Rafe de Crespigny


A Biographical Dictionary of later Han to the Three Kingdoms (23-220 AD)

In 1994 my friend and colleague Michael Loewe of Cambridge University suggested that it might be possible for me to prepare a biographical dictionary of Later Han to serve as a companion to the one he was already engaged on for the Former Han dynasty. His compilation was published by in 2000 as A Biographical Dictionary of the Qin, Former Han and Xin Periods (221 BC-AD 24); the present work is intended to serve in tandem.


Following Dr Loewe's plan, it was agreed that the dictionary should attempt to provide an account of every person recorded as having lived under Later Han or, in the case of non-Chinese people, in relationship to its government. The dynastic period is dated from the proclamation of Liu Xuan 劉玄, the Gengshi 更始 Emperor, in 23, and concludes with the abdication of Liu Xie 劉協, Emperor Xian 獻帝, in 220 AD.

There are some problems of demarcation at the beginning and the end of this nominal period. For the first, we adopted the rule of thumb that if a person was identified as a subject of Wang Mang, his/her details would be provided by the earlier volume. If they served or dealt with the restored Han dynasty, they would be included here. There are of course some overlaps, but in general such men as Wang Yi 王邑, who served Wang Mang and died with him in 23, has a biography only in Dr Loewe's compilation, while Guo Qin 郭欽, who surrendered to Han and was enfeoffed by the Gengshi Emperor, appears also here.

The last years of the dynasty present greater problems. Liu Xie held title as emperor from the middle of 189 until his abdication in favour of Cao Pi 曹丕 and the empire of Wei on 11 December 220, but he never controlled the government, and his reign saw thirty years of division and conflict between contending warlords. Much of Sanguo zhi 三國志, "The Record of the Three Kingdoms," deals with events which took place before the end of Han, and Cao Cao 曹操, one of the great figures of that romantic age and the founder of the state of Wei, remained a subject of the dynasty until his death.

It is in these circumstances that the reference to the Three Kingdoms appears in the title of this work, and I have sought to deal with every person recorded as having made his or her mark on history up to 220. Many leading figures of the third century AD are known or can be assumed to have been born before that year, but they are not normally given an entry unless they played a role in events up to that cut-off date.

It may be observed that the number of people dealt with is more than that for Dr Loewe's volume. His contained six thousand entries; this has more than eight thousand. Material for the history of Qin, Han and Xin is largely contained within the standard histories Shi ji 史記 and Han shu 漢書, but Later Han produced a great number of ancillary writings, such as family and local histories, many of which have survived in whole or in part, while the vast majority of stele from the Han period were composed and carved in the second century AD. I discuss this matter further in the section on sources below.


Where possible, information at the beginning of each entry is provided as follows:

Surname xing + personal name ming, style zi [in brackets], years of birth and death (in brackets), followed by commandery unit of registration;

Sample: Ban Gu 班固 [Mengjian 孟堅 (32-92); Youfufeng.

Where a personal name is not known, but the style is, the style is presented in brackets;

Sample: Gongsun [Boda] 公孫伯逹.

Where a surname is unknown or uncertain, or where both the personal name and the style are unknown, this is indicated;

Samples: Guo [personal name unknown]

Hong [surname unknown]

[Yuanbin (style)] 元賓 [surname and personal name unknown].

Variant characters and or readings are indicated and cross-referenced;

Samples: Liu Desheng 劉德升/ [Junsi 君嗣]

Lao Xiu 牢脩 or Lao Shun ; also, probably mistakenly, as Lao Chuan

Ma Fang 馬訪 also as Zhen Fang 甄訪.

Zhen Fang 甄訪 see Ma Fang 馬訪.

Entries are arranged in alphabetical order of the transcription of the name; where the transcription is the same, they are listed in approximately chronological order.

The style served as an intimate personal name, used only by close friends and relatives, or, in somewhat contradictory fashion, as the literary name, appended to published works. It is not always easy to distinguish a style from a two-character personal name, and I have been conservative in identifying styles as such. Notable indicators are the characters bo , zhong , zi and ji , which were commonly awarded to the first, second, third and subsequent sons of a family. 1

In many texts, people are referred to simply by their style [e.g. Guo Tai 郭泰, frequently as Guo Linzong 林宗], and some are best known by their style [e.g. Liu Yan/Yin 劉縯, usually as Liu Bosheng 伯升]. Because of the importance of these secondary personal names, I attach a List of Styles recorded for men and women of Later Han.

All dates are given as AD unless otherwise specified. Dates of birth and death appear at the head of the entry only when they can be reliably placed within a calendar year; otherwise they are indicated in the body of the item.

While recognising that the Chinese year ends in late January or early February of the West, I have followed the convention which expresses the Chinese year by its major Western equivalent; unless, of course, more precise dating is required. Thus the death of Liu Zhi 劉志, Emperor Huan 桓帝 occurred on 25 January 168, in the winter at the very end of the Chinese year 167/168. His successor Liu Hong 劉宏, Emperor Ling 靈帝, ascended the throne on 17 February 168, at the beginning of the following Chinese year; both the end of Emperor Huan's reign and the beginning of Emperor Ling's are generally referred to as taking place in 168. Fortunately, such confusion does not often arise.

Measurements given by Han Chinese units are generally followed by Western metric approximations. Thus Chen Wu is described as seven feet ( chi) seven inches ( zun) tall [178 cm], and grain is measured by dou [2 litres].

The commandery unit identified at the head of an entry is the person's formal place of registration. 2 Some names of commanderies or kingdoms changed during the course of the dynasty; where this may be significant, variants are given with a dividing stroke;

Sample: Cui Lie 崔烈 [Weikao 威考] (d.192); Anping/Boling.

In a few cases, people were able to change their formal place of residence; this is indicated by an arrow-head;

Sample: Zhang Huan 張奐 [Ranming 然明] (104-181); Dunhuang>Hongnong.

Characters for place-names are not normally provided in the body of the entries, but a Table of Provinces and Commandery Units, including characters, is presented at the end of the biographies. Characters for counties and other units are given in the general index.

Present-day place-names are normally rendered in the style of the tenth edition of The Times Atlas of the World.

A survey of The Administrative Structure of Later Han provides characters for the most important offices and titles; these and some less common ones are also listed in the Index.

Though I have largely followed the system devised by Professor HH Dubs and developed by Professor Hans Bielenstein, I have made amendments, chiefly for the sake of brevity. For example, I render the title taishou 太守, which Dubs and Bielenstein have as "Grand Administrator," simply as "Administrator"; the change from shou to taishou was significant in Qin and Former Han, but taishou was used throughout Later Han. This and other changes, notably the rendering of ministerial titles, are explained in the survey, and I append a table of comparative renderings.

At the end of most entries there are one or more bibliographical citations. These are not intended to be full or in any way comprehensive, but rather provide a first point of reference. When a person has a substantial biography in a Hou Han shu or Sanguo zhi, this is indicated by an asterisk*.

Most bibliographical references relate to early texts but in some cases, where the argument is complex but convincing, I may cite only a modern scholar; readers must examine the evidence that he or she puts forward and decide for themselves.

The rulers of Later Han are best known to history by their posthumous titles, so that the founding sovereign Liu Xiu is commonly referred to as Emperor Guangwu. I usually follow that convention, but entries for each ruler appear under their surname and personal name, with cross-references from the imperial title.


Most of our information on Later Han comes from the standard history Hou Han shu 後漢書, whose annals and biographies were compiled by Fan Ye 范曄 in the early fifth century. The Treatises, which had been composed separately, as part of the Xu Han shu 續漢書 of Sima Biao 司馬彪 in the third century, were later combined with the work of Fan Ye to form the official history, whose first formal edition was prepared in 1022. 3

There were a number of other histories of the Later Han, some of which have completely disappeared, while others survive in part or in fragmentary quotation. The most important are Dongguan Han ji 東觀漢記, compiled as an official record at intervals throughout the dynasty, the Hou Han ji 後漢紀 by Yuan Hong 袁宏 of the fourth century, and the Hou Han shu of Xie Cheng 謝承 of the early third century. Some parts of Xie Cheng's work, and fragments from six others, are collected in the Qijia Hou Han shu 七家後漢書. Twenty-four chapters of Dongguan Hanji survive from an original 143, while the Hou Han ji of Yuan Hong appears largely complete.

Sanguozhi 三國志, compiled by Chen Shou 陳壽 of the third century, was supplemented by the commentary of Pei Songzhi 裴松之, who presented the full work to Liu Yu 劉裕, first emperor of the Song dynasty, in 420. Chen Shou describes the history of the three rival states, Wei, Shu-Han 蜀漢 and Wu 吳, beginning with their founders, Cao Cao, Liu Bei 劉備, and Sun Jian 孫堅 and his son Sun Ce 孫策, and including other leading figures such as the usurper Dong Zhuo 董卓 and the warlords Yuan Shao 袁紹, Liu Biao 劉表 and Liu Zhang 劉璋. The work is formally composed entirely of biographies, but those of the various rulers can take the form of annals.

Pei Songzhi's commentary is more than a set of annotations to the text compiled by Chen Shou, for he put together many extracts from other works to supplement and sometimes contradict Chen Shou's account of events. His eclectic selection included regular histories, family chronicles and, perhaps most notably, local records. Many of these are preserved only in the fragments quoted by Pei Songzhi, but they demonstrate the wide spectrum of literary and historical material which was available during Later Han and in the first centuries which followed. 4

Some works survive in more substantial form. Notable are the Fengsu tongyi 風俗通義 of Ying Shao 應劭, from the end of Later Han, and Huayang guo zhi 華陽國志 by Chang Qu 常璩, a history and gazetteer for present-day Sichuan composed in the mid-fourth century, which is largely intact. The Shuijing zhu 水經注 of Li Daoyuan 酈道元, from the early sixth century, after the time of Pei Songzhi, also preserves geographical and historical information in its descriptions of the courses of the rivers of China. Sources reconstructed from quotations are listed in the bibliography of Early Works Cited, while other titles mentioned in various entries are listed among Literary and Scholarly Works of Later Han.

One feature of the world of Later Han and the centuries which followed is the strong interest in local history and family lineage. This was reflected in works such as Huayang guo zhi, and in a whole genre of writing on worthy men of particular regions, such as Sanfu juelu 三輔決錄 "Evaluative Records of the Three Adjuncts" by Zhao Qi 趙岐/, dealing with gentlemen from the commanderies about the ancient capital of Chang'an 長安, and similar compilations with titles like Qijiu zhuan 耆舊傳 "Venerable Men and Ancient Affairs" or xianxian zhuan 先賢傳 "Accounts of Worthy Men of the Past" for Chenliu 陳留, Shanyang 山陽, Runan 汝南 and Yingchuan 潁川 commanderies and for other territories. 5

In the same way, biographies in Sanguo zhi and its commentary, and in later histories such as Jin shu 晉書, seek to give an account of the ancestry of each individual. Some are exaggerated and unreliable, but a few lineages were recorded through centuries, while many men who might otherwise have remained unknown are remembered through the achievements of their more distinguished descendants. 6


The same sense of local interest and family pride displayed in books and essays in the last years of Later Han was also shown by the erection of stele, whether in commemoration of public works or to honour an individual, family or clan. In her discussion of Later Han stone inscriptions, Ebrey 80 estimated that there were rather more than three hundred such texts still extant; and others have been discovered in the past twenty-five years. Ebrey noted the use made of them by Yen Keng-wang in his study of local government, by Lao Kan in his work on roads and bridges, and by Yang Shu-ta for information on rituals of marriage and death. As she observed, moreover, inscriptions can be of great value to social historians, for the information they provide on individuals of lower rank who seldom appear in official histories, and for the ordering of their lists of sponsors, which indicate the structures of power and responsibility in local communities.

Stele inscriptions survive in two forms. Some stones are physically extant, and a number have been excavated or otherwise discovered in modern times, but the majority are preserved as copies, notably through the work of the Song dynasty scholar Hong Kua/Guo 洪适, whose Li shi 隸釋 and its continuation Li xu 隸續 quoted or recorded 185 such texts, together with his own commentary. A further twenty-six appear in the collected works of Cai Yong 蔡中郞文集, and other compilations which I have referred to, both ancient and modern, are cited in Part III of the bibliography of Early Works Cited.

In dealing with this material, I generally provide entries for individuals mentioned in the body of the texts, though in some cases the transcription is too uncertain to be used. I do not, however, include sponsors of a commemorative stele unless they are mentioned by other sources. Though a full analysis of all the names has yet to be compiled in a Western language, the present work is not the place for it: the extra number of minor entries would add considerably to its length, and such a basic list of names and incidental functions is better kept for a separate study. 7 Similarly, I do not provide entries for the minor officials, soldiers, citizens and convicts whose names appear in the corpus of wooden and bamboo strips and fragments found in the north-western deserts of Gansu and Xinjiang: on these, see notably Loewe 67.

One might imagine that stele, as contemporary and apparently permanent documents, would be more reliable than other texts, written on such ephemeral substances as paper, and subject to the vagaries of copyists. This, however, is not always true. Firstly, as we have observed, a great number of inscriptions survive only because they were copied later, while in some cases the stones themselves are of doubtful authenticity: forgery can be profitable and is not uncommon. So provenance and transmission are sometimes uncertain, while many characters are missing or can be misread, and there are specialised formulae. Stele can thus be difficult to decipher and their interpretation is often unclear. 8

Besides this, one must be aware that the interest and pride in local and family affairs which was taken by the people of Later Han not only produced increasing numbers of stele, but also encouraged exaggerated claims for the ancestry and personal achievements of the men and women who were honoured by such commemorative texts. As Ebrey 80 and Bielenstein 81 have remarked, though inscriptions are generally contemporary, and their information may enhance or correct that which is to be found in the histories, they are frequently eulogistic and omit facts which may reflect badly upon their subject. 9 Stele are important, but some of their composers had a special agenda.


We have already considered the problems of stele inscriptions, which can be affected by copyists'errors, by the difficulty of identifying characters, and by the possibility of forgery. The same considerations, of course, apply to early books: an enormous quantity has been lost; there are many passages, even in the works which survive, which present difficulties of interpretation; and very often one must suspect that the text has been corrupted. Huayang guo zhi, for example, is a well-designed history of present-day Sichuan, with a quantity of detail, albeit sometimes anecdotal or apparently unreliable. Its transmission, however, is far less secure than that of standard histories such as Hou Han shu and Sanguo zhi, and there are many places where names and dates are confused. 10

Similar problems are found in such texts as Shuijing zhu and the Fengsu tongyi of Ying Shao: the text of the latter is often fragmentary and contains many doubtful readings, so it is not always possible to comprehend the significance of an anecdote or incident. And the same may be said of the surviving portions of Dongguan Hanji, the Hou Han shu of Xie Cheng, and the Hou Han ji of Yuan Hong, though the difficulties are seldom so serious as for Huayang guo zhi and Fengsu tongyi.

It is encouraging, on the other hand, to find that the official standard histories, Han shu, Hou Han shu and Sanguo zhi with the Pei Songzhi commentary, present substantially fewer areas of uncertainty. Despite the prejudices of the authors, sometimes evident and sometimes only discovered by research and analysis, their works have benefited from generations of careful scholarship, and modern commentaries, such as those compiled by Wang Xianqian for Han shu and Hou Han shu, and by Lu Bi for Sanguo zhi, ensure that we can have a high level of confidence in the reliability of the texts now before us.

Even when we can be reasonably confident that the material is original, however, there still remain questions on its accuracy and interpretation. This is a problem for students of any period and place, and scholars of Han are well aware of the conscious or unconscious biases of the compilers of the material with which they deal. Early Chinese historians, reflecting Confucian tradition, were concerned to report events honestly, but they also regarded history as a matter of morality. This could affect their understanding and rendering of the truth, while the social and philosophical background common to such literate gentlemen meant that their view of the world reflected that of the landed gentry and traditional leaders of the community. At the end of his admirable essay on the historiography of Hou Han shu, Bielenstein concludes that, while the work is not specifically partial to the gentry or the imperial clan, it is prejudiced against those who placed themselves outside what were considered to be orderly relations, thus opposing the Mandate of Heaven. In similar fashion, Mather comments on the manner in which Yuan Hong, author of Hou Han ji, expressed therein his quasi-naturalistic theory of history, stressing the importance of the relationship between ruler and subject and between father and son. 11 Given such opinions amongst those who recorded events, it is not surprising if eunuchs, and others who were regarded as usurping power, have a very bad press, 12 while those who sought the reform of government on Confucian lines, or who were considered to have behaved in the manner of true gentlemen, were reported with tolerance and approval. 13

There remains, of course, the further question of the authors'own perceptions of reality: What events that are described really happened? What stories have been invented to demonstrate moral qualities? And which accounts of the magical, mystical or supernatural were given credence?

It is not inappropriate that Fan Ye compiled his Hou Han shu at the same time as Pei Songzhi prepared his commentary to Sanguo zhi, and while Liu Yiqing was working on Shishuo xinyu 世說新語 "A New Account of Tales of the World." This last is a collection of anecdotes relating to historical figures, primarily concerned to develop an impression of character and circumstance, and to entertain their readers; as Mather observes, the collection "was, at least in part, a fictionalization rather than sober history" 14 and some anecdotes are better seen as analogies, not necessarily factual. 15 Shishuo xinyu was primarily a work of entertainment, and its lineal descendants are the fictional stories and plays of later centuries, culminating, at least as regards the end of Han, in the Romance of the Three Kingdoms 三國演儀, attributed to Luo Guanzhong 羅貫中 of the late fourteenth century. 16 I seldom cite Shishuo xinyu as a source, but I do provide references to Mather's summary biographies, and these in turn list the items in Liu Yiqing's compilation.

The blurred spectrum from history to anecdote to fiction may likewise be found in Shi ji and Han shu, and certainly within Hou Han shu and Sanguo zhi. Many of the works quoted by Pei Songzhi are accounts of marvels or experiences of the supernatural, such as the Soushen ji 搜神記 "Record of Enquiry about the Spirits" by Gan Bao 干寶, while others provide anecdotes of doubtful authenticity. For its part, Hou Han shu has chapters on the lives of magicians and diviners, full of mysterious and supernatural achievements, and a number of pure ghost stories. 17

Given these conditions, a collection of biographies must include items and incidents of uncertain factual value. My chief purpose, like that of Dr Loewe, is to provide a survey of the information which may be found in the sources, not specifically to compose a series of critical biographies. For the most part, therefore, I accept information at face value, with occasional caveats ("we are told that") and some discussion of contradictions and variant accounts. Any scholar, however, must hold to the motto Caveat Lector.

With all these precautions, I present below A Short History of the Later Han Dynasty, accompanied by a Chronology, to discuss political, economic and social developments of the period, and to provide a historical context to the biographies.


This project has required analysis of a great variety of topics, and I have been fortunate to receive advice and support from many different scholars. In particular, I thank Dr Michael Loewe of Cambridge University, Professor Amanda Brown of the University of Michigan at Ann Arbor, and Professor Liu Ts'un-yan of the Australian National University, for their regular guidance and assistance. Other colleagues whom I have consulted with great profit include Michael Nylan, William H Nienhauser, Robert Joe Cutter, William G Crowell, Burchard Mansvelt Beck, Hans van Ess, Patricia Ebrey, Margaret Pearson, Ch'en Ch'i-yün, J Michael Farmer, Nicola di Cosmo and Igor de Rachewiltz.

I am grateful to the Chiang Ching-kuo Foundation for supporting the editorial work required to make the manuscript suitable for publication, and for the research and editorial assistance of Elizabeth Kat. Patricia Radder and Renee Otto of Brill have been most helpful editors, and have arranged for the maps to be drawn by Mrs F E Derksen-Janssens. My special thanks go to Karina Pelling, for her energy and patience in preparing the final manuscript, and to Greg Young, who has given constant support through his knowledge of Chinese and his skill in computing.

1 The rule is by no means perfect: Cao Pi and Cao Zhi 曹植, for example, sons of Cao Cao, both had the character zi in their styles: Cao Pi had the style Zihuan 子桓, and Cao Zhi's was Zijian 子建.

2 Some leading families, notably those connected to the throne such as the Dou or the Liang , while formally registered in an outlying region, were in practice resident at the capital.

3 On the text history of Hou Han shu and parallel early works, see Bn 54:9-20. On the Treatises of Sima Biao, see MBeck 90.

4 On the historiography of Sanghuo zhi and the commentary of Pei Songzhi, see deC 90:533-589.

5 On this style of local history and biography, see deC 96:563-565. 

6 The best-known example is probably that of the Cui family of Anping 安平, discussed by Ebrey 78, but the Lu family in the southeast held substantial position from the time of Former Han through to the sixth century: deC 96:502-503.

The introductions to the genealogical tables of Xin Tang shu [XTS 71A-75B] summarise the claims of many leading families over several hundred years; in his commentary to HHS, Hui Dong cites a number of these references under the guise Shi xi 世系, and Ebrey 78:157-173 discusses the value and reliability of the tables, noting that they were evidently compiled from several different types of source, including private genealogies and more public inscriptions. Some introductions are summary, others, evidently based upon family records, are voluminous. Compiled many centuries later, however, they can naturally be confused and/or exaggerated, and there are many obvious errors and inconsistencies.

For the present volume, I have sought to identify and provide entries for people whom the Tang and Song historians recorded as having been active during Later Han. One must, however, be cautious in accepting any entry which relies solely upon a reference to XTS: in some cases I indicate specific doubts, and I may not provide entries for all claimed ancestors. For examples, see the entry for Xiao Wangzhi 蕭望之 and the discussions of the family of Xin Bozhen 辛伯? [XTS 73A:2880] and of the lineage ascribed to Jia Xu 賈詡 [XTS 75B:3387-88].

7 The Shike tiba suoyin 石刻題索引 of Yang Dianxun 楊殿珣, in Shike shiliao congshu xinbian 石刻史料叢書新编, provides a general index to stele, and Professor Miranda Brown of the University of Michigan at Ann Arbor is currently studying this great quantity of material; I am extremely grateful for her advice, guidance and most generous help to my work on this part of the project.

8 See Ebrey 80:330, quoting the twentieth-century scholar Huang Kung-shu 黃公渚. For examples of stele whose authenticity is uncertain or whose content is questionable, see the entries for Fei Zhi, Tangxi Dian and Xia Cheng.

9 A number of examples may be observed from a comparison of the biography of the Excellency Qiao Xuan 橋玄 of Liang in HHS 51/41:1695-97 and the fractional account in Xie Cheng 2:15b with the inscription texts compiled by Cai Yong [Cai 1:1-4]: see sub Qiao Ji 橋基 and Bn 81:578-581. On Hu Guang also, see Bn 81:579-580.

After Cai Yong composed the funerary inscription for Guo Tai 郭泰 in 169, he is said to have remarked that it was the only eulogy he had been able to write without any sense of embarrassment: HHS 68/58:2227.

10 The variorum edition of Ren Naiqiang 任乃強, Huayang guo zhi jiaobu tuzhi 校補圖志, indicates how many textual variants, and how many errors, may be found in the present-day work. Where HYGZ is clearly mistaken, I have not indexed or itemised the variants.

11 Bn 54:20-81 at 81; Mather 76:xxi. We may note that Ying Shao used his Fengsu tongyi to give moral commentary; and the relevant chapters of the Zizhi tongjian of the great Song historian Sima Guang [ZZTJ 54-69], while providing an excellent account of the fall of Han, may also be read as a lesson on government morality and personal conduct: deC 96:xiii-xiv.

12 HHS 34/24, for example, contrasts the restraint of the worthy father Liang Shang with the excesses of his son Liang Ji, and similarly distinguishes between the conduct of Liang Na 梁妠, consort of Emperor Shun, and that of her younger sister Nüying 梁女塋, first empress of Emperor Huan.

13 The brutal conduct of Su Buwei 蘇不韋 was widely approved, and there was general acceptance of the principle of vendetta, even when carried out by third parties: e.g. Zhi Yun 郅惲, Dian Wei 典韋, and the Lady Zhao E 趙娥.

14 Mather 76:xv.

15 A notable example is in SSXY xxvii.1; Mather 76:441. It tells how Cao Cao and Yuan Shao, as young men about town, raided a wedding. Cao Cao took the bride, but turned pursuers to chase Yuan Shao. This may perhaps relate to a true incident, and such conduct was surely not uncommon, but it is best understood as an analogy for the manner in which Cao Cao, many years later, took possession of the emperor and then blamed Yuan Shao for attacking him.

16 The Romance cannot be taken as a record of fact. It is a historical fiction, based upon the interests of Southern Song, with strong prejudice in favour of Liu Bei and his state of Shu-Han. See, for example, deC 90:578-589.

17 HHS 82/72A and B have biographies of the Fangshu 方術, while among the men of Exceptional Conduct 獨行 in HHS 81/71, Wang Chun 王忳 had close encounters with the supernatural.

Rafe de Crespigny