Hostname: page-component-77c89778f8-rkxrd Total loading time: 0 Render date: 2024-07-19T07:12:44.715Z Has data issue: false hasContentIssue false

The Evolution of the Tang Political Elite and its Marriage Network

Published online by Cambridge University Press:  21 April 2020

Nicolas Tackett*
University of California, Berkeley
*Corresponding author. Email:


How did the Tang political elite evolve between the seventh and ninth centuries? Using network analysis and a large prosopographic database, this article approaches this question from four perspectives: the marriage network of political elites, the backgrounds of chief ministers, the composition of the capital elite during three time slices, and the makeup of the provincial elite. Despite important continuities in the elite marriage network's basic structure, there were also significant discontinuities. Between the seventh and eighth centuries, Luoyang emerged as a secondary political center, and Luoyang-based families—including so-called “marriage-ban” clans—acquired a renewed significance, partly at the expense of old southern clans, whose political significance declined over the course of the dynasty. In addition, the political divide between capital and provinces grew over time, culminating in the ninth century with capital-based men occupying nearly all significant provincial posts and provincials serving only locally and in second-tier offices.

Research Article
Copyright © The Author(s), 2020. Published by Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)


1 Tackett, Nicolas, The Destruction of the Medieval Chinese Aristocracy (Cambridge, MA: Harvard University Asia Center, 2014)CrossRefGoogle Scholar.

2 For an excellent overview of the economic consequences of the An Lushan Rebellion, see Von Glahn, Richard, The Economic History of China: From Antiquity to the Nineteenth Century (New York: Cambridge University Press, 2016), 210–17CrossRefGoogle Scholar.

3 The use of the term “aristocracy” to describe the Tang political elite remains contentious. For arguments against its use, see Johnson, David, The Medieval Chinese Oligarchy (Boulder, CO: Westview Press, 1977)Google Scholar and Grafflin, Dennis, “The Great Families in Medieval South China,” Harvard Journal of Asiatic Studies 41.1 (1981), 6574CrossRefGoogle Scholar. For arguments in favor, and a discussion of the Tang political elite as a “bureaucratic aristocracy,” see Tackett, Destruction, 11–12, 61–66, as well as the discussion of the “medieval aristocracy” in Nicolas Tackett, “The Social Structure of Sui-Tang China,” in The Cambridge History of China, Vol. 4: Sui and T'ang China, 589–906 AD, Part Two, edited by Anthony Deblasi (New York: Cambridge University Press, forthcoming).

4 See, for example, Yinke, Chen 陳寅恪, Tangdai zhengzhi shi shulun gao 唐代政治史述論稿 (Shanghai: Shangwu yinshuguan, 1947)Google Scholar; Pulleyblank, Edwin G., The Background of the Rebellion of An Lu-shan (New York: Oxford University Press, 1955)Google Scholar; Tang Zhangru 唐長孺, “Menfa de xingcheng ji qi shuailuo” 門閥的形成及其衰落, Wuhan daxue xuebao (renwen kexue ban) 1959.8:1–24; Guodong, Sun 孫國棟, “Tang Song zhi ji shehui mendi zhi xiaorong: Tang Song zhi ji shehui zhuanbian yanjiu zhi yi” 唐宋之際社會門第之消融: 唐宋之際社會轉變研究之一, Xinya xuebao 4.1 (1959), 211304Google Scholar; Mamoru, Tonami 礪波護, “Chūsei kizokusei no hōkai to hekishōsei” 中世貴族制の崩壞と辟召制, Tōyōshi kenkyū 21.3 (1962), 245–70Google Scholar; Mao Hanguang 毛漢光, “Tangdai tongzhi jieceng shehui biandong: cong guanli jiating beijing kan shehui liudong” 唐代統治階層社會變動:從官吏家庭背景看社會流動 (PhD thesis, Guoli zhengzhi daxue zhengzhi yanjiusuo, 1968); Johnson, David, “The Last Years of a Great Clan: The Li Family of Chao chün in Late T'ang and Early Sung,” Harvard Journal of Asiatic Studies 37.1 (1977), 5102CrossRefGoogle Scholar; Johnson, Medieval Chinese Oligarchy; Ebrey, Patricia Buckley, The Aristocratic Families of Early Imperial China: A Case Study of the Po-ling Ts'ui Family (New York: Cambridge University Press, 1978)CrossRefGoogle Scholar; Hanguang, Mao 毛漢光, “Cong shizu jiguan qianyi kan Tangdai shizu zhi zhongyang hua” 從士族籍貫遷移看唐代士族之中央化, Zhongyang yanjiuyuan lishi yuyan yanjiusuo jikan 52.3 (1981), 421510Google Scholar.

5 Sun Guodong, “Tang Song zhi ji shehui mendi,” 245, 283; Hanguang, Mao 毛漢光, “Zhonggu tongzhi jieceng zhi shehui chengfen” 中古統治階層之社會成分, in Zhongguo zhonggu shehui shilun 中國中古社會史論 (Taipei: Lianjing chuban shiye gongsi, 1988), 3150Google Scholar; Johnson, Medieval Chinese Oligarchy, 131–41; Tackett, Destruction, 160–78.

6 For exceptional early studies based on large datasets derived in part from epitaphs, see the work of Hanguang, Mao, esp., “Tangdai tongzhi jieceng shehui biandong,” as well as the articles republished in Mao Hanguang 毛漢光, Zhongguo zhonggu shehui shilun 中國中古社會史論 (Taipei: Lianjing chuban shiye gongsi, 1988)Google Scholar.

7 Shaoliang, Zhou 周紹良 and Chao, Zhao 趙超, eds., Tangdai muzhi huibian 唐代墓誌彙編 (Shanghai: Shanghai guji, 1991)Google Scholar; Shaoliang, Zhou and Chao, Zhao, eds., Tangdai muzhi huibian xuji 唐代墓誌彙編續集 (Shanghai: Shanghai guji, 2001)Google Scholar.

8 Some of the important recent sources of epitaphs from Chang'an and Luoyang that I turned to are Chang'an xinchu muzhi 長安新出墓誌 (Beijing: Wenwu, 2011)Google Scholar; Yuntong, Qi 齊運通, ed., Luoyang xinhuo qichao muzhi 洛陽新獲七朝墓誌 (Beijing: Zhonghua shuju, 2012)Google Scholar; Maoyu, Guo 郭茂育 and Shuisen, Zhao 趙水森, eds., Luoyang chutu yuanyang zhi jilu (2012) 洛陽出土鴛鴦誌輯錄 (2012) (Beijing: Guojia tushuguan, 2012)Google Scholar; Junping, Zhao 趙君平 and Wencheng, Zhao 趙文成, eds., Qin Jin Yu xinchu muzhi souyi 秦晉豫新出墓誌蒐佚 (Beijing: Guojia tushuguan, 2012)Google Scholar; Yangguang, Mao 毛陽光 and Fuwei, Yu 余扶危, eds., Luoyang liusan Tangdai muzhi huibian 洛陽流散唐代墓誌彙編 (Beijing: Guojia tushuguan, 2013)Google Scholar; Liguang, Zhao 趙立光, ed., Xi'an beilin bowuguan xincang muzhi xubian 西安碑林博物館新藏墓誌續編 (Xi'an: Shaanxi shifan daxue, 2014)Google Scholar; Xi'an xinhuo muzhi jicui 西安新獲墓誌集萃 (Beijing: Wenwu, 2016)Google Scholar; Ming, Li 李明, et al. , eds., Chang'an Gaoyang yuan xin chutu Sui Tang muzhi 長安高陽原新出土隋唐墓誌 (Beijing: Wenwu, 2016)Google Scholar; Yutong, Qi 齊運通 and Jianfeng, Yang 楊建鋒, eds., Luoyang xinhuo muzhi 2015 洛陽新獲墓誌 二〇一五 (Beijing: Zhonghua shuju, 2017)Google Scholar. For some of the important recent sources of provincial epitaphs from North China, see notes to Figure 5. Finally, in order to expand the temporal scope of my epitaph survey to pre-Tang times, I included a large number of epitaphs taken from Qiyi, Wang 王其禕 and Xiaowei, Zhou 周曉薇, eds., Suidai muzhiming huikao 隋代墓誌銘彙考 (Beijing: Xianzhuang shuju, 2007)Google Scholar.

9 The database circulates as a Microsoft Access file entitled tbdb015.accdb; it can currently be downloaded from my department website and from

10 One exception is that, largely following the lead of the genealogical tables of chief ministers in the New Tang History, my database treats as distinct the major sub-branches of many of the largest families—branches that typically branched off well before the founding of the Sui.

11 Especially in the capital region, it may be that many of these small patrilines were in fact sub-branches of large politically powerful families, though genealogical evidence is (as of yet) lacking to confirm the connections.

12 For a diagram of the network on the basis of ninth-century data primarily, see Tackett, Destruction, 123.

13 The 87 patrilines include 14,444 known members. If one includes the leaves and minor branches eliminated from the figures (filtering the data only on the giant component), the entire marriage network contains 1257 patrilines; 24,131 individuals; and 1639 marriages. The majority of the excluded patrilines are small in size (often reconstructed from a single tomb epitaph); many of these are likely to be sub-branches of the 87 large patrilines included in the figures, but there is not yet sufficient evidence to confirm the genealogical link.

14 One must also take into account the effect (described below) of a substantially greater number of surviving epitaphs for marriage-ban clans in the later period.

15 See query JCH_note_15 in TBDB ver. 1.5. Specifically, in the period 600–700 CE, individual members of the marriage-ban and imperial-clan cliques were 10.5 times (n=115) more likely to marry within their own clique rather than with a member of the opposite clique. In the periods 700–760 CE and 780–880 CE, intra-clique marriages were 9.8 times (n=129) more likely and 11.7 times (n=190) more likely, respectively. These calculations exclude marriages involving members of patrilines that were within the marriage network but not obviously parts of either of the two cliques.

16 See query JCH_note_16 in TBDB ver. 1.5. By contrast, the discrepancy between the two cliques is less evident if one considers a somewhat shorter span of time: 72 percent of marriage-ban intra-clique marriages and 64 percent of imperial-clan intra-clique marriages involved two patrilines that intermarried with each other across at least two periods of time.

17 In Figure 3c, Guo Ziyi's patriline appears as the node with the greatest number of documented marriage ties to the large node representing the imperial clan. All known marriages between Guo's patriline and the imperial clan date to the period after 760 CE.

18 For examples of intermarriages between the “eastern” branch of the Zhaojun Lis and the “second” and the “great” branches of the Boling Cuis dating to the sixth century; see Tackett, Destruction, 145n83.

19 On why a Chinese elite family's “primary geographic attachment” was near the family cemetery, see Bossler, Beverly J., Powerful Relations: Kinship, Status, and the State in Sung China, 960–1279 (Cambridge, MA: Council on East Asian Studies, 1998), 4243CrossRefGoogle Scholar. Empirical evidence suggests family members indeed spent most of their time at the place of family burial. See Tackett, Destruction, 73–82.

20 As a consequence of the strong desirability to bury family members together, when one finds multiple epitaphs for members of a single large patriline, it is common to find multiple sub-branches of the patriline buried in the same general region. My sense is that, unless a burial is said explicitly to be “temporary” (權), a single epitaph indicates the likely site of burial of the entire patriline in the period of time in question. For an in-depth discussion of this point with concrete examples, see Tackett, Destruction, 113–18.

21 There are also relatively comprehensive lists of the heads of certain ministries, as well as of prefects of the most important prefectures. These, too, would constitute interesting samples for analysis.

22 Moreover, about 65 percent of these families had ten or more known officeholders. See query JCH_note_22 in TBDB ver. 1.5.

23 That is, chief ministers serving more than one tenure in office will appear more than once in the tabulation. Focusing on tenures in office permits one to focus on the moment of selection of chief ministers. One could also tabulate the figures in terms of service years, which produces roughly similar results (data not shown).

24 Johnson, Medieval Chinese Oligarchy, 131–44.

25 The remaining 20–30 percent were from patrilines for whom there is as of yet insufficient data to ascertain their home bases.

26 Aiko, Maeda 前田愛子, “Jotei Bu Sokuten to Tōdai kizoku: Santō gosei o chūshin to shite” 女帝武則天と唐代貴族: 山東五姓を中心として, in Nishijima Sadao Hakushi tsuitō ronbunshū: Higashi Ajia shi no tenkai to Nihon 西嶋定生博士追悼論文集: 東アジア史の展開と日本 (Tokyo: Yamakawa, 2000), 309–36Google Scholar.

27 Johnson, “Last Years of a Great Clan,” 32–40; Ebrey, Aristocratic Families, 91–93; Mao Hanguang, “Cong shizu jiguan qianyi kan Tangdai shizu.”

28 Tackett, Destruction, 226 (Figure 5.4).

29 In addition, early Tang epitaphs are inconsistent in how they identify ancestors. Often they shorten a two-character given name into a single character, or they refer to the ancestor by his style (字) rather than his given name (諱), or, in some cases, even by a combination of the given name and style. As a result, it is easy to miss genealogical ties that exist between individuals mentioned in epitaphs and the patrilines reconstructed in a database.

30 Most of these are listed in Tackett, Destruction, 45n49. By the ninth century, outside of the Capital Corridor, the epitaph record leaves few traces of individuals still based in the prefectures designated by their choronyms, except in the case of the Lower Yangzi, where one still found a population of non-officeholding elites claiming the locale as their place of family origin. See Tackett, Destruction, 44–61.

31 Tackett, Destruction, 16–25.

32 The dip between the 720s and 800s may in part be a consequence of the sample size. Looking at the entire period between 800 and 880 (by resurrecting data I employed in a previous study), one finds 82 percent (rather than 76 percent) of epitaphs are for individuals from families with strong officeholding traditions.

33 The resurgence between the 720s and early 800s may be misleading (see previous note). Looking at the entire period between 800 and 880, the figure for no family tradition of officeholding is 4 percent rather than 7 percent, and the figure for office type = “none” is 9 percent rather than 12 percent.

34 For instance, it is common for an epitaph to identify a man's spouse as, say, Ms. Li of Zhaojun. One knows the surname of the deceased's spouse (Li) and her choronym (Zhaojun), but nothing telling us who her father was or which patriline she belonged to.

35 For more on these lists, see Johnson, Medieval Chinese Oligarchy, 62–70; Tackett, Destruction, esp. 31–35; as well as the footnotes therein.

36 By contrast, row 11 does not count choronyms that appear less prominently in the epitaph, for example as the place of ancestral origin of the deceased.

37 Given the overabundance of extant Luzhou epitaphs, I decided to exclude them from this brief survey. For detailed studies of this corpus, see Xu, Man, “China's Local Elites in Transition: Seventh- to Twelfth-Century Epitaphs Excavated in Luzhou,” Asia Major 3rd ser., 30.1 (2017), 59107Google Scholar; Xu, Man, “Ancestors, Spouses, and Descendants: The Transformation of Epitaph Writing in Song Luzhou,” Journal of Sung-Yuan Studies 46 (2016), 119–68CrossRefGoogle Scholar. Regarding the south, the overwhelming majority of southern epitaphs date to the late eighth century or later. See Tackett, “Elite in the Sui and T'ang,” n. 78. As a consequence, southern epitaphs are not helpful for tracing trends over time—except insofar as their relatively late emergence may provide circumstantial evidence for the rapid accumulation of private wealth in the post-An Lushan south.

38 It is also worth noting that many regions of north China produced too few epitaphs to produce statistically significant results. Examining trends throughout all of north China guarantees a sufficiently large sample size. Historians interested in the Sui-Tang provinces may in the future wish to pay particular attention to the unique characteristics of frontier prefectures (numerous epitaphs survive from both Xiazhou and Yunzhou); the Hebei autonomous provinces, especially Youzhou, which was unique in numerous ways; the Zhaoyi-Weibo region (which would include the corpus of Luzhou epitaphs); other provinces with important military forces (e.g., a number of epitaphs from Pinglu Province survive); as well as the Lower Yangzi and Northern Zhejiang.

39 Not reflected in Figure 4 is a surge in the post-An Lushan period of epitaphs from provinces with major provincial governments—namely the provinces of Youzhou 幽州, Hedong 河東, Chengde 成德, Yiwu 義武, Pinglu 平盧, and Yichang 義昌—undoubtedly reflecting the important impact of provincial governments on mid- to late-Tang provincial society. Epitaphs from these six provinces account for 29 percent of provincial epitaphs up through the 750s, then 62 percent of provincial epitaphs in subsequent decades.

40 For example, based on their epitaphs, the provincials Cheng Junxin 成君信 (812–878), Wang De 王德 (779–856), Li Yuanshun 李元順 (810–867), and Guo Fan 郭璠 (799–872) all appear to have served in the military without holding an identifiable office; and Wang Shang 王尚 (ca. 669–ca. 753) contributed only informally to local governance (receiving an honorary title late in life in recognition of his advanced age). See Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian, 2491; Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian xuji, 1075–76; Shangjun, Chen 陳尚君, Quan Tang wen bubian 全唐文補編 (Beijing: Zhonghua shuju, 2005), 1900Google Scholar; Yingmin, Fan 樊英民, ed., Yanzhou lidai beike lu kao 兗州歷代碑刻錄考 (Jinan: Qi Lu shushe, 2013), 3840Google Scholar; Yuanlong, Miao 苗元隆, ed., San Jin shike daquan 三晉石刻大全: Taiyuan shi Jiancaoping qu juan 太原市尖草坪區卷 (Taiyuan: San Jin, 2012), 315–16Google Scholar.

41 For example, Wang Quan 王詮 (583–668), buried in Xiazhou 夏州 on the northern frontier, had been sent to the far south to fight the “southern barbarians” 南蠻, after which “in recompense for his service to his country, he was awarded by special decree the [merit title of] Commandant of Courageous Guards” 報國酬庸, 蒙授驍騎尉. But his epitaph mentions no substantive military offices. The epitaph of Yan Shen 閻神 (691–756) offers a similar example. See Xin Zhongguo chutu muzhi: Shaanxi (3) 新中國出土墓誌: 陝西 (叁) (Beijing: Wenwu, 2016), 下: 19Google Scholar; Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian xuji, 670. As a general rule, I treat men whose only identifiable official title was a merit title to be non-officeholders. The epitaph of Wang Xuandu 王玄度 (625–695) makes abundantly clear that he “never craved rank and wealth [i.e. an official salary], and never served a lord [i.e. the state]” (不貪榮祿不事王侯) despite holding the title of Supreme Pillar of State (上柱國). Other epitaphs—including those of Song Lü 宋履 (d. ca. 702) and Long Rui 龍叡 (656–741)—identify their subjects both as holders of merit titles and as “recluses” (處士), a common euphemism for a man who never served in office. See Xin Zhongguo chutu muzhi: Shaanxi (3) 下: 26; Jimin, Sun 孫繼民 et al. , eds., Hebei xin faxian shike tiji yu Sui Tang shi yanjiu 河北新發現石刻題記與隋唐史研究 (Shijiazhuang: Hebei renmin, 2006), 290–92Google Scholar; Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian xuji, 578.

42 In the database, within the broad category of “office eligible,” I include exam graduates, tribute scholars (who were not yet technically “eligible”), and candidates for appointment by the civil or military bureaus of appointments (i.e., men identified as 吏部常選 or 兵部常選, respectively). I also include men whose only identifiable official title is a civil prestige title below rank 5, as I have found examples of such men who are explicitly said to have never held a substantive, salaried position. For the moment, I have been treating higher ranked civil prestige titles as well as military prestige titles as evidence of officeholding.

43 In Tang times, such special titles are often identified as “appointments by courtesy” (版授). The man Wang Shang of note 40 died at age 85, but not before “a decree appointed him by courtesy Magistrate of the present [i.e. Yangqu] County in honor of his advanced age” (有制版授本縣令寵耆壽矣). See Miao Yuanlong, San Jin shike daquan: Taiyuan shi Jiancaoping qu juan, 下:315–316. Unfortunately, such honorary offices are often indistinguishable from substantive offices (a problem sometimes also affecting posthumous titles). In the case of Long Run 龍潤 (561–653), we know from his own epitaph that his highest-ranking substantive office involved a relatively modest position as administrator in a fubing garrison. Then, “in the twentieth year of the Zhenguan era [646], his age was lofty, having exceeded eighty; when the imperial carriage visited Jinyang [his home county], [the emperor] personally inquired about the elderly, and with an edict he was appointed by courtesy Prefect of Liaozhou” (貞觀廿年, 春秋廖廓, 已八十有餘, 駕幸晉陽, 親問耆老, 詔板授遼州刺史). In the epitaph of one of his sons, this final office is correctly identified as a courtesy appointment. But four epitaphs for later descendants identify him as Prefect of Liaozhou without revealing the post to be merely honorary. See Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian xuji, 75, 119, 132, 252–53, 332–33, 578.

44 Occasionally one hears of the impact of the Sui–Tang transition on provincial families. The epitaphs of Xue Shigan 薛世感 (626–675) and Ren Suiliang 任遂良 (d. 731) explicitly mention the hardships faced by the family at this time. The epitaphs of Ren Zhong 任忠 (646–698) and Meng Bin 孟賓 (706–753) mention fleeing home during the transition and establishing new roots elsewhere. See Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian xuji, 463, 491–92; Xin Zhongguo chutu muzhi: Shaanxi (3), 下: 35–36, 46–47.

45 The governor in question was Wang Yuankui 王元逵 (812–854). For his epitaph and the epitaph of his daughter, see Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian, 2324–25, 2432–33. The other provincials in question are Wang Shizhen 王士真 (759–809), an earlier Chengde governor unrelated to Yuankui, whose brother and son both married daughters of emperors; and Zhang Youming 張佑明 (788–840), Zhang Feng 張鋒 (808–840), and Feng's wife Ms. Shi 史氏 (821–847). The Zhang's were close relatives of the two governors who had controlled Yiwu Province from 782 to 810. For the epitaphs of these four individuals, see Feng Jinzhong 馮金忠 and Zhao Shengquan 趙生泉, “Hebei Zhengding chutu Tang Chengde jiedushi Wang Shizhen muzhi chutan” 河北正定出土唐成德節度使王士真墓誌初探, Zhongguo guojia bowuguan guankan 2013.5: 80–86; Zhu Xuewu 朱學武, “Hebei Laishui Tang mu qingli jianbao” 河北淶水唐墓清理簡報, Wenwu chunqiu 1997.2:21–24, 26; Zhou Shaoliang and Zhao Chao, Tangdai muzhi huibian, 2255–56, 2270–71.

46 Tackett, Destruction, 91–98, 105.

47 For Wei Xiang's epitaph, see Wencheng, Zhao 趙文成 and Junping, Zhao 趙君平, eds., Qin Jin Yu xinchu muzhi souyi 秦晉豫新出墓誌蒐佚 (Beijing: Guojia tushuguan, 2012), 4: 1021–22Google Scholar. Wei's father had ended his career as Chief Administrator of Bozhou, also in Weibo Province. The epitaph of Li Hong 李洪 appears to constitute a similar case. See Xin Zhongguo chutu muzhi: Beijing (yi) 新中國出土墓誌: 北京(一) 上:20, 下:13–14. Li was the grandson of Crown Prince Zhanghuai 章懷 (killed in Empress Wu's 684 purge), and he had married the daughter of an old Hedong family. All of his (at least) twenty older brothers remained in prominent positions in the capital. But as the youngest of the bunch, he apparently decided at some point (probably during the immediate aftermath of the An Lushan Rebellion) to seek his fortune in “our fief” (我邦), i.e., autonomous Youzhou, where he served in a series of administrative positions. His son too would serve the Youzhou government.

48 Tackett, Destruction, 176–78.

49 A survey of 109 provincial garrison appointments mentioned in the sample of provincial epitaphs suggests that the median distance (between seat of prefecture of burial and seat of garrison, with distance calculated using the Haversine formula) was 324 km. By contrast, capital elites who were appointed to fubing garrisons went a median distance of 182 km away from home.

50 For more on the ninth-century situation, see Tackett, Destruction, 182.

51 Maeda Aiko has previously argued this point. See “Jotei Bu Sokuten to Tōdai kizoku.”

52 Chen Yinke, Tangdai zhengzhi shi shulun gao. For an early critique of the idea of regional blocks as the basis of political factionalism, see Wechsler, Howard J., “Factionalism in Early T'ang Government,” in Perspectives on the T'ang, edited by Wright, Arthur F. and Twitchett, Denis (New Haven: Yale University Press, 1973), 87120Google Scholar.

53 For example, Chōfū, Nunome 布目潮渢, Zui Tō shi kenkyū: Tōchō seiken no keisei 隋唐史研究: 唐朝政權の形成 (Tokyo: Tōyōshi Kenkyūkai, 1968)Google Scholar. Cf. Wechsler, “Factionalism in Early T'ang Government,” who argues that the early Tang emperors sought to balance the power of northwesterners, northeasterners, and southerners. My data corroborates Wechsler's, but shows that the prominence of southerners rapidly declined after the early Tang.

54 Indeed, Mao Hanguang has shown that a small number of families actually increased their influence over time and came to represent by the end of the Tang Dynasty close to half of all known officeholders. See “Tangdai tongzhi jieceng shehui biandong,” esp. 223–24.

55 Tackett, Destruction, 45–55.

56 For concrete examples of the deployment of social capital in the ninth century, see Tackett, Destruction, 129–41.

57 I have also noticed a greater tendency in epitaphs of the seventh century to confuse ancestors (e.g. skip from the grandfather to the great great grandfather) or to confuse ancestors’ offices (e.g. mix up the great grandfather's office with the grandfather's office), suggesting to me that it was in the eighth century that families began to take more care in compiling genealogies and standardizing their contents.

58 For an example, see Tackett, Destruction, 27–28.