109. The quote taken from Emily Wilson’s Times Literary Supplement, April 27, 2016, review of Denis Feeney’s Beyond Greek (Cambridge, MA: Harvard University Press, 2016).
110. WS 3.64.
111. Shimunek, Languages of Ancient Southern Mongolia and North China, 81, calls the Chinese spoken at Pingcheng a “northeastern dialect of Early Middle Chinese.” The multilingual nature of the Taghbach emperors from early times (it is hard to say how early) is another reality obscured by Wei shu; 1,000 years later, under the Manchu regime, it was made clear in the famous trilingual stelae.
10
令人不安的创新
武皇帝涉珪的伟大征服在其二十多岁后半段便告终结。从那时起直到38岁去世，他日益将目光转向内心自我——这对身外之人造成了灾难性后果。有一两年时间，他致力于建造朝堂、设立制度、制定礼仪。但在道武帝统治后期，这些建设让位于与真实或假想敌人的斗争，最终导致真实存在（有时可能也属虚构）的叛乱。史载他建造了一座新都的外壳，但其街道仍大多荒无人烟，居民日益目无法纪。1人们不禁要问：这些新居民——无论是征服者还是被强制迁徙的被征服者——究竟在寻求何种慰藉？
据史料记载，这位皇帝正在转向一种名为"寒食散"的古代版伟哥（Viagra）。2这种兴奋剂由五种矿物质（包括氟石、钟乳石和硫磺）及动植物成分构成，其可靠性值得怀疑。3史载道武帝的服用剂量与日俱增。无论这是否是唯一原因，《魏书》记载他变得越来越不稳定：连续数日不食、彻夜不眠，反复思虑过往成败，畏惧凶兆，并自言自语犹如与鬼神交谈。无论其精神状态的精确性质如何，这位北魏皇帝与朝堂日益疏离，对身边之人愈发猜忌。
道武帝晚年多有杀戮。4 408年，一位莫那娄部首领迎来末日。此人多年效忠道武帝，但数十年前——在386年的关键事件中——曾支持皇位争夺者（即涉珪的叔父）。据记载，当时这位莫那娄人递给涉珪一支箭并说道："三岁犊牛岂能胜重载之车？"尽管少年因被称作"犊牛"而怀恨在心，但在赢得继承权后仍需此人支持，故此后二十年仍任用其为禁卫成员。然而至408年，道武帝在反复思虑过往成败后，遣使至莫那娄人宅邸赠箭并发问："三岁犊牛——今可胜重载车否？"次日清晨此人即遭处决。5除报复旧怨外，道武帝还得以没收其全部财产（此时仍以牲畜为主），颇感快意。6
更具深远影响的是409年处决卫王的命令，据称是对其谋叛计划的惩罚。7以现存有限史料，难辨指控真伪。但基于前文所述原因，道武帝确有理由担忧继承权再起争端。尽管五年前他已至少非正式地指定长子（即未来明元帝）为继承人——封其为齐王，更具深意地任命为相国——但卫王乃皇室近亲且权势显赫。8 408年某夜，卫王曾被道武帝召入宫中见证继承人（即未来太武帝）诞生。皇帝问道："闻中夜召命，可曾惊惧？"据载卫王答曰："臣以诚奉陛下。陛下明察，臣今心安。惊或有之，实无所惧。"9然卫王确有恐惧理由。次年即被赐自尽，自然也未得葬入金陵墓葬群。
此举仅加剧朝中积年累月之恐慌浪潮。被道武帝亲手诛杀官员的尸体——罪名据称包括"步伐错乱"或"奏对失次"——陈列于天安殿外。10赐死卫王后，道武帝设宴邀诸王以缓局势。唯有一王赴宴。余者皆惧受卫王牵连逃亡，部分谋划北投柔然。赴宴亲王获厚待，稍安宗室。11然继承权之争不仅限于皇帝近亲。自诘汾时代起，拓跋家族便是个不断扩张的锥形宗族。诘汾后裔"直勤"们作为祖业共有者，部分人似乎认为所有特勤至少在理论上皆可继承诘汾之位。12北魏君主对此说之否定自不待言，然百年间仍见疏远宗室零星叛乱。13此问题直至孝文帝时代方得最终解决。而404年，为制衡宗亲威胁，道武帝始封非皇室成员"王"爵。14各类身份结构无疑早已存在于过去百年间聚合形成代国的各族群中。但通过创设外部强加的垂直爵位体系（与旧贵族并行），这些封爵——虽非如尔朱部般实授封地的特殊类型——改变了国家形态。借自华夏帝国的爵称体系与军衔结合，以史料未明载之方式，赋予受封者世袭荣誉及在朝堂军中的特权通道。15
道武帝通过任用外戚巩固君权以对抗宗亲的做法，表明他的疯狂中暗含章法。但这颗头脑无疑已病入膏肓。如同所有人类基于自私基因深处的本能，道武帝渴望将基业传予亲生骨肉——而非堂兄弟，更非妻族中人。但显然对儿子能力存疑，我们在其统治末期看到一项尤为令人不安的做法："子贵母死"——正式册立储君后即迫使其生母自尽。在这个一夫多妻制社会，正式皇后与储君生母分属不同个体。需特别指出，所有因此被杀的妇女均非鲜卑人。16通过处决储君生母（并切断皇后与储君的亲密纽带），道武帝试图防止后党及外戚专权（如拓跋历史上卫夫人事件，以及道武帝生母干政之先例）。17需说明的是，后世这项制度反被宫廷权妇用以清除储君生母、掌控继承人并巩固自身权势。18尽管被称为"古制"，这种瓦解家族纽带的做法在早期拓跋史中并无先例，似为道武帝首创；读者可自行判断崔宏为其诵读《汉书》中汉武帝类似故事是否可能成为其灵感来源。19
被立为道武皇后的慕容宝幼女，系中山陷落时与数百族人同被掳获。他们被押解北迁平城，据载慕容氏曾受皇帝"临幸"，但未留下子嗣记载。她早逝（或属自然死亡），具体年份不详，但必在409年前——该年道武帝临终前数月以"谋叛逃亡"罪名诛杀其全部平城族人。20
为道武帝诞下长子（即未来明元帝，409-423年在位）的刘夫人，系匈奴首领刘库仁侄女（亦即道武帝死敌刘显表亲）。道武帝于386年纳其为妃，392年在土默特平原诞下明元帝。但她未能晋位皇后，据称因"铸金人不成"（这项源自内亚的占卜习俗未能流传至本朝之后）。21 409年——虽称深得帝宠——仍被迫自尽，显因皇帝自觉时日无多，不愿儿子受制强妇。时年17岁的储君精神崩溃逃出宫闱。22道武帝遂转向次子拓跋绍，其生母贺兰氏恰为皇帝母族姐妹。（据载少年道武帝初遇贺兰氏即为其美貌倾倒。询问母后时得答："不可——此女过艳非汝良配。况已有夫。"23未几其夫暴毙，姨母遂被纳为妻室。
道武帝虽曾痴恋此女，此刻却欲对绍母实施"古制"。然未及行事，囚禁中的母亲密传其子："汝将何以救我？"当夜，道武帝宿于天安殿时，绍——被《魏书》描述为暴戾少年——率亲随数人并勾结宦官破门而入。警报骤响，拓跋君主跃床而起，摸索剑弓。因不得兵器，遂遭弑杀。24
次日宫门紧闭至正午方开，绍现身问群臣："孤有叔父，亦有长兄——公卿欲奉何人？"25众皆面如死灰，无人应答，唯老臣"拔拔部的嵩"26应曰："唯从殿下。"虽疑云密布，绍仍开启府库广赐布帛（即代国通货）以收买人心。27
响应绍者有其母族贺兰部："旧部余众率子弟召集亲族，四方云集。"28然贺兰终将再尝败绩。明元帝闻父丧匿身盆地环山处静观其变。其联络朝臣之举获热烈响应。当及至逼近都城，禁军擒绍献于明元帝，后者遂迫使绍母子自尽，并处决党羽十余人。据载，曾冒犯道武帝者被带至宫城外广场，遭"群臣"生啖其肉。29虽长子继承制仍存争议，明元帝显然获得广泛支持。其母死后逃离宫廷期间，得护卫王洛儿庇护（此人曾于冰河救主）。王护其隐匿期间联络朝臣，最终助明元帝复位。30
17岁新君以此戏剧性方式登基。虽南朝某史书以"木末（中古音*Muwk-mat）"记其名，但华夏史学传统中其将以明元帝（409-423年在位）之名载入史册。31动人之处在于，新帝本纪首录事迹即为追封被父逼死的生母为皇后。耗时逾年，方将弑妻者葬于盛乐城东金陵墓葬群。32此其间停灵何处，史无明载。
魏收在编纂于六世纪的《魏书》明元帝本纪结尾处写道：道武帝统治末期"朝内多怨"，其子虽遭"遇妖之祸"（此"妖"或指弑父之子，亦可能暗喻其父），终成"内外辑睦"。33实现此局面的举措包括重新起用被道武帝罢黜的重臣，特别是通过授予卫王子嗣王爵恢复该支系地位。34此朝还出现元老重臣权力凌驾君主的"委员会式摄政"现象。以拔拔嵩及崔宏为核心的决策集团，每日坐镇宫城南门止车门外处理政务。35
然帝国与皇室皆未达真正和谐。皇位争夺仍在持续：明元帝即位数月后，某堂兄弟怀揣利刃潜入深宫图谋行刺被捕。36至其统治末期，《魏书》数次隐晦提及皇帝兄弟之死，显系朝廷通过清除手足消除潜在威胁。37不过在这些案例中，爵位并未废除而是由死者子嗣继承。
明元帝在位的十四年间，北魏延续道武末年开启的军事休整期。与满洲地区残存燕政权偶有非决定性接触（该地控制权已落入表面汉化、实则深受鲜卑文化影响的冯氏家族，后文将详述）。更严峻威胁来自柔然草原势力，促使明元帝末年于阴山边境修筑长城，此为阴山以北建立连串军事据点之开端。38
北魏在416年于洛阳附近黄河北岸展示军威，时值南方将领刘裕——他将于420年终结东晋并建立刘宋（420-479年）39——溯流而上从羌族后秦政权手中夺取洛阳（该城自百年前遭劫掠后仅存废墟）及短暂控制长安与关中地区。40（此役期间收集的洛水沿线要塞资料，后收录于北魏地理著作《水经注》。刘裕422年死后不久，明元帝先遣部队后亲率主力南下夺取黄河南岸战略要地，包括洛阳故址及军事重镇虎牢（今河南荥阳西北）——此关隘控制洛阳以东山河夹峙的狭窄通道。41然皇帝随即北返，该区域此后多年沦为争议无人区。42但感知拓跋威胁，羌族首领、河西走廊北凉政权及青藏东部吐谷浑开始努力与建康政权结盟。43
在控制较稳固的疆域内，新帝确曾派遣巡查组调查道武帝时期遗留地方官员的政绩劣迹。此类举措自征服初期即有零星实施：398年中山刚陷落时，道武帝率军穿越平原途中曾作态询问新附民户疾苦，并命令当地官员抚恤贫老。44但系统掌控新占州郡的努力至明元朝方得加强。有巡查官奏报：原属宫廷作坊的裁缝被任命为太原县令后非法从事贸易。45据载该员受惩，州郡状况改善，但腐败与对征服地区控制不彻底的双重问题仍将持续。如第15章将详述，虽从这些地区征收赋税，但皆通过地方豪强中介完成。46
北魏皇室对京畿地区的控制更为稳固。道武帝于398年征服平原后开始营建平城，但直至统治末期仍属在建状态——史载其为大片空地散布零星新建筑群。虽偶尔居留（如最终驾崩于天安殿寝宫），但据建康政权史书《南齐书》记载，道武帝仍保持游牧习性，多数时间"逐水草而居"，真正定居平城宫室始于明元朝。47明元时期完成大部分城建工程，包括422年最终竣工的外城墙。48官员将士开始入住城坊，由强制迁居京畿的民众提供后勤服务。此类迁徙规模在明元帝之子太武帝时期达到顶峰，使京畿人口超百万。49
在东北亚气候整体转冷的背景下，定都高寒干旱地区自有弊端：415年当地即遭逢饥荒。50当迁都平原提议出现时，明元帝咨询崔浩。崔氏反对迁都的核心论据在于拓跋政权依赖神秘感与威慑力："今居北地，假令山东有变，轻骑南出，耀威桑梓（指中原村落）之中，谁知多少？百姓见之，望尘震服。"51宏观而言，平城犹如"山巅城堡"。52皇帝回应："此说深契朕心。"遂维持高海拔都城，仅将极端困苦者暂时南迁平原就食。这些饥民南迁路线，大抵与17年前被强制北徙时相同。短期危机虽解，根本问题持续恶化。
崔浩并非平城重用汉人谋臣的孤例。413年初春，朝廷遣使搜罗"贤俊隐逸"，选拔世家子弟及具决策领导才能者赴京待选。53明元帝礼遇这些人才，自身亦渐染汉学：即位首年即命崔浩讲解经义。如前所述，本纪末尾称其着有《新集》，因不满汉儒刘向著作而编。54不论实际编者何人，值得注意的是该书乃辑录经史章句的汇编，"帝王朝典、祭祀礼仪、军旅誓诰悉备"。55原书虽佚，此例彰显拓跋政权（亦属政权通性）关键特质：需平衡文化资本与武力威慑（后者将于第11章详析）。
1. WS 2.44; ZZTJ 115.3614.
2. WS 2.44; ZZTJ 115.3614.
3. Ute Englehardt, “Cold Food Powder,” in The Encyclopedia of Taoism, 2 vols. (London: Routledge, 2008), 1: 473; and Joseph Needham, Science and Civilization in China, Vol. 5. Chemistry and Chemical Technology: pt. 3. Spagyrical Discovery and Invention: Historical Survey, from Cinnabar Elixirs to Synthetic Insulin (Cambridge: Cambridge University Press, 1954–), 117–19, in which we are provided a passage from Wei shu (114.3049; Needham, 118) describing establishment of the post of an “Alchemist-Royal” who concocted medicines and elixirs, which were tested on individuals sentenced to death. “[S]ince it was not their original intention (to seek for immortality),” the translation tells us, “many died.” This is confirmed by Akahori Akira, in his “Drug Taking and Immortality,” in Taoist Meditation and Longevity Techniques, ed. Livia Kohn (Ann Arbor: Center for Chinese Studies, University of Michigan, 1989), 74, who describes cold-food powder as having been popularized by its association with the philosopher He Yan (d. 249), who used it “to gain relief from depression.” Akahori goes on to say that there was no standard formula, and that “[i]t frequently caused severe adverse reactions and sometimes even led to death.”
4. See the table of those killed or dismissed by Daowu in Li, Bei Wei Pingcheng shi dai, 74–77.
5. WS 28.683–84. The term here translated as “juvenile bull,” du 犊, is the Chinese term that translated whatever the Monalou and Shegui actually said in their own guo yu; it would be fascinating to recover the original, taken as it is from the core of their traditional economy.
6. Zhang, Bei Wei zheng zhi shi, 2: 330.
7. WS 15.372; and see the discussion of this in Zhang, Bei Wei zheng zhi shi, 2: 323–27.
8. On the uncertain nature of Wey’s relationship to Daowu, see Chapter 7 note 61. For the apparent existence already of clear heir apparency, see WS 2.41; and the comments of Li, Bei Wei Pingcheng shi dai, 71. It needs also pointed out that the Chinese name (or title) assigned Mingyuan was “Si” 嗣 (WS 3.49), which simply means “successor.” WS 4A.69 also states that Taiwu was born to Mingyuan in the latter’s “eastern palace,” the Chinese designation of the heir apparent’s residence; and see also discussion of Taiwu’s designation as heir in Li, Bei Wei Pingcheng shi dai, 82. It is true, as Holmgren, “Harem in Northern Wei Politics,” 73, correctly states, that Mingyuan “was never officially proclaimed heir-apparent in the Chinese manner.” This does not, however, mean that he was not designated as heir; instead, here we see yet another sign of how this regime was still largely organized on the basis of Inner Asian tradition (or ad hoc invention), with key aspects not seen in the incomplete record provided in its Chinese history. Vertical succession had already clearly appeared in the Inner Asian world, among the Xiongnu: Di Cosmo, “Aristocratic Elites,” 26. For a general discussion of contested succession in Northern Wei, see Eisenberg, Kingship in Early Medieval China, Chapter 2.
9. BS 15.562 (WS 15.371).
10. WS 2.44.
11. BS 15.575 (WS 15.382).
12. Luo Xin, “Zhi qin,” 104–5; Albert Dien, “Elite Lineages and the T’o-pa Accommodation,” 76–77, citing Marshall D. Sahlins, Tribesmen (Englewood Cliffs, NJ: Prentice-Hall, 1968), 24. On the broadly shared view of corporate ownership of the patrimony, see Biran, “Introduction,” Nomads as Agents of Cultural Change.
13. Eisenberg, Kingship in Early Medieval China, 46.
14. WS 2.41–42, 113.2973. Chen, Shi jia da zu yu Bei chao zheng zhi, 3, points out the importance of this reform of peerage in development of the Taghbach state.
15. Though not receiving land, they did receive bound labor: WS 113.2974; Yan, Bei Wei qian qi zheng zhi zhi du, 180–82. See also Chen, Shi jia da zu yu Bei chao zheng zhi, 12–13; and discussion of the (partial) linkage between peerage and rank within military and administration in Daichi Seiko 大知圣子 (大知圣子), “Guan yu Bei Wei qian qi jue he pin xiang dui ying de ji chu kao cha—yi Nanxun bei wei zhong xin” 关于北魏前期爵和品相对应的基础考察——以南巡碑为中心, in Zhongguo Wei Jin Nan Bei chao shi xue hui di shi jie nian hui ji guo ji xue shu yan tao hui lun wen ji, ed. Zhongguo Wei Jin Nan Bei chao shi xue hui, Shanxi da xue li shi wen hua xue yuan (Taiyuan: Beiyue wen yi chu ban she, 2012), 92–107.
16. Yan, Bei Wei qian qi zheng zhi zhi du, 17. Though a woman of the Helan clan—close affines of the Taghbach—would meet her end this way, as we shall see just below, these were viewed as Xiongnu (see Yao, Bei chao Hu xing kao, 36).
17. Holmgren, “Women and Political Power,” 62, makes the good point that this was linked with a shift from fraternal to filial succession. I do have a somewhat different view than that which she states on p. 63, that “the traditional T’o-pa [Tuoba] leadership structure was based on the concept of a male-dominated military dictatorship which allowed very little room for female or distaff participation in political affairs.” In the opinion of this author, Daowu chose to kill the heir’s mother precisely because women and distaff kinsmen had earlier participated in political affairs, and Daowu did not want them to continue to do so.
18. Tian, Tuoba shi tan, 15–68; and see also Scott Pearce, “Nurses, Nurslings, and New Shapes of Power in the Mid-Wei Court,” Asia Major, 3rd ser., 22.1 (2009): 287–309.
19. WS 24.621; Zhang, Bei Wei zheng zhi shi, 2: 341. A theory was propounded in EMC in 2002 (8 [2002], 1–41) by Valentin Golovachev that the practice had roots in ancient Inner Asian practice (“Matricide among the Tuoba-Xianbei and Its Transformation during the Northern Wei”), to which this author wrote a rebuttal in EMC 9.
20. BS 13.492 (WS 13.325); WS 2.36, 44.
21. BS 13.493 (WS 13.325). See the note on this by James Ware, “An Ordeal among the T’o-pa Wei,” T’oung Pao, 2nd ser., 32.4 (1936): 205–9; and Cheng Ya-ju (Zheng Yaru) 郑雅如, “Han zhi yu Hu feng: chong tan Bei Wei de ‘huang hou,’ ‘huang tai hou’ zhi du” 汉制与胡风：重探北魏的「皇后」、「皇太后」制度, Zhong yang yan jiu yuan li shi yu yan yan jiu suo ji kan 90.1 (2019): 6–7.
22. WS 34.799, 800; ZZTJ 115.3622.
23. WS 16.390.
24. WS 16.389–90.
25. ZZTJ 115.3623. This rendition of the quote is more plausible than that given in Wei shu (16.390), which says, “I have a father, and I have an elder brother.” From the outlines of the story, it seems clear that the assembly knew that Daowu had died.
26. In Wei shu given under the anachronistic “Zhangsun Song.”
27. See discussion of the use of textiles as currency in Wei Wenjiang 卫文江, “Bei Wei shi qi de huo bi liu tong” 北魏时期的货币流通, Bei chao yan jiu 4 (2004): 281–84; and Richard Von Glahn, The Economic History of China: From Antiquity to the Nineteenth Century (Cambridge: Cambridge University Press, 2016), 178. This would continue under Tang.
28. WS 16.390, 83A.1813. See also Tian, Tuoba shi tan, 73.
29. WS 16.390; Yin, “Bei Wei Pingcheng shi lüe,” 194. These events were hopefully embellished by their chronicler.
30. WS 34.799.
31. For the transcribed name Mumo, see SoS 95.2322; NQS 57.983, 984.
32. WS 3.49, 50, 2.44. It may, of course, have taken time to build the tomb; but it will be noted that for Mingyuan himself it took only one month from death to burial: WS 3.64.
33. WS 3.64: 逢枭镜之祸. The terms xiao and jing each refer to a different mythological creature that will eat its own parent, more generally to one lacking all human feeling: see HYDCD 4: 1054b.
34. WS 3.49, 50.
35. ZZTJ 115.3624; WS 24.622, 25.643, 17.698, 113.2975. It is not clear how institutionalized this was; the Wei shu says that “the people of this age called them the eight lords” 世号八公 (WS 25.643). Some have suggested the term “eight lords” was borrowed from Wei and Jin (Liu Jun 刘军, “Bei Wei huang zu zheng zhi ti zhi zhi kao cha—yi zong shi ba gong wei zhong xin” 北魏皇族政治体制之考察——以宗室八公为中心, Nandu xue tan 34.6 [2014]: 21–27), though Hucker, Official Titles, 359 #4371, expresses skepticism. Be that as it may, Liu in the article makes a good case that this and related posts were important bases of power for the broader imperial clan vis-à-vis the emperor and his court. For a good guess of location of the Zhichemen as the main southern gate of the entire palace complex, see Duan and Zhao, Tian xia da tong, 55.
36. WS 3.50; 29.705.
37. Two deaths came in the year 416, and one each in 421 and 422. No detail is provided on how or why the young men died: WS 3.56, 61; WS 16.391, 395, 399. Not all of these necessarily derived from power struggles within the family; for one of the four deaths, the one in 421, Mingyuan is said to have expressed great grief, and to have paid for a well-stocked tomb. See discussion of “sib elimination” during the Mingyuan reign in Eisenberg, Kingship in Early Medieval China, 43.
38. WS 3.63. For broader discussion of the northern garrisons, see Scott Pearce, “The Land of Tai: The Origins, Evolution and Historical Significance of a Community of the Inner Asian Frontier,” in Opuscula Altaica: Essays Presented in Honor of Henry Schwarz, ed. Edward H. Kaplan and Donald W. Whisenhunt (Bellingham: Western Washington University, 1994): 465–98.
39. This was, of course, quite distinct from the early modern Song dynasty of the Zhao family, 960–1279.
40. ZZTJ 117.3689ff.
41. ZZTJ 119, passim; Zhang, Bei Wei zheng zhi shi, 2: 500.
42. ZZTJ 119.3758–59.
43. ZZTJ 118.2724, 119.2753–54; SoS 95.2337.
44. WS 2.31.
45. WS 30.713.
46. Yan, Bei Wei qian qi zheng zhi zhi du, 79–83. And see the 415 edict berating corrupt officials and declaring that taxes from that year still unpaid would as punishment come from the official’s own household budget, and were not to be levied on the peasantry: WS 3.55.
