54. See discussion of this in Joyce Marcus, “The Peaks and Valleys of Archaic States: An Extension of the Dynamic Model,” in Archaic States, ed. Gary M. Feinman and Joyce Marcus (Santa Fe, NM: School of American Research Press, 1998), 59–94; and also in Francis Allard, “Introduction: Power, Monumentality, and Mobility,” in Social Complexity in Prehistoric Eurasia, ed. Bryan K. Hanks and Katheryn M. Linduff (Cambridge and New York: Cambridge University Press, 2009), 328, where Allard points out that “periods characterized by demographic aggregation—as evidenced, for example by public works or military campaigns—are invariably followed by periods that appear to be more decentralized.” Below we will give more attention to definitions of “state.”
55. WS 2.34, 108A.2734. The fabrication of monarchical lineage took place in many states in northeast Asia around this time. See Edward J. Kidder’s discussion of questionable genealogy in the rulers of the archipelago’s emerging state in his Himiko and Japan’s Elusive Chiefdom of Yamatai: Archaeology, History, and Mythology (Honolulu: University of Hawai’i Press, 2007); and on the other end of Eurasia, similar issues for the Goths discussed by Peter Heather, Goths and Romans, 332–489 (Oxford: Clarendon Press, 1991), 19.
56. WS 1.5. The name “Fu” might have been assigned because the fellow’s original, Altaic name was forgotten, but Tamura, Chūgoku shijō no minzoku idōki, 197, expresses skepticism regarding the supposed “successors” of Liwei. It is possible that these have been inserted to construct a link between Liwei and the Yulü line, which we will discuss below.
57. WS 1.5–6; ZZTJ 82.2614. For the idea regarding borrowing of the practice, see Du, Bei Wei shi, 60.
58. HYDCD 1: 646 gives the pronunciation of this character as “yi,” a position supported by Victor Xiong (in his review of “Dawn of a Golden Age,” American Journal of Archaeology Online Publications, January 2005, accessed 13 August 2019). The ZZTJ commentary (85.2701), however, gives “tuo,” which we shall use here.
59. For more on location of this site, see note 67.
60. WS 1.6–7, 23.599. For the guess at a number, see how many together fled back south twenty years later: WS 1.9. For the development of such groups in the midst of the empire’s collapse, see Jin Fagen 金发根, Yongjia luan hou bei fang de hao zu 永嘉乱后北方的豪族 (Taibei: Zhongguo xue shu zhu zuo jiang zhu wei yuan hui, 1964). For several generations, refugees had been fleeing the empire’s slow collapse, some seeking protection among various groups on the northern frontier: SGZ 26.731.
61. WS 1.6.
62. WS 1.6; ZZTJ 85.2701.
63. WS 1.6–7.
64. Zhang Jingming 张景明, “Nei Menggu Liangchengxian Xiaobazitan jin yin qi jiao cang” 内 蒙古凉城县小坝子滩金银器窖藏, WW (2002.8): 50–52, 69; Su Bai 宿白, “Shengle, Pingcheng yi dai de Tuoba Xianbei—Bei Wei yi ji—Xianbei yi ji ji lu zhi er” 盛乐、平城一带的拓跋鲜卑——北魏遗迹——鲜卑遗迹辑录之二 WW (1977.11): 39–40; Adam Kessler, Empires beyond the Great Wall, 79–81. James Watt et al., eds., China: Dawn of a Golden Age, 200–750 ad (New Haven, CT: Yale University Press, 2004), 8–9, 127–28 no. 32, point out this plaque shows Xiongnu influence in depiction of the animals (9). Of course, there is always a possibility that this is a more recent forgery.
65. See the works cited in note 64; and mentions of such seals and their functions among the Southern Xiongnu in Miller, “Southern Xiongnu,” 160–62. For a general overview, see Zhang Jingming 张景明, Zhongguo bei fang cao yuan gu dai jin yin qi 中国北方草原古代金银器 (Beijing: Wen wu chu ban she, 2005), 78–96.
66. Miller, “Southern Xiongnu,” 162, calls the seal “a tool for structured mediation.”
67. Some have suggested that the location of the cache was Yituo’s home base, which we are told was the Shenhe Slope 参合陂; this is a controversy that will be discussed in greater detail in Chapter 7 note 73. Here this author will simply state preference for locating the Shenhe Slope northeast of Datong, on the edge of the open steppe. Confirming this in this particular case is the WS (1.5) statement that Yituo “dwelt north of Dai commandery’s Shenhe slope” 居代郡之参合陂北. The original Dai commandery of the Chinese empire was located to the east of the upper Sanggan valley, in the area of mod. Yuxian, Hebei (and see also ZZTJ 82.2614). An interesting legend tells of how on the Shenhe Slope Yituo had spat out a bug, and in that place grew elms where none had been before: WS 1.7. Such woodlands had long been seen as sacred by frontier populations: near the town of Mayi (“Horse Town”) at the bottom of the Datong Basin had been a grove sacred to the Xiongnu: SJ 110.2892. Similar practices can be seen among the Taghbach: Han Xiang 韩香, “Shi lun zao qi Xianbei zu de yuan shi sa man chong bai” 试论早期鲜卑族的原始萨满崇拜, Heilongjiang min zu cong kan (1995.1): 602; and Holmgren, Annals of Tai, 80 note 8 and 84 note 38. They also had a practical purpose: Shao Zhengkun 邵正坤, “Shi lun Xianbei zao qi de zong jiao xin yang ji qi zhuan bian” 试论鲜卑早期的宗教信仰及其转变, in Zou chu shi ku de Bei Wei wang chao, ed. Jin Zhao and Alede’ertu (Beijing: Wen hua yi shu chu ban shu, 2010), 635, points out that in addition to being sacred, the elm was important to Inner Asians, used to make such things as saddles and arrows.
68. WS 23.599–602. It needs to be noted that the stele itself has not been recovered, though a transcription of this 306 stele is given in the sixth-century Wei shu. If an accurate description of what was on the stele, this contains the first claim that the Taghbach had descended from the Yellow Emperor (599).
69. On Liu Kun, see Fan Zhaofei 范兆飞, “Yongjia luan hou de Bingzhou ju shi—yi Liu Kun ci Bing wei zhong xin” 永嘉乱后的并州局势—以刘琨刺并为中心, Xue shu yue kan (2008.3): 122–30. The reader should note that the term “Xiongnu” is used loosely here, to refer to a variety of groups of disparate ethnicity connected politically with the Southern Xiongnu in regions within the modern Shanxi and Shaanxi provinces.
70. WS 1.7; ZZTJ 87.2752–53. It seems, in fact, that this was a fait accompli, since the history of the Jiankang-based Song regime states that Yilu had begun moving his people into these districts the year before. Liu Kun’s relocation was sped up in 311 by an attack by Yilu on the Bingzhou seat, Taiyuan: SoS 95.1321; JS 5.123; Uchida, Kita Ajia shi kenkyū, 2: 102–3.
71. WS 1.9.
72. For one example of Yilu’s recruitment, see the account of Mo Han in WS 23.603.
73. For general discussion of such issues, see Anthony P. Cohen, Symbolic Construction of Community (London and New York: Tavistock Publications, 1985); and Helms, Craft and the Kingly Ideal.
74. For the nature of a later East Asian hybrid state, with which interesting comparison can be made, see the articles in James Millward et al., eds., New Qing Imperial History: The Making of Inner Asian Empire at Qing Chengde (London and New York: Routledge, 2004).
75. Honeychurch, Inner Asia and the Spatial Politics of Empire, 58–59; and more generally, H·m·l·inen’s Comanche Empire.
76. Richard Tapper, The Conflict of Tribe and State in Iran and Afghanistan (London: Croom Helm; New York: St. Martin’s Press, 1983), 67. At the time the author writes this, Americans are, of course, confronting this reality head on.
77. Robert L. Carneiro, “The Chiefdom: Precursor to the State,” in The Transition to Statehood in the New World, ed. Grant D. Jones and Robert R. Kautz (Cambridge and New York: Cambridge University Press, 1981), 69. And see Di Cosmo, “China-Steppe Relations,” 66, who to the right to tax, draft, and issue laws adds “the authority to represent the whole political community in diplomatic and international relations.”
78. WS 111.2873, 1.9.
79. Di Cosmo, “China-Steppe Relations,” 65, points out centrality of army in the formation of Inner Asian (perhaps all·) states: “There is general agreement on the fact that steppe imperial institutions were, at least initially, all military in nature, since the steppe empire can only be formed by military means. The new armies, which are no longer bands of tribesmen but complex military machines of tens of thousands of people, became the laboratory in which new institutions were born.” See more in Chapter 11, “The Wei Army.”
80. This Wei shu story of struggle between father and son does remind one of how the Xiongnu ruler Modun was treated by his father, and many other such tales; see Di Cosmo, Ancient China, 175–76. Of course, one could similarly try to deny extramarital affairs as mythical construction precisely because of how frequently they appear.
81. WS 1.9; BS 15.545 (WS 14.384); ZZTJ 89.2830.
82. See Chapter 4 note 13; and Huang Lie, “Tuoba Xianbei zao qi guo jia,” 79. Yan, Bei Wei qian qi zheng zhi zhi du, 19, points out that this was a borrowing from the early Sixteen Kingdoms of the fourth century, and perhaps long before that. According to Michael Frachetti, “Differentiated Landscape and Non-uniform Complexity among Bronze Age Societies of the Eurasian Steppe,” in Social Complexity in Prehistoric Eurasia, ed. Bryan K. Hanks and Katheryn M. Linduff (Cambridge and New York: Cambridge University Press, 2009), 19: “Even in local settings, evidence suggests that Bronze Age steppe communities were organizationally heterogeneous—meaning they were not politically or economically centralized under a shared corpus of functional institutions.” For the officers of Northern Wei’s Northern and Southern Units 南北部, see Yan Gengwang 严耕望, “Bei Wei shang shu zhi du” 北魏尚书制度, Li shi yu yan yan jiu suo ji kan 18 (1948), rpt. in Yan Gengwang shi xue lun wen xuan ji (Taibei: Lian jing chu ban shi ye gong si, 1991), 116–22.
83. WS 1.9, 23.603; JS 5.130, 62.1684; ZZTJ 89.2830. See also Tang, “Tuoba guo jia de jian li ji qi feng jian hua,” 199; Zhang, Bei Wei zheng zhi shi, 1: 131–34 (and on p. 135 examining a different figure—300—given in the “Prefatory Annals,” WS 1.9). “Wuhuan” here has lost any particular ethnic meaning, and refers to “various peoples from all over who have come to submit” 诸方杂人来附者 (WS 113.2971).
84. WS 23.603. Holmgren, Annals of Tai, 38, suggests that what this group feared was Pugen and the Yituo line, which wished to keep the polity’s base on the grasslands.
85. JS 62.1685, 1687; WS 23.603; ZZTJ 89.2837–38, 90.2858–59.
86. WS 1.9. For the brief account of Yituo’s wife: BS 13.491 (WS 13.322–23). Her name (as transcribed into Chinese) is given as Wei 惟 in Bei shi; as Qi 祁 in Wei shu. Since this chapter in Wei shu was lost and restored from Bei shi, which had previously been based largely on the original Wei shu chapter, I take the latter’s version of the name—“Wei”—as more reliable. For discussion of variants in her name, see Zhang, Bei Wei zheng zhi shi, 1: 148ff. For discussion of her origins, see Tian, Tuoba shi tan, 27.
87. It is here in the Taghbach lineage that we see particular possibility of manipulation, and the insertion of other interlopers into the interloper’s line. Even if this were true, however, it indicates that Liwei’s clan had become worth claiming membership in. Perhaps with the Serbi principle of noble dynasty as a general model, false claims of genealogy were pervasive during this period. For a later example, see Holmgren, “Lineage Falsification in the Northern Dynasties.”
88. This was done in the time of Xiaowen, when the title Taizu was assigned to Daowu. See discussion in WS 84.1852–53, 108A.2745–46, 2.47 note 14.
89. WS 1.9.
90. WS 1.10; ZZTJ 91.2891; BS 13.491 (WS 13.323). It will be noted that those calling the Taghbach a “woman’s country” would have been members of Shi Le’s court, sneering at the emissaries sent them to establish peace. Here also is a good example of the care we need using description of the doings of one people in the language of another: for though a classical Chinese term is used to describe Madam Wei’s participation in the regime’s decision-making, there certainly was no Chinese-style “court” (chao 朝) that she oversaw (lin 临). And of course, neither was there a “throne,” in the European style at least.
91. HHS 90.2979.
92. See Anne F. Broadbridge, Women and the Making of the Mongol Empire (Cambridge: Cambridge University Press, 2018): Chapters 6 and 7.
93. WS 1.10. In light of this removal of the center of power to the north, Zhang, Bei Wei zheng zhi shi, 1: 142, suggests that the main base of Madam Wei’s power lay with those who had resisted Yulü’s ambition to seize Chinese territory. Holmgren, Annals of Tai, 39; Holmgren, “Women and Political Power,” concurs, saying that opposition to Yulü—including his hostility to Later Zhao—led to support for her leadership. Her son, however, reversed this policy, as we see just below.
94. JS 108.2808; ZZTJ 93.2933; Holmgren, Annals of Tai, 39.
95. For discussion of the Helan, see Zhang, Cong Tuoba dao Bei Wei, Chapter 1; and Tian, Tuoba shi tan, 63–65.
96. WS 1.11.
97. WS 1.11.
98. In classical Chinese in general, and in Chinese accounts of non-Chinese peoples, guo has multiple meanings: “dynasty,” “nation,” “state.” In this book, as will be discussed in Chapter 9, it really refers to all three: the administrative apparatus and pre-modern “nation” that took shape around the Taghbach dynasty.
6
立国
什翼犍（338-376年在位）身材魁伟，史载身高逾六尺半。现代学者卜弼德推测其名——中古汉语音译约作*Dzyip-yik-ken——在其母语中或意为"赤色"。1身处与先祖迥异的时代，其形象亦呈现新貌。拓跋部早期因传统发辫得名"索头"，而什翼犍"发垂委地"，暗示其蓄长发且或为披散状。2此时拓跋贵族已开启剧变之路，此种转变将在未来一两百年间持续演进。
通过终结此前数十年的动荡，什翼犍以三十八年统治期为拓跋部带来全新稳定局面，为北魏的建立铺平道路。但这位十九岁即位的君主，其掌权过程似乎充满波折。在早期拓跋史中，这类被《魏书》记载所隐晦的"不堪之事"，构成了若干值得注意的历史片段。
什翼犍与翳槐皆为郁律之子，据称通过汉文史料中称为"弗"者（参见卷首王朝世系图）传承自沙漠汗血脉。然二人实为异母兄弟：翳槐生于贺兰部首领之妹，什翼犍则生于改汉姓为王氏的乌桓家族女性。3《魏书·本纪》对继承过程的记载兼具简化与曲解：先仅言翳槐第二次在位"经年而崩"——未作任何解释——继而敷衍临终托孤于异母弟的情节；复述其弟拓跋孤亲赴邺城迎归什翼犍之事。4值得注意的是，纪载明示什翼犍即位之地非盛乐亦非草原，而是后赵新占桑干河南之繁畤。
在汉文王朝史籍中，本纪往往承载当权者所需之虚言，而更近真相的线索则偶现于列传。5什翼犍生母王氏传记透露端倪："及崩，帝业垂危，赖皇后之力得兴复大业"。6在拓跋孤残缺传记（可能遭故意删弃，现存为十二世纪辑补本）中可见：诸部大人杀害郁律另一子以阻什翼犍继位，并强推拓跋孤上位。这位名讳可疑的单字人物传记继而塑造其大义形象——坚称必须迎归什翼犍，后者北返后"分国以半授之"。紧随其后仅以一字收场："薨"。7未明示死亡速度；《官氏志》则载其掌北疆军务。8现代学者霍姆格伦（Jennifer Holmgren）提出合理推测：什翼犍实得后赵支持（或由拥立军队在繁畤即位），但"将北魏开国者祖父描述为匈奴傀儡篡位者显然不可接受"9，故史书塑造其"宽仁弘度，喜怒不形于色"的典型成功政客形象。10至于王氏匡扶社稷之具体手段，或系其促成什翼犍与拓跋孤共治协议——该局面终以什翼犍弑杀"忠义"兄长告终。更诡谲处见于拓跋孤子传记：载其"失职"（应指继承权丧失），约四十年后怂恿什翼犍子弑父。11这般政治漩涡令汉人史官如履薄冰，《国记》直书或为崔浩死因（虽存其他可能）。
什翼犍在位的三十八年期间，综合了两位叔祖（若史载亲属关系属实）的事业：征伐者猗·与建制者猗卢。就建制而言，他于338年采用汉式年号"建国"后，确实着手构建汉文书写者所理解的"国"之雏形。12但其官僚体系远未达中原帝国规模，即使借用晋朝官称，组织架构亦大相径庭。339年设立的侍从官（汉文称"近侍"）制度尤具关键意义：这些近臣负责侍奉护卫君主并传达诏令，标志着"内朝"建构的开端——下文将详述此制度对拓跋王权发展的重要性。什翼犍国家建设的另一重要举措是重建猗卢南北二部。其首领汉文称"大人"，管辖对象非全体臣民，而专司王畿内未附属的"杂胡"部落。13这些直属部众与贺兰部等政治独立的附庸群体仍保持区隔。执掌北部的拓跋孤管辖牧民（南部大人则治农耕之民）。
什翼犍延续猗卢立法事业，但摒弃初版的严酷刑律。史未明载法典是否以汉文书写及如何向非汉人群传播，推测最初或为口头法令，类似蒙古的大扎撒。据《魏书》所述，随着国家与王朝定义渐明，法典首要目标转为维护统治：犯"大逆"者诛灭全族，无论男女老幼。14法典亦规范两性道德，对"不以礼交"男女处死。此外律条相对宽缓——至少对富者如此（此乃多数社会通例）。其立法基础仍属以牲畜为主要财富的社会形态，什翼犍时期的王室及此后相当时间内皆如此。15据此法，精英阶层（除大逆外）死罪可用"金马"赎免；平民杀人罪可赔付棺木与四十九匹马了结；盗窃罪中私产保护优先于公产：盗取"官物"（主要指王室畜群）按五倍赔偿，私畜则需十倍偿还。虽较卷帙浩繁的汉律简陋，但此处确见卡内罗(Carneiro)"国家"定义中法制维度的实质性发展。
