84. Note that this is the version given in WS 33.793; the BS 27.981 version is slightly different.
85. ZZTJ 131.4104; BS 13.495 (WS 13.328); BS 15.553 (WS 14.357); WS 33.793. And as for the princes, see notice of the arrival in court of several imperial princes in WS 6.126; pointing this out is Zhang, Bei Wei zheng zhi shi, 5: 184. Modern scholars have different views as to who was the real leader of the plan to get rid of Yi Hun: Song, Bei Wei nü zhu lun, 140, holds it was the empress. Zhang, Bei Wei zheng zhi shi, 5: 186–87; and Eisenberg, Kingship in Early Medieval China, 71, lean more toward the princes, despite the fact that they are little mentioned in the primary accounts. Most likely, it was a bit of both.
86. BS 13.496 (WS 13.328, 329 note 10); Holmgren, “Social Mobility in the Northern Dynasties,” 28–29; Zhang, Bei Wei zheng zhi shi, 5: 194ff.
87. BS 13.495 (WS 13.328).
88. See Ruby Lal’s discussion of “adoption,” or perhaps “taking charge,” of other women’s children by senior females at the Mughal court: Domesticity and Power in the Early Mughal World, 120–23.
89. BS 13.498 (WS 13.331). Some modern scholars have even suggested that Wenming was actually Xiaowen’s biological mother. This is effectively debunked by Li, Bei Wei Pingcheng shi dai, 195–208; Zhang, Bei Wei zheng zhi shi, 6: 62.
90. SoS 72.1871.
91. WS 6.127, 50.1110; SoS 8.159–60; ZZTJ 132.4129–30.
92. ZZTJ 132.4148–49; Chen, Buddhism in China, 154–55; Huang Wen-Yi, “Negotiating Boundaries: Cross-Border Migrants in Early Medieval China” (PhD diss., McGill University, 2017), 28–42.
93. WS 19B.461; ZZTJ 133.4164–65. See the overview of the abdication in Zhang, Bei Wei zheng zhi shi, 5: 308–20; and Eisenberg, Kingship in Early Medieval China, 72–76, who argues that the abdication was a scripted event, in which Wenming played only a minor role.
94. WS 41.921.
95. WS 6.132; ZZTJ 133.4166; WS 114.3038; Hurvitz, Treatise on Buddhism and Taoism, 75. Eisenberg, Kingship in Early Medieval China, 81–84, suggests that though Xianwen had not himself effected the abdication, this being undertaken by powerful members of the imperial family, once he was in the Deer Park he successfully built power, becoming able to evade and from a distance control the court. Song, Bei Wei nü zhu lun, 149–50, describes a sharing of power that she compares to those earlier seen between Mingyuan and his heir, Taiwu; or Taiwu and his son Huang.
96. WS 7A.142; Song, Bei Wei nü zhu lun, 152.
97. WS 7A.138.
98. Song, Bei Wei nü zhu lun, 156–58.
99. This was Li Yi, a son of Li Shun, Cui Hao’s rival in the debate over the Northern Liang campaign whom Taiwu had put to death: WS 36.841.
100. BS 17.629 (WS 19A.441). Of course, toward the end of the dynasty there were emperors who had the opposite problem.
101. BS 13.495 (WS 13.328); ZZTJ 134.4187.
102. Though not so clear in the annals (WS 6.132, 7A.142), we are told Xianwen was poisoned in other parts of Wei shu: 105C.2413; BS 13.495 (WS 13.328). In the last pair of texts, Wenming’s biographies, Bei shi bluntly states the empress dowager “harmed the emperor (Xianwen)” 害帝, while Wei shu says Xianwen “suddenly expired 暴崩. At the time [people] said the empress dowager had done it.” Accompanying the death of Xianwen was the death of his uncle, Zitui, who had been sent off to govern a province and “unexpectedly died” en route: WS 7A.144; BS 17.632 (WS 19A.443); Song, Bei Wei nü zhu lun, 154.
15
双佛
冯太后生于442年，当她的丈夫文成帝于465年去世时，她还只是位23岁的年轻女性。11年后，34岁的她以截然不同的身份重新登上历史舞台，成为平城政坛的核心人物，以崭新的自信姿态执掌大权。她的权力源自三个方面：第一是作为太后的法定地位，即皇室的家长身份，这赋予她掌管后宫事务等诸多权力；第二是她在宫廷中培植的非正式同盟网络，这个集团以兄长冯熙为首——冯熙迎娶皇室公主，晋升为侍中并获封王爵；第三点或许最为重要，是她对皇帝的情感掌控，这位君主自襁褓时期便由她抚养，并将继续接受她的指导直到490年她本人去世。1
一个显而易见的生活现实是：对所有人类群体而言，最重要、最基础的人际关系通常都形成于人生早期。对大多数人来说，这就是与母亲或童年养育者之间的关系。在中国传统中，亲子关系通过"孝"的概念体系化。儒家经典往往更强调父子关系，或许因为这种关系本质上更为脆弱。但随着佛教融入中古东亚的文化体系，母子关系——这种始终存在的人伦——在公共话语领域获得了新的核心地位。2无论如何，孝文帝与冯太后的关系——这位既是他的继祖母也可称为养母的女性——是被明确以"孝"来界定的。对于一个崇尚儒家经典权威的君主（孝文帝无疑属于此类，即便他的族人未必都是如此），这同样赋予了冯太后极大的权威。
作为掌权者，太后始终保持着对监护对象的绝对控制，这是她在宫廷权力的根本基础。据记载，在孝文帝少年时期某个阶段，太后认为这个孩子过于聪慧，担心其将来可能"不利于冯氏"。3这位"睿智而严厉"4的女性随即决定废黜他，其手段包括在隆冬时节将少年囚禁于无供暖的房间，仅着单衣，连续三日断绝饮食。正当冯太后准备用另一位皇子取代他（显然预期他会很快死亡）时，直勤（tigin）元丕（即第13章提到的宴饮宾客）和穆泰的强烈反对使她放弃了该计划——后者多年后曾领导反对孝文帝迁都的叛乱。5
意料之中，孝文帝对这位自命的监护者怀有深刻而复杂的情感。但以最尽孝道的方式，他在公开立场上从未对这位母亲形象表露任何不满——不仅在国事上恭敬从命，甚至在她用粥时亲自侍膳。6事实上，他们被并称为"二圣"，这种关系似乎通过云冈石窟反复出现的双佛并坐造像得到呼应，该题材源自《法华经》。7当冯太后于490年去世时，这位早已实际承担治国重任8、此时无疑已是成熟男性的皇帝，主动断食绝水五日，继而不顾群臣劝阻，执意按照儒家传统进行漫长而繁复的服丧仪式。9他身后被追谥为"孝文帝"可谓实至名归。
* * *
孝文帝统治时期最广为人知的，是通常被称为"改革"的一系列广泛而彻底的政策与制度变革——这些举措基于"使政权更完善"的明确假设。这些改革分两大阶段实施：第一阶段属务实型，发生于480年代中期，当时朝廷仍处于"二圣"共治之下；第二阶段更具理想主义色彩，始于冯太后490年去世后，旨在依照华夏传统建立人间乌托邦。
一个常见论点（或许有人会提出异议）认为：强大国家依赖强大军队，而强大军队需要可靠经济支持。这方面的早期尝试可见于卫王主导的五原农垦区。尽管驻守代地的国人军士大多能自给自足供养家庭，但出征期间仍需国家补给。数年后，北魏掌控的农田面积大幅扩展，旧帝国遗留的农户组织与赋税制度至少以残存形式延续——这些制度曾被苻秦、慕容诸政权采用，并在拓跋部夺取中原后立即接管。10然而这些制度传入北魏时已处于原始且不完善状态，最初数十年间掠夺战利品并进行分配似乎更为重要，此举不仅能巩固士兵忠诚，还可充实国库。11但太武帝完成大规模征服后，可获取的战利品已所剩无几。尽管宏大战役带来大量财富，特别是长江沿岸的远征实际上耗费更巨，导致太武帝统治末期国库几近空虚。12
提升农业生产并使更多产出进入国家财库与粮仓的临时措施，早在太武帝时期就已启动。444年，太武帝颁布法令要求农户在田地竖立名牌，以明确耕地是否产粮。此举可能收效甚微。十一年后，文成帝派遣三十人使团巡查各州，考察当时仍大量存在的"未垦田地"。13军队与国家亟需更可靠的经济支持体系，这一体系最终在一代人之后完全成型。经济需求至少是推动朝廷不断深入其疆域内最具生产潜力与繁荣前景领域的重要因素之一。
480年代改革的核心举措是建立名为"均田"的国家分配耕地新制度。尽管不像某些学者认为的完全基于"普天之下莫非王土"理念，均田制确实运用平城政权的力量，至少部分限制了大面积土地兼并，同时重建了类似秦汉时期纳税编户的小农基础。14在历经长期战乱与动荡的地区，通过将流民安置于过去两个世纪荒废的土地上，这项制度提高了农业产量并增强了社会稳定。
该制度渊源复杂：其雏形可追溯至中国第一帝国末期的土地实验，而更直接的源头则是北魏为安置迁入代地的数十万移民所实施的户籍土地分配政策（计口授田）在帝国全境的推广。15现代学者对此多有论述，简而言之，均田制似乎是融合了中国帝制时期社会控制与生产汲取的模式，通过拓跋部外来者的强制力得以更系统、更高效地推行。
485年，均田制作为正式提案呈交冯太后与孝文帝。当时一位汉族官员上奏请求采取措施解决中原地区两个交织的难题：其一是因灾荒导致的流民问题；其二是豪强大规模逃税现象。16后者实为北魏开国以来——甚至更早时期——长期存在的顽疾。尽管中国帝制时代的郡县地方行政体系在西晋灭亡后得以延续并最终被北魏继承，但其运作已远非先贤设想的治国理念。这套制度非但未能成为中央集权的控制工具，反而沦为第八章所述地方豪强势力的制度支柱（如第八章所述）。
404年，道武帝曾下诏裁撤新征服地区中户数不足百的县。17这些县治有些可能因四世纪战乱导致人口锐减。18但更多县域实际人口远超户籍残存记录，往往数十甚至上百户家庭通过依附"宗主"——这些被指定的地方首领在均田制实施前充当北魏基层行政单位19——来逃避北魏赋役。当然，这些宗主的利益诉求复杂多元。至少在文成帝看来，地方官吏腐败成风："夺百姓之食以营私产，不纳王税"。20随着485年均田制的推行，中央派遣专员督导地方实施新政。此前平城仅通过宗主对黄河流域实施间接管控，试图通过强制移民充实大同盆地京畿地区来建立可靠税基。如今至少在理论上，中央权力将直接延伸至帝国每个村落。21
北魏改革方案的核心内容包括：每位15岁及以上男性分配约6.5英亩用于谷物生产的"露田"（条件允许地区另配等量休耕地）。这些土地属临时授予，待受田者年满70岁时收回重新分配。女性分配面积减半，而拥有奴隶与耕牛者（视经济能力）亦可获得相应配额。所分配土地无需连片——该政策实施于黄河流域自汉以来形成的碎片化土地格局中，显然更适应此类体系。但在某些地区，家庭还可永久获得小块"桑田"用于植桑养蚕。22
赋税体系围绕这些革新重新设计，以夫妇联合申报为基本计税单位，每对夫妇每年需缴纳帛7匹（依地域差异采用丝或麻）、谷物13石。23此外还对未婚家庭成员、奴隶及耕牛征收附加税。徭役义务包含加入地方民兵，这标志着北魏军事性质根本性变革的开端（详见第十一章）。作为将农业人口原子化为小家庭宏大计划的一部分，田亩与赋税现在某种程度上实现了均衡化，自然也便于朝廷管控。至北魏末期，约500万家庭被纳入该征税体系。24
当然，实际制度比此更为复杂且持续演进。虽从未完美运行并始终因地制宜调整——例如官员家庭可拥有远超平民的土地配额——但这仍是划时代的创新。凭借平城政权从强大可靠军队中汲取的新国力，该制度在随后两个多世纪（远迈北魏国祚）持续开发中原农田的生产潜力，为一系列日益强盛的王朝提供稳定财政基础，最终在唐朝臻于顶峰。25
与全面实施土地国家管控和征税密切相关的，是地方行政体系的组织架构。26此前地方官员需通过"宗主"间接征税。但在均田令颁布次年（486年），重臣李冲上奏提出整套改革方案中另一关键环节：建立"三长制"。该制度构建了从邻（5家）至里（25家）再至党（125家）的垂直管理体系。27三长承担户口登记、赋税征收、徭役民兵组织及社区监管等繁重职责。配套改革减轻了实施难度：如484年确立的官员俸禄制度（前文提及北魏长期面临官员无薪导致的腐败问题，高允事例即体现守规者的窘境）28，这显然成为次年（485）推行均田制的动因之一。俸禄制彻底改变了文官处境，使其待遇与武官持平，尤其惠及缺乏军功背景的中下层汉人官员。
均田制作为一项激进的方案，遭到朝中诸多反对中央集权扩张者的抵制。但最终诏令仍得以颁布。尽管史载孝文帝"览而善之"并作为推行依据29，但另一位"佛"显然也深度参与决策。相较于十五年前孝文帝幼年时"事无巨细，一禀于太后……事有疑速，多阙帝旨"引发"朝野闻之莫不惊怪"30的情形，此时政局已有所变化。至485年均田制推行时，孝文帝年届十七八岁，开始更多参与朝政。不过当486年三长制奏议呈递时，仍是冯太后亲自批阅，称善后召集重臣商议。31虽然元丕支持该计划称"公私便之"，但仍有顽固反对者担忧"其实难行"或"恐致纷扰"。冯太后不为所动，断言"立三长则课有常准，赋有恒分，苞荫之户可出，侥幸之人可止。何为不可？"改革遂得实施。
三长制及配套制度的推行，极大增强了北魏对中原农耕人口的控制力，并构建起更高效的财富汲取体系。32部分旧式乡里领袖通过入仕领取新设俸禄适应时变。鉴于新税制仅为旧税率十分之一（针对实际纳税者），终致"海内安之"。33
但490年"二圣"中的长者离世时，孝文帝并未获得安宁。国事因此暂停数年，皇帝以夸张方式持续服丧。34冯太后灵柩安葬于平城东北约20英里方山上的永固陵。该陵始建481年，封土高约75英尺，双墓室结构，属北魏最宏伟陵寝之一。35值得注意的是此举打破传统：太后未入葬大同盆地西北金陵祖陵区，而是独辟兆域。36此举虽有先例可循（近期其他太后亦有别葬），但更无前例的是冯太后生前即决定其监护对象——皇帝——死后亦不归金陵，而将陪葬方山。耐人寻味的是，孝文帝陵墓封土规模明显小于永固陵。最终，这位年轻皇帝终究成为了冯氏家族的助力。
1. See a parallel analysis of her sources of power in Holmgren, “The Harem in Northern Wei Politics,” 89–90; and the comment of Balkwill, “When Renunciation Is Good Politics,” 247, on how the empress dowager “benefitted politically by being the only woman in the emperor’s life.”
2. For xiao in general during this period, see Knapp, Selfless Offspring. For the mother-child relationship in particular, within the Buddhist context, see Alan Cole, Mothers and Sons in Chinese Buddhism (Stanford, CA: Stanford University Press, 1998). A powerful embodiment of such feeling was the much-loved “Tale of Turnip,” in which a young Buddhist monk goes to Hell to save his mother: see Victor Mair, Tun-huang Popular Narratives (Cambridge and New York: Cambridge University Press, 1983), Chapter 2.
3. WS 7B.186.
4. BS 13.496 (WS 13.329).
5. Eisenberg, Kingship in Early Medieval China, 88. Building on the work of Li Ping and other scholars, Eisenberg quite persuasively also demonstrates that in general Yuan Pi had been an ally of Wenming, though the leader of his own faction based in the military: pp. 85–87. Jennifer Holmgren, “The Harem in Northern Wei Politics,” 88, also points out a passage in Zi zhi tong jian (134.4194) describing how in 477 an heir, angry at the dowager empress, tried to poison her, though was in the end dissuaded. This passage, however, has been misread (in part because it is a rare example of sloppy arrangement in the Zhonghua shu ju ed.). The passage does not refer to Xiaowen, but to the Song heir down in Jiankang; see also the same passage in the Wei shu chapter on the “Island Barbarians,” WS 97.2151.
6. BS 13.496 (WS 13.329); WS 7B.186.
7. Yin, “Yungang shi ku suo fan ying de yi xie Bei Wei zheng zhi she hui qing kuang,” 75–77; and Eugene Wang, Shaping the Lotus Sutra: Buddhist Visual Culture in Medieval China (Seattle and London: University of Washington Press, 2005), 7, 10, who cautions against viewing “the twin Buddhas as iconic portraits of the reigning Twin Sages (though such an analogy may have been intended and exploited by the imperial spin doctors).” Also see the very interesting article examining the use of Buddhism by Wenming in Yungang by Guan Furong 管芙蓉, “Fo mu ta dong yu Feng Taihou—Yungang shi ku wen hua nei han jie du” 佛母塔洞与冯太后—云冈石窟文化内涵解读, in 2005 nian Yungang guo ji xue shu yan jiu hui lun wen ji, yan jiu juan (Beijing: Wen wu chu ban she, 2006), 733–36 (rpt. in her Bei chao san lun, 352–59).
8. Though Wenming exerted real power at court, for all his deference Xiaowen was by 485 clearly taking a more significant role himself. See张金龙, “‘Feng shi gai ge’ shuo shang que” “冯氏改革”说商榷, in his Bei Wei zheng zhi yu zhi du lun gao (Lanzhou: Gansu jiao yu chu ban she, 2002), 28–51 (originally in LSYJ 1986.2).
9. BS 13.497 (WS 13.330); ZZTJ 136.4290.
10. See WS 2.28, 2.31; and Wang, Zhuan xing qi de Bei Wei cai zheng yan jiu, 46–47.
11. Wang, Zhuan xing qi de Bei Wei cai zheng yan jiu, 5–8; Han, Bei chao jing ji shi tan, 36–38. And see comments on the economics of “army and state” in WS 110.2851–52.
12. WS 5.123.
13. WS 4B.109, 5.114.
14. See Von Glahn, The Economic History of China, 175, where building on Hori Toshikazu (see note 15) he rejects the notion that Equal Fields was an assertion that the monarch owned all land within the empire.
15. Hori, Kindensei no kenkyū, 99–114; WS 2.47, 3.53. For the Northern Wei Equal Fields, see also Yang Jiping 杨际平,北朝隋唐均田制新探 (Changsha: Yuelu shu she, 2003) 38–60.
16. WS 53.1176; ZZTJ 136.4268.
17. WS 2.41. In fact, one commandery (jun, near the mod. Shaoyuanzhen, Henan) contained four districts but in all just 52 households, with 158 registered subjects on the rolls (WS 106A.2484–85).
18. Lewis, China between Empires, 51, suggests that as many as a million people may have fled the Yellow River region in the midst of the wars that brought Western Jin down. As noted above, however, Andrew Chittick in his Jiankang Empire has expressed doubt at the size of these migrations.
19. WS 53.1180: This description of the status quo was presented in a memorial from the architect Li Chong, from the Gansu Corridor region, who argued here for establishment of the Three Headmen system, and would go on to serve as chief designer for the new capital at Luoyang. Yan, Bei Wei qian qi zheng zhi zhi du, 95–99, suggests this was a new system established by the Taghbach within the Chinese world, and compares the zong zhu to the local headmen (ling min qiu zhang 领民酋长) set up over the High Carts, and other groups.
20. WS 5.117. This does clash with the well-known position of Tanigawa Michio; see discussion in Chapter 2 note 2.
21. Hori, Kindensei no kenkyū, 97.
22. WS 110.2853–55, 7A.156, 53.1176; ZZTJ 136.4268–69. The figure used to measure land is mu 亩; in WS 110.2853 we are told that the main allocation to a man was 40 mu. Hou, “Bei chao de cun luo,” 48, provides the figure of 677.4 square meters per mu. If this is correct—and Northern Wei figures are sparse and not always reliable—forty mu would figure to about 6.5 acres, or about 2.5 hectares.
23. Hori, Kindensei no kenkyū, 121–22. Only a few Northern Wei measuring devices have been found, and none for capacity: Qiu Guangming 丘光明,中国历代度量衡考 (Beijing: Ke xue chu ban she, 1992), 258. Going on the basis of Han measures, for which a dan 石 was about 20 liters, we can say roughly that the 22.9 dan which Hori, including additional taxes to pay official salary, calculates were required of the taxed couple would equal about 460 dry liters, or 13 US bushels. As for the bolts of fabric (pi 匹), these were 4 zhang long (WS 110.2852), which in this time would be about 40 feet (see Qiu, 69), so about 280 feet of fabric in all.
24. See Ge, Zhongguo ren kou shi, 1: 474, drawing on WS 106.2455, where we are told that the Wei population in the early 520s was something like twice that of Jin in Wudi’s Taikang period (280–289), which had been about 2,450,000 households.
25. Denis Twitchett, “Introduction,” in Cambridge History of China, Vol. 3, Sui and T’ang China, 589–906, Part 1, ed. Denis Twitchett (Cambridge: Cambridge University Press, 1979), 24–25; Victor Cunrui Xiong, Emperor Yang of the Sui Dynasty: His Life, Times, and Legacy (Albany: State University of New York Press, 1999), 180–82.
26. See Hori, Kindensei no kenkyū, 124–28, on the linking of Equal Fields and three headman policies.
27. WS 53.1180, 7B.161, 110.2855; ZZTJ 136.4271–72; and Hou Xudong 侯旭东, “Bei chao ‘san zhang zhi’” 北朝三长制, in his Bei chao cun min de sheng huo shi jie (Beijing: Shang wu yin shu guan, 2005), 112–25.
28. WS 7A.153–54. WS 110.2852; Koga Noboru 古贺登, “Hoku-Gi no hōroku sei shikō ni tsuite” 北魏の俸禄制施行について, Tōyōshi kenkyū 24.2 (1965): 152–76; Wang Daliang 王大良,反腐败: 来自古代中国的启示: 以北魏官吏收入与监察机制为例 (Beijing: Min zu chu ban she, 2001), Chapter 2. A year after this, salaries were also established for peerages held by members of the imperial family: WS 7A.155.
29. WS 53.1176.
30. BS 13.496 (WS 13.329).
31. WS 53.1180; ZZTJ 136.4271–72.
32. For its continuing success even at the end of Wei, as the regime collapsed, see the article by Zhou Yiliang 周一良, “Cong Bei Wei ji ge jun de hu kou bian hua kan san zhang zhi de zuo yong” 从北魏几个郡的户口变化看三长制度的作用, rpt. in his Wei Jin Nan Bei chao shi lun ji xu bian (Beijing: Beijing da xue, 1991), 52–66.
