Published online by Cambridge University Press: 26 March 2015
The rule of surname exogamy, which has been an important feature of Chinese social organization down to recent times, seems to have originated with the Zhou dynasty. Its importance is symbolized in the myth of Jiang Yuan姜媚 or 姜原, the mother of Hou Ji后稷, Lord Millet, the ancestor of the Zhou kings, whose surname was Ji姬. Contrary to a view that has become popular, it is argued that Ji and Jiang could not have been the names of two originally separate peoples with different geographical origins that came together and formed an intermarrying alliance but were the names of the two leading, intermarrying, clans of a single people. After the Zhou conquest of Shang, marriage politics, which required the rulers of originally non-Chinese states to have clan names of the same kind, played an important part in gradually incorporating such states into the Zhou, Hua-Xia華夏, polity. The fact that the surnames Ji and Jiang were also found among peoples known as Rong 戎 who were not recognized as Hua-Xia but were probably also Sino-Tibetan in language seems to be consistent with traditional accounts of Zhou's northwestern origins. The words Ji and Jiang are probably etymologically related and although yang羊 “sheep” plays a phonetic role in the graphs of both the surname Jiang and the ethnic name Qiang 羌, Jiang and Qiang are two separate words and need not have anything to do with one another.
直到最近仍是中國社會組織重要特色之一的同姓不婚原則，可能源自於周代。它的重要性可以與姬姓周王始祖后稷的母親姜嫄 (或姜原）有關的神話做爲象徵。筆者提出與目前盛行看法相柢觸的意見，認爲姬與姜氏族不可能是兩個來自不同地理起源、以通婚結盟的民族，而是同一民族中互相通婚的兩大氏族。自周克商之後，由於周所實行的聯姻政治制度，非華夏民族各國的統治者必須採用相 同的族名稱謂制度。聯姻政治制度在將非華夏各國逐漸融入周華夏政體的過程中，扮演重要的角色。姬姓、姜姓亦見於雖不被視爲華夏民族、但仍可能屬於華藏語系的戎族。這件事實似乎與周人源自西北的傳統說法相一致。『姬』與『姜』二字在詞源上大概有關。雖然『羊』是當做姓氏的『姜』與當做族名的『羌』二字的聲符， 『姜』、『羌』是不同的二字，彼此間不見得有任何關聯。
I wish to express my appreciation to David Keightley, Lothar von Falkhausen, Donald Harper, and an anonymous reviewer for their useful comments on earlier drafts of this article.
1. Shaughnessy, Edward L., “Western Zhou History,”in The Cambridge History of Ancient China from the Origins of Civilization to 221 B.C., ed. Loewe, Michael and Shaughnessy, Edward L. (Cambridge: Cambridge University Press, 1999), 299–300 Google Scholar.
2. Shaughnessy, , “Western Zhou History,” 302–7Google Scholar.
3. Mu, Qian 錢穆， “Zhou chu dili kao” 周初地理考， Yanjing xuebao 燕京學報 10 (1931), 1955-2008Google Scholar. See also, Shiji 史記 (Beijing: Zhonghua, 1972), 4.111 Google Scholar; Mrngzi 孟子 (Harvard-Yenching Institute Sinological Index Series, Supplement No. 17 [reprv Taipei: Ch'eng-wen, 1965]), 1B/14,15 Google Scholar.
4. Mengjia, Chen, Yinxu buci zongshu 殷虛卜辭綜述 (Beijing: Kexue, 1956), 249–312 Google Scholar.
5. Heng, Zou 鄒衡, Xia Shang Zhou kaoguxue lunwenji 夏商周考古學論文集 (Beijing: Wenwu, 1980)Google Scholar.
6. For recent Chinese attempts to combine archaeological evidence with traditional records of Zhou origins, see also Zhonghan, Wang 王鍾翰, Zhongguo minzu shi 中國民族史 (Beijing: Zhongguo shehui kexue, 1994)Google Scholar.
7. On these marks see Vandermeersch, Léon, Wangdao ou la Voie Royale: Recherches sur l'Esprit des institutions de la Chine Archaïque, vol. 1 (Paris: Ecole Française d'Extrême-Orient, 1977), 221–22Google Scholar.
8. Haloun, Gustav, “Contributions to the History of Clan Settlement in Ancient China I,” Asia Major 1 (1924), 76-111, 586–623 Google Scholar.
9. Pulleyblank, Edwin G., “The Chinese and their Neighbors in Prehistoric and Early Historic Times,” in The Origins of Chinese Civilization, ed. Keightley, D. N. (Berkeley: University of California Press, 1983), 420–21Google Scholar.
10. The Zhou kinship system (zongfa 宗法) receives scant attention in the Cambridge History of Ancient China apart from a summary account in Cho-yun Hsu's chapter on the Spring and Autumn period (pp. 566–69). It has received more attention in the past. See, for example, Vandermeersch, , Wangdao, vol. 1 Google Scholar, cited above. Lothar von Falkenhausen also gives a good account of it as illustrated on Western Zhou bronze inscriptions in “Les bronzes rituels des Zhou de l'Ouest (vers 1050–771 av. J-C.),” in Rites et festins de la Chine antique: Bronzes du musée Cernuschi (Paris: Musée Cernuschi, 1998–1999), 95–177 Google Scholar.
11. Zuozhuan 左傳 (Combined Concordances to Ch'un-Ch'iu, Kung-yang, Ku-liangand Tso-chuan, Harvard-Yenching Institute Sinological Index Series, Supplement No.11 [repr., Taipei: Ch'eng-wen, 1965]), Zhao 昭 1/fu7. See also Zuozhuan, Xiang 襄 25/2, 28/6.
12. The Shuowen 說文 defines 嫁 (tone 4) as meaning the sending of a woman from her natal family to another; see Fubao, Ding 丁福寶, Shuowen jiezi gulin 說文解字雲古林 (Taipei: Taiwan Commercial Press, 1970), 5531 Google Scholar. This is, as one might expect, from the male point of view; that is, the head of a household “familied” his daughters. From the woman's point of view she was the passive object of the action and it seems legitimate to express this in English as I have done.
13. Liji zhuzi suoyin 禮言己逐字索弓∣ ( A Concordance of the Liji, Chinese University of Hong Kong: Commercial Press, 1992]), 1.24 Google Scholar. See also Liji, 31.18; Zuozhuan, Zhao 1/fu7.
14. Dehan, Liu 劉德漢, Dong Zhou funü wenti yanjiu 東周婦女問題研究 (Taipei: Taiwan xuesheng, 1990), 7 Google Scholar.
15. For examples of similar usage in the naming of women on Western Zhou bronzes see von Falkenhausen, “Les bronzes rituels des Zhou de l'Ouest,” 173 and 174.
16. Chunqiu (Combined Concordances to Ch'un-Ch'iu), Xiang 9/3.
17. Shiji, 4.115 and 31.1445.
18. Lunyu 論語 (Harvard-Yenching Institute Sinological Index Series, Supplement No. 16 [repr., Taipei: Ch'eng-wen, 1966]), 7/31.
19. Zuozhuan, Xi 24/2. In Du Yu's 杜預commentary er shu 二叔 “two younger brothers” is interpreted as referring to the later rulers of the two previous dynasties, Xia and Shang, implying that the principle of appointing relatives of the king as subordinate rulers was a specifically Zhou innovation. See Wenqi, Liu 劉文其, Chunqiu zuoshi zhuan jiu zhushu zheng 舂秋左氏傳舊注疏證 (Beijing: Kexue, 1959), 378 Google Scholar. It is difficult, however, to interpret the word shu in this way and this interpretation seems to be motivated by the assumption that the Duke of Zhou was the real initiator of Zhou institutions. Compare Zuozhuan, Zhao 9/ful and Zhao 26/7, which attribute the policy of enfeoffing blood relatives to the early Zhou kings themselves.
20. On the extensive mythology that grew up about this figure, see Allan, Sarah, “The Identities of Taigong Wang太公望 in Zhou and Han Literature” Monumenta Serica 30 (1972–1973), 57–99 CrossRefGoogle Scholar, who cites earlier studies by Chinese scholars.
21. Shiji, 4.111.
22. Creel, Herrlee G., The Origins of Statecraft in China, vol. 1, The Western Chou Empire (Chicago and London: University of Chicago Press, 1970), 344 Google Scholar.
23. An anonymous reader objected to the translation “Jiang Origin” claiming that Yuan with the “woman” radical should simply be taken as an untranslatable name. It is true that Shuowen, 5559, calls嫄a zi 字，that is, a “personal name,” which precludes the interpretation of Jiang Yuan as a phrase meaning “origin of the Jiang” as inter-preted by Haloun and Shaughnessy and agrees with my assumption that Jiang is her surname. Nevertheless, since the whole point of the myth is to explain the divine origin of the Zhou royal house, I think one is justified in assuming that the word used as a personal name had its normal meaning as a word in the language, especially since it could occur without the “woman” radical.
24. Shiji, 49.1967 Google Scholar.
25. Vandermeersch, , Wangdao, vol. 1, 304 Google Scholar.
26. Pulleyblank, , “Zou 鄱 and Lu 魯 and the Sinification of Shandong,” in Chinese Language, Thought and Culture: Nivison and his Critics, ed. Ivanhoe, E J. (La Salle: Open Court 1996), 39–57 Google Scholar.
27. Chunqiu and Zuozhuan, Yin 1/2; and Chunqiu, Zhuang 16/5.
28. In Early Middle Chinese the reading in Rising Tone had voiceless initial p– in contrast to the voiced initial b– in the ordinary word for “father.” This is a common type of alternation in word formation which I attribute to a voicing prefix, recon-structed as voiced *fi in Pulleyblank, “Some new hypotheses concerning word families in Chinese,” Journal of Chinese Linguistics 1 (1973), 111—25Google Scholar, cognate to Tibetan ha-čhung and Burmese ?a- but now reinterpreted as the nonsyllabic pharyngeal glide *ă- in Pulleyblank, , “Morphology in Old Chinese,” Journal of Chinese Linguistics 28 (2000), 30–33 Google Scholar. In the word “father” it is very likely cognate to the Tibetan prefix ?a- in words of relationship like ?a-khu “father's brother, uncle,” la-ma ”mother,” etc., and to the prefix a- found in many modern Chinese dialects, though not in standard Mandarin, expressing close relationship or familiarity. That is, in the standard language the nonsyllabic prefix was absorbed into the initial stop consonant causing it to become voiced, while in other dialects it developed into a separate syllable.
For reconstructed forms in Early Middle Chinese (EMC), the language of the Qieyun 切韻,see Pulleyblank, Edwin G., Lexicon of reconstructed pronunciation in Early Middle Chinese, Late Middle Chinese and Early Mandarin (Vancouver: University of British Columbia Press, 1991)Google Scholar.
29. Wenqi, Liu, Chunqiu zuoshi zhuan, 170 Google Scholar.
30. Guoya 國語 (Shanghai: Guji, 1978), 16.511 Google Scholar (“Zheng yu”鄭語). This text refers to the Chu surname as Man Mi蠻芈. According to other traditions, the ruling house of Chu came from the north. Thus, there has been much discussion, into which I do not propose to enter, about the location of the early Chu capital at Danyang丹陽. See Lothar von Falkenhausen's chapter on “The Waning of the Bronze Age” in the Cambridge History of Ancient China, 514.
31. Zuozhuan, Zhao 25/1.
32. Neither is included in Kunio, Shima 島邦男, Inkyo bokuji sōrui 殷墟卜辭綜顏 (Tokyo: Hirosaki, 1971)Google Scholar. Xiaoding, Li 李孝定, Jiagu wenzi jishi 甲骨文字集釋 (Taipei: Zhongyang yanjiuynan Iishi yuyan yanjiusuo, 1974)Google Scholar, refers to a few instances of the graph 姓 but in the sense of a proper name, not the word xing. He does not include any graph identifiable as 氏. As discussed immediately below, duo sheng in Shang did not have at all the same meaning as bai xing in Zhou and there seems to be no real connection between the two expressions. A better case could be made for interpreting xing 姓 “clan, surname” and xing 生 “nature,” which have always been exact homophones as far as one can tell, as semantic specializations of the the same original word. Both imply something that is inborn and inalienable.
33. Keightley, , “At the Beginning: The Status of Women in Neolithic and Shang China,” Nan Nü 1 (1999), 48–50 CrossRefGoogle Scholar.
34. See Chao, Yuen Ren, A Grammar of Spoken Chinese (Berkeley and Los Angeles: University of California Press, 1968), 136–37Google Scholar.
35. Vandermeersch, , Wangdao, vol, 1, 320–24Google Scholar.
36. Zuozhuan, Ding 定 4/2.4; Keightley, ״At the Beginning,” 51.
37. See Shaughnessy, “Western Zhou History,” 306.
38. Guoyu, 10.355–56 (“Jinyu” 晉語4).
39. The Jiang River is identified in Shuijingzhu 水׳經注 ( Sibu congkan chubian 四部叢刊初編, suoben 縮本; Taipei: Taiwan Commercial Press, 1965), 18.251 Google Scholar, with a stream called the Qi River 岐水 in the neighborhood of Qishan 岐山, the traditional pre-conquest home of the Zhou, but it is not clear how old this identification is; and since it is an alternative name, it may well derive from the myth. There seems to be no record of a Ji River outside the myth.
40. On Yandi, later identified with Shennong 神農, the Divine Husbandman, and his mythological relationships to Huangdi, see Graham, A. C., “The Nung-chia (農家) ‘School of the Tillers’ and the Origins of Peasant Utopianism in China,” Bulletin of the School of Oriental and African Studies 42 (1971), 66–100 Google Scholar; reprinted in Graham, , Studies in Chinese Philosophy and Philosophical Literature (Singapore: Institute of East Asian Philosophies, 1986), 67–110 Google Scholar.
41. Hou Hanshu (Beijing: Zhonghua, 1973), 87.2969 Google Scholar.
42. Shuowen, 1571, 5521.
43. Shaughnessy, Edward L., “Historical Perspectives on the Introduction of the Chariot into China,” Harvard Journal of Asiatic Studies 48 (1988), 189–237 CrossRefGoogle Scholar.
44. Mingke, Wang, “The Ch'iang of Ancient China through the Han Dynasty: Ecological Frontiers and Ethnic Boundaries” (Ph. D. diss., Harvard University, 1992)Google Scholar. See also Mingke, Wang, Hua Xia bian yuan: lishi jiyi yu zuqun rentong 華夏邊緣：歷史言己憶與族群認同 (Taipei: Yunchen wenhua, 1997)Google Scholar.
45. The identification of Xiongnu with these other names has been a subject of much controversy which cannot be gone into here. Nevertheless, in my opinion there can be no reasonable doubt that it is correct. Unfortunately the article on the Xiongnu which I prepared for the projected Volume 3 of Philologiae Turcicae Fundamenta has never appeared.
46. Sanguo zhi (Beijing: Zhonghua, 1959)Google Scholar, Weizhi 魏志, 30.854.
47. Tsunoda, Ryusaku and Goodrich, L. Carrington, Japan in the Chinese Dynastic Histories: Later Hart through Ming Dynasties (South Pasadena: P.D. and lone Perkins, 1951), 5n.11 and 17n.8Google Scholar.
48. Guoyu, 1.22 Google Scholar; Shiji, 4.144 Google Scholar.
49. Chunqiu and Zuozhuan, Xi 33/3.
50. Let me address the objection made by reviewers of the earlier version of this article to the use of the term “Chinese” as equivalent to Hua or Hua-Xia in Chunqiu times. It is true that, as I have stressed in the past, there is every reason to think that the territory we now refer to as “China,” even if restricted to what used to be called “China proper” (excluding the Northeastern Provinces, Mongolia, Xinjiang, and Tibet) was multi-ethnic in those days, as it is today. Nevertheless, as far as we can tell, Chinese was the only written language and the shared vehicle of elite culture throughout at this formative period. It was a magnet that drew other ethnic and linguistic groups to identify themselves with the Hua-Xia. It seems to me no more objectionable to refer to the culture that called itself “Hua” or “Hua-Xia” as “Chinese” than to use the term “Greek” to refer to those who used the Greek language from Linear B in Crete in the second millennium onward. Through the ages the Chinese script and the Chinese language which it represents have been the most distinctive hallmark of Chinese culture as well as a powerful unifying force politically. We need to be aware that the originally separate ethnic and linguistic groups that were absorbed into the Chinese amalgam probably contributed important elements to it, but that should not prevent us from recognizing the centripetal force that drew all these elements together.
51. Jia was a small state whose rulers had the surname Ji. See Xueqin, Li 李學勤, Eastern Zhou and Qin Civilizations (New Haven and London: Yale University Press, 1985), 78 Google Scholar.
52. Also written 翟,though this is much less frequent. The words represented by the two graphs were homophonous in Middle Chinese, EMC dεjk, but came from different Old Chinese rhyme groups. 狄 can be reconstructed as *lákj, or possibly *ljákj, but 翟 would have been *ljákw. The two rhyme groups were still quite distinct throughout the Western and Eastern Han dynasties; see Changpei, Luo 羅常培 and Zumo, Zhou 周祖護, Han Wei Nan Bei chao yunbu yanbian yanjiu 漢魏南北朝韻部演變研究 (Beijing: Kexue, 1958), 224-25 and 232–33Google Scholar. This is a linguistic puzzle that I do not know how to solve at present.
53. Guliangzhuan (Shisanjing zhushu 十三經注疏 ed. [Beijing: Zhonghua, 1957]), 392 Google Scholar; see also Shiji, 43.1797 Google Scholar (“Suoyin” 索隱 commentary).
54. See n. 59 below.
55. Guowei, Wang, “Guifang, Kunyi, Xianyun kao” 鬼方昆夷搬犹考, Guantang jilin 觀堂集林 (Beijing: Zhonghua, 1959), 583–606 Google Scholar. Shaughnessy, , “Western Zhou History,” 320–22Google Scholar.
56. Pulleyblank, , “The Chinese and their Neighbors,” 448 Google Scholar.
57. di Cosmo, Nicola, “The Northern Frontier in Preimperial China,” in Cambridge History of Ancient China, 960 Google Scholar. According to Di Cosmo, “This term, whatever its origin, soon came to indicate an ‘anthropological type’ rather than a specific group or tribe, which the records allow us to identify as early steppe nomads. The Hu were the source of the introduction of cavalry in China” (951-52). I think one can be more specific. In Han times the term Hu was applied to steppe nomads in general but especially to the Xiongnu who had become the dominant power in the steppe. Earlier it had referred to a specific proto-Mongolian people, now differentiated as the Eastern Hu 東胡, from whom the Xianbei 鮮卑 and the Wuhuan 烏桓 later emerged. Di Cosmo makes the useful point that the famous adoption of “Hu clothing,” that is, riding gear, by the ruler of Zhao was not primarily designed to defend against nomad attacks but “to gain an advantage against other Chinese states” (960).
The Xiongu had, I believe, a quite different ethnic origin. There is good reason to believe that before they were driven north out of the Ordos by the Qin general Meng Tian 蒙׳ in 215 B.C.E., the Xiongnu had been part of a so-called Rong people, called Yiqu 義渠, that had dwelt in the region now known as Shaanbei 陕北 and had been under Chinese influence for centuries. See Pulleyblank, Edwin G., “Ji Hu 稽胡： Indige-nous Inhabitants of Shaanbei and Western Shanxi,” in Opuscula Altaica: Essays Presented in Honor of Henry Schwarz, ed. Kaplan, Edward H. (Bellingham: Western Washington University Press, 1994), 499–531 Google Scholar.
58. Although the Chinese system of correlations between colors and compass directions is part of the theory of the Five Phases (wu xing 五行)that only found explicit development in late Warring States times (see Nivison, David S., “The Classical Philo-sophical Writings,” in the Cambridge History of Ancient China, 809)Google Scholar, it has roots that go back long before. Compare, for example, the description in Guoyu, 19.608 Google Scholar, of the Wu army drawn up to confront the Jin army at Huangchi 黃池 ‘Yellow Pond’ in Henan in 482 B.C.E. The central army on the west side led by the king and facing east had white uniforms, while the armies on the north and south flanks had black and red uniforms respectively.
59. Guoyu, 2.45–46 Google Scholar. Compare Zuozhuan, Xi 24/2.
60. Zuozhuan, Xi 18/2 and 20/5.
61. Zuozhuan, Xi 23/fu 2.
62. Mengzh 4B/1
63. Karlgren, Bernhard, The Book of Documents (reprinted from the Bulletin of the Museum of Far Eastern Antiquities 22 [1950]), 28–29 Google Scholar.
64. Stein, Rolf A., Tibetan Civilization (Stanford: Stanford University Press, 1972), 94–97 Google Scholar.
65. One indication that Shang writing as we know it must have had a prehistory is the relation between the graphs for lai 來 “come” and mai 麥 “wheat.” The simpler graph for “come” is clearly based on a drawing of a plant while that for “wheat” has the addition of a graph for “foot.” The obvious inference is that the simpler graph was originally created to write the word for “wheat” and was then borrowed, with the addition of the “foot” element, for the verb “come” because of similarity in pronunciation. But lai “come” was much more frequent than mai “wheat” so in the course of time the simpler graph was used for the verb and the more complex graph was used for the noun even though the “foot” signifie was inappropriate. When, exceptionally, the graph 來 appears in the Shijing in a context that implies that it is a kind of grain, traditional commentary reads it as lai and interprets it as “a kind of wheat” (mai 麥) instead of recognizing it as the original graph for mat “wheat”; see Pulleyblank, Edwin G., “Early Contacts between Indo-Europeans and Chinese,” International Review of Chinese Linguistics 1 (1996), 14 Google Scholar.
66. Bagley, Robert, “Shang Archaeology,” in Cambridge History of Ancient China, 141 Google Scholar.
67. Bagley, , “Shang Archaeology,” 141 Google Scholar.
68. Early China 16 (1991), 39–80 CrossRefGoogle Scholar.
69. Pulleyblank, , “The Consonantal System of Old Chinese,” Asia Major 9 (1962), 58—144, 206–65Google Scholar; “Some New Hypotheses Concerning Word Families in Chinese,” Journal of Chinese Linguistics 1 (1973), 111–25Google Scholar.
70. Baxter, William H., A Handbook of Old Chinese Phonology (Berlin and New York: Mouton de Gruyter, 1992), 767, 782, 800 CrossRefGoogle Scholar.
71. Baxter, , A Handbook of Old Chinese Phonology, 765 Google Scholar.
72. Pulleyblank, , “The Final Consonants of Old Chinese,” Monumenta Serica 33 (1977/1978), 180–206 CrossRefGoogle Scholar.
73. Pulleyblank, , “Close/open Ablaut in Sino-Tibetan,” in Indo-Pacific Linguistic Studies (Lingua 14 [1965]), ed. Milner, G. B. and Henderson, E. J. A., 230–40Google Scholar.
74. See Pulleyblank, E. G., “Ablaut and Initial Voicing in Old Chinese Morphology: *a as an Infix and Prefix,” Proceedings of the Second International Conference on Sinology. Section on Linguistics and Paleography (Taipei: Academia Sinica, 1989), 1–21 Google Scholar. Note that in the case of both gu 古 and xu 虛, my reconstruction implies that labiovelar initials lost their labialization before *a. As shown in the above paper, this is a general rule corresponding to the loss in Middle Chinese of labialization of final *kw after *a in the Old Chinese *-akw rhyme, as in yao 藥 EMC jiak < *-àkw, le 樂 EMC lak < *-ákw.
75. Kuipers, A. H., Phoneme and Morpheme in Kabardian (s'Gravenhage: Mouton, 1960)Google Scholar.
76. Pulleyblank, Edwin G., “The Locative Particles yü 于, yü 於, and hu 乎,” Journal of the American Oriental Society 106 (1986), 1–12 CrossRefGoogle Scholar.
77. Pulleyblank, Edwin G., “Morphology in Old Chinese,” 26–51 Google Scholar.
