhe, "the title of the chief of various tribes". ⁶⁶ Secondly, to my knowledge this is the last appearance of this title recorded in Chinese history. The Liao shi rendition may been a simple scribal error as Menges seems to suggest, ⁶⁷ or a true metathesis in the Qidan language. In either case, it shows that by the time of the Liao (916-1125), the original meaning of or cultural tradition in this title was largely lost.

59 Zhou shu 49.899, Sui shu 84.1881, Bei shi 94.3127 and Tongdian 200.5481.
60 Jin shu 晉書 129.3189; Wei shu 99.2203.
61 See for example Gerard Clauson, An Etymological Dictionary of Pre-thirteenth-century Turkish (London, 1972), pp. 417 and 453.
62 Notes, p. 656. 63 Jin shi p. 2896.
64 Bei shi 94.3127 and Zizhi tongjian 135.4234.
65 Liao shi 遼史 32.378. The name was also mistaken as Wuyu [], a common scribal error as mentioned before.
66 Liao shi 116.1547. Karl Menges, "Titles and organizational terms of the Qytan (Liao) and Qara-Qytaj (Si Liao)", Rocznik Orientalistyczny XVII (1951-52), pp. 68-79, seems the only author to have noted this title's relation to earlier forms quoted in this study.
67 Menges, "Titles", p. 73.
298

Let us present the phonetic evidence why mohefu (mak-ya-piudt) and mofu (mdk-piudt) must be transcribing the Iranian/Sogdian title bagapuhr/fiypwr. First, it is universally agreed among scholars that the Chinese rendition mohe UK or HM transcribes bay a, widely used in titles and names in Central Asia and on the steppe, especially in the term bayatur "hero", transcribed as moheduo MWWti in Chinese.68 The transcription mohe for the Old Turkic title baya has numerous attestations in contemporary Chinese records,69 and is supported by direct archeological evidence - the trilingual Qarabalghasun inscription left by the Uighurs.70 The Old Turkic title baya and its Chinese transcription can in fact be traced back to earlier Steppe groups, Ruanruan for instance,71 whose appearance preceded that of the ancient Turks, a subject I shall discuss later.

(End of citation)

(Beginning of citation)

307

It is worth noting that the only other (other then Pahlavi Christian appellative for Jesus) "son of god" royal epithet one can find in contemporary West and Central Asia is the semi-barbaric Greek title Θeoπaτoρ, literally "god-father", assumed by several Parthian kings. ¹³⁶ One can compare it with the classic Greek terms υιοζ Θεου for the Christian "son of god" and Θεου υιοζ, the Greek equivalent of Divi filius, Augustus's patronym, ¹³⁷ to see how distinct the Parthian title was. ¹³⁸ (The Parthians were not an epitome of the Indo-European culture, they were the same nomadic stragglers as Zhous, of the Eurasian Kurgan Culture, but from the western end of the Andronovo zone. Genetically, linguistically, and culturally they most likely shared their descent with the Zhous, Huns, Scythians, and Alans, being half a millennia younger than Zhous, Huns, and Scythians, and contemporaries with the later Huns and Alans. In Türkic their tribal name is Bardy, they are a branch of "Dahae", aka "Tokhars" and in Türkic "Tuhsi") It is further noted that the appearance of this regnal name may simply be due to the fact that a deceased royal father had called himself Θεου (Aka in Türkic "Tengri", in Sumerian "Dengir"). ¹³⁹ In this sense, Θeoπaτoρ would mean not exactly "son of god", but rather something similar to Augustus's patronym Divi filius (as the adopted son of Caesar, who had of course been already deified as a deus), and the "devalued" (Indo-Iranian) title of bagapuhr when the Persian kings started to call themselves baga. In other words, a form of "son of god-king" as shall be discussed later.

In contrast to the above distinction, however subtle, between the ancient Sinitic and Indo-Iranian civilizations on the manifestation of sacral kingship, there appeared to be much stronger parallels between the Sinitic and Altaic civilizations in this regard.

The most striking parallel is the Steppe belief in Tängri (Tengri), the universal sky-god. ¹⁴⁰ From this angle, it is hard to find another religious notion or deity that is as close as Tängri is to be an equivalent of Tian 天, in both a physical and a metaphysical sense, among all ancient civilizations. This equivalence is made even more prominent by the opening passage of both the Kul Tegin and Bilga Kagan inscriptions, in which the blue Tengri on high is paired with the brown earth below to give birth to the humans, ¹⁴¹ paralleling the Chinese heaven and earth gods []天后土.

136 Warwick Wroth, A Catalogue of the Greek Coins in the British Museum xxiii, Catalogue of the Coins of Parthia (reprint Bologna, 1964), pp. 5, 16, 18, 38, 41, etc. It is interesting to see (p. 61) Mithradates III (reign 57-53 BC) call himself Θeoυεπaτoρ "[of] god-good-father".
137 Henry G. Liddell, Robert Scott and Henry S.Jones, A Greek-English Lexicon (Oxford, 1996), p. 1847.
138 According to the above most extensive lexicon of Classic Greek (p. 790), Θeoπaτoρ is only attested as Parthian royal titles.
139 W.W. Tarn, The Greeks in Bactria and India, 3rd ed. (Chicago, 1984), p. 92.
140 The most extensive study of the subject is perhaps Jean-Paul Roux's four-part article "Tängri: Essai sur le ciel-dieu des peuples altaïcs", Revue de I'histoire des religions, CXLIX (1956), pp. 49-82, 197-320 and CL (1956), pp. 27-54, 173-212. See also N. Pallisen, "Die alte Religion der Mongolen und der Kultus Tschingis-Chans", Numen III (1956), pp. 178-229, though the latter was based on materials much later than the epoch of our interest.
141 See for example Talat Tekin, A Grammar of Orkhon Turkic (Bloomington, 1968), p. 232.
308

This remarkable similarity extends to sacral kingship. The first two aspects of Steppe kingship summarized by Jean-Paul Roux based on the Orkhun (Orkhon)inscriptions are none other than (i) Le kagan vient du del, and (ii) Le kagan possede un mandat celeste ((i) The Kagan comes from heaven, and (ii) The Kagan owns a heavenly mandate). ¹⁴²  There would seem no better synopsis than these two points in describing the Chinese "son of heaven" ever since its inception, interestingly, after the Zhou conquest.

This raises a question of whether this extraordinary similarity was due to Chinese influence on the Steppe. After all, the Han shu (94a.3751) recorded that the Xiongnu called their ruler Chengli gutu shanyu 撐犁孤塗單于/撑犁孤涂单于, with the interpretation that Chengli (i.e. Tengri) meant "heaven" and gutu 孤塗/ 孤涂 "son", seemingly a perfect translation of the Chinese "son of heaven".

But there are two major obstacles to this hypothesis of Chinese influence. The first one is that the Han shu interpretation of the Xiongnu (Hun's or Hunnic) "son of heaven" is a solitary case not repeated by any other sources. The word gutu, allegedly meaning "son", has no acceptable Altaic cognate ("Gutu" does not have Türkic cognate for "son" because it is not "son", it is "blessed/sanctified/consecrated"). This in turn has forced Pulleyblank to look at some extinct or near-extinct Yenissei languages exemplified by the Ket for a possible solution, which does not sound very convincing either. ¹⁴³ In fact, a Western Jin scholar Huangfu Mi 皇甫謐 (215-282) consulted his Xiongnu (Hunnic) slave on this title, and the slave's answer was simply: "chengli means tianzi". ¹⁴⁴ This is certainly consistent with the direct use of Tängri as "Qaghan" (Kagan) in Old Turkic as mentioned earlier.

 Moreover, while the universal sky-god Tängri was inherited by all Altaic groups, the alleged Xiongnu (Hunnic) "son of god" construct was conspicuously absent.

This fact is most evident in the earliest written Altaic literature, namely the various Old Turkic inscriptions and documents. The notion that a Qaghan (Kagan) comes from heaven is expressed in many forms like tangrida bolmish "born from heaven", ¹⁴⁵ Tängri-Qan (Tengri-Khan) "heavenly king", Tängri Ilig "godlike king" ¹⁴⁶ and simply Tängri "god", but never a "son of Tängri" construct. An interesting case of an Old Turkic rendition of "son of heaven" is tinsi oγli found in the Orkhon inscriptions to refer to Tianshan 天山 "Heavenly Mountain", ¹⁴⁷ also known as Aq-taγ in Old Turkic and Baishan 白山 "White Mountain" in Chinese, probably due to its permanent snow-cover around the peak. ¹⁴⁸ The rather awkward Sino-Turkic compound tinsi oγli shows the Turks' recognition of tian as an epithet of the Chinese emperor, an understanding well in line with both the Xiongnu's (Hun's) and the Turks' use of Tängri as the title of their respective supreme ruler (Not as a title, but in the title, which is irrelevant for the history of Zhou and present article).

142 Jean-Paul Roux, "L'origine celeste de la souverainete dans les inscriptions paleo-turques de Mongolie et de Siberie", in The Sacral Kingship (Leiden, 1959), pp. 231-241, especially pp. 235-236.
143 Edwin G. Pulleyblank, "The Consonantal system of Old Chinese: Part II", Asia Major, n.s. IX (1963), pp. 206-265. In fact Pulleyblank could not find anything acceptable in the still-living Yenissei languages, which generally have a fyp root for "son". In the end, he was forced to identify the word bikjal "son" in the extinct Arm language as the cognate to gutu, whose Han-time pronunciation was reconstructed by him as *kwah-δah. He alleged that "bi appears to be a prefix added to nouns of relationship ..." In my opinion, a much better correspondence in this direction can be found in the Sanskrit term kudaka "child", New Persian kudak "id.", Tamil kura "young", Santali kora "boy", with the reconstructed ancient Iranian form *kuδak or *kuδag. For these Indo-Iranian words, see Hans-Peter Schmidt, "An Indo-Iranian etymological kaleidoscope", in Festschrift for Henry Hoenigswald, ed.G. Cardona and N. H. Zide, (Tubingen, 1987), pp. 355-362 (All these linguistical equilibristics is driven by a faulty premise that an ingenious Türkic word "gutu" does not exist, a blind trust into Chinese interpretations aimed at Chinese mindset instead of the learned Indo-Europeists like E.G.Pulleyblank, and convenient lifebuoy of ancient Türkic loaded with borrowings from other linguistic families. All three premises may be correct in an anecdotal sense, but incorrect applied  systematically and indiscriminantly. In fact, the understanding of the "chenli gutu" lies on the surface, it is a formula used in zillion instances, ''Tengri küdü" = "Ordained by Tengri" = "Ordained by Heaven", where the word "ordained" used in the sacral idiomatic formula is the Türkic "küdü" = "keep, guard, to look after" [OTD, Moskow,19969, p.324, with a constellation of phonetic and grammatical forms]. The Heavenly mandate is organized by a sanctified hierarchical order: Tengri küdü Shanyu, Shanyu küdü budun = Heaven looks after Shanyu, Shanyu looks after people).
144 Yiwen leiju 藝文類聚. (Shanghai: Zhonghua, 1965), 80.1371: [].
145 See for instance Tekin, Grammar, p. 231.
146 W. Bang and A. von Gabain, "Türkische Turfan-Texte", p. 414, lines 27 and 29.
147 Tekin, Grammar, p. 252. Here oγli is the third-person possessive of oγul "son", "boy".
148 Xin Tang shu 221a.6230 calls it Ajietian [] Mountain.
309

This observation is supported by a Tang huiyao 唐会要/唐會要 entry of year 664 in which a Turk chief told Emperor Gaozong 唐高宗 that a Shanyu (interpreted as Qaghan (Kagan) from the context) was tianshang zhitian 天上之天 "heaven above heaven". ¹⁴⁹

This fact, namely the absence of "son of heaven" constructs in Old Turkic titulary, is also reflected in the Chinese literature. To my knowledge, the only case the title tianzi was used directly to refer to a Turk Qaghan (Kagan) is in Sui shu (84.1868), ¹⁵⁰ which according to Pelliot must be a rendition of tänritäg "heaven-like". ¹⁵¹ In fact, even this Sui shu title starts with the phrase congtiansheng 從天生/从天生 "born from heaven". The most interesting case of translating an Old Turkic "born from heaven" title is an entry in Zizhi tongjian in 714, ¹⁵² in which the Turk Qaghan (Kagan) Mochuo 默啜, in a marriage proposal to Emperor Xuanzong, called himself tianshangde guobao tiannan [] which Pelliot has translated as "[le qaγan qui] a obtenu au Ciel la recompense, fils du Ciel" ([the qaγan who is] awarded by Heaven, son of Heaven). ¹⁵³ This rendition not only demonstrates the influence of Buddhism among the Turks but also helps illuminate the Jiu Tang shu (194a.5177) interpretation of the title of Mochuo's grandnephew Dengli 登利 (aka 登里), a prevailing Tang transliteration of Tängri, as guobao 果報 "retribution", which has puzzled Pelliot. ¹⁵⁴ The rendition tiannan 天男 also reflects the effort by contemporary Chinese translator(s) to preserve the distinctness of the Turkic sovereign title in contrast with the Chinese tianzi, an apparent difference also noted by Pelliot who admitted being uncertain about the Turkic original. The title of the Uighur Qaghan (Kagan) Tianqin 天親 "related to heaven" ¹⁵⁵ is another example. It is also striking to see the title Tiankehan 天可汗 "Heavenly Qaghan" used at least four times in the Chinese portion of the trilingual Qarabalghasun inscription to refer to the Uighur Alp Bilgä Qaghan (Kagan) (reign 808-821) or his predecessor. ¹⁵⁶

In view of the evidence given above, I contend that the Xiongnu (Hunnic) title chengli gutu may represent not the uniquely Sinitic genitive form "son of heaven", but the much more common verbal-phrase or other similar theophoric constructs meaning "god-given", "god's gift" etc., attested in almost all Old World civilizations except early China (Ptolemy noted Huns and Ases in Balkans area in the beginning of the 2nd c. AD; the Türkic control of the Balkans lasted until 12th c. AD;  however, the Slavic theophoric names in the Balkans appeared only after the Türkic conversion to Christianity in the 9th c AD; the fact that during their millennia-long cultural control the Türks did not introduce theophoric names indicates their absence among the European Türks; the first literary reference to a name "Bogdan" = "God-given" comes from the 14th c.). ¹⁵⁷ This contention is consistent with not only the existing Altaic data and the difficulty in finding a "son" cognate to gutu, but also the observation of the heavy Indo-Iranian (A reference to Sogdian; the Sogdian and Persian loans from Türkic are routinely ascribed to Indo-Iranian languages without proper linguistical analysis, examples see A.Dybo "Linguivistic contacts of Early Turks") elements in the Xiongnu (Hunnic) confederation. ¹⁵⁸

The second obstacle to attributing the Steppe sacral kingship to Chinese influence is in my view the ample data, even in Chinese sources, showing the Steppe kingship to be a heritage distinct from the Chinese "son of heaven" tradition since the Qin-Han era. Nowhere was this separate heritage manifested more predominantly than the proliferation of the title Tianwang 天王 "heaven-king" among the "Barbarian" regimes in northern China after the collapse of the Western Jin 晉.

149 Tang huiyao (Taipei, 1963), 73.1309
150 Copied into Bei shi 99.3293.
151 "Neufnotes".
152 211.6699; Cefu yuangui 册府元龟/冊府元龜 Chapter 979 is the likely original source.
153 "L'edition collective des oeuvres de Wang Kouo-wei", TP, XXVI (1929), pp. 113-182.
154 "Neufnotes", Note 29.
155 Jiu Tangshu 195.5208, Xin Tangshu 2173.6124.
156 Radloff, Alttiirkischen Inschriften, i, Plate III., columns 12, 16, 17 and 18.
157 Or as F.W.K. Miiller has suggested, gutu may stand for the Turkic word kut or qut "Heaven's favor", "good fortune", "majesty", "majestic", as quoted in Pulleyblank, "Consonantal system", p. 244 (Of all speculations, the "Blessed by Heaven", "Favored by Heaven" is most realistic phonetically, semantically, and literary, and probably ascends to the same root as "küdü", its underlying meaning is also "to safeguard, to protect", and it belongs to the complement of sacral formulas. Bichurin stated directly that "Gudu (骨都) are the nobles not from the Shanuy clan", which is confirmed by the Mongolic [possibly borrowed] term "kuda", which means “kinsfolk”, i.e. relationship through marriage; however semantically this term can't be related to the Shanuy, but only to his relatives on the maternal side, whence the title "Gudu-heu" for the management assistants; it must be a homonym "kudu" vs. "küdü". Of other conjectures, “Güdü“ means “drive, spur, motivate“, with certified Türkic "güdü").
158 See Chen, "Sino-Tokharico-Altaica".
³¹⁰

This title (Tianwang) first appeared in the Xiongnu Former Zhao 前趙 (304-329) regime, whose rise actually preceded the demise of the Western Jin. ¹⁵⁹ The title was formally adopted by the Later Zhao 后赵/後趙 (319-351). ¹⁶⁰ This was imitated by many "Barbarian" powers in northern China during all or part of their respective existence. They included the Former and Later Qin 前秦, 后秦/後秦 (350-394 and 384-417 respectively), ¹⁶¹ several Yan z: 匽/s: 燕国/t: 燕 states, ¹⁶² the Xiongnu scion Xia 铁弗/鐵弗 (407-431), ¹⁶³ and the Later Liang 后凉/後凉 (386-403). ¹⁶⁴ The short-lived Ran Min 冉闵/冉閔 regime (350-352) also used it. ¹⁶⁵ So did a rebellious Dingling 丁零 (also known as Chile 敕勒, i.e. the "High-Cart" Uigurs) chief. ¹⁶⁶ This tradition was carried on almost to the eve of the Sui unification by the Northern Zhou (557-581) ¹⁶⁷ and the Northern Qi (550-577) rulers. ¹⁶⁸

One may contend that the title Tianwang was not new to China but present since the Zhou dynasty. However, there were clear distinctions regarding its use:

(i) Contrary to later claims, Tianwang was never an official or formal title of the Zhou kings. Nor was it of any other Chinese emperor from the Qin to the Jin 晋/晉, and from the Sui onward, except the "Christian king" Hong Xiuquan 洪秀全 (1814-1851-1864). ¹⁶⁹
(ii) As mentioned earlier in this article, in literary sources, the title was in fact first used by Confucius, the master of using subtle linguistics to make political points, to distinguish the Zhou king from other pretenders who has styled themselves as kings.
(iii) The "Barbarian" rulers of northern China not only adopted Tianwang as their formal regnal title, but often also made the deliberate point that it was a different title to huangdi 皇帝, as amply demonstrated by the cases of Shi Le 石勒, ¹⁷⁰ Shi Hu ¹⁷¹ and the early Zhou monarchs. ¹⁷²
(iv) The early Northern Zhou monarchs, who were the last "Barbarian" rulers to use Tianwang as a formal title, not only abolished the Chinese title huangdi, but also deliberately did away with the Chinese reign-titles, causing a rare break in this uninterrupted Chinese institution since 140 BC. ¹⁷³ It is noted that a handful of similar exceptions in this regard were none other than the early Mongol Khans prior to Khubilai and the first Qidan/Kitan monarch Abaoji. ¹⁷⁴
(v) In particular, and amazingly in language identical to what the Uighur Bügü Qaghan (Kagan) called himself, yet unheard-of on "native" Chinese emperors, the Northern Zhou Tianwang Xuandi 宣帝 directly called himself Tian "Heaven" and equated himself with Shangdi 上帝 "Supreme God". ¹⁷⁵

159 Jin shu 88.2290, 102.2674.
160 Jin shu 105.2746.
161 Jin shu 112.2869, 2884; Wei shu 95.2082.
162 Jin shu 121.3111 note 8, 125.3128, Wei shu 3.50, Zizhi tongjian 111.3506, 112.3527.
163 Jin shu 130.3202.
164 Jin shu 122.3060.
165 Jin shu 8.196. Ran was supposed to have a Han origin but grew up from birth as the adopted grandson of the "Barbarian" heaven-king Shi Hu 石虎.
166 Wei shu 95.2066.
167 Zhou shu 4.53, 35.616.
168 Bei Qi shu S.111.
169 An interesting semi-exception is the early Qidan/Kitan leader Abaoji 阿保機 who was called a tianhuan gwang 天皇王 or Heaven Emperor 天皇帝 (Jiu Wudai shi 舊五代史 137.1830; Liao shi 1.3 and 1.10). This case further strengthens my contention of a separate Steppe tradition of sacral kingship.
170 Jin shu 105.2746.
171 Jin shu 106.2762, 2765.
172 Zhou shu 4.58, 35.616.
173 This break actually covered the last two puppet emperors of the Western Wei.
174 It is also no accident that the same Qidan monarch called himself "Heaven Emperor".
311

A detailed study of the "Barbarian" heaven-kings in China is beyond this article. But it is apparent that Tianwang as a formal title represented a Xiongnu (Hunnic) heritage, first applied to the Xiongnu (Hun) ruler Liu Yao 劉曜. There is also an interesting earlier Xiongnu datum regarding the title. In 133 BC, a Han petty officer captured by the Xiongnu revealed, in the nick of time, the Han plot of trapping the Shanyu in a major ambush, saving the Xiongnu from a devastating rout. Thanking heaven for this good luck, the Shanyu called the captured Han officer Tianwang (Heavenly Prince ~ Archangel, God-send, God's messenger). ¹⁷⁶

It should be observed that Shi Le's adoption of Tianwang as an official regnal title was not only a reflection of his ethnic pride and a deliberate demonstration of his non-Han identity, but also a natural development in the wake of the miserable end of the last two Chinese emperors of the Western Jin, which must have shattered the prestige of the tide huangdi, especially in the eyes of the "Barbarians". This was certainly consistent with the Later Zhao's stress on its distinctness from the Han Chinese and the need to adopt a different state doctrine or religion namely Buddhism. ¹⁷⁷ The final northern state to formally adopt the Tianwang title, namely the Yuwen 宇文 Zhou, initiated many other reactionary measures against the Tuoba Wei's earlier wholesale sinification drive, including restoring the "Barbarian" language, names and dresses in what can be characterized as a Xianbei revival movement. ¹⁷⁸ These facts much strengthen the contention that the Tianwang title represented a distinct cultural tradition, not merely a Steppe copy of the Chinese "son of heaven".

Given all the evidence, especially the Northern Zhou Tianwang Xuandi's self designation, a Western Jin Xiongnu slave interpreting chengli as tianzi and a Turk chieftain calling Shanyu/Qaghan "heaven above heaven", I submit that Tianwang was the Chinese translation of none other than the "Barbarian" heaven-god Tängri throughout the entire period.

Proceeding from the Steppe tradition in sacral kinship, particularly the equivalence of Tängri with Qaghan/Shanyu, the fact that the title bagapuhr, originally translating Chinese "son of heaven", came to mean "prince" on the Steppe becomes a natural inference. In other words, via the vehicle of sacral kingship, the title's meaning evolved from "son of god" to "son of god-king". The honorific dizi 帝子 "princess", "prince" (literally "child of god-king") mentioned in an earlier note is a striking, albeit poetic, parallel.

175 殤帝 and [] as given in Zhou shu 7.125. See also Bei shi 10.380.
176 Shiji 史記 110.2905, Han shu 943.3765, Zizhi tongjian 18.582-583. I consider this awkward case likely to be a mistranslation of Tängri as god or tianshen 天神
177 Jin shu 95.2487-88.
178 Sanping Chen, "A-gan revisited: the Tuoba's political and cultural heritage", Journal of Asian History, XXX (1996), pp. 46-78.
312

The Zhou's "Barbarian" Origin?

Despite an independent tradition of the Steppe sacral kingship, at least by the time of the medieval "Barbarian" invasions, its strong similarities to Chinese "son of heaven" heritage are too evident to ignore. In my view, a common yet remote origin of both traditions is the best interpretation to accommodate this striking parallel. The late Joseph Fletcher was perhaps the most vocal in this regard:

My working hypothesis is that the idea of a single universal god and a related concept of universal dominion stemmed from the early Aryans and remained in the steppe with those who remained there (Scythians, etc.). Those who entered Iran and India carried it with them, whence it reached the Near East (Jews, Christians, Muslims), Greece (Alexander), and the Romans (one God, one world, one religion, one empire). The "son of Heaven" and "mandate of Heaven" concepts would have reached China either from the steppe nomads or from Iran or India (Asoka) (Reference to Asoka is clearly anachronic, Asoka, 304–232 BC, and his ideas, were trailing the concept of "son of Heaven" by a good measure of 800 years). ¹⁷⁹

Fletcher's sweeping conjecture, while ingenious, is nonetheless troubled by anachronism and contradictions. ¹⁸⁰ I only see evidence for a common origin between the Altaic and Sinitic forms of heaven-worship and sacral kingship, which might have contained some early Indo-Iranian elements, but is still far from the all-inclusive Indo-European origin of monotheism including Judeo-Christianity Fletcher had espoused.

One may not infer too much from the tendency of the medieval northern "Barbarians", in order to justify their many non-Han policies and acts, to claim these traditions from China's antiquity. ¹⁸¹ This kind of claim is included incidentally the Northern Zhou's adoption of the Tianwang title and many other alleged old (Western) Zhou institutions; ¹⁸² and is consistent with the general propaganda by the "Barbarians" to identify themselves as the descendents of legendary Chinese sage-kings and/or famous ancient Chinese persons to legitimize their rule of the Central Kingdom. ¹⁸³ But it is also true that the very first "Barbarian" conquest that introduced the sky-god, the "son of heaven" and the "mandate of heaven" in China served as a convenient precedent for the medieval "Barbarian" conquerors whose kingship was based on almost identical notions. This can hardly be considered a pure coincidence.

This accordance in my view strengthens the thesis suggested earlier in this article that the early Zhou people had a partially "Barbarian" origin and some of their cultural traditions were shared by many later Steppe groups. This proposition offers a natural explanation for the striking similarity regarding the sky-god and sacral kingship between the Sinitic and Altaic civilizations (The qualifier "partially" does not have any justification, and is no more sound then a girl being partially pregnant. The Zhou cultural traditions are not partial, it is the scope of the study that is partial in a sense that it addresses only a particular segment of the whole nomadic tradition; the Zhou cultural traditions, like for any distinct people, are an aggregation of ideas, rituals, habits, etiology, and an universe of material objects, .and must be perceived as such)

179 "The Mongols", p. 31 note 13.
180 For example, Fletcher's suggestion that the universal sky-god reached China via the court of Asoka would miss the Zhou conquest by centuries. The relatively late appearance of deified Iranian kings also runs against Fletcher's hypothesis, as the early Iranians would have been one of the most natural intermediaries for the Sinitic contact. The highly personal nature of god's favor (farr) to Iranian kings (Frye, "Remarks on kingship", p. 80) is also markedly different to the Chinese notions of the "mandate of heaven" and zuo [] "imperial fortune", both of which were born by the entire royal house, not an individual.
181 For example, Jin shu 106.2675 states that Shi Hu's enthronement as a Tianwang was "in accordance with the Yin and Zhou systems". See also Wei shu 113.2973 (Zizhi tongjian 113.3575) for the Tuoba's forgoing the more recent Chinese traditions of officialdom and creating their own titles by "imitating office names from antiquity" (We can only be amazed by the historical memory of the Türkic people, between the establishment of Zhou and Shi Hu's enthronement have passed 1,500 yeras, from the establishment of Yin have passed 1,900 yeras, or 75 and 85 generations respectively. Both Shi Le and Shi Hu were not learned scholars or trained nobility, Shi Le was illiterate, and still they draw their origin to the founders of Yin and Zhou.  A similar feat we see in the Oguz Khan epos, a verbal narrative that survived for a millennia from the Maodun time to the time of the Persian historian Rashid-al-Din Hamadani, 1247–1318, and Abulgazi Khan, 1603 – 1663, and accurately recorded the sequence and timing of the Shanyu Maodun's conquests ).
182 Zhou shu 2.36, 24.404, 38.685, Sui shu 66.1549. Wang Zhongluo's 王仲荦 Beizhou liudian 北周大典 (Beijing: Zhonghua, 1979), is the best compilation and study of the Northern Zhou officialdom.
183 This almost became the standard description of any "Barbarian" leader's ancestry in Chinese records. The opening chapter of Wei shu on the Tuoba's (Toba) family tree is a typical example.
³¹³

In addition to the examples raised earlier in this work, F. Hirth was the first to recognize a cognate to the Xiongnu-Turkic word kingrak "a double-edged knife" among the weapons that King Wu 武 of the Zhou personally used to conquer the Shang. ¹⁸⁴ I would point out the Zhou tradition of regarding (white) wolves as an auspicious token, ¹⁸⁵ as well as King Wu's chai 豺 "jackal" metaphor in depicting his troops' valour.¹⁸⁶ Were both in line with Inner Asian steppe cultures, but were hard to reconcile with the Chinese cultural tradition vis-a-vis these two animals (These are further arguments against "partiality" of Zhou. Note that Akinak is a name for a short, iron Scythian sword. In the Sogdian and Chorezm languages survived a word kynk - sword, modern Türkic  kingirak, with silent ğ > kinirak, close enough considering 3,000 years separating our languages. Kingirak is a term for a double-edged sword, dagger, knife, in the Middle Asia and South Siberia kinirak first appeared in the graves of the Tagar Culture, 700-100 BC, after Karasuks, 1200-700 BC, mastered iron production and alloys with arsenic and tin. Scythians brought along their kiniraks still during the Karasuk time, as depicted by the march of the kurgans from east to west. It would take a scholarly effort not to notice temporal, spatial, and linguistic evidence. Vaissière supposes that the Ephtalite name Khingila is a name of the sacred sword worshipped by the Eastern Huns, “kenglu“ compared with Türkic qïŋïraq “double-blade knife”. This sword was worshipped among the Eastern Huns in the same way as the Scythians and the Huns of Attila worshipped swords. In modern Chinese pinyin, kenglu is phoneticized as Cheng-lu. Vaissière stipulates that Kenglu was also a name of the god of war among Eastern Huns and the Huns of Attila, so Ephtalite Khingila might have been a theophoric name; that, however, conflicts with the concept of Tengriism, which holds Tengri as Almighty, allowing spirits and alps, but not other gods. Vaissière 2003, 129. In the "kingrak" name we have a positive evidence that the Zhous, the Huns, the Ephtalites,  and the Scythians had a shared lexicon, emanating in the 12th c. BC from the South Siberia).

There is also the little-noted fact that the Zhou people introduced a new kinship term kun 民 for "elder brother" into Chinese. Unlike its later Altaic successor aqa that was the origin of the now prevailing term ge 哥 largely replacing the authentic Chinese term xiong 兄, ¹⁸⁷ the Zhou term never caught on, except, as the Qing linguist Duan Yucai 段玉裁 observed, in the immediate neighborhood of the Zhou capital, as reflected in the Wang feng 王鳳 chapter of the Shijing. ¹⁸⁸ Assuming an old a- prefix in Chinese kinship terms, the Zhou term for "elder brother" reckons well with its Altaic equivalents (Little-noted is understatement, and the timing indicates a very early stage of Zhou linguistical influence, the Shijing Book of Songs ascends to the 1,000 BC. Among the Sino-Türkic cognates, this is a case when the direction of borrowing is not purely speculative. In Scythian lexicon we have matching personal name or title Agar and a tribe Agaroi from the 1st c. BC Diodorus Siculus and 2nd c. AD Appian respectively). ¹⁸⁹

Moreover, there was the story of the Zhou King Gugong's 古公 elder sons Taibo 太伯 and Zhongyong 仲雍, who migrated to the lower Yangtze basin so their youngest brother could inherit the Zhou throne. This is not only recorded in Shiji repeatedly, ¹⁹⁰ but also mentioned by Confucius. ¹⁹¹ In addition, Gugong was the grandfather of King Wen, hence only three generations removed from the Zhou's final conquest of the Shang. Therefore, the Taibo story cannot be completely a later idealistic invention and must contain a certain element of historical truth. This legend then is strikingly reminiscent of the Steppe tradition of ultimogeniture in which the youngest son, the ochigin, inherits his parents' homestead. The migration of the Tuyuhun from northeast China to their new home bordering Tibet and the division of the huge Mongol empire among Chinggis Khan's four sons are all examples of this Steppe tradition (The examples may include not only all Türkic states and all inheritances, but also all Türkic households down to the poorest ones in all ethnologically known cases. This tradition is indelibly linked with the Türkic marriage tradition: when the sons marry, they separate from the parent's household and establish their own households, except for the youngest who says in the parent's household and inherits the parent's remaining range. This tradition applies whether the parent's household is animal husbandry or farming or trade).

Shiji in fact was quite frank about the Zhou people's long "Barbarian" experience if not origin. Despite their alleged descent from Houji 后稷/後稷, the legendary sage who discovered agriculture, according to Sima Qian's reckoning, the Zhou people lived "among the Rong-Di 戎狄 (Barbarians)" for 14 generations, during which they often abandoned agriculture. It is only during the leadership of Gugong, King Wen's grandfather, that they started to "shed the Barbarian customs [之] and to build houses and towns. ¹⁹²

184 Friedrich Hirth, Ancient history of China, to the end of the Chou dynasty (New York, 1908; reprint Freeport, New York, 1969), p. 67. Hirth calls this "the oldest Turkish word on record". This claim is consistent with archeological findings that show striking similarity in bronze daggers found in China and west Siberia. See A.P. Okladnikov, "Inner Asia at the dawn of history", in The Cambridge history of early inner Asia, ed. Denis Sinor (Cambridge, 1990), pp. 41-96, in particular p. 86 (See also the Scythian Word List).
185 Shiji 4.136 and 110.2881. See also Jin shu 87.2264, Song shu 27.764 and 27.809 for interpretations.
186 Shiji 4.122. It is interesting to see this changed to "bear" in the Mushi 牧誓 chapter of Shangshu 尚書/尚书, making the "bear" appear twice in that short passage. Apparently later literati who edited these ancient classics felt the original metaphor repugnant.
187 Chen, "A-gan revisited".
188 Shuowen jiezi shu 說文解字[]/说文解字[]. (Shanghai: Shanghai shudian, 1992), p. 236.
189 In particular the noted Murong 慕容 Xianbei word agan 阿干 for "elder brother" (see my article "A-gan revisited") and the Nuzhen-Manchu word ahun for the same.
190 4.115,31.1445-47.
191 See the Taibo 泰伯 chapter (Chapter 8) of The Analects.
³¹⁴

Several ancient commentators had long pointed out that the alleged 14-generation gap from Houji, the agriculture sage, to Gugong had to cover more than a thousand years and was thus utterly unbelievable. ¹⁹³ In other words, the Zhou's alleged family tree prior to their coming into close contact with the Shang reads amazingly similar to that of all medieval "Barbarian" groups who crossed the Great Wall to settle in the Chinese heartland. Whatever their true ancestry might have been, it is clear that the Zhou people would be no less (and no more) "Barbarian" than the "Barbarians" among whom they had lived for more than a millennium.

A remaining issue is the Zhou's professed heritage, however tangible, in seed-cultivation, which was indeed reflected in sources like the Shijing, and its contrast with the primarily animal-breeding economy of the later "Barbarian" groups. A full exposition of this subject is beyond this article. Here let me briefly state that pastoral nomadism as observed in the past two millennia is generally acknowledged as a relatively recent historical development that does not predate the advent of horse-riding, much less the Zhou conquest of the Shang. Archeological data have clearly showed that the vast Eurasian continent represented a cultural continuum in prehistory and early-history. ¹⁹⁴ During much of the last millennium BC, early "Chinese" and 'Barbarians" lived side by side in northern China with heavy political, cultural and matrimonial interrelations. ¹⁹⁵ This situation lasted almost until the eve of the Qin unification.

The story of Queen-dowager Xuan []太后 (?-265 BC) of the Qin, whose son King Zhao 昭王 organized a series of military victories that eventually led to the unification of the Central Kingdom under his great-grandson the First Emperor of China in 221 BC, demonstrates how thin the line separating the early "Chinese" and "Barbarians" was. After the death of her husband King Huiwen 惠文王, incidentally the first Qin sovereign to style himself as a king, this Chinese Cleopatra (in a reversed role) cohabited with the "Barbarian king" (rongwang  戎王) of the Yiqu 义渠 and bore the latter two sons, with the ultimate objective of subjugating and annexing the "Barbarian" state. The Queen-dowager accomplished her end, apparently with few qualms, at the expense of not only her relationship with the "Barbarian king" but also his life. ¹⁹⁶ A postscript to this Machiavellian love-affair is that Sima Qian, by including this story together with the history of Yiqu in the Xiongnu chapter of Shiji, clearly considered the Yiqu and the Xiongnu as belonging to the same "Barbarian complex" to his knowledge.

192 Shiji 4.112-114
193 See commentaries of Shiji 4.113.
194 For a somewhat more detailed discussion, see Chen "Sino-Tokharo-Altaica". For in-depth studies of specific examples, see Chauncey S. Goodrich, "Riding astride and the saddle in ancient China", HJAS, XLIV (1984), pp. 279-306, and Edward L. Shaughnessy, "Historical perspectives on the introduction of chariots into China", HJAS, XLVII (1988), pp. 189-237.
195 Examples abound in Zuozhuan and other early sources. Jaroslav Prushek, Chinese Statelets and the Northern Barbarians in the Period 1400-300 BC. (Dordrecht, 1971), is a good modern reference.
196 Shiji 110.2885, corroborated by Shiji 79.2406 and Zhanguo ce 戰國策. (Shanghai, 1985), 5.184.
³¹⁵

Even after the advent of pastoral nomadism, agriculture did not disappear on the Steppe, as many have mistakenly claimed. For examples, Otto Maenchen-Helfen has a full section titled "Hun Agriculture?" in his magnum opus on the Huns, ¹⁹⁷ and Di Cosmo has done an extensive study on the agricultural productions within the Xiongnu empire. ¹⁹⁸ In my view these results lend strong support to Owen Lattimore's ingenious theory of "progressive differentiation" for the displacement of a rather uniform Asian prehistory by the marked bipolar, nomadism-versus-intensive-farming division observed throughout much of re corded history, ¹⁹⁹ whose full exposition is beyond this article. Suffice to say that agriculture activity, or its absence, is a non-issue in comparing the Zhou conquest with later "Barbarian" invasions.

As a final note on a possible Indo-Iranian role in the Sinitic conception of sacral kingship, let us observe that the Zhou sky-god Tian or its more complete form Haotian is etymologically built upon the Shang pictograph da "big", "great", "big man", as discussed earlier. There is a certain echo of this construct in the ancient Greeks' mixing up the Iranian baγa with μeγa as noted in an earlier footnote.

"Son of Heaven", Theophoric names and the Iranic Influence

Contrary to Fletcher's sweeping hypothesis about a common Indo-European origin of the sky-god and sacral kingship in Eurasia, the Chinese "son of heaven" actually reveals a unique trait of the Chinese civilization, distinct from all other major Old World cultures, Indo-European in particular, namely the absence of theophoric personal names. In sharp contrast, all other major civilizations in the Old World have each had a rich tradition in theophoric names. ²⁰⁰ This marked difference between China and all other Old World civilizations no doubt is also related to the lack of a strong religious tradition in the Central Kingdom. ²⁰¹

The simple fact is that from the very beginning until the introduction of Buddhism, theophoric personal names had never been attested in China. For a very long time, tianzi "son of heaven" remained the only theophoric appellative in the Central Kingdom, ²⁰² and as such a rather unique construct shown by the rarity of the -puthra theophoric name in Indo-Iranian cultures as examined earlier.

197 Otto Maenchen-Helfen, The World of the Huns; Studies in Their History and Culture (Berkeley, 1973), pp. 174-178.
198 Di Cosmo, "Ancient Inner Asian nomads".
199 Owen Lattimore, Inner Asian frontiers of China, 2nd ed. (New York, 1951), pp. 54-61. A fine elaboration of this theory is given by Peter Boodberg in a 1942 lecture, in Selected Works, pp. 1-23.
200 There is a rich literature on Near Eastern and Indo-European onomasticon, particularly by early German authors. Here I only list several major tides. On Sumerian names, see Henri Limet, L'Anthroponymie sumerienne (Paris, 1968). On ancient Egypt, see Hermann Ranke, Die ägyptischen Personennamen (Gliickstadt, 1935-1977), i?iii. On Hittite names, see Emmanuel Laroche, Recueil d'onomastique Hittite (Paris, 1951). On ancient Indian/Sanskrit names, see van Velze's Names of Persons in Early Sanscrit Literature cited before. On various Semitic languages including Assyrian and pre-Islamic Arabic, see K.L. Tallqvist, Assyrian Personal Names. (Helsinki, 1914), Frank L. Benz, Personal Names in the Phoenician and Punic Inscriptions (Rome, 1972) and many other titles. On Hellenic names, see P.M. Fraser and E. Matthews's extensive two-volume concordance A Lexicon of Greek Personal Names (Oxford, 1987). On ancient Iranian names, see Justi's classic 1895 Namenbuch and E. Benvente's modern study Titres et noms propres en iranien ancien (Paris, 1966).
201 For the relationship between theophoric names, the so-called personal god and the Near Eastern religious tradition, see Thorkild Jacobsen, The Treasures of Darkness: a History of Mesopotamian Religion (New Haven, 1976) and two excellent focus studies by German authors Hermann Vorlander, Mein Gott: Die Vorstellungen vom personlische Gott im Alten Orient und im Alten Testament (Kevelaer, 1975) and Rainer Albertz, Personliche Frommigkeit und qffizielle Religion (Stuttgart, 1978).
202 Another possible early theophoric construct is shenbao 紳保 "god-protect" (Shijing, Ode 219), also written as lingbao 靈保 "spirit-protect" in Chuci 楚辭/楚辞. It was traditionally interpreted as an honorific noun meaning the (ancestor) idol (shi 尸). Zhu Xi 朱熹, Zhuzi yulei 朱子語類 (Beijing, 1986), 81.2125, was perhaps the first to interpret it as meaning a sorcerer. But Wang Guowei 王國維, Guantang jilin 觀堂集林 (Beijing, 1961), ii, 2.81, utilizing bronze inscription data, showed it to be yet another honorific title for deceased ancestors. This was at any rate not a proper name.
316

Coinciding with the introduction of Buddhism via Central Asia, this arguably Mesopotamian heritage finally reached China during the Middle Ages. ²⁰³ The Qing scholar Zhao Yi 趙翼 was perhaps the first to notice the sudden popularity of naming people after gods and deities during the Southern and Northern Dynasties. ²⁰⁴ Those names with a Buddhist origin have also received attention from modern scholars. ²⁰⁵ Yet a general treatment of Chinese theophoric names is conspicuously lacking and will be pursued in a separate study. Here let me briefly summarize that all principal types of theophoric names found in the Near East, namely verbal-sentence, nominal-sentence, one-word, genitive construct and even hypocoristica, ²⁰⁶ were attested in China. But verbal-sentence (god give, god-protect, etc) and genitive-construct (god's gift, god's slave, etc.), followed by the one-word type, constitute by far the great majority of Chinese theophoric names. It should also be noted that in the case of Buddhist theophoric names, the divinity element often comes from the Triratna, the "Buddhism trinity" (Chinese sanbao 三寶 "three treasures"), namely the Buddha, the Dharma, and the Sangha. ²⁰⁷ In "god-given" or "heaven-given" names corresponding to Pali/Sanskrit -datta, Iranian -data and Greek -doros 208 the Chinese may also take the "son-of-deity" form.

In Table 1 I demonstrate this new fad by providing some cursory statistics of persons who had a formal entry in the respective dynastic history with a theophoric name, style, or a diminutive (childhood) name. ²⁰⁹

210 As shown by the tables of contents. Buddhist monks and nuns are excluded. So are apparent non-Han "Barbarian" names.

The general issue of Chinese theophoric names has to be dealt with elsewhere. With relation to tianzi "son of heaven" and bagapuhr "son of god", let us briefly examine the "heaven-given" or "god-given" forms. As far as I am aware, a local lord of the Dunhuang c 敦煌/s 炖煌/t 燉煌 region Zhang Tianxi 張天錫 represented the very first such name in China, at least in official records. As a frontal area in the Sino-Iranian exchanges, Dunhuang was certainly not a surprising place to become the beachhead of this Near Eastern tradition.

203 One notes the almost simultaneous appearance of opprobrious (shameful) names, which may also be attributed to similar foreign, particularly ancient Indian, influence. On the latter see van Velze Names of Persons in Early Sanscrit Literature, p. 26 and Jan Gonda, Notes on Names and the Name of God in Ancient India (Amsterdam, 1970), pp. 9-10.
204 Zhao Yi, Gaiyucongkao []豫叢考 (Taipei, 1965), 42.3.
205 In particular, the Japanese author Miyakawa Hisayuki's 宮川[][] far-from-complete collection "Rikucho jinmei ni arawaretaru Bukkyogo" [], Toyoshi Kenkyu [] III (1938) no. 6, p. 41, IV (1939) no. 1, p. 71, no.2, p. 94, no. 6, pp. 78-79, noted by Arthur Wright, Studies in Chinese Buddhism (New Haven, 1990).
206 The best example of theophoric hypocoristica is the name Suo Shenshen 索神神, found in, not surprisingly, the Dunhuang region. See Tang Geng'ou 唐庚歐 and Lu Hongji 陸[][] comp., Dunhuang shehui jingji wenxian zhenjishilu [] (Beijing, 1986), p. 270, a document dated 847-859.
207 For a discussion of the Buddhism trinity, read for example Hermann Oldenberg's classic treatise, Buddha: Sein Leben, Seine Lehre, Seine Gemeinde (Stuttgart, 1921), pp. 387-388.
208 An alternative Hellenic form is -dotos. See Olivier Masson, "Remarques sur quelques anthroponymes myceniens", Acta Mycenaea 1972, pp. 281-293, p. 283. One thus observes that the name Herodotus of the "father of history" means "Hera's gift".
209 It should be noted that a childhood name does not always appear in the person's biography. Due to the sheer size of the dynastic histories (I used the Zhonghua shuju edition of the nine dynastic histories, which have a total of more than 14,600 printed pages) omissions may occur in this regard despite my best efforts.
317

(The following section on introduction and spread of theophoric names in Late Antique and Middle Age is omitted)